Self-Assembly of DNA Arrays into Multilayer Stacks

Subscriber access provided by HOUSTON ACADEMY OF MEDICINE
Article
Self-Assembly of DNA Arrays into Multilayer Stacks
Alexey Y. Koyfman, Sergei N. Magonov, and Norbert O. Reich
Langmuir, 2009, 25 (2), 1091-1096 • DOI: 10.1021/la801306j • Publication Date (Web): 11 July 2008
Downloaded from http://pubs.acs.org on January 15, 2009
More About This Article
Additional resources and features associated with this article are available within the HTML version:
•
•
•
•
Supporting Information
Access to high resolution figures
Links to articles and content related to this article
Copyright permission to reproduce figures and/or text from this article
Langmuir is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Langmuir 2009, 25, 1091-1096
1091
Self-Assembly of DNA Arrays into Multilayer Stacks
Alexey Y. Koyfman,§,† Sergei N. Magonov,‡ and Norbert O. Reich*,§,†
Department of Chemistry and Biochemistry, Biomolecular Science and Engineering Program, UniVersity of
California - Santa Barbara, Santa Barbara, California 93106-9510, and Veeco Instruments,
112 Robin Hill Road, Goleta, California 93117
ReceiVed April 26, 2008. ReVised Manuscript ReceiVed May 21, 2008
We describe the self-assembly of multilayer hexagonal DNA arrays through highly regular interlayer packing. Slow
cooling of a mixture of three single-stranded DNA sequences with various Mg2+ concentrations leads to the selfassembly of diverse multilayer architectures. The self-assembled aggregates were deposited onto mica surfaces and
examined with atomic force microscopy. The size of the two-dimensional arrays and subsequent stacking to form
multilayer structures are highly dependent on Mg2+ concentration. DNA bilayers and multilayers of defined shape
are favored in 2-5 mM Mg2+ with an average lateral size of 700 nm. Arrays are much larger (up to 20 µm across)
in 10-15 mM Mg2+, although multiple layers still make up 20-60% of the observed structures. Domains within single
layer architectures were identified using Moiré pattern analysis. Distinct structural phases within the multilayer assemblies
include two layers translated by 17.5 nm and interlayer rotations of 20° and 30°. Three layer assemblies have cubic
close packing and taller multilayer architectures of 2D DNA sheets were also identified.
Introduction
DNA is condensed in living organisms up to one million fold
and is subsequently unwound during cell division and RNA
transcription. Due to the highly negatively charged nature of the
DNA backbone, this compaction is largely achieved through the
use of positively charged histones and small molecules such as
spermine. DNA and RNA have recently been used as materials
to study the process of controlled self-assembly. A critical aspect
of such efforts is the inclusion of rigid structural motifs such as
double1 and triple2 crossovers in DNA, or 90° motifs3 in RNA.
Another essential feature is the incorporation of complementary
overhangs to direct the desired assembly.3–8 While the designed
features form on the nanoscale, they have been largely limited
to sizes below a micron.3–5 By introducing symmetry into all
complementary overhangs, much larger assemblies can be
generated.9–14 Such large two-dimensional arrays require building
blocks that are sufficiently flexible to constrain all components
* E-mail: reich@chem.ucsb.edu.
§
Department of Chemistry and Biochemistry, University of California - Santa Barbara.
†
Biomolecular Science and Engineering Program, University of California - Santa Barbara.
‡
Veeco Instruments. Current affiliation - Agilent Technologies.
(1) Li, X.; Yang, X.; Qi, J.; Seeman, N. C. J. Am. Chem. Soc. 1996, 118,
6131–6140.
(2) LaBean, T. H.; Yan, H.; Kopatsch, J.; Liu, F.; Winfree, E.; Reif, J. H.;
Seeman, N. C. J. Am. Chem. Soc. 2000, 122, 1848–1860.
(3) Chworos, A.; Severcan, I.; Koyfman, A. Y.; Weinkam, P.; Oroudjev, E.;
Hansma, H. G.; Jaeger, L. Science 2004, 306, 2068–2072.
(4) Ding, B.; Sha, R.; Seeman, N. C. J. Am. Chem. Soc. 2004, 126, 10230–
10231.
