Screening of a minimal enriched P450 BM3 mutant library for hydroxylation of cyclic and acyclic alkanes

Article abstract of DOI:10.1039/c0cc02924f

A minimal enriched P450 BM3 library was screened for the ability to oxidize inert cyclic and acyclic alkanes. The F87A/A328V mutant was found to effectively hydroxylate cyclooctane, cyclodecane and cyclododecane. F87V/A328F with high activity towards cyclooctane hydroxylated acyclic n-octane to 2-(R)-octanol (46% ee) with high regioselectivity (92%).

Full text of DOI:10.1039/c0cc02924f

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Screening of a minimal enriched P450 BM3 mutant library
for hydroxylation of cyclic and acyclic alkanesw
Evelyne Weber,ab Alexander Seifert,a Mihaela Antonovici,a Christopher Geinitz,a
Ju
¨
rgen Pleissa and Vlada B. Urlacher*b
Received 30th July 2010, Accepted 27th October 2010
DOI: 10.1039/c0cc02924f
A minimal enriched P450 BM3 library was screened for the
ability to oxidize inert cyclic and acyclic alkanes. The F87A/
A328V mutant was found to effectively hydroxylate cyclooctane,
cyclodecane and cyclododecane. F87V/A328F with high activity
towards cyclooctane hydroxylated acyclic n-octane to 2-(R)-
octanol (46% ee) with high regioselectivity (92%).
the strain Rhodococcus ruber CD4 that is able to oxidize
cyclododecane and the corresponding monoketone with
subsequent ring cleavage like that during cyclohexane degra-
dation.7 Whereas numerous Baeyer–Villiger monooxygenases
have been isolated and characterized,10 almost nothing is
known about the corresponding cycloalkane hydroxylases.
Previously we reported a P450 BM3 mutant that hydroxylates
cyclohexane.9 Degradation of cyclic alkanes with a higher
C-atom number, like cyclodecane and cyclododecane are
far less investigated. Thus, there is still a demand in active
(and selective) alkane hydroxylases.
Effective oxidation of readily available cyclic and acyclic alkanes
provides compounds which find applications in manufacturing
soaps, detergents and solvents, as well as in the production of
certain raw materials. For example, cyclododecanol and
cyclododecanone are applied in the synthesis of polyamides
and macrocyclic lactones.1 The important chiral intermediates
such as (R)- or (S)-2-octanol are widely used as versatile
synthons and precursor compounds for the preparation of
pharmaceuticals, agrochemicals, pheromones and liquid
crystals.2 However, the controlled oxidation of the notoriously
inert alkanes is still a challenging task for modern chemistry and
biotechnology.3
In the present report a minimal enriched P450 BM3 mutant
library was tested with four substrates of different shape, size
and chemical reactivity: cyclooctane (C8), cyclodecane (C10),
cyclododecane (C12). This library was previously constructed
by combining five hydrophobic amino acids (alanine, valine,
phenylalanine, leucine and isoleucine) in two positions, 87 and
328, located directly above the heme-group (Fig. 1). The
library contains 24 variants, 11 of those demonstrated a
strong shift in regio- or stereoselectivity during oxidation of
different terpenes and terpenoids.11
Regarding oxidation of acyclic n-alkanes of different chain
lengths several heme- and non-heme iron monooxygenases
have been reported, which hydroxylate alkanes exclusively at
o-position.4 In the area of protein engineering numerous
mutants of CYP101A1 from Pseudomonas putida (P450cam)
and CYP102A1 from Bacillus megaterium (P450 BM3) have
been constructed for oxidation of n-alkanes ranging from
decane to ethane, yielding, however, mixtures of secondary
alcohols.5 The highest regioselectivity towards 2-octanol of
82–89% was reported by Arnold and colleagues for the P450
BM3 mutants with multiple (up to 17) mutations, which
produced either (R)-enantiomer (39% ee) or (S)-enantiomer
(65% ee), correspondingly.6
P450 BM3 is a self-sufficient fusion protein, consisting of a
heme domain and a diflavin reductase domain. For its activity
the enzyme requires only the pyridine cofactor NADPH. All
24 P450 BM3 mutants were actively expressed in E. coli,
purified and tested with target compounds. The wild type
enzyme converted only 8% of cyclooctane and 3% of
cyclodecane with low activity (Table 1). Cyclododecane was
not accepted by wild type P450 BM3 at all. The screening of the
mutant library revealed 20 out of 24 mutants which were able
to convert cyclooctane; 17 mutants accepted cyclodecane as
substrate and 8 mutants were able to oxidize cyclododecane.
