Purification and characterization of a non-kallikrein arginine esterase from dog urine

276

Biochirnica et Biophysica Acta 964 (1988) 276-284 Elsevier

BBA 22872

P u r i f i c a t i o n and c h a r a c t e r i z a t i o n of a non-kailikrein arginine e s t e r a s e from dog urine K i s h o r e K. M u r t h y a n d A. G u i l l e r m o Scicli Hypertension Research Division, Henry Ford Hospital, Detroit, MI (U.S.A.) (Received 21 May 1987) (Revised manuscript received 16 October 1987)

Key words: Arginine esterase; Proteolysis; Kininogenase; Trypsin-like enzyme; (Dog urine)

A non-kallikrein arginine esterase (esterase I) has been purified from dog urine and characterized. The enzyme was purified by a three-step procedure, including ion exchange chromatography on DEAE-Sephacel, affinity chromatography on p-aminobenzamidine-Sepharose, and final gel filtration on UItrogel AcA-54. The purified preparation gave three protein bands on polyacrylamide gel electrophoresis, all of which had esterolytic activity. The enzyme has a specific activity of 601 esterase units/rag protein. It has negligible kininogenase activity. Esterase I gave two closely migrating protein bands on reduced sodium dodecyl sulfate-polyacrylamide gel electrophoresis with molecular weights of 34000 and 33300. Esterase I is a glycoprotein with a pH optimum of 9.5 and a pl of 4.62. The enzyme is strongly inhibited by a host of inhibitors including aprotinin, leupeptin, antipain, soybean trypsin inhibitor, lima bean trypsin inhibitor, and DPhe-Phe-Arg-chloromethyl ketone (150 in the 10-9-10 -8 M range). However, p-aminobenzamidine, Na-p-tosyMysyl chloromethyl ketone and phenylmethylsulfonyl fluoride were weak inhibitors, with 150 values in the 1 0 - 5 - 1 0 - 7 M range. The enzyme preferentially hydrolyzes Pro-Arg bonds. Among fluorogenic substrates used in this study, butyloxycarbonyl-Val-Pro-Arg-methylcoumarinamide (a-thrombin substrate) was found to be the best, with a K m of 1.7/LM and a kcat//Km of 6.3 s " btM -I. However, esterase I does not convert fibrinogen to fibrin nor activate plasminogen to plasmin. Esterase I is immunologically distinct from dog urinary kallikrein, having no cross-reactivity with antibodies against dog kailikrein.

Introduction

Dog urine contains at least two arginine esterhydrolyzing enzymes, as evidenced by ion exchange chromatography. Further purification indicated that only one of these esterases is glandular

Abbreviations: TAME, Na-p-tosyl-arg/lnine methyl ester; PAGE, polyacrylamide gel electrophoresis; TLCK, Na-ptosyl-lysyl chloromethyl ketone; -MCA, -methylcoumarinamide; Boc-, butyloxycarbonyl-; Glt-, glutaryl-. Correspondence: A.G. Scicli, Hypertension Research Division, 2799 W. Grand Boulevard, Detroit, MI 48202, U.S.A.

kallikrein [1,2]. Similar observations on the presence of glandular kallikrein and a non-kallikrein esterase in rat urine have been reported earlier [3-5]. Both kallikrein and the alkaline arginine esterase in rat urine have been purified and characterized [5,6]. The non-kallikrein esterase, named esterase A by Nustad and Pierce [3], has been shown to be a plasminogen activator [6]. This enzyme has been localized in the salivary gland, kidney and other glandular tissues. The esterase is also reported to be regulated by testosterone in the submandibular gland and kidney in rats [7]. However, very little information is presently available regarding the properties of the urinary non-

0304-4165/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

277 kallikrein esterase in the dog, a widely used laboratory animal. Purification of this enzyme and a detailed study of its properties would help localize the enzyme, as well as throw light on its pathophysiological role. As a first step in this direction, we report the purification and characterization of this potent canine urinary arginine esterase. The substrate specificity of this enzyme was also evaluated using synthetic fluorogenic methylcoumarinamide peptide substrate analogues. Materials and Methods

