Efficient bioreduction of bicyclo[2.2.2]octane-2,5-dione and bicyclo[2.2.2]oct-7-ene-2,5-dione by genetically engineered Saccharomyces cerevisiae

Article abstract of DOI:10.1039/b603500k

A screening of non-conventional yeast species and several Saccharomyces cerevisiae (baker's yeast) strains overexpressing known carbonyl reductases revealed the S. cerevisiae reductase encoded by YMR226c as highly efficient for the reduction of the diketones 1 and 2 to their corresponding hydroxyketones 3-6 (Scheme 1) in excellent enantiomeric excesses. Bioreduction of 1 using the genetically engineered yeast TMB4100, overexpressing YMR226c, resulted in >99% ee for hydroxyketone (+)-4 and 84-98% ee for (-)-3, depending on the degree of conversion. Baker's yeast reduction of diketone 2 resulted in >98% ee for the hydroxyketones (+)-5 and (+)-6. However, TMB4100 led to significantly higher conversion rates (over 40 fold faster) and also a minor improvement of the enantiomeric excesses (>99%). The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b603500k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Efficient bioreduction of bicyclo[2.2.2]octane-2,5-dione and
bicyclo[2.2.2]oct-7-ene-2,5-dione by genetically engineered
Saccharomyces cerevisiae
Annika Friberg,a Ted Johanson,b Johan Franze´n,a Marie F. Gorwa-Grauslund*b and Torbjo¨rn Frejd*a
Received 8th March 2006, Accepted 6th April 2006
First published as an Advance Article on the web 8th May 2006
DOI: 10.1039/b603500k
A screening of non-conventional yeast species and several Saccharomyces cerevisiae (baker’s yeast)
strains overexpressing known carbonyl reductases revealed the S. cerevisiae reductase encoded by
YMR226c as highly efficient for the reduction of the diketones 1 and 2 to their corresponding
hydroxyketones 3–6 (Scheme 1) in excellent enantiomeric excesses. Bioreduction of 1 using the
genetically engineered yeast TMB4100, overexpressing YMR226c, resulted in >99% ee for
hydroxyketone (+)-4 and 84–98% ee for (−)-3, depending on the degree of conversion. Baker’s yeast
reduction of diketone 2 resulted in >98% ee for the hydroxyketones (+)-5 and (+)-6. However,
TMB4100 led to significantly higher conversion rates (over 40 fold faster) and also a minor
improvement of the enantiomeric excesses (>99%).
Introduction
Optically active bicyclo[2.2.2]octane derivatives have attracted
interest as rigid building blocks in natural product synthesis1–3 and
have recently been investigated as therapeutic agents for cocaine
abuse.4
The diketones 1 and 2 (Scheme 1) have been used as starting
materials for the synthesis of CNS-modulators,5 and optically
active 2 for the synthesis of chiral ligands.6,7
Optically active 2 has previously been obtained by resolution of
the racemate by fractional recrystallization of its dihydrazone of
(R)-5-(1-phenylethyl)semioxamazide,8 by a 1,2-carbonyl transpo-
sition route from (1R,4S,6S)-6-hydroxybicyclo[2.2.2]octan-2-one9
or from its corresponding enantiomer of 1.10,11
The racemic diketone 1 has been resolved as diastereomeric di-
ethyl(R,R)-(+)-tartrate acetals using HPLC10 and as an inclusion
complex with (S)-(−)-10,10-dihydroxy-9,9-biphenantryl.11 In the
work of Hill et al.10 and Lightner et al.12 the diacid 7 (Scheme 2)
was also resolved as its brucine salt followed by electrolysis to give
(−)-1 of 90.4% ee, however in low total yield. For the synthesis
of the diketones 1 and 2 in multigram quantities of high optical
purity we did not find any of these routes satisfactory. Interestingly,
Lightner et al. also used baker’s yeast for the resolution of ( )-1, to
give hydroxyketone (−)-3 and unreacted diketone (+)-1 of varying
ee values depending on the incubation time. Although baker’s
yeast, Saccharomyces cerevisiae, is readily available, inexpensive
and non-pathogenic,13 drawbacks such as low reductase activities
towards xenobiotic compounds, low yields and low ee values
due to competing enzymes with overlapping activities, have often
limited its applicability.14,15 However, the use of recombinant
DNA technology can deal with many of these shortcomings.
Overexpression of relevant reductases not only increases reaction
Scheme 1 Reagents and conditions: (a) yeast transformation; (b) H2,
Pd/C, EtOH, 1 atm, rt, 3 h; (c) TPAP, NMO, 3 A molecular sieves, CH2Cl2,
rt, 2 h.
˚
rates but also reduces the fraction of unwanted products from
competing enzymes and yeast metabolism thereby increasing both
yield and ee.15 The reductions can often be achieved with smaller
amounts of yeast in a shorter time, thus facilitating product
recovery. Moreover, by employing a strain with a decreased phos-
phoglucose isomerase (PGI) activity, the glucose consumption
can be decreased several fold without affecting the supply of
the needed NADPH cofactor.16,17 This decreases the formation
of ethanol and other metabolic by-products and prolongs the
activity of the yeast by keeping the concentration of inhibitory
metabolites low and by delaying glucose depletion. Previous work
has shown that the open reading frames (ORFs) YMR226c,
aDivision of Organic Chemistry, Center for Chemistry and Chemical
Engineering, Lund University, P.O. Box 124, SE-221 00, Lund, Sweden
bDepartment of Applied Microbiology, Center for Chemistry and Chemical
Engineering, Lund University, P.O. Box 124, SE-221 00, Lund, Sweden
2304 | Org. Biomol. Chem., 2006, 4, 2304–2312
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
In a similar experiment, using approximately twice the amount
of baker’s yeast and an incubation time of 2.5 d, we obtained a
somewhat different result. The hydroxyketone (−)-3 was isolated
in 45% yield and 80% ee, the recovered diketone (+)-1 in 40%
yield and 94% ee and the hydroxyketone (+)-4 was formed in 12%
yield and 80% ee. Also, further reduction experiments revealed, in
agreement with Lightner et al., that the ee values were depending
on the incubation time. Long incubation time led to a higher ee
(and lower yield) of the recovered diketone (+)-1 and a lower ee
(and higher yield) of (−)-3 while the opposite was true for a short
incubation time.
To improve the enantioselectivity and possibly the diastere-
oselectivity in the baker’s yeast reduction and to achieve a
higher reproducibility of the enantiomeric excesses of the isolated
products, a screening of yeast species for the asymmetric reduction
of ( )-1 was performed (Table 1). Non-conventional yeasts and
S. cerevisiae strains expressing various reductase genes were
compared to the control strain TMB4094 and to regular baker’s
yeast. In addition to the hydroxyketones (−)-3 and (+)-4, small
amounts of diols due to over-reduction of the hydroxyketones were
observed for several of the recombinant strains. Excellent ee values
(>99%) and high yields were obtained for (+)-4 using the yeasts
TMB4100 and TMB4091, both expressing the ketoreductase
encoded by YMR226c. However, the overexpression of YDR368w
and YOR120w resulted in lower ee values as compared to the
control strain. This is most likely the cause for the low enan-
tioselectivity observed in the reduction with baker’s yeast, which
should contain a higher relative proportion of these reductases as
compared to YMR226c. The results for TMB4097 was very similar
to baker’s yeast and the control strain TMB4094, indicating that
the overexpressed YOL151w has very little or no activity towards
diketone 1.
