Appl Microbiol Biotechnol (1995) 44 : 7Ð14 ( Springer-Verlag 1995
O R I G I N A L P A P E R
H. S. Shin á P. L. Rogers
Biotransformation of benzeldehyde
to L-phenylacetylcarbinol, an intermediate
in L-ephedrine production, by immobilized Candida utilis
Received revision: 24 February 1995/ Accepted: 1 March 1995
Abstract Biotransformation of benzaldehyde to L-
phenylacetylcarbinol (L-PAC) as a key intermediate for
L-ephedrine synthesis has been evaluated using immo-
bilized Candida utilis. During biotransformation, the
benzaldehyde level and respiratory quotient signiÞ-
cantly a¤ected both L-PAC and by-product benzyl al-
cohol formation. By controlling the benzaldehyde level
at 2 g/l, maintaining a respiratory quotient of 5Ð7 and
pulse feeding glucose, a Þnal concentration of 15.2 g/l
L-PAC was achieved in a fed-batch process. This com-
pares with previous published results of 10Ð12 g/l in
batch culture and 10 g/l L-PAC in a semicontinuous
process with immobilized Saccharomyces cerevisiae. In
a single stage continuous process with immobilized C.
utilis, the steady state L-PAC concentration was signiÞ-
cantly reduced because of the sustained toxic e¤ects of
benzaldehyde.
Introduction
L-phenylacetylcarbinol (L-PAC) is an intermediate in
the production of L-ephedrine and pseudoephedrine,
pharmaceutical compounds used as decongestants and
anti-asthmatics. Reports have indicated also its potential
use in obesity control (Astrup et al. 1992a, b). It is
currently produced via a microbial biotransformation
process using di¤erent species of yeasts with benzal-
dehyde as the aromatic substrate. The following
diagram (Fig. 1) outlines the biotransformation process,
which involves the condensation of an ÔÔactive acetal-
dehydeÕÕ (from pyruvic acid produced by the yeast) with
benzaldehyde. The production of the L-PAC is catalysed
by the enzyme pyruvate decarboxylase (PDC), and is
associated with the formation of benzyl alcohol as a
H. S. Shin á P. L. Rogers (¥)
Department of Biotechnology, The University of New South Wales,
Sydney, N. S. W. 2052, Australia
by-product resulting from the activity of an alcohol
dehydrogenase (ADH) and/or oxidoreductases.
Previous studies have reported concentrations of
10Ð12 g/l L-PAC in batch culture (Vojtisek and Netrval
1982; Culic et al. 1984) and 10 g/l for a semicontinuous
culture using immobilized Saccharomyces cerevisiae
(Mahmoud et al. 1990a, b). Strain-improvement studies
with acetaldehyde-resistant mutants have been re-
ported by Seely et al. (1989) with similar levels of
L-PAC. The role of puriÞed PDC in L-PAC production
has been studied by Bringer-Meyer and Sahm (1988)
and other fundamental investigations have indicated
that oxidoreductases distinct from ADH may be in-
volved in by-product benzyl alcohol formation (Long
and Ward 1989; Nikolova and Ward 1991).
Current commercial practice involves a fed-batch
process with fermentative growth on sugars to produce
biomass, pyruvic acid and induce PDC activity. The
growth phase is followed by a biotransformation phase
involving the further addition of sugars and the pro-
grammed feeding of benzaldehyde to maximize L-PAC
production. Cessation of L-PAC production can occur
as a result of the following factors acting either together
or independently:
1. SigniÞcant reduction of PDC activity due to ben-
zaldehyde or end-product inhibition
2. Pyruvic acid limitation at the end of the biotrans-
formation phase
3. Cell viability loss due to extended exposure to
benzaldehyde and/or increasing concentrations of
benzyl alcohol and L-PAC.
