Glycoprotein Ib in the triton-insoluble (cytoskeletal) fraction of blood platelets

Biochimica et Biophysica Acta, 799 (1984) 209-220 Elsevier

209

BBA 21755

GLYCOPROTEIN lb IN THE TRITON-INSOLUBLE (CYTOSKELETAL) FRACTION OF BLOOD PLATELETS N.O. SOLUM and T.M. OLSEN Research Institute for Internal Medicine, Section on Hemostasis and Thrombosis, University of Oslo Rikshospitalet, Oslo 1 (Norway) (Received November 7th, 1983) (Revised manuscript received March 5th, 1984)

Key words: Glycoprotein lb; Cytoskeleton; (Blood plateleO

Glycoprotein Ib could be demonstrated in the Triton-insoluble (cytoskeletal) fraction of platelets prepared with EGTA by SDS-polyacrylamide gel electrophoresis and staining with the periodic acid Schiff's reagent. Crossed immunoelectrophoresis showed that glycoprotein Ib could be extracted from such Triton-insoluble residues when the extraction solution contained 1% Triton X-100 plus 5 mM CaCI2, but not if it also contained leupeptin. This indicates that glycoprotein Ib was associated to structures in the cytoskeletal fraction in such a way that it could be extracted only after activation of a calcium-dependent protease, and degradation of the actin-binding protein was demonstrated. After crossed immunoelectrophoresis of platelet extracts prepared in the presence of leupeptin or EDTA, a glycoprotein Ib-related, rocket-shaped immunoprecipitate was seen originating from the application well. This was interpreted as being related to glycoprotein Ib associated to actin polymers which did not sediment at low-speed centrifugation. Incubation of platelets with 32p as sodium phosphate led to incorporation of phosphatase-sensitive 32p in all of the glycoprotein Ib-related immunoprecipitates except for that of glycocalicin. This supports the idea that glycoprotein lb traverses the plasma membrane and can be phosphorylated at the inner surface whereas glycocalicin represents the terminal part of the glycoprotein Iba-chain exposed at the outer surface.

Introduction Glycoprotein Ib is one of the platelet membrane glycoproteins which have been in the center of interest in recent years. Many studies indicate that the presence of this protein on the platelet surface is required for the normal interaction between platelets and the von Willebrand factor [1-4]. This interaction is again important for adhesion of platelets to structures in the subendothelial tissue of the vessel wall which are exposed after The nomenclature for platelet membrane glycoproteins in this paper is adapted from Phillips and Poh Agin [9] with the exception that glycoprotein III is referred to as glycoprotein Ilia. 0304-4165/84/$03.00 © 1984 Elsevier Science Publishers B.V.

vessel wall damage [5,6]. Certain studies also indicate that glycoprotein Ib may serve as a receptor for the von Willebrand factor [7,8]. Glycoprotein Ib consists of two polypeptide chains termed Iba and Ibfl [9]. Glycocalicin is considered to represent a proteolytic split product of the a-chain [10-12] formed after activation of a calcium-dependent protease during platelet lysis [13]. It is a highly glycosylated molecule with as much as 60% of its weight consisting of carbohydrate [14]. The carbohydrate moiety has been characterized extensively [14-16]. Whereas glycocalicin behaves as a typical hydrophilic protein, glycoprotein Ib shows the characteristics of an amphiphilic protein both in crossed hydrophobic interaction immunoelectrophoresis [10] and

210

charge shift crossed immunoelectrophoresis [11]. We recently demonstrated that when platelets were solubilized in Triton X-100 in the presence of inhibitors of calcium-dependent proteases, the extracts contained a glycoprotein Ib-related component of less anodal mobility than the regular glycoprotein Ib observed in the absence of such inhibitors [17]. Further, under these conditions a significant amount of the glycoprotein Ib-related material did not move out of the application well during the first dimension electrophoresis [17]. This indicated that some association of glycoprotein Ib to polymeric or insoluble material might exist. The purpose of the present work has been to study this possibility further, and to discuss it in relation to the possibility that glycoprotein Ib may be associated to cytoskeletal structures inside the plasma membrane. Materials and Methods

Commercial materials. Triton X-100, EGTA and alkaline phosphatase (type I from calf mucosa) was from Sigma Chemical Company, St. Louis, U.S.A. EDTA from Merck AG, Darmstadt, F.R.G., N-ethylmaleimide from Schwarz/Mann, New York, U.S.A., the Schiff's reagent (Feulgen) from Raymond A. Lamb, London, U.K., and Insta-Gel II from Packard Instruments, Co., IL, U.S.A. Leupeptin was from Protein Research Foundation, Japan and molecular weight protein standards from Bio-Rad, CA, U.S.A. Washing solutions and standard buffers. These contained 148 mM NaCI, 5 mM glucose, 0.6 mM E D T A and 20 mM Tris-HCl (pH 7.4, 280 mosM) as washing solution for platelets and cytoskeleton, 148 mM NaC1 and 20 mM Tris-HC1 (pH 7.4) as Tris-buffered saline for resuspension of washed platelets, and 38 mM Tris, 100 mM glycine (pH 8.7) and 1% (v/v) of Triton X-100 as standard extraction solution for crossed immunoelectrophoresis. Human platelets. These were obtained from fresh blood drawn in 1 / 1 0 the final blood volume of a solution of 45 mM EDTA and 120 mM NaC1 in plastic tubes. Platelet-rich plasma was obtained from this by centrifugation at 320 × g for 15 rain. The platelet-rich plasma in each tube (approx. 15 ml) was transferred to a conical plastic tube and

