Isolation and Properties of a Thornbin-sensitive Protein of Human Platelets*

SUMMARY A glycoprotein of human platelets is released into the incubation medium upon treatment of the intact cells with 1 unit per ml of thrombin for 2 min. The cellular location and properties of this protein differ from those of other thrombin-labile platelet proteins such as fibrin-stabilizing factor, fibrinogen, and the contractile protein thrombosthenin. It occurs only in the particulate fraction and is released only from intact cells. Thrombin concentrations of 100 units per ml do not release this thrombin-sensitive protein from sonicated platelets. As isolated from the incubation supernatant of thrombin-treated platelets the thrombin-sensitive protein is excluded from Sephadex G-200; it has an apparent molecular weight of 190,000 in sodium dodecyl sulfate polyacrylamide gels, from which it has been purified and partially characterized. No smaller subunits have been found. It has limited and after the

damaged area of a vessel wall in the form of an aggregated platelet plug and (b) a catalytic lipoprotein surface upon which the biochemical interaction of certain clotting factors is thought to occur (1). During this complex series of events platelets themselves undergo many alterations, which can be induced in vitro by thrombin.
Both their intracellular and surface morphology are affected. Their shape changes from that of a discoid spongelike cell with deep invaginations of the surface membrane (2) to a spiny sphere with pseudopodia during aggregation (3)) and their intracellular granules gather to the center of the cell and then disappear.
Prominent intracellular metabolic changes include a burst of lactate production (4) and oxygen comsumption (5), an increase in ATPnse activity, a,nd a fall in ATP and ADP levels (4, 6). A number of platelet constituents including serotonin and ADP (7), sulfated mucopolysaccharides (S), and specific proteins (9) are rapidly released from the cell into the surrounding medium.
Since platelet membranes play a critical role in platelet function in hemostasis, the primary substrate for thrombin action might be expected to be associated with the platelet surface. We have previously shown that thrombin induces changes in the phospholipid synthetic patterns (10) and the adenylate cyclase activity (11) of membranes in intact human platelets.
Such changes are very rapid, having reached their maximum within less than 1 to 2 min of thrombin treatment, and require intact cells as the thrombin substrate. Some of these metabolic alterations are clearly due to proteolytic activity, since they can also be produced by trypsin.
Earlier we described a protein, designated thrombin-sensitive protein, which disappears from the particulate fraction of intact platelets upon thrombin treatment, with a time course and substrate requirement identical with the membrane effects noted above (12). We now report that this protein is released from the platelet, apparently intact, after thrombin treatment.
We have isolated the protein from the platelet particulate fraction and from post-thrombin incubation supernatnnt, solutions and have compared the a,mino acid composition, carbohydrate content, and molecular size of the two preparations.

2724
Thronzbin-sensitive Protein of Human Platelets Vol. 247,No. 9 David Aaronson and was assayed as described previously (14). It is free of plasmin, plasminogen, and Factor Xa activities.' Canalco reagents and apparatus were used for acrylamide gel electrophoresis. I-Dimethylaminonaphthalene sulfonyl chloride (dansyl chloride, B grade) was from Calbiochem and l-fluoro-2,4-dinitro[3H]benzene was from New England Nuclear. Eastman silica gel chromatograms or polyamide layer plates from Chen Chin Trading Co., Ltd., Taiwan, were used in thin layer chromatography.
Phenylmethyl sulfonyl fluoride, DFP, TLCK, and N-acetylneuraminic acid were obtained from Sigma.2 Erythrocyte-agglutinating phytohemagglutinin was isolated from Difco phytohemagglutinin P by the method of Weber et al. (15) and from lentils by the method of Howard and Sage (16).

Nethods
Platelet Preparation-Platelet isolation and washing were carried out at room temperature.
