Asymmetric induction in methyl ketone aldol additions to α-alkoxy and α,β-bisalkoxy aldehydes: A model for acyclic stereocontrol

Article abstract of DOI:10.1021/ja061010o

A systematic study of methyl ketone aldol additions under nonchelating conditions with α-alkoxy and α,β-bisalkoxy aldehydes is described. Additions to aldehydes containing a single α-alkoxy stereocenter generally provide the product diastereomers in accor

Full text of DOI:10.1021/ja061010o

Published on Web 07/06/2006
Asymmetric Induction in Methyl Ketone Aldol Additions to
r-Alkoxy and r,â-Bisalkoxy Aldehydes: A Model for Acyclic
Stereocontrol
David A. Evans,* Victor J. Cee,† and Sarah J. Siska†
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
Received February 20, 2006; E-mail: evans@chemistry.harvard.edu
Abstract: A systematic study of methyl ketone aldol additions under nonchelating conditions with R-alkoxy
and R,â-bisalkoxy aldehydes is described. Additions to aldehydes containing a single R-alkoxy stereocenter
generally provide the product diastereomers in accord with the Cornforth/polar Felkin-Anh models for
carbonyl addition. Vicinal asymmetric induction is sensitive to the aldehyde R-alkyl substituent, but is relatively
insensitive to the nature of the alkoxy protecting group. Aldehyde π-facial selectivity in additions to substrates
containing an additional â-alkoxy-substituted stereocenter exhibits a striking dependence on the relative
configuration of the R- and â-stereocenters. Aldehydes with the R- and â-alkoxy substituents in an anti
relationship in most cases exhibit good diastereoselectivity, while aldehydes with the R- and â-alkoxy
substituents in a syn relationship unexpectedly give product mixtures. A stereochemical model based on
Cornforth-like transition-state arrangements is proposed.
Introduction
enolates. For the anti-aldehyde diastereomers, both R- and
â-stereocenters reinforce addition to the mutually preferred
Asymmetric induction in nucleophilic carbonyl addition
reactions is a powerful control element for the selective
construction of new stereocenters.1 In this article, a systematic
study of stereocontrol in aldol addition reactions of methyl
ketone-derived enolates and aldehydes containing single and
multiple alkoxy stereocenters is presented (eqs 1 and 2). The
alkoxy protecting group, enolate type, and enolate steric
hindrance are systematically varied to give a comprehensive
picture of asymmetric induction in methyl ketone aldol additions
to alkoxy-substituted aldehydes.
aldehyde π-face. In contrast, the corresponding syn-aldehyde
diastereomers exhibit variable selectivity depending on enolate
structure. In this instance, there is a nonreinforcing interplay
between dipolar (â-OR) and steric effects (R-Me) where
sterically demanding enolates respond to steric control and
“smaller” enolates are controlled by dipolar effects.
In previous studies we documented that both R-alkyl and
â-alkoxy stereocenters play contributing roles in dictating
aldehyde π-face selectivity in enol/enolate nucleophilic addition
reactions (eq 3).2 In these processes, the anti diastereomer
exhibits uniformly high diastereoselection with unsubstituted
The goal of this investigation is to determine whether similar
correlations might exist in aldol additions with the corresponding
syn- and anti-R,â-bisalkoxy aldehyde diastereomers (eq 4). Since
these reactions reflect an emerging approach to the synthesis
of extended polyol natural products such as carbohydrates (eq
5),3 the documentation of trends in aldehyde face selectivity
represents an important aspect of this assemblage strategy. In
the following discussion, the numbering system that will be used
† This work is taken in part from the Ph.D. theses of V. J. Cee, Harvard
University, 2003 and S. J. Siska, Harvard University, 2005.
(1) Mengel, A.; Reiser, O. Chem. ReV. 1999, 99, 1191-1223, and references
therein.
(2) Evans, D. A.; Dart, M. J.; Duffy, J. L.; Yang, M. G. J. Am. Chem. Soc.
1996, 118, 4322-4343.
9
10.1021/ja061010o CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 9433-9441
9433

A R T I C L E S
Evans et al.
Chart 1. R-Alkoxy Aldehyde and Enolate Structures
is based on the hexose open-chain product tautomer where the
carbonyl moiety is designated as C1. Accordingly, the carbonyl
center in the aldehyde precursors is designated as C3 (eq 5).
r-Alkoxy Aldehydes. It is well established that a carbonyl
with an adjacent R-alkoxy substituent reacts with a characteristic
stereochemical bias in the absence of chelate organization.4,5
Under such conditions, most nucleophilic additions proceed with
bias for the product diastereomer containing an anti configu-
ration between the newly formed hydroxyl moiety and the
vicinal oxygen heteroatom (Figure 1). Both the Cornforth6 and
polar Felkin-Anh7 transition-state models account for the
preferential formation of the 1,2-anti product diastereomer on
the basis of differing transition state control elements (Figure
1). Recent experimental8 and theoretical9 evidence indicates that
the Cornforth model more accurately describes asymmetric
induction in enolborane additions to R-alkoxy aldehydes.
â-Alkoxy Aldehydes. It is well established that nucleophilic
additions to â-alkoxy aldehydes result in the preferential
formation of the 3,5-anti product diastereomers (eq 6).12 The
selectivity is dependent on the type of enolate nucleophile, with
high levels of selectivity observed in Lewis acid-promoted
additions, moderate levels of selectivity observed in lithium
enolate additions, and little selectivity noted in enolborane
addition reactions. A transition state model based on minimiza-
tion of electrostatic and steric effects has been proposed to
account for the observed sense of 1,3-asymmetric induction
(Figure 2). The 3,5-anti product is proposed to arise from
transition structure C, in which nucleophilic attack occurs anti
to the R-carbon substituent, with the â-stereocenter oriented to
minimize both destabilizing gauche interactions of the â-alkyl
substituent and destabilizing electrostatic interactions between
the â-C-O and CdO dipoles.13 The most likely transition
structures for the formation of the syn product contain either
an unfavorable alignment of C-O and CdO dipoles (D), or an
unfavorable gauche arrangement of the â-alkyl substituent with
the reacting carbonyl (E).14,15
Figure 1. Nucleophilic addition models for R-alkoxy aldehydes.
