Recombinant chlorobenzene dioxygenase from Pseudomonas sp. P51: A biocatalyst for regioselective oxidation of aromatic nitriles

Article abstract of DOI:10.1002/adsc.200505075

An efficient biocatalyst was developed for the cis-dihydroxylation of aromatic nitriles. The chlorobenzene dioxygenase (CDO) genes of Pseudomonas sp. strain P51 were cloned under the strict control of the Palk promoter of Pseudomonas putida GPo1. Escherichia coli JM101 cells carrying the resulting plasmid pTEZ30 were used for the biotransformation of benzonitrile in a 2-L stirred tank bioreactor. Use of a stable expression system resulted in an average specific activity and an average volumetric productivity of 1.47 U/g cdw and 120 mg of product/h/L, respectively. The values represent a three-fold increase compared to the results of the similar biotransformations with E. coli JM101 (pTCB144) where the genes of CDO were expressed under the control of lac promoter. The productivity of the cis-dihydroxylation process was limited by product toxicity. Removal of the products at toxic concentrations by means of an external charcoal column resulted in an additional increase in product concentration by 43%. E. coli JM101 (pTEZ30) was further used for the regio- and stereospecific dihydroxylations of various monosubstituted benzonitriles, benzyl cyanide, and cinnamonitrile. Biotransformations resulted in products with 42.9-97.1% enantiomeric excess. Initial enzymatic activities and isolated yields were obtained in the range of 1.7-4.7 U/g cdw and of 3-62%, respectively.

Full text of DOI:10.1002/adsc.200505075

FULL PAPERS
Recombinant Chlorobenzene Dioxygenase from Pseudomonas
sp. P51: A Biocatalyst for Regioselective Oxidation of Aromatic
Nitriles
Selcuk Yildirim,a Telma T. Franko,b Roland Wohlgemuth,c Hans-Peter E. Kohler,d
Bernard Witholt,a Andreas Schmide,*
a
Institute of Biotechnology, Swiss Federal Institute of Technology Zurich, CH-8093 Zurich, Switzerland
School of Chemical Engineering, UNICAMP, P. O. Box 6066, 13081-970, Campinas, Brazil
Fluka Chemie AG, Biochemistry Department, Industriestr. 25, 4700 Buchs, Switzerland
EAWAG, Swiss Federal Institute of Environmental Science and Technology, 8600 Dꢀbendorf, Switzerland
b
c
d
e
Department of Biochemical & Chemical Engineering, University of Dortmund, Emil-Figge-Strasse 66, 44227 Dortmund,
Germany
Fax: (þ49)-231-755-7382, e-mail: andreas.schmid@bci.uni-dortmund.de
Received: February 1, 2005; Accepted: March 29, 2005
Abstract: An efficient biocatalyst was developed for was limited by product toxicity. Removal of the prod-
the cis-dihydroxylation of aromatic nitriles. The chlor- ucts at toxic concentrations by means of an external
obenzene dioxygenase (CDO) genes of Pseudomonas charcoal column resulted in an additional increase
sp. strain P51 were cloned under the strict control of in product concentration by 43%. E. coli JM101
the Palk promoter of Pseudomonas putida GPo1. Es- (pTEZ30) was further used for the regio- and stereo-
cherichia coli JM101 cells carrying the resulting plas- specific dihydroxylations of various monosubstituted
mid pTEZ30 were used for the biotransformation of benzonitriles, benzyl cyanide, and cinnamonitrile. Bio-
benzonitrile in a 2-L stirred tank bioreactor. Use of transformations resulted in products with 42.9–97.1%
a stable expression system resulted in an average spe- enantiomeric excess. Initial enzymatic activities and
cific activity and an average volumetric productivity isolated yields were obtained in the range of 1.7–
of 1.47 U/g cdw and 120 mg of product/h/L, respec- 4.7 U/g cdw and of 3–62%, respectively.
tively. The values represent a three-fold increase com-
pared to the results of the similar biotransformations
with E. coli JM101 (pTCB144) where the genes of Keywords: aromatic nitriles; asymmetric synthesis; bi-
CDO were expressed under the control of lac promot- ocatalysis; biotransformations; dihydroxylation; in situ
er. The productivity of the cis-dihydroxylation process product removal
Introduction
The use of enzymes as catalyst in organic synthesis pro-
vides new routes to synthesize previously undescribed
compounds.[1–3] Asymmetric dihydroxylation of aro-
matic compounds by bacterial dioxygenases[4,5] (Fig-
ure 1) is such a reaction, where the product cis-dihydro-
diols are valuable chiral synthons used in organic syn-
theses.[6–8] Next to catalyst selectivity, process efficiency
depends on turnover frequency, total turnover number
and volumetric productivity.[9] The rates of cis-dihydrox-
ylation of non-natural (non-physiological) substrates
are generally relatively low, compared to the rates ach-
ieved with natural substrates such as toluene for TDO,
Figure 1. Structural organization of chlorobenzene dioxyge-
nase and designation of the protons and carbons of cis-dihy-
drodiols formed by chlorobenzene dioxygenase. Gene prod-
ucts; tcbAa: large subunit of terminal oxygenase, tcbAb: small
naphthalene for NDO and chlorobenzene for
CDO.[10,11] Recombinant biocatalysts circumvent such
limitations of (mutated) wild-type cells. Volumetric pro-
ductivities can be increased and total reaction times can ductase.
subunit of terminal oxygenase, tcbAc: ferredoxin, tcbAd: re-
1060
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/adsc.200505075
Adv. Synth. Catal. 2005, 347, 1060–1072