Fresh human platelet-rich plasma, prepared at room temperature by the American Red Cross, St. Louis Chapter, was centrifuged 25 min at 100 x g to decrease erythrocyte and leukocyte contamination. The supernatant plasma was centrifuged 17 min at 1,000 x g to sediment the platelets, which were washed twice in 10 to 15 volumes of pH 6.5 buffer containing 0.1.02 M NaCl, 0.0039 M K2HPO+ 0.0039 M Na2HP04, 0.022 M NaHQPO+ and 0.0055 M glucose. Platelets were incubated at 37" with shaking for 7 to 10 min in the absence or presence of 1 unit per ml (0.2 pg per ml) of human thrombin at 1 to 2.5 x log ceils per ml. Incubation buffer contained 0.14 M NaCI, 0.015 M Tris, pH 7.4, and 0.005 M glucose. The incubation mixtures were centrifuged for 10 min at 7,000 x g at 4", the supernatant solutions were respun 10 min at 48,200 x g to remove residual platelets; then boiled 10 nun and chilled immediately. Following overnight dialysis against distilled water the incubation supernatants were lyophilized. Platelet pellets from the 7,000 x g centrifugation were resuspended immediately in the above buffer at the original cell concentration, sonicated 15 to 25 set at 70% intensity with a Biosonik sonifier, and centrifuged 30 min at 48,200 x g. The particulate fraction thus isolated was stored at -20". SDS Solubilization and Gel Electrophoresis-The platelet particulate fractions were suspended by mild homogenization in a Potter-Elvehjem homogenizer in distilled water, assayed for protein, and the suspension immediately diluted with a phos- 7.4, and 0.1% SDS, with 0.035a/0 w/v ammonium pcrsulfate and 0.1% v/v TEMED added immediately before pouring. The solution was poured to a depth of 19 cm in Pyrex tubes, 20 x 0.5 cm, and allowed to polymerize in room light. Bubble formation in the gels was lessened by keeping the gel mixture cool and shielded from light during pouring, which was carried out as quickly as possible. Once polymerized, the gels could be left layered with water and sealed at room temperature for up to 24 hours before use.
Electrophoresis was carried out for 11 hours at 8 ma per gel, with an upper buffer of 0.1 y0 SDS-O.1 M sodium phosphate, pH 7.4-0.1 M 2-mercaptoethanol and a lower buffer of 0.1 M sodium phosphate, pH 7.4. For analytical electrophoreses 230 pg of particulate fraction protein, derived from approximately 0.23 x IO9 platelets, or the incubation supernatant protein derived from 1 to 3 x log control or 0.23 to 0.9 X log thrombin-treated platelets was applied to each gel. Gels were stained for 12 hours in 0.1% Coomassie brilliant blue-12% trichloroacetic acid-50% methanol, destained electrophoretically (17) in 77, acetic acid-10% methanol, and stored in 7% acetic acid. Densitometry of the gels, which attain a length of 20 cm after the staining procedure, was carried out with a Gilford recording spectrophotometer at 555 nm with a model 2410-S Gilford linear transport scanner. The content of thrombin-sensitive protein in each gel was estimated by integrating the area under the protein peak (11).
For preparative electrophoreses up to 1.2 mg of particulate fraction or the incubation supernatant protein from 0.7 to 1 x log platelets was applied to each gel, and the section of gel known to contain the desired protein was cut immediately after electrophoresis. The 8-mm slices were minced into smaller pieces, homogenized in a Potter-Elvehjem homogenizer in 1 ml of 3% SDS-O.1 M sodium phosphate, pH 7.4-l y0 2-mercaptoethanol per four slices, and dialyzed against the same buffer overnight at room temperature. After centrifugation of the homogenate for 15 min at 20,200 x g at 20", the eluted protein was recovered in the supernatant. Recovery was quantitated by comparative densitometric scanning of SDS gels of the eluted protein and of incubation supernatant from a known number of thrombintreated platelets, Amino Acid and Amino Sugar Analyses-Thrombin-sensitive protein isolated by the above elution procedure was precipitated from the eluate by treating with an equal volume of 20% trichloroacetic acid for 30 min at room temperature, followed by freezing for several hours and rethawing. After centrifugation for 30 min at 20" at 20,200 x g, the protein pellet was taken up in 1 to 2 ml of 0.5 M NaHC03 and dialyzed against 2 liters of the same buffer to remove residual SDS, followed by dialysis against distilled water. The protein sample, 0.1 to 0.15 mg, was then lyophilized, hydrolyzed in uacuo at 105' for 22 hours in 1.2 ml of constant boiling HCI, and the hydrolysate analyzed on a Beckman model 120 C amino acid analyzer. Methionine and cysteine were determined as methionine sulfone and cysteic acid after performic acid oxidation (18).
