Biocatalytic oxidation of 2-methylquinoxaline to 2-quinoxalinecarboxylic acid

Article abstract of DOI:10.1021/op025501e

A microbial process using the fungus Absidia repens ATCC 14849 is described for the oxidation of 2-methylquinoxaline to 2-quinoxalinecarboxylic acid. A campaign consisting of three 14000-L runs produced 20.5 kg of the acid with a 28% overall yield. The bioconversion gave a lower yield compared with a three step chemical synthesis (35%), but was carried out in one pot, and avoided safety issues with a di-N-oxide intermediate. Although successfully scaled to produce kilograms of 2-quinoxalinecarboxylic acid for synthesis of a drug candidate, the A. repens bioconversion is unsuitable for further scale-up due to low product concentration (～1 g/L). A second microbial process using Pseudomonas putida ATCC 33015 is also described for the oxidation of 2-methylquinoxaline. The P. putida bioconversion gave an 86% in situ yield at 8-L scale and yielded a product concentration approximately 10-fold greater than that of the A. repens bioconversion.

Full text of DOI:10.1021/op025501e

Organic Process Research & Development 2002, 6, 477−481
Biocatalytic Oxidation of 2-Methylquinoxaline to 2-Quinoxalinecarboxylic Acid
John W. Wong,*,† Harry A. Watson, Jr.,‡ James F. Bouressa,† Michael P. Burns,† James J. Cawley,† Albert E. Doro,†
Donald B. Guzek,† Michael A. Hintz,† Ellen L. McCormick,† Douglas A. Scully,† Joseph M. Siderewicz,†
William J. Taylor,† Susan J. Truesdell,† and Richard G. Wax†
Bioprocess Research and DeVelopment and Chemical Research and DeVelopment, Pfizer Global Research and
DeVelopment, Eastern Point Road, Groton, Connecticut 06340, U.S.A.
Scheme 1. Chemical synthesis of 4
Abstract:
A microbial process using the fungus Absidia repens ATCC
14849 is described for the oxidation of 2-methylquinoxaline to
2-quinoxalinecarboxylic acid. A campaign consisting of three
14000-L runs produced 20.5 kg of the acid with a 28% overall
yield. The bioconversion gave a lower yield compared with a
three step chemical synthesis (35%), but was carried out in one
pot, and avoided safety issues with a di-N-oxide intermediate.
Although successfully scaled to produce kilograms of 2-quin-
oxalinecarboxylic acid for synthesis of a drug candidate, the
A. repens bioconversion is unsuitable for further scale-up due
to low product concentration (∼1 g/L). A second microbial
process using Pseudomonas putida ATCC 33015 is also de-
scribed for the oxidation of 2-methylquinoxaline. The P. putida
bioconversion gave an 86% in situ yield at 8-L scale and yielded
a product concentration approximately 10-fold greater than that
of the A. repens bioconversion.
Scheme 2. Biocatalytic oxidation of 5 to 4
of 4. A recent report describing the aerobic oxidation of
methylpyridines to pyridinecarboxylic acids in the presence
of Co(II), Mn(II), and a radical catalyst suggested that 4 could
be produced by oxidation of 2-methylquinoxaline 5.5 This
chemistry was not evaluated due to our lack of facilities for
performing reactions in an oxygen atmosphere or under high
pressure.
Introduction
2-Quinoxalinecarboxylic acid 4 is used in the synthesis
of a variety of biologically active compounds.1 A recent
project requiring kilograms of 4 prompted our investigation
of scalable methods, since 4 was not commercially available
in kilogram quantities at the time. Compound 4 is also a
metabolite of Carbadox, a feed additive used to promote
growth and control dysentery and bacterial enteritis in swine.2
During investigations conducted in the 1970s, we prepared
4 from the di-N-oxide 1, via the route shown in Scheme 1.
While this route was demonstrated at 25 kg scale (35%
overall yield), this synthesis was deemed unsuitable for
further scale-up due to the mutagenic3 and thermal4 properties
of 1.
