Dystrophin-related protein in the platelet membrane skeleton. Integrin-induced change in detergent-insolubility and cleavage by calpain in aggregating platelets.

The platelet membrane is lined with a membrane skeleton that associates with transmembrane adhesion receptors and is thought to play a role in regulating the stability of the membrane, distribution and function of adhesive receptors, and adhesive receptor-induced transmembrane signaling. When platelets are lysed with Triton X-100, cytoplasmic actin filaments can be sedimented by centrifugation at low g-forces (15,600 × g) but the membrane skeleton requires 100,000 × g. The present study shows that DRP (dystrophin-related protein) sediments from lysed platelets along with membrane skeleton proteins. Sedimentation results from association with the membrane skeleton because DRP was released into the detergent-soluble fraction when actin filaments were depolymerized. Interaction of fibrinogen with the integrin αβ induces platelet aggregation, transmembrane signaling, and the formation of integrin-rich cytoskeletal complexes that can be sedimented from detergent lysates at low g-forces. Like other membrane skeleton proteins, DRP redistributed from the high-speed pellet to the integrin-rich low-speed pellet of aggregating platelets. One of the signaling enzymes that is activated following αβ-ligand interactions in a platelet aggregate is calpain; DRP was cleaved by calpain to generate a 140-kDa fragment that remained associated with the low-speed detergent-insoluble fraction. These studies show that DRP is part of the platelet membrane skeleton and indicate that DRP participates in the cytoskeletal reorganizations resulting from signal transmission between extracellular adhesive ligand and the interior of the cell.

The platelet membrane is lined with a membrane skeleton that associates with transmembrane adhesion receptors and is thought to play a role in regulating the stability of the membrane, distribution and function of adhesive receptors, and adhesive receptor-induced transmembrane signaling. When platelets are lysed with Triton X-100, cytoplasmic actin filaments can be sedimented by centrifugation at low g-forces (15,600 ؋ g) but the membrane skeleton requires 100,000 ؋ g. The present study shows that DRP (dystrophin-related protein) sediments from lysed platelets along with membrane skeleton proteins. Sedimentation results from association with the membrane skeleton because DRP was released into the detergent-soluble fraction when actin filaments were depolymerized. Interaction of fibrinogen with the integrin ␣ IIb ␤ 3 induces platelet aggregation, transmembrane signaling, and the formation of integrin-rich cytoskeletal complexes that can be sedimented from detergent lysates at low g-forces. Like other membrane skeleton proteins, DRP redistributed from the high-speed pellet to the integrin-rich low-speed pellet of aggregating platelets. One of the signaling enzymes that is activated following ␣ IIb ␤ 3 -ligand interactions in a platelet aggregate is calpain; DRP was cleaved by calpain to generate a ϳ140-kDa fragment that remained associated with the low-speed detergent-insoluble fraction. These studies show that DRP is part of the platelet membrane skeleton and indicate that DRP participates in the cytoskeletal reorganizations resulting from signal transmission between extracellular adhesive ligand and the interior of the cell.
Duchenne muscular dystrophy is one of the most common inherited human diseases. It is caused by a defective gene that codes for a 427-kDa protein, dystrophin (1)(2)(3)(4)(5). The deduced amino acid sequence of dystrophin shows that it consists of four domains and suggests that it is a cytoskeletal protein (6). The major rod-shaped domain contains 24 spectrin-like repeats. This domain is flanked on the amino terminus by a domain that has a high degree of homology to the actin-binding domains of spectrin and ␣-actinin, and on the carboxyl terminus by a cysteine-rich domain that shows some homology to a Ca 2ϩbinding region in ␣-actinin. The most carboxyl-terminal end of dystrophin consists of a short domain that has no homology to any known protein and appears to play a role in linking the molecule to the plasma membrane (7,8). Recent studies using purified protein or recombinant fragments containing the putative actin-binding domain (9 -11) have shown that the protein can bind to actin filaments in vitro, supporting the idea that this molecule functions as a cytoskeletal protein. The finding that dystrophin exists in a submembranous location (8,12,13) and that the carboxyl-terminal end of the molecule associates tightly with a complex of membrane glycoproteins (termed dystroglycan) (14 -17) suggests that dystrophin is a component of a submembranous cytoskeleton.
