ISE

J*K

R

ps(oEtlpCbrtoE(wpsEipA

t

wrmtcgcEt

1

Biochemical and Biophysical Research Communications 286, 721–725 (2001)

doi:10.1006/bbrc.2001.5415, available online at http://www.idealibrary.com on

nvolvement of ERK and Protein Tyrosine Phosphataseignaling Pathways in EGCG-Induced Cyclooxygenase-2xpression in Raw 264.7 Cells

ong-Wook Park,* Yoon Jung Choi,* Seong-IL Suh,† and Taeg Kyu Kwon*,1

Department of Immunology and †Department of Microbiology, School of Medicine,eimyung University, 194 DongSan-Dong Jung-Gu, Taegu, 700-712, South Korea

eceived July 5, 2001

hempttIp

soCiptrmPpH(wpst

Ctltp

M

wRcawc

Prostaglandins play regulatory roles in a variety ofhysiological and pathological processes in immune re-ponse and inflammation. Epigallocatechin-3-gallateEGCG) is known to potent antitumor agent with anti-xidant property. We first investigated the effect ofGCG on the production of prostaglandin E2 (PGE2) and

he expression of cyclooxygenase-2 (COX-2), the rate-imiting enzyme in the synthesis of PGE2, using macro-hage cell line, Raw264.7. Our results showed thatOX-2 expression and PGE2 production are upregulatedy EGCG treatment and that this induction of COX-2 isegulated in part at the transcriptional level. In addi-ion, we demonstrated the signal transduction pathwayf mitogen-activated protein kinase (MAP kinase) inGCG-mediated COX-2 expression. The MEK inhibitor

PD098059) prevented EGCG-induced COX-2 expression,hereas sodium orthovanadate (protein-tyrosine phos-hatase inhibitor) significantly enhanced COX-2 expres-ion and PGE2 production. These results suggest thatGCG mediated COX-2 expression and PGE2 production

s associated with the activation of both the ERK androtein-tyrosine phosphatase signaling pathways. © 2001

cademic Press

Key Words: EGCG; COX-2; signal transduction; pros-aglandin, COX-2 promoter.

Tea is one of the most popular beverages in theorld, and the possible beneficial health effects have

eceived a great deal of attention. Polyphenols are theost significant group of tea components, especially

he catechin group of the flavonols. The major teaatechins are epigallocatechin-3-gallate (EGCG), epi-allocatechin (EGC), epicatechin-3-gallate (ECG), epi-atechin (EG), (1)-gallocathechin, and (1)-catechin.GCG has been considered to be a major constituent of

ea (1–3). These polyphenol components are known to1 To whom correspondence should be addressed. Fax: 82-53-255-

398. E-mail: kwontk@dsmc.or.kr.

721

ave antioxidative activities due to their radical scav-nging and metal chelating functions as well as anti-utagenic activities (4–6). Recently, Lin et al. re-

orted that EGCG could prevent the binding of NFkBo the inducible nitric oxide synthase (iNOS) promoter,hereby inhibiting the induction of iNOS transcription.t has been suggested that EGCG may play role inreventing carcinogenesis and anti-inflammation (7).The cyclooxygenase (COX) catalyze the rate-limiting

tep in prostaglandin synthesis. At least two isoformsf the enzyme are expressed in mammalian tissues,OX-1 and COX-2. COX-1 is constitutively expressed

n tissues and is thought to be involved in homeostaticrostanoid biosynthesis (8–10). COX-2 is thought to behe predominant isoform involved in the inflammatoryesponse (8–10). Prostaglandins have a central role inany normal and pathophysiological responses (11).rostaglandin synthesis can be stimulated by varioushysical and chemical agents in mammalian cells (12).owever, the effects of EGCG on prostaglandin E2

PGE2) production still remain uncertain. Therefore,e determined whether EGCG influences COX-2 ex-ression and PGE2 production in macrophage cells,ince these cells play a central role in immune regula-ion and inflammation.

Here we first reported that EGCG upregulatesOX-2 expression and PGE2 production. This induc-

ion of COX-2 is regulated in part at the transcriptionalevel. In addition, we demonstrate the importance ofhe activation of both the ERK and protein-tyrosinehosphatase signaling pathways in COX-2 induction.

