Functional divergence between closely related Baeyer-Villiger monooxygenases from Aspergillus flavus

Article abstract of DOI:10.1016/j.molcatb.2014.05.015

Baeyer-Villiger monooxygenases (BVMOs) catalyse the chemo-, regio- and enantioselective oxidation of ketones to esters and lactones. To date, most of the cloned BVMOs available are derived from bacteria, although Baeyer-Villiger oxidations using fungi have frequently been demonstrated. Here we report the cloning and characterization of four BVMOs from the fungus Aspergillus flavus NRRL3357. Phylogenetic analysis shows these four BVMOs to cluster in a distinct group apart from other well-characterized BVMOs including cyclohexanone, phenylacetone and 4-hydroxyacetophenone monooxygenase. Building on the Grogan classification/clustering of BVMOs, we have designated this new group of BVMOs, Group VI. Group VI BVMOs show an early divergence from the cyclopentanone monooxygenase (CPMO) type BVMOs (Group I). Substrate profiling using cyclic, bicyclic, aliphatic and aryl ketones show a clear divergence in function and specificity not only between this new group of BVMOs and the CPMO-type BVMOs, but also between the four A. flavus BVMO paralogues despite their high sequence similarity. This study not only contributes to the growing number of available BVMOs, but also addresses the current classification of Type I BVMOs, and the usefulness of phylogenetic clustering and prediction of function and selectivity when genome-mining is used to search for new biocatalysts.

Full text of DOI:10.1016/j.molcatb.2014.05.015

Accepted Manuscript
Title: Functional divergence between closely related
Baeyer-Villiger monooxygenases from Aspergillus flavus
Author: F.M. Ferroni M.S. Smit D.J. Opperman
PII:
S1381-1177(14)00165-9
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcatb.2014.05.015
MOLCAB 2963
To appear in:
Journal of Molecular Catalysis B: Enzymatic
Received date:
Revised date:
Accepted date:
3-4-2014
5-5-2014
5-5-2014
Please cite this article as: F.M. Ferroni, M.S. Smit, D.J. Opperman, Functional
divergence between closely related Baeyer-Villiger monooxygenases from
Aspergillus flavus, Journal of Molecular Catalysis B: Enzymatic (2014),
http://dx.doi.org/10.1016/j.molcatb.2014.05.015
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Functional divergence between closely related Baeyer‐Villiger
monooxygenases from Aspergillus flavus
F.M. Ferroni, M.S. Smit, D.J. Opperman (corresponding author)
University of the Free State, Department of Biotechnology, 205 Nelson Mandela Drive,
Bloemfontein, 9300, South Africa
e‐mail: opperdj@ufs.ac.za
Tel: +27 51 401 2714
Fax: +27 51 401 9376
1
Page 1 of 23

Abstract
Baeyer‐Villiger monooxygenases (BVMOs) catalyse the chemo‐ regio‐ and enantioselective oxidation
of ketones to esters and lactones. To date, most of the cloned BVMOs available are derived from
bacteria, although Baeyer‐Villiger oxidations using fungi have frequently been demonstrated. Here
we report the cloning and characterization of four BVMOs from the fungus Aspergillus flavus
NRRL3357. Phylogenetic analysis shows these four BVMOs to cluster in a distinct group apart from
other well‐characterized BVMOs including cyclohexanone, phenylacetone and 4‐
hydroxyacetophenone monooxygenase. Building on the Grogan classification/clustering of BVMOs,
we have designated this new group of BVMOs, Group VI. Group VI BVMOs show an early divergence
from the cyclopentanone monooxygenase (CPMO) type BVMOs (Group I). Substrate profiling using
cyclic, bicyclic, aliphatic and aryl ketones show a clear divergence in function and specificity not only
between this new group of BVMOs and the CPMO‐type BVMOs, but also between the four A. flavus
BVMO paralogues despite their high sequence similarity. This study not only contributes to the
growing number of available BVMOs, but also addresses the current classification of Type I BVMOs,
and the usefulness of phylogenetic clustering and prediction of function and selectivity when
genome‐mining is used to search for new biocatalysts.
Keywords
Baeyer‐Villiger monooxygenase, Aspergillus flavus, biotransformation, desymmetrization, kinetic
resolution
2
Page 2 of 23

1.
Introduction
Baeyer‐Villiger monooxygenases (BVMOs) are flavoproteins that catalyse the conversion of ketones
and cyclic ketones to esters and lactones respectively [1–4]. Oxygen insertion is mediated via a
peroxyflavin species that, analogous to the chemical Baeyer‐Villiger (BV) oxidation, adds
nucleophilically to the carbonyl group forming a tetrahedral Criegee intermediate, followed by
normal migration and cleavage. Unlike chemical BV reactions that require hazardous chemicals like
peracids or hydrogen peroxide as oxidant, BVMOs use dioxygen with water as the only by‐product.
In addition to BV reactions, BVMOs are also able to catalyse epoxidation [5] and heteroatom
oxidations including sulfoxidation [6–8] and N‐oxidation of amines [9–11]. This promiscuous nature,
wide‐substrate scope but high enantio‐, regio‐ and chemoselectivity, makes then promising catalysts
for synthetic organic chemistry. Their usefulness in asymmetric catalysis has further been
demonstrated in the desymmetrization of prochiral substrates [12,13] and the kinetic resolution of
racemic substrates [12–16].
As a result, much effort has gone into the discovery and cloning of BVMOs [17–30] as well as their
evolution towards broader substrate scopes, increased or inverted selectivity as well as stability [31–
43]. With the availability of a multitude of sequenced genomes, computational genome‐mining is
providing an easy alternative to activity screening of wild‐type microorganisms for the identification
of new BVMOs. This approach has been used to identify and characterize multiple new BVMOs from
the bacterium Rhodococcus jostii RHA1 [3,44,45]. Although it was fungal oxidation of steroids that
led to the discovery of Baeyer‐Villiger monooxygenases [46–48], most of the cloned and
characterized BVMOs currently are from prokaryotes. Despite the fact that BVMOs have been
shown to be prevalent in fungi [49–54], the relative ease of working with bacteria compared to fungi
might explain this disproportion in the BVMO literature. In addition, cloning and heterologous
expression of bacterial genes in E. coli can be performed directly from genomic DNA without the
concern of introns and codon usage differences. Only two fungal BVMOs have been cloned and
3
Page 3 of 23

