Cloning and expression of a Baeyer-Villiger monooxygenase oxidizing linear aliphatic ketones from Dietzia sp. D5

Article abstract of DOI:10.1016/j.molcatb.2014.08.020

A Baeyer-Villiger monooxygenase has been identified in the genome sequence of Dietzia sp. D5. Sequence similarity search revealed that the enzyme belongs to a group of BVMOs that are closely related to ethionamide monooxygenase from Mycobacterium tuberculosis (EthA). The BVMO was expressed in E. coli BL21-CodonPlus(DE3)-RP and the best expression was achieved when the E. coli cells were cultivated in terrific broth (TB) at 15 °C and induced with 0.1 mM of IPTG. Since the purified enzyme did not show any measurable activity, the substrate scope of the BVMO has been determined using whole-cell and crude cell extract systems. The enzyme was most active towards linear aliphatic substrates. However, it has shown a moderate degree of conversion for cyclobutanone, 2-methylcyclohexanone, bicyclo[3.2.0]hept-2-en-6-one, phenylacetone and thioanisole. There was no detectable conversion of ethionamide, cyclohexanone and acetophenone.

Full text of DOI:10.1016/j.molcatb.2014.08.020

Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Cloning and expression of a Baeyer–Villiger monooxygenase oxidizing
linear aliphatic ketones from Dietzia sp. D5
a,∗
a
a,b
a
a
Serena Bisagni , Justyna Smu s´ , Georgina Chávez , Rajni Hatti-Kaul , Gashaw Mamo
a
Department of Biotechnology, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden
Instituto de Investigaciones Fármaco Bioquímicas, Av. Saavedra No. 2224, La Paz, Miraflores, Bolivia
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 May 2014
Received in revised form 20 August 2014
Accepted 30 August 2014
Available online 8 September 2014
A Baeyer–Villiger monooxygenase has been identified in the genome sequence of Dietzia sp. D5. Sequence
similarity search revealed that the enzyme belongs to a group of BVMOs that are closely related to
ethionamide monooxygenase from Mycobacterium tuberculosis (EthA). The BVMO was expressed in E. coli
BL21-CodonPlus(DE3)-RP and the best expression was achieved when the E. coli cells were cultivated in
◦
terrific broth (TB) at 15 C and induced with 0.1 mM of IPTG. Since the purified enzyme did not show any
measurable activity, the substrate scope of the BVMO has been determined using whole-cell and crude cell
extract systems. The enzyme was most active towards linear aliphatic substrates. However, it has shown
a moderate degree of conversion for cyclobutanone, 2-methylcyclohexanone, bicyclo[3.2.0]hept-2-en-6-
one, phenylacetone and thioanisole. There was no detectable conversion of ethionamide, cyclohexanone
and acetophenone.
Keywords:
Dietzia
Baeyer–Villiger monooxygenase
Aliphatic ketones
Ethionamide
©
2014 Elsevier B.V. All rights reserved.
1
. Introduction
BVMOs can also be categorized on the basis of their pri-
mary sequence as Type I, Type II and Type O [3]. Among these
three groups, the most studied BVMOs are Type I BVMOs which
contain FAD as a prosthetic group and are NADPH dependent
[4]. Several reports have suggested a further division of Type I
BVMOs based on the degree of sequence similarity [5–7]. One of
such subdivisions contains ethionamide monooxygenase (EthA)
and other enzymes having close sequence similarity with EthA.
EthA monooxygenase was isolated from Mycobacterium tubercu-
losis H37Rv and was identified to be responsible for the activation
of the pro-drug ethionamide into the bio-active sulfoxide inter-
mediate [8,9]. In addition to ethionamide, the enzyme is active
towards aliphatic ketones, such as 2-octanone and 2-decanone, and
phenylacetone [9]. Besides EthA monooxygenase, M. tuberculosis
H37Rv has two other monooxygenases, MO0565c and MO3083,
which are known to catalyze the sulfoxidation of thioanisole
and MO3083 can also oxidize 2-octanone and bicyclohept-2-
en-6-one [10]. Among the 23 BVMOs from Rhodococcus jostii
RHA1, only two, MO13 and MO16, share significant similarity
with EthA monooxygenase. MO16 oxidizes different substrates
(cyclobutanone, 2-metylcyclopentanone, bicyclohept-2-en-6-one
and 2-octanone) but the substrate scope of MO13 is not yet
available [5,7]. A BVMO from Pseudomonas putida KT2440 is the
other characterized enzyme that shows sequence similarity to
EthA monooxygenase; the enzyme oxidizes 4-decanone with the
highest efficiency [11]. A recent report shows the presence of
Baeyer–Villiger monooxygenases (BVMOs) catalyze the oxi-
dation of ketones and heteroatoms (i.e. sulphur, nitrogen,
phosphorous, boron and selenium) to lactones and oxides, respec-
tively [1]. BVMOs are considered promiscuous enzymes which
oxidize a wide variety of substrates. Among the ketones, small and
bulky cyclic ketones, as well as ketones in multicyclic structures and
linear aliphatic ketones with and without aromatic substituents
can be oxidized by BVMOs. However, each BVMO presents a well
defined substrate preference and can oxidize only a limited num-
ber of substrates. A systematic investigation and classification of
the substrate scope of some BVMOs have been recently reported
[
2].