Urine was collected from male mongrel dogs fed a normal diet and water ad libitum. The urine was collected in 0.1% sodium azide and stored at 4 ° C until used. The chemicals used in this study were obtained from the following sources: soybean trypsin inhibitor, lima bean trypsin inhibitor, leupeptin, antipain, phenylmethylsulfonyl fluoride (PMSF) and tosyl-lysyl chloromethyl ketone (TLCK) were from Sigma Chemical Company (St. Louis, MO), tosylarginine [3H]methyl ester (3H-TAME) from Amersham (Arlington Heights, IL), SDS from Bio-Rad Laboratories (Richmond, CA), Pro-PheArg-MCA from Vega Biochemicals (Tucson, AZ), and all other methylcoumarinamide substrates were from Peninsula Laboratories (Belmont, CA).

(urokinase); (3) Boc-Glu-Lys-Lys-MCA (plasmin); (4) Pro-Phe-Arg-MCA (kallikrein); (5) Suc-AlaPro-Ala-MCA (elastase); (6) Suc-Arg-Pro-PheHis-Leu-Val-Tyr-MCA (renin). The enzymes which preferentially cleave each of these substrates are noted in parentheses.

Kininogenase activity This was assayed by incubating the samples with partially purified dog kininogen, the released kinins being assayed as previously described [11].

Caseinolytic activity Casein (1%) in 0.1 M phosphate buffer (pH 7.6) was incubated with 500 ng esterase I or trypsin. An aliquot (200 ml) was removed after 0, 60, 120 and 180 min of incubation at 37 ° C. The proteins were precipitated with 800 /~1 of 15% trichloroacetic acid, kept at 4°C for 1 h, centrifuged, and the A280 of the supernatant was recorded. All enzymatic activity results are the mean of two different determinations, each done in triplicate.

Protein concentration This was estimated by the method of Bradford [12] using bovine serum albumin as standard. The protein concentration in the various fractions during purification was monitored by their absorbance at 280 mm using a 1 cm path length cuvette.

Purification of esterase I Esterolytic activity This was assayed using 3H-TAME as substrate [8]. A preparation of pig pancreatic kallikrein was standardized using unlabeled TAME as substrate [9], one esterase unit (EU) being defined as the concentration of enzyme that cleaves 1 /~mol of TAME per min at 25°C. An activity of 33.4 E U / m g protein was obtained for pig pancreatic kaUikrein. The radiometric assay was also performed using this standardized preparation and all results reported here are after conversion to EU.

Amidolytic activity This was assayed using Pro-Phe-Arg-MCA as substrate [10]. Substrate specificity was determined using the following fluorogenic peptides: (1) Boc-Val-ProArg-MCA (a-thrombin); (2) Glt-Gly-Arg-MCA

Urine from male dogs (15 1) was precipitated with 60% ammonium sulfate and chromatographed on a DEAE-Sephacel column (5 x 30 cm) in 0.01 M Tris-HC1 buffer, pH 8.0, containing 0.05 M NaC1. Bound proteins were eluted using a linear NaC1 gradient (0.05-0.8 M) in the same buffer, at a flow rate of 40 m l / h . Two peaks of esterolytic activity were observed. The fractions corresponding to the first esterolytic peak were pooled, dialyzed against water, lyophilized and used for further purification. This esterase was called esterase I because it appeared first on anion exchange chromatography; peak II was identified as urinary kallikrein [1]. The lyophilized proteins from peak I were taken in 30 ml 0.01 M Tris-HCl buffer, pH 9.0 and mixed with p-aminobenzamidine-CH-Sepharose (30 ml gel) equilibrated with the same buffer

