Baker's Yeast-Mediated Reductions of α-Keto Esters and an α-Keto-β-Lactam. Two Routes to the Paclitaxel Side Chain

Article abstract of DOI:10.1021/jo9900681

Baker's yeast (Saccharomyces cerevisiae) has been used to reduce a series of alkyl esters derived from pyruvate and benzoylformate. Both the yield and enantioselectivities of these reductions were maximized when methyl esters were used, and the (R)-alcohols were isolated in all instances. Yeast-mediated ester hydrolysis was a significant side reaction for products derived from long-chain alcohols. In the case of ethyl benzoylformate, the addition of methyl vinyl ketone increased the enantioselectivity of the reduction. These reductions were applied to two syntheses of the paclitaxel C₁₃ side chain [(2R,3S)-N-benzoyl-3-phenylisoserine]. In the first, a racemic α-keto-β-azido ester was reduced by whole cells of Baker's yeast to afford a diastereomeric mixture in which the desired product predominated and could be isolated chromatographically. In the second, an easily synthesized α-keto-β-lactam was reduced by yeast cells to afford the desired eis isomer as well as the undesired trans diastereomer. Substituting a yeast strain deficient in fatty acid synthase in this reduction suppressed formation of the trans diastereomer. These results suggest that a single enzyme is responsible for both the D- and L-cis-alcohols resulting from reduction of the α-keto-β-lactam. All of the yeast strains used in this project are available commercially, and these biocatalytic reductions require only common laboratory equipment.

Full text of DOI:10.1021/jo9900681

J . Org. Chem. 1999, 64, 6603-6608
6603
Ba k er ’s Yea st-Med ia ted Red u ction s of r-Keto Ester s a n d a n
r-Keto-â-La cta m . Tw o Rou tes to th e P a clita xel Sid e Ch a in
Margaret M. Kayser,*,†,‡ Marko D. Mihovilovic,†,§, J eff Kearns,† Anton Feicht,† and
J on D. Stewart*,§,|
Department of Chemistry, University of New Brunswick, P.O. Box 5050, Saint J ohn,
New Brunswick E2L 4L5, Canada, and Department of Chemistry, University of Florida,
Gainesville, Florida 32611
Received J anuary 14, 1999
Baker’s yeast (Saccharomyces cerevisiae) has been used to reduce a series of alkyl esters derived
from pyruvate and benzoylformate. Both the yield and enantioselectivities of these reductions were
maximized when methyl esters were used, and the (R)-alcohols were isolated in all instances. Yeast-
mediated ester hydrolysis was a significant side reaction for products derived from long-chain
alcohols. In the case of ethyl benzoylformate, the addition of methyl vinyl ketone increased the
enantioselectivity of the reduction. These reductions were applied to two syntheses of the paclitaxel
C13 side chain [(2R,3S)-N-benzoyl-3-phenylisoserine]. In the first, a racemic R-keto-â-azido ester
was reduced by whole cells of Baker’s yeast to afford a diastereomeric mixture in which the desired
product predominated and could be isolated chromatographically. In the second, an easily
synthesized R-keto-â-lactam was reduced by yeast cells to afford the desired cis isomer as well as
the undesired trans diastereomer. Substituting a yeast strain deficient in fatty acid synthase in
this reduction suppressed formation of the trans diastereomer. These results suggest that a single
enzyme is responsible for both the D- and L-cis-alcohols resulting from reduction of the R-keto-â-
lactam. All of the yeast strains used in this project are available commercially, and these biocatalytic
reductions require only common laboratory equipment.
In tr od u ction
tunately, some of these reductases possess overlapping
substrate specificities but with opposite enantioselectivi-
ties, which often diminishes the optical purities of the
alcohol products when a compound is accepted by mul-
tiple enzymes. Traditional approaches to overcoming
these problems have included variations in substrate
structure18 and concentration,19 the use of organic sol-
vents in place of water,20-24 heat treatment,25 and the
addition of exogenous modifiers.12,26-30 With the recent
Baker’s yeast (Saccharomyces cerevisiae) has been the
most popular whole-cell biocatalyst, particularly for
asymmetric reductions of carbonyl compounds.1-5 Reduc-
tions catalyzed by this organism tolerate a large diversity
of carbonyl substrates, and side reactions are rarely
observed. This broad substrate acceptance is due to the
presence of a number of reductase enzymes, several of
which have been isolated and characterized.6-17 Unfor-
(15) Nakamura, K.; Kondo, S.; Kawai, Y.; Nakajima, N.; Ohno, A.
Biosci. Biotechnol. Biochem. 1994, 58, 2236-2240.
(16) Kuhn, A.; van Zyl, C.; van Donder, A.; Prior, B. A. Appl.
Environ. Microbiol. 1995, 61, 1580-1585.
* To whom correspondence should be addressed.
† University of New Brunswick.
‡ Phone: (506)648-5576.Fax: (506)648-5650.E-mail: kayser@unbsj.ca.
§ University of Florida.
(17) Wada, M.; Kataoka, M.; Kawabata, H.; Yasohara, Y.; Kizaki,
N.; Hasegawa, J .; Shimizu, S. Biosci. Biotechnol. Biochem. 1998, 62,
280-285.
Present address: Institute for Organic Chemistry, Vienna Uni-
versity of Technology, Getreidemarkt 9, A-1060 Vienna, Austria.
| Phone: (352) 846-0743. Fax: (352) 846-2095. E-mail: jds2@
chem.ufl.edu.
(1) Neuberg, C. Adv. Carbohydr. Chem. 1949, 4, 75-117.
