- Email: [email protected]
A colorimetric assay for measuring peptidylglycine α-amidating monooxygenase using high-performance liquid chromatography
ANALYTICAL
BIOCHEMISTRY
198,263-267
(1991)
A Calorimetric Assay for Measuring Peptidylglycine a-Amidating Monooxygenase Using High-Performance Liquid Chromatography Toshiyuki
Chikuma,*
Ko-ichi
Hanaoka,*
Y. Peng Loh,t
Takeshi
Kato,$
and Yoko Ishii*
*Department of Pharmaceutical Analytical Chemistry, Showa Collegeof Pharmaceutical Sciences,Machida-shi, Tokyo 194, Japan; iSection on Cellular Neurobiology, Laboratory of Developmental Neurobiology, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892; and SLaboratory of Molecular Recognition, Graduate School of Integrated Science, Yokohama City University, Yokohama 236, Japan Received
April
29, 1991
cessing. The amino acid sequences of many precursors to a-amidated peptides have been elucidated over the past several years. In every case, the amino acid that ultimately bears the a-amide moiety in the final product is followed by a glycine residue in the precursor molecule. Amidation is catalyzed by a copper- and ascorbate-requiring monooxygenase-peptidylglycine a-amidating monooxygenase (PAM,’ EC 1.14.17.3)-which was first identified in porcine pituitary by Bradbury et al. (2). It has been demonstrated that PAM catalyzes oxygenative cleavage at the C-terminal glycine of the precursor peptides to give the amidated peptide and glyoxylate. Enzymes capable of amidation have since been detected in a number of tissues in various species (3-9). The most commonly used assay system for detecting and estimating PAM activity is based on the ability of enzyme preparation to convert the synthetic radioiodinated tripeptide D-Tyr-Val-Gly to the corresponding dipeptide amide D-Tyr-Val-NH,. The degree of conversion to product is determined by cation-exchange chromatography of peptides and quantitation of the radioactive fractions (2). Several other radioactive, synthetic peptides have been prepared and utilized as substrates for the estimation of PAM activity. As above, an analytical system Press, Inc. was required to resolve the assay substrate from any product generated during the assay reaction. These include N-[‘4C]succinyl-Ala-Phe-Gly used in conjunction A number of biologically active peptides have an (Y- with high-voltage paper electrophoresis (lo), [3H]pGluamide structure at their carboxyl termini. In most cases, the presence of the C-terminal a-amide structure is gen‘Abbreviations used: BSA, bovine serum albumin; Dabsyl, 4-dierally essential for the biological activity of their pepmethylaminoazobenzene-4’-sulfonyl; Dansyl, 5dimethylaminotides and may be important in the regulation of peptide naphtbalene-1-sulfonyl; NEM, IV-ethylmaleimide; NLeu, L-norleuhormones (1). This C-terminal a-amide formation is tine; PAM, peptidylglycine cw-amidatiig monooxygenase; PCMB, one of the most important events in prohormone pro- p-chloromercuribenzoate; PMSF, phenylmethylsulfonyl fluoride. In many peptide hormones and neuropeptides, the carboxy-terminal a-amide structure is essential in eliciting biological activity. In the present study, a rapid and sensitive assay method for the determination of peptidylglycine cr-amidating monooxygenase (PAM) activity has been reported. This method is based on the monitoring of the absorption at 460 nm of 4-dimethylaminoazobenzene-4’-sulfonyl-Gly-L-Phe-NH2 (Dabsyl-Gly-Phe-NH,), enzymatically formed from the substrate 4-dimethylaminoazobenzene-4’-sulfonyl-GlyL-Phe-Gly, after separation by high-performance liquid chromatography (HPLC) using a C-18 reversedphase column by isocratic elution. This method is sensitive enough to measure Dabsyl-Gly-Phe-NH2 at concentrations as low as 1 pmol and yield highly reproducible results and requires less than 5 min per sample for separation and quantitation. The concentrations of copper and ascorbic acid required for maximal enzyme activity were 1 PM and 2 nuu, respectively. The pH optimum for PAM activity was 5.0 to 5.5. The Km and V,, values were respectively 3.5 pM and 100 pmol/ag/h with the use of enzyme extract obtained from bovine pituitary. By using this method, PAM activity could be readily detected in a single rat saliva. The sensitivity of this assay method will also aid in the effort to examine the regulation of in Vivo PAM activity. Q 1991 Academic
0003-2697/91$3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form
263 Inc. reserved.
264
CHIKUMA
His-Pro-Cly coupled with thin-layer chromatography (ll), and N-acetyl-[‘251]Tyr-Phe-Gly coupled with liquid-liquid partition of product and substrate (5). On the other hand, nonradioactive, synthetic peptides have also been prepared and utilized as substrates for PAM activity. These substrates are N-Dansyl-D-Tyr-ValGly (12) and N-trinitrophenyl-D-Tyr-Val-Gly (13), and they were based on reversed-phase high-performance liquid chromatographic separation and spectrophotometric detection. In this paper, we describe a new and sensitive assay method for PAM activity using Dabsyl-Gly-Phe-Gly as substrate by HPLC on a reversed-phase column to achieve a rapid and selective separation of substrate and product. Using this method, PAM activity was discovered in rat saliva. MATERIALS
AND
METHODS
Materials. 4-Dimethylaminoazobenzene-4’-sulfonyl chloride, ascorbic acid, l,lO-phenanthroline hydrochloride, bacitracin, and soybean trypsin inhibitor were purchased from Wako (Tokyo, Japan). N-Ethylmaleimide (NEM), oxotremorine, L-norleucine, phenylmethylsulfonyl fluoride (PMSF), pepstatin A, p-chloromercuribenzoate (PCMB), and bovine serum albumin (BSA) were obtained from Sigma Chemical Co. (St. Louis, MO). Other materials and their sources were Gly-LPhe-Gly (Bachem Feinchemikalien AG, Bubendorf, Switzerland), Gly-L-Phe-NH, (Novabiochem, Laufelfingen, Switzerland), beef liver catalase (Boehringer, Mannheim, Germany), and sodium pentobarbital (Dainabot, Tokyo, Japan). Acetonitrile was of chromatographic grade (Wake). Other chemicals and solvents were of analytical reagent grade. Preparation of enzyme source. Two different enzyme preparations were used. (i) Bovine pituitaries were obtained from a local slaughter house. Extracts of bovine pituitaries were prepared and subjected to ammonium sulfate precipitation according to the procedure of Murthy et al. (7). (ii) Male Wistar rats weighting 250300 g were purchased from Charles River (Japan) and housed with ad libitum access to chow and water for 1 week. The animals were anesthetized with sodium pentobarbital (40 mg/kg, ip). After lo-15 min, 2 mg/kg oxotremorine was injected intraperitoneally. The saliva was absorbed with defatted cotton for 30 min and collected from the cotton by centrifugation at 1OOOg for 10 min. synthePeptide synthesis. Dabsyl-Gly-Phe-Glywas sized by the method of Lin and Chang (14) with minor modifications. In brief, 180 pmol of Gly-r.,-Phe-Gly was dissolved in 5 ml of 50 mM sodium carbonate-sodium bicarbonate buffer. To this peptide solution, 45 pmol of Dabsyl chloride in 5 ml of acetone was added. The tightly capped mixture was allowed to react at 70°C for
ET
AL.
