Stereoselection in the corey-bakshi-shibata reduction: Insight from kinetic isotope effects and transition-structure modeling

Article abstract of DOI:10.1002/anie.201206011

Flexible but demanding: Significant steric demand is isolated to the small (iPr) substituent in the oxazaborolidine-catalyzed borane (CBS) reduction of 2',5'-dimethylisobutyrophenone. Computed transition-structure models (see picture) demonstrate that nea

Full text of DOI:10.1002/anie.201206011

.
Angewandte
Communications
DOI: 10.1002/anie.201206011
Stereoselection
Stereoselection in the Corey–Bakshi–Shibata Reduction: Insight from
Kinetic Isotope Effects and Transition-Structure Modeling**
Hui Zhu, Daniel J. OꢀLeary, and Matthew P. Meyer*
The Corey–Bakshi–Shibata (CBS) reduction is one of the
most versatile methods for the enantioselective reduction of
prochiral ketones, exhibiting both high selectivity and a large
substrate range.[1] Enantiofacial discrimination depends upon
the effective discrimination of the large (RL) and small (RS)
substituents (Scheme 1). Steric repulsion is accepted as a key
pro-S, 2’-, and 5’-methyl positions within 1 by using optimized
reactant and transition structures to compare with previously
measured values.[5,6] Two distinct conformeric transition
structures (chair-like and boat-like) were optimized for the
CBS reduction of 1 (Figure 1). The more stable conformer
(2.39 kcalmolꢀ1) exhibits a chair-like configuration of the six
Scheme 1. CBS reduction of 2’,5’-dimethylisobutyrophenone. DMS=di-
methyl sulfide.
Figure 1. Optimized [B3LYP/6-31+G(d,p)] structures for A) chair-like
and B) boat-like transition structures. Cyclic arrangements of reactive
atoms are shown as color-coded spheres (N: blue, O: red, B: pink, C:
gray, and H: white).
director of stereoselection; however, direct measures of steric
interaction at the transition state (TS) are difficult to obtain.[2]
Steric kinetic isotope effects (KIEs) offer a means of
investigating steric interactions quantitatively.[3] Herein, we
Table 1: Comparison of experimentally determined KIEs and those
computed from optimized reactant and transition structures.
[a]
pro-R[b,c] pro-S[b,c] 2’-Me[b]
5’-Me[b]
=
C O
leverage
a combined experimental and computational
Experiment
1.025(1) 0.965(3) 0.974(3) 0.980(5) 0.995(5)
approach toward answering the following questions regarding
stereoselection in the CBS reduction of 2’,5’-dimethylisobu-
tyrophenone (1) using the oxazaborolidine (2) as a catalyst.
1) Is it reasonable to assume that steric repulsion develops
only at the small carbonyl substituent (RS) during enantiofa-
cial discrimination? 2) How can steric KIEs be distinguished
from KIEs arising from other factors? 3) How severe are
conformational constraints in reactions that take place via
cyclic transition structures?
Chair-like TS[d,f] 1.023
Chair-like TS[e,f] 1.029
0.944
0.951
0.938
0.947
0.972
0.966
0.962
0.970
0.973
0.974
0.978
0.981
0.993
0.998
0.999
1.000
Boat-like TS[d,f]
Boat-like TS[e,f]
1.024
1.030
[a] 13C KIE; Ref. [6]. [b] 2H KIE (CH3/CD3). [c] Ref. [5]. [d] Computed at
B3LYP/6-31+G(d,p). [e] Computed at B3LYP/6-31+G(d,p) with
IEFPCM model for THF solvent. [f] Computed using the Bigeleisen
equation with scaled frequencies and a Bell tunnel correction.
Using a modification of a previously reported method (see
“active” atoms. Computed KIEs that result from both
transition structures are reported alongside experimentally
determined KIEs (Table 1). Three results are immediately
apparent from comparisons of the experimental and compu-
2
Experimental Section), we measured the H KIEs at the 2’-
and 5’-methyl positions within 1.[4] We also computed the 13
KIE at the carbonyl position and the H KIEs at the pro-R,
C
2
2
tational data. First, the H KIEs measured at the prochiral
methyl groups and the 2’-methyl position are substantial and
inverse (kH/kD < 1). Second, computed KIEs, with the possi-
ble exception of the effect observed at the pro-R methyl
group, are in excellent accord with measured KIEs. Third,
KIEs computed from transition structures with chair-like and
boat-like arrangements of the six “active” atoms do not differ
significantly. Explaining the physical origins of each of these
observations will provide some insight into the interpretation
of 2H KIE measurements in stereoselective reactions and
provide a unified conceptual understanding of the inherent
properties of oxazaborolidine-catalyzed reductions.
[*] Prof. M. P. Meyer
Department of Chemistry, University of California, Merced
Merced, CA 95343 (USA)
E-mail: mmeyer@ucmerced.edu
Prof. D. J. O’Leary
Department of Chemistry, Pomona College
Claremont, CA 91711 (USA)
Dr. H. Zhu
Department of Chemistry, University of California, Berkeley
Berkeley, CA 94720 (USA)
[**] M.P.M. acknowledges support from the NSF (CHE-1058483 and
CHE-0840505). D.J.O. thanks Pomona College for supporting this
research.
It seems reasonable to expect that methyl positions giving
significant inverse 2H KIEs as a result of steric repulsion
would display greater rotational barriers in computed tran-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201206011.
11890
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 11890 –11893

