From the §Department of Medicine, University of Alberta and $Department of Medicine, Cross Cancer Institute,. Edmonton, Alberta T6G 122, Canada.
Vol. 264, No. 21, Issue of July 25, pp. 12289-12293,1969
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Printed in U.S A
0 1989 by The American Society for Biochemistry andMolecular Biology, Inc
Evidence That Monoclonal Antibodiesagainst CD9 Antigen Induce Specific Association between CD9 and the Platelet Glycoprotein IIb-IIIa Complex* (Received for publication, February 22, 1989)
Joseph R, SlupskyS, Jutta G. SeehaferS, Shou-Ching Tangs, AnnaMasellis-Smiths, and Andrew R. E. ShawSsT From the §Department of Medicine, University of Alberta and $Department of Medicine, Cross Cancer Institute, Edmonton, Alberta T6G 122, Canada
Monoclonal antibodies to theCD9 antigen are powerful platelet agonists. We report here thenovel finding that the anti-CD9 monoclonal antibodies 50H.19 and ALBa promote physical association between CD9 antigen and the glycoprotein IIb-IIIa complex (GPIIbIIIa) component of the platelet fibrinogen receptor. The monoclonal antibodies do not consistently immunoprecipitate proteins other than CD9 from lz6I-1abeled human platelets even if the platelets are first treated with the homobifunctional cross-linking reagent dithiobis(succinimidy1 propionate),indicating that CD9 antigen is not physicallyassociated with other membrane proteins the in resting state. However, the additionof agonistic concentrationsof either monoclonal antibody before cross-linking results in thecoprecipitation of proteins corresponding in mobility and peptide composition to GPIIb, and GPIIIa. The association of CD9 with the GPIIb-IIIa complex is unaffected by a combination of aspirin and ADP scavengers sufficient to abrogate anti-CD9 monoclonal antibody-induced platelet aggregation, and is therefore not dependent upon thromboxane- and ADP-mediated pathways of intracellular signalling. The specificity of the association is demonstrated by the lack of other coprecipitating major proteins, by the requirement for induction by anti-CD9 monoclonal antibodies, and by thefailuretopromotereciprocal associationwith either of the anti-GPIIb-IIIa complex monoclonal antibodies P2 or HuP1-mla.
glycoproteins of the platelet (3). Exposure of the fibrinogen binding site follows platelet activation by a number of agonists, andis thought toinvolve conformational changes in the receptor (4,5), but the mechanismby which this is mediated is obscure. The leukocyte differentiation antigen CD9 is also a major component of the platelet membrane (6). Anti-CD9 mAbs are powerful promoters of platelet aggregation and granule release, indicating that CD9 plays a role in platelet activation (6-11). It has beensuggested that CD9 may be involved in fibrinogen binding since anti-CD9 mAbs are not effective agonists of plateletfunctionsinthrombasthenic patients (who are deficient in GPIIb-IIIa, but not CD9) (79), and their action can be blocked in normal platelets by antibodies directed against the GPIIb-IIIa complex (9). A role in fibrinogen binding is further supported by the synergistic effects of fibrinogen on anti-CD9 mAb-induced platelet aggregation (8, 9), and by the recent demonstration that the anti-CD9 mAb PMA-2 promotes bindingof fibrinogen to the human platelet by an intracellular signal-dependent mechanism (11).However, none of these studies has demonstrated a physical or biochemical relationship between CD9 and the GPIIb-IIIa complex on the platelet membrane. We report here, using a reversible cross-linking agent to detect close spatialrelationships,thatplatelet-aggregating concentrations of the anti-CD9mAb 50H.19 or ALBGpromote specific physicalassociation of the CD9 antigen with the GPIIb-IIIa complex. MATERIALS AND METHODS
Platelet Preparation-Platelet-rich plasma was obtained from the citrated venous blood of healthy volunteers and processed as previFollowing vascular injury, platelets rapidly adhere to the ously described (6). Briefly, prostacyclin (Sigma) was added to a final exposed endothelium, andform aggregatesthat constitute the concentration of 10 ng/ml, and apyrase (grade V, Sigma) to a final primary agent of hemostasis. An essential component of the concentration of 1 unit/ml. The platelet-rich plasma was centrifuged at 550 X g for 10 min, and the platelet pellet was resuspended in aggregation phenomenon is the binding of fibrinogen to the Tyrode’s/Hepes buffer prior to gel filtration on aSepharose 2B fibrinogen receptor (1,21, a