Family clustering of Baeyer-Villiger monooxygenases based on protein sequence and stereopreference

Article abstract of DOI:10.1002/anie.200462964

(Chemical Equation Presented) The identification of enzyme pairs with overlapping substrate specificity and enantiocomplementary transformations is a key challenge in biocatalysis. Enantio and regiodivergent Baeyer-Villiger oxidations were successfully carried out by using a small library of recombinant Escherichia coli strains expressing monooxygenases of various microbial origin (see picture). The clustering of enzymes based on stereopreference is in good agreement with phylogenetic similarity.

Full text of DOI:10.1002/anie.200462964

Angewandte
Chemie
Biotransformations
Family Clustering of Baeyer–Villiger
Monooxygenases Based on Protein Sequence and
Stereopreference**
Marko D. Mihovilovic,* Florian Rudroff, Birgit Grꢀtzl,
Peter Kapitan, Radka Snajdrova, Joanna Rydz, and
Robert Mach
Since its discovery by Adolf von Baeyer and Victor Villiger in
[
1]
1
899, the oxidation process later named after the two
scientists has become a powerful tool in synthesis to break
[
2]
carbon–carbon bonds in an oxygen-insertion process. The
regiochemistry of the reaction is governed by predictable
[
*] Prof. Dr. M. D. Mihovilovic, Dipl.-Ing. F. Rudroff, Dipl.-Ing. B. Grꢀtzl,
Dipl.-Ing. P. Kapitan, MSc R. Snajdrova, MSc J. Rydz
Institute of Applied Synthetic Chemistry
Marie Curie Training Site GEMCAT
Vienna University of Technology
Getreidemarkt 9/163-OC, 1060 Wien (Austria)
Fax: (+43)1-58801-15499
E-mail: mmihovil@pop.tuwien.ac.at
Prof. Dr. R. Mach
Institute of Chemical Engineering
Vienna University of Technology
Wien (Austria)
[
**] Financial support by the European Commission under the Human
Potential Program of FP-5 (contract no. HPMT-CT-2001-00243) and
by the Austrian Science Fund (FWF; project no. P16373) is gratefully
acknowledged. The authors thank Dr. Pierre E. Rouviere (E.I.
DuPont Company) for supporting this project by the generous
donation of six Escherichia coli expression systems for Baeyer–
Villiger monooxygenases. Assistance by Dr. Erwin Rosenberg
(Vienna University of Technology) during the determination of
enantiomeric purity and by Dr. Christian Hametner (Vienna
University of Technology) for NMR-based structure analysis is
acknowledged.
Supporting Information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 3609 –3613
DOI: 10.1002/anie.200462964
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3609

Communications
[
3]
conformational, steric, and electronic effects, and the
rearrangement process of the tetrahedral peroxo Criegee
and characterization of enzymes with overlapping substrate
specificity that yield antipodal products is a key issue for the
further implementation of biocatalytic methods in synthetic
chemistry.
[
4]
intermediate proceeds with strict retention of configuration.
These factors are key prerequisites for performing the
Baeyer–Villiger oxidation in an enantioselective manner.
The conversion of cyclic ketones into optically pure
lactones (Scheme 1), in particular, allows access to highly
Recently, we and others have observed the formation of
antipodal lactones by some representatives of the BVMO
[
13]
enzyme family. This study compares the stereopreference,
with respect to enantio- and regiodivergence, of cyclohex-
anone (CH) and cyclopentanone (CP) monooxygenases
[
14]
originating from Acinetobacter (CHMOAcineto),
Arthro-
[
15]
[16]
bacter (CHMOArthro), Brachymonas (CHMOBrachy), Bre-
[
17]
vibacterium (CHMOBrevi1
,
CHMOBrevi2),
and Rhodococcus (CHMORhodo1, CHMO-
species in recombinant whole-cell-mediated
Comamonas
[
18]
(
CPMOComa),
[
15]
Rhodo2)
Scheme 1. Baeyer–Villiger oxidation of cyclic ketones 1 to form
lactones 2.
