Eur. J. Biochem. 200, 563-571 (1991) 0 FEBS 1991

0014295691005702

A membrane-bound metallo-endopeptidase from rat kidney hydrolyzing parathyroid hormone Purification and characterization

Toru YAMAGUCHI ', Hiroshi KIDO', Masaaki FUKASE', Takuo FUJITA' and Nobuhiko KATUNUMA' ' Division of Enzyme Chemistry, Institute for Enzyme Research, The University of Tokushima, Japan ' Third Division, Department of Medicine, Kobe University School of Medicine, Japan

(Received April 22, 1991) - EJB 91 0525

A metallo-endopeptidase, which appears to be an integral membrane protein of rat kidney, was purified to homogeneity by a series of standard chromatographic procedures. This enzyme significantly hydrolyzed human parathyroid hormone [hPTH(l - 84)] and a synthetic substrate Suc-Leu-Leu-Val-Tyr-Mec (SUC = succinyl, Mec = 4-methyl-coumarinyl-7-amide). The purified enzyme had apparent molecular masses of 250 kDa on gel filtration, and 88 kDa and 245 kDa on sodium dodecyl sulfate/polyacrylamide gel electrophoresis under reducing and non- reducing conditions, respectively. Its pH optimum for activity was 8.0-8.5 and its isoelectric point was pH 4.9. Its activity was inhibited by EDTA, EGTA and o-phenanthroline, but not by phosphoramidon. The metal-depleted enzyme was reactivated by the addition of metal ions. The enzyme was also inhibited by chymostatin and eglin C, and by thiol compounds. Of the synthetic substrates examined, the enzyme hydrolyzed only Suc-Leu-Leu-Val- Tyr-Mec, one of the synthetic substrates for a-chymotrypsin. It did not hydrolyze synthetic substrates with less than four amino acid residues with tyrosine in the P1 position. The enzyme hydrolyzed hPTH and reduced hen egg lysozyme but did not hydrolyze azocasein or [3H]methyl-casein. NH2-terminal amino acid sequence analyses of the degradation products of hPTH( 1 - 84) and reduced hen egg lysozyme by the purified enzyme revealed that the enzyme preferentially cleaved these peptides at peptide bonds flanked by hydrophilic amino acid residues. Amino acid analyses showed that the main degradation products of PTH were hPTH(17 - 29), hPTH(30 - 38) and hPTH(74-84). The ability of the enzyme to hydrolyze peptide bonds flanked by hydrophilic amino acid residues and its inability to degrade azocasein distinguish it from several other kidney endopeptidases reported, such as endopeptidase 24.1 1 and meprin.

Circulating forms of endogeneous parathyroid hormone (PTH) are known to show heterogeneous immunoreactivity [l]. This heterogeneity was the result of proteolysis of PTH, which occurs mainly in the parathyroid gland, Kupffer cells in the liver and the kidney [2 , 31. Cathepsin B and cathepsin D in tissue homogenates from these organs have been shown to cleave PTH specifically into N-terminal and C-terminal fragments [4- 61, and these lysosomal acidic proteinases, es- pecially cathepsin D in Kupffer cells in the liver [7], have been considered to be important in the metabolism of PTH in vivo.

In contrast, several studies on PTH degradation in intact parthyroid [S, 91 and in kidney cells [lo] indicated the occur- rence ofnonlysosomal cleavage of PTH. In our previous study using cultured opossum kidney cells, which retain the charac-

Correspondence to T. Yamaguchi, Division of Enzyme Chemistry, Institute for Enzyme Research, The University of Tokushima, Tokushima, Japan 770

Abbreviations. hPTH, human parathyroid hormone; SUC, suc- cinyl; Boc, N-tert-butyloxycarbonyl; Z, benzyloxycarbonyl; Mec, 4- methyl-coumarinyl-7-amide; iPr2P-F, diisopropylfluorophosphate; Glt, glutaryl.

Enzymes. 5'-Nucleotidase (EC 3.1.3.5); cytochrome-c oxidase (EC 1.9.3.1); lactate dehydrogenase (EC 1.1.1.27); acid phosphatase (EC 3.1.3.2).

ter of proximal renal tubulus [ l l - 131, we found that this cleavage could be attributed to chymotrypsin-like activity in the cell membrane [lo].

For further examination of the enzymatic mechanism of non-lysosomal cleavage of PTH, we isolated a PTH-degrading enzyme from the plasma membranes of rat kidney using hu- man (h) PTH(1- 84) and Suc-Leu-Leu-Val-Tyr-Mec (SUC = succinyl, Mec = 4-methyl-coumarinyl-7-amide) as substrates. This enzyme differed from other known kidney membrane endopeptidases in substrate specificity, and cleaved PTH into biologically inactive fragments. Here we describe the purifi- cation and characterization of this endopeptidase.

MATERIALS AND METHODS

Materials

Kidneys were obtained from male Wistar rats weighing 200-250 g. The reagents used were as follows: Percoll, Q-Sepharose, arginine-Sepharose 4B, and Ampholines (Pharmacia LKB Biotechnology Inc, Uppsala, Sweden) ; 3-[(3- cholamidopropyl)dimethylammonio]-1 -propane sulfonic acid (Chaps) (Dojin, Kumamoto, Japan) ; various proteinase in- hibitors, hPTH(1- 84) and various synthetic substrates (The

5 64

Peptide Institute, Osaka, Japan); Leu-Val-Tyr-Mec, Suc-Leu- Tyr-Mec, Tyr-Mec, Asn-Mec and Gln-Mec (Bachem Fine Chemicals Inc, Budendorf, Switzerland) ; reduced hen egg lysozyme (Seikagaku Kogyo, Tokyo, Japan); [3H]methyl- casein and [3H]diisopropylfluorophosphate (iPr2P-F) (Amersham, Buckinghamshire, England). Eglin C was a gift from Dr H. Fritz (University of Munich, FRG). All other chemicals were of the highest grade available commercially.

