134

Biochimica et Biophysica Acta, 1040 (1990) 134-136 Elsevier

BBAPRO 30270

B B A Report

Binding of the bovine basic pancreatic trypsin imhibitor (Kunitz) to human Glu 1-, Lys 77-, Va1442- and Va1561-plasmin • a comparative study P a o l o A s c e n z i 1, G i n o A m i c o n i 1, M a r t i n o B o l o g n e s i 2, E n e a M e n e g a t t i 3 and Mario Guarneri 3 I C.N.R., Center for Molecular Biology, Department of Biochemical Sciences, Unioersity of Rome "La Sapienza" Rome, 2 Department of Genetics and Microbiology, Section of Crystallography, University of Pavia, Pavia and 3 Department of Pharmaceutical Sciences, Unioersity of Ferrara, Ferrara (Italy) (Received 10 November 1989) (Revised manuscript received 23 April 1990)

Key words: Basic pancreatic trypsin inhibitor; Kunitz inhibitor; Proteinase : inhibitor complex formation; Thermodynamics; Kinetics; (Human plasmin); (Bovine trypsin inhibitor)

Thermodynamic and kinetic parameters for the binding of the bovine basic bancreatic trypsin inhibitor ( B ~ , Kunitz inhibitor) to haman Glut-, Lys77-, Vai 44z- and ValS6t-plasmin (EC 3.4.21.7) have been d e t e ~ n e d between pH 3.0 and 9.5, and from 5.0 to 45.0° C. The inhibitor-binding properties to human Glut-, Lys77-, Va144z- and ValS6t-plmmin suggest a possible role of BPTI in modulating plasmin activity when the inhibitor is used therapeutically.

Plasmin (EC 3.4.21.7) is a serine proteinase that catalyzes mainly the hydrolysis of fibrinogen and fibrin deposits, participating in clot lysis; the enzyme activity has been shown to be modulated by plasma protein inhibitors such as the a2-plasmin inhibitor and a 2macroglobulin (see Refs. 1-5 for review). Considering the presence in human serum of low molecular weight Kunitz-type proteinase inhibitors showing a high degree of homology with the bovine basic pancreatic trypsin inhibitor (BPTI, Kunitz inhibitor) [6], it appeared of interest to investigate, from the thermodynamic and kinetic viewpoints, the inhibition by BPTI of the enzymatic activity of human Glu 1-, Lys77-, Val 442- and ValS61-plasmin. These four plasmin variants are formed during the activation of plasminogen in vivo or generated when the enzyme is prepared in vitro [2]. Possibly, only human Va1561-plasmin (corresponding to the light (B) chain of the three larger human plasmins) has no relevance to the physiological situation, its importance being mainly related to basic and biotechnological research [2]. Thus, the amidase and esterase catalytic

Abbreviations: BPTI, bovine basic pancreatic trypsin inhibitor (Kunitz inhibitor); ZLysONp, N-a-carbobenzoxy,L-lysine p-nitrophenyl ester. Correspondence: P. Ascenzi, C.N.R., Center for Molecular Biology, Department of Biochemical Sciences, University of Rome 'La Sapienza', Piazzale Aldo Moro 5, 00185 Rome, Italy.

parameters of human ValS61-plasmin are less favourable than those of the other three variants, but it can react with the so-called heavy (A) chain to form recombinant human Lys77-plasmin, which resembles the parent enzyme in both physical properties and catalytic functions [2]. The BPTI-binding properties to human Glu 1-, Lys77-, Val 442- and Va1561-plasmin (i) suggest that the BPTI-homologous Kunitz-type proteinase inhibitors present in human plasma are unlikely to be relevant in the fine control of plasmin activity in vivo; and (ii) indicate a possible role of BPTI in modulating the proteinase action when the inhibitor is used therapeutically. Human Glu 1-, Lys77- and Val'~2-plasmin were prepared from GluL, Lys77- and Va1442-plasminogen, respectively [2]. Va1561-plasmin was prepared from tys 77plasmin [2]. BPTI was kindly provided by I~petit (Milano, Italy). ZLysONp was obtained from Sigma (St. Louis, MO, U.S.A.). All other products were from Merck (Darmstadt, F.R.G.). The characterization of human Glu 1-, t y s 77-, MaX442- and Va1561-plasmin, BPTI and ZLysONp was reported previously [2,7,8]. Values of thermodynamic parameters for binding of BPTI to human GluL, Val 4a2- and Va156Lplasmin were determined, between pH 3.0 and 9.5 ( I = 0.1 M), and from 5.0 to 45.0 o C, from the evaluation of the inhibitory effect of BPTI on the enzymatic hydrolysis of ZLysONp, by titration of the proteinase with the inhibitor and measuring the residual activity [9,10]. Values of