Sialic Acid Assay-Sialic acid was hydrolyzed from proteins by boiling samples 1 min in 1 N HCl and was assayed by a modification of the method of Warren (19), halving the reaction volume and extracting with 0.75 ml of cyclohexanone for 0 to 10 nmoles and 1.5 ml for 10 to 70 nmoles of sialic acid. Eluted protein samples were dialyzed exhaustively against 0.5 M NaHC03, followed by distilled water, to remove the bulk of SDS so that Issue of May 10, 1972 N. L. Baenxiger, G. N. Brodie, and P. W. iklajerus 2725 samples could be concentrated easily for the sialic acid assay. Trichloroacetic acid precipitation was not performed on samples taken for sialic acid assay to avoid potential hydrolysis. Residual SDS was assayed by the method of Reynolds and Tanford (20). An identical concentration of SDS was added to the sialic acid standards, since high levels of SDS raise the sample absorbance at 534 nm slightly.

Amino-terminal
Analysis-Dinitrophenylation of eluted particulate fraction thrombin-sensitive protein was done according to the method of Rosenberg and Guidotti (21), with 500 PCi of I-fluoro-2, 4-dinitro[aH]benzene and 35 yl of unlabeled l-fluoro-2,4-dinitrobenzene (100 mg per ml in absolute ethanol) added to 200 pg of protein in 0.2 ml of 1% SDS-O.02 M NaHC03. Dansylation of eluted particulate and released thrombin-sensitive protein was carried out by modification of the method of Gray (22), in which 100 to 300 pg of protein in 1 to 1.4 ml of 3% SDS-O.5 M NaHC03 was stirred overnight in the dark with 0.5 to 1 ml of dansyl chloride, 20 mg per ml in acetone. The modified protein was dialyzed against 0.5 M NaHCO3 and then distilled water to remove SDS; in some experiments it was precipitated out of the reaction mixture in 10% trichloroacetic acid before dialysis. After 16 to 18 hours of hydrolysis in vucuo at 105" in 5.5 N HCl, the dansyl or the ether-extractable DNP-amino acids were chromatographed by the method of Hartley (23), and of Brenner et al. (24), or Wang and Wang (25), respectively. Water-soluble DNP-amino acids were not examined.
Column Chromatography-Crude incubation supernatant from thrombin-treated platelets was incubated with phenylmethyl sulfonyl fluoride (0.4 mM) or TLCK (40 pg per ml) at 4', or with DFP (1 to 2.5 mM) at room temperature, and stored at the same respective temperatures under a toluene atmosphere. After concentration to 1.5 to 3 ml, the TLCK-and phenylmethyl sulfonyl fluoride-treated supernatants from 13 to 25 x lo9 platelets were applied to a Sephadex G-200 column, 59 x 1.5 cm, equilibrated with 0.14 M NaCl-0.15 M Tris-Cl, pH 7.4.0.02(3, NaN3 at 4". The column's void volume was determined with blue dextran. Protein Assays-Protein concentration was determined by absorbance at 280 nm (26), the method of Lowry et al. (27), by a microbiuret assay (28), or by the method of Fruchter and Crestfield (29) where protein determinations were made on solutions containing SDS-mercaptoethanol.

