Part I

TERPENOIDS

1

Sesquiterpenoids

BY J. S. ROBERTS

1 Introduction

This chapter follows the now established biogenetic grouping of sesquiterpenoids. The literature that has been reviewed encompasses a slightly longer period than normal in view of the delay of certain journals in reaching the previous Reporter. The approximate period September 1978 — November 1979 has also witnessed a significant increase in the number of sesquiterpenoid syntheses and structural elucidation studies, and hence this Report is somewhat longer than in previous years. Professor F. Bohlmann and his group have contributed markedly to this increase by the publication of no less than 58 papers in sesquiterpenoid chemistry in the period under review.

2 Farnesane

Several new famesyl sesquiterpenoids have been reported; these include helepuberin acid (1) (Helenium puberulum), caulerpenyne (2) (green alga, Caulerpa prolifera), athanagrandione (3) (Athanasia grandiceps), and the two sweet potato stress metabolites (4) and (5), which are related to the corresponding diketone myoporone. The epimeric sesquiterpenoids (6) and (7) have been isolated from Osmanthus essential oil and both have been synthesized from the known ketones (8) and (9). The furanosesquiterpenoid longifolin (10), which is of both marine and terrestrial origin, has been synthesized (Scheme 1) although the yield in the final coupling step was only 4%.

A number of bicyclofarnesyl hydroquinones have already been isolated from the brown seaweed Dictyopteris undulata. The parent acyclic compound (11) has now been found in the fresh alga. Two papers on the terminal functionalization of farnesyl derivatives have been published. These include the use of 2,4,4,6-tetrabromocyclohexaienone as an alternative to N-bromosuccinimide for the formation of the bromohydrin of the terminal double bond of methyl farnesoate and farnesyl acetate. As is well known the bromohydrin can be converted into the corresponding epoxide with base. Masaki et al. have described the selective chlorosulphenylation of farnesol benzyl ether, which, by further elaboration (Scheme 2), has permitted the synthesis of the naturally occurring diacetate (12).

Still et al. have reported a highly stereoselective synthesis of the alcohol (14) which has been previously converted into the C18 Cecropia Juvenile Hormone (Scheme 3). The key step in this synthesis involves the recently described [2,3] sigmatropic rearrangement of the dilithio dianion derived from (13). Full details of the previously reported C17 and C18 Cecropia Juvenile Hormone syntheses have been published.

3 Bicyclofarnesane

Further biosynthetic studies by Cane et al. have shown that the conversion of farnesyl pyrophosphate (15) into nerolidyl pyrophosphate (16) proceeds by a net syn (suprafacial) process and that the subsequent cyclization to cyclonerodiol (17) occurs in a trans manner (Scheme 4). This careful piece of work was achieved by incorporation studies with doubly labelled nerolidol and mevalonate precursors and then by ascertaining the chirality of the acetate derived by Kuhn–Roth oxidation through enzymatic conversion (malate synthase/fumarase incubation) into labelled malate. In a subsequent series of experiments with labelled precursors, Cane et al. have confirmed that (i) only the C-7 hydroxy-group in cyclonerodiol is derived from water whereas the C-3 hydroxy-group is derived from the intermediate cyclonerodiol pyrophosphate by P — O bond cleavage and (ii) the absolute configuration at C-3 in nerolidyl pyrophosphate is retained through the cyclization and pyrophosphate hydrolysis steps. Finally, by using [1-18O]farnesyl pyrophosphate as the labelled precursor, it was demonstrated that the most probable mechanism for the syn conversion of farnesyl pyrophosphate into nerolidyl pyrophosphate involves a tightly bound ion pair (Scheme 5). This follows from the observed retention of one third of the 18O label in the derived cyclonerodiol. Other possible mechanisms would have demanded zero, one-sixth, or complete retention of the 18O label in cyclonerodiol. Cyclonerodiol has in fact been synthesized in a biomimetic manner by a HII-induced cyclization of nerolidol.

