Syn-anti isomerization of aldols by enolization

Article abstract of DOI:10.1021/jo049626o

A variety of aldol adducts are shown to undergo efficient syn-anti isomerization in the presence of imidazole by an enolization mechanism. Isomerizations are high yielding and occur with little or none of the usual byproducts arising from competing elimination or retroaldol reactions. Most substrates reach equilibrium within 0.3-3 days at ambient temperature in chloroform, benzene, or dichloromethane containing 0.3-1 M imidazole. The process is particularly facile for aldols derived from tetrahydro-4H-thiopyran- 4-one with rate constants for equilibration varying over ca. 1 order of magnitude for the adducts studied; structurally related aldols derived from cyclohexanone isomerized ca. 3-4 times slower. Isomerization of the acyclic aldol 5-hydroxy-4-methyl-5-phenyl-3-pentanone required heating to 60 °C but was achieved with minimal (<5%) retroaldol or elimination. A methoxymethyl ether derivative isomerized 30-40 times slower than the parent aldol. Isomerization of α,α′-disubstituted aldols and α,α′-bisaldols indicated low regioselectivity in the enolization. The synthetic utility of the process was demonstrated with the effective preparation of aldol stereoisomers unobtainable by direct methods.

Full text of DOI:10.1021/jo049626o

Syn -An ti Isom er iza tion of Ald ols by En oliza tion
Dale E. Ward,* Marcelo Sales, and Pradip K. Sasmal
Department of Chemistry, University of Saskatchewan, 110 Science Place,
Saskatoon SK S7N 5C9, Canada
dale.ward@usask.ca
Received March 6, 2004
A variety of aldol adducts are shown to undergo efficient syn-anti isomerization in the presence
of imidazole by an enolization mechanism. Isomerizations are high yielding and occur with little
or none of the usual byproducts arising from competing elimination or retroaldol reactions. Most
substrates reach equilibrium within 0.3-3 days at ambient temperature in chloroform, benzene,
or dichloromethane containing 0.3-1 M imidazole. The process is particularly facile for aldols derived
from tetrahydro-4H-thiopyran-4-one with rate constants for equilibration varying over ca. 1 order
of magnitude for the adducts studied; structurally related aldols derived from cyclohexanone
isomerized ca. 3-4 times slower. Isomerization of the acyclic aldol 5-hydroxy-4-methyl-5-phenyl-
3-pentanone required heating to 60 °C but was achieved with minimal (<5%) retroaldol or
elimination. A methoxymethyl ether derivative isomerized 30-40 times slower than the parent
aldol. Isomerization of R,R′-disubstituted aldols and R,R′-bisaldols indicated low regioselectivity in
the enolization. The synthetic utility of the process was demonstrated with the effective preparation
of aldol stereoisomers unobtainable by direct methods.
In tr od u ction
Isomerization of aldol adducts is an alternative strat-
egy to access stereoisomers that cannot be obtained
directly. A â-hydroxy ketone (or aldehyde) can isomerize
by keto-enol tautomerism or retroaldol-aldol mecha-
nisms.8 The aldol reaction is readily reversible (i.e.,
retroaldol), and when mediated by weak bases (e.g.,
hydroxide, alkoxide, amine), product distribution is gen-
erally under thermodynamic control.9,10 The “directed”
aldol reaction11 using preformed enol(ate) derivatives
typically gives products under kinetic control;2 however,
The aldol addition of a ketone enol or enolate derivative
to an aldehyde can produce up to two new stereogenic
centers, and several stereoisomeric products are generally
possible.1 The development of methods for stereoselective
aldol reactions has been intensively investigated for more
than 20 years, and this reaction is now among the most
powerful in the synthetic chemist’s arsenal for stereo-
controlled carbon-carbon bond formation.2 Aldol reac-
tions have been used extensively in total syntheses of
numerous stereochemically complex natural products.3
Nonetheless, retrosynthetic disconnections involving ste-
reoselective aldol transforms must be made carefully
because the ability to produce any specific aldol stereo-
isomer selectively,2c,4 especially when both substrates are
chiral,5,6 remains a significant challenge.7
(4) Paterson, I.; Channon, J . A. Tetrahedron Lett. 1992, 33, 797-
800.
(5) (a) Masamune, S.; Choy, W.; Petersen, J . S.; Sita, L. R. Angew.
Chem., Int. Ed. 1985, 24, 1-30. (b) Kolodiazhnyi, O. I. Tetrahedron
2003, 59, 5953-6018.
(6) (a) Evans, D. A.; Yang, M. G.; Dart, M. J .; Duffy, J . L.; Kim, A.
S. J . Am. Chem. Soc. 1995, 117, 9598-9599. (b) Evans, D. A.; Dart,
M. J .; Duffy, J . L.; Rieger, D. L. J . Am. Chem. Soc. 1995, 117, 9073-
9074. (c) Evans, D. A.; Yang, M. G.; Dart, M. J .; Duffy, J . L.
Tetrahedron Lett. 1996, 37, 1957-1960.
(7) (a) Evans, D. A.; Kim, A. S. Tetrahedron Lett. 1997, 38, 53-56.
(b) Wu, Z.; Zhang, F.; Danishefsky, S. J . Angew. Chem., Int. Ed. 2000,
39, 4505-4508. (c) Lee, C. B.; Wu, Z.; Zhang, F.; Chappell, M. D.;
Stachel, S. J .; Chou, T.-C.; Guan, Y.; Danishefsky, S. J . J . Am. Chem.
Soc. 2001, 123, 5249-5259. (d) Scheidt, K. A.; Bannister, T. D.; Tasaka,
A.; Wendt, M. D.; Savall, B. M.; Fegley, G. J .; Roush, W. R. J . Am.
Chem. Soc. 2002, 124, 6981-6990. (e) Evans, D. A.; Cote, B.; Coleman,
P. J .; Connell, B. T. J . Am. Chem. Soc. 2003, 125, 10893-10898. (f)
Marco, J . A.; Carda, M.; Diaz-Oltra, S.; Murga, J .; Falomir, E.; Roeper,
H. J . Org. Chem. 2003, 68, 8577-8582.
(1) Up to 16 stereoisomeric products (eight diastereomers) are
possible if both substrates are racemic.
(2) (a) Evans, D. A.; Nelson, J . V.; Taber, T. R. Top. Stereochem.
1982, 13, 1-115. (b) Heathcock, C. H. In Asymmetric Synthesis;
Morrison, J . D., Ed.; Academic Press: New York, 1984; Vol. 3, pp 111-
212. (c) Heathcock, C. H. Aldrichim. Acta 1990, 23 (4), 99-111. (d)
Heathcock, C. H.; Kim, B. M.; Williams, S. F.; Masamune, S.; Paterson,
I.; Gennari, C. In Comprehensive Organic Synthesis; Trost, B. M.,
Fleming, I., Editors; Pergamon Press: Oxford, U.K., 1991; Vol. 2,
Chapters 1.6, 1.7, 1.9, and 2.4. (e) Nelson, S. G. Tetrahedron: Asym-
metry 1998, 357-389. (f) Mahrwald, R. Chem. Rev. 1999, 99, 1095-
1120. (g) Arya, P.; Qin, H. Tetrahedron 2000, 56, 917-947. (h)
Machajewski, T. D.; Wong, C.-H. Angew. Chem., Int. Ed. 2000, 39,
1352-1374. (i) Carriera, E. M. In Modern Carbonyl Chemistry; Otera,
J ., Ed.; Wiley-VCH: Weinheim, Germany, 2000; Chapter 8, pp 227-
248. (j) Palomo, C.; Oiarbide, M.; Garcia, J . M. Chem. Eur. J . 2002, 8,
36-44. (k) Palomo, C.; Oiarbide, M.; Garcia, J . M. Chem. Soc. Rev.