described to date, cycloalkanone monooxygenase (CAMO) from Cylindrocarpon radicicola [55] and
BVMOAf1 from Aspergillus fumigatus [56], both closely related to the well‐characterized
cyclohexanone monooxygenase (CHMO) from Acinetobacter sp. NCIMB9871 and phenylacetone
monooxygenase (PAMO) from Thermobifida fusca.
We focused on a group of four BVMOs from Aspergillus flavus NRRL3357, phylogenetically distinct
from CHMOacinet and PAMO. Herein we describe the cloning, heterologous expression in E. coli and
the characterization of these four closely related BVMOs from Aspergillus flavus and their divergent
substrate specificity as well as divergent regio‐ and enantioselectivities.
2.
Experimental
2.1
Strains, vectors and culture conditions
Aspergillus flavus NRRL3357 was obtained from the Agricultural Research Service (ARS/NRRL) culture
collection and grown on Potato Dextrose Agar (PDA) at 25°C. CHMOacinet and A. flavus BVMOs were
expressed from the pET22b(+) vector (Novagen). E. coli strains were routinely grown in Lysogeny
Broth (LB) medium at 37°C with shaking (200 rpm). Expression of the CHMOacinet and A. flavus
BVMOs were done in E. coli BL21‐Gold (DE3) (Stratagene) in ZYP5052 medium [57] at 20°C (200 rpm,
48h).
2.2
Cloning and heterologous expression
Genomic DNA (gDNA) was isolated from PDA‐grown A. flavus cells using the ZR Fungal/Bacterial DNA
MicroPrep kit (Zymoresearch) from which the complete open reading frames (ORFs) were PCR
amplified. PCR reaction mixtures (50 µL) consisted of 10x KOD Hot Start Polymerase buffer (5 µL),
MgSO4 (1.5 mM), deoxynucleoside triphosphates (0.2 mM each), KOD Hot Start DNA polymerase (1
U), gDNA (50 ng), and both the forward and the reverse primers (Table S1, 0.3 µM). The reaction
conditions consisted of an initial denaturing step at 95°C for 2 min, followed by 18 cycles of
denaturing at 95°C (20 s), annealing at 59°C (10 s), and elongation at 70°C (1 min), with a final
4
Page 4 of 23

extension at 70°C for 4 min. PCR products were purified using the Biospin gel extraction kit (BioFlux)
and blunt‐end ligated into pSMART. Ligations were transformed into E. coli TOP10 competent cells
(Invitrogen) and selected for on LB‐plates containing 30 µg/mL kanamycin. Plasmid DNA was isolated
(Biospin plasmid DNA extraction kit, BioFlux) and verified by DNA sequencing. The complete ORFs
were shuttled to pET22b(+) for expression using NdeI and XhoI.
Constructs were transformed into E. coli BL21‐Gold (DE3) cells (Stratagene) and selected for on LB‐
plates containing 100 µg/mL ampicillin. Expression was performed using auto‐induction media
(ZYP50502) at 20°C for 48 h (200 rpm). Cells were harvested through centrifugation and re‐
suspended in 200 mM Tris‐HCl (pH 8) buffer. Expression levels were evaluated using SDS‐PAGE
analysis [58] with Precision Plus Dual Xtra protein standards (Bio‐Rad) as molecular mass markers
and stained using Coomassie brilliant blue R‐250.
2.3
Biotransformations
Whole‐cell biotransformations were performed in 40 mL amber glass vials using 1 ml reaction
volumes. The biotransformation reaction mixture (BRM) consisted of 0.1 g wet weight/mL in 200
mM Tris‐HCl (pH 8), 100 mM glucose and 100 mM glycerol. The reactions were initiated by the
addition of substrate (10 mM) dissolved in methanol. Reactions were performed at 20°C for 2 h,
where after the reactions were stopped and extracted using an equal volume (2 times 0.5 mL) of
ethyl acetate containing 2 mM 1‐undecanol or 2 mM 3‐octanol as internal standard. GC‐MS analysis
was carried out on a Finnigan Trace GC ultra (ThermoScientific) equipped with a FactorFour VF‐5ms
column (60 m x 0.32 mm x 0.25µm, Varian). Chiral separation (Table S2) was performed using either
a Chiraldex G‐TA or B‐TA column (30 m x 0.25 mm x 0.12 µm, Astec).
2.4
Phylogenetic analysis
Sequences were retrieved from the Aspergillus flavus NRRL3357 genome sequence deposited at the
National Center for Biotechnology Information (NCBI) using BLASTp. Multiple sequence alignments
5
Page 5 of 23

of the predicted amino acid sequences were performed using the MUSCLE EBI web tool
(http://www.ebi.ac.uk/Tools/msa/muscle) with the default parameters [59]. The best amino acid
substitution model was estimated for tree building using the MEGA 5 software [60]. The Whelan
and Goldman (WAG) model [61] was selected with a discrete Gamma distribution with 5 rate
categories and by assuming that a certain fraction of sites are evolutionary invariable. An un‐rooted
maximum likelihood tree was constructed using MEGA 5 using Nearest‐Neighbour‐Interchange (NNI)
with bootstrap support for individual nodes calculated on 500 replicates.
3.
Results and Discussion
3.1
BVMOs in Aspergillus flavus
Aspergillus flavus is a filamentous fungus that produces mycotoxins termed aflatoxins as secondary
metabolites [62]. The biosynthesis of these aflatoxins occur via a highly conserved pathway with
one of the intermediate steps carried out by a Baeyer‐Villiger monooxygenase (moxY) responsible
for conversion of versicolorone to versiconol acetate, or hydroxyversicolorone to versiconal
hemiacetal acetate [63]. Genome mining of the Aspergillus flavus NRRL 3357 through homology
searches using CHMOacinet as search query identified an additional 25 putative Type I BVMOs of
unknown function.
6
Page 6 of 23

