Eur. J. Biochem. 209, 189-194 (1992) 0FEBS 1992

Characterization of a Xenopus laevis skin peptidylglycine a-hydroxylating monooxygenase expressed in insect-cell culture Hiroko SHIMOI', Takashi KAWAHARA', Kenji SUZUKI ', Yasuno IWASAKI ', Arco Y. JENG' and Yoshiki NISHIKAWA'

'

Bio-organics Research Department, International Research Laboratories, Ciba-Geigy (Japan) Limited, Takarazuka, Japan Research Department, Pharmaceuticals Division, Ciba-Geigy Corp. Summit, USA

(Received March 31/July 13, 1992)

-

EJB 92 0458

The C-terminal amide structure of peptide hormones and neurotransmitters is synthesized via a two-step reaction catalyzed by peptidylglycine a-hydroxylating monooxygenase (PHM) and

peptidylhydroxyglycine N-C lyase. A Xenopus laevis PHM expressed in insect-cell culture by the baculovirus-expression-vectorsystem was purified to homogeneity and characterized. Using a newly established assay system for PHM, the kinetic features of this enzyme were investigated. As expected, the enzyme required copper ions, L-ascorbate and molecular oxygen for turnover. Salts like KI and KCI, and catalase stabilized the enzyme in the presence of L-ascorbate. The optimum pH value for the enzyme reaction was around six when Mes buffer was used and around seven when phosphate buffer was used under the same assay condition. Below pH 6, acetate, iodide and chloride ions activated the reaction. The kinetic analysis is consistent with a ping-pong mechanism with respect to peptide and L-ascorbate, and the peptide showed substrate inhibition. The substrate specificity of the enzyme at the penultimate position was examined by competitive assay using tripeptides with glycine at the C-termini and the inhibitory potency of these peptides in descending order was methionine > aromatic > non-polar amino acids.

Many biologically active peptide hormones and neurotransmitters have C-terminal amides. In most cases, the Cterminal amide is essential for biological activity. Bradbury et al. first detected an enzyme activity in porcine pituitary which catalyzed the conversion of glycine-extended peptides to amidated peptides [I]. Since then, the purification and characterization of the enzyme has been reported from frog skin [2, 31, porcine pituitary [4], bovine neurointermediate pituitary [5, 61, rat medullary thyroid carcinoma [7] and porcine atrium [XI. Glembotski et al. [9] named the enzyme peptidylglycine aamidating monooxygenase (PAM). All these enzyme preparations require copper ions and L-ascorbate for catalytic activity, but they differ in molecular mass, 38 - 75 kDa, and in optimal pH [2 - 91. The cDNA clones for the enzymes have been isolated from bovine pituitary [lo], frog skin [I1 - 131, rat atrium [14, 151, rat medullary thyroid carcinoma [16], rat pituitary [I71 and human thyroid carcinoma [18]. The entire coding regions of these cDNA clones encode 850- 1000 amino acids (Fig. 1a), with the exception of the major frog type [amidating enzyme 1 (AE-I) in [ll]; Fig. 1a]. AE-I contains only the N-terminal Correspondence to Y. Nishikawa, Bio-organics Research Department, International Research Laboratories, Ciba-Geigy, Japan Limited, P. 0. Box 1, Takarazuka, Japan 665 Abbreviarions. PAM, peptidylglycine a-amidating monooxygenase; PHM, peptidylglycine a-hydroxylating monooxygenase; PHL, peptidylhydroxyglycine N-C lyase; DBH, dopamine fi-hydroxylase; Dns, dansyl; AE, amidating enzyme. Enzymes. Peptidylglycine a-amidating monooxygenase (EC 1.14.17.3); peptidylglycine a-hydroxylating monooxygenasc (EC 1.14.17.-); peptidylhydroxyglycine N-C lyase (EC 4.3.2.5); dopamine P-hydroxylase (EC 1.14.17.1).

