Evolved CYP102A1 (P450BM3) variants oxidise a range of non-natural substrates and offer new selectivity options

Article abstract of DOI:10.1039/b718124h

The evolution of CYP102A1 variants with enhanced activity and altered specificity characteristics. The Royal Society of Chemistry.

Full text of DOI:10.1039/b718124h

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Evolved CYP102A1 (P450BM3) variants oxidise a range of non-natural
substrates and offer new selectivity optionsw
Christopher J. C. Whitehouse, Stephen G. Bell, Henry G. Tufton,
Richard J. P. Kenny, Lydia C. I. Ogilvie and Luet-Lok Wong*
Received (in Cambridge, UK) 23rd November 2007, Accepted 19th December 2007
First published as an Advance Article on the web 11th January 2008
DOI: 10.1039/b718124h
The evolution of CYP102A1 variants with enhanced activity and
altered specificity characteristics.
(KT5). All showed good stability characteristics, with little
or no activity loss after fifteen months’ storage at À20 1C.
Naphthalene and propylbenzene were again used to assess the
purified enzymes in vitro. WT, the F87A mutant and the highly
active A74G/F87V/L188Q (GVQ) variant developed by
Cytochrome P450 (CYP) monooxygenases have a number of
physiological roles including the biosynthesis of endogenous
compounds and secondary metabolites, and the metabolism of
1
4
Schmid and co-workers were also tested as benchmarks.
1
,2
drugs and xenobiotics. They typically catalyse the insertion
of an oxygen atom from atmospheric dioxygen into the C–H
bonds of a variety of organic compound types, with potential
applications that include the synthesis of fine chemicals and
All four new variants enhanced the product formation rate
6
(PFR) of the wild-type enzyme with naphthalene (3.1 nmol
À1
À1
À1
min (nmol P450) —henceforth abbreviated to min ) by at
least an order of magnitude, A330P being the most effective at
À1
the preparation of drug metabolites. CYP102A1 (P450BM3
)
155 min (Table 1). In all cases 1-naphthol was the only GC-
3
from Bacillus megaterium, a fatty acid hydroxylase, is one of
detectible product. WT was more active towards propylben-
1
zene at 615 min , with 71% coupling of NADPH con-
5
À1
the more extensively researched and engineered members of
4
the CYP superfamily. The wild-type enzyme (WT) turns over
sumption to product formation, but the new variants again
À1
most substrates other than fatty acids at desultory rates. Site-
specific mutagenesis and directed evolution have been em-
ployed, both separately and in tandem, to improve activity
gave higher PFRs, KT2 being the fastest at 2205 min
(Table 1). WT and KT2 yielded B99% 1-phenyl-1-propanol,
but the other new variants directed oxidation away from the
benzylic position. A330P gave 30% o-propylphenol, while
variant KT5, in which mutations F87A and A330P occur in
combination, gave 78% 1-phenyl-2-propanol, leveraging the
selectivity bias of mutation F87A, which formed 53%
1-phenyl-2-propanol when acting alone (Fig. 2). Changes in
haem spin state on substrate addition were determined for
these and other substrates but generally correlated poorly with
NADPH rates (Tables 1, S3 and S5w).
5
,6
7,8
9
towards aromatics, alkanes and also pharmaceuticals, a
1
0
number of which are accepted as substrates by WT. The
challenge of selectivity control has also been addressed, and
compounds not produced by WT have been successfully
1
1
targeted. Various screening procedures have been developed
1
for directed evolution. Indigo formation via indole oxida-
2
1
3
tion, one of the simplest, was first used in CYP102A1 to
1
4
monitor saturation mutagenesis at specific sites. By applying
it as a preliminary activity screen and then sub-screening in
vivo for selectivity as well as activity, we have been able to
identify new variants with much enhanced activity (Fig. 1),
two of which also have the ability to alter product profiles.
Error-prone PCR was carried out on the full haem domain-
coding regions of the wild-type CYP102A1 gene and mutant
F87A. Indigo formation provided a simple assay for identify-
ing variants with improved performance potential. These were
subjected to gene shuffling and a second round of indigo
screening. Those that gave rise to blue colonies were assayed
in vivo for naphthalene and propylbenzene oxidation activity.