the volume made up to approx. 45 ml with washing solution. Thereafter, the platelets were sedimented by centrifugation for 15 min; the first 5 min at 320 × g to preferentially sediment erythrocytes and the last 10 min at 2000 x g to sediment the platelets. In each washing the platelets of each tube were resuspended in approx. 45 ml washing solution by sucking and blowing using a Pasteur pipette discarding the red layer at the bottom of the tube. This was followed by centrifugation as above. After the third washing the platelets were resuspended in an accurately measured volume of the washing solution (or Tris-buffered saline if traces of EDTA in the final platelet suspension were to be avoided), and the platelets were counted in a Thrombocounter-C (Coulter Electronics, Ltd., U.K.). After the last centrifugation, the platelets were either resuspended in the Tris-buffered saline given above, or used directly for extraction in the Triton X-100-containing extraction solution for crossed immunoelectrophoresis as described below, or used for preparation of the cytoskeletal fraction also as described below. Centrifugations and washings were performed at 4-10°C. Alternatively, platelets were prepared from 1-day-old platelet concentrates as described previously [17]. Triton X-IO0 extracts. For crossed immunoelectrophoresis these were prepared by addition of the standard extraction solution to platelet sediments. When leupeptin or EDTA was present, these had been added to the extraction buffer prior to use. If necessary, these solutions were readjusted to pH 8.7. Extraction was performed at room temperature by sucking and blowing using a Pasteur pipette, followed by centrifugation in an Eppendorf centrifuge model 3200 at 8000 × g for two periods of 2 min. For studies of the glycoprorein Ib-related, rocket-like immunoprecipitate at the application point after crossed immunoelectrophoresis, freshly prepared extracts which had not been frozen and thawn, were always used. The cytoskeletal fraction of platelets. This fraction was prepared according to Jennings et al. [18] except that a higher concentration of platelets was used. Washed platelets resuspended to 101° platel e t s / m l in a solution consisting of 138 mM NaC1, 2.9 mM KC1, 12 mM NaHzCO 3, 0.36 mM Na3PO 4, 5.5 mM glucose and 1 mM EDTA at pH 7.4 were mixed at room temperature with an equal

211 volume of a solution of 2% Triton X-100, 10 mM EGTA and 100 mM Tris (pH adjusted to 7.4). Extraction was performed by sucking and blowing the mixture using a Pasteur pipette and terminated by centrifugation in the Eppendorf centrifuge at 8000 x g for 2 × 2 min. The Triton-insoluble residue thus obtained was defined as the cytoskeletal fraction. This was washed as stated in legends to the figures. Antisera. Antisera to glycocalicin/glycoprotein Ib, to whole platelet antigens, and to the glycoprotein l i b / I l i a complex were prepared in rabbits as described in detail elsewhere [10,19,20]. The glycocalicin used for immunization was purified as described in Ref. 10 with the important modification described in Ref. 17. This antiserum was adsorbed with a soluble fraction obtained by freezing and thawing of washed platelets in the presence of EDTA which was added to prevent the formation of the soluble glycocalicin from membrane-bound glycoprotein Ib [10]. Usually, the antiserum was used as such for immunoelectrophoresis without isolation of the immunoglobulin fraction (100-150 /~1 antiserum per ml agarose). The anti whole platelet antiserum was prepared against washed platelets suspended in 0.15 M NaC1. The total material obtained after one cycle of freezing and thawing was used for the immunization. The immunoglobulin fraction was isolated and concentrated after precipitation by addition of 25 g ( N H 4 ) 2 S O 4 to 100 ml antiserum. Finally, this fraction was dialyzed against 0.07 M acetate buffer (pH 5.0) and used for immunoelectrophoresis at 4-6 mg per ml agarose. The antiserum to the glycoprotein l i b / I l i a complex was prepared by injections of the agarose containing the glycoprotein l i b / I l i a immunoprecipitate after crossed immunoelectrophoresis of Triton-solubilized platelet membranes against the anti whole platelet antibodies after the agarose had been cut out of the gel [20]. The antiserum was used as such for immunoelectrophoresis after one third dilution in the agarose. Antiserum to human serum albumin was from Behringwerke AG, F.R.G. Thrombin, bovine, purified to electrophoretic homogeneity was a kind gift from Janson and Johnsen, AS Thrombo, Oslo, Norway. Crossed immunoelectrophoresis in the presence of 1% Triton X-100 was performed as described previously [10,19].

Rocket immunoelectrophoresis was performed according to Laurell [21] as modified by Weeke [22].

