N-peptidyl, O-acyl hydroxamates: comparison of the selective inhibition of serine and cysteine proteinases

BiBBochi~ic~a

ELSEVIER

et Biophysica AEta Biochimica et Biophysica Acta 1295 (1996) 179-186

N-peptidyl, O-acyl hydroxamates: comparison of the selective inhibition of serine and cysteine proteinases Hans-Ulrich Demuth a,*, Angelika Schierhorn b Philip Bryan c Ralph H~Sfke d Heidrun Kirschke d Dieter Br~Smme d a Department of Drug Biochemistry, Hans-Knoell-lnstitute of Natural Product Research Jena, Martin-Luther-University of Halle, D-06120 Halle (Saale), Germany b Department of Biochemistry and Biotechnology, Martin-Luther-University of Halle, D-06120 Halle (Saale), Germany c Maryland Biotechnology Institute, Rockville, MD 20850, USA d Intstitute of Biochemistry, Martin-Luther-Unicersity of Halle, D-06120 Halle (Saale), Germany

Received 6 April 1995; accepted 27 February 1996

Abstract Two series of N-aminoacyl, O-benzoyl hydroxamates were designed to investigate the influence of the substituted benzoyl residue on the hydrolytic stability and the reactivity of these potential inhibitors towards selected cysteine and serine proteinases. The inactivators react more rapidly with cysteine proteinases than with the serine enzymes tested. While Z-Phe-Gly-NHO-Nbz is the most reactive inhibitor of cathepsin L, inhibiting the target protein by a second order rate constant of 932.000M-1 s - t , the bacterial serine proteinase thermitase is inhibited best by Z-Gly-Phe-NHO-Nbz, exhibiting a second-order rate constant of 1.170 M - l s-J. Thiolsubtilisin, having the thiol-group as the reactive nucleophile instead of serine, exhibits specificity constants of the inactivation two orders of magnitude smaller than subtilisin. The degree of selectivity of the inhibitors relative to cathepsin B, cathepsin L, cathepsin S and papain varies up to two orders of magnitude with respect to their second order rate constant of inactivation. The inhibitory reactivity of these compounds varies only up to sixfold depending on the benzoyl substituent. Similarly, the rate constants for the hydrolytic decomposition of the compounds vary by a factor of about 6, suggesting that the structural and mechanistic features of the compounds which are responsible for decomposition as well as for the enzyme inhibition are the same. Comparing both reactions, the data allow the calculation of an acceleration factor of 2.4 × 10 l° for the inhibition of cathepsin L by its most effective inhibitor, clearly characterizing this enzyme inhibition reaction as enzyme-activated. Keywords: Serine proteinase; Cysteine proteinase; Inhibitor; Hydroxamate; Mechanism

I. Introduction Proteins, and among them the proteinases are playing an increasingly important role as targets for drug design [1]. Many diseases or their symptoms originate from a deficiency or an excess of a specific metabolite, which is

Abbreviations: BOC, tert-butyloxycarbonyl; Suc, succinyl; Z, benzyloxycarbonyl; NHMec, 4-methyl-7-coumarylamide; 4-NA, 4-nitroanilide; BzOMe, 4-methoxybenzoyl; BzMe, 4-methylbenzoyl; Bz, benzoyl; BzC1, 4-chlorobenzoyl; Nbz, 4-nitroben:,~oyl; -NHO-, hydroxylamine linkage; Cath.B, cathepsin B (EC 3.4.22.1); Cath.L, cathepsin L (EC 3.4.22.15); Cath.S, cathepsin S (EC 3.4.22.27); pap., papain (EC 3.4.22.2); subt., subtilisin Carlsberg (EC 3.4.21.62). * Corresponding author. Fax: +49 345 5583672; e-mail: hudemuth.hki @t-online.de. 0167-4838/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved PI1 S 0 1 6 7 - 4 8 3 8 ( 9 6 ) 0 0 0 3 8 - 6

either a substrate or product of a proteolytic reaction. Thus, modulation of a single enzyme in a sequence of reaction steps can influence the overall effect of a physiological enzyme cascade. Therefore, research on enzyme inhibitors as potential drugs is a important approach in medicinal chemistry [2]. Especially mechanism-based inhibitors, compounds activated during interaction between a target protein and an inhibitor, have the potential to lower side effects in drug administration. Since the reactivity of the inhibitor is dependent on the catalytic action of the target enzyme on that inhibitor, nonspecific interactions with other proteins can be circumvented. W e have developed a class of such mechanism-based proteinase inhibitors [3,4] which selectively inhibit serine and cysteine proteinases. Peptidyl O-acylhydroxamates can easily comply to the specificity and reactivity of the target enzyme

180

H.-U. Demuth et al. / Biochimica et Biophysica Acta 1295 (1996) 179-186

X--NH--CH\c

[

d' --~--%C--R2 d'

Enz-SH

/

= -.O=CO_R2 ~

X--NH--CH\c__N H

d'