Ca++-ATPase
Assay of Incubation Supernatant and iVembrane Pellets Derived from Thrombinand Nonthrombin-treated Platelets-Ca++-dependent ATPase activity was measured according to the method of Cohen et al. (30) modified by increasing the concentrations of Ca* from 1.1 to 10 mM.
This was found to be the optimal concentration of Ca++ for the assay of Ca++-ATPase activity in platelet membrane preparations.
Isolalion of Thrombosthenin-Thrombosthenin was isolated from platelets by a modification of the method of Nachman et al. (31). Platelets were disrupted according to the hypotonic lysis method of Barber and Jamieson (32) and thrombosthenin extracted from the "membrane pellet" and the soluble cell supernatant as described by Nachman et al. Aliquots of all samples during the extraction procedure were subjected to analysis for Ca++-dependent ATPase as described above and SDS-polyacrylamide disc gel electrophoresis.
Isolation of Released Thronabin-sensitive Protein by Gel Filtration on Sephadex G-200 in XDS-Released thrombin-sensitive protein was obtained from batches of 150 to 200 x log pIateIets as de-scribed above. After thrombin treatment and centrifugation to remove platelets, the supernatant solution was boiled for 10 min, dialyzed overnight against distilled water, and then lyophilized. After five batches of released protein were prepared in this manner, the lyophilized material was dissolved in 10 ml of 3% SDS-l% 2-mercaptoethanol-0.1 M sodium phosphate, pH 7.3, heated 10 min in a boiling water bath and allowed to solubilize overnight at room temperature. This resulting clear solution conta,ining 106 mg of protein in 10 ml was applied to a column, 2.5 x 90 cm, of Scphadex G-200 equilibrated with 0.1% SDS-O.1 M 2-mercaptoethanol-0.1 M sodium phosphate, pH 7.3. The column was eluted with this same buffer and 2-ml fractions were collected and aliquots were assayed for protein, sialic acid, and thrombin-sensitive protein content. The void volume of this column was 130 ml. Thrombin-sensitive protein was measured by performing SDS gels on aliquots of fractions and the thrombin-sensitive protein content quantitated by densitometry of Coomassie blue-stained gels. One unit of thrombin-sensitive protein is defined as an area of 1 cm2 under the densitometer tracing of the protein band which corresponds to approximately I ,ug of protein (11).

RESULTS
Release of Platelet Protein by Thrombin-Our previous studies have shown that a major platelet protein with a molecular weight of 190,000, as analyzed by SDS-polyacrylamide gel electrophoresis, completely disappears from the cells within 15 set to 2 min after the addition of 1 unit per ml of thrombin to whole platelets. Disrupted cells or the extracted protein itself are not affected by thrombin under these conditions, even aft#er incubation times up to 30 min. We have referred to this protein as the thrombinsensitive protein. It occurs exclusively in the particulate fraction of disrupted platelets, and remains in the 48,000 x g pellet through five washes of the particulate fraction. It is not found in the soluble fraction of disrupted platelets, nor in plasma (12). Further there is no protein contained in erythrocyte membranes corresponding to this protein, Electron micrographs of this 48,000 x g particulate fraction have shown it to consist of membrane vesicles, a few intact and many disrupted granules. This fraction contains all of the platelet adenylate cyclase activity, as well as other enzymes associated with cell surface membranes (11). Attempts to localize the thrombin-sensitive protein to the surface membrane fraction of platelets with the platelet membrane fractionation procedure of Barber and Jamieson (32) were unsuccessful due to extensive proteolysis of this protein which occurred during the fractionation procedure. In experiments where sonication of platelets for varying lengths of time were performed it was shown that when greater than 99% of &glucuronidase was solubilized the thrombin-sensitive protein remained in the particulate fraction. Thus while it is not established that this protein is a component of the surface membrane it does appear to be tightly associated with some platelet membranous structure.