Recombinant Chlorobenzene Dioxygenase from Pseudomonas sp. P51
FULL PAPERS
be prolonged on preparative scales by expressing the promoter, and the production of enzymes on cell growth
genes from an efficient and controllable promoter.[12,13]
were examined. The growth of E. coli JM101 (ꢀ), E. coli
We developed a new biocatalyst by cloning the genes JM101 (pRS), and E. coli JM101 (pTEZ30) (Table 1)
of the chlorobenzene dioxygenase (CDO) of Pseudo- was followed for 5 hours in the absence and in the pres-
monas sp. strain P51[14] behind the alk promoter alkBp ence of the inducer [0.05% dicyclopropyl ketone
of Pseudomonas putida Gpo1[15] resulting in pTEZ30. (DCPK)]. The specific growth rates of all three strains
We previously showed that E. coli JM101 (pTEZ30) were identical (0.56 hꢀ1) in the absence of the inducer
can be used for efficient cis-dihydroxylation of cinnamo- (Figure 2). The presence of the inducer caused a slight
nitrile up to technical scales.[16] Here, we describe the decrease in specific growth rates of E. coli JM101 (ꢀ),
construction of the expression plasmid pTEZ30 and and E. coli JM101 (pRS) cells but decreased the specific
evaluate the efficiency of the biocatalyst E. coli JM101 growth rate of E. coli JM101 (pTEZ30) cells significant-
(pTEZ30) based on specific activity of the cells and ly to 0.36 hꢀ1. This shows that the presence of 0.05%
the volumetric productivity of the biotransformation DCPK, and expression of alkBp genes has just a minor
of benzonitrile in a reactor on a 2-L scale. The results effect on cell growth, whereas the enzyme production
are compared to the results of similar biotransforma- decreases the specific growth rate by 36%.
tions with E. coli JM101 (pTCB144)[14] as biocatalyst
The specific activity of E. coli JM101 (pTEZ30) was
where the CDO genes were cloned under the control tested in shake-flasks with naphthalene as a substrate,
of lac regulatory system.[17] Subsequently the new bioca- since CDO showed the highest activity towards this
talyst was used for the stereo- and regiospecific dihy- compound in a previous work[18] and the corresponding
droxylations of various ortho-, meta- and para-substitut- product is commercially available. The cells showed an
ed benzonitriles, benzyl cyanide, and cinnamonitrile. Di- initial specific activity of 2.86 units per gram cell dry
hydroxylation of aromatic nitriles is especially interest-
ing, since it offers the possibility for one-pot multistep
catalysis with a further engineered catalyst carrying,
e.g., nitrilase or nitrile hydratase. All reaction products
were characterized as thoroughly as possible and the re-
action parameters such as specific activity of the cells
and the yields of the reactions were determined.
Results
Effect of Induction on the Growth of E. coli (JM101)
(pTEZ30) and Identification of Dioxygenase Activity
The growth rate of whole-cell biocatalysts has a signifi-
Figure 2. Specific growth rates of induced and non-induced
cant effect on biotransformation efficiency[16] and gener-
E. coli JM101 (ꢀ), E. coli JM101 (pRS), and E. coli JM101
ally decreases with the enzyme production. Therefore
(pTEZ30) cells in 20 mL M9 medium in shaking flasks. * In-
the effects of the presence of inducer, induction of alk duced.
Table 1. Bacterial strains and plasmids used in this study.
Strain or plasmid
Characteristics
Source or reference
Strains of E. coli
DH5a
JM101
[62]
[58]
supE44, DlacU169 (F80lacZDM15) hsdR17 recA1, endA1, gyrA96, thi-1, relA1
supE thi-1 D(lac-proAB) F’ [traD26 pro ABþ lacIq lacZDM15]
Plasmids
pTCB144
pSPZ2MA
pTEZ10
pTEZ20
PBRNSKmD
pTEZ30
[14]
[12]
lac, tcbAa, tcbAb, tcbAc, tcbAd genes Apr
Source of alkBp and AlkS, Apr
pSPZ2MA with tcbAa
This work
pTEZ10 with tcbAa,[a] tcbAb, tcbAc, and tcbAd
Source of Kmr
This work
[12]
AlkBp, AlkS, Kmr, tcbAa, tcbAb, tcbAc, and tcbAd
AlkBp, alkS, Kmr
This work
[30]
PRS
[a]
The plasmid contains only a part of the tcbAa genes.
Adv. Synth. Catal. 2005, 347, 1060–1072
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1061

Selcuk Yildirim et al.
FULL PAPERS
only 2.37 mM in 8.7 hours corresponding to an average
volumetric productivity of 40 mg of product/h/L.
The initial volumetric productivity of E. coli JM101
(pTEZ30) cells was similar to that of E. coli JM101
(pTCB144) cells. In contrast, E. coli JM101 (pTEZ30)
which confers kanamycin resistance was quite stable
during the biotransformation even at the high cell densi-
ties. This resulted in an increase of the volumetric pro-
ductivity to a maximum of 0.26 g of product/h/L after 5
hours. In 8.5 hours 7.5 mM product were formed, yield-
ing an average volumetric productivity of 120 mg prod-
uct/h/L, which is 3-fold the average volumetric produc-
tivity achieved with E. coli JM101 (pTCB144) cells.
The volumetric productivity decreased after five hours
with a significant decrease in specific growth rate, which
is probably due to the toxic effect of the product. We di-
luted the cells taken at this time point and plated them
on LB agar plates with and without kanamycin. Equal
numbers of colonies were obtained in both cases, which
shows high stability of the plasmid. However, the com-
parison of viable cell counts to the spectrometrically de-
Figure 3. Product formation and specific activities of the bio-
transformation of naphthalene to cis-1,2-dihydro-1,2-naph-
thalenediol in 20-mL shaking flasks with E. coli JM101
(pTCB144) (open symbols) and with E. coli JM101
(pTEZ30) (closed symbols). The assay was performed as de-
scribed in Materials and Methods.
weight (U/g cdw) during the biotransformation of naph- termined total number of cells in the reactor indicated
thalene to cis-1,2-dihydro-1,2-naphthalenediol. One that only 60% of the total number of cells measured in
unit is defined as 1 mmol of product formed in one mi- the reactor (data not shown) were living cells. The corre-
nute. During the biotransformation, the specific activity lation between specific growth rate and volumetric pro-
decreased to 1.96 U/g cdw and 0.075 mM of product was ductivity indicates that, when the dioxygenase genes are
formed. Specific activities of the cells harboring efficiently expressed, productivity of the biotransforma-
pTCB144 were slightly higher than the activities of the tions becomes limited by product toxicity. This indicates
cells harboring pTEZ30 (Figure 3).
that use of the expression system E. coli JM101
(pTEZ30) in biotransformations of benzonitrile may re-
sult even in higher volumetric productivities when the
toxic product is removed from the reaction mixture.
At the end of the biotransformation, product concentra-
Biotransformation of Benzonitrile on 2-L Scale
To evaluate the stability of the developed biocatalyst in tion reached 10.74 mM, which is 4.5-times higher than
the bioreactor and the efficiency of the biotransforma- the maximum concentration achieved in biotransforma-
tion with it, we performed the biotransformation of ben- tions with E. coli JM101 (pTCB144). A summary of the
zonitrile to cis-1,2-dihydroxy-3-cyanocyclohexa-3,5-di- comparison of the process parameters is given in Ta-
ene (BNDD) in a bioreactor on a 2-L scale with E. coli ble 2.
JM101 (pTEZ30) (Figure 4). The resultswerecompared
with the results of the biotransformation with E. coli
JM101 (pTCB144) cells. Both E. coli JM101 (pTEZ30) Biotransformation of Aromatic Nitriles
and E. coli JM101 (pTCB144) cells showed similar
growth properties in the reactor during the biotransfor- We used the expression system E. coli JM101 (pTEZ30)
mations (Figure 4A). Although the cell growth behav- to investigate the dihydroxylation of various aromatic
iors were very similar (Figure 4B), the product forma- nitriles. For that purpose we performed the biotransfor-
tion rates differed significantly. With E. coli JM101 mations of the substrates listed in Table 3 in a reactor on
(pTCB144), the initial activity and the volumetric pro- a 300-mL scale. The dihydrodiol metabolites were iso-
ductivity were 0.67 U/g cdw and 0.07 g of product/h/L, lated and identified by GC-MS of their respective n-bu-
respectively, and both decreased significantly during tylboronate derivatives(BB derivatives) and by H NMR
the biotransformation as presented in Figure 4C. Anal- spectroscopy (Table 3). BB derivatives of cis-dihydro-
ysis of colony forming unit by plating E. coli JM101 diol isomers were separated on a GC column with an
(pTCB144) cells taken at the end of the biotransforma- achiral stationary phase and could be distinguished
tion on LB agar plates with and without ampicillin from those of trans-dihydrodiols because of the signifi-
showed that only 13% of the cells contained the plasmid cant mass difference between their molecular ions.[19]
and confirmed the instability of antibiotic resistance Enantiomers of cis-dihydrodiols were separated as their
(data not shown). During the biotransformation the BB derivatives on a GC column with a chiral stationary
product concentration was increased to a maximum of phase and quantified. NMR spectroscopy was used to
1062
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1060–1072

