Origins of 1,6-stereoinduction in torquoselective 6π electrocyclizations

Article abstract of DOI:10.1021/ja400882y

A novel stereoselective electrocyclization developed for the total synthesis of reserpine has been explored by both experiment and theory. A stereocenter six atoms away from the newly forming chiral center is responsible for the diastereoselectivity of the ring closure. This stereogenic center, lying at the junction of two six-membered rings, defines the conformation of the substrates' fused ring skeleton that ultimately distinguishes between the two allowed, disrotatory triene geometries at the transition state. The presence of allylic strain in the disfavored transition state results in a torquoselective ring closure (dr up to 15.7:1).

Full text of DOI:10.1021/ja400882y

Article
pubs.acs.org/JACS
Origins of 1,6-Stereoinduction in Torquoselective 6π
Electrocyclizations
Ashay Patel,† Gregg A. Barcan,†,‡ Ohyun Kwon,* and K. N. Houk*
Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095-1234, United States
S
* Supporting Information
ABSTRACT: A novel stereoselective electrocyclization devel-
oped for the total synthesis of reserpine has been explored by
both experiment and theory. A stereocenter six atoms away
from the newly forming chiral center is responsible for the
diastereoselectivity of the ring closure. This stereogenic center,
lying at the junction of two six-membered rings, defines the
conformation of the substrates’ fused ring skeleton that
ultimately distinguishes between the two allowed, disrotatory
triene geometries at the transition state. The presence of allylic
strain in the disfavored transition state results in a torquoselective ring closure (dr up to 15.7:1).
Scheme 2. Two Disrotatory Modes of Electrocyclization of a
Chiral Hexatriene
INTRODUCTION
■
The 6π electrocyclizations of hexatrienes were an impetus for
the formulation of the ground-breaking orbital symmetry rules
formulated by Woodward and Hoffmann.1 The orbital
symmetry rules predict that the thermal 6π electrocyclization
is disrotatory, while the photochemical process is conrotatory
(Scheme 1).2
Scheme 1. Disrotatory and Conrotatory Modes for the
Electrocyclic Ring Closure of 1,3,5-Hexatriene
cross-coupling/electrocyclization protocol that relied on a
remote stereocenter to induce a diastereoselective ring closure
(Scheme 3). Intrigued by the high levels of selectivity observed
Scheme 3. 1,6-Stereoinduction in a 6π Electrocyclization
Approach to Reserpine Alkaloids
Two modes of disrotation exist for all thermal 6π
electrocyclizations. For a chiral substrate, these two disrotatory
modes of electrocyclization yield diastereomeric products. If
one diastereomer is formed preferentially, such a ring closure is
said to be torquoselective (Scheme 2).3,4
While the orbital symmetry rules clearly explain that thermal
6π electrocyclizations must occur through disrotation, it is less
clear what specific factors control the preference for one
disrotatory mode of 6π electrocyclic ring closure over the other.
Examples of the torquoselective thermal 6π electrocyclization
of hexatrienes date back to 1963,5 and many examples have
been reported since then.6−15 However, the origins of
stereoselectivities in these examples remain unclear.
We recently reported an approach to the reserpine alkaloids
based upon a highly torquoselective thermal 6π electro-
cyclization.16 Pentacyclic structures were prepared with
extremely high levels of selectivity by employing a tandem
Received: January 28, 2013
Published: February 27, 2013
© 2013 American Chemical Society
4878
dx.doi.org/10.1021/ja400882y | J. Am. Chem. Soc. 2013, 135, 4878−4883

Journal of the American Chemical Society
Article
in these reactions, we explored the scope of this chemistry
computationally. Initial results suggested that the methyl ester
at the 3-position of the triene could be engaged in allylic strain
with the adjacent D-ring methylene. We proposed that varying
the steric bulk of groups at the 3-position would have a
significant impact on the diastereoselectivity of the electro-
cyclization. Herein, we report our experimental and computa-
tional results investigating the origins of this remote 1,6-
stereoinduction.
