The degradation of bovine and human prothrombin by human polymorphonuclear leukocyte cathepsin g

THROMBOSIS RESEARCH 44; 339-345, 1986 0049-3848/86 $3.00 t .OO Printed in the USA. Copyright (c) 1986 Pergamon Journals Ltd. All rights reserved.

THE

DEGRADATION OF BOVINE POLYMORPHONUCLEAR

Philip

T. Turkington,

Nigel

AND HUMAN LEUKOCYTE L. Blumsom

PROTHROMBIN BY HUMAN CATHEPSIN G and

Donald

T. Elmore

Departments of Biochemistry and Haematoloqy, Queen's University of Belfast, Belfast BT9 7BL, N. Ireland, UK (Submitted to Editor S. Maqnusson April 1985; Accepted by Editors-in-Chief B. Blomb5ck and A.L. Copley 25.7.1986)*

ABSTRACT Cathepsin G, isolated from human polymorphonuclear leukocytes, was found to effect rapid and specific degradation and biological inactivation of bovine and human prothrombin in the absence of calcium ions with the formation of two peptide fragments from the N-terminal end of the molecule. Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate indicated that the molecular weights of the fragments were 5,000. and 17,500. Proteolysis of prothrombin by cathepsin G was inhibited by calcium ions. Leukocyte proteinases such as cathepsin G may be responsible for haemorrhagic disorders associated with myelocytic leukaemia and septicaemia.

INTRODUCTION We have shown that prothrombin and factor X are rapidly degraded and biologically inactivated by rat mast cell proteinase (RMCPI) (l).. In each case, peptide fragments containing all the residues of Y-carboxyglutamic acid were cleaved from the N-terminal regions of prothrombin and factor X. We now report that cathepsin G from human polymorphonuclear leukocytes also degrades prothrombin rapidly, but the proteolysis is more extensive than with mast cell proteinase.

Key

words:

Prothrombin,

y The acceptance of this exceptional and caused whom it was originally

cathepsin

G, proteolysis

communication by the utter submitted. 339

by the Editors-in-Chief is negligence of the Editor, to

340

PROTHROMBIN AND CATHEPSIN G

MATERIALS

Vol. 44, No. 3

AND METHODS

Reagents were purchased as follows: rabbit brain cephalin, Russell's viper venom, bovine plasma deficient in factor V and bovine plasma deficient in factors VII and X (Diagnostic Reagents Ltd, Oxford, UK), 'Trasylol' (Bayer, West Sussex, UK), human fibrinogen (grade L) (Kabi A.B., Sweden), bovine thrombin (Parke-Davis, Detroit, Mich., USA), turkey ovomucoid, Echis carinatus venom and a-chymotrypsin (Sigma London Chemical Co., Ltd, Poole, UK), aminopeptidase (Miles Laboratories, Ltd, Slough, UK). Isolation and purification of cathepsin G. Cathepsin G was prepared from human polymorphonuclear leukocytes by the method of Baugh and Travis (2) with an additional purification step. Following elution of cathepsin G from the CM-cellulose column, the pH of the eluate was adjusted to 8.0 with 1.0 M-tris. The solution was passed under gravity feed down a column (1.2xlOcm) of Sepharose 4B-turkey ovomucoid conjugate pre-equilibrated in 0.02 M Tris/ chloride buffer, pH 7.0, containing 0.8M NaCl and 0.02 M CaC12. The eluate was concentrated to 2 ml by ultrafiltration and stored in 200 ~1 aliquots. Assay of cathepsin G. Steady-state analyses were carried out with N-methoxycarbonyl-L-phenylalanine 4-methylumbelliferyl ester. To 0.02 M-sodium phosphate buffer, pH 7.0, which was 0.1 M with respect to NaCl (3.0 ml) at 25OC, a 0.2 mM solution of the substrate in t-butanol (20 ~1) was added. Liberation of 4-methylumbelliferone was followed spectronm, hem=445 nm). The specific activity of fluorimetrically (Xex=365 cathepsin G preparations averaged 2,500 u/mg using the foregoing assay. The operational molarity of s+utions of cathepsin G was determined using 4-methylumbelliferyl (4 -NNN-trimethylammonium)cinnamate as spectrofluorimetric titrant (3). Preparation of prothrombin. Bovine prothrombin was prepared from titrated plasma by the method of Bajaj and Mann (4). The product was homogeneous on sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and had an apparent molecular weight of 69,500. It had a specific activity of 1,000 u/mg when assayed by the methodoof Ware and Seegers (5). It was stored in 50% aqueous glycerol at -20 C where it was stable for at least a year. Human prothrombin was Contaminating factor X was column of rabbit antibodies The overall yield was 20%; had a specific activity of

also prepared by the method of Bajaj and Mann (4). removed by immunoaffinity chromatography on a to bovine factor X immobilised on Sepharose 4B. the prothrombin was electrophoretically pure and 922 u/mg.