SDS-acrylamide gel patterns of proteins appearing in the control particulate fraction and the incubation supernatants from control and thrombin-treated human platelets are shown in Fig. 1. Control incubation supernatants contain 0.03 to 0.1 mg of protein, and thrombin supernatants 0.25 to 0.4 mg of protein per IO9 platelets. No change in the electrophoretic mobility of the thrombin-sensitive protein in SDS gels is detectable upon release, so that this process is assumed not to involve any sig- However, a small molecular weight change occurring if the protein were cleaved from the platelet surface, leaving behind a peptide fragment, might not be detectable by this method.
In addition to the thrombin-sensitive protein, an SDS-polyacrylamide gel of the incubation supernatant from thrombintreated platelets contains another dense band at an apparent molecular weight of 68,000. This protein may be albumin, which Davey and Liischer (9) have previously found to be released from platelets by thrombin.
A number of other proteins bands are present in SDS gels of incubation supernatants from both control and thrombin-treated platelets.
Bmong them are three bands at molecular weights of 60,000, 52,000, and 44,000. Other faint bands at molecular weights of 174,000, 156,000, 124,000, and 104,000 also occur in both supernatants.
None of these proteins have been identified. B faint band in Gel b of Fig. 1, at the position of the thrombinsensitive protein in the adjacent gels, indicates the release of some protein in a control incubation supernatant.
The amount of this protein seen in control supernatants varies from none at all to about 5oj, of that seen in incubation medium from thrombin-treated platelets.
Its presence may stem from the reported prothrombin activation on the surface of platelets, possibly during isolation and washing (33). Gel b in Fig. 1 contains the incubation supernatant from 3-fold more control platelets than Gel c does from thrombin-treated platelets.
The quantity of thrombin-sensitive protein in Gel b is negligible compared to that of Gel c, and the other bands, with the exception of that at 68,000 molecular weight, appear of approximately the same density in both gels, suggesting that, t,hese proteins are increased 2-to a-fold in a thrombin supernatant.
Size of Thrombin-sensitive Protein-When incubation supernatant from thrombin-treated platelets is applied to a Sephadex G-200 column in the absence of SDS, the thrombin-sensitive protein appears in the void volume (Fig. 2). In this state it probably consists of more than one 190,000 molecular weight unit, as a protein of one chain that size should be retarded on the column to some degree. When chromatography on G-200 is performed in SDS, this protein is retarded as shown below. Vigorous efforts to break down the apparent 190,000 molecular weight chain into smaller subunits have been unsuccessful.
The following procedures have failed: (a) boiling the membrane fraction in the SDS solubilization medium, either before or after overnight solubilization at room temperature, and SDS solubilization (b) at lower ionic strength or (c) in the presence of 4.3 M urea. The thrombin-sensitive protein band is equally sharp in SDS gels of particulate fraction preparations solubilized in the regular buffer and at lower ionic strength (Fig. 3) ; smearing out of the band might be expected if the latter treatment were disaggregating the protein to subunits which then reaggregated in the higher ionic strength of gels and buffer during electrophoresis.
Because EDTA has been reported necessary for disaggregation of erythrocyte membrane proteins, the platelet particulate fraction has been solubilized by the method of Lenard (34). Overnight dialysis of particulate fraction against 5 InM Na2-EDTA-5 mM 2-mercaptoethanol at pH 7.0 prior to SDS-polyacrylamide gel electrophoresis results in disappearance of most of the thrombin-sensitive protein band.
No new protein bands indicative of subunits are seen farther along the gel (Fig. 4). However, since overnight storage of the particulate fraction in distilled water at 4" prior to SDS-gel electrophoresis gives the same result as EDTA treatment (Fig. 4), and boiling the particulate fraction in distilled water for 10 min before the overnight dialysis against EDTA-Smercaptoethanol results in little or no decrease of this protein in the SDS-gel pattern of the particulate fraction, the disappearance of thrombin-sensitive protein from these preparations may be due to proteolysis rather than to EDTA-mediated disaggregation.
