Highly stereoselective reagents for β-keto ester reductions by genetic engineering of Baker's yeast

Article abstract of DOI:10.1021/ja0027968

While whole cells of baker's yeast (Saccharomyces cerevisiae) are a convenient biocatalytic reducing agent for a wide variety of carbonyl compounds, mixtures of stereoisomeric alcohols are often observed since the organism contains a large number of reductase enzymes with overlapping substrate specificities but differing stereoselectivities. We sought to improve the performance of baker's yeast for β-keto ester reductions by using recombinant DNA techniques to alter the levels of three enzymes known to play important roles in these reactions (fatty acid synthase, Fasp: aldo-keto reductase, Ypr1p: α-acetoxy ketone reductase, Gre2p). A complete set of "first-generation" yeast strains that either lack or overexpress each of these three enzymes was created and tested for improvements in stereoselective reductions of a series of β-keto esters. On the basis of these results, multiply modified ("second-generation") strains were created that combined gene knockout and overexpression in single strains. In some cases, these additional modifications further improved the stereoselectivities of β-keto ester reductions, thereby making several β-hydroxy ester building blocks readily available by reactions that can be performed by nonspecialists. This work also revealed that additional yeast proteins participate in reducing β-keto esters, and further progress using this strategy will require either additional genetic manipulations or the expression of yeast reductases in hosts that lack enzymes with overlapping substrate specificity.

Full text of DOI:10.1021/ja0027968

VOLUME 123, NUMBER 8
FEBRUARY 28, 2001
©
Copyright 2001 by the
American Chemical Society
Highly Stereoselective Reagents for â-Keto Ester Reductions by
Genetic Engineering of Baker’s Yeast
Sonia Rodr ´ı guez,†,§ Margaret M. Kayser,* and Jon D. Stewart*,†
,‡
Contribution from the Department of Chemistry, UniVersity of Florida, GainesVille, Florida 32611, and
Department of Chemistry, UniVersity of New Brunswick, Saint John, New Brunswick E2L 4L5
ReceiVed July 28, 2000
Abstract: While whole cells of baker’s yeast (Saccharomyces cereVisiae) are a convenient biocatalytic reducing
agent for a wide variety of carbonyl compounds, mixtures of stereoisomeric alcohols are often observed since
the organism contains a large number of reductase enzymes with overlapping substrate specificities but differing
stereoselectivities. We sought to improve the performance of baker’s yeast for â-keto ester reductions by
using recombinant DNA techniques to alter the levels of three enzymes known to play important roles in these
reactions (fatty acid synthase, Fasp; aldo-keto reductase, Ypr1p; R-acetoxy ketone reductase, Gre2p). A complete
set of “first-generation” yeast strains that either lack or overexpress each of these three enzymes was created
and tested for improvements in stereoselective reductions of a series of â-keto esters. On the basis of these
results, multiply modified (“second-generation”) strains were created that combined gene knockout and
overexpression in single strains. In some cases, these additional modifications further improved the stereo-
selectivities of â-keto ester reductions, thereby making several â-hydroxy ester building blocks readily available
by reactions that can be performed by nonspecialists. This work also revealed that additional yeast proteins
participate in reducing â-keto esters, and further progress using this strategy will require either additional
genetic manipulations or the expression of yeast reductases in hosts that lack enzymes with overlapping substrate
specificity.
Introduction
the corresponding â-keto esters using chiral ruthenium catalysts.3
This method provides â-hydroxy esters in high yields and with
very high stereoselectivities; however, elevated temperatures and
pressures are sometimes required. Moreover, attempts to extend
this approach to R-substituted â-hydroxy esters have met with
The importance of homochiral â-hydroxy esters as building
blocks for a wide variety of targets has spurred the search for
covenient and highly selective methods for their synthesis.
Several “chemical” strategies that are particularly useful for
specific classes have been developed. For example, R-unsub-
stituted â-hydroxy esters can be prepared via hydrogenation of
1
,2
4
,5
mixed success. By contrast, aldol condensations can deliver
all four possible diastereomers of R-methyl-â-hydroxy esters
*
To whom correspondence should be addressed.
University of Florida.
University of New Brunswick.
†
(3) Burk, M. J. Acc. Chem. Res. 2000, 33, 363-372.
(4) Noyori, R.; Ikeda, T.; Ohkuma, T.; Widhalm, M.; Kitamura, M.;
Takaya, H.; Akutagawa, S.; Sayo, N.; Saito, T.; Taketomi, T.; Kumobayashi,
H. J. Am. Chem. Soc. 1989, 111, 9134-9135.
‡
§
Present Address: Catedra de Microbiologia, Facultad de Quimica, Gral
Flores 2124, 11800 Montevideo, Uruguay.
(
5) Burk, M. J.; Harper, G. P.; Kalberg, C. S. J. Am. Chem. Soc. 1995,
(
1) Mori, K. Tetrahedron 1989, 45, 3233-3298.
117, 4423-4424.
(2) de Souza Pereira, R. Crit. ReV. Biotechnol. 1998, 18, 25-83.
1
0.1021/ja0027968 CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/02/2001

1548 J. Am. Chem. Soc., Vol. 123, No. 8, 2001
Rodr ´ı guez et al.
with exquisite stereochemical control.6,7 Whether these methods
will be useful for larger R-substituents is not clear, however. A
third chemical routesalkylation of the dianion derived from a
chiral â-hydroxy alcoholsprovides the anti-diastereomers for
a wide variety of substrates.8
diminishes the optical purities of the products when whole cells
of the organism are used as the biocatalytic reagent.2
2,23
A
variety of techniques have been developed to overcome these
selectivity problems (vide infra).
The use of purified yeast enzymes for carbonyl reductions
avoids problems associated with competing catalysts with
differing stereoselectivities. A variety of reductase enzymes have
Given the success of “chemical” routes to chiral â-hydroxy
esters, what can biocatalysis offer? Kinetic resolutions of
acylated â-hydroxy esters by lipases and enzymatic reductions
of â-keto esters are the major biocatalytic routes to these
compounds. While lipase-mediated hydrolyses are simple to
perform, the overall conversions require multiple steps (â-keto
ester reduction to form the racemic alcohol, acylation, and then
enzymatic deacylation and separation of the desired product)
24
been purified from baker’s yeast. Unfortunately, yeast alcohol
dehydrogenase, the sole commercially available enzyme, accepts
2
5
only a limited range of ketones as substrates. Moreover,
reductase enzymes also require reduced nicotinamide cofactors,
which must be provided in stoichiometric amounts or via a
regeneration system. These experimental complications have
inspired the search for methods to improve the stereoselectivities
of carbonyl reductions using intact yeast cells since they provide
a source of both the enzyme and cofactor in a simple-to-use
package. The key to improving the stereoselectivities of whole
cell-mediated reductions is to arrange conditions so that only
few yeast enzymes reduce the added carbonyl substrate.