SDS-polyacrylamide gel electrophoresis on cylindrical gels of 5% acrylamide. This was performed essentially as previously described [23]. However, the samples were heated on a boiling water bath for 5 min, and the unreduced samples contained 6 mM N-ethylmaleimide. The cylindrical gels were used for studies of glycoproteins which were stained by the periodic acid Schiff's reagent [23] and scanned in a modified Beckman CDS-100F densitometer at 550 nm [23]. Slab gel electrophoresis with 3% stacking gel and 7% separating gel with staining by Coomassie Brilliant Blue O was performed by the method of Laemmli [24] essentially as modified by Le Stourgeon and Beyer [25]. Phosphorylation with S2p. This was performed by incubation of washed platelets suspended in Tris-buffered saline (pH 7.4) at 1.5.101° platelets/ml with 80 MBq carrier-free [32p]sodium phosphate for 60 min at 37°C in plastic tubes which were turned approx. 25 ° angle from the horizontal position at a constant rate of 34 times per min on a motor-driven tilting device. Thereafter the platelets were sedimented by centrifugation in the Eppendorf centrifuge at 8000 x g for 2 rain, washed twice, and extracted in 1% Triton X-100 for crossed immunoelectrophoresis as described above. After the crossed immunoelectrophoresis of the 32p-containing extracts, the gels were washed 3 times by pressing and swelling in 0.154 M NaC1 to remove non-precipitated proteins. Thereafter, the immunoplates were exposed to X-ray films (Curix RP 1L, Agfa-Gevaert safety film) using a Siemens Titan intensifying screen in a casette stored at -80°C. Under the conditions used, a picture of the GP Ib-immunoprecipitate developed on the film within a few hours. Optimal conditions required exposure for 1-3 days. For removal of [32p]phosphate from the immunoprecipitates by alkaline phosphatase, immunoplates were run in duplicates. Before exposure to the X-ray films one of the plates was incubated in 5 ml of the standard Tris-glycine buffer pH 8.7 (without Triton) containing 1 m g / m l of alkaline phosphatase for about 18 h at room temperature, whereas the other plate was incubated in buffer only. Thereafter, the plates were washed twice as

212

above, and put on x-ray films. Removal of .~2p was also controlled by measuring the radioactivity released into the incubation media. For this, aliquots from the media were counted in a Tri-Carb 460 CD liquid scintillation system (Packard Instruments Co., IL, U.S.A.) using Insta-Gel II from the same company as liquid scintillation cocktail. Total protein. This was determined by the BioRad assay essentially as described by Bradford [26] using human serum albumin as standard. Before determination of total protein in Triton-insoluble fractions or whole platelets, these were solubilized in 1% (w/v) sodium deoxycholate in 0.2 M NaOH. Results

Extraction of unstimulated platelets in 1% Triton X-100 is considered to result in almost total solubilization of the platelet proteins. Even so, a Triton-insoluble residue is always obtained after low-speed centrifugation of such extracts (8000 × g for 4 min). In our studies using washed platelets and a Tris-glycine buffer pH 8.7, this sediment amounted to 340/~g/109 platelets or 13.4% of the total platelet protein (13.4 + 3.0, n = 6). When leupeptin (4.2 mM) was present during solubilization as an inhibitor of calcium-activated proteases, the amount of insoluble residue increased significantly (to 20.5%, i.e. 20.5 + 2.2, n = 6). The uppermost Coomassie Brilliant Blue staining band ( M r 250000) seen on SDS-polyacrylamide gel electrophoresis of reduced whole platelet samples, which is considered to represent the actin-binding protein [27], was not seen on electrophoresis of the Triton-extracts of platelets when these had been prepared in the absence of inhibitors of the calcium-activated proteases (data not shown). This indicates that the actin-binding protein was split by proteolysis during platelet lysis in the absence of such inhibitors. The actin-binding protein is known to cross-link and stabilize actincontaining microfilaments [28,29]. Crossed immunoelectrophoresis of platelet extracts prepared with 1% Triton X-100 in plain buffer using antiserum to glycocalicin showed a continuous immunoprecipitate consisting of two peaks corresponding to glycocalicin and glycoprotein Ib (Fig. 1A). If inhibitors of the calcium-de-

TX WP #

•

•

p~

EDTA I

Fig. 1. Effect of EDTA during solubilization of platelets in 1% Triton X-100 as to the extraction of glycoprotein lb-related material. Washed platelets were extracted with 1% Triton X-100 in the standard buffer at pH 8,7 (101° cells/ml) in the presence or absence of EDTA (3.6 mM), Crossed immunoelectrophoresis with antiserum to glycocalicin (GC) of extracts prepared (A) in the absence and (B) in the presence of EDTA. Note that the area covered by the glycoprotein Ib-related immunoprecipitates in B represents 47.5% of that in A.

pendent proteases were present during solubilization, the appearance of the immunoprecipitate became clearly different (Fig. 1B). Then the bulk of the glycoprotein Ib-related material was seen as a third, more anodal peak, and as a rocket-shaped immunoprecipitate at the application well with the glycocalicin and the regular glycoprotein Ib-peaks clearly reduced (Fig. 1, see also Ref. 17). Also the total area covered by the glycoprotein Ib-related immunoprecipitates was reduced compared to that observed with an extract prepared from the same number of the same platelets without inhibitor, indicating that less glycoprotein Ib-related protein had been solubilized when the calcium-activated proteases had been inhibited. This is demonstrated in Fig. 1 using EDTA as inhibitor, but could also be seen with other inhibitors of calcium-activated proteases.