Recombinant Chlorobenzene Dioxygenase from Pseudomonas sp. P51
FULL PAPERS
Figure 4. Biotransformation of benzonitrile to cis-1,2-dihydroxy-3-cyanocyclohexa-3,5-diene in a reactor on a 2-L scale. Open
and closed symbols represent the values obtained with E. coli JM101 (pTCB144) and E. coli JM101 (pTEZ30), respectively. A.
Cell growth and product formation. B. Growth rate profile during biotransformation. C. Activity and productivity profiles.
Table 2. Process parameters of biotransformation of benzonitrile using E. coli JM101 (pTCB144) and E. coli JM101 (pTEZ30)
in a reactor on a 2-L scale.
E. coli JM101 (pTCB144)
E. coli JM101 (pTEZ30)[a]
Total amount of product produced (g)
Bioconversion yield (%)
0.65
72
2.12
73.5
0.20
2.87
1.47
43.0
14.3
0.26
0.12
Maximum growth rate (hꢀ1
)
0.24
0.67
0.43
10.5
4.56
0.07
0.04
Maximum specific activity (U/g cdw)
Average specific activity (U/g cdw)
Maximum volumetric activity (U/L)
Average volumetric activity (U/L)
Maximum volumetric productivity (g prod/h/L)
Average volumetric productivity (g prod/h/L)
[a]
Values were calculated for the first 9 hours of the biotransformation for a better comparison to the biotransformation using
E. coli JM101 (pTCB144).
determine the structure and the position of the hydroxy NMR measurements of the products showed that the
groups. Figure 1 shows our nomenclature to specify the cis-dihydroxylation occurred on the unsubstituted ring
protons of the dihydrodiol compounds. GC-MS and at the 1,2 positions with all substrates tested.
Adv. Synth. Catal. 2005, 347, 1060–1072
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1063

Selcuk Yildirim et al.
FULL PAPERS
1064
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1060–1072

Recombinant Chlorobenzene Dioxygenase from Pseudomonas sp. P51
FULL PAPERS
Table 4. Biotransformation of aromatic nitriles with E. coli JM101 (pTEZ30) containing CDO.
Substrate
Conversion[a] [%]
Activity[b]
[U/g cdw]
Yield[c]
[%]
Yield[d]
[%]
15 min
30 min
1 h
2 h
3 h
4 h
5 h
6 h
7 h
Benzonitrile
4
3
4
6
1
2
3
3
3
2
7
8
9
11
3
5
5
6
6
6
14
10
11
19
5
7
8
9
12
10
26
15
16
25
12
18
15
21
27
20
33
18
19
28
17
29
19
31
36
24
40
21
21
31
21
37
21
41
43
28
47
24
23
35
26
46
24
51
49
33
51
26
25
36
29
53
25
59
51
35
55
27
26
37
33
58
27
67
52
38
3.9
4.0
4.4
4.7
1.7
2.5
2.9
3.3
3.1
3.0
82
7
62
3
12
4
21
17
11
52
33
10
2-Chlorobenzonitrile
2-Bromobenzonitrile
2-Fluorobenzonitrile
3-Fluorobenzonitrile
4-Chlorobenzonitrile
4-Fluorobenzonitrile
4-Methylbenzonitrile
Cinnamonitrile
27
15
36
47
79
85
56
61
Benzyl cyanide
[a]
Calculated according to the product formation in 20-mL shake flask experiments.
Activity of the cells in the first 30 minutes of the biotransformations in 20-mL shake flask experiments.
Yields of the biotransformations on 300-mL scale
[b]
[c]
[d]
Yield of the biotransformations (on 300-mL scale) after purification of the products.
Specific activity of the cells and the rate of conversion this purpose we tested the well established ISPR con-
are important factors for a scale-up of such biotransfor- cepts[20–25] to remove the product from the reaction me-
mations. Therefore we performed the biotransforma- dium to non-toxic concentrations. First, in situ product
tions in 20-mL shake flasks and determined the reaction extraction using aqueous two-phase systems (ATPS)
rates and the specific activity of the biocatalyst E. coli was tested. Amongst the aqueous two-phase systems
JM101 (pTEZ30). The results are presented in Table 4. (ATPS) investigated [polyethylene glycol (PEG)/phos-
Biotransformations with ortho- substituted benzoni- phate, PEG/ammonium sulfate, PEG/magnesium sul-
triles showed higher activities than with meta- or para- fate and PEG/dextran], PEG/MgSO4 was the least toxic
substituted benzonitriles. The highest activity was ob- system to the cells, and allowed cell growth and substrate
tained with ortho-fluorobenzonitrile. Biotransforma- conversion (data not shown). Concentrations as high as
tion of p-methylbenzonitrile resulted in the highest bio- 10 to 15% of monophasic PEG8000 allowed cell growth
conversion (67%) after 7 hours.
identical to the control experiments on plain culture me-
The yields of the biotransformations performed on a dia. 2.5% to 4.5% MgSO4 reduced the cell growth up to
300-mL scale were in the range of 4–62% (Table 4). 30%. The partition coefficients of the substrate and
The non-optimized biotransformation of 4-methylben- product in PEG4000/MgSO4 and PEG8000/MgSO4 sys-
zonitrile resulted in a significantly higher yield 52%, tems, were elucidated (Table 5). However, the partition
compared to the yield of biotransformation of other sub- coefficients of benzonitrile and BNDD were not high
stituted aromatic nitriles.
enough to keep cells well separated from the substrate
and product, so the option of using ATPS was discarded,
despite the compatibility with the cells. As an alternative
to extractive reactions, adsorptive reactions were stud-
ied. This has been used, amongst others, for ISPR of flu-
Partitioning in Aqueous Two-Phase Systems
The results of the biotransformation of benzonitrile orocatechol, a compound resembling BNDD, during
showed the effect of product toxicity on cell growth conversion of fluorobenzene by Pseudomonas putida,[23]
and volumetric productivity. To examine this effect fur- and in pilot-plant processes for microbial biotransfor-
ther, we aimed at removing the product from the reac- mations.[25–28] We determined preliminary adsorption
tion medium to non-toxic concentrations during the bi- isotherms of benzonitrile and BNDD for various ad-
otransformation. This would allow us to examine if the sorbents and plotted them double-logarithmically to fa-
volumetric productivity of the process could be recov- cilitate comparison (Figure 5). Sepabeads did not ad-
ered by reducing the toxic effects of the product. For sorb either benzonitrile or BNDD significantly. Further-
Table 5. Partition coefficients of benzonitrile and benzonitrile dihydrodiol (BNDD).
16% PEG4000-5.6% MgSO4
15% PEG8000-4% MgSO4
15% PEG8000-6% MgSO4
KBN
KBNDD
2.7
1.6
1.9
1.4
3.9
1.5
Adv. Synth. Catal. 2005, 347, 1060–1072
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1065

