Glycoproteins V and Ib-IX Form a Noncovalent Platelet Membrane ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistryand Molecular Biology, Inc.

Vol. 267, No. 1, Issue of January 5, pp. 364-369,1992 Printed in U.S.A.

Glycoproteins V and Ib-IX Form a Noncovalent Complex in the Platelet Membrane* (Received for publication, June 7, 1991)

Piet W. ModdermanS, Lucia G. Admiraal, Arnoud Sonnenberg, and Albert E. G. Kr. von dem Borne From the Department of Immunohematology, Central Laboratoryof the Netherlands Red Cross Blood Transfusion Service and Laboratory for Experimental and Clinical Immunology, University of Amsterdam, Plesmanlaan 125, 1066 CX Amsterdam, the Netherlands and the Department of Haematology, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherhnds ”

Platelet glycoprotein (GP) V is a M, 82,000 plasma 145,000 a chain disulfide-linked to a M, 24,000 /3 chain) and membrane protein of unknown function that iscleaved GPIX (Mr 22,000) (Ref. 3). The binding site for vWF has by the potent platelet agonist, thrombin, to yield a M, been localized on the M , 45,000 amino-terminal globular 69,600 fragment (GPVtl). Platelet GPIb, a disulfide- domain of the GPIba chain (4, 5). GPIba also has binding linked a@ heterodimer (Mr 160,000) that forms a non- sites for the platelet agonist, thrombin (6), but isit not cleaved covalent complex with GPIX (Mr22,000), functions as by the enzyme. Recent studies show that platelet activation the platelet adhesion receptor for surface-bound von may be accompanied by translocation of GPIb-IX to internal Willebrand factor. Association between GPV and platelet membrane systems (7-9). GPIb-IX has been suggested by the finding that both Platelet GPV is a M, 82,000 membrane protein that is proteins are deficient intheBernard-Souliersyndrome, a bleeding disorder characterized by giant cleaved by thrombin to yield a M, 69,500 fragment (GPVn) platelets and defective interaction withvon Willebrand (10, 11).GPV may also be released as a M , 80,000 fragment factor. Here we report that GPV and GPIb-IX are (GPVs) after cleavage by the platelet calcium-dependent procoprecipitated by monoclonal antibodies (mAbs) tease, calpain (4). Since GPV is the only major platelet against GPV, GPIb, or GPIX when platelets are solu- membrane protein that is a substrate for thrombin, it has bilized in the mild detergent, digitonin. Treatment of been thought to be the thrombin activation receptor. Howdigitonin immunopreciptates with the nonionic deter- ever, there are several arguments against this, e.g. Bernardgent, Nonidet P-40, released GPV from anti-GPIb and Soulier syndrome platelets, which are genetically deficient in anti-GPIX mAb precipitates and GPIb-IX from the GPV, can be activated by thrombin (12), and rabbit anti-GPV anti-GPV mAb precipitate. Removal of the M, 45,000 antibodies that inhibit the conversion of GPV to GPVfl do amino-terminal part of GPIba by treatment with elas- not affect platelet activation by thrombin (13). Moreover, a tase did not abrogate association of GPV with GPIb- thrombin receptor present in platelets, endothelial cells, and IX, showing that the leucine-rich repeat sequences in other cell types has recently been cloned and appears to be GPIba are not required for complex formation. Bind- distinct from GPV (14). ing studies with 1261-labeled mAbs showed the presence GPIb-IX and GPV share several characteristics. 1) GPIb, of 24,370 GPIb-IX complexesand 11,170 molecules of GPIX, and GPV are all absent in the Bernard-Soulier synGPV/platelet (n = 5). These data show that theleucine- drome (15, 16),a bleeding disorder characterized by giant rich glycoproteins GPV and GPIb-IX form a noncova- platelets, reduced ristocetin-dependent binding of vWF, and lent complex in theplatelet membrane. GPV may play reduced adhesion to subendothelium and vWF. 2) cDNA a role in the interaction of platelets with von Wille- sequences of GPIba, GPIbB, and GPIX (17-19) and partial brand factor. amino acid sequences of GPV (20, 21) show that all four proteins belong to the newly described family of leucine-rich proteins. The similarities between GPIb-IX and GPV prompted US GPIb’ is the platelet receptor that mediates the blood shearto investigate the possibility of a physical association between dependent adhesion of unstimulated platelets to von Willebrand factor (vWF) inthe subendothelium of damaged vessel them. To this end, immunoprecipitation experiments were walls (1).The antibiotic ristocetin induces binding of soluble performed with anti-GPIb-IX monoclonal antibodies (mAbs), vWF to GPIb in the absence of bloodflow (2). Platelets as well as with a newly developed mAb against GPV. For contain 20,000-25,000 copies of GPIb-IX, which is a 1:l these studies, plateletswere solubilized in the nonionic deternoncovalent complex between GPIb (consisting of a M , gent digitonin, which lyses cells without disrupting noncovalently associated complexes of membrane proteins, e.g. digi* This work was supported by Centocor, Inc., Malvern, PA. The tonin has been used to study the composition of murine T costs of publication of this article were defrayed in part by the and B cell antigen receptor complexes (22, 23). Our results payment of page charges. This articlemusttherefore be hereby show that GPV and GPIb-IX are coprecipitated from digimarked “advertisement” in accordance with 18 U.S.C. Section 1734 tonin-lysed platelets, indicating that these proteins form a solely to indicate this fact. noncovalent complex in the platelet membrane. $ To whom correspondence should be addressed c / o Publication Secretariat, CentralLaboratory of the Netherlands Red Cross Blood Transfusion Service, P. 0. Box 9406, 1006 AK Amsterdam, The Netherlands. The abbreviations used are: GP, glycoprotein; vWF, von Willebrand factor; mAb, monoclonal antibody; SDS, sodium dodecyl sulfate.

