Novel inhibitors of poly(ADP-ribose) glycohydrolase

Biochimica et Biophysica Acta, 1158 (1993) 251-256 © 1993 Elsevier Science Publishers B.V. All rights reserved 0304-4165/93/$06.00

251

BBAGEN 23857

Novel inhibitors of poly(ADP-ribose) glycohydrolase Kazumasa Aoki a, Koji Nishimura b Hideaki Abe b, Hideharu Maruta b Hiroshi Sakagarni c, Tsutomu Hatano d, Takuo Okuda d, Takashi Yoshida d, Yan-Jyu Tsai e, Fumiaki Uchiumi a and Sei-ichi Tanuma a,b,f,, Department of Biochemistry, Faculty of Pharmaceutical Sciences, Science University of Tokyo, Shinjuku-ku, Tokyo (Japan), b Department of Life Science, Faculty of Bioscienee and Biotechnology, Tokyo Institute of Technology, Yokohama, Kanagawa (Japan), c First Department of Biochemistry, School of Medicine, Showa University, Shinagawa-ku, Tokyo (Japan), a Faculty of Pharmaceutical Sciences, Okayama University, Tsushima, Okayama (Japan), e Department of Pharmacology, Faculty of Medical Sciences, Taipei Medical College, Taipei (Taiwan) and f Research Institute for Biosciences, Science University of Tokyo, Noda, Chiba (Japan) (Received 25 May 1993)

Key words: Poly(ADP-ribose) glycohydrolase;Poly(ADP-ribosyl)ation;Enzyme inhibition; Tannin; Ellagitannin; Structure-activity relationship

The inhibitory effects on poly(ADP-ribose) glycohydrolasepurified from human placenta of three classes of chemically defined tannins; galiotannins, ellagitannins and condensed tannins, were examined in vitro. Oligomeric ellagitannins were found to be most potent inhibitors of poly(ADP-ribose) glycohydrolase, their potencies increasing with increasing number of monomeric residues (dimer < trimer < tetramer). Monomeric ellagitannins and gallotannins were less inhibitory. Condensed tannins, which consist of an epicatechin gallate oligomer without a glucose core, were not appreciably inhibitory. A structure-activity study showed that higher-order conformations of the conjugates with glucose of hexahydroxydiphenoyland valoneoyl groups, which are unique components of ellagitannins, cooperatively potentiated the inhibitory activity.

Introduction

The metabolism of poly(ADP-ribose) has been shown to be involved in gene activation, such as DNA replication, repair and transcription in eukaryotic cells [1-3]. However, the physiological significance of the turnover of poly(ADP-ribose) on chromosomal proteins is not fully understood. De-poly(ADP-ribosyl)ation of specific chromosomal proteins has been reported to occur during initiation of gene expression, DNA replication and cell differentiation [4-6]. Therefore, it is possible that the catabolism of poly(ADPribose) plays an important role in regulation of gene activation. De-poly(ADP-ribosyl)ation of chromosomal proteins is mainly catalyzed by poly(ADP-ribose) glycohydrolase, which hydrolyzes glycosidic (1"-2') linkages of

poly(ADP-ribose) [7-15]. A potent and specific inhibitor for poly(ADP-ribose) glycohydrolase could be a tool for understanding the physiological role of depoly(ADP-ribosyl)ation of chromosomal proteins. So far, ADP-ribose and cAMP have been reported to inhibit poly(ADP-ribose) glycohydrolase [7-16], but they are not potent or specific for the enzyme. Previously, we found that tannin and lignin fractions extracted from green tea leaves and pine cones and chemically defined tannins were inhibitory [17-19]. In the present study, we examined the structure-activity relationship of the inhibitory effects of chemically defined tannins by using homogeneously purified poly(ADP-ribose) glycohydrolase from human placenta [15]. We show that hexahydroxydiphenoyl (HHDP) and valoneoyl residues in oligomeric ellagitannins may be important for inhibition of poly(ADP-ribose) glycohydrolase. Materials and Methods

* Corresponding author. Correspondence to: Department of Biochemistry, Faculty of Pharmaceutical Sciences, Science University of Tokyo, 12 Ichigaya-Funagawara-rnachi,Shinjuku-ku, Tokyo 162, Japan. Fax: 81 3 32683045. Abbreviations: HHDP, hexahydroxydiphenoyl; ECG, (-)-epicatechin gallate; IC50, concentration for half-maximal inhibition; ApaA, diadenosine 5',5"-pl,p4-tetraphosphate.