Selcuk Yildirim et al.
FULL PAPERS
any adsorbents and BN. Therefore, charcoal was used
for the in situ product adsorption for biotransformation
experiments.
Biotransformation of Benzonitrile with Integrated
Product Recovery
To examine the toxic effect of the product on cell growth
and volumetric productivity, we aimed to remove the
product from the reaction medium to non-toxic concen-
trations during the biotransformation. For that purpose,
we performed a biotransformation of benzonitrile in a 2-
L bioreactor equipped with an adsorption column con-
taining 35 g charcoal. To exclude high substrate concen-
trations that could cause toxic effects, benzonitrile was
added repeatedly. The product concentration increased
in the first 7.5 hours of the biotransformation with an
average specific activity of 1.3 U/g cdw. At the end of
Figure 5. Adsorption isotherms of benzonitrile (BN) and cis-
2-dihydroxy-3-cyanocyclohexa-3,5-diene (BNDD) using vari-
ous adsorbents.
more, the adsorption of BNDD by XAD4 was not signif- this period, the specific growth rate decreased to zero
icant. When Charcoal, XAD7 and Lewatit 1065 were due to the toxic effect of the product, which decreased
used as adsorbents, the concentrations of BNDD re- the specific activity to 0.21 U/g cdw (Figure 7). At this
maining in the liquid phase were quite low. With the time point the cell-containing broth was circulated be-
same adsorbents, the concentrations of benzonitrile in tween the reactor and the loop. After the adsorption cy-
the liquid phase were low as well but a few orders of cle, the product concentration decreased from 7.4 to
magnitudes higher (Figure 5). Therefore these adsorb- 1.9 mM and all the substrate in the reaction medium
ents were found suitable for extraction of the BNDD.
was adsorbed. After the addition of new substrate, prod-
After determining the adsorption isotherms of the uct formation started again and the specific activity
products and substrates, we tested the effects of the ad- reached 46% of that observed in the first period. The
sorbents on biomass accumulation in the absence and in product concentration reached again the toxicity thresh-
the presence of benzonitrile. The results are presented in old and a decrease in specific growth rate and specific ac-
Figure 6. Cell amount in the presence of Lewatit 1065 tivity was observed. Therefore, another cycle of adsorp-
reached only 36% of the cell amount without any ad- tion was performed and product concentration de-
sorbent. XAD7, charcoal, and Sepabeads did not have creased to 2.4 mM. After this cycle, only 30% of the ac-
a significant effect on cell growth. Addition of benzoni- tivity was recovered. At the end of the reaction, the total
trile in amounts that would be normallyquite toxic tothe product formation increased up to 15.4 mM giving a bi-
cells did not have a significant effect on growth in the oconversion yield of 61.6% and a product concentration
presence of these adsorbents. In the presence of char- of 2.1 g/L. The cells were still active and the product for-
coal, one of the best adsorbents for benzonitrile (BN) mation was continuing. Toxic concentrations of prod-
and BNDD, biomass (even in the presence of 15 mM ucts (7–8 mM) before the adsorption cycle resulted in
BN) reached the same density as in experiments without a long lag phase of the cells afterwards, which decreased
the specific activity after the adsorption.
Discussion
Recombinant Whole Cells as Biocatalyst
Use of recombinant cells to express dioxygenase genes
for whole-cell cis-dihydroxylations has several advan-
tages. Recombinant cells are well defined and therefore
easy to handle.[29] Expression of genes can be controlled
by inducible promoters.[30,31] This makes it possible to
regulate the expression to optimize enzymatic activities
and to allow the cells to grow to high densities before the
induction phase. Generally, inducers are not substrates
Figure 6. Effects of adsorbents and benzonitrile (BN) on bio-
mass accumulation. Numbers above the bars show the con-
centration (mM) of BN.
1066
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1060–1072

Recombinant Chlorobenzene Dioxygenase from Pseudomonas sp. P51
FULL PAPERS
Figure 7. Biotransformation of benzonitrile to cis-1,2-dihydroxy-3-cyanocyclohexa-3,5-diene using E. coli JM101 (pTEZ30)
cells in a reactor with integrated in situ product removal. Substrate was added repeatedly. After 7.5 and 21.5 hours, bioreactor
*
~
~
medium was passed over an external loop packed with activated charcoal. , cell amount; , product concentration; , total
amount of product formed; &ablock, substrate concentration.
for the dioxygenases and can easily be separated from or tac promoter. Such promoters are known to be not
the dihydrodiols at the end of the reaction. Background tightly regulated and the leakiness of these regulatory
activities can easily be evaluated with control experi- systems have been shown before.[43] Tightly controllable
ments whereas the effects of the other enzymes on ster- promoters enable the separation of the growth and en-
eochemistry of the product in mutant strains are difficult zyme production phases and therefore both phases can
to interpret.[32,33] Therefore, it is generally accepted that be optimized to yield a maximum volumetric productiv-
recombinant cells will be more efficient in whole-cell ity. This will be very useful for cis-dihydroxylations of
cis-dihydroxylation processes especially for the non- non-natural substrates like aromatic nitriles. The non-
natural substrates such as aromatic nitriles. Here we induced cells harboring pTCB144, where the CDO
showed that the use of efficient and stable expression genes were cloned downstream of lac promoter resulted
systems for dioxygenases can result in high volumetric in 26% of the activity of the cells that were induced for 4
productivities for cis-dihydroxylation processes. The de- hours (data not shown). On the other hand, the alk pro-
velopment of efficient and stable recombinant cell ex- moter of the alk regulatory system of Pseudomonas pu-
pression systems will also have a high impact on the syn- tida GPo1, alkBp, allows a tightly regulated gene expres-
thetic applicability of engineered dioxygenases, which sion in E. coli.[31] Recombinant cells containing expres-
are steadily increasing in number.[34–37]
sion systems based on the alk regulatory system have
been successfully used as a biocatalyst for whole cell bi-
otransformations on a preparative scale by several
groups.[31,44–48] In our experiments, CDO gene expres-
sion was tightly controlled by alkBp in pTEZ30 and
the non-induced E. coli JM101 (pTEZ30) cells did not
The alk Regulatory System as an Expression System
for CDO.
Recombinant cells have been used for dihydroxylation show any activity.
processes based on whole cells to produce several differ-
ent cis-cyclohexadienediols.[29,38,39] However, in most ex-
pression systems used so far, such as pDTG601,[40] Selection Markers for a Stable Expression
pKST11,[41]
pDTG141,[42]
p1/1,[39]
pLAC365,
pTAC365,[43] and pTCB144,[14] the dioxygenase genes A drawback of the plasmid encoded production systems
were cloned downstream of promoters such as the lac conferring resistance to b-lactam antibiotics like ampi-
Adv. Synth. Catal. 2005, 347, 1060–1072
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1067

