Structure of the platelet membrane glycoprotein IIb. Homology to the ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Ine.

Vol. 262. No. 18, Issue of June 25. pp. 8476-&182,1987 Printed in U.S.A.

Structure of the Platelet Membrane GlycoproteinIIb HOMOLOGY T O T H E a SUBUNITS OF THE VITRONECTIN MEMBRANERECEPTORS*

AND FIBRONECTIN

(Received for publication, March 6, 1987)

Mortimer PonczSB, Robin EismanS, Randy HeidenreichS, Samuel M. Silvern, GastonVilairen, Saul Surrey& Elias SchwartzS, and Joel S.Bennettli)) From the #Divisionof Hematology, The Children’s Hospital of Philadelphia, the llHematology-Oncology Section, The Hospital of the University of Pennsylvania, and the Departments of Pediatrics, Human Genetics, and Medicine, The University of Pennsylvania School of Medicine, Philndelphia, Pennsylvania 19104

The platelet membrane glycoproteinIIb. IIIa heterodimer complex (GPIIbmIIIa) is the platelet receptor calcium-dependentheterodimer,containsbindingsites for for adhesive proteins, containing binding sites for fi- fibrinogen, von Willebrand factor, and fibronectin and mebrinogen, von Willebrand factor, and fibronectin on diates adhesive reactions between activated platelets (Jenactivated platelets. GPIIb.1IIa also appears to be a nings and Phillips, 1982; Bennett andVilaire, 1979; Marguerie member of a family of membrane adhesive protein et al., 1979; Ruggeri et al., 1982; Ginsberg et al., 1983). Glyreceptors that plays a major role in cell-cell and cell- cosylated glycoprotein IIb (GPIIb)’ has an apparent M, of matrix interactions. GPIIb is the larger component of approximately 136,000 on SDS-polyacrylamide gels and conthis platelet receptor and is composed of two disulfide- tains two disulfide-linked components, a larger subunit of M , linked subunits.In this report describe we the analysis 125,000 and a smaller subunit of M, 23,000 (Jennings and of cDNA clones for human GPIIb that were isolated Phillips, 1982). Glycoprotein IIIa (GPIIIa) is a single-chain from a Xgt 11 expression library prepared using RNA disulfide bond-rich glycoprotein with an unreduced apparent from HEL cells. A total of 3.3 kilobases of cDNA was 110,000 following disulfide bond sequenced, revealing a continuous open reading frame M,of 95,000 that increases to encoding both GPIIb subunits. The cDNA encodes 1039 reduction (Jennings and Phillips, 1982). GPIIb and GPIIIa amino acids: 137 constituting the smaller subunit, 871 are deficient toa similar extent in platelets fromindividuals constituting the larger subunit, and 30 constituting an with the autosomal recessive disorder Glanzmann’s thrombNH2-terminal signal peptide. No homology was found asthenia, suggesting that the synthesis of these proteins is between the larger knd smaller subunits. The smaller under coordinate control (Phillips and Agin, 1977). Interest in the GPIIb. IIIa complex has broadened following speculasubunit contains a 26-residue hydrophobic sequence near its COOH terminus that represents a potential tion that GPIIb.IIIa is a member of a family of membrane transmembrane domain. Four stretches of 1 2 amino adhesive protein receptors that plays major roles in cell-cell acids present in the larger subunitare homologous to andcell-matrixinteractions(Hynes,1987).Thisputative receptor, the the calcium binding sites of calmodulin and troponin family includesthe mammalian fibronectin (FN) C. Northern blot analysis using HEL cell RNA indi- chick embryo fibroblast fibronectin receptor, the vitronectin cated that the mature mRNA coding for GPIIb is 4.1 (VN) receptor, and the LFA-1 and Mac-1 antigens on leukokilobases in size. A comparison of the GPIIb coding cytes (Pytela et al., 1985a, 1985b; Tamkun et al., 1986, Davregion with availablecDNA sequences of the a-chains ignon et at., 1981; Beller et al., 1982; Springer et al., 1985). of the vitronectin and fibronectin receptors revealed Several of these receptors are heterodimers composed of a 41% DNA homology and 74% and 63% aminoacid two-subunit CY chain anda single component (3 chain. Several homology, respectively. Our data establish the amino of these heterodimers also interact with proteins containing acid sequence for the human platelet glycoproteinIIb thetetrapeptide sequenceArg-Gly-Asp-Ser (Pierschbacher and provide additional evidence for the existenceof a and Ruoslahti, 1984; Gartner and Bennett,1985; Ginsberg et family of cellular adhesion protein receptors. al., 1985; Haverstick et al., 1985). Furthermore, comparisons of the aminoacid sequencesof the amino terminiof the larger subunit of GPIIb and the CY chains of LFA-1 and Mac-1 (Charo et al., 1986) and of partial cDNA nucleotide sequences The platelet membrane glycoprotein IIb.IIIa complex,a of the CY chains of the FN and VN receptors(Argraves et al., 1986; Suzuki et al., 1986) suggestthat homology exists among * Supported in part by Grants HL37419, AM16691, HL28157,and these proteins. HL23809 from the National Institutes of Health, Grant 1570 from In order to characterize the structure of platelet GPIIb and The Council for Tobacco Research-U. S. A., Inc., a grant from the University of Pennsylvania, and a Basil OConnor Starter Research GPIIIa, we sought to isolatecDNA for each protein. Because Grant from the March of Dimes. The costs of publication of this circulating human platelets contain little RNA and because article were defrayed in part by the payment of page charges. This bonemarrowmegakaryocytes, theplateletprecursors,are article must therefore be hereby marked “advertisement” in accord- difficult t o obtainin large numbers, we used thehuman ance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paperhas been sthmitted erythroleukemic cell line HEL for our studies (Martin and to theGenBankTM/EMBLDataBankwith accession number(s) Papayannopoulou, 1982). HEL cells have megakaryocyte-like 502764.

