Human Endothelial Cells Synthesize, Process, and

From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
Human Endothelial Cells Synthesize, Process, and Secrete Fibronectin Molecules
Bearing an Alternatively Spliced Type I11 Homology (ED1)
By John H. Peters, Lee Ann Sporn, Mark H. Ginsberg, and Denisa D. Wagner
Cellular fibronectin (Fn) bearing an alternatively spliced
extra type 111 structural repeat (ED11 is normally present at
low concentrations in blood plasma. The source of this
material remains uncertain. In this study, primary cultures
of human umbilical vein endothelial cells (HUVEC) labeled
with 36S-methioninewere observed to synthesize Fn monomers both with and without this segment. Monomers
isolated from cell lysates with antibodies to the ED1
sequence comigrated in nonreduced sodium dodecyl sulfate polyacrylamide gel electrophoresis with the slower
(designated M11, but not the faster (designatedM2). of two
major monomeric populations that were recognized by
antibodies raised t o plasma-derived Fn. The differences
between M I and M2 were not due to glycosylation. since
they were also observed between species of Fn monomer
purified from cells grown in the presence of tunicamycin.
M I and M2 were both observed to incorporate with a
similar rate into dimeric Fn, indicating that Fn monomers
with and without the ED1 site can dimerize with similar
efficiency. Analysis of reduced samples of Fn isolated from
cells with anti-ED1 antibodies indicated the presence of
both M I - M I and MI-M2 dimers. In addition to being
incorporated into extracellular matrix, ED1 Fn was observed to be secreted in soluble form into the medium,
potentially reflecting intravascular release of this protein
by endothelial cells in vivo.
0 1990 by The American Society o f Hematology.
F
the other sites of alternative splicing in primary Fn transcripts.* Although human endothelial cells have been shown
with immunofluorescent microscopy to incorporate ED1 +
Fn into their pericellular matrix? metabolic labeling studies
have not previously been conducted to demonstrate synthesis
of ED1 + Fn. Because endothelial cells may be a source of
circulating ED1 + Fn, we used antibodies specific to the ED1
segment to study the synthesis and intracellular processing of
Fn polypeptides in primary cultures of human umbilical vein
endothelial cells (HUVEC).
IBRONECTINS (Fns) are adhesive glycoproteins which,
as components of extracellular matrices and blood
clots, promote cell-cell and cell-matrix interactions that are
vital to processes of tissue construction and reconstruction
(eg, embryogenesis and wound healing).lS2Three known
regions within the Fn gene encode sequences, designated a t
the protein level as E D l , ED2, and IIICS (also referred to as
EIIIA, EIIIB, and V), which are subject to alternative
splicing a t the level of the primary t r a n ~ c r i p t . Sequences
~.~
encoding the ED1 and ED2 segments may be totally included
or excluded in Fn messenger R N A (mRNA), whereas the
sequence encoding the IIICS segment is subject to partial
inclusion. The ED1 and ED2 sequences are both encoded by
single exons and each constitutes exactly one type I11
structural
Because the bulk of circulating Fn is
produced by the liver,7 and the ED (extra domain) segments
are not detectably encoded by Fn mRNA in normal hepatocytes, these segments have been considered to characterize
cellular or tissue F ~ s . ~However,
.*
low concentrations of Fn
bearing the ED1 segment (ED1 + Fn) are normally present
in blood plasma.’-” Although the source of this material
remains uncertain, the observation that ED1 + Fn appears
by immunohistologic analysis to be localized to blood vessel
endothelia in mature tissues has prompted speculation that
some or all circulating ED 1+ Fn may normally be produced
by endothelial cells.’
Fn biosynthesis has been studied in several types of cells,
including fibroblasts, endothelial cells, hepatocytes, amniotic
fluid cells, as well as transformed and neoplastic cell^.^*'*-'^
Fibroblasts synthesize monomeric Fn polypeptide chains to
which asparagine-linked carbohydrate side chains are rapidly added. Within minutes of synthesis, these monomers are
assembled into dimers and secreted a t the cell surface after
processing of high mannose to complex carbohydrate.” In
most of the cell types that have been studied to date, Fn has
similarly been shown to be secreted in dimeric form rapidly
after synthesis, leaving no large intracellular
ED1 +
Fn has been shown to be synthesized by fibroblasts and
fibrosarcoma
Analysis of mature dimeric Fn secreted
by fibroblasts has shown that these cells produce multiple
similar but nonidentical monomeric subunits bearing the
ED1 segment as a consequence of differential processing at
Blood, Vol 75, No 9 (May 1). 1990: pp 1801-1808
+
MATERIALS AND METHODS
Primary cultures of endothelial cells were obtained from human umbilical veins by mild
proteolytic digestion as described,*’ and cultured in McCoy’s 5A
medium (Flow Laboratories, McLean, VA) containing 20% fetal
bovine serum. For continuous labeling, cells were grown in the
presence of L-(”S)methioninine (New England Nuclear, Boston,
MA) at 20 pCi/mL. For pulse labeling, confluent cells were washed
twice with Hank‘s balanced salt solution and incubated for 15
minutes at 37OC with media lacking serum. The cells were then
Cells and cell culture conditions.
