You are on page 1of 13

Molecular modeling of protein-glycosaminoglycan interactions.

A D Cardin and H J Weintraub

Arterioscler Thromb Vasc Biol. 1989;9:21-32


doi: 10.1161/01.ATV.9.1.21
Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272 Greenville
Avenue, Dallas, TX 75231
Copyright © 1989 American Heart Association, Inc. All rights reserved.
Print ISSN: 1079-5642. Online ISSN: 1524-4636

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://atvb.ahajournals.org/content/9/1/21

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the Copyright
Clearance Center, not the Editorial Office. Once the online version of the published article for which permission
is being requested is located, click Request Permissions in the middle column of the Web page under Services.
Further information about this process is available in the Permissions and Rights Question and Answerdocument.

Reprints: Information about reprints can be found online at:


http://www.lww.com/reprints

Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online
at:
http://atvb.ahajournals.org//subscriptions/

Downloaded from http://atvb.ahajournals.org/ at BOSTON UNIV MED LIBR on October 30, 2013
Molecular Modeling of
Protein-Glycosaminoglycan Interactions
Alan D. Cardin and H.J.R. Weintraub

Forty-nine regions In 21 proteins were Identified as potential heparln-blnding sites


based on the sequence organizations of their basic and nonbaslc residues. Twelve
known heparln-blnding sequences In vttronectln, apollpoprotelns E and B-100, and
platelet factor 4 were used to formulate two search strings for Identifying potential
heparln-blnding regions In other proteins. Consensus sequences for glycosamlno-
glycan recognition were determined as [-X-B-B-X-B-X-] and [-X-B-B-B-X-X-B-X-]
where B Is the probability of a basic residue and X Is a hydropathic residue.
Predictions were then made as to the heparln-blnding domains In endothellal cell
growth factor, purpurln, and antithrombln-lll. Many of the natural sequences con-
forming to these consensus motifs show prominent amphlpathlc periodicities having
both a-hellcal and 0-strand conformations as determined by predictive algorithms
and circular dlchroism studies. The heparln-blnding domain of vttronectln was
modeled and formed a hydrophlllc pocket that wrapped around and folded over a
heparln octasaccharide, yielding a complementary structure. We suggest that these
consensus sequence elements form potential nucleation sites for the recognition of
potyanions In proteins and may provide a useful guide In identifying heparln-blnding
regions In other proteins. The possible relevance of proteln-glycosamlnoglycans
Interactions In atherosclerosis Is discussed.
(Arteriosclerosis 9:21-32, January/February 1989)

T here is considerable interest in elucidating the spe-


cific mechanism(s) by which heparin and other gly-
cosaminoglycans (GAG) interact with proteins to regulate
lipoproteins (LDL). The structure of these regions was of
particular interest as we hypothesized that they could
mediate the interaction of LDL with acidic mucopolysac-
metabolic processes in both normal and disease states. charides of arterial tissue. We showed that LDL contains
Examples of these processes include hemostasis,1'2-3 five to seven heparin-binding sites on the lipoprotein
cell-substrata adhesion, 4 smooth muscle cell surface.15 The heparin-binding cyanogen bromide pep-
proliferation, 56 lipolytic enzyme activity,7 lipoprotein- tides of apo B-100 were isolated, and their amino acid
arterial wall interactions,89.10 and regulation of growth sequences were determined.16 These peptides account
factors in angiogenesis.1112 However, the structural het- for five domains of high positive charge density that we
erogeneity of the acidic mucopolysaccharides with respect suggest mediate the lipoprotein's interaction with glycos-
to size, carbohydrate composition and charge,13 and the aminoglycans. These same regions were also identified
variety of proteins that are known to bind to these mole- by Weisgraber and Rail.17 We noticed16 that these domains
cules have complicated a detailed molecular analysis of in apo B-100 have a sequence similarity to the heparin-
protein-GAG interactions. The most studied system, bind- binding regions of apo E 1819 and human vitronectin (Vn)20
ing of heparin to a specific domain or site on human with respect to the organization of basic and hydropathic
antithrombin-lll (AT-III), involves a unique carbohydrate residues. The present study extends our initial observa-
structure representing only a fraction of the total heparin.14 tions to other proteins whose functions are modified by
The exact amino acid residues on AT-III or other proteins heparin. We show that all these proteins contain unique
which comprise their heparin-binding domains have not consensus sequences of amino acid residues that are
been fully elucidated.
potentially involved in heparin binding.
To understand the molecular recognition of heparin by
proteins, we recently determined the structure of the
heparin-binding regions of apolipoprotein (apo) B-100, the Methods
major protein constituent of human plasma low density Secondary structure predictions of Vn peptide conforma-
tions were performed by the MSEQ program21 based on
From the Merrell Dow Research Institute, 2110 E. GaJbraith Chou-Fasman algorithms22 and by the method of Gamier
Road, Cincinnati, Ohio.
This paper was presented at the 1987 American Heart Asso- et al. 23 Both methods gave comparable predictions of
ciation Meeting, Anaheim, California for the Irvine H. Page predominant /3-strand and 0-tum structures, and circular
Award. dichroism studies on consensus peptides in the presence
Address for reprints: Alan D. Cardin, Ph.D., Merrell Dow of heparin were consistent with these predictions. From
Research Institute, 2110 E. Galbraith Road, Cincinnati, OH
45215. these data, a starting structure was devised for the Vn
Received June 10, 1988; revision accepted September 14, peptide in the three-dimensional modeling studies on hep-
1988. arin binding. Heparin was modeled based on the consJd-