(2) Servi, S. Synthesis 1990, 1-25.
(3) Csuk, R.; Glanzer, B. I. Chem. Rev. 1991, 91, 49.
(4) D’Arrigo, P.; Pedrocchi-Fantoni, G.; Servi, S. Adv. Appl. Micro-
biol. 1997, 44, 81-123.
(5) Stewart, J . D. Curr. Opin. Drug Discov. Dev. 1998, 1, 278-289.
(6) Francke, A. Biochem. J . 1965, 95, 633-640.
(7) King, H. L., J r.; Dyar, R. E.; Wilken, D. R. J . Biol. Chem. 1974,
249, 4689-4695.
(8) Furuichi, A.; Akita, H.; Matsukura, H.; Oishi, T.; Horikoshi, K.
Agric. Biol. Chem. 1985, 49, 2563-2570.
(18) Zhou, B.-N.; Gopalin, A. S.; VanMiddlesworth, F.; Shieh, W.-
R.; Sih, C. J . J . Am. Chem. Soc. 1983, 105, 5925-5926.
(19) D’Arrigo, P.; Pedrocchi Fantoni, G.; Servi, S.; Strini, A. Tetra-
hedron: Asymmetry 1997, 8, 2375-2379.
(20) Nakamura, K.; Inoue, K.; Ushio, K.; Oka, S.; Ohno, A. J . Org.
Chem. 1988, 53, 2589-2593.
(21) Naoshima, Y.; Maeda, J .; Munakata, Y.; Nishiyama, T.; Ka-
mezawa, M.; Tachibana, H. J . Chem. Soc., Chem. Commun. 1990, 964-
965.
(22) Nakamura, K.; Kondo, S.-i.; Kawai, Y.; Ohno, A. Tetrahedron
Lett. 1991, 7075-7078.
(23) Nakamura, K.; Kondo, S.-i.; Kawai, Y.; Ohno, A. Bull. Chem.
Soc. J pn. 1993, 66, 2738-2743.
(24) Medson, C.; Smallridge, A. J .; Trewhella, M. A. Tetrahedron:
Asymmetry 1997, 8, 1049-1054.
(9) Murata, K.; Fukuda, Y.; Simosaka, M.; Watanabe, K.; Saikusa,
T.; Kimura, A. Eur. J . Biochem. 1985, 151, 631-636.
(10) Shieh, W.-R.; Gopalin, A. S.; Sih, C. J . J . Am. Chem. Soc. 1985,
107, 2993-2994.
(25) Kawai, Y.; Kondo, S.-i.; Tsujimoto, M.; Nakamura, K.; Ohno,
A. Bull. Chem. Soc. J pn. 1994, 67, 2244-2247.
(26) Nakamura, K.; Inoue, K.; Oshio, K.; Oka, S.; Ohno, A. Chem.
Lett. 1987, 679-682.
(11) Heidlas, J .; Engel, K.-H.; Tressl, R. Eur. J . Biochem. 1988, 172,
633-639.
(27) Hayakawa, R.; Shimizu, M.; Fujisawa, T. Tetrahedron: Asym-
metry 1997, 8, 3201-3204.
(12) Nakamura, K.; Kawai, Y.; Nakajima, N.; Ohno, A. J . Org. Chem.
1991, 56, 4778-4783.
(28) Hayakawa, R.; Nozawa, K.; Shimizu, M.; Fujisawa, T. Tetra-
hedron Lett. 1998, 67-70.
(13) Ishihara, K.; Nakajima, N.; Tsuboi, S.; Utaka, M. Bull. Chem.
Soc. J pn. 1994, 67, 3314-3319.
(29) Guo, Z.-W.; Sih, C. J . J . Am. Chem. Soc. 1989, 111, 6836-6841.
(30) Nakamura, K.; Kawai, Y.; Oka, S.; Ohno, A. Bull. Chem. Soc.
J pn. 1989, 62, 875-879.
(14) Nakajima, N.; Ishihara, K.; Kondo, S.; Tsuboi, S.; Utaka, M.;
Nakamura, K. Biosci. Biotechnol. Biochem. 1994, 58, 2080-2081.
10.1021/jo9900681 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/10/1999

6604 J . Org. Chem., Vol. 64, No. 18, 1999
Kayser et al.
Sch em e 1
determination of the complete genome sequence of S.
cerevisiae,31 altering the enantioselectivities of whole-cell-
mediated reductions by genetic manipulation has become
a feasible goal. For example, we have already shown that
the catalytic repertoire of Baker’s yeast can be expanded
to include asymmetric Baeyer-Villiger oxidations by
creating a recombinant strain that expresses bacterial
cyclohexanone monooxygenase.32 The use of similar
techniques to improve the enantioselectivity of Baker’s
yeast reductions requires that the yeast enzymes cata-
lyzing the reactions be identified.