We conclude from the foregoing experiments that the 190,000 molecular weight band on SDS gels represents either a single polypeptide chain or an aggregate held together by forces insensitive to the above disruptive techniques.
Sensitivity to Proteolysis-The thrombin-sensitive protein, both within the platelet and after release by thrombin, exhibits a marked sensitivity to proteolysis.
When whole platelets are stored at -2O", little or none of this protein can be found in the subsequently prepared particulate fraction.
A degradation product of the thrombin-sensitive protein appears in the incubation supernatant from thrombin-treated platelets with time, The experiment shown in Fig. 5 shows the degradation of the released protein as a function of storage conditions. Thrombin inhibitors such as TLCK, phenylmethyl sulfonyl fluoride, PMSF, and DFP decrease but do not completely halt this process. Boiling the incubation supernatant immediately after isolation best prevents degradation of the protein.
This breakdown may be due to thrombin or to some nonthrombin proteolytic activity in whole platelets or in the incubation supernatant from thrombin-treated platelets.
The protein appears to contain certain sites particularly susceptible to proteolysis, since it is first converted to a product of molecular weight 170,000. Solubility and Pur$cation-Thrombin-sensitive protein released by thrombin partially precipitates if the incubation supernatant is chilled or concentrated, or its ionic strength decreased: storing and handling it at room temperature causes less precipitation than at 4". Attempts to purify this protein by standard methods of chromatography were abandoned due to the above difficulties.
Elution from SDS-polyacrylamide gels (Fig. 6) proved to be a successful means of purifying this protein from the incubation supernatant of thrombin-treated platelets ("released" thrombin-sensitive protein) and that from the particulate fraction of control platelets ("particulate" thrombin-sensitive protein) in 25 to 407, yield, for purposes of comparing the two proteins.
Caution was necessary in cutting the bands out of gels to avoid contamination with adjacent regions on gels, especially those in the particulate fraction moving directly ahead of the 190,000 molecular weight band, which stain poorly with Coomassie brilliant blue and contain much carbohydrate.
Glycoprotein Nature of Thrombin-sensitive Protein-Thrombin treatment of whole platelets releases 27% of the total platelet sialic acid into the incubation medium, averaging 6.3 nmoles of sialic acid released per 10s platelets.
This incubation supernatant also contains a potent inhibitor of phytohemagglutinininduced human erythrocyte agglutination, at a level of 1000 to 1200 inhibitory units per mg of protein by the assay of Kornfeld and Kornfeld (35). The phytohemagglutinin-inhibitory activity and a peak of sialic acid appear in the Sephadex G-200 void volume containing the thrombin-sensitive protein.
Sialic acid determination on released and membrane protein eluted from SDS gels gave values of 22 residues per 190,000 g of protein for the former and 25 for the latter.
Determination of amino sugar content indicated 16 residues of glucosamine per 190,000 g of protein for the released protein and 13 for the paticulate protein.
Identical recoveries of glucosamine were obtained from 4-and  I  Comparison of amino acid compositions of particulate and released IhrorrLbia-s~rk~ilive pToLein with other platelel p~oleiu The number of residues per 190,000 molecular weight chain were determined by calculating the total micrograms of amino acid in the hydrolysate and converting moles of each amino acid to micromoles per 190,000 micrograms.
The contributions of tryptophan, sialic acid, and hexose to the total molecular weight were not included in the calculations. 22.hour hydrolyses of the protein in constant boiling HCI, so that separate $-hour amino sugar hydrolyses were not used further and glycosamine content was determined from the 22hour amino acid hydrolyses. There was insufficient material for determination of the neutral sugar content of the thrombinsensitive protein.
Amino Acid Composition- Table  I compares the amino acid compositions of released and particulate protein.