21
22 ARTERIOSCLEROSIS V O L 9, No 1, JANUARY/FEBRUARY 1989

orations of Cowan et al. 24 The heparin disaccharide these regions have not been shown experimentally to bind
repeat of 1,4-linked 2-sulfate-a-L-idopyranosyluronic acid heparin. The selection of these sequences was based on
and 2-deoxy-2-sulfamino-a-D-glucose-6-sulfate was con- their similarity to the experimentally defined heparin-
structed from cyclohexane by the addition of the appro- binding domains in Vn, apo E, apo B-100, and PF-4, and,
priate hydrophilic side chains and replacement of one of thus, represent putative heparin-binding regions. The
the cyclic methylene groups with an oxygen. regions shown are characterized by two consensus
Molecular modeling of Vn-heparin interactions was sequences [-X-B-B-X-B-X-] (Table 1) and [-X-B-
performed with an Evans and Sutherland PS-350 Graph- B-B-X-X-B-X-] (Table 2), where B is the probability of
ics System. Molecular structure data on both Vn and occurrence of a basic residue and X is a hydropathic
heparin were analyzed using the Insight and Discover residue. The consensus sequences were determined by
programs25-26 from Biosym Technologies, Incorporated. the matrix analyses shown in Tables 3 and 4. The
Individual structures were built with Insight software and matrices show the frequency of occurrence of any given
geometrically optimized with the force field in Discover.27-33 residue at each position within the [-X-B-B-X-B-X-], i.e.,
The potential energies of the Vn peptide and heparin - 3 to +3 positions, and [-X-B-B-B-X-X-B-X-], i.e., - 4 to
structures were separately minimized. Heparin and pep- +4 positions. With respect to the [-X-B-B-X-B-X-] consen-
tide were then visually docked by using Insight on an sus sequence (see Table 3), Lys is the most common
Evans and Sutherland graphics workstation. The resultant residue at positions - 2 and - 1 , whereas Arg predomi-
complex was geometry-optimized using the Discover force nates at +2. His is infrequently used at these positions.
field and was subjected to molecular dynamics simulation The - 2 , - 1 , and +2 positions show exclusive usage by
on a Vax and a Cray supercomputer. Molecular dynamics basic residues, whereas the region of the consensus has
were performed as follows: the complex was "heated" to virtually no acidic residues. The [-X-B-B-B-X-X-B-X-] con-
600° K and was allowed to equilibrate for 500 fs. The sensus sequence (see Table 4) shows Lys as the pre-
atomic trajectories were then sampled every 100 fs there- dominant basic residue at the - 2 , - 1 , and +3 [-B-]
after for 5 to 7.5 ps. The resulting structure (geometry) positions; Arg is the second most utilized, with His having
obtained at each 100 fs sampling point was cooled to 300° infrequent usage. The sequences within the consensus
K and was allowed to equilibrate over an additional 100 fs are relatively abundant in Asn, Ser, Ala, Gly, lie, Leu, and
of molecular dynamics simulation. The resultant complex Tyr and exhibit a very low occurrence of Cys, Glu, Asp,
was then geometry-optimized to yield a final structure. Met, Phe, and Trp. The - 2 and +3 [-B-] positions contain
Amphipathic periodicities of peptides were analyzed only basic residues, whereas the - 3 and - 1 [-B-] posi-
by the methods of Eisenberg et al. 34 and Schiffer and tions have the greatest variation, with 16 and 17 out of 28
Edmunson. 35 residues being basic, i.e., —60% usage at positions - 3
and - 1 versus 100% at positions - 2 and +3. Although
the - 3 and - 1 positions show a lower frequency of basic
Results residues, either one or both of these positions are used.
Consensus Sequence Analyses of Thus, the consensus represents either a stretch of di- or
Heparln-blndlng Regions tribasic residues separated by two or three hydropathic
The heparin-binding domains of human apo B-100,1617 residues terminated by one or more basic residues. The
apo E, 1819 Vn, 20 and platelet factor 4 (PF-4)24 consist of alignments shown maximize the sequence similarities
extended sequences of basic residues. The three- among the basic regions of the various proteins with the
dimensional molecular interaction of PF-4 with heparin known heparin-binding regions of the reference proteins.
has been modeled based on the X-ray crystallographic Studies with synthetic peptides corresponding to spe-
structure of the tetrameric form. 24 A similar model for the cific consensus sequences show that these structures can
binding of heparin to /3-thromboglobulin based on the function in polyanion recognition.1857 For example, a
structural homology of this protein with PF-4 was also synthetic peptide containing the consensus [-X-B-
proposed.24 The heparin-binding regions of these proteins B-B-X-X-B-X-] that includes residues 3364 to 3371 of apo
are characterized by clusters of arginines and lysines. B-100 (T-R-K-R-G-L-K-L) binds heparin at the same ionic
These clusters form centers of high positive charge den- strength conditions as LDL57 and corresponds to a seg-
sity that electrostatically interact with the acidic groups of ment of a heparin-binding CNBr fragment isolated from
mucopolysaccharides, such as heparin. A feature of these apo B-100.16 Similarly, a synthetic peptide of apo E,
natural sequences is the interspersion of basic and hydro- comprising the LDL receptor binding domain58 and having
pathic residues. We have investigated whether the simi- essentially the same sequence within the consensus
larity in organization of basic and hydropathic residues in region as the apo B-100 synthetic peptide (see Table 2),
the heparin-binding domains of these proteins could be accounts for the major heparin-binding activity of apo
used to identify the basic regions in other heparin-binding E.1819 Using synthetic peptides with successive deletions
proteins that function in polyanion recognition. from the amino terminus, we showed that residues 144 to
As indicated above, sequence similarities among the 147 (L-R-K-R) of apo E were essential for heparin-binding
heparin-binding domains of Vn, apo E, apo B-100, and activity.18 These residues represent the [-X-B-B-B-] ele-
PF-4 with respect to the organization of basic and hydro- ments of the consensus sequence [-X-B-B-B-X-X-B-X-]
pathic residues were noted. These sequences are shown shown in Table 2 and likely account for a significant
in Tables 1 and 2. The basic regions of other known amount of the heparin-binding activity in a major throm-
heparin-binding proteins are also presented; however, bolytic fragment from apo E.18 Monospecific antibodies to
HEPARIN-BINDING PROTEINS Cardin and Weintraub 23

Table 1. Sequence Similarities of Heparin-binding Proteins


Consensus* [X-B-B-X-B-X] Reference
Vnt ^ R P S L A K K QR F R H R N RK G Y RSQ 20
13
Apo B D A T R F KH LRK Y T YN YE A ES S S 17
ApoB "G K A L L K K T KN S E E F AA A MS RY 17
2077
ApoB l D Q F V RK YRA A L GK L PQ QA ND 16, 17
2116
ApoB L TALTKK Y R I T END I Q I A L DD 16, 17
ApoB ""VKAQYKK N K HR H S I T N P L A V L 16, 17
36W
Apo B TTS I GRR QH L R V S T A F V Y T K N 16, 17
119
Fn I G DTWR R P H E T G GY ML E CVC L 36
17ffi
Fn N V S P P RR AR V T D A T E T T I T I S 36
Fn ^ A T P I R H RP R P Y P P N VG E E I Q I 36
Fn ^ K D Q Q R H KVR E E V V T VG N S V N E 36
TS " S L R Q M K K T RG T L LA L E R K D H S 37,38
CT
LpL LCLSCRK N RC N N LG YE I N K V R 39
M
HTGL L C L S C K K GRC N T LG YH V R Q E P 40,41
ECGF/aFGF VLGNYKK P KL L Y C S NG G Y F L R 42,43
M
THR L V R I G K HS R T R Y ERN I E K I SM 44
129
NCAM T I I WK H K G R D V I L K KD V R F I V 45
ApoE ^ A W G E R L R ARM E E MG SR T R D R L 18
128
GDN K A I VSKK N KD I V T V AN A V F V K 46
Table 1 represents the sequence similarity of basic regions in various heparin-binding proteins. The sequences
are compared with the heparin-binding domains of vitronectin, apolipoprotein B-100, apolipoprotein E, and
platelet factor 4.
The consensus sequence [X-B-B-X-B-X] was derived from the usage frequency of basic (B) and nonbasic (X)
residues in the data base (see Table 3).
tOne-letter amino acid code: A=Ala, C=Cys, D=Asp, E=Glu, F=Phe, G=Gly, H=His, l=lle, K=Lys, M=Met,
N=Asn, P=Pro, Q=Gln, R=Arg, S=Ser, T=Thr, V=Val, W=Trp, and Y=Tyr.
Fn=fibronectin, TS=thrombospondin, LpL=lipoprotein lipase, HTGL=hepatic trigryceride lipase, ECGF=endotheliaJ
cell growth factor, aFGF=acidic fibroblast growth factor, THR=thrombin, NCAM=neural cell adhesion molecule, and
GDN=gliaJ-dertved nexin.

