Drug Discovery Strategies and Methods

Transcription

DRUG
DISCOVERY
STRATEGIES
METHODS
EDITED BY
ALEXANDROSMAKRIYANNIS
DIANEB~EGEL
Center for Drug Discovery
University of Connecticut
Storrs, Connecticut, U.S.A.
MAR
.....
C_ E L~
_
MARCEL
DEKKER,
INC.
DEKKER
NEWYORK . BASEL
Although great care has been taken to provide accurate and current information, neither the
author(s) nor the publisher, nor anyone else associated with this publication, shall be liable for
any loss, damage, or liability directly or indirectly caused or alleged to be caused by this book.
The material contained herein is not intended to provide specific advice or recommendations
for any specific situation.
Trademark notice: Product or corporate names may be trademarks or registered trademarks
and are used only for identification and explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress.
ISBN: 0-8247-0691-9
This book is printed on acid-free paper.
Headquarters
Marcel Dekker, Inc.
270 Madison Avenue, New York, NY 10016
tel: 212-696-9000; fax: 212-685-4540
Eastern Hemisphere Distribution
Marcel Dekker AG
Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland
tel: 41-61-260-6300; fax: 41-61-260-6333
World Wide Web
http://www.dekker.com
The publisher offers discounts on this book when ordered in bulk quantities. For more
information, write to Special Sales/Professional Marketing at the headquarters address
above.
Copyright n
n 2004 by Marcel Dekker, Inc. All Rights Reserved.
Neither this book nor any part may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, microfilming, and recording, or by
any information storage and retrieval system, without permission in writing from the
publisher.
Current printing (last digit):
10 9 8 7 6 5 4 3 2 1
PRINTED IN THE UNITED STATES OF AMERICA
Preface
Drug research encompasses diverse branches of science united by a common goal, namely, developing novel therapeutic agents and understanding
their molecular mechanisms of action. This process is a lengthy, exacting,
and expensive undertaking that involves integration of data from different
fields and culminates in the final product—a new drug in the marketplace.
In the past decade, progress in drug research has flourished because of
major contributions from a variety of disciplines.
The material presented in this volume focuses on a number of
research topics that have provided critical information in the field of drug
discovery. Several chapters present techniques that extend our understanding of the three-dimensional structure of macromolecules, principally
proteins, but also nucleic acid polymers and organized lipid and carbohydrate assemblies. As greater structural data on the these molecules become
available, information can be obtained on their interactions with small
endogenous ligand drug molecules as well as on the interactions between
two or more of these biopolymers. Such knowledge enhances our overall
understanding of the biochemical systems of interest and their relevance
for therapeutic discovery. In addition to the basic knowledge gained by
such research, the data provide a solid basis for the development of novel
drugs with greater potencies, higher specificities of action, and reduced side
effects.
Another area of research covered in this book is the in vivo
anatomical localization of potential therapeutics using PET and SPECT
analysis (Chapter 5). These techniques allow researchers to pinpoint the
localization of high-affinity ligands in the living organism with high
accuracy, thus giving researchers a window on the functions of the brain
and other organs and on the sites of action of potential therapeutic agents.
Such studies will provide a blueprint for the design of pharmacological
iii
iv
Preface
agents that will target specific regions of affected organs and deliver therapeutic actions rapidly and with high specificity.
High throughput methods have increased our capacity for appropriate candidate compounds selection and also for developing libraries of
novel compounds from which such candidates can be selected. Chapter 7
discusses the use of solid-phase synthesis for the high throughput production of peptides and other small molecules. In addition, as discussed in
Chapter 6 on peptidomimetics, the swift production of novel leads holds
considerable promise for future discovery of novel therapeutic agents.
The investigation of therapeutic targets for cannabinoid sites of action has already generated considerable interest within the field of drug
discovery, and Chapter 4, which details the results of such studies, highlights the importance of target-based studies. The enhanced appreciation
of the role of stereochemistry in drug action has focused efforts on understanding the conformation of drugs as they bind to their target receptor.
Studies of the diverse effects of cannabinoids and the development of
compounds that employ the information gleaned from the ligand/receptor
data should provide substantial insight into their molecular mechanisms of
action. Future research will promote the development of drugs that are
capable of higher specificity. longer half-lives, and lessened toxicity. In
studies of potential antiviral therapies, the understanding of viral target
molecules is essential for the production of effective medications that interact specifically in the viral life cycle and gene products, which will result
in lowering drug toxicity to the host and enhancing the antiviral activity of
the pharmacotherapy. As the nature of viral infectivity, cell growth, death,
and receptor biology are elucidated, the methods and paradigms for development of highly specific medications will provide superior treatments
for a number of diseases that pose a terrible burden worldwide (Chapters 10
and 11).
From the fields of proteomics and genomics that hold significant
promise for unique medications, several areas of biology have also found
applications in the drug discovery arena. The study of regulatory molecules
and oncogenes has opened new avenues in drug therapy, as discussed in
Chapter 8 on G-protein-coupled receptors and Chapter 2 on SRC homology domains. Research on protein misfolding (Chapter 9), which has been
implicated in neurodegenerative diseases, has highlighted the need to
enhance our understanding of structural alterations in normal proteins
products. Chapter 1 details the development of such research, and asserts
that only as we understand the basic physical mechanisms of such alterations can new therapeutic regimens be proposed and tested.
Preface
v
The topics included in this volume are not intended to be allinclusive. Our approach has been eclectic, in an effort to bring the reader
the most exciting aspects of drug discovery, along with the methods that
show the most promise in enhancing the discovery process.
The chapters presented in this book have been contributed by specialists in their areas of research and will provide a contemporary picture of
the overall field of drug discovery to scientists from diverse disciplines.
Alexandros Makriyannis
Diane Biegel
Contents
Preface
Contributors
1. Protein Crystallography in Structure-Based Drug Design
Xiayang Qiu and Sherin S. Abdel-Meguid
2. Src Homology-2 Domains and Structure-Based, SmallMolecule Library Approaches to Drug Discovery
Chester A. Metcalf III and Tomi Sawyer
3. Three-Dimensional Structure of the Inhibited
Catalytic Domain of Human Stromelysin-1 by
Heteronuclear NMR Spectroscopy
Paul R. Gooley
4. Cannabinergics: Old and New Possibilities
Andreas Goutopoulos and Alexandros Makriyannis
5. Development of PET and SPECT Radioligands for
Cannabinoid Receptors
S. John Gatley, Andrew N. Gifford, Yu-Shin Ding,
Ruoxi Lan, Qian Liu, Nora D. Volkow, and
iii
ix
1
23
61
89
129
vii
viii
6. Structural and Pharmacological Aspects of
Peptidomimetics
Peter W. Schiller, Grazyna Weltrowska, Ralf Schmidt,
Thi M.-D. Nguyen, Irena Berezowska, Carole Lemieux,
Ngoc Nga Chung, Katharine A. Carpenter,
and Brian C. Wilkes
7. Linkers and Resins for Solid-Phase Synthesis: 1997-1999
Pan Li, Elaine K. Kolaczkowski, and Steven A. Kates
8. Allosteric Modulation of G-Protein-Coupled
Receptors: Implications for Drug Action
Angeliki P. Kourounakis, Pieter van der Klein,
and Ad P. I. IJzerman
Contents
147
175
221
9. Protein Misfolding and Neurodegenerative Disease:
Therapeutic Opportunities
Harry LeVine III
245
10. Uncoating and Adsorption Inhibitors of Rhinovirus
Replication
Guy D. Diana and Adi Treasurywala
279
11. Profiles of Prototype Antiviral Agents Interfering
with the Initial Stages of HIV Infection
E. De Clercq
309
Index
337
Contributors
Sherin S. Abdel-Meguid Suntory Pharmaceutical Research Laboratories,
Cambridge, Massachussets, U.S.A.
Irena Berezowska
Quebec, Canada
Clinical Research Institute of Montreal, Montreal,
Katharine A. Carpenter
treal, Quebec, Canada
Ngoc Nga Chung
Quebec, Canada
Clinical Research Institute of Montreal, Mon-
Eric De Clercq Rega Institute for Medical Research, Katholieke Universiteit Leuven, Leuven, Belgium
Guy D. Diana ViroPharma, Inc. Exton, Pennsylvania, U.S.A.
Yu-Shin Ding
U.S.A.
Brookhaven National Laboratory, Upton, New York,
S. John Gatley
U.S.A.
Brookhaven National Laboratory, Upton, New York,
Andrew N. Gifford
York, U.S.A.
Paul R. Gooley
Brookhaven National Laboratory, Upton, New
University of Melbourne, Parkville, Victoria, Australia
ix
x
Contributors
Andreas Goutopoulos Serono Reproductive Biology Institute, Rockland,
Massachusetts, U.S.A.
Ad P. IJzerman
Leiden University, Leiden, The Netherlands
Steven A. Kates
Surface Logix, Inc., Brighton, Massachusetts, U.S.A.
Elaine K. Kolaczkowski
chussetts, U.S.A.
Vertex Pharmaceuticals, Cambridge, Massa-
Angeliki P. Kouranakis
Greece
University of Thessaloniki, Thessaloniki,
Ruoxi Lan
University of Connecticut, Storrs, Connecticut, U.S.A.
Carole Lemieux
Quebec, Canada
Harry LeVine III
University of Kentucky, Lexington, Kentucky, U.S.A.
Pan Li Vertex Pharmaceuticals, Cambridge, Massachusettes, U.S.A.
Qian Liu
U.S.A.
Chester A. Metcalf III
sachusetts, U.S.A.
Thi M.-D. Nguyen
Quebec, Canada
University of Connecticut, Storrs, Connecticut,
ARIAD Pharmaceuticals, Inc., Cambridge, Mas-
Xiayang Qiu SmithKline Beecham Pharmaceuticals, King of Prussia,
Pennsylvania, U.S.A.
Tomi Sawyer
setts, U.S.A.
ARIAD Pharmaceuticals, Inc., Cambridge, Massachu-
Contributors
xi
Peter W. Schiller
Quebec, Canada
Ralf Schmidt
Canada
Clinical Research Institute of Montreal, Montreal, Quebec,
Pfizer Central Research, Groton, Connecticut, U.S.A.
Adi Treasurywala
Pieter van der Klein
Nora D. Volkow
NIDA, Bethesda, Maryland, U.S.A.
Grazyna Weltrowska
Quebec, Canada
Brian C. Wilkes
Quebec, Canada
1
Protein Crystallography in
Structure-Based Drug Design
Xiayang Qiu
SmithKline Beecham Pharmaceuticals, King of Prussia,
Pennsylvania, U.S.A.
Sherin S. Abdel-Meguid
Suntory Pharmaceutical Research Laboratories, Cambridge,
Massachusetts, U.S.A.
I.
INTRODUCTION
Proteins are responsible for a wide variety of important biological functions in living organisms and are commonly used as targets of therapeutic
agents. A unique primary and tertiary structure is a hallmark property of a
protein. Although several related and even unrelated proteins may share
the same overall tertiary structure or fold, each will differ from the others in
the details. Knowledge of the detailed atomic three-dimensional structure
of the protein and/or its ligand complexes should facilitate the design of
novel, high affinity ligands that interact with that protein. The process of
elucidating the atomic structure of proteins and their complexes, and the
design of novel, therapeutically relevant ligands based on these structure
elucidations, is known as structure-based drug design.
Proteins are complex molecules, typically containing several thousand atoms. Although Pauling and Corey proposed the a helix and the h
sheet as the main secondary structural elements of proteins in 1951, and the
crystal structure of myoglobin was reported by John Kendrew in 1958,
1
2
Qiu and Abdel-Meguid
crystal structure determination in the early days were hampered by
numerous technical limitations and usually required many years of hard
work. By the mid-1980s, substantial improvements in data acquisition
software and hardware had considerably accelerated the speed with which
a crystal structure could be determined. Trying to capitalize on the
potential of structure-based drug design, several pharmaceutical companies built their own protein crystallography laboratories, and a number of
structure-based drug design efforts emerged in industrial and academic
laboratories [1].
In the past 10 years, we have experienced a sudden burst in the
number of protein three-dimensional structures determined. By the end of
the twentieth century, merely 40 years since the first protein structure was
solved, there were over 11,000 structures deposited in the Protein Data
Bank (PDB). Although each entry is not a unique protein, the number of
novel structures deposited in the PDB has increased sharply during the last
decade. These proteins include not only soluble proteins, but also a number
of membrane proteins. Furthermore, structures of protein–protein and
protein–nucleic acid complexes, viruses, and the ribosome are also available. This marvelous scientific achievement was mostly credited to the
method of single-crystal x-ray diffraction (protein crystallography),
although a notable number of structures were determined by means of
NMR spectroscopy. Many factors in addition to the incredible advances in
computer hardware and software contributed to the improved efficiency
and precision in protein crystallography: the advent of molecular biology,
which allows for cloning, mutation, and overexpression of many targets
that are difficult to isolate from natural sources; advances in protein
purification that facilitate the production of large amounts of highly
purified proteins; improvement in protein characterization and crystallization strategies; enhancement of data acquisition techniques and equipments; access to powerful synchrotron radiation sources; and introduction
of the selenomethionine multiple-wavelength anomalous diffraction
(MAD) procedure for phase determination. Currently, almost all large
pharmaceutical and numerous biotechnology companies have established
in-house macromolecular crystallography units, and the crystallographic
community is solving thousands of new structures every year. With
structural information becoming more readily available, structure-based
drug design has become an integral part of the modern drug discovery
process and has begun to contribute to a significant portion of the current
drug discovery portfolio.
Protein Crystallography
3
Identifying and bringing a successful small-molecule drug to the
market requires considerable effort, which typically costs millions of
dollars and may span as much as10 years. With this time scale in mind,
one must realize that structure-based drug design is still in its infancy,
having started in earnest in the mid-1980s. While the concept of such a
rational approach has been around for some time, for much of the work
in the field it is still too early to demonstrate market successes. Moreover, although structural knowledge may be used for lead generation
and lead optimization, or even for addressing some developability issues,
it does little to address other important issues in drug development
ranging from the appropriateness of targets or disease models to government regulatory issues or changing market forces. In fact, drug
discovery is a risky business in that only a very small number of compounds are able to find their way to the market. Therefore, the successful structure-based design and the launch of inhibitors of HIV protease
[2] and influenza virus neuraminidase [3] as drugs are particularly
encouraging events for the field [4–9]. In this chapter, we will introduce
the technique of protein crystallography and its use in structure-based
drug design, point out the technical challenges ahead of us, and report
many practical lessons learned during the past decade of structure-based
drug design.
II. THE DRUG DISCOVERY PROCESS
The many steps of the complex and multidisciplinary drug discovery
process can be grouped into four major phases: target identification and
validation, lead identification, lead optimization, and biological testing
(Fig. 1). Choosing an appropriate target is usually the first step in the drug
discovery process. Target selection requires an understanding of human
diseases and the biological processes that lead to a particular disease.
Although historically drugs (e.g., h-lactams) were discovered without
knowledge of their molecular target, knowing the target greatly enhances
one’s ability to discover novel drugs in a timely fashion. Recent advances in
sequencing the human genome, as well as the genomes of many human
pathogens, have provided a large pool of potential novel molecular targets.
Most future drug discovery efforts will start with a relatively unknown gene
selected from a sequence database based on one or more attractive features
that could provide a hint of its function, such as tissue distribution, genome
4
Figure 1
Simplified drug discovery process.
localization, and/or sequence homology or structural analogy to a known
protein. Cloning, expression, purification, and characterization of the
protein target and other tool reagents such as antibodies or receptors will
usually follow, to be used in target validation with a set of appropriate
genetic and biological assays.
The second step is to identify a suitable lead molecule to interact
with the molecular target. This is usually achieved through high throughput screening of available chemical compound libraries and natural
products, typically containing hundreds of thousands of compounds.
Although the size of the library per se is not critical, a library that contains
a large number of molecules is essential to assure molecular diversity.
Novel lead molecules can also be designed by analysis of the threedimensional structure of the target molecule in a process known as de
novo design. A desirable lead should usually have at least low micromolar
binding potency against the target and should be amenable to further
synthetic manipulations.
The third step is to optimize the lead molecule through iterative
chemical synthesis and biological testing, aiming to obtain molecules with
the required potency (typically nanomolar), selectivity, bioavailability,
and DMPK (drug metabolism and pharmacokinetics) properties. This step
usually requires considerable time and resources; usually the synthesis of
hundreds of compounds is needed to deduce a robust SAR (structure–
activity relationship). Such resources can be considerably reduced and the
5
time significantly shortened if optimization employs knowledge from the
three-dimensional structure of complexes of leads with the target.
The last step of the drug discovery process involves the testing of lead
compounds to address issues such as efficacy, bioavailability, and safety.
Testing may include in vitro assays but ultimately would require a suitable
disease model and studies in animals. Many compounds may need to be
designed and synthesized to identify the one compound with all the desired
properties. Such a compound can be advanced to preclinical studies and
eventually to the clinic.
III. THE STRUCTURE-BASED DRUG DESIGN CYCLE
Timely optimization of lead compounds requires knowledge of the threedimensional structure of target–ligand complexes. Protein crystallography
has been the predominant technique used to elucidate the three-dimensional structure of proteins in structure-based drug design. Crystallographic studies usually consume tens of milligrams of pure protein and
take several months to yield the first crystal structure. Therefore, one
should start crystallographic efforts as soon as suitable material is available, preferably even before initiation of high throughput screening. Once
a lead has been identified through high throughput screening or de novo
design, structure determinations of target–ligand complexes should be
pursued. The use of information derived from the structure determination
of the target bound to the initial lead molecule should allow for the design
and synthesis of new ligands with improved properties, as well as the
initiation of further rounds of structure-based design. Through iterations
of structure determination, design, synthesis, and biological testing (Fig. 2)
a drug candidate should emerge.
In addition to lead optimization and lead identification, three-dimensional structures of the target–ligand complexes can contribute to the
traditional drug discovery process in other ways. For example, structural
information combined with genomic sequences may aid in target identification by helping to classify genes with unknown functions. Structures
can be used as templates for de novo design or in silico lead identification
through screening of virtual libraries. Structural information can provide a
basis for the design of directed combinatorial libraries [10,11]. Moreover,
structural studies of leads with serum albumin and various cytochrome
P450s should allow for a better understanding of some of the developability issues that may arise during drug development.
6
Figure 2
Structure-based design cycle.
IV. PROTEIN CRYSTALLOGRAPHY
For most noncrystallographers, protein crystallography tends be a black
box full of jargon. Here, we give a brief overview of the technology in an
attempt to demystify some of the terms used.
A. Crystallization
Obtaining large single crystals that diffract to high resolution remains the
primary bottleneck of protein crystallography. The most widely used
7
crystallization method is the hanging-drop method of vapor diffusion
(Fig. 3), in which a drop (1 or 2 AL) of protein is mixed with an equal
volume of a precipitant on a glass coverslip and is sealed over a well
containing the same precipitant added to the protein. Many factors are
known to be important in protein crystallization: protein purity (preferably >95% pure) and concentration (typically 10 mg/mL), the nature of
precipitant [e.g., poly(ethylene glycol) or various salts] and its concentration, the nature, concentration, and pH of the buffer, the presence or
absence of additives (e.g., metal ions, reducing agents, protease inhibitors,
metal chelators, detergents) and effectors (e.g., ligands, cofactors, substrates, inhibitors), the rate of equilibrium between the protein and the
precipitant, the crystallization temperature, and so on. Since there are no
general rules to correlate all these factors to the eventual success in
obtaining crystals, protein crystallization remains a trial-and-error process
and a significant bottleneck in protein crystallography: failure rate is
typically 50% even with thousands of crystallization trials. Many methods
and techniques have been employed to enhance one’s ability to obtain
protein crystals. Molecular biology and biochemical methods have been
utilized to generate domains of large proteins that may be less flexible and
thus more amenable to crystallization. Biophysical tools such as dynamic
light scattering [12] and ultracentrifugation [13] have been used to study
protein aggregation in solution. Molecular biology has been employed to
generate mutants that do not aggregate or are more soluble. Crystallization trials using incomplete factorial designs [14] allow the screening of a
much wider range of conditions with a modest number of experiments, and
Figure 3 Protein crystallization: diagrammatic representation of the hangingdrop method of vapor diffusion.
8
thus less protein. Miniaturization and automation made possible by the
use of advanced crystallization robots may also have a great impact on the
future success of protein crystallization.
B. X-Ray Diffraction Data Acquisition
The next step is to measure x-ray diffraction data from a single crystal
(Fig. 4). Data are usually measured by means of an area detector
such as a phosphorus image plate or a charge-coupled device (CCD).
Through several steps of computational analysis, the position and
amplitude or intensity of the each diffraction spot can be obtained.
Because diffraction intensities are proportional to the volume of the
crystal and generally decrease at higher resolution, protein crystals must
be reasonably large to give strong enough diffraction signals at high
resolution. While a cube of 0.1 to 0.5 mm in each dimension is still
preferred by most crystallographers, the availability of powerful synchrotron radiation sources has made the analysis of much smaller
crystals feasible. Crystals also must be stable enough in the x-ray beam
to allow the measurement of a complete diffraction data set from a
single crystal. In this regard, flash-freezing of protein crystals under
proper conditions at cryogenic temperatures [15] has virtually eliminated
radiation decay problems.
Figure 4 Diagrammatic representation of single-crystal x-ray diffraction and
data collection.
9
C. Phasing
The ultimate goal of an x-ray diffraction experiment is to produce an
electron density map that is then used to build an atomic model of the
molecule being studied (Fig. 5). The use of single-crystal x-ray diffraction
techniques to determine the three-dimensional structure of molecules
requires the measurement of amplitudes and the calculation of phases
for each diffraction spot. Although amplitudes can be directly measured
from diffracting crystals, as noted earlier, phases are indirectly determined. The inability to directly measure phases is known as the ‘‘phase
problem’’ [16]. In practice, there are several ways to get around the phase
problem. If the protein of interest is small (f100 amino acids) and highresolution data (1.2 A˚ or better) are available, phases can be obtained
computationally by using the so-called direct method. This is basically the
same technique used to determine crystal structures of small organic
molecules. If the protein being studied is known to have a fold similar
to that of a protein with a known three-dimensional structure, one uses the
molecular replacement (MR) method, in which the known structure serves
as a model to generate approximate phases that are then refined against
the experimental data obtained from crystals of the protein under study.
Until recently, multiple isomorphous replacement (MIR) was the most
widely used method for ab initio phase determination. This technique
requires the introduction into the protein under study of atoms of high
atomic number (heavy atoms) such as mercury, platinum, and uranium,
without disrupting the protein’s three-dimensional structure or the
packing in the crystal. This is achieved by soaking crystals in a solution
Figure 5 Steps in the use of protein crystallography for structure determination.
10
containing the desired heavy atom. The binding of one or more heavy
atoms to the protein alters the diffraction of the crystals from that of
the underivatized (native) crystals. If the introduction of heavy atoms is
truly isomorphous, the differences between the diffraction of the
derivative and of the native will represent only contributions from the
heavy atom(s). Thus, the problem of structure determination is reduced
to locating the position of one or a few heavy atoms. Once their
positions have been accurately determined, the heavy atoms are used to
calculate phases for all diffraction intensities. In theory, one needs only
two isomorphous derivatives, but in practice more are needed owing to
errors that are introduced in data measurement as well as the lack of
isomorphism. Multiple-wavelength anomalous dispersion (MAD) phasing, cited earlier, has gained popularity in the last 10 years, and this
more recent technique for ab initio phase determination is now the
predominant method in de novo structure determination. In the MAD
technique, cells that overexpress the protein can be grown in media
containing selenomethionine (Se-Met) instead of methionine, producing
proteins that have Se-Met at all the methionine positions. Because of
the unique absorption quality of Se, diffraction data can be measured
by using a Se-Met-substituted crystal at three or four different wavelengths around the Se absorption edge. These data can be analyzed by
using computational methods to generate phase information, allowing
an electron density map to be calculated [17]. Such an experiment calls
for modern synchrotron facilities.
D. Model Building and Refinement
Once an electron density map has become available, atoms may be fitted
into the map by means of computer graphics to give an initial structural
model of the protein. The quality of the electron density map and
structural model may be improved through iterative structural refinement but will ultimately be limited by the resolution of the diffraction
data. At low resolution, electron density maps have very few detailed
features (Fig. 6), and tracing the protein chain can be rather difficult
without some knowledge of the protein structure. At better than 3.0 A˚
resolution, amino acid side chains can be recognized with the help of
protein sequence information, while at better than 2.5 A˚ resolution
solvent molecules can be observed and added to the structural model
with some confidence. As the resolution improves to better than 2.0 A˚
resolution, fitting of individual atoms may be possible, and most of the
11
Figure 6 Electron density of an a-helix at different resolutions.
amino acid side chains can be readily assigned even in the absence of
sequence information.
E. Understanding Structural Coordinates
Once a crystal structure has been determined, the information is
communicated in the form of an atomic coordinates file. In addition
to a list of the atomic positions, the coordinates file contains other information that deserves an explanation and requires attention by the
user. Some of the terms included in an atomic coordinates file are
explained briefly. It is hoped that the information will provide the reader
with insights to evaluate the quality of the structure, distinguish between
its well-defined and flexible regions, and make sensible decisions in
structural analysis.
The unit cell is the basic microscopic building block of the crystal. A
crystal can be viewed as a three-dimensional stack of identical unit cells,
each defined by three cell edges (a, b, c, in angstroms), and three angles (a,
h, g in degrees) between each pair of edges. Each unit cell may contain one
or more protein molecules related by crystal symmetry. The unique portion
of the unit cell (i.e., the portion that is not related to other portions by
crystal symmetry) is called the asymmetric unit. There are only 230 different
combinations of symmetry elements in crystals; each of these is called a
space group. However, since biological molecules are enantiomorphic,
12
which means that a protein crystal cannot contain mirror planes, the
number of space groups of relevance to protein crystallography is reduced
to 65. It is possible to have more than one copy of the same protein in an
asymmetric unit. However, these will be related by ‘‘noncrystallographic’’
symmetry. Therefore, all atoms of an asymmetric unit, along with the unit
cell dimensions and the space group, must be given in the coordinates file
for subsequent analysis and for regenerating the structure in any portion of
the unit cell or the crystal, which may be important for studying intermolecular ‘‘crystal packing’’ interactions.
The R-factor is probably the single most important number that
provides a sense of the overall quality of the structure. It is defined as
[A||Fobs| k*|Fcalc||] / A|Fobs|, where Fobs is the observed structure factor
(the square root of the measured diffraction intensity or amplitude), Fcalc
is the structure factor calculated from the model, and k is a scaling
factor. The factor R is a measurement of the agreement between the
structural model and the observed diffraction data; the lower the
number, the better. For a refined crystal structure, the R factor is often
approximately 10 times its resolution (e.g., 20% for a 2.0 A˚ resolution
structure). Along with the traditional R factor, most of the recent
structures also report an Rfree value, which is obtained from the part
of the diffraction data (5–10%) set aside and not used during structural
refinement. Generally Rfree is 5–10% higher than R; larger discrepancies
between the two may indicate that there is a problem in the structure
model or diffraction data, or that the structure is overrefined against the
data. Reducing R to below 20% used to be the goal for structural
refinement; but obtaining a sensible Rfree is now considered to be more
important. Therefore, before analyzing a crystal structure on computer
graphics, one should check the R factor and Rfree values to get a sense of
the overall quality of the structure. It is important to note that these
values can be reported as percentages (20%) or as fractions (0.20).
The atomic temperature factor, or B factor, measures the dynamic
disorder caused by the temperature-dependent vibration of the atom, as
well as the static disorder resulting from subtle structural differences in
different unit cells throughout the crystal. For a B factor of 15 A˚2, displacement of an atom from its equilibrium position is approximately 0.44
A˚, and it is as much as 0.87 A˚ for a B factor of 60 A˚2. It is very important
to inspect the B factors during any structural analysis: a B factor of less
than 30 A˚2 for a particular atom usually indicates confidence in its
atomic position, but a B factor of higher than 60 A˚2 likely indicates that
the atom is disordered.
13
For a particular crystal, the number of diffraction data increases as
the resolution increases, which means that more experimental data will
be available for structural refinement. There are four parameters to be
refined for each atom: x, y, z (atomic position), and B (temperature
factor). If the crystal has normal solvent content (i.e., about 50%), the
number of experimental data and refinement parameters will be about the
same at 2.8 A˚ resolution. This suggests that B factors for individual
atoms should be refined only when data have a resolution better than 2.8
A˚. Refinement of atomic B factors at lower resolution will have no
physical meaning, although a lower but meaningless R factor will result.
Identification and refinement of solvent molecules (e.g., waters) become
reliable only when the structure has at least a 2.5 A˚ resolution. Even then,
before a water molecule is used in mechanistic or computational analysis,
it is always wise to check its B factor for the existence of at least one
hydrogen bond to hold the water to the protein. At times, spurious water
molecules are added (such additions will result in a meaningless lower R
factor). Unless the structure has been determined at a reasonably high
resolution, electron density and refinement often do not discriminate
between the oxygen and nitrogen atoms of asparagines and glutamines,
or the alternative conformations of histidine side chains. In a detailed
structural analysis, it may be necessary to check alternative conformations of Asn, Gln, or His side chains and decide which one makes more
sense chemically.
V. IN SILICO LEAD GENERATION
Armed with the crystal structure of the protein–ligand complex and upto-date computer modeling software, one can design additional ligands.
Numerous molecular modeling software programs are available for that
purpose. However, it is important to note that current computational
algorithms have their limitations and utilize many approximations. Therefore, while computer modeling software has been proven useful [4,18],
further testing and structural validations are required to identify the best
possible compound.
A. In Silico Screening of Virtual Compound Libraries
Starting with the crystal structure of the target, it is possible to screen for
leads in three-dimensional compound databases such as the Cambridge
14
Structural Database [19] or the Chemical Abstracts Service Registry [20],
or convert private databases to 3-D structures by programs such as
CONCORD [21]. Several programs are available for such screening. For
example, DOCK [22] works by using a set of overlapping spheres to create
a complementary image of the ligand binding site and essentially matching
the shape of a putative ligand with that of the image to generate a
‘‘goodness of fit’’ score that is then used to rank the hits identified. Instead
of comparing shapes, the program LUDI [23] uses parameters that
describe hydrogen-bonding potential and hydrophobic complementarity
to match the ligand and its binding site. These programs can rapidly search
through three-dimensional databases of small molecules and rank each
candidate. Typically, the 100 to 200 top-scoring compounds are examined
graphically to identify the best 10 to 50 candidates for experimental testing.
In the case of DOCK, 2 to 20% of these in silico hits may show micromolar
binding affinity [4]. Subsequently, crystallography can be used to optimize
any leads identified.
B. Building Leads from Molecular Fragments
Again starting with the crystal structure of the target, another strategy
is to dock small chemical fragments into the ligand binding site, then
grow the fragment to better complement the binding site. Programs
such as GRID [24], AUTODOCK [25], and MCCS [26] can be used
for the docking step. GRID uses small functional groups to probe the
binding site and evaluate interaction energies by using an empirical
Lennard-Jones energy function, as well as electrostatic and hydrogenbonding terms. AUTODOCK uses simulated annealing for ligand
conformational search to dock small ligands of flexible conformations
onto a rigid binding site and a standard force field for rapid grid-type
energy evaluation. MCSS (multicopy simultaneous search) places thousands of copies of functional groups in the binding site and optimizes
them simultaneously to generate energetically favorable positions and
orientations in a flexible binding site. Once selected, suitable binding
fragments can be built into a single compound by manual modeling or
by using linking programs such as CAVEAT [27], which attempts to
identify a suitable cyclic linker from a database. Alternatively, programs like GroupBuild [28] can search compound libraries for potential leads that have the functional fragments identified by the programs
just described.
15
VI. STRUCTURE-BASED LEAD OPTIMIZATION
Once a chemical lead has been identified, the structure-based lead
optimization process goes through several iterations of structure determination, design, chemical synthesis, and biological testing. The goal is
to optimize the lead in terms of electrostatic interactions, van der Waals
contacts, and the fit in the ligand binding pocket. The design process
may be simple and intuitive if one starts with a relatively high affinity
lead. In this case only minor modifications to the existing lead are
introduced at each of the iterations of the drug design cycle. Many of
these modifications may be either proposed from personal knowledge or
derived by computer modeling. However, it is important to note that
computational methods are still not reliable in predicting binding
modes and affinities of ligands, mainly because of inaccuracies in force
fields, limitations in dealing with ligand and target flexibility, and the
lack of a reliable scoring functions, as well as the difficulties in treating
solvent molecules. Therefore, even for seemingly minor modifications of
the leads, it is still necessary to confirm the binding mode experimentally; there are countless examples in which the mode of binding
significantly changes upon introduction of minor modifications to the
original lead.
VII.
EXPERIENCE WITH STRUCTURE-BASED
DRUG DESIGN
Any summary of experience gained during the last 15 years in the area of
structure-based drug design is in some way a work in progress, and clearly
there is much that we still need to learn.
A. Design Should Be Based on Liganded Structures
Many proteins undergo considerable conformational change upon binding
to their ligands. Initiating ligand design based on an unliganded structure
may be misleading if that structure is of a protein that will change its
conformation upon ligand binding. To be on the safe side, one should
always start ligand design based on a liganded structure of the target
protein. An example of a protein that undergoes large conformational
change upon ligation is EPSP (5-enol-pyruvyl-3-phosphate) synthase. The
16
unliganded structure [29] shows a large cavity at the active site (Fig. 7),
much of which disappears upon ligation. Sometimes, different ligands lead
to different conformational changes of the protein target, making the
designing even more challenging.
B. Design of Small Molecules to Interfere with
Protein–Protein Interaction Requires the Structure
of the Complex
Most protein–protein interfaces are large hydrophobic surfaces. For
example, the interface area between growth hormone and its receptor
[30] is about 2100 A˚2. To rationally design a small molecule to interfere
with such large surfaces is a considerable challenge that requires atomic
details of the receptor surface, which may differ for unliganded and
liganded forms. Generally, success in this arena is rare. Occasionally,
protein–protein interactions consists of only a small number of contacts,
such as the RGD interaction with its receptors [31]. In such a case, the
Figure 7 Stereoview of the structure of EPSP synthase in its open conformation.
17
design task becomes essentially a small-molecule–protein interaction
problem and is much more likely to be successful.
C. Allow for Flexibility in the Design of Enzyme
Inhibitors to Assure Optimal Fit in an Often Rigid
Active Site Cavity
It is often very difficult to design a highly constrained ligand that complements and fits snugly in an enzyme active site. Although rigidity of the
ligand is important to reduce entropy and to ensure greater affinity, it is
often wise to initially introduce some flexibility to ensure proper fit in an
often rigid active site. Much of this flexibility could be reduced as much as
possible in later iterations of the drug design cycle.
D. Synthetic Accessibility Is Essential
It is important to design ligands that can be synthesized in a timely
fashion from readily available or easily obtained starting material. Given
that many potential drugs fail for reasons that have nothing to do with
their binding affinity, it is important that one go through a design cycle as
fast as possible to obtain feedback on the suitability of the designed
ligands as drugs.
E. Every Water Molecule Is Special
Incorporation of the position of water molecules that are firmly bound to
the protein can impart affinity and novelty to the designed ligand. A prime
example is the design of a class of HIV protease cyclic urea inhibitors by
DuPont scientists that incorporated a water molecule bound to both flaps
of the enzyme into their ligand [32]. The crystal structure of the HIV
protease–cyclic urea complex [32] shows the urea carbonyl oxygen substituting for the position of the water molecule.
F. Fill Available Space and Maximize Interactions
A major goal of ligand design should be to fill as much of the space in the
binding site as possible without rendering the designed ligand too large.
Ligands greater than 500 Da have lower probability of being orally
bioavailable. It is also important to maximize both polar and nonpolar
interactions with the protein.
18
G. Beware of Crystal Contacts
In the crystal, it is possible for a ligand to make important contacts with
residues from a neighboring molecule producing an artificial mode of
binding that is not possible in solutions. Thus it is important to analyze all
crystal contacts in the vicinity of the ligand prior to proceeding with the
design of new ligands.
H. Use of Surrogate Enzymes Can Lead to Important
Insights, But Optimization Requires the
Target Protein
When the target enzyme is difficult to obtain, related enzymes could be
used to provide insights in the design of novel ligands. For example, papain
was used to design a class of potent cathepsin K inhibitors [33] spanning
both sides of the papain active site. However, fine-tuning these inhibitors to
produce more potent ones required the use of the crystal structure of
cathepsin K itself [34].
I. Iterative Design Is Essential
It is a rarity that the first ligand to be designed is the final one. Thus, it is
common to go through the structure-based design cycle (Fig. 2) several
times with each class of inhibitors being designed. This iteration should
continue until the ideal molecule that will be advanced to development has
been identified.
J. Solubility of Ligands Matters
One of the bottlenecks associated with structure-based design is poor
aqueous solubility of many ligands. If the ligands are insoluble in water, it
is often difficult to form complexes under conditions of crystallization.
Unlike the crystallization of small organic molecules, proteins must be
crystallized from aqueous solutions or using solvents that are highly
miscible with water. Therefore, it is sensible to introduce polar or charged
groups to improve inhibitor solubility, making the target molecule more
amenable to structural studies.
K. No Substitute for Experience
Structure-based drug design is no different from most other areas, in that
experience counts greatly.
19
L. Dedicated Molecular Biology and Protein
Purification Groups Are Essential
Protein crystallography often requires special constructs or mutants to
facilitate crystallization; it also requires large quantities of highly purified
protein. Thus to move forward in a timely fashion, it is important that an
industrial structural biology group employ molecular biologists and
individuals with expertise in protein purification.
VIII. OUTLOOK
Structure-based drug design is now an integral part of most if not all drug
discovery programs. It is a given that structure-based design is part of each
drug discovery effort whenever the target is a soluble protein. However, a
large segment of targets—namely, membrane proteins, particularly Gprotein-coupled receptors—are excluded. It is hoped that this situation will
be remedied in the near future.
REFERENCES
1. Hol WGJ. Protein crystallography and computer graphics-toward rational
drug design. Angew Chem 1986; 25:767–778.
2. Wlodawer A, Erickson JW. Structure-based inhibitors of HIV-1 protease.
Annu Rev Biochem 1993; 62:543–585.
3. Kim CU, Chen X, Mendel DB. Neuraminidase inhibitors as anti-influenza
virus agents. Antiviral Chem Chemother 1999; 10:141–154.
4. Kuntz ID. Structure-based strategies for drug design and discovery. Science
1992; 257:1078–1082.
5. Verlinde LMJ, Hol WGJ. Structure-based drug design: progress, results and
challenges. Structure 1994; 2:577–587.
6. Blundell TL. Structure-based drug design. Nature 1996; 384(suppl):23–26.
7. Babine RE, Bender SL. Molecular recognition of protein–ligand complexes:
application to drug design. Chem Rev 1997; 97:1359–1472.
8. Hol WGJ, Verlinde LMJ. Macromolecular crystallography and medicine.
In: Rossmann MG, Arnold E, eds. International Tables of Crystallography.
Vol. F. Dordrecht, The Netherlands: Kluwer Academic, 2002.
9. Amzel LM. Structure-based drug design. Curr Opin Biotech 1998; 9:366–
369.
20
10. Hogan JC Jr. Directed combinatorial chemistry. Nature 1996; 384 suppl:
17–19.
11. Salemme FR, Spurlino J, Bone R. Serendipity meets precision: the integration of structure-based drug design and combinatorial chemistry for
efficient drug discovery. Structure 1997; 5:319–324.
12. Ferre-D’Amare AR, Burley SK. Use of dynamic light scattering to assess
crystallizability of macromolecules and macromolecular assemblies. Structure 1994; 2:357–359.
13. Hensley P. Defining the structure and stability of macromolecular assemblies in solution: the re-emergence of analytical ultracentrifugation as a
practical tool. Structure 1996; 4:367–373.
14. Carter CW Jr, Carter CW. Protein crystallization using incomplete factorial
experiments. J Biol Chem 1979; 254:12219–12223.
15. Rodgers DW. Cryocrystallography. Structure 1994; 2:1135–1140.
16. Drenth J. Principles of Protein X-Ray Crystallography. New York: SpringerVerlag, 1994.
17. Hendrickson WA, Horton JR, LeMaster DM. Selenomethionyl proteins
produced for analysis by multiwavelength anomalous diffraction (MAD):
a vehicle for direct determination of three-dimensional structure. EMBO J
1990; 9:1665–1672.
18. Bohm HJ. Current computational tools for de novo ligand design. Curr
Opin Biotechnol 1996; 7:433–436.
19. Allen FH, et al. Watson DG. The development of versions 3 and 4 of the
Cambridge Structural Database system. J Chem Inf Comput Sci 1991;
31:187–204.
20. Fisanick W, Cross KP, Forman JC, Rusinko A III. Experimental system for
similarity and 3D searching of CAS Registry substances. 1. 3D substructure
searching. J Chem Inf Comput Sci 1993; 33:548–559.
21. Pearlman RS. Three-dimensional structures: how do we generate them
and what can we do with them? Chem Des Auto News 1993; 8–10:44–47.
22. Kuntz ID, Blaney JM, Oatley SJ, Langridge R, Ferrin TE. A geometric
approach to macromolecule–ligand interactions. J Mol Biol 1982; 161:269–
288.
23. Bohm H-J. The computer program LUDI: a new method for the de novo
design of enzyme inhibitors. J Comput-Aided Mol Des 1992; 6:61–78.
24. Goodford PJ. A computational procedure for determining energetically
favorable binding sites on biologically important macromolecules. J Med
Chem 1985; 28:849–857.
25. Goodsell DS, Olsen AJ. Automated docking of substrates to proteins by
simulated annealing. Proteins 1990; 8:195–202.
26. Miranker A, Karplus M. Functionality maps of binding sites: a multiple
copy simultaneous search method. Proteins 1991; 11:29–34.
27. Bartlett PA, Shea GT, Telfer SJ, Waterman S. CAVEAT: a program to
28.
29.
30.
31.
32.
33.
34.
21
facilitate the structure-derived design of biologically active molecules. In:
Roberts SM, ed. Molecular Recognition in Chemical and Biological
Problems. Cambridge: Royal Society of Chemistry, 1989:182–196.
Rotstein SH, Murcko MA. GroupBuild: a fragment-based method for de
novo drug design. J Med Chem 1993; 36:1700–1710.
Stallings WC, Abdel-Meguid SS, Lim LW, Shieh H-S, Dayringer HE,
Leimgruber NK, Stegeman RA, Anderson KS, Sikorski JA, Padgette SR,
Kishore GM. Structure and topological symmetry of the glyphosphate
5-enol-pyruvylshikimate-3-phosphate synthase: a distinctive protein fold.
Proc Natl Acad Sci USA 1991; 88:5046–5050.
Cunninghum BC, Ultsch M, De Vos AM, Mulkerrin MG, Clauser KR,
Wells JA. Dimerization of the extracellular domain of the human growth
hormone receptor by a single hormone molecule. Science 1991; 254:821–825.
Ku TW, Miller WH, Bondinell WE, Erhard KF, Keenan RM, Nichols AJ,
Peishoff CE, Samanen JM, Wong AS, Huffman WF. Potent non-peptide
fibrinogen receptor antagonists which present an alternative pharmacophore. J Med Chem 1995; 38:9–12.
Lam PYS, Jadhav PK, Eyermann CJ, Hodge CN, Ru Y, Bacheler LT, Meek
JL, Otto MJ, Rayner MM, Wong YN, Chang C-H, Weber PC, Jackson DA,
Sharpe TR, Erickson-Viitanen S. Rational design of potent, bioavailable,
nonpeptide cyclic ureas as HIV protease inhibitors. Science 1994; 263:380–
384.
LaLonde JM, Zhao B, Smith WW, Janson CA, DesJarlais RL, Tomaszek
TA, Carr TJ, Thompson SK, Oh HJ, Yamashita DS, Veber DF, AbdelMeguid SS. Use of papain as a model for the structure-based design of
cathepsin K inhibitors: crystal structures of two papain–inhibitor complexes
demonstrate binding to SV-subsites. J Med Chem 1998; 41:4567–4576.
Thompson SK, Smith WW, Zhao B, Halbert SM, Tomaszek TA, Tew DG,
Levy MA, Janson CA, DAlessio KJ, McQueney MS, Kurdyla J, Jones CS,
DesJarlais RL, Abdel-Meguid SS, Veber DF. Structure-based design of
cathepsin K inhibitors containing a benzyloxy-substituted benzoyl peptidomimetic. J Med Chem 1998; 41:3923–3927.
2
Src Homology-2 Domains
and Structure-Based,
Small-Molecule Library Approaches
to Drug Discovery
Chester A. Metcalf III and Tomi Sawyer
ARIAD Pharmaceuticals, Inc., Cambridge, Massachusetts, U.S.A.
The elucidation of cell-receptor-associated signal transduction pathways
by means of the tools of biochemistry and molecular genetics has
resulted in the identification of a multitude of protein targets for
therapeutic intervention (Table 1) [1]. The fact that many of these
targets have x-ray crystallography and/or NMR spectroscopy to guide
the syntheses of structurally biased single analogues and combinatorial
libraries has ushered the pharmaceutical industry into a new era of drug
discovery. Within cells there exists an enormously diverse data set of
functional proteins and signaling pathways, involving both noncatalytic
and catalytic processes, which are orchestrated through highly specific
protein–protein interactions. In principle, such interactions can be
disrupted or promoted, either directly or indirectly (via enzyme inhibition), through small-molecule intervention driven by structure-based
methods. This chapter discusses the role of Src homology-2 (SH2)
domains as mediators of protein–protein interactions in signal transduction, with a focus on the therapeutic implications of blocking the
SH2 domain of the nonreceptor protein tyrosine kinase Src with
23
Angiotensin (AT1, AT2)
Bradykinin (B1, B2)
Cholecystokinin (CCKA)
Gastrin (CCKB)
Endothelin (ETA, ETB)
a–Melanotropin (MCR1)
Adrenocorticotropin (MCR2)
Substance P (NK1)
Neurokinin-A (NK2)
Neurokinin-B (NK3)
y-opioid (Enkephalin)
A-opioid (Endorphin)
n-opioid (Dynorphin)
Oxytocin
Somatostatin (sst1–sst5)
Vasopressin (V1A, V1B)
Neuropeptide-Y (Y1-Y5)
Calcitonin
G-Protein-Coupled/Integrin
Receptors
Nonreceptor serine/threonine kinases
cAMP-Dependent protein kinase
Phosphoinositol-3-kinase (P13K)
Cyclin-dependent kinases (CDKs)
Receptor serine/threonine kinases
Transforming growth factor
Nonreceptor tyrosine kinases
Src and Src family (Lck, Hck)
Abl, Syk, Zap-70
Epidermal growth factor
Fibroblast growth factor
Insulin
Nerve growth factor
Platelet-derived growth factor
Receptor tyrosine kinases
Receptor/Nonreceptor
Kinases/Phosphatases
Table 1 List of Possible Protein Targets for Therapeutic Intervention
Serinyl proteases
Trypsin
Thrombin
Chymotrypsin-A
Kallikrein
Elastase
Tissue plasminogen activator
Factor Xa
Proteases
Aspartic proteases
Pepsin
Renin
Cathepsins (D, E)
HIV-1 protease
NF-nB, STAT, NFAT, SMAD, CREB
Transcription
Factors/Proteases
24
Metcalf and Sawyer
Source: Ref. 1.
Cell adhesion integrin receptors
avh3 (Fibrinogen)
aIIbh3 or gpIIaIIIb (Fibrinogen)
a5h1 (Fibrinectin)
a4h1 (VCAM-1)
Adenosine (A1-A3)
Cathecholamine (a1, a2, h1-h3)
Histamine (H1, H2)
Muscarinic acetylcholine
Seratonin (5HT1-5HT7)
Melatonin (ML1A, ML1B)
Dopamine (D1, D2, D4, D5)
g-Amino butyric acid (GABAB)
Leukotrienes (LTB4, LTC4, LTD4)
Cysteinyl proteases
Cathepsins (B, H, K, M, S, T)
Proline endopeptidase
Interleukin-converting enzyme
Apopain (CPP-32)
Picornavirus C3 protease
Calpains
Nonreceptor tyrosine phosphatases
PTP1B, Syp
Metallo proteases
Exopeptidase group
Peptidyl dipeptidase-A (ACE)
Aminopeptidase-M
Nonreceptor serine/threonine phosphatases Carboxypeptidase-A
PP-1
Calcineurin
Endopeptidase group
VH1
Endopeptidases (24, 11, 24, 15)
Stromelysin
Gelatinases (A, B)
Collagenase
Receptor/nonreceptor phosphatases
Receptor tyrosine phosphatases
CD45, LAR
Mitogen-activated protein kinase
Protein kinase C (PKC)
Janus family kinases (JAKs)
InB family kinases (IKKs)
Drug Discovery via Src Homology-2 Domains
25
26
Metcalf and Sawyer
designed, nonpeptide small molecules. We highlight ARIAD’s approach
to drug discovery by means of structure-based methods and parallel
synthetic libraries to develop cell-active, in vivo effective inhibitors of Src
SH2-dependent signal transduction pathways, leading to novel drugs for
the treatment of osteoporosis.
I. SIGNAL TRANSDUCTION AND PROTEIN–PROTEIN
INTERACTIONS
The network of protein–protein interactions that define signal transduction pathways in most cells originates at a cellular receptor and is
triggered by the binding of specific external stimuli (e.g., growth factors,
antigens, hormones). Such signal transduction pathways are then propagated within the cell to the nucleus resulting in specific gene activation
and protein synthesis (Fig. 1) [2]. Listed in Table 2 are the modular
domains [3] of various signal transduction proteins and the potential
disease areas providing opportunity for therapeutic intervention through
small-molecule inhibitory drugs [4]. For Src, these disease areas are
osteoporosis and cancer. There are more than 50 known SH2-containing
proteins, of which Src was the first to be identified [5]. The SH2 domain
of Src consists of approximately 100 amino acids and binds cognate
Figure 1 Representation of signal transduction pathways describing highly
specific protein–protein interactions.
27
Table 2 Signal Transduction Proteins as Potential Therapeutic Targets
Protein (Domains)
Src (SH3-SH2-kinase)
Hck (SH3-SH2-kinase)
Syk (SH2-SH2-kinase)
Zap70 (SH2-SH2-kinase)
Syp (SH2-SH2-phosphatase)
STATs (DNA binding-SH3-SH2)
Grb2 (SH3-SH2-SH3)
p85/PI3K [SH3-SH2-SH2 (p85 subunit)]
Bcr/Abl (SH3-SH2-kinase)
Disease area
Osteoporosis, cancer
Immune disease, AIDS
Allergy, asthma
Autoimmune disease
Anemia
Inflammatory disease
Cancer, chronic myelogenous leukemia
Cancer
Chronic myelogenous leukemia
Source: Ref. 4.
phosphotyrosine(pTyr)-containing proteins as well as synthetic peptides in
a sequence-dependent manner [6]. In addition to the SH2 domain, Src
possesses one SH3 domain (f 60 amino acids), which is characterized by
its affinity for proline-rich sequences, and a bilobed tyrosine kinase
catalytic domain (f 300 amino acids) containing N-terminal (NT) and
C-terminal (CT) domains (Fig. 2) that specifically phosphorylates tyrosine
residues of cognate substrate proteins.
II. Src TYROSINE KINASE AND OSTEOPOROSIS
Molecular insight into the protein conformation states of Src kinase has
been revealed in a series of x-ray crystal structures of the Src SH3–SH2–
kinase domain that depict Src in its inactive conformation [7]. This form
maintains a ‘‘closed’’ structure, in which the tyrosine-phosphorylated
(Tyr527) C-terminal tail is bound to the SH2 domain (Fig. 2). The x-ray
data also reveal binding of the SH3 domain to the SH2–kinase linker
[adopts a polyproline type II (PP II) helical conformation], providing
additional intramolecular interactions to stabilize the inactive conformation. Collectively, these interactions cause structural changes within the
catalytic domain of the protein to compromise access of substrates to the
catalytic site and its associated activity. Significantly, these x-ray structures
provided the first direct evidence that the SH2 domain plays a key role in
the self-regulation of Src.
The bone disease osteoporosis results when an imbalance occurs in
the normal course of bone remodeling, a dynamic and highly regulated
28
Metcalf and Sawyer
Figure 2 Depiction of the active (‘‘open’’) and inactive (‘‘closed’’) conformations of Src kinase based on the analysis of x-ray structures of c-Src tyrosine
kinase crystallized in its inactive state [7]. The stabilization of the inactive
conformation is influenced by multiple events including intramolecular binding
of the tyrosine-phosphorylated C-terminus tail to the SH2 domain as well as
interactions between the SH3 domain and the SH2–kinase linker. CT, C-terminal;
NT, N-terminal.
process that involves both bone degradation (resorption) and bone formation. Aberrantly high levels of bone resorption are associated with this
disease, which effects a net decrease in bone density and volume, resulting
in fragile, brittle bones that are subject to premature breaks and fractures
[8]. The most compelling evidence that Src is intimately involved in bone
remodeling comes from genetically engineered Src knockout mice. In these
Src (–/–) mice, the only major phenotype observed is excessive bone
29
formation; a condition termed osteopetrosis (the opposite of osteoporosis). This suggests that selective inhibition of Src, as a therapeutic treatment
for osteoporosis, may shift the bone microenvironment from a state of
perpetual bone degradation to one of normal bone turnover without
deleteriously affecting other Src-associated cellular processes in the body.
The rationale for Src’s involvement in bone processes becomes
apparent when the Src knockout effects are examined at the cellular level
of an osteoclast. Osteoclasts are multinucleated cells that are responsible
for bone resorption. Two different osteoclasts, a wild-type (normal) and
a Src (–/–) osteoclast, are shown schematically in Figure 3. The wild-type
cell shows the characteristics of a bone-resorbing osteoclast, including a
ruffled border and ‘‘pit’’ formed by the bone-degrading actions of an
active osteoclast. These features are absent in the Src (–/–) knockout
osteoclast, albeit they are still viable and adhere to bone. In 2000
(Marzia et al. and Amling et al.) it was suggested that Src plays a
negative regulatory role in osteoblasts (cells that are responsible for the
Figure 3 The effect of an Src (–/–) knockout in mice as shown by differences in
function and appearance of wild-type and Src-minus osteoclasts. The Src-minus
osteoclasts lack the ruffled borders of a normal, resorbing osteoclast, but are
viable and can adhere to bone. (From Ref. 8.)
30
Metcalf and Sawyer
formation of bone) as shown by enhanced bone formation and osteoblast differentiation rates in Src (–/–) mice [8]. Together, these data
provide complementary, mechanistic evidence to validate Src as a viable
therapeutic target for the treatment of bone diseases such as osteoporosis.
III. SH2 DOMAINS AND PHOSPHOPEPTIDE BINDING
The question of whether ligand binding specificity exists among SH2
domains was addressed, quite elegantly, by Cantley et al. [6,9], who used
synthesized combinatorial libraries of pTyr-containing peptides. A majority of the SH2 binding affinity in pTyr-containing peptides can be
attributed to a four-amino acid region represented by the sequence pTyrAaa-Bbb-Ccc. However, binding specificity exists in the three amino acids
directly C-terminal to the pTyr (pY) group, referred to sequentially as
pY+1 (Aaa), pY+2 (Bbb), and pY+3 (Ccc). The preferred pY+1,
pY+2, and pY+3 amino acids for various SH2-containing proteins are
listed in Table 3; the first amino acid listed for each position represents the
most preferred. For Src SH2 this sequence is pTyr-Glu-Glu-Ile. Such
sequence specificity among SH2-containing proteins provides a rationale
for the differentiation of their associated signal transduction pathways
in cells.
The successful design of small molecules to interact with a protein
binding surface is markedly enhanced by an understanding of the target’s
three-dimensional structure, preferably in the context of a bound ligand.
In this regard, early x-ray structures of pTyr-containing peptides bound
to Src SH2 [10,11] paved the way for the discovery of peptide, peptidomimetic, and nonpeptide ligands and the determination of their complexed structures with Src SH2 [12–14] or the highly homologous Lck
SH2 [15,16]. In a landmark paper, Waksman, Kuriyan, and their
colleagues reported [11] the first x-ray structure of a high affinity phosphopeptide (Glu-Pro-Gln-pTyr-Glu-Glu-Ile-Pro-Ile-Tyr-Leu) bound Src
SH2 (KD = 3–6 nM), which uncovered key protein interactions with the
pTyr-Glu-Glu-Ile sequence. In particular, this x-ray structure shows the
bound phosphopeptide oriented perpendicular to a central h sheet and
interacting with two major binding regions of the Src SH2 domain,
namely, one for the ligand pTyr (pY pocket) and another for the ligand
Ile (pY+3 pocket), to provide what has been described as a ‘‘twopronged’’ binding mode. The pY pocket is characterized mainly by
31
Table 3 SH2 Specificity for Phosphotyrosine-Containing Peptides
-Asp-Gly-[pTyr-Aaa-Bbb-Ccc]-Ser-Pro(pY)(pY+1)(pY+2)(pY+3)
SH2 Domain
Group 1Aa
Src
Lck
Group 1Ba
Abl
Syk (N-SH2)
Syk (C-SH2)
Grb2
Group 3b
PLCg (N-SH2)
PLCg (C-SH2)
p85 (N-SH2)
SHC
Aaa
Bbb
Ccc
Glu, Asp, Thr
Glu, Thr, Gln
Glu, Asn, Tyr
Glu, Asp
Ile, Met, Leu
Ile, Val, Met
Glu, Thr, Met
Gln, Thr, Glu
Thr
Gln, Tyr, Val
Asn, Glu, Asp
Glu, Gln, Thr
Thr
Asn
Pro, Val, Leu
Thr
Ile, Leu, Met
Tyr, Gln, Phe
Leu, Ile, Val
Val, Ile, Leu
Met, Ile, Val
Ile, Glu, Tyr
Glu, Asp
Ile, Leu
—
—
Leu, Ile, Val
Pro, Val, Ile
Met
Ile, Leu, Met
a
Group 1 contains Tyr or Phe at hD5.
Group 3 contains Ile, Cys, or Leu at hD5.
Sources: Refs. 6, 9.
b
electrostatic interactions, while the pY+3 pocket involves interactions
that are mostly hydrophobic.
Figure 4 represents the specific Src SH2 binding interactions with
pTyr-Glu-Glu-Ile sequences, as interpreted from x-ray structures [10,11].
The major intermolecular interactions in the pY pocket involve the
phosphate oxygens of the ligand pTyr side chain with the conserved basic
residues Arg158 and Arg178 of Src SH2. It is noted that Arg178 mutation
results in essentially a total loss of binding affinity [17]. Additional
intermolecular hydrogen-bonding interactions are also observed with
Ser180, Thr182, and the backbone NH of Glu181, whereas a hydrophobic
contact occurs between the alkyl side chain of Lys206 and the phenyl ring
of the ligand pTyr residue. The two adjacent glutamic acid residues
(pY+1 and pY+2) form relatively weak interactions (electrostatic and
hydrophobic) with the protein, albeit their extend side chain conformations (oriented away from each other) serve to align and rigidify the
peptide backbone. This is an important feature from a drug discovery
perspective and can be used in the design of rigid, nonpeptide templates
32
Metcalf and Sawyer
Figure 4 Representation of the binding interactions involving the phosphopeptide motif pTyr-Glu-Glu-Ile with Src SH2 as interpreted from complexed x-ray
structures [10,11]. The binding regions of the protein, including the major pY and
pY+3 pockets, are represented by their key binding residues. Also included are
the observed structural waters and their interactions with the pY+1 Glu and pY+3
Ile phosphopeptide residues.
to advance Src SH2 inhibitors (see later: Sec. IV, Lead Discovery and
Combinatorial Chemistry).
The only direct ligand–protein hydrogen bond contact involves the
backbone NH of the pY+1 Glu with the carbonyl oxygen of the His 204
residue. In addition to the hydrophobic interactions involving the Ile
phosphopeptide residue and the pY+3 pocket, there exist potential
hydrogen-bonding possibilities from Tyr205, Ile217, and a buried
Tyr233 residue. Finally, two structural water molecules provide hydrogen-bonding networks between the pY+1 Glu (CO) and pY+3 Ile (NH)
phosphopeptide residues, and the Lys206 (NH) and Ile217 (CO) Src SH2
protein residues, respectively. Such structural waters act as drug design
elements to increase binding affinity (through favorable entropic contributions) and can be exploited by small molecules that bind to or displace
them (see later: Sec. VI, Structure-Based, Small-Molecule Libraries to
Explore Src SH2 Binding).
The importance of the pTyr group for SH2 binding is counterbalanced by the biological instability of the phosphate group to cellular
33
Figure 5 List of pTyr mimics containing nonhydrolyzable and reduced charge
functionality, which were explored in the context of a pTyr-Glu-Glu-Ile peptide.
(From Ref. 12.)
phosphatases as well as low cellular permeability posed by the highly
charged phosphate group [18]. These issues have prompted the pursuit of
pTyr mimics to discover cellulary active inhibitors. In a comparative
binding study involving pTyr mimics, in the context of a pTyr-Glu-GluIle sequence, researchers explored the ability of a variety of functional
groups to act as pTyr replacements (Fig. 5) [12]. The highest affinity,
nonhydrolyzable pTyr replacement was found to be the F2Pmp (difluorophosphonomethyl phenylalanine) group [19]. Although some of the
aforementioned pTyr replacements represent nonhydrolyzable moieties,
the design of a stable pTyr mimic providing both high affinity and adequate
cell permeability has remained challenging [20].
IV. LEAD DISCOVERY AND COMBINATORIAL
CHEMISTRY
The integration of structural biology, drug design (molecular modeling
and ‘‘druglike’’ assessment), and synthetic chemistry to discover novel
small-molecule leads follows the general iterative process outlined in
Figure 6. Available structural knowledge is used to design pharmaceutically driven compounds that will bind a desired protein target; these
34
Metcalf and Sawyer
Figure 6 The iterative drug discovery process integrating structural biology,
drug design, synthetic chemistry, biological testing, and additional input from
other related research areas.
compounds are then synthesized and tested in the appropriate assays. The
biological data are analyzed in the context of available (x-ray or NMR)
structural information to impact the design of the next series of analogues. This process is repeated until a lead compound or series of compounds possessing the desired biological activities are obtained.
The database of available structural information during ARIAD’s
initial investigation into compounds targeting Src SH2 was limited; cases
involving ligand complexes utilized only peptide molecules [21]. Motivated
by an interest to develop orally active Src inhibitors (i.e., nonpeptides) we
adopted an exploratory approach to small-molecule lead discovery, using a
combinatorial chemistry strategy. Combinatorial libraries were biased
with a common phenyl phosphate group and systematically engineered
with diversity elements (selection guided by modeling) to probe the protein
surface for existing and new binding interactions (Fig. 7). Solid phase array
synthesis encompassing a novel phosphate ester linker strategy [22] was
35
Figure 7 A novel, phosphate ester linking strategy [22] was used in the synthesis
of phenyl phosphate–containing compound libraries, accomplished in 96deepwell reaction blocks [23]. Rigid, nonpeptide templates (A-group) and
pY+3 substituents (B-group) satisfied the diversity sites of the molecules.
used to construct the compound libraries in a 96-deepwell plate format
[23]. The diversity elements of the molecules included nonpeptide templates (A-group) and pY+3 substituents (B-group). The A-group diversity
elements were typically rigid and provided access to both the pY and
pY+3 pockets (in a manner similar to the aforementioned Glu-Glu
sequence) as well as directionality for each attached substituent. Binding
interactions targeting the pY+3 pocket were explored through hydrophobic B-group diversity elements. Finally, diversity building blocks
were chosen to target final products in the molecular weight range of
500 to 600. Compounds were screened in a high throughput, fluorescent
polarization (FP) binding assay [24], using estimated concentrations
(relative to a 50 mM DMSO product stock solution assuming 100%
synthetic conversion). To verify final product formation, we performed
qualitative analysis for all compounds using electrospray (+/–) mass
spectroscopy. An HPLC peak area purity assessment was also conducted for selected compounds.
V. SOLID-PHASE PARALLEL SYNTHESIS AND
NONPEPTIDE PHENYL PHOSPHATE LIBRARIES
The combinatorial construction of compound libraries in 96-deepwell
plates is efficiently accomplished by adhering to the two-dimensional grid
36
Metcalf and Sawyer
pattern of the plate. For example, if different diversity elements are
added across the rows of the plate (one diversity element, repeated 8
times, per row), with the 12 columns housing the second set of diversity
elements, 96 discrete compounds (e.g., 8 A-group 12 B-group = 96
compounds; Fig. 7) will result. This format permits the rapid synthesis of
relatively large structurally biased libraries by systematically combining
sets of diversity groups.
To streamline plate synthesis, we developed with Cyberlab, Inc. [25] a
custom high throughput organic synthesizer designed to process the 96deepwell reaction blocks (Fig. 8). This instrument was constructed to
tolerate a wide range of chemistries; therefore, all liquid contacts (syringes,
needles, tubes, and valves) are made of glass, stainless steel, or Teflon.
Coaxial tip needles with N2 inlets (connected to a bubbler) allow inert
dispensing and withdrawal of liquid reagents from the closed vessels
without excessive negative or positive pressure buildup. The instrument
head, which can access all positions on the deck, is fitted with single-needle
and four-needle probes. The tandem use of both needle probes facilitates
the transfer (via 5 mL syringe pumps, not shown) of all reaction intermediates from the reagent vials (left side) to the resin-containing 96deepwell reaction blocks (right side).
The reaction block, which provides a fully enclosed reaction
environment (Teflon, polypropylene, and silicone rubber seals) is a
slightly modified version of a design first disclosed by Sphynx Pharmaceuticals [23]. Figure 9 shows the reaction block and the reagent
vials (100, 30, and 10 mL sizes) in their fully assembled and disassembled states. Holes at the bottom of the wells of the 96-deepwell
polypropylene plate (sealed in fully assembled reaction block) allow the
reaction solutions to be removed from the wells (via a separate vacuum
plenum) and the functionalized resin (retained by Teflon frits) to be
washed with solvents.
The use of the phenyl phosphate group as both a solid support
attachment site and a crucial binding element represents what has been
referred to as a ‘‘pharmacophore-linking’’ strategy [26]. We explored a
variety of phenyl phosphate tether functionalities to provide resins varying
in substitution pattern and in chemical flexibility (Scheme 1 and Table 4)
[22]. All phenyl phosphate resins were synthesized in batch quantities of
20 g or more. Resin synthesis began with the addition of either p-methoxybenzyl alcohol or benzyl alcohol to commercially available bis(diisopropylamino)chlorophosphine, followed by addition of the diversity phenol
[(R1)-OH, DIAT (diisopropylamino tetrazole)]. Displacement of the
37
Figure 8 High throughput organic synthesizer developed in collaboration with
Cyberlab, Inc. [25] and designed to process the 96-deepwell reaction blocks. The
instrument is capable of tolerating a wide range of chemistry (liquid contacts are
glass, stainless steel, or Teflon) and accomplishes the transfer of reagents with
coaxial tip (N2 inlets) single-needle and four-needle probes.
remaining diisopropylamino group from 1 with Wang resin and oxidation
with 4-methylmorpholine N-oxide (NMO) provided the protected phenyl
phosphate resins 2 and 3 in excellent yields, as shown in Table 4. Two types
of functionality, namely, protected carboxy and amino groups, differentiated the starting phenols. Reaction schemes demonstrating compound
synthesis using both phenol types are shown in Scheme 2 [22]. Mild Fmoc
deprotection (1% DBU/DMA) of resin 2a and amide formation using
standard coupling conditions [TBTU, DIEA, p-(CO2H)PhCH2NHFmoc)] resulted in attachment of the first diversity element to provide
resin product 4. A second deprotection followed by a double, one-pot
38
Metcalf and Sawyer
Figure 9 The 96-deepwell reaction block and the reagent vials (100, 30, and 10
mL sizes) used in the organic synthesizer in their fully assembled and
disassembled states.
Scheme 1
39
Table 4 Yields and Loading of Phenylphosphate Resins
Resin
2a
2b
2c
2d
2e
3f
3g
3h
3i
R1
Loading (mmol/g)
Yield (%)
p-(FmocNHCH2CH2)Ph
p-(FmocNHCH2)Ph
m-(FmocNHCH2)Ph
p-(Allyl-O2CCH2CH2)Ph
p-(Allyl-O2CCH2)Ph
m-(Allyl-O2CCH2)Ph
p-[ p-(FmocNHCH2)PhO]Ph
p-(Allyl-O2CCH=CH)Ph
m-(Allyl-O2CCH=CH)Ph
0.659
0.637
0.627
0.730
0.672
0.807
0.552
0.535
0.796
93
89
88
92
84
98
81
66
98
Source: Ref. 22.
Scheme 2 Abbreviations: DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; DMA,
N,N-dimethylacetamide; TBTU, O-benzotriazole-1-yl-N,N,NV,NV-tetramethyluronium tetrafluoroborate; DIEA, diisopropylethylamine; RA, reductive amination; TFA, trifluoroacetic acid; DCM, dichloromethane.
40
Metcalf and Sawyer
reductive amination [Na(OAc)3BH, 2-ethylbutyraldehyde] provided the
fully coupled compound with a branch point at the second diversity
site. Mild conditions (30% TFA/CH2Cl2) to cleave the final compound
from the solid support as well as to remove the p-methoxybenzyl protecting group resulted in the isolation of compound 5 in 86% HPLC purity,
following in vacuo concentration. For the synthesis of compound 7, the
allyl ester of resin 2d was deprotected under palladium-mediated conditions, followed by amide coupling [TBTU, DIEA, m-(NH2CH2) PhCO2allyl] to generate the functionalized phenyl phosphate resin 6. A second
deprotection and coupling [TBTU, DIEA, NH(Me)CH2Ph] provided the
bisamide resin-bound compound, which was cleaved and isolated as
described earlier to yield compound 7 in 66% HPLC purity.
The compound types synthesized by using the foregoing combinatorial approach are represented in Figure 10. Variations in functional
group connectivity (e.g., amides, olefins, sulfonamides) reflect the wide
range of chemistry that was pursued in the generation of these libraries.
Some bifunctional A-group and monofunctional B-group diversity elements used in the coupling reactions are shown in Figure 11. Alkyl
and aryl phosphate ester groups (R; see Fig. 10) were also explored to
Figure 10 Representations of some of the compound types synthesized in the
nonpeptide phenyl phosphate libraries.
41
investigate the binding consequences of reduced charge at the phosphate
group. Initial Src SH2 screening produced hits, which were resynthesized
and then retested in the binding assay. Some of the higher affinity
compounds are shown in Figure 12. Although more than 10,000 compounds were produced by this methodology, only marginal binding
affinities and no high-resolution x-ray or NMR structures were achieved;
the poor aqueous solubility and undesirable physical properties of the
molecules are likely to have hampered these efforts. At this point a decision
was made to pursue a much more structure-based approach. Compounds
Figure 11 List of some of the molecular diversity building blocks used in the
construction of the nonpeptide phenyl phosphate libraries.
42
Metcalf and Sawyer
Figure 12 Resynthesized library hits identified from the high throughput
fluorescence-polarization assay along with their Src SH2 binding data.
43
lending themselves to possible x-ray and/or NMR co structure determination were emphasized.
VI. STRUCTURE-BASED, SMALL-MOLECULE
LIBRARIES TO EXPLORE Src SH2 BINDING
Refocusing our drug-discovery strategy prompted us to revisit the initial
lead compound, pTyr-Glu-Glu-Ile. It was clear that to generate high
affinity, small-molecule compounds for Src, we would likely need to
maintain the key binding interactions of the pTyr-Glu-Glu-Ile motif, as
well as to explore molecules capable of mimicking or interacting with the
structural waters found in the Src SH2-phosphopeptide complexed x-ray
structure. A template would be required that allowed access to both
pockets (pY and pY+3), mimicking the ‘‘two-pronged’’ binding mode
of pTyr-Glu-Glu-Ile. Noteworthy in this regard, a novel, de novo designed
nonpeptide 8 was disclosed [14] with comparable binding to Ac-pTyrGlu-Glu-Ile-NH2 (phosphopeptide 9) (Fig. 13). Significant interactions
involving the benzamide functionality were revealed in an x-ray structure
of 8 bound to Src SH2 [14]. In addition to interacting with several key sites
of Src SH2 (e.g., the pY/pY+3 pockets and the CO of His204), this
compound displaces both structural water molecules and makes a direct
hydrogen bond contact with the backbone NH of Lys206 through its
benzamide CO moiety. The effect of this carboxamide group on Src SH2
binding is demonstrated by the related compounds 10 and 11 [14], in which
the desamide compound 11 binds with over 15-fold lower affinity than 10
(Fig. 13). ARIAD’s strategy was to utilize this high affinity benzamide
template to gain a better understanding of nonpeptide interactions with Src
SH2, and then to advance a database of structure–activity relationships
(SARs) to ultimately develop novel, proprietary Src SH2 inhibitors.
Subsequent to the disclosure of compound 8, a second-generation, higher
affinity compound, containing a methylated benzamide template in the
context of a pTyr group, was reported [27]. A literature procedure [27] was
used to synthesize this compound (12, AP21733) [16], and a 2.5 A˚ x-ray
crystal structure of Lck SH2 (S164C), a protein homologue of Src SH2,
complexed with AP21733, was obtained (M. H. Hatada, unpublished
results). The proposed S-configuration of the benzylic methyl stereocenter
of AP21733 was confirmed through independent asymmetric synthesis
[28]. The Lck SH2–nonpeptide structure reveals adherence to the historical
pTyr-Glu-Glu-Ile interactions in the pY pocket and shows carboxamide
44
Metcalf and Sawyer
Figure 13 Series of de novo designed nonpeptides containing a benzamide
template (exemplified by compound 12, AP21733) designed to interact favorably
with Src SH2 and specifically to displace structural waters found in complexed
Src SH2 structures [14,27]. The Src SH2 binding IC50 is shown for each
compound, as well as a comparative IC50 for Ac-pTyr-Glu-Glu-Ile-NH2
(compound 9).
contacts with Lys182 (206 in Src) and Ile193 (217 in Src). The phenyl ring
of the benzamide template also forms favorable stacking interactions with
Tyr181 (205 in Src). Although the cyclohexylmethyl group interacts with
the pY+3 pocket, the contacts are primarily surface type and do not
extend as deeply into the pocket as the Ile of pTyr-Glu-Glu-Ile. Consequently, SAR exploration of the pY+3 pocket, which had not been
rigorously studied with nonpeptide (peptidomimetic) small molecules
[13,14], became the first objective to be investigated.
Parallel synthesis provides the means of rapidly preparing discrete
analogues for both lead generation and lead optimization strategies,
which makes it an attractive option for developing compound databases
for therapeutic targets. Furthermore, the incorporation of structure-based
methods into the design and evaluation of parallel synthetic libraries has
proven to be a successful strategy for integrating the two drug discovery
technologies [29]. For the synthesis of the benzamide-containing compounds, we devised a hitherto unreported solid phase, parallel synthetic
45
route focusing on pY+3 derivatives based on compound 13 and AP21733
[30]. Our synthetic philosophy adopts an integrated solid and solution
phase strategy that differs from the traditional unidirectional approach by
recognizing the strengths and limitations of each synthetic method and
then devising a route accordingly (Fig. 14) [31]. In addition, this strategy
provides chemical flexibility to incorporate, within the compound’s
molecular design, the necessary functional group complexity dictated by
our structure-based methods. The importance of the carboxamide group
guided our decision to exploit this functionality both as a solid support
attachment site and as a conserved binding element. A Rink amide linkage
was chosen to provide facile coupling of the template, via its benzoic acid,
and eventual generation of the critical benzamide binding moiety upon
cleavage from the solid support. The protected salicylic acid template 14
(synthesized using a modification of the solution phase literature procedure) [27] was coupled to Rink amide AM resin by means of standard
protocols (EDC/HOBt) to provide the benzamide-linked resin 15
Figure 14 Parallel synthetic approaches demonstrating a traditional (unidirectional) strategy and a multifaceted, integrated strategy; the latter utilizes both
solid and solution phase reactions.
46
Metcalf and Sawyer
(Scheme 3) [30]. The pY+3 diversity alcohols (R1)-OH (Fig. 15) were
attached to the template through a Mitsunobu coupling to provide ether
derivatives of 16. Palladium-mediated Alloc deprotection followed by
amide formation using the phosphate-ester-containing diversity acids
(R2)-CO2H provided the fully coupled resin-bound products of 17.
Cleavage from the resin with 95% TFA/H2O, which also afforded benzyl
phosphate deprotection, followed by reversed-phase (RP) semipreparative
Scheme 3 Abbreviations: EDC, 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide)hydrochloride; HOBt, 1-hydroxybenzotriazole; DEAD, diethyl
azodicarboxylate.
47
Figure 15 The diversity alcohols (R1)-OH and carboxylic acids (R2)-CO2H used
to synthesize compounds represented by 18 and 19. (From Ref. 30.)
Table 5 Src SH2 Binding (FP) for Analogs of Compound 18
Source: Ref. 30.
48
Metcalf and Sawyer
HPLC purification generated the final compounds represented by 18
(mixture of diastereomers) and 19.
The selection of the pY+3 diversity R1 groups was guided by a FLO
docking program [32], utilizing 800 commercially available alcohols
(prefiltered by MW, H-bond donors, and reactive groups outside the
OH). The R1 groups were computationally incorporated [33] into the
benzamide template, docked into our Src SH2 binding site model [34], and
then rank-ordered according to favorable fit. The final list of alcohols was
Figure 16 The predicted binding mode of compound 23 in the pY+3 pocket of
the Src SH2 model. The branch point in the pY+3 bisallyl group allows favorable
binding interactions to occur. (From Ref. 30.)