Abbreviations: BVMO, Baeyer–Villiger monooxygenase; BVMO3, Dietzia sp. D5
BVMO; EthA, ethionamide monooxygenase; MO0565c and MO3083, Mycobacterium
tuberculosis BVMOs; MO13 and MO16, Rhodococcus jostii BVMOs; BVMO Pp, Pseu-
domonas putida BVMO; BVMO Ar, Acinetobacter radioresistens BVMO; BVMO Php,
Physcomitrella patens BVMO; BVMO Cm, Cyanidioschyzon merolae BVMO; BVMO Pf,
Pseudomonas fluorescens DSM 50106 BVMO; BVMO Pv, Pseudomonas veronii MEK700
BVMO; AFL 210, AFL 456 AFL 619 and AFL 838 Aspergillus flavus BVMOs; GDH, glucose
dehydrogenase.
∗ Corresponding author. Tel.: +46 46 2224741; fax: +46 46 2224713.
E-mail addresses: serena.bisagni@biotek.lu.se, serena.bisagni@gmail.com
(
S. Bisagni).
http://dx.doi.org/10.1016/j.molcatb.2014.08.020
381-1177/© 2014 Elsevier B.V. All rights reserved.
1

1
62
S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
another BVMO oxidizing ethionamide from Acinetobacter radiore-
sistens [12].
(TB, composed of tryptone 12 g, yeast extract 24 g, glycerol 4 ml,
filtered solution of 0.17 M KH PO4 and 0.72 M K HPO4 100 ml per
2
2
Recently, we have sequenced the genome of a strain of Dietzia.
Four BVMOs have been identified from the draft genome sequence
and one of the BVMOs has been characterized and shown to react
with sulfides and aldehydes with rare regiospecificity [13,14]. One
of the Dietzia sp. D5 BVMOs, named BVMO3, is closely related to
EthA monooxygenase and in this paper, we report the cloning,
expression and substrate scope of this BVMO.
litre) and M9 medium (1 M MgSO4 2 ml, 20% glucose 20 ml, 1 M
CaCl2 0.1 ml, M9 salts 200 ml per litre), respectively. The M9 salts
comprised Na HPO ·7H O l64 g, KH PO 15 g, NaCl 2.5 g and NH Cl
2
4
2
2
4
4
5 g and water to a final volume of 1 l. The media when required
contains 100 ␮g/ml ampicillin, 34 ␮g/ml chloramphenicol, and/or
20 ␮g/ml gentamycin.
Cultivation of E. coli for the recombinant protein expression was
initiated by inoculating the culture media with a pre-inoculum
2
. Experimental
corresponding to 5% of the final culture volume of 200 ml in 1 l
◦
flask. After 3 h of cultivation at 30 C, the OD6 was about 0.6
00
2
.1. Organisms, plasmids and chemicals
and the protein expression was induced adding IPTG or lactose to
final concentration of 0.1 mM and 10 mM respectively. During the
expression phase the cells were cultivated at 15 C for 16 h with
◦
E. coli BL21-CodonPlus(DE3)-RP and ArcticExpress(DE3)-RP
were purchased from Agilent Technologies (Santa Clara, USA). E. coli
NovaBlue, E. coli BL21(DE3), E. coli Rosetta2(DE3) and the plasmid
pET-22b(+) were purchased from Novagen (Darmstad, Germany).
QIAGEN Plasmid Mini Kit and QIAEX II Gel Extraction Kit (Qiagen,
Sollentuna, Sweden) were used to extract plasmids from cells and
DNA from agarose gel, respectively. Genomic DNA was extracted
using ZR Fungal/Bacterial DNA MiniPrep (Zymo Research, Irvine,
USA) from Dietzia sp. D5 which was isolated in our laboratory.
All chemicals used in the study are of the highest available purity
obtained from standard sources.
shaking at 150 rpm. The culture was divided into five aliquots of
equal volume that were harvested by centrifuging for 10 min at 4 C
◦
and 9820 × g using Sorvall RC5C centrifuge. Then each aliquot was
re-suspended in different buffers. The buffers used were: 50 mM
sodium phosphate buffer pH 7.5, 100 mM potassium phosphate
buffer pH 7.5, 100 mM potassium phosphate buffer pH 7.5 with
0.1% Triton X-100 (v/v), 100 mM potassium phosphate buffer pH 7.5
with 10% glycerol (v/v) and 100 mM potassium phosphate buffer pH
7.5 with 1 g/L Bovine Serum Albumin (BSA). The cell suspension was
placed on ice and lysed in three intermittent cycles of 45 s sonica-
tion (Hierscher UP400S Ultrasonicator; amplitude 50%, cycle 0.5)
with 1 min break. The sonicated cell suspension was centrifuged
2.2. Sequence analysis
◦
for 15 min at 4 C and 15,000 × g to remove cell debris and the clear
The multiple sequence alignment analysis of BVMO3 and the
supernatant was used as the enzyme source.
other EthA-like monooxygenases was done using the T-coffee algo-
rithm with default parameters and the output figure was prepared
using CLC Main Workbench. The multiple sequence alignment was
used to derive the phylogenetic tree using Clustal W2 and visual-
ization was obtained using FigTree.
The secondary structures of the BVMOs were predicted using
the software CLC Main Workbench.
The sequence of BVMO3 has been deposited in GenBank
database under the accession number AHE80562.