Although there is now considerable information concerning the structure and interactions of dystrophin, little is known about the way in which the absence of dystrophin leads to muscle cell necrosis. It is well established that in the absence of dystrophin, there are increased concentrations of cytoplasmic Ca 2ϩ and activation of the Ca 2ϩ -dependent protease, calpain (18 -21), suggesting that dystrophin may play a role in stabilizing the sarcolemma or in regulating the activity of Ca 2ϩ channels (22)(23)(24)(25). One of the problems associated with studying the function of dystrophin is that it is found primarily in muscle and brain tissue. In contrast, a related protein, dystrophin-related protein (DRP) 1 (26,27) is present in many different cell types (28). DRP is an autosomal gene product that is 80% homologous to dystrophin (26,29). Recent studies have shown that DRP is present in a submembranous location (30,31) and associates with the same complex of transmembrane glycoproteins as does dystrophin (32). Because the extracellular domain of dystroglycan can bind laminin (11,33) and agrin (34,35), and DRP co-localizes with agrin-induced acetylcholine receptor clusters, it has been suggested that DRP may play a role in the transmission of signals between the extracellular matrix and intracellular cytoskeleton (35). Given the similarities between dystrophin and DRP, it appears that DRP may serve a similar function to dystrophin. However, as with dystrophin, direct evidence that DRP is part of a membrane skeleton in intact cells or that it is involved in transmembrane signaling is lacking.
One cell-type that has a membrane skeleton that can be readily isolated and analyzed is the blood platelet (36). This skeleton coats the plasma membrane and associates with membrane glycoproteins. It has been visualized morphologically (37,38), isolated from detergent-solubilized platelets by centrifugation (37), and shown to be composed of short actin filaments, vinculin, spectrin, actin-binding protein and other unidentified proteins (37)(38)(39). Recent work suggests that the skeleton binds signaling enzymes, and reorganizes following interaction of the integrin ␣ IIb ␤ 3 with its adhesive ligand, fibrinogen, in a platelet aggregate (40 -42). In the present study we have shown that DRP is present in platelets and have used these cells to demonstrate that DRP is associated with a membrane skeleton, that it participates in integrin-induced reorganization of the cytoskeleton, and that it is cleaved by calpain as a consequence of integrin-ligand interactions. These studies provide direct evidence that DRP is a component of a membrane skeleton and point to a role of this protein in mediating the cytoskeletal reorganizations and transmembrane signaling that occur as a consequence of integrin-ligand interactions.

MATERIALS AND METHODS
Preparation of Platelet Suspensions-Venous blood was drawn from healthy donors, from donors with Duchenne muscular dystrophy, and from patients with Glanzmann's thrombasthenia (whose platelets lack ␣ IIb ␤ 3 (43)). Platelets were isolated by centrifugation as described previously (37). Prostacyclin (Sigma) was included in all wash steps at a concentration of 50 ng/ml. Washed platelets were resuspended at a concentration of 1 ϫ 10 9 platelets/ml at 37°C in a Tyrode's buffer containing 138 mM sodium chloride, 2.9 mM potassium chloride, 12 mM sodium bicarbonate, 0.36 mM sodium phosphate, 5.5 mM glucose, 1.8 mM calcium chloride, and 0.4 mM magnesium chloride, pH 7.4. Platelet suspensions were activated with 1.0 NIH unit of thrombin/ml (the thrombin was a generous gift of Dr. John W. Fenton II of the New York Department of Health, Albany, NY). In some experiments, platelets were preincubated with the calpain inhibitors MDL (44) and EST (45). MDL was the generous gift of Dr. S. Mehdi of Merrill Dow (Cincinnati, OH); EST was from Dr. Tamai of Taisho Pharmaceutical (Saitama, Japan). In other experiments, as described in the figure legends, platelets were preincubated with a synthetic peptide consisting of the sequence Arg-Gly-Asp-Ser (RGDS) (Telios Pharmaceuticals, Inc., San Diego, CA). Incubations were terminated by addition of one-third volume of an SDS-containing buffer in the presence of reducing agent (37) and either 1 mg/ml leupeptin or 4 mM EGTA. Samples were analyzed on SDS-polyacrylamide gels according to the method of Laemmli (46). When DRP was under study, samples were electrophoresed through gels containing 6% acrylamide and a ratio of acrylamide:bisacrylamide of 37.5:1. For all other proteins, 7.5% gels and a 29:1 ratio of acrylamide: bisacrylamide were used.
