Regio and enantioselective reduction of t-butyl 6-chloro-3,5-dioxohexanoate with baker's yeast

Article abstract of DOI:10.1016/j.tetasy.2004.07.045

Whole baker's yeast cells reduce t-butyl 6-chloro-3,5-dioxohexanoate regioselectively to the corresponding C5 hydroxy keto ester. While the (R)-alcohol was favored, the enantioselectivity was poor (41% ee). A variety of process conditions were evaluated in order to improve both the enantioselectivity and yield of this reduction. Including a nonpolar resin in the reaction mixture afforded the (R)-alcohol in 94% ee and 50% isolated yield. The enantioselectivity was further improved to >99% ee by substituting purified YGL157w in place of whole yeast cells. This reductase was identified by screening a collection of yeast enzymes uncovered by genome sequence analysis.

Full text of DOI:10.1016/j.tetasy.2004.07.045

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 2825–2828
Regio and enantioselective reduction of t-butyl
-chloro-3,5-dioxohexanoate with bakerÕs yeast
6
a,
Michael Wolberg, Iwona A. Kaluzna, Michael M u¨ ller and Jon D. Stewart
b
a,
b,
*
*
a
Institut f u¨ r Biotechnologie 2, Forschungszentrum J u¨ lich GmbH, D-52425 J u¨ lich, Germany
127 Chemistry Research Building, University of Florida, Gainesville, Florida 32611, USA
b
Received 26May 2004; accepted 15 June 2004
Available online 11 September 2004
Abstract—Whole bakerÕs yeast cells reduce t-butyl 6-chloro-3,5-dioxohexanoate regioselectively to the corresponding C5 hydroxy
keto ester. While the (R)-alcohol was favored, the enantioselectivity was poor (41% ee). A variety of process conditions were eval-
uated in order to improve both the enantioselectivity and yield of this reduction. Including a nonpolar resin in the reaction mixture
afforded the (R)-alcohol in 94% ee and 50% isolated yield. The enantioselectivity was further improved to >99% ee by substituting
purified YGL157w in place of whole yeast cells. This reductase was identified by screening a collection of yeast enzymes uncovered
by genome sequence analysis.
ꢀ
2004 Elsevier Ltd. All rights reserved.
Optically active d-hydroxy-b-keto esters are valuable
chiral building blocks for the synthesis of chiral drugs
This required a reduction biocatalyst with complemen-
tary enantioselectivity. Here, we disclose our results on
the synthesis of (R)-2.
1
,2
4
and natural products of polyketide origin.
We re-
cently described a synthesis of the densely functional-
ized, enantiopure hydroxy keto ester (S)-2 that relies
on a highly regio and enantioselective enzymatic reduc-
A GC–MS-based screen of commercially available bio-
catalysts revealed that fermenting cells of pressed
bakerÕs yeast (BY, Saccharomyces cerevisiae) reduced
3
tion of diketo ester 1 (Scheme 1). While the Lactobacil-
lus brevis reductase has effectively solved the problem of
preparing (S)-2, the impact of this strategy would be en-
hanced significantly by ready access to the antipode of
this chiral building block in enantiomerically pure form.
diketo ester 1 at C with high regioselectivity (Scheme
5
1). Reduction at C3 occurred to a negligible extent
(< 5%). Unfortunately, both the enantiomeric excess
and the reaction yield of (R)-2 were disappointingly
low (41% ee, 52% yield). We, therefore, investigated
reaction parameters known to influence the performance
5
of yeast-mediated ketone reductions (Table 1).
Using resting, rather than fermenting yeast cells, pro-
vided a higher fractional conversion (72%); however,
the enantiomeric excess of (R)-2 increased only slightly
(
entry 1). By contrast, a biphasic water/n-hexane med-
ium significantly improved the enantiomeric purity of
R)-2. Interestingly, the level of improvement depended
(
Scheme 1. Reagents and conditions: (i) see Refs. 3 and 4; (ii) 20mM 1,
À1
on the water/alkane ratio (63–78% ee, entries 2–4). For-
mation of the desired product was depressed by the pres-
ence of the organic phase, however, and conditions
providing the highest enantioselectivity (water/hex-
ane = 25:75) afforded a fractional conversion of only
BY (12g/mmol 1), sucrose (200gL ), H
2
O, 26h, 20 ꢁC.
*
Corresponding authors. Tel.: +49 0 2461 613374; fax: +49 0 2461
63870 (M.M.); tel.: +1 352 8460743; fax: +1 352 8462095
J.D.S.); e-mail addresses: mi.mueller@fz-juelich.de; jds2@chem.