213

In studies using the crossed immunoelectrophoresis technique the platelets are routinely solubilized in 1% Triton X-100 at pH 8.7 [19]. If the extraction is done at pH 7.4 in the presence of EDTA or EGTA, the Triton-insoluble residue corresponds to that previously defined as the cytoskeletal fraction [27]. To study this fraction, platelets washed by our procedure were extracted with Triton X-100 in the presence of EDTA and EGTA according to the procedure of Jennings et al. [18] except that we used a higher concentration of platelets to fit our analytical procedures. SDS-

Z Z

DISTANCE FROM TOP OF I~EL (cm)

Fig. 2. The presence of glycoprotein Ib in the cytoskeletal fraction of platelets as demonstrated by SDS-polyacrylamide gel electrophoresis and staining with the periodic acid Schiff's reagent. (Cyt): cytoskeletal fraction and (WP): whole platelets. Washed platelets (5.109 celis/ml) were solubilized in 1% Triton X-100 in the presence of 5 mM EGTA and 0.5 mM EDTA at pH 7.4. Triton-insoluble residue from 10 TM platelets was washed once in the EDTA-containing washing solution, dissolved in 0.9 ml of SDS-N-ethytmaleimide or SDSmercaptoethanol solutions, heated at 100°C for 5 rain, and subjected to SDS-polyacrylamide gel electrophoresis on cylindrical gels which afterwards were stained as stated and scanned densitometrically. Whole platelets from the same platelet preparation were solubilized and treated identically, and subjected to the same electrophoresis. The material applied to each gel was derived from 1.1.109 platelets for the cytoskeletal fraction and 0.12.109 platelets for the whole platelet samples. Red, reduced samples; Unred, unreduced samples; Opal, opalescence; TSP thrombospondin.

polyacrylamide gel electrophoresis of such cytoskeletal fractions showed the presence of glycoprotein Ib after staining of the gels for glycoproteins with the periodic Schiff's reagent (Fig. 2). The bands of the other well-known platelet glycoproteins were almost invisible, or strongly reduced in intensity relative to that observed with whole platelet samples, using the intensity of the glycoprotein Ib band as a basis of comparison (Fig. 2). Three opalescent bands were clearly seen, one major band corresponding to actin, and two weaker bands in the glycoprotein III region of the gels one of which may represent a-actinin. After isolation and washing of the Triton-insoluble residue in the EDTA-containing washing solution, the residue was suspended in our regular Triton X-100/Tris-glycine buffer at pH 8.7 to which had been added calcium chloride to 5 mM. Crossed immunoelectrophoresis of the supernatant obtained after centrifugation of this suspension showed that associated glycoprotein Ib could be extracted from the cytoskeletal fraction under conditions where residual calcium-dependent protease was activated (Fig. 3). When not only calcium ions, but also leupeptin was present in the extraction buffer, such an extraction of glycoprotein Ib could not be observed, or occurred only to a minor degree (Fig. 4). The use of the anti whole platelet antiserum showed that also some other platelet antigens were present but only in small amounts (Fig. 3) and these were present also if leupeptin was used (Fig. 4). The specificity of the association of glycoprotein Ib was further demonstrated by rocket immunoelectrophoresis according to Laurell [21] using antisera specific to glycoprotein I b / glycocalicin, albumin and the glycoprotein lib-Ilia complex, respectively, and extracts of both the cytoskeletal material and whole platelets. When tested in this system at dilutions which gave similar heights of the rockets formed with antiglycocalicin for both extracts, only small rockets were seen for the other antigens in the cytoskeletal extracts as compared to the whole platelet extracts (Fig. 5). This showed that glycoprotein lb was present in the cytoskeletal fraction, and indicated that this was not due to a general contamination of this fraction by membrane or granule proteins but represented some sort of specific interactions between glycoprotein Ib and some protein(s) in

214

Fig. 3. Demonstration of glycoprotein Ib in the cytoskeletal fraction of platelets by crossed immunoelectrophoresis of Triton X-100 extracts of this fraction prepared in the presence of calcium ions. (A) cytoskeleton extract (CaCI2 present), anti-glycocalicin; (B) whole platelet extract, anti-glycocalicin; (C) cytoskeleton extract (CaCI 2 present), anti whole platelets, and (D) whole platelet extract, anti whole platelets. The cytoskeletal fraction was prepared and washed as in Fig. 2. This residue was then suspended in the Tris-glycine buffer (pH 8.7) which contained 1% Triton X-100 and 5 mM CaCI 2. After stirring at room temperature for 2 min, it was centrifuged at 8000× g for 4 rnin, and the supernatant used as the cytoskeletal extract. For comparisons, a regular Triton X-100 extract of whole platelets was prepared. The extracts were compared by crossed immunoelectrophoresis with antiserum to glycocalicin or whole platelet antibodies. A and C represent material from 1.4.10 s cells, B from 0.2.108 cells and D from 0.6-108 cells.

