Multiple Determinants for the High Substrate Specificity of an ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistryand Molecular Biology, Inc.

Vol. 266,No. 29, Issue of October 15, pp. 19192-19197,1991 Printed in (I. S.A.

Multiple Determinants for theHigh Substrate Specificity of an Angiotensin II-forming Chymase from theHuman Heart* (Received for publication, April 15, 1991)

Akio Kinoshita, Hidenori UrataS, F. Merlin Bumpus, and Ahsan Husains From the Department of Heart andHypertension Research, Research Institute of the Cleveland ClinicFoundation, Cleveland, Ohio 44195-5069

Human heart chymase, a chymotrypsin-like serine Chymases have been implicated in the allergic response and proteinase that hydrolyzes the Phe8-Hiss bond in an- in peptide hormone processing (Reilly et al., 1982; Wintroub giotensin I (Ang I) to yield the octapeptide hormone et al., 1984; LeTrong et al., 1987b; Caugheyet al., 1988; Urata angiotensin 11 (Ang 11) and His-Leu, is the most spe- et al., 1990). cific, efficient Ang II-forming enzyme described. Other In a recent study using peptide 4-nitroanalide substrates, mammalian chymases display a much broader sub- Powers et al. (1985) have analyzed the extended substratestrate specificity. To better define its substratespeci- binding site of mammalian chymases. These studies suggest ficity, we have mapped the extended substrate-binding that human and dog skin chymases and rat chymases I and site of human heart chymase using Ang I analogs. The I1 have similar extended substrate-binding sites. In general, enzyme has a preference for aromatic amino acids these enzymes prefer a hydrophobic aromatic amino acid like phenylalanine, tyrosine, and tryptophanat thePI site. phenylalanine in the PI position of the substrate, and hydroAt the Sz subsite there is a significant preference for proline over hydrophobic or hydrophilic amino acids. phobic amino acids in thePzand P3position.’ Because proline There isno clear preferencefor hydrophobic or hydro- also is tolerated well in thePzposition, this position has been philic amino acids at the Si and S i subsites, but an Ang suggested to be less restrictive than the P1 position. These I analog containinga Pi proline is not hydrolyzed and observations and the findings that rat chymase I can hydroone with a P; proline is hydrolyzed poorly. An increas- lyze a number of naturally occurring peptides and proteins ing reduction in reactivityoccurs when the P position suggest that chymases sharea broad substrate specificity amino acids in Ang I are deleted sequentially from the similar to that of chymotrypsin. However, our recent studies N terminus. An increase or decrease in the length of with human heart chymase using peptide hormones as substrate have revealed a markedly different substrate specificity the His-Leu leaving group also produces a marked decrease in reactivity. No single determinant inAng I between chymases (Urata et al., 1990).For example, the Tyr4Ile‘ bond in the octapeptide Ang 11’ is readily hydrolyzed by is preeminently required for efficient catalysis, but several factors acting synergistically appearbetoim- rat chymase I (Le Trong et al., 1987b), but is not hydrolyzed by human heartchymase. Whereas the Phe7-Phe’ bond in the portant. Thus, we propose that ideal substrates for human heart chymase should contain the structure undecapeptide, substance P, and the Tyrzz-Leuz3 bond in the nXaa-Pro-[Phe, Tyr, or Trpl-Yaa-Yaa, where n 2 6; 28-amino acid residue peptide vasoactive intestinal peptide Xaa = any amino acid; Yaa = any amino acid except are readily hydrolyzed by dog mastocytoma chymase (Caughey proline. This structure existsin Ang I and neurotensin, et al., 1988), these bonds are either not hydrolyzed or poorly both of which are good substrates for human heart hydrolyzed byhuman heartchymase (Urata et al., 1990). This chymase. These findings indicate that the selection of distinction suggests that human heart chymase has a greater the scissile bond bythe extended substrate-binding sitesubstrate specificity than thatof nonprimate chymases. of human heart chymase is more restricted than that Human heart chymase shows a high catalytic efficiency for in other chymases. the hydrolysis of the Phes-Hisg bond in Ang I (Urata et al., 1990), resulting in the formation of the hormone Ang 11. Human heart chymase is the most specific and catalytically efficient Ang II-forming enzyme described (Urata et al., 1990). Chymases are a group of enzymes belonging to anhomolo- To better understand the highly restricted preference of hugous subclass of serine proteinases expressed by mast cells, man heart chymase for the Phe6-Hisgbond in Ang I, we have neutrophils, lymphocytes, and cytotoxic T-cells. Mast cells studied the binding and catalysis by human heart chymase of are a major site of localization of chymases (Woodbury and 36 peptide analogs of Ang I. Our studies show several deterNeurath, 1980). In mast cells, chymases are stored in the minants in Ang I that are important for efficient catalysis of active form within secretory granules (Sayama et al., 1987). the Phe’-Hisg bond and provide critical information for predicting new substrates for human heart chymase. * This work was supported in part by a grant from the Reinberger Foundation and by National Institues of Health Grant HL 33713. The costs of publication of this article were defrayed in part by the payment of page charges. This articlemust therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 4 Recipient of a fellowship from the American Heart Association, North East Ohio Affiliate. § To whom correspondence should be addressed Dept. of Heart and Hypertension Research, Research Institute of the Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195-5069. Tel.: 216-444-2057; Fax: 216-444-9263.