1
38%. Neither entrapping the biocatalyst in calcium alg-
(
inate pellets (immobilized bakerÕs yeast, IBY), adding
sucrose as a source of reducing equivalents, nor includ-
ing additives known to inhibit yeast reductase activities
ufl.edu
Present address: DSM Pharma Chemicals, Donaustaufer Str. 378,
D-93055 Regensburg, Germany.
0
957-4166/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2004.07.045

2826
M. Wolberg et al. / Tetrahedron: Asymmetry 15 (2004) 2825–2828
Table 1. Reduction of diketo ester 1 by bakerÕs yeast (whole cells)
a
b
c
d
e
f
(R)-2 (%)
g
Ee (%)
Entry
Yeast
c (1) (mM)
BY:1
H
2
O:C
6
H
2
O:DBY
Additive (concn)
t (h)
1
2
3
4
5
BY
BY
20
18
18
18
20
12
20
18
18
18
18
18
20
20
20
10
20
20
10
5
12
12
12
12
12
100:0
90:10
50:50
25:75
0:100
––
––
––
––
––
––
––
––
––
––
24
223
22
22
78
29
6.5
29
15
15
15
15
19
19
19
72
766
45
48
BY
BY
73
78
75
38
40
IBY
18
À1
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
IBY
90:10
12
––
Sucrose (100gL
)
6
74
6
IBY
IBY
IBY
IBY
IBY
IBY
100:0
90:10
90:10
90:10
90:10
90:10
100:0
100:0
90:10
––
––
––
––
––
––
10
5
––
––
49
61
12
12
12
12
12
5
84
50
À1
ClCH
ClCH
2
2
COOMe (0.7gL
COOH (0.7gL
)
6
6
À1
0
)
40
53
72
6
À1
1
Crotonic acid (3.0gL
Allyl bromide (4.0gL
)
9
À1
2
)
47
79
57
69
89
80
3
DBY1
DBY1
DBY1
20
––
––
––
4
10
10
90 (8)
89 (8)
89 (10) 89
5
6DBY1
4.5
50:50
2.5
––
21
7
DBY1
DBY1
DBY1
DBY1
DBY2
DBY3
DBY1
DBY1
DBY1
21
10
10
10
10
10
10
10
10
10
35:65
20:80
25:75
13:87
50:50
50:50
50:50
50:50
100:0
1.8
1.0
2.5
2.5
2.5
2.5
2.5
2.5
5
––
––
––
––
––
––
––
––
19
21
78 (4)
72
67
88
88
92
93
72
78
94
8
9
19
19
0
62
1
20
20
19
22
84 (11)
69 (30)
59 (6)
79 (7)
88 (10)
2
h
3
20
14 + 2
6+ 1
6+ 16
15
h
4
20
h
92.5
92
94
5
6DBY1
20
––
XAD-7 (22.5gL
À1
i
10
100:0
4.7
)
50
a
BY: ÔVital GoldÕ (Deutsche Hefewerke GmbH, local grocery); DBY1: ÔWindhagerÕ emulsifier-free (Bio-Zentrale GmbH, local grocery); DBY2: Ôtype
IIÕ (Sigma); DBY3: Ôactive driedÕ (ICN).
Gram cell preparation per mmol 1.
Ratio water/hexane v/v.
b
c
d
e
f
Milliliter water per gram cells.
Preincubation time 0.5–3h.
Fractional conversion determined by GC–MS after extraction of aliquots with CH
brackets = percent dihydroxyhexanoate (if >5%).
2 2 3 2
Cl and derivatization with (CF CO) O/pyridine; numbers in
g
Determined by chiral stationary-phase-HPLC after extraction of aliquots with dichloromethane and cyclization with TFA (Merck ChiraSpher NT,
MTBE/MeOH 98:2, 40ꢁC). Peak identity verified by comparison with both racemic 2 and (S)-2.
Final concentration (feed).
h
i
Isolated yield of purified product.
improved the situation. In all cases, increased enantio-
selectivity was accompanied by decreased formation of
excess of (R)-2 improved to 93% ee under these condi-
tions; however, the fractional conversion remained
unsatisfactory (entries 19 and 20). Yeast cells from two
additional suppliers were also examined (DBY2 and
DBY3). Both performed worse in terms of enantioselec-
tivity, however (entries 21 and 22). Moreover, DBY3 af-
forded 30% of the doubly-reduced product.
(
R)-2 (entries 5–12).