Table 1. Synthesis of Trienes and Experimental
Diastereoselectivities of Electrocyclizations
SYNTHESIS AND ELECTROCYCLIZATION OF
TRIENES
In order to assess the effect of substituent size on the electrocyclization
reaction, we envisioned an expedient route to several trienes with
variations in the size of the functional groups at the 3-position. Our
synthetic plan began with the addition of 3-propynylmagnesium
bromide to the BF3-complexed dihydroisoquinoline 1 (Scheme 4).
■
a
Scheme 4. Synthesis of Key Enynes 3a−f
a
See the Supporting Information for details.
Allylation of amine 2 provided enyne 3a that was functionalized at the
alkyne terminus with groups of varying steric requirement. We then
prepared and studied the electrocyclization of six distinct trienes
derived from the tandem carbopalladation/cross-coupling of enynes
3a−f (Table 1).
The tandem carbopalladation/Suzuki coupling sequence generated
the corresponding trienes, which upon heating, provided the 6π
electrocyclization products. Table 1 provides experimental diaster-
eoselectivities and the A-values of the C3 substituent, as a measure of
size, for each triene examined. All of the electrocyclizations took place
at 120 °C in xylenes, except for R = CO2Me and R = TMS that
underwent cyclization in the cross-coupling reaction at 80 °C. Starting
with the smallest substituent, R = H (Table 1), we observed an
increase in the diastereomeric ratio from 3.5:1 to 6.7:1 for R = SPh
and then to 15.7:1 and 13.3:1 for R = CO2Me and R = Me,
respectively. The diastereomeric ratio is greater for the reaction of
triene 4c(R = CO2Me) compared to the ring closure of 4d (R = Me)
due to differences in reaction temperature; however, the exper-
imentally determined ΔΔG‡ is larger for the triene 4d than 4c (vide
infra). The observation that trienes bearing larger substituents provide
greater levels of diastereoselectivity than trienes with smaller
substitutent is intuitive. Suprisingly, with the larger R = TMS and R
= tBu substituents, we observe that the diastereomeric ratio decreases
to 4.3:1 and 3.2:1, respectively.
a
For entries 3 and 5, the electrocyclization product was isolated
directly from the cross-coupling reaction without isolation of the
triene. All other products were generated by subsequent heating of the
b
isolated triene at 120 °C in xylenes. Diastereomeric ratios determined
c
1
by H NMR. Relative configuration of the new stereogenic center
assigned by analogy to a crystal structure of a pentacyclic reserpine
precursor (see the Supporting Information).16
COMPUTATIONAL METHODS
■
All structures examined computationally were initially prepared using
Avogadro, a freely available molecular editor/builder program that
includes a constrained force field optimization feature useful in
generating reasonable starting geometries for QM optimization.17
Quantum mechanical calculations were performed using Gaussian
09.18 Geometries were initially optimized in vacuo using the meta-
hybrid density functional M06-2X19 with the 6-31+G(d,p) basis set.
Normal-mode analysis confirmed that all optimized reactants and
products were indeed minima and that all transition states located
were first-order saddle points. Activation barriers (ΔH‡ and ΔG‡)
were also determined using M06-2X/def2-TZVPP20//M06-2X/6-
31+G(d,p) single points with thermal corrections determined at the
M06-2X/6-31+G(d,p) level of theory, whereas reaction energies
(ΔHrxn and ΔGrxn) were only calculated at the M06-2X/6-31+G(d,p)
level of theory. All QM calculations were performed using an
“ultrafine” numerical integration grid, consisting of 99 radial shells and
590 angular points per shell. Errors in computed entropies introduced
by the treatment of low frequency modes as the harmonic motions
4879
dx.doi.org/10.1021/ja400882y | J. Am. Chem. Soc. 2013, 135, 4878−4883

Journal of the American Chemical Society
Article
were corrected by raising all harmonic frequencies below 100 cm−1 to
exactly 100 cm−1 as described by Truhlar et al.21 The discussion of
stereoselectivity is restricted to only the lowest energy transition-state
conformers. A more extensive survey of the structures of transition-
state conformers can be found in the Supporting Information.