EXPERIMENTAL PROCEDURES

Purification of Human Heart Chymase-Human heart chymase was purified to homogeneity from human left ventricular tissue The nomenclature used for the individual amino acids (PI, PI’, etc.) of a .substrate and the subsites ( S I ,SI’, etc.) of the enzyme is that of Schechter and Berger (1967). The abbreviations used are: Ang 11, angiotensin 11; Ang I, angiotensin I;Ang 111, angiotensin 111; HPLC, high performance liquid chromatography; TEAP, triethylammonium phosphate.

19192

19193

Human Heart Chymase: Substrate Specificity according to the procedure of Urata et al. (1990). This tissue was obtained from donor hearts unsuitable for transplantation and from the excised hearts of patients undergoing cardiac transplantation; its use was approved by the Institutional Review Board. Peptides-Peptides used in this study were either purchased from Bachem (Torrance, CA) or synthesized by Dr. K. S. Misono, Cleveland Clinic Foundation. Peptides were purified on aCla reverse-phase HPLC column (2.2 X 25 cm; Vydac) with appropriate acetonitrile gradients containing 0.1% trifluroacetic acid and characterized by amino acid analysis and by analytical CIS reverse-phase HPLC. Peptides were purified to a peptide purity exceeding 99%. Peptide concentrations in stock solutions were standardized by amino acid analysis. Enzyme Kinetics-To determine K , and Vmaxvalues for the human heart chymase reaction with human Ang I and Ang I analogs, initial velocities ( u ) were determined. K,,, is the Michaelis-Menten constant is the maximal velocity. Eleven concentrations of substrate and VmaX ranging between 1and 2,000 p M with 0.65 or 13 ng (21.6 or 433 fmol, respectively) of human heart chymase were incubated for 20 min at 37 "C in 20 mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100. Reactions were terminated by the addition of 0.3 ml of ice-cold ethanol, and theresulting solution was evaporated to dryness. The residue was resuspended in 125 pl of distilled water, and 100 pl was applied to a Clareverse-phase HPLC column (Vydac) preequilibrated with 9% acetonitrile in 25 mM TEAP buffer, pH 3.0. The column was developed using a 6-min linear acetonitrile gradient (9-35%) in 25 mM TEAP buffer, pH 3.0, at a flow rate of 2 ml/min. The column effluent was monitored at 214 nm. The elution positions ofAng I, Ang I analogs, Ang 11, and Ang I1 analogs were also determined using pure synthetic standards. The peak area corresponding to Ang I1 or its analog was integrated to calculate Ang I1 or Ang I1 analog formation. Products were separated by reverse-phase HPLC and identified by amino acid analysis. K, and V,, values were calculated by nonlinear regression using the equation, u = Vmax X [S]/(K,,,+ [SI) where u is the initial velocity and [SI is the concentration of substrate. Correlation coefficients were routinely greater than or equal to 0.98. The concentration of human heart chymase was determined by the method of Lowry et al. (1951) using pure bovine a-chymotrypsin as standard. The overall rate constant kea, was calculated by the formula kat= Vm../[Eo] where [Eo]is the total enzyme concentration. 160, the concentration of peptide producing a 50% inhibition of the conversion of Ang I to Ang I1 by human heart chymase, was determined for Ang I analogs not cleaved by human heart chymase. Ang I analogs (11 concentrations ranging between 10 and 3,000 p ~ were ) preincubated with 2.6 ng (87 fmol) of human heart chymase for 120 min at 37 "C in 20mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100 in a final volume of 190 rl. Ten microliters of2.5 mM Ang I was then added to these preincubated enzyme preparations and incubated for 20 min a t 37 "C (final [Ang I] = 125 p ~ )Ang . I1 formed during this incubation was separated and quantitated as described above. To determine the 160 value for inhibition I (final concentration 1 nM) was used as the by Ang 11, [1251]iodo-Ang substrate; the incubation time was 5 min. [1251]Iodo-AngI1 formed during the incubation was separated and quantitated as described by Urata etal. (1990). Iso values were calculated by nonlinear regression using the equation, u = 100 X [S]/(Z, + [SI). K,, the dissociation

constant of the enzyme-inhibitor complex, wasdetermined using the formula I,, = ((E0/2) + K,) + Ki[S]/K, (Cha, 1975). RESULTS AND DISCUSSION

