Linear Competitive Inhibition of Human Tissue Kallikrein by 4-Aminobenzamidine and Benzamidine and Linear Mixed Inhibition by 4-Nitroaniline and Aniline

Home , Competitive inhibition, Mixed inhibition

Brazilian Journal of Medical and Biological Research (2001) 34: 35-44 Human tissue kallikrein inhibition by amidines and anilines 35 ISSN 0100-879X

Linear competitive inhibition of human tissue kallikrein by 4-aminobenzamidine and benzamidine and linear mixed inhibition by 4-nitroaniline and aniline

M.O. Sousa1, Departamentos de 1Análises Clínicas e Toxicológicas, Faculdade de Farmácia, T.L.S. Miranda2, E.B. Costa1, 2Engenharia Química, Escola de Engenharia, and 3Bioquímica e Imunologia, E.R. Bittar3, M.M. Santoro3 Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais, and A.F.S. Figueiredo1 Belo Horizonte, MG, Brasil

Abstract

Correspondence Hydrolysis of D-valyl-L-leucyl-L-arginine p-nitroanilide (7.5-90.0 Key words A.F.S. Figueiredo µM) by human tissue kallikrein (hK1) (4.58-5.27 nM) at pH 9.0 and · Kinetics of human tissue Departamento de Análises Clínicas 37oC was studied in the absence and in the presence of increasing kallikrein inhibition e Toxicológicas · concentrations of 4-aminobenzamidine (96-576 µM), benzamidine Tissue kallikrein Faculdade de Farmácia, UFMG · 4-Nitroaniline (1.27-7.62 mM), 4-nitroaniline (16.5-66 µM) and aniline (20-50 mM). Caixa Postal 689 · Aniline 30123-970 Belo Horizonte, MG The kinetic parameters determined in the absence of inhibitors were: · Benzamidine -1 Brasil Km = 12.0 ± 0.8 µM and kcat = 48.4 ± 1.0 min . The data indicate that · 4-Aminobenzamidine Fax: +55-31-339-7666 the inhibition of hK1 by 4-aminobenzamidine and benzamidine is · Enzyme inhibition E-mail: [email protected] linear competitive, while the inhibition by 4-nitroaniline and aniline is linear mixed, with the inhibitor being able to bind both to the free Presented at the XXVIII Annual enzyme with a dissociation constant K yielding an EI complex, and to Meeting of the Brazilian Society i of Biochemistry and Molecular the ES complex with a dissociation constant Ki, yielding an ESI Biology, Caxambu, MG, Brazil, complex. The calculated Ki values for 4-aminobenzamidine, benzami- May 22-25, 1999. dine, 4-nitroaniline and aniline were 146 ± 10, 1,098 ± 91, 38.6 ± 5.2 and 37,340 ± 5,400 µM, respectively. The calculated Ki values for 4- This work is part of a PhD thesis to be nitroaniline and aniline were 289.3 ± 92.8 and 310,500 ± 38,600 µM, presented by M.O. Sousa to the respectively. The fact that Ki>Ki indicates that 4-nitroaniline and Graduate course in Biochemistry and Immunology, ICB, UFMG. aniline bind to a second binding site in the enzyme with lower affinity than they bind to the active site. The data about the inhibition of hK1 Research supported by FAPEMIG by 4-aminobenzamidine and benzamidine help to explain previous (No. CBS 1173/95) and CNPq observations that esters, anilides or chloromethyl ketone derivatives a (No. 521259/95-9). T.S. Miranda of N -substituted arginine are more sensitive substrates or inhibitors was the recipient of a Recent of hK1 than the corresponding lysine compounds. Doctor fellowship from CNPq (No. 300716/95-8). E.B. Costa was the recipient of a Scientific Initiation fellowship from CNPq (No. 521259/95-9). E.R. Bittar is Introduction likreins. The two groups differ in molecular the recipient of a Graduate weight, isoelectric point, substrate specific- Student fellowship from CNPq. The kallikreins (EC 3.4.21.8) are serine ity, immunological characteristics, the type proteases present in glandular cells, neutro- of kinin released, and functional importance (1). Tissue kallikreins are related to trypsin, Received November 26, 1999 phils and biological fluids. They are divided Accepted October 11, 2000 into two main groups, i.e., plasma (EC but with a higher specificity for the cleavage 3.4.21.34) and tissue (EC 3.4.21.35) kal- site of polypeptide substrates (2). Their prin-

Braz J Med Biol Res 34(1) 2001 36 M.O. Sousa et al.