In this paper, we focus on Baker’s yeast-mediated
reductions of R-keto esters. While this class of compounds
has received relatively less attention than â-keto esters
and acids,33 the product R-hydroxy acids and esters are
equally valuable chiral building blocks, and several
routes to these compounds have been devised.34 Reduc-
tions of R-keto esters catalyzed by Baker’s yeast are
experimentally simple and avoid the yield limitations
associated with kinetic resolutions. The major drawback
is less-than-optimal stereoselectivity, the result of mul-
tiple reductase enzymes produced simultaneously by the
yeast. Three such enzymes have been isolated by Naka-
mura and co-workers,15 although other enzymes may be
important for specific substrates. Our goals in this study
were to investigate further the yeast-mediated reductions
of R-keto carbonyl compounds and apply this knowledge
to synthesize the paclitaxel side chain by reduction of
R-keto carbonyl compounds.35
Ta ble 1. Yea st Red u ction s of r-Keto Ester s Usin g
Com m er cia l Ba k er ’s Yea st (Yea st:su bstr a te ) 20:1 (by
w eigh t); 30 °C, 24 h )
recovered
starting
material product material yield (%) (%)
starting product eea
byproduct(s)
1a
1b
1c
1d
3a
3b
3bb
3bc
3c
2a
2b
2c
2d
4a
4b
4b
4b
4c
<2
<2
41
36
38
24
17
91
68
72
<2
47
92 none
90 none
48 n-C5H11OH, 7%
18 n-C8H17OH, 11%
98 none
44
<2
<2
<2
90
82 none
92 none
none
32
52 n-C5H11OH, 12%;
PhCH2OH, 8%
20 n-C12H25OH, 6%;
PhCH2OH, 5%
>98 none
3e
5
4e
6
30
46
64
<2
a
b
All products possessed the (R)-configuration. Methyl vinyl
Resu lts a n d Discu ssion
ketone (126 mM) was included in the reaction mixture. c Methyl
vinyl ketone (126 mM) and ethyl chloroacetate (66 mM) were
included in the reaction mixture.
Four pyruvate esters and four benzoylformate esters
bearing various alkoxy chains were synthesized and
reduced by Baker’s yeast (Scheme 1), and the results are
shown in Table 1. The (R)-alcohols were isolated in all
cases, although the enantiomeric purities depended
significantly on the length of the alkoxy chain. These
results stand in contrast to earlier reports that the (S)-
alcohols predominated.20-23 This discrepancy likely re-
sults from the very different reaction conditions used in
the present study and the existence of an enantioselective
yeast-mediated hydrolysis of (R)-lactate alkyl esters.23,36
For both 1 and 3, the best results were obtained from
the reductions of methyl esters, suggesting that these are
poor substrates for the yeast esterase(s). Consistent with
this notion, a high yield and optical purity were obtained
for ester 4b, which has been reported to be inert with
respect to Baker’s yeast-mediated hydrolysis.36
We explored two approaches to increasing the enan-
tioselectivities of R-keto ester reductions even further.
In the case of 3b, known inhibitors of yeast reductases
were included in the reaction mixture (Table 1).37 Methyl
vinyl ketone increased slightly both the yield and optical
purity of alcohol 4b, although the simultaneous presence
of methyl vinyl ketone and ethyl chloroacetate was toxic
to the yeast cells, and only starting material was recov-
ered. Two other yeast strains were also tested for the
reduction of 3a (Table 2). Strain ATCC 26403 is unable
to produce fatty acid synthase.38 Substitution of this
strain for commercial Baker’s yeast had little effect on
the enantioselectivity of the reduction, suggesting that
this enzyme is not responsible for the production of 4a .
Another laboratory yeast strain (INVSc1) was also used
for the reduction of 3a . While this strain possessed
markedly different enantioselectivities for reductions of
(31) Mewes, H. W.; Albermann, K.; Bahr, M.; Frishman, D.; Gleiss-
ner, A.; Hani, J .; Heumann, K.; Kleine, K.; Maierl, A.; Oliver, S. G.;
Pfeiffer, F.; Zollner, A. Nature 1997, 387, 7-8.
(32) Stewart, J . D.; Reed, K. W.; Martinez, C. A.; Zhu, J .; Chen, G.;
Kayser, M. M. J . Am. Chem. Soc. 1998, 120, 3541-3548.
(33) Santaniello, E.; Ferraboschi, P.; Grisenti, P.; Manzocchi, A.
Chem. Rev. 1992, 92, 1071-1140.
(34) Chemical syntheses: (a) Evans, D. A.; Morrissey, M. M.; Dorow,
R. L. J . Am. Chem. Soc. 1985, 107, 4346-4348. (b) Corey, E. J .; Link,
J . O.; Shao, Y. Tetrahedron Lett. 1992, 3435-3438. Biocatalytic
routes: (c) Reference 36. (d) Wong, C.-H.; Matos, J . R. J . Org. Chem.
1985, 50, 1992-1994. (e) Effenberger, F. Ang. Chem., Int. Ed. Engl.
1994, 33, 1555-1564. (f) Adam, W.; Fell, R. T.; Hoch, U.; Saha-Mo¨ller,
C. R.; Schreier, P. Tetrahedron: Asymmetry 1995, 6, 1047-1050. (g)
Adam, W.; Lazarus, M.; Boxx, B.; Saha-Mo¨ller, C. R.; Humpf, H.-U.;
Schreier, P. J . Org. Chem. 1997, 62, 7841-7843. (h) Adam, W.; Boland,
W.; Hartmann-Schreier, J .; Humpf, H.-U.; Lazarus, M.; Saffert, A.;
Saha-Mo¨ller, C. R.; Schreier, P. J . Am. Chem. Soc. 1998, 120, 11044-
11048. (i) Sugai, T.; Ohta, H. Tetrahedron Lett. 1991, 7063-7064.
(35) Other synthetic routes: (a) Reference 40. (b) Reference 41. (c)
Ojima, I.; Habus, I.; Zhao, M.; Georg, G. I.; J ayasinghe, L. R. J . Org.
Chem. 1991, 56, 1681-1683. (d) Ojima, I.; Habus, I.; Zhao, M.; Zucco,
M.; Park, Y. H.; Sun, C. M.; Brigaud, T. Tetrahedron 1992, 48, 6985-
7012. (e) Denis, J .-N.; Greene, A. E.; Serra, A. A.; Luche, M.-J . J . Org.