Aldol additions between unsubstituted enolates and R-alkoxy
aldehydes generally favor the formation of the anti product
diastereomer, although significant variations in the magnitude
of asymmetric induction have been reported.10 To generate an
internally consistent data set for this study, R-alkoxy aldehydes
1-311 were studied in aldol addition reactions with methyl
ketone-derived nucleophiles of three distinct types (Chart 1).
Enolsilanes (M ) TMS), enolboranes (M ) 9-BBN), and
lithium enolates (M ) Li) derived from acetone, 3-methyl-2-
butanone, and pinacolone were studied to evaluate the influence
of both the type of enolate and the effects of nucleophile steric
hindrance on diastereofacial selectivity.
(8) The relationship between enolborane geometry and diastereofacial selectivity
in additions to R-alkoxy aldehydes has been interpreted as supporting a
modified Cornforth model: (a) Evans, D. A.; Siska, S. J.; Cee, V. J. Angew.
Chem., Int. Ed. 2003, 42, 1761-1765. For an alternative explanation based
on substituted allylborane additions to R-alkoxy aldehydes, see: (b) Roush,
W. R. In Houben-Weyl; Helmchen, G., Hoffmann, R. W., Mulzer, J.,
Schaumann, E. Eds.; Thieme: Stuttgart, 1995; Vol. E21, pp 1410-1486.
For additional examples of substituted enolborane and allylborane additions
to R-alkoxy aldehydes, see: (c) Hoffmann, R. W. Chem. Scr. 1985, 25
(Special Issue), 53-60. (d) Roush, W. R.; Adam, M. A.; Walts, A. E.;
Harris, D. J. J. Am. Chem. Soc. 1986, 108, 3422-3434. (e) Williams, D.
R.; Moore, J. L.; Yamada, M. J. Org. Chem. 1986, 51, 3916-3918. (f)
Hoffmann, R. W.; Metternich, R.; Lanz, J. W. Liebigs Ann. Chem. 1987,
881-887. (g) Wuts, P. G. M.; Bigelow, S. S. J. Org. Chem. 1988, 53,
5023-5034. (h) Brinkmann, H.; Hoffmann, R. W. Chem. Ber. 1990, 123,
5-2401. (i) Hu, S.; Jayaraman, S.; Oehlschlager, A. C. J. Org. Chem. 1998,
63, 8843-8849.
(3) (a) Davies, S. G.; Nicholson, R. L.; Smith, A. D. Synlett 2002, 1637-
1640. (b) Northrup, A. B.; MacMillan, D. W. C. Science 2004, 305, 1753-
1755. (c) Davies, S. G.; Nicholson, R. L.; Smith, A. D. Org. Biomol. Chem.
2004, 2, 3385-3400. (d) Davies, S. G.; Nicholson, R. L.; Smith, A. D.
Org. Biomol. Chem. 2005, 3, 348-359. (e) Casas, J.; Engqvist, M.; Ibrahem,
I.; Kaynak, B. Cordova, A. Angew. Chem., Int. Ed. 2005, 44, 1343-1345.
(f) Timmer, M. S. M.; Adibekian, A.; Seeberger, P. H. Angew. Chem., Int.
Ed. 2005, 44, 7605-7607.
(9) Cee, V. J.; Cramer, C. J.; Evans, D. A. J. Am. Chem. Soc. 2006, 128, 2920-
2930.
(4) For a review of chelation-controlled nucleophilic additions, see: (a) Reetz,
M. T. Angew. Chem., Int. Ed. Engl. 1984, 23, 556-569. (b) Reetz, M. T.
Acc. Chem. Res. 1993, 26, 462-468. For experimental evidence of chelates
as reactive intermediates see: (c) Chen, X.; Hortelano, E. R.; Eliel, E. L.;
Frye, S. V. J. Am. Chem. Soc. 1992, 114, 1778-1784.
(10) Lithium enolate addition: (a) Heathcock, C. H.; Young, S. D.; Hagen, J.
P.; Pirrung, M. C.; White, C. T.; VanDerveer, D. J. Org. Chem. 1980, 45,
3846-3856. (b) Lodge, E. P.; Heathcock, C. H. J. Am. Chem. Soc. 1987,
109, 3353-3361. (c) Mahler, U.; Devant, R. M.; Braun, M. Chem. Ber.
1988, 121, 2035-2044. (d) Dondoni, A.; Merino, P. J. Org. Chem. 1991,
56, 5294-5301. Mukaiyama aldol: (e) Heathcock, C. H.; Davidsen, S.
K.; Hug, K. T.; Flippin, L. A. J. Org. Chem. 1986, 51, 3027-3037. (f)
Kita, Y.; Tamura, O.; Itoh, F.; Yasuda, H.; Kishino, H.; Ke, Y. Y.; Tamura,
Y. J. Org. Chem. 1988, 53, 554-561. (g) Gennari, C. In ComprehensiVe
Organic Synthesis; Heathcock, C. H., Ed.; Pergamon Press: Oxford, 1991;
Vol. 2, pp 640-647. Enolborane addition: (h) Gennari, C.; Bernardi, A.;
Cardani, S.; Scolastico, C. Tetrahedron 1984, 40, 4059-4065.
(5) For a study of chelation-controlled enolate additions see: Evans, D. A.;
Allison, B. D.; Yang, M. G.; Masse, C. E. J. Am. Chem. Soc. 2001, 123,
10840-10852.
(6) (a) Cornforth, J. W.; Cornforth, R. H.; Mathew, K. K. J. Chem. Soc. 1959,
112-127. The Cornforth model discussed here is modified from its original
form to incorporate contemporary concepts of a staggered arrangement
about the forming C-Nu bond and a >90° angle of attack for the incoming
nucleophile. This is often referred to as the Dunitz-Bu¨rgi angle: (b) Bu¨rgi,
H. B.; Dunitz, J. D.; Shefter, E. J. Am. Chem. Soc. 1973, 95, 5065-5067.
(c) Bu¨rgi, H. B.; Dunitz, J. D.; Lehn, J. M.; Wipff, G. Tetrahedron 1974,
30, 1563-1572.
