1326
J. Saavedra et al. / Tetrahedron Letters 50 (2009) 1324–1327
provides a steric presence that enforces the relative placement of
the large (RL) and small (RS) substituents of the prochiral ketone
in the transition structure (Fig. 3).
KIEs of approximately 5.08–9.06 using a mass conversion based so-
lely on the square root of the masses transferred. Adiabatic trans-
fers are the result of strong electronic mixing of the hydride donor
and acceptor states, typically mediated by hydrogen bonding inter-
actions, electrostatic interactions, or interactions between filled
and unfilled orbitals. The bonding changes that occur during the
rate-determining hydride transfer step involve changes in the lat-
ter two types of interactions during the hydride transfer. The hy-
dride transfer step results in the reduction of the number of
centers bearing formal charge. Analogous to the simultaneous acti-
vation of both nucleophile (BH3) and electrophile (ketone) is the
simultaneous general acid activation of the electrophile (alde-
hydes) and conformational activation of the nucleophile (NADH)
in alcohol dehydrogenases, which essentially weakens the labile
C–H bond upon NADH. The intrinsic primary isotope effect for
the horse liver alcohol dehydrogenase-catalyzed reduction of
benzaldehyde is generally assumed to be around kH/kD = 3.5; how-
ever, measured values and values determined from simulated ki-
netic traces have been found to be 1.9 and 2.8, respectively.26
Unfortunately, apt comparisons with the uncatalyzed reaction
are lacking.
Several observations suggest that the above reasoning is per-
haps too simplistic. Numerous analogs to the (S)-Me–CBS (1) cata-
lyst exist. In testing these analogs, the most often performed
reduction is that of acetophenone to 1-phenylethanol. The two pre-
dominant variations upon 1 are alterations in the geminal substitu-
tion at the carbinol center and the boron substituent. In the
reduction of acetophenone, the boron substituent seems to play lit-
tle role in stereoselection. With H, Me, Et, and n-Bu as substituents
giving nearly quantitative yields and enantiomeric excesses of
97%3, 96.5%4, 96%22, and 96%22, respectively. There does, however,
appear to be a marked effect of the carbinol substitution pattern
upon enantioselectivity, with H, Ph, b-Np, a
-Np, 20-methylphenyl,
20-methoxyphenyl, and thiophenyl groups yielding enantiomeric
excesses of 77%, 97%, 98%, 62%, 76%, 28%, and 82%.5,23 While it is
likely that the simple unsubstituted carbinol lacks the conforma-
tional rigidity necessary for stereocontrol, something more com-
plex is at work in the cases of the
a
-Np, 20-methylphenyl, 20-
methoxyphenyl, and thiophenyl carbinol substitution. The first
three substituents possibly interfere with coordination of the ke-
tone to the catalyst as it enters the catalytic cycle. A potential par-
titioning of the substrate between uncatalyzed and catalyzed
pathways may be responsible for low enantioselectivities in these
cases. While thiophenyl is perhaps slightly less obtrusive than phe-
nyl, this small change in steric presence may be only a secondary
reason for decreased selectivity. More likely is that the thiophenyl
group is capable of coordinating the reductant and may essentially
reduce the concentration of active catalyst-reductant complex in
the active catalytic cycle.
In conclusion, we have measured 13C KIEs for the CBS reduction
of 20,50-dimethylphenyl isopropyl ketone and computed a model
transition structure that yields KIEs that are firmly in agreement
with the experimental measurements. We have also presented an
isotope effect methodology by which 13C KIEs may be measured
for each individual enantiotopic group in reactions where symme-
try breaking makes the groups inequivalent. Finally, the results of
our experimental and computational work have suggested new po-
tential strategies for controlling selectivity in asymmetric catalytic
reactions and have highlighted the similarities and differences be-
tween enzyme and small molecule catalysis.