0167-4838/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

135 the second-order rate constant (kon) for BPTI binding to the four variants of human plasmin (i.e., for proteinase inactivation) were determined, at pH 8.0 (1 = 0.1 M) and 21.0" C, from the decrease of the rate of the enzymatic hydrolysis of ZLysONp over time by mixing the inhibitor with the proteinase and the substrate [11]. Similar to the human Lys77-plasmin/BPTI system [10], inhibitor binding to human Glul-, Va1442- and Va1561-plasmin conforms to simple equilibria as indicated by the Hill coefficient (n) of 1.00 ± 0.02. Values of the association equilibrium constant (K a) for the proteinase: BPTI complex formation were independent of the enzyme concentration. Furthermore, the timecourse of adduct formation demonstrates over 95% correspondence to a pseudo-first-order reaction. The pH titration curves of K a for BPTI binding to human GluL, Val a42- and Va1561-plasmin are superimposable on that obtained for the human Lys77plasmin : BPTI complex formation [10] (see Fig. 1), and thus may be described in identical terms. On increasing the pH from 3.0 to 5.0, the enhancement of BPTI affinity for the four variants of human plasmin (i.e., of Ka values) reflects the acidic midpoint perturbation of a three proton transition. One of the residues likely to affect the proteinase : BPTI complex formation is Asp735. This amino acid side-chain, present in the primary specificity subsite of trypsin-like serine proteinases acting on cationic substrates, modulates their spectral and catalytic as well as ligand and inhibitor binding properties (see Refs. 12 and 13 for review). Also the ionization of Asp645, involved in the proteinase catalytic triad, could affect inhibitor binding. Similar to related serine (pro)enzyme/inhibitor systems [12], no plausible assignment is possible for the third acidic ligand-linked ionizable group which could reflect pH-dependent conformational changes well outside the proteinase/BPTI contact region. On increasing the pH from 5.0 to 9.5, the enhancement of inhibitor affinity for human GIuL, L y s 77-, Va1442- and ValS61-plasmin (i.e., of K a values) reflects the acidic pK shift of a single proton transition. This pK shift might be interpreted as a strengthening of the proteinase hydrogen bond occurring between the oxygen atom, at position G, of Ser 741 and the nitrogen atom, at position E2, of H i s 602. This hydrogen bond is very weak or absent in the inhibitor-free enzyme. The observed different BPTI-binding properties to the four variants of human plasmin (expressed by values of Ka and ko. given in Table I) agree with the proteinase specificity for synthetic substrates [2]. Such behaviour (i) reflects the stabilization state of the human G l u L , L y s 77- and Vala42-plasmin light (B) chain active center by the heavy (A) chain (the effect being mainly exerted by the Ala'~a-Arg 56° domain) [1,2]; and (ii) may indicate that the human ValS61-plasmin specificity subsites, where BPTI binds, are less structured and/or exposed than those of the human Glul-, Lys77- and

10

,

,

9 =!

8

7 t~

v

oo~

6 5 4 i

2

i

3

i

4

|

5

i

i

6

,

,

i

i

7

8

i

i

9

i

10

pH Fig. 1. pH dependence of the association equilibrium constant (Ka) for BPTI binding to human Glu 1- (D), Lys 77- ( 0 ; the data were obtained from Menegatti et al. [10]), Va1442- (zx) and Vai s61- (O) plasmin at 21.0 o C. An average error value of + 8% was evaluated for K a values as the standard deviation. According to Menegatti et al. [10], the unbroken lines were calculated from the following equation (1):

log K a = C - log

([H ÷ ] + r6,~)" ([n+]+

,,

K 'L I G 3" K "L I G • - I o g ~ • xUN

L

~xUN

( [ n ÷ ] + r,~'~,.)

