THE JOURNAL OF BIOLOGICAL CHEMISTRY Val. 257.,No. 17, Issue of September 10, pp. 10458-10466. 1982 Printed L?I U.S.A.

Purification of Glycoproteins IIb and I11 from Human Platelet Plasma Membranes and Characterization IIb-I11 Complex*

of a Calcium-dependent Glycoprotein

(Received for publication, October 13, 1981)

Lisa K. Jennings and David R. Phillips$ From the Departments of Biochemistry, St. Jude Children’s Research Hospital and The University of Tennessee Center for the Health Sciences, Memphis, Tennessee 38101 and Gladstone Foundation Laboratories for Cardiovascular Disease, University of California, Sun Francisco, Sun Francisco, California 94140

Human platelet membrane glycoproteins IIb and I11 are two major integral membrane components that have been identified as sites mediating thrombin-in- duced aggregation. For purposes of our study, glyco- proteins IIb and 111 were solubilized by extracting platelet plasma membranes with a buffer containing 0.1% Triton X-100 and were separated by gel filtration chromatography on Sephacryl S-300, employing Triton X-100-containing column buffers with or without urea or guanidine hydrochloride. The physical properties of the purified glycoproteins were: for glycoprotein IIb,

namic values), Mr = 136,000 (sodium dodecyl sulfate gels); for glycoprotein 111, R, = 67 A, szo,w = 3.2, f / fo = 2.1, M, = 93,000 (hydrodynamic values), Mr = 95,000 (sodium dodecyl sulfate gels). Although the amino acid compositions of the two glycoproteins were similar, antibodies raised against glycoprotein IIb did not cross- react with glycoprotein 111. If divalent cations were not chelated in the Triton extract, glycoproteins IIb and 111 coeluted during gel filtration chromatography (appar- ent Stakes radius of 71 A) and co-sedimented on sucrose gradients (apparent S Z O , ~ of 8.6), from which M, = 265,000 was calculated. Glycoproteins 1Ib and I11 were coprecipitated by an antibody monospecific for glyco- protein IIb. The two glycoproteins dissociated into monomers when EDTA was added to Triton lysates. Readdition of Ca2+ caused them to reassociate into a complex with physical properties similar to those of the complex in the original Triton lysate. The data show that glycoproteins IIb and 111 are a heterodimer complex, that complex formation depends upon the presence of Ca2+, and that chelation of Ca2+ causes dissociation into monomeric glycoproteins.

R, = 61 A, 6 2 0 , ~ = 4.7, f/fo = 1.7, M, 125,000 (hydrody-

The membrane surface of human platelets is responsible for a wide variety of reactions. This surface has receptors for various agonists (e.g. thrombin, ADP, collagen, serotonin) and antagonists (e.g. prostacyclin, prostaglandin,) that affect platelet function (see Ref. 1 for review). It also has receptors to enhance the rate of coagulation (e.g. for factor Xa (2)). In addition, the membrane surface of platelets has receptors for several other proteins (e.g. fibrinogen (3, 4), factor VIII/von

* This work was supported by Grants HL21487 and HL15GlG from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked “advertisement” in accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact. + Recipient of Career Development Award HL00080.

Willebrand factor (5 ,6) , thrombospondin (7,8), and fibronec- tin (9)).

One functional activity of the membrane surface not yet defined is aggregation. This activity is not expressed unless platelets are activated by one or more agonists. In addition, activity requires the presence of Ca2+ in the extracellular space. Platelet aggregation appears to be mediated by specific sites on the membrane surface that can allow for platelet to platelet cohesion with or without the incorporation of other blood cell types (10). Although these sites are measurable in isolated platelet plasma membranes (Il), it is clear that puri- fication and characterization of the sites are essential in de- termining their role in platelet-specific reactions.

Two experimental approaches have been used to identify aggregation sites on the platelet membrane surface. One ap- proach has been to examine platelets from patients with Glanzmann’s thrombasthenia, an inherited bleeding disorder characterized by a long bleeding time, with platelets neither aggregating in response to physiological stimuli nor retracting clots (12). Platelets from these patients are deficient in two major membrane glycoproteins termed glycoprotein IIb and glycoprotein 111, whereas other membrane components are present in apparently normal amounts (13, 14). The absence of these specific glycoproteins has led to the suggestion that they mediate platelet to platelet interaction (see Ref. 15 for review).

Using a different approach to identify platelet membrane aggregation sites, we have found that the addition of Triton X-100 to thrombin-aggregated platelets yields aggregated platelet cytoskeletons, while cytoskeletons from thrombin- stimulated platelets are not aggregated (16). We observed that two membrane glycoproteins, IIb and 111, are selectively re- tained with the aggregated structures but are not present in those that are not aggregated. These data suggest that glyco- proteins IIb and 111 are involved in the direct interaction of platelets during aggregation and that they become Triton- insoluble because of macromolecular associations between the membrane surfaces and actin filaments within platelets. It is not known whether these associations are direct or mediated by other proteins.

Glycoproteins IIb and I11 are prominent integral membrane components. Both glycoproteins are exposed on the outer membrane surfaces, as demonstrated by their labeling on intact cells by probes such as lactoperoxidase-catalyzed iodi- nation (17), transglutaminase (18), and diiododiazosulfanilic acid (19), and by their susceptibility to proteolytic enzymes (20). The carbohydrate moieties of both IIh and 111 are expressed on the platelet membrane surface, since they are labeled by neuraminidase/galactose oxidase/[”H]NaBH4 (21)

10458

Platelet Membrane Glycoproteins IIb and 111 10459

and by periodate/[”H]NaBH4 (22). Both glycoproteins appear to be transmembrane, since neither is hydrolyzed by thrombin on intact cells but both can be hydrolyzed by this protease after membrane isolation (23). The apparent molecular weight of glycoprotein IIb as determined by SDS’-polyacrylamide gel electrophoresis is 142,000, and it consists of two disulfide- linked subunits, glycoprotein IIba (Mr = 132,000), and glyco- protein IIbp (M, = 23,000) (17). Glycoprotein I11 contains at least two intrachain disulfides and has M, = 114,000 when reduced and M , = 99,000 with disulfides intact (17).