heterodlmeric glycoprotein com- column (Pharmacia). plex (GPIIb-IIIa)’which contains two of the major cell surface Antibodies-mAb 50H.19 (anti-CDS), an IgG2, was produced as described (12)and further purified from ascites fluid by ammonium * This work was supported by grants from the Medical Research sulfate precipitation, and dialysis against phosphate-buffered saline. Council of Canada andthe AlbertaCancer Board. The costs of mAbALB, (anti-CD9) and mAb P2 (anti-GPIIb-IIIa complex) are publication of this article were defrayed in part by the payment of IgG, antibodies and were obtained as protein A-Sepharose-purified page charges. This article must therefore be hereby marked “adoerantibodies from Daymar Laboratories (Toronto). mAb HuP1-mla, tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate an IgG, (anti-GPIIb-IIIa complex) was obtained as an ascites fluid this fact. from Cedar Lane Laboratories (Hornby, Ontario) and used without 7 To whom correspondence shouldbe addressed Dept. of Medicine, further purification. Cross Cancer Institute, 11560 University Ave., Edmonton, Alberta Radioiodination of Pbtelets-Platelets were radioiodinated by the T6G 122 Canada. lactoperoxidase technique (13). Briefly, 2 X lo9 platelets were sediThe abbreviations used are: GP, glycoprotein; ASA, acetylsalicylic mented from platelet-rich plasma a t 550 X g for 10 min and resusacid; CP, creatine phosphate; CPK, creatine phosphokinase; DSP, pended in 1 ml of Tyrode’s/Hepes buffer. 0.02 mg/ml of lactoperoxidithiobis(succinimidy1 propionate); mAb, monoclonal antibody; SDS- dase (Boehringer Mannheim) was added followed by addition of 1 PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; mCi of carrier-free NaIz5I (University of Alberta Radiopharmacy) and Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid. 20 d / m l of 0.12% hydrogen peroxide. The suspension was incubated
12289
12290
Antibody-induced Association
for 5 min followed by a second addition of 20 pl/ml of the hydrogen peroxide, and a further 5-min incubation before gel filtration. Cross-linking of Platelet Surface Proteins-Surface radioiodinated platelets ( 2 X 10R/ml)in Tyrode's/Hepes buffer were incubated with the homohifunctional reversible cross-linking reagent DSP (Pierce Chemical Co.) a t 100 pg/ml for 10 min a t room temperature. The reaction was quenched by the addition of ammonium bicarbonate to afinal concentration of 100 mM (14). In some experimentsthe platelets were incubated with C P and CPkinase (final concentrations 5 mM and 40 units/ml, respectively) for 5 min a t 37 "C prior to the addition of cross-linker, or to theaddition of mAb (final concentration of 10 pg/ml) and further incubated for 10 min at 22 "C before crosslinking. Preparation of Nonidet P-40Lysates, Immunoprecipitation, SDSPACE, and Densitometry Scanning-The procedures for detergent lysis, immunoprecipitation, and SDS-PAGE were as previously described (15) with the exception that 5-20% polyacrylamide gradient gels were employed for the separation of proteins. All samples were reduced before application to thegels. Densitometry of the autoradiograms was performed on a Joyce-Loebl Chromoscan 3 densitometer. Quantitation of radiolabeled proteins was determined by integration of the areas under thepeaks. Limited Proteo[ysk-Radiolabeled proteins were immunoprecipitated with mAbs from Nonidet P-40 lysates and separated by SDSPAGE on 5-20% gels. Gel strips were excised, overlaid onto SDS16% polyacrylamide gels, and digested with Staphy/ococcus aureus V8 protease during electrophoresis. Platelet Aggregation Studies-Aggregation was performed using a Peyton Aggregation Module. 0.5 ml of gel-filtered platelets (2 X 10' platelets/ml) were used per experiment. Fibrinogen was added to a concentration of 0.36 mg/ml. Platelets were first incubated for 2 min a t 37 "C with stirring, and then further treated a t 37 'C in oneof four ways: (a) incubated for 6 min; ( b )incubated for 6 min in thepresence of ASA (final concentration of 1 mM); (c) incubated for 5 min before adding CP and CPK (final concentration of 5 mmol, and 40 units/ ml, respectively), and then incubatedfor 1 min; or (d) ASA was added (final concentration of 1 mM), the platelets incubated for 5 min in the presence of ASA (final concentration of 1 mM), CP, and CPK added (final concentrations of 5 mM, and 40 units/ml, respectively), andincubated for 1 min. Then aggregationwas initiated by the addition of mAb 50H.19 (to a final concentration of 10 pg/ml) and monitored up to 12 min.