49
selected according to predicted binding as well as the ability of the R1
group to impart beneficial properties to the molecule as related to low
molecular weight, increase solubility, and other factors.
Table 5 contains the Src SH2 binding results for a selected set of
pY+3-modified nonpeptide analog. Relative to compound 20, which was
synthesized by means of our solid phase method to act as an internal
standard, increases in binding affinity appeared to track the degree of
hydrophobicity at the R1 group as demonstrated by compounds 21
(methyl) and 22 (isopropyl). From a drug design perspective, the result
of 22 is significant because a four-carbon reduction took place, relative to
the cyclohexylmethyl group (MW decrease by 54), without greatly compromising the binding affinity (four-fold).
An extension of the a-branch point of the isopropyl group to a
bisallyl resulted in the highest affinity analog, compound 23. Inspection of
the docked structure of 23 in our Src SH2 model reveals how the branch
point allows one allyl side chain to hug the surface of the protein, while the
other is able to extend deeply into the pY+3 pocket (Fig. 16). A significant
decrease in binding affinity occurs with the incorporation of a morpholine
group, as exemplified by compound 24. Presumably, this result reflects an
incompatibility of the positively charged morpholine group (at pH 7.2 of
the binding assay) in the hydrophobic pY+3 binding pocket of the Src
SH2 domain; structurally, the pY+3 pocket according to our Src SH2
model accommodates this compound.
VII.
DISCOVERY OF AN IN VIVO EFFECTIVE Src
SH2 INHIBITOR
The next logical step in the progression to a cellularly active Src SH2
inhibitor was to incorporate a high affinity, biologically stable pTyr
mimic into the benzamide template. Drug design efforts at ARIAD led to
a novel Src SH2 inhibitors containing 4-diphosphonomethylphenylalanine (Dmp), namely, compound 25 (AP21773; Fig. 17) [16]. The design
concept for the Dmp group evolved from a 1.5 A˚ x-ray structure of Src
SH2, crystallized from citrate buffer, that fortuitously contained a citrate
molecule bound in the pTyr pocket. The x-ray structure reveals a number
of additional hydrogen bonds that citrate makes compared with a pTyr
group; this inspired the design of the Dmp moiety as a novel mimic of the
citrate interactions. Armed with these designed hydrogen bond contact
50
Metcalf and Sawyer
Figure 17 Src SH2 binding IC50 (Fp) for compound 25 (AP21773), which
contains a bone-targeted, 4-diphosphonomethylphenylalanine (Dmp) pTyr
mimic. (From Ref. 16.)
groups, we expected the Dmp to bind with greater affinity than pTyr, and
the Src SH2 binding results for AP21773 (Dmp) and AP21733 (pTyr)
confirm this prediction (Figs. 13 and 17). X-ray and NMR structural
studies involving AP21773 [16] verify these additional Dmp-related
contacts in the pTyr pcket, as well as other key Src SH2 interactions
observed earlier with this benzamide class as already discussed. The Dmp
moiety not only increases Src SH2 binding affinity, but also provides a
mechanism for tissue selectivity by targeting bone [16,35]. This targeting
feature provides a higher local concentration of compound on bone than
Figure 18 Solid phase synthetic scheme and molecular diversity groups for
compound 27. (From Ref. 36.)
51
Table 6 Src SH2 Binding (FP), Rabbit Pit, and Rat TPTX Data for Analogs of
Compounds 27 and 35 (AP22209)
52
Metcalf and Sawyer
in solution, which in addition increases the amount of Src inhibitor
delivered to the resorbing osteoclasts associated with bone. Compounds
containing the Dmp group, including AP21773, also bind to hydroxyapatite (data not shown), a major component of bone [16].
Building on the SAR information obtained from the pY+3 study,
we focused on improving binding affinity and cellular potency by means of
structure-based, parallel synthesis. A resin-bound, enantiomerically
enriched benzamide template 26 (Fig. 18) [28] was synthetically elaborated
in a manner similar to that described in Scheme 3 to provide the desired
Dmp-containing products. A total of 22 structurally biased analogues of
27 were generated (not all combinations synthesized) having specific R1
and R2 groups as shown in Figure 18 [36]. Table 6 shows the SAR results
for a selected series of the benzamide analogues. Similar to the earlier
study, increasing hydrophobicity at the pY+3 position leads to increased
binding affinity, as demonstrated by compounds 28 to 30. Interestingly,
the overall effect on Src SH2 binding of the 3-pentyl group of compound
30 appears to be similar to that of the cyclohexylmethyl group of
AP21773, although the latter group contains two more carbon atoms.
All the derivatives show an approximately 5- to 10-fold reduction in Src
SH2 binding affinity with no substitution (R2 = H, compounds 31–34) at
Figure 19 Solid-phase synthetic scheme and molecular diversity groups for
compound 36.
53
the R2 position. Finally, in an effort to select compounds for testing in an
in vivo thyroparathyroidectomized (TPTX) animal model [37,38], Src SH2
inhibitors were evaluated in a cell-based resorption assay mediated by
rabbit osteoclasts. A potent compound, 35 (AP22209; Table 6), was discovered and showed significant bone resorption inhibition in test animals
(55% inhibition at 25 mg/kg b.i.d.), thus providing in vivo validation for
an Src SH2 inhibitor (C. A. Metcalf III, unpublished results).
A recent series of proprietary, nonpeptide Src SH2 inhibitors synthesized by our solid phase, parallel synthetic method is outlined in Figure 19.
This inhibitor series was based on a set of compounds disclosed earlier
[35,39,40] and containing a novel, high-affinity bicyclic benzamide template designed to interact favorably with the hydrophobic Tyr205 Src SH2
protein residue. A bone-targeting, 3,4-diphosphonophenylalanine (Dpp)
mimic of pTyr was also incorporated [35,40]. The Dpp moiety can be
correlated to both pTyr and citrate (Fig. 20). The biological data for the
library analogs of 36 will be described elsewhere.
Figure 20 Representation of the design rationale for two novel, bone-targeting
pTyr mimics, Dmp and Dpp, relative to an x-ray structure [16] of citrate
complexed with Src SH2.
54
Metcalf and Sawyer
VIII. CONCLUSION
The increasing number of therapeutic targets available to drug discovery
programs has challenged chemists to devise new and efficient strategies for
the advancement of lead compounds to clinical candidate status. One
evolving approach, as described in this chapter, is the integration of
synergistic technologies (e.g., structure-based drug design and combinatorial chemistry) into a focused program that emphasizes the strengths of
each individual method. We have used this philosophy to direct our Src
SH2 program toward achieving novel proprietary Src SH2 inhibitors such
as AP22209, which exhibit promising antiresorptive activity both in an in
vivo animal model and in cell-based osteoclast assays. The use of structure-based, small-molecule libraries allowed us to rationally design compounds relative to predicted binding interactions, while taking advantage
of parallel synthesis to rapidly advance lead optimization. By adopting a
synthetic strategy that utilizes both solid and solution phase chemistries,
we were able to achieve the necessary chemical purity and diversity for
SAR interpretation at all stages of the drug discovery process. This
integrated drug design and combinatorial chemistry strategy is currently
being adapted to other drug discovery programs at ARIAD.
ACKNOWLEDGMENTS
The authors thank all our colleagues at ARIAD Pharmaceuticals,
including Chi Vu, Virginia Jacobsen, Michael Yang, William Shakespeare, Regine Bohacek, Joseph Eyermann, Berkley Lynch, Shelia
Violette, and Manfred Weigele, and especially Mayumi Uesugi, Vaibhav
Varkhedkar, and Chad Haraldson, whose contributions were significant
to the success of this work. We also thank Chris Stearns for her help with
the figures, David Dalgarno for his editorial suggestions, and Jay
LaMarche for allowing us to buy all our expensive toys.
REFERENCES
1. Sawyer TK. Molecular, structural, and chemical biology in pharmaceutical
research. In: Walker JM, Rapley R, eds. Molecular Biology and
Biotechnology, 4th ed. London: Humana Press, 2000:237–260.
2. The reader is referred to two landmark papers discussing, respectively, the
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
55
disruption and promotion of protein–protein interactions in signal transduction and the resulting therapeutic implications: Brugge JS. New intracellular targets for therapeutic drug design. Science 1993; 260:918–919, and
Spencer DM, Wandless TJ, Schreiber SL, Crabtree GR. Controlling signal
transduction with synthetic ligands. Science 1993; 262: 1019–1024.
These are considered to be functionally competent protein domains owing to
their ability to maintain their structure and molecular recognition properties
independent of the full-length protein (Cohen GB, Ren R, Baltimore D.
Modular binding domains in signal transduction proteins. Cell 1995; 80:
237–248).
Botfield MC, Green J. SH2 and SH3 domains: choreographers of multiple
signaling pathways. Ann Rep Med Chem 1995; 30:227–237.
Koch CA, Anderson D, Moran MF, Ellis C, Pawson T. SH2 and SH3
domains: elements that control interactions of cytoplasmic signaling
proteins. Science 1991; 252:668–674.
Songyang Z, Shoelson SE, Chaudhuri M, Gish G, Pawson T, Haser WG,
King F, Roberts T, Ratnofsky S, Lechleider RJ, Neel BG, Birge RB, Fajardo
JE, Chou MM, Hanafusa H, Schaffhausen B, Cantley LC. SH2 domains
recognize specific phosphopeptide sequences. Cell 1993; 72:767–778.
Xu W, Harrison SC, Eck MJ. Three-dimensional structure of the tyrosine
kinase c-Src. Nature 1997; 385:595–602. See also: Xu W, Doshi A, Lei M,
Eck MJ, Harrison SC. Crystal structures of Src reveal features of its
autoinhibitory mechanism. Mol Cell 1999; 3:629–638.
For recent reviews describing Src’s role in the biological processes involving
bone disease, see: Sawyer T, Boyce B, Dalgarno D, Iuliucci J. Src inhibitors:
genomics to therapeutics. Expert Opin. Invest. Drugs 2001; 10:1327–1344,
and Schwartzberg PL. Src in the osteoclast — a potential therapeutic target?
Jikken Igaku 1998; 16(11):1386–1393.
Songyang Z, Shoelson SE, McGlade J, Oliver P, Pawson T, Bustelo XR,
Barbacid M, Sabe H, Hanafusa H, Yi T, Ren R, Baltimore D, Ratnofsky S,
Feldman RA, Cantley LC. Specific motifs recognized by the SH2 domains
of Csk, 3BP2, fps/fes, GRB-2, HCP, SHC, Syk, and Vav. Mol Cell Biol
1994; 14:2777–2785.
Waksman G, Kominos D, Robertson SR, Pant N, Baltimore D, Birge RB,
Cowburn D, Hanafusa H, Mayer BJ, Overduin M, Resh MD, Rios CB,
Silverman L, Kuriyan J. Crystal structure of the phophotyrosine recognition
domain SH2 of v-src complexed with tyrosine-phosphorylated peptides.
Nature 1992; 358:646–653.
Waksman G, Shoelson SE, Pant N, Cowburn D, Kuriyan J. Binding of a
high affinity phosphotyrosyl peptide to the Src SH2 domain: crystal
structures of the complexed and peptide-free forms. Cell 1993; 72:779–790.
Gilmer TM, Rodriguez M, Jordan S, Crosby R, Alligood KJ, Green M,
Kimery M, Wagner C, Kinder D, Charifson P, Hassell AM, Willard D,
56
13.
14.
15.
16.
17.
18.
19.
20.
Metcalf and Sawyer
Luther M, Rusnak D, Sternbach DD, Mehrotra M, Peel M, Shampine L,
Davis R, Robbins J, Patel IR, Kassel D, Burkhart W, Moyer M, Bradshaw
T, Berman J. Peptide inhibitors of Src SH3–SH2–phosphoprotein interactions. J Biol Chem 1994; 269:31711–31719. Also, for pTyr mimics see: Ye
B, Akamatsu M, Shoelson SE, Wolf G, Giorgetti-Peraldi S, Yan X, Roller
PP, Burke TR Jr. L-O-(2-Malonyl)tyrosine: a new phosphotyrosyl mimetic
for the preparation of Src homology 2 domain inhibitory peptides. J Med
Chem 1995; 38:4270–4275.
Plummer MS, Holland DR, Shahripour A, Lunney EA, Fergus JH, Marks
JS, McConnell P, Mueller WT, Sawyer TK. Design, synthesis, and cocrystal
structure of a nonpeptide Src SH2 domain ligand. J Med Chem 1997;
40:3719–3725.
Lunney EA, Para KS, Rubin JR, Humblet C, Fergus JH, Marks JS, Sawyer
TK. Structure-based design of a novel series of nonpeptide ligands that bind
to the pp60c-src SH2 domain. J Am Chem Soc 1997; 119:12471–12476.
Eck MJ, Shoelson SE, Harrison SC. Recognition of a high-affinity
phosphotyrosyl peptide by the Src homology-2 domain of p56lck. Nature
1993; 362:87–91.
Bohacek RS, Dalgarmo DC, Hatada M, Jacobsen VA, Lynch BA, Macek KJ,
Merry T, Metcalf CA III, Narula SS, Sawyer TK, Shakespeare WC, Violette
SM, Weigele M. X-ray structure of citrate bound to Src SH2 leads to a high
affinity, bone-targeted Src SH2 inhibitor. J Med Chem 2001; 44:660–663.
Bibbins KB, Boeuf H, Varmus HE. Binding of the Src SH2 domain to
phosphopeptides is determined by residues in both the SH2 domain and the
phosphopeptides. Mol Cell Biol 1993; 13:7278–7287.
For a discussion on the use of prodrugs to address bioreversibility and cell
permeability for inhibitors of Src SH2, see: Rickles RJ, Henry PA, Guan W,
Azimiora M, Shakespeare W, Violette S, Zoller MJ. A novel mechanismbased mammalian cell assay for the identification of SH2-domain-specific
protein–protein inhibitors. Chem Biol 1998; 5:529–538, and Stankovic CJ,
Surendran N, Lunney EA, Plummer MS, Para KS, Shahripour A, Fergus
JH, Marks JS, Herrera R, Hubbell SE, Humblet C, Saltiel AR, Stewart B,
Sawyer TK. The role of 4-phosphonodifluoromethyl- and 4-phosphonophenylalanine in the selectivity and cellular uptake of SH2 domain ligands.
Bioorg Med Chem Lett 1997; 7:1909–1914.
Burke TR Jr, Smyth M, Nomizu M, Otaka A, Roller PP. Preparation of
fluoro- and hydroxy-4-(phosphonomethyl)-D,L-phenylalanine suitably protected for a solid-phase synthesis of peptides containing hydrolytically stable
analogues of O-phosphotyrosine. J Org Chem 1993; 58:1336–1340.
For recent papers describing the use of pTyr mimics in the context of Src
and Lck SH2 inhibitors, see: (a) Kawahata N, Yang MG, Luke GP,
Shakespeare WC, Sundaramoorthi R, Wang Y, Johnson D, Merry T,
Violette S, Guan W, Bartlett C, Smith J, Hatada M, Lu X, Dalgarno DC,
21.
22.
23.
24.
25.
26.
27.
28.
29.
57
Eyermann CJ, Bohacek R, Sawyer TK. A novel phosphotyrosine mimetic
4V-carboxymethyloxy-3V-phosphonophenylalanine (CCP): exploitation in the
design of nonpeptide inhibitors of pp60src SH2 domain. Bioorg Med Chem
Lett 2001; 11:2319–2323. (b) Lesuisse D, Deprez P, Bernard D, Broto P,
Delettre G, Jean-Baptiste V, Marquette JP, Sarubti E, Schoot B, Mandine
E, Lange G. SAR by crystallography: a new approach combining screening
and rational drug design. Application to the discovery of nanomolar Src
SH2 binders. Chim Nouv 2001: 19:3240–3241. (c) Beaulieu PL, Cameron
DR, Ferland J-M, Gauthier J, Ghiro E, Gillard J, Gorys V, Poirier M,
Rancourt J, Wernic D, Llinas-Brunet M, Betageri R, Cardozo M, Hickey
ER, Ingraham R, Jakes S, Kabcenell A, Kirrane T, Lukas S, Patel U,
Proudfoot J, Sharma R, Tong L, Moss N. Ligands for the tyrosine kinase
p56lck SH2 domain: discovery of potent dipeptide derivatives with monocharged, nonhydrolyzable phosphate replacements. J Med Chem 1999;
42:1757–1766.
For examples of x-ray structures of nonpeptide compounds complexed with
Src SH2, see Refs. 13 and 14, and for Lck SH2, see Ref. 16.
Metcalf CA III, Vu CB, Sundaramoorthi R, Jacobsen VA, Laborde EA,
Green J, Green Y, Macek KJ, Merry TJ, Pradeepan SG, Uesugi M,
Varkhedkar VM, Holt DA. Novel phosphate ester-linked resins: the solidphase generation of phenyl phosphate-containing compounds for SH2
inhibition. Tetrahedron Lett 1998; 39:3435–3438.
Meyers HV, Dilley GJ, Durgin TL, Powers TS, Winssinger NA, Zhu H,
Pavia MR. Multiple simultaneous synthesis of phenolic libraries. Mol
Diversity 1995; 1:13–20.
Lynch BA, Loiacono KA, Tiong CL, Adams SE, MacNeil IA. A
fluorescence polarization based Src-SH2 binding assay. Anal Biochem
1997; 247:77–82.
Cyberlab, Inc., Brookfield, CT.
Backes I, Ellman JA. Solid support linker strategies. Curr Opin Chem Biol
1997; 1:86.
Lunney EA, Para KS, Plummer MS, Prasad JVNV, Saltiel AR, Sawyer TK,
Shahripour A, Singh I, Stankovic CJ. Compounds, composition and
methods for inhibiting the binding of proteins containing an SH2 domain
to cognate phosphorylated proteins. PCT Int Appl WO97/12903, 1997.
Luke GP, Holt DA. Synthesis of (S)-5-(1-aminoethyl)-2-(cyclohexylmethoxy)
benzamide. Tetrahedron: Asymmetry 1999; 10:4393–4403.
For examples of structure-based methods used in combinatorial chemistry,
see: (a) Chang Y-T, Gray NS, Rosania GR, Sutherlin DP, Kwon S, Norman
TC, Sarohia R, Leost M, Meijer L, Schultz PG. Synthesis and application of
functionally diverse 2,6,9-trisubstituted purine libraries as CDK inhibitors.
Chem Biol 1999; 6:361–375. (b) Antel J. Integration of combinatorial
chemistry and structure-based drug design. Curr Opin Drug Discovery Dev
58
30.
31.
32.
33.
34.
35.
Metcalf and Sawyer
1999; 2:224–233. (c) Rohrer SP, Birzin ET, Mosley RT, Berk SC, Hutchins
SM, Shen D-M, Xiong Y, Hayes EC, Parmar RM, Foor F, Mitra SW,
Degrado SJ, Shu M, Klopp JM, Cai S-J, Blake A, Chan WWS, Pasternak A,
Yang L, Patchett AA, Smith RG, Chapman KT, Schaeffer JM. Rapid
identification of subtype-selective agonists of the somatostatin receptor
through combinatorial chemistry. Science 1998; 282:737–740. (d) Szardenings KA, Harris D, Lam S, Shi L, Tien D, Wang Y, Patel DV, Navre M,
Campbell DA. Rational design and combinatorial evaluation of enzyme
inhibitor scaffolds: identification of novel inhibitors of matrix metalloproteinases. J Med Chem 1998; 41:2194–2200. (e) Kick EK, Roe DC,
Skillman AG, Liu G, Ewing TJA, Sun Y, Kuntz ID, Ellman JA. Structurebased design and combinatorial chemistry yield low nanomolar inhibitors of
cathepsin D. Chem Biol 1997; 4:297–307. (f) Combs AP, Kapoor TM, Feng
S, Chen JK, Daude-Snow LF, Schreiber SL. Protein structure-based
combinatorial chemistry: discover of non-peptide binding elements to Src
SH3 domain. J Am Chem Soc 1996; 118:287–288.
Metcal CA III, Eyermann CJ, Bohacek RS, Haraldson C, Varkhedkar VM,
Lynch B, Bartlett C, Violette S, Sawyer TK. Structure-based design and solidphase parallel synthesis of phosphorylated nonpeptides to explore hydrophobic binding at the Src SH2 domain. J Comb Chem 2000; 2:305–313.
Such a strategy follows the philosophy of ‘‘resin capture,’’ as introduced by
Armstrong and Keating: Keating TA, Armstrong RW. Postcondensation
modifications of Ugi four-component condensation products: 1-isocyanocyclohexane as a convertible isocyanide. Mechanism of conversion,
synthesis of diverse structures, and demonstration of resin capture. J Am
Chem Soc 1996; 118:2574–2583. See also: Brown SD, Armstrong RW.
Synthesis of tetrasubstituted ethylenes on solid support via resin capture.
J Am Chem Soc 1996; 118:6331–6332.
FLO97, Graphics and Molecular Mechanics Software for Drug Design.
Available from Colin McMartin: cmcma@ix.netcom.com.
McMartin C, Bohacek RS. QXP: powerful, rapid computer algorithms
for structure-based drug design. J Comput-Aided Mol Design 1997; 11:
333–344.
The Src SH2 binding site model used in this study was developed based on
a high resolution (1.0 A˚) crystal structure of Lck SH2 complexed with
Ac-pTyr-Glu-Glu-Ile-NH2, obtained from the Brookhaven Protein Data
Bank (reference code 1LKK). For a description of the X-ray structure, see:
Tong L, Warren TC, King J, Betageri R, Rose J, Jakes S. Crystal structures
of the human p56lck SH2 domain in complex with two short phosphotyrosyl
peptides at 1.0 A˚ and 1.8 A˚ resolution. J Mol Biol 1996; 256(3):601–610.
Violette SM, Guan W, Bartlett C, Smith JA, Bardelay C, Antoine E, Rickles
RJ, Mandine E, van Schravendijk MR, Adams SE, Lynch BA, Shakespeare
WC, Yang M, Jacobsen VA, Takeuchi CS, Macek KJ, Bohacek RS,
36.
37.
38.
39.
40.
59
Dalgarno DC, Weigele M, Lesuisse D, Sawyer TK, Baron R. Bone-targeted
Src SH2 inhibitors block Src cellular activity and osteoclast-mediated
resorption. Bone 2001; 28:54–64.
Weigele M, Bohacek R, Jacobsen VA, Macek K, Yang MG, Kawahata NH,
Sundaramoorthi R, Wang Y, Takeuchi CS, Luke GP, Metcalf CA III,
Shakespeare WC, Sawyer T. Synthesis of phosphono-containing amino acid
derivatives and peptides as signal transduction inhibitors. PCT Int Appl
W099/24442, 1999.
Frost HM, Jee WSS. On the rat model of human osteopenias and
osteoporosis. Bone Miner 1992; 18:227–236.
Green JR, Muller K, Jaeggi KA. Preclinical pharmacology of CGP 42V446, a
new, potent, heterocyclic biphosphonate compound. J Bone Miner Res
1994; 9:745–751.
Violette SM, Shakespeare WC, Bartlett C, Guan W, Smith JA, Rickles RJ,
Bohacek RS, Holt DA, Baron R, Sawyer TK. A Src SH2 selective binding
compound inhibits osteoclast-mediated resorption. Chem Biol 2000; 7:
225–235.
Shakespeare WC, Yang M, Bohacek R, Cerasoli F, Stebbins K,
Sundaramoorthi R, Azimioara M, Vu C, Pradeepan S, Metcalf CA III,
Haraldson C, Merry T, Dalgarno D, Narula S, Hatada M, Lu X, van
Schravendijk M, Adams S, Violette S, Smith J, Guan W, Bartlett C, Herson
J, Iuliucci J, Weigele M, Sawyer T. Structure-based design of an osteoclastselective, nonpeptide Src homology 2 inhibitor with in vivo antiresorptive
activity. Proc Natl Acad Sci USA 2000; 97:9373–9378.
3
Three-Dimensional Structure of
the Inhibited Catalytic Domain
of Human Stromelysin-1 by
Heteronuclear NMR Spectroscopy
Paul R. Gooley
University of Melbourne, Parkville, Victoria, Australia
I.
INTRODUCTION
With the aid of isotopic enrichment it is now routine to determine the
structure of moderate to large proteins (20 to 40 kDa) by multidimensional
heteronuclear nuclear magnetic resonance (NMR) spectroscopy [1]. The
advantages these heteronuclear experiments offer are spectral simplification and a reduced dependence on narrow proton linewidths. By spreading
the 1H– 1H correlations of a 2-D NMR spectrum into a third and, perhaps,
a fourth dimension, according to the chemical shift of the attached 13C or
15
N nucleus, considerable spectral simplification is achieved. As the proton
of interest is now correlated with its bound 13C or 15N, the information
content for assignment is increased, and as the individual planes of the 3-D
or 4-D spectra contain relatively fewer overlapping peaks, problems with
assignment ambiguities are reduced. These experiments are more efficient
than their homonuclear counterparts because transfer of magnetization
relies on the large one-bond heteronuclear couplings (11 to 140 Hz). The
first stage in solving the structure of a protein requires the acquisition of a
large number of three-dimensional experiments for sequence-specific as-
61
62
Gooley
signment of backbone and side-chain atoms. Typically, a dozen or more 2D and 3-D experiments will need to be acquired, taking a total of 4 to 6
weeks of spectrometer time, thus requiring long-term sample stability or
readily available large milligram quantities. The second stage in solving a
structure remains largely dependent on the acquisition of NOE spectra and
the assignment of many interresidue NOEs. In the special case of a protein
complex, for example, an inhibited enzyme, where one component is
isotopically labeled, the spectrum of the complex can be separated into
subspectra of the two components to aid assignment of both protein and
ligand resonances and most importantly, to determine protein-ligand
contacts [2]. Using these methods, the spectrum of the ligand can be solved
readily providing important information about the conformation of the
ligand in the bound state.
The application of these techniques relies on an abundant source of
13
C- and 15N-enriched proteins and, therefore, the application of heteronuclear NMR spectroscopy to solving the solution structure of proteins
has relied on advances in molecular biology. Many proteins can be overexpressed and isotopically enriched with 13C and 15N by replacing the
carbon source with 13C-glucose and the nitrogen source with a 15NH4+
salt. Efficient enrichment is possible with media supplemented with minerals and vitamins, and using fermentation protocols [3]. Molecular biology
has further contributed to the number of proteins that can be studied by
NMR spectroscopy by overexpressing the catalytic or functional domains
of large proteins, thus truncating the protein to a size (often less than 25
kDa) that is readily amenable to these techniques [4]. This chapter discusses the implementation of these methods to the catalytic domain of
human stromelysin-1 (sfSTR), a matrix metalloendoproteinase (MMP),
complexed to a N-carboxylalkyl inhibitor [5 –8] (Fig. 1). We will focus on
the work where the structure of the protein complex was determined and
compare this structure to other inhibited MMP catalytic domains.
II. THE MATRIX METALLOPROTEINASE FAMILY
The matrix metalloendoproteinases (MMPs or matrixins) are a family of
zinc and calcium dependent extracellular proteases that collectively
degrade most of the protein constituents of the extracellular matrix [9].
There are at least 23 members of this family and are divided primarily on
the basis of sequence homology and substrate specificity into the
following grouping: collagenases (MMP-1, -8, -13, -18) gelatinases
3-D Structure of Stromelysin-1
63
Figure 1 N-carboxylalkyl inhibitor of sfSTR, N-[(R)-carboxyl-ethyl]-(S)-(2phenylethyl) glycyl-L-arginine-N-phenylamide [8]. The convention of Schechter
and Berger [61] is used to describe the specificity subsites of the enzyme S1V, S2V,
S3V which correspond to the side chains P1V, P2V, P3V of the inhibitor. (From Ref. 5.)
(MMP-2, -9), stromelysins (MMP-3, -7, -10, -11), membrane-associated
(MMP-14, -15, -16, -17), and a group of ‘‘others’’ including macrophage
elastase (MMP-12) and enamelysin (MMP-20). These enzymes participate in many normal biological processes such as embryonic development, bone remodeling, wound healing, angiogenesis, and apoptosis.
The pharmaceutical industry has shown considerable attention to the
regulation of these enzymes because pathological proteolysis by these
enzymes accompanies many degradative diseases including arthritis,
ulcerations (corneal, gastric, skin), and periodontal diseases. Degradation
of the basement membrane by one or more of the MMPs is clearly
essential in tumor progression. While the activity of the MMPs is
controlled by endogenous inhibitors such as a-macroglobulin and tissue
inhibitor of metalloproteases (TIMP-1 and -2), the disease state may be
a consequence of an imbalance in the ratio of protease to protease inhibitor. In the disease state, a potent synthetic inhibitor may have therapeutic
effects by restoring the ratio of protease to protease inhibitor to normal
physiological levels.
The MMPs are synthesized as preproproteins and are secreted as
latent proproteins. Most MMPs share a common domain structure of a
propeptide (about 80 amino acids) that has a conserved cysteine ligated to
the catalytic zinc thus maintaining latency [10], a catalytic domain (about
180 amino acids), and a C-terminal domain (about 210 amino acids)
64
Gooley
(Fig. 2). Not all MMPs have these domains, for example, matrilysin lacks
the C-terminal domain and the MT-MMPs do not have the propeptide.
Some MMPs have additional domains, for example, the gelatinases have
three repeats of a fibronectin-type II domain.
The C-terminal hemopexin-like domain has been structurally characterized as a four-bladed h-propellor [11 – 13]. While the structural
homology of this domain is clear, it has varied functions, for example, it
is essential for cleavage of triple helical collagens by the collagenases [14];
however, it is required for activation of pro-MMP-2 by MT1-MMP [15].
The catalytic domain and the C-terminal domain are connected by a
proline-rich linker that modeling experiments suggest may play a role in
recognition and destabilization of collagen [16].
Figure 2 Domain structure of the MMPs: 92 kDa gelatinase-A (MMP-2), 72 kDa
gelatinase-B (MMP-9), the collagenases (MMP-1, -8, and – 13), stromelysin-1
(MMP-3) and matrilysin (MMP-7). Matrilysin is the only known MMP that does
not have a C-terminal hemopexin-like domain.
65
The catalytic domain has been the focal point for drug discovery.
This domain contains the motif HEXXHXXGXXH, which ligates a
catalytic zinc, and a characteristic h-turn that contains a conserved
methionine that is structurally located near the catalytic zinc. These
structural properties have led to the classification of the MMPs with
several other families of metalloproteases (astacins, serralysins, and reprolysins) as ‘‘metzincins’’ [17]. The catalytic domain further contains a
structural zinc and several calcium ions necessary for stability. The
catalytic domain of several MMPs has been expressed in bacterial systems
either as a soluble protein or in inclusion bodies [18, 19]. The proteins have
been purified and/or refolded and shown to be fully active against small
peptide substrates. Structures of the catalytic domains have been solved by
both x-ray crystallography and NMR spectroscopy for collagenase-1, -2
and -3 [20 –25], stromelysin-1 [6,10,26], and matrilysin [27]. Furthermore,
the pro-form of stromelsyin-1 [10], and the full-length proteins collagenase-1 [11] and progelatinase A [28] have been solved by x-ray crystallography. The structure of the catalytic domains appears identical in the
truncated forms to that in the full-length protein, and therefore the smaller
truncated form has been ideal for the structural analysis of the inhibited
forms to aid inhibitor and drug design. In the following discussion we
outline the strategy for determining the structure of the inhibited catalytic
domain of stromelysin-1 by NMR methods.
III. ASSIGNMENT OF THE RESONANCES
OF THE INHIBITED CATALYTIC DOMAIN
OF STROMELYSIN-1
The assignment of the 1H, 13C, and 15N resonances depends on acquiring
a large number of separate 3-D or 4-D triple resonance experiments. The
experiments can be divided into intraresidue and interresidue and, when
combined, lead to sequence-specific assignment through bonds (Fig. 3) [1].
Improvements and new pulse sequences continue; however, a common set
of experiments to assign the backbone (frequently defined as Ha, Ca, N,
HN, C’, Ch) resonances of a protein are: 3-D HNCACB, CBCA(CO)NH,
HCACO, HNCO, (HCA)CO(CA)NH, 4-D HCANNH and HCA(CO)
NNH [29 –32]. Side-chain resonances are assigned using HCCH-TOCSY
and HCCH-COSY [33]. Unambiguous stereospecific assignment of the
methyl groups of Leu and Val are possible by preparing a 10% 13C-labeled
sample and acquiring a 1H,13C-HSQC spectrum [34]. The incorporation
66
Gooley
Figure 3 3-D and 4-D triple resonance experiments correlate interresidue or
intraresidue nuclei. (A) The efficiency of the experiments depend on the large
one-bond couplings; (B) atoms correlated in the 4-D HCANNH, an example of
an intraresidue experiment; (C) the 4-D HCA(CO)NNH, an interresidue experiment; (D) the 3-D HNCO, an interresidue experiment; and (E) the 3-D HCACO,
an intraresidue experiment.
of label is nonrandom such that Leu and Val residues are labeled as
13
Cy2H3-12CgH, 13Cy1H3-13CgH, and 13Cg2H3-12ChH, 13Cg1H3-13ChH,
respectively. Consequently, the 13Cy2H3 of Leu and 13Cg2H3 of Val
groups appear as singlets in the 1H,13C HSQC spectra and are thus readily stereoassigned. Measurement of 3JHNa in 3-D HNHA spectra [35] aids
determination of f torsion angles and stereoassignment of h-methylene
groups requires 3-D HNHB [36] and HACAHB [37] experiments. To
determine the fold of the protein, a large number of interresidue NOEs
must be assigned in 3-D 15N-NOESY [38], 3-D and 4-D 13C-NOESY experiments [39,40]. The assignment of the backbone resonances of the
13 15
C, N-enriched catalytic domain of stromelysin-1 were mostly accomplished with 4-D HCANNH and HCA(CO)NNH experiments (Fig.4) [5].
Side-chain atoms were assigned with 3-D HCCH-COSY and HCCHTOCSY experiments with the carrier located near 35 ppm for aliphatic
side chains and at 124 ppm for aromatic side chains. Stereospecific
assignment of the methyl groups were obtained with a 10% 13C-labeled
67
Figure 4 Sequential assignment of the backbone atoms for the segment Pro-109
to Val-113 of inhibited sfSTR by 4-D HCANNH and 4-D HCA(CO)NNH. Four
planes are shown from each spectrum. The assigned backbone atoms are
indicated in (A). In (B) the upper four planes in solid lines are from the 4-D
HCANNH and the lower four planes in dashed lines are from the 4-D
HCA(CO)NNH. The chemical shifts for the four correlated nuclei in each case
are shown. The correlations continue for the segment Pro-109 to Pro-129. As Pro
lacks a protonated N, this residue serves as a ‘‘stop’’ signal. The correlation of 19
residues with Pro at the N- and C-terminal ends is unique for this segment in the
sequence of sfSTR, therefore these backbone atoms are specifically assigned
without having to further assign side chain atoms. (From Ref. 5.)
68
Gooley
sample and a 2-D 1H, 13C HSQC spectrum, and couplings from 3-D
HNHA and HNHB experiments aided defining f and m1 angles. To
determine which specific ring nitrogens of the six histidine residues ligate
the catalytic and structural zincs a 2-D 1H, 15N HMQC spectrum was
acquired, where the delay for generating antiphase 1H, 15N magnetization
is set to 22 ms [41]. This experiment will favor the weak two- and threebond couplings between the histidine ring nitrogens and the Cy2H and
Cq1H protons (Fig. 5) and clearly showed the Nq2 of His-151, -166, 201,
-205 and -211, and the Ny1 of His-96 are the zinc ligands. Finally, the
critical NOEs that describe the tertiary structure of the protein were
assigned in 3-D 15N-NOESY, 3D and 4-D 13C-NOESY experiments.
IV.
ASSIGNMENT OF THE RESONANCES OF THE
INHIBITOR AND NOEs BETWEEN THE PROTEIN
AND THE INHIBITOR
To understand the interactions between a protein and a small ligand, we
take advantage of the fact that the protein is enriched with 13C and 15N and
the ligand is not. Pulse sequences can be designed to edit the spectrum of the
protein-ligand complex into spectra (2-D 1 H, 1 H-COSY, TOCSY,
NOESY) of the ligand or intermolecular NOEs between the labeled protein
and unlabeled ligand in either 2-D or preferably 3-D NOESY spectra.
These experiments are composed of X-half-filters [2,42,43] and either select
or filter the 13C,15N-attached protons (Fig. 6). Consequently, a 2-D 13C
doubly filtered NOESY spectrum will show intramolecular NOEs for the
ligand, whereas a 3-D 13C-filtered, 13C-selected NOESY will show intermolecular NOEs between the protein and the ligand. The latter experiment
is preferably acquired as a 3-D to minimize the ambiguities in assigning the
protons of the protein that are involved in the interaction with the ligand.
For the complex of the inhibited catalytic domain of stromelysin-1,
2-D doubly filtered 1H,1H COSY and TOCSY experiments performed
poorly. As these experiments depend on 1H, 1H couplings, the linewidths of
the stromelysin-inhibitor complex must be too large for efficient magnetization transfer. On the other hand, 2-D doubly filtered NOESY experiments acquired in 2H2O and H2O showed correlations for all protons of
the inhibitor (Fig. 7), and, as the inhibitor (Fig. 1) was quite simple, the
resonances were readily assigned. 3-D 13C-filtered, 13C-separated NOESY
experiments were also successfully acquired and assignment of these
NOEs were unambiguously obtained (Fig. 8).
69
Figure 5 Part of a 2-D 1H, 15N HMQC spectrum of inhibited stromelysin-1
where the ring nitrogens of His resonate. The delay where antiphase magnetization evolves is set to 22 ms thus favoring the weak two- and three-bond
couplings of Ny1 and Nq2 to the protons of Cy2H and Cq1H [41]. The
deprotonated (h-type) nitrogen typically resonates near 249 ppm. These
resonances for inhibited sfSTR are near 200 to 210 ppm, shifted upfield by
ligation to the zinc ions. For the stable Ny1-H tautomer two strong couplings are
observed from the deprotonated Nq2 nucleus to the Cy2H and Cq1H protons. For
the stable Nq2-H tautomer only one strong coupling is observed from the
deprotonated Ny1 nucleus to the Cq1H proton. For the imidazolium tautomer the
resonances of the ring nitrogens are both near 176 ppm and equivalent couplings
from these nitrogens to both Cy2H and Cq1H protons are observed. For inhibited
sfSTR, His-151, -166, -201, -205 and -211 are in the Ny1-H tautomer, His-179 is
in the Nq2-H tautomer and His-96 and -224 are in the imidazolium tautomer.
Specific labeling of the Ny1 nucleus supports these assignments [5]. (From Ref. 5.)
70
Gooley
Figure 6 X-half filters used for filtering or selecting 13C and 15N-attached
protons. Thick and thin closed rectangles are 180j and 90j pulses, respectively,
open rectangles are spin lock pulses. (A) A simple X-half filter (2). The delay H is
equal to (1/(2[1JXH]) where 1JXH is the one-bond coupling between proton and
either 13C (120 to 140 Hz) or 15N (95 Hz). The second 90j pulse is the editing
71
V. STRUCTURE CALCULATIONS
Peak intensity data from NOE experiments were accumulated and converted to interproton distances by calibrating against the expected short
distances in secondary structure elements. These data were complimented
with coupling constants determined from the 3-D HNHA and HNHB
experiments. A total of 2589 peaks were assigned in all NOE experiments.
After removal of nonconstraining and ambiguous NOEs, typically found
in mobile regions, 1814 meaningful restraints remained: 325 intraresidue,
429 sequential, 324 short-range (i+2 to i+5), 665 long range ( > i+5), and
71 intermolecular. Using a gridsearch program [44] 379 dihedrals (140 f,
140 c, 99 m1) were generated from sequential and intraresidue NOEs and
coupling constant data from HNHA and HNHB experiments. Structures
were calculated using the variable target function algorithm DIANA [45],
but it should be noted that in recent years this method has been replaced by
torsion angle dynamics methods that are far more efficient [46,47]. To
determine the structure of the complex, a residue template of the inhibitor
was built as a single residue covalently linked through an oxygen of the
carboxylate moeity of the inhibitor (Fig. 1) to the zinc which was
covalently bonded to the Nq2 of His-201. The residue template of His151 was created with the structural zinc covalently attached to its Nq2
atom. The structure calculation process is largely iterative with trial structures calculated and incompatible NOEs reassigned or removed and new
NOEs assigned on the basis of agreement with the trial structure. In the
final calculations, and to reduce bias in structure selection, plots of rmsd
and number of structures versus target function [48] of the final 80
pulse. The phase cycling of this pulse with respect to the receiver determines
whether X-nucleus attached protons are selected or filtered. If both signals are
added to the receiver (x,x) X-nucleus attached protons are filtered; and if the
receiver phase is alternated (x,-x) the X-nucleus attached protons are selected. (B)
A doubly tuned half filter for filtering 13C attached protons [42]. In this experiment the filter consists of two delays (H 1,H 2) tuned to different 1JCH values
resulting in superior suppression of artifacts. (C) A doubly tuned time-shared half
filter for 13C/15N (43). In this experiment D = 1/(41JNH), D1 = 1/(41JCH) and D2 =
[1/(41JNH 1/(41JCH)]. Phase cycling the receiver selects or filters both 13C and
15
N attached protons.
72
Gooley
Figure 7 2-D 13C doubly filtered NOESY of inhibited sfSTR using the X-half
filters of Fig. 4A and B. The NOE correlations of the rings of the P1V and the P3V
group are shown. The Hh and Hg protons of the P1V group were distinguished in a
similar 2D 13C doubly filtered TOCSY. The specific assignment of the protons of
the P3V group were determined by NOEs between the H2,6 and the NH of the P3V
in 2D 13C,15N-filtered experiments using the time-shared doubly tuned half filter
of Fig. 4C.
73
Figure 8 Sections of a 3-D 13C-separated, 13C-filtered NOESY of inhibited
sfSTR. Only NOEs between the 12C-attached protons of the inhibitor and the
13
C-attached protons of the protein are observed in this spectrum. These NOEs
describe the S1V, S2V and S3V subsites of sfSTR. Not shown are several NOEs from
Val-197 and His-201 to the ethylene group of P1V. (From Ref. 6.)
structures were used to select structures for energy minimization using
the program FANTOM [49]. In the final calculations, 30 structures were
selected. Table 1 summarizes the DIANA and FANTOM statistics for
these structures.
VI. STRUCTURE OF INHIBITED STROMELYSIN-1
A. The Protein Fold
Superposition of residues 83 to 248 of the family of structures is shown
in Figure 9 viewed along the long axis of the catalytic helix. Residues 249
to 255 are disordered and therefore are not shown. In Figure 10 ribbon
drawings of two views of the molecule are shown, one from above the
h-sheet and the other from below the S1’ subsite. The secondary structure
of sfSTR consists of a five stranded h-sheet with four parallel strands and
74
Gooley
Table 1 Structural Statistics and Residual Violations of the 30 Conformers Used
to Represent the Solution Structure of the Inhibited Catalytic Domain of
Stromelysin-1
Parameter
DIANA
DIANA target function (A˚ )
FANTOM energy (kcal/mol)
Lennard-Jones energy (kcal/mol)
Distance constraint violations (A˚)
sum
maximum
rmsd
Exp. angle constraint violations (j)
sum
Maximum
Rmsd
Rmsd residues 83 – 250 (A˚)
backbone (Ca,N,C’,O)
all heavy atoms
2
10.01 F 0.76
FANTOM
191.0 F 52.8
605.4 F 48.4
35.2 F 1.2
0.48 F 0.06
0.06 F 0.01
61.6 F 1.4
0.39 F 0.03
0.08 F 0.01
93.4 F 11.1
7.5 F 0.9
0.90 F 0.08
112.1 F 19.9
12.6 F 5.0
1.2 F 0.3
0.48 F 0.06
0.94 F 0.06
0.55 F 0.06
0.97 F 0.05
Source: Ref. 7.
one antiparallel strand and the topology 1x, +2x, +2, 1, using
the Richardson nomenclature [50]. The h-sheet lies on two helices (helix
A and B); a third helix (helix C) is near the C-terminus. The molecule
has two zincs: a catalytic zinc is located at the bottom of a cleft, and
a structural zinc above the h-sheet. The overall fold of sfSTR may
be described as follows. The N-terminus is located near the N-terminal
end of helix C. The protein backbone forms a poorly defined irregular
strand for the first 13 residues before entering strand I of the h-sheet,
then descending through helix A. Helix A acts as a backbone to the
protein, spanning its full length. The pronounced amphipaticity of this
helix provides hydrophobic residues for internal packing to helix B and
to the h-sheet, and the hydrophilic residues are exposed to the solvent.
After helix A the protein backbone turns to form strand II of the h-sheet,
which lies parallel to and outside strand I. This strand rises steeply, giving
the h-sheet a distinctly twisted appearance. It is connected by a short
loop to strand III, which is parallel to and inside of strand I. A long loop
connects strands III and IV, crossing over strand V and placing strand IV
along the ligand-binding cleft and antiparallel to strand V. Another small
75
Figure 9 Backbone (Ca, C’, N) trace from residues 83 to 250 of 30 conformers
of inhibited sfSTR. Residues 251 to 255 are disordered and are not included. All
the heavy atoms of the inhibitor are shown. The family of structures are viewed
along the long axis of the catalytic helix B. The inhibitor (I) binds to the protein in
a well-defined cleft and runs antiparallel to the outer strand of the h-sheet with
the ring of P1V homophenylalanine (hP) buried in a bottomless S1V subsite and the
P2V arginine (R) is exposed to the solvent.
loop connects strand IV to V, which runs parallel to strand III. The
structural zinc is ligated by three His, one each from strands IV and V
(His-166 and -179, respectively), and the third (His-151) from the long
loop connecting strands III and IV. The fourth ligand of this zinc appears
to be Asp-153. After strand V the backbone loops to form helix B. The
two His residues of helix B, His-201 and -205, ligate the catalytic zinc. A
short turn then enters an extended strand containing His-211, a third
ligand of the catalytic zinc. From His-211 to Leu-218 several short range
NOEs, in particular between the side chains of Ser-212 and Ala-217,
ChH3 of Ala-217 to the NH protons of Leu-218 and Met-219, and the
backbone atoms of Thr-215 to Ala-217, describes the presence of two
tight turns. An invariant residue, Met-219, which is residue three in one
of these turns, is positioned below the three His residues that ligate the
catalytic zinc and shows NOEs to all three. Except for helix C, the
remainder of the protein is irregular, but well-defined. Helix C runs
perpendicular to helix A; the segments C-terminal to these helices are
near each other.
76
Gooley
Figure 10 Ribbon diagrams of a single conformer of inhibited sfSTR from
residues 83 to 250. (A) The complex is viewed from above the h-sheet. The
positions of the two zincs are indicated as large balls. The strands of the hsheet (I – V) and helices (A – C) are indicated. The heavy atoms of the inhibitor
and residues of the protein that ligate zinc are shown. The inhibitor runs
antiparallel to strand IV. The structural zinc lies above the h-sheet and is
77
B. Conformation of the Inhibitor
The inhibitor binds to the protein in a well-defined cleft (Figs. 9 and 10) and
in an extended fashion, running antiparallel to strand IV of the sheet, as
indicated by the strong CaH-CaH NOE between the P2V residue and Val163 (Fig. 8) and parallel to the nonregular loop region encompassing Pro221 to Tyr-223. One of the most striking features of the structure is that the
S1V subsite appears to pass through the entire structure. Indeed the
aromatic ring of the homophenylalanine group is clearly observed from
below the S1V pocket [7]. The S1V subsite is lined with hydrophobic residues
including Leu-164, Leu-197, His-201, Val-198, Leu-218, Tyr-220, Leu-222
and Tyr-223. The residues Leu-197, Val-198 and His-201 are from the
catalytic helix, whereas Tyr-220 and -223 and Leu-218 and -222 are from
the loop following this helix. Contacts between the protein and inhibitor
are summarized in Figure 11. Despite the P1V group appearing in contact
with a number of residues, the ring of this residue can clearly undergo ring
flips, as indicated by the degeneracy of the H3,5 and H2,6 resonances
(Fig. 7) thus indicating that this ring is not especially restricted. Similarly
not all residues of the S1V are restricted in motion. For example, both
methyls of Leu-197 show intraresidue NOEs to the CaH proton of Leu-197
suggesting that motion around the torsion angles m1, m2 is present. We
note that this residue shows strong NOEs to the homophenylalanine ring
of the inhibitor (Fig. 8) indicating that it is in contact with the inhibitor.
Analysis of spectra with other inhibitors with extensions to the homophenylalanine showed this residue became restricted in motion, and thus
subtle changes to residue mobility is inhibitor dependent.
The family of conformers were analyzed for hydrogen bonds, where
acceptor-donor (N-H. . .O) distance was set to an upper limit of 2.4 A˚ and
ligated by His-166 from strand IV, His-179 from strand V, and His-151 and
Asp-153 both from a 14 residue loop. The catalytic zinc is ligated by His-201
and – 205 from helix B and His-211. (B) The complex is viewed from below S1V
subsite. The heavy atoms of the inhibitor and the residues that are in
intermolecular contact (Leu-164, Leu-197, Val-198, His-201, Leu-218, Tyr-220,
Leu-222, Tyr-223) with the P1V homophenylalanine are shown. To reduce
crowding in the figure not all these residues are labelled. (*) marks Leu-218
and His-201. Val-198 is below Leu-197. Leu-164 is at the N-terminal end of
the h-strand that appears above Leu-197 in this figure. The ribbon diagrams
were produced by MOLSCRIPT [62].
78
Gooley
Figure 11 Potential hydrogen bond partners to the backbone atoms of the
inhibitor and the residues of the S1V subsite that are in intermolecular contact with
the P1V homophenylalanine.
the angle to 35j. The analysis suggests that the NH of the P3V hydrogen
bonds to the carbonyl of Asn-162; the carbonyl of P1V hydrogen bonds to
the NH of the Leu-164 (which is slowly exchanging with deuterium); and
the amine of P1V hydrogen bonds with the carbonyl of Ala-165. The
structures described here do not show hydrogen bonds between the NH
and the carbonyl of the P2V arginine to the protein, which is in contrast to
reported crystal structures which show a hydrogen bond to the NH of Tyr223 [10]. Although Pro-221 and Tyr-223 are near atoms of the inhibitor, for
example, the NH of Tyr-223 shows weak NOEs to the ring of P3V, their
distances in the structure models are not in agreement with these residues
participating in hydrogen bonds. The NH of Tyr-223 does not show slow
exchange with 2H2O and analysis of 2-D saturation transfer difference
1
H,15N HSQC spectra suggested that the exchange rate of the NH of Tyr223 was one to two orders slower than a free amide proton further
79
supporting the lack of a strong hydrogen bond [7]. The modest protection
of the amide from solvent exchange may simply be due to solvent
accessibility. We also note that the conditions of data collection for the
NMR solution structure were 40jC, which may destablize any weak
hydrogen bonds observed in the crystal structure.
VII.
COMPARISON OF INHIBITED STROMELYSIN
TO OTHER MMPS
The structures of the catalytic domain of a large number of MMPs have
been solved by x-ray crystallography and NMR spectroscopy [7,10,20 –
27]. All cases show that this catalytic domain of the MMPs has a common
fold to that described above, suggesting that the design of specific
inhibitors will require detailed structural investigations that take advantage of differences of the specificity pockets. The most significant difference
to date has been the nature of the S1V subsite, which is clearly very deep in
stromelysin and collagenase-3 [7,10,25], deep in collagenase-2 [21], to quite
shallow for collagenase-1 [20,23,25] and matrilysin [27]. In collagenase-1
an arginine residue (equivalent to Leu-197 in stromelysin-1) hydrogen
bonds to a structural water and delimits the S1V subsite. Consequently,
many inhibitors with large bulky P1V groups show poor affinity for
collagenase-1. However, it has been observed that this protein can undergo
a conformational change to accommodate such groups [25]. The catalytic
domain of stromelysin-1 has been studied by NMR spectroscopy as a
complex with a variety of inhibitors [7,10,51], with most binding with
groups in the S1V to S3V subsites. Those with a thiadizole group ligating the
catalytic zinc bind with groups in the nonprime (S1 to S3) subsites. These
inhibitors show NOEs and thus contacts to residues in the S3 subsite
including His-166, Tyr-155, and Tyr-168, which are located near the
structural zinc.
An advantage of NMR spectroscopy is the analysis of protein
dynamics. Measurement and analysis of the relaxation parameters R1,
R2, and the 15N NOE of 15N-labeled proteins leads to an order parameter
(S2) that can describe the relative mobility of the backbone of the protein.
Both collagenase-1 and stromelysin-1 have been studied either as inhibited
complexes or the free protein [19, 52]. Stromleysin-1 was studied with
inhibitors binding to prime or nonprime subsites. Presence or absence of
inhibitors in the nonprime sites had minor effects on the highly ordered
structure of residues in these subsites, which are in contact with the
80
Gooley
inhibitor. Inhibitors binding to the primed subsites induced considerable
order in the regions 191 to 192 and 223 to 224. Most importantly, the
amide proton of His-223 formed a hydrogen bond to a carbonyl group of
the inhibitor. In addition to these changes residues remote from the
inhibitor, but near the binding sites showed increased mobility. These
results suggest that the rigidity of the S1 to S3 subsites are important for
distinguishing between ligands, while the flexible S1V to S3V subsites are
more accommodating to a broad range of residues. The flexibility of the
S1V subsite is in agreement with our observations. Interestingly, similar
studies on collagenase-1 with a hydroxymate inhibitor [19] bound to the
S1V to S3V subsites showed the analogous region to 220-226 of stromelysin1 was disordered in both the presence or absence of inhibitor. All these
results indicate that changes to mobility are complex and mostly unpre-
Figure 12 Catalytic mechanism of thermolysin and stromelysin-1. (A) The
mechanism of thermolysin [54]. (B) The mechanism of stromleysin-1 [10].
Equivalent residues to Tyr-157 and His-231 are not observed for stromelysin-1.
The proposed mechanism for collagenase-1 [53] is similar to stromelysin-1, but
also involves Asn-180 (equivalent to Asn-162 in stromelysin-1). This residue
cannot participate in stromelysin-1 due to an additional residue between Ala-165
and Asn-162. (Adapted from Ref. 10.)
81
dictable: however, such analyses may prove useful in supporting and
monitoring the presence of stabilizing interactions.
The mechanism proposed for fibroblast collagenase [53] and stromelysin-1 (10) is similar to that suggested for thermolysin [54] (Fig. 12). In
thermolysin the zinc ion ligates the carbonyl of the substrate and with Tyr157 and His-231 stabilizes the tetrahedral intermediate. In collagenase-1
and stromelysin-1, however, the stabilization of the carbonyl of the
substrate and the tetrahedal intermediate is by zinc alone. In thermolysin
the NH of the scissile bond is stabilized by a peptide carbonyl of Ala-182
and the side chain carbonyl of Asn-112. For collagenase-1 similar interactions by the peptide carbonyl of Ala-182 and the carbonyl of the side
chain of Asn-180 are suggested. For stromelysin-1, however, only the
carbonyl of Ala-165 would be involved in the stabilization of the NH of
the substrate; the equivalent Asn (Asn-162) is not involved as there is a
residue insertion in the stromelysin-1 sequence compared with the
collagenase-1 sequence. The proposed mechanisms of thermolysin, collagenase-1 and stromelysin-1 suggest that the Glu in the consensus
sequence HEXXH would promote the nucleophilic attack of water on
the scissile bond of the peptide substrate. The solution structure of
stromelysin-1 described here lacks the rigidity expected for the side chain
of this residue, Glu-202. In several members of the family of structures,
however, this side chain does approach a position that is consistent with
the mechanistic role of this residue.
VIII. CONCLUSION
This chapter has discussed the use of heteronuclear NMR and isotope
editing methods to determine the structure of protein complexes of
therapeutically important drug targets. NMR methodology continues
to develop with larger protein complexes being studied, and more
accurate structures being determined. Developments include deuteration
of proteins [1] to enhance relaxation properties, and experiment design,
for example, Transverse Relaxation Optimized Spectroscopy (TROSY)
[55], which takes advantage of favorable relaxation pathways thus
allowing proteins of at least 60 kDa to be studied; inclusion of residual
dipolar couplings as an orientation constraint in structure calculations
[56,57] are increasing the accuracy of solution structures; and combining
deuteration and TROSY experiments has allowed hydrogen bonds to be
directly observed and also included in structure calculations [58]. An
82
Gooley
additional and powerful application of NMR spectroscopy is the method
of ‘‘NMR by SAR’’ developed by Fesik et al. [59], which has been applied
to finding new drug leads for stromelysin-1 [60]. NMR spectroscopy has
clearly become a powerful and essential tool in the design and development of novel drug leads.
REFERENCES
1. Clore GM, Gronenborn AM. Determining the structures of large proteins
and protein complexes by NMR. Trends Biotechnol 1998; 16:22 – 34.
2. Otting G, Wu¨thrich K. Heteronuclear filters in two-dimensional [1H,1H]
NMR spectroscopy: combined use with isotope labelling for studies of
macromolecular conformation and intermolecular interactions. Q Rev
Biophys 1990; 23:39 – 96.
3. Cai M, Huang Y, Sakaguchi K, Clore GM, Gronenborn AM, Craigie R.
An efficient and cost-effective isotope labeling protocol for proteins
expressed in Escherichia coli. J Biomol NMR 1998; 11: 97 – 102.
4. Campbell ID, Downing AK. NMR of modular proteins. N Struct Biol 1998;
5(suppl): 496 – 499.
5. Gooley PR, Johnson BA, Marcy AI, Cuca GC, Salowe SP, Hagmann WK,
Esser CK, Springer JP. Secondary structure and zinc ligation of human
recombinant short-form stromelysin by multidimensional heteronuclear
NMR. Biochemistry 1993; 32:13098 – 13108.
6. Gooley PR, O’Connell JF, Marcy AI, Cuca GC, Salowe SP, Bush BL,
Hermes JD, Hagmann WK, Esser CK, Springer JP, Johnson BA. The
NMR structure of the inhibited catalytic domain of human stromelysin-1.
Nat Struct Biol 1994; 1:111 – 118.
7. Gooley PR, O’Connell JF, Marcy AI, Cuca GC, Hagmann WK, Caldwell
CG, Axel MG, Becker JW. Comparison of the structure of human
recombinant short form stromelysin by multidimensional heteronuclear
NMR and X-ray crystallography. J Biomol NMR 1996; 7:8 – 28.
8. Chapman KT, Kopka IE, Durette PI, Esser CK, Lanza TJ, IzquierdoMartin M, Niedzwiecki L, Chang B, Harrison RK, Kuo DW, Lin T-Y,
Stein RL, Hagmann WK. Inhibition of matrix metalloproteinases by
N-Carboxyalkyl peptides. J Med Chem 1993; 36:4293 – 4301.
9. Nagase H, Woessner JF Jr. Matrix metalloproteinases. J Biol Chem 1999;
274:2149 – 21494.
10. Becker JW, Marcy AI, Rokosz LL, Axel MG, Burbaum JJ, Fitzgerald
PMD, Cameron PM, Esser CK, Hagmann WK, Hermes JD, Springer JP.
Stromelysin-1: Three-dimensional structure of the inhibited catalytic
domain and of the C-truncated proenzyme. Prot Sci 1995; 4:1966 – 1976.
83
11. Li J, Brick P, O’Hare MC, Skarzynski T, Lloyd LF, Curry VA, Clark
IM, Bigg HF, Hazleman BL, Cawston TE, Blow DM. Structure of fulllength porcine synovial collagenase reveals a C-terminal domain containing a calcium-linked, four-bladed h-propellor. Structure 1995; 3:541 –
549.
12. Gomis-Ru¨th FX, Gohlke U, Betz M, Knaüper V, Murphy G, Lo´pez-Otı´ n
C, Bode W. The helping hand of collagenase-3 (MMP-13): 2.7 A˚ crystal
structure of its C-terminal haemopexin-like domain. J Mol Biol 1996;
264:556 – 566.
13. Gohlke U, Gomis-Ru¨th FX, Crabbe T, Murphy G, Docherty AJP, Bode W.
The C-terminal (haemopexin-like) domain structure of human gelatinase A
(MMP-2): structural implications for its function. FEBS Lett 1996; 378:
126 – 130.
14. Bode W. A helping hand for collagenases: the hemopexin-like domain.
Structure 1995; 3:527 – 530.
15. Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA, Goldberg GI.
Mechanism of cell surface activation of 72-kDa type IV collagenase. J Biol
Chem 1995; 270:5331 – 5338.
16. de Souza SJ, Pereira HM, Jacchieri S, Brentani RR. Collagen/collagenase
interaction - does the enzyme mimic the conformation of its own substrate
FASEB J 1996; 10:927 – 930.
17. Bode W, Gomis-Ru¨th FX, Sto¨cker W. Astacins, serralysins, snake venom
and matrix metalloproteinases exhibit idential zinc-binding environments
(HEXXHXXGXXH and Met-turn) and topologies and should be
grouped into a common family, the ‘metzincins.’ FEBS Lett 1993; 331:
134 – 140.
18. Marcy AI, Eiberger LL, Harrison R, Chan HK, Hutchinson NI, Hagmann
WK, Cameron PM, Boulton DA, Hermes JD. Human fibroblast
stromelysin catalytic domain: expression, purification and characterization
of a C-terminally truncated form. Biochemistry 1991; 30:6476 – 6483.
19. Moy FJ, Pisano MR, Chanda PK, Urbano C, Killar LM, Sung M-L,
Powers R. Assignments, secondary structure and dynamics of the
inhibitor-free fragment of human fibroblast collagenase. J Biomol NMR
1997; 10:9 – 19.
20. Bartoki N, Winkler FK, Williams DH, D’Arcy A, Broadhurst MJ, Brown
PA, Johnson WH, Murray EJ. Structure of the catalytic domain of human
fibroblast collagenase complexed with an inhibitor. Nat Struct Biol 1994;
1:106 – 110.
21. Stams T, Spurlino JC, Smith DL, Wahl RC, Ho TF, Qoronfleh MW, Banks
TM. Structure of human neutrophil collagenase reveals large S1’ specificity
pocket. Nat Struct Biol 1994; 1:119 – 123.
22. Bode W, Reinemer P, Huber R, Kleine T, Schnierer S, Tschesche H. The Xray crystal structure of the catalytic domain of human neutrophil
84
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
Gooley
collagenase inhibited by a substrate analogue reveals the essentials for
catalysis and specificity. EMBO J 1994; 13:1263 – 1269.
Lovejoy B, Cleasby A, Hassell AM, Longley K, Luther MA, Weigl D,
McGeehan G, McElroy AB, Drewry D, Lambert MH, Jordan SR. Structure
of the catalytic domain of fibroblast collagenase complexed with an
inhibitor. Science 1994; 263:375 – 377.
Moy FJ, Chanda PK, Chen JM, Cosmi S, Edris W, Skotnicki JS, Wilhelm J,
Powers P. NMR solution structure of the catalytic fragment of human
fibroblast collagenase complexed with a sulfonamide derivative of a
hydroxamic acid compound. Biochemistry 1999; 38:7085 – 7096.
Lovejoy B, Welch AR, Carr S, Luong C, Broka C, Hendricks RT, Campbell
JA, Walker KAM, Martin R, Van Watrt H, Browner MF. Crystal
structures of MMP-1 and MMP-13 reveal the structural basis for selectivity
of collagenase inhibitors. Nat Struct Biol 1999; 6:217 – 221.
Van Doren SR, Kurochkin AV, Hu W, Ye Q-Z, Johnson LL, Hupe DJ,
Zuiderweg ERP. Solution structure of the catalytic domain of human
stromelysin complexed with a hydrophobic inhibitor. Prot Sci 1995; 4:
2487 – 2498.
Browner MF, Smith WW, Castelhano AL. Matrilysin-inhibitor complexes:
common themes among metalloproteases. Biochemistry 1995; 34:6602 –
6610.
Morgunova E, Tuutila A, Bergmann U, Isupov M, Lindqvist Y, Schneider
G, Tryggvason K. Structure of human pro-matrix metalloproteinase-2:
activation mechanism revealed. Science 1999; 284: 1667 – 1670.
Muhandiram DR, Kay LE. Gradient-enhanced triple resonance threedimensional NMR experiments with improved sensitivity. J Magn Reson
1994; 103:203 – 216.
Lo¨hr F, Ru¨terjans H. A triple resonance experiment for the sequential
assignment of backbone resonances in proteins. J Biomol NMR 1995;
6:189 – 197
Boucher W, Laue ED, Campbell SL, Domaille PJ. Improved 4D NMR
experiments for the assignment of backbone nuclei in 13C/15N labelled
proteins. J Biomol NMR 1992; 2:631 – 637.
Powers R, Gronenborn AM, Clore GM, Bax A. Three-dimensional tripleresonance NMR of 13C/15N-enriched proteins using constant-time evolution. J Magn Reson 1991; 94:209 – 213.
Kay LE, Xu G-Y, Singer AU, Muhandiram DR, Forman-Kay JD. A
gradient-enhanced HCCH-TOCSY experiment for recording side-chain 1H
and 13C correlations in H2O samples of proteins. J Magn Reson 1993;
101B:333 – 337.
Senn H, Werner B, Messerle BA, Weber C, Traber R, Wu¨thrich K.
Stereospecific assignment of the methyl 1H NMR lines of valine and leucine
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
85
in polypeptides by nonrandom 13C labelling. FEBS Lett 1989; 249: 113 –
118.
Kuboniwa H, Grezesiek S, Delaglio F, Bax A. Measurement of HN-Ha J
couplings in calcium-free calmodulin using new 2D and 3D water-flip-back
methods. J Biomol NMR 1994; 4:871 – 878.
Archer SJ, Ikura M, Torchia DA, Bax A. An alternative 3D NMR
technique for correlating backbone 15N with side chain Hh resonances in
larger proteins. J Magn Reson 1991; 95:636 – 641.
Grzesiek S, Kuboniwa H, Hinck AP, Bax A. Mutiple-quantum line
narrowing for measurement of Ha-Hh J couplings in isotopically enriched
proteins. J Am Chem Soc 1995; 117:5312 – 5315.
Zhang O, Kay LE, Olivier JP, Forman-Kay JD. Backbone 1H and 15N
resonance assignments of the N-terminal SH3 domain of drk in folded and
unfolded states using enhanced sensitivity pulsed field gradient NMR
techniques. J Biol NMR 1994; 4:845 – 858.
Ikura M, Kay LE, Tsudin R, Bax A. Three-dimensional NOESYHMQC spectroscopy of a 13C-labelled protein. J Magn Reson 1990; 86:
204 – 209.
Vuister GW, Clore GM, Gronenborn AM, Powers R, Garrett DS, Tschudin
R, Bax A. Increased resolution and improved spectral quality in fourdimensional 13C/13C-separated HMQC-NOESY-HMQC spectra using
pulsed field gradients. J Magn Reson 1993; 101B:210 – 213.
Pelton JG, Torchia DA, Meadow ND, Roseman S. Tautomeric states of
the active-site histidines of phosphorylated and unphosphorylated IIIGlc,
a signal-transuding protein from Escherichia coli, using two-dimensional
heteronuclearNMR techniques. Prot Sci 1993; 2:543 – 558.
Petros AM, Kawai M, Luly JR, Fesik SW. Conformation of two nonimmunosuppressive FK506 analogs when bound to FKBP by isotopefiltered NMR. FEBS Lett 1992; 308:309 – 314.
Burgering M, Boelens R, Kaptein R. Observation of intersubunit NOEs in a
dimeric P22 Mnt repressor mutant by a time-shared [15N,13C] double halffilter technique. J Biomol NMR 1993; 3:709 – 714.
Gu¨ntert P, Braun W, Billeter W, Wu¨thrich K. Automated stereospecific 1H
NMR assignments and their impact on the precision of protein structure
determinations in solution. J Am Chem Soc 1989; 111:3997 – 4004.
Gu¨ntert P, Braun W, Wu¨thrich K. Efficient computation of three-dimensional protein structures in solution from nuclear magnetic resonance data
using the program DIANA and the supporting programs CALIBA,
HABAS and GLOMSA. J Mol Biol 1991; 217:517 – 530.
Gu¨ntert P, Mumenthaler C, Wu¨thrich K. Torsion angle dynamics for NMR
structure calculation with the new program DYANA. J Mol Biol 1997;
273:283 – 298.
86
Gooley
47. Bru¨nger AT, Adams PD, Clore GM, DeLano WL, Gros P, GrosseKunstleve RW, Jiang J-S, Kuszewski J, Nilges M, Pannu NS, Read RJ, Rice
LM, Simonson T, Warren GL. Crystallography and NMR system: A new
software suite for macromolecular structure determination. Acta Cryst 1998;
d54:905 – 921.
48. Widmer H, Widmer A, Braun W. Extensive distance geometry calculations
with different noe calibrations - new criteria for structure selection applied to
sandostatin and BPTI. J Biomol NMR 1993; 3:307 – 324.
49. Schaumann T, Braun W, Wu¨thrich K. The program FANTOM for energy
refinement of polypeptides and proteins using a Newton-Raphson minimizer
in torsion angle space. Biopolymers 1990; 29: 679 – 694.
50. Richardson JS. h-Sheet topology and the relatedness of proteins. Nature
1977; 268:495 – 500.
51. Stockman BJ, Waldon DJ, Gates JA, Schaill TA, Kloosterman DA, Mizsak
SA, Jacobsen EJ, Belonga KL, Mitchell MA, Mao B, Petke JD, Goodman
L, Powers EA, Ledbetter SR, Kaytes PS, Vogeli G, Marshall VP, Petzold
GL, Poorman RA. Solution structure of stromelysin complexed to
thiadiazole inhibitors. Prot Sci 1998; 7:2281 – 2286.
52. Yuan P, Marshall VP, Petzold GL, Poorman RA, Stockman BJ. Dynamics
of stromelysin/inhibitor interactions studied by 15N NMR relaxation
measurements: Comparison of ligand binding to the S1-S3 and S1’-S3’
subsites. J Biomol NMR 1999; 15:55 – 64.
53. Spurlino JC, Smallwood AM, Carlton DD, Banks TM, Vavra KJ, Johnson
JS, Cook ER, Falvo J, Wahl RC, Pulvino TA, Wendoloski JJ, Smith DL.
1.56 A˚ Structure of mature truncated human fibroblast collagenase. Proteins
1994; 19:98 – 109.
54. Mathews BW. Structural basis of the action of thermolysin and related zinc
peptidases. Acc Chem Res 1988; 21:333 – 340.
55. Pervushin K, Riek R, Wider G, Wu¨thrich K. Attenuated T2 relaxation by
mutual cancellation of dipole-dipole coupling and chemical shift
anisotropy indicates an avenue to NMR structures of very large biological macromolecules in solution. Proc Natl Acad Sci 1997; 94:12366 –
12371.
56. Tjandra N, Bax A. Direct measurement of distances and angles in biomolecules by NMR in a dilute liquid crystalline medium. Science 1997;
278:1111 – 1114.
57. Olejniczak ET, Meadows RP, Wang H, Cai M, Nettesheim DG, Fesik SW.
Improved NMR structures of protein/ligand complexes using residual
dipolar couplings. J Am Chem Soc 1999; 121:9249 – 9250.
58. Cornilescu G, Hu J-S, Bax A. Identification of the hydrogen bonding
network in a protein by scalar couplings J Am Chem Soc 1999; 121:2949 –
2950.
87
59. Shuker SB, Hajduk PJ, Meadows RP, Fesik SW. Discovering high-affinity
ligands for proteins - SAR by NMR. Science 1996; 274:1531 – 1534.
60. Hajduk PJ, Sheppard G, Nettesheim DG, Olejniczak ET, Shuker SB,
Meadows RP, Steinman DH, Carrera GM, Marcotte PA, Severin J, Walter
K, Smith H, Gubbins E, Simmer R, Holzman TF, Morgan DW, Davidsen
SK, Summers JB, Fesik SW. Discovery of potent nonpeptide inhibitors of
stromelysin using SAR by NMR. J Am Chem Soc 1997; 119:5818 – 5827.
61. Schechter I, Berger A. On the size of the active site in proteases 1. Papain
Biochem. Biophys Res Commun 1967; 27:157 – 162.
62. Kraulis PJ. MOLSCRIPT: A Program to Produce Both Detailed and
Schematic Plots of Protein Structures. J Appl Crystallogr 1991; 24: 946 – 950.
4
Cannabinergics: Old and New
Therapeutic Possibilities
Andreas Goutopoulos
Serono Reproductive Biology Institute, Rockland, Massachusetts, U.S.A.
I.
INTRODUCTION
Cannabis sativa, one of the oldest plants farmed by man, has been known
for its medicinal properties for at least four millennia (Peters, 1999). The
psychoactive–euphoric effects of this plant, as well as its facile and wide
climatic range of cultivation, have rendered it a very popular recreational
drug. Today, cannabis, or marijuana, is still the focus of strong social,
legal, and medical controversy over its therapeutic utility.
Referenda in Arizona and California in 1997, and later, others in
eight additional states, aimed at legalizing marijuana cigarettes for medical
purposes. Two licensed, single-compound, cannabimimetic pharmaceuticals, Marinol (Dronabinol, delta-9-THC from Roxane Labs) and Cesamet
(Nabilone, developed at Eli Lilly, now in use in the United Kingdom), are
marketed for two purposes: to control the nausea and emesis produced by
cancer chemotherapy and to serve as appetite stimulants in AIDS-related
anorexia. In clinical trials with cancer chemotherapy patients, both these
agents have proven to be superior to conventional antiemetics, such as
perchlorperazine (Breivogel, 1998).
Beyond this relatively limited medical use of cannabimimetics, the
current, albeit long-delayed elucidation of their pharmacology is likely to
lead to a wide expansion of the clinical potential and significance of
89
90
Makriyannis and Goutopoulos
these drugs. The oily, noncrystalline nature of the biologically active
terpenoid ingredients of Cannabis sativa contributed to the lag in understanding of cannabinoid biology. The main active ingredient, (–)-delta-9tetrahydrocannabinol (delta-9-THC), was isolated and identified only in
1964 (Mechoulam, 1967), over a hundred years after the isolation of many
important crystalline biologically active natural products, such as morphine and quinine. A second reason for the lack of progress in defining the
biology of cannabimimetics was the long-standing scientific misconception
that the cannabinoid-induced pharmacological actions are mediated by
perturbation of cellular membranes rather than through specific receptors.
This hypothesis was a deterrent in the pursuit of possible specific cannabinoid binding sites. Owing to their high lipophilicity, cannabinoids were
paralleled with general anesthetics in terms of their mechanism of action
(Paton, 1975). Although cannabinoids were found to clearly perturb membranes (Makriyannis, 1987), such effects were never proven to be directly
responsible for their biological activity.
The advent of synthetic cannabimimetics with a high degree of
enantioselectivity (Johnson, 1986; Little, 1988) paved the road for the
identification of specific cannabinoid binding sites in rat brain (Devane,
1988). This discovery marked the onset of a revolution in the understanding of cannabinoid biology.
II. CANNABINOID RECEPTORS
A. The CB1 Receptor
Definitive proof of the existence of the cannabinoid receptor came with
the isolation of the cDNA of a cannabinoid receptor from a rat cerebral
cortex cDNA library and its expression in Chinese hamster ovary
(CHO) cells (Matsuda, 1990). A year later, the corresponding human
receptor, named CB1, was cloned and found to share a 97.3% homology with the rat receptor (Gerard, 1991). The CB1 472 amino acid
sequence revealed (Matsuda, 1990; Gerard, 1991) that it is a member
of the G-protein-coupled receptors (GPCRs). Receptors of this family
are membrane embedded and consist of an extracellular N-terminus,
seven transmembrane helices interconnected with intra- and extracellular
loops, and an intracellular C-terminus. Bramblett et. al. (1995) constructed a model for CB1, using the known structure of bacteriorhodopsin as a starting point.
Cannabinergics
91
The sites involved in interactions with G proteins of the Gi/o family
are the third intracellular loop from the N-terminal side and the Cterminus (Howlett, 1998a). The C-terminus was found to bind with high
affinity to Gi and the synthetic C-terminus peptide was found to individually stimulate GTPgS binding to G protein and to inhibit adenylate
cyclase (Howlett, 1998b). Similarly with other GPCRs, CB1 is allosterically regulated by sodium ions. It has been shown that sodium ions affect
both ligand binding and signal transduction by inducing a receptor
conformational change (Houston, 1998). There is also evidence that an
interhelical H-bonding interaction between helix II Asp and helix VII Asn
is important for the stabilization of a receptor conformational state that
has high affinity for most cannabimimetic ligands (Tao, 1998), (Howlett,
1998a). Sodium ions presumably disrupt this H bond, and thus, result in a
different, low affinity, receptor state.
The CB1 receptor is coupled with Gi (Howlett, 1998a). CB1
activation leads to inhibition of adenylyl cyclase and, therefore, to reduction of cAMP levels. Many eukaryotic cells utilize cAMP as a
second messenger that activates the cAMP-dependent protein kinase A
(PKA), which in turn phosphorylates various proteins, regulating their
function. One of the cAMP-dependent cannabinoid effects is the
enhancement of voltage-sensitive, outwardly rectifying potassium channels, which occurs as a result of decreased phosphorylation of the K+
channel protein by PKA (Deadwyler, 1995). Besides Gi, CB1 is coupled
to Go (Howlett, 1999). Furthermore, apart from inhibition of adenylyl
cyclase, CB1 utilizes several additional effector systems (intracellular
mediators) involving Gi/o proteins: the inhibition of N-type Ca2+ channels (Mackie, 1992); the activation of mitogen-activated protein kinase
(MAP kinase) (Bouabula, 1995a); and the expression of immediate early
genes like Krox-24 (Bouabula, 1995b). Other cannabinoid-induced cellular effects include activation of inwardly-rectifying potassium channels
(Pertwee, 1997) and possibly the activation of phospholipases A, C, or D
(Felder, 1995).
