Proc. Nat!. Acad. Sci. USA Vol. 75, No. 7, pp. 3085-3089, July 1978
Biochemistry
Synthetic peptide derivatives that bind to fibrinogen and prevent the polymerization of fibrin monomers (intermolecular contact sites/protein-ligand interactions)
ANDREW P. LAUDANO AND RUSSELL F. DOOLITTLE Department of Chemistry, University of California at San Diego, La Jolla, California 92093
Communicated by J. Edwin Seegmiller, April 7,1978
ABSTRACT A series of small peptides corresponding to the amino termini of the fibrin a- and a-chains has been synthesized. The Reptides glycyl-L-prolyl-L-arginyl-L-proline and are potent inhibitors of fibrin glycyl-L-prolyl-L-arginylsarcosine polymerization. Moreover, these peptides have a natural stability stemming from their inherent resistance to proteolysis because of the involvement of imino acids in each of their peptide bonds. The peptide glycyl-L-prolylL-arginyl-L-proline binds to fibrinogen and to fragment D, in both cases with an association constant of approximately 5 X 104; it does not bind to fragment E. The number of binding sites is two for fibrinogen and one for fragment D. The tripeptide lycyl-L-prolyl-L-arginine binds less tightly and is less than ha f as effective in preventing polymerization. The peptide glycyl-L-histidyl-L-arginyl-L-proline, which corresponds exactly to the amino terminus of the-fibrin a-chain, does not inhibit the aggregation of fibrin monomers under the conditions used. It does bind weakly to fibrinogen, however, suggesting the involvement of sites other than those binding the a-chain analogues. Various other peptides were found not to inhibit polymerization; these included glycine-L-proline, L-prolyl-L-arginine and glycyl-L-prolyl-Lseryl-L-proline. The last-named corresponds to the serine/arginine amino acid replacement previously reported for a defective human fibrinogen. Vertebrate fibrinogen is transformed into fibrin by the thrombin-catalyzed release of small polar peptides from the amino termini of the a- and i-chains (1). The parent molecules, shorn of these peptides (the fibrinopeptides A and B, respectively), spontaneously polymerize: thrombin
fibrinogen -b' fibrin monomers + fibrinopeptides fibrin monomers
spontaneous
1-'
fibrin
A generation ago Bettelheim and Bailey (2) hypothesized that the newly exposed amino termini were involved as principal contact sites for the polymerization event. At the same time, the most widely accepted polymerization scheme of the day depended on the staggered overlap of rod-like molecules such that the central portion of one molecule interacted with the terminal regions of neighbors (3). Today we know that all six amino termini of vertebrate fibrinogen molecules are tightly gathered in a single domain (4, 5). There is also a plethora of evidence favoring the notion that this domain is equivalent to the central domain of a triglobular molecule of the sort proposed by Hall and Slayter (6). As such, it has been proposed that the release of fibrinopeptides from the central domain allows intermolecular contact between that central domain and the terminal domains of other molecules during the polymerization event (7). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertiement" in accordance with 18 U. S. C. §1734 solely to indicate
this fact.
3085
With this scheme of events in mind, we set out to synthesize a series of peptides corresponding to those regions of fibrin that are exposed by the action of thrombin. We reasoned that these peptides ought to be able to prevent polymerization by binding to the receptor sites on the terminal domains. In particular, we synthesized a family of peptides beginning with the sequence
glycyl-L-prolyl-L-arginine (Gly-Pro-Arg), the terminal tripeptide of the fibrin a-chain. Although Gly-Pro-Arg itself was found to be an inhibitor of polymerization, we found that the addition of another residue increased the binding and inhibitory action significantly. Moreover, by restricting the additional
residue to the secondary amino acids proline and sarcosine, we were able to confer a latent stability on these peptides because all the peptide bonds involved are inherently resistant to attack by most proteases and peptidases. We also synthesized various peptides that did not inhibit
polymerization. These included glycyl-L-proline, L-prolyl-
L-arginine, glycyl-L-histidyl-L-arginyl-L-proline, and glycylL-prolyl-L-seryl-L-proline. The last-named tetrapeptides are of particular interest because one corresponds to the amino terminus of the fibrin ,8-chain and the other, to the amino acid replacement (serine for arginine) reported for the defective human fibrinogen variant known as fibrinogen Detroit (8).
