Directed evolution of the Actinomycete Cytochrome P450 MoxA (CYP105) for enhanced activity

Article abstract of DOI:10.1271/bbb.90013

Actinomycete cytochrome P450 from Nonomuraea recticatena NBRC 14525 (P450moxA) catalyzes the hydroxylation of a broad range of substrates, including fatty acids, steroids, and various aromatic compounds. Hence, the enzyme is potentially useful in medicinal applications, but the activity is insufficient for practical use. Here we applied directed evolution to enhance the activity. A random mutagenesis library was screened using 7-ethoxycoumarin as a substrate to retrieve 17 variants showing <2-fold activities. Twenty-five amino acid substitutions were found in the variants, of which five mutations were identified to have the largest effects (Q87W, T115A, H132L, R191W, and G294D). These mutations additively increased the activity; the quintet mutant had 20-times the activity of the wildtype. These five single mutations also increased in activity toward structurally distinct substrates (diclofenac and naringe- nin). Based on the three-dimensional structure of the enzyme, we discerned that mutations in the substrate recognition site improved the activity, which was substrate dependent; mutations apart from the active site improved the activity as well as the substrates did.

Full text of DOI:10.1271/bbb.90013

Biosci. Biotechnol. Biochem., 73 (9), 1922–1927, 2009
Directed Evolution of the Actinomycete Cytochrome P450 MoxA (CYP105)
for Enhanced Activity
y
y
1;
Hiroki KABUMOTO,1 Kentaro MIYAZAKI,2;3; and Akira ARISAWA
1Bioresource Laboratories, Mercian Corporation, 1808 Nakaizumi, Iwata, Shizuoka 438-0078, Japan
2Institute for Biological Resources and Functions, National Institute of Advanced Industrial Science
and Technology (AIST), Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8566, Japan
3Department of Medical Genome Sciences, Graduate School of Frontier Sciences, University of Tokyo,
Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8566, Japan
Received January 7, 2009; Accepted June 3, 2009; Online Publication, September 7, 2009
[doi:10.1271/bbb.90013]
Actinomycete cytochrome P450 from Nonomuraea
recticatena NBRC 14525 (P450moxA) catalyzes the
hydroxylation of a broad range of substrates, including
fatty acids, steroids, and various aromatic compounds.
Hence, the enzyme is potentially useful in medicinal
applications, but the activity is insufficient for practical
use. Here we applied directed evolution to enhance the
activity. A random mutagenesis library was screened
using 7-ethoxycoumarin as a substrate to retrieve 17
variants showing >2-fold activities. Twenty-five amino
acid substitutions were found in the variants, of which
five mutations were identified to have the largest effects
(Q87W, T115A, H132L, R191W, and G294D). These
mutations additively increased the activity; the quintet
mutant had 20-times the activity of the wildtype. These
five single mutations also increased in activity toward
structurally distinct substrates (diclofenac and naringe-
nin). Based on the three-dimensional structure of the
enzyme, we discerned that mutations in the substrate
recognition site improved the activity, which was
substrate dependent; mutations apart from the active
site improved the activity as well as the substrates did.
natural products, including the anticancer agent doxor-
ubicin,4) the immunosuppressant rapamycin,5,6) the live-
stock antibiotic tylosin,7) and the antiparasitic agent
avermectin.8) These relevant reactions conducted by
P450s can also be exploited ex vivo in the synthesis of
various man-made medicinal compounds. For example,
P450sca from Streptomyces carbophilus has success-
fully been applied to the bioconversion of ML236-B
(compactin) to produce pravastatin.9,10)
Recently, we identified P450 MoxA (P450moxA), a
class I P450, from the actinomycete Nonomuraea
recticatena NBRC 14525, which catalyzes the hydrox-
ylation of fatty acids, steroids, and aromatic com-
pounds.11–13) Many of these reactions were human
P450-like, indicating that the P450moxA-catalyzed
transformation system would be applicable in the
preparation of human or actinomycete-specific P450
metabolites from a diverse set of pharmaceuticals in the
early stages of drug development. The broad substrate
specificity of the enzyme is highly attractive for
medicinal applications. However, the catalytic efficiency
of the wild-type enzyme is insufficient; the conversion
of substrate 7-ethoxycoumarin to 7-hydroxycoumarin
occurs at a rate of only about 0.002 nmol/min/[nmol
enzyme], approximately two orders of magnitude lower
than those of common bacterial three-component P450s.