, , .([n + ]+r,.,o) KL,o) (1) L

where C is a constant that c o r r ~ p o n d s to the alkaline asymptote of log K a, and pKt~NL, pK~'NL, pK~l o and PKt'~G are the p K values of the proton dissociation equilibrium constants for the BPTI-free (KoNL) and BPTI-bound (KLIo) proteinase, respectively. Unbroken lines, calculated with the following sets of parameters: human Glu x-, Lys 7T- and V a l ~ 2 - p l a s m i n - C = 9 . 0 8 , p K t ~ s L = 4 . 5 , p K ~ l G = 3 . 6 , pKt~'NL = 7.0 and p K ~ G = 5.1; and human ValS61-plasmin-C = 8.30, p t re tp PKuN L = 4.4, P K u G = 3.5, pKuN L = 6.9 and pKLi G = 5.0, were obtained with an iterative nonlinear least-squares curve fitting program, which also allowed us to ascribe an average error value of + 125[ to K~JNL, K~I O, K~tNL a n d K ~ O values, as the s t a n d a r d d e v i a t i o n [10]. The pH profile was explored using the following buffers: phosphate (pH 3.0 to 3.5), acetate (pH 3.5 to 6.0), phosphate (pH 6.0 to 8.5) and borate/glycine (pH 8.5 to 9.5); all at I = 0.1 M (sodium salts) [10]. According to Menegatti et al. [10], no specific ion effects were found using different buffers with overlapping p H values. For all the other details, see text.

Va1442-plasmin variants, which turn out to be, in all probability, almost similar. On the other hand, the affinity of the a2-plasmin inhibitor for human GIuL, Lys77- and Va1442-plasmin is very different, and the complex formation proceeds at diverse rates. This finding reflects the involvement of the heavy (A) chain of the proteinase in the binding of the a2-plasmin inhibitor [2,4]. As a whole, the reported results (i) suggest that the human plasmin BPTI binding clefts are localized at the central site of the proteinase; and (ii) indicate that the heavy (A) chain of the enzyme does not participate in BPTI binding.

136 TABLE I Values of thermodynamic and kinetic parameters for B P T I binding to human Glu ~-, Lys 77-, Va1442- and Val TM -plasmin pH 8.0, phosphate buffer, I = 0.1 M. For all the other details, see text. Proteinase

Ka (M-l) ~

AG o (kcal/mol) a

AH o (kcal/mol) a

AS o (entropy units) a

kon (M-1. s - I ) a

Human Human Human Human

9.8.108 1.2-10 9 b 1.3-10 9 1.6.108

- 12.2 -- 12.2 b - 12.2 - 11.0

+ 2.3 + 2.3 b + 2.2 + 2.8

+ 49 -t-49 b + 49 + 47

1.6. l0 s 1.8" 10 s 2.0. l 0 s 1.1. l 0 s

Glul-plasmin Lys77-plasmin Val'~2-plasmin ValS61-plasmin

a Values of Ka, AG °, A S ° and kon have been obtained at 21.0°C; values of A H ° were determined from the linear dependence of log K a on I / T , the temperature ranging between 5.0 and 45.0 ° C. Average error values of + 8% (for K a and kon values) and + 12% (for AG °, A H o and A S o values) were found as the standard deviation. b Values of thermodynamic parameters for BPTI binding to human Lys77-plasmin were obtained from Menegatti et al. [10].

