ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 280, No. 1, July, pp. 137-1461990

Structure-Activity Studies of Fluoroketone Inhibitors of wLytic Protease and Human Leukocyte Elastase’** Chandrika Brandeis

P. Govardhan

University,

Received November

and Robert H. Abeles3

Graduate Department

of Biochemistry,

17,1989, and in revised form February

415 South Street, Waltham, Massachusetts

26,199O

We have synthesized a series of peptidyl fluoroketones that reversibly inhibit the serine proteases human leukocyte elastase (HLE) and cu-lytic protease ((Y-LP). Ac-ambo-AlaCF, (1) inhibits HLE and a-LP with Ki’s of 2.4 and 15 mM, respectively. The effects of structural variations on this parent compound on Ki and the kinetics of inhibition were studied. The acetyl group was replaced by the tripeptide Z-L-Ala-L-Ala-L-Pro to yield the tetrapeptide trifluoroketone (TFK) Z-L-Ala-t-AlaL-Pro-ambo-AlaCF, (2). This extension reduced Ki 3500-fold for HLE and 3000-fold for (Y-LP. Removal of a fluorine atom from a TFK decreases Ki about 15- to 30-fold with both enzymes. Replacement of one fluorine atom of 2 by a residue (-CHz-CHz-COLeuOMe) (6) which can interact with the S; and Sh subsites decreased Ki 30-fold for HLE and 150-fold for (Y-LP compared to Z-L-Ala-L-Ala-t-Pro-ambo-AlaCFzH (3). The Ki of 6 for HLE is approximately equal to that of trifluoroketone 2. For (Y-LP Ki of 6 is lo-fold lower than that for the trifluoroketone 2. Inhibitors with Ki values < lo-’ M exhibit slow binding kinetics. By analogy to cholinesterases and chymotrypsin, it is likely that these enzymes combine with the keto form of the inhibitor to form the enzyme-inhibitor complex. Therefore, k,, and Ki were corrected for the ketone concentration. The corrected k,, values for the slow binding inhibitors are in most cases less than diffusion controlled, ranging between 8.2 X lo4 and 4.68 X lo6 M-’ s-‘. An exception is Z-L-Ala-l;-Ala-L-Pro-ambo-ValCF3 (8) where k,, = 9 X lo’ M-’ s-l, which is nearly diffusion controlled. 0 1990

Academic

Press,

Inc.

Peptidyl tri- and difluoroketones are potent inhibitors of the serine proteases chymotrypsin, porcine pancreatic i This paper is supported by National Institutes of Health Grant GM12633-26. This is publication 1709. ’ A detailed description of the synthesis, purification, and physical properties of peptidyl Auoromethyl ketones is provided as a Miniprint Supplement. 3 To whom correspondence should be addressed. 0003-9861/90

$3.00

Copyright 0 1990 by Academic Press, Inc. rights of reproduction in any form reserved.