At the inception of these studies, we were surprised by the finding that the PheS-Hisgbond in Ang I washydrolyzed readily by human heart chymase, whereas no hydrolysis of the Tyr4-Ile' bond occurred (Urata etal., 1990).Earlier studies with rat mast cell and dog mastocytoma chymases by Powers et al. (1985) had shown that an aromatic hydrophobic amino acid in thePI position was of primary importance for efficient catalysis. Because the amino acids phenylalanine and tyrosine both contain an aromatic hydrophobic side chain, the preference of human heart chymase for the Phe'-Hisg bond suggested that either the Sl substrate-binding pocket of human heart chymase is markedly different from that of other chymases, resulting in either an unusually high specificity for a Pl phenylalanine, or that theextended substrate-binding site of human heart chymase plays a major role in determining substrate specificity. The SI Substrate-binding Pocket of HumanHeart Chymuse-Functionally, the SI binding pocket of human heart chymase appears to be similar to thatof other chymases. Ang I ($, = 160 s-') and its analogs, [Tyr'IAng I (kat= 130 s-') and [Trp'] AngI (kc,,= 210 s-I), all were hydrolyzedefficiently by human heart chymase at the Phea-Hisg, Tyr'-Hisg* and Trp'-Hisg bonds, respectively (Table I). Substitution of a leucine in place of the phenylalanine at theP1position of Ang I led to a slower rate of catalysis by human heart chymase (for [Leu'IAng I, kcat = 46 s-I), but neither [Va18]Ang I nor [Ile'IAng I were cleaved. These data obtained with human heart chymase are similar to the findings of Powers et al. (1985) for the hydrolysis of peptide 4-nitroanalide substrates by rat mast cell, dogskin, and humanskin chymases, indicating that chymases prefer a PI hydrophobic aromatic amino acid. A number of residues are involved in forming the SI pocket of serine proteinases. Amino acids at positions 189 and 226 (chymotrypsinogen numbering) play a significant role in determining the primary substrate specificity (Graf et al., 1987; Murphy et al., 1988). In humanheart chymase the amino acids in positions 1891226 are Ser/Ala (Urata et al., 1991). Similar or identical residues are found in other chymases. In dog chymase (Caughey et al., 1990), rat chymase 1 (Le Trong et al., 1987a), and rat chymase 2 (Benfey et al., 1987) the 1891226positions are occupied by SerIAla, SerIAla, and Ala/Ala, respectively. These comparisons provide structural evidence for the similarity between the primary substrate-binding site of human heart chymase and other chymases. To explain the unusually high specificity of human

TABLEI Effect of PI amino acid residue on the hydrolysis of Ang I by human heart c h y m e The kinetic constants were determined in 20 mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100 at 37 "C. K,,, values were determined by nonlinear regression. K , and kcatvalues are means f S.E. of three separate determinations for each peptide. K, values represent a single determination. Peptide name

Peptide structure"

K,

k

PM

5-1

t

kdKm pM"s"

K,

[Substrate or inhibitor1

pM

PM

PI

AngI 1 2 3 4 5

Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-Leu 1 6 0 f 2600 f 6 A s p - A r g - V a l - T y r - I l e - H i s - P r o - x - 4 - H i s - L e u 46 f 10 Asp-Arg-Val-Tyr-Ile-His-Pro-=-J-His-Leu 110 f 10 Asp-Arg-Val-Tyr-Ile-His-Pro-=-L-His-Leu 310 f 30 Asp-Arg-Val-Tyr-Ile-His-Pro-x-His-Leu NCb Asp-Arg-V 40-3000 a l - T y r - I l e - H i s - P r o640 -G-His-Leu NC

10-700 130 -+ 8 210 f 10 46 f 4 10-1000

2.7 2.8 1.9 0.15 230

10-700 10-700 10-700

Arrows indicate the peptide bond hydrolyzed by human heart chymase. Underlined residues differ from the homologous residue in Ang I. For all peptide substrates no other cleaveage site was observed. * NC, no detectable cleavage. Minimum kc,, detectable was about 0.1 SKI.