2004, 33, 65-75.
(8) A mechanism involving elimination and addition of water would
be thermodynamically unfavorable (with respect to the aldol) and is
without literature precedent. See: Stewart, I. C.; Bergman, R. G.;
Toste, F. D. J . Am. Chem. Soc. 2003, 125, 8696-8697.
(9) (a) Nielsen, A. T.; Houlihan, W. J . Org. React. 1968, 16, 1-438.
(b) Heathcock, C. H. In Comprehensive Organic Synthesis; Trost, B.
M., Fleming, I., Eds.; Pergamon Press: Oxford, U.K., 1991; Vol. 2,
Chapter 1.5, pp 133-180.
(3) Paterson, I.; Cowden, C. J .; Wallace, D. J . In Modern Carbonyl
Chemistry; Otera, J ., Ed.; Wiley-VCH: Weinhein, Germany, 2000;
Chapter 9, pp 249-298.
(10) For an exception, see: List, B. Tetrahedron 2002, 58, 5573-
5590.
(11) Mukaiyama, T. Org. React. 1982, 28, 203-331.
10.1021/jo049626o CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/16/2004
4808
J . Org. Chem. 2004, 69, 4808-4815

Syn-Anti Isomerization of Aldols by Enolization
SCHEME 1
isomerization of metal aldolate intermediates via ret-
roaldol can be facile under certain conditions.12 A ret-
roaldol mechanism has been implicated in several reports
of isomerization of cyclic9a,13 and acyclic12e,14 aldols under
basic conditions. By contrast, examples of isomerization
of aldols consistent with an enolization mechanism are
rare,15 and even in these cases, the possibility of a
retroaldol mechanism cannot be ruled out. Similarly, only
scattered reports of related isomerizations of 3-hydroxy-
carboxylic acid derivatives via enol(ate) intermediates
have appeared.16 Several pertinent examples from the
literature are summarized in Scheme 1. Although the
syn-anti isomerization of aldols is clearly a useful
process, there are few examples presumably because the
basic conditions necessary to generate the required enol-
(ate) often induce retroaldol or elimination reactions. We
previously reported that imidazole efficiently catalyzes
syn-anti isomerization of various â-hydroxy ketones by
an enolization mechanism with negligible retroaldol or
elimination.17 In this paper, the results of our study of
the scope and limitations of this process are presented.
Resu lts a n d Discu ssion
We have been exploring the stereoselectivities of
sequential aldol reactions of 8 and 10 as the foundation
of a thiopyran-based synthetic route to polypropio-
nates.18,19 The propensity of imidazole to isomerize tet-
rahydrothiopyranone-derived aldols was first observed
during a failed attempt to protect the hydroxyl group in
(12) (a) Dubois, J . E.; Fellmann, P. C. R. Acad. Sci., Ser. C 1972,
274, 1307-1309. (b) House, H. O.; Crumrine, D. S.; Teranishi, A. Y.;
Olmstead, H. D. J . Am. Chem. Soc. 1973, 95, 3310-3324. (c) Heath-
cock, C. H.; Buse, C. T.; Kleschick, W. A.; Pirrung, M. C.; Sohn, J . E.;
Lampe, J . J . Org. Chem. 1980, 45, 1066-1081. (d) Heathcock, C. H.;
Lampe, J . J . Org. Chem. 1983, 48, 4330-4337. (e) Swiss, K. A.; Choi,
W.-B.; Liotta, D. C.; Abdel-Magid, A. F.; Maryanoff, C. A. J . Org. Chem.
1991, 56, 5978-5980. (f) Carlier, P. R.; Lo, C. W.-S.; Lo, M. M.-C.; Wan,
N. C.; Williams, I. D. Org. Lett. 2000, 2, 2443-2445.
(13) Kok, P.; Sinha, N. D.; Sandra, P.; de Clercq, P. J .; Vandewalle,
M. E. Tetrahedron 1982, 38, 2279-2283. (b) Vekgova´, H. Collect. Czech.
Chem. Commun. 1986, 51, 930-936. (c) Hui, A.; Fairbanks, A. J .; Nash,
R. J .; de Q. Lilley, P. M.; Storer, R.; Watkin, D. J .; Fleet, G. W. J .
Tetrahedron Lett. 1994, 47, 8895-8898. (d) Seto, H.; Fujioka, S.;
Kamiya, Y.; Yoshida, S. Heterocycles 1998, 48, 2245-2251. (e) Mu-
kaiyama, T.; Shiina, I.; Iwadare, H.; Saitoh, M.; Nishimura, T.;
Ohkawa, N.; Sakoh, H.; Nishimura, K.; Tani, Y.-i.; Hasegawa, M.;
Yamada, K.; Saitoh, K. Chem. Eur. J . 1999, 5, 121-161. (f) Aurreco-
echea, J . M.; Fan˜ana´s, R.; Arrate, M.; Gorgojo, J . M.; Aurrekoetxea,
N. J . Org. Chem. 1999, 64, 1893-1901.
(14) (a) Mahrwald, R.; Costisella, B.; Gu¨ndogan, B. Synthesis 1998,
262-264. (b) Yang, W.; Digits, C. A.; Hatada, M.; Narula, S.; Rozamus,
L. W.; Huestis, C. M.; Wong, J .; Dalgarno, D.; Holt, D. A. Org. Lett.
1999, 1, 2033-2035.
(15) (a) Still, W. C.; Hauck, P.; Kempf, D. Tetrahedron Lett. 1987,
28, 2817-2820. (b) Martin, V. A.; Murray, D. H.; Pratt, N. E.; Zhao,
Y.; Albizati, K. F. J . Am. Chem. Soc. 1990, 112, 6965-6978. (c)
Chowdari, N. S.; Ramachary, D. B.; Cordova, A.; Barbas, C. F.
Tetrahedron Lett. 2002, 43, 9591-9595.
(16) (a) Hanessian, S.; Dube, D.; Hodges, P. J . J . Am. Chem. Soc.
1987, 109, 7063-7067. (b) Woodward, R. B.; Logusch, E.; Nambiar,
K. P.; Sakan, K.; Ward, D. E.; Au-Yeung, B.-W.; Balaram, P.; Browne,
L. J .; Card, P. J .; Chen, C. H.; Cheˆnevert, R. B.; Fliri, A.; Frobel, K.;
Gais, H.-J .; Garratt, D. G.; Hayakawa, K.; Heggie, W.; Hesson, D. P.;
Hoppe, D.; Hoppe, I.; Hyatt, J . A.; Ikeda, D.; J acobi, P. A.; Kim, K. S.;
Kobuke, Y.; Kojima, K.; Krowicki, K.; Lee, V. J .; Leutert, T.; Malchenko,
S.; Martens, J .; Matthews, S. R.; Ong, B. S.; Press: J . B.; RajanBabu,
T. V.; Rousseau, G.; Sauter, H. M.; Suzuki, M.; Tatsuta, K.; Tolbert,
L. M.; Truesdale, E. A.; Uchida, I.; Ueda, Y.; Uyehara, T.; Vasella, A.
T.; Vladuchick, W. C.; Wade, P. A.; Williams, R. M.; Wong, H. N.-C. J .