MATERIALS AND METHODS Human blood plasma was obtained from the San Diego Blood Bank and fibrinogen prepared according to a previously described procedure (9). Fragments D and E were prepared by plasmin digestion of fibrinogen followed by DEAE chromatography along the lines originally described by Nussenzweig et al. (10). A complete description of the digestion and separation conditions appears in a previous article from this laboratory (11). Fibrin monomers were prepared by dispersing fibrin in 1 M NaBr/0.05 M Na acetate, pH 5.3 (12). In this regard, human fibrinogen (10-15 mg/ml) was clotted with bovine thrombin (Parke-Davis, 1 unit/ml final conc.); the final solution contained 150 mM NaCl, 1 mM EDTA, and 2.5 mM phosphate, pH 7.0. After the clot had been allowed to form at room temperature for approximately 30 min, it was wound out on a glass rod and dispersed in a minimal volume of the sodium bromide solution. The dispersion was continued with gentle stirring for 8 hr. at which point it was centrifuged to remove any undispersed material. The A 2o of the dispersed fibrin preparations (fibrin monomers) was generally about 17 (11 mg/ml). Reaggregation of the fibrin was effected by adding 25 Al of the fibrin monomer preparation to 0.5 ml of 80 mM phosphate buffer, pH 6.3, in a 1-cm path-length cuvette. The solution was mixed immediately and reaggregation was monitored by measuring the scattered light in a Zeiss spectrophotometer at 350 nm (13). Peptide additives were present in the phosphate buffer in advance of the addition of fibrin monomers. The
3086
Biochemistry: Laudano and Doolittle
additives included peptides we synthesized ourselves as well as some other commercially prepared peptides used as controls. Peptide Synthesis. Solid-phase peptide synthesis was conducted by established Merrifield procedures (14, 15), and a complete description, including materials, yields, etc., will appear elsewhere. In essence, however, the appropriate Bocamino acids (Boc-proline, Boc-nitroarginine, and Boc-sarcosine) were refluxed (in ethanol containing triethylamine) with chloromethylated polystyrene resin (Bio-Rad); the degree of substitution attained was generally about 0.4 mmol of amino acid per g of resin. The coupling of subsequent amino acids was achieved by the dicyclohexylcarbodiimide method (14, 15). Peptides were cleaved from the resin with HBr/trifluoroacetic acid for 90 min at room temperature (15). The nitro group of nitroarginine-containing peptides was removed by catalytic hydrogenation using 10% Pd/charcoal in acetic acid at a hydrogen pressure of 300 mm Hg for 48 hr. Arginine-containing peptides were purified by paper electrophoresis at pH 6.4. Peptides which did not contain arginine were purified by chromatography on Dowex 50W-X2. Amino acid analysis was conducted on a Spinco model 119 amino acid analyzer after total acid hydrolysis (5.7 M HCO, 108°, 24 hr in sealed tubes). Stock solutions of the various peptides were prepared in H20, and the concentrations were verified by amino acid analysis. The specific activities of radioactive peptides (containing [14C]glycine) were generally about 11 ;Ci/mmol peptide. Equilibrium Dialysis. Equilibrium dialysis was conducted in small test tubes using standard No. 8 Visking dialysis tubing. In the case of fibrinogen binding studies, 1.0 ml of fibrinogen solution (10 mg/ml) was dialyzed against 5 ml of a solution containing 0.2 M NaCl, 0.05 M phosphate buffer (pH 7.0), and various concentrations of the radioactive peptide under study. Binding studies with fragments D and E were performed similarly except that the protein concentration was reduced to 5 mg/ml. These latter studies were confined to the peptide glycyl-L-prolyl-L-arginyl-L-proline; in this case a bovine plasma albumin (10 mg/ml) control was also performed simultaneously. After 8 hr equilibration at room temperature, the dialysis bags were cut open and 0.5 ml aliquots of the inside and outside solutions were removed for scintillation counting. Final concentrations of the protein solutions were also checked at this point by measuring the A28o. Thrombin/Fibrinogen Clotting Assays. These tests were conducted by adding various concentrations of thrombin (Parke-Davis, bovine) to a human fibrinogen solution (1.5 mg/ml) containing various concentrations of peptide additive and then determining the clotting time. The final solution also contained 150 mM NaCl and 5 mM phosphate buffer, pH 7.2.
RESULTS Thrombin/Fibrinogen Clotting Assays. The peptides Gly-Pro-Arg-Sar, Gly-Pro-Arg-Pro, and Gly-Pro-Arg all exhibited significant inhibition of clotting in that they prevented or slowed the appearance of fibrin gels when added as a component of thrombin/fibrinogen mixtures (Table 1). Under the conditions used (fibrinogen = 1.5 mg/ml), the most effective peptides, Gly-Pro-Arg-Pro and Gly-Pro-Arg-Sar, completely prevented clotting at concentrations of about 10-3 M, and, even at 10-4 M, gel formation was significantly delayed. Gly-Pro-Arg exhibited significantly weaker inhibitory action. The inhibition of clotting is likely due in fact to an inhibition of fibrin polymerization.
Fibrin Monomer Reaggregation Studies. Gly-Pro-Arg-Sar
Proc. Natl. Acad. Sci. USA 75 (1978) Table 1. Inhibition and binding properties of synthetic peptides
Peptide Gly-Pro-Arg-Sar Gly-Pro-Arg-Pro Gly-His-Arg-Pro Gly-Pro-Ser-Pro Gly-Pro-Arg Pro-Arg Gly-Pro Gly
Thrombin/ Inhibition of Binding to fibrinogen reaggregation fibrinogen (equilibrium of fibrin clotting time,* dialysis)t monomerst min >200 >200 20 20 42 21 20 ND
++++ ++++
ND ++++
+ ++ -
++ ND
ND = not done; - = no inhibitory effect or no binding. * All peptides at final concentration of 1 mM; control = 20 min. t Various peptide concentrations were studied; see Fig. 1. Other ineffective peptides tested in this system included Gly-Leu-Tyr, Ala-Pro-Gly, tosyl-Arg-Sar-methylester (16), glycylglycine ethylester, and glycine ethylester. All peptides (except Pro-Arg) were labeled with [14C]glycine; see Fig. 2 and Table 2.
and Gly-Pro-Arg-Pro both proved to be powerful inhibitors of fibrin monomer reaggregation. Moreover, the effectiveness of the inhibition depended on both the concentration of the peptide inhibitor and the fibrin monomer concentration. In order to completely prevent polymerization, an approximately 100-fold molar excess of peptide to fibrin monomer was required, although significant retardation of polymerization occurred at ratios as low as 20:1 (these ratios allow for the fact that the fibrin "monomer" is actually dimeric). The peptide Gly-Pro-Arg proved to be less than half as effective as the tetrapeptides used (Fig. 1), and the dipeptides Gly-Pro and Pro-Arg had no detectable effect. Various unrelated peptides and amino acid derivatives (Table 1) and the tetrapeptides Gly-His-Arg-Pro and Gly-Pro-Ser-Pro also had no effect. Gly-His-Arg-Pro is the sequence found at the amino terminus of the fibrin 3-chain, and Gly-Pro-Ser-Pro was modeled on the amino acid replacement that has been reported for the variant human fibrinogen known as fibrinogen Detroit (8). As in the case of the defective fibrinogen, substitution of the arginine residue by serine completely obliterated its effectiveness. Equilibrium Dialysis Studies. Radioactive Gly-Pro-Arg-Pro was found to bind to human fibrinogen with an association constant of about 4 X 104 (Fig. 2). The number of sites at saturation extrapolated to two per 340,000 Mr. The peptide did not bind detectably to a preparation of bovine plasma albumin (10 mg/ml) run simultaneously. The tripeptide Gly-Pro-Arg bound less than half as tightly, but it also indicated the presence of two binding sites per fibrinogen molecule. Radioactive glycine, Gly-Pro, and Gly-Pro-Ser-Pro did not exhibit any binding under similar conditions (Table 1). Gly-His-Arg-Pro bound with an association constant of only 9 X 103 at pH 7.0, and the binding dropped off to less than half that at pH 6.5. In either case the data still indicated two binding sites. Gly-Pro-Arg-Pro bound to fragment D with an association constant of about 6 X 104. The number of sites in this case extrapolated to n = 1, when the observed Mr of 80,000 for our fragment D preparations (11) was used. Concentrations were specified by weighing dry fragment D (5.0 mg/ml) and verified throughout the experiment using an EM of 21 (17). The GlyPro-Arg-Pro preparation did not bind to fragment E.
