Microbial reduction and in situ product crystallization coupled with biocatalyst cultivation during the synthesis of 6R-dihydrooxoisophorone

Article abstract of DOI:10.1002/adsc.200505024

An in situ product crystallization procedure was developed for a crystalline product formed during microbial reduction coupled with cell cultivation. The model reaction was the asymmetric reduction of 4-oxoisophorone (OIP) by baker's yeast (Saccharomyces cerevisiae). Yeast cells were cultivated fed-batch to reach a maximum concentration of 30 gdw·L^-1. The desired product, 6R-dihydrooxoisophorone (DOIP), may be further reduced by baker's yeast to an unwanted by-product; thus, DOIP was removed immediately from the fermenter via an external crystallization loop in this procedure. The OIP reduction rate was five times higher (?0.33 mmol·gdw ^-1·h^-1) as compared to the reduction rate with resting cells. OIP reduction was started when the optimum cell concentration had already been reached in the reactor because the substrate (OIP) at ≥55 mM concentration inhibited cell growth. An appropriate supply of glucose as carbon and energy source was necessary to support the coupled reactions involving cell growth and maintenance and product formation while avoiding formation of metabolic by-products. Final DOIP yield and selectivity were 85% and 99%, respectively, while over 100 g·L^-1 of product was obtained in the crystallizer. The product crystals with favorable properties were readily recovered from the crystallizer. These results indicate that product crystallization is not impaired by the solutes present in the fermentation medium.

Full text of DOI:10.1002/adsc.200505024

FULL PAPERS
Microbial Reduction and in situ Product Crystallization Coupled
with Biocatalyst Cultivation during the Synthesis of
6R-Dihydrooxoisophorone
Evelyn M. Buque-Taboada,a, b Adrie J. J. Straathof,a,* Joseph J. Heijnen,a
Luuk A. M. van der Wielena
a
Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands
Phone: (þ31)-15-278 2330, fax: (þ31)-15-278 2355, e-mail: A. J. J.Straathof@tnw.tudelft.nl
b
Department of Chemical Engineering, University of San Carlos, Talamban, Cebu City 6000, The Philippines
Received: January 19, 2005; Accepted: March 21, 2005
Abstract: An in situ product crystallization procedure tion inhibited cell growth. An appropriate supply of
was developed for a crystalline product formed during glucose as carbon and energy source was necessary
microbial reduction coupled with cell cultivation. The to support the coupled reactions involving cell growth
model reaction was the asymmetric reduction of 4-oxo- and maintenance and product formation while avoid-
isophorone (OIP) by bakerꢀs yeast (Saccharomyces ing formation of metabolic by-products. Final DOIP
cerevisiae). Yeast cells were cultivated fed-batch to yield and selectivity were 85% and 99%, respectively,
reach a maximum concentration of 30 gdw·Lꢀ1. The while over 100 g·Lꢀ1 of product was obtained in the
desired product, 6R-dihydrooxoisophorone (DOIP), crystallizer. The product crystals with favorable prop-
may be further reduced by bakerꢀs yeast to an un- erties were readily recovered from the crystallizer.
wanted by-product; thus, DOIP was removed immedi- These results indicate that product crystallization is
ately from the fermenter via an external crystalliza- not impaired by the solutes present in the fermenta-
tion loop in this procedure. The OIP reduction rate tion medium.
was five times higher (ffi0.33 mmol·gdwꢀ1 ·hꢀ1) as
compared to the reduction rate with resting cells.
OIP reduction was started when the optimum cell Keywords: crystallization; 6R-dihydrooxoisophorone;
concentration had already been reached in the reactor in situ product recovery; 4-oxoisophorone; reduction;
because the substrate (OIP) at ꢁ55 mM concentra- Saccharomyces cerevisiae
Introduction
In previous work,[2] it was shown that with resting cells
of Saccharomyces cerevisiae, an in situ product crystalli-
The range of products that can be produced by microbial zation procedure by an integrated process (Figure 1)
biotransformation is increasing as productivity im- was more efficient than the non-integrated batch and
proves dramatically by metabolic and process engineer- fed-batch configurations. Such an integrated process
ing. In the latter approach, maximizing the biocatalyst
concentration in the reactor and/or immediately remov-
ing inhibiting or degrading products during the process
can raise bioreactor productivity. This requires cell cul-
tivation during the process especially when the biocata-
lyst is not commercially or readily available and/or in-
volves in situ product removal.[1,2] However, simplifying
the product separation and purification steps still re-
mains a challenge as it is the cost-limiting factor in
most cases.[1] When the desired product crystallizes in
the reactor, it is often redissolved using organic solvents,
separated and recrystallized. If the product crystallizes
during fermentation and can be separated from the cells
directly, production costs might be reduced dramatical-
Figure 1. Schematic diagram of the integrated fermentation-
crystallization process for reduction of 1.