From the Department of Immunology, Research Insiitute of
Scripps Clinic, La Jolla; Pulmonary Division, Department of
Medicine. University of California, San Diego; Hematology Unit,
Department of Medicine, University of Rochester School of Medicine and Dentistry, NY: and Hematology Division, Department of
Medicine, Tufts University School of Medicine, Boston, MA
Submitted August 24, 1989: accepted January 2, 1990.
Supported by US Public Health Service Grants GM-08172,
HL-23584 (University of California. San Diego. Specialized Center
for Research), GM-37696. HL-28235. HL-30616, and MOIRR00833. D.D. W. is an Established Investigator of the American
Heart Association. This is publication IMM-5573 from the Department of Immunology, Research Institute of Scripps Clinic.
Address reprint requests to John H. Peters, MD. Center for
Cancer Research, E l 7-225, Massachusetts Institute of Technology,
77 Massachusetts Ave. Cambridge, M A 02139.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
“advertisement” in accordance with 18 U.S.C.section I734 solely to
indicate this fact.
0 1990 by The American Society of Hematology.
0006-4971/90/7509-0010$3.00/0
1801
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
1802
covered with L-(%)methionine at 0.8 to 1.5 mCi/mL in media
without serum for periods varying from 5 to 20 minutes. After pulse
labeling, the cells were washed once with complete media and then
again in Hank’s solution, and lysed immediately or covered with
fresh unlabeled complete media for varying chase periods at 37OC.
Antisera. The preparation and characterization of goat antisera
raised against a 29-amino acid synthetic peptide of the ED1 region of
human Fn and against Fn derived from rabbit plasma were described
previously.” The goat antiserum to rabbit plasma Fn was crossreactive with human Fn as judged by Western blotting and enzymelinked immunosorbent assays (ELISA).
Monoclonal antibody 52/54. Balb C mice were immunized with
the 29-amino acid synthetic ED1 peptide-KLH conjugate used to
raise antiserum to the ED1 in a goat.” The priming dose of antigen
was 25 pg emulsified in complete Freund’s adjuvant and injected
subcutaneously. This was followed in 18, 25, and 32 days by
intraperitoneal injections of 25-pg antigen. Sera were then checked
for preferential antibody titers for cellular versus plasma-derived
Fns in ELISA. Parallel microtiter wells were coated either with Fn
(3 pg/mL in 0.1 mol/L NaHCO,) purified by gelatin chromotography” from the media of GM-1380 fibroblasts grown in the
presence of Fn-depleted fetal bovine serum or from human plasma.
Duplicate dilutions of sera were then applied to the Fn-coated wells,
followed after a washing step with peroxidase-conjugated goat
antibodies to mouse immunoglobulins (Igs) (Tago Inc, Burlingame,
CA). One mouse had serum which, diluted 1:100,000, bound to wells
coated with cellular but not plasma Fn. On the 46th day after initial
immunization, it was given an intravenous booster of 50-pg antigen.
Five days later the spleen was taken and the washed spleen cell
suspension was fused with SP-2 myeloma cells at a 5:l ratio. The
fused cells were grown in hypoxanthine-aminopterin-thymidine
medium before distribution in microtiter wells. Fifty-one colonies of
cells were reactive with human cellular Fn-coated wells by ELISA.
After two sublconing steps at limiting dilutions of 0.5 cell per well, in
which screening was accomplished by reactivity with ED1 peptide in
ELISAs, 25 cell lines were selected. The 52/54 clone, when injected
into pristine-primed mice, produced ascites that contained antibodies which preferentially reacted with cellular Fn in ELISA and
Western blotting assays. When subjected to isotyping using a
ScreenType ELISA kit (Boehringer Mannheim Biochemicals, Indianapolis, IN), supernatant from this clone contained both IgGl and
IgG2b antibodies, both with K light chains. Despite a third subcloning at limiting dilutions of 0.5 cell per microtiter well, it was not
possible to isolate distinct IgGl- and IgG2b-producing clones.
However, both antibodies, purified and eluted by differential pH
from Protein A sepharose (Pharmacia Fine Chemicals, Piscataway,
NJ), specifically reacted with cellular Fn as disclosed in ELISA
assays using second isotype-specific antibodies (Tago Inc), and the
IgGl antibody competed for binding to cellular Fn with the IgG2b
antibody. Therefore, the 52/54 clone produces two different antibodies sharing the same determinant of the EDI.