Am. Chem. Soc. 1981, 103, 3210-3213. (c) Yan, T.-H.; Chu, V.-V.; Lin,
T.-C.; Tseng, W.-H.; Cheng, T.-W. Tetrahedron Lett. 1991, 32, 5563-
5566.
7a 18 by reaction20 with TBDMSCl in the presence of an
excess of imidazole in CH2Cl2; after 5 days, approximately
30% of 7s18 was detected and isolated from the reaction
mixture. That imidazole was responsible for the isomer-
ization was readily demonstrated by monitoring a CDCl3
solution of 7a containing 10 equiv of imidazole (ca. 0.4
1
M) by H NMR; a 1.5:1 equilibrium mixture of 7s and
7a , respectively, resulted within 4 d. Equilibrium was
confirmed by obtaining the same mixture under the
identical conditions but starting from 7s. These solutions
were stable for several weeks with negligible retroaldol
(as judged by the absence of 10) or elimination products
being detected. Subjecting 11a 18 or 11s18 to the same
conditions resulted in a 1.8:1 equilibrium mixture of 11s
and 11a , respectively. In all cases, aqueous workup of
the isomerization reaction provided the mixture of aldols
in >95% yield. Possible isomerization pathways for aldols
7 and 11 are illustrated in Scheme 2. Whereas all possible
stereoisomers can result from a retroaldol-aldol path-
way, keto-enol tautomerism can interconvert only 2
stereoisomers. We conclude that the above isomerizations
of 7 and 11 proceed via imidazole catalyzed enolization21
(18) Ward, D. E.; Sales, M.; Man, C. C.; Shen, J .; Sasmal, P. K.;
Guo, C. J . Org. Chem. 2002, 67, 1618-1629.
(19) Ward, D. E.; Guo, C.; Sasmal, P. K.; Man, C. C.; Sales, M. Org.
Lett. 2000, 2, 1325-1328.
(20) Corey, E. J .; Venkateswarlu, A. J . Am. Chem. Soc. 1972, 94,
6190-6191.
(17) Preliminary communication: Ward, D. E.; Sales, M.; Sasmal,
P. K. Org. Lett. 2001, 3, 3671-3673.
J . Org. Chem, Vol. 69, No. 14, 2004 4809

Ward et al.
TABLE 1. Effect of Solven t on th e Ra te a n d
SCHEME 2
Equ ilibr iu m Ra tio for Isom er iza tion s of 11 a n d 7a
b
starting
aldol
kobs
c
entry
solvent
CDCl3
CDCl3
CD2Cl2
C6D6
Keq (syn/anti) (10-2 h-1
)
t1/2
1
2
3
4
5
6
7
8
9
11a
11s
1.8:1 (11s/11a )
1.8:1 (11s/11a )
1.9:1
6d
0.5 d
12 h
24 h
13 h
130 h
160 h
1 d
33 h
f
5.9
2.9
5.3
0.55
0.43
3d
1.5:1
acetone-d6 1.6:1
DMF-d7
CH3OH
CDCl3
2.1:1
2.0:1
7a
1.5:1 (7s/7a )e
1.6:1e
2.1
f
CD2Cl2
10
11
12
acetone-d6 2.0:1
f
f
f
f
f
f
C6D6
1.4:1
1.7:1
CH3OH
a
Room temperature; [imidazole] ) 0.3 M; [aldol] ) 0.03 M.
The slope of the line obtained by plotting -ln[(Rt - Re)/(Rt + 1)]
b
vs t where Rt is the ratio of aldols (e.g., [11s]/[11a ]; measured by
1H NMR) at time t and Re is the ratio at equilibrium (g8 data
points over the initial 2 half-lives; R2 > 0.99); see the Supporting
Information for details. Relative error estimated at (10%. c Half-
d
life ) t1/2 ) (ln 2)/kobs
.
Determined as above but from only three
data points. e Equilibrium confirmed by obtaining the same ratio
starting from 7s. f Not determined.
solvent effects on the rate and equilibrium ratio were
observed (entries 8-12). For both 7 and 11, the syn
diastereomers (7s and 11s) were predominant at equi-
librium in all solvents examined. Relatively more of the
anti diastereomers (7a and 11a ) were present at equi-
librium in less polar solvents (e.g., C6D6) compared to
polar solvents (e.g., CH3OH) but this effect was modest.
The effect of the base and concentration on the rate of
isomerization of 11s in CDCl3 was also investigated
(Table 2). Isomerizations in the presence of 4-(dimethyl-
amino)pyridine (DMAP) were 1.5-2.5 times more facile
than with imidazole at the same concentration (cf. entry
pairs 1/8, 3/10, 4/11, 7/14). As expected, the rates of
isomerization were independent of the initial concentra-
tion of 11s (cf. entry sets 4/5/6 or 11/12/13) but increased
with increasing concentration of base (cf. entry sets 3/4/7
or 10/11/14). The reaction order in imidazole was deter-
mined to be 1.3 from the slope of the line obtained by
plotting ln(kobs) versus ln([imidazole]) using the data from
entries 1, 3, 4, and 7 in Table 1.25 This noninteger value
implies a complex mechanism for enolization perhaps
involving more than one imidazole molecule.21 Similarly,
the reaction order in DMAP was determined to be near
1 (as expected)21 from the data in entries 9, 10, 11, and
14 in Table 2.26 Although DMAP is somewhat more
effective as an isomerization catalyst, the lower basicity,
lower molecular weight, and much lower cost per mole
of imidazole justify its preferential use.27 By contrast,
reactions were markedly slower in the presence of Et3N
or N-methylimidazole (Table 2, entries 15 and 16) and
isomerization was not observed in the presence of pyri-
dine or 1,2,4-triazole. Attempted isomerization of 11s via
retroaldol under Holt’s conditions (Ti(OiPr)4, CH2Cl2)14b
only gave an elimination product (77%).
because: i) 11 was not detected in the isomerization of
7; ii) 7 was not detected in the isomerization of 11; iii)
imidazole does not catalyze an aldol reaction between 8
and 10 under the same conditions.
Table 1 presents the effect of solvent on the rate22 and
equilibrium ratio for isomerizations of 11 and 7. The rates
of isomerization of 11 in CDCl3 and benzene were ca.
twice those in CD2Cl2 or CH3OH and ca. 10 times greater
than in acetone-d6 or DMF-d7. Reactions in CD3OD were
complicated by the isotope effects resulting from deute-
rium incorporation; however, significant diastereoselec-
tivity in the ketonization of 11e was revealed (i.e., k-1
.
k2) under these conditions. Starting from 11s, the forma-
tion of 3-deuterio 11a was much faster than the forma-
tion of 3-deuterio 11s, and starting with 11a , deuteration
at C-3 in 11a was almost complete (>90%) before the
presence of 11s could be detected (5%). Together, these
results imply that the enolization of 11a is much faster
than the enolization of 11s, at least in methanol solu-
tion.23 The isomerization of 7 was ca. 3 times slower than
11 (cf. entries 2 and 8);24 however, qualitatively similar
(21) For a review on the mechanism of enolization of ketones, see:
(a) Keefe, J . R.; Kresge, A. J . In The Chemistry of Enols; Rappoport,
Z., Ed.; Wiley: New York, 1990, Chapter 7, pp 399-480. (b) Toullec,
J . Adv. Phys. Org. Chem. 1982, 18, 1-77.