cipal known biological function is the highly hK1-catalyzed hydrolysis of Cbz-Tyr-OpNP selective hydrolysis of plasma high and low with a Ki value of 0.2 mM (10). However, molecular mass kininogens at two different there are no reports concerning the inhibi- peptide bonds (Met379-Lys380 and Arg389- tion of the amidase activity of hK1 by 4- Ser390), with residues numbered on the basis aminobenzamidine (4-ABzA), BzA, aniline of the structure of prekallikrein (3), to stoi- (An) and 4-nitroaniline (4-NAn). chiometrically release the vasoactive and The purpose of the present study was to spasmogenic decapeptide kallidin (Lys- examine in depth the kinetics of the inhibi- bradykinin) (2). Human tissue kallikrein tion of hK1 amidase activity by 4-ABzA, (hK1) (4) also hydrolyzes various synthetic BzA, 4-NAn and An in order to identify the a substrates such as N -substituted arginine precise mechanism of inhibition and to de- and lysine derivatives (amides, esters and termine the number of binding sites and their fluorogenic peptides) (5-8), Cbz-Tyr-OpNP accurate inhibition constants (Ki). (9,10) and D-Pro-Phe-Phe-Nan (8). Like other Our results clearly demonstrate that serine proteinases, hK1 is inhibited by diiso- inhibition of hK1 amidase activity by 4- propylfluorophosphate, chloromethyl ke- ABzA and BzA is a linear competitive inhi- tones of arginine and lysine, and basic pan- bition with only one inhibitor molecule bind- creatic trypsin inhibitor (BPTI) (also known ing to one enzyme molecule, forming a bi- as aprotinin, Trasylol® or Kunitz pancreatic nary enzymatic complex EI. On the other trypsin inhibitor) (6,11). In contrast, soy- hand, our results also demonstrate that bean trypsin inhibitor, a potent inhibitor of inhibition of hK1 amidase activity by An trypsin, plasma kallikrein and other serine and 4-NAn is of the linear mixed type, with proteinases, does not inhibit hK1 (6). the inhibitors being able to bind both to the The inhibition of hK1 by BPTI is not a free enzyme to yield a binary enzymatic simple competitive inhibition as first reported complex EI, and to the ES complex to yield (6,10), but is competitive inhibition of the a non-reactive ternary enzymatic complex parabolic type, with two inhibitor molecules ESI. binding to one enzyme molecule, forming a ternary enzymatic complex (12). It is note- Material and Methods worthy that the second BPTI molecule binds to the enzyme with higher affinity, suggest- Material ing that this second binding site was prob- ably created or positively modulated as a An, 4-NAn, 4-ABzA·2HCl, BzA·HCl and consequence of the binding of the first BPTI BPTI were purchased from Sigma Chemical molecule (12). However, the nature of the Co. (St. Louis, MO, USA). D-Valyl-L-leucyl- second binding site for this inhibitor in hK1 L-arginine 4-nitroanilide (D-Val-Leu-Arg- is still unknown (12). Nan) was obtained from Chromogenix AB, There are controversial reports in the (Mölndal, Sweden). literature about the inhibition of hK1 es- All other reagents were of reagent grade terase activity by benzamidine (BzA). BzA from Sigma. has been reported to be a competitive inhibitor of the hK1-catalyzed hydrolysis of Tos- Enzyme Arg-OMe with a Ki value of 6.42 mM, but does not inhibit the hK1-catalyzed hydroly- Homogeneous hK1 was purified in our sis of Cbz-Tyr-OpNP even at a level of 2 mM laboratory from healthy male urine (12). (9). On the other hand, BzA has been re- BPTI-titrated hK1 was used in the enzymatic ported to be a competitive inhibitor of the assays (12).

Braz J Med Biol Res 34(1) 2001 Human tissue kallikrein inhibition by amidines and anilines 37