Over the past decade, various methods of protein
engineering have been applied to P450s, either by
rational processes or evolutionary ones. The active site
residues of CYP101 from Pseudomonas putida
(P450cam) was modified by site-directed mutagenesis,
based on the three-dimensional structure, to enhance
activity.14–17) Directed evolution was used to convert
mammalian CYP 1A2 to variants with 5- and 10-fold
activities toward 7-methoxyresorufin and 2-amino-3,
and 5-dimethylimidazo[4,5-F]quinoline respectively.18,19)
Mammalian CYP 2B1 was modified by directed evolu-
tion to yield a variant with 6-fold activity against
7-ethoxy-4-trifluoromethylcoumarin.20) To date, the
most extensive studies have been done on CYP 102A1
from Bacillus megaterium (P450BM3), a class III P450,
in which the P450 and reductase domains are connected
Key words: actinomycete; cytochrome P450; directed
evolution; hydroxylation; substrate specificity
Actinomycetes show an outstanding ability to produce
a wide variety of biologically active compounds,
reflecting the wide diversity of the enzymes involved
in biosynthesis. Among these enzymes, cytochrome
P450s (P450s), which form an important heme-contain-
ing monooxygenase superfamily, catalyze the incorpo-
ration of a single atom of molecular oxygen into
substrates. Whole genome sequencing of actinomycete
strains has indicated that every actinomycete has
approximately 20–30 P450 genes in its genome.1,2) It
is likely that most of the P450s found in actinomycete
genomes are class I enzymes according to the classi-
fication proposed by Chapman.3) These require ferre-
doxin and ferredoxin reductase as electron transport
cofactors. Many P450 genes found in actinomycetes are
involved in the biosyntheses of commercially important
y
To whom correspondence should be addressed. Kentaro MIYAZAKI, Institute for Biological Resources and Functions, National Institute of
Advanced Industrial Science and Technology (AIST), Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8566, Japan; Tel: +81-29-861-6060;
Fax: +81-29-861-6413; E-mail: miyazaki-kentaro@aist.go.jp; Akira ARISAWA, Tel: +81-538-21-1134; Fax: +81-538-21-1135; E-mail:
arisawa-a@mercian.co.jp

Directed Evolution of Actinomycete CYP105 MoxA
1923
Fig. 1. P450moxA Catalysis of 7-Ethoxycoumarin, Diclofenac Sodium, and Naringenin.
Expression system for P450moxAB. The moxA gene was amplified
by PCR using NdeI-tagged forward primer MoxA-F (50-GCCCATAT-
GACGAAGAACGTCGCCGAC-30) and SacI-tagged reverse primer
MoxA-R (50-GCGAGCTCCTCACCAGGTGACCGGGAGTTC-30),
whereas the moxB gene (a ferredoxin gene located downstream of
moxA) was amplified using SacI-tagged forward primer MoxB-F (50-
GCGAGCTCGACATGATGCGGATCAAAGCG-30) and SpeI-tagged
reverse primer MoxB-R (50-GCACTAGTCAGGCGTCCCGCAGGA-
TGGA-30). The reaction mixture was heated at 94 ꢀC for 2 min and
subjected to 25 cycles of 98 ꢀC for 20 s, 63 ꢀC for 30 s, and 68 ꢀC for
2 min, followed by a final extension at 68 ꢀC for 5 min. The amplicon
was digested with NdeI and SacI for moxA and SacI and SpeI for moxB.
Equal amounts of the fragments were mixed and cloned into the NdeI-
SpeI sites of pT7NS-camAB11,13) to yield expression plasmid
pMox100. This expression plasmid contains ‘‘redox partners,’’ viz.,
ferredoxin (moxB) from N. recticatena and putidaredoxin (pdx) and
putidaredoxin (pdR) reductase both from P. putida, facilitating the co-
expression of all these components under the control of the T7
promoter.11,13) Both pdx and pdR were inserted downstream of moxB
to support electron transport in E. coli.
in a single peptide chain. This unique property makes it
a model enzyme for fundamental studies of P450s,21,22)
but the substrate specificities of this class of P450s are
rather narrow, mostly restricted to fatty acids that may
not be suitable for value-added medicinal uses. In
contrast, P450s from actinomycetes possess intrinsic
activities that are responsible for the production of
biologically active secondary metabolites (e.g., regio-
and stereo-specific hydroxylation of steroids, etc.), but
few protein engineering studies have been performed on
P450s of actinomycete origin.