Baeyer–Villiger oxidations with Escherichia coli as the host
organism.
Initially, desymmetrization of prochiral ketones 1a–i to
the corresponding lactones 2a–i, in part potential precursors
in natural product synthesis, was investigated (Table 1). In a
series of monocyclic ketones with prochiral substitution
patterns, a significant clustering into two groups was
observed: while the majority of BVMOs (“CHMO type”)
gave (À)-2a–d and (+)-2e lactones, CPMOComa and
flexible compounds as platforms for the subsequent synthesis
of bioactive compounds and natural products. Consequently,
enantioselective Baeyer–Villiger oxidations have become a
[
5]
highly active field in asymmetric chemistry in recent years.
Currently, two major strategies are being developed with
implementation of the “green-chemistry” concept aimed at
sustainable, environmentally benign, and atom-efficient proc-
esses. Metal-based, de novo designed chiral catalysts have
been continuously improved and are becoming promising
CHMO
(“CPMO type”) gave the antipodal products
Brevi2
with moderate to excellent enantiomeric excess. This general
trend is only violated by the enzyme CHMOBrevi1, which
displays the stereopreference of a CHMO-type BVMO but
does not accept 4,4-disubstituted ketone 1c. As observed
[
6]
candidates for industrial-scale applications.
By taking
[
19]
advantage of the vast catalytic repertoire of enzymes in
nature, biocatalysis offers alternative entities for stereoselec-
tive oxidation processes with molecular oxygen utilized as the
previously for similar hydroxy compounds, the oxidation of
1c does not yield the expected seven-membered-ring lactone
but rearranges under biotransformation conditions to give the
more stable five-membered-ring system.
[
7]
oxidant. An increasing number of flavin-containing Baeyer–
Villiger monooxygenases (BVMOs) have been identified
during the past decade, and several such proteins display a
remarkably broad acceptance profile for nonnatural sub-
The enantiodivergent trend in biooxidation was also
observed for fused bicycloketones 1 f–h. Generally, moderate
to excellent stereoselectivity was obtained upon biooxidation
with CHMO-type enzymes. CPMO-type BVMOs gave lac-
tones with chirality consistent with the two-enzyme-groups
hypothesis but with lower selectivity for 2 f.
[
8]
strates.
Our approach to overcome some of the BVMO limita-
tions, which have hampered the widespread utilization of
oxidizing enzymes by synthetic chemists, utilizes living whole
cells that are genetically engineered to express the required
Bridged bicyclo precursor 1i is only oxidized by CPMO-
type enzymes to form lactone 2i. Together with previous
[
9]
[20]
protein in high concentration. This concept simplifies the
problem of cofactor recycling, which arises because BVMOs
require nicotinamide adenine dinucleotide (phosphate) in the
reduced form (NAD(P)H) in the initial step of the catalytic
studies of classical kinetic resolutions, this is to some extent
an exception of the hypothesis of stereodivergent biotrans-
formations. However, the clearly differentiated substrate
acceptance again supports the classification of the studied
BVMOs into two groups.
[
10]
cycle. In addition, the tedious process of protein isolation is
overcome and enzyme stability is not a limiting factor. As a
result of the genetic modifications, the overexpressed BVMO
becomes the major fraction in the organismꢀs proteome, and
side reactions by competing enzymes can be essentially
Representatives of this enzyme library exhibited superior
enantioselectivities for desymmetrizations with all ketones
compared to those observed in previously reported bacterial
BVMO oxidations. The potential of enantiodivergent biocat-
alysts with overlapping substrate acceptance for natural
product synthesis is demonstrated by accessing various
[
11]
avoided. This strategy was successfully optimized
and
[
12]
scaled up in pilot-plant industrial fermentation facilities
very recently.
[
21]
indole alkaloids such as alloyohimbane from (À)-2h and
[
22]
The second major challenge in biocatalysis in general is
the aspect of enantiodivergence. Artificial catalytic entities
can be readily tailored to produce antipodal forms of the
required products by inverting the chirality of the inducing
ligand field. This strategy cannot be transferred to biotrans-
formations, as there is no efficient process available to yield
d-amino acid based proteins. Consequently, the identification
antirhine from the antipodal (+)-2h.