Angewandte
Chemie
sition structures than in computed reactant structures. In
a series of computed relaxed rotor scans, the rotational
barriers for the pro-R and pro-S methyl groups are indeed
higher in the transition structure than in the reactant (Fig-
ure 2A&B, respectively). By contrast, the rotational barrier
decreases for the 2’-methyl position in going from the
recent report, OꢀLeary et al. used an enthalpy/entropy
decomposition of the computed 2H KIE for the stereo-
inversion of two axially chiral molecules to demonstrate that
2H KIEs that arise from the development of significant
repulsive interactions at the transition state give significant
normal (kH/kD > 1) entropic contributions and inverse (kH/
kD < 1) enthalpic contributions to the KIE.[10] Analogous with
the findings from Dunitz and Ibberson, OꢀLeary et al. found
that entropic contributions to the KIE can outweigh enthalpic
contributions, leading to an overall normal 2H KIE. Enthalpy/
2
entropy analyses of computed H KIEs offer another view
into the origins of the isotope effects measured here. In
accordance with methyl rotational barriers, one might antici-
pate that enthalpy/entropy decompositions of the CH3/CD3
KIEs at the pro-R and pro-S positions give both significant
normal entropic contributions and significant inverse
enthalpic contributions. Results reported in Table 2 demon-
Table 2: Computational decompositions of 2H KIEs into their enthalpic/
entropic contributions.
pro-R[a]
pro-S[a]
2’-Me[a]
5’-Me[a]
Figure 2. Methyl rotation profiles for reactant (black), chair-like transi-
tion structure (red), and boat-like transition structures (blue) corre-
sponding to the A) pro-R, B) pro-S, C) 2’-, and D) 5’-positions.
Experiment
0.965(3)
0.974(3)
0.980(5)
0.995(5)
Chair-like transition structure
DDG°
0.955
0.925
1.032
0.971
0.946
1.026
0.974
0.973
1.001
0.998
0.995
1.002
DDH°
ꢀTDDS°
optimized reactant structure to the transition structure (Fig-
ure 2C). No significant potential energy barrier for rotation
appears to exist in either the reactant or transition structure
for the 5’-methyl group (Figure 2D). This result might be
expected, given the small barriers associated with methyl
rotation in toluene ( ꢁ 14 calmolꢀ1).[7] Intuitively it seems
reasonable to ascribe the origins of the 2H KIEs at the
prochiral methyl groups to steric origins, considering that the
prochiral methyl groups are conjugatively isolated from the
center of the reaction, that is, hyperconjugation effects seem
unlikely. The scans for methyl rotation further suggest that, if
anything, there should be a normal (kH/kD > 1.0) steric
contribution to the 2H KIE at the 2’-methyl position because
steric repulsion appears to be relieved in the transition state
for this position. Therefore, it is most reasonable to expect
that the g-2H KIE at the 2’-methyl position arises largely from
a reduction in hyperconjugation in going from the reactant to
the transition state. This notion is supported by the observa-
Boat-like transition structure
DDG°
0.950
0.920
1.032
0.974
0.957
1.018
0.978
0.978
1.000
1.000
1.000
1.000
DDH°
ꢀTDDS°
[a] 2H KIE (CD3/CH3). [b] Computed at B3LYP/6-31+G(d,p) with an
IEFPCM model for THF solvent, unscaled frequencies, and no tunnel
correction.
strate that this is indeed the case. Compared to the pro-S
methyl group, the pro-R methyl group exhibits larger
(inverse) enthalpic and smaller (normal) entropic contribu-
tions to the KIE. Steric KIEs at the prochiral methyl positions
offer a contrast to the KIE at the 2’-methyl position, which
appears to be largely dominated by an enthalpic contribution.
These two findings suggest a means of assigning the origins of
2
the substantial inverse H KIEs at the pro-R, pro-S, and 2’-
ꢀ
tion that C H bond lengths within the 2’-methyl rotor are
longer in the optimized chair-like and boat-like transition
structures than in the optimized reactant structure.
methyl groups shown in Table 1. In total, these results suggest
that steric 2H KIEs may have enthalpy/entropy contributions
that differ from KIEs arising from stereoelectronic effects
(e.g., hyperconjugation).
The conventional view of steric isotope effects holds that
ꢀ
ꢀ
the C D bond is effectively shorter than the C H bond,
resulting in inverse 2H KIEs for reactions in which steric
repulsion increases at the transition state.[3a,8] A recent report
by Dunitz and Ibberson challenged this view by demonstrat-
ing that the unit cell of C6D6 is smaller than that of C6H6 below
170 K but larger than that of C6H6 above 170 K.[9] They
further posited that the temperature dependence of the
relative size of these isotopologs of benzene was likely due to
low frequency vibrational modes, the contributions of which
become more important at higher temperatures. In another
What is perhaps most striking about the information in
Tables 1 and 2 and Figures 1 and 2 is that chair-like and boat-
like transition structures appear quite similar in all respects.
The structural similarities between these two transition
structures are evident in a maximal overlap superposition
(Figure 3). Though the boat-like transition structure appears
to place the isopropyl substituent (RS) of 1 in a more axially
confined environment than in the chair-like transition struc-
ture, the borane moiety is removed to a position that is
approximately 0.3 ꢁ more distant from the pro-R methyl
Angew. Chem. Int. Ed. 2012, 51, 11890 –11893
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11891