Fig. 1. Phylogenetic analysis of the four selected Aspergillus flavus BVMOs with previously cloned
and characterized BVMOs. Relationships are shown as an un‐rooted maximum likelihood tree
(Wheland and Gold model) inferred using Nearest‐Neighbour‐Interchange with bootstrap support
(500 replicates) for individual nodes. Groups I – V as previously defined [44]. Sequences used for
comparison are CHMObrevi2: cyclohexanone monooxygenase from Brevibacterium sp. HCU;
CPMOcomamo: cyclopentanone monooxygenase from Comamonas sp. NCIMB9872; MKMpseudo: methyl
7
Page 7 of 23

ketone monooxygenase from Pseudomonas veronii MEK700; ACMOgordon: acetone monooxygenase
from Gordonia sp. TY‐5; CHMOacinet: cyclohexanone monooxygenase from Acinetobacter sp.
NCIMB9871; CHMOpolaro: cyclohexanone monooxygenase from Polaromonas sp. JS666; CHMOrhodoc
cyclohexanone monooxygenase from Rhodococcus sp. HI‐31; CAMOcylind: cycloalkanone
monooxygenase from Cylindrocarpon radicicola ATCC11011; BVMOAF1: BVMO from Aspergillus
fumigates Af293; CHMObrevi1: cyclohexanone monooxygenase from Brevibacterium sp. HCU;
PAMOthermo: phenylacetone monooxygenase from Thermobifida fusca YX; SMOrhodoco: steroid
monooxygenase from Rhodococcus rhodochrous; CDMOrhodoc: cyclododecanone monooxygenase
from Rhodococcus ruber SC1; CPDMOpseudo: cyclopentadecanone monooxygenase from
Pseudomonas sp. HI‐70; HAPMOpseudo: 4‐hydroxyacetophenone monooxygenase from Pseudomonas
fluorescens ACB; AKMOpseudo: aliphatic ketone monooxygenase from Pseudomonas fluorescens
DSM50106; BVMOpseudo: BVMO from Pseudomonas putida KT2440; EtaAmycoba: BVMO from
Mycobacterium tuberculosis; MO1‐MO24: BVMOs from Rhodococcus jostii RHA1; AFL456, AFL838,
AFL210 and AFL619: BVMOs from Aspergillus flavus NRRL3357.
:
Phylogenetic analysis (Figure 1) revealed that four of these A. flavus BVMOs cluster in a
phylogenetically distinct group, apart from CPMOcomamo (Group I), PAMOthermo (Group II), CHMOacinet
(Group III) as well as the other A. flavus BVMOs. Building on the Grogan‐grouping [44] we
designated this new group, Group VI. These four A. flavus BVMOs show an early divergence from
Group I, which contains CPMOcomamo and CHMObrevi2 as well as five of the Rhodococcus jostii BVMOs.
The four BVMOs share approximately 40 – 49% sequence identity and less than 35% sequence
identity to CHMOacinet. Multiple sequence alignments also showed no significant insertions or
deletions between the BVMOs (Figure S1). The two Rossmann‐fold motifs (GxGxxG/A) in the FAD
and NADPH binding domains are absolutely conserved and all four display the BVMO signature motif
(FxGxxxHxxxWP/D) [64] except BVMOAFL456 where tyrosine replaces both the phenylalanine and
tryptophan residues (YxGxxxHxxxYP) within the signature motif. Variation in this sequence motif
has previously been observed and therefore a second BVMO signature motif
8
Page 8 of 23

([A/G]GxWxxxx[F/Y]P[G/M]xxxD) located between the N‐terminal Rossmann‐fold motif and the first
BVMO signature motif have recently been proposed [45]. This motif contains the conserved Asp
residue responsible for proper NADPH binding and/or oxidation during catalysis [65]. Complete
conservation was found amongst the four BVMOs from A. flavus for this motif as well as the
absolutely conserved catalytic Arg residue (Figure S1).
3.2
Cloning and heterologous expression
Because the four BVMOs contained no predicted introns, the complete open reading frames could
be amplified and cloned directly from genomic DNA. Heterologous expression in E. coli yielded
soluble proteins, of approximately 60 kDa in size (Figure 2). BVMOAFL619 showed the lowest
expression levels and additional smaller bands observed on SDS‐PAGE suggests partial degradation
of the enzyme.
Fig. 2. SDS‐PAGE analysis of soluble BVMO production in E. coli BL21(DE3). Lane 1: Molecular weight
marker, lane 2: E. coli extract, lane 3: CHMOacinet, lane 4: BVMOAFL210, lane 5: BVMOAFL456, lane 6:
BVMOAFL619, lane 7: BVMOAFL838
.
3.3.
Substrate scope of BVMOs
Whole‐cell biotransformations using recombinant E. coli strains each expressing one the four BVMOs
(or CHMOacinet for comparison) were used to evaluate substrate scope and selectivity. Glucose and
9
Page 9 of 23

