ORIGINAL PAPER

J. Simmonds á G. K. Robinson

Formation of benzaldehyde by

Pseudomonas putida

ATCC 12633

Received: 16 March 1998 / Received revision: 20 May 1998 / Accepted: 21 May 1998

Abstract Aromatic and heterocyclic aldehydes may be

produced by the mandelate pathway of Pseudomonas

putida ATCC 12633 via the biotransformation of ben-

zoyl formate and substrate analogues. Under optimised

biotransformation conditions (37 °C, pH 5.4) and with

benzoyl formate as a substrate, benzaldehyde may be

accumulated with yields above 85%. Benzaldehyde is

toxic to P. putida ATCC 12633; levels above 0.5 g/l

(5 mM) reduce the biotransformation activity. Total

activity loss occurs at an aldehyde concentration of

2.1 g/l (20 mM). To overcome this limitation, the rapid

removal of the aldehyde is desirable via in situ product

removal. The biotransformation of benzoyl formate

(working volume 1 l) without in situ product removal

accumulates

2.1 g/l

benzaldehyde.

Benzaldehyde

removal by gas stripping produces a total of 3.5 g/l be-

fore inhibition. However, the most ecient method is

solid-phase adsorption using activated charcoal as the

sorbant, this allows the production of over 4.1 g/l

benzaldehyde. Addition of bisulphite as a complexing

agent causes inhibition of the biotransformation and

bisulphite is therefore is not suitable for in situ product

removal.

Introduction

Aromatic aldehydes are commercially signi®cant com-

pounds with applications in the fragrance, ¯avour and

pharmaceutical industries. Biotransformations to pro-

duce these compounds are of interest as the products

may be described as natural and exist in a relatively pure

form, because of the high speci®city of enzyme reactions.

The production of aromatic aldehydes is limited by

the toxicity of the products, therefore the rapid removal

or sequestration of the aldehyde by in situ product re-

moval (Freeman et al. 1993) is desirable. Such tech-

niques previously used for aldehyde removal include gas

stripping (Wecker and Zall 1987), pervaporation (Lamer

et al. 1996), solid-phase adsorption (Berger 1995),

aqueous-organic two-phase systems (Du€ and Murry

1989) and condensation reactions (Shacher-Nishri and

Freeman 1993). In situ product removal reduces the

exposure of the bacteria to the aldehyde product and

therefore decreases the toxic e€ect to the organism.

This paper discusses benzaldehyde production by

Pseudomonas putida ATCC 12633 with a bench-scale

bioreactor for the production of benzaldehyde and in-

dicates the problems encountered when this method is

used, including toxicity of the product and methods of

product removal.

Materials and methods

Growth and maintenance of bacteria

Pseudomonas putida ATCC 12633 (NCIMB 9494), also designated

P. ¯uorescens A.3.12 (Stanier 1947), was maintained on nutrient

agar plates and subcultured bi-weekly. Pseudomonas minimal

medium (PMM), as described by Hegeman (1966), was used for the

liquid culture of the bacteria. Sodium succinate or glucose (5 mM)

was used as a carbon and energy source for uninduced bacteria (no

active mandelate pathway). For bacteria with an active pathway,

racemic mandelic acid (5 mM) was used as the growth substrate.

Appropriately induced bacteria (overnight seed culture grown in

50 ml PMM + 0.5% yeast extract + 5 mM carbon source) were

grown for biotransformation in 2-l ¯asks containing PMM

(500 ml) and mandelic acid (5 mM) and incubated at 30 °C,

200 rpm, to mid-exponential phase (3 h).

Whole-cell biotransformation protocol

Washed bacteria were resuspended to 0.5 g l

)1

in McIlvaine (ci-

trate/phosphate) bu€er (0.1 M, pH 5.4) and equilibrated to 37 °C

before the addition of benzoyl formate, or an alternative substrate.

