1,5-Induction in Reactions Between 2-Alkoxy-(2-tributylstannylethylidene)cyclohexanes and Aldehydes

2-Methoxy(2-tributylstannylethylidene)cyclohexane 16 undergoes stereoselective transmetalation on treatment with tin(IV) chloride to generate the allyltin trichloride 19 which reacts with aldehydes to give the (Z)-(3-hydroxyalkylidene)-2-methoxycyclohexanes 17a-g with excellent 1,5-stereocontrol.

In order to delineate the scope of these reactions, it was necessary to establish whether the chemistry is compatible with the presence of substituents at C(2) and C(3) in the allylstannane. It has been found that the stereoselectivity is not significantly affected by substituents at C(2); indeed it is slightly enhanced. For example, the 2methylpent-2-enyl(tributyl)stannane 7, as an (E)/(Z)mixture, reacts with aldehydes on transmetalation with tin(IV) chloride with excellent 1,5-stereoselectivity 5 (Scheme 2.) However, the effect of a substituent at C(3) is difficult to predict, and has not yet been investigated. Indeed transmetalation of a 3,4-disubstituted alk-2-enylstannane 9 would generate an intermediate tertiary allyltin trichloride 10, and the stability of such an intermediate, e.g. towards elimination of a tin alkoxide, is very uncertain. We now report aspects of the chemistry of the 3,4-disubstituted allylstannane 16. This was found to undergo highly stereoselective transmetalation with tin(IV) chloride giving an intermediate allyltin trichloride which is stable under our reaction conditions and which undergoes usefully stereoselective reactions with aldehydes.

Results and Discussion
The 2-methoxy(2-tributylstannylethylidene) cyclohexane 16 was prepared as outlined in Scheme 3. The condensation between racemic 2-methoxycyclohexanone 11 and triethyl phosphonoacetate gave the unsaturated ester 12 in an 87% yield as a 60 : 40 mixture of (E)-and (Z)-isomers. Interestingly, the corresponding Peterson reaction using ethyl trimethylsilylacetate gave more of the (Z)-isomer, (E) : (Z) = 20 : 80, albeit in a lower yield (47%). The (E)-and (Z)-isomers could be readily distinguished by 1 H NMR, for example 2-H for the (Z)-isomer, at d 5.26, was significantly more deshielded than 2-H for the (E)-isomer, observed at d 3.59. After reduction of the ester, the allylic alcohol 13 was converted into the xanthate 14 with good yields being obtained if the deprotonation of the alcohol was completed by heating in toluene under reflux prior to the addition of the carbon disulfide. The xanthate, still as a 60 : 40 mixture of (E)-and (Z)-isomers, was converted into the dithiocarbonate 15 by heating in toluene under reflux. This rearrangement was quite stereoselective giving an 87 : 13 mixture of diastereoisomers. It is believed that the major product has the vinyl and methoxy groups cisdisposed about the six-membered ring, but this stereochemical assignment was not confirmed. Instead the mixture of dithiocarbonates was taken through to the allylstannane 16 as an 85 : 15 mixture of (E)-and (Z)isomers, by treatment with tributyltin hydride under freeradical conditions, the stereochemical assignment being made on the basis of strong nOe effects between H-2 and the vinylic proton for the major, but not for the minor isomer.
Reactions between the stannane 16 and aldehydes were carried out by treatment of the allylstannane with tin(IV) chloride at -78 o C for 5 min followed by addition of the aldehyde. After a further 45 min at -78 o C the reactions were quenched to give, after chromatography, the products indicated in Scheme 4 6 .
In all cases the reactions appeared to be highly stereoselective with only one product being detected by 1 H and 13 C NMR (>95 : 5). In order to confirm that the 1,5diastereoisomers were distinguishable, the benzaldehyde derived product 17a was converted into its diastereoisomer 18 by inversion using a Mitsunobu reaction followed by saponification (Scheme 5). Although the isomers 17a and 18 were inseparable by TLC, their 1 H NMR spectra were quite different, e.g. 2-H and OCH 3 were at d 4.08 and 3.23 for 17a and at d 4.20 and 3.17 for 18, respectively. Close examination of the crude reaction mixture from the reaction

Scheme 2.
between benzaldehyde and the stannane 16 indicated that only a trace, <5% of 18 was present.
The geometry of the double-bonds in the products 17 was confirmed by nOe studies. For example, for 17a and b significant enhancement of 2'-H 2 but no enhancement of 1'-H was observed on irradiation of 2-H.
The regioselectivity of the aldehyde-stannane reactions, and the selective formation of the (Z)-products, are consistent with transmetalation of the allylstannane 16 on reaction with tin(IV) chloride to give the intermediate allyltin trichloride 19. Reaction with aldehydes through the chair-like transition structure 20 would then give the 1,5-syn-(Z)-products 17 (Scheme 6). This stereoselectivity has precedent in the reactions of aliphatic allylstannanes, e.g. 1 1 . The relative configurations assigned to the stereogenic centres in the products 17 were finally confirmed by an X-ray crystal structure of the pnitrobenzoate ester of the product 17d 7 .
This work has shown that a 3-alkyl substituent does not disrupt the transmetalation using tin(IV) chloride of a 4-alkoxyalk-2-enylstannane and that the allyltin trichlorides generated from 3,4-disubstituted alk-2-enylstannanes can react with aldehydes with excellent stereocontrol. Further work will explore the chemistry of the products, e.g. 21, obtained using related stannanes, and their application to the synthesis of fragments of natural products, e.g. 22 (Scheme 7).