16
B. Schmidt
unprotected RCM precursor 46b gave exclusively the (Z)-diastereomer of nonenolide, (Z)-48 [51]. Oxidative cleavage of the PMB-protecting groups furnished the
natural product nonenolide ((E)-48). This very remarkable protecting group effect
was confirmed in follow-up studies by another research team, who found that
replacing the PMB group by a sterically more demanding TBS group completely
inhibits the RCM reaction and that in the absence of the C-9-methyl group the RCM
takes the same stereochemical course [53]. Effects of protecting groups on the rate
and stereoselectivity of the RCM step were reinvestigated in the most recent total
synthesis of nonenolide. No conversion was observed for a precursor with a TBSprotected OH-group at C-3 and a PMB-protected OH-group at C-6. After cleavage of
the TBS ether at C-3, however, RCM proceeds smoothly and with high (E)-selectivity
[54] (Scheme 12).
It should not be concealed that optimizing RCM reactions by tuning the metathesis
precursors can be laborious and sometimes remains unsuccessful. For example,
cytospolide E, the C-2-epimer of cytospolide D (35, cf. Scheme 6), has remained
elusive so far, although extensive attempts toward its synthesis via macrolactonization and RCM have been undertaken. Metathesis-based syntheses either result in
the formation of cyclic or acyclic dimers, or give the (Z)-isomer of cytospolide E
[55, 56].
2.3 Olefin Cross-Metathesis
Cross-metathesis (CM) is the intermolecular version of an olefin metathesis reaction
[57]. Compared to RCM, CM reactions have only been used sparingly in the total
synthesis of decanolides and in natural product syntheses in general. The reason
for this is most likely the inherent selectivity problem of cross-metathesis reactions: when equimolar amounts of similarly reactive alkenes are subjected to crossmetathesis conditions, three products (without taking (E)/(Z)-isomers into account!)
have to be expected, that is, the dimers of the starting alkenes and the actual crossmetathesis product. Grubbs and co-workers developed a classification scheme for
olefins according to their reactivity in cross-metathesis reactions [58] that allows
one to predict whether a CM reaction will have a selective or a statistical outcome.
Electronically and sterically unbiased olefins are very reactive and undergo fast
homodimerization (the so-called type-1 olefins), but the dimers are subsequently
consumed in the cross-metathesis reaction if a less reactive coupling partner is
present. Selectivity toward the desired CM product can often be improved by using
one coupling partner in excess, which is, however, a major disadvantage if both
CM partners have to be synthesized in multiple steps. An illustrative example is the
following approach to cytospolide E: the substituted pent-4-enoic acid 49 and the
alcohol 50 undergo cross-metathesis in the presence of second-generation catalyst B1
to the envisaged cytospolide E precursor 51 in fair yield, but 50, which was synthesized in six steps from a protected d-glyceraldehyde, had to be used in excess. Apart
from the desired CM product 51, the homodimerization products 52 and 53 were
B. Schmidt
unprotected RCM precursor 46b gave exclusively the (Z)-diastereomer of nonenolide, (Z)-48 [51]. Oxidative cleavage of the PMB-protecting groups furnished the
natural product nonenolide ((E)-48). This very remarkable protecting group effect
was confirmed in follow-up studies by another research team, who found that
replacing the PMB group by a sterically more demanding TBS group completely
inhibits the RCM reaction and that in the absence of the C-9-methyl group the RCM
takes the same stereochemical course [53]. Effects of protecting groups on the rate
and stereoselectivity of the RCM step were reinvestigated in the most recent total
synthesis of nonenolide. No conversion was observed for a precursor with a TBSprotected OH-group at C-3 and a PMB-protected OH-group at C-6. After cleavage of
the TBS ether at C-3, however, RCM proceeds smoothly and with high (E)-selectivity
[54] (Scheme 12).
It should not be concealed that optimizing RCM reactions by tuning the metathesis
precursors can be laborious and sometimes remains unsuccessful. For example,
cytospolide E, the C-2-epimer of cytospolide D (35, cf. Scheme 6), has remained
elusive so far, although extensive attempts toward its synthesis via macrolactonization and RCM have been undertaken. Metathesis-based syntheses either result in
the formation of cyclic or acyclic dimers, or give the (Z)-isomer of cytospolide E
[55, 56].
2.3 Olefin Cross-Metathesis
Cross-metathesis (CM) is the intermolecular version of an olefin metathesis reaction
[57]. Compared to RCM, CM reactions have only been used sparingly in the total
synthesis of decanolides and in natural product syntheses in general. The reason
for this is most likely the inherent selectivity problem of cross-metathesis reactions: when equimolar amounts of similarly reactive alkenes are subjected to crossmetathesis conditions, three products (without taking (E)/(Z)-isomers into account!)
have to be expected, that is, the dimers of the starting alkenes and the actual crossmetathesis product. Grubbs and co-workers developed a classification scheme for
olefins according to their reactivity in cross-metathesis reactions [58] that allows
one to predict whether a CM reaction will have a selective or a statistical outcome.
Electronically and sterically unbiased olefins are very reactive and undergo fast
homodimerization (the so-called type-1 olefins), but the dimers are subsequently
consumed in the cross-metathesis reaction if a less reactive coupling partner is
present. Selectivity toward the desired CM product can often be improved by using
one coupling partner in excess, which is, however, a major disadvantage if both
CM partners have to be synthesized in multiple steps. An illustrative example is the
following approach to cytospolide E: the substituted pent-4-enoic acid 49 and the
alcohol 50 undergo cross-metathesis in the presence of second-generation catalyst B1
to the envisaged cytospolide E precursor 51 in fair yield, but 50, which was synthesized in six steps from a protected d-glyceraldehyde, had to be used in excess. Apart
from the desired CM product 51, the homodimerization products 52 and 53 were