EXPERIMENTALPROCEDURES

Materials-All reagents were obtained from Sigma unless otherwise specified. The megakaryoblastic leukemic cell line MEG-01 was obtained from The Institute for Fermentation, Osaka, Japan. Human leukocyte elastase was obtained from Elastin Products Company,

364

A Complex of Human PlateletGPV and GPIb-IX Inc., Owensville, MO.Reagents for polyacrylamide gel electrophoresis were purchased from Bio-Rad, and Nalz5I and [3H]NaBH4 were obtained from Amersham (Radiochemical Center, UK). MonoclonalAntibodies-SW16 is anIgGl mAb obtained after immunization of a female BALB/c mouse with human platelets, using a described strategy (24). Immunofluoresence testing of the binding of SW16 to peripheral blood cells and cell line cells was performed as described (24). The anti-83 integrin (CD61) mAbC17 binds to a complex-dependent epitope on plateletGPIIb-IIIa (24, 25). AntiGPIV (CD36) mAb IVC7 and anti-GPIbachain (CD42b) mAb MB45 have been described previously (26). MB45 is directed against an epitope on the M, 45,000 amino-terminal domain of GPIba, because 1)MB45 completely inhibits ristocetin-induced agglutination, 2) the MB45 epitope is lost after treatment of platelets with elastase (261, and 3) MB45 immunoprecipitates a M,45,000 polypeptide from the supernatant of elastase-treatedplatelets (data not shown). AntiGPIX (CD42a) mAb FMC 25 was kindly donated by Dr H. Zola, Flinders Medical Centre, Adelaide, Australia. MAb SAM-1 isdirected a t the a581 (VLA-5) integrin (Dr. C. Figdor, Netherlands Cancer Institute, Amsterdam, the Netherlands). SW16 and MB45 were purified by protein G-Sepharose affinity chromatography according to the method suggested by the manufacturer (Pharmacia Fine Chemicals AB, Uppsala, Sweden) and '%I-labeled by the IODO-GEN method, as described (25). Preparation, Surface Labeling, and Stimulation of Platelets-Platelets were isolated by a modification of the method of Legrand et al. (27). Acid/citrate/dextrose platelet-rich plasma was obtained as described (27), and prostaglandin E, a t 100 nM was added to prevent platelet activation. The platelets were pelleted by centrifugation for 10 min a t 1200 X g and washed three times in a buffer containing 36 mM citric acid, 103 mM NaCl, 5 mM KCl, 5 mM EDTA, 5.6 mM Dglucose, 0.35% w/v bovine serum albumin (fraction V, Sigma) and 100 nM prostaglandin E, (pH 6.5). For most experiments, platelets were then resuspended in EDTA-Tyrode, containing 130 mM NaC1, 27 mMKC1, 16 mM NaHC03, 0.36 mM NaH2P04,5.6 mMD-glUCOSe, 5 mM EDTA, and 0.35% bovine serum albumin. For surface labeling, bovine serum albumin was omitted from the platelet washing buffer, and the platelets were resuspended at 2 X 109/ml in EDTA-Tyrode without bovine serum albumin, but containing 100 nM prostaglandin E,. The platelets were then labeled with 1 mCi of 1z61/109platelets by using lactoperoxidase as described (28). Labeling was stopped by washing the platelets three times, as described above. For some experiments, washed platelets were 3Hlabeled by the ~eriodate/[~H]NaBH~ method (29). For stimulation with platelet agonists, unlabeled or 1251-labeled platelets were prepared as described and resuspended at 2 X 10*/ml in Ca-Mg-Tyrode, in which EDTA is replaced by CaClz and MgC12, both at 1mM. For stimulation with thrombin, plateletswere incubated with human a-thrombin (1unit/ml) for 10 min at 37 "C, followed by incubation with hirudin (10 units/ml) for 10 min at 22 "C. For stimulation with calcium ionophore, platelets were incubated with A23187 (1 p ~ for) 10 min at 37 "C. Platelet Aggregation-For aggregation incitratedplatelet-rich plasma, 9 parts of blood were collected in 1 part of 3.8% trisodium citrate and centrifuged for 20 min at 120 x g to obtain platelet-rich plasma, which wasthen diluted to 2.5 X 10' platelets/ml with autologous platelet-poor plasma. Washed platelets for aggregation were prepared in Ca-Mg-Tyrode (2.5 X 108/ml) as described above, except that EDTA in the platelet-washing buffer was replaced by CaC12and MgC12, both at 1mM. Platelet aggregation and theeffect of antibodies were monitored as described (24). Native collagen fibrils from equine tendon for platelet aggregation were purchased from Hormon Chemie (Munchen, Germany). Ristocetin was from Bio/Data Co., Hatboro, PA. Immunoprecipitation-Washed 1251-labeled platelets were lysed for 60 min at 0 "Cin a buffer containing 1%(v/v) Nonidet P-40,25 mM Tris-HC1 (pH 7.5), 150 mM NaCl, 5 mM EDTA, 20 pg/ml soybean trypsin inhibitor, 2 mM benzamidine, 1 mM leupeptin, and 1 mM phenylmethylsulfonyl fluoride. After centrifugation of the lysates (30 min at 13,000 X g at 4 'C), the supernatantswere incubated for 16 h a t 4 "C with mAbs and protein G-Sepharose (Pharmacia). The Sepharose beads were washed with 1%Nonidet P-40, 25 mM Tris-HC1 (pH 7.5), 150 mM NaC1, 5 mM EDTA at 4 "C. For immunoprecipitation from digitonin lysates, Nonidet P-40 was replaced by 1%(w/v) digitonin (Merck, Darmstad, Germany) in the lysis buffer and by 0.1% digitonin in the immunoprecipitation washing buffer. Nonidet P-40 elution of coprecipitated antigens from digitonin immunoprecipitates was performed by incubating the protein G-Sepharose beads in

