Structure and Chemistry (Plart D)

studies in Natural Products Chemistry Volume 17 Structure and Chemistry (Part D) Studies in Natural Products Chemistr...

Author: Atta-ur-Rahman

50 downloads 1224 Views 27MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

studies in Natural Products Chemistry Volume 17 Structure and Chemistry (Part D)

Studies in Natural Products Chemistry edited by Atta-ur-Rahman

Vol. 1 Vol. 2 Vol. 3 Vol. 4 Vol. 5 Vol. 6 Vol. 7 Vol. 8 Vol. 9 Vol.10 Vol.11 Vol.12 Vol.13 Vol.14 Vol.15 Vol.16 Vol.17

Stereoselective Synthesis (Part A) Structure Elucidation (Part A) Stereoselective Synthesis (Part B) Stereoselective Synthesis (Part C) Structure Elucidation (Part B) Stereoselective Synthesis (Part D) Structure and Chemistry (Part A) Stereoselective Synthesis (Part E) Structure and Chemistry (Part B) Stereoselective Synthesis (Part F) Stereoselective Synthesis (Part G) Stereoselective Synthesis (Part H) Bioactive Natural Products (Part A) Stereoselective Synthesis (Part I) Structure and Chemistry (Part C) Stereoselective Synthesis (Part J) Structure and Chemistry (Part D)

studies in Natural Products Chemistry Volume 17 structure and Chemistry (Plart D)

Edited by

Atta-ur-Rahman H.EJ. Research Institute of Chemistry, University of Karachi, Karachi 75270, Paicistan

1995 ELSEVIER Amsterdam - Lausanne - New York - Oxford - Shannon - Tokyo

ELSEVIER SCIENCE B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam, The Netherlands

ISBN: 0-444-82265-8 © 1995 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science B.V., Copyright & Permissions Department, P.O. Box 521,1000 AM Amsterdam, The Netherlands. Special regulations for readers in the USA - This publication has been registered with the Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the USA. All other copyright questions, including photocopying outside of the USA, should be referred to the copyright owner, Elsevier Science B.V., unless otherwise specified. No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. This book is printed on acid-free paper. Transferred to digital printing 2005

FOREWORD

The rapid advances in chromatographic procedures, spectroscopic techniques and pharmacological assay methods have resulted in an increasing number of new and interesting natural products being discovered from terrestrial and marine sources. The present volume contains comprehensive reviews on some of the major advances in this field which have taken place in recent years. The reviews include those on novel metabolites from marine gastropods, the chemistry of marine natural products of the halenaquinol family, secondary metabolites from Echinoderms and Bryozoans, triterpenoids and aromatic compounds from medicinal plants, chemistry and activity of sesquiterpenes from the genus Lactarius, the chemistry of bile alcohols, antifungal sesquiterpene dialdehydes, annonaceous acetogenins, nargenicin macrolides, lignans and diarylheptanoids. Tropane alkaloids and phenolies formed by root cultures are also reviewed. Articles on natural Diels-Alder type adducts, the use of computer aided overlay for modelling the substrate binding domain of HLADH, applications of O NMR spectroscopy to natural product chemistry and the use of biological raw materials in synthesis should also be of interest. It is hoped that the present volume will continue to meet the standards set by the earlier ones of this series and provide much material of interest to a large number of natural product chemists. I wish to express my thanks to Dr. M. Saleh Ajaz and Mr. Athar Ata for their assistance in the preparation of the index. I am also grateful to Mr. Wasim Ahmad, Mr. Asif Khan and Mr. Shabbir Ahmad for the typing work and Mr. Mahmood Alam for secretarial assistance.

December 1994

Atta-ur-Rahman Editor

This Page Intentionally Left Blank

Vll

PREFACE Since days immemorial natural products have had a profound impact on humankind. They were our earliest sources of drugs, derived from traditional herbal medicines. They reaped havoc on man in the form of toxins that would kill or maim people, either in natural disasters, like outbreaks of red tide or ergotism, or in incidences inflicted by man, as the executions in old Greece or the poisonings of adversaries that were often a means of settling power struggles throughout history. And they enriched human life in the form of spices and fragrances. Last not least, they have led to the development of the science of organic chemistry, which started out as the chemistry of natural products. Stimulated by important advances in the biological sciences, particularly in the molecular biology of diseases and in the new field of ecology, the last two decades have seen a tremendous renaissance in the field of natural products. We are now accutely aware of the value of the chemical diversity represented by natural products as a source of new leads for bioactive drugs and of the utility of bioactive natural products as tools in dissecting and analyzing life processes at the molecular level. And we are developing an ever keener sense of the importance of natural products in governing the complex relationships of living organisms in our ecosystems. Concomitantly our view of the role of natural products has changed drastically over the years. While at one time they were considered mere waste products of a luxuriating metabolism, the view now prevails that the synthesis of such compounds represents an evolutionary advantage to the producing organism. With the renewed broad interest in natural products it is most appropriate that a continuing series of publications is dedicated to the topic of natural products chemistry. Professor Atta-Ur-Rahman with his worldwide connections to all the leading natural products chemists of our time is the ideal person for the task of editing this series. He has brought this series to life and has done an outstanding job of sustaining it. The present volume again presents an eclectic mix of articles on many different topics ranging from marine natural products, microbial and plant metabolites all the way to topics like molecular modeling, l^O-NMR spectroscopy or the role of biological raw materials in synthesis. I hope its readers will enjoy this volume as much as I did, and I wish it the same success that its predecessors have enjoyed. Heinz G. Floss University of Washington Seatde, Washington

This Page Intentionally Left Blank

IX

CONTENTS Foreword Preface Contributors

v vii ix

Novel secondary metabolites of marine gastropods M. ALAM AND K.L. EULER

3

Total synthesis and absolute stereochemistry of novel biologically active marine natural products of Halenaquinol family: Theoretical studies of CD spectra NOBUYUKI HARADA AND TATSUO SUGIOKA

33

Bryozoan secondary metabolites and their chemical ecology A.J. BLACKMAN AND .T.T. WALLS

73

Structure and biological activity of triteipenoids and aromatic compounds from medicinal plants R. AQUINO, F. DE SIMONE, N. DE TOMMASI AND C. PIZZA

113

Sesquiteipenes and other secondary metabolites of genus Lcictarius (Basidiomycetes): Chemistry and biological activity G. VIDARI AND P. VlTA-FINZl

153

Stmcture and biosynthesis of bile alcohols: Disorders of choylesterol side-chain oxidation in cerebrotendinous xanthomatosis BISHAMBAR DAYAI., GERALD SALEN AND SARAH SHEFER

207

Antifungal sesquiteipene dialdehydes from the Warhtir^ia plants and their synergists ISAO KUBO

233

Detenuination of relative and absolute configuration in the Annonaceous acetogenins ELIZABETH A. RAMIREZ AND THOMAS R. HO YE

251

The chemistry of the nargenicin macrolides .lAMES KALLMERTEN

283

Some aspects of the chemistry of lignans R.STEVENSON

311

The chemistiy of natural diarylheptanoids G.M. KESERU AND M. NOGRADl

357

Tropane alkaloids in root cuUures of Solanaccous plains M. SAUERWEIN, K. ISHIMARU, K. YOSHIMATSU AND K. SHIMOMURA

395

Phenolics in root cultures of medicinal plants K. ISHIMARU AND K. SHIMOMURA

421

Chemistry and biosynthesis of natural Diels-Alder type adducts from Moraceous plants TARO NOMURA, YOSHIO HANO AND SHINICHI UEDA

451

Modelling the substrate binding domain of horse liver alcohol dehydrogenase, HLADH, by computer aided substrate overlay MAIJA AKSELA AND A.C. OEHLSCHLAGER

479

Applications of '^O NMR spectroscopy to natural products chemistry DAVID W.BOYKIN

549

The role of biological raw materials in synthesis JOHN H.P. TYMAN

601

Subject Index

655

XI

CONTRIBUTORS

Maija Aksela

Depaitmcnt ol' Chemistry, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6

Maktoob Alam

Department of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, University of Houston,, Houston, Texas 77204-5515, U.S.A.

R. Aquino

Dipaitimento di Chimica delle Sostan/e Naturah, Universita Degli Studi di Napoli 'Tederico H", Via D. Montesano 49, 80131 Napoli, Italy

Adrian J. Blackman

Chemistry Department, University of Tasmania, P.O. Box 252C, Hobart, Tasmania-7001, Australia

David W. Boykin

Department of Chemistry, Georgia Slate University, Atlanta, Georgia 30303, U.S.A.

Bishambar Dayal

Department of Medicine, University of Medicine & Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey 07103-2406, U.S.A.

K.L. Eiiler

Department of Medicinal Chemistry and Pharmacognosy, College of Phannacy, University of Houston,, Houston, Texas 77204-5515, U.S.A.

P. Vita-Finzi

Dipartimento di Chimica Organica, Dell' Universita di Pavia, 27100 Pavia, Italy

Yoshio Hano

Toho University, Faculty of Pharmaceutical Sciences, 2-2-1 Miyama, Funabashi, Chiba 274, Japan

Nobuyiiki Harada

Institute for Chemical Reaction Science, Tohoku University, 2-1-1 Kalahira, Aoba Sendai 980, Japan

Thomas R. Hoye

University of Minnesota, Department of Chemistiy, 207 Pleasant Street, S.E. Minneapolis, MN 55455-0431, U.S.A.

K. Ishimaru

Breeding and Physiology Lab., Tsukuba Medicinal Plant Research Station, National Insliule of Hygienic Sciences, I Hachimandai, Tsukuba, Ibaraki-305, .lapan

James Kallmeiten

Syracuse University, Department of Chemistry, Room 1-041, Center for Science & Technology, Syracuse, New York 13244-4KK), U.S.A.

G.M. Keseru

Research Group for Alkaloid Chemisuy of the Hunganan Academy of Sciences, Technical Univeisily of Budapest, H-1521 Budapest P.O.B. 91, Hungaiy

Isao Kubo

Professor of Natural Products Chemistry, Division of Insect and Microbial Ecology, College of Natural Resources, University of California, Berkeley, CalifomL 94720, U.S.A.

Mihaly Nogradi

Research Group for Alkaloid Chemistry of the Hungarian Academy of Sciences, Technical University of Budapest, H-1521 Budapest P.O.B. 91, Hungary

Taro Nomura

Toho University, Faculty of Pharmaceutical Sciences, 2-2-1 Miyama, Funabashi, Chiba 274, Japan

A.C Ochlschlager

Department of Chemistry, Simon Fraser University, Burnaby, British Columbia, Canada V5A IS6

C. Pizza

Dipartimento di Chimica delle Sostan/e Naturali, Universita Degli Studi di Napoli "Federico 11", Via D. Montesano, 49, 8()I31-Napoli, Italy

Elizabeth A. Ramirez

University of Minnesota, Department of Chemistry, 207 Pleasant Street, S.E. Minneapolis, MN 55455-0431, U.S.A.

Gerald Salen

Department ol' Medicine, University ol' Medicine & Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey 07103-2406, U.S.A.

M. Sauerwein

Breeding and Physiology Lab., Tsukuba Medicinal Plant Research Station, National Instiute of Hygienic Sciences, 1 Hachimandai, Tsukuba, Ibaraki-305, Japan

Sarah Shefer

Department o\' Medicine, University of Medicine & Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey 07103-2406, U.S.A.

Koichiro Shimomura

Head of Breeding and Physiology Lab., Tsukuba Medicinal Plant Research Station, National Instiute of Hygienic Sciences, 1 Hachimandai, Tsukuba, Ibaraki-305, Japan

F. De Simone

Dipartimento di Chimica delle Sostanze Naturali, Universita Degli Studi di Napoli 'Tederico \V\ Via D. Montesano, 49, 80131-Napoli, Italy

R. Stevenson

Department of Chemistry, Brandeis University, P.O. Box 9110, Waltham, MA 02254-9110, U.S.A.

Xlll

Tatsuo Sugioka

Inslitule lor Chemical Reaction Science, Tohoku University, 2-1-1 Katahira, Aoba Sendai 980, Japan

N. De Tommasi

Dipartimento di Chimica delle Soslan/e Naturali, Universita Dcgli Studi di Napoli ^'Federico H",Via D. Monlesano, 49, 8()13l-Napoli, Italy

John H.P. Tyman

Department of Chemistry, Brunei, The University of West London, Uxbridge, Middlesex UB8 3PH, U.K.

Shinichi Ueda

Toho University, Faculty of Pharmaceutical Sciences, 2-2-1 Miyama, Funabashi, Chiba 274, Japan

Giovanni Vidaii

Dipartimento di Chimica Organica, Dell' Universita di Pavia, 27100 Pavia, Italy

Justin T. Walls

Zoology Department, University of Tasmania, P.O. Box 252C, Hobart, Tasmania-7001, Australia

K. Yoshimatsu

Breeding and Physiology Lab., Tsukuba Medicinal Plant Research Station, National Insliute of Hygienic Sciences, I Hachimandai, Tsukuba, Ibaraki-305, Japan

This Page Intentionally Left Blank

structure and Chemistry

This Page Intentionally Left Blank

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

Novel Secondary Metabolites of Marine Gastropods M . Alam and K.L. Eiiler

Mollusks have attracted the attention of humans since prehistoric times. People historically have associated certain powers with plants and animals that resembled parts of the human anatomy. Mollusks would be a classical example in thatacertain type of power was associated with cowry (anatomical resemblance to female genitalia) and was thought to be transferred to the possessors of cowry. The first written report about mollusks appeared in Aristotle's "History of animals", which contained a detail discussion of Mediterranean mollusks. During the late fourteen and fifteenth centuries the collecting and studying of shells of marine mollusks became hobbies of gentlemen fix)m well-to-do families with interests in natural history. Since the publication of the first book on marine natural products by Professor Scheuer— Chemistry of Marine Natural Products (1) a number of books (2-6) have been published on marine natural products. Similarly, a number of reviews (7-9) dealing with various aspects of the chemistry of marine mollusks have also appeared in the literature. In the present review the authors have attempted to present a summary of the literature dealing with novel compounds from marine mollusks since 1987. For compounds before 1987 the readers are referred to excellent reviews authored by P. Karuso (10) and H. C. Krebs (11). In order to give readers a broader scope of the novel compounds, examples from all three subclasses of the phylum Mollusca—Prosobranchia, Opisthobranchia and Pulmonata have been selected. During the early seventies the occurrence of a variety of compounds from marine mollusks raised serious questions about their origin. Because gastropods are voracious eaters with virtually every type of feeding habit, it was postulated early on that novel compounds from mollusks may have had their origin in the dietary sources of these invertebrates. One of the earliest reports supporting this hypothesis came from the laboratory of Professor Schantz, who showed that saxitoxin [1 ] (one of a group of neurotoxins commonly known as paralytic shellfish poisons), which was originally isolated from the mollusk Saxidomus gigantius, was actually produced by H2NOCO^

HaNOCOs H R-N

^ ^ ^ OH "" 1

' O" OS03H 2

thedinoflagelbitsGonyaulaxcatenella (12). Similarly, saxitoxin derivatives commonly known as

gonyautoxins [2] isolated initially from the clam Mya arenaria were later determined to be secondary metabolites of another species of Gonyaulax— G. tamarensis (=Alexandrium tamarensis) (13). The first report linking a brominated secondary metabolite of the sea hare Aplysia kurodai with its diet, the red alga Laurencia sp„ appeared in 1967 (14). Similarly, earlier work from Professor Moore's laboratory (15) reported the isolation of deromoaplysiatoxin (3) from the blue green alga Lyngbya gracilis, Debromoaplysiatoxin had been isolated previously from the digestive gland of the stSiharQ Stylocheili4S longicauda (16). The presence of 3 in L. gracilis again suggested a direct relationship between the diet and novel metabolites of the sea hare.

Herbivorous marine prosobranchs of the genus Aplysia feed on red, brown, green or bluegreen algae. During the late seventies and early eighties a number of terpcnoidal secondary metabolites were isolated from various species of Aplysia and were assumed to be accumulated by the mollusks from the dietary sources consisting of red algae (17), brown algae (18-22) and green and blue-green alga (23). During the middle eighties research on the secondary metabolites of various species of Aplysia continued to reward researchers with novel compounds. An examination of various species of Aplysia for the presence of aromatic compounds resulted in the isolation of aplysin and related compounds [4-8] (24), which were also found to be present in the red alga of Laiirencia species that was consumed by the sea hare (25, 26). A chemical investigation (27) of

JUC/N^R2

4. 5. 6. 7. 8.

Ri Ri Ri Ri Ri

=r Br, R2 = H = R2 = H = H, R2 = OH = Br, R2 = OH = H, R2= Br

the mid-gut gland of another Aplysia - A. kurodai, collected from Izy-Shimode Beach, Southwest CI

^^4 Br-^^'^^Br 10

CI

J-^^ ^ Jr""! B / ^ > 1 11

Br-^^'^'"'C1 12

Japan, has resulted in the isolation of four isomeric compounds -aplysiapyranoid A - D [9 -12]. The absolute configuration of aplysiapyranoid B was later established by x-ray crystallography (28) Quite often the type of compound isolated from Aplysia species depends on the location from which they were collected, and therefore, on the algae upon which they feed. A. kurodai collected from Mei Prefecture of Japan was found to contain an extended diterpene (with a prenylated eudismane skeleton)-aplysiadiol [13] and its methyl derivative [14] (29) . It could safely be assumed that a brown alga on which this prosobranch feeds was the actual source of aplysiadiol. Recently a biogenetic type synthesis of 14 has been reported (30).

13. R = H 14. R = CH3

An investigation of the more polar fraction of, presumably, the above mentioned collection of A. kurodai, from the Mei Prefecture of Japan resulted in the isolation of three cytotoxic alkaloids — aplaminone [15], neoaplaminone [16] and neoaplaminone sulfate [17] (31). The biogenetic origin

OCH,

(CH3)3N

(CH3)3N

15

16. R = H

17. R = SO3H

of these alkaloids is still open for discussion. However, they could have been derived from tyrosine ortyramine. An examination of another collection of A. kurodai, presumably from Japan, has resulted in the isolation of aplykurodin A [ 18] and aplykurodin B [19] (32), which could have been derived from a steroidal precursor, which was degraded by the sea hare to produce aplykurodin A. However, its presence in the dietary source of the sea hare (such as a brown alga) can not be ruled out at the present time. A bioassay directedfractionationof A. ^wroda/ collected from Mei Prefecture in Japan has

also resulted in the isolation of a novel compound-aplydilactone [2 0] (33) which is an example example of a cyclopropane-ring-containing fatty acid lactone. Cyclopropaneringcontaining fatty acids are quite rare in nature (34,35). Aplydilactone is reported to increase the activity of phospholipase A2— an enzyme that is responsible for the removal of fatty acids from C2 of phospholipids (33).

A subsequent analysis of specimens of A. kurodai collected in 1985 from the Mei Prefecture in Japan yielded a new labdane-type diterpene, 6/7/-aplysin-20 [21], which is a diastereomer of aplysin-20 [2 2](36), together with the enantiomer of isoconcinndiol [2 3](37).

An investigation of the the chemical defense of Aplysia fasciata collected from the Bay of Naples has resulted in the identification of 4-acetylaplykurodin B[24] and aplykurodinine B [25] (38). Recently a collection and chemical examination oi Aplysia Juliana from the intertidal zone of the Karachi, Pakistan coastline yielded a new diterpene lactone-angasiol acetate [26], the

structure of which was established by x-ray crystallography (39).

CH3OCO

24

25

26

Red algae belonging to the genus Laurencia have been reported to to be a rich source of halogenated sesquiterpenes (40). A species of the sea hare- Aplysia dactylomela, which feeds on Laurencia species, has been reported to concentrate about 25 different chlorochamigrene analogues from its diet. In 1986 Sakaietal (41) reported the isolation and chemical structures of new halogenated chamigrenes[27 - 30] from the digestive glands of A. dactylomela, collected from Hisamatsu Miyako, Okinawa. In addition to halogenated chamigrenes two non-terpenoids— brominated diphenylether [31] and maneonene [32] were also isolated. It could be speculated that 3 1 and 3 2 are derived from the dietary sources of the sea hare. However, anisoles are more common in marine invertebrates, such as sponges of the genus Dysidea, as compared to marine algae.

y^-

^f=n

^

Another collection of i4. dactylomela collected from Kohoma Island, Okinawa, on the other hand was found to contain (42) cuparene-related sesquiterpenes - cyclolaurene [33], laurinterol [34], cyclolaurenol [35], cyclolaurenol acetate [36], cupalaurenol [37] and cupalaurenol acetate [3 8]. The authors failed to comment on the presence or absence of chamigrenes in these samples.

33

34

35

CH3OCO

CH3OCO.

36

37

38

Specimens of A. dactylomela from the Caribbean Sea (43) have been reported to contain an uncommon sesquiterpene-dactylol [3 9] with an 8.5 fused ring skeleton. The likely source of dactylol has been speculated to be the red alga Laurencia poiti (43).

39

A new compound from a new class of diterpenes-dactylomelol [40] was also isolated from the specimens of A. dactylomela, presumably collected from the Canary Islands (44). Dactylomelol could be envisioned arising from the cyclization of the two internal double bonds of geranyllinalol to form carbocyclic rings.

40

An examination (45) of juvenile A. dactylomela feeding on the brown alga Stypopodium zonale from Vega Baja on the north coast of Puerto Rico resulted in the isolation of epitaondiol [41] and 3-ketoepitaondiol [42] along with stypodione [4 3] . A comparison of the A. dactylomela

extract with an extract of S, zonale confirmed the dietary nature of the metabolites of juvenile A. dactylomela.

42

41

43

In order to confirm the biotransformation capability of the sea hare -Aplysia punctata^ Quinoa et al. (46) studied a number of marine algae for the presence of halogenated monoterpenes and compared the hplc and gc profiles of the extracts of the hepatopancreas of A. punctata with algal extracts. The results of this study showed a direct correlation between the chromatographic and gc-ms profile of the secondary metabolites of Plocamium coccineum and A. punctata. However, no biotransformation capability was noticed. Nudibranchs are quite often brightly colored, shelUess mollusks which differ from sea hares in that they do not have a mantle cavity and in their dietary habits by feeding on sponges and corals only. In spite of being brightly colored and shell less nudibranchs have few predators. Many nudibranchs employ a chemical defense to deter predators. While a few species are known to synthesize secondary metabolites (47,48), the majority concentrate secondary metabolites from their dietary sources and use them for their defense. Quite often these nudibranchs are capable of concentrating the most repugnant of the minor metabolites of a sponge or coral, which uses these compounds for its own defense against predators. An investigation of the nudibranch Chromodoris funarea from Palau (49), has resulted in the identification of the bromophenylether [4 4], 0-methylfurodysinin [4 5] and hydroperoxide [4 6]

-OOH

10 However, when the animals were stored in acetone for four weeks, the acetone extract gave furodysinin [47], furodysin [4 8], furodysinin lactone [4 9] and epoxylactone [5 0]. Isolation of the artifacts 47 - 50 illustrates the effect of storage and solvent (acetone) on chemical structures of nudibranch*s metabolites.

48

47

49

50

A comparison of the secondary metabolites of C. funerea collected from different locations of Kaibakku Lake in Kaibakku Island, Palau (50), has resulted in the identification of 12-epi-sca\aiin [51], deoxoscalarin [52], luffariellin C [5 3], lufariellin D [54] and ketodeoxoscalarin [55]. 12-E/7/-scalarin [51] and 52 are metabolites of the sponge Spongia officialis and S. nitens respectively (51,52), while 53 and 5 4 are possibly derived from the sponge Liffariella variabilis, which has been reported (53) to contain luffariellin A [56] and luffariellin B [57]. The metabolite profile of C, funerea collected from Kaibakku Lake differs sharply from the profile of C. funerea collected from Iwayama Bay. An examination of the environment revealed that sponges of the genus Dysidea were absent in Kaibakku Lake, possibly

CX:OCH3

53 . R = H 56 . R = OH

H3COCO " ^ \

54. R = H 57. R = OH

11

H3COCO " ^ V _ o

55. X = 3- or 1-kelo because of the vegetation which heavily shades the shallow side of the lake. The resulting shadow quite possibly prevents Dysidea from growing, since Dysidea depends upon either epiphytic or symbiotic cyanobacteria to provide important metabolites required for their growth. These microorganisms in turn require sunlight for their own growth. Chemical studies on the dorid nudibranch Chromodoris macfarlandU collected from Scripp's Canyon, La Jolla, resulted in the isolation of two aromatic norditepenes-macfarlandin A [58] and macfarlandin B [59](54). Both 58 and 5 9 are closely related to aplysulphurin [60] which is a metabolite of the sponge Aplysilla sulphurea (55). A further investigation (56) of the more polar

58

59

60

fractions of the extract resulted in the isolation of three new rearranged diterpcne acetates, which were identified as macfarlandin C [61], macfarlandin D [6 2] and macfarlandin E [6 3]. A comparison of the concentrations of 58 - 63 in seven individual animals led the authors to conclude that that C. macfarlandi must feed on two different aplysillid sponges, one of which contains 6 1 - 6 3 .

OCOCH:.

OCOCH, CH,OCQ

63

12 Another species of Chromodoris -C. norrisi, collected from a mangrove lagoon on the Island of San Jose in the Gulf of California, has been reported to contain norrisolide [6 4] and macfarlandin E [63] along with rearranged diterpene polyrhaphin A [65] and shahamin C [66](57). A comparison of the concentrations of norrisolide, macfarlandin E, polyrhaphin A and shahamin C from C. norrisi with the concentration of rearranged diterpenes from the sponge Aplysilla polyrhaphis collected from the same location confirmed the dietary origin of these metabolites. The data also supported the assertion that this nudibranch does not preferentially concentrate any specific metabolite. Of the nine diterpenes produced by Aplysilla polyrhaphis, only four, 63 - 66, are retained by the nudibranch; this suggests that these metabolites are of importance to the

OCCX:H3

Hjccxro^

H3COCO

^O

H3COCO—.„

X

.0

nudibranch. Indeed, shahmin C was found to deter predation (feeding) by rainbow wrasse (Thalassoma lucaslu-num ) at a 100 mcg/mg food level. A study of the chemistry of the defense allomones present in C. luteorosea from the Mediterranean Sea (58) resulted in the identification of luteorosin [67], 12-^p/-aplysillin [68], 12epi -12-deacetoxyaplysillin [6 9] and macfarlandin A [58]. Macfarlandin A is a known metabolite of C. macfarlandi (54). Based on a report (59) that C luteorosea feeds on sponges of Spongionella species, the authors speculated that the real source of 67 - 69 is an encrusting species of Spongionella.

OR

67

OCOCH3

68. R = COCH3 69. R = H

The importance of dietary compounds in the determination of the metabolic profile of Chromodoris species is illustrated by the presence of a rearranged diterpene-- chromodorolide A [7 0] from the Indian Ocean's nudibranch Chromodoris cavae (60). A subsequent study of a second

13 collection (61) resulted in the isolation of 7 0 along with chromodorolide B [71]. The biosynthesis of chromodorolides A and B could be envisioned to proceed via the formation of the noirisane skeleton [72], {e.g., the diterpene norrisolide [73](62)) followed by the formation of a new carbon-carbon bond (CI2 -C17). Cyclyzation of the C l l carboxyl group with the bisacetal-oxalane ring would then give the appropriate heterocycles.

CH3OCO..,

CH30C0-«n<

I

JO

H 6COCH3

There are many examples which unequivocally support the theory that most of the defense chemicals of nudibranchs have had their origin in their dietary sources. In one study (63) the dorid nudibranch Chromodoris lachii was found to contain latrunculin A [7 4] and dandrolasin [75] in a ratio that supported the dietary origin of these metabolites even though the sponge on which these nudibranch feeds was not present at that location. Both latrunculin A and dandrolasin have previously

// \^ I OH

HN

^

75

O 74

14 been isolated from the marine sponge Latrunculin magnifica (64) and Oligocerus hemorrhages (65) respectively The nudibranch Cadlina luteomarginata is relatively abundant on the Pacific coast of the United States from Baja California in the south to Punta Eugenia, British Columbia, in the north. It is regarded a close relative of a tropical nudibranch of the genus Chromodoris. It has been suggested that similar to Chromodoris species, the skin chemistry of Cadlina variesfromone collection to the another, thus reflecting differences in the sponge fauna at various locations. During the late seventies and early eighties Faulkner's group conducted a detailed study of the gut contents of the nudibranch Cadlina luteomarginata (66) and sponges found in the environment from which the nudibranch were collected. An examination of the gut contents of the nudibranch C. luteomarginata for the presence of remnants of ten different sponges showed that about 95 percent consumed only one sponge species. The authors failed to speculate on the presence of spicules in the gut content form a sponge of Axinella sp. which had been neither encountered in the habitat nor previously recorded from Southern California. A study (66) of the chemical compositions of an Axinella species and of C. luteomarginata revealed a similarity in the relative concentrations of three compounds, two isonitriles of unknown structures and isonitrile 7 6 in the two organisms (66). However isothiocyanate 77 and two unknown isothiocyanates were present in different concentrations in the sponge and nudibranch studied.

76. R = NC 77. R = NCS

Specimens of C. luteomarginata collected from Sanford Island, British Columbia, were found (67) to contain a rearranged and degraded diterpene-glaciolide [7 8], which was also present as the major component of an extract of the sponge Aplysilla glacialis collected from the same location.

78

In another study C. luteomarginata, collected from two locations (Howe Sound and Barkley Sound) in British Columbia (68), was found to contain albicanyl acetate [79], albicanol [80],

15 furodysin [4 8], and furodysinin [4 7]. One of the collected nudibranchs also had microcionin 2 [81]. Recendy a study of the terpenoids of the sponge Apfysilla glacialis and the nudibranch C. luteomarginata (found grazing on the sponge) has resulted (69) in the identification of glaciolide [78], cadlinolide A [82] and tetrahydroaplysulphurin 1 [83]. That a number of the sponge metabolites were not detected in the nudibranch suggests that a selective sequestering process is utilized by the nudibranch.

^OCOCH

Another dorid nudibranch, Doris verrucosa has been reported (70) to contain an analogue of methylthioadenosine [84] along with diterpenoic acid glycerides, vemicosin A [85] and verrucosin B [86]. The carbon skeleton of the vemicosins could be derived from isocopalane diterpenes which formerly were considered to be the biogenetic precursors of the tetracyclic spongiane diterpenes.

OR|

O^^A./*OR2

85. Ri = H, R2 = COCH3 86. Ri = COCH3. R2 = H

Nudibranchs of the family Phyllidiidae often contain isonitriles which are derived from their dietary sources (sponges) (ref. P. Karuso's review). The Hawaiian nudibranch Phyllidia varicosa and its sponge prey Cioclapta sp. were found to contain 9-siocyanopupukeanane [87] (71,72).

16 Similarly, the nudibranch Phyllidia bourguini, collected from Hachijo-Jima Island, Japan, was found to contain (73) both 9-isocyanopupukeanane [8 7] and its C9 epimer [8 8].

^

:

87. Ri =NC. R2 = H 88. Ri = H, R2 = NC

Is it possible for the nudibranchs to transfer secondary metabolites, which they have sequestered from their dietary sources, to their eggmasses, in order to endow the eggmasses with some kind of protection from predators or marine microorganisms? To answer this question eggmasses of the nudibranch Hexabranchus sanguineus were analyzed (74) for the presence of novel compounds. This investigation resulted in the isolation of antifungal macrolides ulapualide A [8 9] and ulapualide B [90]. At the same time a related antifungal macrolide kabiramide C [91] was also isolated from an unidentified nudibranch*s eggmasses (75).

OHC

OHC^

0CH3

In 1988 the Faulkner group reported the results of a comprehensive study of the macrocyclic metabolites of the nudibranch Hexabranchus sanguineus and sponges of the genus Halichondria . Specimens of the nudibranch collected from subtidal reefs at Kwajalein Atoll, Marshal Islands, were found to contain (76) dihydrohalichondramide [92], tetrahydrohalichondramide [93] and kabiramide C [91]. Both 92 and 9 3 were also found in the eggmasses of the nudibranch (77). It is noteworthy

17 that kabiramide C, along with other halichondramides, has been found in the sponges of the genus Halichondria (76), and thus halichondramides isolated from the nudibranch //. sanduineus could have been derived from sponges of Halichondria species. Another novel antitumor macrolide-

OHC^

sphinxolide [9 4] has been isolated from an unidentified Hawaiian nudibranch (78).

OHC^

The nudibranch Notodoris citrina, which feeds on the yellow sponge Leucetta chagosensis was found to concentrate substituted 2-aminoimidazole alkaloids including naamidine A [9 5],naamine A [9 6], isonaamidine [9 7] and isonaamine A [9 8] (79).

CH3O.

CH30

95

96

CH,0,

CH30

"°tx/:> 97

98

Recently, the yellow nudibranch Notodoris gardineri collected from the Philippines was found to contain dorimidazole A [9 9] and isonaamine A [9 8] (80). A subsequent study of A^. gardineri has resulted in the isolation of four niore imidazole derivatives (81) which were

CH,

"°XX;c^-

99

identified as clathridine [100], zinc clathridine [101], preclathridine A [102] and clathridine B[l 03]. One of the most interesting characteristics of these 2-aminoimidazole alkaloids is the fact that they form complex with metals, However, very few organometalic complexes are known as nutural products from marine sponges (82,83). Zinc clathridine [101] constitute the first organometallic compound which has been isolated from a nudibranch. It could be safely predicted that the N. gardineri is concentrating this and other aminoimidazoles from its sponge diet.

CH3

V - N '

CH3

y-N' Zn'

101

100

CH,

{:Ou^-^o 102

19 Mollusks belonging to the subclass Prosobranchia are characterized by the location of gills, mantle cavity and anus at the anterior of the body of the invertebrate. They respire by the gills and a great majority have a shell in which the head and foot can retract in the presence of a predator. Studies on the chemistry of prosobranch invertebrates have been as common as studies dealing with opisthobranchs. However, since 1984 chemical studies on shelled mollusks have increased in numbers. Before the eighties research dealing with prosobranchs concentrated on toxic compounds present in shellfish. Since the majority of the metabolites that can be isolated from mollusks are of dietary origin, it would be difficult to consider shellfish toxins as secondary metabolites of shellfish only. An investigation of the marine snail Nerrita albicilla, which was collected for the investigation of cytotoxic compounds, resulted in the identification of the known terrestrial antibacterial pigment, fulvoplumierin[104] along with the norisoflavone, albazoin [105] and two isoflavones[106, 107] (84,85).

OCH3 10<>. R = H 107.R=:CH3

The consumption of shellfish (scallops and mussels) harvested during late spring to early summer from the northeastern region of Japan quite often results in what is commonly known as diarrhetic shellfish poisoning. An initial chemical investigation of the toxic mussels resulted in the identification (86) of okadaic acid [108], dinophysistoxin 1 (DTXj) [109] and two toxins of unknown structures. In a later study (87), chemical structures of three new polyether toxins, dinophysistoxin- 3 (DTX3)[110], pectenotoxin-1 [111] and pectenotoxin-2 [112] were reported.

108. Ri = H. R2 = H 109. Ri = H, R2 = CH3 110. R i = CH3CO, R2 = CH3

20 The polycther toxin okadaic acid [108] has previously been reported from the dinoflagellate Prorocentrum lima and sponges of the genus Halichondria (88,89). The origin of [108] and dinophysistoxins in toxic shellfish has been linked to the dinoflagellates Prorocentrum lima and Dinophysis fortii (90). A further investigation of the toxic extracts of the scallop Patinopectin yessoensis, collected from Mutsu Bay, Aomori Prefecture, yielded a new polyether toxin that was identified as pectenotoxin 3 [113](91). Pectenotoxin-3 is an isomer of pectenotoxin 1 and 2.

111. R = CH2OH 112.

R= CH3

113. R= CHO

An extract of the digestive glands of P. yessoensis, collected from the aforementioned location, upon fractionation gave a new polyether toxin, yessotoxin[ll 4]. The structure of which was determined by spectroscopic methods (92) and has the chemical characteristics of dinoflagellate toxins isolated from Gymnodinium breve (93). Recently, toxic mussels collected from Bantry Bay, Ireland, were found to contain dinophysistoxin-2 (DTX2) [115] (94).

NaOjSO 114

21

115

Chemical investigations dealing with the identification of toxic compound(s) responsible for the toxicity of a mussel, Mytilus edulis, from eastern Prince Edward Island, Canada, has resulted in the identification of domdc acid [116] (95). Domoic acid was originally isolated over 30 years ago from the red alga Chondria armata (96). A further investigation of the toxic mussels has resulted (97) in the isolation of domoic acid D [117] and two geometric isomers, isodomoic acid E3 [118] and isodomoic acid E4 [119]. Prosobranch moUusks of the family Lamellariidae are known to be specific predators of colonial ascidians. Although the moUusks are not completely devoid of the shells, their shells are reduced and is completely or in part covered by a fleshy mantle and thus requiring some kind of defense mechanism to ward off predators. In order to explore the defense chemicals of nudi-

CCX)H 117. R

N

H

N H

HOOC.

CH2COOH

"tX"

CH2COOH

R=

118. R =

TroOH

TOOK

116

o^^o CH3O

CH3O

'3 CH3O 120. R = OH 122. R= H

3 CH3O 121. Ri = OCH3, R2 = CH3 123. Ri = H, R2 = H

22 branchs of this family and the relatioship between the mollusks of the family Lamellariidae and their ascidian prey, Faulkner's group invetigated the tunicate Didemnum chartciwn and a moUusk of the genus Lamellari . Both the moUusk and the prey, collected from Koror, Palau, was found to contain aromatic metabolites, lamellarins A - D [ 1 2 0 - 1 2 3 ] (98). Chemical studies dealing with the isolation and structures of novel metabolites of the moUusk Planaxis sulcatus have resulted in the identification of jeunicin [12 4] (99) ll-epi- sinulariolide [125] (100) [12-Epi' sinulariolide has previously been isolated from soft corals and gorgonians (101)], a new cembrane [126] (102) along with a novel epoxy sterol [127](103). In a recent communication Alam er. al. reported (104) the isolation and identification of a novel homoditerpene, planaxool [128] with a modified cembranoid skeleton. The structure determination

124

HO^

=

127

H

128

of 128 was accomplished through the the extensive use of two-dimensional nmr spectroscopy. The mass spectrum of 128 suggested the molecular weight of 348 which was not supported by the chemical shifts of the carbons in the molecules. The presence of a high field carbon (6 102.1) initially suggested the presence of a hemiacetal moiety in 128, which could only be supported by the incorporation of a eunicin nucleusin the molecule. However, the ^H ^H COSY, HDQC spectra suggested fragments A, B and C and the presence of a jeunicin nucleus in 128. The number of oxygenated carbons along with the downfield position of one of the oxygenated carbons suggested the presence of a highly electron withdrawing group in planaxool. The iodine test for the presence of hydroperoxide moiety was positive which when considered in the presence of a band at 3560 c"^ in the ir spectrum, suggested the presence of a hydroperoxide moiety in the molecule. Planaxool was unstable at room tempe- rature and gave a number of degradation products when attempts were made to reduce the hydropoeroxide moiety under mild conditions. A

23

x<^o^v^ 20

final attempt at elemental analysis after hplc purification gave elemental composition of C 65.95, H 8.52, which corresponds to the molecular formula C2iH320^ and supports the assigned structure of planaxool. An examination of the composition of polyaromatic alkaloids of the mollusk Chelynotus semperi and an unidentified tunicate, both collected from Mante channel, Pohnpei, resulted in the identification of shermilamine B [129] and kuanoniamines A-D [130-133] (105). Shermilamine was originally found as a metabolite of a tunicate of the genus Trididemnum (106).

129

131

130

NHCOC2H5

NHCOCH,

132

133

During the last ten years Faulkner's group has extensive investigated marine pulmonate of the genus Sip/ionoria. Siphonares are air-breathing mollusks that resemble limpets with which they co-occur in the intertidal zone. Their diet consists mainly of microalgae. The secondary metabolites that have been isolatedfromSiphonaria species could be divided into three groups: the simple propionates, the a-pyrones and the y-pyrones. During the early eighties a number of

24 investigations dealing with the secondary metabolites of mollusks belonging to the genus Siphonaria have resulted in the identification of: diemenensinA [134] anddiemenensinB[135]

134

135

from the "false limpet" S. diemenensis (107). It is noteworthy that in solution diemenensin B isomerizes to give diemenensin A. Diemenensin A has antibacterial activity and inhibited cell division in fertilized sea urchin eggs. An examination of an acetone extract of S. denriculata, collected from the coast of New South Wales, has resulted in the identification of two isomeric polypropionates denticulatin A [136] and denticulatin B [137](108). No antimicrobial activity was found associated with these denticulatins. However, both showed ichthyotoxicity toward gold fish (108).

A skin extract of the the mollusk Siphonaria pectinata, collected from the sea wall at the entrance of Key Biscayne, Florida, yielded a new propionate derived antibacterial compound which was named pectinatone [138](109). The Chilean pulmonate S. lessoni was found to contain a nor-

138

pectinatone [139] along with both E and Z isomers of furanone 140 and 141(110).

25

140

139

141

A chemical investigation of an acetone extract of four Sinophoria species ~ S, zelandica, S. atra, S. normalis and S. laciniosa, collected from different locations in Australia, has resulted in the characterization of four novel polypropionate metabolites. Siphonarin A [142] and its homologue siphonarin B [143] were isolated from S. zelandica and S, atra, while dihydrosiphonarin A [14 4] and dihydrosiphonarin B [145] were separated from 5. normalis and 5. laciniosa (111). Similarly, S. maura collected from Costa Rica was found to contain two pairs of racemic diastereoisomeric maurapyrone A - D [ 1 4 6 - 1 4 9 ] along witii maurenone [15 0]( 112). Anotiier pulmonate S. baconi collected from Serrento Beach, \^ctoria, Australia was found contain four polypropionate metabolites , baconipyrones A - D [151-154](113). Baconipyrones are unusual among the polypropionate metabolites because they do not have a normal contiguous polypropionate backbone. The presence of siphonarin A in S. baconi led the authors to propose a possible mechanism of conversion of siphonarin A to baconipyrone B and involves an enolate intermediate.

OH 142. R = CH3 143. R = C2H5

146. R = C5Hii 147.

RSI-CBHT

144. R = CH3 145. R = C2H5

148. R = C5Hii 149. R=:i-C3H7

150

26

0

o 151.

R = C2H5

152.

R = CH3

0

0

153.

R = C2H5

154. R = CH3

A chemical study of an acetone extract of the Mexican . maura (collected at Sayulita, Mexico) resulted in the isolation of two polypropionate metabolites, vallartanone A [ 15 5] and vallartanone B [i56](114) which have structures quite different from other polypropionate metabolites of the genus Siphonaria. An unrelated polypropionate, maurenone [150] which was previously isolated from the Mexican S. maura was also present along with vallartanone A and B. It appears that pulmonates of the genus Siphonaria synthesize similar metabolites derived from the condensation

155. R = CH3 156. R = H

of propionyl Co A, irrespective of the geographical locations from which they were were collected. Only one species of Siphonaria— S, maura, is known to produce polypropionate metabolites which differ both in molecular size and structure from other polypropionate metabolites. The chemical defense of the intertidal limpet Collisella limatula collected from southern Califomia has been identified to be a diketone-limatulone [157](115). A proposed biosynthesis of limatulone involves oxidation of two molecules of squalene to give keto aldehydes which then undergo aldol condensation to give 157.

When disturbed the marine limpet Trimusculus reticulatus produces a copious amount of a milky white mucus that repels predatory star fish. A dichloromethane extract of the whole

27 animal and the mucus was found to contain (111) ditcrpencs identified as 6p-isovaleroxylabda-8,13dien-7a,15-diol [158] as the major metabolite while the minor metabolite was identified as 2a,7adiacetoxy-6p-isovaleroxylabda-8,13-dien-15-ol[159]. Both 158 and 159 failed to deter star Hsh in a number of feeding experiments. Marine pulmonate Onchidella henneyi has previously been reported to contain sesquiterpene onchidal [160] (112). In addition to siphonarin A [140], the fijian mollusk Siphonaria normalis has been reported to contain an unusual tricyclic ketal with a trioxaadamantane ring skeleton. The structure of muamvatin 2 [161] was determined by nmr spectroscopy (118). The possible roles of the polypropionate metabolites isolated from Siphonaria species could be associated with its trailfollowing habits along with defense against predators.

OCOCH,

CHaOCO.,,^

CHO

"rr '''"rr

OCOCH,

O

158

160

159

rr^O

OH

161

In conclusion, the natural products isolated from the three subclasses of moUusks are both synthesized by the mollusks and concentrated by them from their dietary sources. There is overwheming evidence that most of the secondary metabolites of opisthobranchs are derived from dietary sources of the invertebrates. These sequestered or concentrated metabolites are then used by the mollusks as chemical defenses against predators. In addition to concentrating organic compounds, a number of opisthobranchs (e.g., Dendrodoris limbata and D, grandiflora ) have been reported (119) to have the capability of biosynthesizing sesquiterpenes, e.g., driman [16 2],which serves as a defensive allomone. Similarly, the sea hare Navanax inermis has been shown to biosynthesize navonones (120). Similar to opisthobranchs, the chemical structures of the secondary metabolites of prosobranchs suggest that they have had their origin in the dietary sources

28

162

of the invertebrates. The isolation of cembranoids such 1 l-ep/-sinuriolide [125](100) and cembrane [126](102) strongly points towards their origin in the dietary sources of Planaxis sulcatus. Unfortunately, at the time of collection of this mollusk no attempts were made to record the fauna or flora present in the surrounding environment. On the other hand, pulmonates belonging to the genus Siphonaria have been identified as invertebrates that have the capability of synthesizing similar polypropionate metabolites, with contiguous carbon skeletons (121), irrespective of the geographical locations from which they have been collected (113). Only one species of Siphonaria-S, baconi is known to produce metabolites which do not contain a contiguous propionate backbone (113). At the present time there is no evidence which supports the dietary origin of pulmonate metabolites. ACKNOWLEDGMENT This work was supported in part by a grant from The University of Houston Coastal Center. REFERENCES 1 2

13

P. J. Scheuer, Chemistry of Marine Natural Products, Academic Press, New York, 1973. H. J. Humm and C. E. Lane ( Eds.), Bioactive Compounds from the Sea, Marcel and Dekker, 1974. J. T. Baker and V. Murphy, Handbook of Marine Science, Compounds from Marine Organisms, Vol. 1, CRC Press Inc., Cleveland, 1976. P. J. Scheuer ( Ed.), Marine Natural Products, Chemical and Biological Perspectives, Vol. I-V, Academic Press, 1978-1983. P. J. Scheuer ( Ed.), Bioorganic Marine Chemistry, Vol. 1, Springer-Verlag, Berlin 1985 and subsequent volumes. D. H. Attaway and O.R. Zaborsky ( Eds.), Marine Biotechnology, Vo. 1. Plenum Press, New York, 1993. D. J. Faulkner and M. T. Ghiselin, Mar. Ecol. Prog. Sc, 1 3 ( 1 9 8 3 ) 295-301. P. J. Scheuer, Ace. Chem. Res., 25 ( 1992) 433 - 439. R. G. Kerr and B. J. Baker, Nat. Prod. Rep., 8 ( 1991 ), 465 - 497. P. Karuso, in: P. J. Scheuer (Ed.) Bioorganic Marine Chemistry, Vol. 1 , SpringerVerlag, New York, 1987, pp. 32 - 59. H. C. Krebs, in: W. Herz, H. Grisebach, G. W. Kirby and Ch. Tamm ( Eds.), Forschritte der Chemie Organischer Naturstoffe, Springer Verlag, New Yorik, 1986, pp. 152 - 363. E. J. Schantz, J. M. Lynch, G. Vayvada, K. Matsumoto and H. Rapoport, Biochemistry, 5 (1966) 1191-1195 A. Prakash, J. Fish, Res. Bd. Canada., 20 ( 1963 ) 1589 - 1592.

14

H. Masuda, Y. Tomie, S. Yamamura and H. Hirata, J. Chem. Soc. Chem. Commun.,

3 4 5 6 7 8 9 10 11 12

29

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

(1967) - 899. J. S. Mynderse, R. E. Moore, M. Kashiwagi and T. R. Norton. Science, 196 (1977) 538 - 540 Y Kato and R J. Scheuer, J. Am. Chem. Soc., 96 (1974 ) 2245 - 2246 K, L. Erickson, in: P. J. Scheuer (Ed.), Marine Natural Products, \bl. V., Academic Pres., New York, 1983, pp. 131-256. A. Amico, G. Oriente, M. Piatelli, C. Tringali, E. Fattorusso, S. Magno and L. Mayo, Tetrahedron, 36 (1980) 1409-1414. A. G. Gonzalez, J. D. Martin, M. Norte, R. Perez, V. Weyler, S. Rafii and J. Claixiy, Tetrahedron lett, 24 (1983) 1075 - 1076. C. Ireland, D. J. Faulkner, J. Finer and J. Clardy, J. Am. Chem. Soc., 98 (1976) 4664 4665. S. L. Midland, R. M. Wing and J.J.Sims, J. Org. Chem., 48 (1983) 1906-1909. H. H. Sun and W. Fenical, Phytochem., 1 8 ( 1 9 7 9 ) 340-341. A. F. Rose, P. J. Scheuer, J . P. Springer and J. Clardy, J. Am. Chem. Soc, 100 ( 1978) 7665-67670. M. Suzuki, K. Kurata and E. Kurosawa, Bull. Chem. Soc. Jpn. (1986) 3982 - 3982. Y Yamamura and Y. Hirata, Tetrahedron, 19 (1963) 1485-1496. J. A. McMillan, I. C. Paul, S. Caccamese and K. L. Rinehart Jr., Tetrahedron Lett., (1976) 4219-4222. T. Kusumi, H. Uchida, Y Inouye, M. Ishitsuka, Y Yamamoto and H. Kakisawa, J. Org. Chem., 52 (1987 ) 4597 - 4600. Y Inouye, H. Uchida, T. Kusumi and H. Kakisawa, J. Chem. Soc. Chem. Commun., (1987) 346-347. M. Ojika, Y Yoshida, M. Okumura, S. leda and K. Yamada, J. Nat. Prod., 53 (1990) 1619-1622. H. Niwa, S. leda, H. Inagaki and K. Yamada, Tetrahedron Lett., 31 (1990) 7157 7158. H. Kigoshi, Y Imamura, K. Yoshikawa and K. Yamada, Tetrahedron Lett., 31 (1990) 4911-4914. T. Miyamoto, R. Higuchi and T. Komori, T. Fujioka and K. Mihashi, Tetrahedron Lett, 27 (1986) 1153-1156. M. Ojika, Y Yoshida, Y Nakayama and K. Yamada, Tetrahedron Lett. 31 (1990) 4907 4910. M. D. Higgs and L. J. Mulheim, Tetrahedron, 37 (1981) 4259 - 4262. H. Niwa, K. Wakamatsu and K. Yamada, Tetrahedron Lett., 30 (1989) 4543 - 4546. S. Yamamura and Y Hirata, Bull. Chem. Soc. Japan, 44 ( 1971 ) 2560 - 2562. M. Ojaka, H. Kiqushi, K. Yoshikawa, Y Kanayama and K. Yamada. Bull. Chem, Soc. Japan, 65 (1992) 2300-2302. A. Spinella, M. Garagnin, A. Crispino, G. Cimino, E. M. Ortea and G. Sodano, J. Nat. Prod. 55 (1992) 989-993. Atta-ur-Rhaman, K. A. Alvi, S. A. Abbas, T. Sultana, N. Shameel, M, I. Choudhary and J. C. Clardy, J. Nat. Prod. 54 ( 1991 ) 886 - 888. J. D. Martin and J. Daria, in: P. J. Scheuer (Ed.), Marine Natural Products, Chemical and Biological Perspectives, Vol.5, Academic Press, New York, 1983, p. 131-257. R. Sakai, T. Higa, C. W. Jefford and G. Bemardinelle, Helv. Chim. Acta , 69 (1985 ) 91 - 105. T Ichiba and T Higa, J. Org. Chem., 51 ( 1 9 8 6 ) 3364-3366. F. J. Schmitz, K. H. HoUenbeak and D. J. Vanderan, Tetrahedron, 34 (1978 ) 2719 2722. D.M. Estrada, J. L. Ravelo, C. R. Perez and J. D. Martin, Tetrahedron Lett, 30 ( 1989 ) 6219 - 6220. W. H. Gerwick and G. Whatley, J. Chem. Ecol. 15 (1989) 677-682. E. Quinoa, L. Castedo and R. Riguera, Comp. Biochem. Phisol. 92B (1989) 99-101. G. Cimino, S. de Rosa, S. De Stefano, R. Morrone and G. Sodano, Tetrahedron, 41 (1985 ) 1093-110 K. Gustafson and R. J. Anderson, Tetrahedron, 41 ( 1985 ) 1101-1108. B. Carte, M. R. Keman, E. B. Barrabee, D. J. Faulkner, G. K. Matsumoto and J. Clardy, J. Org. Chem., 51 (1986) 3528 - 3532. M. R. Keman, E. B. Barrabee and J. D. Faulkner, Comp. Biochem Physiol., 89B ( 1988 ) 275 - 278.

30 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87

G. Cimino, S. de Stefano and L. Minale, Experientia, 29 (1973) 934-936. G. Cimino, S. de Stefano, R. Morrone and G. Sodano, Tetrahedron, 41 (1985 ) 109 3 1100. M. R. Keman, D. J. Faulkner and R. S. Jacobs, J. Org. Chem., 52 ( 1987 ) 3081 3083 T. F. Molinski and D. J. Faulkner, J. Org. Chem., 51 ( 1986) 2601 - 2603. P. Karuso, B. W. Skelton, W. C. Taylor and A. H. White, Aust. J. Chem. 37 ( 1984) 1081 - 1093. T. R Molinski, D. J. Faulkner, H. C-heng, G. D. van Duyne and J. Clardy, J. Org. Chem., 51 ( 1986) 4564 - 4567. S. C. Bobzin and D. J. Faulker. J. Org. Chem., 54 (1989) 3902 - 3907. G. Cimino, A. Crispino and M. Gavagnin and G. Sodano, J. Nat. Prod. 53 ( 1990), 102 106. L. Schmekel and A. Portman, Opisthobranchia de Mittelmeeres, Springer-Verlag, Berlin, 1982, p. 60. E. J. Dumdei, E. D. de Silva, R. J. Andersen, M. I. Cloudhary and J. Clardy, J. Am. Chem. Soc., I l l ( 1989 ) 2712 - 2713. S. A. Morris, E. D. de Silva and R. J. Anderson, Can. J. Chem. 69 ( 1991 ) 768 - 771. J. E. Hochlowski, D. J. Faulker, G. K. Matsumoto and J. Clardy, J. Org. Chem. 48 ( 1983) 1142-1144. Y Kakou, R Crew and G. J. Bakus. J. Nat. Prod., 50 ( 1987 ) 482 - 484. D. J. Vanderah and R J. Schmitz, Lloydia 38 ( 1975 ) 4271 - 27 Y Kashman, A. Groweiss and U. Shmueli, Tetrahedron Lett. 21 (1980) 3629 - 3632. J. E. Thompson, R. P. Walker, S. J. Wratten and D. J. Faulkner, Tetrahedron Lett. 38 ( 1982) 1865-1873. M. Tischler and R. J. Andersen, Tetrahedron Lett. 30 (198 9) 5717 - 5720. J. Hellou, R. J. Andersen and J. E. Thompson, Tetradron 38 ( 1982) 1875 - 1879. M. Tischler, R. J. Anderson, M. L Choudhary and J. Clardy, J. Org. Chem., 56 ( 1991) 42 - 47. G. Cimino, M. Gavagnin, G. Sodano, R. Puliti, C. A. Maltia and L. Mazzarella, Tetrahedron, 44 (1988 ) 2301 - 2310. B. J. Burreson, R J. Scheuer, J . S. Finar and J. Clardy, J. Am. Chem. Soc, 97 ( 1975) 4763. M. R. Hagadone, B. J. Burreson, P. J. Scheuer, J. S. Finar and J. Clardy, Helv. Chim. Acta., 62 ( 1979 ) 2484 -24 N. Fusetani, H. J. Wolstenholme and S. Matsunaga, Tetrahedron Lett., 39 (1990) 56235624. J. A. Roesener and R J. Scheuer, J. Am. Chem. Soc, 108 ( 1986 ) 846 - 874. S. Matsunaga, N. Fusetani, K. Hashimoto, K. Koseki and M. Noma, J. Am. Chem. Soc, 108 ( 1980) 847 - 849. M. R. Keman, T. R Mollinski and D. J. Daulkner, J. Org. Chem., 53 (1988 ) 5014 5020. J. R. Pawlik, M. R. Keman, T. R Molinski, M. K. Harper and D. J. Faulkner. J. Exp. Mar. Biol. Ecol., 119 ( 1988 ) 99 - 109. G. Guella, I. Mancini, G. Chiasera and R Pietra. Helv. Chim. Acta., 72 (1989 ) 237 246. S. Carmely, M. Ilan and Y. Kashman, Tetrahedron, 45 (1989 ) 2193 - 2200. K. A. Alvi, R Crews and D. G. Loughhead, J. Nat. Prod., 54 (1991) 1509 - 1515. K. A. Alvi, B. M. Peters, G. M. Hunter and R Crews, Tetrahedron, 49 (1993 ) 329 336. P. Ciminiello, E. Fattorusso and E. Mangoni, Tetrahedron 46 (1990) 4387 - 4392 and references cited therein. H. He, D. J. Faulkner, A. Y Lee and J. Clardy, J. Org. Chem., 57 ( 1992 ) 2176 2178. R. Sanduja, A. J. Weinheimer, K. L. Euler and M. Alam, J. Nat. Prod. 48 (1985 ) 335 336. R. Sanduja, A. J. Weinheimer and M. Alam., J. Chem. Res. (S), ( 1985 ) 56 - 57. M. Murata, M. Shimatani, H. Sugitani, Y. Oshima and T. Yasumoto, Bull. Jpn. Soc. Sci. Fish., 48 (1982 ) 549 - 552 and refemeces cited therein. T. Yasumoto, M. Murata, Y Oshima, M. Sano, G. K. Matsumoto and J. Clardy, Tetrahedron , 41 (1985 ) 1019 - 1025.

31

88 89 90 91 92 93

94 95

96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120

M. Murakami, Y. Oshima and T. Yasumoto, Bull. Jpn. Soc. Sci. Fish. 48 ( 1982) 69 -72. K. Tachibana, P. J. Scheuer, Y. Tsukitani, H. Kikuchi, D. Van Engen, J. Clardy, Y. Gopichand and F. J. Schmitz, J. Am. Chem. Soc.. 103 ( 1981 ) 2469 - 2471. T. Yasumoto, Y. Oshima, W. Sugawara, Y. Fukuyo, H. Oguri, T. Igarashi and N. Fujita, Bull. Jpn. Soc. Sci. Fish. 46 ( 1 9 8 0 ) 1405-1411. M. Murata, M. Sano, T. Iwashita, H. Naoki and T. Yasumoto, Agric. Biol. Chem., 50 (1986) 2693-2695. M. Murata, M. Kumagai, J. S. Lee and T. Yasumoto. Tetrahedron Lett., 28 (1987 ) 5869 - 5872. Y Shimizu, A. V. K. Prasad, H-N, Chou, G. Wrensford and M. Alam, in: T. R. Tosteson (Ed.), Proceedings of the third International Conference on Ciguatera Fish Poisoning, Puerto Rico, PolyScience Publishers, Quebec, Canada, 1992, p p 3 3 - 4 3 and references cited therein. T. Hu, J. Doyle, D. Jackson, J. Marr, E. Nixon, S. Pleasance, M. A. Quilliam, J. A. Walter and J. L. C. Wright, J. Chem. Soc. Chem. Commun., (1992 ) 39 - 41. J. L. C. Wright, R. K. Boyd, A. S. W. de Freitas, M. Falk, R. A. Foxall, W. D. Jamieson, M. V. Laycock, A. W. McCulloch, A. G. Mclnnes, P. Odense, V. P. Pathak, M. A. Quilliam, M. A. Ragan. P. G. Sim, P. Thibault and J. A. Walter, Can. J. Chem., 67 (1989) 481-490. T. TAkemoto, K. Diago, Y. Icondo and K. J. Kondo, J. Pharm. Soc. Japan , 86 ( 1966) 874 - 877. J. L. C. Wright, M. Falk, A. G. Mclnnes and J. R Walter, Can. J. Chem., 68 (1990) 22 - 25. R. J. Anderson, D. J. Faulkner, H. Cun-hung, G. D. Van Duyne and J. Clardy. J. Am. Chem. Soc., 107 ( 1985 ) 5492 - 5495. R. Sanduja, G. S. Linz, M. Alam, A. J. Weinheimer, G. E. Martin and E. Ezell. J. Heterocyclic Chem., 23 (1986) 529-535. R. Sanduja, S. K. Sanduja, A. J. Weinheimer, and M. Alam, J. Nat. Prod., 49 (1986) 718-719. A. J. Weinheimer, C. W. J. Chang and J. A. Matson, in: W. Hcrz, H. Griebach and G. W. Kirby (Eds.), Fortschritte de chemie organischer Naturstuffe, Vol. 36, Springer-Verlag, New York, 1979, pp. 285 - 387 and references cited therein. G. S. Linz, R. Sanduja, A. J. Weinheimer, M. Alam and G. E. Martin, Tetrahedron lett. 27 (1986) 4833-4836. M. Alam, R. Sanduja and A. J. Weinheimer, Steroids, 52 (1988 ) 45 - 50. M. Alam, G. E. Martin, A. S. Zektzer, A. J. Weinheimer, R. Sanduja and M. A. Ghuman, J. Nat. Prod. 56 ( 1993 ) 774 - 779. A. R. Carrol and R J. Scheur, J. Org. Chem., 55 ( 1 9 9 0 ) 4426-4431. A. R. Carroll, N. M. Cooray, A. Pioner and P. J. Scheuer, J. Am. Chem. Soc., 54 ( 1989) 4231-4232. J. E. Hochlowski and D. J. Faulkner, Tetrahedron Lett. 24 (1983 ) 1917-1920. J. E. Hochlowski, D. J. Faulkner, G. K. Malsumoto and J. Clardy, J. Am. Chem. Soc., 105 (1983 ) 7413 - 7415. J. E. Biskupiak and C. M. Ireland, Tetrahedron Lett., 24 ( 1983 ) 3055 - 3058. R. J. Capon and D. J. Faulkner, J. Org. Chem., 49 ( 1984 ) 2506 - 2508. J. E. Hochlowski, J. C. Coll., D. J. Faulkner, J. E. Bixkupiak, C. M. Ireland, Z. Q-tai, H. Cun-heng and J. Clardy, J. Am. Chem. Soc., 106 (1984 ) 6748 - 6750. D. C. Manker, D. J. Faulkner, J. Org. Chem. 51 ( 1986) 814 -816. D. C. Manker, D. J. Faulkner, T. J. Stout and J. Clardy, J. Org. Chem., 54 (1989 ) 5371-5374. D. C. Manker and D. J. Faulkner, J. Org. Chem., 54 ( 1 9 8 9 ) 5374 - 5377. K. R Alhizati, J. R. Pawli and D. J. Faulkner, J. Org. Chem., 50 (1985 ) 3428 3430. D. C. Manker and D. J. Faulkner, Tetrahedron , 43 (1987 ) 3677 - 3680. C M. Ireland and D.J.Faulkner, Bioorg. Chem. 7 ( 1 9 7 8 ) 12 5 - 1 3 1 . D. M. Roll, J. E. Birkyupiak, C. L. Mayne and C. M. Ireland, J . Am. Chem. Soc. 108 (1986) 6680-6682. G. Cimino, S. De Rosa, S. De Stefano, and G. Sodano, Experientia, 41 (1985) 13351336. H. L. SUeper, V J. Paul anf W. Fenical, J. Chem. Ecol., 6 ( 1 9 8 0 ) 57 - 70.

32

121

D. C. Manker, M. J. Carson and D. J. Faulkner, J. Chem. Soc. Chem. Commun., (1988) 1061 - 1062.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 C) 1995 Elsevier Science B.V. All rights reserved.

33

Total Synthesis and Absolute Stereochemistry of Novel Biologically Active Marine Natural Products of Halenaquinol Family: Theoretical Studies of CD Spectra Nobuyuki Harada and Tatsuo Siigioka

1.

Introduction

In recent years, there have been many reports concerning isolation, structure determination, and biological activity studies of marine natural products. From marine sponges, many novel biologically active compounds have been isolated. For example, Scheuer and coworkers isolated halenaquinone (+)-(!), an antibiotic with a novel pentacyclic skeleton, from a tropical marine sponge of Xestospongia exigua collected in Western Caroline Islands.^ The structure of halenaquinone (+)-! was determined by X-ray crystallographic structure analysis. However, its absolute configuration has remained undetermined, because the X-ray crystallographic studies were carried out for compound 1 itself. Therefore, the X-ray Bijvoet method for determination of the absolute configuration on the basis of the anomalous dispersion effect of heavy atoms could not be applied.

(12b5)-(+).l

(12b5).(+).2

Kitagawa and coworkers isolated halenaquinol (+)-(2), a hydroquinone form of halenaquinone, from the Okinawan marine sponge Xestospongia sapra, together with halenaquinol sulfate (+)-(3).3 Halenaquinol 2 was easily oxidized either by UV-irradiation or by heating in the air to give halenaquinone (+)-!. Furthermore, Nakamura and coworkers isolated xestoquinone (+)-(4) from the same Okinawan sponge Xestospongia sapra as a powerful cardiotonic constituent.^ More recently, Schmitz and his coworker isolated tetrahydroxestoquinol (5), dihydrofuran compound (6), adociaquinone A (H-)-(7), adociaquinone B (+)-(8), and 3-ketoadocia-

34

quinone A (9) from a marine sponge, Adocia sp, from Truk Lagoon, in addition to halenaquinone 1 and xestoquinone 4.5 They also revealed that some of these novel marine natural products showed cytotoxicity. Considering the increasing interest due to the physiological activity of the novel compounds of the halenaquinol family, it is quite significant to determine the absolute stereochemistry of these compounds.

(12bS).(+).3

(12bS).(+)-4

(+)-7

(+)-8

(+)-9

35

To determine the absolute configuration of optically active organic compounds, there are two nonempirical methods. One is the Bijvoet method in the Xray crystallographic structure analysis, which is based on the anomalous dispersion effect of heavy atoms.^'^ xhe X-ray Bijvoet method has been extensively appHed to various chiral organic compounds since Bijvoet first succeeded in determination of the absolute stereochemistry of tartaric acid in 1951. The second method is a newer one based on the circular dichroism (CD) spectroscopy. Harada and Nakanishi have developed the CD dibenzoate chirality rule, a powerful method for determination of the absolute configuration of glycols, which was later generalized as the CD exciton chirality method.8 The absolute stereochemistry of various natural products has been determined by application of this nonempirical method. H

CH3

do cD (8aS)-(+).10

11

(8a/?)-12

(8a5).(+)-13

In addition to the CD exciton chirality method, we have recently reported that the theoretical calculation of the CD spectra by the 7C-electron SCF-CI-dipole velocity MO method8-14 has become an important tool for determination of the absolute configuration of a variety of twisted and conjugated K-electron systems. In fact, we have recently determined the absolute stereochemistry of (8aS)-(+)l,8a-dihydro-3,8-dimethylazulene 10, a labile biosynthetic intermediate for 1,4dimethylazulene 11 isolated from a liverwort, by application of the MO method to the theoretical calculation of the CD spectra of the twisted tetraene system (8a/?)12.15 In that case, we have also succeeded in the experimental verification of the absolute configuration theoretically determined, by comparison of the CD spectra of the natural product with those of synthetic chiral model compounds (8a5)-(+)-

(15aS).(-)-14

(18aS)-(-)-15

36 13.15 Moreover, we have theoretically determined the absolute stereochemistry of novel chiral troponoid spiro compounds (15a5)-(-)-14 and (18a5)-(-)-15 in a similar way; 16 the conclusion of the absolute configuration theoretically obtained was consistent with that of X-ray crystallographic studies. More recently, we have succeeded to determine the absolute stereochemistry of a natural atropisomer of biflavone, 4,'4"',7,7"-tetra-0-methylcupressuflavone (a/?)-(-)-16 isolated from Garcinia mangostane L., Araucaria cunninghamii, and A. cookii, by the molecular orbital calculation of the CD spectrum. 17 The K-electron SCF-CI-DV MO method OH O

CH3O

CH3Q

is thus powerful for non-empirical determination of the absolute configuration of twisted and conjugated 7c-electron systems. In this review, we account for the application of the TC-electron SCF-CI-DV MO method to the more complicated system of the natural products of the halenaquinol family. 18 We achieved the first total synthesis of these chiral halenaquinol, halenaquinone, xestoquinol, and xestoquinone with twisted 7C-electron systems. 1^*20 gy the total synthesis, we experimentally proved that the absolute stereochemistry of the compounds of the halenaquinol family theoretically determined was correct. 2.

Attempt to Apply the CD Exciton Chirality Method

The CD exciton chirality method has been extensively applied to degenerate systems consisting of two identical chromophores such as dibenzoate, bis(naphthalene), and bis(anthracene) compounds.^ In addition to such degenerate systems, the CD exciton chirality method is also useful for determination of the absolute stereochemistry of non-degenerate systems which contain two different chromophores. For example, Harada and Nakanishi determined the absolute configuration

37

of chromomycin A3 by use of the exciton coupling between the long axis polarized iBb transition of naphthalene chromophore and the long axis-polarized transition of /7-methoxybenzoate chromophore (Figure 1).8,21,22 We considered that this method would be applicable to the case of halenaquinol compounds. Namely, we planned to synthesize benzoate derivative 21 and to apply the CD exciton chirality method to the interaction between the naphthalene and benzoate chromophores of compound 21 (Scheme 1).18

+40

250 (+34.0) >^ R- p-Me0-C6H4C0-

+ 20 0

R- C6H5CO-

-20|-40 -60

In EtOH *-27l(-70.6)

-80 200

300

,, . X(nm)

400

Figure 1. CD spectra of chromomycin derivative benzoates.8

Halenaquinol 2 was methylated in refluxing acetone with iodomethane in the presence of potassium carbonate in the dark yielding dimethyl ether (+)-17 as yellow needles (Scheme 1): mp 235 °C; [a]23D+150.1° (CH2CI2). To differentiate the two carbonyl groups at the 3- and 6-positions, halenaquinol dimethyl ether (+)17 was selectively reduced with NaBH4 in the presence of CeCb • 7H2O, which

38

catalyzes the regioselective 1,2-reduction of conjugated enones (Scheme 1).23 Keto-alcohol (+)-18 was obtained as yellow needles, mp 258-259 ""C, and its stereostructure was determined by the ^H NMR coupling constant data. The alcohol was then converted to r^rr-butyldimethylsilyl ether (+)-19: mp 220-221 °C. To reduce the carbonyl group at the 6-position, keto-silyl ether (+)-19 was treated with NaBH4 /CeCls • 7H2O in methanol/dichloromethane. However, the obtained product which was postulated as alcohol 20 was extremely unstable and could not be obtained. Therefore, the preparation of naphthalene-benzoate compound 21 and hence the application of the CD exciton chirality method were unsuccessful.

CH3O

(12bS)-(+)-2 CH3O

CH3O

(+)-18

CH3O

20 Scheme 1.

(12bS)-(+).17

39

3.

The 7c-EIectron SCF-CI-DV Molecular Orbital Method for Calculation of CD Spectra

The CD and UV spectra of the compound with a twisted Tc-electron system can be calculated by the 7C-electron Self-Consistent-Field Configuration-Interaction Dipole-Velocity Molecular Orbital method (the 7C-electron SCF-CI-DV MO method).8-10 In the dipole velocity method, the rotational strength /?ba and dipole strength Dba which govern the sign and intensity of a CD Cotton effect and the intensity of a UV absorption band, respectively are formulated as follows: Rba = 2(\|/alVl\|/b)(\|/alr X VI\|/b)PM2 /(TCGba)

(1)

Dba = 2(X|/al VI\|/b)2pM2 /(TCaba)^

(2)

where V is the del operator, r is a distance vector, PM is the Bohr magneton, and Gba is the excitation wavenumber of the transition a->b. The z-axis component of the electric and magnetic transition moments are expressed, respectively, as^'l^ (ValVI\|/b)z= 2: (CraCsb~CsaCrb)COsZrs bonds

(3)

(\|/alr X Vl\|/b)z = £ (CraCsb - CsaCrb)(Xrs COS Yrs - Yfs cos Xfs) bonds COS Zrs = (Zr ~ Zs)/Rrs

(5)

Xrs = (Xr +Xs)/2

(6)

(4)

where Cra is the coefficient of atomic orbital r in the wave function \|/a, is the expectation value of a dipole velocity vector Vrs which is directed along the bond rs in the direction r->s, Xf, Fr, and Zr are the jc, y, and z coordinates of an atom r, respectively, and /?rs is the interatomic distance between atoms r and s. In a similar way, the x and y components of the electric and magnetic transition moments are calculated. In the TC-electron SCF-CI-CV MO calculation, the following standard values of atomic orbital parameters are employed: for sp2 carbons, Z(C) = 1.0, W(C) = -11.16 eV, (rrlrrXO = 11.13eV, pCC-C, 1.388 A) = -2.32 eV, (C-C, 1.388 A) = 4.70 X 107 cm-l; for ether oxygens, Z(0) = 2.0, H^(0) = -33.00 eV, (rrirr) (O) = 21.53 eV, p(C-O) = -2.00 eV, (C-0) = 6.00 x 10^ cm^l. The electric repulsion integral (rriss) can be estimated by the Nishimoto-Mataga equation. The

40

resonance integral and del value are calculated by employing the following equations, respectively: j8= [5/5(1.388 A)]J3(1.388 A) cos 9 = [(empir, 1.388 A)/(theor, 1.388 A)](theor) cos 9

(7) (8)

where 9 is a dihedral angle. The overlap integral S and (theor) are calculated on the basis of the Slater orbitals. The configuration interactions between all singly excited states are included. The curves of the component CD and UV bands are approximated by the Gaussian distribution Ae(a) = Z Aeic exp[-((a - ak)/Aa)2]

(9)

e(a) = E ek exp[-((a - ak)/AG)2]

(10)

where 2Aa is the \/e width of bands. The Aa value of 2500 cm""l is adopted as a standard value. 4.

Attempt of the Application of the 7t-EIectron SCF-CI-DV MO Method to Halenaquinol Dimethyl Ether

To determine the absolute configuration of halenaquinol (+)-2, we next tried to apply the 7i-electron SCF-CI-DV MO method to halenaquinol dimethyl ether (+)17, because halenaquinol dimethyl ether (+)-17 has a conjugated 7C-electron system composed of a naphthalene-ketone-furan-ketone chromophore which is twisted by the chiral center of the angular methyl group at the 12b position. 1^ Halenaquinol itself was not employed in this case because of its instability for light and heat (even at 40 °C). Furthermore, as a protecting group, a methyl ether group is better than an acetate group, because the 7C-electron system of dimethyl ether (+)-17 containing the lone-pair electrons of ether oxygens is simpler than that of halenaquinol diacetate. In the case of diacetate, the K-electron system becomes complex due to the contribution of the ester carbonyl moieties and its rotational conformation. Although we expected the relatively intense CD Cotton effects for (+)-17, the CD spectrum of (+)-17 showed weak Cotton effects as shown in Figure 2. The weak intensity of the CD Cotton effects may be due to the existence of two carbonyl groups of strong electron-withdrawing nature, which makes the total 7C-electron system of (+)-17 to be less symmetrical and hence the electronic transitions to be more complex and weaker. Therefore, from the view point of the reliability of the

41

determination, the compound (+)-17 was not suitable for the theoretical determination of the absolute configuration, because it is rather difficult to discriminate small positive and negative Ae values. In fact, we actually carried out the theoretical calculation of the CD spectrum of (+)-17, and the obtained results seemed to lead to the 12b5 absolute configuration for (+)-17. However we have not adopted the results as the convincing and unambiguous determination of the absolute configuration because of the small Ae values of CD Cotton effects.

X10

CH3O

500 CH3O (12b5)-(+) 244(4-4.7)

232(-9.3)

347(-h2.8) 303(~-5.6)

400

500 A/nm

Figure 2. CD spectrum of halenaquinol dimethyl ether (12bS)-(+)-17 derived from the natural sample of halenaquinol: solvent, EtOH.

5.

CD Spectra of Naphthalene-Diene Derivatives with Twisted Tc-Electron System

As described above, we could not synthesize alcohol 20, and therefore we failed to determine the absolute configuration of halenaquinol compounds by application of the nondegenerate CD exciton chirality method. However, we were very happy to find that the reductive reaction of ketone (+)-19 discussed above gave the rearranged products (-)-22 and (4-)-23 instead of 20 (Scheme 2). 18 it is considered that these compounds are derived from 20, which undergoes the elimination of the hydroxyl group and simultaneous addition of methanol at the 4position at the stage of working up. In fact, the reduction of ketone (+)-19 and subsequent treatment of the reaction mixture with a catalytic amount of aqueous hydrochloric acid gave trans-methoxy diene (-)-22 and d^-methoxy diene (+)-23

42

,OSi-|-

I

CH3O

I

I ' .....OCH3

CH3O

CH3O

CH3O

(-)-22

(+)-23

Scheme 2. in a moderate yield (Scheme 2). The structures of acetal epimers (-)-22 and (+)23 were determined on the basis the spectroscopic data; especially their relative configurations were unambiguously determined by the IH NMR coupling constant data and NOE enhancement data as shown in Figure 3.

CH3O

H 28 %

30%

Figure 3. NOE data and conformation of naphthalene-diene derivative (+)-23.18

43

The naphthalene-diene compounds (-)-22 and (+)-23 were also derived directly from halenaquinol dimethyl ether (+)-17 by the reduction and subsequent ferf-butyldimethylsilylation (Scheme 3). Diketone (+)-17 was reduced and then treated with hydrochloric acid, as in the case of ketone (+)-19 of Scheme 2, giving ^ranj-methoxy alcohol (-)-24 and cw-methoxy alcohol (-)-25, respectively. Each alcohol was then converted to its fer/-butyldimethylsilyl ether, which was identical with the authentic sample derived from compound (+)-19. CH3O (_).24 + (-).25 CH3O

(12b5)-(+)-17

CH3O

CH,0

CH3O

CH3O (-)-22

(-)-24

CH3O

CH3O

CH3O (-)-2S

(+)-23

Scheme 3. It was quite interesting that naphthalene-diene compounds (-)-22, (+)-23, (-)-24, and (-)-25 exhibited much stronger CD Cotton effects than other halenaquinol derivatives. For example, the UV spectrum of fran^-methoxysilyl ether (-)22 shows two intense n->n* bands (Figure 4): the broad band at 324 nm (e 27 000) with complex vibrational structures and the sharp band at 218 nm (e 42 000). In

44

the corresponding region, the CD spectrum of (-)-22 exhibits three major intense Cotton effects: X^n 338 nm (Ae +6.4), 301 nm (Ae -23.3), and 229 nm (Ae +40.9). Other three naphthalene-diene compounds (+)-23, (-)-24, and (-)-25 also exhibit three major CD Cotton effects of similar intensity and of the same sign as those of (-)-22. These results clearly indicate that the CD Cotton effects observed mainly originates from the 7i-electron chromophore composed of the naphthalene-diene

229(+40.9) + 40 + 30 + 20 + 10 O -10

Obsd In MeOH

vu -20 -30

12

200

300

, , , X(nm)

400

Figure 4. CD and UV spectra of trans-tiher (12b5)-(-)-22.18 moiety which is twisted by the angular methyl group at the 12b position. Namely, the additional chiralities due to the silyloxy group at the 3-position and the methoxy group at the 4-position are less contributory to the CD Cotton effects. In other words, these naphthalene-diene compounds are ideal systems for the theoretical determination of the absolute stereochemistry by application of the K-electron SCFCI-DV MO method.

45

6.

Application of the 7C-EIectron SCF-CI-DV MO Method to Naphthalene-Diene Derivatives

As a model compound for the theoretical calculation of CD spectra, we adopted the molecule (12bS)-26 having the essential part of the 7C-electron system contained in the naphthalene-diene compounds 22-25.1^ Namely, in addition to the naphthalene and conjugated diene chromophores, the lone-pair electrons of the two methyl ether and furan ring oxygens are also included. The absolute configuration of 26 was arbitrarily chosen to be 12b5 for the calculation. The molecular geo-

CH3O

CH3O (12b5).26 metry of the model compound (12b5)-26 was calculated by the molecular mechanics (MMP2)24 as illustrated in Figure 5. The molecular framework of this model compound is relatively rigid, and the D-ring takes a half-chair conformation. These molecular conformations were confirmed by the IH NMR coupling constant and NOE enhancement data of compounds (-»-)-23 and (--)-25 (Figure 3). The part of the double bond and naphthalene chromophores of (12b5)-26 constitutes a clockwise helicity (dihedral angle of 5a-6-6a-7: +170°), while the conjugated diene

(12b5)-26 Figure 5. Molecular conformation of the model compound of naphthalene-diene (12b5)-26 calculated by the molecular mechanics (MMP2).18

46

moiety constitutes a counterclockwise helicity (dihedral angle of 3a-12c-5a-6: -167°). The helical sense of these two moieties is not changed, even if the D-ring takes a boat conformation. Namely, the sense of the twist of the conjugated nelectron system is governed solely by the chirality of the angular methyl group at the 12b position. The theoretical calculation of the CD and UV spectra of (12bS)-26 by the 7Celectron SCF-CI-DV MO method gave the curves illustrated in Figure 6. The UV spectrum curve exhibits two intense 7C~>7C* bands: a broad band at 349 nm (e

223(+35.5)

+40

CH3O

+ 30

r i i ^'^^v^^Y^^/

UJi^^^A^::::^©

+ 20 + 10

CH30

(l2bS)

- / 1 ^^

378(+3.3)

0 I

f 248(-5.7)

-10

/

219(40,300)

-20

o

Calcd

X

322(-22.4)

-30

.349(29,900)

H3

/

1

200

\

UV

1

1

300

X(nm)

\

^^"^-f-^

400

Figure 6. CD and UV spectral curves of the model compound naphthalene-diene (12bS)-26 calculated by the 7C-electron SCF-CI-DV MO method. 18

47

29,900) and a sharp band at 219 nm (e 40,300). These calculated values agree closely with the observed UV data of (~)-22 and other naphthalene-diene derivatives: for (-)-22, A,max 324 nm (8 27,000) and 218 nm (e 42,000) (Figure 4), In the corresponding region, the calculation afforded three principal CD Cotton effects: a weak positive band at 378 nm (Ae +3.3), a negative one of medium intensity at 322 nm (Ae -22.4), and a positive intense one at 223 nm (Ae +35.5). These theoretically obtained CD values are in a good agreement with the observed data of (-)-22 and other naphthalene-diene compounds: for (-)-22, XQXX 338 nm (Ae +6.4), 301 nm (Ae -23.3), and 229 nm (Ae +40.9) (Figure 4). It is thus

+ I00f-

200

300

X(nm)

400

Figure 7. Rotational and dipole strengths of the transitions of the model compound of naphthalene-diene (l2b5)-26 calculated by the 7i-electron SCFCI-DVMOmethod.l8

48

evident that the calculation well reproduced the basic pattern of the CD and UV spectral curve, including the sign, position, intensity, and shape of the bands. Since the absolute configuration of the model compound 26 is fixed to be 12b5, the comparison of the present calculated and observed CD data leads to the unambiguous determination that (-)-22 and other naphthalene-diene compounds have the 12b5 absolute configuration. Accordingly, the absolute stereochemistry of halenaquinol (+)-2 was theoretically determined to be 12b5. Since the UV irradiation of halenaquinol (+)-2 gave halenaquinone (+)-l and the solvolysis of halenaquinol sulfate (+)-3 qualitatively yielded halenaquinol (+)-2, the absolute stereostructures of (+)-l and (+)-3 were also established to be \2hSy respectively. 7.

Circular Dichroic Power of a Twisted Naphthalene-Diene System

In the case of (8a5)-(+)-l,8a-dihydroazulene 10,^5 the composition of the apparent CD and UV bands was rather simple, because each of the apparent bands was composed of a single electronic transition. The case of chiral troponoid spiro compounds (15a5)-(~)-14 and (18a5)-(-)-15 was also simple because of their C2 symmetrical structures. 1^ On the other hand, the 7i-electron chromophores of the twisted naphthalene-diene systems 22-26 are complex and have no symmetric character. 18 Therefore, to clarify the applicability of the 7C-electron SCF-CI-DV MO method to such complicated systems, it is important to analyze the composition of the apparent CD and UV bands theoretically obtained. As illustrated in Figure 7, there are nine major electronic transitions which contribute to the CD and UV bands. The first and second electronic transitions with weak positive rotational strengths at 374.5 and 351.6 nm, respectively, generate the weak positive Cotton effect at 378 nm (Figure 7). Furthermore, the third electronic transition with an intense negative rotational strength at 324.4 nm results in the negative Cotton effect at 322 nm, and the sixth electronic transition with a strong positive rotational strength mainly contributes to the intense positive Cotton effect at 223 nm. The correspondence between the CD rotational strengths theoretically calculated and the apparent CD Cotton effects is thus unambiguous. Therefore, this analysis makes the theoretical determination of the absolute configurations of the halenaquinol compounds to be more reliable. 8.

The Synthetic Strategy for the Total Synthesis of (+)-Halenequinol and (+)-Halenaquinone

As discussed above, we succeeded in the theoretical determination of the absolute stereochemistry of novel marine natural products of halenaquinol family. It is quite natural that chemists, as the next step, want to prove their absolute configurations theoretically determined in an experimental way. So, we started to

49

synthesize halenaquinol and halenaquinone in the natural enantiomeric forms and planned to corroborate their absolute stereostructures by comparison of the CD spectrum of the synthetic sample with that of the natural one.'9 As a synthetic strategy, we performed the retrosynthesis and adopted the synthetic route shown in Scheme 4, where halenaquinol (12b5)-(+)-2 can be prepared by the reduction of halenaquinone (12b5)-(+)-l. The naphthoquinone

CH3O

CH3O > = >

CH3O

CH3O 27

(12bS)-(+)-17

CH3O = > CH3O

29

28

..o-

^ . nc""° ^ Vy^' 31

32

Scheme 4.

(8aR)-(-)-33

31

34

50

moiety of (12b5)-(+)-l can be obtained by the oxidative cleavage of the hydroquinone dimethyl ether (12b5)-(+)-17, and the furan ring of (12b5')-(+)-17 is obtainable by the oxidation of triol 27. The diosphenol moiety of 27 would be obtained by the air oxidation of ketone 28, and the tetracyclic skeleton of 29 is constructed by the Diels-Alder reaction between 3,6-dimethoxybenzocyclobutene 30 and enone 31. The dienophile 31 is derived starting from the WielandMiescher ketone 33 as shown in Scheme 4, where the extra one-carbon unit is introduced as the hydroxymethyl group at the C-5 position of compound 32 by application of the Stork's reductive hydroxymethylation method.25 in this synthetic route, the absolute configuration of the bridgehead methyl group of the optically active Wieland-Miescher ketone is retained as that of the corresponding methyl group of the final product. Therefore, since the absolute configurations of halenaquinol and halenaquinone have been theoretically determined to be 12b5 as discussed above, it indicates that to synthesize the natural enantiomeric forms of halenaquinol and halenaquinone, we should start from the (8a/?)-(-) enantiomer of the Wieland-Miescher ketone 33. This is one of the theoretical requirements for the total synthesis of natural halenaquinol and halenaquinone. 9.

Efficient Preparation of Optically Pure (8a/?)-(-)-WielandMiescher Ketone

Optically active Wieland-Miescher ketone 33 has been used as a key starting material for the total syntheses of various optically active and biologically active natural products.26,27 Among several preparative methods^^ for optically active

(8a/?)-(-)-33

34

33, the most practical method is the asymmetric cyclization of the prochiral triketone 34 with a catalytic amount of optically active proline; the preparation method was independently discovered by the Hajos' group29 and also by the Eder's group.30 Since the catalytic asymmetric synthesis gives compound 33 of moderate optical purity (ca. 70%), Furst and coworkers developed the procedure of enantiomeric purification by recrystallization from diethyl ether.^l In connection with the syntheses of (+)-l,8a-dihydroazulene derivative (8a5)-(+)-13 and related compounds, 15 we used the Furst's method for the preparation of optically pure 33. However, the enantiomeric purification could not be consistently reproduced by

51

recrystallization from diethyl ether. In some cases crystals of high optical purity were obtained, but in other cases crystals of low optical purity were deposited. We found a more reproducible procedure for the enantiomeric purification of optically pure 33^2 and also confirmed its absolute stereochemistry by application of the CD exciton chirality method to bis(p-dimethylaminobenzoate) derivatives.^ The preparative procedure for prochiral triketone 34 was simplified; a solution of 34 and unnatural D-(+)-proline in dimethyl sulfoxide was degassed and then stirred at room temperature for 6 days, although Crabbe and coworkers reported a time saving variation of the preparation by using DMF as solvent.24 The reaction mixture was worked up, and the crude product obtained was purified by a vacuum distillation and by a column chromatography on silica gel to give optically active (8a/?)-(--)-33, [a]D -68.4° (c 1.355, benzene) as a syrup in 82% chemical and 69% optical yields. The syrupy 33 was dissolved in a mixed solvent of diethyl ether and ethyl acetate (10:1), and the solution was kept at -70 °C for about 5 h. In this case, no seed crystals of optically pure (8a/?)-(-)-33 were needed. Crystals deposited were collected by filtration to yield enantiomerically enriched 33: [aJD -95.8° (c 1.10, benzene). The crystals obtained were further enantiomerically purified by two additional recrystallizations, to give optically pure (8a/?)-(-)-33; mp 50.5-51.0 ''C, [a]D -98.96** (c 1.039, benzene). We carried out the preparation of optically pure (8a/?)-(-")-33 with this procedure three times, and all gave satisfactory results. Optically impure 33, [a]D -36° (c 1.12, benzene), obtained from the mother liquor fractions, was similarly crystallized by adding a few crystals of racemic 33 as seeds. The crystals obtained were almost racemic 33, [a]D -5.3° (c 1.25, benzene), and the mother liquor yielded enantiomerically enriched 33, [a]D -80.8° (c 1.13, benzene), which was fed back to the recrystallization steps described above to obtain additional optically pure 33. 10.

Confirmation of the Absolute Stereochemistry of WielandMiescher Ketone by the CD Exciton Chirality Method

The CD exciton chirality method^ is very powerful for the determination of the absolute stereochemistry of chiral organic compounds. To confirm the absolute OR

^cb

OR

job

(l/?,6R,8aR)-(-)-37, R = H

(lR,6S,SaRH-)-38, R = H

(lR,6i?,8aiR)-39, R = Bz-/»-NMe2

(lR,6S,SaR)-40, R = Bz-p-NMCi

52

configuration of (-)-33, the two epimeric bis(p-dimethylaminobenzoates) 39 and 40 were synthesized.^^ x^e relative stereochemistry of hydroxyl and benzoate groups was easily determined from the iH-NMR coupling constant and half-band width data. The p-dimethylaminobenzoate chromophore was selected as the best of the /7-substituted benzoate groups for the observation of of the exciton split CD Cotton effects, and it can be easily introduced by the esterification method using /7-dimethylaminobenzoic acid and 2-chloro-l-methylpyridinium p-toluenesulfonate.33,34 A.S shown in Figure 8, the UV spectrum of bis(/7-dimethylaminobenzoate) 40 shows strong 7i->7C* absorption at A^max 310 nm. In the corresponding region, the CD spectrum of 40 exhibits strong exciton split CD Cotton effects of negative first and positive second signs (A = -88.8). These CD data indicate that two benzoate groups constitute a counterclockwise screw sense, leading to the unambiguous + 50 320.0 ( + 2 3 . 4 ) CD In EtOH

292.0 ( + 1 9 . 5 ) /

A€

0

1

^^j^^^__^

^

\i ^^'^^y I ^ M ^ '^

1

/

2 9 6 . 0 ( - l 4 . 2 ) ---

/

\

1

\

T

-50 - /

310.3(58,800)-

UV

200

/

320.0 (-69.3)

-100 \

/

\

1

X

/

\

. 0^0

o

/

2 2 5 . 0 (13,900)

250

V-^

/ /

X(nm)

\

] \

300

350

Figure 8. CD and UV spectra (solid line) of trans- bis(/7-dimethylaminobenzoate) 40 in EtOH, and CD spectrum (dotted line) of d5-bis(p-dimethylaminobenzoate) 39 in EtOH.32

53

assignment of the (l/?,65,8a/?) absolute configuration to 40. Bis(p-dimethylaminobenzoate) 39 exhibits CD Cotton effects of positive exciton chirality (A = +37.6), leading to a clockwise screw sense, i. e., the (l/?,6/?,8a/?) absolute configuration. The CD intensities of the exciton split Cotton effects of 39 are weaker than those of 40. This phenomenon can be explained by the fact that the two equatorial benzoate groups of 39 are far from each other, lying approximately in the same single plane, while the axial and equatorial benzoate groups of 40 are somewhat closer to each other and are in nearly orthogonal planes, thus constituting a definite negative exciton chirality.^ The (8a/?) absolute configuration of (-)-33 was thus confirmed by the CD exciton chirality method. 11.

Absolute Configuration of (-)-Wie!and-Mlescher Ketone as Established by the X-Ray Crystallographic Method

The optically active Wieland-Miescher ketone 33 has been widely employed as an important key compound as discussed above. However, it was surprising that irrespective of such a significance of compound 33, its absolute configuration has never been directly determined by the X-ray crystallographic method. The absoOR

o—^ (8a/?).(-)-33

Scheme 5.

OR

,..-C0 * „„;Cb

RO^'

-

^

RO

(ll?,6/?,8a/?).(-)-37 R=H

(l/?,6S,8ai?)-H-38 R=H

(li?,6/?,8ai?)-(+)-41 R = Bz./?.Br

(l/?,6S,8a/?).H-42 R = Bz-p-Br

lute configuration of 33 has been originally determined on the basis of the comparison of the optical rotational data of derivatives of 33 with those of steroidal compounds by Prelog and his coworker.^^ In addition to the determination of the absolute configuration of (-)-33 by application of the CD exciton chirality method described above and the indirect determination of the absolute configuration of (+)33 by the X-ray crystallography of a l,8a-dihydroazulene derivative,!^ we carried out the direct determination of the absolute stereochemistry of (-)-33 by the X-ray structure analysis of its bis(/7-bromobenzoate) derivatives (+)-41 and (-)-42.36 Optically pure Wieland-Miescher ketone (-)-33 ([a]D -98.96°) was reduced with LiAlH4 to yield a mixture of epimeric glycols, which were separated by HPLC on silica gel; c/.y-glycol (-)-37 of the first-eluted fraction and trans-g\yco\ (-)-38 of the second-eluted fraction were converted to bis(p-bromobenzoates) (+)-41 and

54 (-)-42, respectively (Scheme 5). The relative stereochemistry of 41 and 42 was determined by the iH NMR data. Both bis(p-bromobenzoates) were recrystallized from ethyl acetate to give suitable crystals for X-ray diffraction. The crystals of (+)-41 were found to be orthorhombic and the space group to be P2i2i2i; a = 12.214 A, b = 31.5143 A, c = 6.076 A, V = 2338.3 A3; p(calcd) = 1.557 g/cm^, p(obsd) = 1.561 g/cm3. The crystal structure was solved by the direct method and by the successive Fourier syntheses. The least-square refinement of positional and thermal parameters, including anomalous scattering factors, led to the final convergence with R = 5.26% for the (l/?,6/?,8a/?) absolute configuration, while a similar calculation for the mirror image structure gave R = 5.62%. So, the absolute stereochemistry of (+)-41 was determined to be (l/?,6/?,8a/?).

Figure 9. ORTEP drawing of (l/?,6/?,8a/?)-(+)-bis(/7-bromobenzoate) 41.36 The crystal structure of (-)-42 was similarly determined; orthorhombic, P2i2i2i; a = 17.251 A, b = 21.155 A, c = 6.540 A, V = 2386.5 A3; p(calcd) = 1.526 g/cm3; p(obsd) = 1.516 g/cm3. Since the final R-value (5.40%) of (l/?,65,8a/?) absolute configuration was smaller than that of the mirror image structure (R = 5.88%), the (l/?,65,8a/?) configuration was assigned to (-)-42. The 8a/? absolute configuration of Wieland-Miescher ketone (-)-33 was thus doubly confirmed by the X-ray crystallographic structure analysis.

Figure 10. ORTEP drawing of (l/?,65,8a/?)-(-)-bis(p-bromobenzoate) 42.36

55

12.

The First Total Synthesis of (+)-Halenequinol and (+)-Halenaquinone

We achieved the first total synthesis of (+)-halenaquinol 2 and (4-)-halenaquinone 1 as follows.^^ The carbonyl group at the 1-position of the optically pure (8a/?)-(-)-Wieland-Miescher ketone 33, [a]25D-98.96° (c 1.039, benzene), was selectively protected to give monoacetal (-)-43,^7 which was then reductively hydroxymethylated according to the procedure of Stork25,38 (Scheme 6); enone (-)-43 was reduced with lithium in liquid ammonia, and the resultant enolate was

^^ (8a/f)-(-)-33

(+)-32

cCb

~^-kJ^

(-)-43

44

(+)-46

(+)-4S

~Jp - Xp -XfK > <

(-)-47

(-)-48

(+)-31

Scheme 6. trapped as trimethylsilyl ether 44. Regeneration of the enolate anion by treatment of 44 with methyllithium and then addition of gaseous formaldehyde gave keto alcohol (+)-32 as a sole stereoisomer in 82% overall yield from (-)-43. Keto alcohol (+)-32 was reduced with lithium tri-^ec-butylborohydride (L-Selectride), yielding cis-glycol (+)-45 in 92% yield, which was then converted to keto glycol (+)-46 in 98% yield by treatment with /?-toluenesulfonic acid in water. The rela-

56

tive stereochemistry of 6(ax)-hydroxyl and 5(eq)-hydroxymethyl groups of compounds (+)-45 and (+)-46 was secured by the ^H NMR coupling constant data of keto acetonide (+)-49 (Scheme 7 and Figure 11): ^H NMR (300.15 MHz, CDCI3) 8 3.884 (IH, ddj = 12.2, 1.1 Hz), 3.961 (1 H, dd, J = 12.2, 2,6 Hz), 4.068 (1 H, ddd, / = 2.6, 2.6, 2.6 Hz).

(+)-45

^

i

l

l

^

(-)-48

^H NMR (CDCI3, 300 MHz) Ha: 4.068 (1 H, ddd, J = 2.6, 2.6, 2.6 Hz) Hb: 3.961 (1 H, dd, J = 12.2, 2.6 Hz) He: 3.884 (1 H, dd, J = 12.2, 1.1 Hz)

Figure 11. ^H NMR coupling constant data of keto-acetonide 49. Formation of the p-toluenesulfonhydrazone of (+)-46, followed by treatment with methyllithium, gave olefin (-)-47 in a quantitative yield. Next, the glycol moiety of (-)-47 was protected as an acetonide to give acetonide olefin (~)-48. We also checked the altemative shortcut path; ketal glycol (+)-45 was converted to olefin acetonide (-)-48 via keto acetonide (-f)-49 (Scheme 7). However, the acetonide moiety of (+)-49 was found to be unstable toward the tosylhydrazone formation reaction, and the yield of (-)-48 was low. Finally, the allylic position of acetonide olefin (-)-48 was oxidized with the reagent of Cr03/3,5-dimethylpyrazole,^^ giving conjugated enone (+)-31 in 63% yield, which was next used as a dienophile of the Diels-Alder reaction with 3,6-dimethoxybenzocyclobutene 30.

57

Although dimethoxybenzocyclobutene 30 had been previously synthesized by the method of photocycloadditionj'^O ^e synthesized it by the pyrolysis of sulfone 53, which was prepared starting from 2,3-dimethy 1-1,4-dimethoxybenzene 50^1 as shown in Scheme 8. Bromination of 50 with N-bromosuccinimide (NBS), followed by treatment of the resulting dibromide 51 with sodium sulfide in aqueous ethanol, gave sulfide 52 in 70% yield. Oxidation of 52 with m-chloroperbenzoic acid in dichloromethane afforded sulfone 53 in 89% yield. For the next thermal elimination reaction of sulfur dioxide, various reaction conditions were examined, and we finally found that the direct heating method of the solid material of sulfone 53 without any solvents afforded the desired 3,6-dimethoxybenzocyclobutene 30 in a moderate yield. The crystals of 53 were pyrolyzed at 305-310 °C in a muffle furnace under a stream of nitrogen to give 30 in 48% yield. CH3O

50

51

52

Schemes. The Diels-Alder reaction of 3,6-dimethoxybenzocyclobutene 30 with some dienophiles was studied (Scheme 9). A mixture of benzocyclobutene 30, N-phenylmaleimide 54, and benzene was heated at 210 °C in a sealed tube. The adduct 55 was obtained in 53% yield. The reaction of 30 and cyclohexenone 56 similarly gave the adduct 57 in 36% yield. In the case of dienone 58, however, the desired product could not be obtained. On the other hand, the Diels-Alder reaction with trans-tnont 59 afforded adduct 60 in a good yield (76% yield). Therefore, we felt confident that the reaction of benzocyclobutene 30 with enone (+)-31 would give the desired adduct with a tetracyclic skeleton.

58

CH3O

55

54

Q o

56

^-

dienone recovered

58

CH3O

The Diels-Alder reaction of compounds 30 and (+)-31 was achieved by heating a benzene solution in a sealed tube at 210-215 °C for 20 h, giving a tetrahydronaphthalene derivative (+)-29 in 33% yield (Scheme 10). To improve the chemical yield of the adduct, we checked the reaction conditions, but all attempts brought no improvement. The l^C NMR spectrum of (+)-29 indicates that the product was

59

composed of a single stereoisomer. However, the relative stereochemistry of the chiral centers newly formed was not investigated further because they vanish at the next dehydrogenation reaction. To dehydrogenate the tetrahydronaphthalene moiety of (+)-29, a benzene solution of (+)-29 was refluxed with 2,3-dichloro-5,6dicyano-l,4-benzoquinone (DDQ) to afford a naphthalene derivative (-)-28 in 89% yield, which was then subjected to the following air oxidation in the presence of a base. Oxygen gas was bubbled through a solution of (~)-28 and potassium tertbutoxide in tert-butyl alcohol for 5 h, and the mixture was worked up with aqueous ammonium chloride to give diosphenol 61 in 90% yield. The structure of 61 was secured by the iH NMR (the sharp singlet at 5 7.60 disappeared when adding D2O), UV (a red shift and a hyperchromic effect of the UV absorption band at the longer

CH3O

CH3O

^ IL J""H CH3O

CH30

(+)-31

30

(+)-29

CH3O

CH3O

CH3O

CH3O (-)-28

61

Scheme 10. wavelength region when adding aqueous NaOH), and high resolution MS data. Deprotection of the acetonide group of 61 by treatment with 60% aqueous acetic acid yielded triol 27, which was subjected to the next reaction without purification (Scheme 11). The oxidation reaction of the primary and secondary hydroxyl groups of 27 and successive cyclization to form a furan ring were accomplished by treatment with dimethyl sulfoxide (DMSO) and 1,3-dicyclohexylcarbodiimide (DCC) in benzene in the presence of trifluoroacetic acid and pyridine, giving halenaquinol dimethyl ether, the desired compound (12bS)-(+)-17 of the furandiketone system, in 44% overall yield from (-)-28. All of the spectroscopic data of the synthetic sample of halenaquinol dimethyl ether (+)-17 were completely

60

identical with those of the authentic sample of (+)-17 derived from natural halenaquinol. The hydroquinone dimethyl ether moiety of (12b5)-(+)-17 was next deprotected by the oxidative cleavage with cerium(IV) ammonium nitrate (CAN) in aqueous methanol affording halenaquinone (12b5)-(+)-l of pale yellow color in 45% yield (Scheme 11). The ^H NMR and UV spectra of the synthetic sample agreed with those of natural halenaquinone. Finally, halenaquinone (12b5)-(+)-l was reduced with aqueous sodium hydrosulfite in acetone to give halenaquinol (12b5)-(+)-2 of yellow color in an almost quantitative yield. As discussed by Kitagawa and coworkers in their paper of the isolation of natural halenaquinol,^

CH3O

61 CH3O

CH3O 27

(12bS)-(+)-l

O

(12bS).(+)-17

(12bS)-(+).2

Scheme 11. compound 2 was very sensitive to light, heat, and air. So, the reaction was carried out in a dark room. The IH NMR spectrum of halenaquinol (12b5)-(+)-2 in DMS0-d6 exhibited two broad singlets at 5 9.6 and 9.8 due to the phenolic hydroxyl groups, which disappeared when D2O was added. The remaining ^H NMR peaks and UV spectrum curve were in a good agreement with those of the natural sample, respectively. The hydroquinone structure of (12b5)-(+)-2 was also secured by the mass spectrum; although the usual measurement procedure of mass spectra afforded only the signals due to the quinone form (12b5)-(+)-l, the peaks of the hydroquinone compound (12b5)-(+)-2 were exclusively obtained by application of the direct injection method of the solid sample. All of the mass spectra of

61

halenaquinol 2, halenaquinone 1, and halenaquinol dimethyl ether 17 show the M CH3 peaks as base peaks, respectively. For example, halenaquinol 2 exhibits miz 319 as a base peak (High-resolution mass spectrum calcd for C20H14O5 - CH3: 319.06064, Found: 319.06161). These results indicate that the halenaquinol and halenaquinone skeletons easily lose the angular methyl group, respectively. The first total synthesis of halenaquinol (12b5)-(+)-2 and halenaquinone (12b5)-(+)-l with a novel polyketide skeleton has been thus accomplished. 13.

CD Spectra of Halenaquinol Dimethyl Ether and Experimental Proof of the Absolute Stereochemistry of the Halenaquinol Family

In the case of the total synthesis of halenaquinol (12b5)-(+)-2 and halenaquinone (12b5)-(+)-l, we started from the Wieland-Miescher ketone (8a/?)-(-)-33, as discussed above. Therefore, it is evident that the synthetic sample of halenaquinol dimethyl ether (+)-17 has the (12b5) absolute configuration. 19 If the theoretical determination of the absolute stereochemistry of the halenaquinol family is correct, the chiroptical data of [a]D and CD spectra of the synthetic sample should be identical with those of the authentic sample of (+)-17 derived from

CH3O

CH3O

Synthetic

(12bS).(+) 244(4-4.5) 244(4-4.5]

j^jL

200

\J

(_9.4) 232(-9.4)

347(4-2.6) 303(-5.6)

loo"

"loo A/nm

Figure 12. CD spectmm of the synthetic sample of halenaquinol dimethyl ether (12b5)-(4-)-17. natural halenaquinol. This was verified as follows: synthetic sample, [a]25D 4-150.3° (c 1.042, CH2CI2), CD (EtOH) Xext 413 nm (Ae 4-1.8), 383 (-H1.4), 363 (4-1.6), 347 (4-2.6), 303 (-5.4), 244 (+4,7), 232 (-9.1); natural sample, [a]23D +150.1° (c 1.124,CH2Cl2), CD (EtOH) ^ext 413 nm (Ae +1.8), 383 (+1.4), 363 (+1.7), 347 (+2.8), 303 (-5.5), 244 (+4.6), 232 (-8.9); compare the CD spectrum of the synthetic sample of halenaquinol dimethyl ether (+)-17 shown in Figure 12

62

with that of the authentic sample of (+)-17 derived from natural halenaquinol depicted in Figure 2. Since the absolute configuration of the angular methyl group is retained throughout the reactions discussed above, these results lead to the experimental and unambiguous determination that the absolute stereochemistry of (+)-halenaquinol and (+)-halenaquinone is 12b5. In addition, these synthetic results also proved that the absolute configurations of halenaquinol compounds theoretically determined were actually correct. 14.

The First Total Synthesis of (+)-Xestoquinone and Xestoquinol

We achieved the first total synthesis of (+)-halenaquinol 2 and (+)-halenaquinone 1, by which the absolute configurations of (+)-l and (+)-2 theoretically determined were experimentally proved to be correct. On the other hand, the absolute stereochemistry of xestoquinone has remained undetermined. Next, we carried out the first total synthesis of xestoquinone (+)-4 and xestoquinol 62 although the latter compound has never been isolated as a natural product.^O

(12bS).(+).4

(12bS)-62

As a synthetic strategy, the route shown in Schemes 12-15 was adopted. Although the absolute configuration of xestoquinone 4 had not been determined, we assumed that xestoquinone had the same absolute configuration as halenaquinone (+)-l. The absolute configuration of (+)-l was determined to be 12b5 by the O.

(8a/?)-(-)-33 Scheme \2.

c

O

(+)-32, R=H +)-63, R=CH20CH3

63

theoretical calculation of the CD spectruml^ and also by the total synthesis,^^ as discussed above. Therefore, we started from hydroxymethyl ketone (4a/?,55,8a/?)(+)-32,19 [a]20D +3.6° (c 1.707, CHCI3), which was derived from WielandMiescher ketone (8a/?)-(-)-33 (Scheme 6). The hydroxy 1 group of (+)-32 was protected as a methoxymethyl ether to give ether (+)-63 in 80% yield (Scheme 12). To reduce the carbonyl group of (+)63 to a methylene moiety of (+)-66 (Scheme 13), we tried the thioacetal reduction method. However, the attempt was unsuccessful because of the exchange of the acetal group at the 1-position by a thioacetal group. Although we also tried the

o.r\ o (+)-63

I

(+)-64, R=CH20CH3, R ' = 0 H

(+)-67, RsCHjOCHj

pl^(+)-65, R=CH20CH3, R*=0(C=S)SCH3 ^ ( + ) - 6 6 , R=CH20CH3, R'=H

RO-^ " (-).68, RrzCHjOCHa

no-" " (+)-69, R=CH20CH3

Scheme 13, Huang-Minion reduction of a similar compound, the method was inapplicable. So, we adopted the method of radical reduction of xanthate (+)-65 instead.^^ Ketone (+)-63 was stereoselectively reduced with lithium tri-^ec-butylborohydride (LSelectride), yielding axial alcohol (+)-64 in 92% yield, which was then con-verted to xanthate (+)-65 in quantitative yield by treatment of lithium alkoxide of the alcohol with carbon disulfide and then with iodomethane. Xanthate (+)-65 was next reduced with tributyltin hydride, giving the desired methylene compound (+)66 in 92% yield. After deprotection of the acetal group of (+)-66 (98% yield), ketone (+)-67 was converted to olefin (-)-68 in 98% yield by formation of tosyl-

64 hydrazone followed by treatment with methyllithium. The allylic methylene part of olefin (-)-68 was oxidized with CrOs and 3,5-dimethylpyrazole,'*3 yielding conjugated enone (+)-69 in 55% yield. The Diels-Alder reaction to construct the molecular skeleton of the tetracyclic system was carried out by heating a mixture of 3,6-dimethoxybenzocyclobutene 30 and enone (+)-69 at 210-220°C for 10 h, giving the desired adduct (+)-70 in 40% yield (Scheme 14). The '3c NMR spectrum of (+)-70 indicated that the product was composed of a single stereoisomer. Tetrahydronaphthalene derivative CH3O

CH3O

(+).69, R=CH20CH3

30

(+)-70, R=CH20CH3 CH3O

CH3O ... ^OR

CH3O

CH3O

(-)-71, R=CH20CH3

72, R=CH20CH3

CH3O

CH3O

CH3O

CH3O 73

(12b5)-(+)-74

Scheme 14. (+)-70 was dehydrogenated with 2,3-dichloro-5,6-dicyano-l,4-benzoquinone (DDQ), giving naphthalene derivative (-)-71 in 80% yield. To construct the furan ring system, we carried out the following reactions; ketone (-)-71 was next subjected to an air oxidation in the presence of a base. Oxygen gas was bubbled into a

65

solution of (-)-71 and potassium tert-butoxidc in tert-buiyl alcohol, yielding diosphenol 72. The IH NMR spectrum of 72 showed a sharp singlet peak at 5 7.095 characteristic of the intramolecular hydrogen-bonded hydroxyl group of a diosphenol moiety; the ^H NMR spectrum of the diosphenol moiety generally exhibits a singlet peak of hydroxyl group proton around 5 6.9-7.6.l^'^^ Deprotection of the methoxymethyl group of 72 by treatment with concentrated hydrochloric acid in methanol gave alcohol 73. To perform the next oxidation reaction of the primary hydroxyl group of 73 to aldehyde and successive cyclization to form a furan ring, we at first tried the dimethyl sulfoxide (DMSO)/l,3-dicyclohexylcarbodiimide (DCC) oxidation method, which was successful in the case of the total synthesis of halenaquinol series. 1^ However, all attempts employing DMSO and various auxiliary reagents (DCC, acetic anhydride, or oxalyl chloride), pyridinium chlorochromate, pyridinium dichromate, and N-bromosuccinimide were unsuccessful. The method of oxidation with activated manganese(IV) oxide, followed by successive treatment with p-toluenesulfonic acid, was finally found to give the desired xestoquinol dimethyl ether (+)-74 as crystals in 24% overall yield from (-)-71.

(+)-74

(12bS).(+)-4

(12bS)-62

Scheme 15. The hydroquinone dimethyl ether moiety of (-h)-74 was oxidatively cleaved with ammonium cerium(IV) nitrate ((NH4)2Ce(N03)6), yielding xestoquinone (12b5)-(+)-4 as crystals: mp 213-216°C dec; natural,^ mp 212-214X dec (Scheme 15). Xestoquinone (12b5)-(+)-4 was finally reduced with sodium hydrosulfite to afford xestoquinol (12b5)-62 in an almost quantitative yield. Although xestoquinol 62 has not been isolated yet as a natural product, the spectroscopic data of the synthetic sample reasonably support its structure. The first total synthesis of (+)xestoquinone and xestoquinol has been thus achieved. Since xestoquinone (+)-4 was converted to adociaquinones A (+)-7 and B (+)-8 by Schmitz^, our total synthesis of (+)-4 implies the formal synthesis of these hexacyclic metabolites 7 and 8.

66

15.

The CD and Absolute Stereochemistry of (+)-Xestoquinone and Xestoquinol

The CD data of the synthetic sample of (+)-xestoquinone were identical with those of the natural xestoquinone (+)-4 (Figure 13).20 Since it is evident that the absolute configuration of the synthetic xestoquinone 4 is 12b5, because we started from the Wieland-Miescher ketone (8a/?)-(-)-33, the results described above lead

Synthetic 347(+2.57)

400

X/nm

237(-13.40)

(12b5)-(+) Natural + 5-J

347(+2.56)

T 310(-2.56)

400

^^ X/nm

236(-11.98)

Figure 13. CD spectra of synthetic and natural samples of xestoquinone (12b5H+)-4inCH3CN. to the conclusion that natural xestoquinone (+)-4 also has the 12bS absolute configuration. Xestoquinone (12b5')-(-f)-4 was converted to xestoquinol 62, the hydroquinone 62 also has the 12b5 absolute configuration. The absolute stereochemistry of (+)-xestoquinone and xestoquinol was thus unambiguously determined.

67

16.

CD Spectrum and Absolute Stereochemistry of Naphthalene-Diene Derivative

As discussed above, the absolute stereochemistry of halenaquinone (+)-l and halenaquinol (+)-2 has been theoretically determined by the calculation of the CD spectra of naphthalene-diene derivatives by means of the 7C-electron SCF-CI-DV MO method. 18 To apply the same method to these xestoquinone compounds, xestoquinol dimethyl ether (+)-74 was converted to naphthalene-diene derivative 75 by reduction with sodium borohydride in the presence of cerium(III) chloride23 and methanol, followed by treatment with pyridinium p-toluenesulfonate and methanol (Scheme 16).20 The product obtained was a mixture of two stereoisomers of the methoxyl group at the 4-position, from which a single isomer 75 was isolated as

2l7.5(+37.5)

OCH,

(4e,l2bS) *—34l.0(+7.2)

I

O

vu

^339.8(25,300)

200

300

x(nm)

400

Figure 14. CD and UV spectra of xestoquinol naphthalene-diene derivative (4^,12b5)-75inCH3CN.20

68

crystals. It was difficult to purify naphthalene-diene derivative 75 because of its instability. The separation of two stereoisomers was checked by an HPLC (Nucleosil 50-5 column; hexane/EtOAc, 30:1). Although the relative stereochemistry of the methoxy group at the 4-position of 75 has remained undetermined, it is a minor problem for the present situation, because the chiroptical properties of the twisted conjugated 7C-electron system composed of naphthalene, conjugated diene, and lonepair orbitals of three ether oxygen atoms are little affected by the chirality at the 4position. In fact, the basic pattern of CD spectrum of 75 shown in Figure 14 is quite similar to those of naphthalene-diene derivatives of halenaquinol series that we have previously prepared starting from natural halenaquinol; compare Figure 14 with Figure 4.18 CH3O

CH3O

OCHo

CH3O

CH3O

(12b5:)-(+)-74

(4^,12bS)-75

Scheme 16. We have previously performed the theoretical calculation of the CD and UV spectra of a model compound 26 with 12b5 absolute configuration by application of the 7t-electron SCF-CI-DV MO methodic (Figure 6). Since the observed CD and UV spectra of xestoquinol naphthalene-diene derivative 75 (Figure 14) are in a good agreement with the theoretically obtained CD and UV curves of the model compound (12bS)-26 (Figure 6), it is evident that compound 75 has a l2hS absolute configuration. The absolute stereochemistry of compounds in the xestoquinone series was thus corroborated. 17.

Concluding Remarks

The absolute stereostructures of halenaquinone (+)-l, halenaquinol (+)-2, halenaquinol sulfate (+)-3, xestoquinone (+)-4, and xestoquinol 62, novel pentacyclic marine natural products isolated from tropical marine sponges, were theoretically determined to be 12b5, respectively, on the basis of the calculation of the CD spectra of naphthalene-diene derivatives by the 7C-electron SCF-CI-DV MO method. These studies also clarified that the theoretical CD method was appUcable to such complex natural products.

69

We have succeeded in the first total synthesis of (+)-halenaquinone 1, and (+)-halenaquinol 2, (+)-xestoquinone 4, and xestoquinol 62 by starting from the (8a/?)-(-)-Wieland-Miescher ketone 33 and also have experimentally determined their absolute configurations to be 12b5. The conclusions on the absolute configuration are in agreement with those derived from the theoretical calculation of CD spectrum. So, these first total syntheses of (+)-halenaquinol, (+)-halena-

(12bS)-(+)-l

(12bS)-(+)-4

(12bS)-(+)-2

(12bS)-62

quinone, (+)-xestoquinone, and xestoquinol provide the experimental proof of the absolute stereostructures of the halenaquinol family theoretically determined. Therefore, this methodology would become a promising tool for the determination of the absolute stereochemistry of various complex natural products with a twisted TT-electron system. Acknowledgement The theoretical determination of the absolute stereochemistry of halenaquinol and halenaquinone has been done in collaboration with Prof. Isao Kitagawa and his coworkers of Osaka University, to whom the authors thank for their contribution. Our studies described here were supported in part by grants from the Ministry of Education, Science, and Culture, Japan, the Suntory Institute of Bioorganic Research, and the Japan Association of Chemistry.

70

References 1

2 3 4 5 6

7 8

9 10 11

12 13 14 15

16 17

(a) Institute for Chemical Reaction Science, Tohoku University, 2-1-1 Katahira, Aoba, Sendai 980, Japan, (b) Pharma Research Laboratory, Hoechst Japan, Ltd., 1-3-2 Minamidai, Kawagoe, Saitama 350-11, Japan. D. M. Roll, P. J. Scheuer, G. K. Matsumoto, and J. Clardy, /. Am. Chem. Soc, 105 (1983) 6177. M. Kobayashi, N. Shimizu, Y. Kyogoku, and L Kitagawa, Chem Pharm, Bull, 33 (1985) 1305. H. Nakamura, J. Kobayashi, M. Kobayashi, Y. Ohizumi, and Y. Hirata, Chem. Le/r., (1985)713. F. J. Schmitz and S. J. Bloor, J. Org. Chem., 53 (1988) 3922. J. M. Bijvoet, A. F. Peerdeman, and A. J. Van Bommel, Nature, 168 (1951) 271. J. M. Bijvoet, A. F. Peerdeman, and A. J. Van Bommel, Proc. K. Ned. Acad. Wet. B 54 (1951) 16. A. J. Van Bommel, Proc. K. Ned. Acad. Wet. B 56 (1953) 268. J. Trommel and J. M. Bijvoet, Acta Crystallogr., 7 (1954) 703. J. M. Bijvoet, and A. F. Peerdeman, Acta Crystallogr., 9 (1956) 1012. See also J. Ibers and W. C. Hamilton, Acta Crystallogr., 17 (1964) 781. N. Harada and K. Nakanishi, Circular Dichroic Spectroscopy - Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, California, and Oxford University Press, Oxford 1983. A. Moscowitz, Tetrahedron 13 (1963) 48. C. Kemp and S. F. Mason, Tetrahedron 22 (1966) 629. A. Brown, C. Kemp, and S. F. Mason, J. Chem. Soc. A (1971) 751. N. Harada, Y. Tamai, Y. Takuma, and H. Uda, J. Am. Chem. Soc, 102 (1980) 501. N. Harada, Y. Tamai, and H. Uda, J. Am. Chem. Soc, 102 (1980) 506. N. Harada, J. Iwabuchi, Y. Yokota, and H. Uda, Croat. Chem. Acta, 62 (1989) 267. C. Rosini, C. Bertucci, P. Salvadori, and M. Zandomeneghi, /. Am. Chem. Soc, 107 (1985) 17. See also: J. Roschster, U. Berg, M. Pierrot, and J. Sandstrom, /. Am. Chem. Soc, 109 (1987) 492. N. Harada, J. Kohori, H. Uda, K. Nakanishi, and R. Takeda, /. Am. Chem. Soc, 107 (1985) 423. N. Harada, J. Kohori, H. Uda, and K. Toriumi, / . Org. Chem., 54 (1989) 1820. N. Harada, H. Uda, T. Nozoe, Y. Okamoto, H. Wakabayashi, and S. Ishikawa, / . Am. Chem. Soc, 109 (1987) 1661. N. Harada, H. Ono, H, Uda, M. Parveen, N. U.-D. Khan, B. Achari, and P. K. Dutta, /. Am. Chem. Soc, 114 (1992) 7687.

71

18

19 20 21 22 23 24 25 26

27 28

29 30 31

32 33

34 35 36 37

M. Kobayashi, N. Shimizu, I. Kitagawa, Y. Kyogoku, N. Harada, and H. Uda, Tetrahedron Lett,, 26 (1985) 3833. N. Harada, H. Uda, M. Kobayashi, N. Shimizu, and I. Kitagawa, 7. Am, Chem, Soc, 111 (1989) 5668. N. Harada, T. Sugioka, Y. Ando, H. Uda, and T. Kuriki, J, Am. Chem. Soc, 110 (1988) 8483. N. Harada, T. Sugioka, H. Uda, and T. Kuriki, / . Org. Chem., 55 (1990) 3158. N. Harada, K. Nakanishi, and S. Tatsuoka, J. Am. Chem. Soc, 91 (1969) 5896. N. Harada and K. Nakanishi, Ace. Chem. Res., 5 (1972) 257. A. L. Gemal and J. L. Luche, / . Am. Chem. Soc, 103 (1981) 5454. N. L. AUinger, /. Am. Chem. Soc, 99 (1977) 8127. N. L. Allinger and Y. H. Yuh, eCP£, 72 (1980) 395. G. Stork and P. D'Angelo, J. Am. Chem. Soc, 96 (1974) 7114. Y. Li, B. Nassim, and P. Crabbe, /. Chem. Soc, Perkin Trans. 1, (1983) 2349. M. Ihara, M. Toyota, K. Fukumoto, and T. Kametani, / . Chem. Soc, Perkin Trans. 7, (1986) 2151. M. Kim, R. S. Gross, H. Sevestre, N. K. Dumlap, and D. S. Watt, / . Org. Chem., 53 (1988) 93. See also (+)-l,8a-dihydroazulene derivatives: ref. 13. V. Prelog and W. Acklin, Helv. Chim. Acta, 39 (1956) 748. G. R. Newkome, L. C. Roach, R. C. Montelaro, and R. K. Hill, / . Org. Chem., 37 (1972) 2098. F. Toda and K. Tanaka, Tetrahedron Lett., 29 (1988) 551. Z. G. Hajos and , D. R. Parrish, /. Org. Chem., 39 (1974) 1615. U. Eder, G. Sauer, and R. Wiechert, Angew. Chem., 83 (1971) 492; Angew. Chem. Int. Ed. Engl., 10 (1971) 496. J. Gutzwiller, P. Buchschacher, and A. Furst, Synthesis (1977) 167. P. Buchschacher and A. Furst, Org. Synth., 63 (1986) 37. P. Buchschacher, A. Furst, and J. Gutzwiller, Organic Syntheses; Wiley: New York, Collect. Vol. 7, page 368 (1990). N. Harada, T. Sugioka, H. Uda, and T. Kuriki, Synthesis, (1990) 53. T. Mukaiyama, M. Usui, E. Shimada, and K. Saigo, Chem. Lett., (1975) 1045. T. Mukaiyama, Angew. Chem., 91 (1979) 798; Angew. Chem. Int. Ed. Engl., 18 (1979) 707. Y. Morimoto, K. Oda, H. Shirahama, T. Matsumoto, and S. Omura, Chem. Lett., (1988) 909. V. Prelog and W. Acklin, Helv. Chim. Acta, 39 (1956) 748. N. Harada, T. Sugioka, H. Uda, and T. Kuriki, Coll. Czech. Chem. Commun., 57 (1992) 1459. G. Bauduin and Y. Pietrasanta, Tetrahedron, 29 (1973) 4225. See also: J. E. McMurry, / . Am. Chem. Soc, 90 (1968) 6821.

72

38

39 40 41

42 43 44

B. M. Trost, Y. Nishimura, K. Yamamoto, and S. McElvain, 7. Am, Chem. Soc. 101 (1979) 1328. J. E. McMurry, A. Andrus, G. M. Ksander, J. H. Musser, and M. A. Johnson, / . Am, Chem, Soc, 101 (1979) 1330. W. G. Salmond, M. A. Barta, and J. L. Havens, /. Org, Chem., 43 (1978) 2057. M. Oda and Y. Kanao, Chem, Lett,, (1981) 37. D. McHale, P. Mamalis, S. Marcinkiewicz, and J. Green, /. Chem, Soc, (1959) 3358. See also: J. G. Nilsson, H. Sievertsson, and H. Selander, Acta Pharm, Suec, 5 (1968) 213. L. I. Smith and F. L. Austin, /. Am, Chem, Soc, 64 (1942) 528. D. H. R. Barton and W. B. Motherwell, Pure Appl. Chem,, 53 (1981) 15. W. G. Salmond, M. A. Barta, and J. L. Havens, J, Org. Chem., 43 (1978) 2057. W. Kreiser and W. Ulrich, Liebigs Ann. Chem., 761 (1972) 121.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

73

Bryozoan Secondary Metabolites and their Chemical Ecology A.J. Blackman and J.T. Walls

1. 1.1.

Introduction General

Bryozoans are sedentary, colonial invertebrates that are widely distributed throughout the marine environment, b u t are l e s s common in freshwater. There are about 4 0 0 0 living s p e c i e s and over 1 0 0 0 0 as preserved fossils. Marine colonies are generally found on rocky shores in the shallow sublittoral zone although they also occur in the ocean depths [11. Colonies are often small (typically l e s s than 5 cm), incons p i c u o u s and infrequent. For t h e s e r e a s o n s and b e c a u s e of the taxonomic problems experienced when working with this phylum, bryozoans are frequently overlooked or ignored. Bryozoan morphology shows great variation with colony form often being related to habitat type. A bryozoan colony is made up of a collection of intercommunicating individuals, zooids, w h i c h are usually l e s s than one millimetre in length. Zooids can specialise (heterozooids) and take on specific roles within a colony, for example, feeding (autozooids), brood or funicular (transport or communication) zooids. A generalised zooid c o n s i s t s of a ring of ciliated oral tentacles which are involved in feeding, respiration and excretion. All bryozoans are voracious filter feeders and therefore need an abundant supply of microplankton to survive. The gut c o n s i s t s of a differentiated U-shaped stomach which terminates near the ring of t e n t a c l e s . The supportive outer body wall is often reinforced by chitin or calcium carbonate or both. A funicular s y s t e m allows communication between adjacent and more widely separated zooids. Bryozoans produce larvae by sexual reproduction which later form a new colony by asexual budding. Brooding of larvae takes place in an enlarged brood chamber, or gonozooid [2]. Bryozoans are of particular interest to researchers for two main reasons: the variety of secondary metabolites that they contain, and b e c a u s e they are common marine fouling o r g a n i s m s . The interesting chemistry they p o s s e s s , combined with a distinct paucity in ecological work on bryozoan secondary metabolites, m e a n s that this area still

74

offers wide scope for investigation. Research in this field comes from two directions. Firstly, the diverse chemistry of many bryozoans has led to a detailed catalogue of their chemical makeup, but lack of complementary biological work has precluded in«depth ecological interpretations. Secondly, the majority of ecological work concerning marine bryozoans has centred around settlement cues and inhibitors of bryozoan larvae, driven undoubtedly by the monetary aspects of antifouling work. Our review covers the chemistry of bryozoan secondary metabolites and direct chemical ecological interactions. Both the documented effects of bryozoan secondary metabolites and the role of adult and larval bryozoans in testing metabolites from other organisms will be discussed. A wide gap in biological knowledge exists and we hope to outline relevant work in this area and highlight directions for subsequent experiments required to bridge this deficiency. 1.2.

Bryozoan taxonomy In most modern works the phylum Bryozoa is defined to exclude the entoprocts and is divided into three major divisions or classes [3). Phylum

Class

Order

- Phylactolaemata (freshwater only)

Bryozoa•

-Stenolaemata

•Gymnolaemata(maln group)

2. 2.1.

-Cyclostomata —Ctenostomata (uncalcifled zooids) —Chellostomata (calcified zooids)

Bryozoan secondary metabolites

Introduction Bryozoans contain a wide variety of secondary metabolites with structural types varying from simple one and two carbon containing compounds to complex macrocycles and alkaloids. The first and only comprehensive survey of bryozoan secondary metabolites was published almost a decade ago [4]. Since that time more specialised aspects of bryozoan chemistry have been reviewed in their own right or have been included in more general reviews. Specialised reviews include the chemistry and biological properties of the

75

bryostatins [5] and the chemotaxonomical challenge and biosynthesis of bryozoan secondary metabolites [6]. Two reviews of marine alkaloids (7-81 and another of marine alkaloids and related compounds 191 have included bryozoan metabolites. The continuing annual reviews of marine natural products includes bryozoan compounds [10]. This section will detail the structures and chemistry of bryozoan secondary metabolites which have been described since the first comprehensive review [4]. Primary metabolites will not be included. Since the distinction between primary and secondary metabolites is rather blurred, some of the compounds included could also be considered to belong to the former rather than to the latter group while other metabolites may appear to have been arbitrarily excluded. Compounds will be presented according to their chemical structures with two main categories being used depending on whether nitrogen is present in the molecule or not. Where appropriate, cross references will be given to different compound types which originate from the same bryozoan.

2.2.

Non-alkaloids

2.2.1 Macrocyclic lactones An important group of secondary metabolites from bryozoans consists of the macrocyclic lactones which have been obtained from Bugula neritina. These compounds have potential as leads for future anticancer drugs. Bryostatins 1-15 are bryopyran lactones, 1-15, which have very selective antineoplastic and cytostatic activity. Bryostatin 1, 1, is undergoing clinical evaluation. Chemical components of B. neritina vary somewhat depending on the area of its collection. The isolation, structural elucidation and biological activity of bryostatins 1-13 have been reviewed 15]. Another bryozoan, Amathia convoluta, has been found which also gives bryostatins 4-6, 3-S, and 8, 7. Examination of voucher specimens showed that some B. neritina occurred on the A. convoluta as a parasitic or epiphytic growth so that the true source of the compounds remains ambiguous although the authors concluded that bryostatin 8 was a genuine constituent of A. convoluta [11]. Investigation of a Tasmanian collection of A. convoluta [12] failed to show any antineoplastic activity which is characteristic of the bryostatins; the bryozoan did however contain a new alkaloid (see Section 2.3.1). Two additional members of this series have been reported recently from Bugula neritina; bryostatin 14. 13, was obtained in 1.02 x 10'^%

76 compound

CHgO

>Ri '•

•

H

O^^

1 2 3 4 5 6 7 8 9 10 11 12 13 14

R B B D A A A D D H H B H OH E

Ri A OH C C D A D A C A D D C A

bryostatin 1 bryostatin 2 bryostatin 4 bryostatin 5 bryostatin 6 bryostatin 7 bryostatin 8 bryostatin 9 bryostatin 10 bryostatin 11 bryostatin 12 bryostatin 13 bryostatin 14 bryostatin 15

OCH3

X.A "- /^o A ^

0=

A

77

yield from a Gulf of Mexico collection while specimens from the Eastern Pacific gave bryostatin 15. 14. in 8.6 x 10"^% yield [131. The structure previously ascribed to bryostatin 3. IB, has been recently revised [141 and then further corrected by members of the same group [151. A stereochemical error was found and the compound should be referred to as 20-epC-bryostatin 3 having structure 16. The authors have also isolated two new compounds. 17. which they have named as bryostatin 3 [151. and bryostatin 3 26-ketone. 18, [141. The synthesis of bryostatin 7, 6, has been reported [161.

16 epibryostatin 3

17 bryostatin 3

Further bioassay-guided examination of a Gulf of Mexico collection of Bugula neritina has yielded neristatin 1, 19. the first bryostatin biosynthetic precursor or degradation product to be reported. Neristatin 1. whose structure was determined by a combination of spectroscopic and X-ray diffraction data, is antineoplastic b u t is not as active as the bryostatins [171. B. neritina also contains fattty acid derivatives (Section 2.2.2 and alkaloids (Sections 2.3.4 and 2.3.5). 2.2.2 Sterols, terpenes and fatty acid derivatives Most of the bryozoan secondary metabolites in this category are not novel, having been isolated previously from other natural sources. Myriapora truncata, collected in the Mediterranean, has been shown to contain five sterols. cholest-7-en-3p.5a.6p-triol. 20. (22E. 24S)-24-methylcholesta-7.22-dien-3p.5a.6p-triol. 2 1 . (22E)-cholesta-

78

CH3O

CH3O

H 19 neristatin 1

18 bryostatin 3 26-ketone

7,22-dien-3p.5a,6p-triol. 22. (22E,24K)-24-methylcholesta-7,22-dien3p.5a.6p-triol, 2 3 . and (22E.24$)-24-ethylcholesta-7.22-dien-3p.5a.6ptriol. 24. the last three of which are novel. The sterols were isolated as diacetyl derivatives after acetylation and separation by reverse phase hplc [181. One of these sterols, 24. has been synthesised 119].

21

20

23

22

24

79

Examination of Cribricellina cribraria from New Zealand has revealed that it contained a range of sterols as well as alkaloids (see Section 2.3.4). While the authors stated that twelve of the sterols were identified, only nine were actually named in the publication:- cholesterol, 25. 24-methylcholesta-5.22-dien-3p-ol, 26. cholestanol. 27. cholest-4-en-3-one. 28, cholesta-5.22-dien-3P-ol. 29, 24-ethylcholesterol. 30. 24-methylenecholesterol. 3 1 . 24-ethylcholesta-5.22-dien3p-ol, 32. 24-methyl-26.27-bisnorcholesta-5.22-dien-3p-ol, 3 3 . 120).

25

26

27

28

29

30

31

32

33

Conopeum seuratum, collected in the Black Sea where it is the most common bryozoan, has been analysed by gas chromatographymass spectrometry. Although over one hundred compounds were detected only a small fraction was identified. Those identified were all known compounds although most were reported for the first time from a bryozoan. Monoterpene alcohols found, either free or as esters, were citronellol, 34, cuminol, 35 and its isomer 36. p-menthan-4-ol, 37, menth-8-en-lO-ol. 38. geraniol, 39. and nerol. 40. The monoterpenes 1,8-cineole, 4 1 . piperitone. 42, pulegone, 4 3 , isopinocamphone, 44,

80

CHzOH •"CHzOH

XH2OH 34 citroneliol

35 cuminol

36

38 menth-8-enIO-0I

37 p-menthan4-0I

.CH2OH CH2OH

-

^ 39 nerol

40 geraniol

^

41 1,8-cineole

42 piperitone

43 pulegone

CHO CHO

CO2H 44 isopinocamphone

45 mirtenal

0$^

46 perillaldehyde

47 dehydro abietatic acid

CO2H

/
49 ursolic acid

50

51 ethyl abietate

mirtenal 4 5 , and perillaldehyde. 4 6 were also found as were the diterp e n e s dehydroabietic acid. 4 7 , and neoabietic acid, 4 8 , (as either the free acid or methyl or propyl esters) and the triterpene ursolic acid, 4 9 . One non-terpene. l . l - d i m e t h y l - 2 - p h e n y l e t h a n o l , 5 0 , was identified [21], The same authors also found ethyl dehydroabietate, 5 1 , from Bugula turrita [21]. Terpenes have been reported previously from Flustra foliacea, where they are largely responsible for the bryozoan*s lemon-like odour [22]. F. foliacea also contains alkaloids (Section 2.3.2). Eight bryozoan species collected from the Sea of J a p a n have been examined for fatty aldehydes. Lipids were extracted and methylated

81 and the aldehydes determined as their dlmethylacetals by capillary column gas-liquid chromatography. Amathia sp., Bugula neritina» Cauloramphus spiniferum, Celleporina irregularis, Exochella areolata okadata and mawatari, Microporella ciliata, Watersiporia aterrima and Schizoporella unicornis contained hexadecanal, 52, heptadecanal, 5 3 . and octadecanal, 54. The taxonomic significance of these findings was also discussed [23]. XHO 52

53

54

2.2.3 Halogen- and sulfur-containing compounds While most bryozoan compounds which contain halogen or sulfur are alkaloids (Section 2.3), several having no nitrogen have been discovered. These compounds have been implicated in various offflavours and bad odours. 2,6-Dibromophenol, 55, has been identified as the cause of an iodoform-like off-flavour in the commercially important Australian prawns Metapenaeus endeavouri and Penaeus plebejus [24]. Lesser amounts of several other bromophenols were also detected in these and several other species of prawns. In an attempt to discover the source of the 2,6-dibromophenol, a selection of eight marine algae, two bryozoans, a hydroid and eight sponges were collected from Exmouth Gulf, Western Australia, the region giving rise to the tainted prawns. Examination of these specimens by gas chromatography-mass spectrometry showed that they all contained bromophenols; both bryozoans, which were not further identified, contained the highest concentrations of 2.6-dibromophenol. Although adult penaeid prawns are thought to be carnivores, the authors concluded that it was unlikely that marine algae, bryozoans, hydroids or sponges contributed greatly to the 2,6-dibromophenol content of the tainted prawns 125]. The malodorous Bijlustra perfragilis which occurs in Bass Strait, the body of water which separates Tasmania from mainland Australia, has been found to give rise to a series of volatile sulfur and halogen compounds,56-61, [26] as well as alkaloids (see Section 2.3.4). Local fishermen are familiar with this bryozoan and its smell since they occasionally encounter it when dredging for scallops. The region where

82 (CH3)2S2

(CH3)2S

CH3SH

56

57

58

CH3OH

CH2CI2

CH3CI

60

61

55

59

t h i s h a p p e n s is colloquially called **shit ground**. The foul smell of t h i s b r y o z o a n w a s e s t a b l i s h e d a s b e i n g d u e to d i m e t h y l disulfide, 5 6 . dimethyl sulfide, 5 7 , a n d m e t h a n e t h i o l , 5 8 . M e t h a n o l , 5 9 , d i c h l o r o m e t h a n e . 6 0 , a n d c h l o r o m e t h a n e , 6 1 , were p r e s e n t a l s o . The relative p r o p o r t i o n s of t h e s e volatile c o m p o u n d s were found to vary at different collection s i t e s a n d over t h e t h r e e - d a y period of t h e a n a l y s i s . The a u t h o r s c o n s i d e r e d it likely t h a t t h e s e volatile c o m p o u n d s h a d a microbial origin.

2.3.

Alkaloids and related compounds

Most b r y o z o a n s e c o n d a r y m e t a b o l i t e s are a l k a l o i d s a n d t h e large majority of t h e m h a v e novel s t r u c t u r e s . Heterocyclic ring s y s t e m s w h i c h a r e r e p r e s e n t e d i n c l u d e p y r r o l e , p y r i d i n e , indole, i s o q u i n o l i n e . p u r i n e a n d p - c a r b o l i n e a s well a s n u m e r o u s new polycyclic s y s t e m s . 2.3.1 p-Phenylethylamine-related alkaloids The Amathia g e n u s h a s a wide g e o g r a p h i c a l d i s t r i b u t i o n a n d its m e m b e r s from different o c e a n s h a v e given a relatively large n u m b e r of a l k a l o i d s w h i c h c o n t a i n t h e p - p h e n y l e t h y l a m i n e s t r u c t u r a l u n i t . The c o m p o u n d s a r e all b r o m i n e - c o n t a i n i n g a m i d e s . A. ivilsoni is p r o b a b l y t h e l a r g e s t a n d m o s t c o m m o n T a s m a n i a n b r y o z o a n . It h a s given rise to a m a t h a m i d e s A-F a s well as a likely b i o s y n t h e t i c p r e c u r s o r . A m a t h a m i d e s A-F, 6 2 - 6 7 . are a s e r i e s of b r o m i n a t e d p - p h e n y l e t h y l a m i n e a n d proline r e l a t e d a l k a l o i d s which differ from e a c h o t h e r by t h e degree of b r o m i n a t i o n or m e t h y l a t i o n a n d by t h e p r e s e n c e of or s t e r e o c h e m i s t r y a b o u t a d o u b l e b o n d [27, 28). A n o t h e r , m o r e r e c e n t , i n v e s t i g a t i o n of A. wilsoni confirmed t h e p r e s ence of t h e a m a t h a m i d e s a n d discovered a likely a m a t h a m i d e b i o s y n thetic precursor 2-(2,4-dibromo-5-methoxyphenyl)ethanamine, 68, [291. The p r o p o r t i o n of t h e total alkaloid a n d t h e relative a m o u n t s of a m a t h a m i d e s from different colonies a t t h e s a m e location were e s s e n tially i d e n t i c a l a n d did n o t show a n y s e a s o n a l v a r i a t i o n . Collections from different localities however did exhibit c o n s i d e r a b l e v a r i a t i o n in b o t h t o t a l alkaloid c o n t e n t a n d w h i c h a m a t h a m i d e s were p r e s e n t [28]. The d i s t r i b u t i o n of t h e a m a t h a m i d e s w i t h i n single colonies h a s also

83

Wa OChfe

OCHJ

Br 62 amathamideA

63 amathamide B

6HJ

64 amathamide C

CKfe 00%

OCH3

OChfe

65 amathamide D

66 amathamide E

67 amathamide F

OCH3

68 69 amathamide G b e e n examined; the alkaloid content varied from nearly 9% dry weight in the outermost tips of the colony to 0% at the b a s e I30J. The ecological significance of t h e s e findings is d i s c u s s e d in Section 3 . Two other Amathia species which occur in Tasmania have been found to also contain amathamides. A new member of the series, amathamide G, 6 9 , occurs in A. convoluta while A. pinnata contains amathamide C. 6 4 , which h a s been previously found in A. wilsoni 131 J. Amathamide G p o s s e s s e s a fully s u b s t i t u t e d aromatic ring and its structure w a s secured, in part, by chemical degradation to yield a known compound. The chemistry of A. convoluta collected from other regions of the world h a s also b e e n described and provides yet another example of geographical variability of secondary metabolites. The possible occurrence of bryostatins in A. convoluta from the Gulf of Mexico, h a s been previously d i s c u s s e d in Section 2 . 2 . 1 . Since this investigation w a s directed towards discovering antineoplastic c o n s t i t u e n t s , the nonactive amathamides would not have been detected. In another recent examination of A. convoluta for secondary metabolites, a series of p-phenylethylamine related alkaloids h a s b e e n found [32]. The

84 CH3 N(CH3)2

CHaO^^^x^^^s^^N

S ^

Br

70

CH3

71

H

NHCH3

CH3

72

H

N(Cli3)2

73

CH3

H

N(Chb)2

compounds, 70-74. are all bromine-containing amides having considerable similarity to the amathamides. Alkaloid 74 possesses the opposite absolute stereochemistry to that of 70. 72. 7 3 and the natural amino acids. This unusual result was determined by an X-ray crystallograpic study and confirmed by a circular dichroism spectrum. The same authors [32] also found that Amathia alternata contained four brominated alkaloids. One. 7 5 . is closely related to one of the metabolites, 74. from A. convoluta, while the others are indoles (see Section 2.3.2). All of the Amathia brominated amides are presumably biosynthesised from amino acids by similar pathways in the related bryozoans. The amathamides are amides derived formally by reaction of 2-(2.4dibromo-5-methoxyphenyl)ethanamine. 68, with proline followed variously by introduction of a double bond, or methyl, methoxy or bromine substituents. The A. convoluta metabolites 70-74, and 75 from A. alternata are all also derived formally from 2-(2.4-dibromo-5-methoxyphenyl)ethanamine. 68—either by direct amide formation with tyrosine, or by having an additional aminopropyl group which is then

85 CHj

H

N(CH3)2

OCH3

H

Ch^

75

'H

o

76

H 77

78

acetylated (71) or acylated by leucine (72 and 73), ent-tyrosine (74) or tyrosine (75)—followed by additional methylation and sometimes bromination. The presence of a methoxy substituent in the meta position of the phenylethylamine moiety in all these compounds is an indication that this part of the alkaloids is derived from phenylalanine rather than tyrosine. 2.3.2 Indole alkaloids Several of the alkaloids from bryozoans can be considered to be indoles—Amathia alternata and Zoobotryon verticillatum contain simple indoles while Chartella papyracea, Flustra foliacea and Hincksinoflustra denticulata give rise to more complex polycyclic compounds. As noted above Amathia alternata contains three brominated indoles, 76-78 as well as the p-phenylethylamine derivative 75 132]. The three indole derivatives from A. alternata are clearly related to the other Amathia compounds: they are amides from indolylethylamine (rather than phenylethylamine) and the amino acid valine—once more with additional methylation and bromination. In this series indolylethylamine is likely to be formed from typtophan in much the same way that phenylethylamine is derived from phenylalanine in the other Am.athia compounds. In an earlier investigation of Zoobotryon verticillatum obtained at San Diego, California, a brominated indole, 7 9 , and its side-chain N-oxide, 80. were found [33]. The same bryozoan is very abundant on

86

CHO

N(CH3)2

CH3

79

CH3

80

81

the southern Atlantic coast of Spain to the extent that it is a common fouling organism in harbours and fish farms. Collections from two different locations in this region have yielded a new brominated indole. 8 1 , as well as 7 9 and 8 0 [34]. The first compound to be isolated from Chartella papyracea (also known as Flustra papyracea) was chartelline A, 8 2 . The structure, a penta-halogenated indole containing a p-lactam ring, w a s secured with the a s s i s t a n c e of X-ray diffraction data [35]. A second collection of the bryozoan from the same location. Roscoff. France, some four years later gave four additional alkaloids, chartellines B. 8 3 . and C. 8 4 . [36]. and chartellamides A. 8 5 . and B. 8 6 . [37]. as well as more chartelline A. Spectroscopic data showed that chartellines B and C had identical skeletons with that of chartelline A and differed by the number of bromine s u b s t i t u e n t s in the benzene ring. Chartellamides A and B are both brominated polycyclic alkaloids with a different, although related skeleton, to that of the chartellines [37]. Another compound, methoxy-

82 chartelline A

83 chartelline B

86 chartellamide B

84 chartelline C

87

87 dechlorochartelline A, 87, was also isolated but shown to be an artefact since it could be prepared by reacting chartelline A with methanol, the solvent used during the extraction [361. Chartelline A as well as the crude mixture of alkaloids was devoid of any significant antimicrobial activity and chartelline A was inactive in the NCI's leukemia test [36J. Prior to the period covered by this review, Flustra foliacea, obtained from Scandinavian waters, had been found to yield ten alkaloids—nine containing the indole ring system (as either physostigmine or tryptamine derivatives) with the remaining one being a quinoline [4]. More recently F. foliacea collected from the same region has been shown to also contain flustramide B. SB. and flustrarine B, 89, which were related to the previously described flustramide A and flustramine

88 flustamide B

89 flustrarine B

90 dihydroflustramine C

H-^\i3

91 flustamine D

92 dihydroflustramine 0 A^oxide

94 isoflustramine D

93 flustramine D A^-oxide

B. Flustrarine B was also synthesised by oxidation of flustramine B with hydrogen peroxide 138]. F. foliacea also occurs in the Bay of Fundy, Canada. Bioassayguided fractionation based on antibacterial activity of the extract of this bryozoan has lead to the isolation of a different series of bromoalkaloids of the physostigmine class, two of which were N-oxides [39]. The major alkaloids were dihydroflustramine C. 90, (-60%), which had been reported previously by these authors [40J and flustramine D, 9 1 . (-40%). Minor alkaloids found were dihydroflustramine C N-oxide, 92, flustramine D N-oxide, 93, and isoflutramine D. 94. The N-oxides were also synthesised by oxidation of the corresponding free bases with m-chloroperbenzoic acid. These alkaloids were shown to be responsible for the wide spectrum of antibacterial activity exhibited by this bryozoan extract. Hincksinoflustra denticulata, which occurs in Tasmania has been found to contain hinckdentine A, a novel pentacyclic tribromo alkaloid [411. The structure and absolute configuration was determined by single crystal X-ray methods. Q

95 hinckdentine A

2.3.3 Pyrrole alkaloids Bryozoans have been found to give rise to dipyrroles and tetrapyrroles. Tambjamines A-D, 96-99, are dipyrroles which occur in Sessibugula translucens. Although these alkaloids were reported in the previous review of bryozoan metabolites [4], full details of their ecological significance, which have been subsequently published were not [42J. The chemical ecology of the tambjamines are discussed in Sections 3.1. and 3.4. Bugula dentata is a dark blue coloured bryozoan that is common in J a p a n . The colour has been ascribed to a tetrapyrrole pigment, 100, which also occurs in a bacterium and an ascidian [43]. Several alkaloids, the amathamides, which contain the pyrrolidine ring were described in Section 2.3.1.