Different G proteins or second messengers may be coupled to CB1
in different brain regions and may mediate different physiological effects
(Howlett, 1999). Utilization of diverse effector systems by CB1 may
explain how the response to cannabimimetics varies across different cell
types. Understanding which physiological responses are mediated by
each of the foregoing intracellular signaling systems is of great significance and may suggest new approaches for the design of selective
cannabimimetic agents.
92
B. The CB2 Receptor
Homology cloning revealed the existence of a second cannabinoid receptor, CB2 (Munro, 1993). This receptor shows 44% homology to the total
CB1 and 68% homology within the transmembrane regions. Not present
in the brain in significant levels, CB2 is found mainly in the periphery and
particularly in tissues of the immune system, such as leukocytes, spleen,
thymus, and tonsils (CB1 is found in some of these cells as well). Localization of CB2 in the immune system suggests an immunomodulatory role
for this receptor. Thus, CB2 is likely to be the mediator of the long-known
immunosuppressive properties of marijuana. Similarly to CB1, CB2 uses
signal transduction pathways, such as inhibition of adenylyl cyclase and
stimulation of MAP kinase. However, unlike CB1, CB2 does not affect ion
channels (Pertwee, 1997).
Although the human genome does not contain genes with high
homology to those of CB1 and CB2, other cannabinoid receptor types
may exist nevertheless. An amino-terminal differentially spliced CB1
variant, CB1A, has been isolated from a human lung cDNA library and,
akin to CB1, is expressed in the brain (Shire, 1995). The biological role and
pharmacological implications of this variant are still unclear.
The existence of a second peripheral CB2-like receptor is supported
by the finding that palmitylethanolamide provided antinociception after
intraplantar injection of formalin solution in mice paws (Calignano, 1998).
This effect was attenuated by a CB2 antagonist, SR144528, but not by
SR14176A (a CB1 antagonist) nor by the opioid antagonist naloxone.
Palmitylethanolamide has no significant affinity for either CB1 or CB2
(Khanolkar, 1996). However, in addition to the findings of Calignano
et al., palmitylethanolamide is shown to have a down-regulating effect
on mast cell activity, presumably mediated through a CB2-like receptor
present in these cells (Facci, 1995).
Mouse vas deferens (MVD) seems to express CB1 and at least one
CB2-like cannabinoid receptor type, as is demonstrated by the presence of
CB1 and CB2-like mRNA as well as by data collected from experiments
with cannabinoid receptor selective agonists and antagonists (Pertwee,
1999). Furthermore, evidence indicates that a CB1-like receptor exists in
vascular endothelium, which upon activation produces significant hypotension (Wagner, 1999). This receptor differs from CB1 in its pharmacological response to some well-characterized cannabimimetics.
None of these possible new CB variant receptors has been cloned
yet; therefore, their existence is still putative. Thus, it is unclear whether
Cannabinergics
93
these observations are indicative of novel cannabinoid receptor types,
results of alternative versions of known receptors coupled with different
effector systems, or even results of different affinity states of a single
cannabinoid receptor. Discovery and characterization of new cannabinoid
receptors with different distribution patterns and ligand affinities is of
major importance because it will provide new targets for the development
of highly selective and clinically useful cannabinergic agents.
C. Cannabinoid Receptor Distribution
The ubiquitous CB1 is found in the central nervous system (CNS), as
well as in the periphery and in both neural and nonneural tissues; it
is one of the most abundant GPCRs in the brain (Breivogel, 1998).
As shown by autoradiographic studies in various mammalian brains
(Herkenham, 1990; Gatley, 1998), CB1 density is highest in basal ganglia: substantia nigra pars reticulata, entopeduncular nucleus, and the
external segment of globus pallidus. Moderately high CB1 density is
found in putamen, cerebellum, and hippocampus, whereas moderate
levels exist in cerebral cortex. The spinal cord shows a range of moderate
densities, while thalamus and brain stem contain low to negligible levels.
Autoradiography studies with [35S]GTPgS revealed that cannabinoid
activity occurs with the same regional distribution as the receptors;
however, the level of activity did not parallel receptor density (Breivogel,
1998). This pattern of CB1 distribution in the brain is similar to that of
D1 receptors, which suggests that the cannabinoid system may be
involved in the modulation of the dopaminergic activity (Gatley, 1998).
In fact, CB1 mediates a negative feedback control over D2 in the striatum
(Giuffrida, 1999).
In the periphery, CB1 is found in the adrenal glands, bone
marrow, heart, lung, prostate, testes, thymus, tonsils, spleen, lymphocytes, phagocytes, smooth muscle, vascular endothelium, peripheral
neurons (e.g., in the gut), kidneys, uterus, and sperm as reviewed by
Schuel et al. (1999).
The CB2 receptor has a more limited distribution, being localized
predominantly in the immune system. Among the human leukocytes, B
lymphocytes express the highest levels of CB2, followed respectively by
natural killer cells, monocytes, polymorphonuclear neutrophils, T8 lymphocytes, and T4 lymphocytes. It is also found in the lymph nodes, spleen,
tonsils, and thymus (Cabral, 1999).
94
D. Cannabinoids and Membranes
Before the first indication of the existence of cannabinoid receptors, the
prevailing theory on the mechanism of cannabinoid activity was that
cannabinoids exert their effects by nonspecific interactions with cell
membrane lipids (Makriyannis, 1990). Such interactions can increase the
membrane fluidity, perturb the lipid bilayer and concomitantly alter the
function of membrane-associated proteins (Loh, 1980). A plethora of
experimental evidence suggests that cannabinoids interact with membrane
lipids and modify the properties of membranes. However, the relevance of
these phenomena to biological activities is still only, at best, correlative. An
important conundrum associated with the membrane theories of cannabinoid activity is uncertainty over whether cannabinoids can achieve in
vivo membrane concentrations comparable to the relatively high concentrations used in in vitro biophysical studies (Makriyannis, 1995). It may be
possible that local high concentrations are attainable under certain physiological circumstances, and, if so, some of the cannabinoid activities may
indeed be mediated through membrane perturbation.
Interactions of cannabimimetics with membranes may be of importance for auxiliary roles such as transport to their sites of action and proper
orientation for optimum interaction with their receptors. The molecular
features of cannabimimetics are shown to govern the manner by which
these molecules cross cell membranes, including the brain–blood barrier
(BBB) (Makriyannis, 1995). Most cannabimimetics are amphipathic, a
property that affects their orientation within the lipid bilayer. Strong
experimental evidence has shown that the phenolic hydroxyl group of
(–)-delta-9-THC anchors it at the polar interface of the membrane,
whereas the tricyclic hydrophobic system remains imbedded in the bilayer
and perpendicular to the fatty acid chains (Martel, 1993; Makriyannis,
1995). Other active dihydroxy THC derivatives adopt similar orientation,
while many inactive analogs assume an orientation parallel to the lipids
and a position deeper, closer to the center of the bilayer. The proper
positioning and orientation of cannabinoids within the membrane may be
crucial for reaching the receptor site, located within the transmembrane
receptor helices, by lateral diffusion (Makriyannis, 1995).
III. THE ENDOGENOUS LIGANDS
The discovery of the cannabinoid receptors and their G-protein-coupled
nature strongly suggested the existence of endogenous cannabimimetic
Cannabinergics
95
Figure 1 The structure of anandamide.
ligand(s) able to exert physiological activity upon binding to these receptors. Initial efforts to identify a possible protein (Nye, 1988) or a watersoluble endogenous cannabimimetic ligand were unsuccessful (Deadwyler,
1995). The hypothesis that such a putative endocannabinoid should be
lipophilic, like the classical exogenous cannabinoids, led Mechoulam et al.
to seek such a ligand in the hydrophobic fractions of porcine brain extracts
(Devane, 1992). Repetitive fractionations and purifications led to the
identification of a substance that bound to CB1 in a saturable fashion.
This compound was the ethanolamide of arachidonic acid (arachidonyl
ethanolamide, AEA) (Fig. 1). The authors named this brain constituent
anandamide from ananda, the Sanskrit word for bliss.
Anandamide is found in human brain: 100 pmol/g in the hippocampus, 75 pmol/g in the thalamus, 60 pmol/g in the cerebellum, and 55
pmol/g in the striatum (Martin, 1999). The concentration of AEA increases
postmortem, especially when the brain is kept at ambient temperature.
Furthermore, AEA surges are observed when cerebellar granule cells are
treated in hypoxic conditions (Hillard, 1997). Although such concentration increases may be artifacts of postmortem brain damage, they may also
occur in living tissue under certain conditions, such as hypoxia.
Outside the CNS, anandamide is found in the spleen and heart at
approximately 10 pmol/g (Martin, 1999). It is also localized in rat testes
and uterus in concentrations significantly greater than those in the brain
Figure 2 The structures of two N-acylethanolamide (NAE) endocannabinoids.
96
Figure 3
The structure of 2-arachidonyl glycerol.
(Schmid, 1997). Very low levels have been detected in serum, plasma, and
cerebrospinal fluid—a fact that indicates that anandamide is not hormonal
in nature but rather is biosynthesized at or near its sites of action.
In addition to anandamide, several other endogenous polyunsaturated fatty acid derivatives were also found to act as cannabimimetics.
They are all now collectively referred to as endocannabinoids. Soon after
the discovery of anandamide, two more fatty acid ethanolamides were
isolated and found to bind to CB1 preparations with affinities similar to
that of anandamide (anandamide CB1 binding affinity Ki = 39.2 nM,
according to Hanus et al., 1993). These were the homo-g-linolenylethanolamide (CB1 Ki = 53.4 nM) and 7,10,13,16-docosatetraenylethanolamide
(CB1 Ki = 34.4 nM) (Fig. 2). All three N-acylethanolamide endocannabinoids were found to be CB1 agonists in the MVD test (Pertwee, 1994).
A different type of endocannabinoid that is also an arachidonic acid
derivative was first isolated from canine gut and identified as 2-arachidonyl
glycerol (2-AG) (Fig. 3) (Mechoulam, 1995a).
Later 2-AG was also found in the brain (Stella, 1997) and spleen
(Di Marzo, 1998). It was shown to be released in a calcium-dependent
manner, reaching concentrations 170 times higher than that of anandamide in the brain (Stella, 1997). Like the other endocannabinoids, 2-AG
was shown to produce the typical tetrad of cannabimimetic behavioral
effects and inhibit electrically evoked contractions of mouse MVD
(Mechoulam, 1995a).
A. Anandamide Pharmacology
Since the discovery of anandamide in 1992, a number of studies have
examined its pharmacological properties. Although its roles are still
elusive, a plethora of data supports the initial postulate that anandamide
is the major endogenous cannabinoid ligand. As mentioned earlier, anandamide binds to CB1 from brain preparations and displaces various well-
Cannabinergics
97
characterized cannabimimetic radioligands (Hillard, 1997). Furthermore,
it binds to CB1 expressed in cells transfected with CB1 DNA (Vogel, 1993).
Its CB1 affinity is comparable to that of delta-9-THC. Anandamide does
not have effects on other than the cannabinoid receptors (Hillard, 1997)
and has cannabimimetic properties, both in vitro and in vivo. Anandamide
acts as a CB1 agonist, as demonstrated when it inhibited forskolinstimulated adenylyl cyclase activity in N18TG2 cells (IC50 = 540 nM)
(Vogel, 1993), CB1 expressing CHO cells (IC50 = 160–322 nM) (Vogel,
1993), and cerebral membranes (IC50 = 1.9 AM) (Childers, 1994). It was
found to have lower efficacy (lower maximal effect) than the high-affinity
cannabimimetics WIN55212-2 and CP-55,940; thus, anandamide was classified as a partial agonist (Vogel, 1993; Childers, 1994). Furthermore, it was
found to have inhibitory effects in N-type calcium currents through a pertussis-toxin-sensitive pathway in N18 neuroblastoma cells (Mackie, 1993).
Anandamide, in vivo, was shown to produce the four characteristic
effects of cannabimimetics, namely, analgesia, hypothermia, hypoactivity, and catalepsy (Smith, 1994; Fride, 1993; Crawley, 1993). These four
effects are not unique to cannabimimetics; when they are produced
together, however, they are highly predictive of cannabimimetic activity
(Martin, 1991). Anandamide was found to be less potent than delta-9THC in producing these behavioral effects in mice (Fride, 1993). It has
quicker onset and shorter duration of action, the latter because of rapid
catabolism. Cross-tolerance studies, in which pretreatment of mice with
delta-9-THC produced tolerance to most of the pharmacological effects
of anandamide and vice versa, indicate that both drugs act on the same
receptor (Jarbe, 1998).
In addition, anandamide was found to parallel classical cannabinoid
pharmacology in a series of nonbehavioral experimental systems. In
isolated MVD, (Pertwee, 1992) and guinea pig ileum, it inhibited electrically evoked twitch responses (Pertwee, 1995). Moreover, anandamide was
shown to decrease intraocular pressure in rabbits (Pate, 1995), to reduce
sperm-fertilizing capacity in sea urchins by inhibition of the acrosome
reaction (Schuel, 1994), and to produce hypotension in rats (Varga, 1995).
All the foregoing pharmacological effects of anandamide, in conjunction with the well-documented existence of specific systems for its
biosynthesis, catabolism, and cellular reuptake to be discussed shortly,
suggest that anandamide is indeed the endogenous cannabinoid ligand.
The other two less studied N-acylethanolamide endocannabinoids and also
2-AG may serve similar functions. The differential roles of each of these
four endocannabinoids are still unclear.
98
Despite all the research conducted on the pharmacological effects of
anandamide and other cannabimimetics, the exact role of the endogenous
cannabinoid system remains elusive. The ubiquitous pharmacology of
cannabimimetics suggests that endocannabinoids have differing functions
depending on the tissue or organ system. A recent report (Giuffrida, 1999)
sheds light on the role of anandamide in the dorsal striatum. It was shown
that anandamide, but not 2-AG, is released after D2 activation and
subsequently suppresses motor activity, possibly by inhibiting postsynaptic GABAergic currents. Therefore, it was suggested that anandamide, at
least in the striatum, plays the role of an autocoid (local neuromodulator)
that has a negative feedforward regulatory effect of D2-mediated locomotor behavior (Giuffrida, 1999).
It may be that 2-AG has different roles in the CNS, for it can reach
170 times higher concentrations than that of anandamide in the brain
(Stella, 1997), even though it was undetectable in the striatum. In the
hippocampus, 2-AG, but not anandamide, was released after glutamatergic activation (Stella, 1997). Sugiura (1999) found 2-AG to be a full CB1
agonist, whereas anandamide is a partial agonist, again pointing to
alternative roles for 2-AG in comparison to anandamide.
B. Endocannabinoid Metabolism
Biosynthesis of Anandamide
Considerable advances have been made during the late 1990s toward
understanding the physiological pathways that are involved in the synthesis and inactivation of endocannabinoids. The first of these pathways to
be observed, an enzymatic activity responsible for anandamide hydrolysis,
led to lower apparent CB1 affinities for anandamide analogs in studies
involving structure–activity relationships (SARs) (Childers, 1994),
(Abadji, 1994). Inclusion in the binding assay of phenylmethanesulfonyl
fluoride (PMSF), a general serine protease inhibitor, protected the anandamide analog from hydrolysis (Abadji, 1994; Khanolkar, 1996). Shortly
after, an enzyme specific for this hydrolytic process was identified and
characterized (Deutsch, 1993; Ueda, 1995). Initially, it was thought that
this hydrolase, named anandamide amidase or fatty acid amidohydrolase
(FAAH), was also responsible for the synthesis of anandamide by acting
reversibly (Devane, 1994). However, the current belief is that anandamide
amidase is unlikely to be physiologically responsible for anandamide
synthesis because of the requirement for significantly higher than normal
Cannabinergics
99
physiological concentrations of arachidonic acid and ethanolamine (up to
160 mM) for this enzyme to catalyze the reverse reaction (Piomelli, 1998).
Therefore, it is currently believed that anandamide is formed from
membrane phospholipids (Fig. 4) through a pathway that involves: (1) a
trans-acylation of the amino group of phosphatidylethanolamine with
arachidonate from the sn-1 position of phosphatidylcholine and (2) a Dtype phosphodiesterase activity on the resulting N-arachidonylphosphatidylethanolamide (NAPE). Synthesis of anandamide is presumably regulated at the levels of both enzymes, the N-acyltranferase and the
phospholipase D, by stimuli that raise intracellular calcium or by receptors
linked with cAMP and PKA. It has been shown that anandamide is formed
when neurons are depolarized and, therefore, the intracellular calcium ion
levels are elevated (Cadas, 1996).
Biosynthesis of 2-AG
Two possible pathways for the biosynthesis of 2-AG have been proposed:
(1) a phospholipase C (PLC) hydrolysis of membrane phospholipids
followed by a second hydrolysis of the resulting 1,2-diacylglycerol by
diacylglycerol lipase or (2) a phospholipase A1 (PLA1) activity that
generates a lysophospholipid, which in turn is hydrolyzed to 2-AG by
lysophospholipase C (Fig. 5) (Piomelli, 1998). Alternative pathways may
also exist from either triacylglycerols by a neutral lipase activity or
lysophosphatidic acid by a dephosphorylase. The fact that PLC and
diacylglycerol lipase inhibitors inhibit 2-AG formation in cortical neurons
supports the contention that 2-AG is, at least predominantly, biosynthesized by the PLC pathway (Stella, 1997). However, a mixed pathway may
also be plausible.
As with the biosynthesis of anandamide, the biosynthesis of 2-AG is
also triggered by increases of intracellular calcium ions that result from
neuronal activity. High frequency stimulation of neurons produced a
fourfold increase of 2-AG accumulation compared with controls, and this
was prevented by sodium ion channel blocking or removal of calcium ions
(Stella, 1997). The concentration of 2-AG in depolarized neurons reached 1
to 2 AM, significantly higher than anandamide and sufficient to substantially activate CB1 (Stella, 1997).
Based on the pathways just proposed for the biosynthesis of anandamide and 2-AG, the formation of these endogenous ligands must also be
dependent on the composition of the precursor lipids. This dependence is
of greater importance for anandamide rather than for 2-AG because
Figure 4
Anandamide biosynthesis.
Cannabinergics
101
Figure 5 Possible 2-AG biosynthesis.
arachidonic acid is rarely esterified at the sn-1 position of phospholipids,
whereas it is commonly found at the sn-2 position.
Endocannabinoid Release
Immediately after synthesis, endocannabinoids are released in the extracellular space, where they then act on the same or neighboring cells as
autocrine or paracrine mediators (Di Marzo, 1999). Experimental evidence
thus far indicates that anandamide and 2-AG, unlike other classical
neurotransmitters, are not stored in vesicles. First, anandamide basal
concentrations are extremely low (5–10 pmol/g), 100 to 10,000 times lower
than those of classical neurotransmitters (Cadas, 1997). Second, stimulus-dependent anandamide release is linked with de novo NAPE and
102
subsequent anandamide biosynthesis (Cadas, 1996; Di Marzo, 1999).
Therefore, it is currently believed that anandamide and 2-AG are produced
and immediately released from the neurons upon demand (Di Marzo,
1999b; Piomelli, 1998).
The poor water solubility of anandamide must preclude extensive
free diffusion in the extracellular space. Since, however, anandamide is
found in brain incubation media or perfusates of brain microdialysis
experiments it obviously exits the cells (Giuffrida, 1999). Additionally, it
is known that striatal astrocytes, which do not produce anandamide, do
respond to it (Cadas, 1996). Therefore, it has been suggested that after
cleavage from NAPE, anandamide is immediately released from the
membrane with the assistance of a membrane transporter (such as a
P-glycoprotein) (Ayotte, Picone, and Makriyannis; unpublished results) or
a lipid binding protein (like a lipocalin) (Piomelli, 1998). Such a lipid
binding protein may also facilitate the passage of anandamide through the
aqueous extracellular space to its sites of action.
Endocannabinoid Inactivation
Anandamide is inactivated in two steps, first by transport inside the cell
and subsequently by intracellular enzymatic hydrolysis. The transport of
anandamide inside the cell is a carrier-mediated activity, having been
shown to be a saturable, time- and temperature-dependent process that
involves some protein with high affinity and specificity for anandamide
(Beltramo, 1997). This transport process, unlike that of classical neurotransmitters, is Na+-independent and driven only by the concentration
gradient of anandamide (Piomelli, 1998). Although the anandamide transporter protein has not been cloned yet, its well characterized activity is
known to be inhibited by specific transporter inhibitors. Reuptake of 2-AG
is probably mediated by the same facilitating mechanism (Di Marzo,
1999a,b; Piomelli, 1999).
Once inside the cell, anandamide is hydrolyzed by a specific hydrolase, anandamide amidase (AEAase) or fatty acid amidohydrolase
(FAAH) (Desarnaud, 1995; Deutsch, 1993). This enzyme is membrane
associated and shows significant specificity for anandamide (Desarnaud,
1995; Lang, 1999).
There is some evidence that in cells with low anandamide amidase
activity, such as platelets and neutrophils, anandamide is inactivated by an
oxidative pathway involving 12(S)-lipoxygenase (Edgemond, 1998).
Metabolism of anandamide by enzymes of the arachidonic acid cascade
Cannabinergics
103
(Fig. 6) may be of physiological importance and may lead to possible
biologically active oxygenated anandamide analogs. These pathways have
not been explored yet. Nevertheless, it was shown that the 11-, 12-, and 15lipoxygenases recognize anandamide and catalyze its hydroxylation in
vitro (Hampson, 1995). Among the resulting oxygenated anandamides,
only the product of 11-lipoxygenase showed affinity to CB1 comparable to
that of anandamide (Hampson, 1995). The physiological relevance of this
finding, if any, is unknown at present.
Less explored is the role and metabolic fate of 2-AG. It is possible
that in many tissues, 2-AG is only an intermediate of a signaling pathway
that generates 1,2-diacylglycerol and arachidonic acid, two well-known
signaling molecules. In the brain however, 2-AG may have regulatory
roles, since it escapes immediate metabolism and accumulates in response
to stimuli-generated Ca2+ surges (Stella, 1997). These differences may arise
Figure 6 Anandamide metabolism: NAPE, N-arachidonylphosphatidyl-ethanolamides; PLD, phospholipase D; AEA, anandamide; AC, anandamide carrier
protein; AT, anandamide transporter; AEAase, anandamide amidase; AA,
arachidonic acid.
104
from differences in the involved isoenzymes or their levels of expression
from tissue to tissue.
Anandamide amidase recognizes and hydrolyzes 2-AG (Goparaju,
1999; Di Marzo, 1999; Lang, 1999); however, there is evidence for the
existence of another specific hydrolase [monoacylglycerol (MAG) lipase]
that hydrolyzes 2-AG (D. Piomelli and A. Makriyannis, 2000, personal
communication). In addition to this pathway, 2-AG diffuses rapidly into
the cell membrane where it could be either hydrolyzed to arachidonic acid
and glycerol or esterified back to phosphoglycerides (Di Marzo, 1999b).
IV. THE ENDOCANNABINOID SYSTEM
It is apparent that a series of critical research breakthroughs during the last
decade have unveiled a new significant biological assemblage, the endocannabinoid system. This system, which is evolutionarily well conserved,
consists of at least two receptor types, each with different localization and
functions; a family of endogenous ligands; and a specific molecular
machinery for the synthesis, transport and inactivation of these ligands.
Although information about this system is now emerging, many significant
questions still remain unanswered. The anandamide transporter and some
of the endocannabinoid metabolic enzymes have yet to be cloned. The
accomplishment of a highly quantitative and detailed mapping of the
endocannabinoid system will produce more information about its physiological roles. The advent of specific cannabinoid receptor antagonists
(Pertwee, 1995c), (Rinaldi-Carmona, 1994) has already facilitated pharmacological studies by enabling reversal of the endogenous cannabinoid
tone, as well as verification of the interaction of various agents with these
receptors. The first inhibitors of anandamide amidase (Boger, 2000;
Deutch, 1997a; Deutch, 1997b; Pertwee, 1995b) and its transporter (Beltramo, 1997; Christie, 2001; Wilson, 2001) are becoming important tools in
understanding the functions of the endocannabinoid system by producing
an hypercannabinoid state. Thorough understanding of this system and its
functions in physiological and disease conditions will likely lead to the
development of new therapeutics.
Invaluable tools for such studies are selective agents capable of
interacting with the protein members of the cannabinoid system and, in
turn, either activating or inhibiting them. Therefore, the study of the SAR
of each of these targets and the identification of differences in ligand
recognition comprises a task of great significance, one that can lead to the
development of highly selective cannabinergic agents. The term ‘‘canna-
Cannabinergics
105
Figure 7 Cannabinoid targets for drug design.
binergic’’ encompasses ligands that act on proteins of the endocannabinoid system, regardless of chemical classification or type of resultant
pharmacological activity. Therefore, this general term includes agents that
act on the cannabinoid receptors, either as agonists or antagonists, as well
as molecules that inhibit AEAase or the anandamide transporter (AT). The
therapeutic potential that emanates from modulating these proteins
renders them important yet unexploited targets for drug design and
development (Fig. 7).
All the aforementioned protein members of the cannabinoid system
are large, membrane-bound proteins; therefore, it is particularly difficult to
obtain direct information about their tertiary structure. Thus, at the
present time, structure-based drug design is not feasible. Detailed exploration of the SAR and subsequent ligand-based design are the most appropriate means for the development of molecular probes for these proteins.
V. MAJOR CLASSES OF CANNABINERGIC LIGANDS
Based on chemical structure, cannabinergic ligands are classified into five
major classes. Structures of representative members from each of the five
chemical classes are shown in Figures 8 to 11.
A. Classical Cannabinoids
Classical cannabinoids (CCs) are tricyclic terpenoid derivatives bearing
a benzopyran moiety. This class includes the natural product (–)-deltanine-tetrahydrocannabinol (Fig. 8, 1) and the other pharmacologically
active constituents of the plant Cannabis sativa.
106
Figure 8
Structures of representative classical cannabinoids.
Figure 9
Nonclassical cannabinoids (NCCs).
Cannabinergics
107
Figure 10 Representative cannabinergic aminoalkylindoles.
Many classical cannabinoid analogs have been synthesized and
evaluated pharmacologically and biochemically (Razdan, 1986; Mechoulam, 1999). The CC structural features that seem to be important for
cannabimimetic activity (Makriyannis, 1990) are as follows:
1.
The phenolic hydroxyl group, can be substituted by an amino
group but not by a thiol group. In contrast to the traditional CC
Figure 11 Structures of representative endocannabinoid analogs.
108
2.
3.
4.
5.
SAR, which considers the phenolic hydroxyl to be one of the
necessary pharmacophoric groups, analogs lacking it or bearing
it in its etherified form retain high receptor binding affinity, (e.g.,
analog 2a) especially for CB2 (Huffman, 1996).
The benzopyran ring is not essential for activity. The pyran
oxygen can be substituted by nitrogen or can be eliminated in
open-ring mono- or bisphenolic compounds. The recently
developed CB2-selective ligand HU-308 (5) is an example of
such a bicyclic cannabinoid (Hanus et al., 1999).
Neither the double bond nor the 9-methyl group are necessary
for activity.
The alkyl chain is probably the most essential CC pharmacophoric group. Increased biological activity results from
elongating the five-carbon delta-eight-THC chain to a sevencarbon chain substituted with 1V,1V- (e.g., 2) or 1V,2V-dimethyl or
with 1V,1V-cyclic moieties (e.g., 3, AMG3). Oxygen atoms
(ethers) and unsaturation (Papahatjis, 1998) within the chain,
or terminal halogens, carboxamido, and cyano groups are
well tolerated (Khanolkar, 2000).
An additional pharmacophore introduced in the nonclassical
cannabinoid series is the southern aliphatic hydroxyl (Makriyannis, 1990). A variation involves the highly potent classical/nonclassical cannabinoid hybrids (e.g., 4, AM919) (Drake, 1998).
B. Nonclassical Cannabinoids
A second class of cannabimimetics was developed at Pfizer, in an effort to
simplify the structure of CCs while maintaining or improving activity
(Johnson, 1986). This class includes bicyclic (e.g., 6) and tricyclic (e.g., 7)
analogs lacking the pyran ring of CCs (Fig. 9). These compounds are
collectively specified as ‘‘non-classical cannabinoids’’ (NCCs). The crystalline CP55,940 (6) and its tritiated analog show high affinity, efficacy, and
stereoselectivity to both cannabinoid receptors and have been used extensively as pharmacological tools. The key compound that led to the
discovery of CB1 was [3H]CP55,940 (Devane, 1988).
The structural resemblance of NCCs and CCs, as well as their
comparable SARs, indicate that they bind to CB1 in a similar fashion.
The side chain and the phenolic hydroxyl of an NCC are crucial for
activity. The hydroxypropyl chain of CP55,940 is not necessary for
Cannabinergics
109
activity. However, when present, its stereochemistry is important and
shows a strong preference for the beta relative configuration.
C. Aminoalkylindoles
The third chemical class of cannabinergics is that of aminoalkylindoles
(AAIs) (Fig. 10). They were developed at Sterling Winthrop as potential
nonsteroidal anti-inflammatory agents (Bell, 1991). These first analogs
exhibited antinociceptive properties that were eventually attributed to
interactions with the cannabinoid receptors. Compound 8 (WIN55212)
is a potent CB1 and CB2 agonist with high stereoselectivity and a slight
preference for CB2. AM630 (9), the first CB2-selective antagonist derived
from this class of compounds, was developed in our laboratory after longterm efforts to obtain such an inhibitor (Pertwee, 1995a). We have recently
reported the development of AM1241, a potent, highly CB2-selective
agonist (Malan, 2001).
This class of compounds differs from the first two by being considerably less lipophilic and more ‘‘druglike.’’ Labeling of CB1 with electrophilic AAIs almost abolished the receptor’s ability to bind to CP55,940,
indicating that AAIs and NCCs (as well as CCs) share at least some points
of interactions with CB1 (Yamada, 1996). Several models have attempted
to define the pharmacophoric equivalency between the functional groups
of AAIs, NCCs, and CCs (Xie, 1995), (Huffman, 1994). Although these
three different classes of cannabimimetics show similarities in their
binding with CB1, they differ considerably in the susceptibility of their
binding affinities to different Na+-modulated allosteric receptor states
(Houston, 1998). They also differ in their affinities to several CB1
mutants (Chin, 1998), as well as in the way they activate the receptor
(Houston, 1998). These differences may be explained by the existence of
more than one ligand binding motif, or by ligand binding to partially
overlapping but distinct receptor binding subsites, or even by induction
of different receptor conformational changes upon binding of different
ligands (Howlett, 1998a). It has been proposed that structurally dissimilar ligands may evoke different receptor–G-protein coupling (Houston,
1998). Therefore, analogs from different cannabinoid ligand classes may
evolve as selective pharmacological agents exhibiting only specific cannabimimetic effects.
Structural features of AAI important for cannabinergic activity are
the 3-aroyl moiety and the 1-chain, which must contain nitrogen, most
often in a heterocyclic ring (e.g., piperidino or morpholino). This chain can
110
be conformationally restricted as part of a six-membered ring fused to the
indole nucleus (D’Ambra, 1992).
D. Endocannabinoids
The class of the endogenous cannabinoids (endocannabinoids) was discovered in 1992 as molecules produced by mammalian cells with affinity for
the cannabinoid receptor (Devane, 1992). This class includes lipid molecules such as fatty acid ethanolamides, monoacylglycerols, and related
synthetic analogs. The two prototypes in this class are the ethanolamide of
arachidonic acid (anandamide) and 2-arachidonyl glycerol (2-AG). Its (R)1V-methylated analog, AM356 (10) (Fig. 11) shows higher affinity and
remarkable metabolic stability (Abadji, 1994). This analog, named Rmethanandamide, has been established as a standard CB1-selective agonist
in the cannabinoid field. The (R,R)-2,1V-dimethyl anandamide was
reported recently to exhibit a threefold improved affinity over R-methanandamide and significant enantioselectivity (Goutopoulos, 2001). Other
modifications that result in high CB1 affinity include the substitution of the
hydroxyl group with halogen, or the methyl group, and the substitution of
the terminal n-pentyl chain with the dimethylheptyl chain, reminiscent of
potent classical cannabinoid ligands (e.g., 12, O-1064) (Pertwee, 2000).
This compound class also includes some fatty acid analogs designed for
endocannabinoid targets other than the cannabinoid receptors. For
instance, arachidonyltrifluoromethylketone (ATFMK) (13) and hexadecylsulfonyl fluoride (14, AM374) are potent inhibitors of anandamide
amidase. The first inhibitor of the anandamide transporter to play an
important role in the discovery of this transport process was AM404 (15)
(Beltramo, 1997).
E. 1,5-Biarylpyrazoles
The fifth class, 1,5 biarylpyrazoles, was developed at Sanofi in 1994 from a
hit generated by high throughput screening for cannabinoid receptor
ligands (Rinaldi-Carmona, 1994). Compounds of this class act as cannabinoid receptor antagonists. Figure 12 shows SR141716A (16), which was
reported, simultaneously with AM630, as the first CB1 antagonist and has
since been used extensively as an important pharmacological tool.
SR141716A shows selectivity for CB1 and often acts as an inverse agonist
rather than a pure antagonist (Pertwee, 2000). Also developed at Sanofi,
SR144528 (17) acts as an antagonist/inverse agonist with selectivity for
CB2. A useful radioimaging agent in PET and SPECT studies [123I]
Cannabinergics
111
Figure 12 1,5 Biarylpyrazole cannabinoid receptor antagonists.
AM281, a 123I-labeled 1,5-biarylpyrazole was synthesized in our laboratory (Gatley, 1998).
VI. THERAPEUTIC POTENTIAL OF CANNABINERGIC
AGENTS
Most known cannabimimetics today have very broad effects on organ
systems, several of which are still not completely delineated. The ubiquitous pharmacology of cannabimimetics is one of the reasons for the
failure, thus far, of the clinical application of these drugs to reach its full
potential. The sections that follow summarize the effects of cannabinergics
on the various physiological systems and the possible therapeutic uses that
may arise from these biological activities.
A. Nervous System
The primary system of cannabimimetic activity is the nervous system. The
CB1 receptor is omnipresent in the brain, especially in areas that control
functions affected by cannabimimetics. One of the functions most pronouncedly influenced by cannabimimetics is motor behavior. Catalepsy,
immobility, ataxia, and impairment of complex behavioral acts after acute
administration of high doses of cannabimimetics are manifestations of
such motor effects (Pertwee, 1997). In lower doses cannabimimetics
produce the opposite effects. The very dense presence of CB1 in the
cerebellum and the basal ganglia, areas responsible for motor activity, is
112
congruent with these observations. The GABA function in the basal
ganglia is enhanced by CB1 agonists (Consroe, 1998). Cannabimimetics
seem to exert an important modulatory action in basal ganglia output
nuclei by inhibiting both inhibitory striatal input, which is tonically
inactive, and excitatory subthalamic input, which is tonically active
(Sanudo-Pena, 1999). The net cannabimimetic effect on motor activity
depends on the level of activity of each of these two functions. This may
explain the biphasic effect of cannabimimetics on motor behavior.
An important recent discovery has advanced the current understanding of how cannabimimetics are implicated in the control of motor
behavior (Giuffrida, 1999). Giuffrida et al. have reported that D2 activation in the striatum results in release of the endocannabinoid anandamide,
which in turn seems to mediate a negative feedback control, counteracting
dopamine-induced facilitation of motor activity (Giuffrida, 1999). Because
of these effects of cannabinergics on the basal ganglia and subsequently on
motor activity, it has been suggested that cannabinergics may be useful
agents in the treatment of motor disorders such as choreas, Tourette’s
syndrome, dystonias, and Parkinson’s disease (Consroe, 1998). In general,
by increasing hypokinetic features in the basal ganglia, CB1 agonists may
alleviate the various hyperkinetic manifestations, such as choreic movements, that characterize basal ganglia disorders. Direct evidence suggesting the involvement of CB1 in Huntington’s chorea is the extensive loss of
CB1 receptors in the substantia nigra and lateral globus pallidus (Glass,
1993). It is still unclear whether these observations are causative of
Huntington’s disease or its results. However, this finding alone argues that
a suitable CB1 ligand could potentially be useful as a diagnostic agent for
this chorea.
Furthermore, the presence of CB1 in the structures and pathways
associated with the pathophysiology of Tourette’s syndrome, and especially the functional link between CB1 and D1, D2, also argues that the
endocannabinoid system may have some involvement in this disorder as
well (Consroe, 1998). In addition, it has been suggested that activation of
CB1 receptors, also owing to their link with the dopaminergic system, may
reduce dyskinesia produced by L-DOPA in patients with Parkinson’s
disease (Brotsie, 1998).
The CB1 receptors present in the hippocampus, amygdala, and
cerebral cortex may be responsible for observations that cannabimimetics
are effective against some types of seizures (Consroe, 1998). The anticonvulsant and antispastic effects of cannabinoids are well documented,
however the mechanisms of these effects are still unclear (Nahas, 1999).
Cannabinergics
113
Orally administered cannabimimetics can relieve some of the symptoms of
multiple sclerosis and spinal cord injury such as muscle spasticity, pain,
tremor, nystagmus, and nocturia (Pertwee, 2000). Recent studies (Baker,
2000; Baker, 2001) have shown that exogenously administered cannabimimetics control spasticity in a multiple sclerosis (MS) model. Possible
implication of both CB1 and CB2 receptors has been suggested. Agents
that elevate anandamide levels by inhibiting AEAase (AM374) or AT
(AM404) also produced these antispastic effects indirectly. Cannabinoid
receptor antagonists blocked these antispastic effects. Respectively,
SR141716A and SR1445228, selective CB1 and CB2 antagonists/inverse
agonists, produced enhanced spasticity when administered alone to the
same animal model (Baker, 2000). Furthermore, it was evident that
endocannabinoids are released during episodes of MS, during which they
alleviate the spastic effects of the disease (Pertwee, 2000). These findings
confirm, at least to some extent, the anecdotal reports that marijuana
smoking alleviates the symptoms in MS patients and establishes cannabimimetics as exciting candidates for the development of agents that control
spasticity and other abnormalities resulting from some neurodegenerative
diseases. These agents may also control spasticity produced by spinal cord
injury by acting on spinal as well as on supraspinal mechanisms (Consroe,
1998). It has been suggested that the effect of cannabimimetics on the
release of glutamate in the substantia nigra appears to be the most
important supraspinal mechanism of cannabimimetic-induced control of
spasticity (Consroe, 1998).
The CB1-mediated inhibition of glutamate release in the hippocampus was also suggested to be the most likely mechanism of the neuroprotective effects of WIN5521,2 observed in both the global and focal
cerebral ischemia animal models (Nagayama, 1999). These effects were
stereoselective and were blocked by SR141716A. Therefore, cannabimimetics may find potential therapeutic utility in the treatment of disorders
resulting from cerebral ischemia, including stroke.
Another neuroprotective activity of cannabimimetics was shown to
be associated with the CB1-mediated inhibition of nitric oxide (NO) release
from rat microglial cells (Waksman, 1999). This study suggests cannabimimetics as potentially useful agents in brain injury resulting from
inflammatory neurodegenerative processes, especially those involving
activation of microglial cells, such as AIDS-encephalitis.
Another significant cannabinoid activity that is mediated by the
nervous system arises from the antinociceptive properties of these agents.
Compelling evidence suggests that cannabimimetics are effective in the
114
control of acute and chronic pain in a variety of antinociceptive tests in
animals (Martin, 1998). Synthetic cannabimimetics have been classified as
equal to morphine in potency and efficacy (Walker, 1999). The mechanism
of the cannabimimetic-induced analgesia is multifaceted and occurs at
several levels: (1) directly on spinal cord mechanisms (Walker, 1999); (2) in
supraspinal mechanisms, specifically in the thalamus and the periaqueductal gray (PAG) matters (Walker, 1999; Martin, 1998); and (3) in the
periphery, possibly involving CB1-like and CB2-like receptors (Calignano,
1998). Other systems, such as n and A opiate receptors, as well as spinal
noradrenergic mechanisms, seem to be involved in the cannabimimeticproduced analgesia (Walker, 1999). Evidence supports the suggestion that
cannabimimetics are effective in animal models of chronic pain, a type of
pain that is poorly managed by opioids (Walker, 1999). It has also been
suggested that CB1 agonists may be superior to morphine in suppressing
pain caused by nerve damage (Pertwee, 2000). This type of pain is signaled
by abnormal discharges of Ah and Adelta fibers, which are much more
populated by CB1 than A-opioid receptors.
Another category of CNS-mediated cannabinoid effects includes
alterations in cognition and memory. Cannabimimetics have been shown
to interfere with the mechanisms of long-term potentiation (LTP), a
candidate mechanism for learning and memory. They also alter presynaptic release of GABA and glutamate from hippocampal neurons (Hampson,
1998). Hippocampus, a structure rich in CB1, plays a major role in memory
processing, especially by enabling memory retrieval, whereas retrohippocampal areas with fewer CB1 receptors are responsible for memory
storage. Hippocampal lesions in rodents impair short-term memory.
Several behavioral studies have exhibited that cannabinoids disrupt information processing in the hippocampus, acting as ‘‘reversible’’ hippocampal lesions (Hampson, 1999). It is suggested that the role of CB1 in
these regions is to regulate storage information by switching hippocampal memory circuits (Hampson, 1998). The role of the cannabinoid
system in memory and cognition renders it a possible target for memory
and cognition enhancing agents. This possibility is strongly supported
by some recent advances in understanding the neurobiology of the
endocannabinoid system (Wilson, 2001; Christie, 2001; Kreitzer, 2001;
Ohno-Shosaku, 2001). Endocannabinoids were found to be the neurotransmitters responsible for the depolarization-induced suppression of
inhibition (DCI) and excitation (DCE). Since DCI enhances memory in
the hippocampus, drugs that inhibit the metabolism and especially the
transport of endocannabinoids are very likely to have a beneficial effect on
Cannabinergics
115
memory by increasing the levels of endocannabinoids at the sites where
DCI takes place (Christie, 2001). Direct cannabinoid receptor agonists
flood the endocannabinoid system, resulting in the well-known overall
disruptive effect in memory and cognition.
Cannabinoids are long known for their psychoactive and euphoric
‘‘high’’ effects and have been used for these properties for centuries. Their
addictive potential and mechanisms appear to be qualitatively and quantitatively different from those of other drugs of abuse. However, recent
studies indicate that cannabimimetics, similar to other addictive drugs,
activate the brain reward/reinforcement circuit (ventral tegmental area,
nucleus pallidus, and ventral pallidum) and produce reward-related
behaviors in laboratory animals (Gardner, 1998). Efforts to separate these
unwanted effects from the desired ones have had only limited success thus
far. This fact, along with the negative social perception of these drugs, has
been a major hindrance to the development of cannabinergic therapeutics.
However, the increasing understanding of the endocannabinoid system
presents us with possibilities for the design of selective agents. Indirect
activation of this system by increasing endocannabinoid levels only at the
sites where they are physiologically produced through inhibition of
endocannabinoid catabolism or transport may lead to increased selectivity
and fewer undesired effects than activation of the cannabinoid receptors
with direct agonists (Pertwee, 2000). Endocannabinoids such as anandamide were shown to have a much lower physical dependence potential
(Aceto, 1998).
Other well-known central cannabimimetic effects that nevertheless
are not well understood are hypothermia, appetite stimulation, and
antiemetic effects. Cannabimimetic-induced hypothermia is thought to
occur by decreasing the thermoregulatory set point through interactions
with the relevant hypothalamic centers (Pertwee, 1995b). Cannabimimetics also stimulate hunger in humans and animals, particularly for solid,
sweet tasting foods (Pertwee, 1995b). For this property, delta-9-THC
(marinol) is clinically used today for the management of AIDS-wasting
syndrome (Nahas, 1999). The advent of potent and CB1-selective ligands
lacking the CB2-mediated immunosuppressive properties may present
significant advantages over the currently used delta-9-THC in the treatment of AIDS patients who are already severely immunocompromised. It
is also conceivable that cannabinoid receptor antagonists may be proven
effective as appetite suppressants, as suggested by the results of a study
showing that SR141716A, a selective CB1 antagonist/inverse agonist,
suppressed rodent appetite for sucrose and ethanol (Arnone, 1998).
116
A second current clinical indication of cannabimimetics is their
antiemetic and antinausea effects, especially in cancer chemotherapy
patients. These effects are mediated above the level of vomiting reflex
and possibly through descending inhibitory connections to the lower brain
stem centers (Levitt, 1986).
B. Immune System
The discovery of the peripheral CB2 receptor, which localizes in cells of the
immune system, is very likely linked to the well-known immunosuppression of marijuana smokers.
Miskin (1985) found that delta 9-THC decreases host resistance to
herpes virus type 2 in mice and guinea pigs by decreasing both cellular and
humoral immunity. In vivo and in vitro studies indicate that macrophages
are the major targets of cannabinoids. delta 9-THC inhibits, in a dosedependent manner, the extrinsic antiviral activity of macrophages (Cabral,
1991). It was also shown that cannabinoids cause morphological changes
in macrophages (Cabral, 1991) and affect their phagocytic and spreading
ability (Spector, 1991).
The involvement of CB2 (and possibly of CB1) receptor(s) in the
immunosuppressive effects of cannabinoids is not proven yet. The localization of CB2 in cells of the immune system and especially in macrophages
and lymphocytes suggests that this receptor serves some immunoregulatory role(s). The first strong piece of evidence that implicates CB2 in such a
function came from Kaminski et al. (1994), who demonstrated that
cannabinoid-induced suppression of humoral immunity was partially
mediated through inhibition of adenylyl cyclase by a G-protein-coupled
mechanism that is pertussis toxin sensitive. Involvement of a membrane
perturbation mechanism in cannabinoid-induced immunosuppression is
also possible, especially in areas exposed to high drug concentrations, such
as lung alveolar macrophages of marijuana smokers (Cabral, 1999). The
involvement of the cannabinoid system in the regulation of the immune
system may suggest that cannabinergics could potentially serve as immunomodulatory agents. Although CB2 selective agents already exist, their
clinical potential in some immunomodulatory role will not be realized until
the CB2 physiological functions are better understood. Cannabidiol, a
cannabis terpenoid ingredient lacking the pyran ring as well as significant
binding affinity for CB1 and CB2, was shown to be an active antiinflammatory agent in the murine model of arthritis (Pertwee, 2000). The
molecular basis of this observation is still unknown.
Cannabinergics
117
C. Cardiovascular System
Cannabinoids reduce platelet aggregation and also produce tachycardia
and orthostatic hypotension due to peripheral vasodilation. A distinct,
CB1-like, receptor is found in the endothelium of rat mesenteric arteries
(Jarai, 1999). This receptor mediates a remarkable vasodilating effect after
activation by any of several CCs, anandamide, or some CB1-inactive CClike analog. This effect is NO independent and is inhibited by the CB1
antagonists, SR141716A and AM251 (Batkai, 2001), and also by cannabidiol. It is possible that exploitation of this new cannabinoid target may
lead to new types of hypotensive agents.
D. Reproductive System
Cannabinoids produce increased ring and chain chromosomal translocations and morphological abnormalities in mouse sperm, as well as reduction of sperm concentration in humans (Zimmerman, 1999). Strong
evidence indicates the presence of functional CB1, or CB1-like receptors,
in human sperm (Schuel, 1999). Furthermore, the endogenous cannabimimetic anandamide is produced in the human uterus and testes (Schuel,
1999). These findings along with several observations on cannabinoidinduced effects on reproductive functions suggest that the cannabinoid
system may be directly involved in the regulation of sperm production,
sperm motility, the acrosome reaction, and prevention of polyspermy
(Schuel, 1999). The endocannabinoid system in the uterus appears to play
a fundamental role in embryo implantation and early development.
Anandamide inhibits these processes and, therefore, regulation of its
levels seems to control the timing of these events (Paria, 1995). These
findings are also in line with recent clinical observations that correlate the
levels of AEAase expression with miscarriages in pregnant women
(Maccarone, 2000). Further understanding of the endocannabinoid functions in the reproductive system will open perspectives for exploitation of
cannabinergics for the treatment of some types of infertility or the development of contraceptives.
Cannabimimetics are also shown to affect reproductive and
metabolic functions indirectly by hormonal modulation through the
hypothalamic and pituitary regulatory centers. They are found to
reduce serum levels of the luteinizing hormone, prolactin, growth
hormone, and thyroid-stimulating hormone, and to increase corticotropin (Murphy, 1998).
118
E. Eye
Cannabinoids reduce intraocular pressure, probably by directly affecting
ocular fluid outflow pathways. The mechanism of this effect is unknown,
and its link to cannabinoid receptors has yet to be established (Green,
1999). Marijuana smoking is allegedly helpful to glaucoma patients, and
the potential use of cannabimimetics for the treatment of glaucoma has
long been recognized. New formulation technologies, as well as the advent
of less hydrophobic cannabimimetics, present us with opportunities to
overcome the challenge of local drug delivery to the eye.
F. Respiratory System
Cannabimimetics are known to produce bronchodilation, which is manifested by a marked increase in airway conductance and reduction in
airway resistance (Vachon, 1973). Although the mechanism of this activity
is not known, it probably does not directly involve adrenergic receptors.
Possible involvement of CB1A (a CB1 variant found in the lung) in
cannabinoid-induced bronchodilation is still unexplored (Shire, 1995).
Recently, it was shown that anandamide is released in the lung upon
Ca2+ stimulation and exerts a dual effect on bronchial response. It strongly
inhibits capsaicin-evoked bronchospasm and cough; however, it causes
bronchoconstriction in vagotomized rodents (Calignano, 2000). These
effects are mediated by CB1 receptors present in axon terminals of airway
nerves since they are blocked by SR141716A. This endocannabinoidmediated control of airway responsiveness may be exploited in the development of new antiasthmatic agents.
G. Gastrointestinal System
Cannabimimetics reduce the intestinal motility by a CB1-mediated inhibitory activity on acetylcholine release from autonomic fibers. An endocannabinoid, 2-AG, was isolated from dog intestine; however, its role there
remains unknown (Mechoulam, 1995a).
VII. CONCLUSIONS
With the discovery of anandamide and 2-arachidonyl glycerol as two new
families of endocannabinoids, cannabinoid research has taken major
Cannabinergics
119
strides toward arriving at an understanding of the molecular mechanism of
cannabinoid action. Currently, there are multiple characterized endocannabinoid proteins [at least two receptors, CB1 and CB2; an enzyme,
arachidonylethanolamide amidohydrolase (AEAase); and a transport
protein, anandamide transporter (AT)] as potential therapeutic targets
for the development of useful medications in the treatment of a multitude
of conditions such as drug addiction, pain, and motor disorders. A number
of ligands (receptor-selective agonists/antagonists, inverse agonists,
enzyme inhibitors, transport inhibitors) are also available which can serve
as important research tools for exploring the endocannabinoid biochemical pathways and their role in the modulation of behavior, memory,
cognition, and pain perception. This is significant progress, considering
that only about a decade ago the sites of action of cannabinoids had not yet
been identified and their molecular mechanism of action was still under
question. The future of endocannabinoid research is undoubtedly very
exciting and full of promise.
REFERENCES
1. Abadji V, Lin SY, Taha G, Griffin G, Stevenson LA, Pertwee RG, Makriyannis A. (R)-Methanandamide: a chiral novel anandamide possessing
higher potency and metabolic stability. J Med Chem 1994; 37:1889–1893.
2. Aceto MD, Scates SM, Razdan RK, Martin BR. Anandamide, an endogenous cannabinoid has a very low physical dependence potential. J Pharmacol Exp Ther 1998; 287:598–605.
3. Arnone M, Maruani J, Chaperon F. Selective inhibition of sucrose nd
ethanol intake by SR141716a, an antagonist of central cannabinoid (CB1)
receptors. Psychopharmacology 1998; 132:104–106.
4. Baker D, Pryce G, Croxford JL. Cannabinoids control spasticity and
tremor in a multiple sclerosis model. Nature 2000; 404:84–87.
5. Baker D, Pryce G, Croxford JL, Brown P, Pertwee RG, Makriyannis A,
Khanolkar A, Layward L, Fezza F, Bisogno T, Di Marzo V. Endocannabinoids control spasticity in a multiple sclerosis model. FASEB J 2001; 15:
300–302.
6. Batkai S, Jarai Z, Wagner J, Goparaju S, Varga K, Liu J, Wang L, Mirshahi
F, Khanolkar A, Makriyannis A, Urbaschek R, Garcia N Jr, Sanyal A,
Kunos G. Endocannabinoids acting at vascular CB1 receptors mediate the
vasodilated state in advanced liver cirrhosis. Nat Med 2001; 7:827–833.
7. Bell MR, D’Ambra TE, Kumar V, Eissentat MA, Herrmann JL, Wetzel
JR, Rosi D, Philion RE, Daum SJ, Hlasta DJ, Kulling RK, Ackerman JH,
Haubrich DR, Luttinger DA, Baixman ER, Miller MS, Ward SJ. Antinociceptive (aminoalkyl)indoles. J Med Chem 1991; 34:1099–1110.
120
8. Beltramo M, Stella N, Calignano A, Lin S, Makriyannis A, Piomelli D.
Functional role of high-affinity anandamide transport, as revealed by selective inhibition. Science 1997; 277:1094–1097.
9. Boger DL, Sato H, Lerner AE, Hedrick MP, Fesik RA, Miyauchi H, Wilkie
GD, Austin BJ, Patricelli MP, Cravatt BF. Exceptionally potent inhibitors
of fatty acid amide hydrolase: the enzyme responsible for degradation of
endogenous oleamide and anandamide. Proc Natl Acad Sci USA 2000; 97:
5044–5049.
10. Bouabula MBB, Rinaldi-Carmona M, Shire D, LeFur G, Casellas P. Stimulation of cannabinoid receptor CB1 induces krox-24 expression in human
astrocytoma cells. J Biol Chem 1995b; 270:13973–13980.
11. Bouabula M, Poinot.-Chazel C, Bourrie B, Canat X, Rinaldi-Carmona M,
LeFur G, Casellas P. Activation of mitogen-activated protein kinases by
stimulation of the central cannabinoid receptor CB1. Biochem J 1995a;
312:637–641.
12. Bramblett RD, Pann AM, Ballesteros JA, Reggio PH. Construction of 3D
model of the cannabinoid CB1 receptor: determination of helix ends and
helix orientation. Life Sci 1995; 56:1971–1982.
13. Breivogel CS, Childers SR. The functional neuroanatomy of brain cannabinoid receptors. Neurobiol Dis 1998; 5:417–431.
14. Brotsie JM. Adjuncts to dopamine replacement, a pragmatic approach to
reducing the problem of dyskinesia in Parkinson’s disease. Mot Disord
1998; 13:871–876.
15. Cabral GA. Marijuana and the immune system. In: Nahas GG et al., eds.
Marijuana and Medicine. Totowa, NJ: Humana Press, 1999.
16. Cabral GA, Vasquez R. Marijuana decreases macrophage antiviral and
antitumor activities. Adv Biosci 1991; 80:93–105.
17. Cadas H, di Tomaso E, Piomelli D. Occurrence and biosynthesis of endogenous cannabinoid precursor, N-arachidonyl phosphatidylethanolamine,
in rat brain. J Neurosci 1997; 17:1226–1242.
18. Cadas H, Gaillet S, Beltramo M, Venance L, Piomelli D. Biosynthesis of an
endogenous cannabinoid precursor in neurons and its control by calcium
and cAMP. J Neurosci 1996; 16:3934–3942.
19. Calignano A, LaRana G, Giuffrida A, Piomelli D. Control of pain initiation
by endogenous cannabinoids. Nature 1998; 394:277–281.
20. Calignano A, Katona I, Desarnaud F, Giuffrida A, LaRana G, Mackie K,
Freund TF, Piomelli D. Bidirectional control of airway responsiveness by
endogenous cannabinoids. Nature 2000; 408:96–101.
21. Childers SR, Sexton T, Roy MB. Effects of anandamide on cannabinoid
receptors in rat brain membranes. Biochem Pharmacol 1994; 47:711–715.
22. Chin CN, Lucas-Lenard J, Abadji V, Kendall DA. Ligand binding and
modulation of cyclic AMP levels depend on the chemical nature of residue
192 of the human cannabinoid receptor 1. J Neurochem 1998; 70:366–373.
Cannabinergics
121
23. Christie MJ, Vaughan CW. Cannabinoids act backwards. Nature 2001; 410:
527–530.
24. Consroe P. Brain cannabinoid systems as targets for the therapy of neurological disorders. Neurobiol Dis 1998; 5:534–551.
25. Crawley JN, Corwin RL, Robinson J, Felder CC, Derane WA, Axelrod J.
Anandamide, an endogenous ligand of the cannabinoid receptor, induces
hypomotility and Hypothermia in vivo in rodents. Pharmacol Biochem
Behav 1993; 46:967–972.
26. D’Ambra K, Estep G, et al. Conformational restricted analogues of pravadoline: nanomolar potent, enanantioselective, (aminoalkyl)indole agonists of the cannabinoid receptor. J Med Chem 1992; 35:124–135.
27. Deadwyler S, Hampson RE, Childers SR. Functional ignificance of cannabinoid receptors in brain. In: Pertwee RG, ed. Cannabinoid Receptors. New
York: Academic Press, 1995: 206–227.
28. Desarnaud F, Cadas H, Piomelli D. Anandamide amidohydrolase activity
in rat brain microsomes. J Biol Chem 1995; 270:6030–6035.
29. Deutsch D, Chin SA. Enzymatic synthesis and degradation of ananadmide,
a cannabinoid receptor agonist. Biochem Pharmacol 1993; 46:791–796.
30. Deutsch DG, Lin S, Hill WAG, Morse KL, Salehani D, Arreaza G, Omeir
RL, Makriyannis A. Fatty acid sulfonyl fluorides inhibit anandamide metabolism and bind to the cannabinoid receptor. Biochem Biophys Res Commun 1997; 231:217–221.
31. Deutsch DG, Makriyannis A. Inhibitors of anandamide breakdown. NIDA
Res Monog 1997; 173:65–84.
32. Devane W, Dysarz FA III, Johnson MR, Melvin LS, Howlett AC. Determination and characterization of a cannabinoid receptor in rat brain. Mol
Pharmacol 1988; 34:605–613.
33. Devane WA, Hanus L, Breuer A, Pertwee RG, Stevenson LA, Griffin G,
Gibson D, Mandelbaum A, Etinger A, Mechoulam R. Isolation and structure of a brain constituent that binds to the cannabinoid receptors. Science
1992; 258:1946–1949.
34. Devane WA, Axelrod J. Enzymatic synthesis of anandamide, an endogenous ligand for the cannabinoid receptor, by brain membranes. Proc Natl
Acad Sci USA 1994; 91:6698–6701.
35. Di Marzo V. 2-Arachidonyl-glycerol as an endocannabinoid: limelight for
a formerly neglected metabolite. Biochem (Moscow) 1998; 63:13–21.
36. Di Marzo V, Bisogno T, De Petrocellis L, Melck D, Orlando P, Wagner JA,
Kunos G. Biosynthesis and inactivation of the endocannabinoid 2-arachidonoylglycerol in circulating and tumoral macrophages. Eur J Biochem 1999;
264:258–267.
37. Di Marzo V, Bisogno T, De Petrocellis L, Melck D, Martin BR. Cannabimimetic fatty acid derivatives: anandamide family and other ‘endocannabinoids.’ Curr Med Chem 1999b; 6:721–744.
122
38. Drake DJ, Tius M, Jensen S, Busch-Petersen J, Kawakami JK, FernandezGarcia MC, Fan P, Makriyannis A. Classical/nonclassical hybrid cannabinoids: southern aliphatic chain-functionalized C-6h methyl, ethyl and
propyl analogues. J Med Chem 1998; 41:3596–3608.
39. Edgemond WS, Hillard CJ, Falck JR, Kearn CS, Cambell WB. Human
platelets and polymorphonuclear leukocytes synthesize oxygenated derivatives of arachidonylethanolamide (anandamide): their affinities for cannabinoid receptors and pathways of inactivation. Mol Pharmacol 1998; 54:
180–188.
40. Facci L, Dal Toso R, Romanello S, Buriani A, Skaper SD, Leon A. Mast
cells express a peripheral cannabinoid receptor with differential sensitivity
to anandamide and palmitoylethanolamide. Proc Natl Acad Sci USA 1995;
92:3376–3380.
41. Felder CC, Joyce KE, Briley EM, Mansouri J, Mackie K, Blond O, Lai Y,
Mitchell RL. Comparison of the pharmacology and signal transduction of
the human cannabinoid CB1 and CB2 receptors. Mol Pharmacol 1995; 48:
443–449.
42. Fride E, Mechoulam R. Pharmacological activity of the cannabinoid receptor agonist, anandamide, a brain constituent. Eur J Pharmacol 1993; 231:
313–314.
43. Gardner EL, Vorel R. Cannabinoid transmission and reward-related events.
Neurobiol Dis 1998; 5:501–533.
44. Gatley SJ, Lan R, Volkow ND, Papas N, King P, Wong CT, Gifford AN,
Pyatt B, Dewey SL, Makriyannis A. Imaging of the brain marijuana
receptor—Development of a radioligand that binds to cannabinoid CB1
receptors in vivo. J Neurochem 1998; 70:417–423.
45. Gerard, CM, Molerau, C, Vassart, G, Parmentier, M. Molecular cloning of
a human cannabinoid receptor which is also expressed in testis. Biochem J
1991; 279:129–134.
46. Giuffrida A, Parsons LH, Kerr TM, Rodriguez de Fonseca F, Navarro M,
Piomelli D. Dopamine activation of endogenous cannabinoid signaling in
dorsal striatum. Nat Neurosci 1999; 2:358–363.
47. Glass M, Faull RLM, Dregnow M. Loss of cannabinoid receptors in the
substantia nigra in Huntington’s disease. Neuroscience 1993; 56:523–527.
48. Goutopoulos A, Fan P, Khanolkar AD, Xie XQ, Lin SY, Makriyannis A.
Stereochemical selectivity of methanandamides for the CB1 and CB2 cannabinoid receptors and their metabolic stability. Bioorg Med Chem 2001;
9:1673–1684.
49. Goparaju SK, Ueda N, Taniguchi K, Yamamoto S. Enzymes of porcine
brain hydrolyzing 2-arachidonoylglycerol, an endogenous ligand of cannabinoid receptors. Biochem Pharmacol 2001; 57:417–423.
50. Green K. Marijuana and intraocular pressure. In: Nahas GE et al, eds.
Marijuana and Medicine. Totowa, NJ: Humana Press, 1999: 581–589.
Cannabinergics
123
51. Hampson, AJ, Hill WA, Zan Phillips M, Makriyannis A, Leung E, Eglen,
RM, Bornheim LM. Anandamide hydroxylation by brain lipoxygenase:
metabolite structures and potencies at the cannabinoid receptor. Biophys.
Biochim. Acta 1995; 1259:173–179.
52. Hampson RE, Deadwyler SA. Role of cannabinoid receptors in memory
storage. Neurobiol Dis 1998; 5:474–482.
53. Hampson RE, Deadwyler SA. Cannabinoids, hippocampal function and
memory. Life Sci 1999; 65:715–723.
54. Hanus L, Gopher A, Almog S, Mechoulam R. Two new unsaturated fatty
acid ethanolamides in brain that bind to the cannabinoid receptor. J Med
Chem 1993; 36:3032–3034.
55. Hanus L, Breuer A, Tchilibon S, Shiloah S, Goldenberg D, Horowitz M,
Pertwee RG, Ross RA, Mechoulam R, Fride, E. HU-308: a specific agonist
for CB2, a peripheral cannabinoid receptor. Proc Natl Acad Sci USA,
1999; 25:14228–14233.
56. Herkenham M, Lynn AB, Little MD, Johnson MR, Melvin LS, DeCosta
BR, Rice KC. Cannabinoid receptor localization in brain. Proc Natl Acad
Sci 1990; 87:1932–1936.
57. Hillard CJ, Cambell WB. Biochemistry and pharmacology of arachidonylethanolamide, a putative endogenous cannabinoid. J Lip Res 1997; 38:
2383–2398.
58. Houston D, Howlett AC. Differential receptor–G-protein coupling evoked
by dissimilar cannabinoid receptor agonists. Cell Signal 1998; 10:667–674.
59. Howlett AC. The CB1 cannabinoid receptor in the brain. Neurobiol Dis
1998a; 5:405–416.
60. Howlett, AC, Mukhopadhyay S, Shim J-Y, Welsh WJ. Signal transduction
of eicosanoid CB1 receptor ligands. Life Sci 1999; 65:617–625.
61. Howlett AC, Song C, Berglund BA, Wilken GH, Pigg JJ. Characterization
of CB1 cannabinoid receptors using receptor peptide fragments and sitedirected mutagenesis. Mol Pharmacol 1998b; 53:504–510.
62. Huffman JW, Dai D, Martin BR, Compton DR. Design, synthesis and
pharmacology of cannabimimetic indoles. Bioorg Med Chem Lett 1994; 4:
563–566.
63. Huffman JW, Yu S, Showalter V, Abood ME, Wiley JL, Compton DR,
Martin BR, Bramblett RD, Reggio PH. Synthesis and pharmacology of a
very potent cannabinoid lacking a phenolic hydroxyl with high affinity for
the CB2 receptor. J Med Chem 1996; 39:3875–3877.
64. Jarai Z, Wagner J, Varga K, Lake KD, Compton, DR, Martin BR, Zimmer
AM, Bonner TI, Buckley NE, Mezey E, Rajdan RK, Zimmer A, Kunos G.
Cannabinoid-induced mesenteric vasodilation through an endothelial site
distinct form CB1 or CB2 receptors. Proc Natl Acad Sci USA 1999;
96:14136–14141.
65. Jarbe TV, Lamb RJ, Makriyannis A, Lin S, Goutopoulos A. Delta 9-THC
124
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
training dose as a determinant for (R)-methanandamide generalization in
rats. Psychopharmacology 1998; 140:519–522.
Johnson MR, Melvin LS, The discovery of nonclassical cannabinoid analgesics. In: Mechoulam C, ed. Cannabinoids as Therapeutic Agents. Boca
Raton, FL: CRC Press, 1986: 121–145.
Kaminski NE, Koh WS, Yang KH, Lee M, Kessler FK. Suppression of the
humoral immune response by cannabinoids is partially mediated through
inhibition of adenylate cyclase by a pertussis toxin-sensitive G-proteincoupled mechanism. Biochem Pharmacol 1994; 48:1899–1908.
Khanolkar AD, A., V.; Lin SY, Hill WAG, Taha G, Abouzid K, Meng Z,
Fan P, Makriyannis A. Head group analogs of arachidonylethanolamide,
the endogenous cannabinoid ligand. J Med Chem 1996; 39:4515–4519.
Khanolkar AD, Palmer SL, Makriyannis A. Molecular probes for the cannabinoid receptors. Chem Phys Lipids 2000; 108:37–52.
Kreitzer AC, Regehr WG. Retrograde inhibition of presynaptic calcium
influx by endogenous cannabinoids at excitatory synapses onto Purkinje
cells. Neuron 2001; 29:717–727.
Lang W, Qin C, Lin C, Khanolkar AD, Goutopoulos A, Fan P, Abouzid
K, Meng Z, Biegel D, Makriyannis A. Substrate specificity and stereoselectivity of rat brain microsomal anandamide amidohydrolase. J Med
Chem 1999; 42:896–902.
Levitt M. Cannabinoids as antiemetics in cancer chemotherapy. In: Mechoulam R, ed. Cannabinoids as Therapeutic Agents. Boca Raton, FL: CRC
Press, 1986: 73–79.
Little PJ, Compton DR, Johnson MR, Melvin LS, Martin BR. Pharmacology and stereoselectivity of structurally novel cannabinoids in mice.
J Pharmacol Exp Ther 1988; 247:1046–1051.
Loh HH, Law PY. The role of membrane lipids in receptor mechanisms.
Annu Rev Pharmacol Toxicol 1980; 20:201–234.
Maccarone M, Valenise H, Bari M, Lazzarin N, Romani C, Finazzi-Agro
A. Relationship between decreased anandamide hydrolase concentrations
in human lymphocytes and miscarriage. Lancet 2000; 235:1326–1329.
Mackie K, Hille B. Cannabinoids inhibit N-type calcium channels in neuroblastoma–glioma cells. Proc Natl Acad Sci 1992; 89:3825–3829.
Mackie K, Devane WA, Hille B. Anandamide, an endogenous cannabinoid, inhibits calcium currents as a partial agonist in N-18 neuroblastoma
cells. Mol Pharmacol 1993; 44:498–503.
Makriyannis A, Banijamali A, Van der Schyf C, Jarrell H. The role of
cannabinoid stereochemistry and absolute configuration and the orientation
of D9-THC in the membrane bilayer. Structure–activity relationships of
cannabinoids. In: Rapaka RS, Makriyannis A, eds. Interactions of Cannabinoids with Membranes. National Institute on Drug Abuse Research
Monograph 79. Rockville, MD: US Department of Health and Human
Services, 1987.
Cannabinergics
125
79. Makriyannis A, Rapaka RS. The molecular basis of cannabinoid activity.
Life Sci 1990; 24:2173–2184.
80. Makriyannis A. In: Pertwee RG, ed. Cannabinoid Receptors. The Role of
Cell Membranes in Cannabinoid Activity. New York: Academic Press,
1995: 87–115.
81. Malan TP Jr, Ibrahim MM, Deng H, Liu Q, Mata HP, Vanderah T,
Porreca F, Makriyannis A. CB2 cannabinoid receptor-mediated peripheral
analgesia. Pain In press.
82. Martel P, Makriyannis A, Mavromoustakos T, Kelly K, Jeffrey KR. Topography of tetrahydrocannabinol in model membranes using neutron diffraction. Biochim Biophys Acta 1993; 1151:51–58.
83. Martin BR, Compton DR, Thomas BF, Prescott WR, Little PJ, Razdan
RK, Johnson MR, Melvin LS, Mechoulam R, Ward SJ. Behavioral, biochemical and molecular modeling evaluations of cannabinoid analogs in
mice. Pharmacol Biochem Behav 1991; 40:471–478.
84. Martin B, Lichtman AH. Cannabinoid transmission and pain perception.
85. Martin BR, Mechoulam R, Razdan RK. Discovery and characterization of
endogenous cannabinoids. Life Sci 1999; 65:573–595.
86. Matsuda LA, Lolait SJ, Brownstein, MJ, Young AC, Bonner TI. Structure
of a cannabinoid receptor and functional expression of the cloned cDNA.
Nature 1990; 346:561–564.
87. Mechoulam R, Ben-Shabat S, Hanus L, Ligumsky M, Kaminski NE,
Schatz AR, Gopher A, Almog S, Martin B, Compton DR, Pertwee RG,
Griffin G, Bayewitch M, Barg J, Vogel Z. Identification of an endogenous
2-monoglyceride, present in canine gut, that binds to cannabinoid receptors.
Biochem Pharmacol 1995a; 50:83–90.
88. Mechoulam R, Devane WA, Glaser R. Cannabinoid geometry and biological activity. In: Nahas GG et al., eds. Marijuana and Medicine. Totowa, NJ:
Humana Press, 1999: 65–90.
89. Mechoulam R, Feigenbaum E. The unpaved road to the endogenous brain
cannabinoid ligands, the anandamides. In: Pertwee RG, ed. Cannabinoid
Receptors. New York: Academic Press, 1995b: 233–258.
90. Mechoulam R, Gaoni Y. Isolation, structure and partial synthesis of an
active constituent of hashish. J Am Chem Soc 1967; 86:1646–1647.
91. Miskin EM, Cabral GA. Delta-9-tetrahydrocannabinol decreases host resistance to herpes simplex virus type 2 vaginal infection in B6C3F1 mice. J
Gen Virol 1985; 66:2539–2549.
92. Munro S, Thomas KL, Abu-Shaar M. Molecular characterization of a peripheral receptor for cannabinoids. Nature 1993; 365:61–65.
93. Murphy LL, Munoz RM, Adrian BA, Villnua MA. Function of cannabinoid receptors in the neuroendocrine regulation of hormone secretion.
94. Nagayama T, Sinor AD, Simon RP, Chen J, Graham SH, Jin K, Green-
126
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
berg DA. Cannabinoids and neuroprotection in global and focal cerebral
ischemia. J Neurosci 1999; 18:2987–2995.
Nahas GG, Sutin KM, Harvey DJ, Agurell S, eds. In: Nahas GG, et al.,
eds. Marijuana and Medicine. Totowa, NJ: Humana Press, 1999.
Nye JS, Voglmaier S, Martenson RE, Snyder SH. Myelin basic protein is
an endogenous inhibitor of the high affinity cannabinoid binding site in
brain. J Neurochem 1988; 50:1170–1178.
Ohno-Shosaku T, Maejima T, Kano M. Endogenous cannabinoids mediate
retrograde signals from depolarized postsynaptic neurons to presynaptic
terminals. Neuron 2001; 29:729–738.
Papahatjis DP, Kourouli T, Abadji V, Goutopoulos A, Makriyannis A.
Pharmacophoric requirements for cannabinoid side chains, multiple bond
and C1V-substituted delta-8-tetrahydrocannabinols. J Med Chem 1998; 41:
1195–1200.
Paria BC, Das SK, Dey SK. The preimplantation embryo is a target for
cannabinoid ligand–receptor signaling. Proc Natl Acad Sci USA 1996; 92:
9430–9464.
Pate E, Jarvinsen K, Utri P, Jarho P, Jarvinen, T. Ophthalmic arachidonylethanolamide decreases intraocular pressure in normotensive rabbits. Curr
Eye Res 1995; 14:791–797.
Paton WDM. Pharmacology of marijuana. Annu Rev Pharmacol 1975; 15:
191–220.
Pertwee RGS, L. A.; Elrick DB, Mechoulam R, Corbett AD. Inhibitory
effects of certain enantiomeric cannabinoids in the mouse vas deferens and
the myenteric plexus preparation of guinea-pig small intestine. Br J Pharmacol 1992; 105:980–984.
Pertwee RG. Pharmacological, physiological and clinical implications of
the discovery of cannabinoid receptors. In: Pertwee RG, ed. Cannabinoid
Receptors. New York: Academic Press, 1995a; 69–83.
Pertwee RG. Pharmacology of cannabinoid CB1 and CB2 receptors. Pharmacol Ther 1997; 74:129–180.
Pertwee RG. Cannabinoid receptor ligands: clinical and neuropharmacological considerations, relevant to future drug discovery and development.
Exp Opin Invest Drugs 2000; 9:1–19.
Pertwee RG, Fernardo SR, Griffin G, Abadji V, Makriyannis A. Effect of
phenylmethylsulphonyl fluoride on the potency of anandamide as an inhibitor of electrically evoked contractions in two isolated tissue preparations.
Eur J Pharmacol 1995b; 272:73–78.
Pertwee RG, Griffin G, Fernando SR, Li X, Hill A, Makriyannis A.
AM630, a competitive cannabinoid receptor antagonist. Life Sci 1995c; 56:
1949–1955.
Pertwee RG, Griffin G, Hanus L, Mechoulam R. Effects of two endogenous fatty acid ethanolamides on mouse vasa deferentia. Eur J Pharmacol
1994; 259:115–120.
Cannabinergics
127
109. Peters H, Nahas GG. A brief history of four millenia (B.C. 2000–A.D. 1974).
In: Nahas GG et al., eds. Marijuana and Medicine. Totowa, NJ: Humana
Press, 1999: 3–7.
110. Piomelli D, Beltramo M, Glasnapp S, Lin SY, Goutopoulos A, Xie XQ,
Makriyannis A. Structural determinants for recognition and translocation
by the anandamide transporter. Proc Natl Acad Sci USA 1999; 10:5802–
5807.
111. Piomelli D, Beltramo M, Giuffrida A, Stella N. Endogenous cannabinoid
signaling. Neurobiol Dis 1998; 5:462–473.
112. Razdan RK. Structure–activity relatioships in cannabinoids. Pharmacol
Rev 1986; 38:75–134.
113. Rinaldi-Carmona M, Bath F, Heaulme M, Shire D, Calandra B, Congy
C, Martinez S, Maruani J, Neliat G, Caput D, Ferrara P, Soubrie P,
Breliere JC, Lefur G. SR14176A, a potent and selective antagonist of the
brain cannabinoid receptor. FEBS Lett 1994; 350:240–244.
114. Sanudo-Pena MC, Tsou K, Walker M. Motor actions of cannabinoids in
the basal ganglia output nuclei. Life Sci 1999; 65:703–713.
115. Schmid P, Paria BC, Krebsbach R, Schmid H, Dey SK. Changes in anandamide levels in mouse uterus are associated with uterine receptivity
for embryo implantation. Proc Natl Acad Sci USA 1997; 94:4188–4192.
116. Schuel H, Goldstein E, Mechoulam R, Zimmerman AM, Zimmerman S.
Anandamide (arachidonylethanlamide), a brain cannabinoid receptor agonist, reduces sperm fertilizing capacity in sea urchins by inhibiting the acrosome reaction. Proc Natl Acad Sci USA, 1994; 91:9460–9464.
117. Schuel H, Chang MC, Burkman LJ, Picone R, Makriyannis A, Zimmerman AM, Zimmerman S. Cannabinoid receptors in sperm. Marijuana and
Medicine. In: Nahas GG et al., eds. Totowa, NJ: Humana Press, 1999:
317–325.
118. Shire DC, Carillon C, Kaghad M, Calandra B, Rinaldi-Carmona M, Le Fur
G, Caput D, Ferrara P. An amino-terminal variant of the central cannabinoid receptor resulting from alternative splicing. J Biol Chem 1995; 270:
3726–3731.