9
,10
and a 50% yield is the maximum that can be obtained.
Enantioselective reductions, on the other hand, allow all of the
starting material to be converted to the desired product. While
a variety of biocatalytic reducing agents are available, baker’s
yeast (Saccharomyces cereVisiae) has been particularly valu-
able.1
1-16
A subset of the approximately 6000 proteins produced
2
6,27
by this organism catalyzes the reductions of ketones and
aldehydes, often with very high efficiencies and stereoselec-
Techniques have included modifying the substrate structure,
15,26,28
carefully controlling the substrate concentration,
changing
the use of organic solvents or two
and the inclusion of enzyme inhibitors
1
7
29,30
tivities. Such reactions have been carried out successfully for
many years on preparatively useful scales. For example, Neuberg
and Nord showed in 1914 that a growing yeast culture reduced
n-pentanal to the corresponding alcohol in 68% yield on a >20
g scale.18 Neuberg also provided the first demonstration that
yeast reduction of a prochiral ketone could yield an optically
active alcohol.19 Since that time, chiral ketone reductions have
been the most common application of baker’s yeast in organic
synthesis. These reductions are simple to perform, the cells are
inexpensive and readily available, and the technique is applicable
to a very broad array of carbonyl compounds. In addition, yeast
reductions help minimize the environmental impact of organic
synthesis: reductions are generally carried out in aqueous
solutions, and the spent catalyst is entirely biodegradable. These
advantages have led to the use of yeast reductions on industrial
the growth conditions,
phase systems,
1
5,31-36
3
7-42
in the culture medium.
While these methodologies favor
one catalyst over competitors, they are based on empirical
findings, and it is difficult to predict a priori their effects.
We have followed a different path in improving the stereo-
selectivities of yeast-mediated carbonyl reductions, using yeast
genome sequence information to rationally design strains with
4
3
predictable stereoselectivities. This approach, outlined in
(
22) Shieh, W.-R.; Gopalin, A. S.; Sih, C. J. J. Am. Chem. Soc. 1985,
1
07, 2993-2994.
(23) Nakamura, K.; Kawai, Y.; Nakajima, N.; Ohno, A. J. Org. Chem.
991, 56, 4778-4783.
1
(24) Stewart, J. D.; Rodriguez, S.; Kayser, M. M. In Enzyme Technology
for Pharmaceutical and Biotechnological Applications; Kirst, H. A., Yeh,
W.-K.; Zmijewski, M. J., Eds.; Marcel Dekker: New York, in press.
(25) Sih, C. J.; Zhou, B.-N.; Gopalin, A. S.; Shieh, W.-R.; VanMiddles-
worth, F. Sel., Goal. Synth. Effic., Proc. Workshop Conf. Hoechst. 14th
2
0,21
scales.
When a single yeast enzyme dominates the reduction of a
particular carbonyl compound, the use of whole yeast cells can
provide the corresponding chiral alcohol in very high optical
purity. In other cases, however, multiple yeast reductase
enzymes with differing enantioselectivities are involved. This
1
983, 184, 251-261.
(26) Zhou, B.-N.; Gopalin, A. S.; VanMiddlesworth, F.; Shieh, W.-R.;
Sih, C. J. J. Am. Chem. Soc. 1983, 105, 5925-5926.
27) Nakamura, K.; Miyai, T.; Nozaki, K.; Ushio, K.; Oka, S.; Ohno,
A. Tetrahedron Lett. 1986, 3155-3156.
(
(
28) Wipf, B.; Kupfer, E.; Bertazzi, R.; Leuenberger, H. G. W. HelV.
Chim. Acta 1983, 66, 485-488.
(
6) Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981, 103,
127-2129.
7) Van Draanen, N. A.; Arseniyadis, S.; Crimmins, M. T.; Heathcock,
C. H. J. Org. Chem. 1991, 56, 2499-2506.
8) Fr a´ ter, G.; M u¨ ller, U.; G u¨ nther, W. Tetrahedron 1984, 40, 1269-
277.
(29) Ehrler, J.; Giovannini, F.; Lamatsch, B.; Seebach, D. Chimia 1986,
40, 172-173.
2
(
(30) Kometani, t.; Yoshii, H.; Kitatsuji, E.; Nishimura, H.; Matsuno, R.
J. Ferment. Bioeng. 1993, 76, 33-37.
(
(31) Nakamura, K.; Kondo, S.-i.; Kawai, Y.; Ohno, A. Tetrahedron Lett.
1991, 7075-7078.
1
5
5
(
9) Akita, H.; Matsukura, H.; Oishi, T. Tetrahedron Lett. 1986, 5241-
(32) Nakamura, K.; Kondo, S.-i.; Kawai, Y.; Ohno, A. Bull. Chem. Soc.
Jpn. 1993, 66, 2738-2743.
244.
(10) Itoh, T.; Kuroda, K.; Tomosada, M.; Takagi, Y. J. Org. Chem. 1991,
(33) Jayasinghe, L. Y.; Smallridge, A. J.; Trewhella, M. A. Tetrahedron
Lett. 1993, 3949-3950.
6, 797-804.
(
(
(
(
11) Neuberg, C. AdV. Carbohydr. Chem. 1949, 4, 75-117.
(34) North, M. Tetrahedron Lett. 1996, 1699-1702.
(35) Medson, C.; Smallridge, A. J.; Trewhella, M. A. Tetrahedron:
Asymmetry 1997, 8, 1049-1054.
12) Servi, S. Synthesis 1990, 1-25.
13) Csuk, R.; Glanzer, B. I. Chem. ReV. 1991, 91, 49.
14) Santaniello, E.; Ferraboschi, P.; Grisenti, P.; Manzocchi, A. Chem.
(36) Cui, J.-N.; Teraoka, R.; Ema, T.; Sakai, T.; Utaka, M. Tetrahedron
Lett. 1997, 3021-3024.
ReV. 1992, 92, 1071-1140.
15) D’Arrigo, P.; Pedrocchi Fantoni, G.; Servi, S.; Strini, A. Tetrahe-
dron: Asymmetry 1997, 8, 2375-2379.
16) Stewart, J. D. Curr. Opin. Drug DiscoVery DeV. 1998, 1, 278-
89.
17) Goffeau, A.; Barrell, B. G.; Bussey, H.; Davis, R. W.; Dujon, B.;
(
(37) Nakamura, K.; Inoue, K.; Oshio, K.; Oka, S.; Ohno, A. Chem. Lett.
1987, 679-682.
(
(38) Guo, Z.-W.; Sih, C. J. J. Am. Chem. Soc. 1989, 111, 6836-6841.
(39) Nakamura, K.; Kawai, Y.; Oka, S.; Ohno, A. Bull. Chem. Soc. Jpn.
1989, 62, 875-879.