ly.[2,3]
Adv. Synth. Catal. 2005, 347, 1147–1154
DOI: 10.1002/adsc.200505024
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1147

Evelyn M. Buque-Taboada et al.
FULL PAPERS
can be improved further by implementing a close-down
phase when substrate feeding is stopped and bioreduc-
tion is continued to convert all the substrate in the reac-
tor. In various cases, the biocatalyst (micro-organism)
needs to be cultivated prior to or concomitant with bio-
transformation and product crystallization. It would be
Figure 2. Model reaction: reduction of 4-oxoisophorone (1)
interesting to incorporate cell growth in this in situ prod- by bakerꢀs yeast.
uct crystallization procedure. Such an integration can
possibly contribute to differences in the biotransforma-
tion and crystallization kinetics as growing cells may fol- in the reactor. Still, it is an interesting case to perform re-
low different metabolic routes compared to resting cells. ductions with pre-cultivated yeast cells. Using whole
Thus, this work aims to demonstrate experimentally cells as biocatalyst would prove economically favorable
the impact of implementing cell cultivation concomitant as this allows for an easy and in vivo cofactor regenera-
with bioreduction and in situ product crystal formation tion in the cell, which sustains catalytic activity for redox
in this integrated fermentation-crystallization process. reactions.[6] However, aeration and nutrient feeding can
This integration strategy should be generally applicable be the most important constraints to consider, which can
to processes that involve cell cultivation, biocatalytic potentially complicate the whole biocatalytic process.
formation of a product, and subsequent product crystal-
In this work, fed-batch cell cultivation was employed
lization. This approach might also be widely suitable for where nutrient (glucose) feed rates were varied to max-
processes where it is required to control and reduce the imize the utilization of the carbon source and to mini-
product concentration in the reactor in order to prevent mize formation of metabolic by-products. Fed-batch
product toxicity, inhibition and/or degradation, and in cell cultivation was also performed concomitant with
addition, to simplify the product separation and recov- batch reduction where the initial concentrations of 1
ery steps.
were varied. These results were used to evaluate which
The chosen model reaction was the asymmetric reduc- conditions would better suit the subsequent integrated
tion of 4-oxoisophorone (OIP; 1) using bakerꢀs yeast (S. experiment involving in situ product crystallization
cerevisiae) as biocatalyst (Figure 2). The desired prod- when a close-down phase is implemented towards the
uct is known as 6R-dihydrooxoisophorone (DOIP; 2), end of the process.
which is a key intermediate in carotenoid synthesis[4]
and in the production of saffron and tobacco flavors.[5]
As bakerꢀs yeast is known to also degrade 2 mainly to Results and Discussion
4S,6R-actinol (ACT; 3), an unwanted by-product, prod-
uct 2 must be removed from the fermenter as soon as it is
Cell Cultivation and Batch Reduction of 1
formed, to prevent low product yield and selectivity. In
this case, in situ product crystallization (ISPC) is appro-
priately applied.
The fed-batch cultivation of bakerꢀs yeast cells is well-
Bakerꢀs yeast is cheap, readily available, and thus is an established in the literature.[7] In this work, bakerꢀs yeast
exceptional organism as it does not need to be cultivated cultivation varied at different glucose feeding rates as
Table 1A. Biomass concentration during cell cultivation at different glucose feed rates for 24 h.
Experiment
Biomass concentration [gdw Lꢀ1
]
Nutrient (glucose) feed rate [mmol hꢀ1
]
Initial, CXi
Final, CXf
Cm0,a
Cm1,a
Cm3,a
Cm5,a
1.3
10.6
10.2
10.6
10.5
12.5
18.3
37.4
0.32
0.31
1.06
1.68
Table 1B. Biomass concentration during cell cultivation and batch reduction at different glucose feed rates for 24 h.