119. Smith PB, Compton DR, Welch SP, Razdan RK, Mechoulam R, Martin
BR, The pharmacological activity of anandamide, a putative endogenous
cannabinoid, in mice. J Pharmacol Exp Ther 1994; 270:219–227.
120. Spector S, Lancz G. Suppression of human macrophage function in vitro
by D9-tetrahydrocannabinol. J Leukocyte Biol 1991; 50:423–426.
121. Stella N, Schweitzer P, Piomelli D. A second endogenous cannabinoid
that modulates long-term potentiation. Nature 1997; 388:773–778.
122. Sugiura T, Kotaka T, Nakane S, Miyashita T, Kondo S, Suhara Y,
Takayama H, Waku K, Seki C, Baba N, Ishima Y. Evidence that the cannabinoid CB1 receptor is a 2-arachidonylglycerol receptor. J Biol Chem
1999; 274: 2794–2801.
123. Tao Q, Abood ME. Mutation of a highly conserved aspartate residue in
128
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
the second transmembrane domain of the cannabinoid receptors, CB1 and
CB2, disrupts G-protein coupling. J Pharmacol Exp Ther 1998; 285:651–
658.
Ueda N, Kurahashi Y, Yamamoto S, Tokunaga T. Partial purification and
characterization of the porcine brain enzyme hydrolyzing and synthesizing
anandamide. J Biol Chem 1995; 270:23823–23827.
Vachon L, Fitzerald MX, Solliday NH, Gould IA, Gaensler EA. Singledose effect of marijuana smoke: bronchial dynamics and respiratory-center
sensitivity in normal subjects. N Engl J Med 1973; 288:985–989.
Varga KLK, Martin BR, Kunos G. Novel antagonist implicates the CB1
receptor in the hypotensive action of anandamide. Eur J Pharmacol 1995;
278:279–283.
Vogel Z, Barg R, Levy D, Saya E Heldman, Mechoulam R. Anandamide, a
brain endogenous compound, interacts specifically with cannabinoid receptors and inhibits adenylate cyclase. J Neurochem 1993; 61:352–355.
Wagner JA, Varga K, Jarai Z, Kunos G. Mesenteric vasodilation mediated
by endothelial anandamide receptors. Hypertension 1999; 33:429–34.
Waksman Y, Olson SM, Carlisle SJ, Cabral, GA. The central cannabinoid
receptor (CB1) mediates inhibition of nitric oxide production by rat microglial cells. J Pharmacol Exp Ther 1999; 288:1357–1366.
Walker JM, Hohmann G, Martin WJ, Strangman NM. The neurobiology
of cannabinoid analgesia. Life Sci 1999; 65:665–673.
Wilson RI, Nicoll RA. Endogenous cannabinoids mediate retrograde signaling at hippocampal synapses. Nature 2001; 410:588–592.
Xie X-Q, Eissentat M, Makriyannis A. Common cannabimimetic pharmacophoric requirements between aminoalkylindoles and classical cannabinoids. Life Sci 1995; 56:1963–1970.
Yamada K, Rice KC, Flippen-Anderson JL, Eissenstat MA, Ward SJ,
Johnson MR, Howlett AC. (Aminoalkyl)indole isothiocyanates as potential electrophylic affinity ligands for the brain cannabinoid receptor. J Med
Chem 1996; 39:1967–1974.
Zimmerman MA, Zimmerman S, Raj AY, Effects of cannabinoids on spermatogenesis in mice. In: Nahas GG et al, eds. Marijuana and Medicine.
Totowa, NJ: Humana Press, 1999: 347–358.
5
Development of PET
and SPECT Radioligands
for Cannabinoid Receptors
S. John Gatley, Andrew N. Gifford, and Yu-Shin Ding
Brookhaven National Laboratory, Upton, New York, U.S.A.
Nora D. Volkow
NIDA, Bethesda, Maryland, U.S.A.
Ruoxi Lan, Qian Liu, and Alexandros Makriyannis
I.
INTRODUCTION
Marijuana is the most commonly illegal drug of abuse in the United
States, but relatively little known about how activation of cannabinoid receptors leads to the psychoactive effects desired by its abusers, or
whether receptor densities are altered in addiction or detoxification, or in
other disease states. Futhermore, the two known G-protein-coupled
receptors for cannabinoids (CB1, which is found in brain and some
peripheral tissues, and CB2, whose distribution is believed to be restricted
to immunological tissues in the periphery) are potentially important
targets for drug development. Various dosage forms of D9-tetrahydrocannabinol (THC), the major psychoactive constituent of marijuana, as well
as novel compounds, are under investigations by several major drug companies. For example, recently developed subtype-selective antagonists
of cannabinoid receptors such as SR141716A may have beneficial drug
129
130
Gatley et al.
actions in treating memory impairment and other disorders. A publication by the Institute of Medicine/National Academy of Sciences entitled
‘‘Marijuana and Medicine: Assessing the Science Base’’ discusses some of
the issues involved in this area.
Positron emission tomography (PET) is an imaging method used to
measure the regional distribution and kinetics of chemical compounds
labeled with short-lived positron-emitting isotopes such as 11C (half-life =
20 min) and 18F (110 min). It thus enables direct measurement of
components of the neurochemical systems in the living human brain
[1,2]. The SPECT (single-photon emission-computed tomography) methodology can also be used to measure some of the same components of
neurochemical systems as PET. As discussed later, SPECT is inferior to
PET in terms of spatial resolution, sensitivity, and quantitation. On the
other hand, SPECT methodology is less expensive than PET and is far
more widely available because of its advantages in clinical nuclear
medicine. These include the use of radionuclides of longer half-life, such
as 123I (13 h) and 99mTc (6 h).
Both PET and SPECT have been used in studies of several drugs to
image functional consequences of acute or chronic drug treatment, using
radiotracers that measure changes in factors such as blood flow, glucose
metabolic rate, or dopamine release. They have also been used to image
changes in the apparent brain concentrations of neuroreceptors with which
the drugs of abuse directly or indirectly interact. For a review of this area,
see Gatley and Volkow [3]. The brain – dopamine system, which may be
involved in the actions of all drugs of abuse, has been an important target
of PET studies [4].
II. IMAGING THE EFFECTS OF CANNABINOIDS
ON METABOLISM AND BLOOD FLOW
Relatively few human imaging studies have evaluated the effects of marijuana or THC on metabolism or blood flow. Acute intravenous THC in
both normal controls and habitual marijuana users led to increased an
increased regional cerebral metabolic rate (CMR) in the cerebellum. This
increase is positively correlated both with concentrations of THC in the
plasma and with the intensity of the subjective sense of intoxication [5].
In a 1997 PET/[15O]water study with 32 abusers [6], THC dose-dependently increased cerebral blood flow (CBF) in the frontal regions, insula
PET Imaging
131
and cingulate gyrus, and subcortical structures, with somewhat greater
effects in the right hemisphere. Self-ratings of THC intoxication were
correlated most markedly with the right frontal region. Most subjects
exhibited increased rCBF in cerebellum. However, those whose cerebellar CBF decreased also had a significant alteration in time sense [7].
The average increase in rCMR after THC administration was less in
marijuana users than in controls, and users had lower cerebellar metabolism than the controls at baseline [8]. Thus the cerebellum shows the
greatest metabolic increase in response to acute THC and responds to
chronic marijuana exposure with a decrease in baseline CMR. Habitual
users but not controls responded to THC administration with increased
rCMR in prefrontal cortex, orbitofrontal cortex, and basal ganglia. In contrast to the robust effects of THC on relative rCMR, changes in global
CMR in response to THC were quite variable, with increases, decreases,
and no changes seen in equal numbers of subjects. There was also variability in subjective effects, which were pleasurable for most subjects but
either minimal or unpleasant (anxiety or paranoia) for others.
The involvement of the cerebellum in the psychoactive effects of
marijuana and in changes in rCMR is consistent with the view that THC
interacts with the high concentration of CB1 receptors in this brain area.
Decreases in the cerebellar rCMR in habitual marijuana users may reflect
the effects of chronic exposure to the drug. Functions known to be
associated with the cerebellum, such as motor coordination, proprioception, and learning, are adversely affected both during acute marijuana
intoxication and in habitual users.
The cannabinoid CB1 receptor is the binding site in the brain for
D9-tetrahydrocannabinol, the active principle of marijuana. This Gprotein-coupled receptor is abundant in specific brain areas including
the cerebellum, the hippocampus, and the outflow nuclei of the basal
ganglia. An in vivo radioligand for the CB1 receptor would allow us to
evaluate disease and drug-induced changes in cannabinoid receptor
densities, and possibly to investigate relationships between receptor
occupancy by agonist and antagonist ligands and their behavioral and
toxic effects. Such studies would contribute not only to our understanding of the neural basis of marijuana abuse, but also to medication
development, since pharmacological manipulation of the CB1 receptor
system might prove useful in conditions such as chronic pain and
multiple sclerosis. In addition to compounds with the classical cannabinoid skeleton, recently developed high affinity agonists and antagonist of
the brain cannabinoid receptor may serve as lead compounds for PET
132
Gatley et al.
Figure 1 Structures of representatives of classes of compounds that blind to
cannabinoid receptors.
and SPECT radiotracers. These have nonclassical cannabinoid, aminoalkylindole, anandamide, and pyrazole strructures (Fig. 1) [9 – 13].
Autoradiographic studies with tritiated CP55,940 and other high
affinity agonists (see, e.g., Ref. 14) demonstrated high concentration of
cannabinoid receptors in the basal ganglia and especially in its outflow
nuclei, the globus pallidus and the substantia nigra. High concentrations
are also found in the hippocampus and the cerebellum. The cerebral
cortex also contains appreciable concentrations of cannabinoid receptors, the highest being the cingulate gyrus. Some other regions including
most of the brain stem and the thalamus, contain low or negligible
concentrations. The pattern of distribution of cannabinoid receptors in
many brain regions is similar to that of dopamine D1 receptors, which
has led to the suggestion that a function of the cannabinoid system may
be to induced modulate brain dopaminergic activity [15].
III. ATTEMPTS TO DEVELOP RADIOLIGANDS
The first attempt to develop a PET radioligand for imaging brain CB1
receptors involved modification of D8-THC by labeling with fluorine-18 in
the hydrocarbon side chain [16,17], as shown in Figure 2. Unfortunately,
PET Imaging
133
this compound did not produce PET images that showed any particular
regional pattern of brain localization when injected into a baboon. It
showed poor uptake but was widely distributed in the brain. Clearance of
radioactivity that did appear to enter the brain was rapid. Furthermore,
uptake of radioactivity in the skull was apparent, which suggested in vivo
decomposition of the radiotracer, leading to the production of labeled
fluoride ion, which then accumulated in bone. It is likely therefore that the
PET images represented only nonspecific uptake of the tracer with a
negligible component due to specific binding to cannabinoid receptors.
Studies with [18F] D8-THC supported the view that a successful
radiotracer must have adequate metabolic stability and a fairly high
affinity for the CB1 receptor to ensure that radioactivity is retained in
brain tissues long enough for tomographic measurement. Furthermore,
a good radiotracer should exhibit high uptake into the brain. This is
likely to be a difficulty for cannabinoid receptor radioligands, since
these molecules are extremely lipophilic. High log P values are generally
associated with poor blood – brain barrier penetrability, presumably
because they remain dissolved in lipid structures in the blood during
transit through the brain capillary bed. The nonclassical cannabinoid
CP55,940, the aminoalkylindole WIN55, 212-2, and THC are reported
to possess log P values of about 6, 5, and 7, respectively. Even the
lowest log P value in this series (5) has been associated with poor brain
penetration in other classes of molecules (see, e.g., Refs. 18, 19).
Furthermore, although [3H]WIN55,212-2 has an affinity about 10-fold
higher than THC, it does not exhibit preferential localization in CB1receptor-rich areas of the mouse brain when injected intravenously
(Gifford et al., unpublished).
Figure 2 Incorporation of fluorine-18 into 5V-[18F]fluoro-D8-THC.
134
Gatley et al.
Figure 3 Relationships between structures of SR141716A, AM251, and AM281.
The introduction of the diarylpyrazole CB1 receptor antagonist
SR141716A [12] immediately suggested that exploitation of this class of
molecules might lead to development of a successful CB1 receptor radioligand. Not only is the affinity of SR141716A of the order of 100-fold
higher than that of THC, but the structure suggested that the log P value
would be considerably lower. In addition, SR141716A is highly selective
for the brain cannabinoid receptor (CB1) relative to the CB2 receptor
found in cells of the immune system. It is also an antagonist, which is a
potential advantage because, in the binding of antagonists of G-proteincoupled receptors, there is no discrimination between receptors in different
affinity states. This is unlike the situation with agonists, which bind predominantly to a high-affinity state of the receptor. Finally, SR141716A
contains three chlorine atoms, suggesting that replacement of one of these
with a radioactive iodine atom might produce a compound with the desired
properties.
Our ‘‘mark I’’ pyrazole radioligand, code-named AM251, was synthesized in nonradioactive and radioactive forms [20,21]. Following
intravenous injection in mice and rats, the radioiodinated compound
concentrated preferentially in brain areas known to contain densities
of CB1 receptors [22]; however, it failed to enter the brain in SPECT
experiments conducted with baboons [23]. On the hypothesis that this
failure was associated with too high a log P value, we synthesized a
related, ‘‘mark II’’ radioligand with an additional structural modification. This was replacement of the piperidine ring of SR141716A and
AM251 with the more polar morpholino ring (Fig. 3). This compound
AM281, was able to visualize CB1 receptors in baboon (Fig. 4) and
rodent (Fig. 5) brains in vivo [23].
PET Imaging
135
Figure 4 Sagittal sections of the brains of two baboons injected intravenously
with [123I]AM251 (left) or [123I]AM281 (right). These experiments indicated that
AM281 is to penetrate the baboon brain much more readily than AM251.
Figure 5 Ex vivo autoradiography of [123I]AM281 in rat brain gave distribution
patterns that were essentially identical to in vitro autoradiographs obtained using
tritiated high affinity agonists.
136
Gatley et al.
Figure 6 Comparison of the behavior of radioidinated AM251 and AM281
in mice.
The degree of failure of AM251 to enter the baboon brain, as assessed
by SPECT scanning, was surprising because in rats and mice AM251 does
clearly enter the brain. Figure 6 presents a comparison of AM251 and
AM281 in mice. In absolute terms, the graphed data should be interpreted
cautiously because the experiments with the two radioligands were not
conducted simultaneously, and other experimental details were not identical [22,23]. However, apparently greater brain uptake of AM281 at early
times is consistent with its smaller log P value. Moreover, it is clear that
there was significant clearance of AM281 between 30 and 120 min, whereas
there was no significant difference between the 30, 60, and 120 min data
points for AM251. The in vivo brain uptake data, which show a more
prolonged retention of AM251, were thus consistent with the in vitro
binding data, which indicated that AM251 has an approximately threefold
higher affinity for the CB1 receptor than AM281. The mouse data shown
in Figure 6, however, did not predict the large difference seen in baboons
(Fig. 4). It may be that tight binding of AM251 to a specific blood protein,
rather than a greater distribution of AM251 into lipophilic blood components, is responsible for its low brain penetration in baboons.
PET Imaging
137
Although our SPECT studies with [123I]AM281 provided a proof of
principle that in vivo imaging of CB1 receptors in primates is possible, the
behavior of this radioligand is far from ideal in that, relative to radiotracers
used to study other neurochemical systems, its uptake in the brain is low
and its clearance is rapid. These factors limit the count rate obtained in
radionuclide imaging studies, and thus the quality of the images. Because
PET is more sensitive than SPECT by an order of magnitude, a CB1
receptor radioligand labeled with fluorine-18 with pharmacokinetic properties to those of similar AM281 would probably be quite acceptable for
human use. On the other hand, a radioiodinated compound for use with
SPECT would be more useful if it had at least the initial brain uptake of
AM281, but a higher affinity, to ensure a longer clearance time. Since our
initial published work with AM281, several other candidate radioligands
have been prepared [24 – 26]. However, to our knowledge no reports of
human studies have appeared.
IV. ONGOING WORK
In our own laboratories, we have continued to synthesize and evaluate new
labeled cannabinoid receptor radioligands. One of these is [18F]AM284,
where the labeled atom is part of a fluoropentyl group on position 1 of the
pyrazole ring (Fig. 7; see also Table 1). Although this (unpublished) study
demonstrated in vivo binding of AM284 to CB1 receptors, as shown by
the fact that co-injection with SR141716A reduced brain binding, the
Figure 7 Structure of an 18F-labeled pyrazole ligand; for brain uptake (see Table
1).
138
Gatley et al.
Table 1 Brain Uptake Data for the
18
F-labeled Pyrazole Ligand of Figure 7
Injected activity in whole brain (%)a
Conditions
15 min
60 min (vehicle)
60 min (+ SR14176A)
AM281
AM284
0.80 F 0.05
0.66 F 0.05*
0.33 F 0.05***
0.07 F 0.008
0.028 F 0.004**
0.020 F 0.001****
*
p < 0.003; cf. 15 min time point.
p < 0.001; cf. 15 min time point.
***
p < 0.001; cf. vehicle.
****
p < 0.035; cf. vehicle.
a
Values are the mean FSD (n = 5).
**
binding was less than one-tenth that of AM281 measured simultaneously
in a dual-isotope experiment. It would therefore not be a practicable PET
radioligand. Other fluorine-18 and radioiodinated cannabinoid receptor
ligands are being synthesized and studied in our laboratories.
While we have not yet started human PET or SPECT studies, we have
used AM281 to conduct fundamental studies of the CB1 receptor system in
Figure 8 Comparison of ability of WIN 55, 212-2 to sedate mice (triangles) and
to block specific binding of [131I]AM281 (squares).
PET Imaging
139
rodents. For example, during experiments in mice designed to measure the
occupancy of CB1 receptors associated with physiological effects of
exogenous cannabinoids, we found that doses of WIN 55,212-2 that
induced profound sedation did not reduce binding of AM281 to cerebellum and hippocampus (Fig. 8). This observation indicates that the
occupancy of the CB1 receptor necessary for physiological effects of
cannabinoids is very low [27].
Experiments in superfused hippocampal slices prepared from rats
were then conducted to compare inhibition of acetylcholine release by the
cannabinoid receptor agonist WIN55,212-2 with inhibition of AM281
binding by this agonist. The results (Fig. 9) show that half-maximal
response is achieved at less than 1% occupancy, confirming that the
agonist occupancy necessary to produce a physiological response is very
low in the CB1 system [27].
A consequence of a very large receptor reserve for CB1 receptors
would be that PET or SPECT could not be used to image the occupancy of
the CB1 receptor by biologically significant doses of agonist drugs. This is
Figure 9 Graph of inhibition of acetylcholine release in superfused hippocampal slices versus CB1 receptor occupancy estimated from inhibition of
[131I]AM281 binding.
140
Gatley et al.
because even high doses of an agonist would not displace an appreciable
fraction of the radiotracer binding. On the other hand, these experiments
indicated that the binding of exogenous CB1 receptor antagonist radioligands should not be affected by changes in the levels of endogenous
ligands such as anandamide. This would thus remove a possible confounding factor in imaging experiments designed to detect changes in
cannabinoid receptor densities.
In the light of these experiments with WIN55,212-2, it is of interest to
speculate on the degree of occupancy of brain CB1 receptors that is
achieved by doses of THC that induce desired effects in human during
the smoking of a marijuana cigarette. THC is a partial agonist with an
efficacy of 20 to 25% [28,29], so that it would act at a higher receptor
occupany than a full agonist like WIN55,212-2. Intravenous doses of 0.5
mg/kg THC are effective in humans [30], and if 1% of the injected dose is
distributed in the brain [31], this would correspond to a concentration of
about 15 nmol/L. Applying the mass action equation and assuming that an
in vitro Kd value for the CB1 receptor of 100 nM is appropriate, and a Bmax
value of 100 nmol/L, an occupancy of 7% is estimated. However, it is likely
that the fraction of THC available for binding to the receptor in vivo is
quite small, since as an extremely lipophilic molecule, it will be distributed
in brain membranes. This is expected to increase the effective Kd value in
vivo and so lower the estimate of occupancy, possibly by more than one
order of magnitude. These considerations, therefore, suggest that only a
very small proportion of the brain CB1 receptors need be activated to
induce psychoactive effects in humans, consistent with our results in mice,
and that PET studies will not be able to measure the degree of occupancy
achieved by marijuana smokers.
Similar studies were done to evaluate the relationship between level
of occupancy of the CB1 receptor by nonradioactive AM281 and the
degree to which the antagonist AM281 was able to reverse the sedative
effect of the agonist WIN55,212-2 [32]. The AM281 effectively restored
the activity to normal levels (Fig. 10). In addition, AM281 alone was
found to significantly stimulate locomotor activity between 1 and 2 h
after its administration (Fig. 11). Both the antagonism of the effect of
WIN55,212-2 and the effect of AM281 alone increased progressively with
doses up to 0.3 mg/kg AM281, but did not further increase at 1 mg/kg.
A 50% occupancy of the CB1 receptor, as assessed by inhibition of
[131I]AM281 binding, was achieved at a dose of 0.45 mg/kg. These data
are consistent with prior in vitro indications that AM281 is a CB1 receptor antagonist or inverse agonist [33] and that AM281 inhibits an
PET Imaging
141
Figure 10 Effect of increasing doses of AM281 on binding of [131I]AM281 in
cerebellum and hippocampus, using brain stem as reference tissue.
endogenous cannabinoid tone. This baseline activity of the CB1 system
might be maintained either by constitutive activity of the receptor [34], or
by endogenous agonists such as anandamide [11]. These experiments [32]
indicate that PET could be used to measure the degree of occupancy of
CB1 receptors by antagonist or inverse agonist drugs in the human brain,
if these drugs turn out to have useful therapeutic effects, such as reducing
memory loss in the elderly [35 –37].
V. CONCLUSIONS
Our development of [123I]AM281, an antagonist radioligand for brain
cannabinoid receptors, has allowed us to image this receptor for the first
time in vivo. Ex vivo autoradiographic experiments have been conducted
in rodents, and SPECT studies have been conducted in baboons. Research
142
Gatley et al.
Figure 11 Stimulation of locomotor activity at CB1 receptor occupancy levels
calculated from the effect of increasing doses of AM281 on binding of
[131I]AM281; squares, antagonism of the sedative effect of WIN55,212-2 at 0
to 15 min; triangles, induction of hyperactivity by AM281 alone at 61 to 120 min.
continues to develop superior radioligands for SPECT research and also
to develop CB1 receptor radioligands that can be labeled with positronemitting nuclides for PET. The results of the animal work to date will
provide the foundation for using AM281 and other cannabinoid receptor
radioligands in human imaging experiments. It is predicted from our
animal data that psychoactive or medicinal doses of agonists such as THC
will not alter CB1 receptor radioligand binding in the human brain. On
the other hand, PET or SPECT is likely to be useful in determining the
degree of CB1 receptor occupancy necessary for therapeutic effects of
antagonist drugs, as well as in evaluating CB1 receptor changes in
addiction and in other diseases.
ACKNOWLEDGMENTS
This research was carried out at the Brookhaven National Laboratory
under contract DE-AC02-98CH10886 with the U.S. Department of
PET Imaging
143
Energy and supported by its Office of Health and Environmental
Research. The research was also supported by awards from the National
Institute on Drug Abuse to AM (DA 07515, DA 09158) and to ANG
(DA 12412).
REFERENCES
1. Wolf AP, Fowler JS. Positron emission tomography. Biomedical research
and clinical application. Neuroimaging Clin N Am 1995; 5: 87 – 101.
2. Fowler JS, Wolf AP. The heritage of radiotracers for positron emission
tomography. Acta Radiol Suppl 1990; suppl 374: 13 – 16.
3. Gatley SJ, Volkow ND. Addiction and imaging of the living human brain.
Drug Alcohol Depend 1998; 51: 97 – 108.
4. Volkow ND, Fowler JS, Gatley SJ, Logan J, Wang GJ, Ding YS, Dewey
SL. PET evaluation of the dopamine system of the human brain. J Nucl
Med 1996; 37: 1242 – 1256.
5. Volkow ND, Gillespie H, Mullani N, Tancredi L, Grant C, Ivanovic M,
Hollister L. Cerebellar metabolic activation by delta-9-tetrahydrocannabinol in human brains. A study with positron emission tomography and 18F-2deoxy-2-fluoro-D-glucose. Psychiat Res 1991; 40: 69 – 80.
6. Mathew RJ, Wilson WH, Coleman RE, Turkington TG, DeGrado TR.
Marijuana intoxication and brain activation in marijuana smokers. Life Sci
1997; 60: 2075 – 2089.
7. Mathew RJ, Wilson WH, Turkington TG, Coleman RE. Cerebellar activity
and disturbed time sense after THC. Brain Res 1998; 797: 183 – 189.
8. Volkow ND, Gillespie H, Mullani N, Tancredi L, Grant C, Valentine A,
Hollister L. Brain glucose-metabolism in chronic marijuana users at baseline and during marijuana intoxication. Psychiatr Res Neuroimaging 1996;
67: 29 – 38.
9. Jansen EM, Haycock DA, Ward SJ, Seybold VS. Distribution of
cannabinoid receptors in rat brain determined with aminoalkylindoles.
Brain Res 1992; 575: 93 – 102.
10. Herkenham M, Lynn AB, Little MD, Johnson MR, Melvin LS, DeCosta
BR, Rice KC. Cannabinoid receptor localization in brain. Proc Natl Acad
Sci USA 190; 87: 1932 – 1936.
11. Felder CC, Briley EM, Axelrod J, Simpson JT, Mackie K, Devane WA.
Anandamide, an endogenous cannabimimetic eicosanoid, binds to the
cloned human cannabinoid receptor and stimulates receptor-mediated signal
transduction. Proc Natl Acad Sci USA 1993; 90: 7656 – 7660.
12. Rinaldi-Carmona M, Barth F, Heaulme M, Shire D, Calandra B, Congy C,
Martinez S, Maruani J, Neliat G, Caput D, Ferrara P, Soubrie P, Breliere
144
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Gatley et al.
JC, Lefur G. SR141716A, a potent and selective antagonist of the brain
cannabinoid receptor. FEBS Lett 1994; 350: 240 – 244.
Devane WA, Breuer A, Sheskin T, Ja¨rbe TU, Eisen MS, Mechoulam, R.
A novel probe for the cannabinoid receptor. J Med Chem 1992; 35:
2065 – 2069.
Herkenham M, Lynn AB, Johnson MR, Melvin LS, de Costa BR, Rice
KC. Characterization and localization of cannabinoid receptors in rat
brain: a quantitative in vitro autoradiographic study. J Neurosci 1991; 11:
563 – 583.
Biegon A, Kerman I. In: Biegon A, Volkow ND, eds. Sites of Drug Action
in the Human Brain. Boca Raton, FL: CRC Press, 1995:65 – 74.
Charalambous A, Marciniak G, Shiue CY, Dewey SL, Schlyer DJ, Wolf
AP, Makriyannis A. PET studies in the primate brain and biodistribution in
mice using ( – )-5V-18F-delta 8-THC. Pharmacol Biochem Behav 1991; 40:
503 – 507.
Marciniak G, Charalambous A, Shiue CY, Dewey SL, Schlyer DJ, Wolf
AP. Fluorine-18 labeled tetrahydrocannabinol: synthesis and PET studies in
a baboon. J Labelled Compd Radiopharm 1991; 30: 413 – 415.
John E, Bott A, Green M. Preparation and biodistribution of copper-67
complexes with tetradentate Schiff-base ligands. J Pharm Sci 1994; 83:
587 – 590.
Welch MJ, Chi DY, Mathias CJ, Kilbourn MR, Brodack JW, Katzenellenbogen JA. Biodistribution of N-alkyl and N-fluoroalkyl derivatives of
spiroperidol: radiopharmaceuticals for PET studies of dopamine receptors.
Appl Radiat Isot 1986; 13: 523 – 526.
Lan R, Gatley SJ, Makriyannis A. Preparation of iodine-123 labeled
AM251, a potential SPECT radioligand for the brain cannabinoid CB1
receptor. J Labelled Comp Radiopharm 1996; 38: 875 – 881.
Lan R, Liu Q, Fan P, Lin S, Fernando SR, McCallion D, Pertwee R,
Makriyannis A. Structure – activity relationships of pyrazole derivatives as
cannabinoid receptor antagonists. J Med Chem 1999; 42: 769 – 776.
Gatley SJ, Gifford AN, Volkow ND, Lan R, Makriyannis A. Iodine-123
labeled AM251: a radioiodinated ligand which binds in vivo to the
mouse brain CB1 cannabinoid receptor. Eur J Pharmacol 1996; 307:
301 – 308.
Gatley SJ, Lan R, Volkow ND, Pappas N, King P, Wong CT, Gifford AN,
Pyatt B, Dewey SL, Makriyannis A. Imaging the brain marijuana receptor:
development of a radioligand that binds to cannabinoid CB1 receptors in
vivo. J Neurochem 1998; 70: 417 – 423.
Mathews WB, Ravert HR, Musachio JL, Frank RA, Rinaldi-Carmona M,
Barth F, Dannals RF. Synthesis of [18F] SR144385: a selective radioligand
for positron emission tomographic studies of brain cannabinoid receptors. J
Labelled Compd Radiopharm 1994; 42: 589 – 596.
PET Imaging
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
145
Margonelli A, Angelini G, Attina A, Cartoni A, Finazzi Agro A,
Maccarrone M. New Effective chemical radiosynthesis of (2)-15,6,8,9,
11,12,14,15, -VH(N)1-Arachidonoyl-glycerol, (2-AG). A Putative endongenous cannabinoid receptor ligand. J Labelled Compd Radiopharm 1999;
suppl 1: S787 – S789.
Tamagnan GXJL, Gao Y, Amici LA, Baldwin RM, Innis RB. Potential
SPECT radioligand for cannabinoid receptor. J Labelled Compd Radiopharm 1999; suppl 1: S191 – S192.
Gifford AN, Bruneus M, Gatley SJ, Lan R, Makriyannis A, Volkow ND.
Large receptor reserve for cannabinoid actions in the central nervous
system. J Pharmacol Exp Ther 1999; 2: 478 – 483.
Sim LJ, Hampson RE, Deadwyler SA, Childers SR. Effects of chronic
treatment with delta-9-tetrahydrocannabinol on cannabinoid-stimulated
[35S]GTPgamma-S autoradiography in rat brain. J Neurosci 1996; 16:
8057 – 8066.
Burkey TH, Quock RM, Consroe P, Roeske WR, Yamamura HI. delta-9Tetrahydrocannabinol is a partial agonist of cannabinoid receptors in
mouse brain. Eur J Pharmacol 1997; 323: R3 – R4.
Hollister LE, HK, G. The benzodiazepine receptor antagonist flumazenil
does not block clinical effects of delta-9-tetrahydrocannabinol. Life Sci 1990;
47: 1655 – 1650.
Schou J, Prockop LD, Dahlstrom G, Rohde C. Penetration of delta-9tetrahydrocannabinol and 11-OH-delta-9-tetrahydrocannabinol through
the blood – brain barrier. Acta Pharmacol Toxicol (Copenh) 1977; 41:
33 – 38.
Cosenza M, Gifford AN, Gatley SJ, Pyatt B, Liu Q, Makriyannis A,
Volkow ND. Locomotor activity and occupancy of brain cannabinoid CB1
receptors by the antagonist/inverse agonist AM281. Synapse 2000; 38: 477 –
482.
Gifford AN, Tang Y, Gatley SJ, Volkow ND, Lan R, Makriyannis A.
Effect of the cannabinoid receptor SPECT agent, AM 281, on hippocampal acetylcholine release from rat brain slices. Neurosci Lett 1997; 238:
84 – 86.
Bouaboula M, Perrachon S, Milligan L, Canat X, Rinaldi-Carmona M,
Portier M, Barth F, Calandra B, Pecceu F, Lupker J, Maffrand JP, Le Fur
G, Casellas P. A selective inverse agonist for central cannabinoid receptor
inhibits mitogen-activated protein kinase activation stimulated by insulin or
insulin-like growth factor 1. Evidence for a new model of receptor/ligand
interactions. J Biol Chem 1997; 272: 22330 – 22339.
Collins DR, Pertwee RG, Davies SN. Prevention by the cannabinoid
antagonist, SR141716A, of cannabinoid-mediated blockade of long-term
potentiation in the rat hippocampal slice. Br J Pharmacol 1995; 115:
869 – 870.
146
Gatley et al.
36. Katona I, Sperlagh B, Magloczky Z, Santha E, Kofalvi A, Czirjak S,
Mackie K,Vizi ES, Freund TF. GABAergic interneurons are the targets of
cannabinoid actions in the human hippocampus. Neuroscience 2000; 100:
797 – 804.
37. Gifford AN, Ashby CR. Electrically evoked acetylcholine release from
hippocampal slices is inhibited by the cannabinoid receptor agonist, WIN
55212-2, and is potentiated by the cannabinoid antagonist, SR 141716A.
J Pharmacol Exp Ther 1996; 277: 1431 – 1436.
6
Structural and Pharmacological
Aspects of Peptidomimetics
Peter W. Schiller, Grazyna Weltrowska, Ralf Schmidt,*
Thi M.-D. Nguyen, Irena Berezowska, Carole Lemieux,
Ngoc Nga Chung, Katharine A. Carpenter,* and Brian C. Wilkes
Clinical Research Institute of Montreal, Montreal, Quebec, Canada
I.
INTRODUCTION
Small linear neuropeptides are structurally flexible molecules, capable of
existing in a number of different conformations of comparably low energy.
This structural flexibility precludes the determination of the bioactive
conformation in solution and, furthermore, may be responsible for the
lack of receptor selectivity of many of the naturally occurring peptide
hormones and neurotransmitters, since conformational adaptation to
different receptor topographies takes place. In recent years, the introduction of conformational constraints into peptides either locally or at a
particular amino acid residue (Na or Ca methylation, substitution of
dehydro- or cyclic amino acids, etc.) or more globally (peptide cyclizations)
emerged as a successful concept in the design of peptide analogues and
peptidomimetics. In many cases, conformationally restricted peptide
analogues showed high receptor selectivity and greatly improved conformational integrity. That different receptor types for a given peptide
hormone or neurotransmitter differ from one another in their conforma-
*Present address: AstraZeneca Research Centre Montreal, St. Laurent, Quebec, Canada.
147
148
Schiller et al.
tional requirements toward the peptide ligand was first unambiguously
demonstrated in the case of opioid receptors by comparing the receptor
affinity profiles of a cyclic enkephalin analogue and its corresponding
open-chain analogue [1]. Furthermore, conformationally restricted peptide analogues often showed enhanced stability against enzymatic degradation because a scissile peptide bond in the native peptide may no longer
be susceptible to enzymatic cleavage when incorporated in a conformationally constrained structure. Conformational restriction of peptides
represents a first step in the rational approach to develop peptidomimetics
because conformationally constrained analogues may serve as relatively
rigid templates for further structural modification aimed at removing some
of the less attractive peptide structural features (e.g., peptide backbone
replacements). In this chapter we illustrate these principles with examples
from the opioid peptide field.
The existence of three major classes (A, y, n) of opioid receptors is
now well established [2,3]. Furthermore, the results of classical pharmacological testing and of opioid receptor binding studies indicate that various
opioid receptor subtypes might also exist [4,5]. Recent work in the cloning
of opioid receptors confirmed the existence of a y [6,7], a n [8], and a A [9]
receptor, but so far has not led to the identification of receptor subtypes.
Like all other G-protein-linked receptors, the three opioid receptors show
seven putative transmembrane helices and considerable sequence similarity (60–70%) among themselves. Highest sequence identity is observed in
the membrane-spanning segments and in the intracellular loops, whereas
lower homology is seen in the extracellular regions. The ligand binding site
is thought to be located in the cavity within the transmembrane domain of
the receptors, whereas the extracellular loops may act as filters for the
ligands, thus possibly playing a role in ligand selectivity. The precise threedimensional structures of the opioid receptors remain elusive because
x-ray diffraction or NMR spectroscopic analysis cannot be used for their
determination. However, theoretical analyses led to approximate models
that are of interest for ligand docking studies and for the testing of
hypotheses related to ligand design [10].
Since the discovery of the enkephalins in 1975 [11] a large number of
endogenous opioid peptides have been detected in mammals, and at
present three distinct families of opioid peptides are known (for a review,
See Ref. 12). These are the enkephalins, the endorphins (a-, h-, and g-), and
the dynorphins and neoendorphins. The recently discovered endomorphins [13] also may represent endogenous opioid peptides. Peptides with
opioid activity have also been isolated from tryptic digests of milk casein
149
(h-casomorphins) [14] and from frog skin—dermorphins [15] and deltorphins [16]. Like the morphine-related opiates, opioid peptides produce a
large spectrum of central and peripheral effects, including analgesia,
tolerance and physical dependence, respiratory depression, euphoria,
dysphoria and hallucinations, sedation, feeding and other behavioral
effects, hypothermia/hyperthermia, miosis, effects on tumor growth, control of release of several peptide hormones and catecholamines, effects on
transit in the gut, and various cardiovascular effects. It has not yet been
possible to establish clear-cut relationships linking specific opioid receptor
types to distinct opioid effects. This is mainly because until recently, potent,
stable agonists and antagonists with high specificity for the various
receptor types have not been available. The results of several studies
suggest that the A receptor plays a primary role in mediating analgesia;
however, there is some evidence that y and n interactions may result in
analgesic effects as well. Among the various isolated organ preparations
used in the in vitro bioassays, the guinea pig ileum (GPI) contains primarily
A receptors but also n receptors; in the mouse vas deferens (MVD) y
receptors are predominant, even though A and n receptors are also present
(for a review, see Ref. 17). It is now well recognized that the binding of
agonists to opioid receptors leads to inhibition of adenylate cyclase via
interaction with a guanine nucleotide regulatory protein [18].
The various endogenous opioid peptides resulting from processing of
the three mammalian precursor molecules display only limited selectivity
toward the different receptor types [12]. The only naturally occurring
opioid peptides with high receptor specificity discovered so far are the yselective deltorphins [16] and, possibly, the endomorphins [13]. Medicinal
chemists and peptide chemists have made numerous efforts to develop
opioid agonists and antagonists with improved receptor selectivity. Substantial progress has been made in the development of selective nonpeptide
opioid receptor ligands (for a review, see Ref. 19). The design of opioid
peptide analogues with high receptor selectivity has also been very
successful (for reviews, see Refs. 20 and 21). Structural modification of
the enkephalins, h-casomorphin (morphiceptin), dermorphin, and the
deltorphins through various amino acid and end group substitutions
produced linear analogues that turned out to be highly selective A agonists,
y agonists, or y antagonists. The preparation of numerous linear analogues
of dynorphin A led to several potent n agonists with high n-receptor
selectivity (for a review, see Ref. 22), whereas potent and selective n opioid
antagonists structurally derived from dynorphin A have not yet been
reported. A most successful strategy in opioid peptide analogue design
150
Schiller et al.
has been the incorporation of conformational constraints through peptide
cyclizations or substitution of conformationally restricted amino acids. In
particular, conformational restriction of the enkephalins or the dermorphins and deltorphins produced potent and highly selective A agonists, y
agonists and y antagonists.
Agonists and antagonists showing high specificity for a particular
opioid receptor class are valuable as pharmacological tools and may also
have potential as therapeutic agents. Futhermore, it has been recognized
that the development of opioid compounds with mixed agonist/antagonist
properties may lead to improved analgesics with minimal side effects. In
this chapter we discuss the development of highly selective A agonists and y
antagonists, the first known compounds with mixed A agonist/y antagonist
properties and a new class of dipeptide y agonists.
II. M OPIOID AGONISTS
A. Cyclic Peptides with M-Agonist Properties
The first conformationally restricted opioid peptide reported in the
literature was the cyclic enkephalin analogue H-Tyr-c[-D-A2bu-Gly-PheLeu-] (A2bu = a,g-diaminobutyric acid), which showed considerable
preference for A receptors over y receptors [23]. Homologues of this
compound containing a D-ornithine or D-lysine residue in place of DA2bu also were A-selective agonists [24]. Analogues of H-Tyr-c[-D-LysGly-Phe-Leu-] having one or two reversed amide bonds showed further
improved A selectivity and excellent stability against enzymatic degradation [25]. One of the most selective cyclic opioid peptides with A agonist
properties reported to date is the dermorphin-related tetrapeptide H-TyrD-Orn-Phe-Asp-NH2 [26]. A theoretical conformational analysis performed with this compound revealed that the 13-membered peptide ring
structure was highly constrained and that the lowest energy conformer was
characterized by a tilted stacking interaction between the Tyr1 and Phe3
aromatic rings [27]. Expansion of the peptide ring structure in this
analogue, as achieved by replacement of Asp with Glu, resulted in the
compound H-Tyr-D-Orn-Phe-Glu-NH2, which showed only slight preference for A receptors over y receptors [28]. The results of a molecular
dynamics simulation carried out with this analogue showed that its 14membered peptide ring structure had moderate structural flexibility, while
the exocyclic Tyr1 residue and the Phe3 side chain enjoyed considerable
orientational freedom [29].
j
j
j
j
151
To limit the structural flexibility of these two aromatic residues,
conformationally restricted analogues of Phe and/or Tyr were substituted
[28,30] (Fig. 1).
Replacement of Phe in the parent peptide with the cyclic phenylalanine analogue 2-aminoindan-2-carboxylic acid (Aic) resulted in a
compound that showed only four times lower A affinity but 65 times lower
affinity for y receptors and, consequently, markedly improved A selectivity
(K yi /K Ai =49.6) (Table 1). The analog H-Tyr-D-Orn-Atc-Glu-NH2, containing the conformationally constrained phenylalanine analogue 2-aminotetralin-2-carboxylic acid (Atc) at the 3 position of the peptide sequence,
also retained high A-agonist potency and showed further improved A
selectivity. Interestingly, the diastereomeric D-Atc3 analogue also displayed good A receptor affinity and high A selectivity. This observation is
in contrast to the weak affinity observed with the D-Phe3 analogue in
comparison to the L-Phe3 parent peptide. Thus, stereospecificity was lost as
a consequence of side chain conformational restriction, presumably
because the D-Atc3 analogue binds to the receptor in a manner different
from that of the D-Phe3 analogue. Replacement of Tyr1 in the cyclic parent
peptide with 6-hydroxy-2-amino-tetralin-2-carboxylic acid (Hat) produced the compound H-Hat-D-Orn-Phe-Glu-NH2, with only about three
times reduced affinity for A and y receptors. Its diastereomer, H-D-Hat-DOrn-Phe-Glu-NH2, was a full agonist at both the A and the y receptor but
showed a substantial decrease in potency. Finally, the conformationally
highly constrained analogue H-( D , L )-Hat-D -Orn-Aic-Glu -NH 2 also
showed high A-receptor affinity and marked A selectivity. This compound
essentially contains only two freely rotatable bonds and represents the
most rigid, rationally designed opioid peptidomimetic reported to date.
j
j
j
j
j
j
Figure 1
j
j
Structural formulas of cyclic analogues of phenylalanine and tyrosine.
152
Schiller et al.
Table 1 Opioid Receptor Affinities of Cyclic Dermorphin Analoguesa
Compound
j
j
H-Tyr-D-Orn-Phe-Glu-NH2
H-Tyr-D-Orn-D-Phe-Glu-NH2
H-Tyr-D-Orn-Phe-Glu-NH2
H-Tyr-D-Orn-Atc-Glu-NH2
H-Tyr-D-Orn-D-Atc-Glu-NH2
H-Hat-D-Orn-Phe-Glu-NH2
H-D-Hat-D-Orn-Phe-Glu-NH2
H-(D,L)-Hat-D-Orn-Aic-Glu-NH2
j
j
j
j
j
j
j
j
j
j
j
j
j
a
j
K Ai (nM)
K yi (nM)
K yi /K Ai
0.981
1660
4.21
8.26
26.3
2.91
54.2
7.68
3.21
14,000
209
1,570
3,510
10.8
74.7
119
3.27
8.43
49.6
190
133
3.71
1.38
15.5
Displacement of [3H]DAMGO (A-selective) and [3H]DSLET (y-selective) from rat brain
membrane binding sites.
j
j
B. Conformational Study of H-Hat-D-Orn-Aic-Glu -NH2
The results of a molecular mechanics study indicated that the lowest
energy conformation of H-Hat-D-Orn-Aic-Glu-NH2 is still characterized
by a tilted stacking interaction of the aromatic rings of the residues in
positions 1 and 3 of the peptide sequence (B. C. Wilkes and P. W.
Schiller, unpublished results) (Fig. 2). It had been suggested that this
tilted stacking arrangement of the two aromatic rings might represent a
structural requirement for high A-receptor affinity of the tetrapeptide HTyr-D-Orn-Phe-Asp-NH2 and structurally related cyclic dermorphin
analogues [31]. An alternative model of the A-receptor-bound conformation based on conformational analysis of morphiceptin analogues is
characterized by a larger distance (f10 A˚) between the Tyr1 and Phe3
aromatic rings [32]. The A-selective morphiceptin analogue H-Tyr-ProPhe(NMe)-D-Pro-NH2 (PL017) in this proposed bioactive conformation
is depicted in Figure 2. A conformer of H-Hat-D-Orn-Aic-Glu-NH2 with
an energy 1.1 kcal/mol higher than that of the lowest energy structure
showed good spatial overlap of its Tyr1 and Phe3 aromatic rings and
N-terminal amino group with the corresponding moieties in morphiceptin
in this proposed bioactive conformation, the root-mean-square deviation
being 1.1 A˚ (Fig. 2). Thus, reasonable low energy conformers consistent
with either one of the two proposed bioactive conformations can be
assumed by H-Hat-D-Orn-Aic-Glu-NH2, and both models remain plausible candidate structures for the A-receptor pharmacophore.
j
j
j
j
j
j
j
j
153
j
j
Figure 2 Spatial overlap of low energy conformers of H-Hat-D-Orn-Aic-GluNH2 (heavy lines) with proposed models of the A-receptor-bound conformation
(light lines) based on conformational analysis of H-Tyr-D-Orn-Phe-Asp-NH2 [24]
(left panel) and H-Tyr-Pro-Phe(NMe)-D-Pro-NH2 (PL017) [29] (right panel).
j
j
III. D OPIOID ANTAGONISTS
The first y antagonists derived from opioid peptides were obtained through
diallylation of the N-terminal amino group of enkephalin analogues. The
best known compound of this type is N,N-diallyl-Tyr-Aib-Aib-Phe-LeuOH (ICI 174,864; Aib = a-aminoisobutyric acid), which is quite y selective
(K Ai /K yi = 128) but not very potent (K yi = 199 nM; Ke = 69 nM in the
MVD assay) [33,34]. The nonpeptide y antagonist naltrindole (NTI) [35] is
highly potent but displays only modest y selectivity (K Ai /K yi = 21.2). A
benzofuran analogue of naltrindole, NTB, showed improved y selectivity
but somewhat lower y-antagonist potency [36]. However, both NTI and
NTB also turned out to be antagonists against A and n agonists in the GPI
assay, with potencies (Ke = 29–48 nM) about 100 to 300 times lower than
those observed against y agonists in the MVD assay (Ke = 0.13 and 0.27
nM, respectively) [36]. The recently discovered TIP(P) peptides represent a
novel class of potent and highly selective y-opioid antagonists [37]. The two
prototype antagonists were the tetrapeptide H-Tyr-Tic-Phe-Phe-OH
(TIPP;Tic = 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid) and the
154
Schiller et al.
tripeptide H-Tyr-Tic-Phe-OH (TIP). TIPP showed high antagonist
potency against various y agonists in the MVD assay (Ke = 3–5 nM),
high y-receptor affinity (K yi = 1.22 nM) in the rat brain membrane receptor
binding assay, and extraordinary y selectivity (K Ai /K yi = 1410). Importantly, TIPP displayed no A- or n- antagonist properties in the GPI assay at
concentrations as high as 10 AM. In comparison with TIPP, the tripeptide
TIP was a somewhat less potent and less selective y antagonist.
A. Structure–Activity Studies of TIP(P) Peptides
Methylation of the N-terminal amino group of TIPP produced a compound, Tyr(NMe)-Tic-Phe-Phe-OH, with four fold enhanced y-antagonist
potency and further improved y selectivity [38] (Table 2). Replacement of
Tyr1 in TIPP with 2V,6V-dimethyltyrosine (Dmt) led to the analog H-DmtTic-Phe-Phe-OH (DIPP), with y-antagonist potency in the subnanomolar
range (Ke = 0.196 nM in the MVD assay) and with still excellent y
selectivity. This compound turned out to be four times more potent than
naltrindole as antagonist in the MVD assay (Table 2) and it represents the
most potent y-opioid antagonist reported to date. The results of stability
studies indicated that TIPP is stable in aqueous solution for extended
periods of time but undergoes slow diketopiperazine formation and
concomitant cleavage of the Tic2—Phe3 peptide bond in DMSO or MeOH
[39]. To prevent this spontaneous degradation, a TIPP analogue containing a reduced peptide bond between Tic2 and Phe3 was synthesized. The
resulting pseudopeptide, H-Tyr-TicC[CH2–NH]Phe-Phe-OH (TIPP[C]),
retained y-antagonist potency comparable to that of the parent peptide and
showed extraordinary y selectivity in the receptor binding assays (K Ai /K yi =
10500) [40], being about 500 times more y selective than naltrindole and 17
times more y selective than [D-Ala2]deltorphin II (Table 2). Moreover,
TIPP[C] was shown to be highly stable against chemical and enzymatic
degradation. It also showed selectivity ratios exceeding 10,000 against all
A- and n- receptor subtypes (A1, A2, n1, n2, n3) [41] and thus represents an
excellent pharmacological tool. The corresponding pseudotripeptide,
H-Tyr-TicC[CH2-NH]Phe-OH (TIP[C]), also retained high y-antagonist
potency, and its y selectivity was 40 times greater than that of its parent,
TIP. Methylation of the secondary amino group of the reduced peptide
bond in TIPP[C] produced the compound H-Tyr-TicC[CH2-NCH3]PhePhe-OH, which retained the high y antagonist potency of the parent
pseudopeptide and showed even higher y selectivity (K Ai /K yi = 15,900,
Table 2). This compound is nearly 300 times more y selective than DPDPE
155
Table 2 Antagonist Potencies and Opioid Receptor Affinities of TIPP Analogues
Compound
H-Tyr-Tic-Phe-Phe-OH (TIPP)
H-Tyr-Tic-Phe-OH (TIP)
Tyr(NMe)-Tic-Phe-Phe-OH
H-Dmt-Tic-Phe-Phe-OH
H-Tyr-TicC[CH2-NH]Phe-Phe-OH
(TIPP[C])
H-Tyr-TicC[CH2-NH]Phe-OH (TIP[C])
H-Tyr-TicC[CH2-NCH3]Phe-Phe-OH
H-Tyr(3V-I)-Tic-Phe-Phe-OH
H-Tyr(3V-I)-Tic-Phe-OH
H-Tyr(3V-I)-TicC[CH2-NH]Phe-Phe-OH
H-Tyr-Tic-Leu-Phe-OH
H-Tyr-Tic-Ile-Phe-OH
H-Tyr-Tic-Cha-Phe-OH (TICP)
H-Tyr-TicC[CH2-NH]Cha-Phe-OH
(TICP[C])
H-Tyr(3V-I)-Tic-Cha-Phe-OH
H-Dmt-Tic-OH
Naltrindole
DPDPE
[D-Ala2]deltorphin II
a
b
Ke
(nM)a
5.86
11.7
1.22
0.196
2.89
K Ai
(nM)b
K yi
(nM)b
K Ai /K yi
1,720
1,280
13,400
141
3,230
1.22
1,410
9.07
141
1.29 10,400
0.248
569
0.308 10,500
9.06
10,800
4.76
13,400
Agonist 5,230
141
12,100
19.2
2,660
7.32
904
12.7
6,460
0.438
3,600
0.219
1,050
1.94
5,570
0.842 15,900
24.8
211
60.0
202
2.08
1,280
2.84
318
4.37
1,480
0.611 5,890
0.259 4,050
12.7
6.55
0.636
Agonist
Agonist
4,010
3.33
1,360
1.84
3.86 0.182
943
16.4
3,930
6.43
1,200
739
21.2
57.5
611
Determined against DPDPE in the MVD assay.
Binding assay based on displacement of [3H]DAMGO (A-selective) and [3H]DSLET
(y-selective) from rat brain membrane binding sites.
and shows even slightly higher y-receptor selectivity than the TIPP[C]
parent peptide.
For the purpose of opioid receptor binding studies, TIPP was also
radioiodinated. Surprisingly, [125I]TIPP binding to y receptors in N4TG1
neuroblastoma cells was substantially reduced in the presence of Na+ and
Gpp(NH)p [42]. These results indicated that substitution of an iodine atom
at the 3V position of Tyr1 in TIPP had turned the y antagonist into a y
agonist. The corresponding ‘‘cold’’ analogue, H-Tyr(3’-I)-Tic-Phe-PheOH, was then synthesized and shown to be a full agonist in the MVD
assay (IC50 = 97 nM). This agonist effect was antagonized by TIPP (Ke =
11 nM) [38]. Corresponding iodination of the Tyr residue in TIP and
TIPP[C] did not result in agonism, but somewhat reduced antagonist
156
Schiller et al.
potency was observed (Table 2). It therefore appears that the astonishing
conversion observed with the tetrapeptide TIPP may be due to an overall
conformational effect rather than to a direct, local effect of the iodine
substituent. Interestingly, substitution of a bromine or chlorine atom at the
3V position of Tyr1 in TIPP produced partial agonists with respective
intrinsic efficacies of 0.16 and 0.12, whereas the Tyr(3V-F)-analogue was
again a pure antagonist (Ke = 13.0 nM) [38]. Thus, systematic substitution
of halogen atoms beginning with iodine and in the order of the periodic
table produced a progressive decrease in intrinsic activity and a concomitant increase in affinity at the y receptor (K yi = 24.2, 3.62, 3.00 and 1.62
nM, respectively).
Replacement of the Phe3 residue in TIPP with the aliphatic amino
acid residues Leu or Ile resulted in analogues that retained high yantagonist potency and considerable y selectivity (Table 2). This result is
in agreement with the weak y-antagonist activity that had been reported for
the tripeptide H-Tyr-Tic-Ala-OH [43]. Obviously, an aromatic residue at
the 3 position of the peptide sequence of TIP(P) peptides is not absolutely
required for y antagonist activity. Most interestingly, saturation of the
Phe3 aromatic ring in TIPP, as achieved through substitution of cyclohexylalanine (Cha), led to H-Tyr-Tic-Cha-Phe-OH [TICP], a compound
showing substantially increased y-antagonist potency and higher y selectivity than the parent peptide [44]. The corresponding pseudopeptide,
H-Tyr-TicC[CH 2 -NH]Cha-Phe-OH (TICP[C]), showed a further
improvement in y-antagonist activity. Its y-antagonist potency is comparable to that of the analogue H-Dmt-Tic-Phe-Phe-OH but, in comparison
with the latter peptide, it is seven times more y selective (K Ai /K yi = 4050)
[44]. Both TIPP[C] and TICP[C] were prepared in tritiated form [45,46]
and should turn out to be valuable new radioligands for y receptor labeling
studies in vitro and in vivo. The analogue H-Tyr(3V-I)-Tic-Cha-Phe-OH
was an antagonist in the MVD assay with a potency about 30 times lower
than that of TICP. Thus, unlike in the case of TIPP, introduction of an
iodine substituent at the 3V position of Tyr1 in TICP did not produce a y
agonist. This result demonstrates once again how a relatively subtle
structural modification, such as the saturation of an aromatic ring, can
have a determinant effect on agonist versus antagonist behavior.
In 1995 the dipeptide H-Dmt-Tic-OH was reported to be a y-opioid
antagonist with unprecedented y-receptor affinity (K yi = 0.022 nM) and y
receptor selectivity (K Ai /K yi = 150,000) [47]. However, in a direct comparison under identical assay conditions, this compound showed about 30
times lower y-antagonist potency and 6 times lower y-receptor selectivity
157
than TICP[C] [48] (Table 2). Similar results were obtained in a more recent
study [49], which confirmed that H-Dmt-Tic-OH had much lower y
receptor affinity [IC50(y) = 1.6 nM] and much lower y selectivity [IC50(A)/
IC50(y) = 558] than had originally been reported. Furthermore, H-DmtTic-OH was found to be unstable in organic solvents owing to diketopiperazine formation (P. W. Schiller and T. M.-D. Nguyen, unpublished
results).
B. Conformational Studies of TIP and TIPP
A molecular mechanics study (grid search and energy minimization) of the
tripeptide y-antagonist TIP resulted in several low energy conformers
having energies within about 2 kcal/mol of that of the lowest energy
structure [50]. The centrally located Tic residue imposes a number of
conformational constraints on the N-terminal dipeptide segment; however, the results of molecular dynamics simulations indicate that this
tripeptide still shows some structural flexibility at the Phe3 residue.
Attempts to demonstrate spatial overlap between the pharmacophoric
moieties of low-energy conformers of TIP and the structurally rigid nonpeptide y antagonist naltrindole were made by superimposing either the
Tyr1 and Phe3 aromatic rings and the N-terminal amino group or the Tyr1
and Tic2 aromatic rings and the N-terminal amino group of the peptide
with the corresponding aromatic rings and nitrogen atom in the alkaloid
structure. In each case the investigators found a conformer of TIP with an
energy very close to that of the lowest energy structure (2.1 kcal/mol
higher). However, the low-energy conformer showing spatial overlap of its
Tic2 aromatic ring with the six-membered aromatic ring of the indole
moiety in naltrindole (Fig. 3) appears to be a more plausible candidate
structure of the y-receptor-bound conformation for two reasons:
1.
2.
The Tic2 aromatic ring has been shown to be of crucial
importance for y antagonist activity [51].
The y-antagonist properties are maintained upon replacement of
the Phe3 residue in the peptide with an aliphatic amino acid
residue (see earlier).
This model of the receptor-bound conformation of TIP is characterized by
a clustered configuration of the three aromatic moieties with the Phe3
aromatic ring sandwiched between the Tyr1 and Tic2 aromatic rings.
A molecular mechanics study of TIPP and TIPP[C] produced about
70 structures within 3 kcal/mol of the lowest energy conformation in each
158
Schiller et al.
Figure 3 Superimposition of a low energy conformer of TIP (heavy lines) with
the minimized structure of naltrindole (light lines). The Tyr1 and Tic2 aromatic
rings and the N-terminal amino group of the peptide are superimposed with the
corresponding moieties in the alkaloid structure. The superimposed molecules
are shown in two different orientations.
case [52]. The lowest energy conformers of both TIPP and TIPP[C]
showed good overlap of their Tyr1 and Tic2 aromatic rings and N-terminal
amino group with the corresponding pharmacophoric moieties of naltrindole. Thus, these results are in agreement with the model of the
receptor-bound conformation of TIP proposed earlier. This model is
characterized by all-trans peptide bonds and was definitely confirmed by
conformational analyses of two TIPP analogues (y antagonists) in which a
cis peptide bond between the Tyr1 and Tic2 residues is sterically forbidden
[53]. Both TIPP and TIPP[C] are very hydrophobic peptides, and the
results of the theoretical conformational analyses clearly indicated that
they enjoy considerable structural flexibility, particularly in their Cterminal dipeptide segment. There is no doubt that their conformations
are quite dependent on the environment. According to our theoretical
analysis, a crystal structure of TIPP published in 1994 [54] is about 3 kcal/
mol higher in energy than the lowest energy structure and shows no
similarity to any of the calculated low energy structures [52,53]. The
crystal structure of TIPP appears to be stabilized by a large number of
intermolecular hydrophobic contacts between layers of TIPP molecules in
the crystal and by several hydrogen bonds to solvent (AcOH) molecules.
There is no reason to believe that it resembles the y-receptor-bound
conformation of TIPP. In an aqueous environment TIP(P) peptides
may undergo a so-called hydrophobic collapse [55]. It is possible that
subtle structural modifications, such as introduction of an iodine sub-
159
stituent at the 3V position of Tyr1, saturation of the Phe3 aromatic ring, or
reduction of the Tic2—Phe3 peptide bond, may produce different patterns
of aromatic ring clustering that could result in either y-agonist or yantagonist activity, as described earlier.
C. Effect of D-Opioid Antagonists on the Development
of Morphine Tolerance and Dependence
Blockade of y receptors with the nonpeptide y antagonist naltrindole
concurrently with chronic morphine treatment has been reported to
attenuate the development of tolerance and the severity of the precipitated
withdrawal syndrome in mice [56]. In an effort to corroborate these
results, the effects of TIPP, TIPP[C], and naltrindole on the development
of morphine tolerance and dependence were examined. Each of the
antagonists was continuously infused into the lateral ventricle of rats
treated chronically with subcutaneous morphine [57]. After a 6-day period
of drug administration, rats treated with TIPP[C] showed no morphine
tolerance and a greatly reduced incidence of withdrawal symptoms
following injection of naloxone. Naltrindole and TIPP also significantly
decreased the amount of time spent in withdrawal but did not attenuate
the development of morphine tolerance. More recently, morphine was
shown to retain its A-receptor-mediated analgesic activity in y-opioid
receptor knockout mice without producing analgesic tolerance upon
chronic administration [58]. These interesting findings clearly demonstrate
that y-opioid receptors play a major role in the development of morphine
tolerance and dependence and suggest the possibility of the combined use
of a A type opioid analgesic and a y-opioid antagonist in the treatment of
chronic pain. Even more interestingly, these results indicate that mixed A
agonist/y antagonists can be expected to be analgesics with low propensity
to produce tolerance and dependence and, therefore, might be of benefit in
the management of chronic pain.
IV. MIXED M AGONIST/D ANTAGONISTS
A. Prototypes and Structure–Activity Relationships
The first known example of a mixed A agonist/y antagonist was a TIPP
analogue in which the free C-terminal carboxylate function had been replaced by a carboxamide function [37]. This compound, H-Tyr-Tic-Phe-
160
Schiller et al.
Phe-NH2 (TIPP-NH2), was found to be a moderately potent A agonist in
the GPI assay and a potent y antagonist in the MVD assay (Table 3).
Replacement of Tyr1 in TIPP-NH2 with Dmt produced the compound
H-Dmt-Tic-Phe-Phe-NH2 (DIPP-NH2), showing an increase in both Aagonist potency and y-antagonist activity by nearly two orders of magnitude [59,60]. In the receptor binding assays DIPP-NH2 displayed very high
y-receptor affinity and still some preference for y receptors over A receptors.
In comparison with DIPP-NH2, the corresponding analogue with a reduced peptide bond between the Tic2 and Phe3 residues, H-Dmt-TicC
[CH2-NH]Phe-Phe-NH2 (DIPP-NH2[C]), was about twice as potent as
agonist in the GPI assay and about half as potent as antagonist in the
MVD assay. Showing A- and y-receptor affinities that were both in the
subnanomolar range, DIPP-NH2[C] was essentially nonselective (K Ai /K yi
= 2.11; Table 3). Therefore, DIPP-NH2[C] represents the first known
opioid compound with balanced A-agonist/y-antagonist properties
[59,60]. In the rat tail flick test, DIPP-NH2[C] given intracerebroventricularly (ICV) produced a potent analgesic effect, being about three
times more potent than morphine. It produced less acute tolerance than
morphine, but still a certain level of chronic tolerance. Unlike morphine,
DIPP-NH2[C] produced no physical dependence upon chronic administration at high doses. Thus, DIPP-NH2[C] fulfilled to a large extent
the expectations based on the mixed A-agonist/y-antagonist concept [60].
Surprisingly, elimination of the C-terminal carboxylate function
of the tripeptide y antagonist TIP resulted in H-Tyr-Tic-NH-(CH2)2-Ph
(Ph = phenyl), a compound that was a moderately potent full y agonist
(Table 3) [61]. Interestingly, lengthening of the phenylethyl substituent by
insertion of an additional methylene group restored y antagonism, as
indicated by the finding that the dipeptide derivative H-Tyr-Tic-NH(CH2)3-Ph was a moderately potent y antagonist in the MVD assay and a
relatively weak partial A agonist in the GPI assay. This remarkable
dependence of y-agonist versus y-antagonist behavior on the length of
the phenylalkyl substituent may be due to conformational effects resulting
in different clustering of the three aromatic moieties present in these
molecules. The analogue H-Dmt-Tic-NH-(CH2)3-Ph showed very high
affinity for both A and y receptors and was a potent y antagonist in the
MVD assay (Ke = 1.69 nM). That this compound displayed relatively
modest agonist potency in the GPI assay suggests that it also may have
partial A-agonist properties and that it may represent a mixed partial A
agonist/y antagonist. In the case of high affinity A-receptor ligands, partial
agonism is not always directly apparent in the GPI assay because the
1700
18.2
7.71
3010
(42 %)c
102
2.14
384
7.88
H-Tyr-Tic-Phe-Phe-NH2
H-Dmt-Tic-Phe-Phe-NH2
H-Dmt-TicC[CH2-NH]Phe-Phe-NH2
H-Tyr-Tic-NH-(CH2)2-Ph
H-Dmt-Tic-NH-(CH2)3-Ph
H-Tyr-c[-D-Orn-Phe-D-Pro-Gly-]
H-Tyr-c[-D-Orn-2-Nal-D-Pro-Gly-]
H-Dmt-c[-D-Orn-2-Nal-D-Pro-Gly-]
4.89
82.0
IC50, (nM)
MVD
233
2.13
41.9
1.69
18.0
0.209
0.537
Ke (nM)b
78.8
1.19
0.943
69.1
160
0.386
0.881
5.89
0.460
K Ai (nM)
b
a
3.00
0.118
0.447
5.22
3.01
0.0871
13.2
17.2
0.457
K yi (nM)
Binding assaysa
Displacement of [3H]DAMGO (A-selective) and [3H]DSLET (y-selective) from rat brain membrane binding sites.
Determined against DPDPE.
c
Maximal inhibition of the contractions at 10 AM.
IC50, nM
Compound
GPI
Table 3 In Vitro Opioid Activities and Receptor Affinities of Mixed A Agonist/y Antagonists
26.3
10.1
2.11
13.2
53.2
4.43
0.0667
0.342
1.01
K Ai /K yi
161
162
Schiller et al.
ileum has a very high A-receptor reserve. Substituted Tyr-Tic-dipeptide
amides with mixed A-agonist/y-antagonist properties are of interest because
their small molecular size and lipophilic character may facilitate their
passage across the blood–brain barrier (BBB). Further efforts aimed at
strengthening the A-agonist component of this class of compounds may
be required.
Another prototype of a mixed A agonist/y antagonist is the cyclic hcasomorphin analogue H-Tyr-c[-D-Orn-2-Nal-D-Pro-Gly-] [62]. This
compound turned out to be a fairly potent A agonist in the GPI assay
and showed relatively modest y-antagonist potency in the MVD assay
(Table 3). The 2-naphthylalanine (2-Nal) residue in this compound is a
key structural determinant for its y-antagonist behavior, since the corresponding Phe3 analogue, H-Tyr-c[-D-Orn-Phe-D-Pro-Gly-] was found to
be a full y agonist in the MVD assay [62]. As expected, an analogue
containing Dmt in place of Tyr1, H-Dmt-c-[-D-Orn-2-Nal-D-Pro-Gly-],
showed greatly increased A-agonist and y-antagonist potency [63]. This
pentapeptide displayed almost equal affinities for A and y receptors in the
subnanomolar range and, thus, represents another example of a balanced
A agonist/y antagonist. In comparison with DIPP-NH2[C], H-Dmt-c[-DOrn-2-Nal-D-Pro-Gly-] has the same A-agonist potency in the GPI assay
and is about four times less potent as a y antagonist in the MVD assay.
The various compounds described in this section represent the only
known mixed A-agonist/y-antagonist substances reported to date. Analgesic testing of all these prototypes will reveal which type of compound
has the greatest potential for the development of viable analgesics.
Further analogues may have to be prepared and examined to determine
the ratio between A-agonist and y-antagonist potency required for
optimal attenuation of tolerance and dependence development. Additional structural modifications may be necessary to increase analgesic
potency and bioavailability.
B. Conformational Study of H-Tyr-c[-D-Orn-2-Nal-DPro-Gly-]
The conformation of the mixed A agonist/y antagonist H-Tyr-c[-D-Orn-2Nal-D-Pro-Gly-] in comparison to that of H-Tyr-c[-D-Orn-Phe-D-Pro-Gly-]
was studied in DMSO-d6 by NMR spectroscopy and by molecular mechanics calculations [62,64]. Neither peptide showed nuclear Overhauser
effects between CaH protons or chemical exchange cross peaks in spectra
obtained by total correlation and rotating frame Overhauser enhance-
163
ment spectroscopy (TOCSY, ROESY). These results indicated that the
average preferred solution conformation of both peptides was characterized by all-trans peptide bonds. The results of temperature-dependence
studies of the amide proton chemical shifts in conjunction with those of
the molecular mechanics studies indicated that the two analogues had
backbone conformations that were both stabilized by Tyr1-COHNPhe3 (or 2-Nal3) and D-Orn2-COHNy-D-Orn2 hydrogen bonds. Furthermore, ROESY experiments revealed a close proximity between the
aromatic moiety of the 3-position residue and the pyrrolidine ring of the
4
D-Pro residue in these two compounds. The comparison of all calculated
low-energy conformations with the various proton NMR parameters led
to proposals for the solution conformation of these two peptides (Fig. 4).
Inspection of the structures reveals that the Phe3- and 2-Nal3analogues have similar backbone conformations and the same side chain
orientation at the 3 position. These results suggest that the y-antagonist
Figure 4 Proposed solution conformations of H-Tyr-c-[-D-Orn-2-Nal-D-ProGly-] (left panel) and H-Tyr-c-[D-Orn-Phe-D-Pro-Gly-] (right panel).
164
Schiller et al.
properties of the 2-Nal3 analogue may not be due to a difference in its
overall conformation in comparison to the Phe3 analogue but rather may
be the result of a direct interference of the 2-naphthyl moiety per se at the
receptor binding site, preventing proper alignment of the peptide such as
required for signal transduction.
V. D AGONISTS
y-Opioid agonists are known to produce analgesic effects and look
promising because they induce less tolerance and physical dependence
than morphine, no respiratory depression, and few or no adverse
gastrointestinal effects [65,66]. Selective peptide y agonists currently
available include the enkephalin analogues H-Tyr-D-Thr-Gly-Phe-LeuThr-OH (DTLET), H-Tyr-c[D-Pen-Gly-Phe-D-Pen]OH (DPDPE), and
H-Tyr-c[D-Cys-Phe-D-Pen-OH]OH (JOM-13), as well a the deltorphins
H-Tyr-D-Met-Phe-His-Leu-Met-Asp-NH2 (dermenkephalin), H-Tyr-DAla-Phe-Asp-Val-Val-Gly-NH2 (deltorphin I), and H-Tyr-D-Ala-PheGlu-Val-Val-Gly-NH2 (deltorphin II) (for reviews, see Refs. 20 and 21).
However, these peptides are of relatively large molecular size and for this
reason their ability to cross the BBB is very limited. Nonpeptide y agonists
that were developed in the early to mid-1990s include the racemic compound BW373U86 [67] and its chemically modified enantiomer SNC80
[68], as well as the compound TAN-67 [69]. However, BW373U86 produced significant toxicity, manifested behaviorally as convulsions and
barrel rolling, in mice [70], and TAN-67 showed no significant antinociceptive activity in the mouse tail flick test [69]. Evidently, there is still a need
for the development of new potent y opioid agonists of low molecular
weight and high lipophilicity.
In an effort to increase the moderate y-agonist potency and the yreceptor selectivity of the dipeptide H-Tyr-Tic-NH-(CH2)2-Ph [61], structural modifications of the C-terminal phenylethyl group were performed
by introduction of an additional substituent either in ortho position of the
phenyl ring or at the h carbon [44] (Table 4). The analogue H-Tyr-Tic-NH(CH2)2-Ph(o-Cl) was a 10-fold more potent y agonist than the parent
peptide in the MVD assay and was five times more y-receptor selective.
Introduction of a second phenyl group at the h carbon of the phenylethylamine moiety led to the compound H-Tyr-Tic-NH-CH2-CH(Ph)2, with 20fold increased y-agonist potency and 2-fold improved y selectivity. The
corresponding N-methylated analogue, Tyr(NMe)-Tic-NH-CH2-CH(Ph)2
165
Table 4 In Vitro Opioid Activities of Dipeptide y-Opioid Agonists
Compound
IC50
(nM)a
K yi
(nM)b
K Ai
(nM)b
K Ai /K yi
H-Tyr-Tic-NH-(CH2)2-Ph(o-Cl)
H-Tyr-Tic-NH-CH2-CH(Ph)2
Tyr(NMe)-Tic-NH-CH2-CH(Ph)2
H-Tyr-Tic-NH-CH2-CH(Ph)COOEt (I)
H-Tyr-Tic-NH-CH2-CH(Ph)COOEt (II)
H-Hmt-Tic-NH-CH2-CH(Ph)2
DPDPE
82.0
8.77
3.77
0.261
1.28
8.64
0.630
5.30
5.22
1.43
0.981
0.581
0.569
3.03
2.00
16.4
69.1
96.9
28.8
12.7
886
153
1670
943
13.2
67.8
29.4
21.9
1560
50.5
835
57.5
a
b
Determined in the MVD assay.
Binding assay based on displacement of [3H]DSLET (y-selective) and [3H]DAMGO
(A-selective) from rat brain membrane binding sites.
displayed subnanomolar y-agonist potency and still marked y-receptor
selectivity. One of the isomers (I) of H-Tyr-Tic-NH-CH2-CH(Ph)COOEt
was also found to be a potent y agonist with very high preference for y
receptors over A receptors. An analogue containing 2V-hydroxy,6V-methyltyrosine (Hmt) in place of Tyr1, H-Hmt-Tic-NH-CH2-CH(Ph)2, turned
out to be particularly remarkable because it showed both subnanomolar y
agonist potency (IC50 = 0.630 nM) and very high y-receptor selectivity
(K Ai /K yi = 835). In a direct comparison under identical assay conditions,
this compound was 8 times more potent than the well-known y agonist
DPDPE and 15 times more y selective. None of these compounds had
significant binding affinity for n-opioid receptors. From these results it can
be concluded that the dipeptide derivatives described here represent a new
class of potent and selective y-opioid agonists. It is expected that these
compounds should be able to cross the BBB to some extent because of their
low molecular weight and high lipophilicity. Therefore, they have potential
as centrally acting analgesics that may produce fewer side effects than the
currently used A type opiates.
VI. CONCLUSIONS
Application of the concept of conformational restriction to opioid peptides
has produced fruitful results, insofar as peptide analogues and mimetics
with interesting opioid activity profiles and high stability against enzymatic
166
Schiller et al.
degradation were obtained. The conformationally restricted analogues
that were developed were amenable to meaningful conformational analysis
permitting the elaboration of models of the bioactive conformation at the A
or y receptor.
The multiple conformational restriction of dermorphin-related tetrapeptide analogues that was performed represents a rational design of
opioid peptidomimetics characterized by a high degree of structural rigidification. This is indicated by the fact that the A-selective agonist H-Hat-DOrn-Aic-Glu-NH2 contains only two freely rotatable bonds, whereas there
are 14 freely rotatable bonds in [Leu5]enkephalin.
The discovery of the TIP(P) peptides and their further structural
modification led to y opioid antagonists with unprecedented potency and
selectivity. The observation that very subtle structural modifications of
these flexible and hydrophobic peptides in some cases converted a y
antagonist into a y agonist and vice versa is most intriguing and unique
in the peptide field. This behavior may be explained with changes in the
patterns of aromatic ring clustering in these hydrophobically collapsed
molecules as a consequence of the minor structural alterations (introduction of a halogen substituent, peptide bond reduction, saturation of an
aromatic ring, etc.) that were performed. The TIP(P) peptides are of
therapeutic interest because y antagonists have been shown to attenuate
the development of morphine tolerance and dependence [56,57] and to
have an immunosuppressive effect [71].
The three prototype mixed A agonist/y antagonists described in this
chapter have excellent potential as analgesics with low propensity to
produce tolerance and dependence. The pseudotetrapeptide DIPPNH2[C] has already been shown to produce a potent analgesic effect, less
tolerance than morphine, and no physical dependence upon chronic
administration. In preliminary experiments, the tetrapeptides DIPP-NH2
and DIPP-NH2[C] were shown to cross the BBB to some extent, but
further structural modifications need to be performed in order to improve
the BBB penetration of these compounds. The Tyr-Tic dipeptide derivatives can also be expected to penetrate into the central nervous system
because they are relatively small, lipophilic molecules. In this context, it is
of interest to point out that the structurally related dipeptide H-Dmt-DAla-NH-(CH2)3-Ph (SC-39566), a plain A-opioid agonist, produced antinociception in the rat by subcutaneous and oral administration [72]. As
indicated by the results of the NMR and molecular mechanics studies, the
conformation of the cyclic h-casomorphin analogue H-Tyr-c[-D-Orn-2Nal-D-Pro-Gly-] is stabilized by intramolecular hydrogen bonds. Therej
j
167
fore, this mixed A agonist/y antagonist has a reduced capacity for intermolecular hydrogen bonding with water molecules and, consequently,
should have a reasonable chance to cross the BBB as well.
The dipeptide y agonists may turn out to be interesting pharmacological tools, since some of them are more potent and more selective than
the y agonists currently in use. Furthermore, these compounds represent a
new class of y agonists and have potential for pain treatment because they
may also be small enough and lipophilic enough to cross the BBB and to
produce a centrally mediated analgesic effect.
ACKNOWLEDGMENTS
The work described in this chapter was supported by operating grants
from the Medical Research Council of Canada (MT-5655) and the National Institute on Drug Abuse (DA-04443).
REFERENCES
1. Schiller PW, DiMaio J. Opiate receptor subclasses differ in their conformational requirements. Nature (London) 1982; 297:74–76.