Immunopurifcation of Fn. Cells were lysed as described so that
the final concentration of ingredients was that of the buffer used for
washing the immunoprecipitate.20Media was centrifuged at 1,000 x
g for 5 minutes to remove cells. To immunopurify Fn, Protein
A-Sepharose C1-4B (Pharmacia) was preincubated with gentle
rocking for 30 minutes with goat anti-Fn antisera, and then
incubated with samples of cell lysate or media for 90 minutes at
room temperature. Each sample of cell lysate was divided such that
75% to 80% of the volume was incubated with Protein A beads (27
mg/25-cm2 flask) that had been preincubated with 3.0 pL/mg
antiserum to the ED1 segment, whereas the remaining 20% to 25%
of the volume was incubated with beads (40 mg/flask) that had been
preincubated with 1.6 pL/mg antiserum to plasma-derived Fn.
Similarly, each sample of media was divided such that 80% of the
PETERS ET AL
volume was incubated with beads (20 mg/mL media) that had been
preincubated with 5.9 pL/mg antiserum to the ED1 segment, while
the remaining 20% was incubated with beads (80 mg/mL media)
that had been preincubated with 1.3 pL/mg antiserum to plasmaderived Fn. After incubation, the beads were washed extensively and
boiled in electrophoresis sample buffer to elute bound Fns for
analysis by gel electrophoresis. These procedures did not result in
quantitative removal of either ED1 + Fn or total Fn from samples, as
judged by the capacity of a second application of each method to
yield additional labeled material possessing identical electrophoretic
mobility to that obtained with either antibody in the first instance.
To demonstrate that immunopurification of labeled ED1 + Fn
from culture medium was specific, Protein A beads were split in half
after preincubation with anti-ED1 antiserum. One resulting portion
was preincubated for 1 hour at room temperature with 1.2 mL of
synthetic ED1 peptide” (167 pg/mL in phosphate-buffered saline
[PBS] corrected to a pH of 7.4 with NaOH), while the other was
incubated with the same volume of PBS (pH 7.4) alone. Equal
volumes of media obtained from cells that had been subjected to
continuous labeling for 3 days were then added to each of the two
portions of beads (plus or minus peptide) and incubated as described
above.
Electrophoresis gels. Five percent sodium dodecyl sulfatepolyacrylamide (SDS-PAGE) gels were prepared as described by
Laemmli.” I4C-labeled proteins (Amersham International, Amersham, UK) were used as molecular weight standards. After electrophoresis, gels were submitted to autoradiography, and the radiolabeled species of Fn within samples that had been immunopurified
with either anti-ED1 or anti-plasma Fn antibodies were quantified
by densitometry.
Tunicamycin treatment. To assess the effect of inhibition of
N-linked glycosylation on newly synthesized Fn, confluent layers of
HUVEC were grown in the presence of culture medium containing a
final concentration of 1 pg/mL tunicamycin (Calbiochem, San
Diego, CA), using a freshly made stock solution at 1 mg/mL in
DMSO. The cells were incubated with tunicamycin for 15 hours
before pulse labeling experiments, as well as during the period of
labeling.
Endoglycosidase H digestion. Purified Fn was diluted in 0.1
mol/L Tris buffer, pH 5.8, so that the final concentration of SDS was
less than 0.5%. Endoglycosidase H (Calbiochem) was added (3
pg/mL) and samples were incubated for 2 hours at 37OC before
analysis on gels.
Immunojuorescence. Cells to be examined were cultured entirely in media containing fetal bovine serum that had been depleted
of Fn by passage of 500 mL serum over a 100”
column of gelatin
sepharose. Cell fixation, permeabilization, and fluorescent staining
were performed as previously described.” Individual coverslips were
double-stained with (1) McAb 52/54 ascites (1:lO dilution) followed
by rhodamine-conjugated goat antibodies to mouse IgG to detect
ED1 + Fn; and (2) rabbit anti-human plasma Fn antiserum (1:lOO
dilution) (Calbiochem) followed by fluorescein-conjugated goat
antibodies to rabbit IgG to stain the total pool of extracellular Fn.
RESULTS
Intracellular processing of Fn. Newly formed Fn monomers have been shown to incorporate into dimers within
minutes of synthesis in fibroblasts.” Therefore, we initiated
our analysis of t h e synthesis a n d processing of Fn polypeptides in primary cultures of HUVEC by pulse-labeling these
cells with ”S-methionine for a 10-minute period immediately
followed by lysis. The total pool of Fn immunopurified from a
portion of the resulting lysate with antibodies to plasmaderived Fn was comprised of three distinct bands when
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
EO1 + FleRONECnN FROM ENDOTWEUAL Qus
1803
a ~ l y n~o d
n d u c d in SDS-PAGE (Fig 1. lane I). All
three bands migrated more slowly than the 200-Kd myosin
molecular weight standard. The slowest migrating of these
comigrated with Fn immunopurified from the medium of
HLEVEC that had been subjected to 3 days of continuous
labeling (not shown). indicating that this band represented
dimeric Fn. The two faster-migratingbands. designated M I
and M2. correspondedto monomeric Fn. Fn immunopurified
from the companion fraction of cell lysate with antibodies to
the ED1 segment contained only material comigrating with
dimeric Fn and the slower-migratingof the two monomeric
species (Fig 1. lane 2). When subjected to gel analysis
reduced. the three species comprising total Fn resolved into
Fb1.