365

200 pl of Nonidet P-40 lysis buffer for 2 h at 4 "C. The beads were washed once with 1 ml of Nonidet P-40 lysis buffer before analysis of bound proteins by SDS gel electrophoresis (see below). The eluted antigens were precipitated with trichloroacetic acid, washed in cold ethanol, dried, and prepared for SDS gel electrophoresis (see below). For immunoprecipitation of released platelet proteins, 1251-labeled platelets were stimulated as described above and pelleted by centrifugation (3 min a t 13,000 X g). Thesupernatants were centrifuged two more times and used for immunoprecipitation after addition of Nonidet P-40 and protease inhibitors to the concentrations described above. For some experiments, only protease inhibitors were added and Nonidet P-40 was omitted from the immunoprecipitation washing buffer or replaced by digitonin. For immunoprecipitation of proteins from stimulated platelets, the platelet pellets (see above) were washed twice before immunoprecipitation was performed as described. Immune complexes were eluted from Sepharose by heating at 95 "C inSDS sample buffer for 5 min (nonreduced samples, containing 10 mM N-ethylmaleimide) or for 10 min (reduction with 2% 2-mercaptoethanol) and analyzed by electrophoresis on SDSpolyacrylamide gels. 1251-Monoclonal Antibody Binding Assay-Platelets were prepared in EDTA-Tyrode as described above (1X 10s/ml, final concentration) and incubated in 200-4 aliquots for 90 min at room temperature with 1251-labeled mAb at theindicated concentrations. Bound radioactivity was then measured after centrifugation of the platelets (in triplicate) through 20% (w/v) sucrose in EDTA-Tyrode (25). Binding data are expressed as specific binding obtained after subtracting from total binding the nonspecific binding observed in the presence of a 50-fold excess of unlabeled antibody. RESULTS

Characterization of the Anti-GPVAntibody, SW16-Monoclonal antibody SW16 was obtained after immunizationof a BALB/c mouse with human platelets. In immunofluorescence testing, SW16 boundto both unfixedand paraformaldehydefixed human platelets. Flow cytometry analysis did not reveal any subpopulations with respect to SW16 binding. SW16 did not bind to red cells, lymphocytes, granulocytes, monocytes, or t o the cells cultured human umbilical vein endothelial cells of a n y of several human cell lines known to express platelet antigens, including the erythroleukemic cell lines HEL and K562 and the megakaryoblastic cell line MEG-01. Purified SW16 at 10 pg/ml (a saturating concentration,see below) did notinfluenceplateletaggregationincitratedplatelet-rich

plasma induced by 20 p~ ADP, 10 PM adrenalin, or 5 pg/ml collagen and also did not affect the aggregation of washed platelets stimulated with 0.1 units/ml thrombin. Furthermore, i t did not inhibit the von Willebrand factor-mediated agglutination of fixed platelets induced by1.25 mg/ml of ristocetin in the presenceof human plasma (data not shown). From Nonidet P-40-lysed, surface-iodinated platelets, SW16 immunoprecipitated a labeled protein with a mobility i n SDS gel electrophoresis (Fig. 1,A and B, lunes 2) that was higher than that of glycoproteins GPIIIa and GPIV (lanes 1 and 5 , respectively). Its mobility relative to these proteins, both underreducing and nonreducingconditions,corresponded to that of GPV (Mr78,000 i n lanes 2). To investigate whether the protein recognized by SW16 represented GPV, surface-labeled platelets were treated with thrombin, which is known to cleave GPV, yielding a M , 69,500 fragment,GPVn (10, 11). From the supernatant of such platelets, SW16 immunoprecipitatedtwopolypeptides,onecorrespondingto GPVfl (Mr64,000 i n Fig. 1, lanes 4 ) and the other one corresponding toGPV, (Mr76,000 i n Fig. 1, lanes 4 ) . The antibody GPV from the thrombinprecipitated only residual intact treated platelets (Fig. 1, lanes 3 ) . To confirm that t h e M, 76,000 polypeptide is identical withGPV,, 1251-plateletsi n CaMg-Tyrodewerestimulatedwiththecalciumionophore, A23187, a treatment that induces the release of GPV, via hydrolysis by the platelet calcium-dependent protease, calpain

