Organic Letters
Letter
Figure 7. Products of the ATH of electron-rich 1-aryl and 1-methyl
DHIQs in this project using catalyst (R,R)-5 and the conditions
a
shown in Figure 5 (overnight reaction time). First report, to our
knowledge, of the formation by ATH using arene/Ru/TsDPEN
b
catalysts. Formate formed.
As expected on the basis of previous literature reports,1,6,7
dimethoxy-substituted 1-aryl imines were reduced with high ee
with (R,R)-5, slightly higher than similar reported examples
using catalyst (R,R)-1 (24: 84% ee, 27: 75% ee);7b however,
the difference was not as significant as that for the electron-
poor products in Figure 6. It should be noted that the imine
precursor of 26 was fully reduced; however, the product was
formed as a mixture of formylated (major) and nonformylated
(minor) amines. In addition, we compared catalyst (R,R)-5
with the reported results for the formation of the methyl-
substituted products 28 and 29, and the products were formed
with 81 and 80% ee, respectively; slightly lower ee values than
have already been reported using (R,R)-1 and (R,R)-2 as ATH
catalysts (Figure 1).1−3,6
It is not exactly clear how the modified catalyst (R,R)-5
controls the asymmetric reduction in these cases. However, the
control of the enantioselectivity of the reduction is believed to
involve a transition state in which the protonated iminium ion
forms a H bond to the SO2 of the tosyl group while a known
η6/CH interaction also operates to stabilize the transition state
(Figure 8A),3 which is analogous to the control of ketone
reductions with this class of catalyst.14 However, the selectivity
is likely to be low because the transitions state (ts) for the
reduction to either enantiomer can be stabilized by similar
interactions. The additional furan group (in (R,R)-5) may
engage in an interaction that serves to stabilize the ts, leading
to the observed major enantiomer (Figure 8B). The lack of
selectivity observed with catalyst (R,R)-4 suggests that this is
an electronic effect involving the heteroatom rather than a
simpler steric or π-stacking effect. Conversely, the additional
steric hindrance in (R,R)-5 results in slower reduction (and
hence incomplete conversions) for more hindered substrates,
that is, those containing ortho-substituted aryl groups.
Figure 6. Products from non-electron-rich DHIQ reduction obtained
in this project using catalyst (R,R)-5 and the conditions shown in
Figure 5 (overnight reaction time). Configurations were assigned by
analogy to 9 (Table 1). aFirst report, to our knowledge, of the
b
formation by ATH using arene/Ru/TsDPEN catalysts. 85% yield,
92% ee reported using (R,R)-1.7b
where comparable, with higher ee than that reported for
catalyst (R,R)-1 (15: 36% ee, 16: 36% ee, 18: 39% ee, 19: 79%
ee).7b Several of the products were reported, to the best of our
knowledge, for the first time with high ee using an arene/Ru/
TsDPEN catalyst in ATH (indicated in Figure 6). Tolerated
substituents included meta- and para-chloro and -methyl and
para-bromo, -iodo, -methoxy, -nitro, and -trifluoromethyl
groups (not all meta-substituted substrates were tested) as
well as meta/para combinations of electron-rich groups. The
synthesis of amine 15 was also carried out on a 1.1 g (5 mmol)
scale and gave a product with 91% ee in 71% isolated yield. In
contrast, the furan catalyst (R,R)-5 is less effective at the ATH
of ortho-substituted aryl-substituted substrates; the reduction
of the ortho-chlorophenyl substrate gave no product 12, and
the ortho-methyl/methyloxyphenyl imines gave products 17
and 19, respectively, in low yield, although with excellent ee.
Hence there is a clear complementary (and mutually exclusive)
pattern of selectivity between Noyori−Ikariya catalysts such as
(R,R)-1 and (R,R)-2 and the heterocycle-functionalized (R,R)-
5. This may reflect the extra steric hindrance around catalyst
(R,R)-5, which is less accommodating to a bulky 2-aryl
substituent; however. the formation of a racemic product using
catalyst (R,R)-4 in the prototype substrate test indicates the
importance of the additional involvement of the furan in the
reaction transition state.
In conclusion, we have demonstrated that the addition of a
heterocyclic group to the basic nitrogen atom of the TsDPEN
ligand in an arene/Ru/TsDPEN ATH complex renders it an
excellent catalyst for the reduction of a previously very
challenging class of DHIQ substrate for this application. The
value of the methodology is highlighted by the formation of
products 9 and 21, which are precursors of pharmaceutical
target molecules (Figure 5). The mode of action remains to be
fully understood, but the presence of a heterocycle is
The study was also extended to a series of electron-rich
substrates and the simpler 1-methyl substrates, with the
resulting THIQs (yields and ee’s) shown in Figure 7.
C
Org. Lett. XXXX, XXX, XXX−XXX