Cells were reciprocally shaken (100 rpm) and samples were taken

Appl Microbiol Biotechnol (1998) 50: 353±358

Ó Springer-Verlag 1998

J. Simmonds á G. K. Robinson (&)

Department of Biosciences, University of Kent at Canterbury,

Canterbury, Kent, CT2 7NJ, England

Tel.: [phone], extension 3530

Fax: [phone]

over 3 h, the bacteria were removed by centrifugation (10 000 rpm,

2 min) and the resulting supernatant was analysed by HPLC.

Benzaldehyde was removed from the biotransformations at this

scale (250 ml, working volume 150 ml) by gas stripping with an

aeration rate of 2.6 vvm. It was reclaimed by passage through two

hexane solvent traps (100 ml). Polytetra¯uoroethylene tubing was

used throughout, as benzaldehyde was shown to be adsorbed by

rubber and silicone.

Large-scale biotransformations

P. putida ATCC 12633 was grown in an Applikon AD1-1012 1.5-l

aerated fermentation vessel (working volume 1 l). The vessel was

heated by an infra-red lamp, connected to a Systag TC1-88L tem-

perature controller, and pH was monitored by a Mostec pH regu-

lator. PMM was autoclaved within the vessel with 0.5 ml

polypropyleneglycol as antifoam; mandelic acid (30 mM) was

sterilized separately. An additional 20 ml minerals supplement

(Hegeman 1966) was also included. The seed culture was grown for

15 h (30 °C, 200 rpm) with yeast extract (0.5%) and mandelic acid

(5 mM) as the carbon and energy source. The reactor was equili-

brated to 30 °C, pH 6.8, agitation 500 rpm, aeration 0.75 vvm,

before inoculation with an overnight seed culture. P. putida ATCC

12633 was grown under these conditions for 15 h. Biotransforma-

tions (1 l) were performed in the growth medium. The accumulation

of benzaldehyde in these large-scale biotransformations was initi-

ated by adjustment of conditions to pH 5.4, 37 °C, 0.2 vvm aeration

and 200 rpm agitation. The organism was allowed to equilibrate to

this environment for 30 min before the addition of substrate.

A number of di€erent methods of benzaldehyde removal were

tested.

Gas stripping

Enhanced aeration at the 1L scale was generated by aeration of the

vessel at 0.8 vvm with agitation at 750 rpm.

Bisulphite addition

Sodium pyrosulphite was added directly to the reactor at intervals

after the formation of 5 mM benzaldehyde (below toxic levels). A

benzaldehyde-bisulphite complex was formed immediately, there-

fore no free bisulphite remained in the reactor. At the end of the

reaction the biomass was separated by centrifugation (10 000 g,

10 min), the complex was precipitated from the supernatant by the

addition of excess bisulphite. The benzaldehyde was regenerated

from the precipitate by the addition of NaOH (®nal concentration

1 M), with a recovery eciency of 85%.

Solid-phase adsorption

Granular activated charcoal, mesh 4±8 mm (Fluka) was added

directly to the reactor (25 g l

)1

), prior to the addition of the bio-

transformation substrate. Sucient stirring of the vessel was

achieved by 0.2 vvm aeration, therefore agitation was discontinued

to prevent attrition of the charcoal pieces. The benzaldehyde was

extracted from the adsorbant with ethyl acetate.

HPLC analysis

A Shimadzu LC-6A HPLC with a UV spectrophotometric detec-

tor, 10-ll loop and a Shimadzu CRGA Chromatopac integrator

was used for all analyses in combination with a 4.6 ´ 250-mm

Spherisorb ODS2 column with 5 lm packing (Jones Chromatog-

raphy, Mid Glamorgan). The mobile phase was 47% acetonitrile,

0.2% orthophosphoric acid in water, with a ¯ow rate of 1 ml/min.

The detection wavelength was 235 nm. All samples were compared

to authentic standards to calculate concentrations.