portion (400 amino acid residues) of the others. One minor form of rat atrium type (rPAM-4 in [15]) is also similar to AE-I in length. Recently, it has been demonstrated that the amidating process is a sequential, two-step reaction catalyzed by two enzymes, peptidylglycine x-hydroxylating monooxygenase (PHM) and peptidylhydroxylglycine N-C lyase (PHL; Fig. 1b) [19 -211. First, the p r o 3 hydrogen on the a-carbon of the C-terminal glycine residue is removed and the ( S ) hydroxyglycine-extended peptide is produced [22]. The product of PHM is then converted to the amide form by PHL [22]. The surprising fact is that these two enzymes are parts of the same precursor (Fig. la). PHM and PHL are linked by one amino acid residue, arginine, which is an endoprotease processing site in the frog [13].Moreover, PHM is processed from the precursor proteins encoded by both AE-I and AE-111. In other words, the nucleotide sequences of the PHM portions in both AE-I and AE-I11 are identical. This suggests that AE-I and AE-111 mRNA are produced by alternative splicing from one transcript of one gene [13]. Despite previous intensive studies characterizing the enzymic amidating reaction, one should now focus on each enzyme, because the former studies were carried out considering the whole amidating reaction, and not the two-step reaction by PHM and PHL. There are several reasons for this. The amidation process was not understood clearly at the beginning; although a hydroxy intermediate was suggested [23], a single enzyme was believed to catalyze the complete amidating reaction. The assay systems used by various laboratories quantified the final amidated product and not the a-hydroxy derivatives, which were not considered to accumu-

190 a

kl

PHM

I I

AE-Ill Is1

PHM

I

PHL

61

PHM

I

PHL

Frog AE-I

Rat PAM-2

peptidylglycine

peptidyl hydroxyglycine

I R -.- ~. -

pepiidylamide

7

1

glyoxylate

Fig. 1. Primary structure of the precursor proteins of amidating enzymes (a) and the relationship between two enzyme domains and a two-step amidating reaction (b). (a) Schematic drawings of precursor proteins of three frog PHM (AE) [ll - 131and a rat enzyme [14]. S, signal sequence; filled region, putative transmembrane domain; shadowed region, repetitive sequence. (b) Two enzymes, PHM and PHL, involved for a sequential two-stcp reaction of amidation are processed from a precursor protein.

late during the reaction. It was quite difficult to obtain enough of the pure enzyme from various sources. to express PHM and PHL we are and measure hydroxylase and lyase activities independently. In this paper* we studied the characteristics Of the monooxygenase, AE-l from Xenopus laevis, which was expressed in insect-cell culture by using the baculovirus expression-vector system [19]. MATERIALS AND METHODS Materials Dns-Tyr-Phe-Gly (Dns, dansyl) was synthesized by the solid-phase method, followed by N-dansylation [19]. Catalase (bovine liver type I) was obtained from Boehringer Mannheim GmbH, L-ascorbic acid from Sigma, sodium bathocuproine disulfonate from Dojin chemicals and Lubrol PX from Amresc. Tripeptides, Gly-Xaa-Gly (see Table 3 for Xaa), were synthesized using the solid-phase procedure and purified by HPLC. Other peptides were from Nova biochemicals. All other chemicals were of reagent grade. PHM was expressed by using the baculovirus-expressionvector system and purified from insect-cell-culture medium as described previously [19], except that the buffer containing 10mM KI was used in all purification steps. The enzyme was stored in 10 mM Tris/Cl, pH 7.5, containing 10 mM KI, 150 mM NaCl and 0.1 YOLubrol PX at -20°C without any loss of activity. We used the same batch of PHM for all the experiments described below. Ultraviolet spectrum of the purified PHM Absorption spectra of the purified enzyme (0.1 mg/ml) were measured in 10 mM TrisjCt, pH 7.5, containing 0.1% Lubrol and 150 mM NaCl with a Beckman DU-62 spectrophotometer. A sample purified without KI was used, because KI interfered with the measurement of spectra.