Isolation and Analysis of Subcellular Fractions-Platelets were lysed by addition of an equal volume of ice-cold lysis buffer containing 2% Triton X-100, 10 mM EGTA, 100 mM Tris-HCl, 2 mg/ml leupeptin, 100 mM benzamidine, and 2 mM phenylmethylsulfonyl fluoride, pH 7.4 (37). Lysates were immediately centrifuged at 15,600 ϫ g at 4°C for 4 min to sediment the cytoplasmic actin filaments. The 15,600 ϫ g supernatant was subsequently centrifuged at 100,000 ϫ g for 2.5 h at 4°C to sediment components of the membrane skeleton (37). In some experiments, as indicated in the text, actin filament depolymerization was induced by omitting EGTA from the lysis buffer, including DNase I (2 mg/ml, Boehringer Mannheim) (39), and incubating lysates at 4°C for 1 h prior to centrifugation. Sedimented material was solubilized by addition of an SDS-containing buffer in the presence of reducing agent (37) and either 1 mg/ml leupeptin or 4 mM EGTA. The 100,000 ϫ g supernatant was solubilized by addition of one-third volume of a four times concentrated SDS buffer (37). All samples were heated to 95°C and electrophoresed through SDS-polyacrylamide gels. In any given experiment, detergent-insoluble fractions originating from the same number of platelets (typically 3 ϫ 10 7 ) were analyzed.
Immunoblotting-Proteins were transferred from SDS gels to nitrocellulose membranes (Bio-Rad) by standard techniques (47). The membranes were stained with Ponceau S solution, blocked overnight in a buffer (T-TBS) containing 0.1% Tween 20, 137 mM NaCl, 25 mM Tris-HCl, pH 8.0, to which 5% nonfat milk (Carnation) was added. Membranes were then incubated overnight at room temperature in the T-TBS buffer containing the primary antibodies and 1.0% bovine serum albumin. Membranes were either incubated in alkaline phosphataseconjugated secondary antibodies (Bio-Rad) and developed with Bio-Rad's 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium color development solutions or were incubated with horseradish peroxidaseconjugated antibody and developed using the enhanced chemiluminescence Western blot detection system (Amersham Corp.).
Polymerase Chain Reaction (PCR) Amplification of Platelet cDNA-Fresh human platelets were isolated and platelet total RNA prepared (48). Approximately 600 ng of RNA were used to synthesize cDNA, using random hexamers and the Geneamp RNA PCR kit (Perkin-Elmer). The resulting cDNA was used in the PCR-amplification reaction. PCR conditions were 0.15 M of each primer, 2.5 units of Taq polymerase, denaturation at 94°C for 1 min, primer annealing at 57°C for 2 min, and extension at 72°C for 4 min for 40 cycles. PCR reaction products were separated on a 2% agarose gel and stained with ethidium bromide. Primer pairs from the amino-terminal end of DRP corresponded to nucleotides 28 -47 (5Ј-AGTCCTGACAATGGGCAGAA-3Ј) and nucleotides 823-842 (5Ј-GCCTCTTCTTCACATTCTTT-3Ј) (29). Primer pairs from the carboxyl-terminal end corresponded to nucleotides 9089 -9111 (5Ј-AACCAGAAATAAGTGTGAAAGAG-3Ј) and 10272-10294 (5Ј-CCTGTGGCCTGCTGGGAACATTT-3Ј) (29). Primers were chosen in separate exons to eliminate possible amplification of any contaminating DNA and were obtained from Operon Technologies Inc. (Alameda, CA). Leukocytes were isolated from human blood on Ficoll (Accurate Chemical and Scientific Corp., Westbury, NY) and leukocyte RNA obtained using the same method as was used to isolate platelet RNA. Primers for the light and heavy chains of IgG were obtained from Operon Technologies Inc.
Preparation of Mouse Muscle Microsomes-Skeletal muscle from freshly killed mice (C57BL/10SNJ, The Jackson Laboratory, Bar Harbor, ME) was homogenized essentially as described by Ervasti et al. (15), with some modifications. Briefly, tissue from one mouse was homogenized on ice in a buffer containing 0.3 M sucrose, 5 mM EGTA, 10 mM iodoacetamide, 20 mM Tris-HCl, 120 g/ml leupeptin, 6 mM benzamidine, and 120 M phenylmethylsulfonyl fluoride, pH 7.4, using a microshaft attachment to the VirtaShear (Virtis Co., Gardiner, NY) high-speed homogenizer. The tissue suspension was centrifuged in microcentrifuge tubes at 14,000 ϫ g for 10 min. Supernatants were centrifuged at 125,000 ϫ g for 30 min. The pellets were resuspended and washed twice in a buffer containing 5 mM EGTA, 50 mM Tris-HCI, 120 g/ml leupeptin, 6 mM benzamidine, and 120 M phenylmethylsulfonyl fluoride, pH 7.4. The washed pellets were resuspended in washing buffer that contained higher concentrations of protease inhibitors (1 mg/ml leupeptin, 50 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride) and solubilized for analysis on SDS-polyacrylamide gels as described above for platelets.
