Examination of the role of Taft-type steric parameters in asymmetric catalysis

Article abstract of DOI:10.1021/jo901698t

(Chemical Equation Presented) We report the use of Taft steric parameters to correlate the substituent size of a ligand with the enantiomeric ratio of a reaction. Linear free energy relationships can be constructed by plotting the log of enantiomeric ratio (er) versus the steric parameters reported by Taft and modified by Charton. Successful correlations were found for aldehyde and ketone allylation under NHK conditions using modular oxazoline ligands developed in our laboratory. Using these correlations, ligands were designed and evaluated for carbonyl allylation reactions. A break in the Charton plot results and is attributed to a global structural change in the catalyst. Additionally, several previously reported enantioselective reactions are analyzed resulting in excellent correlations for both catalysts and substrates. Finally, limitations and issues are presented with illustrative examples.

Full text of DOI:10.1021/jo901698t

pubs.acs.org/joc
Examination of the Role of Taft-Type Steric Parameters in
Asymmetric Catalysis
Matthew S. Sigman* and Jeremie J. Miller
Department of Chemistry, University of Utah, 315 South 1400 East, Salt Lake City, Utah 84112
sigman@chem.utah.edu
Received August 6, 2009
We report the use of Taft steric parameters to correlate the substituent size of a ligand with the
enantiomeric ratio of a reaction. Linear free energy relationships can be constructed by plotting the
log of enantiomeric ratio (er) versus the steric parameters reported by Taft and modified by Charton.
Successful correlations were found for aldehyde and ketone allylation under NHK conditions using
modular oxazoline ligands developed in our laboratory. Using these correlations, ligands were
designed and evaluated for carbonyl allylation reactions. A break in the Charton plot results and is
attributed to a global structural change in the catalyst. Additionally, several previously reported
enantioselective reactions are analyzed resulting in excellent correlations for both catalysts and
substrates. Finally, limitations and issues are presented with illustrative examples.
Introduction
qualitative in nature in contrast to the widely used quanti-
tative methods for exploring electronic effects through linear
free energy relationships (LFER).4 Quantitative methods do
exist for analysis of steric effects (vide infra)5 but are not
commonly employed in the analysis of a reaction mechan-
ism, presumably due to the difficulty in interpreting the
results or synthetically accessing the necessary analogs to
perform the study.
An area of modern organic chemistry which often suggests
a steric effect is asymmetric catalysis.6 Chiral ligands utilized
in enantioselective reactions are often designed to “hinder”
an approach from a certain prochiral face of the substrate to
enhance enantioselection. The incorporation of a larger
substituent is often the first synthetic modification made to
a promising ligand structure. Considering the potential of a
lengthy synthetic endeavor to prepare and evaluate a de-
signed ligand as well as the unanticipated outcomes often
found in asymmetric catalysis, we became interested in
correlating steric parameters to enantiomeric ratios for
In 1894, Emil Fischer proposed the “lock and key” model
for substrate/enzyme interactions implicating the impor-
tance of the size of a substrate in relationship to the active
site of an enzyme.1 Although molecular interactions are
clearly more complex than this model suggests, steric ef-
fects,2 or the repulsive spatial interactions encountered with-
in or between molecules, are commonly evoked in explaining
observed reaction outcomes, as originally suggested by In-
gold in 1930.3 A steric hindrance explanation is generally
(1) Fischer, E. Ber. Dtsch. Chem. Ges. 1894, 27, 673–679.
(2) For a demonstration of the role electrostatics in conformational
analysis, see: Pophristic, V.; Goodman, L. Nature 2001, 411, 565–568.
(3) Ingold, C. K. J. Chem. Soc. 1930, 1032–1039.
(4) (a) For an example text book account of linear free energy relationships,
see : Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic Chemistry;
University Science Books: Mill Valley, CA, 2006. (b) For a review of Hammett
analysis, see: Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165–95.
(5) (a) For a monograph describing different approaches to LFER and steric
effects, see : Balaban, A. T.; Chiriac, A.; Motoc, I.; Simon, Z. Lecture Notes in
Chemistry, Vol. 15: Steric Fit in Quantitative Structure-Activity Relations;
Springer: Berlin, 1980. (b) For a more recent discussion of steric effects in
transition metal chemistry, see: Bunten, K. A.; Chen, L.; Fernandez, A. L.; Poe,
A. J. Coord. Chem. Rev. 2002, 233-234, 41–51.
(6) Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds. Comprehensive
Asymmetric Catalysis I-III; Springer Verlag: New York, 1999; Vols. 1-3.
DOI: 10.1021/jo901698t
r
Published on Web 09/15/2009
J. Org. Chem. 2009, 74, 7633–7643 7633
2009 American Chemical Society

Sigman and Miller
JOCFeatured Article
facilitating the design of new ligands.7 Recent efforts in
asymmetric catalysis design have focused on using computa-
tional methods to predict the outcome of enantioselective
reactions and the design of improved catalysts.8 However,
for catalyst systems which have complicated/unidentified
structures, these methods are difficult to implement. Our
approach is based on a more classic evaluation of steric
parameters using LFERs developed by Taft and Charton
wherein the nature of the catalyst structure or assumptions
related to reactive species are not initially necessary. It
should be noted that Taft steric parameters have been used
to correlate enantioselectivity as a function of substrate
substituents as reported by Porter but not previously as a
function of ligand substituents.9,10 Herein, we describe the
investigation of ligand and substrate effects in asymmetric
catalysis using steric parameters developed by Taft11 and
Charton12 with examples from ligands developed in our
laboratory as well as illustrative examples from previously
reported asymmetric catalytic processes. In addition, the
limitations and potential applications are discussed.
Taft-Type Steric Parameters. The correlation of reaction
rate to electronic parameters is a well-known and widely used
method to probe the mechanism of chemical reactions. Among
the most common methods are Hammett and Brønsted linear
free energy relationships.13 Correlations of this type are mea-
surements of electronic effects and do not take into account
steric parameters that might also contribute to the overall rate
of the reaction in question. In fact, the rates of hydrolysis of
many ortho-substituted benzoates do not fit Hammett ana-
lyses.3 This is in contrast to the often excellent correlations of
the corresponding para- and meta-substituted substrates.14
The rates of hydrolysis of ortho-substituted benzoates de-
crease as the relative size of the ortho-substituents are in-
creased, suggesting a steric effect. Seminal research by Taft
sought to explore this possibility, ultimately allowing for the
correlation of steric parameters to reaction rate.11a
FIGURE 1. Taft’s correlation of methyl ester substituents to the
rate of hydrolysis in the development of Taft’s steric parameters.
In the original studies, the relative rates of methyl ester
hydrolysis were measured for a wide range of aliphatic esters
(Figure 1).11a,b The equation shown in Figure 1 reflects the
steric influence of substituents on the rate of hydrolysis,
where ks is the rate of ester hydrolysis for a given substrate,
kCH is the methyl acetate hydrolysis rate, which is used as the
3
standard, δ is a proportionality constant which is a measure
of sensitivity, similar to the F value in a Hammett plot, and Es
is defined as the Taft steric parameter, analogous to a σ value
in a Hammett plot. The key hypothesis presented by Taft to
isolate the steric impact on hydrolysis rate was based on the
nature of the conditions employed. If hydrolysis is conducted
under basic conditions, the reaction involves the conversion
of a neutral reactant to a negatively charged intermediate.
Therefore, Taft proposed the transition state would be
influenced by polar effects (resonance and induction) and
would be an inaccurate measure of steric influences.8c The
acid-catalyzed pathway, however, involves a charged inter-
mediate formed from a charged precursor and therefore
should be influenced by polar effects to a much lesser degree.
Additionally, the steric influence of the proton was assumed
to be minimal. These assumptions result in a relationship
between the size of the substituent and the rate of hydrolysis,
(7) Miller, J. J.; Sigman, M. S. Angew. Chem., Int. Ed. 2008, 47, 771–774.