of
CD9 with GPIIb-IIIa
- GPllba - GPllla
B
a b
T
. GPllb
- GPllla
2CD9
"
RESULTS
Chemical Cross-linking of PlateletProteins to CD9-To determine the spatial relationships between CD9 antigen and other cell-surface proteins we used the reversible homobifunctional cross-linking agent DSP in conjunction with immunoprecipitation and analysis by SDS-PAGE. DSP reacts covalently wi:h NH2-terminal and lysine amino groups over a span of 12 A, and therefore reportsclose physical proximity between proteins. Since DSP contains an internal disulfide bond, cross-linkingcan be reversed under reducing conditions. When we employed mAbs 50H.19 or ALB6 to immunoprecipitate proteins from Nonidet P-40 lysates of '2sI-labeled platelets, the 22-, 24-, and 27-kDa components of CD9 (6) were obtained, but not other co-precipitating proteins (Fig. L4, lanes la, and I b ) . Since noncovalent molecular interactions or Triton X-100 can often be sensitive toNonidetP-40 demonstrated using chemical cross-linkingreagents andused to stabilize the associations (14, 16-18), we exposed labeled platelets to DSP prior to lysis and immunoprecipitation with anti-CD9 mAb. This treatment gave inconsistent results. Small, but variable amounts of two co-precipitating proteins (Fig. lA, lanes ,?a, b ) were often detectable in the anti-CD9 mAb immunoprecipitates corresponding in mobility, and inrelative intensity of radiolabeling, toplatelet GPIIb, and GPIIIa immunoprecipitated by the anti-GPIIbIIIa complex-specific mAb P2 (Fig. L4, lanes IC and 2c). In some experiments the proteins were not detectable (Fig. lB, lane a). This suggests that the small and variable degree of association detected by cross-linking prior toimmunoprecip-
FIG.1. mAbs 50H.19 and ALBs promote association of GPIIb-IIIa with CD9 on the intact platelet. A, '2sII-labeledplatelets were incubated with mAb (10 pg/ml) (groups 3 and 4), or without mAb (groups 1 and 2) for 10 min a t 22 "C before cross-linking with DSP for 10 min (groups 2and 4). After washing, platelets were lysed, mAb added to groups 1 and 2, and the immunoprecipitated proteins from all groups separated on 5-20% gradient gels by SDS-PAGE. T, total lysate; lane a, mAb 50H.19; lane b, mAb ALBs; and lane c, mAb P2. B, '2511-labeledplatelets were incubated either untreated ( a ) or with mAb 50H.19 (10 pglml) ( b ) for 10 min a t 22 "C before crosslinking with DSP for 10 min. The platelets were lysed, mAb 50H.19 added to group a, and the immunoprecipitated proteins separated on a 5-20% gradient gel by SDS-PAGE. T, total lysate; lanes a and b, mAb 50H.19. A and B represent platelets from different donors. The positions of GPIIba, GPIIbB, GPIIIa, and the three CD9 components of 22, 24, and 27 kDa are indicated. itation results fromCD9-ligand interaction during platelet preparation. In contrast, when platelets were first exposed to a platelet activating concentration (10 pg/ml) of either mAb 50H.19 or ALB6, and then to DSP, the immunoprecipitates contained large quantities of co-precipitating proteins with the mobility characteristics of GPIIb and GPIIIa(Fig. L 4 , lanes 4a and b ) . Since these proteins were undetectable in immunoprecipitates from platelets treated with the mAbs in theabsence of crosslinker (Fig. 1A, lanes 3a, and 6 ) the association between CD9 antigen and GPIIb-IIIa is indeed Nonidet P-40-labile. The quantity of the co-precipitated heterodimeric complex shows a direct dose-response relationship with the amountof cross-
Antibody-induced Association
of
CD9 GPIIb-IIIa with