~3~;NQN’N~~-NH-Gly-L-Phe-Gly R Dabsyl-Gly-Pho-Gly
Peptidylglycine
a-amidating
monooxygenase
( PAM
1
1 a-Hydroxy
derivative
of
C-terminal
glycine ( intermediate
~3)N~N=N~~-NH-Gly-L-Phe-Nli~
+
)
F:zH
3 Dabsyl-Gly-Phe-NH2 (X max
: 460
Glyoxylate nm
)
FIG.
1. Principle of the calorimetric assay of peptidylglycine (Yamidating monooxygenase (PAM) with Dabsyl-Gly-Phe-Gly as the substrate. Recently Tajima et al. suggested that the direct product of the reaction catalyzed by the PAM is a hydroxy derivative at the a-carbon of the carboxyl-terminal glycine (29).
15 min in a water bath with constant shaking, and thereafter the acetone was evaporated under reduced pressure. Dabsyl-Gly-Phe-NH, and 4-dimethylaminoazobenzene-4’-sulfonyl-L-norleucine (Dabsyl-NLeu) were also prepared in a similar manner using Gly-L-PheNH, and L-norleucine as the starting materials, respectively. All synthetic dabsylated products were purified by reversed-phase HPLC on a TSK gel ODS-80TM column (7.8 mm X 30 cm, TOSOH) before use (15). Following purification, the composition of these dabsylated compounds was verified by amino acid analysis. Assay for PAM activity. The principle of the assay method for PAM activity is based on the calorimetric measurement at 460 nm of Dabsyl-Gly-Phe-NH, formed enzymatically from the substrate, Dabsyl-GlyPhe-Gly, after separation by HPLC (Fig. 1). The reaction mixture contained 50 mM sodium acetate buffer (pH 5.4), 200 mM NaCl, 0.5 mM NEM (freshly prepared), 100 pg/ml beef liver catalase, 50 pM Dabsyl-Gly-Phe-Gly, varying concentrations of CuSO, and ascorbic acid (freshly prepared), and enzyme plus water in a total reaction volume of 250 ~1. Incubation was carried out at 37°C and the reaction was terminated by heating at 95°C for 5 min in boiling water. After centrifugation, Dabsyl-NLeu was added to clear supernatant as the internal standard, and an aliquot of the mixture obtained was subjected to HPLC analysis. The peak area of Dabsyl-Gly-Phe-NH, was measured and converted to picomoles from the peak area of Dabsyl-NLeu added as an internal standard. One unit of enzyme activity is defined as the amount of enzyme required to convert 1 pmol of substrate to amidated product in 1 min at 37°C.
COLORIMETRIC
(Y-AMIDATING
2 D
4
3 1
h
L
5
0
Time
( min
1
5
)
FIG. 2. HPLC elution patterns of PAM activity determined using enzyme in bovine pituitary. Conditions were described under Materials and Methods. The concentrations of copper ion and ascorbic acid were fixed at 1 pM and 2 mM, respectively. Peaks: 1, unknown compound; 2, Dabsyl-Gly-Phe-Gly; 3, Dabsyl-NLeu; 4, Dabsyl-GlyPhe-NH,. Three hundred picomoles of Dabsyl-NLeu (internal standard) was added to each sample after incubation. (A) Blank incubation: Dabsyl-Gly-Phe-Gly was incubated without enzyme at 37“C for 4 h. (B) Standard incubation: 300 pmol of Dabsyl-Gly-Phe-NH, was added to a sample tube before incubation as a standard sample. The two peak areas of Dabsyl-Gly-Phe-NH, and Dabsyl-NLeu correspond to 50 pmol. (C) Experimental incubation: Dabsyl-Gly-PheGly was incubated with bovine pituitary extract at 37°C for 4 h. (D) Control incubation: a control tube without the enzyme was incubated, the same amount of active enzyme was added, and the resulting tube was kept in an ice bath before heating at 95’C for 5 min.
ENZYME
265
ASSAY
graphic patterns of the reaction mixture after incubation with 21.4 pg of protein prepared from bovine pituitary for 4 h. The blank incubation (Fig. 2A) contained Dabsyl-Gly-Phe-Gly and Dabsyl-NLeu, and the standard incubation contained exogenous Dabsyl-GlyPhe-NH, in addition to Dabsyl-Gly-Phe-Gly and Dabsyl-NLeu (Fig. 2B). The retention times for Dabsyl-Gly-Phe-Gly, Dabsyl-NLeu, and Dabsyl-GlyPhe-NH, were 2.6, 3.9, and 4.8 min, respectively (Figs. 2A and 2B). The experimental incubation under the standard assay conditions (Fig. 2C) showed a significant amount of Dabsyl-Gly-Phe-NH, at 4.8 min, whereas the control incubation did not show any peak of Dabsyl-Gly-Phe-NH, (Fig. 2D). The enzyme reaction was found to be linear with time at 37°C at least for about 8 h (Fig. 3). The catalytic activity of the enzyme was greatest at a pH of approximately 5.0 to 5.5, with a rapid loss of activity at pH below 4.5 (data not shown). PAM activity was investigated as a function of the amount of enzyme extract obtained from bovine pituitary. Perfect linearity was observed for plots of the amount of Dabsyl-Gly-Phe-NH,, at least from 0.14 to 2.15 units, formed enzymatically from Dabsyl-GlyPhe-Gly against those of enzyme (data not shown). In experiments designed to optimize the PAM assay conditions, various metals were analyzed for their ability to stimulate the PAM activity. Of the metal tested, the only metal that exerted a stimulatory effect on the PAM activity was copper. The optimal copper concentration required for this enzyme was about 1 pM (Fig. 4).