ATERIALS AND METHODS

Cell culture and reagents. The macrophage cell line Raw264.7as obtained from ATCC (Rockville, MD). The cells were cultured inPMI 1640 supplemented with 2 mM L-glutamine, 100 U/ml peni-

illin, 100 mg/ml streptomycin, and 10% FCS. The cells were grownt 37°C, 5% CO2 in fully humidified air and subcultured twiceeekly. Cells were seeded on 6-well plates at 1 3 106 cells/well. The

ells were stimulated for various lengths of time ranging from 1 to

0006-291X/01 $35.00Copyright © 2001 by Academic PressAll rights of reproduction in any form reserved.

24 h in presence of EGCG with or without inhibitors. Anti-COX-2anJRbsS

pEXpf1tDf

(SRMsfCwae

sv2p(Gar2tut

ctwt(p

R

E

mdtwctLst

a2R

C

pcw1rpllapPcd(

prR

tmbRm(es1mPsd

Vol. 286, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

nd anti-Hsp70 antibodies were purchased from Santa Cruz Biotech-ology Inc. (Santa Cruz, CA). Anti-phospho-ERK and anti-phospho-NK were purchased from New England Biolabs Inc. (Beverly, MA).O 31-8220, SB 203580, PD 098059, PDTC (pyrrolidinedithiocar-amate) and LY 294002 were purchased from Biomol (Biomol Re-earch Laboratories, Inc., PA). Other chemicals were purchased fromigma.

Western blot analysis. Cellular lysates were prepared by sus-ending 1 3 106 cells in 100 ml of lysis buffer (137 mM NaCl, 15 mMGTA, 1 mM sodium orthovanadate, 15 mM MgCl2, 0.1% Triton-100, 25 mM Mops, 2 mg/ml proteinase inhibitor E64, adjusted toH 7.2). The cells were disrupted by sonication and extracted at 4°Cor 30 min. Fifty micrograms of cell lysate were electrophoresed on0% SDS–polyacrylamide gels. The proteins were electro-transferredo Immobilon-P membranes (Millipore Corporation, Bedford, MA).etection of the specific proteins was carried out with an ECL kit

ollowing the manufacturer’s instructions.

RNA isolation and reverse transcriptase-polymerase chain reactionRT-PCR). Total RNA was isolated according to Chomczymski andacchi (13). Single-strand cDNA was synthesized from 2 mg of totalNA using M-MLV reverse transcriptase (Gibco-BRL, Gaithersburg,D). The cDNA for COX-2 and actin were amplified by PCR with

pecific primers. The sequences of the sense and antisense primersor COX-2 were 59-CCGTGGTGAATGTATGAGCA-39 and 59-CTCGCTTCTGATCTGTCTT-39, respectively. Conditions for PCRere 13 (94°C, 3 min); 303 (94°C, 45 s; 58°C, 45 s; and 72°C, 1 min);nd 13 (72°C, 10 min). PCR products were analyzed by agarose gellectrophoresis and visualized by ethidium bromide.

DNA transfection and luciferase assay. COX-2 promoter con-tructs were generously provided by Dr. H. Inoue (National Cardio-ascular Center Research Institute, Japan). Briefly, the region from1432 bp to 159 bp of COX-2 promoter was cloned into pGL2. COX-2romoter plasmid was transfected into human embryonic kidneyHEK) 293 cells using the Lipofectamine reagent (Life Technologies,rand Island, NY) according to the manufacturer’s instructions. Tossess the COX-2 promoter luciferase, cells were collected and dis-upted by sonication in lysis buffer (25 mM Tris-phosphate, pH 7.8,mM EDTA, 1% Triton X-100, and 10% glycerol). After centrifuga-

ion, aliquots of the supernatants were tested for luciferase activitysing the luciferase assay system (Promega, Madison, WI) accordingo the manufacturer’s instructions.

PGE2 production assay. Cells were plated in 6-well plates at 106

ell per well and allowed to culture for 24 h. Twenty-four hours later,he culture medium was replaced with fresh serum-free mediumith or without EGCG for various time points. PGE2 secretion into

he culture medium was measured by a PGE2 enzyme immunoassayEIA) kit (Cayman Chemicals, Ann Arbor, MI). All experiments wereerformed in duplicate.