Enzyme assays

Activity hydrolyzing hPTH(1- 84) was assayed by measuring the ability of the enzyme to convert hPTH(1- 84) to a form soluble in 5% trichloroacetic acid [lo, 141. The enzyme was incubated with substrate for 120 rnin at 37°C in a total volume of 500 pl 0.1 M Tris/HC1 pH 8.0 containing 25 ng hPTH(1- 84). Reactions were stopped by adding 500 pl 10% trichloroacetic acid (final concentration, 5%). Bovine serum albumin was added at 2.5 mg/ml as a carrier. The samples were kept at 4°C for at least 15 min and then centri- fuged for 10 rnin at 1800 x g . The supernatant containing hPTH fragments (i. e. acid-soluble products) was washed three times with diethyl ether to remove residual acid, and its hPTH content was then measured by radioimmunoassay with a chicken antibody that recognizes the fragments containing, at least, the amino acid sequence of residues 44- 68 in the hPTH molecule and intact hPTH as well [35]. One unit of hPTH(1- 84)-hydrolyzing activity was defined as the amount catalyzing the production of 1 ng acid-soluble immunoreactive PTH fragments/min.

The hydrolyses of synthetic Mec substrates by the enzyme were examined at pH 8.0 in 50 mM Tris/HCl buffer in a 500-pl reaction volume containing the enzyme and 100 pM substrates. In experiments with inhibitors, the enzyme was preincubated with inhibitors for 20 min at 37 "C before ad- dition of substrates. After incubation for 60 rnin at 37"C, the reaction was stopped by addition of 2 ml 5% acetic acid. The 7-amino-4-methyl-coumarin liberated from the substrate was then measured in a spectrophotofluorometer (Hitachi model 850) using excitation and emission wavelength of 370 nm and 460 nm, respectively. One unit of enzyme activity was defined as that catalyzing the degradation of 1 pmol synthetic substrate/min.

Activities for hydrolyzing azocasein and [3H]methyl-casein were assayed by the method of Butler et al. [I61 and Kirschner and Goldberg [17], respectively.

5'-Nucleotidase, a characteristic enzyme of plasma mem- branes, cytochrome c oxidase, a mitochondria1 marker, lactate dehydrogenase, a cytosol marker, and acid phosphatase, a lysosomal marker, were measured by the methods of Touster et al. [18], Wharton and Tzagoloff [19], Jaeniche and Knof [20] and Appelmans et al. [21], respectively.

Electrophoresis

Sodium dodecyl sulfate/polyacrylamide-gel electro- phoresis was carried out in a 4-20% gradient (mass/vol.) polyacrylamide slab gels (84 x 90 x 1.0 mm) by the method of Laemmli [22] and gels were stained with silver. Molecular masses were determined by comparison with the following standards (Bio-Rad, Richmond, CA) : phosphorylase b (92.5 kDa), bovine serum albumin (66.2 kDa), ovalbumin (45 kDa), carbonic anhydrase (31 kDa), soybean trypsin in- hibitor (21.5 kDa), and lysozyme (14.4 kDii).

Determination of isoelectric point

Isoelectric focusing was performed in tube gels (102 x 5 mm) containing acrylamide (3.9%, mass/vol.), meth- ylenebisacrylamide (0.12%, mass/vol.), glycerol (20%, by vol.), pH 3.5-10 Ampholines (1.6%, by vol.), and pH 2.5- 4.5 Ampholines (0.3%, by vol.). The gels were set in a standard tube-gel-electrophoresis chamber with an upper reservoir con- taining0.02 M NaOH and alower chamber containing 0.01 M H3P04. Samples of the purified enzyme (1 pg) were then ap- plied to the gels and electrophoresis was carried out at 200 V for 14 h and then at 400 V for 1 h at 4°C. After the pH gradient had been fully developed, the gels were cut into 3- mm sections. Each section was transferred to a test tube, 350 p1 ice-cold distilled water was added, and the pH value was measured at 4°C with a microelectrode and Beckman pH meter. For assay of proteolytic activity on Suc-Leu-Leu-Val- Tyr-Mec, the solution was adjusted to pH 8.0 by adding 100 pl 1 M Tris/HCl pH 8.0 and then the slices were incubated with 50 p1 1 mM synthetic substrate at 37°C for 5 h.

Determination of NH2-terminal sequences of products of hPTH [ 1 - 84) and reduced hen egg lysozyme liberated by the purified enzyme

hPTH(1-84) (21 pg) or reduced hen egg lysozyme (100 pg) was incubated with 1 pg of the purified enzyme in 1 mlO.1 M Tris/HCl pH 8.0 for 4 h a t 37°C. Then the samples were fractionated HPLC on a Cosmosil reverse-phase column (5C,-300; 150 x 4.6 mm, Nakarai Tesque, Kyoto, Japan) with a linear gradient of 0-40% acetonitrile in 0.1% trifluoroacetic acid during 80 min at a flow rate of 1 ml/min. Each degradation product was collected in a test tube and freeze-dried. The NH2-terminal sequences of these samples were determined in an Applied Biosystems model 470A gas- phase sequencer and model 120A HPLC analyzer system. Degradation products of hPTH(1- 84) were further subjected to amino acid analysis. Samples were hydrolyzed with 6 M HCl in evacuated, sealed tubes at 110 "C for 24 h. After evap- oration, the hydrolyzates were analysed in a Hitachi 835-50 model amino acid analyzer.

[3H]Di isopropy~uor~pho~phate labeling of the purified enzyme and gel electrophoresis

The purified enzyme (2 pg) was treated with [3H]iPr2P-F (740 kBq, 74 GBq/mmol) at 37°C for 10 h. The reaction was stopped by addition of unlabeled iPr2P-F (10 mM) and pro- teins were precipitated with cold acetone. The resulting pre- cipitate was collected by centrifugation and subjected to SDS/ PAGE (4 - 20% gradient). After electrophoresis, the gel was stained with Coomassie brilliant blue R250, destained, treated with Amplify (Amersham), dried and fluorographed with Fuji Rx film.