Selcuk Yildirim et al.
FULL PAPERS
cillin, penicillin G, and amoxicillin, is the need for addi- alyzes the transformation of benzonitrile to cis-dihy-
tion of antibiotics to maintain the stable production.[49] droxy-cyclohexadienecarbonitrile and benzyl cyanide
In high cell density cultures, the b-lactamase produced to 1-(cis-1,2-dihydroxy-3,5-cyclohexadienyl)-acetoni-
by cells cleaves one bond of the four-membered ring trile (main product 98%), 2-(cis-2,3-dihydroxy-4,6-cy-
of b-lactam antibiotics. This results in the disappearance clohexadienyl)-acetonitrile, and 1-(cis-3,4-dihydroxy-
of the antibiotic resistance and leads to the low plasmid 1,5-cyclohexadienyl)-acetonitrile.[18] In our previous
stability. During the biotransformation of benzonitrile, work we described the cis-dihydroxylation of cinnamo-
the low stability of the plasmid pTCB144, conferring nitrile to trans-3-[(5S,6R)-5,6-dihydroxycyclohexa-1,3-
ampicillin resistance, resulted in decreasing specific ac- dienyl]-acrylonitrile.[16] Here we showed that E. coli
tivities and volumetric productivities. This shortened JM101 (pTEZ30) carrying CDO was able to dihydroxy-
the biotransformation time and resulted in low product late various substituted benzonitriles as well as cinna-
concentrations. On the other hand, the expression plas- monitrile and benzyl cyanide to the corresponding cis-
mid pTEZ30 conferring kanamycin resistance was sta- dihydrodiols. Regiospecific dihydroxylation occurred
ble and enabled the specific activity to increase during at the 1,2 positions. Enantiomeric excesses of the prod-
the biotransformation. Other antibiotics functioning as ucts are in the range of 42.9–97.1%. Differently, the tol-
a protein synthesis inhibitor like tetracycline or neomy- uene dioxygenase-catalyzed dihydroxylation of benzo-
cin can also be used as a stable selection marker.
nitrile and (2-cyanoethyl)benzene results in cis-diols
During the biotransformation of benzonitrile, E. coli with very high enantiomeric excess.[53,54] It was also
JM101 (pTEZ30) efficiently expressed the dioxygenase shown before that the enantiomeric excess of the prod-
genes resulting in significant increases in specific activi- ucts from CDO-catalyzed reactions can be significantly
ties and volumetric productivities. We showed, that once different from those of TDO-catalyzed reactions.[18] Our
the efficient expression system was achieved, the volu- results can be a basis for future work to learn about the
metric productivity of the cis-dihydroxylations became differences in enantiomeric excess. Our ee values were
limited by product toxicity. The toxic effects of the aro- determined by analyzing the BB derivatives of the prod-
matic substrates and the product cis-diols have been well ucts with GC-MS. When a single peak was obtained, we
examined and have been solved by the use of different could not determine whether the peak was the result of a
feeding strategies,[50] regulated substrate addition,[51] pure enantiomer with a very high ee or enantiomers that
two-liquid phase reaction media,[52] and in situ product could not be separated under the analytical conditions.
recovery.[22,23] 46% of the specific activity could be re-
In our previous work the absolute stereochemistry of
covered by removing the product from the reaction me- cinnamonitrile has been determined to be trans-3-
dium by adsorption. The decrease in the specific growth [(5S,6R)-5,6-dihydroxycyclohexa-1,3-dienyl]-acryloni-
rate and specific activity after the adsorption might be trile.[16] The absolute stereochemistry of the other products
due to adsorption of some nutrients that are essential could not be determined because the single enantiomers
to the cells for the activity and/or because of the stressed of the products were not available to us as standards.
or partially permeabilized as a result of high product
The specific activity of E. coli JM101 (pTEZ30) cells
concentrations. Therefore adsorption cycles can be per- towards the aromatic nitriles tested (1.7–4.7 U/g cdw)
formed earlier (before the amount of product reaches were in the same range as the specific activity of the E.
the toxic concentration) which may lead to a further in- coli DH5a (pTCB144) carrying the same enzyme to-
crease and high recovery of the specific activity. The re- wards other substrates like benzene (4.3 U/g protein),
sults showed clearly that the volumetric productivity of toluene (9.1 U/g protein), and chlorobenzene (8.7 U/g
the biotransformations using the expression system E. protein).[18] The isolated yields of the biotransformation
coli JM101 (pTEZ30) could be further increased and of substrates (other than benzonitrile) in reactors were
the reaction time can be prolonged if the product can between 3–52%. One way to increase the yields of these
be removed from the reaction medium to non-toxic con- reactions is to optimize the reaction conditions. In our
centrations during the reaction.
set-up, substrate and product concentrations were not
measured continuously during the reaction and sub-
strate was added repeatedly every hour. To counter
the substrate and product toxicity, controlled addition
of the substrate[16,51] and in situ product removal can be
Substrate Range
To date there are very few reactions described where ar- integrated.[20;23] Another way to increase the yields is
omatic nitriles were converted to the corresponding cis- to improve the isolation and purifications steps. Contin-
dihydrodiols. Toluene dioxygenase (TDO) catalyzes the uous phase extraction with ethyl acetate can be used as
biotransformation of benzonitrile to cis-(1S,2R)-1,2-di- an alternative to the evaporation of the aqueous
hydroxy-3-cyanocyclohexane-3,5-diene,[53]
(2-cyano- phase[29] to improve the yields. The yields after the reac-
ethyl)benzene to (5S,6R)-1-(2-cyanoethyl)-5,6-dihydroxy- tions are significantly higher than those after purifica-
cyclohexa-1,3-diene[54] and trans-cinnamonitrile to tion (Table 4). This shows that purification yields can
(trans-cinnamonitrile)-cis-2,3-dihydrodiol.[55] CDO cat- also be improved significantly.
1068
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1060–1072