this fraction. That the extraction of glycoprotein Ib from the cytoskeletal fraction was related to a proteolysis, was demonstrated by SDS-polyacrylamide gel electrophoresis (Fig. 6). First, the cytoskeletal material showed three major Coomassie Brilliant Blue-staining bands with molecular weights corresponding to those of the actin-binding protein (M r 250000), a-actinin ( M r 105 000) and actin ( M r 43000). After extraction in 1% Triton X-100 in the presence of 5 mM CaC12, the bands corresponding to actin and a-actinin were

still seen at the same M r values as before (Fig. 6D and E) whereas the band corresponding to the actin-binding protein was almost totally absent (Fig. 6D and E). In addition, the extract showed two new bands corresponding to the previously defined proteolytic degradation products of the actin-binding protein [32] called HF-1 (M r 190 000) and HF-2 ( M r 90000), respectively (Fig. 6E). This clearly demonstrated that the extraction of the cytoskeletal fraction by the calcium-containing Triton-solution induced a proteolysis, and that this

215

Fig. 4. Effect of leupeptin on the extraction of glycoprotein Ib from the cytoskeletal fraction of platelets by Triton X-100 in the presence of calcium ions as studied by crossed immunoelectrophoresis with antiserum to glycocalicin or antibodies to whole platelet proteins. (A) leupeptin absent, anti-glycocalicin; (B) ieupeptin present, anti-glycocalicin; (C) leupeptin absent, anti whole platelets and (D) leupeptin present, anti whole platelets. The cytoskeletal fractions were prepared and extracted by 1~ Triton X-100 in the presence of 5 mM CaCI 2 as stated in legend to Fig. 3. In B and D also leupeptin (4.2 raM) had been added to the solution used for extraction. All extracts represent material from 1.7.108 cells.

WP

CYT

A: ANTIGLYCOCALICIN

WP

B: ANTI-

ALBUMIN

CYT

WP

CYT

C : ANTI-

GP lib- Ilia

Fig. 5. The amounts of glycoprotein Ib, albumin and the glycoprotein Ilb-lIla complex in calcium-containing extracts of cytoskeletal fraction relative to the amounts in extracts of whole platelets compared at dilutions of the extracts which contained similar amounts of glycoprotein Ib, using Laurell rocket immunoelectrophoresis with specific antisera. WP, whole

led to a degradation of the actin-binding protein giving the same degradation products as obtained b y incubation of isolated actin-binding protein with purified calcium-activated protease [32]. The glycoprotein Ib-containing rocket-shaped immunoprecipitate at the point of application observed on crossed immunoelectrophoresis of extracts of washed platelets with antiglycocalicin antiserum, was observed only when the extracts were prepared in the presence of inhibitors of the

platelet extract (from 1.1.10 7 cells); Cyt, cytoskeletal extract prepared with CaCI 2 (from 4.10 7 cells). (A) anti-glycocalicim (B) anti-albumin, and (C) anti-glycoprotein Ilb-lIla complex. The extracts were obtained as stated in legend to Fig. 3.

216

MF

ABP HF-1

200,000

1/+5.000 716,250

(x-actinin HF-2

92,500

66,200

~5,000

Acfin

A

B

C

D

compared to those from washed platelets (data not shown). Further, if the washed platelets were activated by thrombin to release stored substances from their intracellular granules, this rocket-shaped immunoprecipitate disappeared or became strongly diminished even if the extracts were prepared with leupeptin and by low-speed centrifugation (Fig. 7). This thrombin-treatment was accompanied by an increase of around 45% in the amount of total protein present in the Triton-insoluble fraction. Activation of platelets by thrombin is known to convert G-actin into F-actin [18,30,31]. Crossed immunoelectrophoresis of Triton-extracts of washed platelets prepared with leupeptin using

E

Fig. 6. Demonstration of proteolysis with degradation of the actin-binding protein during extraction of the cytoskeletal fraction of platelets by Triton X-100 in the presence of calcium ions. SDS-polyacrylamide slab gel electrophoresis with staining by Coomassie Brilliant Blue G. (A) molecular weight protein standards; (B) whole platelets; (C) cytoskeletal fraction; (D) insoluble residue after extraction of the cytoskeletal fraction by 1% Triton X-100 containing 5 mM CaCl 2, and (E) the soluble extract after such extraction of the cytoskeletal fraction in the presence of 5 mM CaCl 2. The cytoskeletal fraction was prepared and extracted as stated in legend to Fig. 3. A Bio-Rad molecular weight marker kit was used at 1.5 ~g of each of the following proteins: Myosin (200000), fl-galactosidase (116 250), phosphorylase B (92500), bovine serum albumin (66200) and ovalbumin (45000). ABP, actin-binding protein. HF-1 and HF-2 mark positions of proteolytic degradation products of ABP (see Ref. 32). Glycoprotein Ib stains very weakly with Coomassie Brilliant Blue. B represents 3.75.107 cells; C, D and E each represent 7.107 cells.