(8) For leading references, see: (a) Oslob, J. D.; Akermark, B.; Helquist,
P.; Norrby, P.-O. Organometallics 1997, 16, 3015–3021. (b) Lipkowitz, K. B.;
D’Hue, C. A.; Sakamoto, T.; Stack, J. N. J. Am. Chem. Soc. 2002, 124,
14255–14267. (c) Kozlowski, M. C.; Panda, M. J. Org. Chem. 2003, 68, 2061–
2076. (d) Alvarez, S.; Schefzick, S.; Lipkowitz, K.; Avnir, D. Chem.;Eur. J.
2003, 9, 5832–5837. (e) Kozlowski, M. C.; Dixon, S. L.; Panda, M.; Lauri, G.
J. Am. Chem. Soc. 2003, 125, 6614–6615. (f) Lipkowitz, K. B.; Kozlowski, M.
C. Synlett 2003, 1547–1565. (g) Chen, J.; Jiwu, W.; Mingzong, L.; You, T.
J. Mol. Catal. A:Chem. 2006, 258, 191–197. (h) Urbano-Cuadrado, M.; Carbo,
J. J.; Maldonado, A. G.; Bo, C. J. Chem. Inf. Model. 2007, 47, 2228–2234.
(9) For the only previous example of the use of Taft steric parameters to
correlate substrate enantioselectivity, see: Wu, J. H.; Zhang, G.; Porter,
N. A. Tetrahedron Lett. 1997, 38, 2067–2070.
log(ks/kCH ) = δEs. This analysis has been applied to various
3
organic and biological chemical processes15 and has been
used in medicinal chemistry via quantitative structure (re)-
activity relationships (QSAR).16 The Es values for several
common substituents are tabulated in Table 1.
Following the development of Taft’s method for steric
correlation there was considerable debate in the 1960s and
(10) An attempted correlation of enantioselectivity to ligand substituent
size using Taft steric parameters has been reported in a substituted proline
catalyzed R-sulfenylation of aldehydes. In this paper, the authors plot % ee
versus Taft Es values which is not a free energy relationship although an
(15) For selected examples, see: (a) Ingold, K. U. J. Phys. Chem. 1960, 64,
1636–42. (b) Eisch, J. J.; Trainor, J. T. J. Org. Chem. 1963, 28, 487–492.
(c) Martin, M. M.; Gleicher, G. J. J. Am. Chem. Soc. 1964, 86, 233–238.
(d) Fife, T. H.; Milstien, J. B. Biochemistry 1967, 6, 2901–7. (e) Milstien, J. B.;
Fife, T. H. J. Am. Chem. Soc. 1968, 90, 2164–8. (f) Sakurai, H.; Hosomi, A.;
Kumada, M. J. Org. Chem. 1969, 34, 1764–1768. (g) Baldwin, S. W.;
Mazzuckelli, T. J. Tetrahedron Lett. 1986, 27, 5975–8. (h) Hansch, C.;
Bjorkroth, J. P. J. Org. Chem. 1986, 51, 5461–2. (i) Nandigama, R. K.;
Edmondson, D. E. Biochemistry 2000, 39, 15258–15265. (j) van den Berg, D.;
Zoellner, K. R.; Ogunrombi, M. O.; Malan, S. F.; Terre’Blanche, G.;
Castagnoli, N.; Bergh, J. J.; Petzer, J. P. Bioorg. Med. Chem. 2007, 15,
3692–3702. (k) Sanhueza, C. A.; Dorta, R. L.; Vazquez, J. T. Tetrahedron:
Asymmetry 2008, 19, 258–264. (l) Ogunrombi, M. O.; Malan, S. F.;
Terre’Blanche, G.; Castagnoli, N.; Bergh, J. J.; Petzer, J. P. Bioorg. Med.
Chem. 2008, 16, 2463–2472.
ꢀ
interesting observation. Franzen, J.; Marigo, M.; Fielenbach, D.; Wabnitz,
T. C.; Jørgensen, K. A. J. Am. Chem. Soc. 2005, 127, 18296–18304.
(11) (a) Taft, R. W., Jr. J. Am. Chem. Soc. 1952, 74, 2729–2732. (b) Taft,
R. W., Jr. J. Am. Chem. Soc. 1953, 75, 4538–9. (c) Taft, R. W., Jr. Steric Eff.
Org. Chem. 1956, 556–675.
(12) (a) Charton, M. J. Am. Chem. Soc. 1969, 91, 615–618. (b) Charton,
M. J. Am. Chem. Soc. 1975, 97, 1552–6. (c) Charton, M. J. Am. Chem. Soc.
1975, 97, 3691–3693. (d) Charton, M. J. Am. Chem. Soc. 1975, 97, 3694–3697.
ꢀ
(13) For reviews and leading references, see: (a) Jaffe, H. H. Chem. Rev.
1953, 53, 191–261. (b) Bender, M. L. Chem. Rev. 1960, 60, 53–113.
(c) McDaniel, D. H.; Brown, H. C. J. Org. Chem. 1958, 23, 420–427.
(d) Advances in Linear Free Energy Relationships; Shorter, J., Chapman, N.
B., Eds.; Plenum Publishing Co. Ltd.: London, 1972. (e) Lewis, E. S. J. Phys.
Org. Chem. 1990, 3, 1–8.
(14) Newman, M. S. J. Am. Chem. Soc. 1950, 72, 4783–4786.
(16) Unger, S. H.; Hansch, C. Prog. Phys. Org. Chem. 1976, 12, 91–118.
7634 J. Org. Chem. Vol. 74, No. 20, 2009

Sigman and Miller
JOCFeatured Article
TABLE 1. Taft and Charton Steric Parameters for Common
Substituents
substituent
-Es (Taft)
v (Charton)
H
Me
-1.24
0
0
0.52
Et
nPr
0.07
0.36
0.39
0.47
0.79
0.38
1.54
0.56
0.68
0.68
0.76
0.87
0.7
1.24
1.33
nBu
iPr
cC6H11
PhCH2
tBu
1-adamantyl
Ph
2.54
0.57 and 1.66
1.54
1.70
CH(nPr)2
CH(iPr)2
CEt3
2.38
1970s concerning Es values and the notion that they were
purely a measure of steric effects.13d,17 In order to address this,
Charton correlated the Es values with the van der Waals radii
of the corresponding group, ruling out both inductive and
resonance effects.12 The resultant equation is log(k/k0) = ψν,
where the log of the relative rate (k/k0) is proportional to the
product of ψ (sensitivity factor) and ν, and ν is the adjusted Es
value based on the van der Waals radii18 of the given sub-
stituent. Some common substituents and their corresponding
ν values are presented in Table 1.
FIGURE 2. Effect of the proline module on facial selectivity in the
Cr-catalyzed addition of allyl bromide to carbonyl compounds.
allow for various modifications to be introduced. To this
end, we have been interested in peptide-derived R-amino-
oxazoline ligands which can be rapidly synthesized from amino
acid building blocks.19 These ligand sets have been success-
fully applied to the enantioselective allylation of aldehydes and
20
€
ketones under Furstner
modified Nozaki-Hiyama-
Kishi (NHK) conditions in which allylic halides are the allyl
source.21 For both aldehyde22 and ketone substrates,23 ex-
cellent enantiomeric ratios (er) can be obtained for aromatic
carbonyl derivatives but reactions with aliphatic carbonyl
substrates lead to only modest er.
Results and Discussion
Applying Taft-Based Steric Parameters to Asymmetric
Catalysis. An ultimate goal of correlating steric parameters
to the outcome of an enantioselective reaction is predicting
what type of modification to the ligand structure would
result in an optimized catalyst system with the least synthetic
effort. To approach this, a modular ligand set derived from
simple and readily available building blocks is optimal to
One intriguing observation in these studies is that the
facial selectivity of aldehyde and ketone allylation directly
correlates to the absolute configuration of the ligand proline
module.22,23 For example, when comparing ligands 1a and 2
that differ only in the configuration of the proline module, a
considerable difference in the enantiomeric ratio was ob-
served for both products (Figure 2). The allylation of ben-
zaldehyde was promoted in nearly 90% ee using both ligands
(96:4 er and 5.5:94.5 er), but with a different sense of
asymmetric induction. A similar effect was also observed
for the allylation of acetophenone (68:32 er and 20:80 er).