Nigaki alcohol (18) has been identified by spectroscopic and chemical means as a constituent of Picrasma ailanthoides Planchon. Latia luciferin (19) has been synthesized in a stereoselective manner. A key step in this synthesis involves the addition of lithium dimethylcuprate to an enol phosphate derived from a β-ketoester to form an α,β-unsaturated ester. Dehydro-β-ionilideneacetic acid (20), an important intermediate in the synthesis of abscisic acid, has been prepared, as have the two nor-abscisic acid derivatives (21). The metabolite (22) of abscisic acid has been identified in the seeds of Robinia pseudacacia L.

A number of rearranged monocyclofarnesyl sesquiterpenoids have been isolated from the sea hare Aplysia dactylomela. These include dactyloxene-A (23), -B (24), and -C (25) and dactylenol (26) together with its acetate. These compounds are related to other marine metabolites isolated from red algae. Aplysistatin (27), a metabolite of the sea hare Aplysia angasi, has been synthesized by a route which involves HgII-mediated brominative cyclization and a novel oxidative debenzylation step (Scheme 6).

The principal component of the defensive secretion of the termite soldiers Ancistrotermes cavithorax has been shown by spectroscopic methods and synthesis to be ancistrofuran (28). Further studies on the components of the defence secretion of the West African termite species Amitermes evuncifer have revealed the presence of the two bicyclic ethers (29) and (30) together with the eudesmane hydrocarbon (31). The latter compound is obviously related to the major component (32) of the secretion.

A detailed study of the 13C n.m.r. spectrum of ascochlorin (33) biosynthesized from [3-13C,4-2H2]mevalonolactone has supported an earlier proposal concerning the concerted nature of the 1,2-hydride and 1,2-methyl shifts (Scheme 7).

The potent biological activity (insect antifeedant, antitumour, antifungal) of warburganal (35) has stimulated considerable synthetic interest in this compound. Three total syntheses of this compound have been recorded in the period under review. The synthesis by Tanis and Nakanishi has additional flexibility since the key intermediate diol (34) can be used in the syntheses of cinnamolide (36), drimenin (37), and polygodial (38). Both norisoambreinolide (39; R = O) and isoambrox (39; R = H2) have been synthesized from (40), the product of the stannic chloride-catalysed cyclization of farnesyl phenyl sulphone. Yahazunol (41), a bicyclofarnesyl hydroquinone, has been identified in the brown seaweed Dictyopteris undulata Okamura, while ilimaquinone (42), which has a rearranged drimane skeleton, is a constituent of a Pacific Ocean marine sponge. Interestingly this latter compound is enantiomeric with respect to a related Mediterranean sponge metabolite. 13C N.m.r. spectral data and chemical evidence have led to a reassignment of the structure of spiniferin-1 (43). This novel bridged sesquiterpenoid co-occurs with spiniferin-2 (44) (in the sponge Pleraplysilla spinifera. This structure has now been confirmed. Two new sponge metabolites from Dysidea herbacea (Keller) are herbadysidolide (45) and the nor-seco-derivative herbasolide (46). The structures of both these compounds have been determined by X-ray analysis. The same species of sponge, collected off the Queensland coast, has also yielded the closely related compound spirodysin (47). In another study of Dysidea species Wells et al. have isolated and identified four more compounds, furodysin (48; R = H), furodysinin (49; R = H) and their corresponding thioacetates (48; R = SAc) and (49; R = SAc). It is interesting to note that BF3-catalysed rearrangement of spirodysin (47) gave a 1:1 mixture of (48; R = H) and (49; R = H).

The structure of karatavic acid (50) has been revised and as such is the first example of a seco-drimane sesquiterpenoid.

Biosynthetic studies have indicated a mixed polyketide-terpenoid origin for the unusual fungal metabolite andobenin (51) (Scheme 8). The two extra methyl groups (*) are derived from methionine. Andobenin co-occurs with andilesin (52) whose structure has been recently elucidated by X-ray and c.d. analysis.