(5) Chelyapov, N.; Brun, Y.; Gopalkrishnan, M.; Reishus, D.; Shaw, B.;
Adleman, L. J. Am. Chem. Soc. 2004, 126, 13924–13925.
(6) Yan, H.; Park, S. H.; Finkelstein, G.; Reif, J. H.; LaBean, T. H. Science
2003, 301, 1882–1884.
(7) Park, S. H.; Pistol, C.; Ahn, S. J.; Reif, J. H.; Lebeck, A. R.; Dwyer, C.;
LaBean, T. H. Angew. Chem., Int. Ed. 2006, 45, 735–739.
(8) Seeman, N. C. Nature 2003, 421, 427–431.
(9) He, Y.; Chen, Y.; Liu, H.; Ribbe, A. E.; Mao, C. J. Am. Chem. Soc. 2005,
127, 12202–12203.
(10) He, Y.; Tian, Y.; Chen, Y.; Deng, Z.; Ribbe, A. E.; Mao, C. Angew.
Chem., Int. Ed. 2005, 44, 6694–6696.
(11) He, Y.; Tian, Y.; Ribbe, A. E.; Mao, C. J. Am. Chem. Soc. 2006, 128,
15978–15979.
(12) He, Y.; Mao, C. Chem. Commun. 2006, 968–969.
(13) He, Y.; Tian, Y.; Chen, Y.; Ribbe, A. E.; Mao, C. Chem. Commun. 2007,
165–167.
in one plane. A new three-point-star motif was recently determined
to have the correct balance of flexibility and stress to generate
rigid and well defined structures.9,12 Systematic increases of the
number of branches in the monomer unit (3 f 4 f 6) generated
arrays which were still restricted to a plane.9–11
One- and two-dimensional nucleic acid architectures have been
used as scaffolds for positioning proteins and inorganic materials6–8,14–17 and for macromolecule detection.18,19 The assembly
of such structures in three dimensions and their temporal control
remain significant challenges. Three-dimensional architectures
can increase the density of guest molecule packing for structure
determination, high-density information storage, localization of
functional proteins next to each other for catalysis, and electronic
and optical applications. Nanometer-scale three-dimensional
objects of defined size have been constructed from DNA20–24
and used to position a protein within the architecture.25 Continuous
three-dimensional DNA lattices with solvent channels that can
accommodate small proteins have been crystallized and function
as molecular sieves.26,27 Enzyme-catalyzed DNA hydrogels
assembled from short strands of cDNA were used for controlled
(14) He, Y.; Tian, Y.; Ribbe, A. E.; Mao, C. J. Am. Chem. Soc. 2006, 128,
12664–12665.
(15) Koyfman, A. Y.; Braun, G.; Magonov, S.; Chworos, A.; Reich, N. O.;
Jaeger, L. J. Am. Chem. Soc. 2005, 127, 11886–11887.
(16) Le, J. D.; Pinto, Y.; Seeman, N. C.; Musier-Forsyth, K.; Taton, T. A.;
Kiehl, R. A. Nano Lett. 2004, 4, 2343–2347.
(17) Zheng, J.; Constantinou, P. E.; Micheel, C.; Alivisatos, A. P.; Kiehl,
R. A.; Seeman, N. C. Nano Lett. 2006, 6, 1502–1504.
(18) Lin, C.; Katilius, E.; Liu, Y.; Zhang, J.; Yan, H. Angew. Chem., Int. Ed.
2006, 45, 5296–5301.
(19) Lin, C.; Liu, Y.; Yan, H. Nano Lett. 2007, 7, 507–512.
(20) Shih, W. M.; Quispe, J. D.; Joyce, G. F. Nature 2004, 427, 618–621.
(21) He, Y.; Ye, T.; Su, M.; Zhang, C.; Ribbe, A. E.; Jiang, W.; Mao, C. D.
Nature 2008, 452, 198–201.
(22) Rothemund, P. W. K. Nature 2006, 440, 297–302.
(23) Goodman, R. P.; Schaap, I. A. T.; Tardin, C. F.; Erben, C. M.; Berry,
R. M.; Schmidt, C. F.; Turberfield, A. J. Science 2005, 310, 1661–1665.
(24) Erben, C. M.; Goodman, R. P.; Turberfield, A. J. J. Am. Chem. Soc. 2007,
129, 6992–6993.