In addition to synthetic approaches, we also examined
biocatalytic methods for the oxidation of 5 to 4. Biocatalytic
oxidation of 5 would likely occur via the sequence shown
in Scheme 2, based on known pathways for aromatic methyl
group oxidation in bacterial and fungal systems. In the
biosynthesis of patulin by the fungus Penicillium griseoful-
Vum, transformation of m-cresol to m-hydroxybenzoic acid
occurs in three steps, beginning with hydroxylation of the
methyl group and followed by sequential oxidation of the
alcohol and aldehyde intermediates.6 The bacterium Pseudo-
monas putida ATCC 33015 can grow on a variety of
methylated aromatic compounds, including toluene, m-
xylene, and p-xylene, as sole carbon and energy source, by
virtue of a degradative pathway that also begins with
hydroxylation of an aromatic methyl group followed by
oxidation to the carboxylic acid.7 P. putida ATCC 33015
has also been shown to oxidize methyl groups in a variety
of five- and six-membered heteroaromatic compounds, and
These safety issues prompted us to examine alternative
chemical methods, none of which led to a suitable synthesis
* To whom correspondence should be addressed. E-mail: john_w_wong@
groton.pfizer.com.
† Bioprocess Research & Development.
‡ Chemical Research & Development.
(1) (a) Marquis, R. W., Jr.; Veber, D. F. PCT Int. Appl. WO Patent 2001034600,
2001. (b) Van Wagenen, B. C.; Moe, S. T.; Smith, D. L.; Sheehan, S. M.;
Shcherbakova, I.; Travato, R.; Walton, R.; Barmore, R.; Delmar, E. G.;
Stormann, T. M. PCT Int. Appl. WO Patent 9926927, 1999. (c) Yabe, Y.;
Watanabe, T.; Nishigaki, T.; Ozawa, Y.; Komai, T.; Nakagawa, A. JP Appl.
Jpn. Kokai Tokkyo Koho, JP Patent 09328428, 1997.
(2) (a) Johnston, J. D. U.S. Patent 3,371,090, 1968. (b) Johnston, J. D. U.S.
Patent 3,433,871, 1969.
(5) Shibamoto, A.; Sakaguchi, S.; Ishii, Y. Org. Process Res. Dev. 2000, 4,
505.
(6) Gaucher, G. M.; Lam, K. S.; Grootwassink, J. W. D.; Neway, J.; Deo, Y.
M. DeV. Ind. Microbiol. 1981, 22, 219.
(7) Gibson, D. T. Microbial Degradation of Organic Compounds; Marcel
Dekker: New York, 1984.
(3) Dobias, L.; Chuda, K.; Volf, J.; Cerna, M. Biol. Chem. ZiVocisne Vyroby
- Vet. 1986, 22, 87.
(4) A DSC evaluation of 1 showed a release of 1733 J/g beginning at 195 °C.
10.1021/op025501e CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/03/2002
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
477

Table 1. Microbial conversions of 5 to 4 in test tube cultures
4 (%)
microorganism
ATCC no. 0.2 g/L 5a 1 g/L 5b
Absidia glauca
Absidia repens
Aspergillus tamarii
Alternaria solani
Penicillium glabrum
Diplodia gossypina
22752
14849
16865
11078
11080
20575
83
83
53
53
47
31
25
40
77
17
1
0
24
23
Cunninghamella echinulata 8983
a In situ yields of 4 4 days after treatment with 5 at 0.2 g/L. b In situ yields
of 4 12 days after treatment with 5 at 1 g/L.
Figure 1. Production of 4 in A. repens bioconversions at 8-L
([), 100-L (9), and 14000-L (2) scale.
is used in the commercial production of 5-methylpyrazine-
2-carboxylic acid from 2,5-dimethylpyrazine.8 To our knowl-
edge, the biocatalytic oxidation of alkyl substituents on
bicyclic heteroaromatic compounds such as 5 has not been
reported. In this contribution, we report methods for the
biocatalytic oxidation of 5 to 4 using bacterial and fungal
whole cell systems.
