Eur. J . Biochem. 204,547-552 (1992) $+ FEBS 1992

A membrane-bound metallo-endopeptidase from rat kidney Characteristics of its hydrolysis of peptide hormones and neuropeptides

Toru YAMAGUCHI, Hiroshi K I D 0 and Nobuhiko KATUNUMA Division of Enzyme Chemistry, Institute for Enzyme Research, the University of Tokushima, Japan

(Received September 17/November 21,1991) - EJB 91 1245

A membrane-bound metallo-endopeptidase that hydrolyzes human parathyroid hormone (1 - 84) and reduced hen egg lysozyme between hydrophilic amino acid residues was isolated from rat kidney [Yamaguchi et al. (1991) Eur. J . Biochem. 200, 563-5711, In this study, the hydrolyses of various peptide hormones and neuropeptides by the metallo-endopeptidase were examined using an automat- ed gas-phase protein sequencer. The purified enzyme hydrolyzed the oxidized insulin B chain and substance P most rapidly, followed by big endothelin 1, neurotensin, angiotensin 1, endothelin 1, rat a-atrial natriuretic peptide and bradykinin, in this order. The enzyme mainly cleaved these peptides at bonds involving a hydrophilic amino acid residue. However, it cleaved bonds between less hydrophilic amino acid pairs in several short peptides, e. g. at the His5 - Leu6 bond in oxidized insulin B chain, the Ile28 - Val29 bond in big endothelin-1 and the Ile5 - His6 and Phe8 - His9 bonds in angiotensin 1. The enzyme cleavage sites of oxidized insulin B chain and angiotensin 1 were different from the reported sites cleaved by meprin and by endopeptidase 2, respectively. Kinetic determination of bradykinin hydrolysis by the purified enzyme yielded values of K , = 18.1 pM and k,,, = 0.473 SKI,

giving a ratio of k,,,/K,,, = 2.62 x lo4 s- l . M-'. The K , value was about 20-fold lower than that reported for meprin and endopeptidase 2. These results indicate that the membrane-bound metallo- endopeptidase from rat kidney is distinguished from meprin and endopeptidase 2 in its substrate specificity and is not parathyroid hormone specific, but has potential capacities to inactivate various biologically active peptide hormones and neuropeptides in vivo.

There is cumulated recent evidence that kidney membrane metallo-endopeptidases are involved in the metabolism of bio- logically active peptides. One of these enzymes, endopeptidase 24,11, which is also present in the central nervous system, lung, intestine and male genital tract [l, 21, has been shown to metabolize atrial natriuretic peptide (ANP) in vitro [3 - 51. In vivo studies in animals and man [6-81 have indicated that administration of inhibitors of endopeptidase 24,ll caused increases in the circulating ANP concentration with a con- comitant increase in diuresis and natriuresis, suggesting a physiological role of endopeptidase 24,11 in ANP degradation it1 r im Moreover, rat endopeptidase 2, another kidney mem- brane metalloproteinase, that differs from endopeptidase 24, l l in its insensitivity to phosphoramidon [9, lo], has been reported to hydrolyze various neuropeptides in vitro [I I].

Like ANP, parathyroid hormone (PTH) is known to be degraded in the kidney and simultaneously to exert biological effects on the organ, such as inhibition of phosphate transport and enhancement of the synthesis of bioactive calcitriol [I 2, 131. Kidney membrane endopeptidase also seems to play some role in the metabolism of PTH. We have reported that human (h) PTH( 1 - 84) is degraded by a non-lysosomal endo- peptidase in intact cultured opossum kidney cells [14] and we have purified and characterized a rat kidney membrane

metallo-endopeptidase that hydrolyzes hPTH(1- 84) between hydrophilic amino acid residues [15].

