Origins of aryl substituent effects on the stereoselectivities of additions of silyl enol ethers to a chiral oxazolinium ion

Article abstract of DOI:10.1021/ol202911v

Density functional theory calculations are reported that reveal the role of aromatic interactions in the additions of aryl-substituted silyl enol ethers to a chiral oxazolinium ion. Aryl-substituted silyl enol ethers give the opposite diastereomer of the

Full text of DOI:10.1021/ol202911v

ORGANIC
LETTERS
2011
Vol. 13, No. 24
6572–6575
Origins of Aryl Substituent Effects on the
Stereoselectivities of Additions of Silyl
Enol Ethers to a Chiral Oxazolinium Ion
Elizabeth H. Krenske*
School of Chemistry, University of Melbourne, Victoria 3010, Australia, and Australian
Research Council Centre of Excellence for Free Radical Chemistry and Biotechnology
ekrenske@unimelb.edu.au
Received October 28, 2011
ABSTRACT
Density functional theory calculations are reported that reveal the role of aromatic interactions in the additions of aryl-substituted silyl enol ethers
to a chiral oxazolinium ion. Aryl-substituted silyl enol ethers give the opposite diastereomer of the adduct than do aliphatic silyl enol ethers, due to
a combination of attractive cationꢀπ and CHꢀπ interactions, reduced steric repulsion, and lower torsional strain in the more “crowded”
transition state.
There is growing interest among synthetic chemists in
the use of aromatic groups as stereodirecting elements.1
Herein, theoretical calculations are used to provide the
first explanation for unexpected aryl-substituent effects
observed in an asymmetric formylation reaction. The
reaction, reported by Hoppe,2 involves the Lewis acid-
catalyzed addition of the chiral 2-methoxy-3-sulfonyl-1,3-
oxazolidine 1 to a silyl enol ether (Scheme 1), and is
representative of a formylation strategythatwas pioneered
by both Hoppe2,3 and Scolastico.4 The Lewis acid-
catalyzed reactions of 1 with silyl enol ethers 2aꢀc afford
the diastereomeric, masked R-formyl ketones 3 and 4,
with selectivities reaching >95:5. Unexpectedly, however,
opposite stereoselectivities were obtained depending on
whether an aliphatic or an aryl-substituted silyl enol ether
was used.2 Cyclohexenol TMS ether 2a gave predomi-
nantly the 2R-adduct (3), but the aryl-substituted analo-
gues 2b and 2c gave predominantly 2S-adducts (4).5
Prompted by the increasing ability of theory to uncover
new forms of aryl-induced stereocontrol, a density func-
tional theory investigation of the aryl effect on Hoppe’s
asymmetric formylation is reported here.
The Lewis acid-catalyzed reactions of the oxazolidine 1
with silyl enol ethers are believed to proceed via inter-
mediate oxazolinium cations of the type 5.6 Hoppe pro-
posed that the transition states giving rise to diastereomers
3 and 4 adopt the conformations TS-3 and TS-4, respec-
tively (Scheme 2). The antiperiplanar arrangement of the
CdC and CdO bonds was suggested on the basis of
Noyori’swork7 on additionsof silyl enoletherstooxonium
cations, where the anti conformation was proposed to
(1) (a) Nishio, M. Tetrahedron 2005, 61, 6923–6950. (b) Nishio, M.;
Hirota, M.; Umezawa, Y. The CH/π Interaction: Evidence, Nature, and
Consequences; Wiley-VCH: New York, 1998. (c) Jones, G. B. Tetrahedron
2001, 57, 7999–8016. (d) Knowles, R. R.; Jacobsen, E. N. Proc. Natl.
Acad. Sci. U.S.A. 2010, 107, 20678–20685.
(2) Conde-Frieboes, K.; Hoppe, D. Tetrahedron 1992, 48, 6011–6020.
(3) (a) Conde-Frieboes, K.; Hoppe, D. Synlett 1990, 99–102.
(b) Winter, E.; Hoppe, D. Tetrahedron 1998, 54, 10329–10338. (c) Steif,
F.; Wibbeling, B.; Meyer, O.; Hoppe, D. Synthesis 2000, 743–753.
(5) A similar reversal of stereoselectivity, with drs reaching >95:5,
was observed with aliphatic and aryl-substituted cyclopentenol TMS
ethers (ref 2).
€
(d) Bruggemann, M.; Holst, C.; Hoppe, D. Eur. J. Org. Chem. 2001,
647–654.
(6) (a) Tomalia, D. A.; Paige, J. N. J. Org. Chem. 1973, 38, 422–430.
(b) Bernardi, A.; Cardani, S.; Poli, G.; Potenza, D.; Scolastico, C.
Tetrahedron 1992, 48, 1343–1352.
(7) Murata, S.; Suzuki, M.; Noyori, R. Tetrahedron 1988, 44, 4259–
4275.
(4) (a) Bernardi, A.; Cardani, S.; Carugo, O.; Colombo, L.; Scolastico,
C.; Villa, R. Tetrahedron Lett. 1990, 31, 2779–2782. (b) Palazzi, C.;
Poli, G.; Scolastico, C.; Villa, R. Tetrahedron Lett. 1990, 31, 4223–4226.
(c) Bernardi, A.; Cavicchioli, M.; Poli, G.; Scolastico, C.; Sidjimov, A.
Tetrahedron 1991, 47, 7925–7936.
r
10.1021/ol202911v
Published on Web 11/22/2011
2011 American Chemical Society