Systematic changes to the configuration of the other chiral
centers in the ligand structure had a less significant impact on
enantioselectivity. These observations prompted an explora-
tion of the role of the proline module in greater detail.7
In order to initiate the investigation, systematic changes to
the nature of the carbamate of the proline module were easily
explored due to the convergent nature of the synthesis and
the ability to modify the size of this group. Ligand 1a, the
optimal diastereomer for benzaldehyde allylation, was cho-
sen as a starting structure into which carbamate modifica-
tions were incorporated. Ligands 1b-e were synthesized
from a common precursor (Figure 3), differentiated by the
size of the carbamate substituent from a small group
(methyl), to a large group (1-adamantyl).19 The ligands were
then evaluated for allylation of an aromatic aldehyde
(17) (a) Fujita, T.; Takayama, C.; Nakajima, M. J. Org. Chem. 1973, 38,
1623–30. (b) For reanalysis and extension of Taft parameters to larger
substituents, see: MacPhee, J. A.; Panaye, A.; Dubois, J.-E. Tetrahedron
1978, 34, 3553–62.
(18) Bondi, A. J. Phys. Chem. 1964, 68, 441–451.
(19) (a) Rajaram, S.; Sigman, M. S. Org. Lett. 2002, 4, 3399–3401. (b)
Rajaram, S.; Sigman, M. S. Org. Lett. 2005, 7, 5473–5475. (c) Miller, J. J.;
Rajaram, S.; Pfaffenroth, C.; Sigman, M. S. Tetrahedron 2009, 65, 3110–
3119.
€
(20) (a) Furstner, A.; Shi, N. J. Am. Chem. Soc. 1996, 118, 12349–12357.
€
(b) Furstner, A. Chem. Rev. 1999, 99, 991–1045.
(21) For examples of catalytic enantioselective Nozaki-Hiyama-Kishi-
type reactions, see: (a) Bandini, M.; Cozzi, P. G.; Melchiorre, P.; Umani-
Ronchi, A. Angew. Chem., Int. Ed. 1999, 38, 3357–3359. (b) Bandini, M.;
Cozzi, P. G.; Umani-Ronchi, A. Chem. Commun. 2002, 919–927. (c) Choi,
H.-w.; Nakajima, K.; Demeke, D.; Kang, F.-A.; Jun, H.-S.; Wan, Z.-K.;
Kishi, Y. Org. Lett. 2002, 4, 4435–4438. (d) Inoue, M.; Suzuki, T.; Nakada,
M. J. Am. Chem. Soc. 2003, 125, 1140–1141. (e) Berkessel, A.; Menche, D.;
Sklorz, C. A.; Schroder, M.; Paterson, I. Angew. Chem., Int. Ed. 2003, 42,
1032–1035. (f) Lombardo, M.; Licciulli, S.; Morganti, S.; Trombini, C.
Chem. Commun. 2003, 1762–1763. (g) Paterson, I.; Bergmann, H.; Menche,
D.; Berkessel, A. Org. Lett. 2004, 6, 1293–1295. (h) Berkessel, A.; Schroeder,
M.; Sklorz, C. A.; Tabanella, S.; Vogl, N.; Lex, J.; Neudoerfl, J. M. J. Org.
Chem. 2004, 69, 3050–3056. (i) Inoue, M.; Nakada, M. Org. Lett. 2004, 6,
2977–2980. (j) McManus, H. A.; Cozzi, P. G.; Guiry, P. J. Adv. Synth. Catal.
2006, 348, 551–558. (k) Xia, G.; Yamamoto, H. J. Am. Chem. Soc. 2006, 128,
2554–2555. (l) Inoue, M.; Nakada, M. Heterocycles 2007, 72, 133–138. (m)
Hargaden, G. C.; McManus, H. A.; Cozzi, P. G.; Guiry, P. J. Org. Biomol.
Chem. 2007, 5, 763–766. (n) Hargaden, G. C.; Guiry, P. J. Adv. Synth. Catal.
2007, 349, 2407–2424. (o) Hargaden, G. C.; Muller-Bunz, H.; Guiry, P. J.
Eur. J. Org. Chem. 2007, 4235–4243. (p) Inoue, M.; Suzuki, T.; Kinoshita, A.;
Nakada, M. Chem. Rec. 2008, 8, 169–181. (q) Zhang, Z.; Huang, J.; Ma, B.;
Kishi, Y. Org. Lett. 2008, 10, 3073–3076. (r) Hargaden, G. C.; O’Sullivan,
T. P.; Guiry, P. J. Org. Biomol. Chem. 2008, 6, 562–566.
(22) Lee, J.-Y.; Miller, J. J.; Hamilton, S. S.; Sigman, M. S. Org. Lett.
2005, 7, 1837–1839.
(23) Miller, J. J.; Sigman, M. S. J. Am. Chem. Soc. 2007, 129, 2752–2753.
J. Org. Chem. Vol. 74, No. 20, 2009 7635

Sigman and Miller
JOCFeatured Article
FIGURE 3. Representative syntheses of oxazoline-proline ligands utilized in this study.
TABLE 2. Systematic Evaluation of Proline Carbamate Ligands in
Nozaki-Hiyama-Kishi Allylation Reaction with Three Carbonyl
Substrates
entry
G
R1
R2
er (R/S)a
ΔΔGqb
1
2
3
4
5
Me
Et
iPr
tBu
Ph
Ph
Ph
Ph
Ph
H
H
H
H
H
60:40
65.5:34.5
82.5:17.5
96:4
0.24
0.38
0.74
1.86
1.86
1-adamantyl
96:4
6
7
8
9
10
Me
Et
iPr
tBu
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
PhCH2CH2
H
H
H
H
H
51:49
0.02
0.06
0.16
0.61
0.60
52.5:47.5
56.5:43.5
74:26
FIGURE 4. Plot of enantiomeric ratio versus Charton values in the
enantioselective allylation of three carbonyl substrates.
1-adamantyl
73.5:26.5
11
12
13
14
15
Me
Et
iPr
tBu
Ph
Ph
Ph
Ph
Ph
Me
Me
Me
Me
Me
25:75
24:76
37.5:62.5
69:31
72:28
-0.65
-0.68
-0.30
0.47
asymmetric induction for the catalyst containing the smallest
group 1a and the largest group 1e was measured, albeit with
the opposite facial selection. Additionally, it is interesting to
inspect the calculated ΔΔGqvalues wherein the addition of
one methyl group (entries 3 and 4) changes this value
∼1.1 kcal/mol.
1-adamantyl
0.56
aER determined by HPLC equipped with a chiral stationary phase.
bEstimated at 25 °C.
Considering that an enantiomeric ratio is (by definition)
the relative rate of the formation of two enantiomers from a
pro-chiral substrate (for kinetically controlled reactions),
correlating steric parameters to the log of enantiomeric ratio
(log(R/S) = ψν) is possible.24 Excellent correlations for the
allylation of benzaldehyde, acetophenone, and hydrocin-
namaldehyde are observedusing eitherChartonorTaft values
(Figure 4, only Charton values depicted). The sensitivity (ψ)
is the greatest for benzaldehyde indicating that changing the
size of the carbamate substituent has the greatest impact
on enantioselectivity as compared to the other substrates
(benzaldehyde), an aromatic ketone (acetophenone), and an
aliphatic aldehyde (hydrocinnamaldehyde). In all cases,
qualitative trends are observed between the relative size of
the carbamate substituents and the product enantioselecti-
vity (Table 2). The range is noteworthy for the allylation of
benzaldehyde (entries 1-5) in that an er of 60:40 was
observed for the methyl substituent (entry 1) versus an er
of 96:4 for the 1-adamantyl substituent (entry 5), and a
similar, but less severe, trend is observed for the allylation
of hydrocinnamaldehyde. A less obvious trend is observed
for the allylation of acetophenone in that a change in facial
selectivity was found (entries 11-15) when progressing
from smaller to larger substituents. A similar magnitude of
(24) It should also be noted that krel values or selectivity factors derived
from kinetic resolutions should also be able to be correlated. Additionally,
ΔΔGq can be used in the correlation.