Chromatographic conditions. Analysis of the product was performed using a Japan Spectroscopic Co. Ltd. HPLC system consisting of a 880-PU pump, 875-uv detector (fixed at 460 nm), 860-CO column oven, 880-50 degasser, 880-02 gradient unit, 802-SC system controller, and 807-IT integrator. The system was operated at 35°C at a flow rate of 0.75 ml/min employing a TSK gel ODS80TM (particle size, 5 pm) reversed-phase column (150 X 4.6 mm i.d.) fitted with a TSK guard gel ODS-80TM (15 X 3.2 mm i.d.; particle size, 5 pm). The mobile phase consisted of 0.01 M sodium acetate buffer (pH 4.00)acetonitrile(41/59, v/v). Protein determination. Protein concentration was measured by the Lowry method as modified by Hartree (16) using BSA as standard protein.
RESULTS This HPLC-calorimetric detection system for the measurement of Dabsyl-Gly-Phe-Gly and DabsylGly-Phe-NH, was found to be very sensitive. The calibration graph for Dabsyl-Gly-Phe-NH, injected showed good linearity from 1 to 1000 pmol. The calibration graph for Dabsyl-NLeu also showed good linearity from 5 to 1000 pmol. Figure 2 shows the chromato-
Time
FIG. 3.
( hours
)
Time-dependent changes of Dabsyl-Gly-Phe-NH, enzymatically formed from the substrate. The standard assay conditions are described under Materials and Methods. The concentrations of copper ion and ascorbic acid in the reaction mixture are 1 pM and 2 mM, respectively. Incubation was carried out at 37°C for the indicated periods.
266
CHIKUMA
0’
.‘..‘m
.....‘m 0.1
0.5
......’ 1 cuso4
5
. ... 10
ET
’
50
( JJM )
FIG. 4. Determination of effects of optimal copper ion concentrations on PAM activity from bovine pituitary. The rates of DabsylGly-Phe-NH, formation by PAM from bovine pituitary extract at various copper ion concentrations were measured as described under Materials and Methods. The concentration of ascorbic acid was fixed at 2 mu. Incubation was done at 37°C for 4 h.
A higher concentration of copper consistently showed an inhibitory effect on a-amidation. The effects of ascorbic acid concentration on enzyme activity were also examined (Fig. 5). The optimal concentration of ascorbic acid for PAM activity was found to be 2 mM. There is a clear loss of the stimulately effect at high concentrations of ascorbic acid. A Lineweaver-Burk plot was obtained from the effect of the concentration of Dabsyl-Gly-Phe-Gly on the rate of formation of Dabsyl-Gly-Phe-NH, by PAM. The Michaelis constant (K,) and the maximum velocity (V-) toward the Dabsyl-Gly-Phe-Gly were calculated to be 3.5 PM and 100 pmol/pg/h, respectively. Finally, we applied this standard assay method for the detection of PAM activity in rat saliva. Under the anesthetized conditions, the muscarinic agonist oxotremorine (2 mg/kg) was intraperitoneally injected to collect saliva. About 0.5 ml of saliva for 30 min per one rat was collected. PAM activity in the saliva was 0.78 + 0.12 nmol/ml/h (mean + SE; n = 5).
AL.
has a few advantages. First, it is very sensitive. The limit of the sensitivity was about 1 pmol of Dabsyl-Gly-PheNH, formed enzymatically. Second, the substrate and the product are separated completely in less than 5 min. Third, more accurate quantitation of the product and better reproducibility were guaranteed in our method by the employment of internal standard (Dabsyl-NLeu) compared to those of the HPLC-fluorometric assay system using N-Dansyl-D-Tyr-Val-Gly as substrate, in which an internal standard had not been used. Finally, an important structural feature of the dabsylated substrate is that the Dabsyl group prevents the ionization of the amino terminus of the peptides. Thus, the net charge of the substrate remains constant in the pH range of 5 to 9, the same as other substrates for PAM, N-Dansyl-D-Tyr-Val-Gly (12,19) and N-trinitrophenyl-D-Tyr-Val-Gly (13). In comparison, the radioiodinated substrate D-Tyr-Val-Gly goes through a transition form in the same pH range. The structural change may complicate the assessment of pH effects on enzyme activity. The use of the dabsylated substrate permits the unambiguous study of pH effects. The stimulating effects of copper ion on PAM activity have been described previously in several tissues, and the addition of copper ion was also essential to the detection of PAM activity in this assay system (optimal concentration was 1 PM). The concentrations of copper ion for optimal PAM activity were very different among tissues. Those from rat pituitary, rat hypothalamus, rat stomach, and rat medullary thyroid carcinoma CA-77 cell culture medium required 2,25,50, and 30 pM copper ion, respectively (3,6,20,21). Therefore, the copper ion requirement for different tissues should be tested individually. Ascorbic acid has been shown to be of widespread importance in a number of key biochemical processes in-
DISCUSSION
The first assay used for the detection of PAM activity was developed by Bradbury and co-workers (2), who utilized ‘2SI-labeled D-Tyr-Val-Gly as a substrate. This assay method has been used by several other laboratories (4,17). Herein we reported a new assay method for PAM activity by the HPLC-colorimetric detection system using Dabsyl-Gly-Phe-Gly as substrate. The sequence of Cterminal Phe-Gly was selected because it appears to be the most reactive substrate for PAM ever reported (18). The proposed sensitive assay method for PAM activity
Ascorbic
acid
( mM )
FIG. 5. Determination of effects of optimal ascorbic acid concentrations on PAM activity from bovine pituitary. The rates of DabsylGly-Phe-NH, formation by PAM from bovine pituitary extract at various concentrations of ascorbic acid were measured as described under Materials and Methods. The concentration of copper ion was fixed at 1 PM. Incubation was done at 31°C for 4 h.