5 To whom correspondence should be addressed Division of Hematology, The Children’s Hospital of Philadelphia, 34th St. and Civic Center Blvd., Philadelphia, PA 19104. I( Established Investigator of the American H e a j Association.

The abbreviations used are: GPIIb, glycoprotein IIb; GPIIIa, glycoprotein IIIa; SDS, sodium dodecyl sulfate; FN, fibronectin; VN, vitronectin; IPTG, isopropyl-1-thio-P-D-galactopyranoside; kb, kilobase; bp, base pair.

8476

Structure of Glycoprotein IIb

8477

1 mM B-mercaptoethanol, and fresh lysozyme added to a concentraproperties and constitutively express a number of platelet tion of 0.2 mg/ml. After 30 min a t 25 'C, 0.1 mM phenylmethanesulproteins, such as platelet factor 4, 8-thromboglobulin, von fluoride was added, and the bacterial lysatewas sonicated for 1 Ib, a n d GPIIb and GPIIIa fonyl Willebrand factor, glycoprotein min a t 4 "C. Immunoblotting was then performed as previously de(Tabilio et al., 1984; Bray et al., 1986; Silver et al., 1987). scribed (Bennett etal., 1982). Briefly,lystates from IPTG-stimulated T h u s , Poncz et al. (1987)isolated several cDNA clones for or unstimulated bacteria and suspensions of washed human platelets plateletfactor 4 from a cDNA library constructed i n the were dissolved in 1% sodium dodecyl sulfate (SDS) and electrophoexpressionvector Xgtll using HEL cell mRNA. We have resed on 0.1% SDS-7.5% polyacrylamide slab gels (Laemmli, 1970). screened this library with polyclonal antibodies against puri- The separated proteins were then electrophoretically transferred to strips of nitrocellulose paper, and the paperwas incubated with 1/20 fied human platelet GPIIb and GPIIIa and identified several dilutions of either rabbit anti-@-galactosidase antisera (Cappel-WorcDNA clones for each protein. We report here t h e c o m p l e t e thington), rabbitpolyclonal anti-GPIIb IgG, or the anti-human platecDNA sequence forthe m a t u r e GPIIb protein and an analysis let GPIIbmonoclonal antibody B1B5 (Silver et al., 1987). '2sI-labeled of several of its structural features. Furthermore, we demon- staphylococcal protein A (Boehringer Mannheim)was used to localize strate homology between GPIIb and the available sequences the sitesof antibody binding, and thelabeled immunoblots were dried and exposed to Kodak X-Omat AR photographic film for 24 h a t f o r t h e a chains of the FN and VN receptors. -70 "C. Northern Blot Analysis of HEL rnRNA-Northern blots were perEXPERIMENTALPROCEDURES formed as described by Thomas (1983). Thirteen micrograms of total Identification of cDNA Clones for CPllb-Total cellular RNA was cellular RNA from HEL cells and from K562 either unstimulated or M phorbol 12-myristate 13-acetate was denaprepared from dimethylsulfoxide-stimulatedHEL cells by the stimulatedwith method of Chirgwin et 01. (1979), and poly(A)' RNA was isolated by tured in thepresence of 1 M glyoxal, 50% dimethyl sulfoxide, electrochromatographyon oligo(dT)-cellulose (Aviv and Leder, 1972). A phoresed on 1% agarose gels, and transferred to nylon membranes (Nytran, Schleicher & Schuell). Membranes were prehybridized for cDNA expression library in Xgtll was constructed from the poly(A)' RNA aspreviously described (Poncz etal., 1987). Briefly, first-strand 20 h a t 42 "C in buffer containing 50% formamide, 5 X SSC (1 X SSC cDNAwas synthesized using oligo(dT)primers(Pharmacia P-L = 150 mM NaCI, 15 mM sodium citrate), 250 pg/ml salmon sperm Biochemicals) and reverse transcriptase (Life Sciences) (Maniatis et DNA (Sigma) and bovine serum albumin, Ficoll, and polyvinylpyra!., 1982). Double-stranded cDNAwas prepared using RNaseH (New rolidone, each at 0.02%. Hybridizationwasperformed for 24 h a t England Biolabs), reverse transcriptase, and Escherichia coli ligase 42 'C in the prehybridization buffer using DNA probes labeled with (Bethesda Research Laboratories) (Gubler and Hoffman, 1983). The '*P a t 10' cpm/ml. Following hybridization, the membranes were double-stranded cDNAs were treated with the Klenow fragment of washed four times with 2 x SSC, 0.1% SDS at 22 "C, twice with 0.1 DNA polymerase (BRL) to produce blunt ends, and EcoRI linkers X SSC containing0.1% SDS at50 "C, and exposed to Kodak X-Omat (Pharmacia P-LBiochemicals) were ligated to the cDNA fragments. AR photographic film a t -70 "C for 12-24 h. mRNA size was deterFollowing EcoRI digestion and removal of the digested linkers by mined using commercially available RNA size markers (BRL). Computer Analyses-Computer analysis was performed using the fractionationon acolumn of SepharoseCL-4B(PharmaciaP-L acid sequences of Biochemicals), thecDNAwas ligated into EcoRI-digested Xgtll, BionetTM system(Intelligenetics).Theamino packaged, and transfected