Enzyme assay by 4-ABzA and BzA and the kinetic data of Bß-TR inhibition by An can be described, Kallikrein amidase activity was assayed according to Plowman (14), as competitive spectrophotometrically (6) at 410 nm to moni- inhibition. e -1 -1 tor the release of 4-NAn ( 410 = 8800 M cm ). On the other hand, the kinetic data of the A 1-cm path-length cuvette containing 50 µl inhibition of hK1 by 4-NAn and An can be of 45.8-52.7 nM hK1 (Mr 31,000) (12) in 200 described by the following scheme: mM glycine/NaOH buffer, pH 9.0, containing @ Ks Km kcat 0.05% (w/v) NaN3 and 100 µl of buffer or 100 E+S ES E+P µl of an adequate dilution of a stock inhibitor solution (10 mM 4-ABzA, 100 mM BzA, 1.66 mM 4-NAn and 100-300 mM An) in the same I Ki I Ki buffer was placed in the thermostated cell compartment of a Shimadzu UV 160 A spec- EI ESI trophotometer at 37oC and pre-incubated for 5 e min. The concentrations of 4-ABzA ( 292 = which, according to Cornish-Bowden (15), -1 -1 e -1 15,280 M cm ) (13), BzA ( 225 = 9390 M is the simplest formal mechanism for mixed cm-1) (13) and 4-NAn solutions were deter- inhibition. According to the author, the in- mined spectrophotometrically. Recently dis- hibitor (I) can bind both to the free enzyme tilled aniline (d = 1.022 g/cm3) was used. The (E) to give a complex EI with dissociation ranges of inhibitor concentrations were se- constant Ki, and to the ES complex to give an lected in order to obtain 20-80% inhibition. unreactive ESI complex with dissociation Then, 350 µl (10.7-128.6 µM) of D-Val-Leu- constant Ki. As shown in this scheme, both Arg-Nan in 200 mM glycine/NaOH buffer, inhibitor-binding reactions are dead-end re- pH 9.0, containing 0.05% (w/v) NaN3 was actions and are therefore equilibria. As both added, and the increase in absorbance at 410 EI and ESI exist, however, it is difficult to nm with time was continuously recorded for 3 see why S should not bind directly to EI to min. The slope of the time-dependent absorbance give ESI. If this reaction is included in the curve extrapolated to zero time was converted mechanism the rate equation becomes much to µM of released 4-NAn per minute. more complicated, because terms in S2 and I2 Bovine ß-trypsin (Bß-TR) (2.91 nM) was appear in it. These terms cancel only if all of also assayed spectrophotometrically at 410 the binding reactions are equilibria, i.e., if nm with the substrate D-Val-Leu-Arg-Nan the substrate- and product-release steps are (200-600 µM) in 100 mM Tris-HCl buffer, pH all fast compared with the reaction that con- 8.1, containing 20 mM CaCl2 and 0.05% (w/v) verts ES into products. In practice, however, NaN3, in the absence and in the presence of An the predicted deviations from simple kinet- (10-40 mM) in order to check for inhibition. ics are difficult to detect experimentally, and Total substrate concentration was determined one cannot use adherence to simple kinetics from the amount of 4-NAn released after com- as evidence that Km, Ki and Ki are true plete hydrolysis by excess Bß-TR. In the inhi- dissociation constants. bition assays with 4-NAn the reference cu- The normalized initial rate (v) will be vette contained the same amount of 4-NAn as given by the following equation: the sample cuvette. Eq. 1 Treatment of kinetic data vi kcat.[S] = v = The kinetic data of the inhibition of hK1 [Eo] Km (1 + [I]/Ki) + [S] (1 + [I]/Ki)

Braz J Med Biol Res 34(1) 2001 38 M.O. Sousa et al.

According to Cornish-Bowden (15), when to the appropriate Michaelis-Menten equa- app linear mixed inhibition occurs, both kcat , and tions, while the Ki values for 4-NAn and An app app kcat /Km vary with the inhibitor concentra- were determined by fitting the experimental tion according to the following equations: data to Equation 1. Whenever used, the linear replots (Equations 3 and 6) were shown kcat app only for diagnostic purposes; they were not kcat = Eq. 2 used to calculate the Ki and Ki values. 1 + [I]/Ki

which can be rearranged to Results

app 1/kcat = 1/kcat + [I]/(kcat.Ki) Eq. 3 The hK1-catalyzed hydrolysis of D-Val- Leu-Arg-Nan followed Michaelis-Menten Additionally, kinetics under the assayed substrate concentration range (7.5-90.0 µM). Figure 1 shows Km (1 + [I]/Ki) the 1/v vs 1/[S] plot for the hydrolysis of D- K app = Eq. 4 m Val-Leu-Arg-Nan (7.5-90.0 µM) catalyzed 1 + [I]/Ki by hK1 (4.58 nM) in the absence and in the and presence of 4-ABzA (1.27-7.62 mM). The

kcat/Km inset shows the replot of the slopes of the app app app app kcat /Km = Eq. 5 lines from Figure 1 (Km /kcat vs 4-ABzA 1 + [I]/Ki concentration) according to Plowman (14). Each point is the mean of 4 determinations. which can be rearranged to Similar results were obtained with BzA. Eq. 6 Figure 2 shows the 1/v vs 1/[S] plot for the app app Km /kcat = Km/kcat + [Km/(kcat.Ki)].[I] hydrolysis of D-Val-Leu-Arg-Nan (7.5-90.0 µM) catalyzed by hK1 (5.27 nM) in the ab- app The kinetic parameters (Km, Km , kcat, sence and in the presence of 4-NAn (16.5-66 app kcat and Ki for 4-ABzA, BzA, 4-NAn and µM). Each point is the mean of 4 determina- An) were calculated based on unweighted tions. Similar results were obtained with An. app app non-linear regression analysis of the data fit Figure 3 shows the replots of Km /kcat app (panel A) and 1/kcat (panel B) vs 4-NAn Figure 1 - Lineweaver-Burk plot for the hydrolysis of D-Val-Leu- 0.25 concentration according to Cornish-Bowden Arg-Nan by hK1 in the absence 1.2 (15) (Equations 6 and 3, respectively). The and in the presence of 4-ABzA. straight lines obtained are consistent with

app app (µM min) 0.8 Inset, Km /kcat vs [4-ABzA]. 0.20 linear mixed inhibition. Statistical analysis Experimental conditions: 200 app

cat 0.4 mM glycine/NaOH, pH 9.0, k of these data using the GraphPad program at / o