19194

Human Heart Chymase: Substrate Specificity

heart chymase for the Phe8-Hisgbond in Ang I, we explored the importance of the extended substrate-bindingsite on catalysis. The ExtendedSubstrate BindingSite of Human Heart Chymase-The catalysis of substrates by serine proteinases involves the formation and thehydrolysis of the acyl-enzyme. The amino acids in the P' positions, which form the leaving group, are potentially involved in interactingwith the enzyme and in forming the acyl-enzyme. Because acylation appears to be the rate-limiting stepfor peptide substrates, the leaving group may play a critical role in determining the rate of substrate catalysis (Polgar, 1989). The Si and SI subsites of human heart chymase were studied by examining the effects of amino acid substitutions at thePi (Table 11) and P,' (Table 111)positions of Ang I on catalysis by human heart chymase. These studies as well as our previous studies (Urata et al., 1990) indicate that humanheart chymase shows a slight preference for hydrophobic amino acids over charged amino acids in the P,' position, and that there is less specificity at the Pi position. Peptides with proline in the Pi position of Ang I are not hydrolyzed by human heart chymase, whereas proline in the P,' position of Ang I decreases the kCat/Km by 95%. Although a preference for hydrophobic amino acids in the Pi position has been described for dog mastocytoma chymase (Caughey et al., 1988), the interaction between the Si and S; subsites of chymases and therespective P' position amino acids in model substrates hitherto have not been analyzed systematically. Thus, directly comparing the S' subsites of human heart chymase with those of other chymases is not possible. Le Trong et al. (1987b) found that rat chymase 1 first hydrolyzes the Tyr4-Ile' bond in Ang I and then the Phe8Hisg bond, whereas we found that humanheart chymase hydrolyzes the Phe8-Hisg but not the Tyr4-Ile' bond. Thus we concluded that the peptide chain length of the C-terminal leaving group and of the N-terminal product may be determinants for catalysis by chymases. In the first setof experiments we examined the effect of the length of the C-terminal leaving group. Ang I and all its analogs which were either extended (by 1 to 4 amino acids based on the structure of the Ang I precursor) or truncated (by 1amino acid) at theC terminuswere cleaved correctly at the Phea-Hisg bond (Table IV). However, the dipeptide His-Leu appears to provide the optimal length for catalysis. When the leaving group was histidine (as for peptide 20, Table IV) or His-Leu-Val-Ile-His-Asn (as for peptide 24, Table IV), the kcat for Ang I1 formation was0.7 and 4%, respectively, of that observed for Ang I. The choice of amino

acids used in the extended Ang I was based on the structure of human angiotensinogen. Increases in the length of the leaving group in Ang I from a dipeptide to a hexapeptide (peptide 24, Table IV) produced an ZlO-fold improvement in binding affinity. This improvement in binding, however, was less productive, because the kcat and the specificity constant kc,,/& decreased concomitantly (compare peptides 21-24 with Ang I; Table IV). In contrast to rat chymase 1 (Le Trong et al., 1987b) and dog mastocytoma chymase (Caughey et al., 1988), human heart chymase clearly prefers a dipeptide leaving group for catalysis, indicating that important differences exist between the S' subsites of various chymases. Based on the crystal structure of rat chymase 2, Remington et al. (1988) have proposed that Pi to Pi residues of substrates interact with residues in the loop region formed by residues 35-41 (chymotrypsinogen numbering) in this enzyme. The differences in functional S' subsite preferences between human heart chymase, dog mastocytoma chymase, and rat chymase 1 are notsurprising because residues 35-41 (chymotrypsinogen numbering) in dog mastocytoma chymase and in rat chymase 1are only 20 and 30%homologous, respectively, compared to those residues in human heart chymase (Fig. 1).Of particular interest may be the location of a proline, a rigid kinked amino acid, situated in the middle of the loopregion formed by residues 35-41 (chymotrypsinogen numbering) of human heart chymase. Proline is not found in the corresponding loop region of mouse, rat, and dog chymases, which could greatly influence the conformation of the S' substrate-binding site region of human heart chymase. As discussed above, a PI hydrophobic aromatic amino acid and a dipeptide leaving group in substrates of human heart chymase produce an optimal specificity constant for catalysis. However, evidenced by the inability of human heart chymase to degrade the Phe4-Ar$ bond in the hexapeptide yMSH(38) (Urata et al., 1990), the requirements for phenylalanine or tyrosine at the P, position, and of a dipeptide leaving group appear not to be preeminent for efficient catalysis by human heart chymase. These observations suggest that additional interactions between regions in the extended substrate-binding site of the enzyme and the P position amino acids in the substrate are important in determining the high substrate specificityof humanheart chymase. The eight potential amino acids in the P positions ofAng I (representing the entire Ang I1 structure) may interact with human heart chymase to enhance binding and catalytic efficiency. Ang I analogs truncated at the N-terminal by one to five amino acids were all cleaved by human heart chymase at thepenultimate

TABLE I1 Effect of PI' amino acid residue on the hydrolysisof A n g I by human heart chymase The kinetic constants were determined in 20 mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100 at 37 ' C . K, values were determined by nonlinear regression. K,,, and kcatvalues are means f S.E. of three separate determinations with each peptide. K, values represent a single determination. Peptide name

Peptide structure"

8

9 10

kcat

PM

S-1

kdKm pM"

s"

K,

[Substrate or inhibitor]

