ARCHIVES
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 286, No. 2, May 1, pp. 546-550, 1991
Reduction of a Disulfide Bond of Thrombospondin Supernatant Solution of Activated Platelets’
in the
Mark V. Speziale and Thomas C. Detwiler2 Department
of Biochemistry,
State University
of New York Health Sciences Center at Brooklyn, Brooklyn,
New York 11203
Received October 4, 1990, and in revised form December 21, 1990
Incubation of the material secreted by activated platelets leads to the formation of disulfide-linked dimers and multimers of one of the proteins, thrombospondin. To determine whether these complexes formed as a result of thiol-disulfide exchange (no change in the number of thiols) or of oxidation of thiols (a decrease in the number of thiols), the number of thiols in TSP was measured during formation of multimers. The number of thiols increased from about 3/mol to 4.8/mol. The half-time for the disappearance of monomers of thrombospondin was fourfold greater than the half-time for appearance of new thiols. The appearance of new thiols, as well as the formation of multimers, was inhibited by Ca2+. The appearance of new thiols was reversible; addition of Ca2’ reversed the process, and at pH 8, but not at pH 6 or 7, the appearance of new thiols spontaneously reversed. No new thiols formed during incubation of partially purified thrombospondin or after the supernatant solution had been treated with activated thiol-Sepharose to remove reactive thiol compounds. It is concluded that thrombospondin has a disulfide bond that is unstable in the absence of Ca2+. It can be attacked by a thiol of another molecule of thrombospondin to form disulfide-linked multimers, by a thiol of the same molecule of thrombospondin to generate isomerization of disulfide bonds or, as observed in this study, by another secreted thiol compound to give a mixed disulfide and a new thiol. o 1991 Academic
Press.
Inc.
Thrombospondin (TSP)3 is a large glycoprotein (M, 420,000) secreted by activated platelets and by a variety 1 This work was supported by Grant HL37250 from the National Institutes of Health, United States Department of Health and Human Services. * To whom correspondence should be addressed. ’ Abbreviations used: TSP, Thrombospondin; Hepes, 4-(2-hydroxyethyl)-1-piperazineethane-sulfonic acid; NEM, N-ethylmalemide; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate.
of cultured cells [for reviews see Refs. (l-3)]. It is considered an adhesive protein with a role in cell-cell and cell-matrix adhesion, but its exact function is unknown. It is composed of three apparently identical disulfidelinked polypeptide chains. The conformation is Ca’+-dependent; the protein is more compact in the presence of Ca2+ (4-6). There are 69 cysteine residues in each polypeptide chain (7), with a single equivalent of free thiol (8). The remainder of the cysteines presumably are in intra- or interchain disulfide bonds. The single equivalent of free thiol is distributed among 12 different cysteine residues in the Ca2+-sensitive part of the molecule; this has been interpreted as evidence of intramolecular thioldisulfide exchange (8). Calcium ions partially protect the thiols from reaction with thiol-specific reagents and stabilize labile disulfide bonds (8-10). Secreted platelet thrombospondin undergoes thiol-dependent formation of disulfide-linked multimers (10). Thrombin becomes disulfide-linked to TSP (9), and it accelerates the formation of TSP multimers (10). Formation of multimers and formation of complexes with thrombin are inhibited by calcium ions (9,10). The ability of TSP to form disulfide-linked multimers and complexes with thrombin may reflect a broader potential of TSP to crosslink other proteins, such as those of the extracellular matrix and cell surface. To investigate the mechanism of the formation of disulfide-linked TSP multimers, we monitored changes in the thiol content of TSP during formation of multimers. A decrease in number of TSP thiols would indicate oxidation of thiols to intermolecular disulfide bonds, while no change in number of thiols would suggest thiol-disulfide exchange. We were surprised to find an increase in the number of thiols. We present evidence that this increase in free thiols is the result of an intermolecular reaction between a disulfide bond(s) in TSP and a separate thiol compound released by activated platelets. MATERIALS
AND
METHODS
Platelets. Human platelets were isolated by differential centrifugation of whole blood obtained from healthy volunteers using 0.15 vol of
546 All
Copyright 0 1991 rights of reproduction
0003.9%x/91 $3.00 by Academic Press, Inc. in any form reserved.