The compositions of both proteins are very similar, the values for most residues agreeing within 11%. Differences outside this range, the discrepancy in tyrosine, cysteine, phenylalanine, and glucosamine content, may reflect differential losses during amino acid analysis of very small samples rather than true differences in composition of the protein from the two sources. Amide content was not determined.
The thrombin-sensitive protein is striking in its content of acidic residues, aspartic and glutamic acid constituting 27r/, of the total residues.
Of the residues, 438 out of 1815 or 24% are nonpolar, including alanine, valine, methionine, isoleucine, leucine, and phenylalanine in the particulate protein, and 23y0 of the protein residues are nonpolar in the released protein.
No amino-terminal amino acid could be found in the protein. With techniques sensitive enough to detect 1 amino-terminal residue per mole of protein, no dansyl amino acid could be de-tected in either particulate or released protein, and dinitrophenylation of 1 mpmole of particulate protein with 500 PCi of l-fluoro-2, 4-dinilro[3H]benzerle (26 @Zi per mole) yield approximately 140 rn$Zi of ether-soluble radioactivity which did not cochromatograph with any known ether-soluble DNP-amino acids. An amino acid which forms a water-soluble DNP derivative has not been excluded, although no dansyl amino acid corresponding to those forming water-soluble DNP derivatives was detected. Isolation of Released Thrombin-sensitive Protein by Chromalogrup/ry o'y1 Sephadex G-200 in SDS-While released and particulate thrombin-sensitive protein isolated by elution from SDS gels appeared to be homogeneous by SDS gels, no independent criteria of purity were available.
Further, the amounts of material obtainable with this technique were very small. We thus attempted to isolate released thrombin-sensitive protein in larger quantities utilizing gel filtration on Sephadex G-200 in SDS as described under "Methods." With this method 106 mg of thrombin-released protein were fractionated as shown in Fig.  7A. While the major proteins of the post-thrombin supernatant are not as well separated here as on as SDS gel, it is clear that thrombin-scnsitivc protein is separated from other glycoprotein material.
The fractions containing this protein (75 to 94) (26 mg) were pooled, concentrated, and reapplied to the G-200 column (Fig. 7B). After this second chromatography on SDS Sephadex G-200, there was correspondence between sialic acid and thrombin-sensitive protein content with a constant ratio of sialic acid to thrombin-sensitive protein throughout the peak. This suggests that this protein is truly a glycoprotein, a conclusion that is further supported by experiments where fractions from both the front and rear of the protein peak were rechromatographed on SDS G-200 with identical positions of protein elution and sialic acid content.
Amino acid analysis of protein isolated in this manner agreed within 10% with results presented in Table I, and the sialic acid content of the protein isolated by gel filtration ranged from 25 to 30 residues per mole which also agrees with the above data on protein eluted from gels. While this technique for protein isolation is quite similar in principle to purification on SDS gels, the apparent identity of the product isolated by the two methods further supports t,he hypothesis that the thrombin-sensitive protein is a single protein which is released relatively intact after thrombin treatment.
As an independent criterion of purity, an antibody to the protein purified by Sephadex chromatography in SDS was produced in rabbits.
Immunodiffusion of this antibody against crude incubation supernatant from thrombin-treated platelets gave a single precipitin line, which was a lint of identity with the purified antigen.
Relationship to Platelet Contractile Protein (Thrombosthenin)- Cohen et al. (30) have reported that thrombosthenin M, the myosin-like component of the platelet-contractile protein which contains Cat+-activated ATPase activity, disappeared from whole equine platelets upon treatment with 10 units per ml of thrombin.
The thrombin-sensitive protein was therefore examined for Ca++-activated ATPase activity to ascertain whether it might in fact be thrombosthenin M, Assay of the incubation supernatant from thrombin-treated platelets revealed no Ca++dependent ATPase activity.
Lack of activity was not due to inhibition by thrombin, as the addition of purified thrombosthenin to the incubation supernatant from thrombin-treated platelets did not affect the Ca+'-ATPase activity of the thrombosthenin.