PM

PM

' 10-700 6 0 f 6 2.7 160 f 20 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-J-His-Leu 2.9 10-700 88 f 3 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-J-E-Leu 3 0 k 3 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-J-E-Leu 2 6 f 0 . 2 21-200 8f0.2 1.10 0.80 10-700 120 f 10 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-J-Q-Leu 150 k 20 A s p - A r g - V a l - T y r - I l e - H i s - P r o - P h e - J - G - L e u 86 f 10 63 f 6 10-700 0.73 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-=-Leu NCb 46 10-700 'Arrows indicate the peptide bond hydrolyzed by human heart chymase. Underlined residues differ from the homologous residue in Ang I. For all peptide substrates no other cleavage site was observed. NC, no detectable cleavage. Minimum kcatdetectable was about 0.1 s-'. PI

AngI 6 7

K"8

19195

Human Heart Chymase: Substrate Specificity TABLE 111 Effect of P2' amino acid residue on the hydrolysisof Ang I by human heart chymase The kinetic constants were determined in 20 mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100 a t 37 "C. K,,, values were determined by nonlinear regression. Values are means -t S.E. of three separate determinations with each peptide. Peptide name

AngI 11 12 13 14 15 16

Peptide structure"

K",

kcat

PM

S"

kcaJKrn

PZ' Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-Leu Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-E Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-e Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-As Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-w Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-Q Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-4-His-Pr

[Substrate] &M" s"

@M

60 f 6 10-700 160 2.7 f 20 1.8 53 f 5 30 f 6 1.5 110 f 1010-7001 6 0 f 6 100 f 10 100 f 2 10-7001.0 350 f 60 2 3 0 f 10 10-700 0.66 110f10 0.66 73 f 6 350f30 50 f 0.8 10-700 0.14

10-700

10-700

Arrows indicate the peptide bond hydrolyzed by human heart chymase. Underlined residues differ from the homologous residue in Ang I. For all peptide substrates no other cleavage site was observed.

TABLEIV Effect of peptide length onthe hydrolysis of Ang I by human heart chymase The kinetic constants were determined in 20 mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100 a t 37 "C. K,,, values were determined by nonlinear regression. Values are means f S.E. of three seDarate determinations with eachDeDtide. Peptide name

Peptide structure"

K?" IrM

Ang I 17 18 19 20 21 22 23 24

Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-L-His-Leu 60 -t 6 190 -t 20 Val-Tyr-Ile-His-Pro-Phe-4-His-Leu 200 -+ 20 Ile-His-Pro-Phe-L-His-Leu 380 -+- 90 His-Pro-Phe-4-His-Leu 310 -t 30 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-A-His 62 -t 3 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-4-His-Leu-Val 48 -C 7 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-~-His-Leu-Val-Ile A s p - A r g - V a l - T y r - I l e - H i s - P r o - P h e - ~ - H i s - L e u - V a l - I l e - H i s 29 f. 1 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-~-His-Leu-Val-Ile-His-Asn 6.2 f 0.5

k*t S"

160 f 20 320 +. 20 110 f 10 56 f. 2 1.1 ? 0.1 45 -C 7 27 -+- 2 29 -+- 1 6.2 f 0 . 3

k c d K m

pM"

[Substrate] -

s"

PM

10-700 10-700 10-700 100-2000 10-700 10-700 10-700 10-700

2.7 1.7 0.55 0.15 0.0035 0.73 0.56 1.oo

1-400 1.00

a Arrows indicate the peptide bond hydrolyzed by human heart chymase. For all peptide substrates no other cleavage site was observed. "

(human hear( chymase numbring)

chymasea heafl Human Rat mast cell protease Ib Rat mast cell protease IIC Mouse mast cell protease IVd Dog mastocytoma chymasee Mouse mast cell protease If Mouse mast cell protease IN Bovine chymottypsinh (chymotrypsinnumbering)

20

L L L L L L L L

E E D E E k k q

I I I I I I f d 35

30

V T S N t T e r V T e k t T e k L T l r I T d r t T k N k T g f

G G G G n G G

P y l f h s s

S k r T l e k

K a v a a d e h

F t i t s r r F

C C C C C C C C

G G G G G G G G

40

FIG. 1. Homologous structural relations among a loop region (residues 35-42; chymotrypsin numbering) proposed to form the S' substrate-binding site of chymases and chymotrypsin. Upper case letters indicate residues that are structurally similar (i.e. conservative substitutions), lower case letters indicate residues that are not structurally similar. 'Urata etal., 1991; bLe Trong et al.,1987a;'Benfey et al., 1987; dSerafin et al.,1991; "Caughey et al.,1990 ; 'Lobe et al., 1986; %erafin et al., 1990 hHartley, 1970.