Antibodies-The polyclonal antibody against DRP was generously provided by Dr. Louis Kunkel (Harvard Medical School, Boston, MA). This antibody was generated against a bacterial fusion protein produced using a PCR-amplified 646-base pair sequence located near the carboxyl terminus of the DRP protein product (27). Dystrophin was detected using a polyclonal rabbit antisera, 6-10, also from Dr. Kunkel, which was raised against a fusion protein generated from dystrophin cDNA encoding a large region of the distal portion of the dystrophin rod domain (49). Monoclonal antibodies, NCL-DYS1, against the mid-rod domain of human dystrophin, and NCL-DYS2, against a synthetic polypeptide consisting of the last 17 amino acids of the carboxyl-terminal region of human dystrophin, were obtained from Novacastra, Inc. (Newcastle upon Tyne, United Kingdom). Monoclonal antibody 327, against pp60 c-src , was a generous gift of Dr. Joan Brugge (ARIAD Pharmaceuticals, Inc., Cambridge MA). Polyclonal antibodies against ␣ IIb ␤ 3 were a gift of Dr. David Phillips of COR Therapeutics (South San Francisco, CA). Polyclonal antibodies against actin-binding protein, talin, and glycoprotein Ib ␣ were raised, affinity-purified, and characterized as described previously (50,51).

Presence of DRP in Platelets-
To determine whether dystrophin or dystrophin-related protein were present in platelets, platelet proteins were electrophoresed through SDS-polyacrylamide gels and analyzed on Western blots. As shown in Fig. 1  (panel A), an antibody (6-10) against the rod domain of dystrophin (49) reacted strongly with dystrophin in mouse muscle but showed negligible reactivity against human platelets. Two other antibodies (NCL-DYS1 and NCL-DYS2) against the midrod domain and carboxyl-terminal amino acids of dystrophin, respectively, also reacted with normal mouse muscle but did not react with a protein in the molecular weight range of dystrophin in platelets (data not shown). In contrast, an antibody against DRP (27) showed weak reactivity against mouse muscle but reacted strongly with a protein in human platelets (Fig. 1, panel B). The reactivity was comparable in platelets from normal controls and from patients with Duchenne mus-cular dystrophy (e.g. Fig. 1, panel B). The protein was of comparable molecular weight to that of the DRP in muscle. Comparison of the Western blots shown in panels B and C of Fig. 1 shows that the protein detected by the DRP antibody in platelets migrated slightly higher than actin-binding protein (the highest molecular weight protein (M r ϳ 300,000) that is readily detected with Coomassie Brilliant Blue on an SDS-polyacrylamide gel of platelet proteins (39)).
To assure that the protein detected by the DRP antibody in platelets was authentic DRP, platelet RNA was isolated and used for reverse transcriptase PCR analysis. Primers from the carboxyl-terminal end of DRP (see "Materials and Methods") generated the expected 1205-base pair fragment (Fig. 2, lane  1). Similarly, primers from the amino-terminal end of DRP generated the expected band of 815 base pairs (Fig. 2, lane 2). Because platelets contain very little RNA as compared to leukocytes and it is often difficult to isolate platelets completely free of leukocytes, the PCR products could conceivably have arisen from contaminating leukocyte RNA. However, this possibility was eliminated by experiments showing that bands of the appropriate size were not generated when the DRP primers were used under the same amplification conditions on isolated leukocyte RNA (data not shown). Furthermore, primers derived from sequences in the light chain and heavy chain of IgG generated bands of appropriate molecular weight from leukocyte RNA but not from the platelet RNA (data not shown).
Association of DRP with the Membrane Skeleton in Platelets-To determine whether DRP was associated with the cytoskeleton, platelets were lysed with a Triton X-100-containing buffer and the detergent-insoluble material isolated by centrifugation. In platelets, many of the cytoplasmic actin filaments are sufficiently cross-linked that they can be sedimented from detergent-lysed platelets by centrifugation at 15,600 ϫ g; however, the membrane skeleton fragments require 100,000 ϫ g to be sedimented (36,37,51). Considerable amounts of the DRP were recovered in the detergent-insoluble fractions (Fig. 3). Essentially all of the detergent-insoluble DRP sedimented in the high-speed pellet along with the membrane skeleton fragments (lane 3).