Materials. Chemically defined tannins (see Fig. 1 for structures) and related polyphenols were isolated from the following plants as reported previously: (-)-epicatechin gallate (ECG), (-)-epigallocatechin and ( - ) epigallocatechin gallate (Thea sinensis L.) [20]; ECG-

252 OH

R~[~o H ~.:~%

.....

~ ' J . J

o~

",0o

•"-OR~

v

OI4

O

OH Tri-O-galloyl-[3-O-glucose

OH

HO

HO

OH

HO

HO~-.~ [ll-

%Oo-~% -°"

,,o

. H .O ~ .".

OH . .

o

~c

CO---.~OCH t~ O'~ .~O

HO" 1 ~

HO~ O OH C ~

1"tl21 OH

OH

[

/..~OH

IF

o/

o ~ O ,

....

0

oc

.o

~ H O

..-~-~,,,ooo-IL R

" ~

oc

I

.o'*-,~-oH ~ ' ~ HO HOHo.~O H

HO HO OH OH HO OH

HO

U

CO

/

.o ~

co

HO

Ho~O H

Nobolanin B

Oenolhein B

.o~2~co.,,

OH

/

6.

,,o~co.Lw_~o ~ 2 ,

~r"OH

HO" J ~ L OH

~o. o .~..~....o

OH

O

X

OH

T °" co--f" ,~roH

o

Codariin A

co.. no"T-~ o ~ -CH,O"CO OH O~ O..~..CO. ~ , , , ~ OH

Oil

OH

j,~oH

,°o~:...~o

.o',/"oH

L .OH

~-f OH

~

OH

Gmaniin

/'-(

OH

O.0oo

O

~

H ' O ~ t

HO

?

.o~

-'Y

0

Casuarlclln

"o's'r"

H ! ~ CCO-OCH~ HO O O HOJ~ /~ ~ O ~ O H

}40"

V

°o °o

OH

HO.~)

)'-%oJ L_%,,

/J..:/

'-%.

.oJ,~ ~o,o~-~o o~__~oH

OH

Cornusiin A

HO

)-

\ ,o oc oc

.o~oH

,

-o.

CO CO OH I H , C O / OCO~rOH

OH

~,~

R-~°c°~

o

Oit

~.

oco:

HO OH Penla. O.galloyH).D.glucose

.o ..~,M- co .~..ocl b O ' ~ OH :

HO

od

OH OH

OCO

NO IIO OH OH

"°'~o"

OH

CO-Oo ~ °

"

O

H O ~ O H

HO

IoHO

COO

OH

Y~

HO Oll CO*-~..OClt~

go

HO

HO-]~.1 Tellimagrandin I HO

OH

HO OH Tetra- O-g alioyl- p,-O-glucose

HO CO'~.O~H) HO co.O)'~-~-~° HO ~H O''/v°'3k''~-o OH /

OH H

OCO

.o o:O:

OH OH

Ho

OH

oH

(3£O OC~

(-)-Eplcalechin: RI= RZ= H (-)-Epicatechin gallala: Rl= H, R)= Gall (-)-Epigallocalechin: R l= OH, R~'= H (-)-Eplgallocalechin _qallate: R~= OH, R~t- Gall

oC

oH

I,IO

Gall = -CO ~ o

OH

OH

(~c-~oH

HO

OH HO

o

o

O

O

OH

~o~.~-0o ~0oc ,o

-,oH

(2)CO I-(O

T

HO'~Nt~" OH

Rugosin D

T°°°~,

°"

.o~o c ~7-C" -ELF'[

OC --

Oil mo

.... o.

~,~..o-oo~"-~o. .o i

--,~

I

~O~-I

CO

o.

~

I

l

~

/_

.o,

o ~o, "°~::,.0o"~H°:.

~-.~. .... -L,'o;"/

,o~-...%..~o,,

O

HO~-

~ ~

;o

I

HO

Y"°-<°V~

HOT

~oH

HO'~

.o-c'YoT.-.~-'-~, o.