this region diminish heparin-binding to apo E, indicating sequence may define a major region of heparin contact.
that this region is utilized in the intact protein.19 These Peterson et al. 59 have shown that pyridoxylation of a
findings with synthetic peptides illustrate examples in single critical lysine, lysine-125, blocks heparin-binding to
which the consensus structures bind heparin and, together antithrombin-lll (AT-III); substantial pyridoxylation also
with the data in Tables 1 and 2 suggest that these occun-ed at lysines-133 and -136, i.e., residues within the
sequence patterns are utilized by numerous proteins in consensus. Moreover, these residues are protected from
polyanion binding. Thus, as shown below, the consensus pyridoxylation by heparin.80 More recently, Smith and
sequences (Tables 1 to 4) may prove useful in identifying Knauer61 isolated a heparin-binding fragment from AT-III
polyanion recognition sites in other proteins. containing the 130L-Y-R-K-A-N-K-S sequence. These data
support, experimentally, a role for the 130L-Y-R-K-A-N-K-S
Predictions of Heparln-Blndlng Based on
Sequence Inspection
Purpurin is a heparin-binding protein that functions in
neuroretinal cell survival and adhesion.49 Its heparin-
binding region has not been determined. However, inspec-
tion of its primary sequence (Table 5) reveals a single AT-III
domain of high positive charge density that fits the con-
sensus sequence [-X-B-B-B-X-X-B-X-] given in Table 2.
The homologous regions in other members of its gene
123
family are also shown; these proteins exhibit reduced or -FAKLNCRLYRKAMKSSKL
no heparin-binding activity. The difference in their heparin 128
reactivities may result in part from the nonconservatjon of 117
positive charges at the - 3 and +3 positions in the -LKKINKAIVSKKNKDIVT h-GDN
consensus; i.e., positions that show 60% and 100% usage 183
-FRKLTHRLFRRNFGYTLR HC-II
by basic residues as determined from the data base of
Figure 1. Proposed heparin recognition regions in antJthrombin
natural sequences summarized in Table 4.
III (AT-III), glial-derived nexin (GDN), and heparin cofactor II
We also examined the sequence of AT-III, which con- (HC-II) based on the consensus sequence analyses of Tables 1
tains the consensus 130L-Y-R-K-A-N-K-S (Figure 1). This to 4.
24 ARTERIOSCLEROSIS VOL 9, No 1, JANUARY/FEBRUARY 1989

Table 2. Sequence Similarities among Basic Regions in Heparln-blndlng Proteins*


Consensus* [X-B-B-B-X-X-B-X] References
Vn ^RPSLAKKQRFRHRNRKGYRSQRGHSRGRNQNSRR 20
Apo B " A L K A G K L K F I I PSPKRPVKLLSGGNTLHLVSTTK 17
33S2
ApoB YK L E G T T R L T R L T R K R G L K L A T A L S L S N K F V E G S 16, 17
1 3 2
ApoE ELRVRLASHLRKLRKRLLRDADDLQKRLAVYQAG 18, 19
y-IP-10 "SQFCPRVE I I ATMKKKGEKRCLNPESKA I KNLLK 47
1 5 0
HTGL A H V S G F A G S S M G G K R K I GR I T G L D P A G P M F E G T S 40,41
ECGF/aFGF "ERLEENHYNTYISKKHAEKHWFVGLKKNGRSKLG 42,43
HRG "'CCHGHGPPPGHLRRRGPGKGPRPFHCRQ I GSVYR 48
1 M
HRG PVLIDFFEDTERYRKQANKALEKYKEENDDFASF 48
a
Pn A V D S F S V K D N F D P K R Y A G K W Y A L A K K D P E G L F LQ 49
/3-TG "GRR I C L D P D A P R I KK I V Q K K L A G D E S A D 50
10
TS NSVFD I F E L T G A A R K G S G R R L V K G P D P S S P A F R I 37
HTGL ^ N C K K G R C N S L G Y D I RR I G H A K S K T L F L I T R A Q S P 40,41
PF-4 " N G R R I C L D L Q A P L Y K K I IKKLLES 47
133
LpL LGAHAAG I A G S L T N K K V N R I T G L D P A G P N F E Y A E 39
LpL ^RCNNLGYEINKVRAKRSSKMYLKTRSQMPYKVFH 39
ECGF/aFGF "I S K K H A E K H W F V G L K K N G R S K L G P R T H F G Q K A I L 42,43
1ia
AT-III Q I H F F F A K L N C R L Y R K A N K S S K L V S A N R L F G D K S 51
HC-II ™ A S S K Y E I T T I H N L F R K L T H R L F R R N F G Y T L R S V N 52,53
Fn ^ D T R T S Y R I GDTWSKKDNRGWLLQC I CTGNGRGE 36
10a4
Fn AQ I TGYRLTVGLTRRGQPRQYNVGPSVSKYPLRN 36
bFGF "GSGAFPPGHFKDPKRLYCKNGGFFLRIHPDGRVD 54
NCAM •°°VVN I K Q D D G G S P I R H Y L I K Y K A K H S S E W K P E I R L 45
a i
LPC FNT I L T T R S Y P Q L R R V F Q K Y T K Y S K H D M N K V L D L 55
LPC " " E K L H Q A M K G V G T R H K A L I R I MVSRSE I DMND I KA 55
THR ^ D G K Y G F Y T H V F R L K K W I Q K V I D Q F G E 44
1 0 4
GDN MVMRYGVNGVGK I L K K L N K A I VSKKNKD I V T V A N 46
PCI ^KMQQVENGLSEKTLRKWLKMFKKRQLELYLPKFS 56
Table 2 represents the sequence similarity of basic regions in various heparin-bindlng proteins. The sequences are compared with the
heparin-binding domains of vttronectin, apollpoprotein EM 00, apollpoproteln E, and platelet factor 4.
T h e consensus sequence [X-B-B-B-X-X-B-X] was derived from the usage frequency of basic (B) and nonbasic (X) residues In the data
base (see Table 4).
Vn=vitronectin, ^IP-10=10 kilodalton gamma-interferon Inducible protein, HRG=histkJine-rich glycoprotein, Pn=purpurln, p-
TG=beta-thromboglobulin, AT-III=arrtithrombln III, HC-ll=heparin cofactor II, bFGF=basic fibroblast growth factor, LPC=lipocortin,
PCI=protein C inhibitor. For other abbreviations, see footnote to Table 1.