2
(
Feldmann, H.; Galibert, F.; Hoheisel, J. D.; Jacq, C.; Johnston, M.; Louis,
E. J.; Mewes, H. W.; Murakami, Y.; Philippsen, P.; Tettelin, H.; Oliver, S.
G. Science 1996, 274, 546-567.
(40) Nakamura, K.; Kawai, Y.; Miyai, T.; Ohno, A. Tetrahedron Lett.
1990, 3631-3632.
(41) Nakamura, K.; Kawai, Y.; Ohno, A. Tetrahedron Lett. 1990, 267-
(18) Neuberg, C.; F., N. F. Biochem. Z. 1914, 62, 482-488.
(19) Neuberg, C.; Lewite, A. Biochem. Z. 1918, 91, 257-266.
(20) Keppler, J. G. J. Am. Oil Chem. Soc. 1977, 54, 474-477.
(21) Crocq, V.; Masson, C.; Winter, J.; Richard, C.; Lemaitre, Q.; Lenay,
270.
(42) Kawai, Y.; Kondo, S.-i.; Tsujimoto, M.; Nakamura, K.; Ohno, A.
Bull. Chem. Soc. Jpn. 1994, 67, 2244-2247.
(43) Rodriguez, S.; Kayser, M. M.; Stewart, J. D. Org. Lett. 1999, 1,
1153-1155.
J.; Vivat, M.; Buendia, J.; Prat, D. Org. Process. Res. DeV. 1997, 1, 2-13.

Genetic Engineering of Baker’s Yeast
J. Am. Chem. Soc., Vol. 123, No. 8, 2001 1549
Scheme 1
Scheme 2
initially to single enzymes (“first-generation” strains), and then
desirable traits can be combined (“second-generation” strains).
For this strategy to be feasible, the identities of all of the yeast
enzymes that participate in the reduction of a given carbonyl
compound must first be known. We therefore focused attention
on asymmetric reductions of â-keto esters because of their
importance as chiral building blocks and because a variety of
yeast enzymes catalyzing these reactions have been isolated and
4
4
characterized. At the outset of this project, it was believed
that three yeast enzymes played the major roles in reducing
22,23,45
â-keto esters (Scheme 2and Figure 1).
Fasp) accepts only R-unsubstituted â-keto esters and yields the
corresponding (R)-alcohols. While both aldo-keto reductase
Fatty acid synthase
(
(
44) Sybesma, W. F. H.; Straathof, A. J. J.; Jongejan, J. A.; Pronk, J.
T.; Heignen, J. J. Biocatal. Biotransform. 1998, 16, 95-134.
45) Shieh, W.-R.; Sih, C. J. Tetrahedron: Asymmetry 1993, 4, 1259-
1269.
Scheme 1, involves disabling undesirable yeast reductases by
gene knockout and overproducing those reductase enzymes with
desirable stereoselectivities. These techniques can be applied
(

1550 J. Am. Chem. Soc., Vol. 123, No. 8, 2001
Rodr ´ı guez et al.
Figure 1. Structures of plasmids used to overexpress yeast reductase enzymes. Fatty acid synthase (Fasp) requires two polypeptide chains, encoded
by the FAS1 and FAS2 genes. Both were placed downstream from GAL1 promoters on separate plasmids with different markers (URA3 and TRP1
in pSRG13 and pSRG29, respectively), allowing co-selection for both plasmids in minimal media lacking uracil and tryptophan. A plasmid designed
to overexpress Ypr1p was created by placing the YPR1 gene downstream from a GAL1 promoter (pSRG14). The plasmid for overexpressing Gre2p
contained the GRE2 gene downstream from the ADH1 promoter (pSRG41).
(Ypr1p) and R-acetoxy ketone reductase (Gre2p) accept R-sub-
Results
stituted â-keto esters and produce the corresponding (3S)-
A variety of strong promoters are available for overexpressing
proteins in baker’s yeast. Since the most convenient are
alcohols, they differ in their diastereoselectivities, the former
48
45,46
yielding syn- and the latter anti-products for most substrates.
controlled by the identity of the carbon source in the growth
medium, for example, glucose or galactose, we first investigated
the effect of these two sugars on the stereoselectivities of â-keto
Although later work showed this scheme to be oversimplified,
it provided a framework for attempting to improve the stereo-
selectivity of â-keto ester reduction by manipulating the levels
of these three proteins.
4
9
ester reductions in an unmodified yeast strain. We chose a
series of nine representative â-keto esters that were used
throughout this study so that meaningful comparisons between
yeast strains and growth conditions could be made (eq 1). In
In an earlier communcation, we reported the construction and
evaluation of baker’s yeast strains in which levels of Fasp and
Ypr1p were manipulated individually and showed that the
43
general approach outlined above could be useful. Our subse-
quent identification of the yeast gene encoding R-acetoxy ketone
reductase (GRE2)2
4,47
has allowed us to create a complete set
of “first-generation” strains involving the three proteins shown
in Scheme 2. Once these had been characterized, multiple gene
alterations were rationally combined to yield “second-genera-
tion” yeast strains that have even higher stereoselectivities in
whole cell-mediated â-keto ester reductions. In addition to
providing simple routes to important chiral building blocks, these
results have also demonstrated that additional yeast enzymes
reduce these substrates.
all cases, the reductions were allowed to reach >85% comple-
tion, and the compositions of the product mixtures were
determined by chiral-phase GC under conditions that gave
baseline resolution of all materials. Absolute configurations were
assigned by comparing GC traces to those from optically
enriched samples with known optical rotation values. These
(
46) Nakamura, K.; Kawai, Y.; Miyai, T.; Honda, S.; Nakajima, N.;
Ohno, A. Bull. Chem. Soc. Jpn. 1991, 64, 1467-1470.
47) Rodriguez Giordano, S. Ph.D. Dissertation, University of Florida,
000.
(
48) Schneider, J. C.; Guarante, L. Methods Enzymol. 1991, 194, 373-
89.
49) S. cereVisiae 15C: MATa, leu2, ura3-52, ∆trp1, his4-80, pep4-3.
(
3
2
(

Genetic Engineering of Baker’s Yeast
J. Am. Chem. Soc., Vol. 123, No. 8, 2001 1551
Table 1. Reductions of â-Keto Esters by Engineered Yeast Strains in Which Single Genes Are Knocked Out or Overexpressed
a
(“First-Generation” Yeast Strains)
fatty acid synthase
Fasp- c Fasp++ d
Ypr1p
Gre2p
carbon
source
unmodified
b
ketone
control
Ypr1p- e
Ypr1p++ f
Gre2p- g
Gre2p++ h
glucose
galactose
glucose
galactose
glucose
galactose
glucose
galactose
glucose
galactose
glucose
galactose
glucose
galactose
glucose
90% ee, S
84% ee, S
70% ee, R
86% ee, R
95% ee, S
84% ee, S
8% ee, R
>98% ee, S
i
90% ee, S
85% ee, S
76% ee, R
86% ee, R
93% ee, S
82% ee, S
14% ee, R
78% ee, R
>98% ee, R
>98% ee, R
74% de, syn
81% de, syn
30% de, anti
1% de, syn
83% de, anti
N.D.
i
38% ee, S
53% ee, S
50% ee, R
95% ee, R
72% ee, S
>98% ee, S
1
1
1
1
1
1
1
1
1
a
b
c
d
e
f
j
N.D.