ESULTS

GCG Induces COX-2 Enzyme Activity

Lipopolysaccharide (LPS) by itself activates mouseacrophages to express COX-2 and produce PGE2. To

etermine whether EGCG, major compound of greenea, was involved in signal pathway transduction path-ay leading to PGE2 releasing and COX-2 expression

aused by LPS, we monitored PGE2 concentrations inhe culture media of cells stimulated with EGCG orPS or EGCG plus LPS. As shown in Fig. 1, LPSignificantly induced COX-2 in Raw264.7 cells. Co-reatment and pretreatment with EGCG did not alter

722

mounts of COX-2 and PGE2 production. Interestingly,5 mM EGCG upregulated COX-2 expression level inaw264.7 cells.

haracterization of COX-2 Expression Inducedby EGCG in Raw264.7 Cells

To investigate whether EGCG could induce PGE2

roduction in the Raw264.7 cells, we monitored PGE2

oncentrations in the culture media of cells stimulatedith EGCG. Treatment with EGCG (25–100 mM, for2 h) caused a concentration-dependent increase in theelease of PGE2 and the expression of a 70 kDa COX-2rotein in Raw264.7 cells (Fig. 2A). There was a corre-ation between the release of PGE2 and the expressionevel of COX-2 protein. To further elucidate the mech-nism responsible for the changes in amounts of COX-2rotein, we determined levels of COX-2 mRNA by RT-CR. Treatment with EGCG resulted in marked in-reases COX-2 mRNA levels, an effect that was in-uced by EGCG in a concentration-dependent mannerFig. 2B).

In further studies of the relationship between COX-2rotein and COX-2 mRNA in Raw264.7 cells, we car-ied out time kinetic studies of EGCG treatedaw264.7 cells. Incubation with EGCG caused a time-

FIG. 1. Effects of EGCG on COX-2 expression and PGE2 produc-ion caused by LPS in Raw 264.7 cells. (A) Cellular lysate protein (50g/lane) was loaded onto a 10% SDS–polyacrylamide gel. Immuno-lot was probed with antibody specific for COX-2. Lysates were formaw 264.7 cell treated with vehicle (lane 1), 25 mM EGCG (lane 2), 1g/ml LPS (lane 3), EGCG pretreated for 2 h and treated with LPS

lane 4), EGCG plus LPS (lane 5), and 10% FBS serum (lane 6). Thequal loading in each lane was demonstrated by the similar inten-ities of Hsp 70. (B) Cells were incubated with vehicle, 25 mM EGCG,mg/ml LPS, and 25 mM EGCG plus 1 mg/ml LPS for 10 h. Theedium was removed, and the production of PGE2 was measured byGE2 enzyme immunoassay kit. Results are expressed as means andtandard deviation of three independent experiments performed inuplicate.

dtacca

iffi1st(c

M

kdtolWM(aPdymi

CsmsiaiWCii

Vol. 286, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ependent increase in COX-2 protein expression. Athe highest peak of COX-2 expression, observed 12 hnd then sustained up to 24 h after treatment. EGCGaused an induction of COX-2 mRNA in Raw264.7ells, which peaked at 8 h and then declined up to 24 hfter the treatment (Fig. 2C).To investigate whether or not EGCG-induced COX-2

nduction is due to promoter activity, transient trans-ection of COX-2 reporter gene construct was per-ormed. Used COX-2 promoters contain several bind-ng sites of the potential transcription factors within432 bases upstream of the COX-2 transcription startite (14–16). EGCG treatment significantly increasedhe promoter activity compared with no treatmentFig. 2D). These results suggest that EGCG treatmentould stimulate the COX-2 promoter region.

FIG. 2. Effects of EGCG on PGE2 production and COX-2 expressiOX-2 expression. Cells were incubated with various concentrationsoluble lysates (50 mg) were subjected to electrophoresis. The blots wedium was removed, and the production of PGE2 was measured by

tandard deviation of three independent experiments performed in dncubated for 12 h with various concentrations of EGCG and RT-PCRconcentration-dependent manner. (C) Time courses of COX-2 prote

ncubated at 37°C for the periods indicated in the presence or absestern blot analysis and RT-PCR, respectively. The results were cOX-2 promoter activity by EGCG treatment. Human COX-2 prom

ncubated with or without EGCG (75 mM) for 12 h. Cells were harvestndependent experiments and bars represent standard deviations.

