Isobutyraldehyde, enol acidity

Preparation of (R)-(+)-3-hydroxy-4-methylpentanoic acid has been reported previously by the submitters.5 Alternative syntheses of (R)-(+)- or (S)-(-)-3-hydroxy-4-methylpentanoic acid rely on aidoi reactions of chiral ketone, ester, or amide enolates,2 8 10 and Lewis-acid mediated additions of chiral silyl ketene acetals to Isobutyraldehyde.3 11 Since both enantiomers of HYTRA are readily available this method enables one to prepare (S)-3-hydroxy-4-methylpentanoic acid as well. 

In certain cases, high levels of selectivity in the asymmetric aldol reaction can be achieved in the absence of a metal salt. The amino acid proUne catalyses the aldol reaction of aldehydes or ketones (which are enolizable) with aldehydes (preferably non-enolizable or branched to disfavour enolization) to give p-hydroxy-aldehydes or ketones. For example, use of acetone (present in excess) and isobutyraldehyde gave the (3-hydroxy-ketone 81 (1.88). The reaction involves an enamine intermediate and is thought to proceed via the usual Zimmerman-Traxler chair-shaped transition state. 

Preliminary formation of the silyl enol ether from either aldehyde, in the absence of the other, would be trouble-free as MejSiCl captures the enolate faster than self-condensation occurs. Here we need the silyl enol ether from isobutyraldehyde. The other aldehyde is now added along with the necessary Lewis acid, here TiClj. The mechanism described on p. 627 gives the aldol after work-up in an excellent 95% yield. No more than 5% of other reactions can have occurred. 

Despite the similarity of the structures of the silicon enolates 186 and 189 and essentially identical reaction conditions, the rationale for the stereochemical outcome, offered by the authors, is completely opposite the predominant approach of silyl ketene acetal 186 to isobutyraldehyde was assumed to occur through a Zimmerman-Traxler-llke transition state 192 where the titanium salt is embedded in the cycle. On the contrary, an open transition state model 193 was proposed for the Mukalyama reaction of silicon enolates 189. Both models of intuitive character give an explanation for the favored topicity the attack of the enolate to the Si-face of the aldehydes. Thus, the fact that ti-configured aldols are formed diastereoselectlvely Is in accordance with the tr ws-enolate/ nti-aldol correlation predicted by the Zimmerman-Traxler model, but an open model might be suitable to explain the stereochemical outcome as well. Both the Helmchen and the Oppolzer auxiliary were applied as acetates to give cx-unbranched-fi-hydroxycarboxylic acids. 

---