All

02254

elastase (PPE)4 (l), and human leukocyte elastase (HLE) (2,3). These inhibitors are believed to be transition state analogs (4) and form an adduct with the active site serine, as has recently been demonstrated by X-ray crystallography (5,6) and NMR spectroscopy (7). (1) (Fig. 1) is an inhibitor of HLE Ac-ambo-AlaCF, and cu-lytic protease (a-LP; K,.‘s are 2.4 and 15 mM, respectively).5 Inhibitor 1 interacts with the S, subsite only.‘j We explored the effect on Ki, as well as on the kinetics of inhibition, of extending the inhibitor 1 in the S and s’ direction. S subsites are well characterized (8) and it is expected that addition of PZ, Ps, etc., residues will decrease Ki. There is also reason to believe that extension of inhibitors in the S’ direction will lower Ki (9). It has previously been shown that extension of chymotrypsin trifluoroketone (TFK) inhibitor in the S’ direction substantially decreased Ki (10). Studies of naturally occurring proteinase inhibitors (4,15,22) and synthetic elastase inhibitors with nonspecific large alkyl chains (11) have provided evidence that S’ interactions are important. Most small molecule inhibitors of serine proteases cannot be extended in the s’ direction due to chemical limitations of the functional groups (e.g., aldehydes, boronic acid derivatives, etc.). An important advantage of inhibitors based on fluoroketones is that they can be extended to provide interactions with S’ subsites. It ’ The abbreviated designation of the peptidyl inhibitors is as defined in IUPAC-IUB Joint Commission on Biochemical Nomenclature (1985). The term ambo designates the residue which follows as being racemic. Other abbreviations used: PPE, porcine pancreatic elastase; HLE, human leukocyte elastase; TFK, trifluoroketone; ol-LP, a-lytic protease; MeoSuccAAPV-pNA, N-methoxysuccinyl-L-alanyl-L-alanyl-L-alanyl-L-prolyl-L-valyl-p-nitroanilide; AcAPA,NA, N-acetyl-Lalanyl-L-prolyl-L-alanyl-p-nitroanilide. 5 K; values refer to the uncorrected K, values unless otherwise noted. 6 The amino acid residues of substrate (or in this case, inhibitor) are designated Pi, P2, etc., numbered from the carbonyl of the scissile amide bond in the direction of the amino terminal. The corresponding subsites on the enzyme are termed Si, SZr etc. The residues in the direction of the carboxyl group from the scissile bond are designated P;, PZ, etc., and the corresponding enzymatic subsites S;, S; etc. (26). 137

138

GOVARDHAN

AND

ABELES

CF3 R3-;

Rs-NH +F2~N+lCH,

R,-NH

CF3

4a

R3-HN 0

R,-NH

CF2

CW

OCH,

“3-I:

0

5

FIG. 1.

Structures

of the peptidyl

should be noted that this extension involves the sacrifice of one fluorine atom. If extended inhibitors are to be useful, interactions at s’ subsites must more than compensate for the increase in Ki due to the loss of one fluorine atom. We also investigated the effect of structural changes on the rate of association (k,,?,,)of inhibitors with HLE and a-LP. This is of particular interest, since association rates for TFK inhibitors with serine proteases is frequently low, i.e., ~10~ M-l s-l. The reason for these slow association rates is not well understood. Finally, the effect of the number of fluorine atoms on Ki was also investigated. This problem is pertinent to the development of inhibitors which interact with the S’ subsites, since in these inhibitors one of the fluorine atoms of the parent TFK is replaced by the structure which provides interaction with the S’ subsites. MATERIALS

AND

METHODS

Enzymes. HLE was a gift of Drs. Philip Stone and W. Ruth. a-LP was a gift of Dr. Bachovchin. Substrates and inhibitors. The substrate N-methoxysuccinyl-Lalanyl-L-alanyl-L-prolyl-L-valyl-p-nitroanilide (MeoSuccAAPVpNA) was purchased from Sigma. N-Acetyl-L-alanyl-L-prolyl-L-ala-

fluoroketones.

nyl-p-nitroanilide (AcAPA-pNA) was obtained from Bachem. The inhibitors l-9 (Fig. 1) were synthesized and a detailed description of the synthesis and physical properties of these compounds is described in the Miniprint Supplement. Analytical methods. The enzymatic activity of HLE was routinely determined spectrophotometrically using the substrate MeoSuccAAPV-pNA at 412 nm (Ac = 8400 M-i cm-‘) in 50 mM Hepes, 0.5 M NaCl, pH 7.5, buffer at 25°C. ol-Lytic protease was assayed with AcAPA-pNA at 412 nm in 50 mM Tris, 0.1 M KCI, pH 8.75, buffer at 25°C. The enzyme concentrations were determined by comparison of the /z~,,values to those previously determined (8, 12). Inhibitors and substrates were dissolved in dimethyl sufoxide when necessary. The content of the organic solvent did not exceed 1% (v/v) in the assay cuvette. Absorption spectra and all spectrophotometric assays were performed in a Perkin-Elmer X-3 or 559 UV/Vis spectrophotometer thermostated at 25°C. Inhibition kinetics. In the case of reversible competitive inhibitors the Ki values were determined from Lineweaver-Burk plots or from initial velocities measured from the linear portion of absorbance vs time progress curves. Inhibitor concentrations used were 2&10X;. The K:s for the slow binding inhibitors were determined according to the method of Cha (13). Progress curves were obtained at several different concentrations of inhibitor and at substrate concentrations 0.X-0.35 mM for HLE and at 0.3-1.5 mM for n-LP. The assays were initiated by addition of enzyme. The curves were successively fit to Eq. [ l] by nonlinear least-squares analysis that provided values for ~0, v, , and kobsd.