COLORIMETRIC
a-AMIDATING
eluding the hydroxylation of dopamine to form norepinephrine (22,23) and collagen biosynthesis (24). In contrast to the optimal concentration of copper ion for PAM activity, those of ascorbic acid were not significantly different among tissues. The inhibitions of PAM activity at supraoptimal concentrations of copper ion and ascorbic acid may be the result of enzymatic inactivation caused by free-radical production induced by ascorbic acid and superoxide in the presence of copper ion (25,26). Mains et al. (27) reported that PAM activity in the submandibular gland was increased after treatment with the a-adrenergic antagonist phenoxybenzamine and decreased after treatment with the a-adrenergic agonist phenylephrine. Moreover, serum levels of PAM activity were decreased after the treatment with phenoxybenzamine. From these data we propose that PAM may be released from salivary glands to the saliva and/ or serum. In the present study, we proved the presence of PAM activity in rat saliva. The effects of various chemical reagents and protease inhibitors on PAM activity in bovine pituitary extract were examined to investigate the possibility that PAM activity was affected by confounding protease activity in enzyme source. PAM activity was not affected by serine protease inhibitor (PMSF), acid protease inhibitor (pepstatin A), and soybean trypsin inhibitor. Furthermore, PAM activity was not affected by bacitracin. However, EDTA and l,lO-phenanthroline inhibited the enzyme activity completely at a final concentration of 0.1 mM, presumably by chelation of the enzyme-bound copper prosthetic group. The enzyme activity was also inhibited by the addition of 1 mM of dithiothreitol, but the inhibitory effects of the thiol compound were restored by the addition of NEM and/or PCMB, suggesting that the addition of typical thiol inhibitor may overcome the inhibitory effects of endogeneous thiol compounds (5,28). These results suggest that PAM activity was not affected by confounding protease activity during long incubation times. PAM is the key enzyme involved in the posttranslational cY-amidation of many biologically active peptides. The understanding of the mechanisms controlling amidated-peptide production will increase through an investigation of the regulation of PAM activity in Go. In conclusion, the ease of handling of this compound compared to the radioactive material makes Dabsyl-GlyPhe-Gly an excellent substrate for the study of PAM. Furthermore, the fast and accurate assay method described in this paper may be a useful means for the purpose described above. ACKNOWLEDGMENTS The authors Pharmaceutical
gratefully Analytical
thank Dr. Chemistry
Akira Tanaka, and Dr. Tomoji
Department of Kocha, Depart-
ENZYME
267
ASSAY
ment of Hygienic Sciences, for helpful
Chemistry, discussions.
Showa
College
of Pharmaceutical
REFERENCES 1. Mains, (1983)
R. E., Eipper,
B. A., Glembotski,
C. C., and Dores,
R. M.
Trends Neurosci. 6, 229-235.
2. Bradbury, A. F., Finnie, M. D. A., and Smyth, 298,686-688. 3. Eipper, B. A., Mains, R. E., and Glembotski,
G. (1982)
D.
Nature Proc.
C. C. (1983)
Natl. Acad. Sci. USA 80,5144-5148. 4. Eipper,
B. A., Myers,
A. C., and Mains,
R. E. (1985)
Endocrinology
116,2497-2504. 5. Mizuno, (1986) 6. Sakata,
K., Sakata,
J., Kojima,
M., Kangawa,
K., and Matsuo,
H.
Biochem. Biophys. Res. Commun. 137,984-991. J., Mizuno, K., and Matsuo, H. (1986) Biochem. Biophys. Res. Commun. 140,230-236. 7. Murthy, A. S. N., Mains, R. E., and Eipper, B. A. (1986) J. Biol. Chem. 261,1815-1822.
8. Kizer, J. S., Bateman, W. H., and Youngblood, 2267.
R. C., Miller, W. W. (1986)
9. Mehta, N. M., Gilligan, J. P., Jones, B. A., and Birnbaum, R. S. (1988) 44-54. 10. Mollay, 254.
C., Wichta,
11. Husain, 12. Jones, Lovato, (1988)
J., and Kreil,
S. S. (1983)
I., and Tate, B. N., Tamburini, S. J., Gilligan,
C. R., Humm,
J., Busby,
Endocrinology
118,2262-
B. N., Bertelsen,
A. H., Roos,
Arch. Biochem. Biophys. 261,
G. (1986) FEBS L&t. 202,251FEBS Lett. 152, 277-281.
P. P., Consalvo, A. P., Young, J. P., Jeng, A. Y., and Wennogle,
S. D., L. P.
Anal. Biochem. 168,272-279. A. G., and May, S. W. (1988) B&hem. Biophys. Res. Commun. 151,499-505. 14. Lin, J. K., and Chang, J. Y. (1975) Anal. Chem. 4’7,1634-1638. 15. Glembotski, C. C., Eipper, B. A., and Mains, R. E. (1984) J. Biol. Chem. 259,6385-6392. 16. Hartree, E. F. (1972) Anal. Biochem. 48,422-427. 17. Glembotski, C. C. (1985) Arch. Biochem. Biophys. 241,673-683. 18. Bradbury, A. F., and Smyth, D. G. (1983) Biochem. Biophys. Res. Commun. 112,372-377. 19. Bendig, M. M. (1986) J. Biol. Chem. 261,11,935-11,937. 20. Emeson, R. B. (1984) J. Neurosci. 4,2604-2613. 13. Katopodis,
21. Jeng, A. Y., Wong, M., Lovato, S. J., Erion, J. P. (1990) Anal. B&hem. 185,213-219. 22. Levin, E. Y., Levenberg, B., and Kaufman,
M. D., and Gilligan,
S. (1960)
J. Biol.
Chem. 235,2080-2086. 23. Friedman, S., and Kaufman, S. (1965) 4773. 24. Murad, S., Sivarajah, A., and Pinnell,
J. Biol. Chem. 240,4763S. R. (1981)
Biochem.
Biophys. Res. Commun. 101,868-875. 25. Davison, A. J., Kettle, 261,1193-1200. 26. Shinar, E., Navok, 14,778-14,783. 27. Mains,
R. E., Myers,
A. J., and Fatur,
T., and Chevion,
D. J. (1986)
M. (1983)
A. C., and Eipper,
J.
J. Biol. Chem.
Biol. Chem. 268,
B. A. (1985)
Endocrinology
116,2505-2515. 28. Mizuno, K., Kojima, M., Sakata, J., and Matsuo, H. (1986) in Peptide Chemistry 1985, pp. 351-356, Protein Research Foundation, Osaka, Japan. 29. Tajima, M., Iida, T., Yoshida, S., Komatsu, K., Namba, R., Yanagi, M., Noguchi, M., and Okamoto, H. (1990) J. Biol. Chem. 265,9602-9605.