Scheme 1. Hoppe’s Asymmetric Additions of Silyl Enol Ethers
to a Chiral 2-Methoxy-3-sulfonyl-oxazolidine2
Scheme 2. Transition State Geometries Proposed2 for Additions
of Silyl Enol Ethers 2aꢀc to the Oxazolinium Ion 5
minimize electrostatic repulsion between the oxonium
oxygen and the developing positive charge on the silyl enol
ether. On this basis, the selectivity for 4 displayed by silyl
enol ethers 2b and 2c is surprising, as TS-4 appears much
more sterically crowded than TS-3. Hoppe tentatively
suggested that TS-4 may be stabilized by an attractive
interaction between the tosyl group and the aromatic ring.
The factors controlling stereoselectivity in the reactions
of 5 with silyl enol ethers were studied here by means of
density functional theory calculations in Gaussian 09.8
Geometry optimizations and conformational searching
were performed at the B3LYP/6-31G(d) level.9 Energies
were calculated both with B3LYP and with three newer
functionals that include corrections for dispersion
(B3LYP-D,10 B97-D,10,11 and M06ꢀ2X12). This type
of approach has previously been used to study aryl effects
on stereoselectivities of cycloaddition reactions13 and has
more generally been shown to be superior to B3LYP alone
for polar reactions of large molecules.14
LUMO of the oxazolinium ring have almost equal coeffi-
cients on N and O, with a slightly greater contribution
from nitrogen.15
When 5 is drawn in its iminium form, the proposed anti-
CdC CO transition state geometries can be seen to
3 3 3
conform to Seebach’s topological rule for additions of π-
systems,16 insofar as the donor CdC π-bond is gauche to
both the acceptor CdN π-bond and the smallest substi-
tuent (H) on the acceptor.
The geometry of the oxazolinium cation 5 is shown in
Figure 1. Although 5 is often drawn with a CdO double
bond, the alternative iminium representation is actually of
slightly greater importance. The CN and CO bond lengths
˚
˚
in 5 (1.31 and 1.29 A, respectively) are both within 0.03 A
of true double bond values, as judged from the model
iminium and oxonium cations 6 and 7. The HOMO and
Figure 1. Calculated geometry of the oxazolinium cation 5 and
related bond lengths in model compounds (B3LYP/6-31G(d)
˚
geometries, distances in A).
(8) Frisch, M. J. et al. Gaussian 09, Revision B.01; Gaussian, Inc.:
Wallingford, CT, 2010.
(9) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789.
(b) Becke, A. D. J. Chem. Phys. 1993, 98, 1372–1377. (c) Becke, A. D.
J. Chem. Phys. 1993, 98, 5648–5652.
(10) Grimme, S. J. Comput. Chem. 2006, 27, 1787–1799.
(11) Antony, J.; Grimme, S. Phys. Chem. Chem. Phys. 2006, 8, 5287–
5293.
Transition states were computed for the reaction of 5
with the aliphatic silyl enol ether 2a. The two lowest-energy
TS conformers (TS-3a and TS-4a) are shown in Figure 2.
Both possess the proposed antiperiplanar CdC CO
geometry. All four density functional methods correctly
(12) (a) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215–
241. (b) Zhao, Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157–167.
(13) (a) Krenske, E. H.; Houk, K. N.; Harmata, M. Org. Lett. 2010,
12, 444–447. (b) Krenske, E. H.; Houk, K. N.; Lohse, A. G.; Antoline,
J. E.; Hsung, R. P. Chem. Sci. 2010, 1, 387–392. (c) Antoline, J. E.;
Krenske, E. H.; Lohse, A. G.; Houk, K. N.; Hsung, R. P. J. Am. Chem.
Soc. 2011, 133, 14443–14451.
3 3 3
(15) Orbital coefficients (2py contributions) calculated at the HF/
6-31G//B3LYP/6-31G(d) level: HOMO, N ꢀ0.23, O 0.19; LUMO, N 0.16,
O 0.16.
(14) Harvey, J. N. Faraday Discuss 2010, 145, 487–505.
(16) Seebach, D.; Golinsky, J. Helv. Chim. Acta 1981, 64, 1413–1423.
Org. Lett., Vol. 13, No. 24, 2011
6573

