De novo chiral amino alcohols in catalyzing asymmetric additions to aryl aldehydes

Article abstract of DOI:10.1021/ol0600640

A de novo structural class of chiral amino alcohol catalysts has been identified through a synergistic effort combining novel architectures from [4 + 3] cycloadditions and quantum mechanical interaction field predictions that closely match subsequent experimental measurements.

Full text of DOI:10.1021/ol0600640

ORGANIC
LETTERS
2
006
Vol. 8, No. 8
565-1568
De Novo Chiral Amino Alcohols in
Catalyzing Asymmetric Additions to Aryl
Aldehydes
1
†
‡
†
,†
Jian Huang, James C. Ianni, Jennifer E. Antoline, Richard P. Hsung,* and
Marisa C. Kozlowski*,
‡
DiVision of Pharmaceutical Sciences and Department of Chemistry, UniVersity of
Wisconsin, Madison, Wisconsin 53705, and Department of Chemistry, Roy and Diana
Vagelos Laboratories, UniVersity of PennsylVania, Philadelphia, PennsylVania 19104
rhsung@wisc.edu; marisa@sas.upenn.edu
Received January 10, 2006
ABSTRACT
A de novo structural class of chiral amino alcohol catalysts has been identified through a synergistic effort combining novel architectures
from [4 3] cycloadditions and quantum mechanical interaction field predictions that closely match subsequent experimental measurements.
+
We have been investigating a new class of chiral amino
1
,2
alcohols that could be prepared from chiral allenamides
Scheme 1. Oxyallyl Cation [4 + 3] Cycloadditions
and serve as ligands in asymmetric catalysis. Specifically,
by employing chiral allenamides 1 as precursors to generate
3,4
nitrogen-stabilized chiral oxyallyl cations 3 via epoxidation
5
(
Scheme 1), we were able to develop a highly stereoselective
6
7,8
inter- and intramolecular (not shown) [4 + 3] cycload-
dition with dienes to give cycloadducts 4 and established
9
an asymmetric 1,3-dipolar cycloaddition using Cu(II)-
†
University of Wisconsin.
University of Pennsylvania.
‡
(
1) For a review on allenamides, see: Hsung, R. P.; Wei, L.-L.; Xiong,
H. Acc. Chem. Res. 2003, 36, 773.
2) Krause, N.; Hashmi, A. S. K. Modern Allene Chemistry; Wiley-VCH
Verlag GmbH & Co. KGaA: Weinheim, Germany, 2004; Vols. 1 and 2.
3) For a recent review, see: Harmata, M. Rec. Res. DeV. Org. Chem.
997, 1, 523.
4) For examples of nitrogen-stabilized oxyallyl cations, see: (a) Myers,
A. G.; Barbay, J. K. Org. Lett. 2001, 3, 425. (b) Sung, M. J.; Lee, H. I.;
Chong, Y.; Cha, J. K. Org. Lett. 1999, 1, 2017. (c) Walters, M. A.; Arcand,
H. R. J. Org. Chem. 1996, 61, 1478 and references therein. (d) See ref 6.
(
(
1
(
bisoxazoline catalysts.10,11 Cycloadducts 4 embody a unique
and structurally novel class of bicyclic R-amido ketone that
(
5) Rameshkumar, C.; Xiong, H.; Tracey, M. R.; Berry, C. R.; Yao, L.
J.; Hsung, R. P. J. Org. Chem. 2002, 67, 1339.
6) Xiong, H.; Hsung, R. P.; Berry, C. R.; Rameshkumar, C. J. Am. Chem.
Soc. 2001, 123, 7174.
(7) Xiong, H.; Huang, J.; Ghosh, S. K.; Hsung, R. P. J. Am. Chem. Soc.
2003, 125, 12694.
(8) Rameshkumar, C.; Hsung, R. P. Angew Chem., Int. Ed. 2004, 43,
615.
(
1
0.1021/ol0600640 CCC: $33.50
© 2006 American Chemical Society
Published on Web 03/14/2006

can be transformed into chiral amino alcohols 5 and 6.6a
However, their capacity as chiral ligands is completely
unprecedented.
18 â-amino alkoxide zinc catalysts that exhibit a wide range
12b,18
of enantioselection.