Chem. 1986, 51, 46. (f) Denis, J .-N.; Correa, A.; Greene, A. E. J . Org.
Chem. 1990, 55, 1957. (g) Denis, J . N.; Correa, A.; Greene, A. E. J .
Org. Chem. 1991, 56, 6939-6942. (h) Commerc¸on, A.; Be´zard, D.;
Bernard, F.; Bourzat, J . D. Tetrahedron Lett. 1992, 5185-5188. (i)
Georg, G. I.; Cheruvallath, Z. S.; Himes, R. H.; Mejillano, M. R.; Burke,
C. T. J . Med. Chem. 1992, 35, 4230-4237. (j) Gou, D. M.; Liu, Y. C.;
Chen, C. S. J . Org. Chem. 1993, 58, 1287-1289. (k) Hattori, K.; Miyata,
M.; Yamamoto, H. J . Am. Chem. Soc. 1993, 115, 1151-1152.
(36) Ushio, K.; Yamauchi, S.; Masuda, K. Biotechnol. Lett. 1991, 13,
495-500.
(37) Nakamura, K. In Stereochemical Control in Microbial Reduc-
tion; Nakamura, K., Ed.; Kluwer Academic Publishers: Boston, 1992;
Vol. 381, pp 389-398.
(38) Schweizer, E.; Bolling, H. Proc. Natl. Acad. Sci. U.S.A. 1970,
67, 660-666.

Yeast-Mediated Reductions of R-Keto Esters
J . Org. Chem., Vol. 64, No. 18, 1999 6605
Ta ble 2. Yea st Red u ction s of Meth yl Ben zoyl F or m a te
reaction mixtures were determined as a function of time
(Table 3).44 When commercial Baker’s yeast was used for
the reduction of 13, the desired diastereomer 14 was the
major product, consistent with the results described
above for acyclic R-keto esters; however, significant
quantities of its enantiomer (ent-14) and trans-alcohol
15 were also formed (Figure 1). All of these are known
compounds.45 The level of ent-15, if formed, was below
our detection limits. The structure of 15, including its
absolute configuration, was further confirmed by X-ray
crystallography after its chromatographic separation
from the product mixture and derivitization as a p-
bromobenzoate ester.46
3a Ca ta lyzed by Differ en t S. cer evisia e Str a in s
reaction
time (h)
product
yield (%)
yeast strain
eea (%)
commercial Baker’s yeast
26403
INVSc1
24
36
24
91
>98
>98
96
46b
91
a
b
The (R)-alcohol was isolated in all cases. Significant quanti-
ties of yeast metabolites were also isolated from this reaction
mixture.
Sch em e 2
The reduction of racemic R-keto-â-lactam 13 allows for
a kinetic resolution, and in the case of commercial
Baker’s yeast, both substrate enantiomers were reduced
to alcohol products. While the enantiomers of 13 were
not resolved by chiral-phase HPLC, it was possible to
calculate its optical purity from the observed product
composition (Table 3). When these values were plotted
as a function of fractional conversion, the data clearly
did not fit the expected pattern for a kinetic resolution
catalyzed by a single enzyme,47 suggesting that more
than one enzyme was responsible for the three observed
products (14, ent-14, and 15). Substitution of the labora-
tory strain INVSc1 in place of commercial Baker’s yeast
improved the optical purity of 14; however, significant
quantities of the undesired trans-alcohol were also
formed by this strain (Table 3). Again, quantitative
analysis of the enantiomeric purity of the starting R-keto-
â-lactam during the course of the reaction was consistent
with more than one enzyme catalyzing the reduction of
13. By contrast, the use of fatty acid synthase-deficient
S. cerevisiae ATCC 26403 almost completely suppressed
formation of the undesired trans isomer 15 (Table 3).
Moreover, a quantitative analysis of the optical purity
of the starting material as a function of fractional
conversion was consistent with reduction by only a single
enzyme (enantioselectivity value 4; Figure 2). This sug-
gests that yeast fatty acid synthase is primarily respon-
sible for the production of trans isomer 15.
â-keto esters,39 these cells displayed nearly identical
behavior toward the reduction of 3a .
Since methyl R-keto esters were reduced by whole
Baker’s yeast cells to (R)-alcohols in high optical purities,
we applied this methodology to a short synthesis of the
paclitaxel C13 side chain [(2R,3S)-N-benzoyl-3-phenyl-
isoserine] (Scheme 2). This work builds on our earlier
route to this compound that featured a yeast-mediated
reduction of a chiral R-keto ester to generate the correct
stereochemistry at C2.40 In the present synthesis, we
attempted to use a Baker’s yeast-mediated reduction to
kinetically resolve racemic â-azido-R-keto ester 7, which
was converted to a separable diastereomeric mixture of
known (R)-alcohols 8a and 8b. Spectral data obtained for
materials prepared by this route were completely con-
sistent with literature data.41 The desired product 8a was
converted to the paclitaxel C13 side chain by the method
of Greene and co-workers.35f While this approach yielded
the desired target compound, the limited diastereoselec-
tivity observed for the enzymatic reduction of 7 was a
significant drawback.
We also explored a second route to the paclitaxel side
chain (Scheme 3). The activated â-lactam derivative of
(2R,3S)-N-benzoyl-3-phenylisoserine allows it to be di-
rectly coupled to readily available 10-deacetylbaccatin
III,42 and we anticipated that the rigid skeleton might
also improve the enantioselectivity of carbonyl reduction.
