alternative approach, we confirmed the presence of the α-alkyl, β-oxo fatty acyl precursors by
converting them to α-alkyl, β-hydroxy fatty acids (mycolates) by prior reduction of bound
mycolates, using NaBH4. This resulted in the appearance of α, a´ and epoxy-MAMEs in
extracts of both whole and delipidated cells from the mutant strain (Figure 3C).
As expected (Minnikin and Polgar, 1966), reduction of the β-oxo mycolate precursors
gave a mixture of separable diastereoisomers and the presence of the β-epimer of α-MAME was
clearly seen (MAME-II, Figure 3C). It would be expected that the β-epimers of α'- and epoxy
mycolates would also be produced, but such minor compounds would not be readily seen on 1D-
TLC (Figure 3C). The epoxy function is also susceptible to NaBH4 reduction and isomeric
hydroxylated derivatives were identified (MAME-I, Figure 3C), corresponding to two artefacts
previously characterised in acid methanolysates (Minnikin et al., 1982). However, reduction of
the delipidated cells gave only the most polar hydroxylated derivative (Figure 3C). This strongly
suggests that access of NaBH4 to cell wall bound epoxy mycolates was restricted. It has recently
been shown that keto mycolates from Mycobacterium bovis BCG adopt a folded “W”
conformation (Villeneuve et al., 2007) with the keto group in a similar hydrophilic environment
as the hydroxy acid unit. It is reasonable to suggest that bound epoxy mycolic acids might also
prefer such a folded “W” conformation that could direct the access of NaBH4 in a regiospecific
manner, resulting in the formation of only a single hydroxylated derivative. It is notable that the
covalently bound β-oxo precursors in the ∆MSMEG4722 mutant also produce only the more
polar derivative. This would suggest that β-oxo precursors also fold in the same way as intact
mycolates, indicating that a β-hydroxy group is not absolutely essential prerequisite for folding
in a “W” conformation. Indeed, the fact that ∆MSMEG4722 mutant cells are viable, with β-oxo
12