7636 J. Org. Chem. Vol. 74, No. 20, 2009

Sigman and Miller
JOCFeatured Article
evaluated. The observation of linear free energy relation-
ships is of great interest in that it implicates that only one
structural perturbation to the diastereomeric transition state
occurs upon a systematic change.
A main question that arises from this observation is what
is the structural origin of this relationship? Unfortunately,
we have been unable to structurally characterize the ligated
Cr complex which would aid in the development of a
predictive stereochemical model and allow for comparison
to the observed LFER.25 Therefore, our attention has been
directed toward three questions: (1) Can this correlation be
utilized to predict the outcome of a proposed catalyst? (2)
Can we extend the steric parameter analysis to a related
catalyst developed in our laboratory? (3) Can we re-evaluate
previously reported enantioselective processes using this
correlation? The following sections discuss our efforts to
explore these questions.
Synthesis of New Catalysts for the NHK Allylation. When
analyzing the LFER described above, larger substituents on
the carbamate could provide for higher enantioselectivity if
the relationship remained linear. This would depend on
whether any global structural changes occurred upon mak-
ing further systematic changes. The linear fits of the Charton
plots for each substrate can be extrapolated to a given
Charton value with the largest substituent of CEt3 predicted
to achieve exceptionally high enantioselection for benzalde-
hyde allylation (Figure 5). With this in mind, catalysts 1f-h
were synthesized and subsequently evaluated in the allyla-
tion of benzaldehyde (entries 1-3) and acetophenone
(entries 4-6). To our surprise, these catalysts performed
considerably poorer than the previous optimal catalyst 1a
with a tert-butyl or catalyst 1e with an adamantyl carbamate.
However, plotting these results shows that these substan-
tially larger groups can be fit linearly with a different ψ value
(0.72 versus 1.51, benzaldehyde allylation) with a break in
the plot (Figure 5). As is seen in Hammett analysis where a
break can indicate a change in mechanism,4,26 the break in
the Charton plot suggests a change in the catalyst conforma-
tion. The linearity following the break again indicates only
one change is occurring by modifying the carbamate group
with these larger groups to achieve the new LFER. While the
initial desire was to predict the outcome of these catalysts,
this study serves as evidence of the carbamate group impact
on catalyst conformation and thus the enantioselectivity.
Even without structural characterization of the catalyst, this
information should aid in the construction of a stereochem-
ical model for this complex system.
FIGURE 5. Use of larger ligand substituents in the NHK allylation
of benzaldehyde and acetophenone and the corresponding Charton
plots.
Charton Analysis of Optimal Ligand for Ketone Allylation.
As briefly mentioned above, we have recently reported the
first highly enantioselective catalytic ketone allylation of
aromatic ketones using allylic halides.23 Our use of a mod-
ular ligand facilitated the discovery of a truncated ligand 3a
responsible for the effectiveness of this system. Based on the
initial success with the Charton analysis for the optimal
ligand for benzaldehyde allylation, we explored the ana-
logous series with the truncated ligand. Ligands 3a-e,
lacking an oxazoline substituent, were synthesized. These
ligands were evaluated in both benzaldehyde and acetophe-
none allylation (Figure 6). Again plotting log of enantio-
meric ratio versus Charton values leads to a LFER for both
carbonyl substrates (Figure 6). In contrast to ligands 1a-e, a
significantly larger ψ value is observed for acetophenone
allylation (1.88) in comparison to benzaldehyde allylation
(0.58). The high sensitivity for acetophenone allylation is
highlighted by the er range of 0.28 to 10.1 (57% ee (S) to 82%
ee (R)) whereas benzaldehyde allylation has an er range of
5.3 to 19 (68% ee to 90% ee). The oxazoline module has
(25) We have kinetic evidence that the reaction is first order in [catalyst],
and we observe no nonlinear effects but have been unable to determine a
precise stereochemical model due to the number of possible orientations of
ligand/substrates/counterions about Cr.
(26) For an example, see: Richards, J. P.; Jencks, W. P. J. Am. Chem. Soc.
1982, 104, 4689–4691.
J. Org. Chem. Vol. 74, No. 20, 2009 7637

Sigman and Miller
JOCFeatured Article
FIGURE 6. Evaluation of analogues of the optimal ligand for
ketone allylation and resulting Charton analysis.
an unanticipated impact in how the substrates interact with
the catalyst as suggested by the differences in substrate
sensitivity for each catalyst class. Even with the significant
modifications to the structure of this catalyst as compared to
1a, a high enantioselectivity can be achieved for benzalde-
hyde allylation with catalyst 3e containing the adamantyl
carbamate. Catalyst 3e is poorer than the original optimal
catalyst 3d for acetophenone allylation suggesting that a
potential break may be occurring in this Charton plot as
larger substituents are incorporated. Considering this ligand
structure has a diastereomeric relationship between the pro-
line and amide module as compared to catalyst 1a, the ability
to extend the Charton analysis is notable.
Successful Correlation of Previously Reported Asymmetric
Catalytic Reactions. An important consideration is whether
the Charton analysis described herein is limited to only one
set of ligands/reactions. To evaluate this, various reactions
types and ligand structures in previously reported enantio-
selective processes were treated using Charton analysis. In
general, reactions performed using a range of ligand mod-
ifications were chosen for analysis and it should be noted
that the analyses presented below are only illustrative, not
FIGURE 7. Charton analysis of Pd-catalyzed allylic alkylation
reactions with oxazoline-based ligands.
comprehensive. Finally, at the end of this section, we will
present some of the unusual sets of data evaluated which
illustrate some of the potential limitations of the method.
Positive ψ Values: Where Larger Substituents Lead to
Better Selectivity. The seminal report by Pfaltz concerning
the use of phosphine oxazoline ligands, a highly useful and
modular ligand class, in asymmetric Pd-catalyzed allylic
alkylation reactions was evaluated (Figure 7).27 We found
that the size of the ligand substituent on the oxazoline can be
correlated to product enantiomeric ratio with a modest ψ =
0.35. This implicates that only substantial changes to the
group off the oxazoline will result in a significantly improved
enantiomeric ratio. Also, it should be noted that the Ph
substituent has two values (Table 1), and the smaller value
(0.57) was utilized. This will be discussed further below. A
similar correlation was found when Park and co-workers’
(27) von Matt, P.; Pfaltz, A. Angew. Chem., Int. Ed. Engl. 1993, 32, 566–
568.
7638 J. Org. Chem. Vol. 74, No. 20, 2009

Sigman and Miller
JOCFeatured Article
reported by Denmark and co-workers (Figure 9).32 Again,
a linear correlation in the Charton plot was observed
wherein the reaction outcome is highly sensitive to the nature
of the substituent size as indicted by a large magnitude
ψ value (-1.8).
Substrate Effects. Correlating the size of the substrate to
product enantiomeric ratio is possible, especially considering
that the origin of Taft/Charton values is based on the effect
of substituent size on the rate of ester hydrolysis. For
example, the enantioselective aziridination of styrenes with
Mn-Salen complexes reported by Komatsu33 was analyzed
wherein styrene has a significantly lower enantiomeric ratio
than substrates with substitution trans in the β-position
(Figure 10). An excellent correlation of the size of the group
in the β-position of the styrene was observed where a ψ value
of 1.42 was calculated, indicating that in this case a larger
substituent leads to a better outcome.
