Enolisation Regioselectivity

Hence ketone (1) might be made by regioselective alkylation of (2) but this is doubtful. A safer route is to disconnect the ethyl group to leave nitrile (3) which can certainly be made by alkylation of nitrile (4) as there is only one site for enolisation. 

Examples and Problems from Eaah Section of the Chapter Self-Condensation There have been many examples of self-condensations in Chapter 19 so two more will be enough to remind you about regioselectivity of enolisation (pTl54). 

Once regioselective and stereoselective controls have been exerted, the cis-decalins must be isomerised to rrans-decalins, the configuration present in the target molecules. Since frans-decalins are thermodynamically more stable than the corresponding cis-decalins, it is possible to isomerise the latter through enolisation, a process that can be favored by the presence of a carbonyl group near to the centre to be inverted. 

Whereas the thermodynamic route described above relied on reagent control to establish the spongistatin C19 and C21 stereocentres, the discovery of highly stereoselective 1,5-anti aldol reactions of methyl ketones enabled us to examine an alternative,16 substrate-based stereocontrol route to 5. Regioselective enolisation of enantiomerically pure ketone 37, derived from a readily available biopolymer, gave end... 

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. 

Another important contribution is to the regioselectivity of enolate formation from unsym-metrical ketones. As we established in chapter 13, ketones, particularly methyl ketones, form lithium enolates on the less substituted side. These compounds are excellent at aldol reactions even with enolisable aldehydes.15 An application of both thermodynamic and kinetic control is in the synthesis of the-gingerols, the flavouring principles of ginger, by Whiting.16... 

Acyl anions require special chemistry because they have umpolung of reactivity specific enols have normal reactivity but the problem of regioselectivity must be solved. The carbonyl compound must enolise under the reaction conditions only on the required side and, as carbonyl compounds are also electrophilic, it must not condense with itself under these conditions. 

The silyl enol ethers of esters, e.g. Ill and lactones, e.g. 114 similarly take part in efficient aldol reactions with enolisable aldehydes and ketones with Lewis acid catalysis, again with complete regioselectivity. Example 113 is particularly impressive as the very enolisable ketone gives a high yield of an aldol product with two adjacent quaternary centres.21... 

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. 

Mechanism may help us. If the required enolate is particularly easy to make, or the electrophile particularly reactive and/or unable to enolise, these are good points. If there are problems of chemo- or regioselectivity we cannot solve, then these are bad points. Stereochemistry may be a problem too. There is not much you can do about the stereochemistry of an enone made by the aldol reaction as you tend to get the thermodynamically more stable isomer (E or Z) and if you want the other you should try a different strategy. 

The point of this is that the proposed starting material 80 can then be made by a simple Wittig reaction with the stable ylid 81, the equivalent of an enolate of MeCHO, as the extra ylid stabilisation avoids any regioselectivity problems in enolisation. The cyanohydrin derivative 82 can lose the marked proton with strong base and alkylation occurs as planned. The product 83 can be hydrolysed... 

The diketo-aldehyde 37 has four electrophilic carbon atoms (A-D) and three positions for enolisation (1-3). Cyclisation could occur in twelve different ways. Three can be regioselectively realised by different conditions.6 With morpholine catalysis, the aldehyde forms an enamine which does a Michael addition on the enone (B) to give 39. Enamine formation with PhNHMe also gives the aldehyde enamine, but this time it does an aldol condensation with the simple ketone (D) to give 38. 

The second 119 clearly came from the ketone 121 and this looks like an aldol product from CH20 and a specific enolate of the ketone 121. This ketone is substituted at only one a-atom, so regioselective enolisation would be possible. But there are problems of chemoselectivity (the other CHO group) and stereoselectivity and a Michael addition approach on the enone 122 looked more promising. Aldol disconnection of 122 seems to have uncovered an excellent symmetrical intermediate 123 which can cyclise only to give 122. 

