Baker's yeast: Improving the D-stereoselectivity in reduction of 3-oxo esters

Article abstract of DOI:10.1016/S0957-4166(99)00025-7

The stereoselectivity of baker's yeast in the reduction of ethyl 3- oxopentanoate was shifted towards the corresponding (R)-hydroxy ester by sugar, heat treatment and allyl alcohol. The highest enantiomeric excesses obtained with baker's yeast with a good reduction capacity, 92-97%, were achieved by combining allyl alcohol and sugar; heat treatment did not increase the stereoselectivity further. With the use of this technique, ethyl (R)-3-hydroxyhexanoate, >99% ee, and ethyl (S)-4-chloro-3-hydroxybutanoate, 82-90% ee, were produced from the corresponding esters, and for the first time an excess of the (R)-enantiomer of ethyl 3-hydroxybutanoate was obtained with ordinary baker's yeast.

Full text of DOI:10.1016/S0957-4166(99)00025-7

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 10 (1999) 551–559
Baker’s yeast: improving the D-stereoselectivity in reduction of
3-oxo esters
Allan C. Dahl,† Morten Fjeldberg and Jørgen Øgaard Madsen ∗
Department of Organic Chemistry, Technical University of Denmark, DK-2800 Lyngby, Denmark
Received 21 December 1998; accepted 19 January 1999
Abstract
The stereoselectivity of baker’s yeast in the reduction of ethyl 3-oxopentanoate was shifted towards the
corresponding (R)-hydroxy ester by sugar, heat treatment and allyl alcohol. The highest enantiomeric excesses
obtained with baker’s yeast with a good reduction capacity, 92–97%, were achieved by combining allyl alcohol
and sugar; heat treatment did not increase the stereoselectivity further. With the use of this technique, ethyl (R)-3-
hydroxyhexanoate, >99% ee, and ethyl (S)-4-chloro-3-hydroxybutanoate, 82–90% ee, were produced from the
corresponding esters, and for the first time an excess of the (R)-enantiomer of ethyl 3-hydroxybutanoate was
obtained with ordinary baker’s yeast. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
Baker’s yeast (Saccharomyces cerevisiae) is a useful ‘reagent’ for the reduction of prochiral 3-oxo
esters to the corresponding chiral 3-hydroxy esters. It is cheap, easy to handle, harmless to man, and it
accepts a wide range of substrates. A frequent problem in the use of baker’s yeast is an unsatisfactory
degree of stereoselectivity. This may be caused by the simultaneous operation of two or more enzymes
with opposite stereoselectivities, since isolated baker’s yeast enzymes have been shown to be highly
stereoselective in the reductions of some 3-oxo esters.1–3 Controlling stereoselectivity therefore becomes
a question of controlling the activity of the individual enzymes.
The physiological state of the yeast, fermenting or resting, and substrate concentration are important
parameters.4 Thus, from ethyl 3-oxobutanoate, -pentanoate and -hexanoate higher proportions of (R)-
products are obtained with fermenting yeast and high substrate concentration than with resting yeast
and low substrate concentration. Both sets of conditions, however, give quite a high excess of ethyl (S)-
3-hydroxybutanoate and ethyl (R)-3-hydroxyhexanoate. From ethyl 3-oxopentanoate, on the other hand,
∗
†
Corresponding author. E-mail: okjom@pop.dtu.dk
Part of PhD thesis by A.C.D.
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00025-7