glycerol were included in the biotransformation mixture and E. coli native metabolism was used for
cofactor regeneration.
None of the A. flavus BVMOs tested (Table 1) was able to convert cyclooctanone or larger
cycloketones (4, 5). Only BVMOAFL210 showed significant activity towards cyclohexanone (2) and
cycloheptanone (3). Correspondingly, only BVMOAFL210 was able to convert the 2‐methyl (7), 3‐
methyl (9) and 4‐alkyl (10 ‐ 12) substituted cyclohexanones. Cyclopentanone (1) and its 2‐methyl
substituted derivative (6) were converted by both BVMOAFL210 and BVMOAFL838, whereas BVMOAFL456
only converted the alkyl substituted substrate. In the case of the 2‐substituted cyclic ketones (6‐8)
oxygen insertion occurred preferentially at the more nucleophilic α‐substituted carbon, to yield the
“normal” lactone (6a‐8a) in large excess. BVMOAFL210 yielded only the proximal lactone product (9a)
from the β‐substituted 3‐methylcyclohexanone (9), whereas CHMOacinet displays no regioselectivity,
with the proximal and distal products formed in near equal ratios.
Table 2 Whole‐cell biotransformations of cyclic alkanes
Scheme 1 Baeyer‐Villiger oxidation of monocyclic ketones
10
Page 10 of 23

Table 1 Whole‐cell biotransformations of cyclic ketones
Conversion [%] (e.e. [%])
Substrate CHMOacinet BVMOAFL210 BVMOAFL456 BVMOAFL619 BVMOAFL838
1
2
>99
>99
48
43.5
27
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
20
n.c.
3
>99
n.c.
n.c.
4
>99
n.c.
n.c.
n.c.
5
n.c.
n.c.
n.c.
n.c.
6
96a/4b
96a/4b
31a
75a/25b
41a/11b
<1a
35a/1b
n.c.
74a/2b
n.c.
7
8
n.c.
n.c.
9
54a/46b
>99
51a
n.c.
n.c.
10
58
n.c.
n.c.
>99 (‐)‐S
>99
91 (+)‐R
58
11
12
n.c.
n.c.
n.c.
n.c.
n.c.
6
87 (‐)
>99
80 (+)
37
94 (‐)
84 (+)
Reaction conditions: [substrates]= 10 mM, [whole‐cells] = 0.1 g/mL wet weight, reaction time = 2 h,
temperature = 20°C, buffer = 200 mM Tris‐HCl (pH 8). Conversions determined by GC‐MS analysis.
n.c. = no conversion. a= “normal lactone” b=abnormal lactone.
Different lengths of alkyl group substitution on the para‐position of cyclohexanone were used to
study the desymmetrization of prochiral substrates by BVMOAFL210. While CHMOacinet produced the
expected S‐enantiomer of the 4‐methylcyclohexanones (10) in enantiomeric excess of >99%,
BVMOAFL210 gave, similar to CPMOcomamo and CHMObrevi2, the enantiocomplentary (R)‐lactone.
Although not optically pure, BVMOAFL210 produced R‐5‐methyloxepan‐2‐one (R‐10a) in enantiomeric
excess of more than 90% compared to CPMOcomamo and CHMObrevi2 that gave only 46% and 44%
respectively [42,66]. Likewise, lactones enantio‐complementary to the ones produced by CHMOacinet
were produced from 4‐ethyl‐ and 4‐n‐propylcyclohexanone (11, 12) in high ee values (>80%, Table
1).
,
11
Page 11 of 23

Racemic 2‐ and 3‐methylcyclohexanone was used to compare the enantiospecificity of BVMOAFL210
with that of CHMOacinet. Unlike CHMOacinet that preferentially converted the S‐ketone of 2‐
methylcyclohexanone (rac‐7), BVMOAFL210 showed no catalytic preference between the enantiomers.
As with chemical Baeyer‐Villiger oxidations, the stereochemistry of the starting ketone is retained
and complete conversion of rac‐7 by CHMOacinet yielded approximately 50% S‐“normal” lactone and
45% R‐“normal” lactone. The less reactive ketone (R) was also converted to small amounts (5%) of
R‐“abnormal” lactone. BVMOALF210 converted the S‐enantiomer without any regioselectivity to equal
amounts of the “normal” and “abnormal” S‐lactone. R‐7 however was converted to a high
regiomeric ratio, yielding more than 96% of the R‐“normal” lactone. CHMOacinet catalyzes the
stereodivergent oxygenation of 3‐methylcyclohexanone to the corresponding regioisomeric
lactones, where the “normal”‐lactone is produced from R‐9, and the “abnormal”‐lactone from S‐9.
This reaction also proceeds with a slight rate preference for R‐9. BVMOALF210 showed a minimal rate
preference for the opposite substrate enantiomer, but absolute proximal regioselectivity, with only
the “normal”‐lactones produced from both S‐ and R‐9, yielding a racemic mixture upon complete
conversion, consistent with CPMOcomamo‐type activity [67,68].
With so little activity observed with typical cyclic ketones we resorted to testing the A. flavus BVMOs
against aliphatic ketones. 2‐Alkanones with chain lengths of C8 – C12 (Table 2) were used to
investigate conversion and regioselectivity. Gratifyingly, all the A. flavus BVMO were active against
all the linear ketones tested, and displayed 100% regioselectivity for the 2‐alkanones substrates,
yielding only the alkyl acetate products. Differences in the chain‐length preference were however
observed, with BVMOAFL619 showing better conversion with C10‐C11, BVMOAFL838 and BVMOAFL456
with C8.
12
Page 12 of 23