(22) See the Supporting Information for details on the determination
of rate constants. The rate for reversible isomerization of two aldols
via their common unstable enol is characterized by kobs which is the
equivalent pseudo-first-order rate constant for equilibration of a
nonequilibrium system (e.g., the rate at which an excess concentration
of 11a reaches its equilibrium value) and is the sum of the rate
constants for the forward and reverse reactions. The forward and
reverse reactions each comprise two steps, enolization (slow) and
ketonization (fast), whose pseudo-first-order rate constants are given
by the product of the rate constant for enolization and the partition
for ketonization. For example, in the isomerization of 11a to 11s via
11e (Scheme 2), kobs ) (k1k2 + k-1k-2)/(k2 + k-1).
We have previous shown that 7a , 7s, and 11s can each
be prepared stereoselectively by aldol reactions of 8 and
(23) If ketonization favors 11a (k-1 . k2) and equilibrium favors
11s (Keq ) (k1k2)/(k-2k-1) ) 1.8), then k1 . k-2 (Scheme 2).
(24) The data does not allow one to distinguish whether these
observed differences are due to different rates of enolization, different
partitions for ketonization, or both.
(25) Linear regression of the four data points used yielded a slope
of 1.28 ( 0.13 (95% confidence interval).
(26) Linear regression of the four data points used yielded a slope
of 0.96 ( 0.23 (95% confidence interval).
4810 J . Org. Chem., Vol. 69, No. 14, 2004

Syn-Anti Isomerization of Aldols by Enolization
TABLE 2. Effect of Ba se a n d Con cen tr a tion on th e Ra te
of Isom er iza tion of 11sa
TABLE 3. Im id a zole-Ca ta lyzed Syn -An ti
Isom er iza tion s of Ald ols 3 a n d 13-19 in CDCl3 Solu tion a
b
[base] [11s]0
kobs
t1/2
(h)
entry
base
imidazole
(M)
(M)
(10-2 h-1
)
1
2
0.10
0.20
0.20
0.40
0.40
0.40
0.80
0.10
0.10
0.20
0.40
0.40
0.40
0.80
0.40
0.015
0.016
0.030
0.030
0.060
0.120
0.030
0.015
0.030
0.030
0.030
0.060
0.120
0.030
0.016
0.040
0.016
0.04
1.5
3.6
3.6
8.1
7.7
7.8
21
3.7
4.4
8.7
19
18
19
32
0.63
0.24
d
f
g
48
19
19
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
8.6
9.0
8.9
3.2
19
16
8.0
3.6
3.9
3.6
2.2
110
289
DMAP
b
starting
aldol
kobs
c
entry
Keq (syn/anti)
(10-2 h-1
)
t1/2
Et3N
1
2
3
4
5
6
7
8
9
13a
14s
15a
16a
17a
3a
1:1.1 (13s/13a ) d
1.5:1 (14s/14a )
3.1:1 (15s/15a ) d
2.0:1 (16s/16a ) d
1:1.6 (17s/17a )
1:2.0 (3s/3a )
8e
0.4 d
0.4 d
0.4 d
3 d
8e
N-methylimidazole 0.40
8e
pyridine
0.22
0.05e
1e
1,2,4-triazole
Ti(OiPr)4
0.036 0.018
2e
1.5 d
4 d
0.7e
g
4.8
1.6
1.7
0.20
0.66
a
b
Room temperature in CDCl3. The slope of the line obtained
by plotting -ln[(Rt - Re)/(Rt + 1)] vs t where Rt is the ratio of
[11s]/[11a ] at time t (measured by 1H NMR) and Re () 1.8) is the
ratio at equilibrium (g8 data points over the initial 2 half-lives;
18sf
17a i
3a i
g
.10 dh
14 h
1:1.6 (17s/17a )
1:2.0 (3s/3a )
1:1.1 (18s/18a )
3.7:1 (12s/12a )d
3.9:1 (12s/12a )d
43 h
41 h
340 h
110 h
10
11
12
18sf,j
12sk
12sl
R2 > 0.99); see the Supporting Information for details. Relative
error estimated at (10%. c Half-life ) t1/2 ) (ln 2)/kobs
.
Only 11s
d
(>95%) after 200 h. e The observed solubility in CDCl3 (ca. 3
mg/mL) as determined by 1H NMR with 11s as internal standard.
a
At room temperature; [imidazole] ) 0.3 M; [aldol] ) 0.03 M.
The slope of the line obtained by plotting -ln[(Rt - Re)/(Rt + 1)]
g
f Only 11s (>95%) after 140 h. Reaction in CH2Cl2 at 0 °C for 5
b
vs t where Rt is the ratio of [11s]/[11a ] at time t (measured by 1H
NMR) and Re () 1.8) is the ratio at equilibrium (g8 data points
h gave an elimination product in 77% yield (cf. ref 14b).
over the initial 2 half-lives; R2 > 0.99); see the Supporting
10 under appropriate conditions.18 Isomerization of 11s
provides a useful route to the elusive anti-anti diaste-
reoisomer 11a , a process that is viable on preparative
scale because these diastereomers can be separated by
fractional crystallization. Thus, MgBr2-mediated aldol
reaction of the trimethylsilyl enol ether of 9 with 10 gave
a 3:1 mixture of 11s and 11a , respectively, after work-
up.18 Crystallization of the crude mixture from methanol
gave 11s in 61% overall yield.28 Isomerization of the so-
obtained 11s with 1 M imidazole in CH2Cl2 for 24 h gave
a 2:1 mixture of 11s and 11a , respectively, in >95% yield.
Crystallization of this mixture gave 11s (ca. 60%) and a
mother liquor highly enriched in 11a (4-8:1). After
subjecting 11s to a second round of isomerization and
crystallization, fractionation of the combined mother
liquors gave 11a in >50% overall yield from 10.
The scope of the imidazole-catalyzed syn-anti isomer-
ization of aldols was examined by treating 3,18 12-17,18
and 1829 with imidazole in CDCl3 solution (Table 3).
Isomerizations of the diastereoisomeric aldols 13-16
gave mutually exclusive products, thereby clearly ruling
out a retroaldol mechanism. The syn diastereomer pre-
dominated at equilibrium for three of the four syn-anti
diastereomer pairs (entries 1-4). The rates of isomer-
ization were not determined accurately for this series of
aldols but were qualitatively similar30 to those for 7 and
11 under similar conditions although the rate of isomer-
Information for details. c Half-life ) t1/2 ) (ln 2)/kobs
.
Equilibrium
d
confirmed by obtaining the same ratio starting from the syn
diastereomer. e Estimated from two to three intermediate data
points and the time required to reach equilibrium (ref 30). f A 9:1
g
h
mixture of 18s/18a was used. Not determined. >80% 18s after
10 days. i [Imidazole] ) 0.7 M. j At 60 °C in benzene-d6; [imidazole]
k
l
) 1 M. [Imidazole] ) 0.4 M. [Imidazole] ) 0.8 M.
ization of 16 was considerably slower. To explore the
importance of the thiopyranone moiety on the isomer-
ization, the structurally related aldols 3, 17, and 18 were
examined (entries 5-10). The isomerization of 17a with
0.3 M imidazole was faster than that of the cyclohex-
anone analogue 3a suggesting that the presence of sulfur
facilitates enolization by a factor of ca. 3. By contrast,
isomerization of the acyclic analogue 18 was exceedingly
slow under these conditions. As expected, the reactions
were faster at higher imidazole concentrations but isomer-
ization of 18s was still very slow at room temperature.