INHIBITION

BY FLUOROKETONES

35,

/I

OF SERINE

139

PROTEASES

RESULTS

I

The structures of the peptidyl fluoroketones used in this study and their inhibitory properties toward (u-LP E and HLE are shown in Fig. 1, and Tables I and II, respec.3 20 i tively. Compounds 5-7 were synthesized from a glutarate derivative and thus are one methylene unit longer than the analogous substrates would be. This modification was adopted to avoid anticipated problems such as HF elimination from an a,a-difluoro succinate derivative. This residue basically serves as a linking unit be15 25 5 10 20 0 tween P, and PZ residues. The choice of leucine at the Time (WC) Pk position for 6 is based on the amino acid sequence for FIG. 2. Reactivation of enzyme-inhibitor complex. The enzymeeglin C, a naturally occurring elastase inhibitor of the inhibitor complex (EI) was preformed by incubating enzyme (1 PM) leech Hirudo me&An& (16). Inhibitor 7 has alanine at and inhibitor (6 @M) for -30 min. Substrate MeoSuccAAPV-pNA the P; position because k,,,/K,,, for the corresponding (470 FM) was added to the EI complex that resulted in displacement substrate indicates favorable interaction of alanine at of inhibitor and reactivation of EI complex. The solid line was generthe PZ site (17). ated by fitting the data to a model of Eq. [Z] and (A) represent the experimental points. Some of the inhibitors tested proved to be slow binding inhibitors (18). The characteristic biphasic reaction progress curve observed with the slow binding inhibitors P = u,t + (u. - udl - exp(-kOb~t))lkob~, 111 is shown for 2 with (u-LP in Fig. 3a. The replot of the data from Fig. 3a (k,bsd vs [I]) is shown in Fig. 3b. The plot is linear, as are all other replots for slow binding where P = product concentration, u0 = uninhibited velocity, u, inhibitors (data not shown). From this replot k,, and koE = steady-state velocity, and kOM is the observed first-order rate constant for the approach to the steady state. The steady-state velocities were determined (see Materials and Methods). The asmeasured and calculated in the presence and absence of inhibitor were sociation rates for two classical competitive inhibitors I

A/

I

25

used to calculate K,. k,, and kofl values were determined from a replot the association and dissociation conOf kobsd values vs [I]. In addition, stants were determined independently. For the kinetic constant maximal standard deviation is +30%. The dissociation rate constant for the slow binding inhibitor 2 with HLE was determined by monitoring the return of activity from a completely inhibited enzyme-inhibitor (EI) complex. The EI complex was preformed and passed through a Sephadex G-25 column to remove excess inhibitor. Aliquots of the isolated EI were withdrawn at different time intervals and diluted into assay cuvette containing substrate and the return of activity was measured. Control experiments demonstrate that there is no detectable loss of enzyme activity under the same conditions during that time period. The return of activity curve thus obtained was fit to an exponential expression (14,15). For classical competitive inhibitors 3 and 6 the off rate (k,J was determined by relaxation kinetics. The EI complex was preformed by addition of enzyme and inhibitor ([I] N 10 KJ such that 90-95% of enzyme was in the EI complex. Addition of substrate (MeOSuccAAPV-pNA) to the EI complex resulted in a new equilibrium. At this new equilibrium, -30% of the enzyme remained in the EI complex. The approach to equilibrium was monitored by following the time course of substrate hydrolysis at 412 nm with a fast recorder. The time course was fit by computer simulation using Eq. [2]: E +S=ES-+E+P ir EI