The next step in the development of the fermentation
process was to evaluate the bioconversion in shake flask
cultures. Flask cultures (300-mL Erlenmeyer flask; 25 mL
of Iowa medium) of A. repens incubated with 1 g/L substrate
produced 4 in 80% yield, but required 24 days of incubation
at 29 °C and 210 rpm agitation. The extreme increase in
reaction time observed in scaling from test tube to flask
cultures was surprising since fermentation parameters,
including inoculum ratio, were kept constant as much as
possible. Although reaction time was a concern, the biocon-
version of 5 to 4 in test tube and flask cultures of A. repens
resulted in a product mixture that appeared clean by HPLC,
with little or no starting material and no organism-related
byproducts. Only one reaction intermediate, 2-hydroxy-
methylquinoxaline 6, was observed in the early stages of
the bioconversion. Aldehyde 3 was not detected in fermenta-
tion broth at any time during the bioconversions. Despite
the potential limitations due to low substrate loading and
long reaction times observed in test tube and flask experi-
ments, 8-L fermentation trials were attempted to assess
further scalability. Initially, bioconversions in 8-L fermentors
were done using the same medium used in flask-scale runs
with 1 g/L substrate added 2 days after inoculation. The
results of these 8-L runs were poor, with in situ yields of
50% after 11 days. However, the process was greatly
improved by changing to a glucose, corn steep solids
medium, and adding a continuous nutrient feed containing
glucose and yeast extract. Under these conditions, 1 g/L
substrate was converted to 4 in 74% yield after 4 days (Figure
1). A further increase in scale to 100 L using the glucose,
corn steep solids medium, gave even better results, reaching
a maximum yield of 80% (0.97 g/L in broth concentration)
after 3 days (Figure 1). While a product concentration of
0.97 g/L is unlikely to be suitable for a commercial process,
it was satisfactory to meet our immediate needs. Thus, we
proceeded to scale the process into 14000-L fermentors.
Three 14000-L bioconversion runs were carried out using a
two-stage fermentation process similar to the one described
for the 100-L bioconversion, except that a total of 1.5 g/L
substrate was added using a fed-batch protocol. This protocol
resulted in a lower overall yield from 5 (57% average in
situ yield) compared to the batch processes used at 8- and
100-L scale, but produced approximately the same product
Results and Discussion
We began our studies by evaluating the ability of P. putida
ATCC 33015 to oxidize 5 to 4. In experiments performed
in flasks and 10-L fermentor cultures, we observed the
production of 4 by P. putida cells grown in the presence of
p-xylene and incubated with 5. These preliminary experi-
ments afforded in situ yields up to 81% starting from an
initial substrate load of 1 g/L, thus demonstrating the ability
of P. putida to perform the desired transformation. However,
scale-up of a P. putida-based process requiring p-xylene
induction was considered problematic for our fermentation
pilot plant facility due to a potential explosion hazard from
p-xylene. Thus, we initiated a parallel effort to identify an
organism that could carry out the transformation without the
addition of a potentially hazardous enzyme inducer. A screen
was performed on over 200 microorganisms from our in-
house culture collection including species of fungi and
bacteria. Test tube cultures of individual organisms were
grown for 2 days in Iowa medium and then incubated with
5 at an initial concentration of 0.2 g/L for 4 days. Extracts
of culture broths were analyzed by HPLC revealing the
presence of 4 in 32% of the cultures. While most of these
organisms produced low yields of 4, seven produced in situ
yields of 25% or higher (Table 1). These organisms were
re-tested at a substrate concentration of 1 g/L to assess their
ability to produce higher product concentrations, a factor that
affects production scale and product recovery and that is
therefore a crucial requirement for commercial feasibility.
