quinuclidine nitrogen increases the observed enantioselectiv-
ity19 by inducing a transition to predominantly open 3, as
the water reaches its limit of miscibility. This result is contrary
to that expected if consideration is given only to the dielectric
constant of water. Previous work has shown that the open 3
configuration of cinchonidine is stabilised by solvents of low
dielectric constant. In this study, an inverse relationship
between enantioselectivity and dielectric constant is clearly
seen with the exception of acetic acid (Fig. 3). Hydrogenation
of ethyl pyruvate in pure water (e: 78.5) yields the expected low
enantioselectivity of 43%. Therefore the addition of water to
CD2Cl2 , a solvent of low dielectric constant (relative permit-
tivity) (e: 8.9) would logically be expected to decrease the sta-
bility of the open 3 conformer. However Burgi and Baiker26
showed by NMR methods that cinchonidine in ethanol had
a much greater proportion of the open 3 conformer present
than was predicted by their ab initio calculations and its dielec-
tric constant (e: 24.3). It was shown by NMR spectroscopy
that the proportion of cinchonidine present in the open 3 con-
figuration was 77%, whereas theoretical calculations based on
dielectric constants predicted a value of not greater than 46%,
the exact value depending on the specific basis set used for the
calculation. It was concluded that this was due to ethanol
hydrogen bonding to the quinuclidine nitrogen and thereby
stabilising the open 3 conformer. In the open configurations
access to the quinuclidine nitrogen was shown to be unhin-
dered, whereas in the closed configuration, such interactions
are hindered as a result of repulsions with the quinoline moi-
ety. Polarised water molecules, present in low concentration
in an apolar solvent would also be expected to coordinate to
an electron rich base such as the quinuclidine nitrogen. The
degree of stabilisation would not be expected to be as high
as that observed for ethanol due to the smaller size of the water
molecule, hence the overall enhancement of open 3 stability is
less, although still significant. Coordination of water to the
alkaloid modifier is also consistent with the decrease in enan-
tioselectivity observed when as-received solvents are used in
this reaction, and provides further evidence for the specific
1 : 1 interaction of the substrate and modifier in a successful
enantioselective reaction. With water ‘‘bracing’’ the modifier
into the open 3 position and, despite this conformation being
preferential to high enantioselectivity, reduced enantioselectiv-
ities are observed as the presence of the water also has an effect
of blocking the adsorption of the substrate and its eventual
coordination to the alkaloid. A slightly enhanced amount of
reactant molecules may therefore react preferentially on the
unmodified Pt sites, known to be in great excess,29 resulting
in a lower than expected enantioselectivity. Pretreatment of
the modifier with ethyl pyruvate reduces this effect by acting
as an ‘‘in situ drying agent’’, removing the water from the
system, negating the blocking effect of the water, but also act-
ing as an agent for further enhancement of the open 3 config-
uration by forming pyruvic acid, which even in small
amounts results in a large increase in the proportion of open
3 alkaloid in solution, overcoming any negative effect due to
site blocking.
3
revealed by the large decrease in the J(H8a–H9) coupling con-
stant and the loss of any cross-peak signal between H1 and
H8a in the NOESY spectra.
Understanding this change in the conformational distribu-
tion of the cinchona alkaloid is crucial to the interpretation
of enantioselectivity since, in all previous conformational stu-
dies it has been assumed that the conformation of the modifier
is purely a function of the solvent in which the modifier is dis-
solved, and is unaffected by the presence of the substrate.
These results show that it is insufficient to consider modifier
conformation in solution alone, and that allowance must be
made for the effects of all the components in the reaction sys-
tem. It is also important to note that the conformational
change is not a result of a direct interaction of the modifier
and reactant in solution as proposed by Margitfalvi and co-
workers,17,18 it is a reaction of the pyruvate ester with dis-
solved water, forming pyruvic acid that in turn interacts with
the cinchonidine causing its conformational distribution to
move in favour of the open 3 conformer; the resulting enan-
tioselectivities are therefore higher as a consequence of the
greater amount of the favoured conformation of the modifier.
However, if the increase in the open 3 configuration of the
alkaloid was solely a product of pyruvic acid produced by
hydrolysis during the pretreatment process, it would be
expected that the initial enantioselectivity observed in dried
solvents would be of a comparable degree to that observed
with as-received solvents in which no pretreatment had been
conducted. The results in Table 2 indicate that this is not the
case. Experiments in all dried solvents gave higher enantios-
electivities than the corresponding values in as-received sol-
vents. This cannot be attributed to the continued presence of
some residual water due to incomplete drying as no pretreat-
ment stage has been used. Despite this, all samples that have
been subjected to the pretreatment stage, i.e. experiments
undertaken in both dried and as-received solvents, resulted
in broadly similar enantioselectivities.
In order to resolve this question, further NMR spectro-
scopic analysis was performed to investigate the conformation
of cinchonidine in solutions containing small quantities of
water. Fig. 8 shows the proportion of cinchonidine present
in the open 3 configuration when dissolved in CD2Cl2 with
progressive addition of water. Cinchonidine (2 mg) was dis-
solved in ultra-dry CD2Cl2 (0.5 ml) with progressive addition
of 0.5 ml aliquots of H2O until saturation was reached. The
coupling constant of the H8a–H9 proton signal was determined
after each water addition.
Clearly, the proportion of the open 3 conformer increases
with water content from 50 to 56% at which concentration
Conclusions
1H NMR spectroscopic experiments, in combination with cat-
alysis studies, have shown that the presence of small amounts
of water in reaction media have a marked effect on the confor-
mation of the cinchona alkaloid modifier and thereby the
resulting enantioselectivity for the hydrogenation of ethyl pyr-
uvate. This is shown to be a result of the base catalysed hydro-
genation of the pyruvate ester resulting in the formation of
pyruvic acid which itself stabilises the open 3 conformation
of cinchona alkaloids by coordination to the quinuclidine
nitrogen. A further unexpected role of water in solution is
that when present in small quantities in apolar solvents it
can stabilise the open 3 conformer by hydrogen bonding to
Fig. 8 The effect on the coformational distribution of cinchonidine of
the successive addition of water to a solution of cinchonidine in
CD2Cl2 .
2844
Phys. Chem. Chem. Phys., 2002, 4, 2839–2845