TABLE

Kinetic Constants for Inhibition

Compound 1” 2* 3” 4” 5” 6” 7” 8’

PI

One such fit is shown in Fig. 2. The association rate constants (k,,) for 2 and 8 with a-LP and for 2 with HLE were determined under pseudo-first-order conditions ([I] > lO[E]) from the rate of onset of slow binding inhibition. This was measured by preincubating a mixture of enzyme and inhibitor and measuring the remaining enzyme activity by addition of substrate. The activity remaining followed pseudo-first-order kinetics.

I

gb

of HLE

K (M) 2.38 (5.28 0.68 (1.5 20 (2 15 (1.5 7.5 (7.5 0.7 (7 2.7 (2.7 3 (6.66 5 (5

X X X x x x x x x x x x X X x x x x

10m3 1O-7) 10-O lo-lo) 10-a 10-a) 10-a 10-a) 10-O 10-7) 10-O lo-s) 10-O W7) 1om9 lo-la) 1o-8 10-9

5

nd

nd

1040 (4.68 X lOa) 6300 (6.3 X 104) nd

0.000713

nd

nd

(PNA)’

27,000 (PNA)

28 (amide)

ndd 46,000 (amide) 182,000 (PNA)

400 (amide)

0.124 nd

8200 (8.2 x 104) nd

0.0055

2 (9 6.8 (6.8

0.00055

x x X X

lo4 107) 10’ 105)

nd

0.0033

Note. Values for k,, and Ki parentheses have been corrected for the concentration of ketone. The correction factors for the trifluoro and difluoro compounds are 4500 and 10, respectively. 0 Classical competitive inhibitor. b Slow binding inhibitor. ‘p-Nitroanilide substrate. d Not determined.

140

GOVARDHAN TABLE Kinetic

Constants

AND

ABELES 0.800 -

II

for Inhibition

b

of a-LP 0.600--

k

kc, /Km Compound

K (MI

1”

15 (3.3 5 (1.11 62.5 (6.3 15 (1.5

2b 3” 4” 5” 6b

0.4 (4 5.5 (5.5 0.2 (4.44 2.5 (2.5

7” 8b 9”

(M-l

x 1o-3 x 10-6) x 1om6 x 10-s) X 1O-6 x 1O-6) x 1om6 x 10-r) nd’ x x x x x x X x

0.0013 (amide) 48 (amide) 48 (amide) nd

ko*

nd

0.00133 nd

nd

nd

nd

nd

1.13 x 104 (1.13 x 106) nd

2100 (amide) 15 (amide) 48 (amide)

(s-l)

nd

240 (1.08 x 106) nd

500 (amide) nd

10-a 10-s) 1om6 10-r) lomr lo-“) 10-s 10-r)

(M-‘O;-‘)

SKI)

0.000

7 0

50

100

150

I 250

200

DloN-4 FIG. 3b. Plot of kobsdvs inhibitor concentration, where kow is derived from the progress curves as described in the legend to Fig. 3a.

0.0043 nd

1 x 103 (4.5 x 106) nd

0.000166 nd

Note. Values for k,, and K, in parentheses have been corrected for the concentration of ketone. The correction factors for the trifluoro and difluoro compounds are 4500 and 10, respectively. ’ Classical competitive inhibitor. b Slow binding inhibitor. ’ Not determined.

(3 and 6) of HLE were determined by relaxation kinetics (Table I). The association rate of the classical inhibitor 3 is 75fold slower than that of the slow binding inhibitor 2 with HLE. The off rate (K,,) of the slow binding inhibitor is 170-fold slower than that of the classical inhibitor.