Recent studies indicate that glycoproteins IIb and 111 exist as a complex. McEver et al. (24) found that a monoclonal antibody copurifies both glycoproteins on an affinity column. This indicates that the two glycoproteins share an antigenic determinant or that they are a complex. Both glycoproteins give a common precipitin line following crossed immunoelec- trophoresis (25-27). Kunicki et al. (28) demonstrated that the common precipitin line occurs only in the presence of Ca’+ and that the addition of EDTA to Triton extracts causes this precipitin line to disappear with the concomitant appearance of two new arcs.

Although neither glycoprotein IIb nor 111 has been isolated in the absence of denaturants, two procedures have been reported for purifying a mixture of the two glycoproteins. Coisolation of these glycoproteins has been accomplished by McEver et al. (24), using an affinity column containing a monoclonal antibody, and by Leung et al. (29), using an affinity column containing Lens culinaris lectin. In the pres- ent study, we have reported a method for purifying glycopro- teins IIb and I11 and separating these glycoproteins in a buffer containing urea and Triton X-100. The properties of these glycoproteins indicate that they form a heterodimer complex, that complex formation depends upon the presence of Ca2+, and that chelation of Ca’+ causes their dissociation into mon- omeric glycoproteins.

MATERIALS AND METHODS‘

RESULTS

Physical Properties of Glycoproteins ZZb and ZII-Glyco- proteins IIb and I11 were purified to apparent homogeneity from isolated platelet plasma membranes by column chro- matographic procedures in buffers containing Triton X-100. Electrophoresis of the nonreduced glycoproteins in 7.5% SDS gels indicated that the molecular weight of purified glycopro- tein IIb is 136,000, while that of glycoprotein 111 is 95,000 (Fig. 6A). The apparent molecular weight of glycoprotein I11 in- creased to 108,000 after reduction (Fig. 6B), an effect attrib- uted to intrachain disulfides in the native glycoprotein (17). After reduction, the larger of the two disulfide-linked subunits had an apparent molecular weight of 116,000 (Fig. 6B). The Stokes radius of the purified glycoproteins was determined by gel fitration chromatography through Sephacryl S-300 (Fig. 7 ) . Glycoprotein 111 purified by column chromatography in urea was used for these studies since the guanidine column protocol produced an oxidized product. Because sedimenta- tion coefficients can be determined in sucrose gradients in the

The abbreviation used is: SDS, sodium dodecyl sulfate. * Portions of this paper (including “Materials and Methods,” Figs.

1-5, and Tables I and 11) are presented in miniprint at the end of this paper. The abbreviations used are: DMSO, dimethyl sulfoxide; Gp, glycoprotein; BSA, bovine serum albumin. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 RockvUe Pike, Bethesda, MD 20814. Request Document No. 81M-2504, cite the authors, and include a check or money order for $5.60 per set of photocopies. Full sue photographs are also included in the m i c r o f i edition of the Journal that is available from Waverly Press.

250{ A

0.2 0.4 0.6 0.8 I O

I

0.2 0.4 0.6 0.8 1.0 R E L A T I V E MOBILITY

FIG. 6. Molecular weight estimates of glycoproteins IIb and I11 by SDS-gel electrophoresis under reducing and nonreduc- ing conditions. Purified glycoproteins and molecular weight stand- ards were electrophoresed through a 7.5% SDS-polyacrylamide gel according to the procedure of Laemmli (41) either (A) without, or ( B ) with prior reduction with /3-mercaptoethanol. Molecular weight standards: 1, myosin, M , = 200,000; 2, P-galactosidase, M , = 116,250; 3, phosphorylase b, M , = 92,500; 4, bovine serum albumin, M , = 66,200; 5, carbonic anhydrase, M , = 31,000.

‘“r- 1

1 I 1 I I

2 4 .6 K D

FIG. 7. Determination of Stokes radius by gel filtration chro- matography. A column (0.9 X 49 cm) of Sephacryl S-300 was equilibrated with 0.05 M sodium phosphate, pH 7.4, 0.05% Triton X- 100, and developed at a flow rate of 2 ml/h. The sample size for all experiments was 0.2 ml and 15-drop fractions were collected into preweighed tubes. The exclusion volume (VO = 15.55 ml) was deter- mined with blue dextran (Pharmacia) and the total volume (V,,,,l = 35.17 m l ) was measured with [14C]sucrose. For calibration, the follow- ing proteins were used: 1, /3-galactosidase, R, = 69 A; 2, phosphorylase a, R, = 65 A; 3, catalase, R, = 52 A; 4, ovalbumin, R, = 27 A.

presence of other proteins (30), it was assumed that the presence of contaminating glycoprotein IIb, which was less than 20% of the amount of glycoprotein 111, did not interfere in these measurements. By these procedures, the apparent Stokes radius of glycoprotein IIb was found to be 61 A, and that of glycoprotein 111 67 A. The sedimentation coefficients of the two glycoproteins were determined by sedimentation through linear 5 to 25% sucrose gradients (Fig. 8). Calibration of these gradients with standard proteins yielded the following values: glycoprotein IIb, s ~ ~ . ~ ~ = 4.7; glycoprotein 111, ~ 2 0 . ~ . = 3.2.

Interaction of Glycoproteins ZZb and ZIZ-The coisolation of glycoproteins IIb and I11 in the initial Sephacryl S-300 column and ion exchange columns suggested that, either these glycoproteins were a complex in solution, or they had similar properties that did not easily permit separation. During the

10460 Platelet Membrane Glycoproteins IIb and IIl

initial stages of this study, anti-glycoprotein IIb antiserum was raised in rabbits by injecting them with glycoprotein IIb that had been extracted from SDS gels. This antiserum was judged monospecific for glycoprotein IIb since only this band was identified when the antiserum was added to SDS gels of platelet plasma membranes (Fig. 9). In addition, this antise- rum formed a precipitin line against purified glycoprotein IIb, but not against purified glycoprotein I11 (data not shown). However, addition of the antiserum to Triton X-100-solubi- lized membranes coprecipitated this glycoprotein with glyco- protein IIb in a ratio similar to that which exists in membranes (Fig. 10). Preimmune serum was without effect. These data were consistent with the notion that glycoproteins IIb and I11 exist as a complex.