2. Martin WR, Eades CJ, Thompson GA, Huppler RA, Gilbert PE. The effects
of morphine- and nalorphine-like drugs in the nondependent and morphinedependent chronic spinal dog. J Pharmacol Exp Ther 1976; 197:517–523.
3. Lord JAH, Waterfield AA, Hughes, J, Kosterlitz HW. Endogenous opioid peptides: multiple agonists and receptors. Nature (London) 1977; 267:
495–499.
4. Clark JA, Liu L, Price M, Hersh B, Edelson M, Pasternak GW. Kappa
opiate receptor multiplicity: evidence for two U50,488-sensitive n1 subtypes
and a novel n3 subtype. J Pharmacol Exp Ther 1989; 251:461–468.
5. Jiang Q, Takemori AE, Sultana M, Portoghese PS, Bowen BD, Mosberg
HI, Porreca F. Differential antagonism of opioid delta antinociception by
[D-Ala2,Leu5,Cys6]enkephalin and naltrindole 5V-isothiocyanate: evidence
for delta receptor subtypes. J Pharmacol Exp Ther 1991; 257:1069–1075.
6. Evans CJ, Keith Jr., DE, Morrison H, Magendzo K, Edwards RH. Cloning of a delta opioid receptor by functional expression. Science 1992; 258:
1952–1955.
7. Kieffer BL, Befort K, Gaveriaux-Ruff C, Hirth CG. The y-opioid receptor:
isolation of a cDNA by expression cloning and pharmacological characterization. Proc Natl Acad Sci USA 1992; 89:12048–12052.
168
Schiller et al.
8. Yasuda K, Raynor K, Kong H, Breder CD, Takeda J, Reisine T, Bell GI.
Cloning and functional comparison of n and y opioid receptors from mouse
brain. Proc Natl Acad Sic USA 1993; 90:6736–6740.
9. Chen Y, Mestek A, Liu J, Hurley JA, Yu L. Molecular cloning and
functional expression of a A-opioid receptor from rat brain. Mol Pharmacol
1993; 44:8–12.
10. Pogozheva ID, Lomize AL, Mosberg HI. Opioid receptor three-dimensional
structures from distance geometry calculations with hydrogen bonding
constraints. Biophys J 1998; 75:612–634.
11. Hughes J, Smith TW, Kosterlitz HW, Fothergill LA, Morgan BA, Morris
RH. Identification of two related pentapeptides from the brain with potent
opiate agonist activity. Nature (London) 1975; 258:577–579.
12. Ho¨llt V. Opioid peptide processing and receptor selectivity, Annu Rev
Pharmacol Toxicol 1986; 26:59–77.
13. Zadina JE, Hackler L, Ge LJ, Kastin AJ. A potent and selective endogenous
agonist for the mu opiate receptor. Nature (London) 1997; 386: 499–502.
14. Henschen A, Lottspeich F, Brantl F, Teschemacher H. Novel opioid peptides derived from casein (h-casomorphins). Hoppe-Seyler’s Z. Physiol
Chem 1979; 360:1217–1224.
15. Montecucchi PC, de Castiglione R, Piani S, Gozzini L, Erspamer V. Amino
acid composition and sequence of dermorphin, a novel opiate-like peptide
from the skin of Phyllomedusa sauvagei. Int J Peptide Protein Res 1981;
17:275–283.
16. Erspamer V, Melchiorri P, Falconieri-Erspamer G, et al. Deltorphins: a
family of naturally occurring peptides with high affinity and selectivity for y
opioid binding sites. Proc Natl Acad Sci USA 1989; 86:5188–5192.
17. Leslie FM. Methods used for the study of opioid receptors. Pharmacol Rev
1987; 39:197–249.
18. Blume A. Interaction of ligands with the opiate receptors of brain membranes: regulation by ions and nucleotides. Proc Natl Acad Sci USA 1978;
75:1713–1717.
19. Zimmerman DM, Leander JD. Selective opioid receptor agonists and
antagonists: research tools and potential therapeutic agents. J Med Chem
1990; 33:895–902.
20. Hruby VJ, Gehring CA. Recent developments in the design of receptor
specific opioid peptides. Med Res Rev 1989; 9:343–401.
21. Schiller PW. Development of receptor specific opioid peptide analogs. In:
Ellis GP, West BG, eds. Progress in Medicinal Chemistry. Vol. 28. Amsterdam: Elsevier, 1991:301–340.
22. Hruby VJ, Agnes RS. Conformation–activity relationships of opioid
peptides with selective activities at opioid receptors. Biopolymers (Peptide
Sci) 1999; 51:391–410.
23. DiMaio J, Schiller PW. A cyclic enkephalin analog with high in vitro opiate
activity. Proc Natl Acad Sci USA 1980; 77:7162–7166.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
169
DiMaio J, Nguyen TM-D, Lemieux C, Schiller PW. Synthesis and
pharmacological characterization in vitro of cyclic enkephalin analogs:
effect of conformational constraints on opiate receptor selectivity. J Med
Chem 1982; 25:1432–1438.
Berman JM, Goodman M, Nguyen TM-D, Schiller PW. Cyclic and acyclic
partial retro-inverso enkephalins: mu receptor selective enzyme resistant
analogs. Biochem Biophys Res Commun 1983; 115:864–870.
Schiller PW, Nguyen TM-D, Maziak LA, Lemieux C. A novel cyclic opioid
peptide analog showing high preference for A-receptors. Biochem Biophys
Res Commun 1985; 127:558–564.
Wilkes BC, Schiller PW. Theoretical conformational analysis of a A-selective
cyclic opioid peptide analog. Biopolymers 1987; 26:1431–1444.
Schiller PW, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN,
Marsden BJ, Wilkes BC. Conformational restriction of the phenylalanine
residue in a cyclic opioid peptide analogue: effects on receptor selectivity and
stereospecificity. J Med Chem 1991; 34:3125–3132.
Wilkes BC, Schiller PW. Molecular dynamics simulations of opioid peptide
analogs containing multiple conformational restrictions. Int J Peptide Protein Res 1992; 40:249–254.
Schiller PW, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN.
Conformational restriction of the Phe3 residue in a cyclic dermorphin
analog: effects on receptor selectivity and stereospecificity. In: Smith JA,
Rivier JE, eds. Peptides: Chemistry and Biology (Proceedings of the 12th
American Peptide Symposium). Leiden: The Netherlands: Escom Science
Publishers, 1992:97–99.
Wilkes BC, Schiller PW. Conformation–activity relationships of cyclic
dermorphin analogues. Biopolymers 1990; 29:89–95.
Yamazaki T, Ro S, Goodman M, Chung NN, Schiller PW. A topochemical approach to explain morphiceptin bioactivity. J Med Chem 1993; 36:
708–719.
Cotton R, Giles MG, Miller L, Shaw JS, Timms, D. ICI 174864: a highly
selective antagonist for the opioid y-receptor. Eur J Pharmacol 1984; 97:
331–332.
Corbett AD, Gillan MGC, Kosterlitz HW, McKnight AT, Paterson SJ,
Robson LE. Selectivities of opioid peptide analogues as agonists and
antagonists at the y-receptor. Br J Pharmacol 1984; 83:271–279.
Portoghese PS, Sultana M, Nagase H, Takemori AE. Application of the
message–address concept in the design of highly potent and selective nonpeptide y opioid receptor antagonists. J Med Chem 1988; 31:281–282.
Portoghese PS, Nagase H, Maloney Huss KE, Lin CE, Takemori AE. Role
of spacer and address components in peptidomimetic y opioid receptor
antagonists related to naltrindole. J Med Chem 1991; 34:1715–1720.
Schiller PW, Nguyen TM-D, Weltrowska G, Wilkes BC, Marsden BJ,
Lemieux C, Chung NN. Differential stereochemical requirements of A vs y
170
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
Schiller et al.
opioid receptors for ligand binding and signal transduction: development of
a class of potent and highly y-selective peptide antagonists. Proc Natl Acad
Sci USA 1992; 89:11871–11875.
Schiller PW, Nguyen TM-D, Weltrowska G, Wilkes BC, Marsden BJ,
Schmidt R, Lemieux C, Chung NN. TIPP opioid peptides: development of
extraordinarily potent and selective y antagonists and observation of
astonishing structure–intrinsic activity relationships. In: Hodges RS, Smith
RS, eds. Peptides: Chemistry, Structure and Biology (Proceedings of the
13th American Peptide Symposium). Leiden, The Netherlands: Escom
Science Publishers, 1994:483–486.
Marsden BJ, Nguyen TM-D, Schiller PW. Spontaneous degradation via
diketopiperazine formation of peptides containing a tetrahydroisoquinoline3-carboxylic acid residue in the 2-position of the peptide sequence. Int J
Peptide Protein Res 1993; 41:313–316.
Schiller PW, Weltrowska G, Nguyen TM-D, Wilkes BC, Chung NN,
Lemieux C. TIPP[C]: a highly potent and stable pseudopeptide y opioid
receptor antagonist with extraordinary y selectivity. J Med Chem 1993;
36:3182–3187.
Visconti, L.M., Standifer, K.M., Schiller, P.W., Pasternak, G.W. TIPP[C]: a
highly selective y ligand. Neurosci Lett 1994; 181:47–49.
Lee PHK, Nguyen TM-D, Chung NN, Schiller PW, Chang KJ. Tyrosineiodination converts the delta opioid peptide antagonist TIPP to an agonist.
Eur J Pharmacol 1995; 280:211–214.
Temussi PA, Salvadori S, Amodeo P, Bianchi C, Guerrini R, Tomatis R,
Lazarus LH, Picone D, Tancredi T. Selective opioid dipeptides. Biochem
Biophys Res Commun 1994; 198:933–939.
Schiller PW, Weltrowska G, Schmidt R, Berezowska I, Nguyen TM-D,
Lemieux C, Chung NN, Carpenter KA, Wilkes BC. Subtleties of structure–
agonist versus antagonist relationships of opioid peptides and peptidomimetics. J Receptor Signal Transduction Res 1999; 19:573–588.
Nevin ST, To´th G, Weltrowska G, Schiller PW, Borsodi A. Synthesis
and binding characteristics of tritiated TIPP[C], a highly specific and
stable delta opioid antagonist. Life Sci (Pharmacol Lett) 1995; 56:PL225–
230.
Szatma´ri I, To´th G, Kerte´sz I, Schiller PW, Borsodi A. Synthesis and
binding characteristics of the tritiated TIPP analogue TICP[C], a highly
specific and stable y opioid antagonist. Peptides 1999; 20:1079–1082.
Salvadori S, Attila M, Balboni G, Bianchi C, Bryant SD, Crescenzi O,
Guerrini R, Picone D, Tancredi T, Temussi PA, Lazarus LH. y Opioidmimetic antagonists: prototypes for designing a new generation of
ultraselective opioid peptides. Mol Med 1995; 1:678–689.
Schiller PW, Schmidt R, Weltrowska G, Berezowska I., Nguyen TM-D,
Dupuis S, Chung NN, Lemieux C, Wilkes BC, Carpenter KA. Conforma-
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
171
tionally constrained opioid peptide analogs with novel activity profiles. Lett
Peptide Sci 1998; 5:209–214.
Page´ D, McClory A, Mischki T, Schmidt R, Butterworth J, St-Onge S,
Labarre M, Payza K, Brown W. Novel Dmt-Tic dipeptide analogues as
selective delta-opioid receptor antagonists. Bioorg Med Chem Lett 2000;
10:167–170.
Wilkes BC, Schiller PW. Theoretical conformational analysis of the opioid y
antagonist H-Tyr-Tic-Phe-OH and the A agonist H-Tyr-D-Tic-Phe-NH2.
Biopolymers 1994; 34:1213–1219.
Schiller PW, Nguyen TM-D, Berezowska I, Weltrowska G, Schmidt R,
Marsden BJ, Wilkes BC, Lemieux C, Chung NN. The TIPP opioid peptide
family: development of a new class of highly potent y-receptor antagonists
with extraordinary y-selectivity. In: Yanaihara N, ed. Peptide Chemistry
1992 (Proceedings of the 2nd Japanese Symposium on Peptide Chemistry.
Leiden, The Netherlands: Escom Science Publishers, 1993:337–340.
Wilkes BC, Schiller, PW. Comparative analysis of various proposed models
of the receptor-bound conformation of TIP(P)-related y opioid antagonists.
Biopolymers (Peptide Sci) 1995; 37:391–400.
Wilkes BC, Nguyen TM-D, Weltrowska G, Carpenter KA, Lemieux C,
Chung NN, Schiller PW. The receptor-bound conformation of H-Tyr-Tic(Phe-Phe)-OH related y-opioid antagonists contains all trans peptide bonds.
J Peptide Res 1998; 51:386–394.
Flippen-Anderson JL, George C, Deschamps JR, Reddy PA, Lewin AH,
Brine GA. X-ray structures of the y opioid antagonist TIPP and a protected
derivative of the y opioid antagonist ICI 174,864. Lett Peptide Sci 1994;
1:107–115.
Wiley RA, Rich DH. Peptidomimetics derived from natural products. Med
Res Rev 1993; 13:327–384.
Abdelhamid EE, Sultana M, Portoghese PS, Takemori AE. Selective blockage
of delta opioid receptors prevents the development of morphine tolerance
and dependence in mice. J Pharmacol Exp Ther 1991; 258: 299–303.
Fundytus ME, Schiller PW, Shapiro M, Weltrowska G, Coderre TJ. The
highly selective y-opioid antagonist H-Tyr-TicC[CH2-NH]Phe-Phe-OH
(TIPP[C]) attenuates morphine tolerance and dependence. Eur J Pharmacol
1995; 286:105–108.
Zhu Y, King MA, Schuller AGP, Nitsche JF, Reidl M, Elde RP, Unterwald
E, Pasternak GW, Pintar JE. Retention of supraspinal delta-like analgesia
and loss of morphine tolerance in y opioid receptor knockout mice. Neuron
1999; 24:243–252.
Schiller PW, Weltrowska G, Schmidt R, Nguyen, TM-D, Berezowska I,
Lemieux C, Chung NN, Carpenter KA, Wilkes BC. Four different types of
opioid peptides with mixed A agonist/y antagonist properties. Analgesia
1995; 1:703–706.
172
Schiller et al.
60. Schiller PW, Fundytus ME, Merovitz L, Weltrowska G, Nguyen TM-D,
Lemieux C, Chung NN, Coderre TJ. The opioid A agonist/y antagonist DIPPNH2[c] produces a potent analgesic effect, no physical dependence and less
tolerance than morphine in rats. J Med Chem 1999; 42:3520–3526.
61. Schiller PW, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN. Novel
opioid peptide analogs with mixed A agonist/y antagonist properties. In:
Maia HLS, ed. Peptides 1994 (Proceedings of the 23rd European Peptide
Symposium) Leiden, The Netherlands: Escom Science Publishers, 1995:
632–633.
62. Schmidt R, Vogel D, Mrestani-Klaus C, Brandt W, Neubert K, Chung NN,
Lemieux C, Schiller PW. Cyclic h-casomorphin analogues with mixed A
agonist/y antagonist properties: synthesis, pharmacological characterization
and conformational aspects. J Med Chem 1994; 37:1136–1144.
63. Schmidt R, Chung NN, Lemieux C, Schiller PW. Development of cyclic
casomorphin analogs with potent y antagonist and balanced mixed A
agonist/y antagonist properties. In: Kaumaya PTP, Hodges RS, eds. Peptides: Chemistry, Structure and Biology (Proceedings of the 14th American
Peptide Symposium). Leiden, The Netherlands: Escom Science Publishers,
1996:645-646.
64. Mrestani-Klaus C, Brandt W, Schmidt R, Neubert K, Schiller PW. Proton
NMR conformational analysis of cyclic h-casomorphin analogues of the
type Tyr-cyclo[-NN-D-Orn-Xaa-Gly-]. Arch Pharm Pharm Med Chem 1996;
329:133–142.
65. Cowan A, Zhu XZ, Mosberg HI, Omnaas JR, Porreca F. Direct dependence
studies in rats with agents selective for different types of opioid receptor.
66. Galligan JJ, Mosberg HI, Hurst R, Hruby VJ, Burks TF. Cerebral delta
opioid receptors mediate analgesia but not the intestinal motility effects of
intracerebroventricularly administered opioids. J Pharmacol Exp Ther 1984;
229:641–648.
67. Chang KJ, Rigdon GC, Howard JL, McNutt RW. A novel, potent and
selective nonpeptidic delta opioid receptor agonist. J Pharmacol Exp Ther
1993; 267:852–857.
68. Calderon SN, Rothman RB, Porreca F, Flippen-Anderson JL, McNutt
RW, Xu H, Smith LE, Bilsky EJ, Davis P, Rice KC. Probes for narcotic
receptor mediated phenomena. 19. Synthesis of (+)-4-[(aR)-a-((2S,5R)-4allyl-2,5-dimethyl-1-piperazinyl)-3-methoxybenzyl]-N,N-diethylbenzamide
(SNC80): a highly selective, nonpeptide delta opioid receptor agonist. J Med
Chem 1994; 37:2125–2128.
69. Kamei J, Saitoh A, Ohsawa M, Suzuki T, Misawa M, Nagase H, Kasuya Y.
Antinociceptive effects of the selective non-peptidic delta-opioid receptor
agonist TAN-67 in diabetic mice. Eur J Pharmacol 1995; 276:131–135.
70. Comer SD, McNutt, RW, Chang, K-J, DeCosta BR, Mosberg HI, Woods
173
JH. Discriminative stimulus effects of BW373U86: a nonpeptide ligand with
selectivity for delta opioid receptors. J Pharmacol Exp Ther 1993; 267:
888–895.
71. Arakawa K, Akami T, Okamoto M, Akioka K, Nakai I, Oka T, Nagase H.
Immunosuppression by delta opioid receptor antagonist. Transplant Proc
1993; 25:738–740.
72. Hammond DL, Stapelfeld A, Drower EJ, Savage MA, Tam L, Mazur RH.
Antinociception produced by oral, subcutaneous or intrathecal administration of SC-39566, an opioid dipeptide arylalkylamide, in the rodent. J
Pharmacol Exp Ther 1994; 268:607–615.
7
Linkers and Resins for Solid-Phase
Synthesis
Pan Li and Elaine K. Kolaczkowski
Vertex Pharmaceuticals, Inc., Cambridge, Massachusetts, U.S.A.
Steven A. Kates
Surface Logix, Inc., Brighton, Massachusetts, U.S.A.
I.
INTRODUCTION
Organic chemistry in the last half of the twentieth century has evolved to a
level of extreme sophistication in which complex macromolecules thought
only to exist in nature were prepared in a laboratory hood. The process
typically involves performing a reaction in an organic solvent followed by
isolating, purifying, and analyzing the compound. This tedious, timeconsuming procedure requires considerable expertise. Bruce Merrifield
was the first to recognize an alternative approach for the preparation of
organic compounds. He applied this method to synthetic peptides and was
awarded the Nobel Prize in 1984 for this discovery [1]. The concept was to
perform the chemistry proven in solution but add a covalent attachment
step that links the target to an insoluble polymeric support. Key advantages to the solid-phase technique include simple filtration, washing without manipulative losses, and ease of automation.
Peptide synthesis was amenable to solid-phase techniques since the
process was repetitive. The C-terminal amino acid is attached to polymeric
surface and the peptide chain is assembled via a two-step process: coupling
of the incoming amino acid that has the alpha-amino group protected
175
176
Pan et al.
and removal of this protecting group. Following chain elongation, the
peptide is liberated from the solid support with concomitant release of the
side-chain-protecting groups. Subsequent to Merrifield’s discovery, other
repetitive processes such as DNA synthesis and protein sequencing were
adapted to solid-phase methods.
Most peptides occur in nature as C-terminal acids or amides and
methods were developed to release a peptide from a solid support to
provide these two chemical functionalities. During the 1980s, research in
peptide synthesis methods focused on releasing peptides from a polymeric
support using various conditions (high acid concentration, low acid
concentration, basic, etc.). During the early 1990s, there was a realization
that solid-phase techniques could be applied to the construction of small
molecules for drug discovery. Since many drugs do not contain carboxylic
acids or amides, there was a need to expand the resulting chemical
functionality at the anchoring position of the molecule following cleavage
from solid support. Thus there has been a recent resurgence in research for
solid-phase methods.
A critical aspect to solid-phase synthesis is the anchoring of the
molecule to the polymeric support. A solid support or resin is required to
possess a functional group that is the starting point for the construction of
the molecule. In addition, resins should possess the following properties:
(1) mechanically robust; (2) stable to temperature variation; (3) good
swelling in a broad range of solvents; (4) acceptable bead size and loadings;
(5) stable with acidic, basic, reducing, and oxidizing conditions; (6)
compatible with radical, carbene, carbanion, and carbenium ion chemistry; (7) biocompatible and swelling in aqueous buffers; (8) little nonspecific
binding to biomolecules; (9) mobile, well-solvated, and reagent-accessible
sites. The two most commonly used supports are polystyrene (PS) functionalized with a chloromethyl 1 (original Merrifield polymer) or amino
group 2 at the terminus (Fig. 1) [2]. For peptide synthesis, the cesium salt of
a protected amino acid is anchored to chloromethyl-PS via a nucleophilic
displacement of chlorine. Following chain elongation, the peptide is
released from the support by treatment with a strong acid such as HF to
provide a peptide acid. Amino acids anchored to aminomethyl-PS form an
Figure 1
Linkers and Resins
177
Figure 2
amide bond and are stable to HF conditions. p-Methylbenzhydrylamine
(MBHA, 3) is an amine functionalized polystyrene that liberates Cterminal amides upon exposure to HF (Fig. 2) [3].
A second strategy is to attach a linker (also referred to as a handle or
anchor) to the resin followed by assembly of the molecule. A linker is
bifunctional spacer that serves to link the initial synthetic unit to the
support in two discrete steps (Fig. 3). To attach a linker to a chloromethylPS resin, a phenol functionality such as handle 4 is used to form an ether
bond (Fig. 4). To attach the same handle to an amino-functionalized
support, acetoxy function 5 or a longer methylene spacer of the corresponding phenol is applied to form an amide bond. Both of these resins
perform similarly and only differ in their initial starting resin [4]. An
alternative approach is to prepare a preformed handle in which the first
building block is prederivatized to the linker and this moiety is attached to
the resin. For peptide synthesis, this practice is common for the preparation of C-terminal peptide acids in order to reduce the amount of
racemization of the a-carbon at the anchoring position [5].
There are three features of a linker that will determine which support
is applicable to a synthetic scheme: (1) the functionality of the molecule at
the anchoring position required for attachment; (2) cleavage conditions;
and (3) the resulting functionality at the anchoring position of the molecule
after the cleavage. As a continuing review on resins and linkers, this
discussion will focus on the work that has been developed from 1997 to
1999 (Refs. 6–9 and references cited therein) and is described according to
Figure 3
178
Pan et al.
Figure 4
the resulting functionality at the anchoring position after the cleavage. In
addition, references and structures will be provided herein for the key
linkers/resins that are commercially available or routinely used in solidphase peptide and organic synthesis.
II. RESINS AND LINKERS FOR CARBOXYLIC
ACID GENERATION
Most peptides contain a carboxylic acid or carboxamide at the Cterminus of the polymer chain. Since Merrifield introduced solid-phase
peptide synthesis in 1963, peptide chemists developed linkers that will
yield these two functionalities upon release from the solid support using a
variety of chemical conditions. The hydroxyl-containing resins (Fig. 5)
based upon alkoxy-substituted benzyl alcohols were developed to supplement chloromethyl- and hydroxymethyl-PS. Carboxylic acids are esterified to the resin typically using N,NV-diisopropylcarbodiimide (DIPCDI).
The acid strength required for release of the molecule from the solid
support is related to the electron donor substituents on the benzene ring
which stabilize the transient resin-bound carbocation. With greater
resonance stabilization conveyed by additional alkoxy groups and aryl
rings, milder acidic conditions are required for cleavage. PAM ( phydroxymethylphenylacetic acid) resin 6 [10] with no electron-donating
groups requires HF treatment while Wang 7 [11], HMPA 8 (4-hydroxymethylphenylacetic acid, also referred to as PAB [ p-alkoxylbenzyl] or
PAC [peptide acid]) [12] and DHPP 9 (4-(1V,1V-dimethyl-1V-hydroxypropyl)phenoxyacetyl) [13] contain resonance stabilizing groups and cleavage is affected with trifluoroacetic acid (TFA). The hydroxymethylbenzyl
Linkers and Resins
179
Figure 5
linker 10 (HMB) [14] contains a carbonyl group para to the ester anchor
and is activated to nucleophilic attack such as hydroxide ion and is stable
toward acid. Alkoxybenzyl derivatives with greater electron donor
strength (Fig. 6) such as SASRIN (super-acid-sensitive resin) 11 [15],
Rink acid 12 [16], and HAL (hyper-acid sensitive) 13 [17] resin allow
carboxylic acids to be cleaved using a lower acid concentration (typically
Figure 6
180
Pan et al.
Figure 7
1–10% TFA) in DCM (dichloromethane). Linkers for carboxylic acids
have also been designed to effect cleavage via photolysis (3-nitro-4hydroxymethylbenzoic acid, ONb 14) [18] and flouridolysis (N-3 or 4)
((4-hydroxymethyl)-phenoxy-t-butylphenylsilyl)phenyl pentanedioic acid
monoamide (PBs) 15 [19] and quinonemethide-based handle 16 [20] (Fig.
7).
Fluorenone derived linker 17 prepared in two steps was coupled to
aminomethyl-PS via DIPCDI [21]. Due to the presence of an electronwithdrawing carboxamide group, the release of carboxylic acids from this
support requires strong acids, such as trifluoromethanesulfonic acid
(TFMSA) (Scheme 1). Insertion of an oxygen adjacent to the biphenyl
rings to the fluorenone scaffold provides xanthene 18 handle [22]. The
oxygen is strategically located to decrease the acid concentration required
Scheme 1
Linkers and Resins
181
Scheme 2
for cleavage and carboxylic acids are released using TFA (Scheme 2). Resin
bound diazo linker 19 was synthesized starting from Wang resin and was
further oxidized to a benzyl aldehyde (Scheme 3) [23]. Carboxylic acids are
anchored to the support in a rapid, colorimetric reaction and are released
upon TFA treatment.
Photolabile linkers play an important role in solid-phase organic
synthesis (SPOS) due to their stability under both acidic and basic
conditions. The ONb photolabile linker was modified to improve cleavage rates and yields; Fmoc-Tos-OH was released in 87% yield after 23 h
(Scheme 4) [24]. Specifically, the primary alcohol was changed to a
secondary benzylic alcohol and the attachment to the resin was through
an alkyl chain as opposed to an amide function. Linker 20 was used for
the production of carboxylic acids or carbohydrates. A second example
Scheme 3
182
Pan et al.
Scheme 4
incorporated a dithiane function to serve as a safety catch against
premature photoreaction [25–26]. A carboxylic acid functionality was
coupled to linker 21 via DIPCDI, the dithiane protecting group was
removed by an S-alkylating reagent such as methyl triflate, and release of
the molecule was accomplished with UV irradiation in THF-methanol
(Scheme 5). Based on 2-pivaloylglycerol, photolabile linker 22 was
prepared in six steps from the dimer of 1,3-dihydroxyacetone (Scheme
6) [27]. The handle was attached to TentaGel S NH2 amino resin, the
protecting groups from the hydroxyl functions were removed, and a series
of peptides were assembled. Cleavage rates were reported to be faster
than other photolabile linkers.
Silyl-based linker 23, cleaved by either basic (TBAF) or neutral
(CsF) fluoridolysis to release carboxylic acids, alcohols, or amines, was
prepared by treatment of a Grignard reagent to an aldehyde resin [28].
To demonstrate the utility of this handle, p-bromobenzoic acid was
Scheme 5
Linkers and Resins
183
Scheme 6
attached to the support and cleavage was accomplished in TBAF in
DMF at 65jC or CsF in DMF at 90jC in 78% and 77% yield, respectively (Scheme 7).
Redox-sensitive resin 24 designed for solid-phase peptide synthesis
(SPPS) [29] was prepared from commercially available 2,5-dimethylbenzoquinone in seven steps [30] and loaded to a support via a Wittig
reaction. Release of the peptide occurs using two sequential mild
conditions, reduction with NaBH4 followed by TBAF-catalyzed cyclic
ether formation (Scheme 8) which provide orthogonality to acid sensitive reactions.
Allylic hydroxycrotyl-oligoethylene glyco-n-alkanoyl (HYCRON)
linker 25 was applied to the synthesis of protected peptides and glycopeptides [31]. HYCRON is stable to both acidic and basic conditions and is
compatible with Boc- and Fmoc-based chemistry. The preparation of this
novel linker is only two steps from commercially available materials.
HYCRON linker can be cleaved under neutral conditions using Pd catalyst
(Scheme 9).
Scheme 7
184
Pan et al.
Scheme 8
III. RESINS AND LINKERS FOR GENERATION
OF AMIDE FUNCTION
Functionalized supports with amino groups such as benzhydrylamine
(BHA) 26 [32] and 4-methylbenzhydrylamine (MBHA) 3 [3] provided Cterminal amides upon HF cleavage (Fig. 2). Polyalkoxyaminobenzyl and
alkoxydiphenylamino resins such as PAL (5-(4-aminomethyl-3,5-dime-
Scheme 9
Linkers and Resins
185
Figure 8
thoxyphenoxy)valeric acid) 27 [33], Rink amide support (RAM) 28 [34]
and 4-(4V-methoxyvenzhydryl)phenoxyacetic acid (Dod) linker 29 [35]
contain more electron-donating groups and were designed on the same
principles as discussed above for the hydroxymethyl supports (Fig. 8).
These three linkers are the most widely used in SPOS and require TFA for
cleavage. Xanthone-based handles XAL (xanthenyl amide linder) 30 [36]
and Seiber 31 [37] resin were designed to release amides using low
concentrations of TFA (Fig. 9). Handles which contain an aminomethyl
and o-nitrobenzyl function (Nb [nitrobenzyl] 32 [38], NBHA [nitrobenzylamine] 33 [39], and a-methyl-6-nitroveratrylamine) 34 [40] are cleaved by
photolysis and are based upon the same principles discussed for hydroxyl
resins (Fig. 10).
Figure 9
186
Pan et al.
Figure 10
In an extension to the xanthenyl theme, the benzyl hydrogen was
replaced with a substituted p-methoxyphenyl ring to give linker 35 (Scheme
10) [41]. Peptide amides were cleaved rapidly and in high purity with TFADCM (1:9) for 15 min or as a protected fragment with TFA-DCM (1:99)
for 3–10 min.
Silyl-derived linker 36 was prepared in three steps from a silyl ether of
serine and incorporated for Fmoc/tBu-based assembly of protected glycopeptide blocks (Scheme 11) [42]. The a-carboxylic acid function of serine
was protected as an allyl ester. Deprotection by a Pd(0) catalyst in the
presence of dimedone liberated the carboxylic acid in order for subsequent
Scheme 10
Linkers and Resins
187
Scheme 11
coupling with amines, alcohols, and carbohydrates. The final glycopeptide
product was released from the support by fluoridolysis (CsF).
As an extension to the p-carboxybenzenesulfonamide ‘‘safety-catch’’
linker [43,44], alkanesulfonamide handle 37 was developed [45]. This linker
tethers carboxylic acids to the solid support to give an acylated sulfonamide which is stable to both basic and acidic conditions (Scheme 12).
Products were released by treatment with iodoacetonitrile followed by the
addition of a nucleophile.
Aryl hydrazide linker 38 stable to both acid and base was utilized in
SPPS [46]. Treatment of the resin with a copper(II) catalyst in the presence
of a base and nucleophile gave the corresponding acid, amide, or ester
(Scheme 13).
Scheme 12
188
Pan et al.
Scheme 13
IV. RESINS AND LINKERS FOR N-SUBSTITUTED
AMIDE GENERATION
Linker 39 with an aldehyde attachment point permits amine anchoring
via reductive amination (Fig. 11) [47]. In peptide synthesis, the handle
attaches the amino as opposed to carboxylic acid function of the Cterminal residue to the support followed by chain elongation (attachment
to the peptide occurs via a backbone nitrogen). The same strategy for
developing handles functionalized with an aldehyde is similar to the
concepts described above. Backbone amide linker (BAL) 40 was prepared from the Fmoc-based tris(alkoxy)benzylamide handle PAL [48]. In
peptide synthesis, BAL allows the preparation of sequences having a
variety of C-terminal functionalities such as alcohols, N-alkyl amides,
and head-to-tail cyclic peptides that are devoid of a trifunctional amino
acid. Due to the electron-donating groups contained in the handle,
release of the peptide is accomplished with a high concentration of
TFA. Based upon the BAL concept, the acid sensitive methoxybenzal-
Figure 11
Linkers and Resins
189
Scheme 14
dehyde polystyrene resin (AMEBA) 41 was reported for the solid-phase
synthesis of sulfonamides, amides, ureas, and carbamates [49]. Reductive amination of aldehydes and ketones with sodium cyanoborohydride
to Rink amide linker generated N-alkyl amines. Acylation followed by
cleavage with TFA provided a method to generate a series of difunctional amines and N-substituted amide derivatives [50]. Subsequently,
backbone linker 42 for Boc-based peptide was developed from a 4alkoxybenzyl derivative in which products were released upon HF
treatment (Scheme 14) [51].
Contrary to an alkoxy benzene scaffold, secondary amides were
generated via novel aldehyde linker 43 based upon an indole scaffold
(Scheme 15) [52]. The indole resin was prepared from indole-3-carboxyaldehyde in two steps and reacted with amines under reductive conditions
to generate resin-bound secondary amines. Treatment of the resin with
Scheme 15
190
Pan et al.
Scheme 16
various reagents followed by TFA yielded amides as well as sulfonamides,
carbamates, and ureas in high yields.
MAMP (Merrifield, Alpha-MethoxyPhenyl) resin 44 is an alternative to aldehyde linkers to construct N-substituted amides [53]. Nucleophilic displacement of the benzylic chloride with an amine followed by
acylation yielded a secondary amide which was released upon a low
(f10%) concentration of TFA (Scheme 16).
V. RESINS AND LINKERS FOR HYDROXYL
AND GENERATION OF AMINO FUNCTION
Dihydropyran (DHP) linker 45 is a common handle that couples an
alcohol to a solid support with subsequent release upon mild TFA conditions (Fig. 12) [54]. An alternative approach is to prepare an active
carbonate linker. N,NV-Disuccinimidyl carbonate (DSC), a valuable
reagent for converting hydroxymethyl-based supports to their corresponding carbonates, was reacted with 4-hydroxymethylpolystyrene 46
and 4-nitrobenzamido (Nbb) 47 resins to anchor alcohols and phenols
(Scheme 17) [55]. The final products were released from the solid support
by HF and photolysis, respectively.
Linkers and Resins
191
Figure 12
p-Benzyloxybenzylamine (BOBA) 48 is a new class of an amine
support and was prepared from Merrifield resin in two steps [56]. BOBA
resin was treated with an aldehyde in the presence of an acid to give an
imine that subsequently reacted with Yb(OTf )3-catalyzed silyl enolates
(Scheme 18). Cleavage with trimethylsilyl triflate (TMSOTf) or 2,3dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) gave either phenols or
amines, respectively.
9-Phenylfluoren-9-yl polystyrene (Phfl) based resin 49 was applied in
the solid-phase synthesis of hydroxyl and amino functions [57,58]. This
resin has higher acid stability compared to the structurally similar trityl
resin. Final release of the product is accomplished with TFA in high purity
(Scheme 19).
Trialkylsilane resin (PS-DES) 50 was incorporated for solid-phase
glycosylation by anchoring a glycosyl donor via their corresponding thiophenyl ether or h-glucopyranosyl fluorides (Scheme 20) [59]. Disaccharides
Scheme 17
192
Pan et al.
Scheme 18
were prepared by reaction with a glycosyl acceptor followed by cleavage
with acetic acid (AcOH).
Phenols were constructed form novel serine-derived handle 51,
which was stable to acids (TFA) and bases (pyridine) (Scheme 21) [60].
The final products were released from the support by fluoride ion.
A variety of cleavage conditions have been reported for the release
of amines from a solid support. Triazene linker 52 prepared from
Merrifield resin in three steps was used for the solid-phase synthesis of
aliphatic amines (Scheme 22) [61]. The triazenes were stable to basic
conditions and the amino products were released in high yields upon
treatment with mild acids. Alternatively, base labile linker 53 synthesized
from a-bromo-p-toluic acid in two steps was used to anchor amino
functions (Scheme 23) [62]. Cleavage was accomplished by oxidation of
the thioether to the sulfone with m-chloroperbenzoic acid followed by helimination with a 10% solution of NH4OH in 2,2,2-trifluoroethanol. A
linker based on 1-(4,4V-dimethyl-2,6-dioxocyclohexylidene)ethyl (Dde)
primary amine protecting group was developed for attaching amino
functions (Scheme 24) [65]. Linker 54 was stable to both acidic and
basic conditions and the final products were cleaved from the resin by
treatment with hydrazine or transamination with n-propylamine.
Scheme 19
Linkers and Resins
193
Scheme 20
REM linker 55 (regenerated after cleavage, functionalized by
Michael addition) is a traceless handle for anchoring secondary amines
(Fig. 13) [64]. Tertiary amines were prepared on this linker via basedinduced Hofmann elimination of the subsequent quaternary ammonium
salt. An analogous vinyl sulfone linker was prepared from Merrifield resin
to perform the identical synthetic strategy (Scheme 25) [65]. Similar to
REM, vinyl sulfone handle 56 was regenerated following cleavage, but
was more stable to acids and nucleophiles such as Grignard reagents than
the former. An extension to the vinyl sulfone theme was demonstrated by
inserting a carbamate function at the anchoring position for the assembly
of 2- and 2,4-substituted pyrrolidines (Scheme 26) [66]. The acid stable,
base labile (final cleavage accomplished with NaOMe) support 57 was
used for N-acyliminium ion reactions.
Scheme 21
194
Scheme 22
Scheme 23
Scheme 24
Pan et al.
Linkers and Resins
Figure 13
Scheme 25
Scheme 26
195
196
Pan et al.
Scheme 27
N-Protected amines were assembled on solid-phase via sulfonamidebased handle 58 (Scheme 27) [67]. Tertiary sulfonamides were generated
upon reaction with allylic, benzylic and primary alcohols under Mitsunobu conditions. Secondary amines were released from the support using
mild nucleophilic conditions such as treatment with thiophenol and
potassium carbonate.
A versatile approach for the solid-phase synthesis of aminopyridazines used the anchoring of 3,6-dichloropyridazine to resin-bound
thiophenol 59 (Scheme 28) [68]. Treatment with nucleophilic amines
released the aminopyridazine products from the solid support without
further oxidation.
Traceless linker 60 based on a benzotriazole scaffold was reacted with
amines and aldehydes to produce Mannich-type amine products [69]. Final
product release was achieved by treatment with Grignard reagents
(Scheme 29).
Scheme 28
Scheme 29
Linkers and Resins
197
Figure 14
VI. RESINS AND LINKERS FOR HYDROXAMIC
GENERATION OF ACID FUNCTIONS
Hydroxamic acids are an important class of compounds targeted as potential therapeutic agents. N-Fmoc-aminooxy-2-chlorotrityl polystyrene
resin 61 allowed the synthesis and subsequent cleavage under mild conditions of both peptidyl and small molecule hydroxamic acids (Fig. 14)
[70]. An alternative hydroxylamine linkage 62 was prepared from trityl
chloride resin and N-hydroxyphthalimide followed by treatment with
hydrazine at room temperature (Scheme 30) [71]. A series of hydroxamic
acids were prepared by the addition of substituted succinic anhydrides to
the resin followed by coupling with a variety of amines, and cleavage with
HCOOH-THF(1:3).
Scheme 30
198
Pan et al.
VII. RESINS AND LINKERS FOR GENERATION
OF SULFONAMIDE, UREA, AND
GUANIDINE FUNCTIONS
Aminosulfonyl ureas were constructed from a sulfonylcarbamate linkage (Scheme 31) [72]. Reaction of chlorosulfonyl isocyanate (CSI) with
Wang resin provided a chlorosulfonylcarbamate 63 which was then
converted to substituted amino sulfonylcarbamate compounds by reaction with excess amines. The final aminosulfonyl urea products were
cleaved from the resin by treatment with amines in HF at reflux
temperature for overnight.
Urea libraries were assembled via thiophenoxy carbonyl linker 64
readily available in two steps from Merrifield resin (Scheme 32) [73].
Treatment of this linker with primary or secondary amines, followed by
basic cleavage with amines generated the ureas. An alternative
approach for the synthesis of ureas was to treat p-nitrobenzophenone
oxime resin with phosgene to give p-nitrophenyl(polystyrene)ketoxime
(Phoxime resin) 65 [74]. The addition of primary amines to the
phosgenated oxime linker gave a resin-bound carbamate. Ureas were
genated by reaction with a second set of amines at temperatures greater
than 80jC (Scheme 33).
Traditional SPPS anchors the peptide to the support via the acarboxylic acid of the C-terminal residue. Novel sulfonyl linker 66 was
prepared to side-chain anchor the guanidine function of arginine
(Scheme 34) [75]. To demonstrate the utility of the linker, tripeptide
H-Phe-Arg-Ala-OMe was assembled in which amino acids were
extended to the anchoring residue in both the C- and N-terminal
directions. HF cleavage released the peptide from the support. Small
molecules containing guanidines were constructed from carbonylimidazole handle 67 generated from Wang resin (Scheme 35) [76]. Treatment
of the carbonylimidazole linker with thiourea basic conditions afforded
Scheme 31
Linkers and Resins
Scheme 32
Scheme 33
Scheme 34
199
200
Pan et al.
Scheme 35
a resin-linked thiourea product. Sulfur displacement of the thiourea
resin with primary and secondary amines followed by TFA cleavage
provided guanidine-containing products. Both mono- and disubstituted
guanidines were prepared in good yields and purities using acyl isothiocyanate resin 68 prepared from carboxypolystyrene in two simple steps
(Scheme 36) [77]. Reaction of a variety of amines with this resin produced
the corresponding acyl thioureas under mild conditions. The guanidine
moiety formation was achieved by exposing the acyl thiourea resin to a
primary or secondary amine. Cleavage of the acyl guanidine was effected
by treatment with TFA-CHCl3-MeOH (1:1:1).
VIII. RESINS AND LINKERS FOR GENERATION
OF ALDEHYDE FUNCTIONS
The Leznoff acetal linker 69 was used to anchor an aldehyde to the solid
support and following a series of reactions, the aldehyde was released by
acidic cleavage [78]. An application of this resin was demonstrated for a
biaryl aldehyde library synthesis which incorporated a Suzuki–Miyaura
reaction (Scheme 37) [79]. Cleavage was effected by a solution of 3 M HCl
Scheme 36
Linkers and Resins
201
Scheme 37
at 80jC to give excellent yields for most of the products. An alternative
strategy implemented serine or threonine as a linker to anchor an aldehyde
to the solid support [80]. Unlike the acetal formation described above, this
linker reacts with an aldehyde to form an oxazoline with release from the
support by aqueous acid (such as HOAc) at 60jC.
A second strategy employed tartaric acid-based linker 70 prepared
from an amino PEGA resin in which C-terminal a-oxo-aldehydes were
generated by an oxidative cleavage (Scheme 38) [81]. Following linear
assembly of the peptide by Fmoc chemistry, TFA treatment removed the
side-chain protecting groups and converted the anchoring acetonide to a
1,2-diol which was oxidized to the aldehyde with NaIO4.
IX. RESINS AND LINKERS FOR GENERATION
OF OTHER FUNCTIONS
Cleavage of all the linkers described above provide a functional group
(carboxylic acid, amide, amine, etc) at the anchoring position. Silyl-based
handles 71,72, and 73 as well as germanium-based handle 74 insert a C-H
bond at the anchoring position and are referred to as traceless (Fig. 15) [82–
Scheme 38
202
Pan et al.
Figure 15
85]. A further extension to this concept was (dimethylsilyl)propionic acid
linker 75 used for the solid-phase synthesis of aryl-containing organic
compounds [86]. The linker was cleaved smoothly with TFA and has been
used for the synthesis of compounds which involved alkylation, acylation,
and Mitsunobu reactions.
Silicon linker 76 was used for direct loading of aromatic compounds to supports for the assembly of pyridine-based tricyclics (Scheme
39) [87]. Following the initial coupling of an aromatic bromide to the
resin by halogen/metal exchange in the presence of tert-butyl lithium, a
Linkers and Resins
203
Scheme 39
series of reactions including TFA deprotection of Boc, alkylation under
strong basic condition, SnCl2 reduction, and ring cyclization were
performed. The final tricyclic products were released from the polymer
via basic fluoride (Bu4N+-F) in THF at room temperature. A similar
trialkylsilane linker was synthesized from Merrifield resin in two steps
[88].
Piperazine linker 77 was treated with propargyltriphenylphosphine
bromide to provide a resin-bound Wittig reagent (Scheme 40) [89]. Base
treatment followed by aldehyde addition produced a resin-bound 2-aminobutadiene which was implemented in [4+2] cycloadditions. Alternatively,
treatment with 3% TFA in CH2Cl2 released a,h-unsaturated methylketones in high yields.
Scheme 40
204
Pan et al.
Scheme 41
Acetal handle 78 synthesized from Merrifield resin and 4-hydroxybenzaldehyde was applied to the solid-phase synthesis of carbohydrates
and 1-oxacephams (Scheme 41) [90]. For the latter, a 1,3-diol was
initially anchored to the support to form a cyclic acetal. A ring opening
reaction with DIBAL generated a resin-bound alcohol which was
converted to the corresponding triflate for N-alkylation with 4-vinyloxyazetidin-2-one. A Lewis acid catalyzed ring closure released 1-oxacephams from the support.
Aryl hydrazide-based linker 79 was developed as a traceless handle
that released products under mild oxidative conditions (Scheme 42) [91].
Polymeric bound p–iodophenylhydrazide was subjected to a variety of
Pd0-catalyzed coupling reactions (Heck, Suzuki, Sonogashira, and Stille).
Oxidation with Cu(OAc)2 in MeOH and pyridine released the final
products in 50–96% yield.
A traceless linker for solid-phase homo- and hetero-Diels-Alder
reactions was based upon resin bound quinodimethane precursors
Scheme 42
Linkers and Resins
205
Scheme 43
(Scheme 43) [92]. Reaction of dienophiles such as 4-nitrobenzaldehyde
with linker 80 at high temperature gave Diels-Alder products. Dihydropyrans were released from the support by Bronsted or Lewis acid-nucleophile combinations in moderate to good yield with stereoselectivity for the
anti isomer.
X. CONCLUSION
The past decade witnessed a renaissance in drug discovery due to the
emergence of solid-phase synthesis. Initially, the solid supports and linkers
used for the repetitive process of biomolecule assembly applied to the
construction of small molecule libraries and scaffolds were required to
contain a carboxylic acid or amide in order to anchor to the polymeric
support. Thus, the linkers from solid-phase peptide synthesis such as Rink,
Wang, and PAL were commonly employed in the synthetic strategy. As
new bond-forming reactions have been adapted for solid phase as well as
the construction of novel lead compounds, synthetic pathways are requiring additional handles that release compounds into solution upon various
cleavage conditions and provide additional functionality at the anchoring
position. In a retrosynthetic analysis of a library, one should plan the
anchoring strategy as it relates to the functionality of the molecule as well
as to insure that the cleavage conditions are compatible with the synthetic
scheme. Although there are now a plethora of linkers that have been
described in the literature, novel handles still provide medicinal chemists
the tools to expand molecular diversity with the ultimate reward of
discovering a new drug candidate.
206
Appendix A. HANDLES AND DERIVATIZED
SOLID SUPPORTS
Pan et al.
Linkers and Resins
207
208
Pan et al.
Linkers and Resins
209
210
Pan et al.
Linkers and Resins
211
212
Pan et al.
Linkers and Resins
213
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Merrifield RB. Solid phase peptide synthesis I. The synthesis of a
tetrapeptide. J Am Chem Soc 1963; 85:2149–2154.
Adams JH, Cook RM, Hudson D, Jammalamadaka V, Lyttle MH,
Songster MF. A reinvestigation of the preparation, properties, and
applications of aminomethyl and 4-methylbenzhydrylamine polystyrene
resins. J Org Chem 1998; 63:3706–3716.
Matsueda GR, Stewart JM. A p-methylbenzhydrylamine resin for improved
solid-phase synthesis of peptides amides. Peptides 1981; 2:45–50.
Albericio F, Barany G. Improved approach for anchoring N a-9-fluorenylmethyloxycarbonylamino acids as p-alkoxybenzyl ester anchoring linkages.
Int J Peptide Protein Res 1985; 26:92–97.
Bernatowicz MS, Kearney T, Neves RS, Koster H. An efficient method for
racemization-free attachment of 9-fluorenylmethyloxycarbonyl-amino acids
to peptide synthesis supports. Tetrahedron Lett 1989; 30:4341–4344.
Blackburn C, Albericio F, Kates SA. Recent Res Devel Org Chem 1997;
1:477–497.
Blackburn C, Albericio F, Kates SA. Drugs Future 1997; 22:1007–1025.
Resins and handles for solid-phase synthesis. In: Kates SA, ed. Biopolymers:
Peptide Science. 1998; 47:309–411.
Kates SA, Albericio F, eds. Solid-Phase Synthesis: A Practical Guide. New
York: Marcel Dekker, 2000.
Mitchell AR, Kent SBH, Engelhard M, Merrifield R. A new synthetic route
to tert-butyloxycarbonylaminoacyl-4-oxymethyl)phenylacetamidomethylresin, an improved support for solid-phase peptide synthesis. J Org Chem
1978; 43:2845–2852.
Wang SS. p-Alkoxybenzyl alcohol resin and p-alkoxybenzyloxycarbonylhydrazide resin for solid-phase synthesis of protected peptide fragments. J Am
Chem Soc 1973; 95:1328–1333.
Sheppard RC, Williams BJ. Acid-labile resin linkage agents for use in solid
phase peptide synthesis. Int J Peptide Protein Res 1982; 20:451–454.
Akaji K, Kiso Y, Carpino LA. Fmoc-based solid-phase peptide synthesis
using a new t-alcohol type 4-(1V,1V-dimethyl-1V-hydroxypropyl)phenoxyacetyl handle (DHPP)-resin (Fmoc=9-fluorenyloxycarbonyl). J Chem Soc
Chem Comm 1990; 584–586.
Atherton E, Logan CJ, Sheppard RC, Peptide synthesis, part 2:
Procedures for solid phase synthesis using Na-fluorenylmethoxycarbonylamino acids on polyamide supports: Synthesis of substance P and of
acyl carrier protein 64–74 decapeptide. J Chem Soc Perkin Trans 1 1981;
538–546.
Mergler M, Tanner R, Gosteli J, Grogg P. Peptide synthesis by a
combination of solid-phase and solution methods. I: A new very acid-labile
214
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Pan et al.
anchor group for the solid-phase synthesis of fully protected peptide
fragments. Tetrahedron Lett 1988; 29:4005–4008.
Rink H. Solid-phase synthesis of protected peptide fragments using a
trialkoxy diphenyl methylester resin. Tetrahedron Lett 1987; 28:3787–3790.
Albericio F, Barany G. Hypersensitive acid-labile (HAL) tris(alkoxy)benzyl
ester anchoring for solid-phase synthesis of protected peptide segments.
Tetrahedron Lett 1991; 32:1015–1018.
Rich DH, Gurwara SK. Preparation of a new o-nitrobenzyl resin for solidphase synthesis of tert-butyloxycarbonyl-protected peptide acids. J Am
Chem Soc 1975; 97:1575–1579.
Mullen DG, Barany G. A new fluoridolyzable linkage for orthogonal solidphase peptide synthesis: Design, preparation, and application of the N-(3 or
4)((4-hydroxymethyl)-phenoxy-tbutylphenylsilyl)phenyl pentanedioic acid
monoamine (Pbs) handle. J Org Chem 1988; 53: 5240–5248.
Ramage R, Barron CA, Bielecki S, Holden R, Thomas DW. Solid phase
peptide synthesis: Fluoride ion release of peptide from the resin.
Henkel B, Bayer E. 9-Hydroxy-9-(4-carboxyphenyl)fluorene—a new linker
for solid phase synthesis. Tetrahedron Lett 1998; 39:9401–9402.
Henkel B, Zeng W, Bayer E. 9-Hydroxy-9-(4-carboxyphenyl)xanthene- a
new linker for the synthesis of peptide amides. Tetrahedron Lett 1997;
38:3511–3512.
Bhalay G, Dunstan AR. Facile solid phase synthesis of an activated diazo
linker. Tetrahedron Lett 1998; 39:7803–7806.
Rodebaugh R, Fraser-Reid B, Geysen HM. A new o-nitrobenzyl photocleavable linker for solid phase synthesis. Tetrahedron Lett 1997; 38: 7653–
7656.
Routledge A, Abell C, Balasubramanian S. The use of a dithiane-protected
benzoin photolabile safety catch linker for solid-phase synthesis. Tetrahedron Lett 1997; 38:1227–1230.
Lee HB, Balasubramanian S. Studies on a dithiane-protected benzoin
photolabile safety catch linker for solid-phase synthesis. J Org Chem 1999;
64:3454–3460.
Peukert S, Giese B. The pivaloyglycol anchor group: A new platform for
a photolabile linker in solid-phase synthesis. J Org Chem 1998; 63:
9045–9051.
Routledge A, Stock HT, Flitsch SL, Turner NJ. New fluoride-labile
linkers for solid-phase organic synthesis. Tetrahedron Lett 1997; 38:
8287–8290.
Zheng A, Shan D, Shi X, Wang B. A model resin linker for solid-phase
peptide synthesis which can be cleaved using two sequential mild reactions.
J Org Chem 1999; 64:7459–7466.
Zheng A, Shan D, Wang B. A redox-sensitive resin linker for the solid
Linkers and Resins
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
215
phase synthesis of C-terminal modified peptides. J Org Chem 1999; 64:
156–161.
Seitz O, Kunz H. HYCRON, an allylic anchor for high-efficiency solid
phase synthesis of protected peptides and glycopeptides. J Org Chem 1997;
62:813–826.
Pietta PG, Marshall GR. Amide protection and amide supports in a solidphase peptide synthesis. J Chem Soc D 1970; 650–651.
Albericio F, Kneib-Cordonier N, Biancalana S, Gera L, Masada RI,
Hudson D, Barany G. Preparation and application of the 5-(4-(9fluorenylmethoxycarbonyl)aminomethyl-3, 5-dimethoxyphenoxy)-valeric
acid (PAL) handle for the solid-phase synthesis of C-terminal peptide
amides under mild conditions. J Org Chem 1990; 55:3730–3743.
Rink H. Solid-phase synthesis of protected peptide fragments using a
trialkoxy diphenyl methylester resin. Tetrahedron Lett 1987; 28:3787–
3790.
Breipohl G, Knolle J, Stuber W. Synthesis and application of acid labile
anchor groups for the synthesis of peptide amides by Fmoc-solid-phase
peptide synthesis. Int J Peptide Protein Res 1989; 34:262–267.
Han Y, Bontems SL, Hegyes P, Munson MC, Minor CA, Kates SA,
Albericio F, Barany G. Preparation and applications of xanthenylamide
(XAL) handles for solid-phase synthesis of C-terminal peptide amides under
particularly mild conditions. J Org Chem 1996; 61:6326.
Seiber P. A new acid-labile anchor groups for the solid phase synthesis of
C-terminal peptide amides by the Fmoc method. Tetrahedron Lett 1987;
28:2107–2110.
Hammer RP, Albericio F, Gera L, Barany G. Practical approach to
solid-phase synthesis of C-terminal peptide amides under mild conditions
based on photolysable anchoring linkage. Int J Peptide Protein Res 1990;
36:31–45.
Ajayaghosh A, Pillai VNR. Solid-phase synthesis and C-terminal amidation
of peptides using a photolabile o-nitrobenzhydrylaminopolystyrene support.
Holmes CP, Jones DG. Reagents for combinatorial organic synthesis:
Development of a new o-nitrobenzyl photolabile linker for solid-phase
synthesis. J Org Chem 1995; 60:2318–2319.
Meisenbach M, Echner H, Voelter W. New methoxy-substituted 9-phenylxanthen-9-ylamine linkers for the solid phase phase synthesis of protected
peptide amides. Chem Commun (Cambridge) 1997; 849–850.
Nakamura K, Hanai N, Kanno M, Kobayashi A, Ohnishi Y, Ito Y,
Nakahara Y. Design and synthesis of silyl ether-based linker for solid-phase
synthesis of glycopeptides. Tetrahedron Lett 1999; 40:515–518.
Kenner GW, McDermott JR, Sheppard RC. The safety catch principle in
solid-phase peptide synthesis. J Chem Soc Chem Comm 1971; 636–637.
216
Pan et al.
44. Backes BJ, Ellman JA. Carbon-carbon bond-forming methods on solid
support. Utilization of Kenner’s ‘‘safety-catch’’ linker. J Am Chem Soc
1994; 116:11171–11172.
45. Backes BJ, Ellman JA. An alkanesulfonamide ‘‘safety-catch’’ linker for
solid-phase synthesis. J Org Chem 1999; 64:2322–2330.
46. Millington CR, Quarrell R, Lowe G. Aryl hydrazides as linkers for solid
phase synthesis which are cleavable under mild oxidative conditions.
47. Boojmara CG, Burow K, Ellman JA. An expedient and high-yielding
method for the solid-phase synthesis of diverse 1,4-benzodiazepine-2,5diones. J Org Chem 1995; 60:5742–5743.
48. Jensen KJ, Alsina J, Songster MF, Vagner J, Albericio F, Barany G.
Backbone amide linker (BAL) strategy for solid-phase synthesis of Cterminal-modified and cyclic peptides. J Am Chem Soc 1998; 120:5441–
5452.
49. Fivush AM, Willson TM. AMEBA: An acid sensitive aldehyde resin for
solid phase synthesis. Tetrahedron Lett 1997; 38:7151–7154.
50. Brown EG, Nuss JM. Alkylation of Rink’s amide linker on polystyrene
resin: a reductive amination approach to modified amine-linkers for the
solid phase synthesis of N-substituted amide derivatives. Tetrahedron Lett
1997; 38:8457–8460.
51. Bourne GT, Meutermans WDF, Alewood PF, McGeary RP, Scanlon M,
Watson AA, Smythe ML. A backbone linker for Boc-based peptide
synthesis and on-resin cyclization: Synthesis of stylostatin 1. J Org Chem
1999; 64:3095–3101.
52. Estep KG, Neipp CE, Stramiello LM, Stephens A, Mavis D, Allen MP.
Indole resin: A versatile new support for the solid-phase synthesis of organic
molecules. J Org Chem 1998; 63:5300–5301.
53. Brown DS, Revill JM, Shute RE. Merrifield Alpha-Methoxyphenyl
(MAMP) Resin; A new versatile solid support for the synthesis of secondary
amides. Tetrahedron Lett 1998; 39:8533–8536.
54. Thompson LA, Ellman JA. Straightforward and general method for
coupling alcohols to solid supports. Tetrahedron Lett 1994; 35:9333–
9336.
55. Alsina J, Chiva C, Ortiz M, Rabanal F. Active carbonate resins for solidphase synthesis through the anchoring of a hydroxyl function. Synthesis of
cyclic and alcohol peptides. Tetrahedron Lett 1997; 38:883–886 and
references cited therein.
56. Kobayashi S, Aoki Y. p-Benzyloxybenzylamine (BOBA) resin. A new
polymer-supported amine used in solid-phase organic synthesis. Tetrahedron Lett 1998; 39:7345–7348.
57. Bleicher KH, Wareing JR. Phfl Polystyrene: A new resin for solid phase
organic synthesis. Tetrahedron Lett 1998; 39:4587–4590.
Linkers and Resins
217
58. Bleicher KH, Wareing JR. Phfl Acetic Acid: A new linker for solid phase
organic synthesis. Tetrahedron Lett 1998; 39:4591–4594.
59. Doi T, Sugiki M, Yamada H, Takahashi T, Porco JA, Jr. Glycosidation of
solid-supported glycosyl donors tethered by a trialkylsilane linker.
60. Chou Y-L, Morrissey MM, Mohan R. Novel serine-based linker for the
solid-phase synthesis of organic compounds. Tetrahedron Lett 1998;
39:757–760.
61. Brase S, Kobberling J, Enders D, Lazny R, Wang M, Brandtner S.
Nitrogen-based linker. 3. Triazenes as robust and simple linkers for
amines in solid-phase organic synthesis. Tetrahedron Lett 1999; 40:2105–
2108.
62. Garcia-Echeverria C. A base labile handle for solid phase chemistry.
63. Chhabra SR, Khan AN, Bycroft BW. Versatile Dde-based primary amine
linkers for solid phase synthesis. Tetrahedron Lett 1998; 39:3585–3588.
64. Gayo KM, Suto MJ. Traceless linker: Oxidative activation and displacement of a sulfur-based linker. Tetrahedron Lett 1997; 38:211–214.
65. Kroll FEK, Morphy R, Rees D, Gani D. Resin-immobilized benzyl and aryl
vinyl sulfones: New versatile traceless linkers for solid-phase organic
synthesis. Tetrahedron Lett 1997; 38:8573–8576.
66. Veerman JJN, Rutjes FPJT, Van Maarseveen JH, Hiemstra H. A novel acid
stable/base labile carbamate linker for N-acyliminium ion reactions on solid
support. Tetrahedron Lett 1999; 40:6079–6082.
67. Kay C, Murray PJ, Sandow L, Holmes AB. A novel, chemically robust,
amine-releasing linker. Tetrahedron Lett 1997; 38:6941–6944.
68. Parrot I, Wermuth C-G, Hibert M. Resin-bound thiophenols as SNARlabile linkers: Application to the solid phase synthesis of aminopyridazines.
69. Schiemann K, Showalter HDH. Development of polymer-supported
benzotriazole as a novel traceless linker for solid-phase organic synthesis.
J Org Chem 1999; 64:4972–4975.
70. Mellor SL, McGuire C, Chan WC. N-Fmoc-aminooxy-2-chlorotrityl
polystyrene resin: A facile solid-phase methodology for the synthesis of
hydroxamic acids. Tetrahedron Lett 1997; 38:3311–3314.
71. Bauer U, Ho W-B, Koskinen AMP. A novel linkage for the solid-phase
synthesis of hydroxamic acids. Tetrahedron Lett 1997; 38:7233–7236.
72. Fitzpatrick LJ, Rivero RA. Solid phase synthesis of substituted aminosulfonylureas using a sulfonylcarbamate linker. Tetrahedron Lett 1997;
38:7479–7482.
73. Dressman BA, Singh U, Kaldor SW. Solid phase synthesis of urea libraries
using a diversifiable thiophenoxy carbonyl linker. Tetrahedron Lett 1998;
39:3631–3634.
218
Pan et al.
74. Scialdone MA, Shuey SW, Soper P, Hamuro Y, Burns DM. Phosgenated pnitrophenyl(polystyrene)ketoxime or phoxime resin. A new resin for the
solid-phase synthesis of ureas via thermolytic cleavage of oxime-carbamates.
J Org Chem 1998; 63:4802–4807.
75. Zhong HM, Greco MN, Maryanoff BE. Solid-phase synthesis of argininecontaining peptides by guanidine attachment to a sulfonyl linker. J Org
Chem 1997; 62:9326–9330.
76. Josey JA, Tarlton CA, Payne CE. Novel linker for the solid-phase synthesis
of guanidines. Tetrahedron Lett 1998; 39:5899–5902.
77. Wilson LJ, Klopfenstein SR, Li M. A traceless linker approach to the solid
phase synthesis of substituted guanidines utilizing a novel acyl isothiocyanate resin. Tetrahedron Lett 1999; 40:3999–4002.
78. Leznoff CC, Wong JY. The use of polymer supports in organic synthesis.
III. Selective chemical reactions on one aldehyde group of symmetrical
aldehydes. Can J Chem 1973, 51:3756–3764.
79. Chamoin S, Houldsworth S, Kruse CG, Bakker WI, Snieckus V. The
Suzuki-Miyaura cross coupling reactions on solid support. Link to
solution phase directed ortho metalation. The Leznoff acetal linker
approach to biaryl and heterobiaryl aldehydes. Tetrahedron Lett 1998; 39:
4179–4182.
80. Ede NJ, Bray AM. A simple linker for the attachment of aldehydes to the
solid phase. Application to solid phase synthesis by the Multipin method.
81. Fruchart J-S, Gras-Masse H, Melnyk O. A new linker for the synthesis of
C-terminal peptide a-oxo-aldehydes. Tetrahedron Lett 1999; 40:6225–6228.
82. Chenera B, Finkelstein JA, Veber DF. Protodetachable arylsilane polymer
linkages for use in solid phase organic synthesis. J Am Chem Soc 1995;
117:11999–12000.
83. Plunkett MJ, Ellman JA. A silicon-based linker for traceless solid-phase
synthesis. J Org Chem 1995; 60:6006–6007.
84. Ellman JA, Plunkett MJ. Germanium and silicon linking strategies for
traceless solid-phase synthesis. J Org Chem 1997; 62:2885–2893.
85. Boehm TL, Showalter HDH. Development of a novel silyl ether linker for
solid-phase organic synthesis. J Org Chem 1996; 61:6498–6499.
86. Newlander KA, Chenera B, Veber DF, Yim NCF Moore ML. Simple silyl
linker for the solid phase organic synthesis of aryl-containing molecules.
J Org Chem 1997; 62:6726–6732.
87. Woolard FX, Paetsch J, Ellman JA. A silicon linker for direct loading of
aromatic compounds to supports. Traceless synthesis of pyridine-based
tricylics. J Org Chem 1997; 62:6102–6103.
88. Hu Y, Porco JA, Jr, Labadie JW, Gooding OW, Trost BM. Novel polymersupported trialkylsilanes and their use in solid–phase organic synthesis.
J Org Chem 1998; 63:4518–4521.
Linkers and Resins
219
89. Hird NW, Irie K, Nagai K. Solid phase synthesis of 2-aminobutadienes
using a piperazine linker. Tetrahedron Lett 1997; 38:7111–7114.
90. Furman B, Thurmer R, Kahuza Z, Voelter W, Chmielewski M. A new
acetal resin valuable for the solid-phase synthesis of 1-oxacephams via a
cyclization/cleavage step. Tetrahedron Lett 1999; 40:5909–5912.
91. Stieber F, Grether U, Waldmann H. An oxidation-labile traceless linker for
solid-phase synthesis. Angew Chem, Int Ed 1999; 38:1073–1077.
92. Craig D, Robson MJ, Shaw SJ. Traceless linkers for solid-phase synthesis.
Homo- and hetero Diels-Alder reactions of ortho-quinodimethanes. Synlett
1998: 12:1381–1383.
8
Allosteric Modulation
of G-Protein-Coupled
Receptors: Implications
for Drug Action
Angeliki P. Kourounakis
University of Thessaloniki, Thessaloniki, Greece
Pieter van der Klein and Ad P. IJzerman
I.
INTRODUCTION
Representing one of the largest superfamilies of proteins in the human
body, G-protein-coupled receptors (GPCRs) play a crucial role in the regulation of a variety of physiological processes, particularly within the
central nervous system and the cardiovascular and endocrine systems. It is
estimated that this superfamily comprises about 500 and possibly over 1000
receptor (sub)types having similar structural and/or sequence motifs, while
operating via common transduction mechanisms to mediate the transmission of extracellular signals into biochemical or electrophysiological
responses in a cell. A specific endogenous molecule, such as a neurotransmitter or hormone, acts as the signaling species that binds to the receptor,
resulting in an activation of intracellular G proteins and signal propagation. Hence, GPCRs are important drug targets; approximately 60% of
current drugs produce their therapeutic actions by binding to GPCRs.
221
222
Kourounakis et al.
The design and development of synthetic drugs is mainly focused on
mimicking (as in the case of agonists) or blocking (as in the case of
antagonists and inverse agonists) the action of the endogenous signaling
molecule by competing at the same site (the ligand binding site) on a
specific receptor. Recently, a new concept of interference with drug action
at GPCRs has emerged for some receptor subclasses such as the muscarinic or adenosine receptor. This concept, namely allosteric modulation of
the receptor by molecules binding at a second (allosteric) site, is thus far
relatively unexplored for GPCRs, although relatively common in the
family of ion channel receptors [1]. This indirect (allosteric) mechanism
(i.e., the modulation of the efficacy or affinity of the endogenous ligand for
its receptor) is the molecular basis of the therapeutic action of benzodiazepines that interact with g-aminobutyric acid A (GABAA) receptors
coupled through ion channels. In contrast, there has been no therapeutic
role found for directly acting agonists or antagonists on this receptor.
Nevertheless, only a few drugs, such as gallamine, alcuronium, or
pancuronium, are known to exert their action at an allosteric site on a
GPCR [2].
The potential advantages or benefits of allosteric drugs over agonists,
antagonists, and inverse agonists may be elaborated as follows.
It is generally found that within the GPCR family, subtypes exist
that bind the same signaling molecule but have different tissue distributions as well as functions. These receptor subtypes have often a high
sequence homology, especially in the regions of the receptor that are
thought to contain the ligand binding site. Thus, in most cases it has been
proven difficult to develop drugs that not only are highly selective for one
receptor subtype but have highly controlled effects on the function of that
receptor and act in those tissues only where their action is desired.
An allosteric drug has the following properties:
1. Has no action when binding on its own to the receptor but only
modulates the actions of the naturally occurring hormone or
neurotransmitter when it is released. Therefore, the temporal
aspects of the natural signaling mechanism are retained and
desensitization is minimized.
2. Has a defined maximum effect that is determined by the
cooperativity associated with its allosterism.
3. Can act selectively at various receptor subtypes not only by
means of its own affinity but also on cooperativity. No pharmacological agent has yet exploited the latter property.
Modulation of G-Protein-Coupled Receptors
223
4. Has enhancing properties that can selectively intensify a
weakened signal from a specific receptor subtype, alleviating the
effects caused by a localized neurotransmitter deficit such as in
Alzheimer’s or Parkinson’s disease.
Although allosteric sites have been characterized on certain biogenic
amine receptors—muscarinic acetylcholine, dopamine, and a-adrenergic
receptors—as well as on adenosine receptors, it is not yet known whether
the presence of an allosteric site is a characteristic of only a few, all, or
subsets of GPCRs. Furthermore, a relevant question is whether these sites
might have a physiological regulatory role as a consequence of binding
with endogenous molecules. Interestingly, there are a few recent reports
showing endogenous ligands affecting GPCR binding and function allosterically. The endogenous tetrapeptide Leu-Ser-Ala-Leu, released from
nerve terminals upon depolarization, inhibits 5HT1B receptor binding and
function at nanomolar concentrations [3,4], an effect specific at 5HT1B and
not at other 5-hydroxytryptamine receptors that were examined. Also, a
recent study reported that binding and function of the human oxytocin
receptor can be inhibited directly by nanomolar concentrations of 5hdihydroprogesterone [5]. Although the actions of the peptide and the
steroid have a number of common features that make their interactions
different from those previously observed, both studies suggest an entirely
unanticipated cross talk between very different signaling mechanisms, the
consequences of which are not yet known.
II. DEFINITION OF ‘‘ALLOSTERIC’’:
RELATED MODELS
The term ‘‘allosteric’’ was first introduced by Monod and Jacob [6], who
referred to an allosteric inhibition (of the synthesis of a tryptophan
precursor by tryptophan) in describing the mechanism underlying the
action of ‘‘an inhibitor that was not a steric analog of the substrate.’’
Thus, first introduced in the field of enzymology, the term ‘‘allosteric’’
(Greek aEEo, other, different; jH eUeo, solid, shape) means ‘‘having a
different shape.’’ It soon referred to the presence in an enzyme of a
(secondary) site of attachment for a substance that modifies enzyme
activity without interacting directly with the active center (primary site).
The allosteric effect, therefore, was attributed to a change in either the
three-dimensional structure of the peptide chain or else a change in
224
Kourounakis et al.
conformation (allosteric transition) that affects the binding of the substrate to the active site [7].
Thus, the emphasis was shifted toward the crucial role of conformational changes of proteins, as elaborated by Monod et al. in 1965 [8]. The
concept extended from enzymology to ‘‘receptology,’’ first to ion channel
receptors and subsequently to GPCRs. The primary site on a receptor is
thus referred to as the (classical) ligand binding site or ‘‘orthosteric site,’’
while the secondary site, or allosteric site, affects binding at the primary
site by inducing a conformational change in the tertiary structure of the
receptor protein.
The simplest model that can describe allosteric interactions at
GPCRs is the ternary complex allosteric model [9]. As shown in
Figure 1, according to this model two parameters define the actions of
allosteric agent (X): its affinity for the unoccupied receptor (Kx) and its
cooperativity (a) with the ligand (A) that interacts at the primary binding
site: a<1 represents negative cooperativity; a=1, no cooperativity; a>1,
positive cooperativity.
However, based on the concept that GPCRs are able to adopt a
variety of conformations, an extended model can also be described, as
shown in Figure 2. In this extended ‘‘cubic ternary complex model’’ of
receptor activation and modulation, the receptor can interconvert between
an active (R*) and an inactive conformation (R), each with a different
Figure 1 Representation of the simple ternary allosteric complex model of
interaction of a ligand A with an allosteric agent X at a receptor R. (From Ref. 2.)
225
Figure 2 Representation of a ‘‘cubic ternary complex’’ model of allosteric
interaction: R, the inactive state of the receptor; R*, the active state of the
receptor; A, ligand; X, allosteric agent. (From Ref. 14.)
affinity for the G protein, ligand A, and allosteric modulator (X). Relative
stoichiometry of the states would be determined by the presence of G
protein and agonists and modified by allosteric modulators [10].
III. MUSCARINIC AND ADENOSINE RECEPTORS
Allosteric interactions on GPCRs have been observed for the muscarinic
[11–13], adenosine A1 [14], a2A-adrenergic [15–17], and dopamine D2
receptor [18]. This chapter focuses only on two allosteric phenomena, as
well as their potential for therapeutic exploitation: that on the muscarinic
receptor and that on the adenosine receptor.
A. Allosteric Modulation on the Muscarinic Receptor
The first and best-studied allosteric site on GPCRs is that on the muscarinic receptor [9,10,12,19,20]. For the five subtypes of these receptors that
have been cloned and pharmacologically defined as M1 to M5, various
agents have been identified that allosterically regulate selectively these
226
Kourounakis et al.
receptor subtypes [19,21]. Gallamine was the first compound shown to
interact allosterically with the muscarinic (M2) receptor, exhibiting negative cooperativity with antagonists [3H]NMS and [3H]QNB [22,23]. The
interactions of gallamine with the M2 receptor were shown to agree with
the ternary complex allosteric model in both binding and functional studies
[23–25]. Since then, a number of ligands have been discovered that interact
with various muscarinic receptor subtypes, confirming that the allosteric
site is present on all five subtypes.
Furthermore, these allosteric effects were shown to be truly subtype
specific, depending on the nature of the allosteric modulating compound.
Thus, alcuronium exerts positive copperativity with [3H]NMS at the M2
and M4 but not at the M1 and M3 receptors [26,27], while other neuromuscular junction blockers such as stercuronium, pancuronium, and dtubocurarine have been shown to exhibit their effects via an allosteric
mechanism specifically on the M2 receptors [28–30].
A growing number of other diverse compounds have also been
shown to bind to an allosteric site on the muscarinic receptors. Among
them are pirenzepine (highly selective for M1 receptor), lidocaine and
verapamil (ion channel blockers), tacrine (anticholinesterase compound), batrachotoxin, and strychnine (glycine receptor antagonist)
[25,31–35].
Although in the cases of gallamine and some of the other agents, a
values for various ligands were all below 1, another group of compounds,
such as brucine and analogues, appear to be allosteric agents exhibiting
positive cooperativity at one or more muscarinic receptor subtypes [36,37].
The interest in agents positively cooperative with ACh at muscarinic
receptors stems from their potential use in the treatment of cognitive
deficits such as Alzheimer’s disease. Brucine and analogues were shown not
only to enhance the affinity of ACh in radioligand binding studies for the
M1, M3, and M4 muscarinic receptors but further to modulate the actions
of acetylcholine in functional studies. First in GTPase and [35S]GTPgS
binding assays (Fig. 3), then in cAMP production and intracellular
Ca2+mobilization assays (Fig. 4), and finally in a tissue model of contraction of the guinea pig ileum strip (Fig. 5) [9]. In all cases, the activity
of these analogues showed ‘‘absolute subtype selectivity’’ with variable
effects on the various muscarinic subtypes. The results suggested the pharmacological feasibility of selectively elevating subnormally functioning
cholinergic neurons in the central nervous system (CNS) by means of an
appropriate allosteric enhancer.
227
Figure 3 Enhancement by brucine of ACh potency at M1 receptors in functional
assays in membranes. Brucine (100 AM) increased the potency of ACh to
stimulate [35S]GTPgS binding to G proteins in m1 CHO cell membranes. In this
experiment, the EC50 value for ACh decreased from 2.8 AM (5) to 0.9 AM (n)
without significantly affecting the basal response or maximal stimulation. (From
Ref. 9.)
B. Allosteric Modulation on the Adenosine Receptor
Extracellular adenosine is regarded as a local hormone that exerts numerous physiological actions in a variety of mammalian tissues. The actions of
this nucleoside in the body are mediated by G-protein-coupled adenosine
receptors subclassified as A1, A2A, A2B, and A3 [38]. The adenosine A1
receptor is interfaced with a Gi protein, which is negatively coupled to the
adenylate cyclase–cAMP signal transduction pathway, and thus, upon
activation, leads to a reduction in intracellular cAMP levels. This receptor
is highly and widely expressed in not only in the CNS but also in other
tissues such as fat cells, bladder, and heart [38–40]. A variety of adenosinemediated efffects (hypotension, inhibition of lipolysis, analgesia) occurs via
the adenosine A1 receptor, rendering it an important target for pharmacological intervention. Nonetheless, the wide distribution of adenosine
228
Kourounakis et al.