.
Angewandte
Communications
5’-methyl positions. The first competition reaction, which
used [2’-D3]-1 and [5’-D3]-1 as competing substrates, gives the
relative rates for the reduction of these isotopomers, KIER
[Eq. (1)]. Six measurements of the relative rates for the
ꢀ
k20ꢀD3
kH6
kH6
KIER
¼
¼
ð1Þ
k50ꢀD3 k50ꢀD3 k20ꢀD3
reduction of [2’-D3]-1 and [5’-D3]-1 were performed. The
second competition reaction employed 1 and [D6]-1 as com-
peting substrates. The ratio of rates for the reduction of 1 and
[D6]-1 gave the product of the KIEs at the 2’- and 5’-positions,
KIEP [Eq. (2)]. Four measurements of the relative rates for
Figure 3. Overlay of computed [B3LYP/6-31+G(d,p)] chair-like (red)
and boat-like (blue) transition structures.
kH6
kH6
kH6
ð2Þ
KIEP
¼
¼
ꢂ
kD6 k50ꢀD3 k20ꢀD3
group. Given the rotational barriers for the pro-R methyl
group and the computed KIEs, it appears that this steric
trade-off is nearly equivalent. Taken as a whole, the physical
and structural similarities between the chair-like and boat-
like transition structures suggest a flexible means of enforcing
stereoselection. The chair-like transition structure is com-
puted to be 2.39 kcalmolꢀ1 lower in free energy than the boat-
like transition structure. While this value cannot be consid-
ered quantitatively reliable, it points to the possibility that 2
can potentially enforce stereoselection through a conforma-
tionally diverse set of transition structures, depending upon
the steric demands presented by the substrate. While this
viewpoint is at variance with the conclusions of previous
computational studies of the CBS reaction, it is consonant
with the extensive substrate range observed for reductions
catalyzed by 2.[11]
In summary, we have found that the transition state for the
highly enantioselective borane reduction of 2 catalyzed by
1 involves steric interactions only at the smaller (isopropyl)
substituent. We have also illustrated two computational
methods by which steric 2H KIEs may be differentiated
from those arising from other physical origins. KIEs that are
measured at the methyl groups and arise from steric
interactions exhibit rotational barriers that increase at the
transition state. Steric KIEs also appear to exhibit substantial
normal contributions as a result of entropy, while KIEs that
arise from hyperconjugation do not have large entropic
contributions. The result of our experimental and computa-
tional work is a conceptual understanding of the origins of the
large substrate range exhibited by 1 in enantioselective
reductions. The key feature of our conceptual model is the
notion that, although the CBS reduction proceeds via a cyclic
transition structure, which is generally assumed to enforce
rigidity, stereoselection can likely be mediated through
a flexible transition structure that conforms to the steric
requirements of the substrate.
the reduction of 1 and [D6]-1 were performed. The rule of the
geometric mean and simple algebra gave the 2H KIEs at both
the 2’- and 5’-methyl positions [Eqs. (3) and (4), respec-
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
kH6
k20ꢀD3
ð3Þ
ð4Þ
¼
¼
KIEP=KIER
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
kH6
k50ꢀD3
KIEP ꢂ KIER
tively].[12] The synthesis of a mixture of [2’-D3]-1 and [5’-D3]-
1 was achieved via Friedel–Crafts acylation of [D3]-p-
xylene.[13] Synthesis of [D3]-p-xylene was achieved using
Suzuki coupling of p-tolylboronic acid and CD3I.[14] Synthesis
of [D6]-1 was accomplished by employing the Friedel–Crafts
acylation of [D6]-xylene. Gaussian09 was used to optimize
reactant and transition structures and to compute vibrational
frequencies resulting from optimized structures.[15] The bicy-
clic nature of the oxazaborolidine catalyst limits accessible
conformational space to rotameric conformers involving
ꢀ
rotations around the C Ar bond in the transition structures.
This conformational space was explored, and the reported
conformers represent the only accessible saddle point struc-
tures for Si attack. Methyl rotation barriers were computed
using a relaxed scan of one of the three dihedrals for which
ꢀ
the C CH3 bond was the central bond. Enthalpic and entropic
2H KIEs were computed using thermochemistry outputs from
Gaussian09.
Received: July 26, 2012
Published online: October 16, 2012
Keywords: asymmetric catalysis · isotope effects ·
.
reaction mechanisms · stereoselection · steric repulsion
[1] a) E. J. Corey, R. K. Bakshi, S. Shibata, J. Am. Chem. Soc. 1987,
109, 5551 – 5553; b) E. J. Corey, R. K. Bakshi, S. Shibata, C.-P.
Chen, V. K. Singh, J. Am. Chem. Soc. 1987, 109, 7925 – 7926;
c) E. J. Corey, S. Shibata, R. K. Bakshi, J. Org. Chem. 1988, 53,
2861 – 2863; d) E. J. Corey, Pure Appl. Chem. 1990, 62, 1209 –
1216; e) E. J. Corey, C. J. Helal, Angew. Chem. 1998, 110, 2092 –
2118; Angew. Chem. Int. Ed. 1998, 37, 1986 – 2012.
Methods Section
2
Experimental measurements of H KIEs at the 2’- and 5’-
methyl positions on 1 were achieved using a variation of
a previously published methodology.[4] Two competition
reactions were used, in conjunction with the rule of the
geometric mean, to extract individual 2H KIEs at the 2’- and
[2] T. Giagou, M. P. Meyer, Chem. Eur. J. 2010, 16, 10616 – 10628.
11892 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 11890 –11893