37 C, 5-min incubation. hK1 con- app 0.15 0.0 the 95% confidence level showed that both centration: 4.58 nM. 4-ABzA m K 0.00 0.20 0.40 0.60 lines have slopes that are significantly dif- concentrations: closed circles, 0 [4-ABzA] (mM) ferent from zero, and the statistical test for mM; open circles, 0.096 mM; 1/v (min)

0.10 ○○○○○○○○ closed triangles, 0.192 mM; departure from linearity gave a negative (non- open triangles, 0.384 mM, and significant) result. Similar results were ob- closed squares, 0.576 mM. Each point in the plot is the mean of 0.05 tained with An. The kinetic parameters for quadruplicate determinations. hK1-catalyzed hydrolysis of D-Val-Leu-Arg- More details are described in Nan in the absence and in the presence of 4- Material and Methods. 0.00 -0.20 -0.10 0.00 0.10 0.20 ABzA, BzA, 4-NAn and An are shown in 1/[D-Val-Leu-Arg-Nan] (µM-1) Table 1. The Bß-TR-catalyzed hydrolysis of D-Val-

Braz J Med Biol Res 34(1) 2001 Human tissue kallikrein inhibition by amidines and anilines 39

Leu-Arg-Nan followed Michaelis-Menten ki- tural relationship to the side chain of the netics in the substrate concentration range tyrosine constituent of the synthetic sub- assayed (200-600 µM) (data not shown). strate CBz-Tyr-OpNP, which was demon- The Michaelis-Menten plot for the hy- strated to be the hK1 substrate (9,10). On the drolysis of D-Val-Leu-Arg-Nan (200-600 other hand, 4-NAn was chosen not only µM) catalyzed by Bß-TR (2.91 nM) in the because of its structural relationship to An, absence and in the presence of An (10-40 but also because it shows a larger dipole mM) showed enzyme inhibition (data not moment than the dipole moment of An (16); shown). The double-reciprocal plot for the it is also one of the products of the hydrolysis An data showed convergent lines crossing at of the chosen substrate. the same point on the 1/v axis, indicating competitive inhibition (data not shown). The hK1 inhibition by 4-ABzA and BzA replot of the slopes of the lines obtained app app from the double-reciprocal plot (Km /kcat ) The double-reciprocal plot indicated that vs An concentration according to Plowman Figure 2 - Lineweaver-Burk plot (14) showed a straight line with r2 = 0.952 0.07 ○○○○○○○○○○○○○○○○○ for the hydrolysis of D-Val-Leu- Arg-Nan by hK1 in the absence (data not shown), indicating that An is a 0.06 and in the presence of 4-NAn. linear competitive inhibitor of Bß-TR ami- Experimental conditions: 200 0.05 dase activity in the concentration range tested. mM glycine/NaOH, pH 9.0, 37oC, 3-min incubation. hK1 No hint of a possible second binding site for 0.04 concentration: 4.58 nM. 4-NAn An can be discerned in the experimental data concentrations: closed circles, 0

1/v (min) 0.03 obtained. The kinetic parameters (Km, kcat, µM; open circles, 16.5 µM; closed triangles, 33.0 µM; open kcat/Km and Ki) for the Bß-TR-catalyzed hy- 0.02 triangles, 50.0 µM, and closed drolysis of D-Val-Leu-Arg-Nan in the ab- 0.01 squares, 66.0 µM. Each point in sence and in the presence of An, determined the plot is the mean of quadru- according to Plowman (14), were 191.7 ± 0.00 plicate determinations. More 0.00 0.02 0.04 0.06 0.08 -1 -1 -0.02 details are described in Material 38.2 µM, 1604 ± 107 min , 8.4 ± 1.8 min 1/[D-Val-Leu-Arg-Nan] (µM-1) and Methods. µM-1 and 10.8 ± 1.0 mM, respectively.

app Discussion Figure 3 - Replots of Km / 0.800 app app A kcat (panel A) and of 1/kcat (panel B) against 4-nitroaniline During the characterization of our hK1 0.600 concentration. Regression coef- ficients: for panel A - intercept: preparation, we decided to check the inhibi- (µM min) 0.400 0.29000 ± 0.01405; slope: tion of its amidase activity by 4-ABzA (pKa2 app 0.006152 ± 0.000347; r2: cat k = pKa of the amidinium group = 12.39), BzA / 0.9906; for panel B - intercept:

app 0.200 0.01898 ± 0.00057; slope: (pKa = 11.41), 4-NAn (pKa = 1.0) (16) and m K 0.00009593 ± 0.00001406; r2: An (pKa = 4.60) (16). At pH 9.0, the opti- 0.000 0.9395. More details are de- mum pH for the hK1-catalyzed hydrolysis of B scribed in Material and Meth- ods. D-Val-Leu-Arg-Nan, 4-ABzA and BzA 0.025 showed a positive charge, while 4-NAn and An had no apparent electric charge, but di- (min) app

cat 0.020 pole moments (µ) of 6.10 D and 1.53 D (16), k respectively. 4-ABzA and BzA were chosen 1/ because they are good models of the side 0.015 chains of arginine and lysine (17) in the 010203040506070 usual human tissue kallikrein substrates (6,7). [4-Nitroaniline] (µM) Aniline was also chosen because of its struc-

Braz J Med Biol Res 34(1) 2001 40 M.O. Sousa et al.