Figure 4 Dose–response curves for the potentiation by brucine of ACh wholecell M1 muscarinic receptor responses. (A) Brucine (104 M) enhanced the
potency of ACh to increase cAMP accumulation in M1 CHO cells by 2.6-fold. (B)
Brucine (100 AM) produced a 3.0-fold increase in ACh potency in Ca2+ response
to ACh. (From Ref. 9.)
229
Figure 5 (A) In a dose-dependent manner, N-chloromethyl brucine (CMB)
enhanced the field-stimulated contractions of isolated guinea pig ileum strips.
The contractions were inhibited by atropine (30 nM). (B) Histogram of the
percentage enhancement of contraction produced in four independent experiments of the type illustrated in A. (From Ref. 9.)
230
Kourounakis et al.
receptors offers both opportunities and drawbacks for therapeutic intervention [38,41]. For example, A1 adenosine agonists, through their interaction with adenosine A1 receptors on fat cells, are able to reduce free fatty
acid levels in the blood. Since this effect sensitizes insulin’s action [42]. It
may be a very useful feature in non-insulin-dependent diabetes mellitus
(type II diabetes). However, serious side effects occur by the concomitant
bradycardia and drop in mean arterial pressure due to interference with
cardiovascular adenosine receptors [43]. Various strategies have been followed to circumvent all or some of these problems, such as the development of partial agonists for that purpose [44–48]. It was shown that some
of these compounds were virtually ‘‘silent’’ on the heart, while keeping a
pronounced, full effect on adipose tissue [49].
On the other hand, among the effects of receptor-bound adenosine
is the ability to protect organs, including the heart and brain, from
ischemic injury [50–52]. The formation of extracellular adenosine as a
breakdown product of ATP is a local phenomenon, induced by a tissue
at risk (e.g., under hypoxic or anoxic conditions: heart failure, stroke,
etc.). As a consequence, compounds that would increase adenosine’s
concentration, and thus its tissue-protective effect, might have a better
therapeutic profile than the agonists described earlier. Marketed nucleoside transport blockers such as dipyridamole and dilazep have already
proven this concept by inhibiting the intracellular uptake of extracellular
adenosine, and thereby effectively increasing its concentration outside
the cell [53,54].
Another interesting approach is to enhance adenosine’s action
locally by means of an allosteric enhancer. In 1990, Bruns and coworkers reported on various 2-amino-3-benzoylthiophene derivatives
capable of enhancing the binding and activity of reference A1 receptor
agonists, such as N6-cyclopentyladenosine (CPA) [14,55]. One of these
‘‘allosteric modulators,’’ PD81,723, or (2-amino-4,5-dimethyl-trienyl)
[3-(trifluoromethyl) phenyl]methanone (Fig. 6), has been investigated
pharmacologically in greater detail by various independent research
groups [56–61].
The modulator PD81,723 enhances two- to threefold the binding
and function of agonists such as CPA, R-PIA, or NECA to adenosine
A1 receptors [62]. As shown in Figure 7, in displacement experiments of
[3H]DPCPX from the human adenosine A1 receptor (wild type), the binding curve of CPA in the presence of PD81,723 is shifted leftward; it seems
that CPA binds more efficiently, since lower concentrations of this agonist are needed to displace the same concentration of radioligand. This
231
Figure 6 Structure of PD81,723, (2-amino-4,5-dimethyl-trienyl)[3-(trifluoromethyl) phenyl]methanone and adenosine A1 agonists/antagonists.
Figure 7 Displacement of 0.4 nM [3H]DPCPX by various concentrations of CPA
from human wild-type (CHO A1) and mutant (CHO A1-mutT277A) adenosine A1
receptors in the absence (n) or presence (5) of PD81,723 (10 AM).
232
Kourounakis et al.
‘‘enhanced’’ activity of CPA is also maintained in second messenger
assays, where, for example, lower concentrations of CPA (in the presence of PD81,723) are needed for the inhibition of forskolin-stimulated
cAMP production in cells bearing adenosine A1 receptors (Fig. 8). It is
known that PD81,723 slows down the kinetics (dissociation) of 3H-labeled
agonists such as [3H]CHA or [3H]CCPA from the receptor as shown in
Figure 9; the half-life of 17 min for the dissociation of CCPA alone from
the rat A1 receptor is increased to 25 min in the presence of 10 AM
PD81,723 [63]. It is postulated that this compound binds to an allosteric
site on the adenosine A1 receptor—which, unlike the muscarinic one, is
not yet so well defined—while at somewhat higher concentration it binds
to the ligand binding site exhibiting antagonistic action. It is presumed
that via its allosteric activity PD81,723 increases the proportion of adenosine receptors in the ‘‘active’’ (R*) conformation that has a high affinity
for agonists and low for antagonists and inverse agonists (Fig. 2). Not
only are these effects selective for the A1 receptor, but they disappear
upon a mutation of the receptor at the proposed agonist binding site [62].
Threonine at position 277 on the A1 receptor is considered to interact with
ribose ring of agonists, since changing it to alanine greatly decreases the
affinity for agonists but not for antagonists. This mutation also eliminated
the activity of PD81,723 (Fig. 7), which no longer can increase the already
Figure 8 Forskolin-stimulated cAMP production of CHO A1 cells after addition
of CPA in the absence (n) or presence (5) of 10 AM PD81,723.
233
Figure 9 Dissociation of agonist [3H]CCPA from rat brain A1 receptors in the
presence (5) or absence (n) of 10 AM PD81,723.
low affinity of agonists [64–67]. This indicates that an intact agonist binding site of the receptor is required for PD81,723 to exert its allosteric
action [62].
Recently we developed a series of novel PD81,723 analogues,
some of which appear to be superior to PD81,723 in their enhancing
activity [68,69]. The synthesis of these derivatives is relatively straightforward, as shown in Figure 10 [68–72]. The 4,5-dimethyl group and
the benzoyl moiety were targets for further modifications, leading to
series of 4,5-dialkyl (1a–g), of tetrahydrobenzo (1h–u) and of tetrahydropyridine (3a–g) derivatives (Fig. 10, Tables 1 and 2). These derivatives were evaluated both as allosteric enhancers of agonist binding to
the rat adenosine A1 receptor and as antagonists on this receptor.
Among them, a number of compounds, in particular 1b, 2e, 1j, 1n,
and 1u (Fig. 11, Table 1), proved to be superior to the reference
compound (PD81,723) in both enhancing activity and diminished antagonistic behavior [68].
Some structure–activity relationships of a further developed R4, R5
alkyl/cycloalkyl series (2a–o, Fig. 10, Table 1) were also investigated.
This study [69] revealed structural features that favored allosteric
enhancing activity, such as benzoyl lipophilic substitution and thiophene
4-alkyl substitution, while other features, such as thiophene 5-bulky
substitution, favored antagonistic properties. Upon further analysis, a
234
Kourounakis et al.
Figure 10 Scheme of synthesis of PD81,723 analogues. Reagents and
conditions: i, DMF; S8, Et3N, RT (or EtOH, S8, Et2NH, 50jC); ii, C6H6, h-alanine,
HOAc; iii, EtOH, S8, Et2NH; iv, BzCl, CH2Cl2, Et3N.
Table 1 Structure and Enhancing/Antagonistic Activity of 2-Amino-3-benzoylthiophenes
Analogues 1a–u and 2a–o
Compound
PD81,723
1a
1b
1c
R0
R4
R5
Enhancement (%)a
Antagonism (%)b
3-CF3
CH3
CH3
100
39 (F4)
CH3
CH3
CH3
CH3
CH3
CH3
8 (F5)
80 (F19)
93 (F32)
H
3-Cl
4-Cl
14 (F3)
19 (F4)
41 (F6)
235
Table 1 Continued
Compound
PD81,723
1d
1e
1f
1g
1h
1i
1j
1k
1l
1m
1n
1o
1p
1q
1r
1s
1t
1u
2a
2b
2c
2d
2e
2f
2g
2h
2i
2j
2k
2l
2m
2n
2o
R0
R4
R5
Enhancement (%)a
Antagonism (%)b
3-CF3
CH3
CH3
100
39 (F4)
H
3-CF3
3-Cl
4-Cl
H
2-Cl
3-CF3
3-Cl
3-I
4-CF3
4-Cl
4-Br
4-I
4-NO2
4-CH3
4-CO2CH3
4-CO2H
3,4-Cl
3-CF3
3-Cl
H
3-CF3
3-Cl
3-Cl
H
3-Cl
3-CF3
H
H
H
3,4-Cl
4-tBu
4-tBu
CH3
CH2CH3
CH2CH3
CH3
CH2CH3
CH3
CH3
CH2CH3
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
—(CH2)4—
H
CH3CH2CH2
H
CH3CH2CH2
H
CH3CH2CH2
H
C5H9
H
C5H9
H
C6H11
H
C6H5
H
C6H5
H
C6H5
H
(CH3)2CHCH2
CH3
CH3CH2
CH3CH2
CH3CH2CH2
CH3
CH3
CH3
CH3
—(CH2)4—
31
112
30
97
47
73
122
93
113
131
123
128
155
34
137
44
29
151
88
67
0
99
52
57
21
38
42
7
13
69
116
125
137
(F4)
(F10)
(F7)
(F25)
(F4)
(F19)
(F19)
(F6)
(F18)
(F11)
(F15)
(F18)
(F21)
(F22)
(F21)
(F9)
(F3)
(F24)
(F8)
(F18)
(F30)
(F25)
(F12)
(F2)
(F5)
(F6)
(F7)
(F14)
(F17)
(F19)
(F7)
(F24)
(F10)
13
5
22
20
35
35
32
51
66
57
40
42
64
19
30
29
35
52
54
50
49
64
64
75
80
58
47
27
17
50
47
40
(F3)
(F11)
(F2)
(F12)
(F6)
(F3)
(F8)
(F5)
(F1)
(F4)
(F5)
(F4)
(F8)
(F2)
(F3)
(n=1)
nd
(F4)
(F3)
(F5)
(F7)
(F2)
(F1)
(F3)
(F2)
(F1)
(F3)
(F5)
(F7)
(F6)
(F1)
(F2)
(F4)
Enhancing activity (at 10 AM of test compound) is expressed as percentage of decrease (FSEM) in
[3H]CCPA dissociation over control (0%) and that of PD81,723 (100%, n = 3).
b
Antagonistic activity is expressed as percentage of displacement (FSEM) of 0.4 nM of [3H]DPCPX by 10
AM of test compound. nd: not determined.
a
236
Kourounakis et al.
Table 2 Structure and Enhancing/Antagonistic Activity of 2-Amino-3-benzoyl4,5,6,7-tetrahydrothieno [2,3-c]pyridines 3a–g and 4
Compound
3a
3b
3c
3d
3e
3f
3g
4
Theophylline
R0
R1
Enhancement (%)a
Antagonism (%)b
H
H
H
H
4-Cl
4-Cl
3,4-Cl
—
H
3-Cl
4-Cl
3,4-Cl
H
3,4-Cl
H
—
53 (F37)
106 (F27)
69 (F23)
57 (F36)
132 (F21)
106 (F31)
174 (F37)
14 (F27)
15 (F7)
67 (F5)
80 (F1)
52 (F2)
4 (F2)
60 (F0)
46 (F2)
51 (F0)
72 (F2)
56 (F5)
Enhancing activity is expressed as percentage of decrease (FSEM) in [3H]CCPA
dissociation over control (0%) and that of PD81,723 (100%, n = 3).
b
Antagonistic activity is expressed as percentage of displacement (FSEM) of 0.4 nM of
[3H]DPCPX by 10 AM of test compound.
a
significant correlation was found between antagonistic activity and hydrophobic fragment constants (k values) [73] of substituent R5 (Fig. 12),
in contrast to a negative correlation with those of R 4. Finally, comparison of low energy conformations (Fig. 13) of some of the 2-amino-3benzoylthiophene derivatives (PD81,723 and 2f ) with known adenosine
A1 receptor antagonists (theophylline and 8-cyclohexyltheophylline)
indicated that thiophene 5-substituents (R5 ) may interact with the same
lipophilic domain of the adenosine A1 receptor that accommodates 8substituents of xanthine antagonists. The separation of the two activities, antagonism and allosteric enhancement, is ultimately necessary for
the development of more potent and selective allosteric enhancers for the
adenosine A1 receptor with potential therapeutic applications.
237
Figure 11 Concentration–effect curves for derivative 1u and PD81,723.
Enhancement of 100% is expressing the maximum decrease in [3H]CCPA
dissociation by the highest concentration of 1u.
Figure 12 Correlation of lipophilicity parameter (k) for substituent R 5 of compounds 1a,d–f,h, j, k, and 2a–o with their antagonistic activity. ***p < 0.0001.
238
Kourounakis et al.
Figure 13 Structure and low-energy conformation with van der Waals surface of
(a) theophylline; (b) CHT; (c) PD81,723; and (d) 2f.
IV. CONCLUSION
The possibility of allosterically modulating receptors offers novel pharmacological means of ‘‘fine-tuning’’ receptor function. Further clarification is
required with respect to whether such modulated receptors are a general
feature of all or only of a subset of GPCRs and whether endogenous agents
regulate via this mechanism receptor function in vivo. Finally, elucidation
of the molecular mechanisms of the allosteric interactions will provide
useful insights for the therapeutic exploitation of this phenomenon in the
design and development of appropriate modulatory drugs.
Abbreviations
ACh
cAMP
[3H]CCPA
CHO
CH3CN
Acetylcholine
Cyclic-3V,5V-adenosine monophosphate
[3H]-2-Chloro-N6-cyclopentyladenosine
Chinese hamster ovary
Acetonitrile
CHT
CMB
CPA
[3H]DPCPX
CPT
DMF
Et3N
Et2NH
GPCR
HOAc
NECA
NMS
PD81,723
QNB
R-PIA
Theophylline
239
8-Cyclohexyltheophylline
N-Chloromethyl brucine
N6-Cyclopentyladenosine
[3H]-1,3-Dipropyl-8-cyclopentylxanthine
8-Cyclopentyltheophylline
N,N-Dimethylformamide
Triethylamine
Diethylamine
G-Protein-coupled receptor
Acetic acid
5V-(N-Ethyl)-carboxamidoadenosine
N-Methyl scopolamine
(2-Amino-4,5-dimethyl-3-thienyl)-[3(trifluoromethyl)
phenyl]methanone
Quinuclidinylbenzilate
N6-[-(R)-1-Methyl-2-phenylethyl]adenosine
1,3-Dimethylxanthine
ACKNOWLEDGMENTS
A. Kourounakis wishes to acknowledge support for this work from the
European Commission financed programs BIO4-CT97-5138 and QLK6CT-1999-51170.
REFERENCES
1. Costa E, Guigotti A. Benzodiazepines on trial: a research strategy for their
rehabilitation. Trends Pharmacol Sci 1996; 17:192–200.
2. Birdsall NJ, Cohen F, Lazareno S, Matsui H. Allosteric regulation of
G-protein-linked receptors. Biochem Soc Trans 1995; 23:108–111.
3. Massot O, Rousselle JC, Fillion MP, Grimaldi B, Cloez-Tayarani I, Fugelli A,
Prudhomme N, Seguin L, Rousseau B, Plantefol M, Hen R, Fillion G. 5Hydroxytryptamine-modulin, a new endogenous cerebral peptide, controls
the serotonergic activity via its specific interaction with 5-hydroxytryptamine
1B/1D receptors. Mol Pharmacol 1996; 50:752–762.
4. Rousselle J-C, Massot O, Delepierre M, Zifa E, Fillion G. Isolation and
characterization of an endogenous peptide from rat brain interacting
specifically with the serotonergic 1B receptor subtypes. J Biol Chem 1996;
271:551–557.
240
Kourounakis et al.
5. Grazzini E, Guillon G, Mouillac B, Zingg HH. Inhibition of oxytocin receptor function by direct binding of progesterone. Nature 1998; 392:509–512.
6. Monod J, Jacob F. General conclusions: teleonimic mechanisms in cellular
metabolism, growth and differentiation. Cold Spring Harbor Symp Quant
Biol 1961; 26:389–401.
7. Monod J, Changeux J-P, Jacob F. Allosteric proteins and cellular control
systems. J Mol Biol 1963; 6:308–329.
8. Monod J, Wyman J, Changeux J-P. On the nature of allosteric transitions:
a plausible model. J Mol Biol 1965; 12:88–118.
9. Birdsall NJ, Farries T, Gharagozloo P, Kobayashi S, Lazareno S, Sugimoto
M. Subtype-selective positive cooperative interactions between brucine
analogs and acetylcholine at muscarinic receptors: functional studies. Mol
Pharmacol 1999; 55:778–786.
10. Christopoulos A, Lanzafame A, Michelson E. Allosteric interactions at
muscarinic cholinoceptors. Clin Exp Pharmacol Physiol 1998; 25:185–194.
11. Waelbroek M. Identification of drugs competing with d-tubocurarine for an
allosteric site on cardiac muscarinic receptors. Mol Pharmacol 1994; 46:
685–692.
12. Tucek S, Proska J. Allosteric modulation of muscarinic acetylcholine
receptors. Trends Pharmacol Sci 1995; 16:205–212.
13. Birdsall NJ, Farries T, Gharagozloo P, Kobayashi S, Kuonen D, Lazareno
S, Popham A, Sugimoto M. Selective allosteric enhancement of the binding
and actions of acetylcholine at muscarinic receptor subtypes. Life Sci 1997;
60:1047–1052.
14. Bruns RF, Fergus JH. Allosteric enhancement of adenosine A1 receptor
binding and function by 2-amino-3-benzoylthiophenes. Mol Pharmacol 1990;
38:939–949.
15. Howard MJ, Hughes RJ, Motulsky HJ, Mullen MD, Insel PA.
Interactions of amiloride with alpha- and beta-adrenergic receptors:
amiloride reveals an allosteric site on alpha 2-adrenergic receptors. Mol
Pharmacol 1987; 32:53–58.
16. Wilson AL, Womble SW, Prakash C, Cragoe EJ Jr, Blair IA, Limbird LE.
Novel amiloride analog allosterically modulates the alpha 2-adrenergic
receptor but does not inhibit Na+/H exchange. Mol Pharmacol 1992;
42:175–179.
17. Leppik RA, Lazareno S, Mynett A, Birdsall NJM. Characterisation of the
allosteric interactions between antagonists and amiloride analogues at the
human alpha2A adrenergic receptor. Mol Pharmacol 1998; 53:916–925.
18. Hoare SR, Strange PG. Regulation of D2 dopamine receptors by amiloride
and amiloride analogs. Mol Pharmacol 1996; 50:1295–1308.
19. Ellis J. Allosteric binding sites on muscarinic receptors. Drug Dev Res 1997;
40:193–204.
20. Birdsall NJM, Lazareno S, Matsui H. Allosteric regulation of muscarinic
receptors. Prog Brain Res 1996; 109:142–151.
241
21. Lee NH, el-Fakahany EE. Allosteric interactions at the m1, m2, and m3
muscarinic receptor subtypes. J Pharmacol Exp Ther 1991; 256:468–479.
22. Clark AL, Mitchelson F. The inhibitory effect of gallamine on muscarinic
receptors. Br J Pharmacol 1976; 58:323–331.
23. Stockton JM, Birdsall NJ, Burgen AS, Hulme EC. Modification of the
binding properties of muscarinic receptors by gallamine. Mol Pharmacol
1983; 23:551–557.
24. Ehlert FJ. Gallamine allosterically antagonizes muscarinic receptor–
mediated inhibition of adenylate cyclase activity in the rat myocardium.
25. Lazareno S, Birdsall NJ. Detection, quantitation, and verification of
allosteric interactions of agents with labeled and unlabeled ligands at G
protein-coupled receptors: interactions of strychnine and acetylcholine at
muscarinic receptors. Mol Pharmacol 1995; 48:362–378.
26. Tucek S, Musilkova J, Neduoma J, Proska J, Shelkovnikov SW, Vorlicek J.
Positive cooperativity in the binding of alcuronium and N-methylscopolamine
to muscarinic acetylcholine receptors. Mol Pharmacol 1990; 38:674–680.
27. Jakubik J, Bacakova L, El-Fakahany EE, Tucek S. Subtype selectivity of the
positive allosteric action of alcuronium at cloned M1–M5 muscarinic
acetylcholine receptors. J Pharmacol Exp Ther 1995; 274:1077–1083.
28. Dunlap J, Brown JH. Heterogeneity of binding sites on cardiac muscarinic
receptors induced by the neuromuscular blocking agents gallamine and
pancuronium. Mol Pharmacol 1983; 24:15–22.
29. Li CK, Mitchelson F. The effect of stercuronium on cardiac muscarinic
receptors. Eur J Pharmacol 1978; 51:251–259.
30. Waelbroeck M, Robberecht P, De Neef P, Christphe J. Effects of dtubocurarine on rat cardiac muscarinic receptors: a comparison with gallamine. J Recept Res 1988; 8:787–808.
31. Roeske WR, Venter JC. The differential loss of [3H]pirenzepine vs [3H] (–)quinuclidinylbenzilate binding to soluble rat brain muscarinic receptors
indicates that pirenzepine binds to an allosteric state of the muscarinic
receptor. Biochem Biophys Res Commun 1984; 118:950–957.
32. Ellis J, Huyler J, Brann MR. Allosteric regulation of cloned m1–m5
muscarinic receptor subtypes. Biochem Pharmacol 1991; 42:1927–1932.
33. Waelbroeck M, Robberecht P, De Neef P, Christophe J. Effects of verapamil
on the binding properties of rat heart muscarinic receptors: evidence for an
allosteric site. Biochem Biophys Res Commun 1984; 121:340–345.
34. Cohen-Armon M, Henis YI, Kloog Y, Sokolovsky M. Interactions of
quinidine and lidocaine with rat brain and heart muscarinic receptors.
Biochem Biophys Res Commun 1985; 127:326–332.
35. Cohen-Armon M, Kloog Y, Henis YI, Sokolovsky M. Batrachotoxin
changes the properties of the muscarinic receptor in rat brain and heart:
possible interaction(s) between muscarinic receptors and sodium channels.
242
Kourounakis et al.
36. Lazareno S, Gharagozloo P, Kuonen D, Popham A, Birdsall NJ. Subtypeselective positive cooperative interactions between brucine analogs and
acetylcholine at muscarinic receptors: radioligand binding studies. Mol
Pharmacol 1998; 53:573–589.
37. Gharagozloo P, Lazareno S, Popham A, Birdsall N. Allosteric interactions
of quaternary strychnine and brucine derivatives with muscarinic acetylcholine receptors. J Med Chem 1999; 42:438–445.
38. DeNinno MP. Adenosine. In: Doherty A, ed. Annual Reports in Medicinal
Chemistry. Vol. 33. San Diego, CA: Academic Press, 1998:111–120.
39. Jacobson KA, van Galen PJM, Williams M. Adenosine receptors:
pharmacology, structure–activity relationships, and therapeutic potential.
J Med Chem 1992; 35:407–422.
40. Jacobson KA, van Rhee AM. Development of selective purinoceptor agonists
and antagonists. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in
Experimental Therapeutics. New York: Wiley-Liss, 1997:101–128.
41. Stiles GL. Adenosine receptor subtypes: new insights from cloning and
functional studies. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches
in Experimental Therapeutics. New York: Wiley-Liss, 1997:29–37.
42. Merkel LA, Hawkins ED, Colussi DJ, Greenland BD, Smits GJ, Perrone
MH, Cox BF. Cardiovascular and antilipolytic effects of GR79236.
Pharmacology 1995; 51:224–236.
43. Mathoˆt RAA, van Schaick EA, Langemeijer MWE, Soudijn W, Breimer
DD, Ijzerman AP, Danhof M. Pharmacokinetic–pharmacodynamic relationship of the cardiovascular effects of the adenosine A1 receptor agonist
N6-cyclopentyladenosine (CPA) in the rat. J Pharmacol Exp Ther 1994;
268:616–624.
44. van der Wenden EM, von Frijtag Drabbe Ku¨nzel JK, Mathoˆt RAA,
Danhof M, Ijzerman AP, Soudijn W. Ribose-modified adenosine analogues
as potential partial agonists for the adenosine receptor. J Med Chem 1995;
38:4000–4006.
45. Mathoˆt RAA, van der Wenden EM, Soudijn W, Ijzerman AP, Danhof M.
Deoxyribose analogues of N6-cyclopentyladenosine (CPA) are partial
agonist at the adenosine A1 receptor in vivo. Br J Pharmacol 1995;
116:1957–1964.
46. Roelen H, Veldman N, Spek AL, von Frijtag Drabbe Ku¨nzel JK, Mathoˆt
RAA, Ijzerman AP. N6,C8-disubstituted adenosine derivatives as partial
agonist for adenosine A1 receptors. J Med Chem 1996; 39:1463–1471.
47. Lorenzen A, Sebastiao AM, Sellink A, Vogt H, Schwabe U, Ribeiro JA,
Ijzerman AP. Biological activities of N6,C8-disubstituted adenosine derivatives as partial agonist at rat brain adenosine A1 receptors. Eur J
Pharmacol 1997; 334:299–307.
48. Van der Wenden EM, Carnielli M, Roelan HCPF, Lorenzen A, Von Frijtag
Drabbe Ku¨nzel JK, Ijzerman AP. 5V-Substituted adenosine analogs as new
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
243
high affinity partial agonist for the adenosine A1 receptor. J Med Chem
1998; 41:102–108.
Van Schaick EA, Tukker HE, Roelan HCPF, Ijzerman AP, Danhof M.
Selectivity of action of 8-alkylamino analogues of N6-cyclopentyladenosine
in vivo: hemodynamic versus anti-lipolytic responses in rats. Br J Pharmacol
1998; 124:607–618.
Liu G-S, Downe JM, Cohen MV. Adenosine, ischemia and preconditioning.
In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental
Therapeutics. New York: Wiley-Liss, 1997:153–172.
Fredholm BB. Adenosine and neuroprotection. Int Rev Neurobiol 1997;
40:259–280.
Knutsen LJS, Murray TF. Adenosine and ATP in epilepsy. In: Jacobson
KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics.
New York: Wiley-Liss, 1997:423–447.
Ijzerman AP, van der Wenden EM. Modulators of adenosine uptake,
release and inactivation. In: Jacobson KA, Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New York: Wiley-Liss, 1997:129–
148.
Geiger JD, Parkinson FE, Kowaluk EA. Regulators of endogenous
adenosine levels as therapeutic agents. In: Jacobson KA, Jarvis MF, eds.
Purinergic Approaches in Experimental Therapeutics. New York; WileyLiss, 1997:55–84.
Bruns RF, Fergus JH, Coughenour LL, Courtland GG, Pugsley TA, Dodd
JH, Tinney JF. Structure–activity relationships for enhancement of
adenosine A1 receptor binding by 2-amino-3-benzoylthiophenes. Mol
Pharmacol 1990; 38:950–958.
Linden J. Allosteric enhancement of adenosine receptors. In: Jacobson KA,
Jarvis MF, eds. Purinergic Approaches in Experimental Therapeutics. New
York: Wiley-Liss, 1997:85–97.
Kollias-Baker C, Rubl J, Dennis D, Bruns RF, Linden J, Belardinelli L.
Allosteric enhancer PD 81,723 acts by novel mechanism to potentiate
cardiac actions of adenosine. Circ Res 1994; 75:961–971.
Musser B, Mudumbl RV, Liu J, Olson RD, Vestal RE. Adenosine A1
receptor-dependent and -independent effects of the allosteric enhancer PD
81,723. J Pharmacol Exp Ther 1999; 288:446–454.
Mizumura T, Auchampach JA, Linden J, Bruns RF, Gross GJ. PD 81,723,
an allosteric enhancer of the A1 adenosine receptor, lowers the threshold for
ischemic preconditioning in dogs. Circ Res 1996; 79:415–423.
Bhattacharya S, Linden J. The allosteric enhancer, PD 81,723, stabilizes
human A1 adenosine receptor coupling to G proteins. Biochim Biophys
Acta 1995; 1265:15–21.
Leung E, Walsh LKM, Flippin LA, Kim EJ, Lazar DA, Seam CS, Wong
EHF, Eglen RM. Enhancement of adenosine A1 receptor functions by
244
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
Kourounakis et al.
benzoylthiophenes in guinea pig tissues in vitro. Naunyn-Schmiedeberg’s
Arch Pharmacol 1995; 352:206–212.
Kourounakis A, Visser C, de Groote M, Ijzerman AP. Differential effects of
the allosteric enhancer PD81,723 on agonist and antagonist binding and
function at the human wild type and a mutant (T277A) adenosine A1
receptor. Biochem Pharmacol 2000; 61:137–144.
Kourounakis A, Visser C, de Groote M, Ijzerman AP. Allosteric modulation of the rat adenosine A1 receptor: differential effects on agonist and
antagonist binding. Biochem Pharmacol 2001; 51:207–215.
Townsend-Nicholson A, Schofield PR. A threonine residue in the seventh
transmembrane domain of the human A1 adenosine receptor mediates
specific agonist binding. J Biol Chem 1994; 269:2373–2376.
Tucker AL, Robeva AS, Taylor HE, Holeton D, Bockner M, Lynch KR,
Linden J. A1 adenosine receptors: two amino acids are responsible for
species differences in ligand recognition. J Biol Chem 1994; 269:27900–
27906.
Rivkees SA, Lasbury ME, Barbahalya H. Identification of domains of the
human A1 adenosine receptor that are important for binding receptor
subtype selective ligands using chimeric A1/A2 adenosine receptors. J Biol
Chem 1995; 270:20485–20490.
Dalpiaz A, Townsend-Nicholson A, Beukers MW, Schofield P, Ijzerman
AP. Thermodynamics of full agonist, partial agonist and antagonist binding
to wild type and mutant adenosine A1 receptors. Biochem Pharmacol 1998;
56:1437–1445.
Van der Klein PAM, Kourounakis AP, Ijzerman AP. Allosteric modulation
of the adenosine A1 receptor. Synthesis and biological evaluation of novel
2-amino-3-benzoylthiophenes as allosteric enhancers of agonist binding.
J Med Chem 1999; 42:3629–3635.
Kourounakis AP, Van der Klein PAM, Ijzerman AP. Elucidation of
structure–activity relationships of 2-amino-3-benzoylthiophenes: Study of
their allosteric enhancing versus antagonistic activity on adenosine A1
receptors. Drug Dev Res 2000; 49:227–237.
Gewald K, Schinke E, Bo¨ttcher H. 2-Amino-thiophene aus methylenaktiven
Nitrilen, Carbonylverbindungen und Scwefel. Chem Ber 1966; 99:94–100.
Jones G. The Knoevenagel Condensation. Vol. 15. Organic Reactions. New
York: Wiley, 1967:204–599.
Tinney FJ, Sanchez JP, Nogas JA. Synthesis and pharmacological
evaluation of 2,3-dihydro-1H-thieno[2,3-e][1,4]diazepines. J Med Chem
1974; 17:624–630.
Hansch C, Leo A. Substituent Constants for Correlation Analysis in
Chemistry and Biology. New York: Wiley, 1979.
9
Protein Misfolding and
Neurodegenerative Disease:
Therapeutic Opportunities
Harry LeVine III
University of Kentucky, Lexington, Kentucky, U.S.A.
I.
DISEASES WITH PROTEIN MISFOLDING
Alzheimer’s disease, a typically late-life dementia and the most prevalent
chronic neurodegenerative disease, is pathologically characterized by the
presence of insoluble h-amyloid peptide in extracellular senile and diffuse
plaques and intracellular accumulation of neurofibrillary tangles (NFTs)
of hyperphosphorylated tau protein as parahelical filaments. These
insoluble protein assemblies are derived from normal cellular proteins
that have deposited in entities that had been recognized histologically for
some 70 years before their main protein constituents were determined. A
potential unifying theme has emerged in the pathology of a number of
chronic neurodegenerative diseases. Improved immunological and microanalytical methods have led to the identification of the constituents of
other proteinaceous deposits associated with neurodegenerative disease.
These deposits are intracellular, however, unlike those of the amyloid-beta
(A h) peptides. A series of genetically dominant trinucleotide repeat
diseases coding for glutamine in which the CAG repeats are expanded
in a different protein in each disease was observed to develop insoluble
polyglutamine-containing inclusions, concentrated in different brain areas
245
246
LeVine
and undergoing cell loss depending on the protein and disease involved [1–
5] (see Table 1). The prototype is Huntington’s disease (HD) which affects
the largest number of people. Polyglutamine stretches can form particularly stable h-sheet structures, which are prone to aggregation [6–9]. First
described in transgenic mice overexpressing exon I of Huntingtin, which
contains the polyglutamine repeat [10], nuclear inclusions of Huntingtin
form in susceptible regions of the brain. The same ubiquitinated inclusions
are also found in human HD tissue when appropriate antibodies are used
[11–14]. Although the correlation of deposits with the synaptic and cell
loss of the disease pathology is imperfect, similar to the situation in
Alzheimer’s disease, their appearance is consistent with a pathway of
protein folding and translocation that leads to cell loss. Similar observations have been made and conclusions reached for types 3 and 7
spinocerebellar ataxia (SCA-3, SCA-7) and dentatorubral–palladoluysian
atrophy (DRPLA), polyglutamine repeat diseases with cerebellar pathology [15–19]. A wide variety of insoluble proteins are associated with
chronic neurodegenerative diseases (Table 2). Familial tauopathies, collectively referred to as FTDP17, are ascribed to mutations in various tau
exonic or intronic sequences that alter mRNA isoform expression, resulting in insoluble fibrillar deposits of the microtubule-associated protein tau
on human chromosome 17. Other tauopathies have been identified,
varying with respect to the brain region affected and the ratio of the
different tau gene splice products deposited [20]. Progressive supranuclear
palsy and Pick’s disease are classic late-onset tau deposition diseases
[21,22]. Tau is a conformationally ambiguous protein that does not adopt
a defined structure in solution [23]. Hyperphosphorylation of tau favors
conformational changes leading to rapid intermolecular h-sheet formation. This inhibits the microtubule-polymerizing activity of this microtubule-associated protein and leads to NFT formation [24]. The
phosphorylation of a specific sequence on tau containing T231 facilitates
binding and depletion of a prolyl isomerase, Pin1, effecting its nuclear
function [25].
Prion diseases resulting in encephalopathy can be transmitted
between individuals within species (more rarely between species) [26–28].
A conformational variant of the normal cellular protein PrPS (PrPC)
(protease-sensitive or cellular) is believed to catalyze [29] or nucleate [30–
33] conversion to the pathological form, PrPR (protease-resistant). This
highly unusual nongenetic mode of transmission of an infectious agent has
been strongly debated [29]. The observation of multiple examples of
nucleated catalysis of aberrant polymerization of protein subunits has
Sites of pathology
a
40–81
36–64
61–84
21–30 VDCCa1Asubunit
37–130 Ataxin-7
40–62
49–88
6–39
15–29
13–42
4–18
7–17
11–34
7–35
Atrophin-1
Androgen receptor
Ataxin-3
Ataxin-2
Ataxin-1
36–121 Huntingtin
Protein
11–34
Normal Disease
NI (n)
NI (n)
NI (memb)
NI (n)
NI (c)
NI (c)
NI (n, c)
NI (c)
Location of
disease (normal)a
NI, nuclear inclusions; (c), normal cytoplasmic localization; (n), normal nuclear localization; (membr), normal membrane localization.
Huntington’s
Striatum (medium, spiny)
disease
Spinocerebellar
Cerebellar cortex (Purkinje cells),
ataxia (SCA), type 1
brain stem
SCA2
Cerebellum, pontine nucleus, substantia
nigra
SCA3 (Machado–
Substantia nigra, globus pallidus,
Joseph disease)
pontine nucleus, cerebellar cortex
SCA6
Cerebellar and mild brain stem atrophy
SCA7
Photoreceptors and bipolar cells,
cerebellar cortex, brain stem
Motor neurons, dorsal root ganglia
Spinal and bulbar
muscular atrophy
(SBMA)
Globus pallidus, dentatorubral and
Dentatorubralsubthalamic nuclei
palladoluysian
atrophy
(DRPLA)
Disease
Repeat number
Table 1 Polyglutamine (CAG Repeat) Neurodegenerative Diseases
Protein Misfolding
247
Paired helical formation, NFT
Lewy bodies and neurites
Neuronal cytoplasm
4R tau
3R tau NF
a-Synuclein
a-Synuclein
a-Synuclein
a-Synuclein
SOD1 mutants
Prion protein
Frontotemporal regions
Cerebrocortical regions,
substantia nigra
Substantia nigra, brain
nuclei
Cerebellum, striatal regions
Brain stem, spinal cord
Brain stem, spinal cord
Neuronal cytoplasm
Extracellular deposits
NFT
4R tau NF
4R tau
Frontotemporal regions,
brain stem, spinal cord
Diffuse and senile plaques,
paired helical formation, NFT
NFT in oligodendroglia and
neurons
NFT in astrocytes and neurons
h Peptide; 4R, 3R tau
Multiple=system tauopathy
(familial)
Progressive supranuclear
palsy (PSP)
Corticobasal degeneration
(CBD)
Pick’s disease
Diffuse Lewy body
disease (DLB)
Parkinson’s
disease
Multiple-system
atrophy (MSA)
Amylotrophic
lateral sclerosis (ALS)
Familial ALS
Creutzfeldt–Jakob
disease (CJD)
New variant CJD
Gerstmann–Straussler–
Scheinker disease
Fatal familial insomnia
Kuru
Deposit
Major protein
Neocortex, hippocampus
Sites of pathology
Alzheimer’s disease
Disease
Table 2 Neurodegenerative Diseases with Insoluble Deposits
248
LeVine
Protein Misfolding
249
markedly decreased the heretical flavor of such concepts. Transmissible
protein aggregation has also been observed with [URE3], the prion form
of Ure2p, a nonchromosomal genetic element regulating nitrogen catabolism, and with [PS1], the prion form of Sup35p in Saccharomyces
cerevisiae [31,34].
The specific etiology of prion diseases in mammalian systems depends on the modified form of the protein [35], of which different variants
display distinguishable conformations [36,37]. Most forms lead to a
spongiform encephalopathy with marked neuronal cell loss in regions that
accumulate the pathogenic protease-resistant conformer of the protein.
Some of the more virulent forms of the protein expressed in Creutzfeldt–
Jakob disease (CJD) are accompanied by classical intracellular amyloid
plaques of PrPR. Although primarily recognized as a rare animal disease
(scrapie), its appearance in the English beef herd in the 1990s and its
potential for transmission to humans after a long latency caused a flurry of
interest in detection and treatment countermeasures [38].
a-Synuclein, a synaptic protein, is deposited in Lewy bodies and
Lewy neurites in Parkinson disease [39–41], in diffuse Lewy body disease
[42], and in the Lewy body variant of Alzheimer’s disease [43]. Multiplesystem atrophy is characterized by intracellular neuronal and glial asynuclein inclusions [44]. The role for a-synuclein in these diseases was
supported by the discovery of mutant forms of a-synuclein, A53T and
A30P, in familial early-onset Parkinson’s disease [40]. Like tau, a-synuclein is a conformationally ambiguous protein with little stable secondary
structure in solution [45]. a-Synuclein, but not the related h- or g-synuclein,
can polymerize in a nucleation-dependent fashion [46–48].
Lou Gehrig’s disease (amyotrophic lateral sclerosis: ALS) displays
motor neuron deposits of hyperphosphorylated neurofilament subunits in
the sporadic disease. Familial ALS, some 20% of all cases of ALS, involves
dominant superoxide dismutase SOD1 mutants that can form h-barrel
aggregates [49–51].
To this list of protein misfolding diseases can be added rare familial
amyloidoses in which the mutated proteins have the classic amyloid fibril
congophilic birefringence and cross-h-sheet structure (Table 3). Many of
these deposits have an impact on the central nervous system (TTR,
cystatin, lysozyme) as well as on other organ systems. A newly described
disease, familial British dementia, is associated with the deposition of Abri,
a 34 amino acid, 4 kDa peptide cleaved from a 277 amino acid precursor
sequence, the last 10 amino acids of which are not normally translated [52].
Familial encephalopathy with neuroserpin inclusion bodies (FENIB) is
250
LeVine
Table 3 Amyloidoses Recognized by the WHO International Nomenclature
Committee on Amyloidosis
Precursor protein
Immunoglobulin light chain
Immunoglobulin heavy chain
Apo-serum amyloid A protein
Transthyretin
h2-Microglobulin
Apolipoprotein AI
Gelsolin
Lysozyme
Fibrinogen a chain
Cystatin C
h-Amyloid precursor
protein
Prion protein
Procalcitonin
Islet amyloid polypeptide
Atrial natriuretic factor
Prolactin
Insulin
Lactoferrina
a
Associated disorder
Plasma cell disorders
Plasma cell disorders
Inflammation-associated, familial
Mediterranean fever
Familial amyloidotic neuropathy,
systemic senile amyloidosis
Dialysis-associated amyloidosis
Familial amyloidotic neuropathy,
aortic amyloidosis
Familial systemic amyloidosis
Familial cerebral hemorrhage
with amyloidosis
Sporadic and familial
Alzheimer’s disease,
familial cerebral hemorrhage
with amyloidosis
Spongiform encephalopathies
C-cell thyroid tumors
Insulinoma, type II diabetes
Atrial amyloidosis
Prolactinomas; pituitary
amyloidosis
Iatrogenic amyloidosis
Corneal amyloidosisa
Preliminary, awaiting confirmation by WHO International Nomenclature Committee
on Amyloidosis. The term amyloidosis is reserved by the committee specifically for
extracellular protein deposits.
another rare hereditary dementing disorder resulting from point mutations
in the neuroserpin gene [53]. FENIB is marked by unique neuronal
inclusion bodies consisting primarily of abnormal aggregated neuroserpin
filaments formed by a mechanism similar to that found in other familial
diseases of serpin conformation, including emphysema and cirrhosis due to
mutant a1-antitrypsin or thromboembolytic disease in antithrombin
mutants [54].
Protein Misfolding
251
The mechanisms of cell loss in these various diseases of protein
deposition may differ in detail, but the association of insoluble protein
inclusions with the pathology suggests that interventions preventing
protein misfolding and deposition may be of therapeutic utility. Alternatively, stabilization of toxic species must be avoided, since the insoluble
form of the protein is one potential strategy for reducing the exposure to
toxic, soluble forms. Approaches similar to those applied to blocking Ah
fibril formation in Alzheimer’s disease may prove fruitful with these other
proteins, possibly extending to some of the same compounds being
developed for AD. There are examples of this for prions [55–58] and for
transthyretin [59].
The purpose of this chapter is to conceptualize the shared molecular
features of protein misfolding in neurodegenerative diseases. By stressing
the commonalities, rational strategies can be devised to target similar
pathways that lead to cellular degeneration and eventually to clinical
symptoms in these diseases. This is one way to maximize the effects of
progress made for the pharmaceutically attractive (relatively large patient
base) neurodegenerative diseases such as Alzheimer’s and Parkinson’s for
application to other serious but less prevalent neurodegenerative diseases.
Such ‘‘piggyback’’ strategies may be a starting point for therapeutics that
already have the appropriate bioavailability, brain penetration, and longterm safety profile required for these applications. Similarly, nonneural
amyloid diseases and diseases with significant amyloid components such as
type II diabetes could also be approached.
II. MECHANISMS OF PROTEIN POLYMERIZATION
Protein homopolymerization is a well-studied process by which cellular
structure is dynamically regulated in response to the environment and
cellular metabolism. Actin and tubulin exist as nucleotide-dependent
(ATP and GTP, respectively) polymers (microfilaments and microtubules) that rapidly elongate and shorten in a reversible manner
regulated by binding proteins that can catalyze either polymerization
or filament shearing. Mathematical analysis of the physical chemistry
of the polymerization of these systems has defined the nucleation and
elongation processes and provided the theoretical basis for models
describing fibril assembly [60–63]. Nature has also provided evidence
that small molecules, such as plant alkaloids and fungal secondary
metabolites, are capable of modulating protein–protein interactions
252
LeVine
(actin depolymerization: cytochalasin, podophylotoxin; tubulin depolymerization: vinca alkaloids; microtubule stabilization: taxol). Another
extremely physiologically important protein polymerization reaction that
has been studied quantitatively is the process of thrombin-catalyzed
fibrinogen fragmentation and assembly of fibrillar fibrin during the
clotting of blood [64–66]. The larger size of the monomeric protein units
in these polymers has simplified detection of the various intermediates
in the assembly process. Atomic level structural resolution of these relatively large proteins has been aided by use of the modulators of
polymerization [67–69].
Protein polymerization can also lead to pathological consequences.
In contrast to physiologically normal assemblies, the pathological polymers are usually poorly reversible or degradable and tend to accumulate
until they cause problems for the surrounding cells or tissue. The polymerization of mutant hemoglobin S inside the red blood cells of individuals
afflicted with sickle cell anemia occurs rapidly and is modulated by the
hemoglobin ligand 2,3-diphosphoglycerate. The mutation decreases the
stability of the deoxygenated form of the protein, leading to exposure of
hydrophobic surfaces and an increased propensity to aggregate. A model
envisioning heterogeneous nucleation along the sides of the polymer and
branching reactions in addition to the standard homogenous nucleation
observed at the ends of growing fibrils was first described for the sickle cell
hemoglobin system [70,71]. The effectiveness of hydroxyurea treatments in
reducing the severity of the sickle cell crisis is ascribed to stabilizing effects
on the mutant hemoglobin conformation [72].
Several pathological self-polymerizing systems have been biophysically characterized sufficiently to permit identification of protein or peptide
species that could serve as molecular targets in a structure–activity
relationship. These include transthyretin (TTR) [73–76], serum amyloid
A protein (SAA) [77], microtubule-associated protein tau [78–80], amylin
or islet amyloid polypeptide (IAPP) [81,82], IgG light chain amyloidosis
(AL) [83–85], polyglutamine diseases [9,86], a-synuclein [47,48] and the
Alzheimer’s h peptide [87–96]. A variety of Ah peptide assay systems
have been established at Parke-Davis to search for inhibitors of fibril
formation that could be therapeutically useful [97].
In the search for fibril formation inhibitors, the self-association to
form amyloid fibrils of the Ah peptides containing 40 and 42 amino
acids can be treated as a coupled protein folding and polymerization
process passing through multiple intermediate peptide species. The in
vitro challenge is (1) to identify the various conformational forms and
Protein Misfolding
253
multimeric species involved, (2) to establish their order and arrangement in
parallel and/or in sequential pathways in a reaction scheme, and (3) to
design assay conditions under which only one intermediate is rate limiting
so that a reasonable structure–activity relationship can be determined. The
in vivo relevance of a particular mechanistic scheme will eventually be
assessed by the activity of bioavailable and brain-penetrant inhibitors of
the defined in vitro reactions in in vivo transgenic models of central
nervous system Ah amyloidosis.
III. INTERMEDIATES IN FIBRIL FORMATION
By analogy to the well-characterized polymerizing tubulin and actin
protein systems, and consistent with experiment, h-peptide aggregation
in the test tube is envisioned as nucleation event rate-limited by the
formation of a multimeric intermediate from the monomeric random coil
peptide in solution. Similar qualitative kinetics hold for light chain (AL)
amyloid, IAPP, TTR, prion protein, tau, polyglutamine diseases, and asynuclein. Figure 1 suggests a schematic view of the process for the hpeptide separated into prenucleation, nucleation, and fibril growth phases.
The techniques used to characterize peptide species at the different stages
of the reaction are noted below the relevant intermediate in the figure. The
presence of the early intermediates has been inferred from the kinetics of
Figure 1 Stages of amyloid aggregation: steps in protein polymerization and the
techniques used to measure them for the Alzheimer’s h peptide.
254
LeVine
Ah fibril formation; an identifiable nucleating species has yet be isolated.
Direct observation has been made difficult by the small size of the h
peptide, which has an effective hydrodynamic radius of 4 nm [98–100], and
by the apparent low abundance of nucleating species due to the low
probability of their formation. Such species would be formally akin to
an enzyme transition state that is usually kinetically inferred or sometimes
trapped with certain kinds of inhibitor. In disaggregated, ultrafiltered
(20 nm pore size) preparations, less than 1% of the molar peptide concentration is inferred to be present as ‘‘seeds’’ or nuclei determined by the
kinetics of fibril formation [101].
There is, however, ‘‘hard’’ evidence for the involvement of transient
species in h-peptide fibril formation. Recent atomic force microscopy
(AFM) [93,94,102–104] and electron microscopic observations [105,106]
have characterized rope like species intermediate between nucleation and
fibril extension. Designated as protofibrils, these species appear to anneal
and to wind around each other. Such a model is consistent with oriented
x-ray fiber diffraction patterns of a triple or higher helix of h sheets
producing a h-helix quaternary fibril structure [107–113]. These protofibrils are on-pathway intermediates in amyloid fibril formation [93,94]
containing h-sheet structure. They are negative with respect to thioflavin
T and apparently toxic to cultured cells [105]. A stiffer, more compact
fibril species (type I) is eventually formed from the initial type II fibril,
10–20 nm high. The dominant fibril form observed is dependent on the
environmental conditions and the initial conformational state of the
peptide [103]. Recent AFM studies indicate that amylin fibrils grow
bidirectionally, from both ends at roughly equal rates [114]. Branched
fibrils and heterogeneous catalysis along the edge of the fibrils [70,71]
have also been observed [94].
The variety of structures of h-peptide species observed by electron
microscopy and by AFM suggest that different surfaces would be available
to bind inhibitors on each species; moreover, the ability of a given inhibitor
to block the fibrillization reaction should depend on the peptide species
present in a particular situation. The rate-limiting species in vivo is
unknown at present. It is also possible for more than one fibrillization
pathway to operate concurrently, depending on the in vitro and in vivo
reaction conditions. A host of molecules have been claimed to inhibit hpeptide amyloid fibril formation on the basis of a variety of assays for
activity. The diversity of structures is represented in Figure 2. Their efficacy
is in general low (IC50 tens of micromolar or higher), corresponding
roughly to the order of magnitude of the amount of peptide present in
Figure 2 Reported inhibitors of Ah aggregation: 1, nicotinamide [156]; 2, Anthranilates [59]; 3, N-alkyl-N-methylpiperidinium bromides
[157]; 4, benzothiazoles [U.S. patent 6,001,331]; 5, Congo Red [142]; 6, melatonin [158]; 7, PPI-558 (Praecis Pharmaceuticals, Inc. patent
WO 9628471); 8, anthracyclines (IDOX) [159]; 9, aza-anthracyclones (WO 9832754-A); 10, iminoaza-anthracyclinones (WO 9832754-A);
11, acridinones (U.S. patent 5,972,956); 12, naphthyl monoazo compounds [U.S. patent 5,955,472]; 13, porphyrins [160]; 14,
naphthalenes (Japanese patent 090954222, Teijin KK); 15, rifamycins [161]; 16, rifampicins [161]; 17, alkylsulfonates/sulfates [162].
Protein Misfolding
255
256
LeVine
most assays, implying a 1:1 compound-to-peptide stoichiometry. In most
cases, however, the stoichiometry for inhibition has not been determined.
A high concentration of Ah peptide is generally used to overcome the
unfavorable kinetics of multiple peptides interacting to form a nucleus
capable of supporting the addition of monomeric peptide. Such reactions
exhibit a lag phase until the nucleus is formed (Fig. 3, curve c). Inhibitors
can affect either the lag phase, the maximal extent of the reaction, or both
(Fig. 3, curve d). Unless both the nucleation and extension reactions are
monitored, inhibitors prolonging the lag phase are poorly distinguished
from those blocking extension from the nucleus, thus muddying any
structure–activity relationships. Quantitative treatment of the reaction
has been proposed to mathematically separate the nucleation and extension reactions [62,115]. Distinguishing true nucleation from various exponential growth mechanisms is actually quite difficult, requiring precise rate
Figure 3 Effect of seeding and inhibitors on aggregation reaction. The lag phase
(curve c) is characteristic of reactions in which formation of nuclei for
polymerization is an unfavorable process. Addition of preformed nuclei or
‘‘seeds’’ (curve a) abolishes the lag phase. Inhibitors may affect the formation of
nuclei and influence either the lag phase, the extension of the nuclei changing the
growth phase, or both (curve d). The inhibitor example (curve d) acts more
strongly at nuclei formation than on the slope or plateau level of the growth phase.
Protein Misfolding
257
determinations of the first 10% of the reaction over a range of reactant
concentrations [116]. Such measurements have not been achieved with the
h peptide.
An extension reaction from a nucleus (Fig. 3, curve a) is pseudo–first
order in peptide concentration and thus more easily analyzed. Inhibitors of
extension would be expected to decrease the reaction rate and/or extent
(Fig. 3, curve b). The nature of the nucleus, however, is a variable as well.
Preformed fibrils can act as one kind of nucleus, adding monomers at the
ends or laterally, which can give rise to more complex kinetics [70].
Branching vs linear addition reactions can be distinguished in some
situations by dynamic light-scattering methods. Another type of nucleus
seems to be present in solubilized aqueous peptide solutions that can pass
through a filter having a pore size of 200 nm, but not 20 nm [101] and
displays linear kinetics of fibril extension [117]. These species are present in
too low an abundance to be observed directly, though they are detectable
kinetically.
Distinguishing the preformed and endogenous nucleus forms is
problematic. The behavior of the accretion of soluble peptide onto AD
plaques in tissue sections [118] or onto sonicated fibrils [119] is kinetically
similar to that of spontaneous soluble nuclei. However, the endogenous
soluble nuclei are not equivalent on a molecular level to fibril or AD plaque
nuclei, since molecules such as Congo Red inhibit endogenous soluble
nuclei extension at 0.25 AM [97], while over 700 AM is required to block
accretion of monomer h peptide onto fibrils or plaques [119]. An assemblycompetent form of the h peptide, h10–35, has been shown to interact with
fibrils and plaques [120] and to adopt a particular conformation in solution
as determined by NMR spectroscopy [121], although no high affinity
inhibitors of this accretion reaction have been reported. Alternatively,
low abundance conformational forms of monomeric peptide may be the
actively associating form of the peptide to endogenous seeds, accounting
for the high (micromolar) amounts of peptide required in vitro for the
extension reaction. These forms may be more abundant in biological
systems, allowing fibril formation to occur at the low bulk concentrations
(nanomolar) of Ah peptides found in vivo.
As a result of the confusion over the identity of nucleating h-peptide
species, prenucleation events remain poorly defined. A variety of methods
possessing different degrees of resolution have been employed to look at
these early stages in fibril formation. Chemical [122] and enzymatic
(transglutaminase) [123,124] cross-linking, electron microscopy (EM)
[105,106,125], AFM [93,94,102,103], ultracentrifugation, dynamic light
258
LeVine
scattering, and fluorescence resonance energy transfer (FRET) with modified beta peptides [89,105] have probed the oligomerization process
preceding nucleus formation but have not yielded definitive structural
information on the species present or the extent of their participation in
nucleus formation. Difficulties arise from the small size (4.3 kDa) of the
monomeric peptide unit, the simultaneous presence of multiple species of
peptide, both conformational and association states, and their transient
nature (since they rapidly form amyloid fibrils as their concentrations
increase). As specific inhibitors of early stages in fibril growth are discovered, peptide species will be better defined, particularly if the intermediates
can be trapped and their structures determined.
Fibril extension from nuclei preformed under defined conditions
has been characterized through a series of nucleus-dependent kinetic
assays. The process of fibril formation from a nucleus in equilibrium with
soluble, mostly monomeric peptide has proved much more amenable to
study than the formation of the nucleus itself. Fibrillar species are readily
detected by growth in size (filtration, sedimentation, static light scattering–turbidity), amyloid-specific reactivity with the optical probes Congo
Red and thioflavin S and T, and by EM and AFM. Endogenous soluble
nuclei or seeds form in aqueous solution, accumulating slowly at low
temperature. Brief treatment with denaturants, organic solvents, and
treatment with neat trifluoroacetic acid (TFA) or concentrated formic
acid breaks down these seed structures, restoring the lag period of
unseeded fibril formation.
The processes of both seed formation and fibril extension are
dependent on temperature and on peptide concentration, with 37jC
being required for establishing equilibrium within 24 h with 30 AM
h1–40. A full description of the assay system may be found elsewhere
[97,117]. A 4 h reaction time is typically within the linear portion of the
time course. This nucleus-dependent assay detects mainly inhibitors that
are substoichiometric with the monomeric peptide, which is present at
high concentration. It is relatively insensitive to inhibitors that target
the monomeric peptide. Whether the inhibitors interact with the growing end of a seed or with a low abundance conformational form of the
h peptide that is competent to add to the seed is difficult to determine at
this time. Similar dose–response curves are obtained for Congo Red as
an inhibitor with either thioflavin T (ThT) fluorescence or filtration of
radioiodinated peptide readouts (Fig. 4) Caveats in the interpretation of
both the ThT and radiometric filtration assays for the evaluation of putative inhibitors are discussed elsewhere [97].
Protein Misfolding
259
Figure 4 Inhibition of nucleated fibril extension by Congo Red: h1–40 fibril
formation detected by filtration of either radiolabeled peptide (circles) or
thioflavin T (ThT) reactivity, (triangles) is inhibited by Congo Red with similar
potency. For assay details, see Ref. 97.
The prediction that fibrillization reactions proceeding via different folding pathways governed by different rate-limiting steps could be
subject to different modes of inhibition appears to be substantiated.
The endogenously seeded type of assay identifies types of inhibitor
different from unseeded assays by using light scattering or turbidity
detection. With the exception of the naphthyl monoazo benzo compounds (12) and the acridinone series (11), the molecules reported in
Figure 2 are ineffective (IC50 > 100 AM) in the presence of 30 AM
h1–40 in seeded assays. In particular, short peptide sequences derived from
the h16–25 amyloidogenic core of the h peptide KLVFFA are ineffective under the seeded assay conditions, although many modifications
of this sequence have been studied [126–129], some of which (e.g., 7)
are being developed as therapeutics. Inhibitors effective in the seeded
assay format such as Congo Red are inactive in an accretion assay
onto immobilized fibrils [119]. Rifampicin and daunomycin are very
weakly active against accretion [130].
260
LeVine
IV. CELLULAR SYSTEMS AND AMYLOID FIBRILS
One of the reported biological effects of h-amyloid and amylin fibrils
is cellular toxicity, inferred in vivo and modeled in various tissue
culture systems. While amyloid fibrils were initially thought to be the
toxic species, it has become increasingly clear that some other entity,
probably soluble oligomers of the h peptide [131,132] that are in
equilibrium with fibrils, are the culprit. Thus, in developing aggregation inhibitors that would be therapeutically useful, it is important to
demonstrate that the nonfibrillar peptide species stabilized by inhibitor
treatment are not toxic to cells. The selection of an appropriate
cellular system is important because the resistance of cell types to
the toxic effects of the Ah peptide varies significantly, often requiring
industrial (50–100 AM) concentrations of peptide or the use of the
nonbiological h25–35 fragment. Mixed neuronal/glial or pure neuronal
embryonic hippocampal or cortical cultures would seem to be the most
relevant cell type, since neuronal cell death and dysfunction are
hallmarks of neurodegenerative disease like AD. Unfortunately, the
embryonic primary cultures are irregularly resistant to the effects of
h1–42 when cell death is monitored. These cultures are heterogeneous
mixtures of neuronal cell types, only some of which seem to be
affected by the Ah peptide. In addition, embryonic mouse neurons
are not the same as the deeply differentiated cells in the brain of an
80-year-old human. Cultured PC12 cell lines have become a favorite
system, with changes in MTT formazan production serving as a
readout. However, the formazan deposition is not related to cell
survival [133–136] and so is not reliable as an indicator of the effects
of amyloid on cell death.
Another prominent site of deposition of h-amyloid fibrils with age
and in AD is within the cerebrovasculature in areas of the brain prone
to parenchymal amyloid deposition [137–139]. The peptide deposits
along the surfaces of the smooth muscle cells of the vascular wall,
resulting in the death of those cells and their replacement by amyloid
fibrils, weakening the vascular wall. Endothelial cells are also affected
[140]. The ‘‘Dutch’’ mutation in the APP precursor protein Q22E,
within the h-peptide sequence, produces a particularly fibrillogenic
and toxic (to smooth muscle cells) peptide associated with primarily
vascular deposition of mutant peptide and hemorrhagic vessel disease
[137]. Thus, in addition to neuronal cells, the brain vascular smooth
muscle cells are a pathologically relevant cell type. While the source of
Protein Misfolding
261
the h peptide in these deposits (brain or smooth muscle cells) is under
debate, the smooth muscle cells in culture generate prodigious amounts
of h peptide and accumulate C-terminal fragments of hAPP [139].
Organotypic cultures of the leptomeningeal blood vessels will accumulate exogenously applied, fluorescently labeled Ah peptide [141]. The
leptomeningeal vascular smooth muscle cells (VSMC) isolated from
either human or canine sources have proved to be reliable indicators for
h-amyloid toxicity. Overnight treatment with 10 AM h1–42 leads to
deposition of fibrillar peptide in ThS-positive strands onto the cell
surface and apoptosis of 70 to 80% of the VSMC assessed by
bisbenzamide staining of condensed nuclei. For this cell type, preformed
fibrils have little effect on cell survival, and the added fibrils remain
scattered over the surface of the culture dish.
Interpreting the effects of amyloid-modulating compounds on
h1–42-induced cellular toxicity and relating the results to in vitro aggregation inhibition is far from straightforward. A number of compounds
are toxic to cells by a variety of routes. Besides interfering with
aggregation, test compounds can block binding to the cell surface,
internalization of h peptide, or any of a myriad of cellular events that
could affect expression of h-peptide toxicity. Lack of effect of a
compound could indicate that it is not blocking the toxic ‘‘site’’ on
the peptide species, that it is not penetrating the cell, or simply that
the compound is adsorbed, sequestered, or metabolized to an inactive
form. In the VSMC system as in other cellular systems [142], Congo
Red blocks both aggregation on the cell surface and h1–42 toxicity at
10 AM, roughly equivalent to the total added peptide concentration.
For optimal effect it must be added either before or along with the h
peptide. Since the IC50 for an antiaggregation effect on 30 AM peptide
in vitro is 0.25 AM, nonspecific adsorption of the compound to
cellular components and to h-peptide fibrils may be mitigating the
effects. Congo Red and other polysulfonate/sulfate polyanions are
known to displace proteins from binding sites on the cell surface
[143]. Congo Red’s practical therapeutic potential is limited because it
does not penetrate the cell membrane or the blood–brain barrier, and
the azo linkages are susceptible to metabolism and carcinogenic
liability. Such difficulties can be addressed by structural modifications
in inhibitors that will likely also improve some of the pharmacokinetic
properties in vivo. However, the connection between cellular effects
and desired in vivo properties of bioavailability and brain penetration
is also not straightforward.
262
LeVine
V. ANIMAL MODELS OF BRAIN DEPOSITION
OF INSOLUBLE PROTEINS
The lack of animal models of Alzheimer’s disease that faithfully
reproduce the human condition has significantly retarded development
of therapies that would be expected to have a significant impact on
the disease or its progression. Such models would provide a test bed
for increasing confidence that a given therapeutic regime would have
the desired effect in animals and would give insights that only
experimental manipulation can provide into the disease process.
Cholinomimetic approaches employed models of cholinergic deficit
to bring the current anticholinesterase drugs to market, even though
they only modeled a subset of the pathology of AD [144]. Serious
discovery and development of therapeutic agents directed at the
deposition of h-amyloid fibrils was put on a firm footing by the
development of mice that overexpressed human APP, and depicted
fibrillar and nonfibrillar h1–40 and h1–42 in the appropriate brain
regions. Various hAPP mutants responsible for early-onset familial
AD have been the most effective, particularly in combination with
transgenic presenilin mutants, another familial AD locus. For a review
of the hAPP mouse and other brain amyloidosis models see Walker [145].
Again, these are only partial models of AD, since no significant neuronal
cell death has been observed with these mice. Importantly, the Hsiao
mouse, Tg 2576 (human APP695, Swedish mutation under the control of
the mouse prion promoter), has been available to both commercial and
academic laboratories, and thus there is a considerable shared pool of
information and experience with this model. Direct comparisons with the
other hAPP mice are limited because those models are not generally
available. The Hsiao mouse shows robust and increasing deposition of h
peptide from age 9 months onward, as well as dystrophic neurites, reactive
astrogliosis, microglial activation around senile plaques, and some phosphorylated (but not tangled) tau immunoreactivity, but no detectable
neuronal cell death or reduction in synaptic counts, despite plaque
densities approaching that of clinical AD.
An important validation of the mice that overexpress human mutant
hAPP as a platform for testing therapeutics targeting h-peptide deposition
has been provided by the Elan company, using their PDAPP mouse [146].
Immunization of the mice, either at an early age or after plaques had
formed, resulted in clearance of immunoreactive plaques and peptide from
the subjects’ brains. Although the elucidation of the mechanism explaining
Protein Misfolding
263
this potential therapy has not been reported, the experiments establish that
deposition of the peptide can be interrupted, and even reversal of preformed plaques is possible in this type of animal model. Human trials of the
safety and efficacy of the immunization protocol are to start in the near
future. The plaques are thus dynamic structures, and therapeutics that
interfere with Ah deposition or production would be predicted to reduce
brain h-amyloid load [147].
Similar partial neurodegenerative disease animal models involving
insoluble protein deposition have been developed for Huntington’s
disease [148], spinocerebellar ataxia type 1 (SCA1) [149,150], and
Machado–Joseph disease (MJD; SCA3) [151], all trinucleotide repeat
disorders. The protein deposits, consisting primarily of the polyglutamine tract, accumulate intracellularly, eventually collecting in the
nucleus, where they are thought to disrupt nuclear function both in
humans and in the animal models. As for Alzheimer’s disease and the
Ah peptide and tau proteins, there is considerable controversy about
the relevance of the deposits to the disease process because observed
neuronal cell loss does not overlap entirely with visible inclusions. The
arguments in favor of relevance are that the toxic species may be
smaller oligomeric species not observed by microscopy and that in some
situations the deposits may serve a protective function by sequestering
potentially toxic material.
The Huntington’s disease models have been studied in considerable
detail. The pathology bears a striking resemblance to the human disease in
a number of respects, although polyglutamine overexpression is not a
complete model. The Bates R6 mice expressing exon I of the human
Huntingtin protein, consisting of the N-terminal 17 amino acids + a
pathological number (115–156 CAGs) of glutamines + 52 more amino
acids under control of the human promoter, develop age-dependent, brainregion-specific cell loss accompanied by nuclear inclusions and behavioral
and motor abnormalities reflecting those in the human disease, leading
eventually to death [152]. Time of onset, severity of the symptoms, and
length of disease are dependent on the number of glutamine repeats in the
observed human pathological range. A longer term mouse model with the
full-length human Huntingtin under its natural promoter, which lacks
the potential diabetic condition of the Bates mouse, may provide a more
realistic picture of the HD process, although it would be less useful for
rapid testing of potential therapies.
Mice expressing high levels of human a-synuclein under the
control of the human PDGFh promoter developed intracellular nuclear
264
LeVine
and cytoplasmic inclusions that were immunoreactive with antibody
specific for the human a-synuclein [153]. As early as 2 to 3 months of
age in the transgenics, inclusions were found in the deeper layers of the
neocortex, the CA3 region of the hippocampus, the olfactory bulb, and
occasionally the substantia nigra, regions typically affected in Lewy
body disease. While nigral tyrosine hydroxylase–positive cell number
was similar to nontransgenic littermates, TH-positive nerve terminals
were significantly reduced, as were TH immunoreactivity and enzymatic
activity in 12-month-old animals. In the 12-month-old mice, neurological impairment similar to that found in Parkinson’s disease was
demonstrable in rotorod performance.
Human prion disease models have also been developed in mice
[154,155]. Crossing the species barrier into an experimentally accessible
animal system, the prions responsible for Creutzfeldt–Jakob disease, new
variant CJD, Gerstmann–Straussler–Scheinker disease, and fatal familial
insomnia produce a reproducible time-dependent neuronal degeneration
leading to death.
VI. CLINICAL TRIALS FOR AD TESTING OF
POSSIBLE DISEASE-MODIFYING AGENTS
While testing of amyloid aggregation inhibitors against AD in human
subjects is a way off, it is worth considering how such trials should be
conducted to establish clinical efficacy. With all the genetic and
biochemical evidence that the h peptide is implicated in AD pathology,
it very well may not be the only relevant pathology in all patients for this
very complex disease. It remains distinctly possible that removal of all
amyloid plaques and/or h peptide from the brains of AD patients will
not restore cognitive function in advanced stages of the disease. Thus, a
paradigm treating significantly cognitively impaired patients to look for
a leveling off in their decline or a reversal to normal may not show the
desired treatment effect with an aggregation inhibitor. To be fair, it may
not show effects with any treatment if too many neurons have died or
become dysfunctional. The cholinomimetic therapies were symptomatic
treatments designed to supplement function (acetylcholine) that had
been lost without regard for the process that caused that functional
loss. In attacking what is believed to be a fundamental process in
disease progression, other measures may be needed to reverse the
degeneration that has already occurred, assuming that not too much
Protein Misfolding
265
damage has been inflicted. Disease-modifying therapies are most likely
to influence progression to disease and/or delay onset of symptoms.
Proper assessment of the effects of a therapy will require clinical trial
designs that make the appropriate measurements. It will be important
to assess plaque load and/or h-peptide level in patients treated with
aggregation inhibitors or other modes of reducing either brain hpeptide content or its effects in addition to the classical cognitive end
points. Whatever the outcome of the trials, in interpreting the results
for the development of new generations of therapeutics it is important
to determine whether the therapy accomplished what it was designed to
do—reduce amyloid peptide deposition. If h peptide is eliminated but
no therapeutic benefit is observed, we should conclude that the h
peptide is not the major player—at least in the patient population selected for study. The answer will be important in justifying future
pharmaceutical investment as well as in guiding future research for
effective therapies against AD.
REFERENCES
1. Davies SW, Beardsall K, Turmaine M, DiFiglia M, Aronin N, Bates GP.
Are neuronal intranuclear inclusions the common neuropathology of
triplet-repeat disorders with polyglutamine-repeat expansions? Lancet
1998; 351:131–133.
2. Kakizuka A. Protein precipitation: a common etiology in neurodegenerative disorders? Trends Genet 1998; 14:396–402.
3. Kim TW, Tanzi RE. Neuronal intranuclear inclusions in polyglutamine
diseases: nuclear weapons or nuclear fallout? Neuron 1998; 21: 657–659.
4. Koshy BT, Zoghbi HY. The CAG/polyglutamine tract diseases: gene
products and molecular pathogenesis. Brain Pathol 1997; 7:927–942.
5. Rubinaztein DC, Wyttenbach A, Rankin J. Intracellular inclusions,
pathological markers in diseases caused by expanded polyglutamine tracts?
J Med Genet 1999; 36:265–270.
6. Scherzinger E, Lurz R, Turmaine M, Mangiarini L, Hollenbach B,
Hasenbank R, Bates GP, Davies SW, Lehrach H, Wanker EE. Huntingtinencoded polyglutamine expansions form amyloid-like protein aggregates in
vitro and in vivo. Cell 1997; 90:549–558.
7. Lathrop RH, Casale M, Tobias DJ, Marsh JL, Thompson LM. Modeling
protein homopolymeric repeats: possible polyglutamine structural motifs
for Huntington’s disease. Proc Int Conf Intelligent Syst Molec Biol 1998;
6:105–114.
266
LeVine
8. Altschuler EL, Hud NV, Mazrimas JA, Rupp B. Random coil conformation for extended polyglutamine stretches in aqueous soluble monomeric peptides. J Peptide Res 1997; 50:73–75.
9. Scherzinger E, Sittler A, Schweiger K, Heiser V, Lurz R, Hasenbank R,
Bates GP, Lehrach H, Wanker EE. Self-assembly of polyglutamine-containing huntingtin fragments into amyloid-like fibrils: implications for Huntington’s disease pathology. Proc Natl Acad Sci USA 1999; 96:4604–4609.
10. Bates GP, Mangiarini L, Davies SW. Transgenic mice in the study
of polyglutamine repeat expansion diseases. Brain Pathol 1998; 8:
699–714.
11. Perez MK, Paulson HL, Pendse SJ, Saionz SJ, Bonini NM, Pittman RN.
Recruitment and the role of nuclear localization in polyglutaminemediated aggregation. J Cell Biol 1998; 143:1457–1470.
12. Gourfinkel AI, Cancel G, Duyckaerts C, Faucheux B, Hauw JJ, Trottier Y,
Brice A, Agid Y, Hirsch EC. Neuronal distribution of intranuclear
inclusions in Huntington’s disease with adult onset. Neuroreport 1998;
9:1823–1826.
13. Becher MW, Kotzuk JA, Sharp AH, Davies SW, Bates GP, Price DL, Ross
CA. Intranuclear neuronal inclusions in Huntington’s disease and
dentatorubral and pallidoluysian atrophy: correlation between the density
of inclusions and IT15 CAG triplet repeat length. Neurobiol Dis 1998;
4:387–397.
14. DiFiglia M, Sapp E, Chase KO, Davies SW, Bates GP, Vonsattel JP,
Aronin N. Aggregation of Huntingtin in neuronal intranuclear inclusions
and dystrophic neurites in brain. Science 1997; 277:1990–1993.
15. Duyckaerts C, Durr A, Cancel G, Brice A. Nuclear inclusions in
spinocerebellar ataxia type 1. Acta Neuropathol Berl 1999; 97:201–207.
16. Hayashi Y, Kakita A, Yamada M, Egawa S, Oyanagi S, Naito H, Tsuji S,
Takahashi H. Hereditary dentatorubral–pallidoluysian atrophy: ubiquitinated filamentous inclusions in the cerebellar dentate nucleus neurons.
Acta Neuropathol Berl 1998; 95:479–482.
17. Holmberg M, Duyckaerts C, Durr A, Cancel G, Gourfinkel An I, Damier
P, Faucheux B, Trottier Y, Hirsch EC, Agid Y, Brice A. Spinocerebellar
ataxia type 7 (SCA7): a neurodegenerative disorder with neuronal
intranuclear inclusions. Hum Mol Genet 1998; 7:913–918.
18. Lieberman AP, Trojanowski JQ, Leonard DG, Chen KL, Barnett JL,
Leverenz JB, Bird TD, Robitaille Y, Malandrini A, Fischbeck KH. Ataxin
1 and ataxin 3 in neuronal intranuclear inclusion disease. Ann Neurol 1999;
46:271–273.
19. Schmidt T, Landwehrmeyer GB, Schmitt I, Trottier Y, Auburger G,
Laccone F, Klockgether T, Volpel M, Epplen JT, Schols L, Riess O. An
isoform of ataxin-3 accumulates in the nucleus of neuronal cells in
affected brain regions of SCA3 patients. Brain Pathol 1998; 8: 669–679.
Protein Misfolding
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
267
Goedert M. Filamentous nerve cell inclusions in neurodegenerative
diseases: tauopathies and a-synucleinopathies. Philos Trans R Soc Lond
B 1999; 354:1101–1118.
Jellinger KA. Movement disorders with tau protein cytoskeletal pathology.
Parkinson’s disease. In: Stern GM, ed. Advances in Neurology. Vol. 80.
Philadelphia: Lippincott, Williams & Wilkins, 1999:393–311.
Tolnay M, Probst A. Review: tau protein pathology in Alzheimer’s
disease and related disorders. Neuropathol Appl Neurobiol 1999; 25:
171–187.
Schweers O, Schonbrunn Hanebeck E, Marx A, Mandelkow E. Structural
studies of tau protein and Alzheimer paired helical filaments show no
evidence for beta-structure. J Biol Chem 1994; 269:24290–24297.
Delacourte A, Buee L. Normal and pathological tau proteins as factors for
microtubule assembly. Int Rev Cytol 1997; 171:167–224.
Lu PJ, Wulf G, Zhou XZ, Davies P, Lu KP. The prolyl isomerase Pin1
restores the functions of Alzheimer-associated phosphorylated tau protein
[see comments]. Nature 1999; 399:784–788.
Foster J, Hunter N. Transmissible spongiform encephalopathies: transmission, mechanism of disease, and persistence. Curr Opin Microbiol 1998;
1:442–447.
Kocisko DA, Priola SA, Raymond GJ, Chesebro B, Lansbury PT Jr,
Caughey B. Species specificity in the cell-free conversion of prion protein to
protease-resistant forms: a model for the scrapie species barrier. Proc Natl
Acad USA 1995; 92:3923–3927.
Kellershohn N, Laurent M. Species barrier in prion diseases: a kinetic
interpretation based on the conformational adaptation of the prion
protein. Biochem J 1998; 334:539–545.
Cohen FE, Prusiner SB. Pathologic conformations of prion proteins. Annu
Rev Biochem 1998; 67:793–819.
Come JH, Fraser PE, Lansbury PT, Jr. A kinetic model for amyloid
formation in the prion diseases: importance of seeding. Proc Natl Acad Sci
USA 1993; 90:5959–5963.
Lansbury PT Jr. Yeast prions: inheritance by seeded protein polymerization? Curr Biol 1997; 7:R617–R619.
Saborio GP, Soto C, Kascsak RJ, Levy E, Kascsak R, Harris DA,
Frangione B. Cell-lysate conversion of prion protein into its proteaseresistant isoform suggests the participation of a cellular chaperone.
Lansbury PT Jr, Caughey B. The chemistry of scrapie infection:
implications of the ‘ice 9’ metaphor. Chem Biol 1995; 2:1–5.
Edskes HK, Gray VT, Wickner RB. The [URE3] prion is an aggregated
form of Ure2p that can be cured by overexpression of Ure2p fragments.