The BVMO was purified from the clear supernatant using immo-
2+
bilized metal ion affinity chromatography on a Ni bound column
TM
HisTrap FF (GE Healthcare, Uppsala Sweden) and eluted with an
imidazole gradient from 0 to 300 mM in 100 mM potassium phos-
phate buffer pH 7.5 containing 200 mM NaCl and 10% glycerol. After
protein elution, FAD was added to the fraction containing the puri-
fied BVMO3. Finally, desalting and concentration was performed
using Vivaspin 20 MWCO 30,000 Da centrifugal concentrators (Sar-
torius Stedim Biotech GmbH, Goettingen, Germany).
The activity of the recombinant enzyme in the crude cell extract
and in the purified form was followed by measuring the conversion
of 25 mM 2-nonanone by gas chromatography as described below.
One enzyme unit is expressed as the amount of enzyme that oxi-
dizes one ␮mol of 2-nonanone per minute under the standard assay
conditions.
2
.3. Cloning of BVMO3 gene
The gene encoding BVMO3 was amplified from the
genomic DNA of Dietzia sp. D5 using
a pair of primers,
BVMO3-F: ATTACCATGGCTGGTAGCACCCACCTC and BVMO3-
R: ATTACTCGAGTGATCGGGCCACCTCGTC, which were designed
based on the BVMO3 gene sequence identified in the draft genome
sequence. The forward and reverse primers had NcoI and XhoI
2.5. Biotransformations using whole cells and crude cell extract
(
underlined) restriction sites, respectively. High Fidelity PCR
enzyme mix (Fermentas, Gothenburg, Sweden) was used to
amplify the gene following the manufacturer instructions. DMSO
Whole-cell biotransformations were performed using both
growing and resting cells. For growing cells, 2 ml TB medium in
a 15 ml Falcon tube was inoculated with a pre-culture of E. coli
BL21-CodonPlus(DE3)-RP cells containing the recombinant plas-
(
5% v/v) was added to the PCR mix to improve the amplification.
After purification with Qiagen PCR cleaning kit, the PCR product
was digested with NcoI and XhoI and ligated to the expression
vector pET-22b(+) digested with the same restriction enzymes.
The ligation product was transformed into electrocompetent
E. coli NovaBlue cells and spread on Luria Bertani (LB) agar plates
containing ampicillin. Colonies were picked from the agar plates
and recombinant plasmids were extracted and sequenced. The
plasmid containing the correct sequence was transformed into the
expression hosts E. coli BL21(DE3), E. coli BL21-CodonPlus(DE3)-RP
and E. coli ArcticExpress(DE3)-RP.
◦
mid with the BVMO gene. The cells were grown for 3 h at 30 C
and with shaking at 150 rpm. Immediately after adding IPTG, the
biotransformation reaction was initiated by adding the substrate to
a final concentration of 5 mM. To boost the regeneration of NADPH,
glucose (25 mM) was also added at the moment of induction. Bio-
◦
transformations were carried out for 16 h at 15 C and at a shake
rate of 150 rpm. The level of conversion was measured by compar-
ing E. coli BL21-CodonPlus(DE3)-RP expressing BVMO3 with cells
from the same E. coli strain carrying pET-22b(+) without the BVMO3
gene, that served also as a blank for side reactions.
2
.4. Protein expression and purification
Biotransformation with resting cells was done using a culture
grown as described above in Section 2.4. At the end of the overnight
expression, the cells were harvested by centrifugation and re-
suspended in an equal volume of 20 mM sodium phosphate buffer
The recombinant E. coli cells were cultivated using low salt LB
(
tryptone 10 g, yeast extract 5 g, NaCl 5 g per litre), Terrific Broth

S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
163
pH 7.4 containing 5 mM substrate and 25 mM glucose and incu-
◦
bated further for 16 h at 15 C and 150 rpm.
Biotransformations with crude cell extract were performed by
incubating mixtures containing 25 mM glucose, 0.4 U of glucose
+
dehydrogenase (GDH), 133 ␮M NADP , 25 mM substrate (from a
1
1
M stock solution in ethanol) and the crude cell extract for 24 h at
5 C and shaking at 150 rpm.
◦
2.6. GC analysis
A sample of the biotransformation reaction (400 ␮l) was
extracted twice with ethyl acetate (500 ␮l twice) and analyzed
using Varian 430-GC gas chromatograph (Agilent Technologies,
Santa Clara, USA). Separation of analytes was done using FactorFour
VF-1 ms 15 m × 0.25 mm (0.25 ␮m) column (Varian, Strathaven,
◦
◦
◦
UK) with 1 min hold at 50 C; heating ramp from 50 C to 225 C
◦
◦
at a rate of 25 C/min; and 2 min hold at 225 C (the injector and
the detector temperature was maintained at 275 C). Identification
◦
of the products was done by comparison with retention time of
commercial standards.
3
. Results
3.1. Sequence analysis
The BVMO3 encoding gene was identified during the draft
genome sequence analysis of Dietzia sp. D5. The protein sequence
similarity search for BVMO3 revealed that it is related to a group of
BVMOs containing EthA [9] and other EthA-like BVMOs [5,10,11].
