Angewandte
Chemie
furan 43 was obtained with a similar enantiomeric excess of
84%, which could be readily improved to enantiopurity by
crystallization.
The reaction sequence was relatively insensitive to steric
bulk around the alcohol functionality, for example, tertiary
alcohol 33 gave tetrahydrofuran 44 in 67% yield. Both a,b-
disubstituted b-hydroxyketone diastereoisomers 34 and 35
were subjected to rhodium catalysis with diazo compound 1 to
give single diastereoisomers 45 and 46, respectively. The latter
results show that, when stereocontrol over positions 4 and 5 of
the final tetrahydrofuran product can be achieved, the
substituent at C4 has a greater effect on the stereocontrol
than that at C5. Additionally, the structures of cyclic
compounds 38 and 43 were confirmed by X-ray crystallog-
raphy (Figures S1 and S2, Supporting Information).
The examples shown in Scheme 3 and 4 clearly establish
the wide scope of this tetrahydrofuran synthesis. However, in
order to rationalize the formation of tetrahydrofurans such as
À
4, as opposed to O H insertion (e.g. 3) products, we
performed a series of control experiments. First, in order to
establish whether the tetrahydrofuran 4 was formed through
À
an initial O H insertion reaction and subsequent intra-
À
molecular aldol cyclization, the ketoester O H insertion
product 3 was exposed to the reaction conditions (i.e. rhodium
octanoate or copper(I) triflate/toluene complex in CH2Cl2 at
reflux; Scheme 5). Under these conditions, no conversion to
Scheme 5. Control experiments. Reagents and conditions: a) [Rh2-
(oct)4] (1 mol%), CH2Cl2 at reflux; b) [Rh2(oct)4] (1 mol%), NEt3
excess, CH2Cl2 at reflux; c) (CuOTf)2·Tol (5 mol%), CH2Cl2 at reflux;
d) NaOMe, MeOH, 08C!RT.
Scheme 4. Synthesis of polysubstituted tetrahydrofurans by rhodium-
the tetrahydrofuran 4 was observed, thus showing that the
metal catalysts were not capable of mediating the intra-
molecular aldol reaction. The reaction was repeated, but this
time in the presence of an excess of triethylamine to
encourage the formation of a reactive aldol intermediate
(enol/enolate), but once again, no tetrahydrofuran 4 was
formed. Finally, we treated the ketoester 3 with a stronger
base (NaOMe) in an attempt to force the aldol process, but
under these conditions only methyl mandelate (44%) was
formed as a result of a retro-Michael elimination of methyl
vinyl ketone and transesterification, and no trace of the
tetrahydrofuran 4 was observed. In order to assess the
reversibility of the formation of the five-membered ring, we
exposed tetrahydrofuran 4 to the original reaction conditions
(Scheme 5). Tetrahydrofuran 4 was stable in the presence of
the metal catalysts, and no retro-aldol product 3 was
observed. Even with the addition of an excess of triethyl-
amine, the tetrahydrofuran 4 was stable, and a stronger base
(NaOMe) was required to cause decomposition of 4. Under
these conditions, methyl mandelate (73%) was formed,
presumably through a retro-aldol reaction to give 3, elimi-
nation of methyl vinyl ketone, and transesterification.
À
and copper-catalyzed diverted O H insertion reaction. Ratios of
À
tetrahydrofuran to O H insertion products (for higher-yielding reac-
tions) as determined by NMR analysis of initial products are given in
parentheses.
85% yield, thus showing that the stereoselectivity of the
reaction could be extended to the substituent at position 4 of
the cyclic ether ring. a-Disubstituted aldol 30 did not give the
desired tetrahydrofuran 41 under copper(I) catalysis, while
single diastereoisomer 41 was obtained in 48% yield under
rhodium catalysis. When allylic alcohol 31 was subjected to
rhodium catalysis, the desired tetrahydrofuran 42 with an
exocyclic double bond was obtained in 67% yield.
Secondary and tertiary alcohols also underwent the
À
diverted O H insertion/cyclization reaction, and in these
cases, the rhodium catalyst was superior. Alcohol 32 gave
a single diastereoisomer of tetrahydrofuran 43 in good yield,
showing that the stereocontrol could be extended to the
position 5 of the tetrahydrofuran product. When enantioen-
riched (R)-aldol 32 (77% ee) was used, no erosion of
stereochemistry was observed, and the product tetrahydro-
Angew. Chem. Int. Ed. 2015, 54, 8485 –8489
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim