Identification of the Src Family Kinases, Lck and Fgr in Platelets
Their Tyrosine Phosphorylation Status and Subcellular Distribution Compared With Other Src Family Members
Abstract We have identified the Src family members, Lck and Fgr in resting human and rodent platelets and compared their subcellular distributions and tyrosine phosphorylation status to those of the other Src family kinases to gain insights into the signal transduction pathways active in maintaining platelets in the circulation. Like Fyn, Lyn, and Yes, most of Fgr and Lck was detergent-insoluble in human and rat platelets. In comparison, Src showed higher detergent solubility than the Src-related kinases. Most all human platelet Src was detergent-soluble, while that of rodent platelets was present in all detergent fractions.
We also compared the tyrosine-phosphorylation status of Lck and Fgr to other Src family members in resting platelets using immunoprecipitation and immunoblotting. All of these Src family members except Fgr exhibited substantial phosphotyrosine antibody labeling. The partitioning of these kinases, with the exception of Src, with the detergent-insoluble fraction, their tyrosine-phosphorylation status, and co-localization with endocytotic vesicles lead us to hypothesize that the Src family kinases are involved in signaling events that drive cytoskeletal reorganization and active endocytosis of plasma proteins by circulating platelets.
Division of Experimental Hematology (T.I.P., S.A.S., N.K.H., C.W.J.), St. Jude Children’s Research Hospital, Memphis, Tenn, 38105, and Departments of Pathology (P.E.S., R.J.B.,) and Hematology and Medical Oncology (B.J.D.), Oregon Health Sciences University, Portland, Ore.
- Received December 27, 1996.
- Accepted September 2, 1997.
Circulating platelets must constantly reorganize their membranes and underlying cytoskeletons to deform and traverse the splenic vasculature. Circulating platelets also very actively endocytose plasma proteins by both receptor-mediated and pinocytotic mechanisms. Both deformation and endocytosis must be driven by signal transduction pathways, most likely involving tyrosine phosphorylation.
Earlier studies reported that human platelets contain high levels of Src as well as Fyn, Hck, Lyn, and Yes,1–3 but not Lck and Fgr.3 Since Lck plays an important role in T cell signaling events4–5 that resemble those during platelet activation, we thought it likely that Lck would be expressed and be active in platelet signaling. Also, Fgr is usually co-expressed with Hck in blood cells,6 and, since platelets were reported to express Hck,3 we were interested in whether Fgr was also expressed in platelets.
The role of the Src family kinases in platelet function remains undefined, and only the subcellular location of Src has been studied.2,7 This study then was designed to determine biochemically whether Fgr and Lck are expressed in platelets, to assess their tyrosine phosphorylation status as compared to the other Src-related kinases in platelets, and to define into which subcellular fractions they and the other platelet Src-related kinases partition. Finally, this study examines whether the Src-related kinases expressed in human platelets also are present in rodent platelets in preparation for analysis of their function using mice with gene disruptions of these genes. Our results indicate that Fgr and Lck along with Fyn, Lyn, and Yes, but not Hck, are expressed in both rodent and human platelets. Src, Fyn, Lck, and Lyn are significantly tyrosine-phosphorylated, and all, except Src, are concentrated in detergent-insoluble fractions, with the detergent solubility of Src showing species variation.