Assay for PHM activity The activity of PHM was measured as described previously [I91 with a slight modification. The following reaction mixture was o.2 Mes,Na, 2o pM Dns-Tyr-Phe-Gly, 2o mM L-ascorbate, 10 mM KI, 0.1 mg/ml and 0.1% Lubrol PHM was first diluted p x , pH 6.0, in a final of 40 with 5o mM Mes,Na (PH 6,0) containing ,o mM KI and 0.1% Lubrol PX, and 2 p1 diluted PHM was used for each test. The activity of the diluted PHM was stable at 4°C for at least 12 h. Reactions were initiated by the addition of PHM, run at 30°C for 5 - 15 min and stopped by the addition of 8 p1 1 M NaOH containing 0.25 M EDTA. During the alkaline treatment for 5 min, the product peptide, Dns-Tyr-PheGlyOH (GlyOH, a-hydroxyglycine) formed by PHM was converted quantitatively to the corresponding amide, which can be measured fluorometrically. For the kinetic experiments using various oxygen concentrations, we used a Thunberg tube to which two arms were attached. The tube was siliconized with dimethyldichlorosilane to avoid the adsorption of the enzyme on the surface of the tube. The reaction mixture in a total volume of 0.5 ml had the same composition as that used in standard assays of PHM. The O2 concentration in the reaction mixture was controlled by sending O 2gas from the arm. The reaction mixture without L-ascorbate was incubated in the closed tube in various O2 concentrations for 5 min. O 2 concentration in the reaction mixture was measured using an O2 electrode, which was set in the tube before beginning the reaction. The reaction was initiated by the addition of L-ascorbate (2 M, 5 pl) with a microsyringe through the rubber stopper on the arm of the tube. After incubation for 10 min, the reaction was stopped by addition of EDTA and the product was quantified as described above. Assay for PHL activity The activity of PHL was measured as described previously 1131.

191 Table 1. Effect of various metal ions on PHM activity in the presence of bathocuproinedisulfonate.The enzyme (6.25 nM) was incubated for 5 min on ice with bathocuproine disulfonate (BD; final concentration, 10 pM) in a standard assay without catalase or Dns-Tyr-Phe-Gly. The indicated metal ion (final concentration, 6 pM) was then added, and finally Dns-Tyr-Phe-Gly was added to initiate the enzyme reaction.

Addition

Salt (25 mM)

Catalase (0.1 mg/ml)

Remaining activity with L-ascorbate at

Specific activity prnol. min-

None

43

Cu2+ (1 pM) BD (10 pM) BD/Cu2+ (6 pM)

43

BD/MgZ+/Mn2+/Fe3+/CoZi / NiZ+/Znzi/Cdz+/BaZ+ (6 pM)

Table 2. Effects of salts and catalase on the stability of PHM. 1 pg PHM was incubated in 40 p10.2 M Mes/Na containing 0.1 % Lubrol PX at 30"C, with or without 2 mM L-ascorbate. The enzyme activity remaining after 15 min was measured by the standard assay method.

0 mM

' . mg-'

0 39

0

Protein assay Protein concentration was estimated by Coomassie-blue protein assay reagent (Pierce) using bovine serum albumin as a standard. The absolute concentration of the purified enzyme was estimated by amino acid analysis of the hydrolyzed enzyme.

RESULTS Purification of recombinant PHM One of the frog amidating enzymes, AE-I, was expressed in insect cell culture using the baculovirus-expression-vector system, and the enzyme was purified from the cell-culture medium. Starting from 500 ml insect cell-culture medium, two purification steps, S-Sepharose and Mono Q , were employed to provide 1.O mg enzyme with a specific activity of 45 pmol . min-' . mg- ' under standard assay conditions. The active preparation was shown to be homogeneous by SDS/ PAGE in the presence of 2-mercaptoethanol. The purified PHM showed a normal protein ultraviolet spectrum with no distinguishing characteristics (data not shown). The preparations did not show any PHL activity. Effects of metal ions and metal-ion chelators on hydroxylating activity The requirement of metal ions for hydroxylating activity was investigated (Table 1). Catalase was not included in the reaction mixture, because it was not necessary to add the enzyme for analysis in a short reaction time (5 min). At 0 - 1 pM CuCI2, enzyme activity was almost constant, but at higher concentrations (> 1 pM), the initial velocity of the reaction was reduced (data not shown). When the enzyme was incubated in the standard assay mixture without catalase containing 10 pM bathocuproine disulfonate, a cupric ion chelator, no hydroxylating activity was observed. Only the addition of CuC1, to the bdthocuproine disulfonate-treated enzyme restored enzyme activity. The addition of other metal ions showed no effect. These results indicate that PHM requires copper ions as an essential element for activity and that other metals cannot replace the copper ions. Stability of purified PHM The purified enzyme was incubated at 30°C in 0.2 M Mes/ Na at pH 6.0 in the presence or absence of KCl, KI or catalase