278

containing 0.1% sodium azide. The mixture was shaken overnight at room temperature, packed into a column and washed with equilibrating buffer until A280 was less than 0.05. Weakly bound proteins were then eluted with 0.1 M phosphate buffer, p H 6.0, containing 1 M NaC1 at a flow rate of 15 m l / h . More strongly bound proteins were eluted with 0.1 M benzamidine dissolved in the same buffer. Esterase I was now eluted by increasing the benzamidine concentration to 0.5 M in pH 6.0 buffer. Three bed volumes of each eluent were used. The 0.5 M benzamidine eluate was dialyzed extensively against water and lyophilized. The lyophilized protein was dissolved in 0.05 M ammonium formate buffer, pH 8.0 and loaded on an Ultrogel AcA-54 column (1.5 × 90 cm) equilibrated in the same buffer. The column was developed at a flow rate of 6 m l / h and 1 ml fractions were collected. The esterolytic activity of individual fractions was assayed. The tubes corresponding to the peak of esterolytic activity were pooled and lyophilized.

Electrophoresis Discontinuous PAGE was done by the method of Davis [13] using the buffer system of Laemmli [14]. The electrophoresis was done in 0.75 mm gel slabs using 10% separating and 4% stacking gels. The gels were stained for protein by the method of Merrill et al. [15] using a Bio-Rad silver staining kit. For esterase activity, gels were stained by a modification of the method of Uriel and Borges [16]. Briefly, a 230 ml solution containing 1 mg of N-acetyl-DL-phenylalanine/3-naphthyl ester and 50 mg of o-anisidine in 0.1 M phosphate buffer, pH 6.8 was prepared immediately before use, and added to the gel on a flat container. After 15 min of incubation, the reaction was stopped with 7% acetic acid. SDS-PAGE was performed in 0.75 mm thick gel slabs using the discontinuous buffer system of Laemmli [14]. An acrylamide concentration of 12.5% for separating and 5% for stacking gels was used. After the run, the proteins were fixed in 40% methanol, 10% acetic acid and later stained with silver stain .as mentioned earlier. The molecular weight was calculated by the method of Weber and Osborn [17] using Pharmacia low molecular weight marker proteins as standards.

Amino acid analysis The amino acid composition was determined by reverse-phase HPLC of the acid-hydrolyzed protein as outlined by Koop et al. [18]. Cysteine residues were analyzed as cysteic acid in a duplicate run.

Carbohydrate analysis The carbohydrate portion of the glycoprotein was analyzed by the method of Perini and Peters [19]. The neutral sugars were separated after conversion to their corresponding glycamines, and the amino sugars were analyzed without modification.

Isoelectric focusing Preparative isoelectric focusing was done using an LKB-2117 Multiphor Unit. The electrofocusing was done in the pH range 3.5-5.0 using Ultrodex as support medium [20].

Plasminogen activator activity Plasminogen activating capability was determined by the method of Schumacher and Schill [21]. Plates were prepared with 1% agar and 1% fibrinogen containing plasminogen. Urokinase standard (1.0-10.0 units) (urokinase, low molecular weight, Calbiochem, San Diego, CA) or esterase I (40-100 ng protein) was added to the wells in the fibrin plate and incubated at room temperature for 48 h.

Clotting ability Fibrinogen (1%) in 0.1 M Tris-HC1 buffer, pH 8.0 was incubated with esterase I (10 ng) or athrombin (1 and 10 ng). The clotting time was determined both in the presence and absence of calcium (100 raM). Results

The purification of esterase I is summarized in Table I. A purification factor of 7210 was obtained after the final step. Anion exchange chromatography of the urinary proteins gave two peaks of esterolytic activity (Fig. 1A) with kininogenase activity associated with peak II, indicating that kallikrein was present in that peak. Peak I, which had very little kininogenase activity, was used for further purification.

279 TABLE 1 P U R I F I C A T I O N OF ESTERASE I Purification step

Total protein (A280)

Initial urine DEAE-Sephacel peak I p-Aminobenzamidine affinity (0.5 M benzamidine eluate) Ultrogel AcA-54

21 163 477

Specific activity (EU/A280)

4.4 0.35

Purification factor

0.1 5.6

1 56

54 721 a

540 7 210

Yield (%) 100 8.9 9.4

a Specific activity of purified esterase I was 601 E U / m g protein.