sequence in heparin-binding. The corresponding homolo- mitogenesis in Balb/C 3T3 fibroblasts. Lys-118 was the
gous regions in glial-derived nexin (GDN) and heparin primary site of dimethylation, and Lys-118 and -112 were
cofactor-ll (HC-II) are also shown (Figure 1). Like AT-III, the major sites of monomethylation. These residues reside
these proteins utilize heparin as cofactor. GDN has two within the predicted heparin-binding sequence 110G-
consensus structures, [-X-B-B-B-X-X-B-X-] and [-X- L-K-K-N-G-R-S-K (Table 2).
B-B-X-B-X], given by residues 117 to 123 (L-K-K-l-N-K-A)
and 126 to 131 (S-K-K-N-K-D), respectively. The 127K- Molecular Modeling of Heparin Binding
K-N-K structure of GDN is identical to the four-residue Limited information is available on the solution confor-
stretch, 3150K-K-N-K, of the experimentally determined mations of heparin-binding domains. Recently we showed
heparin-binding region of apo B-100 (see Table 1). HC-II that the heparin-binding regions of apo E and apo B-100
contains the [-X-B-B-B-X-X-B-X-] consensus given by defined by residues 129 to 169, 202 to 243, and 3345 to
residues 183 to 189 (-F-R-K-L-T-H-R-). Other members of 3381, respectively, increase their helical and ^-strand
the SERPIN gene family, such as arantitrypsin, a,-anfj- contents by >50% upon binding heparin. 5783 These
chymotrypsin, and C-l inhibitor, which do not utilize hep- regions fit the consensus [-X-B-B-B-X-X-B-X-]. The domain
arin as cofactor, do not contain these consensus sequence given by apo E ^ to 169 is known to form amphipathic
elements.62 a-helix upon lipid binding,64 and both apo E129 B iee and
Harper and Lobb 42 recently reported that reductive apo 63345 to 338, show significant amphipathic a-helical
methylation of less than three lysines in acidic fibroblast periodicities as determined by Eisenberg analysis. 345763
growth factor (FGF) significantly reduced its heparin affin- These findings suggest that heparin induces amphipathic
ity, growth factor receptor affinity, and ability to stimulate a-helix in the apo E ^ to ,«, and apo 83359 to 337a peptide
HEPARIN-BINDING PROTEINS Cardin and Weintraub 25

ApoE ApoB-100
135-152 3359-3376

Purpurin Retinol Binding Protein


32-49 28-45

K K

Antithrombin—I ECGF/aFGF
125-142 101-118

K K
RK R K
Figure 2. Helical wheel diagrams of the consensus sequence regions of various proteins.
26 ARTERIOSCLEROSIS V O L 9, No 1, JANUARY/FEBRUARY 1989

Table 3. Distribution of Amino Acids In Heparin- study includes the consensus sequence 130L-Y-R-
bindlng Consensus [X-B-B-X-B-X-] K-A-N-K-S (see Table 2 and Figure 1), and it has amphip-
Consensus sequence athic periodicity by the Eisenberg analysis.34 Based on
helical wheel diagrams,35 acidic FGF shows potential to
[X- B- B- X- B- X] Sum
Amino acid -3 -2 -1 +1 +3 - 3 to +3 form a-helix. Chemical modification of lysine-112 and
+1
lysine-118 by reductive methylation42 might possibly desta-
E 0 0 0 0 0 2 2
bilize the charge interaction with heparin and hence the
D 0 0 0 0 0 2 2 helical conformation.
K 1 10 12 0 4 1 28 Heparin induces amphiphilic /3-strand structure in
R 3 5 5 0 13 0 26 apo E u 2 u . 6 3 The sequences involved are213R-
H 0 3 2 0 2 1 8 L-R-A-R-M and ^R-T-R-P-R-L, which consist of alter-
N 0 0 0 3 0 1 4 nating basic and nonbasic residues of the consensus
Q 0 0 0 2 0 0 2 [-X-B-B-X-B-X-]. Beta-structure is also induced by hep-
C 2 0 0 0 0 2 4 arin in model peptide copolymers having alternating
P 1 0 0 3 0 1 5 lysine and tyrosine residues.66 Thus, heparin may
increase the )3-strand and a-helical character depending
S 1 0 0 1 0 0 2
on the precise organization of basic and nonbasic
Y 2 0 0 2 0 0 4
residues in natural peptide sequences.
A 1 0 0 2 0 1 4
Polyanions may also bind to sites on proteins without
G 2 0 0 2 0 1 5
inducing significant changes in protein secondary struc-
I 0 0 0 0 0 1 1 ture. Proteins may contain exosites contoured to accom-
L 1 1 0 1 0 2 5 modate heparins of specific size and charge. Figure 3
M 1 0 0 0 0 1 2 illustrates one potential representation of the binding of a
F 1 0 0 0 0 1 2 heparin octasaccharide to the heparin-binding region of
T 1 0 0 2 0 1 4 Vn. The heparin and Vn structures were individually
W 1 0 0 0 0 0 1 minimized by the Discover force field26 to locate a stable
V 1 0 0 1 0 1 3 conformation of each molecule. The two molecules were
manually docked, and molecular dynamics simulation
followed by energy minimization of the docked structures
Acidic
0 0 0 0 4 4
refined the fit without significant changes to the original
residues 0
unminimized docked structure of heparin or Vn. As shown
in Figures 3A to 3D, the Vn domain forms a folded
Basic charged pocket into which the heparin octasaccharide fits.
residues 4 18 19 0 19 2 62
The cleft, formed by the twisting of a series of /3-bends
Aromatic
0 0 2 1 7
from the carboxyl-terminal end, is capped by the amino-
residues 4 0
terminal helical portion of the Vn structure. The [-
The residue distributions were compiled from the natural amino X-B-B-X-B-X-] and [-X-B-B-B-X-X-B-X-] elements of the
add sequences of heparin-binding proteins of Table 1.
domain form distinct areas of contact with heparin that
wrap around and fold over the octasaccharide, yielding a
regions. As is shown in Figure 2, helical wheel 35 diagrams structure complementary in charge and molecular con-
of these peptide regions segregate the basic residues tour. Thus, heparin molecules not having the correct
primarily to one side of the helical face, forming a region of complementary structures may not bind. In addition to the
high positive charge density. The [-X-B-B-B-X-X-B-X-]
motif ensures the segregation of the basic residues in the
consensus to one side of the helical face. In addition,
Figure 3. A molecular model of the interaction of a heparin
basic residues flanking the region of the consensus may octasaccharide with the heparin-binding region of vitronectin
also contribute to the charge density. For proteins assum- (Vn). The structure of Vn represents residues 342 to 369 con-
ing a-helical changes upon heparin-binding, the taining the ammo-terminal [-X-B-B-X-B-X-] and carboxyl terminal
[-X-B-B-B-X-X-B-X-] may serve as the primary nucleation [-X-B-B-B-X-X-B-X-] consensus sequence elements. A. This
site for heparin interaction. Subsequent to the structural view represents a look down the z-axis of the two interacting
molecules. As shown, Vn (white) forms a structure that wraps
change, the polyanion may "lock" or stabilize the peptide around the heparin molecule (red). This space-filling view shows
in this configuration. Figure 2 also shows helical wheel the complementarity of surface densities for the two molecules.
projections for purpurin, retinol binding protein, AT-III, and B. A side view of the two interacting molecules. The carbohydrate
FGF. However, the conformations that these regions ring of heparin is shown in blue and the sulfurs and oxygens of
the hydrophilic side chains in yellow and red, respectively. The
assume upon heparin-binding are not known. One possi- a-carbon backbone of Vn is shown in white and the nitrogens of
ble explanation for the reduced heparin affinity in retinol its side chains in blue. C. The same view as in B but with the
binding protein relative to purpurin is a significant decrease added surface densities. The surface densities for Vn are shown
in positive charge density within the region of the consen- by white dots and those for heparin by red dots. Note the
sus. Based on the X-ray crystal structure of a,-antitrypsin, complementarity of the surface densities. D. A180° rotation of the
view shown in C. Note the finger-like projections of the Vn side
Carrell et al. 65 modeled the heparin-binding region of chains that wrap around to cradle the heparin molecule. The color
AT-III as a-helical. The AT-III insertion region used in the codes for the various atoms in views B to D are the same as in A.
HEPARIN-BINDING PROTEINS Cardin and Weintraub 27
28 ARTERIOSCLEROSIS VOL. 9, No 1, JANUARY/FEBRUARY 1989