Biochemistry: Laudano and Doolittle
Time, min
Proc. Nati. Acad. Sci. USA 75 (1978)
3087
Time, min
1.21
at 0.8
0.4
20 Time, min
40
Time, min
FIG. 1. Inhibition of the reaggregation of fibrin monomers by various synthetic peptides. All concentrations given are final concentrations. (Upper left) Increasing concentrations of Gly-Pro-Arg-Pro ranging from 0.1 mM to 0.33 mM (complete inhibition); kcontrol = A. (Upper right) Comparison of effectiveness of Gly-Pro-Arg (*, 0.16 mM) with Gly-Pro-Arg-Pro (0, 0.18 mM); control = A. (Lower left) Comparison of effectiveness of Gly-Pro-Ser-Pro (ED) with Gly-Pro-Arg-Pro (o). Both peptides at 0.27 mM; control = A. (Lower right) Comparison of effectiveness of Gly-His-Arg-Pro (X) with Gly-Pro-Arg-Sar (A). Both peptides at 0.24 mM; control = A.
DISCUSSION The tripeptidyl sequence Gly-Pro-Arg has endured at the amino terminus of fibrin a-chains for more than 400 million yr, being present in the fibrin of the most primitive vertebrates extant (19) as well as in mammalian fibrin (Table 2). This persistence is altogether consistent with a strategic structure-function interdependence. Moreover, replacement of the arginine residue with a serine at this position has been reported in the case of a defective human fibrinogen (8), the polymerization of which is impaired (22). Both of these observations-the one evolutionary and the other genetic-are consistent with early notions that the newly exposed amino terminus ought to be involved in some kind of intermolecular association leading to fibrin polymerization. In formulating the experiments described in this report, we visualized the unmasked sequence as a simple knob which could become involved in a knob-hole interaction. In undertaking the syntheses, we added a fourth residue, because in the natural situation the peptide chain continues on its course, and because the presence of a free carboxyl group might offset the effect of the positively charged arginine side chain. We could have esterified the carboxyl group, of course, but we
wanted something more stable chemically. Similarly, we could have incorporated the fourth residue of the natural sequence, which is a valine in the human molecule but varies in other species (Table 2). The fact that this particular Arg-Val bond has Table 2. Amino acid sequences at the amino termini of fibrin a- and ,-chains from various species Chain Ref. Sequence 1 2 3 4 5 6 7 Humana 18 Gly-Pro-Arg-Val -Val-Glu -Arg. . . Bovine a 19 Gly-Pro-Arg-Leu-Val -Glu-Lys... 20 Dog a Gly-Pro-Arg-Ile -Val -Glu-Arg ... Chicken a 21 Gly-Pro-Arg-Ile -Leu-Glu-Asn... 19 Lamprey a Gly-Pro-Arg-Leu-X* -Glx-Glx ... Human 18 Gly-His-Arg-Pro -Leu-Asp-Lys... Bovine 19 Gly-His-Arg-Pro-Tyr-Asx-Lys... Dog, 20 Gly-His-Arg-Pro-Leu-Asp-Lys... Chicken fi 21 Gly-His-Arg-Pro-Leu-Asp-Lys . . . 19 Lamprey Gly-Val-Arg-Pro--Leu-Pro -X* ... $ Probably Ser or Thr.
3088
Biochemistry: Laudano and Doolittle
Proc. Natl. Acad. Sci. USA 75 (1978)
A
A
X X
U
A
Cc
x A o
20
°0
A
1.0
2.0 R
FIG. 2. Plots of binding data from various equilibrium dialysis studies in which radioactive peptides Gly-Pro-Arg-Pro (A), Gly-Pro-Arg (x), and Gly-His-Arg-Pro (0) were exposed to fibrinogen (see also Table 1). R = moles of bound peptide/mole of fibrinogen; C = concentration of peptide. Fibrinogen concentration = 11 mg/ml (3.2 x 10-2mM).