4.0 . A

J

Peak I

~

5.0

1.0

1.5

ak 2

E

o

cO

1.0~

OJ

8 2.0

0.5 t

c o

r'n

i

1.0

~~':~ 0.5~

---//

5~0

100

150

E~ 0.15

15 ill

I E o

It

0.10

Od

cO

"-r

.~ 0.05

I

5

"7I i

50

100

In the affinity chromatographic step, most of the kininogenase activity (about 95%) was eluted from the p-aminobenzamidine gel upon extensive washing with the equilibrating buffer. Lowering the pH to 6.0 and adding 1 M NaC1 to the buffer eluted the remaining kininogenase activity. Washing with 0.1 M benzamidine was included to elute a contaminant esterase which migrated faster than kallikrein in 10% PAGE. Elution with 0.5 M benzamidine resulted in elution of all of the TAME esterase activity bound to the gel. The affinity chromatographic step resulted in a 10-fold increase in specific activity. Although only very minor contaminants were observed upon gel filtration (Fig. 1B), this purification step increased the specific activity more than 10-fold. The procedure resulted in a yield of 9%, considering the initial separation of DEAE-Sephacel as step 1. The electrophoretic profile of the purified esterase I preparation is given in Fig. 2. Three protein bands, each having esterolytic activity, were observed. This profile is similar to that observed after the affinity chromatography step. Reduced SDS-PAGE of this preparation gave two closely migrating bands with molecular weights of 34000 and 33300. A mean molecular weight of 33 600 was used in all further calculations. A pI of 4.62 was calculated from preparative is•electric focusing.

150

Fraction Number Fig. 1. (A) DEAE-Sephacel chromatography of dog urinary proteins. Ammonium sulfate-precipitated urinary proteins were dissolved in 0.05 M Tris-HC1 buffer, pH 8.0, loaded on a DEAE-Sephacel column (5 x 30 cm) and eluted using a linear gradient of NaCl (0.05-0.8 M). (B) Gel permeation chromatography. The 0.5 M benzamidine eluate from the p-amino-

benzamidine affinity column was dialyzed, lyophilized, dissolved in 0.05 M ammonium formate buffer, pH 8.0 and gel-filtered on an Ultrogel AcA-54 column (1.5x90 cm). ...... , 3H-TAME esterase activity x l 0 - 7 ; • •, osmolality, osmol/kg x 10- 3.

280

be 9-times as potent as esterase when compared to dog urinary kallikrein; esterase I has a specific activity of 601 E U / m g protein as compared to a value of 69.3 E U / m g protein obtained for dog urinary kallikrein. The amino acid composition of esterase I is given in Table II. Carbohydrate analysis indicated the presence of mannose, galactose and glucosamine (Table II). The inhibition of the esterolytic activity of esterase I by various proteinase inhibitors is shown in Table III. Aprotinin, leupeptin, antipain, soybean and lima bean trypsin inhibitors, and Phe-Phe-Arg-chloromethyl ketone inhibited esterase I with 150 values in the 10-~-10 -9 M range, p-Aminobenzamidine, TLCK and PMSF were relatively weak inhibitors with 150 values in the 10-7-10 -5 M range. The kinetic properties of esterase I were studied TABLE II A M I N O ACID A N D CARBOHYDRATE COMPOSITION OF ESTERASE A Residue

Fig. 2. Electrophoresis of purified dog urinary esterase I (10 /~g), in 12.5% polyacrylamide gel electrophoresis. (A) Silver stain for proteins. (B) Esterase activity stain. Substrate used was N-acetyl-DL-phenylalanine/3-naphthyl ester.

Esterase I has a caseinolytic activity of 4.4 A280/nmol per h as compared to an activity of 9.1 A280/nmol per h obtained for bovine trypsin under identical conditions. Thus, esterase l has about 48% of the caseinolytic activity of bovine trypsin. The enzyme has a pH optimum of 9.5 for both esterolytic (3H-TAME) and amidolytic (Pro-PheArg-MCA) activities. Using partially purified dog kininogen as substrate, esterase I generated 1.5/zg bradykinin equiv./min per mg protein as compared to 3 mg bradykinin equiv./min per mg protein released by dog urinary kallikrein. This represents less than 0.1% of the kininogenase activity of dog urinary kallikrein. The esterolytic activity of esterase I and purified dog urinary kallikrein was compared using 3H-TAME as substrate. Esterase I was found to