Product selectivity was also monitored. Four of the most
promising were purified: A330P, a single mutation variant;
A191T/N239H/I259V/A276T/L353I (KT2); F87A/H171L/
Q307H/N319Y (KSK19); and F87A/A330P/E377A/D425N
The R47L/Y51F (RLYF) couplet, which is known to
5
increase CYP102A1 activity towards several substrates,
,16
was then incorporated into variants A330P and KT2. In both
cases NADPH consumption rates were further enhanced.
À1
With naphthalene, RLYF/A330P gave a PFR of 666 min
,
the highest observed, while RLYF/KT2 gave the highest rate
Department of Chemistry, University of Oxford, Inorganic Chemistry
Laboratory, South Parks Road, Oxford, UK OX1 3QR. E-mail:
luet.wong@chem.ox.ac.uk; Fax: +44 1865 272690
w Electronic supplementary information (ESI) available: Experimental
details and additional data tables. See DOI: 10.1039/b718124h
Fig. 1 In vitro activity of CYP102A1 variants with various substrates.
This journal is ꢀc The Royal Society of Chemistry 2008
9
66 | Chem. Commun., 2008, 966–968

View Article Online
a
Table 1 In vitro activity, selectivity and spin shifts of CYP102A1 and variants with naphthalene and propylbenzene
b
Naphthalene
Propylbenzene
Spin shift NADPH
rate
Spin shift NADPH
(%)
1-ol 2-ol
(%) (%)
Phenol
(%)
(
%)
PFR
C (%)
rate
PFR C (%)
WT and generic accelerators
WT
R47L/Y51F (RLYF)
KT2
RLYF/KT2
Phe87 variants
F87A
F87A/KT2
KSK19
KT5
10
5
20
15
80
166
490
3.1
32
113
496
3.9
19
23
38
10
5
25
15
866
2157
2756
3126
615 71
1682 78
2205 80
2688 86
99
99
98
99
—
1
1
—
1
—
1
1306
—
10
10
20
35
20
106
370
511
567
5.5
32
56
31
5.2
8.7
11
5.5
23
20
30
40
40
30
670
1664
2079
1062
3172
248 37
566 34
1081 52
701 66
1015 32
46
54
53
20
77
53
45
46
78
14
—
—
—
—
8
GVQ
2116
487
A330P variants
A330P
RLYF/A330P
o5
483
1306
155
666
32
51
o5
1810
2497
706 39
674 27
68
65
2
5
30
30
c
5
5
a
À1
À1
Rates in nmol min (nmol P450) . C = coupling (total product formation as % of NADPH consumption). Percentages do not always sum to
00 on account of minority products. 1-Naphthol was the only GC-detectible product formed in naphthalene turnovers. Means of at least three
1
b
c
turnovers to within Æ10%. Products were 1-phenyl-1-propanol (1-ol), 1-phenyl-2-propanol (2-ol) and o-propylphenol (Phenol). Includes 7%
p-propylphenol.
À1
with propylbenzene at 2688 min . This approaches the rate
À1 17
(98%) (Table S4, Fig. S1aw). Variant RLYF/A330P also gave
98% o-cresol, but at a PFR of 189 min versus 2.7 min for
À1
À1
reported for WT with fatty acid substrates (3200 min ).
Product profiles were little altered relative to the first genera-
tion variants, though RLYF/A330P gave some p-propylphe-
nol. KT2 was also combined with mutation F87A to give a
second generation variant capable of offering F87A-directed
product profiles at higher rates (Table 1).
WT, coupling efficiency rising from 9.4% to 52%.
Although the new variants were selected on the basis of
their affinity for aromatic substrates, KT2 and A330P also
displayed high activity levels with pentane, particularly in
combination with RLYF. RLYF/KT2 and RLYF/A330P
À1
À1
The range of regioselective outcomes obtained in propyl-
benzene turnovers led us to investigate other alkylbenzenes.
The naturally occurring hydrocarbon, p-cymene is one of the
more interesting as it is a precursor to four distinct flavouring
gave NADPH rates of 2010 min and 1766 min , respec-
À1
tively, with 60% and 67% coupling versus 74 min and 21%
for WT (Table S5w). The GVQ variant also gave a rapid
NADPH rate (2107 min ), though with coupling of just 28%.