In the present study, an immobilized cell system
with Candida utilis has been investigated. The immobi-
lized cell system was selected for detailed evaluation, as
the research by Mahmoud et al. (1990a, b) has sugges-
ted that the toxic e¤ects of benzaldehyde may be mini-
mized by the di¤usional limitations of immobilizing
matrices. The kinetics of both batch and continuous
biotransformation processes have been assessed in the
current investigation.Fig. 1 Mechanism of L-PAC
formation
Materials and Methods
Microorganism and culture media
Candida utilis was kindly provided by ICI Australia Pty. Ltd. The
strain was maintained on culture medium containing (g/l) glucose
20, yeast extract 3.0, (NH
4
)
2
SO
4
2.0, KH
2
PO
4
1.0, MgSO
4
á 7H
2
O
1.0, agar 1.5 with an initial pH of 6.0. For its growth and subsequent
immobilization, this strain was cultivated in a fermentation medium
consisting of (g/l) glucose 60, yeast extract 10, (NH
4
)
2
SO
4
10,
KH
2
PO
4
3.0, Na
2
HPO
4
á12H
2
O 2.0, MgSO
4
á 7H
2
O 1.0, CaCl
2
0.05,
FeSO
4
0.05, Mn SO
4
á 4H
2
O 0.05 at an initial pH of 5.0 and temper-
ature of 25¡ C.
Culture and biotransformation system
The system consisted of several components: fermenter, benzal-
dehyde-feeding pump, exit-gas analysers and computer-linked on-
line respiratory quotient measurement. In this system, the 2-l LH
fermenter (working volume 1.5c) was used for cell growth under
controlled conditions at 25¡ C and pH 5.0 and for biotransformation
at 20¡ C and pH 6.0. For the continuous process, the overßow outlet
was covered with a stainless-steel sieve (mesh size 1.0 mm) to main-
tain the immobilized cell beads in the fermenter. The culture medium
was fed into the fermenter by means of a peristaltic pump (Gilson
Minipul). To feed benzaldehyde into the fermenter, a syringe pump
(Perfusor VII, B. Braun) was used with variable feed rates in the
range of 0.1Ð99.0 ml/h by means of a 50-ml disposable syringe.
Prior to analysis for O
2
and CO
2,
the exit gas was dehumidiÞed
by a cold dehumidiÞer (Komatsu Electronics Inc. Model DH
1052G) to meet the requirements of the gas analysers. The content of
oxygen was measured by the paramagnetic susceptibility of the
sample (Servomix type 1400A), while the content of carbon dioxide
was measured by an infrared gas analyser with single-beam dual
wavelength (Servomix-R type 1410). Output signals (4Ð20 mA) from
both gas analysers were fed into a data interface (Data system FC-4,
Real Time Engineering, Australia) linked to an NEC Powermate
computer. RQ values were calculated instantaneously by this com-
puter and the results were used for control via aeration and/or
agitation.
Immobilization of C. utilis cells
When the level of PDC reached a maximum with pulse feeding of
glucose, the cells were harvested and resuspended in sodium alginate
(3% w/v). Samples of 150 g wet cells (approximately 30 g dry
weight)/300 ml solution were prepared for immobilization. This
preparation was extruded into 2% CaCl
2
solution through a 0.5-mm-
diameter needle, and stabilized in fresh 2% CaCl
2
solution contain-
ing 3% glucose for 1Ð2 h at 4¡ C. For further stabilization and prior
to biotransformation, the immobilized beads were reintroduced into
their own former supernatant with further supplementation of glu-
cose and yeast extract.
Dry cell weight estimation
After centrifugation of a sample of culture broth and resuspension in
isotonic saline, 4 ml of the cell suspension was transferred to pre-
weighed glass tubes and centrifuged at 5000 rpm for 10 min. The
glass tubes containing the cells were dried in an oven at 105¡ C for
24 h, cooled in a desiccator and reweighed. The average values from
three measurements were determined for each sample.
Estimation of glucose concentration
Glucose concentrations were determined by a YSI glucose analyser
(Yellow Springs Instruments Co., model 27).