REDUCTION
OF DISULFIDE
120 mM trisodium titrate/6.6 mM EDTA as anticoagulant. The platelets were washed twice with Hepes-buffered saline (150 mM NaCl/lO mM Hepes, pH 7.4) and resuspended in this buffer to 1.5-4 X 10 platelets/ ml. The experiments described in this paper were performed with the supernatant solution prepared from a suspension of activated platelets. The suspension was made 100 mM with bis-Tris-propane, 2 mM with efluoride, 100 pM aminocaproic acid, 1 mM with phenylmethylsulfonyl with leupeptin, and 100 pM with CaCl,. The platelets were activated by 4 pM ionophore A23187 (Calbiochem, San Diego, CA), and the suspension was gently mixed for 3 min at room temperature. The platelets were removed by centrifugation. The supernatant solution was used immediately or placed on ice under N1. The supernatant solution contained the secreted material, including 0.5-2 mM Ca2+ (11). Measurement of free thiols of TSP. Free thiols were labeled with [3H]NEM (40-60 Ci/mmol; NEN Research Products, Boston, MA) as described (8). The samples were made 8 M with urea or were mixed with nonreducing SDS-PAGE sample buffer containing SDS sufficient to give a final concentration of 2%. The denaturing condition was necessary, because we found that the thiols of TSP are partially protected in the multimeric form of the protein prior to denaturation. NEM (50 pM) was added and the solution was incubated for 30 min at room temperature. A loo-fold molar excess of reduced glutathione was added, and the proteins were resolved by reduced SDS-PAGE. The gels were stained with Coomassie blue, and bands corresponding to TSP were cut out, solubilized with Soluene-350 tissue solubilizer (Packard Instrument Co., Downers Grover, IL), and counted in a liquid scintillation counter with Ecolume scintillation cocktail (ICN Radiochemicals, Irvine, CA). The number of thiols was calculated from the specific activity of the NEM reagent. The number of thiols/mole at time zero was normalized to separate samples in which the value was determined after purification of the protein and determination of the amount of protein by amino acid composition, as described previously (8). Samples for other time points were related to this value. Except for Fig. 1, where statistical information is given, the data shown are from one of two (Figs. 2 and 3) or three (Figs. 4 and 5) identical experiments that gave essentially identical results. Heparin affinity chromatography. TSP was partially purified (approximately 80% homogeneous) for some experiments by chromatography on a column of heparin-agarose (Pierce Chemical Co., Rockford, IL) by the procedure of Lawler et al. (4) with the modifications of Speziale and Detwiler (8). SDS-PAGE. Electrophoresis was essentially as described by Laemmli (12) except for the use of AcrylAide (FMC Bioproducts, Rockland, ME) as an additional crosslinker. Other details have been described (8).
RESULTS Free thiols in TSP were quantified by reaction with labeled NEM followed by separation of labeled TSP by SDS-PAGE (8; see Materials and Methods). We previously showed that NEM reacted specifically with thiol groups in either native or denatured TSP (8). When the supernatant solution from ionophore A23187-activated platelets (i.e., a solution containing substances secreted by activated platelets) was incubated at 37°C after addition of 7 mM EDTA (Fig. l), conditions that lead to multimer formation (lo), the number of NEM-reactive groups in TSP increased to 1.7 + 0.3 times the initial value within 60 min. As with multimer formation (lo), the increase was inhibited by Ca2+ and was temperaturedependent (Fig. 1). The increase in free thiols was also blocked by incubation in 2% SDS (Fig. l), establishing
BONDS
547
IN THROMBOSPONDIN
I .:I: !2’ Y
-----------
J-----J 0
40
SO
120
Time (min) FIG. 1. An increase in the number of TSP thiols during incubation. The supernatant solution from activated platelets in 100 mM bis-Trispropane, pH 7.0, was incubated after addition of 7 mM EDTA (filled symbols, continuous line) or 2 mM CaCl, (open symbols, broken line) at either 25°C (A, A) or 37OC (0,O). At the indicated times, the samples were made 8 M with urea and derivatized with labeled NEM as described under Materials and Methods. TSP was separated by reducing SDSPAGE, and the amount of radioactivity associated with it was measured as described under Materials and Methods. For the experiments at 37’C, the points are averages from five different platelet preparations with bars for standard deviation. There were only two experiments at 25°C. One sample (0) contained 2% SDS added just before incubation; it validates the use of SDS to prevent the change in number of thiols.
the validity of 2% SDS as a means of terminating the reaction in the time-course studies. Identical results were obtained when TSP was resolved by two-dimensional (reduced/nonreduced) electrophoresis, minimizing the possibility that the increase in thiols was due to a comigrating contaminant (data not shown). We confirmed the previous report (10) that incubation led to no change in electrophoretic mobility of TSP on reduced SDS-PAGE and to a decrease in electrophoretic mobility on nonreduced SDS-PAGE due to the formation of multimers. The formation of multimers was reported to exhibit the same Ca2+ sensitivity and temperature dependence as observed here for the appearance of new thiols. The time courses of the two phenomena (Fig. 2) are similar but not identical; the half time for multimerization was about four times greater than that for thiol appearance. The reversibility of the appearance of new thiols in TSP can be demonstrated in two ways. First, addition of excess Ca2+ after incubation in EDTA not only inhibited the appearance of new thiols but also reversed the process (Fig. 3A). The data in Fig. 3B demonstrate that the protection by Ca2+ can also be reversed by addition of EDTA. Second, in a study of the pH dependence of the reaction, we observed that at pH 8 the increase in number of thiols was transient, returning to the initial value by 90 min (Fig. 4).
SPEZIALE
548
,I. Cl
40 SO Time (min)
120
AND
DETWILER
g
FIG. 2. The time courses of TSP multimer formation and the appearance of new thiols. The supernatant solution from activated platelets in 100 mM bis-Tris-propane, pH 7, was incubated at 37°C for the indicated times after addition of 7 mM EDTA. The samples were added to nonreduced SDS-PAGE sample buffer and reacted with labeled NEM as described under Materials and Methods. Half of each sample was analyzed on nonreduced SDS-PAGE for multimer formation, quantified as loss of monomeric TSP (W). To measure monomeric TSP, the band corresponding to monomeric TSP was cut from the Coomassie bluestained gel, the dye was eluted with 25% pyridine overnight, and the absorbance of the eluted dye at 605 nm was measured (18) and compared with the absorbance at zero time. The other half of each sample was analyzed on reducing SDS-PAGE for TSP-associated radioactivity (0).