Part of the role of the carbinol substituents is perhaps hinted at
13
in the C KIE at the 50-methyl position. While all other KIEs are
consistently within one standard error of the computed values,
the 50-methyl group displays a consistent, if marginally significant,
deviation from the predicted 13C KIEs. If this difference is an accu-
rate reflection of what is occurring at the transition state, then it
may hint at an unexpected origin of stereoselectivity in the CBS
system. Computational models of the transition state tend to steer
thinking away from the more correct ensemble picture of the tran-
sition structure. It is possible that torsional motion between the
aryl ring and carbonyl carbon places the 50-methyl group in near
incidence with one of the geminal phenyl groups on the catalyst.
A mixed torsional mode at 23 cmꢁ1 involves significant rotation
of the substrate aryl group in the computed transition structure
for preferred Si attack, signaling that the ensemble of transition
structures germane to a free energy description of the transition
state may contain members with some steric interaction between
the 50-methyl group and the axial phenyl group in the catalyst.
In the simplest terms, the origins of reactivity in the CBS reduc-
tion are well understood: The endocyclic boron serves as a Lewis
acid activator of the electrophile, while the endocyclic nitrogen lo-
cal to the adjacent ring activates the nucleophile (reductant) via
donation of its lone pair. While this thermodynamic argument is
satisfying, further inspection suggests that the CBS catalyst may
employ strategies borrowed from alcohol dehydrogenases to lower
the reaction barrier for hydride transfer. As stated before, the imag-
inary frequency corresponding to reaction coordinate motion at
the transition state is lower than one might anticipate for hydride
transfer (262i cmꢁ1). Furthermore, the diminution of this fre-
quency does not result from a large reduced mass but rather from
a small negative force constant. Such a situation infers that the
transfer is rather adiabatic. The small primary kinetic isotope effect
(kH/kD = 1.7)24 is also indicative of a largely adiabatic transfer. This
value is in contrast to those observed for the borane reduction of
ketones: (kH/kT = 6.22–11.1),25 which correspond to deuterium
Acknowledgments
We thank the ACS Petroleum Research Fund (#48064-G4) for
support of this research and Mike Colvin (UC Merced) for the use
of his Linux Cluster.
Supplementary data
Supplementary data (experimental procedures, NMR integra-
tion results, computational procedures, and energies and geome-
tries of all calculated structures) associated with this article can
References and notes
1. Itsuno, S.; Ito, K.; Hirao, A.; Nakahama, S. J. Chem. Soc., Chem. Commun. 1983,
469–470.
2. Itsuno, S.; Ito, K. J. Org. Chem. 1984, 49, 555–557.
3. Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am. Chem. Soc. 1987, 109, 5551–
5553.
4. Corey, E. J.; Bakshi, R. K.; Shibata, S.; Chen, C.-P.; Singh, V. K. J. Am. Chem. Soc.
1987, 109, 7925–7926.
5. Corey, E. J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1986–2012.
6. Farina, V.; Reeves, J. T.; Senanayake, C. H.; Song, J. J. Chem. Rev. 2006, 106, 2734–
2793.
7. Corey, E. J.; Shibata, S.; Bakshi, R. K. J. Org. Chem. 1988, 53, 2861–2863.
8. Corey, E. J. Pure Appl. Chem. 1990, 62, 1209–1216.
9. Jones, D. K.; Liotta, D. C. J. Org. Chem. 1993, 58, 799–801.
10. Mathre, D. J.; Thompson, A. S.; Douglas, A. W.; Hoogsteen, K.; Carroll, J. D.;
Corley, E. G.; Grabowski, E. J. J. J. Org. Chem. 1993, 58, 2880–2888.
11. Linney, L. P.; Self, C. R.; Williams, I. H. J. Chem. Soc., Chem. Commun. 1994, 1651–
1652.
12. Jockel, H.; Schmidt, R.; Jope, H.; Schmalz, H.-G. J. Chem. Soc., Perkin Trans. 2
2000, 69–76.
13. Zhu, H.; Clemente, F. R.; Houk, K. N.; Meyer, M. P. J. Am. Chem. Soc., in press.
14. Thomas, A. A.; Singleton, D. A. J. Am. Chem. Soc. 1995, 117, 9357–9358.
15. Keith, T. A.; Bader, R. F. W. Chem. Phys. Lett. 1992, 194, 1–8.
16. Keith, T. A.; Bader, R. F. W. Chem. Phys. Lett. 1993, 210, 223–231.