(25) Erben, C. M.; Goodman, R. P.; Turberfield, A. J. Angew. Chem., Int. Ed.
2006, 45, 7414–7417.
(26) Paukstelis, P. J.; Nowakowski, J.; Birktoft, J. J.; Seeman, N. C. Chem.
Biol. 2004, 11, 1119–1126.
(27) Paukstelis, P. J. J. Am. Chem. Soc. 2006, 128, 6794–6795.
10.1021/la801306j CCC: $40.75  2009 American Chemical Society
Published on Web 07/11/2008
1092 Langmuir, Vol. 25, No. 2, 2009
Koyfman et al.
Figure 1. Nucleic acid array formation at various Mg2+ ion concentrations (A) Representative arrays formed at high (10 and 15 mM) Mg2+ concentrations.
(B-D) Representative arrays formed at low (2 and 5 mM) Mg2+ concentrations. (E) Graphical representation of domain sizes formed at various Mg2+
concentrations. Domain size (µm) was determined as the square root of the area occupied by the domain.
drug release.28 DNA has been covalently coupled to organic
materials to produce novel optical, electronic, and mechanical
properties.29,30 An attractive goal with these three-dimensional
architectures is the construction of objects on the micrometer to
millimeter scale.
Our interest lies in the directed incorporation of inorganic
materials within nucleic acid architectures for fundamental studies
as well as various applications. We previously incorporated 5
nm positively charged gold nanoparticles within RNA scaffolds.15
Mao and colleagues9 recently showed that DNA provides an
excellent scaffold to increase the regularity and extent of
nanoparticle assembly on a nucleic acid array. In our efforts to
use this new assembly approach,9 we identified novel architectures
which depended on the assembly conditions. Herein we report
on the formation of highly regular interarray multilayer packing
from hexagonal two-dimensional arrays. The extent of array and
multilayer formation is highly dependent on Mg2+ concentration.
Results and Discussion
The dependence of nucleic acid array formation from three
single-stranded DNA sequences on the concentration of Mg2+
ions is illustrated in Figure 1. Slow cooling of a mixture of three
single-stranded DNA sequences (see Supporting Information)
in the presence of Mg2+ results in formation of 3-point star motifs9
(monomer units of a hexagonal DNA array shown in Figure 3A,
bottom right) that self-assemble into 2D hexagonal DNA arrays
which furthermore form diverse multilayer architectures. Prior
work with the same DNA strands at ∼10 mM Mg2+ using
fluorescent microscopy showed millimeter long arrays.9,10 This
same work demonstrated that arrays observed by atomic force
microscopy were generally smaller than arrays observed by
fluorescent microscopy since the interaction with a mica surface
is potentially destructive to extensive nucleic acid arrays.9–14
Here we show that at high Mg2+ concentrations of 10 and 15 mM
(Figure 1A) DNA strands form micron-sized arrays, 40 ( 20%
(28) Um, S. H.; Lee, J. B.; Park, N.; Kwon, S. Y.; Umbach, C. C.; Luo, D.
Nat. Mater. 2006, 5, 797–801.
(29) Chenoweth, D. M.; Viger, A.; Dervan, P. B. J. Am. Chem. Soc. 2007, 129,
2216–2217.
(30) Janssen, P. G. A.; Vandenbergh, J.; van Dongen, J. L. J.; Meijer, E. W.;
Schenning, A. P. H. J. J. Am. Chem. Soc. 2007, 129, 6078–6079.
Figure 2. Moiré pattern of a single layer array. (A) 5 × 5 µm2 digital
zoom in of a 15 × 15 µm2 scan depicting Moiré patterns. Fast Fourier
transform of image (A) where peaks are outlined by blue, red, and green
boxes (inset). (B) 1.5 × 1.5 µm2 scan of the area outlined within the
white square in (A). Three domains within this scan are indicated by
positioning parallel blue, red, and green lines over them. Inset: Color
coded FFT back transform of the peaks outlined in the inset in (A). Note
that the parallel lines, FFT peaks in inset (A), and domains in inset (B)
are the same colors.
Self-Assembly of DNA Arrays into Multilayer Stacks
Langmuir, Vol. 25, No. 2, 2009 1093
Figure 3. Three predominant orientations of a bilayer assembly. (A) (Top) AFM image showing single and double layer of a DNA hexagonal array.