268
LeVine
35.
DeArmond SJ, Ironside JW. Neuropathology of prion diseases. In:
Prusiner SB, ed. Prion Biology and Diseases. Vol. 38. Cold Spring Harbor,
NY: Cold Spring Harbor Press, 1999:585–652.
Safar J, Wille H, Itri V, Groth D, Serban H, Torchia M, Cohen FE,
Prusiner SB. Eight prion strains have PrP(Sc) molecules with different
conformations [see comments]. Nat Med 1998; 4:1157–1165.
Caughey B, Raymond GJ, Bessen RA. Strain-dependent differences in
beta-sheet conformations of abnormal prion protein. J Biol Chem 1998;
273:32230–32235.
Keohane C. The human prion diseases. A review with special emphasis on
new variant CJD and comments on surveillance. Clin Exp Pathol 1999;
47:125–132.
Farrer M, Gwinn-Hardy K, Hutton M, Hardy J. The genetics of disorders
with synuclein pathology and parkinsonism. Hum Mol Genet 1999;
8:1901–1905.
Polymeropoulos MH, Lavedan C, Leroy E, Ide SE, Dehejia A, Dutra A,
Pike B, Root H, Rubenstein J, Boyer R, Stenroos ES, Chandrasekharappa
S, Athanassiadou A, Papapetropoulos T, Johnson WG, Lazzarini AM,
Duvoisin RC, Di Iorio G, Golbe LI, Nussbaum RL. Mutation in the
alpha-synuclein gene identified in families with Parkinson’s disease. Science
1997; 276:2045–2047.
Polymeropoulos MH. Autosomal dominant Parkinson’s disease and alphasynuclein. Ann Neurol 1998; 44:S63–S64.
Wakabayashi K, Hansen LA, Vincent I, Mallory M, Masliah E. Neurofibrillary tangles in the dentate granule cells of patients with Alzheimer’s
disease, Lewy body disease and progressive supranuclear palsy. Acta
Neuropathol Berl 1997; 93:7–12.
Baba M, Nakajo S, Tu PH, Tomita T, Nakaya K, Lee VM, Trojanowski
JQ, Iwatsubo T. Aggregation of alpha-synuclein in Lewy bodies of
sporadic Parkinson’s disease and dementia with Lewy bodies. Am J Pathol
1998; 152:879–884.
Spillantini MG, Crowther RA, Jakes R, Cairns NJ, Lantos PT, Goedert
M. Filamentous a-synuclein inclusions link multiple system atrophy with
Parkinson’s disease and dementia with Lewy bodies. Neurosci Lett 1998;
251:205–208.
Weinreb PH, Zhen W, Poon AW, Conway KA, Lansbury PT Jr. NACP, a
protein implicated in Alzheimer’s disease and learning, is natively unfolded.
Biochemistry 1996; 35: 13709–13715.
Han H, Weinreb PH, Lansbury PT Jr. The core Alzheimer’s peptide NAC
forms amyloid fibrils which seed and are seeded by beta-amyloid: is NAC a
common trigger or target in neurodegenerative disease? Chem Biol 1995;
2:163–169.
Wood SJ, Wypych J, Steavenson S, Louis JC, Citron M, Biere AL. a-
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
Protein Misfolding
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
269
Synuclein fibrillogenesis is nucleation-dependent. Implications for the
pathogenesis of Parkinson’s disease. J Biol Chem 1999; 274:19509–19512.
Narhi L, Wood SJ, Steavenson S, Jiang Y, Wu GM, Anafi D, Kaufman
SA, Martin F, Sitney K, Denis P, Louis JC, Wypych J, Biere AL, Citron
M. Both familial Parkinson’s disease mutations accelerate alpha-synuclein
aggregation. J Biol Chem 1999; 274:9843–9846.
Durham HD, Roy J, Dong L, Figlewicz DA. Aggregation of mutant Cu/
Zn superoxide dismutase proteins in culture model of ALS. J Neuropathol
Exp Neurol 1997; 56:523–530.
Chou SM, Wang HS, Taniguchi A. Role of SOD-1 and nitric oxide/cyclic
GMP cascade on neurofilament aggregation in ALS/MND. J Neurol Sci
1996; 139 suppl:16–26.
Cleveland DW. From Charcot to SOD1: mechanisms of selective motor
neuron death in ALS. Neuron 1999; 24:525–520.
Vidal R, Fragione B, Rostagno A, Mead S, Revesz T, Plant G, Ghiso J. A
stop-codon mutation in the BRI gene associated with familial British
dementia. Nature 1999; 399:776–781.
Davis RL, Shimpton AE, Holohan PD, Bradshaw C, Feiglin D, Collins
GH, Sonderegger P, Kinter J, Becker LM, Lacbawan F, Krasnewich D,
Muenke M, Lawrence DA, Yerby MS, Shaw CM, Gooptu B, Elliott PR,
Finch JT, Carrell RW, Lomas DA. Famillial dementia caused by
polymerization of mutant neuroserpin. Nature 1999; 401:376–379.
Carrell RW, Lomas DA. Conformational disease. Lancet 1997; 350:134–138.
Baures PW, Peterson SA, Kelly JW. Discovering transthyretin amyloid
fibril inhibitors by limited screening. Bioorg Med Chem 1998; 6:
1389–1401.
Caughey WS, Raymond LD, Horiuchi M, Caughey B. Inhibition of
protease-resistant prion protein formation by porphyrins and phthalocyanines. Proc Natl Acad Sci USA 1998; 95:12117–12122.
Peterson SA, Klabunde T, Lashuel HA, Purkey H, Sacchettini JC, Kelly
JW. Inhibiting transthyretin conformation changes that lead to amyloid
fibril formation. Proc Natl Acad Sci USA 1998; 95:12956–12960.
Tagliavini F, McArthur RA, Canciani B, Giaccone G, Porro M, Bugiani
M, Lievens PM, Bugiani O, Peri E, Dall’Ara P, Rocchi M, Poli G,
Forloni G, Bandiera T, Varasi M, Suarato A, Cassutti P, Cervini MA,
Lansen J, Salmona M, Post C. Effectiveness of anthracycline against
experimental prion disease in Syrian hamsters. Science 1997; 276:1119–
1122.
Oza VB, Petrassi HM, Purkey HE, Kelly JW. Synthesis and evaluation of
anthranilic acid-based transthyretin amyloid fibril inhibitors. Bioorg Med
Chem Lett 1999; 9:1–6.
Oosawa F, Akakura S. Kinetics of Polymerization. Thermodynamics of the
Polymerization of Protein. New York: Academic Press, 1975:41–55.
270
LeVine
61. Oosawa F, Higashi S. Statistical thermodynamics of polymerization and
polymorphism of protein. In: Snell FM, ed. Progress in Theoretical
Biology. New York: Academic Press, 1967:79–164.
62. Bishop MF, Ferrone FA. Kinetics of nucleation-controlled polymerization. A perturbation treatment for use with a secondary pathway.
Biophys J 1984; 46:631–644.
63. Flyvbjerg H, Jobs E, Leibler S. Kinetics of self-assembling microtubules:
an ‘‘inverse problem’’ in biochemistry. Proc Natl Acad Sci USA 1996;
93:5975–5979.
64. Endres GF, Ehrenpreis S, Scheraga HA. Covalent bonding in the reversible polymerization of fibrin monomer. Biochim Biophys Acta 1965;
104:620–623.
65. Bark N, Foldes Papp Z, Rigler R. The incipient stage in thrombin-induced
fibrin polymerization detected by FCS at the single molecule level. Biochem
Biophys Res Commun 1999; 260:35–41.
66. Mosesson MW. Fibrinogen and fibrin polymerization: appraisal of the
binding events that accompany fibrin generation and fibrin clot assembly.
Blood Coagul Fibrinol 1997; 8:257–267.
67. Schutt CE, Myslik JC, Rozycki MD, Goonesekere NC, Lindberg U. The
structure of crystalline profilin-beta-actin. Nature 1993; 365:810–816.
68. Weeds AG, Gooch J, McLauglin P, Pope B, Bengtsdotter M, Karlsson R.
Identification of the trapped calcium in the gelsolin segment 1–actin
complex: implications for the role of calcium in the control of gelsolin
activity. FEBS Lett 1995; 360:227–230.
69. Lowe J. Crystal structure determination of FtsZ from Methanococcus
jannaschii. J Struct Biol 1998; 124:235–243.
70. Ferrone FA, Hofrichter J, Eaton WA. Kinetics of sickle hemoglobin
polymerization II. A double nucleation mechanism. J Mol Biol 1985;
183:611–633.
71. Ferrone FA, Hofrichter J, Eaton WA. Kinetics of sickle hemoglobin
polymerization. I: Studies using temperature jump and laser photolysis
techniques. J Mol Biol 1985; 183:591–610.
72. Eaton WA, Hofrichter J. The biophysics of sickle cell hydroxyurea therapy.
Science 1995; 268:1142–1143.
73. Lashuel HA, Lai Z, Kelly JW. Characterization of the transthyretin acid
denaturation pathways by analytical ultracentrifugation: implications for
wild-type, V30M, and L55P amyloid fibril formation. Biochemistry 1998;
37:17851–17864.
74. Lai Z, Colon W, Kelly JW. The acid-mediated denaturation pathway of
transthyretin yields a conformational intermediate that can self-assemble
into amyloid. Biochemistry 1996; 35:6470–6482.
75. Kelly JW, Colon W, Lai Z, Lashuel HA, McCulloch J, McCutchen SL,
Miroy GJ, Peterson SA. Transthyretin quaternary and tertiary structural
Protein Misfolding
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
271
changes facilitate misassembly into amyloid. Adv Protein Chem 1997;
50:161–181.
Bonifacio MJ, Sakaki Y, Saraiva MJ. ‘‘In vitro’’ amyloid fibril formation
from transthyretin: the influence of ions and the amyloidogenicity of TTR
variants. Biochim Biophys Acta 1996; 1316:35–42.
Yamada T, Kluve Beckerman B, Liepnieks JJ, Benson MD. Fibril formation
from recombinant human serum amyloid A. Biochim Biophys Acta 1994;
1226:323–329.
Friedhoff P, Schneider A, Mandelkow EM, Mandelkow E. Rapid assembly
of Alzheimer-like paired helical filaments from microtubule-associated
protein tau monitored by flourescence in solution. Biochemistry 1998;
37:10223–10230.
Schweers O, Mandelkow EM, Biernat J, Mandelkow E. Oxidation of
cysteine-322 in the repeat domain of microtubule-associated protein tau
controls the in vitro assembly of paired helical filaments. Proc Natl Acad
Sci USA 1995; 92:8463–8467.
Wilson DM, Binder LI. Free fatty acids stimulate the polymerization of
tau and amyloid beta peptides. In vitro evidence for a common effector of
pathogenesis in Alzheimer’s disease. Am J Pathol 1997; 150:2181–2195.
Kayed R, Bernhagen J, Greenfield N, Sweimeh K, Brunner H, Voelter W,
Kapurniotu A. Conformational transitions of islet amyloid polypeptide
(IAPP) in amyloid formation in vitro. J Mol Biol 1999; 287:781–796.
Charge SB, de Koning EJ, Clark A. Effect of pH and insulin on
fibrillogenesis of islet amyloid polypeptide in vitro. Biochemistry 1995;
34:14588–14593.
Wetzel R. Domain stability in immunoglobulin light chain deposition
disorders. Adv Protein Chem 1997; 50:183–242.
Helms LR, Wetzel R. Specificity of abnormal assembly in immunoglobulin light chain deposition disease and amyloidosis. J Mol Biol 1996;
257:77–86.
Stevens PW, Raffen R, Hanson DK, Deng YL, Berrios Hammond M,
Westholm FA, Murphy C, Eulitz M, Wetzel R, Solomon A, et al.
Recombinant immunoglobulin variable domains generated from synthetic
genes provide a system for in vitro characterization of light-chain amyloid
proteins. Protein Sci 1995; 4:421–432.
Hollenbach B, Scherzinger E, Schweiger K, Lurz R, Lehrach H, Wanker
EE. Aggregation of truncated GST-HD exon 1 fusion proteins containing
normal range and expanded glutamine repeats. Philos Trans R Soc Lond B
Biol Sci 1999; 354:991–994.
Burdick D, Soreghan B, Kwon M, Kosmoski J, Knauer M, Henschen A,
Yates J, Cotman C, Glabe C. Assembly and aggregation properties of
synthetic Alzheimer’s A4/beta amyloid peptide analogs. J Biol Chem 1992;
267:546–554.
272
LeVine
88. Fraser PE, McLachlan DR, Surewicz WK, Mizzen CA, Snow AD, Nguyen
JT, Kirschner DA. Conformation and fibrillogenesis of Alzheimer A beta
peptides with selected substitution of charged residues. J Mol Biol 1994;
244:64–73.
89. Garzon-Rodriguez W, Sepulveda-Becerra M, Milton S, Glabe CG. Soluble
amyloid Abeta-(1–40) exists as a stable dimer at low concentrations. J Biol
Chem 1997; 272:21037–21044.
90. Soto C, Castano EM, Kumar RA, Beavis RC, Frangione B. Fibrillogenesis
of synthetic amyloid-beta peptides is dependent on their initial secondary
structure. Neurosci Lett 1995; 200:105–108.
91. Soto C, Castano EM, Frangione B, Inestrosa NC. The alpha-helica to
beta-strand transition in the amino-terminal fragment of the amyloid betapeptide modulates amyloid formation. J Biol Chem 1995; 270:3063–3067.
92. Evans KC, Berger EP, Cho CG, Weisgraber KH, Lansbury PT Jr.
Apolipoprotein E is a kinetic but not a thermodynamic of amyloid
formation: implications for the pathogenesis and treatment of Alzheimer
disease. Proc Natl Acad Sci USA 1995; 92:763–767.
93. Harper JD, Wong SS, Lieber CM, Lansbury PT Jr. Assembly of A amyloid
protofibrils: an in vitro model for a possible early event in Alzheimer’s
disease. Biochemistry 1999; 38:8972–8980.
94. Harper JD, Lieber CM, Lansbury PT Jr. Atomic force microscopic
imaging of seeded fibril formation and fibril branching by the Alzheimer’s
disease amyloid-beta protein. Chem Biol 1997; 4:951–959.
95. Harper JD, Lansbury PT Jr. Models of amyloid seeding in Alzheimer’s
disease and scrapie: mechanistic truths and physiological consequences of
the time-dependent solubility of amyloid proteins. Annu Rev Biochem
1997; 66:385–407.
96. Jarrett JT, Lansbury PT Jr. Amyloid fibril formation requires a
chemically discriminating nucleation event: studies of an amyloidogenic
sequence from the bacterial protein OsmB. Biochemistry 1992; 31:
12345–12352.
97. LeVine H III. Screening for pharmacologic inhibitors of amyloid fibril
formation. Methods Enzymol 1999; 309:467–476.
98. Lomakin A, Teplow DB, Kirschner DA, Benedek GB. Kinetic theory of
fibrillogenesis of amyloid beta-protein. Proc Natl Acad Sci USA 1997;
94:7942–7947.
99. Shen CL, Scott GL, Merchant F, Murphy RM. Light scattering analysis of
fibril growth from the amino-terminal fragment beta(1–28) of beta-amyloid
peptide. Biophys J 1993; 65:2383–2395.
100. Tomski SJ, Murphy RM. Kinetics of aggregation of synthetic betaamyloid peptide. Arch Biochem Biophys 1992; 294:630–638.
101. Lomakin A, Chung DS, Benedek GB, Kirschner DA, Teplow DB. On the
nucleation and growth of amyloid beta protein fibrils: detection of nuclei
Protein Misfolding
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
273
and quantitation of rate constants. Proc Natl Acad Sci USA 1996;
93:1125–1129.
Stine WB Jr, Snyder SW, Ladror US, Wade WS, Miller MF, Perun TJ,
Holzman TF, Krafft GA. The nanometer-scale structure of amyloid-beta
visualized by atomic force microscopy. J Protein Chem 1996; 15:193–
203.
Harper JD, Wong SS, Lieber CM, Lansbury PT. Observation of
metastable Abeta amyloid protofibrils by atomic force microscopy. Chem
Biol 1997; 4:119–125.
Kowalewski T, Holtzman DM. In situ atomic force microscopy study of
Alzheimer’s beta-amyloid peptide on different substrates: new insights
into mechanism of beta-sheet formation. Proc Natl Acad USA 1999;
96:3688–3693.
Walsh DM, Hartley DM, Kusomoto Y, Fezoui Y, Condron MM,
Lomakin A, Benedek GB, Selkoe DJ, Teplow DB. Amyloid h-protein
fibrillogenesis. Structure and biological activity of protofibrillar intermediates. J Biol Chem 1999; 274:25949–25952.
Walsh DM, Lomakin A, Benedek GB, Condron MM, Teplow DB.
Amyloid beta-protein fibrillogenesis. Detection of a protofibrillar intermediate. J Biol Chem 1997; 272:22364–22372.
Blake CC, Serpell LC, Sunde M, Sandgren O, Lundgren E. A molecular
model of the amyloid fibril. CIBA Found Symp 1996; 199:6–15; discussion
15–21, 40–16.
Inouye H, Fraser PE, Kirschner DA. Structure of beta-crystalline
assemblies formed by Alzheimer beta-amyloid protein analogues: analysis
by X-ray diffraction. Biophys J 1993; 64:502–519.
Li L, Darden TA, Bartolotti L, Kominos D, Pederson LG. An atomic
model for the pleated beta-sheet structure of Abeta amyloid protofilaments. Biophys J 1999; 76:2871–2878.
Sunde M, Blake C. The structure of amyloid fibrils by electron microscopy
and X-ray diffraction. Adv. Protein Chem. 1997; 50:123–159.
Sunde M, Serpell LC, Bartlam M, Fraser PE, Pepys MB, Blake CC.
Common core structure of amyloid fibrils by synchrotron X-ray
diffraction. J Mol Biol 1997; 273:729–739.
Nguyen JT, Inouye H, Baldwin MA, Fletterick RJ, Cohen FE, Prusiner
SB, Kirschner DA. X-ray diffraction of scrapie prion rods and PrP
peptides. J Mol Biol 1995; 252:412–422.
Lazo ND, Downing DT. Beta-helical fibrils from a model peptide.
Goldsbury C, Kistler J, Aebi U, Arvinte T, Cooper GJ. Watching amyloid
fibrils grow by time-lapse atomic force microscopy. J Mol Biol 1999;
285:33–39.
Wall J, Schell M, Murphy C, Hrncic R, Stevens FJ, Solomon A.
274
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
LeVine
Thermodynamic instability of human E6 light chains: correlation with
fibrillogenicity. Biochemistry 1999; 38:14101–14108.
Ferrone F. Analysis of protein aggregation kinetics. Methods Enzymol
1999; 309:256–274.
Naiki H, Nakakuki K. First-order kinetic model of Alzheimer’s betaamyloid fibril extension in vitro. Lab Invest 1996; 74:374–383.
Maggio JE, Stimson ER, Ghilardi JR, Allen CJ, Dahl CE, Whitcomb DC,
et al. Reversible in vitro growth of Alzheimer disease h-amyloid plaques by
deposition of labeled amyloid peptide. Proc Natl Acad Sci USA 1992;
89:5462–5466.
Esler WP, Stimson ER, Ghilardi JR, Felix AM, Lu Y-A, Vinters HV,
Mantyh PJ, Maggio JE. Ah deposition inhibitor screen using synthetic
amyloid. Nat Biotechnol 1997; 15:258–263.
Esler WP, Stimson ER, Ghilardi JR, Lu YA, Felix AM, Vinters HV,
Mantyh PW, Lee JP, Maggio JE. Point substitution in the central
hydrophobic cluster of a human beta-amyloid congener disrupts
peptide folding and abolishes plaque competence. Biochemistry 1996;
35:13914–13921.
Lee JP, Stimson ER, Ghilardi JR, Mantyh PW, Lu Y-A, Felix AM,
Llanos W, Behbin A, Cummings M, Criekinge MV, Timms W, Maggio
JE. 1H NMR of Ah amyloid peptide congeners in water solution.
Conformational changes correlate with plaque competence. Biochemistry
1995; 34:5191–5200.
LeVine H III. Soluble multimeric Alzheimer beta(1–40) preamyloid
complexes in dilute solution. Neurobiol Aging 1995; 16:755–764.
Dudek SM, Johnson GV. Transglutaminase facilitates the formation of
polymers of the beta-amyloid peptide. Brain Res 1994; 651:129–133.
Rasmussen LK, Sorensen ES, Petersen TE, Gliemann J, Jensen PH.
Identification of glutamine and lysine residues in Alzheimer amyloid
beta A4 peptide responsible for transglutaminase-catalysed homopolymerization and cross-linking to alpha 2M receptor. FEBS Lett 1994;
338:161–166.
Nybo M, Svehag SE, Holm Nielsen E. An ultrastructural study of amyloid
intermediates in A beta(1–42) fibrillogenesis. Scand J Immunol 1999;
49:219–223.
Poduslo JF, Curran GL, Kumar A, Frangione B, Soto C. Beta-sheet
breaker peptide inhibitor of Alzheimer’s amyloidogenesis with increased
blood–brain barrier permeability and resistance to proteolytic degradation
in plasma. J Neurobiol 1999; 39:371–382.
Pallitto MM, Ghanta J, Heizelman P, Kiessling LL, Murphy RM.
Recognition sequence design for peptidyl modulators of beta-amyloid
aggregation and toxicity. Biochemistry 1999; 38:3570–3578.
Tjernberg LO, Lilliehook C, Callaway DJ, Naslund J, Hahne S, Thyberg J,
Protein Misfolding
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
275
Terenius L, Nordstedt C. Controlling amyloid beta-peptide fibril formation
with protease-stable ligands. J Biol Chem 1997; 272:12601–12605.
Ghanta J, Shen CL, Kiessling LL, Murphy RM. A strategy for designing inhibitors of beta-amyloid toxicity. J Biol Chem 1996; 271:
29525–29528.
Howlett DR, Perry AE, Godfrey F, Swatton JE, Jennings KH, Spitzfaden
C, Wadsworth H, Wood SJ, Markwell RE. Inhibition of fibril formation in
beta-amyloid peptide by a novel series of benzofurans. Biochem J 1999;
340:283–289.
Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos M,
Morgan TE, Rozovsky I, Trommer B, Viola KL, Wals P, Zhang C, Finch
CE, Krafft GA, Klein WL. Diffusible, nonfibrillar ligands derived from
Abeta(1–42) are potent central nervous system neurotoxins. Proc Natl
Acad Sci USA 1998; 95:6448–6453.
Roher AE, Chaney MO, Kuo YM, Webster SD, Stine WB, Haverkamp
LJ, Woods AS, Cotter RJ, Tuohy JM, Krafft GA, Bonnell BS, Emmerling
MR. Morphology and toxicity of Abeta-(1–42) dimer derived from neuritic
and vascular amyloid deposits of Alzheimer’s disease. J Biol Chem 1996;
271:20631–20635.
Liu Y, Schubert D. Cytotoxic amyloid peptides inhibit cellular 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MMT) reduction
by enhancing MTT formazan exocytosis. J Neurochem 1997; 69:2285–2293.
Patel AJ, Gunasekera S, Jen A, Rohan de Silva HA. h-Amyloid-mediated
inhibition of redox activity (MMT reduction) is not an indicator of
astroglial degeneration. Neuroreport 1996; 7:2026–2030.
Isobe I, Michikawa M, Yanagisawa K. Enhancement of MTT, a
tetrazolium salt, exocytosis by amyloid beta-protein and chloroquine in
cultured rat astrocytes. Neurosci Lett 1999; 266:129–132.
Abe K, Saito H. Amyloid beta protein inhibits cellular MTT reduction not
by suppression of mitochondrial succinate dehydrogenase but by
acceleration of MTT formazan exocytosis in cultured rat cortical
astrocytes. Neurosci Res 1998; 31:295–305.
Van Nostrand WE, Melchor JP, Ruffini L. Pathologic amyloid betaprotein cell surface fibril assembly on cultured human cerebrovascular
smooth muscle cells. J Neurochem 1998; 70:216–223.
Wisniewski HM, Frackowiak J, Mazur Kolecka B. In vitro production of
beta-amyloid in smooth muscle cells isolated from amylid angiopathyaffected vessels. Neurosci Lett 1995; 183:120–123.
Frackowiak J, Mazur Kolecka B, Wisniewski HM, Potempska A, Carroll
RT, Emmerling MR, Kim KS. Secretion and accumulation of Alzheimer’s
beta-protein by cultured vascular smooth muscle cells from old and young
dogs. Brain Res 1995; 676:225–230.
Crawford F, Soto C, Suo Z, Fang C, Parker T, Sawar A, Frangione B,
276
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
LeVine
Mullan M. Alzheimer’s beta-amyloid vasoactivity: identification of a novel
beta-amyloid conformational intermediate. FEBS Lett 1998; 436:445–448.
Prior R, D DU, Frank R, Prikulis I, Pavlakovic G. Experimental
deposition of Alzheimer amyloid beta-protein in canine leptomeningeal
vessels. Neuroreport 1995; 6:1747–1751.
Lorenzo A, Yankner BA. Beta-amyloid neurotoxicity requires fibril
formation and is inhibited by Congo Red. Proc Natl Acad Sci USA 1994;
91:12243–12247.
Sadler II, Smith DW, Shearman MS, Ragan CI, Tailor VJ, Pollack SJ.
Sulphated compounds attenuate beta-amyloid toxicity by inhibiting its
association with cells. Neuroreport 1995; 7:49–53.
Knopman DS. Current pharmacotherapies for Alzheimer’s disease.
Geriatrics 1998; 53:S31–S34.
Walker LC. Animal models of cerebral beta-amyloid angiopathy. Brain
Res Brain Res Rev 1997; 25:70–84.
Schenk D, Barbour R, Dunn W, Gordon G, Grajeda H, Guido T, Hu K,
Huang J, Johnson-Wood K, Khan K, Kholodenko D, Lee M, Liao Z,
Lieberberg I, Motter R, Mutter L, Soriano F, Schopp G, Vasquez N,
Vandevert C, Walker S, Wogulis M, Yednock T, Games D, Seubert P.
Immunization with amyloid-h attenuates Alzheimer-disease-like pathology
in the PDAPP mouse. Nat Lond 1999.
Wagner SL, Munoz B. Modulation of amyloid ß-protein precursor
processing as a means of retarding progression of Alzheimer’s disease.
J Clin Invest 1999; 104:1329–1332.
Davies SW, Turmaine M, Cozens BA, DiFiglia M, Sharp AH, Ross CA,
Scherzinger E, Wanker EE, Mangiarini L, Bates GP. Formation of
neuronal intranuclear inclusions underlies the neurological dysfunction in
mice transgenic for the HD mutation. Cell 1997; 90:537–548.
Klement IA, Skinner PJ, Kaytor MD, Yi H, Hersch SM, Clark HB,
Zoghbi HY, Orr HT. Ataxin-1 nuclear localization and aggregation: role in
polyglutamine-induced disease in SCA1 transgenic mice [see comments].
Cell 1998; 95:41–53.
Klement IA, Zoghbi HY, Orr HT. Pathogenesis of polyglutamine-induced
disease: A model for SCA1. Mol Genet Metab 1999; 66:172–178.
Ikeda H, Yamaguchi M, Sugai S, Aze Y, Narumiya S, Kakizuka A.
Expanded polyglutamine in the Machado–Joseph disease protein induces
cell death in vitro and in vivo. Nat Genet 1996; 13:196–202.
Lione LA, Carter RJ, Hunt MJ, Bates GP, Morton AJ, Dunnett SB.
Selective discrimination learning impairments in mice expressing the
human Huntington’s disease mutation. J Neurosci 1999; 19:10428–
10437.
Masliah E, Rockenstein E, Veinbergs I, Mallory M, Hashimoto M, Takeda
A, Sagara Y, Sisk A, Mucke L. Dopaminergic loss and inclusion body
Protein Misfolding
154.
155.
156.
157.
158.
159.
160.
161.
162.
277
formation in a-synuclein mice: implications for neurodegenerative
disorders. Science 2000; 287:1265–1269.
Weissman C, Raeber AJ, Shmerling D, Aguzzi A, Manson JC. Knockouts,
transgenics, and transplants in prion research. In: Prusiner SB, ed. Prion
Biology and Diseases. Vol. 38. Cold Spring Harbor, NY: Cold Spring
Harbor Press, 1999:273–305.
Raeber AJ, Brandner S, Klein MA, Benninger Y, Musahl C, Frigg R,
Roeckl C, Fischer MB, Weissmann C, Aguzzi A. Transgenic and knockout
mice in research on prion diseases. Brain Pathol 1998; 8:715–733.
Salomon AR, Marcinowski KJ, Friedland RP, Zagorski MG. Nicotine
inhibits amyloid formation by the beta-peptide. Biochemistry 1996;
35:13568–13578.
Wood SJ, MacKenzie L, Maleeff B, Hurle MR, Wetzel R. Selective
inhibition of Abeta fibril formation. J Biol Chem 1996; 271:4086–4092.
Pappolla M, Bozner P, Soto C, Shao H, Robakis NK, Zagorski M,
Frangione B, Ghiso J. Inhibition of Alzheimer beta-fibrillogenesis by
melatonin. J Biol Chem 1998; 273:7185–7188.
Merlini G, Ascari E, Amboldi N, Bellotti V, Arbustini E, Perfetti V,
Ferrari M, Zorzoli I, Marinone MG, Garini P, et al. Interaction of the
anthracycline 4V-iodo-4V-deoxydoxorubicin with amyloid fibrils: inhibition
of amyloidogenesis. Proc Natl Acad Sci USA 1995; 92:2959–2963.
Howlett D, Cutler P, Heales S, Camilleri P. Hemin and related porphyrins
inhibit beta-amyloid aggregation. FEBS Lett 1997; 417:249–251.
Tomiyama T, Shoji A, Kataoka K, Suwa Y, Asano S, Kaneko H, Endo N.
Inhibition of amyloid beta protein aggregation neurotoxicity by rifampicin.
Its possible function as a hydroxyl radical scavenger. J Biol Chem 1996;
271:6839–6844.
Kisilevsky R, Lemieux LJ, Fraser PE, Kong X, Hultin PG, Szarek WA.
Arresting amyloidosis in vivo using small-molecule anionic sulphonates
or sulphates: implications for Alzheimer’s disease. Nat Med 1995; 1:143–
148.
10
Uncoating and Adsorption Inhibitors
of Rhinovirus Replication
Guy D. Diana
ViroPharma, Inc., Exton, Pennsylvania, U.S.A.
Adi Treasurywala
Pfizer Central Research, Groton, Connecticut, U.S.A.
I.
INTRODUCTION
Rhinoviruses are responsible for approximately 50% of infections resulting in the common cold [1]. These infections are caused by over 100 distinct serotypes, which vary by exhibiting minor or major changes in
structure. The virus is divided into a major group, consisting of approximately 90%, and a minor group, differing by the mode of attachment of
the virus to the cell. The members of the major group of serotypes have
been shown to bind to domain 1 and 2 of ICAM-1 [2 – 5], while the minor
serotypes appear to have a binding preference for the human low density
lipoprotein receptor (LDLR) [6]. An effective antirhinovirus agent would
be expected to be active against the majority of serotypes, since at any time
one may become infected by any of the 100+ serotypes. The compounds
in the series shown in Figure 1 have demonstrated broad spectrum antirhinovirus activity against both the minor and major group of serotypes
[7 – 12]. These compounds have been shown to inhibit uncoating of the
major group [13] and to block adsorption of the minor serotypes to the cell
[14]. This chapter describes our efforts to determine the mode of binding
279
280
Figure 1
Diana and Treasurywala
General structure of isoxazole series of antipicornavirus agents.
of these compounds and to enhance the activity by utilizing x-ray crystallography coupled with traditional structure –activity methodology.
II. CAPSID BINDING COMPOUNDS
In 1985 Dr. Michael Rossmann and his colleagues determined for the
first time the three-dimensional structure of a human rhinovirus [15].
Their studies, performed with human rhinovirus type 14 (HRV-14), revealed the structure as an eicosahedron consisting of four proteins designated VP1, VP2, VP3, and VP4 forming a protomeric unit, combined
to form a fivefold axis of symmetry (Fig. 2). The surface of the capsid
Figure 2 The three-dimensional structure of HRV-114 consisting of four viral
proteins, VP1, VP2, and VP3; Vp4 is pointing toward the center of the capsid
protein and is not visible.
Inhibitors of Rhinovirus Replication
281
Figure 3 The orientation of disoxaril in the binding site of HRV-14.
protein contains a canyon that was shown to be the cell receptor binding
site [2]. Subsequently, the structure of several additional rhinovirus
serotypes was determined [16 –19]. Although these rhinoviruses share
the same general structure described for HRV-14, the latter appears to be
distinctly different from other rhinoviruses, particularly with respect to
the sequence similarity. Following the elucidation of the structure of
HRV-14, x-ray studies were performed on two members of the series of
compounds shown in Figure 1, disoxaril and WIN52084 [20]. The purpose of this study was to elucidate the nature of the binding of these
compounds to the capsid protein.
Disoxaril was shown to bind in a hydrophobic pocket below the
a depression referred to as the ‘‘canyon,’’ with the oxazoline ring in the
‘‘toe’’ region of the binding pocket. The isoxazole ring resides in the ‘‘heel’’
below the area designated as the pore (Fig. 3). The nitrogen of the isoxazole
Figure 4 Structure of WIN52084.
282
ring was within 3.5 A˚ of asparagine 219, suggesting the possibility of
hydrogen bonding. The phenyl ring was in a stacking conformation with
tyrosine 128 and tyrosine 152. WIN52084, however, was bound in the
opposite orientation, with the isoxazole ring in the ‘‘toe.’’ Subsequently, it
was determined that only two additional compounds that were examined
were bound in the same orientation as WIN52084 (Fig. 4) [21]. This
observation led to the following conclusions:
Analogues with a seven-carbon chain connecting the phenyl and
isoxazole rings, and with a substituent on the oxazoline ring,
were bound with the isoxazole ring in the toe of the hydrophobic pocket.
All other analogues, regardless of the length of the connecting chain,
were bound in the opposite orientation.
A. The Nature of the Binding Site
The influence of substituents connected to the oxazoline ring on the binding orientation of these molecules was intriguing. Since the carbon to
Figure 5 Homologues of WIN52084 illustrating an entaniomeric effect. The
asymmetric center on the oxazoline ring is designated by asterisk.
283
which the alkyl substituents are attached is asymmetric, both enantiomers of WIN52084 as well as a homologous series of compounds were
evaluated against HRV-14 [22] (Fig. 5). In each case, the S isomer was
considerably more inhibitory than the R, which suggested an enantiomeric effect. Examination of WIN52084 in the pocket clearly showed that
the S-methyl group was in close proximity to a hydrophobic pocket
formed by Leu106 and Ser107 (Fig. 6).
To further analyze the interactions of the methyl group of the two
comformers in the binding site, an energy profiling study was performed.
With the x-ray crystal structure of the S isomer of WIN52084 in the virus
pocket serving as a starting point, a window consisting of all residues
within 8 A˚ of any atom was excised from the starting structure. After
charges had been set on the atoms of the resulting pocket and drug
according to a method in Chem-X [23], and after the hydrogen atoms
had been removed, the intermolecular van der Waals energy was calculated via a 6 –12 function for conformations resulting from the rotation of the oxazoline ring about the bond connected to the phenyl ring,
in increments of 10j. A plot of this function versus the rotation angle
Figure 6 WIN52084 bound to HRV-14. The methyl group on the oxazoline ring
is pointing toward a hydrophobic pocket formed by Leu106 and Ser107.
284
showed mainly two peaks at – 90j and 100j. A repeat of this calculation
with the R conformer resulted in a flat valley between 30j and approximately 120j. Similar results were obtained with the R- and S-ethyl
compounds, which showed an even more dramatic pattern that was
significant because the ethyl homologue was more potent (Fig. 7). These
results suggested that the twist angle about the two rings could be an
important factor in determining biological activity. It is possible that
the conformation with the appropriate twist angle may be imposed by
the nature of the binding pocket and that maximum interaction with the
hydrophobic pocket formed by Leu and Ser may also be of importance
[20 – 22].
B. Aliphatic Bridge
The x-ray studies on several analogues in this series of compounds showed
that the chain connecting the isoxazole and phenyl rings adopts a bowed
Figure 7 Plot of energy vs torsion angle from an energy profiling study resulting
from rotating the oxazoline ring of the S isomer of WIN52084 about the phenyl
ring.
285
conformation when bound to HRV-14. It had been assumed that flexibility
of the chain was critical for binding and biological activity. Dynamic
studies by Dr. Andrew McCammon with WIN52084 in HRV-14 revealed
considerable motion of the aliphatic chain during an observation lasting for 10 ps (Fig. 8) (Dr. Andrew McCammon, University of Houston,
personal communication). This result posed several questions regarding
the importance of flexibility vs rigidity of the chain. Would a conformationally rigid chain offer enhanced hydrophobic interactions and consequently improved binding, or are there other factors in the binding
process that would require a flexible chain? To address these issues, several
compounds with rigidity incorporated into the chain were synthesized;
their activity against HRV-14 and HRV-1A examined (Fig. 9) and the
compounds modeled in the respective binding site [24]. WIN54954, which
Figure 8 Molecular dynamics of WIN52084 in HRV-14 during a 10 ps run,
illustrating the movement of the chain. (Courtesy of Andrew McCammon, University of Houston.)
286
Figure 9 Table comparing the activity of the E and Z olefin and butyne analogues of WIN54954.
has been clinically evaluated, was used as a comparator. The Z olefin
demonstrated a two- to threefold reduction in activity in comparison to
WIN54954, while the E isomer showed a threefold enhancement in
activity. The potency of the butyne analogue was more than fourfold
greater than that of WIN54954 against HRV-14 and was comparable to
that of the E isomer.
287
C. Modeling of Conformationally Restricted
Analogues
The structures shown in Figure 9 were constructed using WIN54954 as a
template, since its x-ray conformation in HRV-14 had been determined.
The resulting structures were subjected to the Tripos force field (Maximin
2), using Sybyl version 5.41, with default settings. Rotatable bonds in the
alkyl ether chain were defined, and the structures were flexibly fitted to
WIN54954, in virus-bound conformation, for insertion into the HRV-14
binding site (Fig. 10). The optimized fitted structures were inserted into
each serotype by replacement of virus-bound WIN54954. Since the drugbound conformation of the virus binding site with several of the compounds had been determined, revealing only minor variations in compound structure, insertion of the modeled compounds into the binding site
configuration, derived from WIN54954, appeared reasonable.
Two interesting observations emerged from this study. The acetylene
analogue, which was more than fourfold more potent than WIN54954
Figure 10 Overlay of energy-minimized structures of the E and Z isomers and
WIN54954.
288
against HRV-14, was inactive against HRV-1A, and there was a dramatic
difference between the activities of the E and Z olefins against HRV-14.
Because x-ray studies had shown that the HRV-14 binding site was longer
than the 1A site, the results of this study supported the premise that the
activity is dependent on the length of the molecule. The butyne was
modeled in HRV-14, causing no serious steric interactions. However this
was not the case in HRV-1A, where the chlorine atom appeared to interact
with Ile125.
The difference in activity of the E and Z olefins against HRV-14 was
explained by examining the relatively low energy virus-bound conformations. The result of an overlay of WIN-54954 (based on x-ray crystallography data), minimize E- and Z-olefinic structures and the butyne
analogue, suggested that the E isomer showed a reasonable fit while the
Z isomer did not. Furthermore, when the Z isomer was inserted into the
HRV-14 pocket, unfavorable interactions occurred.
The very high minimal inhibitory concentration (MIC) values for the
Z isomer against HRV-14 and HRV-1A may reflect a slow kon in both
cases. The conformational space accessible to the isoxazole of the E and Z
olefins, the butyne, and the three-carbon chained homologue of WIN
54954 by conformational sweep graph and for the Z olefin disclosed a
significant inaccessible region of space, while the butyne, E olefin, and
alkane do not show this deficit. Consequently, binding to this site may be
dependent on conformational permissibility in this region that is required
for entry into the pocket. These results suggested that the activity of these
compounds against the two serotypes is strongly dependent on the
flexibility of the hydrocarbon chain and the ability of the molecule to fit
into the conformational space of both pockets.
III. PHENYL STACKING
Thus far all the compounds that were examined bound to HRV-14, with
the exceptions noted, are oriented with the phenyl ring in a stacking mode
with Tyr128 and Tyr152. Aromatic –aromatic interactions have been
shown to be quite common in protein – protein interactions [25 – 31],
and in many cases have displayed [32,33] an electrostatic component.
Furthermore, such interactions would be expected to contribute extensively to the binding energy [34]. To determine the nature of the aromatic
stacking interactions, an energy profiling study was performed by twist-
289
ing the phenyl ring about the carbon oxygen bond and examining a
number of parameters such as heat of formation and electronic energy.
Figure 11 shows the torsion angle between the oxazoline and phenyl rings
for each compound after energy minimization and flexible fitting. These
studies were performed with a variety of substituents in the position ortho
to the ether. It was anticipated that if electrostatics were involved in the
stacking of these rings, the correlation of the results of the profiling study
with antiviral activity should relate to the physical and electronic properties of the substituents. There was no correlation between energy maxima
or minima and size or electronic nature of the substituent, however, nor
do the results correlate with biological activity. We concluded that electrostatics play no part in the stacking; rather, the interactions appear to
be hydrophobic. In addition, these results suggest that a planar orientation of the phenyl rings is preferred. The lack of an electrostatic effect
associated with the phenyl –phenyl interactions may be due to the inability of the phenyl ring of these compounds to adapt a true end-surface
orientation as a result of space constraints within the pocket.
Figure 11 Torsion angle between the oxazoline and phenyl rings obtained from
minimized structures fitted to the x-ray structure of WIN54954.
290
IV. HYDROGEN BONDING
The initial x-ray study with disoxaril and WIN52084 in the binding site
revealed that asparagine 219 was within hydrogen-bonding distance of the
nitrogen of the isoxazole or oxazoline rings. Several pieces of information
suggested that this type of binding contributes negligibly if at all to the total
binding energy. Lau and Pettitt [35] examined whether the close approach
of the asparagine and isoxazole ring, which had been observed crystallographically, was indeed an attractive event. By selectively computing the
pairwise attraction of the hydrogen of the asparagine 219 and the nitrogen
of the isoxazole ring, which could conceivably be involved in the hydrogen
bond, and disregarding the contribution of this energy to the overall energy
of the system, the researchers were able to predict that the potential
hydrogen bond was inconsequential.
In addition to the computational studies that argued against the
existence of a hydrogen bond with Asn219, further evidence was obtained
by site-directed mutagenisis of the asparagine in question to an alanine
(Dr. Daniel C. Peaver, Sterling Winthrop Inc., personal communication).
Confirmation of the mutation was accomplished by sequencing. A comparison of the sensitivity of the mutant with the wild type showed that no
change in sensitivity had resulted from the removal of the hydrogen donor
potential. Consequently, these findings were in complete agreement with
the results reported by Lau and Pettitt.
Although the evidence presented strongly suggests the lack of contribution of Asn219 to the binding energy, examination of the x-ray result
of HRV-14-bound compounds revealed the presence of a water molecule in
the vicinity of the isoxazole ring and hydrogen-bonded to the backbone of
Leu106, Ser107, and Asn219 (Fig. 12) [36]. A similar hydrogen-bonding
network has been seen in HRV-50 (Dr. Vincent Giranda, Sterling Winthrop Inc., personal communication). This observation could shed some
light on the relative activity of other heterocyclic replacements for the
isoxazole ring.
A. HRV-14 Model Development
The extensive data generated from x-ray studies with HRV-14 permitted
the development of a model that could define the properties required of this
class of compounds for antiviral activity [37]. This model was dependent on
the orientation and x-ray conformational data for compounds bound to
the viral pocket. Some assumptions were made based on earlier results and
on rules generated for predicting compound orientation. For example, it
291
Figure 12 Hydrogen bonding network involved in the binding of WIN54954
analogues to HRV-14.
was assumed that all the compounds included in the study that had not
been examined by x-ray crystallography behaved in a predictable manner.
Compounds were divided into two groups of seven compounds each. One
group whose conformations were known (Fig. 13) demonstrated various
levels of activity against the virus. The second group consisted of inactive
compounds with related structures (Fig. 14). In the absence of conformational data for this group, one of the active compounds was used as a template for these compounds. A SYBYL (version 5.0) database was created.
All the structures were overlaid in the position found in the binding site
(Fig. 15). Volume maps were then calculated for the Boolean ‘‘union’’ of all
active and inactive compounds, which were then overlaid, and the excess
volume occupied by the inactive compounds, in comparison to the active
compound (Boolean minus), was calculated (Fig. 16). A similar procedure
was followed for the excess volume for the actives (Boolean plus). These
combined results revealed that inactive compounds displayed excessive
bulk around the phenyl ring. Although some bulk is desirable in this
area to enhance hydrophobic interactions, excessive bulk, which leads to
292
Figure 13 Compounds active against HRV-14, which were used in the development of the volume map model.
steric interactions, leads also to inactivity [38]. Conversely, space occupancy in the pore area of the binding site was found to contribute to good
biological activity.
To refine this model qualitatively, the binding of several of these
compounds was subjected to a CoMFA (Comparative Molecular Field
Analysis) [39]. This program examines electrostatic and steric parameters
293
Figure 14 Compounds inactive against HRV-14.
and through a partial-least-squares analysis determines the correlation of
these effects with biological activity. Either rigid structures or fixed
conformations are required to carry out the analysis. Eight compounds,
which had been used in the volume map study and whose binding
conformations were known, were employed. These compounds also had
a reasonable spread of activity against HRV-14. In addition, they offered
some degree of structural diversity. Van der Waals radii for atoms were
taken from a standard Tripos force field. Charges were calculated by the
AMI method by single-point calculations on the receptor-bound conformation of the drug molecule. Point charges on the hydrogen atoms were
not collapsed onto the atom to which they were bound but were left on
the hydrogen atoms. Log p values were calculated using the MedChem
software package (version 3.54). All these parameters, in addition to the
CoMFA field values designated by *, were used in the quantitative
structure – activity (QSAR) analysis (Fig. 17).
294
Figure 15
Overlay of active and inactive compounds.
Structural alignments were obtained from x-ray crystallographic
analysis. The backbone residues within 20 A˚ of any atom of the compound were included in this study. This effectively created a cube 20 A˚ on
a side, which was divided into grid points 1 A˚ apart. A hydrogen atom
and a proton as a probe were used to sample each grid point for both
electrostatic and steric effects. The data were tabulated and cross-validated
along with the physical parameters by means of a partial-least-squares
method, with the following results: good correlation of MIC with CoMFA
data and good predicative capabilities in the case of steric properties
(Fig. 18). No meaningful correlation was seen with electrostatic parameters, either taken in combination with steric factors or evaluated alone.
A regression analysis using all the values shown in Figure 17 revealed no
contribution of any parameters, other than the CoMFA field, to the activity of these compounds.
In addition to the QSAR data, which resulted from this program,
three-dimensional contour maps were generated for both steric and elec-
295
Figure 16 Overlay of volume maps of active compounds and inactive
compounds.
trostatic fields. Cutoffs were used to contour together points where the
correlations were highest and positive and those that were highest and
negative. Although the shapes of the maps coincide with the shape of the
pocket, the structure of the macromolecules was not part of the calculations. The visual results displayed by the contour maps qualitatively
agree with the QSAR results; that is, there is no significant correlation
between electrostatics and biological activity (Fig. 19), despite a strong
correlation between the steric fields and activity, as predicted. Although
a moderate positive effect was seen in the vicinity of the aromatic ring,
in general, this model predicts that excessive bulk in this area negatively
correlates with biological activity. These results are in agreement with the
conclusions empirically generated from the volume map study and also
confirm the lack of electrostatics involved in the phenyl–phenyl stacking
interactions, which had been observed earlier.
296
Figure 17
CoMFA coordinates.
B. Model Development Based on Small-Molecule,
Low-Energy Conformations
Thus far, the model development discussed has been one based on x-ray
conformations. Considering that there are over 100 serotypes, the discovery of broad spectrum antirhinovirus agents would require considerably
more three-dimensional virus structures. We have investigated the possibility of simply using energy-minimized small-molecule conformations
exclusive of the virus structures [40]. The compounds in question were
constructed in SYBYL and minimized by means of Maximin. By using a
template for spatial referencing (Fig. 20) that represented one of the more
potent compounds against HRV-14, it was possible to employ the program
Superimpose (SYBYL version 5.0) to algorithmically overlay the molecules based not on conformational similarity but rather on shape. Volume
maps were then constructed as already described. Figure 21 compares
the volume maps created by this method with those from x-ray structures
in HRV-14. The difference maps clearly show that within certain limits,
increasing chain length increases activity. Although there is a space-filling
requirement for activity, exceeding the appropriate distribution or extent
of bulk results in inactive compounds. These findings essentially duplicate
those obtained from the preceding method.
297
Figure 18 Regression equation obtained from partial-least-squares data crossvalidated in Figure 17.
This procedure was repeated for HRV-1A. Duplicate maps were
generated by means of both procedures and clearly show that shorter
molecules, as measured from the phenoxy to isoxazole moieties, are more
active. Molecules with the correct degree and placement of bulk in the
middle of the volume are also more active. These encouraging results
suggest that this method can be applied to other serotypes without giving
consideration to their three-dimensional structures.
C. Application of Model Development to Drug Design
The results of the model development for HRV-14 and HRV-1A demonstrated that the problem of drug design is complicated by the difference
in the dimensions of the binding sites, at least in the case of these two
serotypes. One solution to this problem is to prepare a compound that
298
Figure 19 Three-dimensional contour map generated from CoMFA analysis of
electrostatic fields showing correlation with antiviral activity. Red indicates a
strong correlation. The grid map is a result of an analysis using a probe atom (charge
0). All charges on the molecules were calculated using the AM1 Hamiltonian
without geometry optimization.
has some degree of flexibility and would be accommodated by the binding sites in both serotypes. We chose to examine the homologous series
shown in Figure 22. The three-carbon bridge structure demonstrated good
activity against HRV-1A but poor activity against HRV-14. As the flexible side chain is increased, a concomitant improvement in activity is seen
against HRV-14, with optimum activity observed against both serotypes
with the three-carbon side chain. This avenue was pursued because further testing of the three-carbon chained analogues against 100 rhinovirus
serotypes indicated that these compounds exhibited a broader spectrum
of activity, suggesting that perhaps HRV-14 is not representative of the
majority of serotypes.
299
Figure 20 Minimized structure serving as a template for spatial referencing in the
model development using energy-minimized small molecules.
D. The Development of a Clinical Candidate
Biological activity is not the only criterion required for drug development,
as anyone who has been involved in this area is aware. Potency, toxicity,
bioavailability, metabolic stability, and plasma half-life are only a few of
the critical issues that must be addressed. Although satisfactory potency
and spectrum activity had been achieved with WIN54954, which has been
clinically evaluated, this compound lacked metabolic stability and consequently displayed a short half-life.
It became clear that the oxazoline ring was metabolically unstable
and was responsible for the generation of crystalurea with disoxaril and for
a drug-induced rash with WIN54954, accompanied by a short plasma halflife. Consequently, a replacement for the oxazoline ring was sought, which
would be metabolically stable and would demonstrate satisfactory bioavailability. After examining several heterocyclic replacements, the 4-
300
Figure 21 Comparison of volume map for HRV-14 generated from x-ray data
(left) and small-molecule energy-minimized structures (right).
methyltetrazole analogue with a three-carbon linker (Fig. 23) appeared to
provide good chemical stability and improved biological activity in
comparison to WIN54954 [41]. However, when this compound was
administerd to dogs, hepatotoxicity was observed which was attributed
to metabolic instability. Further modifications resulted in the synthesis of
the 5-methyl 1,2,4-oxadiazole analogue, which was selected as a possible
Figure 22
Homologous series of compounds.
301
Figure 23 Structure of the 4-methyltetrazole analogues.
development candidate based on potency and spectrum of activity [42]. To
address metabolic stability, however, a monkey liver microsomal assay
was established by means of which the half-life, the extent of metabolism,
and the nature of the metabolic products could be determined [43]. Initially, WIN54954 was incubated at 37jC with a liver microsomal mixture
for 30 min and the incubate was extracted with hexane. The extracts were
analyzed by high performance liquid chromatography (HPLC), which
revealed 18 metabolic products. When the oxadiazole analogue was subjected to the same conditions, two major peaks, metabolites A and B, were
observed by HPLC (Fig. 24), in addition to six minor ones. The rate of
metabolism was similar to that of WIN54954, however, with a half-life of
27 vs 20 min.
302
Figure 24 HPLC spectrum resulting from the incubation of the oxadiazole with
monkey liver microsomes.
The question at this point was whether modifications could be made
to the oxadiazole molecule to enhance metabolic stability and achieve
comparable activity. This approach required knowledge of the site of
metabolism and the nature of the metabolic products. This information
was obtained from ion mass spectrometry. The identity of these products
was determined by comparing the fragmentation pattern of metabolites A
and B with the parent compound and the corresponding daughter ions
(Fig. 25).
Analysis of the metabolic products indicated that hydroxylation
occurred to a greater extent (30%) on the methyl group attached to the
isoxazole ring than to the methyl group on the oxadiazole ring (10%). The
methyl group in this postion was replaced with a trifluormethyl group to
prevent hydroxylation. The result of the incubation of this compound
indicated that although this position was protected, three metabolic
products were produced; in addition, the half-life was not substantially
different from the parent compound.
A similar replacement on the oxadiazole ring (Fig. 26) not only prevented metabolism at this position but also protected the entire mole-
Figure 25 Biotransformation of WIN 61893 and WIN 64172.
Figure 26 Metabolism of the trifluoromethylisoxazole analogue.
303
304
Figure 27
Structure of second properties pleconaril.
cule, resulting in two minor metabolites, and substantially increased the
half-life. In addition, pleconaril has exhibited a broad spectrum of antipicornavirus activity and has shown good bioavailability (Fig. 27) and is
undergoing clinical trials for upper respiratory rhinovirus infections.
V. CONCLUSIONS
x-Ray crystallography has added a new dimension to antirhinovirus drug
design. It has enabled us to examine the molecular interactions within
the compound binding site and to better understand the mechanism of
binding. We have been able to devise a model based on x-ray crystallography that qualitatively describes properties of molecules that are beneficial
for antirhinovirus activity. Also, by comparison to a volume map based
on x-ray conformations, we have developed a comparable model based on
small-molecule energy-minimized structures exclusive of x-ray data.
Finally, we have been able to apply our results to the synthesis of compounds active against both HRV-1A and HRV-14. Aside from the design
aspects, we have dealt with the more practical considerations such as
metabolic stability and bioavailability, which have led to a clinical candidate. Now certain unanswered mechanistic questions can be addressed.
How does the drug enter the binding site? Is there a recognition site, which
may explain some anomalous results that remain a mystery? Hopefully,
305
future work in this regard will eventually lead to an understanding of the
binding process and its relationship to biological activity.
REFERENCES
1. Gwaltney JM Jr. Principles and Practices of Infectious Diseases, 2nd ed.
New York: John Wiley & Sons, 1985.
2. Abraham G, Colonno RJ. Many rhinovirus serotypes share the same cellular receptor. J Virol 1984, 51:814 – 817.
3. Greve JM, Davis G, Meyer AM, et al. The major human rhinovirus receptor
is ICAM-1. Cell 1989; 56:839 – 847.
4. Staunton DE, Merluzzi VJ, Rothlein R, Barton R, Marlin SD, Springer TA.
A cell adhesion molecule ICAM-1 is the major surface receptor for rhinoviruses. Cell 1989; 56:849 – 853.
5. Uncapher GR, Dewitt CM, Colonno RJ. Human rhinovirus serotype families contain all but one human rhinovirus serotype. Virology 1991; 180:814 –
817.
6. Hofer F. Members of the low density lipoprotein receptor family mediate
cell entry of a minor group common cold virus. Proc Natl Acad Sci USA
1994; 91:1839 – 1842.
7. Otto MJ, Fox MP, Fancher MJ, Kuhrt MF, Diana GD, McKinlay MA. In
vitro activity of WIN 51711, a new broad spectrum antipicornavirus agent.
Antimicrob Agents Chemother 1985; 27:883 – 886.
8. Diana GD, McKinlay MA, Otto MJ, Akullian V, Oglesby CJ. [[(4,5-Dihydro-2-oxazolyl)phenoxy]alkyl]isoxazoles. Inhibitors of picornavirus uncoating. Med Chem 1985; 28:1906 – 1912.
9. Diana GD, McKinlay MA, Brisson CJ, Zalay ES, Miralles JV, Salvador UJ.
Isoxazoles with antipicornavirus activity. J Med Chem 1985; 28:748 – 752.
10. Diana GD, Otto MJ, McKinlay MA. Inhibitors of viral uncoating. Pharmacol Ther 1985; 24:287 – 297.
11. Diana GD, Oglesby RC, Akullian V, et al. Structure-activity studies of 5-[[4(4,4-dihydro-2-oxazolyl)phenoxy]alkyl]-3-methylisoxazoles: inhibitors of picornavirus uncoating. J Med Chem 1987; 30:383 – 388.
12. Fox MP, Otto MJ, Shave WJ, McKinlay MA. Prevention of rhinovirus and
poliovirus uncoating by WIN 51711, a new antiviral drug. Antimicrob
Agents Chemother 1986; 30:110 – 116.
13. Fox MP, McKinlay MA, Diana GD, Dutko FJ. Binding affinities of structurally related human rhinovirus capsid binding compounds are related to
their activities against human rhinovirus type 14. Antimicrob Agents Chemother 1991; 30:110 – 116.
14. Pevear DC, Fancher MJ, Felock, et al. Conformational changes in the floor
of the human rhinovirus canyon blocks adsorption to Hela cell receptors. J
Virol 1989; 63:2002 – 2007.
306
15. Rossmann MG, Arnold E, Erickson JW, Frankenberger EA, et al. Structure
of a human common cold virus and functional relationship to other picornaviruses. Nature 1985; 317:145 – 155.
16. Badger, Minor I, Kremer MJ, et al. Structural analysis of a series of antiviral
agents complexed with human rhinovirus 14. Proc Nat Acad Sci USA 1988;
85:3304 – 3308.
17. Oliviera MA, Zhao R, Lee WM, et al. The structure of human rhinovirus 16.
Structure 1993; 1:51 – 68.
18. Kim S, Smith TJ, Chapman MS. Crystal structure of human rhinovirus 1A
(HRV 1A). J Mol Biol 1989; 210:91 – 111.
19. Giranda VL. Structure-based drug design of antirhinoviral compounds.
Structure 1994; 2:695 – 698.
20. Smith TJ, Kremer MJ, Luo M, Vriend G, et al. The site of attachment in
human rhinovirus 14 for the antiviral agents that inhibit uncoating. Science
1986; 233:1286 – 1293.
21. Badger J, Minor I, Kremer MJ, et al. Structural analysis of a series of
antiviral agents complexed with human rhinovirus 14. Proc Natl Acad Sci
USA 1988; 85:3304 – 3308
22. Diana GD, Otto MJ, Treasurywala AM, et al. Enantiomeric effects of homologues of disoxaril on the inhibitory activity against human rhinivirus-14.
J Med Chem 1988; 31:540 – 544.
23. Chem-X, developed and distributed by Chemical Design Limited, Oxford,
England. Geistiger J. Marsili M. Iterative partial equalization of orbital
electronegativity – a rapid access to atomic charges. Tetrahedron 1980;
36:3219 – 3228.
24. Mallamo JP, Diana GD, Pevear et al. Conformationally restricted analogues of disoxaril. J Med Chem 1992; 35:4690 – 4695.
25. Burley SK, Petsko GA. Aromatic-aromatic interactions: a mechanism of
protein structure stabilization. Science 1985; 229:23 – 28.
26. Burley SK, Petsko GA. Weakly polar interactions in proteins. Protein Chem
1988; 39:125 – 192.
27. Levitt M, Perutz MF. Aromatic rings act as hydrogen bond receptors, J Mol
Biol 1988; 201:751 – 754.
28. Singh J, Thornton JM. The interaction between phenylalanine in proteins.
FEBS Lett 1985; 191:1 – 6.
29. Anderson DE, Hurley JH, Nicholson H, Baase WA, Mathews BW Hydrophobic core repacking and aromatic-aromatic interaction in the thermostable mutant of T4 lysozyme Ser 117-Phe. Protein Sci 1993; 2:1285 – 1290.
30. Karlstrom G, Linse P, Wallqvist A, Johnson B. Intermolecular potential for
the H2O-C6H6 and the C6H6-C6H6 system calculated in an ab initio SCF
CI approximation. J Am Chem Soc 1983; 105:3777 – 3782.
31. Fowler PW, Moore GJ. Calculation of the magnitude and orientation of
electrostatic interactions between small aromatic rings in peptides and pro-
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
307
teins: implications for angiotensin II. Biochem Biophys Res Commun 1988;
153:1296 – 1300.
Serrano L, Bycroft M, Fersht AR. Aromatic-aromatic interactions and protein stability. Investigation of double mutant cycles. J Mol Biol 1990;
218:465 – 475.
Pawliszyn J, Szczeskiak MM, Scheiner S. Interactions between aromatic
systems-dimers of benzene and S-triazine. J Phys Chem 1984; 88:1726 – 1730.
Weiner SJ, Kollman PA, Nguyen DT, Case DA. An all atom force field for
simulations of proteins and nucleic acids. J Comput Chem 1986; 7:230 – 252.
Lau WF, Pettitt MB. Selective elimination of interactions: a method for
assessing thermodynamic contributions to ligand binding with application
to rhinovirus antivirals. J Med Chem 1989; 32:2542 – 2547.
Giranda VL, Russo G, Felock P, Draper T, Diana G, Guiles J, Oglesby R,
Long M, Pevear DC. Submitted.
Diana GD, Treasurywala AM, Bailey TR, Oglesby RC, Pevear DC, Dutko
FJ. A model for compounds active against human rhinovirus-14 based on
X-ray crystallography data. J Med Chem 1989; 33:1306 – 1311.
Diana GD, Cutcliffe D, Oglesby RC, Otto MJ, Mallamo JP, Akullian V,
McKinlay MA. Synthesis and structure-activity studies of some disubstituted phenylisoxazoles against human picornavirus. J Med Chem 1989;
32:450 – 455.
Diana GD, Kowalczyk P, Treasurywala AM, Oglesby RC, Pevear DC,
Dutko FJ. CoMFA analysis of interactions of antipicornavirus compounds
in the binding pocket of human rhinovirus-14. J Med Chem 1992; 35:1002 –
1008.
Aldous DJ, Diana GD, Treasurywala AM, Jaeger EP, Pevear DC. Progress
in the development of an antirhinovirus model based on small molecule
conformations, Poster presentation 28 presented at the American Chemical
Society Meeting, San Diego, CA, 1994.
Woods MG, Diana GD, Rogge MC, Otto MJ, Dutko FJ, McKinlay MA.
In vitro activity and in vivo activities of WIN 54954, a new broad spectrum
antipicornavirus drug. Antimicrob Agents Chemother 1989; 33:2069 – 2074.
Diana GD, Volkots D, Nitz TJ, et al. Oxadiazoles as ester replacements in
compounds related to disoxaril: antirhinovirus activity. J Med Chem 1994;
37:2421 – 2436.
Diana GD, Rudewicz P, Pevear DC, et al. Picornavirus inhibitors: trifluoromethyl substitution provides a global protective effect against hepatic
metabolism. J Med Chem 1995; 38:1355 – 1371.
11
Profiles of Prototype Antiviral Agents
Interfering with the Initial Stages
of HIV Infection
Erik De Clercq
Rega Institute for Medical Research, Katholieke Universiteit Leuven,
Leuven, Belgium
I.
INTRODUCTION
The initial stages of the human immunodeficiency virus (HIV) infection
could be defined as the steps of the viral growth cycle that precede the
integration of the proviral DNA into the host cell genome. These stages
occur during the acute phase of the HIV infection, that is, when the virus
has invaded new cells. Once the proviral DNA has been integrated into
the host genome, the host cell and all its progeny cells can be considered to
be persistently or chronically infected. Expression of the integrated viral
genome will follow the classical flow of gene expression: that is, transcription, translation, and post-translational modifications under the concerted regulatory action of both cellular and viral factors.
This chapter reviews prototypes of single chemical entities that
interfere with the initial stages of the acute HIV infection, at steps that
are predominantly, if not solely, determined by specific viral proteins.
Thus, the compounds interacting with these steps in the HIV replicative
cycle may be expected to display a reasonably high specificity in their
mode of action. The targets that could be envisaged for such chemotherapeutic attack are the following: (1) virus adsorption, involving the
309
310
De Clercq
viral envelope glycoprotein gp120, (2) virus – cell fusion, involving both
viral glycoproteins gp120 and gp41, (3) viral uncoating, involving the viral capsid proteins, (4) the substrate (dNTP) binding site of the viral reverse
transcriptase, and (5) an allosteric non – substrate binding site of the HIV-1
reverse transcriptase. HIV inhibitors interacting with these targets have
been the subject of some earlier reviews [1 –4]. The chemokine receptors
CXCR4 and CCR5 used as coreceptors by X4 and R5 HIV-1 strains are
not discussed here; for reviews on inhibition of HIV infection by these receptor antagonists, see Refs. 5 and 6.
The compounds are highlighted from the following viewpoints: antiHIV potency and selectivity, mechanism of action, antiviral activity spectrum, clinical or therapeutic potential, and risk of resistance development.
II. VIRUS ADSORPTION INHIBITORS:
POLYANIONIC SUBSTANCES
Various polyanionic substances (viz., polysulfates, polysulfonates, polycarboxylates, and polyoxometalates) have been reported to block HIV
replication; for a review on the polysulfates, see Ref. 7. These substances
inhibit HIV-induced cytopathicity at a concentration of 0.1 to 1 Ag/mL,
while not being toxic to the host cells at concentrations up to 2 or 5 mg/mL,
thus achieving selectivity indexes of approximately 10,000 [7]. The target of
interaction for the polysulfates would be the V3 loop of the viral gp120
glycoprotein [8 –10]. This loop contains a highly basic region with which
the polyanionic substances could interact electrostatically. Thus, polyanions such as dextran sulfate may be assumed to block virus adsorption by
shielding the viral envelope glycoproteins [8]. Alternatively or additionally, polyanionic substances may also interact with the cellular CD4 receptor [11], thus preventing the viral envelope gp120 from anchoring to
the outer cell membrane.
Depending on their molecular weight, the nature of their anionic
groups, and the density/distribution of their negative charges, the polyanionic substances exhibit an activity spectrum that extends to several
enveloped viruses other than HIV: among the retroviruses, SIV (simian
immunodeficiency virus); among the herpesviruses, HSV (herpes simplex
virus) and CMV (cytomegalovirus); among the orthomyxoviruses, influenza A; among the paramyxoviruses, RSV (respiratory syncytial virus);
and toga-, flavi-, arena-, bunya-, and rhabdoviruses. Among the different
HIV strains, rather striking differences have been noted with regard to
Antiviral Agents in HIV Infection
311
susceptibility to polyanionic substances (e.g., dextran sulfate) [12], and this
differential susceptibility may be related to differences in the composition
of the viral glycoprotein portions with which the compounds interact.
Because of their broad activity spectrum, encompassing various
enveloped viruses, polyanionic substances may be of practical utility in
the prophylaxis and/or therapy of a number of important virus (e.g., HIV,
HSV, CMV, RSV, influenza A) infections. Yet, there is little, if any,
evidence for the in vivo efficacy of these compounds following either
parenteral or topical administration. Polyanions, and dextran sulfate in
particular, are poorly absorbed upon oral administration [13], and, in
addition, sulfated polysaccharides are notorious for their anticoagulant
activity. However, these problems can be overcome by the appropriate
chemical modifications (Fig. 1). Thus, h-cyclodextrin sulfate becomes
orally bioavailable following substitution of benzyl groups at either C-2
or C-6 of the sugar residues mCDS71 [14] and mCDS11 [15], respectively,
and heparin loses anticoagulant activity when acylated at the C-3 position
of the sugar rings [16]. These favorable features (oral bioavailability, loss of
anticoagulant activity) were obtained without impairing the anti-HIV
activity of the products (mCDS71, mCDS11, or O-acylated heparin).
The polyanionic substances may be expected to yield their greatest
promise when put in contact with the virus under the conditions that
Figure 1 Modified sulfated polysaccharides: (A) O-acylated heparin (m = 2, 4,
6, . . .) and (B) mCDS71 (a modified h-cyclodextrin sulfate).
312
De Clercq
mimic as closely as possible the in vitro situation, where these compounds
have proved so clearly effective. Further studies should address such issues
as delivery forms, route(s) of administration, and time of treatment (with
respect to the virus infection), before an appropriate candidate compound
is submitted to clinical trials. Polyanionic substances are not known to
lead to resistance development, although, as mentioned, different HIV
strains may differ markedly ab initio in their susceptibility to this class of
compounds.
III. VIRUS–CELL FUSION INHIBITORS: LECTINS,
ALBUMINS, AND TRITERPENE DERIVATIVES
Because of their interference with the interaction between the viral
envelope gp120 glycoprotein and the cellular CD4 receptor, polyanionic
substances not only inhibit virus adsorption to the cells but also block
syncytium (giant cell) formation between the HIV-infected (gp120+) cells
and uninfected (CD4+) cells. Since syncytium formation results in a
selective destruction of the CD4+ cells, this syncytium formation may
play an important role in the pathogenesis of AIDS (a hallmark of which is
a progressive decline of the CD4+ cells).
There are a number of compounds known to block syncytium
formation without (markedly) affecting virus binding to the cells. These
compounds may therefore be assumed to directly interfere with the virus –
cell fusion process, that is, fusion between the viral envelope and the outer
cell membrane. The compounds that have been postulated to inhibit
virus – cell fusion include the following: mannose-specific lectins (i.e., from
Listeria ovata, Hippeastrum hybrid, Cymbidium hybrid, and Epipactis
helleborine) and N-acetylglucosamine-specific plant lectins (i.e., from
Urtica dioica) [17,18]; a derivative from polyphemusin, a peptide that is
highly abundant in hemocyte debris of the horsehoe crab Limulus polyphemus [19]; succinylated human serum albumin, Suc-HSA [20], and
aconitylated human serum albumin, Aco-HSA [21] (Fig. 2); and triterpene (i.e., betulinic acid) derivatives (Fig. 3) [22]. Research into the efficacy
of a number of natural products that have been described as anti-HIV
agents is reviewed in Ref. 23.
Despite their widely varying origin and structure, these different
classes of compounds seem to be targeted at the virus –cell fusion process,
although the exact mechanism by which the compounds inhibit fusion,
313
Figure 2 Modified human serum albumins from HSA: succinylated human
serum albumin (Suc-HSA) and aconitylated human serum albumin (Aco-HSA).
and the target amino acid sequences (at gp120 and/or gp41) with which
they putatively interact, remain to be elucidated. Most of the compounds
inhibit HIV replication at concentrations of 0.1 to 1 Ag/mL and some
(Aco-HSA and betulinic acid) are even effective within the concentration
range of 0.01 to 0.1 Ag/mL [21,22]. For the plant lectins [17,18] and modified serum albumins [20,21], the inhibitory effects on HIV replication correlated closely with their inhibitory effects on syncytium formation, which
corroborates the hypothesis that their anti-HIV activity is due to inhibition
of virus –cell fusion.
Whereas the plant lectins are inhibitory to HIV-1, HIV-2, and a
number of other (enveloped) viruses (viz., CMV, RSV, influenza A), at
314
De Clercq
Figure 3 Triterpene derivative: betulinic acid RPR103611: N V-{N-[3h-hydroxylup-20(29-ene-28-oyl]-8-aminooctanoyl}-L-statine.
concentrations well below the cytotoxicity threshold, the succinylated and
aconitylated albumins inhibit only HIV, HIV-2 being (much) less susceptible to these compounds than HIV-1 [20,21]. Betulinic acid is even more
restricted in its antiviral activity spectrum in that it is active only against
HIV-1, and not even all HIV-1 strains [22]: specifically, betulinic acid is
not active against the NDK strain of HIV-1. This must point to a highly
specific molecular site for the interaction of betulinic acid [22] and should
help in deciphering the target amino acid sequences (at gp120 or gp41) for
this compound.
The clinical potential of the fusion inhibitors in the therapy and/or
prophylaxis of HIV infections remains a subject for further study. Since
these compounds directly interfere with syncytium formation, they should
be able to block HIV infections generated by both free virus particles and
HIV-infected cells. It is not known how readily the virus may become
resistant to this class of compounds. For betulinic acid, it has been
ascertained that some HIV-1 strains (e.g., NDK) may be resistant ab initio.
IV. VIRUS UNCOATING INHIBITORS: BICYCLAMS
Bicyclams (Fig. 4) consist of two cyclam (1,4,8,11-tetraazacyclotetradecane) units tethered via an aliphatic (i.e., propylene, as in JM2763) or
aromatic bridge [i.e., phenylenebis(methylene), as in JM3100]. While the
bicyclam JM2763 inhibits HIV-1 and HIV-2 replication at a concentration of 0.1 to 1 Ag/mL [24], the bicyclam JM3100 does so at a hundredfold lower concentration, that is, at a concentration that is more than
315
Figure 4 Bicyclam derivatives (A) JM2763 and (B) JM3100, each consisting of
two cyclam (1,4,8,11-tetraazacyclotetradecane) moieties tethered via a propylene (JM2763) or phenylenebis(methylene) bridge (JM3100).
100,000-fold lower than the cytotoxic concentration [25]. In primary T4
lymphocytes and monocytes (macrophages), JM3100 inhibits HIV-1 replication at concentrations lower than 1 ng/mL [25].
The bicyclams represent the only retrovirus inhibitors that have been
postulated to interfere with the viral uncoating process. This assumption
has been based on ‘‘time of addition’’ experiments where the compounds
(JM2763 and JM3100) were found to act at a stage following virus adsorption but preceding reverse transcription; and, since the compounds
did not prove inhibitory to syncytium formation (JM2763) [or were
inhibitory only at a concentration substantially higher than that required
for inhibition of HIV replication (JM3100)], their target of action could be
tentatively identified as a viral uncoating event. This hypothesis was then
corroborated by ‘‘uncoating’’ experiments in which the viral RNA,
recovered from HIV-infected cells that had been exposed to the compounds, was monitored for sensitivity to ribonuclease A: the viral RNA
was protected against degradation by RNase A, as could be anticipated if
the uncoating (i.e., dissociation of the capsid proteins from the viral RNA)
had been impeded [24,25].
Current investigations are attempting to determine with which viral
(capsid) proteins, and which amino acid residues of their target proteins,
the bicyclam interact. The antiviral activity spectrum of the bicyclams is
clearly different from that of other anti-HIV agents in that the bicyclams
are equally effective against HIV-1 and HIV-2 but less effective or not
active against SIV (in human T lymphocytes). Given its high selectivity
index ( >100,000) in vitro, the bicyclam JM3100 offers great potential for
the treatment of HIV-1 and HIV-2 infections in humans. Although the
bicyclams may under some conditions select out drug-resistant variants
from clinical HIV-1 (i.e., HE) strains, it has otherwise proved difficult, or
316
De Clercq
even impossible, to generate resistance to the compounds (i.e., JM3100)
after repeated passages in cell culture [25].
V. REVERSE TRANSCRIPTASE INHIBITORS
INTERACTING WITH THE SUBSTRATE
BINDING SITE
A. Dideoxynucleoside Analogs
In addition to the three anti-HIV agents [AZT (zidovudine), DDI (didanosine) and DDC (zalcitabine)] that have been formally approved by the
U.S. Food and Drug Administration for the treatment of HIV infections,
several other 2V,3V-dideoxynucleoside (ddN) analogues (Fig. 5), including
3V-fluoro-2V,3V-dideoxy-5-chlorouridine (FddClUrd) and 2V,3V-didehydro-
Figure 5 Dideoxynucleoside (ddN) analogues: 2V,3V-dideoxycytidine (DDC), 3Vazido-2V,3V-dideoxythymidine (AZT), 3V-fluoro-2V,3V-dideoxythymidine (FLT),
2V,3V-didehydro-2V,3V-dideoxythymidine (D4T), 3V-thia-2V,3V-dideoxycytidine
(3TC), 3V-thia-2V,3V-dideoxy-5-fluorocytidine (FTC), and 2V,3V-dideoxy-L-cytidine
(L-DDC).
317
2V,3V-dideoxythymidine [D4T (stavudine)], have been reported to inhibit
HIV replication (for review, see Refs. [26 – 28]). In particular, FddClUrd
appears to be an attractive candidate for further development, since it is
much less toxic to the host cells than AZT and most other ddN analogues
[29,30]. Also ranking among the most promising ddN analogues are 3Vthia-2V,3V-dideoxycytidine [3TC (lamivudine)] and 3V-thia-2V,3V-dideoxy5-fluorocytidine (FTC), which are actually more active in their (– )-h- or
L-isomeric form than in the (+)-h- or D-isomeric form [31,32].
All ddN analogues, including 3TC and FTC, act in a similar fashion;
that is, following intracellular phosphorylation to their 5V-triphosphate
form, they serve as competitive inhibitors/alternate substrates of the reverse transcriptase (RT) reaction, thus leading to chain termination, as
has been clearly demonstrated with AZT [33]. The anti-HIV activity of
ddN analogues is critically dependent on their intracellular phosphorylation, the first phosphorylation step being the most crucial. For some
compounds (viz., 2V,3V-dideoxyuridine) and in some cells (viz., monocytes/macrophages), the nucleoside kinase activity of the cells may be
inadequate to satisfactorily accomplish the first phosphorylation step; and
thus prodrugs, including aryl methoxyglycinyl derivatives [34] and bis[S(2-hydroxyethylsulfidyl)-2-thioethyl] esters [35] have been designed that
deliver the 5V-monophosphate form intracellularly, bypassing the first
phosphorylation step.