89 PCH3

pCHa

// \>

PCH3 // \>

// \> Br NH2

NhkX^

NH2

96 tambjamineA

98 tambjamineC

97 tambjamineB Br

CH3O

OCH3

OCH3

// *

// \>

H

y \ w / w cr H

99 tambjamine D

i

I

H

H

100

2.3.4 Pyridine, purine, isoquinoline and p-carboline alkaloids Several bryozoans contain quaternary pyridine compounds. The widely distributed homarine. 101. h a s been found in a mixture of three Cellaria species (C. salicornioides, C. fistulosa and C. sinusoa), Chartella papyracea, Bugula neritina |441 and Cribricellina cribraria [201. Cellaria spp. and C. papyracea also contained the isomeric trigonelline, 102, and other quaternary compounds, see Section 2.3.5. The Cellaria spp. mixture also yielded the three nucleosides 2'-deoxyguanosine, 103, 2'-deoxyadenosine, 104, and inosine, 105, 1441.

a.-

CHa

CHa

101 homarine

102 trigonelline

NH2

HO 103 2'-deoxyguanosine

HO OH 104 2'-deoxyadenosine

105 inosine

In an earlier report phidolopin, 106, a purine which contains the rare nitro functional group, was isolated from Phidolopora pacifica [451. Further investigation of this and four other bryozoans from the

90

NO2

HcxA CH3

106 phidolopin

CH3

107 desmethylphidolopin

^^-^CH20H

108

Northeast Pacific led to additional nitro compounds. Diaperoecia callfornica and P. pacifica contained phidolopin, 106. and desmethylphidolopin, 107, and 3-nitro-4-hydroxybenzyl alcohol. 108. Examination of Heteropora alaskensis, Tricellaria ternata and Hippodiplosia insculpta showed that these three bryozoans contained 3-nitro-4-hydroxybenzyl alcohol, 108. or 3-nitro-4-methoxybenzyl alcohol or both with the latter compound being formed as an artefact. The authors suggested that their failure to detect phidolopin or desmethylphidolopin in extracts of H. alaskensis, T. ternata and H. insculpta was due to the small amounts of bryozoans that could be collected. The isolation of these nitrophenol compounds from five bryozoans. all of which occur in the same habitat but belong to different genera, was interpreted as indicating that the bryozoans obtained the compounds from a dietary or symbiotic micro-organism such as phytoplankton. bacterium or fungus [461. A total synthesis of phidolopin has been reported [47], Costaticella hastata is a common Tasmanian bryozoan; it has been reported as containing a range of 1-substituted p-carbolines. Harman (1-methyl-p-carboline), 109. was shown to be the major compound with lesser amounts of 1-ethyl-p-carboline. 110, (S)-l-(l-hydroxyethyl)P-carboline. 111. and pavettine (1-vinyl-p-carboline). 112. Of these four compounds only (S)-l-(l-hydroxy-ethyl)-P-carboline. 109, was novel. Some geographical variation of the minor alkaloids was found [48]. Re-examination of the voucher specimens has subsequently shown that the collections of C. hasatata contained significant amounts of Orthoscuticella ventricosa, a bryozoan which can not be readily distinguished from C. hastata without the aid of a microscope. Analysis of O. ventricosa revealed that it also contained the same p-carbolines 109-112 [491. The closely related New Zealand bryozoan Cribricellina cribraria has also been found to contain p-carboline alkaloids in addition to steroids (Section 2.2.2) and homarine [20]. Bioassay guided analysis led to the isolation of 113 as the major cytotoxic component. C. cribra-

91

QrQ^ H 109 harman

112

Q^N H 110

113

^CH" H

^

111 pavettine

114

ria also contained a second novel compound, 114, a sulfone, and the known alkaloids harman, 109. 1-ethyl-p-carboline. 110, and pavettine, 111, as minor components. A synthesis of pavettine using a palladiumcatalysed coupling reaction has been described [50]. Biflustra perfragilis, in addition to producing volatile compounds (Section 2.2.3) h a s also given rise to sulfur-containing isoquinolines. A collection of B. perfragilis from Bass Strait yielded the two isoquinolinetriones 115 and 116 [51]. The structure of 115 was secured by a single-crystal X-ray study.

115

116

117

A marine chemistry seminar report made brief reference to an otherwise unpublished investigation of Membranipora perfragilis collected from South Australia. This bryozoan was found to contain 117, another isoquinolinetrione [52]. Membranipora perfragilis is in fact Biflustra perfragilis [53] and the isolation of different compounds from the same species is yet another example of the geographical variation of bryozoan metabolites.

92 2.3.5 Miscellaneous nitrogen-containing compounds In addition to quaternary pyridine compounds homarine, 101. and trigonelline, 102. (Section 2.3.4), Cellaria spp. and Chartella papyracea also contained betaine, 115, taurine, 116, and tetramethylammonium ion, 117. Betaine. 115. was also found in Bugula neritina. These compounds occur in other marine invertebrates as well [44). As also mentioned in Section 2.3.4. Phidolopora pacifica, Diaperoecia californica, Heteropora alaskensis, Tricellaria ternata and Hippodiplosia insculpta contained 3-nitro-4-hydroxybenzyl alcohol. 108. or the artefact 3-nitro-4-methoxybenzyl alcohol or both [461. (CHfe)3i^co2" 118 betaine

2.4.

Hjir^coz119 taurine

(CH3)4N 120

Origin of secondary metabolites

There are three main possible sources for the secondary metabolites that are found in bryozoans. They could be: • synthesised by the animals themselves • obtained from the bryozoans* diet and sequestered • synthesised by symbionts and then possibly passed to the bryozoan and sequestered For the last two possibilities, the compounds so obtained could be further modified by the bryozoan. In a recent review which addressed the problem of the origin of secondary metabolites. Anthoni et al, proposed the hypothesis that many of the compounds obtained from bryozoans are actually synthesised by associated flora or microorganisms [6]. In our unpublished work the sites of production or storage (or both) of the amathamide alkaloids found in Amathia wilsoni were studied. Since the amathamides are brominated compounds, bromine was used as a marker to map distribution within colonies. Energy dispersive X-ray analysis coupled with scanning electron microscopy was used to localise bromine in sections of A. wilsonL Bromine concentrations higher than background were found only on the surface of the bryozoan and not within any of the different internal cell types. The correlation between bromine levels and a rod-shaped bacterium.

93 ubiquitous in the tip region, indicated that the a m a t h a m i d e s are closely associated with this particular bacterium. Evidence of this type does not however prove that bacteria are responsible for the production of the a m a t h a m i d e s . The limits of the above method m u s t be carefully considered before any c o n c l u s i o n s can be made. To detect halogens u s i n g this method requires that sites m u s t contain a moderately intense concentration. Low levels of bromine are undetectable as the signal blends in with the background. It is therefore quite plausible that the a m a t h a m i d e s originate from the bryozoan itself only to be concentrated at the surface. Definite proof linking the production of secondary metabolites with associated microorganisms will only come from obtaining the c o m p o u n d s directly from the isolated microorganisms grown in culture.

3.

Chemical Ecology

3.1.

Introduction

Many ecological interactions have been s h o w n to be mediated by secondary metabolites. In bryozoans t h e s e c h e m i c a l s probably play a vital role in areas as diverse as antifouling. sequestration and antipredation. They may regulate bacterial biofilms and therefore influence settling larvae, reduce levels of predation by both invertebrates and vertebrates and possibly be taken up and stored by specialist predators. In this part of the review we will p r e s e n t an overview of the areas in which bryozoan secondary metabolites are probably involved in ecological interactions. Likely productive topics for future research will be highlighted also.

3.2.

Feeding deterrence

Both chemical defences and morphological features have been shown to reduce levels of predation. Taxonomic work h a s led to the description of possible physical defences in the Bryozoa. Skeletal material, c o n s i s t i n g of calcium carbonate or chitin or both, often constitute a large proportion of each zooid. This material may act to directly deter predators or to reduce the nutritional value of the bryozoan making it an unrewarding prey item. More active physical defences have been documented [54, 551. Membranipora membranacea can rapidly produce large, chitinous s p i n e s within 3 6 hours of attack by trophically specialised nudibranch predators. The presence of s p i n e s is by no m e a n s confined to M. membranacea, their occurrence probably being most common in Cheilostomes. Zooids t h e m s e l v e s can specialise to form a physical defence. Avicularia are specialised zooids

94 which p o s s e s s a modified operculum which can be opened and closed by adductor m u s c l e s . They may serve a protective function discouraging larvae and small predators [IJ. The presence of secondary metabolites in organisms has been proposed as an adaptation against predation [56, 57, 58, 59. 6 0 , 611. Secondary metabolites, toxic or inhibitory to a range of predators, have been isolated from a whole host of different organisms. Secondary metabolites originating from s u c h diverse organisms as dytiscid beetles, [621, tropical seaweeds [63], gorgonians [64, 65) and sea hares [661 have been shown to exhibit anti-predatory properties. Active secondary metabolites are often deterrent to more than one predator. Algal secondary metabolites have been shown to deter both sea urchins and fish [671. Compounds from a Bahamian sponge deter feeding among a natural assemblage of fish [681. Ascidians have also been shown to contain a chemical defence against reef fish in field a s s a y s [611. Little work h a s been performed on the deterrent effect of bryozoan secondary metabolites. The bryozoan Sessibugula translucens contained the tambjamines A-D. which were deterrent to the spotted kelp fish Gibbonsia elegans [421. In a more detailed study [611 several tambjamines and a related tetrapyrrole were isolated from the ascidian Atapozoa s p . and its co-occurring nudibranch predator, Nembrotha s p . Several tambjamines (A, 9 6 . C. 9 8 . E. 1 2 1 , and F, 122) were tested as feeding deterrents towards a variety of carnivorous reef fish in field a s s a y s . Tambjamines A and E were not deterrent when tested at natural concentrations: however a 1:1 mixture of tambjamines E and F reduced feeding when tested below natural concentrations. The other tambjamines present were all significant feeding deterrents at or below natural concentrations. This work highlights the need for the careful control of test concentrations and mixtures when undertaking any kind of feeding experiment.

// \\

'H

/-\

// V

Nhuy

121 tambjamine E

N

K

/

^

122 tambjamine F

In our own unpublished work the palatability of four bryozoan s p e c i e s , two with and two without secondary metabolites, were com-

95 pared in feeding trials. Amathia wilsoni and Orthoscuticella ventricosa each contain a range of alkaloids (see Sections 2.3.1 and 2.3.4 respectively) whilst Bugularia dissimilis and Cellaria pilosa do not. All four bryozoans are common in Tasmanian coastal waters, to depths of 25m, but are more common in the 7-10m range. The coastal rocky reef habitat which they occupy has a high density and diversity of both invertebrate and vertebrate predators, such as the sea urchin Heliocidaris erythrogramma, and various reef fish.

t)U-

^

50-

13

o

•

sz

M- 40. (TJ SZ

\

4

0) 30'

+-»

'JD

^

JD OJ

20-

i

> 10-

Q. '

{ 1

0. ventricosa

,

A. wilsoni

1

1

1

1

1

C. pilosa

1

B. dissimilis

Bryozoan species Figure 1. Relative palatability of whole bryozoan colonies. The error bars represent one standard deviation. Leatherjackets (Acanthaluteres sp.) used in feeding trials showed a distinct feeding preference hierarchy when presented with whole colonies of the four species (Fig. 1). The same hierarchy was apparent when crude (Fig. 2) and purified compounds harman (Fig. 3a) and amathamide C (Fig. 3b) were used. The inhibitory effect harman and amathamide C displayed was less pronounced than when present in a crude extract at the same concentration (A. wilsoni—94% inhibition compared with amathamide C—78% inhibition; O. ventricosa—94% inhibition compared with harman—78% inhibition). The higher relative potency in the crude samples may have been due to a synergistic effect of the secondary metabolites or a slight difference in absolute

96 100

80

E Z5 CO

60

•

ii

I

^ I Control 1 I Treatment 1

d

o O c a) o

i Treatment 2 ^ Control 2

40

20

Q5

I

A. wilsoni

Q. ventricosa

B. disstnmlis

C. pilose

Bryozoan Species Figure 2. The relative palatability of crude bryozoan extracts and controls.

T3 (D

E Z) CO

c: o O

• control 1 vIJ Treatment 1 | rn Treatment 2 '^'control 2

(1)

o

Figure 3a. Relative palatability of pellets containing harman to controls.

Figure 3b. Relative palatability of pellets containing amathamide C to controls.

97 c o n c e n t r a t i o n s . C r u d e e x t r a c t s of O. ventricosa c o n t a i n a r a n g e of a l k a l o i d s ( h a r m a n a n d r e l a t e d c o m p o u n d s ) w h i c h m a y a c t in c o n c e r t to r e d u c e p a l a t a b i l i t y . Similarly A. wilsoni c o n t a i n s a s u i t e of s i m i l a r s e c o n d a r y m e t a b o l i t e s , t h e a m a t h a m i d e s , w h i c h m a y a c t in t h e s a m e way. V a r i a n c e in physiological effect a m o n g s t c o m p o u n d s w i t h s i m i l a r s t r u c t u r e s is well d o c u m e n t e d . Hay et al. t e s t e d a r a n g e of t e r p e n o i d s w h i c h differed in t h e i r c a r b o n s k e l e t o n s a n d c h e m i c a l f u n c t i o n a l i t i e s a n d c o n c l u d e d t h a t c h e m i c a l s t r u c t u r e could n o t b e u s e d a s a p r e d i c tor of feeding d e t e r r e n c e [69]. For e x a m p l e p a c h y d i c t y o l A, 1 2 3 , s t r o n g l y d e c r e a s e d grazing while dictyol E, 1 2 4 , w h i c h h a d one l e s s h y d r o x y g r o u p h a d no significant effect. T h e p r e s e n c e of a wide r a n g e of s i m i l a r c o m p o u n d s in, of e x a m p l e , b r y o z o a n s p e c i e s of t h e Amathia, Bugula, Chartella a n d Flustra g e n e r a (see S e c t i o n 2) m a y b e a r e s u l t of t h e i r differential activity a g a i n s t v a r i o u s p o t e n t i a l p r e d a t o r s .

123 pachydictyol A

124 dictyol E

Variation in the location of secondary metabolites within individual plants and animals has been investigated. This phenomenon has been most commonly documented in algae [70, 7 1 , 72, 73, 74, 75, 76, 77, 78], The "optimal defence** theory predicts that areas most susceptible to predation will be the most defended [79]. That is the portions of the organism most exposed should possess more secondary metabolites than other areas. This seems to be the case with Amathia wilsoni [30). The content and distribution of amathamide alkaloids (Section 2.3.1) within single colonies of the bryozoan varied depending on the location in the colony (Figure 4). The outermost, more exposed, tips of the colony had an alkaloid content of nearly 9% dry weight, while basal parts were apparently devoid of alkaloids. Samples taken midway between tips and base yielded intermediate concentrations of about 1%. As the amathamides have been shown to inhibit feeding of certain fish in laboratory trials this distribution supports the "optimal defence** theory. The feeding ecology of organisms associated with bryozoan colonies such as pycnogonids, nudibranchs, amphipods and ascidians

98 20.4%

15.5%

24.5%

27.770 68.0%

5cm

• Q

^ 2 3

Amathamide A Amathamide B Amathamide C

0

4

Amathamide E

Figure 4. Schematic of a section of a colony of A. wilsoni showing two major branches with associated sub-branching. Numbers represent total amathamides as percentage dry weight at each location. Pie charts represent the proportion of the major amathamides at the position indicated by the arrows. (Branch angles are schematic only.) has been poorly studied. It is not clear, except in the case of Sessibugula translucens [42], whether secondary metabolites dissuade, promote or protect organisms commonly found associated with bryozoans. Similarly the effect of bryozoan secondary metabolites on invertebrate predators has been poorly studied. Both these areas offer potential for collaboration between chemists and biologists.

3.3.

Antifouling

3.3.1 General The marine fouling community includes a well documented range of sessile marine organisms particularly algae, bryozoans, tunicates, bivalves, barnacles, polychaetes and sponges [80]. Invertebrate larvae and algae may settle on and then overgrow other organisms. Such

99 o v e r g r o w t h m a y b e d e t r i m e n t a l b y o b s c u r i n g feeding c u r r e n t s , b l o c k i n g light, c o m p e t i n g for available food a n d i n c r e a s i n g s e d i m e n t a t i o n . Being s e s s i l e filter feeders, b r y o z o a n s c a n often only t o l e r a t e low s e d i m e n t a t i o n r a t e s a n d prefer h i g h w a t e r v e l o c i t i e s . Complex p h y s i c a l a n d c h e m i c a l defensive s t r a t e g i e s h a v e evolved to c o m b a t o v e r g r o w t h in t h e Bryozoa. T h e r e a r e several s t r a t e g i e s t h a t b r y o z o a n s m a y employ to r e d u c e fouling p r e s s u r e a n d i n c r e a s e t h e i r c h a n c e s of s u r v i v a l . P h y s i c a l s u r f a c e p r o p e r t i e s of a n o r g a n i s m m a y affect levels of b a c t e r i a l a n d m a c r o f o u l e r s . Low s u r f a c e t e n s i o n s , different s u r f a c e t e x t u r e s , differe n c e s in s u r f a c e c h a r g e a n d d e g r e e of h y d r o p h o b i c i t y m a y all play s o m e p a r t in r e d u c i n g overall n u m b e r s of fouling o r g a n i s m s . T h e s e p a s s i v e p h y s i c a l p r o p e r t i e s m a y be c o u p l e d w i t h active p h y s i c a l d e f e n c e s t h a t d e t e r or kill s e t t l i n g l a r v a e . S e c o n d a r y m e t a b o l i t e s a r e often u s e d a s p a r t of a n o r g a n i s m ' s defensive s t r a t e g y . S e c o n d a r y m e t a b o l i t e s m a y affect bacterial n u m b e r s or p r o m o t e or d e t e r specific b a c t e r i a t h a t p r o v i d e s e t t l e m e n t c u e s to p o t e n t i a l m a c r o f o u l e r s . S e c o n d a r y m e t a b o l i t e s m a y a l s o d e t e r m a c r o f o u l e r s from s e t t l i n g or kill t h e m o n c e s e t t l e d . 3.3.2 Physical properties in relation to biofilms Biofilm d e v e l o p m e n t is affected by b o t h p h y s i c a l a n d c h e m i c a l f a c t o r s . The a b u n d a n c e a n d c o n d i t i o n of b a c t e r i a in t h e w a t e r c o l u m n p l a y s a major role in initial r a t e of s e t t l e m e n t on a s u r f a c e [81]. S u r f a c e f a c t o r s s u c h a s w e t t a b i l i t y [82] a n d critical s u r f a c e t e n s i o n [83], s u r f a c e h y d r o p h o b i c i t y [84], fluid d y n a m i c forces [85], s h e a r s t r e s s [86], electrolyte c o n c e n t r a t i o n [87] a n d m e t a b o l i c i n h i b i t o r s [88] c a n all affect m i c r o b i a l a t t a c h m e n t , a d h e s i o n or g r o w t h . T h e low s u r f a c e e n e r g y of a g o r g o n i a n o c t o c o r a l h a s b e e n i m p l i c a t e d a s a p a s s ive fouling r e s i s t a n c e m e c h a n i s m u s e d in c o n j u n c t i o n w i t h o t h e r a n t i fouling d e f e n c e s [82]. Active p h y s i c a l p r o c e s s e s s u c h a s l o s s of s u r f a c e t i s s u e s t h r o u g h s l o u g h i n g h a v e also b e e n s h o w n to i n f l u e n c e t h e p r i m a r y fouling film. The p r i m a r y film c o n s i s t s of m o l e c u l a r o r g a n i c a n d m i c r o b i a l c o m p o n e n t s t h a t a r e n o r m a l l y t h e first to **colonise** virgin s u r f a c e s . Microorganisms are lost when epidermal tissue and organic m e m b r a n e s are s l o u g h e d from t h e epibiont-free s u r f a c e of t h e g o r g o n i a n s Leptogorgia virgulata a n d Pterogorgia citrina [89]. Little w o r k h a s b e e n d o n e to c h a r a c t e r i s e t h e p h y s i c a l p r o p e r t i e s of b r y o z o a n s .

surface

100 3.3.3 The influence of biofilms on settling larvae The adsorption of biopolymers onto a surface [901. the attraction and adhesion of bacteria to that surface, their subsequent multiplication and exopolymer production leads to the formation of biofilms [911. The original conditioning film can influence the type and number of settling microorganisms, which in turn can affect the settlement of larvae of marine fouling organisms [92, 931. Larvae of many marine invertebrates require specific environmental stimuli to induce settlement or metamorphosis and such stimuli may be partially derived from the substratum [94, 95, 96, 97). A primary film modifies initial surface properties of an organism and may therefore influence the settling rate and composition of subsequent fouling organisms. Many marine invertebrates including bryozoans prefer to settle on filmed surface, but will also settle on unfilmed surfaces [941. Larvae from several species of the bryozoan Bugula sp. demonstrated preference for substrata that were coated with microorganisms [981. Larvae from the cosmopolitan bryozoan B. neritina were differentially attracted to bacterial films on different substrata [991. This may be due to a response to different bacterial films or a response to a combination of the physical properties specific to each surface and its bacterial film. A biofilm can either promote or reduce settlement rates depending on the organism being considered [911. Settlement experiments have been used not only to establish links between surface properties and settlement, but also as bloassays to isolate compounds possibly responsible for antifouling. Larvae of the bryozoan B. neritina have been used to isolate compounds from the ascidian Eudistoma olivaceum [1001. That larvae can **recognise** different chemical cues has long been accepted. The processes of attraction for the spirorbid Janua (Dexiospira) bioticus and possibly several bryozoan species may be lectin-mediated, where lectins on the larval surface **recognise** and bind to glycoconjugates in the exopolymers of the bacterial surface

[ion. The ability to reduce bacterial surface numbers or promote one type of bacterium over another and thereby encourage certain physical surface conditions, may be a beneficial adaptation employed by some bryozoans against fouling. Extracts from four species of bryozoans. found in Tasmanian coastal waters, have been demonstrated to exhibit selective antibacterial activity [491. The four species showed gradations in fouling by encrusting organisms and differential bacterial numbers.

101 T a b l e 1. I n h i b i t i o n of b a c t e r i a l g r o w t h b y f o u r b r y o z o a n e x t r a c t s . Bacterial strains

Extract 2A

2B

4A

6A

6B

8A

8B

lOA

lOB

U.w. DCM

++

++

++

++

-

++

++

++

++

U.w. MeOH

+++

-

+

+

-

+

+

++

-

a v . DCM

+++

++++

+

++++

+++

++++

•f+++

++++

-

p.v. MeOH

++++

+

++++

+++

++++

+

4-

++++

C.p. DCM

-

-

-

+

-

+

-

\c.p. MeOH

-

-

-

-

-

-

-

iB.d. DCM

-

+

-

+

-

+

-

B.d. MeOH

-

-

-

-

-

-

Gram stain

-

+

-

-

-

+

-

Morphology

rod

rod

cv-rod

rod

rod

cv-rod

rod

rod

rod

Size

var

var

var

var

sml

var

var

sml

vsml

m

m

nm

m

m

m

m

nm

nm

Motility

-

-

1 1

2A-10B indicate plates incubated in sea water between 2 and 10 days. A and B indicate 2 different bacterial strains tested at each interval (at the 4 day interval strain B failed to grow). Size of inhibition zone:

+

0-2 mm

++

2-4 mm

+++

4-6 mm

++++ 6-8 mm A.w.

- Amathia wilsoni

O.v. = Orthoscuticella

C.p.

= Cellaria pilosa

B.d. - Bugularia

DCM = dichloromethane

ventricosa

dissimilis

MeOH = methanol

Morphology: cv-rod = curved rod Size: var = variable size; sml = 5-10 ^m; vsml =< 5 ^m Motility: m = motile; nm = non-motile

The two chemically defended species Amathia wilsoni and Orthoscuticella ventricosa had the most active antibacterial extracts and the lowest levels of fouling. In contrast, extracts of Cellaria pilosa and Bugularia dissimilis, which have no known secondary metabolites, had weak antibacterial properties, and colonies showed large numbers of encrusting fouling organisms. Table 1 (adapted from [49]) shows the selective nature of the antibacterial extracts from the four bryozoans. SEM work conducted at the

102 same time shows the predominance of one type of bacterium on the tips of Amathia wilsoni where the highest levels of the amathamide alkaloids have been found [1021. A specific bacterium can be at least one inducer to larval settlement [103. 104. 105]. Promotion of growth of one bacterium over another may be a way of altering settlement cues to potential foulers in a way that may be beneficial to the bryozoan. The dichloromethane fraction of the aqueous methanol extract of the frondose bryozoan Flustra foliacea showed strong antibacterial activity against an indicator bacterium Bacillus subtilis in a disc diffusion assay (0.5 mg/disc) [39]. Earlier the older regions of Flustra foliacea had been described as having general antibacterial properties [1061. Scrupocellaria reptans would not colonise the older regions of Flustra foliacea, but a constant settlement on the distal growing edge of the frond was noted. This provided circumstantial evidence for bryozoan allelochemicals having an antifouling effect. The range of secondary metabolites that have been described for bryozoans as a group is quite extensive (Section 2). but except for the examples mentioned above, little work into their interactions with bacteria and the surface film has been undertaken. 3.3.4 Secondary metabolites and settling larvae Secondary metabolites themselves can act directly on settling larvae either by being immediately toxic or creating conditions that are not conducive to settlement or attachment once settlement has taken place. Extracts from the sponge Lissodendoryx isodictyalis inhibit settlement of the barnacle Balanus amphitrite at or below a concentration of 400ng/ml kills approximately 25% of settlement-stage larvae at 400ug/ml [107], These concentrations were similar to those found for the extracts of the sponge Xestospongia halichondriods which inhibited the settlement of Bugula neritina [108). A water borne compound released from the ascidian Diplosoma macdonaldi was not toxic to the larvae of the bryozoan Bugula pacifica, but significantly delayed their metamorphosis relative to controls [109]. Two dibromotyrosine metabolites isolated from the sponge Aplysina fistularis were shown to affect feeding in the bryozoan Membranipora membranacea, but had no effect on bryozoan larvae [110. 111). These compounds were exuded by the sponge and collected in both open and closed seawater systems. This example of the specificity of secondary metabolites is by no means isolated. Barnacle settlement inhibitors isolated from the sponge Renilla reniformis and the gorgonian Leptogorgia virgulata were ineffective against the larvae of the bryozoan Bugula neritina, but bryozoan inhi-

103

bitors also isolated from the same organisms had a deterrent effect on barnacle larvae [112|. Few studies have investigated the role that bryozoan secondary metabolites play in mediating the behaviour of would be fouling larvae. Several bryozoans have been screened against larvae of the competitor bryozoan Bugula turbinata [113]. Extracts from Flustra foliacea, which has been shown to contain both terpenes and alkaloids (see Sections 2.2.2 and 2.3.2 respectively) were both larvotoxic and antibacterial [4). These compounds may play an antifouling function in the bryozoan I114I. Bryozoan larvae are often used as test organisms to ascertain activities of extracts from other phyla, but testing bryozoan secondary metabolites for inhibitory or toxic effects on relevant larvae is long overdue. Also it is important not only to consider physical or chemical antifouling strategies in isolation as undoubtedly both mechanisms act in concert. 3.4.

Interspecies competition Competition for space between sessile marine organisms is a common and often important ecological pressure in the intertidal and sub-tidal zones. Overgrowth of one sessile organism by another is a well studied phenomenon [114]. Mechanisms used by benthic marine invertebrates in the competition for space are numerous and varied. These include physical adaptations, such as structures or growth patterns that reduce or hinder overgrowth by adjacent organisms, aggressive behaviour (such as feeding responses) and differential susceptibility to disturbance, and chemical adaptations [115]. Allelochemicals often have a role in the competitive interactions between neighbouring species. The possible chemical responses to overgrowth have been poorly studied in the Bryozoa. To influence an adjacent organism by chemical means, a mode of delivery for the allelochemical is needed that conserves chemical concentrations within their effective ranges and prevents unnecessary loss to the environment. In situ evidence of the release of allelochemicals into the surrounding water is rare. Coll and coworkers detailed a submersible sampling apparatus designed to detect water borne allelochemicals [116]. With the apparatus they demonstrated the release of chemicals from soft corals. Allelochemicals released by soft corals have been shown to produce necrosis in adjacent organisms [117]. We have sampled water around the bryozoan Amathia wilsoni, using a similar submersible sampling apparatus and

104

none of the range of alkaloid amathamides usually found in A, were isolated from the surrounding water [1021.

wilsoni

To our knowledge the possibility that bryozoan allelochemicals may influence competition for space amongst adjacent organisms h a s not been studied. Only sponge-bryozoan interactions have been studied. Bryozoan colonies being overgrown by s p o n g e s may exhibit a band of zooids a few millimetres wide paralleling the growing edge of the sponge [1181. To test the effect of sponge secondary metabolites, four bryozoan species were subjected to extracts from 11 sympatric species of s p o n g e s and colonial a s c i d i a n s . Five of the nine sponge species and one of the ascidian s p e c i e s exhibited species-specific allelochemical effects. The use of whole-organism extracts does not establish the importance of sponge allelochemicals in nature as toxins may be stored in the animal and then released into the water in different concentrations. In situ sampling to determine the concentration of any secondary metabolites present m u s t occur before inferences can be made on ecological significance. These results [119] do. however, demonstrate the potential significance of allelochemicals in competition for space between bryozoans and sponges.

3.5.

Sequestration of bryozoan secondary metabolites

Cases when secondary metabolites are taken in by a specialist predator and then used for their own defence are quite common. Predators may ingest the compounds and detoxify them, as with the mollusc Ovula ovum [1191 or use secondary metabolites directly as a defence. The opisthobranchs are one of the most studied c l a s s e s of marine invertebrates that sequester secondary metabolites. Nudibranchs are known to obtain their secondary compounds from a range of sources taking in compounds primarily from sponges 1120], ascidians [611, bryozoans [42] and coelenterates [121]. The chemical relationship between nudibranchs and sponges h a s been reviewed [122, 122, 123]. Algae are the dietary source of the chemicals found in sea hares and a s c o g l o s s a n s [61, 66, 124]. Several s t u d i e s have shown the feeding deterrent effect these isolated compounds have on predators [61, 125, 126]. Field a s s a y s as well as laboratory t e s t s have also been employed [66]. Whole sea hares {Stylocheilus longicaudia) were used along with crude and isolated extracts to show a deterrent effect on the feeding of reef fish. Bryozoans are a common dietary source for many opisthobranchs, but few investigations have been made to link opisthobranchs, bryozoans and their secondary metabolites.

105

Chemical analysis of the nudibranchs Roboastra tigris, Tambja abdere and Tambja eliora revealed that all three contained a range of tambjamines A-D derived from the bryozoan Sessibugula translucens |42I. These purified compounds w h e n added to freeze-dried e u p h a u siids produced an inhibitory feeding effect in spotted kelpfish. The fish showed a significant feeding avoidance in concentrations ranging between 1 and 5 0 u g / m g . Tambjamines from the nudibranch Nembrotha spp. have b e e n shown to be significant feeding deterrents to a range of carnivorous reef fish [61]. Even though t h e s e tambjamines were derived from a tropical ascidian they are structurally identical to several tambjamines found in Sessibugularia translucens. Pycnogonids are one of the m o s t common co-occurring predators associated with bryozoans, but up till now have not been shown to s e q u e s t e r secondary compounds. Our recent work h a s shown that sequestration of secondary compounds by pycnogonids from their bryozoan diet is widespread [102]. After screening a variety of organisms associated with the two bryozoans Amathia wilsoni and Orthoscuticella ventricosa, several s p e c i e s of pycnogonids were found to contain relatively high concentrations of alkaloidal secondary metabolites. One pycnogonid s p e c i e s , Stylopallene longicauda, present in a high density on A. wilsoni (46 ind/lOOcm^), w a s found to contain a range of amathamides in concentrations higher than the average found in the host colony. Similarly the pycnogonids Achelia s p p . were found to have sequestered secondary metabolites. Achelia s p p . were associated primarily with the bryozoan O. ventricosa and were found to contain harman and a range of related c o m p o u n d s . Another pycnogonid Pseudopallene ambigua was also associated with O. ventricosa and the main alkaloid contained by this particular pycnogonid w a s norharman. Presumably norharman was derived from harman. It is widely a s s u m e d that sequestered metabolites function in antipredator defence. Both amathamide C and harman have been shown to reduce feeding rates (86%-inhibition and 78%-inhibition respectively) of reef fish reinforcing this theory. Feeding trials u s i n g whole pycnogonids showed that P. ambigua was significantly l e s s palatable to a co-occurring fish than Achelia spp.

4.

Conclusion

The metabolites from 3 5 different bryozoan s p e c i e s have been described to date and 3 4 of t h e s e are d i s c u s s e d in this review. Compounds from six of these s p e c i e s together with Alcyonidium gelatinosum. were detailed also in the earlier review of bryozoan chemistry

106 [4]. The overwhelming majority of the bryozoans examined belong to c l a s s Gymnolaemata (25 in order Ctenostomata and 8 in Cheilostomata): there is only one representative from the bryozoan class Stenolaemata (order Cyclostomata). Over 130 secondary metabolites have been identified from bryozoans with 75 compounds being novel. Of these. 56 (75%) are alkaloids. A summary of bryozoan secondary metabolites is presented in Table 2. The origin of bryozoan secondary metabolites remains a mystery; there is some circumstantial evidence that points to a bacterial source b u t the producing organism and the biosynthetic pathways clearly need to be established unambiguously. Many of the compounds from bryozoans exhibit biological activity, most notably the anitineoplastic action of the bryostatins. Despite t h i s , the chemical ecology of compounds from only four bryozoans h a s been investigated, often a s crude extracts rather than pure c o m p o u n d s . One area worthy of more study, not least b e c a u s e of its economic implications, is the effects that bryozoan secondary metabolites have on potential fouling organisms. Also meriting further investigation is the generality of the finding that, in the c a s e of one bryozoan (Section 3.2) there is intercolonial variation of chemical defence. Similarly the two observations that pycnogonids sequester bryozoan metabolites show that this phenomenon deserves further study. Table 2. Summary of bryozoan secondary metabolites. Bryozoan

Order

Structures

Compound Type Reference

Alcyonidium gelatinosum * Ctenostomata

t

S comp.

[4]

Amathia alternata

Ctenostomata

75-78

alkaloid

[32]

Amathia convoluta

Ctenostomata

8

macrocycle

69, 70-74

alkaloid

[11] [31,32]

Amathia pinnata

Ctenostomata

64

alkaloid

[31]

Amathia wilsoni

Ctenostomata

62.66

alkaloid

[27-30]

Amathia sp.

Ctenostomata

52-54

fatty aldehyde

[23]

Biflustra perfragilis

Cheilostomata 56-61

hal.& S comp.

[26]

115-117 Bugula dentata

Cheilostomata 100

alkaloid alkaloid

[43]

107

Bryozoan Bugula neritina *

Structures

Order

Compound Type Reference macrocycle

[4,13-17]

52-54

fatty aldehyde

[23]

101

alkaloid

[44]

Cheilostomata 1-19

Bugula turrita

Cheilostomata 51

terpene

[21]

Cauloramphus spiniferum

Cheilostomata 52-54

fatty aldehyde

[23]

Cellaria flstulosa

Cheilostomata 101-105, 118-12C1 alkaloid

[44]

Cellaria

Cheilostomata 101-105, 118-120 alkaloid

[44]

Cellaria sinusoa

Cheilostomata 101-105, 118-120 alkaloid

[44]

Celleporina

Cheilostomata 52-54

fatty aldehyde

[23]

Cheilostomata 82-87

alkaloid

[4, 35-37] 1

salicornioides

irregularis

Chartella papyracea *

101, 102, 118-120 alkaloid

[44]

Conopeum seuratum

Cheilostomata 34-51

terpene

[21]

Costaticella hastata

Cheilostomata 109-112

alkaloid

[48, 49]

Cribricellina

Cheilostomata 25-33

sterol

[20]

cribraria

101, 109-111,

alkaloid

113-114 Diaperoecia

californica

Cyclostomata

106-108

alkaloid, N comp .[45-47]

1

Exochella areolata

Cheilostomata 52-54

fatty aldehyde

[23]

Flustra foliacea *

Cheilostomata t , 88-94

[4, 46]

t

alkaloid terpene

108

N comp.

[46]

Hincksinoflustra denticulata Cheitostomata 95

alkaloid

[41]

Hippodiplosia

Cheilostomata 108

N comp.

[46]

ciliata

Cheilostomata 52-54

fatty aldehyde

[23]

Myriapora truncata

Cheilostomata 20-24

sterol

[18, 19]

Orthoscuticella

Cheilostomata 109-112

alkaloid

[48, 49]

Phidolopora paciflca •

Cheilostomata 106-108

alkaloid, N comp . [4, 45-47]

Schizoporella unicornis

Cheilostomata 52-54

fatty aldehyde

[23]

Sessibugula translucens *

Cheilostomata 96-99

alkaloid

[4, 42]

Tricellaria ternata

Cheilostomata

N comp.

[46]

Watersiporia aterrima

Cheilostomata 52-54

fatty aldehyde

[23]

Zoobotryon verticillatum *

Ctenostomata

alkaloid

[4, 33, 34]

Heteropora alaskensis

Microporella

insculpta

ventricosa

Ctenostomata

79-81

* Also included in previous review [4] f Not included in this review

[4, 22]

108

5. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

References J . S . Ryland, in: "Bryozoans**, Hutchinson and Co Ltd, London, 1970, pp 11-29. P.E. Bock, in: S.A. Shepherd, and LM. Thomas (Eds) **Marine Invertebrates of Southern Australia. Part 1**: Bryozoans (phylum Bryozoa), South Australian Government. 1982, pp 319-394. F. Borg, ZooL Bidr. Upps.. 10 (1926) 181-507. C. Christophersen, Acta Chem. Scand.. B 39 (1985) 517-529. G.R. Pettit, in: W. Herz. (Ed) "Progress in the Chemistry of Organic Natural Products'*: The bryostatins, Springer-Verlag, New York, 1991, Vol. 57. pp 153-195. U. Anthoni, P.H. Nielsen, M. Pereira and C. Christophersen, Comp. Biochem. Physiol.. 96B (1990) 431-437. C. Christophersen, in A. Brossi. (Ed) "The Alkaloids": Marine alkaloids. Academic Press, New York, 1985. Vol. 24, pp 25-111. J . Kobayashi and M. Ishibashi, in: A. Brossi and G.A. Cordell, (Eds) "The Alkaloids**: Marine alkaloids II, Academic Press, New York. 1992. Vol. 4 1 . pp 42-123. W. Fenical. in: S.W. Pelletier (Ed.) "Alkaloids: Chemical and Biological Perspectives**: Marine alkaloids and related compounds, Wiley. New York. 1986. Vol. 4. pp 275-330. D.J. Faulkner. Nat. Prod. Rep., 9 (1992) 323-364. and earlier reports. G.R. Pettit. Y. Kamano. R. Aoyagi, C.L. Herald. D.L. Doubek. J.M. Schmidt and J . J . Rudloe. Tetrahedron. 41 (1985) 985-994. A.J. Blackman. T.P.D. Eldershaw and S.M. Garland. Aust. J. Chem.. 46 (1993) 401-405. G.R. Pettit. F. Gao. D. Sengupta. J.C. Coll. C.L. Herald. D.L. Doubek. J.M. Schmidt. J.R. Van Camp. J . J . Rudloe and R.A. Nieman. Tetrahedron. 47 (1991) 3601-3610. D.E. Schaufelberger. G.N. Chmurny. J.A. Beutler. M.P. Koleck, A.B. Alvarado. B.W. Schaufelberger and G.M. Muschik, J. Org. Chem.. 56 (1991) 2895-2900. G.N. Chmurny, M.P. Koleck and B.D. Hilton. J. Org. Chem.. 57 (1992) 5260-5264. M. Kageyama and T. Tamura. J. Am. Chem. Soc. 112 (1990) 74077408. G.R. Pettit, F. Gao, D.L. Herald, P.M. Blumberg. N.E. Lewin and R.A. Nieman. J. Am. Chem. Soc. 113 (1991) 6693-6695. F. Cafieri. E. Fattorusso, M. Gavagnin and C. Santacroce. J. Nat. Prod.. 48 (1985) 944-947. N.V. Kovganko and S.K. Ananich. Khim. Prir, Soedin.. 1989. 664-669. M.R. Prinsep, J.W. Blunt and M.H.G. Munro, J. Nat. Prod. 54, (1991) 1068-1076.

109

21 P. Hadjleva. S. Popov. B. Budevska and A. Dyulgerov, Z. Naturforsch., 42C, (1987) 1019-1022. 22 C. Christophersen and J . S . Carle, Naturwissenschaften, 6 5 (1978) 440-441. 23 E.V. Berdyshev. O.E. Getmannnova, V.I. Svetashev and A.A. Kubanin. Comp. Biochem. Physiol., 102B (1992) 639-641. 24 F.B. Whitfield. J.H. Last. K.J. Shaw and C.R. Tindale. J. Set. Food Agric, 46 (1988) 29-42. 25 F.B. Whitfield. K.J. Shaw and D.I. Walker. Wat Set Tech.. 2 5 (1992) 131-138. 26 A.J. Blackman, N.W. Davies and C.E. Ralph, Biochem. Syst. EcoL, 2 0 (1992) 339-342. 27 A.J. Blackman and D.J. Matthews. Heterocycles, 2 3 (1985) 2829-2833. 28 A.J. Blackman and R.D. Green, Aust. J. Chem., 40 (1987) 16551662. 29 A.J. Blackman and S.-L. Fu. J . Nat. Prod.. 52 (1989) 436-438. 30 J.T. Walls. A.J. Blackman and D.A. Rit2. J. Chem. EcoL. 17 (1991) 1871-1881. 31 A.J. Blackman. T.P.D. Eldershaw and S.M. Garland, Aust. J. Chem., 4 6 (1993) 1401-405. 32 W. Fenical, private communication. 1993. 33 A. Sato and W. Fenical. Tetrahedron Lett., 24 (1983) 481-484. 34 M.J. Ortega. E. Zubia and J . Salva. J. Nat. Prod., 56 (1993) 633636. 35 L. Chevolot. A.-M. Chevolot, M. Gajhede. C. Larsen. U. Anthoni and C. Christophersen. J. Am. Chem. Soc, 107 (1985) 4542-4543. 36 U. Anthoni. L. Chevolot, C Larsen, P.H. Nielsen and C. Christophersen, J. Org. Chem. 52 (1987) 4709-4712. 37 U. Anthoni, K. Bock, L. Chevolot, C. Larsen. P.H. Niesen and C. Christophersen. J. Org. Chem., 52 (1987) 5638-5639. 38 P. Keil. E.G. Nielen. U. Anthoni and C. Christophersen. Acta Chem. Scand., B 4 0 (1986) 555-558. 39 M.V. Laycock, J.L.C. Wright, J.A. Findlay and A.D. Patil, Can. J. Chem., 64 (1986) 1312-1316. 40 J.L.C. Wright, J. Nat. Prod.,47 (1984) 893-895. 41 A.J. Blackman, T.W. Hambley, K. Picker, W.C. Taylor and N. Thirasasana. Tetrahedron Lett, 28 (1987) 5561-5562. 42 B. Carte and D.J. Faulkner, J. Chem. EcoL, 12 (1986) 795-804. 43 S. Matsunaga, N. Fusetani. K. Hashimoto. Experientia, 4 2 (1986) 84. 44 U. Anthoni. C Larsen. P.H. Nielsen, C. Christophersen and G. Lidgren, Comp. Biochem. PhysioL, 92B (1989) 711-713. 45 S.W. Ayer, R.J. Andersen, H. Cun-heng and J . Clardy, J. Org. Chem., 50, 3869-2870. 46 M., Tischler, S.W. Ayer and R.J. Andersen, Comp. Biochem. PhysioL, 84B (1986) 43-45.

110

47 K. Hirota. K. Kubo, Y. Kitade, Y. Maki, Tetrahedron Lett, 26 (1985) 2355-2356. 48 A.J. Blackman. D.J. Matthews and C.K. Narkowicz. J. Nat. Prod,, 50 (1987) 494-496. 49 J.T. Walls, D.A. Ritz and A.J. Blackman. J. Exp. Mar. Biol. EcoL, In press. (1993). 50 F. Bracher and D. Hildebrand. Liebigs Ann. Chem., (1992) 13151319. 51 A.J. Blackman. C.E. Ralph. B.W. Skelton and A.H. White. Aust. J. Chem., 46 (1993) 213-220. 52 F.J. Schmitz and T. Yasumoto. J. Nat. Prod., 64 (1991) 1469-1490. 53 P.E. Bock, as quoted in ref. [51]. 54 C D . Harvell. Science, 224 (1984 ) 1357-1359. 55 C D . Harvell. The American Naturalist 128 (1986) 810-823. 56 R.L Vadas. Ecological Monographs. 47 (1977) 337-371. 57 J . C Ogden and P.S. Lobel. Envir. Biol. Fish., 3 (1981) 49-63. 58 J.N. Norris and W. Fenical. in K. Rutzler and I.G. Maclntyre (Eds). The Atlantic Barrier Reed ecosystem at Carrie Bow Cay. Belize I: Structure and communities (Smithsonian Contrib. Mar.Sci.), Washington D . C . 1982. pp 417-431 . 59 W. Fenical. Science 215 (1982) 923-928. 60 V.J. Paul and W. Fenical. Science 221 (1983) 747-748. 61 V.J. Paul. N. Lindquist and W. Fenical, Marine Ecol. Progress Series 59 (1990) 109-118. 62 D.J. Gerhart. Am. Physiol. Soc. 91 (1991) 839-843. 63 M.E. Hay and W. Fenical. Ann. Rev. Ecol. Syst., 19 (1988) 111-145. 64 C D . Harvell. W. Fenical and C H . Greene. Mar. EcoL Prog. Ser. 49 (1988) 287-294. 65 C D . Harvell and W. Fenical. Limnol. Oceanogr., 34 (1989) 382-389. 66 V.J. Paul and S.C Pennings. J. Exp. Mar. Biol. Ecol., 151 (1991) 227-243. 67 M.E.Hay. J.E. Duffy. CA. Pfister and W. Fenical. Ecology, 68 (1987) 1567-1580. 68 S.C. Bobzin and D.J. Faulkner. J. Chem. Ecol., 18 (1992) 309-333. 69 M.E. Hay. J.E. Duffy and W. Fenical. Proceedings of 6th. Int. Coral Reef Symposium, Australia. 1988. Vol. 3. pp 43-48. 70 M.A. Ragan and A. Jensen. J. Expt. Marine Biol. Ecol, 34 (1978) 245-258. 71 D.W. Phillips and G.H.N. Towers. J. Expt. Marine Biol. Ecol, 58 (1982) 285-293. 72 P.D. Steinberg. Science, 2 2 3 (1984) 405-407. 73 W.H. Gerwick. W. Fenical and J.N. Norris. Phytochem., 24 (1985) 1279-1283. 74 K.L.J. Van Alstyne and V.J. Paul. Proceedings of 6th. Int. Coral Reef Symposium, Australia, 1988. Vol. 1. pp 175-186.

in 75 D.J. Carlson, J . Lubcheno, M.A. Sparrow and C D . Trowbridge, J. Chem. EcoL. 15 (1989) 1321-1333. 76 P.D. Steinberg. Oceologia, 7 8 (1989) 373-382. 77 S. Tugwell and G.M. Branch, J. Expt. Marine Biol. EcoL, 129 (1989) 219-230. 78 K.D. Meyer and V.J. Paul, Mar. EcoL Prog. Ser., 82 (1992) 249257. 79 J.K. Nitao and A.R. Zangerl, Ecology, 6 8 (1987) 521-529. 80 S. Yan and Z. Huang. Biofouling, 2 (1990) 229-237. 81 K.C. Marshall, in: R.W.C. Berkeley, J.M. Lynch, J. Melling,P.R. Rutter and B. Vincent (Eds.) Microbial Adhesion to Surfaces Ellis Horwood Limited, Chichester, 1980, pp 187-196. 82 N.H. Vrolijk, N.M. Targett, R.E. Baier and A.E. Meyer. Biofouling, 2 (1990) 39-54. 83 J . J . Zambon, P.S. Huber, A.E. Meyer, J . Slots. M.S. Fornalik and R.E. Baier, Applied Environmental Microbiol, 48 (1984) 12141220. 84 M.C.M van Loosdrecht, J. Lyklema, W. Norde, G. Schra and A.J.B. Zehnder, Applied Environmental Microbiol., 5 3 (1987) 1893-1897. 85 J.D. Bryers and W.G. Characklis, Biotechnol. Bioeng., 2 4 (1982) 2451-2476. 86 T.A. Horbett, J . J . Waldburger, B.D. Ratner and A.S. Hoffman, Journal Biomed. Mat. Res., 22 (1988) 383-404. 87 K.E. Cooksey, Applied Environmental Microbiol., 41 (1981) 1378-1382. 88 J.H. Paul, Applied Environmental Microbiol., 48 (1984) 924-929. 89 W.K. Patton, Bulletin of Marine Science, 2 (1972) 419-431. 90 R Lewin, Science, 224 (1984) 375-377. 91 H.M. Lappin-Scott andH.M. Costerton.H.M. Biofouling, 1 (1989) 323-342. 92 D.J. Crisp and J . S . Ryland, Nature, 185 (1960) 119. 93 J.R. Henschel and P.A. Cook. Biofouling, 2 (1990) 1-11. 94 D.J. Crisp, in: Chemoreception in Marine Organisms. P.T. Grant and A. Mackie (Eds). Academic Press. New York. 1974. pp. 177-265. 95 R.S. Scheltema. Thalassia lugosl., 10 (1974) 263-296. 96 R.D. Burke. Can. J. Zool., 6 1 (1983) 1701-1719. 97 M.G. Hadfield. Bull. Marine Set, 39 (1985) 418-425. 98 M.S. Brancato and R.M. Woollacott. Mar. Biol., 7 1 (1982) 51-56. 99 J.W. Mihm. W.C. Banta and G.I. Loeb. J. Exp. Mar. Biol. EcoL, 54 (1981) 167-179. 100 A.R. Davis and A.E. Wright. J. Chem. EcoL, 16 (1990) 1349-1357. 101 D. Kirchman. S. Graham. D. Reish and R. Mitchell. J. Exp. Mar. BioL EcoL, 56 (1982) 153-163. 102 J.T. Walls. A.J. Blackman and D.A. Ritz. unpublished observations.

112

103 W.A. Miiller in: Proceedings of the 2nd Symposium on Cnidaria (Recent trends in coelenterate biology). T. Tokiota and S. Nishimura (Eds). Seto. Mar. Biol. Lab., 1973, pp. 195-208. 104 R. Neumann, Mar. EcoL Prog, Ser., 1 (1979) 21-28. 105 R.M. Weiner, R.M. Segall, and R.R. Colwell, Applied and Environmental Microbiol, 49 (1985) 83-90. 106 S.M. Al-Ogily and E.W. Knight-Jones, Nature, 265(1977) 728-729. 107 M.A. Sears, D.J. Gerhart and D. Rittschof, J. of Chem. EcoL, 16 (1990) 791-799. 108 G.N. Ware, The patterns and mechanisms of antifouling in some temperate marine sponges, Duke University, Durham, North Carolina, 1984. 109 C M . Young and F.S. Chia, International Journal of Invertebrate Reproduction (Amsterdam), 3 (1981) 221-226. 110 J.E. Thompson, Mar. BioL, 88 (1985) 23-26. 111 R.P. Walker, J.E. Thompson and D.J. Faulkner, Mar. BioL, 88 (1985) 27-32. 112 D. Rittschof, I.R. Hooper and J.D. Costlow, in: Thompson, M. (Ed), Marine Biodeterioration, Oxford and IBW Publishing, Bombay, India, 1988 pp. 599-608. 113 P.E.J. Dyrynda, Dev. Comp. ImmunoL, 7 (1983) 621-624. 114 J.C. Coll, I.R. Price, G.M. Konig and B.F. Bowden, Mar. BioL, 96 (1987) 129-135. 115 G.J. Bakus, N.M. Targett and B. J. Schulte, Chem. EcoL, 12 (1986) 951-985. 116 J.C. Coll, B.F. Bowden and D.M. Tapiolas, J. Exp. Mar. BioL EcoL 60 (1982) 293-299. 117 P.W. Sammarco, J.C. Coll and S. J. La Barre. Exp. Mar. BioL EcoL 91 (1985) 199-215. 118 J.B.C. Jackson and L. Buss, Proc. Nat. Acad. ScL, 72 (1975) 5160-5163. 119 J.C. Coll, L.M. Tapiolas, B.F. Bowden, L. Webb and H. Marsh, Mar. BioL, 74 (1983) 35-40. 120 P.J. Scheur, Naturwissenchaften, 69 (1982) 528-533. 121 P. Karuso, in; P.J. Scheuer (Ed). **Bioorganic Marine Chemistry**, Springer-Verlag, Berlin, 1987, pp. 31-60. 122 J.E. Thompson, R.P. Walker, S.J. Wratten and D.J. Faulkner, Tetrahedron, 38 (1982) 1865-1873. 123 D.J. Faulkner and D.J. Ghislen, Mar. EcoL Prog. Ser., 13 (1983) 295-301. 124 D.J. Faulkner, Nat. Prod. Rep., 7 (1984) 251-280. 125 V.J. Paul and K.L.J. Van Alstyne, Exp. Mar. BioL EcoL. 160 (1992) 191- 203. 126 J.R. Pawlik, M.R. Kernan, T.F. Molinski, M.K. Harper and D.J. Faulkner, Exp. Mar. BioL EcoL, 119 (1988) 99-109.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

113

Structure and Biological Activity of Triterpenoids and Aromatic Compounds from Medicinal Plants R. Aquino, F, De Sinione, N, De TomniasI and C. Pizza Dipartimento di Chimica delle Sostanze Naturali, Via D. Montesano, 49, Napoli, Universitd degli Studi di Napoli

n.

INTRODUCTION

Medicinal plants have always played a key role in world health. In fact plants represent an enormous reservoir of new, undiscovered and bioactive molecules. They are used in modem medicine in four basic ways: a) as sources of direct therapeutic agents; b) as a raw material for the preparation of more complex semi-synthetic drugs; c) the chemical structures of plant metabolites can be used as models for new synthetic compounds; d) plant metabolites can be used as taxonomic markers to establish relationships between groups of plants and to forecast the presence of biologically interesting compounds in families and genera. Over the last two decades there has been a resurgence of interest in the study and use of medicinal plants. The WHO (World Health Organization) has confirmed the importance of traditional medicine to a majority of the. world' s population and encourages all countries to preserve and to use the safe and positive elements of traditional medicine in their national health systems. The WHO Traditional Medicine Programme (1) was inspired by the observation that 80% of the world's population treats diseases exclusively with traditional medicines, and most traditional therapies involve the use of plant extracts or their active constituents. Vegetable species from South America and China are of particular interest in view of their wide use in traditional medicine; they offer local populations immediately accessible safe and effective therapeutic products. However only a small fraction of South American and Chinese medicinal plants have been studied. Therefore it is of general interest to document the experience of traditional healers, to select interesting medicinal species and to identify the constituents responsible for their therapeutic or toxic effects . In the last ten years, in collaboration with Pontificia Universidad Catolica del Peru, Universidad Nacional Mayor de San Marcos, Peni, and other South American institutions and with the Shanghai Institute of Pharmaceutical Industry, our research group has been investigating various species employed in traditional medicines and collecting all information available on the use of plants in a variety of different ailments; the specimens are identified by a qualified botanist and compounds responsible for the therapeutic effects claimed by traditional headers are characterized. The plants studied (Table 1) undergo suitable bioassays in order to locate the pharmacological activity in crude extracts and/or in fractions issuing from various separation steps and/or in pure isolated compounds. Our results are critically analysed and pure isolated compounds undergo pharmacological screening on the basis of their structural relationships with known drugs.

114 In this chapter part of our recent work on the chemistry and biological activity of metabolites from medicinal plants will be summarized. The main characteristic metabolites contained in plants examined are: 1) polyhydroxylated triterpenes and triterpene esters 2) triterpenic glycosides 3) steroidal glycosides 4) aromatic compounds 5) free and glycosydated sesquiterpenes and diterpenes Items 1-4 are analyzed in the following paragraph. Table]. Name of plant and Family

Name of plant and Family

Source

Moniordica balsamina South America (Cucurbitaceae) Wenieria ciliolata South America (Compositae) Tessaria integrifolia South America (Compositae) Senecio pseudotites South America (Compositae) Hcunelia patens South America (Rubiaceae) Guettarda platypoda South America (Rubiaceae)

Be/aria cbmamotnea (Ericaceae) Werneria dactylophylla (Compositae) Eupatoriwn tinifolium (Compositae) Eupatoriwn giudanum (Compositae) Coutarea hexandra (Rubiaceae) Uncaria tofnentosa (Rubiaceae)

South America

Uricaria guaianensis (Rubiaceae) Croton draconoides (Euphorbiaceae) Ta^etes elliptica Detarium microcarpum (Lxguminosae) Combretum micranth. (Combretaceae) Tamils communis (Dioscoriaceae)

South America

Arcytophyllum nitidum South America (Rubiaceae) Mintostachys setosa South America (Labiatae)

Italy

Mercurialis amma (Euphorbiaceae)

Italy

Quercus suber (Cupuli ferae) Calendula arvensis (Compositae) Asparagus cochinchin. (UHaceae) Eriobotrya japonica (Rosaceae)

Italy

Quercus cerris (Cupuliferae) Anchusa officinalis (Boraginaceae) Ardisia japonica (Myrsinaceae)

Italy

1 Source

South America South America Senegal

South America South America South America South America South America

Securidaca longipedun. Senegal (Polygalaceae)

Senegal

Italy China China

Italy China

115 1-1 GENERAL PROCEDURE OF PURIFICATION. The strategy for the separation of pure compounds from plant material, some aspects of which will be discussed here, is crucial for this type of research. The compounds, some of which are biologically active, may be present in very small quantities; many plants have several constituents and it can be very difficult to separate one particular component. Therefore the problems in separation range from the isolation of minute quantities (a few milligrams) for structure determination to the isolation of larger amounts for comprehensive biological tests. No single chromatographic separation method is able to solve all separation problems so multistep chromatographic operations are normally used for the isolation of pure substances. Our standard separation procedure consists of defatting dried plant material with petroleum ether and extraction with solvents of increasing polarity: CHCI3, CHCl3-MeOH, MeOH, H2O; thus hydrophobic metabolites are found in the CHCI3 extract and glycosides in the CHCls-MeOH, MeOH and H2O extracts. The initial stages of extract separation involve traditional column methods with stationary phase as silica gel, XAD ;filtration on dextran gels at different stages of isolation is a complementary method. Subsequent steps for separation of smaller fractions from column chromatography include DCCC (droplet counter current chromatography) and HPLC (high performance liquid chromatography). Selectivity can be adjusted by varying the separation mode (adsorption, partition, ion-exchange, gel filtration) and the eluents. DCCC, a liquid -liquid separation method, does not require a sorbent and involves the partition of a mixture between two immiscible solvents. The proportions of solute which pass into each of the two phases is determined by the respective partition coefficient. This technique offers a number of advantages for the isolation of natural products: total recovery of the introduced sample , no irreversible adsorption , minimal risk of sample denaturation. We use DCCC mainly to separate polar compounds such as glycosides. The choice of a two-phase solvent system is crucial to the success of a DCCC separation and ternary systems are required to increase the selectivity. HPLC is usually applied as a last step in a purification process to isolate pure substances. RP-HPLC with a reversed -phase packed columns is normally used for polar compounds. By varying the per cento of binary eluent system (MeOH-H20) it is possible to separate substances ranging from to terpenic and flavonoidic glycosides. Less polar compounds such as polyhydroxylated triterpenes and their esters and flavonoid aglycones are usually separated from CHCI3 extracts by preliminary fractionation on silica gel column with CHCI3 -MeOH gradients followed by RP-HPLC on a Cig column with MeOH or MeOH-H20 mixtures as the eluent. A typical example is the separation of ursolic acid derivatives from oleanolic acid derivatives and their esters from E, Japonica (23). Medium polar glycosides present in CHCl3-MeOH (9:1) and/or MeOH extracts require a four step procedure: —partition of the extract between n-BuOH and H2O; —gelfiltration on Sephadex LH-20 column with MeOH as the eluent; —DCCC of the Sephadex fractions with a ternary system CHCl3-MeOH-H20 (7:13:8) or of varying composition. Usually the stationary phase is the lower one, and ascending mode is used;

116 — semipreparative HPLC on an RP-18 column with MeOH-H20 in different proportions for final purification. An example of application of this method is the separation of the C-27 glycosyl ester of quinovic acid from its C-28 glycosyl ester isomers from V. tomentosa (4) and G. platypoda (5). For highly polar water-soluble glycosides present in the aqueous extracts, the strategy involves lyophilization of the water extract, passage through an Amberlite XAD-2 column eluted with H2O followed by MeOH, and rechromatography of the dried methanolic eluate on Sephadex column followed by RP-HPLC. A typical example of isolation of highly polar glycosides is that of spirostanol and furostanol glycosides from T, communis (6) and A. cochinchinetisis (7). 11.2 STRUCTURE DETERMINATION METHODS. The metabolites isolated from medicinal plants have a wide variety of chemical structures but complex polar structures such as glycosides of triterpene, steroidal and flavonoid aglycones are of particular interest. Our method for structure determination originally involved a combination of NMR spectroscopy with chemical transformation and enzymatic degradation as well as FAB-MS spectroscopy (in positive and/or negative ion mode); —Acid methanolysis of the glycosides and subsequent GLC analysis of the resulting persilylated methyl sugars provided information on the nature and ratio of the sugars; — FAST ATOM BOMBARDMENT (FAB) mass spectra gave the molecular weight of the whole glycoside and information on the saccharide sequence showing the sequential loss of more external monosaccharide units; —The chemical shift and coupling constant information available from both ^H and l^C NMR spectra confirmed the type of monosaccharide units present in the carbohydrate moiety and indicated their anomeric configurations (for example a and p pyranoid forms are generally distinguishable by chemical shifts of C-1, C-3, and C-5 in the l^C NMR spectra ; in the ^H NMR spectra the anomeric resonance of a-glycosides are at a downfield position by 0.3-0.5 ppm with respect to the corresponding P-anomers; furanose sugars are characterized by distinctive ^^C NMR chemical shifts) (8); — one dimensional pulse sequences techniques such as DEPT (DISTORTIONLESS ENHANCEMENT BY POLARIZATION TRANSFER) allowed distinction between methine, methylene, methyl and quaternary carbons; —empirical evaluation of a and p glycosidation effects and comparison with model compounds, in combination with the analysis of fragments obtained by chemical and enzymatic degradations showed how the monosaccharide units were linked. In the case of a long and branched sugar chain molecular rotation values in the light of Klyne's rules also provided useful information about the absolute and anomeric configurations (9). This approach was used to establish the structures of triterpenoid glycosides from Uticaria tomentosa (4, \0, II),Uncariaguaianensis (12), Guettardaplatypoda

(5, 13), and steroidal glycosides from

117 Tamils communis (6) and Asparagus cochinchinetisis (7) and flavonol glycosides from Mercurialis annua. (14). However the entire process consumes glycosides which are often very difficult to separate and purify and may be better employed in biological evaluation experiments. So the more recent approach to glycoside structural determination is a combination of ID and 2D NMR techniques which are very sensitive and non-destructive . The non-invasive nature of NMR methods allows easy recovery of the intact material by simple removal of solvent for successive biological evaluations. One dimensional NMR methods yield limited information for the determination of the complete structure and stereochemistry of a glycoside. In fact the ID ^H NMR spectrum of a glycoside shows only few recognizable signals such as anomeric protons ( 4.3-5.8 ppm) and methyl doublets of 6deoxy sugars (1.1-1.3 ppm) resonating at an uncrowded region of the spectrum, and the so called " structural reporter resonances" (8). The other proton resonances appear in a very small spectral width (3.0-4.2 ppm) and this creates overlap problems. These difficulties can be overcome by the use of high field techniques. Various two-dimensional NMR techniques enable us to identify the components of a glycoside without relying on analogy with any reference data: —The number of sugar residues and constituent monosaccharides are determined by a combination of COSY (two-dimensional homonuclear correlation spectroscopy), HOHAHA (2D homonuclear Hartman-Hann spectroscopy) and HETCOR (direct ^H-l^C heteronuclear correlation spectroscopy) experiments. COSY allows sequential assignment of most of the resonances of each sugar residue starting from the anomeric signal. If significant overlaps complicate assignment of all the resonances of an oligosaccharide , HOHAHA can be used to resolve the overlapped spectra into a subset of individual monosaccharide spectra which show on the same line signals corresponding to the different spin networks, generally from H-1 to H-6 of each sugar (15). HETCOR correlates all proton resonances to those of the corresponding carbons and can be used to determine interglycosidic linkage, taking into account the known glycosidation shifts. — The anomeric configurations and molecular conformations can be confirmed by NOESY (2D Nuclear Overhouser effect spectroscopy) or ROESY (NOE in Rotating frame) since crosspeaks observed in the spectra are relative to proton pairs that are close together. In general 1,3 diaxial and vicinal eq-ax proton pairs in pyranosyl rings produce intra NOESY cross peaks (e.g. residue crosspeaks for p-glucopyranosyl are observed among H-1, H-3 and H-5 and those for aglucopyranoside between H-1 and H-2) (8). — Linkage sites and sugar sequences can be determined by 2D NMR experiments such as COLOC (long range heteronuclear correlations), NOESY and ROESY which can show correlations between anomeric hydrogens and hydrogens linked to carbon supporting the glycosydation. For smaller quantities of compounds more sensitive inverse detected techniques are available, such as H M Q C (1H-1^C one bond correlation via heteronuclear multiple quantum coherence, analogous to HETCOR) and HMBC (proton detected heteronuclear multiple bond correlation spectroscopy) (15). The last provide, in addition to the intraresidue multiple bond correlations, interresidue correlations between the anomeric carbon and the aglycone protons.We follow this general strategy for the structural determination of triterpenoid saponins of Bupleurum fruticosum (16) andArdisia japonica (9).

118 2. TRITERPENE DERIVATIVES Triterpenes form an important group of natural products widely distributed in the plant kingdom. In the past decade new and complex triterpenes have been isolated from plant material. The chemistry and distribution of triterpenoids have recently been reviewed (17, 18), and many studies have been stimulated by a variety of biological properties exhibited by triterp>enes and triterpene glycosides. Recent developments in NMR spectroscopy such as high field ^H and ^^C NMR techniques have helped greatly in the elucidation of the structures and stereochemical assignments of triterpenoids. Our recent studies on triterpenes from Chinese and South American medicinal plants have provided interesting information about the chemistry of new triterpenes and steroids and their esters and glycosides. All the triterpenes isolated are based on a pentacyclic ring system and can be referred to the basic structure of ursane and oleane. Steroidal

glycosides are based on the

cyclopentaneperhydrophenanthrene ring and can be referred to the basic structure of spirostane or furostane. 2.1 POLYHYDROXYLATED TRITERPENES AND THEIR ESTERS U, tomentosa and E.japonica appear to be generous sources of novel highly hydroxylated urs-12en-28-oic acid derivatives. The 3p, 6a or 6p, 19a hydroxylation pattern is a common feature of the metabolites from both plants; other common hydroxylated sites are C-2 and/or C-23. A first investigation of the CHCI3 extract of U, tomentosa (19) gave three new ursolic acid derivatives: 1 3p, 6p , 19a - trihydroxy-urs-12-en-28-oic acid, 2 which has a formyl rather than methyl group at C-23 and 3, a nor-triterpene which has a hexomethylene group linked at C-4 (Fig.l). A successive investigation of the CHCl3-MeOH (9:1) extract led to the isolation of 4, which has a carbomethoxyl group at C-23, from a Sephadex LH-20 fraction treated with CH2N2 and then separated by HPLC (11). The isolation of 4 is of biogenetic significance in providing the missing link among 1, formyl derivative 2 and nortriterpene 3, and furthermore indicates the sequential biological oxidation at C-23 in triterpenes from U. totnentosa. Studies of the CHCI3 extract of E.japonica resulted in the isolation of a new derivative 3p, 6a , 19a trihydroxy-urs-12-en-28-oic acid, 5. besides the already reported maslinic acid 6, 2a-hydroxyursolic acid 7, 2a, 3a, 19a-trihydroxy-urs-12-en-28 oic acid or tormentic acid 8, 2a, 3p,19a, 23tetrahydroxy-urs-12-en-28-oic acid 9 and ursolic acid 10 ( Fig.l).

119

R

COOH

COOH

COOH

COOMe

CH

1 R=CH3 2R=CHO

R

OH

4 R=COOMe

^ HO- 1

S pCOOMc

HO R

Ri 5 R=H Rl=OH R2=H 9 R=CH20H Rl=H R2=OH

COOH

Fig. I Triterpene 1-4 from U. tomentosa and 6-10 from E. japonica. Since they display a number of structural features frequently encountered in triterpenic natural products, the spectral study of these metabolites may be of general interest to phytochemists.The structures of 1-5 were deduced from EIMS and detailed NMR analysis including 2D (HETCOR) and

120 ID (INAPT and NOEDS) techniques. Particularly useful were the INAPT experiments which delineated the correlation of each methyl carbon with carbon linked via long-range couplings. These experiments enabled us to assign unambiguously the methyl resonances, normally not reported in the literature regarding triterpenoids, showing in compound 1 that the Me-26 signal ( 61.06) correlated with C-8, C-9, C-14, and C-7,Me-25,(6 1.30) correlated with C-10, C-1, C-9, C-5; Me-27 ( 61.26) with C-14, C-8, C-15; Me-29 (6 1.23) with C-19, C-18, and C-20 and finally Me-30 (6 0.95) with C-20, C-19, C-21 resonances. HETCOR experiments also allowed us to distinguish between l^C NMR signals resonating very close specifically signals arising from C-1 ( 40.9 ppm) and C-7 (40.6 ppm); from C-16 (25.6 ppm) and C-21 (26.2 ppm) ; from C-5 (55.9 ppm) and C-18 (53.5 ppm). Table 2. Chemical shift of some key carbons in 1, 2, 3 and 4 (400 MHz, CD3OD/ CHCI3).

1

2

1I

1

4

Illl

3

Position 6C

6H

6C

6H

6C

6H

6C

6H

4

39.7

—

56.4

—

54.0

—

150.9

—

6

68.6

4.54

70.8

3.88

71.9

3.85

69.8

4.44

3

79.2

3.14

72.7

3.76

77.2

72.9

3.98

5

55.9

0.76

49.6

1.39

53.8

3.95 —

1.42

12.7

1.51

52.1 —

1.72 —

24

27.8

1.16

10.1

The A12,13 structures of 1-5 derive from the resonances of the sp2 carbons C-12 (CH by DEPT) at ca 129.5 ppm and C-13 (C by DEPT) at ca 140.2 ppm; this allowed us to distinguish an urs-12-ene fron an olean-12-ene derivative (19). A signal at ca 181.0 ppm (C) appear suggested the presence of a carboxyl group at C-17 in an urs-12-ene skeleton. The presence of two secondary hydroxyl groups and a tertiary hydroxyl group on the basic ursolic structure was derived from NMR and EIMS data. The latter spectra give typical fragments resulting from retro-Diels-Alder cleavages of an ursolic acid which possess a hydroxyl group on ring D or E, and two hydroxyl groups on rings A and/or B. The location of the tertiary hydroxyl group at C-19 is in agreement with the absence of one of the two methyl doublets characteristic of ursane derivatives and with the H-18 resonance at 6 2.59 (s) in the ^H NMR spectrum (in ursolic acid derivatives H-18 resonates as a doublet at ca 6 2.2 ppm) (2). Furthemore the C-19 carbon, supporting a hydroxyl group, resonates at ca 73.9 ppm. The location of two secondary hydroxyl groups at C-3p and C-6p in 1 gives rise respectively to a proton signal at ca 6 3.1 (dd, J=11.5 and 4.5 Hz) ascribable to an axial proton at C-3 and an unresolved signal at 6 4.51 (m, Wl/2 = 6.0 Hz) ascribable to an equatorial H-6 proton. It is interesting that the axial -OH at C-6 has a strong deshielding effect, due to 1,3 diaxial interaction, on Me-24 (6 1.16) , 25 (6 1.30) and 26 (6 1.06) resonances with respect to ursolic acid as a model (6 0.88, 0.99, 0.81, respectively).

121 In compound 5 Me-24, 25 and 26 resonances are virtually unshifted in relation to those of ursolic acid because of the equatorial configuration of the -OH group at C-6 . Furthermore in compound 5, the H-6 signal appears as a ddd (J=10, 10 and 3 Hz) at 6 3.95, thus confirming its axial configuration. It is interesting to note that the -OH group at C-19 induces a downfield shift of the H-16 axial proton which appears at ca 6 2.62 in 1 as an isolated ddd with J=13.5, 13.5 and 4.5 Hz, whereas in ursolic acid it overlaps the other signal at 6 2.0. This shift supports the 19a-OH stereochemistry of triterpenes and is only compatible with a cis-stereochemistry of the ring D/E junction. The NMR spectra of 2 and 4 contain one fewer methyl singlet signal and one more aldehydic signal (s 9.52 in the ^H NMR and 209.3 in the l ^ c NMR) in 2 and a carbomethoxyl signal in 4 (179.5 ppm). The formyl and carbomethoxyl groups at C-23 give rise to downfield shifts of C-4 and C-6 and upfield shifts of C-3, C-5 and Me-24 in 2 and 4 with respect to 1 (Table 2). Similar shifts were reported for gypsogenin ( which has a 23 equatorial -CHO group) and gypsogenic acid (which has a 23 equatorial -COOH group) by comparison with oleanolic acid and methyl pomolate respectively (19,II). Very different A6's were descibed for ilexgenin A methyl ester which has a 24 axial -COOMe. It is therefore possible to distinguish the equatorial or axial -CHO and COOH groups at C-4 via the chemical shifts of key carbons neighboring the substitution positions. The 1H NMR c.s. of H-6, H-3, H-5 and H-24 are also affected by the presence of a —CHO or a — COOCH3 group at C-23. The stereochemistry at C-4 of the -CHO in 2 group was confirmed through NOEds experiments. Strong NOE effects can be observed between the aldehyde signal (6 9.52) and H-5a (6 1.39), H-6a (6 3.88) and H-3a (6 3.76). Compound 3 is a nor-triterpene which shows in the ^H NMR spectrum a complex signal at 6 4.72 (t,J=3.5 ) and 5.27 (brs ) correlating with an sp2 carbon at 104.5 ppm by HETCOR, typical of CH2=C group. The location of the CH2= at C-4 (150.9 ppm) justifies the disappearances of the Me23 and Me-24 signals and the shift values of the C neighboring C-4 position (Table 2). A further investigation of the CHCI3 extract of E,japonica provided a series of triterpene ester derivatives 11-14 (Fig. 2). Compounds 11 and 12 are triterpenes with 2a,3p,19a. 23 tetrahydroxy-urs-12-en-28-oic acid moiety which links a trans-para-coumaroyloxy or a a s -para-coumaroyloxy moiety, respectively, through an ester bond. The presence of 2a, 3p -OH substitution pattern is derived from the characteristic signals ascribable to H-2 (6 3.65 ,ddd, J=10, 13 and 3.5 Hz) and H-3 (6 2.95, d, J=10 Hz). Compounds 11 and 12 exhibited closely comparable spectroscopic data except for signals centered at 6 5.90 and 6.92 (J=13 Hz) in 12 corresponding to a cisiconjugated olefinic system and at 6 6.40 and 7.62 (J=16 Hz) in 11 ascribable to a trans conjugated olefinic system. The chemical shift* of Me-24 and CH20-23 testify that the ester linkages are between the C-23 position and the picoumaroyl moieties. Compound 13 has the same 2a, 3p, 19a, 23 hydroxylation pattern as 11 and 12 but links a caffeoyl moiety at C-3. Compound 14 is a derivative of rotundic acid which links a trans-p-coumaroyloxy moiety at C-3.

122 The ester bond at C-3 shifts downfield by 1.6 ppm the H-3 resonance of 13 (6 4.63 , d, J=10 Hz) with respect to tormentic acid 8. The H-3 resonance in 14 (6 4.45, dd, J=11.5 and 4.0 Hz) is also shifted downfield with respect to rotundic acid.

COOH

COOH

CH2R

CH2OR

13R=H Ri=OH V.2'

11 R

OH

)H

bH 14R=OH R,=H Rt

OH

12 R=

•H

\

/

Fig.2 Tri terpenoid esters from Eriobotrya japonica In the ^^C NMR spectra the ester linkages at C-3 caused a downfield shift of the C-3 ( 6 82.9 in 13) and upfield shift of the C-4 (6 39.9) resonances relative to model compounds such as 7, 10 and tormentic acid. 2.2 TRITERPENIC GLYCOSIDES U. tomentosa, U, guaianensis and G, platypoda are plants of the Rubiaceae family used in traditional Brazilian and Peruvian medicine as potent anti-infiammatory agents. While a number of alkaloids displaying a pronounced enhancement of phagocytosis have been reported in the genus Uiicaria (20), no phytochemical or pharmacological work has previously been done on the non-alkaloidal constituents. These three species contain, in addition to other metabolites, quinovic acid glycosides which have common structural features including a 3p-ol-urs-12-en-27,28-dioic acid aglycone and a sugar moiety made up of one to three monosaccharides. Four groups of quinovic acid glycosides can be discerned including glycosides with the sugar chain attached at C-3 of the aglycone by a glycosidic linkage.; glycosides which link sugars at C-28 by an ester bond and bis-glycosides with a C-3, 28 or C-3,27 glycosilation pattern. The last group of compounds is rare in nature and has been found only in U.tomentosa

(4, 10, l\)U.

guaianensis

(12) and G. platypoda (5,13); they seem to be

123 characteristic metabolites of these two genera of Rubiaceae. Some quinovic acid glycosides with a C3 and/or C-28 glycosylation pattern have been isolated from Cinchona andGuettarda sp. (21) of the same family Rubiaceae. In the course of our work, we have isolated eight new quinovic acid glycosides 15-22 (Fig. 3) from U. tomentosa , two of which (compounds 15 and 22) are also present in U, guaianensis. Seven new quinovic acid glycosides 25-31 (Fig.4) were isolated from G, platypoda and four glycosides 15 and 22-24 from U, guianensis : 23 is a new natural compound while 24 was previously isolated from G. platypoda. Fig. 3 and 4 illustrate the general structures of these compounds. The oligasaccharide portion includes glucose, fucose (6-deoxygalactose), quinovose (6-deoxyglucose) in U. tomentosa and U, guaianensis; glucose, fucose and rhamnose (6-deoxymannose) are present in G.platypoda. All sugars are in their pyranosyl forms with p-anomeric configurations (a in the case of rhamnose). Glycosides 15, 21,22 and 25 from Fig. 3 and 4 have a glucose unit linked through an ester bond at C-27 of the aglycone whereas the other triterpenic glycosides 16,18,19,26,28,29 and 30 have a C-28 ester linkage. Compounds 17,23,24 and 27 possess a sugar moiety linked at C-3 and compound 20 a sugar moity linked at C-28. C-27 and C-28 glycosyl esters of quinovic acid are distinguishible by diagnostic resonances of some key carbons ( C-12, C-13, C-14 and -COOH at C-27 and at C-28) (Table 3) and by some hydrogens such as Me-26 and H-12. In a C-27 glycosyl ester like 15,21, 22 and 25 C-12 resonates at ca 130.9 ppm, C-13 at ca 133.4 ppm and C-14 at ca 57.5 ppm.Thus C-12 is considerably deshielded (+1.8 ppm) whereas C-13 and C-14 are shifted to higher fields (-1.8 and 1.4 ppm, respectively) with respect to quinovic acid derivatives when both C27 and C-28 are unsubstituted as in 17, 23, 24 and 27. Table 3. Characteristic chemical shift of key carbons in quinovic acid derivatives. Carbons C-12 C-13 C-14 C-27 C-28

Quinovic acid derivatives 129.1 ppm 135.2 ppm 58.9 ppm 179.5 ppm 182.0 ppm

C-27 glucosylesters of quinovic acid 130.9 ppm 133.4 ppm 57.5 ppm 178.1 ppm 182.0 ppm

C-28 glucosylesters of quinovic acid 129.9 ppm 134.7 ppm 59.2 ppm 179.5 ppm 178.5 ppm

In a C-28 glycosyl ester like 16, 18,19. 26, 28, 29 and 30, the chemical shifts of C-12 ( 129.9 ppm), C-13 (134.7 ppm) and C-14 ( 59.2 ppm) are virtually identical to those found in 17. The esterification of the C-27 carboxyl groups causes a wide deviation of the chemical shifts of the sp^ carbons (C-12 and C-13) and of C-14 which are in agreement with those reported in quinovic acid and its dimethyl ester derivative (22). Furthermore if an unsubstituted -CCX)H usually resonates at 179.5 ppm (C-27) and 182.0 ppm (C-28), the esterification shifts upfield both C-27 ( 178.1 ppm) and C-28 (178.5 ppm). It is also possible to distinguish C-27 glycosyl esters of quinovic acid via the Me-26 and H-12 resonance in the ^H NMR spectra. Whereas the other aglycone protons are not modified, H-12 is shifted downfield by +0.05 ppm ( from 6 5.59 to 6 5.64) and Me-26 by +0.03 ppm (from 6 0.89 to 6 0.92) in a C-27 glycosyl ester with respect to the C-28 glycosyl ester.

124

000R3

R.O^

R2

OH

16

OH

^OH

OH

^OH

OH

^

HO OH

^OH

OH

HO 19 OH

^ ^ ^ ^ OH

20

^

OH

OH

HO

.

-

H

^

H

OH OH 2

OH

H O - ^ ^ ^ ^

Fig.3 Quinovic acid glycosides from U, tomentosa e t/. guaianemis

125

I^

Rj

H

H

27 OH

H

26 OH

OH

^OH

H

25 OH

H

H 28

OH OH

29

H

H

^OH

.

^OH

OH

30

H OH

OH

31

HO A^O Jf^^I^i:^^ HO--^--^r^

a4

H

H

OH

Fig.4

Quinovic acid glycosides from G. platypoda On basic hydrolysis the C-27 and C-28 glycosyl esters of quinovic acid give the same products with C-27 and C-28 unsubstituted -COOH; when acetylated and treated with CH2N2 they give peracetyl monomethyl ester derivatives which have very similar spectra except for the signal due to -<XIH3 group (3.64 vs 3.63ppm, respectively) (4,5). The identity of the sugar linked at C-3 and/or C28

126 27 was established by acid methanolysis followed by GLC analysis of the TMS methyl glycosides; FAB-MS, iH andl^c NMR data.

COOR, O H . OH OH

OH 32

-CHiOH HO' HO

33

-CHjOH

34

-OOOH

OH H r—OH

HO"

35

-OOOH

OH H r-OH

y—OH 36

-CHjpH

HO-

OH

HO' HO

OH OH

37

-CHjpH

HO' HO

Q OH

OSophorosc

COOH

OH 'CHzOH

Fig 5 Oleanolic glycosides from C. arvensis and A. officinalis The saipe approach was used for the structural determination of five oleanolic acid glycosides from Calendula arvensis 32-36 (23) and 3p, 2ip, 23 trihydroxy-olean-12-en-28-oic acid bisglycoside 44

127 from Aficliusa officinalis (24) (Fig. 5), plants used in Italian folk medicine as anti-inflammatory and antipyretic agents.

42 Ri=OMe 43 Ri=OH

R=

•^1^

38 39 40

Ri=-CHO Rl=-CH(OCH3)2 Ri=CH3

cJ^i

R2=CH3 R2=CH3 R2=-COOH

Fig. 6 Oleanolic glycosides from A,japonica and B.fruticosum Compound 32-36 have a 3 or a 3,28 glycosylation pattern, compound 44 has a 3,21 glycosylation pattern. The sugar moiety linked at C-3 of the aglycones is made up of p-D-galactopyranosyl (l->3)P-D-glucopyranosyl in 32 and 33, p-D-galactopyranosyl (l->3)-p-D-glucopyranosyl uronic acid in

128 34 and 35, p-D-xylopyranosyl in 44. Compound 36 has a more complex trisaccharide chain with an additional glucose unit linked at C-4 of the inner glucose. In addition compounds 32 and 34 link through an ester bond a glucose at C-28, and compound 44 links a sophorose (P-D-glucopyranosyl(I->2)-p-D-glucopyranoside) at C-21. A more modem structural determination approach was used for triterpenoidic glycosides 38-40 (Fig. 6) from Ardisia japonica (9), a Chinese remedy for contusion, rheumatic and neuralgic diseases and 41-43 (Fig. 6) from Blupeurum fruticosum (16), a Chinese anti-inflammatory agent, whose complex and branched oligosaccharides moieties were determined by ID and 2D NMR techniques. The aglycone is cyclamiretin A in 38; 3p,16adihydroxy-13,28 epoxy-30,30-dimethoxy oleane, a new dimethyl acetal at C-30 of cyclamiretin A in 39; and 3p, 16a, hydroxy-13,28 epoxy-olean-29 oic acid in 40. The sugar moiety, linked at C-3 of the aglycone in 38-40, is composed of arabinose, rhamnose and two units of glucose. The nature of the sugars and the positions of the interglycosidic linkages were determined by a combination of COSY, HOHAHA, ROESY and HETCOR experiments (TabA). Table 4. NMR data of the oligosaccharide moieties of 38 in CP3OD. Position

Arabinose 6C

1

105.30

Glucose 1 6C

Rhamnose 6C

104.53

103.71

101.70

Glucose 11 ' 6C

2

79.55

76.40

77.00

71.80

3

72.38

77.70

79.10

72.20

4

76.68

77.90

71.40

74.20

5

64.00

77.70

77.90

70.30

62.90

16.75

6

1

62.90

1

COSY 2md HOHAHA revealed spin correlations for rhamnose and glucose units between each pair from H-1 to H-6, and for arabinose between each pair from H-1 to H-4. The coherence transfer to methylene H-5 of arabinose is not observable because of the small JH4-H5- HETCOR correlating all proton resonances to those of the corresponding carbons, gives the position of the interglycosidic linkages. Thus two sequence are possible: . glucose " Aglycone -

rhamnose

Ara

\ A rhamnose yr rhamnose Aglycone

Ara

\

glucose — ^ rhamnose

129 Roesy experiment (Table 5) allowed differentiation between hypotheses I and II showing spatial correlations between anomeric protons and protons linked to glycosylated carbons.

Table 5. Selected data from ROESY experiments of 38 in CE^OD. Connectivities observed between: H-proton

ROESY (iH-anomeric)

3.17(H-3Aglycone)

4.52 (H-1 Arabinose)

4.09 (H-4 Arabinose)

4.70 (H-l Glucose I)

3.87 (H-2 Arabinose)

4.65 (H-1 Glucose II)

3.44 (H.4 Glucose II)

5.31 (H-4 Rhamnose)

Chemical shifts, multiplicity of the signals, absolute values of the coupling constants and their magnitude in the ^H NMR spectrum as well as NMR data indicated the p-configuration at the anomeric positions for both glucopyranosyl units (JHI.H2~^*^ ^^) ^^^ ^^® a-configuration for the rhamnopyranosyl unit (JHI-H2 "^-^ ^"^J* ^" ^^ibinose in pyranose form was evident from ^^C NMR data, but no further support for the anomeric configuration of the L-arabinopyranose unit could be drawn from the ^H and ^^C nmr data. In fact, the value of its JHI-H2 coupling constant (5.2 Hz) was midway between that reported in the literatura for methyl-p-L-arabinopyranoside (4 Hz) and methyl-a-L-arabinopyranoside (8 Hz). The value of this coupling constant has been reported to be not diagnostic on its own, because of the high conformational mobility of arabinopyranosides (^Ci<->^C4). Evidence supporting an a-L-arabinopyranoside configuration in rapid conformational exchange was obtained from ROESY experiments . Nuclear Overhauser effects were observed from Al to A2 and Alto A3 as expected for ^04 and ^Cj conformations respectively. The nOe A1-A3 would be unlikely for either ^04 or^Cj p-Larabinopyranosides. An nOe was also observed between A l and A5 as expected for an a-Larabinopyranoside in a ^Cj conformation (9). Further evidence suggestig the p-configuration at the anomeric center of the sugars were the molecular rotation values considered in the lights of Klyne's rule. Compounds 41-43 have, as the aglycone, 16p, 23 dihydroxy-13,28epoxy-olean-llen-3pol (saikogenin F), 16p, 23,28-trihydroxy-lla-methoxy-olean-12-en-3Pol, and U a , 16p. 23, 28 tetrahydroxy-olean-12-en-3p ol, respectively.

130

U>...

Arabinopyranosyl

(-OH group were not reported)

For the sugar moiety COSY and HOHAHA experiments revealed all spin correlations from H-1 to H6 of a fucose and two glucose units and in combination with HETCOR allowed us to establish that the two p-D-glucopyranoside are terminal and ^D-fucopyranosyl is directly linked at the aglycone and 2,3 glycosylated. HMBC experiments showing correlations between fucose H-1 and aglycone C-3, between H-1 of one glucose unit and C-2 of fucose, and between H-1 of the other glucose and glucose 2 "^ C-3 of fucose confirmed that the sequence was aglycone —> fucose^ {16). 3 -^ glucose 2.3 STEROIDAL GLYCOSIDES Tamils communis is a temperate species of the family Dioscoreaceae used in traditional southern Italian medicine as anti-inflammatory drug. Asparagus cochinchmensis (Liliaceae) is used as a ionic drug in Chinese medicine. Dioscoreaceae are generally recognized as a source of steroidal sapogenins such as diosgenin, an intermediate for the synthesis of steroidal drugs; several Asparagus sp are used in Indian medicine and contain furostanol oligosides with a (25 S) configuration and a saturated pentacyclic system (7). From the CHCl3:MeOH (9:1) extract of the rhizomes of 7. communis (25) we have isolated two spirostane triglycosides, 45 dioscin and 46 gracillin, which have diosgenin (spirost-5-en-3p-olo) as the aglycone. At C-3 diosgenin links trisaccharide chains made up of two rhamnose units and a glucose unit in 45 and of two glucose units and a rhamnose unit in 46 {(bis-a-L-rhamnopyranosyl(l—>2 and 1—>4)-p-D-glucopyranoside in 45 and a-Lrhamnopyranosyl-(l—>2)-[p-D-glucopyranosyl-(l—>3)]-p-D-glucopyranoside in 46}. The biogenetic precursors of 45 and 46 with furostanol skeleton, 47 (25 R) methylprotodioscin, 49 methyl protogracillin as well as the (25S) epimer of 47, and the methyl protodioscin 48 are present in the MeOH extract (6). Methyl protodioscin 47 is also present in the aqueous extract of the roots of A. cochinchinensis together with the corresponding 20 (22)-dehydro, 20-demethoxy analogue (pseudoprotodioscin 50), recently isolated for the first time from a Palma and with a new glycoside 51 3-0-[a-L-rhamnopyranosyl-(l—>4)-p-D-glucopyranosyl-26-0-(P-D-glucopyranosyl)-25 Rfurosta-5,20-dien-3p,26 diol (Fig. 7). The aglycone is a furost-5-en-3p, 26 diol in 47,48 and 49

131 but in 50 and 51 it is a furosta-5, 20-dien-3p-. 26 diol. It is interesting that all glycosides isolated from A. cochinchinensis are (25 R) A^ or (25 R) A^-^O furostanosides. The negative FABMS spectrum of 47 (methyl protodioscin) gives interesting fragmentations, showing a quasimolecolar anion at m/z 1061 [M-H]' and an intense peak at m/z 1027 due to the elimination of methanol from the 22-methoxy-furost-5-en-3p,26-diol derivative to form A^*^^ derivative, identical to 50. This elimination seems to be a characteristic behaviour of oligofurostanosides and can also be observed in FD (26) and FABMS spectra in positive ions (25). (25R) and (25S) epimers of A^ furostanosides, like methyl protodioscin 47 and methylprotoneodioscin 48, exhibit characteristic differences in the resonances of the Me-21, Me-27 and Me-18, adjacent to the asymmetric centre (Table 6). In fact, in (25 S)-epimers the Me-21 resonance is shifted upfield by 0.02 ppm with respect to (25 R)epimers while Me-27 and Me-18 are shifted downfield by 0.02 ppm and 0.01 ppm respectively. Further evidence is provided by acidic hydrolysis that produces the aglycones yamogenin from (25 S)-epimers and diosgenin from (25 R)-epimers. The configuration at C-25 of A5»20 furostanosides like 50 and 51 is always definible by ^H NMR spectral data of Me-21,27,18 methyl signals. Pseudoprotodioscin 50 and pseudoproneodioscin 52 prepared by acetic acid treatment erf" 47 and 48, gave small but significant differences in the resonances of the Me groups around the C-25 centre (Table 6). Table 6.Characteristic chemical shift values of some hydrogens in (25S) and (25R) epimers of steroids. 5 6 (25R) methylprotodioscin (25S) methylprotoneodioscin Position 48 47 Me-21 Me-27 Me-18 Me-21 Me-27 Me-18 Me-21 Me-27 Me-18

1.03 (3H,d,J=6 Hz) 0.98 (3H,d,J=6 Hz) 0.87 (3H,s) Diosgenin 0.97 (3H,d,J=6.6 Hz) 0.79 (3H,d,J=7 Hz) 0.79 (3H,s) (25R)pseudoprotodiosci n 50 1.63 (3H,s) 0.98 (3H,d,J=6 Hz) 0.74 (3H,s)

1.01 (3H,d,J=6 Hz) 1.00 (3H,d,J=6 Hz) 0.88 (3H,s) Yamogenin 1.01 (3H,d,J=6.7 Hz) 1.09 (3H,d,J=7 Hz) 0.79 (3H,s) (25S)pseudoprotoneodioscin 52 1.62 (3H,s) 1.00 (3H,d,J=7 Hz) 0.75 (3H,s)

l^C NMR data of C-3, by comparison with C-3 of the aglycones, clarify that the sugar chains are linked at this position. The structure elucidation of the carbohydrate moieties are deducible in the usual manner ( methanolysis , 1H and l^C NMR data, FABMS). The linkage of another unit of glucose at C-26 in furostanol glycosides is confirmed by enzymatic hydrolysis with p-glucosidase which determined the splitting of the terminal p-D-glucopyranose from C-26 position . So 50 affords a furosta-5,20-dien-3p,26 diol derivative and 47 gives a spirost-5-en-3p-ol derivative (dioscin 45).

132 ^CH3

OH

45 R=

V"^"

46 R=

"s;^*^°

HO-^^

O-^T^-^—OH HO—^

Ri

R 47

\

R3

R2

R4

-CHj

-OCH3

-CH,

-H

-OCH3

-H

-CH3

-CHj

-H

H

V-OH

OH

.H

48

^?8. 50 HO-'^

5 2

OH O.

H' H( V

OH

49

\)H

OH

N

r-OH

\

-CHj

QH

-OCH3

51

Fig. 7 Steroidal glycosides from T. communis and A. cochinchinensis

-H

133 3.BIOLOGICAL ACTIVITIES Until a few years ago relatively little was known about the metabolites of medicinal plants but in the intervening period a number of active compounds have been isolated . Isolation and identification of the active constituents is essential for the study of their toxicity, stability and effects on metabolism and physiology, for dosages purposes and for structure-activity investigations.The choice of pharmacological screenings of plant metabolites must take into account their uses in traditional medicine, ethnopharmacological literature and the structural relations between isolated compounds and well known drugs. 3 1 ANTIVIRAL ACTIVITY As a part of our investigation into the biological activities of South American and Chinese plants used in traditional medicine, we tested a series of plant metabolites with potential antiviral properties. They were screened test for new active compounds and to find out how small structural differences in these compounds could influence their activities. Whereas various biological activities have been described for triterpenoids, steroids and their glycosides (27), little is known about their antiviral activity: only glycyrrhizin and glycyrrhetinic acid and a few other triterpenoids have been found to inhibit the replication of some DNA viruses. The only one which has been extensively studied is glycyrrhizin, a component of the aqueous extract of licorice root (Glycyrrhiza glabra). This saponin contains two glucuronic acid residues linked at C-3 of 3-p-hydroxy-ll-oxo-18p-olean-12-en-30-oic acid. Glycyrrhizin was found to inhibit the replication of some DNA and RNA viruses in vitro (28) and to have therapeutic and prophylactic effects on chronic viral hepatitis (29). More recently it has been shown to exert an antiviral action in vitro against varicella-zoster virus (VZV) (30) and the human immunodeficiency virus (HIV) (31). For the structural relations between glycyrrhizin and triterpenoids, triterpenic esters, triterpenic glycosides and steroidal glycosides isolated in the course of our work, we studied extensively the possible antiviral activity of the above compounds against two RNA viruses: two enveloped minusstrand RNA viruses (vesicular stomatitis virus, VSV or Sindbis virus, SNV) and a naked plus-strand virus (rhinovirus type IB , HRV IB). These compounds were tested against HRV and VSV infection on HeLA and CER cell cultures respectively. The antiviral activity was determined as % of inhibition of cytopathic effect (CPE) for HRV IB and % of plaque reduction for VSV.The activities are shown in Tables 7-9. An inhibitory effect against VSV infection is evident for quinovic acid glycosides 1520, isolated from U, tomentosa (10), and 25-27 from G. platypoda (10) (Table 7), for oleanolic acid glycosides from C. arvensis 32-37 (34) (Table 8) and for furostane glycosides from T. communis and A. cochinchinensis 45-50 (Table 9) (32), although to a different extent. Quinovic acid glycosides 15-27 are active although at concentrations relatively close to the toxic dose (Tox C50) for CER cell morphology and growth (Table 7). The most active compound is 17 with both unsubstituted -COOH groups at C-27 and C-28 on the basic quinovic acid structure. No relationship was found between the number of sugars and anti VSV activity. The presence of a free C-27 -COOH seems to be important. The nature of the sugar moiety also affects the activity; in fact the only differences in structure between 18 and 19 are the presence of quinovose (in 18) or fucose (in 19).

134 The oleanolic acid glycosides 33, 35, 37, all of them showing the free C-28 -COOH group, are less active and more toxic at maximum non-cytotoxic concentrations (nontoxic concentrations range from 4 to 12 ng/ml). In accordance with general structural activity relationships of the above glycosides, the corresponding derivatives 32, 34 and 36 with an additional sugar moiety at C-28 , are less toxic to the cell and give from 70% to 100% of inhibition at concentrations of 100 ^ig/ml (Table 8). Almost all of these quinovic and oleanolic acid derivatives are inactive against HRV IB infection in HeLA cells.

Table 7. Antiviral activity of quinovic acid derivatives 15-20 and 25-27 against HRV-IB and VSV infection. Compound 1 Concentration ^g/ml 15 40 20 4 16 40 20 4 17 60 20 4 18 40 20 4 19 40 20 4 20 60 20 27 25 26

^*

100 50 |20 100 20 100 20

VSV

j TC50

% plaque

45 36 19 50 33 10 100 40 0 90 46 19 50 37 0 64 37 0 100 75

lo64

HRV IB 2 1 %CPE 1

100 inactive 80 inactive 100 inactive 80 inactive 80 inactive 100

30 50

150 150

5 85 20

Concentration |xg/ml 1

150

100 75

inactive

10

64 5

inactive

20 50 4 10 |0.8 10 ^The maximum non toxic concentrations for HeLA cells of compounds 20 and 26 are 60 ng/ml and 1(X) ng/ml, respectively. 2 CPE= cytopathic effect.

135 Table 8. Antiviral activity of oleanolic acid derivatives 32-37 against HRV-IB and VSV infection. Compound Concentration (jig/ml) VSV % plaque reduction HRV-1B%CPE inhibition 0 100 100 32 0 21 20 0 0 4 0 18 4 33 0 0 0.8 0 0 0.16 NT2 100 100 34 21 20 75 5 4 50 0 0.8 0 25 20 12 35 25 5 4 0 0 0.8 35 70 100 36 0 18 20 0 15 4 25 10 4 37 0 0 0.8 ^ Compounds 33,35 and 37 were tested starting from the maximum non-cytotoxic concentrations. ^NT= not tested It is interesting to note that only 20 and 26 , both containing two glucose units and the free C-27 COOH , as well as 34, containing a glucuronic acid and an esterified C-28 -COOH, reduce the cytopathic effect by 50% at 30,20 and 4 jig/ml, respectively. Thus the infection process by enveloped virus VSV is generally more sensitive to the triterpenic glycosides than that of naked viruses like HRV. In view of the interesting antiviral activity of triterpenoids we extended the antiviral assays to triterpenes 1,8 and 9 and their esters 11-14 isolated from E. japonica (3). The compounds were tested against HRV IB , Sindbis virus (SNV) and human immunodeficiency virus (HIV-1) in Ohio Hela , Hela S3 and C8166 cells, respectively. None of the compounds was effective against SNV when tested at the highest nontoxic concentration. Only compound 13, which has a caffeoyl acid residue in the molecule, is active against HRV IB infection causing a 50% reduction of CPE at 20 mg/ml. At a concentration of 4 |iig/ml its inhibitory effect is -25%, and a complete absense of activity is observed at 0.8 jig/ml. None of the compounds is active against HI V1 but slight differences in the cellular cytotoxicities were observed (data not shown). Thus the single triterpenic skeleton and cinnamic acid residues, rather than sugar residues, render the molecules ineffective against both enveloped and naked viruses. However the sugar moiety seems to be essential to the activity. Only the caffeic acid derivative 13, containing two -OH groups on its aromaticring,exhibited a certain antiviral effect against HRV IB. The well known spirostane triglycosides 45 dioscin and 46 gracillin, which have diosgenin as aglycone, are cytotoxic above 4 and 20 jig/ml respectively , and do not give interesting results below these concentrations in the antiviral screenings (Table 9) (32).

136 Table 9. Effect of spirostane glycosides 45 and 46 and furostane glycosides 47-50 on virus infected cells. VSV Compound

45 46

47 48 49

50

Concentration ^g/ml

14

20 4 0.8

10

% plaque reduction

41 13 0

HRVIB Concentration % CPE M,g/ml reduction

14

20

0 0

100 20 4 100 20

100 16 8 100 0

20 4

25 0

20

50

100 50 20 4 100 20

100 40 35 0 100 0

20 4

25 0

20 4

25 0

4 0.8

25 0

The furostanol tetraglycosides 47-50 (Table 9) are less cytotoxic than the spirostane glycosides and were tested at concentrations of 100 \ig/m\ on CER cells and 20 ng/ml on HeLA cells. 47 (methyl protodioscin ) and 49 (methyl protogracillin) with 25R configurations are quite active against VSV and much less agziinst HRV, while 48 (methyl protoneodioscin ) with 25S configuration, is active against HRV; its cytophatic effect is reduced to 25% at 4 ^ig/ml. It must be noted that 25R and 25S epimers give an inverted intensity of action against enveloped and naked viruses. The tetraglycoside 50 (pseudoprotodioscin) has a 25R configuration like 47 and 49 but it is unsaturated in 20 (22) and therefore lacks the methoxy group at C-22. This structural change may explain why 50 is less active than of the less 47 and 49 against VSV (32). All together the above results indicated that infection by enveloped virus ( VSV o SNV) is generally more sensitive to the steroidal and triterpenic glycosides than infection by naked virus (HRV). Similarly glycyrrhizine was found to inhibit the growth of several enveloped viruses, among which VSV (100% inhibition of CPE at 5.33 mg/ml) (28), HIV (IC50= 404 ng/ml) (31), and VZV at 50% of inhibitory dose 584 jig/ml (33), but not of naked viruses ( poliovirus type 1) (28). However it must be observed that the above concentrations are much higher than those required by oleanolic or quinovic acid glycosides to completely reduce VSV plaque formation. Therefore triterpenoid saponins from U. tomentosa, G. platypoda and C. arvensis seem to possess higher antiviral potential than glycyrrhizine. Glycyrrhizine was found to exhibit no direct inactivating effect on virus particles (28,31), no interferon inducing activity in vitro (31) or in vivo (33), which suggests an action on one or more steps of viral replication cycle.

137 Although the mechanism of action of the above glycosides has not yet been elucidated, they are likely similar. The hypothesis that also the antiviral effect of triterpenoid glycosides from U, tomentosa, G, platypoda and C, arvensis is probably not mediated by a non-specific detergent-like action on virus particles, is supported by two lines of evidence: a) the compounds exhibit different degrees of inhibition towards VSV infection and b) they display a moderate to low inhibition against naked virus infection. 3.2 ANTI-INFLAMMATORY ACTIVITY Uncaria tomentosa is is a Peruvian Rubiacea commonly known as "Una de Gato". In traditional Peruvian medicine its root bark is dried, powdered and boiled in water and this red-hot aqueous extract is reported to be effective in the control of arthritis, gastritis, skin diseases and cancer. A number of alkaloids producing a pronounced enhancement of phagocytosis were isolated by Wagner et al. (20) As a part of our continuing search for new biologically active metabolites from U, tomentosa we preceded to a bioassay-directed fractionation of the extracts from U, tomentosa. The extracts and fractions have been bioassayed by the carrageenan induced edema test on rat paw. This allowed us to identify a quinovic acid derivative with a C-3, 27 glycosidation pattern, 21, as one of the active anti-inflammatory principles of U. tomentosa. (11). The root bark of U, tomentosa was extracted with solvents of increasing polarity (Petroleum ether, CHCI3 , CHCl3-MeOH 9:1, MeOH and H2O). Each extract was tested orally using the carrageenan induced edema in rat paw administrering 2g/Kg of dry bark. In this bioassay the CHCl3-MeOH 9:1 extract (50 mg/Kg) and the H2O extract (84 mg/ Kg) displayed appreciable activity (respectively 69.2 and 41.3% inhibition of maximum edema, at3h,) while the other extracts are not significantly active. Separations of the crude CHCls-MeOH 9:1 extract by Sephadex LH-20 column yielded five main fractions I-V which were tested at doses equivalent to 2g/Kg of dry bark under the same experimental conditions. The most active fractions, I (4.2 m/Kg) and III (2.3 mg/Kg), inhibited edema by 46.8 and 37.4 % (3h) respectively, while the inhibitory rates at 3h of fractions II (25.2 mg/Kg), IV (5.9 mg/Kg), and V (9.1 mg/Kg) were 7.54, 25.6, 26.8%, respectively. By means of RP HPLC compounds 15-22 together with the alkaloid 5a-carboxystrictosidine were isolated from fraction I. The triterpenes oleanolic acid, ursolic acid 10 and 1-4 were isolated from fraction III . All pure compounds, tested by the same procedure, were inactive. Since the tested doses were very low (ranging from 0.04 mg/Kg to 0.51 mg/Kg), a further set of experiments was performed at higher doses (0.014 mmol of each compound/Kg, equivalent to the ED50 of indomethacin); however no significant inhibition of the edema was seen, although the control, indomethacin ( 5mg/Kg po), was active.

138

1.4 1.2 3

1.0

a

0.8

I 0.6 a

0.4-1 0.2 H 0.0 5

6 Hours

Fig. 8 Effects of compounds 1 (A-A), 21 (A-A) and oleanolic acid ( • - • ) at dose of 20 mg/Kg po on carrageenan edema. (O-O) Control. In the last set of experiments compound 21, oleanolic acid, and compound 1 were tested at the highest doses available from the isolated samples, which in the previous screenings had weak but non significant anti-inflammatory effects. As shown in Figure 8 only 21 at 20 mg/Kg caused a 33% of inhibition of the inflammatory response at 3h, while oleanolic acid and the triterpene 1 were inactive at that dose. On the basis of these results, we suggest that the strong anti-inflammatory activity of the extracts and fractions of U, totnentosa may be due to the presence of a combination of compounds. It is possible that some compounds, like 21, have an intrinsic anti-inflammatory effect while others may act synergically or as vehicles enhancing biological activity. However we cannot deny that the activity of the extracts and fractions may be due to a very minor metabolite not isolated or to the

139 isolated metabolites but at higher doses. In fact oleanolic acid is reported to have anti-inflammatory effect in the same test but at higher doses (40 mg/Kg) and by ip or local administration (35). 3.3 HYPOGLYCEMIC EFFECTS The leaves of Eriobotrya japonica (Rosaceae), a small tree commonly known as "loquat", are used in Chinese folk medicine for the treatment of various skin diseases and diabetes melHtus (2). Winter (36) found that the alcoholic extract of the leaves exhibits anti-inflammatory activity in carragenan induced edema in rat paw. More recently Shimizu et al. (37) reported the isolation of some known triterpenes from the Et20 soluble fraction of the ETOH extract of the leaves and, among these, maslinic acid 6 was found to have an anti-inflammatory effect. Furthermore Noreen et al. and Villar et al. have described a significant hypoglycemic effect in rabbits of £". japonica crude alcoholic extract (38) and of tormentic acid (39). For the structural relationship between tormentic acid and triterpenes 5 and 7, isolated from a CHCI3 extract of E. japonica, we tested their hypoglycemic effect in genetically diabetic mice ( C57BL7KS-db/OLA) and normoglycemic rats. Compounds 5 and 7 were tested at 50, 10 and 1 mg/Kg po in genetically diabetic mice; glycosuria was determined by an enzymatic glucose oxidase method. Tolbutamide was used as control at a dose of 500 mg/Kg. While 5 has the same action after 2 hours at any dose tested, 7 is active at the dose of 50 mg/Kg . Both compounds show a marked inhibition of glycosuria at 4 and 7 hours at all three doses tested. The inhibitory effect was completely absent only after 18-24 hours. The activity of 5 and 7 was tested in normoglycemic rats at dose of 0.1, 1, 10 mg/Kg administered orally 30 min before starting the test. Blood glucose level was monitored every 30 min for 2 hours with an enzymatic glucose oxidase method. At a dose of 0.1 mg/Kg, 7 was able to reduce blood glucose level, while a dose of 5 100 times higher was necessary to produce the same effect (40). Thus polyhydroxylated triterpenes from E, japonica are very active as hypoglycemic agents. Noreen and Villar (38,39) proposed that the crude alcoholic extract of E, japonica and tormentic acid act by stimulating Langerhan' s p cells which results in an increased insulin release because the crude extract and tormentic acid are inactive in alloxan-treated animals. The action mechanism of 5 and 7 seems to be analogous. Studies are in progress to delineate this mechanism more precisely. 4. AROMATIC COMPOUNDS 4.1 FLAVONOL Flavonoids, widely distributed in the plant kingdom, are present in many medicinal plants. The widespread flavonol glycosides rutin 79 and narcissin 6S have been isolated from Mercurialis annua L., a poisonous herb endemic in southern Italy and used in veterinary medicine, as are the minor flavonol glycosides 80,67 and 66. The latter is a new natural compound with a glucopyranosyl unit

140 linked to ring B of narcissin. Compound 67 has a glycopyranosyl unit linked to ring A of narcissin and compound 80 has an additional glucose linked to the rutinosil moiety of rutin (Fig. 9) (14). Myricetin-3-O-rhamnoside 58, quercetin-3-O-rhamnoside 59, quercetin-3-O-arabinoside 60, and kampferol-3-O-glucoside 61 are present in the methanol extract of Be/aria cinnamomea, while quercetin-3-O-glucoside 62, quercetin-3-O-galactoside 63 and isorhamnetin 64 are the main metabolites of the MeOH extract of the leaves of Mynthostachys setoscL (41).

OR HO OH

Jt-

OR,

OH

O

k's:^^:^^:^ OH

R2

^3

R4

79

H

H

H

H

65

H

Me

H

H

80

"

^®

^

6^

H

Me

OH HO-^V-^-^R OiK

-OH

j^—t-\^Q

.

H

OH Me

H

'^HO-^-^V^ OH

Fig. 9 Flavonol glycosides from M. annua. Quercetin 56, 63 and 79 and the flavanone naringin have been isolated from the MeOH extract of Tessaria integrifolia (42) used in Peruvian traditional medicine in hepatic and renal insufficiency. Myricetin-3-O-glucoside 81 and myricetin-3-O-rutinoside 82 are present in the diuretic and cholagogic extract of Combretum micranthum (43), a plant used in traditional African medicine. Compounds 61, 63 and 79 together with kampferol-3-O-rutinoside 83 are also the main flavonol glycosides of Eupatorium guayanum (44), a Peruvian plant of the Compositae family widely used for the treatment of asthma, cold, rheumatic pain and hepatic insufficiency. Compounds 63, 75 and

141 79 were also isolated from the MeOH extract of Arcytophyllum nitidum (45). This flavonoidic fraction has been shown to be effective against a series of gram(-)strains tested and E, coli (Table 11). This confirms the antibacterial activity ascribed to this plant by the Peruvian traditional medicine and the validity of its use in folk-medicine against bacterial infection. The antibacterial action mechanism of flavonoids seems to be due to their action on cell membrane permeability as reported in the literature {A6)

Table lO.Anti HIV-1 activity and toxicity of flavanol 51-67.

[""""c~

OGal

5 OH OH OH OH OH OH OH OH OH OH OH OH OH

51 52

1 53

1SS^^ \ \ 56

3 H H H OH OH OH OH

1 ^'^ 1 58

ORha

I 59

ORha

60

OAra

1 ^^ 1 62 1 63

i

64 1

OGlc OGlc OH

I OH

1 65

OGlc-Rha

[ OH

66

OGlc-Rha

1 67

OGlc-Rha

OH OH

4' H H OH OH OH OH OH OH OH OH OH OH OH

6 H OH H H H H H H H H H H H H

7 OH OH OH OH OH OH OH OH OH OH OH OH OH

1 H"

OH

OGlc

OCH3

OGlc

OH

OCH3

1 OH 1 H 1 OH

3' H H H H 2'OH

OH OH OH OH OH H OH OH OH OH

OCH3 OCH3

5' EC50 20 H inactive H inactive H inactive H inactive H inactive H 2 OH 100 OH 50 H inactive H 10 H inactive H inactive H inactive H H 1 inactive H 1 inactive inactive H

"TCiol 50 1 50 1

2 1 10 1 100 1 10 1 40 1 >200 1 >100 1

>ioo 1 100 1

>ioo 1 >ioo 1 >ioo 1 i >200 j >100

1^

142 Table 11. Inhibition growth of test organism. Extract

|

1

Gram (+) Proteus

Pseudomonas aerug.

mirabiiis

Salmonella TYII

Gram (-)

Escherica Bacillus coli

cercus

Bacillus subtilis

Sarcina subflava

Streptococ.l

Staphylococcus aur

faecal is

streptococc. epidermi CHQa

A

I

I

A

A

A

I

CHQa-MeOH

I

I

A

A

I

A

A

MeOH

A

I

I

A

A

A

I

Flavonol -fract.

I

I

A

A 11 A

A

A

1

I=Inactive, A=Active 4.2 ACYLATED FLAVONOL-GLYCOSIDES Two new acylated flavonol glycosides, quercetin-3-(3",6''-diacetyl)-galactopyranoside 84 and quercetin-3-(2", 3",4''-triacetyl)-galactopyranoside 85, have been isolated from Tagetes elliptica together with nine known flavonol glycosides (47) (figure 10). A series of flavonol glycosides esterified on the glucose residue linked at C-3 of the aglycone have been isolated from Quercus cerris L. (87-89) and Quercus suber L. (90-94) (49) (Fig. 10) Compounds 87 and 89 have isorhamnetin as the aglycone, while compounds 88 and 90-94 have kaempferol as the aglycone. Position 6" of the glucose residue is esterified by gallic acid unit in 87, and trans-/7-coumaroyl acid in 88,89 and 91-94. In addition in compound 88 an acetyl group is linked to 0-4"; in 92 another trans-/?-coumaroyl is linked at 0-2"; in 93 there are two acetyl group at C-3" and C-4'' and a trans-p-coumaroyl at C-l"; in 94 a ds-p-coumaroyl unit links C-2". Compound 90 has the C-6'' position free but links a trans-/?coumaroyl at C-2''. The ^H NMR spectra of 84-94 suggest the acetylation sites by the downfield shifts of the related hydrogens when these give isolated, well resolved signals. The ^^C NMR spectra confirm the esterification sites by the typical downfield shifts of the C supporting the substitution and by the upfield shifts (y-effects) of the C neighbouring the esterified carbons, as compared with the corresponding carbon resonances in the unsubstituted glucose and galactose models. Moreover in 87-94, which have different esterification sites, a 2D COLOC spectra is necessary for the identification of the sites of attachment of the various substitution groups: long-range carbonproton shift correlations are observed between the carbonyl carbon of/7-coumaroyl acid residues and H-6'' of the glucose moiety, or between the carbonyl carbon of the trans-p-coumaroyl or the cxs-pcoumaroyl unit and H of glucose supporting the esterification and H-7'" of coumaric acid residue. This work also led to the structural revision of two of these flavonol glycosides, 92 and 93.

143

o"^-i:z^:fe.^ CHPR4 O H II I TPC=—C—C=:C-

O H TPC-Ac=—c--C = C

? V^

OH

U^

OH

CPC=r-C—C«=C. O H H CPC-Ac=—c—C«C.

V-OAc

OAc

OH Galloyl=

—C

Fig.10

OH

OH 1 Compounds

R

87

OCH3

S8 89

OCH3

1 1

90 91 92

1 1

93 94

H H H H H H

Rl H H H TPC H TPC TPC CPC

R2 H H H H H H Ac H

R3 H Ac H H H H Ac H

R4

1

galloyi 1 TPC

TPC

1

H

1

TPC TPC

1 1

TPC TPC

4.3 FLAVANONES The new glycosides 5,7,2'5'-tetrahydroxy-flavanone-7-0-rutinoside 77, narirutin 76, as well as rosmarinic acid 99, are the main metabolites of the anti-inflammatory, antirheumatic and antipyretic extract of Hamelia patens (50). The 5,7,2',5' -tetrahydroxy substitution pattern of the aglycone of 77 is derived from the l^C NMR data which permit its unambiguous distinction from 3',4'dihydroxyderivative as eriodictyol, and 2',4'-dihydroxyderivative as steppogenin. It is interesting that, in the l^C NMR spectrum of 77, the signals assigned to C-2 and C-3 (diagnostic of a flavanone structure) and to C-6' appeared as a couple of signals [80.5, 79.3 , each CH, and 44.0, 43.9, each CH2 and 115.0, 114.8 ppm, each CH, respectively] and each signal was of lower intensity than that of other CH2 or CH in the same spectrum. Also in the ^HNMR spectrum H-3, H-2, Me-6", H-2" signals were split into two resonances. Because 77 is a mixture of (2S) and (2R) forms of the aglycone, as derived from the CD curves (51), and lacking other possible modification of the structure, the observed splits of the NMR signals can be attributed to the presence of (2S) and (2R) isomers. This behaviour is probably related to asymmetric pertubation of the 2'-hydroxylated aromatic ring by the asymmetric centre as reported in the literature (52).

144

Table 12. Structure and anti HIV-1 activity and toxicity of flavanones 74-78. 4' 3' 7 5* 6 5 3 TC50 1 EC50 74 75 |76 [77 |78

H H H H H

OH OH OH OH OH

OH OH OGlc-ORha OGlc-ORha OGlc-ORha

H H H H H

OH H H 2'OH H

H H H H H

OCH3 OH OH OH OCH3

inactive inactive inactive inactive inactive

15

1

10 16 >100 40

4.4 CATECHINS From the blood-red latex of Croton draconoides, used in Peruvian folk-medicine (53) (+) gallocatechin 68 and (-) epigallocatechin 69 have been isolated. (-) Epicatechin 70, (+) catechin 71 and two galloylesters 72 [(-) epicatechin-3-O-gallate] and 73 [(+)-catechin-7-0-gallate] were isolated from the diuretic and antiinflammatory extracts of Detarium microcarpum , a medicinal plant from Senegal (54). 3'

A //; Table 13. Structures, anti HIV-1 activity and toxicity of flavan 68-73. Position 68 69 70 71 72 73

3 (+)0H (-)0H (-PH (+PH (-)gallate (+PH

5 OH OH OH OH OH OH

6 H H H H H H

7 OH OH OH OH OH Ogallate

3' OH OH OH OH OH OH

4' OH OH OH OH OH OH

5' OH OH H H H H

EC50 5 inactive 2 4 1 10

TC50 >80 >100 >100 >100 >100 >100

4.5 QUINIC ACID DERIVATIVES Two caffeoylquinic acids 95 (3,4,5-tri-O-caffeoylquinic acid) and 96 (4,5-di-O-caffeoyIquinic acid), as well as caffeic acid 97 and synapoic acid 98, have been isolated for the first time from Securidaca longipedunculata

(55), and 3,4,5-tri-O-galloylquinic acid 100 has been isolated from Guiera

senegalensis. (56) (Fig. 11).

145 4.6 ANTIVIRAL ACTIVITY Flavonoids are generally known for their anti-inflammatory, antiallergic and anticarcinogenic activity and some of them have mutagenic properties (57). More recently certain flavonoids have been shown to possess antiviral activity. For example quercetin seems to be effective against herpes simplex virus type 1 (HSV-1) , parainfluenza virus type 3 (Pf-3) and Sindbis virus (SV-1) but it was inactive against poliovirus type 2 and 3 and adenovirus type 3 and 4 infections. Morin was shown to be effective against herpesvirus-suis but rutin did not have this activity. Dihydroquercetin (taxifolin) and dihydrofisetin were virucidal against HSV-1 and herpesvirus-suis but Pf-3 and poliovirus type 2 and 3 were resistant to these two flavonoids. Quercetin, morin, luteolin and fisetin were also active against pseudorabies virus. Subsequent studies confirming earlier observations have shown that naturally occurring flavonoids inhibit infectivity and/or replication of certain RNA (RSV, respiratory syncytial virus, Pf-3, poliovirus ) and DNA (HSV-1) viruses (57). Quercetin and hesperitin affect one or more of the biochemical processes involved in the intracellular replication of each of the viruses studied. Quercetin and catechin were active inhibitors of infectivity, while naringenin totally lacked activity. Thus it is evident that important structure-activity relationships exist between flavonoids and they possess a variety of antiviral activities. More recent studies have shown that 4'hydroxy-3-methoxy-flavones such as 3-methyl quercetin block the replication of poliovirus, apparently by selective inhibition of genomic RNA synthesis (58) and certain flavans such as 4'-6dicloroflavan inhibit human rhinovirus replication by interacting specifically with the VIPI capsid protein to prevent virus uncoating (59). In the past few years the inhibitory effects of flavonoids on the reverse trascriptase (RT) of certain retrovirus including human immunodeficiency virus (HIV) as well as cellular DNA polymerase, have been studied (60-61). As a part of our screening of natural compouds as potential anti -AIDS agents, we isolated flavones, flavans and flavanones as well as quinic acid derivatives from medicinal plants and studied their in vitro anti HIV-1 activity. The bioassays were performed on C8166 cells infected with HI V-III-B strain. Formation of syncitia and gpl20 antigen production were observed. Cell viability of infected cells and cytotoxicity of uninfected cell controls were measured by the MTT-FORMAZAN method (62), and gpl20 antigen production was measiired by ELISA (63). EC50 (the concentration of drug which reduces by 50% the production of gpl20 in infected C86166 cells). TC50 (the concentration which causes 50% of cytotoxicity uninfected C8166 cells) were evaluated. Results are shown in tables 10, 12 and 13. The flavans 68, 70,71,72 and 73 exhibit the most pronounced selective anti HIV-1 activity. In particular the 3-O-galloyl ester derivative of (-) epicatechin, 72, consistenly exhibited the greatest activity (EC5o=l ^ig/ml, selectivity index >100) followed closely by 70 (-) epicatechin. Differences between isomers were noted in the lower activity of (+) catechin 71 with respect to (-) epicatechin 70, Substitution of the hydroxyl group at C-7 by a gallate moiety in 73 caused a reduction in the antiviral activity and increased the cytotoxicity (Table 13). Of the seventeen flavones (unsaturated pyrone ring ) tested, only two myricetin 57 and kaempferol-3-O-glucoside 61 caused significant inhibition of HIV-1 infection at non toxic concentrations (Table 10). The selective activity of myricetin 57 (selectivity index =20) contrasted with the inactivity of quercetin 56 which differed only in the absence of a 5'-hydroxyl group, indicating that all three hydroxyl groups at 3',4',5' positions

146 of ring B are required for the activity. Compound 58 (3-O-rhamnoside of myricetin ) and compound 59 (3-O-rhamnoside of quercetin) exhibited only very slight selective antiviral activity. In contrast glycosidation at C-3 of kaempferol 54 , which lacks a further 3' hydroxyl group on ring B, elicited selective anti HIV activity in 61 (selective index=10). None of the five flavanones (carbonyl at position 4 of the saturated pyrone ring) tested exhibited activity against HIV-1 infection (Table 12). Ravanones are generally more cytotoxic than the flavans and flavones studied. The active flavans 6873 and flavons 51, 56 and 57 were also tested against HIV-2, SIV and Herpes simplex virus (HSV) infections and elicited comparable activities (Table 14). Compound 72 was in any case the most active compound (41). Table 14. Antiviral activities of flavonols and flavans against HIV-2, SIV and Herpes simplex virus. SIV Herpes simplex Herpes simplex HIV.2 EC50 EC50 TC50 EC50 68 8 10 32 >80 69 >100 >100 10 >50 70 2 2 10 >50 71 5 5 20 >50 72 1 1 1 >100 73 ND ND 10 1 >100 C8166 cells were infected with HI V-2ROD or SIVMAC and Vero cells were infected with Herpes simplex virus type 1. The caffeoylquinic acids 95 and 96, caffeic acid 97, synapoic acid 98 and the structurally related rosmarinic acid 99 and 3,4,5-tri-O-galloylquinic acid 100 whose anti HIV-1 activity has already been reported, were tested for anti HIV-1 activity in the same experimental conditions used to test flavonoids. The results are presented in Table 15. While 97 and 99 were found to be inactive in inhibiting viral replication, 96 and 100 showed similar antiviral activity and 95 showed much higher selective anti HIV-1 activity. EC50 values are quite comparable for 95,96 and 100 but significant Compound

differences are seen, however, in toxicity of 95 (Table 15). Although similar in potency, and having comparable EC5o's the lower toxicity of 3,4,5-tri-O-caffeoylquinic acid 95 gives a higher selectivity index of about 3(X). This compound also exhibits a highly selective inhibition of HSV type 1 replication, comparable to that of ganciclovir. Thus the antiviral action of these compounds is not peculiar to HIV (55). Table 15. Antiviral activity of compounds 95-100.

HlV-lniB 1HIV-2ROD SIVMAC Herpes 1 Compound EC50* TC50 EC50 EC50 TC50 EC50 95 100 0.32 20 2 200 0.08 96 1 40 0.6 8 2 100 0.16 100 0.15 ND ND ND ND 1 15 97 200 200 >200 1 >200 200 >50 98 200 200 ND ND ND ND 99 100 40 100 80 150 20 >1000 0.01 0.016 0.02 ND ND Azr Ganciclovir 1 ND 1 ND ND 1 ND 100 0.08 '''EC50 values are the concentrations of compound in mg/ml which inhibited by 50% the production of gp 120 of HIV or SIV, or herpes simplex virus type 1 surface antigens.

147 Mechanism of action Like a number of polyanionic compounds, including sulphated polysaccharides, polyhydroxycarboxylates and various tannins, the flavonoids that we tested seem to interact with the surface glycoprotein gpl20 to prevent binding of the virus to the sCE)4 receptor (41). Table 16. Inhibition of gpl20/sCD4 interaction by flavan compound 72. Compound

Concentration ue/ml

% Inhibition Washed * Unwashed 72 20 89 97 4 45 53 0.8 35 38 DS500 10 20 81 2 8 76 0.4 4 42 •Compound removed before addition of sCD4 to immobilised gpl20. Like dextran sulphate, compound 72 was more effective when added prior to or at the time of the virus infection; this indicates that it acts at an early stage of infection. But unlike the action of dextran sulphate, which readily reverses on removal of drug, the flavans irreversibly inactivate virus infectivity. Treatment of immobilized gpl20 with the flavans, irreversibly blocked the binding of sCI>4. Some degree of specificity in the interaction of the various tested flavans with gpl20 was apparent from the selectivity inhibition of antibody binding. Whereas flavonoids blocked in a dose dependent manner the interaction of monoclonal antibodies 358 and 380 with the V3 loop and CD4 binding regions of HIV-1 gpl20, they had no effect on the binding of monoclonal antibodies 360 and 323 to the N and C terminal regions of the molecule (Table 17). Thus there is a correlation between the degree of antibody inhibition and sCD4 binding by various flavonoids and their relative effectiveness in inhibiting virus infection. On the other hand, although it has been reported that some flavonoids can inhibit HIV reverse trascriptase in vitro (61), it is apparent that the inhibitory action of flavans is non specific. In fact we observed that they do not inhibit polymerase activity in the presence of serum albumin or detergents such as Triton X-1(X). The fact that human DNA polymerase a, p and y are inhibited to a similar degree by flavans suggests that flavans bind the polymerase without selectivity. From this evidence it is clear that the inhibition of HIV-1 infection by flavans is principally due to a selective interaction with gpl20. In this respect the anti HIV-1 activity of flavans is similar to that of various tannins and .polyanionic compounds. Similar studies (55) on the mechanism of action of quinic acid derivatives 95-100 suggest that they do not inhibit HIV replication by inhibition of HIV-RT as previously reported for 100 (64, 65), but they inactivate virus infectivity by specifically binding to gpl20 which prevents its interaction with CD4 on t-lymphocytes. The inhibition of HIV infection was in fact more pronounced when compounds were present during virus adsorption than when added after infection, as in the case of dextran sulphate. These compounds reduce syncythium formation between chronically infected and uninfected cells while dextrane sulphate inhibits syncytium formation only when added during the mixing of chronically infected and uninfected cells.

148 Table 17. Inhibition of antibody interaction with gpl20 of flavonols and flavans. Compound

I Concentration (jig/ml)

51 56 57 69 70 71 72 DSsoo

10 10 1 10 1 10 50 5 0.5 50 5 10 2 0.4 10 2 0.4

323 0 0 0 0 0 0 0 0

% Inhibition of antI body binding 358 360 0 0 7 0 0 90 0 50 0 0 96 0 44 17 59 0 24 98 0 78 11 92 0 26 0

Table 18. Inhibition of gpl20sCD4 interaction. Compound

Concentration

% Inhibition

(^ig/ml)

95 96 97 99 DS500

Azr

50 10 2 20 4 0.8 50 10 2 25 5 1 10 2 0.4 50nM 10

Washed* 87 66 45 77 56 39 14 10 2 41 25 10 20 8 4 5 2

Unwashed 97 88 62 95 63 53 22 12 2 63 29 10 81 76 42 3 3

Compound removed before addition of sCD4 to immobilized gpl20.

380 16 40 48 98 97 0 91 87 36 96 24 99 85 37 93 0 0

149

:ooH R2O'//.

OH

Ri

95

R9

cafl

caffeoyl

caffeoyl

caf

caffeoyl

96

100

galloyl

galloyl

gall

COOH .CCX)H

OH 99

Galloyl=

HO

co-

Fig. 11 Quinic acid and cinnamic acid derivatives from S, Longipedunculata and G. senegalensis.

150 REFERENCES 1. 2.

A. Olayiwola, Fitoterapia 63, (1990), 99. Z.Z. Liang, R.- Aquino, V. De Feo, F. De Simone, C. Pizza, Pianta Medica 56 (1990), 330.

3.

N. De Tommasi, F. De Simone, C. Pizza, N. Mahmood, P.S. Moore. C.Conti, N. Orsi, M.L.Stein, J. Nat. Prod. 55 (1992), 1067.

4.

R. Cerri, R. Aquino, F. De Simone, C. Pizza, J. Nat. Prod. 51 (1988), 257.

5.

R. Aquino, F. De Simone, C. Pizza, R. Cerri, J. F. De Mello, Phvtochemistrv 27 (1988), 2927.

6.

R. Aquino, I. Behar, F. De Simone, M.D' Agostino, C. Pizza, J. Nat. Prod. 49, (1986), 1096.

7.

Z.Z. Liang, R. Aquino, F. De Simone, A. Dini, O. Schettino, C. Pizza, Pianta Medica 54 (1988), 344.

8.

P.K. Agrawal, Phvtochemistrv 31 (1992). 3307.

9.

N. De Tommasi, S. Piacente, F. De Simone, C. Pizza, J. Nat. Prod, . in press (1993).

10. R. Aquino, F. De Simone, C. Pizza, C. Conti. M.L. Stein. J. Nat. Prod. 52. (1989), 679. 11. R. Aquino, V. De Feo, F. De Simone, C. Pizza. G. Cirino. J. Nat. Prod. 54. (1991),453. 12. A.M. Yepez, O.L. de Ugaz, CM. Alvarez, R. Aquino, V. De Feo, F. De Simone, C. Pizza, Phvtochemistrv 28 (1991). 1635. 13. R. Aquino, F. De Simone, C. Pizza, J.F. De Mello, Phvtochemistrv 28 (1989), 199. 14. R. Aquino, I. Behar, M.D' Agostino,F. De Simone,0. Schettino, C. Pizza, Biochemical Systematics and Ecologv 15 (1987), 667. 15. A. U. Rahman, M.I. Choudharx, A. Pervil " Principles and Applications of Modem 2D NMR Techniques in Structure elucidation of Complex Method " in A.U. Rahaman Ed. Studies in Natural Product Chemistrv. Elsevier Science Publ. B.V., Amsterdam, 9 (1991), 127. 16. L. Pistelli, A.R. Bilia, A. Marsili, N. De Tommasi, A. Manunta, J. Nat. Prod.55 (1993), 240. 17. M.C. Das, S.B. Mahato. Phvtochemistrv 22 (1983), 1071. 18. G.R. Mallavarapu "Recent Advances in Oleanane Triterpenes" in A.U. Rahman Studies in Natural Product Chemistry. Elsevier Science Publ. B.V., Amsterdam. 7 (1990), 131. 19. R. Aquino, F. De Simone, F.F. Vincieri, C. Pizza, E.Gacs-Bsitz, J. Nat. Prod. 53 (1990), 559. 20. H. Wagner, B. Kreutzkamp, K. Jurcic, Pianta Medica 51 (1985), 419. 21. M.E.O. Matos, M.P. Sousa, M.I.L. Machado, R.B. Filho, Phvtochemistrv 25 (1986), 1419 and references therein cited. 22. A.G. Miana, M.G. Hassan Al-Hazini, Phvtochemistrv 26 (1987), 225. 23. C. Pizza, Z.Z. Liang, N. De Tommasi, J. Nat. Prod. 50 (1987), 927. 24. G. Romussi, S. Cafaggi, C. Pizza, Arch. Pharm. 321 (1988), 753. 25. R. Aquino, I. Behar, F. De Simone, C. Pizza, M. D' Agostino, J. Nat. Prod. 48 (1985), 502. 26. S.B. Singh, R.S. Thakur, H.R. Schulten, Phvtochemistrv 21 (1982), 2079.

151 27. K. Hiller" New results on the structure and biological activity of triterpenoid saponins" in Biologically Active Natural Products Oxford Science Publications. (K. Hostettmann, P. Lea eds.) 12 (1987), 167. 28. R. Pompei, O. Rore, M.A. Marciallis. A. Pani, B. Loddo, Nature 281 (1979), 689. 29. K. Fujisawa. Y. Watanabe, K. Kimura, Asian. Med. J. 23 (1980), 754. 30. M. Baba, S. Shigeta, Antiviral Res. 7 (1980), 99. 31. M. Ito, H. Nakashima, M.Baba. R. Pauwels, E. De Clercq, S. Shigeta, N. Yamamoto Antiviral. Res 7 (1987). 127. 32. R. Aquino, C.Conti, F. De Simone, N. Orsi, C. Pizza, M.L.Stein, Journal of Chemotherapy 3 (1991), 305. 33. N. Abe, T. Ebina, N. Ishida. Microbiol. Immunol. 26 (1982), 535. 34. N. De Tommasi, C. Conti, M. L. Stein, C. Pizza PlantaMedica 57 (1991), 251. 35. M.B. Gupta, T.N.Bhalla, G.P.Gupta, C.R.Nitra, K.P. Bhargave, Eur. J. Pharmacol. 6 (1969), 67. 36. C.A. Winter, E.A. Rislex, G.W. Nuss, Proc. Soc. Exp. Biol. I l l (1962), 544. 37. M. Shimizu, H. Fukumura, H. Tsuji, S. Tanaomi, T. Hayashi, N.Morita, Chem. Pharm. Bull. 34 (1986), 2614. 38. W. Noreen. A. Wadood, H.K. Hydayat, S.A.M. Wahid, Planta Medica 54 (1988), 196. 39. M.D. Ivarra, N. Paya, A. Villar, Planta Medica 54 (1988), 282. 40. N. De Tommasi, F. De Simone, G. Cirino, C. Cicala, C. Pizza, Planta Medica 57 (1991), 414. 41. N. Mahmood, C. Pizza; A. Aquino, N. De Tommasi, S. Piacente, S. Colman, A. Burke, A.J. Hax, Antiviral Research in press (1993). 42. 43. 44. 45. 46. 47. 48. 49. 48. 50. 51. 52.

V. De Feo, M. D' Agostino, F. De Simone. C. Pizza, Fitoterapia 61 (1990), 474. M. D' Agostino, C. Biagi, V. De Feo, F. Zollo, C. Pizza, Fitoterapia 61 (1990), 477. M. D' Agostino. V. De Feo. F. De Simone, C. Pizza. Fitoterapia 61 (1990). 375. V. De Feo. C Delia Valle. F. De Simone. C. Pizza. Annali di Chimica 61 (1990), 474. R.J. Guyglewski, R. Korbut, J. Rodax, J.Jwies. Biochem. Pharm. 36 (1987). 317. M. D' Agostino, F. De Simone, Z.Z. Liang, C. Pizza, Phytochemistry 31 (1992), 4387. G. Romussi, G. Bignardi, C. Pizza, Liebigs Ann. Chem. ( 1988), 989. G. Romussi, G. Bignardi, C. Pizza, N. De Tommasi, Arch. Pharm. 324 (1991), 519. G. Romussi, G. Bignardi, C. Pizza, Liebigs Ann. Chem. ( 1988), 989. R.Aquino, M.L. Ciavatta, F. De Simone, C. Pizza, Phvtochemistrv 29 (1990). 2358. W. Gaffield Tetrahedron 26 (1970). 4093. N.C. Baruah, R.P. Sharma, G.Thyaga Rayani, W. Herz, S. Govidan, Phvtochemistrv 18 (1979). 2003.

53. R.Aquino, M.L. Ciavatta, F. De Simone, Fitoterapia 62 (1991). 454. 54. R.Aquino, M.L. Ciavatta, N. De Tommasi, F. De Simone, C. Pizza. Fitoterapia 62 (1991), 455.

152 55. N. Mahmocxi, P.S. Moore, N. De Tommasi, F. De Simone, C. Pizza, Antiviral Chemistry and Chemotherapy (1993) in press. 56. C. Pizza, personal communication. 57. T.N. Kaul, E. Middleton, RL.Ogra, Journal of Medical Virology 15 (1985), 71 and references therein cited. 58. N. De Meyer, A. Haemers, L. Mishra, H.K. Pandey, L.A.C. Pieters, D.A. Vanden Berghe, A.J. Vlietinck. J. Med. Chemistry 34 (1991), 736. 59. M.A. McKinlay, M.G. Rossmann, Ann. Rev. Pharmacol. Toxicol. 29 (1989), 111. 60. H. Nakane, K.Ono, Biochemistry 29 (1989), 2841. 61. P.S. Moore, C. Pizza, Biochem. J. 288 (1992), 717 and references therein cited. 62. R. Pauwels, J. Balzarini, M. Baba, R. Snoeck, D. Schols, P. Herdewijn, J. Desmyter, E. De Clerq J. Virol. Method. 20 (1988), 309. 63. N. Mahmood, A.J. Hay, J. Immunol. Methods 151 (1992), 9. 64. M. Nishizawa, T. Yamagishi, G.E. Dutschman, W. B. Parker, A.J. Bodner, R.E. Kilkuskie, Y.C. Cheng, K.H. Lee. J. Nat. Prod. 52 (1989), 762. 65. W. B. Parker, M. Nishizawa, M.H. Fisher, N. Ye, K.H. Lee. Biochem. Pharmacol. 38 (1989), 3759.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol 17 © 1995 Elsevier Science B.V. All rights reserved.

153

Sesquiterpenes and Other Secondary Metabolites of Genus Lactarius (Basidiomycetes): Chemistry and Biological Activity G. Vidari and P. Vita-Finzi

Introduction The genus Lactarius (order Agaricales, family Russulaceae), is one of the largest in the subdivision Basidiomycotina of Whittaker's Kingdom of Fungi (1), and it comprises more than 150 species that grow world-wide in different habitats (2). These mushrooms show several morphological and biological features appealing to the chemists of natural products. For example, the fruiting bodies have various sizes and brilliant colours, and exchange organic materials and salts with several host plants, forming important mycorriza. The flesh of some species is mild and edible, while that of most Lactarius mushrooms tastes pungent or bitter, causing irritation to intestinal walls. The burning sensation develops on the lips and the tongue of an unskilled mycologist within a few seconds up to a few minutes, helping him to recognize inedible and toxic species. Moreover, observing a characteristic milky juice which appears on the surface of damaged fruiting bodies, anyone can easily distinguish a Lactarius species from a congener Russula species or other similar mushrooms. In addition, the colour and taste of this latex can be different from species to species and may change in the air, more or less rapidly, even for the same species, a fact that has a significant taxonomic relevance (2). For instance, the milky juice is permanently white and mild in L. volemus, but it is white and rapidly pungent in L. vellereus; it changes from white to yellow in L. scrobiculatuSy while it becomes bitter and red in L. fuliginosus and violet in L. uvidus. As a rule, only the species with a permanently red-orange juice are surely edible and taste mild. Recently, some aspects of these biochemical processes have been investigated and will be discussed later on in this review. Moreover, the observation of an interesting antibiotic activity for some Lactarius extracts (3) stimulated the search of new biologically active compounds among those isolated from these mushrooms. In fact, simple bioassays (4) led to the identification of new products with antimicrobial, cytotoxic, antifeedant and other interesting activities. No less important was the observation that some species seem to withstand attack from parasites such as snails and insects better than others. In fact, it has been shown that resistant species are armed with a chemical defence system which protects the mushrooms from predators and invaders. In this review we will discuss the chemistry and biological aspects of those secondary metabolites that seem more peculiar to Lactarius than to other mushrooms and have, therefore, a taxonomic relevance. By contrast, other important metabolites, such as triterpenoids, sterols, polyisoprenoids, fatty acids, aminoacids, etc., widely distributed in several species of different

154 genera, will not be considered. Moreover, limitation of space prevents us from including here references of papers describing the total synthesis ofLactarius metabolites, when they are not relevant to structure determination. Since excellent reviews on fungal metabolites have been published in the past (5-7), we will discuss in detail the literature published afterwards until the end of 1993. We have also added some recent yet unpublished results from our laboratory. In Tables 1-23 the structures of isolated and synthetic compounds (the latter in italic) are reported, while Table 24 reports the distribution of secondary metabolites in the investigated Lactarius species, which have been listed according to the subdivision of the genus by M. Bon (2). In Table 24 the metabolites have the same number as in the previous Tables 1-23 and the references are reported in parentheses.

Sesquiterpenes isolated from Lactarius species Sesquiterpenes of several kinds are the characteristic metabolites isolated from Lactarius mushrooms. However, other metabolites such as alkaloids, phenols and derivatives have been found in some species. Except humulene and sterpuranes, sesquiterpenes with the other skeletons shown in Scheme 1 have been isolated from Lactarius species. They have been divided into classes according to their biosynthetic origin from farnesylpyrophosphatc. The small class of farncsanc sesquiterpenes is derived directly from the alicyclic precursor, while drimanes, guaianes and the other classes arise by different cyclizations of farnesylpyrophosphatc. Two different cyclizations of a humulene precursor give rise to the classes of cariophyllanes and protoilludanes. The sesquiterpenes formally deriving from a protoilludane precursor constitute the largest group of Lactarius metabolites. Cyclobutane ring contraction of a protoilludane cation may give rise to the marasmane skeleton, whereas further rearrangements of marasmanes lead to the glutinopallane, lactarane and isolactarane skeletons. In principle, the secolactarane skeleton may arise by bond cleavage of a lactarane, however, the results of some biomimetic-like reactions in vitro (vide infra) seem to indicate their direct origin from marasmanes. In ahemative to the protoilludane-marasmane pathway, isolactaranes may originate from a rearrangement of a suitable sterpurane intermediate, even if this route in the Lactarius species has not been corroborated by the isolation of any sterpurane sesquiterpene. Contraction of the seven membered ring of lactaranes, with loss of the C-8 carbon atom, gives rise to the 8-norlactarane skeleton, whereas loss of the C-13 carbon of marasmanes leads to the 13-normarasmane skeleton. The results of a few biosynthetic investigations, discussed later, are consistent with this general scheme. Moreover, the occurrence of sesquiterpenes with different skeletons in the same species, for instance, marasmane, normarasmane, isolactarane, lactarane, and secolactarane sesquiterpenes in L. vellereus, points out their common biogenesis. Drimane, farnesane, glutinopallane, protoilludane, isolactarane, and guaiane sesquiterpenes have been isolated so far in a few Lactarius species; therefore, they may be considered chemotaxonomic markers. By contrast, large quantities of marasmane, lactarane and secolactarane metabolites occur in almost all Sections, as reported in Table 24. The carbons 5 and 13 of the skeletons of many marasmane, lactarane and secolactarane

155 sesquiterpenes are linked by an oxygen bridge forming an extra ring, either a furan or a y-lactone ring. In the latter the carbonyl group may be located either at C-5 or at C-13. Therefore, it is convenient to subdivide these classes of metabolites into the following groups: simple marasmane and lactarane sesquiterpenes (Tables 6 and 10, respectively), heterocyclic marasmanes (Table 7), 5lactaranolides (Tables 11-12), 8,9-seco-5-lactaranolides (Table 14), 13-lactaranolides (Tables 16 and 17), furanolactaranes (Table 18), and 8,9-secofuranolactaranes (Table 19). Compounds with rearranged structures, obtained by chemical reactions, are reported in Tables 13 and 20. Drimane, guaiane, farnesane and cariophyllane sesquiterpenoids are typical products of the plant metabolism. By contrast, sesquiterpenes with the skeletons derived from a protoilludane precursor have been isolated so far only from Basidiomycetes, but they are not unique to Lactarius species. In fact, marasmanes have been found, for example, also in species of the genera Russula, Lentinellus, Auriscalpium, Bondarzewia, Vararia, Dichostereum, Peniophora, Artomyces, Marasmius, and Fomitopsis; protoilludanes have been isolated from Fomitopsis, Clitocybey Laurilia, and Armillaria species; isolactaranes have been isolated from Stereum and Merulius species, while lactaranes and secolactaranes are also present in Russula, Lentinellus and Fomitopsis species. Anatomical characteristics point to the possibility that several of these genera may form a natural group together with the genera Lactarius and Russula (Russulaceae). Farnesane sesquiterpenes Recendy, the farnesane sesquiterpenes 1.1-1.9 (Table 1) have been isolated for the first time from Lactarius porninsis, the only species of the Section Zonarii investigated so far (8). Pominsal (1.1), porninsol (1.2) and the esters 1.3-1.9 show a high thermal and photochemical lability because of the conjugated tetraene system and readily polymerize when their solutions are taken to dryness. Therefore, special mild conditions are required for their isolation and for recording the spectroscopic data. The composition of the ester mixture 1.3-1.9 was established by capillary GC and GC-MS analysis of the methyl esters obtained by transesterification (8). TABLE 1 - Farnesane sesquiterpenes

>f

Name

Substituents

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

Porninsal Porninsol Miristoylpominsol Pentadecanoylpominsol Palmitoleylpominsol Palmitoylpominsol Linoleylpominsol Oleylpominsol Stearoylpominsol

13-oxo 13-OH 13-miristoyloxy 13-pentadecanoyloxy 13-palmitoleyloxy 13-palmitoyloxy 13-linoleyloxy 13-oleyloxy 13-stearoyloxy

Ref.

156

P.O.

2

6

8

10

FARNESANES

13 NORLACTARANES

4

8 ' LACTARANES

13SECOLACTARANES

SCHEME 1 - The proposed biogenesis of Lactarius

Cariophyllane

sesquiterpenes

sesquiterpenes

Alcohol 2.1 (Table 2), the only example of this class, was isolated from Lactarius

camphoratus

157 (9), which belongs to the Section Olentes (Table 24). The structure and absolute configuration of this new cariophyllene oxide (2.1) was determined by a combination of spectral data and single-crystal X-ray analysis of the p-bromobenzoate derivative 2.3. TABLE 2 - Cariophyllane sesquiterpenes

N° 2.1 2.2 2.3

Ref.

Substituents No trivial name No trivial name No trivial name

12-OH 12-OAc 12-OBz-p-Br

9 9 9

Drimane sesquiterpenes Drimanes have only been isolated from two Lactarius species of the Section Uvidi: Lactarius uvidus {10-12) 2LndL.flavidus (23). In addition to uvidins A (3.14) and B (3.36) (10), more recently several new fatty acid esters of drimenol (3.2-3.6) and uvidin A (3.16-3.20) (11), and three new sesquiterpenes (12) with the bicyclofarnesane skeleton have been isolated from L. uvidus (Table 3). The stereostructures of the three uvidins C (3.29), D (3.34), and E (3.22) have been established by spectroscopic data and chemical reactions (12). In the interesting synthesis of uvidin E (3.22) from uvidin A (3.14), the Rubottom's procedure was employed for the regiospecific and stereoselective a-hydroxylation at C-5 of the enone 3.10 (12) (Scheme 2). ^OH

^OTHP

OTHP

^OH

Scheme 2 Exposure of uvidin A (3.14) or the esters 3.16-3.20 to methanolic KOH led to the new lactone 3.38, arising by a Favorskii rearrangement of the a,p-epoxyketone function (11). The same rearrangement was observed for 11-O-ethoxy ethyl uvidin A; however, in this case the lactone ring involved the tertiary OH group of compound 3.39. Sesquiterpenes 3.38 and 3.39 have a new

158 skeleton named isothapsane (11). TABLE 3 - Drimane sesquiterpenes

14 13 3.1-3.37

3.38

O

3.39

N°

Name

Substituents

3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 3.20 3.21 3.22 3.23 3.24 3.25 3.26 3.27 3.28 3.29

Drimenol Palmitoyldiimenol Linoleyldrimenol Oleyldrimenol Stearoyldrimenol 6-Ketostearoyldrimenol No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name Uvidin A No trivial name Palmitoyluvidin A Linoleyluvidin A Oleyluvidin A Stearoyluvidin A 6-Ketostearoyluvidin A No trivial name Uvidin E No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name Uvidin C

ll-OH;7(8)-en;5a-H 1 l-palmitoyloxy;7(8)-en;5a-H 1 l-linoleyloxy;7(8)-en;5a-H I l-oleyloxy;7(8)-en;5a-H II -stearoyloxy ;7(8)-en;5a-H 11 -(6-oxostearoyloxy);7(8)-en;5a-H ll-OH;5a,8a-H 6-OSiMe3;l l-OTHP;5(6),7(8)-dien 1 l-OH;6-oxo;7(8)-en;5a-H 1 l-OTHP;6-oxo;7(8)-en;5a-H ll-OH;6-oxo;5a,8a-H I l-OTHP;6-oxo;5a,8a-H 5a,6a-epoxy;6-OSiMe3;l l-OTHP;7,8-en 7p,8p-epoxy;l l-OH;6-oxo;5a-H 7p,8p-epoxy;l l-OAc;6-oxo;5a-H 7p,8P-epoxy; 11 -palmitoyloxy;6-oxo;5a-H 7p,8p-epoxy; 11 -linoleyloxy;6-oxo;5a-H 7p,8p-epoxy; 11 -oleyloxy;6-oxo;5a-H 7p,8p-epoxy; 1 l-stearoyloxy;6-oxo;5a-H 7p,8P-epoxy; 11 -(6-oxostearoyloxy);6-oxo;5a-H 3p,l l-diOH;6-oxo;7(8)-en;5a-H 5a,l l-diOH;6-oxo;7(8)-en 7,1 l-diOH;6-oxo;7(8)-en;5a-H 5a-OH;l l-OAc;6-oxo;7(8)-en II -OH;7-OAc;6-oxo;7(8)-en;5a-H 7-OH; 1 l-OTHP;6-oxo;7(8)-en;5a-H 3p,l l-diOAc;6-oxo;7(8)-en;5a-H 7-OAc; 11 -OTHP;6-oxo;7(8)-en;5a-H 7p,8p-epoxy;6p,l l-diOH;5a-H

Ref. 10 11 11 11 11 11 10 12 10 12 10 12 12 10 10 11 11 11 11 11 10 12 12 12 12 12 10 12 10,12

159 3.30 3.31 3.32 3.33 3.34 3.35 3.36 3.37 3.38 3.39

No trivial name No trivial name No trivial name No trivial name Uvidin D No trivial name Uvidin B No trivial name No trivial name No trivial name

7p,8p-cpoxy;6p-OH;ll-OAc;5a-H 7p,8p-epoxy;6p,l l-diOAc;5a-H 3p,l l-diOH;6-oxo;5a,8a-H 7a,l l-diOH;6-oxo;5a,8a-H 7p,l l-diOH;6-oxo;5a,8a-H 8p,ll-diOH;6-oxo;5a-H 7p,8p-epoxy;3p,l l-diOH;6-oxo;5a-H 7p,8p-epoxy;3P,l l-diOAc;6-oxo;5a-H (See formula) (See formula)

10,12 12 10 12 12 10 10 10 11 11

Uvidins are attractive chiral starting materials for the synthesis of highly oxidized biologically active drimane-like sesquiterpenes as demonstrated by the syntheses of natural (-)-cinnamodial (115) and (-)-cinnamosmolide (116) from uvidin A (3.14). Guaiane sesquiterpenes Guaiane sesquiterpenes (Table 4) have been isolated so far onlyfromLactarius species of the Section Dapetes (Lactarius deliciosus, L. sanguifluus, etc.) which are characterized by the secretion of a strong coloured milky juice. Usually these mushrooms are edible and of pleasant taste. Each species contains a characteristic mixture of coloured sesquiterpenes responsible for the natural orange, red, green, or even blue colour of the milky juice. The structures of a dozen of guaiane sesquiterpenes were determined by chemical and spectral methods. These compounds are extraordinarily sensitive and could be isolated by employing very mild extraction and purification conditions. For example, two blue pigments of L. indigo were instantaneously converted to an intractable green substance upon addition of MeOH to the acetone solution, or on attempted chromatography (22). Both esters 4.13 and alcohol 4.9 polymerized in air (14). Similarly, delicial (4.5) rapidly polymerized when exposed to light (16). Normal chromatography of delicial was not possible, but small amounts could eventually be obtained by flash chromatography in the dark and with cold solvent, on silica gel prewashed with cold ethyl ether (16). Most isolated guaiane sesquiterpenes show a formyl or free hydroxymethyl group at C-4. However, recent results have shown that they are not present as such in the undamagedfruitingbodies, on the contrary, they are formed enzymatically from fatty acid ester precursors in injured specimens. Interestingly, there are examples of different metabolites produced by the same species collected in different parts of the world. For instance, aldehyde 4.6 has been isolated from Indian (20) but not from European specimens of L. deterrimus, while lactarofulvene (4.1) was isolated from Califomian specimens (13) of L. deliciosuSy but not from European specimens (16). An explanation of these apparent differences between specimens grown in different continents may be the existence of sub-species (16), or a change of the metabolism related to different habitats, or it may be due simply to the formation of artifacts during extraction.