Relative toxicity of benzaldehyde

A method was developed to compare the toxicity of a range of

aldehydes. P. putida ATCC 12633 was grown on a number

of concentrations of di€erent aromatic aldehydes in the presence of

yeast extract (0.5%). The critical concentration was de®ned as the

aldehyde concentration at which the cell density at the stationary

phase was half that achieved when the organism was grown on

yeast extract alone.

Measurement of short-term toxicity

1. The e€ect of benzaldehyde on the benzaldehyde dehydrogenase

isoenzymes was measured at pH 7.0 as these enzymes were not

active at pH 5.4. Bacteria were exposed to di€erent benzalde-

hyde concentrations for 10 min in NaCl (0.9%) at pH 7.0 and

37 °C. The benzaldehyde dehydrogenase activity was measured

by calculating the rate of benzoic acid production, using a

Mettler DL21 titrator.

2. The e€ect of benzaldehyde on the benzoylformate decarboxyl-

ase enzyme was then measured. Conditions in the bioreactor

were mimicked by incubating a suspension of bacteria with

di€erent concentrations of benzaldehyde at pH 5.4 and 37 °C.

Benzoyl formate was added after 5 min and decarboxylase ac-

tivity was measured by the rate of substrate utilisation.

Results

Toxicity of aldehydes

The ability of P. putida ATCC 12633 to grow on benz-

aldehyde was used as an indication of long-term toxicity.

Growth was in PMM at pH 7.0, as at pH 5.4 dehydro-

genase activity was reduced and this would prevent

utilization of the carbon source. The organism was un-

able to grow with benzaldehyde as the sole carbon and

energy source at concentrations above 4 mM. This toxic

e€ect was shown to be related to inhibition of the

pathway by which benzaldehyde was metabolised by

growth on yeast extract (0.5%) in the presence of

benzaldehyde. Toxic e€ects, i.e. a slower growth rate

and a reduction in ®nal cell density, were observed above

5 mM; there was no growth on yeast extract (0.5%) in

the presence of 8 mM benzaldehyde.

Under biotransformation conditions, low levels of

benzaldehyde (below 5 mM) caused an increase in ac-

tivity of the benzaldehyde dehydrogenase enzymes as

previously indicated (Simmonds and Robinson 1997).

Exposure to concentrations between 10 mM and 30 mM

for 10 min had little e€ect on the biotransformation

ability of P. putida ATCC 12633 (Fig. 1a). However, at

concentrations of benzaldehyde above >35 mM there

was no further conversion after 10 min exposure.

The toxicity of benzaldehyde relative to that of other

aldehydes was measured. A range of aromatic and het-

erocyclic aldehydes were tested (Table 1) and indicated

that the toxicity of benzaldehyde was within the range of

other aromatic aldehydes.

Benzoylformate decarboxylase activity was increased

by 50% in P. putida ATCC 12633 exposed to 20 mM

benzaldehyde, compared to the control, which had no

354

exposure (Fig. 1b). However, bacteria were inhibited by

the addition of more than 30 mM benzaldehyde 5 min

before the addition of substrate. The experiment was

repeated with the bacteria being exposed to benzalde-

hyde for 2 h before the measurement of enzyme activity;

there was no increase in activity of the benzoyl formate

decarboxylase in the presence of low concentrations of

benzaldehyde. The rate of biotransformation of benzoyl

formate decreased with increasing benzaldehyde con-

centration (Fig. 1b).

Substrate inhibition

Growth of P. putida ATCC 12633 on benzoyl formate at

concentrations above 20 mM was at a reduced rate,

causing lower cell densities after 4 h. No growth beyond

that achieved with no carbon source added was shown

with a substrate concentration of 160 mM, indicating

total inhibition. In the presence of yeast extract (0.5%)

and benzoyl formate, a similar growth pro®le was ob-

tained with concentrations above 10 mM inhibiting

growth, although to a lesser extent than in the absence of

yeast extract.