bond (i.e. the Phe-His bond) (peptides 17 to 19, Table IV). Any decrease in the peptide chain length at the N terminus ofAng I produced marked sequential decreases in k,,,/K,,, (from 2.7 pL"' s-' for the decapeptide Ang I to 0.15 p ~ " s" for the N-terminal truncated pentapeptide ofAng I). This decrease in the specificity constant was due mainly to a decrease in binding of the N-terminally truncated nona-, octa-, and hexapeptides of Ang I by human heart chymase. This observation is also supported by the observation that shorter C-terminal fragments of the octapeptide Ang 11 bind to human heart chymase with less avidity than longer ones (Table V). These findings clearly suggest the importance of the P position amino acids in Ang I for a high affinity interaction with human heart chymase. The catalysis of the

TABLEV Inhibition of human heart chymase by Ang II, Ang I l l , and Ang II-derived peptides Inhibitor peptides were preincubated with theenzyme for 120 min in 20 mM Tris-HC1, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100, before theaddition of thesubstrate, Ang I. The kinetic constants were determinedinthesame buffer. No cleavage was observed with any of the peptides listed. K, values represent a single determinationusing 11 differentconcentrations of eachinhibitor peptide. Peptide Peptide structure K, [Inhibitor] name

W

PM

Ang 11 Asp-Arg-Val-Tyr-Ile-His-Pro-Phe25 10-500 AngIII Arg-Val-Tyr-Ile-His-Pro-Phe 10-700 37 25 V a l - T y r - I l e - H i s - P r o - P h e 59 10-700 T y r - I l e - H i s - P r o - P h e 99 20-2000 26 27 I l e - H i s - P r o - P h e 100 20-2000 28 H i s - P r o - P h e 620 30-3000

Phes-Hisg bond in Ang I, however, is not greatly dependent on interactionsof Ang I with human heartchymase substratebinding subsites Ss to S5. It has been demonstrated by Yoshida et al. (1980) and Powers et al. (1985) that the Pro-Phe sequence in positions P2-P1of peptide substrates is highly favored by leukocyte chymotrypsin-like proteinases such as chymases and cathepsin G. To study substrate preferences, several peptide hormones were testedassubstratesfor chymases: corticotropin(1-24), luteinizing hormone, [Met'lenkephalin, a-melanocyte-stimulating hormone, y-melanocyte-stimulating

19196

Human Heart Chymase:Substrate Specificity TABLE VI Effect of Pzamino acid residue on the hydrolysisof Ang Z by human heart chymase The kinetic constants were determined in 20 mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100 at 37 "C.K,,, values were determined by nonlinear regression. K , and katvalues are means f S.E. of three separate determinations with each peptide. K, values represent a single determination.

Peptide name

structure"

Peptide

K,

kat

"

s-1

k*t/K. pM"

s"

Ki

[Substrate or inhibitor]

WM

*M

p2

AngI 29 30 31 32 33

Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-l-His-Leu 60-66 160+20 2.7 10-700 Asp-Arg-Val-Tyr-Ile-His-&-Phe-i-His-Leu 140+40 170f30 1.2 10-700 Asp-Arg-Val-Tyr-Ile-His-Arg-Phe-i-His-Leu 230+ 7 110 f 6 0.48 10-700 Asp-Arg-Val-Tyr-Ile-His-e-Phe-l-His-Leu 7 . 6 f 0 . 11 8 & 7 0.42 10-100 A s p - A r g - V a l - T y r - I l e - H i s - ~ - P h e - l - H i s - L e u170 + 10 56 f 3 0.33 10-700 Asp-Arg-Val-Tyr-Ile-His-E-Phe-His-Leu NCb 10-500 29 'Arrows indicate the peptide bond hydrolyzed by human heart chymase. Underlined residues differ from the homologous residue in Ang I. For all peptide substrates no other cleavage site was observed. * NC, no detectable cleavage. Minimum katdetectable was about 0.1 s-'.

TABLEVI1 Kinetic constants for the hydrolysisof Ang I, [TyrS,IlegJANGZ, and neurotensin by human heart chymase The kinetic constants were determined in 20 mM Tris-HC1 buffer, pH 8.0, containing 0.5 M KC1 and 0.01% Triton X-100 at 37 "C. K, values were determined by nonlinear regression. Values are means f S.E. of three separate determinations for each peptide. Peptide name

structure"

Peptide

K,

L,

Lt/Km ~~~

W

Ana I 34 Neurotensin

Asp-Ara-Val-Tvr-Ile-His-Pro-Phe-l-His-Leu 6 0 f 6 A s p - A r g - V a l - T y r - I l e - H i s - P r o - ' l & " ~ - ~ - L e u 20 f 3 p G l u - L e u - T y r - G l u - A s n - L y s - P r o - A r g - A r g - P r o - ~ - J . - G -210 L e uf 30 "

s-1

cM-' s-l

1 6 0 f 210-700 0 2.7 32 f 2 1.6 49 f 8 0.23

[Substrate] ~

FM

10-700 30-3000

Arrows indicate the peptide bond hydrolyzed by human heart chymase. Underlined residues differ from the homologous residue in Ang I. For all peptide substrates no other cleavage site was observed.