Scheme 2 Baeyer‐Villiger oxidation of aliphatic ketones
Table 2 Whole‐cell biotransformations of aliphatic ketones
Conversion (%)
Substrate CHMOacinet BVMOAFL210 BVMOAFL456 BVMOAFL619 BVMOAFL838
13
14
15
16
17
84a
97a/1b
76a
84a
94a
87a
85a
55a
>99a
5a
20a
8a
>99a
>99a
98a
97a
83a
56a
41a
41a
31a
87a
96a
56a
75a
Reaction conditions: [substrates]= 10 mM, [whole‐cells] = 0.1 g/mL wet weight, reaction time = 2 h,
temperature = 20°C, buffer = 200 mM Tris‐HCl (pH 8). Conversions determined by GC‐MS analysis.
n.c. = no conversion.
The importance of the position of the carbonyl group was investigated by comparing the conversion
of 2‐octanone (13) and 3‐octanone (14). Whereas BVMOAFL210 and BVMOAFL838 gave better or similar
conversion with 3‐octanone, the ability of BVMOAFL456 and BVMOAFL619 was greatly reduced. Again
oxygen insertion occurred selectively on the side of the longer alkyl chain, with the exception of
CHMOacinet that produced small amounts of the ethyl ester (14b).
Given their phylogenetic position, the preference for linear alkanones displayed by these four A.
flavus BVMOs was unexpected, since previously characterised members of the closely related
Groups I and VII have either shown preference for cyclic ketones or equal activity. Although not well
documented, CHMOacinet and CPMOcomamo have been shown to be able to convert linear ketones
13
Page 13 of 23

[44,45]. The BVMOs from R. jostii within the Group I all display slight activity towards 2‐octanone
and some of them towards 2‐decanone [45]. Group VII, the second closest related group, contains
methyl ketone monooxygenase (MKM/MekA) and acetone monooxygenase (ACMO). Both of these
BVMOs have shown comparable activity for linear and cyclic ketones [29,69]. On the other hand one
of the most distantly related BVMO groups, Group V, which is represented by the BVMO from
Pseudomonas putida KT2440 [27] and EtaA from Mycobacterium tuberculosis [70], have shown
minimal or no activity with typical cyclic ketones such as cyclohexanone. Group V BVMOs’
regioselectivity is also influenced by substrate size, where short‐chain aliphatic ketones (C8‐C10) are
converted selectively to alkyl acetates, whereas both alkyl acetates and methyl esters are produced
from C11‐C12 aliphatic ketones.
O
18 n=0 R=H
19 n=0 R=OH
20 n=1 R=H
21 n=2 R=H
22 n=2 R=OH
23 n=2 R=OMe
n
O
n
O
R
R
18a-23a
18-23
Scheme 3 Baeyer‐Villiger oxidations of aryl ketones
Table 3 Whole‐cell biotransformations of aryl ketones
Conversion (%)
Substrate CHMOacinet BVMOAFL210 BVMOAFL456 BVMOAFL619 BVMOAFL838
13
14
15
16
17
18
n.c.
n.c.
>99
51
52
16
69
74
46
63
32
89
n.c.
n.c.
35
12
96
>99
64
63
15
96
3.5
16
47
13
>99
59
65
15
Reaction conditions: [substrates]= 10 mM, [whole‐cells] = 0.1 g/mL wet weight for 18‐20 and 0.08
g/mL wet weight for 21‐23, reaction time = 2 h, temperature = 20°C, buffer = 200 mM Tris‐HCl (pH
8). Conversions determined by GC‐MS analysis. n.c. = no conversion.
14
Page 14 of 23

Unlike CHMOacinet, three of the A. flavus BVMOs were able to catalyse oxygen insertion between the
ketone group and the sp2 carbon of acetophenone (18) and 4‐hydroxyacetophenone (19) (Table 3).
Similar to HAPMO, BVMOAFL456 and BVMOAFL838 gave better conversions with the hydroxy substituted
substrate (19) whereas BVMOAFL210 gave higher conversions with acetophenone (18). In addition, all
four A. flavus BVMOs were also able to accept phenylacetone (20). These activities are usually
mutually exclusive with HAPMO (4‐hydroxyacetophenone monooxygenase) not converting
phenylacetone [71] and PAMO (phenylacetone monooxygenases) showing very low activity towards
4‐hydroxyacetophenone [30]. The effect of the distance between the ketone group and the
aromatic ring was also investigated by including benzylacetone (21). Two of the A. flavus BVMOs
(BVMOAFL456 and BVMOAFL619) showed lower activity with increasing distance, while the other two
preferred the longer distance. Hydroxy or methoxy substitutions on the para position of the
aromatic ring of benzylacetone (22 and 23 respectively) generally resulted in similar to lower
conversions.
Table 4 Whole‐cell biotransformations of bicyclic ketones and thioanisole
Conversion (%)
Substrate CHMOacinet BVMOAFL210 BVMOAFL456 BVMOAFL619 BVMOAFL838
24
25
26
1.6a
54a/46b
78a
41a
n.c.
94a
36a
n.c.
13b
2a
1.9a
62a
n.c.
87a/13b
14a/13b
Reaction conditions: [substrates]= 10 mM, [whole‐cells] = 0.1 g/mL wet weight for 24 and 26 and
0.08 g/mL for 25, reaction time = 2 h, temperature = 20°C, buffer = 200 mM Tris‐HCl (pH 8).
Conversions determined by GC‐MS analysis. n.c. = no conversion. a= “normal lactone” b=abnormal
lactone.
The bicyclic ketones α‐tetralone, β‐tetralone and 2‐indanone were not converted by any of the A.
flavus BVMOs, while only BVMOAFL210 gave significant activity with 1‐indanone (24) yielding a single
regioisomer (Table 4). Although all four A. flavus BVMOs were able to convert rac‐
15
Page 15 of 23