BIOCHEMISTRY
198,263-267
(1991)
A Calorimetric Assay for Measuring Peptidylglycine a-Amidating Monooxygenase Using High-Performance Liquid Chromatography Toshiyuki
Chikuma,*
Ko-ichi
Hanaoka,*
Y. Peng Loh,t
Takeshi
Kato,$
and Yoko Ishii*
*Department of Pharmaceutical Analytical Chemistry, Showa Collegeof Pharmaceutical Sciences,Machida-shi, Tokyo 194, Japan; iSection on Cellular Neurobiology, Laboratory of Developmental Neurobiology, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892; and SLaboratory of Molecular Recognition, Graduate School of Integrated Science, Yokohama City University, Yokohama 236, Japan Received
April
29, 1991
cessing. The amino acid sequences of many precursors to a-amidated peptides have been elucidated over the past several years. In every case, the amino acid that ultimately bears the a-amide moiety in the final product is followed by a glycine residue in the precursor molecule. Amidation is catalyzed by a copper- and ascorbate-requiring monooxygenase-peptidylglycine a-amidating monooxygenase (PAM,’ EC 1.14.17.3)-which was first identified in porcine pituitary by Bradbury et al. (2). It has been demonstrated that PAM catalyzes oxygenative cleavage at the C-terminal glycine of the precursor peptides to give the amidated peptide and glyoxylate. Enzymes capable of amidation have since been detected in a number of tissues in various species (3-9). The most commonly used assay system for detecting and estimating PAM activity is based on the ability of enzyme preparation to convert the synthetic radioiodinated tripeptide D-Tyr-Val-Gly to the corresponding dipeptide amide D-Tyr-Val-NH,. The degree of conversion to product is determined by cation-exchange chromatography of peptides and quantitation of the radioactive fractions (2). Several other radioactive, synthetic peptides have been prepared and utilized as substrates for the estimation of PAM activity. As above, an analytical system Press, Inc. was required to resolve the assay substrate from any product generated during the assay reaction. These include N-[‘4C]succinyl-Ala-Phe-Gly used in conjunction A number of biologically active peptides have an (Y- with high-voltage paper electrophoresis (lo), [3H]pGluamide structure at their carboxyl termini. In most cases, the presence of the C-terminal a-amide structure is gen‘Abbreviations used: BSA, bovine serum albumin; Dabsyl, 4-dierally essential for the biological activity of their pepmethylaminoazobenzene-4’-sulfonyl; Dansyl, 5dimethylaminotides and may be important in the regulation of peptide naphtbalene-1-sulfonyl; NEM, IV-ethylmaleimide; NLeu, L-norleuhormones (1). This C-terminal a-amide formation is tine; PAM, peptidylglycine cw-amidatiig monooxygenase; PCMB, one of the most important events in prohormone pro- p-chloromercuribenzoate; PMSF, phenylmethylsulfonyl fluoride. In many peptide hormones and neuropeptides, the carboxy-terminal a-amide structure is essential in eliciting biological activity. In the present study, a rapid and sensitive assay method for the determination of peptidylglycine cr-amidating monooxygenase (PAM) activity has been reported. This method is based on the monitoring of the absorption at 460 nm of 4-dimethylaminoazobenzene-4’-sulfonyl-Gly-L-Phe-NH2 (Dabsyl-Gly-Phe-NH,), enzymatically formed from the substrate 4-dimethylaminoazobenzene-4’-sulfonyl-GlyL-Phe-Gly, after separation by high-performance liquid chromatography (HPLC) using a C-18 reversedphase column by isocratic elution. This method is sensitive enough to measure Dabsyl-Gly-Phe-NH2 at concentrations as low as 1 pmol and yield highly reproducible results and requires less than 5 min per sample for separation and quantitation. The concentrations of copper and ascorbic acid required for maximal enzyme activity were 1 PM and 2 nuu, respectively. The pH optimum for PAM activity was 5.0 to 5.5. The Km and V,, values were respectively 3.5 pM and 100 pmol/ag/h with the use of enzyme extract obtained from bovine pituitary. By using this method, PAM activity could be readily detected in a single rat saliva. The sensitivity of this assay method will also aid in the effort to examine the regulation of in Vivo PAM activity. Q 1991 Academic
0003-2697/91$3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form
263 Inc. reserved.
264
CHIKUMA
His-Pro-Cly coupled with thin-layer chromatography (ll), and N-acetyl-[‘251]Tyr-Phe-Gly coupled with liquid-liquid partition of product and substrate (5). On the other hand, nonradioactive, synthetic peptides have also been prepared and utilized as substrates for PAM activity. These substrates are N-Dansyl-D-Tyr-ValGly (12) and N-trinitrophenyl-D-Tyr-Val-Gly (13), and they were based on reversed-phase high-performance liquid chromatographic separation and spectrophotometric detection. In this paper, we describe a new and sensitive assay method for PAM activity using Dabsyl-Gly-Phe-Gly as substrate by HPLC on a reversed-phase column to achieve a rapid and selective separation of substrate and product. Using this method, PAM activity was discovered in rat saliva. MATERIALS
AND
METHODS
Materials. 4-Dimethylaminoazobenzene-4’-sulfonyl chloride, ascorbic acid, l,lO-phenanthroline hydrochloride, bacitracin, and soybean trypsin inhibitor were purchased from Wako (Tokyo, Japan). N-Ethylmaleimide (NEM), oxotremorine, L-norleucine, phenylmethylsulfonyl fluoride (PMSF), pepstatin A, p-chloromercuribenzoate (PCMB), and bovine serum albumin (BSA) were obtained from Sigma Chemical Co. (St. Louis, MO). Other materials and their sources were Gly-LPhe-Gly (Bachem Feinchemikalien AG, Bubendorf, Switzerland), Gly-L-Phe-NH, (Novabiochem, Laufelfingen, Switzerland), beef liver catalase (Boehringer, Mannheim, Germany), and sodium pentobarbital (Dainabot, Tokyo, Japan). Acetonitrile was of chromatographic grade (Wake). Other chemicals and solvents were of analytical reagent grade. Preparation of enzyme source. Two different enzyme preparations were used. (i) Bovine pituitaries were obtained from a local slaughter house. Extracts of bovine pituitaries were prepared and subjected to ammonium sulfate precipitation according to the procedure of Murthy et al. (7). (ii) Male Wistar rats weighting 250300 g were purchased from Charles River (Japan) and housed with ad libitum access to chow and water for 1 week. The animals were anesthetized with sodium pentobarbital (40 mg/kg, ip). After lo-15 min, 2 mg/kg oxotremorine was injected intraperitoneally. The saliva was absorbed with defatted cotton for 30 min and collected from the cotton by centrifugation at 1OOOg for 10 min. synthePeptide synthesis. Dabsyl-Gly-Phe-Glywas sized by the method of Lin and Chang (14) with minor modifications. In brief, 180 pmol of Gly-r.,-Phe-Gly was dissolved in 5 ml of 50 mM sodium carbonate-sodium bicarbonate buffer. To this peptide solution, 45 pmol of Dabsyl chloride in 5 ml of acetone was added. The tightly capped mixture was allowed to react at 70°C for
ET
AL.