While all of the organisms produced lower yields of 4 at
the higher substrate concentration (Table 1), Absidia repens
ATCC 14849 was selected for further evaluation since it
produced similar yields of 4 at both substrate levels. Further
experiments in test tube cultures of A. repens revealed no
production of 4 at an initial substrate load of 3 g/L,
suggesting toxicity of the substrate to the organism.
(8) Kiener, A. Angew. Chem. Int. Ed. Engl. 1992, 31, 774.
478
•
Vol. 6, No. 4, 2002 / Organic Process Research & Development

Table 2. P. putida bioconversion of bicyclic heteroatomatic
compounds
substrates
yield (%)a
substrates
yield (%)a
5
106
0
9
0
1
8-methylquinoline
3-methylisoquinoline
2-methylindole
17
21
1
2-methylquinoline
3-methylquinoline
4-methylquinoline
6-methylquinoline
5-chloro-2-methylindole
0
a Average in situ yield of carboxylic acids from duplicate flasks after 45 h.
achieving a successful bioconversion. Further improvements
to the process should be possible by implementing feedback
controls for substrate and inducer additions.
Figure 2. P. putida bioconversion of 5 (2) to 4 (9) at 8-L scale.
P. putida ATCC 33015 oxidizes methyl groups in a wide
variety of methylated five- and six-membered heteroaromatic
compounds, and in the case of mono- and dimethylated
pyridine and piperazine analogues, methyl groups at all
positions were susceptible to oxidation except those adjacent
to another methyl group.8 To examine the specificity of P.
putida for methylated bicyclic heteroaromatic compounds,
we incubated P. putida cells with methylated quinoline,
isoquinoline, and indole derivatives. The results (Table 2)
show that methyl groups on 3- and 8-methylquinoline and
3-methylisoquinoline were susceptible to oxidation, but with
much lower yields compared to that from 5. Methyl groups
on 2-, 4-, and 6-methylquinoline were not oxidized, except
for a trace of product from 6-methylquinoline. The lack of
reaction for 2-methylquinoline was very surprising, consider-
ing the ability of P. putida to oxidize methyl groups on 5
and pyridine and piperazine derivatives as described above.
Although these results do not provide a clear picture
regarding the substrate specificity of P. putida for methyl
groups on nitrogen-containing, bicyclic, heteroaromatic
compounds, it appears that compounds with two nitrogens,
such as quinoxaline and piperazine, are better substrates than
quinoline derivatives.
concentration (1.0 g/L average concentration) and therefore
maintained the volumetric productivity observed at lower
scale. Production of 4 at 14000-L scale (Figure 1) was slower
compared to that from the smaller-scale bioconversions,
possibly due to a longer lag phase in the growth of the
organism.
The isolation of 4 from the A. repens bioconversion was
accomplished using the following process. The crude bio-
conversion broth was filtered to remove biomass, acidified
to pH 3, and passed through a column packed with XAD-
16 resin. Compound 4 was then eluted from the resin with
phosphate buffer (pH 8.5) and extracted into methylene
chloride. Four extractions were required to give a 90% yield
for this step. Concentration of the methylene chloride extract
and cooling to 5 °C afforded 4 as an off-white powder. This
process recovered 49% of the 4 present in the crude
bioconversion broth. Recovery of product from the three
14000-L runs yielded 20.5 kg of 4. The overall yield of the
A. repens bioconversion process at 14000 L was 28% based
on a 57% bioconversion yield and a 49% recovery yield.