intoE. coli Y 1090. Monospecific polyclonal GPIIb, the FN receptor, and the VN receptor were initially aligned I g G against human platelet GPIIbwas used to screen theexpression using the FASTP program (Lipman and Pearson, 1985) and GENAwas maximizedusing the library for GPIIb clones (Poncz et al., 1987). These antibodies were LIGN (BionetT"), andthealignment Needleman-Wunschalgorithm(NeedlemanandWunsch, 1970). prepared by immunizing New Zealand White rabbits with purified by the human platelet GPIIb as previously described (Silver et al., 1987). DNA homology was based uponan alignment determined first Clones reacting with the anti-GPIIb I g G were detected using a bio- amino acid comparison. The hydrophobicity profile was calculated as tinylated goat anti-rabbit antiserum coupled withan avidin-horserad- previously described (Kyte andDoolittle, 1982) using segment lengths ish peroxidase complex (Vector Laboratories) and were plaque-puri- of 6 and 11 amino acids. fied. Isolation and Sequencing of GPIIb cDNA Clones-DNA from phage containing putative GPIIb cDNA waspurified and digested with EcoRI. cDNA inserts were then isolated by electroelution and subcloned into M13mp18 or 19 (Yanisch-Perronet al., 1985). These constructs were transfected into JM107, and single-stranded DNA in 2oo both insert orientations wasprepared. Insert orientation and the degree of clone homology was determined by annealing single'I6 stranded DNA from independent isolates and digesting the hybrids 97.4 l? with S1 nuclease (BRL) as previously described (Poncz et al., 1982). T o obtain largercDNA, the Xgtll librarywas screened with "'P168 labeled cDNA isolated from the replicative form of M13. Positive clones were plaque-purified, and DNA inserts were subcloned into M13 as describedabove. The nucleotide sequences of the cDNA inserts were determined by analysis of overlapping Bal31-generated deletion mutants (Poncz et al., 1982) using the dideoxy chain termination techniqueof Sanger etal. (1977) and the universal M13 primer (BRL) (Messing,1983). Complete nucleotide sequences were obtained for both insert orientations. 30 Irnrnunoblotting of Fusion Proteins-To confirm the identity of FIG. 1. Immunoblotting of fusion protein from bacteria putative GPIIb cDNA, immunoblotting of the fusion proteins expressed by E. coli containing Xgtll was performed using bothpolyclo- transfected w i t h a 1.9-kb GPIIb cDNA clone. E. coli Y1089 were nal anti-GPIIb I g G and the monoclonal anti-human platelet GPIIb transfected with a Xgtll clone containinga 1.9-kb cDNA, and fusion protein synthesis was induced with IPTG asdescribed under "Experantibody B1B5 (Silver et al., 1987). Growth of transfected bacteria followed a modification of a previously published procedure (Govin- imental Procedures." Bacterial proteins and proteins from washed human plateletswere then separatedby electrophoresis on 0.1% SDSdan etal., 1985). Fifty microliters of Xgtll stock (10'' plaque-forming units/ml) and 500 pl of an overnight growth of E. coli Y1089 were 7.5% polyacrylamide slab gels, transferred to nitrocellulosepaper, coli @-galactosidase, added to 10 ml of NZCYM medium (Maniatis et al., 1982) containing and immunoblotted using antisera against E. ampicillin (10 pg/ml) and incubated with vigorous agitation a t 35 "C monospecific polyclonal I g G againstplateletGPIIb,andanantifor 3 h. Induction of fusion protein synthesis was carried out by human platelet GPIIb-specific monoclonal antibody BlB5. Antibody incubation at 45 "C for 20 min, followed by the addition of 1 mM binding sites were identified with '2sI-labeled staphylococcal protein isopropyl-1-thio-B-D-galactopyranoside (IPTG) and growth a t 37 "C A. Lane 1 , platelets; lane 2, bacterial cultures induced with IPTG; mobility of native for 2 h. Bacteria were pelleted and resuspended in 0.5 ml of 20 mM lane 3, uninduced cultures. The arrow indicates the Tris acetate buffer, pH 7.5, containing 0.1 mM EDTA, 1 mM MgC12, @-galactosidase.

t t

f

t

8478 1

e 176 237 398 589 620

731 042

953

1064

FIG.2. DNA sequence for GPIIb. 1175 The 3311-bp sequence hasa continuous open reading frame encoding 1039 amino acids includinga probable signal peptide 1286 (amino acids are indicatedby the single 1397 letter code under the nucleotide sequence). The previously reported NH11% terminal amino acids of the large and small subunits of platelet GPIIb are 1619 underlined by the heavy solid lines. A potentialhydrophobictransmembrane 1738 domainis underlined by a dotted line, and the location of a poly(A) addition 1841 signalsequenceis indicated by a box. Possiblecalciumbinding sites are en1952 closed within boxes of dotted lines. Potential N-linked glycosylation sites are 2863 overlined.