12291
linking reagent employed over the range 25-100 pg/ml (data not shown). In contrast to the priming of platelets with anti-CD9mAbs prior to cross-linking with DSP, priming with the anti-GPIIbIIIa mAb P2 at a concentration of 10 pg/ml failed to induce a reciprocal association between the GPIIb-IIIacomplex and CD9 (Fig. l A , lane 4c). The 25-kDa protein immunoprecipitated with mAb P2 is the GPIIbp subunit (Fig. lA, lanes 14c), and isclearly distinguishable from CD9 (Fig. lA, lanes l a L and Ib). The anti-GPIIb-IIIa complex mAb HuP1-mla was F talso unable to promote association an between the GPIIb-IIIa aR complex and CD9 antigen (data not shown). The inabilityof mAbs against the GPIIb-IIIa complex to induce interaction between it and CD9 emphasizes the specificity of the antiCD9-triggered event. Furthermore, since CD9 was never coprecipitated from the lysatesof cross-linked platelets by mAbs P2 (Fig. l A , lane 2c) and HuP1-mla (data not shown), even when the lysates contained small amounts of CD9-GPIIb-IIIa immunoprecipitable by anti-CD9 mAb (Fig. lA, lanes 2a and b ) ,these mAbs do notrecognize the CD9-GPIIb-IIIacomplex. Identification of the Cross-linked Co-precipitants by Limited Time ( m i d Proteolysis-To confirm that the major proteins co-precipitating with CD9 antigen are indeed GPIIb and GPIIIa we FIG.3. Inhibition of mAb 50H.19-induced platelet aggresubjected a gel strip containing anti-CD9 immunoprecipitatedgation by a combination of aspirin, creatine phosphate, and proteins from antibody-primed and cross-linked platelets to creatine phosphokinase. Washed plateletswere untreated (blank), limited proteolysis with S. aureus V8 protease during SDS- or incubated with 100 pmol of ASA, a combination of 5 mmol of CP, described under PAGE ina second gel (Fig. 2, panel 2). Three setsof peptides and 40 units/ml CPK, or both ASA-CP/CPK as and Methods" prior to stimulation with mAb 50H.19 a t were obtained. The one derived from the low molecular weight "Materials 10 pg/ml, and the aggregation recorded. component (Fig. 2, lane 2c) corresponds to that of CD9 (6). The two major co-precipitated '2sI-labeled proteins generated peptides identical to those derived from the proteolysis of GPIIb (Fig. 2, lane la), and GPIIIa (Fig. 2, lane l b ) immunoprecipitated from a platelet lysate by mAb P2. The peptide patterns also closely match the V8 protease peptide pattern described for the GPIIb-IIIa complex (19). Dependence of the Association on CD9-initiated Platelet Activation-Platelet aggregation induced byanti-CD9 mAb is slightly inhibited by the cyclo-oxygenase inhibitor ASA (7, t
l
a b
>
2
I
a b
l
C
2
4
.
6
'
8
10
Distance from origin (cm.)
FIG.2. Identification of GPIIb and GPIIIa from immunoprecipitates by limited proteolysis. GPIIb-IIIa was immunoprecipitated with mAb P2 (10 pg/ml) from Nonidet P-40 lysates of "'1labeled platelets and the proteins separated by SDS-PAGE. A gel strip containing the GPIIb and GPIIIa bands was excised, overlaid onto a further SDS-polyacrylamide gel, and the proteins digested by S. aureus V8 protease during electrophoresis (group I). '"I-Labeled platelets were incubated with the anti-CD9 mAb 50H.19 for 10 min a t 22 "C, cross-linked with DSP for 10 min, and separated by SDSPAGE. The gel strip was excised, and limited proteolysis performed (group 2). The peptide compositions of the two major proteins (2a and 26) exactlycorrespond to those of authentic GPIIb ( l a ) and GPIIIa ( l b ) . Peptides derived from CD9 are indicated (c).