calcium-activated proteases (Fig. 1B with EDTA and Fig. 7A with leupeptin). These extracts had been prepared by centrifugation at low speed (8000 x g for 4 rain). Centrifugations in the ultracentrifuge (100000 × g for 60 min) sometimes sedimented all the material responsible for the rocket-like immunoprecipitate, sometimes reduced the amount significantly without total removal (data not shown). The material which was sedimented by ultracentrifugation, was enriched in cytoskeletal proteins as could be observed by SDS-polyacrylamide gel electrophoresis (data not shown). The rocket-like immunoprecipitate at the application point was less pronounced with extracts prepared from platelets which had been sedimented directly from platelet-rich plasma

¢o-t,ol

tit GP~

: !i~

Fig. 7. The effect of thrombin-treatment of platelets on the immunoprecipitate at the application point after crossed immunoelectrophoresis of Triton-extracts of platelets prepared in the presence of leupeptin and run against anti-glycocalicin. (A) extract of untreated platelets, and (B) of thrombin-treated platelets. Washed platelets suspended in Tris-buffered saline (pH 7.4) were incubated with 5.8 NIH U / m l of thrombin at 37°C for 5 min, whereas the control platelets were incubated without thrombin. Thereafter, the platelets were washed twice and extracted in 1% Triton X-100 in the presence of 4.2 mM leupeptin and 6 ATU/ml hirudin (10 l° cells/ml). Crossed immunoelectrophoresisof the incubation media with anti whole platelet antibodies showed that the thrombin-treated platelets had been activated to undergo the release-reaction. III, see Ref. 17.

217

Fig. 8. 32p in glycoprotein Ib-related immunoprecipitates after crossed immunoelectrophoresis of Triton X-100 extracts of washed platelets which had been incubated with 32p as sodium phosphate at pH 7.4. The following combinations of antisera and detection techniques were used. (A) anti-glycocalicin; Coomassie Brilliant Blue staining; (B) anti-glycocalicin, autoradiography; (C) anti whole platelets, Coomassie Brilliant Blue staining, and (D) anti whole platelets, autoradiography. GC, glycocalicin. A and B represent the same plate, as do C and D.

antiserum to albumin did not reveal a corresponding immunoprecipitate at the application point indicating that this was not due to a general precipitation phenomenon affecting all platelet antigens (data not shown). Incubation of platelets with [32p]phosphate followed by extraction of the platelets in the Tritoncontaining buffer in the presence or absence of leupeptin and crossed immunoelectrophoresis showed that all of the glycoprotein Ib-related immunoprecipitates were radioactively labelled except for that representing glycocalicin (shown in Fgi. 8 with an extract prepared in the absence of inhibitors). This radioactivity could be removed almost totally by incubation of the immunoplates with alkaline phosphatase at pH 8.7 (data not shown).

Discussion

The data presented here are interpreted by the hypothesis that glycoprotein Ib exerts a tendency to adsorb to actin polymers. The polymeric state of these filaments is affected by the presence of the actin-binding protein which is split by the calcium-activated proteases present in platelets [27,32,33] during extraction of platelets in 1% Triton X-100 in plain buffer. This explains, at least in part, the increased amount of protein found in the Triton-insoluble fraction when extraction is performed in the presence of leupeptin, as well as the reduced area covered by the glycoprotein Ib-related immunoprecipitates after crossed immunoelectrophoresis of extracts prepared in the presence of EDTA or leupeptin as compared to those pre-

218 pared without inhibitors. According to this idea the glycoprotein Ib-related, rocket-shaped immunoprecipitate located at the application point after crossed immunoelectrophoresis of such extracts, reflects association of glycoprotein lb to actin-containing polymers which are too large to move in the first dimension electrophoresis but too small to sediment at low speed centrifugation. The presence of this high molecular weight material may reflect some sort of activation of the platelets during washing. Thus, the corresponding immunoprecipitate was much less pronounced when the extracts were made from platelets prior to the washing. The absence of this material from extracts prepared from thrombin-treated platelets may be explained as an increased polymerization of the actin-filaments leading to a greater tendency to sediment during low-speed centrifugation. This idea also fits the observation that thrombin-treatment of platelets increases the amount of protein in the Triton-insoluble fraction as observed here and in previous studies [34]. Association of myosin with actin filaments after thrombin-treatment [35,36] may also be a factor of importance in this respect. The association of radioactive phosphate with the glycoprotein Ib-related immunoprecipirates may reflect a phosphorylation of this protein at the inner surface of the plasma membrane indicating that glycoprotein Ib penetrates the membrane and is available at its cytoplasmic surface. The observation that the glycocalicin part of the immunoprecipitate was not labelled, supports the idea that glycocalicin represents the terminal end of the glycoprotein Iba-chain exposed on the external surface of the plasma membrane. The present data do not tell whether the association of glycoprotein Ib to actin polymers is established during the actual extraction or reflects associations in the intact platelet. However, the initial effect of dibucaine on the agglutination of human platelets by bovine yon Willebrand factor reported previously [17,37] may be explained as the result of intracellular proteolysis of glycoprotein Ib-associated material in the intact platelet [17,38] and thus indicate a biological significance of the observation. In the following, some aspects of these ideas are discussed in more detail and in relation to previous reports. Only few reports exist on the association of