The most likely mechanisms for the appearance of new thiols are the reduction of disulfide bonds or the hydrolysis of a thiol ester, as in C3 and C4 of the complement system and a2-macroglobulin (13). To test the possibility of a thiol ester, TSP samples in either 6 mM EDTA or 8 M urea/6 mM EDTA were preincubated with 100 mM methylamine, a reagent reported to cleave thiol esters (13). No additional thiols were found following this preincubation (data not shown), indicating that there are no thiol esters in TSP. We conclude that the new thiols probably were derived from the reduction of a disulfide bond(s) in TSP. After partial purification of TSP on a heparin affinity column (4,8), the number of thiols/mole of TSP was 2.9 at zero time and 2.8 after 90 min incubation at 37°C in EDTA, while in the control supernatant solution under the same conditions, the number of thiols/mole increased from 3.3 at zero time to 4.9 at 90 min. This suggested that the appearance of new thiols required TSP plus some additional component released by platelets. We used activated thiol-Sepharose, which covalently binds free available thiols, to test whether the additional component was a thiol compound. We incubated thiol-Sepharose or a control Sepharose with the supernatant solution from activated platelets in the presence of Ca2+, which protects the thiols of TSP from reaction with thiol-Sepharose (9,
Time (min) FIG. 3. The reversibility of the effects of EDTA and Ca2+ on the increase in the number of thiols in TSP. The supernatant solution from activated platelets was made 100 mM with bis-Tris-propane, pH 7, and incubated at 37°C. At the indicated times, nonreduced SDS-PAGE sample buffer was added, free thiols were derivatized with labeled NEM, and TSP was analyzed by reducing SDS-PAGE as described under Materials and Methods. (A) The sample contained 7 mM EDTA (0). At 15 min, 15 mM CaCl, was added to part of the sample (0). (B) The sample contained 2 mM CaC12 (plus approximately 2 mM Ca2+ secreted by the platelets) (0). At 15 (0) or 45 min (A), 15 mM EDTA was added.
10). The Sepharoses were removed by centrifugation. Pretreatment with activated thiol-Sepharose, but not with control Sepharose, prevented the increase in thiols (Fig.
I-
1 Cl
t
I
I
t
I
40 Time (min)
I
I
IL
80
FIG. 4. The effect of pH on the time-course of the change in the number of TSP thiols. The supernatant solution from activated platelets in 50 mM bis-Tris-propane was made 7 mM with EDTA and incubated at 37°C at pH 6 (A), 7 (O), or 8 (W). Nonreducing SDS-PAGE sample buffer was added, the samples were reacted with labeled NEM, and TSP was analyzed by reducing SDS-PAGE as described under Materials and Methods.
REDUCTION
OF DISULFIDE
BONDS
549
IN THROMBOSPONDIN
5), supporting the idea that a thiol compound released by activated platelets was responsible for the appearance of new thiols in TSP. DISCUSSION
Activated platelets secrete many things, from metal ions to enzymes and other proteins. This paper and previous papers from our laboratory (8-10) report chemical changes that occur to one of the major secreted proteins, TSP. Incubation of the secreted material leads to formation of disulfide-linked multimers of TSP and to an increase in the number of free thiols in TSP. The new thiols presumably are derived from reduction of one or more disulfide bonds. Over a period of 60 min, there was an appearance of 0.6 new thiols/polypeptide chain, or 1.8 thiols/TSP molecule. While there was some variation in this number, it is clear that there was not a general reduction of the 34 disulfide bonds in each polypeptide chain. The appearance of two new thiols/mole TSP by reduction of disulfide bonds could occur in two ways. There could be two thiol-disulfide exchange reactions involving two disulfide bonds of TSP; each of the reactions would lead to a new thiol in TSP and a mixed disulfide bond (first reaction, Eq. [l]). Alternatively, thiol-disulfide exchange with a single disulfide bond of TSP could proceed to complete reduction of a disulfide bond leading to two new thiols (first reaction plus second reaction, eq. [l]).
Tsp/$
yH\
ä
TSp/S-SR
,““z
l
‘SH
‘S
PI
2
1 TSP‘SH
Time (min) FIG. 5. Pretreatment with activated thiol-Sepharose prevents the increase in the number of TSP thiols. The supernatant solution from activated platelets was made 2 mM with CaCl,. An aliquot was incubated at 25°C for 20 min under nitrogen with no Sepharose (control, l ), with Sepharose 4-B (A) or with activated thiol-Sepharose (W). After removal of the Sepharoses by centrifugation, the samples were made 100 mM with bis-Tris-propane, pH 7.0, and 10 mM with EDTA and incubated at 37’C. At the indicated times, nonreduced SDS-PAGE sample buffer was added, the solutions were reacted with labeled NEM, and TSPassociated radioactivity was measured as described under Materials and Methods. The concentrations of TSP in the Sepharose 4-B and thiolSepharose samples were 60 and 24% of the control sample. Independent experiments demonstrated that this concentration difference would cause a slight difference in the initial rate of reaction but no difference in the number of new thiols formed.