FIG.4. mAb 50H.19-inducedassociationof CD9 antigen and proteins GPIIb and GPIIIa occurs under conditions of complete inhibition of platelet aggregation. Platelet-rich plasma from the donorused for Fig. 3 was incubated with 100pmol of aspirin ASA, radioiodinated, and gel filtered as described under "Materials and Methods." Washed platelets were cross-linked with DSP (100 pg/ml) prior to( a ) ,or following (band c) activationwith mAb 50H.19 in the absence ( b ) ,or the presence (c) of CP (5 mmol) and CPK (40 units/ml). Cross-linked proteins were separated by SDS-PAGE and the autoradiograms scannedby densitometry. 11), but abolished by a combination of ASA and ADP scavengers (1I), suggesting that CD9-induced platelet activation proceeds throughthromboxane amplification of an ADPtriggered pathway. To determine whether the association of CD9 antigen and the GPIIb-IIIa complex was dependent upon intracellular signalling through thesepathways, we employed
12292
Antibody-induced Association
of CD9 with GPIIb-IIIu
a combination of ASA and the ADP scavengers CP/CPK to occur directlyby an antibody-imposed conformational change block mAb-induced platelet aggregation prior to cross-linking in CD9 which exposes a binding site for GPIIb-IIIa, or indiand immunoprecipitation of CD9-associated proteins. Since rectly by the initiation of intracellular signalling pathways we observed donor-specific variation in the extent to which which expose binding sites for CD9 on the GPIIb-IIIa comCD9-mediated platelet aggregation could be blocked by these plex. The latter possibility is very unlikely because we show agents we selected a donor whose platelet aggregation could here that theassociation is not diminished by total blockade be totally prevented by a Combination of ASA and CP/CPK of anti-CD9 mAb-induced platelet aggregation using a com(Fig. 3) to assess the ability of mAb 50H.19 (10 pg/ml) to bination of a thromboxane inhibitor and ADPscavengers. It induce association with GPIIb and GPIIIa under conditions seems more probable that theassociation is a direct result of of complete inhibition of aggregation. Quantitation of the antibody binding to CD9. Furthermore, since a combination relative amounts of the GPIIb-IIIa complex co-precipitating of ASA and ADP scavengerssufficient to abolish platelet with immunoprecipitated CD9 was determined by densitom- aggregation also prevents the binding of fibrinogen (ll), it eter scanning of autoradiograms of SDS gels following anti- can be concluded that the association between CD9 antigen CD9 mAb initiation of platelet aggregation either cross-linked and the GPIIb-IIIacomplex occurs prior to the formationof before the addition of mAb (Fig. 4,a ) , cross-linked after the a functioning fibrinogen receptor, which requires further acor thromboxane-initiated signalladdition of the mAb (Fig. 4, b ) , orcross-linked following tivation by either the ADPsuppression of aggregation by inhibitors of thromboxane- and ingpathways (11).The interactionof CD9 with thefibrinogen ADP-induced platelet activation (Fig. 4, c ) . Whereas no co- receptor under conditions of anti-CD9 mAb-induced platelet precipitating proteins corresponding in mobilityto the GPIIb- activation suggests that the association serves a functional IIIa complex were detectable when DSP was added before the role, perhaps by imposing a new conformation on the GPIIbaddition of mAb 50H.19 (Fig. 4u), a densitometric ratio of co- IIIa complex. In this respect we were struck by our failure to precipitated GPIIb-IIIa to CD9 of 0.67 was obtained under detect even a trace of cross-linked CD9 antigen in immunoconditions of antibody priming (Fig. 4, b ) , and 0.71 when the precipitates containing the CD9-GPIIb-IIIa complex by implatelets were pretreated with a combination of ASA and CP/ munoprecipitation using either of two anti-GPIIb-IIIa comCPK (Fig. 4, c). The association can therefore occur inde- plex-specific mAbs which recognize conformationally deterpendently of platelet activation via thromboxane and ADP- mined epitopes present on GPIIb-IIIa (21, 22). The inability of the anti-GPIIb-IIIa mAbs to detect the CD9-GPIIb-IIIa mediated signalling pathways. complex might therefore report a change in conformation. The finding that Fab fragmentsof the anti-CD9 mAb ALBs DISCUSSION block platelet aggregation