Organic Letters
Letter
loading was decreased to 20 mol %, the yield of the target
product 2b declined to 70% with the ee retained (Table 1,
entry 26); when the catalyst loading was decreased to 10 mol
%
, the yield dropped to 46%, and the 3-butynoate was not
consumed completely even with a prolonged reaction time
Table 1, entry 27). Finally, the optimal conditions were
established for this reaction as 30 mol % P5 as the catalyst and
(
1
,2- dichlorobenzene as the solvent at 10 °C.
Having identified the optimal conditions, we next explored
the generality of the substrate scope of γ-aryl-3-alkynoates. As
shown in Table 2, it was found that different ester groups, such
as ethyl (1a), methyl (1b), and benzyl (1c) were all
compatible with this reaction, delivering the desired products
in 72−85% yield with up to 98% ee. Moreover, substrates with
either an electron-withdrawing or electron-donating group on
the phenyl ring underwent this phosphine-catalyzed tandem
reaction smoothly to furnish the desired products 2d−k and
modified phosphine P1 gave the best results, affording the
product in 90% yield with 90% ee (Table 1, entry 2),
compared with phosphines substituted with other functional
groups such as sulfonamide, thiourea, and pivalamide (Table 1,
entries 3−5). Moreover, 3,5-bis(trifluoromethyl)benzamide-
modified phosphines P5 and P6 derived from L-isoleucine and
L-threonine, respectively, were tested, and both worked well in
this process to give the product 2a in 80% and 69% yield,
respectively, with high enantioselectivity (Table 1, entries 6
and 7). Taking into account the synthesis efficiency of the
phosphine catalyst, we chose P1 and P5 as potential catalysts
to investigate the effect of the solvent and temperature and to
improve the yield and enantioselectivity in this novel
trimerization process (see the SI for details). The catalyst P5
derived from L-isoleucine performed better than P1 in each
issue of various parallel comparisons. Solvent screening
revealed that aromatic solvents gave the highest enantiose-
lectivity (up to 97% ee; Table 1, entries 16−21) and solvents
with moderate or weak polarity led to moderate to good
enantioselectivity (Table 1, entries 8−12) while strongly polar
solvents (DMF, acetonitrile, and acetone) restrained this
reaction (Table 1, entries 13−15). When methyl 4-phenylbut-
2
p in 68−84% yield with 95−99% ee. However, aromatic rings
containing a substituent at the ortho position gave lower yields
2n and 2o) because of the steric hindrance. The reaction was
(
also applicable to 2-naphthyl- and 3-thiophenyl-containing
substrates, providing the corresponding cycloadducts 2q and
2
r in good yields with extremely high enantioselectivity.
Unfortunately, when γ-butyl alkynoate 1s was employed in this
reaction, a complicated mixture was obtained, and HRMS
showed no trimerization product. No reaction took place when
γ-tert-butyl alkynoate 1t was involved. Notably, the stereo-
specificity of this reaction was evident, as only one stereo-
isomer of cyclopentene derivative 2 was detected in all of the
examples despite the presence of two continuous chiral centers
with an all-carbon quaternary stereocenter. The absolute
configuration of product 2a was determined by single-crystal
X-ray diffraction.
Further evaluation of the catalytic efficiency indicated that
this phosphine-catalyzed reaction could be easily enlarged to
the gram scale. 2b and 2o were obtained in 4.2 and 1.8 g,
respectively, under the standard conditions with the yield and
enantioselectivity maintained (Scheme 2a). Regioselective
epoxidation of 2b was realized in the presence of m-CPBA
3
-ynoate (1b) was used instead, the target product 2b was
isolated in 85% yield with 98% ee in 1,2- dichlorobenzene
catalyzed by P5 (Table 1, entry 22). Using a lower temperature
did not increase the enantioselectivity (Table 1, entries 22−
2
4), while using a higher temperature decreased the
14
enantioselectivity slightly to 94% ee, although the yield was
improved slightly to 88%. Furthermore, when the catalyst
to provide product 3 in 83% yield with 98% ee (Scheme 2b).
Interestingly, product 3 could also be obtained by ozonolysis in
C
Org. Lett. XXXX, XXX, XXX−XXX