2174 2285

2 m 2587 2618 2729

2848 2351

3062 3173

3284

RESULTS

Isolution of cDNA Clones Encoding GPIIb-Three clones containing cDNA inserts of 1.0, 1.9, and 2.3 kilobases (kb) were identified with polyclonal anti-human platelet GPIIb IgG during an initial screen of a lawn of E. coli Y1090 infected with 2.4 x lo5 recombinant Xgtll. Following plaque purification of these clones, their cDNA inserts were excised with EcoRI and inserted into the bacteriophage vector M13mp19 (Yanisch-Perron et al., 1985). Single-stranded M13 DNAcontainingdifferentcDNAinserts were cross-hybridized and digested with S1 nuclease (Poncz et al., 1982), demonstrating DNA homology between the inserts (data not shown). To provide preliminary evidence that these cDNA inserts were authentic GPIIb cDNA, thefusion protein synthesized by recombinant phage containing the 1.9-kb cDNA was examined by immunoblotting with the monoclonal antibody

BlB5, an antibodyspecific for the smaller subunitof human platelet GPIIb (Silver et al., 1987). As seen in Fig. 1, E. coli containing the recombinant Xgtll and incubated with IPTG produced a fusion protein containing@-galactosidaseantigens that was larger than thewild-type enzyme. The fusion protein also reacted withpolyclonal anti-human GPIIb IgG and with the monoclonal antibody BlB5, indicating that it contained antigensfoundon at least the smaller subunit of human platelet GPIIb. Bray et al. (1986) provided evidence that HEL cells synthesize a single GPIIb precursor containing both subunits of the mature protein. Because none of the cDNAs we identified with anti-GPIIb IgG were of sufficient size to contain the entire GPIIb sequence, we rescreened the Xgtll library with a radiolabeled 401-bp fragment of the 5' end of the 1.9-kb cDNA. Several overlapping clones with a n overall length of

8479


...................

10

10

10

......................

0

.IO

.10

.IO

FIG. 3. The hydropathic profile of GPIIb. The hydropathic profile of GPIIb was calculated as described the method of Kyte and Doolittle (1982).The hydrophobicity of each under “Experimental Procedures” following 5-bp segment of the GPIIb cDNA sequence is shown on the y axis. The cross-hatched areas are hydrophobic a transmembrane domain.The arrows indicate stretches that may represent an amino-terminal signal peptide and the NH, termini of the two subunits of the mature GPIIb.

acid sequences areencoded by our cDNA sequence beginning at residues 487 and 1,027. Thus, these data confirm that our cDNAs are authentic GPIIb cDNAs and that GPIIb is translated as a larger precursor that is cleaved into two subunits +Pl-+I post-translation. Amino acids 1-31 encoded by the cDNA are FIG. 4. Schematic map of the GPIIb precursor. The larger strongly hydrophobic (Fig. 3), begin with a methionine, and and smaller GPIIbsubunits have been termed CY and p (Jennings and terminate with thesequence Ala-X-Ala that often precedes a Phillips, 1982). The shaded area represents a possible signal peptide, signal peptide cleavage site (Perlman and Halvorson, 1983). whereas the stippled area may represent a transmembrane domain. The four cross-hatched bars are regions homologous to the calcium This suggests that amino acids 2-31 constitute an aminobinding sites of calmodulin and troponin C and may represent calcium terminal signal peptide and that the reading frame for GPIIb binding sites. The region filled with horizontal lines is a likely cyto- begins with the methionine at aminoacid 1. A stretch of 26 plasmic domain. Possible N-linked glycosylation sites are indicated strongly hydrophobic amino acids is also present near the by solid triangles. carboxyl terminus of the smaller subunit and may represent a transmembrane anchor, consistent with the supposition that 3.3 kb were identified during thisscreening. GPIIb is an intrinsic membrane protein (Jennings and PhilGPIIb cDNA Sequence-The cDNAs were subcloned into lips, 1982). Five possible sites for N-linked glycosylation are M13mp18 and mp19, and their nucleotidesequences were determined completely in both orientations.The cDNAs con- present in the sequence, each in the portion of the protein tain 3,311 bp and encode a continuous open reading frameof that is likely to be extracellular (Figs. 2 and 4). Several lines of evidence indicate that GPIIb contains bind1,039 amino acids. The open reading frame is followed by a ing sites for calcium (Fujimura and Phillips, 1983a, 1983b). short 182-bp 3’ untranslated sequence containing a poly(A) Therefore, we examined the GPIIb sequence for amino acid addition sequence beginning 20 bp upstream from a poly(A) stretches homologous to the calcium binding sites of known stretch (Fig. 2). Charo et al. (1986) directly determined the sequences of the 15 amino-terminal amino acids from the calcium binding proteins. Four stretches of 12 amino acids present in the sequence of the larger subunit of GPIIb and larger and smaller subunits of platelet GPIIb. These amino acids are encoded in our cDNA sequence beginning a t amino beginning at residues 274, 328, 395, and 457 (Figs. 2 and 4) acid 32 for the larger subunit sequence and at aminoacid 903 are homologues of the calcium-bindingregions of troponin C for the smaller subunit. This indicates that the maximum size and calmodulin (Table I) (Kretsinger, 1976; Williams, 1986). of the larger subunit of GPIIb is 871 amino acids with a Secondary structure analysis(Chou and Fasman, 1978) (data not shown) also indicates that these amino acids are located calculated molecular weight of 94,507 and that the smaller subunit contains 137 amino acids witha calculated molecular either at the boundary of an a helix and a @ turn or within a weight of 15,514. Hiraiwa et al. (1987) also have published region of turns, suggesting that they would be available to partial amino acid sequences for two human platelet GPIIb coordinate with ionic calcium. fragments produced with lysyl endopeptidase. These amino Comparison of the GPIIb Sequence with Those of Other Signal Peptide

Transmembrane Domain


8480 TABLE I

and theavailable partial sequences for the a chains of the VN and FN receptors (Argraves et al., 1986; Suzuki et al., 1986). The FASTP program (Lipman and Pearson, 1985) and GENALIGN (BionetTM)were used for an initialcomparison of the Residues Protein Aminoacid sequence amino acid and the DNA sequences. To increase alignment, GPIIb 274-285 E-D-DLN-T--V adjustmentsin sequence comparison were needed atthe D-N-DGR-D--V 328-339 amino terminus of the FN receptor (Fig. 5 ) . The GPIIb and 395-406 D-D-DGY-D--V VN receptor a chain sequences are homologous over the entire 457-468 D-D-NGY-D--V 636 amino acids of the partial VN receptor sequence, with 34% identicalamino acidresidues and an additional 40% Troponin C’ 27-38 D-D-GGD-S--E 63-74 D-D-SGT-D--E representing conservative amino acid substitutions. In addi14 D-N-DGY-D--E 1 03-1 tion, 10 of the 11 cysteine residues in that portion of GPIIb 1 39-150 D-N-DGR-D--E homologous to the VN receptor occupy similar positions in the VN receptor sequence. The GPIIb and the partial FN Calmodulinb D-D-N-T-T--E 20-3 1 receptor a chain sequences contain 27% identical amino acids 56-67 D-D-N-T-D--E 93-104 D-D-N-Y-S--G and 36% conservative amino acid substitutions. Again, 4 of N-D-D-E-N--E 1 29-140 the cysteines common to the sequences are in similar posia Kretsinger (1976) tions. Using the same alignment, the VN and FN receptors *Williams 1986 were also compared,revealing 30% identical amino acids and 40% conservative amino acid substitutions.Thisanalysis Proteins-A search for homology of our GPIIb sequence with confirms the relationshipof GPIIb with thea chains of both other genes and proteinswas made using the IFIND program the FN andVN receptors. A DNA sequence comparison of the cDNAs for GPIIb and andtheGenBankandNBRFlibraries available onthe BionetTM system (Intelligenetics). No homology of obvious the VN and FN receptor a chains was based upon the amino acid alignment in Fig. 5 and revealed a41%similarity between biologic significance was found between GPIIbandother sequences present in these sequence libraries. However, there GPIIb and both the VN and FN receptors. When GPIIb is is biochemical evidence that GPIIb is related to the a chains compared to the VN receptor, there is 43% identity for the of the VN and FN receptors (Pytela et al., 1986). Conse- first nucleotide of each codon, 58% for the second, and 22% quently, we made a comparison between the GPIIb sequence for the third, consistent with theknown greater restraint on Comparison of the amino mid sequences of the putative calcium binding sites of GPIIb with the amino acid sequences of the calcium binding sites of troponin C and calmodulin

GPIIb

VNR

399

....

I

....

. . . .

...................

. . . . .

GATGHNI-PQKLSLNAELQLDRQKPRQG-RR~LLGS()9AGTTLNLDLGGKHSP

GP1lb

VNR

:.:...

:.. . ::.:.: :. . . . .....

....

E.

TTnAFLRDEADFRDKLSPIVLSLNVSL----PPTEAGMAPA

.:.::::..:::::.::.. .. .:

. . .........

4 9 o r :.:: :: ::. : 97 k a d g y g v l p r k l n f q v e l l l d k l k q k g a l r r a l f l y s r s p ~ h s k n ~ t ~ ~ ~ eglgi al ~y l r d e s e f r d k l t p I t I f . e y r l d y r t a a d t t g l q p 1

* *

. . . . . . . . . . . . . . . . . . . . .

..... FNR

... . . . ..

... . d l g p a v h

1

*

* *** *

*

***

........ . . ..... .. .. ....... .. .. . . . . . .. . . . .. . . . . . .... ..

... ..

.................. . ... . . . . . . . . . . . . . ... . . .. . . . . pi-npkgleldpegelhhqqkr~~p~~~~~s~~pqil

.*

*

I

*

* * *

**

* t

L Q E U A R G Q R A ~ V T ~ A F L Y L P S L Y Q R - - P L D Q F V L 9 S L E - - E R A l P l Y V ~ V G V L G . . . . . . . . . ,

. . . . . . . . . . . . . . . . . . . . . . . . . . . .. ... . . . . . . . . . . .. . . . . .. vgrldrgksaIlyvksllutet[mnkengnhsyslkssasf~~lefpyknlpiedItnstlvtt--nvtvglqpapnpvpvwillavls . . . .... . . . . . . . . . . . .,

VNR

4 9 4 va

GPIIb

9 7 6 GLLLLTILVLAnVKVGFFKRNRPPLEEDDEEGE

VNR

592

FNR

189 g l l l l g l l l y i l y k l g f f k r s l p y g t ~ ~ ~ k ~ q l k p p a t s d ~

****. * ................... ............................ gllllavlvfumym~gffkrvrppq~~q~~~qlqpheng=g”~~t ......................... ............ ****

FIG. 5. Amino acid homology between GPIIb, the vitronectin receptor, and the fibronectin receptor. The alignment of amino acid sequences for GPIIb, the VN receptor (VNR) (Suzuki et al., 1986), and the FN receptor (FNR) (Argraves et al., 1986) was performed as described under “Experimental Procedures.” Amino acid sequences use the single-letter code. The amino acids of GPIIb are indicated by capital letters and are numbered from the first amino acid of the NH2 terminus of the larger GPIIb subunit. The amino acids for the VN receptor Numbering for the VN and the FNreceptor are indicated by lowercase letters except for the cysteine residues and FN receptors begins with the first amino acid of the published sequences. Identical amino acids are indicated by double dots and conservative substitutions by singk dots. The comparison shown here is between GPIIb and either the VN or the FN receptor. The asterisks identify residues that are identical in all three sequences. Boxes indicate the aligned cysteines.