On the other hand, significant improvements were found
by changing the biocatalyst source. Employing freeze-
dried bakerÕs yeast cells at a comparable cell/substrate
ratio (DBY1) gave a fractional conversion of 79% along
with an enantiomeric excess of 69% under single phase,
aqueous conditions (entry 13). A further step forward
was achieved by increasing the yeast cell/substrate ratio,
which provided a fractional conversion of 90% and an ee
of 89% (entry 14). Under these conditions, 8% of the
doubly-reduced product t-butyl 6-chloro-3,5-dihydroxy-
Based on the above mentioned results, we concluded
that maximum enantioselectivity was observed when
the aqueous phase concentration of 1 was kept low. Di-
keto ester 1 was, therefore, added over a period of 14h,
which improved the enantiomeric excess of (R)-2 to
94%. Unfortunately, the fractional conversion was not
enhanced (59%, entry 23). An acceptable conversion of
79% with an only slightly diminished ee of 92.5% was
realized when the substrate was added over a 6h period
(entry 24). The fractional conversion was further im-
proved by omitting the organic phase (88%, entry 25).
4
hexanoate was also observed (as a 1:1 mixture of dia-
stereomers). Employing DBY1 cells under biphasic
conditions did not improve the enantiomeric purity of
(
R)-2 further (entries 15–18). Moreover, substrate con-
version was decreased by low water content, and a min-
imum of 2.5mL water per gram of dry cell mass was
necessary to rehydrate the cells fully. This meant that
the organic phase fraction could only be increased by
lowering the substrate concentration if a constant
yeast/substrate ratio was maintained. The enantiomeric
With the optimized reaction parameters in hand, the
stage was set for preparative bioconversions (on 0.5–
5.2g scales). We were disappointed to discover that the
isolated yields of (R)-2 were limited to 21–34% under

M. Wolberg et al. / Tetrahedron: Asymmetry 15 (2004) 2825–2828
2827
these conditions, although the reductions proceeded
with 90–91.5% ee. Product recovery was complicated
by the presence of a rather stable emulsion consisting
of n-hexane, water, and a large amount of cell material.
Replacing the liquid organic phase with an adsorber
keto ester (R)-2 could be obtained on >0.5g scales this
2
way.
1
In conclusion, we have demonstrated that bakerÕs yeast
is a favorable biocatalyst for the synthesis of the synthet-
ically valuable hydroxy keto ester (R)-2 by highly regio
and enantioselective reduction of 1. Under optimized
reaction conditions involving whole yeast cells, the
product (R)-2 was obtained in 94% ee and in 50% iso-
lated yield. The enantioselectivity could be improved
even further by employing individual yeast reductases
that were identified empirically following genome analy-
sis. Both modes of reduction can be applied on prepar-
ative scales. Morever, we have shown that the ability to
identify individual biocatalysts that cleanly carry out
ÔdifficultÕ conversions, coupled with technologies for
TM
6,7
resin (Amberlite XAD-7) solved this problem. The
diketo ester substrate was preadsorbed onto the resin,
then it was added to an aqueous suspension of the yeast
cells. Slow substrate addition was unnecessary since the
adsorber resin acted as a reservoir to maintain a low
steady-state aqueous concentration of 1. Because the
product (R)-2 also adsorbed to the resin, it was easily
separated from the cell suspension by filtration at the
end of the reaction. The product (R)-2 was isolated in
a 50% yield after column chromatography under these
8
–10
conditions (94% ee, Scheme 2).
Control experiments
2
2
showed that the modest yield was not due to product
decomposition during the isolation and purification
employing these enzymes under whole-cell or cell-free
4
conditions, demonstrates how collaboration between
chemical and biological methodologies can result in
highly effective processes.
1
1
process.
Acknowledgements
The work carried out by M.W. and M.M. was gener-
ously supported by the Deutsche Forschungsgemein-
schaft (Sonderforschunsgbereich 380 and Graduier-
tenkolleg 440). The skillful technical assistance of
Mrs. Silke Bode is gratefully acknowledged. Generous
financial assistance by the National Science Foundation
Scheme 2. Reagents and conditions: (i) 20mM 1 (18% on XAD-7),
DBY 1 (10g/mmol 1), H
2
O, 20h, 20ꢁC; (ii) 5mM 1, YGL157wp,
NADP , glucose-6-phosphate, glucose-6-phosphate dehydrogenase,
+
cyclodextrin, KPi buffer, pH6, 6h, 30ꢁC.
(CHE-0130315) and a Ruegamer Fellowship supported
the work performed by I.A.K. and J.D.S.
While the process optimization studies described above
provided an acceptable solution to problems of yield
and stereoselectivity in the reduction of 1, our previous
experience suggested that cell free conditions using iso-
lated enzymes might provide an even cleaner bioprocess.