Top layer of the double layer is translated 17.5 nm with respect to the bottom layer where the overlay of two 3-point star motifs is higher. The distance
between the two adjacent overlays is 30 nm. (Bottom) A diagram representation of a 3-point star motif, two layer assembly, and a model depicting
an overlay of two 3-point star motifs. One motif is rotated 180° with respect to another. (B) (Top) AFM image showing two layers of a DNA hexagonal
array rotated 30° with respect to one another. (Inset) Fourier transform of the AFM image indicating that the rotation between the two layers is 30°.
(Bottom) A diagram representation of a two layer assembly rotated 30° with respect to one another. Blue, red, and black dots on the AFM image
and the model correspond to provide the same pattern. (C) (Top) AFM image showing two layers of a DNA hexagonal array rotated 20° with respect
to one another. (Inset) Fourier transform of the AFM image indicating that the rotation between the two layers is 20°. (Bottom) A diagram representation
of a two layer assembly rotated 20° with respect to one another. Blue and red dots on the AFM image and the model correspond to provide the same
pattern.
of the total area of which occurs in multiple layers. Moiré patterns
discussed in Figure 2 are evident in Figure 1A at the scan size
above 10 µm. Different nucleic acid architectures were observed
at low (2 and 5 mM) Mg2+ concentrations (Figure 1B-D). Under
low Mg2+ concentrations, DNA strands form predominantly
submicron bilayer arrays of a defined geometrical shape
resembling rhombuses, triangles, or rectangles with sharply
defined edges. Arrays formed at low Mg2+ concentrations are
mostly composed of two DNA layers positioned on top of each
other at well-defined orientations. Occasionally, pyramidal
architectures composed of multiple layers are observed at low
Mg2+ concentrations (Figures 1B and 5). The array size at low
Mg2+ concentrations is not altered if the Mg2+ is subsequently
increased to 15 mM, suggesting that the arrays are not selectively
bound to the mica surface at different Mg2+ concentrations. It
is interesting to note that the density of array occurrence on the
mica surface is highly Mg2+ dependent. At 2 mM Mg2+ submicron
bilayers are separated tens of microns away from one another
(Supporting Information Figure S1A-C). The array density is
increased when the Mg2+ concentration is increased to 15 mM
indicating a better binding of negatively charged DNA arrays
and the mica surface through Mg2+ ions (Supporting Information
Figure S1D-F).
To further quantify the size of the array surface, we determined
the area of the array using Nanoscope software. Since most of
the arrays have similar lateral dimensions, we assumed that arrays
form perfect squares and estimated an array’s size by taking the
square root of the area. Figure 1E shows the distribution of array
sizes depending on the concentration of Mg2+ ions. Representative
arrays (45) were imaged at each concentration of 2, 5, 10 and
15 mM Mg2+. Under low Mg2+ concentrations, DNA strands
form 88% submicron arrays where sizes range from 230 nm to
6.6 µm. Average sizes are 730 nm at 2 mM and 680 nm at 5 mM
Mg2+. At high Mg2+ concentrations, DNA strands form arrays
that range from 390 nm to 20 µm and are 4.7 ( 4.2 µm at 10
mM Mg2+ and 7.0 ( 4.6 µm at 15 mM Mg2+. At high Mg2+
concentrations, 63% of the arrays are above 3 µm in size. Since
arrays at high Mg2+ concentration are evenly distributed in size,
the larger arrays most likely are formed from a number of smaller
domains as Moiré patterns confirm (Figure 2) as well as from
addition of monomer units. DNA domains are well-ordered
isotropic assemblies of monomer subunits with minimal defects.
The assembly melting temperature (Tm) of DNA is known to
increase with increases in Mg2+ concentration.31,32 Since the
extent of the DNA arrays is in part dependent on the melting
temperature, it is likely that lower Mg2+ concentrations should
result in smaller assemblies. The array edges formed at high
Mg2+ concentrations are poorly defined, implying a constant
addition of monomer units by addition polymerization33 (polydispersity index at high Mg2+ concentrations is close to 1) or
other already formed domains, consistent with the Moiré patterns
(Figure 2). At high Mg2+, lateral extension of DNA arrays occurs
along with the multilayer formation occupying about 40 ( 20%
of the total array area. Multilayer formation could happen by
several mechanisms. For example, the extended arrays could
bend and collapse on the mica surface. Smaller arrays could
form closed tubes and form bilayers once collapsed on the mica
surface. Alternatively, domains could stack on top of each other
during the cooling process with Mg2+ ions counteracting negative
phosphates. For example, at lower concentrations of Mg2+ ions,
the ions which may be limiting are more likely to interact between
phosphates of adjacent domains to neutralize the charge. In
contrast, at higher Mg2+ concentrations, the charges are
counteracted within the same DNA strand, therefore helping to
form extended single arrays with millimeter long structures.