The A. repens bioconversion process gave a lower overall
yield compared with the synthetic process, but was carried
out in a single pot, and eliminated safety hazards associated
with the chemical process. The major drawback of the
bioconversion process is low product concentration, which
limits production capacity and increases solvent consumption
for recovery. Even with 70% solvent recovery, almost 26000
L of methylene chloride was used for the campaign, thus
illustrating the challenge of recovering product from a dilute
aqueous product stream. High product concentration is
therefore a crucial requirement for a commercially feasible
bioconversion process. With this goal in mind we reexamined
the P. putida bioconversion. This time we directed efforts
towards using an inducer that poses less of an explosion
hazard than p-xylene and selected benzyl alcohol for evalu-
ation due to its lower volatility. Trials at 8-L scale showed
that P. putida could indeed catalyze the bioconversion of 5
using benzyl alcohol as inducer and sole carbon source, and
produced greater than 10 g/L 4 (Figure 2). The P. putida
bioconversion of 5 is sensitive to the accumulation of
substrate and benzyl alcohol and performs poorly at substrate
concentrations above 1.5 g/L and benzyl alcohol concentra-
tions above 1 g/L. Therefore, carefully controlled additions
of substrate and benzyl alcohol are very important for
Conclusions
In summary, we demonstrated two bioconversion pro-
cesses for the oxidation of 5 to 4 that eliminated safety
hazards associated with the synthetic process. A fungal
process using A. repens was demonstrated at 14000-L scale,
and while not commercially feasible due to low product
concentration, yielded kilograms of material for the synthesis
of a drug candidate. A second bioconversion process using
P. putida ATCC 33015 was demonstrated at 8-L scale. The
P. putida process yielded a product concentration greater than
10 g/L and is therefore more likely to be commercially
feasible.
Experimental Section
Chemicals. Chemicals, reagents, and solvents were
obtained from commercial sources and used without puri-
fication. All of the substrates for microbial bioconversions
were purchased from Aldrich, as were samples of the
corresponding carboxylic acids, except for 6-quinolinecar-
boxylic acid (Eastman Fine Chemicals). 1 was obtained from
Farchemia.
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
479

Analytical Methods. The amounts of substrates and
carboxylic acid products in broth samples from bioconversion
experiments were determined using HPLC methods described
below. These HPLC methods were calibrated using authentic
samples of the substrates and carboxylic acids.
Method 1: Inertsil C8 column (4.6 mm × 250 mm,
Column Engineering, Inc.) eluted isocratically with aceto-
nitrile:0.05% TFA (1:4, v/v) at a flow rate of 1 mL/min with
UV detection at 236 nm. Retention times for 4 and 5 were
approximately 7.9 and 12.6 min, respectively.
Method 2: Symmetry C18 column (3.9 mm × 150 mm,
Waters Corp.) eluted isocratically with acetonitrile: 0.05%
TFA (1:9, v/v) at a flow rate of 1 mL/min with UV detection
at 236 nm. Retention times for 2-methylquinoline, 2-quino-
linecarboxylic acid, 3-methylquinoline, 3-quinolinecarboxylic
acid, 4-methylquinoline, 4-quinolinecarboxylic acid, 6-meth-
ylquinoline, 6-quinolinecarboxylic acid, 3-methylisoquino-
line, and 3-isoquinolinecarboxylic acid were approximately
2.4, 3.9, 3.1, 4.3, 3.0, 1.5, 3.9, 2.1, 3.0, and 2.5 min,
respectively.
Method 3: Symmetry C18 column (3.9 mm × 150 mm)
eluted isocratically with acetonitrile: 0.05% TFA (1:1, v/v)
at a flow rate of 1 mL/min with UV detection at 236 nm.
Retention times for 2-methylindole, 2-indolecarboxylic acid,
5-chloro-2-methylindole, and 5-chloro-2-indolecarboxylic
acid were approximately 4.8, 2.1, 7.9, and 2.9 min, respec-
tively.
hydroxide. The aqueous sodium hydroxide addition was
completed in approximately 1 h at 17-26 °C. The reaction
mixture was heated to 80 °C for 1 h and then cooled to 25
°C. During the cooling period, an additional 4.3 g of sodium
hydrosulfite was added. The reaction mixture was then
washed with chloroform (4 × 68 L) and the aqueous layer
separated and incubated with 7.5 kg of Darco KB. The
aqueous layer was filtered, adjusted to pH 2.8 with HCl, and
cooled to 10 °C for 2 h. Filtration of this mixture gave 8.95
kg of a brown solid after drying. This material was dissolved
in 227 L of N,N-dimethylacetamide and incubated with 2.7
kg of Darco KB for 1 h. The mixture was filtered, diluted
with 210 L of water, stirred overnight, and then cooled to
10 °C for 1 h. Filtration of this mixture gave 6.25 kg (35%)
of solids (mp 196-196.5 °C) after washing with water and
drying.