-dw-dwFn-h-
-
a h r a d MWEC .*r 0 1 o d " prkd d o x p o a ~ oto
"s-nwdadno. comp#(n d u~B
d0
6%
8os
@ of Fn I"unoputci.d
o d -om
LoF m kmwnop*)(kd ffwn npormo otiqwts of -to
by
Omto M.nu-d.rk.d Fn I,,cm l wd 3j oI to th. Lot
..gm~n
12 .nd4) -0
.rw)yndd ~ . d u d n (mj
g Oc
r.dut(ng ( 0 )u
"
.
)n .db)tb to dknork Fn (0).two -)or
.p.c(..of Fn mo"u (M1wd M21mo w.rr(
Innonoduwd
umpk..Am4-ED1 o d b o d o s hdnd only 0 Jowr -atpopdotion of " u s
t h t camlglotd with M1. U n d u r.dudng
m m th. two - ) ~ r rp.~krd F,,
.O.h
rowhd.
v
m
w
-.
two major b a d that migrated more slowly than the two
monomeric species o k r v e d under nonducing conditions
(Fig 1. lane 3). indicating that at least some intrachain
disulfide b o d . formed before dimerization of Fn monomers.
Although antibodies to the ED1 segment isolated Fn that
comigrated primarily with the slower migrating population
in rtduced electrophotesis (Fig 1. lane 4). densitometric
analysis uniformly showed that a portion (approximately
10%)of this material comigrated with the faster migratingof
the two species (see Fig 4).
It was possible that the difference in apparent molecular
mass betwccn M I and M2 might result from dilierential
carbohydrate processing. sincc Fn is a glycoprotein containing 5% to 7'3 carbohydrate. most or all of which appears to be
in asparagine-linked side chains." To address this p i b i l i t y .
cells were pulse-labeled after having been cultured in the
prcsence of tunicamycin. which inhibits the addition of
aspamginelinkedcarbohydrate side chains.*' The two major
monomeric species of Fn. which were immunopurified from
cells cultured in the presence of this agent for I 5 hours before
and during a 20-minute exgosure to '%-methionine. showed
similar increments in rateof electrophoretic migrationnonrcduccd (Fig 2. left panel). and reduced (Fig 2. right panel).
indicatingthat these species do not dimer in size or antigenicity bared on content of N-linked carbohydrate.
To further examine the relationship between the two
apparent monomericspeciesof Fn ( M 1 and 342) in HUVEC.
pulse-chase studies were conducted. Cells were exposed to
'%-methionine for 5 minutes. then immediately lysed or
incubated in unlabeled medium for 5 or 25 minutes before
lysis. With increasing periods of chase. greater quantities of
label were incorporated into dimeric Fn. whereas lesscr
quantities were present in MI and M2. until the latter
species were no longer evident by 25 minutes of chase (Fig
3A). When the monomeric bands, which were observed
under nonducing conditions at 0-. 5-. and 25-minute chase
timepoints. were quantified by densitometry and plotted
versus time. the radioactivity incarporated into MI and M2
was observed to diminish at similar r a t e (not shown). Of
note. no labeled Fn could be immunopurified from the
medium of cells at the 25-minute chase timepoint. indicating
that the newly synthesi7. Fn dimers were still associated
with thecells.
The entirely dimeric intracellular pool of labeled Fn that
was Ken at late chase timepoints (Fig 3A. 25-minute chase)
included slightly slower migrating material not present at
early chase timepoints (Fig 3h. 5-minutechase). The appearance of this material coincided with loss of resolution of two
major Fn subunit populations in reduced gel analysis (Fig
3R. 25-minute chase). In contrast to the faster migrating
dimers evident at this and earlier chase times. the newly
appearing dimeric material was resistant to endoglyoosidase
H. which cleaves the chitobiosylunit of high mannw but not
complex carbohydrateside chains"." (not shown). indicating
that the appearance of this material at this time resulted
from promzing of carbohydrate side chains from high
manMKelocomp1exform.
The composition of inrrocrllulor Fn dhers. As can be
seen by comparison of Fig I and 2. in which Fn was
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
PETERSETAL
1804
-2.
A r r ) y J . d ~ F n ~ p . . M I n
p r h q cuhuos d H U M C m f t r 0 20-"
prbd d oxposw. . t
-o
e %
compo&.