For the reactions of 4a, 4c, and 4d, a simple relationship
between the steric bulkiness of the C3 substituent, as measured
by A-values, and selectivity exists. Triene 4a, bearing only a
hydrogen atom at C3, undergoes ring closure with little
diastereoselectivity (ΔΔG‡ = 1.0 kcal/mol). Trienes 4c and 4d
with ester and methyl substituents, respectively, yield more
highly selective electrocyclizations (ΔΔG‡ = 1.5−2.0 kcal/
mol). Despite the relative bulkiness of TMS and tBu, the
electrocylizations of trienes 4e and 4f are not very selective,
according to experiment, and this suggests that the neither the
α nor the β transition state can effectively accommodate
extremely bulky C3 substituents such as TMS or tBu.
COMPUTATIONAL RESULTS
The electrocyclizations of trienes 4a−f (Scheme 5) are all
highly exergonic with ΔGrxn values ranging from −18 kcal/mol
■
Scheme 5. Six Model Systems Examined Computationally
Transition states TS4dα and TS4dβ of the electrocyclization
of triene 4d are shown in Figure 2. The triene adopts a boatlike
for the reaction of triene 4a to −27 kcal/mol for the reaction of
triene 4f. Consequently, the stereoselectivity observed exper-
imentally is due to differences in the energies of the two
electrocylicization transition states. The activation energies for
this series of electrocyclization reactions are approximately 30
kcal/mol and are consistent with the elevated temperatures
required experimentally for the electrocyclization. However, the
barriers of the electrocyclic ring closures of trienes 4e and 4f are
lower (ΔG‡ = 26 kcal/mol) than those for the electro-
cyclizations of 4a−d. The computed and experimentally
determined energy differences between the lowest energy
conformers of the α and β transition states for electro-
cyclization of 4a−f are listed in Table 2.
Figure 2. Newman projections of TS4dα and TS4dβ, highlighting the
allylic strain present in the disfavored transition state.
geometry in the disrotatory transition structures, and the two
modes of ring closure can be viewed on the right side of Figure
1. In the favored TS4dα, the C-terminal methyl group rotates
Table 2. A-Values of C3 Substituents, Computed ΔΔH‡ and
ΔΔG‡, and Experimental ΔΔG‡ and Values
b
b
c
comp
comp
comp
A-
value
ΔH‡
ΔΔH‡
ΔΔH‡
exptl
a
a
a
a
a
triene (R)
(ΔG‡)
(ΔΔG‡)
(ΔΔG‡)
ΔΔG‡
4a (H)
0
29.1
(30.2)
0.8 (0.8)
1.0 (1.0)
0.3 (0.8)
1.6 (1.4)
2.4 (2.5)
1.5 (1.6)
1.0
1.5
1.9
2.0
1.0
4b (SPh)
4c (CO2Me)
4d (Me)
0.8
1.2
1.7
2.4
28.8
(29.4)
−0.3 (0.1)
1.6 (1.4)
2.3 (2.4)
1.2 (1.3)
26.1
(27.2)
27.8
(28.8)
4e (TMS)
4f (tBu)
25.8
(26.9)
>4.5 25.7
(26.8)
−0.2 (−0.7) −0.1 (−0.7)
0.9
a
b
c
Values in kcal/mol. Gas phase M06-2X/6-31+G(d,p) energies. Gas
phase M06-2X/def2-TZVPP//M06-2X/6-31+G(d,p) energies.
Figure 1. Top and front views of the α and β disrotatory transition
states of the electrocyclic reactions of triene 4, optimized using M06-
2X/6-31+G(d,p) in vacuo.
In general, computed selectivities for this series of electro-
cyclic reactions are in reasonable agreement with experimental
ones, although we are not able to reproduce the experimental
selectivity of the reaction of triene 4f (R = tBu). As shown in
Table 2, computations overestimate the diastereoselectivity of
the electrocyclizations of trienes 4d and 4e and underestimate
the selectivities of the electrocyclic reactions of 4b, 4c, and 4f.