Since the keto form of the inhibitor is most likely the inhibitory species (19) it was important to establish the hydration state of inhibitors to calculate “true” rate constants and Ki'S. The hydrate:ketone ratio in aqueous solution was determined for the peptidyl difluoroketone 9 by “F and ‘H NMR spectroscopy (Figs. 4a and 4b, respectively). A hydrate:ketone ratio of 9:l was obtained upon integration of the spectrum. Previously, a hydrate: ketone ratio of 1:l was determined for an alkyl difluoroketone, 1,1-difluoro-6,6-dimethyl-heptan-2-one (19). The larger ratio found for 9 is attributed to the inductive effect of the amide group. The keto form of the TFKs was not observable by NMR. However, we have estimated the ratio from the known ratio for trifluoroacetone (5OO:l) and the factor of 9 due to the amide group. This yields a hydrate:ketone ratio of 45OO:l for the peptidy1 trifluoroketones. These values were used to correct rate constants and Ki's in Tables I and II.

0 0

2

4

6

8 Time

10

12

14

16

18

20

22

24

(min)

FIG. 3a. Inhibition of o(-LP by trifluoroketone 2. Progress curves of substrate hydrolysis in the presence of various concentrations of 2. The assay mixture contained 0.35 mM substrate, 10 pM enzyme, and increasing inhibitor concentrations under the standard assay conditions; the reactions were initiated by enzyme addition. The following inhibitor concentrations were used: (W) 56 pM, (V) 112 pM, (7) 167 pM, (0) 223 pM. The curves were successively fit to Eq. [l] by nonlinear least-squares analysis that provided values for k,bd, where kobsdis the first-order rate constant for approach to the steady state.

IIII,,,~~~I/,,,,~~~,/~,,,,“~‘I”“,”’~I~’~’,~”)I”“/““I’~‘~I”~’~~~ -130 -132 -134 -136

-138

-140

-142

PPM

FIG. 4a. “F NMR spectra of 9 in DzO containing 5% CD&N. The spectrum is mainly that of the hydrate species (arrows indicate observable peaks due to the keto form). The set of peaks labeled a belong to one diastereomer and those labeled b belong to the other diastereomer. Chemical shifts are reported in ppm relative to fluorotrichloromethane (0.0 ppm). Isomer a, 6 -131.3, -138.4 (AB of ABX system, doublet of ABq, *Jrr = 282.8 Hz, *Jrn = 55.1 Hz); isomer b, d -131.4, -137.3 (doublet of ABq, ‘Jw = 283.2 Hz, *JFH = 54.6 Hz).

INHIBITION

BY FLUOROKETONES

OF SERINE

141

PROTEASES Kh

=

[Ihl/tIkl

=

k&h

Ik = keto form of inhibitor b

Ih = hydrate

5% FIG. 4b. ‘H NMR of CFzH region of 9 in D,O containing CD&N. Hydrate: isomer c, 6 5.69 (t, ‘JHp = 54.9 Hz); isomer d, 6 5.77 (t, 2JHF = 55.0 Hz). Ketone: isomer e, 6 6.20 (t, ‘5”~ = 54.0 Hz); isomer f, r?6.23 (t, ‘JHF = 54.0 Hz).

DISCUSSION

Before the kinetic data are discussed, a reaction sequence for the formation of the enzyme-inhibitor complex needs to be defined. Analogy with the catalytic reaction suggests a two-step mechanism (Eq. [3]) in which a noncovalent enzyme inhibitor EI* (Michaelis complex) precedes formation of the covalent complex EI: E + I & EI* : EI -1 -2

[31

Our kinetic data, however, do not require formation of a noncovalent intermediate. In all cases where slow binding was observed, the kobsdwas linearly dependent on [I] (Fig. 3b). The failure to obtain kinetic evidence for El* could be due to the magnitude of KF (K,? = k-,/k,). We expect K: to be between 1 and 30 mM, of approximately the same magnitude as KS for amide substrates. In that case, no kinetic evidence for KF would be obtained since inhibitor concentrations used in these experiments did not exceed 200 PM. Studies with chymotrypsin and TFK inhibitors indicate that the keto form of the inhibitor reacts with the enzyme (20). A similar conclusion was reached in studies of TFK inhibitors of cholineesterase (19). Therefore, it is likely that in the reaction with HLE and (u-LP, the keto form of the inhibitor also reacts with the enzyme. The reaction sequence of Eq. [4] takes this fact into account and is a more complete description of the reaction leading to enzyme-inhibitor complex formation. Assuming that the keto form of the inhibitor is the inhibitor species, then I