The phylogenetic relationship of these BVMOs and other well
known and characterized BVMOs is shown in Fig. 1. The tree illus-
trates that BVMO3 shares least similarity within the group. The
sequence identity between BVMO3 and EthA-like proteins is in the
range of 43 and 38% as shown in Table 1 of the Supplementary
Information.
Fig. 1. Phylogenetic tree derived from sequence alignment of representa-
tive BVMOs and BVMO3 from Dietzia sp. D5. The accession number of each
BVMO sequence considered in the alignment is given in bracket. BVMO3
from Dietzia sp. D5 (AHE80562.1), MO0565c-Mt from Mycobacterium tuberculo-
sis H37Rv (NP 215079.1), ETHA3854c-Mt from Mycobacterium tuberculosis H37Rv
(NP 218371.1), MO16-Rj from Rhodococcus jostii RHA1 (YP 702882.1), BVMO-Pp
from Pseudomonas putida KT2440 (NP 744949.1), MO13-Mt Rhodococcus jostii RHA1
(
YP 703731.1), MO3083-Mt from Mycobacterium tuberculosis H37Rv (NP 217599.1),
The multiple sequence alignment of EthA-like enzymes (Fig. 2)
shows that residues that are conserved among the other members
of the group are not maintained in BVMO3 sequence. Although in
some cases the change is simple substitution such as I to L or K to R,
there are several cases where BVMO3 residues differ substantially
from the conserved amino acids. The substituted amino acids are
not part of the conserved cofactor binding motifs but it cannot be
speculated if they are active site residues since the structure of
these enzymes is not known.
BVMO-Ar from Acinetobacter radioresistens (ADF32068), BVMO-Pa from Pseu-
domonas aeruginosa PAO1 (NP 250229.1), BVMO-Pf from Pseudomonas fluorescens
DSM50106 (AAC36351.2), HAPMO-Pf from Pseudomonas fluorescens ACB (Q93TJ5.1),
HAPMO-Pp from Pseudomonas putida JD1 (ACJ37423.1), BVMO4-D from Dietzia
sp. D5, CPDMO-PHI-70 from Pseudomonas sp. HI-70 (BAE93346.1), CHMO-Ac is
CHMO Acinetobacter sp. NCIMB9871 (BAA86293.1), CHMO-RHI-31 from Rhodococ-
cus sp. HI-31 (3UCL A), PAMO-Tf from Thermobifida fusca YX (YP 289549.1), SMO-Rr
from Rhodococcus rhodochrous (BAF48129.1) and CPMO-C from Comamonas sp.
NCIMB9872 (Q8GAW0.3).
The predicted secondary structures of BVMO3 and the other
BVMOs belonging to the EthA-group are compared in Fig. 3. The
aim of this analysis is to compare the secondary structure simi-
larity of the BVMOs in the group. The comparison shows that the
overall secondary structure elements of the different proteins over-
lap to some extent. The structures are more homogeneous in the
N-terminal region in the first part of the protein sequence until
the end of the ␣-helix region 1, instead in the remaining regions
of the protein (␣-helix region 2 and ␤-sheet regions 1 and 2) there
is no unanimous alignment between the protein considered in the
analysis.
Transformation of the BVMO3 gene in the commonly used
E. coli expression host BL21(DE3) did not give any detectable
expression (Fig. 4). This was presumably caused by the high GC
content (>70%) of the gene and the presence of several rare codons.
Thus, E. coli strains designed to overcome codon bias and high GC
content gene expression problems, namely E. coli Rosetta2(DE3),
E. coli BL21-CodonPlus(DE3)-RP and E. coli ArcticExpress(DE3)-RP,
were chosen. Active BVMO3 was expressed by E. coli BL21-
CodonPlus(DE3)-RP and E. coli ArcticExpress(DE3)-RP which carry a
plasmid for the expression of rare tRNAs. The ArcticExpress strain
encodes also for cold induced chaperones (one of them having a
molecular weight of 60 kDa, which in the SDS-PAGE overlaps the
3.2. Cloning, expression and purification
6
1.6 kDa protein band of BVMO3) and hence the BVMO expres-
◦
The BVMO3 gene was cloned in the commercial expression
sion with ArcticExpress cells had to be performed at 10–12 C to
allow the co-expression of the chaperones. However the low tem-
perature for cultivation resulted in low cell density, even when
the expression phase was prolonged to 24 h. Moreover, E. coli
BL21-CodonPlus(DE3)-RP have shown higher biotransformation
conversions than E. coli Arctic-Express(DE3)-RP (data not shown).
Therefore, E. coli BL21-CodonPlus(DE3)-RP was chosen as produc-
tion host for further studies.
vector pET-22b(+). The gene is 1556 nucleotides long and the calcu-
lated mass of the recombinant protein is 61.6 kDa. This calculation
takes into account the mass of the leader peptide PelB and the His-
tag. Although the gene is expressed with PelB which is assumed
to help exporting the protein to the periplasmic region in order to
avoid protein aggregation, the expressed protein remained in the
cytoplasm and resulted mainly in inclusion bodies (Fig. 4).

1
64
S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
Fig. 2. Multiple sequence alignment of EthA-like monooxygenases. Red stars represent the residues that are unique in BVMO3, blue dots highlight the residues with different
physicochemical properties exclusive for BVMO3.
Fig. 3. Predicted secondary structure of BVMO3 aligned to the predictions from the other members of EthA group. White arrows represent ␣-helix and grey arrows represent
␤
-sheets.