Table 1 - Kinetic parameters of human tissue kallikrein.

a200 mM Glycine/NaOH, pH 9.0, at 37oC; enzyme concentration: b4.58 nM, c5.27 nM.

app app app app Experimental Km kcat kcat /Km Ki Ki’ conditions mM (µM) (min-1) (min-1 µM-1) (µM) (µM)

Buffera 0.00 12.0 ± 0.8 48.4 ± 1.0 4.03 ± 0.43 --

4-ABzAb 0.096 18.6 ± 0.6 46.7 ± 0.5 2.51 ± 0.08 146 ± 10 - 0.192 24.3 ± 1.5 45.2 ± 1.0 1.86 ± 0.12 0.384 41.3 ± 4.6 47.9 ± 2.4 1.16 ± 0.14 0.576 46.0 ± 3.6 44.1 ± 1.6 0.96 ± 0.08

BzAb 1.27 24.6 ± 1.6 47.1 ± 1.1 1.91 ± 0.05 1,098 ± 91 - 2.54 40.0 ± 3.6 46.9 ± 1.9 1.17 ± 0.11 5.08 62.3 ± 3.3 47.1 ± 1.3 0.76 ± 0.04 7.62 72.3 ± 4.0 43.2 ± 1.3 0.60 ± 0.04

4-NAnc 0.0165 20.4 ± 2.1 49.3 ± 1.7 2.42 ± 0.26 38.6 ± 5.2 289.3 ± 92.8 0.0330 22.8 ± 1.4 45.7 ± 1.0 2.04 ± 0.20 0.0500 26.0 ± 1.4 43.4 ± 0.9 1.67 ± 0.10 0.0660 26.3 ± 0.8 38.3 ± 0.5 1.46 ± 0.05

Anb 20.0 18.0 ± 0.9 45.8 ± 0.8 2.54 ± 0.13 37,340 ± 5,400 310,500 ± 38,600 30.0 20.6 ± 0.5 43.3 ± 0.3 2.10 ± 0.05 40.0 24.1 ± 0.9 41.8 ± 1.1 1.73 ± 0.13 50.0 24.6 ± 0.9 40.2 ± 0.6 1.63 ± 0.06

the hK1-catalyzed hydrolysis of D-Val-Leu- The data in Table 1 are consistent with com- Arg-Nan is competitively inhibited by 4- petitive inhibition since in the presence of 4- app ABzA (Figure 1). According to Plowman ABzA and BzA the Km values increase (14), competitive inhibition can be linear, with inhibitor concentration, although kcat hyperbolic, or parabolic. Thus, in order to values remain approximately constant (15). distinguish the various types of competitive The present results are consistent with the inhibition, it is necessary to replot the slopes model in which hK1 and 4-ABzA or BzA of the lines obtained from the double-recip- can form enzyme-inhibitor complexes with a rocal plot vs inhibitor concentration. The stoichiometry of 1:1. No hint of a possible result will be linear, hyperbolic or parabolic second binding site for these inhibitors can curves, distinguishing the various types of be discerned in the experimental data ob- inhibition referred to as linear, hyperbolic or tained. Similar results were reported for the parabolic competitive inhibition, respec- 4-ABzA and BzA inhibitions of the amidase tively. In order to clarify the type of inhibi- activity of trypsin, a well-known serine pro- tion of hK1 by 4-ABzA, it was decided to teinase, which showed that 4-ABzA and BzA replot the slopes of the lines obtained from are potent competitive inhibitors with Ki val- the double-reciprocal plot (Figure 1) vs 4- ues of 8.25 µM and 18.4 µM, respectively ABzA concentration. The results obtained (17). showed a linear curve (Figure 1, inset), indi- Comparison of the Ki values for 4-ABzA cating that 4-ABzA is a linear competitive (pKa2 = 12.39) (146 ± 10 µM) and BzA (pKa inhibitor (14) of hK1 amidase activity in the = 11.41) (1,098 ± 91 µM) reveals that 4- concentration range tested. Similar results ABzA is a 7.5-fold more potent hK1 inhibi- were obtained with BzA (data not shown). tor than BzA.