Experiment
Biomass concentration [gdw Lꢀ1
]
Nutrient (glucose) feed rate [mmol hꢀ1
]
Initial, CXi
Final, CXf
Cm0,b
Cm1,b
Cm3,b
Cm5,b
0.8
10.2
9.8
5.3
10.0
10.2
31.6
0.32
0.31
1.06
1.68
10.2
1148
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1147–1154

Synthesis of 6R-Dihydrooxoisophorone
FULL PAPERS
expected (see Table 1A). From an initial cell concentra-
When fed-batch cell cultivation was coupled with
tion CXi ¼1.3 gdw·Lꢀ1, biomass accumulation in the re- batch reduction of 1, then biomass concentration in
actor reached 10.5 gdw Lꢀ1 after 24 h with a glucose feed the reactor changed (Table 1B). This was obviously
rate of 0.32 mmol hꢀ1 (Cm0,a). Increasing the glucose due to product formation, competing in the cell with
feed concentration to over 5 times higher at 1.68 mmol cell growth and maintenance for the carbon and energy
hꢀ1 (Cm5,a) would favor a further increase in biomass sources. The results in Table 1B show that the coupled
from CXi ¼10.6 gdw Lꢀ1 to an amount of >30 gdw Lꢀ1 reactions could be sustained at an approximately five-
after 24 h. This glucose feed rate can apparently support fold increase in medium nutrient feed rate (Cm5,b at
cell growth and maintenance better than the other feed 1.68 mmol hꢀ1) when cell growth as well as bioreduction
rates used (i.e., Cm1,a and Cm3,a in Table 1A). Combining occurred simultaneously. In this case, cell accumulation
strategies involved in experiment Cm0,a and Cm5,a would in the reactor reached ffi 30 gdw Lꢀ1 from an initial
lead to cell cultivation up to 30 gdw Lꢀ1 in 48 h by em- amount of 10.2 gdw Lꢀ1 (Table 1B) concomitant with
ploying the appropriate nutrient feed rates.
batch reduction of 1 at an average biomass-specific
rate of ffi 0.33 mmol gdwꢀ1 hꢀ1 (Figure 3). This condi-
tion was therefore applied subsequently in the integrat-
ed fermentation-crystallization process.
The observed initial biomass-specificreduction rate of
1 of 0.33 mmol·gdwꢀ1 ·hꢀ1 was essentially the same at
different medium nutrient feed rates (Figure 3) and
was about five times higher than in reductions using rest-
ing cells.[2] This suggests that metabolic reactions with
growing cells can be very different as compared with
resting cells, which in this case favors higher reduction
capacity. Glucose as the carbon/energy source was uti-
lized preferably for product formation rather than for
cell growth as no further increase in biomass was ob-
served when the nutrient (glucose) feed rate was
<1.68 mmol·hꢀ1 (Table 1B; Cm1,b and Cm3,b). Further-
more, the final conversion rates varied; complete
(100%) conversion was attained within 24 h with Cm5,b
(Figure 3), as more glucose was available for product
formation and cell growth than with Cm1,b and Cm3,b
,
where degrees of conversion were 85.5% and 92.2%, re-
spectively.
At different initial concentrations of 1, the observed
initial biomass-specific reduction rates varied (Figure
4). A maximum reduction rate of 0.33 mmol·gdwꢀ1
·
hꢀ1 was achieved at an initial concentration (COIP,i) of
53.8 mM (ffi8.2 g·Lꢀ1), but this rate decreased to
0.25 mmol·gdwꢀ1 ·hꢀ1 and 0.23 mmol·gdwꢀ1 ·hꢀ1 at
COIP,i ¼67.5 mM (ffi10.3 g·Lꢀ1) and 37.2 mM (ffi5.7 g·
Lꢀ1), respectively. This indicates that whilst batch reduc-
tion rates were much faster with growing cells than with
resting cells, the maximum rate with the former was ob-
served at a slightly lower substrate concentration
(COIP,i ¼55 mM) than the latter which was at COIP,i
79 mM.[4,8] Consequently, the concentration of 1 in the
reactor must be maintained (i.e., by feeding) at COIP
¼
ffi
55 mM to implement the observed maximum reduction
rate with growing cells. Towards the end of the batch re-
duction (see Figure 4), conversion of 1 was not complete
(only 83–89%) for all experiments at the same nutrient
feed rate (0.33 mmol·hꢀ1) and initial biomass concen-
tration (CXi ¼10 gdw·Lꢀ1). No biomass growth was ob-
served in these cases; rather a slight decrease in biomass
was noted at the end of the 48-h experiment. Obviously,
the glucose supplied in these experiments was not suffi-
Figure 3. Total amounts of (A) 1 and (B) 2 and 3 at different
nutrient feed rates during cell cultivation with simultaneous
**
batch reduction of 1:
þ, Cm1 (0.31 mmol·hꢀ1, Cgluc, feed
¼
102.6 mM, CXi ¼10.2 gdw·Lꢀ1); && , Cm3 (1.06 mmol·hꢀ1
,
*
Cgluc, feed ¼353.1 mM, CXi ¼9.8 gdw·Lꢀ1); ~~*, Cm5
(1.68 mmol·hꢀ1, Cgluc, feed ¼560.9 mM, CXi ¼10.2 gdw·Lꢀ1).