For the hydroxyketone (−)-3, a larger variation of the ee
values was observed as compared to (+)-4. The best ee values
were obtained with baker’s yeast, TMB4097 (YOL151w) and
TMB4094 (control). TMB4100 and TMB4091, both expressing
YMR226c, displayed different ee values of the major enantiomer
(−)-3, varying conversions and amounts of by-products. The ee
dependence on reduction time and conversion could be explained
if 3 was further reduced to form diols. To investigate this
hypothesis, the ee of (−)-3 was monitored during a prolonged
reduction of 1 with TMB4100 (Fig. 1).
Scheme 2 Reagents and conditions (Paquette et al.10,12): (a) 210 ◦C, 2 h
(15%); (b) H2O, 80 ◦C (90–95%); (c) pyridine, H2O, NEt3, 45 V (42%); (d)
H2, Pd/C (10 wt%), EtOH, 1 atm, 3 h (80–100%).
YDR368w (encoding the YPR1 gene) and YOR120w (GCY1)
displayed high activities towards bicyclo[2.2.2]octane-2,6-dione,17
which is structurally similar to 1 and 2. Also, YOL151w (GRE2)
has been shown to reduce many dicarbonyls, often with a different
selectivity, generating other diastereomers and enantiomers than
normally obtained in baker’s yeast catalysed reductions.15 It is also
known that non-conventional yeast species can generate other
products or reduce compounds not accepted by S. cerevisiae15
and it has been seen in screenings of non-conventional yeast
libraries that in particular species of Rhodoturula and Candida
tend to give rise to different stereoisomers.18,19 Based on these
findings we thus anticipated that the bioreduction of ( )-1 could
be improved and that bioreduction of ( )-2 could be achieved. We
here report the screening of several yeast species, both S. cerevisae
strains with overexpressed reductases and non-conventional yeast
strains, for the stereoselective reduction of the diketones ( )-1
and ( )-2 to form the hydroxyketones (−)-3, (+)-4, (+)-5 and
(+)-6 (Scheme 1) in excellent enantiomeric excesses. It should be
noted that these hydroxyketones are not easily obtained directly
by conventional Diels–Alder routes and although seemingly
uncomplicated structures, (+)-4 has not previously been reported
in the literature and 5 only as a racemic mixture of isomers.20
Results and discussion
The racemic diketone 1 was prepared by the method of Lightner
et al.12 (Scheme 2) with minor modification. Although the yield
in the first step was low (15%), the starting materials are
cheap, purification easy and the reaction could be performed
in large scale. Racemic diketone 2 has been reported to be
available via a Diels–Alder reaction of (a) the lithium enolate of
cyclohexenone and 2-(N-methylanilino)acrylonitrile,21 (b) 2-((tri-
methylsilyl)oxy)-1,3-cyclohexadiene and a-acteoxyacrylonitrile22
or (c) 2-((trimethylsilyl)oxy)-1,3-cyclohexadiene and 2-chloro-
acrylonitrile.9 However the dienophile in route (a) requires mul-
tistep synthesis and when routes (b) and (c) were attempted,
the yields were low (typically varying between 10–25%). For
multigram quantities, racemic diketone 2 was most conveniently
prepared from 1 by catalytic hydrogenation of the double bond.11
In the work of Lightner et al., the best yeast reduction of ( )-
1 (preparative scale) afforded (−)-3 of 100% ee (12% yield) and
(+)-1 of 85% ee (28% yield) in a reduction period of 5 d.
The yield of (−)-3 reached a maximum of 42% after 3–4 h after
which it steadily decreased along with the concomitant formation
of diols and a marked drop in ee from 99 to 87%. The yield of
ketoalcohol (+)-4 increased on the other hand until it reached its
theoretical maximum of 50% after about 24 h with excellent ee
(>99%).
Reductions with non-conventional yeasts showed no improve-
ment in ee for (+)-4 as compared to baker’s yeast, though
several Candida strains reached a higher conversion. For (−)-3,
all non-conventional yeasts displayed a lowered ee. A reversed
diastereoselectivity, favoring the formation of (+)-4 over (−)-3,
was observed for Candida intermedia, Rhodoturula gracilis and
Rhodoturula mucilaginosa.
For the generation of (+)-4 and (−)-3 a bioreduction of ( )-1
was performed in a preparative scale (5 g) (Fig. 2).
TMB4100 overexpressing YMR226c and having a 1% PGI
activity was deemed the most appropriate strain. The lowered
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2304–2312 | 2305
©

View Article Online
Table 1 Screening of yeast species for the stereo-selective reduction of ( )-1
Yeast strain
(−)-3 : (+)-4 ratio
Conversion of ( )-1 (%)
(−)-3 ee (%)b
(+)-4 ee (%)b
Saccharomyces cerevisiae strains
Baker’s yeast
TMB4100 (1% PGI, YMR226c)
TMB4091 (YMR226c)
TMB4093 (YDR368w)
TMB4096 (YOR120w)
TMB4097 (YOL151w)
TMB4094 (control strain, empty plasmid)
Non-conventional yeasts
Candida boidini CBS8030
Candida intermedia PYCC4715
Candida tropicalis CBS094
Candida wickerhamii
Rhodoturula gracilis
Rhodoturula mucilaginosa
Pichia pastoris GS115
84 : 16
45 : 55
63 : 37
52 : 48
60 : 40
84 : 16
90 : 10
27
97
84
93
67
64
96
96
81
95 (8)c
76 (3)c
49 (<1)c
49 (<1)c
22
>99
>99
16
48
77
25
86
86 : 14
43 : 57
54 : 46
50 : 50
40 : 60
38 : 62
61 : 39
<2
8
30
21
3
4
4
69
84
76
56
65
66
78
77
31
33
72
54
69
72
a 5 g l−1 dw yeast, 60 g l−1 glucose, 100 mM citric acid buffer pH 5.5, 24 h. b Determined by GC on an a-DEX capillary column. c Figures in parentheses
represent the amount of diols formed.
Fig. 2 Semi fed-batch reduction of diketone 1 (ꢀ) with 15 g l−1 dw
TMB4100 to (−)-3 (ꢁ) and (+)-4 (ꢀ).