0Utor.dkgl.ph d 0 a h g k 6% 80s "
i
d
.
gd d Fn purtfiod lrom cobgrown u n d u norm01
conditkrn ( I 01 In tho prosma d 1 pglml.
tunk.myth(or 16ha*s t + I. Fn wnpuMod from
oliQUo(s d rvut.whh omCPFn t d f n l o r ml-ED1
(OLD11 onttbodlos. E q w I qwmhk. d pvtfiod Fn
m. Wbj0a.d to .ken*.J.
n0nrodUc.d
~ k h or
l roducod (rlghtl. Ompito roductkr, In J.0.
tho two m o b "ark rg.d..rotokwd tho&
r.(.thn Mr.nt. In w o n t md.cu(.r w
t
n d ~ o n t d E D 1 -molocuh.*rctrutmont.
immunoplrified from cells labeled for 10 or 20 minutes.
respectively. the labeled intracellular pool of Fn was obscrvcd to undergo a major shift between timepoints toward
dimeric as opposed to monomeric forms. This shift was
confirmed by densitometric analysis (Fig 4. lefi panels).
Although the labeled Fn that was isolated by anti-ED1
antibodies at either 10 or 20 minutes did not contain free M2
monomers. this material dissociated on reduction to yield a
small prtion of M2 subunits at both timepints (Fig 4. lower
-3.
Fubo&no.rpr(.
nwmrh0rrlno-d
"ukMlwdM2rg.d..
Into dkrmk Fn. R b y CUItwos d W M C m o orposod
l o %-mothhino for 5 " n o s
md chon Incubotod whh modkm Contotnbng no bbd tor 6
or 25 mirwrtos. (A) Compodto
Ma.dkgr.ph of 0 sknglo 5%
SO8 polyocry)omldo 0.1 of
n0nr.duc.d Fn rg.dw knmunopurW from& Iynmswhh
o h h u ontl-ED1 or mmi-pcnnu
Fn omtbod)... -0
compkto
d ( u g ( m o m o ( fostr-m(glotIngMlondM2rg.d..ottho
25-mlnuto chomo tlmopolnt.
A b not0 oddltkn d .Hght)r
. k m r - m l q ~ t ( n gdlmUk -(e
r(.I mt tho 26-mhMo chooo
tkmpdnt. (01com9oJt.0Utor.dkg..ph d akngIo 5% 808
V g d d r . d u t . d
Fn .p.c(.. hnmUn0gutti.d os
In (AI. Not. .dd(tkn d .kmrnJg"t motu&l whh IOUd
r..oknkr, d M1 wd M2 subunit rg.ck. ot tho 2bmirwrt.
dunti"
right panel). indicating the presence within cells of MI-M2
hctdimers. Of note. the primarily dimeric material isb
latcd with anti-ED1 antibodies at the 20-minute timepoint
resolved on reduction into relatively greater proportions of
MI and lesser proportions of 342 subunits than the corm
spnding material isolated with antibodies to plasmaderived
Fn (Fig 4. upper right panel).
Secretion of ED/ c Fn. For endothelial cells to be a
source of soluble t i 0 1 + Fn circulating in the blood plasma,
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
ED1
+ FIBRONECTIN FROM ENDOTHELIAL CELLS
100
=
r
80-
Dimer
oM1
70-
aM2
4
50 .4
E 40 30 20 -
Y
I
Reduced
-
Fig 4. Proportion of labeled
cellular Fn species after 10 and
20 minutes of exposure to 36smethionine. Cells were exposed
to label for 1 0 or 20 minutes
and lysed. Fns were immunopurified from the cell lysate with
antibodies to plasma-derived Fn
(upper panels) or antibodies to
the ED1 segment (lower panels), and the resulting material
was subjected to electrophoresis nonreduced (left panels) or
reduced (right panels) on 5%
polyacrylamide gels. Bands in
autoradiograms were quantified by densitometry. and each
band within a sample of purified Fn was expressed as the
percentage of total density
within that sample. The values
shown at 10 and 20 minutes
represent the average + 1 SD
of 4 and 3 experiments. respectively. Because anti-ED1 antibodies did not recognize free
monomers within the M2 population, this population is not
represented in the lower left
panel.
1805
k
0
oM1
60
4
I=
10
-0
0
Y
Q1
I-
-
o M1
A
M2
E
10
they should have the capacity both to synthesize and secrete
this material in soluble forms. Therefore, we examined the
capacity of cultured endothelial cells to secrete soluble
E D l + Fn after synthesis. Medium was collected from
primary cultures of HUVEC after 3 days of continuous
labeling with 35S-methionine, and portions were subjected to
immunopurification either with antiserum to plasma-derived
Fn or to the ED1 segment. As shown in Fig 5 , material of
similar electrophoretic mobility, constituting a diffuse band
in reduced electrophoresis, was purified by both types of
antibody. Immunopurification of labeled Fn by antibodies to
the ED1 segment was blocked by coincubation of the Protein
A beads with synthetic ED1 peptide before and during
immunopurification, indicating that purification of ED1 +
Fn was specific. Therefore, cultured HUVEC secrete soluble
ED1 + Fn into the medium after synthesis. Of note, the total
pool of labeled Fn that was secreted into the medium by these
cells migrated slightly slower in nonreduced and reduced
electrophoresis than Fn isolated by gelatin affinity
chromatography*’ from human plasma (not shown).