S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
165
Fig. 4. SDS-PAGE comparing the expression of BVMO3 in different E. coli strains. The protein concentration of each soluble crude fraction was measured prior to SDS-PAGE
and a similar amount of protein was loaded for each E. coli strain. Lanes 1 and 10: AllBlue Marker; lanes 2 and 6: soluble and insoluble crude cell extract fractions from E. coli
BL21(DE3); lanes 3 and 7: soluble and insoluble crude cell extract fractions from E. coli Rosetta2(DE3); lanes 4 and 8: soluble and insoluble crude cell extract fractions from
E. coli ArcticExpress(DE3)-RP; lanes 5 and 9: soluble and insoluble crude cell extract fractions from E. coli BL21-CodonPlus(DE3)-RP.
Optimization of some expression parameters including cultiva-
tion temperature, inducer type and concentration, and cultivation
media, was carried out in order to increase the expression of soluble
BVMO3 (data not shown). The highest amount of soluble enzyme in
E. coli BL21-CodonPlus(DE3)-RP was obtained when the cells were
was measured in case of cyclobutanone, 2-methylcylohexanone,
bicyclo[3.2.0]hept-2-en-6-one, phenylacetone and thioanisole by
growing recombinant cells, while acetophenone, cyclohexanone
and even ethionamide were not converted by BVMO3 (Table 1).
On the other hand, excellent conversions were achieved for linear
aliphatic ketones, with the highest degree of conversion measured
for 2-nonanone and 3-decanone by growing cells (Table 2). Inter-
estingly, when the keto-group is on the second carbon there was
only one oxidation product but when it is on the third carbon two
products are observed by gas chromatography. The ratios among
the two products vary for the different 3-keto substrates (Table 2).
Bioconversion of aliphatic ketones was also investigated using the
recombinant resting cells and crude cell extract (Table 2). Resting
cells were prepared by resuspending the recombinant E. coli cells
after overnight expression in sodium phosphate buffer. The conver-
sions achieved by the resting cells were lower than that of growing
cells or crude cell extract (Table 2).
◦
induced with 0.1 mM IPTG, grown at 15 C and in TB medium. These
conditions were used in all the subsequent studies.
It has been suggested that EthA-like BVMOs could be membrane
associated [9,15], therefore the choice of a proper cell lysis buffer is
very important to keep the protein in its active form in the soluble
fraction of the crude cell extract. The choice of the buffer salts and
the additives was based on reports available from scientific liter-
ature [8,9,14,16]. Although the different resuspension buffers did
not influence greatly the amount of BVMO3 present in the soluble
crude cell extract, the SDS-PAGE analysis revealed that the Triton
X-100 had slightly increased the amount of soluble BVMO (Fig. 1
Supplementary Information). However, the activity of the BVMO3
in potassium phosphate buffer containing glycerol was higher than
the ones measured in the other buffers (Fig. 5). Contrary to what was
shown in the SDS-PAGE, the activity of the cell extract containing
Triton X-100 was lower than the other preparations.
The stability of BVMO3 in the crude cell extract was estimated
by measuring the conversion of 2-nonanone after 14 days of stor-
◦
◦
age at 4 C or at −20 C (Fig. 5). The enzyme activity was strongly
◦
reduced after two weeks storage in the refrigerator (4 C), while it
◦
was maintained or decreased slightly when stored at −20 C. There-
fore, potassium phosphate buffer with 10% glycerol was used as cell
resuspension buffer for the following experiments.
The recombinant enzyme was purified using metal ion affinity
chromatography, but the purified enzyme showed very little activ-
ity (<5% conversion) that was completely lost after few days (data
not shown).
3.3. Biotransformations using growing cells, resting cells and
crude cell extract
Fig. 5. Comparison of the enzymatic activity of BVMO3 in different cell lysis buffers.
A range of molecules containing different functional groups
ketones with aliphatic, alicyclic or aromatic moieties and sulp-
hides) that are known to be transformed by other BVMOs were
used as substrates (Tables 1 and 2). Little to moderate conversion
The enzymatic activity was measured immediately after cell lysis (black bars) and
(
◦
◦
after 14 days when the enzyme was stored at 4 C (white bars) and −20 C (grey bars).
The experiment was done twice and standard deviations were calculated from the
experimental results.

1
66
S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
Table 1
◦
Substrates conversion by BVMO3. Whole-cell biotransformations with growing cells were performed with 5 mM substrate. The reaction mixtures were incubated at 15
with shaking at 150 rpm for 16 h.
C
Substrate
Products
Growing cells
Cyclobutanone
34.2%
Cyclohexanone
NC
2
-Methylcyclohexanone
21.0%
Bicyclo[3.2.0]hept-2-en-6-one
13.5%
Phenylacetone
Acetophenone
35.2%
NC
Ethionamide
Thioanisole
NC
32.0%
NC: not converted.
Table 2
Conversion of aliphatic ketone substrates by BVMO3. Whole-cell biotransformations with growing and resting cells and crude cell extract were performed with 5 mM
substrate, while biotransformations with crude cell extract contained 25 mM substrate and glucose/glucose dehydrogenase for cofactor regeneration. The reaction mixtures
◦
were incubated at 15 C with shaking at 150 rpm for 16 h (whole cells and resting cells) or 24 h (crude cell extract). In brackets is the ratio between product a and b, respectively.