4 slowed to a halt after 50–60 h at a concentration of about
15 g l−1 (theoretical maximum 20 g l−1). However, a similar
Fig. 1 Bioreduction of 5 g l−1 diketone 1 to (−)-3 and (+)-4 with 15 g l−1
experiment using half the concentration of 1 (10 + 10 g l−1) in
dw TMB4100 in 5 ml scale (X = enantiomeric excess of (−)-3; ꢀ = 1; ꢁ =
100 mg scale showed an almost full conversion to (+)-4 after 65 h,
demonstrating that the substrate concentration is of importance
for the conversion. In this experiment a decrease of (−)-3 was
observed in the end due to diol formation (data not shown). The
observed difference in conversion to (+)-4 for the two reductions
was likely caused by the inability of the aging yeast to continue
to reduce the substrate for the additional time required when the
higher substrate concentration was used. This effect could be more
pronounced if 1, (−)-3 or (+)-4 have toxic effects on the yeast.
The diastereomers (−)-3 and (+)-4 were easily separated by
column chromatography. Thus, hydroxyketone (+)-4 of >99% ee
was obtained in 32% isolated yield and (−)-3 of 94% ee in 38%
(−)-3; ꢀ = (+)-4; ꢂ = diols).
PGI activity in TMB4100 decreases the glucose consumption
and ethanol formation rates thereby delaying glucose depletion
and allowing NADPH dependent bioreductions to continue to
higher product titers. The reaction was run in semi fed-batch
mode to decrease possible substrate inhibitory effects, which had
been observed for the reduction of a similar bicyclic diketone.16
Reduction of (−)-3 proceeded rapidly while (+)-4 lagged behind
as observed in the previous experiment. After about 15 h, most
diketone was consumed and another batch was added. Formation
of hydroxyketone (−)-3 soon reached its maximum, while (+)-
2306 | Org. Biomol. Chem., 2006, 4, 2304–2312
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
isolated yield. The diketone (+)-1 of 98% ee was recovered in 10%
yield. A higher total yield of the hydroxyketones was obtained
using TMB4100 (70%) as compared to the baker’s yeast (62%),
but the total amount of recovered material (including unconverted
1) after work up and purification was lower. In this case 80% for
TMB4100 as compared to 97% for baker’s yeast. This difference
in material recovery is most likely caused by over-reduction of
the hydroxyketones to diols, which is more pronounced with
TMB4100, which has a higher conversion rate than baker’s
yeast. Extraction of the diols from the aqueous phase, especially
in the preparative scale, was difficult due to their high water
solubility. This implies that diols may be present in larger amounts
than was observed by GC also for the screening experiments,
although a proportionally larger amount of solvent was used for
the extraction of those samples than for the preparative scale
experiments. For the best possible yields of the hydroxyketones
(−)-3 and (+)-4 to be obtained, further optimization of the process
would be required.
Although Candida boidini had a somewhat higher conversion
of 2, it also showed a marked decrease in ee for both (+)-5
and (+)-6. None of the other non-conventional yeasts showed
any improvement of neither ee nor conversion. Similarly as for
diketone 1, a reversed diastereoselectivity (favoring (+)-6 over (+)-
5) was observed with several strains.
Corresponding to diketone 1, reductions of 2 using baker’s yeast
and TMB4100 were performed in a preparative scale. The hydrox-
yketones (+)-5 and (+)-6 were isolated as diastereomeric mixtures
in 86 and 72% total yield for the baker’s yeast and TMB4100,
respectively, with ee values similar to those in the screening
experiments (Table 2). In the baker’s yeast reduction a higher
isolated yield was obtained even though full conversion of the
starting material was not achieved as was the case for TMB4100.
This result can again be explained by the larger amounts of diols
formed as a result of over-reduction using the more efficient
strain TMB4100. The best yields of the diastereomeric mixture of
(+)-5 and (+)-6 were obtained at approximately 90% conversion
of the diketone. Comparing the conversion rates (g diketone
reduced/g yeast × h) TMB4100 was found to be >40-fold faster.
In both experiments, small amounts of the diols were isolated;
predominantly the endo/exo-diol, as confirmed by 1H NMR.
Separation of the diastereomeric mixture (+)-5 and (+)-6 was
not easily achieved. Separation was unsuccessful using column
chromatography and the separation by preparative HPLC was
insufficient. Several ether and ester derivatives of the mixture were
synthesized without improving the ease of purification. Eventually
the diasteroisomers were separated by column chromatography as
their acetal acetates 9a and 9b (Scheme 3).
The screening of yeast species for the asymmetric reduction of
the racemic diketone 2 (Table 2) revealed several strains able to
generate (+)-5 and (+)-6 with excellent ee values (>98%). The
best results were again obtained using TMB4100 and TMB4091,
both overexpressing YMR226c. These two strains also displayed
a somewhat changed diastereoselectivity, favouring (+)-6 over
(+)-5, as compared to the other S. cerevisiae strains. TMB4093
and TMB4096 expressing YDR368w and YOR120w, respectively,
showed a decreased ee for (+)-6. Baker’s yeast and the control
strain also produced >98% ee for both diastereoisomers, but had
a poorer conversion. Also in these cases, diol formation was
observed during prolonged incubation times. However, in contrast
to the diastereomer (−)-3 (from the diketone 1), over-reduction did
not affect the ee of the corresponding isomer (+)-5.
Ketalization followed by esterification (step a) of the di-
astereomeric mixture of (+)-5 and (+)-6 and deprotection (step
b) of the separated diastereomers 9a and 9b were performed
sequentially without purification of the intermediate hydroxy
acetal mixtures 8a and 8b to give (+)-5 and (+)-6 in total
The screening of the non-conventional yeasts did not reveal any
strain with improved properties as compared to baker’s yeast.
Table 2 Screening of yeast species for the stereoselective reduction of ( )-2
Yeast strain
(+)-5 : (+)-6 ratio
Conversion of ( )-2 (%)
(+)-5 ee (%)b
(+)-6 ee (%)b
Saccharomyces cerevisiae strains
Baker’s yeast
TMB4100 (1% PGI, YMR226c)
TMB4091 (YMR226c)
TMB4093 (YDR368w)
TMB4096 (YOR120w)
TMB4097 (YOL151w)
TMB4094 (control strain, empty plasmid)
Non-conventional yeasts
Candida boidini CBS8030
Candida intermedia PYCC4715
Candida tropicalis CBS094
Candida wickerhamii
Rhodoturula gracilis
Rhodoturula mucilaginosa
Pichia pastoris GS115
58 : 42
48 : 52
45 : 55
57 : 43
51 : 49
58 : 42
61 : 39
26 (<1)c
91 (7)c
98 (12)c
28 (<1)c
38 (<1)c
28
99
99
>99
98
91
95
98
99
>99
>99
98
98
>99
98
26
55 : 45
56 : 44
30 : 70
48 : 52
22 : 78
25 : 75
28 : 72
39
8
7
19
6
4
86
98
97
86
87
85
81
88
90
66
95
78
88
81
10
a 5 g l−1 dw yeast, 60 g l−1 glucose, 100 mM citric acid buffer pH 5.5, 24 h. b Determined by GC on an a-DEX capillary column. c Figures in parentheses
represent the amount of diols formed.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2304–2312 | 2307
©

View Article Online
are otherwise not straightforwardly obtained via baker’s yeast
reduction due to poor conversion rates and, for (−)-3 and (+)-
4, only moderate ee values were obtained. In this screening, the
selection of yeast strains was based on the previous knowledge
that certain yeast reductases were able to reduce the related
bicyclic compound bicyclo[2.2.2]octane-2,6-dione.17 As for 1 and
2 in this work, overexpression of YMR226c was found to give
the highest reduction rate. However, for bicyclo[2.2.2]octane-2,6-
dione, YDR368w gave the highest optical purity, >98%, compared
to 93% for YMR226c.17 This demonstrates that even though
reductases may have different selectivities for related substrates,
a limited screening of modified yeast strains could be sufficient to
quickly identify a suitable strain when reductases accepting similar
compounds are already known.