Matrix incorporation of E D l + Fn. To examine the
capacity of endothelial cells to incorporate endogenous ED1 +
Fn into their extracellular matrix, monoclonal anti-ED1
peptide antibodies were used to examine the subcellular
matrix of primary cultures of HUVEC grown in media
containing fetal bovine serum depleted of Fn. As shown in
Fig 6, ED 1 + Fn was specifically detected by immunofluorescence microscopy in a fibrillar pattern within the extracellular matrix of HUVEC, and this pattern was substantially
10
20
20
Time [minutes)
similar to that detected with antibodies raised to plasmaderived Fn.
DISCUSSION
Forms of Fn bearing the ED1 segment appear by immunohistochemical analysis to be confined mainly to the endothelium of larger blood vessels in normal adult tissues.’ This
observation, coupled with the close association between
endothelial cells and the bloodstream, has led to speculation
that the small quantities of E D I + Fn that are normally
present in plasma may be produced by these cells.’ Our
observations strengthen this hypothesis, since we have found
that primary cultures of HUVEC synthesize, secrete, and
deposit ED1 + Fn within their extracellular matrix. However, these cells do not produce only Fn bearing the ED1 site.
We have observed that HUVEC synthesize two major
populations of Fn monomer (designated M1 and M2), which
differ in electrophoretic mobility. Only the M1 species was
observed to contain molecules bearing the ED1 segment,
indicating that the greater apparent molecular mass of this
population can be attributed at least in part to inclusion of
molecules bearing this alternatively spliced site. The observation that HUVEC synthesize monomers both with and
without the ED1 site is consistent with previous studies in
which other types of cells have been shown to contain species
of Fn mRNA differing in content of the sequence encoding
this site,6 and with the finding that fibroblasts synthesize Fn
subunits both with and without the ED1 segment.*
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
--
1806
m R s ET AI.
(within 10 minutes after initiation of synthesis). However.
diflerenccs in si7e and antigenicity between the two species
were not basal on their content of such carbohydrates. since
these relative ditTerences remained between species immunopurified from cells grown in the presence of tunicamycin.
which inhibits the biosynthesisof the dolichol-linkedprecursor of N-linked side chains." After addition of carbohydrate.
the two populations of monomer were both observed to
incorporate simultaneously and completely into Fn dimers of
similar electrophoretic mobility. The latter observation indicates that: ( I ) the M2 population is not a degradation
product of the MI population; and (2) monomers with and
without the ED1 segment can assemble into dimers with
similar efficiency.
The observation that the labeled material isolated from
cells with anti-ED1 antibodies rcsolved into only dimers and
M I monomers in nonreduced electrophoresis. yet on reduction included some M2 subunits. indicated that some of the
dimers isolated by these antibodies were composed of an M I
subunit linked to an M 2 subunit. Therefore. M I-M2 dimers
bearing only one ED1 site are assembled within HUVEC.
Still. the relative scarcity of M 2 subunits within the reduced
material immunopurified with anti-ED1 antibodies at the
20-minute timepoint (Fig 4, lower right panel) indicated that
M I - M 2 dimers comprised only a small portion of the
primarily dimeric material isolated. Therefore. the balance
of dimers within this material must have been of the M I - M I
variety. The prevalence of M I subunits within dimers immunopurified with anti-ED1 antibodies can probably be attributed to the apparently greater quantities of M I as compared
with M2 monomers available for dimeriration within cells
(Fig 4. upper panels). and the capacity of antibodies IO the
ED1 segment to rmgnize only dimers of M I - M I and
M I 4 2 (but not of M2-M2) composition. Of note. M I - M I
dimers isolated from cells with anti-ED1 antibodies may
have contained either one or two ED1 rites since. in the
absence of quantitative immunopurification. only M I monopeptide:
mers purified with these antibodies can be claimed a priori to
uniformly bear the ED1 segment.
Flg E. Munun mdothdW c o b .#.nED1+ Fn bto tho
W w o d i m . W E C mn Comkwoul)y lm48&d wkh "k
Because the human Fn gene contains at least three sites
mhknkw for 3 doyr. ED1 8 m d totot Fn m o 7 that are subject to alternative splicing at the level of the
from &io
in tho p r o a a o~r~
o b~n m o of spttwtk ED1 poptido as
primary transcript with the potential to generate as many as
doscribod in Motriolr and Mothods. Tho rosuhing purifiod Fns
20 diflerent monomeric forms:'
the observation that priworo subjutod to otmrophor..(r
rod&
in 0 6% SDS polymary cultures of HUVEC synthesite Fn monomers that
wrytomido @, a composite outorodkgroph of which is shown.