Amino acid composition Asx Glx Arg His Ser Gly Thr Pro Ala Tyr Val Met Cys lie Leu Phe Trp Lys

Number/mol 39 42 14 6 30 38 13 17 22 12 4 1 1 18 26 9 17

total number of amino acid residues: molecular weight used for calculation: Carbohydrate composition mannose galactose glucosamine

3 2 4

total carbohydrate content = 4.5%

309 33600

281 T A B L E III INHIBITION ESTERASE A

OF

ESTEROLYTIC

ACTIVITY

OF

The reaction was carried out in 20 m M Tris-HCl buffer, p H 8.0, using 3 H - T A M E as substrate. Inhibitor

150

p-Aminobenzamidine Aprotinin Leupeptin Antipain Soybean trypsin inhibitor Lima bean trypsin inhibitor TLCK Phe-Phe-Arg-chloromethyl ketone PMSF

5.4-10 - 7 M 5.2.10 - 9 M 5.1.10 - 9 M 1.1.10- 8 M 5.6.10- 9 M 3.5.10 - 8 M 5.1.10 7 M 2.5-10- 9 M 3.0.10- 5 M

using six different fluorogenic substrates. The enzyme did not have any activity on the elastase substrate Suc-Ala-Pro-Ala-MCA or on the renin substrate S u c - A r g - P r o - P h e - H i s - L e u - L e u - T y r MCA. Among the fluorogenic peptides tested, Boc-Val-Pro-Arg-MCA was the best substrate, having the lowest K m and the highest k c a t / / g m . These results are shown in Table IV. Since a very high Vmax was obtained with GltGly-Arg-MCA, a urokinase substrate, the activity of urokinase on Boc-Val-Pro-Arg-MCA (10 /~M)

was compared with that of esterase I. Both enzymes were used at concentrations (1 / l g / m l ) that had similar enzymatic activity towards the urokinase peptide substrate. While esterase I released 0.1 nmol of aminomethylcoumarine per min of incubation, urokinase could not hydrolyze BocVal-Pro-Arg-MCA, even at a substrate concentration 5-times that used for esterase I. The ability of esterase I to clot fibrinogen was studied by using concentrations of thrombin and esterase I that had identical amidolytic activity on Boc-Val-Pro-Arg-MCA. In the presence of calcium, 10 ng of a-thrombin induced a clot in 3.3 min, and 1 ng thrombin in 6 rain. For esterase I (10 ng), the clotting time (9 min) was equivalent to that of control (saline). In the absence of calcium, thrombin (1 ng) induced a clot in 1 h. No clot was observed with esterase I (10 ng) even after 24 h of incubation. The plasminogen activator activity of esterase I was compared with that of urokinase. Distinct concentration-dependent halos were observed within 48 h around all the points of application of urokinase (10-100 ng). However, no halos were seen around esterase I (40 and 100 ng) even after 7 days of incubation. Esterase I (40 ng) did not alter the clotting time induced by thrombin (10 ng), or the appearances of halos induced by urokinase (50 ng).

T A B L E IV K I N E T I C P A R A M E T E R S O F ESTERASE I F O R D I F F E R E N T F L U O R O G E N I C P E P T I D E SUBSTRATES Each substrate is preferentially cleaved by the enzymes listed in parenthesis. The reaction was carried out in 20 m M Tris-HCl buffer (pH 8.0). The a m o u n t of enzyme used was 20 ng per assay tube (5.9.10 -1° M). Substrate

Km (gM)

Boc-Val-Pro-Arg-MCA (a-thrombin) 1.7 Glt-Gly-Arg-MCA (urokinase) 400.5 Boc-Gly-Lys-Lys-MCA (plasmin) 29.0 Pro-Phe-Arg-MCA (kallikrein) 41.4 Suc-AIa-Pro-AIa-MCA (elastase) Suc-Arg-Pro-Phe-His- Leu-VaI-Tyr-MCA (renin)