160 TABLE 4 - Guaiane sesquiterpenes 14

N°

Name

Substituents

4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14

Lactarofulvene Lactarazulene Lactaroviolin No trivial name Delicial No trivial name DetemDl Sangol No trivial name Stearoyldeterrol Stearoylsangol No trivial name No trivial name Dimers

1(5),2(3),4(15),6(7),9(10),1 l(12)-esaen 13,14 l(10),2(3),4(5),6(7),8(9),ll(12)-esaen 14-16 15-oxo;l(10),2(3),4(5),6(7),8(9),ll(12)-esaen 14,16-18 15-oxo;l(2),3(4),5(6),7(l l),9(10)-pentaen 19 15-oxo;l(2),3(4),5(6),9(10),ll(12)-pentaen 16 15-oxo;l(10),2(3),4(5),6(7),8(9)-pentaen 19,20 15-OH;l(10),2(3),4(5),6(7),8(9),ll(12)-esaen 16 15-OH; 1 (2),3(4),5(6),7( 11 ),9( 10)-pentaen 21 15-OH;1(2),3(4),5(6),9(10),1 l(12)-pentaen 14 15-stearoyloxy;l(10),2(3),4(5),6(7),8(9),ll(12)-esaen 22 15-stearoyloxy; 1 (2),3(4),5(6),7( 11 ),9( 10)-pentaen 21 15-linoleyloxy;l(2),3(4),5(6),9(10),ll(12)-pentaen 16 15-stearoyloxy;l(2),3(4),5(6),9(10),ll(12)-pentaen 14,16 16,22

Ref.

Protoilludane sesquiterpenes Recently, for the first time among the Lactarius sesquiterpenes, the protoilludane skeleton has been assigned to two metabolites oi Lactarius violascens (23) (Table 5). Also this mushroom contains a sesquiterpene alcohol (5.1) and the corresponding 6-oxostearic acid ester 5.2. It is worth noting that this fatty acid (also named lactarinic acid) is peculiar to Lactarius mushrooms where it has been isolated in the form of many sesquiterpenoid esters. The structures of compounds 5.1 and 5.2, particularly the position of the oxygenated group at C-15, have been determined by 2D-NMR spectt-a and NOESY experiments. TABLE 5 - Protoilludane sesquiterpenes O^ 5 J L' ^ 1 4 11^ 13

""^15

N°

Name

Substituents

5.1 5.2

Violascensol 15-(6-Ketostearoyl)violascensol

15-OH 15-(6-oxostearoyloxy)

Ref. 23 23

161 Marasmane and 13-norinarasmane sesquiterpenes Among the not many marasmane sesquiterpenes isolated from Lactarius species (Tables 6 and 7), velutinal esters (7.14, 7.15) deserve a special consideration. Stearoylvelutinal (7.14) was originally isolated by a French group from Lactarius velutinus (39), during the search for the substances that are responsible for the intense blue colour which develops on the gills of a Lactarius specimen by reaction with the sulfo-vanillin reagent. In systematic mycology this reagent is used for identification purposes (40). Independendy, almost at the same time Swedish authors isolated stearoyl- (7.14) and 6-ketostearoylvelutinal (7.15) fromL. vellereus and L. necator (38), in an attempt to clarify the formation of some artifacts. Since then, most Lactarius species have been shown to contain velutinal esters, even if several important exceptions are known (Table 24). Particularly the species of Section Albati, possessing a permanently white milky juice, contain large amounts of velutinal esters. However, these compounds are not unique to Lactarius species, but they have also been found in a number of other genera, for example, in Russula, Lentinellus, Auriscalpium, Artomyces and Peniophora species (40, 113, 117). Catalytic transesterification of esters 7.14 or 7.15 in EtO"/EtOH afforded the free hemiacetal velutinal (7.11) (38), while both velutinal esters gave the methyl acetal 7.13 (38, 40) in HPLC grade methanol. Velutinal (7.11) can be synthesized by selective reduction of isovelleral (6.1) to isovellerol (7.2) using KBH4 in ethanol at r. t , followed by the Sharpless epoxydation of isovellerol (37). Moreover, the treatment of methyl velutinal (7.13) with lithium diisopropylamide afforded, by a p-elimination of the epoxyde, the corresponding 7,13-en-8a-ol derivative 7.9 (36). This product can be easily hydrolysed to isovelleral (6.1) in a THF/H2O mixture containing traces of acid or on prolonged contact with silica gel (36). This conversion may support the biosynthetic pathway proposed for the transformation of velutinals to isovelleral (6.1) in injured mushrooms (36,46). Free velutinal (7.11), its esters 7.14 and 7.15, and methyl acetal 7.13 are labile compounds, and on adsorption on silica gel they yield some of the furanolactarane and secofuranolactarane sesquiterpenes which have been isolated previously from different Basidiomycetes, including Lactarius species. Fast degradation takes place also on dissolving velutinal derivatives in wet acetone or in reagent grade alcohols as under other conditions where traces of acid are probably present (40, 118). Degradation by adsorption on Al203of stearoylvelutinal (7.14) yielded, in addition to the furans, significant amounts of isovellerol (7.2) and lactarol (19.3) (27). The furanoid sesquiterpenes are formed via intermediate dihydrofurans, many of which, in absence of acid, are stable enough to be isolated (85). The formation in vitro of dihydrofurans and furans from velutinal derivatives can be explained by a mechanism via carbocationic rearrangements (Scheme 3) which corroborates the stereochemistry assigned to several Lactarius sesquiterpenes (85, 118). Moreover, this mechanism may be very similar to the enzymatic conversion of velutinal esters to some furanolactaranes and secofuranolactaranes in injured mushrooms (46) {vide infra). One must be aware of the possible formation of artifacts in such conversions and, therefore, strict control must be exerted on any operation where this risk can occur. We must stress that the choice of the solvent for extracting the mushrooms is critical and that preliminary experiments should

162 suggest the best procedure for isolation and chromatographic separation of individual compounds (8, 14, 16, 22, 27, 40, 46). Alcoholic solvents are particularly harmful (119) and instead of them water insoluble solvents like hexane (27), EtOAc (27), ether (21) or CH2CI2 (40) have been recommended. A highly oxygenated marasmane sesquiterpene has been isolated from Lactarius pallidas (29) and named lactaropallidine (6.6). The structure of this compound, including the relative configuration of stereogenic centres, has been elucidated by spectroscopic methods and extensive decoupling experiments. The absolute configuration has been established by the CD measurement for the CO n -> 71* transition.

18.10

18.5

18.17 or 18.19

Scheme 3 - Degradation mechanism of velutinal derivatives Furthermore, reduction of stearoylvelutinal (7.14) with Red-Al (sodium

his{2'

methoxyethoxy)aluminium hydride) in toluene-THF gave directly lactaropallidine in a single step (29). A possible mechanism for this epoxyde-ketone rearrangement is reported in Scheme 4. The absolute configurations of lactaropallidine (6.6) and velutinal esters (7.14, 7.15) were definitively established by an enantioselective synthesis of isovelleral (6.1) (31). This assignment also indicated the absolute configurations of many marasmane, lactarane and secolactarane sesquiterpenes which have been stereochemically correlated to each other and to velutinals or isovelleral (6.1) by chemical reactions. Interestingly, Lactarius sesquiterpenes derived from velutinal esters (7.14, 7.15) have the same configurations as that suggested for the related antibiotic marasmic acid (120), but opposite to that of hirsutic acid, another fungal metabolite (121).

163

Scheme 4 TABLE 6 - Marasmane, isomarasmane and normarasmane sesquiterpenes

6.12-6.22

6.1-6.9

N°

Name

Substituents

6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 6.14 6.15 6.16 6.17 6.18 6.19 6.20 6.21 6.22

Isovelleral [ll'^Hshlsovelleral ^^0-Isovelleral Isovellerdiol No trivial name Lactaropallidine No trivial name No trivial name No trivial name Isoisovelleral No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name

5,13-dioxo;7(8)-en;2a,9a-H 12-2H3;5,13-dioxo;7(8)-en;2a,9a-H i80;5,13-dioxo;7(8)-en;2a,9a-H 5,13-diOH;7(8)-en;2a,9a-H 9a-OH;5,13-dioxo;7(8)-en;2a-H 5,13-diOH;8-oxo;2a,7a,9a-H 5,10a, 13-triOH;7(8)-en;2a,9a-H 5,10a, 13-triOAc;7(8)-en;2a,9a-H 5,7a,8a,13-tetraOH;2a,9a-H 5,13-dioxo;7(8)-en;2a,9a-H 9a-OH;5,13-dioxo;7(8)-en;2a-H 5,7a-diOH;8-oxo;2a,9a-H 5,8a-diOH;7-oxo;2a,9a-H 5,7a-diOAc;8-oxo;2a,9a-H 5,8a-diOAc;7-oxo;2a,9a-H 7a-OH;5,8a-diOAc;2a,9a-H 7a-OH;5,8p-diOAc;2a,9a-H 8a-OH;5,7a-diOAc;2a,9a-H 8P-OH;5,7a-diOAc;2a,9a-H 5,7a,8a-triOAc;2a,9a-H 5,7a,8p-triOAc;2a,9a-H 5,7p,8a-triOAc;2a,9a-H

Ref. 24,25 26 27 25 27,28 29 30 30 25 26,31 28 32 32 32 32 32 32 32 32 32 32 32

164 Other highly oxygenated bicyclic marasmane sesquiterpenes have been isolated from an EtOH extract of L. vellereus, which for many aspects seems an inexhaustible mine of Lactarius sesquiterpenes of any kind (see Table 24). The very unstable 5,10a,13-trihydroxymarasm-7(8)-ene (6.7) is accompanied by the 13-normarasmane isomers 6.12 and 6.13 (32) The latter ketones can derive from lactaropallidine (6.6) by p-elimination and oxidation at C-7. TABLE 7 - Heterocyclic Marasmane sesquiterpenes

NP 7.1 7.2 7.3 lA 7.5 7.6 7.7 7.8

Name No trivial name Isovellerol [n-^Hj]-Isovellerol ^^O-Isovellerol Rubrocinctal A 6-Ketostearoylrubrocinctal A Rubrocinctal B 6-Ketostearoylrubrocinctal B

7.9 No trivial name 7.10 No trivial name 7.11 Velutinal 7.12 in-^Hsl-Velutinal 7.13 Methylvelutinal 7.14 Stearoyl velutinal 7.15 6-Ketostearoylvelutinal 7.16 No trivial name 7.17 No trivial name 7.18 No trivial name 7.19 No trivial name 7.20 No trivial name 7.21 No trivial name 7.22 Isovellerol dimer

Substituents 5-oxo;7(8)-en;2a,9a-H 5-OH;7(8)-en;2a,9a-H 12-2H3;5-OH;7(8)-en;2a,9a-H i80;5-OH;7(8)-en;2a,9a-H 5-OH;12-oxo:7(8)-en;2a.9a-H 5-(6-oxostearoyloxy); 12-oxo;7(8)-en;2a,9a-H 5-OH;7(8)-en;2a,9a-H;12-acid Me ester 5-(6-oxostearoyloxy);7(8)-en;2a,9a-H; 12-acid Me ester 8a-OH;5a-OMe;7( 13)-en;2a,9a-H 8a-OH;5a-0-stearoyloxy;7( 13)-en;2a,9a-H 7a,8a-epoxy;5a-OH;2a,9a-H 12-2H3;7a,8a-epoxy;5a-OH;2a,9a-H 7a,8a-epoxy;5a-OMe;2a,9a-H 7a,8a-epoxy;5a-stearoyloxy;2a,9a-H 7a,8a-epoxy;5a-(6-oxostearoyloxy);2a,9a-H 9a,10a-diOH;5-oxo:7(8)-en;2a-H 9a,10a-diOAc;5-oxo;7(8)-en;2a-H 7a,8a-diOH;5-oxo;2a,9a-H 7a,8MiOH;5-oxo;2a,9a-H 7a-OH;8a-OAc;5-oxo;2a,9a-H 7a-OH;8p-OAc;5-oxo;2a,9a-H

Ref. 33 27,34 35 27 23 23 23 23 36 36 27,37,38 35 38,39 38,39,40 38 32 32 33 41 33 41 27

The following y-lactone sesquiterpenoids with the parent marasmane skeleton have also been isolated from L. vellereus: 13-OH-7(8)-en (7.1) (33), 7a,8a,13-tri-OH (7.18) (33), 7a,8p,13-triOH (7.19) (41), and 9a,10a,13-tri-OH-7(8)-en (7.16) (32) marasman-5-oic acid y-lactones. Their structures have been established by spectroscopic data of the natural products and of their acetyl

165 derivatives. The value of the coupling constant between the protons H-8 e H-9 in compounds 7.18 and 7.19 indicated the trans and the rather unusual cis configuration, respectively. Furthermore, stereoselective cis dihydroxylation with OSO4 of the C-7,8 double bond of 7.1 occurred from the convex side of the molecule, giving the diol 7.18 in good yield (33). The Polish authors suggested that compound 7.18 is formed in Nature by oxidation of lactone 7.1 and that the biogenesis of lactone 7.19 from velutinal ester (7.14) involves the oxidation of the hemiacetal to the lactone ring, which is then followed by trans diaxial opening of the epoxyde (33). Recently, we have isolated the aldehyde rubrocinctal A (7.5), the carboxymethyl ester rubrocinctal B (7.7), and the corresponding 6-oxostearoyl esters 7.6 and 7.8 from Lactarius ruhrocinctus (Section Ichorati) (23) (Table 7). These compounds are the first examples of 12oxygenated isovellerol derivatives from a Lactarius species. Rearranged marasmane skeletons: glutinopallane and isolactarane sesquiterpenes The only two known natural glutinopallane sesquiterpenes (8.2, 8.3) have been isolated from Lactarius glutinopallens (42) (Table 8). Their structures are strictly related to velutinal esters (7.14, 7.15) for the presence of the cyclopropane ring and the 7a,8a-epoxyde, and to rubrocinctals 7.7 and 7.8 for the carbomethoxy group at C-3. The fascinating lactone isolactarorufin (9.5) (Table 9) is the only example of isolactarane sesquiterpenes isolated from a Lactarius specie (43-45). The structure of isolactarorufin has been elucidated by spectroscopic data and confirmed by X-ray analysis of its p-bromobenzoate 9.8 (45). TABLE 8 - Glutinopallane sesquiterpenes 4

N°

Name

8.1 Methylglutinopallal 8.2 Palmitoylglutinopallal 8.3 Stearoylglutinopallal

12

COOMe ^ H

Substituents 5a-OMe 5a-palmitoyloxy 5a-stearoyloxy

Ref. 42 42 42

Lactarane, lactaranolide, secolactarane and related sesquiterpenes New sesquiterpenes of the lactarane (occasionally named also vellerane) group (Table 10) have been isolated by Swedish authors during their pioneering work on the chemical defence system of Lactarius species. Vellerol (10.8) and vellerdiol (10.20) were first isolated from extracts of Lactarius vellereus made at different times after grinding the mushrooms (27). The reduction of either

166 vellerol (10.8) or velleral (10.5) with KBH4 gave the identical diol 10.20 confirming the stereostructures, particularly the biogenetically important C3-H configuration. TABLE 9 - Isolactarane sesquiterpenes 12

12 1

4

14

^

JL

1

14

9.1-9.8

N°

Name

Substituents

9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10

No trivial name No trivial name No trivial name No trivial name Isolactarorufm (Lactarorufin C) No trivial name No trivial name No trivial name No trivial name No trivial name

8-oxo;2(9)-en 8a-OH;2(3)-en;9a-H 8a-OAc;2(3)-en;9a-H 3p-OH;8-oxo;2a,9a-H 3P,8a-diOH;2a,9a-H 3p,8P-diOH;2a,9a-H 3P-OH;8a-OAc;2a,9a-H 3P-OH;8a-OBz-p-Br;2a,9a-H 4,8a,13-triOH 4,8a,13-triOAc

Ref. A'hAA 43,44 43,44 43,44 43,44,45 43,44 43,44 45 44 44

TABLE 10 - Lactarane sesquiterpenes ,12

13 N°

Name

10.1 Chrysorrhedial 10.2 No trivial name 10.3 No trivial name 10.4 No trivial name 10.5 VeUeral 10.6 ^^0-Velleral 10.7 Chrysorrheal (Scrobicalol) 10.8 Vellerol 10.9 i^O-Vellerol

^\^ Substituents 5,13-dioxo;2(9),7(8)-dien;3a,6p-H 5,13-dioxo;3( 12),7(8)-dien;2a,6a,9a-H 5,13-dioxo; 3( 12),7(8)-dien;2a,6P,9a-H 5,13-dioxo;4(6),7(8)-dien;2a,3a,9a-H 5,13-dioxo;4(6),7(8)-dien;2a,3P,9a-H i80;5,13-dioxo;4(6),7(8)-dien;2a,3p,9a-H 5-OH; 13-oxo;2(9),7(8)-dien;3a-6p-H 13-OH;5-oxo;4(6),7(8)-dien;2a,3p,9a-H i80;13-OH;5-oxo;4(6),7(8)-dien;2a,3p,9a-H

Ref. 46 26 26 47 24,47,48 27 46,49 27 27

167 10.10 No trivial name 10.11 No trivial name 10.12 No trivial name 10.13 No trivial name 10.14 No trivial name 10.15 No trivial name 10.16 No trivial name 10.17 No trivial name 10.18 No trivial name 10.19 10.20 10.21 10.22 10.23 10.24 10.25 10.26 10.27 10.28 10.29 10.30 10.31 10.32 10.33 10.34 10.35 10.36 10.37 10.38

Chrysorrhediol VeUeidiol Piperdial Epi-piperdial No trivial name No trivial name No trivial name Piperalol Epi-piperalol No trivial name Pipertriol 7-£p/-pipertriol No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name

10.39 No trivial name 10.40 Vellerol dimer

13-OH;5-diOMe;4(6),7(8)-dien;2a,3a,9a-H 13-OH;5-diOMe;4(6),7(8)-dien;2a,3p,9a-H 5-diOMe;4(6)J(8)-dien;2a,3a,9a-H; 13-acid Me ester 5-diOMe;4(6).7(8)-dien;2a,3P,9a-H; 13-acid Me ester 13-diOMe;4(6),7(8)-dien;2a,3p,9a-H; 5-acid Me ester 13-OH;4(6),7(8)-dien;2a,3p,9a-H; 5-acid 5-OAc;4(6),7(8)-dien;2a,3a,9a-H; 13-acid Me ester 5-OAc;4(6),7(8)-dien;2a,3p,9a-H; 13-acid Me ester 13-OAc;4(6) J(8)-dien;2a,3p,9a-H; 5-acid Me ester 5,13-diOH;2(9),7(8)-dien;3a,6P-H 5,13-diOH;4(6),7(8)-dien;2a,3p,9a-H 8a-OH;5,13-dioxo;4(6)en;2a,3p,7a,9a-H 8a-OH;5,13-dioxo;4(6)en;2a,3p,7p,9a-H 9-OH;5,13-dioxo;4(6),7(8)-dien;2a,3p-H 3a,8a-epoxy;5,13-diOH;6(7)-en;2a,9a-H 3a,8a-epoxy;5,13-diOAc; 6(7)-en-2a,9a-H 8a,13-diOH;5-oxo;4(6)en;2a,3p,7a,9a-H 8a,13-diOH;5-oxo;4(6)en;2a,3p,7p,9a-H 5,8a,13-triOH;2(3)-en;9a-H 5,8a,13-triOH;4(6)-en;2a,3p,7a,9a-H 5,8a,13-triOH;4(6)-en;2a,3p,7p,9a-H 5,8a, 13-triOAc;2(3)-en;9a-H 5,8a,13-triOAc;4(6)-en;2a,3p,7p,9a-H 5,8a, 13-triOCONHCCl3;4(6)-en;2a,3p,7p,9a-H 5,8a,13-triOH;2a,3p,6a,7p,9a-H 5,8a, 13-triOH;2a,3p,6p,7p,9a-H 3a,5,8a,13-tetra-OH;6(7)-en;2a,9a-H 3a-OH;5,8a, 13-tri-OAc;6(7)-en;2a,9a-H 3a,5,8a,13-tetra-OH;2a,9a-H 3a-OH;5,8a, 13-tri-OAc;2a,9a-H

47 47 47 47 47 47 47 47 47 46,49 27,50 51 52 27 53 53 51 52 54 51 55 54 55 55 55 55 54 54 54 54 27

Vellerol (10.8), as well as isovellerol (7.2), readily dimerizes when left in reagent grade solvents. Velleral (10.5) was found to be degraded rapidly on attempted preparative chromatography on AI2O3 (27). Moreover, the dialdehyde 10.5 was slowly oxidized to 9-hydroxyvelleral (10.23) in a hexane solution kept at r. t. for two weeks or even at -30° C when a hexane extract of L. vellereus was kept frozen in the air for months (27). In the same conditions isovelleral (6.1) was oxidized to

168 9-hydroxyisovelleral (6.5) (27). This oxidation also occurred when isovelleral was adsorbed on AI2O3 for 5 h in day light. The equally labile and biologically active dialdehydes piperdial (10.21) and ^p/-piperdial (10.22) were isolated from different Lactarius species (Table 24). During chromatography on silica gel ^/7/-piperdial (10.22) was easily converted into piperdial (10.21), which appears to be the more stable epimer, and into velleral (10.5) (52). Like velleral, both piperdial and epi-piperdial rapidly decomposed within few seconds when chromatography on AI2O3 was attempted (113). Piperalol (10.26) was found together with piperdial (10.21) in the same mushrooms (51), while, in addition to compound 10.22, e/?/-piperalol (10.27) (52) and 7-ep/-pipertriol (10.30) (55) were isolated from different extracts of L. necator. In analogy with isovellerol and vellerol (27), piperalol (10.26) and ep/-piperalol (10.27) readily dimerized when exposed to traces of acid (113). NMR data, particularly NOE results, established the position of the formyl groups and the relative configuration of stereocenters of these sesquiterpenes. Moreover, Li AIH4 reduction of lactarorufm N (11.18) gave 7-e/7/-pipertriol (10.30) (55), and both compounds 10.21 and 10.26 were reduced by KBH4 in EtOH to the same triol named pipertriol (10.29) (51). A similar procedure was followed for establishing the stereostructures of two new pungent-bitter aldehydes, named chrysorrhedial (10.1) and chrysorrheal (10.7), that have been isolated from Lactarius scrobiculatus and L. chrysorrheus (46). UV absorption at 313 nm supported the presence of the dienal system in sesquiterpene 10.1, while the cis relationship between H3-I2 and H-6 in compounds 10.1 and 10.7 was firmly established by hydride reduction of both compounds to the same diol 10.19. The latter compound was also obtained by LiAlH4 reduction of the known lactone 16.2 (46). The positive Cotton effect observed for the n -> 7C* transition of the diene system in the diol 10.19 indicated a positive skewness for the chromophore in accordance with the absolute configurations shown in the formulae 10.1, 10.7, 10.19, and 16.2. Interestingly, although one might expect the 1,4-dialdehyde 10.1 and y-hydroxyaldehydes 7.2, 10.7, 10.26, and 10.27 to be in equilibrium with the corresponding hemiacetal forms, only the ^H-NMR spectra of isovellerol (7.2) showed significant amounts of a cyclic product (27). In fact, in common organic solvents isovellerol (7.2) exists as a mixture of approximately equal amounts of the three forms shown below in Scheme 5 (27).

On the other hand, molecular modelling of chrysorrheal (10.7) (46) clearly suggested that ring closure to the hemiacetal form would require a severe conformational rearrangement at a high energetic cost for the loss of the resonance energy deriving from the conjugation of the carbonyl group to the Cy-Cg double bond.

169 The entire group of lactarane sesquiterpenes 10.1, 10.7,10.19, and 16.2 containing the 2(9),7(8)-cycloheptadiene ring, was submitted to conformational analysis by molecular mechanics and ^H-NMR spectroscopy (46). We observed that conformational mobility of each compound is almost restricted to the interconversion of envelope forms of the cyclopentene ring; by contrast, only a single conformation of the seven membered ring is practically populated, owing to the planarity of either the diene (in 10.19) or the diene-carbonyl double bonds (in 10.1,10.7, and 16.2), and the rigid fusion of the y-lactone ring (in 16.2) (46). However, the geometry of the global minima of sesquiterpenes 10.1,10.7, 10.19 is completely different from that of 13-lactaranolide 16.2. In fact, the orientation of the 3-methyl group is pseudoequatorial in the former three compounds, while it is pseudoaxial in 16.2 (Figure 1).

MM2 computed conformations for compound 16.2

MM2 computed conformations for dialdehyde 10.1, hydroxyaldehyde 10.7, and diol 10.19 Figure 1 Recently, several papers have reported the isolation of new lactarane lactones, possessing the methyl group at C-3 either cis or trans to H-2, and the lactone carbonyl group either at C-5 (5lactaranolides, Tables 11-13) or at C-13 (13-lactaranolides, Tables 16-17). Differentiation between these structural alternatives on the basis of spectroscopic data alone has been often risky, especially when only a single isomer is at hand. Therefore, chemical correlations, synthesis of the possible isomers, and molecular mechanics calculations have always been performed in order to corroborate spectroscopic informations. For instance, the correction of the structure of vellerolactone (11.3) (47) led Daniewski and coworkers to revise the structures of lactarorufm N (11.18) and 3-ep/-deoxylactarorufin A (11.20) (58), that had been correlated with compound 11.3 (Scheme 6). Lactone 11.20 was also compared with the C-3 epimer 11.19 (Scheme 6), whose structure was unambiguously proved by synthesis and further chemical transformations (58).

170

H.IS^AH

11.19

H 6 "

11.3 Scheme 6 Lactarorufins D (11.47) and E (11.49), which are the 4a- and 4p-hydroxy derivatives of lactone 11.20, were isolated from L. necator (78). Single-crystal X-ray analysis demonstrated unequivocally the structure and relative stereochemistry of compound 11.47. The ^H-NMR data of lactarorufin E (11.49) were almost identical with those of epimeric lactarorufin D (11.47), except of course the vicinal coupling constant of the proton H-4 with H-3. A preparation of lactarorufins D and E from the more abundant lactarorufin A (11.44) was attempted in order to test their biological activity (77). Oxidative hydroboration of compound 11.13, followed by hydrolysis of intermediate boric esters with cone. HCl in EtOH, yielded lactarorufin A (11.44) (6.7%), S-^pMactarorufin D (11.46) (84.4%), 3-e/7Mactarorufin E (11.48) (1%), and lactarorufin E (11.49) (5.5%) (Scheme 7).

11.48 HO

: H 11.49 HO

Scheme 7 Several representative lactones showing the 8-OH and 3-Me groups cis to H-2 and H-9, as sardonialactone A (11.50), blennins A (11.17) and D, (11.43), 14-hydroxyblennin A (11.51), and lactarorufins D (11.47) and E (11.49) have been submitted to conformational analysis (74, 78, 81). Molecular mechanics showed that, in spite of the different position of the double bond in the

171 cycloheptene ring, a hinge conformation, in which both OH and Me groups are equatorially oriented, is largely preferred by these compounds. By contrast, the strong intramolecular hydrogen bond between the two cis OH groups at C-3 and C-8 of lactarorufms A (11.44) and B (11.71) is the dominant steric factor which affects the overall molecular shape of these two lactones. In fact, in order to form this intramolecular bond, the cycloheptene rings of lactarorufms A (11.44) and B (11,71) must still assume a hinge conformation but folded in an opposite direction to that of 11.47 and 11.49 (Figure 2). For 3-ep/-lactarorufin D (11.46) for which a conformation similar to lactarorufin A (11.44) was expected, dynamic NMR studies indicated the existence of two conformations in solution at r. t. (77).

R = H R' = OH (11.47) R = OH R» = H (11.49)

R = R»=R"=H (11.17) R = R'=H R" = OH (11.43) R'=R" = H R = OH (11.50) R = R"==H R'= OH (11.51)

R = H (11.44) R = OH (11.71)

MM2 computed conformations for compounds 11.17,11.43,11.44,11.47,11.49,11.50,11.51, and 11.71 Figure 2 Additional new identified lactarorufms are 3,12-anhydrolactarorufin A (11.11) from L. necator (63), and 15-hydroxylactarolide A (13-hydroxylactarorufm B) (11.83) from L. mitissimus (83). The structure of the latter compound was confirmed (83) by NaBH4 reduction of the lactol group yielding lactarorufin B (11.71), whose structure had been confirmed by X-ray analysis (72). Furthermore, compound 11.83 was synthesized from lactarorufin B (11.71) (82) by introducing the C-13 hydroxy group in two steps. DIBAL reduction of lactarorufin B gave furantriol (18.27), which was acetylated to the corresponding diacetate 18.28, and this was then oxidized with MCPBA to 11.83. Interestingly, the ^H-NMR spectrum of lactol 11.83 showed only one signal for the proton H-13, owing to a mixture of fast equilibrating epimers. The rate of equilibration was much higher in MeOH than in CHCI3 (82). The 3,8-internal ether of lactarorufin A (11.15) was isolated from L. necator (59), and it was identical with the compound previously obtained from lactarorufin A (11.44) by dehydration with MsCl-Py (53). By comparison with the synthetic isomer 13-oxolactone 16.10 (59), the ^H-NMR spectra of compound 11.15 showed small differences in the chemical shifts, that were attributed to the shorter distance of the protons of 11.15 from the ether oxygen. In addition, the Polish authors found that the acid catalysed dehydration of furandiol (18.14) and 5-deoxylactarolide B (16.15) to the corresponding 3,8-internal ether 18.5 and 16.10, respectively, could be achieved in modest yields by the azeotropic method of removal of water (59) (Scheme 8).

172 The same authors assigned the structure 11.1 to a new polyunsaturated 5-lactaranolide sesquiterpene isolated from an ethanol extract of L. vellereus (56).

p-TsOH CfiHT

Scheme 8

16.10

TABLE 11 - 5-Lactaranolide sesquiterpenes 12

N°

Name

Substituents

11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 11.10 11.11 11.12 11.13 11.14 11.15 11.16 11.17 11.18 11.19 11.20 11.21 11.22

No trivial name Pyrovellerolactone Vellerolactone No trivial name Lactarotropone No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name Blennin A Lactarorufin N No trivial name No trivial name No trivial name No trivial name

2(3),6(7),8(9)-trien 3(4),6(7)-dien;2a,9a-H 4(6)J(8)-dien;2a,3p,9a-H 8,9-en;2a,3a,6a,7a-H 8-oxo;2(9),3(4),6(7)-trien 3a-OH;6(7).8(9)-dien;2a-H 8a-0H; 1 (2),6(7)-dien;3a,9a-H 8a-OH;l(2),6(7)-dien;3a,9P-H 8a-OH;2(3),6(7)-dien;9a-H 8a-OH;3(4),6(7)-dien;2a,9a-H 8a-OH;3(12),6(7)-dien;2a,9a-H 8a-OAc;2(3),6(7)-dien;9a-H 8a-OAc;3(4),6(7)-dien;2a,9a-H 8a-OAc;3( 12),6(7)-dien;2a,9a-H 3a,8a-epoxy;6(7)-en;2a,9a-H 15-D3;3a,8a-epoxy;6(7)-en;2a,9a-H 8a-OH;4(6)-en;2a,3p,7a,9a-H 8a-OH;4(6)-en;2a,3p,7p,9a-H 8a-OH;6(7)-en;2a,3a,9a-H 8a-OH;6(7)-en;2a,3p,9a-H 8a-OH;6(7)-en;2p,3a,9a-H 8a-OH;6(7)-en;2p,3p,9a-H

Ref. 56 47,57 47,57 58 29 59 60 60 61,62 62 63 61 62,64,65 63 53,59 66 67,68 58,62 58,69 58,69 60 60

173

11.23 11.24 11.25 11.26 11.27 11.28 11.29 11.30 11.31 11.32 11.33 11.34 11.35 n.36 11.37 11,38 11.39 11.40 11.41 11.42 11.43 11.44 11.45 11.46 11.47 11.48 11.49 11.50 11.51 11.52 11.53 11.54 11.55 11.56 11.57 11.58 11.59 11.60 11.61 11.62 11.63 11.64 11.65

No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name Blennin B No trivial name Blennin D Lactaronifin A No trivial name No trivial name Lactarorufin D No trivial name Lactarorufin E Sardonialactone A No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name

8p-OH;6(7)-en;9a-H 29 8a-OAc;3(4)-en;2a,3p,7p,9a-H 62 8a-OAc;4(6)-en;2a,3p,7a,9a-H 67 8a-OAc;6(7)-en;2a,3a,9a-H 58,69 8a-OAc;6(7)-en;2a,3p,9a-H 69 8a-OMs;4(6)-en;2a,3p,7a,9a-H 68 8-oxo;2a,3a,6a,7a,9a-H 58,62 8a-OH;2a,3a,6a,7a,9a-H 58,62 8p-OH;2a,3a,6a,7a,9a-H 58 8a-OAc;2a,3a,6a,7a,9a-H 58,62 2,9-epoxy;8-oxo;3(4),6(7)-dien 70 3a,4a-epoxy;8a-OAc;6(7)-en;2a,9a-H 71 3a,8a-epoxy;15-(OBz-/?-Br);6(7)-en;2a,9a-H 66,72 3a,8a-epoxy;15-OMs;6(7)-en;2a,9a-H 66 3a,8a-epoxy;15-OTs;6(7)-en;2a,9a-H 66 3a-OH;8-oxo;2(9),6(7)-dien 70 3a-OH;8-oxo;6(7)-en;2a,9a-H 53 3a-OEt;8-oxo;6(7)-en;2a,9a-H 73 8a, 13-diOH;2(3),6(7)-dien;9a-H 67 8a,15-diOAc;3(4),6(7)-dien;2a,9a-H 66 2a,8a-diOH;4(6)-en;3p,7a,9a-H 68,74 3a,8a-diOH;6(7)-en;2a,9a-H 53,64,72,75,76 3a,8p-diOH;6(7)-en;2a,9a-H 53 4a,8a-diOH;6(7)-en;2a,3a,9a-H 77 4a,8a-diOH;6(7)-en;2a,3p,9a-H 78 4p,8a-diOH;6(7)-en;2a,3a,9a-H 77 4p,8a-diOH;6(7)-en;2a,3p,9a-H 77,78 7a,8a-diOH;4(6)-en;2a,3p,9a-H 29,78,79 8a,14-diOH;4(6)-en;2a,3p,7a,9a-H 80,81 8a-OH;3a-OEt;6(7)-en;2a,9a-H 54,73 2a-OH;8a-OAc;4(6)-en;3P,7a,9a-H 68 3a-OH,8a-OAc;6(7)-en;2a,9a-H 64,65,82 3a-OH;8p-OAc;6(7)-en;2a,9a-H 53 4a-OH;8a-OAc;6(7)-en;2a,3a,9a-H 77 4P-OH;8a-OAc;6(7)-en;2a,3a,9a-H 77 7a-OH;8a-OAc;4(6)-en;2a,3p,9a-H 79 3a-OH;8a-stearoyloxy;6(7)-en;2a,9a-H 71 3a-OEt;8a-OAc;6(7)-en;2a,9a-H 73 3a,8a-diOAc;6(7)-en;2a,9a-H 64 3a,8a-diO(S02);6(7)-en;2a,9a-H 53 3a-OH;8-oxo;2a,6a,7a,9a-H 53,62 3a,8a-diOH;2a,6a,7a,9a-H 64 3a,8a-diOH;2a,6p,7P,9a-H 64

174 11.66 No trivial name 11.67 No trivial name 11.68 No trivial name 11.69 No trivial name 11.70 Lactarolide A 11.71 Lactarorufin B 11.72 3-Ethyl-lactarolicie A 11.73 No trivial name 11.74 No trivial name 11.75 No trivial name 11.76 No trivial name 11.77 No trivial name 11.78 No trivial name 11.79 No trivial name 11.80 No trivial name 11.81 No trivial name 11.82 No trivial name 11.83 No trivial name 11.84 No trivial name 11.85 No trivial name 11.86 No trivial name

3a-OH;8a-OAc;2a,6,7,9a-H 64 8a-OH;3a-OAc;2a,6,7,9a-H 64,69 3a,8a-diOAc;2a,6,7,9a-H 64 3a,13-diOH;8-oxo;6(7)-en;2a,9a-H 54 3a,8a,13-triOH;6(7)-en;2a,9a-H 54,65,82,83 3a,8a, 15-triOH;6(7)-en;2a,9a-H 64,66,72,84 8a,13-diOH;3a-OEt;6(7)-en;2a,9a-H 54 3a, 13-diOH;8a-OAc;6(7)-en;2a,9a-H 54,65,82 3a,8a-diOH; 15-OBz-/7-Br;6(7)-en;2a,9a-H 66 3a,8a-diOH; 15-OTs;6(7)-en;2a,9a-H 66 3a-OH;8a,13a-diOAc;6(7)-en;2a,9a-H 54,83 3a-OH;8a,13p-diOAc;6(7)-en;2a,9a-H 54,83 3a-OH;8a,15-diOAc;6(7)-en;2a,9a-H 66 3a-OH;8a, 15-diOBz-/?-Br;6(7)-en;2a,9a-H 66 3a-OH;8a,15-diOTs;6(7)-en;2a,9a-H 71 3a-OEt;8a, 13-diOAc;6(7)-en;2a,9a-H 54 3a,8a,15-triOAc;6(7)-en;2a,9a-H 66 3a,8a,13,15-tetraOH;6(7)-en;2a,9a-H 82,83 3a,13-diOH;8a,15-diOAc;6(7)-en;2a,9a-H 82 3a-OH;8a, 13a, 15-triOAc;6(7)-en;2a,9a-H 82 3a-OH;8a, 13p, 15-triOAc;6(7)-en;2a,9a-H 82

TABLE 12 - 5-Lactaranolide derivatives /12

rsp

Name

Substituents

12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8 12.9

No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name

5-OMe;4(6),7(8)-dien;2a,9a-H 3a,8a-epoxy;5-OH;6(7)-en;2a,9a-H 3a,8a-epoxy;5P-OMe;6(7)-en;2a,9a-H 8a-OH;5p-OMe;2(3),6(7)-dien;9a-H 8a-OH;5p-OMe;3(12),6(7)-dien;2a,9a-H 3a,5,8a-triOH;6(7)-en;2a,9a-H 3a,8a-diOH;5-OMe;6(7)-en;2a,9a-H 8a,5-diOH;3a-OMe;6(7)-en;2a,9a-H 8a-OH;3a,5a-diOMe;6(7)-en;2a,9a-H

Ref. 47 85 85 85 85 85 85 85 85

175 Three new 8-oxolactarane lactones 11.5, 11.33 and 11.38 were isolated from L. scrohiculatus (70) and L. pallidas (29). Ketones 11.5 and 11.38 could also be obtained by PDC oxidation of 2,3-anhydrolactarorufin A (11.9), while the a,P-cpoxyketone 11.33 was considered a biogenetic precursor of the 8-norlactarane sesquiterpene 15.1 (70). Expulsion of the C-8 carbonyl group from the lactarane skeleton of compound 11.33 has been suggested to occur via a benzylic-like rearrangement, followed by a decarboxylative aromatization (70) (Scheme 9).

0

y^-^v

^OFT

0

0 11.33

Scheme 9 The relative configurations of lactaroscrobiculide A (16.2) and the corresponding 2,9-epoxyde 16.9 have been definitively established by NOE experiments and molecular modelling performed by means of the MM2 program (46). To this purpose, the conformational spaces of the stereostructures 16.2 and 16.9 were explored and compared with the corresponding diastereomers 16.2 A and 16.9 A. It was then found that the experimental ^H-NMR vicinal coupling constants of the two natural sesquiterpenes matched the values calculated for stereoisomers 16.2 and 16.9 instead of those computed for compounds 16.2 A and 16.9 A.

O

1^-^

O

16.9 A

Epoxyde 16.9 is strongly suspected of being an artifact, because it was not found any more in

176 fresh extracts of L. scrobiculatus (46) and it was obtained by oxidation in air of dienelactone 16.2. This oxidation, as well as that with MCPBA, showed a rather surprising diastereoselectivity, since oxidizing agents approached the apparently more hindered face of the C2-C9 double bond of compound 16.2.

TABLE 13 - Rearranged 5-lactaranolide sesquiterpenes

NO

Name

Substituents

13.1 13.2 13.3

No trivial name No trivial name No trivial name

(See formula) 13-OH 13-OAc

Ref. 68 58 58

TABLE 14 - 8,9-Seco-5-lactaranolide sesquiterpenes and derivatives .12

rsp

Name

Substituents

14.1 14.2 14.3 14.4 14.5 14.6

No trivial name No trivial name No trivial name Lactardial Blennin C (Lactaronecatorin A) No trivial name

5-oxo;6(7)-en 5-oxo 5,8-dioxo;2(9),6(7)-dien 5-OH;8-oxo;2(9),6(7)-dien 8-OH;5-oxo;2(9),6(7)-dien 5a-OMe;8-oxo;2(9),6(7)-dien

14.7

No trivial name

8-OAc;5-oxo;2(9),6(7)-dien

Ref. 67 67 49,60 51,85, 61,67,86 85 61

The extremely labile triene-enol-lactone 16.1 was found to be involved in the rapid yellowing of the milky juice and flesh of L. chrysorrheus and L. scrobiculatus, and it could be isolated from the mushrooms under special mild conditions (46). The strong UV absorption of compound 16.1 at

177 370.4 nm was that expected for the cross-conjugated dienone-triene chromophores, while the further unsaturation in the furanone ring was indicated by comparison of the NMR data of compound 16.1 with lactone 16.2. Biosynthetic considerations suggested for lactone 16.1 the same absolute configuration of sesquiterpenes 10.1, 10.7, and 16.2. The new lactone 16.6, one of the few known natural 13-lactaranolides, has recently been isolated from L. vellereus (87). The simulated ^^C-NMR spectra of compound 16.6 suggested that the configuration at C-3 was opposite to that of isomeric lactaroscrobiculide A (16.2). This stereochemistry was established unequivocally by correlation of sesquiterpene 16.6 with 3-deoxy-3^/7/-lactarorufin A (11.20), as shown in Scheme 10 (87). H $.

. > ^ 1)DIBAL / \ 2)AC20/Py ^ ' AcO 1) MCPBA 2)NaBH4 3) Separation

11.20 H 5^

MeOH/H O

: H AcO 11.27 (52.5%)

16.6 Scheme

6

AcO 15.13 (18.9%)

10

For comparison, 3-deoxylactaroscrobiculide B (16.5), the C-3 epimer of natural lactone 16.6, was synthesized from 8-acetyl-5-deoxylactarolide B (16.17), as shown in Scheme 11 (87).

MeOH/H"^ reflux

S c h e m e 11

O

Comparing the ^H-NMR spectra of the epimeric pairs 1 1 . 2 6 , 11.27 and 1 6 . 1 2 , 1 6 . 1 3

178 Daniewski and coworkers noticed that the signal of the proton H-3 cis to the C-8 acetoxy group was shifted to lower field than that of the corresponding trans proton (87). TABLE 15 - 8-Norlactarane sesquiterpenes

O—'12

N°

Name

Substituents

15.1

No trivial name

(See formula)

Ref. 70

TABLE 16 - 13-Lactaranolide sesquiterpenes ,12

N'

Name

Substituents

16.1 16.2 16.3 16.4 16.5 16.6 16.7 16.8 16.9 16.10 16.11 16.12 16.13 16.14 16.15 16.16

Chrysorrhelactone LactaroscrobicuHde A No trivial name No trivial name No trivial name No trivial name LactaroscrobicuHde B No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name

2(9),5(6),7(8)-trien;3a-H 2(9),7(8)-dien;3a,6p-H 3(4),6(7)-dien;2a,9a-H 4(6).7(8)-dien;2a,3a,9a-H 6(7),8(9)-dien;2a,3a-H 6(7),8(9)-dien;2a,3p-H 3a-OH;6(7),8(9)-dien;2a-H 8a-OAc;3(4),6(7)-dien;2a,9a-H 2p,9p-epoxy;7(8)-en;3a,6p-H 3a,8a-epoxy;6(7)-en;2a,9a-H 3a-OH;6(7)-en;2a,9a-H 8a-OAc;2a,3a,9a-H 8a-OAc;2a,3p,9a-H 3a-OEt;8-oxo;6(7)-en;2a,9a-H 3a,8a-diOH;6(7)-en;2a,9a-H 8a-OH;3a-OEt;6(7)-en;2a,9a-H

Ref. 46 86,46 47 47 87 87 82,88 65 46,89 59 54,88 87 87 73 54,59,65,82 54,73

179 16.17 16.18 16.19 16.20 16.21 16.22 16.23 16.24 16.25 16.26 16.27 16.28 16.29 16.30

No trivial name No trivial name No trivial name LactarolideB 3-Ethyl-lactarolide B No trivial name No trivial name No trivial name No trivial name No trivial name N^ rnv/a/ name No trivial name M? rrzv/fl/ «ame No trivial name

3a-OH;8a-OAc;6(7)-en;2a,9a-H 3a-OEt;8a-OAc;6(7)-en;2a,9a-H 3a,5-diOH;8-oxo;6(7)-en;2a,9a-H 3a,5,8a-triOH;6(7)-en;2a,9a-H 5,8a-diOH;3a-OEt;6(7)-en;2a,9a-H 3a,5-diOH;8a'OAc;6(7)-en;2a,9a-H 3a-OH;5,8a-diOAc;6(7)-en;2a,9a-H 3a-OH;5a,8a-diOAc;6(7)-en;2a,9a-H 3a-OH;5p,8a-diOAc;6(7)-en;2a,9a-H 3a-OH;8a, 15-diOAc;6(7)-en;2a,9a-H 3a-OEt;5,8a-diOAc;6(7)-en;2a,9a-H 3a,5-diOH;8a, 15-diOAc;6(7)-en;2a,9a-H 3a-OH; 8a,5a, 15-triOAc;6(7)-en;2a,9a-H 3a-OH;8a,5p,15-triOAc;6(7)-en;2a,9a-H

54,65,82 73 54 54,65,82 54 54,82 54 82 82 82 54 82 82 82

TABLE 17 - 13-Lactaranolide derivatives ,12

OEt N°

Name

Substituents

17.1

No trivial name

2(9),7(8)-dien;3a,6p-H

Ref. 46

Furanolactarane and secofuranolactarane sesquiterpenes Recently, relatively few new furanolactarane sesquiterpenes have been isolated from Lactarius species (Table 18). Furanodiene (18.1), previously obtained by synthesis (Scheme 12) (60), is also a true metabolite of Lactarius scrobiculatus (46), while the structure of the highly oxidized dioxofuran 18.26, isolated from L. vellereus (56), was confirmed by single crystal X-ray diffraction analysis (95). 3-E/7/-furandiol (18.15) and 4a,8a-dihydroxyfuran (18.16), two isomers of the more widespread 3a,8a-dihydroxyfuran (18.14), have been isolated from L. scrobiculatus (70), and from L. piperatus (94), L. torminosus (94), L. necator (94) andL. circellatus (52), respectively. The relative configuration of furan 18.16 was established by comparing the experimental ^H-NMR coupling constants of sesquiterpene 18.16 with those calculated for all possible stereoisomers (94). The molecular mechanics (MM2) computed conformational mixture of compound 18.16 comprises three conformers, two of which, 18.16 A and 18.16 B, differ mainly in a twist of the cyclopentane ring, whereas the third conformer 18.16 C (40%) shows an entirely different folding

180 of the seven membered ring, which facilitates the formation of an intramolecular hydrogen bond across the ring (94) (Figure 3). The results of molecular modelling (74) also definitively proved the stereostructure 18.13 for furoscrobiculin D, correcting a previous assignment based on the NMR data alone (88). In the preferred conformation 18.13 A of furoscrobiculin D, accounting for more than 95% the entire population, the C-3 methyl and the C-8 hydroxy groups have an equatorial orientation, as in conformers 18.16 A and 18.16 B (Figure 3).

18.16B

18.16C

MM2 computed conformations for compounds 18.13 and 18.16

Figure 3 Furantriol (18.27), isolated from L. mitissimus (84), is one of the few lactarane sesquiterpenes in which one of the gem-methyl groups at C-11 is oxidized and it was chemically correlated (82) with hictarorufin B (11.71), another example of this kind. The Polish authors suggested that lactone 11.71 was enzymatically formed from furan 18.27, and that a C-15 oxidized sesquiterpene of the velutinal type was the common precursor of both compounds in the mushroom (84). Actually, the possibility for the C-15 methyl group to be oxidized at an eariy stage of the lactarane biosynthesis seems to be confirmed by the recent finding of C-15 hydroxylated protoilludane sesquilerpenoids (5.1 and 5.2) in L. violascens (23) (Table 5). Further chemical correlations put stereochemical assignments of most Lactarius sesquiterpenes on a solid basis. Attempted formation of the bromide from the furanosesquiterpene 18.10 with PhaP and CBr4 gave furanether A (18.5) and pyrovellerofuran (18.3), as main products (93). The latter compound had previously been obtained by thermal rearrangement of isovelleral (6.1) (90), while furanol 18.10, isolated from L. vellereus (93), was also formed when velutinal derivatives were decomposed on silica gel (85). This experiment correlated the absolute configuration of isovelleral (6.1) with that of velutinal esters and, indirectly, with the stereostructures of many oihtr Lactarius

181 sesquiterpenes (93). TABLE 18 - Furanolactarane sesquiterpenes ,12

N°

Name

Substituents

18.1 18.2 18.3 18.4 18.5 18.6 18.7 18.8 18.9 18.10 18.11 18.12 18.13 18.14 18.15 18.16 18.17 18.18 18.19 18.20 18.21 18.22 18.23 18.24 18.25 18.26 18.27 18.28

No trivial name No trivial name Pyrovellorofuran No trivial name FuranetherA FuranetherB Furoscrobiculin B Furosardonin A Furanol No trivial name No trivial name Furoscrobiculin A Furoscrobiculin D Furandiol 3-£p/-furandiol No trivial name 3-O-Methylfurandiol Furoscobiculin C 3-0-Ethylfurandiol No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name Furantriol No trivial name

l(2),8(9)-dien;3a-H 2(3),8(9>dien 3(12),8(9)-dien;2a-H 3(4)-en;2a,9a.H 3a,8a-epoxy;2a,9a-H 3p,8p-epoxy;2a,9a-H 3a-OH;8(9)-en;2a-H 8a-OH;l(2)-en;3a,9a-H 8a-OH;2(3)-en;9a-H 8a-OH;3(12)-en;2a,9a-H 8a-OAc;2a,3p,9a-H 2,9-epoxy;8-oxo 2p,8a-diOH;3a,9a-H 3a,8a-diOH;2a,9a-H 3p,8a-diOH;2a,9a-H 4a,8a-diOH;2a,3p,9a-H 8a-OH;3a-OMe;2a,9a-H 3a-OH;8a-OEt;2a,9a-H 8a-OH;3a-OEt;2a,9a-H 2P-OH,8a-OAc;3a,9a-H 3a-OH;8a-OAc;2a,9a-H 3p-OH;8a-OAc;2a,9a-H 3a,8a-diOEt;2a,9a-H 3a-OEt;8a-OAc;2a,9a-H 4a,8a-diOAc;2a,3p,9a-H 4,8-dioxo;3P-OH;2(9)-en 3a,8a,15-triOH;2a,9a-H 3a-OH;8a,15-diOAc;2a,9a-H

Ref. 46,60,85 85 90 47 59,79,88 88 88 60,79 91,92 93 87 88 74,88 86,88,91 70 94 85,92 88 54,73,88,92 74,88 65,86 70 88 73 94 56,95 82,84 82,84

The series of reactions shown in Scheme 12 proved the same absolute configuration at C-3 for all natural lactone and furan secolactarane sesquiterpenes (60).

182

18.8

0

.

v^ 14.5

(^

V ^

DIBAL

\^ c c

OH O O^

1 r

14.3

o

11.8

OH

Scheme 12 TABLE 19 - 8,9-SecofuranoIactarane

sesquiterpenes

.12

CHOHCHjCOMe 19.1-19.4

19.5

N°

Name

Substituents

19.1 19.2 19.3 19.4 19.5

Lactaral No trivial name Lactarol No trivial name No trivial name

8-0X0

8-COOMe 8-OH 8-OTHP (See fomiula)

Ref. 96,97 96 51,60,97 97 88

Noteworthy among these conversions are the first successful cyclizations of lactone and furan secolactaranes to the corresponding lactarane sesquiterpenes, which were obtained by a Me2AlCl catalysed ene reaction (60). Under these conditions lactaral (19.1) yielded direcdy the diene 18.1, identical with the dehydration product of furosardonin A (18.8), while smooth cyclization of aldehyde 14.3 gave the lactone 11.8 in which the protons H-8 and H-9 have the "unnatural" cis stereochemistry. This result could be anticipated by examination of the Dreiding models of the two possible transition states 11.8 A and 11.8 B, which showed that unfavourable steric interactions developing between the C-3 methyl group and the bulky >C=0-"A1~ complex are minimized in the

183 transition state 11.8 B leading to lactone 11.8, O CH3

11.8 A

Structure elucidation of many Lactarius sesquiterpenes often requires interconversions of yhydroxybutenolide, butenolide, and furan rings for confirming spectroscopic assignments. Examples of DIBAL reductions of y-lactones to the corresponding furans include the conversions of blennin C (14.5) to lactarol (19.3) (Scheme 12) (60) and of compound 11.20 to 3-deoxy-3-e/7i-furandiol (87) (Scheme 10).

18.21 AcO

O AcO

11.73

HO AcO 29.6 •*" HO _ O _ J ^ H

16.22 25.3 %

Scheme 13

AcO

11-54 22.6 %

184 TABLE 20 - Rearranged furanolactarane sesquiterpenes

Q H N°

Name

Substituents

20.1

No trivial name

(See formula)

Ref. 25

The reverse transformation of a furan to a butenolide ring has been achieved in fair to good yields with NBS in aq. dioxane (Wiesner procedure) (122), as in the following conversions: 18.8 to 11.7 (60), 18.9 to 11.9 (60), and 18.21 to 16.22, 11.54 and 11.73 (82) (Scheme 13). The observed moderate sitoselectivity of the furan ring oxidation at C-5 was attributed to a coordination of the electrophilic Br"^ species with the allylic C-8-OR group, prior to the attack on the aromatic ring (60).

RO 18.19 R = H 18.24 R = Ac

: H RO 11.52 R = H 11.60 R = Ac

o 1 : 3.5 1 :5.7

RO 16.16 R=H 16.18 R=Ac

Scheme 14 MCPBA oxidation of several furanolactarane sesquiterpenes to the corresponding lactarolides (mixture of lactol epimers) has been studied in details by Daniewski and coworkers (59, 65, 73, 82, 84, 87). The sitoselectivity of this oxidation is only moderate and the directing effect of neighbouring oxygenated groups is often unpredictable, so that variable mixtures of C-5 and C-13 lactols are usually obtained. Smooth NaBFit reduction of the separated lactols afforded the corresponding ylactones with the carbonyl group either at C-13 or at C-5. The entire sequence of reactions (MCPBA (or NBS) furan oxidation - NaBH4 lactol reduction) allowed several important correlations of furanolactarane and lactaranolide sesquiterpenes (59, 60, 65, 73, 82, 83, 84, 87), as already reported in Scheme 10 and further illustrated by the examples of Scheme 14.

185

Dibenzonaphthyridinone alkaloids, prenylated phenols, benzofurans and chromenes Sesquiterpenes, as already reported in the previous sections, are the most widespread Lactarius metabolites; however, a few species possess a particular metabolism which leads to secondary metabolites of other classes. Moreover, interesting new compounds with a different biogenesis have been isolated also from species producing large quantities of sesquiterpenes. Interest in the considerable mutagenicity of extracts of Lactarius necator, a mushroom often cited in this review for the occurrence of several lactarane sesquiterpenes, led to the isolation of a highly mutagenic alkaloid named necatorin (4.8 mg from 30 kg of mushroom), for which the structure of 7hydroxycoumaro[5,6-c]cinnoline was originally proposed (101). Necatorin was then shown by direct comparison (100) to be identical with necatorone, isolated almost at the same time by Steglich (99) as one of the pigments of the fruiting bodies of L. necator. Spectroscopic data of this unstable alkaloid established the unusual 5,10-dihydroxydibenzo[de,h][l,6]-naphthyridin-6-one structure (21.2), which was confirmed by total synthesis (100). Necatorone forms red needles which dissolve in DMSO to produce a grass-green solution showing strong green-yellow fluorescence. With aq. ammonia, successive deprotonations of compound 21.2 produce blue and purple anions. Therefore, necatorone is believed to be partially responsible for the change to a deep purple of the dark olivebrown colour of the caps and stalks of L. necator on exposure to ammonia vapours. TABLE 21 - Dibenzonaphthyridinone alkaloids

21.1-21.3

N°

Name

Substituents

21.1 21.2 21.3 21.4 21.5 21.6

10-Deoxynecatorone Necatorone (Necatorin) No trivial name 10,10'-Dideoxy-4,4'-binecatorone 10-Deoxy-4,4'-binecatorone 4,4'-Binecatorone

5-OH 5,10-diOH 5,10-diOMe (See formula) lO'-OH 10,10'-diOH

Ref. 98 99, 100,101 99 98 98 98

186 Necatorone (21.2) was methylated by CH2N2 in methanol/H20 to yield the dimethyl ether 21.3 as the main product. More recendy, other two new necatorone-type alkaloids isolated from L. necator have been identified as 4,4'-binecatorone (21.6) and 10-deoxy-4,4'-binecatorone (21.5) (98). From L. atroviridis, a dark-green North American species, in addition to compounds 21.2, 21.5 and 21.6, 10,10'-dideoxy-4,4'-binecatorone (21.4) was obtained as main alkaloid (98). The structures of all these alkaloids have been established by spectroscopic data and confirmed by synthesis (98). The occurrence of the same alkaloids in L. necator and in L. atroviridis indicates the close taxonomic relationship of both species. Like necatorone, the colour of the DMSO solutions of alkaloids 21.5 and 21.6 changes to purple on addition of alkali, while that of compound 21.4 gives a dove-grey colour with alkali. It is noteworthy that in young, light brown fruiting bodies of L. necator about equal amounts of pigments 21.2 and 21.6 are present, whereas in aged, dark brown specimens the ratio between these compounds becomes 5 : 95. In search for the compounds responsible for the antimicrobial and immunosuppressive activities of L. flavidulus, an edible mushroom in spite of the bitter taste, three geranylphenols have been isolated and named flavidulols A (22.8), B (22.13) and C (22.15) (103, 104). The structure of flavidulol A (22.8) is very similar to that of wigandol isolated from Wiganda kunthii Choisy, the former compound being the methyl ether and the latter the acetate of the same phenol. Flavidulol B (22.13) could be an artifact derived from flavidulol A by a Cope-type rearrangement. The structures of all the flavidulols and their acetyl derivatives (Table 22) could be determined by spectroscopic studies. Particularly, NOE and ^^C-^H-COLOC NMR techniques allowed to establish the configuration of the double bonds in the geranyl moiety of compounds 22.8 and 22.15 as well as the cis stereochemistry at C-2 and C-7 of flavidulol B (22.13) (104). Catalytic hydrogenation of compound 22.8 afforded dihydro and tetrahydro derivatives, 22.11 and 22.12, respectively, while on treatment with 2N HCl in MeOH flavidulol A (22.8) gave two linear tricyclic products 22.17 and 22.18 (104). Recently, geranylgeranylhydroquinone (22.6) and a mixture of fatty acid esters 22.7 have been isolated from L. lignyotus (23). Clearly, these phenols are biogenetically related to flavidulols A-C and to compound 22.1. Compound 22.6 could also be obtained by hydrolysis of the esters 22.7. The acids esterified in 22.7 were identified by GC-MS analysis of the mixture of methyl esters obtained by transesterification (23). Interestingly, the free hydroquinone 22.6 has previously been isolated from the sponge Ircinia muscarum (123) and from plants of the genus Phacelia (124). In a preliminary study on the metabolites of the Lactarius species of the Section Plinthogali the fruiting bodies were extracted by grinding under solvents at r. t.. Unexpectedly, on TLC plates sprayed with the sulfo-vanillin reagent, the metabolites of these species were revealed as green spots, and, therefore, they could easily be differentiated from the metabolites of the ox\\tr Lactarius species. In fact, separation of EtOAc extracts of L. fuliginosus and L. picinus by silica gel column chromatography led to the isolation of benzofuran and chromene derivatives, unprecedented among Basidiomycetes metabolites (102,105) (Tables 22 and 23).

187 TABLE 22 - Prenylated phenols

22.6 and 22.7

OMe

OMe 22.8-22.12

OMe

10

OMe

22.15-22.16

22.13-22.14

OMe 22.17-22.18

N^

Name

Substituents

22.1 22.2 22.3 22.4 22.5 22.6 22.7 22.8 22.9 22.10 22.11 22.12 22.13 22.14 22.15 22.16 21 Al 22.18

No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name Flavidulol A No trivial name No trivial name No trivial name No trivial name Flavidulol B No trivial name Flavidulol C No trivial name No trivial name No trivial name

4'-OH;2(3)-en 4'-stearoyloxy;2(3)-en 2,3-epoxy;4'-OH 4'-[2-OH-3-(3-Me-2-butenyl)-5-OMe-phenoxy];2(3)-en 4'-[2-OAc-3-(3-Me-2-butenyl)-5-OMe-phenoxy];2(3)en 2',5'-diOH 2',5'-diOAcyl (*) 4*-OH;Z-2{3),E-6(7)-dien 4'-OMe;Z-2(3),E-6(7)-dien 4'-OAc;Z-2(3),E-6(7)-dien 4'-OH;Z-2(3)-en 4'-0H 4'-0H 4'-0Ac 4',4'-diOH 4',4'-diOAc 6P-0H 6P-0Me

Ref. 102 102 102 102 102 23 23 103,104 103,104 103,104 104 104 103,104 103 103,104 103 104 104

(*) Mixture of esters of the following acids: miristic, pentadecanoic, palmitoleic, palmidc, linoleic, oleic, and stearic acid (23). The structures of the new compounds 22.4, 23.2, 23.5, and 23.8-23.11 have been elucidated by spectroscopic methods; particularly the structures of chromenes 23.9 and 23.11 have

188 been established by NOE experiments and biosynthetic considerations (102). 6-Methoxy-2,2-dimethylchromene (23.5) and benzofuran 23.2 have also been synthesized by alkylation of 4-methoxyphenol to 2-(3-methyl-2-butenyl)-4-methoxyphenol (22.1), followed by acid catalysed cyclization of the corresponding epoxyde 22.3 to 23.3 and 23.6. Dehydration of 23.6 withp-TsOH gave 23.5, while NBS dehydrogenation of 23.3 afforded 23.2 (102). MeO^

MeO,

22.4

OMe

oMe

Scheme 15 - C-C and C-O phenol dimerizations It is worth noting that only one compound, the stearate of 4-methoxy-2-(3-methylbutenyl)phenol (22.2) could be isolated from young intact fruiting bodies of L.fuliginosus and L. picinus extracted in the cold (102). On the other hand, in injured mushrooms the stearate 22.2 was rapidly hydrolysed by lipases to free phenol 22.1. Therefore, the ester 22.2 is the biogenetic precursor not only of compounds 23.2 and 23.5, but also of 22.4 and 23.8-23.11, which can be considered dimerization products of 22.1. Oxidative dimerizations of phenolic compounds occur in Nature by one-electron transfer C-C and C-O couplings which are catalysed by phenol oxidase enzymes. Reactions of the same kind are probably responsible for the reddening of the flesh and milky juice of damaged mushrooms of the Section Plinthogali. In fact, the same change of colour was observed when synthetic phenol 22.1 was added to a mush of Lfuliginosus from which the original metabolites had been washed out with CH2CI2. Moreover, this experiment afforded the same mixture of chromenes and benzofuran as originally isolated from damaged fresh fruiting bodies. The structures of the red pigments are still unknown as they remain irreversibly adsorbed on the top of chromatographic columns.

n The oxidative dimerization of phenol 22.1 was simulated in vitro. Exposure of this compound to the complex Cu(N03)2-pyridine gave rise to dimers 23.8 and 23.9 by a C-C coupling reaction. Compound 23.9 could be cyclodehydrogenated to 23.8 by reaction with DDQ. On the other hand, exposure of phenol 22.1 to K3Fe(CN)6 gave the product 22.4 of a C-O coupling, which was then transformed into 23.10 by DDQ cyclodehydrogenation (102) (Scheme 15). In addition to dimers, a natural trimer 23.12 of phenol 22.1 has been isolated from L. fuliginosus and its structure has been elucidated by accurate and extensive NOEDS experiments (102). Furthermore, a qualitative evaluation of the contents of different extracts of the same or different Lactarius species of the Section Plinthogali has been carried out by GC and GC-MS analysis, using a Dexsil 300 column (125). By this method, the simple 2,2-dimethylchromene (23.4) has been identified in an extract of L. picinus (102). TABLE 23 - Benzofurans and Chromenes MeO

«X3<"^©0c 23.4-23.7

N°

Name

Substituents

23.1 23.2 23.3 23.4 23.5 23.6 23.7 23.8 23.9 23.10 23.11 23.12

No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name No trivial name

r-oxo;2(3)-en r-OH;r-Me;2(3)-en

r-OH;r-Me

8 23.8-23.12

3(4)-en 6-OMe;3(4)-en 3-OH;6-OMe 3-OAc;6-OMe 8-(2,2-diMe-6-OMe-2-H-chromen-8-yl) 8-[2-OH-3-(3-Me-2-butenyl).5-OMe-phenyl] 8-[2-(3-Me-2-butenyl)-4-OMe-phenoxy] 8-[2-OH-4-OMe-6-(3-Me-2-butenyl)phenoxy](*) 8- (2-[2-(3-Me-2-butenyl)-4-OMe-phenoxy]4-OMe-6-(3-Me-2-butenyl)phenoxy)

Ref. 102 102 102 102,105 102,105 102 102 102,105 102 102 102 102

(*) Alternatively, compound 23.11 may have the following substituent: 8-[2-OH-3-(3-Me-2-butenyl)5-OMe-phenoxy] (102).

Biological activities of Lactarius metabolites With such a great variety of compounds isolated from Lactarius species, some of them were expected to have a definite function in wild mushrooms and to exhibit other biological activities not strictly related to the evolutionary process. However, the investigations in this field have been fragmentary, and the biological activity of most isolated compounds has not been tested at all. Moreover, many compounds discussed in this section may possess still other, so far unknown biological properties.

190 A detailed study of structure-antifeedant activity relationship has been published by Polish researchers (59, 71, 73, 126-129). They have tested more than 30 compounds with lactarane, isolactarane, marasmane and 13-normarasmane skeletons, either isolated from species oiLactarius or their synthetically prepared derivatives. A test with wafer disks was used to examine the efficiency of these sesquiterpenes to inhibit feeding in adults of storage pests Sitophilus granarius L. and Triholium confuswn Duv. as well as in larvae of 7. confusum and Trogoderma granariwn Ev. (126, 129). Considering all the results reported in the literature, we could not find a clear-cut structureactivity correlation or a minimum requirement for obtaining antifeedant activity. Considering the average intensity of deterrence of lactarane sesquiterpenes, the highest activity was found for lactarorufin A (11.44) and 3-0-ethylfurandiol (18.19). However, even these sesquiterpenes can not be considered universally antifeedant for the storage pests, as both compounds showed at least one modest deterrence coefficient towards tested storage pests. Interestingly, compound 18.19 is not a metabolite of Lactarius mushrooms but is formed during the extraction of the fruiting bodies with ethanol (92). A significant trend of the structure-activity relationship can be found for the lactarorufins series. The inversion of configuration of the hydroxy group at C-8 decreased the activity of lactarorufin A (see compound 11.45) (71). Similarly, the saturation of C6-C7 double bond of compound 11.44 to give lactones 11.64 and 11.65, as well as acylation of the C-8 hydroxy group of 11.44 to yield esters 11.54 and 11.59, decreased the deterrent activity (71). The same decrease of activity was observed when the C-15 methyl group or the C-13 methylene group of lactarorufin A were oxidized, as in lactarorufin B (11.71) or in lactarolide A (11.70) (71), or when the tertiary C3 oxygenated group was etherified, as in the 3,8 internal ether 11.15 (59) or in the O-ethyl ether 11.52 (73). In the latter case, oxidation of the C-8 secondary hydroxy group restored the original deterrent activity (see compound 11.40) (73). Comparing the average antifeedant activity of the couples 11.40/16.14 (73), 11.15/16.10 (59), 11.52/16.16 (73), and 11.44/16.15 (71) one can observe a decrease of activity on moving the lactarane carbonyl group from C-5 to C-13. In the furanoid series of lactarane sesquiterpenes the weakest antifeedant activity was detected (73) for the free 3,8-diol 18.14, which is much less deterrent than the corresponding lactone 11.44. The specific activity of furandiol (18.14) increased on dehydration of the C-3 hydroxy group to yield furanol (18.9) (73) or on etherification to give the internal ether 18.5 (59) or the 3-0ethyl ether 18.19 (73). Some compounds showing a modest average deterrent activity are selective antifeedant against individual pest species. For instance, the diketohydroxyfuran 18.26 was a very strong antifeedant against Sitophylus granarius, and a good antifeedant against Triholium confusum and Trogoderma granarium larvae, but it became an attractant for Triholium confusum adults (71). Marasmane and lactarane dialdehyde sesquiterpenes are among the strongest known antifeedant compounds. Isovelleral (6.1) is an antifeedant against mammals (108), insects and fishes (113), while velleral (10.5) is an antifeedant against fishes (113). Chrysorrhedial (10.1), a very pungent dialdehyde recently isolated from L. chrysorrheus (46), is very likely responsible for the high unpalatable taste of this mushroom to the opossum, a natural fungivore (108). Interestingly, uvidin A (3.14), which has no aldehyde group, has the same deterrent activity, at a concentration of 100 ppm, as polygodial against the phytophagous insect Spodoptera littoralis (11). On the other hand, uvidin B

191 (3.36), the 3-p-hydroxy derivative of uvidin A, is much less active in the same test, pointing out that even small molecular changes may have significant effects on the biological activity. Because of the relatively large and potentially increasing use of wild mushrooms as food, the mutagenic effect of several common mushrooms, including a dozen of Lactarius species (L. necator Persoon ex Fr., L. torminosus Schff ex Fr., L. helvus Fr., L. vellereus Fr., L. resimus Fr., L. deliciosus Fr., L. deterrimus Groger, L. trivialis Fr., L. glyciosmus Fr., L. mitissimus Fr., L. rufus Scop, ex Fr., L. quietus Fr.) have been tested in the Ames Salmonella typhimurium tester strains (130-132). A few Lactarius species under investigation showed a weak but significant mutagenicity. L. necator had the highest mutagenic activity, which was not only maintained but even doubled when samples of crude extracts were placed in a boiling water bath to simulate the effects of food processing (131). In another study (133) Sterner found that the mutagenic activity of commercially preserved (pickled) L. necator was much higher than that observed for fresh samples. The toxic agent was later isolated in crystalline form by HPLC of an aqueous mushroom extract and its mutagenicity was found to be comparable with aflatoxin Bl (134). The toxin concentration in mushrooms was estimated about 2 mg/kg. Moreover, the toxin was found stable for at least one day at r. t. and at pH values between 1.0 and 11.0 (134). The strong mutagen, called necatorin (135), was first assigned a coumaro-cinnoline structure (101), but this assignment was later revised by Steglich, who found necatorin identical with a sample of necatorone (21.2) (99, 100). Isovelleral (6.1) and the artifact 9-hydroxyisovelleral (6.5) have been shown to be potent direct-acting mutagens in the Salmonella/micTosomt assay (132, 136). Therefore, isovelleral seems to be responsible, at least in part, for the mutagenicity of some pungent Lactarius species (for instance, L. vellereus and L. rufus). However, the dialdehyde functionality is not an essential requisite for conferring a significant mutagenic activity to any compound. In fact, velleral (10.5), piperdial (10.21) and lactardial (14.4), as well as other unsaturated sesquiterpenes aldehydes, are not mutagenic at all (27,113,132,136). Isovellerol (7.2) was also found to be mutagenic, although more than ten times less than isovelleral (6.1), and it requires metabolic activation by rat liver microsomal enzymes (27). As this in vitro metabolism easily performs oxidations, it was suggested that the mutagenicity of isovellerol is due to its oxidation back to isovelleral (27). The much lower activity of isoisovelleral (6.10) and 9-hydroxyisoisovelleral (6.11) than compounds 6.1 and 6.5, respectively, is also remarkable (136), as these differences indicate that the cyclopropane ring stereochemistry strongly affects the mutagenic activity. Stearoylvelutinal (7.14), vellerol (10.8) and vellerdiol (10.20) were found not to be mutagenic at all (27). Three guaiane sesquiterpenes, the ester (4.13), lactaroviolin (4.3), and deterrol (4.7) showed a weak mutagenic activity in the Ames Salmonella assay (137). In addition, deterrol (4.7) exhibited moderate cytotoxicity towards Ehrlich ascitic tumor cells (EGA cells) and weak toxicity towards L1210 cells (lymphocytic leukaemia mouse) (137). Lactaroviolin (4.3) showed a weak inhibitory effect towards EGA cells, while the ester 4.13 did not impair either cell line at 50 )J.g/ml (137). The Swedish authors pointed out that the greater toxicity of deterrol than of lactaroviolin was rather

192 unexpected, since an aldehyde functionality is normally more reactive and thereby more toxic than the corresponding alcohol functionality; they suggested that the azulene portion contributes significantly to the biological activities of compounds 4.3 and 4.7 (137). One should note that guaiane sesquiterpenes are typical metabolites of a few Lactarius species, as L. deliciosus Fr., L. deterrimus Groger and L. sanguifluus Paulet ex Fr., which are commonly appreciated as human food in many European countries. However, these toxic but thermolabile compounds are not considered harmful to human health, if mushrooms are thoroughly cooked before ingestion (137). Testing the antibacterial, antifungal, cytotoxic, phytotoxic, and algaecidal activities of twenty unsaturated sesquiterpene dialdehydes from plants and mushrooms. Sterner and Anke found (136) the highest activity for velleral (10.5), isovelleral (6.1) and isoisovelleral (6.10), which is a synthetic derivative of compound 6.1. Isovelleral (6.1), which is probably the most biologically active secondary metabolite ever isolated from Lactarius mushrooms, showed minimal inhibitory concentrations (MIC) in the range of 0.5-5 |ig/ml against Gram-positive and Gramnegative bacteria, and in the range of 0.1-5 |ig/ml against some fungi (136). Cytotoxic effects of isovelleral against ECA and L 1210 cells were in the range of 2 p.g/ml, while the MIC value against the alga Chlorella vulgaris was determined as 1 |ig/ml (136). Furthermore, isovelleral (6.1) showed 100 % hemolysis of bovine erythrocytes at a concentration of 100 |ig/ml and suppressed completely germination of Lepidium sativum and Setaria italica at a concentration of 20 and 10 |ig/ml, respectively (138). Velleral (10.5) is somewhat less active than isovelleral (6.1) (136), and comparable with that of velleral is the antimicrobial activity of piperdial (10.21) (51). By contrast, lactardial (14.4) is considerably less active (about 10-500 times less) than isovelleral (136). Sterner observed that modification of the structure of a natural sesquiterpene aldehyde to less polar compounds usually increased the antimicrobial and cytotoxic effects and reduced mutagenicity, while the introduction of a hydroxy group had the reverse effect (136). These results clearly show that the antimicrobial and mutagenic activities do not parallel each other and can be separated by derivatization. In addition, the reduction of one or both the aldehyde groups of dialdehydes 6.1, 10.5 and 10.21 to isovellerol (7.2), vellerol (10.8) and vellerdiol (10.20), and piperalol (10.26), respectively, was accompanied by a drastic decrease of the antibacterial and antifungal activities (27, 51). Moderate antimicrobial activity was also observed for the following compounds: deterrol (4.7) against Acinetobacter calcoaceticus (137); lactaroviolin (4.3) against tubercule bacilli (139); chrysorrhedial (10.1), chrysorrheal (10.7) (46), and rubrocinctal A (7.5) (23) against Bacillus subtilis and Staphylococcus aureus; phenol 22.1 against Candida albicans (102). The brine shrimp {Anemia salina) lethality assay (140) has been used by us routinely for the detection and isolation of mushroom bioactive compounds. Toxicity against this organism was observed for uvidin A (3.14) (LC50 48.8 ppm) (11), lactaroscrobiculide A (16.2) (LC50 12.9 ppm), chrysorrhedial (10.1) (LC50 15.7 ppm), chrysorrheal (10.7) (LC50 43.6 ppm) (46), and geranylgeranylhydroquinone (22.6) (LC50 0.47 ppm) (23). Moreover, 50 y/ml and 0.05 y/ml of uvidin A (3.14) inhibited 97% and 21% ^H-tymidine incorporation, respectively, into HL-60 human leukaemia cells (11). Phenol 22.6 was reported to be a potent contact allergen too (124). The related compound flavidulol A (22.8) exhibited a good antimicrobial activity in vitro against Staphylococcus

193 aureus. Bacillus subtilis, and other bacteria and fungi (103). In addition, flavidulols A (22.8), B (22.13) and C (22.15) appreciably suppressed proliferation of mouse lymphocytes stimulated with mitogens such as concanavalin A (Con A) and lipopolysaccharide (LPS) (104). The IC50 values of the last three compounds were calculated to be 8.9, 4.9 and 36.3 Hg/ml, respectively, in the former test, and 6.7, 3.9 and 28.3 |ig/ml, respectively, in the latter assay (104). Besides showing various biological activities, several compounds isolated from Lactarius species make unpalatable the flesh and latex of the mushrooms and are suspected to cause gastrointestinal upset in humans. Isovelleral (6.1), velleral (10.5), piperdial (10.21), epi-piperdial (10.22), chrysorrhedial (10.1), lactardial (14.4), and phenol 22.1 have a hot taste for the human tongue, while lactaroscrobiculide A (16.2) and chrysorrheal (10.7) are bitter and slightly astringent. Porninsal (1.1) and porninsol (1.2) have a disgusting taste. The unsaturated dialdehyde functionality, that is common for many biologically active sesquiterpenes isolated from Lactarius species, appears to be responsible for the hot taste and the biological activities of several biogenetically totally different terpenoids, produced by organisms that have little if any evolutionary relationship. In fact, these compounds have been isolated from such different organisms as fungi, higher plants, liverworts, molluscs, termites, and algae (113, 141). Several successful attempts have been reported to demonstrate a quantitative structure-activity relationship (QSARS) for important unsaturated dialdehydes of this kind, including velleral (10.5), isovelleral (6.1), isoisovelleral (6.10), and their corresponding 9-hydroxyderivatives 10.23, 6.5 and 6.11. The multivariate PLS-method gave a good correlation between chemical and molecular features of unsaturated dialdehydes and (a) the mutagenic activity in the Ames Salmonella test (142), (b) the membrane toxicity measured as the induction of ATP leakage in ELD cells, (c) the effect on cell membrane permeability in human neuroblastoma SH-SY5Y cells (143), and (d) the inhibitory effect on the dopamine Dl receptors (144). The most important structural descriptors were found to be the dipole moment, the LUMO-HOMO energy difference, the lipophilicity, the atomic charges of the unsaturated dialdehyde functionality, and geometrical properties such as the distance and the dihedral angle between the dialdehyde groups. Several observations indicate that the biological activities of unsaturated dialdehydes are caused by their covalent binding to nucleophilic structures present in living cells, for instance, amino groups of enzymes and receptors or RNA/DNA, or sulfydryl groups of proteins. In support of this hypothesis Cimino et al. observed (141) that selected natural and synthetic pungent 1,4-dialdehydes reacted in vitro with methylamine under "biomimetic" conditions (phosphate buffer, pH 9, MeCN), forming pyrrole derivatives. Under these conditions, both isovelleral (6.1) and isoisovelleral (6.10) reacted very slowly with MeNHa, giving a product formulated as P I for the reaction of the dialdehyde 6.1, and as P2 for the reaction of the stereoisomer 6.10 (Scheme 16). However, large quantitative and qualitative activity differences between the unsaturated dialdehydes have been reported, and other molecular mechanisms have been suggested besides the formation of pyrroles (145).

194

Scheme 16

Occurrence of metabolites in Lactarius significance

species and their biological

In spite of the large number of metabolites isolated from Lactarius species, their significance for the mushrooms has generally been understood only recently. Some of these compounds have been reported for various species of Lactarius, as well as for other genera, while others seem to be specific to one or few species. Moreover, a few metabolites seem to be irregularly present in the same species, even when apparently identical isolation procedures have been used. For instance, isovelleral (6.1) was found in Lactarius torminosus by American (109) but not by Swedish authors (51); isovelleral (6.1) and velleral (10.5) were reported to be isolated from L. piperatus {L. pergamenus) (38) and L, rufus (38), respectively, but in later investigations these mushrooms were not found to produce even traces of these sesquiterpenes (36, 51); neither the isolation of vellerolactone (11.2) and pyrovellerolactone (11.3) from L. vellereus (57) nor that of blennin A (11.17) from L. torminosus (110) could be repeated in later investigations (27, 51). An explanation of these irregularities may be the existence of easily confused sub-species, especially when the mushrooms grow in different habitats. Large qualitative and quantitative differences in the pattern of isolated metabolites were observed especially when different extraction and isolation procedures were used. We have already underlined the possible formation of artifacts during the extraction of fruiting bodies with alcoholic solvents or during the handling and storage of the extracts. For example, the alkoxy derivatives 11.72,16.21,18.17, and 18.19 are surely artifacts formed during the extraction with alcohols. Moreover, freezing-thawing the mushrooms prior to extraction as well as the collection of the specimens and/or the transport from their habitats to the laboratory were found to affect the enzymatic processes and to trigger enzymatic conversions of the original metabolites (113). For these reasons early taxonomic conclusions (63, 106, 107, 110, 112, 146) drawn on the basis of the identified metabolites, must be considered now with caution, especially when results obtained in different laboratories have been compared.

195 After several investigations, it is now firmly established that any physical injury to the fruiting bodies of Lactarius species gives rise to a more or less rapid dramatic change of the contents in secondary metabolites. This may be perceived also by human senses as a change, for example, of the taste and colour of the milky juice. To extract the compound(s) originally present in intact mushrooms, young and apparently parasites unaffected fruiting bodies were frozen, immediately after collection, with dry ice (14) or by immersion into liquid nitrogen (16, 38), and extracted with a suitable solvent such as diethyl ether, hexane or ethyl acetate. Alternatively, the mushrooms were frozen at -20° C and extracted with CH2CI2 (40,102). In addition, immediately after an injury to the fruiting bodies, a few drops of the milky juice were quickly collected with a capillary tube, suspended in CH2CI2, and analyzed (8, 46, 102). Following this procedure carefully during the investigations on intact mushrooms, several authors (8, 11, 14, 16, 21, 22, 23, 27, 38, 40, 42, 46, 49, 51, 52, 102, 113) isolated only fatty acid esters of very few compounds, or even of a single compound, depending on the species. The most common acids of these esters are the C-16 and C-18 saturated and unsaturated fatty acids usually occurring in many living organisms, like palmitic, stearic, oleic and linoleic acids. Other homologues have been found only occasionally, while esters of the new acid 6-oxostearic acid (lactarinic acid) have often been isolated alone or in a mixture with other esters. These compounds are contained in an emulsified form in the lacticiferous hyphae of the intact mushrooms, where they seem to be stable. This was shown, for instance, by the investigation of old but still intact specimens of L. vellereus (113) and L. fuliginosus (102). In fact, upon extraction, old specimens yielded the same metabolites as young specimens, although considerable quantitative differences were observed. Stearoylvelutinal (7.14) and 6-ketostearoylvelutinal (7.15) are by far the most widespread of all sesquiterpenoid esters found in Lactarius species, as shown in Table 24. They occur either alone, as 7.14 in L. vellereus (38, 40) and 7.15 in L. mitissimus (23), or mixed as, for instance, in L. circellatus and L. necator (52). Other two velutinal esters have been detected, but they are still unidentified (40). Interestingly, the mild and colourless velutinal esters 7.14 and 7.15 are typical metabolites not only of species with permanently white latex, but also of mushrooms, as L. scrobiculatus and L. chrysorrheuSy in which the colour of the latex changes from white to yellow (46). Fatty acid esters of other marasmane sesquiterpenoids, as the two rubrocinctal derivatives 7.6 and 7.8, have been isolated so far only from L. ruhrocinctus (23). Of limited distribution are also the farnesane fatty acid esters 1.3-1.9 (8), drimenol (3.2-3.6) and uvidin A (3.16-3.20) esters (11), violascensol lactarinate 5.2 (23), and the two glutinopallane esters 8.2 and 8.3 (42). Further unprecedented sesquiterpenoids are the guaiane esters 4.10-4.13 contained in intact specimens of Lactarius species of the Section Dapetes (14, 16, 21, 22). Various mixtures of yellow esters 4.12 and 4.13, red stearoylsangol (4.11), and blue stearoyldeterrol (4.10) are responsible for the beautiful colours of the latex of these species. Fatty acid esters of metabolites different from sesquiterpenes have been isolated from intact fruiting bodies of particular Lactarius species: the colourless phenol stearate 22.2 is a typical metabolite of L. fuliginosus and L. picinus (102), while L. lignyotus contains a mixture (22.7) of fatty acid esters of geranylgeranylhydroquinone (23).

196 To investigate the chemical transformations occurring in disrupted tissues, several specimens are usually minced in absence of any organic solvent, and portions of the mush are extracted with a suitable solvent (usually hexane or CH2CI2) at increasing times, and analyzed for their contents. Several observations indicate that when the lacticifers get broken the original fatty acid esters can come into contact with enzymes that convert the esters in different times (seconds to hours) into other compounds with free hydroxy or aldehyde groups. These transformations are considered to be truly enzymatic, as they have never been observed in vitro, and the kind of compounds formed depends on the species and the metabolite(s) originally present in the fruiting bodies. The first observations made on the biochemical processes occurring in L. vellereus (27) are for many aspects a landmark in the long history of Lactarius chemistry. Wickberg and coworkers found that in injured L. vellereus the tasteless and inactive stearoylvelutinal (7.14) was converted in a few seconds into the two pungent and potent antimicrobial and antifeedant (see Section on Biological activity) sesquiterpene dialdehydes isovelleral (6.1) and velleral (10.5). These compounds were then gradually (minutes to hours) reduced by the mushroom enzymes to the much less toxic and non-pungent compounds isovellerol (7.2) and vellerol (10.8), and eventually to vellerdiol (10.20) (27). The sesquiterpenes 6.1, 7.2, 10.5, and 10.8 were typically found to constitute 80-90% of the extracts made 5 min. or later after grinding the fruiting bodies at 22° C, while only small amounts of stearoylvelutinal (7.14) were detected by TLC. Free velutinal (7.11) was not originally present in the mushrooms, although small amounts could be observed by TLC shortly after grinding at 4° C (27). The fact that dialdehydes 6.1 and 10.5 are highly toxic, and that they are not formed until the mushroom has been injured, have both been interpreted as if these sesquiterpenes are the active principles of a chemical defence system that protects injured fruiting bodies from parasites and predators (27, 108). The subsequent enzymatic reductions of velleral and isovelleral to the considerable less toxic compounds 7.2, 10.8 and 10.20 have been suggested to have the effect of saving the mushroom from unnecessarily prolonged contact with its own defence chemicals, because the two dialdehydes 6.1 and 10.5 may be also toxic to the mushrooms itself, as indicated by their antifungal activity (27). The fate of velutinal esters 7.14 and 7.15 in other pungent Lactarius species is similar to that occurring in L. vellereus; however, besides velleral (10.5) and isovelleral (6.1), other four hot taste unsaturated dialdehydes are produced by different mushrooms as a response to a physical injury. Each species shows a fingerprint pattern of pungent dialdehydes comprising different amounts of velleral (10.5, isovelleral (6.1), piperdial (10.21), ept-piperdial (10.22), lactardial (14.4), and chrysorrhedial (lO.l). Kinetic studies as accurate as those on the enzymatic reactions occurring in L. vellereus (27) have not been carried out for other species; however, several observations suggest that conversions of velutinal esters to dialdehydes are always fast, taking a few seconds in L. scrobiculatus (46) up to a few minutes in L. quietus (27). Interestingly, this time corresponds very well to that required by a sample of the mushroom for developing a sharp taste in the mouth. The efficiency of the conversions of sesquiterpenes depends on the species. In ground fruiting bodies of L. chrysorrheus (46), L. torminosus, and L. piperatus (51) the esters of velutinals disappear completely after a few minutes; by contrast, the main part of the esters remains unchanged in the

197 ground fruiting bodies of L. circellatus and L. necator (52), even after several hours. However, given the high specific biological activity, even the limited amounts of free dialdehydes formed in mushroom tissue are considered sufficient to ensure an adequate deterrence against predators. As velleral and isovelleral in L. vellereus, the other pungent dialdehydes appear to be formed in mushrooms almost simultaneously with the corresponding hydroxy aldehydes,. In addition to the aldehydes, small amounts of furanosesquiterpenes of enzymatic origin have been isolated from the following Lactarius species: furanodiene (18.1) from L. scrohiculatus (46); furanol (18.9) from L. chrysorrheus (46), L. scrohiculatus (49), L. circellatus, and L. necator (52); lactarol (19.3) from L. piperatus and L. torminosus (51); furandiol (18.14) from L. scrohiculatus (49, 86, 88), L. piperatuSyL. torminosus, L. necator and L. circellatus (51, 52, 94); the hydroxyfuran 18.16 from the latter four species (51, 52, 94). These results are particularly interesting since the furans have long been considered merely artifacts formed by the degradation of velutinal esters during extraction and work-up (27, 85). Actually, this is certainly true for most findings of the furanosesquiterpenes reported in Tables 18 and 19,. By means of the standard procedure described above, several sesquiterpene lactones have also been isolated from extracts of injured fruiting bodies and are considered true mushroom metabolites. For instance, blennin C (14.5), lactaroscrobiculide A (16.2), chrysorrhelactone (16.1) and 8,9seco-5-lactaranolide 14.3 all appear in the fruiting bodies of L. scrohiculatus (46, 49) and L. chrysorrheus (46) within the first minutes after an injury. The rapid formation of the light sensitive chrysorrhelactone (16.1) is responsible for the dramatic yellowing of the latex and flesh of these two mushrooms. Interestingly, in fruiting bodies of L. scrohiculatus cut in the dark and under Na, yellowing occurred as promptly as in the air, demonstrating that atmospheric free oxygen is not directly involved in the colour change (46). Large quantities of sesquiterpene lactones have especially been isolated when minced fruiting bodies of different Lactarius species were kept in an EtOH solution at r. t. for several hours, or even for a few months (64). As EtOH causes the formation of artifacts and furandiol (18.14) was autoxidized to lactarorufin A (11.44) when an EtOH solution was exposed to air in light for a few days (54), it would be desirable to repeat the extractions of the mushrooms in other solvents, following the standard procedure described above. Moreover, in order to better understand the biological significance of sesquiterpene lactones, it would be interesting to determine the kinetics of their formation and disappearance in injured mushroom tissues. However, the formation of most of these lactones is believed to be assisted by specific mushroom enzymes, as they have never been obtained as degradation products of velutinal ester (85, 118), and different patterns of lactarane and secolactarane lactones have been found in Lactarius species containing only velutinal esters 7.14 and 7.15 as the original sesquiterpenes (63, 106, 107). Moreover, it has been suggested that other factors, besides injury, regulate the formation of the lactones, and that they can be derived biogenetically from oxidation, not only of the furans (54), but also of lactarane dialdehydes or hydroxy aldehydes (52, 113). Compared with the pungent Lactarius species containing velutinal derivatives, the enzymatic conversions taking place in injured specimens of other species show both similarities and differences. In fact, the original esters are first quickly hydrolysed by lipases to give free alcohols or phenols.

198 TABLE 24 - Occurrence of secondary metabolites in Lactarius species Sub-genus (Eu) Lactarius Section ALBATI (Bat) Sing. L. controversus Pers. ex Fr. 7.14 (23), 11.9 (106), 11.10 (106). 11.11 (106), 11.44 (107) L.deceptivus Peck

6.1 (108, 109)

L. glaucescens Crossland ss. Blum 7.14 (40) L.pergamenus^ L.pergamenus^

6.1 (25), 10.5 (48), 11.2 (57), 11.3 (57), 18.9 (92), 18.17 (92), 19.1 (96) 11.9 (106), 11.10 (106), 11.11 (106)

L. piperatus (L. ex Fr.) S. F. Gray 6.1 (36. 38, 51), 7.14 (51), 10.5 (36, 38, 51). 10.8 (51). 10.21 (36, 51), 10.26 (51), 14.4 (51), 18.14 (51), 18.16 (94), 19.3 (51) L.subvellereusVeck

6.1 (109), 10.5 (109)

L. vellereus (Fr.) Fr.

6.1 (24, 25, 27, 38), 6.5 (27). 6.7 (30). 6.12-6.13 (32, 41), 7.1 (33), 7.2 (27). 7.11 (27), 7.14 (38. 40), 7.16 (32, 41). 7.18 (33), 7.19 (41). 7.22 (27). 9.5 (32, 41). 10.5 (24. 27. 36. 38, 48), 10.8 (27), 10.20 (27), 10.23 (27), 10.40 (27), 11.1 (56, 87), 11.2 (27, 57. 87), 11.3 (27, 57), 11.9 (106), 11.10 (106), 11.11 (106), 11.20 (106), 11.44 (32, 41, 107), 14.5 (106), 16.6 (87), 18.9 (56, 92), 18.10 (93), 18.26 (56, 95), 18.14 (32. 41. 82, 107), 18.17 (92), 18.19 (56, 73, 92), 19.1 (96), 19.3 (87)

L. velutinus Bert.

7.14 (39, 40)

Section TRICHOLOMOIDEI Fr. L pubescens Fr.

7.14 (23)

L. scrobiculatus (Scop, ex Fr.) Fr. 7.14 (40, 46, 49), 10.1 (46). 10.7 (46, 49), 11.5 (70). 11.9 (70). 11.11 (70), 11.33 (70). 11.38 (70). 11.70 (54). 14.3 (49). 14.4 (46. 49). 14.5 (46. 49. 86). 15.1 (70). 16.1 (46). 16.2 (46. 49. 86). 16.7 (88), 16.9 (46, 89), 16.20 (54). 18.1 (46), 18.5 (88), 18.6 (88), 18.7 (88), 18.9 (49), 18.12 (88), 18.13 (88), 18.14 (49, 86, 88). 18.15 (70). 18.18 (88). 19.1 (88) L. torminosus (Schff. ex Fr.) Fr. 6.1 (51. 109). 7.14 (51). 10.5 (38. 51, 109), 10.8 (51), 10.21 (51). 10.26 (51), 11.9 (63, 110, 111), 11.10 (63), 11.11 (63), 11.17 (63, 81.110, 111). 11.18 (63). 11.44 (107. 110, 111), 11.47 (107), 11.51 (80, 81. 110, HI), 14.4 (51), 14.5 (63, 110, HI), 18.14 (51, 107, 110, 111). 18.16(94), 19.3 (51) Section ZONARII Quel. em. K. R. L. porninsis Roll

1.1-1.9 (8)

Section DAPETES Fr. L. deliciosus (L. ex Fr.) S. F. Gray 4.1 (13. 14). 4.2 (14-16). 4.3 (14. 16. 17). 4.5 (16). 4.7 (16). 4.9 (14. 16). 4.12 (16), 4.13 (14, 16). 4.14 (16) L.deterrimus GrOger

4.2 (16. 112). 4.3 (16). 4.5 (16). 4.6 (20). 4.7 (16). 4.9 (16). 4.12 (16), 4.13 (16). 4.14 (16)

199 L. indigo Schw. ex Fr.

4.3 (22), 4.10 (22), 4.14 (22)

L. salmonicolor Heim-Leclair 4.9(112). 4.13 (112) L. sanguifluus (Paulet ex Fr.) Fr. 4.3 (112), 4.4 (19), 4.6 (19), 4.8 (21), 4.9 (112), 4.11 (21), 4.13 (21.112) L. semisanguifluus Heim-Leclair 4.2 (112), 4.3 (112), 4.9 (112), 4.13 (112) Section UVIDI (Konr.) Bon Lflavidus Bond. 3.1 (23), 3.14 (23) L. uvidus (Batsch. ex Fr.) Fr. 3.1 (10), 3.2-3.6 (11), 3.14 (10), 3.16-3.20 (11), 3.22 (12), 3.29 (12), 3.34 (12), 3.36 (10) L. violascens (Otto) Fr.

5.1-5.2 (23)

Section GLUTINOSI Qu. L.blennius{Vx.)VT.

7.14 (40). 11.9 (106), 11.10 (106), 11.11 (106), 11.17 (67, 68, 106), 11.20 (106). 11.41 (67), 11.43 (68, 74), 11.44 (67, 107), 11.70 (54), 11.72 (54), 14.5 (67, 106), 16.20 (54), 16.21 (54), 18.9 (67), 18.14 (67, 107)

L.circellatusVx.

7.14 (52), 7.15 (52), 10.5 (52), 10.8 (52), 10.22 (52), 10.27 (52), 11.18 (52), 14.4 (52), 14.5 (52), 18.9 (52), 18.14 (52), 18.16 (52)

L. flexuosus FT.

6.1(113)

L. glutinopallens Lmgc 8.2-8.3 (42) L. necator (Bull. em. Pers. ex Fr.) Karst. 7.14 (38. 40. 52). 7.15 (38, 52), 9.5 (55), 10.5 (36, 52), 10.8 (52), 10.22 (36, 52). 10.27 (52), 10.30 (55), 11.9 (61. 63, 106). 11.10 (63. 106), 11.11 (63, 106), 11.15 (59), 11.18 (52, 62, 63. 106), 11.20 (63, 69, 106), 11.44 (55, 78, 107), 11.47 (78, 107), 11.49 (78, 107), 11.50 (78, 107), 14.4 (52). 14.5 (52. 61, 63, 106), 18.5 (59), 18.9 (52). 18.14 (52. 65. 82. 94). 18.16 (52. 94), 19.3 (113). 21.2 (99. 100. 101). 21.5 (98). 21.6 (98) L. pallidas Persoon ex Fr. 6.6 (29). 11.5 (29). 11.9 (29). 11.44 (29). 11.50 (29). 11.70 (54), 14.5 (29), 16.20 (54). 16.21 (54). 18.14 (29). 18.19 (29), 19.1 (29) L, trivialis (Fr.) Fr.

7.14 (113). 7.15 (113). 11.44 (110). 14.5 (110). 18.14 (110)

L. turpis^* c

11,9 (106), 11.11 (106). 11.20 (106), 14.5 (106)

L.vietus^

11.9 (106), 11.10 (106), 11.11 (106), 11.18 (106), 11.20 (106), 14.5 (106)

Section COLORATI Bataille L./MJCM5Rolland

7.14 (23)

L. glyciosmus^

11.9 (63), 11.10 (63), 11.17 (63), 11.18 (63), 11.20 (63), 14.5 (63)

L. helvus Fr.

11.9 (63), 11.11 (63), 18.9 (92), 18.17 (92)

L. rufus (Scop, ex Fr.) Fr.

200 6.1 (38). 7.14 (40, 113), 7.15 (113), 9.5 (44, 64, 107), 10.5 (36, 38, 52), 11.44 (53, 64, 107), 11.50 (107), 11.71 (64, 66, 82), 18.14 (82, 107), 18.19 (73) L. spinosolus^

11.9 (106). 11.10 (106). 11.11 (106), 11.20 (106), 11.44 (107), 18.14 (107)

Section RUSSULARES Fr. L. chrysorrheus FT.

7.15 (46). 10.1 (46). 14.4 (46). 14.5 (46), 16.1 (46). 16.2 (46). 18.9 (46)

L. decipiens Quel

7.15 (23)

L. fulvissimus Romdign. 7.15(23) L. mitissimusFr. L. quietus FT. L. subdulcis^

7.15 (23). 11.44 (83. 84). 11.70 (83). 11.71 (83. 84). 11.83 (83). 18.14 (83. 84). 18.27 (83. 84) 6.1 (27). 7.2 (27), 7.14 (23), 11.9 (63), 11.10 (63), 11.11 (63). 11.44 (107), 18.14 (107), 18.19 (73) 11.9 (63), 11.20 (63), 11.44 (107), 14.5 (63), 18.14 (107) Sub-genus Rhysocybe Nhf. ex B.

Section PLINTHOGALI (Bull.) Sing. L. fuUginosus FT.

22.1 (102), 22.2 (102), 22.4 (102), 23.5 (102), 23.8 (102), 23.9 (102), 23.10 (102), 23.11 (102), 23.12 (102)

L. lignyotus Fr.

22.6 (23), 22.7 (23)

L.picinusFT.

IIA (102), 22.2 (102), 23.2 (102), 23.4 (102), 23.5 (102), 23.8 (102), 23.9 (102), 23.10 (102)

Section ICHORATI (Nhf.) Bon L. rubrocinctus Fr.

7.5 (23), 7.6 (23), 7.7 (23), 7.8 (23)

Section OLENTES Bat. L. camphoratus (Bull, ex Fr.) Fr. 2.1 (9) Section TABIDI Fr. L. hepaticus (PI.) Boud. 18.7 (114) L thejogalus^

11.9 (63), 11.11 (63), 11.17 (63), 11.18 (63), 14.5 (63)

Not included in the above classification L.atroviridis{Fcck)

21.2 (98), 21.4 (98), 21.5 (98), 21.6 (98)

L.flavidulus Imai

22.8 (103, 104), 22.13 (103. 104). 22.15 (103, 104)

L. tomentoso-marginatus Hesler-Smith 6.1 (109) ^) Investigations have probably been carried out on specimens of L. piperatus (L. ex Fr.) S. F. Gray, reported as L. pergamenus in earlier papers (38,113). t>) In the original paper the taxonomic authority has been omitted (106). c) In the original paper this species has been considered different from L. necator (106), whereas in modem mycology (2) L. turpis is synonymous of L. necator.

201 which are then converted to aldehydes or other products in different times. Fresh extracts of injured species of the Section Dapetes gave deterrol (4.7), sangol (4.8) and dihydrodeterrol (4.9), and the corresponding aldehydes, lactaroviolin (4.3), compound 4.4 and delicial (4.5) (16, 21). The green colour that the latex of L. deliciosus and L, deterrimus assumes in time, is due to the formation of violet (lactaroviolin) and blue (deterrol) compounds and their mixing with yellow compounds (the alcohol 4.9, the esters 4.12 and 4.13, and delicial) originally present or formed afterwards (16). In injured specimens of L. fuliginosus and L. picinus the strongly acrid free phenol 22.1, rapidly released from the tasteless stearic acid ester 22.2, is gradually oxidized to a mixture of benzofuran 23.2, chromenes 23.5, 23.8-23.12, and polymeric red pigments (102). The latter compounds, which are responsible for the flesh and milky juice reddening of L. fuliginosus and L. picinus, are formed by C-C and C-0 phenol couplings. Similar phenol oxidase catalysed reactions are probably involved in the dimerization of flavidulol A (22.8) to flavidulol C (22.15) in L. flavidulus (103, 104), and of necatorone (21.2) to 4,4'-binecatorone (21.6) in L. necator (98). None of the compounds formed from precursors different from velutinal esters displays the wide spectrum of strong biological activities as, for instance, isovelleral (6.1) and velleral (10.5). However, a few of these metabolites possess an interesting antimicrobial and/or antifeedant activity and the similarities of the biochemical transformations occurring in the two groups of mushrooms can not be merely fortuitous. Therefore, it is likely that also the Lactarius species not producing pungent dialdehydes are somewhat protected from predators and parasites by the secondary metabolites formed in injured fruiting bodies. In conclusions, it appears that evolution has introduced several biochemical differences in different Lactarius species, such as : - the kind and amounts of compound(s) stored as fatty acid ester(s) in intact mushrooms; - the nature of the fatty acids esterified in these esters; - the rate and kinds of the enzymatic reactions occurring in injured mushroom tissues; - the kinds and amounts of secondary metabolites formed in injured fruiting bodies by conversions of the original fatty acid ester(s); - the biological activities of the so formed compounds. All such characteristics may be useful and, along with the morphological features, may facilitate taxonomic correlation of the Lactarius species. The mechanisms of the enzymatic conversions taking place in injured Lactarius fruiting bodies are still poorly understood. Of course, the co-isolation of metabolites with related structures from the same mushroom has suggested possible biosynthetic routes (35, 36, 46, 63, 70, 74, 78, 90, 102, 113, 118). The conversions of velutinal esters 7.14 and 7.15 to lactarane and secolactarane sesquiterpenes appear particularly interesting, as they involve a rearrangement of the marasmane skeleton with concomitant creation of new stereocenters. Therefore, stereochemical details of isolated sesquiterpenes have indicated possible biosynthetic intermediates. For example, the findings of compounds with opposite configuration at C-3 (compare velleral (10.5) with chrysorrhedial (10.1)) have suggested the existence of separate biosynthetic routesft-omthe velutinal esters to the different compounds (35,46). It seems that at least four divergent routes must exist (46): one pathway would

202

lead to isovelleral (6.1), another to the velleral (10.5)-piperclial (10.21)-ep/-piperdial (10.22) group of sesquiterpenes, another pathway would lead to chrysorrhedial (10.1) and related compounds, and the fourth one to secolactarane and furanoid sesquiterpenes. The mechanism of the latter enzymatic conversions is very likely similar to the non enzymatic degradation of velutinal esters in vitro (Scheme 3) (85, 118). Recently, some of these hypotheses have been confirmed by experiments with labelled compounds. Sterner has investigated the enzymatic conversions of stearoylvelutinal (7,14) in L. vellereus by feeding, in separated experiments, ^^O or [I2-2H3]-labelled isovelleral (6.3, 6.2) or velleral (10.6) to a freshly prepared mush of the fruiting bodies (27, 36). By determining the label content of the different sesquiterpenes formed. Sterner found that isovelleral and velleral are the precursors of isovellerol (7.2) and vellerol (10.8), respectively, and that isovelleral is not a precursor of velleral and the lactaranes. Several observations indicate that the enzymatic conversions of velutinal ester to other marasmane and lactarane sesquiterpenes start with the opening of the epoxyde ring, which is almost simultaneous with the enzymatic hydrolysis of the ester group (36, 46). In fact, free velutinal (7.11) was identified in the first hexane extract of L. vellereus obtained at 4° C (27). Moreover, when [12-2H3]-Iabelled velutinal (7.12) (35) was fed to injured fruiting bodies of L. vellereus, analysis of the enzymatic reactions products by ^H-NMR spectroscopy showed that free velutinal was a good substrate for the conversions and both the isovellerol (7.2) and vellerol (10.8) fractions contained a C-12 deuteriated compound (35). In another enlightening experiment, when methylvelutinal (7.13), which itself has never been isolated as a natural product, was fed to L. vellereusy the ene-acetal 7.9 could be identified in an EtOAc extract of the mush made 15 min. later, suggesting a possible mechanism for the enzymatic opening of the epoxyde (36). In conclusion, while the first steps of the velutinal ester conversions are rather certain, the details of further enzymatic transformations are still hypothetical, and their investigation will require more sophisticated experiments, possibly with isolated enzymes and labelled advanced intermediates. Acknowledgements - We acknowledge the continuous financial support of our research project on mushroom metabolites by the Italian MURST (Funds 40%) and CNR (Progetto Finalizzato Chimica Fine II). Most of the credit for our scientific contributions must be given to numerous coworkers whose names appear in the references. We wish also to remind our long friendship with W. M. Daniewski, B. Wickberg, O. Sterner, and the late M. Koc5r, who all have shared our passion for the Lactarius chemistry, and have collaborated with us for many years. Note added in proof Flavidulol D was isolated from L.flavidulus, and its structure was identified as the stearate of fiavidulol A (22.8) on the basis of spectroscopic data (147). The new marasmane lactone P3 and diketopiperazine P4 were isolated from the most polar fraction of an ethanolic extract of L. vellereus (148). The structures of compounds P3 and P4 were established by extensive NMR studies together with acetylation reactions. The isolation of lactone P3 suggested that the possible oxidation of the 15methyl group, typical of a few lactarane sesquiterpenes, could take place at the early velutinal stage

203 before its transformation to the final sesquiterpenes (148).

HO i H OR P3 R = H P3' R = Ac

References

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

R. H. Whittaker, Science, 163, 150 (1969). M. Bon, Documents Mycologiques, Tome X, Fascicule n. 40, Groupe de Micologie Fondamentale et appliquee, Lille (1980). M. Mordaski, Archiwum Immunologii i Terapii Doswiadczalnej, 4, 335 (1956). K. Hostettmann (Ed.), Assays for Bioactivity, Vol. 6 in the Series Methods in Plant Biochemistry, Academic Press London (1991). W. A. Ayer and L. M. Browne, Tetrahedron, 37, 2199 (1981). W.Tumer, Fungal Metabolites, Academic Press, London 1971. W.Turner and D. Aldridge, Fungal Metabolites II, Academic Press, London 1983. A. Gamba-Invernizzi, L. Garlaschelli, A. Rossi, G. Vidari and P. Vita-Finzi, J. Nat. Prod., 56, 1948 (1993). W. M. Daniewski, P. A. Grieco, J. C. Huffman, A. Rymkiewicz and A. Wawrzun, Phytochemistry, 20, 2733 (1981). M. De Bernardi, G. Mellerio, G. Vidari, P. Vita-Finzi and G. Fronza, / . Chem. Soc. Perkin I, 221 (1980). L. Garlaschelli, G. Mellerio, G. Vidari and P. Vita-Finzi, /. Nat. Prod, in press. M. De Bernardi, G. Mellerio, G. Vidari, P. Vita-Finzi and G. Fronza, /. Chem. Soc. Perkin I, 2739 (1983). D. J. Bertelli and J. H. Crabtree, Tetrahedron, 24, 2079 (1968). K. Vokac, Z. Samek, V. Herout and F. Sorm, Coll. Czech, Chem. Commun., 35,1296 (1970). V. Benesova, V. Herout and F. Sorm, Coll. Czech, Chem. Commun., 19, 1351 (1954). O. Bergendorff and O. Sterner, Phytochemistry, 27, 97 (1988). E. Heilbronner and R. W. Schmid, Helv. Chim. Acta, 37, 2018 (1954). F. Sorm, V. Benesova, J. Krupicka, V. Sneberk, L. Dolejs, V. Herout and J. Sicher, Coll. Czech, Chem. Commun., 20, 227 (1955) and references cited therein. S. De Rosa and S. De Stefano, Phytochemistry, 26, 2007 (1987). S. K. Koul, S. C. Taneja, S. P. Ibraham, K. L. Dhar, and C. K. Atal, Phytochemistry, 24, 181 (1985). O. Sterner, O. Bergendorff and F. Bocchio, Phytochemistry, 28, 2501 (1989). A. D. Harmon, K. H. Weisgraber and U. Weiss, Experientia, 36, 54 (1980). G. Vidari and P. Vita-Finzi, unpublished results. P. H. List and H. Hackenberg, Arch. Pharmaz., 302, 125 (1969). G. Magnusson, S. Thoren and B. Wickberg, Tetrahedron Lett., 1105 (1972). T. Hansson, O. Sterner, B. Wickberg and R. Bergman, J. Org. Chem., 57, 3822 (1992). O. Sterner, R. Bergman, J. Kihlberg and B. Wickberg, J. Nat. Prod., 48, 279 (1985). O. Sterner, R. E. Carter and L. Nilsson, Mutation Res., 188, 169 (1987). M. De Bernardi, G. Fronza, G. Mellerio, V. Valla, G. Vidari and P. Vita-Finzi, Gazz. Chim. It., 114, 163 (1984). W. M. Daniewski, M. Gumulka, K. Ptaszynska, P. Skibicki, G. Fronza and G. Vidari, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 37, 283 (1989). R. Bergman, T. Hansson, O. Sterner and B. Wickberg, /. Chem. Soc, Chem. Commun., 865

204

32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

(1990). W. M. Daniewski, W. Kroszczynski, P. Skibicki, M. De Bernardi, G. Fronza, G. Vidari and P. Vita-Finzi, Phytochemistry, 27, 187 (1988). W. M. Daniewski, M. Gumulka, K. Ptaszynska, P. Skibicki, J. Krajewski and P. Gluzinski, Phytochemistry, 31, 913 (1992). S. Nozoe, R. Kobayashi, S. Urano and J. Furukava, Tetrahedron Lett,, 1381 (1977). T. Hansson, Z. Pang and O. Sterner, Acta Chem. Scand., 47, 403 (1993). T. Hansson and O. Sterner, Tetrahedron Lett., 32, 2541 (1991). O. Sterner, R. Bergman and B. Wickberg, Finn. Chem. Lett., 116 (1985). O. Sterner, R. Bergman, E. Kesler, L. Nilsson, J. Oluwadiya and B. Wickberg, Tetrahedron Lett.,24, 1415 (1983). J. Favre-Bonvin, K. Gluchoff-Fiasson and J. Bernillon, Tetrahedron Lett., 23, 1907 (1982). K. Gluchoff-Fiasson and R. Kuhner, C. R. Acad. Sci. Ser. Ill, 294, 1067 (1982). W. M. Daniewski, M. Gumulka, P. Skibicki, U. Jacobsson and T. Norin, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 35, 251 (1987). J. Favre-Bonvin and K. Gluchoff-Fiasson, Phytochemistry, 27, 286 (1988). W. M. Daniewski, M. Kocor and S. Thor^n, Heterocycles, 5, 77 (1976). W. M. Daniewski, M. Kocor and S. Thoren, Polish J. Chem., 52, 561 (1978). A. Konitz, M. Bogucka-Ledochowska, Z. Dauter, A. Hempel and E. Borowski, Tetrahedron Lett., 3401 (1977). M. De Bernardi, L. Garlaschelli, L. Toma, G. Vidari and P. Vita-Finzi, Tetrahedron, 49, 1489 (1993). J. Froborg and G. Magnusson, J. Am, Chem. Soc, 100, 6728 (1978). G. Magnusson, S. Thoren and T. Drakenberg, Tetrahedron, 29, 1621 (1973). Z. Pang, F. Bocchio and O. Sterner, Tetrahedron Lett., 33, 6863 (1992). G. Magnusson and S. Thoren, Acta Chem. Scand., 27, 2396 (1973). O. Sterner, R. Bergman, C. Franzen and B. Wickberg, Tetrahedron Lett., 26, 3163 (1985). O. Sterner, Acta Chem. Scand., 43, 694 (1989). W. M. Daniewski and M. Kocor, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 19, 553 (1971). M. De Bernardi, G. Fronza, G. Mellerio, G. Vidari and P. Vita-Finzi, Phytochemistry, 18, 293 (1979). W. M. Daniewski, M. Gumulka, K. Ptaszynska, G. Vidari, L. Garlaschelli, G. Fronza and M. Budesinsky, Phytochemistry, 11, 3315 (1988). W. M. Daniewski, M. Gumulka, P. Skibicki, J. Krajewski and P. Gluzinski, Phytochemistry, 30, 1326 (1991). G. Magnusson and S. Thoren, Acta Chem. Scand., 27, 1573 (1973). W. M. Daniewski and J. Krol, Polish J. Chem., 55, 1247 (1981). W. M. Daniewski, M. Gumulka, E. Pankowska, K. Ptaszynska, E. Bloszyk, U. Jacobsson and T. Norin, Phytochemistry, 32, 1499 (1993). M. De Bernardi, G. Fronza, A. Scilingo, G. Vidari and P. Vita-Finzi, Tetrahedron, 42, 4277 (1986). W. M. Daniewski, M. Koc5r and J. Krol, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 23, 637 (1975). W. M. Daniewski, M. Kocor and J. Krol, Rocz. Chem., 50, 2095 (1976). W. M. Daniewski, M. Koc5r, T. Januszewski and A. Rymkiewicz, Polish J. Chem., 55, 807 (1981). W. M. Daniewski and M. Kocor, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 18, 585 (1970). W. M. Daniewski, W. Kroszczynski and J. Krol, Polish J. Chem., SI, 483 (1983). W. M. Daniewski, M. Koc5r and B. Zoltowska, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 21, 785 (1973). G. Vidari, M. De Bernardi, P. Vita-Finzi and G. Fronza, Phytochemistry, 15, 1953 (1976). M. De Bernardi, G. Fronza, G. Mellerio, G. Vidari and P. Vita-Finzi, Phytochemistry, 19, 99 (1980). W. M. Daniewski, M. Kocor and J. Krol, Rocz. Chem., 51, 1395 (1977). A. Bosetti, G. Fronza, G. Vidari and P. Vita-Finzi, Phytochemistry, 28, 1427 (1989). W. M. Daniewski, M. Gumulka, K. Ptaszynska, P. Skibicki, E. Bloszyk, B. Drozdz, S. Stromberg, T. Norin and M. Holub, Eur. J. Entomol., 90, 65 (1993). M. Bogucka-Ledochowska, A. Hempel, Z. Dauter, A. Konitz, E. Borowski, W. M. Daniewski, M. Kocor, Tetrahedron Lett., 2267 (1976). W. M. Daniewski, M. Gumulka, P. Gluzinski, J. W. Krajewski, E. Pankowska, K.

205

74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117

Ptaszynska, P. Sitkowski and E. Bloszyk, Polish J. Chem., 66, 1249 (1992). L. Garlaschelli, L. Toma, G. Vidari and D. Colombo, Tetrahedron, 50,1211 (1994). W. M. Daniewski, A. Eichart, J. Jurczak, L. Kozerski and J. S. Pyrek, Bull. Acad. Polon. ScL, Ser. Sci. Chim., 20, 131 (1972). E. Baranowska and W. M. Daniewski, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 20, 313 (1972). W. M. Daniewski, M. Gumulka, K. Ptaszynska, B. Kamienski, P. Skibicki, U. Jacobsson and T. Norin, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 37, 289 (1989). W. M. Daniewski, A. Wawrzun, M. De Bernardi, G. Vidari, P. Vita-Finzi, G. Fronza and G. Gatti, Tetrahedron, 40, 2757 (1984). D. Andina, M. De Bemardi, A. Del Vecchio, G. Fronza, G. Mellerio, G. Vidari and P. VitaFinzi, Phytochemistry, 19, 93 (1980). K.-G. Widdn and E.-L. Seppa, Phytochemistry, 18, 1226 (1979). A. Talvitie, K.-G. Widdn and E.-L. Seppa, Finn. Chem. Lett., 62 (1980). W. M. Daniewski, M. Gumulka, K. Ptaszynska, P. Skibicki, U. Jacobsson and T. Norin, Polish J. Chem., 66, 791 (1992). W. M. Daniewski, M. Gumulka, K. Ptaszynska, P. Sitkowski, P. Skibicki, U. Jacobsson and T. Norin, Bull. Acad. Polon. Sci., Ser. Sci. Chim., 39, 251 (1991). W. M. Daniewski, M. Gumulka and P. Skibicki, Phytochemistry, 29, 527 (1990). O. Sterner, R. Bergman, J. Kihlberg, J. Oluwadiya, B. Wickberg, G. Vidari, M. De Bemardi, F. De Marchi, G. Fronza and P. Vita-Finzi, / . Org. Chem., 50, 950 (1985). M. De Bernard!, G. Fronza, G. Vidari and P. Vita-Finzi, Chim. e Ind., 58, 177 (1976). W. M. Daniewski, M. Gumulka, K. Ptaszynska, P. Skibicki, U. Jacobsson and T. Norin, Phytochemistry, 31, 3933 (1992). R. Battaglia, M. De Bernardi, G. Fronza, G. Mellerio, G. Vidari and P. Vita-Finzi, / . Nat. Profif., 43, 319(1980). G. Vidari, L. Garlaschelli, M. De Bernardi, G. Fronza and P. Vita-Finzi, Tetrahedron Lett., 1773 (1975). J. Froborg and G. Magnusson, Tetrahedron, 34, 2027 (1978). S. Nozoe, H. Matsumoto and S. Urano, Tetrahedron Lett., 3125 (1971). G. Magnusson, S. Thoren, J. Dahmen and K. Leander, Acta Chem. Scand., B 28, 841 (1974). J. Kihlberg, R. Bergman, L. Nilsson, O. Sterner and B. Wickberg, Tetrahedron Lett., 24, 4631 (1983). O. Sterner, O. Wik and R. E. Carter, Acta Chem. Scand., B 42, 43 (1988). W. M. Daniewski, P. Gluzinski, J. W. Krajewski and P. Skibicki, J. Cryst. Spectr. Research, 21, 407 (1991). G. Magnusson and S. Thoren, Tetrahedron, 30, 1431 (1974). J. Froborg, G. Magnusson and S. Thordn, Acta Chem. Scand., B 28, 265 (1974). J.-D. Klamann, B. Fugmann and W. Steglich, Phytochemistry, 28, 3519 (1989). B. Fugmann, B. Steffan and W. Steglich, Tetrahedron Lett., 25, 3575 (1984). C. S. Hilger, B. Fugmann and W. Steglich, Tetrahedron Lett., 26, 5975 (1985). T. Suortti, A. von Wright and A. Koskinen, Phytochemistry, 22, 2873 (1983). M. De Bernardi, G. Vidari, P. Vita-Finzi and G. Fronza, Tetrahedron, 48,7331 (1992). A. Takahashi, G. Kusano, T. Ohta and S. Nozoe, Chem. Pharm. Bull., 36, 2366 (1988). H. Fujimoto, Y. Nakayama and M. Yamazaki, Chem. Pharm. Bull., 41, 654 (1993). E. Conca, M. De Bernardi, G. Fronza, M. A. Girometta, G. Mellerio, G. Vidari and P. VitaFinzi, Tetrahedron Lett,, 22, 4327 (1981). W. M. Daniewski, W. Kroszczynski, A. Wawrzun and A. Rymkiewicz, / . Liquid Chrom., 7, 2915 (1984). W. M. Daniewski, W. Kroszczynski and A. Wawrzun, Polish J. Chem., 61, 123 (1987). S. M. Camazine, J. R. Resch, T. Eisner and J. Meinwald, J. Chem. Ecol, 9, 1439 (1983). S. M. Camazine and A. Lupo, Mycologia, 76, 355 (1984). E.-L. Seppa and K-G. Widen, Ann. Bot. Fennici, 17, 56 (1980). H. Pyysalo, E.-L. Seppa and K-G. Widen, J. Chrom., 190, 466 (1980). J. A. Schmitt, Z. Pihkunde, 39, 219 (1974). O. Sterner, Ph. D. Dissertation, University of Lund (1985). J. Gamier, J. Mahuteau and M. Plat, Plant. Med. Phytother., 24, 87 (1990). L. Gariaschelli and G. Vidari, Tetrahedron, 45, 7371 (1989) L. Gariaschelli, P. de Tullio and G. Vidari, Tetrahedron, 47, 6769 (1991) W. Steglich and O, Sterner, Z. MykoL, 54, 175 (1988).

206 118 M. De Bernard!, G. Vidari, P. Vita-Finzi and K. Gluchoff-Fiasson, Tetrahedron Lett., 23, 4623 (1982). 119 P. List and H. Hackenberg, Zeitschr. f. Pilzkunde, 39, 97 (1973). 120 P. D. Cradwick and G. A. Sim, Chem, Comm., 431 (1971). 121 F. W. Comer, F. Mc Capra, J. H. Qureshi and A. I. Scott, Tetrahedron, 23, 476 (1967). 122 J. M. Ferland, Y. Lefebvre, R. Deghenghi and K. Wiesner, Tetrahedron Lett., 3617 (1966). 123 G. Cimino, S. De Stefano and L. Minale, Experientia, 28, 1401 (1972). 124 G. W. Reynolds and E. Rodriguez, Phytochemistry, 20, 1365 (1981). 125 C. Allievi, M. De Bemardi, F. De Marchi and G. Mellerio, / . Chrom., 261, 311 (1983). 126 J. Nawrot, E. Bloszyk, H. Grabarczyk and B. Drozdz, Pr. Nauk lOR, XXIV, 27 (1982). 127 J. Nawrot, E. Bloszyk, H. Grabarczyk, B. Drozdz, W. M. Daniewski and M. Holub, Pr. Nauk 70/?, XXV, 91 (1983). 128 J. Nawrot, B. Drozdz and M. Holub, Herba Polonica, XXXI, 209 (1985). 129 J. Nawrot, E. Bloszyk, J. Harmatha, L. Novotny and B. Drozdz, Acta entomol. hohemoslow., 83, 327 (1986). 130 J. Knuutinen and A. von Wright, Mutation Res., 103, 115 (1982). 131 A. von Wright, J. Knuutinen, S. Lindroth, M. Pellinen, K.-G. Wid^n and E.-L. Seppa, Fd. Chem. Toxic, 20, 265 (1982). 132 O. Sterner, R. Bergman, E. Kesler, G. Magnusson, L. Nilsson, B. Wickberg, E. Zimerson and G. Zetterberg, Mutation Res., 101, 269 (1982), 133 O. Sterner, R. Bergman, C. Franzen, E. Kesler and L. Nilsson, Mutation Res., 104, 233 (1982). 134 T. Suortti and A. von Wright, J. Chromatogr., 255, 529 1983). 135 A. von Wright and T. Suortti, Mutation Res., Ill, 103 (1983). 136 H. Anke and O. Sterner, Planta Medica, SI, 344 (1991). 137 H. Anke, O. Bergendorff and O. Sterner, Fd. Chem. Toxic, 27, 393 (1989). 138 H. Anke, E. Hillen-Maske and W. Steglich, Z. Naturforsch., 44c, 1 (1989). 139 H. Willstaedt and B. Zetterberg, Svensk. Kern. Tidskr., 58, 306 (1946). 140 B. N. Meyer, N. R. Ferrigni, J. E. Putuam, L. B. Jacobson, D. E. Nichols and J. L. McLaughlin, Planta Medica, 45, 3\ (1982). 141 G. Cimino, G. Sodano and A. Spinella, Tetrahedron, 43, 5401 (1987). 142 L. M. Nilsson, R. E. Carter, O. Sterner and T. Liljefors, Quant. Struct.-Act. Relat., 1, 84 (1988). 143 A. Forsby, E. Walum and O. Sterner, Chem. Biol. Interactions, 84, 85 (1992). 144 F. Bocchio, S. Kalf-Hansen, K. Dekermendjian, O. Sterner and R. Witt, Tetrahedron Lett., 33, 6867 (1992). 145 G. L. Fritz, G. L. Mills. J. D. Wanhen and R. M. Waters, J. Chem. EcoL, 15, 2607 (1989). 146 M. De Bemardi, M. A. Girometta, G. Mellerio, G. Vidari and P. Vita-Finzi, Micologia Italiana, 11, (1) 25 (1982). 147 A. Takahashi, G. Kusano, T. Ohta and S. Nozoe, Chem. Pharm. Bull., 41, 2032 (1993). 148 P. Skibicki, XV Conference on Isoprenoids, Zakopane (Poland), Abstract book, 27 (1993).

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

207

Structure and Biosynthesis of Bile Alcohols: Disorders of Cholesterol Side-Chain Oxidation in Cerebrotendinous Xanthomatosis Bishambar Dayal, Gerald Salen and Sarah Shcfcr Introduction Bile alcohols are polyhydroxy C27 sterols that serve as intermediates in the biosynthesis of cholic acid and chenodeoxycholic acid from cholesterol (1, 2). Recently several studies have shown that increased amounts of bile alcohols namely 27-nor-5p-cholestane-3a,7a,12a,24, 25-pentol and 5P-cholestane-3a, 7a,12a,25,26-pentol are excreted (as glucuronides in urine of patients with liver diseases such as primary biliary cirrhosis (3), liver cirrhosis (4, 5) and aantitrypsin deficiency (6). Ichimiya et. al., described the occurrence of 5Pcholestane-3a,7a,12a,26,27-pentol (5P-cyprinol) and 5P-cholestane-3a,7a, 12a,26-tetrol (27-deoxy-5P-cyprinol) in addition to 27-nor-5P-cholestane-3a, 7a,12a,24,25-pentol and 5P-cholestane-3a,7a,12a,25,26-pentol (5p-bufol), in the urine of a patient with obstructive jaundice (7). In other studies they demonstrated that in complete extrahepatic cholestasis 5P-cholestane-24,26pentol, 5P-cholestane-25,26-pentol and 5P-cholestane-3a,7a,12a,26-tetrol were excreted in increased amounts (8). Synthesis and absolute stereochemistry of 26 (or 27)-nor-5p-cholestane-3a,7a,12a,24S,25-pentol isolated from the urine and feces of a sitosterolemic patient were recently reported from our laboratory (9). It has been suggested that the appearance of these compounds might reflect the alternate pathway of biosynthesis of bile acids from cholesterol (2-5, 10, 11). The presence of multiple defects at different levels of the synthetic pathway of bile acid in liver diseases and the reactions leading to bile alcohol formation in several subcellular compartments may well be reflected in the bile alcohol profile which, in turn, may give further insight to the subcellular function of the liver (11). In 1971, Salen reported (12) that the rare inheirted lipid storage disease, cerebrotendinoux xanthomatosis (CTX), was associated with defective bile acid synthesis. The major and prominent clinical features in CTX syndrome were tendon xanthomas, juvenille cataracts, dementia, pyramidal paresis, cerebellar ataxis, abnormal electroencephalogram (EEG), and cerebral computed tomographic (CT) scans, premature atherosclerosis, pulmonary dysfunction and osteoporosis. Low serum levels of 25-hydroxyvitamin D3 and 24,25dihydroxyvitamin D3 were also detected in these patients in association with osteoporosis and frequent bone fractures (13,14). The disease is inherited as an autosomal recessive trait, but is usually detected in adults when cholesterol and cholestanol have accumulated over many years (13-16). Major biochemical

208

abnormalities included increased levels of plasma and tissue cholestanol (5aderivative of cholesterol) and defective bile acid synthesis that manifests in the virtual absence of chenodeoxycholic acid in the bile and the excretion of large amounts of unusual bile alcohols (17,18). Isolation and Characterization of Bile Alcohols in Cerebrotendinous Xanthomatosis (CTX ) Chromatographic analysis of the polar sterol fractions of the bile and feces, in conjunction with gas-liquid chromatography (GLC)-mass spectrometry indicated two major components, 5P-cholestane-3a,7a,12a,25-tetrol and 5Pcholestane-3a,7a,12a,23,25-pentol, and a minor component, 5P-cholestane-3a, 7a,12a,24,25-pentol(18-21). Only minute amounts of 5P-cholestane-3a, 7a, 12a,23R-tetrol, 5P-cholestane-3a,7a, 12a,24R-tetrol, 5P-cholestane-3a,7a, 12a,24S-tetrol and 5P-cholestane-3a,7a,12a,25S,26-pentol have been detected (20-24). The presence of 5P-cholestane-3a,7a,12a,25-tetrol was positively identified by comparison with the synthesized sample prepared in our laboratory (20-22,24-27) (Fig.l). The predominent bile alcohol of the pentol fraction was 5P-cholestane-3a,7a,12a,23,25-pentol, amounting to approximately 80% by weight, while 5P-cholestane-3a,7a,12a,24,25-pentol accounted for approximately 20% of this fraction. The less abundant pentol was shown to be identical with 5P-cholestane-3a,7a,12a,24a,25-pentol, which had been prepared from 5P-cholestane-3a,7a,12a,25-tetrol (20,21).

CH-N2

H

IV H FIG. 1. Synthesis of 5P-Cholestane-3a, 7a, 12a, 25-tetrol. I, Choiic acid; II, 3a, 7a, 12a-Triformoxy-cholanicacid; III, 3a, 7a, 12a-Triformoxy-24-oxo25-diazo-25-homo-5P-cholane; IV, Triformoxy-methyl homocholate; V, 5P-Cholestane-3a, 7a, 12a, 25-tetrol.

209 25-OH

3-OH12-OH 7-OH

n

"12PH 7P

J

-Cholostano-3a,7a,12a,25-te'trol

(PPM)

3.0 4 , 3 4 . 2 4 . 1 4 . 0 3 . 9 3 . 8 3 . 7 3 . 6 3 . 5 3 . 4 3 . 3 3 . 2 3 . 13.0 r 2 (PPM) ,CH3

18

26.27

CH3

c^^

19 2U Solvent 25-OH 12-OH I H7a&H3p

( b )5p-Choles!ane-3a,7p.12a.25-totr6l

^V'l

UL -

7-OH Hi2&

'r

•

4.5

-

4.0

-

3.5

-

3.0

-

2.5

-

2.0

-

1.5

-

1.0

-

0.5

4.5

• • • %

-

. i4_.

. r

F

•

• '

•

•

•

• • •

.

1

". I Jt. 4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

r

F2 (PPM)

Fig.2

400MHz iH-'H Correlated 2D NMR Spectrum

210

The biosynthetic 5p-cholestanetetrol and pentol had the same melting point, TLC and GLC properties, and infrared and mass spectra as the reference compound (19-21). In addition, the recently described two-dimensional ^HNMR studies (28-30) from our and other laboratories have further elucidated their structure and stereochemistry. Fig. 2 illustrates 2D-NMR ^H-^H COSY spectra of 5p-cholestane-3a,7a, 12a 25-tetrol (for comparison 2D-NMR ^H-^H COSY spectra of 5P-cholestane-3a,7P, 12a,25-tetrol is also provided). The 13C-DEPT spectra of (24R) and 5p-cholestane-3a,7a,12a,24a,25-pentol and (23R)-5P-cholestane-3a,7a,12a, 24a,25-pentol has been recently described (21). Furthermore, since knowledge of the C-24 and C-25 stereochemistry in these bile sterols is of great biosynthetic relevance in our studies, we have utilized Sharpless modified asymmetric dihydroxylation approach (21,31-36) to synthesize (24R and 24S)-5p-cholestane-3a,7a,12a,24,25-pentols and (25R and 25S)-5P-cholestane-3a,7a,12a,25,26-pentols from the olefinic mixture of 3a,7a,12a-triacetoxy-5P-cholest-24-ene-triol and its corresponding A^S-triol (Fig. 3). These Sharpless AD-mix reagents as described offered high levels of enantioselectivity and practicality in introducing chiral 1,2-glycol functionality into olefinic substrates (33-36). This synthetic approach, which we selected not only was stereoselective, but also established the absolute configuration of the two 5p-cholestane pentols required for biosynthetic studies. Very high optical purities thus obtained in (24R, 24S) and (25R, 25S)-5P-cholestanepentols and their derivatives were confirmed by the results of lanthanide-induced CD Cotton effect measurements and ^H and ^^C (DEPT) nuclear magnetic resonance studies (33). This aspect of research on the asymmetric dihydroxylation of 5Pcholestene-triols using catalytic amounts of Os04 in the presence of chiral catalysts have recently been reported from our laboratory (21,33). With regard to the identification of the predominating pentol as 5Pcholestane-3a,7a,12a,23,25-pentol electron impact mass spectrometric studies (19) performed on the underivatized bile alcohol had molecular ion at m/e 452 and a prominent peak at m/e 253, suggessting a 5P-cholestanepentol with three hydroxy 1 substituents in the ring system (19). The mass spectrum of the TMSi derivative indicated that there were five hydroxyl groups, three in the ring system and two in the side-chain. The presence of a hydroxyl group at C-25 seemed highly probable because of the fragment ion at m/e 131. The presence of a hydroxyl group at C-23 was suggested by the series of fragment ions at m/e (TMS derivative) 667, 577, 487,397, and 307 (19,27). Recent FAB-MS studies of the underivatized 23,25-pentol exhibited (M+H)"^ 453 and (M+Na)"^ 475 which confirmed unambiguously its molecular weight (Fig. 4) (37). Additional confirmation of the structure of the 23, 25 pentol was obtained by NMR. The chemical shift of the C-21 methyl group in the bile alcohol series appears as a doublet ( at 102.6 Hz in 5P-cholestane-3a,7a,12a,25-tetrol and at 103.6 and 104.8 Hz in the 5p-cholestane-3a,7a,12a,24a,25-pentol and in the 23,25-pentol respectively ) with a coupling constant of 6 Hz.

211

HO

AoO

t

,, HOAc/Ac,0 OH ^ AoO^' Anhyd. NaOAc

HO^'

i-CHj AD-mix-p *-

HO"''

if

"'OH

IV ^DHpH OH

-M.,Kl.pi.ll...lhtl.,UI.,.l..,..l.,IU.,../.|.lM|.ll.,UJ.,....,..M,..w|Ho,....,ol.,r...,I.M.....,....,^

300

M20

MGO

500

5M0

Fig. 4.

590

{>'•!*

G20

212

The presence of a hydroxyl group at C-22 (a to the C-21 methyl group ) in the 23,25-pentoI was excluded for the following reasons. 1. The C-21 doublet was not shifted appreciably ( approx. 1 Hz. ) compared with its position with 5pcholestane-3a,7a,12a,24a, 25-pentol. One would have expected a shift of 7 to 10 Hz for the C-21 doublet if the hydroxyl group had been situated at C-22 (deshielding effect ) (38). 2. The ^H-NMR spectrum in Fig. 2a shows that in the presence of a hydroxyl group at the C-25 position only ( as in 5(3cholestane-3a,7a,12a,25-tetrol) the C-26 and C-27 methyls appear as a sharp single peak because of their equal environment and free rotation around the C24 bond. The introduction of a hydroxyl group at the C-24 position ( as in 5Pcholestane-3a,7a,12a,24a,25-pentol, for its str. see Fig. 3 ) produces a difference in the chemical shift of C-26 and C-27 methyls of 2-3 Hz. This suggests that there is a restricted rotation along the C-24,25 bond so that the two methyl groups do not have equal environment. The same situation would have been exhibited by a 23-hydroxylated pentol. The inspection of a molecular model of 5P-cholestane-3a,7a,12a,23,25-pentol shows that the introduction of a hydroxyl group at C-23 causes the methyl groups at C-26 and C-27 to be shielded to a greater extent than by the introduction of a hydroxyl group at C24. If the 23,25 pentol had a C-22 substitution instead of C-23 as in ecdysone, then the C-26 and C-27 methyls would again be equivalent because their environment is not affected by the C-22 hydroxyl group, which is four carbon atoms distant. One would then expect free rotation around the C-24 bond, as in 5p-cholestane-3a,7a,12a,25-tetrol, and a sharp singlet for the C-26 and C-27 methyls. The presence of a hydroxyl group not vicinal to the C-25 hydroxyl was further suggested by the failure of lead tetraacetate to cleave the side chain. In contrast, lead tetraacetate oxidation of 5P-cholestane-3a,7a,12a,24a,25pentol produced a cleavage of the side-chain with the formation of cholyl aldehyde. The occurrence of bile alcohols hydroxylated at position 25 in CTX patients indicated the presence of an alternate pathway of bile acid synthesis from cholesterol (via 25-hydroxylated intermediates ). In particular, the identification of 5P-cholestane-3a,7a,12a,24a,25-pentol indicated that cholic acid arised from the cleavage of a 24,25- glycol. Absolute Configuration of Pentahvdroxv Bile Alcohols by Lanthanide-Induced Circular Dichroism (CD) Studies. During the course of these studies, the two 5P-cholestane-3a,7a, 12a,24,25-pentols, epimeric at C-24, and 5p-cholestane-3a,7a,12a,23,25pentol epimeric at C-23 were further resolved by analytical and preparative TLC, and were either synthesized or isolated and characterized. In addition, the absolute stereochemistry at C-23 of 5P-cholestane-3a,7a,12a,23,25-pentol and at C-24 of 5P-cholestane-3a,7a,12a,24,25-pentol (39,40) was established by circular dichroism (CD) studies employing the method of Nakanishi (41,42). These experiments conclusively defined the chirality of these pentahydroxy bile alcohols having 1,2 and 1,3 glycol systems in the side chain. Using Eu(fod)3

213

under anhydrous conditions, we obtained the desired CD spectra exhibiting very large induced split Cotton effects. On the basis of the empirical rule the bile alcohols 5P-cholestane-3a,7a,12a,24a,25-pentol (I) and 5p-cholestane-3a,7a, 12a,24p,25-pentol (II) were assigned the 24R and 24S configurations respectively and 5p-cholestane-3a,7a,12a,23a,25-pentol (III) was shown to possess the 23R configuration. These assignments were fully confirmed by comparison with "24(R), 25-dihydroxycholesteror(IV) (Fig. 5) a model compound whose single-crystal X-ray structure has been determined (Table 1) (43). Similarly lanthanide-induced CD and ^^c NMR studies elucidated the absolute configuration at C-25 in 5p-cholestane-3a,7a,12a,25,26-pentol as 25S (44). pt\

HO'

HO

IV

Fig. 5. Structures ofisomeric bile alcohols

With regard to the chirality of this at C-25 (Primary-tertiary glycol system in the side-chain) recently Partridge, et al. (45) have shown unequivocally, that for primary-tertiary a-diols, the R compounds have positive Cotton effects and the S compounds have negative Cotton effects at 318 nm. Therefore 25,26-pentol was assigned S-configuration at C-25. The configuration at C-24 of (24R)-pentol, (24S)-pentol and (23S)pentol was further established by NMR experiments and verified by CD analysis The results of ^H- and 13c- NMR studies are summarized in Table 2. The C24H in (24R) pentol shows triplet at 6 3.4 and in (24S) shows doublet at 3.2 in 1H NMR. The C-23H in (23R) pentol also shows triplet at 5 4.1 in 1H NMR. The l^c spectrum in CDCI3 exhibits 30 signals between 140 and 10 ppm (data not shown) and analyzed by DEPT-135 spectrum in CDCI3 which allowed the identification of 7 quaternary carbons, 5 CH's, 10 CH2's, 8 CH3*s for three compounds. However, these ^^C data could not differentiate 24S-pentol from 24R-pentol (33).

214

Table 1. Circular dichroism of 5P-cholestanepentols epimeric at C-23 and C-24. Entry Compound 5P-Cholestane-3a,7a ,12a,24a,25-pentol 5P-Cholestane-3a,7a ,12a,24P,2 5-pentol 5P-Cholestane-3a,7a ,l2a,23a,2 5-pentol 5P-Cholest5-ene-3p, 24R,25-triol

Orign of sample

Molar CD Solvent ratio AE X, nm Substate Eu(fod)3

Chirality

Isolated from CTX patients and synthesized.

1:1

CHC13

-13.5 309 +9.2 285

24 R

Synthesized in our lab.

1:1

CHC13

+9.5 308 -5.9 283

24 S

Isolated from CTX patients

1:1

CHCI3

-2.7 320 +2.7 290

23 R

Synthesized

1:1

CHCI3

-11.5

308

24 R

The sign of the longer wave length Cotton effect (first Cotton effect) is in agreement with the chirality of the acyclic a-glycol. Table. 2 iR- and l^C-NMR data for (24R) and (24S) pentols (in CDCI3 solution).

13c

Types 5P-cholestane-3a. 7a, 12a, 24R,25-pentol. 5P-cholestane-3a, 7a, 12a, 24S,25-pentol. 5|3-cholestane-3a, 7a, 12a, 23R,25-pentol.

3.4 (t) C-24H" 3.2 (d) C-24H 4.1(t)C-23H

78.6 (C-24) 79.6 (C-24) 68.0(C-23)

The insertion of hydroxyl groups into the 23- or 24-position of 5Pcholestane-3a,7a,12a,25-tetrol was found to be stereospecific. Although all these compounds were potential precursors of bile acid, studies in vivo and in vitro experiments using [3P-^H] and (24-l^C) 5p-cholestane-3a,7a,12a,25tetrol (46) (Figs.6, 7), (24-^^c) 5p-cholestane-3a,7a,12a,24R,25-pentol and (24-^^C) 5P-cholestane-3a,7a,12a,24S,25-pentol demonstrated the existence of a new 25-hydroxylation pathway for the transformation of cholesterol to cholic acid in these patients (2,10). The reaction sequence involved the stereospecific formation of a 24S-hydroxy pentol, 5P-cholestane-3a,7a,12a,24S,25-pentol, 3a7a,12a,25-tetrahydroxy-5P-cholestan-24-one and did not involve 5Pcholestanoic acids as intermediates (Fig. 8). The two bile pentols, 5Pcholestane-3a,7a,12a,24R, 25-pentol and 5P-cholestane-3a,7a,12a,23R,25-

215

pentol apparently accumulated because CTX liver has a markedly diminished capacity (one-third as great as normal liver) to 24S-hydroxylate 5P-cholestane3a,7a,12a,25-tetrol to form 5p-cholestane-3a, 7a,12a,24S,25-pentol. HO

HO^'

m

80/20«a/p

"^

H

—- IV

Fig. 6. Preparation of 5p-cholestane-3a,7a,12a,25-tetrol[3P-^l

HO^^

ffl

+

^^CHjNz

1 Equiv.

1 Equiv.

+

(C2H5)3N

IV + CH,OH

V

i

VI

Fig. 7. Synthesis of 5P-cholestane-3a,7a,12a,25-tetro1 [24 -^^^C]

216 K

HO^

H

vOH

OH

5P-Cholestane-3a,7a,12a,25-tetrol

5P-Cholestanc-3a,7a,12a,24P,25-pentol ((24 S)-pentol)

25% I HO^ vH "^

HO^

COOH

H

5P-Cholestane-3a,7a,12a,24a,25-pentoI ((24 R).pentol)

Cholic Acid

Fig. 8. In V/VroTransformationof 5P-cholestanc-3a,7a,12a,25-tetrol.

Further studies from our laboratory demonstrated that in CTX the impaired synthesis of bile acids is due to a defect in the biosynthetic pathway involving the oxidation of the cholesterol side-chain (47). Where as the synthesis of both primary bile acids, cholic and chenodeoxycholic acids, is subnormal, that of chenodeoxycholic acid is more markedly affected. As a consequence of the inefficient side-chain oxidation, increased 23,24 and 25hydroxylation of bile acid precursors occurs with the consequent marked increase in bile alcohol glucuronides secretions in bile, urine and plasma while in feces free bile alcohols are obtained (48-52, 22). Only ether glucuronides were observed in this study and there was no detectable accumulation of bile alcohol glucuronides lacking 12a-hydroxyl group (53). These compounds were isolated and characterized by capillary coulmn gas-liquid chromatography (GLC) and gas chromatography-mass spectrometry (GC-MS) after treatment with P-glucuronidase (53). Intact bile alcohol glucuronides from bile , urine and plasma were also analyzed by secondary ion and fast atom bombardment mass spectrometry (FAB-MS) (53-55). A combination of TLC and positive ion FAB-MS for the characterization of underivatized major biliary and plasma bile alcohol glucuronide, 5P-cholestane-3a,7a,12a,25-tetrol-3-0-P-D-glucuronide and the minor compound 5P-cholestane-24,25-pentol glucuronide isolated from bile were recently reported from our laboratory (53,54). The specta were characterized by abundant ions formed by attachment of a proton, [M + H]"^,or of alkali ions,[M + Na]"*" and [M + K]"*",to the glucuronide salt. These ions allowed an unambiguous deduction of the molecular weight of the sample. Positive ion FAB-MS spectra of underivatized biliary bile alcohol glucuronide.