The key step in our approach is the Baker’s yeast-
mediated reduction of racemic R-keto-â-lactam 13, which
was assembled in two steps from readily available
starting materials in 77% overall yield by a modification
of the procedure of Cainelli et al.43 The ketone was
reduced by Baker’s yeast, and the compositions of the
If a single enzyme is responsible for the formation of
both 14 and ent-14, it must be capable of producing both
a
D- and an L-alcohol. This initially surprising lack of
enantioselectivity may be a consequence of the nearly
symmetrical structure of â-lactam 13. The calculated
structures of 14 and ent-14 can be overlaid on one
another so that the aromatic groups are nearly coincident
(Figure 3). The hydrides added during the enzymatic
reduction of 13 are within 0.5 Å and aligned nearly
parallel to one another, suggesting that a single enzyme
active site could accommodate production of both 14 and
ent-14, as suggested by our data. The identity of the
enzyme(s) responsible for the production of 14 and ent-
14 remains to be established, however.
(44) Initial experiments indicated that the low solubility of 13 in
the aqueous reaction medium posed difficulties in accurate sampling
for analysis by chiral-phase HPLC. On the other hand, complete
extractions of reaction mixtures provided >90% mass recovery. Parallel
reactions were therefore started simultaneously, and then each reac-
tion mixture was extracted completely after a given period of time.
All samples were analyzed by chiral-phase HPLC, which cleanly
resolved 13 and the four possible reduced products.
(45) Srirajan, V.; Deshmukh, A. R. A. S.; Bhawal, B. M. Tetrahedron
1996, 52, 5585-5590.
(46) Details of this crystal structure will be published elsewhere.
(47) Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J . J . Am. Chem.
Soc. 1982, 104, 7294-7299.
(39) Kayser, M. M.; Mihovilovic, M. D.; Rodriguez, S.; Stewart, J .
D. Unpublished results.
(40) Kearns, J .; Kayser, M. M. Tetrahedron Lett. 1994, 2845-2848.
(41) Thijs, L.; Porskamp, J . J . M.; van Loon, A. A. W. M.; Derks, M.
P. W.; Feenstra, R. W.; Legters, J .; Zwanenburg, B. Tetrahedron 1990,
46, 2611-2622.
(42) Holton, R. A.; Liu, J . H. Bioorg. Med. Chem. Lett. 1993, 3, 2475-
2478.
(43) Cainelli, G.; Panunzio, M.; Giacomini, D.; DiSimone, B.; Cam-
erini, R. Synthesis 1994, 805-808.

6606 J . Org. Chem., Vol. 64, No. 18, 1999
Kayser et al.
Sch em e 3
Ta ble 3. Ba k er ’s Yea st-Med ia ted Red u ction s of Ra cem ic r-Keto-â-la cta m 13
starting 14 ent-14 15
ent-15
reaction material conversionb ee (13)c (3R,4S) (3S,4R) ee (14)d (3R,4R) (3S,4S) ee (15)c
time (h)
13a (%)
(%)
(%)
(%)
(%)
(%)
(%)
(%)
(%)
commercial yeast
3
9
20
72
3
61
43
13
<1
77
71
47
21
5
81
61
53
47
<1
39
57
87
100
23
29
53
79
95
17
18
26
38
25
34
46
48
17
22
41
47
49
13
29
36
39
55
5
9
16
14
2
2
4
15
14
4
67
58
48
55
79
83
82
52
56
53
57
60
59
22
9
14
25
38
4
5
8
17
32
<1
2
<1
>89
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
>93
>96
>97
>75
>80
>88
>94
>97
INVSc1 (laboratory strain)
14
21
62
1
60
10
31
46
53
6
9
20
72
21
46
68
119
96f
26403 (FAS deficient)
39
48
53
100
8
9
10
35
>50
>67
>75
>90
3
4
10
a
b
Reaction compositions were obtained by integrations of HPLC peaks. Calculated as (14 + ent-14 + 15 + ent-15)/(13 + 14 + ent-14
+ 15 + ent-15). c Calculated as ([14 + ent-15] - [15 + ent-14])/(100 - [14 + ent-14 + 15 + ent-15]). Calculated as (14 - ent-14)/(14 +
d
ent-14) × 100. e Calculated as (15 - ent-15)/(15 + ent-15) × 100. f These data were obtained from a separate reaction that was allowed
to proceed to completion. This reaction required less time than the previous set.
F igu r e 2. Enantioselectivity of the reduction of 13 by fatty
acid synthase-deficient Baker’s yeast. Values for the enantio-
meric excess and fractional conversion were calculated from
the data in Table 3. The data were fit by nonlinear least-
squares methods to an equation derived from that of Sih and
co-workers in which the enantioselectivity was the only
adjustable parameter.44 The close correspondence between the
calculated and observed data is consistent with the production
F igu r e 1. Reduction of R-keto-â-lactam 13 by Baker’s yeast
strains. Compositions of the reaction mixtures and values of
fractional conversion were determined by HPLC analysis.