Animals, Antibodies, and Cell Lines
Retired-breeder male Wistar rats were obtained from Harlan Industries (Indianapolis, Ind), and were used for all rat studies. Male C57BL/6 mice were obtained from The Jackson Laboratory, Bar Harbor, ME. The monoclonal antibody (MoAb) 4G10 against anti-phosphotyrosine and MoAb 4G10 covalently linked to Protein A-Sepharose were from Dr Brian Druker (Oregon Health Sciences University). MoAb 327 Src antibody was from Dr Joan Brugge, ARIAD Pharmaceuticals, Cambridge, Mass. The specificity of this antibody for Src was confirmed by its absence of reactivity by immunoblotting with Src-/- platelets. Affinity-purified polyclonal rabbit antibodies (PoAb) against Src gene family tyrosine protein kinases including Blk, Fyn, Hck, Lck, and Lyn were obtained from Santa Cruz Biotechnology, Inc, Santa Cruz, Calif. The specificity of the Fyn and Lyn antibodies were verified by absence of their reactivity with Fyn-/- and Lyn-/- platelets, respectively. The specificity of the Lck antibody was substantiated as per Bijlmakers et al8 by its reactivity with the appropriate molecular weight band in immunoblots of Jurkat cells, but not JCAM cells; JCAM cells are a subline of Jurkat cells that lack Lck.4 Moab 3H9 to human Yes9 was kindly provided by Dr Tadashi Yamamoto, The Institute of Medical Science, The University of Tokyo, or was purchased from Wako Bioproducts. The specificity of this Yes antibody was previously demonstrated by absence of its reactivity with brain of Yes-/- mice.10 Affinity-purified polyclonal rabbit antimouse Fgr and Hck antibodies were provided by Dr Clifford Lowell, University of California, San Francisco, Calif. The specificity of the Fgr and Hck antibodies was previously demonstrated by absence of their reactivity with marrow of mice with gene deletions of these proteins.6 Monoclonal antibody to cortactin p80/85 (4F11) was from Dr Albert Reynolds (Vanderbilt University). Antihuman PoAb against Shc was purchased from Upstate Biotechnology, Inc. MoAb to actin (C4) was purchased from Boehringer Mannheim Corp (Indianapolis, Ind). Affinity-purified PoAb to P-selectin was kindly provided by Dr Michael Berndt, Baker Medical Research Institute, Monash University, Prahran, Victoria, Australia. Antihuman PoAb against platelet membrane glycoprotein (GP) IIb was generously provided by Dr Peter Newman, Blood Center of Southeastern Wisconsin, Milwaukee, Wis. MoAb against human platelet membrane GP IV was kindly supplied by Dr Narenda Tandon, American Red Cross, Rockville, Md. Protein G Sepharose 4 Fast Flow was from Pharmacia LKB Biotechnology, Inc.
The Jurkat T cell line and the Jurkat cell subline, JCAM-1, that lacks Lck4 were obtained from the American Type Cell Collection.
Collection and Preparation of Platelets for Biochemical Studies
Blood was collected from the abdominal aorta of metofane-anesthetized rats and mice, and from the antecubital vein of normal human volunteers into syringes containing acid citrate dextrose (ACD) (0.13 mol/L citric acid; 0.15 mol/L sodium citrate; 0.1 mol/L dextrose) plus 8 or 40 μmol/L prostaglandin E1 (PGE1) anticoagulant (1:9, anticoagulant/blood). Platelet-rich plasma (PRP) was prepared by differential centrifugation. Then, three methods of platelet isolation from PRP were compared including gel filtration, pelleting from PRP without washing, and pelleting from PRP followed by washing (data not shown). Platelets to be solubilized for electrophoresis without washing were pelleted from PRP by centrifugation (2950g/20 minutes/room temperature). Another set of platelets was solubilized after pelleting from PRP and washing three times in EHS buffer (1 mmol/L Na2EDTA, 10 mmol/L HEPES, and 0.15 mol/L NaCl), pH 7.6.11 Both the unwashed and washed platelet pellets were suspended in 60 volumes of EHS buffer for solubilization. For most experiments, platelets were solubilized after isolation from PRP by gel-filtration through Sepharose 2B columns that were washed and eluted with a HEPES-modified Tyrodes buffer: 12 mmol/L NaHCO3, 138 mmol/L NaCl, 5.5 mmol/L glucose, 2.9 mmol/L KCl, 10 mmol/L HEPES, with or without 5 mmol/L EGTA, with or without 0.1% bovine serum albumin (pH 6.5). In all cases, one mL of platelet suspension was solubilized by boiling after addition of one-third volume of gel sample buffer containing 0.125 mol/L Tris-HCl, pH 6.8, 40% glycerol, 8% SDS, 160 mmol/L dithiothreitol, and 0.01% bromphenol blue.12 The profile of phosphorylated proteins was the same in all cases. However, proteins of gel-filtered platelets demonstrated a one-third lower level of total labeling with the 4G10 phosphotyrosine antibody as determined by densitometry of immunoblots. This difference could be explained by some activation of platelets by the additional steps of isolation inherent in the other methods. Platelet activation was assessed by ultrastructural immunogold analysis of the localization of the granule membrane protein, P-selectin. In gel-filtered platelets, P-selectin was restricted primarily to granule membranes, indicating minimal granule release (data not shown). Platelets isolated by the other methods showed varying degrees of P-selectin on their surface membrane, indicating some degree of granule release (data not shown).
One-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed under reducing conditions in 7.5% to 15% linear gradients of acrylamide using the discontinuous buffer system of Laemmli.12 Approximately 30 μg of protein was loaded per lane.