217

5p-cholestane-3a,7a,12a,23,25-pentol-3-0-P-D-glucuronide did not provide molecular ion peak but a series of fragment ions diagnostic of the glucuronide residue attached at the C-3 position of the bile pentol were exhibited. Using negative ion FAB-MS Egestad et al. (55) has also reported the analysis of intact conjugated bile alcohols from the urine of a CTX patient. These studies indicated that the more polar metabolite (23R) 5P-cholestane-3a,7a,12a,23,25pentol was excreted in greatest quantities in urine and 5P-cholestane-3a,7a, 12a,25-tetrol comprised 15% of the total bile alcohol glucuronides (55). The difference between the patterns of urinaiy,biliary,and plasma excretion of bile alcohols and the predominance of the 5p-cholestane-3a,7a,12a,23R,25-pentol in urine is most probably due to more efficient renal excretion of the more polar 5P-cholestane-3a,7a,12a,23R,25-pentol or renal C-23 hydroxylation of 5p-cholestane-3a,7a,12a,25-tetrol(52). Transfprmatipng pf Bile Alcphqlg; Biosynthetic Pathways of Bile Acid Synthesis from cholesterol: According to current information, bile acid synthesis begins with hydroxylation of the cholesterol nucleus followed by shortening and oxidation of the side chain (2,56-58). The major reactions in these transformations are shown in (Figs. 9 and 10). The first reaction is catalyzed by hepatic cholesterol 7 a monooxygenase (cholesterol 7a-hydroxylase, a cytochrome P-450-dependent hydroxylase (56). Under normal circumstances, this step is considered ratecontrolling for bile acid synthesis (56-58). Further metabolism of 7a-hydroxy cholesterol involves oxidation of the 3P-hydroxy 1 group to a ketone and isomerization of the double bond from C-5,6 to C-4,5 by hepatic microsomal 3P-hydroxy D4-5 steroid dehydrogenase-isomerase to yield 7a-hydroxy-4cholesten-3-one (III). This compound is the last intermediate which is common to both cholic acid and chenodeoxycholic acid synthesis and is a favored substrate for 12a-hydroxylase, which is also a cytochrome P-450 dependent enzyme. Thus for cholic acid formation, 12a-hydroxylation occurs to give 7a,12a-dihydroxy-4-cholesten-3-one which then undergoes a stereospecific reduction of the double bond at C-4 and the 3-oxo group to form 5P-cholestane3 a , 7 a , 1 2 a - t r i o l (soluble fraction of the liver cell) (56-58). For chenodexycholic acid formation, 7a-hydroxy-4-cholesten-3-one is similarly reduced to produce 5p-cholestane-3a,7a-diol (2,56-58). At this piont, two pathways have been proposed for degradation of the cholestane side chain in the biosynthesis of bile acids. These differ in the site proposed for the first hydroxylation step in the side-chain oxidation of cholesterol for the formation of cholic acid(2,56-58). 25-Hydroxylation pathway : This pathway has been demonstrated in both rat and human liver (2,10,56). It involves the 25-hydroxylation of 5P-cholestane3a,7a,12a-triol to give 5P-cholestane-3a,7a,12a,25- tetrol (XIV) followed by stereospecific 24S-hydroxylation to yield 5P-cholestane-3a,7a,12a,24S,25pentol (XV, Fig. 9). The pentol is then oxidized to 5P-cholestane-3a,7a,12a ,25-tetrahydroxy-5p-cholestan-24-one (XVI) (59,60), which is degraded by

218

cytosolic enzymes to give cholic acid and acetone(10,56) (Fig. 10a and 10b) . It was demonstrated that CTX liver microsomes have 20% of the capacity of normal liver microsomes to form 5P-cholestane-3a,7a,12a,24S,25-pentol. Thus, 5P-cholestane-3a,7a,12a,25-tetrol accumulates in the liver and plasma where, during passage through the kidney, further hydroxylations at C-22, C23, C-24, and C-26 likely occur (56). It is important to emphasize that once glucuronidation takes place bile alcohol glucuronides cannot be transformed to bile acids. Therefore, according to this formulation, the major block in bile acid synthesis involves 24S hydroxylation of 5P-cholestane-3a,7a,12a,25-tetrol (a deficiency of 5P-cholestane-3a,7a,12a,25-tetrol-24S hydroxylase) in cholic acid synthesis (56). The deficiency of chenodeoxycholic acid is believed to result from the diversion of virtually all precursors to the cholic acid pathway because 12a-hydroxylation is more active in CTX than normal liver. Most evidence favors defective bile acid synthesis that involves incomplete oxidation of the cholesterol side chain as the primary abnormality in CTX. However, the precise location of the enzymatic abnormality remains controversial. Recent studies by Duane et al; demonstrated normal dominance of the 26-hydroxylation pathway but he cautions that this is not to say that 24-hydroxylation is not impaired in CTX patients (61,62). Indeed, impaired activity of this enzyme remains one possible reason that so little of the 25-hydroxy bile alcohols produced in CTX is further oxidized to bile acid (62). 26-Hvdroxvlation pathwav: An alternative explanation for the bile acid synthetic defect in CTX has been proposed by Oftebro and colleagues which starts via 26hydroxylation of 5P-cholestane-3a,7a,12a-triol (IX, Fig. 10a and 10b). In this pathway the mitochondrial fraction of both human and rat liver contains a 26hydroxylase enzyme (63) which can convert 5P-cholestane-3a,7a,12a-triol (IX ) to 5P-cholestane-3a,7a,12a,26-tetrol (XI) (Fig. 10a and 10b ). This tetrol is oxidized to 3a,7a,12a-trihydroxy-5p-cholestan-26-oic acid (THCA, XII) by liver cytosol (2,64). Further hydroxylation at C-24 forms varanic acid (XIV) and its side chain is shortened with oxidation at C-24 to yield cholic acid (X,Fig. 10 a). These investigators demonstrated diminished mitochondrial 26hydroxylation of 5p-cholestane-3a,7a,12a-triol and 5P-cholestane-3a,7a-diol, possible precursors for cholic acid and chenodeoxycholec acid in CTX liver. As a consequence, neither 26-hydroxylated intermediates can be formed so that total primary bile acid synthesis would be diminished. Accordingly, the accumulation of 5P-cholestane-3a,7a,12a,25-tetrol arises from 25hydroxylation of 5P-cholestane-3a,7a,12a-triol by the alternative microsomal 25-hydroxylation mechanism.

219

HO'

III. 7a-Hydroxycholest-4-en-3-one HO

IV. 5P-Cholestane-3a,7a-diol

I

Vin. 7a, 12a-Dihydroxycholest-4-cn-3-onc

CH2OH

IKT

V. 5P-Cholestane-3a,7a,26-triol

IX. 5P-Cholestane-3a,7a,12a-triol

i CXXDH

r'''y'y^^^ jj^»'^Vsx'''jv^'//^^ VI. Chenodeoxycholic acid

+

COOH

CH3CH2C00H Vn. Propionic acid

+

CH3CXX:H3

XI. Acetone X. Cholic acid

Fig. 9. The elaboration of the cholesteol nucleus in bile acid synthesis. (Cholic acid and chenodeoxycholic acid biosynthesis pathway).

220

IX. 5P-Cholestane-3a,7a,12a-triol

XIII. 5P-Cholestane-3a,7a,12a,25-tetrol

X. 5P-Cholestane-3a,7a,12a,26-tetrol

COOH

HO'

XI. 3a,7a,12a-Trihydroxy-5P-cholestanoic acid

XIV. 5P-Cholestane-3a,7a,12a,24P,25-pentol ((24 S)-pentoI)

XV. 3a,7a,12a,25-Tetrahydroxy.5pcholestan-24-one

XII. 3a,7a,12a-Trihydroxy-5Pcholestanoyl-CoA

Fig. 10 (a).

221

XV. 3a,7a,12a,25.Tetrahydroxy-5Pcho1estan-24-one

XII. 3a,7a,12a-Trihydroxy-5Pcholestanoyl-CoA

XIII. 3a,7a,12a,24Metrahydroxy-5Pcholestanoyl-CoA (varanoyl-CoA)

X. Cholic acid

XI. Acetone

COOH +

HO^

CH3CH2CCX3H

Vn. Propionic acid

X. Cholic acid Fig. 10 (b). The sequence leading to the oxidation and cleavage of the side chain of 5P-cholestane-3a,7a, 12a-triol; pathway for side-chain cleavage in cholic acid biosynthesis.

However, the precise localization of the enzymatic defect in bile acid synthesis in CTX (microsomal 24S hydroxylation or mitochondrial 26hydroxylation) awaits the determination of which side chain oxidative mechanism is quantitatively most important in bile acid synthesis. Until then, both mechanisms must be considered (56,64,65). While these studies were in progress, Reshef et al reported the analysis of the sterol 27-hydroxylase gene mutant allele in a Jewish family of Algerian origin (64). They further stated that the CTX disorder which is characterized by extensive nervous system involvement, juvenile cataracts, tendon xanthomatosis and premature atherosclerosis was caused by the sterol 27-

222

hydroxylase mutation. An additional mutant allele found in a Jewish family of Algerian origin was characterized. Sequence anaylsis revealed a C to T transition at cDNA position 1037 which predicted a threonine to methionine substitution at residue 306 (designated T306M). It was suggested that this transition was the mutation causing CTX in that family. The three sterol 27hydroxylase gene mutations accounted for all 10 CTX families and their presence suggested the existance of positive selective forces that lead to an increased prevalence of the relatively rare disease in Jews from North Africa (66-69). In summary, these studies demonstrated that in CTX the impaired synthesis of bile acids is due to a defect in the biosynthetic pathway involving the oxidation of the cholesterol side-chain. As a consequence of the inefficient side-chain oxidation, increased 23, 24 and 25-hydroxylation of bile acid precursors occurs with the consequent marked increase in bile alcohol glucuronides secretions in bile, urine, plasma and feces (free bile alcohols). These compounds were isolated, synthesized and fully characterized by various spectroscopic methods. In addition, their absolute stereochemistry determined by Lanthanide-Induced Circular Dichroism (CD) and Sharpless Asymmetric Dihydroxylation studies. Further studies demonstrated that (CTX) patients transform cholesterol into bile acids predominantly via the 25-hydroxylation pathway. This pathway involves the 25-hydroxylation of 5P-cholestane-3a,7a, 12a-triol to give 5P-cholestane-5P-cholestane-3a,7a,12a,25- tetrol followed by stereospecific 24S-hydroxylation to yield 5P-cholestane-3a,7a,12a,24S,25pentol which in tum was converted to cholic acid. Finally, treatment of these patients with chenodeoxycholic acid (750 mg per day) decreased cholestanol production and plasma concentrations of bile alcohols and suppressed abnormal bile acid synthesis (70). Materials 9nd Methods Melting points: were determined on a Thermolyne apparatus (Thermolyne Corp., Dubuque, lA) model MP-12600 and are uncorrected. Thin-laver chromatographv (TLC): All tetrahydroxy bile alcohols and their corresponding isomers were separated on Silica gel G plates (Analtech, Uniplates, Newark, NJ, 0.25-mm thickness) in the solvent system: CHCl3/(CH3)2CO/CH30H, 70:50:10 (v/v/v). For pentahydroxy sterols and their corresponding isomeric analogs, the solvent system CHCl3/(CH3)2CO/ CH3OH, 70:50:15 (v/v/v) was used (9). The spots were detected with phosphomolybdic acid (3.5 % in isopropanol), sulfuric acid (10 %) and heating for 1 minat HOC. Conjugated bile acids and bile alcohols were analyzed by TLC using silica gel G plates (Brinkmann Instruments, Westbury, NJ; 0.25-mm thickness) with the solvent system chloroform/methanol/acetic acid/water 26:8:4:2 (v/v/v/v) (5,7,9). Bile salts were made visible with a spray reagent which consisted of 10 % phosphomolybdic acid in ethanol. Detection with naphthoresorcinol (0.2 % in

223

acetone), aqueous 10% phosphoric acid ( 5 : 1 ) , and development at 100 C for 10 min gave a blue color, indicating the presence of bile alcohol glucuronides. ^H-NMR spectrometry: Nuclear magnetic resonance spectra were recorded on a Varian Associates XL-400 spectrometer equipped with Fourier transform mode. All NMR spectra were taken in (CDCI3 + CD3OD) solution (unless otherwise indicated) with Me4Si as the internal standard and the degree of substitution at each carbon was determined by experiments in the singlefrequency off-resonance decoupled mode. The nature of each carbon in the bile alcohols was deduced through Distortionless Enhancement by Polarization Transfer (DEPT) experiment (21,28) performed by using polarization pulse of 90^ and 135^, respectively, obtaining in the first case, only signals for CH groups and in the second case , positive signals for CH and CH3, and negative signals for CH2 groups. The two-dimensional I H - I R chemical shift correlated (2D NMR, COSY-45) data for isomeric 5P-cholestane-3a,7a,12a, 25-tetrol and 5P-cholestane-3a,7P, 12a,25-tetrol were acquired at a sweep width of 3200 Hz using a standard pulse sequence (collection of 256 free induction decays, FIDs) as described previously (28-30). A transform size of 2K x 2K data points was obtained after zerofilling. The heteronuclear 1 H - 1 3 C chemical shift correlated (HETCOR, 2D NMR) experiments were acquired (21) with sweep widths dictated by the appearance of the *3c and ^H spectra by using 2K data points in the 256 free induction decays (FIDs) collected on a 3200 Hz spectral width. Gas-liquid chromatographv: Capillary GLC analysis of bile sterols and bile acid methyl esters (as their trimethylsilyl derivatives) was performed on Hewlett-Packard model No. 4890 (equipped with a flame ionization detector) and a split column injector using a CP sil 5 (CB) WCOT capillary column (25 m X 0.22 mm with 0.13-mm film thickness). Helium was used as a carrier gas at a flow rate of 20.2 mL/min (135 kPa). Optical measurements: The circular dichroism (CD) measurements were carried out on a Jasco-500A spectropolarimeter at 24C, under a stream of high purity, dry N2. The coefficient of dichroic absorption, Ae, was calculated from the molar ellipticity [6] by the following equation: molar ellipticity [6] = 3300 Ae. Both the molar ellipticity [q] and Ae are expressed in degree x cm^ x dmol' 1(9). Positive Ion Fast AtpmBombardment Maigg SpQctrpmctry (FAB-Mg): Direct probe mass spectrometric analysis of the free bile alcohols and their intact conjugates (25-tetrol/pentol glucuronides) was performed by FAB-mass spectrometry using a ZAB-IF mass spectrometer (VG, Manchester) equipped with a standard FAB source. The FAB technique has emerged as a simple and extremely versatile ionization process that is particularly well suited to biomedical applications. Although negative ion FAB-MS is much more satisfactory than the positive ion mode, this was not available at that time. In the FAB mode (positive ions) the sample was dissolved in thioglycerol and a

224

small aliquot (1-2 |uL) placed on the mass spectrometer direct probe (54). After insertion into the mass spectrometer, the sample was bombarded with a neutral atom beam of xenon having approximately 6 kV of translational energy. Electron Ionization Mass spectra (EI-MS): of the bile alcohols were obtained with a Varian MAT-5 and Varian MAT-III gas chromatograph-mass spectrometer as described previously (9). High-performance liquid chromatography (HPLC): of the bile alcohols was carried out on a Waters Associates ALC 201 system employing a Waters model 401 refractive index detector and a Waters 4-mm ID x 30-cm in Bondapak Cjg column (Waters Associates Inc., Milford, MA). The solvent system was 80 % MeOH/H20 (v/v) with a flow rate of 1.5 mL/min. REFERENCES 1.

Danielsson, H., (1973).Mechanisms of bile acid biosynthesis. In The Bile Acids: Chemistry, Physiology and Metabolism. Physiology and Metabolism, Vol. 2. P.P. Nair and D. Kritchevsky, editors. Plenum Publishing Corporation, New York, 1-32.

2.

Salen, G., Shefer, S., and Berginer, V.D. (1983). Familial disease with storage of sterols other than cholesterol: cerebrotendinous xanthomatosis and sitosterolemia with xanthomatosis. In The Metabolic Basis of Inherited Disease. J. B. Stanbury, J. B. Stanbury, J. B. Wyngaarden, D. S. Fredrickson, J. L. Goldstein, and M. S. Brown, editors McGraw-Hill, New York. 713-730.

3.

Karlaganis, G., Alme, B., Karlaganis, V. and Sjovall J. (1981). Bile alcohol glucuronides in urine. Identification of 27-nor-5p-cholestane-3a, 7a, 12a, 24, 25-pentol in man. /. Steroid Biochem. 14: 341-345.

4.

Ludwig-Kohn, H., Henning, H. V., Sziedat, A., Mattaei, D., Spitteller, G., Reiner, J. and Egger, H. J. (1983). The identification of urinary bile alcohols by gas chromatography-mass spectrometry in patients with liver disease and in healthy individuals. Eur. J, Clin, Invest 13: 91-98.

5.

Karlaganis, G., Karlaganis, V. and Sjovall, J. (1984). Identification of 27nor-5P-cholestane-3a,7a,12a,24,25,26-hexol and partial characterization of the bile alcohol profile in urine. / Lipid Res, IS: 693-702.

6.

Karlaganis, G., Nemeth, A., Hammarskjold, B., Strandvik, B. and Sjovall, J. (1982). Urinary excretion of bile alcohols in normal children and patients with a-antitrypsin deficiency during development of liver disease. Eur, J, Clin, Invest, 12: 399-405.

225

7.

Ichimiya, H., Yanagisawa, J. and Nakayama, F. (1984). Significance of bile alcohols in urine of a patient with cholestasis: identification of 5Pcholestane-3a,7a,12a,26,27-pentol (5p-bufol) and 5P-cholestane-3a,7a, 12a,26-tetrol (27-deoxy-5P-cyprinol). Chem. Pharm. Bull 32: 28742877.

8.

Ichimiya, H., Yanagisawa, J. and Nakayama, F. (1987). Altered metabolism of bile alcohol and bile acid in complete extrahepatic cholestasis: qualitative and quantitative aspects. /. Lipid Res. 2S: 10281037.

9.

Dayal, B., Tint, G. S., Toome, V., Batta, A. K., and Shefer, S., and Salen G. (1985). Synthesis and structure of 26(or 27)-nor-5P-cholestane-3a, 7a,12a,24S, 25 -pentol isolated from the urine and feces of a patient with sitosterolemia and xanthomatosis. /. Lipid Res, 26: 298-305.

10.

Shefer, S., Cheng, F. W., Dayal, B., Hauser, S., Tint, G. S., Salen, G. and Mosbach, E. H. (1976). A 25-hydroxylation pathway of cholic acid biosynthesis in man and rat. /. Clin. Invest. 57: 897-903.

11.

Patteson, T. E., Vlahcevic, Z. R., Schwartz, C. C , Gustafsson, J., Danielsson, H. and Swell, L. (1980). Bile acid metabolism in cirrhosis. VI. Sites of blockage in the bile acid pathways to primary bile acids. Gastroenterology. 79: 620-628.

12.

Salen, G. (1971). Cholestanol deposition in cerebrotendinous xanthomatosis: a possible mechanism. Ann. Intern. Med. 75: 843-851.

13.

Berginer, V. M. and Abeliovich, D. (1981). Genetics of cerebrotendinous xanthomatosis: An autosomal recessive trait with high gene frequency in Sephardium of Moroccan origin. Am. J. Med. Genet 10: 151-157.

14.

Berginer, V. M., Shany, S., Alkalay, D., Berginer, J., Dekel, S., Salen, G., Tint, G. S., and Gazit, D. (1993). Osteoporosis and increased bone fractures in cerebrotendinous xanthomatosis (CTX) . Metabolism 42: 69-74.

15.

Berginer, V. M., Berginer, J,, Salen, G. et al. (1981). Computed tomography in cerebrotendinous xanthomatosis. Neurology (NY) 3 1 : 1463-1465.

16.

Menkes, J. H., Schimschok, J. R. and Swanson, P. D. (1968). Cerebrotendinous xanthomatosis: The storage of cholestanol within the nervous system. Arch. Neurol. 19: 47-53

226

17.

Salen, G. and Grundy S. M. (1973). The metabolism of cholestanol, cholesterol, and bile acids in cerebrotendinous xanthomatosis. /. Clin, Invest 52: 2822-2835.

18.

Setoguchi, T., Salen, G., Tint, G. S. and Mosbach, E. H. (1974). A biochemical abnormality in cerebrotendinous xanthomatosis: incomplete oxidation of the cholesterol side chain. / Clin. Invest 53: 1393-1401.

19.

Shefer, S., Dayal, B., Tint, G. S., Salen, G., and Mosbach, E. H. (1975). Identification of pentahydroxy bile alcohols in cerebrotendinous xanthomatosis: characterization of 5P-cholestane-3a,7a,12a,24e,25pentol and 5p-cholestane-3a,7a,12a,23e,25-pentol. /. Lipid Res, 16: 280-286.

20.

Dayal, B., Shefer, S., Tint, G. S., Salen, G., and Mosbach, E. H. (1976). Synthesis of 5P-cholestane-3a,7a,12a,25-tetrol and 5p-cholestane-3a, 7a, 12a,24e,25-pentol. /. Lipid Res, 17: 74-77.

21.

Dayal, B., Salen, G., Padia, P., Shefer, S., Tint, G. S., Williams, T. H., Toome, V., and Sasso,.G. (1992). Stereoselective synthesis of (24R,24S) 5P-Cholestane-3a,7a,12a, 24,25-pentol and (25R,25S ) 5P-cholestane3a, 7a,12a,25, 26-pentol using a modified osmium catalyzed Sharpless Asymmetric Dihydroxylation Process. Chem, Pbys, Lipids, 61: 271-281

22

Dayal, B., Batta, A. K., Shefer, S., Tint, G. S. and Salen, G. (1978). Synthesis of biological precursors of cholic acid. Steroids. 32: 337-344.

23.

Yashuara, M., Kuramoto, T. and Hoshita, T. (1978). Identification of 5P-cholestane-3a,7a,12a,23P-tetrol, 5P-cholestane-3a,7a,12a,24atetrol and 5P-cholestane-3a,7a,12a,24p-tetrol in cerebrotendinous xanthomatosis. Steroids, 31: 333-345.

24.

Dayal, B., Tint, G. S., Greeley, D. N., Williams, T. H., and Salen, G. (1983). Identification of 5p-cholestane-3a,7a,12a,25,26-pentol in cerebrotendinous xanthomatosis. Steroids. 42: 441-448.

25.

Dayal, B., Bagan, E., Speck, J., and Salen, G. (1980). A facile synthesis of 5P-cholestane-3a,7a,12a,25-tetrol. Steroids.35: 439-444.

26.

Dayal, B., Tint, G. S., Batta, A. K., Shefer, S., and Salen, G. (1981). Synthesis of biological precursors of cholic acid. Steroids,37: 205-211.

27.

Tint, G. S., Dayal, B., Batta, A.K., Shefer, S., Cheng, F.W., Salen, G., and Mosbach, E.H. (1978). Gas-liquid chromatography-mass

227

spectrometry of trimethylsilyl ethers of bile alcohols. /. Lipid Res. 19: 956-966. 28.

Dayal, B. and Salen. G. (1991). A stereospecific synthesis and twodimensional 1H-NMR investigation of isoursocholic acid. /. Lipid Res. 32: 1381-1387 and references cited theirin.

29.

Doddrell, D. M., Pegg, D. T., and Bendall, M. R.. (1982). Distortionless enhancement of NMR signals by polarization transfer. /. Magn. Reson. 48: 323-327.

30.

Bendall, M. R., and Pegg, D. T. (1983). Complete accurate editing of decoupled C13 spectra using DEPT and a quaternary-only sequence. / Magn. Reson. 53: 272-296.

31.

Wai, J. S. M, J. Marko, J. S. Svendsen, M. G. Finn, E. N. Jacobsen and K. B. Sharpless. (1989). A mechanistic insight leads to a greatly improved osmium-catalyzed asymmetric dihydroxylation process. /. Am. Chem. Soc. I l l : 1123-1125.

32.

Jacobsen, E. N., Marko, J., Mungall, W. S., Schroder,G., and Sharpless, K. B. (1988). Asymmetric dihydroxylation via ligand accelerated catalysis. /. Am. Chem. Soc. 110: 1968-1970.

33.

Dayal, B., Keshava Rao, Seong W. M., Pramanik, B. N., Huang, E. C, Toome, V., and G. Salen. (1994). Asymmetric syntheses and lanthanideinduced CD studies of (24R and 24S) 5P-cholestane-3a,7a, 12a,24,25pentols. Int. J. Pure and Applied Chemistry (m press).

34.

Takano, S., Yoshimitsu, T., and Ogasawara, K. (1994). Asymmetric dihydroxylation of a ineso-symmetric cyclic diene Using AD-mix reagents: A new enantiocontrolled route to conduritol E. /. Org. Chem. 59: 54-57.

35.

Xu, D., Crispino, G. A., and Sharpless, B, K. (1992). Selective asymmetric dihydroxylation of dienes /. Am. Chem. Soc. 114: 75707571.

36.

VanNieuwenhze, M. S., and Sharpless, K. B. (1994). The Asymmetric dihydroxylation of cis-AUylic and Homoallylic Alcohols. Tetrahedron Letters 35, 843-846.

37.

Dayal, B. and Salen, G. (1990). Fast atom bombardment mass spectrometry (FAB-MS) studies of conjugated bile alcohols. Presented in part at the 17th International Symposium on the Chemistry of Natural

228

Products (17th lUPAC Meeting) held in New Delhi, India, February 410. 38.

Hoffmeister, H., Rufer, H., Keller, H. H., Schairer, H. and Karlson, P. (1965). Vergleichende spektrometrische Untersichungen an a,Pungesattigten Steroidketonen. Chem. Ber. 98: 2361-2375.

39.

Dayal, B., Salen, G., Tint, G. S., Toome, V., Shefer, S., and Mosbach, E.H. (1978). Absolute configuration of pentahydroxy bile alcohols excreted by patients with cerebrotendinous xanthomatosis: a circular dichroism study. /. Lipid Res. 19: 187-190.

40.

Dayal, B., Tint, G. S., Shefer, S., and Salen, G. (1979). Configurational assignment of 5p-cholestane-3a,7a,12a,23,24-pentol excreted by patients with cerebrotendinous xanthomatosis (a circular dichroism study). 5rero/c/5.33: 327-338.

41.

Nakanishi, K. and Dillon, J. (1971). A simple method for determining the the chirality of cyclic a-glycols with Pr(DPM)3 and Eu(DPM)3. /. Am. Chem. 5cx:.93: 4058-4060.

42.

Nakanishi, K. and Dillon, J.(1975). Absolute configurational studies of vicinal glycols and amino alcohols. II. With Pr(DPM)3. J. Am. Chem. Soc. 97: 5417-5422.

43.

Partridge, J.J,, Toome, V. and Uskokovic', M.R. (1976). A stereospecific synthesis of the 24(R),25-dihydroxy-cholesterol side chain. /. Am. Chem. 5oc. 98: 3739-3741.

44.

Dayal, B., Salen, G., Toome, V., and Tint, G.S. (1986). Configuration at C-25 in 5P-cholestane-3a,7a,12a,25,26-pentol excreted by patients with cerebrotendinous xanthomatosis: circular dichroism and ^^C-NMR studies. J. Lipid Res. 21: 1328-1332.

45.

Partridge, J. J., Shivey, S.J., Chada, N.K., Baggiolini, E.G., Blount, J.F. and Uskokovic', M.R. (1981). Synthesis and structure proof of a vitamin D3 metabolite, 25(S),26-dihydroxycholecalciferol. /. Am. Chem. Soc. 103: 1253-1255 (and references cited therein).

46.

Dayal, B., Bagan, E., Tint, G. S., Shefer, S., and Salen, G. (1979). Preparation of [3P-3H] bile acids and bile alcohols. Steroids. 34: 259271.

229

47.

Salen, G., Shefer, S., Cheng, F.W., Dayal, B., Batta, A.K., and Tint, G.S. (1979). Cholic acid biosynthesis: The enzymatic defect in cerebrotendinous xanthomatosis. /. Clin. Invest 63: 38-44.

48.

Hoshita, T., Yasuhara, M., Une, M., Kibe, A., Itogu, E., Kito, S., and Kuramoto, T. (1980). Occurrence of bile alcohol glucuronides in bile of patients with cerebrotendinous xanthomatosis. /. Lipid Res. 21: 10151021,

49.

Wolthers, B. G., Volmer, M., van der Molen, J., Koopman, B. J., de Jager, A. E. J., and Waterreus ,R. J. (1983). Diagnosis of cerebrotendinous xanthomatosis (CTX) and effect of chenodeoxycholic acid therapy by analysis of urine using capillary gas chromatography. Clin. Chim. Acta. 131: 53-65.

50.

Shimazu, K., Kuwabara, M., Yoshi, M., Kihira, K., Takeuchi, H., Nakano, I., Ozawa, S., Onuki, M., Hatta, Y., and Hoshita, T. (1986). Bile alcohol profiles in bile, urine, and feces of a patient w i t h cerebrotendinous xanthomatosis. J. Biochem. (Tokyo) 99: 477-483.

51.

Batta, A. K., Shefer, S., Batta, M., and Salen, G. (1985). Effect of chenodeoxycholic acid on biliary and urinary bile acids and bile alcohols in cerebrotendinous xanthomatosis; monitoring by high performance liquid chromatography. /. Lipid Res. 26: 690-698.

52.

Batta, A. K., Salen, G., Shefer, S., Tint, G. S., and Batta, M. (1987). Increased plasma bile alcohol glucuronides in patients with cerebrotendinous xanthomatosis: effect of chenodeoxycholic acid. /. Lipid Res.2%: 1006-1012.

53.

Dayal, B. and Salen G. FAB-MS and Proton NMR analysis of 5Pcholestane-3a,7a, 12a,24R,25-pentoI-3-0-P-D glucuronide. Presented in part at the 200th American Chemical Society National meeting (Carbohydrate Division) held in Washington D.C. August 24-29, 1990.

54.

Dayal. B., Salen, G., Tint, G. S., Shefer, S. and Benz, S.W. (1990). Use of positive ion fast atom bombardment mass spectrometry for rapid identification of a bile alcohol glucuronide isolated from cerebrotendinous xanthomatosis patients. Steroids. 55: 74-78.

55.

Egestad, B., Pettersson, P., Skrede, S., and Sjovall, J. (1985). Fast atom bombardment mass spectrometry in the diagnosis of cerebrotendinous xanthomatosis. Scand. J. Clin. Lab. Invest. 45: 443-446.

230

56.

Salen, G., Shefer, S., Chen, F.W., Dayal, B., Batta, A. K. and Tint, G. S. (1985). Biosynthesis of bile acids in cerebrotendinous xanthomatosis. Relationship of bile acid pool sizes and synthesis rates to hydroxylation at C-12, C-25, and C-26. /. Clin Invest 76: 744-751.

57.

Danielson, H. and Sjovall, J. (1975). Bile acid metabolism. Annu. Rev. Biochem. 44: 233-253.

58.

Gibbons, G. F., Mitropoulos, K. A. and Myant, N. B. (1982). Biochemistry of cholestrol. Amsterdam: Elsevier Biomedical Press, 189203.

59.

Dayal, B., Tint, G. S., Batta, A. K., Shefer, S., Salen, G., Bose, A. K. and Pramanik, B. N. (1983). Synthesis of 3a,7a,12a,25-tetrahyd^oxy-5Pcholestane-24-one, an intermediate in the 25-hydroxylation pathway of cholic acid biosynthesis from cholesterol. /. Lipid Res, 24: 208-210.

60.

Shefer, S., Salen, G., Cheng, F. W., Dayal, B., Batta, A.K., Tint, G.S., Bose, A.K. and Pramanik, B.N. (1982). Chemical lonization-Mass Spectrometric Approach to Structure Determination of an Intermediate in Bile Acid Biosynthesis. Anal Biochem, 121: 23-30.

61.

Duane, W.C, Bjorkhem, I., Hamilton, J.N. and Mueller, S.M. (1988). Quantitave Importance of the 25-Hydroxylation pathway for bile acid biosynthesis in the rat. Hepatology, 8:613-618.

62.

Duane, W.C, Pooler, P.A. and Hamilton., J.N. (1988). synthesis in man. / Clin, Invest 82: 82-85.

63.

Oftebro, H., Bjorkhem, I., Skrede, S., Screiner, A. and Pedersen, J.I. (1980). Cerebrotendinous xanthomatosis: a defect in mitochondrial 26hydroxylation required for normal biosynthesis of cholic acid. / Clin, Invest 65: 1418-1430.

64.

Bjorkhem, I., Fausa, O., Hopen, G., Oftebro, H., Pedersen, J. I., and Skrede, S. (1983). Role of the 26-hydroxylase in the biosynthesis of bile acids in the normal state and in cerebrotendinous xanthomatosis. /. Clin, Invest! I: 142-148.

65.

Skrede, S., Bjorkhem, I., Krittingen, E.A., Buchmann, M.S., Lie, S.O., East, C , and Grundy, S. (1986). Demonstration of 26-hydroxylation of C27-steroids in human skinfibroblastsand a deficiency of this activity in cerebrotendinous xanthomatosis. /. Clin. Invest 78: 729-735.

Bile acid

231

66.

Reshef, A., Meiner, V., Berginer, V. M., and Leitersdorf, E. (1994). Molecular genetics of cerebrotendinous xanthomathosis in Jews of North African origin. / Lipid Res. 35: 478-483.

67.

Leitersdorf, E., Reshef, A., Meiner, V., Levitzki, R., Schwartz, S. P., Dann, E. J., Berkman, N., Cali, J. J., Klapholz, L. and Berginer, V. M. (1993). Frameshift and splice-j-junction mutations in the sterol 27hydroxylase gene cause cerebrotendinous xanthomathosis in Jews of Moroccan origin., /. Clin. Invest 91: 2488-2496.

68.

Cali, J. J., Hesieh, C-L., Francke, U., and Russell, D. W. (1991). Mutations in the bile acid biosynthetic enzyme sterol 27-hydroxylase underlie cerebrotendinous xanthomathosis. /. Biol. Chem. 266: 77797783.

69.

Cali, J. J., Russell, D. W. (1991). Characterization of the human sterol 27-hydroxylase: a mitochondrial P-450 that catalyzes mutiple oxidation reactions in bile acid biosyntheis. /. Biol. Chem. 266: 1114-111%.

70.

Berginer, V. M., Salen, G., and Shefer, S.. (1984). Long-term treatment of cerebrotendinous xanthomatosis with chenodeoxycholic acid. N. Engl. J. Med. 3X1: 1649-1652.

This Page Intentionally Left Blank

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

233

Antifungal Sesquiterpene Dialdehydes from the Warburgia Plants and their Synergists Isao Kubo

1. INTRODUCTION We now can control many human pathogenic microorganisms with the antibiotics that are presently available. Nevertheless, the need for new antimicrobial agents still exists. For example, systematic infections caused by filamentous fungi, especially in patients with impaired host defense mechanisms, have become increasingly serious worldwide. Thus, various antifungal agents have been introduced, but control of many of the fungal diseases has not yet been achieved. Hence, in our continuing search for antimicrobial agents from tropical plants, an emphasis was placed on searching for antifungal agents against Candida albicans, one of the most important fungi responsible for human systemic infections (1). Tropical plants are exposed throughout the entire year to attack by various parasites such as bacteria, fungi, and insects. They are, therefore, confronted with harsh conditions for survival. This leads to efficient built-in defense mechanisms, and thus tropical plants offer a rich and intriguing source of secondary metabolites possessing attractive biologically active properties. There is no doubt that plants are still a good source for new antifungal agents (2). However, we now need to investigate them from a different point of view, since isolation and structure determination of biologically active principles are no longer a main part of the study. This is largely because of the recent rapid developments of various instruments. The plants used for this study were collected primarily based on information provided by medicine men in East Africa, mainly in Kenya and Tanzania (3). Botanically identified plants were extracted with methanol at ambient temperatures. The extracts were first tested for their antimicrobial activity against four representative microorganisms at 100 /xg/ml (4). The active extracts were then tested against more microorganisms. As a result, interestingly, the information gathered from medicine

234 men was found to be useful. The plants collected based on their information had a much higher probability of containing active extracts than those of the plants collected randomly (5). Thus, from 79 extracts, which included 72 species of plants distributed among 35 families, 40 extracts initially gave positive results indicative of antimicrobial activity at 100 ftg/ml against one or more of the microorganisms tested. The current study has already resulted in a number of publications. Noticeably, as a result, few phytochemicals showed activity against fungi, especially C. albicans. It would be timely to review the study so far accomplished. I would like to focus on a particular group of compounds in this chapter instead of writing a catalog paper. More specifically, antifungal sesquiterpene dialdehydes from the Warburgia trees are reviewed based on our own study. These sesquiterpene dialdehydes are the rare phytochemicals that exhibited potent antifungal activity against C. albicans. The aim of this review is to describe an example of what can be done after identification of biologically active principles. 2. INITIAL STUDIES I first came to know of the W, ugandensis tree because of its strong hot taste when I was on a tenure appointment at the International Centre of Insect Physiology and Ecology (ICIPE), located in Nairobi, Kenya. The genus Warburgia belonging to a small family, Canellaceae, is endemic to East Africa and consists of only two species, namely W, ugandensis and W, stuhlmannii. The methanol extract of the bark of W. ugendensis collected in the ICIPE ground was first found to exhibit potent insect antifeedant activity against the African armyworm, Spodoptera exempta (6). Based on fractionation guided by the assay three antifeedants were isolated from the bark, leaves and fruit of this tree (7,8). The active compounds were isolated after repeated various chromatographies and characterized as unique hot tasting sesquiterpene dialdehydes (9), warburganal (1) (10) and muzigadial (canellal) (2) (11,12), in addition to a known congener, polygodial (3). Most of the initial chemistry was done at Professor K. Nakanishi's lab together with his colleagues at Columbia University in New York (13). Incidentally, these antifeedant sesquiterpene dialdehydes received much attention by synthetic organic chemists and their various synthetic pathways have already been reported (14-16) including those of the optical active (-)-warburganal from I-abietic acid (17).

235

CHO

CHO

(1) R=OH (3) R=H

(2)

In addition, since the Warburgia plants are widely used in folk medicines in East Africa (18), their extracts were also submitted together with other plant extracts for various available pharmacological assays. Among the plant extracts tested, the two Warburgia species were found by Professor M. Taniguchi of Osaka City University to exhibit a broad antimicrobial spectrum, including against Candida utilis (4). This is how the project was initiated. The initial biochemical investigation of the antifungal principles has been achieved in collaboration with Professor Taniguchi and his co-workers (1923) . 3 ISOLATION AND IDENTIFICATION OF ANTIFUNGAL PRINCIPLES A bioassay directed fractionation of the n-hexane extract from the bark of IV. ugandensis

using Bacillus

subtilis

and

Saccharomyces cerevisiae has resulted in the isolation of three active sesquiterpene dialdehydes, warburganal (1), muzigadial (2) and polygodial (3) which were previously identified as the insect antifeedants from the same sources. Their structures were identified by means of spectroscopic methods. Incidentally, the Warburgia barks were originally extracted with methanol. However, methanol, at least partly if not totally, inactivated the antifungal dialdehyde compounds through their acetal (or hemiacetal) formation. Therefore, the use of alcohols was avoided throughout the isolation procedure. This also limits their practical application. Polygodial was first isolated as a hot tasting substance from the sprout of Polygonum hydropiper (Polygonaceae) (24), which has been used in folk medicines and food spices in several Asian countries including Japan and Vietnam. For example, the sprout of P. hydropiper is a well known relish for "sashimi" in Japan (25,26). Subsequently, warburganal was also isolated from the same source in minute amounts (27). In addition to these three antifungal sesquiterpene dialdehydes (1-3), a number of

236 congeners such as mukaadial (4) (28), ugandensidial (cinnamodial) (5) (29), epipolygodial (6) (13) as well as confertifolin (7), 9Of-hydroxycinnamolide (8) (30), cinnamosnolide (9) (13), colorata-4(13),8-dienolide (10), and bemadienolide (11) (13) were also isolated from various parts of W. ugandensis, but none of them exhibited any antifungal activity up to 100 /xg/ml. Their structures were all identified by means of spectroscopic methods, particularly by NMR spectroscopy. All the sesquiterpenoids isolated (1-11) are considered to be oxidation products of the drimane skeleton. The same sesquiterpenoids were also isolated from the bark of W. stuhlmannii (10,11,31) as well as several other plants (30-32). This result is in general agreement with those reported recently (33). It should be noted that each dialdehyde compound gives the corresponding lactones by treatment of 2i\r-HCl and IN-NaOH. For example, the alkaline treatment with warburganal (1) at room temperature gave 9a-hydroxycinnamolide (8) through intramolecular Cannizzaro reaction, and the acid treatment yielded futronolide (12). Similarly, by alkaline treatment of ugandensidial (cinnamodial) (5) yielded the lactone (13) which was converted to cinnamosnolide (9) by acetylation (28). Nevertheless, the lactones were found to be no longer active. 4. ANTIFUNGAL ACTIVITY Besides warburganal (1), muzigadial (2) and polygodial (3), none of the other sesquiterpenoids (4-11) isolated, showed any noticeable antifungal activity up to 100 /xg/ml, although they were all isolated from the active fractions. Nevertheless, most of the purified sesquiterpenoids were tested against eleven additional microorganisms. The results are listed in Table 1. Warburganal, muzigadial and polygodial exhibited activity against all fungi tested (5). In particular, they were highly

active against Candida utilis, Saccharomyces cerevisiae, Penicillium chryaogenuuif Hansenula anomala, and Sclerotinia lihertiana. Based on this finding, the activity of polygodial (3) and warburganal (1) was also tested against Candida albicans, As expected, both were found to be highly active against this important pathogenic fungus; their minimum inhibitory concentrations (MICs) were 3.13 and 6.25 /xg/ml, respectively (34). Among the three antifungal sesquiterpene dialdehydes (1-3), structurally the simplest polygodial exhibited the most potent activity (19). It was 2-8 times more active than warburganal and

237

CHO CHO

CHO

(8)

CHO

(10)

R=H

(9) R=OAc

OH

(11)

"OH

(12) CHO

OH" (8)

238 muzigadial (2) against all species of fungi tested. Noticeably, the potency of polygodial against these fungi is almost comparable to that of amphotericin B, which is one of the most potent antibiotics known against filamentous fungi (21), although its high toxicity limits its wide use. Therefore, polygodial may be potent enough to be considered for practical application. Interestingly, by contrast to polygodial, its C-9 epimer, epipolygodial (6) , did not show any activity up to 100 /xg/ml (19). In addition to 6, mukaadial (4) possessing two additinal hydroxyl groups at C-6 and C-9 to polygodial, did not exhibit any activity up to 100 /xg/ml. Table 1. Antimicrobial activity of the Warburgia (6,35). MIC

Microorganisms tested 1 >100 Staphylococcus aureus >100 Bacillus subtilis Micrococcus luteus >100 Escherichia coli >100 Proteus vulgaris >100 Pseudomonas aeruginosa >100 Saccharomyces cerevisiae 3.,13 Schizosaccharomyces pombe' 12.,5 Hansenula anomala 12.,5 Candida utilis 3.,13 C. albicans 6.,25 3.,13 Sclerotinia libertiana Mucor mucedo 25 Rhizopus chinensis 100 Aspergillus niger 50 50 Penicillium crustsum

sesquiterpenoids

(/xg/nil)

2

3

>100 >100 >100 >100 >100 >100 1.,56 25 25 3.,13

>100 >100 >100 >100 >100 >100 0..78 6..25 1..56 1..56 3..13 1.,56 6.,25 12.,5 25 25

-

3.,13 25 100 50 50

4 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100

-

100 >100 >100 >100 >100

6 >100 >100 >100 >100 >100 >100 >100 >100 >100 >100

-

>100 >100 >100 >100 >100

-, Not tested. A large number of phytochemicals have been isolated as antifungal agents (2,35). These phytochemicals are biodegradable and, more importantly, they are renewable. The efficient utilization of such renewable natural products is becoming increasingly important worldwide. However, their activity is usually not potent enough to be considered for practical application, even though each of the antifungal phytochemicals may play an important role in the defense against fungal attacks in living plants. This is always a dilemma when the antimicrobial activities of phytochemicals are considered. Hence, studies to enhance their biological activity are needed. This strategy seems to be a most promising approach for efficient utilization of

239 renewable natural products. Therefore, an attempt to enhance antimicrobial activity of some purified antimicrobial agents was made. As a model experiment, polygodial was first combined with several antibiotics such as actinomycin D and rifampicin. As a result, polygodial significantly enhanced the antifungal activity of these antibiotics against C. utilis and S. cerevisiae, but not vice versa (19,21,22,36). Polygodial also synergyzed the antifungal activity of maesanin (14) against specifically C. utilis (37). Maesanin was isolated from the fruit of an East African medicinal plant Maesa lanceolata (Myrsinaceae) (38). The reason for these combination effects seems to be caused by an increase in the permeability of the plasma membrane to the

antibiotics brought about by polygodial (39,40). Thus, cells of S. cerevisiae were treated with polygodial; ultrastructural changes in the cell membrane followed (21), as seen Figure lb. These morphological data suggest that the primary site of action of polygodial is at the plasma membrane with the simultaneous involvement of organelle disorganization followed by the fatal loss of cellular constituents such as proteins and polysaccharides, as illustrated in Figure 2 (36). 5. COMBINATION EFFECTS In our continuing search for antimicrobial agents from tropical plants, several phenylpropanoids, for example, anethole (15) from the seeds of Pimpinella anisum (Umbelliferae) (41), and eugenol (16), methyleugenol (17) and safrole (18) in addition to anethole from the seeds of Licaria puchuri-major (Lauraceae) (42), were isolated as antimicrobial principles in rather large quantities. All exhibited a moderate but broad spectrum. Their MICs ranging from 200 to 400 /xg/ml are not potent enough to warrant further studies. However, it is still worthwhile to investigate the possibility of their practical use as antifungal agents, since they were all isolated from various food spices which have long been consumed by many people. Hence, they were first combined with a sublethal amount of polygodial to enhance

240

Figure 1. a. Section of untreated control cell of S, cerevisiae. b. Section of S. cerevisiae cell treated with 50 /xg/ml of polygodial for 10 min. CW, cell wall; PM, plasma membrane (cell membrane); N, nucleus; M, mitochondrion; V, vacuole.

241

Incubation time (hr)

Incubat ion time (hr) Figure 2. Leakage of a f^^i-

242 their antifungal activity against several fungi such as C. albicans

and S.

Pityrosporum

cerevisiae

ovale.

as well as a dermatomycotic fungus

Unexpectedly, polygodial did not synergize

the antifungal activity of any of these phenylpropanoids tested. In contrast, the antifungal activity of polygodial was significantly increased when combined with one of the phenylpropanoid compounds. Thus, noticeably, a dramatic increase in the antifungal activity of polygodial occurred when it was combined with a sublethal amount of anethole. The activity of polygodial against C. albicans

and S,

cerevisiae

was increased

32- and 64-fold, respectively. In other words, the MIC against C. albicans

was lowered from 3.13 to 0.098 /xg/ml and in the case of

5. cerevisiae,

from 1.56 to 0.024 /ig/ml, when polygodial was

combined with 100 /xg/ml of anethole (MMIC for both C.

and S, cerevisiae)

albicans

(41) .

6. FUNGICIDAL ACTIVITY The above combination effects are discussed based on the MIC. However, the MIC, which is obtained after a 48 hr incubation, does not fully characterize the antifungal activity of the combination. For example, it was not clear if the combination of polygodial and anethole had fungicidal or fungistatic activity. Hence, the viable count method to analyze the growth curve of C. albicans was used in order to obtain the minimum fungicidal concentration (MFC). The growth curve of C. albicans in the presence of polygodial and anethole is illustrated in Figure 3 (43). Both polygodial and anethole exhibited fungicidal activity against C. albicans at 3.13 and 200 /xg/ml; however they no longer showed any fungicidal activity at 1.56 and 100 /xg/ml. Thus, C. albicans was not viable with 18 and 42 hrs at 3.16 /xg/ml of polygodial and 200 /xg/ml of anethole. The lethal concentration of both polygodial and anethole against C. albicans was the same as their MIC values. The combination of polygodial and anethole was found to possess fungicidal rather than fungistatic activity. Fungicidal activity against C. albicans with the combination of polygodial and anethole is shown in Figure 3. When 100 /xg/ml of anethole (^IC for C. albicans) was combined with more than 0.098 /xg/ml of polygodial, C. albicans was not viable within 6 or 12 hr. Thus, the fungicidal activity of polygodial against C. albicans was increased 32-fold by anethole. By contrast, the fungicidal activity of anethole was enhanced only 4-fold by polygodial.

o

o 41!

0)

O

O PU

•H O rH O 04

o Xi

CO

CD

^

CNJ

Xui jad SXI90 asqmnu Boq

U Xi

%

•H V

a o

u

•H JJ cd ,Q :3

0

IW

c

0) V 0)

ro
ft (1)

,c|

c PJ 0

4J

CO

m

4J

•H •H

•i

S-H

o u u

-q 0 M CO -H u• a)

O

X! M-l XJ

iH

Tl CQ C 0) rd

t

:^-Ha u Tl

>1

0 ^4J tD ^ Or-4 M 0 O ft

, ^ r-\

ro (U

4J

(U o M43

ns

g>
243

244 Following this discovery, warburganal (1) was also examined in combination with anethole to see if anethole had the same enhancing activity (34). As expected, anethole also significantly increased the activity of warburganal against both C. albicans and S, cerevisiae. In this combination, the activity against C. albicans and S, cerevisiae was enhanced 32- and 256-fold when warburganal was combined with 100 /xg/ml of anethole. The MIC of warburganal against C. albicans was lowered from 6.25 to 0.20 /xg/ml and that against S. cerevisiae from 6.25 to 0.024 /xg/ml, respectively. Anethole also enhanced the activity of polygodial and warburganal against P. ovale but not as much as against C. albicans and S, cerevisiae. The MICs were reduced only from 50 to 6.25 /xg/ml of polygodial, and from 25 to 3.13 /xg/ml of warburganal in combination with 50 /xg/ml anethole (MMIC for P. ovale). Similarly, the antifungal activity of polygodial was significantly increased when combined with sub-inhibitory concentrations of methyleugenol (17) and safrole (18), while eugenol (16) did not seem to have any meaningful enhancing activity, as shown in Table 2. Table 2. Antifungal activity of polygodial (3) in combination with MMIC of phenylpropanoids (15-18) . MIC (/xg/ml)

Compounds combined

C. None (polygodial alone) Anethole (15) Eugenol (16) Methyleugenol (17) Safrole (18)

albicans 6.25 0.20 3.13 1.56 0.78

S,

cerevisiae 3.13 0.049 1.56 0.78 0.39

P.

ovale 50 6.25 25 0.39 6.25

Noticeably, methyleugenol showed a 128-fold enhancement of the activity of polygodial against P. ovale. The MIC was lowered from 50 to 0.39 /xg/ml. Phenolic compounds that do not possess a free phenolic group, such as anethole, safrole, and methyleugenol, seem to exhibit more enhancing effects with polygodial (3) than phenolic compounds that have a free phenolic group, such as eugenol. Anethole was very effective in enhancing the antifungal activity of polygodial and warburganal against C.

albicans,

C. utilis

and S. cerevisiae,

most effective against P. ovale

and methyleugenol was the

(41).

In addition, anethole was also combined with other antifungal agents, such as amphotericin B, to see if it had the same enhancing activity. This combination was based on the fact

245 that the mode of action of amphotericin B is also known to damage the plasma membrane by interacting with sterols (40) in fungal 9CH3

OR

,—Q

H3C0>. ^^

0;

(15)

(16) R=H (18) (17) R=CH3 cell membranes (39) . However, anethole did not enhance the antifungal activity of amphotericin B against C. albicans and S. cerevisiae. The antifungal activity of amphotericin B was somewhat antagonized by anethole as shown in Table 3 (41). Table 3. Antifungal activity of amphotericin B in combination with anethole (15). Concetration of anethole ( 1 5 ) iixg/ml) 0 6.25 12.5 25 50 100

MIC

C.

albicans 0.78 1.56 3.13 12.5 6.25 0.20

S,

(/xg/ml) cerevisiae 0.78 1.56 3.13 6.25 12.5 3.13

P.

ovale 3.13 3.13 0.78 0.78 0.39

-

Increasing the amount of anethole increased its MICs. More specifically, the activity of amphotericin B against C. albicans and S, cerevisiae was decreased 16-fold when it was combined with 25 and 50 ^g/ml of anethole, respectively. Thus, the MICs were increased against both fungi from 0.78 to 12.5 /xg/ml. In contrast, anethole did increase the antifungal activity of amphotericin B against P. ovale 8-fold. Thus, the MIC was lowered from 3.13 to 0.39 /xg/ml when amphotericin B was combined with 50 /ig/ml anethole. These results do not seem to indicate that a chemical reaction between anethole and amphotericin B occurred prior to the assay. The enhancing activity of anethole depends on the species of fungi being tested and antifungal agents being combined. CONCLUSION The above observations suggest polygodial itself may be a

246 promising antifungal agent because of its effectiveness. Moreover, the combination of more than two compounds may be superior to the use of a single antimicrobial compound in order to enhance and broaden the total biological activity. More importantly, it may hinder the development of resistant mechanisms in microorganisms. For the otherwise healthy person, fungal infections are more of a nuisance than a health threat and are normally kept in check by a strong immune system and by innocuous bacteria of the throat and gut. However, when outside forces such as cancer chemotherapy or heavy doses of antibiotics derange the body's natural defenses, fungal populations can sharply increase and cause serious health problems. Substances that enable physicians to use lower, safer antibiotic dosages to kill the fungi would be a useful addition to the therapeutic arsenal. Further studies of increased potency due to polygodial as well as warburganal and muzigadial may pave the way for the development of new and extraordinarily powerful antifungal agents, especially against one of the most important filamentous fingi, C. albicans. Acknow1edgemen

t

The work has involved a number of scientists. I am greatly indebted to my colleagues cited in the references.

247 REFERENCES 1. Y. Fukuzawa and K. Kagawa, Cooperation of humoral and cellmediated immunities against experimental candidiasis and cryptococcosis, in Filamentous Microorganisms (T. Arai, ed) Japan Scientific Societies, Tokyo, (1985) 247-253. 2. L. A. Mitscher, S. Drake, S. R. Gollapudi and S. K. Okwute, A modern look at folkloric use of antiinfective agents, J. Nat. Prod, 50 (1987) 1025-1040. 3. I. Kubo, Pharmacies in the Jungle, in Science Year 1982, World Book-Childcraft International, Chicago, (1981) 126-137. 4. M. Taniguchi, A. Chapya, I. Kubo and K. Nakanishi, Screening of East African plants for antimicrobial activity I, Chem. Pharm. Bull., 26 (1978) 2910-2913. 5. M. Taniguchi and I. Kubo, Ethnobotanical drug discovery based on medicine men's trials in African savanna: Screening of East African plants for antimicrobial activity II, J". Nat. Prod, in press. 6. I. Kubo, Screening techniques for plant-insect interactions, in Methods in Plant Biochemistry Vol. 6 (K. Hosttetmann, ed) Academic Press, London (1991) 179-193. 7. I. Kubo and K. Nakanishi, Insect antifeedants and repellents from African plants, in ACS Symposium Series 62, The Chemical Basis for Host Plant Resistance to Pest (P. A. Hedin, ed), American Chemical Society, Washington, (1977) 165-178. 8. I. Kubo, Insect control agents from tropical plants, in i?ecent Advances in Phytochemistry, Vol. 27, Phytochemical Potential of Tropical Plants (K. R. Downum, J. T. Romeo and H. A. Stafford, eds.), Plenum, New York, (1993) 133-151. 9. I. Kubo and I. Ganjian, Insect antifeedant terpenes, hottasting to humans, Experientia, 37 (1981) 1063-1064. 10. I. Kubo, Y. W. Lee, M. Pettei, F. Pilkiewicz and K. Nakanishi. Potent army worm antifeedants from the East African Warburgia plants. J. Chem. Soc, Chem. Commun. , 10131014. 11. I. Kubo, I. Miura, M. Pettei, Y. W. Lee, F. Pilkiewicz and K. Nakanishi, Muzigadial and warburganal, potent antifungal antiyeast and African armyworm antifeedant agents. Tetrahedron Lett., (1977) 4553-4556. 12. F. S. El-Feraly, T. McPhail and K. D. Onan, X-ray crystal structure of canellal, a novel antimicrobial sesquiterpene from Canella winterana, J. Chem. Soc, Chem. Commun., (1978) 75-76. 13. K. Nakanishi and I. Kubo, Studies on warburganal, muzigadial and related compounds, Israel J. Chem., 16 (1978) 28-31. 14. S. P. Tanis and K. Nakanishi, Stereospecific total synthesis of (±) warburganal and related compounds, J". Amer. Chem. Soc, 101 (1979) 4399-4400. 15. M. Jalani-Naini, D. Guillerm and J. Y. Lallemand, Synthese totale du (±) polygodial, de la drimenine et de composes apparent^s a jonction de cycle cis et trans. Tetrahedron, 39 (1983) 749-758 and references cited therein. 16. K. Mori and H. Watanabe, Synthesis of both the enantiomers of polygodial, an insect antifeedant sesquiterpene. Tetrahedron, 42 (1986) 273-281. 17. H. Okawara, H. Nakai and M. Ohno, Synthesis of (-)warburganal and 4a-methoxycarbonyl congener from 1-abietic acid, Tetrahedron Lett., 23 (1982) 1087-1090. 18. J. O. Kokwaro, Medicinal Plants of East Africa, East African Literature Bureau, Nairobi, (1976) 45.

248 19. M. Taniguchi, T. Adachi, S. Oi, A. Kimura, S. Katsumura, S. Isoe and I. Kubo, Structure-activity relationship of the Warburgia sesquiterpene dialdehydes, Agric, Biol. Chem, 48 (1984) 73-78. 20. M. Taniguchi, T. Adachi, H. Haraguchi, S. Oi and I. Kubo, Physiological activity of warburganal and its reactivity with sulfhydryl groups, J. Biochem, 84 (1983) 149-154. 21. M. Taniguchi, Y. Yano, E. Tada, K. Ikenishi, S. Oi, H. Haraguchi, K. Hashimoto and I. Kubo, Mode of action of polygodial, an antifungal sesquiterpene dialdehyde, Agric, Biol, Chem,, 52 (1988) 1409-1414. 22. M. Taniguchi, Y. Yano, K. Motoba, S. Oi, H. Haraguchi, K. Hashimoto and I. Kubo, Polygodial-induced sensitivity to rifampicin and actinomycin D in Saccharomyces cerevisiae, Agric, Biol, Chem,, 52 (1988) 1881-1883. 23. Y. Yano, M. Taniguchi, T. Tanaka, S. Oi and I. Kubo, Protective effects of Ca^* on cell membrane damage by polygodial in Saccharomyces cerevisiae, Agric, Biol, Chem,, 55 (1991) 603-604. 24. C. S. Barnes and J. W. Loder, The structure of polygodial: a new sesquiterpene dialdehyde from Polygonum hydropiper L., Aust, J, Chem,, 15 (1962) 322-327. 25. A. Ohsuka, The structure of tadeonal and isotadeonal components of Polygonum hydropiper L, Nippon Kagaku Zasshi, 84 (1963) 748-752. 26. I. Kubo, Synergistic effect of polygodial on antifungal activity. Drug News Persp, , 2 (1989) 292-295. 27. Y. Fukuyama, T. Sato, I. Miura and Y. Asaka, Drimane-type sesqui- and norsesquiterpenoids from Polygonum hydropiper, Phytochemistry, 24 (1985) 1521-1524. 28. I. Kubo, T. Matsumoto, A. B. Kakooko and N. K. Mubiru, Structure of mukaadial, a molluscicide from the Warburgia plants, Chem. Lett,, (1983) 979-980. 29. C. J. W. Brooks and G. H. Draffan, Sesquiterpenoids of Warburgia species-II, Ugandensolide and ugandensidial (cinnanodial), Tetrahedron, 25 (1969) 2887-2898. 30. D. Kioy, A. I. Gray and P. G. Waterman, Further drimane sesquiterpenes from the stem bark of CaneiJa winterana, J, Nat, Prod., 52 (1989) 174-177. 31. R. F. McCallion, A. L. J. Cole, J. R. L. Walker, J. W. Blunt and M. H. G. Munro, Antibiolic compounds from New Zealand plants II. Polygodial, an anti-Candida agent from Pseudowintera colorata, Planta Medica, 44 (1982) 134-138. 32 M. S. Al-Said, S. M. El-Khawaja, F. S. El-Feraly and C. H. Hufford, 9-Deoxy drimane sesquiterpenes from Canella winterana, Phytochemistry, 29 (1990) 975-977. 33. D. Kioy, A. I. Gray and P. G. Waterman, A comparative study of the stem-bark drimane sesquiterpenes and leaf volatile oils of Warburgia ugandensis and W. stuhlmannii, Phytochemistry, 29 (1990) 3535-3538. 34. I. Kubo and M. Himejima, Potentiation of antifungal activity of sesquiterpene dialdehydes against Candida albicans and two other microorganisms, Experientia, 48 (1992) 1162-1164. 35. L. A. Mitscher, R. P. Leu, M. S. Bathala, W. N. Wu, J. L. Beal and R. White, Antimicrobial agents from higher plants I. Introduction, rationale and methodology, Lloydia, 35 (1972) 157-166. 36. I. Kubo and M. Taniguchi, Polygodial, an antifungal potentiator, J". Nat. Prod., 51 (1988) 22-29.

249 37. M. Taniguchi, Y. Yano, E. Tada, T. Tanaka, S. Oi, H. Haraguchi, K. Hashimoto and I. Kubo, Potentiation by polygodial of the respiration inhibiting activity of maesanin, Agric, Biol, Chem., 53 (1989) 1525-1530. 38. I. Kubo, T. Kamikawa and I. Miura, Isolation, structure and synthesis of maesanin, a host defense stimulant from an African medicinal plant Maesa lanceolata, Tetrahedron Lett, (1983) 3825-3828. 39. S. C. Kinsky, Antibiotics interaction with model membranes, Annu. Rev, Pharmacol,, 10 (1970) 119-142. 40. J. M. T. Hamilton-Miller, Fungal sterols and the mode of action of the polyene antibiotics, Adv, Appl. Microbiol., 17 (1974) 109-134. 41. I. Kubo and M. Himejima, Anethole, a synergist of polygodial against filamentous microorganisms, J". Agric, Food Chem,, 39, (1991) 2290-2292. 42. M. Himejima and I. Kubo, Antimicrobial agents from Licaria puchuri-major and their synergistic effect with polygodial, J, Nat, Prod,, 55 (1992) 620-625. 43. M. Himejima and I. Kubo, Fungicidal activity of polygodial in combination with anethole and indole against Candida albicans, J, Agric, Food Chem., in press.

This Page Intentionally Left Blank

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

251

Determination of Relative and Absolute Configuration in the Annonaceous Acetogenins Elizabeth A. Ramirez and Thomas R. Hoye

I. INTRODUCTION The Annonaceae is a large group of trees and shrubs, found mostly in tropical and subtropical regions. There Is wide botanical diversity within this class, including 120 genera and over 2000 species. Although these plants have long been used in various folk remedies, It is only recently that the chemical source of these medicinal properties has been investigated. A wide variety of natural products has been isolated from the Annonaceae.^ However, the report in 1982 of the Isolation of the acetogenin uvaricin (1) from Uvaria accuminata^ was the first example of what is now a large and growing series of acetogenins found in these sources. More than one-hundred examples have been reported in the intervening decade.3.4 That these compounds often exhibit remarkably potent cytotoxic and other biological activity has fueled interest in this class.

2 (Qigantecin)

Figure 1. Examples of the Tetrahydrofuranyl Structure in Annonaceous Acetogenins: an Adjacent bis-THF [Uvaricin (1)], a non-Adjacent bis-THF [Gigantecin (2)], and a mono-THF [Annonacin (3)].

252 The Annonaceous acetogenins have some common structural features. Most of the compounds reported so far contain one or two tetrahydrofuran (THF) rings situated along a long, unbranched aliphatic chain. The only reported exceptions to this structure, molecules containing appropriately located epoxides and/or alkenes, can be construed as biogenetic precursors to the THF backbone."* One additional anomaly, muricatacin.s can be envisioned as arising by oxidative cleavage of monotetrahydrofuran acetogenins and has been suggested as a product of plant catabolism.6 Of those acetogenins containing two THF rings, these rings can be either adjacent (i.e., directly linked, cf. 1) or separated by a four-carbon chain [cf., gigantecin (2)]7 (Figure 1). Also, one end of the aliphatic chain is invariably terminated by a ylactone, either a,p-unsaturated [cf. 1- 3, annonin I (4),8 and bullatacin (5)^] or saturated with a hydroxyl group at 0(33) as in itrabin (7)^^ In some cases, this functionality is rearranged to an acetonylbutanolide with an oxygen at 0(4) [cf., bullatacinone (6),^ Figure 2]. Hydroxyl groups are nearly always found adjacent to the THF rings; various other hydroxyls may be scattered along the carbon chain, with the most common location at 0(4) (cf., 5), two carbons removed from the lactone moiety.

Figure 2. Examples of Various Lactone Moieties in Annonaceous Acetogenins: an a, p-Unsaturated-y-lactone [Annonin I (4)], a 0(4)-Hydroxylated a, pUnsaturated-y-lactone [Bullatacin (5)], a Rearranged Acetonyl Butanolide [Bullatacinone (6)], and a Saturated p-Hydroxy-y-lactone [Itrabin (7)].

253 The constitutional stmctural features just described were deduced by classical methods. Both ""H and '•^C NMR spectroscopy have proven valuable in unraveling these complex structures, and Interpretation of mass spectrometric fragmentation patterns is often used to pinpoint the location of substituents along the carbon chains. In recent years, two-dimensional NMR techniques have provided even more structural detail, and spectral trends among the many known compounds has made determination of connectivity a relatively straightforward task."* Determination of the stereochemical features within this new class of molecules, however, is a much less simple matter. Their waxy, amorphous, or microcrystalline nature has so far made them unsuitable for direct X-ray diffraction studies. However, given the biological properties of the Annonaceous acetogenins and their analogs, their relative and absolute configuration is an important issue. Several members of this class, possessing the same connectivity but all diastereomeric in nature, have been found to have significantly different bioactivltles,^.^ indicating that their mode of action is configuration dependent. Currently many generally applicable methods for determining configuration are available, and the three-dimensional structure of these molecules is no longer a complete mystery. Before these methods were available, however, structure determination resulted in the assignment of only a handful of stereochemical attributes In a few individual compounds.

HQ^

2)H202

OH S-Lactic Acid

Figure 3. Degradation of Uvaricin for Determination of Configuration at C(36). The first stereochemical feature reported for the Annonaceous acetogenins was the absolute configuration of C(36) in uvaricin (1 ).'•'' Uvaricin was subjected to degradation by ozonolysis with an oxidative workup (Figure 3), yielding, among other products, lactic acid. The mixture, after derivatlzation with CF3CH2N=C=0, was subjected to gas chromatography on a chiral column, and the degradation product was thus determined to be S-lactic acid. It therefore follows that C(36), from which this stereogenic center derived, is also S. Although the acetogenins presumably arise by similar blosynthetic pathways, In the absence of similar degradation experiments [or alternative methods for

254 deducing C(36) configuration] it is dangerous to conclude that all acetogenins share this configuration.3 Although, as previously mentioned, the Annonaceous acetogenins have thus far proven largely unsuitable for X-ray crystallographic studies, stmctures have been reported for two acetogenin derivatives. Pettit et al.12 conducted X-ray crystallographic studies on the 15-p-bromophenylurethane derivative 9 of rolliniastatin I (8). Later, Born et al.6 converted annonin I (4) to the saturated, crystalline potassium salt 10 (Figure 4) and successfully determined Its X-ray structure, providing the connectivity and relative configuration of this molecule. Although to date no other Annonaceous acetogenins have been similarly studied, the results of these two experiments have proven indispensable for validation of other methods for deducing stereochemical features. Neither of these studies provided the absolute configuration of the acetogenin.

O,

OK*

HO 15

MS

b 4 (Annonin I)

/. H2> Pd/C

'

2. KOH. CH^H

b

b b

OH 24

[24

HO

Figure 4. Two Acetogenin Derivatives, 9 and 10, with Single Crystal, X-ray Structures.

With an ever-increasing number of Annonaceous acetogenins being reported, there have been many efforts to develop methods for elucidation of configuration that are more generally applicable. Table 1 shows a time-line chronicling the major contributions that address various stereochemical issues in these molecules. Beginning with our report in 1986 describing the relative configuration of the bis-THF portion of uvaricin,i3 research has yielded many complementary methods for

255

determining configuration within these natural products. This chapter presents an overview of these methods (through 1993). The classification/labeling scheme for the structural subunits of the Annonaceous acetogenins proposed^ and later expanded^ by McLaughlin will be used in this chapter. 1"* The terminal lactone moiety is labeled "A", and this subunit is designated A1, A2, A3, A4, or A5 depending on Its structure (Figure 5). For example, the a,punsaturated lactone without other functionality on the carbon chain, such as found in uvaricin (1), is known as A1. The letter "B" refers to the THF subunit, labeled B1, 82, 83, or 84 ; for example, 83 refers to the non adjacent bis-tetrahydrofuranyl subunit. Other, miscellaneous functionality is represented by the letter "C": hydroxyl (C1), carbonyl (C2), acetate ester (C3), vicinal diol (C4), epoxide (C5), and olefin (C6). This scheme for organization and labeling of the acetogenins simplifies reference to molecules possessing common structural features; for example, instead of referring to "a monoTHF subunit bearing only one adjacent hydroxyl group", one can simply say "a 84 subunit." o Vo Rx.--V..>**,>/^

o OH V - o R^sX^s.^'SV^

i^ o—f o "vX^'^AsX^

o V-o n^^X-Ss..A^

o V-o ^^.y^-^^y'^Kf^

OH

A1

OH

A2

^—'

B1

A3

OH

OH

OH

A4

OH

OH

B2

B3

^—'

A5

OH

84

OH

Y

^r >" v^

Figure 5- Lactone (A#), Tetrahydrofuran (8#), and Miscellaneous (C#) Substructural Units Found in the Annonaceous Acetogenins.^

The strategy we have used to develop general methods for the determination of both relative and absolute configuration within various substructural units of the acetogenins consists of four parts, i) An appropriate set of model compounds is identified and synthesized so as to provide a complete set of diastereomers of unambiguously known

256 Table 1 . Time-line of Important Events, Including Major Advances in Determination 1982

First Annonaceous acetogenin, uvaricin, reported.

Cole et al. J. Org. Chem. 47,3151.

1985

C(36) of uvaricin determined to be S by degradation to lactic acid.

Cole et al. J. Nat. Prod. 48, 644.

1986

Model compounds of bis-THF structure prepared for NMR correlation studies; observation that ^H NMR shifts of acetate methyl groups correlate with relative configuration of C(15)/C(16) and C(23)/C(24) in uvaricin (later proven to be correct).

Hoye et al. Tetrahedron 42, 2855.

1987

""H NMR correlation studies suggest that bis-THF portion of uvaricin possesses threo/trans/threo/trans/erythro relative configuration among C{15)-C(24). Establishes general method for determining relative configuration.

Hoye et al. J. Am. Chem. Soc.

1987

Rolliniastatin I reported, and successful X-ray studies on p-bromophenylurethane derivative establish complete relative configuration.

Pettit et al. Can. J. Chem. 65, 1433.

1988

''H NMR correlation method further validated by comparison of rolliniastatin I NMR data with the now known relative configuration. Method made more quantitative, relying less on visual inspection. Relative configuration of bis-THF moiety of aslmicin verified.

Hoye et al. J. Org. Chem. 53, 5578.

1989

Rollinicin reported, containing a vicinal diol along Sneden et al. one carbon chain; relative configuration assigned as J. Nat. Prod. 52, 822. erythro based on cims fragmentation pattern.

1989

Bullatacin and bullatacinone reported; absolute configuration at C(4) assigned as S based on ORD spectral data (later proven to be incorrect); bullatacin successfully converted to bullatacinone, proving that they possess the same relative configurations along the adjacent bis-THF backbone.

McLaughlin et al. J. Nat. Prod 52, 463.

1990

Annonin I (squamocin) reported; successfully studied by X-ray analysis of a derivative. Previously described NMR correlation method gives results in accordance with structure determined by X-ray. Complementary 1H as well as i^C NMR correlation method developed to determine the configurational relationship between a THF ring and an adjacent hydroxylated carbon.

Born et al. Planta Med. 56,312.

109, 4402.

257

of Configuration, in the Development of Annonaceous Acetogenin Chemistry. 1991

"^H NMR-based method for assigning relative configuration (i.e., cis vs. trans) of 2-acetonyl-4alkylbutanolides.

Hoye et al. J. Org. Chem. 56, 5092.

1991

Synthesis of 15,16,19,20,23,24-^exep/-uvaricin (a diastereomer of the natural product) confirms relative configuration and establishes absolute configuration (via Mosher esters) of the first acetogenin, uvaricin.

Hoye et al. J. Am. Chem. 113,9369.

1992

Gigantetronenin and gigantrionenin reported, first Annonaceous acetogenins found to contain a double bond along one of the aliphatic chains. Configuration in both determined to be cis from ^H NMR coupling constants.

McLaughlin et al. J. Nat Prod. 55, 1655.

1992

Relative configuration of some mono-THF acetogenins confirmed by correlation of ^H and "^^C NMR chemical shifts with two mono-THF model compounds of known configuration.

Figaddre et al. Tetrahedron Lett. 33, 5749.

1992

Absolute configuration of C(4) in C(4)-hydroxylated compounds determined by NMR analysis of Mosher ester derivatives and comparison to model compounds. All configurations studied determined to be R at C(4). Method also applicable for determining the relative configuration between C(4) and C(36), as well as the absolute configurations of carbinol centers adjacent to THF rings.

McLaughlin, Hoye, et al. J. Am. Chem. Soc. f 74, 10203.

1993

General method for determining relative configuration of mono-THF acetogenins by correlation of 1H chemical shifts with mesitoylated model compounds.

Cassady et ai. Tetrahedron Lett. 34, 5847/5851

1993

Total synthesis of enf-bullatacin, the enantiomer of the bis-THF natural product, confirms absolute configuration of bullatacin.

Hoye et al. Tetrahedron Lett. 34, 5043.

1993

Total synthesis of solamin and reticulatacin, two mono-THF acetogenins, confirms their absolute configuration.

Keinan et al. J. Am. Chem. Soc.

115,4891.

Soc.

258 relative and/or absolute configuration, ii) An appropriate battery of spectral data from this set is carefully collected, interpreted, and tabulated, iii) Trends In these data are observed, iv) Relevant data from the natural products themselves, or appropriate derivatives, are collected and compared with those from the set of model compounds to deduce the relevant configurational relationships. Many of the general methods described herein make use of comparisons of NMR chemical shift data between a molecule with an intact natural product skeleton and another, skeletally simpler, model compound. It is more convenient to draw parallels between the two if the numbering scheme used to refer to the atoms involved are the same in both stmctures. Therefore, wherever possible, atoms on the carbon skeleton of the model compound(s) will be numbered corresponding to the natural product(s) they are intended to mimic, regardless of the "proper" numbering for the model structure.

II. THF BACKBONE 1. Adjacent bis-THF Structures (B1) In 1986, this laboratory''^ described the synthesis of a series of twelve acetylated model compounds 11a-l (Figure 6) for the bis-THF structure of uvaricin (1), one of only a handful of Annonaceous acetogenins known at the time. Among other things, we noted that the ''H NMR chemical shifts of the acetate methyl groups on the models showed a clear correlation with the relative configuration (either three or erythro) between the carbon bearing the acetoxy group and the adjacent carbon in the THF ring. Specifically, an erythro relationship between C(15)/C(16) or C(23)/C(24) in the model compounds led to a 5 of 2.051 ± 0.007 ppm, while a three relationship placed the methyl group at 2.075 ± 0.008 ppm. Since the acetate methyl group in uvaricin (1) resonates at 2.049 ppm, while the acetate derivative of uvaricin (the diacetate 12) showed acetate signals at 2.049 and 2.074 ppm, we concluded that the relationship between C(23)/C(24) in uvaricin is erythro, while the C(15)/C(16) relationship is three (Figure 7). Notice that very small differences in chemical shift were meaningful in the trend just described--the acetate methyl groups for each set of six diastereomeric compounds having either both erythro or both three terminal diastereomeric relationships all fell within a range of just over one one-hundredth of a ppm. Moreover, the two different diastereomeric environments led to a difference of only slightly more than two onehundredths of a ppm (i.e., A5 = 0.024). This requires a certain degree of care in measuring and reporting chemical shift data. To ensure reproducibility and confidence in our measured 8 values, we always include TMS as an internal standard in our samples, and we always set the TMS resonance to 5 = 0.00 ppm before printing spectral peak positions. Although this may seem obvious, it Is apparent to us that many

259

< ^Me

AcO,.,^ n-CsH,^

AcO,,,^ n-CgH^^ AcO,

fl5 ^16 O

erythrocis

n-CsH,, AcO,,.^ n-CsHn

P

I P

AcO,,.^n-C5Hii

I P

"MQ

threo-

J 20

cis

>

123 l24

erythro-

"^Mo

AcO**'^n-C5Hii

AcO^^^n-CsHn AoO^''^n-CsHn AcO^^nCgHt,

AcO*'*^n-CgHi,

\AB

11a

lib

11c

erythro/cis/threo/cis/erythro « er/c/lh/c/er

erlVth/der

er/t/th/t/er

lid er/c/er/c/er

lie erA/er/c/er

111 erfV&rfVer

AcO^^n-CsHii

AcO^^n-CsHii AcO^^n-CsHn

AcO^^n-Crf-lii

AcO^^n-CsHn

AcO^^n-CsHn

AcO*

AcO^^n-CgHt^ AcO*'

AcO

AcO**

AoO

n-CgHii

n-CsHn

n-C^n

iig

11h

111

Ih/c/th/c/th

th/VtfVc/th

iy

tlWtfWth

th/c/er/cmi

n-CsHn

tt/t th/Ver/c/th

n-CsHn

111 fhn/erfm

Figure 6- Twelve (of Twenty Possible) Diastereomeric, Synthetic, Model Compounds for the bis-THF Core of B1 Acetogenins. researchers are in the habit of referencing spectral resonances to some standard value of the solvent peak (e.g., residual CUCb in the CDCI3 to 8 = 7.26). This is dangerous because the solvent chemical shift is solute-dependent (e.g., CHCI3 is a weak hydrogen-bond donor); the inert TMS standard much less so.

or both down (i.e.. threo)

r^]"^

O

HgC^ O S 2.049

5 2.049

Figure 7. Acetate CJbis Chemical Shifts of Uvaricin Acetate (12) Define the C(15)/C(16) and C(23)/C(24) Relative Configurations for Uvaricin (1).

260 Further comparison of the model compounds H a - P ^ with uvaricin acetate (12) supported the assignment of C(23)/C(24)-erythro and C(15)/C(16)-threo. This analysis contributed additional information about the three-dimensional structure of uvaricin (1). Each stereorelationship along the THF backbone (three and erythro for pairs of adjacent oxygenated carbons, or cis and trans across THF rings) was correlated with a distinctive set of chemical shifts for the protons along the bis-tetrahydrofuranyl structure. The differences observed were in some cases small, but still significant enough to impart confidence In deducing the relative configuration along the B1 subunit. The chemical shift patterns of the twelve model compounds and of uvaricin and uvaricin acetate are recorded in Table 2 and shown graphically in Figure 8. (Due to the symmetry of the model compounds, and to simplify the graphic, we here revert to the numbering scheme used in the original paper.) In addition to the previously noted acetate methyl shifts, key observations were: i) H(5) and H(2) each appeared 0.04 - 0.08 ppm downfield in the trans/trans models compared to the cis/cis Isomers (where cis and trans refer to the substitution pattern on each THF ring); ii) if the configuration of the model is either cis/cis or trans/trans, H(2) is shifted farther downfield for the relationship C(2)/C(2') = three than for C(2)/C(2') = erythro; and iii) in the unsymmetrical (cis/trans) models, the resonances for H(2) and H(2') are nearly superimposed for C(2)/C(2') = erythro but significantly separated when C(2)/C(2') = three. Visual comparison of the ''H NMR spectrum of uvaricin acetate (12) to the model compounds yielded the closest match with the .../trans/threo/trans/... models; this, coupled with the previous information about the C(15)/C(16) and C(24)/C(25) relationships led to the conclusion that uvaricin has a three/trans/threo/trans/erythro configuration along the THF backbone, proceeding from carbon 15 to carbon 25 (see Figure 8). Table 2.

^H NMR Chemical Shift Values (in ppm) for the Methine Protons Associated with Oxygenated Carbons In the Model Compounds l l a - l .

1 ^ Configuration

H(2)

H(2')

H(5)

H(5')

H(6)

H(6')

Ac

Ac'

11a

er/c/th/c/er

3.81

3.81

3.94

3.94

4.90

4.90

2.045

2.045

lib

er/t/th/c/er

3.76

3.88

3.93

4.01

4.91

4.91

2.053

2.048

11c

er/tAh/t/er

3.88

3.88

3.98

3.98

4.91

4.91

2.045

2.045

lid

er/c/er/c/er

3.71

3.71

3.91

4.95

er/t/er/c/er

3.80 3.84

3.91

3.91 3.97

4.95

lie

4.91

4.96

2.045 2.058

2.053

3.99

3.99

4.92

4.92

2.050

2.050

3.86

3.93

3.93

4.94

4.94

2.069

2.069 2.077

2.045

lit

er/t/er/t/er

3.80 3.84

iig 11h

th/c/th/c/th

3.86

th/t/th/c/th

3.84

3.93

3.91

4.08

4.88

4.88

2.077

111

th/t/thAAh

3.90

3.97

3.97

th/c/er/c/th

3.77

3.93

3.93

4.85 4.84

4.85 4.84

2.074 2.074

11| 11k

3.90 3.77

th/t/er/c/th

3.82

3.82

4.850

th/t/er/t/th

3.84

3.84

3.93 3.97

3.97

111

3.97

4.84

2.073

2.073

4.852

2.080

2.075

4.84

2.071

2^07;^J

261

er/c/tfi/c/er er/t/th/c/er

o^Ms

er/t/th/t/er

nc f/ef

er/t/er/c/er er/t/er/t/er

-J5-J ;

th/t/th/c/th .n-CsHii t h / t / t h / l / t h

15

Figure 8.

th/t/th/l/th

I

A

2

A_..

I..5. 11k 111

: 5.0

.5._5lL...J2 5 i 2

6

4.9

r 4.8

f2 74

A

5 !

... i.. 6 !-«•

....lij... 2i^15

H(24)

A

•.A...

2415

I

...A.

i.

.A;A_.

4.1

-4-

th/c/th/c/er

.A:A.

I

5(ppm)

th/t/th/t/er

.-_.A.

J5-...J.?....:.

* __..5

..6..;. 6 f^ ....6!..

;.

i..2i.;.

^

._.j5l

..e..i :«--^

Ih/c/th/c/th

Ih/c/er/c/th th/t/er/c/th th/t/er/l/lh

t d .....^-.

rs 24/15;

.;

4.0

3.9

-4—

-4—

"T" 3.8

u.-Ji 23/16 20/19

.; ;

3.7

-I 1-—

i ...i

J.J.J AcO

'1623 i 20/19' 23/16201/19

\

AcO

•

i.Ul..LJ.L -.:-l-—1-

2.08 2.06 2.04

i

-^-

AcO

AcO

AcO's!

Graphical Comparison of Proton Chemical Shift Data for Model bisAcetates 11a-l and the Peracetate Derivatives of Uvaricin (12), Rolliniastatin I (14), and Asimicin (15).

In that early work w e relied much more on proton than carbon chemical shift trends. This was, in part, driven by the limited quantities of some of the twelve pure, synthetic, model compounds. However, the proton shift trends were also more meaningful than the carbon for this particular set of model compounds. It is our contention that proton chemical shift data should be used more frequently for this purpose and that this underutillzation Is largely a bias of technological origin. From the advent of '•^c NMR spectroscopy chemical shift trends were recognized to be of primary importance. Relatively large field dispersion and the routine lack of coupling data predisposed

262 researchers to rely heavily on "^^c chemical shifts. In the case of protons, only in the last decade has the routine availability of spectra recorded at Increasingly higher magnetic fields provided relatively complete assignments of the majority of resonances in spectra of complex molecules. Thus, proton chemical shift trends in complex molecules now warrant very careful attention. The 1987 report of the X-ray structure of a derivative of rolllniastatin I (8)''2 permitted us to further validate this ''H NMR correlation method."'^ The relative configuration along the THF rings in rolllniastatin I (8) (I.e., threo/cis/threo/cis/erythro) as determined by chemical shift correlation (Figure 8) matched exactly with the crystallographically determined structure. In the course of this analysis, some refinements to the method were made. Unlike uvaricin acetate (12), the triacetylated rolllniastatin I (14) did not exhibit a clear correlation with a single set of model compounds; two possibilities for the relative configuration were identified by simple visual inspection. Therefore, it was necessary to make the method more quantitative to arrive at an unambiguous conclusion. This was accomplished by comparing each of the eight measured "'H chemical shifts [H(15), H(16), H(19), H(20), H(23), H(24), and the two acetate methyl groups] for the natural product derivative with the analogous resonances for each of the model compounds, and taking the sum of the observed chemical shift differences. (For the six, unsymmetrical diastereomers of the model compounds that were JTJQI made, the expected chemical shifts were extrapolated from the relevant symmetrical model compounds.) The model having the smallest sum of the absolute values of chemical shift differences (I|A5's|) compared with the natural product represents the most likely relative configuration. The results of this comparison both for rolllniastatin I (8) and for aslmicin (13), another recently (at the time) discovered Annonaceous acetogenin,''^ are summarized In Table 3. Rolllniastatin I triacetate (14) shows the best match with a hypothetical erythro/cis/threo/cis/threo model, which corresponds exactly with the relative configuration determined by X-ray crystallography on the derivative of the natural product. Aslmicin triacetate (15) was determined to be threo/trans/threo/trans/threo, which is the same conclusion reached by visual inspection. A comment must be made at this point about the limitations of this method. It leaves open the question of directionality of the stereochemical relationships. For example, were the complete relative configuration of rolllniastatin I (8) not known from the x-ray crystallographic study, it would not be possible to tell whether the order of relative configuratlonal relationships proceeding from C(15) to C(24) was threo/cis/threo/cis/erythro or erythro/cis/threo/cisAhreo. We refer to this as the "endedness" problem, and it is an issue of structural ambiguity that has been overlooked In a number of instances.

263

Table 3.

Quantitative Comparison of ''H NMR Chemical Shifts for the Peracetates of Uvaricin (12), Rolliniastatin I (14), and Asimicin (15) with each of the Twelve Model Compounds lla-l as Well as with Eight Additional Extrapolated Unsymmetrical Isomers. 2:|A5's| model

12

14

15

er/c/th/c/er er/lAh/c/er er/l/th/t/er

tra lib 11c

0.11 0.17 0.21

0.36 0.52

er/c/er/c/er

lid

0.30 0.28 0.08 0.62

5 6 7

er/l/er/c/er er/t/er/t/er th/c/th/c/th

lie 11f

0.34

8 9 10 11

thMh/cAh th/lAh/tAh th/c/er/c/th

0.44 0.28 0.36 0.32 0.02

th/l/er/c/th

111 11k

0.15 0.19 0.19 0.23 0.29 0.27

th/l/er/t/lh

111

0.22

0.15 0.21

0.22

12 13

er/c/th/c/th

11a/11g

0.26

0.09

0.36

entry

descriptor

1 2 3 4

119 11h 111

0.18 0.26 0.32 0.12 0.44 0.28

0.41

0.16 0.72

0.38 0.16

14

er/t/th/c/th

11b/11h

0.19

0.15

0.27

15

th/t/th/c/er

11h/11b

0.35

0.26

0.41

16 17

er/t/th/l/lh

0.05 0.47

0.09 0.54

18 19

er/l/er/c/lh th/t/er/c/er

20

er/l/er/t/lh

11C/11I 11d/111 lle/11k 11k/lie 11f/111

0.22

er/c/er/c/th

0.26 0.26 0.14

0.30 0.19 0.16 0.17

0.34 0.32 0.22

Born et. al.^ have reported a complementary technique to determine the relative configuration between a carbon in a THF ring and an adjacent carbinol center. This approach is applicable to all B1, 82, B3, or B4 substructures. The model compounds 16-ef and 16-f/i (Figure 9) were synthesized as a mixture of diastereomers and separated chromatographically as their acetate derivatives. These acetates were assigned as three or erythro by the observed ^H NMR coupling constant between H(15) and H(16). The isomer with the smaller coupling constant (J15/16 = 5.2 Hz) was assigned the three relative configuration while that with the larger (J15/16 = 6.0 Hz) was assigned as erythro. Such a small difference in J's suggests that this assignment was somewhat tenuous. However, it has since been confirmed by subsequent stereospecific synthesis and correlation of model mono-THF compounds."^^^o jhe

264

i^

O I

i^-'CioHai

(^

7\

63.84(82.29)"

^\1^15^^10^21

O ""5 3.84(71.83)

6 3.79(82.47)'

^ ^ , ^ " " 5 3.40 (73.87)

Figure 9. Diagnostic Proton (and Carbon) Chemical Shift Data for Simple erythroand threo-a-HydroxyalkyI Tetrahydrofurans. acetates were then reconverted to the free alcohols 16, which were studied by ''H and "•^c NMR spectroscopy. A correlation was found between the three or erythro configuration of the models and the chemical shifts of nearby ''H and ""^c nuclei, particularly C(15) and H(16). These results are summarized in Table 4 along with the relevant data for annonin I (4), the discovery of which was reported in the same paper. It is clear from these data that annonin I possesses one three and one erythro relationship between C(15)/C(16) and C(23)/C(24). The question of which was which (I.e., the endedness), however, was resolved only through X-ray crystallographic structural analysis of the derivative 10. This work provided another verification of our original approach to assignment of bis-THF relative configuration. Table 4.

Correlation of ''H and "•^c NMR Chemical Shift Values Between Annonin I (4) and the Diastereomeric Pair of Model Compounds 16-er and 16-th. 16-er

4 (Annonin 1)

16-th

H(15)

3.84

3.40

H(15), H(24)

3.40, 3.87

H(16) C(15)

3.84 71.83

3.79 73.87

H(16), H(23) C(15). C(24)

3.88, 3.87 71.7,74.1

1 C(16)

82.29

82.47

C(16), C(23)

83.4, 82.9

All of the methods discussed so far, however useful, still leave open the question of absolute configuration. In 1992, we described a study carried out in collaboration with the McLaughlin group^"' detailing our studies of Mosher ester (i.e., methoxytrifluoromethylphenylacetate) derivatives22-24 of various carbinol centers in the acetogenins. Since all THF-contalning acetogenins have at least one hydroxyl group in the B subunit, Mosher derivatization of these groups provides an opportunity to draw conclusions about the absolute configuration in this portion of the molecule. The principle behind the Mosher ester technique is illustrated in Figure 10. The two enantiomers of the Mosher acid chloride, (f?)-MTPA-CI and (S)-MTPA-CI, are used to derivatize a stereo-

265 genie carbinol center to the (S)- and (f?)-MTPA esters, 17-Sand 17-/?, respectively.25 Assuming that the preferred conformation is as shown, with the trifiuoromethyl group eclipsed with the carbonyt, conclusions can be drawn regarding the absolute configuration of the carbinol center based on "^H and ''^F NMR spectroscopic data.26 Since the phenyl group will tend to have a shielding effect on nearby atoms, protons in the L3 portion of the ester should appear farther upfield in the "• H NMR spectrum of 17Sthan in 17-/?, while those in the L2 substituent should display the opposite trend. more highly shielded

less highly shielded OMe

^<='S less highly shielded

more highly shielded

17S

t7fl

Figure 10. Dominant Conformations of Diastereomeric Mosher (MTPA) Esters of Generic Carbinols Indicating Various Shielding Environments for Protons. To confirm the applicability of this method to the tetrahydrofuran portion of the Annonaceous acetogenins, the symmetric model bis-THF 18, where both C(15) and C(24) were known to be f?, was synthesized and converted to the (S)- and (f?)-MTPA esters, S-MTPA-18 and f?-MTPA-18.2i The chemical shift differences of selected protons in these diastereomeric esters are shown in Figure 11. As expected, the sign of

63.81

HO/, n-CgH^

53.96--.^"

\ n-CgH^,

OCH3 14C^v^Bp^^(CH2)7C^^

81.66-1.58

^AOH^HQCH^CH^TC^

-¥ 0.04 A ^(14)

18

S'MTPA'18

61.62-1.54

R'MTPA-18

Figure 1 1 . Chemical Shift Differences (A6's) for Key Protons in the MTPA Esters of the Synthetic Model f?,f?-Diol 18.

266

the chemical shift difference (A5 = 6S - 6R) is negative for H(16) and H(19) in the L^ portion of the molecule and positive for H(14) In the L^ portion. Results for hexepiuvarlcin (19), a synthetic acetogenin of known absolute configuration,^^ were similar. This study2i resulted In determination of absolute configuration of certain stereogenic centers for several acetogenins. Among them were bullatacin (5) and bullatacinone (6), where the B1 subunit of the molecule possesses "unlike ends" (i.e. three/.../.../.../erythro or erythro/.../.../.../three). The mono-MTPA derivatives were made and the position of the ester determined by mass spectrometry. This allowed both unambiguous determination of absolute configuration and a solution to the endedness problem described earlier. This work, carried out in In the course of our synthesis of /7exep/-uvaricin (19),27 represented the first instances for which the entire stereostructure of any acetogenins was deduced (Figure 12).

Figure 12. First Acetogenins for Which the Entire Stereostructure was Determined.

2. mono-THF Structures (B2) Until recently, researchers wishing to assign the relative configuration of subunit B2 relied on the NMR correlation methods developed for other subunits, as just described. For example, the mono-THF-containing annonacin (3) was assigned a three configuration between C(15)/C(16) and C(19)/C(20) by the methods of Born et al, and designated as trans across the THF ring by comparing the ''H chemical shifts of its acetate with our previously described bis-THF model compounds 11a-l.28 This threo/trans/threo assignment was later verified."'^ OH

HO,..

''^sX^^s/\/\ IS

b 20 -

3 (Annonacin)

OH

o. 'VO

.4^>ur

HO,..^ MS

L '° 1 20

Figaddre et al. reported in 199229 the first use of mono-THF model compounds containing two a-hydroxyalkyi substituents to study the stereochemical features of

267

mono-THF acetogenins. Two synthetic Intermediates, 20a and 20b, of known relative (and absolute) configuration and both possessing a trans-substituted THF ring were studied by iH and ^^c NMR spectroscopy. The chemical shift patterns in these free alcohols were compared directly with those of natural products of known relative configuration (as determined by extrapolation of the bis-THF stereostructure determination methods described above), in hopes of finding a diagnostic pattern. This pattern appeared in the i^C and "^H chemical shifts at positions 15,16,19, and 20 (Table 5). These shifts showed excellent correlation between the natural products examined, murisolin (21) and annonacin A (22), and the models possessing the same relative configuration. Figadere's model compounds can be used to distinguish acetogenins having threo/trans/threo or erythro/trans/threo structures, but only these two out of the eight possible diastereomeric relationships were modeled.

Table 5.

^H and "i^C NMR Chemical Shifts of Synthetic Intermediates 20a and 20b and Two mono-THF Acetogenins, Murisolin (21) and Annonacin A (22). H(15)| H(16)

H(19)

H(20)

C(15)|C(16)|C(19)

C(20) 1

20a 1 (threoArans/threo)

3.41

3.80

3.80

3.41

74.04

82.71

82.71

74.01

20b (erythro/trans/threo)

3.82

3.82

3.82

3.38

71.56

83.24

82.15

74.33

Murisolin {21) (threo/trans/threo)

3.38

3.76

3.76

3.38

74.2

82.7

82.7

74.2

Annonacin A (22) 1 (erythro/trans/threo)

3.82

3.82

3.82

3.40

71.62

83.32

82.31

74.361

A year later, Cassady published the synthesis^^ and study^^ of a set of model compounds representing the full spectrum of possible stereochemical relationships in dihydroxylated mono-THF acetogenins. Six^O mesitoylated compounds (23a-f, Fig. 13) were synthesized and subjected to 1H NMR spectroscopic studies. In these bismesltoates, good correlation between the relative configuration and the chemical shifts of selected protons was observed (Table 6). Specifically, It was found that: (I) where a three relationship existed between C(15) and C(16) [or C(19) and C(20)],

268

MesO^^ C 4 H g

cIUo" o

M G S O *^ C ^ H g

MesO,, ^C4H9

MesO^

C4H9

120 C4Hfl

MesO*

[20 C4H9

c/I"

MesO*

c/I"

MesO^^^C4H9 |15

MesO^'*

MesO /,.^C4Hg *fl5

120 C4H9

MesC^

[20 C4H9

MesC

120 C4H9

23a

25/7

23c

23d

23e

23f

(th/c/th)

(er/c/er)

(th/c/er)

(ttVt/th)

(er/t/er)

(th/t/er)

Me

O

0-1

MesO- s mesitoate Me'

Me

Figure 13. Six Mesitoylated Model Compounds for the B2 Subunit. H(15) [or H(20)] displayed a chemical shift of <5.37 ppm, while an erythro relationship gave rise to a 5 of >5.43 ppm, (ii) where a carbinol center is involved In a three relationship, the 2,6-methyl groups of the attached mesitoate ester will have a chemical shift of >2.45 ppm, while an erythro relationship produces a chemical shift of <2.40 ppm for these methyls, and (ill) a cis relationship across the THF ring will cause the THF ring protons, H(16) and H(19), to display a chemical shift of 3.89-3.97 ppm, while a trans relationship causes a downfield shift to 4.00-4.11 ppm. Table 6.

Important Chemical Shifts in the Cassady Model mono-THF Mesitoates 23.

Compound

Relative Configuration

6H(16). H(19)

5H(15), H(20)

5 H(Me's)

23a

threo/cls/threo

3.97

5.29

2.45

23b

erythro/cis/erythro

3.93

5.45

2.40

23c

threo/cis/erythro

3.89, 3.97

5.37, 5.43

2.39, 2.48

23d

threo/trans/threo

4.09

5.35

2.45

23e

erythro/trans/erythro

4.05

5.48

threo/trans/erythro

4.00,4.11

5.30, 5.52

[ 2 3 1

2.40

2.39.2.51 J

To apply this method to mono-THF acetogenlns, a "difference minimization" scheme similar to that earlier described''^ was used. That is, the differences in chemical shifts between a model and the natural product for all relevant hydrogens were added, and this value was compared for each of the models. The model with the smallest sum of differences is concluded to have the same relative configuration as the natural product.

269

In this case, the natural product annonacin (3) was first determined to have a symmetrical structure (I.e. not threoArans/erythro or threo/cis/erythro) by the fact that protons in analogous positions were superimposed in the 1H NMR spectrum (as in Table 6, the protons are resolved in an asymmetric structure). It then remained only to determine if the relationships In annonacin were threo/cis, erythro/cis, threo/trans, or erythro/trans. The ring methine [H(19)-H(20)] and ester methine [H(15)-H(20)] protons in annonacin per-mesitoate were compared with the model compounds 23 (Table 7). These results clearly show that annonacin (3) possesses the threo/trans/threo relative Table 7.

1

Difference Minimization Data for Identifying the Best Fit (threo/trans/threo) Between Mesitoate Models 23 and Annonacin (3).

ZA5H = |6Hannonacin - SHmodellRing Methine + ISHannonacIn " 8HmodellEster Methine Ring Methine [C(16)/C(19)]

Model SHannonacIn 8H model Configuration threo/cis

4.13

|A8H|

Ester Methine [C(15)/C(20)] |A8H|

ZASH

5.29

0.03

0.19

SHannonacin SHmodel

3.97

0.16

5.32

erythro/cis

4.13

3.93

0.20

5.32

5.45

0.13

0.33

1 threo/trans

4.13

4.09

0.04

5.32

5.35

0.03

1 erythro/trans

4.13

4.05

0.08

5.32

5.48

0.16

0.07 0.24 1

configuration. This method is also applicable to the B4 subunit, containing only one hydroxyl group a to the THF ring. Cassady's observations are consistent with similar observations we have made on the analogous acetate esters.^o

Figure 14. Acetogenins for Which the Absolute Configuration of the B2 Subunit Has Been Determined.

270 Again, these correlation methods leave the question of absolute configuration unanswered. However, the absolute configuration of the B2 unit in reticulatacin (24), isoannonacin-10-one (25), annonacin-10-one (26), and annonacin (3) has been determined by the Mosher ester method described earlier (Figure 14).2t The configuration of all carbinol centers flanking THF rings in these particular compounds was determined to be R. Given the previously determined threo/trans/threo relative configuration of all these molecules, the absolute configuration of the entire THF portion was assigned.

3. non-Adjacent bis-THF Acetogenins (B3) The several acetogenins containing the B3 or non-adjacent bis-THF substructure all have one THF ring with two adjacent hydroxyl groups, and a second ring flanked by one hydroxyl group [e.g., gigantecin (2, Figure 1)]. Given the above constitution, methods already described can be applied to each mono-THF portion of the B3 acetogenins. Indeed, the relative configuration of some nonadjacent bis-THF molecules has been proposed based on such methods.'^ However, since the two THF's are separated by only two carbons, it is possible that each subunit exerts an influence on the chemical shift patterns of the other, which could possibly perturb the data. In addition, the known stereostructure determination methods can only treat the two portions of the THF backbone as isolated entities and provide no means of determining the stereochemical relationship between the two. That is to say, there is yet no way to confidently distinguish, e.g., isomers 27a and 27b (Figure 15). It would be useful to have a method to completely and unambiguously assign the configuration of this type of structure.

27a

27b

Figure 15. Ambiguity Exists (cf., 27a vs. 27b) in the Relative Configuration Between Internal Carbinol Centers in All non-Adjacent bis-THF Acetogenins (B3).

III. TERMINAL y-LACTONE (AND ADJACENT 4.0H) As stated earlier, virtually all Annonaceous acetogenins possess a y-lactone at the terminus of the carbon chain. The known structural variations within this lactone ring ("A") are summarized in Figure 5. Although the absolute configuration at C(36) in

271 uvaricin was the first stereochemical feature to be deduced for the Annonaceous acetogenins, further stereochemical studies have focused almost exclusively on the tetrahydrofuran backbone. Attempts have been made to find a general method to determine the stereostructure of the lactone moiety. These include the use of Hudson's rule and/or optical rotary dispersion methods.9.3i-32 However, we view these approaches as tenuous since they are based upon data from less than ideal model compounds. As mentioned earlier, the absolute configuration of C(36) in uvaricin has been unambiguously determined to be S^^ Presuming a similar biosynthetic pathway for all Annonaceous acetogenins, it is tempting to assume that all such natural products possess this absolute configuration. However, this need not be the case. It is necessary either to perform a chemical degradation on each acetogenin, which may not be practical in all cases, or to develop a new method for unambiguous assignment of the lone, remote C(36) stereocenter In A1 acetogenins. Another early assumption was that the C(36)/C(4) relative configuration in 4-hydroxy butenolide-containing (A2) acetogenins could be determined by direct comparison of NMR shift data. For example, bullatacin (5) was assumed to have the same relative configuration (4S*,36f?*) determined for rolliniastatin I (8) because they exhibited "essentially the same" ^H and ^^C NMR signals for subunit A2.9 However, in the course of preparing model compounds for the structure A2, we observed that the ''H and "'^C NMR behavior of this structural unit was virtually identical regardless of the relative configuration.3'33 Clearly, similarity by NMR spectroscopy is not enough to establish the relative configuration [fT.R* (or like) or fT.S* (or unlike)] of C(36)/C(4).

1. Unsaturated Lactones Having a Hydroxyl Group at C(4) [subunit A2] Recently, general methods have been developed to assign both the absolute configuration of C(4) and the relative configuration between C(4) and C(36) in acetogenins possessing the A2 subunit.2''.34 in these acetogenins, which have a hydroxyl group at C(4), the absolute configuration of this carbinol center can be determined by the Mosher ester method described earlier in this chapter.2i Model compound 28SS, of known (S,S) configuration at C{4) and C(36), was prepared and derivatized with both enantlomers of the Mosher acid chloride. The A8H and A8F obtained for these derivatives are consistent with the known S configuration at C(4). Extension of this method to the A2 subunit in several acetogenins (Figure 16) has shown that the configuration at C(4) in all these compounds is R. Note that this result disproved a previous assumption that bullatacin (5) has the S-configuration at C(4).9

272

OH ^ V - O

28-SS

28-RS

OH

>-0

28'SR

Figure 15. Absolute Configuration of the A2 Subunit In Bullatacin (5), Asimicin (13), RoHiniastatin I (8), Annonacin (3), and Annonacin-10-one (26) Is (4H,36S) [or (4f?,34S) for 3 and 26] from Comparison of their Mosher Esters with Those of Synthetic Model Compounds 28.

Further work has enabled us to extend this method to determine the relative configuration between C(4) and C(36) in the A2 subunit .20,34 Whereas the method just described (and indeed, the conventional use of Mosher ester analysis) relied on the sign of the chemical shift differences to determine the absolute configuration, we have found that the C(4)/C(36) relative configuration is reflected in the magnitude of these differences. This phenomenon was observed for all the Annonaceous acetogenlns we studied. In addition to the model compound 28-SS described above, its stereoisomers 28-f7Sand 28-S/? were synthesized and derivatized with (R)- and (S)-MTPA-CI, respectively, and the Individual sets of chemical shift differences In their ""H and ""^F NMR spectra were compared. These A6 data, along with those from the relevant portion of bullatacin (5), are shown as absolute values graphically in Figure 16. It is clear from the graphic that the magnitudes of the chemical shift differences fall into definite patterns depending on the relative configuration between C(4) and C(36). Thus bullatacin (5),

273 along with the other Annonaceous acetogenins we studied, was determined to possess FT.S^ or unlike relative configuration between these stereogenic centers. Since all of these acetogenins had been determined to have an /? configuration at C(4) by the methods described earlier, it follows that C(36) is S. This information establishes the complete stereostructure for all five natural products studied, with the exception of C(10) In annonacin, 3 (Figure 15). • ^HData

«

» ^^hUata ||A5FI

|A8HI

(ppm)

(S)-7SS/(fl)-7SS

(S)-7S/?/(S)-7f?S

(ppm)

Bullatacin(l)

1 r

U.J

,A n +U.O

k-+0.4 U-+0.3

0.2

0.1

. -- . .

.-..._...« -.....-

0.0

j]

\lu

1—1—1—\—1—1

37 36 35 3a/b 4

Like

5

. _1 L......---.

J . . . . J U-+0.2 h-+0.1

m "Tl

1 1U—-on 1 1LLJ I l l 1—1—1—1—1—1 r"1—1—1—11—1 37 36 35 3sJb 4

Unlike

5

\37 36 35 3a/b ^

5

36S4f?

Figure 16. Graphical Representation of Absolute Values of A8H and A8F Data from Models 28 vis-a-vis Bullatacin (5): Magnitudes, Not Signs, Permit Assignment of C(4) vs. C(36) Relative Configuration. 2. Rearranged Acetonylbutanolides (A3) In 1991, we reported a method^s for assigning the relative configuration between C(2) and C(4) in the acetogenins containing an acetonylbutanolide moiety (A3). First, cis and trans model compounds 29c and 29f, respectively, were synthesized, and the relative configuration of each was unambiguously assigned through nOe studies and conformational analysis to explain the observed values of coupling constants. With these assignments in hand, the model compounds were examined by ''H NMR spectroscopy in both CDCI3 and CeDe. In both solvent systems, the two diastereomers were found to contain many characteristic differences in their chemical shift and coupling patterns (Table 8). Most diagnostic are the following: (i) H(4) occurs >0.1 ppm downfield in the trans isomer compared with the cis, (ii) the geminal hydrogens H(3a) and H(3p) are closer together (A6, of --0.2 ppm vs. ~1.1 ppm) in the spectrum of the trans compounds vs. the cis, and

274 (iii)

H(4) and H(3|3) in the trans diastereomers exhibit relatively small coupling (J = 3.9 Hz for [29f|) while in the cis isomer the large coupling (J = 9.8 Hz for [29c]) reflects their trans-diaxial-like orientation. o o—f

.o o

o' ' y 3< 6^^

29c Table 8.

29t

""H NMR Data Reflecting Diagnostic Differences in Model Diastereomeric Acetonyl Butanoiides 29c and 29t. Compound #

|

In CDCI3

c/s-Model 29c

frans-Model 29t

Isoannonacin 31

lsoannonacin-10-one 25

5H(4)

4.41

4.55

4.54

4.55

JH(33)/H(4)

A5H(3a) - H(33) ^Configuration

9.8

3.9

3.6

3.6

+ 1.09

-0.22

-0.27

- 0.24

cis

trans

trans

trans

Compound # In CeDe

c/s-Model 29c

frans-Model 29t

5H{4)

3.69

4.02

Bullatacinone Bullatacinone Squamone (Minor) (Major) 30 6b 6a 4.05

3.72

JH(33)/H(4)

9.8

3.9

A5H(3a)/H(33) 1 Configuration

+ 1.10

-0.31

-0.30

+ 1.08

cis

trans

trans

cis

4.00

trans

|

Comparison of these results with the published spectral data of acetogenins with the A3 subunit and their derivatives yields strong evidence for the relative configuration of the lactone moiety. These results (Table 8) show that the mono-THF acetogenins squamone (30), isoannonacin (31), and isoannonacin-10-one (25) possess a trans relationship between C(2) and C(4); and in the bis-THF acetogenin bullatacinone (6), which was isolated as a mixture of cis and trans diastereomers, the trans configuration predominates (Figure 17). Rollinone (32) was shown to be an --1:1 mixture of cis and trans isomers. Notice that in those structures the absolute configurations of C(15)-C(24) is known only for 6a and 25; for all other structures in Figure 17 only the relative configurations along the THF core and, independently, across the butanolide ring are known.

275

Figure 17. Acetogenins Containing the Rearranged Acetonyl Butanolide (A3). 3. Saturated Lactones with a |3-Hydroxyi Group (A5) In the recently discovered A5 lactones-those bearing a hydroxyl group on the lactone moiety~the relative configuration within the lactone has been reported by Cortes et alJO In itrabin (7) and jetein (33), the protons on the lactone were found to exhibit ^H NMR nOe's of 2% between H(2) and H(33) and 1% between H(33) and H(34), suggesting an all-cis stnjcture (Figure 18). Coupling constants of -5.5 Hz between H(2) and H(33) and -7 Hz between H(33) and H(34) were measured for jetein 33;36 these data are consistent with the reported configuration.

Figure 18. Examples of p-Hydroxybutanolide-containing (A5) Acetogenins.

276 III. OTHER

FUNCTIONALITY

Although all Annonaceous acetogenlns possess some variant of the lactone subunit, and most have THF rings, other functional groups are often present along the carbon chain (or in place of the THF ring structure in presumed biological precursors). One of these substructures, the carbonyl moiety (C2), adds no stereochemical complexity to the natural product; the rest, however, have stereochemical Issues that must be addressed in determining the complete three-dimensional structures of the natural products. Since most new research has focused on other parts of the acetogenlns, this is an area that is still being explored. However, since these functional groups are not unique to this class of natural products, some classical methods do exist for determining their configuration. In some acetogenlns, a lone hydroxyl group (C1) has been found along the carbon chain. In one such compound, the absolute configuration of this carbinol center has been successfully determined. McLaughlin et. al.^^ prepared the tris-MTPA esters of bullacin (34) and concluded, from the data summarized in Table 9, that C(6) possesses an S configuration. Note that even protons relatively remote from C(6) show useful A5's.

Table 9.

Mosher Ester A5H Data for the Tris-MTPA Derivative of Bullacin (34). 5

H(35)

(S'MTPA'34)

5

(R'MTPA'34)

A5H

= 5S - 6R

1.39

1.39

0

H(34)

4.97

4.96

+ 0.01

H(33)

6.95

6.86

H(3)

2.18

+ 0.09 + 0.09

H(4)

2.27 1.54

H(5)

1.67

H(6) H(7)

5.08 1.57

1.55 5.07 1.67

1.43

+ 0.11 + 0.12 -0 -0.10

277 Some acetogenins, such as rollinicin (35), have vicinal did moieties (C4) in their structure. Sneden et. al.38 employed an interesting method for determining the relative configuration of the diol in rollinicin. By a method previously reported by Murata.^Q the tris-trimethylsilyl ether of the natural product was prepared and analyzed by ci mass spectrometry. According to Murata, if the [MH+ - MeaSiOH - Me2Si=CH2]+ ion (MH - 90 - 72) is in much greater abundance than the [MH+ - MeaSiOH - Me3SiOH]+ ion (MH - 90 - 90), then the relative configuration of the diol is erythro; if the opposite is observed, it is three. With rollinicin, the MH - 90 - 72 ion was found to be twice as abundant as the MH - 90 - 90 ion, suggesting an erythro relationship for the diol In 35.

Another known subunit of the acetogenins is an isolated 1,2-disubstituted alkene moiety (C6). The geometry of the double bond can be determined by inspection of coupling constants In "^H NMR spectroscopy. For example, the vinylic protons, H(9) and H(10), in giganenin (36) were shown to be coupled to each other with J = 10.9 Hz, diagnostic for the Z-olefin geometry.-^o Alternatively, the geometry can be deduced from the •'^C NMR shifts of the allylic carbons. For example, in epomuricenin (37),41 C(18) and C(21) were found to have chemical shifts at 824.0 and 827.0; a trans geometry would have resulted in a downfield shift for these carbons.

37

(Epomuricenin A)

A special case of structure determination is that of muricatacin (38). This unique molecule was isolated from Annona muricata seeds,^ along with more conventional acetogenins, and is considered to be a metabolic product of the plant. By comparing sign and magnitude of optical rotation of 38 with the known, enantiomerically pure analog, (4S,5S)-5-hydroxypentadecan-4-olide (39), an intermediate in the synthesis of disparlure, it was concluded that 38 exists as a mixture of the (4R,5R) and (4S,5S) enantiomers, with the (4f?,5f?) slightly in excess (Figure 19). This hypothesis was later supported by syntheses of both enantiomers of muricatacln.6.42-44

278

38

(MurJcatacin)

[alo^ - +29.2° Mixture of (R,R) and (S,S) isomers

(S.S)

Figure 19. Comparison of Rotation Data for Muricatacin (38) and Analog (39).

IV. VALIDATION OF STRUCTURE DETERMINATION METHODS In several cases, evidence for the validity of the methods described in this chapter has come by chemical conversion of one molecule into another, known product, thus demonstrating that they possess the same relative (and in some cases, absolute) configuration. The first example of such a conversion was the treatment of bullatacin (5) to produce the rearranged product, bullatacinone (6) (Figure 20). The product so obtained proved to be indistinguishable from the natural product.^ These results confirmed the finding that bullatacin and bullatacinone have the same relative configuration of the THF subunit, thus providing support for the method used to make those independent assignments. A similar result was obtained in the conversion of 25desoxy-4-hydroxyneorollinicin (40) to rollinone (32).45

Figure 20. Base-catalyzed Conversion of 4-Hydroxy Butenolide Acetogenlns (A2) Into Rearranged Butanolides (A3).

279 In another experiment, annonacin (3) was oxidatively cleaved with mCPBA to muricatacin (38) (Figure 21), the absolute configuration of which had been established by optical rotatlon.5 This conversion suggested the /? configuration at C(20) of annonacin. This fact was later borne out by the Mosher ester method described earlier in this chapter, lending support to the validity of the method.

^^'.. p H

38 (Muricatacin)

HO

Figure 21. Oxidative Cleavage of Annonacin (3) to Muricatacin (38).

Further proof has come through the recent total synthesis of several Annonaceous acetogenins. The first of these to be constnjcted was (+)-(15,16,19,20,23,24)-/)exep^ uvaricin (19), an unnatural diastereomer of uvaricin (1) reported from our laboratories in 1991.27 Knowing the relative, but not the absolute, configuration among C(15)-C(24) of

HO^

]15

0^ 1

41

(en^Bullatacin)

24

(Reticulatacin)

jP l24

M7

r^'o HO

O

38 (Muricatacin)

•H

Figure 22. Acetogenins for Which Total Syntheses Have Been Described.

280 uvaricin (1), we arbitrarily prepared one of the two possible enantiomers of the THF backbone. This was eventually coupled with an A1 fragment having 36-S configuration, as known for the natural product. The resulting acetogenin had ""H and '•^c NMR data identical with the natural product, but a slightly different optical rotation ([a]D = +11.3° for the natural product and +9.5° for the synthetic product). Definitive proof of the difference between these two molecules came with the Mosher ester derivatization of C(15) in both compounds. The (f?)-MTPA derivative of 1 had different i H and ^^F NMR spectra from the (/?)-MTPA derivative of 19, but Identical spectra to the (S)-MTPA derivative of 19. Since the absolute configuration of 19 was known based on the method of its synthesis, this provided conclusive proof of the relative and absolute configuration of uvaricin (1). Further efforts by several groups have resulted in total syntheses of muricatacin 38,6, 42-44 (-).bullatacin (41) (the enantiomer of the natural product),^^ solamin 42,^^7 and reticulatacin 24,^*7 (Figure 22). All of the above have been synthesized by methods that establish unambiguously their absolute configuration. All of the stereochemical information obtained so far by total synthesis is in complete agreement with the conclusions reached earlier by the methods described in this chapter. V.

CONCLUSION Although tremendous progress has been made in this area, several stereochemical

issues remain unaddressed in the Annonaceous acetogenins. The most obvious of these is the fact that there is still no general way to determine the configuration of C(36) in the A1 subunit, that is, the a,p-unsaturated lactone having no hydroxyl groups in Its vicinity. The absolute configuration at this center has been determined for uvaricin by degradation, but as mentioned earlier, this approach may not be practical in all cases. However, the absence of other stereogenic centers on the A1 subunit and the remoteness of this substituent from other functionality presents no small challenge to development of a more general method. In addition, although stereochemical information for portions of the B3 unit (nonadjacent bis-THF) is available, the complete relative configuration of the subunit as a whole remains unsolved {vide supra). Future work will undoubtedly provide more information for molecules of this structure. Despite these limitations, however, a wealth of information now exists on the stereochemical features of the Annonaceous acetogenins, and the complete relative and absolute configuration has been solved for several members of this class. The tremendous value and potential of NMR-based strategies and arguments is clear. Chemical shift trends rather than coupling constant analysis have proven much more powerful for this class of non-rigid molecules. It Is our hope that this collection of structure elucidation methods will prove valuable in the future study of this important family of natural products.

281 REFERENCES 1.

M. LebcBuf, A. Cave, P. K. Bhaumik, B. Mukherjee, and R. Mukherjee, Phytochemistry, 21 (1982) 2783-2813.

2.

S. D. Jolad, J. J. Hoffmann, K. H. Schram, and J. R. Cole, J. Org. Chem., 47 (1982) 3151-3153.

3.

J. K. Rupprecht, Y.-H. Hui, and J. L McLaughlin, J. Nat. Prod., 53 (1990) 237-278.

4.

X.-P. Fang, M. J. Rieser, Z.-M. Gu, G.-X. Zhao and J. L McLaughlin, Phytochem. Anal., 4(1993)27-48.

5.

M. J. Rieser, J. F. Kozlowski, K. V. Wood, and J. L. McLaughlin, Tetrahedron Lett., 32 (1991) 1137-1140.

6.

B. Figaddre, J.-C. Harmange, A. Laurens, and A. Cav6, Tetrahedron Lett., 32 (1991) 7539-7542.

7.

A. Alkofahi, J. K. Rupprecht, Y.-M. Liu, C.-J. Chang, D. L Smith and J. L McLaughlin, Experentia, 46 (1990) 539-541.

8.

L. Born, F. Lieb, J. P. Lorentzen, H. Moeschler, M. Nonfon, R. Sollner, and D. Wendisch, Planta Med., 56 (1990) 312-316.

9.

Y.-H. Hui, J. K. Rupprecht, Y. M. Liu, J. E. Anderson, D. L. Smith, C.-J. Chang, and J. L McLaughlin, J. Nat. Prod., 52 (1989) 463-477.

10. D. Cortes, S. H. Myint, M. Leboeuf, and A. Cav§, Tetrahedron Lett., 32 (1991) 61336134. 11. S. D. Jolad, J. J. Hoffmann, J. R. Cole, C. E. Barry, R. B. Bates, G. S. Linz, and W. A. Konig, J. Nat. Prod., 48 (1985) 644-645. 12. G. R. Pettit, G. M. Cragg, J. Polonsky, D. L Herald, A. Goswami, C. R. Smith, C. Moretti, J. M. Schmidt, and D. Welsleder, Can. J. Chem., 65 (1987) 1433-1435. 13. T. R. Hoye and J. C. Suhadolnik, Tetrahedron, 42 (1986) 2855-2862. 14. For an alternative classification scheme see reference 36. 15. T. R. Hoye and J. C. Suhadolnik, J. Am. Chem. Soc, 109 (1987) 4402-4403. 16. T. R. Hoye and Z. Zhuang, J. Org. Chem., 53 (1988) 5578-5580. 17. J. K. Rupprecht, C.-J. Chang, J. M. Cassady, J. L. McLaughlin, K. L Mikolajczak, and D. Welsleder, Heterocycles, 24 (1986) 1197-1201. 18. J. B. Gale, J.-G. Yu, X. E. Hu, A. Khare, D. K. Ho, and J. M. Cassady, Tetrahedron Lett., 34(1993)5847-5850. 19. J. B. Gale, J.-G. Yu, A. Khare, X. E. Hu, D. K. Ho, and J. M. Cassady, Tetrahedron Lett., 34 (1993) 5851-5854. 20. P. R. Hanson, Ph.D. Thesis, University of Minnesota, 1993. 21. M. J. Rieser, Y.-H. Hui, J. K. Rupprecht, J. F. Kozlowski, K. V. Wood, J. L McLaughlin, P. R. Hanson, Z. Zhuang, and T. R. Hoye, J. Am. Chem. Soc, 114 (1992) 10203-10213. 22. J. A. Dale and H. S. Mosher, J. Am. Chem. Soc, 95 (1973) 512-519. 23. G. R. Sullivan, J. A. Dale, and H. S. Mosher, J. Org. Chem., 38 (1973) 2143-2147.

282 24. I. Ohtani, T. Kusumi, Y. Kashman, and H. Kakisawa, J. Am. Chem. Soc, 113 (1991) 4092-4096. 25. Notice that Cahn-lngold-Prelog priority interchange accompanies each of the chemical conversions of the Mosher acid (MTPA-OH) to Mosher acid chloride (MTPA-CI) to Mosher ester (MTPA-OR). This is a particularly crucial detail since both enantiomers of MTPA-CI as well as of MTPA-OH are now commercially available. 26. See references 24 and 21 for a discussion of some of the ambiguities and pitfalls in relying on 19F data alone for assignment of configuration. 27. T. R. Hoye, P. R. Hanson, A. C. Kovelesky, T. D. Ocain, and Z. Zhuang, J. Am. Chem. Soc, 113 (1991) 9369-9371. 28. F. Lieb, M. Nonfon, U. Wachendorff-Neumann, and D. Wendisch, Planta Med., 56 (1990) 317-319. 29. J.-C. Harmange, B. Figad^re, and A. Cav6, Tetrahedron Lett., 33 (1992) 57495752. 30. Because of the skeletal symmetry in compounds 23 the total number of possible diastereomers is reduced from eight (i.e., 2" where n = 3 stereogenic centers) to six. 31. W. Klyne, P. M. Scopes, and A. Williams, J. Chem. Soc, (1965) 7237-7242. 32. N. Franca, O. R. Gottlieb, and D. T. Coxon, Phytochemistry, 16 (1977) 257-262. 33. Z. Zhuang, unpublished observations and ref. 20. 34. T, R. Hoye, P. R. Hanson, L E. Hasenwinkel, E. A. Ramirez, and Z. Zhuang, Tetrahedron Letters, in press. 35. T. R. Hoye and P. R. Hanson, J. Org. Chem., 56 (1991) 5092-5095. 36

A. Cave, D. Cortes, B. Figaddre, R. Hocquemiller, O. Lapr^vote, A. Laurens, and M. Leboeuf, in: K. R. Downum, K. R. Downum, J. T. Romeo, and H. A. Stafford, (Eds), Phytochemical Potential of Tropical Plants, Plenum, New York, 1991. pp. 167-202.

37

Z.-M. Gu, X.-P. Fang, L Zeng, K. V. Wood, and J. L McLaughlin, Heterocycles, 36 (1993) 2221-2228.

38

M. J. Abreo and A. T. Sneden, J. Nat. Prod., 52 (1989) 822-828.

39

T. Murata, T. Ariga, and E. Araki, J. Lipid Res., 19 (1978) 172-176.

40. X.-P. Fang, J. E. Anderson, D. L Smith, K. V. Wood, and J. L McLaughlin, Heterocycles, 34 (1992) 1075-1083. 41.

F. Roblot, T. Laugel, M. Leboeuf, A. Cav6, and O. Laprevote, Phytochemistry, 34 (1993) 281-285.

42. G. Scholz and W. Tochtermann, Tetrahedron Lett., 32 (1991) 5535-5538. 43. J. A. Marshall and G. S. Welmaker, Synlett, (1992) 537-538. 44. W. Tochtermann, G. Scholz, G. Bunte, C. Wolff, E.-M. Peters, K. Peters, and H. G. von Schnering, Liebigs Ann. Chem., (1992) 1069-1080 45. M. J. Abreo and A. T. Sneden, J. Nat. Prod., 53 (1990) 983-985. 46. T. R. Hoye and P. R. Hanson, Tetrahedron Lett., 34 (1993) 5043-5046. 47. S. C. Sinha and E. Keinan, J. Am. Chem. Soc. 115 (1993) 4891-4892.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

283

The Chemistry of the Nargenicin Macrolides James Kallmerten

I. INTRODUCTION The search for natural products possessing antibiotic activity against drug-resistant microorganisms continues to uncover a fascinating array of biologically active compounds. The recent discovery of a new structural class of macrolide antibiotics, the nargenicins, demonstrates the continuing role of natural products isolation as a catalyst for research in a multidisciplinary host of areas, including bioorganic, medicinal and synthetic organic chemistry. This report reviews recent efforts to define the structure, activity and biological origin of the nargenicins, and examines progress towards the total synthesis of naturally-occurring nargenicin macrolides. 1.

Isolation and Characterization of the Nargenicin Macrolides

The nargenicins comprise a small family of macrolide antibiotics characterized by a highlyfunctionalized decenolide system and a novel oxa-bridged octalin nucleus. The first nargenicins were independently discovered by groups at Pfizer and Upjohn. A 1977 patent issued to Pfizer describes the isolation of an antibiotic complex from the soil-dwelling microorganism Norcardia argentisisA In a subsequent report, Celmer and co-workers detailed the isolation and preliminary characterization of the major component of the complex, which they named nargenicin Ai.^ Based on an elegant series of NMR studies, the Pfizer group established the connectivity of the macrolide and the stereochemistry of the novel ll-oxatricyclo[4.4.1.1»60^»7]undecene nucleus; however, ambiguity surrounding the C14-C15 olefin geometry precluded further refinement of the gross structure. Meanwhile, studies by Whaley and coworkers at Upjohn had resulted in the isolation of a new macrolide antibiotic from cultures of Saccharopolyspora hirsuta, the structure of which was established as 2 by X-ray crystallography.^ The Upjohn workers termed their new antibiotic nodusmicin (from the Latin for knot, noduSy an appellation which reflects the complex Figure 1

=Y^ MeO.

T H I JL OH 1 ^ Me ^v^ H ^

22

JL. ^^ 20

Me,,

2; nodusmicin

1;R«0H. nargenicin A, 3; RaH, 18-deoxynargenicin A,

284 connectivity of the macrolide) and demonstrated the structural relationship of 2 to the Pfizer macrolide by a three step synthetic conversion of nodusmicin to nargenicin Ai (Figure 7), confirming the original Pfizer connective assignments and defining the structure of the latter compound as 1.^ The absolute configuration of an 11-O-p-nitrophenacyl derivative of nargenicin Ai has been determined using the nonempirical CD exciton chirality method of Nakanishi,^ suggesting that the configuration of the parent compound is that shown.^ Figure 2

r-^'°v^ MeO

4; nargenicin Ci 5; R,- R3. OMe, H^ OH, nargenicin B, 6; Ri= OMe. Rg- H. R3. OH. nargenicin Bg 7; R i - H. R2- OH. Rg. OMe, nargenicin B3

HO"i8

8; coloradocin (luminamicin)

Subsequent to the isolation of nargenicin Ai and nodusmicin, reports of the identification of structurally-related nargenicin antibiotics have appeared, although details of the isolation and characterization of these macrolide congeners have yet to be recorded in the primary literature. 18-Deoxynargenicin Ai 3 has been isolated from a nodusmicin-producing strain of 5. hirsuta by the Upjohn groups The structure of the 18-deoxy congener was established by synthetic conversion of nargenicin Ai to 3 via the Barton deoxygenation protocol (Figure 7).^ The Pfizer group has reported that further analysis of their N. argentisis antibiotic complex has led to the isolation of the C19-oxygenated nargenicin Ci, 4,9 as well as a series of C2-epimeric nargenicins that includes nargenicins Bi, 5, B2,6 and B3, 7.10 Recently, workers at Abbott Laboratories reported the isolation and preliminary characterization of a new nargenicin-type ^w-macrolide antibiotic, coloradocin, 8, from cultures of Actinoplanes coloradoensis .^^ The novel bislactonic structure proposed for 8 was based on a detailed spectroscopic examination which included extensive homo- and heteronuclear 2D NMR studies. 1^ Although these investigations failed to unequivocally defiine the relative configuration of the macrocyclic chirality, the key structural elements of 8, including an oxa-bridged octalin

285 nucleus, a maleic anhydride chromophore and a unique 10,14 ^«-macrolide system have been firmly established. Coloradocin is apparently identical to luminamicin, isolated by Omura and coworkers from a producing organism tentatively identified as belonging to the genus NocardioidesA^ The Japanese group has identified a structurally related compound, lustromycin, fit>m a Streptomyces strain; details of the structure have not yet been reported. ^^ 2. Biological Activity The nargenicins exhibit activity against gram-positive bacteria, especially Staphylococcus aureus (Table 1). A comparison of activities indicate that nargenicin Ai 1 is considerably more active in vitro than the C9-hydroxy congener nodusmicin, 2 J 18-Deoxynargenicin Ai is less active than 1 against staphyloccocus strains but shows pronounced activity against streptococci. Table 1. Agar-diffusion Antimicrobial Spectrum of 1, 2 and 3J Organism Bacillus subtilis Staphylococcus aureus Micrococcus luteus Klesiella psewnoniae Mycobacterium avium Penicillum oxalicum Saccharomyces pastorianus Bacteroidesfragilis Clostridium perfringens

Nodusmicin 0 — 35 27 26 0 0 33 33

INHIBrnON ZONE SIZE* Nargenicin Ai 18-Deoxynargenicin Ai 0 36 42 32 30 0 0 37 48

15 27 31 18 26 0 0 21 36

* In mm from 12.6 mm disc dosed with 0.08 ml of 1mg/ml solution. The activities of nargenicin Ai and Bi against S, aureus strains are comparable to that of erythromycin. Studies by the Pfizer group have indicated that both nargenicins retain activity against multiple-drug resistant Staphylococcus strains (Table 2). 10 Table 2, Comparative In Vitro Spectrum and Potency of 1 and 5.^® Organism Staphylococcus aureus Staphylococcus aureus^ Staphylococcus epidermis Staphylococcus epidermis*"^ Neisseria sicca Escherichia coli Pasteurella multocida

ANTIBIOTIC and MIC (Hg/ml) Nargenicin Ai Nargenicin Bi Erythromycin 0.1 0.2 0.8 0.8 25 2.0 25

0.8 1.6 1.6 6.3 0.8 12.5 200

*Multiple-dmg resistant strain. **Methicillin/erythromycin resistant strain.

0.1 >200 0.1 >200 1.6 1.6 0.4

286 Studies by the Upjohn group confumed the potent activity of nargenicin Ai against multiple-drug resistant strains of S. aureus and demonstrated the in vivo activity of the antibiotic in mice (Table 3)7 Table 3. Activity of 1, 2 and 3 Against Drug-Resistant S. aureus Strains^ In VitrO'MlC (^ig/ml) Sa Resistant Strain UC76 (control) UC6685 (P.T,C,N,K,E)* UC6686 (G,K)* UC6687 (P,K)* UC6686 (P,T,C,Ch,E)* UC76 (subcutaneous) UC76 (oral)

Nodusmicin >100 >100 >100 >100 >100

>320

Nargenicin Ai 18-Deoxynargenicin Ai 0.2 0.2 0.2 0.2 0.2

1.0 1.5 3.1 1.5 3.1

In Vivo CD50 (mg/kg) 17.4 MOO 50

* Resistant by Kirby-Bauer disc diffusion method to: penicillin G (P), Tetracycline (T), Clindamycin (C), Novobiocin (N), Kanamycin (K), Chloroamphenicol (Ch), Erythromycin (E). Originally isolated from a screen for antibiotic activity against anaerobic bacteria, coloradocin, 8, exhibits pronounced and selective activity against pathogenic anaerobes and microaerophiles, displaying only limited activity against aerophilic microorganisms. ^^ The biological spectrum of 8 contrasts those of the structurally-related nargenicins, which are primarily active against aerobic targets. In vitro studies have demonstrated that coloradocin has activity comparable to vancomycin against selected anaerobes, including Clostridium difficile, the causative organism for pseudomembranous colitis.^ 1 Coloradocin is particularly active against the niicroaerophilic organisms Neisseria gonorrhoeae, Haemophilus influenzae and Legionella strains, and is effective against ampicillin-resistant strains ofN. gonorrhoeae and H, influenzae. Acute toxicity of coloradocin is low (LD50 in mice injected intra-peritoneally is >500 mg/kg). 3. Biosynthesis The polyketide biosynthetic origin of the nargenicins has been independentiy confirmed by Cane 15 and Rinehardt.l^. Feeding experiments with l^C-labeled sodium acetate indicate that carbon pairs C1-C2, C3-C4, C5-C6, C7-C8 and C11-C12 of nargenicin Ai are acetate-derived; analogous studies with 13C-labeled propionate have shown that the C17-C18-C19, C15-C16C20, C13-C14-C21 and C9-C10-C22 triads of 1 derive from propionate (Figure 3). The C23 methyl group of nargenicin Ai originatesfromL-methionine; the pyrrole 2-carboxylic acid moiety of 1 derivesfrompropionate/acetate via the succinate-a-ketoglutarate-dehydroproline pathway. Cane has demonstrated that advanced di-, tri- and tetraketide fatty acid precursors are incorporated directiy into the nargenicins, and has noted that an early intermediate in the nargenicin pathway is common to the biosynthetic scheme leading to erythromycin A and

287 Figure 3

=

A Me"

CH0. r S ^ W ' S " "

^ y ^

=

'O-

M e ^ ^ ^ .

o isl ,^^^v^ HO

20 Incorporation of Acetate and Propionate Subunits Into Nargenicin A

19

methymycm.l7 A three stage sequence has been proposed for the overall biosynthetic pathway leading to the nargenicin macrolides, consisting of: (1) elaboration of an extended, hranched-chain fatty acid 9 from propionate-acetate condensations, (2) cyclizations to establish the cis-fused octalin and macrolide systems of 11, and (3) final oxidations and introduction of the C23 methyl and C9-0-acyl groups of 1 (Figure 4). Labeling studies with 180-Iabelled propionate-acetate precursors reveal that oxygen atoms at CI, C9, C l l and C17 are derived from propionate and acetate, while fermentation studies conducted in an ^^02 atmosphere indicate that the C2 and C18 oxygen substituents and the C8-C13 ether bridge of nargenicin Ai originate from molecular oxygen, an event that presumably occurs late in the biosynthetic pathway. While mechanistic details of the biosynthetic elaboration of tiie octalin and macrolide systems have yet to be defined, Cane has suggested that the nargenicin octalin nucleus may be generated by the intramolecular Diels-Alder cyclization of an oxygenated tetraene intermediate 10.^ Figure 4 COS-Enz

S-Enz

10 .OH

O2

11

1; nargenicin A^

288 4. Chemistry of the Naturally-Occurring Nargenicins The novel structural features of the nargenicins impart a unique chemistry to the macrolides. Not unexpectedly, the decenolide system is relatively sensitive and readily undergoes acid or base-catalyzed ring cleavage (Figure 5). Whaley and coworkers have reported that hydrolysis of nodusmicin by methanolic sodium hydroxide affords the corresponding seco acid 12; further degradation with sodium periodate affords acetaldehyde and aldehydic acid 13.3 Solvolysis of nargenicin Ai 1 in basic ethanol initially affords a ring-expanded lactone 14; at extended reaction times 14 undergoes lactone cleavage to yield ester 15.^ Alternatively, treatment of nargenicin Ai with acidic ethanol affords seco-tsitx 15 directiy. Figure 5

MeO.

2; nodusmicin

12

13 .O-CP

1; nargenicin A^ CP« 2-carboxypyrrole

Reagents: (a) IH NaOH, 1:1 H20:MeOH; (b) aq NaI04; (c) NaOEt, EtOH; (d) HCl, EtOH. The sensitivity of the nargenicin decenolide system to cleavage is fortunately not reflective of the ease with which the macrolide can be regenerated from seco acid derivatives. The presence of three sp^-hybridized carbons and the fusion to the rigid 1 l-oxatricyclo[4.4.llA0^''7]undecene nucleus impart considerable rigidity to the lactone system and reduces the number of degrees of rotational freedom available to seco derivatives. Steliou has demonstrated a facile, tin-mediated closure of the tetrahydropyranyl-protected nodusmicin seco acid 16 to lactone 17 (Figure 6)}^ Early synthetic studies by Magerlein and co-workers established that the rate of reactivity of the nargenicin hydroxyl substituents with a variety of electrophilic reagents follows the general trend C 1 8 > C 9 » C 1 1 . ^ The structural relationship between nodusmicin, 2, and nargenicin Ai, 1, was confirmed by silylation of the CIS hydroxyl group of 2, followed by selective acylation of the C9 hydroxyl of 18 and deprotection (Figure 7).^^ This strategy has been used to prepare a series of 9-C)-acyl analogs of nargenicin Ai from the C18-silylated intermediate 18; in general.

289 Figure 6 ,OTHP

a.b MeO.

Reagents: (a) 10 eq. dihydropyran, CSA, CH2Q2; (b) 1.1 cq KOH, 1:1 HiOiMeOH, reflux; (c) 1 eq Me2SnO, mesitylene, reflux. diminished antibiotic activity is observed for this series (Table 4). Finally, nodusmicin has been transformed to 18-deoxynargenicin Ai, 3, by thioacylation and reductive deoxygenation of thioimidazole 19 via the Barton protocol^ Figure 7 ,o-cp

CPa 2-carboxypynrolel

3; 18-deoxynargenicin A

Reagents: (a) rBuMe2Sia, imidazole, CH2CI2; (b) DCC, pym>le-2-carboxylic acid, pyridine; (c) aq BU4NF; (d) l,r-tiiiocarbonyldiimidazole, THF; (e) nBusSnH, THF. A notable aspect of the chemistry of the nargenicins is the remarkably inert C14-C15 trisubstituted olefin. In their original isolation smdies, Whaley and coworkersreportedthat while the C5-C6 olefin of 2 undergoes rapid hydrogenation, the C14-C15 olefin was uneffected by hydrogenating conditions, even at extended reaction times.3 Subsequent studies on the seco derivative 20 and tetraacetate 21 have revealed that the C14-C15 olefin of these compounds is similarly unreactive towards bromination and oxidation with a variety of reagents, including osmium and ruthenium tctroxides and ozone (Figure 8).^^ The lack of reactivity of the C14-C15 olefin is presumably a consequence of the extreme steric environment of this group, which is

290

Table 4.

Activity of 9-0-Acyl Esters of Nodusmicin Against S.

9-0-Sub^tituent

MIC (^g/ml)

H (nodusmicin, 2) Pyrrole-2'-carbonyl (nargenicin Al, 1) Pyrrole-3'-carbonyl Benzoyl Thiophene-2'-carbonyl Thiophene-3'-carbonyl Furan-2'-carbonyl Furan- 3 '-carbonyl Nicotinoyl Isonicotinoyl 4-MethylpyrTole-2'-carbonyl N-Methylpynole-2'-carbonyl L-Prolyl A.3'-L-Prolyl Pyrrole-2'-acetyl PyiTole-2'-acryloyl

125 0,125 0,39 >250 3,9 0,5 7,8 0.5 250 >250 0,78 >250 62,5 12.5 >100 >50

AureusA^

blocked from approach by external reagents by both the C4 and CI 1 substituents. In an attempt to mitigate the steric environment at C14-C15, iodolactone 24 was prepared from 21; however, the lack of reactivity of 24 towards oxidizing and other reagents parallels that of 20 and 21.^1 Figure 8

MeO

23a; R» H 23b; R« Ac

Reagents: (a) K2CO3, MeOH; (b) AciO, DMAP; (c) Br2, CCI4; (d) OSO4, THF; (e) I2, MeCN. As part of an effort to prepare compounds for correlation with advanced synthetic intermediates, Plata examined the stepwise degradation of the nargenicin C19-C14 macrolide subunit.^i Following the earlier work of Whaley,^ ester 20 was subjected to periodate oxidation and the remaining C9 and C l l hydroxyls protected as methoxymethyl ethers (Figure 9),

291 Hydrogenation of the C5-C6 olefin of 25 and oxidative decarbonylation afforded enone 26. Attempts to oxidize the C14-C15 olefin of 26 using a variety of reagents (O3, OSO4, MCPBA) were unsuccessful, as were efforts to generate a p-alkoxy ketone for an anticipated retro-aldol reaction by conjugate addition of oxygen nucleophiles to 26. Exposure of 26 to aqueous acid resulted in the selective cleavage of the Cll ether and intramolecular addition to the enone, yielding tetracyclic 27 as a single isomer of undetermined stereochemistry. Figure 9 PMOM

25 .OMOM

Reagents: (a) NaI04, THF, H2O; (b) MeOCHaCl, (iPr)2NEt; (c) H2,5% Pd/C, MeOH; (d) DABCO, O2, Cu(0Ac)2,2,2'-bipyridyl; (e) HQ, H2O. n. SYNTHETIC STUDIES OF THE NARGENICIN MACROLIDES The unique structural features of the nargenicin macrolides pose an intriguing synthetic challenge and several groups have recorded efforts directed at the total synthesis of these compounds.^2-24 Critical structural and stereochemical elements which must be addressed in an effective synthetic route to the nargenicins include: (1) the novel, highly-functionalized 11-oxatricyclo[4.4.1.l»^»'7]undecene nucleus, (2) the remote stereogenic centers at C2, CI6, C17 and CI 8 of the macrolide system, and (3) the presence of an acid/base sensitive decenolide ring. Steliou's lactonization of a protected 5ec
292 C8-C13 ether bridge spans two axial sites of the £ii-fused octalin system, and molecular mechanics calculations predict arelativelystrain-free tricyclic array. Based on these observations, our initial strategic considerations focused on approaches in which the nargcnicin ether bridge would be established by intramolecular addition of C8 or C13 alcohols to complementary electrophilic sites at CI3 or C8 (Figure 10). The diaxial orientation of the vicinal oxygens at C8 and C9 further suggested that intramolecular addition to a C8-C9 epoxide would serve as an efficient vehicle for tricycle construction with concurrent development of C9 stereochemistry. Figure 10 HO—A-/~OH

„ /

"^^

HO

In 1984, we reported the realization of this general strategy in the preparation a highlyfunctionalized ll-oxatricyclo[4.4.1.0]undecene model containing key structural and stereochemical elements of the nargenicin system.26 Initial construction of the requisite ds-fused decalin nucleus via Diels-Alder reaction of bcnzoquinonc and 1-trimethylsilyloxybutadiene afforded the crystalline enedionc 28; stereoselective reduction of the dionc and ketalization to effect differentiation of theresultingtriol yielded 29 (Figure 11). At this juncture, we examined Figure 11

MeaSiO

OPO(NMe2)2

Me

Me

Reagents: (a) PhH, 250; (b) DIBAL. PhH. 0°; (c) Me2C(OMe)2, PPTS; (d) PDC, CH2CI2; (e) Me2CuLi, Et20. -78 -> 0^; (0 ClPO(NMe2)2; (g) Na, THF, NH3, rBuOH.

293 strategies for introduction of the nargenicin CIO methyl substituent with concurrent olefin migration, a conceptually simple transformation which proved to be unexpectedly challenging. Oxidation of 29 afforded enone 30, which, when treated with lithium dimethylcuprate, yielded the anticipated 31, the product of exclusive a-face addition of the organocopper reagent. The intermediate enolate resulting from cuprate addition to 30 could be intercepted as the enol phosphoamidate 32; however, attempts to reduce this intermediate to the olefin via the Ireland protocol27 were thwarted by preferential reduction of the allylic ketal. Figure 12 ,NHTs

Me

34

1 b.c

31

1'

37

y

1' H 8

Me

°x° 35

38

36

Reagents: (a) LiAlH4, EtiO, (P; (b) MsCl, pyridine; (c) DBU, THF; (d) pTsNHNHa, EtOH; (e) 2.5 eq. MeU, EtiO, -78 -> 0^; (f) NaOMe, MeOH. Subsequent efforts to establish the required C8-C9 olefin focused on elimination of a suitably derivatized C8 alcohol (Figure 12). Hydride reduction of ketone 31 gave a single product, alcohol 34, which underwent mesylation/elimination to afford the undesired olefin regioisomer 35. We next examined the possibility of generating a C8-C9 olefin by means of tosylhydrazone elimination.28 Treatment of the tosylhydrazone derived from ketone 31 with methyl lithium afforded a C8-C9 olefinic product; however, proton NMR analysis suggested that the observed product was actually the trans-fused decalin, 36. In an effort to identify the origin of the undesired ring junction epimerization, ketone 31 was treated with sodium methoxide, resulting in a rapid conversion to the trans-fused ketone 38, the tosylhydrazone of which was identical to that obtainedfrom31. These results confirmed that epimerization of 31 occurs under the conditions of tosylhydrazone formation and that the tosylhydrazone formed from 31 was in fact the SMS-fused 37. We next considered introduction of the CIO methyl substituent via SN2' addition to a suitably-functionalized derivative of allylic alcohol 29. Treatment of acetate 39 with lithium dimethylcuprate afforded 40, the product of SN2 addition to the allylic system (Figure 13). In

294 contrast, mesylation of 29 and treatment of the sensitive mesylate 41 with Yamamoto's mixed cuprate reagent,29 followed by an acidic workup to effect deketalization, afforded an excellent yield of the desired diol 42. In retrospect, the failure of early efforts to generate a C8-C9 olefin may be attributable to the highly-strained nature of the target system, as evidenced by the fact that attempts to isolate the intermediate isopropylUdene from cuprate addition to 41 invariably resulted in decomposition of the product via ketal elimination. With diol 42 in hand, our goal was the elaboration of substrates with which to examine ether bridge formation. Allylic oxidation of 42, followed by protection of the C l l hydroxyl as the methoxymethyl ether gave enone 43. Epoxidation yielded the desired a-44, accompanied by the diastereomeric p-44. Figure 13 OAc

Me

40

94% from

29

Me HO ^ OH

42

e,

•'I

H

57%

,0

a?.-cp 8"

Me O " OMOM

u

O

OMOM

a-44

86%

Me O " OMOM 43

P-44 Reagents: (a) AC2O, pyridine; (b) MeiCuLi, Et20, -780; (c) nBuLi, MsCl, Et20, OO; (d) MeCu-BF3Et20, O^, a q N H 4 a ; (e) Mn02, CH2CI2; (f) MeOCH2Cl,EtN(iPr)2; (g)MCPBA,CH2Cl2 0o. The addition of simple organometallic reagents to a-44 was investigated as a prelude to incorporation of more advanced nucleophiles (Figure 14). Not surprisingly, the outcome of nucleophilic addition to a-44 was highly sensitive to the nature of the reagent. Treatment with excess methyl lithium, followed by low temperature workup, gave as the only isolated product tertiary alcohol 45. In contrast, addition of methyl Grignard reagent to a-44 was accompanied by a secondary, intramolecular addition of the initially formed magnesium alkoxide to the C8-C9 epoxide, to give tricyclic ether 46 in a single step. The magnesium cation clearly plays a critical role as a Lewis acid in the intramolecular epoxide opening, as demonstrated by the rapid closure of 45 to 46 upon treatment with an equivalent of Grignard reagent.

295 Figure 14 b y -.0 Me H O •' OMOM a-44 Coupling constants for 46 (Hz) J7.8 Ja.g *^9,io Jio.ii ^11.12 J7.12

« 0 *• 5.3 " 5.0 =11.5 * 3.0 « 0

"12 O R

Reagents: (a) 1.5 eq. MeLi, Et20, -78 -> O®, then aq NH4a; (b) MeMgBr, THF. -78 -> (P. Figure 15 O^ ^OMOM 'Me

.OMOM

Reagents: (a) 1.1 eqLiC(OEt)=CH2. MgBr2Et20, THF, -78^; (b) MeOCH2Cl, EtN(iPr)2; (c)pTsOH, MeOH; (d) NBS, CCI4, (PhC02)2; (e) Me2CuLi, THF, -780; (f) MeMgBr, Et20, (P; (g) BusSnH, AIBN, ethyl acrylate; (h)pTsOH, allyl alcohol; (i)rtBuaSnH,AIBN, PhH.

296 Our preliminary studies having established a reliable body of chemistry for stereocontroUed elaboration of the nargenicin ll-oxatricyclo[4.4.1.0]undecene ring system, we tumed our attention to the examination of strategies for introduction of C4 and CI3 substituents that would ultimately become the macrolide system of the antibiotics.30 Incorporation of a latent acetyl fragment into the oxa-bridged tricyclic nucleus was readily accomplished by treatment of ketone a-44 with cthoxyvinyllithium in the presence of magnesium salts, which yielded the tricyclic 47 (Figure 15) Exposure of 47 to acidic methanol in an attempt to transform the sensitive vinyl ether into a more stable ketal derivative resulted in an unexpected cleavage of the C l l methoxymethyl ether and fomiation of the tetracyclic 48. Figure 16

X-ray crystal structure of bromide 49 (selected hydrogen atoms have been omitted for clarity)

We next examined protocols for introduction of a three-carbon substiment representing the C1-C3 subunit of the nargenicins. Stereoselective bromination of 48 was accompanied by migration of the olefinic group to afford, as the only observed product, allylic bromide 49, the structure of which was unambiguously established by X-ray crystallography (Figure 16).^l Initial investigations of the coupling of 49 with organometallic reagents were not encouraging. Treatment of 49 with lithium dimethylcuprate yielded only the SN2* adduct 50; addition of methyl Grignard reagent proceeded in the desired SN mode but afforded an equimolar mixture of epimeric adducts 51. Attempts to introduce the requisite three-carbon fragment via capture of an allylic radical^^ were also disappointing. Debromination of 49 in the presence of ethyl acrylate gave only the reduced 52, indicating that bimolecular addition reactions of the allylic radical derived from 49 were not competitive with H atom scavenging. To enhance the coupling probability, allyl ketal 53 was prepared; however, attempts to induce cyclization of 53 via allyl radical formation yielded the product of simple reductive dehalogenation, olefin 54. Nucleophilic introduction of a C4 substituent proved to be more productive. Oxidation of olefin 48 via the Salmond procedure^^ gave enone 55, which underwent stereoselective addition by organocuprate reagents to yield only p adducts (Figure 17). Addition of the mixed cyanocuprate derived from ethoxyethyl-protected 3-bromopropanol to 55, followed by enolate phosphorylation and reduction^^ afforded 56, incorporating the desired C1-C3 subunit. The stereochemistry of addition was confirmed by comparison of the allyl derivative 59 with material prepared by an alternate route, consisting of: (1) reduction of 55 to alcohol 57, (2) transformation of 57 to ester 58 via orthoester Claisen rearrangement, and (3) homologation to 59 via a standard

297

b.C

EEO.

53%

HOo OMOM

Me—I " O OMe 58

59

Cu--Li.

Reagents: (a) Cr03-3,5-dimcthylpyrazole, CH2CI2, -20O; (b) LiCu(CN)(CH2)30CH2(Me)OEt, THF, -780, then (NMe2)2POa; (c) Na, NH3, THF, rBuOH; (d) LiEtsBH, THF; (e) (EtOsCMe, TsOH, xylene,reflux;(0 UAIEU, Et20,0®; (g) PDC, CH2CI2; (h) Ph3P=CH2, THF; (i) aq HQ, THF; (j) MsQ, pyridine, then rBuOK, DMSO. Wittig olefination sequence. The highly stereoselective p-face addition of organocuprates to 55 is remarkable considering the steric demands typically associated with such reagents, and is presumably a consequence of directed addition to the enone by precoordination of the organocupratereagentto the C8-C13 ether bridge, to give an intennediate complex A. Figure 18 OMOM EEO.

OMe

Reagents: (a) Cr03, aq H2SO4, acetone, -2(F; (b) CH2N2, Et20; (c) LDA, MoOPH, THF, (P; (d) KH, Mel.