The amount of glycoprotein IIb-I11 complex in the eluate from the initial Sephacryl S-300 column used for glycoprotein purification was determined by centrifugation through 5 to 25% sucrose gradients. Although some complexed material with a sedimentation coefficient of approximately 8.6 s was observed, the amount present varied, representing between 10 to 40% of the glycoprotein present (range of four preparations, data not shown). A recent publication by Kunicki et al. (28) showed that the presence of Ca2+ was required for the copre- cipitation of glycoproteins IIb and I11 during crossed immu- noelectrophoresis. If Ca" mediates the interaction of glyco- protein IIb with glycoprotein 111, variations in Ca2+ concen- trations could account for the variable amounts of complex present. Fig. 11A shows the distribution of glycoproteins IIb and I11 in 5 to 25% sucrose gradients in the presence of 2 mM Ca'+. Essentially all of the glycoprotein IIb and I11 present cosedimented with a sedimentation coefficient of 8.6 s. This result was obtained even with membranes previously treated with a buffer containing 0.01 M EDTA, 0.01 M Tris, 0.15 M sodium chloride) for 1 h at 25 "C. A similar sedimentation coefficient was observed for the two glycoproteins in Triton extracts of freshly isolated platelets (data not shown). Addi- tion of leupeptin (1 mg/ml) to inhibit the calcium-dependent protease in platelets (31, 32) had no effect on the yield or physical properties of the complex. No detectable binding (<5% of the protein mass) of Triton X-100 to the complex was

10 20 30 40 FRACTION NUMBER

FIG. 8. Determination of sedimentation coefficients by cen- trifugation on linear sucrose gradients. Linear sucrose gradients (5 to 258 sucrose) containing 0.05 M sodium phosphate, pH 7.4, and 0.05% Triton X-100 were centrifuged for 6.25 h at 49,000 rpm and 22 "C in a SW 50.1 rotor (Beckman). The sample size for all experi- ments was 0.2 ml and the gradient was analyzed by collecting 8-drop fractions from a punctured tube (46 to 47 fractions/gradient). The following proteins were used for calibration: I , catalase, sm.. = 11.3; 2, y-globulin, sa,,,,. = 7.12; 3, lactoperoxidase, ~ m . , ~ = 5.37; 4, bovine serum albumin, sa,... = 4 .6 5, ovalbumin, sa).,,. = 3.66.

"

1 2 3 FIG. 9. Immuno-overlay of SDS gels of human platelet

plasma membranes with anti-glycoprotein IIb antibody. Plate- let plasma membranes were solubilized in 2% SDS and electropho- resed under nonreducing conditions through 7.5% acrylamide gels (lane I ) . Glycoprotein IIb was identified by incubating the acrylamide gel with the immuno-overlay described by Adair et al. (42), with slight modification (see Materials and Methods). Lane 2, immuno-overlay with preimmune serum; lane 3, immuno-overlay with immune serum.

A. 0. -

- II b, - 111

1 2 3 4 1 2 3 4

FIG. 10. Coprecipitation of glycoproteins IIb and 111 by mon- ospecific anti-glycoprotein IIb antibody. Membrane proteins were labeled with Iz5I by the lactoperoxidase-catalyzed iodination procedure and then solubilized with 1% Triton X-100 in a buffer containing 0.001 M EDTA, 0.01 M Tris, pH 7.4 at 22 "C. Various concentrations of either (A) preimmune serum, or ( B ) immune serum (lane I , 0 pl; lane 2, 20 pl; lane 3, 40 pl; lane 4, 50 pl) were incubated with the solubilized sample for 4 h at 4 "C. The antigen-antibody complex was adsorbed onto protein A-coated Staphylococcus aureus solubilized in 2 8 SDS, and electrophoresed through exponential 5 to 20% acrylamide gels.

observed when the complex was sedimented through a gra- dient containing "-labeled Triton X-100 as described by Clarke (33). The presence of 5 mM EDTA in the extraction buffer and in the sucrose gradient caused the two glycopro- teins to sediment at slower rates (Fig. 11B), which were similar to those of each of the purified glycoproteins. The apparent Stokes radius of the calcium-dependent glycoprotein IIb-I11 complex was 71 A, as determined by gel filtration chromatog- raphy through Sephacryl S-300 in a buffer containing 2 mM CaCI2, 0.05% Triton X-100,O.l M Tris-HCI, pH 7.4. From these values we calculated M , = 265,000 for the complex. Table I11 summarizes these data and those of the individual glycopro- teins. Addition of 10 mM CaC12 to the membranes extracted in

Platelet Membrane Glycoproteins IIb and III 10461

TABLE 111 Summary of the physical properties of glycoproteins (GP) IIb and 111 and the calcium-dependent glycoprotein IIb-III complex

Value" Property

GP IIb GP 111 GP IIb-111 complex

Stokes radius (A)h 61 f 4 67 f 5 Sedimentation coefficient, sm,, (S)' M,, calculated" Frictional ratio, f/fo" 1.7 2.1 1.5 M,, SDS electrophoresis" 136,000 f 13,000 95,000 f 8,000 (231,000 f 21,000)'

71 f 2 8.6 f 0.2 4.7 f 0.2 3.2 f 0.2

125,000 f 13,000 93,000 f 12,000 265,000 f 15,000

~

" All values were obtained for nonreduced samples. Error limits for Stokes radii, sedimentation coefficients, and M , from SDS gels rep- resent 95% confidence limits of the least squares regression lines.

From gel fdtration. ' From sedimentation on linear sucrose gradients. " Calculated from equations:

M , = 6nNR,~zo.t, 1 - up2o.u.

and

I 1

3000 c. I 2000/

1( 5000 4000

2000

6000

4000 .r 6

% 2000 z

- - -

6000

'4000

2000

(25%) Tube Number (5%)

FIG. 11. Sedimentation of Triton X-100-solubilized glycopro- teins (GO IIb and I11 in the presence and absence of Ca2+. Proteins and glycoproteins in isolated platelet plasma membranes were iodinated by lactoperoxidase-catalyzed iodination and solubi- lized at 22 "C in a solution containing 2 mM CaC12, 1% Triton X-100, and 0.1 M Tris, pH 7.4. The Triton-insoluble material was removed by centrifugation and 0.2-ml aliquots were loaded on a 5 to 25% linear gradient of sucrose and sedimented 5.1 h at 54,000 rpm and 22 "C in a SW 55 rotor (Beckman). Each fraction (20 drops/fraction) was treated with SDS and electrophoresed through 7.5% acrylamide gels. Glycoproteins IIb and 111 were cut from dried gels and the amount of radioactivity in each determined in a gamma counter. A, solubilized

sample treated with 5 mM EDTA for 1 h at 22 "C and loaded on a sample loaded on a gradient containing 2 mM CaC12. B, solubilized

gradient containing 10 mM EDTA. C, sample B treated with 10 mM Ca'+ for an additional hour at 22 "C and loaded on a gradient containing 2 mM CaC12.