induced by the agonists ADP,colIn this report we present the novel finding that the antilagen, and thrombin (7) supports the possibility that CD9 CD9 mAbs 50H.19 and ALBs promote physical association could play a central role in platelet activation. The abilityof between the CD9 antigenand aglycoprotein heterodimer mAb to induceCD9 to associate with the GPIIb-IIIacomplex identified as the plateletfibrinogen receptor GPIIb-IIIa comin the absence of effective intracellular signalling suggests plex. The reliability of chemical cross-linking as a means of that the interaction is conformationally determined, and that identifying specific molecular associations is dependent upon the antibodies mimick the action of a natural ligand. Such a is not merely a ligand has yet to be identified, but couldbe a membrane thedemonstrationthatthecross-linking function of physical proximity engendered byinteraction with component, or extracellular matrix protein which provides a high abundance proteins in a stochastic manner. Since GPIIb- coactivation signalnecessary forthe formationof a functional IIIa isa major component of the plateletsurface it isconceiv- fibrinogenreceptor.CD9 isreported to bea nonintegral able that cross-linking of the complex to CD9 could reflect component of theplasmamembrane(23),appearstobe such nonspecific molecular interaction. However, the specific- anchored by ester-linked long chain fatty acid ligands (241, ity of cross-linking between CD9 and GPIIb-IIIa is indicated and lacks a glycosyl phosphatidylinositol anchor.’ Yet CD9 by the lack of correlation between the relative densities of the can induce phosphoinositide metabolism following mAb bindother labeled proteins in the total lysates (Fig. 1, A, lane T; ing (25). CD9 must therefore interact with integral compoB, lune T ) , and those appearing as co-precipitants (Fig. lA, nents of the plasma membrane in order to transduce signal. its lunes 4u, b; B, lune b). Furthermore, theassociation is actively The GPIIb-IIIa complex may have signalling capacity since induced since it is dependentupon the presence and sequence monovalent mAb directed against oneof the components, the of addition of anti-CD9 mAb. It is conceivable, however, that GPIIbP chain, can initiate platelet activation (26). Our finding CD9 might be cross-linked to GPIIb-IIIa via the immuno- that platelet-aggregatingdoses of two mAbs directed against globulin itself. That this is not the case is demonstrated by CD9 stimulate association of CD9 antigen with the GPIIbthe fact that, although CD9 and GPIIb-IIIa are present in IIIa complex therefore provides the firstevidence for a mechapproximately equal amounts on the platelet membrane (the anism whereby antibodies to CD9 might effect intracellular number of fibrinogen binding sites on the activated platelet signalling pathways. Although an association between CD9 surface averages between 38,000 (20) and 45,000 (4), whereas and thefibrinogen receptor has notpreviously been reported, the number of anti-CD9 mAb binding sites are estimated to it has been proposed on the basis of co-precipitation studies be 46,000 (11) and 65,000 (9)), CD9 cannot be demonstrated with the anti-CD9mAb AG-1 that CD9 may weakly associate by immunoprecipitation with anti-GPIIb-IIIacomplex mAbs. with another platelet glycoprotein GPIb (9). However, only Therefore a combination of immunoglobulin and cross-link- 5% of the GPIb appeared tobe CD9 associated in that study ing agent are not sufficient to produce the reciprocal associ- and could only be visualized following periodate [3Hlborohyation even when the IgG are of the same isotype. dride labeling of the carbohydrate moieties, and not after The necessity to add the antibody to the platelets before surface lZ5I-labeling, making it very unlikely that we could the chemical cross-linking agent distinguishes the dynamic have detected GPIbassociation under our conditions. associationbetween CD9 antigenandGPIIb-GPIIIa from In conclusion, our studies provide direct evidence for an those between other cell surface molecules which are consti* J. R. Slupsky, J. G. Seehafer, S . 4 . Tang, A. Masellis-Smith, and tutively associated (14, 17, 18). How the antibody effects the association is an interesting question. Theoretically this could A. R. E. Shaw, unpublished results.