In a typical experiment, Proteolysis of prothrombin by cathepsin G. prothrombin (1 umol) in 0.1 M Tris-chloride buffer (30 ml), pH 7.4, which was 20 mM with respect to NaCl, ;as digested with a solution (200 ~1) of cathepsin G (126 pmol) at 37 C. Aliquots were removed at intervals during 45 min for clotting assays (5,6) and electrophoretic analysis. The residue after 45 min was subjected to ion-exchange chromatography. Other experiments were carried out with enzyme: substrate ratios of 1:lOO to 1:10000.

Vol. 44, No. 3

PROTHROMBIN AND CATHEPSIN G

Polyacrylamide gel electrophoresis in the Other analytical techniques. absence and presence of sodium dodecyl sulphate, amino-acid analysis and N-terminal residue analysis were carried out essentially as before (1). N--Terminal sequences were also identified using immobilised aminopeptidase

341

M.

RESULTS AND DISCUSSION Digestion of bovine prothrombin with cathepsin G at 37OC caused rapid loss of coagulant activity (Fig. 1) when aliquots were assayed by the two-stage method of Ware and Seegers (5). Similaroresults were obtained with bovine prothrombin and humanoprothrombin at 28 C, although the rate of inactivation was lower than at 37 C. The rate of inactivation was also strongly dependent on the enzyme: substrate ratio (Fig. 1). Although the two-stage assay of prothrombin showed that almost complete inactivation occurred, when the first stage was prolonged from 6 min. to 30 min., about 25% of the potential thrombogenic activity was still detected. Additionally, when a cathepsin G digest of prothrombin was treated with the procoagulant protein from Echis carinatus venom, the full potential thrombin activity was liberated (Fig. 2). The presence of Ca2+ ions (4-10 mM) afforded almost complete protection to the action of cathepsin G. The reported failure of a chymotrypsin-like proteinase, presumably cathepsin G, from human granulocytes to inactive prothrombin (7) may have been due to the inclusion of Ca2' ions in the incubation mixture although this is not clear from the published experimental protocol. A separate control experiment showed that Ca2+ ions had no appreciable effect on the rate of hydrolysis of N-methoxycarbonyl-L-phenylalanine 4-methylumbelliferyl ester by cathepsin G. Bovine thrombin was unaffected by cathepsin G. The results of the foregoing experiments strongly resemble those reported when prothrombin was exposed to RMCP I (1) and indicate that proteolysis is limited to the N-terminal region not including the thrombin moiety. 100

Fig. 1 The effect of cathepsin G on the clotting activity of bovine prothrombin as assayed by the two-stage method of Ware and Seegers (5). Hydrolysis of prothrombin by cathepsin G was carried out in 0.05 M tris chloride buffer, pR 7.4, containing 0.1 M NaCl at 37'C. Enzyme: substrate ratios were l:lOO(Ol:lOOO(~ -0) and l:lOOOO(*-). 0) 9 1:500(_),

342

PROTHROMBIN AND CATHEPSIN G

Vol. 44, No. 3

8 4

8

12

16

20

24

28

TIME (MIN)

Fig. 2 The effect of cathepsin G on the activity of bovine prothrombin as assayed by the conventional two-stage method (5) (o----O) and as assayed The by the procoagulant fraction of Echis carinatus venom (6) (A--d). digestion of prothrombin vith cathepsin G was carried out in 0.05 M tris chloride buffer, pH 7.4, containing 0.1 M NaCl at 28'C with enzyme: substrate ratio of 1:500. Analysis of cathepsin G digests of prothrombin by SDS-PAGE revealed the presence of two products, '1' and Pa', (1, 65000 and 52000 respectively). In contrast, PAGE revealed the presence of four products. Two bands, later shown to be Pl' and Pa', had lower anodic mobilities than prothrombin; the other two bands, Pl" and P2", had anodic mobilities greater than that of prothrombin. The failure zo detect the small anionic peptides by SDS-PAGE is in agreement with the results obtained with digests of prothrombin with RMCP I. These results indicate that cathepsin G cleaves prothrombin at two alternative sites and the smaller highly anionic fragments originate from the N-terminal region of the prothrombin molecule.

A partial separation of the fragments formed by the action of cathepsin G on prothrombin was achieved by ion-exchange chromatography on DEAEcellulose at pH 8.4 using a linear salt gradient (Fig. 3). Prothrombin assays using the procoagulant fraction of Echis carinatus venom and analysis by PAGE revealed that the first peak eluted was Pl', the second peak was P2' and the third peak was a mixture of Pl" and Pa". Attempts to separate the latter were unsuccessful and identification of the sites of cleavage in the prothrombin molecule by cathepsin G was achieved by determination of the N-terminal sequence of Pl' and P2' by dansylation followed by acid hydrolysis, using 3 M toluene-p-sulphonic acid containing 0.2% (w/v) 3-(2-aminoethyl)indole in the case of Trp (8), and by amino-acid analysis of digests

Vol. 44, No. 3

PROTHROMBIN AND CATHEPSIN G

343

-0.6

*0.4 z ij : -0.2

A0

* f 20

_--I_

~._

.

.