Adv. Synth. Catal. 2005, 347, 1147–1154
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1149

Evelyn M. Buque-Taboada et al.
FULL PAPERS
a 0.2-mm filter.[2] Two process schemes were designed.
Process scheme A was an integrated fermentation-crys-
tallization process where the reduction stage was carried
out concomitant with further biomass cultivation after
24 h from the start of the experiment. Process scheme
B was a modification of process scheme Awhere the re-
duction stage was implemented when the biomass con-
centration in the reactor had already reached an opti-
mum level. Further elaborations of these processes are
given subsequently.
Process Scheme A
This combines the cell cultivation, bioreduction and in
situ product crystallization stages. As shown in Figure 5,
the first stage of 24 h consisted of yeast cell cultivation
from an initial concentration (CXi) of 1.23 gdw·Lꢀ1 to
9.84 gdw·Lꢀ1 when the glucose feed rate was
0.32 mmol·hꢀ1. In the second stage (from 24 to 48 h), re-
duction was started by addition of 55.5 mmol (8.4 g) 1
and feeding at 0.5 mL·hꢀ1 (3.4 mmol·hꢀ1). The glucose
feed rate was increased to over five-fold (1.67 mmol·
hꢀ1) to further support cell growth and product forma-
tion. In the third stage, feeding of 1 was increased to
1.5 mL·hꢀ1 (10.2 mmol·hꢀ1) and the glucose feed rate
was raised to 6.01 mmol·hꢀ1 as biomass was expected
to increase to a maximum amount of about 30 gdw·Lꢀ1
.
However, the results in Figure 5 show that biomass did
not increase in stage 2 when reduction of 1 was imple-
mented. Rather, it remained constant during this period
(contrary with batch reduction when cell dry weight si-
multaneously increased as shown earlier). Eventually,
Figure 4. Total amounts of (A) 1 and (B) 2 and 3 at different
initial concentrations of 1 during cell cultivation with simulta-
**
*
neous batch reduction:
þ, COIP,i ¼67.5 mM; && ,
COIP,i ¼53.8 mM; ~~*, COIP,i ¼37.2 mM.
cient to sustain all the required metabolic reactions in
the cell.
Integrated Fermentation-Crystallization Procedure for
Reduction of 1
This process was performed as shown in Figure 1 where
the fermenter was attached to the crystallizer and in be-
tween the two units an ultrafiltration membrane was
placed in an external loop to recycle the biomass.
From the crystallizer, the mother liquor (excluding the
&
*
Figure 5. Process scheme A: Total amounts of 1 ( ), 2 ( ), 3
~
*
( ) and biomass ( ) in the reactor. Reduction was started af-
ter 24 h (stage 2) and no further increase in biomass was ob-
crystals) was recirculated back to the fermenter through served after this time.
1150
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1147–1154

Synthesis of 6R-Dihydrooxoisophorone
FULL PAPERS
it decreased in stage 3 when 1 started to accumulate in
the reactor. Biomass disintegration and lysis occurred
at this stage, resulting to the formation of biomass (de-
bris) aggregates, which caused severe clogging in the ul-
trafiltration membrane. After 72 h, the biomass concen-
tration (CX) was 6.28 gdw·Lꢀ1
.
In this case, when 1 was fed such that its concentration
was constant in the reactor (i.e., COIP ꢁ 55 mM), cell
growth was inhibited as illustrated in Figure 5. The ex-
pected maximum amount of cells in the reactor was
not attained. Nevertheless, reduction of 1 occurred as
2 increased in time. This indicates that the cells favored
the utilization of glucose for product formation (reduc-
tion of 1) rather than cell growth. Hence, feeding of 1
should be started at the moment that the targeted max-
imum biomass concentration has been reached in order
to maximize bioreactor productivity.