A Complex of Human Platelet GPV and GPIb-IX

366 A

Ilb-

Illa, lvV

i

"S

Vll-

-

e-

VI1

1

2

3

4

5

FIG. 1. Immunoprecipitation of GPV and GPV-relatedpoly-

peptides by monoclonal antibody SW16. Washed human plate-

1 2 3 lets were surface-labeled with ''''I and either lysed directly in Nonidet FIG. 2. Immunoprecipitation of CPV. from the supernatant 1'-40 lysis buffer (lanes 1,2. and 5 ) or resuspended in Ca-Mg-Tyrode, of calcium ionophore-treatedplatelets. "'I-Laheled platelets treated with thromhin/hirudin, pelleted, and washed as described were either lysed directly or resuspended in Ca-Mg-Tyrodeand under "Experimental Procedures." The stimulated platelets were then stimulated with the calcium ionophore, A23187. or with thrombin/ lysed (lane 3 ) , Nonidet P-40 and protease inhibitors also added to hirudin as described under "Experimental Procedures." The control the platelet supernatants(lanes 4 ) , and both lysates and supernatantsplatelets and the supernatants of stimulated plateletswere treated as then used for immunoprecipitation. For comparison, immunoprecip- described in the legend to Fig. 1 and used for immunoprecipitation itates of major platelet membrane proteins with electrophoretic mo- with SW16. Lune 1, lysates of untreated platelets; /anP 2, supernatant hilities similar to that of GPV are also shown. Immunoprecipitation of A28187-treated platelets; lane -7, supernatant of thrombin-treated was performed with the following mAbs: anti- GPIIh-IIIa mAb C17 platelets. The sampleswere electrophoresed in the ahsence of reduc(lane I ) , anti-CPV mAb SW16 (lanes 2 4 ,and anti-CPIVmAh IVC7 ing agent ona 7.5% SDS-polyacrylamide gel. Molecular weight mark(lane fj). Samples were electrophoresed in the absence (panel A ) or ers are indicated asin the legend to Fig. 1. presence (panel R ) of reducing agent on 7.5% SDS-polyacrylamide gels. Molecular weight markers indicate M,200,000, 116,250, 97,400, 66,200, and 42,670.

(30). Fig. 2 (lane 2 ) shows that SW16 immunoprecipitated GPV. but no GPVn or intact GPV from the supernatant of ionophore-activated platelets. Analysis after surface labeling by the periodate/['H]NaBH, method showed that the proteinrecognized by SW16 had the same mobility as the 9-labeled antigen and appeared tobe strongly labeled by this method (data not shown), which is consistent with thecharacteristics of GPV (11, 12). The antibody did not recognize any platelet proteins in nonreduced or reduced immunoblotting.Fromthecombineddata, we conclude that SW16 is directed against a conformation-dependent epitope on platelet GPV that is also present on its proteolytic fragments, GPV. and GPVn. Surface Expression of Platelet GPV-To quantifythe amount of GPV present on the platelet surface, binding of "'1-labeled SW16 to resting and thrombin-stimulated platelets wasmeasured. SW16 binding to resting platelets was time- and concentration-dependent and, as shown for a representative experiment in Fig. 3, was saturated at a concentration of 1-2 pg of SW16/108 platelets. Scatchard analysis (31) revealed the presence of 12,100 binding sites/platelet with a Kd = 0.64 nM (Fig. 3, inset). For the platelets from four donors, 13,860 & 1,240 SW16 binding sites were found (mean f S.D.), with a mean Kd = 0.57 nM. When the same platelets were treated with 1 unit/ml of thrombin (10 min a t 37 "C) before addition of SW16, antibody binding was greatly reduced (Fig. 3). The mean number of sites S.D. recognized by SW16 on these plateletswas 590 270 ( n= 4), a reduction of 96%. Flow cytometry analysisof thrombin-treated platelets did not show the presenceof subpopulations of platelets with respect to SW16 binding (data not shown). Go-precipitation of GPV and GPIb-IX from Digitonin-lysed

*

*

cn 5

-A-A-A-A-A-

0 0

0.5

1

.o

1.5

2.0 4

SW 16 added (ug/ml) FIG. 3. 1zDI-SW16 binding to resting and thrombin-stimu-

lated platelets. Washedplatelets ( 1 X 10"/ml) in Ca-Mg-Tyrode were eitherincubated in thepresence of theplateletantagonist, prostaglandin E,, a t 100 nM ( 0 )or stimulated with thrombin ( A ) , as described under "Experimental Procedures." The platelets were then incubated with various concentrations of ""I-SW16 for 90 min at 22 "C. Roundreactivity wasmeasured aftercentrifugation of the platelets through 20% sucrose. The i n s d shows the Scatchard plot of SW16 binding to unstimulated platelets. Platelets-The absence of both CPIb-IX and GPVfrom platelets of patients with the Bernard-Soulier syndrome (15, 16) suggested that the expression of GPIb-IX might require the presence of GPV, andvice versa, and that CPIh-IX and GPV might bephysicallylinked in theplatelet membrane. To investigate whether such an association exists, immunoprecipitations were performed with""I-platelets lysed in digitonin, a mild nonionicdetergent.Undertheseconditions,