platelet membrane glycoproteins to the cytoskeletal fraction. Phillips et al. [34] using SDSpolyacrylamide gel electrophoresis primarily looked for such associations after thrombin-treatment, and did not report any detection of glycoprotein Ib in the cytoskeletal fraction. However, the techniques used (e.g. 125I-labelling by lactoperoxidase and Coomassie Brilliant Blue staining) are far from optimal for detection of glycoprotein lb. Rotman et al. [39] described a protein of the cytoskeletal fraction which bound wheat germ agglutinin after SDS-polyacrylamide gel electrophoresis in the reduced state, and claimed that this was glycoprotein Ia based on its electrophoretic mobility. Both glycoprotein Ia and glycoprotein Ib, as well as glycocalicin, bind wheat germ agglutinin [40]. Zucker and Masiello [41] recently reported that a membrane protein which could be 125I-labelled by lactoperoxidase was present in the cytoskeletal fraction of unstimulated platelets in as high an amount as 56.2% of its total amount. Their values for the apparent molecular weight for this protein from SDS-polyacrylamide gel electrophoresis were 158000 in the reduced state and 170000 in the unreduced [41]. The corresponding values for glycoprotein Ib are 143 000 and 170 000, respectively [9]. Variations in shape and microspike formations and thus in actin polymerization due to differences in the platelet isolation techniques may be important for the amount of protein found in the Triton-insoluble fraction. In addition, it has been shown that the composition of the solution used for the extraction is of major importance. Thus, EDTA favours inclusion of myosin whereas increasing concentrations of KCI reduces the amount of actin and actin-binding protein in the Triton-insoluble residue [42]. Certain drugs may lead to depolymerization of the microfilaments [43]. Our value for total protein in the Triton-insoluble fraction after extraction of washed platelets at pH 8.7 in the presence of leupeptin (20.5%) fits the value of 20.7% found by Phillips et al. [34] for extracts prepared at pH 7.4 in the presence of EGTA. It has been shown that a prior incubation of platelets with certain local anesthetics, gives a Triton-insoluble fraction which is granular instead of fibrillar [43]. The local anesthetics are considered to induce a proteolytic splitting of the actin-

219

binding protein, at least at higher concentrations of the drugs [38]. The present studies show that the actin-binding protein is absent from extracts if they are prepared without inhibitors of the calcium-activated proteases. A corresponding observation has been reported previously [27]. If our hypothesis of the adsorption of glycoprotein Ib to actin polymers is correct, such extracts do not contain the high molecular weight material which manifests itself as the glycoprotein Ib-related immunoprecipitate at the application point discussed above, or glycoprotein Ib is not associated to this material in extracts prepared without the proteolytic inhibitors. The present work shows that glycoprotein Ib can be extracted from whole platelets or the cytoskeletal fraction by 1% Triton X-100 in the presence of calcium ions with only a small portion being degraded to glycocalicin (Figs. 1A and 3A) whereas under the same conditions the actin-binding protein is grossly degraded by calciumactivated protease(s) (Fig. 6). However, if platelets are lysed by freezing and thawing or treated with a low concentration of Triton X-100, glycoprotein Ib is almost totally degraded to yield glycocalicin [10,17] and the degradation is prevented by inhibitots of the same proteases. For some reason, the actin-binding protein therefore seems to be more sensitive to proteolytic degradations giving HF-1 and HF-2 [32] than glycoprotein Ib to give glycocalicin. Phosphorylation of the glycoprotein Ib (achain) has been reported previously by Marchesi and Chasis [44] using SDS-polyacrylamide gel electrophoresis to analyze the proteins. From these studies the authors suggested that glycoprotein Ib traverses the bilipid layer of the platelet membrane, bearing reactive groups on both outer and cytoplasmic surfaces [44]. Our data support this idea which again represents one of the basic requirements for a biological role of the association of glycoprotein Ib to the cytoskeletal fraction described in the present paper. One such role might be to regulate receptor interactions [45]. Perhaps such interactions may be involved in the increased von Willebrand factor-induced agglutination of platelets observed during dibucaine-incubations or storage [17,37].

Acknowledgements This work was supported by the Norwegian Council on Cardiovascular Diseases and Anders Jahres fond til vitenskapens fremme.