2 mM

%

KI KC1 -

KI KC1

-

41

-

62 47 70 93 85

-

+ + +

0.3 17 3.9 73 95 73

for 15 min. The remaining activity was measured in the standard assay method described in Materials and Methods (Table 2). In the absence of L-ascorbate, the enzyme was unstable without additives. KI and catalase showed stabilizing effects. In the presence of L-ascorbate, the enzyme incubated without catalase was unstable. Catalase stabilized the enzyme significantly in that case. The greatest stability in the presence of L-ascorbate was obtained after the addition of both KI and catalase. KC1 showed partial stabilizing effects. pH profile The optimum pH of the reaction was around pH 6 when Mes buffer was used and around pH 7 when phosphate buffer was used, as shown in Fig. 2a. The pH profiles varied depending on the assay buffer at lower pH, suggesting that not only pH, but also the kind of buffer, affected hydroxylating activity. To study the effect of the buffer, we further investigated the activation of hydroxylating activity by addition of various concentrations of anions at pH 5.3 (Fig. 2b). The acetate, iodide and chloride ions activated the hydroxylation at pH 5.3 in 0.2 M Mes/Na buffer, whereas, sodium acetate (10- 100 mM) did not show any effect in 0.2 M sodium acetate buffer. At pH 6.0, we detected some inhibitory effect of the salts (data not shown). Kinetics A series of steady-state kinetic analyses were conducted to investigate the reaction mechanism of PHM. Double-reciprocal plots of initial velocity against Dns-Tyr-Phe-Gly concentrations in the presence of various fixed concentrations of L-ascorbate are shown in Fig. 3a. The parallel plots obtained with varying L-ascorbate concentrations at high l/[Dns-TyrPhe-Gly] are the same as those observed in double reciprocal plots for enzymes following a ping-pong mechanism. When 1/[Dns-Tyr-Phe-Gly] was decreased, the plots passed through a minimal value and bent up, suggesting substrate inhibition. As the concentration of L-ascorbate increased, the position of the minimum moved closer to the x-axis. This pattern suggests that the enzyme reaction proceeds through a ping-pong mechanism with competitive substrate inhibition. If L-ascorbate binding and reduction of the enzyme is required before binding of the other substrates, as in the case of dopamine P-hydroxylase (DBH) [24], binding of Dns-Tyr-Phe-Gly to the oxidized enzyme would lead to unproductive complex forma-

192 0.5

7

-

Interaction Qf PHM with tripeptides

a

There were several reports studying the structure/activity relationship of peptide substrates for PAM. However, since only amidated products were quantified, it is not clear whether the data reflected the catalytic competence of PHM, PHL or a combination. We examined the binding affinity of several glycine extended peptides to PHM in order to study the structure/activity relationship. The ability of various glycineextended peptides to inhibit the hydroxylation of Dns-TyrPhe-Gly was determined in the standard assay mixture. K, for various peptides given in Table 3 was determined using Dixon plots. Different amino acid residues adjacent to the C-terminal glycine influenced their inhibiting ability in the hydroxylating reaction of PHM differently. Inhibitory potency was the highest in methionine and decreased through aromatic, nonpolar and polar amino acid residues and proline or glycine in that order.

0.4-

F

r

.-e

0.3-

-E 0

E, 0.2N

s v

>

0.1

-

o ! 3

5

7

PH

DISCUSSION

' I 0 0

40

80

120

Concentration (mM)

Fig. 2. pH profile and effect of anions on the hydroxylating reaction. (a) pH dependence of apparent activity of PHM. Enzyme was incubated for 5 min at 30°C with 20 pM Dns-Tyr-Phe-Gly, 20 mM I -ascorbate, 10 mM KI, 0.1 mg/ml catalase and 0.1 % Lubrol PX in the following buffers (0.2 M); (0)sodium acetate; ( A ) Mes/Na; (fl) sodium phosphate. (b) Effect of salt concentration on the activity of PHM at pH 5.3. The assay was performed under thc same conditions as described in (a) with 0.2 M sodium acetate, pH 5.3, containing different concentrations of Mes/Na buffer, pH 5.3 (0),and with 0.2 M Mes/Na. pH 5.3, containing different concentrations of sodium acetate buffer, pH 5.3 (A), KI ( O ) ,KCI (17).

tion. No substrate inhibition by L-ascorbate was observed in the range 0.5-8 mM L-ascorbate. From the intercept replot (Fig. 3 b,c), we obtained an apparent K , for Dns-Tyr-Phe-Gly of 7.0 pM, an apparent K, for L-ascorbate of 4.8 mM and an apparent V,,, of 160 pmol . min-' . mg-'.