To elucidate the biological role of the metallo- endopeptidase in the metabolisms of various peptide hor- mones and neuropeptides, we investigated the hydrolytic properties of the enzyme. This paper reports that the enzyme hydrolyzed a diverse range of peptides tested and that it hy- drolyzed not only peptide bonds flanked by hydrophilic amino acids, but also certain less hydrophilic amino acid pairs in short peptides. The cleavage sites of the purified enzyme were found to differ from those of meprin and endopeptidase 2. The biological role of this enzyme in peptide metabolism is discussed.

MATERIALS AND METHODS

Materials

Rat ANP, big endothelin 1 (porcine, residues 1 -39), endothelin 1, human angiotensin 1, bradykinin, neurotensin and substance P were obtained from the Peptide Institute, Osaka, Japan. Oxidized bovine insulin B chain was purchased from Sigma Chemical Co., St. Louis, USA. All other chemi- cals were of the highest grade commercially available.

Correspondence to T. Yamaguchi, Third Division, Department of

Ahhwviations. h, human; PTH, parathyroid hormone; ANP, Medicine. Kobe University School of Medicine, Kobe, Japan 650

atrial natriuretic peptide.

Purification of the enzyme

The rat kidney membrane metallo-endopeptidase hy- drolyzing PTH was purified as described previously [15].

548

1 Insulin B-chain

., 2

1 I I I I

ANP 4

a l l I I I I

f Endothelin 1 t

Angiotensin 1 l k

Bradykinin .1

I 1 I I 1

Neurotensin 1 J .

Substance P 2

i 1 1 I I

Retention Time (min) 10 20 30 40 50

1

lo‘ 20 30 $0 50 I

Fig. 1. HPLC analysis of peptide products formed by incubation of peptides with the purified rat enzyme. Peptides (20 pM) were incubated with 2.6 nM purified enzyme in 100 mM Tris/HCl, pH 8.0, for 4 h at 37’C. Then, each sample was fractionated by HPLC as described in Materials and Methods. Thc product peaks are numbered in order of yield, and numbers correspond to those in Table 2 and Fig. 2.

Hydrolysis of peptides and peptide-sequence analyses

The incubation mixture (volume 1 ml) contained 100 mM TrisiHCI, pH 8.0, 20 pM peptide and 7.10 nM (625 ng) rat enzq’me. Samples were incubated at 37°C for 0 h or 4 h and reactions were terminated by adding 50 pl 100% (by vol.) acetic acid. The mixtures were then fractionated by HPLC on an ODSl20T reverse-phase column (CIS; 4.6 minx 250 mm; Toyo Soda, Tokyo, Japan) with a linear gradient of 0-45% acetonitrile in 0.1% trifluoroacetic acid for 40 min at a flow rate of 1 ml/min. Elution was monitored by the absorbance at 215 nm and rates of hydrolysis were assessed as rates of disappearance of substrate peaks, measured with a LKB 2220 recording integrator linked to a HPLC system (Pharmacia LKB Biotechnology Inc., Uppsala, Sweden). Degradation products were collected in test tubes and freeze-dried. The amino acid sequences of these samples were determined with an Applied Biosystems model 470A gas-phase sequencer and model 120 A HPLC analyzer system.

Kinetic determination for bradykinin hydrolysis

Kinetics for bradykinin hydrolysis by the enzyme were measured by quantitative HPLC analysis of reaction mixtures. The incubation mixture (volume 250 p1) contained I00 mM Tris/HCl, pH 8.0, 2.17 nM (48 ng) rat enzyme and 5-80 pM bradykinin. Samples were incubated at 37°C for 2 h and reactions were stopped by adding 12.5 ~ 1 1 0 0 % (by vol.) acetic acid. The samples were then analyzed by HPLC as described above. Product peak areas were converted to concentration by defining the areas measured after complete hydrolysis as equal to the initial substrate concentration. Hydrolysis rates were assessed by the disappearance of the substrate peak or by the appearance of the products/unit time, which agreed quite well. Kinetic constants were calculated using the Michaelis-Menten equation and the linear transform methods of Lineweaver-Burk. The values for k,,, were calculated by assuming a catalytic subunit of 88 kDa. Three separate kinetic studies were performed using different enzyme preparations.