The mean absolute deviation (MAD) between the computed
ΔΔG‡ and corresponding experimental values is 0.6 kcal/mol
for this series of electrocyclizations.
downward through a “concave up” boat geometry. The key
difference between the two transition states can be found in the
dihedral highlighted in green in Figure 1. This dihedral angle is
nearly planar (7°) in the disfavored transition structure
(TS4dβ) and staggered (73°) in the lower-energy TS4dα.
The Newman projections of TS4dα and TS4dβ in Figure 2
show more clearly this difference between the two transition
4880
dx.doi.org/10.1021/ja400882y | J. Am. Chem. Soc. 2013, 135, 4878−4883

Journal of the American Chemical Society
Article
structures. The planar arrangement in TS4dβ introduces allylic
strain in this transition state between the C3 substitutent and
the methylene hydrogen of the piperidine (highlighted). This
effect includes contributions from steric clashes between the
piperidine hydrogen and the methyl group as well as torsional
strain between the two coplanar bonds.
The electrocyclic reactions of tricyclic trienes 4b and 4f merit
further discussion, as factors beside allylic strain are also
important in explaining the torquoselectivity of these ring
closures. Triene 4b possesses a flexible thiophenyl substituent
at the C3 position of the triene, introducing the possibility that
several rotameric α and β transition states exist. The lowest
energy transition-state rotamers leading to the α and β
diastereomers, TS4bα and TS4bβ, respectively, are shown in
Figure 3. TS4bα is 0.8 kcal/mol lower in free energy than
would exist in the absence of these intramolecular dispersion
interactions, we replaced the phenyl group with a hydrogen
atom or methyl group and performed single point energy
calculations on these structures. ΔΔE between these model
trienes are 5.9 and 4.9 kcal/mol, respectively, favoring the
formation of the α diastereomer.
According to computations, the reaction of 4f proceeds with
a ΔG‡ value of 25 kcal/mol and a ΔGrxn value of −27 kcal/mol.
These values are similar to corresponding values determined for
the reaction of 4e. In comparison to the electrocyclizations of
trienes 4a−d, 4e and 4f undergo ring closures with a 3−5 kcal/
mol lower barrier and are also 3−7 kcal/mol more exergonic.
The diminished barrier and the increased exergonicity of the
electrocyclizations of both 4e and 4f are due to a decrease in
steric demand for the bulky C3 substituent as the reactant
progresses along the reaction coordinate toward the product.
Computations also indicate that the electrocyclic reaction of
triene 4f (R = tBu) is essentially nonselective, with a slight
preference for the β stereoisomer (minor product). Illustrated
in Figure 4 are the α and β transition states for this
Figure 3. Top and front views of the α and β disrotatory transition
states of the electrocyclic reaction of triene 4b optimized using M06-
2X/6-31+G(d,p) in vacuo. Note the S-phenyl group is made
transparent for clarity where necessary.
Figure 4. Top and back views of the α and β disrotatory transition
states of the electrocyclic reaction of triene 4d optimized using M06-
2X/6-31+G(d,p) in vacuo.
TS4bβ according to the large basis set calculations. This ΔΔG‡
value is smaller than the 1.5 kcal/mol value observed
experimentally. The difference in the corresponding H−C−
C−C dihedral between TS4bα and TS4bβ is even larger than
that of TS4dα and TS4dβ, as the dihedral angle in TS4bβ is 0.7°
presumably because the thiophenyl substituent is conjugated
with the triene in this structure. Allylic strain in TS4bβ may be
attenuated by electrostatic attraction between the electro-
negative sulfur atom and the positively charged hydrogen atom;
however, torsional strain will still destabilize TS4bβ relative to
TS4bα.