(7) (a) Che´rest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968, 9, 2199-
2204. (b) Che´rest, M.; Felkin, H. Tetrahedron Lett. 1968, 9, 2205-2208.
(c) Anh, N. T.; Eisenstein, O. NouV. J. Chim. 1977, 1, 61-70. (d) Anh, N.
T. Top. Curr. Chem. 1980, 88, 145-162.
(11) For experimental details concerning the construction of aldehydes 1-17,
and stereochemical proofs of the products, see the Supporting Information.
(12) Evans, D. A.; Duffy, J. L.; Dart, M. J. Tetrahedron Lett. 1994, 35, 8537-
8540, and references therein.
(13) This assumption is supported by semiempirical calculations of ground-
state aldehyde conformations. For aldehyde-BF3 complexes (AM1), see
ref 2. For uncomplexed aldehydes (AM1 and PM3), see: Bonini, C.;
Esposito, V.; D’Auria, M.; Righi, G. Tetrahedron 1997, 53, 13419-13426.
9
9434 J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006

Model for Acyclic Stereocontrol
A R T I C L E S
Chart 2. R,â-Alkoxy Aldehyde Structures Studied
Figure 2. 1,3-Polar model for asymmetric induction in â-alkoxy aldehydes.
r,â-Bisalkoxy Aldehydes. There is no unified body of
literature that might be used to predict the diastereoselectivities
resulting from nucleophilic additions to the diastereomeric syn-
and anti-aldehyde diastereomers depicted below.
whereas the para-methoxybenzyl (PMB) and TBS protecting
groups were chosen for the â-oxygen substituent in accord with
the study of 1,3-asymmetric induction. Three additional aldehyde
structural types were selected to evaluate the effect of an
unbranched â-alkyl group (12,13), a cyclic protecting group
(14,15), and a methyl-substituted â-stereocenter (16,17).
Results and Discussion
r-Alkoxy Aldehydes. The π-facial selectivities exhibited by
aldehydes 1-3 in aldol addition reactions with methyl ketone-
derived enolates is presented in Table 1. In the majority of cases,
the 3,4-anti product diastereomer predominates in accord with
expectation. Since the 3,4-syn adduct may be rationalized
through a chelate-controlled addition, it is noteworthy that the
lithium enolates (THF, -78 °C) exhibit a neglible tendency to
respond to this control element. The steric hindrance of the
nucleophile appears to affect diastereofacial selectivity only in
the Lewis acid-promoted enolsilane addition reactions (M )
TMS/BF3‚OEt2). Comparison of the benzyl-substituted R-ben-
zyloxy aldehyde 1 and the iso-propyl-substituted R-benzyloxy
aldehyde 2 reveals that â-branching results in improved dia-
stereofacial selectivity in the lithium enolate addition reactions.
The corresponding iso-propyl-substituted R-silyloxy aldehyde
3 exhibits comparable results to 2, with a slight improvement
observed for enolborane additions. Lithium enolate additions
afford an effective strategy for the construction of anti-â,γ-
alkoxy carbonyl compounds in high diastereomeric purity,
provided the R-alkyl substituent is branched.16
The impact of multiple stereocenters on aldehyde diastereo-
facial selectivity might be analyzed in terms of the stereocontrol
elements provided by the individual substituents. Such an
approach has proven successful in describing trends in π-facial
selectivity for nucleophilic addition reactions of R-methyl-â-
alkoxy aldehydes (eq 3). Consideration of the individual
observed face selectivities for additions to R- and â-alkoxy
aldehydes leads to the conclusion that the two stereocenters in
the syn-aldehyde diastereomer should be mutually reinforcing,
since the R- and â-configurations promote nucleophilic addition
to the same aldehyde π-face (Figure 3). Alternatively, the anti-
aldehyde diastereomer appears to be nonreinforcing where the
R- and â-configurations promote nucleophilic addition to
opposite aldehyde π-faces. On the basis of this simple analysis,
one would predict high levels of stereoselectivity for nucleo-
philic additions to the syn-aldehyde, and low levels for the anti-
aldehyde.
Table 1. Aldol Reactions of R-Alkoxy Aldehydesa
Figure 3. Predicted impact of R- and â-alkoxy stereocenters.
To investigate these qualitative projections, aldehydes 4-11
were selected for study in aldol addition reactions (Chart 2).
The structural features of these aldehydes were chosen to
correspond to the 1,2- and 1,3-asymmetric induction studies
(vide supra). These aldehydes share a branched iso-propyl
substituent at the â-position. Benzyl and silyl groups were
selected to protect the R- and â-oxygen atoms due to their
common use in synthesis design and their significant steric and
electronic differences. The benzyl (Bn) and tert-butyldimeth-
ylsilyl (TBS) protecting groups were selected for the R-oxygen
substituent in accord with the study of 1,2-asymmetric induction,
3,4-anti:3,4-synb (yield)c
aldehyde
R
′
M
)
TMS/BF3
‚
OEt2
M
)
9-BBN
M ) Li
1
Me
i-Pr
t-Bu
77:23 (67)
47:53 (59)
42:58 (77)
68:32 (88)
69:31 (93)
68:32 (98)
71:29 (83)
64:36 (75)
62:38 (73)
2
3
Me
i-Pr
t-Bu
80:20 (67)
53:47 (60)
64:36 (65)
75:25 (79)
79:21 (89)
80:20 (81)
94:06 (88)
91:09 (84)
89:11 (76)
Me
i-Pr
t-Bu
82:18 (66)
75:25 (69)
50:50 (66)
85:15 (83)
85:15 (85)
82:18 (76)
85:15 (84)
88:12 (53)
91:09 (78)
a All reactions were conducted at -78 °C in CH2Cl2 except when M )
Li (-78 °C in THF). All isolable products were unambiguously character-
ized. b Ratios were determined by HPLC analysis of the unpurified reaction
mixture, or by GLC analysis of the derivatized (silylated or acetylated)
unpurified reaction mixture. c Yields are reported for the mixture of isolated
diastereomeric adducts.
(14) The same control elements appear to be operating in the reduction of
â-alkoxy ketones: Evans, D. A.; Dart, M. J.; Duffy, J. L. Tetrahedron
Lett. 1994, 35, 8541-8544.