N.D.
i
N.D.
i
N.D.
i
88% ee, R
i
93% ee, S
i
36% ee, R
i
>98% ee, S
i
36% ee, R
i
97% ee, R
i
96% de, syn
i
81% de, syn
i
65% de, syn
72% ee, S
N.D.
95% ee, S
N.D.
98% ee, S
N.D.
>98% ee, S
N.D.
l
l
l
l
l
l
l
l
53% ee, S
k
>98% ee, S
k
81% ee, S
k
34% ee, R
k
71% de, syn
k
38% de, anti
k
>98% de, anti
k
>98% de, anti
k
68% ee, S
25% ee, R
88% ee, R
67% ee, R
>98% ee, R
67% de, syn
98% de, syn
45% de, syn
61% de, syn
12% de, anti
28% de, syn
49% de, syn
95% de, syn
70% ee, R
>98% ee, R
>98% ee, R
>98% de, syn
83% de, syn
10% de, syn
59% de, syn
59% de, anti
17% de, syn
30% de, syn
28% de, syn
N.D.
l
l
l
l
l
l
l
l
g
h
i
galactose
glucose
galactose
42% de, anti
N.D.
i
>98% de, syn
a
Values reported are for enantiomeric or diastereomeric excess, as appropriate. Yeast strain 15C. c Yeast strain 2B. d Yeast strain 15C(pSRG13,
b
e
f
g
h
i
pSRG29). Yeast strain JS3. Yeast strain 15C(pSRG14). Yeast strain 10495B. Yeast strain 15C(pSRG41). Glucose is not compatible with
the promoter used to overexpress the protein. Not determined. Galactose is not compatible with the promoter used to overexpress the protein.
These ketones are not substrates for fatty acid synthase.
j
k
l
experiments revealed that the identity of the carbon sources
glucose or galactoseshad a strong effect on the stereoselectivity
of the reduction (Table 1). For R-unsubstituted â-keto esters
stereochemical assignments were made from chiral-phase GC
measurements of reactions that had proceeded to >85%
completion.
1
a-1e, galactose gave higher amounts of the (3R)-alcohol
Fatty acid synthase (Fasp) reduces R-unsubstituted â-keto
esters to (3R)-alcohols, and its elimination would therefore be
expected to increase formation of the (3S)-enantiomer. This was
observed in each case when ketones 1a-e were reduced by
whole cells of the Fasp knockout strain (Table 1). For example,
compared to glucose. The effect of carbon source on the
diastereoselectivities observed for ketones 1f-1i was less
consistent, although galactose appeared to favor formation of
the syn-(2R,3S)-alcohols at the expense of the anti-(2S,3S)-
diastereomer. Note that no alcohol products with the (3R)-con-
figuration were observed in the reductions of ketones 1f-1i.
1
e was reduced to (3R)-2e in >98% ee by the parent strain;
however, substituting the Fasp knockout strain completely
The results of these preliminary experiments dictated the
choice of promoter used for overexpressing the three yeast
enzymes (Fasp, Ypr1p, and Gre2p) so that the effect of the
carbon source would augment the stereoselectivity of the
overexpressed protein. The galactose-inducible GAL1 promoter
reversed the stereoselectivity so that (3S)-2e was obtained in
>98% ee. These results suggested that Fasp was the key (3R)-
selective yeast â-keto ester reductase, at least for ketones 1a-
e. Unfortunately, the strain overexpressing Fasp was somewhat
unstable and prone to loss of one or both plasmids unless grown
(PGAL1) was chosen for Fasp and Ypr1p (which yield the (3R)-
51
in minimal medium. For the one case examined with the Fasp
and syn-(2R,3S)-alcohols, respectively), while the ADH1 pro-
moter (PADH1) along with glucose as a carbon source was used
for overexpressing Gre2p (which produces anti-(2S,3S)-alcohols
from most R-substituted â-keto esters). While the reasons for
the observed changes in stereoselectivity as a function of carbon
source were not explored in detail, it is likely that gene
overexpression strain, the reduction of ketone 1d, use of these
cells increased the amount of (3R)-alcohol, as expected. Neither
the Fasp knockout or overexpression strains were examined for
reductions of R-substituted â-keto esters 1f-i since these are
not substrates for this enzyme.
The Ypr1p reduces R-unsubstituted â-keto esters to (3S)-
alcohols and R-substituted â-keto esters to syn-(2R,3S)-alcohols,
and its absence was predicted to diminish the amounts of these
products. Contrary to these expectations, for reductions of 1a-
e, the loss of Ypr1p had essentially no effect, regardless of
carbon source. For reductions of R-substituted â-keto esters 1f-
i, the absence of Ypr1p either modestly decreased (1g, 1i) or
had little influence (1f, 1h) on the amount of syn-alcohol
produced. The possibility that a functional YPR1 gene was still
present in the genome of the deletion strain was ruled out
expression levels of yeast reductase enzymes differ between cells
grown in glucose or galactose.50
“
First-Generation” Yeast Strains. Yeast knockout strains
lacking either Fasp, Ypr1p, or Gre2p were created by standard
methods. In addition, strains that individually overexpress one
of three reductases (Fasp, Ypr1p and Gre2p) were constructed.
The impacts of these genetic manipulations on the stereoselec-
tivities of â-keto ester reductions were assessed by the same
techniques that had been used previously to examine the
unmodified host cells. Each of the six engineered strains was
used to reduce the â-keto esters examined previously (eq 1),
and the results are collected in Table 1. As before, reactions
were performed with whole cells of the indicated strains, and
(51) While minimal medium could have been used throughout this study
to ensure plasmid maintenance, biotransformations were performed in rich
medium since this shortened reaction times significantly. Plasmid loss during
the bioconversions was not a serious problem with any strain other than
the Fasp overexpression strain. Episomal plasmids were used in this study
since the higher gene dosages (as compared to single integrated copies)
were expected to lead to higher protein expression levels.
(
50) Lahkari, D. A.; DeRisi, J. L.; McCusker, J. H.; Namath, A. F.;
Gentile, C.; Hwang, S. Y.; Brown, P. O.; Davis, R. W. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 13057-13062.

1552 J. Am. Chem. Soc., Vol. 123, No. 8, 2001
Rodr ´ı guez et al.