A Complex of Human Platelet SW16 immunoprecipitated GPV, as well as a protein with a mobility similar to that of GPIb (Fig. 4, lane 2). Conversely, mAbs against GPIb (MB45) and against GPIX (FMC 25) coprecipitated a protein with a mobility similar to that of GPV in addition to GPIb-IX (lanes 4 and 6, respectively). Only GPV was observed in Nonidet P-40 immunoprecipitates of SW16 (lane 1) and only GPIb and GPIX, butno GPV, in Nonidet P-40 precipitates of MB45 (lane 3) and FMC 25 (lane 5). Similar results were observed when platelets were solubilized inthe nonionic detergent octyl glucoside (not shown). GPV and GPIb-IX were not precipitated from digitonin lysates by mAb SAM-1 against the a5pl (VLA-5) integrin (lane 7) or by C17 against GPIIb-IIIa (lane 8 ) . As shown by the precipitation of glycocalicin bythe mAbs against GPIb and GPIX (lanes 3-6), some proteolysis of the GPIba chain had apparently occurred during the immunoprecipitation procedure. Also, variable amounts of a labeled polypeptide with an electrophoretic mobility slightly higher than that of GPV were occasionally precipitated by FMC 25 (lane 5 ) and, to a lesser extent, by MB45 from Nonidet P-40 lysates. Comparison with immunoprecipitated GPV fragments showed that the polypeptide was neither GPV. nor GPV" (not shown). These results suggested that GPV and GPIb-IX form a noncovalent complex that can be dissociated by Nonidet P40. To identify the coprecipitated proteins, Sepharose-bound digitonin precipitates of SW16, MB45, and FMC 25 (Fig. 5, lanes 1,6, and11) were incubated in NonidetP-40 lysis buffer for 2 h at 4 "C. Analysis of the proteins remaining bound to the Sepharose beads (lanes 2, 7, and 12) and of the eluted proteins (after trichloroacetic acid precipitation, lanes 3, 8, and 13) shows that a polypeptide with the mobility of GPIb was eluted from the anti-GPV mAb precipitates, whereas a GPV-like polypeptide was preferentially eluted from the antiGPIb and anti-GPIX mAbs. Immunoprecipitation with MB45 (lane 5) and with SW16 (lanes 10 and 15) from the Nonidet P-40 eluates confirmed the identity of the eluted molecules as GPIb and GPV, respectively. Neither of these molecules was precipitated by a control mAb (10Gll) against a2pl. Association of Proteolytic Fragments of GPV and GPIb-To investigate which parts of GPV and GPIb-IX areinvolved in

GPV and GPIb-IX

367

-

GPlb-

.c

GCGPV-

~

em

. *

* a *

-

GPIX4321

5

6 7 8 9 10

-

111213 15 14

FIG.5. Noncovalently bound GPV and GPIb-IX are identified by immunoprecipitation after dissociation of their complex by Nonidet P-40.""I-Labeledplatelets were lysed in digitonin andused for immunoprecipitationwithSepharose-bound mAbs against GPV (SW16, lanes 1-5), GPIbcv (MB45, lanes 6-10), and GPIX (FMC 25, lanes 12-15). Aliquots of the precipitates were analyzeddirectly (lanes I, 6, and 11), whereas the remainderwas incubated inNonidet P-40 lysis buffer for 2 hat 4 "C.Aliquots of the proteins remaining bound to Sepharose (lanes 2, 7, and 12) and of the proteins eluted by Nonidet P-40 and collected by trichloroacetic acid precipitation (lanes 3,8, and 13) were also analyzed. Aliquots of the remainder of the Nonidet P-40 eluate were used for immunoprecipitation with anti-a2pl integrin mAb lOGll (lanes 4, 9, and 1 4 ) , with MB45 (lane 5),and with SW16 (lanes 10 and 15). The proteins were analyzed under nonreducing conditions on a 5-15% SDS polyacrylamide gel. Molecular weight markers and abbreviations are as indicated in the legend to Fig. 4.

complex formation, '2sI-platelets were subjected to proteolytic treatments, and theassociation between the digested components of GPIb-IX andGPV was analyzed. Platelets were first incubated with the calcium ionophore, A23187, a treatment resulting in thecalpain-mediated release of GPV, and theM, 130,000 fragment of the GPIb, chain, glycocalicin (30). When the supernatant of A23187-treated platelets was used for immunoprecipitation, SW16 immunoprecitated only GPV, (Fig. 6, lane 4 ) and MB45 precipitated only glycocalicin ( l a n e 5). FMC 25 did notreact with anyproteinunderthese conditions, asGPIX is not released by calpain (lane 6). . .. ".-. -. . Similar results were observed after addition of digitonin to A23187 supernatants (lanes 1-3). These resultsshow that the protein domainsresponsible for the association between GPV GPlband GPIb-IX are either not contained in the glycocalicin and GPV, moieties of GPIba andGPV, respectively, or that their GCbinding properties were changed after release by calpain. Platelets were also treated with various concentrations of GPVhuman leukocyte elastase, which releases the M,45,000 amino terminus of GPIba, leaving the MI 100,000 membrane-bound fragment disulfide-linked to GPIbp(4). This M,125,000GPIb remnant remains complexed to GPIX (see below). Immunoprecipitation with MB45 showed that theM,45,000 fragment was present in the supernatantof platelets treated with elasGPIX"" tase for 5 min a t 37 "C at concentrations of 0.1 pg/ml and 1 2 3 4 5 6 7 8 higher. GPV has also been shown to be a substrate for elastase FIG.4. Immunoprecipitation of a complex ofGPV and (4), but an increase of the SW16-precipitated amount of GPV, GPIb-IX from platelet digitonin lysates. "'I-Labeled platelets above background levels was only detectable with platelets were lysed in Nonidet P-40 lysis buffer (lanes I, 3, and 5) orin a buffer containing 1% digitonin (lanes 2,4, 6, 7, and 8 ) and then used treated with elastase at concentrations of 1 pg/ml or higher for immunoprecipitationwith anti-GPV mAb SW16 (lanes 1 and 2), (data not shown). When platelets treated with 0.1 pg/ml of anti-GPIba mAb MB45 (lanes 3 and 4 ) , anti-GPIX mAb FMC 25 elastase were solubilized in digitonin, GPIX, GPV, and both (lanes 5 and 6 ) ,anti-a5pl mAb SAM-1 (lane 7),and anti-GPIIb-IIIa intact GPIb and the MI 125,000 remnant of GPIb (running mAbC17 (lane 8). The samples were analyzed under nonreducing at MI 140,000 in this type of gel) were present in the immuconditions on a 5-15% SDS-polyacrylamide gel. Molecularweight markers indicate M,200,000, 116,250, 97,400, 66,200, 42,670, 31,000, noprecipitate of anti-GPIX mAb FMC 25(Fig.7, lune 3). and 21,500. GC, a polypeptide comigratingwith the GPIba proteolytic However, almost all of GPIb appeared to be digested after treatment with elastase at 1 pg/ml (lane 4 ) . Under these product, glycocalicin.