Preparation of antibody

Antisera against the rat enzyme was prepared by injecting a purified enzyme preparation into rabbits (New Zealand White). For this, the enzyme was emulsified with an equal volume of Freund's complete adjuvant and 2 x lo9 inactivated Bordetellu pertussis cells/rabbit and injected intracutaneously at a dose of 80 pg antigen/rabbit. From 3 weeks later, the rabbits were given subcutaneous booster injections of 50 pg of the purified enzyme in complete Freund's adjuvant, and

565

were bled 10 days after the last injection. IgG was prepared by the method of Johnstone and Thorpe [23].

Phuse separation of integral membrane proteins of rat kidney in Triton X - I 14 solution

The effect of phase-partitioning in Triton X-114 on the present enzyme was investigated according to the method of Bordier [24]. In brief, plasma membranes of rat kidney (200 pg) were treated with 200 pl 10 mM Tris/HCl pH 7.4, 150 mM NaCl, and 1 .O% Triton X-114 solution at 0 "C over- night. Integral membrane proteins were isolated from the hydrophilic proteins by phase separation above the cloud point of the detergent (30 "C). After centrifugation, a detergent phase containing amphiphilic integral membrane proteins and an aqueous phase containing hydrophilic proteins were obtained; Triton X-114 and buffer were added to each phase in order to yield equal volumes (500 pl) and approximately the same salt and surfactant content for both samples. Aliquots (20 pl) of each samples were analyzed together by Western immunoblotting using antibody against the purified enzyme.

Western inzmunohlotting

For Western blot analysis, the samples treated with Triton X-114 were subjected to electrophoresis on SDSjPAGE (4- 20% gradient) and transferred electrophoretically to nitrocel- lulose; excess sites were blocked with 1 % bovine serum albu- min. The membrane was then incubated for 12 h at room temperature with 1 : 100 dilution of antibody against the purified enzyme (15.1 mgiml). Bound antibodies were detect- ed by the anti-rabbit alkaline phosphatase method (Strata- gene) according to the manufacturer's instructions. The pre- stained SDSjPAGE low-range standards (Bio-Rad) used as molecular mass markers were phosphorylase h (106 kDa), bovine serum albumin (80 kDa), ovalbumin (49.5 kDa), car- bonic anhydrase (32.5 kDa), soybean trypsin inhibitor (27.5 kDa) and lysozyme (18.5 kDa).

Protein determinution

Protein was measured with bovine serum albumin as a standard using a BCA protein assay kit (Pierce, Rockford, IL) according to the method of Redinbaugh and Turley [25].

RESULTS

Purification of membrane-bound proteinuse

Plasma membranes were purified from adult rat kidney on a Percoll gradient by the methods of Belsham et al. [26] and Record et al. [27]. Briefly, tissue homogenates were centrifuged at 5000 x g for 10 min. The pellet, containing nuclei, heavy mitochondria and plasma membranes, was dispersed in iso- osmotic Percoll solution and centrifuged at 10000 x g for 1 h. Then the clear band of membranes just below the surface was collected, washed, and used as purified plasma membranes. The purity of the renal plasma membranes used in this study was assessed by measuring the specific activities of 5'-nucleo- tidase, a marker enzyme for plasma membranes, as well as cytochrome c oxidase, lactate dehydrogenase, and acid phos- phatase. The specific activity of 5'-nucleotidase in the plasma membranes was increased 11.55-fold over that in the homogenate, while the specific activities of cytochrome c oxi- dase, lactate dehydrogenase, and acid phosphatase were in-

creased 2.23-fold, 0.01-fold and 0.10-fold, respectively (data not shown).