298 Of the procedures examined for introduction of the C2-methoxy substituent, the most successful proved to be enolate oxidation as described by Vedjas.34 Deprotection of 56, oxidation and esterification gave 60. Treatment of the enolate derived from 60 with M0O5HMPA-pyridine complex and methylation of the resulting hydroxyesters afforded 61, comprising the fully-functionalized "northern half' of the nargenicins, as a mixture of C2 cpimers. While the stereoselectivity of the C2 oxidation protocol was far from optimal, the sequence leading to 61 demonstrated the basic chemistry necessary for the synthesis of naturally-occuiring nargenicins. 2. Total Synthesis of 18-Deoxynargenicin Ai A logical extension of our model studies was direct nucleophilic introduction of the nargenicin C14-C19 subunit by addition of a suitably-functionalized vinyl nucleophile to ketone a-44. In 1988, we recorded the experimental realization of this strategy in the context of the first total synthesis of a naturally-occurring nargenicin macrolide, (+)-18-deoxynargenicin Ai, 3 (Figure 19).35 Our synthesis of a precursor to the nargenicin C14-C19 subunit, 66, derives from Corey's preparation^^ of the analogous silyl-protected olefin. Addition of lithium acetylide to optically active epoxide 62 afforded diol 63 which was homologated by coupling of the derived tosylate with lithium dimethylcuprate. Protection of alcohol 64 and alkylation of the terminal alkyne furnished 65; hydrozirconation and halogenation of the intermediate vinyl zirconium species gave the desired iodide 66. Figure 19 HO

e,f HO

O^^^^^OMe

HO

62 HO

Me

MOMO

Me

64

Me

63 Me

Me

65

Me

MOMO

Me

Me

Me

66

Reagents: (a) Ti(iPrO)4, (-)-DIPT, rBuOOH, -20O; (b) 2,2-dimethoxypropane, POCI3; (c) Li-CCH, DMSO; (d) Amberlite IR-120, MeOH, H2O; (e)/7TsCl, pyridine; (f) LiCuMe2, Et20,0°; (g) MeOCHiCl, iPriNEt; (h) LDA, Mel, THDF, -780; (i) Cp2ZrHCl, PhH, then I2. Addition of the reagent derived from lithiation of optically active iodide 66 to the previously prepared, racemic ketone a-44 gave the expected 1:1 mixture of diastereomeric adducts 61 and 68. In contrast to our earlier model studies, intramolecular epoxide opening proceeded to completion without the addition of a Lewis acid (Figure 20). While diastereomers 61 and 68 were readily separable by chromatography, an unequivocal determination of relative stereochemistry by spectroscopic methods was not possible. Each isomer was therefore carried

299

OMOM

MOMO.

OMOM

PMOM

Me HOzO

MOMO. Me

I d,€ OMOM

^Me

74a:p-OMe 74b: a-OMe

U - 3; R= 2-carboxypyrrole

Reagents: (a) 66,2 eq ©uLi, -78°; (b) MeOCHgCI, iPrgNEt; (c) CrOa. 2,5-dimethylpyra2ole, CH2CI2; (d) LiCu(CN)(CH2)30CH(OEt)Me, THF. -78**. then (NMegJzPOCI; (e) Li. NH3, THF, ©uOH; (f) CrOa. HzO-acetone; (g) CHgNg. EtgO; (h) LDA. MoOPH, THF. -78**; (i) AggO. Mel. DMF; (i) aq HCI; (k) 2-thiopyridyldisulfide, PhaP, xylene; (I) pyrrole-2-carboxylic acid, DCC, THF.

300 separately through our sequence for introduction of the C1-C3 sidechain. Fortunately, we were able to obtain suitable crystals of intermediate 70, derived from adduct 67, for characterization by X-ray crystallography.31 The X-ray structure revealed acid 70 to be an intermediate from the "unnatural" diastereomer series (Figure 21). Thus assured of the relative stereochemistry of adducts 67 and 68, we focused our attention on further synthetic development of the latter. Figure 21

X-ray crystal structure of acid 70 ("unnatural" diastereomer series). Hydrogen atoms have been omitted for clarity.

Oxidation of 68 to the cnone 71 and addition of a three-carbon C1-C3 subunit using our previously established scheme afforded 72, comprising the completed carbon framework of the nargenicins. Oxidation and csterification gave 73, the enolatc of which oxidized and alkylated to yield a mixture of C2-mcthoxy cpimers. Given the sensitivity of the nargenicin decenolide system to acid, we elected at this juncture to deprotect the C9, C l l and C17 hydroxy 1 groups to give acid triol 74. While interference with the desired macrocyclization by cither the C9 or CI 1 hydroxyl seemed unlikely, we nevertheless required cyclization conditions compatible with free hydroxyl groups. The Corey "double activation" protocol3*7 proved satisfactory in this regard; surprisingly, conversion of 74 to the thiopyridyl ester followed by thermolysis in refluxing xylene afforded a single lactonic product 75, in addition to a mixture of recovered acids 74a and 74b which was enriched in the latter compound. The reluctance of seco acid 74b to undergo lactonization is presumably a consequence of a transannular steric interactions (Figure 22) which preclude cyclization. A steric compression between the C2 and C14 substituents of thioester £ destablizes this conformation relative to that of the epimeric fi, which readily undergoes lactonization to 75. Figure 22 O

JU/^°Me^ 74a

74b

—*-

C

301 That decenolide 75 was in fact 18-deoxynodusmicin was confirmed by acylation at C9 with pyrrole-2-carboxylic acid under Magerlein's conditions'^ to give 18-deoxynargenicin Ai which was identical to an authentic sample of 3. The optical rotation of synthetic 3 was in agreement with that of the natural macrolide, providing experimental verification of the proposed absolute configurations of nargenicin Ai, 1, and nodusmicin, 2.^ The above scheme represents the first, and to date only, recorded preparation of a naturally occurring nargenicin macrolide. 3. A Second Generation Synthesis of 18-Deoxynargenicin Ai. Our synthesis of 18-deoxynargenicin Ai 3 demonstrated a short, convergent assembly of the nargenicin carbon skeleton and established a reliable foundation for further synthetic studies. A number of key stereochemical issues remained unresolved from this effort, the most critical of which was the lack of stereocontrol at C2 and C16-C18 of the macrolide system. In a second approach to the nargenicins, we sought to examine a stereorational alternative for development of the macrolide system which would provide for control of C16-C18 stereochemistry and furnish a functionalized templatefromwhich C18-oxygenated nargenicins could be elaborated.^^ Figure 23

^A^'

Me

.OR' LDAorBuU

MeaSiCI

THF. -78*0

-78->0'*

CpzSiMea MOMO lOMO

OR*

MOMO

76

"X

Me

LOAorBuLi

Me

OMe

MOMO Me

77

C ^

/

Me

THF.-78'»C WOMO MOMO

Me

H

OH

CH,

Our "second generation" approach to the nargenicins was based on the well-documented ability of [3,3] and [2,3] sigmatropic rearrangements to translate stereochemical information along a nascent acyclic framework. Of particular relevance to the problem at hand were two diastereoselective sigmatropic variants that had been the focus of recent studies in our laboratories: the [3,3] Claisen rearrangement of glycolate esters 76,^9 and the [2,3] Wittig rearrangement of tertiary allylic ethers 77,^0 a reaction that provides a uniquely stereoselective entry to highlyfunctionalized, E-trisubstituted olefins, a key structural unit of the nargenicin macrolide system and of other polypropionate-derived natural products (Figure 23).^^ Our analysis of the nargenicin macrolide system suggested that the [2,3] Wittig rearrangement of tertiary allylic ether 83 would deliver the correct C16-C17 stereochemistry. Towards this goal, a careful hydrolysis of enol ether 47 yielded die corresponding methyl ketone 78 (Figure 24). While the diastereofacial preference for nucleophilic addition to ketone 78 was initially unclear, we envisioned the a-chelation between the ketone carbonyl and the ether bridge of n as the most probable scenario for addition (Figure 25). Consistent with this expectation.

302

Figure 24 ,OMOM

OMOM

x-ray crystal structure of 79

KH, DME

—i-N, CI 82

.OMOM

OMOM

83

Reagents: (a) 1 M aq HCl, THF; (b) MeCC-MgBr, EtiC -780; (c) LDA, THF, MeCHO, -780; (d) MsCl, NEt3, THF; (e) DBU. THF; (0 MeLi, Et20, -780; (g) LDA, THF, -78o treatment of 78 with propynyl Grignard reagent afforded the crystalline alcohol 79, the structure of which was confirmed by X-ray crystallography."^^ Figure 25

mucj —•^ B

303 Preparation of the desired 81 thus required a net inversion of the tertiary alcohol center of 79, a goal that could be accomplished by reversing the order of substituent introduction. Ketone 78 was transformed to the a,|3-unsaturated 80 by aldol condensation and elimination. Addition of methyl Grignard reagent to 80 resulted in preferential 1,4 addition, reflecting the hindered environment of the C14 ketone; in contrast, the 1,2 adduct, 81, was obtained as the major product from treatment of 80 with methyl lithium. With the tertiary allylic substrate 81 in hand, sigmatropic homologation of the nargenicin sidechain was effected by 0-alkylation with chloromethyloxazoline 82 to give ether 83, followed by [2,3] rearrangement of the derived lithium enolate to give oxazoline 84. That we had indeed established the desired C14-C15 olefin geometry and stereochemistry at C16 and C17 was confirmed by X-ray crystallographic analysis of 84 (Figure 26).^^ The diastereoselectivity of the [2,3] Wittig rearrangement of 83 is consistent with that observed for related acyclic substrates,^^*^^ suggesting that a multidentate complex £ of the enamide counterion serves as the control element for the sigmatropic event Figure 26

[2.3]

H*

X-ray crystal structure of oxazoiine 84

Hydrolysis of oxazoline 84 and esterification of the resulting acid affords 85, an advanced intermediate which should prove useful for synthetic studies directed at the CISoxygenated nargenicin congeners. Intersection with our earlier scheme leading to 18deoxynargenicin Ai was accomplished by protection of 85 and reduction to the alcohol 86, whereupon tosylation and homologation by lithium dimethylcuprate displacement afforded 68, an intermediate in our previously described synthesis of 3 (Figure 27).

304 Figure 27 0

0

,,.OMOM

a.b

"°Y^Me O^N

c,d

MON/

"•^Y^Me COgMe

V- «

L

85

18-deoxynargenlcin A^

]-

86 e,f

Figure 20 68

Reagents: (a) 10 cq McI, NaOH, H2O, MeOH; (b) CH2N2, Et20; (c) MeOCH2Cl, (iPr)2NEt; (d) LiAlH4, Et20,00; (e)pTsCl, pyridine; (0 Me2CuLi, THF, -78o. Our "second generation" linear route to 3, in which remote chirality at C16 and C17 of the nargenicin macrolide system is introduced by a stereoselective sigmatropic event, offers a considerable advantage in efficiency over our initial convergent sequence. Of particular significance is the potential for further stereocontroled development of the CI8 nargenicins from advanced intermediates such as 85 and 86. 4. Synthetic Studies Directed at Coloradocin. In contrast to the decenolide nargenicins, the ^w-macrolide coloradocin, 8, has received relatively littie synthetic attention. The demanding structural and stereochemical features of the nargenicins are amplified in the biS'la.ctomc coloradocin system, which additionally presents a number of unique considerations, including oxidation at C19 and C22, a enol ether linkage in the 14-membered macrocycle and an extended maleic anhydride chromophore. These challenging structural elements, the chemical sensitivity of the macrolide and a number of undefined stereogenic centers combine to make coloradocin a formidable synthetic target Not unexpectedly, coloradocin is a difficult compound to handle. Chromatographic purification of the macrolide is accompanied by significant material loss, and pure samples of the macrolide slowly decompose to uncharacterized products.'*^'^^ xhe chemical sensitivity of coloradocin has complicated degradation studies to some extent. One notable observation is that alkaline solvolysis of the maleic anhydride system of 3 results in an equilibrium between esters 87 and the parent anhydride, suggesting that an analogous anhydride formation could constitute the penultimate step in a synthetic sequence leading to coloradocin (Figure 28).'^^ The synthetic challenges noted above notwithstanding, we have recendy initiated synthetic studies directed at coloradocin and have developed an effective approach to the oxa-bridged octalin nucleus of 3. The unusual ll-oxatncyclo[5.3.1.l'^03«8]undecene ring system of coloradocin is unusual, but not unknown; at least two other natural products, the sesquiterpene

305 Figure 28

NaOMe MeOH

MeO

f JL

O

(^ 1

OH Me n

HCis

87a; Ri= C02Me, Ra* CO2H 87b; Ri= CO2H, R2= COgMe

8; coloradocin

crytofaurDnol^*^ and a diterpene, chantacin,^^ incorporate a similar framework. Our analysis of the coloradocin system parallels that employed for the nargenicins, in that two complementary strategies for ether bridge fomiation were envisioned (Figure 29). In one approach, nucleophilic displacement of a C8 leaving group by the C13 alcohol or alkoxide moiety of £ would establish the tricyclic undecene of 3; alternatively, intramolecular capture of a C13 cationic intermediate fi by a C8 alcohol would generate the requisite ether bridge Hnkage. Figure 29 HO

<^ i

^ ^'^

HO

'^

Our initial efforts were directed at the first of these approaches and focused on the preparation of a suitably-functionalized equivalent of E-^^ Towards this end, we required a stereocontroUed entry to oxirane p-44, the minor product obtained from epoxidation of 43. Oxidation of diol 42, an intermediate from our original nargenicin sequence, afforded the sensitive ketol 88, whereupon directed epoxidation gave the desired |J-44 as the exclusive product. Addition of ethoxyvinyllithium to P-44 yielded tertiary alcohol 89, which upon treatment with LiAlH4 underwent regioselective oxirane cleavage to give the C9 alcohol 90. We anticipated that mesylation of 90 and exposure to iodide would result in a net inversion of C9 conHguration; subsequent intramolecular SH2 displacement of the C9 iodide would establish a C9-C13 ether linkage. Unexpectedly, the major product of this sequence was olefin 93, presumably formed from the elimination of iodide 91. In retrospect, the close proximity of the Cll proton to the CI3 alkoxy substituent no doubt contributes to the facility of this elimination. Attempts to salvage the scheme by generation of a C9 cationic species from olefin 93 were uniformly unsuccessful; for example, alkoxymercuration of 93 and reduction of the resulting organomercurial afforded only the tricyclic 94, the product of intramolecular addition to the more stable CIO cation.50

306 Figure 30

Reagents: (a) MnOi, CH2a2; (b) («PiO)3V0, rBuOOH, CH2a2; (c) MeOCH2a, /PrzNEt; (d) LiC(0Et)=CH2, TEiF. -780; (e) LiAlH4, Et20; (f) aq HQ; (g) MsCl, pyridine; (g) nBu4NI, benzene; (h) Hg(OAc)2, CH2CI2; (i) /iBusSnH, ATON, toluene. Another intermediate from our earlier nargenicin model studies, ether-bridged octalin 46, provided an attractive starting point for an examination of our second approach to the coloradocin tricyclic nucleus.^9 Oxidation of 46 furnished ketone 95, which underwent reductive cleavage of the 0-C8 bridge linkage upon treatment with samarium diiodide to give a 3:1 equilibrium mixture of ketone 96 and the corresponding hemiketal 97. Reduction and acetylation of the C9 alcohol afforded 98, from which diene 99 was prepared by acid-catalyzed dehydration. Subsequent studies led to the discovery of a more efficient, three step route to 99 from epoxide p-44. Installation of the diene system by Peterson olefination of P-44 gave diene 100, whereupon regioselective hydride addition to the C8-C9 epoxide yielded 99. Having secured a reliable entry to diene 99, we examined procedures for intramolecular addition to establish the C9-C13 ether bridge. Oxymercuration of 99 by treatment with mercury (II) trifluoroacetate afforded a poorly characterized organomercurial product, the reduction of which regenerated the starting diene 99. Selenoetherification of 99 was somewhat more successful, providing a low yield of the tricyclic selenide 101; the major product from this reaction appears to be that resulting fix>m 1,4 addition of HCl to the diene system of 99. Attempts to suppress this undesired side reaction by buffering the reaction medium resulted in recovery of starting diene. Final reductive deselenation was accomplished by exposure of 101 to tributyltin hydride to give the ether-bridged octalin 102, incorporating the key structural and stereochemical elements of the coloradocin 11-oxatricycloundecene nucleus.

307

Figure 31

4iC^

.OH OH

Me OMOM

Me

46

OMOM

Me

95

c,d

H

5

Me OMOM

Me

96 .OH

e.c

H

Me OMOM

97

h.l

• ^

t^'-'Me

Me OMOM

Me OMOM 99

98

101;R=SePh 102;R=:H

tg O

Me OMOM

p-44 Reagents: (a) (a) PDC, CH2CI2; (b) Snil2, THF; (c) LiAlH4, Et20; (d) AciO, THF, DMAP; (e)pTsOH, benzene; (f) Me3SiCH2Li, THF, -780; (g) KH, THF; (i) PhSeCl, THF; (j) /iBusSnH, AIBN, toluene, 1 lO^. In the absence of further stereochemical details regarding the coloradocin oxa-bridged system, we assumed at the outset of our model studies that the disposition of octalin substituents for 8 would parallel those of the structurally-related nargenicins. This assumption was substantiated, as hydrogenation of 102 afforded a saturated analog 103 with spectroscopic properties closely resembling those of coloradocin. Of particular note are chemical shifts and coupling constants for the C9 and CIO protons of 103, which are in good agreement with values reported for the parent macrolide 8,1^ an observation that suggests a C9-C10 stereochemical correspondence between coloradocin and the simple decenolide nargenicins. Figure 32 O

H H2, Pd-C Me H Me OMOM 102

MeOH

LI

H Me H Me OMOM 103

^>4!>2/-^ ""^

OMOM

Jioii(103)«7.2Hz J10.11 (8) « 7.5 Hz

308 Conclusion. Considerable progress has been recorded towards the synthesis of naturally occiming nargenicins, and an efficient, reliable and preparadvely useful strategy for development of nargenicin carbon frameworks has emerged. Key issues in this area that remain to be explored include (1) effective, asymmetric entries to the oxa-bridged nargenicin octalin systems, (2) reliable procedures for control of macrocycle chirality, especially the C2 stereogenic center, (3) efficient strategies for addressing the synthesis of highly-oxygenated nargenicin congeners such as the B series of nargenicins and coloradocin. Acknowledgements. The author wishes to thank Dr. Walter Celmer (Pfizer), Dr. Howard Whaley (Upjohn) and Dr. James McAlpinc (Abbott) for stimulating and informative discussions and for supplying biological activity data and generous gifts of naturally-occurring nargenicins. The synthetic efforts cited above represent the experimental and intellectual contributions of a group of extremely talented Syracuse University graduate students and postdoctoral associates, including Drs. Daniel Plata, Lucius Rossano, Simon J. Coutts and Mr. Jeffrey Evans. The collective enthusiasm and commitment of these individuals to the nargenicin program has made our collaboration truly memorable.

REFERENCES 1. W. D. Celmer, W. P. Cullcn, C. E. Moppctt, M. T. Jefferson, L. H. Huang, R. Shibakawa and J. Tone, U.S. Patent 4,148,883 (to Pfizer, Inc.) August 18, 1977. 2. W. D. Celmer, G. N. Chmumy, C. E. Moppett, R. S. Ware, P. C. Watts and E. B.Whipple, J. Am. Chem. Soc. 1Q2 (1980) 4203. 3. H. A. Whaley, C. G. Chidester, S. A. Mizsak and R. J. Wnuk, Tetrahedron Letts. 21 (1980) 3659. 4. B. J. Magerlcin, Abstract # MEDI72fromthe 182nd National Meeting of the American Chemical Society, New York, NY, August 23-28,1981. 5. N. Harada and K. Nakanishi, in Circular Dichroic Spectroscopy. Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA (1983). 6. D. E. Cane and C.-C. Yang, J. Antibiot. IS (1985) 423. 7. H. A. Whaley and J. H. Coates, Abstract #187 from the 21st Interscience Conference of Antimicrobial Agents and (Chemotherapy, November 4,1981. 8. B. J. Magerlein and R. J. Reid, J. Antibiot. 31 (1982) 254. 9. W. D. Celmer, W. P Cullen, R. Shibakawa and J. Tone (to Pfizer Inc.) U. S. Patent 4,436,747; see Chem. Abstracts IQl (1984) 53332. 10. J. Tone, R. Shibakawa, H. Maeda, Y. Yamauchi, K. Niki, M. Saito, K. Tsukuda, E. B. Whipple, P. C. Watts, C. E. Moppett, M. T. Jefferson, L. H. Huang, W. P. Cullen and W. D. Celmer, Abstract #62fromthe 20th Interscience Conference of Antimicrobial Agents and Chemotherapy, September 22,1980.

309 11. M. Jackson, J. P. Karwowski, R. J. Theriault, P. B. Femandes, R. C. Semon and W. H. Kohl, J. Antibiot. 4Q (1987) 1375. 12. R. R. Rasmussen, M. H. Scherr, D. N. Whittem, A. M. Buko and J. B. McAlpine, J. Antibiot. 4Q (1987) 1383. 13. S. Omura, R. Iwata, Y. Iwai, S. Taga, Y. Tanaka and H. Tomoda, J. Antibiot. 3S (1985) 1323. 14. H, Tomoda, R. Iwata, Y. Takahashi, Y. Iwai, R. Oiwa and S. Omura, J. Antibiot. 32 (1985) 1205. 15. D. E. Cane and C.-C, Yang, J. Am. Chem. Soc 1Q6 (1984) 784. 16. W. C. Snyder and K. L. Rinehart, J. Am. Chem. Soc. IQfi (1984) 787. 17. D. E. Cane, W. Tan and W. R. Ott, J. Am. Chem. Soc. 115. (1993) 527. 18. K. Steliou and M.-A. Poupart, J. Am. Chem. Soc. IQS (1983) 7130. 19. B. J. Magerlein and S. A. Mizsak, J. Antibiot. 2S (1982) 112. 20. J. Kalknerten, Syracuse University, unpublished results. 21. D. J. Plata, Ph.D. dissertation, Syracuse University (1987). 22. R. C. F. Jones and J. H. Tunnicliffe, Tetrahedron Letts. 2fi (1985) 5845. 23. W. R. Roush and J. W. Coe, Tetrahedron Letts. 2£ (1987) 931. 24. J. W. Coe and W. R. Roush, J. Org. Chem. 54 (1989) 915. 25. T. Mukaiyama, Angew. Chem. Int. Ed. Engl. 1& (1979) 707, and references therein. 26. J. Kallmerten, Tetrahedron Letts. 21 (1984) 2843. 27. R. E. Ireland, D, C. Muchmore and U. Hentgartner, J. Am. Chem. Soc 24, (1972) 5098. 28. R. H. Shapiro and M. J. Heath, J. Am. Chem. Soc. S2 (1967) 5734. 29. Y. Yamamoto and K. Maruyama, J. Am. Chem. Soc 22 (1977) 8068. 30. J. KaUmerten and D. J. Plata, Heterocycles 25 (1987) 145. 31. C. E. Pfluger, D. J. Plata and J. Kallmerten, Syracuse University, unpublished results. 32. M. D. Bachi and C. Hoomaert, Tetrahedron Letts. 22 (1981) 2689. 33. W. G. Salmond, M. A. Barta and J. L. Havens, J. Org Chem. 42 (1978) 2057. 34. E. Vedjas and J. E. Telschow, J. Org Chem. 41 (1976) 740. 35. J. Kallmerten and D. J. Plata, J. Am. Chem. Soc. Ufl (1988) 4041. 36. E, J. Corey, E. J. Trybulski, L. S. Melvin, K. C. Nicolau, J. A. Secrist, R. Lett, P. W. Sheldrake, J. R. Falck, D. J. Brunelle, M. F. Haslanger, S. Kim and S. Yoo, J. Am. Chem. Soc IQQ, (1978) 4618.

310 37. E. J. Corey and K. C. Nicolau, J. Am. Chem. Soc. 26 (1974) 5614. 38. L. T. Rossano, D. J. Plata and J. Kallmerten, J. Org Chem. 52 (1988) 5189. 39. T. J. Gould, M. Balestra, M. D. Wittman, J. A. Gary, L. T. Rossano and J. Kallmerten, J. Org. Chem. 52 (1987) 3889. 40. J. Kallmerten and M. D. Wittman, J. Org. Chem. 51 (1988) 4631. 41. J. Kallmerten and M. D. Wittman, Studies in Natural Products Chemistry, Vol. J, A. urRahman, ed., Elsevier, Amsterdam (1989) pp. 233-285. 42. C E. Pfluger, J. Kallmerten and D. J. Plata, Acta. Cryst. £45 (1989) 1031. 43. R. Ostrander, J. Kallmerten and L. T. Rossano, Acta. Cryst. £41 (1991) 2410. 44. R. H. Chen, J. E. Hochlowski, J. B. McAlpine and R. R. Rasmussen, J. Liq. Chrom. 11(1988)191. 45. R. R. Rasmussen and M. H. Scherr, J. Chrom. 2Sfi (1987) 325. 46. S. J. Coutts, J. M. Evans and J. Kallmerten, unpublished results. 47. H. Hikino, Y. Takeshita, Y. Hikino and T. Takemoto, Chem. Pharm Bull. 14 (1966) 735. 48. M. Sugano, T. Shindo, A. Sato, Y. lijima, T. Oshima, H. Kuwano and T. Hata, J. Org. Chem. 55 (1990) 5803. 49. J. M. Evans and J. Kalhnerten, Synlett (1992) 269. 50. L. T. Rossano, Ph.D. dissertation, Syracuse University (1990).

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

311

Some Aspects of the Chemistry of Lignans R. Stevenson

1.

INTRODUCTION The term lignan(e) was introduced by R.D. Haworth in 1936 in a review article

dealing with natural resins (1). He noted the widespread natural occurrence of "propylbenzenederivatives" (e.g., safrole, eugenol, caffeic acid), designated asQ-Cj units, and recognized the existence of a group of phenolic resinols whose structures might be formally derived from two such QCj units. For the latter, of which all known at that time possessed a bond between the j8-carbon atoms of the C3 chain, the general name lignan has become established.

Cpf C5-C3 Unit

Lignan Carbon Framework

Many related natural products have subsequently been identified with the significant structural difference of having the C5-C3 units bonded at sites other than iS^S'. For these, the useful designation ''neolignan'' was introduced (2). Ambiguity has since arisen however from use of the same term with a conflicting definition, based upon possible biogenetic rather than structure considerations (3). Most workers in this field, nevertheless, express a preference for the original Haworth definition of lignan, and this article is restricted to that understanding. The first significant reviews of lignans appeared in 1955, with that of Erdtman (4) having particular significance in

312

laying the foundation of consideration of the biogenesis of these products, and that of Hearon and MacGregor (5) offering comprehensive coverage (314 references) of the chemistry of the 33 members known at that time. Later reviews emphasizing systematics and nomenclature followed (6,7). The first extensive multi-authored monograph was published in 1978 (8) and a volume surveying the chemical, biological and clinical properties has recently become available (9). Recognition of the increasing significance of biological properties and potential pharmacological applications (10-13) has further stimulated interest in this field, with timely literature reviews by Whiting (14). A review devoted to the synthesis of lignans (15) has additionally been up-dated with emphasis on asymmetric methodology (16). An accurate accounting of the total number of natural lignans of well-defined structure takes on an element of a Sisyphean task since scarcely a month passes without reports of newly isolated products, particularly in the pages of Phytochemistry and The Journal of Natural Products, Useful compilations, although inevitably out-dated, do however exist (13,17,18). Suffice it to say that, within the Haworth lignan definition and excluding natural glycoside and ester variants, about 500 natural compounds are known. With few exceptions, these can be conveniently and unambiguously sub-classified within the structural sub-classes [A-J, Scheme 1] corresponding to A B C D E F G H I J

Dibenzylbutanes Dibenzylbutyrolactones Furans Tetrahydrofurans (in three sub-groups, D(a), D(b), and D(c)) Arylnaphthalenes Aryldihydronaphthalenes Aryltetralins Tetrahydrofurofurans Dibenzocyclo-octadienes Diarylcyclobutanes

The aryl group substitution patterns found in lignans are delineated in Scheme 2. The rings may be mono-, di-, tri-, or tetrasubstituted by the common phenolic hydroxy 1, methyl ether and methylenedioxy groups. With a few exceptions (isolated notably from mammalian sources), all lignans have an oxygenated substituent/?am

313

to the C3 sidechain precursor unit. This survey is organized to provide a balanced overview of the above ten structural sub-classes (A-J). The choice of topics is of necessity both subjective and selective, but includes within each group representative members which have been of significance in the development of the field or isolated from natural sources with demonstrated biological activity. Emphasis is also given to synthetic methods

S£ll£iTi£l. Lignan Structure Sub-Classification

?

•0: Ar'

"TJt

A

B

O'

O

D(a)

D(b)

ir^-^r^y^Ar'

D(c)

Ar

AT

E .0^

K

Ar

o

H

V-/ /—^ Ar

Ar-

314 Scheme 2. Substitution on C5-C3 Unit Mono-

XT

HO,

jy

-cr

Di-

"XT

HO

"Xr

MeO'

H O - ^

OMe

HQ

<^i

MeO'

HO'

TriHO

McO,

MeO'

MeO'

<

"^6 OH MeO.

HO

V_(^

V,6 Tetra-

OMe

<

O

r <

<xi:.

Br

315

which appear to have general applicability, unless they have been extensively and recently reviewed elsewhere. Since this article is directed to the non-specialist, questions of systematic nomenclature are avoided, in the interest of simplicity and clarity within each group, and since no comprehensive system yet appears to have gained general approval (19). A.

DIBENZYLBUTANES When last reviewed as a lignan group fifteen years ago (20), there were 17

members of reasonably well-established structure; that number has now more than tripled. The most venerable is (-)-guaiaretic acid (1), isolated from the gum and heartwood of "lignum vitae." Of particular significance was the establishment of the configuration of the sole chiral centre (3R), initially by chemical correlation with (-)3,4-dihydroxyphenylalanine of known absolute configuration (21). It has also served as a useful standard for chemical correlation of natural members of other lignan structural sub-classes. MeO,

OH

M^.so-nordihydroguaiaretic acid (2) occurs in high concentrations in the resinous exudate of Larrea divaricata (the creosote bush of south-western U.S.A. and Mexico). Recognized about 50 years ago as an excellent natural fat anti-oxidant, numerous other pharmacological applications and industrial (non-food) uses have been discovered and were interestingly and succinctly reviewed in 1972 (22). Effectiveness against a variety of oxidative-reductive and other enzymes has been demonstrated and it is used extensively as a standard for comparison of other lipoxygenase inhibitors now being widely developed (23). A synthesis of NDGA aimed to be economically competitive with natural

316

extraction has been developed in the Hoffmann-LaRoche laboratories and outlined in Scheme 3 (24). Friedel-Crafts propionylation of veratrole in chloroform gave the ketone (2), which on bromination yielded the a-bromoketone (4). An improved procedure for alkylation of (4) with the sodium enolate of (2), by conducting the reaction in liquid ammonia at -33°, produced the racemic diketone (5). The yield in each of these steps surpassed 90%, as did the acid-catalyzed dehydration of (5) to the furan (6). A survey of catalytic hydrogenation conditions provided reliable conversion to the cw-tetrahydrofuran (7) or the tetramethyl ether hydrogenolysis Scheme 3. Synthesis of NDGA

MeO>, MeO-^^^^

""

A r ^ ^

8

7 Ar = 3,4-Diinethoxyphenyl.

•*- 2

317

product (8) which on demethylation yielded NDGA (2). The principal functional group variations found in this family are hydroxyl groups (at C-1 and -4) or the corresponding ethers or esters. These are exemplified by (-)-dihydrocubebin (2), a constituent of Piper guineense (West African Black Pepper) used in treatment of coughs, intestinal disease, venereal disease, and rheumatism (25), ariensin (10) occurring in the exudate of the bark of Bursera ariensis and reportedly used as a cicatrizing agent (26), and (-f )-niranthin (11). one of many lignans isolated from Phyllanthus niruri (27). MeO,

...CHiOMe

MeO

CHjOMe

9 (-)-Dihydrocubebin, R=H 1 0 Ariensin, R=COCH3

Syntheses of (9) and (10) were conveniently effected (Scheme 4) starting from diethyl succinate (12) (28). The resultant dienolate (13). obtained by the action of two equivalent of lithium di-isopropylamide (29) gave on alkylation with methylenedioxybenzyl bromide in excellent yield a mixture of the (±)-ester (14) and me.yo-ester (15) which by alkaline hydrolysis yielded the dicarboxylic acid mixture (16) and (17). Without separation, this mixture on heating wih acetic anhydride gave the known rran^-dipiperonylsuccinic anhydride (18) (30). Reduction of (18) with lithium aluminium hydride gave (±)-dihydrocubebin (2), acetylation of which yielded (±)-ariensin (10). A synthesis of (±)-niranthin (11) (31) (Scheme 5) addresses the issue of the two aryl rings having different substituents. j8-Benzylbutyrolactones of type (19) are now readily available in racemic and enantiomeric forms (see next section) via classical Stobbe reaction procedures. The lithium enolate of (12) (32) reacted

318 Scheme 4. Synthesis of (±) - Dihydrocubebin (9) and (±)- Ariensin (10) H

CHjCOjEt

CO2R

CHjCOjEt

"I e,? CH—C—OEt CHr^C—OEt

0'i

13

Ar = 3,4-Methylenedioxyphenyl Scheme 5. Synthesis of ( ± )- Niranthin CHO

6^ ..CHjOH

11 CH2OH

b~/

319

readily with the aldehyde (20) (33) to give a mixture of epimeric alcohols (21, X = OH) which on catalytic hydrogenolysis gave the dibenzylbutyrolactone (21, X = H). Standard lithium aluminium hydride reduction to the diol (22) and methylation yielded (±)-niranthin (11). An interesting approach in which the thiophenes (23) and (24) are used as the C4-butane building block has been communicated (Scheme 6) (34). Cross-coupling of (23) with benzylmagnesium halides or (24) with arylmagnesium halides gave the respective dibenzylthiophenes (25) and dimethyldiarylthiophenes (26).

These

transformations were effected by use of a nickel phosphine complex, and transformation to the dibenzylbutanes (27) accomplished by Raney nickel desulfurisation. The overall pathway has also been modified for the synthesis of unsymmetrically substituted lignans.

A disadvantage appears to be lack of

stereospecificity in the terminal reduction step. Scheme 6. Thiophene Based Synthesis of Dibenzylbutanes Br,

,Br

Ai

v

23 Me

>—Ar

25 Me

v

y

Br-^g-/^Br

Ar-^^gX^Ar

24

26

•

^

Ar

27

Additional structural features found among members of this lignan group include benzylic hydroxyl and carbonyl groups, carboxylate ester functions and alkene unsaturation. Examples of such members (Scheme 7) include furoguaiaoxidin (28, from Guiacum officinale (35)), saururinone (22, from Saururus cemuus (36)), hydroxybuphthalmin

(2fi,

from

Heliopsis buphthalmoides (37))

veratrylpiperonylbutanol (31), from Virola elongata (38)).

and the

320

OMe

b^

OMe

B.

DIBENZYLBUTYROLACTONES Over sixty natural dibenzylbutyrolactone lignans are known; about two thirds

can be incorporated by the general formula (32): Y,Z may be -H, -OH or = 0 (Scheme 8). Representative examples are pluviatolide (33), oxomatairesinol (34), oxohinokinin (25), podorhizol (36) and parabenzlactone (37).

Others, such as

wikstromol (38) and thujastandin (22) are a-hydroxy or ai3-dihydroxybutyrolactones. Benzylidene and dibenzylidebutyrolactones such as suchilactone (40) and taiwanin A (41) are also known. j3-Benzyl-7-butyrolactones (44) for which convenient preparative procedures are available, and improved techniques for their a-alkylation and a-hydroxyalkylation, provide the most common synthetic route for these lignan sub-classes (39). The Stobbe condensation (40) of aryl aldehyde with dimethyl succinate (Scheme 9) leads to the half-ester (42) which can be catalytically hydrogenated at atmospheric pressure to give the dihydro half-ester (43). Selective reduction of the potassium salt of the latter can be effectively achieved by calcium borohydride (41)

321 Scheme 8. Dibenzylbutyrolactone Lignan Structures Y

H

MeO,

p

H O

MeO. HO

32

Ar'

l^ 9 H

V P

MeO, MeO

0-—/

Me

»<XxX' <Xfji40

38 Y=OH,

41

Z=H 39Y=Z-OH

^^

"OMe

OH

to yield the lactone (44).

^OMe OMe

h~J

Treatment of the lithium enolate with the necessary

aldehyde (ArCHO) leads to the epimeric alcohols (22, Y = H, Z = OH) or with the benzyl bromide (Ar'CHjBr) gives the lactone (22, Y = Z = H). More detailed discussions of application of the Stobbe reaction and useful variations pertinent to lignan synthesis have been provided (42,43). The occurrence and identification of enterolactone (45) in human pregnancy urine and other mammalian sources (44,45) attracted immediate interest in

322 Scheme 9. CHzCOiMe CHjCOzMe

^e ArCH

®CHC02iMe

COzMe

_^AKp

^OMe

Ar, iX^COzMe

-en ii2-^

_^

ArCH=C-Cp2Me

m

^

ArCH=C:—COjlVie Cf rHzCQzH

CHiCpi®

42

ArCHj-CH-COiMe CH2CO2H

Ca(BH4)2

32

*"

-

43

^

44

biological function and potential medical use. These aspects have recently been summarized (46). From a chemical viewpoint, this compound presented unusual features, e.g., occurring from other than a plant source, in a racemic form and lacking/7-phenolic (or ether) sub-unit functionality. Several syntheses were reported shortly thereafter, including two based on Stobbe procedures (47,48). As based on the Scheme 9 outline, a straightforward synthesis (48) is available. The benzyl ether (47) obtained from m-hydroxybenzaldehyde (46) was condensed with dimethyl succinate to give the half-ester (48) which was converted to the benzylidene lactone (49) and thence the benzyl lactone (50) by catalytic hydrogenation. Alkylation of the derived benzyl ether (51) with (47) gave the mixture of epimeric alcohols (52); alternatively, alkylation with m-benzyloxybenzyl bromide yielded (53). Both (52) and (53) produced enterolactone (45) on hydrogenolysis (Scheme 10). A second dibenzylbutyrolactone synthesis procedure which has found wide application utilizes a "tandem conjugate addition" to butenolide (49,50,51) and has

323

^^\«J«''''\/^"^

^ ' h C H ^ O . ^ ^ ^ ^ ^ ^ i CO,Me PhCH2

48

46R=H

COjH

4 7 l^CH2Ph

PhCHjO,

49

^

"OR

5 2 R=CH2Ph, X=OH 53R=CH2Ph,X=H SOl^H ^ 51 R=CH2Ph

been applied to a preparation of enterolactone (45) (52) (Scheme 11). Thus the anion of m-methoxybenzaldehyde phenylthioacetal (54) added in Michael fashion to butenolide and the product (55) trapped by m-methoxybenzyl bromide to produce enterolactone dimethyl ether (56), which on Raney nickel desulftirisation and demethylation with boron tribromide (53) yielded (45). A measure of the interest in the biological activity of these dibenzylbutyrolactone lignans is evinced in the recent spate of publications dealing with the total synthesis of the natural optically active products.

Again, the Stobbe

condensation pathway (Scheme 9) has been usefully exploited for this purpose. In a series of papers, resolution of the intermediate hemisuccinate esters (43) by chiral bases has been described (54), as has asymmetric hydrogenation (55), and the optically active lignan products synthesized in the usual way (43 -* 44 -^ 45).

324 Scheme 11. Tandem Conjugate Addition Route to Enterolactone

rx

ArC(SPh)2

PhS>,^^SPh

I

54

O

^ o

Ar

55 Ar=m-Methoxyphenyl

O

ArCH^Br ^

b (i) Raney Ni (ii) BBr3

45 Other methods of preparation and diastereoselective alkylation

of chiral

butyrolactones (44) are summarized in the recent review of asymmetric synthesis of lignans (16). C.

FURANS Numerically, this lignan class is presently the least significant, and all known

members have been isolated from tne same source, the heartwood of Guaiacum officinale.

Originally (56) ftiroguaiacin (57) (also known as a-guaiaconic acid) (57)

was isolated as the dimethyl ether derivative (58) and methylfuroguaiacin (59) was isolated as the ethyl ether (60). Later (58), from a more polar extract, furoguaiacidin

RO"

OR

57 58 59 60

R-R*=H R=R=IVle R=Me, R»«H R=Me, R'=Et

61 R-H 62 R«Et

325

(61) was isolated as the diethyl ether (62). These compounds are clearly amenable to synthesis by the standard fiiran heterocycle construction procedures. A synthesis of furoguaiacidin diethyl ether (62) is outlined in Scheme 12 (59). Alkylation of the j8-keto ester (63) with the a-bromopropiophenone (64) yielded the diketo ester (65) which was converted to the furan (66) by acid-catalyzed dehydration. Reduction of (66) with lithium aluminium hydride, and methylation of the resultant alcohol yielded (^. Scheme 12. Synthesis of Furoguaiacidin Diethyl Ether (62)

? ? ArCCH2C02Et + ArCCHCHj —

Et02C^

^

K*

^Ar

Br

63

64

65

Et02C>^^ Ai

^

^Q-'-^^Ar

66 1 (i) LiAlH4 (u) Mel

Ar = 4-Ethoxy-3-niethoxyphenyl

62 D.

TETRAHYDROFURANS

a. 2.5-BisaryItetrahydrofurans When a review of this class appeared in 1987, there were 24 natural members listed (60). Well over double this number of compounds of reasonably established constitution are now known. As represented by the generic formula (67). with the exception of a few members which bear hydroxymethyl groups at C-3 or C-3 and -4, R = R' = CH3 in all natural products of this group. This allows for six diastereomers, represented here as three 3,4-rraAW-dimethyl (68 A,B,C) and three 3,4-c/.s-dimethyl (68 D,E,F) tetrahydrofurans. In the common situation in which R = R' == CH3 and Ar = Ar', ten stereoisomers may exist, consisting of diastereomeric pairs (A,B,C,F and their enantiomers) and two meso forms (D,E) (Scheme 13). Where Ar = Ar' = 3,4dimethoxyphenyl, all are now known and the trivial names, veraguensin (§7A), galbelgin (67B), saucernetin (67C), tetrahydrofuroguaiacin B (67D), galgavin (£ZE) and ganschisandrin (67F) assigned.

326 Scheme 13.

Ar-^_>*^ O'

^'''

67

67A c-t-t

•^O^^*-' 67D c-a-c

67B H-t

Ar O 67E t'Q-t

^j^ c-t-c

Ar'

A r ^ 3 o 67F

Ar'

This lignan group gained significance in providing a lead in the development of potent platelet-activating factor antagonists. Platelet-activating factor (PAF) is a highly potent phospholipid found in a variety of cells (platelets, basophils, neutrophils, eosinophils, mast cells, endothelial cells, macrophages) implicated in inflammatory processes and eliciting biological response by interaction with specific receptors. It has become increasingly likely that PAF has multiple effects which may be relevant in many human diseases, and enormous recent activity in studying these effects has been an anticipated consequence. Much diverse information on these topics is now available in multi-author books or lengthy reviews (e.g., 61-65). To gauge the effect of PAF in a wide range of pathophysiological states and to develop drugs for use in human disease, an extensive search has ensued, particularly in major pharmaceutical company laboratories, to identify PAF antagonists.

A screening

programme initiated by the Merck Research Laboratory based upon an assay measuring the inhibition of binding of pH] PAF to rabbit platelet membrane preparations (66) revealed in 1982 that (±)-veraguensin (67A) (67) was a potent PAF antagonist. The natural product (H-)-veraguensin was first isolated from the Mexican

327

tree Ocotea veraguensis Mez. (Lauraceae) (68).

This discovery led to the

preparation (67,69-73) and inhibitor assay of all veraguensin stereoisomers. It was concluded that the all-cfa isomer (67D) was in fact the most potent antagonist of this group (72). The six isomers required for assay were prepared (72) as summarized in Scheme 14.

r--( (i) LiAlH4 (ii)Pd-Cy ' H2.

Ar^^Q^

67F

Ar

/--Ar

o o

|(i) LiAIH4 |(i)LiAIH4 (ii) MsCI (ii)MsCI

\r-'''^^^

Ar

67E

A r ^ ^ ^ ^ A r O |pd.C,H2 I ^

Ar^^^^-^Ar

67D

Ar«3,4-Dimethoxyphenyl In assay of the four 3,4-bisnor analogues (68-71) of these lignan tetrahydrofurans, it emerged that rraw5-2,5-bis(3,4,5-trimethoxyphenyl) tetrahydrofuran (L-652,731) (71) was the most potent of all ten compounds tested (Scheme 15). Further chemical elaboration led to (±)-rraAW-2-(3-methoxy-5-methylsulfonyl-4propoxyphenyl)-5-(3,4,5-trimethoxyphenyl) tetrahydrofuran (L-659,989) (72) (74), which was biochemically and pharmacologically characterized (75), and also obtained

328 Scheme 15. Tetrahydrofuran PAF Antagonists

x/^. OMe

OMe

MeO OMc

OMe

OMe

70, R«H

68, R=H 69, R=OMe

7 1 , R-OMe (L-652,731)

o p 9H

:::^9C OMe

OMe

OMe

OMe

73(MK287)

7 2 (L-659,989)

MeO

MeO

in both optically active forms (76,77).

In order to achieve improved metabolic

stability and pharmacokinetic profile, polar group modifications were investigated, from which the (-)-2S,5S-/ra/w-isomer of MK287 (73) emerged as a potent, specific and orally active PAF receptor antagonist, and chosen for clinical trial for asthma (78). Most recently, the development of (-)-rrfl/w-(2S,5S)-2-[3-(2oxopropyl)sulfonyl]-4-/i-propoxy-5-(3-hydroxypropoxy) phenyl-5-(3,4,5-trimethoxyphenyl) tetrahydrofuran (74) withftirtherimprovement of in vivo potency and drug characteristics has been described (79). The synthesis of this "third generation tetrahydrofuran derivative" is outlined in Scheme 16. In each of these diaryltetrahydrofuran syntheses, the significant starting material was a diarylbutane-1,4-dione, conveniently prepared by the Stetter reaction involving catalyzed addition of an aldehyde to an activated alkene (80). 3-Benzyloxy-4hydroxybenzaldehyde

(75) was converted

to 3-(methylthio)-4-/z-propoxy-5-

(benzyloxy)benzaldehyde (76) by three standard steps and reacted with 3,4,5-

329

Scheme 16. Syntheslsof74 0)Brj

OCHjPh

OCHjPh

OMe

75

76

77

-0^^ 78

SOjMe OPr

OCHjPh

SOiMe

Ar

S^OP, OCHjPh

80

OPr CHjPh

"'C^V^''''^^^ OR 8 1 R=CH2Ph 7 4 R«(CH2)30H

Ar=3,4,5-Triniethoxyp hen yl-

trimethoxyphenyl vinyl ketone (77) under Stetter conditions to yield the diketone (78). Oxidation of (78) with m-chloroperbenzoic acid and reduction with the chiral reducing agent S-BINAL-H (81) in a regio- and enantioselective manner yielded the S-alcohol (79). Reduction of (79) with sodium borohydride andcyclization with trifluoroacetic acid gave chirally pure trans-iM) and the cw-isomer, which could be equilibrated under acid conditions to increase the yield of (80).

A three step

sequence completed the synthesis of 74. The methylsulfone anion of (80) with ethyl acetate gave the ketosulfone (81) which on hydrogenolytic debenzylation and realkylation with 3-bromopropanol yielded (74). The preparation of a water-soluble phosphate ester pro-drug derivative equipotent to (74) in vivo was also disclosed.

330

b.

3.4-Bisbenzyltetrahydrofurans This is the smallest group of lignan tetrahydrofurans with fewer than ten

members.

All of reasonably well defined constitution are 3,4-rrflAW-disubstituted.

Of these, most attention has been directed towards burseran, a constituent of Bursera microphylla with tumour-inhibiting properties (82). Optically pure (-)-rraAW-burseran (82) and (-f )-d5'-burseran (83) were stereoselectively synthesized from chiral butyrolactones and gas chromatographic comparison indicated that the natural Scheme 17. Synthesis of (+)- and (-)- Burseran MeO.

"i

MeO

83 Ar y—SnMe3

\f

Ar =3,4,5-Trimethoxyph enyl Ar'=3,4-Methylenedioxyphenyl R=PhCH(OMe)-

(-)- Burseran

82

(+)- Burseran

84

331

anti-tumour lignan is the trans isomer (83,84). An interesting coupling reaction pathway to (±)-burseran has been devised (85) and routes to both (-)-burseran (82) and (-h)-burseran (84) have recently been described (86,87). The latter, in which a radical-mediated carbocyclization is the key step, is outlined in Scheme 17. Allylation of 3,4,5-trimethoxycinnamyl alcohol (85) gave the diene ($©, which underwent cyclization on radical stannylation (88) to yield (87).

A selective

oxidative cleavage of the trimethylstannyl group of (87) with eerie ammonium nitrate gave the aldehyde mixture (88) with same stereochemical ratio. The ^ron^-isomer (82), obtained in high excess (>23:1) after equilibration was treated with 3,4methylenedioxyphenyllithium to give the separable epimeric adducts (90). Resolution of (90) with (S)-O-methylmandelic acid gave the diastereomeric esters (91) and (92), which were separately hydrogenolyzed to give (-)-burseran (82) and (+)-burseran (84) respectively, c. 2-Aryl-3-methyl-4-bQnzyltetrahydrofurans At this time, there are known about 30 members of this sub-class (93) which Scheme 18.

2-Aryl-3-methyl-4-benzyltetrahydrofurans

•Ar'

--yf 93

HO'

>^^V^Ar'

H 94

HO

332

bears a close structural resemblance to theftirofurangroup (94) (Scheme 18). A close biogenetic correspondence may also be presumed from the observations that in the natural tetrahydrofurans the C-3 substituent is almost invariably a hydroxymethyl group and that the C-3 and C-4 substituents have a cw-relationship.

As a

comparison, (-f )-lariciresinol (25), an extract of the resinous exudate of the European larch {Larix decidua) is seen as a dihydro-derivative of (4-)-pinoresinol (96) which displays inhibitory activity against cyclic adenosine monophosphate phosphodiesterase (89).

Little attention has been devoted to syntheses of this class per se but

preparations by benzylic hydrogenolysis (Pd/Hj or NaNHj) of suitable fiiroftirans (e.g., (96) -* (95)) are well known (90-91). The principal substituent variations found naturally are hydroxyl functions (at locations denoted by arrows in (93)) or benzylic ketone groups as in (H-)-arborone (97). isolated from the heartwood of Gmelina arborea (92). E.

ARYLNAPHTHALENES This group of over 40 known natural members can be conveniently subdivided

on a structure basis into naphthalene-2-carboxylic acid lactones ("down" lactones) (28), naphthalene-3-carboxylic acid lactones ("up" lactones) (22) and non-lactonic naphthalenes (100). In this last group, in the most common situation, R = R' = Me, but examples are known in which R and R' are hydroxymethyl groups or amide groups.

Additionally R' may be -CHO, -COjH or H.

Although there were

numerous structure mis-assignments made in the early literature and ultimately corrected by unambiguous synthesis, proton magnetic resonance spectra determination has enormously simplified constitution assignment (93). A convenient route to lignans of this class is apparent from the known transformation of the phenylpropiolic acid to the anhydride of phenylnaphthalene-2,3dicarboxylic acid by the action of acetic anhydride (Scheme 19). This reaction was discovered about a century ago (94); the interesting early history has been briefly reviewed (95) and a wide generality of the reaction demonstrated (96).

It is

particularly useful for products bearing identical substituents in rings A and C. Thus, it has been long known that 3,4-methylenedioxyphenylpropiolic acid (101) on

333 Scheme 19. Arvlnaphthalcne Lignans : Lactone Synthesis

OX" Cyyi" cpc Ar

T

98

Ar 99

1

Ar

100

h-J Taiwanin C 1 0 5

heating with acetic anhydride yields the anhydride (102) (97).

Reduction with

lithium aluminium hydride and oxidation of the resultant diol (103) with F^tizon's reagent yielded both lactones, justicidin E (104), a piscicidal constituent of Justicia procumbens, and taiwanin C (105) which had been isolated from Taiwania cryptomeroides (93). As anticipated, oxidation occurred with high stereoselectivity of the less hindered hydroxymethyl group (104:105^ ca 4:1). A shorter procedure, which yields both lactone products with the opposite stereoselective outcome consists of direct reduction of anhydride with sodium borohydride (98).

334

When 2-bromo-4,5-methylenedioxyphenylpropioIic acid (106) failed, on heating with acetic anhydride, to yield the desired intramolecular cyclized anhydride (107), the problem was overcome by the use of dicyclohexylcarbodiimide in dimethoxyethane solution below 0° (99,100).

The product (107) was subsequently used

(101,102) as an intermediate for the synthesis of helioxanthin (108) (Scheme 20) which had been isolated and identified as a constituent of Heliopsis helianthoides (103,104). Scheme 20. Synthesis of Helioxanthin

PCC

^

CO2H

\yJ UAIH4 CH.OH P

^ Ag^COj CH2OH

h-J

In a series of papers under the rubric of Intramolecular Diels-Alder Reactions during 1963-1976, the cyclization inter alia of phenylpropargyl phenylpropiolate esters, i.e., functionality designed to produce arylnaphthalene lactones essentially in one step, was extensively examined (105). With the aryl rings differently substituted by the common substituents, there is little regioselectivity in the cyclization step. A

335

useful application is the recently reported synthesis (106) (Scheme 21) of the natural lactones, daurinol (112) and retrochinensin (113). The former, a constituent of Haplophyllum dauricum (107-109), bears in ring A the isovanillyl fragment; this is extremely rare among lignans and in some cases is based on equivocal evidence. The latter was first isolated as a constituent from an anti-depressant extract from Justicia prostata (110). By heating the di-ynic ester (109) in xylenes, the lactones (110) and (111) were obtained and separated.

Debenzylation of (110) by a catalytic

hydrogenation procedure (111) gave daurinol (112). Similar debenzylation of (111) followed by methylation of the resultant phenol yielded retrochinensin (113). Scheme 21. Synthesis of Daurinol and Retrochinensin BnO Me O

COX" <XXp6

"O

OW

109

OMe

llGR^Bn

lllR«Bn

112R=H

113R-Me

The overall modest yields achieved in these syntheses have recently been markedly improved by the use of the solid-phase copolymer of 4-vinylpyridine (P4VP) (112) in the formation of the starting di-ynic esters. For example, when a suspension of P4-VP polymer in dichloromethane was stirred with the acid chloride from (101) and then the propargyl alcohol (114), the ester (115) was obtained excellent yield. By heating in xylene, (115) underwent intramolecular cyclization to yield justicidin E(104) and taiwanin C(105) as the major products; in addition, the isomers helioxanthin (108) and retrohelioxanthin (116) could also be isolated (Scheme 22) (113). Increased interest in these four lactone products has resulted from an assay which indicates 5-lipoxygenase inhibitory activity (114).

336 Scheme 22. Intramolecular Cyclizalion of Arylpropargyl Arylpropiolate Esters

lOi^^Si-

-COCl (i) P4.VP (u) A r — = — C H 2 O H

114 O ArO Ar-

—

/

115

Ar=3,4-Methylenedioxyphenyl

A general two-step synthesis of arylnaphthalenes from 0-/-butyldimethylsilylcyanohydrins involving a tandem conjugate addition-aldol reaction, followed by Schema 23. OTB TBS

"CN

NC

OTBS

MeO.

MeO, ^O CFiCO.H^

I " O

117

118 TBS='^Butyldilne(hylsilyl

iTl

119 Justicidin B

337

acid-catalyzed construction of the naphthalene ring has recently been described (115) and an example shown in Scheme 23. The known r-butyldimethylsilylcyanohydrin (111) derived from veratraldehyde gave on successive treatment with lithium diisopropylamide, 2-butenolide and piperonal the butyrolactonef 118). which on refluxing with trifluoracetic acid gave the natural lactone, justicidin B (119). F.

ARYLDIHYDRONAPHTHALENES About 14 natural compounds of this lignan sub-class are known, among which

are examples of 1,2-dihydro, 1,4-dihydro and 3,4-dihydro-l-arylnaphthalenes. The double bond locations are also designated as a-apo, ^-apo and y-apo respectively Scheme 24. AryWihydronaphthalene Lignans

CL-apo

COjH '^-apo

CHjOH y-apo

120

OMe

MeQ HO

122 OMe OH

U/

338

(120) (Scheme 24).

A representative of each group is thomasic acid (121). an

extractive of the heartwood of Ulmus thomasii Sarg. ("rock elm") (116), jSapoplicatitoxin (122) from Thuja plicata Donn, (Western Red Cedar) (117), and collinusin (123) from the leaves of Cleistanthus collinus (118). a. 1,2-Dihydro-l-arylnaphthalene Lignans The structure (121) of thomasic acid was established by extensive spectroscopic and chemical degradation studies (116,119,120) and presents several interesting features. It is one of the few lignans occurring naturally in racemic form and in possessing a free carboxyl group. In addition, the trans vicinal substituents (at C1,2) adopt a diaxial conformation. The synthesis of this product (Scheme 25) was Scheme 25. Synthesis of Thomasic Acid OMe OH

OMe

OMe 124

MeO,

CO,R

R'O

•CO2R

JL J Ho-"^

125

OMe

126 R«=R'«H 1 2 7 R-Me,R'=H 1 2 8 R-Me,R=CH2Ph

MeO

CHO

McO, R'O

OCH,Ph

129

1 3 0 R-Me,R'-CHiPh 1 2 1 R=R'-H

based on biomimetic speculation (121). Phenolic oxidation of sinapic acid (124) gave the dilactone (125) (122), which with hydrochloric acid in aqueous dioxan gave the

339

congener, thomasidioic acid (126): similar treatment in methanol gave the dimethyl ester (127). Reduction of the dibenzyl ether (128) with lithium aluminium hydride followed by selective allylic oxidation with manganese dioxide yielded the aldehyde (129). The aldehyde ftinction was selectively oxidized in the presence of the primary alcohol function (123) to give the methyl ester (130) which on standard debenzylation and hydrolysis produced thomasic acid (121). b. jg-ApQlignans A significant synthesis pathway for such products (Scheme 26) involves specific photoconversion of 2,3-dibenzylidene-butyrolactones, usually prepared by Stobbe Scheme 26, p-ApoUgnans MeO

MeO

h^

h^ condensation

methods

(124).

For

example,

condensation

of

dimethyl

piperonylidenesuccinate with veratraldehyde gave the half ester (131), which on selective reduction (LiAlH4 at -25**) followed by acidification yielded the lactone (132). Photoirradiation (light filtered through borosilicate glass) of (132) gave a

340

mixture from which the jS-apolignans (133. 56% yield) and (134. 28% yield) were isolated. Of further interest, air oxidation of jS-opolignans provides an additional pathway to the corresponding arylnaphthalene and 4-hydroxyarylnaphthalene lactones, c. 7"/4/7<7lignans The 7-apolignan, collinusin (123) was isolated from Cleistanthus collinus, a highly poisonous plant reportedly used for insecticidal, piscicidal and suicidal purposes (125). proceeds

with

Unlike the cyclization of arylpropargyl arylpropiolates which little

regioselectivity,

cinnamyl

arylpropiolates

give

aryldihydronaphthalene-2-carboxylic acid lactones with excellent regioselectivity (105). A synthesis of collinusin (123) using this method (126) has recently been markedly improved (127) (Scheme 27). Polymer (P4-VP)-mediated esterification Scheme 27.

MeO

MeO

CHjOH P4-VP

135 -COCI

^o DMF, A

MeO

h~J

h~J

341

of 3,4-methyIenedioxypropiolyl chloride with 3,4-dimethoxycinnamyl alcohol (135) gave the ester (136) in 89% yield, which on heating in dimethylformamide gave collinusin (123) in 84% yield and the isomer (137) in 5% yield. Excellent agreement in the spectroscopic data indicates that the latter is one of the lignans isolated from the tumor-inhibiting extract of Polygala polygama (128,129) which had previously been considered to have the j8-a/7(?polygamatin structure (134). G.

ARYLTETRALINS About 90 aryltetralin lignans are presently known of which one, podophyllotoxin

(138). may be claimed to have generated most interest within the entire lignan field. Podophyllum is the name given to the dried roots and rhizomes derived commercially from Podophyllum peltatum L. (a North American plant known variously as "May apple," "American mandrake," "Indian apple," "wild lemon," "duck's foot" and vegetable calomel) and Podophyllum emodi Wall, (a related Indian species). The water-insoluble alcohol-soluble fraction of podophyllum possesses most of the recognized biological activity of the original root and is referred to as podophyllum resin or podophyllin. The early history incorporating medicinal attributes is available in intriguing reviews (130,131). The isolation of podophyllotoxin from the resin is usually dated to 1880 (132) with structure elucidation being completed in 1951, after particularly notable contributions from Borsche, Spath, Hartwell and their coworkers (for comprehensive review coverage, see references (5) and (131)).

The first

pioneering classical total synthesis of (±)-podophyllotoxin was communicated in 1962 (133) with experimental details being provided in 1966 (134). The anti-neoplastic activity of podophyllotoxin and derivatives has prompted continuous development into clinical agents for treatment of human neoplasia. The semi-synthetic 4'-demethylepipodophyllotoxin derivatives, Etoposide (139) and Teniposide (140), developed by a Sandoz (Basel) group have attracted considerable attention (135-137) (Scheme 28). They have established antitumour activity with lesser toxicity and mechanism of action differing from podophylloxin itself. Enormous activity has been exerted dealing with syntheses of podophyllotoxin and the related diastereomeric forms (Scheme 29) and a masterly summarizing review

342 Scheme 28.

"^-^^V^^ "^^^^^--^^^o

MeO

MeO'

^r OMe OMe

^^

OMe

OH

138

1 3 9 R=CH3 Etoposide(VP-16-213) 1 4 0 R= B

\ Teniposide (VM-26 )

Scheme 29. Podophyllotoxin and Diastereomers

Ar Podophyllotoxin

Picropodophyllin

I;!

r O -

H

O

Epipodophyllotoxin

^'^^

XJ^ ? ft

Ar

Epiisopicropodophyllin

OH

OH

OH

Ar

Epiisopodophyllotoxin

o0

r

>k" = H Ar

o o

Epipicropodophyllin Isopodophyllotoxin Ar == 3,4,5-TrimethoxyphenyI

H b Ar Isopicropodophyllin

343

has just appeared (138). A paper encompassing a comprehensive route to all eight diastereomeric Podophyllum lignans is particularly worthy of attention (139). In certain cases, structure elucidation of aryltetralin lignans has only been established by total synthesis of the (±)-compounds.

Noteworthy in this respect

have been the constituents of Phyllanthus niruri Linn. (Euphorbiaceae), extracts of which have been used medicinally (in treatment of asthma, jaundice and bronchial infections) (140).

Considerable confusion resulted mainly from

differing

interpretations of spectroscopic data, and at least five different structures were proposed for the major constituent, hypophyllanthin. The structure of the aryltetralin constituents established by unequivocal synthesis (141) are shown in Scheme 30: they were given the names hypophyllanthin (141). nirtetralin (142), phyltetralin (143) and lintetralin (144). Scheme 30. Aryltetralin Lignans of Phyllanthus niruri

MeO,

CHjOMe

CHzOMe

CHjOMe

CHjOMe

OMe

OMe

CHjOMe

CHjOMe

CHjOMe

CHjOMe

OMe

OMe

Related syntheses of (141) and (142) are outlined in Scheme 31. The starting aldehyde (145) (142) was converted by now standard procedures to the benzylic butyrolactone (146) and thence to the mixture of epimeric alcohols (147). Acid-

344 Scheme 31. Synthesis of HypophyDanthin and Nirtetralin

MeO^

^CHO

MeO

MeOs,

JMcO,

MeO. Hypophyllanthin

Nirtetralin

142

141 OMe

aluminium hydride reduction of (148) and methylation of the resultant diol gave (±)nirtetralin (142).

Bromination of (146) proceeded as expected (electrophilic

substitution ortho- to the methoxyl substituent) to produce (150) which by the same procedures via (151) gave (±)-hypophyllanthin (141) (143).

The same general

pathway provided (±)-phyltetraIin (143) and (±)-lintetralin (144) (144). H.

TETRAHYDROFUROFURANS This lignan sub-class of 2,6-diaryl-3,7-dioxabicyclo[3,3,0]octanes, is one of the

largest and most widely distributed groups. About 100 members are known, if natural glycoside derivatives are included.

As expected, the five-membered

heterocyclic rings are constrained in cis fusion. When the aryl groups are identically

345

substituted, three distinct stereochemical series are possible and well established. Representatives are eudesmin (152) (most stable with "equatorial" aryl groups), ^preudesmin (153) (of lesser stability with one "equatorial" and one "axial" aryl group) and diatudcsmin (154) (of least stability with two "axial" aryl groups). They may undergo epimerisation under acidic conditions (145). Principal structural group variation within the standard structures are additional hydroxyl groups at one or more of positions 1, 2 and 4. Scheme 32.

Tetrahydroturofurans

8

Ar'

....Ar'

-H Ar

152

Ar=Ar*=3,4-Dlmcthoxyphetiyl

HO-

155

f

so,

tr-

McOjC Ar*"

160

O^

HOHjC

r-'O O'

161

'~~^

^OH

H.4-4...H Ar-

162

^o^

8

Ar*

...../

O

...H

163

346

Structure elucidation of the natural tetrahydrofurofurans by chemical methods, particularly hydrogenolytic cleavage of the heterocyclic ring to simpler identifiable moieties has been well reviewed (146,147). Reliable consignment of configuration has now been considerably aided by PMR, CMR and NOE experiments and significant data has been usefully tabulated (148). The monographs (8,9) also present clear expositions of methods of synthesis. A recent communication addressing a short approach to both symmetrical and unsymmetrical tetrahydrofurofuran lignans is outlined in Scheme 32 (149). Michael addition of the sodium salt of the silyl monoprotected diol (155) to the asulfonylcinnamate (156) gave the ether (157) which on desulfurization and hydrolysis gave the hydroxy acid (158). Lactonization of (158) by the Mukaiyama method (150) to (159) was followed by a Claisen rearrangement under specific conditions (151) leading to the tetrahydrofuran (160) following esterification. Reduction to the tetrahydrofuranmethanol (161) followed by oxidative cleavage of the vinyl group yielded the known (±)-samin (162) which with arylmagnesium halide and dehydration gave the lignan structure (163). The latter stages of this synthesis scheme had previously been reported (152). An alternative approach to that outlined in Scheme 32, with the aim of ready adaptation to asymmetric synthesis, has been recently reported (153). I.

DIBENZOCYCLO-OCTADIENES About 50 members of this class (including esters of natural alcohols) are known;

most have been isolated from the fruit, leaves and seeds of Schizandra species, extracts of which have long been used in Asia for medicinal purposes.

These

extracts are the basis of drugs known in China as Wu-Wei-Zi and in Japan as Kitagomisi. The isolated products are usually denoted by names or variants (including alphabetical letter attachments) of schizandrin, gomisin and wuweizisu. Two other significant sources are Kadsura species (with typical product names as kadsurin and kadsuranin) and Steganotaenia araliaceae (giving rise to steganes). The structurally simplest natural members have phenol, phenolic ether and/or phenolic ester functional groups at C-1,2,3,12,13,14 and methyl groups at C-7 and

347

8. In addition, alcohol or derived ester functional groups (benzoates, angelates, tiglates) may be located at C-6 and/or C-7 (164) (Scheme 33). The Steganotaenia products are modifications with a //ww-lactone group at C-7 and 8 and alcohol, ester or carbonyl group at C-9 (165). Scheme 33. Dibenzocyclo-octadienes

MeO MeO,

MeC Med 1 6 6 R-(4-)-Goinisin H

MeO

168

S-(>>-Steganone

A particularly interesting feature of this lignan sub-class is that the aryl groups, with rotation restricted by locking within the cyclo-octane ring, confer dissymmetry to the molecule. Both R- and S- configurations are well represented in the natural products, e.g., R-(-f)-gomisin H (166), S-(-)-gomisin N (167) and S-(-)-steganone (168). The earliest work on the isolation and structure elucidation was performed by Kochetkov and co-workers (154,155).

Particularly noteworthy have been the

continuing efforts of Ikeya and Taguchi and their co-workers on the isolation, structure elucidation and absolute configuration determination of the schizandrins and gomisins beginning in 1979 (156). Application of a wide range of techniques (UV,

348

MS, CD, X-ray crystal analysis, PMR, CMR and NOE difference spectroscopy) used significantly by these workers has been useftilly summarized (157). The use of vanadium oxyfluoride as a reagent for the intramolecular oxidation of non-phenolic substrates (158) provided a convenient synthesis pathway to dibenzocyclo-octadienes (Scheme 34). It was shown that the racemic diaryIbutane Scheme 34.

170 MeO,

MeO

OMe OMe

MeO

171

RC),

RO

172R=Me

173 R=Me

174R=-CH2-

175R=-CH2-

349

(169) with this reagent gave the trans-dimethyl product (170) and the meso-dindAogue (8) give the cw-dimethyl analogue (171) (159). With this established, oxidation of the hexamethoxydiaryl-butane (172) with vanadium oxyfluoride in trifluoroacetic acid gave (±)-deoxyschizandrin (173) (159) and similar treatment of (174) gave (±)-wuweizusin C (175) (142). This intramolecular oxidation route is also applicable to the natural lactones from dibenzylbutyrolactones. A variety of alternative oxidants used for tris(trifluoroacetate)

(160),

the same purpose and ruthenium

including

thallium

tetrakis(trifluoroacetate)

(161),

bis(trifluoroacetoxy)iodobenzene (162) and ferric perchlorate (163) have subsequently been introduced. The discovery of antileukemic activity possessed by the lactones isolated from Steganotaenia araliaceae (164) has prompted considerable efforts directed towards their synthesis. These have recently been admirably reviewed (165). J.

DIARYLCYCLOBUTANES All known members of this group have the comparatively rare 2,4,5-

trimethoxyphenyl group as the aryl component. Heterotropan (176) was isolated from Heterotropa takaoi with the assigned configuration supported by NOE measurements, and was synthesized by photo-dimerization of (E)-asarone (179) (166). A stereoisomer isolated from Magnolia salicifolia was named magnosilin and assigned structure (177) (167). Both products were subsequently isolated from the same source, Piper cubeba (168).

A third stereoisomer, isolated from Piper

sumatranum var. andamanica with the M-trans structure (178) was named andamanicin, and also obtained from the complex mixture obtained from irradiation of (179) (169) (Scheme 35). Acoradin (from Acorus calamus) (170) and bisasaricin (from Acorus gramineus) (171) have been reported as diarylcyclobutanes, but with unassigned configurations. Although this subclass has structures in accordance with our definition of lignan, it seems probable that a distinctly different mode of biogenesis is involved. The acid-catalyzed dimerization of (179) yields an arylindane neo-lignan product (180) (172,173).

350 Scheme 35. Diarylbutancs

x^-y <

Ar

176 Heterotropan

•

•

—

^

Ar

Ar''

177 Magnosilin

Ar'

Ar

178 Andamanicin

Ar=2,4,5-Trimethoxyphenyl

ACKNOWLEDGEMENTS I am deeply indebted to Dr. Emile Al-Farhan for his preparation of the structure schemes, and the Merck Research Laboratories for their support of our work.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9.

R.D. Haworth, Annu. Rep. Prog. Chem., 33 (1936) 266. O.R. Gottlieb, Phytochemistry, 11 (1972) 1537. O.R. Gottlieb, Rev. Latinoam. Quim., 5 (1974) 1. H. Erdtman, Moderne Methoden der Pflanzenanalyse, K. Paesch and M.V. Tracey (Eds.), Springer-Verlag, 1955, p. 3. W.M. Hearon and W.S. McGregor, Chem. Rev., 55 (1955) 957. K. Freudenberg and K. Weinges, Tetrahedron, 15 (1961) 115. K. Weinges and R. Spanig, Oxidative Coupling of Phenols, W.I. Taylor and A.R. Battersby (Eds.), M. Dekker, New York, 1967, p. 323. C.B.S. Rao (Ed.), Chemistry of Lignans, Andhra University Press, Waltair, India, 1978. D.C. Ayres and J.D. Loike, Lignans, Cambridge University Press, Cambridge, 1990.

351

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.

O.R. Gottlieb, New Natural Products and Plant Drugs with Pharmacological, Biological and Therapeutic Activity, Springer-Verlag, Berlin, 1977, p. 227. W.D. Macrae and G.H.N. Towers, Phytochemistry, 23 (1984) 1207. Ref. (9), pp. 85-112. O.R. Gottlieb and M. Yoshida, Natural Products of Woody Plants, Vol. 1, J.W. Rowe (Ed.), Springer-Verlag, 1989. D.A. Whiting, Nat. Prod. Rep., (a) 2 (1985) 191, (b) 4 (1987) 499, (c) 7 (1990) 349. R.S. Ward, Chem. Soc. Rev., 11 (1982) 75. R.S. Ward, Tetrahedron, 15 (1990) 5029. G.M. Massanet, E. Pando, F. Rodriguez-Luis and E. Zubia, Fitoterapia, 60 (1989) 3. Ref. (9), pp. 12-84. Ref. (9), pp. 4-7. Ref. (8), pp. 65-94. A.W. Schrecker and J.L. Hartwell, J. Amer. Chem. Soc, 79 (1957) 3827. E.P. Oliveto, Chem. Ind., London, 1972, p. 677. J. Rokach (Ed.), Leukotrienes and Lipoxygenases, Elsevier, Amsterdam, 1989, p. 438. C.W. Perry, M.V. Kalnins and K.H. Deitcher, J. Org. Chem., 37 (1972) 4371. D. Dwuma-Badu, J.S.K. Ayim, T.T. Dabra, H.N. ElSohly, J.E. Knapp, D.J. Slatkin and P.L, Schiff, Lloydia, 38 (1975) 343. J.D. Hernandez, L.U. Roman, J. Espineira and P. Joseph-Nathan, Planta Med, 47 (1983), 215. A.S.R. Anjaneyulu, K.T. Rao, L.R. Row and C. Subrahmanyam, Tetrahedron, 29 (1973) 1291. J.M. Burke and R. Stevenson, Planta Medica, (1985) 450. R.N. Long and M.W. Rathke, Synth. Comm., 11 (1981) 687. R.D. Haworth and D. Woodcock, J. Chem. Soc, (1938) 1985. G.E. Schneiders and R. Stevenson, Org. Prep. Proced. Int., 14 (1982) 1. P.A. Ganeshpure and R. Stevenson, J. Chem. Soc, Perkin 1, (1981) 1681. G.E. Schneiders and R. Stevenson, J. Org. Chem., 46 (1981) 2969. A. Minato, K. Tamao, K. Suzuki and M. Kumada, Tetrahedron Lett., 21 (1980) 4017. P. Majumder and M. Bhattacharyya, Chem. Comm., (1975) 702. K.V. Rao and N.S. Prakasa Rao, J. Nat. Prod., 53 (1990) 212. J. Jakupovic, A. Schuster, F. Bohlmann, R.M. King and H. Robinson, Planta Medica, (1986) 18. M.J. Kato, M. Yoshida and O.R. Gottlieb, Phytochemistry, 29 (1990) 1799. P.A. Ganeshpure and R. Stevenson, J. Chem. Soc, Perkin 1, (1981) 1681, and references cited therein. W.S. Johnson and G.H. Daub, Org. React., 6 (1951) 1. E. Brown, J.-P. Robin and R. Dhal, J.C.S. Chem. Comm., (1978) 556. G. Bagavant and P.A. Ganeshpure, Current Sci., 47 (1978) 338. Ref. (9), pp. 306-313.

352

44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71.

S.R. Stitch, J.K. Toumba, M.B. Groen, C.W. Funke, J. Leemhuis, J. Vink and G.F. Woods, Nature, 238 (1980) 738. K.D.R. Setchell, A.M. Lawson, F.L. Mitchell, H. Adlercreutz, D.N. Kirk and M. Axelson, Nature, 238 (1980) 740. Ref. (9), pp. 95-98. G. Cooley , R.D. Farrant, D.N. Kirk and S. Wyn, Tetrahedron Lett., (1981) 349. P.A. Ganeshpure and R. Stevenson, Chem. and Ind., (1981) 778. R.E. Damon, R.H. Schlessinger and J.F. Blount, J. Org. Chem., 41 (1976) 3772. F.E. Ziegler and J.A. Schwartz, J. Org. Chem., 43 (1978) 985. G.B. Mpango and V. Snieckus, Tetrahedron Lett., (1980) 4827. A. Pelter, P. Satyanarayana and R.S. Ward, Tetrahedron Lett., (1981) 1549. M.B. Groen and J. Leemhuis, Tetrahedron Lett., (1980) 5043. K. Lalami, R. Dhal and E. Brown, J. Nat. Prod., 54 (1991) 110 and earlier papers in series of which this is Part 12. H. Kawano, Y. Ishii, T. Ikariya, M. Saburi, S. Yoshikawa, Y. Uchida and H. Kumobayashi, Tetrahedron Lett., 28 (1987) 1905. F.E. King and J.G. Wilson, J. Chem. Soc, (1964) 4011. J.F. Kratochvil, R.H. Burris, M.K. Seikel and J.M. Harkin, Phytochemistry, 10 (1971) 2529. P.L. Majumder and M. Bhattacharyya, Chem. Ind., London, 1974, 78. F.G. Schreiber and R. Stevenson, J. Org. Chem., 40 (1975) 386. T. Biftu and R. Stevenson, Phytotherapy Research, 1 (1987) 97. F. Snyder (Ed.), Platelet-Activating Factor and Related Lipid Mediators, Plenum Press, New York, 1987. P.J. Barnes, C.P. Page and P.M. Henson (Eds.), Platelet Activating Factor and Human Disease, Blackwell, London, 1989. D.A. Handley, R.N. Saunders, W.J. Houlihan and J.C. Tomesch (Eds.), Platelet-Activating Factor in Endotoxin and Immune Dieases, M. Dekker, New York, 1990. P. Braquet, L. Touqui, T.-Y. Shen and B.B. Vargaftig, Pharmacological Reviews, 39 (1987) 97. M. Koltai, D. Hosford, P. Guinot, A. Esanu and P. Braquet, Drugs, 42 (1991) 19 and 174. S.-B. Hwang, C.-S. Lee, M.J. Cheah and T.Y. Shen, Biochemistry, 22 (1983) 4756. R. Ahmed, F.G. Schreiber, R. Stevenson, J.R. Williams and H.M. Yeo, Tetrahedron, 32 (1976) 1339. N.S. Crossley and C. Djerassi, J. Chem. Soc, (1962) 1459. T. Biftu, B.G. Hazra and R. Stevenson, J. Chem. Soc, Perkin 1, (1979) 2276. T. Biftu, B.G. Hazra, R. Stevenson and J.R. Williams, J. Chem. Soc, Perkin 1, (1978) 1147. R. Stevenson and J.R. Williams, Tetrahedron, 33 (1977) 285.

353

72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104.

T. Biftu, N.F. Gamble, T. Doebber, S.-B. Hwang, T.-Y. Shen, J. Snyder, J.P. Springer and R. Stevenson, J. Med. Chem, 29 (1986) 1917. T. Biftu, T.W. Doebber, S.-B. Hwang, T.R. Beattie, T.-Y. Shen and R. Stevenson, U.S. Pat. (1986) 4,595,693. M.M. Ponpipom (and 12 co-authors), Biochem. Biophys. Res. Commun, 150 (1988) 1213. S.-B. Hwang, M.-H. Lam, A.W. Alberts, R. Bugianesi, J.C. Chabala and M.M. Ponpipom, J. Pharmacol. Exp. Ther., 246 (1988) 534. M.M. Ponpipom, R.L. Bugianesi and J.C. Chabala, Tetrahedron Lett., 29 (1988)6211. T. Biftu, J.C. Chabala, J. Acton, C.-H. Kuo and R. Stevenson, Drugs of the Future, 14 (1989) 359. S.P. Sahoo (and 18 co-authors), Biorg. Med. Chem. Lett., 1 (1991) 327. N.N. Girotra (and 22 co-authors), J. Med. Chem., 35 (1992), 3474, H. Stetter, Angew. Chem. Int. Ed., Engl., 15 (1976) 639. R. Noyori. L Tomino, Y. Tanimoto and M. Nishiwaza, J. Am. Chem. Soc, 106 (1984) 6709. J.R. Cole, E. Bianchi and E.R. Trumbull, J. Pharm. Sci., 58 (1969) 175. K. Tomioka, T. Ishiguro and K. Koga, J. Chem. Soc, Chem. Comm., (1975) 652. K. Tomioka, T. Ishiguro and K. Koga, Chem. Pharm. Bull., 33 (1985) 4333. J.L. Belletire, D.F. Fry and S.L. Fremont, J. Nat. Prod., 55 (1992) 184. N. Rehnberg and G. Magnusson, J. Org. Chem., 55 (1990) 4340. S. Hanessian and R. Leger, Synlett., (1992) 402. G. Stork and P.M. Scher, J. Am. Chem. Soc, 108 (1986) 303. H. Tsukomoto, S. Hisada and S. Nishibe, Chem. Pharm. Bull., 32 (1984) 4482. K. Weinges, Chem. Ber., 94 (1961) 2522. Ref. (8), pp. 189-192. P. Satyanarayana, P.K. Rao, R.S. Ward and A. Pelter, J. Nat. Prod., 49 (1986) 1061. T.L. Holmes and R. Stevenson, J. Org. Chem., 36 (1971) 3450. A. Michael and J.E. Bucher, Chem. Ber., 28 (1895) 2511. A.W. Johnson, Acetylenic Compounds, Vol. II, p. 71, Arnold, London (1950). F.G. Baddar, J. Chem. Soc, (1947) 224, and subsequent papers in series. R.D. Haworth and W. Kelly, J. Chem. Soc, (1936) 745. Z.-I. Horii, M. Tsujiuchi, K.-I. Kanai and T. Momose, Chem. Pharm. Bull., 25 (1977) 1803. D. Brown and R. Stevenson, Tetrahedron Lett., (1964) 3213. D. Brown and R. Stevenson, J. Org. Chem., 30 (1965) 1759. T.L. Holmes and R. Stevenson, Tetrahedron Lett., (1970) 199. T.L. Holmes and R. Stevenson, J. Chem. Soc. (C), (1971) 2091. R.S. Burden, L. Crombie and D.A. Whiting, Tetrahedron Lett., (1968) 1035. R.S. Burden, L. Crombie and D.A. Whiting, J. Chem. Soc. (C), (1969) 693.

354

105. L.H. Klemm, Vinh Tarn Tran and D.R. Olson, J. Heterocycl. Chem., 13 (1976) 741, and earlier papers in series. 106. P.T. Anastas and R. Stevenson, J. Nat. Prod., 54 (1991) 1687. 107. D. Batsuren, E.Kh. Batirov, V.M. Malikov, V.N. Zemlyanskii and M.R. Yagudaev, Chem. Nat. Compd. (Engl. Transl.), 17 (1981) 223. 108. D. Batsuren, M.R. Yagudaev, E.Kh. Batirov and V.M. Malikov, Chem. Nat. Compd. (Engl. Transl.), 19 (1983) 17. 109. N.D. Abdullaev, M.R. Yagudaev, E.Kh. Batirov and V.M. Malikov, Chem. Nat. Compd. (Engl. Transl.), 23 (1987) 63. 110. S. Ghosal, S. Banerjee and A.W. Frahm, Chem. Ind. (London), (1979) 854. 111. T. Bieg and W. Szeja, Synthesis, (1985) 76. 112. W.K. Fife and Z.-D. Zhang, Tetrahedron Lett., 27 (1986) 4933, 4937. 113. R. Stevenson and J.V. Weber, J. Nat. Prod., 52 (1989) 367. 114. Private communication from Dr. Tesfaye Biftu (Merck Research Laboratories). 115. T. Ogiku, M. Seki, M. Takahashi, H. Ohmizu and T. Iwasaki, Tetrahedron Lett., 31 (1990) 5487. 116. M.K. Seikel, F.D. Hostettler and D.B. Johnson, Tetrahedron, 24 (1968) 1475. 117. B.F. MacDonald and G.M. Barton, Can. J. Chem., 51 (1973) 482. 118. T.R. Govindachari, S.S. Sathe, N. Viswanathan, B.R. Pai and M. Srinivasan, Tetrahedron, 25 (1969) 2815. 119. F.D. Hostettler and M.K. Seikel, Tetrahedron, 24 (1969) 2325. 120. A.F.A. Wallis, Tetrahedron Lett., (1969) 5287. 121. R. Ahmed, M. Lehrer and R. Stevenson, Tetrahedron, 29 (1973) 3753. 122. K. Freudenberg and H. Schraube, Chem. Ber., 88 (1955) 16. 123. E.J. Corey, N.W. Gilman and D.E. Ganem, J. Am. Chem. Soc, 90 (1968) 5616. 124. T. Momose, K. Kanai and K. Hayashi, Chem. Pharm. Bull., 26 (1978) 3195. 125. T.R. Govindachari, S.S. Sathe and N. Viswanathan, Tetrahedron Lett., (1967) 4183. 126. E. Block and R. Stevenson, J. Org. Chem., 36 (1971) 3453. 127. R. Stevenson and J.V. Weber, J. Nat. Prod., 54 (1991) 310. 128. G.C. Hokanson, Lloydia, 41 (1978) 497. 129. G.C. Hokanson, J. Nat. Prod., 42 (1979) 378. 130. M.G. Kelly and J.L. Hartwell, J. Natl. Cancer Inst., 14 (1954) 967. 131. J.L. Hartwell and A.W. Schrecker, Fortschritte der Chemie Organischer Naturstoffe, L. Zechmeister (Ed.), 15 (1958) 78. 132. V. Podwyssotsky, Arch. Exptl. Pathol. Pharmakol, 13 (1880) 29. 133. W.J. Gensler and CD. Gatsonis, J. Am. Chem. Soc, 84 (1962) 1748. 134. W.J. Gensler and CD. Gatsonis, J. Org. Chem., 31 (1966) 3224, 4004. 135. L Jardine in. Anticancer Agents Based on Natural Product Models, J.M. Cassady and D. Douros (Eds.), Academic Press, New York, (1980) pp. 319351. 136. T.W. Doyle in, Etoposide (VP-16) Current Status and New Developments, Academic Press, New York, (1984). 137. Ref. (9), pp. 90-95, pp. 113-137. 138. R.S. Ward, Synthesis, (1992) 719.

355

139. S.P. Dorsey, D. Rajapaksa, N.J. Taylor and R. Rodrigo, J. Org. Chem., 54 (1989) 4280. 140. L.M. Perry, Medicinal Plants of East and Southeast Asia, The M.I.T. Press, Cambridge, Mass., (1980) pp. 83, 150, 281, 292. 141. P.A. Ganeshpure, G.E. Schneiders and R. Stevenson, Tetrahedron Lett., 22 (1981) 395. 142. G.E. Schneiders and R. Stevenson, J. Org. Chem., 46 (1981) 2969. 143. G.E. Schneiders and R. Stevenson, J.C.S., Perkin I, (1982) 999. 144. P.A. Ganeshpure and R. Stevenson, J.C.S., Perkin I, (1981) 1681. 145. Ref. (8), p. 230 and references cited therein. 146. Ref. (8), Chapter 7. 147. Ref. (9), pp. 213-216. 148. Ref. (9), pp. 208-212 and references cited therein. 149. H.M. Bradley and D.W. Knight, J. Chem. Soc, Chem. Commun., (1991) 1641. 150. T. Mukaiyama, M. Usui and K. Saigo, Chem. Lett., (1976) 49. 151. R.E. Ireland and D.W. Norbeck, J. Am. Chem. Soc, 107 (1985) 3279. 152. S. Takano, T. Ohkawa, S. Tamori, S. Satoh and K. Ogasawara, J. Chem. Commun., (1988) 189. 153. H.M. Bradley, R.G. Jones and D.W. Knight, Synlett., (1992) 479. 154. N.K. Kochetkov, A. Khorlin and O.S. Chizov, J. Gen. Chem. (U.S.S.R.), 31 (1961) 3454 and subsequent papers in series. 155. N.K. Kochetkov, A. Khorlin, O.S. Chizov and V.L Sheichenko, Tetrahedron Lett., (1961) 730. 156. Y. Ikeya, H. Taguchi, L Yosioka and H. Kobayashi, Chem. Pharm. Bull., 27 (1979) 1383 and subsequent papers in series. 157. Ref. (9), pp. 170, 230-242, 257. 158. S.M. Kupchan, A.J. Liepa, V. Kameswaran and R.F. Bryan, J. Am. Chem. Soc, 95 (1973)6861. 159. T. Biftu, B.G. Hazra and R. Stevenson, J.C.S., Perkin I, (1979) 2276. 160. M. Taafrout, Y. Landais and J.P. Robin, Tetrahedron Lett., 27 (1986) 1781. 161. Y. Landais and J.P. Robin, Tetrahedron Lett., 27 (1986) 1785. 162. A. Pelter, R.S. Ward and A. Abd-El-Ghani, J. Chem. Soc, Perkin I, (1992) 2249. 163. M. Tanaka, H. Mitsuhashi and T. Wakamatsu, Tetrahedron Lett., 33 (1992) 4161. 164. M. Kupchan, R.W. Britton, M.F. Ziegler, C.J. Gilmore, R.J. Restivo and R.F. Bryan, J. Am. Chem. Soc, 95 (1973) 1335. 165. Ref. (9), pp. 344-354. 166. S. Yamamura, M. Niwa, M. Nonoyama and Y. Terada, Tetrahedron Lett., (1978)4891. 167. T. Kikuchi, S. Kadota, K. Yanada, K. Tanaka, K. Watanabe, M. Yoshizaki, T. Yokoi and T. Shingu, Chem. Pharm. Bull., 31 (1983) 1112. 168. L.P. Badheka, B.R. Prabhu and N.B. Mulchandani, Phytochemistry, 26 (1987) 2033.

356

169. S. Malhotra, S.K. Koul, S.C. Taneja, P. Pushpangadan and K.L. Dhar, Phytochemistry, 29 (1990) 2733. 170. A. Patra and A.K. Mitra, Indian J. Chem., 17B (1979) 412. 171. Y. Yuan, C. Wang and X. Zhou, Zhongcaoyao, 13 (1982) 387. 172. D.B. Saxena, Phytochemistry, 25 (1986) 553. 173. E. Al-Farhan, P.M. Keehn and R. Stevenson, J. Chem. Research (S), (1992) 36.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

357

The Chemistry of Natural Diarylheptanoids G.M. Keserii and M. Nogradi Research Group for Alkaloid Chemistry of the Hungarian Academy of Sciences Technical University of Budapest H-1521 Budapest P.O.B. 91, Hungary 1. Introduction 2. Phytochemistry of diarylheptanoids 3. Biosynthesis of diarylheptanoids 4. Biological activity of diarylheptanoids 5. Synthesis of diarylheptanoids 1. INTRODUCTION Diarylheptanoids constitute a distinct group of natural plant metabolites characterized by two aromatic rings linked by a linear seven-carbon aliphatic chain. They may be divided into two subgroups, i.e. open chain and macrocyclic diarylheptanoids. In the latter the aromatic rings are connected to form a diarylether or a biaryl moiety.

<0^^-^

Curcumin (1), yellow dye of Curcuma tinctoria and other Curcuma species was first mentioned as early as in 1808 by Trommsdorff and isolated in 1815 by Vogel and Pelletier (1) was not only the first diarylheptanoid ever to be discovered, but also among the first natural organic compounds prepared in a more or less pure state.

358 O

HO^

^^

O

1

^^

"OH

The constitution of this relatively simple and broadly used compound was established surprisingly late by Lampe and Milobedzka (2), first in 1910 by degradation followed in 1913 by a synthesis (3,4). Curcumin remained the only representative of the group as long as until 1964. In the following 30 years, however, about 70 linear and 35 macrocylic diarylheptanoids were isolated. In the present review we are dicussing their chemistry, phytochemistry, biosynthesis, biological activity, and synthesis. 2. PHYTOCHEMISTRY OF DIARYLHEPTANOIDS 2.1 Open chain diarylheptanoids Open chain diarylheptanoids were isolated from Acer nicoense (Aceraceae), from Alnus and Be tula species (Betulaceae), jfrom Alpinia, Curcuma and Zingiber species belonging to the family of gingers (Zingiberaceae) and finally from certain Centrolobium species (Leguminosae). Individual compounds will be classified according to the plant families in which they occur. 2.\A. Aceraceae From the stem bark of Acer nicoense Maxim, indigeneous in Japan, Nagai and his coworkers isolated two new glycosides of already known (-)-centrolobol (2) (5) i.e. aceroside VII (3) and aceroside VII (4) (6)'.

HO^ ^^

^-^

OH

2 R=H 3 R = p-D-Glcp 4 R = p-D-Apif-(l->6)-p-D-Glcp

* Glcp = glucopyranosyl, Apif = apiofuranosyl, Araf = arabofiiranosyl

359 Originally an S configuration has been assigned to (-)-centrolobol by Albuquerque et al (5), but NMR studies connected with the isolation of the glycosides indicated an R configuration for the aglycon (6). lA.lBetulaceae In the family of Betulaceae open chain diaryl-heptanoids were found in Alnus and Be tula species. 2.1.2.1. Alnus species Two optically active compounds (5 and 6), for which the fancy names yashabushiketol and dihydro-yashabushiketol were coined, were isolated by Asakawa et al. from A. flrma (7,8) and A. sieboldiana species (9), while from the former source Urarova and her coworkers isolated a levorotatory saturated diol 7 (10). Absolute configuration of (-)-7 was established to be S,S by Asakawa et al (11) by correlation with (iS)-6. In A. firma Suga et al discovered a dienone (8) (12) which could not be derived from diarylheptanoids isolated earlier. O

OH

{S.S)'l

O h

Ph

h

Ph

OH

8

In the flowers of A. sieboldiana Asakawa et al found a series of diols, triols, and ketodiols (13), namely the epimeric diols yashabushidiol A [(i^,*S)-7] and B [(/?,/?)-7], the epimeric ketodiols yashabushiketodiol A [(li?,55)-9] and B [(l.S',55)-9)], as well as yashabushitriol (10). Configuration of all of them has been established by NMR spectroscopy and correlation with (iS)-6.

(li?,55)-9

(l/?,27?,55)-10

Oregonin, an optically active ketol [(iSJ-11) is the diarylheptanoid component of several Alnus species and was first isolated as its xyloside in 1974 from A. rubra (14), but the absolute configuration of the aglycon remained unknown until it was also isolated from A.

360 serriilatoides and the problem settled by an X-ray analysis (15). Recently Lee and his coworkers isolated the glucoside of oregonin (12) (16) from A. hirsuta along with its interesting conjugate with ellagitannine (13, see on next page). O

ORi

J R2

HO^

'V^CN^r^^^-^-^V^^

HO"

KJJ

(^-11

lOl O

vOR2

H

H

(^-^2 P-D-Glcp

H

OGlcp

(^13

ooc

OH

OH

In a study on the colouring matters of Alnus species Nomura and his coworkers isolated a series of new open chain diaryIheptanoids from A. japonica (17), such as hannokinol (14) (18), hannokinin (15) (19) both of undetermined stereochemistry, and the enone 16, a dehydration product of 15. OH

HO

HO

OH

OH

361

OH

HO

Whereas 14 and 15 could also be isolatedfromA. hirsuta and from Betula platyphylla (20), the occurrence of the enone in these plants was not reported and therefore 16 may be an artefact formed during isolation. The dextrorotatory antipode of the widely distributed (-)-centrolobol (2) was first discovered by Sasaya in A. hirsuta (21) in the company of an interesting tetrahydronaphthopyrone derivative (17), being in evident biogenetic relationship to open chain diarylheptanoids.

17

2.1.2.2. Betula species The fnst diarylheptanoid from B. platyphylla, the glucoside platyphylloside (18) was isolated by M. Terazawa et ah (22). Configuration of the aglycon, i.e. platyphyllone (15) was determined by NMR studies by Ohta and his coworkers (23).

OH

(5)-18 R=p-D.Glcp (5)-19 R=p-D-Apif(l->6).p.D-GIcp (•S)-20 R = P-D-Apif(l->2)-p-D-Glcp

From B. pendula Smite et al isolated a two new apiose containing glycosides of (5)-15, namely 19 and 20 (24) along with two known (3 and 4) and one new glycosides (21) of (/?)centrolobol (2) (6). H^ vOR

HO

o

a

OH

(i?)-21 R = p-D-Apif(l->2)-[P-D-Apif-(l->6)]-p-D-Glcp

362 2.2.2. Zingiberaceae 2.2.2.1. Alpinia species Examination of A. officinamm by Itokawa and his group provided besides the known compound 6 (8) and the ketol 22 (25) two new enones, 28 and 29 (26). Circular dichroism studies revealed the interesting fact that, in contrast to the known components of Alnus firma andy4. sieboldiana^ which contained (S)-6 and (5)-22, in Alpinia officinamm the same alcohols were present as the R enantiomers. On further investigation the same plant also yielded a total of eight new achiral or racemic diaryIheptanoids: 23, 24, 31 (27), 25, 26, and 30, (28), 33 (29), 27 and 32 (30). Since the highly susceptible P-hydroxyketones can readily undergo dehydration and subsequent addition of methanol, the enones 28, 29, and 30, as well as the p-methoxyketones 23, 25, and 26 may be artefacts. This assumption is also supported by the racemic nature of 23, 25, and 26. O

ORi 22 23 24 25 26 27

II

o

Ph

P h - ^v ^

R3

OH OH OH H OH OH

0

O

O

28 29 30

R2 Rl H OMe Me OMe H H Me H H Me H OH

II ^J<^ "N^

^R2

-R3

Rl R2 H H OMe OH H OH

31 32 OH

Rl R2 OMe OH OMe H

OH

(/?./?>33

OH

Itokawa and his coworkers isolated from A. oxyphylla the closely related ketones yakuchinone A and B (34 and 35) (31, 32), while A. katsumadai yielded, apart from the Almis constituents 7 and 8, four new compounds (36-39) (33).

363 O

OMe

o

Ph

OH

OMe OH

35

34

QH

O

2.2.2.2 Curcuma species As we have mentioned in the introduction the first diarylheptanoid, curcumin (1) was isolated in 1815 by Vogel and Pelletierfi-omC. longa (1). Following the elucidation of its constitution and its synthesis (2-4) it was also submitted to X-ray analysis (34). Prompted by its colouring properties, its extended use in spices and later also by its promising biological actitvities C longa extracts were the subject of several studies. Thus Srivanasan reported in 1953 the isolation of two minor components, demethoxycurcumin (40) and bis-demethoxycurcumin (41) (35), their constitution was determined by Whiting et al. (36). The first saturated curcumin derivative, hexahydrocurcumin (S)-(41\ was discovered by Murata and his coworkers (37). Its enantiomer was later found in Alpinia officmamm and based on its CD spectrum an R configuration was assigned to this compound. Structural variation in the curcumin family was enriched by the discovery of dihydrocurcumin (43) by Ravindranath (38), and by Kiuchi, who isolated the first cyclic derivative, cyclocurcumin (44) (39). O

O

HO

OH

O

(5)^2

OH

40 41

Rl H H

R2

OMe H

364 O

MeO HO

O

^^r^--^:

MeO

OMe

HO Curcumin, as well as compounds 40-43 were also isolated from C xanthorrhiza by Uehara et a/. (29), who obtained from this source also two new compounds, the optically active diol {S,S)'4S and the racemic diketoalcohol 46. While formation of 45 can be envisaged by reduction of hexahydrocurcumin, racemic 46 might have been the result of addition of water onto the double bond of curcumin. Masuda and his coworkers reported in 1992 the isolation of an unsynunetrical curcumin derivative, 5-methoxycurcumin (47) from C xanthorrhiza (40). OH OH MeO^ ^^ ^^^ J^ J^ ^^ ^^^ ^OMe (^.^-45

OH

O

^OH

O

MeO

OMe OH

It is interesting that compound 8 isolated previously from Alnus firma (Betulaceae) and its reducedfrom(48) are also present in C xanthorrhiza (Zingiberaceae) (41). OH (48)

365 2.2.2.3. Zingiber species Diarylheptanoids from Zinziber officinale, the common ginger, the gingerones A, B, and C (49-51), as well as isogingerone B (52) were isolated by Endo and his coworkers in 1990 (42). Later Kikuzaki et al. isolated additional components, such as 53, the demethyl derivative of 49, further (5)-hexahydrocurcumin (42), from which 49 can be derived by dehydration (43). It is noteworthy that apart from the diastereomeric diols {S,S)-45 and (R,S)'45, arising from the reduction of 42, the corresponding acetates, 54 and 55, could also be isolated (44). The same authors later described two more diaiylheptanoid acetates (56, 60) and a series of racemic P-hydroxyketones (57-59) (45). Considering that P-hydroxyketones readily eliminate water, it cannot be excluded that the enones 49-53 are artefacts.

,OR»

MeO

OH

49 50 51 52 53

OAc

R3 R2 Rl H OMe H OMe OMe H

H

Rl H

H

54

.OR2

meso >-55

HO

O MeO^

HO"

56

OH

R2

H Me OMe H

OH

0r ^X'V^

O

OMe H

OAc

MeO

H

OMe H OH H

^R^

T^ k^ O H 1 R2

Rl H

R2

R3

57

H

H

58 59

OMe H

H OMe

OMe OMe

OAc OAc

HO

OH

HO

OH

Recently another species of the same family Z cassumar was investigated by Masuda et al. who isolated from this plant three novel diarylheptanoids, the cassumins A, B, and C (61-63) (46), in which the diarylheptanoid sceleton was extended by an aiylbutenyl unit.

366 Three more compounds of presumably similar biogenetic origin, cassumunarin A, B, and C (64-66) were found in the same plant by Jitoe and his associates (46).

O

O OMe

MeO OMe

OMe 61 R = H 62 R = OMe

OMe MeO

OMe

Me

OMe OMe

'

^

•

•

'

^

OMe

OH

OH

OMe

OMe

^

"OMe OMe

65 66

R=H R = OMe

OMe

2.2.3. Leguminosae Among leguminous Centrolobium species diarylheptanoids were first isolated from C rohustim in 1964 (5). It contains (-)-(/^)-centrolobol (2), found later in several other plants too {Acer nicoense (Aceraceae) (6), Betula pendula (Betulaceae) (24)) as well as two cyclic components, (3/?,75)-centrolobin (67) and (3/?,75)-de-0-methylcentrolobin (68) (47). First

367 an S configuration was assigned to (-)-2 (5), but this was corrected later to R (6). Stereochemistry of the cyclic components 67 and 68 were cleared as late as in 1984 by Jurd and his coworkers (48). A systematic study by Gottlieb et al established the surprising fact that, in contrast to C robustum, C paraense and C sclerophyllum contains centrolobol and de-O-methylcentrolobin in the antipodal form [(5)-2, (3^,77?)-68] (49), while (3SJRy centrolobin (67) was only identified in C. tomentosum.

67 R = H 68 R = Me HO

2.2. Macrocyclic diarylheptanoids Macrocyclic diatylheptanoids can be derived from their open chain congeners by oxidative phenol coupling resulting in macrocyclic biaryls or biaryl ethers. Usually they were named after the plant source and can be conveniently classified according to the plant families in which they occur. l.lA.Aceraceae From Acer nicoense, a maple indigeneous in Japana series of macrocyclic diarylethers, mainly in form of their glycosides were isolated predominantly by Inoue, Nagai and his coworkers. Thus acerogenin A (72) (50) is the common aglycon of acerosides I (69) (50), III (70) (51), and VI (71) (51). On further study of A. nicoense Kubo and his group discovered a second aglycon, acerogenin B (73) (52) but none of its glycosides. Aceroside IV (74) and its aglycon acerogenin C (75) (53), as well as aceroside V (76) and its aglycon acerogenin D (77) isolated 10 years later, are also diaiylether type macrocyclic diarylheptanoids.

RiO

69 70 71 72

Rl p-D-Glcp H H H

R2

H p-D-Apif-( 1 ->6)-p-D-Glcp p-D-Glcp H

368

74 R = p-D-Glcp 75 R = H 73

76 77

R = p-D-Glcp R=H

78 R = p-D-Glcp 79 R.= H

From Acer nicoense the first biaryl type macrocycle, aceroside XI (78) and its aglycon acerogenin E (79) was isolated by Nagumo et al in 1993 (54). Although in other plants several diaryl type macrocyclic diarylheptanoids have been isolated, the cooccurrence of 79 and its ether type analogues confirmed the hypothesis about their biogenesis by oxidative phenol coupling. 2.2.2. Betulaceae 2.2.2.1. Alnus species From Alm4S species i.e. varieties of alder, containing also a large variety of open chain diarylheptanoids (7-21), some biaryl type macrocyclic compounds have also been isolated. In Alnus japonica Steud. indigeneous in Japan Nomura et al. found four diphenols of this kind. The constitution of alnuson (80) and alnusoxide (81) was elucidated in 1975 (20), while alnusdiol (82) and its oxidation product, alnusonol (83) were characterized in 1981 (17). It was observed that 81 could only be isolated from the dried plant and may be therefore an artefact. Alnusdiol was optically active and therefore a /mm-diol but its absolute configuration remained unknown. From A. hirsuta only the enone 82 and the ketol 83 could be isolated (21) also suggesting that the oxide 81 may be an artefact.

369

80

81

OH

M

82 X =
^^

«3 X = =0

pjo

] ^

1

Pk,

VzzO /

ftd

^

ot

O

Me

2.2.2.2. Carpinus species Compounds 80 and 93 were reisolated by Sawa and his coworkers from C. cordata along with a new biaryl type macrocycUc diarylheptanoid (84) containing an unusual diol acetonide ftmction. 2.2.2.3. O^/rya species As the first known representative of macrocycUc diarylheptanoids the ketotriol asadanin (85) as well as the epimeric tetrols, asadanin I (86) and II (87) derived by reduction of the keto group were isolated from O. japonica by Yasne et al in 1965 (56). The complete stereostructure of asadanins is still unknown, but interestingly it has been established that in the tetrols configuration at C-4 is the same, but different at C-2. Therefore it can be assumed that asadanin itself may occur in Nature in two epimeric forms. OH

OH OH

85 86 87

X= =0 X = OH X = epi'OH

370

2.2.3. Burseraceae Garuga pinata Roxb. and Garuga gamblei King, are indigeneous in India and are widely used for the preparation of traditional medicines. Garuganins and garugamblins are constituents of these plants and both biaryl and diaryl ether type macrocyclic compounds can be found among them. Of diaryl ether type are garuganin I (88) (57), III (95) (58), IV (89) and VI (93) (59), further garugamblin 1 (90) and 2 (91) (60-61), while in garuganin II (92) and its isomer, garuganin V (94) (59) the aromatic rings are linked by a C-C bond.

88 89 90 91

R3 Rl R2 OMe H OMe H OMe H H OMe H OCH2O

OMe MeO

^OMe

OMe 92 The constitution of 88 (62), 91, 92 (63) and 93 (64) has been confirmed also by X-ray crystallography. Based only on NMR evidence to garuganin III Mishra et al. assigned structure 89, but a total synthesis of this compound and its isomers by Keserii et al (65) proved that the correct constitution of garuganin III was in fact 95. OMe MeO

95

371 2.2.4. Casuarianaceae From C junghuhniana Aoki and his coworkers isolated besides alnusoxide (81) another macrocyclic compound casuareinondiol (96) (66). 2.2.5. Myricaceae From Myrica nagi Whiting and his coworkers isolated in 1970 two biaryl type diarylheptanoids, myricanon (97) and myricanol (101), as well as the glucoside of the latter (104) (67). Maltenid et al. studied M gale L. a species growing in Scandinavia and found three closely related compounds, porson (98), galeon (99) and hydroxygaleon (100) (69). Porson was later reisolated from M rubra by Takeda et al (70) in which also five myricanol glycosides (104-108) all from M rubra (71, 72) and 5-deoxymyricanon (102) (70) were found. 13-Oxomyricanol (103), a constituent of M nagi (73) is the most highly oxygenated compound in this series. R6

R7

R40 R3a

MeO

97 98 99 100

Rl H OMe H H

101

H

102

H

R2

R3

H OMe H H H H

H Me Me Me H H

R4 Me H H H Me Me

R5

R6

OMe H H H OMe OMe

OH H H H H OH

R7 H OH H OH H H

X H H

=o =o H H

Y

=o =o H H

=o

OH

R=H R = p-D-Glcp R = p-D-Glcp.OCC6H2-3,4,5.(OH)3 R = p-D-Glcp-(l->6)-P-D-Glcp R = P-D-Glcp-(l-^3)-P-D-Glcp R = P-D-Glcp-(l->6)-a-L-Araf

Z H H H H H =0

372

Whiting and his coworkers observed that myricanon (97) rearranged under the action of Lewis acids giving isomyricanon (72). Based on NMR spectra the structure suggested for isomyricanon by Whiting et al. (109) has been later revised by Sakurai et al. to 110 (72). OH

OMe MeO

109

110

2.3. 9-Phenylphenalenones At first sight the title compounds characterized by the general formula 111, are rather unrelated to diarylheptanoids. At closer look, however, the two classes of compounds can be linked by a very plausible biosynthetic hypothesis first forwarded by Thomas (80) and shown in the following scheme below. 9-Phenylphenalenones were isolated from Haemodorum (74), Lachnates (75, 76), Xiphidium {11\ Wachendorfia (78), and Anigozanthos (79) (Haemodoraceae) species.

OH A diaryIheptanoid dienone (112) biosythesized from tyrosine and phenylalanine which, according to Bazan et al. (81), undergoes oxidation to form from the oxygenated aromatic ring an ortho-(\\x\nonQ (113). The latter, being an active dienophile, undergoes intramolecular Diels-Alder cycloaddition. Aromatization of the adduct finally gives a 9phenyl-phenalenone. In fact periodate oxidation of 112 gave lachnanthocarpone (114), a typical 9-phenylphenalenone (82).

373

3. BIOSYNTHESIS OF DIARYLHEPTANOIDS Two hypotheses have been proposed for the biosynthesis of diarylheptanoids. Geissmann pointed out that a common feature of linear diarylheptanoids is oxygenation of the seven membered aliphatic chain at positions 3 and/or 5, further double bonds at positions 1 and/or 6. This suggested that in the course of their biosynthesis diarylheptanoids are formed by the attachment of two cinnamate units to the central carbon atom of a malonate (82). Scheme HO^ ^

MeO

^

I L - ^

HO, ^^

.S.CoA

"^

Jj^ , . ijCOS.CoA

w

MeO

CO2H Js!s-/J

OH CoA.S.

OMe OH

^^A^^y^yjOC OMe

In order to check Geissmann's proposal by experimentally Whiting et al. (36, 83) injected [l-l^C] and [2-l^C] sodium acetate and malonate, [l-^^C]- and [3-l4c]-phenylalanine, and sodium ^H-ferulate into the roots of Curcuma longa. After a few days of incubation

374

curcumin was isolated from the roots. Alkaline degradation established the incorporation of all of the l^^C labelled precursors (see Table 1). Table 1. Incorporation of ^"^C labelled precursors into curcumin and acerogenin A Compound

Incorporation (%)

[ l-14C]-phenylalanine [3 -1 ^C]-phenylalanine Na-[2-14c]-acetate Na-[2-14C]-malonate Na-[l-l^C]-acetate Na-[l-l4c]-malonate

curcumin (1)

acerogenin (79)

0.103 0.354 0.032 0.015 0.038 1 0.043

0.036 0.052 0.039 0.054 <0.001 -

Alternating distribution of radioactivity along the aliphatic chain led to the conclusion that after the condensation of one cinnamate unit with five malonate units the second aromatic ring was formed by cyclization of the three terminal polyacetate units. Final decision between the two hypotheses requires further experiments. Ar

S.CoA

CO2H CH2CO2S.C0A

OH

TYY'u OH

OMe

Macrocyclic diarylheptanoids can be derived from linear ones by oxidative cyclization. This is supported by the fact that in all of the biaryl type representatives the aryl-aryl bond is in meta position relative to both ends of the seven-carbon chain, while in those containing a diarylether bond a meta,para bridging can be identified. This substitution pattern corresponds to activation of the position ortho to the phenolic hydroxy group. Biosynthesis of cyclic diarylheptanoids was investigated by Inoue et al. (84). They fed [l-^^C] and

375 [2-14C] sodium acetate and malonate, [1-l^C]- and [3-l^C]-phenylalanme, and [l^CJ-methyl methionine to yoimg shoots of Acer nikoense. After a few days acerogenin A (72) was isolated and incorporation of [l-l'^C] acetate and malonate but not that of [l-^^C] acetate was found (see Table 1). Elimination of the aliphatic chain by permanganate oxidation gave a diacid (115) wich exhibited only 3% of the radioactivity of the parent compound (72) obtained from [l-l'^C] acetate and malonate. This indicated that radioactivity from these sources appears in the central atom (C-10) of the seven carbon chain. Poor incorporation of labelled methionine confirmed that C-10 did not come from a methyl group. A comparison of incorporation of [l-^'^C] and [2-l^C] acetate into curcumin and acerogenin A suggests that the biosynthesis of the two compounds may be different. With curcumin both [l-l'^C] and [2-14C] acetate and malonate were incorporated, thus presumably one cinnamate and five malonate units were joined, while acerogenin A should be formed by oxidative cyclization of (-)-centrolobol (2) built upfromtwo p-coumarate and one malonate imits.

4. BIOLOGICAL ACIVITY OF DL\RYLHEPTANOIDS Several of the plant species containing diarylheptanoids are widely used in traditional medicine. Ginger, the rhizome of Zingiber officinalis, is one of the best known ingredients of preparations used in the folk medicine in the Far East (44). Similarly to Zingiber species, ctrXdimAlnus species also yield characteristic aromatic extracts (13) and the seeds oiAlpinia katsumadai have been applied as aromatic stomachic (33). The rhizomes of Alpinia officinarum are used both in traditional Chinese medicine (26) and in Japanese "kampo" medicine for the treatment of gastrointestinal disorders (85). An extract called "yakuchi" prepared from the fruits oiAlpinia oxyphylla had been used originally in China for the same purpose, but its use later has become universal in the entire oriental medicine (31). In European ethnomedicine certain Curcuma species were recommended as choleretic drugs, but were also used as spices. Their use as cosmetics in Indonesia was also reported. From the leaves of Acer nikoense an eyewash was prepared in Japan, but the plant was also known as a remedy for certain hepatic disorders (54). Several applications of Garuga species, indigeneous in India have been described. Thus the extracts of the leaves of G. pinnata and G. gamblei were used in admixture with honey for the treatment of asthma, while the bark extract was applied against the opacity of cornea (57). The decoction of the roots proved to be an effective remedy of pulmonary infections (58). Some Myrica species, also common in India, are well known ingredients of traditional medicine in the Far East and their piscidal activity has been reported (86). The isolation of plant constituents in the pure state prompted several pharmacological studies, which will now be summarized according to their activity profile.

376 4.1. Antiinflammatory activity Since the introduction of non-steroidal antiinflammatory agents and the elucidation of their mechanism of action it has become well known, that such compounds (as e.g. aspirin or indomethacin) act through the inhibition of fatty acid cyclooxygenase (87). Following the elucidation of the first steps in prostaglandin (PG) biosynthesis several non-classical PG modulators of arachidonic acid metabolism (thromboxanes, prostacyclin, fatty acid peroxides, etc.) have been identified. These compounds are not only antinflammatory agents, but exhibit several other biological activities too. For this reason the other biological activities of PG inhibitory diarylheptanoids will also be discussed here. Inhibition of PG biosynthesis by open chain diarylheptanoids was first reported by Itokawa and his coworkers (26, 31). Their results were implemented by studies on additional compounds first by Flynn (88) and later by Kinchi et al. (85). The latter group extended the assay of inhibitor activity also to the 5-lipoxygenase enzyme system. Their results are summarized in Table 2. Among macrocyclic diarylheptanoids only for garuganins (88-95), isolated from Garuga pinnata was some antiinflammatory activity claimed, but no specific data were disclosed (57, 58). Besides inhibiting PG synthetase and 5-lipoxygenase some open chain diarylheptanoids also inhibit the 15-lipoxygenase mediated peroxidation of linolenic acid as well. Curcumin (1), further compounds 40 and 41 inhibited at concentrations of 0.3, 6.7, and 13.3 mg/ml the activity of the 5-lipoxygenase enzyme system (89-92). Using a mixture of the above compounds Toenensen demonstrated their synergism, and that the mechanism of action of 40 and 41 was based on mutual noncompetitive activation (93). Masuda and his associates reported 98, 90 and 97% inhibition of 15-lipoxygenase by cassumunin A (61), B (62), and compound 63, at 135 mM concentration (46). Several open chain diarylheptanoids, such as 1,14, 35, and 36 also inhibited the cyclooxygenase enzyme system (88). Table 2. Inhibition of PG synthase and 5-lipoxygenase (5-LO) by some diarylheptanoids (IC50 values in \\M) Compound 1 6 11 7,2 23 24 27

PG-synthase 5-LO -

8.0

170

-

-

4.4 1.6

4.4 2.3 19

-

-

0.26

Compound 29 31 34 35 36 42

PG-synthase 2.0 2.0 0.5 2.3 170 23

5-LO 0.41 5.4 4.0 3.2

The inhibition of arachidonic acid metabolism by diarylheptanoids can be attributed to their antioxidant properties. Studies on structure-activity relationships demonstrated that hydroxyl groups were responsible for the antioxidant activity, while an unsubstituted phenyl

377

ring in unsymmetrical diarylheptanoids was an essential condition for a hydrophobic interaction between the enzyme and its inhibitor (85). According to Osawa (93) and Larson (94) the antioxidant potential of curcumin (1) and its congeners is due to the chelation of metal ions by the p-diketone moiety, while in the opinion of Cuvelier and his coworkers (85) this effect can be explained by derealization between the keto group and the phenoxy radical generated during oxidation. By ESR Schaich and his associates identified the presence of a phenoxy and an alkyl radical (96). The former arose by the homolysis of a phenolic hydroxy group while thelatter may be formed by electron transfer from the double bond of the enol form. 4.2. Antihepatotoxic activity The extract of the rhizomes of Curcuma longa, called "ukon" showed significant hepatoprotective activity when rat hepatocytes where challenged with CCI4 or galactosamine (97). This observation encouraged Mikino and his group to determine the antihepatotoxic activity of pure components isolated from C. longa and other members of the Zingiberaceae family (98) (see Table 3). Table 3. Effect of some diarylheptanoids on the cytotoxic effect of CCI4 and galactosamine (GalN) in rat hepatocites at 1 mg/ml concentration Compound

Control 6 7 8 11 15 22 24

Activity (GPT%) CCI4

GalN

100 80 63 92 32 44 31 27

100 13 70 74 7 72 66 54

Compound

29 31 35 36 38 39 42 54

Activity (GPT%) CCI4

GalN

44 47 29 93 72 76 33 79

50 76 132 69 71 102 84 8

It can be seen from the table that compounds lacking a free phenolic hydroxyl group were effective protectants against CCI4 induced citotoxicity. Unsaturation of the seven-carbon chain and its oxygenation did not affect activity. For compounds substituted at the aromatic ring it was found that compounds with a free hydroxy group were more active than analogues with an imsubtituted phenyl ring or those in which the hydroxy group was methylated. Interestingly, in contrast to the antiinflammatory activity associated with the inhibition of PG synthase, as well as 5- and 15-lipoxygenase, for antihepatotoxic activity the presence of afreehydroxy group is not essential.

378 4.2.3. Antifungal and antibacterial activity There is only a single report on the antifungal activity of diarylheptanoids namely in the case of gingerenones A, B, and C (49-51) as well as isogingerenone B (52) isolated from Zingiber ojjicinarum which Endo et al. found to be moderately effective (42). In vitro at 10 ppm concentration 49 showed 20% anticoccidal inhibitory activity in the case of Eimeria tenella NAIH, complete antifungal inhibiton of elongation of invading hyphae of Pyricularia orizae above 10 ppm concentration and halved the appressoria formation at 10 ppm. The only diarylheptanoid showing antibacterial activity so far is centrolobin (68), as reported by Goncalves and his coworkers (99). An interesting observation is that on intensive irradiation by light curcumin (1) became highly phototoxic against Salmonella typhimurium and Escherichia coli (100). Among cyclic diarylheptanoids garuganins and garugamblins isolated from Garuga pinnata and G. gamblei (88-95) exhibited antibacterial activity (57, 58). Activity and mechanism of action of the pure isolated constituents is unknown. A rationalization of the antibacterial activity of Garuga components has been attempted by Keseruand Nogradi (101). Garuga components having antibacterial activity contain a 15-membered macrocycle which shows a suprising structural similarity to ansamycin antibiotics. Incorporation of an aromatic system into a single macrocyclic ring joining the aromatic rings at distant points is a common feature of both the ansamycin antibiotics, such as e.g. rifamycin SV (116) and garuganins and garugamblins. CH3

CH3

CH3

OCH2COOH

Rifamicin B

OH

Rifamicin S

Rifamicin SV (116)

Computational studies established that garuganin I (88) as well as the conformationally rather similar garuganin III (95), further garugamblin-1 and -2 (90 and 91) exhibited several

379 structural features similar to ansamycins. Thus fitting of structures 88 and 116 was 0.7 A/atom, i.e.quite close. The 1,3-dioxy function, regarded as the pharmacophore in ansamycins (102, 103) is well mimicked by 88 and therefore it can be assumed that it is an antibacterial agent with a similar mechanism of action. Electrostatic potential maps of 88 and 116, were, however different, suggesting that the level of activity of Garuga components was lower. Steric fit of the ansa bridge in the diaryl type component 92 and 116 was much less (1.2 A/atom) in accordance with the fact that no antibacterial effect was found for this compomid (64). 4.4. Pungency From plants belonging to the Zingiberaceae family several pungent compounds (gingerones, shogaols, aparadols) were isolated (104, 105). The first pungent diarylheptanoids, yakuchinones A and B (34, 35) were isolated by Itokawa et al from Alpinia oxyphylla (21, 31, 32). Limiting concentrations for pungency of 7.85 and 6.34 fiM were found for 34 and 35 respectively (106). Tests with synthetic derivatives of 34 and 35 revealed that unsaturation in the chain decreased, while phenolic hydroxy groups increased pungency. A complete lack of free hydroxy groups or their exchange for methoxy groups diminished or completely eliminated pungency. 4.5. Other effects The total extract of the rhizomes of Curcuma longa showed nematocidal activity against second stage larvae of dog roundworm (Toxocara canis) (107). Inactivity of the pure components (1, 40, 41, 44) in contrast to activity of mixtures in various proportions prompted Kiuchi and associates to postulate synergism of the constituents, Itokawa et al reported that the extract oi Alpinia ojficinarum inhibited the contraction of guinea pig ileum (26) induced by BaCl2 and histamine at 0.1 mg/ml concentration. Curcumin (1) and its derivatives increased the stability of ftuosemide, clonazepam, and nifedipin in serum (108), most conspicuously in the case of nifedipin, half-life of which was increased sixfold. This stabilizing effect can be exploited in the formulation of these light sensitive compounds (109). 5. SYNTHESIS OF DIARYLHEPTANOIDS 5.1. Synthesis of open chain diarylheptanoids In view of the constitution and chemical properties of diarylheptanoids a genersal strategy for their synthesis should satisfy the following critetia (110): i) Ready variation of the aryl end groups with either a free or blocked phenolic hydroxyl at both or either ends of the chain.

380

ii) incorporation of 3- and/or 5-oxygen functions into the chain and iii) possibility of introducing radiolabels into the chain for the purposes of biosynthetic studies. Methods for the synthesis of natural diarylheptanoids may be classified according to the ways the seven membered aliphatic chain is elaborated. The first synthesis of a natural diarylheptanoid, i.e. of curcumin (1) (the only compound of this type known at that time), was described by Lampe and his coworkers (4). They condensed carbomethoxyferuloyl chloride (117) with ethyl acetoacetate, decarboxylated the product (118) to the diketone 119 which was then repeatedly acylated with the acid chloride 117. Acid catalyzed deacetylation to 119 followed by removal of the blocking groups gave finally curcumin (1). O

O

CH. 02Et

O

^)H^ ,

O

l)anisol/Na 2) 117

rrV^ 119

OMe O

^ ^

O

A ^ AcOH Me02C0

)2Me

J^ )Me

120 O

)Me O

121 R = C02Me 1 R=H

A more generally applicable scheme was elaborated by Whiting and his group (110). The aryl-C4 fragment was obtained by Friedel-Crafts acylation of a suitably substituted benzene derivative. The resulting keto acid (122) was subjected to Clemmensen reduction and the product (123) was carried through a series of standard transformations giving the Grignard compound 126 which was reacted with a variety of aryl-C3 partners, e.g. with the

381 cinnamaldehyde 128 thus giving access to derivatives oxygenated at C-3. Alternatively the dithioacetal 129 was alkylated with a iodo compound (127) to give, after deprotection, the same end products. R3

AlCU succinic anhydride ^IQ

RiQ

R3 R2 RiQ

Q

X

O

CO2H 122

l)MeOH/H+ r 123 2)LiAlH4 1

pl24 PBr3| -125 Mg, Et20 ^126 ^127

RO

Zn/Hg^ HCl

X = CO2H X = CH20H -1 X = CH2Br

l)TosCl, 2)LiI

X == CHzMgBr X = CH2I

-^

382 Aldehyde 128 was also utilized for the preparation of the oxyrane 131, which gave, when reacted with the anion of the dithioacetal 129, 3,5-dioxygenated compounds such as 132 (111) (see preceding page).

DNaH/DMSO, "'

2)Me3S-I-

f f V ^ ^ ^ ^ > V ^

"»

OR HO

,

THF,BuLi

HgCl2 HgO

O

RO For the purpose of biosynthetic studies Whiting el al. prepared some labelled compounds too. First [l-^'^C] acetic acid was transformed to the dihydrooxazin 133 which was then alkylated with a benzyl chloride (134). Methylation of the product (135) gave an oxazinium salt (136). Reaction of the salt with the Grignard compound 126 concluded the construction of the diarylheptanoid sceleton labelled at the oxygenated carbon of the C7 chain. By this strategy several open chain diarylheptanoids were prepared by Whiting and his coworkers, among them mc-aO-dimethylcentrolobol (138, R = R^ = Me, R2 = R3 = H).

383 Syntheses by Sakakibara and associates (112) was based on substituted arylpropanoic acids (139) which were transformed in several steps into the ylide (140). Wittig reaction with substituted cinnamaldehydes gave the enones 141. In this way several natural diarylheptanoids were synthesized (17, 81).

Ar2-CH=CH-CHO BuLi 5.1.2. C5 + C2 schemes This strategy was first exploited by Miyashita and his group for the synthesis of yashabushiketol (5), yashabushiketodiol A (meso-l) and B {rac-l), and yashabushitriol (10) (113). First 5-phenyl-2-penten-l-ol (142) was subjected to Sharpless epoxidation in the presence of diethyl (/?,/?)-tartrate [(+)-DET] to give the (2iS',35)-epoxyalcohol 143. Swem oxidation of 143 yielded the aldehyde 144 as the C5 component to which styrylcerium(II) chloride was added providing the epoxyalcohol 145. Repeated Swem oxidation gave an epoxyketone (146), which could be converted either by chemoselective reduction of the epoxy function to 5 or, via the diepoxyde 147 to a mixture of epimeric ketodiols (9) separable by chromatography. Finally reduction of 9 gave the triol 10 and as C-3 epimers in a ratio of 2:1. •V-'CC^M

ph-^^^'

OH

(COCl)2,DMSO pj^ CH2Cl2,-60OC

o

O

146

147

Ph

I Na+[PhSeB(OEt)3]5

Na+[PhSeB(OEt)3]-1 10 -^

9 + e/7i-9

Kato and his coworkers constructed the aryl-Cs unit by DEPC [(EtO)2P(0)CN)] mediated C-acylation of/-butyl cyanoacetate with phenylpropanoic acid leading, as the key intermediate, to the acetalized beta-ketoaldehyde 148 (114). Addition of 2-phenylethyl magnesium bromide followed by deprotection of the carbonyl group completed the synthesis of the racemic P-ketol 6. Acid catalyzed elimination of water from 6 finally provided the enone 28.

384

NCCH2C02tBu "^^^H DEPC, Et3N

Ph'

O C02tBu

Ph

1)145 PC 2) (CH20H)2

N O

Jo

iBujAlH

O^

^O

PhCH2CH2MgBr

148

PhTosOH, C6H6

Treatment of the aldehyde 148 with lithium phenylacetylide yielded another versatile intermediate (149), which could be transformed, after stereoselective partial reduction to the corresponding trans allylic alcohol (150) , either by simple deprotection to racemic 36 or first by oxidation and then deprotection to the diketone 39. Elimination of water from the ketol 36 gave 8, which could be reduced with diisobutyl aluminum hydride without affecting the double bond to the dienol 48 .

PhC^CLi

385 5.1.3. C^ + Ci schemes Itokawa et al. based their synthesis of yakuchinone A (34) and B (35) on a C^ + C] scheme (106, 115). Acetone and cinnamaldehyde was condensed to a dienone (151) giving after hydrogenation the Ar-C5 unit (152). Another way to 152 was the reaction of phenylbutanoyl chloride with dimethyl cadmium. Condensation of the dienone with vanilline yielded 35, and hydrogenation of the latter gave 34 (see next page). Looking at the condensation of the ketone 152 with a variety of aromatic aldehydes it was foimd that while with aldehydes lacking a free hydroxy group the traditional method using concentrated alkali was successful, in the presence of a free hydroxyl a mixture of acetic acid and pyrrolidine proved to be the best. The method was exploited by Itokawa and his group for the synthesis of several natural diarylheptanoids. Another example for the €5 + Cj strategy is the method published by Arrieta et al (116) in which the enedione 153 was reacted with aromatic aldehydes. The yield of this condensation could be improved by Mann et al. (117) using the magnesium enolate of the enedione 153 in the presence of tributyl borate and butylamine (117). O CHO Ph - ^ ^ ^

MeXO ^^2'

cs. J L ^^>^

Me

H2/Pd ^

•

OMe H2/Pd

vaniUine AcOH, pyrrolidine

OH

OMe OH O ArCHO P h / % / \ / ^ M e 153

O

O

Ph^^^^^>^"^^--^^^

5.1.4. C1 + C5 +Ci schemes The use of curcumim as a dye required the elaboration of a practical industrial synthesis of the compound. The first procedure with this aspiration was based on the fmdings of

386 Pavolini (118) who prepared curcumin by condensation of acetylacetone with vanilline. He observed that Knoevenagel condensation, one of the main side reactions, could be suppressed by adding boric oxide. This method was further developed by Pabon (119) who first prepared a complex from acetylacetone and boric anhydride and reacted the product with vanilline in the presence of triisopropyl borate and butylamine. Yields calculated on vanilline were thus improved from 10% to 80%. Besides several other indutrial applications (120-123) the method was also used for the synthesis of dihydroyashabushiketol (6) (11) and the tetramethylether of oregonin (11) (14). By selective alkylation of acetylacetone at C-3 Pedersen and associates exploited the above method for the preparation of a series of non-natural diarylheptanoids substituted at C-4 (124).

Me^^^--^^MeB203,B(OBu)3,BuNH2

^ri^^^^^^:^-^^^

5.2. Synthesis of macrocyclic diarylheptanoids 5.2.1. Biaryl type compounds The central problem of the synthesis of macrocyclic compounds in general is cyclization of the macro ring. In its pioneering synthesis of alnusone dimethylether (111) and alnusone (80) Semmelhack (125) coupled macrocylization with the formation of the diaryl bond using Ni(0)[PPh3]4 mediated coupling of aromatic halogen compounds (111). Whiting et ai applied the method for the synthesis of myricanone (97) and myricanol (98) (126). The main features of this scheme will be illustrated by the synthesis of myricanone. First applying their C4 + C3 scheme (see earlier) the ketone 154 was prepared and transformed using l2/AgOCCF3 reagent to the diiodo compound 155. Ni(0) catalyzed cyclization to 156 and finally debenzylation gave myricanone (97). Reduction of the ketone 155 to the alcohol and carrying through the corresponding acetate through the same sequence of reactions yielded racemic myricanol (106). It has to be noted that 97 and 98 could not be obtained, by a biomimetic way, i.e. oxidative cyclization of the corresponding open chain phenols 137 and 138 respectively. In contrast the bromo compound 157 prepared from 154 could be induced to undergo photocyclization yielding a mixture of myricanone and its mono and dibenzyl ethers. Whiting's method for cyclization was applied by Semmelhack et al on the linear diaryIheptanoid 132, obtained by the C4 + C3 scheme elaborated by the same authors.(lll).

387

MeO Ji AgOCCFj

MeO

OBn

MeO Ni(0)[Ph3P]4 MeO

X

OBn

156 R = Bn 97 R = H

OBn 154

Br, AcOH

MeO MeO

OBn hv NaOH/MeOH myricanone and its benzyl ethers

5.2.2. Diaiylether type compounds The first synthesis of a natural diarylether type macrocyclic diarylheptanoid was reported by Keseru et al. (127). Although Whitmg and his coworkers using Ti(02CCF3)3 succeeded in the oxidative cyclization of the phenol 137 to give a macrocyclic diarylether (126), the method could not be extended to natural compounds. Schemes for diarylether t3pe macrocyclic diarylheptanoids will be illustrated by the synthesis of some Garuga components.

388 5.2.2.1. Synthesis of garugamblin-1 (91) by Wurtz type cyclization [127] After the failure of cyclization reactions by intramolecular condensations, a modified version of the Wurtz reaction was fnst selected as the key reaction in synthesis of Garuga components. 1,3-dioxo functionality was introduced in a masked form as an isoxazole ring. First the diarylether 158 was prepared by reacting isovanilline with methyl 4-bromobenzoate. 158 was condensed with the phosphonium salt 159 obtained in three steps from the readily available methyl 5-methylisoxazol 3-carboxylate. CHO

UeO

^r^

CuQ, 140^C ^ pyridine

Br"'^^-^

^C02Me

/ o. //-bromo-succin- ^ 9 ^v phjP M e - V ^ C O . M e -'''«-^v. CCl, g ^ ^ ^ ^ ^ c O ^ i ^ i T M e

9 ^\ A A c O . M e 159 With the aid of this practical C5 synthon introduction of the remaining carbon atoms of the C7 chain can be realized in a single step. Wittig reaction of the phosphonium salt 159 with the aldehyde 158 gave in good yield an olefm (160), which was first hydrogenated to 161

^^ l^« " 1 "

^

^

O—N >^

^ 1 6 0 X - X = CH=CH,R = CO,Me

DMSO 161 X-X = CH2-CH2, R = C02Me

MeO

LiAlH^ r 1^162 X-X = CH2-CH2, R = CH2OH

PBr3 [^ 163 X-X = CH2-CH2, R = CH2Br

\

MeO Na Ph,C=CPh3

N < ^

i

O

NH2

O

O U

O U

OMe O O OMe ,^,^,

(£)-168

O

I

OMe

XJ^

MeO

389 and then transformed t in two steps to a dibromide (161->162~>163). The dibromide was cyclized under high dilution conditions in a modifed Wurtz reaction mediated by the radical anion formed from sodium and tetraphenylethene. Cyclization was accompanied by the reductive cleavage of the isoxazole ring to give, as the primary product, the enamine 164 which was hydrolysed to the diketone 165. The characteristic P-methoxyenone functionality of garugamblin-1 (99)was elaborated by methylation of the diketone 165 with diazomethane. This reaction was nonselective and gave a mixture of all possible regio and stereoisomers (99,166-168). The compound 167, i.e. the unnatural Z stereoisomer of garugamblin-1 (99), spontaneously isomerized to the natural product on standing in chloroform. 5.2.2.2. The synthesis of garuganin III (95) by intramolecular Wittig reaction (65) The synthesis started from the ester-aldehyde 169 obtained by Ullmann reaction of 4-fluorobenzaldehyde and methyl 3,4-dimethoxy-5-hydroxybenzoate. Following the protection of the aldehyde group by acetalization (170) the ester group was transformed to an aldehyde in two steps (170->171->172). After introduction of a C5-unit as described in Section 5.2.2.1 (173), hydrogenation (to 174) and transformation of the ester group to a phosphoniomethyl group (174-^175-^176), accompanied by deprotection of the original aldehyde group, set the stage for an intramolecular Wittig reaction, which gave in fair yield the desired cyclized product (178). Cleavage of the N-O bond of the izoxazole ring gave the enamine 179. The concluding steps of the synthesis were identical to those described above for garugamblin-1. Detailed NMR studies of the methylation products (89, 95,181 and 182) revealed that, contrary to the the literature it was the (J5)-12-keto-enolether 95 which showed an NMR spectrum identical to that of natinal garuganin III (58). Thus the correct structure of garuganin III is 95. The strategy based on intramolecular Wittig reactiuon was successfiiUy applied also to the synthesis of garugamblin -2 (91) (128). MeO

L

MeO

Ri R2 169 C02Me CHO 170 C02Me CH(OMe)2

LiAlH4 171 CH2OH CH(OMe)2 Mn02 172 CHO

CH(OMe)2 ^l^^ » 173 ^ ^2 KOtBu

390 X-X r- 173 -CH=CH-

O—N

MeO

a.

^X-'^^V^^^

Ri COjMe

R2 CH(0Me)2

|:^174 -CH2CH2- COjMe

CH(0Me)2

U - 1 7 5 -CH2CH2- CH2OH

CH(0Me)2

HC(0Me)3 LiAlH^ SOBr 176

R2

CH2CH2- CHjBr

CHO

Ph,P L^177 -CH2CH2- CHjP+Phj

CHO

MeO^ H2>Pt02 , Raney-Ni

MeO

178 ^ ^

^Z^ O

^>

NH^ (£>89

179

XX.

^OMe

180

O

O

OMe O

(2)-182

O

OMe

MeO

REFERENCES 1 Trommsdorff, Trommsdorffs Journal der Pharmacie, 16 (1808) 96; E. Vogel and S. S. Pelletier, J. Pharm. 2 (1815) 50 (quoted in Beilstein VIII, 554). 2 V. Lampe, J. Milobedzka and St. V. Konstanecki, Ber. Dtsch. Chem. Ges. 43 (1910) 2163. 3 V. Lampe and J. Milobedzka, Ber. Dtsch. Chem. Ges. 46 (1913) 2235. 4 V. Lampe, Ber. Dtsch. Chem. Ges. Dtsch. Chem. Ges. 51 (1918) 1347-1361. 5 I. L. Albuquerque, C. Galeffi, C. G. Casinovi, G. B. Marini-Bettolo, Gazz. Chim. Ital. 94 (1964) 287. 6 M. Nagai, N. Kenmochi, M. Fujita, N. Furukawa and T. Inoue, Chem. Pharm. Bull. 34 (1986) 1056-1060. 7 Y. Asakawa, F. Genjida, S. Hajashi and T. Matsuura, Tetrahedron Lett. (1969) 3235. 8 Y. Asakawa, Bull. Chem. Soc. Jpn. 43 (1970) 2223. 9 Y. Asakawa, Bull. Chem. Soc. Jpn. 43 (1970) 575. 10 N. I. Uvarova, G. I. Oshitok, A. K. Dzizenko and G. B. Elyakov, Khim. Prir. Soedin 6 (1970)463.

391 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Y. Asakawa, Bull. Chem. Soc. Jpn. 45 (1972) 1794. T. Suga, Y. Asakawa and N. Iwata, Chem. Ind. (1971) 766. T. Hashimoto, M. Tori and Y. Asakawa, Chem. Pharm. Bull. 34 (1986) 1846-1849. J. J. Karchesy, M. L. Laver, D. F. Barofsky and E. Barofsky, J. Chem. Soc. Chem. Comm. (1974)649-650. T. Suga, S. Ohta, T. Hirata and T. Aoki, Chem. Lett. (1982) 895-898. M. Lee, T. Tanaka, G. Nonaka and L Nishioka, Phytochemistry 31 (1992) 967-970 M. Nomura, T. Tokoroyama and T. Kubota, Phytochemistry 20 (1981) 1097-1104. N. Sasatami and G. Izumiya, Bull. Coll. Exp. For. Coll. Agric. Hokkaido Univ. Jpn. 31 (1974)23. M. Terasawa, T. Koga, H. Okuyama and M. Miyake, Nippon Mokuzai Gakkaishi 19 (1973)47. M. Nomura, T. Tokoroyama and T. Kubota, J. Chem. Soc. Chem. Comm. (1975) 316317. T. Sasaya, Hokkaido Daigaku Nogakubu 42 (1985) 191, Chem. Abstr. 103 (1985), 157293 . M. Terazawa, T. Koga, H. Okuyama and M. Miyake, Mokuzai Gakkaishi 30 (1984) 391, Chem. Abstr. 101 (1984) 187934b. S. Okta, M. Koyama, T. Aoki and T. Suga, Chem. Soc. Bull Jpn. 58 (1985) 2423. E. Smite, L. N. Lundgren and R. Andersson, Phytochemistry 32 (1993) 365-369. T. Inoue, T. Shinbori, M. Fujioka, K. Hashimoto and Y. Masada, Yakugaku Zasshi 98 (1978) 1255. H. Itokawa, M. Morita and S. Mihashi, Chem Pharm. Bull 29 (1981) 2383-2385. F. Kiuchi, M. Shibuya and U. Sankawa, Chem. Pharm. Bull. 30 (1982) 2279-2282. H. Itokawa, H. Morita, I. Midorikawa, R. Aiyama and M. Morita, Chem. Pharm Bull. 33(1985)4889-4893. S. Uehara, I. Yasuda, K. Akiyama, H. Morita, K. Takeya and H. Itokawa, Chem. Pharm. Bull. 35 (1987) 3298-3304. F. Kinchi, S. Iwakami, M. Shibuya, F. Hanaoka and U. Sankawa, Chem. Pharm. Bull. 40 (1992) 387. H. Itokawa, R. Aiyama and A. Ikuta, Phytochemistry 20 (1981) 769-771. H. Itokawa, R. Aiyama and A. Dkuta, Phytochemistry 21 (1982) 241-243. M. Kuroyanagi, T. Noro, S. Fukushima, R. Aiyama, A. Ikuta, H. Itokawa and M. Morita, Chem. Pharm. Bull. 31 (1983) 1544-1550. H. N. Toennesey, J. Karlsen and A. Mostad, Acta Chem. Scand. Ser. B, B36 (1982) 475. K. R. Srinavasan, J. Pharm. Pharmacol 5 (1953) 448. P. J. Roughley and D. A. Whiting, J. Chem. Soc. (1973) 2379-2388. T. Murata, M. Shinohara and M. Miyamoto, Chem. Pharm. Bull. 20 (1972) 2291. V. Ravindranath and M. N. Satyanarayana, Phytochemistry 19 (1980) 2031-2032. F. Kiuchi, Y. Goto, N. Sugimoto, N. Akao and K. Kondo, Y. Tsuda, Chem. Pharm. Bull. 41 (1993) 1640-1643. T. Masuda, J. Isobe, A. Jitoe and N. Nakatani, Phytochemistiy 31 (1992) 3645-3647. P. Claeson, A. Panthong, P. Tuchinda, V. Rentracul, D. Kanjanapothi, W. C. Taylor and T. Sautisuk, Planta Medica 59 (1993) 451. K. Endo, E. Kanno, Y. Oshima, Phytochemistry 29 (1990) 797-799. D. J. Harvey, J. Chromat. 212 (1981) 75. H. Kikuzaki, J. Usuguchi and N. Nakatani, Chem. Pharm, Bull. 39 (1991) 120-122. H. Kikuzaki, M. Kobayashi and N. Nakatani, Phytochemistiy 30 (1991) 3647-3651.

392 46 T. Masuda, A. Jitoe and N. Nakatani, Chem Lett. (1993) 189-192. 47 A. Aragdo, A. Crareiro, A. da Costa Prado, O. R. Gottlieb, P. C. Welerson and I. L. Albuquerque Phytochemistry 9 (1970) 1869. 48 L. Jurd and R. Y. Wong, Aust. J. Chem. 37 (1984) 1127. 49 L. V. Alegrio, R. Braz-Filho and O. R. Gottlieb, Phytochemistry 28 (1989) 2359-2362. 50 M. Nagai, M. Kubo, M. Fujita and T. Inoue, Chem. Pharm. Bull. 26 (1978) 28052810. 51 M. Nagai, M. Kubo, K. Takahashi, M. Fujita and T. Inoue, Chem. Pharm. Bull. 31 (1983) 1923-1928. 52 M. Kubo, T. Inoue and M. Nagay, Chem. Pharm. Bull. 28 (1980) 1300-1303. 53 M. Kubo, M. Nagai and T. Inoue, Chem. Pharm. Bull. 31 (1983) 1917-1922. 54 S. Nagumo, N. Kaji, T. Inoue and M. Nagai, Chem. Pharm. Bull. 41 (1993) 12551257. 55 Y. Saxa, T. Sasaya and S. Ozawa, Mokuzai Gakkaihi 34 (1988) 162 Chem. Abstr. 109 (1988)51715. 56 M. Yasue, J. Japan Wood Res. Soc. 11 (1965) 146, 153; 12 (1966) 226, 231. 57 M. M. Haribal, A. K. Mishra and B. K. Sabata, Tetrahedron 41 (1985) 4949-4951. 58 A. K. Mishra, M. M. Haribal and B. K. Sabata, Phytochemistry 24 (1985) 2463-2465. 59 G. Venkatraman, A. K. Mishra, P. S. Thombare and B. K. Sabata, Phytochemistry 33 (1993) 1221-1225. 60 H. Kalchhauser, H. G. Krishnamurty, A. C. Talukdar and W. Schmid, Monatsh. Chem. 119(1988) 1047-1050. 61 H. Kalchhauser and H. G. Krishnamurty, Magn. Res. Chem. 27 (1989) 635. 62 V. Pattakhi, S. Krishnaswami and E. J. Gabe, Acta Cyst. C40 (1984) 832. 63 M. Nethaji, V. Pattabhi and H. G. Krishnamurty, A. C. Talukdar, Acta Cyst C46 (1990)307. 64 S. Krishnaswami, V. Pattabhi and E. J. Gabe, Acta Cyst. C43 (1987) 527. 65 G. M. Keseru, Z. Dienes, M. Nogradi and M. Kajtar-Peredy, J. Org. Chem. 58 (1993) 6725. 66 T. Aoki, S. Okta and T. Suga, Phytochemistry 29 (1990) 3611. 67 M. J. Begley, R. V. M. Campbell, L. Crombie, B. Tuck and D. A. Whiting, J. Chem. Soc (C) (1971) 3634-3639. 68 T. Anthonsen, G. B. Lorentzen and K. E. Malterud, Acta Chem. Scand B29 (1975) 529. 69 K. E. Malterud, T. Anthonsen and J. Hjortas, Tetrahedron Lett. (1976) 3069. 70 Y. Takeda, T. Fujita, T. Shingu and C. Ogimi, Chem. Pharm. Bull. 35 (1987) 25692573. 71 Y. Yaguchi, N. Sakurai, M. Nagai and T. Inoue, Chem. Pharm. Bull. 36 (1988) 14191424. 72 N. Sakurai, Y. Yaguchi, T. Hirakawa, M. Nagai and T. Inoue Phytochemistry 30 (1991)3077-3079. 73 K. E. Malterud and T. Anthousen, Phystochemistry 19 (1980) 705. 74 R. G. Cooke and W. Segal, Aust.J. Chem. 8 (1955) 107, 413. 75 J. M. Edwards and U. Weis, Phytochemistry 9 (1970) 1653. 76 J. M. Edwards and U. Weis, Phytochemistry 13 (1974) 1597. 77 T. Cremona and J. M. Edwards, Lloydia 37 (1974) 112. 78 J. M. Edwards, Phytochemistry 13 (1974) 290. 79 R. G. Cooke and R.L. Thomas, Aust. J. Chem. 28 (1975) 1053. 80 R. Thomas, Biochem J. 78 (1961) 807.

393 81 A. C. Bazan, J. M. Edwards and U. Weiss Tetrahedron 34 (1978) 3005-3015. 82 T. A. Geissmann and D. H. G. Cront, "Organic Chemistry of SecondayPlant Products" Freeman Cooper S. F. (1969). 83 P. J. Roughley and D. A. Whiting, Tetrahedron Lett. 40 (1971) 3741-3746. 84 T. Inoue, N. Kenmochi, N. Furukawa and M. Fujita, Phytochemistry 26 (1987) 14091411. 85 F. Kinchi, S. Iwakami, M. Shibuya, F. Hanaoka and U. Sankawa, Chem Pharm. Bull 40(1992)387. 86 R. N. Chopra, S. L. Nayer and I. C. Chopra, "Glossary of Indian Medicinal Plants" C. S. I. R. O. Publications (1956) p. 172. 87 J. Garicia-Rafaneil and J. Fom, Arzneim. Forsch. 29 (1979) 630. 88 D. L. Flynn, M. F. Rafferty and A. M. Boctor, Prostaglandins, Leukotrienes Med. 22 (1986) 357. 89 H. M. Toemesen, Int. J. Pharm. 50 (1988) 670. 90 S. Toda, M. Ohniski, M. Kimura and K. Nakashima, J. Ethnopharmacology 23 (1988) 105. 91 S. Toda, T. Miyase, M. Arichi, H. Tanizawa and Y. Takino, Chem. Pharm. Bull. 33 (1985) 1725-1728. 92 T. Masuda, A. Jitoe, J. Isobe, N. Nakatani and S. Yonemori, Phytochemistry 32 (1993) 1557. 93 H. H. Toennesen, hit. J. Pharm. 51 (1989) 179. 93 T. Osawa and M. Namiki, J. Agric. Food. Chem. 33 (1185) 780. 94 R. Larson, Phytochemistry 27 (1988) 969. 95 M. E. Cuvelier, H. Richard and C. Berset, Biosci. Biotech. Biochem. 56 (1992) 324. 96 K. M. Schaich, C. Fischer and R. King, A.C.S. Symp. Ser. 547 (1994) 204. 97 Y. Kiso, Y. Suzuki, N. Watanabe, Y. Oshima and H. Hikino, Planta Medica 49 (1983) 185. 98 H. Hikino, Y. Kiso, N. Kato, Y. Hamada, T. Shioiri, R. Aiyama, H. Hokawa, F. Kiuchi and U. Sankawa, J. Ethnopharmacology 14 (1985) 31. 99 O. G. deLima, M. M. de Albuquerque and M. H. D. Maia, Rev. Inst. Antibiot. Univ. Recife 2 (1959) 19. 100 H. N. Toennesen, H. De Kies, J. Karlsen and G. Beijersbergen van Henegauwen, J. Pharm. Sci. 76(1987)371. 101 G. M. Keseru and M. Nogradi, J. Mol. Stud. THEOCHEM 286 (1993)259. 102 W. Wehrli and M. Staehelin, Bacteriol. Rev. 35 (1971) 290. 103 O. Ghisalba and P. Traxler and H. Fuhre and W. J. Richter, J. Antibiot Tokio 33 (1980)847. 104 D. W. Connel and M. D. Sutherland, Aust. J. Chem. 22 (1969) 1033. 105 D. W. Connel, Aust. J. Chem. 23 (1970) 369. 106 H. Hokawa, R. Aiyama and A. Kuta, Chem. Pharm. Bull. 31 (1983) 2491. 107 F. Kinchi, Y. Goto, N. Sugimoto, N. Akao, K. Kondo and Y. Tsuda, Chem. Pharm. Bull 41 (1993) 1640. 108 H. M. Toennsen and J. Karlsen, Int. J. Pharm. 41 (1987) 75. 109 H. N. Toennsen and J. Karlsen, Int. J. Pharm. 38 (1987) 247. 110 P. Henley-Smith, D. A. Whiting and A. F. Wood, J. C. S. Perkin 1 (1980) 614-622. 111 M. F. Semmelhack and L. S. Ryono, J.Am. Chem. Soc. 97 (1975) 3873-3875. 112 M. Sakakibara, K. Mori and M. Matsni, Agric. Biol. Chem. 36 (1972) 1825. 113 M. Miyashita, M. Hoshino and A. Yoshikoshi, Chem. Lett. (1990) 791-794. 114 N. Kato, Y. Hamada and T. Shioiri, Chem. Pharm. Bull. 32 (1984) 3323-3326.