--

40 FRACTION

I___

60 NUUBER

60

100

@ML)

Fig. 3 Separation of the fragments produced by the action of cathepsin G on bovine prothrombin (enzyme:substrate = 1:8000) in 0.1 M tris chloride buffer, pH 7.4, containing 20 mM NaCl at 37OC for 45 min. The digest was applied to a column (2.2 x 20 cm) of DEAE-cellulose (DE-50) in 0.1 M tris chloride buffer, pH 8.4, at room temperature and the optical density was monitored at 2171~1 (A-4) and 28Onm (0-e). Fractions of 5 ml were collected. Peaks I (Pl') and II (Pa') were separately pooled and concentrated by ultrafiltration.

obtained by the action of aminopeptidase identified are as follows: Bovine Pl' Bovine Pp' Human Pl' Human P2'

M.

The N-terminal

sequences

Trp-Ala....... Leu-Glu-Thr... Trp-Ala....... Glu-Gln.......

We conclude that the bonds cleaved are Phe41-Trp42 (bovine and human), Leu167-Leu168 (bovine), Leu168-G1u16g (human) based on the primary structures of bovine and human prothrombin (9-11).

344

PROTHROMBIN AND CATHEPSIN G

Vol. 44, No. 3

Comparison of the results reported herein with those obtained with RMCP I shows that both enzymes cleave the Phe41-Trp42 bond releasing a peptide, Pl", containing all the residues of y-carboxyglutamic acid. In contrast to the action of RMCP I, however, cathepsin G alternatively generates a rather larger moiety, Pa", comprising all the sequence contained in fragment 1 together with the N-terminal sequence of fragment 2 of prothrombin. This fragment is derived by cleavage of a bond lying between kringles 1 and 2. Failure to detect fragments comprising residues 42-167 (bovine) or 42-168 (human) indicates that cathepsin G does not cleave both bonds in the same molecule. This suggests that cleavage of one bond leads to a conformational change that causes the other to be partially or completely buried. The differences in behaviour of RMCP I and cathepsin G towards prothrombin probably reflect differences in subsite specificity of the two proteinases. We consider that the in vitro experiments reported here may have an in vivo parallel in pathological conditions such as myelocytic leukaemia and septicaemia when the abnormally high population of leukocytes could release amounts of cathepsin G greater than those which could be controlled by Consequent degradation of prothrombin as circulating proteinase inhibitors. described above would lead to the well established clinical manifestation of haemorrhagic complications.

REFERENCES 1. WYLIE, A.R.G., LONSDALE-ECCLES, J.D., BLUMSOM, N.L. and ELMORE, D.T. Proteolysis of bovine and human prothrombin and of bovine factor X by rat mast cell proteinase, (preceding paper) 2. BAUGH, R. and TRAVIS, J. Human leukocyte granule elastases: Biochemistry, 15, 836-841, 1976. isolation and characterisation.

rapid

JAMESON, G.W., ROBERTS, D.V., ADAMS, R.W., KYLE, W.S.A. and ELMORE, D.T. 3. Determination of the operational molarity of solutions of bovine a-chymotrypsin, trypsin, thrombin and factor Xa by spectrofluorimetric titration. Biochem. J., 131, 107-117, 1973. BAJAJ, S.P. and MANN, K.G. The simultaneous purification of bovine 4. Biol. Chem., 248, 7729-7741, 1973. prothrombin and factor X. 2. -WARE, A.G. and SEEGERS, W.H. Two-stage 5. determination of prothrombin concentration. 471-482, 1949.

procedure for the quantitative Amer. J.. Clin. Pathol., g, ---

DENSON, D.W. Clot-inducing substances 6. Res reference to Echis carinatus. -Thromb. -.J

in snake venom with particular 8 351-360, 1976. _'

7. SCHMIDT, W., EGBRING, R. and HAVEMANN, K. Effect of elastase-like and chymotrypsin-like neutral proteases from human granulocytes on isolated clotti_ng factors. ~Thromb. L-J Res 6 315-326, 1975. a. LIU, T.Y. and CHANG, Y.H. Hydrolysis of proteins with p-toluenesulphonic acid. -J. Biol. e., 246, 2842-2848. 9. MAGNUSSON, S., SOTTRUP-JENSEN, L. and CLAEYS, H. Complete primary structure of prothrombin: Isolation, structure and reactivity of ten carboxylated glutamic acid residues and regulation of prothrombin activation by thrombin. In: Proteases and Biological Control, eds. Reich, E., Rifkin, D.B. and Shaw, E., pp. 123-149, 1975.

Vo'l.44, No. 3

PROTHROMBIN AND CATHEPSIN G

345

10. WALZ, D.A., HEWETT-EMMETT, D. and SEEGERS, W.H. Amino acid sequence of human prothrombin fragments 1 and 2. --Proc. Nat. Acad. E., U.S.A., 74_, 1969-1972, 1977. 11. BUTKOWSKI, R.J., ELION, J., WWNING, M.R. and MANN, K.G. Primary structure of human prethrombin 2 and a-thrombin. 2. ~Biol. Chem., 252, 4942-4957, 1977.