Immunoblotting of Antibodies to Electrophoretically Separated Proteins
Separated proteins were electrophoretically transferred to nitrocellulose sheets in a transfer buffer consisting of 25 mmol/L Trizma base, 192 mmol/L glycine, and 20% methanol. The nitrocellulose strips were incubated overnight with the primary antibodies at room temperature. The blots were then washed five times with rinsing buffer containing 10 mmol/L Trizma base and 150 mmol/L NaCl, pH 8.0. After that, the blots were incubated with secondary antibody conjugated to horseradish peroxidase (HRP-conjugated donkey-antimouse IgG or donkey-antirabbit IgG; Jackson ImmunoResearch Laboratories, Inc) for antibody detection by the enhanced chemiluminescence (ECL) method (Amersham Corp). Two percent casein was included in all incubations and washes to reduce nonspecific binding.13
Quantitation of Bands in Coomassie Blue-Stained Electrophoretic Gels and ECL Films of Immunoblots of Platelet Proteins
The optical density of platelet protein bands in SDS-polyacrylamide gels stained with Coomassie blue and ECL films of immunoblots was determined using an image analyzer (Bio Image System, Millipore Corp).
Detergent Extraction and Fractionation of Platelets
Platelets were extracted 30 minutes on ice by the addition of an equal volume of detergent buffer containing 100 mmol/L Tris-HCl, pH 7.4, 2% Triton X-100, 10 mmol/L EGTA, 200 U/mL kallikrein inhibitor units of aprotinin, 1, 2 or 7.5 mmol/L sodium orthovanadate (all prevented tyrosine dephosphorylation), 1 mmol/L sodium fluoride, 0.5 mg/mL leupeptin, 8 mg/mL benzamidine-HCL, and 2 mmol/L 4-[2-aminoethyl]-benzenesulfonyl fluoride, HCl (AEBSF). The Triton-insoluble residues were pelleted (15 850g/20 minutes/4°C) in an angle-head rotor. The supernatant was removed and centrifuged (100 000g/2.5 hours/4°C) to obtain a high speed pellet.14 The low-speed and high-speed pellets and high-speed supernatants then were solubilized for SDS-PAGE as described above.
Immunoprecipitation of Proteins
For immunoprecipitations, platelets isolated by gel filtration were extracted with Triton X-100 as described above, and the Triton-insoluble material pelleted by centrifugation (26 816g/10 minutes/4°C). The Triton-insoluble pellets were extracted 30 minutes with RIPA buffer on ice (2 mL of RIPA buffer/Triton pellet from each Wistar rat). RIPA buffer consisted of 10 mmol/L Tris, pH 7.4, 158 mmol/L NaCl, 5 mmol/L EGTA, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 100 U/mL Kallikrein Inhibitor Units of aprotinin, 0.5 mg/mL leupeptin, 8 mg/mL benzamidine-HCL, 2 mmol/L AEBSF, 7.5 mmol/L sodium orthovanadate, and 1 mmol/L sodium fluoride. The RIPA-extracted samples were centrifuged (26,816g/20 minutes/4°C) in an angle-head rotor. One-half mL of the RIPA-soluble and RIPA-insoluble fractions were solubilized in gel sample buffer for SDS-PAGE. The Triton X-100 and RIPA supernatant were used for immunoprecipitation.
Platelet Triton-soluble and Triton-insoluble/RIPA-soluble fractions were preabsorbed with protein G-Sepharose for 1 to 2 hours at 4°C. The Sepharose beads were removed by centrifugation (14 000g/1 minute/4°C). The antibodies used for immunoprecipitation included the phosphotyrosine MoAb 4G10 covalently linked to Protein A-Sepharose, and Src MoAb 327, cortactin p80/85 MoAb 4F11, and affinity-purified PoAb antibodies to the Src gene family tyrosine kinases Fgr, Fyn, Lck, and Lyn coated onto protein G-sepharose by incubation for 1 hour at 4°C. Beads precoated with antibodies were pelleted, and the beads washed once with RIPA buffer. To collect the immune complexes, the Triton or RIPA supernatants were usually incubated 4 hours (in some cases overnight) with the antibody-coated Sepharose beads at 4°C. Immunoprecipitates were pelleted, washed five times with the appropriate detergent buffer, RIPA or 2% Triton. The immune complexed-beads were resuspended in 350 μL of EHS buffer and solubilized in one-third volume of Laemmli gel sample buffer. After boiling for 8 minutes, the samples were centrifugated (14 000g/1 minute/room temperature) to remove the beads, and the supernatant was electrophoresed.
Detection of Lck mRNA in Platelets by Reverse Transcription-Polymerase Chain Reaction and Nucleotide Sequencing.