723

APK Signal Pathway after EGCG Treatment

It is well established that mitogen-activated proteininase (MAPK) signaling pathways mediate COX-2 in-uction in a number of cell types (17–19). To examinehe role of the cell signaling pathway in the inductionf COX-2 by EGCG, we measured phosphorylationevel of MAPK in EGCG treated Raw264.7 cells by

estern blotting using a phosphorylation specificAPKs antibodies (Fig. 3A). By treatment with EGCG

75 mM), phosphorylation of ERK was induced 10 minfter incubation and then declined. Preincubation withD 098059 (50 mM) in culture medium before the ad-ition of EGCG significantly decreased ERK phosphor-lation. EGCG-induced phosphorylation of ERK at 30in was not suppressed by SB 203580 (p38 MAPK

nhibitor) at 10 mM. Phosphorylation of c-Jun

(A) Concentration-dependent increases in the PGE2 production andGCG for 12 h and then harvested in lysis buffer. Equal amounts ofanalyzed using a specific antibody against COX-2 and Hsp70. TheE2 enzyme immunoassay kit. Results are expressed as means and

licate. (B) COX-2 mRNA induction by EGCG treatment. Cells werealysis was performed. The expression of COX-2 mRNA increased inxpression and COX-2 mRNA induced by EGCG (75 mM). Cells weree of EGCG. Levels of COX-2 protein and mRNA were detected byrmed by repeating two independent experiments. (D) Induction ofr construct (21432/159) was transfected into HEK 293 cells andnd assayed for luciferase. Data are mean values obtained from three

on.of EerePGupan

in eenconfiote

ed a

NaEmiRsPCip

T

piirReectp

D

sctbbte2

Cvo6hwsmid

EsmEmaJwc(2ilai

Vol. 286, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

-terminal kinase (JNK) was also observed at 10 minnd sustained until 30 min. These results suggest thatGCG induce activation of ERK and JNK. To deter-ine whether or not the MAPK signaling pathways,

s involved in EGCG-induced COX-2 upregulation,aw264.7 cells were pretreated with or without MAPKpecific inhibitors in the presence of EGCG (Fig. 3B).D 098059 significantly inhibited the induction ofOX-2 protein expression, suggesting that EGCG-

nduced COX-2 expression was mediated, at least inart, through the ERK signaling pathway.

FIG. 3. Effects of SB 203580, PD 98059, and RO 31-8220 onGCG-induced phosphorylation of MAP kinase and COX-2 expres-ion. (A) Raw 264.7 cells were incubated for indicated time to deter-ine phosphorylation of ERK and JNK in the presence or absence ofGCG and indicated inhibitor. Equal amounts of soluble lysates (50g) were subjected to electrophoresis. The blots were analyzed usingspecific antibody against phosphorylated ERK and phosphorylatedNK. The lower panel shows the same blot stripped and reprobedith ERK and JNK antibodies as an internal control of the protein

ontents per lane. (B) Cells were pretreated with various inhibitors50 mM PD 98059, 10 mM SB 203580, 2 mM RO 31-8220, 25 mM LY94002, and 100 mM PDTC) for 90 min and then washed. Cells werencubated with EGCG (75 mM) for 8 h. Equal amounts of solubleysates (50 mg) were subjected to electrophoresis. The blots werenalyzed using a specific antibody against COX-2. The equal loadingn each lane was demonstrated by the similar intensities of Hsp 70.

724

he Effect of Sodium Orthovanadateon COX-2 Expression

Sodium orthovanadate is generally regarded as arotein tyrosine phosphatases inhibitor (20). To exam-ne whether protein tyrosine phosphatase was involvedn the signal transduction pathway leading to PGE2

elease and COX-2 expression caused by EGCG,aw264.7 cells were treated with EGCG in the pres-nce or absence of sodium orthovanadate, and COX-2xpression was measured (Fig. 4). When Raw264.7ells were cotreated with EGCG and sodium or-hovanadate, further potentiations of the COX-2 ex-ression were observed after 8 h of treatment.