Table 4. Distribution of Amlno Acid Residues in Heparin-binding Consensus [X-B-B-B-X-X-B-X]


Consensus sequence
[X- B- B- B- X- X- B- -X] Sum
Amino acid -4 -3 -2 -1 +1 +2 +3 +4 - 4 to+4

16 11 17 53
11 40
H 0 1 1 1 0 0 2 1 6
N 0 2 0 0 1 5 0 1 9
Q 0 0 0 1 1 3 0 1 6
C 0 0 0 0 0 1 0 0 1
P 2 1 0 0 2 1 0 0 6
S 2 1 0 0 2 1 0 3 9
Y 1 2 0 2 1 1 0 2 9
A 1 1 0 1 4 0 0 3 10
G 2 0 0 3 3 6 0 2 16
I 3 1 0 1 4 3 0 3 15
L 5 3 0 1 5 3 0 2 19
M 1 0 0 0 0 0 0 3 4
Li.

0 1 0 0 1 0 0 0 2
T 5 0 0 0 0 1 0 0 6
W 1 0 0 1 1 0 0 0 3
V 0 0 0 1 2 1 0 1 5
Acidic
residues* 1 0 0 0 1 2 0 1 5
Basic
residues* 4 16 28 17 0 0 28 6 104
Aromatic
residues* 2 3 0 3 3 1 0 2 13
The residue distributions were compiled from the natural amino acid sequences of heparin-binding proteins of
Table 2.
T h e percent composition of basic, acidic, and aromatic residues at each position within the consensus is given by:
[X-B-B—B-X-X-B---X]
%Basic 14 57 100 61 0 0100 21
%Acidic 3 0 0 0 3 7 0 3
%Aromatic 6 11 0 1 1 1 1 3 0 6

Table 5. Proposed Heparin Recognition Region of Purpurln: Comparison to Homologous


Regions In Other Sequence-related Proteins
Proteins [X-B-B-B-X-X-B-X] Reference
Purpurin "PKRYAGKWYALAKK 49
S
Retinol-binding protein KARFSGTWYAMAKK 49
ia
a,-Microglotxjlin l SR I Y G K W Y N L A I G 85
12
/5-Lactogtobulin I Q K V A G T W Y S LAMA 85

consensus regions, the folding is such that basic residues (LpL) and endothelial cell growth factor (ECGF)/acidic
outside the consensus are brought into contact with the FGF, for example, the consensus sequence elements are
hydrophilic side chains of heparin. The presence of these not close together within the primary sequence. 39 ' 4243
domains in Vn and possibly other proteins may facilitate However, two or more of these regions may come together
the interaction of neighboring amino acid residues with the by folding of the protein structure to form a topological
heparin molecule. In this way, different heparin structures domain or crevice into which the heparin molecule binds.
may bind different heparin-binding proteins. In addition to In contrast, other proteins may not exhibit a particular
electrostatic binding, the model shown in Figure 3 implies specificity for heparins differing in size, charge, and car-
attractive forces due to van der Waals interactions and bohydrate composition, depending on the organization of
formation of intermolecular hydrogen bonds. Table 6 lists their polyanion recognition domains.67 Such interactions
some of the quantitative parameters of the model. should not be de-emphasized in importance as they may
The Vn sequence contains both heparin-binding con- serve a biological function. These studies point out the
sensus sequence elements in series. In lipoprotein lipase unique sequence commonalities and conformational orga-
HEPARIN-BINDING PROTEINS Cardin and Weintraub 29

Table 6. Minimized Interatomic Distances between Potential Interacting Side Chains of Vitronectin and Heparin
Consensus Residue Amino acid Interatomic Heparin side
sequence number residue distance (A) chain
342 R 1.99 O-sulfate
X 346 A —
B 347 K —
B 348 K 2.26 carboxylate
X 349 Q —
B 350 R —
X 351 F —
352 R 4.12 carboxyiate
353 H 6.46 O-sulfate
X 354 R 10.05 O-sulfate
B 355 N —
B 356 R 7.19 O-sulfate
B 357 K 2.01 O-sulfate
X 358 G —
X 359 Y —
B 360 R 2.93 O-sulfate
X 361 S —
363 R 9.41 O-sulfate
365 H 3.18 O-sulfate
369 R 3.84 O-sulfate
These interatomic distances are representative of one of the sampled geometries from the molecular dynamics trajectory. The
interactions listed yield minimal energy contributions.

nizations specifying polyanion recognition among proteins possibly residues 3352 to 3378, which resemble the
of diverse biological function. receptor-binding sequence of apo E. Alternatively, hep-
arin may bind to other sites on the LDL surface that inhibit
Discussion receptor binding. Presently, it is not known what effect
increased circulating levels of endogenous GAG 72 have
Significance to the Pathogenesls of
on LDL catabolism.
Atherosclerosis
We previously showed that proteolysis of the LDL
The purpose of this study was to obtain structural
surface with trypsin significantly diminished receptor-
insight into the interactions of GAG with proteins, as
binding without appreciably affecting heparin binding.73 It
these molecules are involved in numerous cardiovascu-
is possible that one or more mutations affecting the
lar processes, including lipoprotein-proteoglycan inter-
receptor-binding region of apo B-100 that inhibits LDL
actions in the arterial wall, 7 - 10 cell substratum adhesion
receptor binding may not affect the ability of the particle to
and spreading,4 regulation of smooth muscle cell prolif-
bind arterial proteoglycan. Such a molecule would be
eration, 56 hemostatic processes, 12 ' 3 and neovascular-
ization.11'12 Our studies have focused primarily on the particularly atherogenic since it would not be removed
binding of LDL with heparin as a model of lipoprotein- from the circulation by the liver via the LDL receptor
arterial wall GAG interactions. pathway. Such lipoproteins may complex with GAG and
A limited number of regions in apo E and apo B-100 be taken up by macrophages to form foam cells. During
bind heparin. 15 - 19 The LDL receptor-binding region in progression of atherosclerosis, the changes in the GAG
apo E56 corresponds to a heparin-binding region18'18 and composition of the arterial wall that are known to occur
is structurally similar to the heparin-binding region defined may further potentiate the interaction with LDL particles.74
by residues 3352 to 3378 of apo B-100.1«'17 Studies with Polymorphisms in the protein and DNA structure of
monoclonal antibodies to apo B-100 that block LDL bind- apo B-100 have been detected, 757e and in two recent
ing to the receptor have been interpreted as a single studies structural mutations in LDL apo B-100 have been
domain for receptor recognition.68'69'70 In contrast, studies implicated in primary moderate hyperlipoproteinemia and
on the interaction of heparin with LDL show that there are atherosclerosis.77'78 These observations may be explained
five to seven heparin contact sites on the LDL surface,15 by mutations that alter receptor binding but not the lipo-
and approximately this number of heparin-binding pep- protein's interaction with arterial wall glycosaminoglycans.
tides have been purified from apo B-IOO.1617 It is known Srinivasan et al. 79 recently reported that heparin inhibits
that heparin releases LDL from cell receptors.71 Heparin LDL accumulation in rabbit aorta, suggesting the possibil-
may inhibit the binding of LDL to the receptor by binding to ity that heparin occupancy of the GAG-binding domains
the region of apo B-100 involved in receptor-recognition, on the LDL surface may block LDL reactivity with arterial
30 ARTERIOSCLEROSIS VOL 9, No 1, JANUARY/FEBRUARY 1989