As the reaction is carried out in acidic solution, we can expect modest regioselectivity for enolisation on the more substituted side of the carbonyl group. An example from Vogel12 is the oxidation of butanone 96 to give the monoxime 98 via the nitroso compound 97. The dioxime (a useful ligand for metals) 99, the diketone, or the monoxime 98 can be made by this route. Reduction of the oximes gives primary amines. 

This approach can use the inherent regioselectivity of silyl enol ether formation (chapter 3) using kinetic or thermodynamic enolisation. Hence kinetic enolisation of enones (chapter 11) occurs on the a side leading to 2-Me3SiO-butadienes such as 222. Epoxidation of this silyl enol ether gives the unstable silyloxy ketone 223 which can be desilylated by fluoride ion and hence transformed into the hydroxyketone 225 or acetoxy ketone 224. These transformations are useful because the hydroxy ketones can be unstable34 (see below). 

The reasons for the choice of reagent for each reaction are instructive. Epoxidation of a silyl enol ether was chosen for the hydroxylation for two reasons. It was known that the potentially difficult (the a-atom on both sides is primary) regioselectivity of enolisation of ketones 228 could be solved by using LDA and HMPA. The lithium enolate is formed away from the quaternary ring junction. Then epoxidation of related alkenes was known to go on the bottom face and the stereochemistry of the epoxidation determines the stereoselectivity of the hydroxylation 232. 

In a typical reaction, the lithium enolate of the carbonyl compound 247 is prepared in the usual way and reacted with yellow MoOPH to give the hydroxylated product 248 and blue molybdenum compounds so that the reaction is self-indicating. The best feature of the method is that it uses lithium enolates since the regioselective preparation of these intermediates is much easier to control than that of other metal enolates. Hence 2-phenylcyclohexanone 249 gives a single regioisomer of the hydroxyketone 250 from kinetic enolisation with good stereoselectivity. 

The one-carbon fragment is ethyl formate. This reaction is important as a method of control since it occurs only on one side of the carbonyl group that is it is regioselective. The reason is that this product can itself enolise in... 

The order of events may be changed. The regioselectivity problem disappears in the alkylation of an ester since esters can enolise on one side only. Therefore, if we complete all our alkylations before making the ketone, we can start from malonate and make unsymmetrical ketones. This means disconnecting the ketone (9) first by a 1,1 C-C process (see Chapter 13). Is later chapters devices like this will become more important as target molecules get larger. 

For an unambiguous condensation, there must be no doubt as to which compound enolises and which compound acts as the electrophile (chemo-selectivity) nor any doubt about the regioselectivity of the enolisation. We need definite and favourable answers to three questions ... 

Regioselectivity of enolisation is still important. Acid (33) was used in a (very early ) stage of Woodward and Eschenmoser s vitamin B 2 synthesis. Disconnection gives an unsymmetrical ketone (8) and reactive, non-enolisable (34). Reaction in acid (page 154) ensured enolisation on the more substituted side of (8). 

The first two examples allow for control over the regioselectivity of enolisation of an unsymmetrical ketone if only one side can dehydrate or form an anion. Formaldehyde reacts with (84) at its most enolisable site, ° but the product (85) reacts at its methyl group as the alternative, and no doubt favoured, product (86) cannot dehydrate. The final product (87) was needed in an alternative approach to patulin (see page 158). 

The regioselective reaction of HC02Et or CO(OEt>2 with an unsymmetrica ketone is important as it activates that side of the ketone towards enolisation (Chapters 14 and 20). Thus ketone (89) gives (90) and not (88) on condensation with CO(OEt)2 as only (90) can form a stable enolate ion (91). The product was needed for a synthesis of the ant pheromone (92) (see Chapter 29). 

This approach was discussed in Chapter 6. The synthesis of the spasmolytic Diphepanol (28) is an example. Disconnection of one or both phenyl groups gives a-amino carbonyl compounds which can be made from the corresponding a-halo compounds (see Chapter S). Halogenation of both (29) and (30) is regioselective as the other side of the carbonyl group cannot enolise. 

---