Lithium enolates chemoselectivity

We need now to look at situations where both compounds might enolise and see how specific enolates can be used to control which compound does so (chemoselectivity) before looking at how we control which side of an unsymmetrical ketone forms the enolate (regioselectivity). We met two specific enol equivalents in chapter 13 (5-dicarbonyl compounds and lithium enolates and they are the keys to this section. 

C with LDA results in the chemoselective formation of an aza-enolate D, as in the case of the analogous aldimine A of Figure 10.30. The C=C double bond of the aza-enolate D is fnms-configured. This selectivity is reminiscent of the. E-preference in the deprotonation of sterically unhindered aliphatic ketones to ketone enolates (Section 10.1.2, paragraph Stereocontrol in the Formation of Lithium Enolates ) and, in fact, the origin is the same both deprotonations occur via six-membered ring transition states with chair conformations. The transition state structure with the least steric interactions is preferred in both cases, which is the one that features the C atom in the /3 position of the C,H acid in the pseudoequatorial orientation. 

The best results are obtained with trisyl azide, which again leads to high yields of the azide transfer product 2, especially if the enolate 1 is added to trisyl azide (see entries 1 and 2). Interestingly, the best chemoselectivity and, in addition, identical yields of azide (73%) result from the reaction of the lithium enolate with trisyl azide (entry 3). The reaction of the ester enolate 1 with trisyl azide is less sensitive to the nature of the enolate metal than is the corresponding imide enolate reaction (see Section 7.1.1.). Acetic acid quench, on the other hand, again proved to be useful. Unfortunately, bis-azidation to 3 and diazo transfer to 4 are also observed. 

We find that we need a crossed aldol condensation between two ketones so tve chemoselectivity. We also need to make one enol(ate) from an unsymmetrical ketone so v. r regioselectivity too. The obvious solutions are a lithium enolate, a silyl enol ether, or a come with an extra ester group. 

Lithium enolates do not even solve all problems of chemoselectivity most notoriously, they fail when the specific enolates of aldehydes are needed. The problem is that aldehydes self-condense so readily that the rate of the aldol reaction can be comparable with the rate of enolate formation by proton removal. Fortunately there are good alternatives. Earlier in this chapter we showed examples of what can go wrong with enamines. Now we can set the record straight by extolling the virtues of the enamines 96 of aldehydes.17 They are easily made without excessive aldol reaction as they are much less reactive than lithium enolates, they take part well in reactions such as Michael additions, a standard route to 1,5-dicarbonyl compounds, e.g. 97.18... 

We shall discuss further aspects of the aldol reaction in the next two chapters where we shall see how to control the enolisation of unsymmetrical ketones, and how to control the stereochemistry of aldol products such as 121. We shall return to a more comprehensive survey of specific enol equivalents in chapter 10. In this chapter we are concerned to establish that chemoselective enolisation of esters, acids, aldehydes, and symmetrical ketones can be accomplished with lithium enolates, enamines, or silyl enol ethers, and we shall be using all these intermediates extensively in the rest of the book. 

The synthesis was planned around the reaction of a specific enolate of ester 136 with the epoxide 137. This reaction was expected to give mainly trans 138 and is chemoselective both because of the usual enolate problem and because 137 contains a terminal alkyne. The lithium enolate was too basic and the aluminium enolate was used instead. The reaction gave an 85 15 mixture of trans and cis 138 and also an 85 15 mixture of trans and cis 139 after cyclisation. Dihydroxylation by osmylation gave a mixture of diols 140 this was deliberate so that they could determine the stereochemistry at C-2 . To the surprise of the chemists, natural rubrynolide was identical to one of the minor (i.e. cis) diols in the 15% part of the mixture. Careful NMR analysis showed that it was 135a. 

Reactions in which the enol or enolate (or equivalent) of one carbonyl compound reacts with an electrophilic carbonyl compound (usually both are aldehydes or ketones) are often loosely called aldol reactions. In the last chapter we saw how the use of lithium enolates and other specific enolate equivalents conquers the problem of chemoselectivity in enolisation of aldehydes and acid derivatives. In this chapter, we are going to use the same intermediates to solve the problem of regioselectivity in crossed aldol reactions in which the enolising component is an unsymmetrical ketone. 

Treatment with BuLi generates chemoselectively the lithium enolate of the less substituted lactim and electrophiles attack the face opposite the branched isopropyl group. Selectivity is good the purified diastereoisomers can be isolated in over 80% yield and hydrolysis requires only dilute aqueous acid as these are easily protonated imines. These examples show a benzylic and an allylic halide. The first 50a is unnatural (R) -phenylalanine and the second 50b is an unnatural amino acid. 

The aldol reactions of titanium enolates have been the best studied of all the transition metal enol-ates."- In many cases they show higher stereoselectivity and chemoselectivity in their reactions than lithium enolates and are easily prepared using inexpensive reagents. They also promote high levels of diastereofacial selectivity in reactions of chiral reactants. The Lewis acidity of the titanium metal center can be easily manipulated by variation of the ligands (chloro, alkoxy, amino, cyclopentadienyl, etc.) attached to titanium, which leads to enhanced selectivity in appropriate cases. Moreover, the incorporation of chiral ligands on titanium makes possible efficient enantioselective aldol reactions. 

However, the more hindered, less basic lithium hexamethyldisilazamide reacts slowly with 1 at 0 °C to provide chemoselectively the desired enolate species 5. The a-protons of these rhenium-acyl complexes are believed to have a lower pKa than the cyclopentadienyl protons, but unless treated with hulky, selective hases the cyclopentadienyl protons exhibit greater kinetic acidity due to statistical factors and an earlier, reactant-like transition state since minimal rchybridiza-tion is required at the anionic center after cyclopentadienyl deprotonation. Equilibration of the cyclopentadienyl anion to the thermodynamically more stable enolate species cannot compete with the rapid acyl migration84. 

Methyl ester 431 is tethered by an alkyl chain to an acrylate Michael-acceptor and activated toward decarboxylation by a C-2-ethoxycarbonyl group. As a result of this, chemoselective Sn2-dealkylation of the methyl ester, decarboxylation and cyclization of the enolate by Michael addition occurs upon exposure to lithium chloride in DMEU, affording chroman 432 in excellent yield and diastereoselectivity (Equation 178) <1998JOC144>. 

Dicarbonyl compounds continue to attract attention as precursors to unsaturated ketones. Readily prepared trimethylsilyl enol ethers react well with alkyl-lithiums, but poorly with Grignard or dialkylcopper-lithium reagents, to give enones [equation (23)]. a-Oxoketene thioacetals undergo chemoselective... 

---