This parametrization set was previ-
ously shown to provide internally consistent models,1
2b,18
but
Thus, given that we now have an efficient access to this
class of substrates, we first assessed the ability of these new
ligands as asymmetric catalysts in the alkylation of aryl
aldehydes using a quantum mechanical QSAR analysis based
on its excellent predictive potential.12,13 Catalytic asymmetric
the true test of the method lies in a priori predictions. For
ee
all the analyses, the enantioselectivities are converted to ∆G
ee
using the relationship ∆G ) -RTln[(S)/(R)], so that the
variables used in the correlation possess an underlying linear
relationship.
additions to aldehydes and ketones represent an important
For new chiral catalysts, prediction of enantioselec-
tivities commences with calculation of the lowest energy
catalyst conformers. For the benzaldehyde alkylation reac-
tion, the dimeric catalyst structures (Figure 1) were found
to provide the most robust models previously18 and were
subsequently employed in this analysis. Initially, con-
formers of the compounds were constructed and computed
using the semiempirical method PM3 in Spartan.19 These
structures were used with the QMQSAR program. For the
GQSAR program, the lowest energy conformers were
subsequently geometry optimized using RHF/3-21G* and
verified as ground-state minima by an additional frequency
analysis.
venue in organic synthesis.1
4-16
We report here our prelimi-
nary efforts unveiling these de novo chiral amino alcohols
in asymmetric catalysis.
The enantiomeric excesses (ee) of the heretofore unknown
â-amino alkoxide zinc catalysts (Figure 1) were predicted
All of the resultant dimers were then aligned about a set
of common atoms. For QMQSAR, the alignment employed
the Zn-O-N atoms. For GQSAR, the alignment employed
the Zn-O-Zn-O atoms. A different alignment was utilized
with QMQSAR since this program has a protocol to mask
parts of the structure.
Figure 1. Asymmetric catalysis with chiral amino alcohols.
with 3D-QSSR (quantitative structural selectivity relation-
1
7
Both the QMQSAR and GQSAR programs employ
quantum mechanical interaction fields in the form of
electrostatic potential field (EPF) values computed at ordered
grid points encompassing the compound. Either single-point
ship) methods employing quantum molecular interaction
fields as implemented in the programs QMQSAR and
GQSAR. Although detailed descriptions on how both these
programs compute and utilize molecular fields are described
2
0
12,13
PM3 semiempirical calculations with Divcon (QMQSAR)
elsewhere,
a short introduction is provided here. Two sets
21
22
or B3LYP/6-31G** calculations with Gaussian03 (GQSAR)
afforded the requisite electron densities employed in generat-
ing the EPF values for each compound along a common grid.
Field spacing in the EPF was initially 0.35 Å and was
adjusted during the course of the model building to a finer
grid around correlated EPF points according to a MAXMIN
of catalysts are employed in generating predictions using this
approach: a parametrization set to generate a model, and a
prediction set to which this model is applied. For this
asymmetric reaction, the parametrization set is composed of
(
9) For recent reviews, see: (a) Harmata, M.; Rashatasakhon, P.
Tetrahedron 2003, 59, 2371. (b) Harmata, M. Acc. Chem. Res. 2001, 34,
23
diversity algorithm. The EPF values represent the pool of
5
95. (c) Davies, H. M. L. In AdVances in Cycloaddition; Harmata, M., Ed.;
JAI Press: Greenwich, CT, 1998; Vol. 5, p 119. (d) West, F. G. In AdVances
in Cycloaddition; Lautens, M., Ed.; JAI: Greenwich, CT. 1997; Vol. 4, p
independent variables from which the multi-linear regression
(
MLR) models were built.
1
. (e) Rigby, J. H.; Pigge, F. C. Org. React. 1997, 51, 351. (f) Harmata, M.
Simple MLR models between the EPF points and the ∆Gee
Tetrahedron 1997, 53, 6235.
(
10) Huang, J.; Hsung, R. P. J. Am. Chem. Soc. 2005, 127, 50.
values of the parametrization set were optimized by a
(11) Harmata, M.; Ghosh, S. K.; Hong, X.; Wacharasindu, S.; Kirch-
hoefer, P. J. Am. Chem. Soc. 2003, 125, 2058.
12) (a) Dixon, S., Jr.; Mertz, K. M., Jr.; Lauri, G.; Ianni, J. C. J. Comput.
Chem. 2005, 26, 23. (b) Kozlowski, M. C.; Dixon, S.; Panda, M.; Lauri, G.