Microbial Screen for Bioconversion of 5 to 4. Cultures
of various microorganisms were obtained from the American
Type Culture Collection (Manassas, VA) and stored as
glycerol suspensions of spores or vegetative cells at -70
°C. Microbial cultures were grown in glass culture tubes (16
mm × 125 mm) containing 2.5 mL of Iowa medium (20
g/L glucose, 5 g/L nutrisoy flour, 5 g/L yeast extract, 5 g/L
NaCl, 5 g/L K2HPO4; adjusted to pH 7 before sterilization).
Individual tubes were inoculated with spores or vegetative
cells of various microorganisms and incubated at 29 °C on
a rotary shaker (210 rpm) for 48 h. Each culture was then
incubated with 0.5 mg of 5 (0.05 mL of 10 mg/mL DMSO
solution) and incubated for an additional 4 days. The culture
broths were then acidified to pH 2 with 4 N HCl, extracted
with EtOAc, and analyzed by HPLC (method 1).
A. repens ATCC 14849 Bioconversion of 5 to 4.
Procedure A (8-L BioconVersion). A 14 L fermentor (New
Brunswick Scientific, New Brunswick, NJ) containing 8 L
of glucose-corn steep solids medium (20 g/L cornsteep
solids, 20 g/L glucose; adjusted to pH 4.85 before steriliza-
tion) was autoclaved at 121 °C for 0.5 h. The fermentor was
inoculated with 80 mL of a glycerol suspension of A. repens
ATCC 14849 spores harvested from surface cultures grown
on rice, and was operated at 29 °C with agitation at 600
rpm, aeration at 8 L/min, and pH below 7.5. The biocon-
version was started 44 h after inoculation by the neat addition
of 8 g of 5. A feed consisting of glucose (160 g/L) and yeast
extract (32 g/L) was started immediately after substrate
addition and maintained at a rate of 8 mL/h for the duration
of the fermentation. Aliquots of fermentation broth were
removed at various times, extracted with methanol, and
analyzed by HPLC to determine yields of 4 and 5. The
bioconversion reached a maximum concentration of 0.89 g/L
4 at 93 h after substrate addition with no residual substrate
detected. This concentration corresponds to a 74% in situ
yield.
Method 4: Discovery RPamide C16 column (4.6 mm ×
150 mm, Supelco) eluted isocratically with acetonitrile:
0.05% TFA (1:9, v/v) at a flow rate of 1 mL/min with
detection at 236 nm. Retention times for 8-methyquinoline
and 8-quinolinecarboxylic acid were 2.4 and 2.9 min,
respectively.
Preparation of 2-Quinoxalinecarboxylic Acid 4 via
Scheme 1. 2-Quinoxalinecarboxaldehyde, Dimethylacetal
2. (CAUTION: Explosion hazard) An aqueous solution of
sodium hydrosulfite (35.7 kg in 146 L of water) was added
to 24.4 kg of 1 dissolved in 110 L of ethanol. This addition
was completed within 0.5 h during which the reaction
temperature increased from 20 to 50 °C. After 1 h, the
reaction mixture was cooled to 35 °C and extracted with
chloroform (5 × 69 L). The combined chloroform extracts
were incubated with 2.7 kg of Darco KB for 0.5 h and filtered
after the addition of 0.9 kg of Celite. The filtrate was
concentrated to a volume of approximately 38 L and
transferred to the next step without further purification.
2-Quinoxalinecarboxaldehyde 3. HCl (85 L, 4 N) was
added to the crude 2-quinoxalinecarboxaldehyde, dimethyl-
acetal 2, and heated to between 50 and 60 °C for 0.5 h. The
reaction mixture was cooled to 26 °C and extracted with
chloroform (4 × 68 L). The combined chloroform extracts
were incubated with 2.5 kg of Darco KB for 0.5 h, filtered,
concentrated, and transferred to the next step without further
purification.