Table 2. Reductions of â-Keto Esters by Engineered Yeast Strains in Which Levels of Two or More Enzymes Are Manipulated by
Overexpression or Knockout (“Second-Generation” Yeast Strains)
-
Fasp++, Gre2p ,
Gre2p++,
Ypr1p++,
Gre2p++,
-
a
- b
- d
- e
ketone
Ypr1p
Fasp
ketone
Gre2p
Ypr1p
carbon source
yield
galactose
28%
72%, S
+31.4 (c 0.7)
galactose
80%
glucose
76%
>98%, S
+40.1 (c 2.5)
glucose
carbon source
yield
galactose
90%
98%, syn
+3.68 (c 2.5)
galactose
89%
83%, syn
+9.32 (c 2.5)
galactose
92%
65%, syn
+6.36 (c 2.5)
galactose
75%
glucose
86%
70%, syn
+8.04 (c 2.5)
glucose
c
1
1
1
1
1
a
b
c
1f
%
[
ee
D
% de
R]2
5
[R]
25
D
carbon source
yield
carbon source
yield
85%
73%
1g
1h
1i
%
[
ee
D
88%, R
>98%, S
+29.4 (c 2.5)
glucose
% de
67%, anti
+5.56 (c 2.5)
glucose
R]2
5
-12.5 (c 2.5)
galactose
66%
[R]
25
D
carbon source
yield
carbon source
yield
83%
67%
%
[
ee
D
64%, S
>98%, S
+32.3 (c 2.5)
glucose
% de
>98%, anti
+15.04 (c 2.5)
glucose
70%
>98%, anti
+5.16 (c 2.5)
R]2
5
+15.8 (c 2.5)
galactose
78%
[R]
25
D
carbon source
yield
carbon source
yield
87%
d
e
%
[
ee
D
91%, R
>98%, S
+26.2 (c 2.5)
glucose
90%
>98%, S
+16.5 (c 2.5)
% de
>98%, syn
+14.8 (c 2.5)
R]2
5
-21.8 (c 2.5)
[R]
25
D
carbon source
yield
galactose
c
26%
%
[
ee
D
>98%, R
-15.8 (c 2.5)
R]2
5
a
Yeast strain 24B(pSRG29). b Yeast strain 2B(pSRG41). c Isolated yield of purified alcohol. d Yeast strain 10495(pSRG14). e Yeast strain
JS3(pSRG41).
conclusively by Southern blotting with two different probes.52
Another possibility is that other enzymes with similar selectivity
compensate for the loss of Ypr1p, and we have shown recently
that at least two other yeast â-keto ester reductase enzymes have
sufficient to overcome a lack of stereoselectivity in whole cell-
mediated reductions. More importantly, these studies showed
that our original notion that only two yeast enzymes reduced
â-keto esters to (3S)-alcohols was oversimplified. For example,
if Ypr1p were the only yeast enzyme capable of producing syn-
(2R,3S)-alcohols, then the corresponding knockout strain should
have yielded only the anti diastereomers. This was not observed.
The situation with Gre2p is similar, with the knockout mutant
still yielding significant quantities of anti-alcohols, contrary to
our initial expectations.
53
similar stereoselectivities as Ypr1p. In contrast to these
disappointing results for the Ypr1p knockout strain, some
reductions by the strain overexpressing Ypr1p were quite
stereoselective, particularly for 1a and 1c where the (S)-alcohols
were produced in very high optical purities. Reductions of other
R-unsubstituted â-keto esters (1b, 1d, and 1e) still yielded the
(R)-alcohols as the major products, although overexpression of
Second-Generation Yeast Strains. Although our “first-
generation” engineered yeast strains provided good improve-
ments in stereoselectivities for certain substrates, obtaining
mixtures of isomers was still a common outcome. We sought
to overcome this problem by combining overexpression and
knockout strategies within single yeast strains. This was
particularly important for Ypr1p and Gre2p knockouts, since
our results from “first-generation” strains showed that their
deletions attenuated, but did not completely eliminate, alcohol
stereoisomers that we had originally ascribed solely to these
enzymes. We hoped to further improve stereoselectivity by
coupling the attenuations achieved by single enzyme deletions
with increased levels of enzymes with desirable specificities.
After considering the results described above, four “second-
generation” yeast strains were created. Twosthe “(3R)-selec-
tive“ and “(3S)-selective” strainsswere optimized for stereo-
selective reductions of R-unsubstituted â-keto esters. The other
two strainss“syn-selective” and “anti-selective”swere designed
for high selectivities in reducing R-substituted â-keto esters.
Ypr1p increased the level of (S)-alcohol. For R-substituted
â-keto esters 1f-i, the expected syn-(2R,3S)-alcohols were
always the major products, with the largest increases observed
for the reductions of 1f and 1i, which gave 2f and 2i in 96 and
>98% de, respectively.
The Gre2p reduces R-unsubstituted â-keto esters to (3S)-
alcohols and most R-substituted â-keto esters to anti-(2S,3S)-
alcohols. Deletion of this protein generally led to small or
modest reductions in the proportions of these products from
ketones 1a-i with two exceptions: ketones 1b and 1e actually
gave slightly higher amounts of (3S)-alcohols as compared to
reductions by the parent strain under the identical conditions.
These observations were both reproducible and contrary to
expectations. By contrast, overexpression of Gre2p had more
useful influences on the stereoselectivities of â-keto ester
reductions, increasing the proportions of the expected alcohol
products in each case. Ketones 1h and 1i provided two excellent
examples of these changes, with the anti-(2S,3S)-alcohols being
the only observed products.
Strains optimized for R-unsubstituted â-keto esters were used
to reduce ketones 1a-e, and the results are collected in Table
While the above studies revealed a few instances of highly
stereoselective â-keto ester reductions, they also revealed two
key shortcomings in the “first-generation” yeast strains. First,
manipulating levels of single yeast reductase enzymes was rarely
2
. The performance of the “(3R)-selective” strain was incon-
sistent, yielding the expected (3R)-alcohols from ketones 1b,
d, and 1e but producing the (3S)-alcohols from 1a and 1c. By
1
contrast, reductions by the “(3S)-selective” strain afforded only
the desired alcohol enantiomers from ketones 1a-e. While
yields of these reductions were variable, losses were mainly
due to evaporation of the products since no products other than
the expected alcohols were present in the crude reaction mixtures
according to GC. This was particularly true for the “(3R)-
(
52) One Southern blotting probe annealed to the YPR1 gene and the
other to the URA3 gene. The former gave no signal with the YPR1 deletion
strain but showed a band of the expected size when used to probe genomic
DNA from the parent strain. The latter probe showed the expected band
for the YPR1 deletion strain but no signal for the parent strain, as expected.
(53) Rodriguez, S.; Schroeder, K. T.; Kayser, M. M.; Stewart, J. D. J.
Org. Chem. 2000, 65, 2586-2587.

Genetic Engineering of Baker’s Yeast
J. Am. Chem. Soc., Vol. 123, No. 8, 2001 1553
Figure 2. Improvements in the stereoselectivities of R-unsubstituted
â-keto ester reductions by engineered yeast strains. Enantiomeric excess
values for alcohols derived from ketones 1a-e are grouped by substrate.