With respect to cyclic alkanes several reports describe
degradation of cyclohexane either by single bacterial strains
of Nocardia, Pseudomonas, Acinetobacter and Xanthobacter or
by cyclohexane-degrading consortia.7 The pathway proposed
for the degradation of cycloalkanes includes the initial ring
hydroxylation by a monooxygenase, followed by formation of
the corresponding ketone by a dehydrogenase, and finally by
the insertion of one oxygen atom into the ring, catalyzed by a
Baeyer–Villiger monooxygenase.8 A previous study describes
a Institute of Technical Biochemistry, Universitaet of Stuttgart,
Allmandring 31, 70569 Stuttgart, Germany
Fig. 1 The substrate binding cavity of P450 BM3 F87A/A328V (a)
and P450 BM3 wild type (b) in complex with cyclododecane after 3 ns
of unrestraint MD simulation. The mutated positions are depicted in
red. Positions 87 (left) and 328 (right) stabilize the substrate in the
active site cavity. The activated oxygen of the heme is shown in orange.
b Institute of Biochemistry, Heinrich-Heine University Dusseldorf,
¨
Universitatsstr. 1, 40225 Dusseldorf, Germany.
¨
¨
E-mail: Vlada.Urlacher@uni-duesseldorf.de
w This article is part of the ‘Enzymes and Proteins’ web-theme issue for
ChemComm.
c
944 Chem. Commun., 2011, 47, 944–946
This journal is The Royal Society of Chemistry 2011

View Article Online
Table 1 Oxidation ratesa and conversionsb of three cyclic alkanes measured with selected P450 BM3 mutantsc
Cyclooctane
Cyclodecane
Cyclododecane
Ox. rates (eq. minÀ1
)
Conv (%)
Ox. rates (eq. minÀ1
)
Conv (%)
Ox. rates (eq. minÀ1
)
Conv (%)
Enzyme
WT
F87A
2
20
130
102
141
87
46
230
73
8
65
90
85
75
80
70
75
82
52
58
87
o1
13
23
28
78
20
67
35
16
20
13
106
3
57
60
70
50
60
60
14
23
19
23
53
—
4
—
20
6
14
—
46
6
—
—
—
3
F87A/A328F
F87A/A328I
F87A/A328L
F87A/A328V
F87V
F87V/A328F
F87V/A328I
F87V/A328L
F87V/A328V
A328V
1.5
3.6
—
18
2.4
—
—
—
1
64
34
200
—
—
a
Oxidation rates given in nmol product per nmol P450 per min (referred to as eq. minÀ1) were measured by GC/MS after 15 min. Reactions
b
contained 1 mM P450, 100 mM NADPH, 200 mM substrate in 2% DMSO and potassium phosphate buffer, pH 7.5. Conversions in % were
measured with 1 mM P450, 600 mM NADPH and 200 mM substrate after 1 h. All experiments were performed in triplicates on two independent
occasions. Errors are not higher than 8%.
c
In all the cases the single oxidation product was the
corresponding alcohol. The detailed GC-analysis revealed,
however, that in many cases conversion of cyclic substrates
was less than 20%. Only 6 mutants converted >70% of
cyclooctane to cyclooctanol; 5 mutants demonstrated >50%
conversion of cyclodecane to cyclodecanol, and at least 2 were
able to convert >20% cyclododecane. In some cases higher
initial oxidation activity did not lead to higher conversions,
which might be due to lower mutants’ stability.
substrate approached the heme in mutant F87A/A328V and
remained stable in close proximity to the heme centre with the
closest C–HÁ Á ÁO–Fe distance of 2.95 A. A fluctuating root
mean square deviation of the protein backbone between 1.0
and 1.2 A from the crystal structure further indicates that there
is no major impact by the substrate on the protein backbone.