Enz-OH

X--NH--CH NH

\s--~

I

o//C~'O--Eaz

Scheme 1. Proposed mechanisms of inhibition of serine and cysteine proteinases by N-peptidyl, O-acyl hydroxamates.

simply by modifying the N-acyl and the O-acyl residue of the hydroxylamine linkage. These compounds inactivate serine proteinases with second-order rate constants of 2to 10000 [4,5] depending on the length of the N-peptidyl residues. Such inhibitors containing substrate analog dipeptide residues in particular, have been shown to be very potent and selective towards cysteine proteinases as well [6,7]. In designing inhibitors for aminopeptidases, we could demonstrate that only covalent catalyzing enzymes are inhibited by this new class of inhibitors [8]. They inhibit the cysteine proteinase papain by forming a sulphenamidadduct [9]. In contrast, crystallization studies indicate that the serine proteinase subtilisin is inhibited by carbamylation of the serine in its active site, after rearrangement of

the inhibitor molecule [ 10]. Both reaction paths involve the modification of the active-site nucleophile by a reactive chemical intermediate (Scheme 1). Electron-withdrawing residues of R 2 increase the reactivity of the inhibitors [4]. By selecting proper substituents these residues may also be used to modulate the affinity of the inhibitor towards the target enzyme [4,6,7]. Since mammalian lysosomal cysteine proteinases as well as bacterial and viral thiol-dependent invasion factors represent attractive targets for selective inhibition testing, the electronic influence of the O-acyl residue of the peptidyl hydroxamates on the selectivity of the inhibitors is of interest. So far, in the case of the proline-specific aminopeptidase DP IV an appropriate systematic study has been performed. The aim of the present study is to use two series of model compounds to investigate whether the choice of a certain benzoyl substituent influences the reactivity and selectivity of the inhibitors towards selected serine and cysteine proteinases.

2. Materials and methods

2.1. General, synthesis and product analysis All reagents for synthesis were purchased from commercial sources and solvents were dried using common procedures, oL-N-Z-peptidyi methyl esters were obtained according to known standard procedures [11]. The hydroxamic acids were obtained by treatment of the corresponding methyl esters with 3.5 M hydroxylamine in a methanol solution followed by O-benzoylation according to Demuth

Table 1 Analytical data of N-dipeptidyl, O-benzoyl hydroxamates No.

Inhibitor

Mr

C

H

N

1

Z-Phe-GIy-NHO-BzOMe b (C 27H 27N30 7) Z-Phe-Gly-NHO-BzMe b (C 27H 27N3 06 ) Z-Phe-GIy-NHO-Bz b (C 26H 25N3 O6) Z-Phe-GIy-NHO-BzC1 b (C26 H 24N306C1) Z-Phe-GIy-NHO-Nbz b (C 26H 24N40 8) Z-Gly-Phe-NHO-BzOMe b (C 27H 27N30 7) Z-GIy-Phe-NHO-BzMe b (C 27H 27N306 ) Z-Gly-Phe-NHO-Bz b (C 26H 25N3 06 ) Z-GIy-Phe-NHO-BzCI b (C26 H24N306CI) Z-GIy-Phe-NHO-Nbz b (C 26H 24N408 )

505.17 found: 489.18 found: 475.16 found: 509.11 found: 520.13 found: 505.17 found: 489.18 found: 475.16 found: 509.11 found: 520.13 found:

64.14 62.98 66.24 66.04 65.67 64.99 61.24 61.10 60.00 59.84 64.14 64.43 66.24 65.91 65.67 65.71 61.24 60.15 59.99 59.84

5.30 5.48 5.56 5.54 5.299 5.44 4.74 4.72 4.65 4.69 5.299 5.46 5.56 5.55 5.299 5.30 4.74 4.69 4.65 5.00

8.84 8.51 8.58 8.51 8.84 9.01 8.24 8.14 10.76 10.72 8.84 8.35 8.58 8.53 8.84 8.76 8.24 8.28 10.77 10.84

2 3 4 5 6 7 8 9 10

MS(M + H) +

M.p. (°C) a

506.10

157 - 160

490.16

144-147

476.10

128-136

510.04

118-120

521.04

1 l 8-122

506.11

136-139

490.10

128 - 130

476.16

138-139

510.08

149-152

521.12

139-142

a Uncorrected values. b BzOMe, 4-methoxybenzoyl; BzMe, 4-methylbenzoyl; Bz, benzoyl; BzCI, 4-chlorobenzoyl; Nbz, 4-nitrobenzoyl; -NHO-, hydroxylamine linkage.