The enzyme hydrolyzing both hPTH and Suc-Leu-Leu- Val-Tyr-Mec appeared to be an integral membrane protein (as described below) and it could be extracted with 1 % Chaps or 4% octyl glucoside. Therefore, the plasma membranes were suspended in 100 ml 10 mM Tris/HCl pH 7.0 containing 1% Chaps, stirred overnight and centrifuged for 1 h at 25 000 x g. Then (NH4)2S04 was added to the resulting supernatant at 20% saturation (4°C) with stirring for 20 min and the mixture was centrifuged at 19000 x g for 20 min. The enzyme was precipitated from the supernatant by further addition of (NH4)2S04 at 60% saturation. The precipitate was dissolved with a small volume of 10 mM Tris/HCl pH 7.0, dialyzed against 10 mM Tris/HCl pH 7.0 overnight and applied to a Q- Sepharose column (2.0 x 4.2 cm), which had previously been equilibrated with the dialysis buffer. The column was washed with the same buffer until the absorbance of the eluate at 280 nm had returned to the base line, and then the enzyme was eluted with 240 ml of a linear gradient of 0 - 0.5 M NaCl in 10 mM Tris/HCI pH 7.0 (Fig. 1A). Using Suc-Leu-Leu- Val-Tyr-Mec as a substrate, major enzyme activity was eluted with about 280mM NaCl and a small peak with about 200 mM NaCl. When assayed in the presence of 100 pM of both phosphoramidon and amastatin, the latter activity was completely suppressed. The major enzyme activity, which was not affected by either inhibitor (data not shown), was submit- ted to further purification. The active fractions were pooled, dialyzed against 10 mM Tris/HCl pH 7.0 overnight and con- centrated to 1 ml by ultrafiltration in a Diaflo YM-10 mem- brane (Amicon). Then the enzyme solution was fractionated by HPLC on a TSKgel DEAE-5PW column ( 7 5 ~ 7 . 5 mm, Toyo Soda) with a linear gradient of 0-0.25 M NaCl in 10 mM Tris/HCI pH 7.0 in 60 min at a flow rate of 1 ml/min, as shown in Fig. 1B. The active fraction eluted with about 150 mM NaCl was collected, concentrated to 1 ml by ultrafiltration and fractionated by HPLC on double-linked TSK-Gel G3000SW columns (600 x 0.75 mm each, Toyo Soda, Tokyo, Japan) in 50 mM sodium acetate pH 5.5 con- taining 0.2 M Na2S04 and 0.6% Chaps at a flow rate of 0.4 ml/min (Fig. 1 C). The active fractions were pooled, dialyzed against 50 mM sodium phosphate pH 7.5 containing 1.8 M (NH4)2S04 overnight, and fractionated by HPLC on a hydrophobic interaction column (Biofine 75 x 7.5 mm, Jasco, Hachiohzi, Japan) with a decreasing linear gradient of 1.8 - OM (NHJ2S04 in 50mM sodium phosphate pH 7.5 in 60 min at a flow rate of 1 ml/min, as shown in Fig. 1 D. The active fraction eluted with about 720 mM (NH4)2S04 was collected, dialyzed against 10 mM Tris/HCI pH 7.5 overnight and applied to an arginine-Sepharose 4B column (1.0 x 3.0 cm), which had been equilibrated with the dialysis buffer. The column was washed with the same buffer and then the enzyme was eluted with 200 ml of a linear gradient of 0-0.1 M NaCl in 10 mM Tris/HCI pH 7.5, as shown in Fig. 1 E. The enzyme was eluted with about 65 mM NaC1. Aminopeptidase N ac- tivity, which had been co-purified with the enzyme until this step, was eluted earlier than the enzyme and clearly separated from it (data not shown). At all steps, the activity hydrolyzing Suc-Leu-Leu-Val-Tyr-Mec was inseparable from the hPTH- hydrolyzing activity, and no other hPTH-hydrolyzing activity was detected. The purified enzyme was collected, concentrated by ultrafiltration and stored at - 20 "C for characterization.

The results of a typical purification of the enzyme are summarized in Table 1. Starting with 295 g kidney from 160 male Wistar strain rats, we obtained 3.4 pg purified enzyme

566

3i 2.

ad .- A Q-Sepharose -_

UI

Fractlon Number (5 ml/tube)

t Fractlon Number ( 0 . 8 mi/tube)

B TSKgei DEAE-5PW

Fraction Number ( 2,ml/tube)

Fraction Number ( 2 mlltube)

Fig. 1 . Column-chromatographic steps in the purifcation of the enzyme f rom rat kidney. Details of the procedure are described in the text. (A) Q-Sepharose. (B) TSKgel DEAE-5PW (C) TSK-Gel G3000SW. (D) Hydrophobic interaction HPLC column (Bio-fine). (E) Arginine-Sepharose 4B. Bars at the top indicate the fractions pooled. (-) (protein A I B 0 ) ; (-- - -) gradient; ( 0 ) activity with Suc-Leu-Leu-Val-Tyr-Mec; (0) activity with hPTH(1-84)

Table 1. Purification of the metalloproteinuse f rom rut kidney membranes The enzyme was purified from 295 g kidney from 160 rats

Fraction Suc-Leu-Leu-Val-Tyr-Mec hydrolysis hPTH(1- 84) hydrolysis

protein total specific purifica- yield total specific purifica- yield activity activity tion activity activity tion

Pg PU

Plasma membranes 62803 141 077 1 % Chaps extract 15 829 74 679 20 - 60% saturated (NH4)* SO4 precipitation 12130 61 236 Q-Sepharose 4 598 24969 TSKgel DEAE-5PW 1311 8 904 TSK-Gel G3000SW 198 4253 Bio-fine 9.9 3159 Arginine-Sepharose 4B 3.4 1527

@img

2 246 4718

5 048 5430 6 792

21 479 319090 449118

-fold O/'o U Uimg

1 100 256 4.1 2.1 53 110 7.0

2.2 43 129 31 2.4 18 70 15 3.0 6.3 28 21 9.6 3.0 16 80

142 2.2 10 1037 200 1 .I 4.3 1259

-fold YO

100 1.7 43

2.6 51 3.1 27 5.1 11

19 6.1 253 3.9 307 1.7

567

A B 1 2 3 1 2

92.5

66.2

-106.0

-80.0

45.0 -49.5

31 .O

21.5 14.4

-32.5

-27.5

-18.5

Fig. 2. SDSjPAGE of the puriyied enzyme and phase-separated mem- brane proteins. (A) The purified enzyme (0.5 Fg) shown in Fig. 1 E was boiled for 5 min with 1% (mass/vol.) SDS in the presence (lane 2) or absence (lane 3) of 1% (by vol.) 2-mercaptoethanol. Samples were then subjected to SDSjPAGE and silver-staining. Gels were calibrated with phosphorylase b (92.5 kDa), bovine serum albumin (66.2 kDa), ovalbumin (45 kDa), carbonic anhydrase (31 kDa), soy- bean trypsin inhibitor (21.5 kDa), and lysozyme (14.4 kDa) (lane 1). (B) Phase separation of integral membrane proteins of rat kidney in Triton X-114 solution. Plasma membranes from rat kidney were mixed with Triton X-I 14 and submitted to phase separation at 30°C. Aliquots of the obtained detergent and aqueous phases were analysed together by Western immunoblotting under reducing conditions using an antibody against the purified enzyme as described under Materials and Methods. Gels were calibrated with prestained SDSjPAGE low- range standards (Bio-Rad). Lane 1, detergent phase; lane 2, aqueous phase. Number on the side show standard molecular masses in kDa

(1.1 O h yield) and achieved 200-fold purification from the step of plasma membrane extraction, as judged by assay with Suc- Leu-Leu-Val-Tyr-Mec as substrate. With hPTH(1- 84) as substrate, the yield was 1.7% with 307-fold purification.