~3~;NQN’N~~-NH-Gly-L-Phe-Gly R Dabsyl-Gly-Pho-Gly
Peptidylglycine
a-amidating
monooxygenase
( PAM
1
1 a-Hydroxy
derivative
of
C-terminal
glycine ( intermediate
~3)N~N=N~~-NH-Gly-L-Phe-Nli~
+
)
F:zH
3 Dabsyl-Gly-Phe-NH2 (X max
: 460
Glyoxylate nm
)
FIG.
1. Principle of the calorimetric assay of peptidylglycine (Yamidating monooxygenase (PAM) with Dabsyl-Gly-Phe-Gly as the substrate. Recently Tajima et al. suggested that the direct product of the reaction catalyzed by the PAM is a hydroxy derivative at the a-carbon of the carboxyl-terminal glycine (29).
15 min in a water bath with constant shaking, and thereafter the acetone was evaporated under reduced pressure. Dabsyl-Gly-Phe-NH, and 4-dimethylaminoazobenzene-4’-sulfonyl-L-norleucine (Dabsyl-NLeu) were also prepared in a similar manner using Gly-L-PheNH, and L-norleucine as the starting materials, respectively. All synthetic dabsylated products were purified by reversed-phase HPLC on a TSK gel ODS-80TM column (7.8 mm X 30 cm, TOSOH) before use (15). Following purification, the composition of these dabsylated compounds was verified by amino acid analysis. Assay for PAM activity. The principle of the assay method for PAM activity is based on the calorimetric measurement at 460 nm of Dabsyl-Gly-Phe-NH, formed enzymatically from the substrate, Dabsyl-GlyPhe-Gly, after separation by HPLC (Fig. 1). The reaction mixture contained 50 mM sodium acetate buffer (pH 5.4), 200 mM NaCl, 0.5 mM NEM (freshly prepared), 100 pg/ml beef liver catalase, 50 pM Dabsyl-Gly-Phe-Gly, varying concentrations of CuSO, and ascorbic acid (freshly prepared), and enzyme plus water in a total reaction volume of 250 ~1. Incubation was carried out at 37°C and the reaction was terminated by heating at 95°C for 5 min in boiling water. After centrifugation, Dabsyl-NLeu was added to clear supernatant as the internal standard, and an aliquot of the mixture obtained was subjected to HPLC analysis. The peak area of Dabsyl-Gly-Phe-NH, was measured and converted to picomoles from the peak area of Dabsyl-NLeu added as an internal standard. One unit of enzyme activity is defined as the amount of enzyme required to convert 1 pmol of substrate to amidated product in 1 min at 37°C.
COLORIMETRIC
(Y-AMIDATING
2 D
4
3 1
h
L
5
0
Time
( min
1
5
)
FIG. 2. HPLC elution patterns of PAM activity determined using enzyme in bovine pituitary. Conditions were described under Materials and Methods. The concentrations of copper ion and ascorbic acid were fixed at 1 pM and 2 mM, respectively. Peaks: 1, unknown compound; 2, Dabsyl-Gly-Phe-Gly; 3, Dabsyl-NLeu; 4, Dabsyl-GlyPhe-NH,. Three hundred picomoles of Dabsyl-NLeu (internal standard) was added to each sample after incubation. (A) Blank incubation: Dabsyl-Gly-Phe-Gly was incubated without enzyme at 37“C for 4 h. (B) Standard incubation: 300 pmol of Dabsyl-Gly-Phe-NH, was added to a sample tube before incubation as a standard sample. The two peak areas of Dabsyl-Gly-Phe-NH, and Dabsyl-NLeu correspond to 50 pmol. (C) Experimental incubation: Dabsyl-Gly-PheGly was incubated with bovine pituitary extract at 37°C for 4 h. (D) Control incubation: a control tube without the enzyme was incubated, the same amount of active enzyme was added, and the resulting tube was kept in an ice bath before heating at 95’C for 5 min.
ENZYME
265
ASSAY
graphic patterns of the reaction mixture after incubation with 21.4 pg of protein prepared from bovine pituitary for 4 h. The blank incubation (Fig. 2A) contained Dabsyl-Gly-Phe-Gly and Dabsyl-NLeu, and the standard incubation contained exogenous Dabsyl-GlyPhe-NH, in addition to Dabsyl-Gly-Phe-Gly and Dabsyl-NLeu (Fig. 2B). The retention times for Dabsyl-Gly-Phe-Gly, Dabsyl-NLeu, and Dabsyl-GlyPhe-NH, were 2.6, 3.9, and 4.8 min, respectively (Figs. 2A and 2B). The experimental incubation under the standard assay conditions (Fig. 2C) showed a significant amount of Dabsyl-Gly-Phe-NH, at 4.8 min, whereas the control incubation did not show any peak of Dabsyl-Gly-Phe-NH, (Fig. 2D). The enzyme reaction was found to be linear with time at 37°C at least for about 8 h (Fig. 3). The catalytic activity of the enzyme was greatest at a pH of approximately 5.0 to 5.5, with a rapid loss of activity at pH below 4.5 (data not shown). PAM activity was investigated as a function of the amount of enzyme extract obtained from bovine pituitary. Perfect linearity was observed for plots of the amount of Dabsyl-Gly-Phe-NH,, at least from 0.14 to 2.15 units, formed enzymatically from Dabsyl-GlyPhe-Gly against those of enzyme (data not shown). In experiments designed to optimize the PAM assay conditions, various metals were analyzed for their ability to stimulate the PAM activity. Of the metal tested, the only metal that exerted a stimulatory effect on the PAM activity was copper. The optimal copper concentration required for this enzyme was about 1 pM (Fig. 4).