J. Am. Chem. Soc. 2003, 125, 6614.
(
(17) For other work using QSAR analyses for synthetic reactions, see:
(a) Oslob, J. D.; Åkermark, B.; Helquist, P.; Norrby, P.-O. Organometallics
1997, 16, 3015. (b) Alvarez, S.; Schefzick, S.; Lipkowitz, K.; Avnir, D.
Chem.sEur. J. 2003, 9, 5832. (c) Hoogenraad, M.; Klaus, G. M.; Elders,
N.; Hooijschuur, S. M.; McKay, B.; Smith, A. A.; Damen, E. W. P.
Tetrahedron: Asymmetry 2004, 15, 519. (d) Van der Linden, J. B.; Ras,
E.-J.; Hooijschuur, S. M.; Klaus, G. M.; Luchters, N. T.; Dani, P.; Verspui,
G.; Smith, A. A.; Damen, E. W. P.; McKay, B.; Hoogenraad, M. QSAR
Comb. Sci. 2005, 24, 94. (e) Sciabola, S.; Alex, A.; Higginson, P. D.;
Mitchell, J. C.; Martin J. Snowden, M. J.; Morao, I. J. Org. Chem. 2005,
70, 9025.
(
13) Phaun, P.-W.; Ianni, J. C.; Kozlowski, M. C. J. Am. Chem. Soc.
004, 126, 15473.
14) For reviews, see: (a) Pu, L.; Yu, H.-B. Chem. ReV. 2001, 101, 757.
b) Pu, L. Tetrahedron 2003, 59, 9873. (c) Walsh, P. J. Acc. Chem. Res.
003, 36, 739.
15) For earlier examples, see: (a) Oguni, N.; Matsuda, Y.; Kaneko, T.
2
(
(
2
(
J. Am. Chem. Soc. 1988, 110, 7877. (b) Kitamura, M.; Suga, S.; Kawai,
K.; Noyori, R. J. Am. Chem. Soc. 1986, 108, 6071.
(
16) For leading references, see: (a) Jeon, S.-J.; Li, H.; Garc ´ı a, C.;
(18) Ianni, J. C.; Phuan, P. W.; Annamalai, V.; Panda, M.; Kozlowski,
M. C. Angew. Chem., Int. Ed., accepted.
(19) Spartan ’02; Wavefunction, Inc.: Irvine, CA, 2002.
(20) Dixon, S. L.; K. M. Merz, J. J. Chem. Phys. 1997, 107, 879.
(21) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(22) Gaussian 03, revision A.1; Gaussian, Inc.: Pittsburgh, PA, 2003.
See Supporting Information for the full citation.
LaRochelle, L. K.; Walsh, P. J. J. Org. Chem. 2005, 70, 448 and refs 1-62
cited therein. (b) Hari, Y.; Aoyama, T. Synthesis 2005, 583. (c) Li, H.;
Walsh, P. J. J. Am. Chem. Soc. 2004, 126, 6538. (d) Wei, C.; Mague, J. T.;
Li, C.-J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5749. (e) Xu, M. H.; Pu,
L. Org. Lett. 2002, 4, 4555, also see 4143. (f) Wipf, P.; Kendall, C. Chem.s
Eur. J. 2002, 8, 1778. (g) Seebach, D.; Beck, A. K.; Heckel, A. Angew.
Chem., Int. Ed. 2001, 40, 92. (h) Frantz, D. E.; F a¨ ssler, R.; Carreira, E. M.
J. Am. Chem. Soc. 2000, 122, 1806.
(23) Kirkpatrick, S.; Gelatt, C. D., Jr.; Vecchi, M. P. Science 1983, 220,
671.
1566
Org. Lett., Vol. 8, No. 8, 2006

simulated Monte Carlo approach.12 Overall, the minimal
MLR models that provided highly cross-validated results of
Table 2. QMQSAR and GQSAR Enantioselectivity Predictions
for Ligands 5a-d, 6a, and 6d
the parametrization set in the aldehyde alkylation were
comprised of two EPF points per catalyst.1
2b,18
In other words,
average predictions (18 Models Total)
two regions of the catalysts were found to account for most
of the variance in ee; groups of different sizes or electronic
aspects at these positions modify the enantioselection.
These models were subsequently used in generating pre-
dictions in the asymmetric alkylation of benzaldehyde for the
six new amino alcohols 5a-d and 6a and 6d (Scheme 2)
incorporating the novel scaffold found in amines 5 and 6.