552
A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
approximately 50% ee of (R)- and approximately 50% ee of (S)-product may be obtained with fermenting
or resting yeast, respectively, as shown in this study. So this ester may be useful for testing additional
means of improving stereoselectivity.
The use of inhibitors has been successful in some cases. Thus, from methyl 3-oxopentanoate Nakamura
and coworkers5,6 obtained the (R)-product in 96% ee by incubating the yeast suspension with 1.0 g l−1
of allyl alcohol for 30 minutes prior to addition of the oxo ester and glucose, compared to an ee of 59%
without the inhibitor. Under similar conditions, using the ethyl ester, Ushio and coworkers7 prepared the
(S)-alcohol with an ee of 98% or better with allyl bromide as inhibitor at a concentration of 4–5 g l−1.
In a more recent work, heat treatment (50°C) of baker’s yeast was introduced as a means of controlling
stereoselectivity.8 The method, which obviously rests on selective thermal denaturation of oxidoreduc-
tases, has been successfully used in the reduction of 3-oxo ester substituted in the 2-position8 and 2,4-
alkanediones.9
In a previous study we have shown that the stereochemical outcome of the reduction of 3-oxo esters
is also influenced greatly by the growth conditions of the yeast and the use of co-substrates other than
sucrose and glucose:10 with baker’s yeast harvested while growing under anaerobic conditions ethyl (R)-
3-hydroxypentanoate was obtained in 95% ee in the presence of sucrose; the (S)-enantiomer was obtained
in 81–93% ee by slowly adding the oxo ester to ordinary baker’s yeast stirred with gluconolactone.
With the same techniques, ethyl (S)-4-chloro-3-hydroxybutanoate in 98% ee and the (R)-enantiomer
in 89–94% ee were obtained, respectively. An enantiomeric excess of ethyl (S)-3-hydroxybutanoate of
99.2% was realized with ordinary baker’s yeast and gluconolactone as a co-substrate, but it was not
possible to obtain an excess of the other enantiomer with anaerobically grown baker’s yeast.
In the present work we have performed experiments with heat and allyl alcohol treatment of ordinary
baker’s yeast to improve the D-stereoselectivity in reductions of 3-oxo esters. The heat treatment
method has not been used previously with unsubstituted 3-oxo esters. The use of allyl alcohol has been
investigated with such substrates. We wanted, however, to test these methods in our laboratory in order
to directly compare the results with the findings of our previous study.10
2. Results and discussion
The key experiments, involving heat treatment and allyl alcohol inhibitors, were performed with ethyl
3-oxopentanoate as a substrate. As reference experiments, this ester was reduced with resting and with
fermenting baker’s yeast under the basic conditions used in this study: baker’s yeast (Danisco Distillers,
Danisco A/S, Denmark) from a local shop, 25 g, was suspended in tap water, 125 ml, at room temperature
and ethyl 3-oxopentanoate, 0.10 ml, was added. In experiments with fermenting yeast sucrose, 25 g, was
added and the mixture was stirred for at least 0.5 hour prior to addition of oxo ester to ensure that
fermentation was in progress. After stirring for approximately 24 hours a sample was withdrawn for
chiral GC as described previously.10 All of the oxo ester had been reduced in the reference experiments,
and ethyl (S)-3-hydroxypentanoate in 50% ee and the (R)-enantiomer in 33% ee was produced with
resting and fermenting baker’s yeast (preincubated 1 hour), respectively (entries 1 and 2 in Table 1).
To study the effect of heat treatment on the stereoselectivity, baker’s yeast was suspended in water and
heated to different temperatures for 30 minutes. The mixture was cooled to room temperature and ethyl 3-
oxopentanoate was added. After 20–23 hours the degree of conversion and the enantiomeric composition
of the product was analysed by GC (Fig. 1). Heat treatment changed the stereochemical course of the
reduction with resting yeast from mainly (S)- to mainly (R)-hydroxy ester in an enantiomeric excess
as high as 64–69% when the yeast had been heated to 48–51°C. Yeast heated to 54°C again afforded