This work also presents a route to the diketones 1 and 2 in
high enantiomeric excesses through the oxidation of their corre-
sponding hydroxyketones. The results presented herein may allow
for improvements of conversion rates, yields and enantiomeric
excesses for substrates where regular baker’s yeast previously has
been used with limited success.
Scheme 3 Separation of the diastereoisomers (+)-5 and (+)-6. Reagents
and conditions: (a) (i) ethylene glycol, PTS, benzene, reflux, 12 h; (ii) acetic
anhydride, pyridine, 24 h and (iii) separation of 9a and 9b by column
chromatography; (b) (i) K2CO3, MeOH, H2O, 24 h and (ii) 2M HCl,
acetone, 2 h.
yields of 39 and 38%, respectively, over the entire separation
sequence.
As we needed large amounts of the hydroxyketone (+)-6 for fu-
ture applications, it was desirable to avoid this separation sequence.
Therefore the following route was applied. Hydroxyketone (+)-6
of >99% ee was obtained in 99% yield by catalytic hydrogenation
of (+)-4 and also by baker’s yeast reduction of optically pure (+)-
2 (obtained by resolution with baker’s yeast of (+)-1 followed by
catalytic hydrogenation of the double bond). Treating the diketone
(+)-2 with 12 g l−1 dw TMB4100 for 6 h selectively gave (+)-6 of
>99% ee in 85% yield. In this experiment, the endo,endo-diol was
also isolated in 5% yield. Treating optically pure (+)-2 with baker’s
yeast resulted in slow conversion of the starting material. After a
reduction period of 7 d with yeast and glucose additions every 48 h,
the reaction was stopped and (+)-6 was isolated in only 65% yield.
It should be noted that although the diketone (+)-1 obtained from
the baker’s yeast reduction is not optically pure, the enantiomeric
excess could be increased to >99% ee by recrystallization from
diisopropyl ether. Although these routes resulted in hydroxyketone
(+)-6, the most efficient one was to reduce diketone ( )-1 to (+)-4
using TMB4100 followed by catalytic hydrogenation of the double
bond (Scheme 1).
Experimental
General
Compressed commercial baker’s yeast (Kronja¨st from Ja¨stbolaget
AB, Sollentuna, Sweden), was purchased from a local distrib-
utor. Saccharomyces cerevisiae CEN.PK113-7A [MAT a, MAL-
8c; SUC2; his3-D1] was a gift from Dr P. Ko¨tter, Institute of
Microbiology, Frankfurt, Germany. S. cerevisiae strains over-
expressing YMR226c (TMB4091), YDR368w (TMB4093),23
YOR120w (TMB4096)17 and a control with an empty plasmid
(TMB4094)23 were previously constructed from a CEN.PK back-
ground. TMB4100 (YMR226c) was previously constructed from
S. cerevisiae RBY 7–1 (ENY.WA-1A, with 1% PGI activity,
encoded by YBR196c) as described elsewhere.17
The non-conventional yeast strains used in the screening
were the following: Candida wickerhamii obtained from the
Yeast Culture Collection of the University of the Free State,
Bloemfontein, South Africa, Candida tropicalis CBS094, Candida
boidini CBS8030 Candida intermedia PYCC4715, Rhodotorula
mucilaginosa and Rhodotorula gracilis obtained from the Division
of Food Biotechnology and Microbiology, Warsaw Agricultural
University, Poland and Pichia pastoris GS115 (Invitrogen). Es-
cherichia coli DH5a (Life Technologies, Rockville, MD) was used
for all subcloning.
The formation of only the diastereoisomer (+)-6 from (+)-
2 also confirms the absolute configuration of (+)-6. Lightner
et al. previously determined the absolute configuration of (−)-
3 and thus the absolute configuration of (+)-5 was confirmed by
hydrogenation of the double bond of (−)-3.
Oxidation of the hydroxyketones 3–6 using tetrapropylam-
moniumperruthenate (TPAP) smoothly afforded the optically
active diketones (+)-1, (−)-1, (+)-2 and (−)-2 in 91–100% yield
with ee values corresponding to their precursor hydroxyketones
(Scheme 1).
GLC (Perkin-Elmer AutoSystem XL) was performed with
a Factor Four capillary column (Varian, 30 m × 0.25 mm,
0.25 lm film thickness) and with an a-DEX 120 permethylated
a-cyclodextrin fused silica capillary column (Supelco, 30 m ×
0.25 mm, 0.25 lm film thickness) for determination of◦enan-
tiomeric compositions. The samples were analysed at 100 C (60
min) −2 ◦C min−1 −180 ◦C. The retention times were: 79.34 min
((+)-3), 79.64 min ((−)-3), 83.84 min ((−)-4), 84.17 min ((+)-
4), 84.45 min ((−)-5), 84.80 min ((+)-5), 86.67 min ((−)-6) and
86.98 min for (+)-6.
Conclusions
The screening of the non-conventional and genetically modified
yeasts identified several yeast strains that markedly improved the
bioconversions of the diketones 1 and 2 to their corresponding
hydroxyketones. The genetically modified yeast TMB4100 was
successfully applied for the synthesis of the optically active
hydroxyketones 3–6 with high ee values. These compounds
HPLC analyses were performed on a Chiralcel OD–H column
(250 × 4.6 id, 5 lm) (Daicel) on a Varian Prostar system equipped
with a PDA detector. Flow rate: 1.0 ml min−1, detection at 220 nm.
2308 | Org. Biomol. Chem., 2006, 4, 2304–2312
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Solvent: hexane–2-propanol 90 : 10, 13.6 min ((−)-1), 19.3 min
((+)-1).
and the combined aqueous phase was repeatedly extracted with
EtOAc and dried. The organic phase was filtered and the solvent
removed at reduced pressure.
All NMR spectra were recorded on a Bruker ARX 300
spectrometer, unless otherwise indicated, using CDCl3 (CHCl3 d
7.26 (1H) and 77.9 (13C)) or C6D6 (C6H6 d 7.16 (1H) and 128.39
(13C)) as solvents. Column chromatography was performed on
Matrex (25–70 lm) silica gel. TLC was carried out on silica gel
(60 F254, Merck), the plates were impregnated with a solution of
KMnO4 (10 g), K2CO3 (50 g), 5% NaOH (20 ml) and H2O (900 ml)
and the compounds visualized upon heating. All Rf values are
given using heptane–EtOAc 1 : 1 as the mobile phase. Melting
points were taken on a Sanyo Gallenkamp melting point apparatus
(MPD.350.BM3.5) and are uncorrected. Optical rotations were
measured on a Perkin-Elmer 241 polarimeter at 22 ◦C and are
given in 10−1 deg cm2g−1. Elemental analyses were performed by
H. Kolbe Mikroanalytisches Laboratorium. Organic extracts were
dried using Na2SO4 throughout.