31
However, heating a solution of 18s in benzene-d6
containing 1.0 M imidazole at 60 °C for 9 days produced
1.1:1 equilibrium32 mixture of 18a and 18s, respectively.
Despite these much harsher conditions, elimination was
negligible and only a small amount of retroaldol occurred
(as indicated by the presence of benzaldehyde, ca. 4%).
Finally, isomerization of the methoxymethyl (MOM)
ether derivative 12s was 30-40 times less facile than
11s (cf. entries 4 and 7 in Table 2 with entries 11 and 12
in Table 3) demonstrating the importance of the hydroxyl
(27) For some substrates, DMAP was found to be much more
effective than imidazole: Akinnusi, O. T. Unpublished results.
(28) Silica gel chromatography (25-50% ethyl acetate in hexane)
of the mother liquor at this stage gave 11s (6%) and 11a (22%).
(29) Heathcock, C. H.; Davidsen, S. K.; Hug, K. T.; Flippin, L. A. J .
Org. Chem. 1986, 51, 3027-3037.
(31) Chloroform solutions of imidazole were not stable to prolonged
heating.
(32) The ratio was unchanged after 13 days. Holt et al. (ref 14b)
obtained a 5:4 mixture of 18a and 18s by isomerization in the presence
of Ti(O-i-Pr)4 (retroaldol mechanism).
(30) Assuming a pseudo-first-order process and at least 5 half-lives
required to reach equilibrium.
J . Org. Chem, Vol. 69, No. 14, 2004 4811

Ward et al.
SCHEME 3
ally isomerized in CDCl3 solution in the presence of
imidazole (0.4 M), and the relative concentration of the
components as a function of time was monitored by H
1
NMR. In each case, the same equilibrium ratio of 36:31:
22:11 of 19a c, 19sc, 19a t, and 19st, respectively, was
obtained within 5-7 days. Isomerizations of the four
diastereomers of 20 under the same conditions were
much slower than with 19. An equilibrium mixture of
62:28:7.5:2.5 of 20a c, 20sc, 20a t, and 20st, respectively,
was obtained from 20st within 3 weeks; however, more
than 6 weeks were required to achieve the same mixture
starting from 20sc or 20a t and 20a c had not reached
equilibrium after 2 months. For both 19 and 20, the 1,3-
cis diastereomers (i.e., a c and sc) predominated over the
1,3-trans diastereomers (i.e., a t and st) with the anti
diastereomer predominating over the syn diastereomer
(i.e., a c > sc and a t > st) at equilibrium. Comparing the
equilibrium ratios reveals a greater difference in ther-
modynamic stability among the diastereomers of 20 than
among those of 19 presumably because of the reduced
steric interaction of the axial substituent in the 1,3-trans
diastereomers 19a t and 19st compared to 20a t and 20st
due to presence of the sulfur atom.
SCHEME 4
Exposure of individual diastereomers of 21-24 to
imidazole led in each case to an equilibrium mixture
containing only those diastereomers that could result
from keto-enol tautomerism (Table 4).35 These isomer-
izations proceeded without significant elimination or
retroaldol even after prolonged reaction (>100 days in
some cases).36 A number of synthetically useful aspects
were noted from the effect of structure and reaction
conditions on the process.
SCHEME 5
The isomerization of 21a a with 1 M imidazole in CDCl3
gave an equilibrium mixture within 10 days where the
21ss diastereomer predominated (Table 4, entry 1).37
1
However, monitoring the progress of the reaction by H
NMR revealed that 21a s comprised >65% of the mixture
after 1 day indicating that isomerization of 21a a to 21a s
was much faster than from 21a s to 21ss (Scheme 4). This
observation was exploited by stopping the reaction after
1 day whereby 21a s could be obtained in good yield from
21a a (Table 4, entry 2). Isomerization of 22a a in several
solvents revealed that the ratio of the Cs-symmetrical
aldols (22ss, 22a a ) at equilibrium was solvent dependent
(1.3-3.8:1) but the unsymmetrical bisaldol 22a s consis-
tently comprised 60-65% of the mixture (Table 4, entries
3-7).38 Optimization of the conditions resulted in a
convenient preparation of 22ss from 22a a (Table 4, entry
7). Sequential two-directional aldol reactions of 8 with
10 allows for rapid (two to four steps) and stereoselective
syntheses of 21a a , 21ss, 22a a , and 22a s.39 Although we
have been unable to obtain 21a s or 22ss directly by this
approach, isomerizations of the readily available 21a a
group to the facility of reaction and suggesting the
possibility of intramolecular catalysis,21 perhaps via
hydrogen bonding to the carbonyl group.
To further explore the generality of process and the
effects of structure on reactivity, isomerizations of the
R-methyl aldols 19 and 2033 (Scheme 3)34 and the bisaldol
derivatives 21-2419 (Schemes 4 and 5) were investigated.
In these cases, isomerization can occur through regio-
isomeric enols producing a four-component equilibration.
The diastereomers 19st, 19a c, and 19a t were individu-
(35) A retroaldol mechanism is clearly ruled out. For example, there
are a total of 20 diastereomers possible for bisadol 21 (or 22). In
principle, all diastereomers could be produced from 21 or 22 via a
retroaldol-aldol mechanism whereas enolization can only interconvert
the three diastereomers shown in Schemes 4 and 5 (note: 21a s is
identical to 21sa and 22a s is the enantiomer of 22sa ).
(33) Duhamel, P.; Cahard, D.; Quesnel, Y.; Poirier, J .-M. J . Org.
Chem. 1996, 61, 2232-2235.
(36) In some cases, 1-5% of benzaldehyde (signifying retroaldol) was
observed after prolonged reaction.
(34) Aldols prepared by reactions of benzaldehyde with the LDA
generated Li enolates from 2-methylcyclohexanone and from tetrahy-
dro-3-methylthiopyran-4-one. The individual diastereomers were sepa-
rated by chromatography. The relative configurations of 20a t and 20st
were incorrectly assigned in the literature; see the Supporting Infor-
mation for details.
(37) Entropy of symmetry favors 21a s (C1) by a factor of 2 over 21a a
and 21ss (C2); thus. neglecting differences in enthalpy among the
diastereomers, a 1:2:1 mixture of 21a a /21a s/21ss would be expected.
(38) Neglecting differences in enthalpy among the diastereomers,
a 1:2:1 mixture of 22a a /(()-22a s/22ss would be expected.
4812 J . Org. Chem., Vol. 69, No. 14, 2004

Syn-Anti Isomerization of Aldols by Enolization
TABLE 4. Im id a zole-Ca ta lyzed Syn -An ti Isom er iza tion s of Bisa ld ols 21-24a
entry
starting aldol
[imidazole] (M)
solvent
ratio of aldols at equilibrumb
t (d)c
1
2
3
4
5
6
7
8
9
21a a
21a a
22a a
1.0
1.0
0.8
0.4
0.8
0.4
1.0
0.4
0.4
CDCl3
CH2Cl2
CDCl3
CDCl3
CH2Cl2
C6D6
acetone
CDCl3
CDCl3
6:39:55 (21a a /21a s/21ss)
10
1
4
29:60:11 (21a a /21a s/21ss)d
18:60:22 (22a a /(()-22a s/22ss)e
18:61:21 (22a a /(()-22a s/22ss)e
14:66:20 (22a a /(()-22a s/22ss)
12:61:27 (22a a /(()-22a s/22ss)
9:63:28 (22a a /(()-22a s/22ss)f
13:24:23:40 (23a a /23a s/23sa /23ss)e
7:47:46 (24a a /(()-24a s/24ss)e
8
10
10
8
35
120
23a a
24a a
a
b
d
At room temperature. Measured by 1H NMR. c The approximate time required to reach equilibrium. A nonequilibrium mixture.