181

H.-U. Demuth et aL /Biochimica et Biophysica Acta 1295 (1996) 179-186

et al. [4,12,13]. The Z-protected inhibitors were purified by crystallization from methanol/water resulting in 50-96% yield of the expected product (Table 1). After drying and recrystallization from methanol/diethyl ether the purity of the products was checked by TLC. HPLC-analysis of all compounds was perfornaed using a Merck-Hitachi-system equipped with a photodiode array detector. The structures and molecular weights were confirmed by mass spectroscopy and elemental analysis. Mass spectra were taken in positive ion mode on a Fisons VG-BIO-Q triple quadrupole tandem mass spectrometer equipped with an electrospray atmospheric pressure LC-interface (Manchester, UK). Data were acquired and processed on a Intel486 personal computer system.

Table 2 Pseudo-first-order rate constants of decomposition of N-dipeptidyl, Obenzoyl hydroxamates No. Inhibitor Rate constant, trp-Substituent 103k (rain- i ) constants l 2 3 4 5 6 7 8 9 10

Z-Phe-GIy-NHO-BzOMe Z-Phe-Gly-NHO-BzMe Z-Phe-GIy-NHO-Bz Z-Phe-GIy-NHO-BzC1 Z-Phe-GIy-NHO-Nbz Z-GIy-Phe-NHO-BzOMe Z-Gly-Phe-NHO-BzMe Z-Gly-Phe-NHO-Bz Z-Gly-Phe-NHO-BzCI Z-GIy-Phe-NHO-Nbz

0.31 0.42 0.45 0.68 2.27 1.38 1.77 2.89 4.62 7.04

- 0.27 - 0.17 0.0 0.23 0.78 - 0.27 - 0.17 0.0 0.23 0.78

2.2. Enzymes

Papain and subtilisin Carlsberg were purchased from SIGMA (USA) and used without further purification. The double-mutant enzyme thiolsubtilisin (subtilisin $221C N218A) was prepared according to Bryan et al. [14]. Cathepsin B (EC 3.4.22. l) and cathepsin L (EC 3.4.22.15) were prepared from the lysosomal fraction of rat liver as described by Kirschke et ~d. [15] and Barrett and Kirschke [16]. Bovine spleen cathepsin S (EC 3.4.22.27) was isolated according to Kirschke et al. [ 17]. All cathepsins used were in an electrophoretically homogeneous form. 2.3. Substrates

Z-Phe-Arg-NHMec and Z-Val-Val-Arg-NHMec were synthesized as in Brrmme et al. [7,18]. Benzoyl-Arg-4-NA and Suc-Ala-Ala-Ala-4-NA were products of Bachem (Switzerland). Suc-Ala-Ala-Phe-NHMec was a gift of Prof. S. Fittkau, Institute of Biochemistry, Martin-Luther-University of Halle. 2.4. Decomposition o f the inhibitors

The stability of the hydroxamates in aqueous solution has been tested by means of UV-spectroscopy according to Demuth et al. [4]. Inhibitors (10-200 o~M) were incubated in 40mM Tricine buffer (pH 7.6) adjusted to an ionic strength of 0.125. The total volume in a typical experiment was 2.0ml containing 5% acetonitrile or dimethyl sulfoxide (v/v). Kinetic runs were recorded over a range of 225 to 300 nm using a Carl-Zeiss-Jena microprocessor-controlled M 40 spectropbotometer equipped with a jacketed cell compartment, containing electrical heat and temperature control at 30 + 0.1°C. The data were collected and stored to an internal R i l l - b u f f e r and analyzed using software packages provided on a ROM-card 'reaction kinetics' for the instrument. Parameters of decomposition reactions with half-lives longer than 200 min were calculated by fitting the data collected at 30 min time intervals to a first-order-reaction model using non-linear regression

programs running at PC-compatible computers. Pseudofirst-order rate constants of the decomposition compiled in Table 2 represent the mean of triplicate measurements with a standard error less than 10%. The standard error of the pseudo-first order rate constants estimated as described was in all cases less than 5%. Measurements where performed at least over the time-course of 5 half-times of decomposition. 2.5. Enzyme assays

Inactivation of proteinases with substrate analogous inhibitors proceeds according to Eq. (1): ki

k.

E+I~E.I~E-I

(1)

k i

where E- I, E-I, k i and k_i represent the enzyme-inhibitor complex, the inactivated enzyme and the rate constants for the formation of the non-covalent enzyme-inhibitor complex. k 2 is the first-order rate constant for the formation of modified enzyme when enzyme is saturated with the inhibitor. This rate constant may be estimated according to Kitz and Wilson [19] by preincubation of enzyme and inhibitor, and subsequent estimation of residual activity in an assay with substrate. Tian and Tsou [20] introduced a more convenient method for evaluating inactivation rates in the presence of substrate according to Eq. (2): ki

kz

I~E.I~E-I k-i

+ E

(2) + kI

kcat

S~E'S~E+S k-i

where substrate and inhibitor are competing for the enzyme's binding site. Since the decrease in enzyme concentration during incubation with inhibitor follows pseudo-first order kinetics the enzyme-catalyzed substrate turnover fol-