bicyclo[3.2.0]hept‐2‐en‐6‐one (25), the activity as well as regioselectivity differed significantly
between the different biocatalysts (Table 4). This is in contrast to previous results showing a high
conservation between phylogenetic grouping based on sequence data and the regioselectivity for 25
[72]. We therefore decided to investigate the kinetic resolution of 25 through time‐course reactions.
Similar to CHMOacinet, BVMOAFL210 showed no kinetic resolution of rac‐bicyclo[3.2.0]hept‐2‐en‐6‐one
(25), whereby both enantiomers of 25 was converted at the same rate (Figure 3). The remaining
three BVMOs however all preferentially oxidized (1R,5S)‐25, with (1S,5R)‐25 resolved in 99% ee (57%
conversion), 98% ee (56% conversion) and 93% ee (53% conversion) for BVMOAFL456, BVMOAFL838 and
BVMOAFL619 respectively. Oxidation of rac‐25 by CHMOacinet occurs via an enantiodivergent route,
whereby (1R,5S)‐25 is converted selectively to the “normal” or 2‐oxabicyclic lactone and (1S,5R)‐25
is converted to the “abnormal” or 3‐oxabicyclic lactone. BVMOAFL210 yielded the “normal” lactone
from (1S,5R)‐25 and both the regioisomers from (1R,5S)‐25, resulting in low ee for 25a. BVMOAFL456
and BVMOAFL838 both showed excellent regioselectivity with only the “normal” (25a) lactones
produced. BVMOAFL619 gave the regiocomplementary “abnormal” lactone (25b) from (1R,5S)‐25,
with the normal lactone (1R,5S)‐25a produced from (1S,5R)‐25 at a very slow rate. To our
knowledge this is the only BVMO described to date that preferentially gives (1S,5R)‐25b. Of the four
BVMOs from Rhodococcus josti in Group I, all preferentially formed the 2‐oxabicyclic lactones (25a)
from rac‐25 with modest to low kinetic resolution of the racemic starting ketone, with MO3 and
MO11 displaying opposite enantiopreference to MO4 and MO20 [44]. The remaining members of
Group I, CHMObrevi2 and CPMO, both display similar high regioselectivity for the formation of the
“normal” lactones but without any kinetic resolution of the substrate, resulting in near racemic
mixtures of the “normal” lactones [72].
16
Page 16 of 23

Fig. 3. Baeyer‐Villiger oxidation of rac‐bicyclo[3.2.0]hept‐2‐en‐6‐one (25) by CHMOacinet and A. flavus
BVMOs.
Lastly, heteroatom oxidation was investigated using thioanisole (26) as a substrate. Again different
profiles were observed (Table 4). BVMOAFL210 and BVMOAFL619 displayed comparable sulfoxidation
activity towards thioanisole, however BVMOAFL210 was able to further convert the sulfoxide (26a)
product to the corresponding sulfone (26b).
4.
Conclusions
Four new BVMOs from the fungus A. flavus have been functionally expressed in E. coli. These
BVMOs constitute a new subgroup (Group VI) of Type I BVMOs and show an early divergence from
the CPMOcomamo‐type (Group I) of BVMOs. Unlike BVMOs from the CPMOcomamo‐cluster, three of the
four BVMOs were largely unable to convert cyclic ketones and all four BVMOs efficiently converted
17
Page 17 of 23

the aliphatic ketones tested, reminiscent of the more distantly related EtaA‐type BVMOs (Group V).
Although multiple sequence alignments and phylogenetic groupings often serve as guides to group
and predict substrate preferences of BVMOs, substrate profiling for BV‐type reactions as well as
sulfoxidation revealed these four closely related BVMOs to display different activities and
selectivities.
Apart from the BVMOs from Rhodococcus jostii, most of the new BVMOs cloned and characterized
are orthologues, identified through genome‐mining using existing characterized BVMO sequences.
Orthologues generally retain their function with most of the sequence differences considered
neutral. Paralogues within the same organism that originated through gene‐duplication and
functional divergence, however often evolve new functions and specificity [73,74]. Although the
natural function of these four A. flavus BVMOs are unknown, apart from retaining BV‐type activity, a
clear divergence in their substrate specificity and selectivity is observed. This is comparable to what
will be observed when mutants from a directed evolution experiment are included in a phylogenetic
analysis, where phylogenetic grouping will highlight only the close evolutionary history of the wild‐
type and the mutant enzymes, but changed substrate scope and/or specificity will not be reflected.
We therefore propose that biocatalyst discovery through cloning paralogues from microorganisms,
and in particular fungi, containing multiple BVMOs should result in a more diverse biocatalytic
toolbox.
Acknowledgements
This work was supported by the National Research Foundation South Africa and SASOL Ltd. The
authors thank Mr. Sarel Marais for GC analysis.
References
[1]
K. Balke, M. Kadow, H. Mallin, S. Saß, U.T. Bornscheuer, Org. Biomol. Chem. 10 (2012) 6249–
6265.
18
Page 18 of 23