Estimation of ethanol concentration
Ethanol concentrations were estimated using a gas chromatograph
(Packard, series 427). The relevant column and its operation were as
follows: column material, 6.4-mm glass 1.5 m long; packing material,
Porapak Q in mesh range 100Ð200 lm; carrier gas, nitrogen
(30 cm3/min); oven temperature, 180¡ C (isothermal); injector tem-
perature, 220¡ C; detector temperature, 220¡ C with ßame ionization
detector; injection sample, 3 ll. The ethanol concentrations of the
sample were estimated by comparison with standard samples.
Estimation of benzaldehyde, L-PAC and benzyl alcohol
concentrations
Concentrations of benzaldehyde, L-PAC and benzyl alcohol were
determined by gas chromatography. Samples were prepared by
extraction into dichloromethane (sample:solvent"1:5). The bio-
transformation sample (0.2 ml) was mixed with 1 ml dichloro-
methane in a microcentrifuge tube and vortexed for 2 min. A sample
from the bottom organic layer was injected into a gas chromato-
graph with the column and its operating conditions as follows:
column material, 6.4-mm glass 1 m long; packing material, Chromo-
sorb W. Hr/SE 30WTX 10 in the mesh range of 80Ð100 lm; carrier
gas, nitrogen (30 cm3/min); oven temperature, 115¡ C (isothermal);
injector temperature, 180¡ C; detector temperature, 180¡ C with
ßame ionization detector; injection sample, 3 ll. The concentrations
8of benzaldehyde and benzyl alcohol were determined by comparison
with standard samples (from Aldrich) and L-PAC (from ICI Austra-
lia Pty. Ltd).
Pyruvic acid determination
Determination of pyruvic acid was carried out by an enzymatic
analysis (Boehringer-Mannheim analytical kit no. 718 882). In the
presence of NADH#H`, lactate dehydrogenase reduces pyruvic
acid to lactic acid and the amount of NADH#H` oxidized to
NAD` corresponds stoichiometrically to the amount of pyruvic
acid. The decrease in NADH#H` was determined by di¤erence in
sample absorbance at 340 nm.
Analysis of enzyme activities
Extraction of enzymes from free cells
Cells from 1 ml of culture broth were harvested by centrifugation
(Eppendorf Centrifuge) at 12000 rpm for 1 min and washed twice
with 30 mM TRIS bu¤er (pH 6.5). Cells were resuspended in the
same bu¤er and the volume adjusted to 0.4 ml. Approximately 1 g
glass beads (size 0.5 mm, B. Braun, catalogue no. 854 170/1) were
mixed with 0.4 ml cell suspension and vortexed at maximum speed
for 2 min. For every 30 s of vortexing, the sample was cooled for
1 min in an ice bath. Cell debris were removed by centrifugation at
12 000 rpm for 3 min. The supernatant were collected for subsequent
enzyme assays and protein determination.
Extraction of enzymes from immobilized cells
To extract enzymes from immobilized cells, 3 ml immobilized beads
containing C. utilis were put into a ceramic hammer mill and gently
extracted with 3 g pretreated Þne sand (which was washed three
times with 3 M HCl and Reverse Osmosis (RO) water until neu-
tralized). Crushed immobilized beads were suspended in 20 ml water
and centrifuged at 1000 rpm for 2Ð3 min. From the resultant super-
natant, yeast cells were harvested and washed with 30 mM TRIS
bu¤er (pH 6.0) by centrifugation at 5000 rpm for 10 min. The har-
vested cells were resuspended into 3.0 ml 30 mM TRIS bu¤er, and
then 1 ml yeast suspension was centrifuged at 12 000 rpm for 1 min,
and the enzymes were extracted by ball milling.
Pyruvate decarboxylase
The activity of PDC was assayed by coupling the decarboxylation
reaction with the ADH-mediated reaction and monitoring the oxi-
dation of NADH#H` to NAD` at 340 nm (Bergmeyer 1974).