182

H.-U. Demuth et aL / Biochimica et Biophysica Acta 1295 (1996) 179-186

lows proportionately. By applying chromogenic or fluorogenic substrates, and establishing steady-state conditions during the inactivation time, first-order rate constants may be obtained by fitting absorbency and time values to an exponential function. Measurements made at constant inhibitor and different substrate concentrations give different kobs-Values. Extrapolation of these values to a substrate concentration of zero, gives the apparent rate constant kapp. By fitting rate constants evaluated at different inhibitor concentrations to a hyperbola, the inactivation parameters K i ( = k _ J k l) and k 2 may be obtained. This method was applied here. The progress curves of the inactivation of the cysteine proteinases cathepsin B, L, S, thiolsubtilisin and the serine proteinase thermitase in the presence of substrate were monitored at 25°C using a Shimadzu spectrophotometer UV-300 or a Carry 2200 spectrophotometer equipped with a fluorescence detection unit at an excitation wavelength of 383 nm and with an emission filter of 450nm. Kinetic experiments were carried out using a constant enzyme concentration in 50 mM-acetate buffer (pH 5.5) for cathepsin L, in 50 mMphosphate buffer containing 0.01% ( v / v ) Triton X-100 (pH 6.5) for cathepsin S, and in 50 mM phosphate buffer (pH 6.0) for cathepsin B. Enzyme concentrations in all experiments were 0.7nM, 2.3nM and 0.9nM for the cathepsins L, S and B, respectively. For the activation of the cathepsins L, B and S the enzymes were incubated for 5min at 25°C with 2.5mM dithioerythritol, 2.5mM EDTA* Na 2 and 0.005% ( v / v ) Brij-35 in assay buffer. The reaction was started by addition of the activated enzyme (0.5 ml) to 1 ml substrate in assay buffer containing seven different inhibitor concentrations (per one substrate concentration). Substrates were Z-Phe-Arg-NHMec (3 txM and 8 txM) for cathepsin L, Z-Phe-Arg-NHMec (10 txM and 50 IxM) for cathepsin B and Z-Val-Val-ArgNHMec (10 IxM and 50 txM) for cathepsin S. Thermitase (0.1 to 0.5 IxM in the assay) was analyzed without activation in 50mM phosphate buffer (pH 7.6) using Suc-Ala-Ala-Phe-NHMec (10 IxM and 50 ixM) as substrate. Thiolsubtilisin (5 to 50 jxM in the assay) was analyzed in 50mM Tris-HC! buffer (pH 7.5) using SucAla-Ala-Phe-NHMec (10 IxM and 50 jxM) as substrate. The activities of papain and subtilisin Carlsberg were analyzed spectrophotometrically on a Kontron 930 UV-Vis spectrophotometer using the chromogenic substrates benzoyl-arginyl-4-NA (0.5 and 1.1 mM) and Suc-AIa-Aia-Ala4-NA (1.4 mM and 2.9 mM), respectively. Papain (1.16 txM in the assay) was investigated after a 10min incubation at 25°C in the assay buffer containing 2.0mM cysteine and 1.0mM-EDTA* Na 2. 0.5 ml of the activated enzyme was added to 1.0 ml of the buffer solution containing substrate and different inhibitor concentrations. Subtilisin (0.24 IxM in the assay) was analyzed at 25°C in 50 mM Tricine buffer (pH 7.6) without activation. Progress curves reflecting the enzyme-catalyzed substrate hydrolysis in the presence of inactivator were followed for

3 to 5 half-times of the occurring enzyme inactivation and obtained in duplicates. The parabolic progress curves were analyzed according to the method described above using non-linear regression programs running on a PC-compatible computer. In cases where no saturation of the enzyme by the inhibitor was achieved ([I] << K i) the rate constants have been calculated according to Crawford [21]. The resulting pseudo-first order rate constants (kob~) were used for secondary plotting if their standard deviations were smaller than 5%. 2.6. Stability of the inhibitors

N-peptidyl, O-benzoyl hydroxamates decompose in aqueous solutions yielding peptide hydroxamic acids and the benzoic acid [4,12]. Spontaneous decomposition of diacyl hydroxamates used in this work has been followed UV-spectrometrically in the range of 225to 400nm at 30°C. Data were collected at several wavelengths as functions of time and the pseudo-first-order rate constants calculated using non-linear regression programs as described by Demuth et al. [4,12]. Reactions were monitored in 10mm silica cells of 2.5 ml buffer solution consisting of 0.1 mM inhibitor and 0.04 M sodium phosphate or Tricine buffer (pH 7.6) with ionic strength adjusted to 0.125 with KC1. Thus, enzyme kinetic experiments could be planned to maintain constant inhibitor concentrations by taking the inhibitor decomposition rates into account. 2.7. Reversibility of the modification

To demonstrate chemical modification of the target enzymes by the inhibitors, enzyme was incubated for 15 min with 10 to 50 IxM inhibitor solutions. The reaction mixture was then placed in Centricell molecular filter tubes (Polyscience) having an exclusion molecular weight of 10000Da. Protein and inhibitor were separated by repetitive centrifugation for one hour at 5000rpm and subsequent washing with buffer solution. The activity of the protein solution and of controls was estimated as described above.