Substrate
Product
Conversion
Growing cells
Resting cells
Traces
Crude cell extract
73.6%
2
-Heptanone
-Octanone
41.1%
60.8%
2
6.6%
78.3%
2
2
2
-Nonanone
-Decanone
-Undecanone
89.8%
80.8%
63.9%
3.2%
10.5%
NC
82.4%
83.7%
57.9%
a
3
-Octanone
44.3% (1:1)
5.9% (2:3)
43.6% (1:1)
b
a
b
a
b
3
3
-Nonanone
-Decanone
64.2% (1:1)
85.9% (2:1)
1.6% (2:3)
45.9% (1:2)
59.0% (1:1)
Traces (ND)
NC: not converted.
ND: not determined.

S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
167
The reactions prepared with crude cell extract exhibited com-
plete conversion of 5 mM substrate after overnight incubation,
hence a higher substrate concentration (25 mM) was used in the
conversion studies. The cofactor NADPH was enzymatically regen-
erated by glucose dehydrogenase and the co-substrate glucose
was added in stoichiometric amount to the substrate. The high-
est substrate conversion by crude cell extract was achieved for
2
-nonanone and 2-decanone (Table 2).
When the enzyme was assayed using 2-nonanone as substrate,
the activity in the crude cell extract and the growing cells prepara-
tions was about 14.3 U/ml and 4.7 U/ml, respectively.
4
. Discussion
Dietzia sp. strains are known to have a variety of oxygenases
[
17–19]. This work presents cloning, expression and substrate
scope of a BVMO isolated from Dietzia sp. D5. Sequence analysis of
this monooxygenase revealed that it is an EthA monooxygenase-
like enzyme and shares high sequence similarity with other
characterized BVMOs from M. tuberculosis H37Rv [9,10], P. putida
KT2440 [11] and R. jostii RHA1 [5]. There has been a great interest
on EthA monooxygenase mainly because of its physiological role
of converting the pro-drug ethionamide to the biologically active
form. Mutations on this gene and its regulator EthR are the cause
of the formation of ethionamide-resistant strains of M. tuberculosis
[
20].
BVMO3 amino acid sequence differs at sites that are highly con-
served in the EthA and EthA-like monooxygenases (Fig. 2). Further
analysis highlighting these differences is given by the prediction of
the secondary structure of BVMO3 and the other EthA-like BVMOs,
presented in Fig. 3. It seems that the secondary structure is not
strictly conserved among this group of enzymes. On the N-terminus
side, BVMO3 secondary structure resembles that of EthA monooxy-
genase, MO16 and BVMO Pp. The common pattern is present up to
the amino acid residue 200, corresponding to the chunk of sequence
containing the two Rossmann folds and the so-called BVMO motif
[
21]. A similar analysis made for other groups of BVMOs, namely
cyclohexanone monooxygenase-like BVMOs and phenylacetone
monooxygenase-like BVMOs shows that there is a much higher
degree of consensus among the secondary structures and a pattern
of ␣-helix and ␤-sheets throughout the whole secondary struc-
ture alignment can easily be identified (Fig.2a and b Supplementary
Information). This observation suggests that the sequence similar-
ity of EthA-like BVMOs might not correspond to similar protein
folding, however, to fully further support this statement, determi-
nation of crystal structure for one or more enzymes of this type is
necessary.
Although members of this BVMO group are expected to have
a similar substrate scope, the results from previously published
characterizations do not support this assumption (Table 3). Inter-
estingly, BVMO3 did not oxidize ethionamide but catalyzes the
oxidation of other sulphides, such as thioanisole. The ability to
oxidize thioanisole seems to be a common feature among this
group of BVMOs and several other BVMOs which do not belong
to this group. The oxidation of cyclohexanone and acetophenone
by EthA-like BVMOs is very poor, while different levels of con-
versions are reported for the oxidation of bicyclohept-2-en-6-one,
phenylacetone and 2-octanone. BVMO3 exhibited strong prefer-
ence towards linear aliphatic ketones such as 2-nonanone and
3
-decanone; similar behaviour was observed for P. putida BVMO
and EthA monooxygenase from M. tuberculosis but other enzymes
of the group (i.e. MO3083 and MO0565c) do not show such a prefer-
ence (Table 3). The reason why the other enzymes did not convert
aliphatic ketones might be related to their poor expression, such
as in the case of MO13 from Rhodococcus [7], due to choosing

1
68
S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
a restricted number of substrates to investigate the reactivity of
the BVMOs (i.e. in case of MO3083c and MO0565 the only linear
aliphatic ketone tested was 2-octanone) or due to intrinsic proper-
ties.
The substrate scope of EthA-like and non EthA-like BVMOs
which are able to convert linear aliphatic ketones is compared in
Table 3. The non EthA-like BVMOs share less than 20% sequence
identity with BVMO3 but can oxidize linear aliphatic ketones with
good or moderately good conversion. When substrates other than
linear aliphatic ketones are considered, differences in the conver-
sion obtained emerge among these enzymes. Interestingly, AFL 456,
AFL 619 and AFL 838 are all unable to oxidize cyclohexanone that
is a characteristic property of EthA-like BVMOs. In particular AFL
process, especially when the gene is high GC content and con-
tains several rare codons. This could be solved by synthesizing
the gene of interest, however screening of commercially available
E. coli strains that overcame the problem could be an alternative.
E. coli BL21-CodonPlus(DE3)-RP and E. coli ArcticExpress(DE3)-RP
demonstrated to produce soluble protein. The enzymatic activity
measurement indicated that E. coli BL21-CodonPlus(DE3)-RP was
a better expression host for BVMO3. The protein expression level
was further increased by optimizing a number of parameters, such
as culture media, inducer type and concentration and cultivation
temperature. The expression level might be further increased if
the expression vector could be changed, for instance a vector with
fusion tags that increases protein solubility [23].