394 115 116 117 118 119 120

H. Hokawa, R. Aiyama and A. Ikuta, Phytochemistry 21 (1982) 241. A. Arrieta, G. Mann, L. Beyer and H. Wilde, Bol. Soc. Quim. Perm. 53 (1987) 85. G. Mann, L. Beyer and A. Arrieta, Z. Chem. 27 (1987) 172. T. Pavolini, Riv. Ital. Essenze, 19 (1937) 167. H. J. J. Pabon, Rec. Trav. Chim. 83 (1964) 379. A. Sieglitz and O. Horn, Ger. Pat. (1952, Dec. 11) 859145, Chem. Abstr. 52 (1958) P12916. 121 E. Graf, Ger. Pat. 1280849, Chem. Abstr. 70 (1969) P37475 and Ger. Pat. 1282642 Chem. Abstr. 70 (1969) P47111. 122 W. Liebenow, J. Grafe, Ger. Pat. 2501220. 123 M. Mimura, Y. Takahara and T. Osawa, Jap. Pat. 9069431 Chem. Abstr. 113 (1990) P130920. 124 U. Pedersen, P. B. Rasmussen and S.-O. Lawerson, Liebigs Ann. Chem. (1985) 15571569. 125 M. F. Semmelhack, P. Helquist, L. D. Jones, L. Keller, L. Mendelson, L. S. Ryono, J. G. Smith and R. D. Staufer, J. Am. Chem. Soc 103 (1981) 6460-6471. 126 D. A. Whiting and A. F. Wood, J. Chem. Soc. Perkin 1 (1980) 623-628. 127 B. Vermes, G. M. Keserii, G. Mezey-Vandor, M. Nogradi, and G. Toth, Tetrahedron, 49(1993)4893-4900. 128 G. M. Keseru, M. Nogradi, and M. Kajtar-Peredy, Liebigs Ann. Chem.(1994) 361-364.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

395

Tropane Alkaloids in Root Cultures of Solanaceous Plants K. Shimomura, K. Yoshimatsu, K. Ishimam and M. Sauerwein

1.

INTRODUCTION Solanaceous plants, such as genus Atropa, Datura, Duboisia, Hyoscyamus and Scopolia, are regarded as rich sources of alkaloids, namely of the pharmaceutically interesting tropane derivatives [1]. The principal alkaloids of medicinal interest in this group are hyoscyamine and scopolamine (Fig. 1), which act as anticholinergic agents in parasympathetic nervous system and are used as mydriatics and spasmolytics [1]. Although the chemical synthesis of these alkaloids has been described [2], the drugs are still produced exclusively from plant materials [3]. There are considerable literatures on the production of tropane alkaloids in tissue and cell cultures derived from various parts of intact plants [4]. In a number of cases, root differentiation is required for enhanced tropane alkaloid biosynthesis [3, 5]. The production of the economically valuable tropane alkaloids, scopolamine and hyoscyamine, by these cultures has not been commercially successful, however, root cultures are so far the best system to investigate the production and biosynthesis of tropane alkaloids. In addition to the adventitious root cultures, recendy, the transformed root cultures, so called "hairy root", have been popularly used for the research. Hairy roots, developed from the plant cells genetically transformed with Agrobacterium rhizogenes, have some advantages for production of secondary metabolites, namely, good growth, stable productivity of secondary metabolites and no requirement of plant growth regulators in the culture medium [6]. In this article, we demonstrate the establishment of the adventitious and the hairy root cultures of several solanaceous plant species including Datura, Duboisia, Hyoscyamus, and Scopolia etc. for the production and biosynthetic studies of tropane alkaloids. In addition, the isolation and structural elucidation of the new tropane alkaloid 7p-hydroxyhyoscyamine and the piperidone alkaloid hyalbidone from them is also presented. 2.

PLANT MATERIALS AND ESTABLISHMENT OF CULTURES Plant materials discussed in this article are listed in Table L

2.1. Induction of adventitious (non-transformed) roots The adventitious roots were established from leaf segments (ca. 5 x 5 mm) of axenic shoot culture or intact plants (in the case of Duboisia) (surface-sterilized with 75 % EtOH for 10 sec and 2 % NaClO containing 1 drop/40 ml Tween 20 for 10 minutes) on Murashige-Skoog (MS) [7] solid medium containing 1 mg/1 lAA or 0.1 mg/1 NAA in the dark at 25 X^, except for Hyoscyamus species. The roots were subcultured every 4-8 weeks in liquid medium (50 ml/100 ml flask) containing the same phytohormone as used for their induction (0.5 mg/1 lAA or 0.1 mg/1 NAA). Root cultures of Hyoscyamus species were established from the roots of axenic plants in vitro and subcultured in phytohormone-free MS liquid medium (50 ml/l(X) ml flask).

396 Table 1. Solanaceous plants studied in this article. Cultures established

Plant species

Adventitious roots (non-transformed roots)

Datura innoxia Duboisia myoporoides-D. leichhardtii hybrid (M-II-8-6) D. myoporoides-D. leichhardtii hybrid (M-II-8-14) Hyoscyamus albus H. aureus H. muticus H. niger H. pusillus Scopolia tangutica

Hairy roots (transformed roots)

Datura innoxia Duboisia myoporoides-D. leichhardtii hybrid (M-II-8-6) Hyoscyamus albus H. niger Scopolia tangutica

2.2. Induction of hairy roots The hairy roots were induced by co-culture of the leaf segments (ca. 5x5 mm) of axenic plants or intact plants (in the case of Duboisia) in a half-strength MS (1/2 MS) liquid medium (2% sucrose) with A. rhizogenes strain A4, strain 15834 and strain MAFF 03-01724 [8] (pre-cultured in YEB liquid medium overnight) at 10" bacteria/20 ml liquid medium for 2 days at 25*^ in the dark. The bacteria were eliminated on 1/2 MS solid medium containing 0.5 g/1 Claforan®. The axenic hairy roots thus obtained were subcullured in phytohormone-free MS or Woody Plant (WP) [9] liquid medium (50 ml/100 ml flask). 2.3. Opine analysis To prove the transformation, the opines (agropine and mannopine for the hairy roots transformed with A. rhizogenes A4 and 15834, mikimopine for the ones transformed with A . rhizogenes MAFF 03-01724) were extracted, separated by paper electrophoreses and detected with silver staining (agropine and mannopine) [8, 10] or Pauly's reagent (mikimopine). In addition, half of the aqueous extract was acidified with IN HCl and the mikimopine (from neutral extract) and mikimopine lactam (from acid extract) were separated by thin layer chromatography (TLC) (cellulose plate, 80% acetone) and detected with Pauly's reagent. Mikimopine standard as well as the lactam were prepared by synthesis according to the method of Isogai et al. [8]. 2.4. Culture of adventitious and hairy roots in liquid medium Root cultures were maintained on a rotary shaker at 100 rpm in the dark or under the light at 25^. All media contained 3% sucrose and the pH was adjusted to 5.7 before autoclaving at 121T;

397 for 15 min. For the culture experiments and the time courses ca. 30 mg/flask of the roots were inoculated. 3.

PROCEDURE FOR PHYTOCHEMICAL STUDIES ON TROPANE ALKALOIDS 3.1. Extraction and Isolation The lyophilized hairy roots were extracted with CHCl3-MeOH-28 % NH4OH (15 : 5 : 1). After filtration and concentration, the extract was acidified to pH 2 with IN H2SO4 and partitioned with CHCI3. The aqueous layer was then made alkaline (pH 10) with 28 % NH4OH and partitioned with CHCI3. The organic layer was evaporated and the residue was subjected to column chromatography on Fuji-gel 0DS-G3 (H20/MeOH - MeOH/5% CH2CI2 gradient). For further purification of the fractions repetitive column chromatography on silica gel (CHCI3 / MeOH gradient) and Bondapak Cjg Porasil B (MeOH-H20 gradient) were used. 3.2. Synthesis of HPLC reference substances (littorine and 7p-hydroxyhyoscyamine) Littorine was synthesized from phenyllactic acid and tropine-HCl (Ig of each) using the same method as established for atropine [11]. The obtained O-acetyllittorine (900 mg) was dissolved in 10 ml MeOH, 9(X) mg NaHC03 was added, and the reaction mixture was refluxed in a sealed tube for 90 min. After filtration, MeOH was evaporated, the residue was dissolved in IN H2SO4 and partitioned with CHCI3. To the aqueous layer 28 % NH4OH was added (pH 10) and littorine was isolated with CHCI3. For final purification silica gel column chromatography (CHCI3 / MeOH ,4:1) was used. 7P-Hydroxyhyoscyamine was also synthesized by reduction of scopolamine [12]. A solution of (-)-scopolamine HBr (1 g) in H2O (20 ml) was treated with Raney Ni (W4) (1.5 g) at 2 0 ^ for 19 hr with stirring under H2. After removal of the catalyst by filtration, the filtrate was adjusted to pH 10 with 28 % NH4OH and extracted with CHCI3. The CHCI3 layer (dried with Na2S04), after evaporation to dryness, was applied to column chromatography over Fuji-gel ODSG3 (30%-40% MeOH) to afford hyoscyamine (290 mg) and the mixture of hydroxyhyoscyamines. The mixture was separated on column chromatography over Bondapak Cjg Porasil B [MeOH-10 mM Na 1-heptanesulfonate (pH 4 with AcOH) 3:17-1:3] to afford 6p-hydroxyhyoscyamine (150 mg) and 7P-hydroxyhyoscyamine (150 mg). 3.3. Sample preparation and HFLC analysis The cultures or other plant materials were harvested and fresh weight and dry weight, after lyophilization, were determined individually. Ca 50 mg of each sample was extracted with 5 ml CHCl3-MeOH-28 % NH4OH (15:5:1) using sonication (10 min). After filtration through filter paper the solvent was evaporated. The residue was acidified with 2 ml of IN H2SO4 and partitioned with CHCI3. The aqueous layer was alkalized with 28 % NH4OH and the alkaloid bases were extracted with CHCI3 three times. After evaporation of the solvent by N2 gas stream the alkaloid extracts were dissolved in 150-400 ^1 MeOH and 5 \i\ of which was injected into HPLC. The TSK-ODS 120A column (4.6 x 250mm) kept at 40 X! was used, and eluted isocratically with acetonitrile-lOmM SDS (pH 3.3, adjusted with 1 % H3PO4) (2 : 3). The flow rate was 1.1 ml/min throughout. The effluent was detected by UV at 215 nm.

398 3.4. Feeding experiment (Sigma) (200 mg) was added aseptically to hairy roots of Hyoscyamus albus MAFF 03-01724 cultured in a 2L air-lift type fermenter with phytohormone-free WP liquid medium for 21 days. After an additional 7 days of culture, the hairy roots were harvested. These treatments were done 6 times. From the lyophilized roots (3(X) g dry wt.) the alkaloids were isolated as described in section 3.1. After final purification of the alkaloids by column chromatography on silica gel (EtOAc-MeOH-NH40H gradient), ^H and ^^C NMR spectra were recorded at 300 and 75.5 MHz, respectively, locked to the major deuterium resonance of the solvent (CDCI3). The intensity of the enhanced signal at C3 of the tropane moiety of each alkaloid (tropinone, tropine, littorine, hyoscyamine, 6p-hydroxyhyoscyamine and scopolamine) was normalized to the signal of the methine in the side chain of the respective alkaloid itself or, in the case of tropinone and tropine, to that of a known amount of unlabelled hyoscyamine added to the sample. Due to the overlapping of the signals in some cases the diminution of deuterated C2 and C4 was not quantified. 4.

NEW ALKALOIDS FROM SOLANACEOUS PLANTS 4.1. 7p'Hydroxyhyoscy amine from Hyoscyamus albus hairy roots The new tropane alkaloid 7p-hydroxyhyoscyamine was isolated together with five tropane alkaloids hyoscyamine, norhyoscyamine, scopolamine, 6p-hydroxyhyoscyamine and tropine (Fig. 1) from the hairy roots of//, albus which have been transformed with A. rhizogenes 15834 and cultured in phytohormone-free WP liquid medium. This alkaloid was positive to Dragendorffs reagent (a reddish brown coloration) and its chemical structure was determined by the spectral and chemical data. The ^H-nuclear magnetic resonance (NMR) spectrum (Table 2) showed one methyl (6 2.44), three methylene (5 1.33, 2.03; 6 1.40, 2.11; 5 1.73, 2.31) and four methine (5 2.80, 3.17, 3.70 and 5.00) signals whose chemical shifts and coupling constants were closely correlated to those of 6p-hydroxyhyoscyamine. The ^-^C-NMR spectrum (Table 3), also similar to that of 6Phydroxyhyoscyamine, revealed the presence of a tropane diol skeleton (5 28.4, 30.1, 36.4, 40.2, 58.0, 66.6, 67.9, 75.5) and one phenylhydroxypropionic acid moiety (6 54.4, 64.0, 127.9, 128.1, 129.0, 135.5, 172.1). Furthermore, the fast atom bombardment (FAB) mass spectrum (MS) with the prominent [M+H]"*" peak at m/z 306 indicated the same molecular mass as that of 6phydroxyhyoscyamine. From these spectral data, this new o^opane alkaloid was presumed to be the structural isomer of 6p-hydroxyhyoscyamine. Acid hydrolysis of this new alkaloid with 10% HCl gave (-)-tropic acid and a hydrolysate whose spectral data, ^H-and ^^C-NMR and electron-impact (EI) MS, were completely identical with that of 3a, 6p-dihydroxytropane HCl. This evidence indicated the hydrolysate to be the enantiomer of 3a, 6 p-di hydroxy tropane HCl. Concomitant with the structural confirmation by preparing the alkaloid from (-)-scopolamine HBr using Raney Ni, this new tropane alkaloid was characterized as 7p-hydroxyhyoscyamine [12].

399 Table 2 ^H-NMR spectral data of 7p-hydroxyhyoscyamine and 6P-hydroxyhyoscyamine at 270 MHz (6 values, in CDCI3). H

7p-hydroxyhyoscyamine

6p-hydroxyhyoscyamine

1

2.80 (br s)

3.05 (m)

2

1.33(brd,J=16.1Hz) 2.03(ddd.J=16.1,4.6,4.5Hz)

1.20(brd,J=16.1Hz) 2.01 (dt,J=16.1.4.5Hz)

3

5.00 (t. J=5.4 Hz)

5.01 (I, J=5.4 Hz)

4

1.40(brd.J=16.1Hz) 2.11(dt,J=16.1,4.5Hz)

1.53(bfd,J=16.1Hz) 2.11(ddd,J=16.1,4.6,4.5Hz)

5

3.17 (m)

2.92 (br s)

6

1.73(brdd.J=13.6,7.4Hz) 2.31(dd.J=13.6.7.0Hz)

4.31(dd,J=7.0,2.4Hz)

7

3.70(dd,J=7.0,2.4Hz) 1.76 (dd,J=7.0. 13.6 Hz) 3.79 (m)

1.55(brd,J=13.6Hz) 3.78 (m)

y

3.79 (m) 4.16 (dd, 1=9.7,7.6 Hz)

3.78 (m) 4.14(dd,J=9.7,7.6Hz)

Ph

7.23-7.38 (5H, m)

7.22-7.36 (5H, m)

T

N-Me

2.44 (3H, s)

2.43 (3H, s)

Table 3 ^^C-NMR spectral data of 7p-hydroxyhyoscyamine, 6|i-hydroxyhyoscyamine, 3a, 6P-dihydroxytropane HCl and tropic acid at 67.5 MHz (5 values, ^ in CDCI3, ^ in acetone-d^ + D2O, ^ in acetone-d^. C

7P-OH-hyoscyamine^ 6p-OH-hyoscyamine^ 3a, 6p-(OH)2-tropaneHCl^

tropic acid^

1 2 3 4 5 6 7

66.6 28.4 67.9 30.1 58.0 40.2 75.5

57.9 29.8 67.9 28.5 66.7 75.8 39.9

r

172.1 54.4 64.0

172.1 54.4 64.1

174.5 55.7 65.6

127.9 128.1 129.0 135.5

127.8 128.1 129.0 135.5

128.5 129.7 129.8 138.4

36.4

36.3

T y Ph

N-Me

65.0 37.4 72.3 37.7 62.1 41.4 73.1

36.6

400 4.2. Hyalbidone from Hyoscyamus albus hairy roots The new piperidone alkaloid hyalbidone [2a, 2'p -(1,2 ethanediyl) bis (N-niethyl-3piperidone)] was isolated from the hairy roots of H, albus transformed with A. rhizogenes MAFF 03-01724. Together with the piperidone alkaloid eight tropane alkaloids were isolated and identified from their NMR data by comparison with those of authentic samples.

NCH3

NCH3

Tropinone

NCH3

-OH

\ - 0 H

H'-Tropine

Troplne

CH2OH

Littorine Norhyoscyamine

CH2OH

7p-Hydroxyhyoscyamine

Hyoscyamine

CH2OH JMCH3

CH2OH d\

—Ml O

NCH3

o

HO Scopolamine

6p-Hydroxyhyoscyamine

CH3 Hyalbidone

Fig 1. Alkaloids isolated from hairy root cultures of Hyoscyamus albus. After purification by Bondapak CI8 Porasil B column chromatography, hyalbidone was obtained as a yellow oil, which showed a strong positive reaction to Dragendorffs reagent. Neither EI-MS nor FAB-MS gave the molecular ion peak but m/z 222 (for M"^-30). The ^H-NMR

401

spectrum gave resonances for two N-Me groups, two methines and eight methylenes. The ^^CNMR spectrum showed in addition to the two methines, two N-Me and eight methylene resonances the signals for two isolated carbonyl carbons. Two dimensional-NMR experiments were carried out to verify the structure [13]. The tropane alkaloids isolated together with hyalbidone were identified as hyoscyamine, 6phydroxyhyoscyamine, 7P-hydroxyhyoscyamine, scopolamine, littorine, tropine, ^'-tropine and tropinone (Fig.l).

5.

TROPANE ALKALOIDS IN ADVENTITIOUS ROOT CULTURES

5.1. Distribution of tropane alkaloids in solanaceous species The adventitious root cultures of Datura innoxia, Duboisia hybrid M-II-8-14 (a cross-bred between D. myoporoides and D. leichhardtii), and Scopolia tangutica were established from the axenic shoot cultures or intact plants (in the case of Duboisia) on MS solid medium containing 0.1 mg/1 NAA or 1.0 mg/1 lAA. The adventitious roots were maintained in MS liquid medium containing the same phytohormones (0.1 mg/1 NAA or 0.5 mg/1 lAA) in the dark. Addition of auxin in the culture medium has been employed for the maintenance of adventitious root cultures [14], however, the adventitious roots of //. albus and H. niger were induced and maintained in hormone-free 1/2 MS medium [15]. The adventitious roots of H. albus and H. niger showed the highest level of 6P- and 7Phydroxyhyoscyamines. The amount of hyoscyamine in the adventitious roots of H. albus increased remarkably when the roots were cultured in MS or 1/2 MS medium and it became the main alkaloid. In contrast, scopolamine was the main constituent of the adventitious roots of H. albus in Gamborg B5 (B5) [16] and WP media. Thus, in Hyoscyamus species, the ratio of tropane alkaloids produced during the culture period seemed to depend mainly on the culture conditions and not on the species, as described by Hashimoto et al. [17]. In adventitious roots of//, albus the production of 7P-hydroxyhyoscyamine was at its highest level when the roots were cultured in WP medium, whereas the adventitious roots of D, innoxia and S. tangutica did not produce any detectable level of 7p-hydroxyhyoscyamine. In the Duboisia hybrid a trace amount of 7p-hydroxyhyoscyamine was detected. In the adventitious roots of //. aureus^ H. muticus and //. pusillus, traces of 7Phydroxyhyoscyamine were detected (Table 4). 5.2. Alkaloid production by adventitious root culture ofDuboisia hybrid The genus Duboisia consists of three species, D. leichhardtii F. Muell, D, myoporoides R. Br. and D. hopwoodii F. Muell. The leaves of the former two species are a major source of the tropane alkaloids, scopolamine and hyoscyamine [1]. The alkaloid content in Duboisia can be affected by the environment, and a seasonal decline was noted in the scopolamine content between May and September in Australia [18]. Callus and root cultures of Duboisia have been studied with a view to developing a new method to obtain tropane alkaloids more efficiently and the results obtained supported the view that root organ culture is so far the best system for the production of tropane alkaloids [5]. Therefore suitable culture conditions for growth and alkaloid production must be established. 4-Chloro-indole-3-acetic acid (4-Cl-IAA) is the second naturally occurring auxin to be found

402 in immature seeds of Pisum sativum [19] and 5,6-dichloro-indole-3-acetic acid (5,6-Cl2-IAA) is an auxin synthesized by Hatano and coworkers [20]. These two lAA derivatives exhibit strong hormonal activity compared to lAA and this strong activity is partly ascribed to their resistance to peroxidase oxidation [20]. In this section, we describe the effect of auxins, including 4-Cl-IAA and 5,6-Cl2-IAA, on growth and tropane alkaloid production by adventitious root cultures of a hybrid between D. myoporoides and D, leichhardtii [21].

Table 4 Distribution of tropane alkaloids in several adventitious root cultures. Alkaloids [mg/g dry weight] Plant species Datura innoxia

Culture conditions

hyoscyamine scopolamine 6p-OH-hyoscyamine 7p-OH-hyoscyamine

MS. 0.5 mg/l lAA. 6wa 3.60

Duboisia hybrid (M-II-8-14) MS. 0.5 mg/1 lAA,4w

0.06

n.d.

n.d.

3.01

1.67

0.76

trace

8.27 11.47 4.45 3.92

4.40 5.84 5.68 6.40

2.45 3.98 1.13 3.06

0.19 0.23 0.23 0.28

Hyoscyamus albus

MS. 4w 1/2MS. 4w B5.4W WP.4w

H. aureus

MS. 3w B5. 3w WP.3w

0.82 1.32 0.68

2.46 4.14 1.81

0.42 0.63 0.42

0.23 0.27 0.14

H. muticus

MS. 3w B5. 3w WP.3w

0.67 0.77 1.17

2.87 2.71 3.22

0.47 0.45 0.45

0.17 0.19 0.22

li. niger

A, 1/2MS. 6w B. 1/2MS. 6w

4.61 3.51

2.71 4.97

1.83 1.02

0.04 0.03

H. pusillus

MS. 3w B5. 3w WP. 3w

2.65 2.64 2.14

0.73 0.38 0.53

0.58 0.50 0.41

0.06 n.d. n.d.

ScopoUa tangutica

MS. 0.5 mg/1 lAA. 8w 0.11 MS. 0.1 mg/1 NAA. 8w 0.04

0.05 0.05

0.05 0.04

n.d. n.d.

^: Weeks.

Adventitious roots could be cultured in MS liquid medium containing lAA (0.1-0.5 mg/1), NAA (0.1-0.5 mg/1), 4-Cl-IAA (0.01-0.5 mg/1) or 5,6-Cl2-IAA (0.001-0.05 mg/1), but not in 2,4-D below 0.1 mg/1 (Fig. 2). On the other hand, roots formed calli in the presence of 2,4-D over 0.1 mg/1. The effects of auxins on growth and alkaloid production by adventitious root cultures of Duboisia hybrid are shown in Fig. 3 and Fig. 4. Treatment with lAA at 0.5 and 1 mg/1 was the most effective in scopolamine production. 2,4-D which promoted callus formation showed an inhibitory effect on alkaloid production (Fig. 3).

403

Fig. 2. Adventitious roots of Duboisia hybrid cultured in MS liquid medium containing 0.5 mg/1 lAA (left), 0.1 mg/14-Cl-IAA (center) or 0.01 mg/1 5,6-Cl2-IAA (right) for 4 weeks.

1106.71

1200.

1106.71

n

10004

(158.01

146.71 "9 o 15

scopolamine In roots

pi

hyoscyamlne In roots

^

scopolamine in medium

H

hyoscyamlne In medium

(108.71

60O

(155.31

fn

400

200J

11 0.1

D.

0.5 1.0 lAA (mg/1)

(114.01 (1.31

EE 0.1 0.5 NAA (nng/l)

1.0

0.01 0.1 2,4-D (mg/1)

(107.31 1.0

Fig. 3. Effects of lAA, NAA and 2,4-D on growth and alkaloid production. Roots were cultured in the dark for four weeks. The numbers in brackets represent growth index* on a dry weight basis. Bars represent standard deviation of the mean, n=3. *growthindex= final dry weight initial dry weight (ca. 1.5 mg)

4-Cl-IAA and 5,6-Cl2-IAA enhanced growth of the roots compared with lAA at low concentrations. The optimum concentrations for alkaloid production were 0.1 mg/1 for 4-Cl-IAA and 0.01 mg/1 for 5,6-Cl2-IAA. In these conditions, decrease of scopolamine and a slight increase of hyoscyamlne yield were obtained as compared with lAA.

404 700(168.71

600.

[156.0]

m

(166.71

m (208.7]

1 400. (20.71

§ 300. [200.01 (138.01 l ^ ^ ' * ^ !

0.001

Fig. 4.

0.01

0.05 0.1 0.5 4-CI-IAA (mg/l)

1.0

0.001 0.01 0.05 0.1 5,6-Ci2-IAA(mg/l)

0.5

1.0

Effects of 4-Cl-IAA (a) and 5,6-Cl2-IAA (b) on growth and alkaloid production. The explanation of the numbers in brackets and columns is the same as in Fig. 3.

The time course of alkaloid production were also examined with different concentrations of lAA (0.5 mg/l), 4-Cl-IAA (0.1 mg/l) and 5,6-Cl2-IAA (0.01 mg/l) (Fig. 4). The alkaloid yield in roots cultured with 4-Cl-IAA and 5,6-Cl2-IAA increased more rapidly than those with lAA after two weeks of culture. Scopolamine was also detected in the culture medium, especially after four to five weeks, and its yield increased significantly in the presence of 4-Cl-IAA and 5,6-Cl2-IAA as compared to lAA. A continual recovery of useful secondary metabolites from a culture medium might be important for the industrial production. These results indicate that lAA derivatives (4-CllAA and 5,6-Cl2-IAA) may be applied to the industrial production of scopolamine in cultures. 6.

ALKALOID PRODUCTION IN HAIRY ROOT CULTURES 6.1. Distribution oftropane alkaloids in solanaceous species Hairy roots of Datura innoxia, Duboisia hybrid M-II-8-6, Scopolia tangutica and the Hyoscyamus species were established by the co-culture method using their leaf discs with A . rhizogenes 15834. In addition, H. albus was transformed with A. rhizogenes strain MAFF 0301724 [15]. The hairy roots of H. albusy transformed with A. rhizogenes 15834 and cultured in WP liquid medium, produced the highest level of scopolamine and 7p-hydroxyhyoscyamine. On the other hand, in the hairy roots transformed with A. rhizogenes MAFF 03-01724, totally less alkaloids were produced overall, and hyoscyamine was the main constituent. The hairy roots of//. niger cultured in 1/2 MS medium accumulated the highest contents of hyoscyamine and 6phydroxyhyoscyamine. It is interesting that the hairy roots of D./ViAK^Jc/a and S. tangutica produced a trace amount of 7j3-hydroxyhyoscyamine, while the alkaloid could not been detected in these two adventitious root cultures (Tables 4, 5). These differences of the production between the hairy roots and the adventitious roots might be caused by the transformation with A. rhizogenes [22].

405 900

scopolamine in roots hyoscyamine in roots scopolamine in medium hyoscyamine in medium

Fig. 5. Time course of alkaloid production by root cultures of Duboisia hybrid in the presence of 0.5 mg/1 lAA (a), 0.1 mg/14-Cl-IAA (b) and 0.01 mg/l 5,6-Cl2-IAA (c).

406 Table 5 Distribution of tropane alkaloids in several hairy root cultures. Alkaloids [mg/g dry weight] Plant species

Culture conditions hyoscy amine scopolamine 6p-OH-hyoscyamine 7P-OH-hyoscyamine

Datura innoxia

MS, 4w2

1.72

0.35

0.37

trace

Duboisia M-II-8-6

1/2MS. 8w

2.14

2.47

1.23

trace

Hyoscyamus albus 15834 WP, 5w

3.38

4.55

1.66

0.14

//.fl/^itsMAFF 03-01724 WP,4w

5.38

0.42

0.68

0.10

//. niger

A (1/2MS, 5w) B (1/2MS, 5w) C (MS, 4w)

7.95 4.59 12.51

0.59 0.86 0.18

2.46 0.71 1.55

0.05 0.04 0.07

Scopolia tanguiica

A (1/2MS, 8w) B (1/2MS. 8w) C (MS, 4w)

0.41 0.52 0.46

0.13 0.18 0.19

0.14 0.14 0.03

trace trace trace

^: Weeks. 6.2. Alkaloid production in hairy root cultures of Hyoscyamus albus Hyoscyamus plants, namely henbane, have been known to man from ancient times as a remedy for various diseases. One of the allied species of Hyoscyamus niger (described as Hyoscyamus leaf in the European Pharmacopoeia), Hyoscyamus albus is grown on the European Continent, particularly in France, and in the Indian subcontinent [1]. Quantitatively and qualitatively its alkaloids appear similar to those of//, niger [1]. The hairy root cultures of Hyoscyamus species have been studied for alkaloid production by several researchers [3, 23, 24]. However, the influence of phytohormones and light on the production of tropane alkaloids in transformed roots of this genus has not been reported in detail. In this section, we describe the alkaloid production by //. albus hairy roots cultured under different conditions. 6.2.1. Alkaloid pattern in hairy roots transformed with different strains of A. rhizogenes Leaf-disks of in vitro cultures of //. albus were transformed with A. rhizogenes strain 15834, strain A4 and strain MAFF 03-01724 as described above [13, 15, 25]. The hairy roots and the adventitious (non-transformed) roots were subcultured on phytohormone-free WP solid medium with 3% sucrose in the dark or in the light. After 30 days of culture the content of 7p-hydroxyhyoscyamine, 6P-hydroxyhyoscyamine, scopolamine, hyoscyamine and littorine in the roots was determined by HPLC [25]. The alkaloid content of hairy roots transformed with A. rhizogenes 15834 and of adventitious roots cultured under light condition was much higher than that cultured in the dark (Fig.6). In contrast the alkaloid production in the transformants with A. rhizogenes strains A4 and MAFF 03-01724 was slightly increased only under light condition. Especially the proportion of the oxygenated tropane derivative scopolamine was enhanced in all light grown root cultures. Its maximal amount was detected in the adventitious root cultures. In addition, the content of 7P-hydroxyhyoscyamine and 6p-

407 hydroxyhyoscyaminc, which have been presumed to be the biogenetic intermediates from hyoscyamine to scopolamine, was elevated to some degree under light condition. However the content of hyoscyamine slighriy increased and littorine showed a tendency to decrease in the light. Thus the oxygenating step in the biosynthesis of the tropine moiety seems to be enhanced by the light, but the light is not stricdy required. In consequence the contents of scopolamine in the adventitious roots cultured under the light were up to 4-fold compared to the dark grown roots. On the other hand the main alkaloid in all cultures was hyoscyamine. Its maximal content was detected in the hairy roots transformed with A, rhizogenes A4 cultured in the dark as well as in the normal roots cultured in the light (Fig. 6). The variation of the alkaloid content in those hairy roots was presumed to be caused by the insertion of the different Ri plasmids [26,27].

Adventitious roots

Fig. 6. Alkaloid production in the hairy roots and the adventitious roots off/, albus after 4 weeks in WP liquid medium. Bars indicate the standard error. We compared the time course of tropane alkaloid production in the adventitious and two types of hairy roots of H. albus, transformed with the different strains of i4. rhizogenes [15]. In the adventitious roots of H. albus, hyoscyamine was the main alkaloid during the first 6 weeks of culture in phytohormone-free MS liquid medium. At the seventh week of culture, with the beginning of the stationary growth, scopolamine became the major constituent (Fig. 7). In the hairy roots, transformed with A. rhizogenes 15834, hyoscyamine was the main alkaloid only for the first two weeks of culture in phytohormone-free WP liquid medium. After that time scopolamine became the major alkaloid. The production of scopolamine increased rapidly and it became almost double the amount of hyoscyamine at the end of the culture period. The yield of 6P-hydroxyhyoscyamine and 7P-hydroxyhyoscyamine produced in these hairy roots was comparable to that in the adventitious roots (Fig.7). In contrast, in the hairy roots transformed with A. rhizogenes MAFF 03-01724 hyoscyamine was the main product during the entire culture period and over 5 mg/lOOml flask of hyoscyamine

408 were obtained after 22 days of culture. This was approximately double the amount compared to the yield in the adventitious roots after 42 days of culture and 3-fold the amount compared to that in the hairy roots transformed with A. rhizogenes.

15834 (Fig 7).

Adventitious roots MS medium

¥5 E,4

J -^ h)^oscyamine 1 - ^ ^ S(;opolamine J

- d t - 6 3-OH-hyoscyamine

1

- i U . 7 3-OH- hyoscya mine

re o < 2H

10

15

20 25

30

,^. ^\

35 40

15 20

45 50

25

30

Fig. 7. Yield of alkaloids in the adventitious roots (left) and hairy roots (center and right) of H. albus cultured in phytohormone-free MS liquid medium (adventitious roots) or in phytohormone-free WP liquid medium (hairy roots).

6.2.2. Influence of phytohormones on hairy root cultures The morphology of the adventitious and hairy roots changed to callus-like structures with the addition of phytohormones. Calli with roots were obtained when the roots were cultured on the medium with the addition of lAA. The growth of the adventitious root cultures was enhanced when the medium contained lAA (2 or 4 mg/1) in combination with 1 mg/1 kinetin. transformed with A. rhizogenes

Similarly the growth rate of the hairy roots

A4 cultured in the light was enhanced with the addition of low

concentrations of phytohormones, while only friable calli were obtained when 2 or 4 mg/1 of lAA were added to the medium. rhizogenes

Growth and morphology of the hairy roots transformed with A .

15834 was similar to that of the adventitious root cultures. Here again the maximal

growth rate was observed in the medium containing both lAA and kinetin at a ratio of 1:1. On the other hand the growth of the hairy roots transformed with A. rhizogenes

MAFF 03-01724 was

reduced with the addition of any combination of phytohormones. The influence of phytohormones and light on the production of tropane alkaloids in transformed roots of Hyoscyamus albus was investigated [25]. In adventitious roots cultured in the dark, the addition of lAA (up to 4 mg/1) to the culture medium only slightly affected alkaloid production (Fig. 8a). On the other hand, when the adventitious roots were cultured under light, the alkaloid content decreased with the addition of lAA. Any combination of lAA with kinetin rapidly

409 reduced the alkaloid production in the adventitious roots cultured in the dark as well as in those cultured in the light. The addition of a single phytohormone or their combination to the hairy root cultures transformed with A. rhizogenes A4 reduced the alkaloid formation substantially (Fig. 8b). The inhibitory effect of phytohormones on alkaloid production was larger when these hairy roots were cultured in the light. In contrast to the transformants with strain A4, in the hairy roots transformed with strain 15834, the addition of lAA enhanced the alkaloid production when the roots were cultured in the dark (Fig. 8c). Maximal alkaloid content in the hairy roots was obtained with 2 mg /1 lAA. On the other hand, alkaloid production was slightly depressed in light grown cultures. Kinetin reduced alkaloid production in the hairy roots. The addition of phytohormones to the hairy root cultures transformed with A. rhizogenes MAFF 03-01724 had a somewhat different results (Fig. 8d). Kinetin, which inhibited the alkaloid production in all other cultures, in this case enhanced the alkaloid formation. Especially in the dark, the addition of 1 mg /1 kinetin or 1 mg /1 lAA or their combination resulted in higher productivity. The highest content of littorine was obtained under these culture conditions. Under light conditions, the phytohormones did not strongly affect the alkaloid production, except for the combination of lAA with kinetin (2+1 mg/ 1). Here the content of hyoscyamine was almost doubled. Other combinations of phytohormones resulted in suppressed scopolamine production. The biosynthesis of tropane alkaloids is known to be located in the roots of the plants. From there hyoscyamine is transported via the xylem to the aerial parts of the plants, where the transformation into scopolamine takes place [28, 29]. In this context it is remarkable that in adventitious roots and hairy roots of H. albus cultured in the light the alkaloid content and especially the formation of scopolamine and its intermediates were increased compared to the cultures grown in the dark. On the other hand, in the dark larger amounts of littorine were produced. Littorine is supposed to be the pool of tropine in the biosynthetic route of the tropane alkaloids [30, 31; See section 6.2.3.]. Therefore its increased content in the dark grown roots may be attributed to its lower utilization as a precursor for the biosynthesis of the oxygenated tropanes occurring in the light. The effect of phytohormones on tropane alkaloid production varied according to the strain of A. rhizogenes used for transformation. This might be caused by the different insertion of the Riplasmid or clone [26, 27]. The different contents of endogenous phytohormones in the hairy roots, originated to the genes coding for phytohormone biosynthesis in the Ri-plasmid, may influence the response of the hairy roots to the exogenous phytohormones in the medium. 6.2.3. AIkaloid production by H. albus hairy roots transformed MAFF 03-01724 As the hairy roots of//, albus transformed with A. rhizogenes strain MAFF 03-01724 were growing very rapidly they were used for further investigation [31]. The hairy root cultures were tested for growth and alkaloid production in various media. The axenic hairy roots were cultured as usual in phytohormone-free WP liquid medium with 5% sucrose in the dark.

410

a: Adventitious roots

? ^ o) 4. "

3.

m

7fJ-OH-hyoscyannine

J2I

hyoscyamine

B

6B-OH-hyoscyamine

Q

littorine

n

scopolamine

1

c: Hairy roots 15834 e-5 ^3 6 2H

^^

fi

J em,

0/1

0/4

1/0

QJ?L

rfL 1/2

—fu

;tn, H h

FhH .

FTI

0/0

Phytohormones: kinetin/IAA [mg/l]

Fig 8. Influence of phytohormones on alkaloid production by adventitious and hairy roots of Hyoscyamus alhus cultured in WP solid medium. Phytohormones were added either alone or in combination. Bars indicate standard error.

411 The alkaloid content in the hairy roots cultured for 19 days in four different media, containing 3% and 5% sucrose each, was examined (Fig. 9). The media tested were: 1/2 MS, MS, B5 and WP liquid media. The hairy roots cultured in the media with low sucrose concentration showed generally higher contents of alkaloids. On the other hand, the growth was better in the media with the higher sucrose concentration. Hyoscyamine was the main alkaloid in any case, but the ratio of hyoscyamine to the other alkaloids was highly dependent on the culture media. The highest content of 7p-hydroxyhyoscyamine was observed in B5 media, while in MS media higher yields of 6P-hydroxyhyoscyamine and scopolamine were obtained. In WP medium the hairy roots showed the fastest growth, but the alkaloid production was relatively low, except for the comparably high yield of littorine.

H

7B-OH-hyoscyamine

0

hyoscyamine

g

6Q-OH-hyoscyamine

Q littorine

Q scopolamine

Fig. 9. Alkaloid production in H. albus hairy roots MAFF 03-01724 cultured in different liquid media for 19 days. Numbers in brackets show the fresh weight. Bars indicate standard error.

To increase the growth of hairy roots, the influence of the sucrose concentration in the WP liquid medium was tested (Fig. 10). The fastest growth was obtained with 8% sucrose (8.91 g fresh weight after 19 days). A similar effect was observed to that described in the previous experiment, i. e. the total alkaloid content decreased with the higher sucrose concentration in the medium, whereas the amount of 7P-hydroxyhyoscyamine increased up to 0.012% dry weight in the media with 8% and 10% sucrose. The content of littorine produced by the hairy roots particularly decreased with higher sucrose concentrations in the medium.

412

Sucrose concentration [%]

Fig. 10. Alkaloid production in H. albus hairy roots MAFF 03-01724 cultured in WP liquid media with different sucrose concentrations for 19 days. The explanation of the numbers in brackets and columns is the same as in Fig. 9.

It has been reported that low concentrations of nitrate (10 to 20 mM) in the culture medium could stimulate the synthesis of tropane alkaloids in Datura stramonium hairy roots [32]. Therefore, we added 1 to 50 mM KNO3 to WP (3% sucrose) liquid medium, originally containing 9.7 mM KNO3. The addition of KNO3 ^^ ^^^ culture medium enhanced the growth of the //, albus hairy roots remarkably (Fig. 11). In WP (3%) liquid medium with additional 15 mM KNO3 up to 13.7 g fresh weight was obtained from only two individual root tips (ca 10 mg fresh weight) after 19 days of culture. The addition of higher concentrations of KNO3 (20 - 50 mM) also stimulated the growth of the hairy roots, but to a lesser extent. On the other hand, the addition of 1 - 25 mM KNO3 ^^ ^^^ culture medium did not strongly affect the alkaloid content in the hairy roots. 7pHydroxyhyoscyamine was produced in a larger amount with increasing concentrations of KNO3 ^" the medium. In contrast, the contents of scopolamine decreased from 0.031 to 0.019% dry weight. In further experiments we compared the time course of the alkaloid production in the hairy roots cultured in four different media [B5 (3% sucrose), WP (5% sucrose), WP (3% sucrose) and WP containing 3% sucrose and 15 mM KNO3 (WPN03)] (Fig. 12). As mentioned above, the fastest growth was observed in the WPN03 liquid medium. These cultures reached the logarithmic growth stage later (after 12 days) compared to the hairy roots cultured in the other three media (logphase after 8 days). On the other hand, the hairy roots cultured in B5, WP (3% sucrose) and WP (5% sucrose) media turned blueish-grey and died after 25 - 28 days of culture, while with the addition of 15 mM KNO3 the hairy roots could be cultured for over 6 weeks.

413

Fig. 11. Alkaloid production in H. albus hairy roots cultured in WP liquid media with 3 % sucrose and additional potassium nitrate (KNO3) for 19 days. The explanation of the numbers in brackets and columns is the same as in Fig. 9.

1412.

• 0

B5,3% sucrose

%

^^^^"""^

WP, 3% sucrose

y/^

—^5t— WP, 5% sucrose

g10-

-jie-

y^

WPN03,3% sucrose

. / ' ^ > v

J l ^

^ S ^

o) 8-

"^ 4. 201 1

^

,

1^

10

15 Days

1^

20

,

1

25

30

Fig. 12. Growth of //. albm hairy roots MAFF 03-01724 in different liquid media. In the hairy roots cultured in WPN03 medium, over 30 mg of tropane alkaloids was produced in one 100 ml flask (50 ml medium) within 28 days of culture (Fig. 13). This was 5 to 6fold largest amount to that produced in the hahy roots cultured in the other media. Hyoscyamine was the main alkaloid produced by the hairy roots during the entire culture period in all media tested. In the hairy roots cultured in B5 medium the highest content of scopolamine was obtained, while the content of 7p-hydroxyhyoscyamine and littorine was comparably low during the entire culture period. On the other hand littorine was produced in larger amounts in all WP media. Until 18 days

414 of culture the content of littorine increased in the cultures. After 18 days the amount of littorine declined, especially in the WPN03 liquid medium. Concomitant to the decrease of littorine the yield of the other alkaloids increased. From these phenomena littorine might be an intermediate or storage form of tropine for the biosynthesis of other tropane alkaloids.

20 18 ||c: WP, 3% sucrose |

id: WPN03. 3% sucrose |

; 16 ^14 -12 •10 : 8 !

6

1

;

4 2 Oi

10

Fig. 13. Time course of alkaloid production in H. albus hairy roots MAFF 03-01724 in different liquid media (50 ml/100 ml flask, inoculum weight: 10 mg).

7.

ELUCIDATION OF BIOSYNTHETIC ROUTE OF TROPANE ALKALOIDS USING HYOSCYAMUS ALBUS HAIRY ROOTS During the last few years, the biosynthesis of tropane alkaloids has been intensively studied [28]. The origins of both the acidic and basic moieties have been investigated and some steps in the biosynthesis were examined at the enzymatic level [33]. On the other hand, the last steps of tropic acid biosynthesis, i.e. those of the acidic moiety of the medicinal important alkaloids scopolamine and hyoscyamine, are still not completely understood [34, 35]. The mechanism of esterification of these acids with tropine is still unknown. On the other hand, the conversion of the amino acid ornithine into the N-methylpyrrolidium ion has been confirmed for many tropane alkaloids [29 and references cited therein]. The origin of the remaining carbons of the tropane moiety has been attributed to acetoacetate, malonate or two units of acetate [28, 36, 37]. The fast-growing and highly-productive hairy root cultures of H. albus transformed with A. rhizogenes MAFF 03-01724 (Fig. 14) were used to investigate the incorporation of C^H3^^COONa into several tropane alkaloids such as tropinone, tropine, littorine, hyoscyamine, 6[3-hydroxyhyoscyamine and scopolamine (for formulae see Fig. 1) [38].

415

Fig. 14. Hyoscyamus albus hairy roots transformed with Agrobacteriwn rhizogenes MAFF 0301724. The roots were cultured in a 21 air-lift type fermenter with phytohormone-free WP liquid medium for 28 days.

C%3^^COONa (100 mg/1) was added to the cultures at day 21, when tropane alkaloid formation became most active. The roots were harvested after another 7 days of culture. Alkaloids were extracted, isolated by column chromatography and analyzed by NMR. The labelled acetate was incorporated into the tropane moiety of all tropane alkaloids as tropinone, tropine, littorine, hyoscyamine, 6P-hydroxyhyoscyamine and scopolamine, whereas only little labeling was observed in hyalbidone. The ^^C-label appeared only at the C-3 position of tropane, as was expected. According to their integration by -NMR the incorporation of 1-acetate into the tropane moiety was at a different level for each alkaloid (example given for hyoscyamine in Fig 14a). The highest rate of ^^C-labeling was found in tropinone, followed by tropine, littorine, hyoscyamine, 6P-hydroxyhyoscyamine and scopolamine (Table 6).

416 Table 6. Incorporation of 1-^-^C-acetalc into iropane alkaloids as analyzed by ^-^C-NMR alkaloid

amount [mg]

C-3 shift [ppm] normalized on [area = 100%] C-3 [% peak area]

1

tropinone

22

178.1

C2' of 4 added

921 %

2

tropinc

11

56.9

C2' of 4 akkxl

874%

3

littorine

550

69.1

C2'

419%

4

hyoscyamine

930

68.3

C2'

306%

5

6P-hydroxyhyoscyaminc

160

68.0

C2'

252%

6

scopolamine

75

66.8

C2'

177%

The differences in the incorporation of 1-^^C-acetate into the tropane moiety might reflect their position in the biosynthetic sequence: Earlier alkaloids show a higher incorporation than those at later steps. Because littorine shows a higher incorporation of 1- ^C-acetate it may be synthesized earlier than hyoscyamine. Also circumstantial evidence supports this view. In the time course of alkaloid production by hairy roots (see Fig. 15c) the amount of littorine decreased after 18 days, whereas the amount of the alkaloids hyoscyamine, 6P-hydroxyhyoscyamine and scopolamine increased. From these observations, we assume that littorine might be biosynthetic intermediate of hyoscyamine, 6p-hydroxyhyoscyamine and scopolamine (Fig. 16). Although degradation, turnover or different pool sizes may influence the incorporation pattern significantly, it has been reported that radioactive hyoscyamine was obtained by feeding dual labeled littorine, 3a (2hydroxy-3-phenylpropionyloxy-[l- ^C]-tropane-[3p- H]), in Datiira stramonium plants [31]. This supports our hypothesis. However, the drastic change in the -^H:'^C ratio between feeding littorine and hyoscyamine (relatively higher radioactivity was found in the tropine moiety of hyoscyamine) was found in the same experiment and it may indicate that littorine is not converted directly to hyoscyamine. For understanding the biosynthesis of the hyoscyamine involved intermediate, littorine, further investigations are required. In hyalbidone, a little labeling was observed, indicating that this alkaloid is either not related to the tropane alkaloids or is synthesized at later growth stage when the ^^C-acetate was already consumed. Concomitant to the enhancement of the signal for C-3 in the ^-^C NMR spectrum, the signals for H2 and H4 decreased in the *H-NMR specdaim of all the tropine moieties compared with those in authentic samples. This indicates the incorporation of deuterium at C2 and C4 derived from As the diminution of the signals for H2 and H4 was not identical at both positions, we suppose a sequential incorporation of labeled acetate (example given in Fig. 15b). From these data we assume that the biosynthesis from N-methyl-pyrrolidinium ion to tropine is a two-step process which does not involve a four carbon unit (acetoacetyl coenzyme A) but two units of acetyl coenzyme A, which were added sequentially as has been suggested for the biosynthesis of cocaine [28, 37]. On the other hand it has been reported that l,2-^^C2-acetate was incorporated with an equal efficiency at C2 and C4 by non-transformed root cultures of H. albus [36]. Further investigations are required to clarify this matter.

417

(a) 6 2 CHjOH

Hyoscyamioe

T*"»" .-yM|, •• t , . - ^ jiMiAt,»s^ •M*. *<-.Mh'« '^^->v,.<^'\"i. >t/v»^tMV

^**'>VVMVWWV.»VAVT*M

2.4 3

6.7

1.5

NMe

'^•rV'^fc>^-v^t>"'n.<^i'**-»-i«y'

180

170

160

150

140

130

',*vis*«yyA'*T'Wvivf

120

110

100

90

80

70

60

50

40

30

ppm

(b) (I)

y yL JtAj) / 7.5

7.0

6.5

6.0

55

5.0

4.5

4.0

3.5

3.0

2.5

2.0

15

ppm

2.3

2.2

Fig. 15 a), b)

2.1

2.0

1.9

2.3

2.2

2.1

2.0

1.9

Incorporation of C%3^^0C)0' in hyoscyamine by hairy root cultures of//, albus. Numbers indicate the signals of the tropane moiety, a) ^^C NMR; I) after feeding; II) control b) ^H NMR; I) after feeding; II) control, enlargement of H2a, H4a; III) after feeding, enlargement of H2a, H4a

418 COOH H2l

2CH^^COONa

M-CH3

(^H2 Ornithine

iV^

^^1 NCH3

P^OH

NCH3

^Gro

Tropinone

'^/'H

Tropine

•v^'" NCH3

Cl^OH

^3c-0

Littorine

CH2OH

Fig. 16.

Possible biosynthetic pathway of tropane alkaloids in Hyoscyamus albus hairy roots transformed withi4. rhizogenes MAFF 03-01724.

REFERENCES 1 W. C. Evans, Pharmacognosy, 13ih Ed. Bailliere Tindall, London, 1989. 2 H. Ida, Y. Watanabe, C. Kobayashi, J. Org. Chem., 50 (1985) 1818-1825. 3 A. Strauss, in Y. P. S. Bajaj (Ed.), Biotechnology in agriculture and forestry 7, Medicinal and aromatic plants II. Hyoscyamus spp.: In vitro culture, and the production of tropane alkaloids. Springer-Verlag, Berlin Heidelberg, 1989, pp. 286-314. 4 G. Petri and Y. P. S. Bajaj, in Y. P. S. Bajaj (Ed.), Biotechnology in agriculture and forestry 7, Medicinal and aromatic plants II. Datura spp.: In vitro regeneration, and the production of tropanes. Springer-Verlag, Berlin Heidelberg, 1989, pp. 135-161. 5 Y. Kitamura, in Y. P. S. Bajaj (Ed.), Biotechnology in agriculture and forestry 4, Medicinal and aromatic plants I. Duboisia spp.: In vitro regeneration, and the production of tropane and pyridine alkaloids. Springer-Verlag, Berlin Heidelberg, 1988, pp. 419-436. 6 S. Sauerwein, K. Yoshimatsu and K. Shimomura, Plant Tissue Culture Letters, 9 (1992) 1-9.

419 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

T. Murashigc and F. Skoog, Physiol. Plant., 15 (1962) 473-497. A. Isogai, N. Fukuchi, M. Hayashi,.H. Kamada, H. Harada and A. Suzuki, Phytochemistry, 29(1990)3131-3134. G.B. Lloyd and B.H. McCown, Inter. Plant Prop. Soc., 30 (1980) 421-427. A. Petit, C. David, G. A. Dahl, J. G, Ellis, P. Guyon, F. Casse-Delbart and J. Temp^, Mol. Gen. Genet., 190 (1983) 204-214. G. C. Schmidt, T. E. Eling and J. C. Drach, J. of Pharm. Sci., 56 (1967) 215-222. K. Ishimaru and K. Shimomura, Phytochemistry, 28 (1989) 3507-3509. M. Sauerwein, K. Ishimaru and K. Shimomura, Phytochemistry, 30 (1991) 2977-2978. G. Payne, V. Bringi, C. Prince and M. Shuler, in Plant Cell and Tissue Culture in Liquid System : Root Cultures, Hanser Publishers, Munich Vienna New York Barcelona, 1992, pp 225-279. K. Shimomura, M. Sauerwein and K. Ishimaru, Phytochemistry, 30 (1991) 2275-2278. O. L. Gamborg, R. A. Miller and K.Ojima, Exp. Cell. Res., 50 (1968) 151-158. T. Hashimoto,.Y. Yukimune and Y. Yamada, J. Plant Physiol, 124 (1986) 61-75. O. Luanratana, W. J. Griffin, J. Nat. Prod., 43 (1980) 552-558. S. Marumo, H. Hattori, H. Abe, Nature, 219 (1968) 959-960. T. Hatano, M. Katayama, S. Marumo, Experientia, 43 (1987) 1237-1239. K. Yoshimatsu, T. Hatano, M. Katayama, S. Marumo, H. Kamada and K. Shimomura, Phytochemistry, 29 (1990) 3525-3528. A. J. Parr and J. D. Hamill, Phytochemistry, 26 (1987) 3241-3245. R. J. Robins, A.J. Parr and N. J. Walton, Planta, 183 (1991) 196-202. A. J. Parr, J. Payne, J. Eagles, B. T. Chapman, R. J. Robins and M.J.C. Rhodes, Phytochemistry, 29 (1990) 2545-2550. M. Sauerwein, K. Shimomura and M. Wink, J. Plant Physiol., 140 (1992) 147-152. Y. Mano,.S. Nabeshima, C. Matsui and H. Ohkawa, Agric. Biol. Chem., 50 (1986) 27152722. Y. Mano, H. Ohkawa and Y. Yamada, Plant Sci., 59 (1989) 191-201. E. Leete, Planta Med., 56 (1990) 339-352. H. W.Liebisch and H. R. Schutte, in: K. Mothes, H. R. Schutte and M. Luckner (Eds.), Biochemistry of alkaloids, Deutscher Verlag der Wissenschaften, Berlin, 1985, pp. 106-136. E. Leete, E. P. Kirven, Phytochemistry, 13 (1974) 1501-1504. M. Sauerwein and K. Shimomura, Phytochemistry, 30 (1991) 3277-3280. J. Payne, J. D. Hamill, R. J. Robins and M. J. C. Rhodes, Planta Med., 53 (1987) 474-478. Y. Yamada, T. Hashimoto, T. Endo, Y. Yukimune, J. Kohno, N. Hamaguchi and B. Drager, in: B. V. Charlwood and M. J. C. Rhodes (Eds.), Secondary Products from Plant Tissue Culture, Clarendon Press, Oxford, 1990 pp. 227-254. W. C. Evans and J. G. Woolley, Phytochemistry, 15 (1976) 287-289. P. Bachmann, Y. Yamada and R. Robins, Planta med., 57 suppl.(1991) 9. U. Sankawa, H. Noguchi, T. Hashimoto and Y. Yamada, Chem. Pharm. Bull., 38 (1990) 2066-2069. E. Leete, Heterocycles, 28 (1989) 481 -483. M. Sauerwein, K. Shimomura and M. Wink, Phytochemistry, 32 (1993) 905-909.

This Page Intentionally Left Blank

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

421

Phenolics in Root Cultures of Medicinal Plants K. Ishimaru and K. Shimomura

1.

INTRODUCTION From ancient times, numerous plants have served as sources of natural products used for foods, pigments, flavors and drugs (occasionally as toxicants). Particularly, the plants producing secondary metabolites which have some pharmacological interaction with animals (including human beings), have been carefully delivered and used as "medicinal plants". The major classes of useful compounds produced in the medicinal plants are, for example, alkaloids, saponins, lipids, phenolics and so on. With the recent development of analytical methods, the chemical structures of phenolics, especially polyphenols such as tannins, whose chemistry has remained unsolved for a long time, have gradually been elucidated (1) and tannins have been proved to possess several new important medicinal activities (2-8). Particularly, the importance of tannins as virustatic pharmaceuticals is continuously increasing (9-11). Many approaches for the production of useful secondary metabolites by planttissuecultures (cell and organ cultures) have been developed and have led, in several plants, to the production of the objective chemicals in large amounts (with higher contents compared to those of the intact plants). Most of these were alkaloids and terpenoids (12). Concerning the production of polyphenols by tissue culture methods, sufficient research has not been done except for some examples using cell suspension cultures (13-25). In most of the plant species, undifferentiated cell cultures, in comparison with organ cultures, have tended to be unsatisfactory in producing the desired secondary metabolites in high amount. Recently, in many laboratories, the cultures of hairy roots, genetically transformed with Agrobacterium rhizogenes, have been employed for the production of useful compounds in medicinal plants because of their good and rapid growth as well as stable and high productivity of secondary metabolites in hormone-free culture conditions. In this article, some examples of recent work in the production of medicinally important polyphenols in some medicinal plants. Sanguisorha officinalis (Rosaceae) (26, 27), Geranium thunbergii (Geraniaceae) {2%), Swertia japonica (Gentianaceae) (29, 30), Rheum palmatiim (Polygonaceae) (31) and Phyllanthus niruri (Euphorbiaceae) (32) using the hairy root cultures in addition to the adventitious root (not transformed root) cultures are described. 2.

PHENOLIC PRODUCTION IN ROOT CULTURES OF ADVENTITIOUS AND HAIRY ROOT CULTURES OF SANGUISORBA OFFICINALIS L. Sanguisorha officinalis L. (Rosaceae) is one of the traditional plants whose air-dried roots (Japanese name 'ziyu') are used medicinally as hemostatic, antiphlogistic and astringent. The effective constituents of the plant have been shown to be high molecular weight (Mr) polyphenols of

422 tannins (hydrolyzable one) such as sanguiins (33-36). 2.1 Phenolics in adventitious root cultures 2.1.1 Establishment of adventitious root cultures The seeds collected from S. officinalis L. plants were surface sterilized in 2% NaClO with Tween 20, rinsed with sterilized water, and germinated aseptically on 0.5% agar medium containing 0.5% sucrose in light (3000 lux). The roots of the axenic plants thus obtained were cut off and transferred into Murashige-Skoog (MS) liquid medium (37) (50 ml per 100 ml Erlenmeyer flask) containing 1 mg/1 indole-3-acetic acid (lAA). These adventitious roots were maintained by subculturing at 2-month intervals in the dark at 25°C on a rotary shaker at 100 rpm. 2.1.2 Isolation ofphenolics from adventitious root cultures The adventitious roots cultured for 4 weeks under the above conditions were lyophilized and extracted with 80% aqueous acetone. The extract, after concentration and filtration of the precipitation, was partitioned with EtOAc. The EtOAc layer, after evaporation of the solvent, was subjected to column chromatography with Sephadex LH-20 (elution with EtOH, followed by waterMeOH) to give seven phenolics, gallic acid (1), (-H)-catechin (2), (+)-gallocatechin (3) (38), procyanidin B-3 (4) (39), l,2,3,6-tetra-(9-galloyl-P-D-glucose (5) (40), l,2,3,4,6-penta-<9-galloylP-D-glucose (6) (40) and pedunculagin (7) (36). The aqueous layer was also purified by a combination of MCI-gel CHP-20P (water-MeOH), Sephadex LH-20 (80% EtOH) and Avicel Cellulose (2% AcOH) column chromatographies to afford P-glucogallin (8) (41), 4,6hexahydroxydiphenoyl (HHDP)-D-glucose (9) (36), sanguiin H-6 (10) (36) and sanguiin H-11 (11) (36). The chemical structures of 1-11 are shown in Fig. 1. Among these compounds, 11 is one of the highest Mr phenolics (Mr. 3738) whose chemical structures have been clearly elucidated so far. 2.1.3 Production ofphenolics in adventitious root cultures The production of 1, 2, 5, 6, 10 and 11 in the adventitious root culture (1-8 weeks) in MS liquid medium containing 1 mg/1 lAA was determined by high-performance liquid chromatography (HPLC) (Fig. 2). HPLC conditions were as follows; column: Nucleosil 100-5Cjo (4.6 mm i. d. x 250 mm), mobile phase: MeCN-50 mM H3PO4 (1:19-1:3), flow rate: 0.8 ml/min, column temperature: 40*'C, detect: UV (280 nm),Rt(min): 1 (6.52), 2 (12.46), 10(19.93), 11 (21.29), 5 (22.54) and 6 (27.04). In this culture, the roots grew well for 6 weeks of culture and after week 4 they became light brown, presumably due to the complexation of the phenolics in the roots and Fe contained in the culture medium. The production of 10 increased rapidly from the beginning of the culture and after it reached a maximum level (7.3 mg / flask at week 4) it began to decrease gradually. On the other hand, the amount of 11 continued to increase until the end of the culture time. This observation might support the presumption of the biosynthetic conversion from 10 into 11 (11 was presumed to be synthesized by the oxidative condensation of two molecules of 10 through C-O coupling) (36). The production of 5 and 6 increased remarkably after week 4, while the levels of 1 and 2 were almost on a plateau (1: ca 0.1 mg / flask, 2: ca 0.5 mg / flask) in this culture.

423

•p-

- ^ ,:. &

"~9Ci. HO

1 (gallic acid)

2 (catechin)

3 (gaiiocatechin) HO

OH

CO

••OCH,

"°^Si::^oco-Q--o

-•pa. •

>rt'''°

\°

OH

HO

4 (procyanidin B-3) 5 (1,2,3,6-tetra- O-galloyi-P-Dgiucose)

•—v

CO^

^OCH,

V-x

J = ^

^OH

^ o W V o c o ^ - 0 ^^°

Xo

OH HO

HOCH,

OH

OH

7 (pedunculagin)

6 (1,2,3,4,6-penta- O-galloyi-p. D-glucose)

^OH

Ho\.--vA-*oco-/

HO

•CO..OCH,

po^o-^^o.

V-o

8 (p-glucogaliin) 9 (4,6-HHDP-D-glucose)

10 (sangulin H-6) „^

^

HO

HO OH

O

^^_y

OH

OC CO

HO

HO OH

^ / ^

OH

11 (sanguiinH-11)

Fig. 1 Phenolics in root cultures of Sanguisorba officinalis

424

Fig. 2 Time course of tannin production of adventitious root cultures of Sanguisorba officinalis, 2.1 A

Effects of various media on growth and phenolic production in adventitious root cultures. To determine the effects of various media on the growth and phenolic production (1, 5, 6, 10 and 11), five basal liquid media [MS, half strength MS (1/2 MS) consisting of a half of macro salt formulation of MS, Gamborg B5 (B5) (42), Woody Plant (WP) (43) and Root Culture (RC) (44)] containing 1 mg/1 lAA were prepared. The adventitious roots were inoculated [ca 100 mg, fresh weight (fw)] into these five media (50 ml / 1(X) ml flask) and cultured for 4 weeks in the dark. The growth [dry weight (dw)] and phenolic production (% as dw) in the roots are shown in Table 1. Root growth was observed to be satisfactory except for RC medium. The roots, cultured in MS medium, showed the highest level of phenolic content especially in the production of 10. There was a positive correlation between growth and the production of phenolics, indicating the usefulness of this root culture for the production of secondary metabolites. 2.1.5 Effects of auxins on growth and phenolic production in adventitious root cultures The adventitious roots (ca 100 mg, fw) were also inoculated in MS liquid media (50 ml / 100 ml flask) containing lAA (0.1, 0.5 or 1 mg/1), indole-3-butyric acid (IBA) (0.1, 0.5 or 1 mg/1) or 1naphthaleneacetic acid (NAA) (0.01, 0.1 or 1 mg/1) and harvested after 4 weeks culture. The growth and phenolic production of these roots are shown in Table 2. The growth rate of the roots cultured with 0.5 or 1 mg/1 lAA was almost twice as that of those cultured with 0.1 mg/1 lAA. However, the levels of phenolic contents of these roots (cultured with 0.1, 0.5 or 1 mg/1 lAA) were almost the same. Addition of IBA or NAA, especially at the concentration of 1 mg/1, enhanced the growth of the roots remarkably, which increased ca 10-20 times as compared with roots cultured with low concentrations of IBA (0.1 mg/1) or NAA (0.01 mg/1). The phenolic production also slightly increased in proportion to IBA or NAA concentration. The highest production of phenolics was obtained when 1 mg/1 NAA was supplemented to the culture medium.

425 Table 1

Effects of various media on growth and phenolic production in adventitious root cultures of Sanguisorba officinalis

1 dw (mg)*

1 (% as dw) 1 5 6 10 11

MS 99.4

1/2 MS 66.9

0.10 0.10 0.04 5.87 0.49

0.06 0.17 0.03 3.73 0.46

B5 70.9 0.10 0.49 0.03 4.49 0.73

WP 82.0 0.14 0.19 0.14 5.15 0.51

RC

1

18.2 0.56 2.82 0.28 1.67

2.33

1

' dry weight of roots per a flask

Table 2

Effects of auxins on growth and phenolic production in adventitious root cultures of Sanguisorba officinalis

material dw (mg)^ lAA (mg/1) 91.8 0.1 120.7 0.5 1 153.3 |lBA"(mg^) 0.1 43.3 ^ 152.4 0.5 1 183.3 [NAACmg/T) 0.01 ''"Al.r "" 138.8 0.1 1 314.5

1 0.03 0.05 0.03 0.02 0.02 0.03 0.03 0.05

2 0.21 0.34 0.30 0.08 0.26 0.20 -

% dry weight 6 5 0.06 0.15 0.07 0.16 0.08 0.15 0.02 0.10 0.01 0.03 0.01 0.06 0.03 0.10 0.08 0.13 0.34 0.17

10 3.23 3.27 3.07 2.26 4.38 4.21 3.01 3.01 2.99

11 0.43 0.39

0.39 J 0.25 1 0.31

0.46 J 0.29 1 0.47 0.82

dry weight of roots per a flask

2.1.6 Effects of gibberellic acid (GA3) on growth and phenolic production in adventitious root cultures Since Kamada etal. (45) and Ohkawa etal. (46) reported that GAo enhanced the growth and alkaloid production in the hairy root culture of Datura innoxia, the effects of GA^ v^ lAA, IB A or NAA combinations on the growth and phenolic production were determined. The roots (ca 50 mg, fw) were inoculated in MS liquid media (50 ml / 100 ml flask) containing various combinations of GA3 (0, 0.001,0.01, 0.1 or 1 mg/1) and 1 mg/1 lAA, IBA or NAA and cultured for 4 weeks (Table 3). In the combination with 1 mg/1 IBA or NAA, GA^ strongly inhibited the growth of the roots. GAo also exhibited no prominent effect on phenolic production when 1 mg/l lAA, IBA or NAA was present in the culture medium.

426 Table 3

Effects of GAo on growth and phenolic production in adventitious root cultures of Sanguisorba officinalis % dry weight

iw (mg)^ GA3 (mg/1)! < material lAA (1 mg/1) 0 ! 191.6 0.001 I 153.2 0.01 I 150.0 0.1 I 156.9 1 ! 159.7 1 IBA'(lmg/l) """"o"""T 222.4 ^"" 0.001 1 196.4 0.01 I 128.3 0.1 I 137.6 1 1 122.4 |NAA(lmg/l) """'o"'"T 408.1 ""r 0.001 ! 271.2 ! 0.01 I 170.1 ! 0.1 I 156.2 I 1 ! 126.2 1

1 0.09 0.04 0.02 0.03 0.03 0.03 0.09 0.07 0.03 0.05 0.04

2 0.28 0.14 0.19 0.19 0.12 0.13 0.08 0.07 0.18 0.16 0.22

5 0.33 0.14 0.07 0.05 0.15 0.14 0.05 0.04 0.22 0.16 0.06 0.05 0.09

6 0.16 0.04 0.01 0.02 0.02 0.07 0.07 0.04 0.05 0.04

| 10 3.92 5.16 4.36 2.65 2.88 3.82 2.89 3.11 3.29 2.60 3.88 3.32 3.80 3.17 4.01

11 1 0.79 0.47 0.28 0.24 0.28 0.58" 0.44 0.33 0.45 0.37 0.68" 0.65 0.30 0.21 0.31

^ dry weight of roots per a flask

2.2 Phenolics in hairy root cultures I.IA Establishment of hairy root cultures For the induction of the hairy roots from S. officinalis, A. rhizogenes A4 strain harbouring root-inducing (Ri) plasmid was used, but the infection of the bacteria to the plant tissues was fairly difficult due to the strong antibiosis of its phenolic constituents. The direct infection method with A. rhizogenes A4 did not succeed because the infected sites of the cut ends of the plant, exposed to the air, readily turned brown and then began to die. Therefore, we selected the co-culture method for the induction of the hairy roots. The leaf segments cut from the axenic plants obtained above (see 2.1.2) were used for the explants for transformation. The bacteria of A. rhizogenes A4 subcultured on YEB agar medium (47) was transferred to YEB liquid medium (20 ml / 100 ml flask) and precultured for one day in the dark at 25 *C on a rotary shaker (100 rpm). The solution of this A. rhizogenes A4 (200 jil) and the leaf segments of the plant were inoculated to 1/2 MS liquid medium (20 ml / flask) and co-cultured for 2 days in the same conditions as above. The infected leaf segments, after rinsing with sterile water, were transferred to 1/2 MS medium (solidified with 0.2 % gelrite containing 0.5 g/1 Claforan®) and incubated at 25*C in the dark. After 3 to 4 weeks, 12 hairy roots appeared on the segments. The tips of the hairy roots were cut off and cultured on the same medium to eliminate the bacteria. The axenic hairy roots thus obtained were maintained in hormone-free 1/2 MS liquid medium on a rotary shaker in the dark. Three clones (SoH-1-3) which showed sufficient growth were selected and used for the experiment. In 1985, Stachel etal reported that acetosyringone was the signal molecule which activated the virulence gene expression of Agrobacterium (48). Thereupon, for the induction of the hairy root, we also used A.

427 rhizogenes A4 which was precuhured in the medium (YEB) containing acetosyringone (100 mM). Procedures for the infection with the bacteria and the maintenance of the hairy roots were the same as mentioned above. In this case 10 hairy roots were obtained, three of which (So-H-4-6) were selected. Opines (agropine and mannopine) of the hairy roots were extracted and detected using high voltage paper electrophoresis (49) (data not shown). 2.2.2 Phenolics in hairy root cultures So-H-1-6 (ca 50 mg, fw) were inoculated into hormone-free 1/2 MS liquid medium (50 ml / flask) and cultured under the same conditions as above for 4 weeks. The growth and phenolic (1, 5 , 6 , 1 0 and 11) production of these hairy roots are shown in Table 4. The growth rates of So-H-1, 3, 4 and 6 in this medium were fairy superior to those of the adventitious roots cultured in MS, 1/2 MS, B5, WP and RC liquid media containing 1 mg/1 lAA (Table 1). Particularly, So-H-3 showed prominent growth which was almost 3 to 5 times larger than those of other hairy roots. In five clones (So-H-1, 2 and 4-6) the major phenolic was 10 (0.217-0.569 % as fw) whose levels were higher than that of the parent plants (0.206 % as fw). So-H-5, in spite of its poor growth, showed the highest level of 10. On the other hand, So-H-3 which showed the fastest growth, indicated a high production of 11 (0.221 % as fw) and 5 (0.332 % as fw). Tanaka etal reported that in the intact plant the content of 11 was not so high (0.119 % as fw) and 5 was not detected (36). Therefore So-H-3 might have developed the capability to produce specifically the secondary metabolites which were not biosynthesized as much in the intact plant. This clone was also morphologically different from the other clones (Fig. 3). These results indicate the importance (and necessity) of the selection of clones as well as the culture media in the use of hairy and adventitious roots for the production of secondary metabolites.

Table 4

Growth and phenolic production in hairy root cultures of Sanguisorba officinalis

material

1

1 "^ 1

1

fw (g)^ ' (Iw (mg)^ [1 2.21 So-H-1 130.7 ; 0.003 (0.05) 1.71 101.8 1 0.002 '^ (0.04) 8.92 624'.6 ]" 0.113 (1.62) 220".r T 0.005 3.95 '^ (0.08) T" • 0.81 " T 83.3 T 0.005 (0.05) 1.45 'nil T 0.007 (0.08)

1 "^

1 '^

^ weight of roots per a flask

% as fw ; (%) as dw 5 6 10 0.001 0.228 0.008 (3.84) (0.14) (0.01) 0.264 0.001 0.003 (0.05) (0.01) (4.43) 0.040 0.322 0.089 (0.57) (4.60) (1.28) 0.011 0.001 0.217 (0.19) (0.02) (3.89) 0.012 0.011 0.569 (0.12) (5.53) (0.11) 0.021 0.264 0.007 (0.25) (0.08) (3.09)

11 0.085

(1.44) 0.048 (0.80) 0.221

1 1 1 1

(3.16) J 0.124 1

(2.21) 1 0.074 1

(0.72) 1 0.166 1 (1.94)

428

Fig. 3 Hairy roots of Sanguisorba officinalis (left: So-H-1, right: So-H-3)

2.3 Perspective of adventitious and hairy root cultures The adventitious and hairy root cultures of S. officinalis seem to be useful not only for the production of high Mr hydrolyzable tannins such as 10 and 11 but also for biosynthetic studies. The phenolic content in the hairy roots was almost the same as that of the adventitious roots. Taking into account the rapid growth, the hairy root cultures of this plant are more valuable than its normal root cultures for the production of these phenolics. 3.

HAIRY ROOT CULTURES OF GERANIUM THUNBERGU Geranium thunbergii Sieb. et Zucc. (Geraniaceae) is one of the most important medicinal plants generally used in Japan against diarrhea. This plant has been shown to be a rich source of ellagitannins such as geraniin (1 2) (50), corilagin (13) (51) and elaeocarpsin (14) (52-53) (Fig. 4) etc. With the discovery of some antiviral actions of 12, the major constituent of G. thunbergii, and its related phenolics in recent work (11), this plant is also expected to be used as a new antiviral agent. 3.1 Establishment of hairy root cultures The axenic plants (shoot culture) of G. thunbergii were obtained by the same method as above (see 2.2.1). The petioles were cut off from the shoot culture and used for the explants for the direct infection with A. rhizogenes A4. The bacteria was applied by a needle to the cut ends of the petioles and, 2-3 weeks after inoculation, several hairy roots appeared at the infected sites. The hairy roots, after elimination of the bacteria on 1/2 MS solid medium containing an antibiotic (Claforan®), were transferred and maintained in hormone-free 1/2 MS and B5 liquid media in the dark at 25*C. Among 20 clones of the hairy roots, one clone which showed the fastest growth was used for the experiment. 3.2 Isolation of phenolics from hairy roots 3.2.1 From the hairy roots cultured in 1/2 MS medium The hairy roots, cultured in 1/2 MS liquid medium for 4 weeks, were lyophilized and extracted with 80% aqueous acetone. The extract, after concentration, was subjected to column

429 chromatography of Sephadex LH-20 (water-MeOH) and seven phenolics, 1, 2, 5, 6, 8, 13 and 1,6-di-O-galloyl-p-D-glucose (15) (40) (Fig. 4) were isolated.

HO O H

HO,

OH OH HO OH

HO,

?o^^°

/=<

M

9

OH

Y" /

OH

HaCj

?

CO (HO

^ // O

O:

-OH

OH

OH

OH

OCO-^A-OH

HO O H

OH

13 (corilagin)

12 (geraniin)

HO^

OH

CO

= \ " /==\ OH

OH

HO

OH

PH

OH

Ho\,...^-^-.\^OCO—/

OH V-0 OH

15 (1,6-dl- O -galloyl-p-D-glucose)

OH

O-^O O

'; HO

';'*0 OH

14 (elaeocarpsin) Fig. 4 Phenolics of Geranium thunbergii

16 (ellagic acid)

430 3.2.2 From the hairy roots cultured in B5 medium Lyophilized hairy roots, cultured in hormone-free B5 liquid medium for 4 weeks, were extracted as above and subjected to the combination of column chromatographies using Sephadex LH-20 and MCI-gelCHP-20P to afford six phenolics,!, 2 , 8 , 12, 13 and ellagic acid (16) (Fig. 4). These phenolics obtained from the hairy roots were also isolated from the leaves of the intact plant of G. thunbergii. 3.3 Production ofphenolics in the hairy root cultures The hairy roots (ca 70 mg, fw) were inoculated into hormone-free 1/2 MS and B5 liquid media (50 ml / flask) and cultured in the dark on a rotary shaker (1(X) rpm). The growth is shown in Fig. 5. In both media, the hairy roots grew successfully and reached the maximum levels (3.86 g, fw in 1/2 MS and 2.22 g, fw in B5) at week 6. In these cultures, after 2 to 3 weeks, the tips of the hairy roots began to show a blueish purple colour which was presumed to be caused by a complex between the phenolics in the hairy roots and metal ions contained in the culture medium (Fig. 6). The appearance of this colour also indicated that the hairy roots were rich in phenolics such as tannins. As shown in Fig. 6, the hairy roots grew with rich branching only when they were cultured in 1/2 MS medium.

'53

UH

Time (weeks) Fig. 5 Growth of hairy roots of Geranium thunbergii

431

Fig. 6 Hairy roots of Geranium thunbergii (left: cultured in 1/2 MS liquid medium, fight: cultured in B5 liquid medium)

The contents of phenolics (1, 2, 5, 6, 12 and 13) in the hairy roots were determined by HPLC (Fig. 7 and 8). HPLC conditions were the same as those mentioned above (see 2.1.3); Rt (min): 1 (6.52), 2 (12.46), 13 (14.63), 12 (17.34), 5 (22.54)and 6 (27.04). In the hairy roots cultured in 1/2 MS medium (Fig. 7), galloylglucoses were the major products; 6 was the major compound whose maximum content was ca 0.4 % (as dw) at week 3. On the other hand, in B5 medium (Fig. 8), the major constituent was 12 (0.67 % as dw at week 3) the content of which was ca 70 % of that detected in the roots of the parent plants. This observation was very interesting because it suggested the possibility of biosynthetic regulation of ellagitannins in G. thunbergii hairy roots in accordance with the constituents of the culture medium.

O.8• 0.6

0.4

":.

2

----i---

5

......... --~.........

6

......... •-~,:-........

12

...........~ , .......

13

t ':..

/

N

0.2

0.0

1

2

3

4

5

6

7

8

9

Weeks

Fig. 7

Contents of phenolics in Geranium thunbergii hairy roots cultured in 1/2 MS medium.

432 0.8

:

I 2

0.6

5

0.4

13

12

,a

0.2

0.0

i

,.r

~

T

"r

I

2

3

4

5

6

7

""

/

8

9

Weeks

Fig. 8 Contents of phenolics in Geranium thunbergii hairy roots cultured in B5 medium. 3.4 Effects of the constituents in the culture medium on phenolic production in the hairy root

cultures To determine the effects of the constituents in the culture medium on phenolic (tannin) production in G. thunbergii hairy roots, two culture media were prepared; one consisted of I/2 MS major and minor elements (inorganic elements) and B5 vitamins (named as "1/2 MS:vB5") and the other B5 major and minor elements and MS vitamins (B5:vMS). The hairy roots (ca 70 mg, fw) subcultured in 1/2 MS and B5 liquid media, respectively, were inoculated into four hormone-free liquid media (1/2 MS, B5, 1/2 MS:vB5 and B5:vMS). After four week-culture in the dark at 25 *(2, the contents of 1, 2, 5, 6, 12 and 13 in these hairy roots were determined by HPLC (Table 5). Table 5

Contents of phenolics in hairy roots of Geranium thunbergii cultured in various media and in the roots of the intact plant ,

,

medium a

, '

medium b

B5

i

B5

B5 B5 B5 1/2 MS 1/2 MS 1/2 MS 1/2 MS

, ,

, , ,

content (% as dw) 1

i 0.012

2

5

6

12

13

0.001

0.02i

0.046

0.443

0.031

I

,

1/2 MS

, 0.019

0.001

0.126

0.241

0.087

, , ,

1/2MS:vB5 B5:vMS 1/2MS B5 B5:vMS 1/2MS:vB5

, I , ~ 0.025 ,0.048 , 0.013 ', 0 . 0 1 3

0.001 0.024 0.002

0.109 0.017 0.068

0.239 0.040 0.394

0.042 0.507 -

0.091

0.044 0.073

0.048 0.043

0.624 0.648

0.056

0.264

0.401

0.173

-

-

-

0.988

I"

,

', ', I

, r

roots of the plant c , I

,,=

_

0.138 0.048

r

'I 0.038

aMedium in which the hairy roots were maintained. bMedium in which the hairy roots were transferred and cultured for 4 weeks. CCultivated in the field (in Japan) and collected in September. -: not detected.

0.155

433 Without reference to the original media in which the hairy roots were subcultured (1/2 MS or B5), the hairy roots cultured in 1/2 MS and 1/2 MS:vB5 media contained mainly 6, while the hairy roots cultured in B5 and B5:vMS media produced 12 as the major product. The pattern of the production of tannins in the hairy roots cultured in 1/2 MS:vB5 medium was approximately the same as that of the ones cultured in 1/2 MS medium. Similarly, the tannin pattern in the hairy roots cultured in B5:vMS medium was almost identical to that cultured in B5 medium. These results suggested that the tannin production in G. thunbergii hairy roots was not regulated by the vitamins but by inorganic constituents in the media. The main difference in the constituent of inorganic elements between 1/2 MS and B5 media was NH4 + (10.3 mM in 1/2 MS and 2 mM in B5). Previously, Deno etal reported that NH4 + in culture medium had an effect on the metabolism of tropane alkaloids in the adventitious root culture in Duboisia myoporoides R. Br. (54) and Yazaki et al reported that it inhibited shikonin production in Lithospermum erythrorhizon cell culture (55). Based on these effects of NH4 + on secondary metabolism in plant tissue cultures, we suggest that the tannin metabolism (especially ellagitannin) in G. thunbergii hairy roots might be regulated by NH4 + acting as an inhibitor in the oxidative pathway from galloylglucoses into 12 (Fig. 9). Therefore, the accumulation of 6 in the hairy roots cultured in 1/2 MS medium (Fig. 7) might be due to the inhibition of its further metabolism. This example of G. thunbergii hairy root is the first one which succeeded in the production of tannins by transformed plant cells [recently, Morris et al reported the formation of condensed tannin by the hairy roots of Lotus corniculatus (56)]. G. thunbergii hairy roots seemed to be very useful for the biosynthetic experiments of tannins (ellagitannins).

Glucose

OH

+ HO H

OH

HO I10 OH OH

COOH

•.

140

".

,co i ~/o

~"~

gallic acid (1)

"~1 3 ° ° ~ /

~

OH

/1~ HOOH O ~ H

_

.o

o

co

o.

o

o.

~ ~

1,2,3,4,6-penta-O -

galloyl-13-D-glucose (6)

to

.

OH

o.

•

"

NH4 ÷

HO HO Oll OH

'.

°"

~HO H-OOH ~ " ~ OH H,-,--,---,------*---OH

~o / o. o ,co ~ . ~ o, ~ c o ~ - , ~ o .

oH

o.

corilagin 1131

HO HO OH OH ~ HO OH

.o~ co .o-~/ ' ~ ) - - c o~'OCHI -o--~ o / . ~__OH o.~.~.o~o ~ -o. .

/_~

in 1/2 MS medium ~" in B5 medium

Fig. 9 Biosynthesis of geraniin in Geranium thunbergii hairy roots

"

~co o~~~OH HO"~)H"~Of

o, ¢o O==~~~OH ~O. q4[-- H0 OH

geranlin (12)

°-~7

"o~

434 HAIRY ROOT CULTURES OF SWERTIA JAPONICA

4.

The whole plant of Swertiajaponica Makino (Gentianaceae) is an important bitter stomachic in Japan where it is called "senburi". The plant is also claimed to be effective in the treatment of hepatitis (57). Components of this plant such as the bitter secoiridoids (58-63), phenyl glucosides (64), flavonoids (65) and xanthones (66-68) have been intensively studied mainly by Japanese researchers but, to date, there are few reports on the production of these constituents in vitro. Callus culture produced no bitter principles (secoiridoids) (69, 70), xanthones and flavones but coumarin derivatives, scopoletin and its glucoside were detected (71).

4.1 Establishment of hairy root cultures The plant S. japonica (2 years old), grown in the field, was collected and used for the establishment of the shoot cultures. The deleaved fresh shoots, after surface sterilization with 2 % NaCIO with Tween 20 (see 2.2.1), were cultured on hormone-free 1/2 MS medium in the light (16 hr photoperiod/day). A. rhizogenes 15834 strain harbouring Ri plasmid (pRi 15834) was inoculated by a needle onto the cut ends of the axenic stems. Two to 4 weeks after inoculation, hairy roots appeared at the inoculated sites. The hairy roots, after removal of the bacteria, were transferred and maintained on hormone-free RC solid medium in the dark. 4.2 Growth of the hairy roots Ca 20 mg (fw) of hairy roots were inoculated into hormone-free 1/2 MS, 1/2 B5 and RC liquid media and cultured at 25°C on a rotary shaker (80 rpm) in the dark. The growth rates were shown in Fig. 10. The amount of the hairy roots increased satisfactorily after week 2 in these media, particularly the growth rates observed in 1/2 B5 and 1/2 MS media were almost 3 times larger than that in RC medium. In these cultures, the hairy roots showed pale yellow colouration suggesting the enrichment of yellow pigment (xanthones) in the roots. In addition, the hairy roots were slightly bitter to the taste which indicated the presence of bitter principles. 12

I/2

10 ~ e~

B5

I / 2 MS RC

~--

8

"3

6

,.-,

4

0

1

2

3

4

Fig. 10 Growth of hairy roots of Swertia japonica

5

6

7 8 Time [weeks]

435 4.3 Constituents of hairy roots 4.3.1 Isolation from hairy roots Lyophilized hairy roots cultured in RC liquid medium for 6 weeks were extracted with MeOH. The extract, after concentration, was mixed with sufficient amount of water and partitioned with CHCI 3" The CHCI 3 layer, after evaporation to dryness, was subjected to a silica gel column chromatography using C6H6-EtOAc to afford two xanthones, bellidifolin (17) (67) and methylbellidifolin (18) (67) (Fig. 11). The aqueous layer was chromatographed over Sephadex LH-20 (with 60 % MeOH) and MCl-gel CHP-20P (water-MeOH) to give swertianolin (19) (68), the new xanthone 20, 1-O-sinapoyl-13-D-glucopyranoside (21) (72, 73), two new phenyl glucosides 2 2 and 23 (Fig. 11) and one fraction A which contained the bitter principles.

R2.O

O

OH Me

17 18 19 20

.

HOCH2

(bellidifolin) R1, R2=H RI=Me, R2=H RI=H, R2=-GIc RI=H, R2=-GIc-Xyl

_

OH

o_Q

OH OMe

21

.O o_Q

coo.

HOC.H20

COO--~OH

COOMe

22 HOCH2

OMe

HOC"H2 H~OC/O_~-" OH

R~O

22a

o- o OMe

OMe

23

O

O

0 . °.

GIc-C~ H

OH

24 (amarogentin) R=H 25 (amaroswerin) R=OH Fig. 11 Phenolics in Swertiajaponica hairy roots

436 4.3.2 Xanthones A new xanthone 2 0 was obtained as pale yellow needles (mp 265 *C) and its 1H-nuclear magnetic resonance (NMR) (in DMSO-d6) spectrum showed methoxyl (8 3.89), meta-coupled aromatic (8 6.37, 6.58) and ortho-coupled aromatic (8 7.19, 7.28) signals. These signals were closely correlated to those of 19. The 13C-NMR spectrum of 2 0 (Table 6) also demonstrated the presence of a similar xanthone (1,3,5,8-oxygenated) structure to that of 19. Furthermore, the 1H. NMR spectrum indicated two anomeric proton signals [8 4.77 (J=7.5 Hz), 8 4.21 (J=7.5 Hz)]; the coupling constant and chemical shift of the former signal were closely related to the glucose H-1 signal of 19. The presence of a glucose moiety was also suggested by the six carbohydrate signals (8 65.8, 69.8, 73.5, 75.9, 76.5 and 103.8) in the 13C-NMR spectrum.

Table 6

13C.NM R spectral data of 17-2 0 (5 values at 67.5 MHz, in DMSO-d 6 + D20 ) C

20

17

18

19

1 2 3 4 4a 5 6 7 8 8a 9 9a 10a -OMe

162.7 97.5 166.6 92.6 156.7 140.9 121.8 112.6 149.7 112.0 181.3 103.0 145.3 56.4

161.9 97.5 167.1 92.9 157.4 137.3 123.8 109.5 151.8 107.5 184.0 102.1 143.3 56.2

161.8 97.8 167.2 93.1 157.4 139.7 120.7 109.2 152.7 107.6 183.9 102.3 144.6 56.4 56.7

162.7 97.2 166.3 92.2 156.4 141.0 121.1 112.3 149.4 111.9 181.1 103.1 145.0 56.1

1 2 3 4 5 6 1 2 3 4 5

103.8 73.5 75.9 69.8 76.5 65.8 104.3 73.5 76.3 69.7 68.7

103.5 73.5 76.1 69.7 77.4 60.8

437 Five additional carbohydrate signals (8 68.7, 69.7, 73.5, 76.3 and 104.3) were also observed and the chemical shifts of these signals suggested the presence of a xylose moiety. Acid hydrolysis of 2 1) with HCI gave D-glucose, D-xylose and an aglycone which was identical to 17. The linkage of the glucose moiety to the C-8 position was indicated by the close similarity of the 13C-NMR signals for the xanthone moiety and glucose C-1 in 2 0 with those of 19. The glucose C-6 signal was observed relatively downfield (8 65.8) compared to that of 19 (8 60.8) which indicated that the xylose C-1 position was connected to the glucose C-6 position. The configuration of the anomeric centres of glucose and xylose was concluded to be based on the J value (7.5 Hz) of their H-1 signals. Together with further support from the fast atom bombardment (FAB) mass spectrum (MS) of 21) which showed a prominent ion peak at m/z 591 [M+Na] +, 21) was concluded to be 8-0primeverosylbellidifolin. It is biosynthetically interesting that the hairy roots produce only 1, 3, 5, 8-oxygenated xanthones although in vivo plants produce both 1, 3, 5, 8- and 1, 3, 7, 8-oxygenated derivatives which were presumed to be biogenetically equivalent (67). 4.3.3 Bitter principles The bitter principles produced in the hairy roots were detected and identified to be amarogentin (2 4) (63) and amaroswerin (2 5) (63) (Fig. 11) by HPLC analysis of the aqueous layer and fraction A (see the part of isolation). Samples (aqueous layer and fraction A) were separately subjected to HPLC on a Nucleosil 100-5C18 (4.6 mm i. d. x 250 mm) column, mobile phase MeOH-water (9:10), flow rate 0.8 ml/min, column temperature 35 *C, Rt (min): 2 5 (11.1) and 2 4 (13.1). Although the contents of these compounds were fairy low (below 0.0001% as dw), this S. japonica hairy root culture was the first example which succeeded in the production of bitter principles by plant tissue cultures. 4.3.4 New phenyl glucosides 2 2 showed [M+Na] + ion peak at m/z 459 in the FAB-MS. The 1H-NMR (in DMSO-d6) spectrum exhibited an anomeric [8 4.95 (d, J=7.2 Hz)], ABX-type aromatic [8 6.68 (1H, d, J=8.8 Hz), 7.04 (1H, dd, J=8.8, 3.0 Hz), 7.47 (1H, d, J=3.0 Hz)] and ABCD-type aromatic [8 7.40 (1H, br. dd, J=7.9, 2.6 Hz), 7.52 (1H, t, J=7.9 Hz), 7.71 (IH, br. dd, J=2.6, 1.6 Hz), 7.75 (1H, br. d, J=7.9 Hz)] proton signals. The 13C-NMR spectrum (Table 7) showed signals due to one glucose moiety (8 60.5, 69.5, 73.2, 76.3, 77.0, 100.7), two aromatic rings (12 carbons), one carbonyl (8 164.7) and one carboxyl (8 170.4) carbons. Furthermore, the 1H-13C long range shift correlation spectrum (JcH=I0 Hz) indicated the correlation of the carbonyl carbon at 8 164.7 (C-7') with the aromatic proton at 8 7.71 (H-2') through a three-bond coupling, suggesting that the carbonyl carbon was connected to C-1' position (but the cross peak between C-7' and H-6' was not observed). Methylation of 22 with diazomethane gave a methyl ester 22a whose FAB-MS exhibited an intense [M+Na] + ion peak at m/z 337. The 1H-NMR (in DMSO-d6) spectrum of 22a showed an anomeric [8 4.90 (d, J=7.1 Hz)], one methoxyl (8 3.85) and ABCD-type aromatic [8 7.33 (1H, dd, J=8.0, 2.2 Hz), 7.46 (1H, t, J=8.0 Hz), 7.57 (1H, br. d, J=2.2 Hz), 7.61 (1H, br. d, J=8.0 Hz)] proton signals. In this spectrum, the signals due to the ABX-type aromatic ring, present in 2 2, were absent. This suggests that 2 2, on methylation with diazomethane, was readily cleaved at the ester linkage (C-7' position). The 1H-IH nuclear Overhauser enhancement and

438 exchange spectroscopy (NOESY) spectrum of 22a revealed NOE correlation cross peaks between the glucose C-1' proton and two aromatic (H-2 and H-4) protons, and between methyl proton (H-8) and two aromatic (H-2 and H-6) protons. This observation suggested that the glucose moiety was linked to the C-3 position through its C-1' position and that the COOMe group was connected to the C-1 position. The configuration of the anomeric centre was concluded to be 13from the J value (7.1 Hz) of the H-I' signal in the 1H-NMR spectrum of 22a. Therefore, 22a was characterized as 3-013-D-glucopyranosyl benzoic acid methyl ester. Acid hydrolysis of 2 2 also furnished two phenolics, one of which was isolated and identified as gentisic acid. In the 1H-NMR spectrum of 2 2, the aromatic signals attributed to H-6 (5 7.47) and H-4 (5 7.04) were considerably shifted downfield compared with those of gentisic acid [5 7.15 (H-6), 6.96 (H-4)], indicating that in 2 2 the 3glucosyl benzoic acid moiety was located at the C-5 position through its carbonyl carbon. From the chemical and spectral data above, 22 was concluded to be 5-(3'-O-13-D-glucopyranosyl) benzoyloxygentisic acid. Table 7

13C.NMR spectral data of 21-2 3 (8 values at 67.5 MHz) C

22 a

23 b

21 b

1 2 3 4 5 6 7 8 9 1' 2' 3' 4' 5' 6' 7'

120.6 160.4 116.0 124.5 140.2 122.1 170.4

129.4 154.6 94.6 155.9 94.6 154.6

140.1 107.4 149.4 126.3 149.4 107.4 148.0

2 3 4 5 6 OMe a in DMSO-d 6 b in acetone-d6 + D20

115.7 166.9 130.5 117.4 157.5 121.5 130.0 123.0 164.7 73.2 76.3 69.5 77.0 60.5

75.6 77.5 70.8 77.9 62.3 57.0

74.0 78.1 71.3 78.7 62.6 57.1

}

439

23 exhibited a prominent [M+Na] + ion peak at m/z 355 in the FAB-MS. The 1H and 13CNMR (Table 7) spectra showed the presence of one glucose moiety, a symmetrically substituted aromatic ring [~i 94.6, 129.4, 154.6, 155.9, 8 6.10 (2H, s)] and two methoxyl groups [5 3.70 (6H, s)]. By comparing these spectral data with those published [Saijo et al obtained 2 3 by tannase hydrolysis of the gallate isolated from Mallotus japonicus (Euphorbiaceae) (74)], 2 3 was concluded to be 2,6-dimethoxy-4-hydroxyphenol 1-O-13-D-glucopyranoside. As mentioned above, the hairy roots of S. japonica produced several glycoside compounds [phenyl glucosides (21-2 3), glycosylated xanthones (19 and 2 0) and bitter principles (2 4 and 2 5)], three of which (2 0, 2 2 and 2 3) have not been detected in natural plants. This observation, indicating the strong glycosylitic activity in S. japonica hairy roots, also suggests the possibility of the application of the hairy roots for biotransformation (glycosylation in secondary metabolism). 5.

ADVENTITIOUS ROOT CULTURES OF RHEUM PALMATUM Rhubarb ('daio' in Japanese), the rhizome and root of Rheum plants (Polygonaceae), is one of the most important traditional Chinese medicines used in combination with other crude drugs for bloodstasis syndrome, hypertension, renal disorder and so on. Recently, some new biological activities such as psychotropic (7), improvement of nitrogen metabolism (3-6), and inhibition of angiotensin-converting enzyme (75) were discovered and proved to be due to its polyphenol (condensed tannin) constituents such as rhatannins (2 6 and 2 7) (6) (Fig. 12). In spite of many phytochemical, biochemical and pharmacological examinations, there have been few reports [formation of anthraquinones (76) and sennosides (77)] on tissue cultures of rhubarb plants.

F

/

L :o OC o.O.3 H

OH

HO

Fig. 12 Chemical structures of rhatannins

___/

OH

26 (rhatannln I) R = - O C O - ~ O H OH 7-8

27 (rhatannin Ii) R=-OH

440 5.1 Establishment of adventitious root cultures The petioles, cut from the plants cultivated in the field, were surface sterilized (see 2.1.2) and aseptically inoculated on MS solid media containing various concentrations of 2,4-D, IAA, or NAA and/or benzyladenine (BA) in the dark at 25 *C. After about 4 weeks of culture, several adventitious roots appeared on the segments cultured with 0.5 or 2 mg/1 NAA. The roots were cut off and transferred to MS liquid medium (50 ml / flask) supplemented with 2 mg/1 NAA and maintained over 1 year on a rotary shaker (100 rpm) in the dark. 5.2 Constituents of adventitious roots 5.2.1 Isolation from adventitious roots The adventitious roots, cultured in MS liquid medium containing 2 mg/l NAA for 4 weeks, were lyophilized and extracted with 80% aqueous acetone. The extract, after concentration, was subjected to Sephadex LH-20 column chromatographies using EtOH and 60 % MeOH to give a phenolic compound 2 which is one of the structural elements of rhatannins. 5.2.2 Production of 2 in adventitious root cultures The roots (ca 100 rag, fw) were inoculated to three different media (MS medium containing 2 mg/l NAA and B5 medium containing 0.5 mg/l NAA or 0.5 mg/l IAA, each 50 ml/flask) and cultured for 8 weeks in the dark. The growth rate and production of 2 in these root cultures are shown in Fig. 13. ,,

6"

B5 IAA 0.5

56.a

d= 09 t.l.,

4

"

B5 NAA 0.5

3"

---

MS NAA 2.0

2" 1

t

r

0 0

I

I

I

i

I

I

I

!

1

2

3

4

5

6

7

8

0.25 0.20 I~ B5 IAA 0.5

~. 0.15

II B5 NAA 0.5 !~1 MS NAA 2.0

0.10 0.05 0.00 1

2

3

4

5

6

7

8

(Weeks) Fig. 13 Growth and catechin content in adventitious roots of Rheum palmatum cultured in MS (2 mg/l NAA) and B5 (0.5 mg/1NAA, 0.5 mg/1 IAA) liquid media.

441 Among the three media tested, MS liquid medium supplemented with 2 mg/1 NAA exhibited the most preferable effect on the growth of the roots. The growth rate of the roots cultured in this medium was approximately 3 times larger than that of those cultured in B5 medium. Comparison of B5 media supplemented with 0.5 mg/1 NAA with those containing 0.5 mg/1 lAA, showed very little difference in the effects on the growth of the roots. On the other hand, the content of 2 in the roots cultured in B5 liquid medium with 0.5 mg/1 NAA was over 2 times larger than that of the one cultured in B5 medium with 0.5 mg/1 lAA. The roots, cultured in MS liquid medium containing 2 mg/1 NAA showed the highest content of 2 for the most part of the culture period (except at week 2). High levels of the content of 2 were observed both at the early (until week 2) and last (after week 7) periods of the culture. During the rapid growth of the roots cultured in MS liquid medium with 2 mg/1 NAA, the highest content of 2 appeared within the first week of the culture. The cell suspension cultures of this plant also showed good growth and formation of 2 (maximum level, 0.38 % as dw). Although the contents of 2 in tissue culture cells (both adventitious root and cell suspension cultures) were comparatively lower than those in the plants of Rhubarb (R. palniatuni-6.06 % as dw; R. officinale-lAS % as dw) (78), taking into account the rapid growth of the root and cell cultures, these cultures seemed to be useful for the production of 2. 6.

ADVENTITIOUS AND HAIRY R(X)T CULTURES OF PHYLLANTHUS NIRURI Phyllanthus rdruri L., a small plant which grows mainly in tropical and subtropical regions in Central and South American countries, and in India, is one of the most important traditional medicines used for the treatment of jaundice, asthma, hepatitis and urolitic disease. Intensive chemical examinations of this plant have been carried out and several constituents such as lignans (79-81), alkaloids (82-84), flavonoids (85), tannins (86) and phthalic acid (87) have been identified. In addition, especially in this decade, several pharmacological experiments have also been reported (88, 89). In spite of many phytochemical and biochemical investigations, there have been only a fewreportson tissue culture of this plant. 6.1 Establishment of hairy root cultures Seeds of P. niruri collected in Peru were sterilized and germinated aseptically on hormonefree 1/2 MS solid medium (see 2.2.1). Axenic plants were subcultured and used as explants. For the induction of the hairy roots, two types of Agrobacterium strains were used; one was the A4 strain harbouring pRi A4b, the other ^4. tumefaciens R-1000+121 having two plasmids, pRi A4b aiid a mini Ti plasmid (pBI 121) containing genes encoding for neomycin phosphotransferase II (NPT-II) and P-glucuronidase (GUS) on the T-DNA. These Agrobacterium strains were inoculated via a needle onto the stems of axenic plants in vitro. After 2 to 3 weeks with 16 hr light (2(X)0 lux) a day, several hairy roots appeared at the infected sites. These Agrobacterium strains showed no clear difference for the induction of the hairy roots. After removal of the bacteria on hormone-free MS solid medium containing an antibiotic (see 2.1.4), the axenic hairy roots were subcultured in hormone-free 1/2 MS liquid medium (50 ml / flask) in the dark on a rotary shaker (100 rpm). Among some hairy roots established, eight clones (Pn-H-1-5 induced by A. rhizogenes A4 and PnH-6-8 induced by A. tumefaciens R-1000+121) which showed sufficient growth in this medium were selected and used for the experiment. Detection of opines (agropine and mannopine) from

442 these hairy roots was carried out by paper electrophoresis. In Pn-H-6-8, GUS expression was confirmed with X-Glu (5-bromo-4-chloro-3-indolyl-P-D-glucuronide) as a substrate (90). 6.2 Establishment of adventitious root cultures Stem segments from axenic plants were cut off and cultured on 1/2 MS solid medium supplemented with 0.1 mg/1 NAA. After 2 to 3 weeks of culture in the dark, several adventitious roots appeared at the cut ends with some calli. The roots were cut off and maintained in 1/2 MS liquid medium containing 0.1 mg/1 NAA on a rotary shaker (100 rpm) under light (2000 lux, 16 hr light per a day). 6.3 Phenolics in P. niruri 6.3.1 Isolation ofphenolics from hairy roots Pn-H-6, harvested after 6 weeks of culture in hormone-free 1/2 MS liquid medium on a rotary shaker, was lyophilized and homogenized with MeOH. The MeOH extract was evaporated to dryness and subjected to Sephadex LH-20 column chromatography eluting with EtOH containing increasing proportions of water to afford four compounds, 1, (-)-epicatechin (2 8), (-)-epicatechin 3-(9-gallate (2 9), (-)-epigallocatechin 3-0-gallate (30) and fraction B. Fraction B was chromatographed over MCI-gel CHP-20P (elution with water-MeOH) to give 3 and (-)epigallocatechin (31) (Fig. 14). 6.3.2 Isolation ofphenolics from intact plants For the determination of the phenolic constituents of R niruri plant, we collected two materials; one was grown on a field in Peru while the other was grown under hydroponics at 20 "C in 14 hr light/day conditions. (a) Isolation from leaves and stems of the plant grown on the field Dried leaves and stems of the parent plants collected in Peru were mashed and extracted with MeOH. The extract was subjected to a combination of Sephadex LH-20, MCI-gel CHP-20P and Bondapak C|g Porasil B column chromatographies to give six compounds, 1, 12, 13, brevifolin carboxylic acid (32) (91), l,2-di-0-gaIloyl-3,6-(/?)-HHDP-p-D-glucose (33) (92) and terchebin (34) (93) (Fig. 14). (b) Isolation from plants grown under hydroponics The axenic plants obtained above (see 2.5.2) were transferred and cultivated under hydroponics for 3 months. The mature plants were harvested and separated into two parts (one was the leaf and stem portion while the other was the root portion). These parts were separately extracted with 80% aqueous acetone (leaf and stem) or MeOH (root) and subjected to the same column chromatographies mentioned above. From the extract of the leaf and stem portion, fiwc phenolics, 1, 12, 13, 1,2,4,6-tetra-O-galloyl-P-D-glucose (35) (40) and phyllanthusiin D (36) (94, 95) (Fig. 14) were isolated. On the other hand, from the extract of the root portion, five compounds 3 and 2 8-31 all of which were the flavan-3-ols produced in the hairy roots (Pn-H-6) were isolated.

443

OH .OH

HO,^^^s^^O.

28 (epicatechin)

^9 (epicatechin 3- O^allate) OH

3^ (,p,g3„oca,ech.n 3- Ogallate)

CO- o

T

p l^^Sk^.

iJ

HO

^

HO COOH

32 (brevifolln carboxylic acid)

31 (epigallocatechin) HO^

HO OH

^oH

OH OH

-^-CO-OH,cJ

OCO-Q^OH

/fe7

HO^

S?

^OH

P

/=\}* / = \ = = 4 V H ^ " ° OH O^'^OH

H ^

I -*-

_<-

I

OH

/=\}* / = \ ^ = H / H / " ° "

^ ^ ' ""

i

iiT

W

33 (1,2-dl-0-galloyl-3,6-HHDPB-D^lucose)

^»

34 (terchebin) HO

HO O H ^ O H

HO—^ HO.

V—Vi

CO

"^\^ HO

HO'^

/^OH

OH

/

6 ,co CO^

"a9?

°x

CO

M

OH

^OH

OH

^ ' ?^°'""4^*^

??

? ?

CO

CO

«>=^»°xM //^°" ""'CH.-O

35 (1,2,4,6-tetra- O-galloyl-p-D-glucose)

OH

COM« 36 (phyllanthusiin D)

Fig. 14 Phenolics of Phyllanthus niruri

444 The leaves and stems of P. niruri plants, cultivated under hydroponic conditions, contained similar kinds of hydrolyzable tannins to those produced in the parent plants collected in Peru. This result demonstrates the profitable use of hydroponics for a good supply of this important medicinal plant. The hydrolyzable tannins observed mainly in the aerial parts (leaf and stem) could not be isolated from its roots. This tissue specific distribution of phenolic compounds (hydrolyzable tannins and flavan-3-ols which were the structural elements of condensed tannins) observed in R niruri was also interesting, considering the biosynthetic position of secondary metabolites in this plant. 6.3.3 Phenolics in hairy root cultures The clones Pn-H-1-8, cultured in hormone-free 1/2 MS liquid medium for 6 weeks in the dark, were harvested, lyophilized and extracted with MeOH. The phenolics (1-3 and 2 8-31) in the extracts were determined by HPLC. HPLC conditions were as follows; column: Nucleosil 1005Cjg (4.6 mm i. d. x 250 mm), mobile phase: MeCN-50 mM NaH2P04 (1:19-2:3), flow rate: 0.9 ml/min, column temperature: 40t:, detect: UV (280 nm), R^ (min): 1 (5.3), 3 (12.5), 31 (16.1), 2 (17.3), 28 (20.0), 3 0 (20.6) and 2 9 (24.4). The contents of these phenolics in Pn-H-1-8 are shown in Table 8. The main phenolics contained in Pn-H-1-5 induced by A. rhizogenes A4 was 3 0 (0.177-0.449 % as dw). In these five clones, the ratio of the contents of 1-3 and 28-31 did not differ very much. Although the contents of 2 (0.002 %) and 29 (0.002 %) in Pn-H-3 were lower than those of the other hairy roots, this clone produced the highest amount of 3 0 (0.449 %). Three clones, Pn-H-6-8, induced by A. tumefaciens R-1000+121 also produced similar flavan-3-ols to those observed in Pn-H-1-5. There was no clonal variation in the production of these phenolics among the eight clones induced by the two types of Agrobacterium. Among the hairy roots, clone Pn-H-8 produced the highest amounts of 2 8 (0.238 %), 2 9 (0.066 %) and 3 0 (0.525 %). 6.3.4 Phenolics in adventitious root cultures The adventitious root (Pn-A) of P. niruri was inoculated into 1/2 MS liquid medium containing 0.1 mg/1 NAA and cultured for 6 weeks in the dark. The contents of phenolics in the roots were determined by HPLC by the same method as above (Table 8). The contents of flavan-3ols (especially 2 , 2 9 and 31) produced in Pn-A were fairy low. The content of 30 (0.026 %) in Pn-A was almost one-ninth to one-twentieth less than those in the hairy roots. Therefore, in this plant, the hairy root culture was a more useful method for the production of flavan-3-ols than the adventitious (normal) root culture. 6.4 Semi-large scale culture of hairy root To establish a semi-large scale cultivation of the hairy roots, Pn-H-6, which showed the strongest GUS expression, was cultured in a 6 1 air-lift type fermenter and the production of phenolics was examined. Ca. 7.9 g (fw) of Pn-H-6 was inoculated into hormone-free 1/2 MS liquid medium in the fermenter equipped with a paddle (20 rpm) and supplied with air (300 ml/min) in the light. The hairy roots grew satisfactorily and the weight of the roots reached ca 694 g (fw) after 7 weeks of culture. In this culture, some hairy roots (Pn-H-6a) were cut by the paddle and formed a globular mass of ca 3 cm in the diameter. The contents of phenolics both in Pn-H-6a and

445 in the hairy roots (Pn-H-6b) which were not clumped in this culture were examined by HPLC (Table 8). The contents of 2 8-3 0 in Pn-H-6a and Pn-H-6b were almost one-third to one-twentieth to that produced in Pn-H-6 which was cultured in 100 ml flasks on a rotary shaker. Therefore, this semilarge scale propagation of the hairy roots was not so effective, especially for the production of galloyl derivatives of flavan-3-ols, such as 2 9 and 3 0. In this experiment, we also tried the shoot cultures of this plant using a 6 1 air-lift type fermenter (Pn-S-1) and a 21 rotating drum type fermenter (I*n-S-2) (both in hormone-free 1/2 MS liquid medium) and crown gall cell (Pn-C) cultures, which were induced by A, tumefaciens C58 harbouring the mutant-type tumour-inducing (Ti) plasmid pGV 2215 (96), in hormone-free MS liquid medium under light (20(X) lux). The contents of phenolics produced in these shoot and crown gall cell cultures are also shown in Table 8. Both in Pn-S-1 and Pn-S-2, only flavan-3-ols were produced and the hydrolyzable tannins which were observed in the leaves and stems of the parent plants were not detected. The contents of flavan-3-ols in these shoot cultures were similar to those of Pn-H-6a and Pn-H-6b showing low contents of the galloyl derivatives 2 9 and 3 0. On the other hand, in the crown gall cells, phenolics (1 -3 and 2 8-31) were not detected. From the above results, it was apparent that for the production of hydrolyzable tannins using such liquid cultures (hairy root, adventitious root and shoot) further investigations will be required. Concerning the production of flavan-3-ols contained in the roots of this plant, the hairy root culture was the most useful method of those tested. The constitution of the flavan-3-ols found in the plant and the hairy roots and shoot cultures of P. niruri are very similar to that observed in leaves of Thea sinensis (green tea) (97). Cultures of P. niruri might also have the same pharmacological effects as those expected from tea leaves (eg. 98,99).

Table 8

Contents of 1 -3 and 2 8-31 in different cultures of Phyllanthus niruri

cultures Pn-H-1 -2 -3 -4 -5 -6 -7 1

1 -8

1 0.006 0.004 0.002 0.003 0.002 0.005 0.006 0.004

[ Pn-A ""r

d.oo"i

|Pn-H-6a"^r 0.014 1 -6b ! 0.005

1 Pn-S-l^

1 -2

! 0.004

1 Pn-C ""1 -: not detected

2 0.005 0.045 0.002 0.008 0.006 0.005 0.010 0.020 0.010 0.010 0.007

3 0.056 0.054 0.067 0.024 0.015 0.078 0.040 0.074 0.007 0.040 0.022 0.037 0.031

% dry weight 29 28 0.042 0.100 0.014 0.065 0.002 0.021 0.014 0.039 0.011 0.072 0.050 0.110 0.054 0.008 0.066 0.238 0.014 0.024 0.015 0.008 0.006 0.009 0.028 0.014 0.033

| 30 0.379 0.306 0.449 0.267 0.177 0.424 0.314 0.525 0.026 0.055 0.036 0.050 0.043

31

1

0.118 0.132 0.111 0.110 0.092 0.109 0.084

0.106 J

J

0.08"4 1 0.021 J 0.127 1 0.058 J

446 7.

CONCLUSION AND PROSPECTS Biosynthetic study of plant secondary metabolism which can lead to a stable and good supply of desired constituents is very important particularly for medicinal plants. Plant tissue cultures (both cells and organized cultures) have been successfully applied to several medicinal plants and resulted in the successful production of some useful secondary metabolites such as alkaloids and terpenoids. Recendy, in some solanaceous plants (Atropa belladonna etc), the artificial improvement of tropane alkaloid metabolism, directed at the high production of scopolamine, has been tried by Agrobacterium-mtdidXtd gene transfer (100). However, for the production of phenolics, especially for polyphenols in medicinal plants, sufficient research has not been carried out on tissue cultures and only cell suspension cultures have been usually employed. Some examples of root cultures (both adventitious and hairy roots) of medicinal plants demonstrated in this article indicate the useful application of root cultures for biosynthetic study of plant polyphenols as well as their production. The adventitious and hairy roots of Sanguisorba officinalis produced high Mr polyphenols such as sanguiins (10: Mr. 1870, 11: Mr. 3738) and were usable for the determination of the effects of some chemicals (medium constituents, auxins, GAo etc) on the production of these polyphenols. Swertia japonica hairy roots, producing several glycoside compounds as well as the bitter principles and yellow pigments, were clarified to possess strong glycosylation ability and may be expected to be used for biotransformation. Geranium thunbergii hairy roots seemed to be useful for the elucidation of the mechanism in hydrolyzable tannin metabolism. The adventitious roots of Rheum palmatum succeeded in the formation of 2, an important structural element of condensed tannins. Both adventitious and hairy roots of Phyllanthus niruri produced relatively high levels of flavan-3-ols which were similarly observed in the roots of the intact plant. From the results of these root cultures of medicinal plants, the system of hairy root culture seemed to be generally superior to normal root cultures for the production of secondary metabolites because of their good growth and stable productivity under simple culture conditions requiring no plant growth regulators. In addition, in the near future, hairy root cultures, followed by the establishment of the system (conditions) for regeneration from the roots, is likely to gain in importance for the modification of medicinal plants (preparation of transgenic plants) in order to improve the production of their useful secondary metabolites . REFERENCES 1 I. Nishioka, Yakugaku Zasshi, 103 (1983) 125-142. 2 J. M. Sieburth and J. T. Conover, Nature, 208 (1965) 52-53. 3 T Nagasawa, S. Shibutani and H. Oura, Yakugaku Zasshi, 98 (1978) 1642-1650. 4 T. Nagasawa, S. Shibutani, H. Oura, Y. Shoyama and I.Nishioka, Chem. Pharm. Bull., 28 (1980) 1736-1739. 5 S. Shibutani, T. Nagasawa, T. Yokozawa and H. Oura, Yakugaku Zasshi, 100 (1980) 434442. 6 S. Shibutani, T. Nagasawa, H. Oura, G. Nonaka and I. Nishioka, Chem. Pharm. Bull., 31 (1983) 2378-2385. 7 S. Ueki, G. Nonaka, I. Nishioka and M. Fujiwara, Journal of Medicinal and Pharmaceutical Society for Wakan-Yaku, 2 (1985) 502-503. 8 J-M. Delaisse, Y. Eeckhout and G. Vaes, Biochemical Pharmacology, 35 (1986) 3091-3094. 9 M. Nishizawa, T. Yamagishi, G. E. Dutschman, W. B. Parker, A. J. Bodner, R. E. Kilkuskie, Y-C. Cheng and K-H. Lee, J. Nat. Prod., 52 (1989) 762-768.

447 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

G. Nonaka, I. Nishioka, M. Nishizawa, T. Yamagishi, Y. Kashiwada, G. E. Dutschman, A. J. Bodner, R. E. Kilkuskie, Y- C. Cheng and K-H. Lee, J. Nat. Prod., 53 (1990) 587-595. J. Corthout, L. A. Pieters, M. Claeys, D. A. V. Berghe, A. J. Vlietinck, Phytochemistry, 30 (1991) 1129-1130. M. Sauerwein, K. Yoshimatsu and K. Shimomura, Plant Tissue Culture Letters, 9 (1992) 19. F. Constabel, Planta, 79 (1968) 58-64. D. J. Durzan, S. C. Chafe and S. M. Lopushanski, Planta, 113 (1973) 241-249. R. J. Westcott and G. G. Henshaw, Planta, 131 (1976) 67-73. E. E. Oberthur, R. L. Nicholson and L. G. Butler, J. Agric. Food Chem., 31 (1983) 660662. G. L. Lees, Plant Cell Reports, 5 (1986) 247-251. T. Cai, B. Daly and L. Butler, Plant Cell, Tissue and Organ Culture, 9 (1987) 245-252. Y Moumou, F. Trotin, M. Pinkas, J. Dubois and J. Vasseur, Ann.pharmaceutiques francaises, 45 (1987) 255-260. G. L. Lees, Plant Cell Reports, 7 (1988) 166-169. K. Yazaki and T. Okuda, Plant Cell Reports, 8 (1989) 346-349. K-H, Law and N. R Das, Phytochemistry, 28 (1989) 1099-1100. K. Yazaki and T Okuda, Phytochemistry, 29 (1990) 1127-1130. K. Yazaki, T. Yoshida and T Okuda, Phytochemistry, 30 (1991) 501-503. Y Moumou, J. Vasseur, F. Trotin and J. Dubois, Phytochemistry, 31 (1992) 1239-1241. K. Ishimaru, M. Hirose, K. Takahashi, K. Koyama and K. Shimomura, Phytochemistry, 29 (1990)3827-3830. K. Ishimaru, M. Hirose, K. Takahashi, K. Koyama and K. Shimomura, Plant Tissue Culture Letters, 8 (1991) 114-117. K. Ishimaru and K. Shimomura, Phytochemistry, 30 (1991) 825-828. K. Ishimaru, H. $udo, M. Satake, Y. Matsunaga, Y. Hasegawa, S. Takemoto and K. Shimomura, Phytochemistry, 29 (1990) 1563-1565. K. Ishimaru, H. Sudo, M. Satake and K. Shimomura, Phytochemistry, 29 (1990) 38233825. K. Ishimaru, M. Satake and K. Shimomura, Plant Tissue Culture Letters, 7 (1990) 159-163. K. Ishimaru, K. Yoshimatsu, T. Yamakawa, H. Kamada and K. Shimomura, Phytochemistry, 31 (1992) 2015-2018. G. Nonaka, T Tanaka and I. Nishioka, J. Chem. Soc. Perkin I, (1982) 1067-1073. G. Nonaka, T. Tanaka, M. Nita and I. Nishioka, Chem. Pharm. Bull., 30 (1982) 2255-2257. T Tanaka, G. Nonaka and I. Nishioka, Chem. Pharm. Bull., 32 (1984) 117-121. T. Tanaka, G. Nonaka and I. Nishioka, J. Chem. Res (M), (1985) 2001-2029. T. Murashige and R Skoog, Physiol. Plant., 15 (1962) 473-497. W. Mayer and G. Bauni, Liebigs Ann. Chem., 611 (1958) 264-268. R. S. Thompson, D. Jacques, E. Haslam and R. J. N. Tanner, J. Chem. Soc. Perkin Trans I (1972) 1387-1399. E. A. Haddock, R. K. Gupta, M. K. A-S. Sabah, E. Haslam and D. Magnolato, J. Chem. Soc. Perkin Trans I (1982) 2515-2524. Y Kashiwada, G. Nonaka and I. Nishioka, Chem. Pharm. Bull., 32 (1984) 3461-3470. O. L. Gamborg, R. A. Miller and K. Ojima, Exp. Cell Res., 50 (1968) 151-158. G. B. Lloyd and B. H. McCown, Int. Plant. Prop. Soc., 30 (1980) 421-427. In EMBO Course The Use of Ti Plasmid as Cloning Vector for Genetic Engineering in Plants, 4-23 August (1982) p. 109. H. Kamada, H. Ohkawa, H. Harada and K. Shimomura, in Proc. of the Plant Growth Reg. Soc. of America, 2-6 August (1987) p. 227. H. Ohkawa, H. Kamada, H. Sudo and H. Harada, J. Plant Physiol., 134 (1989) 633-636.

448 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

G. Vervliet, M. Holsters, H. Teuchy, M. Van Montagu and J. Schell, J. Gen. Virol, 26 (1975) 33-48. S. Stachel, E. E. Messens, M. Van Montagu and P. Zambryski, Nature, 318 (1985) 624-629. A. Petit, C. David, G. A. Dahl, J. G. Ellis, P. Guyon, R Casse-Delbart and J. Tempe, Mol. Gen. Genet., 190 (1983) 204-214. T. Okuda, T. Yoshida andT. Hatano, J. Chem. Soc. Perkin Trans I, (1982) 9-14. T. Tanaka, G. Nonaka and I. Nishioka, Phytochemistry, 24 (1985) 2075-2078, T. Tanaka, G. Nonaka, I. Nishioka, K. Miyahara and T. Kawasaki, J. Chem. Soc. Perkin Trans I, (1986) 369-376. G. Nonaka, S. Morimoto and I. Nishioka, Chem. Pharm. Bull., 34 (1986) 941-943. H. Deno and S. Uhara, Japan Patent, 62-6675 (1987) p. 347. K. Yazaki, H. Fukui, M. Kikuma and M. Tabata, Plant Cell Report, 6 (1987) 131-134. P. Morris and M. P Robbins, J. Exp. Bot., 43 (1992) 221-231. H. Hikino, Y. Kiso, M. Kubota, M. Hattori and T. Namba, Shoyakugaku Zasshi, 38 (1984) 359-360. H. Inouye and Y. Nakamura, Tetrahedron Letters, 47 (1968) 4919-4924. T. Kubota and Y. Tomita, Tetrahedron Letters, 5 (1961) 176-182. H. Inouye, T. Yoshida, Y. Nakamura and S. Tomita, Tetrahedron Letters, 42 (1968) 44294432. H. Inouye, S. Ueda and Y. Nakamura, Chem. Pharm. Bull., 18 (1970) 1856-1865. Y Ikeshiro and Y. Tomita, Planta medica, 48 (1983) 169-173. H. Inouye and Y Nakamura, Tetrahedron, 27 (1971) 1951-1966. Y Ikeshiro, T. Kubota and Y Tomita, Planta medica, 47 (1983) 26-29. M. Komatsu, T. Tomimori, Y. Makiguchi and K. Asano, Yakugaku Zasshi, 88 (1968) 832837. T. Tomimori and M. Komatsu, Yakugaku Zasshi, 89 (1969) 410-417. M. Komatsu, T. Tomimori and N. Mikuriya, Chem. Pharm. Bull., 17 (1969) 155-162. I. Sakamoto, T. Tanaka, O. Tanaka andT. Tomimori, Chem. Pharm.Bull., 30 (1982) 40884091. Y. Kitamura, M. Dono, H. Miura and M. Sugii, Chem. Pharm. Bull., 36 (1988) 1575-1576. H. Miura, K. Kawashima, Y. Kitamura and M. Sugii, Shoyakugaku Zasshi, 40 (1978) 4043. H. Miura, M. Ida, Y Kitamura and M. Sugii, Shoyakugaku Zasshi, 32 (1978) 90-95. J. B. Harborne and J. J. Corner, Biochem. J., 81 (1961) 242-250. K. Herrmann, Prog. Chem. Org. Nat. Prod., 35 (1978) 73-132. R. Saijo, G. Nonaka and I. Nishioka, Phytochemistry, 28 (1989) 2443-2446. J. Inokuchi, H. Okabe, T. Yamauchi, A. Nagamatsu, G. Nonaka and I. Nishioka, Chem. Pharm. Bull., 33 (1985) 264-269. T. Furuya, S. Ayabe and K. Noda, Phytochemistry, 14 (1975) 1457. Y Ohshima, K. Takahashi and S. Shibata, Planta Medica, 54 (1988) 20-24. Y Kashiwada, G. Nonaka and I. Nishioka, Chem. Pharm. Bull., 7 (1989) 999-1004. G. E. Schneiders and R. Stevenson, J. Chem. Soc. Perkin Trans I, (1982) 999-1003. B. Singh, P K. Agrawal and R. S. Thakur, J. Nat. Prod., 52 (1989) 48-51. P. Satyanarayana, P. Subrahmanyam, K. N. Viswanatham and R. S. Ward, J. Nat. Prod., 51 (1988) 44-49. M. S. Tempesta, D. G. Corley, J. A. Beutler, C. J. Metral, R. A. Yunes, C. A. Giacomozzi and J. B, Calixto, J. Nat. Prod., 51 (1988) 617-618. B. S. Joshi, D. H. Gawad, S. W. Pelletier, G. Kartha and K. Bhandary, J. Nat. Prod., 49 (1986) 614-620. N. B. Mulchandani and S. A. Hassarajani, Planta Medica, 50 (1984) 104-105. O. Hnatyszyn, G. Ferraro and J. D. Coussio, J. Nat. Prod., 50 (1987) 1156-1157.

449 86

H. Ueno, S. Horie, Y. Nishi, H. Shogawa, M. Kawasaki, S. Suzuki, T. Hayashi, M. Arisawa, M. Shimizu, M. Yoshizaki, N. Morita, L. H. Berganza, E. Ferro and I. Basualdo, J. Nat. Prod., 51 (1988) 357-359. 87 B. Singh, P. K. Agrawal and R. S. Thakur, Indian J. Chem. 25B (1986) 600-602. 88 S. P. Thyagarajan, K. Thiruneelakantan, S. Subramanian andT. Sundaravelu, Indian J. Med. Res., 76 (1982) 124-130. 89 K. V. Syamasundar, B. Singh, R. S. Thakur, A. Husain, Y. Kiso and H. Hikino, J. Ethnopharmac, 14 (1985) 41-44. 90 R. A. Jefferson, T. A. Kavanagh and M. W. Bevan, EMBO J., 6 (1987) 3901-3907. 91 O. T. Schmidt and R. Eckert, liebigs. Ann. Chem., 618 (1958) 71-82. 92 R. Saijo, G. Nonaka and I. Nishioka, Chem. Pharm. Bull., 37 (1989) 2063-2070. 93 O. T. Schmidt, J. Schulz and R. Wurmb, Liebigs Ann. Chem., 706 (1967) 169-179. 94 T. Yoshida, H. Itoh, S. Matsunaga, R. Tanaka and T. Okuda, Chem. Pharm. Bull., 40 (1992) 53-60. 95 L. Y Foo and H. Wong, Phytochemistry, 31 (1992) 711-713. 96 J. Leemans, R. Deblaere, L. Willmitzer, H. De Greve, J. P. Hemalsteens, M. Van Montagu and J. Schell, EMBO J., 1 (1982) 147-152. 97 G. Nonaka, O. Kawahara and I. Nishioka, Chem. Pharm. Bull., 31 (1983) 3906-3914. 98 Y-J. Ahn, T. Kawamura, M. Kim, T. Yamamoto and T. Mitsuoka, Agric. Biol. Chem., 55 (1991) 1425-1426. 99 T. Yokozawa, H. Oura, S. Sakanaka and M. Kim, Biosci. Biotech. Biochem., 56 (1992) 896-899. 100 T. Hashimoto, Nippon Nogeikagaku Kaishi, 66 (1992) 1-8.

This Page Intentionally Left Blank

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol 17 © 1995 Elsevier Science B.V. All rights reserved.

451

Chemistry and Biosynthesis of Natural Diels-Alder Type Adducts from Moraceous Plants Taro Nomura, Yoshio Hano and Shinichi Ueda

Introducliion [4 + 2]Cycloaddition, well known as the Diels-Alder reaction^ is a widely-used reaction in the synthesis of organic compounds including natural products. On the other hand, many possible biosynthetic Diels-Alder constructions from natural sources have been reported [1 - 25]. Such constructions in natural are mostly stereo- and regiospecific, suggesting the pivotal step of the cycloaddition reaction in the biosynthesis to be enzymatic. However, the evidence for a biological Diels-Alder reaction is rare. Mulberry tree is a rich source of intermolecular Diels-Alder type adducts comprising two molecules of isoprenylphenols [26]. This article describes the chemistry of the Diels-Alder type adducts from moraceous plants and the biosynthesis of the adducts in Worus alha. cell cultures. 1. Mulberry Diels-Alder type adducts Mulberry tree, a moraceous plant of the genus Morus, has been widely cultivated in Japan and China in order to serve its leaves for sillcworms. On the other hand, the root bar]c of the tree has been used as a Chinese crude drug so called "Sang-Bai-Pi", for an antiphlogistic, a diuretic, and an expectorant. A few pharmacological studies on the mulberry tree demonstrated a hypotensive effect of the extract in rodents. Considering these reports, it was suggested the active principle to be a mixture of many phenolic components. Kuwanon G (1) is the first isolation of the active substance exhibiting the hypotensive action from the Japanese Morus root baric. Kuwanon G (1) [27], molecular formula C40H36OH, was assumed to be a biogenetic intermolecular Diels-Alder type construction formed from dehydrolcuwanon C (2) and a chalcone derivative (3) based on the NMR spectroscopic studies (Figure 1). An evidence of the

452

HO

Figure 1

250**C,

Z 0^-

Figure 2 regiochemistry for 1 was obtained by the following results: DielsAlder

reaction

of

trans-chalcone

and

2-methyl-3-phenyl-l,3-buta-

diene gave two cycloproducts, one of which is the all-trans type adduct

(4) in relative configuration among three substituents on

the methylcyclohexene ring and another is the cis-trans (5) in relative configuration

(Figure 2 ) .

type adduct

The X-ray analyses of

these cycloproducts revealed that the regioselectivity in the [4 + 2]

reaction

was

kuwanon G (1).

coincided

with

that

in

the

A pyrolysis of kuwanon G octamethyl

afforded tra/3s-2,2',4^4'-methoxychalcone tetramethyl

estimated

ether

(2a)

(Figure

3).

case

ether

of

(la)

(6) and dehydro-kuwanon C Subsequently,

the

two

fragmentation products 6 and 2a gave, when heated in toluene at 160 ®C for 61 h in a sealed tube, two [4 + 2]cycloadducts

(Figure 4 ) .

One of the adducts was identified with (±)-la and the other was a

453

OCH, HjCO

HjCO

H3CO

OCH3

HjCO^^^v^^OCH, H3CO

l a

CH,0

O

6 Figure 3

Pyrolysis of kuwanon G octamethyl ether (la)

HsCO

H3CO,

OCHs OCH9 H3CO

HsCO^^^^Ss^OCHj HjCO. H3CO.

OCH3

Figure 4

Diets-Aider reaction of the pyrolysis products 2a and 6

454

.'•vvi-° X OH O

10: 11:

12: 3"-Hp 1 3 : 3".Ha

3".Ha 3"-Hp

14: 3".Ha 15: 3"-Hp

F i g u r e 5 Typical Diels-Alder type adducts from Moms root bark cis-trans type cycloadduct (7a) as was observed in the case of the synthesis of the model compounds described above. The structure of kuwanon G (1) has thus been established as depicted in Figure 1 [28]. Kuwanon G (1) is optically active ([a]D - 534°) and considered to be formed through an enzymatic Diels-Alder reaction of a chalcone derivative as a dienophile and dehydrokuwanon C as a diene (Figure 1 ) . About forty kinds of optically active Diels-Alder type adducts, which are presumably formed from a dehydroprenylphenol and a chalcone derivative, have been isolated from Japanese Morns root bark and Chinese crude drug ''Sang-Bai-Pi" [26]. The mulberry Diels-Alder type adducts may be divided into the following four types on the basis of the phenol nuclei; a) adducts of a chalcone and a dehydroprenylflavonoid {e.g. (1), kuwanon H (8) [29] and sanggenon C (9) [30]}, b) adducts of a chalcone and a dehydroprenylchalcone {e.g. kuwanons I (10) [31] and J (11) [32]}, c) adducts of a chalcone and a dehydroprenyl-2-arylbenzofuran {e.g. mulberrofurans C (12) [33] and J (13) [34]}, and d) adducts of a chalcone and a dehydroprenylstilbenes {e.g. kuwanons X (14) [34] and Y (15) [35]} (Figure 5 ) . It is very interesting that a pair of

455 isomers, such as 10 and 11 and others, coexist in the Morus root bark as in the case of the Diels-Alder reaction of a trans-chalcone and a butadiene derivative giving rise to a pair of isomers, alltrans and cis-trans type adducts. On the other hand, callus tissues of Morus alba L. induced from the seedlings or the leaves were cultured under specified conditions and subjected to selection over nine years, giving rise to cell lines producing characteristic Diels-Alder type adducts at high levels [32]. Eight Diels-Alder type adducts, kuwanons J (11) [32, 36], Q (16) [36], R (17) [36], V (18) [36], mulberrofurans E (19) [37], T (20) [38], chalcomoracin (21) [32, 39], and kuwanol E (22) [38], have been isolated from the callus tissues along with morachalcones A (23) [36, 39], B (24) [36], and moracin C (25) [36, 40] (Figure 6 ) . Assuming that compounds (23), (24), and (25) are a dienophile or an equivalent of a diene, kuwanon J (11) and kuwanon V (18) are composed of two molecules of 23 and 24, respectively. While, both kuwanons Q (16) and R (17) are formed from 23 and 24. Similarly, chalcomoracin (21) is composed of 23 and 25 and mulberrofuran E (19) is formed from 24 and 25. It is noteworthy that all combinations of these monomers, 23, 24, and 25, could be isolated from the M. alba callus tissues.

OH

11: RisR2=OH 16: Ri^sOH Ri^H 17: R,sH R2=^OH 18: Rt^RjssH

19 ! RirrRjsH/ 2 0 : R i = - - l ^ R2=OH 2 1 : RisH R2-OH

F i g u r e 6 Phenolic components of Morus alba cell cultures

22

456 2. Other optically active Diels-Alder type adducts from moraceous plants The compounds structurally similar to the mulberry Diels-Alder type adduct have also been isolated from the other species of the plants of the family Moraceae. Brosimones A (26) [41], B (27) [14] and D (28) [14] have been isolated from Brosimopsis oblongifolia, a Brazilian moraceous plant (Figure 7). They are expected to be formed through the saune way as the mulberry Diels-Alder type adducts. Of these compounds, compound (26) is a unique adduct which may be formed through an intramolecular [ 4 + 2 ] cycloaddition reaction of the isoprenyl portion and the a, p-double bond of the chalcone slceleton of compound (28) (Figure 8). Artonins C (29) [42], D (30) [42] and I (31) [43] isolated from Artocarpus heterophyllus, an Indonesian moraceous plant, can also be regarded

OH

HO

OH

26

27

28

F i g u r e 7 Optically active Diels-Alder type adducts from Brosimopsis oblongifolia

HO-%^

OH

28 Figure 8

26 A hypothesis of the formation of 26 from 28 through Diels-Alder reaction

457

30:R=

F i g u r e 9 Diels-Aider type adducts from Indonesian and Argentine moraceous plants

ny

I

MO

OH

TO'

HO^...^s^Oj,0^...;s^CHO

HO^...'fe^O

37

Figure 10

Ketalized compounds from moraceous plants

MO

458 as typical Diels-Alder type adducts of a chalcone derivative and a dehydroprenylphenol (Figure 9 ) . Mulberrofuran U (32) [44] has been isolated from an Argentine moraceous plant Morns

(Figure

insignis

9). Ketalized Diels-Alder type adducts^ mulberrofurans F (33) [45]^ G

(34)

[45], K

(35)

[46] and

Icuwanol A

(36)

isolated from the Moras root barJc (Figure 10).

[47], have

(34) are presumably formed through an intramolecular reaction

of

chalcomoracin

respectively.

(21)

Indeed, compounds

stereospecifically

from

the

and

original

)cetalization

mulberrofuran

(33) and

C

(37) [20], soroceins A to

the

above

isolated from Sorocea ure 10).

(38) [20] and B described

bonplandii,

(12),

(34) could be derived adducts

21

and

respectively, under acidic conditions (Figure 11) [45]. similar

been

Compounds (33) and

(39)

Jcetalized

12,

Soroceal

[20], structurally

compounds,

have

been

a Brazilian moraceous plant (Fig-

These compounds are also considered to be modifications

of original adduct through an intramolecular Icetalization reaction.

H^ -H2O

12: R=H 21:R= Figure 11

40

Figure 12

34: R=H

y

y

33:R=:

.^^

The formation of the ketalized compounds from the original adducts under acidic conditions

D*

Colouration mechanism of40 under acidic conditions

459 Mulberrofuran I (40) [48] has been isolated as a red pigment from the root bark of the mulberry tree. The compound (40) colours red under acidic conditions* The ^H-NMR spectrum of 40 in the presence of deutelized trifluoroacetic acid revealed that only the proton at C-2" position was exchangeable for deutelium. This result suggested that the colouration mechanism of 40 is depicted as in Figure 12 [26]. Compound (40) is considered to be one of the modifications derived from mulberrofuran C (12). As shown in Figure 13, the two ways of oxidative cyclizations of a hemilcetal intermediate derived from 12 may afford Icetalized compound (34) and the red pigment (40), respectively. Aromatized compound albanol B (41) [49], presumably derived from 34 through a dehydrogenation, has been isolated from Indian Morns alba L. However, this compound is not optically active, indicating that the sole chriral center at C-8" position is racemic [50]. Albanol B pentamethyl ether (41a) formed from mulberrofuran G pentamethyl ether (34a) through the dehydrogenation reaction with DDQ is optically active ([ctJo + US'*) [50]. This fact indicated albanol B (41) to be an artefact formed during the isolation procedures. As a result, it was concluded that compound (41) is derived from mulberrofuran I (40), a red pigment, through an auto-oxidation in the presence of an acid or Lewis acid-li]ce catalyst such as Si02 (Figure 14) [50]. Aromatized

40

H2O, -H2

Figure 13 The formations of 34 and 40 through two ways (a and b) of oxidative cyclizations of a hemilcetal intermediate derived from 12

460 OCHg

DDQ

H3CO

H3CO

OCH3

OCH3

-2H,

34a

41a (optically active)

HO^^^s^v^OH

OH

40: R=H 43 : R=OH

41: R = H (racemate) 42: R = O H (racemate)

-H2O2

HO

Figure 14 compound

mulberrofuran

Japanese Morus that

compound

compound

corresponding

( 4 2 ) [51] has also

root bark as a blue pigment.

aromatized

pigment,

P

red

such

isolated

to

be

Indeed,

an

the

from

Considering the result

as 4 1 is an artefact

( 4 2 ) seems pigment.

been

from

artefact

a red

from

corresponding

the

compound

mulberrofuran S ( 4 3 ) [50] has been isolated from the same source as a

red pigment

modifications

(Figure

14).

biosynthetically

Many

other

derived

Alder type adduct such as mulberrofuran from Japanese Morus

root bark [ 2 6 ] .

substances

from

regarded

the original

C ( 1 2 ) have been

as

Dielsisolated

461 3. Absolute Configuration of Mulberry Diels-Alder Type Adducts As described above^ the stereochemistries of the mulberry DielsAlder type adducts could be divided into the following two groups; one is an all-trans in relative configuration and the other is a cis-^trans configuration. All-trans type adduct may correspond to an exo-addition product in the Diels-Alder reaction of a chalcone and a dehydroprenylphenol, whereas a cis-trans type adduct corresponds to an endo-addition product in the reaction. Absolute stereochemistries of the mulberry Diels-Alder type adducts were confirmed by the circular dichroism (CD) spectroscopic evidence and by X-ray analysis [26, 52, 53].

13a: R= <

12a: R= < Ae

Ae

+20-

+5—

• 1 •? 1 • ±z j

\

: 12a

+101

\

/

'*

/ /*

0"

k,-/

0

r^ i \ -10'

1

1 t

-20-^

-5-

r r——1 r 250 300 350 400 nm F i g u r e 15

I

,

,

,

: 13a

,

250 300 350 400 nm

CD spectra of 12,13,12a, and 13a

1

462

J

Both CD spectra of a pair of isomers, mulberrofurans C ( 1 2 ) and ( 1 3 ) , showed a large magnitude of Ae values more than those of

the other adducts [ 5 3 ] . reduced compounds

On the other hand, the CD spectra of the

( 1 2 a ) and ( 1 3 a ) , which

from 1 2 and

are formed

1 3 , respectively, by LiAlH4 reduction, are mirror

images of each

other in the Jt - Ji* region due to the 2-arylbenzofuran and

13

at the C - 3 " chiral

chromophore

chromophore

This indicates that the stereochemistries of 12

(Figure 15) [ 5 3 ] .

are antipodal

center

to each

mulberrofuran

G

( 3 4 ) could

be

conditions.

Monobromomulberrofuran

bearing other.

derived G

the

2-arylbenzofuran

As from

descrived

above,

12

acidic

under

pentamethyl

ether

prepared from 3 4 a by the treatment with NBS w a s converted aromatized 16).

compound

( 3 4 c ) through dehydrogenation

by D D Q

(34b) to an (Figure

T h e X-ray crystallographic analysis of 3 4 c revealed that the

absolute configuration of the sole chiral center at C-8" is R [ 5 3 ] . OCH3

OCHa

H3CO

OCHa

H3CO

NBS 34b DDQ R=Br — ^

34a: R = H

OCH3

g^^y;. 0CH3 34c (8"./?)

Figure 16

OH 12:3"S,4"/?,5"5

Figure 1 7

OH 13 : 3"/?, 4"/?, 5"S

Absolute stereochemistries of 12 and 13

463

H3CO, HaCO

i)OH OH

OH O 44 (3"1?, 4"i?, 5"S)

ii)(CH3)2S02

iy)P'BtBzC\

45: R=H 45a: R^CHa

46: RisrRjsH 46a: Ri^H Rjsp-BrBz 46b: Ri=R2sp-BrBz

Figure 18 As the relative configuration of the four chiral centers in 34 has been confirmed by the X-ray analysis of its pentamethyl ether (34a) [49], the absolute configuration of 34 was specified as 3"S, 4"K, 5"S, B"R [53]. Hence, the absolute stereochemistry of 12 was also determined to be 3"S, 4",JR, 5"S [53]. Since the stereochemistry of 13 at the C-3" position is antipodal to that of 12, the absolute configuration of 13 was expressed as 3"R, A"R, 5"5 [53]. The absolute stereochemistries of a pair of isomers, all-trans type and ciS'-trans type adducts, has thus been determined as shown in Figure 17. On the other hand, the absolute configuration of kuwanon L (44) [54], an all-trans type adduct, was confirmed by the following result: Alkali degradation followed by methylation of 44 gave its pentcunethyl ether (45a) by way of 45. Treatment of 45a with OSO4 gave a cis-diol product (46), which was converted to the mono-pbromobenzoate (46a) and the di-p-bromobenzoate (46b) (Figure 18). The CD spectrum of 46b exhibited a positive Cotton effect owing to a typical exciton coupling between the two p-bromozenzoyl chromophores [52]. Thus the absolute configuration of the three chiral centers of the methylcyclohexene ring of 44 has been specified as 3"R, 4t"R, 5"S [52]. This stereochemistry was in accordance with that of an all-trans type adduct, mulberrofuran J (13). Optical rotations ([a]p) of typical mulberry Diels-Alder type adducts are summarized in Table 1. The all-trans type adducts exhibit negative optical rotation, while the cis-trans type adducts exhibit positive values. Considering the absolute stereochemi-

464 Table 1 Optical rotation values ([a]^) all-trans type adducts mulberrofuran J (13) -341® kuwanon X (14) -322* kuwanon I (10) -454* kuwanon G (1) -534* kuwanon H (8) -536* kuwanon L (44) -277*

cis-trans

type adducts

mulberrofuran C (12) kuwanon Y (15) kuwanon J (11) kuwanon Q (16) kuwanon R (17) kuwanon V (18)

+153** +172** +85** +160** + 56** +145**

HO

T

11 4" 13"

OH 0

^jto-addition

^OH

ksAs^

OH

OH O

chalcone

* 7--

aW'trans 3"/?, 4"i?, 5"5

M

dehydroprenylphenol

HO

L C -> 0

e/i^o-addition Figure 19

p M

OH 0

1 M

cis-trans 3"5,4"/?,5"5

Absolute stereochemistries of the mulberry Diels-Alder type adducts

stries of a pair of isomers, mulberrofurans C (12) and J (13), the chirality at the C-3" position influences the sign of the rotation. Namely, the absolute configuration of the all-tra/is type adducts is the same as that of 13, while that of the cis-trans type adducts is the same as that of 12. The absolute stereochemistries of three chiral centers in the methylcyclohexene ring of the all-trans type

465 adducts have thus been determined as Z^'R, 4"!?, 5"S, whereas those of the cis-trans type adducts as 3"S, ^"R, 5"S (Figure 19) [53]. 4. Biosyniihesis of the mulberry Dlels-Alder type adducts 4.1. Biosynthesis of kuwanon J (11) and chalcomoracin (21) As described in section 2, some cell strains of Morus alba callus tissues induced from the seedlings or the leaves have a high productivity of the mulberry Diels-Alder type adducts. The yields of major adducts chalcomoracin (21) and kuwanon J (11) by the cell strains are about 100 - 1000 times more than those of the intact plant [32]. The biosynthesis of the mulberry Diels-Alder type adducts has been studied with the aid of the excellent cell strains. Administration of [1-^^C]-^ [2-l3c]-, or [1^2-l3c2]acetates to the Morus alba cell cultures revealed the early stages of biosynthesis of 11 and 21^ in which both adducts are originated from two molecules of cinnamoylpolylcetide skeletons (Figure 20) [55]. The chalcone skeleton of both adducts 11 and 21 has been considered to be formed through deoxygenation at C-5 of the cinnamoylpolyketide skeleton [56], followed by the Claisen-type condensation at C-4 and C-9 of the skeleton (Figure 21). On the other hand, the 2-arylbenzofuran skeleton of 21 is formed through an aldol-type condensation at C-3 and C-8 of the cinnamoylpolyketide skeleton, followed by decarboxylation (Figure 21). The other interesting result was obtained with respect to the biosynthesis of the isoprenyl units of 21. Administration of i^clal^elled acetate to the cell cultures resulted in the highly ^^Cenriched aromatic carl)ons of 21 (about 18 % ) , while two isoprenyl units of 21 were lal^elled to a lesser extent (about 0.4 % ) . In addition, on the basis of 13^.13^ spin-spin coupling, the labelling of [ 2-l^c]acetate takes place in the contiguous carbons at the starter acetate unit with regard to the mevalonate biosynthesis (Figure 20b). On the contrary, [l-^^C]acetate was not incorporated into the isoprenyl units of 21 [55], unlike the incorporation into the cinnamoylpolyketide skeletons. These findings suggest the participation of tricarboxylic acid (TCA) cycle to the biosynthesis of the isoprenyl units of 21. In the experiment of [2-i3c]acetate, the contiguous ^^C atoms can be derived from the two methyl groups of the intact acetates administered by way of at least two passages through the TCA cycle, while in the case of [l-^^C]acetate the i^C atom was removed as carbon dioxide. Accordingly, the acetate

466

?a HO

^A^OH HO 11 (a)

OH

I

• •

21

CH3COSC0A

(b)

F i g u r e 2 0 ^^C-Labelling patterns of 11 and 21 from "C-labelied acetates

• • CHjCOGNa

C2X3

"*•

CoAS^ O OH

x-'^^v^OH „ SCoA SCoA ,X5^^'

CoAS

OH chalcone skeleton Figure 21

2-arylbenzofuran skeleton

Biosyntheses of the chalcone and 2-arylbenzofuran skeletons in Moras alba cell cultures

467

CH3COSC0A •/2 HOOC^^O

•12 HOOC^^O

HOOC'V''°°"

•^^COOH •/2 •/2 CO2

•^^COOH •/2

"^COOH

/ • ^

^COOH

HO

^COOH

1:

•/2 HOOC^^OH

ji/2

T C A cycle

h

m/2

HOOC^ ^OH

Tr •/2

•/2 •/2 HOOC HOOC^

r

•/2 HOOC

O^COOH

CO2

COOH •/2

/

«\1

COOH •/2

•/2 ^ HOOC^^O •/2CH3

h

•/2 CO2

• / 2 1/2 CH3COSC0A

Figure 22 incorporated into the isoprenyl units of 21 was not the intact acetate administered, but [1,2-^^02]acetate reorganized from the methyl group of the intact acetate through the TCA cycle (Figure 22) [55]. This hypothesis was reinforced by the administration experiment with (2-1^0]acetate in a pulsed manner (three times, every 12 h) [57]. In this case, however, the ^^c-labelling at the isoprenyl units of the resulting 21 was observed not only at the starter acetate carbons, but also at the third acetate carbons at C-7" and C-25", which showed i^c-^^c coupling with the adjoining carbons at C-1" and C-23", respectively (Figure 23). In spite of the laclc of ^^C-laljelling at the second acetate, the third acetate carbon was apparently enriched here with ^^C. This fact can be explained in terms of the isomerization of the two 3,3dimethylallyl groups via 3-methyl-l,3-butadienyl groups (Figure 23). Namely, the i^C-labelling at the third acetate carbon can be attributable to the transfer of the i^C-labelling from the cismethyl carbon to the trans-methyl carbon involving an isomerization of a cisoid and a transoid of a diene (Figure 23). This finding

468

11^1

fi—iT^

Ho^ jii^ JL JL.<^ 7- ^^ 2"

HO / \

Y

\J

-K

3^^^^,X*s.jS!^

X^o°" JIL,OH

1

HO

I T 22" 23T 21" 1 ^ OH • 24"

21

Figure 23

,-Hr

V

^ ' ^

^^C-Labelling pattern of 21 by pulse-administration experiment with [2."CJacetete

gave a confirmative evidence on the formation of the diene structure at the isoprenyl portion for the Diels-Alder type cycloaddition reaction. It is interesting that the isomerization takes place not only at the isoprenyl unit participating in the Diels-Alder type cycloaddition reaction, but also at the other isoprenyl unit, which did not participate in the reaction (Figure 23). Thus the administration experiment with i^c.^^belled acetate to the M. alba cell cultures revealed that the Diels-Alder type adducts 11 and 21 are presumably biosynthesized through the [4 -«2]cycloaddition reaction between two molecules each of cinnamoylpolyketide-derived skeleton and mevalonate. Final confirmation of the biosynthesis of the mulberry Diels-Alder type adducts was performed by the administration experiment with Omethylchalcone derivatives, as modified precursors, to the M. alba cell cultures [58]. 0-Methylated chalcones or 0-methylated DielsAlder type adducts have not been detected in the cell cultures. Administration of 0-methylated chalcone (47) to the cell cultures yielded the metabolites (48), (49), (50), (51), and (52) (Figure 24) [58]. The conversion of the precursory chalcone (47) into (48) indicated that isoprenylation takes place after aromatization of the chalcone skeleton derived through the Claisen-type condensation of the cinnamoylpolyketide. The metabolites (49), (50), (51), and (52) revealed that the precursory chalcone (47) was incorporated

469 OCHa

YV^o^^^c

HaCO

50 : R1-CH3 R2=H R3SOH 5 1 : Ri=CHj R2=R3=H 5 2 : Ri=:R2=CH3 RjsOH

Figure 24

H3C0

HaCq

HaCO

Figure 25

470 intact into the Diels-Alder type adducts, as a diene or a dienophile. An analogous experiment with synthesized (48), one of the metabolites from (47) in the cell cultures, yielded the same Diels-Alder type metabolites (49), (50), (51), and (52), as in the case of the experiment employing (47) (Figure 24) [58]. Subsequently, administration of tri-0-methylated chalcone (53) to the cell cultures afforded the Diels-Alder type metabolite (54) (Figure 25) [58], in which the precursory chalcone (53) was incorporated as a dienophile. These results strongly indicate that dehydrogenation at the isoprenyl portion of (48) followed by a [4 + 2]cycloaddition reaction with the a, P-double bond of another molecule of isoprenylchalcone leads to the formation of the DielsAlder type metabolites. Furthermore, the Diels-Alder type metabolites from the precursory chalcones (47), (48), and (53) are all optically active, having the same stereochemistries as those of kuwanon J (11) and chalcomoracin (21). This fact revealed the [4 + 2]cycloaddition step to be enzymatic. Administration of 0methylated precursory chalcone to the M. alba cell cultures has thus demonstrated that optically active mulberry Diels-Alder type adducts such as 11 and 21 are biosynthesized through an enzymatic intermolecular [4 + 2]cycloaddition reaction. Another interesting result concerning the enzyme system of the M, alba cell cultures was obtained at the structure determination experiment of artonin I (31) [43], which has been isolated from an Indonesian moraceous plant Artocarpus heterophyllus, as described in section 2. Artonin I (31) was regarded as a typical Diels-Alder type adduct formed from a chalcone and a flavone, artocarpesin (55)

[^"""1 kOHJ 11

HO^^jTs^O^ "N^ HO

y

h=\ HO

JLoH OH

'

artonin I (31)

Figure 26

Bioconversion of 55 to 31 using Morus alba cell cultures

471 which co-occurs with 3 1 . Administration of 55 into the M» alba cell cultures yielded the metabolite (31) in high yield. The structure of artonin I has thus been established as formula 3 1 . This finding also corroborates the biosynthetic pathway simultaneously (Figure 26) [43]. This is the first example of determining the structure of a target substance with the aid of an enzyme system of cell cultures of other plant.