Properties ofthe purified enzyme

Size. The molecular mass of the native enzyme was deter- mined by gel filtration on double-linked TSK-Gel G 3000 SW columns in the presence of 0.6% Chaps (Fig. 1 C). Removal of the detergent did not affect the elution volume of either the enzyme or the standards. The molecular mass of the enzyme was calculated to be 250 kDa. The standards used were bovine thyroglobulin (670 kDa), bovine y-globulin (1 58 kDa), chicken ovalbumin (44 kDa), horse myoglobin (17 kDa), and vitamin BI2 (1.35 kDa) (Bio-Rad). The purified enzyme (0.5 pg) was subjected to SDSjPAGE and silver-staining (Fig. 2A). The molecular masses of the protein band were 88 kDa and 245 kDa under reducing and non-reducing con- ditions, respectively.

Membrane nature of the enzyme. To clarify the membrane nature of the enzyme, the effect of phase-partitioning in Triton X-114 on this proteinase was examined. Plasma membranes from rat kidney were treated with Triton X-114 and separated into detergent and aqueous phase. Each phase was analysed by Western immunoblotting under reducing conditions using antibody against the purified enzyme. The 88-kDa band co- inciding with the purified enzyme was strongly stained in the detergent phase but not in the aqueous phase (Fig. 2B). Thus

1 0 1 10 20

Slice number

Fig. 3. Isoelectric focusing of the proteinase. Isoelectric focusing was performed at 4°C in tube gel containing 1.6% (by vol.). Ampholines of pH 3.5-10 and 0.3% Ampholines (by vol.) of pH 2.5-4.5 as described in Materials and Methods. Proteolytic activities in slices of gel with Suc-Leu-Leu-Val-Tyr-Mec ( 0 ) as substrate and the pH values (0) were determined. Proteolytic activities were expressed as percent- ages of the maximum activity observed in slice 11

Table 2. Activities of the purified enzyme on synthetic substrates Values are activities of the purified enzyme as percentages of that with Suc-Leu-Leu-Val-Tyr-Mec. Substrate concentration was 100 pM

Substrate Relative activity

Suc-Leu-Leu-Val-Tyr-Mec Leu-Val-Tyr-Mec Suc-Leu-Tyr-Mec Tyr-Mec Suc- Ala- Ala-Pro-Phe-Mec Glt-Phe-Mec Z-Phe- Arg-Mec Z- Arg-Arg-Mec Boc-Val-Leu-Lys-Mec Boc-Glu-Lys-Lys-Mec Sue- Ala- Ala- Ala-Mec Suc- Ala-Pro- Ah-Mec Suc-Gly-Pro-Mec Leu-Mec Asn-Mec Gln-Mec

Yo

100.0 5.4 9.9 2.6 2.0 2.2 0.3 2.8 1.6 2.4 3.3 2.8 2.3 3.1 0.2 0.2

the enzyme was shown to retain the nature of integral mem- brane proteins.

Dependence onpH. The pH activity curve for the hydrolysis of Suc-Leu-Leu-Val-Tyr-Mec was examined using 100 mM concentrations of sodium acetate pH 3.0 - 5.0, potassium phosphate pH 6.0- 7.5, borate pH 8.0- 9.5 and Tris/HCI pH 10.0. The optimum pH with the substrate was 8.0-8.5 (graphic data not shown).

Isoelectric point. When the purified enzyme was subjected to isoelectric focusing, the enzyme activity was detected in a symmetrical peak at pH 4.9 (Fig. 3).

Substrate specificity. The activities of the purified enzyme on various synthetic peptide derivatives are shown in Table 2.

568

/ E I i g ul N d r

% I 1 I I

Retention Time (min) 0.64 1

E i B 7 n

Retention Time (min)

Fig. 4. HPLC analysis of the fragments of hPTH(1-84) ( A ) and reduced lysozyrne ( B ) produced by the purified enzyme. hPTH(1- 84) (21 pg) and reduced hen egg lysozyme (100 pg) were subjected to HPLC after incubation with the enzyme in 1 ml 0.1 M Tris/HCl pH 8.0 for 4 h at 37°C. Arrows indicate the retention times of the intact forms. The NH2-terminal amino acid sequences of the fragments, designated as a-i and 1 - 7, were determined in an automated gas- phase sequencer

Of the compounds tested, Suc-Leu-Leu-Val-Tyr-Mec was the best substrate. Compounds with three or less amino acid resi- dues with tyrosine in the P1 position, such as Leu-Val-Tyr- Mec, Suc-Leu-Tyr-Mec and Tyr-Mec, and compounds with phenylalanine in the P1 position, such as Suc-Ala-Ala-Pro- Phe-Mec and Glt-Phe-Mec (Glt = glutaryl) which are all good substrates for a-chymotrypsin (data not shown), were scarcely hydrolyzed by the enzyme. These results suggest that the substrate specificity of the enzyme is distinct from that of chymotrypsin-type proteinases, and that the enzyme requires more than three amino acid residues for hydrolytic activity. The enzyme scarcely hydrolyzed substrates of trypsin-like proteinases, elastase, prolyl-endopeptidase and leucyl aminopeptidase. We also analysed the NH2-terminal amino acid sequences of the major degradation products of hPTH(1- 84) and reduced hen egg lysozyme produced by the purified enzyme after their separation by reverse-phase HPLC as described in Materials and Methods. As shown in Fig. 4A and B, the retention times of intact hPTH and reduced lysozyme were 64 min and 69 min, respectively; the intact forms of both these compounds were completely hydrolyzed by the enzyme during incubations for 4 h. In both cases, the degradation products were eluted before the intact forms. The

NH,-terminal amino acid sequences of the products, shown in Table 3, indicated that the major sites of cleavage were His9-Asnl0, Asnl6-Serl7, Gln29-Asp30, Gly38-Ala39, Ala73-Asp74 and Ser77-Met78 in hPTH(1- 84) and Tyr23- Ser24, Gly26-Asn27, Ala42-Thr43, Ser91 -Va192, Ser100- Asp101 and Gly137-Thr118 in reduced hen egg lysozyme. Amino acid analyses of the three major degradation products of hPTH, designated as peaks b, d and g in Fig. 4A, revealed that these fragments were hPTH(17 -29), hPTH(30- 38) and hPTH(74- 84), respectively (Table 4). These results suggest that the purified enzyme has a preference for peptide bonds that are flanked by hydrophilic amino acid residues. The en- zyme did not hydrolyze several other proteins tested, such as azocasein and [3H]methyl-casein, during incubation for up to 5 h (data not shown).