-

A Complex of Human Platelet

368

GPV and GPIb-IX centration of elastase used (lane 5 ) , under which conditions SW16 precipitated large amounts of a GPVn-like GPV fragment from the platelet supernatant (notshown). Comparison of Anti-GPV and Anti-GPIb mAb Binding to Platelets-To compare the number of GPV molecules and the number of GPIb-IX complexes/platelet, binding of '251-SW16 and l2'II"B45 to platelets was measured. The platelets from five donors bound 11,170 f 590 (mean f S.D.) SW16 molecules/platelet with a mean K d = 0.55 nM and 24,370 & 1,010 MB45 molecules/platelet with a mean Kd = 0.81nM. The mean ratio of the number of GPIb and GPV molecules/ platelet was 2.18 (range 2.10-2.31).

GPlbGCGPV F GPV,

DISCUSSION 1

2

3

4

5

6

7

8

9

FIG. 6. Absence of a complex between GPV. and the GPIb fragment, glycocalicin, both in the absence and presence of digitonin. The supernatant of A23187-stimulated 12sI-labeledplatelets was used for immunoprecipitation with monoclonal antibodies, either in the absence of detergent (lanes 4-6) or after addition of digitonin to a concentration of 1% (lanes 1-3). Forcomparison, immunoprecipitates from digitonin-lysed control platelets are shown (lanes 7-9). MAbs usedfor immunoprecipitationwere anti-GPV mAb SW16 (lanes 1,4, and 7), anti-GPIb mAb MB45 (lanes 2 , 5 , and 8), and anti-GPIX mAb FMC 25 (lanes 3, 6, and 9). The samples were analyzed under nonreducing conditions on a 7.5% SDS-polyacrylamide gel. Molecular weight markers are indicated as in the legend t o Fig. 1. GC, a polypeptide comigrating with the GPIba proteolytic product, glycocalicin.

GPlb-

GPV-

GPIX1 2 3 4

5

FIG. 7. The M, 45,000 amino-terminal portion of GPIbba is not required for the association between GPV and GPIb-IX. 12sII-labeled platelets (2.5 X 10R/ml) in EDTA-Tyrodewere incubated at 37 "C for 5 min, either in the absence (lane 1 ) or in the presence of human leukocyte elastase at 0.01 pg/ml (lane 2 ) , 0.1 pg/ml (lane 3), 1 pg/ml (lane 4 ) , or 10 pg/ml ( l a w 5). Soybean trypsin inhibitor and phenylmethylsulfonyl fluoride were then added to all samples to final concentrations of 400 pg/ml and 2 mM, respectively, and digitonin lysatesof the plateletswere used for immunoprecipitation with FMC 25. The immunoprecipitateswere analyzed under nonreducing conditions on a 5-15% SDS-polyacrylamide gel. Molecularweight markers are asindicated in the legend to Fig. 4. The arrow indicates the position of the disulfide-linked complex of the M,100,000 GPIba remnant and GPIbP.

conditions, FMC 25 still precipitated an appreciable amount of GPV (lane 4). This shows that the association between GPV and GPIb-IX does not require the presence of the M , 45,000 amino-terminal fragment of GPIba, which contains the GPIba leucine-rich repeat sequences (17), as well as the binding sites for von Willebrand factor and thrombin (4, 5). GPV itself was no longer coprecipitated at the highest con-