Procedure B (100-L BioconVersion). An 8-L seed culture
of A. repens was prepared in a 14 L fermentor as described
in procedure A and operated for 23 h without substrate
addition or nutrient feed. This seed culture was used to
inoculate 100 L of glucose-corn steep solids medium
containing 1 g/L polypropylene glycol P2000 (Dow). The
2-Quinoxalinecarboxylic Acid 4. Water was added to
the 2-quinoxalinecarboxaldehyde 3 to give a total volume
of 363 L. This mixture was incubated with 40.9 L of 35%
hydrogen peroxide followed by 73 L of 10% aqueous sodium
480
•
Vol. 6, No. 4, 2002 / Organic Process Research & Development

100-L fermentor was operated at 28 °C, 250-300 rpm
agitation, 57 L/min aeration, and pH below 7.5. The
bioconversion was initiated 21 h after inoculation by the
addition of 100 g of 5 dissolved in 0.5 L of ethanol. A
nutrient feed (100 mL/h) containing glucose (160 g/L) and
yeast extract (32 g/L) was initiated 3 h after substrate addition
and maintained throughout the bioconversion. Aliquots of
fermentation broth were removed at various times, extracted
with methanol, and analyzed by HPLC to determine yields
of 4 and 5. This bioconversion produced a maximum
concentration of 0.97 g/L 4 (80% in situ yield) at 72 h after
substrate addition. A small amount of 5 (0.02 g/L) was
detected at 72 h, but not in later samples.
Procedure C (14000-L BioconVersion). A seed culture of
A. repens was prepared in a 2000-L fermentor containing
1100 L of glucose-corn steep solids medium. This fermentor
was inoculated with 100 mL of an A. repens spore suspension
and operated at 28 °C, 550 L/min aeration, 175 rpm agitation,
and pH 5.0 for 37 h. The seed culture was then transferred
to a 20000-L airlift fermentor containing 14000 L of the
glucose-corn steep solids medium. This transfer was
conducted while the seed culture was in exponential growth
with a respiration rate of 24 mmol/L/h. Aeration was set at
11200 L/min, and the temperature was controlled at 28 °C.
The bioconversion was started 30 h after inoculation by the
addition of 14 kg of 5. An additional 7 kg of 5 was added
137 h after inoculation. A continuous nutrient feed containing
25 g/L ardamine PH and 125 g/L glucose was initiated along
with the first substrate addition at a rate of 28 L/h. The
bioconversion was stopped 168 h after the initial substrate
addition with a final concentration of 1.15 g/L 4 and a
residual concentration of 0.075 g/L 5.
by inoculating three 2.8 L Fernbach flasks, each containing
200 mL of mineral salts medium (6 g/L Na2HPO4, 3 g/L
KH2PO4, 1 g/L NH4Cl, 0.5 g/L NaCl, 0.12 g/L MgSO4, 0.01
g/L CaCl2, 1 mL/L trace elements solution, and 15 mL/L
FeEDTA solution; trace elements solution: 10 g/L di-sodium
EDTA.2H2O, 9 g/L ZnSO4‚7H2O, 4 g/L MnCl2‚4H2O, 2.7
g/L H3BO3, 1.8 g/L CoCl2‚6H2O, 1.1 g/L CuCl2‚2H2O, 0.18
g/L NiCl2‚6H2O, 0.2 g/L, Na2MoO4‚2H2O; FeEDTA solu-
tion: 5 g/L di-sodium EDTA‚2H2O, 2 g/L FeSO4‚7H2O) and
0.2 mL of benzyl alcohol with a glycerol suspension of P.