&
*
~
Figure 6. Process scheme B: Amounts of 1 ( ), 2 ( ), 3 ( )
Process Scheme B
*
and biomass ( ) in the reactor. Reduction was started after
48 h (stage 3), when CX ¼30 gdw·Lꢀ1. Feeding of 1 was
done at a rate of 1.5 mL·hꢀ1, implementing a reduction
rate of about 0.33 mmol·gdwꢀ1 ·hꢀ1 as concentration level
of 1 was constant in the reactor. During the close-down phase
(stage 4), feeding of 1 was stopped but reduction was contin-
ued for another 24 h.
In this experiment, the observations and results ob-
tained in process scheme A were considered. As shown
in Figure 6, the reduction of 1 was started in stage 3 when
the cell concentration in the reactor had already reached
30 gdw·Lꢀ1 to avoid growth inhibition and maximize bi-
oreactor productivity. Compound 1 was added in the re-
actor at 56.5 mmol (8.6 g) and at the same time, fed at
1.5 mL·hꢀ1 (10.2 mmol·hꢀ1) to maintain the concentra- moval during the integrated process. In addition to the
tion of 1 in the reactor that allows the maximum reduc- dissolved 2, such as shown in Figure 6 for the reactor,
tion rate. The glucose feed rate was also raised, to crystals of 2 (60.50 g) accumulated in the crystallizer
6.04 mmol·hꢀ1, to further sustain product formation and were readily recovered. Final washing with ice-
and cell maintenance. The results (Figure 6) show that cold water, filtration and drying gave product crystals
the biomass concentration in the reactor was indeed of 99.5% purity. The enantiomeric excess (ee) of 2 was
maintained during reduction of 1 in stage 3 as the aver- ꢁ 98%, making the aforementioned procedure an at-
age biomass dry weight was 30.5 g. The concentration tractive option for enantioselective reduction of 1. The
of 1 in the reactor was constant at COIP ffi 60 mM, which biocatalyst consumption rate was 0.42 kg per kg of 2 ob-
indicates that the feed rate of 1 (1.5 mL·hꢀ1
¼
tained; this can further be reduced when the reduction
10.2 mmol·hꢀ1) was equal to the observed reduction process is continued up to several days more before
rate (ffi0.33 mmol·gdwꢀ1 ·hꢀ1). This was a favorable the close-down phase. This is a potential benefit in par-
situation for an integrated process because the reduc- ticular for other processes with more expensive microbi-
tion rate was maximal by keeping the concentration of al biocatalysts. The volumetric productivity was 0.55 g·
1 in the reactor at a constant level while avoiding sub- Lꢀ1 ·hꢀ1 and would certainly increase when the biore-
strate inhibition and cell lysis.
duction process would be optimally lengthened and/or
At different stages during the experiment, the glucose the biocatalyst cultivation period reduced.
feed rate was raised appropriately to sustain the re-
quired metabolic reactions in the cell, and support bio-
catalytic activity. Dissolved oxygen during the experi- Morphology and Crystal Size Distribution of 2
ment was constantly above 50% of air saturation. Meta-
bolic by-productsin the samplessuch asethanol andace- It is important to assess the morphology and crystal size
tate were negligible.
distribution of the resulting product as its properties
In Table 2 are shown the important process param- (i.e., dissolution rates, viscosity, color, shape, impurity
eters usually used to assess the efficiency of a certain bi- content) have a large impact on its applications and han-
ocatalytic process.[9] In this process scheme, the final dling.[10] The crystals of 2 obtained were white and rod-
yield and selectivity of 2 were 85% and 99%, respective- like (Figure 7A). They were mechanically quite stable
ly. The latter indicates a negligible degradation of 2, ren- and tended to form large complex aggregates (Fig-
dering ISPC an effective method for in situ product re- ure 7B). The results in Figure 8 show that most of the
Adv. Synth. Catal. 2005, 347, 1147–1154
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1151

Evelyn M. Buque-Taboada et al.