+RS-SR
Either case involves reaction of a thiol compound, RSH, with TSP. Several observations support the required involvement of RSH. The intermolecular nature of the reaction was demonstrated by the fact that there was no increase in the number of thiols during incubation of TSP partially purified from the supernatant solution by heparin affinity chromatography. Further support for a bimolecular reaction comes from our observation (unpublished) that the rate of appearance of new thiols is only 20% as fast after a X-fold dilution of the supernatant solution. The fact that the other compound contained a reactive thiol group was demonstrated by the absence of the reaction in supernatant solutions pretreated with activated thiol-Sepharose (Fig. 5). The reversibility of the appearance of new thiols (Figs. 3 and 4) is also consistent with thiol-disulfide exchange reactions, as shown in Eq. [l]. This reversibility can be understood in two ways. First, Ca2+ stabilizes a conformation of TSP in which certain disulfide bonds in the Ca2+-sensitive region of the
molecule are more stable (i.e., the redox potential of the thiol-disulfide redox couple is more negutiue) as was shown previously (9), so that a disulfide bond that can be reduced by a redox system in EDTA may be reoxidized by the same redox system in Ca2+. This was seen as a Ca2+-induced loss of new thiols in Fig. 3A. In the second example of reversibility, higher pH, reduction was followed by reoxidation without any deliberate change in the system; there must have been a change in the redox state of the non-TSP redox system, perhaps due to slow oxidation by air at the higher pH. It is likely that the increase in the number of thiols in TSP is related to two other phenomena we have reported. First, we observed that a single equivalent of thiol in TSP is distributed among 12 different cysteine residues in the Ca2+-sensitive part of the TSP polypeptide chain (6); we inferred that labile disulfide bonds (8, 10) in the Ca2+sensitive region of TSP must be susceptible to attack by thiols, leading to isomerization of the disulfide bond pairings. Second, we observed that TSP forms disulfide-linked complexes with other TSP molecules and with other proteins (e.g., thrombin) (9). These two observations and the observations reported in this paper can be understood as manifestations of the attack by three different thiols on
550
SPEZIALE
AND
a labile disulfide bond. (i) Attack by a thiol on the same molecule would lead to isomerization of disulfide bonds, manifested as a distribution of the thiol among different cysteines without a change in the total number of thiols, as reported (8). (ii) Attack by a thiol on a different molecule of TSP would result in the formation of disulfidelinked dimers and multimers of TSP, as described (10). (iii) Attack by a thiol on some other released molecule, such as reduced glutathione or a small thiol protein, would result in the formation of a mixed disulfide bond and the appearance of a new thiol on TSP, as reported here. The evidence that the three phenomena are related is that each shows the same temperature dependence and the same Ca2+ dependence. While the formation of multimers appeared slower than the appearance of thiols (Fig. 2), that may be because we quantified multimer formation as the disappearance of TSP monomers, without consideration of the progression of the reaction from dimers to trimers and higher multimers; we thus measured a minimum rate of multimerization. We have no evidence about the type of thiol compound that is secreted and that reacts with TSP. It is likely that a low molecular weight thiol-disulfide redox system is the ultimate source of reducing equivalents, while reduction of disulfide bonds in TSP may possibly be mediated by a protein. As an example of a possible thiol-disulfide reductant, platelets contain large amounts of reduced glutathione (14-16), some of which may be released, as reported for other cells [ref. (17) for a review]. We considered the possibility that glutathione was derived from contaminating red cells, another rich source of reduced glutathione, but comparisons of platelet suspen-
DETWILER
sions without detectable red cells with those deliberately contaminated with red cells showed no difference in the appearance of new thiols of TSP. REFERENCES 1. Lawler, J. (1986) Blood 67, 1197-1209.
2. Frazier, W. A. (1987) J. Cell Biol. 106, 625-632. 3. Mosher, D. F. (1990) Annu. Reu. Med. 41,85-97. 4. Lawler, J., Chao, F. C., and Cohen, C. M. (1982) J. Biol. Chem. 267, 12,257-12,265.
5. Lawler, J., and Simons, E. R. (1983) J. Biol. Chem. 258, 12,09812,101.
6. Lawler, J., Derick, L. H., Connolly, 7. 8. 9. 10.
J. E., Chen, J.-H., and Chao, F. C. (1985) J. Biol. Chem. 260, 3762-3772. Lawler, J., and Hynes, R. 0. (1986) J. Cell Biol. 103,1635-1648. Speziale, M. V., and Detwiler, T. C. (1990) J. Biol. Chem. 265, 17859-17867. Danishefsky, K. J., Alexander, R. J., and Detwiler, T. C. (1984) Biochemistry 23,4984-4990. Turk, J. L., and Detwiler, T. C. (1986) Arch. Biochem. Biophys.
245,446-454. 11. Detwiler,
T. C., and Feinman, R. D. (1973) Biochemistry
12, 282-
289. 12. Laemmli, U. K. (1970) Nature 227, 680-682. 13. Sottrup-Jensen, L. (1987) in The Plasma Proteins (Putman, F. W., Ed.) Vol. 5, pp. 191-291, Academic Press, Orlando, FL. 14. Matsuda, S., Ikeda, Y., Aoki, M., Toyama, K., Watanabe, K., and Ando, Y. (1979) Thromb. Haemost. 42,1324-1341. 15. Hofmann, J., Losche, W., Till, W., Boscia, A., and Arese, P. (1980) Artery 8,431-435. 16. Dethmers, J. K., and Meister, A. (1981) Proc. Nutl. Acad. Sci. USA
78,7492-7496. 17. Meister, A., and Anderson, M. E. (1983) Annu. Reu. Biochem. 52, 711-760. 18. Fenner, C., Traut, R. R., Mason, D. T., and Wikman-Coffelt, J. (1975) Anal. Biochem. 63.603-606.