been reported to be cleaved by thrombin (23) and by plasmin (24), however, convinced us to incorporate secondary amino acids (imino-) at this position in order to prevent undesirable proteolysis during our experiments. It also did not escape our attention that these compounds have potential therapeutic value as antipolymerants, and long-term stability and invulnerability to proteases would obviously be advantageous in such a setting. Previous reports from this laboratory have dealt with the effectiveness of the Arg-Sar bond in peptides designed as thrombin inhibitors (16). As a result, we realized that the synthetic tetrapeptides reported in this article could have antithrombin activity independent of the fibrin antipolymerant action. As a consequence we turned to fibrin monomer reaggregation studies in order to evaluate their effectiveness. Thus, Gly-Pro-Arg ought not to be a thrombin inhibitor per se, because it not only has an unprotected carboxyl group but is itself a product of thrombin cleavage. Ironically, Gly-Pro-Arg has been synthesized by others in the past and reported to be a thrombin inhibitor (25). The thrombin plasma assay used by those workers obviously precluded their discovering its actual mode of action, its "anti-thrombin" effect most likely being entirely due to the inhibition of fibrin polymerization. The observation that the peptides which inhibit fibrin polymerization bind to fragment D was consistent with the proposed scheme whereby a reciprocal association occurs between the central and terminal domains of neighboring molecules (7). It is also consistent with experiments in which fi-
brinogen and fragment D can be absorbed to immobilized fibrin (26). The fact that the tetrapeptide Gly-His-Arg-Pro, which corresponds to the terminal sequence exposed on the E-chain of fibrin, does not impair fibrin reaggregation is especially interesting, although the data have to be carefully considered. First, it must be remarked that technical difficulties make fibrin reaggregation studies difficult above pH6.5. Under the conditions of our study (pH 6.3), the histidine was presumably protonated. In line with this, the binding of Gly-His-Arg-Pro to fibrinogen was significantly less at pH 6.5 than at pH 7.0 or 7.3. On the other hand, Gly-His-Arg-Pro had no effect on the clotting time in the thrombin/fibrinogen assays conducted at pH 7.3, suggesting that the binding involves a set of sites that are different from those that are involved in initial polymerization. Although further binding and competition studies ought to resolve this point, the present observations are consistent with the well-known fact that fibrin polymerization in mammals can occur after release of the fibrinopeptide A only (27, 28). The strength of the association of the tetrapeptides GlyPro-Arg-Sar and Gly-Pro-Arg-Pro with fibrinogen is sufficient to ensure that affinity labeling studies (29) using similar compounds with functional reactive groups attached are feasible. It will be of great'interest to learn which portions of the individual chains are involved in the actual sites. Also of interest will be studies with lamprey fibrinogen because that material can be clotted by the exclusive removal of the fibrinopeptide B (19, 30). Interestingly enough, its (3-chain tetrapeptide varies from
Biochemistry: Laudano and Doolittle the mammalian sequences by a valine/histidine replacement (Table 2). We thank Marcia Riley for technical assistance in preparing fibrinogen and fragments D and E, Dennis Trovato for the operation of the amino acid analyzer, and Cheryl Yokahama for assistance with some of the equilibrium dialysis studies. These studies were supported by U.S. Public Health Service Grants HL-18,576 and GM-17,702. 1. Bailey, K., Bettelheim, F. R., Lorand, L. & Middlebrook, W. R., (1951) Nature 167,233-234. 2. Bettelheim, F. R. & Bailey, K. (1952) Biochim, Biophys. Acta 9, 578-579. 3. Ferry, J. D., Katz, S. & Tinoco, I., Jr. (1954) J. Polym. Sci. 12, 509-516. 4. Marder, V. (1971) Scand. J. Haematol. Supp, 13,21-36. 5. Blombick, B. & Blombick, M. (1972) Ann. N.Y. Acad. Scd. 202, 77-97. 6. Hall, C. E. & Slayter, H. D. (1959) J. Biophys. Biochem. Cytol. 5, 11-16. 7. Doolittle, R. F. (1973) Adv. Prot. Chem. 25,1-109. 8. Blombeck, M., Blombick, B., Mammen, E. F. & Prasad, A. S. (1968) Nature 218, 134-137. 9. Doolittle, R. F., Schubert, D. & Schwartz, S. A. (1967) Arch. Blochem. Blophys., 118,456-467. 10. Nussenzweig, V., Seligman, M., Pelmont, J. & Grabar, P. (1961) Ann. Inst. Pasteur, Paris 100, 377-389. 11. Doolittle, R. F., Cassman, K. G., Cottrell, B. A., Friezner, S. J. & Takagi, T. (1977) Biochemistry 16, 1710-1715. 12. Donnelly, T. H., Laskowski, M., Jr., Notley, N. & Scheraga, H. A. (1955) Arch. Biochem. Biophys. 56,369-387. 13. Latallo, Z., Fletcher, A. P., Alkljaersig, N. & Sherry, S. (1962) Am. J. Physiol. 202,675-680. 14. Merrifield, R. B. (1964) J. Am. Chem. Soc. 85,2149-2154.
Proc. Natl. Acad. Sci. USA 75 (1978)
3089
15. Stewart, J. M. & Young, J. D. (1969) Solid Phase Peptide Syntheis, (Freeman, San Francisco, CA). 16. Weinstein, M. J. & Doolittle, R. F. (1972) Biochim. Biophys. Acta 258,577-590. 17. Marder, V. J., Shulman, N. R. & Carroll, W. R. (1969) J. Biol. Chem. 244,2111-2119. 18. Iwanaga, S., Wallen, P., Grondahl, N. J., Henschen, A. & Blombick, B. (1967) Biochim. Biophys. Acta 147,606-609. 19. Cottrell, B. A. & Doolittle, R. F. (1976) Biochim. Biophys. Acta 453,426438. 20. Birken, S., Wilner, G. D. & Canfield, R. E. (1975) Thromb. Res. 7,599-610. 21. Murano, G., Walz, D., Williams, L., Pindyek, J. & Mosesson, M. W. (1977) Thromb. Res. 11, 1-10. 22. Mammen, E. F., Prasad, A. S., Barnhart, M. L. & Au, C. C. (1969) J. Clin. Invest. 48,235-249. 23. Blombdck, B., Blombeck, M., Henschen, A., Hessel, B., Iwanaga, S. & Woods, K. R. (1968) Nature 218,130-134. 24. Takagi, T. & Doolittle, R. F. (1975) Biochemistry 14, 940946. 25. Dorman, L. E., Cheng, R. C. & Marshall, F. N. (1972) Chemistry and Biology of Peptides (Ann Arbor Science Publishers, Ann Arbor, MI), pp. 455-459. 26. Kudryk, B. J., Collen, D., Woods, K. R. & Blomback, B. (1974) J. Biol. Chem. 249,3322-325. 27. Blomback, B., Blomback, M. & Nilsson, I. M. (1957) Thromb. Diath. Haemorrh. 15,76-86. 28. Doolittle, R. F., Oncley, J. L. & Surgenor, D. M. (1962) J. Biol. Chem. 237,3123-3127. 29. Wofsy, L., Metzger, H. & Singer, S. J. (1962) Biochemistry 1, 1031-1039. 30. Doolittle, R. F. (1965) Biochem. J., 94,735-741.