O2 dependence The [O,] dependence of enzyme activity under standard assay conditions was investigated. From a double-reciprocal plot of initial velocity against [O,], as shown in Fig. 4, the apparent K,,, was determined to be 71 pM.

Previous work [19] on a frog PHM (AE-I), which was expressed in insect-cell culture medium, showed that the product of the enzyme reaction was the a-hydroxyglycine form of the substrate and not the amidated peptide. In this paper, further characterization of this enzyme was performed. Several groups reported the purification and characterization of PAM 12, 4, 5, 7, 81. However, PAM was treated as an enzyme catalyzing the complete amidating reaction described above. Before the existence of PHL became apparent [19,21], the hydroxy derivative of the glycine-extended peptide was suggested to be an intermediate of the amidating reaction by Bradbury and Smith [23], and it was believed to be converted to the amide by the same enzyme [25].However, as long as the whole amidating activity was followed, one could not be aware of the presence of the second enzyme, PHL. As a result, the coexistence of both enzyme activities in PAM remained undiscovered for a long time. In reality, the a-hydroxyglycine product of PHM reaction is stable below pH 6, and this allowed its structural analysis [19,26]. The assay method eStdblished here is based on the fact that the hydroxylated product can be instantly converted to the amide structure under alkaline conditions. Since the initial report of PAM, it was suggested that the characteristic features of this enzyme are similar to those of another copper-ascorbate-dependent enzyme, DBH [27]. DBH, which is also involved in hormone synthesis in secretory granules. was purified from bovine adrenal medullae and has been well characterized both kinetically and physicochemically (reviewed by Stewart and Klinman [24]). In addition, the high similarity of nucleotide and amino acid sequence between AE-I and DBH in the putative copper-binding sites also supports the similarity of the two enzymes in the reaction mechanism [28]. In the present study, we focused on a frog PHM (AE-I) as a hydroxylating enzyme. not an arnidating enzyme, and examined the characteristics of the enzyme. The instability of PHM in some in vitrn conditions drew our attention. Metalloenzymes, including copper-containing enzymes, are known to be inactivated by active oxygen [29]. Although the instability of PHM in acidic conditions containing L-ascorbate is reported [4, 301, we found that PHM is relatively stable in the presence of high concentration of catalase and KI at pH 5 - 6, which is the physiological pH in secretory vesicles [31].

!0.0 193

a 0.04

-

20.01 r

0.03

F c

x

g

.7-

’

0 -0.5

E

-

0

0.5

[ascorbate] -l(mM

0.02

2

-’ 1.0 )

3-

v

r

>

0.01

0 0

,

I

0.1

0.2

I

I

0.3

l4

-0.5

0

0.5

1.0

[ascorbate]~’ (mM” )

[Kdansyl-Tyr-Phe-Gly1-l ( pM -l) Fig. 3. Double-reciprocal plots of the dependence of initial velocity of activity of PHM on L-ascorbate and Dns-Tyr-Phe-Gly. (a) Double-reciprocal plots of initial velocity against Dns-Tyr-Phe-Gly concentrations at a series of fixed concentrations of L-ascorbate. The concentrations of L-ascorbate are indicated in the figure. (b) l/Vaxis-interceptreplot of the data in (a). (c) llapparent-K,,,replot of the data in (a).

Table 3. Effect of a series of peptides (Gly-Xaa-Gly) on PHM activity. The capacity of various tripeptides to inhibit the hydroxylation of Dns-Tyr-Phe-Gly by PHM was determined in the standard assay mixture. Ki was calculated using a Dixon plot. r is the correlation coefficient for the Dixon plot used in the determination of Ki.

J

0.6

-

c .-

Xaa

E

T-

-

: 0

Ki

r

mM 0.4-

Met

v

Phe TYr

T-

Ser TrP

>

Val Arg

Ala

LYS

-16

-a

0

a

16

Fig. 4. Double-reciprocal plot of initid velocity against O2 concentration. The reaction conditions were as described in Materials and Methods.