E*I h? EI 2

[41

Ki and, more importantly, k,, have to be corrected for the ketone concentration. The corrected values are shown in parentheses in Tables I and II. For elastase, corrected k,, values range from 6.3 X lo4 M-l s-l (6) to 9 X lo7 Mpl s-l (8), which is nearly diffusion controlled. It is noteworthy that the k,,‘s for classical competitive inhibitors 3 and 6 are relatively slow while that for the slow binding inhibitor 8 is nearly diffusion controlled. The phenomenon designated as slow binding (14) is a consequence of experimental conditions. Since Ki’s of slow binding inhibitors are relatively low, experiments are conducted at high substrate concentrations and low inhibitor concentrations, i.e., under conditions so chosen that the fully inhibited velocity is not reached within a few seconds after the start of the reaction. In other words, conditions are chosen so that the concentration of free enzyme and that of inhibitor are low, so that the rate of formation of the enzymeinhibitor complex can be observed without the use of rapid reaction techniques. What is the reason for slow on-rates, i.e., on-rates less than diffusion controlled? With TFKs one of the contributing factors is the low concentration of the keto form of the inhibitor. However, after correction of k,, for ketone concentration, the on-rates for most of the inhibitors in Tables I and II are still less than diffusion controlled. Other effects must contribute to the low onrate. Before considering these effects it is useful to examine the nature of k,,, which is a complex constant. For the mechanism described by Eq. [3], k,, = k, k2/ (k-, + k2) at inhibitor concentrations below saturation. Two limiting cases can be considered: kz > k-l, then k,, x kl, and k2 < k, , then k,, N k,/K,, where KS = k-Jkl. On chemical grounds the addition of the active site serine to a TFK is expected to be very fast compared to the addition of serine to amide or ester substrates. However, k2 could be decreased due to unfavorable interactions of the CFB group, or due to nonproductive binding in the formation of El. The k,, for the HLE inhibitor 8 is diffusion limited or nearly so, and therefore k,, m kl. For most other inhibitors listed in Tables I and II k,, is considerably below 10’ M-’ s-l. It is of interest to compare k,, with the association rate of substrates with serine proteases. As far as we know, k,, values for substrates of HLE or (u-LP have not been determined. It has been determined for amide substrates of chymotrypsin. The association rate of N-acetylphenylalaninamide and chymotrypsin is 6 X lo7 M-l 5~’ (21) and that of N-acetyl-L-tryptophan methyl ester is 1.1 X lo7 m-l s1 (22). For many inhibitors listed in Tables I and II, k,, is considerably slower. It is therefore likely that k,, f k,. For