QMQSAR
GQSAR
expt. values
mean ee
(%)
SD ee
(%)
mean ee
(%)
SD ee
(%)
ee
(%)
yield
(%)
L
228 K ) -45 °C
5
a
b
c
99.2
99.0
99.2
99.1
99.0
98.8
0.1
0.2
0.1
0.2
0.1
0.1
98.1
98.0
98.0
97.9
97.8
97.6
0.7
0.6
0.6
0.7
0.5
0.5
95
81
5
5
5
6
6
97
97
81
81
d
a
d
Scheme 2. Chiral Amino Alcohols
95
90
73
90
2
78 K ) 0 °C
5
5
5
5
6
6
a
b
c
d
a
d
98.0
97.6
98.1
97.9
97.6
97.3
0.2
0.3
0.2
0.4
0.2
0.2
95.9
95.7
95.8
95.6
95.5
95.1
1.2
1.0
1.0
1.1
0.9
0.9
93
91
93
79
92
90
88
86
298 K ) 25 °C
5
5
5
5
6
a
b
c
d
a
97.0
96.5
97.2
96.9
96.6
96.1
0.2
0.4
0.3
0.5
0.2
0.2
94.3
94.1
94.3
94.0
93.8
93.3
1.5
1.3
1.2
1.4
1.1
1.1
89
93
89
89
87
93
For both QMQSAR and GQSAR, average ∆G predic-
1
2,18
tions
(Table 1) were obtained from 18 separate models
6d
44
53
QMQSAR and GQSAR models give highly predictive
Table 1. QMQSAR and GQSAR ∆G Predictions for Ligands
a-d, 6a, and 6d
results, but that the GQSAR results exhibit slightly better
5
13,18
correlations with experimental values.
On this basis,
average predictions (18 models total)
QMQSAR GQSAR
mean ∆G SD mean ∆G
(kcal/mol) (kcal/mol) (kcal/mol) (kcal/mol)
greater than 95% ee values were expected for the catalysts
derived from 5a, 5c, 5d, 6a, or 6d in the asymmetric
alkylation of benzaldehyde at -45 °C (see Scheme 2).
Encouraged by these predictions, we prepared the deriva-
tives illustrated in Scheme 2 from [4 + 3] adduct 4 (Scheme
1) and examined these compounds in this transformation.24
Reactions were performed in toluene at various temperatures
expt. valuesa
SD
∆G
(kcal/mol)
L
5
5
5
5
6
6
a
b
c
d
a
d
2.47
2.37
2.51
2.45
2.39
2.32
0.09
0.15
0.11
0.18
0.08
0.07
2.09
2.06
2.08
2.05
2.03
1.99
0.33
0.26
0.26
0.29
0.22
0.21
1.66
1.80
1.76
1.83
1.66
using 2.2 equiv of Et Zn and 5-10 mol % of the respective
2
25
ligands, 5a, 5c, 5d, 6a, or 6d. Gratifyingly, all the additions
were achieved in high yields (Table 2), indicating that these
new chiral amino alcohols form highly active catalysts. More
significantly, g90% ee was obtained experimentally in
almost all cases, closely matching the QMQSAR and
GQSAR predictions (Tables 1 and 2). Reasons for lower but
reproducible ee values obtained using 6d at 0 °C and room
temperature are not clear at this point; however, the 95% ee
observed at -45 °C indicates that the results at higher
temperature reflect a secondary effect rather than the inherent
enantioselection. The difference in ee between 5 and 10 mol
a
Values obtained from ee results listed in Table 2 by using ∆Gee
RTln[(S)/(R)].
)
-
obtained with 17 parametrization compounds each. As
indicated by the small standard deviations (SD) of the
averaged results from these 18 models, it is clear that no
single parametrization compound is providing an overwhelm-
ing bias. For convenience, the predicted energies are
converted into enantioselectivities computed at the reaction
temperatures employed, 228, 278, and 298 K (Table 2). The
standard deviations (SD) from the average predictions also
provide a measure of the error in the predictions, which is
low for both the QMQSAR and GQSAR models. The
enantioselectivity predictions for all six amino alcohols using
both models are high. Prior work indicates that both the
%
ligand is negligible, indicating little competition from other
less selective pathways. In line with prior results, QMQSAR
(24) See Supporting Information. Due to low yields in the preparation
and purification, 5b was not examined. We appreciate one reviewer’s
suggestion of White’s protocol as a viable option: White, J. D.; Wardrop,
D. J.; Sunderman, K. F. Org. Synth. 2002, 79, 125 and 130.
(25) Enantiomeric excess values were determined using HPLC with a
Chiralcel-OD column at 254 nm.