549

RESULTS HPLC analysis of peptide products hydrolyzed by the purified enzyme

Figure 1 shows the HPLC patterns after 4 h incubation of the oxidized insulin B chain, rat ANP, endothelin 1, big

Table 1. Hydrolysis of peptides by the purified enzyme during 4 h incubation. Peptides at 20 pM were incubated with 7.10 nM (625 ng) purified enzyme a t 37 ;C for 4 h in 1 ml 100 mM Tris/HCI, pH 8.0. Peptide products were fractionated by HPLC on a CI8 reverse-phase column with a linear gradient of 0-45% acetonitrile in 0.1% trifluoroacetic acid for 40 min at a flow rate of 1 ml/min. Rates of hydrolysis after 4 h were assessed as rates of disappearance of sub- strate peaks as described in Materials and Methods.

Peptides Hydrolysis rates

Oxidized insulin B chain Substance P Big endothelin 1 Neurotensin Angiotensin 1 Endothelin 1 ANP Bradykinin

YO 98 96 83 81 60 57 52 50

endothelin 1, human angiotensin 1, bradykinin, neurotensin and substance P with the purified enzyme. All these peptides were hydrolyzed by the enzyme and the degradation products were detected by their ultraviolet absorbance at 215 nm in HPLC profiles. The rates of hydrolysis were calculated from the rates of disappearance of the substrate with the retention times indicated by arrows in Fig. 1, and as determined in Materials and Methods. As shown in Table 1, the oxidized insulin B chain and substance P were hydrolyzed most rapidly, followed by big endothelin 1, neurotensin, angiotensin 1, endothelin 1, ANP and bradykinin, in this order.

Determination of amino acid sequences of peptide products liberated by the enzyme

The individual peak fractions shown in Fig. 1 were col- lected and their amino acid sequences were determined in an automated gas-phase sequencer (Table 2) . The sites of hydrolysis of each of the peptides predicted from these analy- ses are shown in Fig. 2.

Oxidized insulin B chuin

The oxidized insulin B chain was hydrolyzed to many fragments by the enzyme and the cleavage sites of the major degradation products were determined. This pcptide was

Table 2. Amino acid sequences of degradation products of peptides. The individual peak fractions, designated in Fig. 1, were collectcd and their amino acid sequences were determined in an automated gas-phase sequencer as described in Materials and Methods.

Peptide Peak Amino acid sequence Cleavage site

Insulin B chain (bovine) 1 Glu-Arg-Cly-Phe-Phe-Tyr Gly20 - Glu2l Tyr26-Thr27

2 Thr-Pro-Lys-Ala Tyr26-Thr27

Val1 8 - Cysl9 3 GI y-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val cys7 - Gly8

4 Phe-Val-Asn-Gln-His His5 - Leu6 5 Phe-Val-Asn-Gln-His-Leu- 6 GI y-Glu-Arg-GI y-Phe-Phe-Tyr Cysl9 - Cly20

Tvr26 - Thr27 7 Gly-Glu-Arg-GI y-Phe-Phe-Tyr-Thr-Pro-Lys- A h c-ysl9 - Gly20

Atrial natriuretic peptide (rat) 1 Gly-Leu-GI y-C ys- Asn

2 Ser-Leu-

Endothelin I 1 Cys-Ser-Cys-Ser-Ser- Leu-Met- 2 Asp-Ile-Ile-Trp

Big endothelin 1 (porcine. 1 ~ 39) 1 Cys-Ser-Cys-Ser-Ser-Leu-Met- Asp- 2 Asp-Ile-He-Trp- 3 Cys-Ser-Cys-Ser-Ser-Leu-Met- 4 Ser-Pro-Ser-Arg-Scr 5 Val-Pro-Tyr-Gly-Leu-Gly

Angiotensin 1 (human)