Isolated yields (13 mg scale): 21a a (25%), 21a s (45%), 21ss (10%). e Equilibrium confirmed by obtaining the same ratio starting from the
a s or ss diastereomers. f Isolated yields (46 mg scale): 21a a (8%), 21a s (63%), 21ss (26%).
and 22a a constitute synthetically viable routes to these
diastereomers. Thus, the combination of stereoselective
aldol reactions and isomerization allows the synthesis of
any the six diastereomers 21 and 22 in only a few steps
from common building blocks.
mechanistic subtleties have been revealed.21 Compara-
tively few mechanistic studies of enolization of â-oxy
ketones (i.e., hydroxy, alkoxy, acyloxy) have appeared,
and most of these are concerned with the ensuing
elimination reaction.41 Our results clearly suggest that
enolization is facilitated by a â-hydroxy group, and this
is most easily accommodated by assuming hydrogen
bonding to the carbonyl group. Similarly, the enhanced
enolization observed for anti aldol 11a relative to syn
aldol 11s can be rationalized by “tighter” hydrogen
bonding in the anti diastereomer.42 Although intramo-
lecular acid- and base-catalyzed enolization of â-keto
acids and o-hydroxyacetophenones have been studied,21,43
we are unaware of any reports of similar catalysis by a
â-hydroxy group. However, base-catalyzed enolization of
hydroxyacetone (at the CH3 group)44 is 3.4 times faster
than acetone45 which, in turn, is 10-20 times faster than
methoxyacetone46 establishing the beneficial effect of an
R-hydroxy group. The order of reactivity observed for
isomerizations of 18, 3, and 17 (relative rates ca. 0.1, 1,
and 3, respectively) can be explained by consideration of
the rates for base-catalyzed enolization of 3-pentanone,
cyclohexanone, and cyclohexanones with polar substitu-
ents at the 4 position (e.g., -OH, -OMe, -OCOMe,
etc.).47 An interesting question concerns the requirements
for isomerization in preference to retroaldol or elimina-
tion. Base-catalyzed retroaldol requires abstraction of the
proton from the hydroxy group and is facilitated by
stronger bases and in protic solvents.9 Specific base-
catalyzed eliminations from â-oxy ketones with poor
Comparing the isomerizations of 22-24 with 0.4 M
imidazole in CDCl3 showed that the proportion of the
syn-syn diastereomer at equilibrium increased with
increasing MOM protection (22ss, 21%; 23ss, 40%; 24ss,
46%) suggesting that manipulation of the hydroxy group
could be a useful synthetic tool in these processes.
However, the isomerizations of the MOM-protected de-
rivatives 23 and 24 were much slower than 22. These
trends are consistent with those observed in the isomer-
izations of 11 and 12. We were interested in determining
if the hydroxy groups in 19-23 imparted any kinetically
significant regioselectivity or stereoselectivity to the
isomerizations because such effects could be exploited
synthetically. Obtaining the pseudo-first-order rate con-
stants for isomerization (i.e., kn, k-n; n ) 1-4) for the
four-component systems 19-24 is much more difficult
than for the two-component systems (e.g., 11-18).40
Although these rate constants must be interpreted with
caution,24 the data suggest that enolizations of 19, 20,
and 23 occur with only modest regioselectivity (<2.5:1).
Our data firmly establish that isomerization proceeds
by an enolization mechanism; however, the mechanistic
details of the enolization step are uncertain. Acid- and
base-catalyzed enolization of ketones has been exten-
sively studied for more than a century, and various
(41) (a) Noyce, D. S.; Reed, W. L. J . Am. Chem. Soc. 1959, 81, 624-
628. (b) Stiles, M.; Wolf, D.; Hudson, G. V. J . Am. Chem. Soc. 1959,
81, 628-632. (c) Fedor, L. R. J . Am. Chem. Soc. 1967, 89, 4479-4482.
(d) Fedor, L. R. J . Am. Chem. Soc. 1969, 91, 908-913. (e) Cavestri, R.
C.; Fedor, L. R. J . Am. Chem. Soc. 1970, 92, 4610-4613. (f) Fedor, L.
R.; Glave, W. R. J . Am. Chem. Soc. 1971, 93, 985-989. (g) Hupe, D.
J .; Kendall, M. C. R.; Spencer, T. A. J . Am. Chem. Soc. 1972, 94, 1254-
1263. (h) Hupe, D. J .; Kendall, M. C. R.; Sinner, G. T.; Spencer, T. A.
J . Am. Chem. Soc. 1973, 95, 2260-2270. (i) Hupe, D. J .; Kendall, M.
C. R.; Spencer, T. A. J . Am. Chem. Soc. 1973, 95, 2271-2278. (j) Hupe,
D. J .; Wu, D. J . Am. Chem. Soc. 1977, 99, 7653-7659. (k) Perera, S.
K.; Fedor, L. R. J . Am. Chem. Soc. 1979, 101, 7390-7393. (l) Mayer,
B. J .; Spencer, T. A. J . Am. Chem. Soc. 1984, 106, 6349-6354.
(42) Kitamura, M.; Nakano, K.; Miki, T.; Okada, M.; Noyori, R. J .
Am. Chem. Soc. 2001, 123, 8939-8950.
(39) The chiral bisaldols used in this study were racemic (ref 19).
The stereoselective syntheses of 21a a , 21ss, 22a s, and 22sa require
the use of enantiomerically pure 10. This work will be reported
separately.
(40) See the Supporting Information for details on the determination
of rate constants. The kinetic model (Scheme 3 or 5) assumes each
isomerization reaction is first order with respect to the particular
“starting” aldol concentration at a fixed concentration of imidazole, as
shown. The forward and reverse reactions each comprise two steps,
enolization (slow) and ketonization (fast), whose pseudo-first-order rate
constants are given by the product of the rate constant for enolization
and the partition for ketonization. For example, in the isomerization
of 19a t to 19a c via 19e (Scheme 3), k1 ) (kekk′)/(kk + kk′). The ratios
kn/k-n (n ) 1-4) are governed by the corresponding equilibrium ratios
for the component pair of aldols. The sums kn + k-n (n ) 1-4) can be
used to compare the ease of the isomerization between two components
within a four-component equilibration or to compare with the rate of
isomerization in a two-component equilibration (i.e., kobs) (ref 22). In
contrast to a two-component system, the rate that equilibrium is
achieved in a four-component system depends on the initial distribution
of the components. The average of kn + k-n (n ) 1-4) can be used as
an overall indication of the facility of equilibration in a four-component
system and can be compared to the rate of isomerization in a two-
component equilibration (i.e., kobs) (ref 22).
(43) Kirby, A. J . Adv. Phys. Org. Chem. 1980, 18, 183-278.
(44) Crugeiras, J .; Richard, J . P. J . Am. Chem. Soc. 2004, 126, 5164-
5173.
(45) (a) Keeffe, J . R.; Kresge, A. J .; Schepp, N. P. J . Am. Chem. Soc.
1990, 112, 4862-4868. (b) Hine, J .; Kaufmann, J . C.; Cholod, M. S. J .
Am. Chem. Soc. 1972, 94, 4590-4595.
(46) Hine, J .; Hampton, K. G.; Menon, B. C. J . Am. Chem. Soc. 1967,
89, 2664-2668.