Inhibitors. Table 5 shows the effects of various inhibitors and other compounds on the activities of the purified enzyme. The enzyme activity was markedly inhibited by chelating agents such as EDTA, EGTA, and o-phenanthroline, sug- gesting that it is a metallo-endopeptidase, but was not inhibit- ed by phosphoramidon. The enzyme was also inhibited by chymostatin and eglin C, inhibitors of chymotrypsin-type ser- ine proteinases, and by thiols, such as dithiothreitol and re- duced glutathione. However, the enzyme was not inhibited by inhibitors of trypsin, elastase, aminopeptidase, aspartic or cysteine proteinases.

[3H]iPr2P-F-labeling of the enzyme. Since the purified enzyme partially resembled a seryl chymotrypsin-like enzyme in being inhibited by chymostatin and eglin C, we investigated whether this purified enzyme was labeled by [3H]iPr2P-F as described in Materials and Methods. A fluorogram of the gel after SDSjPAGE of [3H]iPr2P-F-treated samples showed no radioactive band at a position coinciding with the enzyme, indicating that the enzyme is clearly not a serine proteinase (data not shown). This assumption was supported by the finding that this enzyme activity was not inhibited by iPr2P- F as shown in Table 5 .

Metal requirement. Treatment of the purified enzyme with 10 mM EDTA followed by extensive dialysis reduced the en- zyme activity to 7.9% of that without EDTA treatment. The inactive apoenzyme was reactivated by cations in the following order: Zn2+ > Mn2+ > Co2+ > Ca2+ (Table 6). Thus, the enzyme probably contains a metal ion, that is essential for its activity.

DISCUSSION

In the present study, a metallo-endopeptidase hydrolyzing both hPTH(1- 84) and Suc-Leu-Leu-Val-Tyr-Mec was purified from renal plasma membranes and characterized. No other enzyme hydrolyzing hPTH was detected in the extract of renal plasma membranes at neutral pH. Moreover, the enzyme activity in crude preparations seems to be attributed to the purified enzyme itself, but not to other coexisting pro- teinases, such as endopeptidase 24,ll and aminopeptidase, since the activity in fractions from Q-Sepharose chrorna- tography was unchanged irrespective of the presence of phos- phoramidon or amastatin (see Results). The purified enzyme appears to have an oligometric structure. It has a molecular mass of 250 kDa as judged by gel filtration, and the molecular mass of its subunits is 88 kDa as judged by SDSjPAGE under reducing conditions. In non-reducing conditions, it gave a single band with a molecular mass of 245 kDa. The enzyme showed the properties of a metallo-endopeptidase: it was sen-

569

Table 3. NH2-terrninal amino acid sequences of degradation products of hPTH and reduced lysozyme The NH2-terminal sequences of major fragments of hPTH(1-84) (peaks a- i in Fig. 4A) and reduced hen egg lysozyme (peaks 1-7 in Fig. 4B) were determined in an automated gas-phase sequencer as described in Materials and Methods

Protein ~~

Peak N H 2-terminal sequence Cleavage site

hPTH a b

d e f g h i

1 2 3 4 5 6 7

C

Lysozyme

Ala-Pro-Leu- Ala-Pro- Asp-Val- Asn-Val-Leu-Thr- Ser-Val-Ser-Glu-Ile-Gln- Asp-Val-His- Asn-Phe- Ala-Pro-Leu-Ala-Pro- Arg-Asp- Met-Glu-Arg-Val-Glu-Trp-Leu- Ser-Met-Glu- Arg-Val-Glu-Trp-Leu- Ser-Val-Ser-Glu-Ile-Gln-Leu- Asn-Leu-Gly-Lys-His-Leu- Asn-Ser-

Val- Asn-Cys-Ala-Lys-Lys-Ile-Val- Thr- Asn- Arg-Asn-Thr-Asp-Gly-Ser-Thr-Asp- Asn-Trp-Val-Cys-Ala- Ala-Lys-Phe-Glu- Ser-Leu-Gly - Asn-Trp-Val-C ys- Ala- Ala-Lys-Phe- Thr-Asp-Val-Gln-Ala-Trp-Ile- Arg-GI y-C ys- Arg- Asp-Gly-Asp-Gly-Met-Asn- Ala-Trp-Val- Ala-Trp- Lys-Val-Phe-Gly-Arg-Cys-Glu-

Gly38-Ah39 Ala73-Asp74 NH2-terminal Gln29-Asp30 Gly38-Ala39 Serl7-Metl8 Asnl6-Serl7 NH2-terminal His9-Am10

Ser91 -Val92 Ala42-Thr43 Gly26-Am27 Tyr23-Ser24 Glyll7-Thrl18 Serl 00-Asp1 01 NH2-terminal

Table 4. Amino acid compositions of peaks b, d and g compared with the theoretical amino acid compositions of hPTH( 74 - 84), hPTH(30 - 38) and hPTH(17-29), respectively Peak fractions b, d and g were obtained after incubation of hPTH(1-84) with the purified enzyme (Fig. 4A). These peak fractions were prepared for amino acid analysis as described in Materials and Methods