Chromatographic conditions. Analysis of the product was performed using a Japan Spectroscopic Co. Ltd. HPLC system consisting of a 880-PU pump, 875-uv detector (fixed at 460 nm), 860-CO column oven, 880-50 degasser, 880-02 gradient unit, 802-SC system controller, and 807-IT integrator. The system was operated at 35°C at a flow rate of 0.75 ml/min employing a TSK gel ODS80TM (particle size, 5 pm) reversed-phase column (150 X 4.6 mm i.d.) fitted with a TSK guard gel ODS-80TM (15 X 3.2 mm i.d.; particle size, 5 pm). The mobile phase consisted of 0.01 M sodium acetate buffer (pH 4.00)acetonitrile(41/59, v/v). Protein determination. Protein concentration was measured by the Lowry method as modified by Hartree (16) using BSA as standard protein.
RESULTS This HPLC-calorimetric detection system for the measurement of Dabsyl-Gly-Phe-Gly and DabsylGly-Phe-NH, was found to be very sensitive. The calibration graph for Dabsyl-Gly-Phe-NH, injected showed good linearity from 1 to 1000 pmol. The calibration graph for Dabsyl-NLeu also showed good linearity from 5 to 1000 pmol. Figure 2 shows the chromato-
Time
FIG. 3.
( hours
)
Time-dependent changes of Dabsyl-Gly-Phe-NH, enzymatically formed from the substrate. The standard assay conditions are described under Materials and Methods. The concentrations of copper ion and ascorbic acid in the reaction mixture are 1 pM and 2 mM, respectively. Incubation was carried out at 37°C for the indicated periods.
266
CHIKUMA
0’
.‘..‘m
.....‘m 0.1
0.5
......’ 1 cuso4
5
. ... 10
ET
’
50
( JJM )
FIG. 4. Determination of effects of optimal copper ion concentrations on PAM activity from bovine pituitary. The rates of DabsylGly-Phe-NH, formation by PAM from bovine pituitary extract at various copper ion concentrations were measured as described under Materials and Methods. The concentration of ascorbic acid was fixed at 2 mu. Incubation was done at 37°C for 4 h.
A higher concentration of copper consistently showed an inhibitory effect on a-amidation. The effects of ascorbic acid concentration on enzyme activity were also examined (Fig. 5). The optimal concentration of ascorbic acid for PAM activity was found to be 2 mM. There is a clear loss of the stimulately effect at high concentrations of ascorbic acid. A Lineweaver-Burk plot was obtained from the effect of the concentration of Dabsyl-Gly-Phe-Gly on the rate of formation of Dabsyl-Gly-Phe-NH, by PAM. The Michaelis constant (K,) and the maximum velocity (V-) toward the Dabsyl-Gly-Phe-Gly were calculated to be 3.5 PM and 100 pmol/pg/h, respectively. Finally, we applied this standard assay method for the detection of PAM activity in rat saliva. Under the anesthetized conditions, the muscarinic agonist oxotremorine (2 mg/kg) was intraperitoneally injected to collect saliva. About 0.5 ml of saliva for 30 min per one rat was collected. PAM activity in the saliva was 0.78 + 0.12 nmol/ml/h (mean + SE; n = 5).
AL.
has a few advantages. First, it is very sensitive. The limit of the sensitivity was about 1 pmol of Dabsyl-Gly-PheNH, formed enzymatically. Second, the substrate and the product are separated completely in less than 5 min. Third, more accurate quantitation of the product and better reproducibility were guaranteed in our method by the employment of internal standard (Dabsyl-NLeu) compared to those of the HPLC-fluorometric assay system using N-Dansyl-D-Tyr-Val-Gly as substrate, in which an internal standard had not been used. Finally, an important structural feature of the dabsylated substrate is that the Dabsyl group prevents the ionization of the amino terminus of the peptides. Thus, the net charge of the substrate remains constant in the pH range of 5 to 9, the same as other substrates for PAM, N-Dansyl-D-Tyr-Val-Gly (12,19) and N-trinitrophenyl-D-Tyr-Val-Gly (13). In comparison, the radioiodinated substrate D-Tyr-Val-Gly goes through a transition form in the same pH range. The structural change may complicate the assessment of pH effects on enzyme activity. The use of the dabsylated substrate permits the unambiguous study of pH effects. The stimulating effects of copper ion on PAM activity have been described previously in several tissues, and the addition of copper ion was also essential to the detection of PAM activity in this assay system (optimal concentration was 1 PM). The concentrations of copper ion for optimal PAM activity were very different among tissues. Those from rat pituitary, rat hypothalamus, rat stomach, and rat medullary thyroid carcinoma CA-77 cell culture medium required 2,25,50, and 30 pM copper ion, respectively (3,6,20,21). Therefore, the copper ion requirement for different tissues should be tested individually. Ascorbic acid has been shown to be of widespread importance in a number of key biochemical processes in-
DISCUSSION
The first assay used for the detection of PAM activity was developed by Bradbury and co-workers (2), who utilized ‘2SI-labeled D-Tyr-Val-Gly as a substrate. This assay method has been used by several other laboratories (4,17). Herein we reported a new assay method for PAM activity by the HPLC-colorimetric detection system using Dabsyl-Gly-Phe-Gly as substrate. The sequence of Cterminal Phe-Gly was selected because it appears to be the most reactive substrate for PAM ever reported (18). The proposed sensitive assay method for PAM activity
Ascorbic
acid
( mM )
FIG. 5. Determination of effects of optimal ascorbic acid concentrations on PAM activity from bovine pituitary. The rates of DabsylGly-Phe-NH, formation by PAM from bovine pituitary extract at various concentrations of ascorbic acid were measured as described under Materials and Methods. The concentration of copper ion was fixed at 1 PM. Incubation was done at 31°C for 4 h.