ISCUSSION

Polyphenolic compounds in green tea have been as-ociated with lower risk of some diseases, includingancer (21, 22). Based on many studies, it is believedhat the biological responses of green tea are mediatedy its major polyphenolic constituent EGCG that haseen shown to possess exceptional antioxidant poten-ial. In additional, EGCG has anticancer effects and isffective for reducing free radical-mediated injury (23–5). These findings suggest that EGCG has anti-

FIG. 4. Effects of sodium orthovanadate on EGCG-inducedOX-2 expression and PGE2 production. Cells were incubated withehicle, 60 mM EGCG, 10 mM sodium orthovanadate, 50 mM sodiumrthovanadate, 60 mM EGCG plus 10 mM sodium orthovandate, and0 mM EGCG plus 10 mM sodium orthovandate for 8 h. Cells werearvested in lysis buffer. Equal amounts of soluble lysates (50 mg)ere subjected to electrophoresis. The blots were analyzed using a

pecific antibody against COX-2 and Hsp70. The medium was re-oved, and the production of PGE2 was measured by PGE2 enzyme

mmunoassay kit. Results are expressed as means and standardeviation of three independent experiments performed in duplicate.

ioag

cPstbropeTsC1bpvirwipitiiwhvssit(fSpv

irakcio

A

(

R

111

1

1

1

1

1

1

1

2

222

2

2

2

2

2

2

3

3

Vol. 286, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

nflammatory effects. However, the present study dem-nstrated that EGCG upregulates COX-2 expressionnd PGE2 production in Raw264.7 macrophage cells, sug-esting that EGCG enhances inflammatory processes.In the present study, COX-2 protein induction coin-

ided with the secretion of PGE2 into the culture medium.revious studies have demonstrated that COX-2 expres-ion can be regulated by transcriptional as well as post-ranscriptional mechanism (26–27). Our results obtainedy RT-PCT assay and transient transfection assay usingeport constructs, in which the report is under the controlf various lengths of the COX-2 regulatory region, sup-ort the conclusion that EGCG regulates COX-2 genexpression at least in part at the transcriptional level.he region from 2840 bp to 1123 bp includes the bindingites for many transcriptional factors such as NFkB,/EBP, CRE/ATF, E-box, STAT3, AP2 and NF-IL6 (14–6). Recently, it was reported that COX-2 expression cane regulated through different MAP kinase signalingathways and that the particular signaling pathway in-olved is dependent on the type of stimuli (17–19). Thenduction of COX-2 by PDGF has been demonstrated toequire activation of the ERK signaling pathway (17),hile constitutively active MEKK1 has been shown to

nduce COX-2 expression by activating the SEK1/MKK4-38 kinase pathway (18). Our data suggested that EGCGnduces PGE2 production via activation of PKC and par-ially by activation of ERK. The suppression of EGCG-nduced COX-2 expression by MEK inhibitor PD 98059 isn agreement with the concept that ERK signaling path-ay is important in the regulation of COX-2 by EGCG. Itas been reported that ERK can phosphorylate and acti-ate CRE/ATF, E-box and NF-IL6 (28–30). These findinguggest that EGCG treatment may activate these tran-criptional factors through ERK signaling pathway, lead-ng to COX-2 expression and subsequent PGE2 produc-ion. In addition, a role for protein-tyrosine kinasesPTKs) in control of COX-2 expression was also reportedor interleukin-1a treated human endothelial cells (31).ynergistic enhancement of expression of the COX-2 ex-ression was observed on stimulation of Raw264.7 withanadate plus EGCG.In conclusion, the present study is first report show-

ng that COX-2 expression and PGE2 production areegulated by EGCG. We provide evidence showing thatctivation of ERK MAP kinase and protein-tyrosineinase signaling pathways an important role in theontrol of COX-2 expression. The EGCG may regulatenflammatory processes by modulating PGE2 as well asther inflammatory mediators.

CKNOWLEDGMENT

This work was supported by a Korea Research Foundation GrantKRF-2000-041-F00103).

725

EFERENCES

1. Yang, C. S., and Wang, Z., Y. (1993) J. Natl. Cancer Inst. 58,1038–1049.

2. Stoner, G. D., and Mukhtar, H. (1995) J. Cell Biochem. 22,169–180.

3. Katiyar, S. K., and Mukhtar, H. (1996) Int. J. Oncol. 8, 221–238.4. Lin, Y. L., Juan, I. M., Chen, Y. L., Liang, Y. C., and Lin, J. K.