Intrigued by the significantly different behavior of the two
BVMO clusters in the desymmetrization reactions, we inves-
tigated the regiodivergent transformation of fused ketones
bearing a cyclobutanone structural motif (Scheme 2). This
conversion is considered to be one of the “benchmark”
[
23]
reactions for asymmetric Baeyer–Villiger oxidations and
3
610
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 3609 –3613

Angewandte
Chemie
[
a]
Table 1: Enantiodivergent Baeyer–Villiger oxidation of prochiral ketones 1 by recombinant E. coli cells producing BVMOs of bacterial origin.
Strain
Product
[
9b]
[b]
CHMOAcineto
CHMOArthro
CHMOBrachy
CHMOBrevi1
CHMOBrevi2
CPMOComa
CHMORhodo1
CHMORhodo2
53%
61%
59%
86% ee (À) >99% ee (À)
42% 35%
65%
54%
92% ee (+)
38%
96% ee (+)
51%
50%
89% ee (À)
66%
78%
>99% ee (À)
55%
33%
n.c.
[
c]
6
8
9
9
3
3
5
5
2% ee (À)
98% ee (À)
5% ee (À)
n.a.
[
b]
54%
50%
46%
n.c.
n.a.
[
c]
7% ee (À)
45%
>99% ee (À) 92% ee (À) >99% ee (À)
82% ee (À)
>99% ee (À)
60% ee (À)
56%
[
b]
69%
48%
40%
99% ee À)
61%
71%
45%
n.c.
n.a.
[
c]
3% ee (À)
>99% ee (À) 97% ee (À)
94% ee (+)
70%
91% ee (À)
21%
>99% ee (À)
55%
85% ee (À)
[
b]
[b]
73%
65%
>99% ee (À)
59%
n.c.
10%
n.c.
[
c]
[c]
8% ee (À)
50%
9% ee (+)
66%
7% ee (+)
58%
n.a.
97% ee (À)
>99% ee (+) 65% ee (À)
95% ee (À)
71% ee (À)
92%
94% ee (+)
76%
n.a.
19%
37%
61% ee (+)
54%
56%
44%
>99% ee (À)
63%
42%
0% ee
89%
9% ee (+)
62%
59%
44% ee (+)
99% ee (+)
58%
60% ee (+)
92%
48% ee (+)
51%
93% ee (+)
53%
[
13a]
68%
46% ee (+)
76% ee (+)
59%
91% ee (+)
59%
99% ee (À)
>99% ee (+) 95% ee (+)
[
b]
72%
75%
47%
n.c.
[
c]
2% ee (À)
>99% ee (À) 94% ee (À)
96% ee (À)
96% ee (+)
60%
96% ee (+)
85% ee (À)
98% ee (À)
73% ee (À)
n.a.
[
b]
63%
67%
47%
64%
75%
71%
51%
n.c.
n.a.
[
c]
0% ee (À)
95% ee (À)
94% ee (À)
90% ee (À)
75% ee (À)
95% ee (À)
73% ee (À)
[a] Yields are given for products isolated after flash column chromatography; ee values were determined by chiral-phase gas chromatography; the sign of specific
rotation is given. [b] n.c.=no conversion. [c] n.a.=not applicable.
resolution that led to the formation of regioisomeric lactones
and 5 in approximately 1:1 ratio and with high optical
4
purities. By contrast, CPMOComa and CHMOBrevi2 (CPMO-
type enzymes) yielded predominantly “normal” lactone 4 in
nearly racemic form. Trace amounts of “abnormal” product 5
were obtained in good enantiomeric excess.
Scheme 2. Regiodivergent Baeyer–Villiger oxidation of fused ketone 3
to “normal lactone” 4 and “abnormal lactone” 5.