Recombinant Chlorobenzene Dioxygenase from Pseudomonas sp. P51
FULL PAPERS
agarose (Difco). M9 medium contained glucose as a carbon
source 0.5% (wt/vol), thiamine0.001% (wt/vol), andantibiotic:
ampicillin (100 mg/L) for cells containing pTCB144; kanamy-
cin (50 mg/L) for cells containing pTEZ30. Cultivations for
cloning procedures were carried out at378C. Other precultures
and cultures of the E. coli strains were grown at 308C on a hor-
izontal shakers at 200 rpm. The cells containing pTCB144 and
pTEZ30 were induced with 0.5 mM isopropyl-b-d-thiogalacto-
sidase (IPTG) or 0.05% (vol/vol) dicyclopropyl ketone
(DCPK), respectively.
Conclusion
Several improvements in process design for cis-dihy-
droxylations have been achieved based on the use of dif-
ferent feeding strategies,[50] regulated substrate addi-
tion,[51,56,57] two-liquid phase reaction media,[52] and in
situ product recovery.[22,23] However, these advances
need to be complemented with biocatalyst design where
the tailored whole cells are used as biocatalyst. Here, we
showed that the volumetric productivity of the cis-dihy-
droxylation processes could be increased significantly
by use of an efficient and stable recombinant expression
system as biocatalyst. Once the expression is stable, the
volumetric productivity becomes limited by product
toxicity. Well-established ISPR approaches such as ad-
sorption on a solid matrix[20,22,23] can be applied to re-
move the toxic effect of the product. This may enable
one to further increase the efficiency of the overall proc-
ess based on the recombinant biocatalyst.
Construction of the Expression Plasmid pTEZ30
The cloning and DNA modification protocols used have been
described elsewhere.[58] 0.97 kb of the tcbAa genes of pTCB144
were amplified by polymerase chain reaction (PCR) using the
two primers (P1: 5’ ACGTGAGGAGCATATGAATCA-
We applied the biocatalyst E. coli JM101 (pTEZ30)
carrying CDO for the regio- and stereoselective dihy- CACCGAC 3’) and (P2: 5’ CGTTCGGCGCGCCTGGATCC-
CATGTCCGAACC 3’). Three cloning sites were introduced
with these primers: an NdeI site in the forward primer (in ital-
ics), and a BamHI, and AscI site in the reverse primer (in italics
and bold, respectively). The PCR product and the plasmid
pSPZ2MA (carrying alkS and alkBp) were digested with
NdeI and AscI followed by insertion of the PCR product into
pSPZ2MA (Figure 8). The resulting plasmid was called
pTEZ10. The plasmid pTCB144 was digested with ApaI and
BamHI and the 3.17 kb fragment was inserted into the plasmid
pTEZ10 after digestion with ApaI and BamHI. The resulting
plasmid carrying alkS, alkBp, tcbAa, tcbAb, tcbAc, and tcbAd
and Apr genes was called pTEZ20. Then pTEZ20 was partially
digested with NotI and the resulting 8.27 kb fragment was iso-
lated. The unique NotI site of pBRNSKMD (carrying a Kmr
gene) was cut and ligated with the 8.27 kb fragment of the par-
tial digestion. The final plasmid carrying alkS, alkBp, tcbAa,
tcbAb, tcbAc, tcbAd, and Kmr was called pTEZ30.
droxylation of benzonitrile, ortho-, meta- and para-sub-
stituted benzonitriles, as well as cinnamonitrile and ben-
zyl cyanide. cis-Dihydroxylations were at the 1,2 posi-
tions and the products have 42.9 to 97.1% enantiomeric
excess. Further work aimed at screening for a nitrilase
for the selective hydrolysis of the nitrile group is under
way. This may enable the production of new chiral syn-
thons in significant yields and amounts.
Experimental Section
Strains, Plasmids, Materials, and Chemicals
The strains and plasmids used and constructed are listed in Ta-
ble 1. Restriction and DNA modification enzymes were ob-
tained from Roche Molecular Biochemicals (Rotkreuz, Swit-
zerland) and New England Biolabs (Schwalbach, Germany).
The High Pure Plasmid Isolation Kit of Roche Diagnostics
GmbH (Indianapolis, IN, USA) was used for small-scale plas-
mid DNA isolations following the supplierꢂs protocol. Amber-
lite resins XAD4 and XAD7 were obtained from Rohm &
Haas (Frankfurt, Germany), activated charcoal (TYP: DGF
0.5–2.5) was a gift from CarboTech Aktivkohlen GmbH (Es-
sen, Germany), Lewatit 1065 and Sepabeads were obtained
from Bayer AG (Leverkusen, Germany), and Mitsubishi
Chemicals (New York, USA), respectively. 3-Fluorobenzoni-
trile, and dicyclopropyl ketone were obtained from Aldrich
and other chemicals were obtained from Fluka AG (Buchs,
Switzerland).
Media and Cultivation Conditions
We used Luria-Bertani (LB) broth (Difco, Detroit, MI.) com-
plex or M9 minimal medium[58] supplemented per liter with
1 mL of trace elements solution US*, the composition of which
is described elsewhere.[59] Solid media contained 1.5% (wt/vol)
Figure 8. Structure of the genetic elements constructed in this
study. tcbAaþ: The first 0.97 kb of the tcbAa gene.
Adv. Synth. Catal. 2005, 347, 1060–1072
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1069

Selcuk Yildirim et al.
FULL PAPERS
ry evaporator at room temperature to a volume of approxi-
mately 50 mL. Then it was diluted to 1 liter with ice-cold aceto-
nitrile for the precipitation of salts and proteins. The superna-
tant was then evaporated to approximately 3 mL and the prod-
uct was purified on a 1.9 cm i.d.ꢂ30 cm silica gel (60) column
with 80% ethyl acetate and 20% hexane as the eluent.
Cell Growth and Biotransformations in Shake-Flasks
A single colony was transferred into 5 mL LB medium and in-
cubated for 10 hours. 100 mL M9 medium were inoculated with
1 mL of the LB preculture and incubated overnight. Two times
40 mL M9 medium were inoculated to an A450 of 0.1 using M9
preculture and incubated. Cells were grown to an A450 of ꢁ0.3
and one of the two cultures was induced. Cell growth of both
cultures was followed spectrophotometrically. For biotransfor-
mation experiments in shake- flasks, cultures were centrifuged
at 48C after 4 hours of induction, washed with an equal volume
of M9 medium without carbon source, and resuspended in
20 mL M9 medium to a cell concentration of 0.29 g cdw/L
and 1.3 g cdw/L for the biotransformation of naphthalene,
and aromatic nitriles, respectively. Naphthalene and solid aro-
matic nitriles were added from stock solutions dissolved in
methanol to a final concentration of 1 mM and 2 mM, respec-
tively. Liquid aromatic nitriles were added directly to a concen-
tration of 2 mM. Samples were taken and centrifuged at 48C
and at 20,000ꢂg for 5 minutes. Supernatants were frozen at
ꢀ208C until the measurements. All assays were performed
twice.
Analytical Procedures
HPLC Analysis: The formation of metabolites was measured
by a Hewlett Packard HP 1050 Ti reversed phase HPLC with
a Nucleosil C18 RP column (pore size, 100 ꢃ; particle size
5 mm; inner diameter, 12.5 cmꢂ2 mm) (Macherey-Nagel AG,
Oensingen, Switzerland). The UV detector (HP 1040M Series
II Diode-Array) was set at a wavelength range of 210–600 nm.
Product formation from the naphthalene biotransformation
was quantified with the help of an authentic standard. Products
from biotransformations of aromatic nitriles were purified,
identified, and used as standards for quantifications.
GC-MS Analysis: After purification of the products, their n-
butylboronate derivative (BB derivative) was prepared as de-
scribed elsewhere[18] and 1 mL of the diluted samples was inject-
ed at 508C into a series 8000 GC, combined with an MD800
mass spectrometer (Fisons Instruments, Beverly, MA, USA),
on a SPB-1 column (SUPELCO, length, 30 m; inner diameter,
0.32 mm; film thickness, 0.25 mm fused silica). The column
temperature was increased to 2508C at 108C/min. Electron im-
pact (EI) spectra were obtained at ꢀ70 eV. GC-MS raw data
were analyzed using the software package MassLab (Fisons).
The enantiomers were separated and quantified by GC-MS.
The enantiomers were separated as their BB derivatives on a
cyclodextrin column BGB-172 (length, 30 m; inner diameter
0.25 mm; film thickness, 0,25 mm), obtained from BGB Analy-
tik AG (Anwil, Switzerland). Samples of the BB derivatives
were diluted with cyclohexane and 1 mL of the diluted samples
was injected on column at 1208C. The column temperature was
programmed as follows: 28C/min to 2308C, and 108C/min to
2508C. All samples were analyzed by electron ionization (EI,
ꢀ70 eV) using full scan monitoring (m/z¼50 to 400). The
enantiomeric excess (ee) was defined as (A1 ꢀA2)/(A1 þ
A2)ꢂ100, where A1 and A2 are the peak areas of the BB deriv-
atives of the two cis-dihydrodiol enantiomers, and A1 was the
larger peak area.
Biotransformation of Benzonitrile on 2-L Scale
Biotransformation of benzonitrile in a reactor on 2-L scale was
performed as described elsewhere.[16] The cells were induced
with the addition of 0.5 mM IPTG or 0.05% DCPK to the reac-
tor. Benzonitrile was added to a final concentration of 2 mM
and then added repeatedly to keep the substrate concentration
between 0.5 and 2 mM. Samples were taken at different time
points, diluted 10 times with ice-cold acetonitrile and centri-
fuged at 48C. Supernatants were filtered through a cellulose
membrane filter 0.2 mm (Spartan, Schleicher & Schuell
GmbH, Germany) and analyzed. Biotransformations were car-
ried out twice and nearly identical results were obtained.
Biotransformation of Aromatic Nitriles in Reactors on
300-mL Scale
E. coli JM101 (pTEZ30) cells were grown and induced in a bi-
oreactor on a 2-L scale as described above. Four hours after in-
duction, cells were transferred into six sterile Sixfors bioreac-
tors) Infors, Bottmingen, Switzerland) (300 mL cell suspension
to each). The temperature was kept at 308C and the pH was ad-
justed to 7.2 with 25% NH4OH and 25% phosphoric acid. The
reactors were aerated at a rate of 2 liters per minute and stirred
at 300 rpm. Biotransformation was started by addition of sub-
strate. An additional amount of 0.3 mmol substrate was added
repeatedly every hour from a 100 mM stock solution in MeOH.
Liquid substrates were added directly to the same concentra-
tion. The reaction was continued for 7 hours. At the end of
the biotransformations, samples were taken from each reactor,
treated as described above and analyzed via HPLC.
1H NMR Spectroscopy: NMR spectra were recorded at
300 MHz using the Bruker Avance 300-MHz instrument with
QNP probehead and autosampler at ambient temperature.
Tetramethylsilane was used as an internal reference and fully
deuterated DMSO was used as solvent, unless stated other-
wise. Chemical shifts are given in parts per million relative to
tetramethylsilane (at 0 ppm).
Partitioning in Aqueous Two-Phase Systems
Aqueous two-phase systems were composed according to
standard procedures.[60] 30 mg benzonitrile (BN) or cis-1,2-di-
hydroxy-3-cyanocyclohexa-3,5-diene (BNDD) were added to
a vial containing about 5 mL upper phase and 5 mL lower
phase. The vial was sealed and shaken at 228C. After 90 mi-
nutes both phases were separated by centrifugation and ana-
lyzed with HPLC for benzonitrile or BNDD.
Product Recovery and Purification
The medium in the reactor was centrifuged at 500 rpm and at
48C for 15 minutes to separate the cells from the medium. To
decrease the volume, the supernatant was evaporated in a rota-
1070
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1060–1072