In thisreport, the characteristics of a monoclonal antibody against human platelet GPV are described, and it is shown that GPV forms a noncovalent complex with the GPIb-IX complex. To our knowledge, SW16 represents the first monoclonal antibody against GPV. As shown by immunoprecipitation, SW16 recognizes an epitope that is present on intact GPV, on the calpain-released M , 76,000 fragment, GPV,, and on thethrombin-released M , 64,000 fragment, GPVn (Figs. 1 and 2). As thrombin was recently shown to cleave the carboxyl-terminal part of GPV, to yield GPVn (20), the SW16 epitope must be located on the M, 64,000 amino-terminal portion of GPV. GPV was complexed with platelet GPIb-IX when platelets were lysedin digitonin but not in Nonidet P-40. The complex wasrecognizedbymAbs against GPV, GPIba, and GPIX (Fig. 4) and against GPIbP (not shown) but not bymAbs against other platelet antigens, including GPIIb-IIIa and the a5pl integrin (Fig. 4), a2D1, a6D1, the CD31 antigen, and von Willebrand factor (data not shown). Addition of Nonidet P40 to digitonin-derived immunoprecipitates of the complex of GPV and GPIb-IX released all of the GPV from the antiGPIb and anti-GPIXprecipitates and nearly all of the GPIbIX from the anti-GPV mAb precipitates (Fig. 5). This further indicates that the association between GPV and GPIb-IX is sensitive to disruption by Nonidet P-40 and therefore probably not the artifactual result of platelet solubilization in digitonin. The noncovalent interaction between GPV and GPIb-IX appears to be different from that between GPIX and GPIb, since, under the experimental conditions, the former is dissociated by Nonidet P-40 whereas the latter is not (3). The nature of the association between these proteinsis not known, but it has been suggested that "leucine-rich glycoprotein" repeats, which are present inGPV and in all three chains of GPIb-IX, might be involved in protein-protein interactions (32). However,glycocalicin and GPV., which both contain such repeats, were not coprecipitated from the supernatantof A23187-stimulated platelets (Fig. 6). Furthermore,removal of the M , 45,000 amino-terminal part of GPIba, which contains all of the GPIba leucine-rich repeats (17), did not prevent coprecipitation of GPV with the complex of GPIX, GPIbD, and theM , 100,000 GPIba remnant (Fig. 7). The stoichiometry of the complex between GPV and GPIbIX remains to be determined. Our monoclonal antibody binding studies indicatethat each platelet containsapproximately 11,000 molecules of GPV and 24,000 molecules of GPIb, the latter number closely corresponding to values previously determined with anti-GPIb and anti-GPIX antibodies (3, 33). Attempts to directly determine the complexation ratio between GPV and GPIb-IX by covalent cross-linking of the members of the complex with dithiobis(succinimidy1propionate) andbis-[/3-(4-azidosalicylamido)ethyl]disulfidewere un-