Direct nucleotide sequencing was performed on a PCR product prepared from cDNA derived from human platelet mRNA. Blood for preparation of platelets was collected from a healthy adult volunteer, the platelets separated by differential centrifugation and washed in phosphate-buffered saline with acid citrate dextrose. Messenger RNA was isolated from 108 platelets using the Micro-Fast Track mRNA Isolation Kit (Invitrogen Corp), and cDNA prepared by reverse transcription. A 326 bp PCR product spanning the region corresponding to amino acids 374 to 482 at the carboxy terminus of human Lck p56 was prepared. The amplifications were performed using a Perkin-Elmer Cetus thermal cycler under the following conditions: 1 minute at 94°C, 1.5 minutes at 55°C, and 2 minutes at 72°C for 40 cycles. The PCR product was directly sequenced using the fmol DNA Sequencing System (Promega Corp) with 33P-labeled primers.
Ultrastructural Immunogold Analyses
Platelets were prepared for ultrastructural immunogold analyses as previously described.15 P-selectin was detected using an affinity-purified rabbit antihuman platelet P-selectin antibody, followed by goat anti-rabbit IgG bridging antibody (Jackson ImmunoResearch Labs, Inc), and then rabbit anti-goat IgG conjugated to 10 nm gold particles (Sigma Chemical Co) for 20 minutes. The grids were washed 5 times in the same buffer (5 minutes each) and then in double-distilled water 4 times (5 minutes each). They were then stained and embedded in a mixture of methylcellulose and 0.1 to 0.4% uranyl acetate, and viewed with a Philips 301 transmission electron microscope.
Detection of Lck and Fgr Along With Other Src Family Kinases in Resting Platelets and Their Detergent Solubility
We first established the SDS-PAGE protein profile and detergent solubility of rat platelet proteins. Representative Coomassie blue-stained protein profiles of whole rat platelets in parallel with the relative distributions of specific proteins among the high-speed supernatant, and high-speed and low-speed Triton pellets are presented in Fig 1⇓. (Most of the protein in the band at 66 kDa in the first three lanes is bovine serum albumin from the gel filtration platelet isolation buffer.) Actin is the major protein in platelets and is concentrated in the low speed Triton-insoluble pellet (Fig 1⇓ and Fig 2⇓, lower panel), along with actin-associated proteins including actin-binding protein, α-actinin, tropomyosin (Fig 1⇓) and cortactin (Fig 2⇓, lower panel). In contrast, myosin is the major band in the high-speed Triton pellet, which also contains other cytoskeletal proteins such as spectrin, talin, and vinculin as well as actin-binding protein and α-actinin.
Next, the expression of Lck and Fgr as well as the other Src family kinases in rodent and human platelets was examined by immunoblotting with affinity-purified polyclonal or monoclonal antibodies. As depicted in Fig 2⇑ (lower panel) human and rat platelets as well as mouse platelets (not shown) contained Src and the Src-related kinases, Fyn and Lyn. Human and rat platelets also contained Fgr and Lck, which had not been detected by kinase assays in a previous study of human platelets.3 Subsequent immunoblot analyses with the same affinity-purified antibodies, indicated that Fgr and Lck are in mouse platelets as well (not shown).
Expression of Lck in platelets was further demonstrated by production and nucleotide sequencing of a 326 bp RT-PCR product from human platelet mRNA spanning the region of amino acids 374 to 482 at the carboxy terminus of human Lck p56. The nucleotide sequence of this PCR product matched the published sequence (data not shown).
The blot for Yes in Fig 2⇑ is with human platelets, since blotting of rodent platelet proteins with the 3H9 Yes monoclonal antibody yielded a barely detectable reaction. In contrast, Hck had been reported in human platelets;3 however, we did not find Hck in rat, mouse or human platelets with either affinity-purified rabbit antimouse Hck antibody (from Dr Clifford Lowell) or affinity-purified rabbit antihuman Hck antibody (from Santa Cruz Biotechnology). Both of these Hck antibodies recognized an appropriately sized band in immunoblots of SDS-solubilized mouse bone marrow electrophoresed in parallel lanes as a positive control for platelets (data not shown).
The detergent solubility studies revealed that substantial portions of Lck and Fgr as well as Fyn, Lyn, and Yes are Triton X-100-insoluble (Fig 2⇑, lower panel). The protein loads for all fractions were the same as those in Fig 1⇑. Fgr and Lck along with Fyn, Lyn and Yes were found in the low-speed, Triton-insoluble fraction (15 850g pellet) in both rat and human platelets (Fig 2⇑, lower panel). Consistent with earlier reports,16–18 we found Src of resting human platelets in the 100 000g detergent-supernatant and 100 000g-pellet, but not in the low-speed pellet (Fig 2⇑, lower left panel). In comparison, a portion of Src was present in the low speed detergent-insoluble pellet of resting rat platelets (Fig 2⇑, lower right panel), indicating that the detergent solubility of Src differs between human and rat platelets. This difference in Src detergent solubility of human and rat platelets was not due to activation of rat platelets, since ultrastructural immunogold analysis of P-selectin localization demonstrated only α-granule, and not plasma membrane association (data not shown).