4.2. Shlkimate pathway and Isoprenoid biosynthesis in Morus alba cell cultures As described above, administration experiment with i^C-labelled acetates revealed that both the mulberry chalcone and 2arylbenzofuran skeletons originate from cinnamoylpolyketide. In order to confirm the biosynthesis of the cinnamoyl moiety derived from shikimate via aromatic amino acid, phenylalanine or tyrosine, further experiments administering phenylalanine and tyrosine to the M. alba cell cultures were carried out [59]. In the administration experiment with L-[3-i3C]phenylalanine to the cell cultures, a high i^C-enrichment (about 33 %) from the amino acid was observed at the

11

21

F i g u r e 2 7 '^C-Labeling from L-[3-"C]phenylalanine and L-[3->^C]tyrosine

472 C-3 and C-5" positions of chalcomoracin (21) as well as at the C-P and C-5" positions of kuwanon J (11) (Figure 27). This finding indicates that both the cinnamoyl moieties in the chalcone and 2arylbenzofuran moieties originate from L-phenylalanine. An analogous experiment with L-[ 3-^^C]tyrosine also showed the same ^^C-labelling pattern as in the case of L-[3-^3c]phenylalanine (Figure 27). The ^^C-incorporation was about 18 %. Both Lphenylalanine and L-tyrosine, intermediates on the shikimate pathway, are thus precursory to the mulberry chalcone and 2arylbenzofuran skeletons. Transformation of L-phenylalanine into L-tyrosine by direct hydroxylation has been well established in mammal cells [60]. An enzyme system isolated from spinach leaves also caused the transformation [61]. In higher plants, however, direct conversion of L-phenylalanine into L-tyrosine has been rare [62]. It has also been reported that the overlapping of the secondary metabolites from L-phenylalanine with those of L-tyrosine is restricted, because of the presence of two independent metabolic pathways for these aromatic amino acids. L-Phenylalanine has been known to be converted to traus-cinnamate, which, in turn, is subjected to 4-hydroxylation resulting in the formation of pcoumarate through the action of cinnamate 4-hydroxylase in higher plant [64]. In Morus alba cell cultures, it was suggested that both

O

NH2

r^COOH

6

X phenyl pyruvate

COOH

r^COOH

6

L-phenylalanine

cinnamate

prephenate

[O]

\

COOH

NH2

u COOH

OH

4-OH-phenylpyruvate

I^COOH

Q OH

L-tyrosine

•

OH

/7-coumarate

F i g u r e 2 8 The shikimate pathway leading to p-coumarate from L-phenylalanine and L-tyrosine

473 L-phenylalanine and L-tyrosine are converted into p-coumarate, which are conclusively incorporated into the chalcone and 2arylbenzofuran skeletons (Figure 28) [59]. On the other hand, in the experiment with ^^c.^abelled acetates, the acetate incorporated into the isoprenyl units of chalcomoracin (21) was reconstructed acetate from the methyl group of exogenous acetate through the TCA cycle, as described in section 4.I. On the basis of this novel finding, further studies with respect to the biosynthesis of the isoprenyl unit of 21 were carried out by administering dI-[ 2-1^0 Jmevalonate or L-t2-^^C]leucine, the candidates for isoprenoid precursor, to the M» alba cell cultures [57], In both experiments, no ^^C-incorporation was observed at the isoprenyl carbons of 21. In the case of L-[2-^3c]leucine, however, polyketide-derived aromatic carbons were enriched with ^^C in the same labelling pattern as in the case of [l-^^c] acetate with different degrees of ^^C-enrichment. This result indicated that L[2-1^0]leucine was catabolized to [l-^^cjacetyl CoA via 3-hydroxy-3-

K ^—2^1•

I ?

HjCpHO

^ \

COOH

H,0

. HMG-CoA

CO2

CH3COSC0A '^

CH3

/ TCA \ I cycle

)

CH3COSC0A

li 1rS V

rr°"

HO.

^^OH

OH OH 0

Figure 29

HO

Fate of L-leucine in Morus alba cell cultures

474 methylglutaryl CoA (HMG CoA) followed by the participation of the polyketidG biosynthesis (Figure 29) [57]. Such a fate for Lleucine has been reported in the case of sesquiterpene paniculide biosynthesis in Andrographis paniculata tissue cultures [64]. Morus alba cell cultures also yield P-sitosterol (56) [32], which is a good target for the examination of isoprenoid biosynthesis from isoprenyl precursors in the cell cultures. The ^^C-labelling of 56 in the experiments with [l-^^C]-, [2'^^C]-, or [1,2-^^02]acetates was in accordance with Ruziclca's biogenetic rule [57], as was verified in the case of 56 in Rabdosia japonica tissue cultures (Figure 30) [65]. Accordingly, the exogenous acetates were incorporated intact into the isoprenyl units of 56. dl-[2^3c]Mevalonate was also incorporated into the expected positions of 56 (Figure 31), in spite of non-incorporation of this precursor into the isoprenyl unit of 21 [57]. This result indicated that non-incorporation of the precursor into the isoprenyl units of 21 was not due to the impermeability across the cell walls. Thus the incorporation manner of the precursors, including acetate, into the isoprenyl units of 21 is different from that observed in 56. It is most lilcely that at least two independent isoprenoid biosynthetic pathways, that for sterols and that for isoprenylphenols, operate in the Morus alba cell cultures.

Figure 30

"C-LabcIIIng pattern of 56 from [1-*^C]- (•), [2-^^C]- (•), and [l,2-'X2]acetates {i

475

F i g u r e 3 1 ^^C-Labelling pattern of 56 fk*om [2-^^C]mevalonate

5. Conclusion Optically active Diels-Alder type adducts isolated from Morus sp. were found to be biosynthesized through an enzymatic intermolecular [4 + 2]cycloaddition reaction between an isoprenyl portion of an isoprenylphenol as the diene and an a, ^-double bond of a chalcone as the dienophile. This is the first example demonstrating a biological intermolecular Diels-Alder reaction. Although (-)-flavoskylin [2] produced by Penicillium sp., ageliferins [17] isolated from an Okinawan marine sponge, and the sesquiterpene-monoterpene adducts [18] isolated from Artemisia AerJba-alba have also been considered to be formed via biosynthetic intermolecular Diels-Alder reaction, the biological reaction has not yet been demonstrated. On the other hand, a few examples of biological intramolecular Diels-Alder reactions have been unambiguously found. Betaenone B and solanapyrone A, phytotoxins from Phoma tetae and Altenaria solani, respectively, were found to be formed through intramolecular Diels-Alder reaction [7, 12, 15]. In the biosynthetic studies of these phytotoxins, specific cytochrome P450 inhibitors have been used effectively for the identification of the putative precursors. (+)-Brevianamides A and B, mycotoxins from Penicillium brevicompactum, were also found to be biosynthesized through the pathway involving intramolecular Diels-Alder reaction, on the basis of experiments administering supposed precursors labelled with ^H or ^^C [66]. It is noteworthy that enzymes catalysing a Diels-Alder reaction actually occur in biological systems.

476 References 1. S. Ito and Y. Hirata, Tetrahedoron Lett., 1972, 2557 - 2560. 2. S. Seo, U. Sankawa, Y. Ogihara, Y. litaka, and S. Shibata, Tetrahedron, 29 (1973), 3721 - 3726. 3. W. Hofheinz and P. Schoenholzer, Helv. Chim. Acta., 60 (1977), 1367 - 1370. 4. A. C. Bazan and J. M. Edwards, Tetrahedron, 34 (1978), 3005 3015. 5. W. M. Bandaranayake and J. E. Banfield, J. Chem. Soc. Chem. Commun., 1980, 902 - 903. 6. D. E. Cane and C. Yang, J. Am. Chem. Soc, 106 (1984), 784 787. 7. H. Oikawa, A. Ichihara, and S. Sakamura, J. Chem. Soc. Chem. Commun., 1984, 814 - 815. 8. W. R. Roush, S. M. Peseckis, and A. E. Walts, J. Org. Chem., 49 (1984), 3429 - 3432. 9. R. N. Moore, G. Bigam, J. K. Chan, A. M. Hogg, T. T. Nakashima, and J. C. Vederas, J, TVm. Chem. Soc, 107 (1985), 3694 - 3701. 10. D. E. Cane and C. Yang, J. Antibiot., 38 (1985), 423 - 426. 11. A. Ichihara, M. Miki, H. Tazaki, and S. Sakamura, Tetrahedron Lett., 28 (1987), 1175 - 1178. 12. H. Oikawa, A. Ichihara, and S. Sakamura, J. Chem. Soc. Chem. Commun., 1988, 600 - 602. 13. R. M. Williams, E. Kwast, H. Coffman, and T. Glinka, J. Am. Chem. S o c , 111 (1989), 3064 - 3065. 14. I. Messana, F. Ferrari, J. Francisco de Mello, and M. do Carmo Mesquita de Aranjo, Heterocycles, 29 (1989), 683 - 690. 15. H. Oikawa, T. Yokota, T. Abe, A. Ichihara, S. Sakamura, Y. Yoshizawa, and J. Vederas, J. Chem. Soc. Chem. Commun., 1989, 1282 - 1283. 16. H. Oikawa, T. Yokota, A. Ichihara, and S. Sakamura, J. Chem. Soc. Chem. Commun.,1989, 1284 - 1285. 17. J. Kobayashi, M. Tsuda, T. Maruyama, H. Nakamura, Y. Ohizumi, M. Ishibashi, M. Iweunura, T. Ohta, and S. Nozoe, Tetrahedron, 46 (1990), 5579 - 5586. 18. J. A. Marco, J. F. Sanz, E. Falco, J. Jakupovic, and J. Lex., Tetrahedron, 46 (1990), 7941 - 7950. 19. C. Chen and D. J. Hart, J. Org. Chem., 55 (1990), 6236 - 6240. 20. I. Messana, F. Ferrari, F. Delle Monache, R. A. Yunes, J. B. Calixto, and T. Bisognin, Heterocycles, 32 (1991), 1287 - 1296. 21. H. Oikawa, Y. Murakami, and A. Ichihara, J. Chem. Soc. Perkin Trans 1., 1992, 2949 - 2954 22. H. Oikawa, Y. Murakami, and A. Ichihara, J. Chem. Soc. Perkin Trans 1., 1992, 2955 - 2959 23. I. C. Parsons, A. I. Gray, T. G. Hartley, and P. G. Waterman, Phytochemistry, 33 (1993), 479 - 482. 24. D. E. Cane, W. Tan, and W. R. Ott., J. Am. Chem. Soc, 115 (1993), 527 - 535. 25. W. A. Carroll and P. A. Grieco, J. Am. Chem. S o c , 115 (1993), 1164 - 1165. 26. T. Nomura, Fortschr. Che. Org. Naturst., 53 (1988), 87-201. 27. T. Nomura and T. Fukai, Chem. Pharm. Bull., 28 (1980), 2548 2552. 28. T. Nomura, T. Fukai, T. Narita, S. Terada, J. Uzawa, Y. litaka, M. Takasugi, S. Ishikawa, S. Nagao, and T. Masamune, Tetrahedron Lett., 22 (1981), 2195 - 2198. 29. T. Nomura and T. Fukai, Heterocycles, 14 (1980), 1943 - 1951. 30. T. Nomura, T. Fukai, Y. Hano, and J. Uzawa, Heterocycles, 16

477 (1981), 2141 - 2148. 31. T. Nomura, T. Fukai, J. Matsumoto, A. imashimizu, S. Terada, and M. Kama, Planta Med., 46 (1982), 167 - 174. 32. S. Ueda, T. Nomura, T. Fukai, and J. Matsumoto, Chem. Pharm. Bull., 30 (1982), 3042 - 3045. 33. T. Nomura, T. Fukai, J. Matsumoto, and T. Ohmori., Planta Med., 46 (1982), 28 - 32. 34. K. Hirakura, Y. Hano, T. Fukai, T. Nomura, J. Uzawa, and K. Fukushima, Chem Pharm. Bull., 33 (1985), 1088 - 1096. 35. Y. Hano, H. Tsubura, and T. Nomura, Heterocycles, 24 (1986), 2603 - 2610. 36. J. Ikuta, T. Fukai, T. Nomura, and S. Ueda, Chem. Pharm. Bull., 34 (1986), 2471 - 2478. 37. S. Ueda, J. Matsumoto, and T. Nomura, Chem. Pharm. Bull., 32 (1984), 350 - 353. 38. Y. Hano, T. Nomura, and S. Ueda, Heterocycles, 29 (1989), 2035 - 2041. 39. M. Takasugi, S. Nagao, T. Masamune, A. Shirata, and K. Takahashi, Chem. Lett., 1980, 1573 - 1576. 40. M. Takasugi, S. Nagao, S. Ueno, T. Masamune, A. Shirata, and K. Takahashi, Chem. Lett., 1978, 1239 -1240. 41. I. Messana, F. Ferrari, and M. do Carmo Mesquita de Aranjo. Tetrahedron, 44 (1988), 6693 - 6698. 42. Y. Hano, M. Aida, and T. Nomura, J. Nat. Prod., 53 (1990), 391 - 395. 43. Y. Hano, M. Aida, T. Nomura, and S. Ueda, J. Chem. Soc. Chem. Commun., 1992, 1177 - 1178. 44. P. Basnet, S. Kadota, S. Terashima, M. Shimizu, and T. Namba, Chem. Pharm. Bull., 41 (1993), 1238 - 1243. 45. T. Fukai, Y. Hano, K. Hirakura, T. Nomura, J. Uzawa, and K. Fukushima, Chem. Pharm. Bull., 33 (1985), 3195 - 3204. 46. Y. Hano, H. Kohno, M. Itoh, and T. Nomura, Chem. Pharm. Bull., 33 (1985), 5294 - 5300. 47. Y. Hano, M. Itoh, and T. Nomura, Heterocycles, 23 (1985), 819 824. 48. Y. Hano, T. Fukai, T. Nomura, J. Uzawa, and K. Fukushima, Chem. Pharm. Bull., 32 (1984), 1260 - 1263. 49. A. V. Rama Rao, V. H. Deshpande, R. K. Shastri, S. S. Tavele, and N. N. Dhaneshwar, Tetrahedron Lett., 24 (1983), 3013 3016. 50. Y. Hano, Y. Miyagawa, M. Yano, and T. Nomura, Heterocycles, 28 (1989), 745 - 750. 51. Y. Hano and T. Nomura, Heterocycles, 24 (1986), 1381 - 1386. 52. Y. Hano, S. Suzuki, H. Kohno and T. Nomura, Heterocycles, 27 (1988), 75 - 81. 53. Y. Hano, S. Suzuki, T. Nomura, and Y. litaka, Heterocycles, 27 (1988), 2315 - 2325. 54. T. Nomura, T. Fukai, Y. Hano, K. Nemoto, S. Terada, and T. Kuramochi, Planta Med., 50 (1984), 151 - 156. 55. Y. Hano, T. Nomura, and S. Ueda, Chem. Pharm. Bull., 37 (1989), 554 - 556. 56. S. Ayabe, A. Udagawa, and T. Furuya, Arch. Biochem. Biophys., 261 (1988), 458 - 462. 57. Y. Hano, A. Ayukawa, T. Nomura, and S. Ueda, Naturwissenschaften, 79 (1990), 180-182. 58. Y. Hano, T. Nomura, and S. Ueda, J. Chem. Soc. Chem. Commun., 1990, 610 - 613. 59. Y. Hano, T. Nomura, and S. Ueda, Can. J. Chem., 72 (1994), in press. 60. A. R. Moss and P. Schoenheimer, J. Biol. Chem., 135 (1940), 415 - 429.

478 61. P, M. Nair and L. C. Vining, Phytochemistry, 4 (1965), 401 411. 62. U. Weis and J. M. Edwards in : The Biosynthesis of Aromatic Compounds, John Wiley & Sons, New York, 1980, pp. 144 - 184. 63. H. G. Teutsch, M. P. Hasenfratz, A. Lesot, C. Stoltz, J. -M. Gamier, J. -M. Jetsch, F. Durst, and D. Werck-Reichhart, Proc, Natl, Acad. Sci. U. S. A., 90 (1993), 4102 - 4106. 64. P. Anastasis, I. Freer, K. H. Overton, D. Picken, D. S. Roycroft, and S. B. Singh, J. Chem. Soc. Perkin Trans 1., 1987, 2427 - 2436. 65. S. Seo, A. Uomori, H. Ebizuka, H. Noguchi, and U. Sankawa, J. Chem. Soc. Perkin Trans 1, 1988, 2407 - 2414. 66. J. F. Sanz-Cervera, T. Glinka, and R. M. Williams, J. Am. Chem. S o c , 115 (1993), 347 - 348.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

479

Modelling the Substrate Binding Domain of Horse Liver Alcohol Dehydrogenase, HLADH, by Computer Aided Substrate Overlay Maija Aksela and A.C. Oelilschlager

6.1.

INTRODUCTION

Enzymes are catalytically active proteins that are involved in every in vivo transformation. They enhance the rates of biochemical reactions by 10^ to 10^2 by reduction of the free energy of activation J Two distinctive properties of enzymes are their high substrate specificity and the narrow range of conditions under which they are effective. They usually catalyze one reaction of a few substrates. Activities are dependent on pH, temperature, the presence of cofactors, as well as concentrations of substrates and products. Enzymes perform specific reactions because they possess cavities in which substrates are oriented while they are transformed (Figure 1). This process Involves interaction of the substrate with amino acids of the enzyme. Enzyme-substrate complexes have been studied by kinetic analysis, chemical modification, inhibition of enzymes by specific compounds that Interact with active sites, detection of characteristic spectral absorption bands during reaction of enzymes with substrates, and X-ray crystallographic analysis of enzymes combined with compounds which are in similar structure to the natural substrates. The Interaction between enzymes and substrates has been analyzed by the concepts of "lock-andkey" and "Induced fit". The former presumes that the substrate surface must fit the enzyme surface like a key In a lock, while the latter refined theory assumes that binding of the substrate induces conformational changes in the enzyme to provide a better fit. 6.2.

Enzymes as Catalysts in Organic Synthesis

The potential of enzymes as catalysts in asymmetric synthesis has been recognised for many years.2-12 Rate acceleration and stereoselectivity, together with techniques for the low-cost production and the rational alteration of their properties, make enzymes attractive as chiral catalysts in organic synthesis. Enzyme-catalyzed reactions have been categorised into six main groups''^ as shown In Table 1. Three of them, oxldoreductases, hydrolases, and lyases have been found useful In organic synthesis.

480 About 50 of the 200 enzymes produced industrially have been shown to be important to synthesis.I"* Active site

© Enzyme

Substrate

Enzyme-Substrate Complex

Product

E-Product

Figure 1. Schematic diagram of an enzyme-catalyzed reaction. Table 1. Enzyme classification. Class

Action

Oxido-reductase

Oxidation-reduction reactions.

Transferase

Transfer of a functional group (e.g. amino, methyl) from one substrate to another.

Hydrolase

Bond cleavage by addition of water.

Lyase

Reactions involving additions to double bonds or cleavages with formation of double bonds.

Isomerase

Ligase

Racemization of optical or geometric isomers and certain intramolecular oxidation- reductions. Formation

of

new

triphosphate (ATP).

bonds, cleavage

of

adenosine

481 The chemo-, regio- and enantloselectivlty of enzymes make them Ideal asymmetric catalysts. Most Important Is the differential recognition of diastereotopic groups of chlral and prochlral substrates. The mild conditions under which most enzymes operate minimize Isomerizatlon, epimerlzation and racemization associated with many other chemical processes.^.Q Maturation of the petro-chemlcal Industry, environmental pressures for "clean chemistry" and the explosive development of biotechnology have increased interest in the application of enzymatic processes to organic synthesls.3 Enzymatic processes play an increasing role In the generation of chlral pharmaceutical Intermediates, watersoluble materials and Copolymers. One problem In the development of enzymatic reactions for organic synthesis Is the prediction of the stereochemistry of reaction. Reliable models for prediction of stereochemistry are needed to broaden the application of enzymes to organic synthesis. 6.3.

Horse Liver Alcohol Dehydrogenase as a Catalyst

Horse liver alcohol dehydrogenase, HLADH, (also abbreviated as ADH or LADH) is the most extensively studied oxido-reductase. It plays a central role in ethanol metabolism and has been one of the main tools for understanding the mechanism of this process.'iS it was crystallized from horse liver In 1948 by Bonnlchen and Wassen and Is commercially available. Three Isozymes EE, ES and SS are formed by dimeric combination of two different, E or S (E "ethanol-active" and S "steroid active"), protein chains.''6 The EE- Isozyme of HLADH has been used in organic synthesis. HLADH Is nicotinamide coenzyme {NAD+/NADH) dependent and catalyzes the redox equilibrium between a large number of alcohols and ketones or aldehydes (Figure 2). The equilibrium is overwhelmingly in favour of the reduction reaction.^ Phosphate analogs of the coenzymes, NADP+ and NADP may also serve as coenzymes in very limited situations. The oxidation-reduction takes place through a ternary complex in which the substrate and coenzyme are simultaneously bound In the active site of enzyme. A zinc Ion at the active site binds the substrate and facilitates hydrogen transfer by acting as a Lewis acld.17 it has been assumed that the alcohol forms a zinc bound alkoxide during the reaction.is His-51 has been suggested to be important In transfer of a proton between the active site and surrounding solvent.''® Which amino acids are

482

9

HH..

R

H.N'Y^ OII '^ i A- --fNADH

NAD-"

NH2 HO

OH

^ O

O

^

OH

OH

N ^T^ll

^ 3 HO

f OH

Figure 2. Schematic view of the oxido-reduction catalyzed by HLADH. responsible for the observed pH dependencies of coenzyme association and dissociation as well as substrate binding and hydrogen transfer steps are not yet certain.19 The oxidation has been suggested to occur by an ordered mechanism in which the productive substrate binding site is formed after coenzyme binding. Dissociation of the enzyme-NADH is the rate limiting step.20 OH

f

CH3OH R

HO'

RCH2OH

R'

bnoH R'

d 6T Q5^

YADH HLADH Steroid alcohol dehydrogenase

Figure 3. Structural

specificity

of

HLADH

compared

with

dehydrogenase, YADH and steroid alcohol dehydrogenase.

yeast

alcohol

483 Over one hundred compounds have been examined as substrates for HLADH. It has a broad structural specificity and is well suited for organic synthesis. The enzymes reacts with acyclic, mono-, bi-, tri- and tetracyclic (steroidal) substrates. Figure 3 shows a comparison of the substrate specificity of common dehydrogenases.^^ HLADH has a wider use than yeast alcohol or steroid alcohol dehydrogenases. It operates on most substrates with high enantiomeric specificity. Examples in which HLADH exhibits superior selectivity include preparation of stereoisomers of bridged bicyclic

(±)

O unreactive enantiomer

83% e.e.

2.

100% e.e. H

H

(±) P

^OH

3. XH2CH2OH

'CH2CH2OH

(±)

4.

97% e.e.

J 10'''^ HO-^

L ^OH ^OH (±)

O

^

J^ J a'^^^ O'^'^X)-^ 100% e.e. OH

^ ^ unreactive enantiomer

"^ " ^

macrolide and polyether antibiotics

OH

5. (±)

100% e.e.

Figure 4. Examples of HLADH-catalyzed reactions.

100% e.e.

484 compounds (entry 1 in Figure 4).22 Stereospecific reduction of highly symmetrical diketones is also facile (entry 2, Figure 4).23 it is also possible to combine several different kinds of specificity to achieve in a single step a transformation that usually requires several reactions (entry 3, Figure 4).24 Chiral lactones, which are useful starting materials for many natural products, can be formed by stereospecific oxidations of mesodlols (entry 4, Figure 4).25 Even thioketones are reduced stereospecifically (entry 5, Figure 4).26 Reactions with HLADH typically occur at temperatures between 4°C and 25°C and in the pH range of 5 to 10. For catalysis of a reduction the optimum pH is ~7 while for the reverse oxidation it is -8.^ Reaction times vary from a few hours in the most favourable substrates and 2-3 weeks for the slowest. The disadvantage of HLADH has been the high cost of coenzymes.

Fortunately, several recycling methods are available that

allow reduction of substrates at the research scale (up to 1 kg of substrate).27-30 Por example, the ethanol-coupled method has been used for reduction and flavin mononucleotide (FMN) recycling for oxidation. 6.4.

X-Ray Crystallographic Studies of HLADH

Understanding of protein structure and function has been greatly enhanced by X-ray crystallography. At low resolution (4-6 Angstrom), the electron density map reveals the folding of the polypeptide chain, but few other structural details. At 3.0 A , it is possible, in favourable cases, to resolve amino acid chains, while at 2.5 A the positions of atoms may often be given with an accuracy of ± 0.4 A. In order to locate atoms to 0.2 A, a resolution of about 1.9 A and very well ordered crystals are necessary.3i.32 Three different crystal forms of HLADH have been studied crystallographically: orthorhombic (C222) crystals of the apoenzyme (native enzyme), triclinic (PI) or monoclinic (P2-|) crystals of the holoenzyme (apoenzyme with its coenzyme) and of ternary complexes (holoenzyme with substrate or inhibitor).''S j h e X-ray structure of the apoenzyme (EE-isozyme) was determined to 2.4 A resolution in 1976.33 The structures of the ternary complexes with NADH and dimethyl sulfoxide,^^ NAD+ and p y r a z o l e , 3 5 NAD+ and p-bromobenzyl alcohol,36 tetrahydro-NAD and dimethylamlnocinnamaldehyde37 and NAD+ and 2,4-(4-pyrazolyl)butylisoythiourea,38 have been determined to 2.9 A resolution. HLADH (EE) has been found to be a dimer comprised of two identical subunits of Mr 40,000 (Figure 5). Each subunit contains both a coenzyme and a substrate binding

485 domain. The single polypeptide chain of one subunit contains 374 amino acids of known sequence.^d These two subunits are separated by a cleft, containing a deep pocket. Both subunits of the dimer bind coenzyme and substrate in essentially the same manner; thus these two active sites are assumed to be the same.^s The two subunits of the dimer are linked by interactions within the coenzyme binding domains which form a core of 143 residues in the middle of the molecule. The catalytic domains are at opposite ends of the dimer and comprise 231 amino acid residues.33

Figure 5. Schematic of dimeric HLADH.^^ The coenzyme binding domain has a structure very similar to corresponding domains in other dehydrogenases and some kinases. The coenzyme binds in the cleft across the edge of the coenzyme binding domain via several hydrogen bonds, charged and hydrophobic interactions.'^^ The nicotinamide moiety of coenzyme is positioned in the active site with the A-side of the ring facing the zinc atom.^^ Coenzyme binding induces a conformational change from an open to a closed form. The enzyme may exist in both conformations, but the equilibrium favours the open conformation in the absence of coenzyme binding.i^

486 The major conformational difference between the open and closed forms is a rigid body rotation of each catalytic domain by about 10° with respect to the coenzyme binding domain. Thus, a cleft between the catalytic and coenzyme domains is closed, making the active site less accessible to solution and more hydrophobic. During this change some amino acid side chains, especially that of, Val-294 change their orientation. The side chains of Leu-57 and Val-294 which are about 15 A away from each other in the apoenzyme form, are only 4-5 A apart in the ternary complex. No new residues are brought into the substrate binding domain. The conformational change also brings the catalytic zinc 1 A closer to the nicotinamide binding domain and moves the substrate and nicotinamide binding domains closer. The interactions between the protein and the nicotinamide ring of coenzyme are assumed to be involved In Inducing the conformational change of the enzyme during coenzyme binding.18 COENZYME

174

CYS

Figure 6.

48

Schematic of ethanol in substrate binding domain of HLADH.

The substrate binding domain has two zinc atoms, one of which, the catalytic zinc, is located at the bottom of the hydrophobic substrate binding domain, 20 A from the surface of the molecule. Cys-46, Cys-174 and His-67 are ligated to this zinc which binds the substrate in a position relative to the coenzyme that facilitates direct hydride

487 transfer.""^ The function of the second zinc atom and the protein lobe that surrounds it, Is unknown. A deep pocket accommodates the substrate and the nicotinamide moiety of the coenzyme. It is 5-10 A wide and --20 A long, extending from the protein surface to the catalytic zinc atom. The part of this pocket associated with substrate binding in HLADH Is much larger and more hydrophobic than in other dehydrogenases. The substrate binding domain consists of a large hydrophobic barrel lined with non-polar side chains of Leu-57. Phe-93, Phe-110. Leu-116, Phe-140. Leu-141, Pro-295, Pro-296, lle-318 and Val-294, which are in the same subunlt as the ligands to the catalytic zinc that resides at the bottom of the barrel.33 The only polar groups in catalytic site are Ser-48 and Thr-178 (Figure 6). In the open apoenzyme form several firmly bound water molecules are located in the inner part of the pocket outside the co-ordination sphere of zinc. In the closed form of the ternary complex the coenzyme and the substrate occupy this space.''S According to the X-ray crystallographlcally determined structures of ternary complexes of HLADH the side chains of amino acid residues bordering the substrate binding domain can adopt chain conformations compatible with the volume occupied by the bound llgand. In the binding of p-bromobenzyl alcohol and the inhibitor, dimethyl sulfoxide, there are changes in the positions of the side chains of Leu-116 and Ser48.36 The conformation of Leu-116 is also different in the NAD+-4-lodopyrazole and (dlmethylamlno)cinnamaldehyde complexes.37 in the NAD+ and 2,4-(4pyrazolyl)butylisothlourea complex side chains of Met-306, Leu-57, Leu-116 and lle318 assume different conformations than in the HLADH-NADH-dlmethyl sulfoxide complex.38 6.5.

Previous Models for the Substrate Binding Domain of HLADH

Several models have been developed for the substrate binding domain of HLADH from kinetic studies and/or X-ray crystallographic data. The substrate binding domain has been analyzed using diamond lattice and cubic section models (Figure 7). The first attempt to describe the topology of the substrate binding domain was by Prelog In 1964 In terms of the diamond lattice model.^"' This work was refined by Jones et al. In 1976 and 19778.42,43 and Horjales and Branden in 1985.^4 The first cubic section model of HLADH was developed by Jones and Jacovac in 1982.'^5 other types of

488 models have been developed by Ringold et al. in iges,'*^ Nakazaki et al. in 1980,^-^ Duller and Branden in 1981^8 and Lemiere et al. in 1982 (Table 2)."^^ The diamond lattice model (Figure 7) was developed using six-membered ring ketone substrates. The determination of forbidden and undesirable positions was achieved by analysis of the relative rates of reduction of a series of cyclohexanones and decalones of known absolute configuration.8 The geometry indicated at the C-0 centre was considered to resemble the structure of the alcohol rather than that of the ketone in the transition state. It was assumed that all substrate molecules bound with oxygen in the

("-

\\A] O — undesirable position # - forbidden position

0

OH - forbidden cube

Figure 7. Diamond lattice and Jones cubic section model. same orientation and direction of the forming C-H bond. The back of the lattice section was thought to be a flat coenzyme binding site and the lower and left sides of the lattice were thought to be bounded by the enzyme. The oxidation and reduction was assumed to be forbidden if binding of a substrate superimposed a group in one of the enzyme occupied locations. If the group occupied an undesirable location, the rate of reaction would be very slow. Horjales and Branden (1985) constructed a diamond lattice model by docking cyclohexanol and its monoethyl derivatives into the experimentally determined active site of the enzyme (X-ray crystallographic structure, 1982), using computer graphics and energy minimization methods."*"^ The lattice positions were classified as allowed, forbidden or boundary depending their distances to protein atoms (Figure 8). The

489 lattice positions were considered as allowed (A), if all distances to neighbouring atoms are larger than 2.9 A, forbidden (F), If any distance to a fixed atom Is less than 2.6 A or boundary (B). Boundary regions were sub-grouped into B 1 , consisting of those positions that Interact only with side chains of Leu-57, Leu-116 or Met-306 and B2 comprising all other boundary positions. The volume available to substrates was similar to the substrate binding domain derived from analyses of previous kinetic data of the methyl cyclohexanones (Dutler and Branden, 1981^^) and the diamond lattice model (Figure 9a).s Using this method, Horjales and Branden extended the lattice model to extend the limits of definition of the substrate binding domain (Figure 9b). Table 2.

A summary of previous models and methods HLADH, presented

chronologically Author

Year

Models

1. Prelog

1964

2. Ringolde/a/.

1965

3. Jones et al.

1976

4. NakazakI ef a/.

1980

5. Dutler & Brandon

1981

6. Jones & Jacovac

1982

7. Lemlereet al.

1982

the diamond lattice model, using kinetic data composite structures of satisfactory and unsatisfactory substrates In front and side views the refined diamond lattice model the C2 -ketone rule oxidation and reduction cyclohexanol rings superimposed and compared; kinetic data; X-ray of apoenzyme the cubic section model (manual) "flat" cyclohexanone model

Substrates cyclohexanone and decaione derivatives 3- and 4-alkyl cyclohexanones, 10methyl-2-decalones cyclohexanone and decaione derivatives the cage-shaped C2ketones 2-,3-,4-alkylcyclohexanol with alkyl groups, methyl, ethyl, l-propyl and t-butyl. alkyl cyclohexanones 2-,3-,4-alkyl cyclohexanones

8. Horjales & Branden 1985

the diamond lattice

cyclohexanol and Its

model, using docking

monomethyl derivatives

& computer graphics

490

asfi(\

utnoi

Figure 8. Stereo diagram of the substrate binding domain using the substrate docking method.

Figure 9. Figure 9a: A comparison between two diamond lattice models. Positions A-J are derived by kinetic studies as forbidden or hindered. Positions K-R represent boundary or forbidden positions defined by Horjales and Brandon. Figure 9b shows the extended diamond lattice model. In 1982 Jones and Jacovac mechanically constructed a cubic section model using Framework Molecular Models.-* The cubic section model was conceived because

491 substrates other than those containing a cyclohexyl ring, e.g. cyclobutanones, and cyciopentanones were not easily fitted to the diamond lattice model. The lattice was based on sp^^-hybridised carbon lengths and angles. Thus, accurate predictions for numerous substrates were difficult to make. Models that resemble the cubic section approach have been developed for mono-oxygenases (Johnson in 197850) and microbial reductases (Nagazaki etal. in 198051). Jones and Jacovac employed the kinetics of alkyl cyclohexanone reductions (Duller et al. 1977,1981 and Lemi^re et al. 1980) for the estimation of forbidden and limited access volumes. In addition, they used the X-ray derived structure of the apoenzyme, the enzyme-NAD complex and several ternary complexes to derive the shape of the substrate binding domain (Figure 10a). The model, used 1.3 A3 cubes and the C-0 function is defined as an alcohol as in the diamond lattice model. The oxygen of substrate Is assumed to ligate to the catalytic zinc and hydride delivery or abstraction occurs from the front. Jones and Jacovac defined forbidden and limited access regions as A1-2, El-3, G4-5, L4-6, M7-8 and Q7-9, the left-hand halves of B1, H4 and J4, P7 and the right-hand halves of B1, H4 and N7 as well as cubes 0 9 , P8, P9. Binding in the limited access area results in a reduced rate of reduction of substrate. Cubes below the plane of the paper reflect the character of cubes immediately above them. The allowed region was suggested to correspond to the substrate binding domain of HLADH as identified by X-ray (Figure 10b). However, the boundaries between the allowed, limited and forbidden regions were assumed to be flexible due to the movement of amino acid side chains during substrate binding. The forbidden cube

y/ipg/^ y^ y%

**.^T"" "^y^

[M

N

o

p

Q

R

G

H

1

J

K

L

A

B

0

D

E

F,-- ..t.

5

>

-aC

^^

U

Figure 10.

U(F1)

£9

8

/ 4

2

»

3J

V

Leu57 Ser 48

\Z^ ^

OH

Figure 10a, Jones cubic section model. Figure 10b, Correlation with X-ray crystallographic studies.

492 E1 corresponds to the Phe-93 location; K4 and Q7 to lle-116; E2, E3 and K5 to Phe110; 0 9 , P9 and Q9, to lle-318; B1 to Ser-48; G4, M7 and half of the cubes, H4 and N7 to Leu-57 while J4 and P7 are limited because of the nicotinamide coenzyme. O

6 (2R)-]

,"^2- tt::p=^ — H - ^ : i i j 2 : i £ ! i ^ o I N7

07

P7

Q7

H4

14

J4

K4

B^

El

B1 >^1

KJ-^H

O

II

HD' ^ H

6' (2S)-1

HLADH (2S,2S)-2

I

iV^^X

Figure 11. Prediction of stereochemistry of reduction for 2-alkyl cyclohexanones by Jones cubic model. The direction of hydride delivery from NADH is indicated by the arrows > and > for reactive and unreactive product or orientations, respectively. Only conformation IV can bind without penetrating forbidden regions.

493 The prediction of product fit into the HLADH active site was conducted separately for each enantiomeric product, using the mechanical model. For the product to fit well, the bound substrate must not occupy forbidden positions. Penetration of substrate Into two or more cubes of limited access was judged to be equivalent to a forbidden interaction. The Jones group have applied this model to more than 50 different cyclic and bicyclic compounds (e.g. Figure 11). The predictions for both reductions and oxidations use the same model and are reliable. 6.6. Present Cubic Section IVIodei for HLADH 6.6.1.

Generai

The goal of the present study was to develop a computer-based cubic section model of the substrate binding domain of HLADH. It was considered that the Jones cubic section model could be refined by use of computer assisted substrate overlay in combination with kinetic data on a wide variety of substrates. As in the Jones approach we used the alcohol products as the surrogate substrate structures. Thus, we determined the low energy conformation of alcohols produced from ketones that have been reported to be reduced by HLADH and for which comparative kinetic data vs cyclohexanol could be calculated. As well, we determined the preferred conformations of all alcohols that would have been produced from ketones subjected to but falling to undergo HLADH reduction. These calculations utilised molecular mechanics (MACROMODEL) and yielded accurate co-ordinates for all atoms in each alcohol. Where enantiomeric or stereoisomeric alcohols were produced or capable of production, the co-ordinates of each were calculated. Alcohol products were assigned a priority number based on their rate of production vs cyclohexanol. Relative reduction rates, enantiomeric excesses of alcohol products, extent of conversion of substrates, yields and absolute configurations of alcohols were used In calculating the priority number for each substrate with the faster reacting substrates receiving the higher priority numbers. Co-ordinates of atoms in the energy minimised alcohols were transferred to ENZYME, a program that allowed the structures to be identically oriented in a cubic section model. Using ENZYME program the energy minimised substrates were placed in a Cartesian co-ordinate (x, y, z) system with C-0 aligned with the carbon at the origin, the oxygen on the negative y axis and the ahydrogen the yz-plane.

Conducting this operation on all alcohols gives a map of

acceptable and forbidden cubes according to the average priority numbers for each

494 cube. Forbidden areas were assumed to be occupied or blocked by amino acid residues or coenzyme. The model is directly comparable to that of the Jones group in that we have used cubes of 1.3 A on a side and placed the carbon-oxygen bond as well as the carbon-hydrogen bond of the carblnyl carbon In the same orientation. We have compared the present model with that of Jones as well as with the X- ray crystallographic structure of the enzyme. 6.6.2.

MACROMODEL and ENZYME Programs

MACROMODEL and ENZYME allow facile structure entry, energy minimization, reorientation, atom to cube assignment and calculation of average cube priority values for each cube.

MACROMODEL, developed by Clark Still at Columbia University

employs a recent version of Allinger's popular force field minimization routine. Each molecule is described in the program according to the atom types It contains and the connections between atoms. The energy of each conformation of each structure is calculated by evaluation of the Van der Waals, bond stretching, bond angle bending, torsional and dipole interactions. In Appendix 1 there is a description of those parts of MACROMODEL that were useful in this work. ENZYME SITE PREDICTION SYSTEM V 1.6

• • Bi m •

-

Forbidden Priority < 0.5 Priority < 1 Priority < 5 Priority < 10

•

- Priority >= 10 Layer 1

Cube file is col 3.cf

READ CF READ DATA ADD TO CF REMOVE CF ALIGNMENT ROTATE ON X ROTATE ON Y ROTATE ON Z PREDICT DISPLAY CF ROTATE BONDS

Press

PRINT CF SAVE EXIT CSl.30 CORNER DISCOL 6

Figure 12.

Menu of the ENZYME program. One layer of constructed substrate

binding domain is presented.

495 The ENZYME program was developed at Simon Eraser University In Pascal on a VAX 750 using an IBM personal computer with an enhanced graphics adapter and graphics tablet running a VT 100/Tektronix 4107 terminal emulator. The source code program is available from the authors. It is fully compatible with the MACROMODEL program and was designed to be easily transferred between computers. ENZYME allows construction of cubic section substrate binding domains for enzymes other than HLADH. The program executes various procedures through menu driven processes. Representative procedures are shown in Figure 12 are read the product file, add the information for the molecule to the cube file, remove information for the molecule from the cube file, align the molecule in space, rotate about an axis, predict the averaged priority for the product, display cube file data, rotate about a bond, change the size of the cube, shift the origin from the center of the cube to the corner, print the cube file data, save the cube file and exit the program. An additional program, COMPARE, allows plots and cube by cube comparison of two constructions of substrate binding domains. COMPARE was also written in Pascal on a VAX 750 using an IBM personal computer configured as above and the source code is available from the authors. The documentation of ENZYME and COMPARE menus are presented in Appendix 2. The details of the use of the programs, MACROMODEL and ENZYME are described in Chapter 7.5. 6.6.3. Criteria for Choice of Substrate Surrogates As in the Jones protocol the cubic section model of the substrate binding domain of HLADH were constructed using structures of alcohol products rather than ketone substrates. The alcohol products were originally chosen by the Jones group because the transition state the geometry for the reduction was considered to resemble that of the alcohol rather than that of the ketone. The relative rate of reduction of substrate vs cyclohexanone for each ketone was required to be known. Furthermore, configurations of alcohol products, enantiomeric excess values, yields and % conversion of substrate required for calculation of the priority number for each enantiomer of product should be measured under comparable conditions (i.e. pH, temperature, concentration of enzyme, coenzyme and substrate, etc.). According to Alderweireldt et al. (1988) HLADH models are valid only for reaction conditions used in the reactions from which the models are constructed.52 Furthermore, the model is only reliable if the reactions have been conducted under kinetic control.

496 Table 3. The substrates used in construction of the active site picture. OH

OH

OH ,

OH ,

^48.53

2^.53

38

OH

OH

OH

g26

754

g54

OH

OH

^353,55

^^53,55

OH

OH

^48

5S4

g54

OH

OH

OH

,0^

„26

^^26

OH

cis & trans

cis & trans

^^

trans

2 1 ^ ' 22^=*

23^'

cis & trans

,ge

-.8 ,78

cis & trans

,«8 ^g8

^g„

2o„

OH trans 24"

2 ^

2 ^

05 OCT A„ ^ 27®

32

28®

29^

33

34

X"

30^^

31®

35

36 OH

OH HO

^

HO 37

38

39

57

40

41 OH

^ 4 OH

OH

42"

43

,58.56

OH

44

45^

497 Using these criteria 45 substrates for which experimental information was available for reaction at pH 7 were chosen to construct the model (Table 3). There is a significant amount of literature data for substrates of HLADH which lack information on enantiomeric excess values and absolute configurations of products and where the relative rates of reaction have been measured under different reaction conditions (e.g. pH 8.5). Substrates falling Into this category (46-69, Table 4) were not used. Heterocyclic bicycllc substrates 70-73 were only used for testing.

O

Example 1: (±)

OH OH

O

Example 2:

OH

OH

ir &

^ (±)

OH

OH OH

X::^oH^^^

\::^

^=^/X:::poH

Figure 13. The possible conformations of two substrate surrogates.

498 Table 4. Product structures used in testing model for prediction of stereoselectivity. OH

OH

OH

OH

OH

46^^ 46''

4/^ 47'

48^

49^°

50^°

OH

OH

OH

OH

OH

6 ^ 6 ^ O^ Cx. 6^ .60

51^

60

52'

OH

53

OH

6

OH

.60.49

g^60.49

62

OH

^

-OH

^60.49

^H

6< 6: A^.o^ 67^

66"

HO"^-^^

71

68"

HO-^^T"^

72'

69"

70"

HO-^^-^TT"^

73'

Structurally rigid substrate surrogates were added to the cubic section model before more flexible molecules as the following order signifies: pentacyclic, tetracyclic, tricyclic, bicyclic ketones, trans/cis-decalones, methyl cyclohexanones and alkyl cyclohexanones. The hydroxyls of cyclohexanols were oriented axlally with respect to cyclohexyl rings consistent with the Jones protocol and with the observation that

499 reduction of conformationaily locked trans-decalins (16,18, 20, 22 and 23) produced exclusively axial alcohols. All low energy conformations of conformationaily flexible alcohol products were added included (Figure 13). 6.6.4.

Calculation of Priority for Alcohol Products

The priority number for an alcohol product was determined by the relative rate and the stereoselectivity of reduction. The priority number for each substrate surrogate was calculated from the rates of production of each enantiomeric alcohol product using the total relative rate of reduction vs cyclohexanol and the enantiomeric ratio (E) formula derived by Sih (Table 5).^^ Table 5. Formulas for calculations of the priority numbers for product enantiomers.

ln[1-c(1-fee(P))] *=~ln[1-c(1-ee(P))]

Sih's formula For product 1:

(1) (2) (3)

For product 2 :

P+ = R ( Y , + Y 2 ) ^ E 2 + I )

p-=f^(v^)(E?n-) where

(4)

(5)

E = the enantlometric ratio, the discrimination between two competing enantiomers by enzyme c = the extent of conversion of racemic substrate e e ( P ) = the optical purity, the enantiomeric excess of the product P = priority number for major enantiomer p = priority number for minor enantiomer Y = yields of products R = relative rate vs. cyclohexanone

The relative rate for cyclohexanone reduction was set at 100. This would give, for example, a total relative rate of 24 for the reduction of 2,3,5,6-tetrahydro-2-

500 isopropylpyran-4-one (precursor of 7, Figure 14)7 ENZYME contains a subroutine to calculate the priority numbers. It will request ail necessary values for the equations (Table 5). The ratio of enantiomers (major or minor) are to be given by the user. The priority values are calculated and displayed by using the PREDICT opWon of the menu. The priority number, 22.6 is given for the major product, (2S,4S-7). For the enantlomer, the priority number is near zero because the enantiomeric excess for this reaction is 1.00 (Figure 14). OH

QH

0.r

R = 24

c = 0.5 (2S,4S) Yi=32 ee=1.00

(2R,4S) Yi=2 ee=0.28

(2S.4S) Product 1:

P+ = 22.6 E = 15200

'

(2R.4R)

p. = C

(2R.4S) Product 2:

E = 2.3

I

P+=1.0

(4)

p. = 0.4

(5)

>

(2S.4R)

Figure 14.

(2)

^

Priority number calculation for enantiomers of 2,3,5,6-tetrahydro-2-

isopropyl-pyran-4-ol.^ 6.6.5.

Construction of Substrate Binding Domain for HLADIH

A global energy minimised conformation using MACROMODEL was obtained for each compound to be used In construction of the model. The torsional angle constraint was set at zero degrees for the atoms needed in this orientation then the minimization process was executed using the Block Diagonal Newton Rapson minimization method, BDNR.

Compounds to be included in the model were reoriented according to the

Jones protocol, using the MACROMODEL ANALYZE mode (ALIGN X, ROTATE Y options). The C-0 bond is aligned with the carbon at the origin, the oxygen on the negative y axis and the a-hydrogen in the yz -plane (Figure 15). The coenzyme is assumed to be in front of the origin and the catalytic active zinc atom is in the lower portion, directly ligated to the hydroxyl group of the substrate.

501 Each compound data file was transferred in ASCII format from MACROMODEL to the ENZYME program by execution of MM FORMAT on the file produced by MACROMODEL. Compound files in ENZYME are called for entry Into cube files by using READ DATA option. At this point the program requests the priority number of the compound which was calculated by the process, explained in Section 7.4. Using the ALIGNMENT option each compound was aligned with reference to the axes with the carbon on the origin (0, 0, 0). It was helpful at this juncture to check the orientation of each structure by rotation about the y-axis by ±10 degrees. Correctly minimised and oriented structures were added to the cube file of a chosen name using the ADD CUBEFILE option. After each stmcture was added the cube file was saved. The model produced by this process could be viewed, using the DISPLAY CUBEFILE option. To obtain the co-ordinates of the cubes and average priority value of each cube the PRINT CUBEFILE option was used. Addition of a structure to an empty cube file assigns each occupied each cube of the model the calculated priority of the structure. If atoms of a second structure occupy some previously marked cubes the average of the priorities for the entered structures will be calculated for these cubes. When a structure known not to be produced (i.e. Its priority is zero) is added to the model all cubes not previously identified as being occupied and which are occupied by that molecule will be marked forbidden. The forbidden classification will be removed If a subsequently added structure has atoms occupying those cubes. The program does not use priority values of zero to calculate average priorities. Addition of many structures to a cube file creates a cubic section model in terms of allowed (with averaged priority numbers) and forbidden (priority 0) cubes. Each cube is identified according to the Cartesian co-ordinates (x, y, z) of its corner farthest from the origin. The modelled volume contains several layers of cubes of user defined size. ENZYME allows average priority values of cubes to be viewed in different colors (represented as shades in Figure 12). Display of cube priorities in two colors can be chosen. Green is the default color for allowed cubes and red for forbidden cubes. A hard copy of the average priority values for each cube can be obtained by using the PRINT CUBEFILE -menu option. The method allows openended improvements since new structures can be added to refine cavity topology.

502

^

Reaction

_^h^£L,

J^

(±) rel rate vs cyclohexanone = 26 conversion of the reaction * 0.46 an enantiomeric excess of product = 0.64 [^ a yield of product = 39

E - 7.76

Computer simulation H

(0-^

MM2

A T H -^--'- y^H

DRAWN

I

PRIORITY = Acceptance ranking

REORIENT

OH 2.97

23.03 (01lY

I

yo

z:

m

I cube - atom match P = cube acceptance ranking

cube acceptance ranking = P

A = average acceptance ranking

I repeat for additional substrate Figure 15. Procedure of defining topology of the substrate binding domain of HLADH.

503 Using the same equivalently oriented structures 12 different cube files were constructed (Table 6). The effects of location of the origin in the corner or In center of the center cube, size of cubes (1.3 A or 0.65 A) and inclusion of hydrogens in the structures were studied. Special cube files, coded co13*, co065*, ce13* and ce065* were constructed without the hydrogen on oxygen (Table 6). Usually this hydrogen can occupy two side by side cubes in layer -1. This is due to very slight differences in orientation produced during the minimization process. Table 6. Twelve different cube files constructed. orii jjln corner center

Icode of cube file

1. co13

X

2. CO065

X

3. 0013*

X

4. CO065

X

5. co13h

X

6. co065h

X

X

X

X

X X

X

X

X

11. ce13h

X

12. ce065Hi

X

X X

X

X

10. ce065*

6.6.6.

X X

8. ce065

1

X

X

X

X

a-H&H on SL b s t r a t e s oxygen g-H aii hydrogens

X X

7. ce13

9. e e l 3*

cub ) size 1.3A 1 0.65A

X X

X

X

X

X X

X

Model for Substrate Binding Domain of HLADH

A goal of this work was to refine the cubic section model developed by Jones"^^ using computer modelling. Thus, equivalent orientations of compounds, cube size, 1.3A. and origin location were used to construct the cubic section models of HLADH. Because the jacks and plastic tubing of the mechanical model occupy space, Jones changed

504 the edges of cubes C, D, I, and J in layer 0 to 1.4 A. Computer modelling allowed the Present Model (co13) Jones Model

layer 4

layer 3

layer 2

M^ N

0

P

Q

R 1

p

H

1

J

K

L

r

B

c p

E

F

u

u

u

u

u

u

1

505

layer 1

[N

0

H

1

B

c p

u

U

^M J

u

layer 0

0.02

^ ^

1

layer-1

4

34

1

lofH

I1

45 10 9 11

506

layer-2

fA

N

0

P

Q

R1

r

H

1

J

K

L

r

B

c p

E

F

u

u

u

u Cubes:

u

u

n

•

|58j allowed in present model I

>

(5 = averaged priority number, 8 = number of atoms)

I allowed in Jones model forbidden

1 1

cube size 1.4 A

forbidden or limited in Jones model limited in Jones model allowed for some 3-alkylcyclohexanols or heterocyclic bicyclic alcohols (usually forbidden) allowed for some 4-alkylcyclohexanols (before unknown) Figure 16. Cubic section model for the substrate binding domain of HLADH: Jones model (left) and the present (co13, right). Jones assigned each cube alphabetically and we have used his convention to allow comparison. The priority numbers for each cube were rounded in the present model. size of ail cubes to be of the same. The construction of the original cubic section model used only the kinetic data of alkyl cyclohexanones. In the present model 166 different conformations of alcohol products and potential products were used to construct each cube file. Conformations of 55 had positive priority values, while 111 represented non reactive conformations.

In both the Jones and the present models substrates

contained only one carbon bound hydrogen which is the a-hydrogen and a hydrogen on oxygen, in our model other hydrogens of modelled structures were deleted after minimization. The previous division of cubic space into allowed, limited and forbidden layers spaces has been refined in that ENZYME generates individual acceptability rankings for each cube (Figure 16). ENZYME gives a printed list of the cube file: the contents of each cube, the average priority of cubes and the total number of atoms eventually in each cube. For model c o l 3 the origin is in cube OD (cube D in layer 0),

507 the oxygen In cube OD-2 (2 cubes under cube D in layer 0 ), the a-hydrogen In cube -1 D (cube D In layer -1 ) and the hydrogen on oxygen In cubes -1C-2 or -1D-2. In both the Jones and col 3 models the allowed area opens most significantly to the left of the origin. This can be seen more clearly by the present model because layers 3

Table 7. Comparison between cubes in Jones and col 3 models. 1 layer, 1 the code Jones model Present model the contents of cube of cube col 3 & priority ( ) layer -1 H K 0

F F F

34 4 0.02

substrate 30a(39) 30b(4) 31(0.02)

layer 0 G M N H e J Q K

o aM-aR 1

F F F&L F&L L L F F A ?

0.2 0.8 0.01 1 2 12 1 1 0.2 17 1 For A*

! 1

15a(0.2) 2b(0.57).14a(0.98) 45b(0.01) 10 substrates 33d, 41b, 90a. 30b 15 substrates 45a(1.29) 38d(0.27),37d(0.05) 38c, 30a *aO: 70a

layer 1 A & uA G M H N uB o

J p Q K E uE F—2 aN-aQ

F F F A A AorL A A. F or L F F F F F ?

For A* 0-2 unk 2 2 2 0.5 2 5 4~ 0.1 13 1 0.4 For A*

*52a, 53b, 55a, 48a I5a(0.2) 38a, 37a, 40a 37c. 41b, 44b 12a(3.6),39b(1.01) 37c, 37d, 42b 11 substrates 37c, 37d, 42a 44a(94.9),37d(0.27) 38b(0.08),37b(0.1) sulphur-compounds 39a(1.49) 7d(0.43) VO to 73

1

508

layer 2 K

F

Q E

F F

uE

F

uA bG uC B-2 N G 0 P

A A AorL A A A ForL ForL

0.4 unk 2 0.2 2 2 1 F A* 13 4 For A*

45a(1.49),45b(0.57) 38d(0.87),37d(0.05) 38b(0.08). 44b(5.13) 45b(O.Oi) 7d(0.43), 6d(0.08), 5d(0.26) 2x120(2.39) 2x120(2.39) 39b(1.01) *58 110(23.2). 120(2.39) 17a(7.5). 41b. 19a 60a, 63b. 66b, 65a

layer 3 bG bA A B

? ? ? ?

61 61 31 11

C D C-2 D-2 D-l C-l J 1 aP aQ E

? ? ? ? ? ? ? ? ? ? ?

3 4 28 28 A* A*

1

1

A* A* A* A* For A*

8a(120.5), 120(2.39) 8a(120.5). 120(2.39) IOo(39.3). 110(23.2) 18a(3.0), 6a(14.8), 7a(22.6). 20a(3.5) 18a(3.0), 20a(3.5) I9a(4.0) 80(27.53) 80(27.53) *58a *59a 58a, 58b. 59a, 59b 58a, 59a 66b, 65a, 63b 65a 70-73

j

layer 4

1

4c

?

A*

61a. 63a. 65b. 66a

and 4 are more fully defined. Due to the greater number of compounds used the allowed area in the present model is slightly larger than in the earlier version. A list of average priority values for eaoh oube are given in Appendix 3. comparison of each cube in the two models is presented in Table 7.

The detailed

509 The cubes with priority numbers of 2 or lower are considered boundary areas (Table 2). They are allowed for some substrates and forbidden for others. The priority of structures that occupy these cubes is usually very low. Some bulky structures were found to occupy low priority cubes G, N, Q and K in layer 0, cubes G, K and uE in layer 1 and cubes K and E in layer 2. These are forbidden in the Jones model. Bulky substrates are assumed to change the orientation of amino acid side chains lining the substrate binding cavity. It is suggested that except for bulky structures these cubes will be forbidden as suggested by Jones. The high priority value of cube 1E (13.0) is caused mainly by sulphur containing substrates. In every case C2 of these heterocycles affects this cube making it allowed. The high value of cube 1Q (48.0) is due to one cage-shaped substrate (44a) with a priority value of 94.9. Another substrate (37d) occupying the cube has a priority value of only 0.27. It is suggested that this cube also belongs in a boundary area. Cube, OM, forbidden in the earlier model, is a boundary area with average priority value, 0.8 in the present (col3) model. The cube contains two atoms of 2-ethyl cyclohexanol (2b) and its sulphur analog (14a). Cubes 1H, 1N, 10 and U are allowed in the earlier model and each have an average priority value of 2 or lower 0013. They are occupied by atoms of bulky substrates. In the earlier model cube OJ was limited access, while In the present model its average priority value is high (12.0) and it contains 15 atoms of as many substrates. Cube IP, with a priority value of 5 (occupied mostly by bulky substrates), is a boundary area In both models. Cube OO, with a priority value of 17 contains bulky substrates (38c, 30a). Usually areas above cubes OM to OR are forbidden in col3. This volume was not defined In the Jones model. According to the Jones model regions below the origin reflect the character of cubes immediately above them. Thus, those cubes below limited or allowed cubes were assumed to be limited access. Cubes, 2D-2, 2D-3, 2E-2, 2E-3 and 2E-4 should then be limited although they are forbidden in col 3. The average priority value for cube 1M is undefined with the data we used. It could belong to a boundary area as does the neighbouring cube IN. Cube 2B-2 could be allowed if more data were available. Cubes 1F-2 and 2uE have low priority values and are suggested to be in a boundary area. Bridged substrates, 30a, 30b and 31 effect some cubes in layer - 1 , making them allowed. It is possible that these substrates change the orientation of amino acid side chains or coenzyme near this layer to allow binding. The cubes 3bA, 3bG, 3C"2, 3C-3,

510 2bA. 2A-1, 2C-2, 2C-3, 2B-3, 1B-2, 1A-2 and 1A-3 are occupied by atoms of phenyl groups of one or two substrates (8,12, Figure 16). Some structures were examined which were not included in the cubic section model used for testing. For example, linear and branched 3-alkyl cyclohexanols can change some forbidden cubes in col 3 to allowed. Also 4-alkyl cyclohexanols were found to extend the allowed area In layers 2, 3, and 4. These substrates were not used in the construction because they were studied under different reaction conditions or no values of enantiomeric excess were available.

Addition of substrates containing

hetereocyclic bicyclic rings (70-73) would change cubes OaO, 1aN, 1aP, 1aQ, and 3E to allowed (Figure 16). After construction of model col 3 wherein the origin was on the corner of a cube It was of interest to visualise the effect of placing the origin on the center of a cube (eel3, Figure 17).

This process was executed for the same set of structures used for

construction of the corner centred cubic sector model. Placing the origin at the center of a cube alters border areas and changes the average priority values on most cubes as well as the number of substrate atoms in each cube. Appendix 4 gives the printed list of cube priorities for the center cube origin model (ce13). The effect of cube size was also studied. Thus, the cubic section model was constructed with the same set of substrates as model co13 assigning the origin to the corner of the center cube but with a cube size of 0.65 A. The printed list of the cube priorities for this model (co065) is in Appendix 5. Altogether 14 different layers were defined. A combined view of both co13 (1.3 A) and co065 (0.65 A) models allows refinement of the 1.3A model (Figure 18). For example, only half of the cubes OK, OH, IK, I E , 2E and 2K allowed in the 1.3A model are only allowed in the 0.65 A model. The cubes OH, ON, I N , and 1H are regarded to be on a boundary area. Jones assumed that the right half at the cubes OB, OH, and ON were limited and half were forbidden. The present analysis allows clearer definition of the border of forbidden and boundary areas.

511

3.3^

o 28'

layer 5

p

61^

layer 4

23«

6^

3

61^

r

21

6l2

18«

l'

4^

5*

0.01

23^

0.22|

10^

"^

4'^

15 1 11^= 3='

iiH 9«

o

71 12H 4^

12 1

^^0.24^

2

4^

16

4'

16

2

28'

layer Z

layer 3

^ 4 11 ^0.5 11 0.5= 52 ^

1

11

12 5 8 4 0

15 6

p ^

3 P

11 J2

1 4

layer 1

layer 0

512

59

O

iQ> .55 10

layer-2 layer-! I

I allowed

^ ^

forbidden

o=origin

Figure 17. Cubic section model (ce13) of the substrate binding donnain of HLADH.

0

layer 4

layer 3

513

•ii

layer 2

layer 1

layer -1 layer 0

allowed forbidden

D

forbidden position in 0.65 A map

unknown

layer -2 Figure 18. Cubic section models of the substrate binding domain of HLADH using 1.3A (col3) and 0.65A (co065) cubes.

514 6.6.7.

Testing of the Models

All twelve cubic section models listed Table 6 were tested to determine which constituted the best model for prediction of substrate reactions with HLADH. The effects of the origin placement, cube size and structure of substrates were tested. In some analyses compounds that had been previously used in construction of the models were used. If this were the case the entries relating to these products were individually removed from the model before a prediction was executed. For new products the minimization, orientation and alignment processes were conducted before execution of a prediction. The PREDICT option (mouse driven) gives the average priority of cubes occupied by each atom of the test molecule and identifies those atoms in cubes of undefined and zero priority. PREDICT (pressing the second button of the mouse) will write the above information to a file with the same file name as the molecule file but with the extension '.pre'. Another way to visualise the results is to use the cubefile graphic display. The positions of each atom of the test molecule are seen as white circles in the individual layers of the model. The results of prediction process were analyzed using both the data in the prediction table (average priority value of each cube, average priority value of cubes over molecule, forbidden sites and unknown sites) and the graphic display. Generally, if the average priority value is quite high (approximately 10 and over) and no forbidden cubes or no occupied cubes with low (A<1 ) priority are present the test molecule will be formed rapidly by HLADH. If there are two atoms of test molecule In a boundary area (A< 2) formation is slow or does not occur. Figure 19 shows an example of a prediction table and a display of positions of the atoms of a test molecule in the cubic section model (co13). The test molecule is known not to be formed by HLADH reduction (the priority number is zero). Although no forbidden cubes are occupied one occupied cube has a very low priority value, 0.24. This suggests that the test molecule would not be expected to be produced by HLADH. The cubic section map for this molecule further showed that some carbons are near to forbidden cubes. It is usually better to check all visualisation methods when using ENZYME for prediction of reactivity, in practice it is found that the full prediction table gives the most accurate prediction. The cubic section model, co13 (Table 6) was used for prediction of the relative rates of formation for all 73 products in their different conformations listed in Table 3. Of these,

515

the more rigid were used in construction of the model but the less rigid (28) were not. For example. 3- and 4-alkyl cyclohexanols and heterocyclic bicycllc alcohols (Table 4) were not used in model construction. Because of the huge amount of data generated only twenty of the more interesting prediction results have been presented (Table 8). OH

layer 2 Si

IC H

9b

C I C -4-0—h-

layer 1

minimized structure

Type hoh.pre Substrate Original Priority: 0.00 Original State: forbidden ATOM ATOM # X

m - < »iiM

layer 0

H

layer -1

C 0 0 0 0 S O H C H

1 2 3 4 5 6 7 8 9 10

Y

Z

-1 1 0 0 0 0 -1 0 -2 -1 -1 -2 0 -2 0 -1 -2 1 1 -1 -2 0 0 1

CURRENT TOTAL AVERAGE #HITS PRIORITY PRIORITY 13 55 55 55 15 16 55 45 3 55

67.2 567.5 567.5 567.5 166.2 111.4 567.5 482.0 0.7 567.5

12.1 10.3 10.3 10.3 11.1 7.0 10.3 10.7 0.2 10.3

Total # atoms in molecule: 10 Average Priority of cubes over molecule: 9.3 Total # hits in cubes: 367 Average Priority for cubes aver all hits: 0.2 Number of unknown sites: 0 Number of forbidden sites: 0

Figure 19. Prediction table and part of displayed cubic section map for a test stoicture. The model col3. Only the forbidden cubes close to product atoms have been assigned.

516 Table 8. The prediction results for twent]^ test molecules using model co13. P=prlorlty A=average forbidden Test molecules Unk/Low 1 (expt) priority cube priority cube OH

.

^

,1,

10.1

OH

0

9.1

OK

0

9.1

OB

0

9.8

OE

0.05

8.5

OaM

0

7.6

OR.OaR

-2bF-'«=unk

0

7.6

0F,-2F

-2bA-i=unk

8.0 11.4

0A.-2A

0.9 OH OH

OH

OH

2.

(;V'^(46)

OH

OH

OH

3.

^

,S4, 0 52.67 OH

9.2

2E-1

0.03

10.0

2E

0.97

10.8

0 OH

\ n

517 [ T e s t molecules

P=priority (expt)

1 forbidden cube

A=average priority

Unk/Low 1 j priority cube

65.5

9.5

1A

2bA=unk

0

8.3

IF

2bF=unk

0

8.4

3E-2

2E-^

1.6

8.7

14.8

11.2

0

9.1

OH OH

^

3B"^=unk

OH

2F

2E

OH

1.2 0.02

120

10.6 1

8.6

1

2uE

2M=^unk

8.1 (28.8)

Ph

1

_''* AH

0

6.2

2F,3bF,3bL,2R

27.5

6.6(15.9)

5 cubes unk

0

6.5

5 cubes

518 A=average priority

P=priority (expt)

Test molecules

forbidden cube

, Unk/Low 1 priority cube

OH

32.6

13.9

0

9.3

12.4

10.9

0

9.3

2E-1

9.5

OH, OM

OH OH

'

OH OH

2F

y w '"' 1

1.0

1

^ OH

0

8.5

OR

0

8.9

-IB,-2A

0

8.5

-1E.2F

^ OH OH

9.

k J (62) OH

28.6 OH

1 -^J^

21.0

11.2 10.2

OK

519 1

[ Test molecules

A=average priority

P=priority (expt)

1

forbidden cube

lOnk/Low 1 priority cube

OH

J^

10.

(58) OH 1.3

7.8

0.06

8.5

3D"^

31, 3J=unk

OH

wax

3J, 2N=unk

(18)

ft

OH

3.0

10.6

0

7.7

3E

.0

1

10.1

1 laQ, laP, laO

0

1

6.7

laO, laP, laN

0

5.6

IF. OF

0

10.0

OA,OB, 1A

0

7.2

OA, OB

0

9.1

OF, OE, OL

(16)

12- k X j H

^

^ OH

1 ^^i

520 Test molecules

P=priority (expt)

00

A=average priority

forbidden cube

0

7.0

1L.0F.-1E

0

7.5

OA,-IB

0

5.2

IL.OF.OE.OB

0

5.7

OA, 08, OE

Unk/Low

1

priority cube

OH

06 06

r- A„'"' A>i,H

23.0

9.8(10.5)

3.0

9.2

U

0

9.4

U , OH

0

8.2

OP,OK

33.6

7.3(12.8)

-1H=unk

4.1

7.8(8.1)

-1 K=unk

0

9.3

U , IE-1

0

8.7

-IB-1

OH

A™ A^H MI

JO="

AV"

(30)

^

^

1

A:>OH

1

521 1 Test molecules

[

P=prJority (expt)

[

A=average [ priority

forbidden cube

rUnk/Low priority cube

r ^^''^ H O ^

1.5

8.5

1E'"'«unk

^

1.0

9.3

1B''*=unk

H

U

0

10.1

OH

0

8.9

OK, OP

13.6

9.7

IP

1.1

10.4

10

^

^ OH

17.

(42)

OH

OH

OH

L.

^

,44, OH

®

94.9

7.3(14.6)

1Q

5.1

9.9

1N.2E

522 Test molecules

A=average priority

forbidden cube

Unk/Low priority cube

(40)

19.

20.

P=priorJty (expt)

6.2

9.2

OH, U

1.3

9.6

OP, OH, 1H

9.7

IK, OK

9.4

0K.2K, U

x:b^

HO'

(70)

CO

91.3

11.5

laP, laQ

0.8

7.6

3E

81.6

7.2

OaO, laN

16.3

11.2

HOT'

xx>

Ha

For 2-methyl cyclohexanone (1), the first case in Table 8, the average priority value allows one to easily predict the main product of the reaction as the (IS, 2S) isomer, although one atom of the substrate occupies a low priority cube (OH).

The

enantiomeric (1R, 2R) product may be formed in small quantities, but formation of the other diastereoisomer (IS, 2R) and its enantiomer (1R, 2S) are clearly less likely due

523 to occupation of forbidden cubes. Our results agree with Jones prediction results for 2alkyl cyclohexanols (Figure 11). For test molecules cited In entries 2-4, 9, and 10 the assigned priorities are more difficult to compare with experiment because different reaction conditions were used in measuring their relative rates of formation. In entries 2 and 4 although many cubes of unknown priority are found, the calculated average priority values quite cleariy predict the relative proportions of products. For entry 10 cis/trans isomer ratios cannot be predicted. Only for 4-methyl cyclohexanols (entry 9) do the priority values predict cis/trans isomer ratios In agreement with experiment. The model (co13) accurately predicts preference for product formation from heterocyclic substrates with alkyl substitution. Good examples are entries 5, 7, and 8. If a phenyl group is appended to a heterocyclic ring predictions are less accurate because of the number of cubes of undefined priority occupied by this residue (entry 6). For bicycllc molecules (entries 11,12 and 13) the model accurately predicts which isomers which are not formed. Removal of bicyclo[3.2.1]octan-2-ol, entry 15, from model 0013 renders some cubes in layer-1 undefined priority, thus the priority values calculated for this case do not accurately predict exo/endo isomers. For bicyclo[2.2.1) heptan-2-ol, entry 14, priority values do accurately predict exo/endo isomer > 1 . Bicyclo[3.2.1]octan-2-one may effect the orientation of some amino acid side chains and the coenzyme which allows it to react with HLADH. The stereochemical course of reductions of some cage-shaped molecules (entries 16-19) is not well predicted. For entries 16 and 19 the products which are not formed (priority value is 0) occupy many low priority cubes. The major product with priority value 91.3 in entry 20 (new substrate) has a higher average priority value compared to other conformations but two atoms are in forbidden cubes. Based on this it can be assumed that these cubes could be allowed for some substrates instead of forbidden as shown in model co13. Models in which the origin was moved to the center of the center cube, the cube size was reduced to 0.65 A and hydrogens were added to the compounds used to construct the model and to the test compounds were next examined with 15 substrates to determine if improvements were achieved by these model modifications (Table 6, 9 and 10). Table 9 shows the predictions made by models of 1.3A and 0.65A cube size with the origin at the corner of the center cube with and without added hydrogens. Table 10 shows results obtained using the same models constructed with the origin in the center of the center cube. The average priority values of cubes over the substrates as well as the number of unknown sites and forbidden sites obtained for each substrate from each model were compared with experimentally derived priority values.

524 Table 9.

Prediction of acceptance by HLADH using cubic section models with 1.3A

or 0.65A cubes with origin at the corner of center cube, with and without added hydrogens. Average prority value, unknown and forbidden sites. Test molecule OH

P(expt)

32.6

co13

CO065

0013*^ C O 0 6 5 * co13h i 5o065h

13.9,0,0 11.5,0,0

14.4.0.0 11.7,0.0

11.4,0.0

10.5,0,0

9.3.0,1

9.2.0,1

8.0.0,0

7.9,0,3

OH OH 0

n

12.4

9.5,0,1

9.4,0.1

10.9.0,0 11.1.0.0

11.0,0,0 11.1.0.0

11.1.0,0

11.2,1.0

9.3.0.0

9.2,0,0

9.9,0.0

7.5.0.0

OH 0

9.5.0.0

9.4,0,0

OX'"' OH

J^

10.6.0.0 10.3,1.0

10.2,1.0

9.3.5,0

0

7.7.0.1

7.5.0.1

7.2,0,3

8.0.0,2

7.2.0.7

0

10.1.0.3 7.6,0.4

10.1.0,3 7.3,0,4

6.9.0,8

5.5.0,14

0

6.7.0.3

7.5.0.3

6.6.0.3

7.4,0,3

5.9,0.7 ! 4.6,0,13

33.6

7.3.1.0

6.8.3.0

7.5.1.0

6.9.3.0

5.0.0,0

3.7,10.0

4.1

7.8.1.0

7.8.3.0

7.5,1.0

7.5.1.0

6.5,0.0

3.7,12.0

0

9.3.0,0

7.2,0,4

9.1.0.0

6.8.0.4

8.2.0.0

6.0,0,6

8.8.0,0

5.6,0.4

8.8.0.0

7.5.0.3

(30)

/Kk°"

AT 1

10.6.0.0 10.4.1.0

3.0

1

^

/k^H

0

5.3,0.4

8.4.0.0

6.2,0,0

1

525 P(expt^

Test molecule

co13

CO065

co13*

CO065*

co13h |:o065h 1

^=fV^ (41) 1—1 ^

9.7

11.0.0,0 16.6,0,0

0.3

8.1,0,0

11.0,0,0 17.1,0,0

9.5,0,0

10.2,3,0

7.8,0,0

10.2,0,0

6.5,5,0

12.4,0,1 5.5,0,3

9.5,0,0

4.1,0,7

8.9,0,0

9.6,0,0

7.0,0,5

n 6.7.2,0

6.3,2,0

^ 0

12.1,0,1 5.8,0,3

^

1

*

_ X)H

1

0

1 9.0.0,0

9.2,0,2

9.1,0,3

J

In the first entry of Tables 9 and 10 predictions agree with the original model (co13), except for the enantiomer of the minor isomer (priority 0). For this compound ENZYME gives one or more forbidden sites for models coded co13h, co065h, ce065, ce065* and ce065h. In models co065h, ce065, ce065* and ce065h (Table 6) the priority values obtained do not accurately predict isomer preference. Entry 2 In Tables 9 and 10 is a good example of a case in which priority numbers accurately predict the outcome of reduction. For the bicyclic substrate (entry 3, Tables 9 & 10) the interesting result is that models with a 0.65 A cube size give many forbidden sites for endo products. In the case of cube file, eel 3 the same result was found. Many forbidden sites also are given for the enantiomer with a priority number of 0 in entry 4. We conclude that there are no significant differences between the predictive value of cubic section models wherein the origin is at the center or the corner of the center cube. Reducing the cube size from 1 .sA to 0.65A allows more accurate predictions to be made because positions of atoms in test compounds are more accurately defined. The presence or absence (coded with *) of a hydrogen on oxygen does not significantly Influence predictive value of the models. Thus, model col 3 can confidently be used to construct the shape of the substrate binding domain for HLADH and is probably the best model format for other dehydrogenases.

526 Table 10. Prediction of acceptance by HLADH using cubic section models with 1.3A or 0.65A cube sizes with the origin at center of the center cube, with and without added hydrogens. Average priority values, unknown and forbidden sites. P(expt) co13 C O 0 6 5 C 0 1 3 * C O 0 6 5 * co13h )o065h Test molecule OH

6."' 32.6

11.2,0,0 10.8,0.0

11.3.0,0 10.9.0,0

11.1.0,0

10.9,1.0

9.6,0.1

9.6,0,1

9.5.0.2

10.8.0,2

OH OH 0

12.4 '

10.4,0.1

10.4.0.1

10.5.0.0 10.8,0.0

10.5,0,0 10.9,0.0

10.9,0.0

11.2,0,0

9.7.0.0

9.6.0.0

9.5,0,0

9.4,0,3

11.5.0.0 10.1,1.0

10.2.2,0

10.1.2,0

OH OH 0

9.1.0.1

9.0.0.1

ccrH

"

OH

^

3.0

OH

11.4.0.0 10.6,1.0

0

7.7,0,3

7.9.0.3

7.5,0.3

7.6,0,3

8.5.0,5

7.1.0,8

0

7.2.0.4

5.4.0.4

6.9.0,4

5.0.0.4

8.7,0,5

5.5,0.11

0

7.2,0.3

5.4.0.3

6.9,0,3

5.0.0.3

8.1,0,5

7.5,0,9

33.6

8.2.0.0

6.2.1.0

8.0,0.0

5.8.1.0

6.3.3.0

7.2.6.0

4.1

8.5.0.0

6.6.2.0

8.3.0.0

6.2.2,0

7.3.3.0

7.5.6,0

0

9.6,0.2

7.0.1.1

9.5.0.2 ' 6.7,0.2

9.3,0,0

5.6,0.3

0

9.1,0.2

5.3.1,1

9.0.0.2

4.8.0.2

10.0.0,0

5.4.0,0

kr""' A>™ 1 1

\

1

h j^H

527

PrexDt^

Test molecule

co13

CO065

C013*

CO065* c o 1 3 h

bo065h

^^fY <*^) •-SOH

1—1

-S;

9.7

10.6.0,0 6.4.3.0

10.6.0,0 6.1.3.0

8.3.0.0

7.3,5,0

0.3

8.4.0.0

6.0.3.0

8.3.0.0

5.6.3.0

8.2.0,0

4.9,5,0

0

6.8.0,1

6.1.0.2

6.5.0.1

5.8.0.2

8.2.0,0

5.2,0.5

0

8.6.0.1

8.1,0,2

8.5,0.1

7.9.0.2

9.1,0,0

6.2.0.6

^

^

1^ 6.6.8.

Comparison of Present Models with X-Ray Crystallographic

Structure The X-ray crystallographic structure of the ternary complex, HLADH-NADH-dlmethyl sulfoxide, DIVISO, has been determined to 2.9 A.34 The crystallographic co-ordinates available from the Brookhaven protein data bank were entered into MACROMODEL and the amino acid residues surrounding the active site identified. Using MACROMODEL and ENZYME the substrate binding domain of HLADH was mapped as a cubic section model. The model obtained from this approach were compared with model col 3 obtained by substrate surrogate overlay. In the comparative study X-ray crystallographic data extending 15 A around the catalytic zinc atom was employed. Using MACROMODEL the sulphur of DMSO in the X-ray structure of the HLADH-NADH-DMSO complex was changed to a carbon reformulating this component as a 2-propanoxy anion with the carbon-oxyanion bond along the previous S-O axis and the carbon-hydrogen bond along to sulphur-lone pair axis. This alcohol structure positioned so that the oxyanion was located 2.2 A from the zinc. Distances between the substrate atoms and the HLADH-NADH residue atoms were determined using MACROMODEL. Table 11 shows the distances between the oxyanion (1), central carbon (1) and the neighbouring HLADH-NADH atoms within 4 A In this complex compared to similar distances for the HLADH-NADH-DMSO and

528 HLADH-NAD-p-bromobenzyl alcohol complexes. The locations of oxygen (1) and carbon (1) In the constructed 2-propanoxy anion-NADH-HLADH complex compared well with analogous distances in the other complexes. The distance from oxyanion (1) to the side chain atoms of HLADH differ only 0-0.5 A between the 2-propanoxyHLADH-NADH and DMSO-HLADH-NADH complexes and 0-0.6 A between the 2propanoxy-HLADH-NADH and PBBA-HLADH-NADH complexes. The location of

Table 11.

Interactions of three ternary complexes of HLADH and NADH. For the pbromobenzyl alcohol complex both observed and productive distances have been given 36, HIadh-Nadh ours HIadh-Nad1 -Dmso(A) Pbba(A) atom atom residue (A) 01

Zn Cys-46 Ser-48 His-67

Cys-174 CoE

C1

Zn Ser-48 His-67

SG OG CB CE-1 CD-2 NE-2 SG C4N C5N C6N

2.2 3.7 2.7

2.2

2.2 (2.2P)

3.6

3.4 (3.5p)

3.7 3.3 3.2

3.2 3.7 3.9

3.3 (3.4p) 3.8 (3.8p)

2.6 3.4

3.1

3.2 (3.2p)

3.6 3.0 3.2 3.8 2.9 3.3

3.5 (3.5p) 3.6 (3.5p) 3.4 (3.5p) 3.9(3.8p) 3.0 (3.1p) 3.4 (3.7p)

3.8 5.0 4.4

3.4

4.9

3.6 (3.6p)

3.6 OG CB CE-1 NE-2

'

3.7

CD-2 Phe-93

CE-2

3.5

4.2

Leu-141

CD-I

3.5

4.3

CZ

3.5

4.1

3.5 3.1

(3.5p)

Cys-174

SG

3.4

(3.6p)

CoE

C3N

3.2

(3.6p)

C4N C5N

5.1

2.4

(3.4p)

3.2

(3.7p)

1

529 carbon (1) differs 0.6-0.8 A In the case of DMSO-HLADH-NADH (except CN4 of NADH where a 2.7 A difference was obtained) and 0.1-1.3 A in the case of the PBBAHLADH-NADH complex from 2-propanoxy-HLADH-NADH complex. The embedded 2-propanoxy group the HLADH-NADH-2-propanoxy complex was oriented as previously described substrate surrogates using MACROMODEL Carbon (1),a-hydrogen and oxygen (1) of the 2-propanoxy anion were used to align the complex, other atoms of this product surrogate were deleted. The MACROMODEL. data file thus created was transferred to ENZYME using MMFORMAT. The complex was aligned as previously described for substrate surrogates and a cubic section model was created with a priority value of zero for all occupied cubes. The coordinates of cubes and atom types of cubes in this model were obtained by using PREDICT-and PRINT CUBEFILE-opWons. It contains 24 layers from-13 to 10. A printed version is available from the authors. We next determined the location in the cubic sector model of those residues lining the substrate binding domain of HLADH. Using the ANALYZE and DISPLAY modes of MACROMODEL (FINDNAME RESIDUE -options, MOLECULE, SET 1 and MONO -options) the residues near the active site pocket were identified and marked. Each marked residue (e.g. Phe-93) was separated from other residues using the DELETE ATOM / BOND command. Each separated residue was connected to the surrogate product to create a rigid complex as needed in the alignment process of ENZYME. Each of the files created by these operations was stored individually. The following residue files were formed: His-67, Cys-174, Cys-46, Leu-57, Ser-48, Phe-93, His-51, Leu-141, Met-306. Phe-110. lle-318, Pro-295, Val-294, Thr-94, Leu-92, Pro-91, Pro-95, Asp-49 and Val-52. These residue files were transferred to ENZYME, using MMFORMAT. The DISPLAY CUBEFILE option was used to determine the position of each residue within the cubic section model generated by ENZYME (shown as white circles each occupied cube). The co-ordinates of atoms in each residue were obtained from this process by creation of individual files for each residue and each of these was converted to a cubic section model. Addition of the residue cubic section models gave a cubic sector model of the substrate binding domain consisting only of the residues lining the this pocket. Model 0013 and the model constructed with all hydrogens in substrate surrogates (co13h) were compared with the cubic section model derived from display of the cubes occupied by residues of HLADH lining the substrate binding domain. Using

530 COMPARE the X-ray based model was overlaid with the substrate surrogate derived models.

Forbidden cubes in c o 1 3 . c o 1 3 h and the X-ray derived model which

overlaid were identified and displayed (Figure 20). The co-ordinates of each cube and atom types in each cube were also identified. This allowed location of the residues lining the substrate binding domain that interfered with substrate binding. Only layers between -3 to 4 have significant overlap between the models. We commenced analysis of layers at layer -3, then proceeded to layer -2 and so on. The coenzyme, NADH was located in the front layers. The forbidden cubes in layers -3, -2 and -1 are close to the coenzyme and residues Phe-93, lle-318, Cys-174 and His-67. Residues Leu-309, Val-309 and Thr-94 with attached hydrogens may effect layer 0. The coenzyme and Phe-93 are the closest to the cubes -1H, -1K and -10 In model c o l 3 . For some bridged substrates (30, 31) these cubes are allowed in this model. In layer 0 the residues Leu-309, lle-318 and Val-294 affect the cubes above OM to OR and may be In the boundary area. Residues Phe-93, lle-318 and possibly Hls-67 affect the forbidden cubes on the right side of model co13. Val-294, Ser-48 and coenzyme are the residues nearest to the forbidden left edge. According to the Xray derived cubic section model the left side of the substrate binding domain (in layers 1-4) Is more open because Val-294 and Leu-309 can change orientation. In layer 1 Phe-93, Thr-94, His-67 and lle-318 affect the forbidden right side and lle-318 affects the cubes above ON to OQ in model co13. Ser-48 and His-51 are quite close to the left edge of co13. This edge is assumed to be the boundary area. It is forbidden for some substrates and allowed for others. In layer 2 the right edge of co13 Is affected by Hls-67, Phe-93, Leu-116, Leu-141, lle-318 and Phe-110. The Van der Waals interactions from other layers (1 and 3) have an especially important role in layer 2 definition. His-51 and Ser-48 may affect the left edge, an area that seems to be allowed for substrates with phenyl groups. The upper part of the left side of co13 is the most open with Val-294, Leu-309 and Pro-295 as the nearest residues. The same residues affect layers 2 and 3. Whereas residues Leu-57 and Phe-140 affect layer 4. The cubes in the upper edge of the left side of layers 2 and 3 are defined from larger substrate surrogates. They may affect orientations of the side chains of Val-294 and Leu-309. Leu-116 and Met-306 have also been suggested to change their side chain conformations with more bulky substrates (38) however, Met-306 Is quite deep In layers 3 and 4 In present X-ray based model. Comparison of the shapes of the substrate binding domain derived from these two approaches provides a good view of the domain.

531 Jones manually derived positions of amino acid residues in the apoenzyme structure and some ternary complexes published before 1977. Figure 21 shows a comparison between the present results with the earlier definition. The boundary region of model (0013) has also been assigned in the Figure. Ser-48, Phe-93 and lle-318 are in the same location while Leu-57 is in layers 3, 4 and 5 In our X-ray derived model. Using the present method positions of residues were positioned more accurately.

(H)

(C)

(N)

(N)

nil [TJ G

(Ip) CoE (C,H)

(0)

(Ip)

CoE

CoE

CoE

A

0

P

Q

Tn

H

1(H) J

K

L

B

CoE D C CoE^)

E

F

uE

uF

uA(C •uB(C) uC(K) uD |coE CoE CoE

(C)

(S) C174 (H) (C) (H) C174 a? A

(C)

(H) C-46

layer -3 (C)

CI 74 stoiTi type (C) and the amino acid residue, Cys-174 forbidden allowed for some 3-alkylcyclohexanols or heterocyclic bicyclic alcohols, usually forbidden allowed for 4-alkylcyclohexanols, usually unknown Figure 20. The overlaid layers with X-ray based model and present models oo13 (upper) and oo13h (lower).

532

(0) 1-318

(0)

1-318 (H) (H) CoE 1-318 (C)

(H)

(0)

(H) CoE

F M ^ ^ G (C) CoE BCoB

[uA

(C) CoE CoE

(N) CoE

acog adoB J

"(C) '(C) CoE CoE C D B

uB

(H)

(H)

o u C ^ uD

1'1 K 1

E

uE

""(Hi P-93J uF

(H)

|p-93 (H) CoE (0)

(H)

(C)

(H)

C-46 (S)

C-46 (H) (C) C-46 C-46

layer - 2

(H)

(H)

533

RcJo

pcoTl

1-318 (C) h-318 1 (C)

pi (H) CoE

G

[0 1 0.02 J H 1 1 34

C

A

(Ip) CoE

(C) ! uA CoE

(0.1 P: qoE

TH

[N

NJB

K 1 L 4

p 55 ^ ^

uC ^ uD 45 11

uE

uF

(C,C) |P-93 (H) P-93

10 9 (H) H-67

(H)

(c)

(C)|

(N) H-67

(C) (C,H)i (N)

(c) 1 (H)

(0)1 H-67

C-46 C-46

layer -1 (C,H) 1-318

(0)

(C) 1-318 (C)

(H) CoE

[G A

(Ip) CoE

0 2 P 2^ ^ R 3 17 JHIC 1 9 J 6 K 3 2 4 4 16 D 55 E 3 B 2 c 2

(C) CoE

Ko.ip;

12

^ ^ u C

^ uD 1 uE

44 11

10 8

JP93| KC^ P-93

WPS

(H)

(H) H-67

(C) (C.H)

mi

1 I

(C)

1

(N) H-67

(N)

(C) (H) C-46 C-46

(0) |H-67

534

layer 0

535

layer 1

536

537

ric)

(H)

rw

M-306 M-306

(H)

P-110

P-295

p-no

(H)

(H) V-294

[ M ^ pT"

PTHT \Q

0

[R

1

L-116

H G 61'^ V-294 H-294 2 A "Z B 'A 61 11 31 (H)

uA

c

ZTD 1

•(d) P-93

A^yV.

uB

rTc)

u E ( H ^ ^ a (H) (C) L-141 (H)

1 28

H-51

1 28 , (H)

(0)

H-51 (N)j

S-48 (H)|

H-5l!

A-49

L-141

(Ip.O)

{H)| P-14(^

layer 3 (C)\

(C)

M-306 M-306

p-no (H)

(H) (H)

p-no

P-29S 1 (C) |L-n6

11 121 (H) V-294

IM

N

\ 0

R

^J»57

(C)

3 4

(H) H-51 (N) H-51

1

1.3 H 4 1

G 5J 4 61^ 61^ 7 1 25 ^ ^ ^ 4 V-294 H-294 4 A 7 B 13 C 5| D 8 E 4 1 121 46 17 16 1 1 8^ 0.6^ 19 10 2 2

H-51

^P

28^ 0,48

"loTJ uF 1 m

p-93

^1

(C) l " 2 16 ^ ^ 0.22 L-141 (H) 1 28 L-141 1 4

(H)

(H)

A-49

P-140

moi

538 IHTI

(H)

(H)

M-306

M-306

p-no

(H) L-116

(H) L-ne

[M

(H)

(H)

N

0

P

1

•

G

H

A

B (H

>oo?

uA

L-57 uB(C) L-57

L-57

lr-57

Q(C) L-ne K

D

"Wl

(H)

"mlflcT 1 L-n 6 1-116 L

F(H)

(H)

L-14l| | L - 1 4 1

^uD

(H)

• 1

(H) L-141 (C)

L-141

H-51

(C) L-141

(C)

(H)

(C,H)

(C,H)

H-51

H-51

P-140

L-141

(H)

(C) 1 ^^^ P-140

H-51 (H) V-52

layer 4 (H)

(H)

(H)

M-306

M-306

p-110

(H) L-116

(H) L-ne

\IA

(H)

(H)

r^

0

H

A

B 1 C 1 3.5 3.0 H,57 uB(C]

uA

L-57|

P

K

1

^ ^ ^^^ ^^^

L

F(H)

(H)

L-14l| L-141 uF

(H)

1 57K

L-141 (C)

(H)

(C)

L-141 L-141

H-51 (C)

(H)

1 L-116 L-116| [1-116

G

L-57

(HTI

(H)

H-51 H-51

(C.H)

(C,H)

P-140

L-141

(C)

(C)

H-51

p-140 (H) V-52

(H)

539 Jones model 309

[306

3od sial am

p!95

cW\ ^

294 294 294

layer 2

Ju r

mo

>^;

294

110

51

1''

^

51

1411 48

48

I 93

1141

48

HO]

48

110^

309

309

bisj

295

318 318

i94^H

294

layer 1

94

294 93 MM

k^

1

1

51

48

309

48

no^

bisl

309

318

309

3ji

COEI

iK

c^

S — < ,—1

• H ESil JS? mm w^m mm ^mm >S$^ layer 0

u

93

CoE

|H-5J

CoE

^

M

—<

MP

48 48

-^^^ . . • torDidden [ ^ BMhJ or limited

93

93

48 67

, „ „. allowed boundary \ ^ in some KX^ ^^^^«

94

^Mji"

[2*94^^

294

^.^.^^

Ur

67

^

1

forbidden

Figure 21. Comparison of Jones map of amino acid residues with present definition.

540 6.7.

CONCLUSIONS

The present computer-based substrate surrogate overlay method gave a refined cubic section model for the substrate binding cavity of HLADH. Allowed and forbidden regions were accurately defined using the origin in the corner of a cube and a cube size 1.3 A or 0.65A (model co13 and co065). Boundary regions (Figure 21) which can be either allowed or forbidden were defined. The average acceptance value in these regions was usually 2 or lower. Low average priority values for some cubes (Figure 16: OG, ON, OQ, OK, 1 G, 1 K, l u E , 2K, and 2E) are due to the inclusion of bulky substrate surrogates with low priority values in construction of the model. These are assumed to change the orientation of the side chains of amino acid residues lining the substrate binding domain. Reducing the cube size to 0.65A (co065) from l.sA (co13) primarily affects cubes OH, OK, ON, OQ, 1K, 1H, and IN in the 1.3A model by defining those parts of these boundary cubes that are forbidden or allowed. The cubic section models, co13 and co065, proved to be accurate in predicting acceptance of substrates and the stereochemical course of HLADH mediated reductions. The average priority value for all cubes, number of forbidden sites and number of low priority cubes occupied by atoms in a possible alcohol are used to estimate the acceptability of the corresponding ketone by HLADH. The prediction method developed can be applied to other enzymes (e.g. other dehydrogenases, esterases, or lipases) if kinetic and specificity data is available. Comparison of the X-ray based cubic section model with the substrate surrogate derived cubic section model gave a view of the amino acid residues around the substrate binding domain (Figure 21). Forbidden and boundary regions defined by the substrate surrogate overlay method are near the side chains amino acid residues. Substrate interactions with surface of the amino acid residues could be estimated by molecular mechanics (MACROMODEL). The close resemblance of the present models, developed with rather sophisticated computer techniques, to the original Jones model, developed with hand held stick models, is a testimony to the ingenuity of the Jones group. 6.8.

ACKNOWLEDGEMENTS The authors acknowledge the financial support of Suomen Kulttuurirahasto,

Neste Oy and Kemira Oy of Finland and the Natural Sciences and Engineering

541 Research Council of Canada. As well the authors thank Mr. L. Corey and Ms. Sharon Cyprik for development of the ENZYME and COMPARE programs and Clint Surry and Dr. All Mohammed for help in preparation of the manuscript. M.A. sincerely thanks Professor H. Krieger for his support and encouragement of her studies in Canada.

6.9. REFERENCES 1. E. L Smith, R. L Hill, I. R. Lehman, R. J. Lefkowitz, P. Handler & A. White, in: R. S. Laufer, E. Warren & D. Mclvor (Eds), Principles of Biochemistry: general aspects, R. P. Donnelley & Sons Company, 1983. 2. C.-H. Wong, Science, 244 (1989) 1145. 3. A. Akiyama, M. D. Bednarski, M. J. Kim, E. S. Simon, H. Waldmann & G. M. Whitesides, Chemtech, (October 1988) 627. 4. J. B. Jones, Tetrahedron, 42 (1986) 3351. 5. G. M. Whitesides & C.-H. Wong, Angewandte Chemie, 24 (1985) 617. 6. M. Schneider, Enzymes as Catalysts in Organic Synthesis, Reidel, Dordrecht,1986. 7. J. B. Jones, in: P. Dunnill, A. Wiseman and N. Blakeborough (Eds), Enzymes and Nonenzymic Catalysis, Horwood/Wiley, Chichester/New York,1980. 8. J. B. Jones, C. J. Sih & D. Perlman, Applications of Biochemical Systems in Organic Chemistry, John Wiley & Sons, inc., 1976. 9. J. Retey & J. A. Robinson, Stereospeciflcity in Organic Chemistry and Enzymology, Verlag Chemie, Weinhelm/Deerfield Beach, Florida/Basel,1982. 10. A. Fischli, in: R. Scheffold (Ed.), Modern Synthetic Methods, Vol. 2, SalleSauerlander, Frankfurt, 1980. 11. J. B. Jones, in: J. R. Morrison (Ed.), Asymmetric Synthesis, Vol. 5, Academic Press, New York, 1980. 12. P. Sonnet, Chemtech, (Fehruary 1988) 94. 13. Enzyme Nomenclature, Elsevier, Amsterdam, 1973.

542 14. K. Martinek & J. J. Semenov, Appl. Biochem., 3 (1981) 93. 15. H. Theorell, in: R. G. Thurman, T. Yonetani, J. R. Williamson & B. Chance (Eds.), Alcohol and Aldehyde Metabolizing Systems, Academic Press, Inc.,1974. 16. C.-l. Branden, H. Jornvall, H. EkIund & B. Furugen, in: P.D. Boyer (Ed.), The Enzymes, Vol. XI, Part A, Academic Press, New York, 1975. 17. H. EkIund & C.-l. Branden, in: T. G. Spiro (Ed.), Zinc Enzymes, John Wiley & Sons, Inc., 1983. 18. H. EkIund & C.-l. Branden, in: F. A. Jurnak and A. McPherson (Eds), Biological Macromolecules and Assembles, Vol. 3, John Wiley & Sons, Inc., 1983. 19. V. C. Sekhar & B. V. Blapp, Biochemistry, 27 (1988) 5082. 20. J. P. Klinman, CRC Critical Reviews in Biochemistry, 10 (1981) 39. 21. J. B. Jones, Enzyme Engineering, 6 (1982)107. 22. A. I. Irwin & J. B. Jones, J. Am. Chem. Soc, 98 (1976) 8476. 23. D. R. Dodds & J. B. Jones, J. Chem. Soc, Chem. Commun., (1982).1080; J. Am. Chem. Soc, 110(1988)577. 24. A. J. Irwin & J. B. Jones, J. Am. Chem. Soc. 99 (1976),1625. 25. I. J. Jacovac, G. Ng, K. P. Lok, & J. B. Jones, J. Chem. Soc, Chem. Commun., (1980) 515. 26. J. Davies & J. B. Jones, J. Am. Chem. Soc, 101 (1979) 5405. 27. M. A. Findeis & G. M. Whitesides, Ann. Rep. Med. Chem., 19 (1984) 263. 28. J. B. Jones & K. E. Taylor, Can. J. Chem., 54 (1976) 2969. 29. C.-H. Wong & G. M. Whitesides, J. Amer. Chem. Soc, 103 (1981) 4890. 30. W. H. Baricos, R. P. Chambers & N. Cohen. Anal. Lett., 9 (1976) 257. 31. A. R. Fersht, in: Freeman (Ed.), Enzyme Structure and Mechanism, New York, 1985.

543 32. R. W. Hay, Bio-inorganic Chemistry, Ellis Norwood, Halsted Press, 1989. 33. H. Ekiund, B. Nordstrom, E. Zeppezauer, G. Soderlund, I. Ohisson, T. Boiwe, B.-O. Soderberg, O. Tapia, C.-l. Branden, and A. Akerson, J. Mol. Biol., 102 (1976) 27. 34. H. Ekiund, J.-P. Samama, L Wallen, C.-l. Brand6n, A. Akeson & T. A. Jones, J. Mol. Biol., 146(1981)561. 35. H. Ekiund, J.-P. Samama & L Wallen, Biochem., 21 (1982) 4858. 36. H. Ekiund, B. V. Plapp, J.-P. Samama & C.-l. BrSnd^n, The J. of Biol. Chem., 257 (1982)14349. 37. E. Cedergren-Zeppezauer, J.-P. Samama & H. Ekiund, Biochem., 21 (1982) 4895. 38. E. Horjales, H. Ekiund & C.-l. Brarlden, J. Mol. Biol., 197 (1987) 685. 39. H. Jornvall, Eur. J. Biochem., 16 (1970) 25. 40. H. Ekiund, Biochem. Soc. Trans., 17 (1989) 293. 41. V. Prelog, Pure Appl. Chem., 9 (1964) 119. 42. A. J. Irwin & J. B. Jones, J. Am. Chem. Soc, 99 (1977) 556. 43. A. J. iHA^in & J. B. Jones, J. Am. Chem. Soc, 99 (1977) 1625. 44. E. Horjales & C.-l. Branden, J. Biol. Chem,, 260 (1985) 15445. 45. J. B. Jones & I. J. Jacovac, Can. J. Chem., 60 (1982) 19. 46. J. M. H. Graves, A. Clark & H. J. Ringold, Biochem., 4 (1965) 2655. 47. M. Nakazaki, H. Chlkamatsu, K. Naemura, Y. Sasaki & T. FujIjI, J. Chem. Soc, Chem. Commun., (1980) 626. 48. H. Duller & C.-l. Branden, Bioorg. Chem., 10 (1981) 1. 49. G. L Lemiere, T. A. Van Osselaer, J. A. Lepolvre & F. C. Alderweireldt, J. Chem. Soc, Perkin Trans II, (1982) 1123.

544 50. R. A. Johnson, in: W. S. Trahanovsky (Ed.), In Oxidation in Organic Chemistry, Part C. Academic Press, New York, 1978. 51. M. Nakazaki, H. Chikarnatsu, K. Naemura & M. Asao, J. Org. Chem., 45 (1980) 4432. 52. J. J. Willaert, G. L. Lemiere, L. A. Joris, J. A. Lepoivre & F. C. Alderweireldt, Bioorganic Chemistry, 16 (1988) 223. 53. J. B. Jones & T. Takemura, Can. J. Chem., 60 (1982) 2950. 54. J. A. Haslegrave & J. B. Jones, J. Am. Chem. Soc, 104 (1982) 4666. 55. T. Takemura & J. B. Jones, J. Org. Chem., 48 (1983) 791. 56. M. Nakazaki, H. Chikamatsu & Y. Sasaki, J. Org. Chem., 48 (1983) 2506. 57. M. Nakazaki, H. Ckikamatsu & Y. Sasaki, J. Org. Chem., 48 (1983) 4337. 58. M. Nakazaki, H. Chikamatsu, K. Naemura, T. Suzuki, M. Iwasaki, Y. Sasaki & T. Fujiji. J. Org. Chem., 46 (1981) 2726. 59. T. A. Van Osselaer, G. L Lemiere, J. A. Lepoivre & F. C. Alderweireldt, Bull. Soc. Chim. Belg., 89(1980)389. 60. T. A. Van Osselaer, G. L, Lemiere, J. A. Lepoivre & F.C. Alderweireldt, Bull. Soc. Chim. Belg., 89(1980) 133. 61. J. Van Luppen, J. A. Lepoivre, G. L. Lemiere & F. C. Alderweireldt, Heterocycles, 22 (1984) 749 62. A. R. Krawczyk & J. B. Jones, J. Org. Chem., 54 (1989)1795. 63. L K. P. Lam, I. A. Gair & J. B. Jones, J. Org. Chem, 53 (1988) 1611. 64. C.-S. Chen, Y. Fujinloto, G. Girdaukas & C. J. Sih, J. Am. Chem. Soc, 104 (1982) 7294.

545 6.10.

Appendices

Appendix 1 : MACROIVIODEL Description After connecting to the VAX and setting the correct sub directory (e.g. set default [site]), the program MACROMODEL is started by typing RMMOD. When the display is shown on the screen, active buttons are shown in green. The first mouse key is shown as the pick key. The second mouse key acts the same for ail buttons except for the DRAW button. The third mouse key will display the help screen for that button. The molecule on the screen will be drawn by using the DRAW button active. The first key draws connecting bonds to subsequent points whereas the second key restarts the drawing of bonds. The hydrogens are added by choosing the H ADD button three times. Non carbons are replaced by choosing the appropriate symbol and picking the atom to replace. Functional groups can also be appended in this manner. The structure is then minimized in energy by selecting the ENERGY mode (choose MM2 force field) and picking start.

Constraints are added by selecting the

CONSTRAINT submode and selecting the appropriate choice of constraint (CATM, CDIS, CBOND, CTOR) and choosing the atoms involved. To orient the structure, pick the ANALYZE mode and pick ALIGN X, ALIGN Y, ALIGN Z, ROTX, ROT / a n d ROTZ as needed. The aline buttons need two atoms picked after the command to align on the axis and the rotate buttons need a typed input on the number of degrees to rotate on that axis. The structure is stored by picking the WRITE button an supplying a filename (the extension defaults to .DAT) and give a blank line for sequence structure number. When finished, exit the program by picking the STOP button and confirming that you wish to exit. Answer "Y" to the question for deleting the log file.

546 Appendix 2:

ENZYME and COMPARE Documentation

To execute the ENZYME program, type RUN ENZYME and the graphics screen with a menu is displayed as in Figure 12. The choice of options are: READ CUBEFILE, READ DATA, ADD TO CUBE FILE, REMOVE FROM CUBEFILE, ALIGNMENT, ROTATE ON X AXIS, ROTATE ON Y AXIS, ROTATE ON Z AXIS, PREDICT PRIORITY FOR MOLECULE, DISPLAY CUBEFILE, ROTATE ABOUT BOND, PRINT CUBEFILE, SAVE CUBEFILE, EXIT, CHANGE CUBE SIZE, CHANGE ORIGIN IN CUBE, and DISPLAY CUBE WITH DIFFERENT PRECISION. The desired option is chosen by depressing a mouse button while pointing the cursor at the option and the exact operation to be performed is dependent on which of three buttons was pressed. 1. Read cube file {READ CF): This option asks for the cube file to read; the extension .cf will be added to the file name. 2. Read data (READ DATA): This option reads in the data generated by MACROMODEL program (the extension must be '. dat' and the file must be in ASCII format which is made by running MMFORMATor\ the file made by MACROMODEL and supplying the filename) and asks for the necessary information to determine its priority. The first button will direct the program to follow this section by the alignment section; the second button directs the program to return to the menu. 3. Add to cube file {ADD TO CF): This option adds the molecule to the cube file; It first checks to see if the molecule has already been aliened. 4. Remove from cube file {REMOVE CF): This option removes the priority from the cubes that are occupied by the molecule. It does not check the cube file for the presence of the molecule since it is impossible to confirm the previous addition of the molecule. The molecule must have the same priority ranking as it had when it was added. 5. Alignment {ALIGNMENT): The option aligns the molecule in the reference axes with one of the atoms on the origin with a relative co-ordinate of (0, 0, 0). It will ask for the atom to be placed on the origin, the atom to be on an axis which is chosen by one of the buttons, and the atom to be placed in a plane which is chosen by one on the buttons. When choosing this section with the second button, the atom chosen will be translated to the origin and the alignment of the rest molecule is not done. 6. Rotate on x axis {ROTATE ON X), rotate on y axis {ROTATE ON Y), rotate on z axis {ROTATE ON Z): These options will rotate the molecule on the specified axis by an

547 amount which is typed in. Depression of the second button will rotate the molecule on the axis by 180**. All rotations are positive in the right handed co-ordinate system. 7. Predict priority for molecule (PREDICT): This option determines the priority of the molecule based on the information in the cube file. The second button will write the information, and how each atom fits, in a file with the same name as the molecule but with the extension '.pre*. 8. Display cube file (DISPLAY CF): This option displays the cube file by displaying each layer of the cube file starting from the near side. It will either display the cube file with 3 colors or 6 colors depending on the value shown for the last option. The displayed layers will be limited (from -5 to +10 inclusive) if the first button is pressed and will be unlimited if the second button is pressed. If a molecule shown on the screen, then circles within the cubes are drawn to indicate which cubes the molecule occupies. 9. Rotate about bonds (ROTATE BOND): This option rotates a group of atoms the specified bond and returns the orientation with the best fit for the cube file; it does not add the molecule to the cube file. 10. Print cube file (PRINT CF): This creates a file which contains the cube coordinates and the average priority for the cube. It also includes the number of atoms in the cube. 11. Save cube file (SAVE CF): This saves the cube file. 12. Exit (EXIT): This option also saves the cube file if it has been changed and exits the program. 13. Change cube size (CF number): This option allows the size of the cube to be changed. It does not recalculate the cube file to reflect a change in size of the cube so a different cube file should always be used with each cube size. If you wish a size different from the default, specify the change in size before doing any operations on the cube file so the cube file will not contain garbage. 14. Change origin in cube (CORNER or CENTER): This option moves the origin of the axes to the corner of the cube or the center of the cube. The place on the origin is reflected by location of the origin, specify the change before doing any operations on the cube file. 15. Display cube with different precision (DISCOL number): This option determines if the cube file Is displayed In 3 or 6 colors with different ranges.

548 To Start to execution of COMPARE, type RUN COMPARE after exiting ENZYME program. Then choose option READ CUBEFILE and give the name of first cube file and also the name of second cube file. After it choose option DISPLAY CUBEFILE and it gives the overlaid picture of these cube files. The overlaid cubes can be seen with different colors: forbidden in first cube file (black), forbidden in both cube files (red), conflict in first cube file (blue), conflict in second cube file (yellow), and allowed in both cube files (green). In the end of the program it gives the number of conflicts in both cube files.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

549

Applications of ^^O NMR Spectroscopy to Natural Products Chemistry David W. Boykin

INTRODUCTION The investigation of natural products chemistry has been and continues to be an integral part of the development of Organic Chemistry. Of paramount importance to the study of natural products, whether isolated from plants, mammals, microorganisms, marine life or insects, is compound identification or, if a new substance, structure elucidation. Today, natural products chemists have an array of highly sophisticated instrumentation available to perform structure elucidation. Of the more powerful technologies mass spectrometry and nuclear magnetic resonance (NMR) spectroscopy are the most widely used. A combination of the many forms of 1 H and l^c NMR spectroscopy can provide extraordinary amounts of structural information and, arguably, in most cases is sufficient for characterization of new structures. The entire December 1992 issue of Magetic Resonance in Chemistry was devoted to articles using iH and 13c NMR spectroscopy for structure elucidation of natural products and serves to illustrate the power of this approach. Use of NMR spectroscopy of some of the more difficult to detect nuclei such as oxygen, nitrogen and sulfur is rapidly growing [1-3]. l^O NMR spectroscopy is emerging as a useful adjuvant to other spectroscopic methodologies for acquiring structural information. Oxygen, the most abundant element on earth, widely occurs in many types of natural products [4]. Detection of oxygen by NMR spectroscopy can be achieved for the l^O isotope. Because the l^o isotope is quadrupolar, its spin quantum number is 5/2, its natural abundance is 0.037% and its receptivity relative to ^^C is 0.06, it is sometimes considered to be difficult to study by NMR spectroscopy. However, by use of modem instrumentation and wise choices of experimental conditions (pulse repetition time, solvent, temperature and concentration), spectra with acceptable signal-to-noise ratio can be obtained generally in less than four hours (and often in less than one hour) for molecules with molecular weights of less than 300. A representative natural abundance l^o NMR spectrum for the natural product santonin obtained in approximately two hours on a Varian VXR 400 NMR spectrometer at 54.22 MHz is shown in Figure 1. Of course, enrichment with the l^O isotope, even at low levels(ca. 1%), dramatically reduces the instrument time required to obtain high quality data [5]. l^O NMR linewidths increase with molecular weight and currently the practicality of the experiment

550

becomes questionable for many compounds of molecular weight above 500. The purpose of this review is to provide an overview of the l^O NMR spectroscopy characteristics of organic molecules, to give illustrations of previous use of this method in natural products chemistry and , when possible , to offer some suggestions for future development of the methodology.

-CH3

o b

I I M

T I r I t I t T T > I I I T r I I I I I I I > I 1 J I 1 I I I t I I r'"| i t i i | i i » i i i i i t f 7 t t r i

500

400

300

200

100

T

> t

0 PPM

Fig. 1. l^O NMR spectrum of santonin, 0.5M in acetonitrile at IS^Cj internal reference 2-butanone 558 ppm, external reference water 0 ppm.

551

BACKGROUND During the last decade I ^ Q nuclear magnetic resonance (NMR) spectroscopy has emerged as a valuable method for organic chemists concerned with structure and conformation elucidation as well as for those interested in probing electronic distribution [1, 6-10]. Despite some of the problems with detection, the large chemical shift range for common functional groups (ca. 800 ppm) makes I ^ Q NMR spectroscopy useful for detecting subtle changes in structure. It has become well established that l^o NMR chemical shifts are more sensitive to structural changes than those of 13c and 15N [1,10]. Figure 2 illustrates the 17o NMR chemical shift range of some common functional groups.

FUNCTIONAL GROUPS HgO (Internal)

HgO (External)

Sulphlnylimlnes

Anhydrides (c-0) Isocyanates

Anhydrides (C=0)

PiiBH3HEaH 900

800

700

600

•••••• 500

400

300

pUffl^HH 200

TW

-100

17 0 NMR Chemical Shift (PPM) Fig. 2. General ^^O NMR chemical shift ranges for important functional groups.

552 l^O NMR chemical shifts are perhaps best understood in terms of an additive dependency of the paramagnetic, GoP, and diamagnetic, Oo^ screening constants Equation (1) [8].

8o = aoP + aod

(1)

A number of investigations have led to the conclusion that most ^^O NMR chemical shifts are largely dependent upon the paramagnetic term [1,8,9]. The description of Karplus and Pople is often used to describe the paramagnetic term, shown as Equation 2, for an oxygen nucleus (O) bound to another nucleus (X) [11]. GoP = -e2h2 / 2m2c2AE (r-3)2po[IQox]

(2)

E is the "average excitation energy," frequently approximated as AE, the difference in energy between ground state and the first maximum in the electronic spectra, (r'3)2po is the inverse of the mean volume of 2p orbitals on oxygen, and Qox is the charge density bond order matrix. Even though these terms may be interrelated, a number of empirical correlations between 1^0 NMR chemical shifts and other data taken to represent various terms in the KarplusPople expression have been reported [8]. The major structural factors which influence l^O NMR chemical shifts are electronic, steric and hydrogen bonding. The influence of each of these factors will be briefly illustrated here; however, the reader is referred to several reviews for more in-depth coverage of these topics [1, 6-8]. Generally, double-bonded oxygen, with significant n bond character, is more sensitive to substituent effects than single bonded oxygen. Large changes in l^O NMR chemical shifts, both shielding and deshielding, can be observed by changes in the electronic and steric environment for a given functional group. Large shielding effects are observed for hydrogen bond formation to multiple bonded oxygen; smaller effects are observed for hydrogen bond formation to single-bonded oxygen. Electronic effects A study of substituted acetophenones by Brownlee and co-workers [12] nicely illustrates the effect of electronic factors on carbonyl 1^0 NMR chemical shifts. Substitution with the electron donating amino group results in shielding and introduction of the electron withdrawing nitro group produces deshielding of the acetophenone carbonyl l^O NMR signal (cf. 1-3). The l^O NMR chemical shift range for the carbonyl signal of the /^^fra-substituted acetophenones that was observed is approximately 50 ppm. The acetophenones l^O NMR chemical

553

shifts were correlated with a+ values, DSP values and by the more fundamental property the calculated n electron density[12]. Recently, the l^o NMR chemical shift of related carbonyl systems (methyleneindanones) covering an even wider chemical shift range (ca. 100 ppm) have been shown to be correlated with AMI calculated electron densities [13]. 564 ppm

NHi 1

NO2 2

3

Dahn and co-workers have systematically examined the effect of substituents on a wide range of carbonyl functional groups and have interpreted the magnitude of the rho values for carbonyl l^O NMR chemical shift-sigma plus correlations as a measure of the "electron-demand" or polarity for different classes of carbonyl groups [14-18]. Table 1 contains rho plus values for the more common carbonyl compounds which have been studied to date. The rho plus values range from approximately 29 for the highly electron deficient -COCF3 carbonyl group to ca. 7 for the more electron rich, non-hindered amide carbonyl system. The rho plus value noted for the N,N-dimethylbenzamides, ca. 2, reflects the large torsion angle of the functional group and is a classical example of a significant decrease in conjugation between aryl ring and functional group as a result of loss of planarity [19]. Dahn, et al. [15] have also shown that the l^o NMR chemical shift of the parent carbonyl compound is linearly related to the magnitude of the p+ value and the chemical shifts are thought to be dependent upon the electron density/bond order of the carbonyl group (Figure 3 is a plot analogous to that of Dahn). Deviations from the line were suggested to arise from other contributions to the chemical shift, probably the AE term of the Karplus Pople expression[15]. The values in Table 1 and Figure 3 reflect the electron demand of various functional groups that organic chemists have intuitively employed for decades; however, this work by Dahn, et al elegantly places them on a quantitative scale [14-18].

554

Table 1 Sensitivity of 4-X-Ar-COY I^Q N M R chemical shifts to substituents on the aryl ring 5(X=H)^

p+b

Ref.

CF3

544

29

14

H

562

26

14

Br

513

24

14

CH3

549

22

14

CI

491

20

14

SEt

489

16

14

F

353

14

14

OCOAr

386

13

14

CHN2

440

10

15

OCH3

337

8

14

NH2

326

7

14

NHCH3

318

7

19

coo-

265

5

14

N(CH3)2

348

2

19

Y

^ The carbonyl I ^ Q NMR chemical shift(ppm) for the parent compound in each series. b Rho plus value from correlation using Hammett Brown sigma plus constants.

555 600

g 500 H a

SEt

a

O

13

u

400 OCOAr

z O N(CH3)2

I ^-^

NHCH3

200

T"

Rho-Plus Values

Fig. 3. Plot of Carbonyl 170 NMR chemical shift yalues(ppin) for ArCOX versus rho-plus values for chemical shift sigma-plus correlations(see ref 15). The l ^ o NMR chemical shifts of functional groups other than carbonyl ones, including aromatic nitro groups [20], pyridine N-oxides [21,22], phenols [23], and anisoles [24] have also been found to be predictably quite sensitive to electronic effects. Interestingly, the l^O NMR chemical shifts of aryl sulfones [25,26] and sulfoxides [25,27] are essentially insensitive to substituent effects.

556

Steric effects In addition to providing insights into electronic distribution, other important structural information can be readily deduced from l^O NMR data. Conformational structure of both aromatic [28,29] and aliphatic series [30] of oxygen containing compounds has been studied extensively by 1^0 NMR methodology. However, more recently, work from our laboratories has shown that l^O NMR spectroscopy is a valuable method for detection of the effects of steric interactions on molecular structure for organic compounds. It is useful to divide the steric interactions into two categories: one involves systems in which steric interactions result in rotation of functional groups around single bonds to relieve van der Waals interactions, and the other deals with rigid systems in which steric interactions are partially accommodated by bond angle and bond length distortions. van der Waals effects: conformationally flexible systems l^O NMR spectroscopy has been successfully used to gain information about the conformational structure of both aromatic and aliphatic compounds. The conformation of functional groups attached to aromatic rings can be significantly influenced by introduction of bulky groups near such a functional group. Repulsive van der Waals interactions can be reduced by torsional rotation around the single bond connecting the functional group and the aromatic ring. Such interactions are often discussed in terms of steric inhibition of resonance. Large differences in l^O NMR chemical shifts noted for ortho and para nitrobenzenes in the pioneering work of Christ and Diehl were attributed to steric inhibition of resonance [28]. However, only in recent years have quantitative relationships between l^O NMR chemical shifts and functional group torsion angles been developed [6,7,31-33]. As was previously pointed out [1], the effect of torsion angle rotation on ^^O NMR chemical shift can potentially lead to three different results: (i) deshielding, (ii) shielding, or (iii) no change. A large number of examples of the deshielding effect have been observed [6,7,31-33], three cases in which shielding was noted have been reported [34-36] and one example in which no change on I ^ Q NMR chemical shift was observed on torsion angle rotation has been published [35]. Only the deshielding cases (illustrated by 2, 4-6) will be discussed in this review.

557

Good quantitative correlations between l^o NMR chemical shifts (deshielding) and torsion angles have been reported for aryl nitro compounds [31], aryl ketones [32,33], aryl acids [33], aryl esters [33], aryl acid chlorides [37], and aryl amides [33], and a more qualitative correlation has been noted for aryl sulfinyl compounds [38]. The 1^0 NMR carbonyl chemical shift for several different aryl systems plotted versus the torsion angle defined by the intersection of the plane of the carbonyl group and the plane of the aryl ring deduced from molecular mechanics (MM2) energy minimized structures are shown in Figure 4. These results are representative of the general relationships that have been developed between torsion angles and ^^O NMR chemical shifts. MM2 calculations predict that for many different carbonyl systems rotation around the single bond connecting a functional group to an aryl ring results in repulsive van der Waals energy of near zero [7]. Consequently, the change in electron density on the carbonyl oxygen as a result of torsion angle rotation is considered to be the major contributor to the observed ^7o NMR chemical shift variation. It is important to note that the origin of the deshielding shifts detected in conformationally mobile systems is fundamentally different from the origin of deshielding shifts observed for the rigid planar systems reviewed below.

558

700

Torsion Angle (Degrees)

Fig. 4. Plot of the carbonyl ^'O NMR chemical shift for aryl acids O (33), aryl amides^ (33), aryl esters # ( 3 3 ) , aryl acid chlorides ffl(37), and aryl ketones ^ (32) verus torsion angle between the aryl ring and the carbonyl functional group.

van der Waals effects: rigid planar systems Large downfield shifts of the N-oxide ^'O NMR signal for hindered pyridine-N-oxides (7,8) and quinoline-N-oxides (9,11) systems in which torsion angle twist is not thought to be likely led to a broader study of rigid planer systems[39,40].

559

HjCrCHa CH3

341 ppm

o .CH,

":)

10

370 ppm

11

12

In an investigation of the 1^0 NMR properties for a series of 2- and 4-alkyl substituted pyridine-N-oxides[cf. 7,8], it was found that the 2-alkyl compounds were deshielded relative to the 4-alkyl isomers [39], Large deshielding effects were noted for highly hindered compounds such as 8. Quinoline N-oxides exhibited similar effects; the deshielding effect of the methyl group in 2methylquinoline-N-oxide (10) is similar to that noted for 2-methylpyridine-Noxide [40]. However, 8-methylquinoline-N-oxide (11) displayed a much larger effect, presumably as a consequence of the peri-like relationship between the methyl and N-oxide groups [40]. In contrast, a large upfield shift, a consequence of intramolecular hydrogen bonding, was noted for 8-hydroxyquinoline-N-oxide (12) (vide infra).

560

Several series of hindered carbonyl compounds were systematically investigated in an effort to understand the origin of the downfield shifts of the 17o NMR signals in rigid planar systems. Selected compounds and the corresponding 1^0 NMR data from these studies are shown below. A 27 ppm deshielding effect is noted (cf. 13 and 14) for the hindered isomer of the tbutylphthalide system [41,42]. MM2 calculations for these isomers predicted carbonyl aryl ring torsion angles of essentially zero. However, in the case of 14, the calculations predict moderate carbonyl bond angle deformations. The I ^ Q NMR data for a similar pair of compounds, the indanones 15 and 16, have been obtained [43]. The downfield position of the l^o NMR signal for 16 is viewed as a consequence of repulsive van der Waals interactions between the t-butyl group and the carbonyl oxygen. MM2 calculations for 16 reveal no torsion angle twist; however, bond angle deformations are predicted.

HaC^

CH3 346 ppm .CH3

^ O

170 ppm

O

173 ppm

CHj

13

H3C..

CH3 541 ppm CH3

16

Reasonable models (isomers) for unhindered phthalimides which would allow direct comparison of chemical shifts for the hindered phthalimides do not exist. Nevertheless, the progressive deshielding of the carbonyl l^o NMR shift of 1720 as the N-alkyl groups become larger is consistent with increasing steric interaction and presumably increasing repulsive van der Waals interactions. MM2 calculations predict no torsion angle changes for these compounds but do suggest in-plane bonding deformations [44].

561 379 ppm

374 ppm

O N-H • <

O

Cc

N-CH3

^ \ ^ -<

394 ppm 0

0

o

O:^"^" \ ^ = ^

0 18

17

383 ppm

CH3

f

1;N-4-CH3

^ ^ - ^ ^

0

CH3 0

20

19

The flavone pair (21 and 22) and the t-butylanthraquinone examples (23 and 24) show pronounced deshielding shifts for the hindered isomers (e.g., 50 ppm; compare 23 and 24). MM2 calculations for these sets of compounds also predict bond angle deformations without torsion angle twist. 451 ppm

21

523 ppm

23

22

573 ppm

24

The origin of the deshielding effects in these systems appears to be repulsive van der Waals interactions. Chesnut has shown for several heteroatom nuclei that

562

repulsive van der Waals interactions are responsible for deshielding shifts. For example, deshielding shifts of ^'^O, ^^N and 31p NMR data were correlated with repulsive van der Waals energies [45-47]. An approach for accessing this possibility for the effects observed for the carbonyl systems described above was developed which compared MM2 calculated van der Waals energies for hindered and unhindered isomers with l ^ o NMR chemical shift differences [7]. A plot of estimated "local" repulsive van der Waals energies versus 17o NMR chemical shift differences yielded a line with a slope of approximately 13 ppm/kcal [7]. This correlation was considered a reflection of the role of the r'3 term of the Karplus- Pople equation in determining the deshielding effects on ^^O NMR chemical shifts in these systems. In summary, it is reiterated that large deshielding effects on 1^0 NMR chemical shifts can be observed in hindered carbonyl systems which are conformationally flexible as well as in rigid planar systems. The origin of the downfield effects in the two systems are different consequences of repulsive van der Waals interactions. Futhermore, semi-flexible systems have been studied and, in general, the observed deshielding effects could be quantitatively factored into torsion angle and rigid planar components [48]. Hydrogen bonding effects The significant influence of hydrogen bonding on carbonyl ^^O NMR chemical shifts was noted by early workers in the field [28,49,50]. A 52 ppm shielding shift was observed for the acetone 1^0 NMR signal on intermolecular hydrogen bonding with water [50]. Large shielding shifts of carbonyl l^O NMR signals resulting from intramolecular hydrogen bonding have been reported for salicylaldehydes, ortho-hydroxy -acetophenones [51], and -quinones and related compounds [52,53]. Some of the earlier l^O NMR spectroscopy literature for intramolecular hydrogen bonding systems is confusing, since often the observed upfield shifts were described as if arising exclusively from hydrogen bonding. To determine the magnitude of the contribution from hydrogen bonding to an acceptor functional group l^Q NMR chemical shift, it is essential to factor out any contributions from electronic and/or torsion angle effects. A method of analysis was developed to deal with these factors during a study of 6^r//z^-aminoand amido-acetophenones [54]. The validity of the factoring method for determination of torsion angle contributions in conformationally flexible systems was confirmed by a study of rigid planar amino- and amido-fluorenones and anthraquinones [55]. This factoring approach has led to the deduction of A 5 H B values (hydrogen bonding component to the functional group l^O NMR chemical shift) for a number of intramolecular hydrogen bonding systems [23,54-58].

563

Not surprisingly, the A5HB values generally depend on the basicity of the hydrogen bond acceptor and the acidity of the hydrogen bond donor. Figure 5 illustrates the average change in shielding shift (A8HB values) observed for aryl ketone acceptors as a function of acidity of the donor. Note that as the acidity varies from the moderately strong phenolic donor to the considerably weaker amino donor the shielding influence is reduced. Figure 6 contains average hydrogen bonding induced shielding data (A5HB values) for systems in which the hydrogen bond donor is held constant (phenolic) and the basic character of the acceptor is varied. In this case the magnitude of the shielding observed decreases with decreasing basicity of the acceptor functional group. Other factors such as the geometry of the intramolecular hydrogen bond have also been shown to be extremely important [59]. For 7-hydroxyindanones [59] where the hydrogen bond distance is significantly longer as a consequence of the geometric constraints of the 5-ring-6-ring fusion, the A8HB value is dramatically reduced to ca. 10 ppm in comparison to the average values of 52 ppm noted for shorter hydrogen bonding distances arising from 6-ring-6-ring fusion [56,60].

60

e a,

M

ArOH 50 H

40

.s

oe m

30 H

Ji

20

^

10

ArNH2

^ Donor Group

Fig. 5. Hydrogen-bonding shielding effects on ArCOR 170 NMR signals on intramolecular hydrogen bonding to various donors.

564 60-

s

a a

"wD

ArCOR 50-

G

40H

ArCOH

n

30

iJ

20H

ArN02 ArCOOR lOH

Acceptor Functional Group

Fig. 6. Hydrogen-bonding shielding effects on different functional group 170 NMR signals on intramolecular hydrogen bonding to a phenolic donor. Interestingly, the effect on the 170 NMR chemical shift of multiple hydrogen bonding to a single carbonyl group acceptor appears to be roughly additive. For example, in systems capable of forming two intramolecular hydrogen bonds to a single carbonyl acceptor, the A6HB values observed are approximately twice those noted for systems which can form only a single hydrogen bond (60). These effects are apparent from 170 NMR data obtained from toluene solutions for 1-hydroxyanthraquinone (25), 1,8-dihydroxyanthraquinone (26) and 3-methyl-l,8-dihydroxyanthroquinone (27) [60]. 88 ppm OH

456 ppm O

25

91 ppm OH

395 ppm ^ HO

92 ppm 389 ppm OH ^ HO

27

565

A comparison of l^O NMR results for 2',4,-dihydroxyacetophenone (28) and 2',6'-dihydroxyacetophenone (29) from two different solvents, acetonitrile and toluene, lead to a better understanding of hydrogen bonding in these phenolic systems [56], 463 ppm

29

In both cases, 28 and 29, formation of two intramolecular hydrogen bonds to the carbonyl group is not possible for geometric reasons. Larger A 5 H B values were anticipated as a result of an expected intermolecular hydrogen bonding component to the chemical shift; however, in acetonitrile A 5 H B values of 50 and 51 ppm, consistent with values for monohydroxyacetophenones, were noted. In addition, a small (ca. 2 ppm) upfield shift of the signal for the internal reference, 2-butanone, was observed. These results suggest that the phenolic OH group not participating in intramolecular hydrogen bonding is primarily involved in intermolecular hydrogen bonding with the solvent (acetonitrile). The A 6 H B values for 28 and 29 obtained from toluene solution are 55 and 57, in good agreement with the average value of 54 found for toluene solutions of other systems which can form only one intramolecular hydrogen bond. In the toluene solution, l^O NMR study of 28 and 29 revealed that the 2-butanone signals (internal reference) were shielded by 11-13 ppm. These results suggest that in toluene solution phenolic OH groups which are not participating in intramolecular hydrogen bonding are free to form intermolecular hydrogen bonds. Apparently, in the case of toluene as the solvent, phenolic hydrogen bonds with solvent are weak. Consequently, by performing l^o NMR spectroscopy measurements in toluene solution and by observing the chemical shift of an internal ketone reference, "free" phenolic groups can be detected. It should be noted that proton to oxygen coupling is often directly observed for intramolecular hydrogen bonded phenolic systems in acetonitrile solutions [56,57]. Direct proton-to-oxygen coupling has also been observed for simple alkyl alcohols [61]. The magnitude of the coupling constant ^JOH appears to be fairly constant and these values are in agreement with those obtained by lineshape analysis [8,62]and from l^o INEPT experiments[63]. The values are

566

typically 80 ± 25 Hz (see Table 2). The U Q H values have been shown to be dependent upon concentration, temperature and solvent [61]. Currently, there exists only an incomplete understanding of the factors which influence UoH values; consequently, further investigations are needed. Table 2 Representative values of directly observed proton to oxygen coupling (UoH) in acetonitrile solution. Compound

IJOH(HZ)

Ref.

2'-hydroxyacetophenone

86

56

2'-hydroxy-4'-methoxyacetophenone

75

56

2'-hydroxypropiophenone

91

56

2-acetyl-1 -naphthol

80

56

methyl salicylate

73

57

ethyl salicylate

104

57

salicylaldehyde

92

23

5-acetoxy-2-hydroxybenzaldehyde

58

23

2-nitrophenol

96

58

2-methoxyphenol

72

23

3-pentanol

79

61

cyclopentanol

79

61

cyclohexanol

73

61

^In each case the solvent was anhydrous acetonitrile and data were collected at 75°C l^O NMR SPECTROSCOPY OF NATURAL PRODUCTS A recent survey of application of NMR of chalcogen nuclei to natural products chemistry included some interesting examples of ^^o NMR results for

567

selected natural products [64]. A number of reports have now appeared describing the l^Q NMR properties of several natural product systems. Even more specialized usage of l^o NMR spectroscopy includes a study on coal distillates [65] and l^o NMR imaging of plant stems [66]. In this section of the chapter an overview of the l^o NMR chemical shift characteristics of some important classes of natural products will be presented. Occasionally, for illustrative purposes, 1^0 NMR spectra of selected natural products obtained in the author's laboratory will be included. Data for certain key compounds needed to facilitate discussion, also obtained in the authors laboratory, are included. The new data presented here were measured under our standard operating conditions [60] using acetonitrile as solvent and, except for artemisinin(vWe infra), at TS^C. Ketones The ketone functional group appears in a wide variety of natural products ranging from the simple terpenes to complex macrocycles and polyfunctional molecules. This section will focus on simple, naturally-occurring ketones, largely from the terpene family, which illustrate relationships between structure and ^^O NMR chemical shifts. Other more complex ketone containing structures will be examined later according to their natural product family classification, e.g. chromones, flavones, etc. Double bonds in conjugation with the ketone function result in shielding of the l^O NMR carbonyl signal. The magnitude of the shielding shift is modest in acyclic systems(typically 5 ppm or less) whereas, in cyclic ones the shift is larger[67-70].Note the 12 ppm difference in l ^ o NMR chemical shift for 30 and 31, even larger differences are observed for 5-membered ring pairs[67]. Analysis of the apparently competing factors which cause these differences between cyclic and acyclic systems is incomplete. The effect of substituents on the l^O NMR chemical shifts of simple cyclohexenones (31-33) has been investigated [67]. Both a and p alkyl groups cause shielding of the cyclohexenone carbonyl signal (cf. 32 and 33 with 31).

30

31

32

33

568

The 1^0 NMR data for the terpenes isophorone (34), nopinone (35) and verbenone (36) are shown below. The presence of a g^/n-dimethyl group in isophorone causes only a small shielding effect, compare values of 32 and 34, whereas, in the saturated system the chemical shift of 3,3-dimethylcyclohexanone is downfield of 30 by about 12 ppm [71]. The l^o NMR chemical shift value for nopinone (35) is upfield of that of 30 by 14 ppm. In the unsaturated system verbenone (36) the presence of a double bond causes a 22 ppm shielding shift relative to 35. The direction of the shift is consistent with results previously discussed, however the magnitude of the shift is greater than noted for simple cyclohexanone/cylohexenone comparisons. Perhaps conformational changes also contribute to the larger upfield shift. Additional information is required for complete analysis of the shielding change noted for 36.

34

35

36

The naturally occurring ionone pair 37 and 38 demonstrate again that the effect of conjugation on acyclic carbonyl l^o NMR chemical shift is smaller than that for cyclic a,p-unsaturated ketones.

37

38

569

A recent report describes the ^^o NMR chemical shifts of various oxidation products 39-42 of (+)-3-carene [72]. The chemical shifts of the two a,Punsaturated ketones 39 and 40 are substantially shielded relative to the simple cyclohexenones 32 and 33. This shielding effect doubtlessly involves conjugation with the cyclopropane ring[68,73], however detailed analysis requires additional data to appropriately account for the influence of the ^em-dimethyl group. The quinone-like molecule 41 gives only one carbonyl signal, apparently the l^O NMR chemical shift of the two different oxygen atoms overlap. The carbonyl l ^ o NMR chemical shift for 41 is downfield of its analogs 39 and 40 consistent with the quinone-like structure(vWe infra)A^O NMR data for other bicyclic ketones related to naturally occurring ones have been reported[71]. 518.8 ppm

571.7 ppm

546.8 ppm

39

Lactones A number of important lactones occur in nature and consequently an understanding of l^o NMR chemical shift values for this family of cyclic esters may prove useful in characterization of such compounds. This section focuses on relatively simple lactones, other lactones such as courmarins are treated later as a separate class of natural products. l^O NMR results for a number of simple lactones have been published [74]. The signals for both the carbonyl group oxygen and the single-bonded oxygen are sensitive to substituents; however, no systematic variation with ring size is apparent(cf. 43-46). Substitution at the 5position of the butyrolactone ring results in deshielding of the single-bonded oxygen l^o NMR signal, however the carbonyl resonance is not effected(cf. 44 and 47). Introduction of a double bond in conjugation with the carbonyl group leads to shielding of both lactone l^O NMR resonances(cf. 44 and 48), whereas introduction of a double bond in conjugation with the single-bonded oxygen results in deshielding of both signals(cf. 47 and 49).

570 348.5 ppm

340.5 ppm

o-o CI

O 241.0 ppm

o

178.5 ppm 44

43

340.3 ppm

377.0 ppm

o

167.0 ppm 45

46

326.7 ppm

351.3 ppm

^>

C^o

O

O

367.0 ppm

206.6 ppm

172.7 ppm

47

o

239.0 ppm

48

49

Data for the natural products sclareolide (50) (see Figure 7) and the ascaricide santonin (51) (see Figure 1) are consistent with that for simpler lactones. The carbonyl oxygen signals for 50 and 51 and the single-bonded P 353 ppm ^O

243 ppm

CH, a\CH3

509.3 ppm O H,C

CH,

CH3

O-

188.8 ppm 50

O 345.4 ppm

51

oxygen signal for 50 are substantially deshielded compared to 44. The changes in chemical shift are generally reasonable based upon substituent effects observed of simple lactones[74], however there are too many structural variations for which adequate models are unavailable to allow a detailed analysis of the l ^ o NMR chemical shifts of 50 and 51 .

571

I It I M II II I I I I I j I I I I I I I II j r 400 300 500

I I I I I I I I I T> ' '» I I I 1 t I I I I I ) I I I 100 0 PPM 200

Fig. 7. 1 7 o NMR spectrum of sclareolide (SO), 0.5 M in acetonitrile at 75^0, internal reference 2-butanone 558 ppm, external reference water 0 ppm. Butenolides constitute a major class of lactone natural products which have been shown to play important roles in cancer chemotherapy [75]. It was previously noted that introduction of a double bond in conjugation with the lactone carbonyl results in shielding of the carbonyl l^o NMR chemical shift; for example compare the carbonyl chemical shift of 44 and 48 [74]. However, l^Q NMR data for lactones with conjugated exocyclic double bonds do not appear to have been reported. The l^o NMR values for the simple butenolide 52 and parthenolide (53) (see Fig. 8) were obtained in acetonitrile solution. Note that the l ^ o NMR chemical shifts of the carbonyl groups of 52 and 53 are similar. Interestingly, the conjugation of an exocyclic double bond causes larger shielding than noted for an endocyclic double bond(cf. 48 and 52); perhaps , in part due to

572

a p-effect. The single-bonded lactone oxygen of 53 is deshielded compared to that of 52, consistent with the deshielding effect of substituents on the 5-position of the lactone ring. The signal at 40 ppm for 53 is attributed to the epoxide oxygen, in agreement with value previously reported for this functionality [76,77]. CH3 \

^<^

315,8 ppm

-| ^ 40.0 ppm

o-^o 71.7 ppm

H3C

0

» 0 —

189.0 ppm

•1

I I I I I [ I » [ I I I I I 1 I 1 I I i

500

400

n

Q 318.0 ppm

I I I t I I I I »» I { I f I I j I I I I I ) F f I I I I I I I I I I I I 1 I 1

300

200

100

0 PPM

Fig. 8. 1 7 o NMR spectrum of parthenolide (53), 0.5 M in acetonitrile at 750C, internal reference 2-butanone 558 ppm, external reference water 0 ppm.

573

Lactams Although lactams, perhaps, are not as ubiquitous in nature as lactones, a brief mention of the l^o NMR chemical shift of simple lactams seems appropriate. 17o NMR results for four lactams(54-57) are given below with their structures [78]. The amide carbonyl signal varies with ring size in a manner similar to that noted for lactones. The chemical shift of the lactam carbonyl signal is displaced, as expected due, to greater electron density on the oxygen, by approximately 40 ppm upfield of that of the lactones. As is the case with any amide containing the NH function, care should be taken in interpretation of small chemical shift differences for lactams since these acyclic amides can associate by hydrogen bonding. 314.3 ppm

301.1 ppm

O. B 54

55

o

320.6 ppm

336.6 ppm

0O

56

57

Quinones The wide occurrence of quinones [79] makes this a particularly important class of compounds and this fact is reflected by several ^^O NMR investigations of quinones and related compounds [52,53,55,60,67,80,81]. Data for representative simple quinones and one quinone methide are shown below. 624 ppm O

572 ppm O

574

The chemical shift of the quinone ^^O NMR signal is shielded by fusion of aromatic rings, compare 59 and 60 with 58, consistent with greater delocalization of the carbonyl group's 7C-electron density. Interestingly, the 1^0 NMR chemical shift for the quinonemethide 61 is shielded by 35 ppm compared to anthraquinone which is consistent with differences in their calculated electron densities [81,82]. The observed chemical shifts for quinones studied to date suggest that the quinone-carbonyl group responds to substituent effects as do other carbonyl groups. The electronic effect of substituents on the carbonyl chemical shift appears normal (electron donators cause shielding; electron withdrawing groups produce deshielding). The ^^o NMR signals of hindered quinones are deshielded relative to non-hindered isomers. Intramolecular hydrogen bonding from phenolic OH groups cause large (ca. 50 ppm per H-bond) shielding shifts of quinone carbonyl l^o NMR signals(see 25-27). A study of the biosynthesis of citrinin(62) in Aspergillus terreus used several isotopically labeled acetates including [l-13c, 17o]-acetate. The ^^O NMR spectrum of citrinin obtained from [l-l^c, l^oj-acetate gave three signals (148, 179 and 279 ppm) (83). These signals were assigned to 0-2 and to the oxygen atoms attached to C-8 and C-6, respectively. The structure of citrinin was described in terms of a tautomeric equilibrium involving the atoms at C-6 and C-8[83].

62

Coumarins and chromones A special family of lactones, the coumarins, which widely occur in plants[84], have been investigated by ^^O NMR spectroscopy in some detail. Values for l^o NMR chemical shifts for representative members of the class are shown as 64 and 65; data for dihydrocoumarin 63 is given to illustrate the influence of conjugation on the l^O NMR signal for the lactone carbonyl group.

575 375.8 ppm 204.0 ppm

63

351 ppm 219 ppm

347 ppm 225 ppm

64

65

Tables 3 and 4 contain results from a study of a series of coumarins and furocoumarins in, 1,2-dibromoethane, which found that the lactone carbonyl signal consistently appeared near 350 ppm; more variability was seen for the single bonded oxygen, especially for the furocoumarins, typically appearing at 220 ppm [85]. Synthetic 3-aryl coumarins also exhibit 8(C=0) and 6(-0-) values near those mentioned above [86]. Recently, a study of a series of 7-substituted-4methylcoumarins, with a wide range in electronic character of the substituents, demonstrated that the carbonyl signal is quite sensitive to substituent effects and that the l^O NMR chemical shift is reasonably well correlated with the carbonyl oxygen AMI estimated electron density [87]. Table 3 17o NMR chemical shifts (ppm) of coumarins in 1,2-dibromoethane at 70oC[85]

:ixx. No. Rl 64 H 65a H 66 CH3 67 H ^ Signal for OCH3 group 62 ppm.

R2 H OCH3 H CH3

5(C=0) 351 347 351 352

8(-0-) 219 225 220 221

576

Table 4 17o NMR chemical shifts (ppm) of furocoumarins (psoralens) in 1,2-dibromoethane at 7 0 T [ref. 85] R2

No.

Rl

R2

R3

5(C=0)

68 69 70 71 72 73 74

H 0CH3 H OCH3 0CH3 H e

H H 0CH3 0CH3 Br H H

H H H H H d H

352 350 351 351 352 351 351

5(-0-)a 5b(-0- ) 5(other) 219 200C 220 200C 200C 215 200C

200 200C 202 200C 200C 207 200C

14 45 25C

44

^ Lactone single bonded oxygen, b Furan single bonded oxygen. ^ Broad overlapping signals. ^

^-"M

l^o NMR data for a few chromones were also reported earlier[88,89]. As can be seen from the data (from chloroform solutions ) given with the structures 66-68 the ketone chemical shift is sensitive to substitiuents like most aryl ketones. However, the vinyl ether signal is relatively insensitive to

66

67

68

577

substituents, except in the more complicated chromone visnagin(69). The ^^O spectra for visnagin and its analog khellin(70) are shown in Fig, 9. The chemical shift value differences between 69 and 70 are consistent with the change in structure (the addition of the 9-methoxy group in khellin). The carbonyl l^O

' \ \ \ i-fTnm-i 1 i M p ^ M p M t \-t M I p » M I I 1 < A I t M I \ M I v\\ x \ \ \ \ \ \ \ \y \ 500

400

300

200

100 —

0 PPM

Fig, 9. 17 o NMR spectrum of visnagin(69)(bottom) and khellin(70)(top) , 0,5 M in acetonitrile at 75^ C, internal reference 2butanone 558 ppm, external reference water 0 ppm.

578

NMR chemical shifts, as expected, are essentially the same [only small effects due to ''meta "-substitution are anticipated]. The 5-methoxy signal is shielded by ca. 10 ppm in 70 as a result of the "para " -methoxy, consistent with results from anisoles[24]. Both ring oxygens are shielded due to the introduction of the 9methoxy group in 70 as expected due to their "orr/zo"-relationship[90]. The 9methoxy signal is significantly shielded in accord with results for the analogous methoxy group of l,2,3-trimethoxybenzene[90]( see also the aryl ether section below).

40.4 ppm

30.9 ppm

459.3 ppm

vs OCH3 0

OCH3 0

11

^ ^

0

460.6 ppm

^ CH3

O'

YST

ll11 0

CH3

0CH3 200.4 ppm

166.2 ppm 69

192.7 ppm

5.5 ppm

154.2 ppm

70

A recent report describes the use of l^o NMR spectroscopy to differentiate between coumarins and their isomers, the chromones [91]. This study reports data for five isomeric pairs of coumarins and chromones in dimethylsulfoxide at lOOX. Results from two pair are shown as 71-74 to illustrate the difference between the data for the isomers. After investigating the use of several solvents these authors selected dimethylsulfoxide as the solvent of choice for their studies, despite the large solvent signal, due to the greater solubility of the coumarins and the chromones . These investigators suggest that the difference in chemical shift of the two different type carbonyl groups (enone and lactone) and the two different single-bonded oxygens (vinyl ether and lactone) can be reliably used to distinguish between chromones and coumarins. Previous investigators had noted the characteristic region of 1^0 NMR absorption for the coumarin and chromone systems; however, Nagasawa and co-workers[91] were first to recognize the importance of this approach for natural products structure elucidations. This method appears to be a valuable way to distinguish between courmarins and chromones, however as with all structural assignments based on chemical shift

579 differences, each case should be carefully examined and the effect of all substituents must be taken into account. CH,

CH,

CH,

O 440.2 ppm 165.4 ppm

339.7 ppm 213.7 ppm H,C

H,C

72

71

352.8 ppm 216.8 ppm

73

439.9 ppm 165.4 ppm

74

Substituents influence carbonyl I ^ Q N M R chemical shifts of the chromanones listed in Table 5, as they do other aryl ketones[89]. Introduction of a methoxy group "para" to the carbonyl in 76 causes shielding analogous to that for acetophenones[12].On the other hand, location of a methyl group in the 5-position causes a 25 ppm deshielding shift(cf.79).The deshielding shift was attributed to the peri interaction between the methyl and carbonyl groups[64,89]. It was suggested that the interaction results in loss of coplanarity of the aryl ring and the carbonyl group which would account for the deshielding. However, molecular mechanics calculations predict a carbonyl-aryl ring torsion angle of only 3 degrees, which is insufficient to account for a 25 ppm downfield shift[48] . Repulsive van der Waals interactions, as described earlier in this review, are likely to contribute to the observed downfield shift[48]. The large shielding shift of the carbonyl signal of 81, a consequence of intramolecular hydrogen bonding, is consistent with results from other hydrogen bonded systems [60]. Data are not available that would allow an estimation of the electronic effect of the OH group in the chromanone system, hence predicting the portion of the shielding effect solely attributable to hydrogen bonding is not currently possible. The cyclic oxygen signal for the chromanones is essentially unaffected by substituents.

580

except for 80 for which the 8-methyl group causes significant shielding, arising from a y-gauche interaction[89]. Table 5 ^^O NMR chemical shifts of substituted chromanones in chloroform at 55''C [89]

Rl

O

R4

Compound 75 H No. Rl 76 H 77 H 78 H 79 CH3 80 H 81 OH

H R2 H H H H H H

H H R3 R4 0CH3 H OEt H H OiPr 0CH3 H OCH3 CH3 OCH3 H

518 117 5(C=0) 8(.0-) 500 118 497 119 495 119 525 120 499 108 443 120

8(other) 71 100 119 67 70 70(OCH3)93(OH)

Flavones and flavanones Flavones and their partially saturated analogs, flavanones, appear widely in plant sources and, consequently, continue to be extensively investigated[92]. A l^O NMR study of ten methoxy substitued flavones in chloroform solution has appeared[93]. However, the l^o NMR data shown below for some selected flavones and flavanones were obtained from acetonitrile solutions for these compounds. The difference in 1^0 NMR chemical shift for the flavone and flavanone carbonyl groups is large cf. 68 and 82. The shielding of 68 relative to 82 (ca. 80 ppm) is attributable to an increase in electron density on the carbonyl oxygen of 68 as a result of both oxy-enone and 2-phenyl conjugation. Introduction of a methoxy group in the 7-position of both systems(83 and 84) produces shielding of 13-18 ppm as expected of this electron donating group, based upon the early studies on acetophenones[12]. However, introduction of a methoxy group in position 5, which is electronically, but not sterically, equivalent to the 7-position produces deshielding shifts of 24 and 34 ppm(see 85 and 86).

581

The downfield shifts are consistent with repulsive van der Waals interactions between the 5-methoxy group and the carbonyi oxygen [48]. The I^Q N M R

O

Ph

158 ppm 68

HaCO

82

H3CO'

66.1 ppm

158.6 ppm 83

"^

65.1 ppm

O

Ph

95.5 ppm 84

chemical shift of the carbonyi group of the 5-hydroxyflavone 88 is 380 ppm , which represents substantial shielding and doubtlessly arises from a significant contribution from intramolecular hydrogen bonding. The isomeric 7hydroxyflavone 87 is poorly soluble in acetonitrile. In order to obtain data for 87, enrichment by exchange with H2^^0 was necessary. The carbonyi signal for 87 appears at 432 ppm. Consequently, the hydrogen bond induced shift for the carbonyi signal of 87 is deduced to be 52 ppm. This hydrogen bonding shift is in agreement with those of other aryl carbonyi groups[60].