The pH optimum of the enzyme is 6 - 7, depending on the buffers used. Under acidic conditions in Mes/Na buffet, below pH 6, some anions such as acetate, iodide and chloride stimulated the enzyme reaction. This has not been observed before, but in the case of DBH, anion activators (i.e. fumarate) [32, 331 and also pHdependent subunit dissociation [34J were reported. Fumarate also stimulated the reaction of PHM at pH 5.3 but did not at pH 6.0 (data not shown). The steady-state kinetic studies of the enzyme showed that hydroxylation probably proceeds through an ordered pingpong mechanism with respect to the peptide substrate and L-ascorbate. Our data demonstrate thc occurrence of an irre-

His

Pro Thr Ile Leu GlY

0.22 0.62 0.62 1.o 1.4 1.8 1.9 2.3 2.8 3.0 6.8 7.1 4.9 9.9 30

0.981 0.994 0.994 0.964 0.944 0.995 0.996 0.989 0.986 0.998 0.974 0.990 0.987 0.999 0.987

versible step between the binding of L-ascorbate and peptide substrate. This is the first report on the mechanism of substrate binding to PHM as a hydroxylating enzyme, although the porcine and bovine enzymes were reported to have kinetics similar to amidating enzymes [4, 61. The mechanism implied by our results is consistent with that reported for DBH [32, 351. In the case of DBH, studies showed negative cooperativity against L-ascorbate which may be caused by a reversible tetramer/dimer dissociation depending on pH and the concentration of L-ascorbate [34, 361. Our results are not explainable by such a mechanism for PHM, although PHM is believed to be a monomeric enzyme.

194 Compared to studies on the substrate structure/activity relationship of PAM [2, 37, 381, the properties of PHM were similar. These observations suggest the possibility that the specificity of peptide amidation is determined by PHM activity or that the substrate specificity for PHM is similar to that of PHL. Jn the present work, we confirmed that the mechanism of reaction of PHM is the same as that of DBH, in terms of cofactor requirement. We also showed that other properties of these enzymes such as stability and activation were similar. It is interesting to note that both peptide hormones and neurotransmitters are processed by enzymes with a similar mechanism. We thank Mr Michihiro Taka;, Ms Keiko Yamasaki and Ms Hiroko Togame for synthesis of glycine-extended peptides. We are also grateful to Dr Andreas G. Katopodis for critical reading of the manuscript and Dr Andrew James for English correction.

REFERENCES 1 . Bradbury, A. F., Finnie, M. D. A. & Smyth, D. G. (1982) Nature 298,686-688. 2. Mizuno, K., Sakata, J., Kojima, M., Kangawa, K. & Matsuo, H. (1986) Biochem. Biophys. Res. Commun. 137,984-991. 3. Mizuno, K., Kojima, M., Sakata, J., Nakamura, T. & Matsuo, M. (1988) in Peptide chemistry 1987 (Shiba, T. & Sakakibara, S., eds) pp. 427-430, Protein Research Foundation, Osaka. 4. Kizer, J. S., Bateman, Jr R. C., Miller, C. R., H u m , J., Busby, Jr W. H. & Youngblood, W. W. (1986) Endocrinology 118, 2262 -2267. 5. Murthy, A. S. N., Mains, R. E., & Eipper, B. A. (1986) J . Biol. Chem. 261, 1815-1822. 6. Murthy, A. S. N., Keutmann, H. T. & Eipper, B. A. (1987) M d . Endocrinoi. 1,290 - 299. 7. Mehta, N. M., Gilligan, J. P.? Jones, B. N., Bertelsen, A. H., Roos. €3. A. & Birnbaum, R. S. (1 988) Arch. Biochem. Biophys. 261,44-54. 8. Kojima, M., Mizuno, K., Kangawa, K. & Matsuo, H. (1989) J . Biochem. (Tokyo) 105,440-443. 9. Glembotski, C. C., Eipper, B. A. & Mains, R. E. (1984) J . B i d . Chem. 259,6385-6392. 10. Eipper, B. A., Park, L. P., Dickerson, I. M., Keutmann, H. T., Thiele, E. A,, Rodriguez, H., Schofield, P. R. & Mains, R. E. (1981) Mol. Endocrinol. 1,177 -790. 11. Mizuno, K., Ohsuye, K., Wada, Y.,Fuchimura, K., Tanaka, S. & Matsuo, H. (1987) Biochem. Biophys. Res. Commun. 148,546 552. 12. Ohsuye, K., Kitano, K., Wada, Y., Fuchimura, K. & Tanaka, S. (1988) Biochem. Biophys. Res. Commun. 150, 1275- 1281. 13. Iwasaki, Y., Kawahara, T., Shimoi, H., Suzuki, K., Ghisalba, O., Kangawa, K., Matsuo, H. & Nishikawa, Y. (1991) Eur. J . Biochem. 201, 551 - 559.