It is interesting to note that the DNA arrays are rather stable.
These arrays were imaged on mica surfaces in air before
degradation becomes visible for up to a month (Supporting
Information Figure S2), unlike RNA arrays where damage
becomes apparent after one week.15 No damage to the DNA
(31) Schildkraut, C.; Lifson, S. Biopolymers 1965, 3, 195–208.
(32) Tan, Z.; Chen, S. Biophys. J. 2006, 90, 1175–1190.
(33) Rubinstein, M.; Colby, R. H. Polymer Physics; Oxford University Press:
New York, 2003.
1094 Langmuir, Vol. 25, No. 2, 2009
Koyfman et al.
Figure 4. Modeling and AFM of two- and three-layer assemblies. (A) Folding of a representative DNA sheet along the blue line (top) leads to dense
packing in which three-point-star motifs are positioned on top of each other, rotated 180° and predicted to be separated by 30 nm (shown in red).
Bottom shows a folded DNA sheet positioned on another DNA sheet to generate two- and three-layer assemblies. The distance between two
three-point-star motifs is predicted to decrease to 15 nm. (B) AFM and cross sections of two- and three-layer assemblies. For the two-layer assembly,
the spacing between adjacent three-point-star motifs varies between 30 and 32 nm and the heights are 1.2 and 2.3 nm for single and double layers
respectively (blue line). For the three-layer assembly, spacing between two three-point-star motifs varies between 14 and 19 nm and the heights are
1.3 and 2.4 nm for the single and three-layer assemblies respectively (red line). Green triangles enclose six and ten three-point-star motifs for double
and triple layers respectively; the corresponding models are discussed in the text. DNA is cubic close-packed within the triple-layer assemblies. (C)
Distributions of distances separating adjacent locations of the two three-point-star motifs positioned on one another for the double layer (blue, n )
276) and triple layer (red, n ) 291). Distance separating adjacent locations of the two three-point-star motifs positioned on one another is 16.2 nm
for the triple layer and 32.2 nm for the double layer.
arrays is observed when an aliquot of the solution containing
DNA arrays is placed on the mica surface and dried after being
stored at 4 °C for multiple months.
Figure 2 shows a Moiré pattern of a single layer array where
we observe distinct orientations of DNA domains within the
single layer of a DNA sheet. Moiré pattern analysis provides a
robust approach to identify long-range order as well as the size
and orientation of domains. Moiré patterns are observed at the
scan size of 10-20 µm due to the interference between the AFM
scan lines which act as a reference grating, and closely packed
hexagonal array architectures which form the sample lattice
grating34 (Figure 1A). To obtain Moiré patterns, the AFM grating
pitch PG should be similar to the integer multiple of the sample
lattice pitch PL, PG ≈ nPL. The 15 µm by 15 µm image was
scanned at a sampling rate of 512 pixels. Therefore, the grating
pitch was PG ) 15 µm/512 pixels ) 29.3 nm/pixel, very close
to the 30.3 nm sample lattice pitch PL. Analysis of similar patterns
was used to detect structural defects formed in the microphase(34) Hexemer, A.; Stein, G. E.; Kramer, E. J.; Magonov, S. Macromolecules
2005, 38, 7083–7089.
separation pattern common for thin films of block copolymers.34
Figure 2A is a 5 × 5 µm2 digital magnification of a 15 × 15 µm2
scan of Moiré patterns. The inset in the bottom right corner is
a fast Fourier transform (FFT) of the image in Figure 2A. The
colored rectangles in Figure 2A were iteratively optimized (size
and position) to maximize both the individual and overall coverage
of oriented domains shown in Figure 2B. Three distinctive FFT
peaks are outlined by blue, red, and green boxes and correspond
to distinct orientations of DNA domains within the single layer
of a DNA sheet. The inset at the top left corner of Figure 2B
shows a color-coded inverse FFT of only the FFT peaks in which
the peak outlines and domains have the same color. Figure 2B
shows a 1.5 × 1.5 µm2 scan of the area outlined within the white
square in Figure 2A. Three domains within this scan are indicated
by positioning parallel blue, red, and green lines over them which
correspond to the colors of the FFT peak outlines and domains
in the inverse FFT.