In this study, we used directed evolution to enhance
P450moxA activity. Because the prominent property of
the enzyme is broad substrate specificity, as described
above, we investigated to determine how directed
evolution for one specific substrate would affect activity
against others. First we used 7-ethoxycoumarin as a
substrate in primary screening because it generates a
fluorescent product, 7-hydroxycoumarin (Fig. 1), and
is therefore suitable for high throughput screening.
Variants with improved activity were then subjected to
a second assay using structurally different diclofenac
sodium and naringenin. We identified several variants
that showed enhanced activities without compromising
intrinsic broad specificity. Finally, we discuss the
structure–function relationship based on the recently
determined three-dimensional structure of the enzyme.23)
Construction of a random mutagenesis library and screening.
Random mutagenesis of the moxA gene was carried out by error-prone
PCR using the Diversify PCR random mutagenesis kit (Clontech,
Mountain View, CA). The reaction mixture (50 ml) contained 5 ml 10ꢁ
Titanium Taq Buffer, 1 ml 50ꢁ Diversify dNTP mix, 32 nmol MnSO4,
2 nmol dGTP, 1 ng pMox100, 10 pmol each of forward (50-GTGAG-
CGGATAACAATTCCCCTC-30) and reverse primer (50-GTACCGA-
CGATGACCACGTTGTC-30), and 1 ml Titanium Taq polymerase. The
reaction mixture was heated at 95 ꢀC for 5 min and subjected to 25
cycles of 98 ꢀC for 20 s, 63 ꢀC for 30 s, and 68 ꢀC for 2 min, followed
by a final extension at 68 ꢀC for 5 min. The amplicon was digested with
NdeI and SacI, cleaned, and ligated into the corresponding sites on
pMox100. Competent E. coli DH5ꢀ cells were then transformed with
the ligation mixture and transformants grown on Luria–Bertani (LB)
agar plates containing 50 mg/l of ampicillin (approximately 1,000
clones) were collected. The mutant pool was seeded in LB medium and
cultivated at 37 ꢀC overnight. Plasmid DNA was isolated and used as a
randomly mutated moxA gene library. A small portion of the plasmid
library was then used to transform E. coli BL21(DE3). Gene
expression was carried out essentially as described previously.11) The
transformants were grown on LB agar plates containing 50 mg/l of
ampicillin at 37 ꢀC. Colonies were selected and inoculated in 1 ml of
M9mix-carbenicillin medium (6.78 g/l Na2HPO4, 3 g/l KH2PO4,
0.5 g/l NaCl, 1 g/l NH4Cl, 10 g/l casamino acids, 4 g/l D-glucose,
0.1 mM CaCl2, 1 mM MgCl2, 0.1 mM FeSO4, 20 mg/l thymine, and
50 mg/l carbenicillin) in separate wells of a 96-well plate. Each 96-
Materials and Methods
Strains and reagents. N. recticatena NBRC 14525 was used as a
source of moxAB genes.12) Escherichia coli DH5ꢀ was used as a
general cloning host, and BL21(DE3) was used for gene expression. 7-
Ethoxycoumarin and naringenin (40,5,7-trihydroxyflavanone) were
purchased from Sigma-Aldrich (St. Louis, MO), diclofenac sodium
was from Wako (Osaka, Japan), restriction enzymes and a DNA
ligation kit were from Takara Bio (Shiga, Japan), KOD-plus DNA
polymerase was from Toyobo (Osaka, Japan), PCR purification and gel
extraction kits were from Qiagen (Hilden, Germany), and the
BugBuster and Overnight Express Autoinduction reagent was from
Novagen (Madison, WI).