A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
553
Table 1
Baker’s yeast, 25 g, was suspended in tap water, 125 ml. If heat treated, the temperature and time are
shown. After cooling to room temperature allyl alcohol and/or sugar were added (at the same time) and
the yeast was preincubated for the indicated period. Ethyl 3-oxopentanoate was added and a sample
was withdrawn for analysis by chiral GC after the indicated reduction time. Duplicates are separated
by a slash
aUnless otherwise stated.
bGC analysis showed a strongly reduced amount of product.
mainly the (S)-hydroxy ester. In the experiments where the yeast had been heated to temperatures above
48°C the reduction was incomplete. The time of treatment at 48°C also affected the stereoselectivity
(Fig. 2) as well as the degree of conversion; the enantiomeric excess of the product reached a level of
82–89% at 90–180 minutes of heat treatment but if the time exceeded 110 minutes the reduction did not
go to completion. As a compromise between enantiomeric excess of product and enzymic activity heat
treatment at 48°C for 60 minutes was used as standard.
In some heat treatment experiments the enantiomeric excess of the produced hydroxy ester was
followed during the reduction and was found to decrease. It was suspected that thermally denatured
enzymes might undergo renaturation when the temperature had been lowered.11 In order to see if this was
the case, some experiments were run, where the yeast was treated at 48°C for 60 minutes and then stirred
at room temperature for periods of 2, 5 or 21 hours before addition of ethyl 3-oxopentanoate. Products
of (R)-configuration with an enantiomeric excess of 74, 69, and 37%, respectively, were obtained at full
conversion, compared with values of 79 and 76 in the standard experiments (entry 3 in Table 1). Thus,
renaturation appears to take place to some extent. In an attempt to avoid this, an experiment was run in

554
A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
Figure 1. Enantiomeric excess of ethyl 3-hydroxypentanoate produced by baker’s yeast heat treated at different temperatures.
Baker’s yeast, 25 g, was suspended in tap water, 125 ml, and heated to different temperatures for 30 minutes. After cooling
to room temperature ethyl 3-oxopentanoate, 0.10 ml, was added and the mixture was stirred for 20–23 hours. A sample was
extracted for chiral GC
Figure 2. Enantiomeric excess of ethyl 3-hydroxypentanoate produced by baker’s yeast heat treated for different time periods.
Baker’s yeast, 25 g, suspended in tap water, 125 ml, was heated to 48°C for different time intervals and cooled to room
temperature. Ethyl 3-oxopentanoate, 0.10 ml, was added and 20–24 hours later a sample was withdrawn for analysis by chiral
GC
which the yeast suspension was kept at 41°C after the heat treatment, and here an enantiomeric excess of
85% was obtained versus 76 or 79% (Table 1, entry 3). Renaturation seems to be prevented at an elevated
temperature, but we realize that a higher rate of reduction may explain the improved result. Since the
effect is small if the oxo ester is added immediately after heat treatment, experiments were performed at
room temperature as standard.