Other candidate tyrosine-phosphorylated proteins (P-TyrP) also displayed different detergent solubilities. The adapter P-TyrP, Shc, with a molecular weight similar to the Src family kinases,19 was most prevalent in the Triton-soluble fraction. In comparison, the cytoskeletal protein, cortactin, which is reported to be tyrosine-phosphorylated on platelet activation,20 was found exclusively in the low-speed Triton pellet (Fig 2⇑, lower panel).
In contrast to the Src family P-TyrP, very little of the platelet membrane glycoproteins, GPIIb and GPIV, were detected in the low-speed Triton pellet (Fig 2⇑, lower right panel), indicating that the Triton X-100 extraction procedure was sufficient to solubilize these two integral membrane proteins. However, these two membrane proteins did differ somewhat in their Triton solubility in that some GP IIb was associated with the high-speed Triton pellet, while this fraction contained only a trace of GP IV. Of interest here is that GP IV was reported to be complexed with Fyn, Lyn, and Yes in human platelets as defined by reciprocal immunoprecipitation experiments.3 Our observation that GP IV is largely Triton soluble, coupled with our finding that small amounts of Fyn, Lyn, and Yes are Triton-soluble, suggests that the reported complexes between GP IV, Fyn, Lyn, and Yes were Triton-soluble, and therefore, not associated with Triton-insoluble membrane domains.
Profile Comparison of P-TyrP of Resting Rodent Versus Human Platelets and Their Detergent Solubility
The tyrosine phosphorylation profile of resting platelet proteins was next evaluated to gain clues as to which of the Src family or their substrates, may exhibit this post-translational modification. Immunoblotting of solubilized, gel-filtered rat platelet proteins with the 4G10 phosphotyrosine antibody identified eight phosphorylated bands with apparent molecular weights (Mrs) of 135, 85, 80, 62, 60, 56, 54 and 40 kD (Fig 2⇑, right top panel). Mouse platelets had a similar P-TyrP profile (data not shown). Gel-filtered human platelets demonstrated P-TyrPs with apparent Mr of 145 kD and Mrs of 29 to 35 kD in addition to those present in rat and mouse platelets (Fig 2⇑, left top panel).
We next examined the detergent solubility of tyrosine-phosphorylated proteins of resting rat (Fig 2⇑, top right panel) and human platelets (Fig 2⇑, top left panel) to characterize the subcellular associations of the P-TyrP with respect to the Src family kinases and other candidate proteins studied above. Tyrosine-phosphorylated proteins were present in all Triton fractions; however, they displayed some variation in their detergent solubility. All of the tyrosine-phosphorylated bands except one in rat and two in human platelets, showed the highest proportion of phosphotyrosine label in the low-speed Triton pellet. The first exception was a band of approximately 60 kD, that displayed the highest proportion of tyrosine phosphorylation in the Triton-soluble fraction for both rat and human platelets. The other exception seen only in human platelets was a band of 70 to 75 kD, which was most prominent in the high-speed Triton pellet.
While some portion of most phosphotyrosine-containing protein bands was present in all Triton fractions, two bands with apparent kDs of 80 and 85 were detected only in the Triton-insoluble fraction. These two tyrosine-phosphorylated bands were shown by reciprocal immunoprecipitation and immunoblotting to be the two isoforms of cortactin (data not shown).
Examination of the Phosphorylation Status of Src Family Proteins
We next used immunoprecipitation and immunoblotting to determine whether Lck and Fgr and the other Src-family kinases were tyrosine-phosphorylated in resting platelets, as well as to identify other P-TyrP in the resting platelet. RIPA extracts of the Triton X-100-insoluble fraction were used for most immunoprecipitations, since the tyrosine-phosphorylated proteins were concentrated in the Triton-insoluble fraction. MoAb 4G10 immunoblots of 4G10 MoAb immunoprecipitates from Triton-insoluble, RIPA-soluble platelet extracts revealed six prominent phosphotyrosine-containing bands (Fig 3⇓).
Immunoblotting of the 4G10 immunoprecipitates with specific antibodies showed that the 55 to 60 kDa phosphotyrosine region contained Lck, Fyn, and Lyn (Fig 3⇑). Reciprocal immunoprecipitation with affinity-purified Lck, Fyn, and Lyn antibodies followed by immunoblotting with the 4G10 MoAb, confirmed that these Src-related kinases were tyrosine-phosphorylated in extracts of resting platelets (data not shown). Cortactin was present in the 80 and 85 kDa bands of MoAb 4G10 immunoprecipitates as identified by cortactin MoAb. Reciprocal immunoprecipitates with cortactin MoAb probed with MoAb 4G10, confirmed tyrosine-phosphorylation of cortactin in resting platelet extracts.