Brad ykinin

Neurotensin

1 Asp- Arg-Val-Tyr-Ile-His-Pro-Phe- 2 His-Pro-Phe

3 Asp-Arg-Val-Tyr 1 Ser-Pro-Phe- Arg 2 Arg-Pro-Pro-GI y-Phc

Serl9-Gly20 Asn24-Scr25

Leul7-Asp18

L c ~ l 7 - A ~ p l 8

Gly34-Ser35 He28 - Val29 Gly34-Scr35

Ile5 - His6 Phc8 - I Iis9 Tyr4- Ile5 Phe5 - Scr6 Phe5 - Ser6

1 GIu-Asn-Lys-Pro- Arg-Arg- Pro-Tyr- Ile-Leu Tyr3-Glu4 2 Clu- Asn- Lys-Pro- Tyr3 - Glu4

Substancc P 1 Arg-Pro-Lys-Pro-Gln-Gln-Phe-Phe Phc8 - Gly9 2 Arg-Pro-Lys-Pro-Gln-Gln Gln6- Phe7 3 Phe-Phe-Gly-Leu-Met Gln6- Phc7

5 50

N

C

Endothelin I ANP (Rat)

C

'"?

Big Endothelin I (porcine. 1-39)

1 Arg-Pro-Pro4 y-h-Ser -Pro-Phe-Arg

1 1 1 Aq-Arg-Val -1yr- lle-His-Pro-Rw-His-La

4 c * Angiotensin I (Human) Bradykinin

4 a w ? -4 -

1 1 1 R/r-Leu-Tyr-Glu-Asl~Lyys-Prolirg-Arg-Pro-lyr-lle-L~ Arg-Pro-Lys-Pro-Gln-Gln~~~~~Gly-L~-Met-NH*

4 3 b 4- b -Z-? 4---*--:3,-

Neurotensin Substance P

Fig. 2. Identities of the products formed by incubation of peptides with the purified enzyme. The numbered peptides corresponding to the HPLC peaks in Fig. 1 were identified by determining their amino acid sequences in an automated gas-phase sequencer. Arrows indicate the bonds attacked by the enzyme.

hydrolyzed at His5 - Leu6, Cys7 - Gly8, Val18 - Cysl9, Cysl9 - Gly20, Gly20 - Glu21 and Tyr26 - Thr27, producing peptides with sequences of 4 - 11 amino acids. The cleavage site of the main products (Fig. 1, peaks 1 and 2) was Tyr26 - Thr27.

Rat A NP

This hormone was hydrolyzed at Serl9 - Gly20 and Asn24 - Ser25, to produce a mid-portion fragment, ANP(20 - 24), designated as peak 1 in Fig. 1. The N-terminal portion of ANP appeared as another prominent peak (Fig. 1, peak 2) in the HPLC profile, but the C-terminal fragment was not detected as a major peak.

Endothelin I

Endothelin 1 was cleaved at the Leul7-Asp18 bond, liberating its N-terminal and C-terminal fragments (Fig. 1, peaks 1 and 2, respectively).

Big endothelin I One of the cleavage sites of big endothelin 1 was at

Leu17 - Aspl8, corresponding to that of endothelin 1. This peptide was also cleaved at Ile28 -Val29 and Gly34- Ser35 in the extended C-terminal chain.

Angiotensin 1

and PheB - His9. Angiotensin 1 was hydrolyzed at Tyr4 - Ile5, Ile5 - His6

Bradykinin Bradykinin appeared to be cleaved at one site, Phe5-

Ser6, to produce fragments 1-5 and 6-9 (Fig. 1, peaks 2 and 1, respectively).

Neurotensin This peptide was hydrolyzed at Tyr3 - Glu4, liberating two

C-terminal fragments with different retention times (Fig. 1, peaks 1 and 2). Its N-terminal fragment was not detected.