Angewandte
Chemie
[3] a) R. E. Carter, Adv. Phys. Org. Chem. 1973, 10, 1 – 27; b) T.
Felder, C. A. Schalley, Angew. Chem. 2003, 115, 2360 – 2363;
Angew. Chem. Int. Ed. 2003, 42, 2258 – 2260; c) T. Hayama, K. K.
Baldridge, Y.-T. Wu, A. Linden, J. S. Siegel, J. Am. Chem. Soc.
2008, 130, 1583 – 1591; d) J. S. Mugridge, R. G. Bergman, K. N.
Raymond, Angew. Chem. 2010, 122, 3717 – 3719; Angew. Chem.
Int. Ed. 2010, 49, 3635 – 3637.
[4] J. D. West, S. E. Stafford, M. P. Meyer, J. Am. Chem. Soc. 2008,
130, 7816 – 7817.
[5] M. P. Meyer, Org. Lett. 2009, 11, 4338 – 4341.
[6] J. Saavedra, S. E. Stafford, M. P. Meyer, Tetrahedron Lett. 2009,
50, 1324 – 1327.
[8] a) L. S. Bartell, Tetrahedron Lett. 1960, 1, 13 – 16; b) L. S. Bartell,
J. Am. Chem. Soc. 1961, 83, 3567 – 3571; c) L. S. Bartell,
Tetrahedron 1962, 17, 177 – 190.
[9] J. D. Dunitz, R. M. Ibberson, Angew. Chem. 2008, 120, 4276 –
4278; Angew. Chem. Int. Ed. 2008, 47, 4208 – 4210.
[10] D. J. OꢀLeary, P. R. Rablen, M. P. Meyer, Angew. Chem. 2011,
123, 2612 – 2615; Angew. Chem. Int. Ed. 2011, 50, 2564 – 2567.
[11] a) D. K. Jones, D. C. Liotta, I. Shinkai, D. J. Mathre, J. Org.
Chem. 1993, 58, 799 – 801; b) G. Alagona, C. Ghio, M. Persico, S.
Tomasi, J. Am. Chem. Soc. 2003, 125, 10027 – 10039.
[12] J. C. Decius, E. B. Wilson, Jr., J. Chem. Phys. 1951, 19, 1409 –
1412.
[13] H. Zhu, M. P. Meyer, Chem. Commun. 2011, 47, 409 – 411.
[14] L. J. Gooßen, Appl. Organomet. Chem. 2004, 18, 602 – 604.
[15] Gaussian09, RevisionA.02, M. J. Frisch, et al., see the Support-
ing Information for the full reference.
[7] W. A. Kreiner, H. D. Rudolph, B. T. Tan, J. Mol. Spectrosc. 1973,
48, 86 – 99.
Angew. Chem. Int. Ed. 2012, 51, 11890 –11893
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11893