Biochemistry
Synthetic peptide derivatives that bind to fibrinogen and prevent the polymerization of fibrin monomers (intermolecular contact sites/protein-ligand interactions)
ANDREW P. LAUDANO AND RUSSELL F. DOOLITTLE Department of Chemistry, University of California at San Diego, La Jolla, California 92093
Communicated by J. Edwin Seegmiller, April 7,1978
ABSTRACT A series of small peptides corresponding to the amino termini of the fibrin a- and a-chains has been synthesized. The Reptides glycyl-L-prolyl-L-arginyl-L-proline and are potent inhibitors of fibrin glycyl-L-prolyl-L-arginylsarcosine polymerization. Moreover, these peptides have a natural stability stemming from their inherent resistance to proteolysis because of the involvement of imino acids in each of their peptide bonds. The peptide glycyl-L-prolylL-arginyl-L-proline binds to fibrinogen and to fragment D, in both cases with an association constant of approximately 5 X 104; it does not bind to fragment E. The number of binding sites is two for fibrinogen and one for fragment D. The tripeptide lycyl-L-prolyl-L-arginine binds less tightly and is less than ha f as effective in preventing polymerization. The peptide glycyl-L-histidyl-L-arginyl-L-proline, which corresponds exactly to the amino terminus of the-fibrin a-chain, does not inhibit the aggregation of fibrin monomers under the conditions used. It does bind weakly to fibrinogen, however, suggesting the involvement of sites other than those binding the a-chain analogues. Various other peptides were found not to inhibit polymerization; these included glycine-L-proline, L-prolyl-L-arginine and glycyl-L-prolyl-Lseryl-L-proline. The last-named corresponds to the serine/arginine amino acid replacement previously reported for a defective human fibrinogen. Vertebrate fibrinogen is transformed into fibrin by the thrombin-catalyzed release of small polar peptides from the amino termini of the a- and i-chains (1). The parent molecules, shorn of these peptides (the fibrinopeptides A and B, respectively), spontaneously polymerize: thrombin
fibrinogen -b' fibrin monomers + fibrinopeptides fibrin monomers
spontaneous
1-'
fibrin
A generation ago Bettelheim and Bailey (2) hypothesized that the newly exposed amino termini were involved as principal contact sites for the polymerization event. At the same time, the most widely accepted polymerization scheme of the day depended on the staggered overlap of rod-like molecules such that the central portion of one molecule interacted with the terminal regions of neighbors (3). Today we know that all six amino termini of vertebrate fibrinogen molecules are tightly gathered in a single domain (4, 5). There is also a plethora of evidence favoring the notion that this domain is equivalent to the central domain of a triglobular molecule of the sort proposed by Hall and Slayter (6). As such, it has been proposed that the release of fibrinopeptides from the central domain allows intermolecular contact between that central domain and the terminal domains of other molecules during the polymerization event (7). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertiement" in accordance with 18 U. S. C. §1734 solely to indicate
this fact.
3085
With this scheme of events in mind, we set out to synthesize a series of peptides corresponding to those regions of fibrin that are exposed by the action of thrombin. We reasoned that these peptides ought to be able to prevent polymerization by binding to the receptor sites on the terminal domains. In particular, we synthesized a family of peptides beginning with the sequence
glycyl-L-prolyl-L-arginine (Gly-Pro-Arg), the terminal tripeptide of the fibrin a-chain. Although Gly-Pro-Arg itself was found to be an inhibitor of polymerization, we found that the addition of another residue increased the binding and inhibitory action significantly. Moreover, by restricting the additional
residue to the secondary amino acids proline and sarcosine, we were able to confer a latent stability on these peptides because all the peptide bonds involved are inherently resistant to attack by most proteases and peptidases. We also synthesized various peptides that did not inhibit
polymerization. These included glycyl-L-proline, L-prolyl-
L-arginine, glycyl-L-histidyl-L-arginyl-L-proline, and glycylL-prolyl-L-seryl-L-proline. The last-named tetrapeptides are of particular interest because one corresponds to the amino terminus of the fibrin ,8-chain and the other, to the amino acid replacement (serine for arginine) reported for the defective human fibrinogen variant known as fibrinogen Detroit (8).