142

GOVARDHAN

difluoroketones ko,, is decreased 75 to 130-fold compared to the corresponding TFK (compare 8 and 9; 2 and 3). It is unlikely that substitution of the trifluoromethyl group by a difluoromethyl group reduces k, . Most likely the decrease in ko, reflects a change in k2, In summary, it is likely that for most inhibitors of Tables I and II ko,, # kl and that k2 makes a strong contribution to k,, . If that is the case, changes in kccat/Kmshould parallel changes in k,, . We do not have sufficient data to test this point. The data of inhibitory potencies (Ki) shown in Tables I and II show qualitatively that structural changes affect Ki for the two enzymes similarly: (i) Replacement of an alanine residue by a valine residue at the P, position decreases Ki. With HLE, the effect is very pronounced. The Ki of the valyl compounds 8 and 9 are 226- and 400fold lower than the corresponding alanyl compounds 2 and 3. This trend has been observed in the k&K,,, of pnitroanilide substrates (8) and in the second-order rate constant of inhibition of HLE by chloromethylketones (23) with increases of 7- and 50-fold, respectively. With (r-LP, Ki decreases 25fold on replacement of an alanyl residue in P1 by valine. cr-LP differs from HLE in that (u-LP exhibits a 2-fold preference for an alanine residue over valine in the P1 position of its substrates (24). The reversed specificity observed here was also seen with the boronic acid inhibitors (12). (ii) On extending the inhibitor 1, which consists of only a P1 component in the P direction, Ki decreases markedly with both enzymes. Comparison of 1 and 2 shows that extension of 1 in the P direction by three peptide residues decreases Ki 3500-fold for HLE. Addition of a dipeptide residue to 3 to extend in the P’ direction decreases Ki for HLE 30-fold. The lesser effect seen in the P’ direction is consistent with the observation of Stein et al. (9) that there might not be important interactions past S;, which is hydrophobic. To assess the importance of the hydrophobic pocket in the S; position we tested the fluoroketone 4. The similarity of its Ki value to the difluoromethyl analog without the P; substituent (3) indicates that structural features other than hydrophobicity may be important. With LU-LP, extension by three residues in the P direction lowers Ki 3OOO-fold. Extending the inhibitor in the P’ direction to include interactions up to P; produces a 150-fold decrease in Ki (compare 2 and 6, Table II). Extension in the P’ direction lowers Ki in comparison to the parent difluoromethylketone. For practical considerations, i.e., development of an inhibitor with the lowest possible Ki, comparison should be made to the trifluroketone (compare 2 with 6 and 7). For HLE the Ki of the extended inhibitor (6) is approximately equal to that of the trifluoroketone 2. If the goal is an inhibitor with minimal Ki then the modification which we introduced was not useful. Possibly, a more potent inhibitor can be obtained by introduction of an amino acid other than leucine. For (u-LP, the Ki of the extended inhibitor 6 is IO-fold lower than that of 3.

AND

ABELES

In this case introduction of a P’ residue resulted in an improved inhibitor. The effect of chain length on Ki can be compared to the effect of similar structural alterations of the corresponding substrates k&K,,, (25). With HLE and p-nitroanilide substrates, k,,JK,,, increases 103-fold upon extending a substrate from Pi to P, (8). On extending the tetrapeptide substrate corresponding to 3 in the P’ direction to the hexapeptide substrate corresponding to 7, kcat/Km increases 115-fold (17). Thus the effect of peptide chain extension on kJK, parallels its effect on Ki. Such a correlation was observed previously by Stein et al. (3) with a series of trifluoroketones. With a-LP, similar effects of peptide chain length are observed (24). These results establish that extended binding interactions can be used to increase the efficacy and, possibly, the specificity of inhibition as was observed with chymotrypsin (10). (iii) Increasing the number of fluorine atoms from two to three decreases Ki approximately 15- to 30-fold for both HLE and a-LP. An effect of similar magnitude was found for chymotrypsin (lo), acetylcholinesterase, and pseudocholinesterase (19). We attribute this effect to stabilization of the hemiketal alkoxide which results from the addition of a serine-OH group to the carbonyl group of the inhibitor (6).

REFERENCES 1.

Imperiali, B., and Abeles, R. H. (1986) Biochenistry

25, 3760-

3767. 2. Dunlap, R. P., Stone, P. J., and Abeles, R. H. (1987) &o&em. Biophys. Res. Commun. 145,509-X3. 3. Stein, R. L., Strimpler, A. M., Edwards, P. D., Lewis, J. J., Manger, R. C., Schwartz, J. A., Stein, M. M., Trainor, D. A., Wildonger, R. A., and Zottola, M. (1987) Biochemistry 26,2682-2689. 4. Westerik, J. O., and Wolfenden, R. (1972) J. Biol. Chem. 247, 8195. 5. Takahashi, L. H., Radhakrishnan, R., Rosenfeld, R. E., Meyer, F. E., Trainor, D. A., and Stein, M. (1988) J. Mol. Biol. 201,423428. 6. Brady, K. D., Ringe, D., and Abeles, R. H. (1990) Biochem., in press. 7. Liang, T. C., and Abeles, R. H. (1987) Biochemistry 26, 76037608. 8. Stein, R. L., Strimpler, A. M., Hori, H., and Powers, J. C. (1987) Biochemistry 26,1301-1305. 9. Stein, R. L., and Strimpler, A. M. (1987) Biochemistry 26, 22382242. 10. Imperiali, B., and Abeles, R. H. (1987) Biochemistry 26, 44744477. 11. Lentini, A., Farchiare, F., Ternai, B., Krena-Ongarjnukool, N., and Tovivick, Phichai (1987) Biol. Chem. Hoppe-Seyler 368,369378. 12. Kettner, C. A., Bone, R., Agard, D. A., and Bachovchin, W. (1988) Biochemistry 27,7682-7688. 13. Cha, S. (1975) Biochem. Pharmacol. 24,2177-2185. 14. Morrison, J. F. (1982) Trends Biochem. Sci. (Pers. Ed.) 7, 102105.