Assay The fractions were assayed for trated 5-fold yielded Ca++-ATPase activity of less than 0.3 pmole of Pi liberated per hour, compared with 6.5 ,umoles of Pi liberated from ATP per hour in the particulate fraction of the counterpart platelets, indicating that less than 5% of the membrane-associated Ca++-ATPase activity was released from the cells. SDS-gel electrophoresis showed that 80% of the platelet thrombin-sensitive protein had been released into the incubation supernatant assayed above. Thrombosthenin isolated from both platelet membranes and the soluble cell supernatant showed high specific activity Ca++-ATPase activity (1.9 and 4.03 FM Pi liberated per mg of protein per hour, respectively). SDSpolyacrylamide gel electrophoresis of both these preparations showed no band corresponding to the thrombin-sensitive protein. DISCUSSION Davey and Liischer (9) have previously described the types of proteins released from intact platelets by thrombin, which were examined by agarose gel electrophoresis and immunoelectrophoresis. They found released proteins in the prealbumin, albumin, &lipoprotein, fibrinogen, and y-globulin classes, as well as platelet-specific cr-and fl-globulins: Enzyme studies have shown the release of acid phosphatase and fi-glucuronidase (9), and a procoagulant activity believed to be either a lipoprotein (9) or phospholipid micelles (38). In the studies of Davey and Ltischer (9) most of the proteins and cnzymc activities were present in incubation supernatants from both control and thrombin-treated platelets, with amounts in the latter supernatant 2-to a-fold higher than those of the former.
Albumin was an exception, being increased 7-fold in the supernatant from thrombin-treated platelets.
A P-globulin fraction containing a lipoprotein with procoagulant activity comprised 37% of the thrombin supernatant.
We cannot correlate the protein bands on SDS-polyacrylamide gel electrophoreses of control and thrombin supernatan.ts with the various proteins known to be released other than the band at a molecular weight of 68,000, which is greatly increased in thrombin over control supernatants and is probably albumin (the molecular weight for human serum albumin is 65,800 (39)) The other electrophoretic bands, excluding the thrombin-sensitive protein, appear to be roughly 2-to a-fold increased in thrombin supernatants.
The procoagulant lipoprotein found by Davey and Liischer has properties similar to the thrombin-sensitive protein and could well be the same protein, since the lipoprotein seems from their data to be excluded from Sephadex G-200 and the thrombin-sensitive protein comprises over 90 % of the protein appearing in our G-200 void volumes. Davey and Liischer found the P-globulin fraction of the thrombin supernatant containing the procoagulant lipoprotein to be increased only 2.6-fold over that of the control supernatant, yet the 190,000 molecular weight band on an SDS gel of a thrombin supernatant contains far more than 3 times the amount of protein as the same band in a control supernatant.
However, the experimental conditions used by these authors differed from ours; they incubated platelets at an 8-to 20-fold higher concentrations than in our studies, with a IO-fold greater thrombin concentration, and apparently did not treat their supernatants with proteolysis inhibitors. This may account for the discrepancy, since Davey and L&her (9) would have lost much of the protein through precipitation from an excessively concentrated solution and through proteolytic degradation.
A number of investigators have reported thrombin-labile proteins, different from fibrinogen, occurring in platelets. Such proteins disappear from crude platelet extracts or from the soluble supernatant fraction of the cell upon treatment with fairly high thrombin concentrations (> 10 units per ml), as analyzed by acrylamide and starch gel electrophoresis and immunodiffusion. The protein which we are describing, in contrast to these proteins, is located in the platelet particulate fraction and in intact cells is sensitive to 0.1 to 1 unit per ml of thrombin Its location in the particulate fraction and its thrombin sensitivity in intact cells, along with its insusceptibility to thrombin in sonicates, distinguish it from the thrombin-labile proteins of Salmon and Bounameaux (40) and Nachman (41), which disappear from platelet sonicates, and the fibrin-stabilizing factor described