Org. Lett., Vol. 8, No. 8, 2006
1567

tends to overpredict the enantioselectivity with a range of
9
9
7-99% ee, whereas GQSAR predictions with a range of
4-98% ee are closer to the actual experimental values
(Table 2).
On the basis of the assignment of absolute configuration
of the resulting alcohol, a model using ligand 5a is proposed
in Figure 2. The regions in red outline the portions of the
Figure 3. Results with additional substrates and 5a.
the aryl group of the aldehyde increased (Figure 3, entry 1
vs entries 2-4).
Along the same line, when the size decreased from Ph to
Bn, the ee decreased noticeably (entry 1 vs entry 5).
Intriguingly, the ee did not suffer any further loss and actually
increased slightly with an additional methylene unit as shown
in entry 6. This finding also implies that these new ligands
can be useful in asymmetric additions to simple alkyl
aldehydes (entries 5 and 6), which remain a challenge in
the field.
Finally, the result in entry 7 suggests that a more electron-
rich and Lewis basic aldehyde carbonyl oxygen (para-
methoxybenzaldehyde) with tighter coordination to Zn further
enhances the difference between TS-anti-pro-R and TS-anti-
pro-S and leads to higher ee. Thus, a less electron-rich and
Lewis basic carbonyl oxygen, as in para-fluorobenzaldehyde,
should provide lower ee, as was seen in entry 8. It should
be noted here that, although not shown, enantiomers of these
new chiral amino alcohols could be prepared from either
diastereoselective or enantioselective [4 + 3] cycloadditions
and, thus, in principle, utilized to provide the antipodes (R-
enantiomers) of the addition products.
Figure 2. Proposed stereochemical models.
catalyst that the QSSR models identified as important to
enantioselection; in general, moderate steric bulk at these
positions increases ee. The model reveals interesting struc-
tural features of this new catalytic system. The unique oxa-
bicyclic scaffold of 5a provides an inflexible backbone
suitable for rigid complexation with Zn. Such rigidity allows
for an excellent transfer of the chiral information from the
indicated ligand substituents to the transition state (TS)
maximizing the energetic difference between pro-R and pro-S
transition states and leading to high enantioselectivities.
Of the four canonical transition states, anti-pro-R/S and
syn-pro-R/S,26 the two anti transition states are more favor-
able due to the large bulk of the chiral ligand that compresses
the syn transition states leading to disfavorable steric
interactions. Among the anti transition states, a more severe
steric interaction between Ar and Et groups in TS-anti-pro-R
results in greater stability for TS-anti-pro-S, which accounts
for the observed S-enantiomer.
This proposed model is consistent with several additional
experiments. First, disruption of coordination to the zinc by
silylation of the alcohol oxygen atom in 5a as in 7 (Scheme
Herein, we have described the identification of a class of
structurally novel chiral amino alcohols for use in asymmetric
catalysis through an approach that utilizes simple, rapid, and
highly predictive 3D-QSSR analyses.
Acknowledgment. This work is supported by NIH-
NIGMS (GM066055 and GM59945). Computing resources
were provided by the NCSA (CHE020059) and the NSF
CRIF (CHE0131132) programs. This work was in part
carried out at the University of Minnesota.
2
) clearly diminished the ability of the ligand to provide
asymmetric induction in the addition to benzaldehyde (90%
yield, 88% ee vs 65% yield, 0% ee). Second, the above
models indicate that large aldehyde substituents amplify the
steric repulsion in TS-anti-pro-R and lead to higher ee. In
accord with this observation, the ee increased as the size of
Supporting Information Available: Experimental de-
tails, NMR spectra for all new compounds, and the full
citation of reference 22. This material is available free of
charge via the Internet at http://pubs.acs.org.
(
26) (a) Yamakawa, M.; Noyori, R. J. Am. Chem. Soc. 1995, 117, 6327.
(
b) Yamakawa, M.; Noyori, R. Organometallics 1999, 18, 128. (c) Goldfuss,
B.; Houk, K. N. J. Org. Chem. 1998, 63, 8998. (d) Vazquez, J.; Pericas,
M. A.; Maseras, F.; Lledos, A. J. Org. Chem. 2000, 65, 7303. (e) Rasmussen,
T.; Norrby, P.-O. J. Am. Chem. Soc. 2003, 125, 5130.
OL0600640
1568
Org. Lett., Vol. 8, No. 8, 2006