(15) For a recent double stereodifferentiating Mukaiyama aldol addition see:
Keck, G. E.; Knutson, C. E.; Wiles, S. A. Org. Lett. 2001, 3, 707-710. In
this paper, Keck presents a clear three-dimensional representation for
merged stereoinduction that is adopted herein.
9
J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006 9435

A R T I C L E S
Evans et al.
Table 3. Aldol Reactions of anti-R,â-Bisalkoxy Aldehydesa
Table 2. Aldol Reactions of syn-R,â-Bisalkoxy Aldehydesa
3,4-anti:3,4-synb (yield)c
3,4-anti:3,4-synb (yield)c
OEt2 9-BBN
aldehyde
R
M
)
TMS/BF3
‚
OEt2
M
)
9-BBN
M
)
Li
aldehyde
R
M
)
TMS/BF3
‚
M
)
M ) Li
4
Me
i-Pr
t-Bu
31:69 (50)
18:82 (83)
25:75 (62)
40:60 (95)
44:56 (91)
54:46 (89)
55:45 (98)
69:31 (93)
71:29 (77)
8
Me
i-Pr
t-Bu
99:01 (86)
97:03 (84)
97:03 (77)
93:07 (90)
92:08 (94)
96:04 (94)
99:01 (83)
99:01 (95)
98:02 (87)
5
6
7
Me
i-Pr
t-Bu
87:13 (77)
49:51 (86)
45:55 (61)
75:25 (93)
78:22 (98)
83:17 (99)
74:26 (86)
79:21 (87)
72:28 (76)
9
10
11
Me
i-Pr
t-Bu
90:10 (81)
78:22 (87)
75:25 (79)
86:14 (99)
80:20 (96)
80:20 (99)
>99:01 (99)
99:01 (94)
>99:01 (67)
Me
i-Pr
t-Bu
67:33 (78)
49:51 (67)
18:82 (37)
45:55 (86)
36:64 (83)
33:67 (86)
63:37 (90)
84:16 (87)
66:34 (86)
Me
i-Pr
t-Bu
65:35 (83)
41:59 (95)
09:91 (89)
91:09 (88)
86:14 (92)
81:19 (90)
>99:01 (95)
>99:01 (95)
99:01 (98)
Me
i-Pr
t-Bu
34:66 (69)
43:57 (61)
14:86 (61)
07:93 (90)
07:93 (87)
03:97 (91)
68:32 (79)
68:32 (83)
66:34 (70)
Me
i-Pr
t-Bu
95:05 (82)
87:13 (82)
47:53 (69)
98:02 (83)
99:01 (78)
97:03 (83)
>99:01 (89)
>99:01 (87)
99:01 (90)
a All reactions were conducted at -78 °C in CH2Cl2 except when M )
Li (-78 °C in THF). All isolable products were unambiguously character-
ized. b Ratios were determined by HPLC analysis of the unpurified reaction
mixture, or by GLC analysis of the derivatized (silylated or acetylated)
unpurified reaction mixture. c Yields are reported for the mixture of isolated
diastereomeric adducts.
a All reactions were conducted at -78 °C in CH2Cl2 except when M )
Li (-78 °C in THF). All isolable products were unambiguously character-
ized. b Ratios were determined by HPLC analysis of the unpurified reaction
mixture, or by GLC analysis of the derivatized (silylated or acetylated)
unpurified reaction mixture. c Yields are reported for the mixture of isolated
diastereomeric adducts.
anti-r,â-Bisalkoxy Aldehydes. Reactions of the anti-
configured aldehydes 8-11 with methyl ketone-derived enolate
nucleophiles are presented in Table 3. The immediate conclusion
is that this aldehyde family exhibits improved reaction dia-
stereoselectivities when compared to their syn-aldehyde coun-
terparts (Table 2). Under Mukaiyama aldol conditions (M )
TMS/BF3‚OEt2), the diastereoselectivity was significantly af-
fected by both the steric hindrance of the enolate component
and the identity of the oxygen protecting groups. Aldehyde 8
exhibits uniformly high selectivity with all enolsilanes, whereas
the diastereoselectivity in the reactions of the other aldehydes
is sensitive to enolsilane steric hindrance. Enolborane additions
show a moderate protecting group dependence, with aldehydes
bearing a â-OTBS substituent (M ) 9-BBN, 8 and 11) giving
superior selectivity relative to â-OPMB aldehydes (M ) 9-BBN,
9 and 10).17 All aldehydes reacting with lithium enolates (M )
Li) afford outstanding selectivity for the 3,4-anti diastereomer
(>98:02). In addressing the issues of synthesis design, the results
indicate the relative ease of establishing the 3,4-anti/4,5-anti
stereotriad by an aldol-based strategy.
Unbranched â-Alkyl Substituent. The effect of the relative
size of the â-alkyl substituent on diastereofacial selectivity was
examined in aldehydes 12 and 13 (Table 4). The syn-aldehyde
12 is observed to provide aldol adducts favoring the 3,4-syn
diastereomer for both enolsilane and enolborane nucleophiles,
while modest selectivity for the 3,4-anti diastereomer is observed
with the lithium enolate. In contrast, the anti-configured
aldehyde 13 is observed to provide aldol adducts with good to
excellent selectivity for the 3,4-anti diastereomer. While the
differences in selectivity between syn- and anti-aldehydes is still
syn-r,â-Bisalkoxy Aldehydes. The π-facial selectivities
exhibited by syn-R,â-bisalkoxy aldehydes (4-7) in aldol reac-
tions with methyl ketone-derived enolate nucleophiles are
presented in Table 2. Although it had been anticipated that the
syn diastereomer might contain stereoreinforcing control ele-
ments (Figure 3), the levels of selectivity for the 3,4-anti product
are surprisingly low. The Lewis acid-promoted enolsilane
additions (M ) TMS/BF3‚OEt2) are the only conditions under
which selectivity was significantly affected by the steric
encumbrance of the enolate component, with selectivity gener-
ally decreasing with the increasing steric size of the enolate
substituent (R). This trend is also observed in reactions of
R-alkoxy aldehydes (Table 1). On average, the enolsilane
additions slightly favor the 3,4-syn product diastereomer. The
enolborane aldol reactions (M ) 9-BBN) are characterized by
a significant protecting group dependence, in which the presence
of two TBS protecting groups results in the nearly exclusive
formation of the 3,4-syn product. Lithium enolate additions are
relatively consistent across all aldehyde and enolate structures,
and while the 3,4-anti product is favored, the diastereoselectivity
is lower than for aldehydes containing a single R-alkoxy-
substituted stereocenter (Table 1). Regarding synthetically useful
transformations, the present study reveals the unexpected
difficulty in obtaining the 3,4-anti/4,5-syn diastereomer from
an aldol-based approach. However, useful levels of selectivity
(g93:07 dr) are available for the construction of the 3,4-syn/
4,5-syn diastereomer via enolborane addition to the bis-TBS
protected aldehyde 7 (M ) 9-BBN).