A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
555
Since fermentation greatly favours production of (R)-hydroxy esters, as seen from the reference
experiments (Table 1, entries 1 and 2), we examined whether the addition of sugar to the heat treated yeast
prior to addition of the oxo ester could increase the stereoselectivity. Sugar was observed to have no effect
(compare entries 3 and 4 in Table 1). The complete absence of an effect of sugar after heat treatment is
very likely to be caused by thermal denaturation of enzymes in the glycolysis system. At this point it may
be concluded that the enzymes of baker’s yeast that are capable of reducing ethyl 3-oxopentanoate cannot
be thermally denatured so selectively that heat treated yeast can produce ethyl 3-hydroxypentanoate in a
fully satisfactory enantiomeric excess (we believe most workers want an enantiomeric excess of at least
90%).
We turned to investigating the use of allyl alcohol as inhibitor. As mentioned above, the use of this
inhibitor in baker’s yeast reductions of 3-oxo esters was introduced by Nakamura and coworkers.5 In
their work, baker’s yeast was stirred with allyl alcohol, typically 1 g l−1, for 30 minutes at 30°C. Then
substrate and glucose were added. Initially, we performed a series of experiments on the reduction of
ethyl 3-oxopentanoate with resting yeast that was preincubated with varying amounts of allyl alcohol
for different times. As shown in Fig. 3 the (R)-selectivity increased with an increasing amount of allyl
alcohol. With short preincubation times, 0.5–2 hours, a maximum of 12–43% ee of the (R)-product
(interval includes all values), was obtained with 0.4–0.5 ml of the inhibitor. It is seen also that there is
no clear effect of the preincubation time in the range 0.5–3 hours. Longer preincubation times, 1–2 days,
led to a considerable rise in selectivity, again with a clear dependence on the amount of inhibitor. At 0.3
ml of allyl alcohol an enantiomeric excess of ≥89% was reached. There was no significant improvement
of selectivity by applying a longer preincubation time than about 1 day.
Since allyl alcohol shifts the stereoselectivity in the same direction as heat treatment, it was obvious
to test a possible additive effect of these techniques. In a number of experiments baker’s yeast was heat
treated at 48°C for 1 hour, the mixture was cooled to room temperature and a varying amount of allyl
alcohol was added. After preincubation for 0.5 or 1 hour, ethyl 3-oxopentanoate was added (Table 1,
entries 5–9). In this way the enantiomeric excess of the product became as high as 98%; unfortunately,
the reduction did not go to completion under these conditions, except for the experiment where only 0.15
ml allyl alcohol was used (Table 1, entry 5). When sugar was added all the oxo ester was reduced and the
enantiomeric excess was as high as 95 and 96% in two experiments (Table 1, entries 10 and 11), but, as
shown below, there is actually no need for heat treatment when allyl alcohol is used in combination with
sugar.
Even though the combination of heat treatment and addition of sugar and allyl alcohol appears to be
quite good for preparing (R)-hydroxy esters, a method without the time-consuming step of heat treatment
of the yeast would be preferred. Therefore a series of experiments was run where baker’s yeast was
preincubated with sugar and varying amounts of allyl alcohol for different times before addition of ethyl
3-oxopentanoate (Fig. 4). Two important observations were made as compared with the results from
the experiments with resting yeast (Fig. 3). Firstly, full conversion to product of very high enantiomeric
excess was obtained, and secondly, the full effect of allyl alcohol was realized in a very short time.
A rather constant level of 88–93% ee (interval includes all values) with 0.15–0.5 ml allyl alcohol was
reached with a preincubation time of 0.5 hour. Increasing the preincubation time to 1 hour raised the ee
values to a level of 92–97% (interval includes all values) with 0.2–0.5 ml allyl alcohol. The enantiomeric
excess was now as high as with sugar, allyl alcohol and heat treatment (Table 1, entry 10 and 11).
Preincubation for more than 1 hour did not raise the enantiomeric excess further. Since the additive
effect of sugar and allyl alcohol is very high it is tempting to suggest that allyl alcohol interferes with
regeneration of cofactors in the cell. It has been reported that the allyl alcohol inhibits yeast alcohol
dehydrogenase;12 this enzyme is important in the fermentation of sugar since it converts NADH produced

556
A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
Figure 3. Influence of preincubation time and amount of allyl alcohol on the ee of ethyl 3-hydroxypentanoate produced by
resting baker’s yeast. Baker’s yeast, 25 g, was stirred with allyl alcohol, 0–0.5 ml, in tap water, 125 ml, for 0.5–48 hours. Ethyl
3-oxopentanoate, 0.10 ml, was added and stirring was continued for 17–25 hours. The mixture was analysed by chiral GC. The
points represent an average of one to four experiments
in the glycolysis to NAD+ by reducing acetaldehyde to ethanol.13 If NAD+ is not regenerated in this way,
the glycolysis will stop, and we did indeed observe a reduction in evolution of CO2 with increasing
amounts of allyl alcohol.
The stereoselectivity in reductions with fermenting yeast (Fig. 4) are very sensitive to the preincubation
time in the absence of allyl alcohol; on going from 0.5 to 2 hours of preincubation the ee of the
product dropped from 50 to 13% (both values are averages of two experiments). This could be due
to the considerable decrease in the pH that is caused by the fermentation process. This assumption was
supported by some experiments in which the pH was kept at 5 with NaOH (pH-stat). Thus, without
allyl alcohol an enantiomeric excess of 46% was obtained after a preincubation time of 1 hour (Fig. 5),
versus 33% without control of the pH (Fig. 4). With 0.2 ml of allyl alcohol added, on the other hand,
a hardly significant increase in ee was observed: 95 and 96% ee in two experiments at pH 5, versus
92 and 95% ee in two experiments without control of the pH. The independence of stereoselectivity on
preincubation time in experiments with allyl alcohol added may be explained by a strong inhibitory effect
of this alcohol, which anchors the stereoselectivity at a high level, combined with the above mentioned
inhibition of the fermentation process, which reduces the decrease of pH.
The work with ethyl 3-oxopentanoate as substrate was concluded by some experiments (Table 1,