Braz J Med Biol Res 34(1) 2001 Human tissue kallikrein inhibition by amidines and anilines 41

Comparison of the Ki values for hK1 hK1 inhibition by 4-NAn and An inhibition by 4-ABzA (146 ± 10 µM) and BzA (1098 ± 91 µM) with the Ki for trypsin The Michaelis-Menten plot for the hy- inhibition by 4-ABzA (8.25 µM) and BzA drolysis of D-Val-Leu-Arg-Nan (7.5-90 µM) (18.4 µM) (17), respectively, reveals that 4- catalyzed by hK1 (4.58-5.27 nM) in the ab- ABzA and BzA bind much more weakly to sence and in the presence of 4-NAn (pKa = hK1 than to trypsin. These results agree with 1.00) (16.5-66 µM) and An (pKa = 4.60) (20- a published report stating that BzA binds 50 mM) showed enzyme inhibition (data not much more weakly to hK1 (Ki ~15 mM) and shown). The double-reciprocal plot for the pK1 (Ki ~1 mM) than to trypsin (Ki = 20 µM) 4-NAn data (Figure 2) showed convergent (18). lines crossing approximately at the same The present results can explain previous point in the second quadrant, indicating lin- observations regarding the hK1 substrate ear mixed inhibition (15). Similar results specificity, which indicate that arginine es- were obtained with An. The data for 4-NAn ter or anilide derivatives are more sensitive and An were also analyzed by the Dixon plot substrates for the enzyme than the corre- (1/v vs [I]) and by the Cornish-Bowden plot sponding lysine compounds (19). Similarly, ([S]/v vs [I]) (15), respectively. The straight our data explain previous observations about lines obtained from the Dixon plots inter- the reactivity of arginine and lysine sected approximately at the same point in the chloromethyl ketones in inactivating hK1 second quadrant, while the straight lines ob- which revealed that the enzyme was 10-fold tained from the Cornish-Bowden plots inter- more reactive with the Arg chloromethyl sected approximately at the same point in the ketones than with the Lys ones (11). Thus, third quadrant (data not shown). These re- since the pKa value of the guanidinium group sults also indicate linear mixed inhibition localized on the side chain of Arg is 12.5 and (15). According to Segel (25), the Dixon the pKa value of the amino group localized plots for partial and most mixed-type inhibi- on the side chain of Lys is 10.0, it is easy to tion systems are curved. However, when the explain why hK1 has a significant prefer- ESI complex is not catalytically active, the ence for Arg over Lys residues at the P1 plot is linear. The data for 4-NAn and An in position (20) of their substrates (21). As Table 1 are also not consistent with competi- app app previously reported, the stability of ion pairs tive inhibition. As both Km and kcat vary increases with the difference in pKa of the with inhibitor concentration, linear mixed groups involved (22). In this way, ion pairs inhibition is suggested (15). formed by a given anion (for instance, car- Thus, in order to further clarify the inhi- boxylate) with Arg will be more stable than bition type, we decided to replot the values app app app those formed by Lys (22). There is evidence of Km /kcat and 1/kcat vs [I], respec- that this is true in many circumstances of tively, according to Equations 6 and 3, re- biological interests (23,24). The interactions spectively (Figure 3, panels A and B, shows between hK1 and 4-ABzA and BzA seem to the results obtained with 4-NAn). Similar app follow the same rule. results were obtained with An. As both Km app Comparison of the 4-ABzA and BzA inhi- and kcat vary with [I], linear mixed inhibi- bition of the amidase activities of human tis- tion was indicated (15). The Ki values for 4- sue kallikrein (this work) and trypsin (17) NAn (38.6 ± 5.2 µM) and for An (37,340 ± reveals that with these small molecules the 5,400 µM) and the Ki values for 4-NAn inhibition mechanism of these two serine pro- (289.3 ± 92.8 µM) and for An (310,500 ± teinases is similar - both enzymes are inhibited 38,600 µM), respectively, were calculated by a linear competitive mechanism. according to the fit of Equation 1 into the

Braz J Med Biol Res 34(1) 2001 42 M.O. Sousa et al.