(16) Heathcock and co-workers have previously reported highly diaster-
eoselective aldol addition reactions between the lithium enolate of pina-
colone and R-alkoxy aldehydes. See ref 10b.
(17) For a report of a â-oxygen protecting group effect in the addition of boron
enolates to R-methyl-â-alkoxy aldehydes, see: Gustin, D. J.; VanNieu-
wenhze, M. S.; Roush, W. R. Tetrahedron Lett. 1995, 36, 3443-3446.
9
9436 J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006

Model for Acyclic Stereocontrol
A R T I C L E S
Table 6. Aldol Reactions of anti-R,â-Bisalkoxy Aldehyde
quite large, aldehyde 13 does exhibit slightly lower selectivity
relative to the corresponding â-iso-propyl-substituted aldehyde
8 (Table 3).
Acetonidesa
Table 4. Aldol Reactions of â-Ethyl-R,â-bisalkoxy Aldehydesa
46:47b (yield)c
9-BBN
R
M
)
TMS/BF3
‚
OEt2
M
)
M ) Li
Me
i-Pr
t-Bu
83:17 (81)
63:37 (61)
25:75 (69)
67:33 (98)
76:24 (85)
78:22 (74)
96:04 (81)
96:04 (85)
94:06 (90)
3,4-anti:3,4-synb (yield)c
a All reactions were conducted at -78 °C in CH2Cl2 except when M )
Li (-78 °C in THF). All isolable products were unambiguously character-
ized. b Ratios were determined by HPLC analysis of the unpurified reaction
mixture, or by GLC analysis of the derivatized (silylated or acetylated)
unpurified reaction mixture. c Yields are reported for the mixture of isolated
diastereomeric adducts.
aldehyde
R
,
â
-
M
)
TMS/BF3
‚
OEt2
M
)
9-BBN
M ) Li
12
13
syn
anti
07:93 (83)
84:16 (91)
22:78 (96)
83:17 (88)
71:29 (88)
95:05 (76)
a All reactions were conducted at -78 °C in CH2Cl2 except when M )
Li (-78 °C in THF). All isolable products were unambiguously character-
ized. b Ratios were determined by integration of the 1H NMR spectra of
the unpurified reaction mixtures. c Yields are reported for the mixture of
isolated diastereomeric adducts.
Model for Asymmetric Induction. Our observations (Tables
2 and 3) reveal that the anti-aldehyde diastereomers exhibit
significantly higher reaction diastereoselectivities than their syn-
aldehyde counterparts. This trend runs counter to preliminary
expectations based on a summation of the individual contribu-
tions of the R- and â-stereocenters to π-facial selectivity (Figure
3). It is apparent that the presence of multiple oxygen stereo-
centers affects aldehyde π-facial selectivity in a manner that is
not encountered in aldehydes with single oxygen-substituted
stereocenters.
The aldehydes under study have considerable conformational
flexibility, and the generation of transition state models that
correlate with our data can be simplified by the application of
well-established paradigms for asymmetric induction to select
appropriate OdC-CR torsion angles. Both the Cornforth and
PFA torsion angle constraints are applied in this regard. To
determine the likelihood of a transition structure contributing
to product formation, the following assumptions are made: (A)
fully developed syn-pentane interactions between the R-OP
substituent and the non-hydrogen â-substituents are prohibi-
tive;19 (B) developing syn-pentane interactions between the
nucleophile and substrate are prohibitive;20 (C) fully developed
syn-pentane interactions are energetically more costly than
developing syn-pentane interations within the reacting electro-
phile; (D) developing syn-pentane interactions between CdO
and â-C-R are energetically more costly than between CdO
and â-C-O.13 Prohibitive syn-pentane interactions are high-
lighted in red, while developing syn-pentane interactions are
shown in blue.
Acetonide Protecting Group. The syn- and anti-R,â-bis-
alkoxy aldehyde acetonides corresponding to erythrose and
threose18 were investigated to determine the effect of a cyclic
protecting group and relative configuration on diastereofacial
selectivity (Tables 5 and 6). Acetonide 14, corresponding to a
syn-R,â-bisalkoxy configuration, reacts with low to moderate
diastereoselectivity in Lewis acid-promoted and enolborane aldol
reactions. Again, the selectivity of the lithium enolate aldol
reaction is significantly better, and is the first example of high
levels of 3,4-anti diastereoselectivity from a syn-R,â-bisalkoxy
aldehyde in this study. Acetonide 15 corresponding to an anti-
R,â-bisalkoxy configuration exhibits similar trends, with the
average diastereoselectivity being slightly higher than that
observed for the syn-configured aldehyde 14. The similar
behavior of the syn- and anti-acetonide aldehydes is noteworthy,
since the majority of syn- and anti-aldehydes studied exhibit
significant differences in π-facial selectivity.