Vmax ( g m o l . m i n - 1-mg -1 )

kca t (s -1 )

kca t / K m ( s - l . t t M -1 )

12.45

7.09

4.30

77.00

43.88

0.11

9.75

5.55

0.19

8.85

3.33

0.08

no detectable activity

-

no detectable activity

-

282

Discussion A potent T A M E esterase that is different from urinary kallikrein has been purified with dog urine and characterized. The purification procedure involved anionic exchange chromatography, affinity chromatography on p-aminobenzamidineSepharose gel and a molecular sieving step using Ultrogel AcA-54. The first two steps were needed to separate the esterase from urinary kallikrein and also from another unidentified urinary T A M E esterase. The last step helped in increasing the specific activity 10-fold, probably by the removal of residual benzamidine from the affinity step. We arbitrarily named the purified enzyme esterase I because it eluted first on anionic exchange chromatography under the conditions used. Even though both dog urinary kallikrein and the esterase I are inhibited weakly by p-aminobenzamidine, the degree of inhibition of esterase I is relatively higher; its 150 was 5 . 1 0 - 7 M as compared to 4.7-10 -4 M for urinary kallikrein [1]. This difference in the 150 values helped in separating esterase I from kallikrein, since the latter was only weakly bound to the gel. We first attempted to use affinity chromatography on immobilized aprotinin, since this peptide was a good inhibitor of the non-kallikrein esterase activity in canine urine. Esterase I was tightly bound to the immobilized aprotinin, and drastic conditions such as acetate buffer pH 3.0, 6 M guanidine or 3 M KSCN had to be used to elute the enzyme from the gel. However, irreversible inactivation of the enzymatic activity was observed upon elution with these agents, suggesting denaturation of the enzyme. Thus, affinity chromatography on immobilized aprotinin was not used as a purification step. Using a p-aminobenzamidine-Sepharose column, the enzyme could be eluted under mild conditions (0.5 M benzamidine at pH 6.0) and found to retain its enzymatic activity. Esterase I is an acidic protein which has an isoelectric point of 4.62 with an optimum p H for proteolytic activity in the alkaline range. The data obtained on reduced SDS-PAGE and its carbohydrate content suggests that it is a single-chain glycoprotein. The purified esterase 1 was microheterogeneous, since three protein bands were observed in alkaline PAGE, all bands being enzymatically active.

There have been earlier reports on the presence of a non-kallikrein esterase in rat urine. This enzyme was first reported by Nustad and Pierce [3], who named it esterase A, and was later characterized by Chao [6]. McPartland et al. [5] further extended these studies and reported that rat urine contains at least two non-kallikrein TAME esterases, which they named A 1 and A 2. Esterase A and esterase A 2 a r e apparently the same enzyme. There are both similarities and differences between dog urinary esterase I and the rat nonkallikrein esterases. Dog esterase I gave three protein bands on PAGE similar to rat esterase A [6]. Similar microheterogeneity has been reported in the case of rat urinary esterase A 2, where four bands were seen [5]. Dog esterase I has a molecular weight of 33600 as compared to a molecular weight of 23000-24500 reported for rat esterase A and that of 41000 reported for rat esterase A 2. Dog esterase I is a single-chain enzyme while rat esterase A is composed of a heavy and a light chain linked by a disulfide bridge. Furthermore, dog esterase I has negligible kininogenase activity using canine kininogen substrate, unlike rat esterase A, which is an active kininogenase, or rat esterase A 2, which has half of the kinin-generating activity of rat glandular kallikrein, a potent kininogenase. Dog esterase I also is not a plasminogen activator, while rat esterase A is able to convert plasminogen to plasmin [6]. Thus, esterase I from dog urine appears to differ from rat urinary esterase A. Since we did not analyze urine from female dogs, we can only speculate about whether the enzyme described here is the canine equivalent of rat esterase A 1, an androgen-dependent enzyme [22]. Another non-kallikrein esterase was separated from kallikrein during the affinity chromatography step. It may be that this unknown esterase is similar to the rat esterase A. The isoelectric point, molecular weight and amino acid composition of dog esterase I do resemble those previously reported [23] for a canine acidic arginine esterase. However, this arginine esterase was not further characterized by these investigators. Esterase I is inhibited by a number of proteinase inhibitors. The inhibitory profile indicates that esterase I is a serine proteinase, and that it does not resemble urinary kallikrein. The trypsin-