551

Table 3. Kinetic parameters for bradykinin hydrolysis,

kc,,/K,,, Reference Proteinase Kin kc,,

PM S - l s - 1 . M-1

Purified rat enzyme 18 0.5 2.6 x lo4 this paper Meprin 425 21.8 5.1 x104 18 Endopeptidase 2 325 1.5 4.5 103 11

3 L 1

3 1

I I I I I I

0 10 20 30 40 50 60 70 80

s (fl> Fig. 3. Dependence of velocity of hydrolysis on bradykinin concen- tration. Velocities for hydrolysis of various concentrations of bradykinin by the purified enzyme were determined by quantitative HPLC analysis as described in Materials and Methods. Inset, Lineweaver-Burk double-reciprocal plot of the same data.

Substance P Substance P was cleaved at Gln6 - Phe7 and Phe8 - Gly9.

Kinetics for bradykinin hydrolysis

When 20pM bradykinin was incubated with 2.17 nM enzyme at 37°C in 100 mM Tris/HCl, pH 8.0, hydrolysis was linearly increased time-dependently for up to 4 h (data not shown). Values for K, and k,,, were determined for bradykinin hydrolysis by measuring velocities by quantitative HPLC analysis at varying substrate concentrations in 2 h incu- bations. As shown in Fig. 3, kinetic constants were calculated using the Michaelis-Menten equation and the linear-transform methods of Lineweaver-Burk. Graphic analyses (Fig. 3 ) yielded values of K , = 18.1 pM and k,,, = 0.473 s - (V,,, = 1.03 nM . s-'; [enzyme] = 2.17 nM at the subunit molecular mass of 88 kDa), giving a ratio of k,,,/K, = 2.62 x lo4 s-' . M-I.

DISCUSSION

In the present study, the purified rat kidney membrane metallo-endopeptidase was found to hydrolyze various peptides with polypeptide chains shorter than PTH (Fig. 1). Thus, the purified enzyme was not specific for PTH, but also hydrolyzed other biologically active peptides in vitro.

The purified enzyme tended to cleave these peptides mainly at bonds involving a hydrophilic amino acid residue, such as those in the insulin B chain, ANP and neurotensin (Table 2 and Fig. 2). However, cleavage at bonds between less hydrophilic amino acid pairs was also observed, such as at His5 - Leu6 in the insulin B chain, Ile28GVa129 in big endothelin 1, and Ile5 - His6 and Phe8 - His9 in angiotensin 1. These results show that hydrolysis of short peptides containing less hydrophilic amino acid pairs is not limited to peptide bonds flanked by hydrophilic amino acids.

Two other metallo-endopeptidases, mouse meprin and rat endopeptidase 2, have also been purified from the kidney membrane and characterized [9 - 1 1, 16 - 181. These two en- zymes have quite similar properties, such as an oligometric structure, insensitivity to phosphoramidon and preference for

hydrophobic amino acid residues in peptide cleavage [9 - 1 1, 16,171. Hence, it has been proposed that these mouse and rat enzymes should be classified under the same EC number [lo]. The peptides tested in this study have also been used in studies on the substrate specificities of meprin and endopeptidase 2 [ l l , 171 and thus the results on the substrate specificities of these three purified enzymes can be compared directly. Our enzyme differs from meprin and endopeptidase 2 in its ability to hydrolyse the oxidized insulin B chain and angiotensin 1. It hydrolyzed the insulin B chain at Cys7-Gly8, Va118- Cysl9, Cysl9 - Gly20 and Tyr26 - Thr27 bonds and angio- tensin 1 at Ile5-His6 and Phe8-His9 bonds (Table 2 and Fig. 2) . These sites in the insulin B chain and in angiotensin 1 were not hydrolyzed by meprin and endopeptidase 2, respec- tively [I l , 171. However, in the metabolism of other neuropeptides, such as bradykinin, neurotensin and substance P, the substrate specificity of the purified enzyme resembles that of endopeptidase 2. Like the present enzyme, endopeptidase 2 cleaves Phe5 - Ser6 in bradykinin, Tyr3 - Glu4 in neurotensin and Gln6 - Phe7 and Phe8 - Gly9 in sub- stance P [l I]. Nevertheless, the present enzyme seems to differ from endopeptidase 2 in its substrate specificity, because our previous study using hPTH(1- 84) and reduced hen egg lysozyme, which are likely to be more suitable than short neuropeptides for investigating cleavage sites, indicated that the enzyme preferentially cleaves peptides between hydro- philic amino acid residues [I 51.