5 mM EDTA (approximately 5 mM free Ca") caused 40% of the glycoprotein to again cosediment (Fig. 11C). When the complex isolated from the Ca2'-containing gradient (Fig. 11A) was dialyzed uersus 10 mM EDTA, the two glycoproteins

where d was assumed to be 0.2 g of solvent/g of protein (43). Since SX,.D~O of the three proteins did not differ appreciably from the values in H20, the average U of the standards employed, 0.738, was used for these calculations (38).

Fig. 6A. 'The sum of the values for glycoproteins IIb and 111.

sedimented at rates similar to those of the purified glycopro- teins (data not shown). No polypeptides other than glycopro- teins IIb and I11 were observed in the gradient after dissocia- tion of the complex.

DISCUSSION

This study has presented a method for the purification of glycoproteins IIb and 111 from Triton X-100-solubilized plate- let plasma membranes. The method involved solubilization of both glycoproteins from isolated platelet plasma membranes followed by gel filtration chromatography through Sephacryl 5-300 employing Triton X-100-containing column buffers with or without either urea or guanidine hydrochloride. Initial solubilization resulted in a 4-fold purification: neither glyco- protein was solubilized in cold Triton X-100 solutions contain- ing 0.1 M salt, and both were quantitatively solubilized by Triton X-100 in low ionic strength solutions. This differential extraction permitted initial separation of glycoproteins IIb and I11 from other membrane proteins that were soluble at higher salt concentrations in the presence of Triton X-100. After gel Titration chromatography of the extract through a single column, the eluted material consisted almost exclusively of a mixture of glycoprotein IIb and I11 (47 and 49%, respec- tively).

One objective of the present study was to separate glyco- protein IIb from glycoprotein 111 without using SDS, as was used in a previous study (29). Even though both glycoproteins were monomeric when divalent cations were chelated by EDTA (see below), the two glycoproteins had similar prop- erties in Triton X-100 and were, therefore, difficult to separate by chromatographic procedures. The Stokes radii of the two glycoproteins, 61 A for glycoprotein IIb and 67 A for glycopro- tein 111, caused them to elute at comparable positions on gel filtration columns. Their similar isoelectric points (approxi- mately 5.3) (34) most likely accounted for their coelution from ion exchange columns. Although urea, guanidine hydrochlo- ride, and SDS brought about their separation, urea was se- lected as an agent in the final purification scheme, since other studies have shown that functional activity can be restored when proteins are removed from solutions containing urea (35). Using this medium, we purified glycoprotein IIb to apparent homogeneity and prepared glycoprotein 111 contam- inated with approximately 20% glycoprotein Ilb. At the pres- ent time, no assay is available for establishing the functional role of these glycoproteins.

Glycoprotein I11 was purified to homogeneity as determined by SDS-gel electrophoresis only after reduction of the glyco-

10462 Platelet Membrane Glycoproteins IIb and IIl

protein. Treatment of glycoprotein 111 with P-mercaptoetha- no1 in guanidine hydrochloride produced a high molecular weight form of the glycoprotein, which eluted near the void volume of a Sephacryl S-300 column and failed to enter an SDS-acrylamide gel. Although this property permitted isola- tion of glycoprotein I11 without detectable glycoprotein IIb, glycoprotein I11 purified by this procedure is obviously not suitable for studies requiring native disulfides within the mon- omeric glycoprotein. It was of interest that a high molecular weight product formed. Glycoprotein 111 solubilized in 2% SDS was reduced by 1% P-mercaptoethanol, and in 6 M guanidine hydrochloride a disulfide-linked complex was formed. While the basis for the formation of this product remains obscure, a possible explanation is that in guanidine hydrochloride the disulfide exchange occurs preferentially over reduction once the protein is partially reduced. If true, glycoprotein I11 must self-associate under these conditions.

The calculated molecular weights for the individual glyco- proteins were in close agreement with the values obtained from electrophoresis in SDS gels: glycoprotein IIb, 125,000 uersus 136,000, respectively; glycoprotein 111, 93,000 versus 95,000, respectively. The values determined by the two tech- niques were remarkably consistent. Any discrepancy in the two values may have been due to an underestimation of the molecular weight from physical parameters by underestimat- ing the Stokes radius of an asymmetric protein by gel filtration chromatography (36), and overestimating the molecular weight of a glycoprotein by SDS gels (37). The data suggested that each glycoprotein exists as a monomer in the Triton X- 100-containing buffer. The calculated frictional ratios of the two glycoproteins were large (f/fo = 1.7 for glycoprotein IIb; f / fo = 2.1 for glycoprotein 111), indicating that both glycopro- teins are asymmetric. Glycoprotein I11 eluted before glycopro- tein IIb on gel filtration columns but sedimented slower than glycoprotein IIb in linear sucrose gradients, indicating that glycoprotein I11 is probably more asymmetric than glycopro- tein IIb.

A recent report by Leung et al. (29) showed that glycopro- teins IIb and I11 yield different sets of peptide fragments from trypsin hydrolysis, suggesting that each glycoprotein is a separate molecular entity. Our data support this conclusion. First, the physical properties of each glycoprotein as described above were markedly different. Second, while the amino acid compositions of the two glycoproteins were similar, significant differences in the mole percentages of several of the amino acids were observed. Third, antibodies raised against glyco- protein IIb did not cross-react with glycoprotein 111. Further work is required to determine whether these observed differ- ences in glycoprotein behavior are in fact due to differences in polypeptide structure or to differences in glycosylation.