While membrane skeleton proteins sediment in the highspeed detergent-insoluble fraction because they are associated with the detergent-insoluble membrane skeleton fragments, other proteins can sediment at these g-forces because they are inherently insoluble in Triton X-100. To distinguish between these possibilities, a lysis buffer that induces depolymerization of actin filaments (39,51) was used (the buffer contained DNase I and free Ca 2ϩ ) and the effect of this on the solubility of DRP determined. Analysis of the high-speed detergent-insoluble pellets confirmed that the amount of filamentous actin was decreased in lysates containing free Ca 2ϩ and DNase I (Fig. 4, compare the first and second lanes of panel A). Depolymerization of actin filaments in the detergent lysates was accompanied by decreased sedimentation of several proteins known to be associated with the membrane skeleton in unstimulated platelets (42,51): glycoprotein Ib ␣ (Fig. 4, panel B), ␣ IIb ␤ 3 (Fig. 4, panel C), and pp60 c-src (Fig. 4 panel D). Depolymerization of actin was also accompanied by decreased sedimentation of DRP (Fig. 4, panel E).
Incorporation of DRP into Integrin-rich Cytoskeletal Complexes-When platelets are activated, actin polymerization occurs and the cytoskeleton reorganizes such that increased actin, myosin, and other cytoskeletal proteins sediment at low g-forces from detergent-lysed platelets (36,52,53). In addition, fibrinogen is secreted from intracellular granules and binds to ␣ IIb ␤ 3 . Because it is a bivalent molecule, fibrinogen binds to ␣ IIb ␤ 3 on adjacent platelets and induces platelet aggregation. This cross-linking of ␣ IIb ␤ 3 on adjacent platelets results in  2-4). Lysates were centrifuged for 4 min at 15,600 ϫ g. The resulting pellet was solubilized in SDS-containing buffer (lane 2) and the Triton X-100 supernatant was centrifuged for a further 2.5 h at 100,000 ϫ g. The resulting high-speed pellet (lane 3), and the high-speed supernatant (lane 4) were solubilized in SDS-containing buffer. All samples were electrophoresed through SDS-polyacrylamide gels and transferred to nitrocellulose paper. Blots were incubated with antibody against DRP. Antibody-antigen complexes were detected by enhanced chemiluminescence.
signaling across the ligand-occupied integrin which in turn results in a second set of cytoskeletal reorganizations (36,40,42,52). The result of this second reorganization is that ␣ IIb ␤ 3 (Fig. 5, panel A) and membrane skeleton proteins (e.g. talin and pp60 c-src shown in Fig. 5, panels B and C) no longer require high g-forces to be sedimented from detergent-lysed platelets, they can be sedimented at low g-forces (36,40,42,52). The redistribution of integrin and membrane skeletal proteins to the low-speed detergent-insoluble fraction occurs slowly in unstirred platelet suspensions (in which cell-cell contact and, therefore, aggregation is minimized) (Ref. 42 and left-hand panels of Fig. 5) and much more rapidly in suspensions that are stirred (middle panels in Fig. 5); redistribution of the integrin and membrane skeletal proteins occurs more slowly in stirred suspensions if aggregation is inhibited by inclusion of the tetrapeptide RGDS (that binds to the ligand-binding site of ␣ IIb ␤ 3 , preventing binding of dimeric adhesive ligand) (Ref. 42 and Fig.  5, right-hand panels). As shown in Fig. 5, the amount of DRP that sedimented at low g-forces increased as platelets were activated with thrombin. The redistribution to the low-speed pellet occurred slowly in unstirred suspensions (left-hand panel) and more rapidly in stirred suspensions (middle panel); inclusion of RGDS in stirred suspensions decreased the rate of redistribution (right-hand panel). There was a close correlation between the rate at which DRP was incorporated into the membrane skeleton and that at which ␣ IIb ␤ 3 and the other components of the membrane skeleton were incorporated.
Platelets from patients with Glanzmann's thrombasthenia are deficient in ␣ IIb ␤ 3 (43). Thus, although they undergo other activation-induced events (42), they do not undergo ␣ IIb ␤ 3induced transmembrane signaling. Therefore, the integrindependent redistribution of membrane skeleton proteins to the low-speed detergent-insoluble fraction does not occur when these platelets are activated (42). In the present study, we observed that although DRP redistributed to the low-speed detergent-insoluble pellet of normal platelets (Fig. 6A), it did not redistribute to the low-speed detergent-insoluble fraction of thrombasthenic platelets that were activated in the same way (Fig. 6B). This finding is consistent with the idea that DRP is part of a submembranous skeleton that is incorporated into integrin-rich cytoskeletal complexes as a result of integrininduced transmembrane signaling in platelets.