hormone-(3-8), substance P, and vasoactive intestinal peptide. A proline does not precede a phenylalanine, tyrosine, or tryptophan in these peptide hormones. Despite the absence of a Pz proline, many of these peptide hormones are cleaved readily by rat and dog chymases but not by human heart chymase (Le Trong et al., 198713; Caughey et al., 1988; Urata et al., 1990). Because positions Pz-P1of Ang I contain the Pro-Phe sequence, it is likely that a Pzproline is a particularly important determinantfor catalysis by human heartchymase. Ang I analogs, in which the Pzproline was replaced by alanine, arginine, or histidine, were cleaved efficiently, albeit a little less efficiently than peptides with a Pz proline, at the PhesHisg bond by human heart chymase (Table VI). This finding disputes the contention that proline in theP2position with a PI hydrophobic aromatic amino acid is absolutely required for hydrolysis by human heart chymase. Interestingly, the PhesPheg bond in [Pheg]AngI containing aPzproline is catalyzed efficiently (kat= 88 s-') by human heart chymase, but the Phe7-Phee bond in [Phe7]Ang I containing no Pz proline is not cleaved (peptide 33, Table VI). The Phe7-Phe' bond in the undecapeptide hormone, substance P, containing no P2 proline, also is not cleaved by human heart chymase. These collective data show that if other factors are optimal, a PZ proline is not critical for the catalysis of peptide substrates by human heart chymase, but when other factors such as the C-terminal leaving group are notoptimal, a Pzproline appears be a critical determinant for catalysis. The main role of the Pz proline, however, is not to increase the free energy of binding ( [Phe7]AngI containingno PZproline binds relatively well to human heart chymase; Ki= 29 p M ) , but to increase the free energy necessary for the acylation of human heart chymase, and hence, the catalysis of peptide substrates. The importance of the P2proline for catalysis is anunusual feature of human heart chymase not sharedby dogand ratchymases.

In the course of these investigations with human heart chymase it became apparent that no single determinant in the substrate, Ang I, is absolutely required for efficient catalysis, although important roles seem to be played by several determinants acting synergistically: 1) a PI phenylalanine, tyrosine, or tryptophan; 2) a P p proline; 3) the presence of P,positioned amino acids, where n > 6; and 4) a dipeptide Cterminal leaving group containing no prolines. Relating these observations to why the Phes-Hisg but not theTyr4-Ile' bond in Ang I was readily hydrolyzed by human heart chymase, we propose that the Tyr4-Ile' bond in Ang I is not agood target for catalysis by human heart chymase because the structural environment of the Tyr4-Ile' bond, compared to that of the Phes-Hisg bond, is far from optimal. Specifically, with reference to theTyr4-Ile' bond in Ang I, the valine in the putative Pz position is not optimal, the putative C-terminal leaving group is too long, and the peptide chain length of the Ppositioned amino acids is too short. To prove this hypothesis we synthesized an Ang I analog in which the PheO-Hisgbond was replaced with a TyrS-Ileg bond. [Tyrs,Ileg]Ang I was cleaved readily (kcat= 32 s-') by human heart chymase at the Tyra-Ilegbond but notat theTyr4-Ile5bond (peptide 34, Table VII). This finding indicates that the extended substratebinding site in human heart chymase is important in the selection of the scissile bond within Ang I. In initial substrate specificity studies, Urata et al. (1990) described the inability of human heart chymase to cleave various peptide hormone substrates. The substrates used in that study were chosen randomly from a list of peptides used to address the specificity of rat and dog chymases. Based on key factors suggested above for the optimal catalysis of peptide substrates by human heart chymase, we searched a protein database for peptide hormones other than Ang I having the structure, nXaa-Pro-[Phe, Tyr, or Trpl-Yaa-Yaa, where