The cyclododecane molecule is in close contact with the amino
acid side chains of the two hotspot positions (Fig. 1a).
Interestingly, the small side chain of alanine in the mutant
F87A/A328V creates a space close to the heme which is
occupied by the substrate. In the wild type enzyme this space
is occupied by a more bulky side chain of phenylalanine in this
position (Fig. 1b). As a consequence the activated oxygen is
less accessible for the substrate. This observation might explain
why the double mutant converts cyclododecane, while the wild
type does not.
The detailed analysis of all P450 BM3 mutants, which resulted
in more than 20% conversion with at least one of the substrates
(Table 1), has revealed that most of them contain alanine or
valine at position 87, independent on the substrate tested.
However, the mutants containing F87V resulted generally in
lower conversions of cyclodecane and cyclododecane compared
to cyclooctane. Only the F87A/A328V mutant was able to
oxidize the largest substrate cyclododecane with a relatively
high activity of 18 nmol product per nmol P450 per min and
resulted in 46% conversion. Interestingly, it was not sufficient to
solely increase the accessibility of the heme by replacing the
bulky residues in the hotspot positions by smaller amino acids to
maximize conversion of these large substrates as indicated by the
higher conversions by the mutant F87A/A328V compared to
the F87A. Obviously, the mutant with the better fitting shape of
the substrate binding cavity can stabilize the substrate more
effectively during catalysis.
A recent report on degradation of cyclododecane by the
R. ruber CD4 strain suggests that the larger the carbon ring of
the cyclic alkane, the more flexible it is and the easier it can take a
conformation like that of an open-chain alkane. This observation
was also made during docking of cyclododecane. Since fixation
of inert alkanes in the binding pocket of P450 BM3 occurs almost
exclusively via hydrophobic interactions, and as there is no
‘‘terminal’’ part of a ring, the reaction with acyclic alkanes can
be considered as ‘‘analogous’’ to a subterminal attack on an
acyclic substrate.7 Consequently, a preference for larger rings
therefore may indicate a good potential for attack at a certain
subterminal position on a linear aliphatic structure. Our previous
calculation of activation energy barriers for H-abstraction of
different positions of n-octane revealed a slightly lower activation
energy for o-1 position (corresponding to 2-octanol) compared
to all other positions.15 Based on both the hypotheses we tested
the same set of the P450 BM3 mutants with n-octane for
improved regioselectivity towards 2-octanol formation.
The relatively large size of cyclododecane prompted us to
study by molecular dynamics simulations how this substrate is
accommodated in the active site cavity of the most active
mutant F87A/A328V in comparison to the wild type enzyme.
The mutations F87A and A328V were introduced to the
structure of P450 BM3 (PDB entry 1bu7A) with the Pymol
0.99 program.12 The cyclododecane molecule was placed at a
distance of 6 A to the activated heme oxygen by hand. The
complex was simulated in explicit water using the Amber 8
molecular dynamics simulation program package.13 The
system was coupled to an external pressure and temperature
bath.14 The force field ff03 was used and the partial charges of
cyclododecane were taken from corresponding atom types in
amino acids. During the time course of the simulation the
The P450 BM3 wild type enzyme demonstrated very low
activity and regioselectivity upon n-octane hydroxylation.
After 30 min reaction substrate conversion achieved ca. 5%,
which correlates with the previously reported data.16 The ratio
of 2-octanol in the product mixture was 15%. Besides that
3-octanol (37%), 4-octanol (43%) as well as 5% ketones were
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 944–946 945

View Article Online
Table 2 Oxidation ratesa and conversionsb of n-alkane with selected
including A328V.6a Though being less regioselective than the
here reported F87V/A328F, that mutant demonstrated higher
activity. This confirmed our suggestion that both positions 87
and 328 in combination control enzyme regioselectivity, while all
other mutations influence mainly enzyme activity and stability.