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 286, No. 2, May 1, pp. 546-550, 1991
Reduction of a Disulfide Bond of Thrombospondin Supernatant Solution of Activated Platelets’
in the
Mark V. Speziale and Thomas C. Detwiler2 Department
of Biochemistry,
State University
of New York Health Sciences Center at Brooklyn, Brooklyn,
New York 11203
Received October 4, 1990, and in revised form December 21, 1990
Incubation of the material secreted by activated platelets leads to the formation of disulfide-linked dimers and multimers of one of the proteins, thrombospondin. To determine whether these complexes formed as a result of thiol-disulfide exchange (no change in the number of thiols) or of oxidation of thiols (a decrease in the number of thiols), the number of thiols in TSP was measured during formation of multimers. The number of thiols increased from about 3/mol to 4.8/mol. The half-time for the disappearance of monomers of thrombospondin was fourfold greater than the half-time for appearance of new thiols. The appearance of new thiols, as well as the formation of multimers, was inhibited by Ca2+. The appearance of new thiols was reversible; addition of Ca2’ reversed the process, and at pH 8, but not at pH 6 or 7, the appearance of new thiols spontaneously reversed. No new thiols formed during incubation of partially purified thrombospondin or after the supernatant solution had been treated with activated thiol-Sepharose to remove reactive thiol compounds. It is concluded that thrombospondin has a disulfide bond that is unstable in the absence of Ca2+. It can be attacked by a thiol of another molecule of thrombospondin to form disulfide-linked multimers, by a thiol of the same molecule of thrombospondin to generate isomerization of disulfide bonds or, as observed in this study, by another secreted thiol compound to give a mixed disulfide and a new thiol. o 1991 Academic
Press.
Inc.
Thrombospondin (TSP)3 is a large glycoprotein (M, 420,000) secreted by activated platelets and by a variety 1 This work was supported by Grant HL37250 from the National Institutes of Health, United States Department of Health and Human Services. * To whom correspondence should be addressed. ’ Abbreviations used: TSP, Thrombospondin; Hepes, 4-(2-hydroxyethyl)-1-piperazineethane-sulfonic acid; NEM, N-ethylmalemide; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate.
of cultured cells [for reviews see Refs. (l-3)]. It is considered an adhesive protein with a role in cell-cell and cell-matrix adhesion, but its exact function is unknown. It is composed of three apparently identical disulfidelinked polypeptide chains. The conformation is Ca’+-dependent; the protein is more compact in the presence of Ca2+ (4-6). There are 69 cysteine residues in each polypeptide chain (7), with a single equivalent of free thiol (8). The remainder of the cysteines presumably are in intra- or interchain disulfide bonds. The single equivalent of free thiol is distributed among 12 different cysteine residues in the Ca2+-sensitive part of the molecule; this has been interpreted as evidence of intramolecular thioldisulfide exchange (8). Calcium ions partially protect the thiols from reaction with thiol-specific reagents and stabilize labile disulfide bonds (8-10). Secreted platelet thrombospondin undergoes thiol-dependent formation of disulfide-linked multimers (10). Thrombin becomes disulfide-linked to TSP (9), and it accelerates the formation of TSP multimers (10). Formation of multimers and formation of complexes with thrombin are inhibited by calcium ions (9,10). The ability of TSP to form disulfide-linked multimers and complexes with thrombin may reflect a broader potential of TSP to crosslink other proteins, such as those of the extracellular matrix and cell surface. To investigate the mechanism of the formation of disulfide-linked TSP multimers, we monitored changes in the thiol content of TSP during formation of multimers. A decrease in number of TSP thiols would indicate oxidation of thiols to intermolecular disulfide bonds, while no change in number of thiols would suggest thiol-disulfide exchange. We were surprised to find an increase in the number of thiols. We present evidence that this increase in free thiols is the result of an intermolecular reaction between a disulfide bond(s) in TSP and a separate thiol compound released by activated platelets. MATERIALS
AND
METHODS
Platelets. Human platelets were isolated by differential centrifugation of whole blood obtained from healthy volunteers using 0.15 vol of
546 All
Copyright 0 1991 rights of reproduction
0003.9%x/91 $3.00 by Academic Press, Inc. in any form reserved.