The biotransformation activity of P. putida ATCC

12633 was also a€ected by high benzoyl formate con-

centrations. A suspension of bacteria presented with a

high (above 75 mM) concentration of benzoyl formate

had an initial biotransformation activity the same as

those challenged with a low (below 10 mM) concentra-

tion. However, after 10 min incubation with the high

concentration of substrate, there was no further bio-

transformation and no more benzaldehyde was pro-

duced (data not shown).

Production of benzaldehyde

A suspension of P. putida ATCC 12633, grown on

mandelate (5 mM) was used for the fed-batch bio-

transformation of benzoyl formate. The substrate was

fed at a constant rate and further benzoyl formate was

added on substrate exhaustion; product removal was by

gas stripping (Fig. 2a). The total amount of benzalde-

hyde (Fig. 2b) was distributed between the reactor and

the two solvent traps, which were placed in series.

Consequently, benzaldehyde was produced to levels of

3.5 g l

)1

over 3 h with a productivity of 1.81 g benzal-

dehyde h

)1

g dry weight

)1

with fed-batch feeding of the

substrate, 2.6 vvm aeration and polypropylene glycol

(0.03%). However, after this time the rate of biotrans-

formation of benzoyl formate was reduced.

Process optimization

The above results indicated that both benzoyl formate

and benzaldehyde levels should be kept low during the

biotransformation, to prevent either substrate or prod-

uct inhibition. Di€erent feeding regimes in fed-batch

culture were investigated, to ®nd which gave the most

ecient conversion to benzaldehyde. These included 5,

10 and 40 mM spikes at de®ned intervals and continu-

ous substrate addition (Table 2). When 10 mM spikes of

benzoyl formate were added at 30-min intervals, the

Fig. 1 Biotransformations P. put-

ida ATCC 12633 after incubation

with benzaldehyde. (a) Benzalde-

hyde dehydrogenase activity mea-

sured after 10 min exposure to

benzaldehyde at pH 7.0, 37 °C by

biotransformation of benzalde-

hyde. (b) Benzoylformate decar-

boxylase activity measured after

5 min and two hrs exposure to

benzaldehyde at pH 5.4, 37 °C by

biotransformation of benzoylfor-

mate

Table 1 The relative toxicity of a number of aromatic and het-

erocyclic aldehydes

Aldehyde

Critical concentration (mM)

p-Tolualdehyde

11.1

Thiophene-3-carboxaldehyde

7.2

Benzaldehyde

7.1

Pyridine-3-carboxaldehyde

5.1

Thiophene-2-carboxaldehyde

1.3

355

yield was 40%, with 33% of the substrate remaining.

The frequency of additions was decreased to every

40 min, consequently the yield increased to 82%, and

only 1% of the substrate was not biotransformed.

However, as less substrate can be added within the same

period, the amount of benzaldehyde formed is reduced

and thus productivity decreases.

The continuous addition of benzoyl formate, at a rate

equivalent to that of the biotransformation, gave the

highest molar yield of benzaldehyde. However, the bio-

transformation activity of benzoylformate decarboxyl-

ase was reduced after 2.5 h. The decrease in

biotransformation rate occurred after an aldehyde con-

centration of 17 mM (‹2 mM) has been reached within

the reactor. A more ecient method of aldehyde re-

moval was required to prevent product inhibition and

extend the viability of the biotransformation. This was

studied on a larger scale.

Larger-scale benzaldehyde production

Growth of P. putida ATCC 12633 in a bioreactor (1.5 l)

with mandelic acid as the sole carbon and energy source

gave a cell concentration of 1.9 g dry weight/l over a

15-h period. The growth rate was initially rapid, with a

doubling time of 60 min, and then slowed as the culture

became oxygen-limited. The rate of biotransformation

of benzoyl formate by P. putida ATCC 12633 was 30%

less after a period of oxygen starvation, compared to the

activity during exponential growth.

The biotransformation of benzoyl formate was car-

ried out without removal of the cells from the growth

medium. The accumulation of benzaldehyde was initi-

ated by reduction in pH to 5.4 and an increase in tem-

perature to 37 °C.