Addition of poptido blockd immunopurificmlon of Fn only
resolve into two major species in ondimensional electroanti-ED1 antibodies. confirming thoir spuificlly.
phoresiswas unexpected. However. there is precedent for this
finding. Hepatocyte-derived Fn. the predominant form in
plasma. classically resolves in reduced electrophoresis into a
Both populations of Fn monomer were observed to possess
doublet comprised of two distinct bands corresponding with
rates of electrophoretic migration that were faster under
two populations of mature subunits. Analogous to the situanonreducing as compared with reducing conditions. consistion for M2. the smaller of the two major Fn subunit
tent with a r l y acquisition by Fn polypeptides of intrachain
populations prescnt in the plasma of rodents can be distindisulfide bonds in the endoplasmic reticulum preceding
guished from its larger counterpart by an absence of moledimeri7ation. Of interest. the three regions of Fn that are
cules bearing a specific alternativelyspliced sequence. in this
subject to alternative splicing do not contain sites for
case localized to the IllCS region." Although the twospecies
intrachaindisulfide linkage.'" and therefore diflenntly spliced
of monomer that we have observed to be prcsent in HUVEC
monomers should exhibit similar shifts in mobility between
may each be heterogeneous. potentially including difkrently
nonreduced and reduced electrophoresis. Also. consistent
spliced subspecies of Fn not rcsolved by the gel system uxd.
with previous studies of Fn synthesis in fibroblasts." both of
i t is of interest to note that we have used this same system to
the major monomeric species of Fn were observed to carry
identify as many as four species of Fn monomer. at least WIO
asparagine-linked carbohydrate at very early timepoints
aPFn
aEOl
2
-
250K-
1OOK-
-
+
-
+
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
ED1 + FlBRONECnN FROM ENDOTHELIAL CELLS
1807
Flg6. H u m m n m d o t h . ( i . ( a e N o ~ E D l +
Fnln th.lrp.ricr(kJ.r motrlx In0 p”. M l w t o
“1
oxtrocdhhr Fn. Prkrury c u h r of MUVEC
p b t d on camlips and cutturd for 8 dovm
(wing only modi. aontalning F n - d . p k t d fwal
boyin. swum). Ind)vldllol covrr.lips
MI. thm
fixod ond double-staid whh McAb 62/64 &tom
and anti-humon plosmo Fn antisorum. Sngk W s
wwo thm oxaminod for dhtributkn of toto1 Fn (0
and c) and € 0 1 . Fn Ib ond d). Eoth primary
antibodkr were .ppH.dln M u o k m
oxaminod in a and b) or in tho p r m o f 33 pg/mL ED1
poptld. (W
ouminod in c and d). Only the
staining of ED1 i. Fn W.8 Moclcod by to(mubetion
with ED1 peptido (dl. indlating .9.d(ic noining of
ED1 Fn. B.r = 10 pm.
MI.
o f which contained molecules bearing the ED1 segment. in
similar pulse-chase experiments with human fetal foreskin
fibroblasts (our unpublished results). Therefore. the population of Fn monomers synthesized by primary cultures of
HUVEC appears to be simpler than that which has been
demonstrated to be produced by fibroblasts.’
In this qualitative study. we have found that primary
culturesof human umbilical vein endothelial cells synthesize.
secrete. and deposit forms o f Fn bearing the ED1 segment.
The extent to which vascular endothelial cells might similarly produce this form o f Fn in vivo. thereby potentially
contributing to the circulating pool of ED1 + Fn, will require
funher experimentation. Although the significance of inclu-
sion of ED segments in circulating Fn remains uncertain.
these sites could potentially confer specific functions. panicularly in view o f recent observations indicating that the
alternatively spliced type 111 connecting segment o f Fn
contains cell type-spccific adhesion
ACKNOWLEDGMENT
We thank Dt Charles G. Cochram for his continuing guidtncc
and for providing rrsouroes toward production of IIW~OCIOIUII rntibody 52/54: Dr Tanya N. Mayadas for her generous advice: Sarah
Lawrence. Catherine Case, and Margarita Nodelman for their
expert technical assistance: and Monica Rar~lcttand Lynn LaCivita
for assistance in preparationof the manu.script.
REFERENCES
1. H y n a R O Fibromctins.Sci Am 25442.1986
2. Ruorlahti E Fibroncctin and its receptors. Ann Rev Biochem
57:375. 1988
3. Hynes R O Molecular biology of fibrano*in. Ann Rev Cell
B
i
o
l 1:67. 1985
4. Kornblihtt AR. Umexrwa K. VibaPedmen K. Baralle FE:
Primary structure of human fibromctin: DilTerential splicing may
gemrate at least IO polypeptides from a single gem. EMBO J
4:1755.1985
5. Gutman A. Kornblihtt AR: Identificationof a third region of
all-specific alternativesplicing in human fibromctin mRNA. Roc
Nail Acad Sci USA 8471 79.1987
6. Schwanbauer J& Patcl RS. Fonda D. Hyms R O Multiple
sites of altemative splicing of the rat fibromctin gem transcript.