~-y.o-oo~oH ~

O ~

OH O

H

Nobolanin E CO~

oH

Oil

,-~

oo

Ho.~

m , -o-

CO"'~CMI;Hr

.o.~,~ ~o.O..~..-o

CO HO

HO

'~-~o,,

~.?o-~o~o. ~°~.-~ ~ I oH ~ " ~ " ° - ~ o - ~ ) - ~

CO

H

C~;H;r

o~o~,O -3-C34-

HO

OCHII F ' o OH .oHHO o-,,~..-o,

/

"° ° ~ ' : ~ ° ° HO-~ ~'c°°'J"°'J--c",~

"°"-c'-°°

ECG-lrimer

Nobotanin K

Fig. 1. Structures of tannins and related compounds.

o'~ oo.

~ O - ° ° T 2 , °"°" ,o'~,r"

,~,oH

-Iv

~o-

o.

o, L(_

253 dimer, ECG-trimer and ECG-tetramer (Saxifraga stolonifera Meerb.) [20]; tri-, tetra- and penta-o-galloyl/3-D-glucose (produced from tannic acid) [18]; geraniin (Geranium thunbergii Sieg. et Zucc.) [21]; casuarictin (Casuarina stricta Ait.) [21]; tellimagrandin I and cornusiin A (Comus officinalis Sieb. et Zucc.) [20]; oenothein B (Oenothera erythrosepala Borbas.) [23];

rugosin D (Rosa rugosa Thunb.) [20]; coriariin A (Coriaria Japonica A. Gray) [24]; nobotanin B and E (Tibouchina semidecandra Cogn.) [25,26]; nobotanin K (Heterocentron roseum A. Br. et Bouch) [27]. (-)-Epicatechin and [adenine-2,8-3H]NAD (30 Ci/mmol)were purchased from Sigma and DuPont-New England Nuclear, respectively. DEAE-cellulose paper (DE-81) was

TABLE I

Effect of tannins and related compounds on poly(ADP-ribose) glycohydrolase Compound

IC50 (~M)

Number of functional groups Galloyl

Gallotannins Trigalloylglucose Tetragalloylglucose Pentagalloylglucose Ellagitannins Monomer Tellimagrandin I Casuarictin Geraniin Dimer Cornusiin A Oenothein B Rugosin D Coriariin A Nobotanin B Trimer Nobotanin E Tetramer Nobotanin K

a

HHDP b

Valoneoyl ~

(1) (2) (3)

31.8 _+2.8 d 24.2 _+ 1.9 18.9 + 1.1

3 4 5

0 0 0

0 0 0

(4) (5) (6)

11.9 +0.5 11.7 + 0.6 15.5 _+0.6

2 0

1 2

0 0

1

1

0

(7) (8) (9)

7.1 _+0.3 4.8 +_0.4 6.1 _+0.4

(10)

8.5 _+0.5

(11)

4.4 _+0.3

3 2 5 6 3

1 0 1 2 2

1 2 ! 0 1

(12)

1.8 _+0.2

5

2

2

(13)

0.38 _+0.03

6

3

3

Condensed tannins ECG-dimer ECG-timer ECG-tetramer

> 100 >100 > 100

2 3 4

0 0 0

0 0 0

Related compounds ( - )-Epicatechin ( - )-Epicatechin gallate ( - )-Epigallocatechin ( - )-Epigallocatechin gallate

> > > >

0 0 1 0 1

0 0 0 0 0

0 0 0 0 0

OH - C O - ~ O H OH I

I

CO b H

O HO

-

~

CO ~

OH HO I

CO

I

CO

OH OH I

OC

HO OH HO OH HO OH a Mean _+S.D. for duplicate assays (n = 3).

100 100 100 100

254 from Whatman. Red-Sepharose was from Pharmacia. Butyl-Toyopearl 650M, heparin-5PW and G2000SW were from Tosoh, Japan. Preparation of [3H]poly(ADP-ribose). [3H]Poly (ADP-ribose) was prepared from [3H]NAD using the chromatin system of HeLa $3 cells as described previously [8-10,14,15]. The poly(ADP-ribose) obtained had an average chain length (n) of 15. The specific radioactivity of the polymer was calculated to be (2.5-3.0) • 106 cpm//xmol ADP-ribose residues. Assay of poly(ADP-ribose) glycohydrolase. Poly(ADPribose) glycohydrolase activity was assayed by measuring the decrease in radioactivity of [3H]poly(ADPribose), which was adsorbed to DE-81 paper as described previously [8-10,14,15]. The reaction mixture contained 50 mM potassium phosphate (pH 7.2), 10 mM 2-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, 100 t x g / m l of bovine serum albumin, 20 p~M [3H]poly(ADP-ribose) (n = 15) and an appropriate amount of enzyme in a total volume of 30 /xl. The reaction was carried out at 37°C. One unit of enzyme activity is defined as the amount liberating 1 #mol of ADP-ribose from poly(ADP-ribose) per min under these conditions.