(c).


8481

GPIIb mRNA containsa long 5’ untranslated region. The cDNAsequence for GPIIb confirms recent in vitro experiments indicating that GPIIb is translated as a single polypeptide and then cleaved into two disulfide-linked subunits (Brayet al.,1986; Rosa et al.,1986). However, in contrast 9.5 to an earlier report based on limited amino acid sequence 7.5 information (Charo et al., 1986), we found no similarity between the amino acidsequences of the larger and smaller 4.4 GPIIbsubunits.Ourestimated molecular weight for the GPIIb polypeptide is 110,021 with an M , 94,507 larger subunit and an M , 15,514 smaller subunit. Assuming that processed GPIIb contains 14% carbohydrate (McEveret al.,1982), our 2.4 data suggest a molecular weight of 128,000 for the intact glycoprotein and molecular weights of 110,000 and 18,000 for 1.4 the larger and smaller GPIIb subunits, respectively. The M , 128,000 value for intact glycosylated GPIIb is similar to the valuecalculatedfrom gel filtration chromatography but is somewhat less than that estimatedfrom SDS-polyacrylamide gel electrophoresis (Jennings andPhillips, 1982),possibly due to the effect of glycosylation on protein mobility in SDS0.3polyacrylamide gels (Banker and Cotman,1972). The hydrophobicity plot for GPIIb suggests the presence of a transmembrane domain in the smaller subunit. Consistent with this interpretation is the distribution of possible N linked glycosylation sites, all of which would be present in the extracellular domain of the protein. Electronmicrographs I 2 3 of the platelet GPIIb. IIIa heterodimer complex have sugFIG.6. Northern blot analysis of RNA from HEL and K562 gested that thecomplex forms a signet ring-like structure on cells. Total RNA from HEL cells, from K562 cells stimulated with lo-” M PMA, and from unstimulatedK562 cells was denatured in the the outer membrane surface (Carrel1 et al., 1985; Parise and presence of glyoxal and electrophoresed on 1% agarose gels. RNA Phillips, 1985). Based on these electron micrographs, it was two hydrophobic domains that from K562 cells was used as a control in this analysisbecause PMA- proposed that GPIIIa contains stimulated K562 cells express platelet GPIIIa, but not GPIIb-like position the structurein the membrane and that GPIIb forms proteins (Silveret al., 1987). Following electrophoresis, the RNA was the signet portion of the structure without contact with the transferred tonylon paper and hybridized to the nick-translated3.2- membrane. However, the amino acid sequence of a GPIIIaLune kb GPIIbcDNA as described in the “Experimental Procedures.” like protein from chick fibroblasts, integrin, contains a single 1 , 1 3 pg of HEL RNA; lane 2,13 pg of RNA from unstimulated K562 cells; lane 3, 13 pg of RNA from PMA-stimulated K562 cells. Com- hydrophobic core near the carboxyl terminus of the protein mercial RNA size markers (BRL)were electrophoresed concurrently (Tamkun et al., 1986), as doesa GPIIIa-like protein from and used to calculate RNA size. human endothelial cells (Fitzgerald et al., 1987). If human platelet GPIIIa also contains a single transmembrane domain, sequence divergence in the first andsecond coding positions. the interpretationof these electron micrographic images must Northern Blot Analysis of HEL mRNA-Northernblot be reexamined in light of the presence of possible membraneanalysis was performed using total RNA from HEL and the spanning domains in both members of the GPIIb-IIIa comradiolabeled 3.2-kb GPIIb cDNA. K562 cell RNA was used as plex. A hydrophobic stretch of amino acids also precedes the a negative control because these cells do not containa GPIIb- amino terminus of GPIIb and likely represents a signal peplike protein (Silver et al., 1987). A single 4.1-kb species of tide. Because this sequence begins with a methionine, transmRNA was present in the HEL cell blot, whereas no bands lation of the GPIIb mRNA may be initiated at this point, were detected in the blots usingK562 RNA (Fig. 6). suggesting that our DNA sequence encodes the entire GPIIb precursor. DISCUSSION The integrityof the platelet membrane GPIIb. IIIa heterodimer complex requires the presence of submicromolar conWe have isolated and determined the DNA sequence for human platelet GPIIb using cDNA obtained from an HEL centrations of ionic calcium (Kunicki et al.,1981; Brass et al., cell expression library. Confirmation that these clones repre- 1985). The basis for this requirement is not clear, although sent platelet GPIIb cDNA isbased on two lines of evidence. Fujimura and Phillips have shown that calcium ions protect First, thefusion proteins produced by E. coli containing these platelet GPIIb from cleavage by thrombin (1983), that the insertsreactwith polyclonal and monoclonal anti-human migration of GPIIb in SDS-polyacrylamide gels increases in the presence of calcium (1983b),and thatradioactive calcium platelet GPIIb antibodies. Second, the open reading frame (1983b). We found four encodes previously determined aminoacid sequences forboth binds to GPIIb, but not to GPIIIa stretches of 12 amino acids near the midportion of the large the larger and smaller subunitsof human platelet GPIIb. We calculated a molecularweight of approximately 110,100 subunit of GPIIb that are similar to calcium-binding sites andtroponin C. Thesestretchesare for the polypeptide encoded by the open reading frameof the foundincalmodulin of GPIIb that are likely to be extracellular mature GPIIbcDNA, a value similar to that of the polypeptide present on portions would permit calcium ion precursor for GPIIb directly translated in vitro using HEL and that are in conformations that cell mRNA (Silver et al., 1987). This suggests that the3.3-kb binding. Thus, it is possible that calcium ions bound to these conformation of GPIIb such that cDNA sequence contains the entirecoding region for GPIIb. sites on GPIIb maintain the it interacts with complementary portions of GPIIIa to form Northern blot analysis, however, indicates a GPIIb mRNA size of approximately 4.1 kb, indicating that the full-length the GPIIb. IIIaheterodimer.