Results from selected yeast strains are arranged with the most highly
modified for (3R)-selectivity on the left and that most highly modified
Figure 3. Improvements in the stereoselectivities of R-substituted
â-keto ester reductions by engineered yeast strains. Diastereomeric
excess values for alcohols derived from ketones 1f-i are grouped by
substrate. Results from selected yeast strains are arranged with the most
highly modified for syn-(2R,3S)-selectivity on the left and that most
highly modified for anti-(2S,3S)-selectivity on the right: black,
-
-
for (3S)-selectivity on the right: black, Fasp++, Gre2p , Ypr1p ; dark
-
-
-
gray, Ypr1p ; striped, Gre2p ; light gray, Fasp ; white, Gre2p++,
-
Fasp .
-
-
-
Ypr1p++, Gre2p ; dark gray, Gre2p ; light gray, Ypr1p ; white,
-
Gre2p++, Ypr1p .
selective” strain since long reaction times were required to
completely consume the ketone substrates.
under the same conditions (Table 1). Slightly better results were
obtained by combining these two knockouts with overexpression
of Fasp in the cases of 1b, 1d, and 1e (Figure 2). Unfortunately,
the same strain reduced acetoacetates 1a and 1c mainly to the
The two yeast strains designed for diastereoselective reduc-
tions of R-substituted â-keto esters were tested for their abilities
to reduce ketones 1f-i. The “syn-selective” strain generally gave
high levels of the expected alcohol diasteromers, although
complete suppression of the anti-alcohols was only achieved
for ketones 1f and 1i. As anticipated, the “anti-selective” strain
gave mainly anti-alcohols from ketones 1g-i. The only excep-
tion was 2f, which gave mainly the syn-alcohol. This behavior
was anticipated since purified Gre2p has been shown to give
(3S)-alcohol. Clearly, this strain overexpressed Fasp poorly or
not at all; further progress will require a more efficient means
of overproducing this protein. Generally better results were
obtained for diastereoselective reductions of 1f-i. In these cases,
elimination of either Ypr1p or Gre2p had modest impacts;
however, combining these knockouts with overexpression of
the other reductase gave good to excellent diastereoselectivities
4
5,46
mainly the syn-2f,
and it was therefore not surprising that
this outcome was also observed when the engineered yeast cells
were used.
(Figure 3). Thus, by choosing the appropriate engineered yeast
strain, it was possible to prepare many of the alcohols derived
from 1a-i in high stereochemical purities and good chemical
yields.
Discussion
One goal of these studies was to show that it was possible to
genetically engineer a set of baker’s yeast strains with each
member optimized for yielding only a single stereoisomer from
a range of â-keto esters while still maintaining the experimental
simplicity inherent in using whole baker’s yeast cells. This goal
was partially reached. Eliminating enzymes that lead to undes-
ired stereoisomers was the first step toward enhancing the
selectivities of yeast strains, and in some cases, knockout of a
single gene was sufficient. For example, in the reductions of
R-unsubstituted â-keto esters 1a-e, genetically disabling Fasp
virtually eliminated production of (3R)-alcohols except for that
derived from 1b (Figure 2). This latter problem was solved by
overexpressing a (3S)-selective reductase (Gre2p) in the Fasp
knockout strain to give yeast cells that reduce ketones 1a-e to
the corresponding (3S)-alcohols in >98% ee in each case. Efforts
to create whole yeast cells with (3R)-selectivity were only
partially successful, however. Individually eliminating either
Gre2p or Ypr1p had generally little effect on the enantiomeric
purities of 2a-e when compared with unmodified cells grown
Our results also demonstrated conclusively that yeast enzymes
other than the three targeted in this study can also participate
in reducing â-keto esters. For example, knocking out Ypr1p
reduced, but did not completely eliminate, production of the
syn-(2R,3S)-alcohols derived from 1f-i. This would have been
expected if only the enzymes shown in Scheme 2 were involved.
Since studies with purified Gre2p have shown that this enzyme
is highly selective for production of anti-(2S,3S)-alcohols,4
there must be other yeast enzymes with substrate- and stereo-
selectivities similar to those of Ypr1p. We have recently
analyzed the complete S. cereVisiae genome and identified 49
proteins likely to catalyze carbonyl reduction, and five of these
proteins share high amino acid sequence similarity with Ypr1p.24
Two of these five Ypr1p relativessGre3p and Gcy1pshave
been expressed in Escherichia coli, and whole cells of these
engineered bacterial strains not only reduced â-keto esters but
also showed the same stereoselectivity as Ypr1p for 1f-i.53
Whether the remaining three Ypr1p relatives share similar
properties is currently under investigation. The results from the
5,46

1554 J. Am. Chem. Soc., Vol. 123, No. 8, 2001
Rodr ´ı guez et al.
Gre2p knockout strain also indicate clearly that baker’s yeast
produces additional enzymes with substrate- and stereoselec-
tivities similar to those of Gre2p. We have identified three
additional S. cereVisiae putative proteins that share high amino
acid sequence similarity with Gre2p,24 and it is tempting
to speculate that one or more of these may be responsible for
the incomplete stereoselectivity found for Gre2p knockout
strains. By contrast, fatty acid synthase appears to be virtually
the only yeast enzyme that reduces â-keto esters to (3R)-alcohols
(20 mg/L), L-leucine (30 mg/L), L-methionine (20 mg/L), L-tryptophan
-
(20 mg/L), and uracil (2 g/L). Yeast Fasp mutants required emulsified
fatty acids in the growth medium: 1% Tween 40 and 0.007% each of
myristic, palmitic, and stearic acids.56 Yeast expression vector pYES2
was purchased from Invitrogen, pAAH5 was a generous gift from
Benjamin Hall (University of Washington), YCplac111 and YCplac22
were provided by R. Daniel Gietz (University of Manitoba), and YEp13
was a gift from Henry Baker (University of Florida). S. cereVisiae strain
1
5C was generously provided by Andrew Buchman (Penn State
University), and INVSc1 and 10495B were purchased from Invitrogen
and Euroscarf, respectively.
(Figure 2).
Genetic engineering represents an alternative method to
â-Keto esters 1a-g were obtained from commercial suppliers and
used as received. Ketones 1h and 1i were synthesized according to
Dauben and Hart57 with slight modifications. Flash chromatography
improving the stereoselectivity of whole-cell-mediated yeast
reductions. Compared to prior strategies such as changes in
culture conditions or selective enzyme inactivation or inhibition,
increasing stereoselectivity by recombinant DNA techniques
requires a greater initial time investment for strain construction.
On the other hand, once the multiply modified yeast strains have
been prepared, their use is as simple as wild-type cells. More-
over, in most cases, they provide alcohols 2a-i in stereochem-
ical purities at least as high as those achieved by previous
biocatalytic methods without requiring special growth condi-
tions, changes in substrate structure, or toxic additives. Our
results also demonstrate the complexity inherent in using whole
S. cereVisiae cells for ketone reductions. At least eight enzymes
capable of â-keto ester reduction have been purified from
58
was carried out with 60 Å silica gel according to the procedure of Still.