Purification of poly(ADP-ribose)

glycohydrolase.

Poly(ADP-ribose) glycohydrolase was purified from the nuclear fraction of human placenta by sequential chromatographies on butyl-Toyopearl 650M, Red-Sepharose, single stranded DNA-cellulose, heparin-5PW and G2000SW [15]. The specific activity of the final preparation of poly(ADP-ribose) glycohydrolase was 57 t z m o l / m i n per mg protein, showing a 55 000-fold increase over that in the cell homogenate [15]. Results

Effects of chemically defined tannins on poly(ADPribose) glycohydrolase activity The chemically defined tannins tested (Fig. 1) can be classified according to their structures into gallotannins, ellagitannins and condensed tannins (Table I). The inhibitory activity of each tannin on poly(ADPribose) glycohydrolase purified from human placenta was evaluated by adding various concentrations to the standard assay mixture containing 20 ~ M [3H]poly (ADP-ribose). The enzyme reaction was performed at 37°C for 10 min to maintain the linearity of the reaction and hydrolyze less than 30% the radioactive substrate. Fig. 2 shows representative titration curves of the inhibition by chemically defined tannins of poly(ADPribose) glycohydrolase activity. The inhibition curves of ellagitannins (Fig. 2 b - d ) were shifted to lower concentrations than those of gallotannins (Fig. 2a). Condensed tannins (ECG-dimer, ECG-trimer and ECG-tetrainer) had little effect on poly(ADP-ribose) glycohy-

100

a

I

I

I

I

I

I

I

I

I

50

0

100

¢f I

¢

b

0 o

~

50

u,I

'O

0

//

~oo

12

.Q L.

d

5C

121 < o

o_

0 I00

d

0

4/

0

I

L

i

i

0.1

1

10

100

Concentration (/JM) Fig. 2. Effects of chemically defined tannins on poly(ADP-ribose) glycohydrolaseactivity.The activityof poly(ADP-ribose)glycohydrolase was determined as described in Materials and Methods. The reaction mixture contained various concentrations of ECG (~), ECG-tetramer (•), tri- (o) and penta-(e) galloylglucose (a); the monomeric ellagitannins tellimagrandin I (e) casuarictin (o) and geraniin ( • ) (b); the dimeric ellagitannins nobotanin B (e), oenothein B (o), rugosin D ( • ) and cornusiin A (zx) (c); or the trimeric and tetrameric ellagitannins nobotanin E (o) and K (e) (d). drolase activity, even at 100 /xM (Fig. 2a). ( - ) - E p i catechin, ECG, (-)-epigallocatechin and ( - ) - e p i g a l locatechin gallate, which are important components and structural units of condensed tannins, had no appreciable inhibitory effect on the enzyme activity (Table I). Thus, only hydrolyzable tannins (gallotannins and ellagitannins) inhibited poly(ADP-ribose) glycohydrolase activity.

Inhibition of poly(ADP-ribose) glycohydrolase activity by hydrolyzable tannins The ICs0 values of hydrolyzable tannins determined from these titration curves are listed in Table I. Gal-

255

30

3

20

s

10

OO

3 Number

5

4

of galloyl residue

m gallotannin Fig. 3. Plot of IC,, values versus number of galloyl residues of gallotannins. The numbers in parentheses correspond to those shown in Table I.

lotannins showed lower inhibitory activities (IC,, = 18.9-31.8 PM) than ellagitannins (IC,, = 0.38-15.5 PM). The degree of inhibitory activity of gallotannins increased with their molecular weight or number of galloyl residues (tri- < tetra- < penta-galloyl). Penta-galloylglucose, in which the glucose core is saturated with galloyl groups, was the most active. Oligomeric ellagitannins were more inhibitory than monomeric ellagitannins. The most effective inhibition was the tetrameric ellagitannin nobotanin K, with a IC,, value of 0.38 PM, which was about one order of magnitude lower than that of the dimeric ellagitannin nobotanin B (IC,, = 4.4 PM). The inhibitory activity of oligomeric ellagitannins increased with the number of monomeric residues (dimeric < trimeric < tetrameric), then IC,,, values ranging from 0.38 to 8.5 PM.