Values of A H o and A S o given in Table I indicate that the human Glu 1-, Lys77-, Val '~2- and ValStLplas min:BPTI complex formation is mainly an entropydriven process. Moreover, the positive AS o values could reflect the removal of the proteinase- and/or inhibitorbound water molecules during the complexation (see Refs. 10, 12, 14, 15 for review). BPTI-homologous Kunitz-type proteinase inhibitors do not appear to be physiological modulators of human plasmin activity in either health or disease [4,6] because of their low concentration ([•] = 0.015 #M (see Ref. 6)), even though in principle it could be. In fact, (i) the high association equilibrium constants (i.e., K a values; see Fig. 1 and Table I) make the human plasmin:BPTI complexes practically irreversible, and (ii) the half-time of enzyme inactivation ( t l / 2 = 0.693/(kon • [I]) = 240 s) is high enough to enable human plasmin to react with fibrin(ogen), and low enough to enable BPTI-homologous Kunitz-type proteinase inhibitors to inactivate human plasmin within a reasonable period of time after it has accomplished its physiological degradative function (thus preventing proteolytic attack at undesirable substrates). On the other hand, in the case of the therapeutic use of BPTI (see Ref. 4), the administration of 500 000 KIU (kallikrein inhibitory units) of the drug (corresponding to about 2.3 /tM inhibitor plasma concentration) decreases the half-time of human plasmin inactivation by 100-times (t~/2 = 1.4 s), so that BPTI may be used to quickly control fibrinolysis. Thus, upon complete activation of human plasminogen (plasma concentration about 1/~M (see Ref. 4)), human plasmin is effectively scavenged by the ae-plasmin inhibitor (with t l / 2 -~ 0,018 s), and secondly by a2-macroglobulin (with i l l 2 = 3.0 s) and extrinsic BPTI (with t l / e = 1.4 s) (see Refs. 4, 6). In conclusion, the BPTI-homologous Kunitz-type proteinase inhibitors, present in human plasma [6], are unlikely to play any role in the fine control of plasmin

activity in vivo. However, since BPTI can act as back-up irdaibitors at therapeutic concentrations, it may be used in circumstances where protection is needed against the deleterious effects of rapid and complete activation of plasminogen. References 1 Castellino, F.J. and Powell, J.R. (1981) Methods Enzymol. 80, 365-378. 2 Robbins, K.C., Summaria, L. and Wohl, R.C. (1981) Methods Enzymol. 80, 379-387. 3 H~rn, H. and Heidland, A., eds. (1982) Proteases: Potential Role in Health and Diseases, Plenum Press, New York. 4 Barrett, A.J. and Salvesen, G., eds. (1986) Proteinase Inhibitors, Elsevier, Amsterdam. 5 Cunningham, D.D. and Long, G.L., eds. (1987) Proteinases in Biological Control and Biotechnology, Alan R. Liss, New York. 6 Fioretti, E., Angeletti, M., Citro, G., Barra, D. and Ascoli, F. (1987) J. Biol. Chem. 262, 3586-3589. 7 Kassel, B. (1970) Methods Enzymol. 19, 844-852. 8 Ascenzi, P., Sleiter, G. and Antonini, E. (1982) Gazz. China. Ital. 112, 307-317. 9 Fritz, H., Schult, H., Meister, R. and Werle, E. (1969) HoppeSeyler's Z. Physiol. Chem. 350, 1531-1540. 10 Menegatti, E., Guarned, M., Bolognesi, M., Ascenzi, P. and Amiconi, G. (1986) J. Mol. Biol. 191, 295-297. 11 Vincent, J.-P. and Lazdunski, M. (1972) Biochemistry 11, 29672977. 12 Amiconi, G., Ascenzi, P., Bolognesi, M., Menegatti, E. and Guarneri, M. (1988) in Macromolecular Biorecognition: Principles and Methods (Chaiken, I., Chiancone, E., Fontana, A. and Neff, P., eds.), pp. 117-130, The Humana Press, Clifton. 13 Me~aegatti, E., Guarneri, M., Bolognesi, M., Ascenzi, P. and Amiconi, G. (1988) in Macromolecular Biorecognition: Principles and Methods (Chaiken, I., Chiancone, E., Fontana, A. and Neff, P., eds.), pp. 101-115, The Humana Press, Clifton. 14 Amiconi, G., Ascenzi, P., Bolognesi, M., Guarneff, M. and Menegatti, E. (1987) Adv. Biosciences 65, 177-180. 15 Bolognesi, M., Ascenzi, P., Amiconi, G., Menegatti, E. and Guarneri, M. (1988) in Macromolecular Biorecognition: Principles and Methods (Chaiken, I., Chiancone, E., Fontana, A. and Neff, P., eds.), pp. 81-100, The Humana Press, Clifton.