MATERIALS AND METHODS Human blood plasma was obtained from the San Diego Blood Bank and fibrinogen prepared according to a previously described procedure (9). Fragments D and E were prepared by plasmin digestion of fibrinogen followed by DEAE chromatography along the lines originally described by Nussenzweig et al. (10). A complete description of the digestion and separation conditions appears in a previous article from this laboratory (11). Fibrin monomers were prepared by dispersing fibrin in 1 M NaBr/0.05 M Na acetate, pH 5.3 (12). In this regard, human fibrinogen (10-15 mg/ml) was clotted with bovine thrombin (Parke-Davis, 1 unit/ml final conc.); the final solution contained 150 mM NaCl, 1 mM EDTA, and 2.5 mM phosphate, pH 7.0. After the clot had been allowed to form at room temperature for approximately 30 min, it was wound out on a glass rod and dispersed in a minimal volume of the sodium bromide solution. The dispersion was continued with gentle stirring for 8 hr. at which point it was centrifuged to remove any undispersed material. The A 2o of the dispersed fibrin preparations (fibrin monomers) was generally about 17 (11 mg/ml). Reaggregation of the fibrin was effected by adding 25 Al of the fibrin monomer preparation to 0.5 ml of 80 mM phosphate buffer, pH 6.3, in a 1-cm path-length cuvette. The solution was mixed immediately and reaggregation was monitored by measuring the scattered light in a Zeiss spectrophotometer at 350 nm (13). Peptide additives were present in the phosphate buffer in advance of the addition of fibrin monomers. The
3086
Biochemistry: Laudano and Doolittle
additives included peptides we synthesized ourselves as well as some other commercially prepared peptides used as controls. Peptide Synthesis. Solid-phase peptide synthesis was conducted by established Merrifield procedures (14, 15), and a complete description, including materials, yields, etc., will appear elsewhere. In essence, however, the appropriate Bocamino acids (Boc-proline, Boc-nitroarginine, and Boc-sarcosine) were refluxed (in ethanol containing triethylamine) with chloromethylated polystyrene resin (Bio-Rad); the degree of substitution attained was generally about 0.4 mmol of amino acid per g of resin. The coupling of subsequent amino acids was achieved by the dicyclohexylcarbodiimide method (14, 15). Peptides were cleaved from the resin with HBr/trifluoroacetic acid for 90 min at room temperature (15). The nitro group of nitroarginine-containing peptides was removed by catalytic hydrogenation using 10% Pd/charcoal in acetic acid at a hydrogen pressure of 300 mm Hg for 48 hr. Arginine-containing peptides were purified by paper electrophoresis at pH 6.4. Peptides which did not contain arginine were purified by chromatography on Dowex 50W-X2. Amino acid analysis was conducted on a Spinco model 119 amino acid analyzer after total acid hydrolysis (5.7 M HCO, 108°, 24 hr in sealed tubes). Stock solutions of the various peptides were prepared in H20, and the concentrations were verified by amino acid analysis. The specific activities of radioactive peptides (containing [14C]glycine) were generally about 11 ;Ci/mmol peptide. Equilibrium Dialysis. Equilibrium dialysis was conducted in small test tubes using standard No. 8 Visking dialysis tubing. In the case of fibrinogen binding studies, 1.0 ml of fibrinogen solution (10 mg/ml) was dialyzed against 5 ml of a solution containing 0.2 M NaCl, 0.05 M phosphate buffer (pH 7.0), and various concentrations of the radioactive peptide under study. Binding studies with fragments D and E were performed similarly except that the protein concentration was reduced to 5 mg/ml. These latter studies were confined to the peptide glycyl-L-prolyl-L-arginyl-L-proline; in this case a bovine plasma albumin (10 mg/ml) control was also performed simultaneously. After 8 hr equilibration at room temperature, the dialysis bags were cut open and 0.5 ml aliquots of the inside and outside solutions were removed for scintillation counting. Final concentrations of the protein solutions were also checked at this point by measuring the A28o. Thrombin/Fibrinogen Clotting Assays. These tests were conducted by adding various concentrations of thrombin (Parke-Davis, bovine) to a human fibrinogen solution (1.5 mg/ml) containing various concentrations of peptide additive and then determining the clotting time. The final solution also contained 150 mM NaCl and 5 mM phosphate buffer, pH 7.2.
RESULTS Thrombin/Fibrinogen Clotting Assays. The peptides Gly-Pro-Arg-Sar, Gly-Pro-Arg-Pro, and Gly-Pro-Arg all exhibited significant inhibition of clotting in that they prevented or slowed the appearance of fibrin gels when added as a component of thrombin/fibrinogen mixtures (Table 1). Under the conditions used (fibrinogen = 1.5 mg/ml), the most effective peptides, Gly-Pro-Arg-Pro and Gly-Pro-Arg-Sar, completely prevented clotting at concentrations of about 10-3 M, and, even at 10-4 M, gel formation was significantly delayed. Gly-Pro-Arg exhibited significantly weaker inhibitory action. The inhibition of clotting is likely due in fact to an inhibition of fibrin polymerization.
Fibrin Monomer Reaggregation Studies. Gly-Pro-Arg-Sar
Proc. Natl. Acad. Sci. USA 75 (1978) Table 1. Inhibition and binding properties of synthetic peptides
Peptide Gly-Pro-Arg-Sar Gly-Pro-Arg-Pro Gly-His-Arg-Pro Gly-Pro-Ser-Pro Gly-Pro-Arg Pro-Arg Gly-Pro Gly
Thrombin/ Inhibition of Binding to fibrinogen reaggregation fibrinogen (equilibrium of fibrin clotting time,* dialysis)t monomerst min >200 >200 20 20 42 21 20 ND
++++ ++++
ND ++++
+ ++ -
++ ND
ND = not done; - = no inhibitory effect or no binding. * All peptides at final concentration of 1 mM; control = 20 min. t Various peptide concentrations were studied; see Fig. 1. Other ineffective peptides tested in this system included Gly-Leu-Tyr, Ala-Pro-Gly, tosyl-Arg-Sar-methylester (16), glycylglycine ethylester, and glycine ethylester. All peptides (except Pro-Arg) were labeled with [14C]glycine; see Fig. 2 and Table 2.
and Gly-Pro-Arg-Pro both proved to be powerful inhibitors of fibrin monomer reaggregation. Moreover, the effectiveness of the inhibition depended on both the concentration of the peptide inhibitor and the fibrin monomer concentration. In order to completely prevent polymerization, an approximately 100-fold molar excess of peptide to fibrin monomer was required, although significant retardation of polymerization occurred at ratios as low as 20:1 (these ratios allow for the fact that the fibrin "monomer" is actually dimeric). The peptide Gly-Pro-Arg proved to be less than half as effective as the tetrapeptides used (Fig. 1), and the dipeptides Gly-Pro and Pro-Arg had no detectable effect. Various unrelated peptides and amino acid derivatives (Table 1) and the tetrapeptides Gly-His-Arg-Pro and Gly-Pro-Ser-Pro also had no effect. Gly-His-Arg-Pro is the sequence found at the amino terminus of the fibrin 3-chain, and Gly-Pro-Ser-Pro was modeled on the amino acid replacement that has been reported for the variant human fibrinogen known as fibrinogen Detroit (8). As in the case of the defective fibrinogen, substitution of the arginine residue by serine completely obliterated its effectiveness. Equilibrium Dialysis Studies. Radioactive Gly-Pro-Arg-Pro was found to bind to human fibrinogen with an association constant of about 4 X 104 (Fig. 2). The number of sites at saturation extrapolated to two per 340,000 Mr. The peptide did not bind detectably to a preparation of bovine plasma albumin (10 mg/ml) run simultaneously. The tripeptide Gly-Pro-Arg bound less than half as tightly, but it also indicated the presence of two binding sites per fibrinogen molecule. Radioactive glycine, Gly-Pro, and Gly-Pro-Ser-Pro did not exhibit any binding under similar conditions (Table 1). Gly-His-Arg-Pro bound with an association constant of only 9 X 103 at pH 7.0, and the binding dropped off to less than half that at pH 6.5. In either case the data still indicated two binding sites. Gly-Pro-Arg-Pro bound to fragment D with an association constant of about 6 X 104. The number of sites in this case extrapolated to n = 1, when the observed Mr of 80,000 for our fragment D preparations (11) was used. Concentrations were specified by weighing dry fragment D (5.0 mg/ml) and verified throughout the experiment using an EM of 21 (17). The GlyPro-Arg-Pro preparation did not bind to fragment E.