General description 3
For the screening of yeast species for asymmetric reduction of 1
and 2 to give 3–6, the following procedure was applied: Screening
for whole-cell reduction of ( )-1 and ( )-2 was made with 5 g l−1
dry weight yeast in 1 ml medium composed of 100 mM citric acid
buffer, pH 5.5 and 60 g l−1 of glucose using 5 mg of ( )-1 or ( )-2.
◦
The experiments were carried out at 30 C on a rocking table in
1.5 ml Eppendorf tubes provided with syringe holes for outlet gas
for 24 h. The reaction mixtures were extracted with EtOAc (5 ml),
dried (Na2SO4), filtered and analyzed by GC.
Molecular biology methods
Plasmid DNA was prepared with the Biorad miniprep kit (Her-
cules, CA). DNA enzymes were obtained from Fermentas (St.
Leon-Rot, Germany) or Life Technologies (Rockville, MD). DNA
extractions from PCR reactions and agarose gels were made using
the QIAquick extraction kit (Qiagen, Hilder, Germany).
Competent DH5a E. coli cells were prepared using the method
of Inoue et al.24 Yeast transformation was made as described
elsewhere.25 E. coli transformants were selected on Luria–Bertani
(LB) medium26 plates with 100 lg ml−1 ampicillin (IBI Shelton
Scientific, Inc., Shelton, CT). E. coli transformants were grown
over night in LB medium with 100 lg ml−1 ampicillin. S. cerevisiae
transformants were selected on minimum medium agar plates
containing 40 g l−1 glucose, 6.7 g l−1 nitrogen base and 20 g l−1
agar.
General description 1
For the baker’s yeast catalysed reductions performed on a prepar-
ative scale, the following procedure was applied: Sucrose (35.5 g)
and Baker’s yeast (35.5 g compressed yeast corresponding to 9.6 g
dry weight) was added to a solution of diketone (22.0 mmol) in
water (185 ml). The resulting suspension was slowly stirred at 30 ◦C
for 60 h (( )-1) (48 h for ( )-2 and 7 d for (+)-2) and then filtered
through a pad of Celite. The resulting yeast cake was incorporated
in the Celite by stirring with a spatula to remove lumps before it
was carefully rinsed with EtOAc. The aqueous phase was extracted
with EtOAc (5 × 150 ml) and dried. The organic phase was filtered
and the solvent removed at reduced pressure.
General description 2
Construction of strain TMB4097
For the preparative scale reductions using the genetically engi-
neered yeast TMB4100, the following procedure was applied:
bioreductions were performed at 30 ◦C in 100 mM citric acid
buffer, pH 5.5 supplemented with 60 g l−1 of glucose. Reductions
in 5 ml scale were run in glass vials with rubber stoppers equipped
with magnetic stirrers and syringes for outlet gas and sampling.
Reductions in 90 and 120 ml scale were performed in a 150 ml
water tempered fermentor equipped with a magnetic stirrer and a
cotton plugged syringe for outlet gas. Cell free medium from the
reductions was collected by centrifugation at 4000 g for 5 min.
The reductions were performed with 15 g l−1 dry weight TMB4100
except for the reduction of (+)-2 which used 12 g l−1 dw. The
reduction times were 67.5 h (( )-1) (6 h for ( )-2 and 6.5 h
for (+)-2). The yeast was separated from the reaction mixtures
by centrifugation and the resulting yeast pellets were washed
once with a small amount of Milli-Q water (10% of the reaction
volume). The wash water was combined with the reaction medium
The YOL151w (GRE2) gene was amplified from chromosomal
S. cerevisiae DNA using primers containing specific restriction
sites (Table 3). The shuttle vector p423-GPD (with HIS3 marker)
and the YOL151w PCR product were digested with BamHI, SpeI
and ligated at 16 ◦C overnight. The resulting plasmid p423GPD-
YOL151w (HIS3) was amplified in E. coli DH5a and used to
transform S. cerevisiae CEN.PK113-7A generating TMB4097.
Yeast growth
Yeast strains were kept at −80 ◦C and streaked on rich agar
plates containing 20 g l−1 glucose, 20 g l−1 tryptone, 10 g l−1 yeast
extract, 15 g l−1 agar and 100 mM phosphate buffer pH 6.3, except
TMB4100 which was streaked on the same medium but with 20 g
l−1 fructose and 1 g l−1 glucose. Colonies were taken from the agar
plates and used to inoculate 25–200 ml medium in shake flasks with
sterile cotton applicators. Non-conventional yeasts were grown in
Table 3 Primers and restriction sites (bold) for the construction of the S. cerevisiae strain TMB4097 overexpressing the reductase encoded by YOL151w
Primer for amplification of YOL151w (GRE2)
Restriction site
5ꢁ-AAACTAGTAACAGATAGCAGTATCACACGCCCGTAAAT-3ꢁ
SpeI
BamHI
5ꢁ-AAGGATCCGAAGAGAAAAATGCGCAGAGATGTACTAGATGAT-3ꢁ
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2304–2312 | 2309
©

View Article Online
20 g l−1 glucose, 20 g l−1 tryptone, 10 g l−1 yeast extract in 100 mM
phosphate buffer pH 6.3. TMB-strains were grown in 20 g l−1
glucose and 7 g l−1 yeast nitrogen base in 100 mM phosphate buffer
pH 6.3, except TMB4100 which was grown in the same medium
but with 20 g l−1 fructose, 1 g l−1 glucose and supplemented with
50 mg l−1 of tryptophan, uracil and leucine. The cells were grown
over-night at 30 ◦C and 200 rpm, except TMB4100 which was
grown for 40 h. Cells were harvested by centrifugation at 4000 g
for 5 min and washed once with Milli-Q water. Cell concentration
was determined by dry weight. Culture broth (1–5 ml) was filtered
and the filter (0.45 lm membrane filter, Pall) was washed with
3 volumes of distilled water and dried in a microwave oven to
constant weight (7 min, 350 W). Dry weight measurements were
carried out in duplicate.