The report by Zhou and co-workers using chiral 1,2,3,4-
tetrahydroquinolinyloxazoline ligands in the Ru-catalyzed
asymmetric transfer hydrogentation of ketones was analyzed
(Figure 10).34 Excitingly, both the substituents on the ligand
and the ketone substrate could be correlated leading to linear
free energy relationships. Of interest, a significant negative ψ
value is observed for the substrate indicating that smaller
substrates perform better and a modest positive ψ value is
found for the ligand.
Potential Issues/Limitations of Charton Analysis. Consid-
ering Charton values were established for the rate of methyl
ester hydrolysis, which can be considered a mechanistically
simple reaction, it is exciting that direct application to
asymmetric catalysis was successful. However, as with any
type of LFER, there are issues and limitations that one must
consider when implementing this tool. Below are examples to
demonstrate some of the common limitations we found in
our investigations.
Complexity Associated with Phenyl Groups. Unlike acid-
catalyzed methyl ester hydrolysis, asymmetric catalysis can
be influenced by electronic factors, such as π-stacking, which
were not viable concerns in Taft’s original work. As men-
tioned above, this is observed in the fact that two Charton
values exist for Ph, 1.66 and 0.57. These two values are
believed to originate for the in-plane and out-of-plane
orientations of the phenyl group to the carbonyl. The in-
plane orientation (0.57) would offer very little steric repul-
sion to any incoming nucleophile to the carbonyl. If the
phenyl group is placed out-of-plane or perpendicular to the
carbonyl (1.66), the steric effect is greatly increased as
the nucleophile approaches the carbonyl, which could ac-
count for the slower rate of hydrolysis. As demonstrated
above, it is common to have to check if each of these values
when using Charton correlations, and one can imagine that
this conflict in calculated values can sometimes result in
unexplained lack of correlation of the Ph group.
FIGURE 8. Charton analysis of diethylzinc additions to benzalde-
hyde using TADDOL-based ligands.
reported use of a chiral 10-substituted oxazolinylferrocene as
a chiral ligand for an analogous reaction was analyzed.28 The
change in enantiomeric ratio as a factor of steric size was
more pronounced in this ligand class, which is indicated by
the larger ψ value (0.75).
Exploring other asymmetric organometallic reactions also
proved fruitful, where a LFER is observed for Seebach and
co-workers’ reported asymmetric addition of diethyl zinc to
benzaldehyde, catalyzed by chiral Ti-TADDOL derivatives
(Figure 8).29 In this case, a substantial ψ value (1.85) indi-
cates that this reaction is highly sensitive to the nature of the
ligand substituents. Also, it should be noted that the larger
value for Ph is used in this case (ν =1.66).
Negative ψ Values: Where Smaller Subtituents Lead to
Greater Selectivity. Chiral hydroxamic acids were also found
to correlate well (Figure 9) as demonstrated by Uang and co-
workers for the asymmetric epoxidation of allylic alcohols
using Vanadium salts.30 However, in this case the ψ value is
negative, consistent with a smaller ligand substituent leading
to higher enantiomeric ratio. This is in contrast to the
common conception that bigger substituents are generally
better in ligand structure. Two other examples are presented
to illustrate a negative ψ value. The first is Imamoto and co-
workers’ reported enantiomeric ratios for the asymmetric
hydrogenation of R-(acylamino)acrylic derivatives, cata-
lyzed by chiral Rh-BisP* complexes (Figure 9)31 wherein
the range of Charton values in this correlation (ν =
0.87-2.38) is noteworthy. The second is the enantioselective
cyclopropanation of allylic alcohols with bis(sulfonamide)
derived chiral diamines and bis(halomethyl) zinc reagents
Other Unusual Observations. In addition to Ph, there
are a number of substituents that have questionable Char-
(28) Park, J.; Quan, Z.; Lee, S.; Han Ahn, K.; Cho, C.-W. J. Organomet.
Chem. 1999, 584, 140–146.
n
ton values. Analyzing the steric bulk of Pr, Bu, and Bu
n
i
(29) Ito, Y. N.; Ariza, X.; Beck, A. K.; Bohac, A.; Ganter, C.; Gawley, R.
E.; Kuehnle, F. N. M.; Tuleja, J.; Wang, Y. M.; Seebach, D. Helv. Chim. Acta
1994, 77, 2071–2110.
(30) Wu, H.-L.; Uang, B.-J. Tetrahedron: Asymmetry 2002, 13, 2625–
2628.
(31) Imamoto, T.; Watanabe, J.; Wada, Y.; Masuda, H.; Yamada, H.;
Tsuruta, H.; Matsukawa, S.; Yamaguchi, K. J. Am. Chem. Soc. 1998, 120,
1635–1636.
(32) Denmark, S. E.; Christenson, B. L.; O’Connor, S. P. Tetrahedron
Lett. 1995, 36, 2219–2222.
(33) Minakata, S.; Ando, T.; Nishimura, M.; Ryu, I.; Komatsu, M.
Angew. Chem., Int. Ed. 1998, 37, 3392–3394.
(34) Zhou, Y.-B.; Tang, F.-Y.; Xu, H.-D.; Wu, X.-Y.; Ma, J.-A.; Zhou,
Q.-L. Tetrahedron: Asymmetry 2002, 13, 469–473.
J. Org. Chem. Vol. 74, No. 20, 2009 7639

Sigman and Miller
JOCFeatured Article
FIGURE 9. Charton analysis of various reaction types wherein a negative ψ value is observed.
substituents, we consider iBu to occupy more volume
n
n
than Pr or Bu. Surprisingly, all three substituents have a
Charton value of 0.68.12 Considering that these values were
calculated for the attack of a nucleophile on the carbonyl
of a methyl ester substrate this can be rationalized, as all
three groups contain a methylene carbon R to the carbo-
nyl. When attempting to employ these substituents in
correlations a nonlinear dependence can result. Charton
analysis of the report by Braga and co-workers’ asym-
metric addition of diethylzinc to aldehydes reveals an exam-
ple of nonlinearity (Figure 11).35 We found that a good
n
n
n
i
correlation is observed for the Pent, Hex, Hep, and Pr
substrates only if the nBu substrate was excluded. It is
difficult to rationalize why this substrate deviates to such
a large degree but again highlights the complex nature
of asymmetric catalysis. As an aside, Charton analysis
of the effect of the ligand substituent reveals a LFER with
a ψ value of -0.59.
“Break” in the Charton Plot. A break in the Charton plot
can also occur as demonstrated in our chemistry described
above where we attribute this to a change in ligand con-
formation. A similar result is observed by conducting a
Charton analysis on the chiral oxazoline-based ligand used
by Brunner and co-workers in their Rh-catalyzed asym-
metric hydrosilylation reaction (Figure 12).36 A linear free
energy relationship is observed for five substituents with a
t
ψ value of 0.48 only when the Bu substituent is excluded
from the analysis. One possible explanation for the signifi-
cant difference is the substituent size forces the chelation
environment about Rh to change where the change may
result in the ligand slipping from bidentate to monodentate.
On this issue, one must be very careful in analysis of limited
data sets. While we present a few examples above with only
three data points, the conclusions that arise from such a
data set must be taken very carefully, and it is suggested
that either substrates with a broad range of Charton values
(35) Braga, A. L.; Rubim, R. M.; Schrekker, H. S.; Wessjohann, L. A.; de
Bolster, M. W. G.; Zeni, G.; Sehnem, J. A. Tetrahedron: Asymmetry 2003, 14,
3291–3295.
FIGURE 10. Charton analysis in the effect of substrate size on the
enantioselective outcome.
(36) Brunner, H.; Stoeriko, R. Eur. J. Inorg. Chem. 1998, 783–788.
7640 J. Org. Chem. Vol. 74, No. 20, 2009

Sigman and Miller
JOCFeatured Article
catalytic processes through linear free energy relationships.