[2]
[3]
[4]
[5]
H. Leisch, K. Morley, P.C.K. Lau, Chem. Rev. 111 (2011) 4165–4222.
A. Riebel, G. de Gonzalo, M.W. Fraaije, J. Mol. Catal. B Enzym. 88 (2013) 20–25.
D.E. Torres Pazmiño, H.M. Dudek, M.W. Fraaije, Curr. Opin. Chem. Biol. 14 (2010) 138–144.
S. Colonna, N. Gaggero, G. Carrea, G. Ottolina, P. Pasta, F. Zambianchi, Tetrahedron Lett. 43
(2002) 1797–1799.
[6]
[7]
[8]
S. Colonna, N. Gaggero, P. Pasta, G. Ottolina, Chem. Commun. (1996) 2303–2307.
S. Colonna, N. Gaggero, G. Carrea, P. Pasta, Chem. Commun. (1998) 415–416.
F. Zambianchi, P. Pasta, G. Carrea, S. Colonna, N. Gaggero, J.M. Woodley, Biotechnol. Bioeng.
78 (2002) 489–496.
[9]
S. Colonna, V. Pironti, G. Carrea, P. Pasta, F. Zambianchi, Tetrahedron. 60 (2004) 569–575.
[10] S. Colonna, V. Pironti, P. Pasta, F. Zambianchi, Tetrahedron Lett. 44 (2003) 869–871.
[11] G. Ottolina, S. Bianchi, B. Belloni, G. Carrea, B. Danieli, Tetrahedron Lett. 40 (1999) 8483–
8486.
[12] M.D. Mihovilovic, G. Chen, S. Wang, B. Kyte, F. Rochon, M.M. Kayser, J.D. Stewart, J. Org.
Chem. 66 (2001) 733–738.
[13] M.J. Taschner, D.J. Black, Q.‐Z. Chen, Tetrahedron: Asymmetry. 4 (1993) 1387–1390.
[14] A. Kirschner, U.T. Bornscheuer, Angew. Chem. Int. Ed. Engl. 45 (2006) 7004–7006.
[15] K. Geitner, A. Kirschner, J. Rehdorf, M. Schmidt, M.D. Mihovilovic, U.T. Bornscheuer,
Tetrahedron: Asymmetry. 18 (2007) 892–895.
[16] C. Rodríguez, G. de Gonzalo, M.W. Fraaije, V. Gotor, Tetrahedron: Asymmetry. 18 (2007)
1338–1344.
[17] P.C. Brzostowicz, K.L. Gibson, S.M. Thomas, M.S. Blasko, P.E. Rouvière, J. Bacteriol. 182 (2000)
4241–4248.
[18] P.C. Brzostowicz, D.M. Walters, S.M. Thomas, V. Nagarajan, P.E. Rouvière, Appl. Environ.
Microbiol. 69 (2003) 334–342.
[19] Y.‐C.J. Chen, O.P. Peoples, C.T. Walsh, J. Bacteriol. 170 (1988) 781–789.
[20] H. Iwaki, Y. Hasegawa, S. Wang, M.M. Kayser, P.C.K. Lau, Appl. Environ. Microbiol. 68 (2002)
5671–5684.
[21] N.M. Kamerbeek, M.J. Moonen, J.G. Van Der Ven, W.J. Van Berkel, M.W. Fraaije, D.B.
Janssen, Eur. J. Biochem. 268 (2001) 2547–2557.
19
Page 19 of 23

[22] Y.‐M. Kim, S.‐H. Jung, Y.‐H. Chung, C.‐B. Yu, I.‐K. Rhee, Biotechnol. Bioprocess Eng. 13 (2008)
40–47.
[23] A. Kirschner, J. Altenbuchner, U.T. Bornscheuer, Appl. Microbiol. Biotechnol. 73 (2007) 1065–
1072.
[24] K. Kostichka, S.M. Thomas, K.J. Gibson, V. Nagarajan, Q. Cheng, J. Bacteriol. 183 (2001) 6478–
6486.
[25] B.G. Kyte, P. Rouvière, Q. Cheng, J.D. Stewart, J. Org. Chem. 69 (2004) 12–17.
[26] S. Morii, S. Sawamoto, Y. Yamauchi, M. Miyamoto, M. Iwami, E. Itagaki, J. Biochem. 126
(1999) 624–631.
[27] J. Rehdorf, A. Kirschner, U.T. Bornscheuer, Biotechnol. Lett. 29 (2007) 1393–1398.
[28] D. V. Rial, P. Cernuchova, J.B. van Beilen, M.D. Mihovilovic, J. Mol. Catal. B Enzym. 50 (2008)
61–68.
[29] A. Völker, A. Kirschner, U.T. Bornscheuer, J. Altenbuchner, Appl. Microbiol. Biotechnol. 77
(2008) 1251–1260.
[30] M.W. Fraaije, J. Wu, D.P.H.M. Heuts, E.W. van Hellemond, J.H.L. Spelberg, D.B. Janssen, Appl.
Microbiol. Biotechnol. 66 (2005) 393–400.
[31] H.M. Dudek, M.J. Fink, A. V Shivange, A. Dennig, M.D. Mihovilovic, U. Schwaneberg, M.W.
Fraaije, Appl. Microbiol. Biotechnol. (2013).
[32] H.L. van Beek, G. de Gonzalo, M.W. Fraaije, Chem. Commun. 48 (2012) 3288–90.
[33] Z.‐G. Zhang, L.P. Parra, M.T. Reetz, Chemistry (Easton). 18 (2012) 10160–72.
[34] D.J. Opperman, M.T. Reetz, Chembiochem. 11 (2010) 2589–2596.
[35] M.T. Reetz, B. Brunner, T. Schneider, F. Schulz, C.M. Clouthier, M.M. Kayser, Angew. Chemie.
116 (2004) 4167–4170.
[36] M.T. Reetz, Angew. Chem. Int. Ed. Engl. 50 (2011) 138–174.
[37] M.T. Reetz, F. Daligault, B. Brunner, H. Hinrichs, A. Deege, Angew. Chem. Int. Ed. Engl. 43
(2004) 4078–4081.
[38] M.T. Reetz, S. Wu, Chem. Commun. (2008) 5499–5501.
[39] M.T. Reetz, S. Wu, J. Am. Chem. Soc. 131 (2009) 15424–15432.
[40] A. Kirschner, U.T. Bornscheuer, Appl. Microbiol. Biotechnol. 81 (2008) 465–472.
[41] C.M. Clouthier, M.M. Kayser, M.T. Reetz, J. Org. Chem. 71 (2006) 8431–8437.
[42] C.M. Clouthier, M.M. Kayser, Tetrahedron: Asymmetry. 17 (2006) 2649–2653.
20
Page 20 of 23