Synthesis of (−)-3 and (+)-4 by reduction with the genetically
engineered yeast TMB4100
Following general description 2, ( )-1 (5.00 g, 36.7 mmol) was
reduced with TMB4100. The reduction was performed in semi
fed-batch mode with a starting concentration of 20 g l−1 of ( )-1
(125 ml scale). Another batch of substrate together with 60 g l−1 of
glucose was added when the reaction had slowed down and most
diketone was reduced. The residue was purified as described for
the baker’s yeast reduction above to give recovered (+)-1 (0.49 g,
10%) of 98% ee (HPLC, Chiralcel OD–H) as a white solid (mp 88–
96 ◦C; [a]D +1252 (c 0.55, CHCl3)), (+)-4 (1.60 g, 32%) of >99%
ee (GC, a-DEX 120) as a white solid (mp 149–152 ◦C; [a]D +542
(c 0.52, tBuOMe)) and (−)-3 (1.91 g, 38%) of 94% ee (GC, a-DEX
120) as a white solid (mp 163–165 ◦C; [a]D −554 (c 0.41, CHCl3)).
1H NMR data were consistent with those reported for the baker’s
yeast reduction above.
( )-Bicyclo[2.2.2]oct-7-ene-2,5-dione (1). Compound 1 was
prepared by modification of the literature procedure12 as follows. A
solution of ( )-5,7-dioxobicyclo[2.2.2]octan-2,3-dicarboxylic acid
(18.5 g, 81.8 mmol) in pyridine (150 ml), water (150 ml) and
triethylamine (25 ml) was electrolyzed using Pt electrodes at 45 mV
(0.8 A) in an open water-cooled (5–10 ◦C) vessel for 6 h. The
reaction mixture was concentrated at reduced pressure to give a
black tar where after aqueous sat. NaCl (100 ml) was added. The
resulting mixture was acidified to pH 1 with conc. HCl and worked
up as follows: extraction with EtOAc (3 × 100 ml) followed by
drying, filtration and removal of the solvent at reduced pressure.
The residue was column chromatographed (SiO2, heptane–EtOAc
2 : 1) and the fractions containing 1 (Rf 0.46) were concentrated
to give impure 1 as a yellow oil. Recrystallization from isopropyl
ether g◦ave 1 (4.70 g, 42%) as white needles: mp 95–98 ◦C (lit.10: mp
95–99 C). 1H NMR data were consistent with those reported.10
Synthesis of (+)-5 and (+)-6 by baker’s yeast reduction
Following general description 1, 2 (2.50 g, 18.1 mmol) was
reduced with baker’s yeast. The residue was purified by column
chromatography (SiO2, heptane–EtOAc 1 : 1, TLC Rf 0.15) to
give an inseparable mixture of the diastereomers (+)-6 and (+)-5
(2.19 g, (86%) as a white solid; (mp 235–238 ◦C; [a]D +5.99 (c 1.84,
EtOH)).
Separation of the diastereoisomers (+)-5 and (+)-6 via the acetals
(+)-9a and (+)-9b
Ethylene glycol (12.0 ml, 214 mmol) and PTS (140 mg, 810 lmol)
was added to a mixture of (+)-5 and (+)-6 (2.00 g, 14.3 mmol)
in benzene (50 ml) under an argon atmosphere in a Dean–Stark
apparatus. Water removal was performed at reflux temperature
over night and then cooled to ambient temperature where after
diethyl ether (150 ml) was added. The resulting solution was
washed sequentially with sat NaHCO3 (2 × 30 ml) and water
(30 ml) followed by drying, filtration and removal of the solvent
at reduced pressure to give a diastereomeric mixture of hydroxy
acetals 8a and 8b (2.20 g, 84%) as a transparent oil; TLC Rf 0.25,
which was used directly in the next step. An analytical sample
was column chromatographed (pentane–diethyl ether 3 : 2) and
a pure fraction of the hydroxy acetal (+)-8a was obtained as a
transparent oil.
Synthesis of (−)-3 and (+)-4 by baker’s yeast reduction
Following the general description 1, ( )-1 (3.00 g, 22.0 mmol) was
reduced with baker’s yeast. The residue was purified by column
chromatography (SiO2, heptane–EtOAc 4 : 1 until (+)-1 eluted
and then 1 : 1, TLC Rf 0.46 for (+)-1, 0.21 for (+)-4 and 0.11 for
(−)-3) to give recovered (+)-1 (1.19 g, 40%) of 94% ee (HPLC,
◦
Chiralcel OD–H) as a white solid (mp 97–100 C; [a]D +1116 (c
0.22, CHCl3)) (lit.12: mp 98–100 ◦C; [a]D +1005 (c 0.107, CHCl3)).
(1R,4R,5S)-5-Hydroxybicyclo[2.2.2]oct-7-en-2-one
((+)-4).
Hydroxyketone (+)-4 (0.38 g, 12%) of 80% ee (GC, a-DEX 120)
(1R,2S,4R)-[Spiro[bicyclo[2.2.2]octan-5,2ꢀ-[1,3]dioxolane]]-2-ol
((+)-8a). [a]D +11.8 (c 2.99, tBuOMe). IR (NaCl) 3387 cm−1.
1H NMR (C6D6) d 4.04 (1 H, m), 3.63 (2 H, m), 3.50 (2 H, m),
2.46 (1 H, m), 2.20–1.95 (4 H, m), 1.89 (1 H, dt, J 14.3 and 2.8),
1.76 (1 H, q, J 3.1), 1.70–1.61 (2 H, m), 1.55–1.32 (2 H, m). 13C
NMR (C6D6) d 110.9, 68.5, 64.3, 64.3, 39.7, 34.9, 33.7, 33.6, 21.9,
18.1. HRMS (FAB+, direct inlet) [M + H] calcd for C10H16O3:
184.1099; found 184.1111. (Found: C, 65.03; H, 8.76%. C10H16O3
requires C, 65.19; H, 8.75%).
was obtained as a pale yellow semi-solid; ([a]D +397 (c 1.00,
1
tBuOMe)). IR (KBr) 3426, 1715 cm−1. H NMR (CDCl3) d 6.39
(1 H, t, J 7.3), 6.19 (1 H, t, J 7.1), 4.10 (1 H, m), 3.07 (1 H, m),
2.96 (1 H, m), 2.65 (1 H, dd, J = 18.5 and 1.3), 2.14 (2 H, m),
1.94 (1 H, dd, J 18.5 and 1.8), 1.51 (1 H, dt, J = 14.0 and 2.5). 13
C
NMR (CDCl3) d 212.5, 135.3, 129.8, 67.2, 49.0, 40.5, 33.3, 32.7.
HRMS (FAB+, direct inlet) [M + H] calcd for C8H11O2: 139.0759;
found 139.0743. (Found: C, 69.68; H, 7.24%. C8H10O2 requires C,
69.54; H, 7.30%).