3. Results and discussion

3.1. Stability of the inhibitors

In designing inhibitors for potential diagnostic purposes it is crucial to predict the life-time of these compounds in biological fluids. This often depends on the biodegradability of these compounds and on their chemical stability under physiological conditions. N-peptidyl, O-acyl hydroxamates have been designed with these factors in mind. Affinity to the target enzyme is optimized by choosing the

H.-U. Demuth et aL / Biochimica et Biophysica Acta 1295 (1996) 179-186 018 -

0,8

p = 0.793

Z

N

n

2

p-C .~

0,2

t~ o

0,0

,

.

~

-

j

= .

-0,2 -0,4 -0,4

. I

-0,2

i

0,0

i

0,2

[

0,4

I

0,6

I

0,8

O"

Fig. 1. Dependence of the pseudo-first-order rate constants of the decomposition of differently substitutecl N-Z-Phe-Gly- and N-Z-Gly-Phe-, Obenzoyl hydroxamates on the the substituent constants cr (Hammet-correlation). (The pseudo-first-order rate constants of the decomposition were taken from Table 2, the value found for the decomposition of the O-benzoyl derivative is set k0, and cr-values are taken from reference [28]). ( O ) Z-Phe-Gly-NHO-derivatives, ( 0 ) Z-Gly-Phe-NHO-derivatives.

corresponding substrate-analog peptide residue. Modulation of chemical stability and additional binding specificity is achieved by selecting the proper O-acyl residue [2]. In the range of pH 4to pH 9the rate constants of degradation of N-terminally protected N-peptidyl, O-acyl hydroxamates only increase within one order of magnitude showing pKa-values of 4.8 to 5.5 [4,5]. The main reaction product, peptidyl hydro,~amic acid, is formed in a monomolecular decomposition reaction [22] that is not catalyzed by buffer, water, hydroxyl or hydronium ions [ 12]. The acid proton of the hydroxamate linkage results in the observed pK,-values for this reaction [22,23]. In the present study we investigated the stability of the N-termi-

183

nal Z-protected dipeptide derivatives at pH 7.6 and compared the electronic influence of the substituent of the benzoyl leaving group on the decomposition reaction with the influence of this residue in inactivating the particular serine and cysteine proteinases. The results are listed in Table 2. The stability of the N-terminally protected O-benzoyl hydroxamates is comparable to what is known from previous work [4,7,12,13,22]. Compounds having Gly in position PI appear to be more stable than the inhibitors having Phe, Ala or Val- in this position. The linear free-energy (LFE) relationship of the reactions shows a modest positive dependence of the reaction rate on the electronwithdrawing forces of the substituents (Fig. 1). The Hammet correlation coefficients p are estimated at 0.793 ( r = 0.993) and 0.708 ( r = 0.959) for the pseudo-first-order rate constants of the decomposition of Z-Phe-Gly-NHO-Bz and Z-Gly-Phe-NHO-Bz with the benzoyl substituents 4-OMe, 4-CH 3, H, 4-C1 and 4-NO 2, respectively. These values are consistent with the value of p = 0.62, found for the decomposition of substituted H-Gly-Pro-NHO-Bz hydroxamates and what has been found for the decomposition of dibenzoyl hydroxamates (p = 0.86) by Renfrow and Hauser [241.

3.2. Reactions of peptidyl hydroxamates with cysteine proteinases Incubation of cathepsin B, cathepsin L, cathepsin S and papain with the Z-protected compounds 1-5resulted in time-dependent inhibition of their enzyme-catalyzed substrate hydrolysis. While saturation for all cathepsins was observed, the solubility limits of the N-Z-Phe-GIy-, Obenzoyl hydroxamates did not allow the estimation of the Ki-values in the case of the inactivation of papain. How-

Table 3 Parameters of the inactivation of several cysteine proteinases by substituted N-Z-Phe-GIy-, O-benzoyl hydroxamatesN-Z-Phe-Gly-, O-benzoyl hydroxamates No.

Inhibitor

1

Z-Phe-Gly-NHO-BzOMe

2

Z-Phe-Gly-NHO-BzMe

3

Z-Phe-Gly-NHO-Bz

4

Z-Phe-Gly-NHO-BzC1

5

Z-Phe-Gly-NHO-Nbz

kapp/[I]" b kapp" e Inhibitor concentration.

k~ ( s - n) Ki (tzM) k 2 / K i (M -1 S - 1 ) k 2 (s - l ) K~ (i.zM) k 2 / K i ( M - 1 s - i) k 2 ( s - 1) K~ (p.M) k 2 / K i (M- 1 s - I) k2 ( s - I ) K~ (I~M) k2/K i (M-I s-I) k 2 ( s - n) K i (ixM) k~./K i ( M - 1 s - I )

Cath.B

Cath.L.

Cath.S.