kb

-- * -

-

Structure of Xycoprotein IIb

8482

The amino acid sequence of GPIIb is strikingly similar to the limited sequence information available for the (Y chains of the VN and the FN receptors. For example, 74% of the amino acids in the VN receptor sequence are either identical or represent conservative amino acid substitutions, and virtually all of the cysteine residues appear to occupy homologous positions in the molecules. The hydrophobic profiles also reveal significant homology. Both have a hydrophobic domain near their carboxyl termini followed by a short hydrophilic stretch of amino acids. In thedirection of the amino terminus there is additional similarity, although the two profiles are not identical. This degree of similarity between GPIIb and the 01 chains of the VN and FNreceptors suggests that these proteins divergedfrom a common ancestor andthatthe divergence of all three proteins occurred at approximately the same time. However, there are anumber of differences between GPIIb and the (Y chains of these receptors. Potential N-linked glycosylation sites differ in both number and position, and the larger GPIIb subunit is richer in proline residues (12 out of 44 residues) in its carboxyl terminal region. Because proline residues are thought to inhibit the formation of (Y helical regions, this area is likely to assume a unique structure of GPIIb. It has been postulated that theplatelet membrane GPIIb . IIIa complex is a member of a family of membrane adhesion receptors. The DNA sequence determination of cDNA coding for the full-length mature GPIIb has permitted acomparison of GPIIb with other members of this putative family and has provided further evidence for this hypothesis at both the DNA and protein level. However, the specificity of each of these receptors is unique. Thus, whereas the VN receptor and FN receptors interact exclusively with their respective ligands, the GPIIb-IIIacomplex interacts with each of these ligands plus fibrinogen and von Willebrand factor (Pytela et al., 1986). With the information provided by the sequences for these receptors, the task nowwillbe to relate the sequences to function andit reconcile the structuralsimilarities and differences among the receptors with their unique specificities. Acknowledgments-The data bases used for this manuscript were the most recent (January, 1987) versions of the National Institutes of Health DNA sequence library (GenBank) and the National Biomedical Research Foundation's protein sequence database (NBRF) published on the BionetTMsystem. Computer resourcesused to conduct our studies were provided in part by the National Institutes of Health sponsored BionetTMNational Computer Resource for Molecular Biology (NIH 1U41 RR-01685). REFERENCES Argraves, W. S., Pytela, R., Suzuki, S., Millan, J. L., Pienchbacher, M. D., and Ruoslahti, E. (1986)J. Biol. Chem. 261,12922-12924 Aviv, H., and Leder, P. (1972)Proc. Natl. A c d . Scr. U. S. A. 69, 1408-1412 Banker, G. A,, and Cotman, C. W. (1972)J. Biol. Chem. 247,5856-5861 Beller, D. I., Springer, T. A,, and Schreiber, R. D. (1982)J. Exp. Med. 156, 1000-1009 Bennett, J. S., and Vilaire, G. (1979)J. Clin. Inuest. 64. 1393-1401