A biotransformation (1 l) in growth medium, with no

removal of benzaldehyde, produced 2.17 g l

)1

benzal-

dehyde over 3 h (a productivity of 0.41 g benzalde-

hyde h

)1

g dry weight

)1

). The reaction stopped after the

benzaldehyde concentration had reached 20 mM, indi-

cating inhibition of the synthetic pathway because of the

toxicity of the benzaldehyde. A more ecient method of

benzaldehyde removal was required to maintain the

product at low levels within the biotransformation.

Benzaldehyde removal

Gas stripping was tested as a method of benzaldehyde

removal at the 1-l scale, with a combination of aeration

at 0.8 vvm and agitation at 750 rpm. The rate of

Fig. 2 Biotransformation of ben-

zoylformate by Pseudomonas put-

ida ATCC 12633 grown with

mandelic acid (5 mM) as a carbon

and energy source. Reaction in

McIlvaine bu€er at 37 °C, pH 5.4

with removal of benzaldehyde by

enhanced aeration, 2.6v/v/min

Table 2 E€ect of feeding regime on the eciency and yield of benzaldehyde production from benzoyl formate by Pseudomonas putida

ATCC 12633: 3 h biotransformation at 37 °C, pH 5.4, 2.6 vvm aeration, polypropyleneglycol (0.03%)

Benzoyl formate

spike size (g l

)1

)

Total number

of spikes

Interval between

spikes (min)

Remaining benzoyl

formate (g l

)1

)

Total benzaldehyde

Produced (g l

)1

)

Molar yield

benzaldehyde (%)

7.95

1

±

9.45

0.80

10

4.24, 2.12

1+1

60

0.90

4.64

65

1.06

6

30

3.00

3.57

40

1.06

4

40

0.07

3.50

82

0.53

9

20

0.00

3.39

71

Continuous addition, rate: 1.33 g l

)1

h

)1

0.02

3.28

82

356

removal from this type of reactor was much slower than

from the 150-ml reactor used previously. Only 55% of a

5 mM solution of benzaldehyde was removed over a 3-h

period, compared to 77% from an equivalent concen-

tration at the smaller scale. The total benzaldehyde

produced was 1.5 g l

)1

, with a yield of 78%.

Bisulphite addition was also examined as a method of

in situ product removal. However, the addition of

bisulphite to a biotransformation (1 l) of benzoyl for-

mate caused immediate cessation of the biotransforma-

tion. No further substrate was metabolised and no more

benzaldehyde was formed. The relative toxicity factors

for bisulphite and the bisulphite-benzaldehyde addition

complex were 6 mM and 12 mM respectively. Thus,

bisulphite was as toxic as benzaldehyde to P. putida

ATCC 12633 and the adduct only slightly less toxic.

Solid-phase adsorption was the ®nal method of al-

dehyde removal tested. Activated charcoal was used in

the bioreactor as a model sorbant, as it was non-toxic

and cheaply available in variable mesh sizes. The sub-

strate, benzoyl formate, did not sorb to the activated

charcoal. The total benzaldehyde, after extraction, was

compared to that produced in an equivalent biotrans-

formation with no activated charcoal present (Fig. 3).

As seen previously, the concentration of benzaldehyde

produced by the standard biotransformation reached a

plateau when the free benzaldehyde concentration in the

reactor reached 2.3 g l

)1

(20 mM). With activated

charcoal the biotransformation continued beyond this

point, as the free benzaldehyde concentration remained

below 5 mM, even when the total benzaldehyde, in-

cluding adsorbed compound, was above 40 mM. The

rate of benzaldehyde production was constant to the

termination of the biotransformation at 3 h. The total

benzaldehyde produced was 4.8 g l

)1

, with a molar yield

of 88% and a productivity of 0.82 g benzaldehyde h

)1

g

dry weight

)1

. Ethyl acetate was used to recover the

benzaldehyde from the sorbant with a recovery eciency

of 39%.

