262 R. Baumes
ac3MH levels in wines range from sub olfactive threshold values, sometimes below
the analytical limit of quantification, to maxima of 100 ng/L for 4MMP and of a
few hundreds ng/L for ac3MH (Bouchilloux et al. 1998; Darriet et al. 1993, 1995;
Fretz et al. 2005; Guth 1997; Kotseridis and Baumes 2000; Lopez et al. 2003; Murat
et al. 2003; Schneider et al. 2003; Tominaga et al. 1996, 1998a, 2000a).
During the alcoholic fermentation, a yeast carbon-sulfur lyase with -elimination
activity releases the odorous thiols from theS-cysteine conjugates by cleavage
of the carbon-sulfur bond of the cysteine moiety (Tominaga et al. 1998b). Thus,
this enzymatic activity of yeast is necessary to cleave theS-cysteine conjugates
and it was shown that fermentation with differentSaccharomyces cerevisiaeyeast
strains generate very different amounts of the thiol odorants (Murat et al. 2001b;
Howell et al. 2004). However, although almost the whole amounts of precursors
are degraded, the yields in thiol odorants fromS-cysteine conjugates are always
low, whatever be the precursor and thefermentation medium (model or natu-
ral), ranging from 0.06% to l 0.6% for 4MMP (Murat et al. 2001b), and from
0.6% to 10.2 % for 3MH (Murat et al. 2001a). With regard to ac3MH, it has
no direct precursor in grape, but the yeast acetylates partly the alcohol func-
tion of 3MH to give ac3MH, as it acetylatesthe fusel alcohols to generate their
acetates. Thus, the biogeneration of these thiols is associated with theS- -lyase
and acetyl transferase activities of yeast, which depends greatly on the composi-
tion of the must, the conditions of fermentation and the yeast strain. Some strains
ofSaccharomyces bayanusvar.Uvarumwere reported to release efficiently these
thiols (Murat et al. 2001b; Masneuf-Pomarede et al. 2002; Howell et al. 2004;
Dubourdieu et al. 2006; Masneuf-Pomarede et al. 2006). However, yeast could
hardly degrade directly theS-glutathione conjugates without hydrolyzing the tripep-
tide into the cysteine conjugate, as previous studies have shown that a bacteria
S- -lyase extract does not release any thiol from this kind of precursor (Tomi-
naga et al. 1995). Regarding the stereoselectivity of 3MH release by certainS- -
lyases, recent studies showed that the degree of discrimination observed was low for
S- -lyases of enological yeasts, and moderate forS- -lyases of other sources
(Wakabayashi et al. 2003, 2004; Tominaga et al. 2006).
An alternative biogenetic pathway starting from (E)-2-hexenal in must and lead-
ing to 3-mercaptohexanol after yeast fermentation was recently evidenced
(Schneider 2001; Schneider et al. 2006). Similarly, various yeasts were shown
to produce 3-mercaptohexanol from both addition products of cysteine and hex-
2-enal,S-(1-hydroxyhex-3-yl)-L-cysteine (P3MH) and 2-(2-S-L-cysteinylpentyl)-1,
3-thiazolidine-4-carboxylic acid (Wakabayashi et al. 2004). The mechanism of
the pathways involved in the experiments reported by Schneider (2001) are still
unknown, but they should involve the addition of the sulfhydrylgroup of a sulfur
compound to the unsaturated hex-2-enal, leading to adducts which could be those
reported by Wakabayashi et al. (2004). Then, the enzymatic reduction by yeast of the
aldehydic group, and further reactions, such as a -elimination by yeastS- -lyase,
if the sulfur compound is different from hydrogen sulfide, would yield the thiol.
The formation of 4MMP from mesityl oxide during must fermentation by yeast
was similarly observed, but this compound was not reported as a must constituent