(47) Hegarty, A. F.; Dowling, J . P.; Eustace, S. J .; McGarraghy, M.
J . Am. Chem. Soc. 1998, 120, 2290-2296.
J . Org. Chem, Vol. 69, No. 14, 2004 4813

Ward et al.
leaving groups (e.g., methoxy, hydroxy) with rate-limiting
breakdown of an equilibrium concentration of the enolate
have been observed.48 For these compounds, keto-enol
tautomerism is enhanced under conditions that minimize
the equilibrium concentration of enolate (i.e., use of weak
bases). Further, the rate of elimination from enolates of
intramolecularly hydrogen bonded aldols should be at-
tenuated because of poor π-σ* orbital overlap. Finally,
the observed influence of the base on the rates of
isomerization is difficult to accommodate. The rates of
general base-catalyzed enolization (rate-limiting abstrac-
tion of the R-C-H proton) typically increase with in-
creasing pKa of the base.21 Low basicity and steric
arguments can be invoked to account for the low reac-
tivities of pyridine (pKa 5.21) and Et3N (pKa 10.75) with
11s (Table 2), respectively. This approach seems less
appropriate to explain the much lower reactivity of
N-methylimidazole (pKa 7.06) compared to imidazole (pKa
6.95)49,50 and the reactivity of DMAP (pKa 9.2) seems to
rule out a requirement for a hydrogen bond donor.
Although imidazole and DMAP may catalyze enolization
by different mechanisms, their unique ability to catalyze
isomerization of aldols raises the possibility of nucleo-
philic catalysis.51
unavailable by direct methods (e.g., 11a , 21a s, and 22ss).
Although the equilibrium ratios are typically modest,
synthetically useful amounts of material can be obtained
by recycling if necessary.
Exp er im en ta l Section 53
(3S)-r el-3-[(R)-(6R)-1,4-Dioxa -8-th ia sp ir o[4.5]d ec-6-yl-
(h yd r oxy)m eth yl]tetr a h yd r o-4H-th iop yr a n -4-on e (11a ).
Crystallization of the crude product from aldol reaction of 10
(3.47 g, 18.5 mmol) and the trimethylsilyl enol ether of 8 (7.00
g, 37.0 mmol) in the presence of MgBr2‚OEt2 (14.3 g, 55.4
mmol) as previously described18 gave solid 11s (3.49 g, 62%)
and a mother liquor containing a ca. 5:1 mixture of 11a /11s.
A solution of the solid 11s (3.49 g, 11.5 mmol) and imidazole
(6.8 g, 0.10 mol) in CH2Cl2 (100 mL) was allowed to stand for
24 h. The mixture was washed with aqueous citric acid (0.2
M) and the aqueous layer extracted with CH2Cl2. The com-
bined organic layers were dried over Na2SO4 and concentrated
to give a 2:1 mixture of 11s/11a (3.40 g, 98%). Crystallization
of the mixture from methanol gave solid 11s (2.08 g, 61%) and
a mother liquor containing a ca. 8:1 mixture of 11a /11s. The
solid 11s (2.08 g) was subjected to isomerization as above to
give solid 11s (1.22 g, 58%) and a mother liquor containing a
ca. 6:1 mixture of 11a /11s. The combined mother liquors from
the above crystallizations (3) (4.1 g; a 6:1 mixture of 11a /11s)
were fractionated by MPC (25-50% ethyl acetate in hexane)
to give 11s (437 mg; a total of 1.66 g of 11s isolated, 30%
overall from 10) and 11a (2.88 g, 51% from 10). Spectral data
was identical to that reported previously.18
Con clu sion
(3R,5S)-r el-3,5-Bis[(S)-(6R)-1,4-d ioxa -8-t h ia sp ir o[4.5]-
d e c-6-ylh yd r oxym e t h yl]t e t r a h yd r o-4H -t h iop yr a n -4-
on e (21a s). Bisaldol 21a a (13 mg, 0.026 mmol) was added to
a solution of imidazole (136 mg, 2.00 mmol) in CH2Cl2 (2.0 mL).
After 20 h at room temperature, the mixture was diluted with
aqueous citric acid (0.1 M) and extracted with CH2Cl2 (×3).
The combined organic layers were dried over Na2SO4 and
concentrated to give a 6.2:3:1 mixture (12 mg) of 21a s, 21a a ,
and 21ss, respectively. Fractionation of the mixture by PTLC
(2% MeOH in CH2Cl2; multiple elution) gave 21ss (1.5 mg,
12%), 21a a (3.5 mg, 27%), and the titled compound (6 mg,
46%): IR (DRIFT) νmax 3518, 2917, 1694, 1427, 1153, 1131,
1101, 1042 cm-1; 1H NMR (500 MHz, C6D6) δ 4.69 (1H, ddd, J
) 1.5, 3.5, 6 Hz, HC-1′), 4.17 (1H, ddd, J ) 3.5, 7.5, 10 Hz,
HC-1′′), 3.45-3.24 (6H, m, H2CO × 4), 3.24-3.02 (9H, m,
H2CO, HC-2, HC-3, HC-5, HC-6, HC-7′, H2C-7′′), 3.13 (1H, d,
J ) 1.5 Hz, HOC-1′), 2.85 (1H, dd, J ) 11.5, 13 Hz, HC-2 or
HC-6), 2.77 (1H, d, J ) 10 Hz, HOC-1′′), 2.72 (1H, br d, J )
13.5 Hz, HC-7), 2.58 (1H, m, J ) 4, 8, 13 Hz, HC-9′′), 2.52
(1H, ddd, J ) 3.5, 3.5, 13 Hz, HC-2 or HC-6), 2.49 (1H, ddd, J
) 3, 10.5, 13.5 Hz, HC-9′), 2.25-2.24 (2H, m, HC-9′, HC-9′′),
2.21 (1H, ddd, J ) 3.5, 4, 9.5 Hz, HC-6′), 2.16 (1H, ddd, J )
3.5, 7, 7.5 Hz, HC-6′′), 1.63 (1H, ddd, J ) 3, 6, 13.5 Hz, HC-
10′), 1.48 (1H, ddd, J ) 3.5, 10.5, 13.5 Hz, HC-10′), 1.45-1.36
(2H, m, H2C-10′′); (500 MHz, CDCl3) δ 4.53 (1H, br d, J ) 8.5
Hz, HC-1′), 4.17-3.89 (9H, m, HC-1′′, H2CO ×4), 3.33 (1H,
ddd, J ) 3.5, 4, 13.5 Hz, HC-2), 3.27 (1H, ddd, J ) 3.5, 3.5, 12
Hz, HC-5), 3.21 (1H, dd, J ) 12, 13 Hz, HC-6), 3.10 (1H, d, J
) 1 Hz, HOC-1′), 3.09-2.50 (9H, m, HC-3, H2C-7′, H2C-7′′,
H2C-9′, H2C-9′′), 2.97 (1H, d, J ) 9.5 Hz, HOC-1′′), 2.91 (1H,
ddd, J ) 3, 3.5, 13 Hz, HC-6), 2.85 (1H, dd, J ) 11.5, 13.5 Hz,
HC-2), 2.22 (1H, ddd, J ) 3, 7.5, 7.5 Hz, HC-6′′), 2.16-2.07
(2H, m, HC-6′, HC-10′ or HC-10′′), 1.88 (1H, ddd, J ) 3,
8.5, 13.5 Hz, HC-10′ or HC-10′′), 1.78-1.71 (2H, m, HC-10′,
HC-10′′); 13C NMR (125 MHz, CDCl3) δ 215.5 (s), 110.2 (s),
109.0 (s), 70.9 (d, C-1′), 66.9 (d, C-1′′), 64.9 (t), 64.8 (t), 64.3
(t), 64.0 (t), 57.5 (d), 57.0 (d), 47.5 (d), 46. (d), 37.1 (t), 35.8 (t),
35.5 (t), 34.6 (t), 29.3 (t), 26.8 (t), 26.7 (t), 26.6 (t); LRMS (EI)
m/z (relative intensity) 492 ([M]+, 1), 304 (8), 188 (15), 159
In conclusion, imidazole has been shown to be a very
effective catalyst for syn-anti isomerization of aldols via
an enolization mechanism. DMAP is also effective isomer-
ization catalyst but the lower basicity, molecular weight,
and cost per mole of imidazole justify its preferential use
for most substrates. The process is applicable to a broad
range of substrates52 and is particularly facile for tet-
rahydro-4H-thiopyran-4-on-derived aldols. The presence
of the aldol hydroxy group greatly facilitates isomeriza-
tion but has only a small effect on enolization regio-
selectivity. Most substrates reach equilibrium within
0.3-3 days at ambient temperature in chloroform or
benzene containing 0.3-1 M imidazole. A more rapid
equilibration can be achieved with gentle warming but
the generality of this approach was not examined.