Commercial Baker’s yeast: b, 14; O, 15. S. cerevisiae IN-
VSc1: 9, 14; 0, 15. S. cerevisiae ATCC 26403: 2, 14; 4, 15.
of both 14 and ent-14 by a single enzyme.
chain R-keto esters, enzymatic reductions can provide
high enantioselectivity, although the levels of diastereo-
selectivity can be relatively modest, e.g., in the reduction
of 7 to 8. For an R-keto-â-lactam such as 13, the
commercially available laboratory strain INVSc1 pro-
vided the desired alcohol diastereomer 14 with the
highest optical purity (up to ca. 80% ee at 50% conver-
sion), although significant quantities of the trans dias-
Con clu sion s
Taken together, our results show that yeast-mediated
reduction of R-keto carbonyl compounds provides a simple
route to optically active alcohols. In the case of the open-

Yeast-Mediated Reductions of R-Keto Esters
J . Org. Chem., Vol. 64, No. 18, 1999 6607
on YPD medium (1% Bacto-Yeast Extract, 2% Bacto-Peptone,
2% glucose). For culturing ATCC 26403, the YPD medium was
supplemented with K2HPO4 (5 g/L), KH2PO4 (5 g/L), TWEEN
40 (10 g/L), myristic acid (70 mg/L), palmitic acid (70 mg/L),
and stearic acid (70 mg/L).
Gen er a l P r oced u r e for th e P r ep a r a tion of Ben zoyl-
for m a te a n d P yr u va te Ester s. To a stirred solution of R-keto
acid (30 mmol) in dry alcohol (150 mmol) at -10 °C (for methyl
and ethyl alcohols; 0 °C for pentyl alcohol; 10 °C for octyl and
dodecyl alcohols) was added freshly distilled thionyl chloride
(33 mmol) dropwise over a period of 15 min. The solution was
then brought to room temperature, and stirring was continued
for an additional 30 min. Excess thionyl chloride was removed
by rotary evaporation, and the esters were purified by distil-
lation.
Gen er a l P r oced u r e for th e Red u ction of r-Keto Ester s
Usin g Com m er cia l Ba k er ’s Yea st. The ester (1.0 g) was
added with vigorous stirring to a suspension of dry yeast (20
g) in distilled water (800 mL) in a round-bottom flask whose
volume was at least twice the total volume of the reaction
mixture. Sucrose (20 g) was then added over a period of 1 h
(to minimize foaming). The reaction was allowed to proceed
at room temperature for 24 h. After this time, the reaction
mixture was centrifuged at 4000g for 15 min. Alternatively,
Celite was added to the reaction mixture suspension, and then
this was vacuum-filtered through no.1 filter paper. The
isolated supernatant (or filtrate) was saturated with NaCl,
acidified with 10% HCl, and extracted continuously with
dichloromethane for 16 h. The organic layer was dried over
MgSO4 and the solvent removed in vacuo to give a yellow
liquid. The composition of the crude product was established
by 1H NMR, followed by the addition of diazomethane to
determine whether any acids were present. Chromatography
on a short silica gel column using 10% ethyl acetate-hexanes
afforded pure reduction products. The enantiomeric excess
values were determined by optical rotation (sodium D-line, 25
°C, 2 dm path length, c ) 2 g/100 mL, CHCl3). For those
compounds where a literature value for the specific rotation
was unavailable, the ester was hydrolyzed by being stirred for
10 min in an aqueous solution containing 10% potassium
hydroxide and 5% zinc chloride. The solution was made slightly
acidic by the addition of 10% HCl and extracted with dichlo-
romethane. The organic layer was dried over MgSO4, filtered,
and evaporated. The residual acid was converted to the ester
with diazomethane, and the specific rotation was determined
for the methyl ester.
Yea st-Med ia ted Red u ction s in th e P r esen ce of In h ibi-
tor s. The general procedure for yeast reductions described was
followed. Ethyl chloroacetate was added to the reaction
mixture at a final concentration of 66 mM, and methyl vinyl
ketone was used to a final concentration of 126 mM. The
reactions were worked up as usual, and the inhibitors were
removed during the final chromatographic separation.
Gen er a l P r oced u r e for Red u ction s Usin g La bor a tor y
Str a in s of S. cer evisia e. YPD medium (40 mL) was inocu-
lated with the appropriate S. cerevisiae strain, and the culture
was incubated at 30 °C in an orbital shaker for 24 h. A 100
mg portion of the ketone was added, and the reduction was
carried out at 30 °C. After complete consumption of the
substrate (as determined by HPLC), the cells were separated
by centrifugation at 3000g for 5 min, and then they were
washed once with water and once with EtOAc. The aqueous
layers were extracted with EtOAc, and the combined organic
layers were dried and evaporated. Products were isolated by
flash chromatography.
F igu r e 3. Overlay of calculated structures for 14 and ent-14.
Both structures were sketched and energy-minimized using
the Tripos force field (Sybyl 6.4). Dummy atoms were created
in the centers of both aromatic rings, and these, along with
the C3 proton, were used for rigid-body superpositioning.
Atoms of 14 are shown in light gray, while those of ent-14 are
depicted in dark gray. The hydrides delivered by the reductase-
(s) are shown in black and indicated by an arrow. Note the
pseudosymmetry inherent in these structures.
tereomer 15 were also formed. Interestingly, while
inactivation of the fatty acid synthase by genetic muta-
tion largely suppressed the formation of 15, it also
diminished the optical purity of the desired alcohol 14.
From a quantitative analysis of the data, it appears that
both 14 and ent-14 are products of a single reductase
enzyme with relatively low enantioselectivity that favors
(S)-13. In the presence of fatty acid synthase, however,
the apparent enantioselectivity of this enzyme is in-
creased because fatty acid synthase competes effectively
for the undesired (R)-13 and converts it to a different
diastereomer that can be removed by simple chromato-
graphic separation. This suggests that further genetic
manipulations of yeast reductases might lead to whole-
cell catalysts with greater enantio- and diastereoselec-
tivities for reductions of such compounds.
Exp er im en ta l Section
Chemicals used were purchased from Aldrich Chemical Co.