To this end, we have illustrated that linear free energy
relationships can be constructed by plotting log of enantio-
meric ratio versus Charton steric parameters for multiple
substrates using modular oxazoline ligands developed in our
laboratory for aldehyde and ketone allylation under NHK
conditions. In an attempt to design an improved catalyst for
carbonyl allylation, we found incorporation of a larger
group into our ligand structure resulted in an unanticipated
break in the linearity of the Charton plot. Interestingly, a
new LFER is observed after the break most likely indicating
the size of the substituent forces the catalyst to adopt a
different conformation or constitution.
We have also been able to extend the Charton analysis to
the optimal catalyst for ketone allylation under NHK con-
ditions developed in our laboratory where, in this case, the
sensitivity as measured by the ψ value is greatest for the
allylation of acetophenone in comparison to benzaldehyde.
This is contrast to ligand 1a. While understanding the origin
of these effects is of great importance, the LFER observed
for these catalysts will be clearly central as we utilize this
ligand class in optimizing catalysts for difficult substrate
classes and extending it to new reaction development.
Finally, we have investigated the use of Charton analysis
for previously reported enantioselective processes. LFERs
can be successfully observed for a wide variety of ligands and
reaction types indicating the potential broad applicability of
the quantitative use of steric parameters in asymmetric
catalyst design. It should be pointed out that issues exist in
terms of the accuracy of some of the Charton values. How-
ever, in our opinion, insight can be garnered not only from
the successful correlations but also from the outliers. Future
work in our laboratory is focused on applying the quantita-
tive use of steric parameters to aid in the design of improved
and new catalysts for asymmetric catalytic reactions as well
as understanding the structural and mechanistic origins of
the observed linear free energy relationships.
FIGURE 11. Charton analysis of a diethylzinc addition to alde-
hyde derivatives using an oxazoline-derived ligand.
Experimental Section
Preparation of Heptan-4-yl 4-Nitrophenyl Carbonate (SI-1).
To a stirring solution of 4-heptanol (495 μL, 4.26 mmol,
1 equiv), CH2Cl2 (10 mL), and pyridine (520 μL, 6.39 mmol,
1.5 equiv) at -5 °C was added p-nitrophenyl chloroformate
(1.03 g, 5.11 mmol, 1.2 equiv) in four equal portions. The
reaction flask was removed from the cooling bath and allowed
to warm to room temperature with stirring. After the reaction
mixture was stirred for 3 h, the precipitated pyridine hydro-
chloride salt was removed by filtration; H2O (15 mL) was added
to the filtrate, and the solution was extracted with Et2O (3 ꢀ
30 mL). The organics were washed with 1 M HCl (3 ꢀ 30 mL),
saturated aqueous NaHCO3 (2 ꢀ 30 mL), and saturated aqu-
eous NaCl (30 mL). The organics were dried over Na2SO4,
filtered, and then concentrated under reduced pressure. The
crude mixture was taken on without further purification: Rf =
0.68 (33% EtOAc/hexanes, stained with KMnO4).
FIGURE 12. Charton analysis of a hydrosilylation reaction of
substituted benzaldehyde derivatives using a pyridyl-oxazoline
ligand.
Preparation of (S)-2-Benzyl 1-Heptan-4-ylpyrrolidine-1,2-
dicarboxylate (SI-4). To a stirring solution of SI-1 (1.11 g, 3.96
mmol, 1.1 equiv), THF (6 mL), and saturated aqueous NaHCO3
(6 mL) was slowly added Bn-Pro-OH (870 mg, 3.6 mmol,
1 equiv) in three equal portions. The reaction mixture was
then heated to 50 °C. After being stirred for 12 h, the mixture
was concentrated under reduced pressure to remove THF. The
resulting mixture was then extracted with CH2Cl2 (3 ꢀ 40 mL). In
order to remove the residual p-nitro phenol, the organics were
or a significant number of substrates with a limited range of
Charton values are evaluated to achieve statistically robust
results.
Conclusions
In summary, we have applied steric parameters originally
developed by Taft to quantitatively analyze asymmetric
J. Org. Chem. Vol. 74, No. 20, 2009 7641

Sigman and Miller
JOCFeatured Article
washed repeatedly with 10% aqueous NaHCO3 until the yellow
color did not persist. The organic layer was dried over Na2SO4,
filtered, and then concentrated under reduced pressure. Purifica-
tion was accomplished by flash chromatography on a 3.5 ꢀ 13 cm
column, eluting with 8% EtOAc/hexanes, collecting 9 mL frac-
tions. The product-containing fractions were combined and
concentrated under reduced pressure to give the desired product
SI-4 (950 mg, 76% yield) as a clear viscous oil: Rf = 0.47 (33%
EtOAc/hexanes, stained with KMnO4); [R]23D -38.2 (c = 0.55,
CHCl3); 300 MHz 1H NMR (CDCl3) at 50 °C δ 7.36-7.29 (m,
5H), 5.26-5.00 (m, 2H), 4.85-4.70 (m, 1H), 4.49-4.27 (m, 1H),
3.65-3.36 (m, 2H), 2.31-2.10 (m, 1H), 2.05-1.77 (m, 3H),
1.63-1.16 (m, 8H), 0.96-0.79 (m, 6H); 75 MHz 13C NMR
(CDCl3) at 50 °C δ 172.7, 154.5, 135.8, 128.5, 128.2, 127.9,
75.0, 66.5, 58.8, 46.7, 36.4, 30.9, 23.4, 18.4, 13.9; IR (neat) 2957,
2934, 2873, 1746, 1698, 1406, 1336, 1162, 1114, 1086, 982,
737, 696 cm-1; HRMS (ES) calcd. C20H29NO4Na (M þ Na)þ
370.1994, obsd. 370.1995.
Preparation of Benzyl (S)-1-((R)-4-Benzyl-4,5-dihydrooxazol-
2-yl)-2-methylpropylcarbamate (SI-7). To a stirring solution of
Cbz-Val-OH (2.06 g, 8.19 mmol, 1 equiv) and CH2Cl2 (33 mL),
at -5 °C, was slowly added N-methylmorpholine (1.08 mL, 9.83
mmol, 1.2 equiv) via syringe. After the solution was stirred for
10 min at -5 °C, isobutyl chloroformate (1.08 mL, 8.19 mmol,
1.0 equiv) was added dropwise, via syringe, to the reaction
mixture. After the solution was stirred for an additional
45 min at -5 °C, H-D-phenylalaninol (1.36 g, 8.99 mmol,
1.1 equiv) was added in ca. three equal portions to the reaction
mixture. The reaction mixture was then allowed to warm to
room temperature. After 2 h, TLC analysis indicated complete
consumption of starting material. The reaction was quenched by
addition of 1 M HCl (15 mL) and diluted with CH2Cl2 (40 mL).
The layers were separated, and the aqueous layer was removed.
The organic layer was washed with H2O (3 ꢀ 20 mL) and brine
(10 mL). The organic layer was dried over Na2SO4, filtered, and
then concentrated under reduced pressure. The crude mixture
was taken on without further purification.
HRMS (ES) calcd. C22H27N2O3 (M þ H)þ 367.2022, obsd.
367.2019.
