V. Gotor et al.
The higher reactivity of ketones bearing EWGs has tradi-
tionally been attributed to a higher electrophilicity of the
carbonyl carbon because of the inductive effect of these sub-
stituents. However, the initial rates of reaction do not
always correlate with the conversions and the normalized
IR frequencies of the carbonyl groups not only suggest the
opposite but show a good correlation with the experimental
conversions, thus affording a straightforward and easy pre-
dictive method for determining the position of the equilibri-
um that will be attained for a given ketone. Therefore, it
seems that the extent of the reduction of these ketones is
not under kinetic control and, thus, it should be under ther-
modynamic control. This is in agreement with the good cor-
relation observed for the conversions and the calculated
Gibbs free energy ketone/alcohol differences against 2-prop-
anol/acetone. This behavior has been explained on the basis
that the presence of EWGs would destabilize the ground
resonance state of the ketones by decreasing the contribu-
tion of the dipolar resonant form. This hypothesis explains
the general higher reactivity of these substrates and suggests
that the equilibrium constants of these reactions are mainly
driven by the different energies of the ketones. Moreover,
the higher contribution of the resonant double-bond form
leads to higher frequencies in the absorption band, which
was further confirmed by the calculated values of different
suitable descriptors of the C=O bond.
HT protocol by using Al
(
(
of a 0.1m solution in dry toluene) were added under N
(
A
H
U
G
R
N
U
3 3
: In a sealed Schlenk tube, Al ACHTUNGTRENNUNG( tBuO)
2
2
2
with one drop of water, and Na
mined by GC analysis.
2 4
SO was added. Conversions were deter-
General protein–ligand computational methods: All molecular mechanics
computations were performed by using the molecular modeling package
Molecular Operating Environment (MOE) 2007.09 (Chemical Comput-
[
27]
[28]
ing Group). In all cases the Amber99 force field, the corresponding
dictionary charges as implemented in MOE, and the Generalized Born
[
29]
solvation model with relative dielectric constant values of 2 and 80 for
[
30]
the protein matrix and the solvent, respectively, were selected. A non-
bonded cut-off of 8 ꢅ with a smoothing function of between 8 and 10 ꢅ
were used. In all molecular dynamics simulations, the NVT ensemble and
the Nosꢂ–Poincarꢂ–Anderson equations were selected. The initial and
simulation temperatures were set to 0 and 300 K, respectively, a tempera-
ture relaxation time of 10 fs was selected, and the length of the heating
and simulation periods was 1 ps. No constraints were imposed on any
bond and a step size of 1 fs was used. The convergence criterion of the
energy minimizations (EM) was set to a gradient value of 0.01 or
ꢀ
1
0.001 kcalmol if only parts of the protein or the whole enzyme–sub-
strate–cofactor complex, respectively, were minimized. The rest of pa-
rameters were set to their default values. High-quality pictures of repre-
[31]
sentative structures were generated with PyMOL 0.99.
Ab initio calculations: Gibbs free energies of reactants and products cor-
responding to the reaction described in Table 1 were calculated according
to the following Equation (1):
D Gꢀ aq ¼ D Gꢀ aq þ Eꢀ MP2=CBSþD Gꢀ solv
HF
COSMOS
Once the lack of reactivity of halohydrins was clarified, it
became evident that the oxidation of halohydrins and relat-
ed alcohols should be feasible by using another ketone bear-
ing an EWG as a hydrogen acceptor. Such reactions did
indeed proceed to a relatively high extent but they did not
reach equilibrium in the time frames used. Therefore, in the
case of alcohols bearing EWGs, the oxidation is slowed
down as compared with other hydrogen donors, and longer
reaction times or higher temperatures are required to reach
the equilibrium concentrations.
ꢀTSconf
ð1Þ
ꢀ
HF
aq
in which DG represents the thermal contributions (translational, rota-
tional and vibrational) to Gibbs free energy, E¯ MP2/CBS is the MP2 electron-
ꢀ
COSMOS
solv
ic energy evaluated at the complete basis set (CBS) limit, DG
is
the solvation free energy and Sconf is the conformational entropy. In this
expression, the energy components (except Sconf) are averaged over all
the existing conformers of each particular structure according to a Max-
well–Boltzmann distribution.
For the isodesmic transformations, energies were averaged over the most
stable conformers (Nconf) of each particular structure according to a Max-
well–Boltzmann distribution. For the sake of completeness, relative free
¯
energies of reaction were calculated in solution (DGaq) and in the gas
¯
min
phase (DGgas). Relative free energies (DGaq ) evaluated over the most
stable conformers in solution are also given. Further details of the ab
initio calculations are provided in the Supporting Information.
Experimental Section
General experimental methods: Alcohol dehydrogenases and ketones
1
2
a–i were purchased from commercial sources. Racemic alcohols (ꢁ)-
a–i were either synthesized by conventional reduction of the corre-
sponding ketones (NaBH
sources. All other reagents, including catalysts and solvents, were of the
highest quality available. Flash chromatography was performed by using
4
, MeOH, RT) or purchased from commercial
Acknowledgements
F.R.B. is supported by Programme Alban, the European Union Program
of High Level Scholarships for Latin America (scholarship no.
E07D402519AR). I.L. thanks Principado de Asturias for personal fund-
ing (Clarꢀn Program). Financial support from the Spanish Ministerio de
Ciencia e Innovaciꢆn (MICINN; project CTQ2007-61126/PPQ) is grate-
fully acknowledged. Prof. F. Lꢆpez Ortiz is gratefully acknowledged for
helpful discussions.
1
13
silica gel 60 (230–400 mesh). H and C NMR and DEPT spectra were
1
13
obtained by using
a DPX-300 spectrometer ( H: 300.13 MHz, C:
7
5.5 MHz) for routine experiments. Gas chromatography (GC) analyses
were performed by using a standard gas chromatograph with nitrogen as
the carrier. UV and IR spectra were performed by using standard UV/
Vis and IR spectrophotometers, respectively.
Enzymatic HT protocol by using LBADH: Generally, LBADH (3 U)
was added to Tris-HCl buffer (600 mL, 50 mm, pH 7.5; 1 mm MgCl
NADPH) in an Eppendorf vial (1.5 mL). Then, the corresponding ketone
25 mm) and 2-propanol (50 mm, 1.8 mL, 2 equiv) were added to the mix-
ture. Reactions were shaken at 308C and 150 rpm for 24 h. The reactions
were stopped by extraction with ethyl acetate (2ꢄ0.6 mL). The organic
layer was separated by centrifugation (1.5 min, 13000 rpm) and dried
2
, 1 mm
[2] K. G. Denbigh, The Principles of Chemical Equilibrium, Cambridge
(
over Na
2
SO
4
. Conversions were determined by GC analysis.
11018
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11012 – 11019