Figure 2. Diastereomeric transition states for reaction of oxa-
zolinium cation 5 with the silyl enol ether 2a (B3LYP/6-31G(d)
geometries). The bottom view shows the Newman projection
down the forming CꢀC bond. Values of ΔHqrel (kcal/mol) were
obtained using the 6-31G(d) basis set, except for the values in
parentheses, which employ the 6-311þG(d,p) basis. Distances
Figure 3. Diastereomeric transition states for reaction of oxa-
zolinium cation 5 with the silyl enol ether 2b. Basis sets are as
described in Figure 2.
˚
are in A.
over TS-3b. This is in keeping with the underlying role
of dispersion in CHꢀπ interactions in general.1,18 The
dispersion-inclusive methods predict TS-4b to be favored
by 0.3ꢀ0.6 kcal/mol. B3LYP, which lacks dispersion
corrections, favors TS-3b by 0.6 kcal/mol.
Transition states for the reaction of5 withsilyl enol ether
2c are shown in Figure 4. Here, both aromatic interactions
and torsional differences control the stereoselectivity. TS-
predict 3a to be the favored product.17 The computed
stereoselectivity, based on the two pictured conformers alone,
is ΔΔHq = 1.9ꢀ2.9 kcal/mol (ΔΔGq = 1.8ꢀ2.7 kcal/mol).
The stereoselectivity for 2a is primarily related to torsional
strain about the forming CꢀC bond. Steric clashing in the
more “crowded” transition state TS-4a is actually no more
severe than in TS-3a, but the avoidance of steric clashing in
TS-4a is achieved only at a cost of rotating the forming CꢀC
bond into a partially eclipsed arrangement. The Newman
projections at the bottom of the Figure show the small
dihedral angles in TS-4a (31ꢀ36°). In contrast, the favored
transition state (TS-3a) is quite well staggered (50ꢀ51°).
Transition states for the corresponding reaction with the
aryl-fused silyl enol ether 2b are shown in Figure 3. Here,
3c adheres to the anti CdC CO geometry, but TS-4c
3 3 3
adopts a gauche arrangement with the CdC bond lying
underneath the oxazolinium ring. The gauche conforma-
tion avoids destabilizing steric clashes and allows for a
weakly stabilizing interaction between the aromatic π-
system and the oxazolinium ring. Proton H-5 in TS-4c lies
closest to the aromatic π-cloud, albeit with a longer
˚
Carom separation than in TS-4b (2.92 and 2.71 A,
H
3 3 3
respectively). TS-4c is further stabilized by torsional strain
differences; its forming bond is well staggered (dihedrals
55ꢀ60°) whereas TS-3c is partially eclipsed (37ꢀ43°).
All four functionals correctly predict TS-4c to be
favored, but B3LYP underestimates the selectivity com-
pared to the other three functionals. The stereoselectiv-
ity based on the two pictured TS conformations is
1.8ꢀ3.9 kcal/mol.
the transition states again adopt antiperiplanar CdC CO
3 3 3
arrangements, but the partial eclipsing that was present in
TS-4a is not found in TS-4b. The stereoselectivity instead
originates from interactions between 5 and the aromatic
ring of 2b. Whereas TS-4a contained a close steric contact
involving the CH2 group adjacent to the forming bond, in
TS-4b this is replaced by a weak attractive interaction
between the aromatic π-cloud and the positively charged
atoms of the oxazolinium ring. The H-5 proton of the
oxazolinium ion lies closest to the π-cloud of the aromatic
ring (hashed line).
(17) Additions of 2aꢀc to 5 were found to be exothermic by at least
28 kcal/mol (M06ꢀ2X), and likely to be under kinetic control under the
reaction conditions (0 °C, CH2Cl2).
(18) Tsuzuki, S.; Fujii, A. Phys. Chem. Chem. Phys. 2008, 10, 2584–
2594.
Only the three density functionals that include terms to
represent dispersion correctly predict TS-4b to be favored
6574
Org. Lett., Vol. 13, No. 24, 2011