In summary, several P450 BM3 mutants were identified
which enable oxidation of C8–C12 cycloalkanes. To our
knowledge this is the first example of active cycloalkane’s
monooxygenases. Furthermore, our results demonstrate the
functional flexibility of P450 BM3 mutants with minimal
number of mutations. Single and double P450 BM3 mutants
with substitutions in the active site were engineered, accepting
and hydroxylating inert alkanes to corresponding alcohols,
and in the case of acyclic alkanes—with high regio- and
moderate enantioselectivity.
P450 BM3 mutantsc
n-Octane
Conversion
(%)
Ratio of 2-octanol
(%)
Ox. rates
(eq. minÀ1
)
Enzyme
WT
F87A
F87A/
A328F
4
5.5
5.2
5
15
8
15
12
49
F87A/A328I
F87A/
A328L
7.2
18
13
10
60
25
F87A/
A328V
F87V
F87V/A328F
F87V/A328I
F87V/A328L
F87V/
9.1
10
7
9
22
11
15
12
5
15
13
14
12
20
92
51
85
47
This work was supported by grants from the German
Research Foundation (DFG, SFB 706), the Baden-
A328V
A328V
Wurttemberg Ministry of Science, Research and the Arts and
¨
the Fonds der Chemischen Industrie (Germany).
144
33
67
a
Oxidation rates given in nmol product per nmol P450 per min
(referred to as eq. minÀ1) were measured by GC/MS after 15 min.
Reactions contained 1 mM P450, 100 mM NADPH, 200 mM substrate
Notes and references
b
in 2% DMSO and potassium phosphate buffer, pH 7.5. Conversions
1 Evonik Degussa GmbH, High performance building blocks. 2009,
http://www.degussa/hpp.com/dl/brochure/buildingblocks_eng.pdf.
2 (a) A. Liese, T. Zelinski, M. R. Kula, H. Kierkels, M. Karutz,
U. Kragl and C. Wandrey, J. Mol. Catal. B: Enzym., 1998, 4,
91–99; (b) L. Xue, D. J. Zhou, L. Tang, X. F. Ji, M. Y. Huang and
Y. Y. Jiang, React. Funct. Polym., 2004, 58, 117–121.
3 A. E. Shilov and G. B. Shul’pin, Chem. Rev., 1997, 97, 2879–2932.
4 (a) C. L. Chen, K. H. Chen, S. C. Ke, S. S. Yu and S. I. Chan,
J. Inorg. Biochem., 2004, 98, 2125–2130; (b) D. L. Craft,
K. M. Madduri, M. Eshoo and C. R. Wilson, Appl. Environ.
Microbiol., 2003, 69, 5983–5991; (c) T. Fujii, T. Narikawa,
K. Takeda and J. Kato, Biosci., Biotechnol., Biochem., 2004, 68,
2171–2177; (d) E. G. Funhoff, U. Bauer, I. Garcia-Rubio,
B. Witholt and J. B. van Beilen, J. Bacteriol., 2006, 188, 5220–5227.
5 (a) R. Fasan, M. M. Chen, N. C. Crook and F. H. Arnold, Angew.
Chem., Int. Ed., 2007, 46, 8414–8418; (b) F. Xu, S. G. Bell,
J. Lednik, A. Insley, Z. Rao and L. L. Wong, Angew. Chem., Int.
Ed., 2005, 44, 4029–4032.
in % were measured with 1 mM P450, 600 mM NADPH and 200 mM
substrate after 30 min. All experiments were performed in triplicates
on two independent occasions. Errors are not higher than 10%.
c
formed. Eight out of 24 mutants from the minimal library
produced 2-octanol in a ratio of >25%. The strongest shift in
regioselectivity towards 2-octanol (product ratio of >50%)
was observed with mutants, containing a bulky amino acid at
one of the two hotspots, if the second one was represented by a
smaller alanine or valine residue. When both positions 87 and
328 were occupied by valine and/or alanine, the ratio of
2-octanol dropped to 7% (Table 2).