582 61.2 ppm

472.0 ppm

OCH3 O

62.5 ppm

543.6 ppm

OCH3 O

86

86.4 ppm OH

O

87

380.4 ppm O

Ph

88

Carbohydrates One of the major families of natural products, carbohydrates, which contain multiple oxygen functions, present difficulties in acquiring interpretable natural abundance l^O NMR spectra. Application of 1^0 NMR to the study of carbohydrates is quite challenging not only due to the numerous alcohol and ether-like oxygens, which results in signal assignment difficulties, but also because of the general necessity of obtaining the data in water solution. Despite these difficulties significant progress is being made in the evaluation of the ^^O NMR spectroscopic properties of sugars by Lauterwein and co-workers[94,95]. Good quality spectra are obtained at high temperature (90'*C)[96], using 1^0depleted water[96] and water suppression techniques[97]. Assignments have been made by use of shift reagents[95,98] and by a computational strategy which solves multiple linear equations using observed chemical shift differences between

583

signals for sugars and a model compound expressed as a sum of y and 6 shielding terms(94). -10.6 ppm 1

-9.9 ppm 1

63.7 ppm

HOfi;jC

I.O

HO^

L6 ppm U Q ^

^OH

9.6 ppm

9.6 ppm *

36.5 ppm 47.9 ppm

6.3 ppm

HO|LC ^ 9 Q 52.6 ppm 6.3 ppm HO V l ' ' ^ - L N HO j L g J ^ ^ ^ ^^ O H 6.3 ppm 1 1 44.0 ppm

89

90

-7.3 ppm H O

OH 34.4 ppm 50.5 ppm

91

The assignments for the l^o NMR signals of the common sugars Dglucose(89), D-mannose(90) and D-galactose(91) have been made based on chemical shift arguments, selective enrichment (or depletion), and shift reagent studies [94,98]. The l^O NMR chemical shift assignments for 89-91 are given with their structures[94]. Note primary OH signals (C-6) appear between -7 and -11 ppm, generally the secondary OH signals (C-2, C-3, C-4) overlap between 6 and 10 ppm (except for C-4 for galactose); however, the secondary OH's at the anomeric position (C-1) appear between 36 and 50 ppm, depending on the a or p configuration, [assignments to a and p resonances were not made]. The I ^ Q NMR chemical shift of the 6-membered ring oxygen for these sugars appears between 50 and 64 ppm.

584 -10.2 ppm 1 9.6ppmHO

-11.4 ppm 1

55.4 ppm

HOB^iC

HOB

I.O

^

8.6 ppm HO

0CH3

HO-

HO-

OHp 9.6 ppm

9.6 ppm

50.4 ppm

uo

^

OH|

8.6 ppm

8.6 ppm

9.6 ppm *

OCH3 8.6 ppm

92

93

-10.7 ppm 1

65.4 ppm1

HOjLiC 10.8 ppm HO ' y H3CO. -10.7 ppm

.0 •^

OHT'

10.8 ppm 1

OH 37.7 ppm 48.9 ppm

94

The assignment of the l^O NMR chemical shifts of three O-methyl derivatives of D-glucose (92-94) have been made using similar approaches[94]. For both methyl a- and P- glucopyranosides 92 and 93 the methoxy signal overlaps the resonances of the secondary OHs (C-2,C-3,C-4); the coincidence of the four signals was inferred from signal intensity[94]. In the case of the 3-0methyl-D-glucopyranose (94 ) the 3-methoxy signal overlaps that of the 6hydroxy, again overlap was inferred from signal intensity [94]. Recently, all eight diastereoisomeric 1,6-anhydro-P-D-hexopyranoses and three related model compounds have been studied in l^O-depleted water using l^O NMR spectroscopy methodology [94,97]. Chemical shift assignments for all resonances in the eleven compounds were made. The results for glucosan ( 95), mannosan( 96), and galactosan ( 97) as well as the three model compounds 98100 are shown below.

585 44.7 ppm I

65.5 ppm

63.0 ppm

- t — O 38.9 ppm

p

\ - O 27.9 ppm

P^ I 19.2Sppm ppm I 24.1 ppm OH OH 16.7 ppm

95

I OH

HO—^

4.1.6 ppm -0.2 ppm

*V— O 38.2 ppm

1

M^ 5.9 ppm

2.8 ppm

•

OH

20.7 ppm

16.5 ppm

96

97

42.2 ppm

I

^o

I 62.2 ppm

HO

I OH

-03 ppm

98

99

\

15.6 ppm

100

Furans Compounds containing the furan ring appear widely in plants[99]. Some l ^ o NMR data for the furan ring system have already been presented in the discussion of the spectral characteristics of coumarins and related compounds. In this section the focus will be on the basic furan system and some fused ring analogs[8,76,100,101]. A shielding effect of ca. 42 ppm on the l^O NMR signal of the furan oxygen is noted on fusion of a benzene ring(cf. 101-102 and 104105) and it appears to be additive (cf. 102 and 103).

586 240 ppm

199 ppm

101

102

247.5 ppm

104

103

203 ppm

105

The l^O NMR chemical shift data for series of 2-substituted and 2,5disubstituted furans failed to correlate with Hammett-type constants[101]. However, the data were well correlated with the parameter Q which depends, in part, upon the polarizability and ionization potential of the substituent. Phenols and aryl ethers Numerous naturally occurring aromatic systems ranging from alkaloids to lignans contain phenolic and aryloxy groups[102-105]. Several examples containing these functional groups have already been treated in connection with review of the ^^O NMR properties of coumarins, chromanones, etc.(v/d^ supra). The purpose of this section is to briefly review the l^Q NMR characteristics of such functional groups and, in particular, examine the chemical shifts of structures containing more than one of the single bonded oxygen functional groups, for example, systems which contain a 1,2,3-trimethoxyaryl ring. Multialkyloxyaryl rings appear in numerous important naturally occurring compounds such as podophyllotoxin, glyfoline and derivatives. As yet, none of the more complex structures have been studied by 1^0 NMR spectroscopy but much of the

587

needed basic results on simple phenols and multi-methoxy aryl have been reported. Additionally, also included in this section are l^o NMR results obtained in our laboratory on selected trimethoxyaryl systems which were selected to facilitate discussion. 1^0 NMR chemical shift vales of various phenolic compounds have been reported in earlier l^o NMR investigations[28,49,51,53,60,106], however a systematic study of phenols did not appear until recently [23 ]. Substituent electronic effects on the phenolic OH l^o NMR chemical shift is illustrated by the results for 108-110 and is similar to substituent effects on the methoxy signal of anisoles [24]. Electron donating groups causes shielding and electron withdrawing groups results in deshielding of the phenolic OH I ^ Q NMR chemical shift. Data for other phenols are given with the structures 111-114.

NH, 108

109

OH

111

NO2 110

OCH3

112

113

114

It is apparent by comparing the data for ortho and para isomers 111-112 and 113-114 that formation of an intramolecular hydrogen bond to the single bonded oxygen of both the OH and OCH3 groups causes shielding of the 1^0

588

NMR chemical shift. The hydrogen bond shift of 8-14 ppm is considerably smaller than noted for hydrogen bond induced chemical shift for carbonyl group 17o NMR chemical shifts(yide supra). The influence of both electronic and steric factors on ^^O NMR chemical shifts of anisoles have been extensively studied[24,90]. The electronic effect of substituents is analogous to that of phenols which was previously discussed. The effect of orthO'SAkyl groups of varying bulk has been assessed and is illustrated in Table 6. Table 6 1^0 NMR chemical shift data of methoxy groups of substituted anisoles in chloroform solution^ . 0CH3

Ri.

No.

Rl

R2

115 116 117 118 119

Me Et i-Pr t-Bu t-Bu

Me Et i-Pr t-Bu Me

JL

.R2

5(ppm) 16.5 16.7 13.5 27.3 27.0

^ Obtained at room temperature. The deshielding shift for the highly hindered anisoles 118-119 relative to the less hindered ones 115-117 was suggested to arise from repulsive van der Waals interactions and/or bond angle deformation in these systems [90]. Results from studies on multi-methoxy aryl systems are indicated below with the structures 120-123. The greater shielding of the 1^0 NMR chemical shift for the ortho isomer 122 than that of the para isomer 121 was interpreted as arising from rotation of the methoxy groups from the plane of the aromatic ring as a result of repulsion between non-bonding electrons of the oxygen atoms[90]. Recent investigations employing GAUSSIAN-86 calculations and using the GIAO approach have resulted in the conclusion that the origin of the deshielding is repulsive van der Waals interactions between the two oxygen atoms[107,108]. An even greater shielding shift is noted for the signal of the 2-methoxy group of l,2,3,-trimethoxybenzene(123). Using the ortho-para equivalency assumption the

589

electronic effect of an ortho methoxy group is -8 ppm(cf. 120 and 121), consequently the chemical shift of 123 would be expected to be 32 ppm if only electronic effects were involved. The highly shielded value for for the 2-methoxy group of 123 is greater than two times the difference in chemical shifts of 121 and 122, presumably arising from even greater repulsive van der Waals interactions. -6.1 ppm 37.1ppin OCH3

H3CO^

OCH3 1

37.1 ppm OCH3

OCH,

120

121

122

123

Data collected for this review in our laboratory from acetonitrile solutions for several substituted 1,2,3-trimethoxybenzenes are given below with their structures. The difference in chemical shift for 123 given below and that taken from the literature(see above) reflects the differences in solvent and temperature used for the two measurements. The ^^O chemical shift of the two equivalent methoxy groups(at the 3 and 5 positions) are essentially unaffected as the substituent at the 1-position is changed(124-128). In contrast, the signal for the methoxy group at the 4-position is influenced by the substituent at the 1-position . In the case of methyl and hydroxymethyl substitution(124 and 125) small shielding shifts consistent with those noted for anisoles[24] are observed. Substitution of vinyl and carbonyl groups(126-128) in the 1-position cause substantial downfield shifts of the 4-methoxy signal. The downfield shift of approximately 22 ppm for the l^O NMR signal for the 4-methoxy group of the natural product asarone(126) is significantly greater than the downfield shifts noted for the 4-methoxy signal for the carbonyl substituted compounds 127 and 128. In fact, the result is in contrast to that found on substitution of a vinyl group in the para position of anisole where a small shielding shift was noted[109]. The large downfield shift noted for the methoxy signal for 126 is consistent with increased double bond character of the 4-methoxy group(increased conjugation). If this is the case, a change in conformation of the 4-methoxy group might be expected which also would be likely to cause

590

conformational changes for the other two methoxy groups, however only small changes in the l^o NMR chemical shifts of the signal for the 3,5-dimethoxy groups are noted. Additional study is required for a more complete understanding of the factors which determine the ^^O NMR chemical shifts in these important 1,2,3-trimethoxyaryl systems. 1.3 ppm OCH3

-0.9 ppm

35.4 ppm .OCH3

35.0 ppm

H3CO,

OCH3

JL

-0.2 ppm 35.0 ppm

.OCH3

36.1ppm

OCH3 26.1 ppm ' OCH3

5.0 ppm CH2OH

123

124

125

127

128

23.1 ppm 36.4 ppm OCH3

126

Miscellaneous natural products A recent report describes the use of l^o NMR spectroscopy for confirmation of stereochemistry of a dimethoxy derivative of abietic acid methyl ester [110]. The assignment of the axial stererochemistry to both of the methoxy groups in 129 is based upon ID and 2D ^H and ^^C NMR spectroscopic measurements.

591

l^O NMR chemical shifts and their assignment for 129 are shown with the structure below. The l^o NMR signals for the two methoxy groups in this diterpene apparently overlap at 1.6 ppm. The chemical shift of 1.6 ppm is viewed by the authors as consistent with values of methoxy groups in axial positions of simple cyclohexane ring systems and the assignment was made on that basis[l 10].

129

The antimalarial arteminisin(130) contains an endo cyclic peroxide group , a functional group which rarely appears in natural products. 267.2 ppm „

CH, V^

H,C 88.0 ppm I

n

H

II CH3 371.1 ppm

130

Fig. 10 contains the l^o NMR spectrum of arteminisin obtained at 40oC in acetonitrile. The resonance for the peroxy group is readily recognized as the signal at 267 ppm due to its large line width, approximately 1900Hz[lll,112]. The lactone carbonyl signal appears at 370 ppm and the signal at 200 ppm is assigned to the lactone single bond oxygen. The remaining signal at 88 ppm is

592

attributable to the ketal-like oxygen and is in the region noted earlier for ketal type oxygens[76,113].

I I r [ f 400

I I I I I I t I ' I ' l l

300

I r I I M

200

f j F T |-

100

0 PPM

Fig. 10. 1 7 o NMR spectrum of artemisinin (130), 0.5 M in acetonitrile at 40<>C, internal reference 2-butanone 558 ppm, external reference water 0 ppm. Pyrylium salts provide the brilliant color in many flowers and consequently these cationic oxygen containing heterocyclic compounds have been extensively studied[l 14]. While numerous physical studies on these systems have been performed the first l^O NMR investigation was reported only recently[115]. ^^O NMR chemical shifts of a series of isobenzopyrylium salts were found to be sensitive to substitution at the 1- and 3-postion which is illustrated by the data given with the structures 131-134. It is interesting that a larger shielding shift is noted for replacement of the 3-methyl by a phenyl group than the substitution of

593

the 1-methyl group by a phenyl one (cf. 131-133). The origin of this difference is not clear, but it may be a result of different conformations of the phenyl group at the two different postions. The 1-phenyl group is likely to be more twisted out of the plane of the isobenzopyrylium ring than a phenyl group at the 3-position. The effect of substitution of phenyl groups appears to be approximately additive since the l ^ o NMR chemical shift value of 134 is shielded by 30 ppm compared to 131. If the effect of phenyl group substitution were exactly additive the shielding shift of 134 would have been predicted to be 27 ppm . 300 ppm

289 ppm

284 ppm

270 ppm

133

134

CH,

131

132

Several other natural products systems have been studied, some quite extensively, by l^o NMR methodology, however space constraints prohibt detailed discussion of these systems therefore only important leading references will be given. Extensive l^o NMR studies on amino acids and small peptides have been performed by Lauterwein and coworkers [116-119], by Fiat and coworkers [120-123] and others [124]. Several studies have used l^O-enriched dioxygen and carbon monoxide to study by I ^ Q NMR techniques the interactions of these biochemically important small molecules with various proteins [125128]. A number of investigators have explored the properties and interactions of nucleic acid bases [129,130], nucleosides [131,132], nucleotides [133-138] and one report has appeared in which l^o NMR spectroscopy approaches were applied to the study of small molecule-DNA interactions [139]. A recent report describes the careful analysis of the effect of structure on l^o NMR chemical shifts of over forty hydroxyterpenoids [140]. A study of the 1^0 NMR spectroscopy of over thirty steroid ketones, acids, esters and alcohols enriched with l^O has recently appeared [141]. CONCLUSIONS The 17o NMR chemical shift patterns for the common and not so common functional groups encountered in naturally occurring compounds provides a

594 Strong foundation for assisting with structure elucidation and analysis of stereochemical relationships. Even though most of the relationships developed between structure and 1^0 NMR chemical shift are empirical they are valuable predictive tools. Increasing interest in and the study of the origins of l^O NMR chemical shifts should place these relationships on a more solid theoretical foundation in the years ahead. The rapid growth in numbers of publications which apply ^^O NMR spectroscopy to natural product problems clearly indicates the growing importance of this methodology to the field.

ACKNOWLEDGMENTS Acknowledgment is made to the Donors of the Petroleum Research Fund, administered by the American Chemical Society, for partial support of this research, to NSF (CHE-8506665), and most especially to my colleagues, whose names are cited in the appropriate publications in the reference section, without whom this chapter would not have been possible.

REFERENCES 1. D.W. Boykin (Ed), l ^ o NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, PL, 1991. 2. M. Witanowski, L. Stefaniak and G.A. Webb, in: G.A. Webb (Ed), Annual Reports on NMR Spectroscopy, Academic Press, London, 1993, Vol. 25, pp 1-468. 3. J.F. Hinton, in: G.A. Webb (Ed), Annual Reports on NMR Spectroscopy, Academic Press, London, 1987, Vol. 19, pp 1-33. 4. R.H. Thomson (Ed), The Chemistry of Natural Products, Blackie, Glasgow, 1985. 5. G.W. Kabalka and N.M. Goudgaon, in: D.W. Boykin (Ed), "17o NMR Spectroscopy in Organic Chemistry," CRC Press, Boca Raton, FL, 1991, Chapt. 2, pp 21-37. 6.

A.L. Baumstark and D.W. Boykin, in: A.L. Baumstark (Ed), Advances in Oxygenated Processes, JAI Press, Greenwich, Conn., 1992, Vol. 3, pp Mi-

7. 8.

D.W. Boykin and A.L. Baumstark, Tetrahedron, 45 (1989) 3613-3651. J.-P. Kintzinger, in: P. Diehl, E. Fluck, R. Kosfield (Eds), NMR-17, Oxygen-17 and Silicon-29, Springer-Verlag, Berlin, 1981, pp 1-64. 9. J.-P. Kintzinger, in: P. Laszlo (Ed), NMR of Newly Accessible Nuclei, Academic Press, New York, 1983, Vol. 2, pp 79-104. 10. W.G. Klemperer, in: J.B. Lambert and E.G. Riddell (Eds), Reidel, Dordrecht, Holland, 1983, pp 245-260. n . M. Karplus and J.A. Pople, J. Chem. Phys., 38 (1963) 2803-2807. 12. R.T.C. Brownlee, M. Sadek and D.J. Craik, Org. Magn. Reson., 21 (1983) 616-620.

595

13. A. Kumar and D.W. Boykin, J. Mol. Struct., 296 (1993) 95-101. 14. H. Dahn, P. Pechy and V.V. Toan, Angew. Chem., Int. Ed. Engl., 29 (1990) 647-648. 15. H. Dahn and P. Pechy, J. Chem. Soc. Perkin Trans. 2, (1993) 67-70. 16. H. Dahn and P. Pechy, J. Chem. Soc. Perkin Trans. 2, (1991) 1721-1723. 17. H. Dahn and P. Pechy, J. Chim. Phys., 89 (1992) 1683-1687. 18. H. Dahn, P. Pechy and V.V. Toan, Magn. Reson. Chem., 28 (1990) 883-887. 19. J. Spychala and D.W. Boykin, J. Chem. Research, in press. 20. D.J. Craik, G.C. Levy and R.T.C. Brownlee, J. Org. Chem., 48 (1983) 1601-1606. 21. D.W. Boykin, A.L. Baumstark and P. Balakrishnan, Magn. Reson. Chem., 23 (1985), 276-279. 22. M. Sawda, Y. Takai, S. Kimura and S. Misumi, Tetrahedron Letters, (1986), 3013-3016. 23. D.W. Boykin, S. Chandrasekaran and A.L. Baumstark, Magn. Reson. Chem., 31 (1993)489-494. 24. M. Katoh, T. Suguwara, Y. Kawada and H. Iwamura, Bull. Chem. Soc. Jpn., 52 (1977) 3475-3476. 25. S.A. Evans, Jr., in: D.W. Boykin (Ed), I ^ Q N M R Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, PL, 1991 Chapt. 10, pp 263-320. 26. J.W. Kelly and S.A. Evans, Jr., Magn. Reson. Chem., 25 (1987) 305-308. 27. G. Modena, U. Quintily and G. Scorrano, J. Am. Chem. Soc, 94 (1972) 202-208. 28. H.A. Christ and P. Diehl, Helv. Phys. Acta, 36 (1963) 170-182. 29. D.J. Sardella and J.B. Strothers, Can. J. Chem., 47 (1969) 3089-3092. 30. E.L. Eliel, K.-T. Liu and S. Chandrasekaran, Org. Magn. Reson., 21 (1983) 179-182. 31. P. Balakrishnan and D.W. Boykin, J. Org. Chem., 50 (1985) 3661-3663. 32. M.G. Oakley and D.W. Boykin, J. Chem. Soc. Chem. Commun., (1986) 439441. 33. A.L. Baumstark, P. Balakrishnan, M. Dotrong, C.J. McCloskey, M.G. Oakley and D.W. Boykin, J. Am. Chem. Soc, 109 (1987) 1059-1062. 34. H. Cerfontain, C. Kruk, R. Rexwinkel and F. Stunnberg, Can. J. Chem., 65 (1987) 2234-2237. 35. D.W. Boykin, G.H. Deadwyler and A.L. Baumstark,Magn. Reson. Chem., 26 (1988) 19-23. 36. D.W. Boykin, Spectroscopy Letters, 25 (1992) 1199-1205. 37. D.W. Boykin and A. Kumar, Spectroscopy Letters, 24 (1991) 723-731. 38. R. Benassi, U. Folli, D. larossi, A. Mucci, L. Schenetti and F. Taddei, Magn. Reson. Chem., 28 (1990) 702-710. 39. D.W. Boykin, P. Balakrishnan and A.L. Baumstark, Magn. Reson. Chem., 23 (1985) 695-697. 40. D.W. Boykin, P. Balakrishnan and A.L. Baumstark, J. Heterocycl. Chem., 22 (1985) 981-984. 41. A.L. Baumstark, P. Balakrishnan and D.W. Boykin, Tetrahedron Letters, (1986) 3079-3082.

596

42. D.W. Boykin, A.L. Baumstark, M.M. Kayser and CM. Soucy, Can. J. Chem., 65 (1987) 1214-1217. 43. D.W. Boykin, R.L. Hertzler, J.K. Delphon and E.J. Eisenbraun, J. Org. Chem., 54 (1989) 1418-1423. 44. A.L. Baumstark, M. Dotrong, M.G. Oakley, R. Stark and D.W. Boykin, J. Org. Chem., 52 (1987) 3640-3643. 45. S. Li and D.B. Chesnut, Magn. Reson. Chem., 24 (1986) 96-100. 46. S. Li and D.B. Chesnut, Magn. Reson. Chem., 23 (1985) 625-638. 47. D.B. Chesnut, D.W. Wright and B.A. Krizek, J. Mol. Structure, 190 (1988) 99-111. 48. D.W. Boykin, B. Dewprashad and E.J. Eisenbraun, J. Org. Chem., 55 (1990)425-429. 49. H.A. Christ, P. Diehl, H.R. Schneider and H. Dahn, Helv. Chim. Acta, 44 (1961) 865-880. 50. J. Reuben, J. Am. Chem. Soc, 91 (1968) 5725-5729. 51. T.E. St. Amour, M.I. Burgar, B. Valentine and D. Fiat, J. Am. Chem. Soc, 103(1981) 1128-1136. 52. S. Chandrasekaran, W.D. Wilson and D.W. Boykin, Org. Magn. Reson., 22 (1984) 757-760. 53. G. Jaccard and J. Lauterwein, Helv. Chim. Acta, 69 (1986) 1469-1485. 54. A.L. Baumstark, S.S. Graham and D.W. Boykin, J. Chem. Soc. Chem. Commun., (1989) 767-768. 55. A.L. Baumstark, S.S. Graham and D.W. Boykin, Tetrahedron Letters (1990) 957-960. 56. D.W. Boykin, A.L. Baumstark and M. Beeson, J. Org. Chem., 56 (1991) 1969-1971. 57. D.W. Boykin and A. Kumar, J. Heterocycl. Chem., 29 (1992) 1-4. 58. D.W. Boykin, J. Mol. Struct., 295 (1993) 39-45. 59. D.W. Boykin and A. Kumar, J. Mol. Structure, in press. 60. A.L. Baumstark and D.W. Boykin, New J. Chem., 16 (1992) 357-360. 61. S. Chandrasekaran and D.W. Boykin, Heteroatom Chem., 3 (1992) 63-67. 62. C. Delseth, J.-P. Kintzinger, T.T.T. Nguyen and H. Neiderberger, Org. Magn Reson., 11 (1978)97-110. 63. M. Schumacher and J. Lauterwein , J. Magn. Reson., 83 (1989) 97-110. 64. H. Duddeck, in: Atta-ur Rahman (Ed), Studies in Natural Products Chemistry, Vol. 9, Elsevier, Amsterdam (1991), pp 109-126. 65. D.W. Grandy, L. Petrakis, D.C. Young and B.C. Gates, Nature, 308 (1984) 175-177. 66. G.D. Mateescu, G.M. Yvars, D.I. Pazara, N.A. Alldridge, J.C. LaManna, W.D. Lust, M. Mattingly and W. Kuhn, in: T.A. Baillie, J.R. Jones (Eds), "Synthesis and Applications of Isotopically Labelled Compounds, 1988." Proceedings of the Third International Symposium, Innsbruck, Austria, 1721 July 1988, Elsevier, Amsterdam, 1989, pp 499-508. 67. D.W. Boykin, A.L. Baumstark, A. Mehdizadeh and M.K. Venkatramanan, Magn. Reson. Chem., 28 (1990) 305-310. 68. C. Delseth and J.-P. Kintzinger, Helv. Chim. Acta, 59 (1976) 466-475. 69. C. Delseth, T.T. T. Nguyen and J.-P. Kintzinger, Helv. Chim. Acta, 63 (1980) 498-503.

597

70. D.W. Boykin and A.L. Baumstark in: D.W. Boykin(Ed), 17o NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, FL, 1991, pp. 205-231. 71. J. K. Crandall, M.A. Centro, and S. Borresen, J.Org. Chem., 44 (1979)11841186. 72. E. Kolehmaninen, K. Laiha, M. Heinanen, K. Rissanen,R. Frohlich, J. Korvola, P. Manttari and R. Kauppinen, J. Chem. Perkin Trans. 2, (1993) 641-648. 73. D.W. Boykin, P. Balakrishnan and A. L. Baumstark, Magn. Reson Chem. 25 (1987) 248-250. 74. D.W. Boykin, D.W. SuUins and E.J. Eisenbraun, Heterocycles, 29 (1989) 301-306. 75. J.M. Cassady and M. Suffness, in: J.M. Cassady and J.D. Douros (Eds), "Anticancer Agents Based on Natural Product Models," Academic Press, New York, 1980, pp 201-269. 76. S. Chandrasekaran in: D.W. Boykin(Ed), l^o NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, FL, 1991, pp. 141-203. 77. J. P. Monti, R. Faure, A. Sauleau and J. Sauleau, Magn Reson. Chem., 24 (1986) 15-20 . 78. D.W. Boykin, D.W. SuUins, N. Phourhmady and E.J. Eisenbraun, Heterocycles, 29 (1989) 307-31. 79. R.H. Thomson, "Naturally Occurring Quinones", 2nd Ed, Academic Press, London, 1971. 80. A. L. Baumstark, M. Dotrong, R.R.Stark and D.W. Boykin ,Tetrahedron Letters, (1988) 2143-2146. 81. D.W. Boykin, A. Kumar and S.M. Shevchenko, Magn. Reson. Chem., 30 (1992) 16-19. 82. S. G. Semenov and S.M. Shevchenko, Zh. Org. Khim., 21 (1985) 368-372. 83. U. Sankawa, Y. Ebizuba, H. Noguchi, Y. Isikawa, S. Kitaghawa, T. Yamamoto, T. Kobayashi, Y. litak and H. Seto, Tetrahedron, 39 (1983) 1061-1068. 84. R.D.H. Murray, J. Mendez and S. A. Brown, "The Natural Coumarins", John Wiley, New York, 1982. 85. H. Duddeck, D. Rosenbaum, M. Hani, A. Elgamal and N.M.M. Shalaby, Magn. Reson. Chem.,25 (1989) 489-491. 86. Y. Ming and D.W. Boykin, Heterocycles, 26 (1987) 3229-3231. 87. A. Kumar and D.W. Boykin, Indian J. Heterocyclic Chem., in press. 88. H. Duddeck and A. Levai, Magn. Reson. Chem., 30 (1992) 65-67. 89. H. Duddeck, J.C. Jaseberenyi, A. Levi, T. Timar and M.H.A. Elgamal, Magn, Reson. Chem., 27 (1989) 170-172. 90. M.A. Wysocki, P.W. Jardon, G.J. Mains, E.J. Eisenbraun and D.W. Boykin, Magn. Reson. Chem., 25 (1987) 331-334. 91. K. Nagasawa, Y. Higuchi, K. Ito, M. Imanari and N.Fujii, Chem. Pharm. Bull., 41 (1993)211-213. 92. K.R. Markham,"Techniques of Flavonoid Identification", Academic Press, New York, 1982.

598

93. J.-C. Wallet, E. M. Gaydou and M.T. Cung, Bull. Liaision Groupe Polyphenols,(1990) 207-210. 94. J. Lauterwein, J.Schulte, M. Schumacher,and M. Cerny,, Magn. Reson. Chem.,30 (1992) 312-319. 95. I.P.Gerothanassis, J. Lauterwein and N. Sheppard, J. Magn. Reson., 48 (1982)431-446. 96. J. Lauterwein and LP. Gerothanasis, J. Magn. Reson. 51 (1983) 153-156. 97. J. Shulte and J. Lauterwein, Magn. Reson. Chem. 30 (1992) 334-337. 98. P.A.L. Gorin and M. Mazurek, Carbohydrate Research, 67 (1978) 479-483. 99. P.M. Dean, "Naturally Occurring Oxygen Ring Compounds", Butterworths,London, 1963. 100. S. Chimichi, R. Nesi and F. DeSio, Org. Magn. Reson. 22 (1984) 5556. 101. E.L. Clennan and M.E. Mehrsheik-Mohammadi, Magn, Reson. Chem., 23 (1985) 985-987. 102. D. C. Ayres and J.D. Loike, "Lignans", Cambridge University Press, Cambridge, 1990. 103. E. Haslam, "Plant Polyphenols",Cambridge University Press, Cambridge, 1989. 104. G.A. Cordell, "Introduction to Alkaloids", John Wiley and Sons, New York,1981. 105. T.-L. Su and K. A. Wantanabe in: Attar-ur Rahman and F.Z. Basha(Eds),Studies in Natural Products Chemistry,Vol. 13, Elsevier,Amsterdam, 1993, pp 347-432. 106. V.V. Lapachev, I.Y. Mainagashev, S.A. Stekhova, M. A. Fedotov, V.P. Krivopalov and V.P. Mamaev, J. Chem. Soc, Chem. Commun., (1985) 494-495. 107. J.C. Facelli, R.C. Contreras and M.F. Tufro, J. Mol. Structure(Theochem), 281 (1993) 61-66. 108. R.C. Contreras, R. R. Biekofsky, D. G. de Kowalewski, A. M. Orendt and J. C. Facelli, J. Phys. Chem., 97 (1993) 91-93. 109. G.A. Kalabin, D.F. Kushnarev, R.B. Valeyev, B.A. Torofimov and M.A. Fedotov, Org. Magn. Reson. 18 (1982) 1- 9. 110. H. Kahlig, P.P.Nimmerrichter and W. Robien, Monatshefte Chemie, 124(1993)71-75. 111. A. L. Baumstark, P.C. Vasquez and P. Balakrishnan, Tetrahedron Letters, (1985)4485-4488. 112. M.G. Zagorski, D. S. Allan, R.G. Salomon, E. L.Clennan, P.C. Heah, and R. P. L'Esperance, J. Org. Chem., 50(1985)4484-4490. 113. M. Manoharan and E. L. Eliel, Magn. Reson. Chem., 23 (1985) 225 231. 114. A.T. Balaban, A. Dinculescu, G. N. Dorofenko, A.V. Koblik, U.V. Mezheritskii and W. Scroth,"Pyrylium Salts, Synthesis and Physical Properties", Supp. 2, Advances in Heterocyclic Chemistry, Academic Press 1982.

599

115. A. Saba, F. S. Sib and J-P. Aycard, Spectroscopy Letters, 26 (1993) 491-496. 116. R. Hunston, LP. Gerothanassis and J. Lauterwein, Org. Magn. Reson.,18 (1982) 120-121. 117. I. P. Gerothanassis, R. N. Hunston and J. Lauterwein, Magn. Reson. Chem., 23 (1985) 659-665. 118. J. Lauterwein, LP. Gerothanassis and R. N. Hunston, J. Chem. Soc, Chem. Commun., (1984) 367-369. 119. R. N. Hunston, LP. Gerothanassis and J. Lauterwein, J. Am. Chem. Soc.,107 (1985) 2654-2661. 120. J. Tritt-Goc and D. Fiat, Magn. Reson. Chem., 29 (1991) 156-163. 121. B. Valentine, A. Steinschneider, D. Dhawan, M. L Burgar, T. St. Amour and D. Fiat, Int. J. Peptide Protein Res., 25 (1985) 56-68. 122. H. Gilboa, A. Steinschneider, B. Valentine, D. Dhawan and D. Fait, Biochim Biophys. Acta, 800 (1984) 251- 257. 123. M.L Burgar and D. Fiat, J. Phys. Chem., 85 (1981) 502-510. 124. C. Sakarellos, LP. Gerothanassis, N. Birlirakis, T. Karayanis, M. Sakarrllos-Daitsiotis and M. Marraund, Biopolymers, 28 (1989) 15-26. 125. H. C. Lee, K. Cummings, K. Hall, L. P. Hager and E. Oldfield, J. Biol. Chem., 263 (1988) 16118-16124. 126. L P. Gerothanassis, M. Momenteau and B. Loock, J. Am. Chem. Soc, 111(1989)7006-7012. 127. K. D. Park, K. Guo, F. Adebodun,M. L. Chiu, S. G. Sligar and E. Oldfield, Biochemistry, 30 (1991) 2333-2347. 128.1. P. Gerothanassis, B. Loock and M. Momenteau, J. Chem. Soc, Chem.. Commun., (1992) 598-600. 129. M. I. Burgar, D. Dhawan and D. Fait, Org. Magn. Reson., 20 (1982) 184-190. 130. S. Chandresekaran, W. D. Wilson and D.W. Boykin, J. Org. Chem., 50 (1985)829-831. 131. J. A. Coderre, S. Mehdi, P.C. Demmou, R. Weber, D.D. Traficante and J. A. Gerlt, J. Am. Chem. Soc, 103 (1981) 1870-1872. 132. H.M. Schwartz, M. MacCross and S. S. Danyluk, J. Am. Chem. Soc, 105(1983)5901-5911. 133. M.D. Tsai in: D. L. Purich(Ed), Methods of Enzymology, Academic Press, New York, 87 (1982) 235- 279. 134. S. L. Huang and M. D. Tsai, Biochemistry, 21 (1982) 951-959. 135. M. Petershiem, V. W. Miner, J.A. Gerlt and J.H. Prestegard, J. Am. Chem. Soc, 105 (1983) 6357-6358.

600

136. S. Mehdi and J.A. Gerlt, Biochemistry, 23 (1984) 4844-4852. 137. D.A. Wisner, C.A. Steginsky, Y.-J. Shyy and M.D. Tsai, J. Am. Chem. Soc, 107(1985)2814-2815. 138. M. A. Reynolds, J.A. Gerlt, P.C. Demon, N. J. Oppenhiemer and G. L. Kenyon, J. Am. Chem. Soc, 105 (1983) 6475-6481. 139. W.D. Wilson, Y.-H. Wang, S. Kusuma, S. Chandresekaran and D. W. Boykin, Biophys. Chem., 24 (1986) 101-109. 140. E. Kolehmainen and K. Laihia, Magn. Reson. Chem., 31 (1993) 763768. 141. L.L. Smith, J.E. Herz, and E.L. Ezell, Steroids, 58 (1993) 260-267.

Atta-ur-Rahman (Ed.) Studies in Natural Products Chemistry, Vol. 17 © 1995 Elsevier Science B.V. All rights reserved.

601

The Role of Biological Raw Materials in Synthesis John H.P. Tyman

1.

INTRODUCTION Although in total synthesis the nature of the starting material has been of less significance since many different intermediates have often led to the same target molecule, 'partial** and 'semisynthetic'* methodologies are selective with regard to this factor. The term semi-synthesis* implies the construction of an organic molecule by the modification of a natural product, generally of replenishable rather than fossil fuel origin, as the starting material. The majority of synthesised products have indeed in general involved the use of petrochemical and coal-derived intermediates but increasingly biological raw materials are displacing these. Semisynthesis can be considered to embrace a whole range of approaches including the building up of molecules, the degradation of a large molecule to give a smaller desired one, and even the employment of microbiological reagents to effect biotransformations. At its simplest it could be regarded as the modification of natural products by minor structural changes perhaps with the objective of structure/property studies and at the more ambitious level as a synthetic strategy for its own sake, the starting material being selected for its stereochemical rather than its economic value. The procedure of semi-synthesis has been in progress for many decades in both industrial and academic laboratory syntheses and in the latter context might be deemed historically to be the forerunner of the 'Chiron' approach (ref.l), the alternative strategy being that of asymmetric synthesis (ref.2). The chiron approach progressively and perhaps mainly through the use of carbohydrate and amino acid intermediates together with organic methodology has imaginatively expanded the semi-synthetic route and represents its logical *The terms 'semi-synthesis' and 'partial synthesis' (for this see ref .3,p66 and ref .44,p285 for another undefined use) may be conceived of as different, the former relating to a unnatural industrial material and the latter to total synthesis although both commence with a natural product. No formal definition exists and a distinction is not made here; the natural starting point seems more important.

602 conclusion. It is of interest to note that in a compilation of thirty five syntheses of major natural products, seven proceeded from the use of a natural intermediate namely those of ascorbic acid, cycloartenol, /j-carotene, penicillin, cephalosporin, bradykinin and patchouli alcohol (ref.3). Although the possibilities of organic synthesis have been enormously advanced in the last decade and it is now inevitably dominated by computer methodology, the intuitive procedure of semi-synthesis had been in existence for many years already. Increasingly it has proved essential in the pharmaceutical and food industries to examine the relative properties of diastereoisomers and enantiomeric forms and in this, semi-synthesis from a natural intermediate of known stereochemistry can be valuable (ref.4). The organic chemical industry can be beset with environmental problems and conceivably semi-synthesis can offer for at least some of these issues an alternative pathway (ref.5). The objective of this review is to recall some of the achievements of this approach and to perhaps emphasise that natural organic resources are the most varied and widespread in all countries of the world. Semi-synthesis is intrinsically associated with the use of natural intermediates of ready availability. The target molecule has invariably although not always been another natural product and it is perhaps in the context of unnatural target molecules often closely related to a natural model that semisynthesis may be most familiar.Classical examples of semi-synthesis in the preparation of CIO alcohols and other terpenoids, in penicillin and cephalosporin chemistry, transformations of morphine alkaloids, of steroids (considered as a group separate from terpenoids), of lipid resources, and the use of natural carbohydrate molecules as template molecules are described. Although it has been customary to employ standard organic reagents in these procedures, biological systems have been used notably in biotransformations on steroids. Although this aspect is strictly outside the scope of the general conception of semi-synthesis, the admission of biological reagents can be considered as a logical conclusion to the sequence, nonnatural intermediate transformed with an non-natural reagent, natural intermediate with an non-natural reagent, natural intermediate with a natural reagent. In this review examples are given both of those achievements which were adopted industrially, even in some cases if now superseded, and of less well known syntheses which have stimulated academic work and illustrated the scope of semi-synthesis.

603 The term semi-synthesis appears to imply that only half the usual synthetic effort is involved through the gratuitous availability of the right natural intermediate. It may well be generally thought of as simply making ingeniously neat single-stage but trivial functional group changes to a given carbon skeleton. It is suggested in this review that it has a more profound significance and can be conceived as an approach to multi-step synthesis which the following examples outlined may exemplify. Advocacy of semi-synthesis as a viable economic proposition presumes that the starting natural product is readily available or does not involve elaborate separational problems, an assumption which may be unjustified. In this context there is a competitive situation with conventional synthesis and biological/botanical approaches, the latter incorporating genetic and biotechnological methodology. Clearly this present account of the author's interpretation of the true nature of semi-synthesis could be directed either to show its use in the construction of members of the major classes of compounds such as terpenes, alkaloids, lipids etc.or as an illustration of the usage of stereochemically defined amino acids, hydroxy acids, carbohydrates or terpenes as intermediates in the synthesis of these various members. Accordingly a summary has been given at the end of each major section to meet both these aspects of the subject. It comprises a retrospective view effectively covering a period of nearly fifty years. The achievements in the semi-synthesis of natural products have invariably consisted of total syntheses from bio-materials due to the enantiomeric nature of the starting material. With achiral intermediates racemic products resulted. 2. THE SYNTHESIS OF SOME MONO-, SESQUI-, DI-, TRI- and TETRATERPENES 2.1 Monoterpenes (i) The acyclic CIO alcohols such as linalool, geraniol, nerol, citronellol and the cyclic series of terpineols are widely used in perfumery while in the cyclic series menthol has a particular interest in the flavour industry. The traditional source of these fragrant materials has been from essential oils such as lavender, lavendin, petitgrain, geranium, citronella and others while peppermint oil has been used for obtaining menthol. The CIO alcohols isolated from such natural sources have a characteristic odour associated with trace but intensely odorous minor components, for example, 'rose oxide* in geraniol from oil of geranium (ref.6). In

604 the period 1955-65 the greatly increased use of compounded perfumes in many newly developed consumer products, the tendency to use land for food crops, ravages of essential oil crops by adverse climatic and sometimes even political changes tended to focus attention on the possibility of readly available synthetic sources although paradoxically the absence of the trace components referred to led to the evaluation of these by perfumers as new sources rather replacements. The classical synthesis of these alcohols, such as linalol (ref.7), was already well known but had attracted little practical industrial interest until extensive development work in acetylenic chemistry arising from research in carotenoids by Hoffmann La Roche which is referred to later, led as a side development to 'pure' linalol, while from a totally different and natural source 'pine oil' (Pinus sylvestris), the Glidden Co. of Florida had separated the •^-' and (?. -pinenes and in association with the English group, A. Boake Roberts Co.,(now BBA) experimented with their comprehensive semi-synthetic conversion to the complete range of CIO alcohols. In the original innovation (ref.8) (-)-p.-pinene (1) separated from the t>-isomer by fractional distillation was pyrolysed at 600C to afford myrcene (2) in 80-90% yield together with some pseudo-myrcene Reaction of the purified former compound with hydrogen chloride at -IOC in the presence of cuprous chloride gave a complex mixture rich however in linalyl, geranyl and neryl chlorides. Nucleophilic substitution with sodium acetate in an appropriate solvent resulted in the correspondingacetates which were separated to give racemic linalyl acetate (3) as the major product, together with geranyl (4) and neryl acetates (5) as illustrated in the following scheme. Scheme 1

605 For the synthesis of (+)-citronellol (6), the mixed pinenes were catalytically

hydrogenated

to give

(-)-cis-pinane

(7) which

was

pyrolysed to (-)-citronellene. Application of the Ziegler reaction with aluminium hydride proceeded selectively at the more reactive disubstituted double bond and following atmospheric oxidation and aqueous work-up, (+)-citronellol was isolated identical with that derived by the reduction of natural

citronellal by the Ponndorf-

Meerwein-Verley method (ref.9) as shown. Scheme 2

(6) For the synthesis of (-)-menthol (8), (+)-.:^-pinene was autoxidised by aeration in the presence of a transition metal salt catalyst at the secondary allylic position and the resulting hydroperoxide reduced with alkaline sodium sulphite to afford the corresponding secondary alcohol (9)trans-verbenol together with certain isomers. Pyrolysis of the former yielded (+)-trans-isopiperitenol (10) which upon catalytic hydrogenation resulted in (+)-isomenthol (11). The desired (-)-menthol which alone possesses the required flavour properties was obtained by equilibration of the material from hydrogenation with sodium menthoxide to give a mixture of this isomer, (-)-menthol (62%), neomenthol (23%), isomenthol (12%) and neoisomenthol (3%), from which fractional distillation afforded the purified product. The traditional route to menthol involved the oxidation of (+)-citronellol to citronellal (which was also obtained from oil of citronella) its thermal cyclisation to isopulegol (12) and reduction of this to afford menthol isomers. Scheme 3

606 Scheme 4

By the use of the readily available pinenes all of

the acyclic CIO

alcohols were derived by the schemes outlined. (ii) The original procedures gave complex mixtures resulting in difficult fractional distillations on account of the stringent olfactory requirements in acceptance of the final products. Developmental work resulted in process simplifications and more selective procedures and one of the improvements was the use of pyrolytic reactions generally on the hydroxypinenes (ref.lO). The separation of the pinene isomers was avoided by their hydrogenation to cis-pinane as described and its autoxidation to selectively give a new hydroxypinane, *pinanol' (13), pyrolysis of which afforded linalol and thence linalyl acetate as shown. Allylic rearrangement of linalol gave geraniol/nerol while oxidation of the mixture resulted in the two citrals. Their reduction at the first double bond afforded citronellal, which although of little perfumery interest/ gives access to citronellol, hydroxycitronellal and the menthols. Thus a combination of semi-synthesis from pinenes together with improvements in traditional routes has given these CIO alcohols in greater availability. Scheme «^ ^s

'

^ •^''^

607

By contrast, acetylenic chemistry has been developed by Hofmann La Roche based on the use of propanone, diketene and acetylene as a pathway to linalol from the precursors »methylheptenone' and •dehydrolinalol* which itself has in turn been utilised for the synthesis of the (CI3) ionones and subsequently in related methodology for the C15 alcohol, nerolidol all of which are referred to again subsequently. 2.2 Sesquiterpenes Although advances in monoterpene chemistry have dominated the industrial scene, significant developments have taken place in other categories. Sesquiterpene chemistry contains many examples where semi-synthesis has been adopted. A useful synthesis (ref.ll)of patchouli alcohol, an important fragrant constituent of patchouli oil, from (+)-camphor, that onetime important natural product which was employed as a plasticiser for nitrocellulose (itself a semi-synthetic polymer), was complicated by structural revision of the sesquiterpene alcohol. Dihydrocarvone (14) obtained by saturation of the ring double bond in carvone, a major constituent of oil of spearmint has been employed for two very different sesquiterpenes, the ketone campherenone (15) and the alcohol, occidentalol (16). In the first case an enol acetate was converted to a bicyclic Intermediate by earlier established methodology and the route emulated a plausible biogenetic sequence giving racemic campherenone (ref.l2) as shown. Any chirality in (14) is apparently lost.

Scheme 6

(viif)

'^ (ix) (15)

(Appendix 1 lists the reagents used)

(epimer at *)

608 In the synthesis of occidentalol (ref.l3), a eudesinane-type compound consisting of a cis-fused decalin containing a homoannular 1,3-diene system, dihydrocarvone was converted by a typical Robinson annellation reaction to the basic required bicyclic structural unit. (It is of interest that a related bicyclic methyldecalenone structure, the Wieland-Miescher ketone, has been employed for the synthesis of longifolene (ref.l4), copaene (ref.15) and sativene (ref.l6) by three totally different strategies outside the present concept of the semi-synthetic approach). Scheme 7

(i)

m

(ii)

H

OH

a 6)

(S)-(+)-Carvone (17) has itself served as the starting material in the synthesis of the tricyclic lactone, eucannabinolide (18)(ref.l7) and in that of the wheat plant fungal toxin helminthospoal (19) (ref.18). By contrast,(R)-(-)-carvone has been engaged as the intermediate for the formation of the epoxydilactone, picrotoxinin (20) (ref.l9). Scheme 8

H (18)

(19) ( 17)

*(Appendix 1 lists the reagents used)

(70)

609 Another instance of the use of a natural intermediate in the sesquiterpene series is illustrated by the synthesis of a-cedrene (21)^ a constituent of cedar oil, from the acyclic alcohol nerolidol (22) a major component of cabreuva oil (ref.20). The route proceeds by way of ^T-bisabolene (23) by four steps, probably simulating the biogenetic route as shown giving racemic (21) and some epiisomer(21a). Scheme 9

(0

(ii)

(Hi) ( y- 3 ]

(22;

H (2ia) (i)HC02H,n-C5Hi2,25C,(ii)CF3C02H,n-C5H<,2,25C,(iii)H* Scheme 10: An alternative synthesis has been described also from nerolidol and although it is lengthier it comprises some unusual chemistry (ref.21). The isomer of nerolidol, farnesol in the 3(2),7(E) form has been employed in a short route for the construction of sesquicarene (24) (ref.22) the bis-methyl analogue of the bis-hydroxymethyl compound, sirenin (25), the first plant sex hormone to be characterised. Scheme 10 (»)

(24)

(25>

(i)MnO2,Hex.,0C,(ii)NH2NH2,Et3N,EtOH,(iii)MnO2,CH2Cl2,0C(iv)CuI,THF, 35C. In the two preceding examples involving an acyclic intermediate the same number of carbon atoms are present in the starting material and the product. This adjustment of functionality might be deemed semisynthesis in the conventional view. The elegant conversion of santonin (26) to dihydrocostunolide (27) represents an example in the fully cyclic series (ref. 23).

610 Scheme 11

L-'-H

,

( ?7 1

(i)Pd-SrC03 ,H2 ,EtOAc, (ii)Br2, CCl^, CHCl,, OC, (iii)LiBr,Li2C03, DMF,125C, (iv)Al(OiPr)3,i-PrOH;Al2O3,Pyr.220C,(v)hu,MeOH,-18C,(vi)RaNi,H2,MeOH, -ISC. 2.3 Diterpenes In

the

diterpene

series

a notable

recent

example

synthesis is that (ref.24) of pachiclavularolide

of

semi-

(28) a substance

isolated from the soft coral, Pachiclavularia violacea. starting from S-citronellal readily available from S-citronellol. The key step in the formation catalysed

of the macrocycle

diyne

cyclisation

rearrangement of a f

is an intramolecular

following

sulphinyl ester from

initially

a

palladium-

Mislow-Evans

S-citronellal. *

Scheme 12

^°-'^^^^-;i>(5^!^

•Experimental details on this synthesis are awaited. The tricyclic pyronolactone, forskolin (29), a polycyclic isoprenoid, has been synthesised by several different approaches (ref.25) from »^-ionone, an intermediate actually more easily obtainable from citral and propanone by Claisen condensation and cyclisation rather than from natural sources.

611 Scheme 13 (29)

2,4 Triterpenoids The synthesis of malabaricanediol (30)(ref .26) from the natural hexaene, squalene follows a biogenetic-type route a feature of many serai-syntheses. The product was obtained in racemic form. Scheme 14

(30) (i)MeC03H,CH2Cl2;H30* ,NH2CSNH2,2,3-diolsepn., (ii)aq.HCl04,*BuOH,sepn. AgN03-Si02;AC20,Pyr.,(iii)NBS/BuOH,(iv)H20,NBS;MeC03H,(v)Zn,AcOH;KOH, MeOH,(vi)Picric acid,MeNO2r20C. 2.5 Tetraterpenoids In this series the achievement of the now well-established industrial syntheses of carotenoids notably that of B-carotene has been of practical interest because of the 'permitted' colourant nature of the compound and for the legally stipulated vitaminisation

612 of certain foodstuffs. It can of course be regarded as a colourant or as a lipid, in the sense that it is a precursor of vitamin A. The first synthesis of :3-carotene (31) from fi-ionone (ref.27) involved a Reformatsky reaction with propynyl magnesium bromide to give a CI 6 intermediate two moles of which as the acetylenic Grignard reagent reacted with oct-4-ene-2,7-dione and thence (31) was obtained (Scheme 15), Although this type of synthesis was applied to lycopene and to other pigments, Hoffmann La Roche sought a more high-yielding route which would be both applicable to ^^-carotene and to vitamin A. The manufacturing route (ref .28) developed in intensive work proceeds from ':^-ionone obtained, in fact, synthetically from dehydrolinalol and diketene although the classical route based on the Claisen condensation of citral and propanone giving pseudoionone and cyclisation to the ionones represents the semi-synthetic pathway and perhaps justifies the inclusion of this example in the present account. It is evidence of the potentially competitive situation in these different methodologies. By the Darzens reaction .:-ionone afforded a C14 aldehyde which as the diethyl acetal underwent addition to ethyl vinyl ether in the presence of boron trifluoride etherate to yield a C16 acetal. After hydrolysis, loss of ethanol and formation of the diethyl acetal as before, reaction under acidic condiions with ethyl propenyl ether gave the unsaturated CI9 aldehyde after hydrolysis and removal of ethanol. Reaction of two moles with ethyne dimagnesium bromide produced the C40 chain and dehydration of the diol, selective catalytic hydrogenation followed by isomerisation completed a remarkable 'technical* synthesis of ?.-carotene. Further variations have involved the use of two moles of the CI4 aldehyde and a CI 2 divinyl ether. An independent approach (ref.29) has utilised vitamin A (32) converted to a phosphonium salt, thence to the corresponding phosphoran, autoxidation of which afforded ::^-carotene ( Scheme 16). Scheme 15

|Av^<^-^ {Jl^ A^Y^^y^^^

(i) ^:^^^9Br ^Etp ;PhCH3, 4-TSA.

(ii

(ii) 2EtMgBr, (iii) Qikx'^5^..^Y*^/Et20 , (iv) Pd-C,H2, '

613

(31)

(RCH=CHR)

H

(37)

Scheme 16

RCHOHy^ ^ RCH^RWX-^ RCH=P(Phl^RCH=:CHR 2 2.6 Summary Table 1^ collects the preceding information together with regard to the natural starting material, the product, and the nature of the synthesis whether single or multi-step. Table 1 Semi-Synthesis of various Terpenoid Compounds Natural Intermediate

Product

Pinenes (CIO)

Linalol(ClO)

Pinanes (CIO)

Geraniol/Nerol

Synthesis Type Short multi-step

Reference

8 10

Citronellol (+)-Camphor (CIO)

Patchouli Alcohol

Dihydrocarvone

Campherenone

(CIO) * "

11

(C15)

12

(C15)

,from

R(-)-carvone

Occidentalol

•Stereochemistry not stated

13

614 (C15)

Picrotoxinin

19

Eucannabolide (C15)

17

S(+)-Carvone

18

Nerolidol(C15)

Helminthosporal (C15) Cedrene (C15)

Farnesol (C15)

Sesquicarene (C15)

22

S-Citronellol

Patchiclavularolide

24

"

(CIO)

20

(C20)

»<.-Ionone (C15)

Forskolin

25

Squalene (C30)

Malabaricanediol (C30)

26

^-lonone

/j-Carotene (C40)

27

(Natural, from citral) >9-Ionone (Synthetic)

"

28

Vitamin A (C20)

29

SEMI-SYNTHESIS OF SOME HETEROCYCLIC COMPOUNDS 3.1 Penicillins, Semi-synthetic penicillins and cephalosporins (i) The Penicillins and Semi-synthetic penicillins The

chemistry

of

the

semi-synthetic

penicillins

represents

an

outstanding development in the fifties both for the pharmaceutical industry, in the form of the then Beecham Group, the medical world and for humanity in general. To some it epitomises semi-synthesis and might be considered to embody the origin of this methodology. In truth of course pharmaceuctical natural products anti-malarial

research has systematically used

as models to emulate as seen for example with the

chloroquine

compared

with

quinine.

However

a

combination of perception, multidisciplinary skills, and commercial flair has given the group concerned a unique place in pharmaceutical

615 history. The structural elucidation of penicillin in the late thirties and forties was protracted and hampered because of lack of recognition of its inherent instability due to its highly susceptible >8-lactam ring and heterogeneous composition, although an effort (ref.30) by groups, on both sides of the Atlantic, unparalleled in the history of organic chemistry was expended to solve the problem. Penicillin consisted of five components having the structure shown although by the use of improved strains of PeniciIlium notatum and chrysogenum submerged culture methods, the use of corn steep liquor and the incorporation of acids, RCOgH, in the medium, the proportion of certain component members could be selectively enhanced. Thus addition of phenoxyacetic acid afforded pencillin V, while phenylacetic acid gave penicillin G, benzyl penicillin.

RCX)

(x]^

3^

F X

Ir-HOCM CO

d'

^

An incentive for the structure determination was the expectation that synthesis would be feasible, an aim that was complicated by the finding that three different structures appeared valid and in any case the lack of appropriate protective groups, of non-acidic reagents for the formation of the ^-lactam ring and the presence of three chiral centres presented unusual synthetic problems in a superficially simple bicyclic molecule. The synthesis was achieved in 1957 (ref.31) and although it was not industrially feasible compared with highly developed fermentation routes it had an influence in drawing attention to the structure, 6-aminopenicillanic acid (33), the acyl derivatives of which constitute the penicillins. In the early fifties it was becoming apparent that penicillins V and G were losing their efficacy against several important pathogenic bacteria. This was attributed mainly to the ability of these bacteria to produce an enzyme, 3-lactamase, which was capable of cleaving the four-membered ring. The Beecham scientists investigating this inactivation reasoned that a bulky acylamino group at the 6-position

616 in the penicillin might sterically block this degradation of the lactam ring. The rate of hydrolysis by the enzyme could be influenced by the nature of the acyl group, RCO. However to obtain such alternative structures required the availability of the parent compound, 6-aminopenicillanic acid (6-APA) which, strangely, had never previously been encountered in the classical degradative work on penicillin. Concurrent with the American synthetic work the Beecham group had isolated this substance in 1959 by firstly using precursor-free broths and secondly by enzymic hydrolysis of natural penicillin by amidase and recovery of the required product, 6-APA with the aid a of liquid anion exchanger. Following development work on the acylation of 6-APA a great range of unnatural or semisynthetic penicillins (ref.32) was produced typified by structures with the acyl group RCO, as follows,

R'CO

These

PhCH(NH2)C0

Ampicillin

HOC^H^CH (NH2) CO

Amoxycil lin

2,6 - (OMe) gC^H^CO

Broxi 1

compounds

possesed

a

low

toxicity, were

capable

administration with reasonable stability to /J -lactamase

of

oral

and were

effectve against Gram positive and Gram negative organisms as indeed were the natural penicillins. Subsequently further compounds were introduced with different side chains at the 6-position which were an advance on the initial series. A contemporary account has been given also covering material dicussed in the next section (ref.33). For

the

production

of

semi-synthetic

penicillins,

Penicillium

chrysoqenum is now employed and following fermentation with a complex formulated broth containing selected ingredients, iso-penicillin N (having RCO= "02CCH(NH3*) (CH2)3CO),

is isolated by solvent extraction

and after enzymic hydrolysis, the resultant 6-aminopenicillanic acid is acylated under enzymic conditions with the appropriate side-chain acid to afford the respective final product. Scheme 17

617 As indicated, the synthesis of pencillin V (ref.31) afforded certain intermediates which were later converted to 6-aminopenicillanic acid (ref.34) but the low yield could not enable the process to compete with the fermentation route. Since the synthesis of Penicilin V commenced with valine it could of course in the present context itself be viewed as an aspect of semi-synthesis. Although loss of chirality occurred at an early stage due to the formation of an oxazolone which gave racemic penicillamine and the final product had to be resolved with loss of the chiral value of the starting material, the remarkable steps involved are shown. Scheme 18

A

4But

^

(X^

CC^^It

CC^H

(i)ClCH2COCl,NaOH,(ii)AC2O,60C,(iii)H2S,NaOMe,(iv)HCl,H2O,(v)HCO2Bu * ,Na0Bu^, (vi)A,Na0Ac,Hp0,Et0H, (vii)Pyr., (viii)N2H4;HCl;PhOCH2COCl, EtjN (ix)HCl/CH2Cl2;Pyr., (x)K0H,DCC,Resolve.

(ii) The Cephalosporins In 1956 Cephalosporin C (34) had been isolated and it was also obtained in 1961 from a variety of Cephalosporixim in the vicinity of a Sardinian sewage outlet. Although it had appeared for some time that only penicillins and cephalosporins containing A -lactam in association with S-containing rings were endowed with anti-bacterial action, the isolation of thienamycin (35) a very active and broad spectrum compound from a Streptomyces and of nocardin (36) from a Nocardia strain (ref.35) disproved this generalisation.

618

( .-. 6')

^3 4)

(37 )

Cephalosporin C was found to possess greater stability to acidic conditions and to

A" lactamase than the penicillins yet a weaker

anti-bacterial action which however was enhanced of

7-aminocephalosporanic

acid

(37)(7-ACA),

in acyl derivatives obtained

by

acidic

hydrolysis, separation through the classical method of precipitation at the iso-electric pH, succeeded by reacylation.

Refined chemical

procedures were developed, by the Eli Lilly group, based on the use of nitrosyl chloride in formic acid which gave improved yields of 7ACA and thence a range of unnatural or semi-synthetic cephalosporins. Their broad spectrum activity, often against organisms which had become resistant to penicillins through lactam ring cleavage, nontoxicity and acid stability led to many new compounds of this type. An expectation that it would penicillins

generally be

feasible

to transform

to cephalosporins (ref.36) has proved possible for those

lacking the 3-acetoxy group as shown by the conversion of the phenoxy compound,

penicillin V to the 3-methyl compound. Cephalexin (38).

Scheme 19

.-4P< P^^ CO^H

FhO

^^^

C9^l3

(iii)

¥^^

(vi)

ic^^v-ca^

619

(38)

OC^H (i)Cl3CCH20H,DCC,Pyr.,(ii)MCPBA,(iii)DMF, A (iv)PhH,PCl5;MeOH;H20, (v) CljCCHgOgCNHCH (Ph) COgH, DCC, (vi) PG removal. Since no straightforward enzymatic or fermentation method is accessible comparable to that used for the penicillin series the formation of semi-synthetic cephalosporins has generally consisted in the chemical removal of the acyl side-chain from fermentationproduced Cephalosporin C to afford 7-ACA, the carboxyl group being protected as an ester at this stage, succeeded by reacylation and modification of the group at the 3-position. Many semi-synthetic cephalosporins have been obtained by this strategy which resembles the original approach adopted. An example is the derivation of ceftazidime (39) a powerful broad spectrum compound.

(39)

HO^C Cephaloporin C is readily obtained through fermentation from the organism Cephalosporium acremonium in a medium of selected ingredients as with the production of penicillin although the work-up procedure involves more unit operations including an adsorption stage. It has been synthesised (ref.37) by a twelve stage route and since this commences with L(4-)-cysteine it constitutes a facet of semi-synthesis. In this approach the chirality of the starting amino acid is preserved throughout, as illustrated in the following scheme. Scheme 20

—•- H - 7 V + ' ^ > - ^ 'I"'

620 hi H

I^CCl^

(viii) a

OOpH

(Appendix 1 lists the reaqents used) (iii) Vitamins Vitamin chemistry contains a variety of structures the majority of which are heterocyclic or lipidic in type and apart from Vitamin B^2 they are obtained synthetically although with the exception of the vitamin D series few start with a natural product intermediate. For certain of the heterocyclic members attention has turned inevitably to the amino acids as possible sources of starting material. (a). L(+)-Cysteine has been utilised in a compact seven-stage synthesis (ref.38) of the important nutritional substance (+)biotin by the steps illustrated below. Scheme 21 * HCl-H^M^

''^ ^

HNTTvlBn

Hir^JHBn

(v)

^"\

H N r \ j B n (iiO^

H^lAsiBu (iv)

HN

^

H'

H'^

^.^^^^ O

Hr/S>©n

(viii

H^sr\H

H-4

HH ^"^^^"^xX^H

( b ) . The f i r s t step in Karrer's synthesis(ref.39) and the current method for the synthesis of riboflavin (vitamin 83) namely the formation of a Schiff's base from D(-)-ribose and 3,4-dimethylaniline lAppendix 1 l i s t s the reaqents used)

621 comprises an element of semi-synthetic methodology. In the previtamin era the oxidation of nicotine (from Nicotiana tabacmn) with nitric acid (ref.40) to obtain nicotinic acid, an intermediate for the synthesis of nicotinamide, must be one of the earliest degradative aspects of semi-synthesis and could conceivably have relevance at the present time as a partial alternative use of an agricultural product in a society which is increasingly rejecting smoking.

O^— or' (c). The Vitamin D series The anti-rachitic factors resulting from the ultraviolet irradiation of cholesterol (40) and ergosterol (41) namely vitamin Dj^ (42) cholecalciferol and Dg, ergocalciferol (43) respectively are important dietary materials the chemistry of which was only elucidated by the investigations of many chemists (ref.41). Vitamin D3 is most easily derived by semi-synthesis from cholesterol through formation firstly of 7-dehydrocholesterol by reaction with Nbromosuccinimide followed by dehydrobromination with collidine. Ultraviolet light irradiation affords previtamin D3 which is thermally isomerised to the endo compound shown and thence to the exo final product (chair form).

(i)NBS;collidine,(ii)(iii)hv,(iv)A,

622 Ergosterol undergoes similar changes although with the 7,8 unsaturation already present ring B opening occurs facilely with the result that in addition to vitamin D2, lumisterol, tachysterol and the 5,6-trans vitamin are formed. A summary is given in Table 2 of the examples of semi-synthesis discussed. Table 2 Semi-Synthesis of some Heterocyclic Compounds Natural

Product

Synthesis Type

Reference

Semi-synthetic

Two-stage

32,33

Intermediate Penicillin N

Penicillins Valine

Penicillin V

Multi-step

31

Penicillin V

Cephalexin

Multi-step

34,36

L(+)-Cysteine

Cephalosporin C

Multi-step

37

L(+)-Cysteine

Biotin

Multi-step

38

D(-)-Ribose

Riboflavin

Multi-step

39

Nicotine

Nicotinic acid

Single step

40

Cholesterol

Cholecalciferol

Two-stage

41

(vitamin D3)

4. SEMI-SYNTHESIS OF STEROIDS 4.1 Formation of cycloartenol Although cycloartenol (44) belongs to the tetracyclic triterpenoids as the parent compound, it is implicated as an important biosynthetic intermediate both in that class and in plant steroids and is thus a relevant introduction to the latter group which have been so significant in semi-synthetic applications. Following the enzyme-catalysed cyclisation of squalene epoxide, it is the first intermediary compound encountered and in the plant series it has a comparable position to lanosterol (45) in the animal

623 world. The conversion of lanosterol to cycloartenol (ref.42) affords a link between the two groups and introduces the ingenious methodology of the Barton reaction, namely the photolysis of nitrite esters. The scheme consisted in acetylation, protection of the double bond by bromination, oxidation to afford the 6,11-dione, selective reduction to give the 6~methylene,ll-ol, liberation of the 3-ol and reduction of the dibromo compound. After selective 3benzoylation, formation of the 11-nitrite ester, iodination afforded the 19-iodomethyl,ll-ol,which upon oxidation gave back the 11-one. Strong base treatment yielded an anion at the 9-position and following formation of the cyclopropane ring, reduction of the 11one yielded the final product as shown. Scheme 23

(44) .C20;Br2,(ii)Cr03,Zn,AcOH(iii)NH2NH2,KOH>LAH(iv)BzCl,Pyr. (i)Ac NOCl ,Pyr., (V) hv, I2 ,(vi) CrOj/Cvii) -QBu^ HOBu^/(viii) LAH. 4.2 Steroidal Hormones from Sapogenins and Other Natural Sources (i) Gestogens Semi-synthesis of steroids has been a rich field for the derivation of a whole range of structures at a time when total synthesis was unknown and although subsequently for certain compounds notably estrone remarkable developments have led to total synthesis, the use of natural intermediates has proved a major contribution. In this area, work initiated on semi-synthetic transformations of plant

624 sapogenins enabled many compounds in the gestogen, estrogen and androgen classes to be obtained readily for the first time. These contributions typified by the synthesis of progesterone (47) from diosgenin (46) (ref.43) from a Mexican Dioscorea and the isolation of other sapogenins had a dramatic influence on hormone production at that time leading amongst other developments to the setting-up of the Syntex group. The structure of the saponins and the sapogenins has been comprehensively discussed (ref.44). The conversion of diosgenin (ref.43) by acidic cleavage of the glycoside ring ,oxidation of ring E, further steps of side-chain elimination and catalytic reduction to afford pregnenolone in high overall yield is shown below, the remaining steps of alkaline hydrolysis/isomerisation and Oppenauer oxidation giving (+)-progesterone. Scheme 24

( 41 !

(i)Ac20,H30% (ii)Cr03,AcOH, (iii)AcOH,H2Pd-C, (iv)"OH;Oppenauer.

Prior to the use of diosgenin the semi-synthetic alternative of obtaining progesterone from cholesteryl acetate involved unattractive and low-yielding stages such as oxidation of the dibromide with sidechain degradation. By contrast and subseguent to the above pathways, stigmasterol (48) from the soya bean could be converted as shown, in two steps in higher overall yield, than even by the diosgenin route, to progesterone with however the problematic use of two ozonisation stages although the oxidation of the enamine could also be effected with sodium dichromate. Nevertheless a good route to progesterone proved at that time of value in the early stages of the development of methods for the corticosteroids to improve on the lengthy syntheses which had been embarked upon and which commenced with the naturally-derived intermediate methyl desoxycholate.

625 Scheme 25 ( 48 47

(i)03,CH2Cl2,(ii)C5H„N, A,(iii)NapCrp07,AcOH, PhH. The cardiac aglycone^strophanthidin has been investigated (ref.45) in semi-synthesis for obtaining in seven straightforward steps, 19norprogesterone a compound with eight times the activity of progesterone itself. Mention should be made of the high activity found with the more recent introduction of the related compounds norethynodrel and norhisterone readily obtainable from estrone methyl ether and of the angular ethyl analogue, norgestrel available from a number of asymmetric synthetic routes (ref.46) rather than semisynthesis.

(ii) Estrogens The formation of 19-norsteroids such as estrone by the aromatisation of ring A in andostradiene-3,17-dione and andostratriene derivatives from cholesterol intermediates led finally to 6-dehydroestrone which gave estrone in high yield by hydrogenation. Following the first total synthesis of equilenin by Bachmann et.al., interest had turned to that of estrone and although semi-synthetic methodology gave early success in achieving this, advances in a variety of total syntheses of estrone and analogues superseded this initial situation. In work by Marker and the Syntex group (ref.47) diosgenin was converted to a spiro l,4,6-trien-3-one which was aromatized by pyrolysis at 600C to a 1,3,5(10),6-tetraene. Hydrogenation to the l,3,5(10)-triene and degradation of the side-chain of the methyl ether afforded 3-methoxy-17-acetylestra-l,3,5(10),16-tetraene which proved to be a general intermediate for obtaining 19-norhormones. For estrone the foregoing tetraene was oximated, submitted to Beckmann rearrangement with 4-acetamidobenzenesulphonyl chloride and the resulting enamine acetate hydrolysed by acid to (+)-estrone methyl ether (49) as shown.

626

(Appendix 1 lists the reagents used) Recent interest has been directed to using cheaply available cholesterol in a combination of chemical and microbiological procedures. Chemical formation of the 10-hydroxymethyl derivative followed by microbiological, rather than lengthy chemical degradation of the side-chain by incubation with a soil microorganism (CSD-10), afforded estrone in 72% yield. More recent innovations include the use of hydroxylating microorganisms to replace the first chemical steps and the use of microbiological transformations throughout (ref.48). Scheme 27

(i)4 steps(chem.), (ii)l step(biol.). (iii) Androgens For the formation of testosterone (51), androsten-5-3-ol-17-one (50) obtained by oxidation from cholesterol was employed although it was also derivable from diosgenin by way of 16-dehydropregnenolone as described for the preparation of progesterone. The synthesis of the androgen followed the route shown and the acetate was isolated in an overall yield of nearly 40% in the Syntex procedure from the androstenone intermediate. Scheme 28 ^^ ^ p pAc (51)

(i)HC02H, (ii)NaBH4;Ac20,4-TSA, (iii)Oppenauer.

627 In a diosgenin-based route, 16-dehydropregnenolone was converted to the 16,17-epoxide which readily afforded the 17-hydroxy compound that as the acetate-oxime followed by Beckmann rearrangement, as described for estrone afforded an androstenolone intermediate and thence testosterone (ref.47). In more recent developments androstenolone from various sources has been converted micobiologically to the corresponding dione and then testosterone obtained by selective reduction with sodium borohydride. Because of their high androgenic activity testosterone and its 19-nor analogue have also become important targets by total synthesis. Numerous other sapogenins indigenous to the USA and Mexico were investigated by Marker for conversion of their C29 structure to that of the Cg-iSteroids and so thorough was this search that it stimulated field work in other continents particularly when larger-scale corticosteroid investigations gathered pace, (iv) Corticosteroids Although the hormones of the cortex or outerstructure of the adrenal gland had been worked on for many years by American and Swiss groups, their potential for the alleviation of rheumatoid arthritis raised expectations and requirements for their increased production in the period of the early fifties. The requirement for oxygenated groups at C-i^and C-,^ in steroid structures presented new synthetic problems. Important compounds were corticosterone (52), cortisone (53) and Cortisol (54).

^-OH

I 54)

For the preparation of substantial amounts of cortisone, a semi-synthetic multi-stage route comprising more than twenty steps was developed to a production scale by the Merck laboratories (ref.49) commencing with methyl desoxycholate from the corresponding acid, a relatively minor component together with the major material, cholic acid in ox bile, the former posessing a useful hydroxyl group at C-,2- The key stages of these transformations are shown.

628 Scheme 29

(i) BzCl, Pyr.; CrOj ; SeO^,; KOH; MeOH, H*; Pt-Hg ;MeOH, H*, (ii) HCl, CHCI3 ; NaHCOj , (iii)Br2,-78C;Ag2CrO^;Cr03,(iv)PhMgBr,-H20;HBr;Zn-AcOH;NBS;HBr,(v)NBS; KOAc;ACpO;Cr03, (vi)HCN;POCl,,Pyr. ;KOH, (vii)AC20;Os04;Cr03;Na2S03, (viii) AC2O; Br2, (ix) 2 , 4-DNP, NaOAc-HOAc; MeC0C02H, HOAc, CHCI3, HBr A large number of other semi-synthetic routes were pioneered from 3-/3-hydroxypregnane-11,20-dione obtained from cholesterol, ergosterol, stigmasterol or diosgenin. An attractive route was based on hecogenin (55), freely available from the East African sisal plant, (ref.50) by conversion to the dibromide which was transformed to the 12-hydroxy-11-one and thence as the diacetate by reduction to the 11-keto compound shown in an overall yield of 55% from hecogenin. Side-chain Marker degradation afforded cortisone (ref.51) after further steps. Scheme 30

(i)AC2 0 ; B r 2 ; ' 0 H , ( i i ) Z n / H 3 0 * ; A c 2 0 ; C a / N H 3 ( l i q . ) , ( i i i ) A s Scheme24,(iv) PBA, (v)HBr;Br2/(vi)KI;NaoAc;Ni/EtOH;"OH.

for

629 Many advances were achieved during this period by a variety of chemical approaches. A dramatic microbiological finding by the Upjohn group that progesterone could be selectively converted to the ll-ahydroxy compound in 50% yield by the soil microorganism Rhizopus arrhizus and subsequently in 80-90% yield with Rhizopus nigricans was reported (ref .52) and presaged a fresh approach. The formation of the 11-^'hydroxy compound led to a synthesis of cortisone acetate through the steps shown consisting of hydrogenation of the 4,5-double bond, oxidation of the 11-hydroxyl group, conversion to the 20-enol acetate, epoxidation and cleavage of the epoxide to give the 17hydroxy derivative, bromination to afford the 21-bromomethyl leading in two steps to the acetoxy derivative. The final stage involved restoration of the 4,5-unsaturation (ref.53). Scheme 31

'"C^ (53)

(vii)

( i ) R h i 2 o p u s a r r h i z u s , ( i i ) H 2 ,Ni,"OH, ( i i i ) C r 0 3 ; N a B H 4 , P y r . ;Ac20, ( i v ) A c 2 0 , 4-TSA;PBA;H30* , (v)Br2 ,Ac0H;"0H;MeC0NHBr,t-Bu0H;AC20, ( v i ) B r 2 rAcOH, ( v i i ) NHgCGNHNHg; MeCOCOgH; H3O*,

In the second approach, 16-dehydropregnenolone was employed, the double bond of which at the 16,17- position was epoxidised, the epoxide cleaved with hydrogen bromide to the bromohydrin which was reduced and following formation of the 3-formoxy derivative, acetylation of the 17-hydroxy compound gave an intermediate which was oxidised by the Oppenauer reaction to give cortexolone (56) as the diacetate to which the microbiological oxidation could be applied. The process is outlined. Scheme 32 (56)

OHCO^ ( i ) PBA; HBr; Raney N i ; Brg; K i ; "OAc, ( i i ) HCOgH; ACgO, P y r . ; Oppenauer.

630 During

the latter stages of work on the components

extracts a highly active substance, aldosterone having

the

structure

corticosterone

shown.

Starting

from

in adrenal

(57) was isolated the

21-acetate

of

a compact synthesis of this compound was developed

(ref .54) as shown involving initial treatment with nitrosyl chloride followed by nitrous acid. Scheme 33

^NA-

QAC

H9

QAc

liii) f 57) (i)N0Cl,Pyr.,(ii)hu,(iii)HN02. Many more active compounds such as prednisone, prednisolone and several f luorosteroids have been introduced involving microbiological dehydrogenation at the 1-position in the first two and in the majority of compounds, the use of semi-synthetic routes. The

preceding information is summarised as follows.

Table 3 Semi-Synthesis of Steroids Natural Intermediate

Product

Lanosterol

Cycloartenol

Synthesis

Reference

Type Multi

42

Progesterone

Four steps

43

norProgesterone

Multi

45 47

Diosgenin Strophanthidin Diosgenin

Estrone testosterone

II II

II

Cortisone

II

49

II

II

50

M

Desoxy cholic Acid Hecogenin Progesterone

11-Hydroxy

Single step

52

631

11-Hydroxy progesterone Corticosterone acetate

Cortisone

Multi-step

53

Aldosterone

Three steps

54

5. SEMI-SYNTHESIS IN ALKALOID CHEMISTRY In this area some interest has lain firstly in the semi-synthesis of alkaloids from certain natural intermediates and secondly in functional group transformations in structural/property studies. 5,1 Synthesis from Steroidal and Alkaloidal Intermediates Steroidal intermediates have been employed in the synthesis of certain alkaloids such as dihydroconessine (58)(ref,55), one of the 3-aminoconanine bases, from the bile acid compound 3-Bacetoxybisnorcholenic acid, a substance obtained from appropriatelyprotected coprostanol by oxidation and two Barbier-Wieland degradations. The method consisted in converting the carboxyl into a formylamino group, the acetoxy into a tosylate which was then substituted by a dimethylamino group, reduction of the formylamino to methylamino and formation of the five-membered ring E by chlorination of the CIS methyl and cycloalkylation of the methylamino group as illustrated in the following scheme. Conessine and its dihydro derivative are used for treating amoebic dysentry. Scheme 34

^-^o

(58)

(i)S0Cl2 ,Pyr. ;NaN3;NaOH,H20, A;ACOCHO, (ii)K2CO3,MeOH;TsCl,Pyr,0C(iii) HOAc,Et2O,(iv)Me2NH,DMF;LAH,THF,A,(v)PtO2,H2,AcOH;NCS,Et2O,(vi)hv,90% HgSO^, [ (n-Bu)2NCl] ,0C. In the indole series or Rauwolfia alkaloids, corynantheine (59) a

632

( 60^

( 6 ]. •)

^O^C

"(y^

^^°2^

OMc

(62:

relative of yohimbine (60) and reserpine (61) has been synthesised (ref.56) from yohimbone which itself was derived from yohimbine through oxidation by the Oppenauer procedure (ref.57). The methodology involved successively the introduction, adjacent to the carbonyl group, of an a-hydroxymethylene function, conversion to a methylthio group, and following oxime formation, a novel Beckmann rearrangement. The remaining steps are as shown in Scheme 35 .

(i)HC02Et,NaOMe,PhH,(ii)TsSCH3,NaOMe,(iii)NH20H,"OH;SOCl2,Et20,(iv) Raney N i , EtOH; "OH; HCl, MeOH, (vi) HC02Me, Ph3C", THE, (vi) MeOH, HCl. Yohimbine,

a

synthesised

(ref.58) in an elegant six-step route from

compound

with

aphrodisiac

properties,

has

been

tryptamine

derivable in theory from tryptophan a major constituent amino acid from casein. Other alkaloids of the indole group such as vincamine

633 (62) have been obtained by way of this amine (ref.59). The chemical and biological studies on reserpine, the most important member of the large group of alkaloids from Rauwolfia have led to the development of drugs acting on the central nervous system and to investigations in neuropharmacology. The synthesis itself from 6-methoxytryptamine (ref .60), although outside the ambit of semi-synthesis, was modified to effect commercial production when the great demand for the compound as an anti-depressant made this competitive with its natural extraction. Remarkably, although many modifications by semisynthesis were studied no new derivative has displaced the parent alkaloid, a situation which contrasts with the position of morphine in the analgesic area. 5.2 Semi-Synthetic Transformations of the Morphine Alkaloids Although it has been synthesised (ref 61), (now by four distinct routes) morphine (63), a powerful analgesic is readily available from the natural source, opium obtained from Papaver somniferum as the major alkaloid present (approx. 10%) and indeed the first to be isolated by Serturner in 1803, while the methyl ether, codeine (64) which comprises only 0.5% is relatively weakly active. Thebaine (65) a third component of opium is non-analgesic but is important as an intermediate for the formation of codeine, a somewhat scarce commodity (ref .62) but a valuable mild analgesic and anti-tussive in great demand. It is best prepared from morphine by methylation with phenyl trimethy1ammonium ethoxide the by-product consisting of dimethylaniline (ref.63). Since for general clinical use morphine has a number of disadvantages the main one of which is its addictive property, many attempts have been made to modify the structure by semi-synthesis even before the structure had been fully established (ref.64). In parallel with the use of morphine and thebaine as natural intermediates many compounds having a simple structural resemblance to morphine have been studied for their potential analgesic activity without addictive and other side effects (ref.65), a search which continues to this day. The contribution of semi-synthesis can be categorised into a first generation group of compounds certain of which are still in use and a second group introduced in more recent years (refs.65,66). For the first group, high analgesic activity (between two to five-fold activity in terms of the weight required by comparison to morphine, with considerably lower ED50values) has been found for the diacetyl

634 compound (heroin)(66), 7,8-dihydro-6-oxomorphine (dilaudid)(67), the 5-methyl analogue of dilaudid (metopon)(68), the 14-hydroxy analogue of dilaudid

(numorphan) (69), and the 6--desoxy analogue of dilaudid

(dihydrodesoxymorphine)(70),

while

in

the

codeine

series,

dihydrocodeine was twice as active as codeine which has only

7,8one

tenth the activity of morphine. All the transformations from morphine result from single or two stage, or in the case of numorphan, threestage reactions. Addictiveness was still evident in these compounds but a remarkable finding

was that the N-allyl analogue of morphine

(nalorphine) (71) was an antagonist.

(65)

f 63)

r 64)

AcO (

HOv

f:f.)

AcO

( 69)

In the second generation of morphine-based analgesics certain of the structural features of the first group have been preserved and a wider variety of N-alkyl groups studied. A remarkable finding has been the high activity exhibited by the Diels-Alder adduct of thebaine (ref.67) with methyl vinyl ketone which has a comparable activity to morphine, although thebaine itself is totally devoid of analgesic properties, and furthermore the enormously enhanced analgesic potency, some 500 times that of morphine, of the tertiary alcohols (ref,68) obtained by the action of Grignard reagents on the ketone. Demethylation of the phenolic ether afforded still further

635

increased effect (ref.69). The reactions involved are shown and essentially consist of the employment of acrylonitrile, acrolein, methyl vinyl ketone, diethylmaleate and 1,4-benzoquinone with thebaine to afford 6,14-endo-ethenotetrahydrothebaine derivatives. Both the methyl vinyl adduct and its ethano analogue were converted to t-alcohols by reaction with a variety of Grignard reagents and thence demethylated under basic conditions to give a range of novel analgesics having remakable properties. One of this group buprenorphine (72), prepared from thebaine is an N-cyclopropylmethyl derivative.(Scheme 36) In the group of other purely N-alkyl compounds may be mentioned Nallylnorcodeine (73) (ref.70), and both N-alkyl and N- cycloalkyl norcodeine derivatives (74) (ref.71) were compounds described many years ago and more recent compounds in this series although considered new, have been a rediscovery.

McO ^72 J

(74)

Nv.^ HO' The precursor was obtained by demethylation of the parent N-Me compound by the use of well-established reagent, cyanogen bromide. Others are butorphanol (the only compound derived from a non-natural source) a 14-hydroxy-N-cyclobutylmethyl compound in which the oxide

636 ring is no longer present, naloxone an N-allyl analogue of numorphan and naltrexone the coresponding N-cyclopropyl compound. Many of these compounds exhibit good analgesic activity with a low abuse potential and freedom from the side effects found with morphine. Table 4 Semi-Synthesis of certain Derivatives of Morphine Natural Intermediate

Product

Synthesis

Reference

Type

Morphine

Codeine

single

63

Codeine codeine

N-allylnorcodeine N-cycloalkylnorcodeine

step Two-step "

70 71

Thebaine

Keto derivatives of 6,14 Single-step

67

-endo-ethenotetrahydo thebaine "

t-Alcohols derived

Single-step

68

two-step

69

from ketones "

6.

Phenols from t-alcohols (cf.Buprenorphine)

SEMI-SYNTHESIS OF CARBOHYDRATES AND RELATED STRUCTURES

This area represents the essence of semi-synthesis, perhaps in its most esoteric aspect. The anticipated success of using stereochemically-defined starting materials has often entailed much experimentation to select the most appropriate compound for the task. As with other groups of natural products there are many classic examples of the use of semi-synthesis as well as more recent highly developed cases in what has come to be termed the chiron approach. 6.1 Synthesis of Ascorbic acid and of Disaccharides (i) The synthesis of Vitamin C (75), ascorbic acid from Dgalactose (ref.72) and that from D-glucose (ref.73) were described in the same year which also saw the development of the the latter as an industrial route. Sorbitol obtained by the reduction of D-glucose was oxidised with the bacterium Acetobacter suboxvdans to L-sorbose which as the bis-acetonylidene derivative was next oxidised to the corresponding acid, L-2-oxogulonic acid, the methyl ester of which was base-catalysed to afford the final product as illustrated.

637 Scheme 37

HO

OH OH

H HO OH

H

(15)

(i)H2 ,Pd-C, (ii) Acetobacter suboxydans , ( iii )Me2 CO , H* / iV)KMnO^/KOH, (vi) MeOH/HCl, (vii)NaOMe. A two-step microbiological route to 2-oxogulonic acid has recently been devised (ref.74) in which D-glucose is converted by the organism, Erwinia herbicola, to D-2,5-dioxogulonic acid which is then transformed by Corynebacterium to L-2-oxogulonic acid convertible by warm acidic treatment to ascorbic acid (vitamin C). In the onestep procedure D-glucose is directly transformed to 2-oxogulonic acid by the generation of a recombinant Erwinia herbicola containing the gene for 5-ketoreductase obtained from Corynebacterium. (ii) The synthesis of ( + )-sucrose (76), (a-D-glucopyranosyl j3-Dfructofuranoside) has been achieved (ref.75) from the reaction of 1,3,4-tri-O-acetyl-D-glucose and 1/2,3,4-tetra-O-acetylfructofuranoside, followed by deacetylation, chromatography, reacetylation and chromatography. The yield of sucrose was very low by this procedure although the success following so many earlier attempts has been referred to as 'the Mount Everest of organic chemistry'. Scheme 38

(i)A,100C,-AcOH; Prep.Chrom.;ACgO;Prep.Chrom.

-AcOH.

Recently there has been interest in the enzymatic conversion of glucose to an equal mixture with fructose since this has the same sweetness as sucrose. Glucose, readily available from starch by the action of amylolytic enzymes, has been submitted on a large scale to the action of glucose isomerase in an insolubilised form (ref,76) to achieve this transformation.

638 6.2 Semi-synthesis of Pyranose derivatives Small acyclic components, C5 and C6 alicyclic compounds have been employed as intermediates in addition to the readily available monosaccharides such as glucose or mannose. In this context these intermediates are template structures. Thus, (S,S)-tartaric acid (77) has been employed for constructing part of the structure of the antibiotic, indanomycin (ref.77).

Butenolides derived from ribonolactone (78) have been valuable in the construction of several bicyclic natural products (ref.78). D(+)Ribonolactone as its trityl derivative was converted to the 0,0thiocarbonate and thence by reduction to the butenolide shown which with various buta-l,3-dienes readily formed bicyclic structures. Scheme 39

"SXr^Xr-^^^ Y H6 6H

(i)Ph3CCl,Pyr.,(ii)0,0-Thiocarbonate,(iii)Raney Ni. Comparable six-membered rings have been derived in extensive studies (ref.79). Scheme 40

(i) KI, MegCO,(ii)Zn,(iii)hydrolysis;MnOg.

In more complex examples, diacetone derivatives of glucose, of mannose , galactose and glucoheptonolactone have played a vital role for the specific synthesis of polyhydroxylatedpiperidines a group

639

OH

9*^

QH

(79) SH A ^ rBo^V^-^^^^^ TT^^ H H of compounds which are powerful inhibitors of glycosidases, enzymes which can selectively hydrolyse one sugar from an attached one. Deoxynojirimycin (79) and deoxymannojirimycin (80), having the structures shown, are amino sugar derivatives structurally closely related to D-glucose and D-mannose respectively (ref.80) and the problems involved in their semi-synthesis and that of deoxyfuconojirimycin (81) from D-lyxonolactone have been discussed (ref.81). In practice the route from mannose involving a double inversion at C5 and that from glucose, involving a pyranose pathway, were unrewarding although with a furanose analogue an improvement was effected. By contrast L~gulonolactone (82), a substance readily available from the hydrogenation of vitamin C, offered a convenient intermediate (ref .82) for the connection of C5 and CI and was found to involve few protective groups although the overall yield of (80) was slightly lower than by the use of D-glucose (as the bisacetonylidene derivative. The preferred route is illustrated. Scheme 41 HO QH (Mel^BJSiO QH(Me]^BjSi<

(i) ^"%^vo iHj ^~N^>o (rn)

(iv) -^

180)

(i)MepBu*SiCl;Me2CO,H*,(ii)(CF3SO)20,NaN3,(iii)H2,Pd-C,(iv)BH3Me2S,H\ As was evident in the semi-synthetic chemistry of steroids many natural starting matrials are predictable as potentially available, yet only detailed experimentation can resolve the choice. The field of monosaccharides, disaccharides and polysaccharids contains many industrial examples of semi-synthesis, which have entered chemical history, such the hydrolysis of starch to form Dglucose, the esterification of cellulose, the chemistry of rayon and, more recently, advances in the chemistry of carbohydrate surfactants (ref.83).

640 Table 5 Semi-Synthesis of certain Carbohydrate Derivatives Natural Intermediate

Product

Glucose Galactose

Vitamin C

Glucose,Fructose Glucose (S,S)-Tartaric acid

Sucrose Fructose Indanomycin Structure Bicyclic Produts Natural products Polyhydroxy Piperidines

Ribonolactone

L-Gulonic acid

11

Synthesis Type Multi-step II

Single-step 11

Reference

72 73 75 76

Multi-step

77

Multi-step

78

It

79

H

82

7.

SEMI-SYNTHESIS IN LIPID CHEMISTRY Lipid chemistry contains many instances of semi-synthesis in the chemistry of the fatty vitamins, prostaglandins and leukotrienes, the phospholipids, and the glycerides with their structural relatives. 7.1 Vitamins The lipidic vitamins (ref.84) include vitamin A (32), a substance intrinsic to the physiology of vision, vitamin E (83), a natural protective antioxidant, and vitamins K^ (84) with K2 (85), antihemorrhagic compounds, each of which is derivable from an initial natural product intermediate. Although traditionally n ionone obtained from citral (a major constituent of lemon grass oil) was used for the synthesis of vitamin A, a synthetic source has now replaced this in a process which also gives /g-carotene. In one method the CI4 aldehyde in that process is reacted with a C6 eneyne component and selective hydrogenation followed by dehydration and isomerisation affords the final product (ref.85). Scheme 42

(i)ClCH2C02Et,NaOEt, (ii)-OH, (-COp) , (iii)A,2EtMgBr; H,0* (iv)Pd-C,Hp,(v) ACgO, (Vi)l2.

^

641

BrM30^

^^Br

Vitamin E, a-tocopherol, an effective antioxidant, is obtained as a diastereoisomeric mixture in the 2,3'(S),7*(R),11*(R), form and the natural 2,3*(R),7'(R), 11'(R), form (83) by the reaction of phytyl bromide with trimethylhydroquinone in the presence of zinc chloride (ref,86), the phytol required being obtained from chlorophyll.

Yk-o>p^^^^^Y^^-^^^

'83)

Scheme 43

Vitamin K^(2-methyl-3-phyty1-1,4-naphthoquinone (84) has been unambiguously prepared from l-acetoxy-2-methyl-4-hydroxynaphthalene and phytol under mildly acidic conditions and the intermediate product hydrolysed and oxidised (ref.87). (Scheme 43)

(84)

( Compound B is a by-product if the 1,4-diacetate is used)

After some initial structural uncertainties, vitamin Kg, 2-methyl-3farnesylgeranyIgerany1-1,4-naphthoquinone, was obtained following a route similar to that for the phytyl analogue (ref.88).

rss)

642 7.2 Prostaglandins, Thromboxanes and Leukotrienes This important group of substances (ref.89,90) contains a number of instances where a natural product intermediate has been selected as the starting point for one reason, to avoid the problem of a resolution at some stage. Terpenoids, carbohydrates and poyunsaturated fatty acids have been employed at the outset for this purpose. (i) Terpenoids For the prostaglandins (-)-pulegone has afforded a pathway to the cyclopentano compound shown (86), a compound which in turn with a variety of different protective groups has enabled other members of this class to be derived (ref. 91). Scheme 44

(viiO

9 ^

Bn

! S6 }

PbzO (i)PhMgBr,CuCl;(ii)Na,i-PrOH,PhCH3A,(iii)H2C=CHCOCl,Et3N,(iv) (^)'^OBn ,AlClj,CH2Cl2, -55C,(v)LDA,THF,-78C;02, (EtO)3P ,THF,-78C, (vi) LAH, Et20, (vii)NaIO^,t-BuOH/H2O,pH7;H2O2NaHCO3;Kl3,NaHCO3,H2O,0C, (viii)PBzCl,Pyr.;(n-Bu)3SnH,AIBN,PhH,55C;H2,Pd-C,HCl,EtOAc/EtOH. (ii) Carbohydrate intermediates In the thromboxane group (ref.92) ,or-methyl-D-glucoside has afforded a seven-step synthesis of Thromboxane B2 (87) by the stages shown. Scheme 45 ^.. r^^ QMS

.-Ph BzO^. (ii) MeO

(iv)

JOMS

& (iii)

643

OH MeO (Appendix 1 lists the reagents used) In another case (ref.93), triacetyl-D-glucal was used for the preparation of Leukotriene B5 (88). Scheme 46

^^ |iv,v)

IT

(A)

I

(88) (5CiH) (Appendix i lists the reagents used) (iii) Arachidonic acid For the synthesis of hydroperoxides (HPETE compounds)^ the corresponding alcohols (HETE Compounds) and the selective formation of epoxides, arachidonic acid (89) has been primarily involved (ref.94) as depicted. Scheme 47 (89)

(i)Kl3,KHCO3,THF,H2O,0C,(ii)DBU,PhH,(iii)Et3N,MeOH,(iv)MsCl,Et,N, CH2Cl2,-65C;H2O2,Et2O,-110C;LiOH,(v)NaBH4,H2O,pH9,(vi)LiOH,DME,H2O.

644 7.3 Phospholipids Phospholipid chemistry contains several well-authenticated cases of semi-synthesis particularly with respect to recent developments in the area of alkoxylipids and the chemistry of substances such as platelet activating factor (ref.95) a mediator of not only platelet activation but of other biological reactions. The synthesis of the active racemic / l-0-alkyl-2-acetyl-sn-glycero-3-phosphocholine (90) in which the alkyl group is saturated C16 and C18 has been described by a number of workers (ref. 96) from a 1-0-alkyl-sn-glcerol as illustrated. Scheme 48 CH^OR

^.

iHOPOCHCHBr

9^^^

CH^Of^HCH N(M?) CH^OTOCKCI^NlMel

(i)TrCl,Et3N,PhCH,,(ii)NaH,DMF,BnCl,(iii)(EtO),B,H,BO, ,(iv)

Cl2P(0)OCH2CH2Br, (v^MejN.AggCOj, (vi)Pd-C,H2, (vii)AC20,Et3N.

Certain naturally-occurring ether lipids as for example ethanolamine plasmalogens (ref.97), or their choline analogues (ref.98) as well as neutral ether lipids (ref. 99) afford starting materials for semisynthesis by chemical procedures for the first source and chemical/biological methods for the other two. Chemical procedures (ref.98) involve the steps, hydrogenation of the alkenyl group, methanolysis of the acyl group and acetylation. 7.4 Glycerides It is in this class of compounds that the practice of semi-synthesis particularly in the industrial sphere of single or two-step operations is pre-eminent and traditional. The preparation of soap, monoglycerides, or their sulphates, the fatty alcohols, amides and numerous other functional compounds many of which are examples of nucleophilic substitution almost entirely devolve on the reactions of glycerides many of which are examples of nucleophilic substitution. The range of these tranformations particularly with reference to recent developments in surfactant chemistry has been summarised (ref. 100) and discussed with regard to edible applications (ref.101).

645 Methyl linoleate derived by interesterification from trilinolein has been converted to a pheromone 'bouquet' (ref.102)/ ZZZ-3,6,9eicosane and ZZZ-3,6,9-heneicosane, (of value in combatting the soya bean moth)(ref.103) by the following reactions. Scheme 49

(i)LAH,Et20;4-TSCl,Pyr.,(ii)Li(n-Pr)2Cu + LiEtgCu; NH^Cl. 7.5 Phenolic Lipids In the phenolic lipids such as the alkylphenols from Anacardium occidentale^ semi-synthetic transformations have been described which exploit the functionality of the component phenols, cardanol/ cardol and anacardic acid (ref. 104). A number of reactions are shown for

§^r Scheme 50

f^'^V^ (ii)

m ^Y

f^^yR

\ Polymer

rf^, ,v,-rr ^^

6H

^"'

6(ci^a^o}H

(94)

—^Fblynier

(96) J

(91)

{ i ) K o l b e , ( i i ) L A H , E t 2 0 , (iii)CH20,NaOH, (iv)Me2NH,CHpO,H,0,A, ( v ) K O H , c a t . , Et.Ox.,(vi)SCl2,PhH,(vii)EtMgBr,HMPA,CH20,H30*,(viii)H2NOH.HCl,Pyr.

646 cardanol including formation

of an aldoxime (91) useful

solvent extraction of copper

for the

(11), a bis-thiophenol

(92), an

ethoxylate of interest as a biodegradable detergent (93) (ref. 83), a

Mannich base (94) and a diol (95)

representing the first stage

in the polymerisation of cardanol with formaldehyde. Anacardic acid

similarly forms an isomeric aldoxime and a polymer

with formaldehyde.

8.

SEMI-SYNTHESIS IN ANTIBIOTIC, POLYMER AND PEPTIDE CHEMISTRY There are of course many other examples of semi-synthesis in

various classes of compounds such as in the antibiotics, illustrated by tetracycline chemistry (ref,105) where minor modifications have been obtained from the tetracyclic structure.

H.^H^y

In

natural

polymer

chemistry

a vast

number

of

derivatives

and

technologies involving cellulose including the process for rayon the chromatographic esters

material

diethylaminoethylcellulose

such as the triacetate

indicate both

and

numerous

the historical

and

present-day role of this important polymeric carbohydrate. In the synthesis of peptides such as bradykinin* the use of the configurationally correct aminoacid is obligatory and semi-synthesis and total synthesis here become indistinguishable. By contrast

the

unnatural dipeptide sweetening agent, aspartame (ref .106) illustrates the scope for semi-synthesis. *Bradykinin is a nonapeptide Table 6

(ref.3).

Semi-Synthesis in Lipid Chemistry

Natural

Product

Intermediate /f-lonone

Synthesis Type

Vitamin A

Multi-step

Reference 85

647 Phytol If

Farnesol, Geraniol (-)-Pulegone •^-Methyl-D-glucoside Triacetyl-D-glucal Arachidonic acid Monoglyceride Ether Plasmalogens Glycerides Natural and Technical Cashew Phenols

Vitamin E Vitamin K^ Vitamin Kg Prostaglandins Thromboxane Leukotriene Hydroperoxides Alcohol Platelet Activation Factor II

Surfactants Edible Products Polymerisation Products Industrial Chemicals

Two-step II

Multi-step II

II

Three-step II

Multi-step

II

Single-step

II

86 87 88 91 92 93 94 94 96

97 100 101

102

CONCLUSIONS It is appropriate to draw some conclusions from the numerous quite different examples given in this review. The circumstances in which semi-synthesis has arisen in a given area appear to be unique for each class of compound. For the CIO acyclic alcohols the availability of vast quantities of cheap pine oil at a time when essential oil sources were increasingly costly appears to have been a stimulus. For the steroids, progesterone, estrone and cortisone, the absence of cheap total syntheses and the resource of available sapogenins and cheap cholesterol seems to have been a driving force. For the antibiotics, particularly the penicillins the inadequacy of the existing members of the group and a curiosity to find the reason, suggests itself as the reason for the origin of the new compounds. In the case of aspartame the discovery seems to have been accidental like the original detection of saccharin by I.Remsen. Thus the background to these cases reflects applied and possibly economic considerations. By contrast total syntheses, regarded as aspects of semi-synthesis, from natural source intermediates which have been included in this review were undertaken for their scientific challenge and few could have been said to have had an economic motive. Primarily they have advanced organic methodology without a developmental approach as a factor since in most cases the economic

648 value of the product was either not known or was a secondary matter. Technology transfer is not evident to the extent existent in the three cases listed given at the outset. Increasingly over recent years total synthesis as a general endeavour has been directed to ever more complex challenging structures and even the viability of extraction and isolation of the natural material itself is uncertain. In those cases where these are facile it would seem that traditional semi-synthetic operations could play a part when structure/property studies show interesting results in the compound concerned. Although multi-stages and the yield 'arithmetical factor* tend to work against the viability of total synthesis, conceived in the widest sense, it remains the great channel through which organic chemistry advances and thus must always be a valid study. The prospects however for semi-synthesis are also very good for a number of reasons. It has been estimated that more than 90% of the world's natural product resources are still untapped. With the aid of molecular modelling it is possible to investigate the prosthetic or functional group area in which the activity of the system under examination is thought to reside. Functional group transformatios are the province of semi-synthesis. Specific reagents or enzyme systems may thus render these located specific sites vulnerable to appropriate transformations. The proposed mechanism for the mode of action of the enediyne antibiotics calicheamicin and dynemicin (ref.107) indicate the possibilities for other systems. REFERENCES 1

S.Hanessian, J. Franco and B. Laroche, Pure and App. Chem., 1990, 62, 1887; S. Hanessian, * Total Synthesis of Natural Products: The Chiron Approach', Pergamon, Oxford, 1983.

2

C.H. Heathcock, in 'Asymmetric Synthesis', ed. J.D. Morrison, Academic Press, New York, vol. 3, p.Ill, U 9 8 4 .

A. Koskinen,

'Asymmetric Synthesis of Natural Products', Wiley, 1993. 3

I. Fleming, 'Selected Organic Syntheses', Wiley, london, 1973

4

E.J. Ariens, Med. Res. Rev., 1988, 8, 309.

5

R.A.Sheldon, Chem. and Ind., (London), 1992, 903.

6

F. Seidel and A. Stoll, Helv. Chim. Acta, ;J.H.P. Tyman and B.J. Willis, Tetrahedron Letters, 1970, 4507.

7

L. Ruzicka and V. Fornasir, Helv. chim. Acta, 1919, 2, 182.

8

M.F. Carroll, Rep. Progr. Appl. Chem., 1961, 500; Glidden Co., USP 2871271 (further refs.); Van Ameringen-Haebler Inc.,

649

9 10 11 12 13

14 15

USP 2882323; B,D. Sully, Chem. and Ind., 1964, 263. J.H.P. Tyman, J. App. Chem. and Biotechnol,, 1972, 22, 465. R.E. Close, Manuf. chem.. Aerosol News, 1980, 41. G. Buchi, W.D. MacLeod and J. Padilla, J. Am. Chem. Soc, 1964, 86, 4438. G.L. Hodgson, D.F. MacSweeney and T. Money, J. Chem. Soc. Chem. Commun., 1971,766. Y. Amano and C. Heathcock, Can. J. Chem., 1972, 50, 340; M. Sergent, M. Mongrain and P. Deslongchamps, Can. J. Chem.,1972,50, 336. E.J. Corey et.al., J.Am. Chem. Soc, 1964, 86, 478. c.H. Heathcock et. al., J. Am. Chem. Soc, 1966, 88, 4110.

16 j.E. McMurry et. al., J. Am. Chem. Soc, 1968, 90, 6821. 17 w.C. Still, S. Murata, G.Revial, and K. Yushihara, J.Am. Chem. Soc, 1983, 105, 625. 18 E.J. Corey and S. Nozoe, J. Am. Chem. Soc, 1965, 87, 5728. 19 E.J. Corey and H.L. Pearce, J. Am. Chem. Soc, 1979, 101, 5736. 20 N.H. Anderson and D.D. Syrdal, Tetrahedron Letters, 1972, 2455. 21 E. Demole, P. Enggist and C. Borer, Helv. Chim. Acta, 1971, 54, 1845. 22 E.J. Corey and K. Achiwa, Tetrahedron Letters, 1969, 3257. 23 E.J. Corey and A.G. Hortmann, J. Am. Chem. Soc, 1965, 87, 5841. 24 B.M. Trost and S. Mallart, Abstract 426, 18th lUPAC Symposium on Chemistry of Natural Products, Strasbourg, Aug. 1993. 25 E.J. Corey, P. Da Silva Jardine and J.C. Rohloff, J. Am. Chem. Soc, 1988, 110, 3672; E.R. Loft A.S. Kotnis and T.A. Broadbent, Tetrahedron Letters, 1987, 28, 2799; K.C. Niolaou and W.S. Li, J. Chem. Soc. Chem. Commun., 1985, 425. 26 K.B. Sharpless, J. Am. Chem. Soc, 1970, 92, 6999. 27 P. Karrer and C.H. Eugster, Compt. Rend., 1950, 230, 1920 28 O. Isler, H. Lindlar, M. Montavon, R. Ruegg, and P. Zeller, Helv. Chim. Acta, 1956, 39, 249. 29 H.J. Bestmann and 0. Kratzer, Angew. Chem., 1961, 73/ 757. 30 H.T. Clarke, J.R. Johnson and R. Robinson, (Eds.), 'The Chemistry of Penicillin', Princeton Univ. Press, 1949. 31 J.C. Sheehan and K.R. Henery-Logan, J. Am. Chem. Soc, 1959, 84, 2983. 32 F.P. Doyle, G.R. Fosker, J.H.C. Nayler and H. Smith, J. Chem.

650 S o c , 1962, 1440; F.P. Doyle, K. Hardy, J.H.C. Nayler, M.J. Soulal, E.R. Stone and H.R.J. Waddington, J.Chem. S o c , 1962, 1453. 33

M.S. Manhas and A,K. Bose, 'Synthesis of Penicillins, Cephalosporins, and their analogues', M. Dekker, New York, 1969. 34 J.C. Sheehan and K.R. Henery-Logan, J. Am. Chem. S o c , 1962, 87, 2983. 35 P.G. Sairanes (Ed.), 'Topics in Antibiotic Chemistry', Vol. 3, Ellis Horwood, Chichester, 1969. 36 R.M. Morin and M. Gorman, (Eds.), 'Chemistry and Biology of a -lactam Antibiotics, Vol.1, Penicillins and Cephalosporins', Academic Press, New York, 1982 37 R.B. Wooward, K. Heusler, J. Gosteli, P. Naegeli, W. Oppolzer, R. Ramage, S. Ranganathan and H. Vorbruggen, J. Am. Chem. S o c , 1966, 88, 852; R.B. Woodward, Science, 1966, 153,487. 38 E.J. Corey and M.M. Mehrota, Tetrahedron Letters, 1988, 29, 57. 39 P. Karrer,B.Becker, F. Benz, P. Frei, H. Salomon, and K. Schopp, Helv. Chim.Acta, 1935, 1^, 1435. 40 C. Huber, Ann. d. Chem. u Pharm. Bd., 1867, 1£1, S.271. 41 L.F. Fieser and M. Fieser, 'Topics in Organic Chemistry', Reinhold, New York, 1963. 42 43

44

D.H.R, Barton, D. Kumari, P. Welzel, L.J. Danks, and J.F. McGhie, J. Chem. Soc.(C), 1969, 332. R.E. Marker and J. Krueger, J. Am. Chem. S o c , 1940, 62, 3349; R.E. Marker, T. Tsukamoto and D.L. Turner, J. Am. Chem. S o c , 1940, 62, 2525.

45

C.W. Shoppee,'Chemistry of The Steroids', Butterworths, London, 1964. Ref.41, p251.

46

M. Rosenberger,T.P.Fraher and G. Saucy, Helv. Chim. Acta, 1971, 54,. 2857;

R.T.

Blickerstaff,

A.C.

Ghosh,

'Total

Synthesis of Steroids', Vol 30, Academic Press, New York,1974. 47

G. Rosenkanz, 0, Mancera, F. Sondheimer and C. Djerassi, J. Org. Chem., 1956, 21, 520; USP 2335616.

48

E.R.H. Jones, Pure and App. Chem., 1973, 32r 39.

49 50

L.H. Sarett, J. Am. Chem. S o c , 1948, 70, 1454. R.K. Callow,J.W. Cornforth and P.C. Spensley, Chem. and Ind., 1951, 699.

651 51

52 53

54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

73 74 75

A.F.B. Cameron, R^M. Evans, J.C. Hamlet, J.S. Hunt, P.F. Jones and A.G. Long, J. Chem. Soc, 1955, 2807; J.H. Chapman, J. Elks, F.H. Phillips, T. Walker, and L.J. Wyman, ibid., 1956, 4330,4344. D.H. Peterson and H.C. Murray, J. Am. Chem. Soc, 1952, 1±, 1871. G. Rosenkrantz, J. Pataki, and C. Djerassi, J. Am. Chem. S o c , 1951, rz, 4055; T.H. Kritchevsky, D.L. Garmaise and T.F. Gallager, J. Am. Chem. Soc, 19*52, 74, 483. D.H.R. Barton, G.M. Beaton, G.M. Geller, and M.M. Pechet, J. Am. Chem. S o c , 1961, 77, 2400. E.J. Corey and W.R. Hertler, J. Am. Chem. Soc, 1958, 8£, 2903; ibid., 1959, 81, 5209. R.L. Autrey and P.W. Scullard, J. Am. Chem. Soc, 1968, 90, 4917. B. Witkop, Annalen, 1943, 554, 83. G. Stork and R.N. Guthikonda, J. Am. Chem. Soc, 1972, 9£, 5109. M. Kuchne, Lloydia, 1964, 27, 435. R.B. Woodward, F.E. Bader, H. Bickel, A.J. Frey and R.W. Kierstead, Tetrahedron, 1958, 2^, 1. M. Gates and G. Tschudi, J. Am. Chem. Soc, 1956, 78, 1380. R.B. Barber and H. Rapoport, J. Med. Chem., 1976, 19, 1175. K.W. Bentley, 'The Chemistry of the Morphine Alkaloids', Oxford University Press, London, 1954, p., 59. J.M. Gulland and R. Robinson, Mem. Proc Manchester Lit. Phil. Soc, 1925, 69, 79. W.F. Michne, 'Analgesics*, Kirk-Othmer Encyclopedia of Chemical Technology, 3rd. Edn., 1978, Vol.2, p,574. K.W. Bentley, 'The Morphine Alkaloids', in 'Chemistry of the Alkaloids', Ed. S.W. Pelletier, Reinhold, New York, 1970, 140. K.W. Bentley and D.G. Hardy, J. Am. Chem. Soc, 1967, 89^, 3267, K.W. Bentley, D.G. Hardy and B. Meek, ibid., 3273. K.W. Bentley and D.G.Hardy, ibid., 3281. J. Pohl, Z. Exp. Path. Ther., 1915, 17, 370. J. von Braun, M. Kuhn and S. Siddiqui, Chem. Ber., 1926, 59, 1081. R.G. Ault, D.K. Baird, H.C.Carrington,W.N. Haworth, R. Herbert, E.L. Hirst, E,G.V. Percival, F. Smith and M. Stacey, J.Chem. Soc, 1933^ 1419. T.Reichstein, A. Grussner and R. Oppenauer, Helv. Chim. Acta, 1922, li6, 1019. T. Sonoyama et. al., Appl. Environ. Microbiol., 1982, « , 1064; A.J. Pratt, Chem. Brit., 1989, 25^, 282. R.U. Lemieux and G. Huber, J. Am. Chem. Soc, 1953, 75, 4118.

652 76 77

j.L. Meers, Biotechnol. Letters, 1981, 3, 131. M.P. Edwards, S.V. Ley, S.G. Lister, B.D. Palmer and D.J. Williams, J. Org. Chem., 1984, 49, 3503.

78 79 80 81 82

J. Mann and A. Wilson, J.Chem. Soc. Chem. Commun., 1985, 7 37. B.J. Fitzsimmons and B. Fraser-Reid, Tetrahedron, 1984, 4jO, 1279. H. Paulsen and A. Todt, Angew. chem. Int. Ed. Engl., 1966,5,495. G.W.J. Fleet, Chem. Brit., 1989, 25, 287. G.W.J. Fleet, N.G. Ramsden and D.R. Witty, Tetrahedron, 1989, 45, 319.

83

E.N. Vulson, 'Enzymatic Synthesis of Surfactants', in 'Surfactants in Lipid Chemistry', ed., J.H.P. Tyman, Royal Society of chemistry,

Cambridge, 1992. 84 S.F. Dyke, 'The Chemistry of the Vitamins', Interscience, London, 1965, p.271. 85

o.Isler and P. Schudel, Advan. Org. Chem., Eds. R.A. Raphael, E.C. Taylor, Interscience, 1963, 4, 1.

86 87 88 89 90 91 92 93 94

P.Karrer and V. Demole, Schweiz. Med. Wochschr., 1938, 68, 964. R.Hirschman, R. Miller and N.L. Wendler, J.Am. Chem. S o c , 1954, 76, 4592. O.Isler, R. Ruegg, L.H. Chopard-Dit-Jean, Chimia (Switz.) 1958, 12, 69. S.M. Roberts and F. Scheinmann, Eds., 'New Synthetic Routes to Prostaglandins and Thromboxanes', Academic Press, Lomdon, 1982. E.J. Corey and Xue-min Cheng, 'The Logic of Chemical Synthesis', Wiley, New york, 1989, p259 E.J. Corey and H.E. Ensley, J. Am. Chem. S o c , 1975, 97, 6908. E.J. Corey, M. Shibasaki and J. Knolle, Tetrahedron Letters, 1977,1625. E.J. Corey, S.G. Pyne and W.-g. Su, Tetrahedron Leters, 1983, 24, 4883. ref.90, p.339, p.343.

95

J. Beneveniste, P.M. Henson and C.G. Cochrane, J. Exp. Med, 1972, 136, 1356.

96

H, Eibl, Angew. Chem. Int. Ed. Engl., 1984, 23, 257.

97 J. Beneveniste, M. Tence, P. Varenne, J. Bidault, C. Boullet, and J. Polonsky, Comptes Rend., Ser.D, 1979, 289, 1037. 98

C.A. Demopoulos, R,N, Pinckard and

D.J. Hanahan, J. Biol. Chem.,

1979, 254, 9355. 99

T. Muramatsu, N. Totani and H.K. Mangold, Chem. Phys. Lipids, 1981, , 29, 121. 100 K. Coupland, ref.83,p.i.

653 101

102 103 104

105 106 107

R,J. Hamilton and A. Bhati, Eds.^ 'Fats and Oils: Chemistry and Technology'/ Applied Science publishers^ London^ 1980; P. Tooley, 'Fats. Oils and Waxes', Murray, London, iy/l. Hua-min Yu and H.K, Mangold, Chem. and Ind. (London), 1988, 787. R.R. Heath, J.H. Tomlinson, N.C. Leppla, J.R. Mclaughlin, B. Dueben, E. Dundulis, R.H. Guy, J. Chem. Ecol., 1983, 9, 645. J.H.P. Tyman, 'Non-isoprenoid Phenolic Lipids', in 'Studies in Natural Products Chemistry', Ed. Atta-ur-Rahman, Elsevier, Amsterdam, Vol.9, p.313. J.H. Boothe and J.J. Hlavka, Kirk Othmer Enc. Chem Tech., Vol.3, 3rd. Edn., p.66. J. Am. Chem. Soc, 1963, 81, 3943. R.H. Mazur, J.M. Schlatter and A.H. Goldkamp, J. Med. Chem., 1969, 91, 2684. K.C. Nicolaou, W.-M. Dai, S.-C. Tsay, V.A. Estevez and W. Wrasidlo, Science, 1992, 256, 1172.

Abreviations AIBN a-Azoiso-butyronitrile Bn Benzyl BzCl Benzoyl Chloride DBU 1,5-diazabicyclo[5.4.0]undecene-5 DCC 1,3-Dicyclohexylcarbodiimide DIBAL Di-iso-butylaluminium hydride DME 1,2-Dimethoxyethane DMF N, N-Dimethy 1 f ormamide DNP 2,4~Dinitrophenylhydrazone HMPA Hexamethylphosphorictriamide LAH Lithiiim aluminium hydride LDA Lithium diethylamide MCPBA m-Chloroperbenzoic Acid NBS N-Bromosuccinimide NCS N-Chlorosuccinimide PBA Perbenzoic Acid THF Tetrahydrofuran 4-TSA 4-Toluenesulphonic Acid TsCl 4-Toluenesulphonyl Chloride PG Protective Group

654 APPENDIX Scheme 6: (i)HOCH2CH20H,H30*(ii)Oxidn . (iii)Ph3P=CHCH2CH2 0Thp;H3 0* , ( i v ) C C l 4 ( n Oct)3P;H30* , (v)TsOH,AcO(Me)C=CH2, (vi)BF3,H20,CH2Cl2, ( v i i ) (CH20H)2,H* ; NaI,Me2C0 ( v i i i ) PPh3, ( i x ) CH3SOCH2" /Me2C0; H3O*. Scheme 7: (i)HgOAc,H20/HF;NaBH^,NaOH,(ii)CH2=CHCOEt,KOH,(iii)HCl,THF,(iv)PdC,Hp (v)HC02Et,NaH,Et2 0 , ( v i ) B r 2 ,NaOH, ( v i i ) L i C l ,Li2C03 ,DMF, 125C, ( v i i i ) (PhP)3RhCl, A , ( i x ) L A H , E t 2 0 , (X)TS0H,C6H6. Scheme 2 0 : (i)Me2CO/BuOCOCl,(ii)Me02CN=NC02Me,(iii)Pb(OAc)^;NaOAc,MeOH,(iv) MeS02Cl/^Pr2NEt;NaN3, (v)Al/Hg/MeOH, (vi)^'BU3Al, ( v i i ) (OHC)2C=CHC02CH2C CI3/ ( v i i i ) C F 3 C 0 2 H , ( i x ) A /DCC,(x)B2H6;Ac20,Pyr.,(xi)Zn/AcOH/H20. * Scneme 2 1 : (i)COCl2,PhCH3, A;BnNH2/CH2Cl2/(ii)Ph3P,DME,Et2O,H2O;6MHCl,100C,(iii) SClC^H3(N02)2/CH2Cl2,(iv)(CeCl3,^:r'''^x^-v.cl ,n-BuLi)THF,--78C,(v)DIBAL-H ,PhCH3,-78C;(PhS)2,(n-Bu)3P,CHpClp,(vi}(C6H„)3SnH,AlBN,PhH, A,(vii) NaCN,EtOH/H20;Pd-C,H2,EtOAc, ( v i i i ) 2MNaOH, A;HBr,H2U. Scheme 2 6 : (i)Oxidn.;2,3-desatn.(ii)A,600C,min.oil,(iiijPd-C^Hg,(iv)asin Scheme24; (v)H2N0H.HCl,Pyr., (vijAcNHC^H^SOgCl^Pyr. , 2 0 C , (vii)H30*. Scheme 4 5 : (i)PhCH(OMe)2,TsOH,DMF,A;BnCl,Pyr.,CH2Cl2, ( i i ) P t 0 2 , H2/cat.H2S04,MeOH; BnCl,Pyr. ,CH2Cl2,-78C;H30*;MsCl,Pyr. , - l O C , ( i i i ) N a I , Z n - C u , D M F , A/'EtjN, MeOH/H20,(iv)(MeO)2(Me)CNMe2,diglyme,25C,A,(vjig/THF/HgO^OC;(nBu)3SnH, NaBH^,PhH,(vi)(Ref.E.J.Corey,M.Shibasaki,J.Knolle,T.Sugahara, Tetrahedron L e t t e r s , 1977,785 ) . Scheme 4 6 : (i)MCPBA,BF3.Et20,CH2Cl2, -IOC, (ii)Zn~Hg,HCl,Et2O,0C, (iii)DBU,THF, ( i v ) LiOH,DME/H20;Pd-C,H2 ,EtOAc, (v)PDC,MgSO^ ,mol . s i e v e s CH2CI2 , ( v i ) l 2 ,Ph3P, i m i d , Et20/MeCN; BrMgC=CCH2OMgBr, CuBr, THF65C, ( v i i ) LAH, E t j 0 , ( v i i i ) ( S , S ) ( - ) - D M T , T i ( O i ~ P r ) ^ ,t-BuOOH,CH2Cl2/ -20C, ( i x j H , , Pd-CaC03, Et3N ,THF;Cr03. 2Pyr.,CH2Cl2,(x)Ph3P=CHCH2CH20C(Me)20Me,THF,PhCH3,-78C;AcOH,H20/MeCN, (xi)l2,Ph3P,(i-Pr)2NH,imid,Et2O/MeCN;Ph3P,MeCN,60C,(xii)n-BuLi,THF,7 8 C ; H M P A ; A , - 7 8 C , ( x i i i ) i - P r O K , i - P r O H ,0C;LiOH,MeOH/H2O.

CO2H

655 SUBJECT INDEX Araucaria cookii 36 Anchiisa officinalis 126 Absolute configuration 212A^^ Absolute stereochemistry 51 Acer nicoense 358,367-368 Acemikoense 375 Aceraceae 358,366-367 AcerogeninA 367,375 biosynthesis of 375 Acerogenin A-D 367 Acerogenine-E 368 Acerosidel 367 AcerosidelV 367 AcerosideV 367 AcerosideVII 358 AcerosideXI 368 Acetobacter suboxydans 636-637 Acetone 562 *^ONNfR spectrum 562 Acetonylbutanolides 273 Actinoplanes coloradoensis 283 Acylated flavonol-glycosides 142 y4t/ocia species 34 Adociaquinone A-B 33 Agrobacterium rhizogenes 395,421 Albati 198 Albazoin 19 Albicanol 14 Albicanyl acetate 14 Alcohol dehydrogenase 479 Alexandrium tamarensis 4 Alkaloids 79-80,8,89,404-406,409,411-412, 441,631 P-carboline 89 isoquinoline 89 production of 404,409,411-412 pyridine 89 purine 89 semi-synthesis of 631 Alnusfirma 359,364 Alnus hirsuta 360-361 Alnusjaponica 360,368 Alnus rubra 359 Alnus serrulatoides 359-360 Alnus species 358-359,368,375 Alnusdiol 368 Alnuson 368 Alnusonol 368 Alnusoxide 368,371 >4//7/mti species 358

Alpinia katsumadai 362,375 Alpinia qfficinarum 362-363,375 Alpinia oxyphylla 362,375,379 Alpinia sieboldiana 362 Amathamide A-G 83 Amathamides 82,84,88,93,95.97 antifeeding activity of 97 Amathia altemata 85 Amathia convoluta 75 Amathia genviS 82 Amathia wilsoni 92,95,97,101 Amphotericin B 245 antifungal activity of 245 Anacardium occidentale 645 Analgesic 633 Androgens 626-627 semi-synthesis of 626-627 A ndrographis paniculata 472 Angasiol acetate 6 Anigozanthos 372 Annona muricata 277 Annonaceous acetogenins 251-286 Annonaceae 251 Annonacin 266,272,279 AnnonacinA 267 AnnoninI 252 Annonacin-lO^one 272 Antibacterial activity 19,24, Antibiotic activity 153,283-284,286,378, ofmacrolide 283 Anticholinergic agents 395 Antifeedant 153,234,237 Antifouling 93,98 Antifungal 16,196,235,239-240,378 agents 239 activity of 378 macrolide 16 sesquiterpene dialdehydes 233 Antihepatotoxic activity 377 Antiinflammatory activity 137,376 Antiinflammatory activity 376 Antiinflammatory drug 130 Antileukemic activity 348 Antimicrobial activity 24,153,233,235,238, 241,285 Antiphlogistic 451 Antipredation 93 Antirachitic factors 623 Antitumor 17 Antitussive 633 Antiviral activities 133-135,145-146 Herpes simplex virus 146

656 HIV-2 146 offlavonoids 145 of quinovic acid derivatives 134 of oleanolic acid derivatives 135 Aplinia species 358 Aplydilactone 6 Aplykurodin A 5 Aplysia dactylomela 7,8 Aplysia faciata 6 Aplysia kurodai 4,6 Aplysia Juliana 6 Aplysia punctata 9 Aplysiapyranoid A-D 5 Aplysilla glacialis 14,15 Aplysilla sulphurea 11 12-ep;-Aplysillin 12 Aplysina fistularis 103 Apoenzyme 487 p-Apolignans 339 synthesis of 339 Araucaria cunninghamii 36 A rdisiajaponica 117,127 Ariensin 318 synthesis of 318 Arteminisin, 591 '^O-NMR spectrum of 591 Artemisia herba-alba 475 Arthritis 137 Aryl ethers 586 '^O-NMR spectrum of 586 Aryldihydronaphthalenes 337 2-AryI-3-methyl-4-benzyltetrahydrofurans 331 Arylnaphthalenes 332 Aryltetralins 341,343 Asadaninl&Il 369 Ascorbic acid 636 synthesis of 636 Asimicin 262 triacetate 262 Asparagus sipQcits 130 Asparagus cochinchinensis 116-117,130.132 Asymmetric, catalysts 323,479,481 hydrogenation 323 synthesis of 479 A tropa germs 395 Auxins 424 efiTect on root culture growth 424 Avicularia 193 Baconipyrones A-D 25 Bacterial growth inhibitors 101

Barnacle settlement inhibitors 103 Basidiomycetes 153 Benzyl alcohols 92 BetaenoneB 475 Betaine 92 Betula species 358.361 Betula pendula 361,366 Betula platyphylla 361 Betulaceae 358,359,364,366 Bijlustra perfragilis 81,89,91 Bile alcohols 207,212 CD studies 212 Bioactivities 253 Bioconversion 470 using Morus alba cell 470 Biofilms 100 Biological activity 133 Biological Diels-Alder reaction 451 Biological raw materials 601 Biologically active compounds 153 Biopolymers 481 Biosynthesis 207,465 ofcholicacid 207 ofkuwanon 465 Biosynthetic route 314 of tropane alkaloids 314 5-BisaboIene 609 Bisaryltetrahydrofiirans 325 Bisbenzyltetrahydrofurans 330 Bis-THF acetogenins 270 Bis-THF structures 258 (+)-Brevianamides A and B 475 Brominated compounds 4,9,81,84 amides 84 diphenylether 9 phenols 81 phenylether 9 Brosimones 456 Brosimopsis oblongifolia 456 Broxil 616 Bryostatins 75 Bryozoans 73,74 taxonomy of 74 Bugula dentata 88 Bugula neritina 75,77,89,92 Bugula trubinata 102 Bugula turrita 80 Bugularia dissimilis 95,101 Bullacin derivatives 276 Bullatacin 252,266,271-272 BuUatacinone 266,278

657 Bupleurum jruticosum 117,127 Burseracea 370 Bursera microphylla 330 Burseran 330 synthesis of 330 Bursera ariensis 317 /-Butyl anthraquinone 561 ^'ONMR spectrum of 561 p-CarboHne alkaloids 90 Cabreuvaoil 609 CadlinolideA 15 Calendula ctrvensis 126 Cancer 137 Candida albicans 233 Candida utilis 235 Canellaceae 234 Caibohydrate 542,636-640 *'0-NMR spectrum of 542 semi-synthesis 636-640 Cardiotonic constituent 33 P-Carotene 611,612 Carotenoids 73 Carpinus species 369 Carpinus cordata 369 (+)-Carvone 608 Cassumins A, B, and C 365 Cassumunarin A, B and C 366 Cassumunin A 376 Casuareinondiol 371 Casuarianjunghuhniana 371 Casuarianaceae 371 Catechins 144 CD Cotton effects 44 CD excition chirality method 35,36,51,53 CD spectra 37,44,61,461 of chromomycin 37 a-Cedrene 609 Ceilaria species 89,92 Cellaria pilosa 95,101 Cembrane 22,28 Cembranoids 28 (3/?,75>Centrolobin 366 (37?,7/?>Centrolobin 367 Centrolobium species 358,366 Centrolobium paraense 367 Centrolobium robustum 367 Centrolobium sclerophyllum 367 Centrolobium tomentosum 367 (->Centrolobol 358,361 (-H^>Centrolobol 361

Centrolobol 367 Cephalosporins 614,617-619 semi-synthesis of 617-619 Cephalosporium acremonium 619 Cerebrotendinous xanthomatosis 207 Chalcomoracin 465 biosynthesis of 465 Charonia lampas 88 Chartella papyracea 85,86,89,92 ChartellamideA&B 86 Chartelline A,B & C 86 Chelynotus semperi 23 Chenodeoxycholic acid 207 biosynthesis of 207 Chiral and prochiral substrates 481 Chiral, bases 323 Chiral catalysts 479 Chiron approach 601 Sharpless oxidation 211 Cholestanepentols 214 CD data of 214 Cholestane tetrol 213-215 synthesis of 213-215 Cholesterol 207 Cholicacid 219,221 biosynthesis of 219,221 Chondria armata 21 Chromones 574 *^ONMR spectrum of 574 Chromodoris species 12,14 Chromodoris cavae 13 Chromodorisfunarea 9,10 Chromodoris lachii 13 Chromodoris macfarlandi 11 Chromodoris norrisi 12 Chromodorolide A 13 Chromodorolide B 13 Chromomycin 37 CD spectra of 37 Cioclapta species 16 Circular dichroic power 48 .y-Citronellal 610 (->Citronellene 605 (+>Citronellol 605 Clathridine 18 ClathridineB 18 "C NMR spectra 116,416,436,438 of phenyl glycosides 438 of xanthones 436 Collinusin 340 synthesis of 340

658 Collisella limatula 26 Coloradocin 284,286,304-306 synthesis of 305-306 COMPARE 495 Conopeum seuratum 79 Copaene 608 Corticosteroids 627-631 semi-synthesis of 627-631 Costaticella hastata 90 Coumarins 574-575 ''O-NMR spectrum of 574-575 ChbhceUina cribraha 79,89-90 Cubic section model,488,503,506,513-514,540 for substrate binding domain of HLADH 503 Col3 514 Jones 488 of HLADH 503 Col3andCo065 540 Culture medium 432 effect on prod, of polyphenols 432 Cupalaurenol 7 Cupalaurenol acetate 7 Cuparene-related sesquiterpenes 7 Curcuma species 358 Curcuma longa 363,373,377,379 Curcuma tinctoha 357 Curcuma xanthorrhiza 364 Curcumin 357,363,374,379 biosynthesis of 374 Cycloartenol 622 formation of 622 Cyclocurcumin 363 Cyclolaurene 7 Cyclolaurenol 7 acetate of 7 Cytochrome P450 475 inhibitors of 475 Cytotoxicity 251 Dactylomelol 8 Dandrolasin 13 Dapetes 198 Darzens reaction 612 Datura genus 395 Daurinol 335 synthesis of 335 12-epi' 12-Deacetoxyaplysillin 12 Dehydrogenation 59 by DDQ 59 Dendrodoris grandiJJora 28 Dendrodoris limbaia 28

Denticulatin A and B 24 antimicrobial activity of 24 Deoxoscalarin 10 5-Deoxymyricanon 371 18-Deoxynargenicin 298,301 synthesis of 298,301 DEFT 116 2a,7a-Diacetoxy-6p-isovaleroxylabda-8,13dien-15-ol 27 Diaperoecia califomica 90,92 Diarylbutanes 349 Diarylcyclobutanes 348 Diarylether 388 Diarylheptanoids 372,373,375,379 activity of 375 biosynthesis of 373 dienone 372 Diastereoselective alkylation 324 Diastereotopic groups 481 Dibenzocyclooctadienes 346 Dibenzylbutanes 315 Dibenzylbutyrolactones 320 2,6-Dibromophenol 81 Dictyols 97 Didemnum chartcium 22 Diels-Alder adducts 451 Diemenensin A 24 antibacterial activity of 24 Diemenensin B 24 1,2-Dihydro-l-arylnaphthalene lignans 338 Dihydrocostunolide 609 Dihydrocubebin 318 synthesis of 318 Dihydrocurcumin 363 Dihydrohalichondramide 16 Dihydrosiphonarin A and B 25 Diketoalcohol 364 Diketone 26 Dinoflagellates 20 Dinophysis fortii 20 Dinophysistoxin-1 19 Dinophysistoxin-2 20 Dinophysistoxin-3 19 Dioscoreaceae 130 Diphenols 368 Dirman 28 Disaccharides 636 semi-synthesis of 636 Diterpenes 8,11-15,27,148 Diuretic 451 Domoicacid 21

659 DomoicacidD 21 Dorimidazole A 18 Doris verrucosa 15 Drimanes 154 DSP values 118 Duasmodactyla kurilensis 153 Duboisia gtnyis 395 Dysidea geraxs 10,11 Effect of torsion angle rotation 556 Effects of Nif 433 on secondary metabolism 433 Eimeria tenella 376 7t-EIectron SCF-CI-DV MO method 39-40,45,47 Electron-demand 553 Ellagitannine 360 ENZYME program 493-494,500 Enzyme system 470 of Morus alba 470 Epitaondiol 8 Epomuricenin 277 Epoxy sterol 22 Epoxydilactone 608 Epoxylactone 10 Eriobotryajaponica 115,118-119 Erwinia herbicola 637 Escherichia coli 378 Estrogens 625-526 semi-synthesis of 625-626 Eucannabinolide 608 Eudistoma olivaceum 100 Euphorbiaceae 343,439 Expectorant 451 FAB 116 Farnesane sesquiterpenes 155 Fatty acid ester 201 Fatty acids 221,263 Flavanones 143-144,580 anti HIV-1 activity of 144 *'0-NMR of 580 Flavones 561,580 "O-NMR 561,580 Flavonoids 341 Flavonol 117,139-141 anti HIV-1 activity of 141 glycosides of 117,140 Flustrafoliacea 80,85,87,102,103 Flustra papyracea 86 Flustramines 87 Forskolin 610

Framework molecular models 490 Fuloplumierin 19 antibacterial activity of 19 Fungicidal activity 243 Furanosesquiterpenes 197 Furans 324,585 *'0-NMR spectrum of 585 Furodtysin 10,15 Furodysinin 10,15 lactone 10 Furoguaiacidin diethyl ether 325 synthesis of 325 Furostane glycosides 136 Furostanol glycosides 116,135 tetraglycosides 135 Galeon 371 Garcinia mangostane L 36 Gargugamblins 370 Garw^a species 375 Garuga gamblei 370, Garuga pinnata 3 70,3 75-3 76 Garugamblins 378 Garugamblin-1 386,389 synthesis of 386 Garugamblin-2 386 Garuganins 370,378 Garuganinlll 386 synthesis of 386 Gastritis 137 Gentianaceae 421,434 Geraniaceae 421 Geraniin 433 biosynthesis of 433 Geranium thunbergii 421,428,433 root culture 428 Gestogens 623 semi-synthesis of 623 Gibberellic acid 425 growth effect on root culture 425 Gibbonsia elegans 94 Giganenin 277 Gingerenones A, B and C 378 Glaciolide 14,15 GLC analysis 116 Glutinopallane 154 Glutinosi 199 Glycosides 115 purification of 115 Gmelina arborea 332 Gonyaulax tamaremis 4

660 Growth inhibition 142 Growth 413,434 of Hyoscyamus albus 413 of hairy roots 434 Guaiacum officinale 319,324 Guaianes 154 Guettarda platypada 116,125 Gymnodinium breve 20 Gymnolaemata 106 Gypsogenin 121 'H-NMR spectra 116,416,437 Haemodoraceae 372 Hairy root cultures 426 Halenaquinol dimethyl ether 40 Halenaquinone 33 (+)-Halenequinol 48-50,55-61 total synthesis of 48-50,5-61 (+)-Halenequinone 48-50,55-61 total synthesis of 48-50,55-61 Halichondria species 17 Halogen-containing compounds 7,9,81 chamigrenes 7 monoterpenes 9 sesquiterpenes 7 Hannokinin 360 Hannokinol 360 Haplophyllum dauricum 335 Heliocidaris erythrogramma 95 Heliopsis buphthalmoides 319 Heliopsis helianthoides 334 Helioxanthin 334 synthesis of 334 Heterocyclic bicyclic substrate 497 Heterocyclic compound 614-622 semi-synthesis of 614-622 Heteropora alaskensis 90,92 Heterotropa takaoi 348 Hexabranchus sanguineus 16 Hexahydrocurcumin 363,368 (5>-Hexahydrocurcumin 365 Hinckdentine A 88 Hincksinoflustra denticulata 85,88 Hippodiplosia insculpta 90,92 HLADH 479,481 HLADH-catalyzed reactions 483 Holostane 118 Holothuria tubulosa 134 Homoditerpene 22 Hyalbidone 400 Hydride transfer 486-487

Hydrogen bonding 562 effects on *'0-NMR 562 Hydrolase 480 Hydroperoxide 9 Hydroxygaleon 371 7p-Hydroxyhyoscyamine 397-399 6p-Hydroxyhyoscyamine 398-399 p-Hydroxyketones 365 Hyoscyamine 395,398-400,417 anticholinergic agent 395 Hyoscyamus ^^nws 395 Hyoscyamus albus 398 Hypoglycemic effects 139 Hypopyhyllanthin 344 synthesis of 344 Hypotensive activities 451 Ichorati 200 Ichthyotoxicity 24 Indole alkaloids 85 Influence of phytohormones 408,410 P-lonone 612 Isoannonacin-10-one 270 9-Isocyanopupukeanane 16 Isodomoic acid E3 21 E4 21 Isoflavones 19 Isogingerenone B 378 Isolactarane 154 Isomenthol 605 Isomerase 480 Isomyricanon 372 Isonaamidine 17 Isonaamine A 17,18 Isonitrile 14 Isoprenoid biosynthesis 471 Isoprenylphenols 451,474 biosynthesis of 474 Isoquinolines 91 Isothiocyanate 14 6p-Isovaleroxylabda-8,13-dien-7a, 15-diol 27 Isozymes 481 Janua bioticus 100 Jeunicin 22 Jones map 539 of amino acid residues 539 Jones model 507 Justicia procumbens 333 Justicia prostata 335

661 KabiramideC 16,17 antifungal activity of 16 Kadsura species 346 Karpluspople expression 552 Ketalized Diels-Alder type adducts 458 3-Ketoadociaquinone A 33 Ketodeoxoscalarin 10 3-Ketoepitaondiol 8 Ketones 567 *'0-NMR spectrum of 567 Ketotriol asadanin 369 Kuanoniamines A-D 23 KuwanonG 451,464-465 biosynthesis of 465 optical rotation of 464 KuwanonsJ 455 Labdane-type diterpenes 6 Lachnanthocarpone 372 Lachnates 372 Lactams 573 *'0-NMR 573 Lactarane 154 Lactarius ^erms 153,156,159 sesquiterpenes from 159 biogenesis of sesquiterpenesfrom156 Lactanus camphoratus 156 Lactarius chrysorrheus 196 Lactanus circellatus 197 Lactarius controversus 198 Lactariusdeceptivus 198 LactariIts deliciosus 198 Lactariusdeterrimus 198 LactariusJJavidulus 198,201 Lactariusfuliginosus 153,201 Lactarius fulvissimus 200 Lactarius glaucescens 198 Lactarius glutinopallens 199 Lactariusglyciosmus 199 Lactariushelvus 199 Lactarius indigo 199 Lactarius lignyotus 200 Lactarius, metabolites 154 Lactariusmitissimus 200 Lactariusnecator 199 Lactarius pergamenus 198 Lactanuspicinus 196,201 Lactanuspiperatus 197,198 Lactanus quietus 196 Lactariusrufus 199 Lactarius scrobiculatus 153,196,197

Lactarius subveilereus 198 Lactarius thejogalus 200 Lactarius torminosus 196,197 Lactarius vellereus 153,196,197 Lactones 270,271,275,569 Y-Lactone 270 "ONMR 569 saturated 275 unsaturated 271 Lamellariidae 21,22 Lamellarins A-D 22 Larixdecidua 332 Larrea divaricata 315 Larvotoxic 102 Latrunculin A 13 LatruncuUn magniflca 14 Lauraceae 240,327 Laurencia species 4,7 Laurinterol 7 Leguminosae 358,366 Leptogorgia virgulata 99 Leucetta chagosensis 17 Leukotrienes 642 semi-synthesis of 642 Ligase 480 Lignans 311,313,441 Liliaceae 130 Limatulone 26,27 Linalyl acetate 604 Lipid 640 semi-synthesis of 640 5-Lipoxygenase inhibitory activity 335 Lissodendoryx isodictyalis 168 Littorine 397 Longifolene 608 LuJfjfarieUa variabilis 10 Luffariellin A-D 10 Luteorosin 12 Lyase 480 Lyngbya gracilis 4 Macfarlandin A-E 11,12 Macrocyclic diarylethers glycosides 367 Macrocyclic diarylheptanoids 367,387 synthesis of 387 Macrocyclic lactones 75 Macrofoulers 99 MACROMODEL 494,500,529,540 Maesa lanceolata 240 Magnolia salicifolia 348 Malabaricanediol 611

662 Maneonene 7 Maurapyrone A-D 25 Maurenone 25,26 Medicinal plants 421 Membranipora membranacea 93 Membranipora perjragilis 91 (-VMenthol 605 Mercurialis annua 117 Mesitoylated model compounds 268 Metabolite 4 5-Methoxycurcumin 364 (3/?,75^-de-0-Methylcentrolobin 366 Methylthioadenosine 15 Microbial reductases 491 Microcionin-2 15 Minimization method 500 BDNR 500 MoUusks 3 Monooxygenase 491 Monoterpene 9,79,603 alcohols 79 synthesis of 603 Moraceae 456 Moraceous plants 451 Morphine 633,636 semi-synthesis 633,636 Morusalba 451,465,471 cell cultures 471 Mosher ester method 264,271,276 Mulberroflirans 459 Mulberry, Diels-Alder type adducts 451 Mulberry tree 519 Multiple hydrogen bonding 564 "ONMR 564 Muricatacin 252,277-279 Murisolin 267 Mushrooms 153 Muzigadial 235 Mycotoxins 475 Mydriatics 395 Myrcene 604 Myriapora truncata 77 Myrica species 375 Afyricagale 371 Myrica nagi 371 Afyrica rubra 371 Myricaceae 371 Myricanon 372 Myrsinaceae 240 Mvtilus edulis 21

Naamidine A 17 Naamine A 17 Naphthalene-diene derivatives 41,45,47,48 CD, spectra of 41,45 Nargenicins 283,286,288,291 biosynthesis 286 chemistry of 286 macrolides 283 synthesis 291 Natural products 566 *^0-NMR spectrum of 566 Navanax inermis 28 Navonones 28 NDGA 316 synthesis 316 Neotignan 311 Neomenthol 605 Neurotoxins 3 Nicotinamide coenzyme 481 Niranthin 318 synthesis of 18 NMR,data 128,131 of oligosaccharides 128 of steroid epimers 131 Nocardioides 285 Nodusmicin activity 290 against Staphylococcus aureus 290 NOE spectrum of 129 oligosaccharides 129 Norcardia argentisis 283 Norditerpenes 11 Norhyoscyamine 398 Norisoflavone 19 Norrisolide 12,13 Notodoris citrina 17 Notodoris gardineri 18 Nudibranchs 105 '^O-NMR spectroscopy 549,557 of aryl derivatives 557 O-methylftirodysinin 9 Okadaicacid 19 Oleanolic acid 135,138 antiviral activity of 135 against carrageenan edema 138 Oleanolic glycosides 126-127 Olentes 200 Oligocerus hemorrhages 14 Onchidal 27 Onchidella benneyi 27 Opisthobranchia 3

663 Oregonin 360 Orthoscuticella ventricosa 90,95,101 Ostrya species 369 Ostryajaponica 369 Oxidation 207

of cholesterol 207 Oxido-reductase 480,481 13-OxoinyricanoI 371 Pachiclavularia violacea 610 Papaver somniferum 633 Patinopectin yessoensis 20 Pauly's reagent 396 Pectenotoxin-1 and2 19,20 Pfectenotoxin-3 20 Penicillins 614-617 semi-sjmthesis of 614-617 PemcillJum species 475 PenicilUum brevicompactum 475 PeniciUium chrysogenum 616 Penicilli urn notatum 615 Pentamethyl ether 459 Peptide 646 semi-synthesis of 646 Phenolic components 455 of Morus alba 455 Phenolic lipids 645 Phenolics in adventitious root cultures 444 in hairy root cultures 444 Phenols 586 *'0-NMR 586 Phenyl glucosides 437 9-Phenylphenalenones 372 Phidolopora pacifwa 92 Phospholipids 644 Phototoxic 378 Phthalicacid 441 Phyllanthus niruri 317,343,421,441,443 Phyllidia bourguini 16 Phyllidia varicosa 15 Phyllidiidae 15 Phytotoxins 475 Picrotoxinin 608 (-)-ci5-Pinane 605 Pinus sylvestris 604 Piper guineense 317 Piper sumatranum 348 Pisum sativum 402 Planaxis sulcatus 22,28 Planaxool 22,23 Platelet-activating factor 326

Platyphyllone 361 Plinthogali 200 Plocamium coccineum 9 Podophyllum emodi 341 Podophyllum pellatum 341 Polyaromatic alkaloids 23 Polybrominated biphenylethers 10 Pblyether toxin 20 Polygala polygama 341 Pblygodial 235,244 antifungal activity of 244 Pblygonaceae 237,421,439 Polygonum hydropiper TM Polymer 646 semi-synthesis of 646 Polyphenols 421,427-429 Polypropionates 25,26 PolyrhaphinA 12 Ponndorf-Meerwein-Verley method 605 Porninsal 155 Porninsol 155 Porson 371 Preclathridine A 18 Prorocentrum lima 20 Prosobranchia 3,19 Prostaglandins 642 semi-synthesis of 642 Pterogorgia citrina 99 Pulmonata 3 Pungency 379 Pycnpgonids 105 Fyanose 638-640 derivative of 638 semi-synthesis of 638-640 F^olysis 455 ofkuwanonG 455 Pyrrole alkaloids 88 Quinicacid 144 derivatives of 144 Quinones 573 *'0-NMR spectrum of 573 Quinovicacid 123-125,134 derivatives of 123 glycosides of 124-125 antiviral activity of 134 Reduction of substrate 495 Reticulatacin 270 Retrochinensin 335 synthesis of 335

664 Rhatannins 439 Rheum palmatum 421,439 Rhizopus arrhizus 629 Rhizopus nigricans 629 Rho plus values 553 ROESY spectrum 129 of oligosaccharides 129 Rolliniastatin I 262,272 Rollinicin 277 Rollinone 278 Rosaceae 421 /?w.v.vw/a species 153 Russulares 200 Saccharomyces cerevisiae 241 Sanguisorba officinalis 421,423 polyphenols 423 Sativene 608 Saururus cernuus 319 Saxidom us giganti us 3 Saxitoxin 3 12-fi'/?/-Scalarin 10 Scopolamine 395 antichlolinergic agent 395 Scopoiia genus 395 Sequestration 93,104 Sesquicarene 609 Sesquiterpene 8,27,28,153-154, 196-197,607 Sessihugula translucens 88,94,98 Shahamin C 12 Shermilamine B 23 Shikimate pathway 471-472 Silver staining 396 Sindbis virus 135 12-
Sphinxolide 17 antitumor activity of 17 Spirostane glycosides 136 Spongia nitens 10 Spongia ojjicialis 10 Spongionella species 12 Staphylococcus aureus 285 Steganotaenia araliaceae 346 Stereoisomers 483 Stereospecific oxidation/reduction 484 Steric effects 556 '^O NMR spectrum of 624 Steroidal glycosides 483 substrates 483 Steroids 622-631 semi-synthesis of 622-631 Sterols 78-79,153,207,474 Stylocheilus longicauda 4,104 Stypodione 8 Substrate binding domain 487,494,500,502-503 forHLADH 500,502-503 surrogates 498 Sulfones 91 Sulfur-containing compounds 81 1 l-e;?/-Sinulariolide 22 Swertiajaponica 421 Tabidi 200 Taiwania cryptomeroides 333 Tambjamines 89,94,105 A , B , C & D 89 E & F 94 Tamus communis 130,132 Taurine 90 Terpenoids 4,15,613,642 metabolites 4 Tetrahydroaplysulphurin 15 Tetrahydrofurans 325 Tetrahydrofurofurans 344 Tetrahydrohalichondramide 17 Tetrahydronaphthopyrone 361 derivatives of 361 Tetraterpenoids 611 Thomasicacid 338 synthesis of 338 Thromboxanes 642 semi-synthesis of 642 Th uja plicata Donn. 338 Toxocara can is 379 Traditional medicine 113 (+)-7>*f7w.y-isopiperitenol 605

665 Transferase 480 Transformations 217 of bile alcohols 217 Tricellaria temata 92 Tricholomoidei 198 Tricyclic ketal 27 Tricyclic lactone 608 Tridiemnum gems 23 Trimusculus reticulatus 27 Triterpene 80,118 derivatives 118 Tropane alkaloids 395,406,416,418 biosynthesis of 395 "C-incorporation 416 distribution of 401,406 Tumor-inhibiting 341 Twisted 7r-electron system 41 Ulapualide A and B 16 Ulmus thomasii 338 (/ncana genus 122 Uncaria guaianensis 116,124 Uncaria tomentosa 116,118,124 Uvaria accuminata 251 Uvaricin 271 erf Values 553 Vallartanone A 26 Vallartanone, B 26 Van der Waals efiFects 556,558 Verrucosin A-B 15 Virola elongata 319 Virustatic pharmaceuticals 421 Vitamins D 621 semi-synthesis of 621 Vitamins 620,640 semi-synthesis of 620,640 Wachendorfia 372 Warbugia stuhlmannii 234 Warbugia ugandensis 234 Waiburganal 235 Water-soluble materials 481 Wieland-Miescher ketone 50,51,53,608 absolute configuration of 53 X-ray crystallography 53,88,91,254,262,296,303, 461,463,484,488 brominated, nargenicine 296 ofoxazoline 303

Xanthones 436 "C-NMR spectrum of 436 Xestoquinol 62 total synthesis of 62 CD and absolute stereochemistry of 66 Xestoquinone 34,62,66 total synthesis of 62 CD and absolute stereochemistry of 66 Xestospongia exigua 33 Xestospongia sapra 33 Xiphidium 372 YADH 482 Yakuchinone A and B 362 Yashabushidiol A 359 Yashabushiketodiol 359 Yeast alcohol dehydrogenase 482 Yessotoxin 20 Zinc bound alkoxide 481 Zinc clathridine 18 Zingiber spQCies 365 Zingiber officinalis 365 Zingiber qfficinarum 378 Zingiberaceae 358,362,364,377,379 Zinziber cassumar 365 Zinziber officinale 365 Zonarii 198 Zoobotryon verticillatum 85

Printed in the United Kingdom by Lightning Source UK Ltd. 106298UKS00001BD/21

780444 822659'