Table 1. Computed Stereoselectivities in the Gas Phase and in
Dichloromethane for Additions of 2aꢀc to the Oxazolinium
Ion 5a
reactant
ΔΔGq (gas phase)
ΔΔGq (CH2Cl2)
2a
2b
2c
2.0
0.0
3.4
0.7
ꢀ2.0
ꢀ0.3
a ΔΔGq(TS-4ꢀTS-3), M06ꢀ2X/6-311þG(d,p)//B3LYP/6-31G(d),
kcal/mol. Solution values incorporate CPCM solvation energies com-
puted at the M06ꢀ2X/6-31G(d) level.
The stabilization provided by the cationꢀπ and CHꢀπ
interactions in TS-4b/c is expected to be sensitive to
solvation. Solvation energies in dichloromethane were
computed with the CPCM method,20 and the results are
shown in Table 1. The computed ΔΔGq values capture the
solution-phase stereoselectivity for 2c but not for 2b (a full
treatment would require consideration of other low-lying
TS conformers). They do, however, reflect the expected
attenuation of aromatic interactions in solution.
In summary, DFT calculations provide a clear explana-
tion for the unexpected aryl-substituent effects on stereo-
selectivity in Hoppe’s asymmetric formylation reaction.
The aryl substituent controls stereoselectivity by alleviat-
ing torsional strain, and by replacing repulsive steric
interactions with attractive aromatic interactions. These
stereocontrol elements are likely to be relevant to various
other asymmetric reactions. Theory continues to reveal
an increasing repertoire of aryl-substituent effects on
stereoselectivity.
Figure 4. Diastereomeric transition states for reaction of oxa-
zolinium cation 5 with the silyl enol ether 2c (B3LYP/6-31G(d)
geometries). Basis sets are as described in Figure 2.
To estimate the degree to which aromatic interactions
contribute to the overall transition-state stabilization for
2b/c, the aromatic ring [(CH)4 fragment] was deleted, and
the two free valences were filled with H atoms.19 Removal
of the aromatic ring from 2b completely reversed the
selectivity, such that TS-3 was favored by 1.0 kcal/mol
(ΔΔEq, B3LYP-D). For 2c, the selectivity for TS-4 per-
sisted, but was 0.5 kcal/mol smaller. The aromatic inter-
actions are the primary influence on stereoselectivity for
2b, while for 2c they act in concert with the torsional strain
effects.
Acknowledgment. This work was performed with the
financial support of the Australian Research Council
(DP0985623) and ARC Centre of Excellence for Free
Radical Chemistry and Biotechnology, and computational
support from the NCI National Facility and University of
Melbourne School of Chemistry. EHK thanks Professor
Ken Houk (UCLA) for valuable discussions.
˚
(19) H atoms were appended at a distance of 1.09 A, with the same
Supporting Information Available. Calculated geo-
metries and energies, and a complete citation for ref 8.
This material is available free of charge via the Internet at
http://pubs.acs.org.
bond angles and dihedrals as the original carbon atoms. All other
remaining geometrical parameters were left unchanged.
(20) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995–2001.
(b) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem.
2003, 24, 669–681.
Org. Lett., Vol. 13, No. 24, 2011
6575