4 Bisabolane, Sesquicarane

A large number of oxygenated bisabolane sesquiterpenoids (53) — (71) have been isolated and identified by Bohlmann's group (see also ref. 342). A third member of the marine-derived halogenated bisabolanes, deodactol (72), has been isolated from the sea hare Aplysia dactylomela. Its structure and absolute stereochemistry have been determined by X-ray analysis and are closely related to those of caespitol and isocaespitol.

Stereospecific syntheses of both (E)-and (Z)-α-bisabolenes, (73) and (74) respectively, have been carried out (Scheme 9) and the spectral data for each diastereoisomer have been recorded. The enantiomers of each have also been made starting from (+)- and (-)-limonene and as a result the β-bisabolene present in the essential oil of Opoponax has been shown to be the (+)-(S,Z)-isomer, thus correcting a previous report. As might be expected the odour characteristics of the (E)- and (Z)-isomers are subtly different. Both diastereoisomeric racemic α-bisabolols (75) and (76) (only one enantiomer of each is shown) have been prepared from the two isoxazolidines (77) and (78) which, in turn, were derived from intramolecular cyclization of the nitrones of (6E)- and (6Z)-farnesal respectively. From this work it is suggested that natural (-)-α-bisabolol must be (75) in contradiction to a recent report which held that (76) is the correct structure of the natural isomer. Further work will be needed to resolve this question.

New syntheses of ([+ or -])-ar-turmerone (79) and ([+ or -])-nuciferal (80) have been reported (Schemes 1048 and 1149) whereas Meyers and Smith have used the (+)-oxazoline (81) to good effect in an asymmetrically induced synthesis of (+)-ar-turmerone (82) (Scheme 12). A neat one-pot synthesis of β-curcumene (83) has been developed which involves only two steps (Scheme 13). In a synthesis of the aromatic analogue, (-)-α-curcumene (84), Kumada et al. have used an asymmetrically induced cross-coupling Grignard reaction in the presence of a nickel complex of (85) to produce (84) in 66% enantiomeric excess (Scheme 14). A Vilsmeier-Haack-Arnold formylation of (+)-limonene has been used as the key step in a synthesis of (+)-α-atlantone (86) (Scheme 15). In another use of monoterpenoids for sesquiterpenoid synthesis, the ene products (87)–(89) from (+)-limonene and (-)-carvone with methyl vinyl ketone and methyl propiolate respectively have been used to prepare (+ )-β-bisabolene (90), (-)-cryptomerion (91), (+)-β-atlantone (92), and (+)-α-atlantone (86) respectively. (-)-Limonene also features in a short synthesis of (-)-E-lanceol (96) in which the key step is conjugate addition of the lithio anion of (-)-limonene (93) to the keten dithioacetal (94) to give, after hydrolysis, the aldehyde (95), which could then be reduced to (-)-(E)-lanceol (96).

Spectroscopic evidence has been used to deduce the structures of isosesquicafone (97) and the derivative of sesquisabinene (98) isolated from Haplopappus tenuisectus and Arctotis grandis respectively. Two isoprenylogues of α-phellandrene, namely the methyl esters of 3,4-dihydronidorellaurin acid (99) and nidorellaurin (100), have been identified in Nidorella auriculata DC.

Sesquipinane, Sesquicamphane

Following on from Money's important work on the use of a monocyclic precursor for the synthesis of bicyclic and tricyclic sesquiterpenoids, Noyori et al. have now gone one step further and shown that the dibromo-ketone (101) (prepared from E,E-farnesol) undergoes a [3 + 2] intramolecular cycloaddition in the presence of pentacarbonyliron to give a mixture of campherenone (102) and epicampherenone (103) via the reactive 2-oxyallyl cation species.