283 like nature of esterase I is further suggested by its caseinolytic activity. The specific activity of the purified esterase I using TAME as a substrate was 601 U E / m g protein, which is similar to the value of 556 EU/A280 reported for rat urinary esterase A by Chao [6]. This value, however, is much higher than the TAME esterase activity of dog urinary kallikrein (69.3 E U / m g protein) and pig pancreatic kallikrein (33.4 E U / m g protein). Studies were done using different fluorogenic substrates to determine the substrate specificity of the canine esterase I. These studies revealed that peptide bonds involving basic amino acids at the carboxyl part of the bond were more susceptible than those involving neutral or aromatic amino acids. The best substrate for this enzyme was Boc-Val-Pro-Arg-MCA. This study also revealed the importance of the second amino acid in the peptide chain for proper hydrolysis. Even though arginine was present as the carboxyl terminal amino acid in each of these sequences, the peptides Glt-GIy-Arg-MCA and Pro-Phe-Arg-MCA were not such good substrates as Boc-Val-Pro-ArgMCA. Boc-Val-Pro-Arg-MCA is a good substrate for a-thrombin [10,24], and, as shown here, for dog urinary esterase I; however, esterase I could not clot fibrinogen. Thus, esterase I does not behave as thrombin when tested on a natural thrombin substrate, despite high affinity for an artificial thrombin substrate. Esterase I had a similar K m but a higher Vr,ax than urokinase for the fluorogenic peptide GltGIy-Arg-MCA, a very good urokinase substrate [10,24]. Hence, it was of interest to compare the activity of urokinase on the a-thrombin substrate Boc-Val-Pro-Arg-MCA. Urokinase did not cleave the peptide even at a substrate concentration of 50 /~M. These results further confirm that esterase I differs from urokinase, a-thrombin and glandular kallikrein, and is also different from the rat nonkallikrein esterase called esterase A by Nustad and Pierce [3] and by Chao [6]. Esterase I is a glycoprotein with potent TAME esterase activity. Approximately 22% of the total TAME esterase activity in initial urine was recovered at the end of the purification procedure. However, final proteins were less than 0.35 A280

units compared with more than 21 163 .4280 units of the initial urine. Although esterase I appears to be responsible for a substantial proportion of the TAME esterase activity present in dog urine, its concentration in terms of proteins is quite small, as suggested by the high purification factor obtained. The biological function and the source of the esterase present in dog urine are unknown. Esterase I did not cross-react with antisera against dog urinary kallikrein when tested by radioimmunoassay and double immunodiffusion (not shown). Thus, this enzyme is not similar to the multiple non-kallikrein esterases found in rat submandibular gland which are recognized by kallikrein antisera [25,26]. Attempts to determine whether esterase I entered the urine at either the proximal or distal nephron by using the stop-flow technique [27] were unsuccessful, possibly because esterase I is unstable in highly diluted urine. Further characterization of esterase I regarding cellular localization will require the production of specific antibodies. High-titer antibodies could not be obtained here because of the scarcity of immunogenic material. The results obtained with the artificial substrates indicate that esterase I cleaves peptide bonds containing basic amino acids, and that peptides containing the sequence Pro-Arg would be preferred as substrates over peptides having the sequences G l y - A r g or Phe-Arg. Using artificial peptide substrates, a kallikrein-like enzyme and a thrombin-like enzyme have been reported to be present in the neurointermediate lobe of the rat pituitary [28]. It is possible that dog esterase I and the rat pituitary thrombin-like enzyme are related. Further work will be needed to elucidate the sites of synthesis and storage of esterase I, its hormonal regulation, its natural substrate(s) and its physiological role.