References 1 Nurden, A.G. and Caen, J.P. (1975) Nature 255, 720-722 2 Jenkins, C.S.P., Phillips, D.R., Clemetson, K.J., Meyer, D., Larrieu, M.-J. and Li~scher, E.F. (1976) J. Clin. Invest. 57, 112-124 3 Solum, N.O., Hagen, I. and Gjemdal, T. (1977) Thromb. Haemostas. 38, 914-923 4 Fauvel, F., Grant, M.E., Legrand, Y.J., Souchon, H.. Tobelem, G., Jackson, D.S. and Caen, J.P. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 551-554 5 Weiss, H.J., Baumgartner, H.R., Tschopp, T.B., Turitto, V.T. and Cohen, D. (1978) Blood 51,267-279 6 Legrand, Y.J., Fauvel, F., Gutman, N., Muh, J.P., Tobelem, G., Souchon, H., Karniguian, A. and Caen, J.P. (1980) Thromb. Res. 19, 737-739 7 Okumura, T. and Jamieson, G.A. (1977) Thromb. Res. 8, 701-706 8 Cooper, H.A., Clemetson, K.J. and L~scher, E.F. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 1069-1073 9 Phillips, D.R. and Poh Agin, P. (1977) J. Biol. Chem. 252, 2121-2126 10 Solum, N.O., Hagen, I. Filion-Myklebust, C. and Stabaek, T. (1980) Biochim. Biophys. Acta 597, 235-246 11 Solum, N.O., Hagen, I. and Sletbakk, T. (1980) Thromb. Res. 18, 773-785 12 Clemetson, K.J., Naim, H.J. and Liascher, E.F. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 2712-2716 13 Phillips, D.R. and Jakabova, M. (1977) J. Biol. Chem. 252, 5602-5605 14 Okumura, T., Lombart, C. and Jamieson, G.A. (1976) J. Biol. Chem. 251, 5950-5955 15 Judson, P.A., Anstee, D.J. and Clamp, J.R. (1982) Biochem. J. 205, 81-90 16 Tsuji, T., Tsunehisa, S., Watanabe, Y., Yamamoto, K., Tohyama, H. and Osawa, T. (1983) J. Biol. Chem. 258, 6335-6339 17 Solum, N.O., Olsen, T.M., Gogstad, G.O., Hagen, i. and Brosstad, F. (1983) Biochim. Biophys. Acta 729, 53-61 18 Jennings, L.K., Fox, J.E.B., Edwards, H.H. and Phillips, D.R. (1981) J. Biol. Chem. 256, 6927-6932 19 Hagen, 1., Bjerrum, O.J. and Solum, N.O. (1979) Eur. J. Biochem. 99, 9-22 20 Gogstad, G.O., Hagen, I., Krutnes, M.B. and Solum, N.O. (1982) Biochim. Biophys. Acta 689, 21-30 21 Laurell, C.B. (1972) Scand. J. Clin. Lab. Invest. 29. Suppl. 124, 21-37 22 Weeke, B. (1973) Scand. J. Immunol. 2, Suppl. 1, 15-56 23 Solum, N.O., Hagen, I. and Peterka, M. (1977) Thromb. Res. 10, 71-82 24 Laemmli, U.K. (1970) Nature 227, 680-685

220 25 Le Stourgeon, W.M. and Beyer, A.L. (1977) Methods Cell Biol. 16, 387-406 26 Bradford, M.M. (1976) Anal. Biochem. 72, 248-254 27 Rosenberg, S., Stracher, A. and Lucas, R.C. (1981) J. Cell Biol. 91,201-211 28 Hartwig, J.H., Tyler, J. and Stossel, T.P. (1980) J. Cell Biol. 87, 841-848 29 Weeds, A. (1982) Nature 296, 811-816 30 Carlsson, L., Markey, F., Blikstad, I., Persson, T. and Lindberg, U. (1979) Proc. Natl. Acad. Sei. U.S.A. 76, 6376-6380 31 Carroll, R.C., Butler, R.G., Morris, P.A. and Gerrard, J.M. (1982) Cell 30, 385-393 32 Truglia, J.A. and Stracher, A. (1981) Biochem. Biophys. Res. Commun. 100, 814-822 33 White, G.C. (1980) Biochim. Biophys. Acta 631,130-138 34 Phillips, D.R., Jennings, L.K. and Edwards, H. (1980) J. Cell Biol. 86, 77-86 35 Daniel, J.L., Molish, I.R. and Holmsen, H. (1981) J. Biol. Chem. 256, 7510-7514

36 Fox, J.E.B. and Phillips, D.R. (1982) J. Biol. Chem. 257. 4120-4126 37 Coller, B.S. (1982) Blood 60, 731-743 38 Nachmias, V.T., Sullender, J.S. and Fallon, J.R. (1979) Blood, 63-72 39 Rotman, A., Heldman, J. and Linder, S. (1982) Biochemistry 21, 1713-1719 40 Clemetson, K.J., Pfueller, S., Li~scher, E.F. and Jenkins, C.S.P. (1977) Biochim. Biophys. Acta 464, 493-508 41 Zucker, M.B. and Masiello, N.C. (1983) Blood 61,676-683 42 Rosenberg, S., Lawrence, J. and Stracher, A. (1982) Cell Motil. 4, 317-332 43 Nachmias, V., Sullender, J. and Asch, A. (1977) Blood 50, 39-53 44 Marchesi, S.L. and Chasis, J.A. (1979) Biochim. Biophys. Acta 555, 442-459 45 Geiger, B. (1983) Biochim. Biophys. Acta 737, 305-341