Heart Human

Chymase:Specificity Substrate

19197

n 2 6; Xaa = any amino acid; Yaa = any amino acid other Hartley, B. (1970) Philos. Trans. R. SOC.London B Biol. Sci. 267, 77-87 than proline. Neurotensin, a tridecapeptide hormone with the Kobayashi, K., and Katunuma, N. (1978) J. Biochem. (Tokyo) 8 4 , structure PGlu-Leu-Tyr-Glu-Asn-Lys-Pro-Arg-Arg-Pro-Tyr65-74 Ile-Leu, fulfills this criterion. As expected, the Tyr1'-Ile12bond Kobayashi, K., Sanada, Y., and Katunuma, N. (1978) J. Biochem. in neurotensin is cleaved efficiently by human heart chymase (Tokyo) 84,477-481 (kat= 49 s-') (Table VII). This observation strongly supports Le Trong, H., Parmelee, D. C., Walsh, K. A., Neurath, H., and Woodbury, R. G. (1987a) Biochemistry 26,6988-6994 the importance of multiple synergistic factors in determining Le Trong, H., Neurath, H., and Woodbury, R. G. (1987b) Proc. Nutl. the substratespecificity of human heart chymase. Acad. Sci, U. S. A . 84,364-367 Summary and Conclusions-Human heart chymase is a Lobe, C. G., Havele, C., and Bleackley, R. C. (1986) Proc. Natl. Acad. highly efficient and specific enzyme which catalyzes the conSci. U.S. A. 83,1448-1452 version ofAng I to Ang I1 (Urata et al., 1990). The high Lowry, 0.H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) specificity of human heart chymase is an unusual characterJ. Biol. Chem. 1 9 3 , 265-275 istic for the chymase family of chymotrypsin-like proteinases, Murphy, M. E. P., Moult, J., Bleackley, R.C., Gershenfeld, H., Weissman, I. L., and James, M. N. G.(1988) Proteins Struct.Funct. which are known for their broad substrate specificity (KobaGenet. 4,190-204 yashi et al., 1978; Kobayashi and Katunuma, 1978; Powers et Polgar, L. (1989) Mechunims of Protease Action, pp. 87-122, CRC al., 1985;Le Trong et al., 1987,a and b).We report here that Press, Inc., Boca Raton, FL the high substrate specificity of human heart chymase for Powers, J. C., Tanaka, T., Harper, J. W., Minematsu, Y., Barker, L., Ang I is due to thepresence of multiple determinants in the Lincoln, D., Crumley, K. V., Fraki, J. E., Schechter, N. M.,Lazarus, G. G., Nakajima, K., Nakashino, K., Neurath, H., and Woodbury, substrate primary structurewhich are necessary for efficient R. G. (1985) Biochemistry 2 4 , 2048-2058 catalysis. Based on the broad specificity of rat and dog chyReilly, C. F., Tewksbury, D.A., Schechter, N. M., and Travis, J. mases, some investigators have proposed that chymases in (1982) J.Biol. Chem. 267,8619-8622 different species play similar but diverse roles in the allergic Remington, S. J., Woodbury, R. G., Reynolds, R.A., Matthews, B. response and in peptide hormone processing. Our findings W., and Neurath, H. (1988) Biochemistry 27,8097-8105 support the hypothesis that in humans, particularly in human Sayama, S., Iozzo, R. V., Lazarus, G. S., and Schechter, N. M. (1987) J.Biol. Chem. 262,6808-6815 heart tissue, a chymase is present which has evolved to a Schechter, I., and Berger, A. (1967) Biochem. Biophys. Res. Commun. highly specific enzyme with more discrete functions. Acknowledgments-We thank Dr. Robert M. Graham for his helpful and constructive criticism of the manuscript. We wish to acknowledge the excellent technical assistance provided by Dennis Wilk and editorial assistanceby Suzanne Hazan. REFERENCES Benfey P. N., Yin, F. H., and Leder, P. (1987) J. Biol. Chem. 262, 5377-5384 Caughey, G. H., Leidig, F., Viro, N. F., and Nadel, J. A. (1988) J. Pharmacol. Exp. Ther. 244,133-137 Caughey, G. H., Raymond, W. W., and Vanderslice, P. (1990) Biochemistry 29,5166-5171 Cha, S. (1975) Biochem. Phurmacol. 24,2177-2185 Graf, L., Craik, C. S., Patthy, A., Roczniak, S., Fletterick, R. J., and Rutter, W. J. (1987) Biochemistry 26,2616-2623

27,157-162 Serafin, W. E., Reynolds, D. S., Rogelj, S., Lane, W. S., Conder, G. A., Johnson, S. S., Austen, K. F., and Stevens, R. L. (1990) J. Biol. Chem. 266,423-429 Serafin, W. E.,Sullivan, T. P., Conder, G. A., Ebrahimi, A., Marcham, P., Johnson, S. S., Austen, K. F., and Reynolds, D. S. (1991) J. Biol. Chem. 266,1934-1941 Urata, H., Kinoshita, A., Misono, K. S., Bumpus, F. M., and Husain, A. (1990) J. B i d . Chem. 266,22348-22357 Urata, H., Kinoshita, A., Perez, D., Misono, K. S., Bumpus, F. M., Graham, R. M., and Husain, A. (1991) J.Biol. Chem. 266,1717317179 Wintroub, B. U., Schechter, N. M., Lazarus, G. S., Kaempfer, C. E., and Schwartz, L. B. (1984) J. Invest. Dermatol. 83,336-339 Woodbury, R. G., and Neurath, H. (1980) FEBS Lett. 114,189-196 Yoshida, N., Everitt, M. T., Neurath, H., Woodbury, R. G., and Powers, J. C. (1980) Biochemistry 19,5799-5804