Figure 3 shows three predominant orientations of a two-layer
assembly. Figure 3A (top) shows an AFM image of a single
layer of a DNA hexagonal array in the bottom right corner. A
Self-Assembly of DNA Arrays into Multilayer Stacks
Figure 5. AFM and cross section of two-layer and multilayer assemblies.
Multilayer assemblies form at low Mg2+ concentration. The cross section
(black line) was shown to represent the height of a two-layer assembly
(2.5 nm) rotated 30° with respect to one another and a taller multilayer
assembly that is 8.8 nm in height.
double layer is formed when a DNA sheet is folded on itself
along the edge in the top left corner of the image. This is a face
centered cubic orientation,5 since a 3-point-star motif is located
within the hexagon formed by other six 3-point-star motifs. This
orientation results from the translation or 180° rotation of the top
layer (blue) with respect to the bottom layer (green) (Figure
3A(bottom)). A model of the overlay of two 3-point-star motifs
rotated 180° with respect to another is shown in Figure 3A
(bottom). A zoom in of the overlay of two 3-point-star motifs
rotated 180° with respect to another is shown in Figure 3A
(bottom). The distance between the two opposite edges of a
hexagon formed from six 3-point-star motifs is predicted to be
30.3 nm calculated by assuming 0.33 nm/base pair pitch and 2
nm diameter in a DNA helix.9 Assuming a 30.3 nm distance
between the two opposite edges of a hexagon, the top of the
double layer is translated 17.5 nm (15.15 nm/cos 30) with respect
to the bottom layer. The overlay of two 3-point-star motifs
generates a double layer which is higher than a single layer. The
distance between the two adjacent overlays of 3-point-star motifs
of similar height (indicated in red on the model in Figure 3A
bottom) overlays is predicted to be 30.3 nm. The actual distance
between adjacent 3-point-star motifs is consistent with the model
and was 32.1 ( 2.2 nm (N ) 130).
Figure 3B (top) shows an AFM image of two layers of a DNA
hexagonal array rotated 30° with respect to one another. The
inset is a Fourier transform of the AFM image indicating that
the rotation between the two layers is 30°, since 12 dots are
equidistant from one another. The bottom of Figure 3B shows
Langmuir, Vol. 25, No. 2, 2009 1095
a diagram of a two-layer assembly where two sheets are rotated
30° with respect to one another. Blue, red, and black dots on the
AFM image and the model correspond and show the same pattern.
The two sets of 12 red dots are enclosed within the two sets of
5 dots. The first set of 5 dots is colored blue and is indicated for
the set of 12 red dots on the right. Only two members of the
second set of 5 dots are indicated in black.
Figure 3C (top) shows an AFM image of two layers of a DNA
hexagonal array rotated 20° with respect to one another. The
inset is a Fourier transform of the AFM image indicating that
the rotation between the two layers is 20°, since the two sets of
six dots are separated by 20°. The bottom diagram represents a
two-layer assembly where the two layers are rotated 20° with
respect to one another. Blue (12) and red (5) dots in the AFM
image and the model reveal the same patterns.
Figure 4A represents a model of two- and three-layer
assemblies. The top part of Figure 4A depicts a model of a DNA
sheet that is folded on itself along the blue line. Upon folding,
DNA becomes more densely packed, where three-point-star
motifs are positioned within hexagons and on top of other threepoint-star motifs and are separated by 30.3 nm. The overlay of
two three-point-star motifs is indicated in red. The bottom of
Figure 4A shows a model of a folded DNA sheet positioned on
another DNA sheet to generate an assembly of two and three
layers. The unoccupied three-point-star motif, located within
formed hexagon, was positioned on another three-point-star
motif of the third layer via a 180° rotation. Upon the formation
of the third layer, the distance between the adjacent interactions
of the two three-point-star motifs is predicted to be half as much
at ∼15 nm.