Only a small amount of Fgr was immunoprecipitated from RIPA extracts of rat or mouse platelets by 4G10 MoAb (Fig 4⇓, right panel). Furthermore, Fgr, immunoprecipitated by affinity-purified polyclonal Fgr antibodies from the Triton-soluble fraction, did not show reactivity with 4G10 MoAb (Fig 4⇓, left panel). These observations suggest that Fgr has only a low level of tyrosine-phosphorylation in resting platelets, in comparison to the other Src-related kinases, or that the affinity of the P-Tyr antibody for Fgr is low.
Src immunoprecipitated by Src MoAb from Triton-soluble and insoluble fractions was tyrosine-phosphorylated (Fig 5⇓). However, in contrast with the Triton-soluble fraction, little if any of Src in the Triton-insoluble fraction was immunoprecipitated by 4G10 phosphotyrosine MoAb (Fig 6⇓). The inability of Src to be immunoprecipitated from this fraction could be explained by (1) the phosphotyrosine residues of Src being complexed and not available to the 4G10 MoAb, or (2) the 4G10 MoAb has a higher affinity for the other tyrosine-phosphorylated proteins in the RIPA fraction.
Several Src-related kinases including Fyn, Hck, Lyn, and Yes3,21,22 have been reported in resting human platelets, but Fgr and Lck were not detected even though they are present in other hematopoietic cells. Here, we have analyzed platelets for expression of Fgr and Lck and compared their detergent solubility and tyrosine phosphorylation status to those of the other Src family kinases of platelets. Contrary to an earlier report,3 we found that Fgr and Lck are expressed in rodent and human platelets along with Src, Fyn, and Lyn. The presence of both Fgr and Lck in rodent and human platelets presents further possibilities for participation of Src family kinases in platelet intracellular signaling. We also confirmed the presence of Yes in human platelets.22 The Yes-specific monoclonal antibody we used yielded only a very weak reaction in immunoblots of rodent platelets. However, we did not detect Hck in human platelets as previously reported,3 nor in rodent platelets, using an Hck antibody whose specificity for Hck was shown by its lack of reactivity with Hck-/- mouse marrow.6 All of these kinases except Src were largely detergent-insoluble, and all except Fgr showed considerable tyrosine-phosphorylation.
Our detection of Fgr and Lck in platelets, while an earlier study did not,3 may relate to differences in detection methodology. We used the direct approach of immunoblotting with affinity-purified antibodies, and in addition, RT-PCR in the case of Lck. In contrast, Huang et al detected the Src family kinases based on the ability of immunoprecipitated kinases to autophosphorylate themselves.3
The Src family kinases are membrane-associated, and detergent solubility has been used as an approach to analyze that association. The subcellular distributions of the Src family kinases have been studied in nucleated cells, but except for Src and Yes, their subcellular localizations have not hitherto been examined in platelets. With the exception of Src and Blk, the Src family members in nucleated cells are detergent insoluble.23–28 In agreement with those results, we found substantial amounts of platelet Fgr, and Lck as well as Fyn, Lyn, and Yes in the low-speed detergent-insoluble fraction. In comparison, Fox et al18 reported that Yes in resting human platelets was detergent-soluble. This discrepancy may be related to Yes antibody specificity, since the specificity of the 3H9 Yes monoclonal antibody we used was verified by immunoblotting of Yes-/- mice.10 However, Yes-/- platelets reacted with the 2-7 Yes monoclonal antibody which was used in the Fox et al study (T.I.P. et al, unpublished results), questioning the specificity of the 2-7 monoclonal antibody for Yes.
In an ultrastructural immunogold analysis conducted in parallel with this study, we found Fgr, Fyn, Lck, and Lyn co-localized with clathrin over electron-dense cytosolic compartments of platelets.29 This co-localization of Src-related kinases with clathrin suggests that these kinases are associated with endocytotic vesicles in platelets. In contrast, the majority of Src was in the high-speed pellet and supernatant, and associated ultrastructurally with plasma membranes, membranes of the surface-connected canalicular system, and α-granule membranes.29
The difference in detergent solubility between Src and the Src-related kinases most likely relates to the inability of Src to be palmitylated.30–32 The fact that Src is not palmitylated, but a portion is still Triton-insoluble in resting rat platelets, suggests that Src in resting rat platelets is not bound to the same Triton-insoluble macromolecules as the other Src family members. Our finding that Triton-insoluble Fyn, Lck, and Lyn were immunoprecipitated by the phosphotyrosine antibody, but Triton-insoluble Src was not, suggests that Src may be bound to Triton-insoluble macromolecules via a phosphotyrosine group in the absence of its ability to be palmitylated.