putida ATCC 33015 cells. After agitation on a rotary shaker
(200 rpm) for 24 h at 30 °C, the contents of the three first
stage cultures were combined, and 400 mL used to inoculate
8 L of mineral salts medium in a 14-L fermentor. The mineral
salts medium used in the fermentor was identical to the first
stage medium except for the addition of P2000 (1.5 mL/L)
to prevent foaming. The fermentor was operated at pH
7-7.5, 29-30 °C, 8 L/min airflow, and 600-850 rpm
agitation. During the bioconversion, (NH4)2SO4 (20 g) and
NH4Cl (20 g) were added at 25 and 75 h after inoculation,
respectively. Immediately after inoculation, a continuous feed
of benzyl alcohol was initiated at a rate of 0.58 g/h and
increased incrementally to a rate of 4.5 g/h by 76 h after
inoculation. A total of 258 g of benzyl alcohol was added
over the course of the bioconversion. A total of 82 g of 5
was added to the fermentor in increments of 2, 2, 6, 4, 2, 8,
4, 8, 6, 6, 8, 4, 4, 8, 4, and 6 g at 2, 5, 11, 18, 23, 35, 42, 46,
51, 56, 60, 66, 75, 83, 92, and 96 h after inoculation,
respectively. Aliquots of fermentation broth were removed
at various times, extracted with methanol, and analyzed by
HPLC to determine yields of 4 and 5. The bioconversion
reached a maximum concentration of 10.7 g/L 4 (86% in
situ yield) after 99 h.
Recovery of 4 from 14 000-L A. repens Bioconversion.
4 was recovered from a 14000-L A. repens bioconversion
using the following general procedure. The fermentation
broth from a 14000-L bioconversion was filtered through a
0.2 µm tangential flow ceramic filter (U.S. Filter), diluted
with 10 mM aqueous potassium phosphate (∼3500 L), and
adjusted to pH 3 with phosphoric acid. The acidified broth
filtrate was loaded onto a column (2.44 m × 0.61 m, i.d.)
containing 743 L of XAD-16 resin. The resin was washed
with three bed volumes of 10 mM potassium phosphate (pH
3) and then eluted with five bed volumes (∼3700 L) of 100
mM potassium phosphate (pH 8.5). The eluate was adjusted
to pH 2 with phosphoric acid, charged with sodium chloride
(360 kg) to produce a 10% solution, and extracted with
methylene chloride (4 × 2580 L). The methylene chloride
extract was concentrated to approximately 1000 L and cooled
to 5 °C for 4 h, and the resulting solids were collected and
dried to give 4 as an off-white solid. This process was carried
out on three 14000-L bioconversion runs resulting in the
recovery of 20.5 kg of 4 (28% yield).
P. putida ATCC 33015 Oxidation of Bicyclic Het-
eroaromatic Compounds. An 8-L fermentor culture of P.
putida ATCC 33015 was prepared and inoculated essentially
as described in the previous section and operated at pH 7,
29 °C, 8 L/min aeration, and 600 rpm agitation. A continuous
feed of benzyl alcohol was initiated after inoculation at a
rate of 0.55 g/h, and increased to 1.0, 1.8, and 2.5 g/h at 22,
30, and 47 h, respectively. Ammonium chloride (10 g) was
added to the fermentor 31 h after inoculation. At 53 h after
inoculation, 200-mL aliquots of the P. putida culture broth
were transferred to 18 Erlenmeyer flasks (1 L). Duplicate
flasks were then incubated with 0.1 g of 5, 2-methylquino-
line, 3-methylquinoline, 4-methylquinoline, 6-methylquino-
line, 8-methylquinoline, 3-methylisoquinoline, 2-methylin-
dole, and 5-chloro-2-methylindole, and agitated (210 rpm)
for 45 h at 29 °C. Samples were then centrifuged to remove
cells, diluted with methanol, and analyzed by HPLC using
methods 1-4.
P. putida ATCC 33015 Bioconversion of 5 to 4. A two-
stage fermentation process was used for the bioconversion
of 5 to 4 by P. putida. The first-stage culture was prepared
Received for review January 7, 2002.
OP025501E
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
481