The reaction mixture consisted of (ll) 200 mM sodium citrate bu¤er
(pH 6.0) 950, 10 mg/ml NADH (sodium salt) 10, 100 mg/ml sodium
pyruvate 32, 10 mg/ml alcohol dehydrogenase (Sigma Chem. Co.,
Product no. A-3263) 3, enzyme sample 5. One unit of enzyme activity
is deÞned as that activity which converts 1.0 lmol of pyruvate to
acetaldehyde/min at pH 6.0 and 25¡ C. The activity of the enzyme
was monitored as NAD` formation by changes in absorbance at
340 nm.
Alcohol dehydrogenase for ethanol
The basic reaction for determination of ADH activity is the oxida-
tion of ethanol to acetaldehyde with monitoring of the reduction of
NAD` to NADH#H` (modiÞed from Bergmeyer 1974). The re-
action mixture consisted of (ll): 35 mM Trizma base (pH 8.5), 935,
20 mg/ml NAD` 30, absolute ethanol 30, enzyme sample 5. One
unit of enzyme activity is deÞned as that activity which converts
1.0 lmol ethanol to acetaldehyde/min at pH 8.5 and 25¡ C. The
activity of the enzyme was monitored as NADH formation by
changes in absorbance at 340 nm.
Protein determination
Protein determinations of cell-free crude extract, following enzyme
extraction were carried out by the Bradford method (Bradford 1970)
with lyophilized bovine serum albumin as a reference.
Results
Optimal fed batch culture of C. utilis and its
fermentative enzyme proÞles
In order to enhance PDC activity prior to cell immobil-
ization, an extended fed-batch culture under partially
fermentative conditions was developed. Initially, respir-
atory metabolism was maintained for the Þrst 8Ð9 h to
obtain a high biomass concentration for immobiliz-
ation, and then a switch from aerobic respiration to
fermentative growth was made by reducing agitation
(from 1000 rpm to 500 rpm) and aeration rate (0.6 vvm
to 0.3 vvm). Before the initial glucose was completely
exhausted, pulse feeding of a supplement containing
glucose and yeast extract (approximate concentrations
30 g/l and 5 g/l respectively) was initiated. As shown in
Fig. 2, this resulted in enhanced PDC and ADH activ-
ities to maximum values of 0.59 unit/mg and 0.83
unit/mg protein respectively.
Immobilization of C. utilis cells and associated
enzyme proÞles
Following cell immobilization (which involved cell har-
vesting and entrapment in calcium alginate), there was
a decline in enzyme activities because of reduced levels
of cellular metabolism. As a result, a glucose-feeding
protocol was initiated, which resulted in the levels of
PDC and ADH in immobilized C. utilis increasing as
shown in Fig. 3. The highest activities of PDC, ADH
were 0.61 unit/mg and 0.93 unit/mg protein respective-
ly, after 12 h incubation. Comparison of enzyme pro-
Þles indicated that ADH activity was a little higher in
the immobilized cells compared to the free cells.
Comparison of biotransformation by free and
immobilized cells with various initial concentrations
of benzaldehyde
Biotransformation studies were carried out in shake
ßasks with addition of benzaldehyde and 30 g/l glucose
9Fig. 2a,b E¤ect of pulse feeding of glucose on (a) kinetics of Candida
utilis growth: h biomass, d glucose,s ethanol, j pyruvate, and (b)
fermentative enzyme proÞles: s alcohol dehydrogenase, h pyruvate
decarboxylase
with free and immobilized cells after PDC had been
fully induced in both systems. Results for various initial
concentrations of benzaldehyde on L-PAC and by-
product benzyl alcohol formation are shown in Fig. 4
with data expressed on a millimolar basis to illustrate
molar conversion of benzaldehyde to products. No
evidence of benzoic acid production was found follow-
ing biotransformation.
With free and immobilized cells below 30 mM ben-
zaldehyde, benzyl alcohol was preferentially produced
instead of L-PAC. However, L-PAC was preferentially
formed above 40 mM benzaldehyde in both systems.
Higher concentrations of benzaldehyde were accom-
panied by higher L-PAC formation and by an enhanced
molar ratio of L-PAC formation to benzyl alcohol until
benzaldehyde inhibition occurred.