Recombinant Chlorobenzene Dioxygenase from Pseudomonas sp. P51
FULL PAPERS
[7] D. R. Boyd, N. D. Sharma, C. C. R. Allen, Curr. Opin.
Biotechnol. 2001, 12, 564–573.
[8] A. N. Phung, M. T. Zannetti, G. Whited, W. D. Fessner,
Angew. Chem. Int. Ed. 2003, 42, 4821–4824.
[9] J. M. Keith, J. F. Larrow, E. N. Jacobsen, Adv. Synth. Cat-
al. 2001, 343, 5–26.
[10] G. Bestetti, D. Bianchi, A. Bosetti, P. Di Gennaro, E.
Galli, B. Leoni, F. Pelizzoni, G. Sello, Appl. Microbiol.
Biotechnol. 1995, 44, 306–313.
Determination of Adsorption Isotherms
Adsorbent (XAD4, XAD7, charcoal, and Lewatit 1065, Sepa-
beads) was added (0.10–1.00 g) to a 10 mL aqueous solution of
benzonitrile (30 mg) or BNDD (30 mg), the mixture was
sealed and stirred at 228C. After 90 min, the liquid phase was
analyzed for benzonitrile or BNDD using HPLC as described
above. Their concentrations in the solid phase were derived
from the mass balance.
[11] M. G. Quintana, H. Dalton, Enzyme Microb. Technol.
1999, 24, 232–236.
[12] S. Panke, A. Meyer, C. M. Huber, B. Witholt, M. G.
Wubbolts, Appl. Environ. Microbiol. 1999, 65, 2324–
2332.
[13] S. Panke, M. G. Wubbolts, A. Schmid, B. Witholt, Bio-
technol. Bioeng. 2000, 69, 91–100.
[14] C. Werlen, H.-P. E. Kohler, J. R. van der Meer, J. Biol.
Chem. 1996, 271, 4009–4016.
Determination of Effects of Adsorbents and
Benzonitrile on Biomass Accumulation
The E. coli JM101 (pTEZ30) cells were grown in 19 mL M9
medium to a cell concentration of 0.35 g cdw/L and then 0.5 g
adsorbent was added. The suspension was incubated on a shak-
er at 308C and after 9 h the amount of cells was determined.
These experiments were repeated with the addition of benzo-
nitrile together with the adsorbents.
[15] A. Grund, J. Shapiro, M. Fennewald, P. Bacha, J. Leahy,
K. Markbreiter, M. Nieder, M. Toepfer, J. Bacteriol.
1975, 123, 546–556.
[16] S. Yildirim, J. Zezula, T. Hudlicky, B. Witholt, A. Schmid,
Adv. Synth. Catal. 2004, 346, 933–942.
Biotransformation of Benzonitrile with Integrated
Product Recovery
Biotransformation of benzonitrile in a 2-L reactor was per-
formed as described. 7.5 h after biotransformation had started,
the bioreactor medium was passed over an external loop,
packed with 35 g of activated charcoal by means of a peristaltic
pump. Details about the dimensions, set up, and preparation of
the external loop are described elsewhere.[61] A second adsorp-
tion cycle was performed after 17 h. Thereafter the biotransfor-
mation continued for another 7 hours.
[17] J. Brosius, A. Holy, Proc. Natl. Acad. Sci. USA 1984, 81,
6929–6933.
[18] H. Raschke, M. Meier, J. G. Burken, R. Hany, M. D.
Muller, J. R. van der Meer, H. P. E. Kohler, Appl. Envi-
ron. Microbiol. 2001, 67, 3333–3339.
[19] J. M. Sugihara, C. Bowman, J. Am. Chem. Soc. 1957, 80,
2443–2446.
[20] M. D. Lilly, J. M. Woodley, J. Ind. Microbiol. 1996, 17,
24–29.
[21] G. J. Lye, J. M. Woodley, Tibtech 1999, 17, 395–402.
[22] A. A. Garcia, D. H. Kim, G. Whited, L. Kwart, W. An-
thony, C. Downie, Isolation & Purification 1994, 2, 19–
25.
[23] R. M. Lynch, J. M. Woodley, M. D. Lilly, J. Biotechnol.
1997, 58, 167–175.
Acknowledgements
This work was supported by the Swiss Priority Program Bio-
technology (SPP Biotechnology) and Fluka Corporation
(Buchs, Switzerland). FAPESP (Brazil) is acknowledged for
funding a sabbatical stay of T. T. F. at the Department of Bio-
technology, ETH Zurich. We thank Dr. Adrie J. J. Straathof
for critical and most helpful discussions.
[24] D. Stark, U. von Stockar, Adv. Biochem. Eng. Biotechnol.
2003, 80, 149–175.
[25] A. Freeman, J. M. Woodley, M. D. Lilly, Bio-technology
1993, 11, 1007–1012.
[26] M. Held, S. Panke, H. P. E. Kohler, W. Suske, B. Witholt,
M. G. Wubbolts, Bio World 1999, 5, 3–9.
[27] J. T. Vicenzi, M. J. Zmijewski, M. R. Reinhard, B. E.
Landen, W. L. Muth, P. G. Marler, Enzyme Microb.
Technol. 1997, 20, 494–499.
[28] R. P. Chauhan, J. M. Woodley, Chemtech 1997, 27, 26–
30.
[29] M. A. Endoma, V. P. Bui, J. Hansen, T. Hudlicky, Org.
Process Res. Dev. 2002, 6, 525–532.
[30] B. Bꢀhler, A. Schmid, B. Hauer, B. Witholt, J. Biol.
Chem. 2000, 14, 10085–10092.
[31] M. Nieboer, M. Gunnewijk, J. B. van Beilen, B. Witholt,
J. Bacteriol. 1997, 179, 762–768.
[32] D. T. Gibson, S. M. Resnick, K. Lee, J. M. Brand, D. S.
Torok, L. P. Wackett, M. J. Schocken, B. E. Haighler, J.
Bacteriol. 1995, 177, 2615–2621.
References
[1] S. M. Roberts, Biocatalysts for fine chemicals synthesis,
(Eds.: M. S. Roberts, et al.), John Wiley & Sons, Chices-
ter, 1999.
[2] A. Schmid, J. S. Dordick, B. Hauer, A. Kiener, M. Wub-
bolts, B. Witholt, Nature 2001, 409, 258–268.
[3] K. M. Koeller, C.-H. Wong, Nature 2001, 409, 232–240.
[4] D. T. Gibson, V. Subramanian, (Eds.), Microbial degra-
dation of organic compounds. Microbial degredation of
aromatic hydrocarbons, (Ed.: D. T. Gibson), Marcel
Deckker, Inc., New York, 1984, pp. 181–252.
[5] D. R. Boyd, G. N. Sheldrake, Nat. Prod. Rep. 1998, 15,
309–324.
[6] T. Hudlicky, D. Gonzales, D. T. Gibson, Aldrichimica
Acta 1999, 32, 35–62.
Adv. Synth. Catal. 2005, 347, 1060–1072
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1071