A Complex of Human Pla telet GPV and GPIb-IX

369

4. Wicki, A. N., and Clemetson, K. J. (1985) Eur. J. Biochem. 153, successful, both with intact platelets andwith digitoninplate1-11 let lysates (data not shown). 5. Handa, M., Titani, K., Holland, L. Z., Roberts, J. R., and Ruggeri, The immunoprecipitation data presented in Fig. 4 suggest Z. M. (1986) J. Biol. Chem. 261,12579-12585 that, in digitonin lysates, part of GPIb-IX is not complexed 6. Harmon, J. T., and Jamieson, G.A. (1986) J. Biol. Chem. 261, with GPV, as clearly more GPIb is precipitated by the anti13224-13229 GPIb and anti-GPIX mAbs than by the anti-GPV mAb, 7. George, J. N., Pickett,E. B., Saucerman, S., McEver, R. P. Kunicki, T. J., Kieffer, N., and Newman, P. J. (1986) J . Clin. SW16. This was confirmed by experiments in which SW16 Inuest. 78, 340-348 was used for preclearing of lysates, followed by immunopre8. Michelson, A. D., and Barnard, M. R. (1987) Blood 70, 1673cipitation with MB45. Vice versa, sequentialprecipitation 1678 with MB45 and SW16 showed that digitonin lysates also 9. Hourdill6 P., Heilmann,E., Combrie, R., Winckler, J., Clemetson, contain some uncomplexed GPV (data not shown). These K. J., and Nurden, A. T. (1990) Blood 76, 1503-1513 10. Phillips, D. R., and Poh-Agin, P. (1977) Biochem. Biophys. Res. experiments do not permit conclusionswithrepect to the Commun. 75,940-947 extent of complexation of GPIb-IX and GPV in the platelet membrane, however, as it is possible that some dissociation 11. Mosher, D. F., Vaheri, A., Choate, J. J., and Gahmberg, C.G. (1979) Blood 53,437-445 of the complex also occurs in digitonin. 12. Jamieson, G. A., and Okumura, T. (1979) J. Clin. Inuest. 61, To date, the function of GPV is unknown. Although it is 861-864 the only major platelet membrane protein that is a substrate 13. Bienz, D.,Schnippering, W., and Clemetson, K. J . (1986) Blcod 68,720-725 for thrombin,most evidence contradicts a role for GPV as the platelet activation receptorfor thrombin (see introductory 14. Vu, T. H., Hung, D. T., Wheaton, V. I., and Coughlin, S. R. (1991) Cell 64, 1057-1068 remarks). Theobserved association between GPV and GPIb- 15. Clemetson, K. J., McGregor, J . L., James, E., Dechavanne, M., IX now suggests that GPV might be involved in platelet and Luscher, E. F. (1982) J. Clin. Inuest. 70,304-311 interactions with von Willebrand factor. There is some evi- 16. Berndt, M. C., Gregory, C., Chong, B. H., Zola, H., and Castaldi, P. A. (1983) Blood 62, 800-807 dence for such a function. Although neither polyclonal rabbit 17. Lopez, J. A,, Chung, D. W., Fujikawa, K., Hagen, F. S., Papayanti-GPV antibodies (14) nor anti-GPV mAb sW16 (this annopoulou, T., and Roth, G. J. (1987) Proc. Natl. Acad. Sci. study) affected ristocetin-induced platelet agglutination, one U. S. A. 84,5615-5619 report in abstractform (34) has described anti-GPV antibod- 18. Lopez, J. A., Chung, D. W., Fujikawa, K., Hagen, F. S., Davie, E. ies in the serum of a patient with the Bernard-Soulier synW., and Roth, G. J. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, drome thatdid inhibit ristocetin-induced agglutination. How2135-2139 ever, GPV is probablynot absolutely required for GPIb bind- 19. Hickey, M. J., Williams, S. A., and Roth, G. J. (1989) Proc. Natl. Acad. Sci. U. S.A. 86,6773-6177 ing to vWF, shown as by the capacity of the soluble M,45,000 T., Fujimura, K., Maehama, S., Takemoto, M., Oda, 20. Shimomura, amino-terminalfragment of GPIba to bind to vWF in a K., Fujimoto, T., Oyama, R., Suzuki, M., Ichihara-Tanaka, K., ristocetin-dependentmanner without participation of any Titani, K., and Kuramoto, A. (1990) Blood 75, 2349-2356 other component of the GPIb-IX complex (35). 21. Roth, J. R., Church, T. A., McMullen, B. A., and Williams, S. A. One possibility would be that GPVis involved in the shear(1990) Biochem. Biophys. Res. Commun. 170, 153-161 22. Oetggen, H. C., Pettey, C. L., Maloy, W. L., and Terhorst, C. dependent exposure of vWF-binding sites on the GPIba chain. (1986) Nature 320, 272-275 Normally, platelets only interact with vWFif it is bound toa J., Leclerq, L., Radbruch, A., Rajewski, K., and Reth, solidsurface and only at high blood shearrates, such as 23. Hombach, M. (1988) EMBO J. 7,3451-3456 encountered in arteries and arterioles (36). Bindingof soluble 24. Tetteroo, P. A. T., Lansdorp, P. M., Leeksma, 0. C., and Von vWF to platelet GPIb is induced by the positively charged dem Borne, A. E. G. Kr. (1983) Br. J. Haematol. 55,509-522 antibiotic ristocetin (2). There is some evidence that platelet- 25. Modderman, P. W., Van Mourik, J. A., Van Berkel, W., Cordell, J. L., Morel, M. C., Kaplan, C., Ouwehand, W. H., Huisman, vWF interactions involve the exposure of a vWF-binding site J . G.. and Von dem Borne. A. E. G. Kr. (1989) Br. J . Haematol. on GPIb that is normally cryptic, e.g. in platelet-type von 73,514-521 Willebrand disease, GPIb on affected platelets can bind nor- 26. Von dem Borne. A. E. G. Kr.. Modderman. P. W.. Admiraal. L. mal vWF in the absence of flow or ristocetin (37, 38). From G., and Nieuwenhuis, H. K. (1989)in ieukocyte Typing'IV the association between GPIb-IX and GPV, it is conceivable (Knapp, W., Darken, B., Gilks, W. R., Rieber, E. P., Schmidt, R. E., Stein, H., and Von dem Borne, A. E. G. Kr., eds) pp, that GPV might function in regulating the accessibility of 951-966, Oxford University Press, Oxford this binding site. C., Dubernard, V., and Nurden, A. T. (1985) Biochim. Although more research is required to elucidate function 27. Legrand, B i ~ p h yActa ~ . 812,802-810 and extent of the complexation between GPV and GPIb-IX, 28. Sonnenberg, A., Hogervorst, F., Osterop, A., and Veltman, F. E. as well as theprecise architecture of the complex, we conclude M. (1988) J. Biol. Chem. 263, 14030-14038 from theresults described in this report that GPV and GPIb-29. Steiner, B., Clemetson, K. J., and Luscher, E. F. (1983) Thromb. Res. 29,43-52 IX form a noncovalent complex in the platelet membrane, a finding that may be relevant in the study of the phenomena 30. McGowan, E. B.,Yeo, K.-T., and Detwiler, T. C. (1983) Arch. Biochem. Biophys. 227, 287-301 of shear-dependent plateletadhesion and of the deficiency of 31. Scatchard, G. (1949) Ann. N. Y. Acad. Sci. 51,660-672 the four membrane proteins in the Bernard-Soulier syndrome. 32. Roth, G. J. (1991) Blood 77, 5-19 Acknowledgment-We thank Dr. Heddy Zola for his generous gift of monoclonal antibody FMC 25. REFERENCES 1. George, J . N., Nurden, A. T., and Phillips, D. R. (1984) N. Engl. J. Med. 311, 1084-1098 2. Coller, B. S., and Gralnick, H. R. (1977) J. Clin. Inuest. 60, 302312 3. Du, X., Beutler, L., Ruan, C., Castaldi, P. A., and Berndt, M. C. (1987) Blood 69, 1524-1527

33. Coller, B. S., Peerschke, E. I., Scudder, L. E., and Sullivan, C. A. (1983) Blood 6 1, 99-1 10 34. Drouin. J.. McGreeor. J. L., Clemetson. K. J.. and Hashemi. S. (1989) Blood 74,-31a (abstr.) 35. Vicente, V., Kostel, P. J., andRuggeri, Z. M. (1988) J . Biol. Chem. 263,18473-18479 36. Weiss, H. J., Turitto, V. T., and Baumgartner, H. R. (1978) J . Lab. Clin. Med. 92, 750-762 37. Weiss. H. J., Mever. D.. Rabinowitz, R., Pietu, G., Girma, J.-P., Vicic, W. J., and Rogers, J . (1982)'N. Engl. J. Med. 306, 326333 38. Miller, J. L., and Castella, A. (1982) Blood 60, 790-794