proteoglycan. Ye et al. 80 reported that the lysine residues glycans and proteoglycans in physiological and pathological
on apo B-100 are involved in an interaction with the processes of body systems. Basel: Karger, 1982:214-230
10. Radhakrishnamurthy B, Srlnlvasan SR, Vljayagopal P,
plasminogen-like domains (Kringle 4) on apo(a). These
Darferes ER Jr, Berenson GS. Mesenchymal injury and
researchers proposed that the association of Lp(a) with proteoglycans of arterial wall in atherosclerosis. In: Varma
lysine-rich regions in apo B-100 may diminish the interac- RS, Varma R, eds. Glycosaminoglycans and proteoglycans
tion of LDL with the LDL receptor, as reported for apo(a),81 in physiological and pathological processes of body systems.
and direct the particles into other catabolic pathways. Basel: Karger, 1982:231-251
11. Maclag T. Angiogenesis. In: Spaet TA, ed. Prog Hemost
Such complexes, however, may still bind GAG. In this
Thromb 1984:7:167-182
regard, Kostner et al. 82 reported that Lp(a) and LDL 12. Lobb RR, Harper JW, Fett JW. Purification of heparin-
readily formed complexes with GAG that caused massive binding growth factors. Anal Biochem 1986;154:1-14
cholesteryl ester accumulation in mouse peritoneal mac- 13. Llndahl U, Felngold DS, Roden L. Biosynthesis of heparin.
Trends in biochemical sciences 1986;11:221-225
rophages. Most recently, Horn-Brahimi et al. 83 reported a 14. Lam LH, Sllbert JE, Rosenberg RD. The separation of
role for proteoglycan in the binding of LDL to cultural active and inactive forms of heparin. Biochem Biophys Res
arterial smooth muscle cells and proposed that the LDL- Commun 1976:69:570-577
proteoglycan interaction leads to lipoprotein cholesterol 15. Cardln AD, Randall CJ, Hlrose N, Jackson RL. Physical-
chemical interaction of heparin and human plasma low
accumulation in vascular tissues. Camejo et al. 84 showed density lipoproteins. Biochemistry 1987;26:5513-5518
that three of nine synthetic peptides corresponding to 16. Hlrose N, Blankenshlp DT, Krivanek MA, Jackson RL,
different hydrophilic regions of apo B-100 inhibited the Cardln AD. Isolation and characterization of four heparin-
binding cyanogen bromide peptides of human plasma apo-
binding of LDL to an intima-media chondroitin sulfate lipoprotein B. Biochemistry 1987;26:5505-5512
proteoglycan. Two of the three peptides, P-2 (3359 to 17. Weisgraber KH, Rail SC Jr. Human apolipoproteln B-100
3377) and P-11 (2106 to 2121), represent apo B-100 heparin-binding sites. J Biol Chem 1987;262:11097-11103
heparin-binding domains,18'17 and the third, P-1 (4230 to 18. Cardln AD, Hlrose N, Blankenshlp DT, et al. Binding of a
high reactive heparin to human apolipoprotein E: Identifica-
4254), contained a heparin-binding consensus sequence. tion of two heparin-binding domains. Biochem Biophys Res
With these points in mind, the heparin-binding domains of Commun 1986:134:783-789
apo B-100 should be considered important regions that 19. Weisgraber KH, Rail SC Jr, Mahley RW, Milne RW,
Marcel YL, Sparrow JT. Human apolipoprotein E. Determi-
may affect LDL metabolism in the arterial wall.
nation of the heparin binding sites of apolipoprotein E3. J Biol
Chem 1986:261:2068-2076
20. Suzuki S, Oldberg A, Hayman EG, Plerschbacher MD,
Acknowledgments Ruoslahtl E. Complete amino acid sequence of human
We thank Susan Treadway and Rita Hoeper for preparing the vitronectin deduced from cDNA. Similarity of cell attachment
manuscript and Steve Biedenbach for photography, Richard L. sites in vitronectin and fibronectin. EMBO J1985;4:2519-2524
Jackson, John L. Krstenansky, and Larry R. McLean for their 21. Black SD, Glorloso JC. MSEQ: A microcomputer-based
helpful suggestions regarding this study, and Hepar Industries for approach to the analysis, display, and prediction of protein
providing the heparin. Cray computer time was provided by Cray structure. BioTechniques 1986:4:448-451, 454-460
Research, Incorporated, Mendota Heights, MN. David A. Deme- 22. Chou PY, Fasman GD. Prediction of protein conformation.
ter (MDRI) and Erich Wimmer and John Mertz (Cray Research, Biochemistry 1974;13:222-245
Incorporated) provided invaluable assistance in these calcula- 23. Gamier J, Osguthorpe DJ, Robson B. Analysis of the
tions. accuracy and implications of simple methods for predicting
the secondary structure of globular proteins. J Mol Biol 1978;
120:97-120
References 24. Cowan SW, Bakshl EN, Machln KJ, Isaacs NW. Binding
1. Long WF, Williamson FB. Glycosaminoglycans and the of heparin to human platelet factor 4. Biochem J 1986;
control of cell surface proteinase activity. Med Hypotheses 234:485-488
1983;11:285-308 25. Struthers RS, Rlvler J, Hagler AT. Design of peptide
2. Rosenberg RD. Biological actions of heparin. Semin Hema- analogs: Theoretical simulation of conformation. Energetics
tol 1977;14:427-440 and dynamics in conformationally directed drug design:
3. Wight TN. Vessel proteoglycans and thrombogenesis. In: Peptides and nucleic acids as templates or targets. Vida JA,
Spaet TH, ed. Prog Hemost Thromb New York: Grune & Gordon M, eds. Washington, DC: American Chemical Soci-
Stratton, 1980;5:1-39 ety, 1984:239
4. Knox P. The cell surface in health and disease. Mol Aspects 26. Dayringer HE, Tramontano A, Sprang SR, Ftetterick RJ.
Med 1984:7:177-311 Interactive program for visualization and modelling of pro-
5. Castellot JJ Jr, Addonlzlo ML, Rosenberg R, Kamovsky teins, nucleic acids and small molecules. J Mol Graphics
MJ. Cultured endothelial cells produce a heparinlike inhibitor 1986:82-87
of smooth muscle cell growth. J Cell Biol 1981^0:372-379 27. Hagler AT, Huler E, Ufson S. Energy functions for peptides
6. Majack RA, Cook SC, Bomsteln P. Control of smooth and proteins. I. Derivation of a consistent force field including
muscle cell growth by components of the extracellular matrix: the hydrogen bond from amide crystals. J Am Chem Soc
autocrine role for thrombospondin. Proc Natl Acad Sd USA 1974:96:5319-5327
1986:83:9050-9054 28. Hagler AT, Llfson S. Procedure for obtaining energy pa-
7. Robinson DS. The clearing factor lipase and its action in the rameters from crystal packing. Acta Crystallogr 1974;
transport of fatty acids between the blood and the tissues. 30:1336-1341
Adv Lipid Res 1963;1:1-133 29. Alllnger NL Conformational analysis. 130. MM2. A hydro-
8. Hollander W. Unified concept on the role of acid muco- carbon force field utilizing V, and V2 torsional terms. J Am
polysaccharides and connective tissue proteins in the accu- Chem Soc 1977;99:8127-8134
mulation of lipids, lipoproteins, and calcium in the atheroscle- 30. Hagler AT, Stern PS, Llfson S, Ariel S. Urey-Bradley force
rotic plaque. Exp Mol Pathol 1976:25:106-120 field, valence force field, and ab initio study of intramolecular
9. Morrison LM, Varma R. Glycosaminoglycaris and coronary forces in tri-tert-butyimethane and isobutane. J Am Chem
heart disease. In: Vaima RS, Varma R, eds. Glycosamino- Soc 1979;101:813-819
HEPARIN-BINDING PROTEINS Cardin and Weintraub 31