REDUCTION
OF DISULFIDE
120 mM trisodium titrate/6.6 mM EDTA as anticoagulant. The platelets were washed twice with Hepes-buffered saline (150 mM NaCl/lO mM Hepes, pH 7.4) and resuspended in this buffer to 1.5-4 X 10 platelets/ ml. The experiments described in this paper were performed with the supernatant solution prepared from a suspension of activated platelets. The suspension was made 100 mM with bis-Tris-propane, 2 mM with efluoride, 100 pM aminocaproic acid, 1 mM with phenylmethylsulfonyl with leupeptin, and 100 pM with CaCl,. The platelets were activated by 4 pM ionophore A23187 (Calbiochem, San Diego, CA), and the suspension was gently mixed for 3 min at room temperature. The platelets were removed by centrifugation. The supernatant solution was used immediately or placed on ice under N1. The supernatant solution contained the secreted material, including 0.5-2 mM Ca2+ (11). Measurement of free thiols of TSP. Free thiols were labeled with [3H]NEM (40-60 Ci/mmol; NEN Research Products, Boston, MA) as described (8). The samples were made 8 M with urea or were mixed with nonreducing SDS-PAGE sample buffer containing SDS sufficient to give a final concentration of 2%. The denaturing condition was necessary, because we found that the thiols of TSP are partially protected in the multimeric form of the protein prior to denaturation. NEM (50 pM) was added and the solution was incubated for 30 min at room temperature. A loo-fold molar excess of reduced glutathione was added, and the proteins were resolved by reduced SDS-PAGE. The gels were stained with Coomassie blue, and bands corresponding to TSP were cut out, solubilized with Soluene-350 tissue solubilizer (Packard Instrument Co., Downers Grover, IL), and counted in a liquid scintillation counter with Ecolume scintillation cocktail (ICN Radiochemicals, Irvine, CA). The number of thiols was calculated from the specific activity of the NEM reagent. The number of thiols/mole at time zero was normalized to separate samples in which the value was determined after purification of the protein and determination of the amount of protein by amino acid composition, as described previously (8). Samples for other time points were related to this value. Except for Fig. 1, where statistical information is given, the data shown are from one of two (Figs. 2 and 3) or three (Figs. 4 and 5) identical experiments that gave essentially identical results. Heparin affinity chromatography. TSP was partially purified (approximately 80% homogeneous) for some experiments by chromatography on a column of heparin-agarose (Pierce Chemical Co., Rockford, IL) by the procedure of Lawler et al. (4) with the modifications of Speziale and Detwiler (8). SDS-PAGE. Electrophoresis was essentially as described by Laemmli (12) except for the use of AcrylAide (FMC Bioproducts, Rockland, ME) as an additional crosslinker. Other details have been described (8).
RESULTS Free thiols in TSP were quantified by reaction with labeled NEM followed by separation of labeled TSP by SDS-PAGE (8; see Materials and Methods). We previously showed that NEM reacted specifically with thiol groups in either native or denatured TSP (8). When the supernatant solution from ionophore A23187-activated platelets (i.e., a solution containing substances secreted by activated platelets) was incubated at 37°C after addition of 7 mM EDTA (Fig. l), conditions that lead to multimer formation (lo), the number of NEM-reactive groups in TSP increased to 1.7 + 0.3 times the initial value within 60 min. As with multimer formation (lo), the increase was inhibited by Ca2+ and was temperaturedependent (Fig. 1). The increase in free thiols was also blocked by incubation in 2% SDS (Fig. l), establishing
BONDS
547
IN THROMBOSPONDIN
I .:I: !2’ Y
-----------
J-----J 0
40
SO
120
Time (min) FIG. 1. An increase in the number of TSP thiols during incubation. The supernatant solution from activated platelets in 100 mM bis-Trispropane, pH 7.0, was incubated after addition of 7 mM EDTA (filled symbols, continuous line) or 2 mM CaCl, (open symbols, broken line) at either 25°C (A, A) or 37OC (0,O). At the indicated times, the samples were made 8 M with urea and derivatized with labeled NEM as described under Materials and Methods. TSP was separated by reducing SDSPAGE, and the amount of radioactivity associated with it was measured as described under Materials and Methods. For the experiments at 37’C, the points are averages from five different platelet preparations with bars for standard deviation. There were only two experiments at 25°C. One sample (0) contained 2% SDS added just before incubation; it validates the use of SDS to prevent the change in number of thiols.
the validity of 2% SDS as a means of terminating the reaction in the time-course studies. Identical results were obtained when TSP was resolved by two-dimensional (reduced/nonreduced) electrophoresis, minimizing the possibility that the increase in thiols was due to a comigrating contaminant (data not shown). We confirmed the previous report (10) that incubation led to no change in electrophoretic mobility of TSP on reduced SDS-PAGE and to a decrease in electrophoretic mobility on nonreduced SDS-PAGE due to the formation of multimers. The formation of multimers was reported to exhibit the same Ca2+ sensitivity and temperature dependence as observed here for the appearance of new thiols. The time courses of the two phenomena (Fig. 2) are similar but not identical; the half time for multimerization was about four times greater than that for thiol appearance. The reversibility of the appearance of new thiols in TSP can be demonstrated in two ways. First, addition of excess Ca2+ after incubation in EDTA not only inhibited the appearance of new thiols but also reversed the process (Fig. 3A). The data in Fig. 3B demonstrate that the protection by Ca2+ can also be reversed by addition of EDTA. Second, in a study of the pH dependence of the reaction, we observed that at pH 8 the increase in number of thiols was transient, returning to the initial value by 90 min (Fig. 4).
SPEZIALE
548
,I. Cl
40 SO Time (min)
120
AND
DETWILER
g
FIG. 2. The time courses of TSP multimer formation and the appearance of new thiols. The supernatant solution from activated platelets in 100 mM bis-Tris-propane, pH 7, was incubated at 37°C for the indicated times after addition of 7 mM EDTA. The samples were added to nonreduced SDS-PAGE sample buffer and reacted with labeled NEM as described under Materials and Methods. Half of each sample was analyzed on nonreduced SDS-PAGE for multimer formation, quantified as loss of monomeric TSP (W). To measure monomeric TSP, the band corresponding to monomeric TSP was cut from the Coomassie bluestained gel, the dye was eluted with 25% pyridine overnight, and the absorbance of the eluted dye at 605 nm was measured (18) and compared with the absorbance at zero time. The other half of each sample was analyzed on reducing SDS-PAGE for TSP-associated radioactivity (0).