Papain

0.09 8.9 10100 0.11 16 6900 0.19 16 8900 0.26 15 17300 0.19 12 15 800

0.02 b (1.3) ~ 152400 ~ 0.047 0.15 313300 0.031 0.082 378000 0.08 b (1.3) c 615515 a 0.041 0.044 931 818

0.05 0.76 65800 0.085 2.0 42500 0.036 0.85 42300 0.047 0.48 97900 0.086 0.9 95 500

0.0103 b (9.9) c 1028 ~ 0.009 b (5.2) c 1784 a 0.0123 b (5.2) ~ 2433 a 0.0134 b (5.2) ~ 2589 a 0.0114 b (2.01) c 5670 a

184

H.-U. Demuth et al. /Biochimica et Biophysica Acta 1295 (1996) 179-186

0,4

p = 0.659

~

o

2

0,2. ~,~,-~ 0,0-

; C Hp -3 C~ I H

0

--

-.0,2. -o,4. -0,4

p = 0.616

~p-OCH3 I

-0,2

I

0,0

I

0,2

i

0,4

i

0,6

I

0,8

Fig. 2. Dependence of the second-order rate constants of the inactivation of cathepsin L and papain by differently substituted N-Z-Phe-Gly-, O-benzoyl hydroxamates on the substituent constants t~ (tr-values are taken from [28]). (The inactivation rate constants were taken from Table 3, and the value found for the inactivation of the enzymes by the O-benzoyl derivative is set k 0, ((3) inactivation of cathepsin L, ( 0 ) inactivation of papain).

ever, considering the Ki-value of 0.9 mM found for the inactivation of papain by N-Boc-Tyr-Gly-NHO-Bz in the comparable range, the kapp/[I]-values obtained should give a reasonable estimate of the specificity parameter k 2 / K ~. The inactivation parameters are listed in Table 3. As is easily seen from Table 3, the inactivation rate constants of all cysteine proteinases tested are relatively independent of the electronic nature of the substituent of the benzoyl leaving group of the inhibitors. Half-lives for the inactivation reaction range from 7.7 to 3.7 seconds and from 67 to 52 seconds for cathepsin B and papain, respectively. Generally, cathepsin B and cathepsin S show a poor tendency to be rapidly inactivated by the inhibitors having more electro-negative benzoyl substituents. The correlation coefficients,, of the linear-free-energy relationship for the second-order rate constants of the inactivation are 0.354 ( r = 0.82) and 0.347 ( r = 0.8). However, the second-order rate constants for the inactivation process of cathepsin L and papain result a moderate Hammet-correlation (Fig. 2). The correlation coefficients are calculated to be 0.66 ( r = 0.926) and 0.616 ( r = 0.973) for cathepsin L and papain. The results show that the rate determining step of the inactivation (k 2) is, if at all, only minimally influenced by the nature of the O-benzoyl residue of the hydroxamates. That means that the departure of the O-benzoyl residue takes place at a distinct step before the target enzyme is chemically modified by the inhibitor. This is reflected by the dependence of the bimolecular reaction constant ( k 2 / K i) on the electro-negativity of the benzoyl substituents. While the reactions of cathepsin B and cathepsin S are not very sensitive to the nature of the leaving group, the inactivation of papain and cathepsin L is almost equally influenced by the inhibitors leaving group as the sponta-

neous decomposition of the compounds. These differences might be explained by the different impacts of single steps in the catalytic reaction path on the overall-rate constant. It has been shown by Krantz and coworkers, that especially in the inactivation of cathepsin B, steric elements of the departing leaving benzoyl residue greatly influence the specificity constant of enzyme inactivation [6]. Considering the H 2 0 concentration of 55M, the second-order rate constants for the hydrolytic decomposition of the N-Z-Phe-Gly-, O-benzoyl hydroxamates are between 4.13 × -5 for the 4-nitrobenzoyl derivative and 5.64 × - 6 for the 4-methoxybenzoyl derivative. Since in both enzyme inhibition and inhibitor decomposition, N-O-bond fission is the key chemical step, it is of significance that the cysteine proteinases tested in this study enhance this step by a factor of up to ten orders of magnitude. Therefore, this can be considered as 'enzyme-activated' inhibition.