EMBO J 6:2573.1987
7. Tamkun JW. Hyms R O Plasma fibromctin is synthesized and
secreted by hepatocytes. J Biol Chem 258:4641.1985
8. Paul JI.Schwanhuer JE. Tamkun JW. Hyms R O Cell-typespecific fibromclin subunits gencmted by alternativesplicing. J Biol
Chem 261:I2258. 1986
9. Vartio T. Laitinen L. Narvanen 0. Cut& M. fhormll L
Zardi L. Vinanen I: DitTmntial expression of the ED q u e m
amtaining form of cellular fibrwKcrin in embryonic and adult
human tissues. J CellSci 88:419.1987
IO. Peters JH. Ginsberg MH. Case CM. Cochram CH: Releare
of soluble fibrwcrin containing an extra type 111 domain (EDI)
during acute pulmonary injury mediated by oxidantsor leukocytain
vivo. Am Rev R a p Dis 138: 167.1988
I I . Peters JH. Maunder R. Woolf A. Ginsberg MH: Inrmsed
plasma levels of ED1 + (“‘cellular”) fibromctin in patienu with
vascular injury. J LnbClin Med I 13:586.1989
12. Choi MG. Hyms RO: Biosynthesis and processing of fibroneclinin NIL.8 hamster cells. J Biol Chem 254:12050.1979
13. Olden K. Pmrt RM. Yamada KM: Role of carbohydratesin
protein recmion and turnover: Effects of tunicamycin on the ma*
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
1808
cell surface glycoprotein of chick embryo fibroblasts. Cell 13:461,
1978
14. Paul JI, Hynes RO: Multiple fibronectin subunits and their
post-translation modifications. J Biol Chem 259:13477, 1984
15. Jaffe EA, Mosher DF: Synthesis of fibronectin by cultured
human endothelial cells. J Exp Med 312:122, 1978
16. Birdwell CR, Gospodarowicz D, Nicolson G: Identification,
localization, and role of fibronectin in cultured bovine endothelial
cells. Proc Natl Acad Sci USA 75:3273, 1978
17. Crouch E, Balian G, Holbrook K, Kuksin D, Bornstein P:
Amniotic fluid fibronectin. Characterization and synthesis by cells in
culture. J Cell Biol78:701, 1978
18. Wagner DD, Ivatt R, Destree AT, Hynes RO: Similarities
and differences between fibronectins of normal and transformed
hamster cells. J Biol Chem 256:11708, 1981
19. Bumol TF, Reisfeld RA: Biosynthesis and secretion of fibronectin in human melanoma cells. J Cell Biochem 21:129, 1983
20. Wagner DD, Olmstead JB, Marder VJ: Immunolocalization
of von Willebrand protein in Weibel-Palade bodies of human
endothelial cells. J Cell Biol95:355, 1982
21. Engvall E, Ruoslahti E Binding of soluble form of fibroblast
surface protein, fibronectin, to collagen. Int J Cancer 20:1, 1977
22. Laemmli UK: Cleavage of structural proteins during the
assembly of the head of bacteriofage T,. Nature 227:680,1970
PETERS ET AL
23. Struck DK, Lennarz WJ: Evidence for the participation of
saccharide-lipids in the synthesis of the oligosaccharide chain of
ovalbumin. J Biol Chem 252:1007, 1977
24. Robbins PW, Hubbard SC, Turco SJ, Wirth D F Proposal for
a common oligosaccharide intermediate in the synthesis of membrane glycoproteins. Cell 12393, 1977
25. Tarentino AL, Trimble RB, Maley F Endo-beta-N-acetylglucosaminidase from Streptomyces plicatus. Methods Enzymol 50:
574,1978
26. Schwarzbauer JE, Paul JI, Hynes RO: On the origin of
species of fibronectin. Proc Natl Acad Sci USA 8 1:1424,1985
27. Humphries MJ, Akiyama SK, Komoriya A, Olden K, Yamada KM: Identification of an alternatively spliced site in human
plasma fibronectin that mediates cell type-specific adhesion. J Cell
Biol 103:2637, 1986
28. Humphries MJ, Komoriya A, Akiyama SK, Olden K, Yamada KM: Identification of two distinct regions of the type. 111
connecting segment of human plasma fibronectin that promote cell
type-specific adhesion. J Biol Chem 262:6886, 1987
29. Humphries MJ, Akiyama SK, Komoriya A, Olden K, Yamada KM: Neurite extension of chicken peripheral nervous system
neurons on fibronectin: Relative importance of specific adhesion sites
in the central cell-binding domain and the alternatively spliced type
I11 connecting segment. J Cell Biol 106:1289, 1988
From www.bloodjournal.org by guest on January 12, 2015. For personal use only.
1990 75: 1801-1808
Human endothelial cells synthesize, process, and secrete fibronectin
molecules bearing an alternatively spliced type III homology (ED1)
JH Peters, LA Sporn, MH Ginsberg and DD Wagner
Updated information and services can be found at:
http://www.bloodjournal.org/content/75/9/1801.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.