Acknowledgements The authors thank Dr. Oscar A. Carretero, M.D. for his useful advice, and Dr. Fulvio Perini for helping with the amino acid and carbohydrate analysis. The excellent secretarial help of Miss Elizabeth G. Podgorny is appreciated.

284

References 1 Murthy, K.K., Carretero, O.A. and Scicli, A.G. (1986) Arch. Biochem. Biophys. 244, 563-571. 2 Matsuda, Y., Miyazaki, K., Fujimoto, Y., Hojima, Y. and Moriya, H. (1982) Chem. Pharm. Bull 30, 2512-2520. 3 Nustad, K. and Pierce, J.V. (1976) Biochemistry 13, 2312-2319. 4 Chao, J. and Margolius, H.S. (1980) Fed. Proc. 3, 852. 5 McPartland, R.P., Rapp, J.P., Joseph, M.K. and Sustarsic, D.C. (1983) Biochim. Biophys. Acta 762, 100-108. 6 Chao, J. (•983) J. Biol. Chem. 258, 4434-4439. 7 Chao, J., Ross, C., Shimamoto, K., Miller, D. and Margolius, H. (1982) Agents Actions Suppl. 9, 322-330. 8 lmanari, T., Kaizu, T., Yoshida, H., Yates, K., Pierce, J.V. and Pisano, J.J. (1976) Chemistry and Biology of the Kallikrein-Kinin System in Health and Disease (Pisano, J.J. and Austen, K.F., eds.), pp. 205-213, DHEW Publication No. (NIH) 76-791, Washington, DC. 9 Roberts, P.S. (1958) J. Biol. Chem. 232, 285-291. 10 Morita, T., Kato, H., lwanaga, S, Takada, K., Kimura, T. and Sakakibara, S. (1977) J. Biochem. 82, 1495-1498. I 1 Carretero, O.A., Oza, N.B., Piwonska, A., Ocholik, T. and Scicli, A.G. (1976) Biochem. Pharmacol. 25, 2265-2270. 12 Bradford, M.M. (1976) Anal. Biochem. 72, 248-254. 13 Davis, B.J. (1964) Ann. N.Y. Acad. Sci. 12I, 404-427. 14 Laemmli, U.K. (1970) Nature 227, 680-685. 15 Merrill, C.R., Goldman, D., Sedman, S.A. and Ebert, M.H. (1981) Science 211, 1437.

16 Uriel, J. and Borges, J. (1968) Nature 218, 578-580. 17 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406-4412. 18 Koop, D.R., Morgan, E.T., Tarr, G.E. and Coon, M.J. (1982) J. Biol. Chem. 257, 8472-8480. 19 Perini, F. and Peters, B.P. (1982) Anal. Biochem. 123, 357-363. 20 Winter, A., Perlmutter, H. and Davis, H. (1975) LKB Application Note 198. 21 Schumacher, G.F.B. and Schill, W.B. (1972) Anal. Biochem. 48, 9-26. 22 McPartland, R.P., Sustarsic, D.L. and Rapp, J.P. (1981) Endocrinology 108, 1634-1638. 23 Matsuda, Y., Fujimoto, Y., Akihama, S. and Moriya, H. (1984) Chem. Pharm. Bull. 32, 2371-2379. 24 Iwanaga, S., Morita, T., Kato, H., Harada, T., Adachi, N., Suga, T., Masuyama, I., Takada, K., Kinura, T. and Sakakibara, S. (1979) Kinins II - Biochemistry, Pathophysiology and Clinical Aspects (Fugii, S., Moriya, H. and Suzuki, T., eds.), pp. 147-163, Plenum Publishing Company, New York. 25 Bradtzaeg, P., Gautvik, K.M., Nustad, K. and Pierce, J.V. (1976) Br. J. Pharmacol. 56, 155-167. 26 Khullar, M., Scicli, G., Carretero, O.A. and Scicli, A.G. (1986) Biochemistry 25, 1851-1857. 27 Scicli, A.G., Carretero, O.A, Hampton, A., Cortez, P. and Oza, N.B. (1976) Am. J. Physiol. 230, 533-536. 28 Powers, C.A. (1986) J. Neurochem. 47, 145-153.