31. Hagler AT, Lltson S, Dauber P. Consistent force field localization, restriction fragment length polymorphism, and
studies of intermolecular forces in hydrogen-bonded crys- expression in Escherichia coli. Biochemistry 1988;27:752-759
tals. 2. A benchmark for the objective comparison of alter- 54. Esch F, Balrd A, Ling N, et al. Primary structure of bovine
native force fields. J Am Chem Soc 1979;101:5122-5130 pituitary basic fibroblast growth factor (FGF) and comparison
32. Hagler AT, Dauber P, LJfson S. Consistent force field with the amino-terminal sequence of bovine brain acidic
studies of intermolecular forces in hydrogen-bonded crys- FGF. Proc Natl Acad Sci USA 1985;82:6507-6511
tals. 3. The C=O H-O hydrogen bond and the analysis of the 55. Wallner BP, Mattallano RJ, Hesslon C, et al. Cloning and
energetics and packing of carboxylic acids. J Am Chem Soc expression of human lipocortin, a phospholipase A2 Inhibitor
1979;101:5131-5141 with potential anti-inflammatory activity. Nature 1986;
33. Dauber P, Hagler AT. Crystal packing, hydrogen bonding, 320:77-81
and the effect of crystal forces on molecular conformation. 56. Suzuki K, Deyashlkl Y, Nlshloka J, et al. Characterization
Ace Chem Res 1980;13:105-112 of a cDNA for human protein C inhibitor: A new member of
34. Elsenberg D, Weiss RM, Terwllllger TC. The hydrophobic the plasma serine protease inhibitor superfamily. J Biol
moment detects periodicity in protein hydrophobicity. Proc Chem 1987;262:611-616
Natl Acad Sci USA 1984;81:140-144 57. Cardin AD, Jackson RL, Sparrow JT. Interaction of glycos-
35. Schrffer M, Edmunson AB. Use of helical wheels to repre- aminoglycans with lipoproteins. Ann NY Acad Sci 1988 (in
sent the structures of proteins and to identify segments with press)
helical potential. Biophys J 1967;7:121-135 58. Welsgraber KH, Innerartty TL, Harder KJ, et al. The
36. Hynes R. Molecular biology of fibronectin. Annu Rev Cell receptor-binding domain of human apolipoprotein E. Mono-
Biol 1985; 1:67-90 clonal antibody inhibition of binding. J Biol Chem 1983;
37. Dlxrt VM, Hennessy SW, Grant GA, Rotweln P, Frazler 258:12348-12354
WA. Charactenzation of a cDNA encoding the heparin and 59. Peterson CB, Noyes CM, Pecon JM, Church FC, Black-
collagen binding domains of human thrombospondin. Proc bum MN. Identification of a lysyl residue in antlthrombin
Nat) Acad Sci USA 1986;83:5449-5453 which Is essential for heparin binding. J Biol Chem 1987;
38. Lawler J, Hynes RO. The structure of human thrombospon- 262:8061-8065
din, an adhesive glycoprotein with multiple calcium-binding 60. Pecon JM, Blackburn MN. Pyridoxylation of essential lysines
sites and homokxjies with several different proteins. J Cell in the heparin-binding site of antrthrombin III. J Biol Chem
Biol 1986;103:1635-1648 1984;259:935-938
39. Wlon KL, Klrchgessner TG, Lusls AJ, Schotz MC, Lawn 61. Smith JW, Knauer DJ. A heparin binding site in antrthrom-
RM. Human lipoprotein lipase complementary DMA sequence. bin III. Identification, purification, and amino add sequence.
Science 1987;235:1638-1641 J Biol Chem 1987;262:11964-11972
40. Martin GA, Busch SJ, Meredith GD, et al. Isolation and 62. Bock SC, Skrlver K, Nielsen E, et al. Human C1 inhibitor
cDNA sequence of human postheparin plasma hepatic tri- primary structure, cONA cloning, and chromosomal localiza-
glyceride lipase. J Biol Chem 1988;263:10907-10914 tion. Biochemistry 198625:4292-4301
41. Komaromy MC, Schotz MC. Cloning of rat hepatic lipase 63. Cardin AD, Jackson RL Structural properties of the heparin-
cDNA: Evidence for a lipase gene family. Proc Natl Acad Sci binding domains of human apolipoprotein E. In: Malmendier
USA 1987;84:1526-1530 CL, Alaupovic P, eds. Eicosanoids, apolipoproteins, lipopro-
42. Harper JW, Lobb RR. Reductive methylatton of lysine res- tein particles and atherosclerosis, (in press)
idues in acidic fibroblast growth factor: Effect on mitogenic 64. Sparrow JT, Sparrow DA, Culwell AR, Gotto AM Jr.
activity and heparin affinity. Biochemistry 1988;27:671-678 Apolipoprotein E: Phospholipid binding studies with synthetic
43. Esch F, Ueno N, Balrd A, et al. Primary structure of bovine peptides containing the putative receptor binding region.
brain acidic fibroblast growth factor (FGF). Biochem Biophys Biochemistry 1985;24:6984-6988
Res Commun 1985;133:554-562 65. Carroll RV, Chrlstey PB, Boswell DR. SERPIN:
44. Fenton JW II, Bing DH. Thrombin active-site regions. Semin Antithrombin-lll and other inhibitors of coagulation and fibrino-
Thromb Hemost 1986;12:200-208 lysis. Evidence from amino acid sequence. In: Verstraete N,
45. Cunningham BA, Hemperly JJ, Murray BA, Predlger EA, Vermylen J, LJgmen HR, Arnout J, eds. Thrombosis and
Brackenbury R, Edelman GM. Neural cell adhesion mole- hemostasis. International Society of Thrombosis and Hemo-
cule: Structure, immunoglobulin-like domains, cell surface stasis, Leuzen: Leuzen University Press, 1987:1-15
modulation, and alternative RNA splicing. Science 1987; 66. Stone A L Optical conformation of heparin and heparin
236:799-806 complexes with catlonic dyes, amines and protein models.
46. Sommer J, Gloor SM, Rovelll GF, Hotsteenge J, Nick H, Fed Proc 1977;36:101-106
' Meier R, Monard D. cDNA sequence coding for a rat 67. Lubec G. Interactions of N-CAM with heparin-like molecules.
glial-derived nexin and its homology to members of the Nature 1986:323:743-744
serpin superfamily. Biochemistry 1987;26:6407-6410 68. Tlkkanen M, Schorrfeld G. The recognition domain for the
47. Luster AD, Unkeless JC, Ravetch JV. y-lnterferon tran- low density lipoprotein cellular receptor is expressed once on
scriptionally regulates an early-response gene containing each lipoprotein particle. Biochem Biophys Res Commun
homology to platelet proteins. Nature 1985;315:671-676 1985:126:773-777
48. Kolde T, Foster D, Yoshrtake S, Davle EW. Amino acid 69. Milne RW, Marcel YL. Monoclonal antibodies against human
sequence of human histidine-rich glycoprotein derived from low density lipoprotein. Stoichiometric binding studies using
the nucleotide sequence of its cDNA. Biochemistry 1986; Fab fragments. FEBS Lett 1982;146:97-100
25:2220-2225 70. Marcel YL, Hogue M, Theolis R Jr, Milne RW. Mapping of
49. Berman P, Gray P, Chen E, et al. Sequence analysis, antigenic determinants of human apolipoprotein B using
cellular localization, and expression of a neuroretina adhe- monoclonal antibodies against low density lipoproteins. J
sion and cell survival molecule. Cell 1987;51:135-142 Biol Chem 1982;257:13165-13168
50. Begg GS, Pepper DS, Chesterman CN, Morgan FJ. Com- 71. Goldstein JL, Basu SK, Brunschede GY, Brown MS.
plete covalent structure of human £-thromboglobulin. Bio- Release of low density lipoprotein from its cell surface
chemistry 1978;17:1739-1744 receptor by sutfated glycosaminoglycans. Cell 1976;7:85-95
51. Bock SC, Wlon KL, Vehar GA, Lawn RM. Cloning and 72. Nakashlma Y, DIFerrante N, Jackson RL, Pownall HJ.
expression of the cDNA for human antithrombin-lll. Nucleic The interaction of human plasma glycosaminoglycans with
Acids Res 1982;10:8113-8125 plasma lipoproteins. J Biol Chem 1975:250:5386-5392
52. Ragg H. A new member of the plasma protease inhibitor 73. Cardin AD, Ranganathan S, HI rose N, et al. Effect of
gene family. Nucleic Acids Res 1986;14:1073-1087 trypsin treatment on the heparin- and receptor-binding prop-
53. Blinder MA, Marasa JC, Reynolds CH, Deaven LL, Tolkrf- erties of human plasma low-density lipoproteins. Biochemis-
sen DM. Heparin cofactor II: cDNA sequence, chromosome try 198625:5258-5263
32 ARTERIOSCLEROSIS VOL 9, No 1, JANUARY/FEBRUARY 1989