Previous studies by others have indicated that glycoproteins IIb and I11 interact when solubilized in Triton X-100: they are present in a common arc in crossed immunoelectrophoresis plates (26) and they are co-isolated by a monoclonal antibody raised against human platelets (24). In support of these obser- vations, we have found that both glycoproteins were precipi- tated by monospecific glycoprotein IIb antiserum in Triton X- 100 extracts of platelet plasma membranes. Recently, Kunicki et al. (28) found that the presence of Ca’+ was required for glycoproteins IIb and I11 to co-precipitate during crossed immunoelectrophoresis. In the present study, we have iden- tified this complex as a species with a sedimentation coeffi- cient of 8.6 s and a Stokes radius of 71 A. In agreement with the findings of Kunicki et al. (28), the presence of Ca2+ was required for complex formation. Glycoproteins IIb and I11 were stained by Coomassie blue to a similar intensity in SDS gels of the complex from linear sucrose gradients prepared in

the presence of Ca”, indicating that the two glycoproteins are present in equal amounts. Thus far, we have been unable to detect any proteins in the complex other than glycoproteins IIb or 111. Although a small protein like calmodulin might have gone undetected, no calmodulin was detected in the purified complex isolated from sucrose gradient, nor was any binding of 1251-caLmodulin found in the reconstituted complex (L. K. Jennings, R. Wallace, and D. R. Phillips, unpublished observations). Because the sums of the individual molecular weights of the two glycoproteins determined from hydrody- namic measurements (218,000) and from SDS electrophoresis (231,000) were similar to the calculated molecular weight of the complex (265,000), and since the apparent stoichiometry of glycoprotein IIb and glycoprotein I11 in the complex was 1:l as determined by densitometric scans of Coomassie blue- stained gels, we concluded that the complex is a heterodimer consisting of 1 molecule of glycoprotein IIb and 1 molecule of glycoprotein 111.

Our experiments suggest that the calcium-dependent gly- coprotein complex exists in membranes similar to what is observed in Ca2+-containing buffers and is not a result of detergent solubilization of the platelet membrane. First, the stoichiometry of glycoproteins IIb and I11 in membranes, as determined by Coomassie blue staining of SDS gels, was similar to that in the complex isolated either from Ca2+- containing sucrose gradients or gel filtration columns (approx- imately 1:l). Second, when a Triton X-100 extract of platelet membranes was run on sucrose gradients containing EDTA as in Fig. 11, separation of the subunits was observed. There- fore, solubilization of platelet membranes by Triton X-100 did not account for complex formation of glycoproteins IIb and 111. Third, while essentially all of the glycoprotein IIb and I11 was in a complex in membranes extracted with Triton in the presence of Ca2+, we were unable to reform the complex completely in extracts where the complex had been dissoci- ated with EDTA.

The data presented herein permit speculations concerning orientation of the glycoprotein IIb-I11 complex within mem- branes. First, the high frictional ratios of the individual gly- coproteins indicate that each is highly asymmetric. Second, the similarity of the Stokes radii of the glycoprotein IIb-111 complex to those of the individual glycoproteins implies that the two glycoproteins interact with extensive overlap along their long axis. Finally, our inability to detect bound Triton X-100 suggests that hydrophobic domains constitute only a small portion of the glycoprotein, even though both glycopro- teins appear to be transmembrane (23). Presumably, hydro- phobic interactions with membranes would be limited to these domains, as has been suggested for the adenylate cyclase of the rat renal medulla (38) and the nerve growth factor receptor of adult rabbit superior cervical ganglia (39).

Previously, we considered several possibilities to explain the deficiency of glycoproteins IIb and I11 in platelets of patients with the inherited bleeding disorder Glanzmann’s thrombasthenia: 1) that the polypeptide structures of glyco- proteins IIb and I11 are the products of the same gene; 2) that they have a post-translational modification in common; and 3) that the synthesis of both glycoproteins is required before either intercalates in plasma membranes (14). Although the fist two possibilities have not been eliminated, it seems likely that a defect in the synthesis of one subunit of a membrane glycoprotein complex would affect the assembly of another subunit into membranes. An example of this is shown in the work of Cabral and Schatz (40), who used a genetic approach to study the synthesis of cytochrome c oxidase subunits in yeast mitochondrial membranes. They found that mutants defective in the synthesis of cytochrome c oxidase subunits

Platelet Membrane Glycoproteins IIb and III 10463

coded by nuclear DNA cause both a loss and a defective assembly of cytochrome c oxidase subunits coded by mito- chondrial DNA. If defective glycoprotein synthesis in mega- karyocytes affects assembly of the glycoprotein IIb-III com- plex in platelets, it is possible that phenotypic expression of thrombasthenia may not result from an aberrant expression of the same gene product in all affected individuals.

While the functional studies described above suggest that glycoprotein IIb or glycoprotein III is an aggregation site on the membrane surface, it remains to be determined which glycoprotein subunit is responsible for this activity. It is our objective to use the purified glycoproteins to determine whether one of them performs all functional activity, or if they must exist in a complex before the functional activity is expressed.

Acknowledgments-we thank Becky Naukam and Anne Baughan for their enthusiastic and capable technical assistance, Dr. Stanley Rail for determining the amino acid analysis, Dr. Michael Dockter for writing the computer program used to integrate the areas under peaks of densitometer scans, and Dr. Martin Morrison for his gift of bovine lactoperoxidase. We thank Joe F. Andres and Russell Levine for their editorial assistance and Richard A. Wolfe for manuscript. preparation.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10. 11.

12.

13.

14. 15.

16.

17.

REFERENCES Berndt, M. C., and Phillips, Il. R. (1981) in Platelets in Biology

and Pathology (Gordon, J., ed) pp. 43-75, Elsevier/North- Holland Biomedical Press, Amsterdam

Miletich, J. P., Jackson, C. M., and Majerus, P. W. (1978) J. Biol. Chem. 253,6908-6916

Bennett, J. S., and Vilaire, G. (1979) J. Cl&. Znuest. 64, 1393- 1401

Marguerie, G. A., Plow, E. F., and Edgington, T. S. (1979) J. Biol. Chem. 254. 5357-5363

Kao, K. J., P&o, S. V., and McKee, P. A. (1979) J. Clin. Inuest. 63,656~664

Schneider-Trip, M. D., Jenkins, C. S. P., Kahle, L. H., Sturk, A., and ten Cate, J. W. (1979) Br. J. Haematol. 43,99-112.