Fig. 3a,b Kinetics of immobilized cells: (a) d glucose consumption,
s ethanol, j pyruvate production; (b) fermentative enzyme proÞles
with pulse feeding of glucose: s alcohol dehydrogenase, h pyruvate
decarboxylase
It was evident also that higher L-PAC concentra-
tions could be produced with immobilized cells com-
pared to free cells, an observation consistent with the
results of Mahmoud et al. (1990a). However, benzyl
alcohol production with immobilized cells was higher
than for free cells over the range of benzaldehyde con-
centrations. From these results it is evident that selec-
tion of an optimum level of benzaldehyde is necessary
to enhance L-PAC formation as well as to minimize
benzyl alcohol formation.
Biotransformation kinetics with various sustained
concentrations of benzaldehyde
Investigations of the e¤ect of a relatively constant ben-
zaldehyde level on L-PAC formation were carried out
10Table 1 Summary of
biotransformation products
with immobilized cells with
various benzaldehyde levels
maintained in the fermanter
Benzaldehyde L-PAC benzyl alcohol dp/dt L-PAC Reaction Molar yield
level (g/l) (g/l) (g/l) (g/l/h) time (h) for L-PAC ! (%)
0.8 7.0 6.4 0.35 20 44.1
1.5 9.5 6.1 0.43 22 52.8
2 10.8 6.0 0.54 20 56.4
4 7.3 4.8 0.45 16 52.3
! Molar conversion yield based on benzaldehyde utilized
Fig. 4a,b Comparison of biotransformation products for (a) free
cells and (b) immobilized cells after 16 h incubation with various
initial concentrations of benzaldehyde in shake ßasks at 20¡ C and
180 rpm: j benzaldehyde, s benzyl alcohol, d L-phenylacetylcar-
binol (¸-PAC)
with four di¤erent benzaldehyde concentrations in
a controlled fermenter following a period of adaptation
for 3Ð4 h. A short acclimatisation phase was used with
addition of 0.8 g/l/h benzaldehyde for this adapta-
Fig. 5 E¤ect of various levels of benzaldehyde (s, 0.8 g/l, d 1.5 g/l,
h 2.0 g/l, j 4.0 g/l) on L-PAC formation as a function of time
tion. During this time, it appeared that the cells
adapted to the toxic substrate, which resulted in min-
imizing viability and/or enzyme activity loss following
later extended exposure. After this acclimatisation, bio-
transformation kinetics were evaluated at various feed-
ing rates of benzaldehyde until L-PAC concentrations
reached their peak values.
To maintain approximately constant benzaldehyde
levels, samples were taken every hour, and benzal-
dehyde concentrations were measured immediately by
gas chromatography. Through this analysis, levels of
benzaldehyde were maintained at relatively constant
values in the range of 0.8Ð4 g/l by controlled feeding.
As shown in Fig. 5, the highest level of L-PAC
(10.8 g/l) was achieved at 2 g/l benzaldehyde. At 4 g/l
benzaldehyde, 7.3 g/l L-PAC was obtained within the
relatively short period of 16 h. Further evaluation of
the kinetics, as summarized in Table 1, supports the
conclusion that increasing the level of benzaldehyde
(up to 2 g/l) resulted in higher L-PAC formation and
a relative reduction in benzyl alcohol formation. Both
L-PAC and benzyl alcohol production were inhibited
at 4 g/l benzaldehyde.
Besides the inhibiting e¤ect of benzaldehyde, it is
possible also that benzyl alcohol accumulation could
11Table 2 E¤ect of aeration rate
and respiratory quotient on
the biotransformation of
benzaldehyde to L-PAC
Aeration RQ range Maximum conc. L-PAC Benzyl alcohol Molar yield
rate (vvm) pyruvate (g/l) (g/l) (g/l) for L-PAC (%)
0.3 12Ð20 4.2 10.7 6.0 56.2
0.6 7Ð12 3.8 12.4 4.9 64.6
0.75 5Ð7 3.5 12.5 4.7 65.7
1.0 1Ð4 2.2 10.1 5.2 57.1
inhibit both L-PAC and benzyl alcohol formation.