NMR spectra were taken on Varian Gemini or VXR300 instruments
operating at 300 MHz, and all spectra were obtained in CDCl and
(77.0 ppm,
C). IR spectra were recorded on a Perkin-Elmer 1600 FT-IR from
3
1
3 3
referenced to residual CHCl (7.26 ppm, H) or to CDCl
13
thin films. Mass spectra (EI, 70 eV) were obtained from a Hewlett-
Packard 5890 series II GC interfaced with a 5971A mass-selective ion
detector. Optical rotations were measured from CHCl solutions using
3
a Perkin-Elmer 341 polarimeter operating at room temperature.
Analytical GC was carried out with a Hewlett-Packard 5890 instrument
equipped with flame ionization detectors and a 0.32 mm × 30 m DB-
17 column (J&W Scientific) for nonchiral separations and a 0.25 mm
×
25 m Chirasil-Dex CB column (Chrompack) for enantiomer
separations. Methyl benzoate was used as an internal standard. Samples
for GC analysis of biotransformation reactions were prepared by mixing
4
4
baker’s yeast, and few of these have been linked with the
corresponding gene.24 Exhaustive knockout of all competing
2
1
00 µL of the reaction mixture with 600 µL of EtOAc, vortexing for
0 s, and then centrifuging in a microcentrifuge for 2 min. Racemic
reductases is therefore difficult or impossible, even when it is
believed that all of the participants have been identified. For
example, Oliver and co-workers recently prepared a yeast strain
in which all seven putative aryl alcohol dehydrogenase genes
were simultaneously knocked out in a single cell; surprisingly,
this septuple mutant catalyzed the reduction of aryl aldehydes
as efficiently as the parent strain, showing that this reduction
was catalyzed by even more enzymes than had been suspected
previously.54 For this reason, we have recently pursued an
alternative strategy: expression of individual S. cereVisiae
4
alcohols were prepared from ketones 1a-i by reduction with NaBH ,
and conditions that gave baseline resolution of all products were used
for analyzing products from whole cell-mediated reductions. Alcohols
2a and 2g required acetylation prior to GC analysis for resolution of
the optical isomers, which was carried out with excess Ac
2
O in DMF
catalyzed by DMAP overnight at room temperature.47
Construction of Gene Knockout Strains. A haploid strain lacking
59
one of the two genes required for Fasp (2B) was derived by sporulation
and asci dissection of the heterozygous Research Genetics strain 21061
6
0
43
as described by Kaiser et al. In our earlier communication, we had
reductases in E. coli, which allows the whole bacterial cells to
-
be used as a stereoselective reducing agent.53 The combination
used a Fasp strain that contained an uncharacterized point mutation
61
in the FAS2 gene that largely inactivated Fasp; however, the complete
FAS2 gene deletion strain constructed here completely eliminated Fasp
and was found to give higher stereoselectivities. A yeast strain unable
to produce Ypr1p (JS3) was constructed by replacing the YPR1 gene
with a URA3 insertion using the microhomology PCR method. The
of these engineered E. coli and baker’s yeast strains can thereby
make a variety of S. cereVisiae ketone reductase enzymes
available for preparative purposes in a form suitable for
nonspecialists. The increasing availability of genome sequences
from several microorganisms also opens the possibility of
applying this metabolic engineering strategy to overcoming
problems of low stereoselectivity in other whole-cell biocatalysts
as well.
62
63
desired gene replacement was confirmed by PCR and Southern blot
43
hybridization as described earlier. A haploid mutant strain lacking
64
Gre2p (10495B) was obtained commercially from the Euroscarf
consortium.65
(
56) Schweizer, E.; Bolling, H. Proc. Natl. Acad. Sci. U.S.A. 1970, 67,
Experimental Section
6
2
60-666.
General Methods. Recombinant DNA procedures were carried out
(57) Dauben, W. G.; Hart, D. J. J. Org. Chem. 1977, 42, 3787-3793.
(58) Still, W. C.; Kahn, M.; Mitra., A. J. Org. Chem. 1978, 14, 2923-
925.
55
essentially as described by Sambrook et al. Restriction endonucleases
and T4 DNA ligase were purchased from New England Biolabs or
Promega. Oligonucleotides were obtained from Integrated DNA
Technologies or Gemini Biotechnology. All DNAs amplified by PCR
were sequenced to ensure that no spurious mutations were introduced.
Standard media and techniques for growth and maintenance of E. coli
and S. cereVisiae strains were used. YPD contained 1% Bacto-Yeast
Extract, 2% Bacto-Peptone, and 2% glucose and YP-Gal contained 1%
Bacto-Yeast Extract, 2% Bacto-Peptone, and 2% galactose. Minimal
(59) Diploid strain 21061 (MATa/MATR, his3∆1, leu2∆0, ura3∆0,
lys2∆0, met15∆0, fas2::kanMX4; heterozygous in the last three mutations)
was obtained from Research Genetics and sporulated, then haploids unable
to grow in the absence of exogenous fatty acids but which survived Geneticin
were selected. One with the desired phenotype was designated strain 2B.
(60) Kaiser, C.; Gottschling, D. E.; Stearns, T.; Adams, A. Methods in
Yeast Genetics: Laboratory Course Manual; Cold Spring Harbor Labora-
tory: Cold Spring Harbor, 1990.
(61) Burkl, G.; Castorph, H.; Schweizer, E. Mol. Gen. Genet. 1972, 119,
(
SD) medium contained 0.67% Bacto-Yeast Nitrogen Base without
315-322.
amino acids. When appropriate, this was supplemented with L-histidine
(62) S. cereVisiae JS3: MATa, leu2, ura3-52, ∆trp1, his4-80, pep4-3,
ypr1::URA3.
(
54) Delneri, D.; Gardner, D. C. J.; Bruschi, C. V.; Oliver, S. G. Yeast
999, 15, 1681-1689.
55) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning. A
Laboratory Manual, 2nd ed.; Cold Spring Harbor: Cold Spring Harbor,
989.
(63) Manivasakam, P.; Weber, S. C.; McElver, J.; Schiestl, R. H. Nucleic
Acids Res. 1995, 23, 2799-2800.
1
(
(64) S. cereVisiae 10495B: MATa, ura3-52, his3, leu∆1, lys2, trp1∆63,
gre2::kanMX4.
1
(65) Hajji, K.; Clotet, J.; Arino, J. Yeast 1999, 15, 435-441.