Preparation of (S)-Heptan-4-yl 2-((S)-1-((R)-4-Benzyl-4,5-
dihydrooxazol-2-yl)-2-methylpropylcarbamoyl)pyrrolidine-1-
carboxylate (1f). To a stirring solution of (S)-1-((heptan-4-
yloxy)carbonyl)pyrrolidine-2-carboxylic acid (obtained by de-
protection of the SI-4 with 10 mol % Pd/C at 1 atm H2 for 5 h)
(141 mg, 0.55 mmol, 1 equiv) and CH2Cl2 (3 mL), at -5 °C, was
slowly added N-methylmorpholine (72 μL, 0.66 mmol, 1.2
equiv) via syringe. After the solution was stirred for 20 min at
-5 °C, isobutyl chloroformate (72 μL, 0.55 mmol, 1.0 equiv) was
added dropwise, via syringe, to the reaction mixture. After the
solution was stirred for an additional 45 min at -5 °C, the free
oxazoline amine (140 mg, 0.60 mmol, 1.1 equiv), obtained by
deprotection of the SI-7 shown above (10 mol % Pd/C at 1 atm
H2 for 5 h), was added dropwise to the reaction mixture as a
solution in CH2Cl2 (0.5 mL). The reaction mixture was then
allowed to warm to room temperature. After 3 h, TLC analysis
indicated complete consumption of starting material. The reac-
tion was quenched by addition of 1 M HCl (1 mL) and diluted
with CH2Cl2 (6 mL). The layers were separated, and the aqueous
layer was removed. The organic layer was washed with H2O (3 ꢀ
5 mL) and brine (5 mL). The organic layer was dried over
Na2SO4, filtered, and concentrated under reduced pressure.
Purification was accomplished by flash chromatography on a
2 ꢀ 15 cm column, eluting first with 200 mL of 50% EtOAc/
hexanes, followed by 40% acetone/hexanes, collecting 9 mL
fractions. The product containing fractions were combined and
concentrated under reduced pressure to give the desired ligand
(119 mg, 46% yield) as a clear colorless oil: Rf = 0.30 (66%
EtOAc/hexanes, stained with PMA); [R]23 -36.7 (c = 0.39,
D
1
CHCl3); 400 MHz H NMR (CDCl3) at 50 °C δ 7.29 (t, J=
7.2 Hz, 1H), 7.21 (d, J = 9.2 Hz, 2H), 7.19 (d, J = 6.8 Hz, 2H),
4.83 (t, J = 5.6 Hz, 1H), 4.57 (dd, J=8.6, 5.6 Hz, 1H), 4.42-
4.29 (m, 2H), 4.20 (t, J = 8.8 Hz, 1H), 3.97 (t, J = 8.2 Hz, 1H),
3.58-3.33 (m, 2H), 3.12 (dd, J = 13.9, 5.1 Hz, 1H), 2.62 (dd, J =
13.9, 8.9 Hz, 1H), 2.44-1.80 (m, 4H), 1.63-1.21 (m, 7H),
0.97-0.85 (m, 14H); 100 MHz 13C NMR (CDCl3) at 50 °C δ
171.4, 166.3, 156.3, 137.8, 129.3, 128.5, 126.5, 75.6, 72.1, 67.1,
60.3 52.4, 46.9, 41.7, 36.6, 31.6, 27.8, 24.5, 18.5, 17.7, 14.0; IR
(neat) 3314, 2958, 2933, 2872, 1690, 1664, 1533, 1412, 1190,
1116, 982, 702 cm-1; HRMS (ES) calcd C27H41N3O4Na (M þ
Na)þ 494.2995, obsd 494.2990.
To a stirring solution of the crude mixture in CH2Cl2 (60 mL)
and triethylamine (40 mL), at -5 °C, was slowly added
p-toluenesulfonyl chloride (1.85 g, 9.83 mmol, 1.2 equiv) in ca.
three equal portions. After addition was complete, the flask was
removed from the cooling bath and the reaction mixture was
allowed to warm to room temperature. After ca. 45 min, TLC
analysis indicated complete conversion of the starting material
to the p-toluenesulfonate. The reaction mixture was heated at
reflux for 15 h. The reaction mixture was allowed to cool to
room temperature before diluting with CH2Cl2 (25 mL). To this
mixture was slowly added saturated aqueous NaHCO3 (25 mL).
The resulting organic layer was washed with saturated aqueous
NaHCO3 (2 ꢀ 10 mL), H2O (20 mL), and brine (15 mL). The
organic layer was dried over Na2SO4, filtered, and then con-
centrated under reduced pressure. Purification was accom-
plished by flash chromatography on a 4.75 ꢀ 13 cm column,
eluting first with 300 mL of 50% Et2O/hexanes, followed by
35% acetone/hexanes, collecting 18 mL fractions. The product
containing fractions were combined and concentrated under
reduced pressure to give the desired Cbz protected oxazoline
amine (2.52 g, 84% yield over two steps) as a white solid: mp
96-98 °C; Rf = 0.57 (35% acetone/hexanes, stained with
phosphomolybdic acid); [R]23D -11.7 (c = 0.325, CHCl3); 400
MHz 1H NMR (CDCl3) at 50 °C δ 7.37-7.17 (m, 10H), 5.39 (d,
J = 8.8 Hz, 1H), 5.15 (d, J = 12.3 Hz, 1H), 4.44-4.35 (m, 2H),
4.23 (t, J = 8.8 Hz, 1H), 4.00 (t, J = 8.0 Hz, 1H), 3.08 (dd, J =
13.7, 5.1 Hz, 1H), 2.63 (dd, J = 13.7, 8.4 Hz, 1H), 2.11 (m, 1H),
Preparation of (S)-Isopropyl 2-((R)-1-methoxy-3-methyl-1-
oxobutan-2-ylcarbamoyl)pyrrolidine-1-carboxylate (SI-8). To a
stirring solution of (S)-1-(isopropoxycarbonyl)pyrrolidine-2-
carboxylic acid (103 mg, 0.51 mmol, 1 equiv) in CH2Cl2 (2 mL),
at -5 °C, was slowly added N-methylmorpholine (84 μL,
0.76 mmol, 1.5 equiv) via syringe. After the solution was stirred
for 20 min at -5 °C, isobutyl chloroformate (67 μL, 0.51 mmol,
1 equiv) was added dropwise, via syringe, to the reaction mix-
ture. After the solution was stirred for an additional 30 min at
-5 °C, N-methylmorpholine (64 μL, 0.58 mmol, 1.15 equiv) was
added slowly, via syringe, to the reaction mixture. This was
followed by the addition of R-valine methyl ester hydrochloride
(94 mg, 0.56 mmol, 1.1 equiv) in ca. three equal portions to the
reaction mixture. The reaction mixture was then allowed to
warm to room temperature. After 2 h, TLC analysis indicated
complete consumption of starting material. The reaction was
quenched by addition of 1 M HCl (5 mL) and diluted with
CH2Cl2 (5 mL). The layers were separated, and the aqueous
layer was removed. The organic layer was washed with H2O (3 ꢀ
10 mL) and brine (5 mL). The organic layer was dried over
Na2SO4, filtered, and then concentrated under reduced pres-
sure. Purification was accomplished by mixed solvent recrystal-
lization (1:4, Et2O/hexanes) to yield SI-8 as colorless crystals
(138 mg, 86% yield): mp 85-86 °C; Rf = 0.40 (66% EtOAc/
hexanes, stained with KMnO4); [R]23D -81.8 (c = 2.89, CHCl3);
0.95 (d, J = 6.8 Hz, 3H), 0.90 (d, J = 6.8 Hz, 3H); 100 MHz 13
C
NMR (CDCl3) at 50 °C δ 166.7, 156.1, 137.6, 136.4, 129.2,
128.5, 128.4, 128.1, 128.0, 126.5, 72.2, 67.0, 66.9, 54.3, 41.7, 31.5,
18.8, 17.4; IR (thin film) 2977, 2935, 2863, 1382, 1123 cm-1
;
7642 J. Org. Chem. Vol. 74, No. 20, 2009

Sigman and Miller
JOCFeatured Article
500 MHz 1H NMR (CDCl3) at 50 °C δ 4.96 (pent, J = 6.3 Hz,
1H), 4.53 (dd, J = 8.8, 4.9 Hz, 1H), 4.36 (d, J = 7.8 Hz, 1H), 3.72
(s, 3H), 3.57-3.36 (m, 2H), 2.35-1.84 (m, 5H), 1.27 (t, J = 5.4
Hz, 6H), 0.96 (d, J = 6.8 Hz, 3H), 0.91 (d, J = 6.8 Hz, 3H); 125
MHz 13C NMR (CDCl3) at 50 °C δ 172.3, 172.1, 155.9, 69.4,
60.8, 57.2, 52.0, 47.2, 30.8, 29.4, 24.2, 22.3, 19.1, 17.8; IR (neat)
3262, 3070, 2973, 1742, 1710, 1659, 1552, 1412, 1202, 1176, 1143,
1122, 768 cm-1; HRMS (ES) calcd C15H27N2O5 (M þ H)þ
315.1920, obsd 315.1919.