Integrin-induced Cytoskeletal Reorganizations Lead to the Cleavage of DRP by Calpain-One consequence of the formation of integrin-rich cytoskeletal complexes in aggregating platelets is that calpain is activated (54). Several of the proteins that are cleaved by the protease (spectrin, talin, and actin-binding protein) are components of the integrin-rich cytoskeletal complexes (39,42,50). Since the findings described FIG. 5. Western blots showing that dystrophin-related protein redistributes along with membrane skeletal proteins to the lowspeed detergent-insoluble fraction of aggregating platelets. Suspensions of platelets (1 ϫ 10 9 platelets/ml) were incubated with thrombin for the indicated times. Suspensions were either agitated occasionally (left-hand panels) or stirred (middle-and right-hand panels). The samples shown in the right-hand panels had been preincubated with 0.5 mM RGDS for 5 min prior to thrombin addition. Incubations were terminated by addition of Triton X-100 lysis buffer. Lysates were centrifuged for 4 min at 15,600 ϫ g to obtain the low-speed detergent-insoluble pellet. All samples were electrophoresed through SDS-polyacrylamide gels and transferred to nitrocellulose paper. Blots were incubated with antibodies against ␣ IIb ␤ 3 (A), talin (B), pp60 c-src (C), or DRP (D) as shown. Antibody-antigen complexes were detected by enhanced chemiluminescence .   FIG. 4. Effect of actin depolymerization on the recovery of DRP and membrane skeleton proteins from the high-speed detergentinsoluble fraction of platelets. Suspensions of platelets (1 ϫ 10 9 platelets/ml) were lysed by addition of an equal volume of a Triton X-100 lysis buffer that contained 10 mM EGTA to chelate Ca 2ϩ present in the Tyrode's buffer and platelet extracts (lanes 1), or the same buffer lacking EGTA and containing 2 mg/ml DNase I (lanes 2). Lysates were incubated at 4°C for 1 h and then centrifuged for 2.5 h at 100,000 ϫ g. The resulting pellets were solubilized in SDS-containing buffer, samples were electrophoresed through SDS-polyacrylamide gels, and transferred to nitrocellulose paper. above suggest that DRP is a component of these cytoskeletal complexes, we determined whether DRP was also cleaved by calpain. As shown in Fig. 7, the amount of DRP-reactive protein in platelets decreased when platelets were stirred with thrombin (compare lane 2 with lane 1). The decrease in intact DRP was accompanied by the appearance of a fragment of ϳ140 kDa (lane 2 of Fig. 7). The cleavage of DRP did not occur in thrombasthenic platelets (see Fig. 6). Furthermore, a concentration of the tetrapeptide RGDS that partially inhibited fibrinogen binding to ␣ IIb ␤ 3 (as shown by a partial inhibition of aggregation (data not shown)) partially inhibited the cleavage of DRP (Fig. 7, lane 3; see also Fig. 5).
The finding that cleavage of DRP could be inhibited with RGDS and did not occur in thrombasthenic platelets is consistent with cleavage being induced by calpain. To directly test this, platelets were incubated with agonist in the presence of the membrane permeable inhibitors of calpain, MDL (44) and EST (45). The concentration of MDL used was such that it partially inhibited the activity of calpain, as shown by the partial inhibition of the appearance of the degradation products of actin-binding protein (Fig. 8, panel A) (actin-binding protein is cleaved to generate fragments of ϳ200 and ϳ100 kDa; the ϳ100-kDa fragment is then cleaved further to generate a fragment of ϳ91 kDa (50)). The concentrations of EST used were such that they almost completely prevented degradation of actin-binding protein in thrombin-treated platelets (Fig. 8, panel A). Similarly, MDL partially inhibited degrada-tion of DRP while EST completely prevented any detectable degradation (Fig. 8, panel B). The amount of calpain activation in thrombin-activated platelets can be quite variable. The ex-FIG. 6. Western blots showing that dystrophin-related protein does not redistribute to the low-speed pellets from lysates of platelets of patients with Glanzmann's thrombasthenia. Suspensions of platelets (1 ϫ 10 9 platelets/ml) were agitated with thrombin for the indicated times. Incubations were terminated by addition of Triton X-100 lysis buffer. Lysates were centrifuged for 4 min at 15,600 ϫ g to obtain the low-speed detergent-insoluble pellets. The Triton X-100 supernatants were centrifuged for a further 2.5 h at 100,000 ϫ g to obtain the high-speed detergent-insoluble pellets and the detergent-soluble fractions. All samples were electrophoresed through SDS-polyacrylamide gels, transferred to nitrocellulose paper, and blotted with antibodies against DRP. Antibody-antigen complexes were detected by enhanced chemiluminescence .   FIG. 7. Western blots showing the cleavage of dystrophin-related protein in thrombin-activated platelets. Suspensions of platelets (1 ϫ 10 9 platelets/ml) were preincubated alone (lanes 1 and 2), or with 0.5 mM RGDS (lane 3) for 5 min. Suspensions were then stirred in the absence (lane 1) or presence of thrombin (lanes 2 and 3) for 30 min. Incubations were terminated by addition of an SDS-containing buffer. Samples were electrophoresed through SDS-polyacrylamide gels and transferred to nitrocellulose paper. Blots were incubated with antibody against dystrophin-related protein. Antibody-antigen complexes were detected by enhanced chemiluminescence.