Bennett, J. S., Vilaire, G., and Cines, D. B. (1982)J. Biol. Chem. 257,80498054 Brass, L. F., Shattil, S. J., Kunicki, T. J., and Bennett, J. S. (1985)J. Biol. Chem. 260,7875-7881 Bray, P. F., Rosa, J.-P., Lingappa, V. R., Kan, Y.W. McEver, R. P., and Shuman, M. A. (1986)Proc. Natl. Acnd. Sci. U. S. A. 83, 1480-1484 Carrell, N. A,, Fitzgerald, L. A., Steiner, B., Erikson, H. P., and Phillips, D. R. (1985)J. Biol. Chem. 260.1743-1749 Charo, I. F., Fitzgerald, L. A., Steiner, B., Rall, S. C. Bekeart, L. S., and Phillips, D. R. (1986)Proc. Natl. Acnd. Sci. U.S. A. 83,8351-8355 Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutter, W. J. (1979) Biochemistry 18,5294-5299 Chou, P. Y., and Fasman, G. D. (1978)Adu. Enzyrnol. 47,25-148 Davignon, D., Martz, E.,Reynolds, T., Kurzinger, K., and Springer,T.A. (1981) J. Immuwl. 127,590-595 Fitzgerald, L. A., Steiner, B., Rall, S. C., Lo, S.-S., and Phillips, D. R. (1978)J. Biol. Chem. 262,3936-3939 Fujimura, K., and Phillips, D. R. (1983a)J. Biol. Chem. 258, 10247-10252 Fujimura, K., and Phillips, D. R. (1983b)Thromb. Huemostasis 50,251a Gartner, T. K., and Bennett, J. S. (1985)J. Biol. Chem. 260, 11891-11894 Ginsberg, M. H., Forsyth, J., Lightsey, A., Chediak, J., and Plow, E. F. (1983) J . Clin. fnuest. 71, 619-624 Ginsberg, M. H., Pierschbacher, M. D., Ruoslahti, E., Marguerie, G., and Plow, E. (1985)J. Biol. Chem. 260,3931-3936 Govindan, M. V., Devic, M., Green, S., Gronemeyer, H., and Chambon, P. (1985)Nucleic Acids Res. 13,82933304 Gubler, U., and Hoffman,B. J. (1983)Gene (Amst.) 25,263-269 Haverstick, D. M., Cowan, J. F., Yamada, K. M., and Santoro, S. A. (1985) Blood 66,946-952 Hiraiwa, A,, Matsukage, A., Shiku, H., Takahashi, T., Naito, K., andYamada, K. (1987) Rhod 69.560-S64 .~~~ ,~ . ~ ~ ~ . . Hynes, R. 0. (1987)Cell 4-8;549-554 Jennings, L. K., and Phillips, D1. R. (1982)J. Biol. Chem. 257,10458-10466 Kretsinger, R. H. (1976)A,nnu. Rev. Biochem. 45.239-266 Kunicki, T. J., Pidard, D., Rosa, J.-P., and Nurden, A. T.(1981)B M 58, 268-278 K3el J., andDoolittle, R. F. (1982)J. Mol. Biol. 157, 105-132 Laemmli, U.K . (1970)Nature 227,680-685 Lipman, D. J., and Pearson, W. R. (1985)Science 227, 1435-1441 Maniatis, T.,Fritsch, E. F., and Sambrook, J. (1982)Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, ""

NV

M&&erie, G. A., Plow, E. F., and Edgington, T. S. (1979)J. BioL Chem. 254, 5357-5363 Martin, P., and Papayanno oulou, T. (1982)Nature 216, 1233-1235 McEver, R. P., Baenziger, U., and Majerus, P. W. (1982)Blood 59,8045 Messing, J. (1983)Methods Enzymol. 101, 20-79 Needleman, S. B., and Wunsch. C. D. (1970)J. Mol. Biol. 48,443-453 Parise, L. V., and Phillips, D. R. (1985)J. Biol. Chem. 260, 1750-1756 Perlman, D.,and Halvorson, H. 0. (1983)J. Mol. Bid. 167,391-409 Phillips, D. R., and Agin, P. P. (1977)J . Clin. Inuest. 60,535-545 Pierschbacher, M. D., and Ruoslahti, E.(1984)Nature 309,30-33 Poncz, M., Solowiejczyk,D., Ballantine, M., Schwartz, E., and Surrey,S. (1982) Proc. Natl. Acnd. Set. U. S. A. 79, 4298-4302 Poncz, M., Surrey, S., LaRocco, P., Weiss, M. J., Rappaport,E. F., Conway, T. M., and Schwartz, E. (1987)Blood 69,219-223 Pytela, R., Pierschbacher, M. D., Ginsberg, M. H., Plow, E. F., and Ruoslahti, E. (1986)Science 231,1159-1162 Pytela, R., Pierschbacher, M. D., and Ruoslahti, E. (1985a)Proc. Natl. A c d . Sci. U. S. A. 82,5766-5770 Pytela, R., Pierschbacher, M. D., and Ruoslahti, E. (1985b)Cell 40, 191-198 Rosa, J.-P., Cevallos M., and McEver R. P. (1986)Blood 68.Suppl. 1,325a Ruggeri, Z. M., Baddr, R., and DeMarho, L. (1982)Proc. Natl. A c d . Sci. U. S. A. 79, 6038-6041 Sanger, F., Nicklen, S., and Coulson, A. A. (1977)Proc. Natl. Acnd. Sci. U. S. A. 74,5463-5468 Silver, S. M., McDonough, M. M.. Vilaire, G., and Bennett, J. S. (1987)Blood 69,1031-1037 Springer, T. A,, Teplow, D. B., and Dreyer, W. J. (1985)Nature 314,540-542 Suzuki, S., Argraves, W. S., P la, R., Aral, H., Krusms, T., Pierschbacher, M. S. A. 83,8614-8618 D., and Ruoslahti, E. (1986$roc. Natl. A c d . Sci. Tabilio, A,, Rosa J.-P., Testa, U., Kieffer, N., Nurden, A. T., Del Canizo, M. C., Breton-Go&, J., and Vainchenker,W. (1984)EMBO J.3,453-459 Tamkun, J. W., DeSimone, D. W., Fonda, D., Patel, R. S., Buck, C., Horowitz, A. F., and Hynes, R. 0. (1986)Cell 46,271-282 Thomas, P. S. (1983)Methods Enzymol. 100, 255-266 Williams, R. J. P. (1986)in Calcium and the Cell (Evered, D., and Whelan, J., eds) Ciba Foundation Symposium 122, pp. 145-161, John Wiley & Sons, Chichest,er, UK Yanisch-Perron, C., Vieira, J., and Messing, J. (1985)Gene (Amst.) 33, 103119

!.

k