Table 5. Aldol Reactions of syn-R,â-Bisalkoxy Aldehyde
Acetonidesa
44:45b (yield)c
As the aldehydes under consideration contain vicinal alkoxy
substituents, it is possible that aldehyde conformations contain-
ing a gauche arrangement of alkoxy substituents may be
stabilized relative to aldehydes containing an anti arrangement
of these groups.21 By means of NMR spin-spin coupling
measurements of 1,2-dimethoxyethane in the liquid phase, it
has been established that the gauche OCCO tgt conformer is
0.5 kcal/mol more stable than the trans OCCO ttt conformer.22
R
M
)
TMS/BF3
‚
OEt2
M
)
9-BBN
M ) Li
Me
i-Pr
t-Bu
65:35 (61)
67:33 (69)
54:46 (28)
49:51 (74)
54:46 (91)
54:46 (72)
89:11 (62)
92:08 (48)
90:10 (58)
a All reactions were conducted at -78 °C in CH2Cl2 except when M )
Li (-78 °C in THF). All isolable products were unambiguously character-
ized. b Ratios were determined by HPLC analysis of the unpurified reaction
mixture, or by GLC analysis of the derivatized (silylated or acetylated)
unpurified reaction mixture. c Yields are reported for the mixture of isolated
diastereomeric adducts.
(19) The syn-pentane arrangement for methyl propyl ether (COCC and OCCC
dihedral angles of +60 and -60°) has been calculated to be +6.2 kcal/
mol: Wiberg, K. B.; Murcko, M. A. J. Am. Chem. Soc. 1989, 111, 4821-
4828.
(20) For the reaction of acetaldehyde enolborane with 2-methoxypropanal (ref
9), it has not been possible to generate calculated structures which contain
a syn-pentane relationship between the forming C-C bond and the O-CH3
bond, suggesting that this arrangement is prohibitively high in energy.
(21) Wolfe, S. Acc. Chem. Res. 1972, 5, 2-111.
(18) Acetonides corresponding to aldehydes 4-11 were initially studied.
Unfortunately, it was found that when subjected to a BF3‚OEt2-promoted
enolsilane addition reaction, the anti-R,â-alkoxy aldehyde acetonide
underwent acetonide migration. This rearrangement greatly complicated
the analysis of the reaction diastereoselectivity. Fortunately, the aldehyde
acetonides related to erythrose and threose were not observed to undergo
migration.
9
J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006 9437

A R T I C L E S
Evans et al.
Figure 4. Cornforth transition structures for nucleophilic addition to syn- and anti-R,â-alkoxy aldehydes.
Unlike 1,2-dimethoxyethane, however, the aldehydes under
consideration contain a C3-C6 butane fragment, and we believe
that the conformational preference of this fragment is more
significant, as trans-butane is known to be 0.9 kcal/mol more
stable than gauche-butane, a value considerably larger than the
gauche preference for 1,2-dimethoxyethane. The finding that
the trends in diastereoselectivity observed in R,â-bisalkoxy
aldehydes persist among R-alkoxy, â-methyl aldehydes 16 and
17 (Table 7), in which vicinal alkoxy substituents are not
present, supports this view.
arrangement of â-C-O and CdO. This would likely make
addition to the syn-aldehyde via CS1 a very favorable situation,
were it not for the prohibitive syn-pentane interaction between
O-P1 and the forming C-C bond. This interaction was not
accounted for in our simple prediction (Figure 3), and is likely
responsible for the failure of this prediction. Rotation about the
C4-C5 bond results in additional transition structures CS2 and
CS3 in which P1 does not experience prohibitive interactions
with the nucleophile, but this comes at the expense of additional
gauche interactions involving the â-alkyl substituent. According
to criterion D, transition structure CS2 is proposed to be the
most likely for addition to the syn-aldehyde. Corresponding
transition structures for addition to the anti-aldehyde (CA2 and
CA3) can also be considered, but suffer from significant
destabilizing interactions relative to CA1. Comparison of the
most likely transition structures for addition to the syn- and anti-
aldehydes, CS2 and CA1, respectively, reveals an identical
developing syn-pentane interaction between â-C-O and Cd
O. The suboptimal position of the â-alkyl substituent in the
transition state for addition to the syn-aldehyde, compared to
an optimal positioning of this group for addition to the anti-
aldehyde, suggests that addition to the anti-aldehyde should be
more faVorable. This finding is consistent with our observations
(Tables 1 and 2) and we conclude that evaluation of the
Cornforth transition structures of syn- and anti-R,â-bisalkoxy
aldehydes leading to the 3,4-anti product diastereomer provides
a sufficient explanation for the observed differences in dia-
stereofacial selectivity.
Table 7. Aldol Reactions of R-Alkoxy-â-Methyl Aldehydesa
3,4-anti:3,4-synb (yield)c
aldehyde
R
,
â-
M
)
TMS/BF3
‚OEt2
M
)
9-BBN
M ) Li
16
17
syn
anti
47:53 (63)
70:30 (87)
78:22 (87)
87:13 (93)
81:19 (83)
98:02 (90)
a All reactions were conducted at -78 °C in CH2Cl2 except when M )
Li (-78 °C in THF). All isolable products were unambiguously character-
ized. b Ratios were determined by GLC analysis of the silylated unpurified
reaction mixture. c Yields are reported for the mixture of isolated diaster-
eomeric adducts.
Cornforth Model. Conformational representations of transi-
tion structures based on the Cornforth model are depicted in
Figure 4. In all representations, the reacting π-face of CdO
corresponds to the formation of 3,4-anti products, and the
R-C-O and CdO bonds are in an antiparallel arrangement. In
this orientation, three transition structures are possible due to
rotation about the C4-C5 bond for both the syn-configured
aldehyde (CS1-CS3) and the anti-configured aldehyde (CA1-
CA3). A unique consequence of the R-oxygen substituent is that
the protecting group P1 is coupled to rotation about the C4-C5
bond due to potential syn-pentane interactions with the non-
hydrogen â-substituents. Beginning with the â-alkyl group in
an optimal position anti to the reacting carbonyl, transition state
CS1 contains the same favorable nonparallel arrangement of
â-C-O and CdO that is implicated in the 1,3-polar model of
asymmetric induction (Figure 2) while the corresponding
transition state for the anti-aldehyde CA1 contains a parallel
Polar Felkin-Anh Model. Conformational representations of
transition structures based on the Polar Felkin-Anh model are
illustrated in Figure 5. In all representations, the reacting π-face
of CdO corresponds to the formation of 3,4-anti products, and
the R-C-O and CdO bonds are in a perpendicular arrangement.