corresponding data in a Michaelis-Menten able to interact with a site in the enzyme. plot. According to the authors, the hydroxyl group 183 Comparison of the Ki values for hK1 of the reactive Ser is the most probable inhibition by 4-NAn (pKa = 1.00) (38.6 ± 5.2 candidate for the dipole of the enzyme that µM) and 4-ABzA (pKa2 = 12.39) (146 ± 10 interacts with the dipole at the para position µM), respectively, reveals that 4-NAn binds in substituted benzamidines. On the other 3.8-fold more strongly to the active center of hand, electron-withdrawing groups generate hK1 than 4-ABzA. Since at pH 9.0 the a dipole of opposite polarity that decreases amidinium group of 4-ABzA is bearing a full binding by a dipole-dipole repulsion. positive charge, while the amino group of 4- We do not know which group is a pos- NAn shows a positive charge induced by sible candidate for the dipole of the enzyme intramolecular transfer of electrons from the that interacts with the dipole in 4-NAn. amino group to the nitro group by isovalent Comparison of the Ki values for 4-NAn resonance (16,26), it would be reasonable to (38.6 ± 5.2 µM) and for An (37,340 ± 5,400 expect that 4-ABzA would interact better µM) reveals that 4-NAn binds 967-fold more with the S1 subsite of the active center of strongly to the active center of hK1 than An. hK1 (20) than 4-NAn. However, the data This result is partially consistent with the obtained do not concur with this reasoning. larger dipole moment of 4-NAn (6.10 D) As a speculation, we may assume that 4- over An (1.53 D) (16). However, the larger NAn, after binding to the S1 subsite of the dipole moment of 4-NAn over An is not active center of hK1 (20), possibly could sufficient to explain the data obtained. As a participate in an additional enzyme-inhibitor speculation, we may state that, after binding interaction of the dipole-dipole type, involv- to the S1 subsite of the active center of hK1 + ing the negative charge of the oxygen atom (20) through its =NH2 group, An could not of the nitro group at the C-4 position of the participate in an additional dipole-dipole in- aromatic ring, induced by resonance (16), teraction as 4-NAn can. with some group in the neighborhood of the Comparison of the Ki values for 4-NAn active center region of hK1, while 4-ABzA, (289.3 ± 92.8 µM) and for An (310,500 ± which shows an uncharged-NH2 group, also 38,600 µM) reveals that 4-NAn binds 1073- at the C-4 position of the aromatic ring, fold more strongly to a second binding site would not participate. The additional dipole- on hK1 than An. The second binding site for dipole interaction would reinforce the bind- these molecules is not known, but it is quite ing of 4-NAn to the active center of hK1. probable that the negative charge of the oxy- A similar interaction was suggested by gen atom of the nitro group of the 4-NAn Mares-Guia et al. (26) to explain their results molecule is more available for an additional about the electronic effects in the interaction dipole-dipole interaction with it than the of para-substituted benzamidines with tryp- negative charge at the C-4 position of the An sin. According to these authors, their data molecule. can be interpreted in terms of an enzyme- The fact that Ki>Ki indicates that both 4- inhibitor interaction of the dipole-dipole type. NAn and An bind to a second binding site in A dipole will appear in the inhibitor as a the hK1 molecule with lower affinity than consequence of an intramolecular charge they bind to the S1 subsite of the hK1 active transfer from the substituent to the ring or center. vice-versa. Their data concur with the model The presence of a second BPTI binding in which intramolecular charge transfer ren- site in hK1 has been already demonstrated ders positive the electron-donating substitu- (12). We do not know the location of this ent, thereby giving origin to a dipole that is second binding site for 4-NAn or An in hK1,

Braz J Med Biol Res 34(1) 2001 Human tissue kallikrein inhibition by amidines and anilines 43 and also whether it is different from the Bß-TR inhibition by An second binding site for BPTI. 4-NAn is a special case since it is present Comparison of the kcat/Km values for the in the substrate molecule D-Val-Leu-Arg- hK1- and Bß-TR-catalyzed hydrolysis of D- Nan where it shows neither a positive nor a Val-Leu-Arg-Nan (4.03 ± 0.43 and 8.4 ± 1.8 negative charge. In the ES complex, the Nan min-1 µM-1, respectively) reveals that D-Val- (4-NAn) group is accommodated at the S1 Leu-Arg-Nan is a slightly better substrate for position (20) of the hK1 active center. After Bß-TR than for hK1. substrate hydrolysis the released 4-NAn be- Additionally, comparison of the Ki val- comes a dipolar molecule. As a dipolar mol- ues for An inhibition of hK1 (37,340 ± 5,400 ecule 4-NAn is able to bind to the anionic µM) and of Bß-TR (10,800 ± 1,000 µM) site of hK1 as a competitive inhibitor through reveals that An interacts better with the S1 the positive charge on the amino group, and subsite of the active center of Bß-TR than is also able to bind to a second binding site in with the S1 subsite of the active center of hK1, possibly through the negative charge of hK1. the oxygen atom of the nitro group. Thus, 4- NAn is a product of the reaction that is able Acknowledgments to inhibit hK1 as a mixed inhibitor. The presence of a second inhibitor bind- We thank Prof. Giovanni Gazzinelli for ing site in hK1 seems to be clear and may reading the manuscript and offering con- have important implications in the physi- structive criticism. ological activity of this enzyme.