A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
557
Figure 4. Influence of preincubation time and amount of allyl alcohol on the ee of ethyl 3-hydroxypentanoate produced by
fermenting baker’s yeast. Baker’s yeast, 25 g, was preincubated with sugar, 25 g, and allyl alcohol, 0–0.5 ml, in tap water, 125
ml, for 0.5–24 hours before ethyl 3-oxopentanoate, 0.10 ml, was added. The mixture was stirred for 17–24 hours and analysed
by chiral GC. Each point represents an average of one to three experiments
entries 12 and 13) that demonstrate a good reduction capacity coupled with a high stereoselectivity of
fermenting yeast treated with allyl alcohol. Reduction of 1 ml of the oxo ester gave products of 91% ee
without control of the pH; at pH 5, slightly higher values, 92 and 93%, were obtained.
Finally, some other esters that were used as substrates in our foregoing study on reductions with
anaerobically growing or grown baker’s yeast,10 namely ethyl 3-oxobutanoate and -hexanoate and ethyl
4-chloro-3-oxobutanoate, were also subjected to reduction with baker’s yeast under the best conditions
of the present work: fermenting yeast with allyl alcohol added. However, the heat treatment methods
was also tested with ethyl 3-oxobutanoate in a single experiment (Table 1, entry 14), where the product
showed a large excess of the (S)-enantiomer. It appears that the (S)-enzyme, which is particularly active
in reduction of this oxo ester,7 is thermally more stable than the (S)-enzyme, which is responsible for
the (S)-product from ethyl 3-oxopentanoate. With the allyl alcohol technique an excess, though small,
amount of ethyl (S)-3-hydroxybutanoate was still obtained in an experiment without pH control (entry
15). With the pH adjusted to about 5 the stereoselectivity shifted to the (R)-side: in two experiments
with addition of 1 ml of substrate an ee value of 18% was obtained, but the conversion was incomplete
(Table 1, entry 16). At a pH of about 8 the enantiomeric excess of (R)-product increased to 87% (Table 1,
entry 17), but these conditions are unfortunately not useful because a high proportion of substrate and/or