(1996) J. Agric. Food Chem. 44, 1387–1394.5. Katiyar, S. K., Agarwal, R., Zaim, M. T., and Mukhtar, H. (1993)

Carcinogenesis 14, 849–855.6. Shiraki, M., Hara, Y., Osawa, T., Kumon, H., Nakauama, T., and

Kawakishi, S. (1994) Mutat. Res. 323, 29–34.7. Lin, Y. L., and Lin, J, K. (1997) Mol. Pharmacol. 52, 465–472.8. Smith, W. L., and DeWitt, D. L. (1996) Adv. Immunol. 62, 167–

215.9. Griswold, D. E., and Adams, J. L. (1996) Med. Res. Rev. 16,

181–206.0. Herschman, H. R. (1996) Biochim. Biophys. Acta 1299, 125–140.1. DeWitt, D. L. (1991) Biochim. Biophys. Acta 1083, 121–134.2. Weksler, B. B., Eldor, A., Falcone, D., Levin, R. I., Jaffe, E. A.,

and Minick, C. R. (1982) Cardiovascular Pharmacology of theProstaglandins (Herman, A. G., Vanhoutte, P. M., Denolin, H.,and Goossens, A., Eds.), pp. 137–148, Raven Press, New York.

3. Chomczymski, P., and Sacchi, N. (1987) Anal. Biochem. 162,156–159.

4. Inoue, H., Yokoyama, C., Hara, S., Tone, Y., and Tanabe, T.(1995) J. Biol. Chem. 270, 24965–24971.

5. Lukiw, W. J., Pelaez, R. P., Martinez, J., and Bazan, N. G. (1998)Proc. Natl. Acad. Sci. USA 95, 3914–3919.

6. Kim, Y., and Fischer, S. M. (1998) J. Biol. Chem. 273, 27686–27694.

7. Xie, W., and Herschman, H. R. (1996) J. Biol. Chem. 271, 31742–31748.

8. Guan, Z., Buckman, S. Y., Pentland, A. P., Templeton, D. J., andMorrison, A. R. (1998) J. Biol. Chem. 273, 12901–12908.

9. Adderley, S. R., and Fitzgerald, D. J. (1999) J. Biol. Chem. 273,27686–27694.

0. Li, L., Eisen, A. Z., Sturman, E., and Seltzer, J. L. (1998) Bio-chim. Biophys. Acta 1405, 110–120.

1. Bushman, J. L. (1998) Nutr. Cancer 31, 151–159.2. Ahmad, N., and Mukhtar, H. (1999) Nutr. Rev. 57, 78–83.3. Garbisa, S., Biggin, S., Cavallarin, N., Sartor, L., Benelli, R., and

Albini, A. (1999) Nat. Med. 5, 1216.4. Garbisa, S., Sartor, L., Biggin, S., Salvato, B., Benelli, R., and

Albini, A. (2001) Cancer 91, 822–832.5. Ahmad, N., Gupta, S., and Mukhtar, H. (2000) Arch. Biochem.

Biophys. 376, 338–346.6. Mestre, J. R., Subbaramaiah, K., Sacks, P. G., Schantz, S. P.,

Tanabe, T., Inoue, H., and Dannenberg, A. J. (1997) Cancer Res.57, 1081–1085.

7. Daen, J. L., Brook, M., Clark, A. R., and Saklatvala, J. (1999)J. Biol. Chem. 274, 264–269.

8. Nakajima, T., Kinoshita, S., Sasagawa, T., Sasaki, K., Naruto,M., Kishimoto, T., and Akira, S. (1993) Proc. Natl. Acad. Sci.USA 90, 2207–2211.

9. Prasad, K. S., and Brandt, S. J. (1997) J. Biol. Chem. 272,11457–11462.

0. Davis, S., Vanhoutte, P., Pages, C., Caboche, J., and Laroche, S.(2000) J. Neurosci. 20, 4563–4572.

1. Hirai, K., Takayama, H., Tomo, K., and Okuma, M. (1997) Bio-chem. J. 322, 373–377.