Biochemistry: Laudano and Doolittle
Time, min
Proc. Nati. Acad. Sci. USA 75 (1978)
3087
Time, min
1.21
at 0.8
0.4
20 Time, min
40
Time, min
FIG. 1. Inhibition of the reaggregation of fibrin monomers by various synthetic peptides. All concentrations given are final concentrations. (Upper left) Increasing concentrations of Gly-Pro-Arg-Pro ranging from 0.1 mM to 0.33 mM (complete inhibition); kcontrol = A. (Upper right) Comparison of effectiveness of Gly-Pro-Arg (*, 0.16 mM) with Gly-Pro-Arg-Pro (0, 0.18 mM); control = A. (Lower left) Comparison of effectiveness of Gly-Pro-Ser-Pro (ED) with Gly-Pro-Arg-Pro (o). Both peptides at 0.27 mM; control = A. (Lower right) Comparison of effectiveness of Gly-His-Arg-Pro (X) with Gly-Pro-Arg-Sar (A). Both peptides at 0.24 mM; control = A.
DISCUSSION The tripeptidyl sequence Gly-Pro-Arg has endured at the amino terminus of fibrin a-chains for more than 400 million yr, being present in the fibrin of the most primitive vertebrates extant (19) as well as in mammalian fibrin (Table 2). This persistence is altogether consistent with a strategic structure-function interdependence. Moreover, replacement of the arginine residue with a serine at this position has been reported in the case of a defective human fibrinogen (8), the polymerization of which is impaired (22). Both of these observations-the one evolutionary and the other genetic-are consistent with early notions that the newly exposed amino terminus ought to be involved in some kind of intermolecular association leading to fibrin polymerization. In formulating the experiments described in this report, we visualized the unmasked sequence as a simple knob which could become involved in a knob-hole interaction. In undertaking the syntheses, we added a fourth residue, because in the natural situation the peptide chain continues on its course, and because the presence of a free carboxyl group might offset the effect of the positively charged arginine side chain. We could have esterified the carboxyl group, of course, but we
wanted something more stable chemically. Similarly, we could have incorporated the fourth residue of the natural sequence, which is a valine in the human molecule but varies in other species (Table 2). The fact that this particular Arg-Val bond has Table 2. Amino acid sequences at the amino termini of fibrin a- and ,-chains from various species Chain Ref. Sequence 1 2 3 4 5 6 7 Humana 18 Gly-Pro-Arg-Val -Val-Glu -Arg. . . Bovine a 19 Gly-Pro-Arg-Leu-Val -Glu-Lys... 20 Dog a Gly-Pro-Arg-Ile -Val -Glu-Arg ... Chicken a 21 Gly-Pro-Arg-Ile -Leu-Glu-Asn... 19 Lamprey a Gly-Pro-Arg-Leu-X* -Glx-Glx ... Human 18 Gly-His-Arg-Pro -Leu-Asp-Lys... Bovine 19 Gly-His-Arg-Pro-Tyr-Asx-Lys... Dog, 20 Gly-His-Arg-Pro-Leu-Asp-Lys... Chicken fi 21 Gly-His-Arg-Pro-Leu-Asp-Lys . . . 19 Lamprey Gly-Val-Arg-Pro--Leu-Pro -X* ... $ Probably Ser or Thr.
3088
Biochemistry: Laudano and Doolittle
Proc. Natl. Acad. Sci. USA 75 (1978)
A
A
X X
U
A
Cc
x A o
20
°0
A
1.0
2.0 R
FIG. 2. Plots of binding data from various equilibrium dialysis studies in which radioactive peptides Gly-Pro-Arg-Pro (A), Gly-Pro-Arg (x), and Gly-His-Arg-Pro (0) were exposed to fibrinogen (see also Table 1). R = moles of bound peptide/mole of fibrinogen; C = concentration of peptide. Fibrinogen concentration = 11 mg/ml (3.2 x 10-2mM).