Acetic anhydride (3.50 ml, 36.9 mmol) was added to a solution
of the crude mixture of 8a and 8b (1.70 g, 9.23 mmol) in pyridine
(7.8 ml) under an argon atmosphere. The resulting solution was
kept at room temperature for 24 h and then filtered through a
10 cm plug of silica (heptane–EtOAc 9 : 1) for removal of excess
pyridine. The solvent was removed at reduced pressure and the
(1S,4S,5S)-5-Hydroxybicyclo[2.2.2]oct-7-en-2-one ((−)-3). Hy-
droxyketone (−)-3 (1.37 g, 45%) of 80% ee (GC, a-DEX 120)
was obtained as a white solid; (mp 149–158 ◦C; [a]D −463 (c 0.27,
CHCl3)) (lit.12: 140–157 ◦C; [a]D −528 (c 0.123, CHCl3)). 1H NMR
data were consistent with those reported.12
2310 | Org. Biomol. Chem., 2006, 4, 2304–2312
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
residue was column chromatographed (SiO2, heptane–EtOAc 87 :
13) to give (+)-9a (1.01 g, 50%) as an oil (TLC Rf 0.63) and (+)-9b
(0.94 g, 47%) as an oil (TLC Rf 0.57).
batch mode at a concentration of 20 g l−1 of ( )-2 (5 ml scale). The
residue was purified as described for the baker’s yeast reduction
above to give an inseparable mixture of the diastereomers (+)-5
and (+)-6 (73 mg, (72%), both of >99% ee (GC, a-DEX 120), as a
white solid; (mp 243–244 ◦C; [a]D +8.55 (c 1.45, EtOH)).
(1R,2S,4R)-[Spiro[bicyclo[2.2.2]octan-5,2ꢀ-[1,3]dioxolane]]-2-yl
acetate ((+)-9a). [a]D +9.1 (c 0.55, tBuOMe). IR (KBr) 2949,
2874, 1734, 1367, 1248, 1121, 1020 cm−1. 1H NMR (C6D6) d 5.10
(1 H, m), 3.57–3.32 (4 H, m), 2.49 (1 H, m), 2.00–1.74 (5 H, m),
1.69 (3 H, s), 1.54 (1 H, p, J 3.1), 1.52–1.42 (1 H, m), 1.35 (2 H,
m). 13C NMR (C6D6) d 170.1, 110.4, 72.0, 64.4, 64.3, 39.4, 33.4,
31.6, 31.0, 21.7, 21.2, 18.8. HRMS (FAB+, direct inlet) [M + H]
calcd for C12H19O4: 227.1283; found 227.1276. (Found: C, 63.78;
H, 7.94%. C12H18O4 requires C, 63.70; H, 8.02%).
Synthesis of (+)-6 by baker’s yeast reduction
Following the general description 1, (+)-2 (1.23 g, 8.91 mmol) was
reduced with baker’s yeast. Additional yeast (12.5 g) and sucrose
(14 g) was added every 48 h. The residue was purified by column
chromatography (SiO2, heptane–EtOAc 1 : 1, TLC Rf 0.14) to give
(+)-6 (807 mg, (65%) of >99% ee (GC, a-DEX 120) as a white
solid; (mp 151 ◦C (subl.); [a]D +13.1 (c 3.57, EtOH).
(1S,2S,4S)-[Spiro[bicyclo[2.2.2]octan-5,2ꢀ-[1,3]dioxolane]]-2-yl
acetate ((+)-9b). [a]D +17 (c 0.90, tBuOMe). IR (KBr) 2939,
1
2872, 1732, 1371, 1242, 1132, 1024 cm−1. H NMR (400 MHz,
Synthesis of (+)-6 by reduction with the genetically engineered
yeast TMB4100
C6D6) d 4.86 (1 H, m), 3.51 (2 H, m), 3.37 (2 H, m), 2.39 (1 H,
m), 1.98 (2 H, m), 1.88 (1 H, m), 1.81 (1 H, m), 7.73 (1 H, ddd,
J 14.3, 3.3 and 1.6), 1.67 (3 H, s), 1.53 (1 H, p, J 3.0), 1.48
(1 H, m), 1.26 (1 H, m), 1.05 (1 H, m). 13C NMR (C6D6) d 170.3,
110.5, 71.6, 64.4, 64.2, 35.8, 33.7, 32.0, 31.6, 22.6, 21.2, 20.8.
HRMS (FAB+, direct inlet) [M + H] calcd for C12H1942: 227.1283;
found 227.1286. (Found: C, 63.49; H, 7.88%. C12H18O4 requires
C, 63.70; H, 8.02%).
Following the general description 2, (+)-2 (1.80 g, 13.0 mmol) was
reduced with TMB4100. The reduction was performed in batch
mode at a concentration of 20 g l−1 of (+)-2 (90 ml scale).
The residue was purified as described for the baker’s yeast
reduction above to give (+)-6 (1.56 g, 85%) of >99% ee (GC, a-
DEX 120) as a white solid (mp 151 ◦C (subl.); [a]D +13.6 (c 2.52,
EtOH)) and the endo,endo-bicyclo[2.2.2]octane-2,5-diol (87 mg,
5%) (as determined by comparison to an authentic sample by 1H
and 13C NMR9).
Deprotection of (+)-9a and (+)-9b to give (+)-5 and (+)-6.
K2CO3 (2.26 g) was added to a mixture of (+)-9a (1.13 g, 5.00
mmol) (or (+)-9b (1.00 g, 4.42 mmol)) in MeOH (29 ml) and water
(11 ml). The resulting mixture was stirred at room temperature for
24 h where after the MeOH was removed at reduced pressure.
The mixture was concentrated to approximately 29 ml followed
by acidification with 2 M HCl where after acetone (29 ml) was
added. The reaction mixture was stirred at room temperature
for 2 h before it was concentrated to approximately 35 ml and
then extracted with EtOAc (5 × 50 ml). Drying of the organic
extracts, filtration and removal of the solvent at reduced pressure
was followed by column chromatography (SiO2, heptane–EtOAc
1 : 2).
Synthesis of (−)-1, (+)-1, (−)-2 and (+)-2 by TPAP oxidation
˚
N-Methylmorpholine-N-oxide (51 mg, 0.43 mmol), 3 A crushed
molecular sieves (150 mg) and tetrapropylammonium perruthen-
ate (3.8 mg, 5 mol%) were added to a solution of the hydrox-
yketones (3–6) (0.22 mmol) in CH2Cl2 (5 ml) under an argon
atmosphere. The resulting mixture was stirred at room temperature
for 2 h and then filtered through a pad of silica (bottom layer) and
celite (top layer), rinsing with EtOAc. The solvent was removed
at reduced pressure and the residue was column chromatographed
(SiO2, heptane–EtOAc 1 : 1) to give (−)-1, (+)-1, (−)-2 and (+)-2 in
91–100% yield. For (−)-1 (from (−)-3 of 94% ee): mp 101–103 ◦C;
[a]D −1222 (c 0.55, CHCl3), (+)-1 (from (+)-4 of >99% ee): mp
89–94 ◦C; [a]D +1256 (c 0.55, CHCl3), (−)-2 (from (+)-5 of 88%
ee): mp 208–213 ◦C (melt and subl.); [a]D −48 (c 0.42, CHCl3) and
(+)-2 (from (+)-6 of >99% ee): mp 212–213 ◦C (melt and subl.);
[a]D +50 (c 0.44, CHCl3), respectively.