TS4bα and TS4bβ are both stabilized by intramolecular C−H
π interactions that could provide stabilization by dispersion. In
TS4bα, the C−H π interaction occurs between the S-phenyl
group and the C2 hydrogen of the triene (2.85 Å). A similar
interaction occurs in TS4bβ between the S-phenyl group and an
aromatic hydrogen of the aromatic A-ring (2.84 Å). Because the
reactions are performed in xylenes, the solvent may be
competing with the S-phenyl group for these weak interactions
in such way that the two transition states may not be
differentially stabilized. In order to estimate the selectivity that
electrocyclization. Both TS4fα and TS4fβ suffer from a number
of steric clashes, some of which are quite severe and range in
terms of interatomic (H−H) distances from 1.94 to 2.07 Å.
The introduction of additional steric clashes in the α transition
state of 4f leads to the erosion of selectivity observed
experimentally.
ORIGINS OF DIFFERENTIAL ALLYLIC STRAIN IN
THE ELECTROCYCLIZATION TRANSITION STATES
OF TRIENES 4A−F
■
The piperidine ring (C-ring, Scheme 5) directly fused to the
triene discriminates between the two modes of disrotation. This
ring prefers to adopt a half-chair conformation, which places
the hydrogen at the stereocenter in an axial position.
Overlaying the piperidine rings (heavy atoms only) of the α
and β transition structures (TS4α and TS4β) reveals the
conformations of this heterocycle in either transition state are
nearly identical (RMSD 0.17 Å).
4881
dx.doi.org/10.1021/ja400882y | J. Am. Chem. Soc. 2013, 135, 4878−4883

Journal of the American Chemical Society
Article
Author Contributions
How does this piperidine ring cause the two disrotatory
modes to have different energies? In the α transition states, the
triene adopts a “concave up” boat geometry. This arrangement
places the C3 substituent (see Figures 1 and 2) between the
pseudoaxial and pseudoequatorial hydrogens of the piperidine
ring. Eclipsing strain is avoided as the C3−C4 bond of the
triene also lies between these hydrogens. In the β transition
states, the triene features a “concave down” boat that forces the
C3−C4 bond into the plane of the pseudoequatorial C−H
bond (Figures 1 and 2), introducing eclipsing strain as well as
allylic strain between the C3 substituent and the pseudoequa-
torial hydrogen.
Remote stereoinduction in this manifold is only possible
because the triene adopts one of two well-defined, clearly
distinguishable boatlike geometries (i.e., “concave up” or
“concave down”) at the transition state. These boatlike
transition states have rigid geometries that do not distort
easily, since the boat arrangement of the triene maximizes
orbital overlap22 within the cyclic array of p orbitals, while
conforming to the rules of orbital symmetry. Overlays of the
triene portions of TS4aβ−TS4fβ confirm this recalcitrance
toward distortion as the triene geometries are conserved
throughout the series of trienes, even for TS4eβ (R = TMS)
and TS4fβ (R = tBu).
†These authors contributed equally.
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENTS
■
A.P. and K.N.H. acknowledge the financial support of the NIH
(GM-36700 to K.N.H.). A.P. thanks the Chemical-Biology
Interface Training Program for its support (T32 GM 008496).
The following computational resources were used in this study:
the Hoffman2 cluster at UCLA and the Extreme Science and
Engineering Discovery Environment (XSEDE) supported by
the NSF (OCI-1053575). O.K. and G.A.B. acknowledge
funding from the NIH (R01GM071779) and the NSF
(equipment grant CHE-1048804).
REFERENCES
■
397.
(1) Woodward, R. B.; Hoffmann, R. J. Am. Chem. Soc. 1965, 87, 395−
(2) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. 1969, 8,
781−932.
(3) Rondan, N. G.; Houk, K. N. J. Am. Chem. Soc. 1985, 107, 2099−
2111.
(4) Kirmse, W.; Rondan, N. G.; Houk, K. N. J. Am. Chem. Soc. 1984,
106, 7989−7991.
(5) Corey, E. J.; Hortmann, A. G. J. Am. Chem. Soc. 1963, 85, 4033−
CONCLUSIONS
■
4034.