To rationalize this significantly different biocatalytic
activity, we compared the results from the biooxidations
with sequence-analysis data for the genes and proteins of all
eight BVMOs. While only minor correlation was observed on
the DNA level of the structural genes, a significant trend in
enzyme similarity was found at the amino acid level.
Phylogenetic tree analysis of the representatives of the
BVMO enzyme family vis-ꢁ-vis a remote reference sequence
also resulted in clustering into two groups, which to a high
degree reflected the reaction profiles of the biocatalysts
(Figure 1). CPMOComa and CHMOBrevi2 form a distinctly
different cluster while the other six enzymes form the
[
24]
has been studied in detail with CHMOAcineto
.
Racemic
compound 3 is transformed into two types of regioisomeric
lactones in a resolution process: migration of the more-
substituted carbon atom generates the expected “normal”
lactone 4, while “abnormal” lactone 5 is formed by migration
of the less-substituted carbon atom.
Again, we observed a divergent trend for the two enzyme
[
25]
groups (Table 2). CHMO-type proteins displayed a clean
Table 2: Regiodivergent Baeyer–Villiger oxidation of racemic fused ketone 3 by recombinant E. coli cells producing BVMOs of bacterial origin.
[
a]
[b]
Yield 4+5 [%]
Ratio 4:5
ee (1S,5R)-4 [%]
ee (1R,5S)-5 [%]
CHMOAcineto
CHMOArthro
CHMOBrachy
CHMOBrevi1
CHMOBrevi2
CPMOComa
CHMORhodo1
CHMORhodo2
74
86
73
85
61
61
83
81
51:49
53:47
50:50
51:49
98:2
97:3
50:50
50:50
95
88
94
96
0
0
99
97
>99
>99
>99
>99
99
>99
>99
>99
[a] Combined yields are given for the mixture of 4 and 5 after single flash column chromatography. [b] Ratio and ee values were determined by chiral-
phase gas chromatography.
Angew. Chem. Int. Ed. 2005, 44, 3609 –3613
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3611

Communications
currently being addressed in our laboratory—seem necessary
for the development of a comprehensive and predictive
model for this enzyme family to successfully expand the
biocatalytic armament in the field of Baeyer–Villiger oxida-
tions.
Received: December 16, 2004
Revised: March 1, 2005
Published online: April 29, 2005
Keywords: biocatalysis · bioorganic chemistry ·
.
monooxygenases · oxidation · phylogenetic analysis
Figure 1. Phylogenetic tree of BVMOs originating from Acinetobacter,
Arthrobacter, Brachymonas, Brevibacterium, Comamonas, and Rhodococ-
4
cus species with the N -diaminopropane monooxygenase from Sinorhi-
[
[
1] A. Baeyer, V. Villiger, Chem. Ber. 1899, 32, 3625 – 3633.
2] G. R. Krow, Org. React. 1993, 43, 251 – 798; M. Renz, B.
Meunier, Eur. J. Org. Chem. 1999, 737 – 750.
zobium meliloti (DNMOSino) as the outgroup (1000 bootstraps).
[
3] R. Noyori, T. Sato, H. Kobayashi, Bull. Chem. Soc. Jpn. 1983, 56,
2661 – 2679.
“
CHMO-type” group. CHMOBrevi1 is located at the borderline
[4] K. Mislow, J. Brenner, J. Am. Chem. Soc. 1953, 75, 2318 – 2322.
[5] M. D. Mihovilovic, F. Rudroff, B. Grꢂtzl, Curr. Org. Chem. 2004,
of the two clusters closer to the CHMO group, a fact which
again agrees with its stereopreference and slightly modified
substrate-acceptance profile.
A related phylogenetic tree analysis with biomolecular
interpretation has been reported previously for a large
general set of monooxygenases.
our results to be the first connection of primary protein
sequence with biocatalyst performance for BVMOs.
8, 1057 – 1069.
[
[
[
6] C. Bolm in Peroxide Chemistry (Ed.: W. Adam), Wiley-VCH,
Weinheim, 2000, pp. 494 – 510; G. Strukul, Angew. Chem. 1998,
110, 1256 – 1267; Angew. Chem. Int. Ed. 1998, 37, 1198 – 1209.