(-)-trans-β-Bergamotene (104) has been reported to be a constituent of the aerial parts of Clibadium cf. asperum. This compound may not be new since a similar hydrocarbon was recorded by Cane and Nozoe although no optical rotations were given.

Two new syntheses of β-santalene (106) have been reported. In the first one, the starting material (105) was obtained from the Diels-Alder reaction between cyclopentadiene and methyl buta-2,3-dienoate followed by hydrogenation (Scheme 16). The second synthesis (Scheme 17) starts from camphene and involves a rearrangement of the γ-lactone by an exo-3,2-methyl shift with subsequent Wagner-Meerwein and hydride shifts to produce the δ-lactone (107). This lactone could also be converted into β-santalol (109) by standard methodology. Another synthesis of β-santalol (109) involves the construction of (108) by a Diels-Alder reaction and subsequent side-group transformations (Scheme 18). In the synthesis of the naturally occurring tricyclo-eka-santalol (110), the homocon jugate addition of phenylsulphenyl chloride was used to add the third ring (Scheme 19).

Full papers on the syntheses of ([+ or -])-isoalbene (111) and (-)-albene (112) have been published. Although there is no doubt that in this nice piece of work Kreiser et al. have synthesized naturally occurring albene from (+)-camphenilone, there still must remain some doubt about the absolute configuration of this interesting tricyclic olefin. This follows from the fact that a rare endo-3,2 methyl shift is proposed in the synthetic conversion of (113) into (114) which is ultimately transformed into (-)-albene (112). A more circuitous but better precedented route (Scheme 20) would produce the ketone (115) with the opposite absolute stereochemistry and hence it may be that the absolute stereochemistry of naturally occurring albene requires closer examination.

6 Cuparane, Trichothecane

Another simple synthesis of α-cuparenone (116) based on a [3 + 2] cyclo-addition has been published (Scheme 21).

The bicyclic lactone (117) has been considered as a useful synthon for the synthesis of verrucarol (118). A second synthesis of this lactone has been described (Scheme 22). Starting from verrucarol (118), Tamm et al. have appended the two requisite side-chains (119) which can be lactonized with di-(2-pyridyl) disulphide and triphenylphosphine to give tetrahydroverrucarin J (120). A recent investigation of the biosynthesis of trichodermin (121) using [1-13C]acetate is in accord with an earlier result although some reassignments of certain chemical shifts have been suggested.

7 Chamigrane, Widdrane, Thujopsane

A new synthesis of α-chamigrane (122) has been reported (Scheme 23).

The halogenated chamigranes and related metabolites from the marine algae of the genus Laurencia continue to be actively investigated, with X-ray analysis playing a crucial role in determining both relative and absolute stereochemistry. With the aid of this technique the absolute stereochemistries of iso-obtusol and obtusol have been reassigned as (123) and (124) respectively. González et al. have also examined a large number of these bromo-chloro-sesquiterpenoids and their derivatives by 13C n.m.r. spectroscopy and they have shown that subtle changes in chemical shifts can be used to assign the relative position and stereochemistry of the bromo- and chloro-groups (Scheme 24). The structures (125) and (126) have been assigned by this method to obtusane and isofurocaespitane respectively. A key biosynthetic intermediate in the sequence chamigrane -> perforatane -> perforane has now been isolated as a minor constituent of Laurencia perforata. The structure of this compound, perforenol (127), was deduced from chemical and X-ray crystallographic evidence. Two related compounds, guadalupol (128) and epiguadalupol (129), have been identified as metabolites of Laurencia snyderiae var. guadalupensis. An examination of the red alga Laurencia majuscula Harvey has resulted in the isolation of the bromo-chamigrane derivatives (130) and (131), and once again X-ray analysis of a derivative of (130) has led to the determination of absolute stereochemistry. Suzuki et al. have also established the absolute stereo-chemistries of the metabolites (132) — (136) from Laurencia glandulifera Kützing by an X-ray analysis of (132) and subsequent chemical interconversions with the other four compounds.