been reported to be cleaved by thrombin (23) and by plasmin (24), however, convinced us to incorporate secondary amino acids (imino-) at this position in order to prevent undesirable proteolysis during our experiments. It also did not escape our attention that these compounds have potential therapeutic value as antipolymerants, and long-term stability and invulnerability to proteases would obviously be advantageous in such a setting. Previous reports from this laboratory have dealt with the effectiveness of the Arg-Sar bond in peptides designed as thrombin inhibitors (16). As a result, we realized that the synthetic tetrapeptides reported in this article could have antithrombin activity independent of the fibrin antipolymerant action. As a consequence we turned to fibrin monomer reaggregation studies in order to evaluate their effectiveness. Thus, Gly-Pro-Arg ought not to be a thrombin inhibitor per se, because it not only has an unprotected carboxyl group but is itself a product of thrombin cleavage. Ironically, Gly-Pro-Arg has been synthesized by others in the past and reported to be a thrombin inhibitor (25). The thrombin plasma assay used by those workers obviously precluded their discovering its actual mode of action, its "anti-thrombin" effect most likely being entirely due to the inhibition of fibrin polymerization. The observation that the peptides which inhibit fibrin polymerization bind to fragment D was consistent with the proposed scheme whereby a reciprocal association occurs between the central and terminal domains of neighboring molecules (7). It is also consistent with experiments in which fi-
brinogen and fragment D can be absorbed to immobilized fibrin (26). The fact that the tetrapeptide Gly-His-Arg-Pro, which corresponds to the terminal sequence exposed on the E-chain of fibrin, does not impair fibrin reaggregation is especially interesting, although the data have to be carefully considered. First, it must be remarked that technical difficulties make fibrin reaggregation studies difficult above pH6.5. Under the conditions of our study (pH 6.3), the histidine was presumably protonated. In line with this, the binding of Gly-His-Arg-Pro to fibrinogen was significantly less at pH 6.5 than at pH 7.0 or 7.3. On the other hand, Gly-His-Arg-Pro had no effect on the clotting time in the thrombin/fibrinogen assays conducted at pH 7.3, suggesting that the binding involves a set of sites that are different from those that are involved in initial polymerization. Although further binding and competition studies ought to resolve this point, the present observations are consistent with the well-known fact that fibrin polymerization in mammals can occur after release of the fibrinopeptide A only (27, 28). The strength of the association of the tetrapeptides GlyPro-Arg-Sar and Gly-Pro-Arg-Pro with fibrinogen is sufficient to ensure that affinity labeling studies (29) using similar compounds with functional reactive groups attached are feasible. It will be of great'interest to learn which portions of the individual chains are involved in the actual sites. Also of interest will be studies with lamprey fibrinogen because that material can be clotted by the exclusive removal of the fibrinopeptide B (19, 30). Interestingly enough, its (3-chain tetrapeptide varies from
Biochemistry: Laudano and Doolittle the mammalian sequences by a valine/histidine replacement (Table 2). We thank Marcia Riley for technical assistance in preparing fibrinogen and fragments D and E, Dennis Trovato for the operation of the amino acid analyzer, and Cheryl Yokahama for assistance with some of the equilibrium dialysis studies. These studies were supported by U.S. Public Health Service Grants HL-18,576 and GM-17,702. 1. Bailey, K., Bettelheim, F. R., Lorand, L. & Middlebrook, W. R., (1951) Nature 167,233-234. 2. Bettelheim, F. R. & Bailey, K. (1952) Biochim, Biophys. Acta 9, 578-579. 3. Ferry, J. D., Katz, S. & Tinoco, I., Jr. (1954) J. Polym. Sci. 12, 509-516. 4. Marder, V. (1971) Scand. J. Haematol. Supp, 13,21-36. 5. Blombick, B. & Blombick, M. (1972) Ann. N.Y. Acad. Scd. 202, 77-97. 6. Hall, C. E. & Slayter, H. D. (1959) J. Biophys. Biochem. Cytol. 5, 11-16. 7. Doolittle, R. F. (1973) Adv. Prot. Chem. 25,1-109. 8. Blombeck, M., Blombick, B., Mammen, E. F. & Prasad, A. S. (1968) Nature 218, 134-137. 9. Doolittle, R. F., Schubert, D. & Schwartz, S. A. (1967) Arch. Blochem. Blophys., 118,456-467. 10. Nussenzweig, V., Seligman, M., Pelmont, J. & Grabar, P. (1961) Ann. Inst. Pasteur, Paris 100, 377-389. 11. Doolittle, R. F., Cassman, K. G., Cottrell, B. A., Friezner, S. J. & Takagi, T. (1977) Biochemistry 16, 1710-1715. 12. Donnelly, T. H., Laskowski, M., Jr., Notley, N. & Scheraga, H. A. (1955) Arch. Biochem. Biophys. 56,369-387. 13. Latallo, Z., Fletcher, A. P., Alkljaersig, N. & Sherry, S. (1962) Am. J. Physiol. 202,675-680. 14. Merrifield, R. B. (1964) J. Am. Chem. Soc. 85,2149-2154.
Proc. Natl. Acad. Sci. USA 75 (1978)
3089
15. Stewart, J. M. & Young, J. D. (1969) Solid Phase Peptide Syntheis, (Freeman, San Francisco, CA). 16. Weinstein, M. J. & Doolittle, R. F. (1972) Biochim. Biophys. Acta 258,577-590. 17. Marder, V. J., Shulman, N. R. & Carroll, W. R. (1969) J. Biol. Chem. 244,2111-2119. 18. Iwanaga, S., Wallen, P., Grondahl, N. J., Henschen, A. & Blombick, B. (1967) Biochim. Biophys. Acta 147,606-609. 19. Cottrell, B. A. & Doolittle, R. F. (1976) Biochim. Biophys. Acta 453,426438. 20. Birken, S., Wilner, G. D. & Canfield, R. E. (1975) Thromb. Res. 7,599-610. 21. Murano, G., Walz, D., Williams, L., Pindyek, J. & Mosesson, M. W. (1977) Thromb. Res. 11, 1-10. 22. Mammen, E. F., Prasad, A. S., Barnhart, M. L. & Au, C. C. (1969) J. Clin. Invest. 48,235-249. 23. Blombdck, B., Blombeck, M., Henschen, A., Hessel, B., Iwanaga, S. & Woods, K. R. (1968) Nature 218,130-134. 24. Takagi, T. & Doolittle, R. F. (1975) Biochemistry 14, 940946. 25. Dorman, L. E., Cheng, R. C. & Marshall, F. N. (1972) Chemistry and Biology of Peptides (Ann Arbor Science Publishers, Ann Arbor, MI), pp. 455-459. 26. Kudryk, B. J., Collen, D., Woods, K. R. & Blomback, B. (1974) J. Biol. Chem. 249,3322-325. 27. Blomback, B., Blomback, M. & Nilsson, I. M. (1957) Thromb. Diath. Haemorrh. 15,76-86. 28. Doolittle, R. F., Oncley, J. L. & Surgenor, D. M. (1962) J. Biol. Chem. 237,3123-3127. 29. Wofsy, L., Metzger, H. & Singer, S. J. (1962) Biochemistry 1, 1031-1039. 30. Doolittle, R. F. (1965) Biochem. J., 94,735-741.