7] Z. Li, J. B. van Beilen, W. A. Duetz, A. Schmid, A. de Raadt, H.
Griengl, B. Witholt, Curr. Opin. Chem. Biol. 2002, 6, 136 – 144;
S. G. Burton, Trends Biotechnol. 2003, 21, 543 – 549.
[
26]
However, we consider
8] N. M. Kamerbeek, D. B. Janssen, W. J. H. van Berkel, M. W.
Fraaije, Adv. Synth. Catal. 2003, 345, 667 – 678; M. D. Mihovi-
lovic, B. Mꢃller, P. Stanetty, Eur. J. Org. Chem. 2002, 3711 – 3730;
S. M. Roberts, P. W. H. Wan, J. Mol. Catal. B 1998, 4, 111 – 136.
9] Expression systems for cyclohexanone monooxygenase from
Acinetobacter (CHMOAcineto): a) G. Chen, M. M. Kayser, M. D.
Mihovilovic, M. E. Mrstik, C. A. Martinez, J. D. Stewart, New J.
Chem. 1999, 23, 827 – 832; b) M. D. Mihovilovic, G. Chen, S.
Wang, B. Kyte, R. Rochon, M. M. Kayser, J. D. Stewart, J. Org.
Chem. 2001, 66, 733 – 738; c) for an alternative Escherichia coli
based whole-cell biocatalyst, see: S. D. Doig, L. M. OꢀSullivan, S.
Patel, J. M. Ward, J. M. Woodley, Enzyme Microb. Technol. 2001,
When the alignment of protein sequences of BVMOs
included in this study is compared with the recently described
point mutations in CHMOAcineto
[
27]
,
two striking similarities
[
can be identified. The Leu143Phe mutation exactly follows
the separation in the two main enzyme groups. The CPMO
type in both cases has a phenylalanine, whereas the CHMO
type, with the exception of CHMOBrevi1, has a leucine residue
in this position, a fact again that reflects the borderline
position of the latter enzyme. Phenylalanine is conserved in
position 432 throughout the studied sequences with two
exceptions: CHMOBrevi1 and CHMOArthro. The Phe432Tyr
mutation exactly mimics the amino acid composition of
CHMOBrevi1 at this position. This mutation has been indicated
to significantly increase stereoselectivity and, interestingly, a
similar trend was observed in this study.
28, 265 – 274; d) for a preceeding Saccharomyces cerevisiae
expression system, see: J. D. Stewart, K. W. Reed, M. M.
Kayser, J. Chem. Soc. Perkin Trans. 1 1996, 755 – 757.
[10] For an isolated two-enzyme system, see: S. Rissom, U. Schwarz-
Linek, M. Vogel, V. I. Tishkov, U. Kragl, Tetrahedron: Asym-
metry 1997, 8, 2523 – 2526.
[
11] H. D. Simpson, V. Alphand, R. Furstoss, J. Mol. Catal. B 2001,
Recently, the first X-ray structure for a moderately related
1
1
6, 101 – 109; A. Z. Walton, J. D. Stewart, Biotechnol. Prog. 2002,
8, 262 – 268; S. D. Doig, P. J. Avenell, P. A. Bird, P. Gallati, K. S.
[
28]
BVMO from an extremophilic microorganism was solved.
This work gave valuable suggestions for the molecular
mechanism of the enzymatic oxidation. However, the
enzyme structure was determined in the absence of
NADPH and no cocrystallization with a substrate is available,
so far. As the authors suggest that the protein undergoes
extensive conformational changes in the biocatalytic cycle,
further conclusions for distant members of the BVMO family,
such as those included in this study, seem rather speculative.
However, when we consider the information together with
recent results in modifying the enantioselectivity of CH-
MOAcineto by random mutagenesis, we can begin to identify
BVMO regions with major impact on biocatalytic behavior
and stereopreference. Further structural and biotransforma-
tion studies on BVMOs more closely related to the two
clusters outlined herein and on a larger set of ketones—
Lander, G. J. Lye, T. Wohlgemuth, J. M. Woodley, Biotechnol.