their IC,, values against their number of functional groups (galloyl, HHDP and valoneoyl) (Figs. 3-5). As shown in Fig. 3, the plot of IC,, values of gallotannins versus their number of galloyl residues was linear. In contrast, the plot of IC,,, values of ellagitannins versus their number of monomeric units was hyperbolic (Fig. 4). The plot of IC,, values of ellagitannins versus their number of HHDP groups (Fig. 5a) or galloyl residues (not shown) did not show a clear relationship. A curve was obtained when the IC5,, values were plotted versus the number of valoneoyl groups, but the slope was not noticeably hyperbolic (Fig. 5b). Interestingly, a hyperbolic curve similar to that in Fig. 4 was obtained when the IC,, values were plotted versus the sum of the numbers of HHDP and valoneoyl groups (Fig. 5~).

I

a

1

O(6)

15 -

. (5)

l (41

10 -

l C10) O(7) l (9)

5 l(a)

l (n) .(12)

l(13)

I 2

1

OO

3

Number

of HHDP

group

Number

of valoneoyl

Structure-activity relationship of hydrolyzable tannin-induced inhibition of poly(ADP-ribose) glycohydrolase

To study the structure-activity relationship of the inhibitory activities of hydrolyzable tannins, we plotted

I

I

I

0

(

I

(6)

group

(4) (5)

l(10) O(7) \

(9) (8). (1 1). \,

l(12) (

I

1 Number

2

4

3

of monomeric

unit

in ellagitannin

Fig. 4. Plot of IC,, values versus number of monomeric units of ellagitannins. The numbers in parentheses correspond to those shown in Table I.

“0 Number

1

2 of HHDP

3

4

5

6

plus Valoneoyl group

Fig. 5. Plot of IC,, values versus number of HHDP and valoneoyl groups. HHDP groups (a), valoneoyl groups (b) or HHDP plus valoneoyl groups (cl. The numbers in parentheses correspond to those shown in Table I.

256 Discussion

References

The present study provides information about novel inhibitors of poly(ADP-ribose) glycohydrolase in vitro. Of three classes of chemically defined tannins tested, only hydrolyzable tannins inhibited poly(ADP-ribose) glycohydrolase. Since condensed tannins and tannin-related compounds, which are structural units of tannins, were not appreciably inhibitory, conjugation of the functional galloyl, HHDP and valoneoyl groups with sugar (glucose) is necessary for the inhibitory activity. In particular, oligomeric ellagitannins were more potent inhibitors than the previously known inhibitors, ADP-ribose (IC50= 1-3 raM), cAMP (IC50=5-10 raM), Ap4A (ICs0 = 30 raM) [7-15] and DNA intercalators, such as ethacridine, daunomycin and ethidium bromide (ICs0 = 50-100/~M) [16]. The inhibitory activity of gallotannins was almost completely dependent on the number of their galloyl residues, the plot of their ICs0 values against their number of galloyl groups being linear. Ellagitannins were more inhibitory than gallotannins, suggesting that the inhibitory effects of HHDP and valoneoyl groups, which are unique components of ellagitannins, may be greater than that of the galloyl group. Moreover, the inhibitory effects of oligomeric ellagitannins were greater than those of monomeric ellagitannins. These results indicate that a specific higher-order conformation of HHDP and valoneoyl groups in oligomeric ellagitannins may be important in tight interaction with the glycohydrolase molecule. It is noteworthy that the slope of a plot of the ICs0 values of oligomeric ellagitannins against their number of monomeric units was hyperbolic, suggesting cooperative effects of HHDP and valoneoyl groups. This idea is supported by the fact that a plot of the ICs0 values against the number of HHDP plus valoneoyl groups was also hyperbolic. A plot of ICs0 values against the number of valoneoyl groups was slightly hyperbolic, but a similar plot against HHDP group was not, so the contribution of valoneoyl group to the inhibitory activity may be greater than that of HHDP group. These results suggest that the stereochemical compositions of HHDP and valoneoyl groups in oligomeric ellagitannins may be important for inhibition of poly(ADPribose) glycohydrolase. The exact mechanism of this inhibitory effect is, however, not known at the present time and remains to be determined.