When benzyl alcohol was above 6Ð7 g/l, there was little
further production of either benzyl alcohol or L-PAC,
indicating a degeneration of catalytic activity resulting
from continuous contact with both toxic substrate and
by-product.
E¤ect of respiratory quotient on L-PAC production
The metabolism of C. utilis is signiÞcantly a¤ected by
available oxygen, and the respiratory quotient (RQ)
has been used as a good indicator of the metabolic
status of the yeast. For normal respiratory growth,
RQ"1.0 is maintained with glucose as substrate while,
with increasing fermentation, RQ rises. Biotransforma-
tion of benzaldehyde to L-PAC involves the production
of CO
2
from decarboxylation of pyruvate to acetal-
dehyde and the extent of biotransformation may be
indicated by the RQ value, although the e¤ect is com-
plex with high RQ being related also to higher levels of
fermentation of glucose to ethanol.
The e¤ect of aeration rate and RQ on L-PAC
formation was evaluated with 2 g/l benzaldehyde level
in order further to identify critical parameters for the
biotransformation. Table 2 shows that a low aeration
rate of 0.3 vvm resulted in a high RQ value. Even
though this was accompanied by adequate pyruvate
accumulation, a relatively high level of benzyl alcohol
was formed, presumably because of higher levels of
ADH (and/or other oxidoreductases). By contrast, the
high aeration rate of 1.0 vvm resulted in a lower RQ
and lower L-PAC production, due to decreased accu-
mulation of pyruvate and reduced PDC activity. From
Table 2 it can be concluded that conditions for which
the RQ value was maintained between 5 and 7, are
likely to be the most favourable to maximise L-PAC
and minimize benzyl alcohol formation.
Biotransformation kinetics for L-PAC formation
A detailed biotransformation kinetic evaluation was
carried out with an immobilized cell density (cell dry
weight) of 15 g/l at 2 g/l benzaldehyde and 30 g/l glu-
cose pulse feeding (Fig. 6a). Aeration was controlled to
maintain the RQ value in its optimum range of 5Ð7,
temperature was controlled at 20¡ C and pH at 5.0. As
shown in Fig. 6b, L-PAC production occurred up to
15.2 g/l. The increased L-PAC formation compared to
the previous results can be ascribed to various factors,
such as programmed feeding of benzaldehyde at the
optimum level (2 g/l), RQ values maintained in the
range of 5Ð7, and pulse feeding of glucose to facilitate
pyruvate production (up to 8 g/l). ProÞles of enzyme
activities showed that the PDC activity declined
more rapidly with increasing reaction time than did
ADH (Fig. 6c). The PDC activity remained at about
0.65 unit/mg protein during the early stages of the
biotransformation, but declined to about 0.2 unit/
mg protein by the end, indicating that Þnal cessation
of L-PAC formation resulted from depletion of
pyruvate (with no further production) rather than
complete loss of PDC activity. The resultant decline in
benzaldehyde feeding proÞle is shown in Fig. 6d. At
the end of biotransformation, the cells appeared to
have lost metabolic activity completely (monitored by
CO
2
evolution) presumably because of the increasing-
ly inhibitory e¤ects of benzaldehyde and biotrans-
formation products.
Evaluation of a continuous immobilized cell
process for L-PAC production
A continuous process with immobilized C. utilis was
evaluated in a continuously stirred-tank reactor
(CSTR) with low-level aeration using immobilized cells
(approximate cell density"15 g/l). Prior to benzal-
dehyde addition, a continuous culture was established
at a dilution rate D"0.15 h~1 with 60 g/l glucose-
based medium, and maintained for a su¦cient period
to reach steady state.