3.3. Reactions of peptidyl hydroxamates with serine proteinases N-peptidyl-O-acyl hydroxamates react less effectively with serine proteinases than with cysteine proteinases [2,7]. While the rate constants of inactivation are around 1 - 2 orders of magnitude smaller and exhibit half-lifes of around 100 seconds, the specificity constants of the inhibition reactions are especially influenced by the weaker binding between the compounds and the serine enzymes. Thermitase inhibition by Z-GIy-Phe-NH-O-Nbz was found most effective, resulting in a k 2 / K i value of about 1000 s - 1 M - 1 (Table 4). As with the cysteine proteinases investigated in this study, the rate determining inactivation step for serine proteinases is not greatly influenced by the nature of the benzoyl leaving group of the inhibitors. While for thermitase inactivation, the rate constants of inactivation, k 2, reflecting this step, are weakly dependent on the nature of the leaving benzoyl residue, the rate constants for subtilisin inactivation show no significant coherence to the different substitutions on the inhibitors (Table 4). However, the Hammet-correlation of the second-order inactivation rate constant k 2 / K ~ of thermitase results a substantial correlation coefficient of p = 0.974 ( r = 0.989). The correlation of k 2 / K i for subtilisin inactivation shows a modest correlation, resulting in a Hammet-correlation coefficient of p = 0.40 ( r = 0.934). This strongly suggests an influence of the departing benzoyl residue on reaction steps of the interaction between enzyme and inhibitor before the inactivation of the protein occurs. The linear free-energy relationship analysis of the inhibition of serine and cysteine proteinases in this paper demonstrate that the variation of the leaving O-acyl residue might influence the overall inactivation reaction to some extent depending on the individual target proteinase. Thus,

H. -U. Demuth et al. / Biochimica et Biophysica Acta 1295 (1996) 1 7 9 - 1 8 6

185

Table 4 Parameters of the inactivation c f subtilisin-type serine proteases and the mutant cysteine protease thiolsubtilisin by substituted N-Z-GIy-Phe-, O-benzoyl hydroxamates No.

Inhibitor

1

Z-GIy-Phe-NHO-BzOMe

2

Z-Gly-Phe-NHO-BzMe

3

Z-Gly-Phe-NHO.-Bz

4

Z-GIy-Phe-NHO.-BzC1

5

Z-Gly-Phe-NHO-Nbz

103k2 ( s - i ) K i (I.zM) k 2 / K i M -1 s - l ) 103k2 ( s - I ) K i (IxM) k 2 / K i M -1 s -~) 103k2 ( s - l ) K~ ( ~ M ) k 2 / g i M -1 s - l ) 103 k 2 ( s - t) K i (I.LM) k 2 / K i M - t s - 1) 103 k 2 ( s - I) K i (txM) k2/K i M -I s -1)

Thermitase a

Subtilisin b

Thiolsubtilisin c

3.0 30 100 4.0 25 170 5.0 31 161 6.0 14 428 7.0 60 1170

2.9 69 42 3.8 56 68 7.0 73 96 2.0 23 87 14.0 11 128

0.12 f (100) g 1.2 e 0.18 f (100) g 1.75 e 0.25 f (100) g 2.5 ~ 0.38 f (100) g 3.8 e 0.21 f (100) g 2.1 de

a In 50 mM phosphate buffer (pH 7.6). b In 50 m M Tricine buffer (pH 7.6). c In 50 mM Tris-HC1 buffer (pl-I 7.5). d Partial decomposition of the inhibitor during reaction time. e kapp/[i] ' f k~pv" g Inhibitor concentration.

in each individual case an appropriate study must be performed to match the most suitable inhibitor with a given enzyme.

3.4. Reactions of peptidyl hydroxamates with thiolsubtilisin Thiolsubtilisin was previously prepared and kinetically analyzed to study mechanistic similarities and differences of serine and cysteine proteinases (Philipp and Bender [25]). The exchange of the nucleophilic heteroatom of the active-site nucleophile influences the deacylation step of the catalytic cycle to only a small extent. But, this mutation drastically changes the acylation mechanism compared to native subtilisin. Thiolsubtilisin hydrolyzes amide substrates 100-fold more slowly than the serine enzyme. This may be explained by the fact that thiol leaving groups are better than amine leaving groups. This could account for a higher energy barrier for acylation of thiol-subtilisin and thus result in a prefelxed regeneration of substrate and free enzyme instead of the formation of the acyl enzyme. With these results in mind we compared the inactivation of subtilisin and thiolsubtilisin by N-Z-Gly-Phe-, O-benzoyl hydroxamates, finding that thiol-subtilisin is inactivated about 40-times slower than subtilisin (Table 4). In contrast to our results, inactivation of thiolsubtilisin by substrate analog chloromethyl-ketones occurred approx. 100-times faster than with subtilisin (Tsai and Bender [29]). There, the reaction with the serine enzyme leads to the alkylation of histidine in the active-site, while the alkylation on thiolsubtilisin occurs at the nucleophilic sulfhydryl residue of the active site. Though during sub-

strate hydrolysis the sequence of reactions steps and the nature of the intermediates are essentially the same, the inhibition by chloromethyl ketones results different products. The product formed in the inactivation of the cysteine proteinase papain by diacyl hydroxamates is a sulphenamid-derivative, while the inhibition product of the serine proteinase subtilisin is a carbamoyl-derivative [10,26]. However, for both reactions the initial chemical step of inactivation is likely the formation of a tetrahedral adduct [6,21,27,30]. If this is so, then a similar energy barrier in the formation of the tetrahedral intermediate (as argued for the acylation in amide substrate hydrolysis by subtilisin and thiolsubtilisin) could be responsible for the less effective inactivation of thiolsubtilisin by the N,O-diacyl hydroxamate inhibitors. Product analysis of the chemical species formed in the reaction between inhibitor and the serine or cysteine subtilisins under acid conditions will give further insight in the reaction mechanism of this new class of proteinase inhibitors [30].