FIG. 8. Western blots showing the calpain-induced cleavage of actin-binding protein (ABP) and dystrophin-related protein in thrombin-activated platelets.
Suspensions of platelets (1 ϫ 10 9 platelets/ml) were preincubated at 37°C with carrier (no addition), or with one of the membrane-permeable inhibitors of calpain, MDL (25 mM; see Ref. 45) and EST (100 g/ml; see Ref. 45), for 20 min. Suspensions were then stirred with thrombin for the indicated times. Incubations were terminated by addition of an SDS-containing buffer. Samples were electrophoresed through SDS-polyacrylamide gels and transferred to nitrocellulose paper. Blots were incubated with antibody against actin-binding protein (panel A) or dystrophin-related protein (panel B). Antibody-antigen complexes were detected by enhanced chemiluminescence. ABP; 200, 100, and 91 indicate calpain-induced actin-binding protein fragments of ϳ200, ϳ100, and ϳ91 kDa (50) respectively; 140 indicates a fragment of DRP of ϳ140 kDa. periment presented in Fig. 8 is one in which considerable activation of calpain occurred, as indicated by the extensive cleavage of actin-binding protein (panel A). In this experiment, there was also considerable cleavage of DRP as shown by the almost complete loss of intact DRP (panel B); when cleavage of DRP was extensive, the fragment of ϳ140 kDa was no longer detected (Fig. 8, panel B). When calpain activity was partially inhibited (with MDL) the ϳ140-kDa fragment was detected again (panel B). Thus, the ϳ140-kDa fragment presumably represents an intermediate cleavage product of DRP that contains the epitope in the carboxyl-terminal domain against which the antibody was raised; further cleavage of the ϳ140-kDa fragment results in loss of this epitope. Like DRP, the ϳ140-kDa hydrolytic fragment of the protein was present in the low-speed detergent-insoluble fraction from aggregating platelets (see Figs. 5 and 6).

DISCUSSION
Although the molecular cause of muscular dystrophy is now known, little information is available about the way in which the absence of functional dystrophin leads to muscle necrosis. One of the problems in elucidating the function of dystrophin is that it is present predominantly in muscle and brain. In contrast, a related protein, DRP, is present in many non-muscle cells. In the present study, we show that DRP exists in platelets. In this cell, the membrane is lined by a skeleton that is readily isolated from detergent-lysed platelets by centrifugation (37) and has been visualized by electron microscopy (37,38). Several lines of evidence show that DRP is a component of the platelet membrane skeleton. First, it was recovered along with the membrane skeleton in the high-speed fraction from detergent-lysed platelets. Second, like other membrane skeleton proteins, it was released from the detergent-insoluble material when actin filaments were depolymerized. Third, DRP redistributed, along with other membrane-skeleton proteins, to the low-speed detergent-insoluble fraction from aggegating platelets. Finally, like other membrane-skeleton proteins (40,42), the redistribution from the high-to the low-speed pellet in aggregating platelets was dependent on binding of adhesive ligand to ␣ IIb ␤ 3 and did not occur in platelets that lacked this integrin. The observations that dystrophin binds to actin in vitro (9 -11), that dystrophin and DRP associate with membrane glycoproteins (14 -17, 32), and that dystrophin and DRP exist in a submembranous location (8,12,13,30,31), have provided circumstantial evidence that these proteins are part of a submembranous cytoskeletal structure. The present study provides direct evidence that DRP is indeed a component of a membrane skeleton in an intact cell.
The finding that DRP is a component of the platelet membrane skeleton suggests a number of potential functions for this protein. For example, the skeleton coats the entire plasma membrane and it is thought that it may regulate the function and distribution of membrane glycoproteins (37,52), and also stabilize the membrane, preventing microvesicles from being shed (55). It also binds signaling molecules and appears to be involved in transmembrane signaling following integrin-ligand interactions (42). In other cells, binding of extracellular ligand to dystroglycan has been implicated in inducing changes in the organization of a membrane skeleton (34,35). It is not known whether dystroglycan is present in platelets; the reorganizations of the DRP-containing membrane skeleton detected in the present study were initiated by interaction of extracellular ligand with the integrin ␣ IIb ␤ 3 . In platelets, at least two adhesive receptors are associated with the membrane skeleton (glycoprotein Ib-IX and ␣ IIb ␤ 3 ) (42, 51) and transmembrane signaling is induced as a consequence of ligand binding to both of them (42, 56 -59). Thus, it appears possible that DRP is a component of a skeletal structure that reorganizes in response to interaction of associated glycoproteins with their extracellular adhesive ligands. Additional work will be needed to investigate this possibility and to identify the membrane glycoprotein with which DRP associates in platelets.