INHIBITION

BY FLUOROKETONES

15. Williams, J. W., and Morrison, J. F. (1979) in Methods in Enzymology (Purich, D. L., Ed.), Vol. 63, pp. 437-467, Academic Press, San Diego. 16. Seemuller, U., Eulitz, M., Fritz, H., and Strobl, A. (1980) HoppeSeyler’s 2. Physiol. Chem. 361,1841-1846. 17. McRae, B., Makajima, K., Travis, J., and Powers, J. C. (1980) Biochemistry 19,3973-3978. 18. Morrison, J. F., and Walsh, C. T. (1988) in Advances in Enzymology (Meister, A., Ed.), Vol. 61, p. 202, Wiley, New York. 19. Allen, 8473.

K. N., and Abeles, R. H. (1989) Biochemistry

20. Brady, K. D. (1989) Ph.D. thesis, Brandeis

28, 8466-

University.

21. Hess, G. P. (1971) in Enzymes (Boyer, P. D., Ed.), 3rd ed., Vol. 3, pp. 218-219, Academic Press, San Diego. 22. Brouwer, 1307.

A. C., and Kirsch,

J. F. (1982) Biochemistry

21, 1302-

OF SERINE

PROTEASES

143

23. Powers, J. C., Guptin, B. F., Dale Harley, A., Nishino, N., and Whitley, R. J. (1977) Biochim. Biophys. Acta 485,156-166. 24. Bauer, C. A., Brayer, G. D., Sielecki, A. R., and James, M. N. G. (1981) Eur. J. Biochem. 120,289-294. 25. Bartlett, P. A., and Marlowe, C. K. (1983) Biochem. 22, 18411846. 26. Schechter, I., and Berger, A. (1967) Biochem. Biophys. Res. Commm. 27,157-162. 27. Bolognesi, M., Gatti, G., Menegatti, E., Guarneri, M., Marguart, M., Papamokos, E., and Haber, R. (1982) J. Mol. Biol. 162,839868. 28. Delbaere, L. T. J., Brayer, G. D., and James, M. N. G. (1979) Nature (London) 279,165-168. 29. Laskowski, M., Jr., Tashiro, M., Empie, M. W., Park, S.-J., Kato, I., Ardelt, W., Wieczorek, M. (1983) in Proteinase Inhibitors (Katunama, N., Umezawa, H., and Holzer, H., Eds.), pp. 55-68, Springer-Verlag, Berlin. 30. Read, R. J., Fuginaga, Y., Sielecki, A. R., and James, M. N. G. (1983) Biochemistry 22,4420-4433.

144

GOVARDHAN MINIPRINT

AND

ABELES

SUPPLEMENT

INHIBITION

BY FLUOROKETONES

OF SERINE

PROTEASES

145

146

GOVARDHAN

AND

ABELES

Structure-activity studies of fluoroketone inhibitors of alpha-lytic protease and human leukocyte elastase.

We have synthesized a series of peptidyl fluoroketones that reversibly inhibit the serine proteases human leukocyte elastase (HLE) and alpha-lytic pro...
908KB Sizes 0 Downloads 0 Views