Prog. 2002, 18, 1039 – 1046.
12] I. Hilker, V. Alphand, R. Wohlgemuth, R. Furstoss, Adv. Synth.
Catal. 2004, 346, 203 – 214.
13] a) H. Iwaki, Y. Hasegawa, P. C. K. Lau, S. Wang, M. M. Kayser,
Appl. Environ. Microbiol. 2002, 68, 5681 – 5684; b) M. D.
Mihovilovic, B. Mꢃller, A. Schulze, P. Stanetty, M. M. Kayser,
Eur. J. Org. Chem. 2003, 2243 – 2249; c) M. D. Mihovilovic, F.
Rudroff, B. Mꢃller, P. Stanetty, Bioorg. Med. Chem. Lett. 2003,
13, 1479 – 1482; d) M. D. Mihovilovic, F. Rudroff, B. Grꢂtzl, P.
Stanetty, Eur. J. Org. Chem. 2005, 809 – 816.
[
[
[
27]
[14] N. A. Donoghue, D. B. Norris, P. W. Trudgill, Eur. J. Biochem.
976, 63, 175 – 192.
1
[
15] P. Brzostowicz, D. M. Walters, S. M. Thomas, V. Nagarajan, P. E.
Rouviere, Appl. Environ. Microbiol. 2003, 69, 334 – 342.
16] M. G. Bramucci, P. C. Brzostowicz, K. N. Kostichka, V. Nagar-
ajan, P. E. Rouviere, S. M. Thomas (E. I. DuPont de Nemours &
[
3
612
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 3609 –3613

Angewandte
Chemie
Co., USA), WO2003020890, 2003 [Chem. Abstr. 2003, 138,
33997].
2
[
17] P. C. Brzostowicz, K. L. Gibson, S. M. Thomas, M. S. Blasko,
P. E. Rouviere, J. Bacteriol. 2000, 182, 4241 – 4248.
[
[
18] M. Griffin, P. W. Trudgill, Eur. J. Biochem. 1976, 63, 199 – 209.
19] M. Taschner, D. J. Black, J. Am. Chem. Soc. 1988, 110, 6892 –
6
893.
20] B. G. Kyte, P. Rouviere, Q. Cheng, J. D. Stewart, J. Org. Chem.
004, 69, 12 – 17.
21] R. Riva, L. Banfi, B. Danieli, G. Guanti, G. Lesma, G. Palmisano,
J. Chem. Soc. Chem. Commun. 1987, 299 – 300.
[
[
[
2
22] B. Danieli, G. Lesma, M. Mauro, G. Palmisano, D. Passerella,
Tetrahedron 1994, 50, 8837 – 8852.
[
[
23] V. Alphand, R. Furstoss, J. Org. Chem. 1992, 57, 1306 – 1309.
24] D. R. Kelly, C. J. Knowles, J. G. Mahdi, I. N. Taylor, M. A.
Wright, J. Chem. Soc. Chem. Commun. 1995, 729 – 730.
25] M. D. Mihovilovic, P. Kapitan, Tetrahedron Lett. 2004, 45, 2751 –
[
[
[
2754.
26] M. W. Fraaije, N. M. Kamerbeek, W. J. H van Berkel, D. B.
Janssen, FEBS Lett. 2002, 518, 43 – 47.
27] M. T. Reetz, B. Brunner, T. Schneider, F. Schulz, C. M.
Clouthier, M. M. Kayser, Angew. Chem. 2004, 116, 4167 – 4170;
Angew. Chem. Int. Ed. 2004, 43, 4078 – 4081.
[
28] E. Malito, A. Alfieri, M. W. Fraaije, A. Mattevi, Proc. Natl.
Acad. Sci. USA 2004, 101, 13157 – 13162.
Angew. Chem. Int. Ed. 2005, 44, 3609 –3613
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3613