14. Stoffers, D. A., Green, C . B.-R. & Eipper, B. A. (1989) Proc. Natl Acad. Sci. USA 86,735-739. 15. Stoffers, D. A., Ouafik, L'H. & Eipper, B. A. (1 991) J . Biol. Chem. 266, 1701- 1707. 16. Bertelsen, A. I$., Beaudry, G. A,, Galella, E. A., Jones, B. N., Ray, M. L. & Mehta, N. M. (1990) Arch. Biochem. Biophys. 279, 87 -96. 17. Kato, I., Yonekura, H., Yamamoto, H. & Okamoto, H. (1990) FEBS Lett. 269,319- 323. 18. Glauder, J., Ragg, H., Rauch, J. & Engels, J. W. (1990) Biochem. Biophys. Res. Commun. 169, 551 -558. 19. Suzuki, K., Shimoi, H., Iwasaki, Y . , Kawahara T., Matsuura, Y. & Nishikawa, Y. (1990) EMBU J. 9,4259-4265. 20. Takahashi, K., Okamoto, H., Seino, H. & Noguchi, M. (3990) Biochem. Biophys. Res. Commun. 169, 524 - 530. 21. Katopodis, A. G., Ping, D. & May, S. W. (1990) Biochemistry 29, 61 15 - 6120. 22. Kawahara, T.. Suzuki, K., Iwasaki, Y., Shimoi, H., Akita, M., Moro-oka, Y. & Nishikawa, Y. (1992) J . Chem. SOC. Chem. Commun. 8,625 - 626. 23. Bradbury, A. F. & Smyth, D. G. (1987) Eur. .I. Biochem. 169, 579 - 5x4. 24. Stewart, L. C. & Klinman, J. P. (1988) Annu. Rev. Biochem. 57, 551 - 592. 25. Young, S. D. & Tamburini, P. P. (1989) J . A m . Chem. Soc. 111, 1933- 1934. 26. Tajima, M., lida, T., Yoshida, S., Komatsu, K., Namba, R., Yanagi, M., Noguchi, M. & Okamoto, H. (1990) J . Biol. Chem. 265,9602-9605. 27. Eipper, B. A., Mains, R. E. & Glembotski, C. C. (1983) Proc. Natl Acad. Sci. USA 80, 5144- 5148. 28. Southan, C. & Kruse, L. I. (1989) FEBS Lett. 255, 116- 120. 29. Shinar, E., Navok, T. & Chevion, M. (1983) J. Biol. Chem. 258. 14778-14783. 30. Beaudry, G. A., Mehta, N. M., Ray, M. L. & Bertelsen, A. H. (1991) J . Biol. Chem. 265, 17694- 17699. 31. Arvan, P., Rudnick, G. & Castle, J . D. (1984) J . Riol. Chem. 259. 13 567- 13 572. 32. Goldstein, M., Joh, T. H. & Garvey, T. Q . (1968) Biochemistry 7,2724-2730. 33. Ahn, N. & Klinman, J. P. (1983) Biochemistry 22, 3096-

3106. 34. Saxena, A., Hensley, P., Osborne, Jr J. C. & Fleming, P. J. (1985) J . Biol. Chem. 260, 3386-3392. 35. Fitzpatrick, P. F., Harpel, M. R . LkVillafranca, J. J. (1986) Arch. Biochem. Biophys. 249, 70- 75. 36. Stewart, L. C. & Klinman, J. P. (1991) J . Biol. Chem. 266,11537 11 543. 37. Bradbury, A. F. & Smyth D. G.(1983) Biochem. Biophys. Res. Commun. I 12, 372 - 317. 38. Tamburini, P. P., Jones, B. N.. Consalvo. A. P., Young, S. D., Lavato, S. J., Gilligan; J. P., Wennogle, L. P., Erion, M. & Jeng, A. Y. (1988) Arch. Biochem. Biophys. 267,623-631.