A Common Approach to the Synthesis of Monocyclofarnesyl Sesquiterpenes

In such natural products the trimethylsubstituted ring is often additionally functionalized in the C3-neopentyl position (e.g. 2 4) as well as at the methyl-bearing tertiary carbon atom (e.g. 2 7). In view of the placement of functionality in the well known trans-decalones 8a and 8b it was quite obvious that they could serve a role as common starting materials for many monocyclofarnesylsesquiterpenes syntheses. More specifically, Baeyer-Villiger oxidation of the B-ring ketone would provide 9a/b containing the correctly functionalized carbocyclic ring with appropriate stereochemistry for a variety of natural products. Introduction of two carbons atoms α to the lactone-carbonyl (alternatively in the open chain form) would complete the construction of the monocyclofarnesyl skeleton, (Scheme 1). We are currently engaged in using this strategy for the synthesis of several sesquiterpenes. Recently we disclosed the total synthesis of (±)-farnesiferol-C 2 which closely followed along these lines. Here we would like to present a review and update on our ongoing efforts to synthesize aplysistatin 4, ambilol-A 5, 7-hydroxy-6,11-cyclofarnes3(15)-en-2-one 6 and ancistrofuran 7.


Introduction
Sesquiterpenes having a monocyclofarnesyl skeleton 1 are ubiquitous in nature 1 .
In such natural products the trimethylsubstituted ring is often additionally functionalized in the C3-neopentyl position (e.g. 2 -4) as well as at the methyl-bearing tertiary carbon atom (e.g. 2 -7 # ).In view of the placement of functionality in the well known trans-decalones 8a 2 and 8b 2a,3 it was quite obvious that they could serve a role as common starting materials for many monocyclofarnesylsesquiterpenes syntheses.More specifically, Baeyer-Vil-liger oxidation of the B-ring ketone would provide 9a/b containing the correctly functionalized carbocyclic ring with appropriate stereochemistry for a variety of natural products.Introduction of two carbons atoms α to the lactone-carbonyl (alternatively in the open chain form) would complete the construction of the monocyclofarnesyl skeleton, (Scheme 1).
We are currently engaged in using this strategy for the synthesis of several sesquiterpenes.Recently we disclosed the total synthesis 4 of (±)-farnesiferol-C 2 which closely followed along these lines.Here we would like to present a review and update on our ongoing efforts to synthesize aplysistatin 4 5 , ambilol-A 5 #6 , 7-hydroxy-6,11-cyclofarnes-3(15)-en-2-one 6 7 and ancistrofuran 7 8 .
The apparently straightforward B-ring ketone oxidation to the ε-lactones 9a/b bore some surprises.Thus, performing the oxidation on the hydroxy ketone 8b provided the un-conventional product 10 arising through participation of the 3β-hydroxy group.Substrates bearing an acid labile OH protecting group which is removed under the Baeyer-Vil- The mechanism of this rearrangement is currently the subject of investigation.Thus, the question of whether transanular oxa-ring formation occurs stepwise (8b → 9b → 12 → 10) or in a concerted fashion with direct participation of the 3β-hydroxyl (8b → 13 → 10) is being addressed by us by means of a computational analysis of the thermodynamic and kinetic aspects of the possible reaction paths 9 (see Scheme 2).
Whilst the interesting rearrangement in the Baeyer-Villiger reaction of 8b giving 10 allowed the synthesis of farnesiferol-C 2 4 , the other (lactone) products in hand (9a and 9c/d) open the way for our synthetic approaches to the natural products 4 -7.
The decalones 8a/b are prepared 2,3 from the well known Wieland-Miescher ketone 17 11 which has served as starting material in many syntheses 2a,2b,3a,4,12 .The advent of the availability of the Wieland-Miescher ketone in enantiomerically pure form 13 has widened its scope so as to permit the synthesis of natural products in their optically pure state 14 .We have recently found an improved and very efficient method for a rapid, selective and high-yielding preparation of the monoketal 18 of this diketone, uncontaminated by bisketal (which is extremely tedious to separate!) 15 .In summary, these developments allowed an expedient preparation of the common decalone intermediates 8a-d in appreciable quantity along established lines 3b and in optically pure form.
The decalone 8d was oxidised using trifluoroperoxyacetic acid / trifluoroacetic acid in dichloromethane providing exclusively the lactone 9d in 75% yield.The final two carbons of the monocyclofarnesyl skeleton were then incorporated by alkylation of the lithium enolate (LDA/TMEDA/THF/-78 °C) with 1,2-dibromoethyl ethyl ether, cleanly providing the ethoxyethylated product 19d (79%) as a 2:1 mixture of diastereoisomers.Although the stereochemistry at the alkylated carbon in 19d was of no consequence for subsequent synthetic manipulations, the alkylation was shown to have occurred exclusively from the α-face (opposite to the angular methyl substituent) by means of an X-ray crystal structure of the minor diastereomer; (Fig. 1).With a conclusive structure proof of the C-15 monocyclofarnesyl skeleton now in hand, the mixture of diastereomers 19d was treated with 5% aqueous K2CO3 in THF/acetone resulting in a one-pot conversion to the hydroxy butenolide 20d involving ε-lactone hydrolysis, butyrolactone formation of the intermediate γ-bromo acid and β-ethanol elimination thus completing the overall transformation of decalone 8d to the advanced sesquiterpene intermediate 20d in just three steps and 47% overall yield.The hydroxybutenolides 20 serve as intermediates for aplysistatin 4 17   furan 7 (from 20a).The latter compound was prepared in a similar fashion from 8a in 40% overall yield (see Scheme 4).Several routes from 20a to ancistrofuran can be envisaged.Our initial results indicate, that phenylselenation of the dienolate derived from the TBDMS-ether derivative gives, albeit in low yield (amongst other products) the deconjugated phenylseleno derivative 21.Clearly, oxidative elimination of this compound will provide the exocyclic α,β-unsaturated lactone 22, deprotection and ring closure of which would result in the butenolide 23 which has been previously converted to ancistrofuran 8g .Therefore the transformation of 21 into 23 would complete a formal total synthesis of ancistrofuran.At the time of writing, our efforts in this regard are continuing and we will report the results at an appropriate opportunity.Owing to the fact, that the absolute configuration of ancistrofuran is not known, we are in a position to answer this question by virtue of the use of chiral material stemming from the (+)-Wieland-Miescher ketone.
If the introduction of the final two carbon atoms of the monocyclofarnesyl skeleton is delayed until a later point in the synthetic sequence, one may open the way to farnesiferol-C 2 4 , ambilol-A 5 and 7-hydroxy-6,11-cyclofarnes-3( 15)-en-2-one 6 17 .In a procedure analogous to that used for the synthesis of farnesiferol-C, 4 a standard Barbier-Wieland approach (PhMgBr / Et2O then p-TsOH / CHCl3) provided diphenylolefin 24 from lactone 9a.At the time of deadline for submission we are involved in the oxidative cleavage of this compound to the hydroxy acid 25.Treatment of this intermediate with methyllithium followed by TBDMS protection and vinylmagnesium bromide addition (final two monocyclofarnesyl carbons) should result in the allylic alcohol 26 4 .It then remains to prepare the allylic bromide 27 4 and couple it with the 3-methylfuran Grignard reagent 28 16 in order to complete the synthesis of ambilol-A.