558
A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
Figure 5. Effect of allyl alcohol on the stereoselectivity of fermenting baker’s yeast in reduction of ethyl 3-oxopentanoate at
pH 5. Baker’s yeast, 25 g, was preincubated at pH 5 (pH-stat) with sugar, 25 g, and allyl alcohol, 0–0.2 ml, for 1 hour. Ethyl
3-oxopentanoate, 0.10 ml, was added and stirring was continued for 19–21 hours. A sample was withdrawn for chiral GC. Each
point represents an average of two experiments
product is lost. However, these experiments gave an excess of ethyl (R)-3-hydroxybutanoate for, to the
best of our knowledge, the first time with ordinary baker’s yeast.
When fermenting baker’ yeast preincubated with allyl alcohol was used to reduce 1 ml of
ethyl 3-oxohexanoate the selectivity for the corresponding (R)-product was very high; ethyl (R)-3-
hydroxyhexanoate with >99% ee was obtained in two experiments with control of the pH (Table 1, entry
18).
The last compound used as substrate in this study, ethyl 4-chloro-3-oxobutanoate, was firstly reduced
with fermenting baker’s yeast inhibited by allyl alcohol in an experiment without control of the pH; all
the substrate, 0.10 ml, was reduced to the (S)-hydroxy ester with 87% ee (Table 1, entry 19). In three
experiments with control of pH, 1 ml of the substrate was fully converted to the D-(S)-product with
82–90% ee (Table 1, entry 20).
Our results from reduction of ethyl 3-oxobutanoate and ethyl 4-chloro-3-oxobutanoate by fermenting
baker’s yeast inhibited with allyl alcohol are similar to, or slightly better than, those of Nakamura
and coworkers,5,6 who obtained ethyl (S)-3-hydroxybutanoate in 40% ee and ethyl (S)-4-chloro-3-
hydroxybutanoate in 85% ee. They work with a higher substrate/yeast ratio and a lower sugar (glu-
cose)/yeast ratio, and, probably more importantly, they do not preincubate the yeast with allyl alcohol
together with sugar for a period prior to addition of substrate.
With respect to ethyl 3-oxobutanoate and -hexanoate, the present results with allyl alcohol and sugar
are superior to our previous results obtained with anaerobically grown or growing yeast,10 where the
configuration of produced ethyl 3-hydroxybutanoate could not be shifted from mainly (S) to mainly (R),
and the enantiomeric excess of the produced hexanoate could not be increased above what is obtained
with fermenting ordinary baker’s yeast. On the other hand, anaerobically grown or growing baker’s yeast
produces ethyl (R)-3-hydroxypentanoate of 90–96% ee10 which equals the results obtained in this work
with allyl alcohol and sugar. Ethyl (S)-4-chloro-3-hydroxybutanoate of 95 or 98% ee was obtained with

A. C. Dahl et al. / Tetrahedron: Asymmetry 10 (1999) 551–559
559
anaerobically grown baker’s yeast,10 which is superior to the results obtained with ordinary baker’s yeast
with allyl alcohol and sugar. These two techniques therefore complement each other.
References
1. Shieh, W.-R.; Gopalan, A. S.; Sih, C. J. J. Am. Chem. Soc. 1985, 107, 2993–2994.
2. Heidlas, J.; Engel, K.-H.; Tressl, R. Eur. J. Biochem. 1988, 172, 633–639.
3. Nakamura, K.; Kawai, Y.; Nakajima, N.; Ohno, A. J. Org. Chem. 1991, 56, 4778–4783.
4. Nielsen, J. K.; Madsen, J. Ø. Tetrahedron: Asymmetry 1994, 5, 403–410.
5. Nakamura, K.; Inoue, K.; Ushio, K.; Oka, S.; Ohno, A. Chem. Lett. 1987, 679–682.
6. Nakamura, K.; Kawai, Y.; Oka, S.; Ohno, A. Bull. Chem. Soc. Jpn. 1989, 62, 875–879.
7. Ushio, K.; Hada, J.; Tanaka, Y.; Ebara, K. Enzyme Microb. Technol. 1993, 15, 222–228.
8. Kawai, Y.; Kondo, S.-i.; Tsujimoto, M.; Nakamura, K.; Ohno, A. Bull. Chem. Soc. Jpn. 1994, 67, 2244–2247.
9. Cui, J.-N.; Teraoka, R.; Ema, T.; Sakai, T.; Utaka, M. Tetrahedron Lett. 1997, 38, 3021–3024.
10. Dahl, A. C.; Madsen, J. Ø. Tetrahedron: Asymmetry 1998, 9, 4395–4417.
11. Thermostability of Enzymes; Gupta, M. N., Ed. Springer–Verlag: Berlin, 1993.
12. Rando, R. R. Biochem. Pharmacol. 1974, 23, 2328–2331.
13. Stryer, L. Biochemistry; 3rd edn, W. H. Freeman: New York, 1988; p. 363.