The simultaneous operation of multiple yeast reductases
with differing stereoselectivities is a major reason that
reductions by whole bakerÕs yeast cells often lead to
multiple products. Moreover, the large number of
potential yeast reductases encoded by the S. cerevisiae
References
1
. Wess, G.; Kesseler, K.; Baader, E.; Bartmann, W.; Beck,
G.; Bergmann, A.; Jendralla, H.; Bock, K.; Holzstein, G.;
Kleine, H.; Schnierer, M. Tetrahedron Lett. 1990, 31,
2545–2548; Brower, P. L.; Butler, D. E.; Deering, C. F.;
Le, T. V.; Millar, A.; Nanninga, T. N.; Roth, B. D.
Tetrahedron Lett. 1992, 33, 2279–2282.
2. Evans, D. A.; Gauchet-Prunet, J. A.; Carreira, E. M.;
1
2
genome (ca. 50) makes it difficult to eliminate un-
3,14
1
wanted competitors by selective inhibition
1
or gene
This situation also makes it dif-
5–17
knockout strategies.
Charette, A. B. J. Org. Chem. 1991, 56, 741–750.
. Wolberg, M.; Hummel, W.; Wandrey, C.; M u¨ ller, M.
Angew. Chem., Int. Ed. 2000, 39, 4306–4308.
3
4
5
ficult to discover which yeast enzyme(s) participate in
reducing specific substrates. Since the classical approach
to isolating and identifying the yeast reductase(s)
responsible for the (R)-selective conversion of diketo
ester 1 was expected to be a troublesome and lengthy
process, we instead took advantage of a library of
bakerÕs yeast ketone reductases that was assembled from
. Spectral data for (R)-2: Wolberg, M.; Hummel, W.;
M u¨ ller, M. Chem. Eur. J. 2001, 7, 4562–4571.
. (a) Gottwald, M., In Methods of Organic Chemistry, 4th
ed.; Houben-Weyl, 1995; E21d, pp 4143–4197; (b) San-
taniello, E.; Ferraboschi, P.; Grisenti, P.; Manzocchi, A.
Chem. Rev. 1992, 92, 1071–1140; (c) Csuk, R.; Gl a¨ nzer, B.
I. Chem. Rev. 1991, 91, 49–97; (d) Servi, S. Synthesis, 1990,
1–25.
1
8,19
an analysis of the fully sequenced yeast genome.
This collection contained 12 of the most promising
reductases, each expressed as a fusion protein with
glutathione S-transferase to allow one-step purification.
6 . DÕ Arrigo, P.; Pedrocchi Fantoni, G.; Servi, S.; Strini, A.
Tetrahedron: Asymmetry 1997, 8, 2375–2379.
7
. Vicenzi, J. T.; Zmijewski, M. J.; Reinhard, M. R.; Landen,
B. E.; Muth, W. L.; Marler, P. G. Enzyme Microb.
Technol. 1997, 20, 494–499.
Screening our collection of 12 S. cerevisiae reductases re-
vealed that three NADPH-dependent members of this
library reduced diketo ester 1: YGL157w, YOL151w,
and YDR386w. All three enzymes provided the hydroxy
keto ester 2 in the desired (R)-configuration and with
enantiomeric excess values of >99%, 96.5%, and
8
. Vicario, J. L.; Job, A.; Wolberg, M.; M u¨ ller, M.; Enders,
D. Org. Lett. 2002, 4, 1023–1026.
9
. Enders, D.; Vicario, J. L.; Job, A.; Wolberg, M.; M u¨ ller,
M. Chem. Eur. J. 2002, 8, 4272–4284.
10. Representative procedure: Commercially available
2
0
TM
>
99%, respectively (Scheme 2). Enantiopure hydroxy
Amberlite XAD-7 resin was washed prior to use with

2828
M. Wolberg et al. / Tetrahedron: Asymmetry 15 (2004) 2825–2828
water, acetone, and ethyl acetate, and dried under reduced
pressure until the weight remained constant. Dry resin
9.0g) was added to a solution of diketo ester 1 (2.00g,
.5mmol) in ethyl acetate (70mL), and the solvent was
thoroughly evaporated under reduced pressure. The
charged resin was added to a suspension of DBY-1
14. Nakamura, K.; Kawai, Y.; Oka, S.; Ohno, A. Bull. Chem.
Soc. Jpn. 1989, 62, 875–879.
15. Zhou, B.-N.; Gopalin, A. S.; VanMiddlesworth, F.; Shieh,
W.-R.; Sih, C. J. J. Am. Chem. Soc. 1983, 105, 5925–
5926.