The most likely mechanisms for the appearance of new thiols are the reduction of disulfide bonds or the hydrolysis of a thiol ester, as in C3 and C4 of the complement system and a2-macroglobulin (13). To test the possibility of a thiol ester, TSP samples in either 6 mM EDTA or 8 M urea/6 mM EDTA were preincubated with 100 mM methylamine, a reagent reported to cleave thiol esters (13). No additional thiols were found following this preincubation (data not shown), indicating that there are no thiol esters in TSP. We conclude that the new thiols probably were derived from the reduction of a disulfide bond(s) in TSP. After partial purification of TSP on a heparin affinity column (4,8), the number of thiols/mole of TSP was 2.9 at zero time and 2.8 after 90 min incubation at 37°C in EDTA, while in the control supernatant solution under the same conditions, the number of thiols/mole increased from 3.3 at zero time to 4.9 at 90 min. This suggested that the appearance of new thiols required TSP plus some additional component released by platelets. We used activated thiol-Sepharose, which covalently binds free available thiols, to test whether the additional component was a thiol compound. We incubated thiol-Sepharose or a control Sepharose with the supernatant solution from activated platelets in the presence of Ca2+, which protects the thiols of TSP from reaction with thiol-Sepharose (9,
Time (min) FIG. 3. The reversibility of the effects of EDTA and Ca2+ on the increase in the number of thiols in TSP. The supernatant solution from activated platelets was made 100 mM with bis-Tris-propane, pH 7, and incubated at 37°C. At the indicated times, nonreduced SDS-PAGE sample buffer was added, free thiols were derivatized with labeled NEM, and TSP was analyzed by reducing SDS-PAGE as described under Materials and Methods. (A) The sample contained 7 mM EDTA (0). At 15 min, 15 mM CaCl, was added to part of the sample (0). (B) The sample contained 2 mM CaC12 (plus approximately 2 mM Ca2+ secreted by the platelets) (0). At 15 (0) or 45 min (A), 15 mM EDTA was added.
10). The Sepharoses were removed by centrifugation. Pretreatment with activated thiol-Sepharose, but not with control Sepharose, prevented the increase in thiols (Fig.
I-
1 Cl
t
I
I
t
I
40 Time (min)
I
I
IL
80
FIG. 4. The effect of pH on the time-course of the change in the number of TSP thiols. The supernatant solution from activated platelets in 50 mM bis-Tris-propane was made 7 mM with EDTA and incubated at 37°C at pH 6 (A), 7 (O), or 8 (W). Nonreducing SDS-PAGE sample buffer was added, the samples were reacted with labeled NEM, and TSP was analyzed by reducing SDS-PAGE as described under Materials and Methods.
REDUCTION
OF DISULFIDE
BONDS
549
IN THROMBOSPONDIN
5), supporting the idea that a thiol compound released by activated platelets was responsible for the appearance of new thiols in TSP. DISCUSSION
Activated platelets secrete many things, from metal ions to enzymes and other proteins. This paper and previous papers from our laboratory (8-10) report chemical changes that occur to one of the major secreted proteins, TSP. Incubation of the secreted material leads to formation of disulfide-linked multimers of TSP and to an increase in the number of free thiols in TSP. The new thiols presumably are derived from reduction of one or more disulfide bonds. Over a period of 60 min, there was an appearance of 0.6 new thiols/polypeptide chain, or 1.8 thiols/TSP molecule. While there was some variation in this number, it is clear that there was not a general reduction of the 34 disulfide bonds in each polypeptide chain. The appearance of two new thiols/mole TSP by reduction of disulfide bonds could occur in two ways. There could be two thiol-disulfide exchange reactions involving two disulfide bonds of TSP; each of the reactions would lead to a new thiol in TSP and a mixed disulfide bond (first reaction, Eq. [l]). Alternatively, thiol-disulfide exchange with a single disulfide bond of TSP could proceed to complete reduction of a disulfide bond leading to two new thiols (first reaction plus second reaction, eq. [l]).
Tsp/$
yH\
ä
TSp/S-SR
,““z
l
‘SH
‘S
PI
2
1 TSP‘SH
Time (min) FIG. 5. Pretreatment with activated thiol-Sepharose prevents the increase in the number of TSP thiols. The supernatant solution from activated platelets was made 2 mM with CaCl,. An aliquot was incubated at 25°C for 20 min under nitrogen with no Sepharose (control, l ), with Sepharose 4-B (A) or with activated thiol-Sepharose (W). After removal of the Sepharoses by centrifugation, the samples were made 100 mM with bis-Tris-propane, pH 7.0, and 10 mM with EDTA and incubated at 37’C. At the indicated times, nonreduced SDS-PAGE sample buffer was added, the solutions were reacted with labeled NEM, and TSPassociated radioactivity was measured as described under Materials and Methods. The concentrations of TSP in the Sepharose 4-B and thiolSepharose samples were 60 and 24% of the control sample. Independent experiments demonstrated that this concentration difference would cause a slight difference in the initial rate of reaction but no difference in the number of new thiols formed.