FULL PAPERS
Table 2. Process parameters for the integrated fermentation-
crystallization experiment (process scheme B).
in favor of the larger particles, especially when these
crystals have undergone storage. When the produced
crystals underwent final washing and drying to increase
the purity up to 99.5% (the same purity as that of the
standard sample), the crystal morphology and size dis-
tribution were comparable. This indicates that recrystal-
lization is not required, despite the presence of many
solutes in the fermentation medium used for cell growth.
This process strategy might be applicable to many bio-
catalytic (e.g., reduction) systems where crystalline
products are obtained. Many examples of reactions
where ISPR via crystallization might be employed are
known.[2,10a,11] Products from this process can include
specialty biochemicals, pharmaceuticals, and fine chem-
icals such as amino acids, steroids, proteins, esters and
Parameter [units]
Results
Final yield (2/1 fed), [%]
85
98.7
Final selectivity (2/(2þ3)), [%]
Final concentration of 2 [g·Lꢀ1
]
8.98[a]/101.45[b]
0.42
Biocatalyst consumption [kg/kg 2]
Volumetric productivity [g·Lꢀ1 ·hꢀ1
]
0.55[c]/0.92[d]
[a]
In fermenter.
In crystallizer.
[b]
[c]
During cultivation and reduction stages.
During reduction stage only.
[d]
crystals have a length range of 1–30 mm and a diameter ketones. The biocatalyst involved may range from
range of 1–20 mm. Considering a crystal population of wild-type bakerꢀs yeast (S. cerevisiae) or bacteria (Es-
1500, the average length and diameter were 20 and cherichia coli) to genetically engineered microorgan-
12 mm, respectively. However, due to the tendency to isms. However, when microbial cultivation is necessary,
form aggregates, the crystal size distribution may vary substrate/product inhibition, toxicity and degradation,
sufficient supply of energy and carbon sources, as well
as aeration and agitation, must be taken into account
during the process design, amongst others, in order to
obtain an economically promising biocatalytic process.
Conclusion
An efficient procedure was developed for the synthesis
of 2involving cell cultivation coupled with reduction of 1
Figure 7. (A) Typical crystals of 2, which are rod-like in
shape, and tend to form large complex aggregates (B).
and subsequent in situ product crystal formation. Proc-
ess implementation is feasible, however, biocatalyst cul-
tivation must be done prior to bioreduction and in situ
product crystallization as 1 inhibits cell growth when
its concentration in the reactor is constantly high (ꢁ
55 mM). An appropriate supply of the carbon and ener-
gy source such as glucose is necessary as it is preferably
utilized for product formation rather than cell growth.
Furthermore, the reduction rate of 1 using growing cells
was five times higher than using resting cells. A concen-
tration of over 100 g·Lꢀ1 of 2 was obtained in the crys-
tallizer at the end of the integrated experiment. The
product crystals were readily recovered and did not re-
quire recrystallization.
Experimental Section
Chemicals
Compound 1 (>98%) was supplied by Fluka Biochemika
(Buchs, Switzerland). Standard samples of 2 and 3 were kindly
provided by DSM (Basel, Switzerland). The purity of all other
Figure 8. Crystal size distribution of 2.
chemicals used was at least laboratory grade.
1152
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1147–1154

Synthesis of 6R-Dihydrooxoisophorone
FULL PAPERS
Cell Cultivation and Batch Reduction of 1
Analytical Methods
Cells were cultivated aerobically in a 2-L bioreactor (Appli-
kon, Schiedam, The Netherlands). The fermenter was equip-
ped with two six-blade Rushton-type impellers (diameter
4.5 cm; impellers 0.5 cm above each other) and with an air out-
let condenser. The total working volume was 1 L, the air flow
was 0.25 vvm and the stirrer speed was 800 rpm. The tempera-
tures of the fermenter and its attached condenser were set at
308C and 2–48C, respectively. The control of pH at 5.5 with
1 M H2SO4 and 2 M KOH was done via a biocontroller (Appli-
kon, ADI 1010). BIOEXPERT software (Applikon, NL) was
used as data-acquisition program. The parameters measured
on-line during the experiment were the dissolved oxygen con-
centration, pH, temperature, and stirrer speed.