References

1. Bhoola KD, Figueroa CD & Worthy K kallikrein. Methods in Enzymology, 80: Souza EP, Rogana E, Santoro MM & (1992). Bioregulation of kinins: kallikreins, 466-492. Figueiredo AFS (1995). Kinetic mechan- kininogens and kininases. Pharmacologi- 7. Antonini E, Ascenzi P, Menegatti E, ism of the inhibition of human urinary kal- cal Reviews, 44: 1-80. Bortolotti F & Guarneri M (1982). Catalytic likrein by basic pancreatic trypsin inhibi- 2. MacDonald RJ, Margolius HS & Erdös EG properties of human urinary kallikrein. tor. Brazilian Journal of Medical and Bio- (1988). Molecular biology of tissue kal- Biochemistry, 21: 2477-2482. logical Research, 28: 505-512. likrein. Biochemical Journal, 253: 313- 8. Chagas JR, Portaro FCV, Hirata IY, 13. Rogana E, Nelson DL & Mares-Guia M 321. Almeida PC, Juliano MA, Juliano L & (1975). Characterization of the ultraviolet 3. Del Nery E, Chagas JR, Juliano MA, Prado Prado ES (1995). Determinants of the un- absorption spectra of para-substituted de- ES & Juliano L (1995). Evaluation of the usual cleavage specificity of lysyl-bradyki- rivatives of benzamidine. Journal of the extent of the binding site in human tissue nin-releasing kallikreins. Biochemical Jour- American Chemical Society, 97: 6844- kallikrein by synthetic substrates with se- nal, 306: 63-69. 6848. quences of human kininogen fragments. 9. Hial V, Diniz CR & Mares-Guia M (1974). 14. Plowman KM (1972). Inhibitor studies. In: Biochemical Journal, 312: 233-238. Purification and properties of a human uri- Hume DN, Stork G, King EL, Herschbach 4. Berg T, Bradshaw RA, Carretero OA, Chao nary kallikrein (kininogenase). Biochemis- DR & People JA (Editors), Enzyme Kinet- J, Chao L, Clements JA, Fahnestock M, try, 13: 4311-4318. ics. McGraw-Hill Book Company, New Fritz H, Gauthier F, MacDonald RJ, 10. Geiger R, Mann K & Bettels T (1977). York, 56-75. Margolius HS, Morris BJ, Richards RI & Isolation of human urinary kallikrein by 15. Cornish-Bowden A (1981). Inhibitors and Scicli AG (1992). A common nomencla- affinity chromatography. Journal of Clini- activators. In: Cornish-Bowden A (Editor), ture for members of the tissue (glandular) cal Chemistry and Clinical Biochemistry, Fundamentals of Enzyme Kinetics. kallikrein gene families. Agents and Ac- 15: 479-483. Butterworth & Co., Ltd., London, 73-98. tions, 38 (Suppl): 19-25. 11. Kettner C, Mirabelli C, Pierce JV & Shaw 16. Cram DJ & Hammond GS (1964). Physical 5. Schachter M (1980). Kallikreins (kinino- E (1980). Active site mapping of human properties. In: Cram DJ & Hammond GS genases) - A group of serine proteases and rat urinary kallikreins by peptidyl (Editors), Organic Chemistry. McGraw-Hill with bioregulatory actions. Pharmacologi- chloromethyl ketones. Archives of Bio- Book Company, New York, 183-225. cal Reviews, 31: 1-17. chemistry and Biophysics, 202: 420-430. 17. Mares-Guia M & Shaw E (1965). Studies 6. Geiger R & Fritz H (1981). Human urinary 12. Miranda TLS, Ramos CHI, Freire RTS, on the active center of trypsin. Journal of

Braz J Med Biol Res 34(1) 2001 44 M.O. Sousa et al.

Biological Chemistry, 210: 1579-1585. 21. Chen Z & Bode W (1983). Refined 2.5 Å EM (1976). Structural basis for the speci- 18. Katz BA, Liu B, Barnes M & Springman X-ray crystal structure of the complex ficity of phosphorylcholine-binding immu- EB (1998). Crystal structure of recombi- formed by porcine kallikrein A and the noglobulins. Immunochemistry, 13: 945- nant human tissue kallikrein at 2.0 Å reso- bovine pancreatic trypsin inhibitor. Jour- 949. lution. Protein Science, 7: 875-885. nal of Molecular Biology, 164: 283-310. 25. Segel IH (1975). Rapid equilibrium partial 19. Ascenzi P, Menegatti E, Guarneri M, 22. Weber G (1992). Transfer of proteins to and mixed-type inhibition. In: Segel IH Bortolotti F & Antonini E (1982). Catalytic apolar media and the dynamic interactions (Editor), Enzyme Kinetics. John Wiley & properties of serine proteases. 2. Com- of proteins and membranes. In: Weber G Sons, New York, 161-226. parison between human urinary kallikrein (Editor), Protein Interactions. Routledge, 26. Mares-Guia M, Nelson DL & Rogana E and human urokinase, bovine ß-trypsin, Chapman & Hall, Inc., New York, 161- (1977). Electronic effects in the interac- bovine thrombin, and bovine a-chymo- 176. tion of para-substituted benzamidines trypsin. Biochemistry, 21: 2483-2490. 23. Riordan JF, McElvany KD & Borders Jr CL with trypsin: the involvement of the P- 20. Schechter I & Berger A (1967). On the (1976). Arginyl residues: anion recogni- electronic density at the central atom of size of the active site in proteases. I. Pa- tion sites in enzymes. Science, 195: 884- the substituent in binding. Journal of the pain. Biochemical and Biophysical Re- 886. American Chemical Society, 99: 2331- search Communications, 27: 157-162. 24. Paddlan EA, Davis D, Rudikoft S & Potter 2336.

Braz J Med Biol Res 34(1) 2001