COLORIMETRIC
a-AMIDATING
eluding the hydroxylation of dopamine to form norepinephrine (22,23) and collagen biosynthesis (24). In contrast to the optimal concentration of copper ion for PAM activity, those of ascorbic acid were not significantly different among tissues. The inhibitions of PAM activity at supraoptimal concentrations of copper ion and ascorbic acid may be the result of enzymatic inactivation caused by free-radical production induced by ascorbic acid and superoxide in the presence of copper ion (25,26). Mains et al. (27) reported that PAM activity in the submandibular gland was increased after treatment with the a-adrenergic antagonist phenoxybenzamine and decreased after treatment with the a-adrenergic agonist phenylephrine. Moreover, serum levels of PAM activity were decreased after the treatment with phenoxybenzamine. From these data we propose that PAM may be released from salivary glands to the saliva and/ or serum. In the present study, we proved the presence of PAM activity in rat saliva. The effects of various chemical reagents and protease inhibitors on PAM activity in bovine pituitary extract were examined to investigate the possibility that PAM activity was affected by confounding protease activity in enzyme source. PAM activity was not affected by serine protease inhibitor (PMSF), acid protease inhibitor (pepstatin A), and soybean trypsin inhibitor. Furthermore, PAM activity was not affected by bacitracin. However, EDTA and l,lO-phenanthroline inhibited the enzyme activity completely at a final concentration of 0.1 mM, presumably by chelation of the enzyme-bound copper prosthetic group. The enzyme activity was also inhibited by the addition of 1 mM of dithiothreitol, but the inhibitory effects of the thiol compound were restored by the addition of NEM and/or PCMB, suggesting that the addition of typical thiol inhibitor may overcome the inhibitory effects of endogeneous thiol compounds (5,28). These results suggest that PAM activity was not affected by confounding protease activity during long incubation times. PAM is the key enzyme involved in the posttranslational cY-amidation of many biologically active peptides. The understanding of the mechanisms controlling amidated-peptide production will increase through an investigation of the regulation of PAM activity in Go. In conclusion, the ease of handling of this compound compared to the radioactive material makes Dabsyl-GlyPhe-Gly an excellent substrate for the study of PAM. Furthermore, the fast and accurate assay method described in this paper may be a useful means for the purpose described above. ACKNOWLEDGMENTS The authors Pharmaceutical
gratefully Analytical
thank Dr. Chemistry
Akira Tanaka, and Dr. Tomoji
Department of Kocha, Depart-
ENZYME
267
ASSAY
ment of Hygienic Sciences, for helpful
Chemistry, discussions.
Showa
College
of Pharmaceutical
REFERENCES 1. Mains, (1983)
R. E., Eipper,
B. A., Glembotski,
C. C., and Dores,
R. M.
Trends Neurosci. 6, 229-235.
2. Bradbury, A. F., Finnie, M. D. A., and Smyth, 298,686-688. 3. Eipper, B. A., Mains, R. E., and Glembotski,
G. (1982)
D.
Nature Proc.
C. C. (1983)
Natl. Acad. Sci. USA 80,5144-5148. 4. Eipper,
B. A., Myers,
A. C., and Mains,
R. E. (1985)
Endocrinology
116,2497-2504. 5. Mizuno, (1986) 6. Sakata,
K., Sakata,
J., Kojima,
M., Kangawa,
K., and Matsuo,
H.
Biochem. Biophys. Res. Commun. 137,984-991. J., Mizuno, K., and Matsuo, H. (1986) Biochem. Biophys. Res. Commun. 140,230-236. 7. Murthy, A. S. N., Mains, R. E., and Eipper, B. A. (1986) J. Biol. Chem. 261,1815-1822.
8. Kizer, J. S., Bateman, W. H., and Youngblood, 2267.
R. C., Miller, W. W. (1986)
9. Mehta, N. M., Gilligan, J. P., Jones, B. A., and Birnbaum, R. S. (1988) 44-54. 10. Mollay, 254.
C., Wichta,
11. Husain, 12. Jones, Lovato, (1988)
J., and Kreil,
S. S. (1983)
I., and Tate, B. N., Tamburini, S. J., Gilligan,
C. R., Humm,
J., Busby,
Endocrinology
118,2262-
B. N., Bertelsen,
A. H., Roos,
Arch. Biochem. Biophys. 261,
G. (1986) FEBS L&t. 202,251FEBS Lett. 152, 277-281.
P. P., Consalvo, A. P., Young, J. P., Jeng, A. Y., and Wennogle,
S. D., L. P.
Anal. Biochem. 168,272-279. A. G., and May, S. W. (1988) B&hem. Biophys. Res. Commun. 151,499-505. 14. Lin, J. K., and Chang, J. Y. (1975) Anal. Chem. 4’7,1634-1638. 15. Glembotski, C. C., Eipper, B. A., and Mains, R. E. (1984) J. Biol. Chem. 259,6385-6392. 16. Hartree, E. F. (1972) Anal. Biochem. 48,422-427. 17. Glembotski, C. C. (1985) Arch. Biochem. Biophys. 241,673-683. 18. Bradbury, A. F., and Smyth, D. G. (1983) Biochem. Biophys. Res. Commun. 112,372-377. 19. Bendig, M. M. (1986) J. Biol. Chem. 261,11,935-11,937. 20. Emeson, R. B. (1984) J. Neurosci. 4,2604-2613. 13. Katopodis,
21. Jeng, A. Y., Wong, M., Lovato, S. J., Erion, J. P. (1990) Anal. B&hem. 185,213-219. 22. Levin, E. Y., Levenberg, B., and Kaufman,
M. D., and Gilligan,
S. (1960)
J. Biol.
Chem. 235,2080-2086. 23. Friedman, S., and Kaufman, S. (1965) 4773. 24. Murad, S., Sivarajah, A., and Pinnell,
J. Biol. Chem. 240,4763S. R. (1981)
Biochem.
Biophys. Res. Commun. 101,868-875. 25. Davison, A. J., Kettle, 261,1193-1200. 26. Shinar, E., Navok, 14,778-14,783. 27. Mains,
R. E., Myers,
A. J., and Fatur,
T., and Chevion,
D. J. (1986)
M. (1983)
A. C., and Eipper,
J.
J. Biol. Chem.
Biol. Chem. 268,
B. A. (1985)
Endocrinology
116,2505-2515. 28. Mizuno, K., Kojima, M., Sakata, J., and Matsuo, H. (1986) in Peptide Chemistry 1985, pp. 351-356, Protein Research Foundation, Osaka, Japan. 29. Tajima, M., Iida, T., Yoshida, S., Komatsu, K., Namba, R., Yanagi, M., Noguchi, M., and Okamoto, H. (1990) J. Biol. Chem. 265,9602-9605.