The best mutants regarding 2-octanol production were F87V/
A328L (85%), F87/A328V (single mutant, 67%) and F87V/
A328F (92%). Probably, the flexible n-octane molecule should
be ‘‘fixed’’ more tightly, for example by the bulky phenylalanine
or isoleucine either at position 87 or 328, in order to be oxidized
preferably at one position. The similar correlation between
structural changes in the substrate binding pocket and enzyme
regioselectivity upon n-octane oxidation was observed pre-
viously for the homologous CYP102A3 from B. subtilis.17
The same mutants, F87V/A328L, F87/A328V and F87V/
A328F demonstrated high conversions of cyclooctane, signifi-
cant conversions of cyclodecane, but low or no conversion of
cyclododecane. These results suggest that high activity towards
a cyclic alkane may indeed indicate a potential specificity for
a subterminal attack on a corresponding acyclic structure.
However, no direct correlation can be built without additional
experiments on other cyclic and acyclic substrates. Moreover, in
all the cases the more flexible n-octane was converted much
slower than the cyclic substrates (Table 2).
6 (a) M. W. Peters, P. Meinhold, A. Glieder and F. H. Arnold, J. Am.
Chem. Soc., 2003, 125, 13442–13450; (b) P. Meinhold,
M. W. Peters, M. M. Chen, K. Takahashi and F. H. Arnold,
ChemBioChem, 2005, 6, 1765–1768.
7 J. D. Schumacher and R. M. Fakoussa, Appl. Microbiol.
Biotechnol., 1999, 52, 85–90.
8 E. H. Lee and K. S. Cho, Chemosphere, 2008, 71, 1738–1744.
9 S. C. Maurer, K. Kuhnel, L. A. Kaysser, S. Eiben, R. D. Schmid
and V. B. Urlacher, Adv. Synth. Catal., 2005, 347, 1090–1098.
10 M. Kataoka, K. Honda, K. Sakamoto and S. Shimizu, Appl.
Microbiol. Biotechnol., 2007, 75, 257–266.
11 A. Seifert, S. Vomund, K. Grohmann, S. Kriening, V. B. Urlacher,
S. Laschat and J. Pleiss, ChemBioChem, 2009, 10, 853–861.
12 W. L. DeLano, The PyMOL Molecular Graphics System, DeLano
Scientific, San Carlos, CA, USA, 2002, http://www.pymol.org.
13 D. A. Case, T. A. Darden, T. E. Cheatham III, C. L. Simmerling,
J. Wang, R. E. Duke, R. Luo, K. M. Merz, B. Wang,
D. A. Pearlman, M. Crowley, S. Brozell, V. Tsui, H. Gohlke,
J. Mongan, V. Hornak, G. Cui, P. Beroza, C. Schafmeister,
J. W. Caldwell, W. S. Ross and P. A. Kollman, AMBER 8,
University of California, San Francisco, CA, 2004.
14 H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren,
A. Dinola and J. R. Haak, J. Chem. Phys., 1984, 81, 3684–3690.
15 K. A. Feenstra, E. B. Starikov, V. B. Urlacher, J. N. Commandeur
and N. P. Vermeulen, Protein Sci., 2007, 16, 420–431.
16 D. Appel, S. Lutz-Wahl, P. Fischer, U. Schwaneberg and
R. D. Schmid, J. Biotechnol., 2001, 88, 167–171.
Investigation of the enzyme enantioselectivity demonstrated
that the most regioselective mutant F87V/A328F produced up
to 92% of (R)-2-octanol with 46% ee. In comparison, the wild
type enzyme produced 15% (R)-2-octanol with 20% ee. The
P450 BM3 mutant that produced up to 82% of (R)-2-octanol
(39% ee) as reported by Peters et al., contained 13 mutations,
17 O. Lentz, A. Feenstra, T. Habicher, B. Hauer, R. D. Schmid and
V. B. Urlacher, ChemBioChem, 2006, 7, 345–350.
c
946 Chem. Commun., 2011, 47, 944–946
This journal is The Royal Society of Chemistry 2011