As summarized in Table 3, 60 g/l of glucose was
converted basically to ethanol (27.1 g/l) and pyruvate
(2.1 g/l) prior to benzaldehyde addition. The enzyme
activities indicated a lower PDC activity than that
obtained in fed-batch culture. When benzaldehyde was
added to the medium, glucose utilization decreased, as
did the pyruvate and ethanol levels, and signiÞcant
inhibition of ADH and PDC activities was evident.
While the highest benzaldehyde feed rate of 1.5 ml/h
resulted in increased L-PAC production, a ÔÔpseudoÕÕ-
steady state was maintained only for 48Ð50 h.
With 1.0 ml/h feed rate, operation stability could be
12Fig. 6aÐd Biotransformation time course with immobilized cells: (a)
kinetics of (d) glucose consumption, (s) ethanol and (j) pyruvate
production; (b) biotransformation kinetics: j benzaldehyde, s benzyl
alcohol, (d) L-PAC; (c) enzyme proÞles, s alcohol dehydrogenase,
h pyruvate decarboxylase; (d) controlled feeding rate proÞle for
benzaldehyde
maintained for at least 110Ð120 h, however, the reduced
benzaldehyde resulted in benzyl alcohol production
exceeding L-PAC production.
From the data it is evident that a continuous L-
PAC biotransformation process with immobilized cells
at the higher benzaldehyde levels would have signiÞ-
cant di¦culties in long-term operation because of the
steady decline in PDC activity. This is likely to result
from continuous exposure to the toxic substrate and
the possible inhibition e¤ects of L-PAC and/or benzyl
alcohol. The result is consistent with that of Mahmoud
et al. (1990b) who reported signiÞcant inhibition e¤ects
for a semicontinuous process (operating for only a lim-
ited number of cycles) for L-PAC production using
immobilized S. cerevisiae.
Discussion
Previous studies (Mahmoud et al. 1990a, b) have sug-
gested that an immobilized cell process may o¤er signi-
Þcant advantages for a biotransformation involving
a toxic substrate. In the present investigation of the
biotransformation of benzaldehyde to L-PAC, several
interesting characteristics emerged. First, in a shake-
ßask comparison between free and immobilized cells it
was demonstrated that the immobilized cells could
tolerate higher initial benzaldehyde concentrations (up
to 70 mM, or 7.4 g/l) before substrate inhibition. For
free cells, the level was 50 mM (or 5.3 g/l), indicating
that the substrate proÞles resulting from benzaldehyde
di¤usion into the calcium alginate beads have provided
13Table 3 Kinetic parameters of
continuous process with
immobilized cells in 1.5
l controlled fermenter at 20¡ C
and pH 6.0. (BZ benzaldehyde,
BA benzyl alcohol, PDC
pyruvate decarboxylase, ADH
alcohol dehydrogenase)
BZ feed Concentration (g/l)
rate
(ml/h) S
*/
S
065
P P S
065
P P
(glucose) (glucose) (EtOH) (pyruvate) (BZ) (L-PAC) (BA)
0 60 1.35 27.1 2.1 0 0 0
0.5 60 10.5 23.5 0.9 0.1 0.7 1.6
1.0 60 16.9 19.5 0.6 2.2 2.3
1.5 60 35.2 10.1 0.4 0.45 4.0 3.6
Kinetic values (g/g/h) Activities (mg protein)
!q
(BZ)
q(L-PAC)
q
(BA)
Productivity PDC ADH
of L-PAC
(g/l/h)
0 0 0 0 0.47 0.94
0.021 0.007 0.016 0.11 0.40 0.87
0.040 0.022 0.023 0.33 0.30 0.85
0.063 0.040 0.036 0.60 0.26 0.71
some protection against substrate inhibition. Second, it
was demonstrated that the yield of L-PAC compared to
the production of the major by-product, benzyl alco-
hol, was dependent on the benzaldehyde concentration
and the available oxygen in the microenviroment (as
measured by RQ). Higher benzaldehyde levels favoured
L-PAC production, a result consistent with the obser-
vation by Long and Ward (1989) that the PDC of
S. cerevisiae was more resistant to benzaldehyde than
was ADH (and presumably other oxidoreductases).