[43] M.D. Mihovilovic, F. Rudroff, A. Winninger, T. Schneider, F. Schulz, M.T. Reetz, Org. Lett. 8
(2006) 1221–1224.
[44] C. Szolkowy, L.D. Eltis, N.C. Bruce, G. Grogan, Chembiochem. 10 (2009) 1208–1217.
[45] A. Riebel, H.M. Dudek, G. de Gonzalo, P. Stepniak, L. Rychlewski, M.W. Fraaije, Appl.
Microbiol. Biotechnol. 95 (2012) 1479–1489.
[46] G.E. Turfitt, Biochem. J. 42 (1948) 376–383.
[47] J. Fried, R.W. Thoma, A. Klingsberg, J. Am. Chem. Soc. 75 (1953) 5764–5765.
[48] D.H. Peterson, S.H. Eppstein, P.D. Meister, H.C. Murray, H.M. Leigh, A. Weintraub, L.M.
Reineke, J. Am. Chem. Soc. 75 (1953) 5768–5769.
[49] A. Keppler, A. Porto, I. Schoenleincrusius, J. Comasseto, L. Andrade, Enzyme Microb. Technol.
36 (2005) 967–975.
[50] G. Fantin, P.P. Giovannini, A. Guerrini, S. Maietti, A. Medici, P. Pedrini, Biotechnol. Lett. 28
(2006) 805–810.
[51] M. Kawamoto, T. Utsukihara, C. Abe, M. Sato, M. Saito, M. Koshimura, N. Kato, C.A. Horiuchi,
Biotechnol. Lett. 30 (2008) 1655–1660.
[52] T. Kołek, A. Szpineter, A. Swizdor, Steroids. 73 (2008) 1441–1445.
[53] B. Ratuś, W. Gładkowski, C. Wawrzeńczyk, Enzyme Microb. Technol. 45 (2009) 156–163.
[54] A.C. Hunter, H. Khuenl‐Brady, P. Barrett, H.T. Dodd, C. Dedi, J. Steroid Biochem. Mol. Biol. 118
(2010) 171–176.
[55] F. Leipold, R. Wardenga, U.T. Bornscheuer, Appl. Microbiol. Biotechnol. 94 (2012) 705–717.
[56] M.L. Mascotti, M. Juri Ayub, H. Dudek, M.K. Sanz, M.W. Fraaije, AMB Express. 3 (2013) 33.
[57] F.W. Studier, Protein Expr. Purif. 41 (2005) 207–234.
[58] U.K. Laemmli, Nature. 227 (1970) 680–5.
[59] R.C. Edgar, Nucleic Acids Res. 32 (2004) 1792–7.
[60] K. Tamura, D. Peterson, N. Peterson, G. Stecher, M. Nei, S. Kumar, Mol. Biol. Evol. 28 (2011)
2731–9.
[61] S. Whelan, N. Goldman, Mol. Biol. Evol. 18 (2001) 691–9.
[62] M.A. Klich, Mol. Plant Pathol. 8 (2007) 713–22.
[63] Y. Wen, H. Hatabayashi, H. Arai, H.K. Kitamoto, K. Yabe, Appl. Environ. Microbiol. 71 (2005)
3192–8.
21
Page 21 of 23

[64] M.W. Fraaije, N.M. Kamerbeek, W.J.H. van Berkel, D.B. Janssen, FEBS Lett. 518 (2002) 43–47.
[65] R. Orru, H.M. Dudek, C. Martinoli, D.E. Torres Pazmino, A. Royant, M. Weik, M.W. Fraaije, A.
Mattevi, J. Biol. Chem. 286 (2011) 29284 –29291.
[66] M.D. Mihovilovic, F. Rudroff, B. Müller, P. Stanetty, Bioorg. Med. Chem. Lett. 13 (2003) 1479–
1482.
[67] 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.
[68] S. Wang, M.M. Kayser, H. Iwaki, P.C.K. Lau, J. Mol. Catal. B Enzym. 22 (2003) 211–218.
[69] T. Kotani, H. Yurimoto, N. Kato, Y. Sakai, J. Bacteriol. 189 (2007) 886–893.
[70] M.W. Fraaije, N.M. Kamerbeek, A.J. Heidekamp, R. Fortin, D.B. Janssen, J. Biol. Chem. 279
(2004) 3354–3360.
[71] N.M. Kamerbeek, A.J.J. Olsthoorn, M.W. Fraaije, D.B. Janssen, Appl. Environ. Microbiol. 69
(2003) 419–426.
[72] M.D. Mihovilovic, F. Rudroff, B. Grötzl, P. Kapitan, R. Snajdrova, J. Rydz, R. Mach, Angew.
Chem. Int. Ed. Engl. 44 (2005) 3609–3613.
[73] M. Soskine, D.S. Tawfik, Nat. Rev. Genet. 11 (2010) 572–82.
[74] T. Gabaldón, E. V Koonin, Nat. Rev. Genet. 14 (2013) 360–6.
Graphical Abstract
22
Page 22 of 23

Highlights
•
•
•
•
Four new fungal Baeyer‐Villiger monooxygenases cloned and characterized
Phylogenetic subgroup diverged from CPMO‐type BVMOs
Paralogues showing different substrate scope and specificity
Kinetic resolution of rac‐bicyclo[3.2.0]hept‐2‐en‐6‐one
23
Page 23 of 23