(
8
16. Rodriguez, S.; Kayser, M. M.; Stewart, J. D. Org. Lett.
1999, 1, 1153–1155.
(
85g) in deionized water (400mL), and the mixture was
shaken at 20ꢁC (130rpm, 1L shake flask, horizontal
shaker). After 15h, the resin was collected on a sintered
glass funnel (porosity 0), washed with a minimal amount
of water, and extracted with acetone (4 · 50mL) and ethyl
acetate (50mL). The extract was concentrated under
reduced pressure, and the residue dissolved in ethyl
acetate. The solution was washed with aq NaHCO (5%)
3
4
and brine, dried over MgSO , and the solvent evaporated
under reduced pressure. Flash chromatography on silica
gel (ethyl acetate/isohexane 40:60 v/v) gave hydroxy keto
ester (R)-2 as a pale yellow oil. Yield: 1.01g (50%).
17. Rodriguez, S.; Kayser, M. M.; Stewart, J. D. J. Am. Chem.
Soc. 2001, 123, 1547–1555.
18. Kaluzna, I.; Andrew, A. A.; Bonilla, M.; Martzen, M. R.;
Stewart, J. D. J. Mol. Catal. B: Enzym. 2002, 17, 101–105.
19. Kaluzna, I. A.; Matsuda, T.; Sewall, A. K.; Stewart, J. D.
J. Am. Chem. Soc., submitted for publication.
20. The following yeast enzymes were screened (as GST fusion
proteins): YOL151w, YDR368w, YDR541c, YGL039w,
YGL157w, YDL124w, YJR096w, YCR107w, YNL331c,
YBR149w, YNL274c and YPL113c. Assay mixtures
+
contained NADP (0.15mg, 0.20lmol), glucose-6-phos-
1
1. Two control experiments were carried out in order to
phate (4.3mg, 14lmol), glucose-6-phosphate dehydro-
genase (Sigma type XV, 5lg), 5mM 1, 1.0equiv of
hydroxypropyl-b-cyclodextrin and 100lL of purified
GST-fusion protein in a final volume of 1.0mL of
100mM KPi, pH7. Reactions were incubated for 24h at
30ꢁC and product formation was monitored by GC after
extraction with CH Cl . The YOL151w, YDR368w and
YGL157w GST-fusion proteins reduced 1; none of the
remaining nine displayed catalytic activity toward this
substrate.
show that neither the reaction nor isolation procedures
diminished the yield or enantiomeric excess of (R)-2. To
probe product stability under aqueous reaction condi-
tions, a 50mg portion of (S)-2 (>99% ee) was dissolved in
1
(
.5mL of ethanol, then this was added to 50mL of KP
pH6.0) and the mixture was incubated at 30ꢁC for 18h.
After extraction with CH Cl , the combined organics were
dried with MgSO and concentrated to afford 49mg of (S)-
(shown to be >99% ee by chiral-phase HPLC analysis).
i
2
2
2
2
4
2
The stability of the product toward chromatography was
demonstrated by filtering a 50mg portion of (S)-2 through
a 1 · 5cm silica column using 35:65 ethyl acetate/hexanes.
The eluent was concentrated to yield 50mg of (S)-2 whose
enantiomeric excess was >99% ee by chiral-phase HPLC
analysis.
21. Representative procedure: Reaction mixtures contained
NADP (0.13g, 0.17mmol), glucose-6-phosphate (3.5g,
+
10.4mmol), glucose-6-phosphate dehydrogenase (Sigma
type XV from bakerÕs yeast, 2mg), hydroxypropyl-b-
cyclodextrin (9.6g, 6.4mmol), 1 (0.65g, 2.8mmol; dis-
solved in 50mL of EtOH) and purified YGL157w fusion
protein (180mg) and 630mL of 100mM KPi, pH6.0. The
mixture was incubated at 30ꢁC for 6h, then extracted with
1
1
2. Stewart, J. D.; Rodriguez, S.; Kayser, M. M. In Enzyme
Technologies for Pharmaceutical and Biotechnological
Applications; Kirst, H. A., Yeh, W.-K., Zmijewski, M.
J., Eds.; Marcel Dekker: New York, 2001; pp 175–208.
3. Nakamura, K.; Inoue, K.; Oshio, K.; Oka, S.; Ohno, A.
Chem. Lett. 1987, 679–682.
CH
2
Cl
2
(3 · 600mL). The combined organics were dried
and concentrated in vacuo.
22. Walton, A. Z.; Stewart, J. D. Biotechnol. Progr. 2004, 20,
with MgSO
4
403–411.