To further examine the difference in substrate specificity between our enzyme and meprin/endopeptidase 2, we deter- mined kinetics for bradykinin hydrolysis by the purified en- zyme. Kinetic parameters for bradykinin hydrolysis were also measured by meprin [18] and endopeptidase 2 [ll], and their published values are compared with the present data in Ta- ble 3. The K,,, value for the purified enzyme (18 pM) is about 20-fold lower than those for meprin and endopeptidase 2. The specificity ratio, k,,,/K, = 2.6 x lo4 s - l . M-', is similar to that for meprin but about sixfold higher than that for endopeptidase 2. These kinetic parameters for bradykinin hy- drolysis also distinguish our enzyme from meprin and endopeptidase 2.

This study showed that the purified enzyme can hydrolyze various biologically active substances in vitro. However, the actual substrates of this enzyme in vivo and in physiological intact kidney cells require further study.

This study was supported by a grant from the Yamanouchi Foun- dation for Research on Metabolic Disorders and research grants from the Japanese Ministry of Education, Science and Culture.

REFERENCES 1. Kenny, A. J. (1986) Trends Biochem. Sci. 11,40-42. 2. Ronco, P., Polland, H., Galceran, M., Delanche, M., Schwartz.

J . C. & Verroust, P. (1988) Lab. Invest. 58, 210-217.

552

3. Kochn. J. A., Norman, J . A,, Jones, B. N., LeSueur, L., Sakane, Y. & Ghai, R. D. (1 987) J . B id . Chem. 262, 11 623 - 11 627.

4. Olins, G. M., Spear, K . L., Spiegel, N . R., Reinhard, E. J . & Zurcher-Neely, H. A. (1987) Eur. J . Biochem. 170, 431 -434.

5. Sonnenberg, J. L., Sakane, Y., Jeng, A. Y., Loehn, J. A., Ansell, J. A,, Wennogle, L. P. & Ghai, R. D. (1988) Peptides 9, 173- 1x0.

6. Sybcrtz, E. J., Chiu, P. J. S., Vemulapalli, S., Pitts, B., Foster, C. J., Watkins, R. W., Barnett, A. & Hanslanger, M. F. (1989) J . Phurmacol. Exp. Ther. 250,624 - 63 1.

7. Olins, G. M., Krieter, P. A., Trapani, A. J., Spear, K. L. & Bovy, P. R. (1989) Mol. Cell. Endocrinol. 61, 210-218.

8 . Danielewicz, J. C . , Barclay, P. L., garnish, I. T., Brown, D., Campbell, S. F., James, K., Samuels, G. M. R., Terrett, N . K . & Wythes, M. J. (1989) Biochem. Biophys. Res. Commun. 164, 58-65.

9. Kenny, A. J. & Ingram, J. (1987) Biochem. J . 245,515-524.

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

1. Stephenson, S. L. & Kenny, A. J . (1988) Biochem. J . 255, 45- 51.

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

335 - 346.

13. Armitage, E. K. (1986) Clin. Chem. 32,418-424. 14. Yamaguchi, T., Fukase, M., Nishikawa, M., Fujimi, T. & Fujita,

T. (1988) Endocrinology 123,2812-2817. 15. Yamaguchi, T., Kido, € I . , Fukasc, M., Fujita, T. & Katunuma.

N. (1991) Eur. J . Biochem. 200, 563-571. 16. Beynon, R. J., Shannon, J. D. & Bond, J. S . (1981) Biochern. J .

17. Butler, P. E., McKay, M. J. & Bond, J . S. (1987) Biocheni. J . 241.

18. Wok, R. L. & Bond, J . S. (1990) Anal. Biochem. 191, 314-

IYY, 591 -598.

229 - 235.

320.