Based on the fact that the integrin-rich cytoskeletal complexes in platelets associate with cytoplasmic actin (42) and that they contain a number of the proteins present in focal contacts of cultured cells (e.g. talin and vinculin (42)), we have suggested that they may be analagous to focal contacts in adherent cells (36,42). Interestingly, there have been previous indications that dystrophin may be present in integrin-rich domains in other cells (60). As in focal contacts, a number of signaling molecules appear to associate with the integrin-cytoskeletal complexes in platelets (e.g. pp60 c-src , pp125 FAK , phosphoinositide 3-kinase, and calpain (41,42,57,58,(61)(62)(63)(64)). At least in some cases, the recruitment of signaling molecules to the integrin-rich cytoskeletal complexes appears to be involved in activation of the enzymes (41,64). The specific protein-protein interactions that mediate the recruitment of the signaling enzymes to the integrin-cytoskeletal complexes are not known. A number of the proteins present in these cytoskeletal complexes contain the SH2 and SH3 domains that have been implicated in protein-protein interactions. Interestingly, an additional motif present in a number of signaling molecules has recently been identified and shown to be present in DRP (65,66). It will be of interest to determine whether DRP plays a simple structural role in the integrin-cytoskeletal complexes or whether it is also involved in binding and regulating signaling molecules.
One enzyme that is recruited to focal contacts in cultured cells (67) and is incorporated into the detergent-insoluble integrin-rich cytoskeletal fraction in aggregating platelets (64) is calpain. This protease is selectively activated at sites where the integrin clusters with cytoskeletal proteins (64). Thus, we have suggested that recruitment of the protease to the "focal-contact-like" structures in platelets is the first step in activation of this protease (64). The finding in the present study that DRP is part of the integrin-rich detergent-insoluble cytoskeletal fraction is of interest because one of the characteristics of muscle from patients with Duchenne's muscular dystrophy is activation of calpain and subsequent degradation of muscle proteins. In muscle, it is thought that the absence of dystrophin may result in decreased membrane stability and thus, increased Ca 2ϩ concentrations and calpain activation. An alternative idea is that dystrophin normally serves to directly regulate Ca 2ϩ fluxes (21,24,25). In platelets, integrins have been implicated in the regulation of Ca 2ϩ fluxes (68,69) and calpain activation (54). An increased understanding of the role of the integrin-cytoskeletal complexes and of DRP in regulating calpain activation in platelets may shed light on the way in which the absence of dystrophin leads to increased calpain activation in muscle.
The fact that DRP is cleaved by calpain suggests that it plays an active role in inducing the cytoskeletal remodeling that is induced by integrin-ligand interactions. Previous studies have shown that dystrophin is a substrate for calpain in vitro (70, 71) but it has not been known whether it is cleaved by this protease in an intact cell. The major DRP fragment detected in the present study was one of ϳ140 kDa that reacted with an antibody against the carboxyl-terminal end of the molecule. While this end of the molecule contains the binding site for dystroglycan, it does not contain the binding site for actin; despite this, the fragment remained associated with the cytoskeleton in aggregating cells. Future work will be needed to determine whether additional proteolytic fragments remain associated with the cytoskeleton and to identify cytoskeletal and membrane proteins that mediate the interaction of DRP and its calpain-induced fragment with the integrin-rich cytoskeletal fraction.
In summary, it is becoming increasingly apparent that the membrane skeleton in platelets binds signaling molecules and is involved in transmitting signals from extracellular adhesive proteins to the interior of the cell. The present study shows that DRP is part of this structure. The finding that integrin-induced transmission of signals from extracellular ligand to the interior results in cleavage of DRP by calpain, suggests that DRP may play an important role in mediating integrin-induced cytoskeletal remodeling and transmembrane signaling. Because the platelet membrane skeleton can be readily obtained from detergent-solubilized platelets, signaling events can be rapidly induced, and DRP-containing integrin-rich cytoskeletal complexes can be isolated, the platelet may provide a useful model in which to characterize the interactions of DRP and to identify the function of this protein. The high degree of homology between dystrophin and DRP suggests that these proteins may serve the same function. Thus, studies on the platelet could lead to an increased understanding of the way in which the absence of dystrophin in patients with Duchenne's muscular dystrophy leads to cell necrosis.