(1R,4R,5S)-5-Hydroxybicyclo[2.2.2]octan-2-one ((+)-5). Hy-
droxyketone (+)-5 of >99% ee (GC, a-DEX 120) was obtained
as a white solid in 92% yield. TLC Rf 0.14; mp 241–242 ◦C (melt
and subl.); [a]D +4.15 (c 2.73, EtOH). IR (KBr) 3426, 1727 cm−1.
1H NMR (CDCl3) d 4.11 (1 H, m), 2.40–2.10 (6 H, m), 2.06 (1 H,
br s), 1.93 (1 H, m), 1.78 (1 H, m), 1.62 (1 H, m), 1.49 (1 H, m).
13C NMR (CDCl3) d 216.6, 68.0, 42.6, 41,7, 35.6, 33.9, 23.1, 17.5.
HRMS (FAB+, direct inlet) [M + H] calcd for C8H13O2: 141.0916;
found 141.0917. (Found: C, 68.42; H, 8.70%. C8H12O2 requires C,
68.54; H, 8.63%).
Synthesis of (+)-5 and (+)-6 by catalytic hydrogenation
The hydrogenation catalyst, 10 wt% Pd/C (327 mg) was added
to a mixture of (−)-3 (100 mg, 0.72 mmol) (or (+)-4 (300 mg,
2.17 mmol)) in ethanol (10 ml for (−)-3) or 30 ml for (+)-4) at room
temperature. The resulting slurry was stirred under a hydrogen
atmosphere at room temperature at atmospheric pressure for 3 h
and then filtered through a pad of silica (bottom layer) and celite
(top layer). The solvent was removed at reduced pressure and
the residue purified by column chromatography (SiO2, heptane–
EtOAc 1 : 1) to give (+)-5 (or (+)-6) in 99% yield. For (+)-5 (from
(−)-3 of 80% ee): mp 238–239 ◦C; [a]D +1.45 (c 1.45, EtOH) and
(1S,4S,5S)-5-Hydroxybicyclo[2.2.2]octan-2-one ((+)-6). Hy-
droxyketone (+)-6 of >99% ee (GC, a-DEX 120) was obtained
as a white solid in 96% yield. TLC Rf 0.14; mp 151 ◦C (subl.); [a]D
1
+13.1 (c 2.70, EtOH). H NMR data were consistent with those
reported.9
Synthesis of (+)-5 and (+)-6 by reduction with the genetically
engineered yeast TMB4100
◦
Following the general description 2, ( )-2 (100 mg, 0.72 mmol)
was reduced with TMB4100. The reduction was performed in
(+)-6 (from (+)-4 of >99% ee): mp 151 C (subl.); [a]D +13.6 (c
2.61, EtOH).
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2304–2312 | 2311
©

View Article Online
11 T. Kinoshita, K. Haga, K. Ikai, K. Takeuchi and K. Okamoto,
Tetrahedron Lett., 1990, 31, 4057–4060.
Acknowledgements
12 D. A. Lightner, L. A. Paquette, P. Chayangkoon, H. S. Lin and J. R.
Peterson, J. Org. Chem., 1988, 53, 1969–1973.
13 W. Sybesma, A. Straathof, J. Jongejan, J. Pronk and J. Heijnen, Biocatal.
Biotransform., 1998, 16, 95–134.
The work was financially supported by grants from the Swedish
Foundation for Strategic Research, the Swedish Research Council,
the Royal Physiographic Society in Lund, the Crafoord Founda-
tion, the Knut and Alice Wallenberg Foundation and the Research
14 S. Servi, Synthesis, 1990, 1–25.
15 T. Johanson, M. Katz and M. F. Gorwa-Grauslund, FEMS Yeast Res.,
2005, 5, 513–525.
16 M. Katz, I. Sarvary, T. Frejd, B. Hahn-Hagerdal and M. F.
Gorwa-Grauslund, Appl. Microbiol. Biotechnol., 2002, 59, 641–
648.
17 M. Katz, T. Frejd, B. Hahn-Hagerdal and M. F. Gorwa-Grauslund,
Biotechnol. Bioeng., 2003, 84, 573–582.
¨
School in Medicinal Science at Lund University (FLAK). We
thank Jermina Naganjac for helping with the construction of
TMB4097, and Nora Bieler for assisting with the preparative scale
reductions of the diketones 1 and 2.
18 A. L. Botes, D. Harvig, M. S. van Dyk, I. Sarvary, T. Frejd, M. Katz, B.
Hahn-Haegerdal and M. F. Gorwa-Grauslund, J. Chem. Soc., Perkin
Trans. 1, 2002, 1111–1114.
19 M. Katz, T. Johanson and M. F. Gorwa-Grauslund, Yeast, 2004, 21,
1253–1267.
References
1 P. Yates, D. J. Burnell, V. J. Freer and J. F. Sawyer, Can. J. Chem., 1987,
65, 69–77.
2 J. T. Njardarson and J. L. Wood, Org. Lett., 2001, 3, 2431–2434.
3 L. A. Paquette and H.-C. Tsui, J. Org. Chem., 1996, 61, 142–145.
4 S. Moore, dissertation, Georgia Institute of Technology, 2004.
5 D. Peters, G. M. Olsen, E. Nielsen, P. K. Ahring and T. Jo¨rgensen, WO
Pat., 2002-DK3472002096911, 2002.
20 D. Davalian, P. J. Garratt and R. Riguera, J. Org. Chem., 1977, 42,
368–369.
21 H. Ahlbrecht, M. Dietz, C. Schoen and V. Baumann, Synthesis, 1991,
133–140.
6 R. Shintani, K. Okamoto, Y. Otomaru, K. Ueyama and T. Hayashi,
J. Am. Chem. Soc., 2005, 127, 54–55.
22 N. H. Werstiuk, S. Yeroushalmi and G. L. Hong, Can. J. Chem., 1992,
70, 974–980.
7 N. Tokunaga, Y. Otomaru, K. Okamoto, K. Ueyama, R. Shintani and
T. Hayashi, J. Am. Chem. Soc., 2004, 126, 13584–13585.
8 Y. Otomaru, K. Okamoto, R. Shintani and T. Hayashi, J. Org. Chem.,
2005, 70, 2503–2508.
23 M. Katz, B. Hahn-Ha¨gerdal and M. Gorwa-Grauslund, Enzyme
Microb. Technol., 2003, 33, 163–172.
24 H. Inoue, H. Nojima and H. Okayama, Gene, 1990, 96, 23–
28.
9 F. Almqvist, N. Ekman and T. Frejd, J. Org. Chem., 1996, 61, 3794–
25 R. D. Gietz and A. Sugino, Gene, 1988, 74, 527–534.
26 F. M. Ausabel, B. R. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith
and K. Struhl, Current Protocols in Molecular Biology, New York, John
Wiley & Sons, 1995.
3798.
10 R. K. Hill, G. H. Morton, J. R. Peterson, J. A. Walsh and L. A. Paquette,
J. Org. Chem., 1985, 50, 5528–5533.
2312 | Org. Biomol. Chem., 2006, 4, 2304–2312
This journal is
The Royal Society of Chemistry 2006
©