À1
À1
À1
compounds. WT (PFR 168 min ) gave predominantly p,a,a-
An NADPH turnover rate of 3250 min with pentane has
trimethylbenzyl alcohol, along with small quantities of the two
possible aromatic hydroxylation products, thymol (3%) and
carvacrol (7%), and just 2% 4-isopropylbenzyl alcohol (Table
S3w). Variant A330P, by contrast, yielded 19% thymol, 16%
carvacrol and 19% 4-isopropylbenzyl alcohol. The new var-
been reported by Arnold and co-workers for variant 139-3,
1
8
though no PFR data are available. A330P variants did not
show altered pentane product profiles relative to WT, but
selectivity shifts were observed with octane, A330P giving 53%
2-octanol versus 15% for WT, which yielded primarily 3-
and 4-octanol (Table S6, Fig. S1bw). An identical level of
2-octanol formation (53%) was reported for variant 9-10A, a
derivative of 139-3, rising to 89% when three additional
mutations (V78F, A82S and A328F) were designed into
iants also enhanced activity, with KSK19 giving a PFR of
À1
1
442 min (Fig. 1). A more dramatic selectivity shift was
observed with toluene, KT5 yielding 95% benzyl alcohol and
just 5% o-cresol where WT gave almost exclusively o-cresol
7
,19
the active site.
KT2 thus increased NADPH consumption rates and cou-
pling relative to WT across a range of non-natural substrates
while leaving product profiles almost unaltered. F87A/KT2
and KSK19 displayed the same pattern of behaviour relative
to base mutation F87A, increasing PFRs while giving F87A-
like product distributions. The mutations comprising KT2 and
F87A-free KSK19 may therefore be classified as generic rate
accelerators for the substrates studied. KSK19 is generally
more active than F87A/KT2, implying that its F87A-free
derivative may prove more potent than KT2. The mutated
residues in KT2 and KSK19 are all outside the active site
2
0
(
Fig. 3), though Leu353 lies next to a substrate access
channel residue, Met354, while Ala191 is close to the channel
2
entrance. KT2 mutations N239H and I259V have been
1
2
2,23
Fig. 2 Gas chromatography analysis of propylbenzene oxidation.
This journal is ꢀc The Royal Society of Chemistry 2008
reported previously.
Chem. Commun., 2008, 966–968 | 967

View Article Online
This work was supported by the EPSRC and BBSRC, UK
Grant ref. EP-D048559-1).
(
Notes and references
1
Cytochrome P450: Structure, Mechanism and Biochemistry, ed. P.
R. Ortiz de Montellano, Kluwer Academic/Plenum Publishers,
New York, 3rd edn, 2005.
2
The Ubiquitous Roles of Cytochrome P450 Proteins, ed. A. Sigel, H.
Sigel and R. K. O. Sigel, John Wiley & Sons, Chichester, Metal
Ions in Life Sciences, 2007, vol. 3.
3
4
M. J. Cryle, R. D. Espinoza, S. J. Smith, N. J. Matovic and J. J. De
Voss, Chem. Commun., 2006, 2353–2355.
S. G. Bell, N. Hoskins, C. J. C. Whitehouse and L. L. Wong, in The
Ubiquitous Roles of Cytochrome P450 Proteins, ed. A. Sigel, H.
Sigel and R. K. O. Sigel, John Wiley & Sons, Chichester, Metal
Ions in Life Sciences, 2007, vol. 3, ch. 14, pp. 437–476.
A. B. Carmichael and L. L. Wong, Eur. J. Biochem., 2001, 268,
5
6
7
8
9
3
117–3125.
Q. S. Li, J. Ogawa, R. D. Schmid and S. Shimizu, Appl. Environ.
Microbiol., 2001, 67, 5735–5739.
M. W. Peters, P. Meinhold, A. Glieder and F. H. Arnold, J. Am.
Chem. Soc., 2003, 125, 13442–13450.
R. Fasan, M. M. Chen, N. C. Crook and F. H. Arnold, Angew.
Chem., Int. Ed., 2007, 46, 8414–8418.
B. M. Vugt-Lussenburg, E. Stjernschantz, J. Lastdrager, C. Oos-
tenbrink, N. P. Vermeulen and J. N. Commandeur, J. Med. Chem.,
2007, 50, 455–461.
20
Fig. 3 The locations of the mutation sites.
Variant A330P displayed more complex behaviour, always
enhancing NADPH rates, but coupling significantly better
than WT with certain substrates and less well with others.