Acknowledgements This work was supported in part by the German Research Foundation: 'Deutsche Forschungsgemeinschaft', Grant No.: De 471/1-2. The authors are grateful to R.P. Pauly, UBCA, for critical reading of the manuscript.

186

H.-U. Demuth et al. / Biochimica et Biophysica Acta 1295 (1996) 179-186

References [1] Walkinshaw, M.D. (1992) Med. Res. Rev., 12, 317-372. [2] Demuth, H.-U. (1990) J. Enzyme Inhib. 3, 249-278. [3] Fischer, G., Demuth, H.-U. and Barth, A. (1983) Pharmazie, 38, 249-250. [4] Demuth, H.-U., Baumgra6, R., Schaper, C., Fischer, G. and Barth, A. (1988) J. Enzyme Inhib. 2, 129-142. [5] Schmidt, C. Schmidt, R. and Demuth, H.-U. (1991) in Proc. 21th Eur. Peptide Symposium (Giralt, E. and Andreu, D., eds.), pp. 761-762, Elsevier, Amsterdam. [6] Smith, R.A., Coles, P.J., Spencer, R.W., Copp, L.J., Jones, C.S. and Krantz, A. (1988) Biochem. Biophys. Res. Commun. 155, 12011207. [7] BriSmme, D., Schierhorn, A., Kirschke, H., Wiederanders, B., Barth, A., Fittkau, S. and Demuth, H.-U. (1989) Biochem. J. 263, 861-866. [8] Demuth, H.-U., StiAckel, A., Schierhorn, A., Fittkau, S., Kirschke, H. and Br~mme, D. (1993) Biochim. Biophys. Acta 1202, 265-270. [9] M6nard, R., Feng, R., Storer, A.C., Robinson, V.J., Smith, R.A. and Krantz, A. (1991) FEBS Lett. 295, 27-30. [10] Steinmetz, A.C.U., Demuth, H.-U. and Ringe, D. (1994) Biochemistry 33, 10535-10544. [11] Wiinsch, E. (1974) in Synthesis of Peptides, Houben-Weyl, Vol. 15/I, Methods in Organic Chemistry (MUller, E., ed.), GeorgThieme, Stuttgart. [12] Demuth, H.-U., Neumann, U. and Barth, A. (1989) J. Enzyme Inhib. 2, 239-248. [13] Demuth, H.-U., Schtinlein, C. and Barth, A. (1989) Biochim. Biophys. Acta 996, 19-22.

[14] Bryan, P., Alexander, P., Strausberg, Schwarz, F., Lan, W., Gilliland, G. and Gallagher, D.T. (1992) Biochemistry 31, 4937-4945. [15] Kirschke, H., Langner, J., Wiederanders, B., Ansorge, S. and Bohley, P. (1977) Eur. J. Biochem. 74, 293-301. [16] Barrett, A.J. and Kirschke, H. (1981) Methods Enzymol. 80, 535561. [17] Kirschke, H. and Shaw, E. (1981) Biochem. Biophys. Res. Commun. 101,454-458. [18] Br~mme, D., Steinert, A., Friebe, S., Wiederanders, B. and Kirschke, H. (1989) Biochem. J. 264, 475-481. [19] Kitz, R. and Wilson, I.B. (1962) J. Biol. Chem. 237, 3245-3249. [20] Tian, W.-X. and Tsou, C.-L. (1982) Biochemistry 21, 1028-1032. [21] Crawford, C., Mason, R.W., Wikstri3m, P. and Shaw, E. (1988) Biochem. J. 253, 751-758. [22] Demuth, H.-U., Fischer, G., Barth, A. and Schowen, R.L. (1989) J. Org. Chem. 54, 5880-5883. [23] Bauer, L. and Exner, O. (1974) Angew. Chem. 13, 376-384. [24] Renfrow, W.B. and Hauser, C.R. (1937) J. Am. Chem. Soc. 59, 2308-2314. [25] Philipp, M. and Bender, M.L (1983) Mol. Cell. Biochem. 51, 5-32. [26] Mentlein, R. (1988) FEBS Lett. 234, 251-256. [27] Robinson, V.J., Coles, P.J., Smith, R.A. and Krantz, A. (1991) J. Am. Chem. Soc. 113, 7760-7761. [28] Jaff6, H.H. (1953) Chem. Rev. 17, 191-207. [29] Tsai, I.-H. and Bender, M.L. (1979) Biochemistry 18, 3764-3768. [30] Ding, X.C, Rasmussen, B.F., Demuth, H.-U., Ringe, D., Steinmetz, A.C.U. (1995) Biochemistry 34, 7749-7756.