In this orientation, three transition structures are possible due
to rotation about the C4-C5 bond for both the syn-configured
aldehyde (FS1-FS3) and the anti-configured aldehyde (FA1-
FA3). Transition structures FS2, FS3, FA1, and FA3 are observed
to contain prohibitive syn-pentane interactions between the
forming C-C bond and non-hydrogen â-substituents, and it is
unlikely that these contribute to product formation. The remain-
ing transition structures, FS1 for addition to the syn-aldehyde,
and FA2 for addition to the anti-aldehyde, exhibit developing
syn-pentane interactions between CdO and either OP2 or R,
respectively. To the extent that CdOTR is more destabilizing
than CdOTOP2 (criterion D), the polar Felkin-Anh transition
structures suggest that addition to the anti-aldehyde is less
favorable than the analogous addition to the syn-aldehyde. This
(22) (a) Viti, V.; Indovina, P. L.; Podo, F.; Radics, L.; Nemethy, G. Mol. Phys.
1974, 27, 541-559. (b) Tasaki, K.; Abe, A. Polym. J. 1985, 17, 641-655.
(c) Abe, A.; Tasaki, K.; Mark, J. E. Polym. J. 1985, 17, 883-893.
9
9438 J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006

Model for Acyclic Stereocontrol
A R T I C L E S
Figure 5. Polar Felkin-Anh transition structures for nucleophilic addition to syn- and anti-R,â-alkoxy aldehydes.
finding is inconsistent with our observations (Tables 1 and 2)
and we conclude that evaluation of the PFA transition structures
of syn- and anti-R,â-bisalkoxy aldehydes leading to the 3,4-
anti product diastereomer is insufficient in accounting for the
observed differences in diastereofacial selectivity.
subjected to aldol addition reactions (Table 7). The syn-
configured aldehyde exhibits lower diastereoselectivity for the
3,4-anti product diastereomer in every case. This finding lends
additional support to the steric interactions identified in the
Cornforth transition states for addition to R,â-bisalkoxy alde-
hydes.
Cornforth Model: Acetonide-Protected Aldehydes. In
contrast to the dramatic differences in diastereofacial selectivity
observed for syn- and anti-aldehydes with independent protect-
ing groups (Tables 2 and 3), the syn- and anti-acetonides (Tables
5 and 6) exhibited comparable levels of selectivity in aldol
addition reactions. Examination of the Cornforth transition
structures for addition to syn- and anti-aldehyde acetonides
(Figure 6) indicates that the acetonide linkage serves to remove
the offending NuTP1 interaction in the transition state for
nucleophilic addition to the syn-aldehyde (compare CS1, Figure
4, with Cs, Figure 6). The steric environment in the vicinity
of the nucleophile is now similar for both syn-(Cs) and
anti-aldehydes (CA), consistent with the comparable aldehyde
π-facial selectivity.
The preceding analysis has established that Cornforth transi-
tion structures leading to the formation of 3,4-anti product
diastereomers provide a plausible explanation for the relative
differences in asymmetric induction observed for syn- and anti-
R,â-bisalkoxy aldehydes. For the syn-aldehydes, the 3,4-syn-
diastereomer often constitutes a significant part (and in some
cases, major part) of the observed product distribution. This
implies the existence of transition structures leading to the 3,4-
syn-product that compete effectively with the favored transition
structure for the formation of the 3,4-anti product (CS2, Figure
4). Figure 8 illustrates several transition structures for the
formation of the 3,4/4,5-syn-product in which prohibitive syn-
pentane interactions are absent. NCS and NFS are formally
related to the Cornforth and polar Felkin-Anh models, respec-
tively, due to the relationship of CdO and R-C-O. NS1 and
NS2, on the other hand, contain a synparallel arrangement of
CdO and R-C-O. The developing syn-pentane interactions are
so similar, and the orientations of the R-stereocenter so different,
that it is difficult to determine the most likely transition structure
without performing a more sophisticated computational analysis.
Observations in Related Systems. Other groups have
reported the addition of enolate nucleophiles23 to R,â-bisalkoxy
aldehydes.24 While syn- and anti-configured aldehydes of the
same structure are not directly compared, the following examples
highlight the importance of the principles established by this
systematic study.
Figure 6. Cornforth transition structures for nucleophilic addition to R,â-
alkoxy aldehyde acetonides.
syn-r,â-Bisalkoxy Aldehydes. Kobayashi and co-workers
have reported that addition of a polymer-supported silylketene
thioacetal to a syn-R,â-silyloxy aldehyde provides exclusively
the 3,4-syn-product diastereomer (eq 7),25 a result that is
inconsistent with the Cornforth/polar Felkin-Anh models for
asymmetric induction. A similar case has been documented by
Paterson and co-workers (eq 8).26 These unexpected syn
selectivities can now be understood as a consequence of the
aldehyde configuration, in which the â-alkoxy substituent forces
the large R-silyloxy group to project into the path of the
nucleophile at what would otherwise be the preferred π-face of
the aldehyde (CS1, Figure 4). In contrast, syn-R,â-alkoxy
Additions to r-Alkoxy-â-Methyl Aldehydes. The influence
of the configuration at the â-stereocenter on aldehyde π-facial
selectivity is proposed to be due in part to the conformational
constraint imposed by the â-alkoxy substituent on the R-oxygen
protecting group, which effectively removes any contribution
from transition state CS1 (Figure 4). Since this is primarily a
steric effect, it follows that a methyl substituent in place of the
â-alkoxy substituent should result in the same trend in π-facial
selectivity (Figure 7), with lower 3,4-anti selectivity for the syn-
aldehyde. To investigate these predictions, syn- and anti-R-
silyloxy-â-methyl aldehydes 16 and 17 were constructed and
9
J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006 9439

A R T I C L E S
Evans et al.
Figure 7. Cornforth transition structures for nucleophilic addition to syn- and anti-R-alkoxy-â-methyl aldehydes.
in the proposed transition state (CA1, Figure 4) for nucleophilic
addition to anti-R,â-bisalkoxy aldehydes.
Figure 8. Transition structures for the formation of the 3,4-syn/4,5-syn-
product diastereomer.
aldehyde acetonides are observed to provide aldol adducts with
high selectivity for the anti product diastereomer,27 as the
example from Kobayashi and co-workers demonstrates (eq 9).25
As noted previously (SC, Figure 6), the acetonide protecting
group minimizes destabilizing interactions with the nucleophile,
resulting in selectivity for the 3,4-anti product diastereomer.