Genetic Engineering of Baker’s Yeast
J. Am. Chem. Soc., Vol. 123, No. 8, 2001 1555
Construction of Overexpression Strains for Fasp, Ypr1p, and
Gre2p. An overexpression system for both subunits of Fasp was created
in two compatible S. cereVisiae expression plasmids (pYES2 and a
ing the Gre2p knockout strain with the overexpression plasmid for
Ypr1p (pSRG14). Finally, the “anti-(2S,3S)-selective” strain (Gre2p++,
-
Ypr1p ; JS3(pSRG41)) utilized the Ypr1p knockout strain and the
66
derivative of YEplac112 ) in which both genes (FAS1 and FAS2) were
under control of GAL1 promoters. PCR amplifications were used to
introduce suitable restriction sites immediately preceding and following
the ends of the coding regions of FAS1 and FAS2. To avoid the need
to amplify and subsequently sequence the entire FAS1 and FAS2 genes,
PCR amplifications were carried out on shorter portions near the 5′
and 3′ ends. After the amplified portions were cloned and sequenced,
they were combined with the middle regions of the appropriate genes
to reconstruct the complete FAS1 and FAS2 genes flanked by SacI and
XhoI restriction sites. The complete FAS1 gene was cloned directly
between these sites in expression vector pYES2 to afford pSRG13.
While expression of the FAS2 gene utilized the same promoter and
terminator as pSRG13, the remaining portion of the plasmid was derived
from YEplac112 in the final expression plasmid pSRG29.53 The
overexpression plasmid for Ypr1p, pSRG14, has been described in our
existing overexpression plasmid for Gre2p (pSRG41). For all of these
“second-generation” strains, the choice of carbon source used during
whole-cell-mediated reductions was based on the promoter in the
overexpression plasmid.
General Procedure for â-Keto Ester Reductions with Yeast
Strains. S. cereVisiae strains were maintained on minimal plates
supplemented with the appropriate nutrients, for example, 2B and 2B-
56
(
pSRG41) required emulsified fatty acids. Fresh plates were streaked
from frozen stocks and a single colony was used to inoculate 50 mL
of YPD in a sterile 250 mL Erlenmeyer flask. The culture was incubated
at 30 °C in a rotary shaker at 200 rpm until the OD600 value was between
4 and 6 (early log phase). Cells were then harvested by centrifuging at
3000g for 10 min at 4 °C, the cell pellet was resuspended in 20 mL of
10 mM Tris-Cl, 1 mM EDTA (pH 7.5) by vortexing, and the cells
were collected by centrifugation as above. This washing procedure was
repeated two additional times. The final cell pellet was resuspended in
4
3
earlier communication. The gene encoding Gre2p (GRE2) was
amplified from S. cereVisiae genomic DNA using one primer that
annealed at the 5′ end of the coding region and another that bound
10 mM Tris-Cl, 1 mM EDTA (pH 7.5), 15% glycerol at a concentration
of 0.1 g/mL (wet weight). At this stage, cells were either used directly
for a reaction or frozen in aliquots at -80 °C for later use. Standard
reaction mixtures for preparative biotransformations contained 100 mL
of YPD or YP-Gal (depending on the promoter) along with 10 mM
â-keto ester substrate. Freshly prepared or frozen yeast cells were added
to an initial concentration of 2 mg/mL. Reaction flasks were shaken at
∼
70 bp downstream from the stop codon. Both primers contained
HindIII sites, and the amplified gene was cloned into this site in
5
3
pAAH5. Once constructed, plasmids were used to transform S.
cereVisiae 15C by the lithium acetate method.
Construction of “Second-Generation” Yeast Strains. The “(3R)-
selective” strain (Fasp++, Gre2p , Ypr1p ; 24B(pSRG29)) was
constructed after several steps. A yeast strain devoid of both Ypr1p
6
7
-
-
2
00 rpm and 30 °C, and the conversion was monitored by GC. After
the reactions were complete, the mixtures were centrifuged at 3000g
for 10 min at 4 °C, the supernatant was extracted with CH Cl
(3 × 50
mL), and the cell pellet was extracted with 20 mL of CH Cl . The
organic extracts were combined, washed with brine, and dried with
Na SO . When emulsions formed, they were broken by centrifugation
-
-
and Gre2p was created by mating JS3 (Ypr1p ) and 10495B (Gre2p )
in minimal solid medium lacking both uracil and histidine. Once cells
of the diploid strain developed, they were sporulated, and the resulting
spores were dissected and screened for clones able to grow in the
absence of uracil and in the presence of 200 µg/mL Geneticin. Three
colonies with these properties were further assayed for altered stereo-
selectivity in whole-cell â-keto ester reductions and one of the two
2
2
2
2
2
4
at 3000g for 10 min at 4 °C. After concentration by rotary evaporator,
the â-hydroxy esters were purified by flash chromatography on 1 ×
15 cm silica columns using 1:1 ether:hexanes as mobile phase. All of
with the expected phenotype was designated 24B68 and used for further
the â-hydroxyesters are known compounds, and the spectral data agreed
studies. Yeast strain 24B was then transformed with pSRG29 (which
with the literature reports.8
,69,70
In the case of ethyl 2-alkyl 3-hydroxy
directs the overexpression of the fatty acid synthase R-subunit) to
-
-
esters, the spectral data also provided information on the diasteromeric
purity.
complete construction of the Fasp++, Gre2p , Ypr1p strain. In the
-
-
absence of the overexpression plasmid for Fasp, the Gre2p , Ypr1
double knockout strain slowly reduced 1c and 1d, providing (S)-2c
and (R)-2d in 50 and 98% ee, respectively when galactose was used as
the carbon source. The behavior of this strain toward other â-keto ester
substrates was not examined.
The “(3S)-selective” strain (Gre2p++, Fasp ; 2B(pSRG41)) utilized
the previously constructed overexpression plasmid for Gre2p (pSRG41)
Acknowledgment. Financial support by the National Science
Foundation (CHE-9816318), the University of Florida and the
CONICYT-Uruguay is gratefully acknowledged. We also thank
Henry Baker and Mike Thompson for their helpful discussions
and Maurice Swanson and Ron Hector for assistance in asci-
dissection experiments.
-
in a host strain incapable of producing Fasp. The “syn-(2R,3S)-selective”
-
strain (Ypr1p++, Gre2p ; 10495(pSRG14)) was created by transform-
JA0027968
(
66) Gietz, R. D.; Sugino, A. Gene 1988, 74, 527-534.
(67) Gietz, R. D.; Woods, R. A. In Molecular Genetics of Yeast. A
Practical Approach; Johnston, J. R., Ed.; Oxford University Press: New
York, 1994; pp 121-134.
(69) Nakamura, K.; Miyai, K.; Nagar, A.; Oka, S.; Ohno, A. Bull. Chem.
Soc. Jpn. 1989, 62, 1179-1187.
(70) Heathcock, C. H.; Pirrung, M. C.; Sohn, J. E. J. Org. Chem. 1979,
44, 4294-4299.
(
68) S. cereVisiae 24B: MATa, his3, his4, leu2, lys2, trp1, gre2::kanMX4,
ypr1::URA3.