Antibody-induced Associationof
CD9 with GPIIb-IIIa
12293
8. Higashihara, M., Maeda, H., Shibata, Y., Kume, S., and Ohashi, T. (1985) Blood 6 5 , 382-390 9. Miller, J. L., Kupinski, J. M., and Hustad, K. 0. (1986) Blood 68,743-751 10. Gorman, D. J., Castaldi, P. A,, Zola, H., and Berndt,M. C. (1985) Nouv. Rev. Fr. Hematol. 2 7 , 255-259 11. Hato, T., Ikeda, K., Yasukawa, M., Watanabe, A., and Kobayashi, Y. (1988) Blood 7 2 , 224-229 12. MacLean, G. D., Seehafer, J., Shaw, A. R. E., Kieran, M. W., and Longenecker, B. M. (1982) J. Natl. Cancer Inst. 69,357-364 13. Nachman, R. L., Hubbard, A., and Ferris, B. (1972) J.Biol. Chem. 248,2928-2936 14. Blue,M-L., Craig, K. A., Anderson, P., Branton, K. R., and Schlossman, S. F. (1988) Cell 5 4 , 4 1 3 4 2 1 15. Seehafer, J., Longenecker, B. M., and Shaw, A. R. E. (1984) Znt. J. Cancer 3 4 , 821-829 16. Brenner, M. B., Trowbridge, I. S., and Strominger, J. L. (1985) Cell 4 0 , 183-190 17. Samson. M.. Cousin. J.-L.. and Fehlmann., M. (1986) . , J.Zmmunol. Acknowledgments-We would like to thank Dr. M. Longenecker, 137,2293-2298 ' Department of Immunology, University of Alberta, for kindly supplying us with mAb, and thestaff of the Audiovisual Department of the 18. Bonnefov. J. Y.. Guillot. 0.. SDits. H.. Blanchard. D.. Ishizaka. K., and'Banchereau, J: (198g) J. 'Exp. Med. 1 6 7 ; 57-72 CrossCancer Institute for theirpreparation of the photographic 19. Burns, G. F., Cosgrove, L., Triglia, T., Beall, J. A., Lopez, A. F., material. Werkmeister, J. A., Begley, C. G., Hadda, A. P., D'Apice, J. F., Vadas, M. A., and Cawley, J. C. (1986) Cell 4 5 , 269-280 REFERENCES 20. Marguerie, G. A., Edgington, T. E., and Plow, E. F. (1980) J. Biol. 1. Mustard, J. F., Kinlough-Rathbone, R. L., Packham, M.A., Chem. 255,154-161 Perry, D. W., Harfenist, E. J., and Pai, K. R. M. (1979) Blood 21. McGregor, J. L., Brocher, J., Wild, F., Follea, G., Trzeciak, M. C., James, E., Dechavanne, M., McGregor, L., and Clemetson, 54,987-993 K. (1983) Eur. J. Biochem. 1 3 1 , 427-436 2. Marguerie, G. A,, Plow, E. F., and Edgington, T. S. (1979) J. Biol. 22. Thurlow, P. J., Barlow, B., Connellan, J. C., and McKenzie, I. F. Chem. 254,5357-5363 C. (1983) Br. J. Hematol. 5 5 , 123-134 3. Jennings, L. K., and Phillips, D. R. (1982) J. Biol. Chem. 2 5 7 , 23. Newman, R. A., Sutherland, D. R., LeBien, T. W., Kersey, J. H., 10458-10466 and Greaves, M. F. (1982) Biochim. Biophys. Acta 7 0 1 , 3184. Shattil, S. J., Hoxie, J. A,, Cunningham, M., and Brass, L. F. 327 (1985) J . Biol. Chem. 2 6 0 , 11107-11114 24. Seehafer, J. G., Tang, S. C., Slupsky, J. R., and Shaw, A. R. E. 5. Coller, B. S. (1985) J. Clin. Invest. 7 6 , 101-108 (1988) Biochim. Biophys. Acta 9 5 7 , 399-410 6. Seehafer, J. G., Slupsky, J. R., Tang, S. C., and Shaw, A. R. E. 25. Higashihara, M., Maeda, H., Yatomi, Y., Takahata, K., Oka, H., (1988) Biochim. Biophys. Acta 9 5 2 , 92-100 and Kume, S. (1985) Biochem. Biophys. Res. Commun. 1 3 3 , 7. Boucheix, C., Soria, C., Mirshahi, M., Soria, J., Perrot, J. Y., 306-313 Fournier, N., Billard, M., and Rosenfeld, C. (1983) FEBS Lett. 26. Jennings, L., Phillips, D. R., and Walker, W. S. (1985) Blood 6 5 , 161,289-295 1112-1119
association between CD9 antigen and a configurable member of the integrin family, and constitute the first biochemical evidence for the spatial interaction of a surface protein with the GPIIb-IIIa complex. Exposure of the fibrinogen binding site is known to depend upon a complex multistep mechanism involving the interplay of signalling pathways and conformationalevents. The requirement for intact pathways of intracellular signalling for the anti-CD9mAb-mediated binding of fibrinogen to the fibrinogen receptor (ll),in combination with the data presented here, establish CD9 as a functional and physical component of the GPIIb-IIIa complex. The characterization of the biochemical relationships between these moleculesis likely to provide new insight intothe functioning of the fibrinogen receptor.