An AFM image and cross sections of a two- and three-layer
assembly are shown in Figure 4B. The blue line crosses over a
two-layer assembly to show the height of the DNA architectures.
The spacing between adjacent locations of the two three-pointstar motifs is consistent with modeling and varies between 30
and 32 nm. The height of a single layer is 1.2 nm, while the
height of the double layer is 2.3 nm. The red line crosses over
a three-layer assembly and shows that the spacing between
adjacent locations of the two three-point-star motifs is consistent
with modeling and varies between 14 and 19 nm. The height of
a single layer is 1.3 nm, while the height of the triple layer is
2.4 nm. Since the heights of double and triple layers are similar,
we suggest that the third layer is packed within cavities of the
bottom two layers. This arrangement of three DNA layers is
similar to the cubic close-packing (CCP) of metal atoms since
the third layer of the model indicated in black in the triangular
inset is rotated 180° with respect to the first sheet (green). The
top DNA sheet lies in a unique position different from the bottom
two sheets creating an “ABC” close-packed layer sequence. Green
triangles enclose six and ten three-point-star motifs positioned
on one another for double and triple layers respectively.
Furthermore, DNA height measurements of DNA helices by
AFM often reveal values below 2 nm, presumably as a result of
the forces exerted by the vibrating tip in tapping mode.
To further quantify the distance separating adjacent locations
of the two three-point-star motifs that occur in different layers,
276 and 291 measurements were taken along the double layer
and triple layers respectively. The distribution of these distances
is shown in Figure 4C. Adjacent locations of the two threepoint-star motifs in different layers but positioned on one another
is 16.2 nm with a standard deviation of 2.0 nm for the triple
layer. For the double layer, the distance was twice as much, 32.2
nm, with a standard deviation of 1.7 nm. The measured distances
are in good agreement with the values of the model where the
1096 Langmuir, Vol. 25, No. 2, 2009
overlay of two three-point-star motifs are 30.3 and 15.15 nm for
the double and triple layers respectively.
Figure 5 demonstrates the occurrence of pyramidal architectures
containing more than two or three layers at low concentrations
of Mg2+ ions. These multilayer assemblies appear in about 3%
of the total number of structures. The number of layers in such
a multilayer structure cannot be determined due to compaction
of the DNA sheets. Based on the height of two and three layer
structure (Figure 4) we estimate the multilayer structures in
Figures 5 and 1B to contain anywhere from 5 to 11 layers.
Conclusions
In conclusion, we describe the formation of multilayer DNA
structures from three-point-star motifs which self-assemble into
hexagons and hexagonal 2D sheets. The formation of smaller
multilayer structures of a defined geometrical shape was observed
at low Mg2+ concentration while the formation of larger singlelayer and multilayer 2D sheets was observed at higher Mg2+
concentration. Moiré patterns were observed when the images
of single layer architectures were taken on the 10-20 µm scan
and were useful for determination of the domains within the
architectures. The most predominant orientations of a two-layer
assembly were determined to be translation of 17.5 nm of one
layer with respect to another where 3-point-star motifs are
positioned 30 nm with respect to one another, 30° and 20° rotations
Koyfman et al.
of the two layers. Three-layer assemblies with cubic close-packing
are even more densely packed where 3-point-star motifs are
separated by 15 nm. A small portion of assemblies at low Mg2+
concentration form multilayer structures with more than three
layers. High resolution visualization of these structures is beyond
the capability of atomic force microscopy, but we believe threedimensional architectures to be one of the exciting directions of
the nucleic acid nanotechnology. Unlike the formation of DNA
crystals or DNA hydrogels, layer by layer assembly has an
advantage as various layers can pack within three-dimensional
structures, thereby providing different pore sizes for guest
molecule incorporation.
Acknowledgment. Funding for this work was provided by
the Institute for Collaborative Biotechnologies (UCSB) to N.
Reich. We wish to thank Gary Braun and Dr. Michael Rubinstein
for helpful discussions and critical review of the manuscript, and
Dr. Sergey Belikov for help with LabView for the Moiré pattern
analysis.
Supporting Information Available: Materials and Methods,
density dependence of DNA arrays, and degradation of multilayer DNA
arrays are presented. This material is available free of charge via the
Internet at http://pubs.acs.org.
LA801306J