Phillips, D. R., Jennings, L. K., and Prasanna, H. R. (1980) J. Biol. &em. 255, 11629-11632

George, J. N., Lyons, R. M., and Morgan, R. K. (1980) J. Clin. IRuest.66,1-9

Plow, E. F., Birdwell, C., and Ginsberg, M. H. (1979) J. Clin. Invest. 63,540-543

Baumgartner, H. R. (1973) Microuasc. Res. 5,167-179 Prasanna, H. R., Edwards, H. H., and Phillips, D. R. (1981) Blood

57.305-312 Caen, J. I’., Castaldi, P. A., Leclerc, J. C., Inceman, S., Larrieu,

M.-J., Probst, M., and Bernard, J. (1966) Am. J. Med. 41, 4-27 Nurden, A. T., and Caen, J. P. (1974) Br. J. Haematol. 28, 253-

260 Phillips, D. R.. and Agin, P. P. (1977) J. Clin. Inuest. 60,535-f&5 Phillips, D. R. (1980) in Progress in Hemostasis and Thrombosis

(Spaet, T. H., ed) Vol. 5, pp. 81-109, Grune and Stratton, New

19.

20. 21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33. 34.

35

36.

37.

38. 39.

40. 41. 42.

43.

44.

45.

46.

47.

48. 49.

5944-5949 George, J. N., Potterf, R. D., Lewis, P. C., and Sears, D. A. (1976)

J. Lab. Clin. Med. 88, 232-246. Phillips, D. R. (1972) Biochemistry l&4582-4588 Phillips, D. R., and Agin, P. P. (1977) Biochem. Biophya. Res.

Common. 75,940~947 Gahmberg, C. G., and Andersson, L. C. (1977) J. Biol. Chem. 252,

5888-5894 Phillips, D. R., and Agin, P. P. (1974) Biochim. Biophys. Acta

352,218-227 McEver, R. P., Baenziger, N. L., and Majerus, P. W. (1980) J.

Clin. Invest. 66, 1311-1318 Hagen, I., Bjerrum, 0. J., and Solum, N. 0. (1979) Eur. J.

Biochem. 99,9-22 Hagen, I., Nurden, A., Bjerrum, 0. J., Solum, N. O., and Caen, J.

(1980) J. Clin. Inuest. 66, 722-731 Shulman, S., and Karpatkin, S. (1980) J. Biol. Chem. 255,4320-

4327 Kunicki, T. S., Pidard, D.. Rosa, J.-P., and Nurden, A. T. (1981)

Blood 56,268-278 Leung, 2. L. K., Kinoshita, T., and Nachman, H. L. (1981) J. Biol.

Chem. 256,1994-1997 Martin, R. G., and Ames, B. N. (1961) J. Biol. Chem. 236,1372-

1379 Phillips, D. R., and Jakabova, M. (1977) J. Biol. Chem. 252,5602-

5605 Truglia, J. A., and Stracher, A. (1981) Biochem. Biophys. Res.

Commun. 100,814-822 Clarke, S. (1975) J. Biol. Chem. 250.5459-5469 &m&son, K. J., Capita&, A., and Liischer, E. F. (1979)

Biochim. Biophys. Acta 553, 11-24 Kundig, W., and Roseman, S. (1971) J, Biol, Chem. 246, 1407-

1418 Nozaki, Y., Schechter, N. M., Reynolds, J. A., and Tanford, C.

(1976) Biochemistry l&3884-3890 Banker, G. A., and Cotman, C. W. (1972) J. Biol. C&m. 247,

5856-5861 Neer, E. J. (1974) J. Biol. Chem. 249,6527-6531 Costrini, N. V., Kogan, M., Kukreja, K., and Bradshaw, R. A.

(1979) J. Biol. Chem. 254, 11242-11246 Cabral, F.. and Schatz, G. (1978) J. Biol. Chem. 253,4396-4401 Laemmli, U. K. (1970) Nature (Land.) 227, 680-685 Adair, W. S., Jurevich, D., and Goodenough, U. W. (1978) J. Cell

Biol. 79,281-285 Tanford, C. (1961) in Physical Chemistry ofMacromolecules, pp.

364-396, John Wiley and Sons, Inc., New York Barber, A. J., and Jamieson, G. A. (1970) J. Biol. Chem. 245,

6357-6365 Oknmura, T., and Jamieson, G. A. (1976) J. Biol. Chem. 251,

5944-5949 Jennings, L. K., Fox, J. E. B., Edwards, H. H., and Phillips, D. R.

(1981) J. Biol. Chem. 256,6927-6932 Bennett, V., and Stenbuck, P. J. (1979) J. Biol. Chem. 254, x%3-

2541 Jon=n, S., and Kronvall, G. (1974) Eur. J. Immunol. 4.29-33. Brown, J. P., Wright, P. W., Hart, C. E., Woodbury, R. G.,

Hellstriim, K. E., and Hellstrtim, R. D. (1980) J. Biol. Chem. 255,4980-4983

York Phillips, D. R., Jennings, L. K., and Edwards, H. H. (1980) J. Cell

50. Greenwood, F. C., Hunter, W. M., and Glover, J. S. (1963) Bio- them. J. 89, 114-123

Biol. 86, 7’7-86 Phillips, D. R., and Agin, P. P. (1977) J. Biol. Chem. 252, 2121-

51. Lowry, 0. H., Rosehrough, N. J., Farr, A. L., and Randall, R. J. (1951) ?J. Biol. Chem. 193, 265-275

ZlZti 18. Okumura, T., and Jamieson, G. A. (1976) J. Biol. Chem. 251,

52. Dulley, J. R., and Grieve, P. A. (1975) Anal. Biochem. 64, 136- 141

10464 Platelet Membrane Glycoproteins IIb and 111

Platelet Prepdratlon

but less thdn 14 days fmn renlrectlon) were gemmusly donated by the Mld-South Blmd COIlectlO" Center. MemDhll. Tennessee. I" a tYDICI1 DRDlnt lo" . 100 D la t r l e t

