.
Angewandte
Communications
of TMSCHN2 (entry 4).[13] Other polar aprotic solvents like
acetone and dichloromethane gave lower yields of 2a after
longer reaction times and, in addition, small quantities of the
vinyloxirane 3a, as a mixture of Z/E isomers, were observed
(entries 5 and 6). Interestingly, while the primary alcohol
MeOH gave a complex mixture of products (entry 7), use of
the secondary alcohol iPrOH allowed the isolation of 2a as
the Z isomer in good yield (71%) after a short reaction time
(entry 8).
This encouraging result led us to explore the scope and
limitations of the cyclization (Table 2). The alkynals 1b,c (R1,
R2 = CH2OBn, CH2OAc), bearing two Csp3-type substituents
at C3, also gave fairly good yields of the corresponding
vinyldihydropyrans 2b,c. The monosubstituted alkynals 1d–g
allowed a study of the diastereoselectivity of the cyclization.
Thus, the methoxycarbonyl alkynal 1d (R1 = CO2Me, R2 = H)
cyclized to give 2d as a single syn diastereomer, but unfortu-
nately, the yield was low. To our delight, the benzyloxymethyl
and acetoxymethyl alkynals 1e,f (R1 = CH2OBn and
CH2OAc, R2 = H) were smoothly converted into 2e,f in
good yields in a completely diastereoselective syn fashion.
Owing to the critical role of C4 oxygenated substituents in
natural dihydropyrans with biological activity,[1] we decided to
evaluate the cyclization of the 3-silyloxyalkynal 1g (R) (R1 =
H, R2 = OTBS). To our initial surprise, the oxygenated
dihydropyran 2g was obtained in good yield, but the
diastereoselectivity was rather low in both iPrOH and Et2O
(syn/anti 3.3:1 and 1.7:1, respectively). We believe that the
oxygenated substituent in 1g might coordinate to the key
ruthenium intermediate and therefore modify the diastereo-
selectivity (see Scheme 3). Even the more bulky 3-silyloxy-
alkynal 1h (R1,R2 = H, OTIPS) gave only a slightly higher
diastereoselectivity for the dihydropyran 2h (syn/anti 4:1 in
iPrOH). Interestingly, the C2-monosubstituted 2-propyl-5-
hexynal exclusively gave the 5-propyl-2-vinyl-3,4-dihydro-
pyran (4) in moderate yield, thus showing the influence of the
nature of the reacting conformer in the reaction course.[14]
Notably, cyclization of alkynal 1i, with substituents at C3 and
C4, again leads to high levels of diastereoselectivity and gave
rise exclusively to the syn vinyldihydropyran 2i in fairly good
yield. Chemoselectivity of the reaction was analyzed during
cyclization of the difunctionalized enynal, 3-(prop-2-ynyl)-
hex-5-enal, in which the major isolated product 5 derives from
the cyclization of the enyne[10] and the minor dihydropyran 5’
derives from the cyclization of the ynal.
Table 2: Ruthenium-catalyzed cyclizations of the alkynals/alkynones 1 to
the 2-vinyldihydropyrans 2.
The cyclization of alkynones was subsequently explored.
Initially, the alkynyl methyl ketone 1j (R1, R2 = CO2Me, R =
Me) was subjected to the typical reaction conditions in iPrOH
and this gave moderate yields of a mixture of the vinyl-
dihydropyrans 2j (R = TMS) and 2j’ (R = H). It was sus-
pected that the enolizable ketone could cause partial desily-
lation of TMSCHN2 in iPrOH and give rise to the mixture of
cyclized products. As a result, the reaction was also performed
in MeOH and this exclusively gave the expected desilylated
2j’ in similar overall yield.[10] By contrast, the non-enolizable
alkynyl phenyl ketone 1k (R1, R2 = CO2Me, R = Ph) cyclized
more smoothly and cleanly than 1j in both iPrOH and CH2Cl2
to exclusively give the silylated dihydropyran 2k in fairly
good yield. In contrast, the desilylated dihydropyran 2k’ was
obtained in very good yield when the cyclization was
performed in MeOH. Pleasingly, the methoxycarbonyl alky-
none 1l (R1 = CO2Me, R2 = H, R = Ph) cyclized to the
dihydropyran 2l as a single syn diastereomer in moderate
yield as compared to the low yield obtained with 1d. To our
delight, the enantiomerically pure silyloxy phenyl and meth-
ylketones (R)-1m and (R)-1n (R1 = H, R2 = OTBS) cyclized
diastereoselectively to give the corresponding syn vinyldihy-
dropyrans (R,R)-2m and (R,R)-2n in rather good yields.
Remarkably, steric factors in the most stable conformer of the
key ruthenium intermediate probably control the diastereo-
selectivity of the cyclization process (2m or 2n versus 2g).
Conformationally locked bicyclic morpholines (dihydro-
1,4-oxazines), for example, oxabispidines and 8-oxa-3-
azabicyclo[3.2.1]octanes, have been shown to display
[a] Typical conditions: [Cp*RuCl(cod)] (10 mol%), TMSCHN2
(1.8 equiv), iPrOH, RT, [1]=0.15m. [b] Et2O as solvent. [c] MeOH as
solvent. [d] CH2Cl2 as solvent. TBS=tert-butyldimethylsilyl, TIPS=tri-
isopropylsilyl.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 5959 –5963