We have explored the torquoselective 6π electrocyclizations of
a series of isoquinoline-derived trienes differentially function-
alized at the 3-position of the triene experimentally and
computationally. For the six trienes examined, there is a
qualitative agreement between theory and experiment. The
observed diastereoselectivities for this series of triene electro-
cyclizations are sensitive to substitution at the C3 position of
the triene. Trienes bearing moderately bulky substituents such
as methyl or methyl ester at this position undergo 6π
electrocyclization with the greatest levels of diastereoselectivity,
whereas trienes with more bulky C3 substituents such as R =
TMS or R = tBu undergo a much less stereoselective ring
closure. Our results also illustrate that the conformation of the
piperidine ring (C-ring) in conjunction with the strict
geometric requirements of the triene at the transition state
are responsible for the relay of stereochemical information from
a distal stereogenic center to the forming stereocenter six atoms
away. Allylic strain in the β transition states explains the
observed preference for the α diastereomer in this series of
electrocyclizations.
(6) Okamura, W. H.; Peter, R.; Reischl, W. J. Am. Chem. Soc. 1985,
107, 1034−1041.
(7) Bamba, M.; Nishikawa, T.; Isobe, M. Tetrahedron Lett. 1996, 37,
8199−8202.
(8) Bamba, M.; Nishikawa, T.; Isobe, M. Tetrahedron 1998, 54,
6639−6650.
(9) Whitesell, J. K.; Minton, M. A. J. Am. Chem. Soc. 1987, 109,
6403−6408.
(10) Benson, C. L.; West, F. G. Org. Lett. 2007, 9, 2545−2548.
(11) Jung, M. E.; Min, S.-J. Tetrahedron 2007, 63, 3682−3701.
(12) Hayashi, R.; Feltenberger, J. B.; Hsung, R. P. Org. Lett. 2010, 12,
1152−1155.
(13) Hayashi, R.; Walton, M. C.; Hsung, R. P.; Schwab, J. H.; Yu, X.
Org. Lett. 2010, 12, 5768−5771.
(14) Katsunori, T.; Katsumura, S. J. Am. Chem. Soc. 2002, 124, 9660−
9661.
(15) Sklenicka, H. M.; Hsung, R. P.; McLaughlin, M. J.; Wei, L.-L.;
Gerasyuto, A. I.; Brennessel, W. B. J. Am. Chem. Soc. 2002, 124,
10435−10442.
(16) Barcan, G. A.; Patel, A.; Houk, K. N.; Kwon, O. Org. Lett. 2012,
14, 5388−5391.
(17) Hanwell, M. D.; Curtis, D. E.; Lonie, D. C.; Vandermeersch, T.;
Zurek, E.; Hutchison, G. R. J. Cheminf. 2012, 4, 17.
(18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
revision C.01; Gaussian, Inc.: Wallingford, CT, 2009.
(19) Zhao, Y.; Truhlar, D. Theor. Chem. Acc. 2008, 120, 215−241.
ASSOCIATED CONTENT
■
S
* Supporting Information
Relevant NMR data, synthetic details, coordinates for
computed structures in xyz format, details regarding the
conformational analysis of the transition states, and computa-
tional methods. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
houk@chem.ucla.edu; ohyun@chem.ucla.edu
Present Address
‡Gregg A. Barcan: Department of Chemistry, Stanford
University, Stanford, California, 94305-5080, United States.
4882
dx.doi.org/10.1021/ja400882y | J. Am. Chem. Soc. 2013, 135, 4878−4883

Journal of the American Chemical Society
Article
(20) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7,
3297−3305.
(21) Ribeiro, R. F.; Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J.
Phys. Chem. B 2011, 115, 14556−14562.
(22) Houk, K. N.; Li, Y.; Evanseck, J. D. Angew. Chem., Int. Ed. Engl.
1992, 31, 682−708.
4883
dx.doi.org/10.1021/ja400882y | J. Am. Chem. Soc. 2013, 135, 4878−4883