Compounds 1, 2, and 3 in the supernatant of the reaction mix-
ture were analyzed as described previously.[2] The enantiomer-
ic excess (ee) of 2 was checked using a gas chromatograph (Shi-
madzu GC 17A) equipped with an FID and a chiral diAc-tBu-
Si-b-cyclodextrin column (MEGA, Legnano, Italy) 25 mÂ
0.25 mm i.d., film thickness 0.25 mm, carrier gas helium, flow
rate 0.58 mL·minꢀ1, split ratio 1:50, temperatures of injector
and detector 2508C and 2808C, respectively. Retention times
for 1, 2, 6S-dihydrooxoisophorone (6S-DOIP), 4R,6R-actinol,
and 3 were 2.7, 5.0, 5.2, 11.1 and 11.3 min, respectively. Glu-
cose, ethanol, acetate as well as the biomass dry weight in the
reaction mixture were analyzed as described elsewhere.[13]
The measured amounts of 1, 2, 3, glucose and biomass were cor-
rected for the actual volume in the reactor as well as the
amounts taken out during sampling.
A mineral growth medium for bakerꢀs yeast cultivation[12]
was prepared. The standard procedure was to add 0.7 L me-
dium solution to the fermenter followed by a known amount
of bakerꢀs yeast (1 g active dry bakerꢀs yeast, Fermixꢂ; 97.5%
dry weight, DSM-Gist, Delft, The Netherlands) suspended in
0.30 L medium solution to make up the 1-L total working vol-
ume. The mixture was stirred and aerated for 30 min to accli-
matize (pre-incubate) the yeast. Subsequently, glucose was
supplied at a rate of 0.33 mmol·hꢀ1 with CGluc, feed ffi 110 mM
in medium solution. Other glucose feed rates (as specified)
were also used in various experiments. The reactor system
was allowed 2 h to attain stationary oxygen consumption at
308C. Samples of the reactor liquid were analyzed for biomass
dry weight and glucose concentration. Ethanol and acetate
concentrations in the samples were also checked occasionally.
For batch reductions, the reaction was started by addition of
a known amount in the range of 5–10 g of pure 1, after the 2-h
acclimatization period when oxygen consumption in the reac-
tor was stabilized. For these experiments, the liquid samples
were analyzed for 1, 2 and 3 in addition to biomass dry weight
and concentrations of glucose, ethanol, and acetate.
The product crystal morphology and the crystal size distri-
bution (CSD) were determined using an Image Analyzer
(IA) consisting of a Sony CCD video camera module (XC-
77CE), an Olympus Stereo zoom microscope (SZH) and a
PC with IA software LEICA Qwin version 3 (Olympus).
Acknowledgements
This research is funded by the MHO-USC-TUD Project in
Chemical Engineering. We kindly thank DSM (Delft, The Neth-
erlands) for providing bakerꢀs yeast strains and F. Hoffman-La
Roche-VFCD, now DSM (Basel, Switzerland) for DOIP and
actinol. We would like to thank Max Zomerdijk and Stef van
Hateren (TU Delft), Victoria Napisa and her staff (USC) and
Florie Gil Pino (research assistant, USC) for all the valuable
support and assistance in conducting this research work.
References
[1] a) G. J. Lye, J. M. Woodley, Trends Biotechnol. 1999, 17,
395–402; b) K. Schꢃgerl, Biotechnol. Adv. 2000, 18,
581–599; c) D. Stark, U. von Stockar, Adv. Biochem.
Eng. Biotechnol. 2003, 80, 149–175; d) A. J. J. Straathof,
Biotechnol. Prog. 2003, 19, 755–762.
[2] E. M. Buque-Taboada, A. J. J. Straathof, J. J. Heijnen,
L. A. M. van der Wielen, Biotechnol. Bioeng. 2004, 86,
795–800.
[3] a) J. Bao, K. Koumatsu, K. Furumoto, M. Yoshimoto, K.
Fukunaga, K. Nakao, Chem. Eng. Sci. 2001, 56, 6165–
6170; b) J. P. Cardoso, Biotechnol. Bioeng. 1993, 42,
1068–1076; c) M. Furui, N. Sakata, O. Otsuki, T. Tosa,
Biocatalysis 1988, 2, 69–77; d) S. Takamatsu, D. Y. Y.