6
19 from Aspergillus flavus converts approximately the same sub-
An interesting finding of this study is the influence of the resus-
pension buffer on the recovery of functional enzyme in the crude
cell extract. It was already suggested that additives can increase the
stability of BVMOs after cell lysis and in particular these additives
can be fundamental when the protein interacts with the cell mem-
brane, as it has been speculated for this type of BVMOs [8,9,15]. The
presence of potassium appears to have a slightly beneficial effect.
In agreement with previous reports [8,9], Triton X-100 slightly
increased the amount of protein in the crude cell extract but it has
a negative effect on the activity of BVMO3 (Fig. 5). Glycerol proved
to be a good stabilizer for the enzyme, especially when the enzyme
strates as BVMO3, despite their low overall sequence similarity.
This observation suggests that BVMO’s sequence similarity does
not always indicate substrate scope similarity and that the experi-
mental characterization of the enzyme is the only tool to determine
the substrate preference of BVMOs.
Comparison of the predicted secondary structure of the aliphatic
ketone-converting BVMOs with EthA-like BVMOs (Fig. 2c Supple-
mentary Information) shows that, in addition to low sequence
identity, there is great heterogeneity in the distribution of the
secondary structure motifs along the protein sequence. This obser-
vation hints that, despite the low overall sequence similarity and
differences in the secondary structure, BVMOs may exhibit similar
substrate scope if the active site area is similar. The information
about which residues form the active site in BVMO3 or other EthA-
like BVMOs is still limited as none of these enzymes has been
crystallized and its three dimensional structure determined.
Among the substrates tested, BVMO3 has shown high conver-
sions for linear aliphatic ketones. The maximum conversion of
◦
was stored at −20 C (Fig. 5). Glycerol and other polyols are known
to provide stabilizing effect through preferential exclusion mecha-
nism [24]. Irrespective of the resuspension buffer used, prolonged
◦
storage of BVMO3 at 4 C of BVMO3 reduced the enzymatic activity,
suggesting that this BVMO is rather unstable. In fact, the purified
BVMO3 lost all of its activity rapidly. Such loss of activity could be
reduced upon understanding the mechanism of inactivation of the
enzyme and thereby modifying the purification strategy (i.e. using
purification buffers with detergents or hydrophobic molecules,
testing other types of protein chromatography) and enzyme for-
mulation for storage. This would allow further characterization of
BVMO3.
8
9.8% and 80.8% was reached for 2-nonanone and 3-decanone,
respectively by growing cells (Table 2).
In whole-cell biotransformations, higher conversion yields were
obtained with growing cells as compared to the resting cells, even
though it was previously reported that resting cells biotransfor-
mation resulted in higher conversion [11]. It has of course to be
underlined that when using whole-cells for biotransformations
such as the E. coli strain used, the metabolic efficiency of the cells
after expression phase, the efficiency of the cofactor regeneration
efficiency or the loss of the cofactors due to cell permeabilization
and oxidative stress can affect the oxidation efficiency [22].
In case of the crude cell extract, the amount of substrate used
was five-times higher compared to that used for the whole-cell
biotransformations. When 5 mM of the linear ketones were added
to the crude cell extract the conversion was complete within a
short time, possibly because of the higher enzyme load or the
reduced mass transfer limitations due to absence of cell barriers
and easy mixing. Aliphatic ketones were confirmed to be the most
favoured substrates for BVMO3. The differences between conver-
sion by whole-cell and in crude cell extract systems were negligible
when the aliphatic 2-ketones are considered. Nevertheless, the con-
version of 3-nonanone and 3-decanone by the crude cell extract
was less efficient. Moreover the ratio between the oxidation prod-
ucts varied between biotransformation methods.
5. Conclusions
BVMO3 from Dietzia sp. D5 has been characterized for its sub-
strate scope and linear aliphatic ketones are shown to be the
preferred substrates. Although this enzyme shows high sequence
similarity with previously reported EthA and EthA-like BVMOs, it
was shown that they have differences in conserved amino acids
and BVMO3 could not oxidize ethionamide. Furthermore, a com-
parison with other BVMOs with similar substrate scope suggested
that there is no clear correlation between sequence similarity and
substrate scope. This opens up the challenge to identify the residues
that determine substrate specificity in this group of BVMOs, which
could be achieved with the help of X-ray crystallography and sys-
tematic mutagenesis studies.
Acknowledgments
This research was supported by Marie Curie Networks for Initial
Training fellowship in the project “BIOTRAINS” (FP7-PEOPLE-ITN-
There was no detectable conversion of cyclohexanone, ace-
tophenone and ethionamide. Cyclohexanone and acetophenone are
expected to be poor substrates for this group of BVMOs (Table 3);
however, ethionamide which is a known substrate for EthA [8,9]
was not converted by BVMO3. This evidence suggests once again
that sequence similarity does not ensure similar substrate scope for
this group of enzymes.