74. Hoff HF, Wagner WD. Plasma low density llpoproteln accu- protein retention by aortic tissue. Contribution of extracellular
mulation In aortas of hypereholesterolemic swine correlates matrix. Atherosclerosis 1986;62:201-208
with modifications in aortic glycosaminogtycan composition. 80. Ye SQ, Trleu VN, Stlers DL, McConathy WJ. Interactions
Atherosclerosis 1986;61:231-236 of low density lipoprotein^ and other apolipoprotein B-
containing lipoproteins with lipoprotein(a). J Biol Chem 1988;
75. Schumaker VN, Robinson MT, Curtlss LK, Butler R, 263:6337-6343
Sparkes RS. Anti-apoprotein B monoclonal antibodies detect 81. Armstrong VW, Walll AK, Seldel D. Isolation, characteriza-
human low density lipoprotein polymorphism. J Biol Chem tion, and uptake in human fibroblasts of an apo{a)-free
1984:259:6423-6430 lipoprotein obtained on reduction of lipoprotein(a). J Lipid
Res 1985:26:1314-1323
76. Chan L, VanTulnen P, Led better DH, Dakjer SP, Gotto AM
82. Kostner GM, Varga MB, Rotheneder M, Gruber E, Zech-
Jr, Chen SH. The human apolipoprotein B-100 gene: A highly ner R. The interaction of lipoprotein-a with glycosaminogly-
polymorphic gene that maps to the short arm of chromosome cans: Possible mechanism for its atherogenictty. Circulation
2. Biochem Biophys Res Commun 1985;133546-255 1986;74(suppl ll):ll-25
77. Rapacz J, Hasler-Rapacz J, Taylor KM, Checovlch WJ, 83. Hom-Brahimi M-C, Breton M, Berrou E, Deudon E, Plcard
Artie AD. Lipoprotein mutations in pigs are associated with J. Role of proteoglycan in the binding of low-density lipopro-
elevated plasma cholesterol and atherosclerosis. Science tein to cultured arterial smooth muscle cells. Artery 1988;
1986:234:1573-1577 15:122-139
78. Vega GL, Grundy SM. In vivo evidence for reduced 84. Camejo G, Olofsson S-O, Lopez F, Carlsson P, Bondjers
G. Identification of apo B-100 segments mediating the inter-
binding of low density lipoproteins to receptors as a cause action oftowdensity lipoproteins with arterial proteoglycans.
of primary moderate hypercholesterolemia. J Clin Invest Arteriosclerosis 1988;8:368-377
1986;78:1410-1414 85. Pervalz S, Brew K. Homology of /3-lactoglobulin, serum
79. Srlnlvasan SR, Vl|ayagopal P, Datferes ER Jr, Abbate B, retinol-binding protein and protein HC (a,-microglobulin).
Radhakrishnamurthy B, Berenson GS. Low density lipo- Science 1985;228:335-337

Index Terms: heparin • glycosaminoglycans • consensus sequences lipoproteins • apolipoproteins


• molecular dynamics