Ryu, Biotechnol. Bioeng. 1988, 32, 184–191; e) S. Taka-
matsu, D. Y. Y. Ryu, Enzyme Microb. Technol. 1988,
10, 593–600; f) T. Tosa, M. Furui, N. Sakata, O. Otsuki,
I. Chibata, Ann. N. Y. Acad. Sci. 1988, 542, 440–443.
[4] a) H. G. W. Leuenberger, in: Biocatalysts in Organic Syn-
thesis, (Eds.: J. Tramper, H. C. van der Plas, P. Linko),
Elsevier Science Publishers, Amsterdam, 1985, pp. 99–
118; b) H. G. W. Leuenberger, W. Boguth, E. Widmer,
R. Zell, Helv. Chim. Acta 1976, 59, 1832–1849.
Synthesis of 2
In these experiments, the same set-up and protocols were em-
ployed as in cell cultivation and batch reductions with the fol-
lowing modifications. The initial cell concentration employed
was about 1 gdw·Lꢀ1. Glucose feed rate was varied (as speci-
fied) at different stages of the experiment. After 24 h, reduc-
tion was started by addition of about 8 g of pure 1 and, at the
same time, 1 was fed at 0.50 mL·hꢀ1 (3.4 mmol·hꢀ1) to main-
tain a concentration level in the reactor that avoids substrate
inhibition.[4,8] The feed rate of 1 was increased to 1.5 mL·hꢀ1
(10.2 mmol·hꢀ1) after 48 h as the cell concentration was ex-
pected to increase three-fold at this time. Subsequently, the in-
tegrated fermentation-crystallization procedures described
previously[2] were followed. In the course of the experiment,
liquid samples were taken from the reactor and the crystallizer
and analyzed for 1, 2, and 3. Concentrations of glucose, ethanol
and acetate as well as biomass dry weight in the liquid samples
from the reactor were also determined. The crystallizer volume
was over 0.6 L.
Adv. Synth. Catal. 2005, 347, 1147–1154
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1153

Evelyn M. Buque-Taboada et al.
FULL PAPERS
[5] a) N. Hori, T. Hieda, Y. Mikami, Agric. Biol. Chem. 1984,
48, 123–129; b) K. Sode, K. Kajiwara, E. Tamiya, I. Kar-
ube, Biocatalysis 1987, 1, 77–86.
[6] a) P. DꢀArrigo, G. Pedrocchi-Fantoni, S. Servi, Adv. Appl.
Microbiol. 1997, 44, 81–123; b) W. F. H. Sybesma, A. J. J.
Straathof, J. A. Jongejan, J. J. Heijnen, Biocatal. Bio-
transform. 1998, 16, 95–134.
[7] a) S. L. Chen, M. Chiger, in: Comprehensive Biotechnol-
ogy, Vol. 3, (Ed.: M. Moo-Young), Pergamon Press, Ox-
ford, 1985, pp. 429–461; b) R. F. Beudeker, H. W. van -
Dam, J. B. van der Plaat, K. Vellenga, in: Yeast Biotech-
nology and Biocatalysis, (Eds.: H. Verachtert and R.
De Mot), Marcel Dekker, New York, 1990, pp. 103–
146; c) P. M. van Hoek, PhD thesis, Delft University of
Technology (NL), 2000.
[9] A. J. J. Straathof, S. Panke, A. Schmid, Curr. Opin. Bio-
technol. 2002, 13, 548–556.
[10] a) A. McPherson, Crystallization of Biological Macro-
molecules, CSHL Press, New York, 1999; b) J. W. Mullin,
Crystallization, 4th edn., Butterworth-Heinemann, Ox-
ford, 2001; c) N. S. Tavare, Industrial Crystallization:
Process Simulation, Analysis and Design, The Plenum
Chemical Engineering Series, New York, 1995.
[11] A. Liese, K. Seelbach, C. Wandrey, Industrial Biotrans-
formations, Wiley-VCH, Weinheim, 2000; b) R. N. Patel,
Stereoselective Biocatalysis, Dekker, New York, 2000.
[12] C. Herwig, U. von Stockar, Biotechnol. Bioeng. 2003, 81,
837–847.
[13] a) E. M. Buque, I. Chin-Joe, A. J. J. Straathof, J. A. Jon-
gejan, J. J. Heijnen, Enzyme Microb. Technol. 2002, 31,
656–664; b) I. Chin-Joe, A. J. J. Straathof, J. T. Pronk,
J. A. Jongejan, J. J. Heijnen, Biotechnol. Bioeng. 2001,
75, 29–38.
[8] E. M. Buque-Taboada, A. J. J. Straathof, J. J. Heijnen,
L. A. M. van der Wielen, Enzyme Microb. Technol.
2005, accepted for publication.
1154
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1147–1154