Amino acid Peak b hPTH Peak d hPTH Peak g hPTH (74 - 84) (30-38) (17-29)

residues/molecule

Asp/Asn Ser Thr Glu/Gln GlY Ala Val Leu Phe LYS Arg

2.0 1.1 0.8 1.2 0.2 0.8 1.9 1 .o 0.3 2.1 0

2 1 1 1 0 1 2 1 0 2 0

2.1 0.2 0 0.2 1.3 0.9 1.7 1 .o 1 .o 0.2 0

2 0 0 0 1 1 2 1 1 0 0

0 1.2 0 2.9 0.3 0 1 .o 2.1 0.6 2.1 1.7

0 1 0 3 0 0 1 2 0 2 2

sitive to metal-ion-chelating agents, and the EDTA-treated apoenzyme was reactivated by metal ions. In addition, it had some properties of a chymotrypsin-type proteinase: it prefer- entially hydrolyzed Suc-Leu-Leu-Val-Tyr-Mec and was in- hibited by chymostatin and eglin C (Table 5). However, the purified enzyme is obviously not a serine proteinase since it was insensitive to phenylmethylsulfonyl fluoride or iPr2P-F (Table 5 ) and was unlabeled by [3H]iPr,P-F. In addition, the enzyme was clearly distinguishable from a-chymotrypsin in that it did not hydrolyze substrates containing three amino acids or less with tyrosine in the P1 position or those with phenylalanine in the PI position.

The purified enzyme was shown to be sensitive to both thiol reagents (dithiothreitol and reduced glutathione) and chelating agents (EDTA, EGTA, and o-phenanthroline) (Table 5). These sensitivities partly resemble other metallo- proteinases, such as endopeptidase 24.1 5 [28] and Pz-peptidase

[29], both of which were recently proved to be identical [30]. However, endopeptidase 24.1 5/Pz-peptidase has a smaller mo- lecular mass of about 70 kDa [28, 291 and is distinct from the present enzyme.

In addition, the purified enzyme was clearly distinguish- able from endopeptidase 24.11, which is the most readily iden- tified and well-characterized kidney membrane endopeptidase [31], in its insensitivity to phosphoramidon. Another mem- brane endopeptidase in mouse kidney, meprin [16, 321, which has also been named ‘endopeptidase-2’ in rats [33, 341, is also insensitive to phosphoramidon, and requires distinction from the present enzyme. Like the present enzyme, rat endo- peptidase-2 has an oligomeric structure, the molecular mass of its subunits is 80 kDa and the molecular mass of its native form appears to be 220 kDa on SDSjPAGE under non-reduc- ing conditions [34]. However, substrate specificity is promi- nently different between meprin/endopeptidase-2 and the

5 70

Table 5. Inhibitor sensitivity of the purified enzyme The purified enzyme (50 ng) was preincubated with each compound at the indicated concentration for 20 min at 37'C. Then the activity with Suc-Leu-Leu-Val-Tyr-Mec was assayed as described in Materials and Methods. Relative activity was calculated as a percentage of that of the enzyme without inhibitor

Addition Conc. Relative activity

mM None Phenylmethylsulfonyl fluoride 20 Diisopropylfluorophosphatc 20 Soybean trypsin inhibitor 0.1

Antipain 0.1 Leupeptin 0.1 Eglin C 0.1 Chymostatin 0.1 Elastatinal 0.1 Ep475 0.1 E64 0.1 Pepstatin A 0.1 Bestatin 0.1 EDTA 0.1

EGTA 0.1

o-Phenan throline 0.1

Bowman-Birk soybcan protcinase inhibitor 0.1

1

1

1 Phosphoramidon 0.1

Reduced glutathione 0.1

Dithiothrcitol 0.1 1

1

Yo 100 102 110 116 113 89 94 51 16

137 100 116 96 83 52 16 56 16 61 18 91 28

3 18 3

Table 6. Abilities of divalent metul ions to restorC uctivity of metal- depleted enzyme The purificd enzyme (3 pg) was incubated at 37°C for 20 min in 2 ml 10 mM Tris/HCI pH 8.0 containing 10 mM EDTA. The EDTA was then removed by dialysis against 10 mM Tris/HCl pH 8.0, for 5 h with a change of the buffer once an hour at 4°C. The activities of aliquots of the enzyme (50 ng) were then assayed in the absence and presence of the indicatcd concentrations of mctal ions with Suc-Leu- Leu-Val-Tyr-Mec as substrate, as described in Materials and Methods. Relative activity was calculated as a percentage of the activity of the enzyme without EDTA treatment

Metal Relative activity at

0 1 pM 10pM 100 pM 1000 pM

%

ZnS04 7.9 4.5 21.4 23.9 36.0 MnCl, 7.1 13.6 16.7 39.6 COCI, 5.3 7.2 15.2 19.9 CaCl, 5.6 1.2 6.7 23.2

purified enzyme. Meprin/endopeptidase-2 efficiently degrade azocasein [I 6, 331 and preferentially hydrolyze peptides be- tween hydrophobic amino acid residues [16, 351, whereas the present enzyme showed no activity towards azocasein and had a preference for hydrophilic amino acids in peptides (Table 3). These findings make the present enzyme distinct from other

The purified enzyme cleaved hPTH(1- 84) into many frag- ments (Fig. 4A). The cleavage sites and amino acid sequences of the major products were determined (Tables 3 and 4). The results showed that three major degradation products with sequences of 9 - 13 amino acid residues were generated by cleavage of the bonds between Asnl6-Serl7, Gln29-Asp30, Gly38-Ala39 and Ala73-Asp74. Since PTH needs at least the continuous peptide sequence from Val2 to Arg25 to exert its biological effects on bone and kidney (361, cleavage of the Asnl6-Serl7 bond would completely abolish its biological activity. Hence these data suggest that the physiological role of the purified kidney plasma membrane proteinase in PTH metabolism, if any, is in inactivating the circulating hormone rather than in converting it into an active form. Indeed, kidney has generally been accepted to play a major role in the metab- olism of PTH by inactivating the hormone and removing it from the circulation [2, 31. Thus this endopeptidase appears to be a potential candidate for the PTH-degrading enzyme acting in the kidney in vivo.