Kwong et al have hypothesized that the Triton-insoluble structure with which Src-related kinases associate is glycophosphoinositol (GPI),33 which anchors proteins to membranes.27,34 However, recent evidence indicates that cholesterol- and sphingolipid-rich membranes also are highly detergent insoluble,32 so that detergent insolubility does not necessarily equate with GPI-anchoring. This point has been made even more directly by a recent study showing that Lck co-immunoprecipitated with a GPI-anchored protein when Triton extraction and immunoprecipitation were performed at the customary 4°C, but not when immunoprecipitation was performed at room temperature after 5 minutes incubation of the detergent extract at 37°C.27 These authors concluded that Lck and GPI-anchored proteins are not directly complexed, but simply anchored in the same membrane domain.
These findings also bring into question the reported association of Triton-insoluble Src family members with caveolae, since this conclusion was based on the co-sedimentation of these protein tyrosine kinases with caveolae in density gradients after Triton extraction of whole cells.26 Caveolar association of the Triton-insoluble Src-related kinases also is questionable because these kinases are also Triton-insoluble in cell lines lacking caveolin,35,36 a caveolae-associated protein, or caveolae. We did not detect caveolin in platelets (C.W.J. et al, unpublished work). The apparent association of the Src-related kinases with endocytotic vesicles, and not with caveolae, in platelets in our ultrastructural analysis,29 suggests that the functional roles of the Src-related kinases in platelets may differ from those in other cell types.
Of the Src family members, only the tyrosine-phosphorylation status of Src had been reported in platelets.17 Our analysis indicates that Src, Fyn, Lck, Lyn are tyrosine-phosphorylated. In comparison, Fgr showed little or no tyrosine-phosphorylation. This lack of Fgr tyrosine-phosphorylation does not support the current dogma that Src family kinases are usually tyrosine-phosphorylated with the kinase activity determined by whether the proteins are tyrosine-phosphorylated on the kinase-activating (for example, Tyr-416 in the case of Src) or inactivating tyrosine (for example, Src Tyr-527).30
In summary, we have (1) identified the Src family kinases, Fgr and Lck as well as Fyn, Lyn, Src, and Yes in rodent as well as human platelets, (2) determined that Fgr, Fyn, Lck, Lyn, and Yes are substantial amounts of detergent-insoluble, (3) found that Src detergent-solubility differed in resting rat and human platelets, and (4) demonstrated that Src family members except for Fgr are tyrosine-phosphorylated in resting rodent and human platelets.
The large repertoire of Src family kinases present in platelets along with the differences in their detergent solubility, tyrosine-phosphorylation, and ultrastructural localization suggest that individual Src family tyrosine kinases of resting platelets function in different signaling events required for reorganization of their cytoskeletons during circulation and for their active endocytosis of plasma proteins. These studies now provide the basis for analysis of the role of Src family members in platelet formation and function using genetically-engineered mice with deletions of Src family members.
Selected Abbreviations and Acronyms
|ACD||=||acid citrate dextrose|
|PCR||=||polymerase chain reaction|
|SDS-PAGE||=||sodium dodecyl sulfate-polyacrylamide gel electrophoresis|
This work was supported in part by R01 Grant HL 51546 (C.W.J.and P.E.S.) from the National Heart, Lung, and Blood Institute, P30 CA21765 Cancer Center Support Grant and P01 Calif 20180 from the National Cancer Institute, Public Health Service, Department of Health and Human Services, and by American Lebanese Syrian Associated Charities.
Golden A, Nement SP, Brugge JS. Blood platelets express high levels of the pp60c-src-specific tyrosine kinase activity. Proc Natl Acad Sci U S A.. 1986;83:852–856.
Rendu F, Lebret M, Danielian S, Fagard R, Levy-Toledano S, Fischer S. High pp60c-src level in human platelet dense bodies. Blood. 1989;73:1545–1551.
Huang M-M, Bolen JB, Barnwell JW, Shattil SJ, Brugge JS. Membrane glycoprotein IV (CD36) is physically associated with the Fyn, Lyn, and Yes protein-tyrosine kinases in human platelets. Proc Natl Acad Sci U S A.. 1991;88:7844–7848.
Lowell CA, Soriano P, Varmus HE. Functional overlap in the Src gene family: Inactivation of Hck and Fgr impairs natural immunity. Genes Dev. 1994;8:387–398.