1 Mandel, P., Okazaki, H. and Nidergang, C. (1982) Proc. Nucleic Acid Res. Mol. Biol. 27, 1-51. 2 Sugimura, T. and Miwa, M. (1983) Carcinogenesis 4, 1503-1506. 3 Ueda, K. and Hayaishi, O. (1985) Annu. Rev. Biochem. 54, 73-100. 4 Tanuma, S., Johnson, L.D. and Johnson, G.S. (1983) J. Biol. Chem. 258, 15371-15375. 5 Tanuma, S., Kawashima, K. and Endo, H. (1986) J. Biochem. 99, 915-922. 6 Althaus, F.R., Lawrence, S.D., He, Y.-Z., Sattler, G.L., Tsukada, Y. and Pitot, H.C. (1982) Nature 300, 366-368. 7 Miwa, M., Tanaka, M., Matsushima, T. and Sugimura, T. (1974) J. Biol. Chem. 249, 3475-3482. 8 Tanuma, S., Kawashima, K. and Endo, H. (1986) J. Biol. Chem. 261,965-969. 9 Tanuma, S., Kawashima, T. and Endo, H. (1986) Biochem. Biophys. Res. Commun. 135, 979-986. 10 Tanuma, S. and Endo, H. (1990) Eur. J. Biochem. 191, 57-63. 11 Tavassoli, M., Tavassoli, M.H. and Shall, S. (1983) Eur. J. Biochem. 135,449-455. 12 Hatakeyama. K., Nemoto, Y., Ueda, K. and Hayaishi, O. (1986)J. Biol. Chem. 261, 14902-14911. 13 Thomassin, H., Jacobson, M., Guay, J., Verreault, A., Aboul-ela, N., Menard, L. and Poirier, G.G. (1990) Nucleic Acids Res. 18, 4691-4694. 14 Maruta, H., Inageda, K., Aoki. T., Nishina, H. and Tanuma, S. (1991) Biochemistry 30, 5907-5912. 15 Uchida, K., Suzuki, H., Maruta, H., Abe, H., Aoki, K., Miwa, M. and Tanuma, S. (1993) J. Biol. Chem. 268, 3194-3200. 16 Tavassoli, M., Tavassoli, M.H. and Shall, S. (1985) Biochim. Biophys. Acta 827, 228-234. 17 Tanuma, S., Sakagami, H. and Endo, H. (1989) Biochem. Int. 18, 701-708. 18 Tanuma, S., Tsai, Y.-J., Sakagami, H., Konno, K. and Endo, H. (1989) Biochem. Int. 19, 1395-1402. 19 Tsai, Y.-J., Aoki, T., Maruta, H., Abe, H., Sakagami, H., Hatano, T., Okuda, T. and Tanuma, S. (1992) J. Biol. Chem. 267, 14436 14442. 20 Hatano, T., Edamatsu, R., Hiramatsu, R., Mori, A., Fujita, Y., Yasuhara, T., Yoshida, T. and Okuda, T. (1989) Chem. Pharm. Bull. 37, 2016-2021. 21 Okuda, T., Yoshida, T., Asida, M. and Yazaki, K. (1983) J. Chem. Soc. Perkin. Trans. 1, 1765-1772. 22 Hatano, T., Ogawa, N., Kira, R., Yasuhara, T. and Okuda, T. (1989) Chem. Pharm. Bull. 37, 2083-2090. 23 Hatano, T., Yasuhara, T., Matsuda, M., Yazaki, K., Yoshida, T. and Okuda, T. (1990) J. Chem. Soc. Perkin. Trans. 1, 2735-2743. 24 Hatano, T., Hattori, S. and Okuda, T. (1986) Chem. Pharm. Bull. 34, 4092-4097. 25 Yoshida, T., Haba, K., Shingu, T. and Okuda, T. (1987) Heterocycles 26, 2845-2848. 26 Yoshida, T., Haba, K., Arata, R., Okuda, T. and Shingu, T. (1987) Natural Products 29, 676-683. 27 Yoshida, T., Hatano, T. and Okuda, T. (1989) J. Chromatogr. 467, 139-147.

Acknowledgement This work was supported in part by a Grant-in-Aid from the Ministry of Education, Science and Culture of Japan.