Concentrates of h a n plate le ts prepared for c l I n l c r l tranrfuslon (me t M n 3

Glycopmteln Purlf lcatlon

m d l f fe ren t la l ex t rac t lm of p la te le t plasm &rare wlth the nonlonlc detergent

c h m a t o g r a m In v a ~ l o u s Trlton x-IW-contalnlng bufferr. M r r m fm plate le t

rcn l rect lon) were used I n t h l s study because glycopmtelns I l b and I l l I n OutdatCd p late le ts or fresh platelets had Identlcal electrophowtlc pmpertles m both "amen- slondl SOS gels and nonreduced-reduced tm-dlmenrlanal SOS gels (data mt shorn). To extract glycoproteln I l b and glycoproteln I l l d r d n c s f m 100 platelet mncentrates prepared with I T S buffers were suspended I n 26 nl O f E I S buffer contalnlng 0.11 Trlton X-IW and incubated m Ice fop 30 m n . The suspension was centr l fvged at 2B.wO r p Type 64 I1 m t o r ) for 1 hr at 4'C. The pel le t conta lnd g lycopmtelns I lb and Ill

the suspension was incubated on ice for I h and centr l fugcd at 28.000 qm (Type €4 T I F lg . I . Lane 3) and was suspended I n 25 mI Of ET buffer mntalnlng 0.11 Tr l ton X-100;

m t o r ) for I hr at 4'c. m e Trlton x-IW extract ~n fT buffer mntalned glycopmtelns I l b and 111 (Flg. 1. Lane 4 ) . A feature Of th ls extraction. AICh resulted I n a marked

wIth I l l t o n X-IO0 i n the lb leMe of added sa l t a t 0-C but were mt IOIubI l lmd by I r l t o n PYrlflcatiOn. 111 that both glycopmtclns were solubll lzed k n mxbraner were treated

X-100 I n the presence of 0.1 M sodlm chlorlde at O T . Although the effect of temperature m g lycepmte ln so l~b l l l za t lon was mt systmatlcal ly studled we obscned that Mth glycopmtelns were solub l l l red by the 0.1 M MC1-contalnlng bufier at temperatures hlgher thdn that obtdlnd fm Imrrlm of the sample In w l t l n g Ice. l c t l n and y o r l n were two other mjor pmtelns In Isolated b r a n e l ( Ident l f led by Crlterla prerlourly establ l lhed (16.46) and they m a l n c d l r l ton- Insolub le In both hlgh and 10" sa l t Concentratlonl.

Pw l f l ca t l on of glycoprotelns I l b and I l l fm h u r n plate le t d r a n e s us based

TrltOn X-IW I n the presence and absence O f added sal t . fo l lorcd by mlm

COncentvdtes that were w t d d t M for cl ln lcal t ransfur lon (mre t M n three days fm

extract *.I applled t o a col- (4.7 x 92 (II) of SepMcryl S-300 equll lbrated wlth 0.05 Solubl l tzed gl jmpmteln U S subjected to gel f l l t r a t l o n Chmatography. me ET

M %dl- phosphate contalnlng 0.051 l r l t o n x-100. pH 7.4. and u s eluted at a flow ra te Of 25 ml lh (Flg. 2). COlm ch-tWJraphlC p m c e d u ~ s were performed at 22 t 2.C. 611 f rac t l an r were collected In 13 x I00 m glass tubes. Each f ract lon US mnltored by absot'bance meaw-nts at 280 n. uslng the mlm buffer as the reference. and then anslyzed by SDS-polyacrylmlde gel elwtmphorcrlr. Glycopmtclnr I lb and 111

denr l tme t r l c scan$ O f Y)S gels under el ther Rduclng or mnreduclng condltlonl. glYC0- coplr l f led at an apparent constant molecular ra t l o by t h l l pmcedure. Bared on

glycopmtelnr I lb and I l l bound to D U E Sephlcryl I n a buf fer mnt l ln lng 0.51 Trlton. pmte ln r I l b and 111 con$tl tuted 961 to 981 O f the pmte ln I n the pooled fraction. Both

0.05 M r o d l u phosphate. pH 7.4. and m e l u t e d b e t r t n 0.1 and 0.25 M Iodlyn chlorlde. but since Increased Durlflcdtlon us mt dDDarent. Ion exChan9e chraatwraDhY U s mt Included i n the f l M 1 p l r l f l c a t l m S C M .

CDclutlon of g l ympmte lns I l b and I l l by the above chmatographlc pmcedure

. .

suggested that there glycopmtclns exlsted as I cmplex In the Tr l ton X-100 r$~raCt lOn buffer. Although wbreqvent e z p r l m t r s h w d t h a t t h l s m p l e x requlred Ca , add l t lm O f 5 mM EOTA. rh lch dlSlOCldteS t h l l cmplex (ILC Results). Itlll did M)t

cbrened. however. thdt a d d l t i m o f 2 M urea. 6 M guanldlne hydrochlorldc. or 0.51 SDS peml t Ypara t lon Of the two glycopmtelns by chrmtographlc pmcedures. It u s

t o the 0.051 TrltOn X-100 buffer a l l o r d for the Kpl rd t lon Of g lycopmteln I lb fmn glycopmteln I l l by Sephacryl S-NO ch-tqraphy. Fractions enrlched r l t h glycopmtelnr I lb and I l l fm four prepardtlons descrlbed I n Flgure 1 were pooled and concentrated by lyaphl l lzat lan. The l yoph l l l r ed rmp lc US dlssolred I n 50 nl Of 2 M w e d and aw l led t o a c o l m (2.5 x 170 a) o f Sephacryl S-300 equl l lbmted wlth 0.05 M rodlm phosphate contalnlng 0.051 Trlton X - 1 0 0 and 2 M urea. pH 7.4, and eluted at a

enrlched I n elthe? glycopmteln I lb or glycopmteln 111 were pwled. dldlyzed versus flew ra te of 15 mI/h. Flgure 3 IhO- SOS gels O f the eluted pmteln. FrdCtlonl

d l a l y s i l buffer. and then concentrated under presswe to 6 "1 urlng I 11110 F l l t e r

t o a C O I M ~ (1.5 x 100 n) Of Sephacryl S - M o equll lbrated r l t h 0.05 M sdlu ln phosphate ( k l c a n Corpmt lon. Lerlngton. M) . The enrlched glycopmteln I lb smplc u s reaw l led

contdinlng 0.051 Tr l ton 1-100 and 2 M u r n . pH 7.4. Fractions contalnlng glympmteln I l b were pmled. cmcentmted to 6 .I and reapplled to the same Sephrcryl S-yx l colmn.

de ten lmd by Sm gels. Reduced gels shored that plrlfled glycoprotein I I conllsted Of The peak f rac t lmr fm t h l l mlm Contdlnd apomxl ldtely 97s glycoproteln I l b as

tm chalnl. glycopmteln I l b and glycoproteln I lbe. Ib effort was made t o Separate these dlrulf lde-llnked rubunhr.