Though not part of the active site, this mutation frequently
altered selectivity patterns, generally in a different sense to
F87A. When in combination with F87A, however (as in KT5),
rather than giving A330P-like product profiles it gave F87A-
like profiles with selectivity shifts that exceeded those given by
other F87A variants. The crystal structure of the A330P haem
domain may help to elucidate the structural consequences of
introducing a proline residue directly next to Pro329, a sub-
strate access channel residue. It will be interesting to explore
the effect of alternative substitutions at Ala330, and to estab-
lish how A330P combines with selectivity-directing mutations
1
0 G. Di Nardo, A. Fantuzzi, A. Sideri, P. Panicco, C. Sassone, C.
Giunta and G. Gilardi, J. Biol. Inorg. Chem., 2007, 12, 313–323.
1 V. B. Urlacher, S. G. Bell and L. L. Wong, in Modern Biooxidation,
ed. R. D. Schmid and V. B. Urlacher, Wiley-VCH, Weinheim,
1
2
007, ch. 4, pp. 99–122.
1
1
2 K. L. Tee and U. Schwaneberg, Comb. Chem. High Throughput
Screening, 2007, 10, 197–217.
3 E. M. Gillam, A. M. Aguinaldo, L. M. Notley, D. Kim, R. G.
Mundkowski, A. A. Volkov, F. H. Arnold, P. Soucek, J. J. DeVoss
and F. P. Guengerich, Biochem. Biophys. Res. Commun., 1999, 265,
4
69–472.
1
1
1
4 Q. S. Li, U. Schwaneberg, P. Fischer and R. D. Schmid, Chemistry,
2000, 6, 1531–1536.
5 Q. S. Li, J. Ogawa, R. D. Schmid and S. Shimizu, FEBS Lett.,
5
,7,24
such as A264G, I263A, A82L and A328V.
2
001, 508, 249–252.
In summary, our screening protocol located CYP102A1
variants with significantly enhanced activity levels across an
array of substrates. Typically, they improved coupling
efficiencies and also combined well with site-specific muta-
tions, attributes not shared by all activity-enhancing muta-
6 V. B. Urlacher, A. Makhsumkhanov and R. D. Schmid, Appl.
Microbiol. Biotechnol., 2006, 70, 53–59.
17 J. H. Capdevila, S. Wei, C. Helvig, J. R. Falck, Y. Belosludtsev, G.
Truan, S. E. Graham-Lorence and J. A. Peterson, J. Biol. Chem.,
1
996, 271, 22663–22671.
1
8 A. Glieder, E. T. Farinas and F. H. Arnold, Nat. Biotechnol., 2002,
20, 1135–1139.
19 P. Meinhold, M. W. Peters, M. M. Chen, K. Takahashi and F. H.
4
tions. The amino acid substitutions involved showed little
overlap with those discovered using other methods, suggesting
that directed evolution, as applied to CYP102A1, has yet to
exhaust its potential. The new variants should make good
starting-points for further protein evolution studies. A330P,
with its striking selectivity characteristics and enhanced pro-
duct formation rates, could be particularly useful in the pur-
suit of altered product profiles, e.g. in the development of
Arnold, ChemBioChem, 2005, 6, 1765–1768.
2
2
0 H. Li and T. L. Poulos, Nat. Struct. Biol., 1997, 4, 140–146.
1 M. D. Paulsen and R. L. Ornstein, Proteins: Struct. Funct. Genet.,
1995, 21, 237–243.
2
2 P. C. Cirino and F. H. Arnold, Angew. Chem., Int. Ed., 2003, 42,
3
299–3301.
2
2
3 US Pat., 2003100744, 2003. World Pat., 2003008563, 2003.
4 R. J. Sowden, S. Yasmin, N. H. Rees, S. G. Bell and L. L. Wong,
Org. Biomol. Chem., 2005, 3, 57–64.
2
5
26
phenol-forming variants or terminal alkane oxidases. The
unusual selectivity patterns displayed by KT5 suggest that the
F87A/A330P couplet could also make a promising evolution-
ary platform.
2
5 T. S. Wong, N. Wu, D. Roccatano, M. Zacharias and U. Schwa-
neberg, J. Biomol. Screening, 2005, 10, 246–252.
6 O. Lentz, A. Feenstra, T. Habicher, B. Hauer, R. D. Schmid and
V. B. Urlacher, ChemBioChem, 2005, 7, 345–350.
2
9
68 | Chem. Commun., 2008, 966–968
This journal is ꢀc The Royal Society of Chemistry 2008