Discussion

The partial inactivation of the mandelate pathway

benzaldehyde dehydrogenase isoenzymes of P. putida

ATCC 12633 at pH 5.4 allows the accumulation of

benzaldehyde. In this study in situ product removal de-

creased the in¯uence of aldehyde toxicity and also pre-

vented the enzymic conversion of the aldehyde to

benzoic acid by separating the compound from the

biomass.

P. putida ATCC 12633 was unable to grow on high

(above 8 mM) concentrations of benzaldehyde with or

without yeast extract (0.5%), indicating that benzalde-

hyde toxicity was a general phenomenon unrelated to

the speci®c pathway for metabolism of the compound.

This long-term toxicity also probably contributed, in

part, to the 15% lower activity seen after 2 h exposure to

concentrations of benzaldehyde above 10 mM. The loss

of activity was proportional to the increase in benzal-

dehyde concentration. There was no reduction of ac-

tivity in cells that had not been exposed to the aldehyde.

However, there was a more signi®cant e€ect on the

biotransformation rate in the reactor; after 2 h exposure

there was no activation associated with the low con-

centrations of benzaldehyde (Simmonds and Robinson

1997). Thus, there was a 60% di€erence in the bio-

transformation rate between cells that had been exposed

to 10 mM benzaldehyde for 5-min and those exposed for

2 h.

Benzoyl formate decarboxylase was a€ected by lower

concentrations of benzaldehyde than the benzaldehyde

dehydrogenase isoenzymes. Therefore, high concentra-

tions of benzaldehyde (above 20 mM) were avoided in

the bioreactor, as this would prevent the rapid bio-

transformation of benzoyl formate, although benzalde-

hyde was oxidised to benzoic acid at higher

concentrations (below 30 mM). The ``deactivation'' of

pyruvate decarboxylase, a related enzyme, by benzal-

dehyde has been demonstrated (Chow et al. 1995). The

process was shown to be ®rst-order with respect to

benzaldehyde concentration, in the range 100±300 mM,

with a square-root dependence on time.

Benzoyl formate was also shown to be inhibitory at

levels above 20 mM. Therefore, the maximum benzoyl

formate concentration in the reactor was restricted by

the fed-batch feeding of substrate; with substrate added

to a level of 10 mM at 40-min intervals, a molar yield of

benzaldehyde of 82%, with a concentration of 3.5 g l

)1

was achieved in 3 h. This gave a productivity of 1.81 g

benzaldehyde h

)1

g dry cell weight

)1

. Therefore, if the

feeding regime and the rate of removal of benzaldehyde

are balanced to reduce the toxic and inhibitory e€ects to

a minimum, this method may be used for the ecient

production of aromatic aldehydes.

Fig. 3 Benzaldehyde production by Pseudomonas putida ATCC

12633 from benzoylformate. E€ect of benzaldehyde sequestration by

activated charcoal on the total benzaldehyde produced. Experiment

performed in duplicate, error bars show the range

357

Gas stripping was an e€ective method of in situ

benzaldehyde removal at the 150-ml scale, raising the

total amount of benzaldehyde produced from 2.1 g l

)1

to 3.5 g l

)1

. However, the method did not remove the

aldehyde rapidly enough at the 1L scale. This was partly

because of the relative reduction in the amount of air

passing through the medium, 2.6 vvm in the small re-

actor to 0.8 vvm on the larger scale. Increasing the air-

¯ow would increase the eciency of the removal but

would also become progressively more dicult and ex-

pensive to scale-up further. Gas stripping also caused a

decrease in productivity; this was related to slight ¯oc-

culation of the organism exacerbated by the enhanced

aeration at pH 5.4 (data not shown).