Bijlmakers M-J JE, Isobe-Nakamura M, Ruddock LJ, Marsh M. Intrinsic signals in the unique domain target p56lck to the plasma membrane independently of CD4. J Cell Biol. 1997;137:1029–1040.
Stein PL, Vogel H, Soriano P. Combined deficiencies of Src, Fyn, and Yes tyrosine kinases in mutant mice. Genes Develop. 1994;8:1999–2007.
Berndt MC, Phillips DR. Purification and preliminary physicochemical characterization of human platelet membrane glycoprotein V. J Biol Chem. 1981;256:59–65.
Fox JEB. Linkage of a membrane skeleton to integral membrane glycoproteins in human platelets. Identification of one of the glycoproteins as glycoprotein Ib. J Clin Invest. 1985;76:1673–1683.
Pestina TI, Jackson CW, Stenberg PE. Abnormal subcellular distribution of myosin and talin in Wistar Furth rat platelets. Blood. 1995;85:2436–2446.
Oda A, Druker BJ, Smith M, Salzman EW. Association of pp60src with Triton X-100-insoluble residue in human blood platelets requires platelet aggregation and actin polymerization. J Biol Chem. 1992;267:20075–20081.
Clark EA, Brugge JS. Redistribution of activated pp60c-src to integrin-dependent cytoskeletal complexes in thrombin-stimulated platelets. Mol Cell Biol. 1993;13:1863–1871.
Fox JEB, Lipfert L, Clark EA, Reynolds CC, Austin CD, Brugge JS. On the role of the platelet membrane skeleton in mediating signal transduction. Association of GP IIb-IIIa, pp60c-src, pp62c-yes, and the p21ras GTPase-activating protein with the membrane skeleton. J Biol Chem. 1993;268:25973–25984.
Oda A, Druker B, Miyakawa Y, Watanabe R, Handa M, Watanabe K, Ikeda Y. SHC is a major tyrosine phosphorylated protein in human platelets and is associated with Triton X-100 insoluble residue. Blood. 1994;84(suppl 1):534a.
Zhao Y-H, Krueger JG, Sudol M. Expression of cellular-Yes protein in mammalian tissues. Oncogene. 1991;5:1629–1635.
Štefanová I, Hořejsí V, Ansotegui IJ, Knapp W, Stockinger H. GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Science. 1991;254:1016–1019.
Thomas PM, Samelson LE. The glycophosphatidylinositol anchored Thy1 molecule interacts with the p60fyn protein tyrosine kinase. J Biol Chem. 1992;267:12317–12322.
Draberova L, Draber P. Thy-1 glycoprotein and src-like protein-tyrosine p53/p56fyn are associated in large detergent-resistant complexes in rat basophilic leukemia cells. Proc Natl Acad Sci U S A.. 1993;90:3611–3615.
Sargiacomo M, Sudol M, Tang Z, Lisanti MP. Signal transducing molecules and glycosyl-phosphatidylinositol-linked proteins form a caveolin-rich insoluble complex in MDCK cells. J Cell Biol. 1993;122:789–807.
Rodgers W, Crise B, Rose JK. Signal determining protein tyrosine kinase and glycosyl-phosphatidylinositol-anchored protein targeting to a glycolipid-enriched membrane fraction. Mol Cell Biol. 1994;14:5384–5391.
Shenoy-Scaria AM, Dietzen DJ, Kwong J, Link DC, Lublin DM. Cysteine3 of Src family protein tyrosine kinases determines palmitylation and localization in caveolae. J Cell Biol. 1994;126:353–363.
Stenberg PE, Pestina TI, Barrie RJ, Jackson CW. The Src family kinases, Fgr, Fyn, Lck, and Lyn, co-localize with coated membranes in platelets. Blood. 1997;89:2384–2393.
Cooper JA. The Src family of protein-tyrosine kinases. In: Peptides and Protein Phosphorylation. Kemp BE, ed. Boca Raton, Fla: CRC Press; 1990:85–104.
Resh MD. Myristylation and palmitylation of src family members: the facts of the matter. Cell. 1994;76:411–413.
Casey PJ. Protein lipidation in cell signaling. Science.. 1995;14:221–225.
Fra AM, Williamson E, Simons K, Parton RG. Detergent-insoluble glycolipid microdomains in lymphocytes in the absence of caveolae. J Biol Chem. 1994;269:30745–30748.
Gorodinsky A, Harris DA. Glycolipid-anchored proteins in neuroblastoma cells form detergent-resistant complexes without caveolae. J Cell Biol. 1995;129:619–627.