Preparation of (S)-Isopropyl 2-((R)-1-(2-Hydroxyethylamino)-
3-methyl-1-oxobutan-2-ylcarbamoyl)pyrrolidine-1-carboxylate
(SI-12). To a stirring solution of SI-8 (141 mg, 0.45 mmol,
1 equiv), toluene (800 μL), and tetrahydrofuran (800 μL) was
added 2-aminoethanol (137 μL, 2.28 mmol, 5 equiv) via syringe.
After 3 d at reflux the mixture was diluted with CHCl3 (10 mL).
The organic layer was then washed with H2O (3 ꢀ 10 mL) and
brine (10 mL). The organic layer was dried over Na2SO4,
filtered, and then concentrated under reduced pressure. Purifi-
cation was accomplished by mixed solvent recrystallization (1:5
CH2Cl2/hexanes) to yield a white solid (141 mg, 91% yield): mp
6.3 Hz, 3H), 1.19 (d, J = 6.3 Hz, 3H), 0.89 (d, J = 6.8 Hz, 3H),
0.86 (d, J = 6.8 Hz, 3H); 125 MHz 13C NMR (CDCl3) at 50 °C δ
172.1, 167.2, 155.9, 69.5, 68.2, 61.2, 54.5, 52.8, 47.5, 31.9, 31.2,
29.7, 24.3, 22.6, 22.5, 19.2, 18.0; IR (neat) 3213, 3045, 2960,
2874, 1687, 1657, 1549, 1418, 1380, 1209, 1110, 933, 769, 697
cm-1; HRMS (ES) calcd C16H27N3O4Na (M þ Na)þ 348.1899,
obsd 348.1908.
Preparation of Homoallylic Alcohols. To a 1.5 dram vial were
added CrCl3 (7.9 mg, 0.05 mmol, 0.1 equiv), Mn(0) (325 mesh)
(55 mg, 1.0 mmol, 2 equiv), and 2a (17 mg, 0.05 mmol, 0.1 equiv).
The vial was fitted with a permeable Teflon cap, charged with
tetrahydrofuran (2.5 mL), and allowed to stir. After ca. 2 min,
triethylamine (15 μL, 0.1 mmol, 0.2 equiv) and chlorotrimethyl-
silane (255 μL, 2 mmol, 4 equiv) were added to the reaction
mixture. After the solution had stirred for 20 min and the color
had changed to a deep blue/purple color, allyl bromide (85 μL,
1.0 mmol, 2 equiv) was added to the reaction mixture. After
30 min, carbonyl (0.5 mmol, 1 equiv) was added, and the mixture
was allowed to stir at room temperature. After 24 h, the reaction
was quenched by addition of a saturated aqueous NaHCO3
solution (1.5 mL). Caution: Take care to add the NaHCO3
solution dropwise as a large amount of gas evolves. After being
stirred for 20 min, the mixture was passed through a Celite plug,
eluting with diethyl ether. The layers were separated, and the
aqueous layer was extracted with diethyl ether (3 ꢀ 10 mL). The
combined organic layers were dried over Na2SO4, filtered, and
then concentrated under reduced pressure. Purification was
accomplished by flash chromatography on a 2 ꢀ 15 cm column,
eluting with 8% EtOAc/hexanes, collecting 9 mL fractions. The
product containing fractions were combined and concentrated
under reduced pressure to give the desired homoallylic alcohol
as a colorless oil (64-97% yield). The enantiomeric excess of
acetophenone allylation was determined by HPLC analysis
(Chiralcel OJ-H, hexanes/i-PrOH = 97/3, flow rate=1.0 mL/
min): tmajor=9.6 min (R), tminor=11.5 min (S). The enantiomeric
excess of benzaldehyde allylation was determined by HPLC
analysis (Chiralcel OD, hexanes/i-PrOH = 98/2, flow rate =
1.0 mL/min): tmajor=15.9 min (R), tminor=18.4 min (S).1 The
enantiomeric excess of hydrocinnamaldehyde allylation was de-
termined by HPLC analysis (Chiralcel OD, hexanes/i-PrOH=
185-187 °C; Rf = 0.43 (10% MeOH/CH2Cl2, KMnO4); [R]23
D
þ1.6 (c = 0.49, CHCl3); 500 MHz 1H NMR (CDCl3) at 50 °C δ
4.52 (dd, J = 8.8, 4.9 Hz, 1H), 4.32 (d, J = 5.9 Hz, 1H), 3.72 (s,
3H), 3.53-3.32 (m, 2H), 2.34-1.81 (m, 14H), 1.66 (bs, 6H), 0.97
(d, J = 6.8 Hz, 3H), 0.92 (d, J = 6.8 Hz, 3H); 125 MHz 13C
NMR (CDCl3) at 50 °C δ 172.4, 80.5, 60.6, 57.3, 52.1, 47.3, 42.0,
36.3, 31.2, 30.6, 24.2, 19.2, 17.9; IR (neat) 3285, 2966, 2872,
1664, 1638, 1546, 1421, 1383, 1345, 1231, 1173, 1109, 930, 768
cm-1; HRMS (ES) calcd C16H29N3O5Na (M þ Na)þ 366.2005,
obsd 366.2005.
Preparation of (S)-Isopropyl 2-((R)-1-(4,5-Dihydrooxazol-2-
yl)-2-methylpropylcarbamoyl)pyrrolidine-1-carboxylate (3c). To
a stirring solution of SI-12 (117 mg, 0.34 mmol, 1 equiv) and
tetrahydrofuran (2.25 mL) was added triphenylphosphine
(108 mg, 0.41 mmol, 1.2 equiv) in one portion. This was followed
by dropwise addition of diisopropyl azodicarboxylate (86 μL,
0.41 mmol, 1.2 equiv), via syringe, to the reaction mixture. The
reaction mixture was allowed to clear between drops. The
progress of the reaction was monitored by TLC analysis. After
2 h of stirring, the mixture was concentrated under reduced
pressure. Purification was accomplished by flash chromatogra-
phy on a 1.5 ꢀ 12 cm column, eluting with 75 mL or 35%
EtOAc/hexanes followed by 40% acetone/hexanes, collecting
9 mL fractions. The product-containing fractions were com-
bined and concentrated under reduced pressure. The solid was
then recrystallized by mixed solvent recrystallization (1:4 acet-
one/hexanes) to yield a colorless solid (97 mg, 88% yield): mp
112-114 °C; Rf = 0.31 (50% acetone/hexanes, ninhydrine);
[R]23D -71.7 (c = 0.29, CHCl3); 500 MHz 1H NMR (CDCl3) at
50 °C δ 4.88 (sept, J = 6.3 Hz, 1H), 4.54 (dd, J = 8.8, 4.9 Hz,
1H), 4.34-4.25 (m, 1H), 4.20 (sext, J = 9.3 Hz, 2H), 3.76 (t, J =
9.8 Hz, 2H), 3.53-3.27 (m, 2H), 2.26-1.76 (m, 5H), 1.20 (d, J =
98/2, flow rate = 1.0 mL/min): tmajor = 19.3 min (S), tminor
34.5 min (R).
=
Acknowledgment. This work was supported by the Na-
tional Science Foundation (CHE-0749506). M.S.S. thanks
the Dreyfus foundation (Teacher-Scholar) and Pfizer for
their support.
Supporting Information Available: Ligand analogue syn-
1
thesis and H and 13C spectra. This material is available free of
charge via the Internet at http://pubs.acs.org
J. Org. Chem. Vol. 74, No. 20, 2009 7643