All solvents and ethyl pyruvate were distilled immediately
prior to use. Melting points are uncorrected. Proton and carbon
NMR spectra were recorded from CDCl3 solutions unless
otherwise indicated. Optical rotations were measured from
CHCl3 solutions. Chiral-phase GC analyses were performed
using a chiral Cyclosilb column (0.25 mm × 30 m, film
thickness 0.25 µm). Chiral-phase HPLC analyses were per-
formed using a nonchiral Econosphere silica 5U column (4.6
× 250 mm) connected in series with a Chiracel OD-H column
(4.6 × 150 mm) using a gradient of petroleum ether: i-propanol
from 95:5 to 80:20 at 1% per min as the mobile phase.
Detection was accomplished by UV monitoring at 254 nm.
Commercial Baker’s yeast was obtained from a local bakery
in a vacuum-packed form, and these large blocks were divided
into 20 g portions and stored under argon at 4 °C. S. cerevisiae
strains INVSc1 and ATCC 26403 were obtained from Invit-
rogen, Corp. and the American Type Culture Collection,
respectively. For routine culturing, these strains were grown
Meth yl 2-Oxo-3-a zid o-3-p h en ylp r op ion a te (7). Sodium
azide (3.801 g, 58.45 mmol) was added in a single portion to a
solution of methyl phenylglycidate (2.006 g, 11.26 mmol) in a
mixture of methanol (40 mL), water (5 mL), and dimethyl
formamide (8 mL). The mixture was warmed to 60 °C and
stirred vigorously for 46 h. The crude product was extracted
with ether, dried over MgSO4, and concentrated by rotary
evaporation, and then the mixture was chromatographed on
silica gel using EtOAc-hexanes (1:9) to afford 2.048 g of
methyl 2-hydroxy-3-azido-3-phenylpropionate as an oil (82%

6608 J . Org. Chem., Vol. 64, No. 18, 1999
Kayser et al.
yield) whose spectral data were identical to those reported
previously.35f A portion (1.23 g, 10.5 mmol) of this material
was treated with J ones’ reagent (3 mL) in a 60 °C water bath
and stirred for 12 h. The reaction mixture was filtered,
neutralized with 10% sodium carbonate solution, and extracted
with diethyl ether. The organic layer was washed with brine
and dried over MgSO4, and the solvent was evaporated. The
crude material was purified by silica gel chromatography using
EtOAc-hexanes (1:9) to yield 0.402 g of the title compound
as a yellow oil (27% yield). IR (CHCl3): 2130, 1750, 1730, 1720
the title compound as yellow crystals (85%, mp 124-126 °C).
1H NMR: δ 3.78 (s, 3H), 5.55 (s, 1H), 6.87 (d, J ) 8.8 Hz, 2H),
7.29-7.40 (m, 5H), 7.42 (d, J ) 8.8 Hz, 2H) ppm. 13C NMR: δ
55.5 (q), 74.9 (d), 114.8 (d), 119.8 (d), 126.4 and 129.4 (2d),
129.5, (d), 129.9 (s), 131.7 (s), 158.0 (s), 160.0 (s), 190.7 (s) ppm.
Red u ction of 13 by S. cer evisia e ATCC 26403. A frozen
culture of ATCC 26403 containing 0.1 g of yeast cells was
thawed and added to 100 mL of YPD medium containing fatty
acids in a baffled 250 mL flask. The mixture was shaken at
30 °C for 30 min at 250 rpm on an orbital shaker, and then
solid 13 (270 mg, 1.0 mmol) was added. The mixture was
shaken at 30 °C at 250 rpm and sampled periodically for HPLC
analysis.48 When all of the starting material had been con-
sumed, the mixture was transferred to a 500 mL centrifuge
bottle. Solid NaCl was added to saturation followed by 60 mL
of EtOAc. After shaking, the mixture was centrifuged at 3000g
at 0 °C for 10 min. The supernatant was decanted carefully
from the cell pellet into a separatory funnel, and the aqueous
layer was removed and subsequently extracted with EtOAc
(2 × 60 mL). The combined organic layers were washed with
brine (75 mL), dried with anhydrous Na2SO4, and concentrated
in vacuo. The crude product was purified by silica gel chro-
matography using a gradient elution from hexanes to EtOAc-
hexanes (1:4) to afford 120 mg of 14 as beige crystals (45%
yield, mp 204-208 °C; lit.45 mp 198-201 °C). Spectral data
were consistent with those reported previously.45
cm-1 1H NMR (CDCl3): δ 3.75 (s, 3H), 6.16 (s, 1H), 7.11-
.
7.46 (m, 5H) ppm. 13C NMR (CDCl3): δ 51.5, 124.3-133.8,
159.5, 169.6 ppm. MS (EI, 70 eV): m/ z 219 (M+, 2%), 192 (7%),
135 (8%).