The antiviral activity spectrum of the ddN analogues should, in
principle, extend to all retroviruses as well as hepadnaviruses [i.e., hepatitis B virus (HBV)], since HBV, like retroviruses, replicates through an
RNA template-driven RT process. Indeed, various ddN analogues (particularly, the L-enantiomeric forms 3TC, FTC, and L-DDC) have been
shown to inhibit HBV replication [36 –38]. Consequently, 3TC is, at present, pursued as a potential drug candidate for the treatment of both HIV
and HBV infections.
Prolonged AZT therapy of HIV-infected individuals leads to a reduction of virus sensitivity to the drug [39]. This reduced sensitivity, generally termed ‘‘resistance,’’ appears to be based on the following mutations
in the HIV-1 RT [40,41]: 41 Met ! Leu, 67 Asp ! Asn, 70 Lys ! Arg,
215 Thr ! Phe/Tyr, and 219 Lys ! Gln. Of these mutations, the
215 Thr ! Tyr mutation has been the most frequently detected among
AZT-resistant HIV isolates from patients under prolonged AZT therapy
[42]. The 74 Leu ! Val mutation is responsible for resistance to DDI [43],
and the 184 Met ! Val mutation confers resistance to 3TC, FTC, DDC,
and DDI [44 – 46].
318
De Clercq
The mutations at position 74 (Leu ! Val) and 184 (Met ! Val) of
the HIV-1 RT do not lead to cross-resistance to AZT. Nor would the 215
Thr ! Tyr mutation lead to cross-resistance to 3TC, FTC, DDC, or DDI.
In fact, the mutations at positions 74 and 215 seem to counteract each
other, and so do the mutations at positions 184 and 215. Based on this
‘‘mutually counteracting mutation’’ principle [47], drug combinations
could be envisaged that, if combined, might counteract emergence of resistance to one another: namely, combinations of AZT with either DDI,
3TC, FTC, or DDC. As will be explained further, these two-drug combinations may be extended to three-drug or four-drug combinations, following the addition of one or more of the HIV-1-specific nonnucleoside RT
inhibitors (NNRTIs).
B. Acyclic Nucleoside Phosphonates
Acyclic nucleoside phosphonates (ANPs) (Fig. 6) may be regarded as
analogous to the ddN monophosphates, thus allowing us to circumvent the
first phosphorylation step required for the intracellular activation of the
compounds. After they have been taken up as such by the cells, the acyclic nucleoside phosphonates (PMEA, PMEDAP, PMPA, PMPDAP,
FPMPA, and FPMPDAP) are converted intracellularly to their respective
diphosphate form (PMEApp, PMEDAPpp, PMPApp, PMPDAPpp,
FPMPApp, and FPMPDAPpp) and, in such form they interact as competitive inhibitors, alternate substrates, or chain terminators with the reverse transcriptase [48 –50].
PMEA and its congeners are more effective in vivo than could be
predicted from their in vitro potency. While less potent as an antiretrovirus agent than AZT in vitro, PMEA proved clearly superior to AZT
when the two drugs were compared for their effectiveness in vivo, in mice
infected with murine Moloney sarcoma virus [51,52]. PMEA was also
shown to be effective against various other retrovirus infections, including
Friend leukemia virus (FLV), Rauscher leukemia virus (RLV), and LPBM5 (murine AIDS) virus infection in mice, feline leukemia virus (FeLV)
or feline immunodeficiency virus (FIV) infection in cats, and SIV infection
in macaque (rhesus) monkeys (for review, see Ref. 53). In the latter model
[54], again PMEA proved far superior to AZT in suppressing several
parameters of the disease.
The antiviral activity spectrum of PMEA, PMEDAP, and their
congeners is not confined to retroviruses but also extends to hepadnaviruses (e.g., HBV). PMEA has proved effective against duck HBV infec-
319
Figure 6 Acyclic nucleoside phosphonates (ANPs): 9-(2-phosphonylmethoxyethyl)-adenine (PMEA) and -2,6-diaminopurine (PMEDAP), (R )-9-(2-phosphonylmethoxypropyl)-adenine (PMPA) and -2,6-diaminopurine (PMPDAP), (S )-9-(3fluoro-2-phosphonylmethoxypropyl)-adenine (FPMPA) and -2,6-diaminopurine
(FPMPDAP), and the bis(pivaloyloxymethyl) ester of PMEA [Bis(pom)-PMEA].
tion in both duck hepatocytes and Pekin ducks [55]. For PMEA and
PMEDAP, but not for PMPA, PMPDAP, FPMPA, or FPMPDAP, the
activity spectrum also extends to herpesviruses (e.g., HSV, CMV). This
would make PMEA and PMEDAP particularly attractive as therapeutic
modalities in AIDS patients, since they might be useful not only for the
treatment of the underlying HIV infection but also for the therapy/
prophylaxis of the intercurrent HSV or CMV infections.
320
De Clercq
Another attractive feature of PMEA, and, in fact, all ANPs, is
prolonged antiviral action, lasting for several days, or even one week or
longer, after a single-dose administration. This long-lasting antiviral
action may be related to the long half-life of the active metabolites (e.g.,
PMEApp) within the cells and may permit infrequent (e.g., weekly) dosing
of the ANPs in the prophylaxis and/or therapy of (retro)virus infections.
Little is known on how readily or rapidly retro- or herpesviruses may
develop resistance to the ANPs. In the in vitro and in vivo experiments
done so far with PMEA, PMEDAP, or any of the other ANPs, resistance
development did not seem to occur, but further studies are needed to
address this issue.
Since the ANPs are only slowly taken up by the cells and poorly
absorbed following oral administration, some efforts have been directed
toward the development of prodrugs (esters) that would be better taken up
by the cells. These efforts have yielded the bispivaloyloxymethyl [bis(pom)]
derivative of PMEA (Fig. 6) [56]. Bis(pom)-PMEA shows a cellular uptake increased more than a hundredfold, as well as fivefold better oral
bioavailability than the parent compound [57]. Both PMEA (given intravenously) and bis(pom)-PMEA (given perorally) are now in clinical trials
in patients with AIDS.
VI. REVERSE TRANSCRIPTASE INHIBITORS
INTERACTING WITH A NONSUBSTRATE
BINDING SITE: NON-NUCLEOSIDE REVERSE
TRANSCRIPTASE INHIBITORS
The identification of the HIV-1-specific non-nucleoside reverse transcriptase inhibitors (NNRTIs) as a separate class of HIV inhibitors was
heralded by the discovery of the tetrahydroimidazo[4,5,1-jk][1,4]benzodiazepin-2(1H )-one and -thione (TIBO) derivatives (Fig. 7) [58,59] and
1-(2-hydroxyethoxymethyl)-6-(phenylthio)thymine (HEPT) derivatives
(Fig. 8) [60,61]. The first TIBO derivatives (R82150, R82913) were the
first NNRTIs [58] postulated to act as inhibitors of HIV-1 RT [59]. For the
HEPT derivatives it became evident that they also interact specifically
with HIV-1 RT after a number of derivatives (i.e., E-EPU, E-EBU, and
E-EBU-dM) had been synthesized that were more active than HEPT
itself [62,63]. Following HEPT and TIBO, several other compounds, i.e.,
nevirapine, pyridinone, and bis(heteroaryl)piperazine (BHAP), were
321
Figure 7 Tetrahydroimidazo[4,5,1-jk][1,4]benzodiazepin-2(1H )-one (TIBO) derivatives (A) R82913 and (B) R86183 (with a chlorine substituted in the 9- or 8position, respectively).
Figure 8 (A) 1-(2-Hydroxyethoxymethyl)-6-(phenylthio)thymine (HEPT). (B) 5Isopropyl-1-(ethoxymethyl)-6-benzyluracil (I-EBU, MKC-442).
322
De Clercq
described as HIV-1-specific RT inhibitors (for a review on the HIV-1specific RT inhibitors, see Refs. 28 and 64).
The HEPT and TIBO derivatives were discovered as the result of a
systematic evaluation for anti-HIV activity in cell culture. They were later
found to achieve their anti-HIV-1 activity through an interaction with the
HIV-1 RT. In contrast, nevirapine, pyridinone, and BHAP emerged from
a screening program for HIV-1 RT inhibitors. The anti-HIV-1 activity
of these compounds was subsequently confirmed in cell culture. Like the
HEPT and TIBO derivatives, the 2V,5V-bis-O-(tert-butyldimethylsilyl)-3Vspiro-5VV-(4VV-amino-1VV,2VV-oxathiole-2VV,2VV-dioxide)-pyrimidine (TSAO) derivatives (Fig. 9) [65,66] and a-anilinophenylacetamides (a-APA) (Fig. 10)
[67] were discovered through the evaluation of their anti-HIV activity in
cell culture. Subsequently, they were found to act as specific inhibitors of
HIV-1 RT.
Yet other compounds have been found to inhibit HIV-1 replication
through a specific interaction with HIV-1 RT (i.e., quinoxaline S-2720 [68],
5-chloro-3-(phenylsulfonyl)indole-2-carboxamide [69], dihydrothiazoloisoindolones [70] and a number of natural substances (e.g., calanolide A
and inophyllums, from the tropical rain forest trees Calophyllum lanigerum
and Calophyllum inophyllum, respectively) [71,72]. All these and yet other
compounds could be considered to be NNRTIs. The most potent among
the NNRTIs, some of the HEPT derivatives (E-EBU-dM) [63] and a-
Figure 9 2V,5V-Bis-O-(tert-butyldimethylsilyl)-3V-spiro-5W-(4W-amino-1W,2W-oxathiole-2W,2W-dioxide)pyrimidine (TSAO) derivatives TSAO-T, TSAO-m3T, and
TSAO-e3T.
323
Figure 10 a-Anilinophenylacetamide (a-APA) derivatives (A) R18893, (B)
R88703, and (C) R89439.
APA derivatives (R89439) [67], inhibit HIV-1 replication at a concentration of approximately 1 ng/mL, that is, 100,000-fold below the cytotoxicity
threshold.
While the ddNs and ANPs must be converted intracellularly to
their 5V-triphosphates (ddNTPs) or diphosphate derivatives before they
can interact as competitive inhibitors/alternate substrates with regard to
the natural substrates (dNTPs), the NNRTIs do not need any metabolic
conversion to interact, noncompetitively with respect to the dNTPs, at
an allosteric, non –substrate binding site of the HIV-1 RT. Through the
analysis of NNRTI-resistant mutants, combined with site-directed mutagenesis studies, it has become increasingly clear which amino acid
residues are involved in the interaction of the NNRTIs with HIV-1 RT,
and, since the conformation of the HIV-1 RT has been resolved at 3.0
A˚ resolution [73], it is now possible to visualize the binding site of the
NNRTIs [74].
The antiviral activity spectrum of the NNRTIs is limited to HIV-1,
probably because only HIV-1 RT contains a pocket site at which the
324
De Clercq
NNRTIs may bind. The high specificity displayed by the NNRTIs in
their binding to HIV-1 RT signals that it should, a priori, be relatively
easy for the enzyme (and the virus) to escape the inhibitory effects of the
NNRTIs through mutations of the amino acid residues that either are
directly involved in the binding of the NNRTIs or contribute to the configuration of the pocket that is ideal for NNRTI binding.
From pilot studies carried out in the clinic with the NNRTIs TIBO
R82913 [75] and pyridinone L-697,661 [76], it appears that the compounds
are well tolerated and do not cause toxic side effects. Most of the HIV-1
isolates obtained from the patients treated with TIBO R82913 appeared to
be as sensitive to the compound as wild-type virus; only two HIV-1
variants were isolated, showing a sensitivity that was reduced 20-fold or
more than 100-fold, the latter being caused by a mutation (Tyr ! Leu) at
position 188 of the RT [77]. In fact, the latter mutation was lost upon
passaging the virus in vitro in cord blood lymphocytes. Following treatment of the patients with pyridinone L-697,661, drug-resistant HIV-1
variants appeared that contained mutations at the RT positions 103 (Lys
! Asn) and 181 (Tyr ! Cys) [76].
HIV-1 resistance to NNRTIs rapidly arises following passage of the
virus in cell culture in the presence of the compounds. The 181 Tyr ! Cys
mutation is most commonly seen, and it leads to resistance, or at least to
reduced sensitivity, to most of the NNRTIs (i.e., TIBO, HEPT, nevirapine, pyridinone, BHAP, TSAO, a-APA) [78 – 84]. The 188 Tyr ! His mutation is associated with resistance to TIBO [85], but not nevirapine [82].
The 103 Lys ! Asn mutation is associated mainly with resistance to TIBO
and pyridinone [78,85]. The 100 Leu ! Ile mutation is associated mainly
with resistance to TIBO [85,86]. The 106 Val ! Ala mutation mainly leads
to resistance to nevirapine and HEPT [83,84,87]. The 138 Glu ! Lys
mutation is responsible for resistance to TSAO [88,89]. The 190 Gly ! Glu
mutation accounts for resistance to quinoxaline [68], while also leading to a
dramatic reduction in RT activity [90]; and the 236 Pro ! Leu mutation is
responsible for resistance to BHAP [91].
The rapid emergence of drug-resistant HIV-1 mutants under selective
pressure of the HIV-1-specific RT inhibitors has been generally viewed as a
limitation for, if not an argument against, the clinical usefulness of these
compounds. Yet, several aspects of virus – drug resistance, particularly
with respect to the NNRTIs, remain to be addressed before the problem
of resistance can be fully assessed. For example, how pathogenic are drugresistant variants in comparison to wild-type virus? How readily are such
drug-resistant variants transmitted from one person to another? Do virus-
325
resistant variants persist when the drug is withdrawn, or do they readily
revert to the wild type?
Assuming that the development of drug resistance may indeed
compromise the clinical usefulness of the NNRTIs, how might this
problem be prevented or circumvented? If resistance develops to one of
the NNRTIs, treatment could be switched to any of the other NNRTIs to
which the virus has retained sensitivity. For example, 5-chloro-3-(phenylsulfonyl)indole-2-carboxamide [69] is active against the HIV-1 strains
that, because of the 103 Lys ! Asn mutation or 181 Tyr ! Cys mutation,
have acquired resistance to various other NNRTIs (i.e., TIBO, nevirapine,
pyridinone, BHAP). The a-APA derivative R89439 [67] is active against
the 100 Leu ! Ile mutant, which is resistant to the TIBO derivatives
R82913 and R86183. Within the TIBO class, a minor chemical modification, the shifting of the chlorine atom from the 9-position (R82913) to the
8-position (R86183), suffices to restore activity against the 181 Tyr ! Cys
mutant [92]. Similarly, pyridinone L-702,019, which differs from its
predecessor L-696,229 only by the addition of two chlorine atoms (in
the benzene ring) and substitution of sulfur for oxygen (in the pyridine
ring), remains remarkably active against HIV-1 mutants containing the
103 Lys ! Asn or 181 Tyr ! Cys mutation [93]. In some instances
resistance to one of the NNRTIs may even be accompanied by hypersensitivity to others: the 236 Pro ! Leu mutation, which causes resistance
to BHAP, confers 10-fold increased sensitivity to TIBO, nevirapine, and
pyridinone [91].
The 181 Tyr ! Cys mutation, which is responsible for resistance to
most NNRTIs, has been found to suppress the 215 mutation (Thr ! Phe/
Tyr), which is responsible for resistance to AZT [94], and, vice versa, the
181 Tyr ! Cys mutation can be suppressed by AZT, which thus means that
the mutations at positions 181 and 215 counteract each other. Yet other
mutations have proved to counteract each other: 236 Pro ! Leu vs 138 Glu
! Lys, and, as mentioned, 215 Thr ! Phe/Tyr vs 184 Met ! Val, and 215
Thr ! Phe/Tyr vs 74 Leu ! Val [47]. Based on the resistance mutations
that counteract each other, combinations of different drugs could be
envisaged—namely, combinations of AZT with either TIBO, a-APA,
HEPT, nevirapine, or pyridinone—and these two drug combinations
could be extended to three- or four-drug combinations by the addition of
another ddN analogue (such as 3TC) and/or another NNRTI (such as
BHAP or TSAO).
What would seem to be an attractive approach to the prevention of
resistance development is the ‘‘knocking-out’’ strategy [95]. If NNRTIs,
326
De Clercq
such as BHAP (U-88204 or U-90152), are used from the start at a
sufficiently high concentration (i.e., 1 or 3 AM, respectively), they completely suppress virus replication [96,97], with the result that the virus is
‘‘knocked out’’ and does not have the opportunity to become resistant. If
U-90152 is combined with AZT, the concentrations of the individual drugs
can be lowered to achieve total virus clearance [97]. Five NNRTIs (TIBO,
HEPT, nevirapine, pyridinone, and BHAP) have been shown to ‘‘knock
out’’ HIV-1 in cell culture when used at concentrations (1 –10 Ag/mL) that
are nontoxic to the cells [95]. That the virus was really knocked out, and
thus the cell culture cleared (‘‘sterilized’’) from the HIV-1 infection by the
NNRTIs, was ascertained by two successive rounds of 35-cycle PCR
(polymerase chain reaction) analysis, which failed to reveal the presence
of any proviral DNA [95]. Thus, when used at ‘‘knocking-out’’ concentrations, the NNRTIs may be expected to effect a long-lasting suppression
of HIV-1 replication. This ‘‘knocking-out’’ phenomenon could be obtained at lower concentrations if the NNRTIs were combined with each
other, or with any of the ddN analogues (i.e., AZT), particularly if selected
on the basis of the ‘‘mutually counteracting mutation’’ principle.
VII. CONCLUSION
An acute HIV infection can be blocked at any of the following stages of
the infection: virus adsorption, virus – cell fusion, viral uncoating, and reverse transcription. At the reverse transcriptase (RT) level, chemotherapeutic intervention could be envisaged at either the substrate or a non –
substrate binding site. Polyanionic substances (i.e., sulfated polysaccharides) prevent virus adsorption; plant lectins, succinylated (or aconitylated) albumins, and triterpene (i.e., betulinic acid) derivatives interfere
with virus – cell fusion; bicyclams inhibit viral uncoating; 2V,3V-dideoxynucleosides (ddNs) and acyclic nucleoside phosphonate analogues, following
intracellular conversion to their phosphorylated derivatives, interact with
the substrate binding site of the RT; and the nonnucleoside reverse transcriptase inhibitors (NNRTIs) are targeted at a non –substrate binding site
of HIV-1 RT. Some of these compounds (viz., bicyclams) and, among the
NNRTIs, some of the HEPT and a-APA derivatives, were found to inhibit
HIV-1 replication at concentrations (f1 ng/mL) that were 100,000-fold
or more below the cytotoxicity threshold. As a rule, it may be postulated
that the more specific the antiviral action, the more likely the development of virus – drug resistance; hence, NNRTIs, which engage in a highly
327
specific interaction with HIV-1 RT, rapidly lead to the emergence of drugresistant virus strains. To prevent such drug-resistant virus strains from
emerging, several strategies could be envisaged, the most attractive being
the combination of several drugs at concentrations high enough to ‘‘knock
out’’ the virus from the start. This ‘‘knocking-out’’ phenomenon has been
achieved with the NNRTIs, regardless of whether combined with any of
the ddN analogues, and it may be extended to combinations of drugs that
interact at targets other than the reverse transcriptase.
ACKNOWLEDGMENTS
The original investigations of the author are supported by the Biomedical Research Programme of the European Community, the Belgian
Nationaal Fonds voor Wetenschappelijk Onderzoek, the Belgian Fonds
voor Geneeskundig Wetenschappelijk Onderzoek, the Belgian Geconcerteerde Onderzoeksacties, and the Janssen Research Foundation. I
thank Christiane Callebaut for her dedicated editorial assistance.
REFERENCES
1. De Clercq E. New perspectives for the chemotherapy and chemoprophylaxis
of AIDS (acquired immune deficiency syndrome). Verh K Acad Geneeskd
Belg 1992; 54:57 – 89.
2. De Clercq E. Anti-HIV agents interfering with the initial stages of the HIV
replicative cycle. In: Morrow WJW, Haigwood NL, eds. HIV Molecular
Organization, Pathogenicity and Treatment. Amsterdam: Elsevier Science
Publishers, 1993:267 – 292.
3. De Clercq E. The emerging role of fusion inhibitors in HIV infection. Drug
Dev Res 1999; 2:321 – 331.
4. De Clercq E. Inhibition of HIV infection by bicyclams, highly potent and
specific CXCR4 antagonists. Mol Pharmacol 2000; 57:833 – 839.
5. De Clercq E, Schols D. Inhibition of HIV infection by CXCR4 and CCR5
chemokine receptor antagonists. Antiviral Chem Chemother 2001; 12 (suppl
1):19 – 31.
6. De Clercq E. New developments in anti-HIV chemotherapy. Curr Med
Chem 2001; 8:1543 – 1572.
7. De Clercq E. Anti-HIV activity of sulfated polysaccharides. In: Yalpani M,
ed. Carbohydrates and Carbohydrate Polymers, Analysis, Biotechnology,
328
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
De Clercq
Modification, Antiviral, Biomedical and Other Applications. Mount Prospect, IL: ATL Press, 1993:87 – 100.
Schols D, Pauwels R, Desmyter J, De Clercq E. Presence of class II histocompatibility DR proteins on the envelope of human immunodeficiency
virus demonstrated by FACS analysis. Virology 1992; 189:374 – 376.
Callahan LN, Phelan M, Mallinson M, Norcross MA. Dextran sulfate
blocks antibody binding to the principal neutralizing domain of human
immunodeficiency virus type 1 without interfering with gp120-CD4 interactions. J Virol 1991; 65:1543 – 1550.
Batinic D, Robey FA. The V3 region of the envelope glycoprotein of human
immunodeficiency virus type 1 binds sulfated polysaccharides and CD4derived synthetic peptides. J Biol Chem 1992; 267:6664 – 6671.
Schols D, Baba M, Pauwels R, Desmyter J, De Clercq E. Specific interaction
of aurintricarboxylic acid with the human immunodeficiency virus/CD4 cell
receptor. Proc Natl Acad Sci USA 1989; 86:3322 – 3326.
Schols D, Pauwels R, Witvrouw M, Desmyter J, De Clercq E. Differential
activity of polyanionic compounds and castanospermine against HIV replication and HIV-induced syncytium formation depending on virus strain and
cell type. Antiviral Chem Chemother 1992; 3:23 – 29.
Lorentsen KJ, Hendrix CW, Collins JM, Kornhauser DM, Petty BG,
Klecker RW, Flexner C, Eckel RH, Lietman PS. Dextran sulfate is poorly
absorbed after oral administration. Ann Intern Med 1989; 111:561 – 566.
Moriya T, Saito K, Kurita H, Matsumoto K, Otake T, Mori H, Morimoto
M, Ueba N, Kunita N. A new candidate for an anti-HIV-1 agent: modified
cyclodextrin sulfate (mCDS71). J Med Chem 1993; 36:1674 – 1677.
Otake T, Schols D, Witvrouw M, Naesens L, Nakashima H, Moriya T,
Kurita H, Matsumoto K, Ueba N, De Clercq E. Modified cyclodextrin
sulphates (mCDS11) have potent inhibitory activity against HIV and high
oral bioavailability. Antiviral Chem Chemother 1994; 5:155 – 161.
Baˆrzu T, Level M, Petitou M, Lormeau J-C, J. Choay J, Schols D, Baba M,
Pauwels R, Witvrouw M, De Clercq E. Preparation and anti-HIV activity of
O-acylated heparin and dermatan sulfate derivatives with low anticoagulant
effect. J Med Chem 1993; 36:3546 – 3555.
Balzarini J, Schols D, Neyts J, Van Damme E, Peumans W, De Clercq E. a(1-3)- and a-(1-6)-d-mannose-specific plant lectins are markedly inhibitory
to human immunodeficiency virus and cytomegalovirus infections in vitro.
Balzarini J, Neyts J, Schols D, Hosoya M, Van Damme E, Peumans W, De
Clercq E. The mannose-specific plant lectins from Cymbidium hybrid and
Epipactis helleborine and the (N-acetylglucosamine)n-specific plant lectin
from Urtica dioica are potent and selective inhibitors of human immunodeficiency virus and cytomegalovirus replication in vitro. Antiviral Res 1992;
18:191 – 207.
329
19. Nakashima H, Masuda M, Murakami T, Koyanagi Y, Matsumoto A, Fujii
N, Yamamoto N. Anti – human immunodeficiency virus activity of a novel
synthetic peptide, T22 ([Tyr-5,12,Lys-7]polyphemusin II): a possible inhibitor of virus – cell fusion. Antimicrob Agents Chemother 1992; 36:1249 –
1255.
20. Jansen RW, Molema G, Pauwels R, Schols D, De Clercq E, Meijer DKF.
Potent in vitro anti – human immunodeficiency virus-1 activity of modified
human serum albumins. Mol Pharmacol 1991; 39:818 – 823.
21. Jansen RW, Schols D, Pauwels R, De Clercq E, Meijer DKF. Novel, negatively charged, human serum albumins display potent and selective in vitro
anti – human immunodeficiency virus type 1 activity. Mol Pharmacol 1993;
44:1003 – 1007.
22. Mayaux J-F, Bousseau A, Pauwels R, Huet T, Heńin Y, Dereu N, Evers M,
Soler F, Poujade C, De Clercq E, Le Pecq J-B. Triterpene derivatives that
block entry of human immunodeficiency virus type 1 into cells. Proc Natl
Acad Sci USA 1994; 91:3564 – 3568.
23. De Clercq E. Current lead natural products for the chemotherapy of human
immunodeficiency virus (HIV) infection. Med Res Rev 2000; 20:323 – 349.
24. De Clercq E, Yamamoto N, Pauwels R, Baba M, Schols D, Nakashima H,
Balzarini J, Debyser Z, Murrer BA, Schwartz D, Thornton D, Bridger G,
Fricker S, Henson G, Abrams M, Picker D. Potent and selective inhibition
of human immunodeficiency virus (HIV)-1 and HIV-2 replication by a class
of bicyclams interacting with a viral uncoating event. Proc Natl Acad Sci
USA 1992; 89:5286 – 5290.
25. De Clercq E, Yamamoto N, Pauwels R, Balzarini J, Witvrouw M, De Vreese
K, Debyser Z, Rosenwirth B, Peichl P, Datema R, Thornton D, Skerlj R,
Gaul F, Padmanabhan S, Bridger G, Henson G, Abrams M. Highly potent
and selective inhibition of human immunodeficiency virus by the bicylam
derivative JM3100. Antimicrob Agents Chemother 1994; 38:668 – 674.
26. De Clercq E. HIV inhibitors targeted at the reverse transcriptase. AIDS Res
Human Retrovir 1992; 8:119 – 134.
27. De Clercq E. Hamao Umezawa Memorial Award Lecture: An odyssey in
the viral chemotherapy field. Int J Antimicrob Agents 2001; 18:309 – 328.
28. De Clercq E. New developments in anti-HIV chemotherapy. Farmaco 2001;
56:3 – 12.
29. Balzarini J, Van Aerschot A, Pauwels R, Baba M, Schols D, Herdewijn P,
De Clercq E. 5-Halogeno-3V-fluoro-2V,3V-dideoxyuridines as inhibitors of human immunodeficiency virus (HIV): potent and selective anti-HIV activity of
3V-fluoro-2V,3V-dideoxy-5-chlorouridine. Mol Pharmacol 1989; 35:571 – 577.
30. Van Aerschot A, Herdewijn P, Balzarini J, Pauwels R, De Clercq E. 3VFluoro-2V,3V-dideoxy-5-chlorouridine: most selective anti-HIV-1 agent
among a series of new 2V- and 3V-fluorinated 2V,3V-dideoxynucleoside analogues. J Med Chem 1989; 32:1743 – 1749.
330
De Clercq
31. Schinazi RF, Chu CK, Peck A, McMillan A, Mathis R, Cannon D, Jeong
L-S, Beach JW, Choi W-B, Yeola S, Liotta DC. Activities of the four optical
isomers of 2V,3V-dideoxy-3V-thiacytidine (BCH-189) against human immunodeficiency virus type 1 in human lymphocytes. Antimicrob Agents Chemother 1992; 36:672 – 676.
32. Schinazi RF, McMillan A, Cannon D, Mathis R, Lloyd RM, Peck A,
Sommadossi J-P, St Clair M, Wilson J, Furman PA, Painter G, Choi
W-B, Liotta DC. Selective inhibition of human immunodeficiency viruses
by racemates and enantiomers of cis-5-fluoro-1-[2-(hydroxymethyl)-1,3-oxathiolan-5-yl]cytosine. Antimicrob Agents Chemother 1992; 36:2423 – 2431.
33. Huang P, Farquhar D, Plunkett W. Selective action of 3V-azido-3V-deoxythymidine 5V-triphosphate on viral reverse transcriptases and human DNA
polymerases. J Biol Chem 1990; 265:11914 – 11918.
34. McGuigan C, Pathirana RN, Balzarini J, De Clercq E. Intracellular delivery
of bioactive AZT nucleotides by aryl phosphate derivatives of AZT. J Med
Chem 1993; 36:1048 – 1052.
35. Puech F, Gosselin G, Lefebvre I, Pompon A, Aubertin A-M, Kirn A, Imbach J-L. Intracellular delivery of nucleoside monophosphates through a
reductase-mediated activation process. Antiviral Res 1993; 22:155 – 174.
36. Doong S-L, Tsai C-H, Schinazi RF, Liotta DC, Cheng Y-C. Inhibition of
the replication of hepatitis B virus in vitro by 2V,3V-dideoxy-3V-thiacytidine
and related analogues. Proc Natl Acad Sci USA 1991; 88:8495 – 8499.
37. Furman PA, Davis M, Liotta DC, Paff M, Frick LW, Nelson DJ, Dornsife
RE, Wurster JA, Wilson LJ, Fyfe JA, Tuttle JV, Miller WH, Condreay L,
Averett DR, Schinazi RF, Painter GR. The anti-hepatitis B virus activities,
cytotoxicities, and anabolic profiles of the ( – ) and (+) enantiomers of cis-5fluoro-1-[2-(hydroxymethyl)-1,3-oxathiolan-5-yl]cytosine. Antimicrob
Agents Chemother 1992; 36:2686 – 2692.
38. Lin T-S, Luo M-Z, Liu M-C, Pai SB, Dutschman GE, Cheng Y-C. Antiviral
activity of 2V,3V-dideoxy-h-L-5-fluorocytidine (h-L-FddC) and 2V,3V-dideoxyh-L-cytidine (h-L-ddC) against hepatitis B virus and human immunodeficiency virus type 1 in vitro. Biochem Pharmacol 1994; 47:171 – 174.
39. Larder BA, Darby G, Richman DD. HIV with reduced sensitivity to zidovudine (AZT) isolated during prolonged therapy. Science 1989; 243:1731 – 1734.
40. Larder BA, Kemp SD. Multiple mutations in HIV-1 reverse transcriptase
confer high-level resistance to zidovudine (AZT). Science 1989; 246:1155 –
1158.
41. Kellam P, Boucher CAB, Larder BA. Fifth mutation in human immunodeficiency virus type 1 reverse transcriptase contributes to the development of
high-level resistance to zidovudine. Proc Natl Acad Sci USA 1992; 89:1934 –
1938.
42. Mayers DL, McCutchan FE, Sanders-Buell EE, Merritt LI, Dilworth S,
Fowler AK, Marks CA, Ruiz NM, Richman DD, Roberts CR, Burke
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
331
DS. Characterization of HIV isolates arising after prolonged zidovudine
therapy. J AIDS 1992; 5:749 – 759.
St Clair MH, Martin JL,Tudor-Williams G, Bach MC, Vavro CL, King
DM, Kellam P, Kemp SD, Larder BA. Resistance to ddI and sensitivity
to AZT induced by a mutation in HIV-1 reverse transcriptase. Science 1991;
253:1557 – 1559.
Gao Q, Gu Z, Parniak MA, Cameron J, Cammack N, Boucher C, Wainberg
MA. The same mutation that encodes low-level human immunodeficiency
virus type 1 resistance to 2V,3V-dideoxyinosine and 2V,3V-dideoxycytidine confers high-level resistance to the ( – )enantiomer of 2V,3V-dideoxy-3V-thiacytidine. Antimicrob Agents Chemother 1993; 37:1390 – 1392.
Tisdale M, Kemp SD, Parry NR, Larder BA. Rapid in vitro selection of
human immunodeficiency virus type 1 resistant to 3V-thiacytidine inhibitors
due to a mutation in the YMDD region of reverse transcriptase. Proc Natl
Acad Sci USA 1993; 90:5653 – 5656.
Schinazi RF, Lloyd RM Jr, Nguyen M-H, Cannon DL, McMillan A, Ilksoy
N, Chu CK, Liotta DC, Bazmi HZ, Mellors JW. Characterization of human
immunodeficiency viruses resistant to oxathiolane – cytosine nucleosides.
De Clercq E. HIV resistance to reverse transcriptase inhibitors. Biochem
Pharmacol 1994; 47:155 – 169.
Balzarini J, Hao Z, Herdewijn P, Johns DG, De Clercq E. Intracellular
metabolism and mechanism of anti-retrovirus action of 9-(2-phosphonylmethoxyethyl)adenine, a potent anti – human immunodeficiency virus compound. Proc Natl Acad Sci USA 1991; 88:1499 – 1503.
Balzarini J, Holy´ A, Jindrich J, Dvorakova H, Hao Z, Snoeck R, Herdewijn
P, Johns DG, De Clercq E. 9-[(2RS)-3-Fluoro-2-phosphonylmethoxypropyl]
derivatives of purines: a class of highly selective antiretroviral agents in vitro
and in vivo. Proc Natl Acad Sci USA 1991; 88:4961 – 4965.
Balzarini J, Holy´ A, Jindrich J, Naesens L, Snoeck R, Schols D, De Clercq
E. Differential antiherpesvirus and antiretrovirus effects of the (S) and (R)
enantiomers of acyclic nucleoside phosphonates: potent and selective in
vitro and in vivo antiretrovirus activities of (R)-9-(2-phosphonomethoxypropyl)-2,6-diaminopurine. Antimicrob Agents Chemother 1993; 37:332 –
338.
Balzarini J, Naesens L, Herdewijn P, Rosenberg I, Holy´ A, Pauwels R, Baba
M, Johns DG, De Clercq E. Marked in vivo antiretrovirus activity of 9-(2phosphonylmethoxyethyl)adenine, a selective anti – human immunodeficiency virus agent. Proc Natl Acad Sci USA 1989; 86:332 – 336.
Balzarini J, Sobis H, Naesens L, Vandeputte M, De Clercq E. Inhibitory
effects of 9-(2-phosphonylmethoxyethyl)adenine and 3V-azido-2V,3V-dideoxythymidine on tumor development in mice inoculated intracerebrally with
Moloney murine sarcoma virus. Int J Cancer 1990; 45:486 – 489.
332
De Clercq
53. Naesens L, Balzarini J, De Clercq E. Therapeutic potential of 9-(2-phosphonylmethoxyethyl)adenine (PMEA) as an antiviral drug. Rev Med Virol
1994; 4 :147 – 159.
54. Tsai C-C, Follis KE, Sabo A, Grant RF, Bartz C, Nolte RE, Benveniste RE,
Bischofberger N. Preexposure prophylaxis with 9-(2-phosphonylmethoxyethyl)adenine against simian immunodeficiency virus infection in macaques.
J Infect Dis 1994; 169:260 – 266.
55. Heijtink RA, De Wilde GA, Kruining J, Berk L, Balzarini J, De Clercq E,
Holy´ A, Schalm SW. Antiviral activity of 9-(2-phosphonylmethoxyethyl)
adenine (PMEA) on human and duck hepatitis B virus infection. Antiviral
Res 1993; 21:141 – 153.
56. Starrett JE Jr, Tortolani DR, Hitchcock MJM, Martin JC, Mansuri MM.
Synthesis and in vitro evaluation of a phosphonate prodrug: bis(pivaloyloxymethyl)-9-(2-phosphonylmethoxyethyl)adenine. Antiviral Res 1992;
19:267 – 273.
57. Srinivas RV, Robbins BL, Connelly MC, Gong Y-F, Bischofberger N, Fridland A. Metabolism and in vitro antiretroviral activities of bis(pivaloyloxymethyl) prodrugs of acyclic nucleoside phosphonates. Antimicrob Agents
Chemother 1993; 37:2247 – 2250.
58. Pauwels R, Andries K, Desmyter J, Schols D, Kukla M-J, Breslin HJ,
Raeymaeckers A, Van Gelder J, Woestenborghs R, Heykants J, Schellekens
H, Janssen MAC, De Clercq E, Janssen PAJ. Potent and selective inhibition
of HIV-1 replication in vitro by a novel series of TIBO derivatives. Nature
1990; 343:470 – 474.
59. Debyser Z, Pauwels R, Andries K, Desmyter J, Kukla M, Janssen PAJ, De
Clercq E. An antiviral target on reverse transcriptase of human immunodeficiency virus type 1 revealed by tetrahydroimidazo-[4,5,1-jk][1,4]benzodiazepin-2(1H)-one and -thione derivatives. Proc Natl Acad Sci USA 1991;
88:1451 – 1455.
60. Baba M, Tanaka H, De Clercq E, Pauwels R, Balzarini J, Schols D, Nakashima H, Perno C-F, Walker RT, Miyasaka T. Highly specific inhibition of
human immunodeficiency virus type 1 by a novel 6-substituted acyclouridine
derivative. Biochem Biophys Res Commun 1989; 165:1375 – 1381.
61. Miyasaka T, Tanaka H, Baba M, Hayakawa H, Walker RT, Balzarini J, De
Clercq E. A novel lead for specific anti-HIV agents: 1-(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine. J Med Chem 1989; 32:2507 – 2509.
62. Baba M, De Clercq E, Tanaka H, Ubasawa M, Takashima H, Sekiya K, Nitta
I, Umezu K, Walker RT, Mori S, Ito M, Shigeta S, Miyasaka T. Potent and
selective inhibition of human immunodeficiency virus type 1 (HIV-1) by 5ethyl-6-phenylthiouracil derivatives through their interaction with the HIV-1
reverse transcriptase. Proc Natl Acad Sci USA 1991; 88:2356 – 2360.
63. Baba M, De Clercq E, Tanaka H, Ubasawa M, Takashima H, Sekiya K,
Nitta I, Umezu K, Walker RT, Mori S, Ito M, Shigeta S, Miyasaka T.
64.
65.
66.
67.
68.
69.
70.
71.
72.
333
Highly potent and selective inhibition of human immunodeficiency virus
type 1 by a novel series of 6-substituted acyclouridine derivatives. Mol
Pharmacol 1991; 39:805 – 810.
De Clercq E. HIV-1-specific RT inhibitors: highly selective inhibitors of
human immunodeficiency virus type 1 that are specifically targeted at the
viral reverse transcriptase. Med Res Rev 1993; 13:229—258.
Balzarini J, Pe´rez-Pe´rez M-J, San-Fe´lix A, Schols D, Perno C-F, Vandamme
A-M, Camarasa M-J, De Clercq E. 2V,5V-Bis-O-(tert-Butyldimethylsilyl)-3Vspiro-5VV-(4VV-amino-1VV,2VV-oxathiole-2VV,2VV-dioxide)pyrimidine (TSAO) nucleoside analogues: highly selective inhibitors of human immunodeficiency
virus type 1 that are targeted at the viral reverse transcriptase. Proc Natl
Acad Sci USA 1992; 89:4392 – 4396.
Balzarini J, Pe´rez-Pe´rez M-J, San-Fe´lix A, Velazquez S, Camarasa M-J, De
Clercq E. [2V,5V-Bis-O-(tert-Butyldimethylsilyl)]-3V-spiro-5VV- [4VV-amino-1VV,2VVoxathiole-2VV,2VV-dioxide) (TSAO) derivatives of purine and pyrimidine nucleosides as potent and selective inhibitors of human immunodeficiency
virus type 1. Antimicrob Agents Chemother 1992; 36:1073 – 1080.
Pauwels R, Andries K, Debyser Z, Van Daele P, Schols D, Stoffels P, De
Vreese K, Woestenborghs R, Vandamme A-M, Janssen CGM, Anne´ J,
Cauwenbergh G, Desmyter J, Heykants J, Janssen MAC, De Clercq E,
Janssen PAJ. Potent and highly selective human immunodeficiency virus
type 1 (HIV-1) inhibition by a series of a-anilinophenylacetamide derivatives targeted at HIV-1 reverse transcriptase. Proc Natl Acad Sci USA 1993;
90:1711 – 1715.
Kleim J-P, Bender R, Billhardt U-M, Meichsner C, Riess G, Ro¨sner M,
Winkler I, Paessens A. Activity of a novel quinoxaline derivative against
human immunodeficiency virus type 1 reverse transcriptase and viral replication. Antimicrob Agents Chemother 1993; 37:1659 – 1664.
Williams TM, Ciccarone TM, MacTough SC, Rooney CS, Balani SK, Condra JH, Emini EA, Goldman ME, Greenlee WJ, Kauffman LR, O’Brien JA,
Sardana VV, Schleif WA, Theoharides AD, Anderson PS. 5-Chloro-3-(phenylsulfonyl)indole-2-carboxamide: a novel, non-nucleoside inhibitor of
HIV-1 reverse transcriptase. J Med Chem 1993; 36:1291 – 1294.
Mertens A, Zilch H, Ko¨nig B, Scha¨fer W, Poll T, Kampe W, Seidel H, Leser
U, Leinert H. Selective non-nucleoside HIV-1 reverse transcriptase inhibitors. New 2,3-dihydrothiazolo[2,3-a]isoindol-5(9bH)-ones and related compounds with anti-HIV-1 activity. J Med Chem 1993; 36:2526 – 2535.
Kashman Y, Gustafson KR, Fuller RW, Cardellina JH, McMahon JB,
Currens MJ, Buckheit RW Jr, Hughes SH, Cragg GM, Boyd MR. The
calanolides, a novel HIV-inhibitory class of coumarin derivatives from the
tropical rainforest tree, Calophyllum lanigerum. J Med Chem 1992; 35:2735 –
2743.
Patil AD, Freyer AJ, Eggleston DS, Haltiwanger RC, Bean MF, Taylor PB,
334
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
De Clercq
Caranfa MJ, Breen AL, Bartus HR, Johnson RK, Hertzberg RP, Westley
JW. The inophyllums, novel inhibitors of HIV-1 reverse transcriptase isolated from the Malaysian tree, Calophyllum inophyllum Linn. J Med Chem
1993; 36:4131 – 4138.
Jacobo-Molina A, Ding J, Nanni RG, Clark AD Jr, Lu X, Tantillo C,
Williams RL, Kamer G, Ferris AL, Clark P, Hizi A, Hughes SH, Arnold
E. Crystal structure of human immunodeficiency virus type 1 reverse transcriptase complexed with double-stranded DNA at 3.0 A˚ resolution shows
bent DNA. Proc Natl Acad Sci USA 1993; 90:6320 – 6324.
Nanni RG, Ding J, Jacobo-Molina A, Hughes SH, Arnold E. Review of
HIV-1 reverse transcriptase three-dimensional structure. Implications for
drug design. Perspect Drug Dis Des 1993; 1:129 – 150.
Pialoux G, Youle M, Dupont B, Gazzard B, Cauwenbergh GFMJ, Stoffels
PAM, Davies S, De Saint Martin J, Janssen PAJ. Pharmacokinetics of
R82913 in patients with AIDS or AIDS-related complex. Lancet 1991;
338:140 – 143.
Saag MS, Emini EA, Laskin OL, Douglas J, Lapidus WI, Schleif WA,
Whitley RJ, Hildebrand C, Byrnes VW, Kappes JC, Anderson KW, Massari
FE, Shaw GM, the L-697,661 Working Group. A short-term clinical evaluation of L-697,661, a non-nucleoside inhibitor of HIV-1 reverse transcriptase. N Engl J Med 1993; 329:1065 – 1072.
Vandamme A-M, Debyser Z, Pauwels R, De Vreese K, Goubau P, Youle
M, Gazzard B, Stoffels PA, Cauwenbergh GF, Anne´ J, Andries K, Janssen
PAJ, Desmyter J, De Clercq E. Characterization of HIV-1 strains isolated
from patients treated with TIBO R82913. AIDS Res Hum Retrovir 1994;
10:39 – 46.
Nunberg JH, Schleif WA, Boots EJ, O’Brien JA, Quintero JC, Hoffman JM,
Emini EA, Goldman ME. Viral resistance to human immunodeficiency virus
type 1-specific pyridinone reverse transcriptase inhibitors. J Virol 1991;
65:4887 – 4892.
Richman D, Shih C-K, Lowy I, Rose J, Prodanovich P, Goff S, Griffin J.
Human immunodeficiency virus type 1 mutants resistant to nonnucleoside
inhibitors of reverse transcriptase arise in tissue culture. Proc Natl Acad Sci
USA 1991; 88:11241 – 11245.
Mellors JW, Dutschman GE, Im G-J, Tramontano E, Winkler SR, Cheng
Y-C. In vitro selection and molecular characterization of human immunodeficiency virus-1 resistant to non-nucleoside inhibitors of reverse transcriptase. Mol Pharmacol 1992; 41:446 – 451.
De Vreese K, Debyser Z, Vandamme A-M, Pauwels R, Desmyter J, De
Clercq E, Anne´ J. Resistance of human immunodeficiency virus type 1
reverse transcriptase to TIBO derivatives induced by site-directed mutagenesis. Virology 1992; 188:900 – 904.
Sardana VV, Emini EA, Gotlib L, Graham DJ, Lineberger DW, Long WJ,
83.
84.
85.
86.
87.
88.
89.
90.
335
Schlabach AJ, Wolfgang JA, Condra JH. Functional analysis of HIV-1
reverse transcriptase amino acids involved in resistance to multiple nonnucleoside inhibitors. J Biol Chem 1992; 267:17526 – 17530.
Balzarini J, Karlsson A, De Clercq E. Human immunodeficiency virus type1
drug-resistance patterns with different 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine derivatives. Mol Pharmacol 1993; 44:694 – 701.
Byrnes VW, Sardana VV, Schleif WA, Condra JH, Waterbury JA, Wolfgang JA, Long WJ, Schneider CL, Schlabach AJ, Wolanski BS, Graham
DJ, Gotlib L, Rhodes A, Titus DL, Roth E, Blahy OM, Quintero JC,
Staszewski S, Emini EA. Comprehensive mutant enzyme and viral variant
assessment of human immunodeficiency virus type 1 reverse transcriptase
resistance to nonnucleoside inhibitors. Antimicrob Agents Chemother 1993;
37:1576 – 1579.
Balzarini J, Karlsson A, Pe´rez-Pe´rez M-J, Vrang L, Walbers J, Zhang H,
O¨berg B, Vandamme A-M, Camarasa M-J, De Clercq E. HIV-1-specific
reverse transcriptase inhibitors show differential activity against HIV-1 mutant strains containing different amino acid substitutions in the reverse transcriptase. Virology 1993; 192:246 – 253.
Mellors JW, Im G-J, Tramontano E, Winkler SR, Medina DJ, Dutschman
GE, Bazmi HZ, Piras G, Gonzalez CJ, Cheng Y-C. A single conservative
amino acid substitution in the reverse transcriptase of human immunodeficiency virus-1 confers resistance to (+)- (5S)-4,5,6,7-tetrahydro-5-methyl6-(3-methyl-2-butenyl)imidazo[4,5,1-jk][1,4]benzodiazepin-2(1H)-thione
(TIBO R82150). Mol Pharmacol 1993; 43:11 – 16.
Balzarini J, Karlsson A, Pe´rez-Pe´rez M-J, Camarasa M-J, Tarpley WG, De
Clercq E. Treatment of human immunodeficiency virus type 1 (HIV-1)-infected cells with combinations of HIV-1-specific inhibitors results in a different resistance pattern than does treatment with single-drug therapy. J Virol
1993; 67:5353 – 5359.
Balzarini J, Velazquez S, San-Fe´lix A, Karlsson A, Pe´rez-Pe´rez M-J, Camarasa M-J, De Clercq E. Human immunodeficiency virus type 1-specific
[2V,5V-bis-O-(tert-butyldimethylsilyl)-h-D-ribofuranosyl]-3V-spiro-5VV-(4VV-amino-1VV,2VV- oxathiole-2VV,2VV-dioxide)-purine analogues show a resistance spectrum that is different from that of the human immunodeficiency virus type 1specific non-nucleoside analogues. Mol Pharmacol 1993; 43:109 – 114.
Balzarini J, Karlsson A, Vandamme A-M, Pe´rez-Pe´rez M-J, Zhang H,
Vrang L, O¨berg B, Ba¨ckbro K, Unge T, San-Fe´lix A, Velazquez S, Camarasa
M-J, De Clercq E. Human immunodeficiency virus type 1 (HIV-1) strains
selected for resistance against the HIV-1-specific [2V,5V-bis-O-(tert-butyldimethylsilyl)-3V-spiro-5VV-(4VV-amino-1VV,2VV-oxathiole-2VV,2VV-dioxide)]-h-D-pentofuranosyl (TSAO) nucleoside analogues retain sensitivity to HIV-1-specific
nonnucleoside inhibitors. Proc Natl Acad Sci USA 1993; 90:6952 – 6956.
Kleim J-P, Bender R, Kirsch R, Meichsner C, Paessens A, Riess G. Muta-
336
91.
92.
93.
94.
95.
96.
97.
De Clercq
tional analysis of residue 190 of human immunodeficiency virus type 1
reverse transcriptase. Virology 1994; 200:696 – 701.
Dueweke TJ, Pushkarskaya T, Poppe SM, Swaney SM, Zhao JQ, Chen ISY,
Stevenson M, Tarpley WG. A mutation in reverse transcriptase of bis(heteroaryl)piperazine-resistant human immunodeficiency virus type 1 that confers increased sensitivity to other nonnucleoside inhibitors. Proc Natl Acad
Sci USA 1993; 90:4713 – 4717.
Pauwels R, Andries K, Debyser Z, Kukla MJ, Schols D, Breslin HJ, Woestenborghs R, Desmyter J, Janssen MAC, De Clercq E, Janssen PAJ. New
tetrahydroimidazo[4,5,1-jk][1,4]-benzodiazepin-2(1H)-one and -thione derivatives are potent inhibitors of human immunodeficiency virus type 1
replication and are synergistic with 2V,3V-dideoxynucleoside analogs. Antimicrob Agents Chemother 1994; 38:2863 – 2870.
Goldman ME, O’Brien JA, Ruffing TL, Schleif WA, Sardana VV, Byrnes
VW, Condra JH, Hoffman JM, Emini EA. A nonnucleoside reverse transcriptase inhibitor active on human immunodeficiency virus type 1 isolates
resistant to related inhibitors. Antimicrob Agents Chemother 1993; 37:947 –
949.
Larder BA. 3V-Azido-3V-deoxythymidine resistance suppressed by a mutation conferring human immunodeficiency virus type 1 resistance to nonnucleoside reverse transcriptase inhibitors. Antimicrob Agents Chemother
1992; 36:2664 – 2669.
Balzarini J, Karlsson A, Pe´rez-Pe´rez M-J, Camarasa M-J, De Clercq E.
Knocking-out concentrations of HIV-1-specific inhibitors completely suppress HIV-1 infection and prevent the emergence of drug-resistant virus.
Virology 1993; 196:576 – 585.
Vasudevachari MB, Battista C, Lane HC, Psallidopoulos MC, Zhao B,
Cook J, Palmer JR, Romero DL, Tarpley WG, Salzman NP. Prevention
of the spread of HIV-1 infection with nonnucleoside reverse transcriptase
inhibitors. Virology 1992; 190:269 – 277.
Dueweke TJ, Poppe SM, Romero DL, Swaney SM, So AG, Downey KM,
Althaus IW, Reusser F, Busso M, Resnick L, Mayers DL, Lane J, Aristoff
PA, Thomas RC, Tarpley WG. U-90152, a potent inhibitor of human immunodeficiency virus type 1 replication. Antimicrob Agents Chemother
1993; 37:1127 – 1131.
Index
N-acylethanolamide, 97
N-acyltransferase, 99
Actin, 251, 253
depolymerization, agents of, 252
Adenyl cyclase, 92, 116, 149
AIDS patients, 319, 320
AIDS-encephalitis, 113
Alcuronium, 222, 226
see also G-protein-coupled
receptors, muscarinic
receptors
Alzheimer’s disease, 245, 246, 249,
251–254, 257–263
animal models of, 262
cholinergic deficit models, 262
mice, overexpressing human
hAPP, 262
Hsiao mouse, 262
PDAPP mouse (Elan mouse
model), 262–263
transgenic presenilin mutants,
262
b-amyloid (Ah), amyloid precursor
protein (APP), 260
b peptide b1–42, 261
b peptide b10–35, 257, 260
b peptide b25–35, 260
[Alzheimer’s disease]
peptide plaques (Ah), 245, 252–
253, 254, 257
Ab fibril formation, 257,259,
260
early events, methods of
study, 257–258
fibril extension,
characterization of, 258
inhibitors of Ah fibrilization,
nucleation, aggregation,
254–257, 259, 260
Congo Red, 257, 258, 259,
261
prenucleation, 257
cell systems for testing inhibitors,
260
embryonic mouse neurons,
260
mixed neuronal/glial embryonic
cultures, 260
organotypic culture, 261
leptomeningeal blood vessel,
261
leptomeningeal vascular
smooth muscle cells
(VSMC), 261
337
338
[Alzheimer’s disease]
PC12 cell lines, 260
smooth muscle cells, 261
cerebral vasculature and deposition
of h-amyloid fibrils, 260–
261
clinical trials for therapeutics, 264–
265
amyloid aggregation inhibitors,
264
Dutch mutation, 260
familial AD, 260, 262
presenilin, 262
Lewy body variant, 249
neurofibrillary tangles (NFTs), 245
Tau protein, hyperphosphorylated, 245, 252
pathological characterization, 245
therapeutic interventions, 251
AM251, 117, 134, 136
SPECT primate studies, 134
AM281, 136 137, 139, 140–141, 142
[123I] AM281 SPECT primate
studies, 136, 137, 142
[18F] AM284, 137
AM356 [(R)-1V-methanandamide],
110
AM374 hexadecylsulfonyl fluoride,
110
AM404, 110
AM630, 109, 110
AM1241, 109
Aminoalkylindoles (AAIs), 109–110
Amylin (islet amyloid polypeptide,
IAPP), 252, 253
Analgesia, 97, 149, 159
Anandamide amidase or arachidonylethanolamide amidohydrolase (AEAase), 102,
104, 105, 110, 117, 119
Anandamide transporter (AT), 105,
119
Aracidonic acid, 99, 101, 103
Index
Arachidonyl ethanolamide (AEA) or
anandamide, 95–98, 101–
103, 110, 115, 118, 140
biosynthesis, 98
D2 activation, 98, 112
localization, 95-97
role in embryo production and
early development, 117
production in the human
reproductive system, 117
2-Arachidonyl glycerol (2-AG), 96,
97–98, 101–102, 103–104,
110 118
biosynthetic pathways, 99
N-Arachidonylphosphatidylethanolamide (NAPE), 99, 101–
102
Archidonyltrifluoromethylketon
(ATFAK), 110
Arthritis, 63
Astroctyes, 102
Ataxia, 111
Atomic force microscopy (AFM), 254
Biarylpyrazoles, 110
Benzodiazepines, 222
Cannabimimetics, 89–90, 96, 97–98,
107, 111, 114–115
analgesia, 114
antiemetic and antinausea effects,
116
antinociceptive effects, 114
chronic pain, 114
appetite stimulatory effects, 115
and blood-brain barrier, 94
CC (see Cannabinoid analogs )
and membrane interactions, 94
NCC (see Cannabinoid analogs)
reproductive and metabolic effects
of, 117
respiratory effects, 118
Cannabinergics, 91, 104–105, 111, 115
Index
[Cannabinergics]
treatment of motor disorders, 112
and multiple sclerosis, 113
Cannabinoids, 90, 115, 116
and the cardiovascular system, 116
and membrane interactions, 94
anticonvulsant effects, 112
antispastic effects, 112
intraocular pressure effects, 118
memory effects on, 114
and the reproductive system, 117
Cannabinoid analogs, 107
classical cannabinoid analogs (CC)structural features important for cannabimimetic
activity, 107–108, 109
nonclassical cannabinoid analogs
(NCC), 108, 109
Cannabidiol, 116
Catalepsy, 97, 111
Catecholamines, 149
CB1 receptor, 90–91, 92, 93, 103, 108–
109, 111, 112, 114,117, 118,
119, 129, 140–141, 142
brain CB1 receptor occupancy,
140–141
central nervous system localization,
93
cAMP levels, 91
cannabinmemetic ligands to, 91
CMR, effect on, 131
see also Cerebral metabolic rate
Gi coupled, 91
and Huntington’s chorea, 112
mediated inhibition of nitric oxide
release, 113
and Parkinson’s disease, 112
and peripheral somatic localization,
93
potassium channels, 91
radioligand labeling, 137
sodium ion affect, 91
and Tourette’s syndrome, 112
339
CB1A, 92, 118
CB2 receptor, 92, 108–109, 119, 129,
134
immunomodulatory action, 92, 116
signal transduction pathway, 92
somatic localization, 93
Cannabis sativa, 89, 105
Cesamet, 89
Cerebral blood flow (CBF), 130–131
Cerebral ischemia, 113
Cerebral metabolic rate (CMR), 130–
131
Cerebellum, 131
Chemokine receptors, CXCR4 and
CCR5
coreceptors for HIV, 310
Classical cannabinoids (CC), 105, 117
Combinatorial chemistry and Src,
33–36
CMV, cytomegalovirus, 310, 313, 319
CP-55,940, 97, 108–109
autoradiographic studies, 132–133
Cytochrome P450s, 5
Deltorphins, 149, 164
see also Opioids
Dermorphin, 149
cyclic analog of, 152
see also Opioids
Diabetes, type II, 230, 251
Disoxaril, 281
binding to HRV-capsid protein,
281–282
Dopamine, 225
D1 receptors, 132
D2 activation, 112
see also Arachidonyl ethanolamide
1,2-Diacylglycerol, 103
(R,R)-2,1V-Dimethyl anandamide,
110
4-Diphosphonomethylphenylalanine
(DMP), 50–52
Dihydroxyl THC derivatives, 94
340
7,10,13,16-Docasatetraenylethanolamide, 96
Dynamic light scattering, 257
Dynorphin A analogues, 149
Dysphoria, 149
Endocannabinoids, 95, 101–102, 104,
110, 113–114, 115
depolarization-induced
suppression of inhibition
(DCI), 114
depolarization-induced excitation
(DCE), 114–115
Endomorphins, 148–149
Endorphins, 148
Enkephalins, 148, 149
cyclic analogue of, 148, 150
see also Opioids
homologues, 150
other analogues, 150
5-Enol-pyruvyl-3-phosphate
synthase, 15–16
Enthanolamine, 99
Enzyme surrogate, 18
Euphoria, 149
Fatty acid amidohydrolase (FAAH),
98
Fluorescence resonance energy
transfer (FRET), 258
GABA, 111
GABAA, 222
Gallamine, 222, 226
receptors
muscarinic receptors
G-protein-coupled receptors
(GPCRs), 90–91, 109, 116,
221
a2A-adrenergic receptors, 225
adenosine, 230
extracellular production, 230
Index
[G-protein-coupled receptors
(GPCRs)]
adenosine receptors, A2a, A2B, A3,
225, 227
A1 receptor, 227
agonists, metabolic effects, 230
CCPA, 232–233
N 6-cyclopentyladensine,
(CPA), 230–232
NECA, 230
R-PIA, 230
allosteric enhancers of, 230
PD81,723, (2-amino-4,5dimethyl-trienyl)[3-(trifluoromethyl)
phenyl]methanone, 230, 232–
234
analogs of, 233–236
expression of, 227, 230
mechanism of action, 227
partial agonists of, 230
agonists of, 222
allosteric, definition, 223–224
cooperativity, 224
negative cooperativity, 224
positive cooperativity, 224
compounds exhibiting on
muscarinic receptors, 226
allosteric modulation of, 222, 227
drug properties, 222
allosteric sites, characterization of,
223
antagonists of, 222
endogenous ligands, allosteric
function on GPCRs, 223
inverse agonists of, 222
muscarinic receptors, M1–M4,
225–226
compounds binding to an
allosteric site on muscarinic receptors, 226
receptology, 224
subtypes, 222
Index
[G-protein-coupled receptors
(GPCRs)]
ternary complex allosteric model,
224
cubic ternary complex model,
224–225
Hanging drop crystallization, 7
HBV, 317
High throughput organic synthesizer,
36
HIV (human immunodeficiency virus)
acute phase of infection, 209
AIDS, 312
CD4 receptor, 310, 312
initial stages of infection, 309
persistant or chronic infection, 309
protease inhibitors, 3
proviral DNA, 309
replicative cycle, 309–310
reverse transcriptase inhibitors,
316
acyclic nucleoside phosphonates,
318–320
PMEA and congeners, 318
half-life, 320
non-HIV antiretroviral
activity of, 318
non-retroviral antiviral
activity of, 318–319
uptake and absorption, 320
dideoxynucleoside analogues,
316–318
antiviral spectrum, 317
AZT (zidovudine), 316, 317,
318, 326
D4T (stavudine), 317
DDC (zalcitabine), 316, 317,
318
DDI (didanosine), 316, 317,
318
FddCIUrd, 316
FTC, 317, 318
341
[HIV (human immunodeficiency
virus)]
3TC (lamivudine), 317, 318,
325
viral resistence via RT
mutation, 317–318
NNRTIs, 318–326
a-APA, 322–323, 324
antiviral spectrum, 323
BHAP, 318, 320, 323, 324, 325,
326
binding to RT, 323–324
Calophyllum lanigerum, 322
5-chloro-3-(phenylsulfonyl)
indole-2-carboxamide,
322
HEPT derivatives, 318, 322–
323, 324, 325, 326
HIV-1 resistance and drug
switching, 324, 325
nevirapine, 318, 322, 324, 325,
326
pyridinone, 318, 322, 324, 326
pyridinone L-697,661, 324, 325
quiloxaline S-2720, 322
TIBO derivatives, 318, 322,
324, 325, 326
TSAO derivatives, 322, 324
syncytium, 312, 315
virus-drug resistance, 324
viral envelope protein, gp41, 313
viral envelope protein, gp120, 310,
312
virus adsorption inhibitors, 310
polyanionic substances, 310–312
alterations for higher
bioavailability, 311
dextran sulfate, 310
inhibition of non-HIV virus,
311
polysulfates, 310
342
[HIV (human immunodeficiency
virus)]
virus-cell fusion inhibitors, 312–314
albumins, 313–314
aconitylated human serum
albumin (Aco-HAS), 313,
314
succinylated human serum
albumin (Suc-HSA), 312,
314
lectins, 312–314
N-acetylglucosamine-specific
plant lectins, 312
mannose-specific, 312
triterpene derivatives, 312
betulinic acid, 312, 314
virus uncoating inhibitors, 314–
316
bicyclams, 314
JM2763, 314
JM3100, 314–316
Homo-g-linolenylethanoamide, 96
HSV (Herpes Simplex Virus), 310, 319
Huntington’s disease, 245–246
animal models of, 262–263
Bates R6 mice, 263
CAG repeats, genetically determined, 245
polyglutamine-containing brain
inclusions, 245–246
aggregated h-sheets, 246, 253
Hypoactivity, 97
Hypothermia, 97, 115, 149
see also Cannabimimetic
Influenza A virus, 310, 313
IgG light chain amyloidosis (AL),
252, 253
Lead molecule, 4–5
selection, 4
testing, 5
optimization, 5
Index
Ligand complexes, 1
and crystal contacts, 18
Ligand design, 17
iterative, 18
solubility, 18
12(S)-Lipooxygenase, 102–103
a-Macroglobin, 63
Marijuana, 89, 129, 140
and glaucoma, 118
immunomodulatory properties, 92,
116
and multiple sclerosis, 113
Marinol, 89, 115
see also D9-THC,
MAP kinase, 92
Matrixins or matrix metalloendoproteinase (MMP), 62–64
and biological processes pathologic, 63
and biological processes normal, 63
basement membrane degradation,
63
catalytic domain, 65–66
collagenases, 63
enamelysin, 63
elastase, macrophase, 63
gelatinases, 63
matrilysin, 64
membrane-associated, 63
stromelysins (see Stromelysin)
Merrifield, B., 175–176
Molecular biology,
and protein crystallography, 2
and drug design, 19
Molecular replacement method
(MR), 9
Monoacylglycerol lipase (MAG), 104
Morphiceptin, 149
analogs of, 152–153
Morphine, 90, 159
Multiple isomorphous replacement
(MIR), 9–10
Index
Myoglobin, 1
multidimensional, heteronuclear,
61
Naloxone, 92
Neurodegenerative disease
amyloidoses, familial, 249
Familial British dementia and
Abri deposition, 249
amyotrophic lateral sclerosis (ALS
or Lou Gehrig’s disease),
249
familial ALS, 24
superoxide dematase (SOD1)
mutations, 249
Alzheimer’s disease, 245
see also Alzheimer’s disease
familial encephalopathy with
neuroserpin inclusion
bodies (FENIB), 249–250
Parkinson’s disease (see Parkinson’s
Disease)
‘‘piggyback’’ strategies for
therapeutics, 251
polyglutamine, trinucleatide repeat
diseases, 245–246, 252, 253
(see also Huntington’s disease
(HD))
Dentatorubral-palladoluysian
atrophy (DRPLA), 246
Machado-Joseph disease (MJD;
SCA 3), 246
animal models for, 263
spinocerebellar ataxia, SCA-3
SCA-7, 246
animal models for SCA 1, 263
prion diseases, 246, 249, 264
animal models of, 264
Creutzfeldt-Jakob disease
(CJD), 249
PrPc, or PrPs, cellular protein,
protease-sensitive or
cellular, 246
343
[Neurodegenerative disease]
PrPR, protease resistant, 246,
263
PS1, 249
scrapie, 249
spongiform encephalopathy,
249
Sup35p, 249
RE3, 249
tau and tauopathies, 246, 263
see also Alzheimer’s disease
familial tauopathies
(FTDP17), 246
mRNA isoform expression
alterations, 246
Pick’s disease, 246
progressive supranuclear
palsy, 246
Neuropeptides, 147
NMR, 2, 23, 163, 167
NOE spectra, 62
Nucleoside transport blockers, 230
dipyridamole, 230
dilazep, 230
Opioids
blood-brain barrier, 162, 164, 167
deltorphins, 149, 164
dermorphin, cyclic analog of, 152
enkephalin analogue, 150, 153
homologue, 150
mixed A agonists/ y antagonists,
160–163, 166
H-Dmt-Tic-Phe-Phe-NH2
(DIPP-NH2), and
analogues of, 160–162, 166–167
morphiceptin, 149
analog of, 152–153
naltrindole (NTI), 153–154
naltrindole, benzofuran analogue
(NTB), 153–154, 159
peptide analogs, 149–151, 166
peptidomimetics, 148–149, 151
344
[Opioids]
receptors, 148
receptors n and A, 114, 149, 165
receptor y, agonists, 150, 151,
160, 163–164, 165–166
Bw373u86, 164
H-Tyr-D-Thr-Gly-Phe-LeuThr-OH (DTLET), 163–
164
H-Tyr-c[D-Pen-Gly-Phe-DPen]OH (DPDPE), 164,
165
H-Tyr-c[D-Cys-Phe-D-PenOH]OH (JOM-13), 164
SNC80, 164
TAN-67, 164
receptor y antagonists, 150, 153,
160
and morphine tolerance and
dependence, 159
receptor A, agonists, 150–151,
153, 160, 167
H-Dmt-D-Ala-NH-(CH2)3-Ph
(SC-399566), 167
H-Tyr-Tic-Phe-OH (TIP), TIP
peptides, 154, 157–159,
166
H-Tyr-Tic-Phe-Phe-OH
(TIPP) TIPP peptides,
154–159
Osteoclasts, 29,53
Osteopetrosis (see SRC)
Osteoporosis (see SRC)
P-glycoprotein, 102
Pain, chronic management of, 159
Palmitylethanolamide, 92
Pancuronium, 222, 226
receptors and muscarinic
receptors
Parallel synthesis integrated, 45–49,
52
Index
Parkinson’s disease, 249
familial early-onset Parkinson’s
disease, 249
A54T and A30P, mutant forms of
a-synuclein, 249
a-synuclein, 249, 252, 253
A54T and A30P, mutant forms
of, 249
Lewy neurite, deposited in, 249
Lewy body disease, 249
mice expressing high levels of
human a-synuclein, 263–
264
h-synuclein, 249
g-synuclein, 249
Peptide analogues, 147
conformationally restricted, 148
Peptiomimetics, 147
Pharmacophore-linking strategy, 37
Phenylmethoanesulfonyl fluoride
(PMSF), 98
Phenyl phosphate resins, 37–43
Phospholipase D, 99
Polymerase chain reaction (PCR), 326
Positron emission tomography (PET),
130, 132, 137–138, 140,
142
rodent studies (see also AM281),
139, 141
Prolyl isomerase (PIN 1), 246
Protein Data Bank (PDB), 2
Protein crystallography (see Single
crystal x-ray diffraction)
Protein kinase A, 91
Protein-protein interactions,
interfaces protein-protein, 16–17
isotopic enrichment, 62
protein-ligand contacts, 62
and signal pathway induction, 23,
26
Proteins, a and h pleated sheets, 1
atomic structure, 1
three-dimensional structure, 1
Index
Reproductive system, 117, 149
Rhinoviruses, 279
antirhinoviral agents, 279
see also Disoxaril
aromatic-aromatic interactions,
288
oxazoline and phenyl rings of
inhibitors of rhinovirus
replication, 289
torsion angle of, 289
stacking interactions, 288
blockers of adsorption of minor
group, 279–280
capsid binding compounds, 280–
282
comparative molecular field
analysis (CoMFA), 292–
293
energy profiling study, 283
enantiomers of WIN52084,
283–284
R- and S-ethyl compounds, 284
hydrogen bonding of compounds to
binding sites, 290
inhibitors of uncoating of major
group, 279–280
model development, 296
drug design, 297–298
clinical candidate development, 299, 300–301
SYBYL use of, 291, 296
pleconaril, 304
QSAR (quantitative structureactivity analysis), 293,
294
WIN52084, 281, 290
binding to HRV-14 capsid
protein, 281–282
dynamic studies, 285
WIN54954, 285–287, 299, 301
acetlylene analogue of, 287
clinical candidate, 299, 300
E isomer of, 286
345
[Rhinoviruses]
metabolic stability, 301
HPLC determination of, 301
miminal inhibitory concentration
(MIC) values, 288
oxadiazole analogue of, 301–302
HPLC determination of
metabolite, 301
ion mass spectrometry metabolite determination, 302
template for novel compounds
in virus-bound conformation, 287
Tripos force field, 287, 293
volume maps, 291
Z olefin of, 286
human rhinovirus 14 (HRV-14),
280–281, 285–288, 290,
293
model development, 296–207,
297–298, 304
three-dimensional structure,
280–281, 293
contour maps, 294–295
HRV-1A, 285, 288, 297-298, 304
major group, 279
binding to ICAM-1, 279
minor group, 279
binding to human low density
lipoprotein receptor
(LDLR), 279
x-ray crystallographic studies of
HRV-14 binding with
active compounds, 284,
290–294
other rhinoviral serotype
structures, 281
VP1–VP4, 280–281
Rotational Overhauser effect spectroscopy (ROESY), 163
Reverse-phase (RP) semipreparative
HPLC, 47
Rhabdoviruses, 310
346
RSV (respiratory syncytial virus), 310,
313
Signal transduction pathways, 26
Single crystal x-ray diffraction, 2,
6–15, 23
and data acquisition, 8
flash freezing, 9
phasing, 9
SIV (simian immunodeficiency virus),
310, 315
Sedation, 149
Serum albumin, 5
see also HIV
Serum amyloid A protein (SAA), 252
Selenomethione multiple-wave length
anomalous diffraction, 2
Sickle-cell anemia, 252
hemoglobin S aggregation, 252
hydroxyurea treatment, 252
Single-photon emission-computed
tomography (SPECT),
130, 137, 140, 142
Solid-phase parallel synthesis, 36–43
Solid-phase synthesis,
aldehyde moiety containing
products, 200–201
amides at C-terminus, 184, 201
amines, yield after release, 192, 201
secondary amines, 193, 196
Mannich-type amine products,
196
N-protected amines, 196
tertiary amines, 193
amino function, generation of, 190–
191
aminopyridazines synthesis, 196
aminosulfonyl ureas, 198
anchor (see Linker)
aryl-containing compounds, 202
C teriminus of polymer chain
C-terminal attachment, 175–176,
177, 198
Index
[Solid-phase synthesis]
carbohydrate synthesis, 204
carboxamide functional group after
release, 178
carboxylic acid functional group
after release, 178, 201
cleavage from resin, 176, 178
fluoridolysis, 180, 182
photolysis or photolabile linkers,
180–182, 190
dihydropyrans, synthesis of, 205
DNA, synthesis of, 176
guanidine moiety containing
products, 198–200
handle (see Linker)
hydroxamic acids, synthesis of, 197
linker, 177–178, 180–181
aldehyde attachment point
linker, 188
see also Resins
alkanesulfonamide handle, 187
aryl hydrazide linker, 187
attachment to resin, 177
backbone amide linker (BAL),
188
DHP (dihyrdopyran) linker, 190
Dod linker, 185
HMB, 179
hydroxycrotyl-oligoethylene
glycol-n-alkanoyl
(HYCRON), 183
piperazine linker, 203
preformed, 17
silyl-derived linker, 186, 201
trialkylsiane linker, 203
Wang, 205
xanthone-based handles, 185–186
1-oxacephams synthesis, 204
peptides, 175–176
protein sequencing, 176
pyridine-based tricyclics, 202–203
resins, 176, 178
see also Linkers
Index
[Solid-phase synthesis]
p-benzyloxybenzylamine
(BOPA), 191
4(1V,1V-dimethyl-1V-hydroxypropyl)phenoxyacetyl)
(DHHP), 178
4-hydroxymethylphenylacetic
acid HMPA), 178
p-hydroxymethylphenylacetic
acid (PAM), 178
PAB (see HMPA)
MAMP (Merrifield, AlphaMethoxyPhenyl) resin,
190
5-(4-aminomethyl-3.5-dimethylthoxyphenoxy)valeric
acid (PAL), 184–185, 205
9-phenylfluoren-9-yl polystyrene
(Phfl), 191
SASRIN (super-acid-sensitive
resin), 179
redox-sensitive resin, 183
rink acid, 179, 205
rink amide support (RAM), 185
HAL (hyper-acid sensitive), 179
trialkysilane resin (PS-DES), 191
see also Resins
Solid support, 176
see also Resins
polystyrene (PS)
functionalized with chloromethyl, 176
with amino group at terminus,
176
release after peptide synthesis,
176
with strong acid, 180, 190
Sperm, 117
SR14176A (CB1 antagonist), 92, 110,
113, 117, 118, 129, 134,
137
appetite suppression, 115
SR144528 (CB2 antagonist), 92, 110
347
Src,
and cancer, 26
homology-2 domains, 23, 26
inhibitors nonpeptides, 54
inhibitors, orally active, 34
nonreceptor protein kinase src, 23
and osteopetrosis, 29
and osteoporosis, 26, 28
and knockout mice, 28
and SH2 domain, 26–27
Src SH2-phosphopeptide complex,
43
Stercuronium, 226
see also G-protein-coupled receptors, muscarinic receptors
Stromelysin-1, human (sfSTR), 62–79
catalytic domain, 62, 79
complexed to inhibitor, 62,
resonance assignment of
inhibited catalytic
domain, 65–68
resonance assignment of inhibitor and NOE between
inhibitor and protein, 71
structure of inhibitor protein, 73–
76
inhibitor and conformation, 77–79
Structure-based drug design, 2
and protein crystallography, 5
Tau protein, 246, 253
see also Neurodegenerative
diseases, tauopathies
hyperphosphorylation, 246
()-D8-Tetrahydrocannabinol, 105,
132–133
[18F] D8– THC and CB1 receptor,
133
D9-THC, 90, 94, 97, 116, 129, 140
effect on cerebral metabolic rate,
130–133
and cross-tolerance studies with
anandamide, 97
348
[D9-THC]
management of AIDS-wasting
syndrome, 115
Tissue inhibitor of metalloproteases
(TIMP-1 and 2), 63
Thermolysin, 80–81
Total correlation spectroscopy
(TOCSY), 163
Tolerance and physical dependence,
149
Transthyretin (TTR), 252
Transverse relaxation optimized
spectroscopy (TROSY),
81
Tubocuranine, 226
see also G-protein-coupled receptors, muscarinic receptors
Index
Tubulin, 251, 253
depolymerization, agents of,
252
microtubule stabilization agent of,
252
Tumor growth, 26, 63, 149
Water molecule, and importance in
ligand design, 17
WIN55212-2, 97, 109, 133, 139,
140–141
[3H] WIN55212-2, 133
possible mechanism of neuroprotective effect, 112
X-ray diffraction (see single crystal
x-ray diffraction)

Drug Discovery Strategies and Methods

Transcription

Similar documents

The Pharmaceutical Industry and The Process of Drug Discovery

6. Endocrinology 2010

Touch and Tactile Perception

Prezatide Copper Acetate (GHK-Cu)

Price Request for Calendar Finishing (pdf file)

GBC WIREBIND W15 BINDING MACHINE

gbc thermabind t400 binding

Full Brochure

Signal transmission with cell surface and intracellular receptors Prof.Dr.Gönül Kanıgür