Isomerizations are high yielding and occur with little or
none of the usual byproducts arising from competing
elimination or retroaldol reactions even after prolonged
reaction. Consequently, this isomerization process can
provide convenient synthetic access to diastereomers
(48) Methoxy-2-butanone,41d cis-octahydro-8a-hydroxy-4a-methyl-
2(1H)-naphthalenone,41h and prostaglandin E241k (note: intramolecular
hydrogen bonding is not possible for these aldols).
(49) The rate constants for imidazole- and N-methylimidazole-
catalyzed enolizations of â-oxy ketones were very similar (see ref 41g
and h).
(50) For imidazole-catalyzed enolization of acetone, see: (a)
Mel’nichenko, I. V.; Yasnikov, A. A. Ukr. Khim. Zh. 1964, 30, 723-
728. (b) Bender, M. L.; Williams, A. J . Am. Chem. Soc. 1966, 88, 2502-
2508. For a discussion of the possibility of imidazole acting as a
bifunctional enolization catalyst, see refs 41g, 51a, and: (c) Banks, B.
E. C. J . Chem. Soc. 1962, 63-71. (d) Breslow, R.; Graff, A. J . Am.
Chem. Soc. 1993, 115, 10988-10989.
(51) (a) Bruice, P. Y.; Bruice, T. C. J . Am. Chem. Soc. 1978, 100,
4793-4801. (b) Bruice, P. Y. J . Am. Chem. Soc. 1983, 105, 4982-4996.
(c) Bruice, P. Y. J . Am. Chem. Soc. 1990, 112, 7361-7368.
(52) For other examples, see: (a) Notz, W.; Tanaka, F.; Watanabe,
S.; Chowdari, N. S.; Turner, J . M.; Thayumanavan, R.; Barbas, C. F.
J . Org. Chem. 2003, 68, 9624-9634. (b) Williamson, R. T.; Marquez,
B. L.; Sosa, A. C. B.; Koehn, F. E. Magn. Reson. Chem. 2003, 41, 379-
385.
(53) See the Supporting Information for general methods and
procedures.
4814 J . Org. Chem., Vol. 69, No. 14, 2004

Syn-Anti Isomerization of Aldols by Enolization
(12), 132 (73), 99 (100), 86 (24), 55 (21); HRMS m/z calcd for
Hz, HC-6′, HC-6′′), 2.09 (2H, ddd, J ) 3, 6.5, 14 Hz, HC-10′,
HC-10′′), 1.77 (2H, ddd, J ) 3.5, 10, 14 Hz, HC-10′, HC-10′′);
13C NMR (125 MHz, CDCl3) δ 213.9 (s, C-4), 109.9 (s × 2, C-5′,
C-5′′), 66.9 (d × 2, C-1′, C-1′′), 64.6 (t × 2, CH2O), 64.5 (t × 2,
CH2O), 57.8 (d × 2, C-3, C-5), 46.6 (d × 2, C-6′, C-6′′), 35.5 (t
× 2, C-10′, C-10′′), 34.3 (t × 2, C-2, C-6), 27.5 (t × 2, C-7′, C-7′′),
26.8 (t × 2, C-9′, C-9′′); LRMS (EI) m/z (relative intensity) 492
([M]+, 1), 188 (14), 159 (11), 132 (73), 100 (10), 99 (100), 86
(24), 55 (8); HRMS m/z calcd for C21H32O7S3 492.1310, found
492.1311 (EI).
C
21H32O7S3 492.1310, found 492.1310.
(3R ,5S )-r el -3-[(R )-(6S )-1,4-D io x a -8-t h ia s p ir o [4.5]-
d e c-6-ylh yd r oxym e t h yl]-5-[(S )-(6R )-1,4-d ioxa -8-t h ia -
spiro[4.5]dec-6-ylhydroxymethyl]tetrahydro-4H-thiopyran-
4-on e (22ss). Bisaldol 22a a (46 mg, 0.093 mmol) was added
to a solution of imidazole (680 mg, 10.0 mmol) in acetone (10
mL) at room temperature. After 8 days, the mixture was
diluted with aqueous citric acid (0.1 M) and extracted with
CH2Cl2 (×3). The combined organic layers organic were dried
over Na2SO4 and concentrated to give a 6.2:3.0:1 mixture of
22a s, 22ss, and 22a a , respectively. Fractionation of the
mixture by PTLC (2% MeOH in CH2Cl2; multiple elution) gave
22a s (29 mg, 63%), 22a a (4 mg, 8%), and the titled compound
22ss (12 mg, 26%): IR (DRIFT) νmax 3511, 2916, 1702, 1427,
1153, 1102, 1038, 734 cm-1; 1H NMR (500 MHz, CDCl3) δ 4.57
(2H, ddd, J ) 2.5, 4.5, 6.5 Hz, HC-1′, HC-1′′), 4.09-3.96 (8H,
m, H2CO × 4), 3.25 (2H, br d, J ) 13.5 Hz, HC-2, HC-6), 3.09
(2H, ddd, J ) 4.5, 6.5, 11.5 Hz, HC-3, HC-5 [J HC-2-HC-3 ) 4.5,
11.5 Hz]), 3.06 (2H, d, J ) 2.5 Hz, HO × 2), 2.98 (2H, dd, J )
9.5, 14 Hz, HC-7′, HC-7′′), 2.91 (2H, dd, J ) 11.5, 13.5 Hz,
HC-2, HC-6), 2.78-2.71 (4H, m, HC-7′, HC-7′′, HC-9′, HC-9′′),
2.64 (2H, m, HC-9′, HC-9′′), 2.11 (2H, ddd, J ) 3.5, 4.5, 9.5
Ackn owledgm en t. Financial support from the Natu-
ral Sciences and Engineering Research Council (Canada)
is gratefully acknowledged.
Su p p or tin g In for m a tion Ava ila ble: General experimen-
tal and procedures for 12, 19, 20, 23, and 24, determination
of relative configurations, determination of rate constants, and
1H NMR spectra for all reported compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O049626O
J . Org. Chem, Vol. 69, No. 14, 2004 4815