Meth yl (2R,3S)-2-Hyd r oxy-3-a zid o-3-p h en ylp r op ion a te
(8a ). Sucrose (20 g) was added slowly to a stirred suspension
of dry yeast (20 g) in water (400 mL). Once active fermentation
had begun, 7 (0.505 g, 2.29 mmol) was added. After 48 h, Celite
(5 g) was added, and then the mixture was filtered through
no. 1 filter paper. The crude product was isolated using the
general procedure described above to give 0.334 g (67% yield)
of a 7:3 ratio of (2R,3S)-8a and (2R,3R)-8b (as determined by
1H NMR analysis). The two diastereomers were separated by
spinning-plate chromatography on a 2 mm silica plate using
EtOAc-pentane (15:85) as the solvent. Spectral data were
consistent with those reported previously.41
3,3-Dieth oxy-1-(4-m eth oxyp h en yl)-4-p h en yl-2-a zetid i-
n on e (12). A solution of diisopropylamine (4.21 g, 41.6 mmol)
in 42 mL of dry THF under nitrogen was treated with n-BuLi
(41.6 mmol, 1.6 M hexanes solution) at -20 °C. After the
solution was warmed to 0 °C, stirring was continued for 15
min to complete formation of LDA. A solution of ketal 10 (6.65
g, 37.8 mmol) in 35 mL of dry THF was added at -70 ( 5 °C,
and the resulting reaction mixture was stirred for 2.5 h at this
temperature. After addition of a solution of 11 (4.00 g, 18.9
mmol) in 20 mL of dry THF, the solution was stirred for an
additional 1.5 h at -70 ( 5 °C, and then it was allowed to
warm slowly to room temperature. After stirring for an
additional 18 h at room temperature, the reaction mixture was
quenched with brine and extracted with Et2O. The combined
organic layers were washed with 2 M HCl, saturated NaHCO3,
and brine, then dried over MgSO4, and concentrated in vacuo.
Removal of all volatiles by Kugelrohr distillation enabled
recovery of some excess 10. The crude product was purified
by recrystallization from diisopropyl ether to afford 5.77 g of
the title compound as colorless crystals (89%, mp 98-99 °C).
1H NMR: δ 0.85 (t, J ) 7.0 Hz, 3H,), 1.30 (t, J ) 7.0 Hz, 3H),
3.10 (dq, J 1 ) 7.0 Hz, J 2 ) 9.6 Hz, 1H), 3.55 (dq, J 1 ) 7.0 Hz,
J 2 ) 9.6 Hz, 1H), 3.65 (s, 3H), 3.80 (dq, J 1 ) 7.0 Hz, J 2 ) 9.6
Hz), 4.00 (dq, J 1 ) 7.0 Hz, J 2 ) 9.6 Hz, 1H), 5.05 (s, 1H), 6.70
(d, J ) 9.2 Hz, 2H), 7.20 (d, J ) 9.2 Hz, 2H), 7.32-7.39 (m,
5H) ppm. 13C NMR: δ 14.7 (q), 15.2 (q), 55.4 (q), 59.8 (t), 60.7
(t), 69.4 (d), 107.9 (s), 114.2 (d), 119.0 (d), 128.0 and 128.5 (2d),
128.6 (d), 130.4 (s), 133.7 (s), 156.3 (s), 163.1 (s) ppm.
1-(4-Meth oxyph en yl)-4-ph en yl-azetidin e-2,3-dion e (13).
A suspension of 12 (0.492 g, 1.44 mmol) in 5 mL of water was
cooled to 5 °C and slowly treated with 20 mL of concentrated
H2SO4. After complete addition of the acid, 2 vol % of acetone
was added, and the suspension was stirred at the above
temperature. Another 20 vol % of acetone was added after 2
h, and stirring was continued until TLC analysis indicated
complete hydrolysis (usually 1-1.5 h). The suspension was
poured into ice/water and extracted with EtOAc. The combined
organic extracts were neutralized with saturated NaHCO3,
dried over MgSO4, and concentrated in vacuo to give 0.33 g of
tr a n s-3-Hydr oxy-1-(4-m eth oxyp h en yl)-4-ph en yl-2-a zet-
in d in on e (15). The general procedure for Baker’s yeast-
mediated reductions of 13 was followed, and the crude products
from three separate reactions were combined to afford 304 mg
of a mixture containing 14, ent-14, and 15. Chromatography
on silica gel using petroleum ether-EtOAc (3:1) afforded 90
mg of 15 as colorless crystals (11% yield, mp 152-155 °C; lit.49
mp 155-156 °C). 1H NMR (DMSO-d6): δ 3.71 (s, 3H), 4.70 (d,
J ) 1.5 Hz, 1H), 4.87 (d, J ) 1.5 Hz, 1H), 6.69 (d, J ) 9.1 Hz,
2H), 7.13 (d, J ) 9.1 Hz, 2H), 7.29-7.36 (m, 5H) ppm. 13C NMR
(DMSO-d6): δ 55.3 (q), 65.7 (d), 76.7 (d), 114.2 (d), 119.0 (d),
126.1 (d), 129.8 (d), 128.7 (d), 130.1 (s), 136.0 (s), 156.4 (s),
166.8 (s) ppm.
Ack n ow led gm en t. Financial support by the Natu-
ral Sciences and Engineering Research Council and the
University of New Brunswick (M.M.K.), the National
Science Foundation (J .D.S.; Grant CHE-9816318), and
the FWF for a Schro¨dinger Fellowship (M.D.M.; Grant
J 1471-CHE) is gratefully acknowledged. We also thank
Cerestar, Inc. for supplying the cyclodextrins used in
this work. J .D.S. is a New Faculty Awardee of the
Camille and Henry Dreyfus Foundation (1994-1999).
We thank Gang Chen for his assistance throughout this
project and Dr. Fernande Rochon for performing the
X-ray crystallographic analysis associated with this
work.
Su p p or tin g In for m a tion Ava ila ble: NMR spectra for the
compounds used in this study. This material is available free
of charge via the Internet at http://pubs.acs.org.
J O9900681
(48) Analytical samples were prepared by vortexing 0.5 mL of the
reaction mixture with 0.5 mL of EtOAc. After the mixture was
centrifuged at 7500 rpm for 1 min, the organic layer was removed,
filtered, and analyzed by HPLC.
(49) Coss´ıo, F. P.; Palomo, C. Tetrahedron Lett. 1985, 4239-4242.