Highly fermentative conditions, such as high RQ, were
less favourable to L-PAC production and an optimum
RQ range of 5Ð7 (partially aerobic) was identiÞed. The
results demonstrated also that respiratory metabolism
(RQ"1Ð4) resulted in a marked reduction in L-PAC
production presumably because of low PDC activity.
Finally it was established, with optimal control of ben-
zaldehyde and RQ levels, that an L-PAC concentration
of 15.2 g/l could be achieved in 22 h in a fed-batch
culture. This compares with 10 g/l L-PAC produced by
immobilized S. cerevisiae in a semi-continuous process
(Mahmoud et al. 1990b), and reßects the capacity of C.
utilis as a suitable yeast for biotransformation of ben-
zaldehyde, as well as the optimal conditions used.
A continuous immobilized-cell process with C. utilis
was evaluated also, and found to produce low L-PAC
levels (no more than 4 g/l in sustained operation). Such
a process would be unsuitable for the biotransforma-
tion of toxic substrates such as benzaldehyde.
References
Astrup A, Breum L, Toubro S, Hein P, Quaade F (1992a) The e¤ect
and safety of an ephedrine-ca¤eine compound compared to
ephedrine, ca¤eine and placebo in obese subjects on an energy
restricted diet: a double blind trial. Int J Obesity 16:269Ð277
Astrup A, Buemann B, Christensen NJ, Toubro S, Thoebek G,
Victor OJ, Quaade F (1992b) The e¤ect of ephedrine/ca¤eine
mixture on energy expenditure and body composition in obese
women. Metabolism 41: 686Ð688
Bergmeyer HU (1974) Methods of enzymatic analysis, vol. 1. Aca-
demic Press New York and London, pp. 429Ð509
Bradford MM (1970) A rapid and sensitive method for the quantiÞ-
cation of microgram quantities of protein, utilization of the princi-
ple of protein dye binding. Anal Biochem 72:248Ð254
Bringer-Meyer S, Sahm H (1988) Acetoin and phenylacetylcarbinol
formation by the pyruvate decarboxylase of Zymomonas mobilis
and Saccharomyces carlsbergensis. Biocatalysis 1:321Ð331
Culic K, Netrval J, Souhadra J, Ulbrecht S, Vojtisek V, Vodnansky
MM (1984) Method of cost reduction in the production of
D-(!)-1-phenyl-1-hydroxy-2-propane for the production of L-
(!)ephedrine. Czechoslovakian patent 22,941
Long A, Ward OP (1989) Biotransformation of benzaldehyde by
Saccharomyces cerevisiae: characterization of the fermentation
and toxicity e¤ects of substrates and products. Biotechnol Bio-
eng 34:933Ð941
Mahmoud WM, El-Sayed AdH MM, Coughlin RW (1990a) Pro-
duction of L-phenylacetylcarbinol by immobilized yeast cells. I.
Batch fermentation. Biotechnol Bioeng 36:47Ð54
Mahmoud WM, El-Sayed AdH MM, Coughlin RW (1990b) Pro-
duction of L-phenylacetylcarbinol by immobilized yeast cells. II.
Semicontinuous fermentation. Biotechnol Bioeng 36:55Ð63
Nikolova P, Ward OP (1991) Production of L-phenylacetylcarbinol
by biotransformation: product and by-product formation and
activities of the key enzymes in wild-type and ADH iso-enzyme
mutants of Saccharomyces cerevisiae. Biotechnol Bioeng
38:493Ð498.
Seely RJ, Hefner DL, Hageman RW, Yarus MJ, Sullivan SA (1989)
Process for making phenylacetylcarbinol (PAC), microorganisms
for use in the process and a method for preparing the microor-
ganisms. US Patent, PCT/US 89/04423
Vojtisek V, Netrval J (1982) E¤ect of PDC activity and pyruvate
concentration on production of 1-hydroxy-1-phenylpropanol in
Saccharomyces carlsbergensis. Folia Microbiol 27:173Ð177