Aldehyde sequestration by bisulphite was also tested

as a method of in situ product removal. However, the

biotransformation stopped immediately the bisulphite

was added and further experiments showed that

bisulphite was more toxic than benzaldehyde to the

biotransformation. Therefore the biomass must be sep-

arated from the bisulphite sequestration stage. This

separation could be achieved in a hollow-®bre reactor,

or a recycling system, where the biomass is returned to

the reactor and the benzaldehyde removed from the

aqueous phase by bisulphite, either free or immobilised.

This step would also act as a concentration stage, be-

cause a saturated solution of bisulphite would result in

the formation of an easily separated, insoluble, benzal-

dehyde-bisulphite adduct.

Solid-phase adsorption was the most ecient

method of in situ product removal tested. The rapid

rate of benzaldehyde extraction from the aqueous

phase prevented build-up of the compound to toxic

levels, and the activated charcoal did not appear to

a€ect the eciency of the reaction. However, for this

approach to be useful a di€erent method of recovering

the benzaldehyde from the solid phase must be devel-

oped. The method could also be improved by use of a

di€erent solid-phase adsorbant. A potentially more ef-

®cient method of extraction would use an external

vessel containing the solid phase, through which the

biotransformation mixture could be pumped. This

would allow for change of adsorbant on blockage or

saturation without interrupting the biotransformation

(Berger 1995). The rate of removal of product may also

be controlled by the size or number of extraction

vessels. A diculty with this approach is the rapid rate

of recycling required to maintain the correct conditions

in the medium during the recycling process, any vari-

ation of pH or temperature would reduce yields from

the biotransformation.

The mandelate pathway of P. putida ATCC 12633

was used to produce benzaldehyde by biotransforma-

tion. The reaction gave good yields and, with e€ective

product removal, was not terminated by aldehyde tox-

icity. The major limitation of this technique for the

production of aromatic aldehydes was the diculty of

obtaining substrates with an aromatic or heterocyclic

ring and an appropriate glyoxylate moiety. However, the

range could be signi®cantly increased by the methyl

group oxidation of acetophenone derivatives. This re-

action is currently under study.

Acknowledgements We are grateful to Lonza Ltd. and BBSRC for

funding this research and to Dr. Olwen Birch for assistance during

the work at Lonza Ltd., Visp, Switzerland.

References

Berger R (1995) Bioprocess technology. In: Aroma biotechnology.

Springer Berlin Heidelberg New York, pp 139±162

Chow YS, Shin HS, Adesina AA, Rogers PL (1995) A kinetic-

model for the deactivation of pyruvate decarboxylase (PDC) by

benzaldehyde. Biotechnol Lett 17: 1201±1206

Du€ JB, Murray WD (1989) Oxidation of benzyl alcohol by a

whole cells of Pichia pastoris and by alcohol oxidase in aqueous

and non aqueous reaction media. Biotechnol Bioeng 34: 153±

159

Freeman A, Woodley JM, Lilly MD (1993) In situ product removal

as a tool for bioprocessing. Biotechnology 11: 1007±1011

Hegeman GD (1966) Synthesis of the enzymes of the mandelate

pathway by Pseudomonas putida. I. Synthesis of the enzymes of

the wild type. J Bacterial 91: 1140±1154

Lamer T, Spinnler HE, Souchon I, Voilley A (1996) Extraction of

benzaldehyde from fermentation broth by pervaporation. Pro-

cess Biochem 31: 533±542

Shacher-Nishri Y, Freeman A (1993) Continuous production of

acetaldehyde by immobilized yeast with in situ product trap-

ping. Appl Biochem Biotechnol 39: 387±399

Simmonds J, Robinson GK (1997) Novel biotransformations to

produce aromatic and heterocyclic aldehydes. Enzyme Micro-

biol Technol 20: 367±374

Stanier RY (1947) Simultaneous adaptation: a new technique for

the study of metabolic pathways. J Bacteriol 54: 339±348

Wecker MA, Zall RR (1987) Fermentation strategies: acetaldehyde

or ethanol? Process Biochem 22: 135±138

358