+RS-SR
Either case involves reaction of a thiol compound, RSH, with TSP. Several observations support the required involvement of RSH. The intermolecular nature of the reaction was demonstrated by the fact that there was no increase in the number of thiols during incubation of TSP partially purified from the supernatant solution by heparin affinity chromatography. Further support for a bimolecular reaction comes from our observation (unpublished) that the rate of appearance of new thiols is only 20% as fast after a X-fold dilution of the supernatant solution. The fact that the other compound contained a reactive thiol group was demonstrated by the absence of the reaction in supernatant solutions pretreated with activated thiol-Sepharose (Fig. 5). The reversibility of the appearance of new thiols (Figs. 3 and 4) is also consistent with thiol-disulfide exchange reactions, as shown in Eq. [l]. This reversibility can be understood in two ways. First, Ca2+ stabilizes a conformation of TSP in which certain disulfide bonds in the Ca2+-sensitive region of the
molecule are more stable (i.e., the redox potential of the thiol-disulfide redox couple is more negutiue) as was shown previously (9), so that a disulfide bond that can be reduced by a redox system in EDTA may be reoxidized by the same redox system in Ca2+. This was seen as a Ca2+-induced loss of new thiols in Fig. 3A. In the second example of reversibility, higher pH, reduction was followed by reoxidation without any deliberate change in the system; there must have been a change in the redox state of the non-TSP redox system, perhaps due to slow oxidation by air at the higher pH. It is likely that the increase in the number of thiols in TSP is related to two other phenomena we have reported. First, we observed that a single equivalent of thiol in TSP is distributed among 12 different cysteine residues in the Ca2+-sensitive part of the TSP polypeptide chain (6); we inferred that labile disulfide bonds (8, 10) in the Ca2+sensitive region of TSP must be susceptible to attack by thiols, leading to isomerization of the disulfide bond pairings. Second, we observed that TSP forms disulfide-linked complexes with other TSP molecules and with other proteins (e.g., thrombin) (9). These two observations and the observations reported in this paper can be understood as manifestations of the attack by three different thiols on
550
SPEZIALE
AND
a labile disulfide bond. (i) Attack by a thiol on the same molecule would lead to isomerization of disulfide bonds, manifested as a distribution of the thiol among different cysteines without a change in the total number of thiols, as reported (8). (ii) Attack by a thiol on a different molecule of TSP would result in the formation of disulfidelinked dimers and multimers of TSP, as described (10). (iii) Attack by a thiol on some other released molecule, such as reduced glutathione or a small thiol protein, would result in the formation of a mixed disulfide bond and the appearance of a new thiol on TSP, as reported here. The evidence that the three phenomena are related is that each shows the same temperature dependence and the same Ca2+ dependence. While the formation of multimers appeared slower than the appearance of thiols (Fig. 2), that may be because we quantified multimer formation as the disappearance of TSP monomers, without consideration of the progression of the reaction from dimers to trimers and higher multimers; we thus measured a minimum rate of multimerization. We have no evidence about the type of thiol compound that is secreted and that reacts with TSP. It is likely that a low molecular weight thiol-disulfide redox system is the ultimate source of reducing equivalents, while reduction of disulfide bonds in TSP may possibly be mediated by a protein. As an example of a possible thiol-disulfide reductant, platelets contain large amounts of reduced glutathione (14-16), some of which may be released, as reported for other cells [ref. (17) for a review]. We considered the possibility that glutathione was derived from contaminating red cells, another rich source of reduced glutathione, but comparisons of platelet suspen-
DETWILER
sions without detectable red cells with those deliberately contaminated with red cells showed no difference in the appearance of new thiols of TSP. REFERENCES 1. Lawler, J. (1986) Blood 67, 1197-1209.
2. Frazier, W. A. (1987) J. Cell Biol. 106, 625-632. 3. Mosher, D. F. (1990) Annu. Reu. Med. 41,85-97. 4. Lawler, J., Chao, F. C., and Cohen, C. M. (1982) J. Biol. Chem. 267, 12,257-12,265.
5. Lawler, J., and Simons, E. R. (1983) J. Biol. Chem. 258, 12,09812,101.
6. Lawler, J., Derick, L. H., Connolly, 7. 8. 9. 10.
J. E., Chen, J.-H., and Chao, F. C. (1985) J. Biol. Chem. 260, 3762-3772. Lawler, J., and Hynes, R. 0. (1986) J. Cell Biol. 103,1635-1648. Speziale, M. V., and Detwiler, T. C. (1990) J. Biol. Chem. 265, 17859-17867. Danishefsky, K. J., Alexander, R. J., and Detwiler, T. C. (1984) Biochemistry 23,4984-4990. Turk, J. L., and Detwiler, T. C. (1986) Arch. Biochem. Biophys.
245,446-454. 11. Detwiler,
T. C., and Feinman, R. D. (1973) Biochemistry
12, 282-
289. 12. Laemmli, U. K. (1970) Nature 227, 680-682. 13. Sottrup-Jensen, L. (1987) in The Plasma Proteins (Putman, F. W., Ed.) Vol. 5, pp. 191-291, Academic Press, Orlando, FL. 14. Matsuda, S., Ikeda, Y., Aoki, M., Toyama, K., Watanabe, K., and Ando, Y. (1979) Thromb. Haemost. 42,1324-1341. 15. Hofmann, J., Losche, W., Till, W., Boscia, A., and Arese, P. (1980) Artery 8,431-435. 16. Dethmers, J. K., and Meister, A. (1981) Proc. Nutl. Acad. Sci. USA
78,7492-7496. 17. Meister, A., and Anderson, M. E. (1983) Annu. Reu. Biochem. 52, 711-760. 18. Fenner, C., Traut, R. R., Mason, D. T., and Wikman-Coffelt, J. (1975) Anal. Biochem. 63.603-606.