However, this enzyme does not seem to be specific for hPTH because it also hydrolyzed reduced hen egg lysozyme (Fig. 4B). Therefore the possibility that this enzyme might also participate in the metabolism of other hormones or bio- logically active substances requires further study.

We thank Dr K. Tanaka (Division of Enzyme Pathology, Institute for Enzyme Research, the University of Tokushima) and Dr Y . Nishii (Research Laboratories of Chugai Pharmaceutical Co. Ltd, Tokyo) for useful suggestions. This study was supported by a grant from the Yamanouchi Foundation for Research on Metabolic Disorders and research grants from the Japanese Ministry of Education, Science and Culture .

REFERENCES 1. Berson, S. A. & Yalow, R. S. (1968) J . Clin. Endocrinol. Metub.

2. Martin, K. J., Hruska, K. A,, Freitag, J . J., Klahr, S. & Slatopolsky, E. (1979) N . Engl. J . Med. 301, 1092-1098.

3. Armitage, E. K. (1986) Clin. Chem. 32,418-424. 4. Hamilton, J. W., Jilka, R. L. & MacGregor, R. R. (1983) Endo-

crinology 113, 285-292. 5. MacGregor, R. R., Hamilton, J. W., Kent, G. N., Shotstall, R.

E. & Cohn, D. V. (1979) J . Biol. Chem. 254,4428-4433. 6. Botti, R. E. & Zull, J. E. (1983) Endocrinology 112, 393-395. 7. Diment, S., Martin, K. J. & Stahl, P. D. (1989) J . Biol. Chem.

8. MacGregor, R. R., Jilka, R. L. & Hamilton, J. W. (1986) J . B id .

9. MacGregor, R. R., McGregor, D. H., Ridgeway, R. D., Lee, S.

10. Yamaguchi, T., Fukase, M., Nishikawa, M., Fujimi, T. & Fujita,

11. Teitelbaum, A. P. & Strewler, G. J. (1984) Endocrinology 114,

12. Pollock, A. S., Warnock, D. G. & Warnock, G. J. (1986) Am. J .

13. Caverzasio, J., Rizzoli, R. & Bonjour, J.-P. (1986) J . Biol. Chem.

14. Yamaguchi, T., Baba, H., Fukase, M., Kinoshita, Y., Fujimi, T. & Fujita, T. (1986) Biochem. Biophys. Res. Commun. 141,

15. Fukase, M., Fujitd, T., Matsumoto, T., Ogata, E., Iijima, T., Takezawa, J., Saito, K., Ishige, H. & Fujimoto, M. (1989) Folia Endocrinol. Jpn 65, 801 - 827.

16. Butler, P. E., McKay, M. J. & Bond, J . S. (1987) Biochem. J . 241,

17. Kirschner, R. J. & Goldberg, A. L. (1983) J . Biol. Chem. 258,

28, 1037-1047.

264,13403-13406.

Chem. 261, 1935-1940.

H. & Hamilton, J. W. (1986) Bone Mia. 1,41-50.

T. (1988) Endocrinology 123,2812-2817.

980-985.

Physiol. 250, F22 7 - F225.

261, 3233 - 3237.

162 - 768.

229-2235

known kidney membrane endopeptidases. 967 - 976.

571

38. Touster, O., Aronson, N. N. Jr, Dulaney, J. T. & Hendricken, H.

19. Wharton, D. C. & Tzagoloff, A. (1967) Methods Enzymol. 10,

20. Jaenicke, R. & Knof, S. (1968) Eur. J . Biochem. 4, 157-163. 21. Appelmans, F., Wattiaux, R. & deDuve, C. (1955) Bicochem. J .

22. Laemmli, U. K. (1970) Nature 227,680-685. 23. Johnston, A. & Thorpe, R. (1982) Immunochemistry in practice,

pp. 44 - 46, Blackwell Scientific Publications, London. 24. Bordier, C. (1981) J . Bid. Chem. 256, 1604-1607. 25. Rcdinbaugh, M. G. & Turley, R. B. (1986) Anal. Biochem. 153,

26. Belsham, G. J., Denton, M. & Tanner, M. A. (1980) Biochem. J .

27. Record, M., Bes, J X . , Chap, H. & Douste-Blazy, L. (1982)

(1970) J . Cell. Bid. 47, 604-618.

245 - 250.

59,438-445.

267 - 271,

192,457-461.

Biochim. Biophys. Acta 688, 57-65.

28. Orlowski, M., Reznik, S., Ayala, J. & Pierotti, A. R. (1989)

29. Tisljar, U. & Barrett, A. J. (1989) Arch. Biochem. Biophys. 274,

30. Barrett, A. J. & Tisljar, U. (1989) Biochem. J . 261, 1047-

31. Kenny, A. J. (1986) Trends Biochem. Sci. 11, 40-42. 32. Beynon, R. J., Shannon, J. D. & Bond, J. S. (1981) Biochem. J .

33. Kenny, A. J. & Ingram, J. (1987) Biochem. J . 245, 515-

34. Barnes, K., Ingram, J. & Kenny, A. J. (1989) Biochem. J . 264,

35. Stephenson, S. L. & Kenny, A. J. (1988) Biochern. J . 255, 45-

36. Habener, J. F., Rosenblatt, M. & Potts, J. T. (1984) Physiol. Rev.

Biochem. J . 261, 952 -958.

138 - 144.

1050.

199, 591 -598.

524.

335-346.

51.

64,985-1053.