2
008-238531) and the Swedish Agency for Research Co-operation
with Developing Countries (Sida-SAREC).
Appendix A. Supplementary data
Although a satisfactory level of soluble recombinant protein
expression was achieved using E. coli BL21-CodonPlus(DE3)-RP,
this report shows that BVMO expression can be a cumbersome
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.molcatb.
2014.08.020.

S. Bisagni et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 161–169
[14] S. Bisagni, R. Hatti-Kaul, G. Mamo, AMB Express 4 (2014) 1–10.
169
References
[
[
15] L.N. Britton, A.J. Markovetz, J. Mol. Chem. 252 (1977) 8561–8566.
16] A. Völker, A. Kirschner, U. Bornscheuer, J. Altenbuchner, Appl. Microb. Bio-
technol. 77 (2008) 1251–1260.
17] L. Procópio, V. Alvarez, D. Jurelevicius, L. Hansen, S. Sørensen, J. Cardoso, M.
Pádula, A. Leitão, L. Seldin, J. Elsas, Antonie van Leeuwenhoek 101 (2012)
[
[
1] G. de Gonzalo, M.D. Mihovilovic, M.W. Fraaije, ChemBioChem 11 (2010)
208–2231.
2] M.J. Fink, D.V. Rial, P. Kapitanova, A. Lengar, J. Rehdorf, Q. Cheng, F. Rudroff,
M.D. Mihovilovic, Adv. Synth. Catal. 354 (2012) 3491–3500.
3] H. Leisch, K. Morley, P.C.K. Lau, Chem. Rev. 111 (2011) 4165–4222.
4] D.E. Torres Pazmi n˜ o, M.W. Fraaije, in: M. Tomoko (Ed.), Future Directions in
Biocatalysis, Elsevier Science B.V., Amsterdam, 2007, pp. 107–127.
2
[
289–302.
[
[
[
[
[
[
[
18] J.A. Linares-Pastén, G. Chávez-Lizárraga, R. Villagomez, G. Mamo, R. Hatti-Kaul,
Enz. Microb. Technol. 50 (2012) 101–106.
19] Y. Nie, J.L. Liang, H. Fang, Y.Q. Tang, X.L. Wu, Appl. Microb. Biotechnol. 98 (2014)
[
[
[
[
[
5] C. Szolkowy, L.D. Eltis, N.C. Bruce, G. Grogan, ChemBioChem 10 (2009)
163–173.
1
208–1217.
20] F. Brossier, N. Veziris, C. Truffot-Pernot, V. Jarlier, W. Sougakoff, Antimicrob.
Agents Chemother. 55 (2011) 355–360.
21] M.W. Fraaije, N.M. Kamerbeek, W.J.H. van Berkel, D.B. Janssen, FEBS Lett. 518
6] M.J. Fink, T.C. Fischer, F. Rudroff, H. Dudek, M.W. Fraaije, M.D. Mihovilovic, J.
Mol. Catal. B: Enzym. 73 (2011) 9–16.
7] A. Riebel, H. Dudek, G. de Gonzalo, P. Stepniak, L. Rychlewski, M.W. Fraaije,
Appl. Microbiol. Biotechnol. 95 (2012) 1479–1489.
(
2002) 43–47.
22] F. Heuser, K. Marin, B. Kaup, S. Bringer, H. Sahm, Metab. Eng. 11 (2009)
78–183.
8] T.A. Vannelli, A. Dykman, P.R. Ortiz de Montellano, J. Biol. Chem. 277 (2002)
1
1
2824–12829.
[
[
[
23] D. Esposito, D.K. Chatterjee, Curr. Opin. Biotechnol. 17 (2006) 353–358.
24] K. Gekko, S.N. Timasheff, Biochemistry 20 (1981) 4667–4676.
25] E. Beneventi, M. Niero, R. Motterle, M.W. Fraaije, E. Bergantino, J. Mol. Catal. B:
Enzym. 98 (2013) 145–154.
9] M.W. Fraaije, N.M. Kamerbeek, A.J. Heidekamp, R. Fortin, D.B. Janssen, J. Biol.
Chem. 279 (2004) 3354–3360.
10] D. Bonsor, S.F. Butz, J. Solomons, S. Grant, I.J.S. Fairlamb, M.J. Fogg, G. Grogan,
[
Org. Biol. Chem. 4 (2006) 1252–1260.
[
[
26] A. Kirschner, J. Altenbuchner, U. Bornscheuer, Appl. Microb. Biotechnol. 73
[
[
11] J. Rehdorf, A. Kirschner, U. Bornscheuer, Biotechnol. Lett. 29 (2007) 1393–1398.
12] D. Minerdi, I. Zgrablic, S.J. Sadeghi, G.G. Gilardi, Microb. Biotechnol. 5 (2012)
(
2007) 1065–1072.
27] F.M. Ferroni, M.S. Smit, D.J. Opperman, J. Mol. Catal. B: Enzym. 107 (2014)
7–54.
7
00–716.
4
[
13] S. Bisagni, B. Summers, S. Kara, R. Hatti-Kaul, G. Grogan, G. Mamo, F. Hollmann,
Top. Catal. 57 (2014) 366–375.