Ilb-

111-

40 45 40 50 55 57 60 65

Flg. 3. h r l f l c l t l o n Of glycoproteln I l b by gel f l l t r a t l o n chrmatography In 2 M urea. Pooled peak fractlonr fm four prepdratlons (F lg . 2 ) were Iyophl l l led. d lssolred I n 2 II Urea to 1/10 O f the Or ig ina l w lme. ad applled to a calm of Scphacryl S-Mo (1.5 x IW an1 e w l l l b r d t e d d t h 0.05 M sodlm Dhosohdte Contdlnlna 0.051 l r l t o n X-100 and 2 M

Platelet Membrane Glycoproteins IIb and 111 10465

rere Waled and rechmatcqraphed on the S m c o l m . glycoprotein 111 could mt be Yhem tk glycoprotein I11-enriched fractions fm the Trlton urea c o l m (Fig. 3)

separated cmplete ly fm glycoprotein Ilb. Further. *hm the glycoprotein I l l -enrlched

hydrochlorlde. 0.05 M sodlm phosphate. pH 7.4, glycoprotein I l b was present as a fractlonr were ChvmatOgraphCd on a c o l u n contalnlng 0 . 0 1 Triton. 6 M guanidine

disulfide reduction muld allow for bct ter separation. fractions cnrlchd I n contanlnant throughout the g lympmtein Ill peak (data not shorn). To d c t c n l n d w t h e r

glycopmteln 1 1 1 frm the f i r s t urea c o l m e r e poled. Incubated for 1 h a t lnbient tmpemt~re r l t h I1 s-mercdptoethaml and 6 M guanldlm hydrochloride. and applied t o a COImn 11.5 x 1W cn) Of Sephacryl S - D I equil ibrated with 0.05 M s c d m phOsphate Contalnlng 0.05% Trlton X-100. 6 II guanidine hydrochloridc. and 5 M onwrcaptoethaml.

buffer as the reference. and by 50s-wlyacrylmlde gels after the f lact ions were dla- pH 7.4. Fractlonr were analyzed by abwwbdnce measu-nts at 281 m, using the c o l u n

wl th 2% B~ercaptoethaml. Glycoprotein Ill eluted as a h m g e m u s peak near the lyzed versus d i d l y I l I buffer t o m v c the guanidine hydrochlorlde. and were reduced

v o l m of th is co lmn precedlng the e lu t lon po l l t lon Of glycoprotein I lb. resultlng in

additiondl reduction (data mt shorn1 revealcd that olwoomteln I l l "as disu l f ide the pvrlflcdtlm of glycoproteln 111 (Fig. 4). SO5 gels Of the eluted frdctlons without

"". .

1 1 4 I

48 50 52 54 56 58 60 62

J

Flg. 4. Purification of glycoprotein 111 by gel filtration chroutograply in 6 II guanidlm hydrochloride. Fractions enriched i n glycoprotein 111 fm the init lal m a co lun ( f rac t ions 40 through 52 in Fig. ?A) were concentrated to 1/10 r o l u e and dldlyzed versus 0.01 M sodim phorylate buffer containing 0.051 TvIton X-100. pH 1.4. Guanldlne hydmchlorldc ( 6 M) and )-.rrrcaptathaml (1%) .Med and the solut ion Incubated for 1 h a t 2 2 7 . This solution us applied t o a colm of scphacryl S - 3 w e w l l l b r a t c d r l t h 0.051 Triton X-100. 6 M guanidine hydmchlorldc. 5 .)I o-rrcapto- ethdnol. 0.05 M s o d l u phosphate. pH 7.4. Fractlonr e m analyzed by 7.5% p o l ~ a c ~ ~ l m l d c gels d e r reducing mndltlons.

P1.tclet hbrmes d 8 9 0.11 Trite" x-IW Extract 120 21 28

--- loo 3.3 98

Sephacryl 5-300 Co lun Peak 27 41 49 5.8 39

bp I l b Purification:

I n I T buffer

Trlton-urea calm I Triton-Urea Colun 2

19 66 8.1 18 4.2 91 11

Trlton-Urea Column 3 1.4 91 12 12 4.3

CP 111 Pur l f lcdt ion: Trlton-urea Colyn 1 Gumldine Wdmchloride 2.6

9.2 98 61 1.0 I1

12 7.4

TAeLE I 1

h l m , k l d ColporltiomS Of Glycoprotein I l b and GlYCOpT'Otdn 111'

hino k i d - GP I l b - CP l l l C Lysine 4.15 6.03 Hist id ine 2.19 AP4l"iW

Aspdrtic k i d Glutami~ k i d 12.48

1.13

4.82 4.39 8.84 10.11

11.86

Thmnine 4.25 Serin Proline Glyclnr A1 ani ne

5.80 9.06 9.22 6.89

10.54 5.66 9.52

1.32 6.05

Methionine valine 7.99 6.84

Isolcucin 2.81 1.21 1.85

4.39

Leucine 11.21 9.30 rymsine 2.78 3.22 Phenylalanine 3.45 3.95

umssed as .OILS/IOO d e s total a im acfds. Avevage of 3 detenlnations. AVer.gc O f 5 deteniMtiOllS.

1 2 3

Flg. 5. Exponmtial 5 t o MI acrylmidc-sM gcl of p l r i f i e d glycoprotein LID fm Fig. 3 (Lm 1). t k miztum frm Flg. 2 (Lane 2). and glycoprotein 111 f m Fig. 4 (Lam 3).

10466 Platelet Membrane Glycoproteins IIb and III