Conclusions
A systematic study of asymmetric induction in aldol addition
reactions of R-alkoxy and R,â-bisalkoxy aldehydes has been
presented. We find that asymmetric induction in lithium enolate
additions to R-alkoxy aldehydes is superior to that obtained from
the corresponding enolborane and enolsilane nucleophiles. The
levels of asymmetric induction in the lithium enolate additions
are relatively insensitive to both the identity of the R-oxygen
protecting group and the steric hindrance of the enolate
nucleophile. These additions are, however, sensitive to the nature
(23) Only enolate nucleophiles reacting under conditions where chelation is
unlikely are included in this discussion. While a large number of examples
exist for the addition of organometallic reagents to R,â-alkoxy aldehydes,
the analysis of the observed diastereoselectivity is greatly complicated by
the ambiguous nature of the nucleophile and the unknown contribution
from chelated transition states.
(24) Aldol addition reactions of R,â-alkoxy aldehydes in which the R-oxygen
is part of a tetrahydrofuran or tetrahydropyran ring system have been
reported. Due to the lack of comparable cases from our study, these
examples will not be discussed. For R-THP aldehydes see: (a) Dondoni,
A.; Ianelli, S.; Kniezo, L.; Merino, P.; Nardelli, M. J. Chem. Soc., Perkin
Trans. 1 1994, 1231-1239. (b) Sasaki, M.; Nonomura, T.; Murata, M.;
Tachibana, K. Tetrahedron Lett. 1995, 36, 9007-9010. For an R-THF
aldehyde, see: (c) Anderson, O. P.; Barrett, A. G. M.; Edmunds, J. J.;
Hachiya, S.-I.; Hendrix, J. A.; Horita, K.; Malecha, J. W.; Parkinson, C.
J.; VanSickle, A. Can. J. Chem. 2001, 79, 1562-1592.
anti-r,â-Alkoxy Aldehydes. The azaspiracid class of natural
products28 has inspired a number of aldol-based approaches for
the construction of the C34-C35 bond. Three research groups
have independently found that the addition of a C35-C40 methyl
ketone enolate to the indicated structurally diverse anti-R,â-
bisalkoxy aldehydes (eqs 10-12) results in extremely high
selectivity for the unnatural 33,34-anti diastereomer.29 These
examples are consistent with the favorable features identified
(25) Kobayashi, S.; Wakabayashi, T.; Yasuda, M. J. Org. Chem. 1998, 63,
4868-4869.
(26) Enolborane and lithium enolate additions are also reported: Paterson, I.;
Di Francesco, M. E.; Kuhn, T. Org. Lett. 2003, 5, 599-602.
9
9440 J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006

Model for Acyclic Stereocontrol
A R T I C L E S
of the â-alkyl substituent. For R,â-bisalkoxy aldehydes, the
relationship between relative configuration and π-facial selectiv-
ity has been established. Aldehydes with an anti-R,â-bisalkoxy
configuration reacted with methyl ketone-derived enolates to
give 3,4-anti products with consistently superior diaster-
eoselectivity relative to the syn-R,â-bisalkoxy aldehydes. This
trend is observed in aldehydes containing a wide range of
protecting groups, branched or unbranched â-alkyl substituent,
and is even evident in nucleophilic additions to R-alkoxy-â-
methyl aldehydes. The only exception is the case of the
acetonide-protected R,â-bisalkoxy aldehydes, which exhibit
similar levels of diastereoselection regardless of relative con-
figuration. A Cornforth transition-state model is proposed to
account for these observations in which the â-alkoxy substituent
dictates the position in space occupied by the R-oxygen
protecting group, which in turn governs aldehyde π-facial
selectivity due to its proximity to the approaching nucleophile.
The systematic study presented here provides a comprehensive
data set for the synthesis of vicinal alkoxy stereotriads by an
aldol-based approach, and should lead to greater sophistication
in the synthesis of poly-hydroxylated structures.
(27) Lithium enolate nucleophile: (a) Dondoni, A.; Merino, P. Synthesis 1993,
903-908. (b) Dondoni, A.; Merino, P. J. Org. Chem. 1991, 56, 5294-
5301. Silyloxyfuran nucleophile: (c) Rassu, G.; Spanu, P.; Casiraghi, G.;
Pinna, L. Tetrahedron 1991, 47, 8025-8030.
(28) (a) Satake, M.; Ofuji, K.; Naoki, H.; James, K. J.; Furey, A.; McMahon,
T.; Silke, J.; Yasumoto, T. J. Am. Chem. Soc. 1998, 120, 9967-9968. (b)
Nicolaou, K. C.; Vyskocil, S.; Koftis, T. V.; Yamada, Y. M. A.; Ling, T.;
Chen, D. Y.-K.; Tang, Wenjun; Petrovic, G.; Frederick, M. O.; Li, Y.;
Satake, M. Angew. Chem., Int. Ed. 2004, 43, 4312-4318. (c) Nicolaou,
K. C.; Koftis, T. V.; Vyskocil, S.; Petrovic, G.; Ling, T.; Yamada, Y. M.
A.; Tang, W.; Frederick, M. O. Angew. Chem., Int. Ed. 2004, 43, 4318-
4324.
(29) Equation 10: (a) Travis Dunn, Harvard University, unpublished result.
Equation 11: (b) Forsyth, C. J.; Hao, J.; Aiguade, J. Angew. Chem., Int.
Ed. 2001, 40, 3663-3667. Equation 12: (c) Nicolaou, K. C.; Pihko, P.
M.; Diedrichs, N.; Zou, N.; Bernal, F. Angew. Chem., Int. Ed. 2001, 40,
1262-1265.
Acknowledgment. Support has been provided by the National
Institutes of Health (GM-33328-18). Fellowship support from
the NSF for V.J.C. and Lilly for S.J.S. is also acknowledged.
Supporting Information Available: Experimental details,
analytical data, and stereochemical proofs for all new com-
pounds (PDF, CIF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA061010O
9
J. AM. CHEM. SOC. VOL. 128, NO. 29, 2006 9441