Selcuk Yildirim et al.
FULL PAPERS
[33] K. Lee, D. T. Gibson, Appl. Environ. Microbiol. 1996, 62,
3101–3106.
[34] R. E. Parales, K. Lee, S. M. Resnick, H. Jiang, D. J. Less-
ner, D. T. Gibson, J. Bacteriol. 2000, 182, 1641–1649.
[35] K. Pollmann, V. Wray, H. J. Hecht, D. H. Pieper, Micro-
biology-Sgm 2003, 149, 903–913.
[48] O. Favre-Bulle, E. Weenink, T. Vos, H. Preusting, B.
Witholt, Biotechnol. Bioeng. 1993, 41, 263–272.
[49] L. E. Hꢀsken, H. H. Beeftink, J. A. M. de Bont, J. Wery,
Appl. Microbiol. Biotechnol. 2001, 55, 571–577.
[50] R. O. Jenkins, G. M. Stephens, H. Dalton, Biotechnol. Bi-
oeng. 1986, 29, 873–883.
[36] H. Suenaga, T. Watanabe, M. Sato, Ngadiman, K. Furu-
kawa, J. Bacteriol. 2002, 184, 3682–3688.
[51] C. J. Hack, J. M. Woodley, M. D. Lilly, J. M. Liddell, En-
zyme Microb. Technol. 2000, 26, 530–536.
[37] N. Zhang, B. G. Stewart, J. C. Moore, R. L. Greasham,
D. K. Robinson, B. C. Buckland, C. Lee, Metab. Eng.
2000, 2, 339–348.
[38] J. Reddy, C. Lee, M. Neeper, R. Greasham, J. Zhang,
Appl. Microbiol. Biotechnol. 1999, 51, 614–620.
[39] L. P. Wahbi, D. Gokhale, S. Minter, G. M. Stephens, En-
zyme Microb. Technol. 1996, 19, 297–306.
[52] A. M. Collins, J. M. Woodley, J. M. Liddell, J. Ind. Micro-
biol. 1995, 14, 382–388.
[53] D. R. Boyd, N. D. Sharma, B. Byrne, M. V. Hand, J. F.
Malone, G. N. Sheldrake, J. Blacker, H. Dalton, J.
Chem. Soc. Perkin Trans. 1 1998, 1935–1943.
[54] T. Hudlicky, M. A. A. Endoma, G. Butora, J. Chem. Soc.
Perkin Trans. 1 1996, 2187–2192.
[40] G. J. Zylstra, D. T. Gibson, J. Biol. Chem. 1989, 264,
14940–14946.
[55] W. K. Keener, M. E. Watwood, W. A. Apel, Appl. Micro-
biol. Biotechnol. 1998, 49, 455–462.
[41] C. C. R. Allen, D. R. Boyd, H. Dalton, N. D. Sharma,
S. A. Haughey, R. A. S. McMordie, B. T. McMurray,
G. N. Sheldrake, K. Sproule, J. Chem. Soc. Chem. Com-
mun. 1995, 119–120.
[42] W.-C. Suen, D. T. Gibson, Gene 1994, 143, 67–71.
[43] L. P. Wahbi, P. Phumathon, A. Brown, S. Minter, G. M.
Stephens, Biotechnol. Lett. 1997, 19, 961–965.
[44] S. Panke, M. Held, M. G. Wubbolts, B. Witholt, A.
Schmid, Biotechnol. Bioeng. 2002, 80, 33–41.
[45] M. G. Wubbolts, O. Favre-Bulle, B. Witholt, Biotechnol.
Bioeng. 1996, 52, 301–308.
[56] J. M. Carragher, W. S. McClean, J. M. Woodley, C. J.
Hack, Enzyme Microb. Technol. 2001, 28, 183–188.
[57] A. Amanullah, C. J. Hewitt, A. W. Nienow, C. Lee, M.
Chartrain, B. C. Buckland, S. W. Drew, J. M. Woodley,
Biotechnol. Bioeng. 2002, 80, 239–249.
[58] J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular cloning
a laboratory manual. 2nd edn., (Ed.: C. Nolan), New
York, Cold Harbor Spring Laboratory Press, 1989.
[59] B. Buhler, I. Bollhalder, B. Hauer, B. Witholt, A. Schmid,
Biotechnol. Bioeng. 2003, 81, 683–694.
[60] M. E. Silva, T. T. Franco, Brazil. J. Chem. Eng. 2000, 17.
[61] M. Held, A. Schmid, H. P. E. Kohler, W. Suske, B. With-
olt, M. G. Wubbolts, Biotechnol. Bioeng. 1999, 62, 641–
648.
[46] S. A. Rothen, M. Sauer, B. Sonnleitner, B. Witholt, Bio-
technol. Bioeng. 1998, 58, 356–365.
[47] B. Bꢀhler, I. Bollhalder, B. Hauer, B. Witholt, A. Schmid,
Biotechnol. Bioeng. 2003, 82, 833–842.
[62] D. Hanahan, J. Mol. Biol. 1983, 166, 557–580.
1072
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1060–1072