1924
H. KABUMOTO et al.
Saturation mutagenesis.
well plate contained 94 mutants, and one positive (carrying the wild-
type moxA gene) and one negative control (pT7NS-camAB only). The
well plates were agitated at 800 rpm at 30 ꢀC overnight in an Iwashiya
(Tokyo) incubator shaker (SCF-98-8). The culture was then transferred
using a 96-pin replicator to another 96-well plate containing 1 ml
of fresh M9CPX medium (6.78 g/l Na2HPO4, 3 g/l KH2PO4, 0.5 g/l
NaCl, 1 g/l NH4Cl, 10 g/l casamino acids, 0.1 mM CaCl2, 0.1 mM
FeSO4, 20 mg/l thymine, 80 mg/l 5-aminolevulinic acid, and 50
mg/l carbenicillin), and Overnight Express Autoinduction solutions
(Novagen). After incubation at 25 ꢀC for 24 h, the cells were collected
by centrifugation at 1;200 ꢁ g for 10 min. The supernatant was
discarded and the cells were resuspended in 1 ml of CV-2 buffer
(50 mM sodium phosphate buffer, pH 7.4, 2% v/v glycerol, 50 mg/l
carbenicillin, 0.1 mM isopropyl-ꢁ-D-thiogalactopyranoside, and 0.1 mM
7-ethoxycoumarin). After 4 h of incubation at 28 ꢀC, the reaction
mixture was centrifuged at 1;200 ꢁ g for 10 min. The supernatant,
200 ml, was transferred to a 96-well black plate and the fluorescence
intensity of the product (7-hydroxycoumarin) was measured on a
Molecular Devices (Sunnyvale, CA) fluorescent plate reader (Spec-
tramax GEMINI XS) with an emission wavelength of 460 nm under
excitation at 380 nm.
A set of complementary degenerate
primers, SA-F (50-CTACTTCACCATCGCCGACNNKGTCACCT-
CCCGGC-30) and SA-R (50-GGCCAGCCGGGAGGTGACMNNGT-
CGGCGATGGTG-30) were designed, where N ¼ A, G, C, T; K ¼ G
or T; and M ¼ A or C, to randomize the 294th amino acid of
P450moxA (Gly in the wildtype). Two separate PCRs were carried
out to amplify the N- (895-bp) and C- (355-bp) terminal fragments
using primers MoxA-F and SA-R for the N-fragments, and SA-F and
MoxA-R for the C-fragments. The products were combined at a 1:1
molar ratio and used as the template for the third PCR using MoxA-F
and MoxA-R primers to obtain the full-length gene. The PCR mixture
(50 ml) contained an equimolar mixture of the PCR products, 20 pmol
of each primer, 10 nmol of each dNTP, 50 nmol of MgSO4, 1ꢁ KOD
buffer, and 1 unit of KOD-plus DNA polymerase (Toyobo). PCR was
performed at 95 ꢀC for 5 min, followed by 30 cycles of 98 ꢀC for 20 s,
63 ꢀC for 30 s, and 68 ꢀC for 2 min, followed by a final extension at
68 ꢀC for 5 min. This amplicon was cloned back into the NdeI-SacI
sites of pMoxA100 to produce a site-saturation mutagenesis library.
Activity screening and spectral analysis were carried out as described
above.
DNA sequencing. Nucleotide substitutions included in variant
enzymes were determined by DNA sequencing for both strands of
DNA on an Applied Biosystems 3130 sequencer using the Big Dye
Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Foster
City, CA).
Determination of enzyme activities. Enzyme activities were deter-
mined by in vivo whole-cell assays. Cells carrying P450moxA variants
were grown in 25 ml of medium, as described in the previous section.
The cells were re-suspended in 5 ml of CV-2 buffer and divided into two
15-ml tubes (each a 2-ml aliquot). One of the cell suspensions was used
in biotransformation and the other in determination of the P450
concentration. For the reaction with 7-ethoxycoumarin, the substrate
was added to the cell suspension at a final concentration of 0.1 mM and
the tube was agitated at 120 rpm at 28 ꢀC for 4 h. After the reaction was
terminated by the addition of 200 ml of 20% trichloroacetic acid, the
reaction mixture was centrifuged at 1;200 ꢁ g for 10 min. The super-
natant was measured on a fluorescent plate reader with emission
wavelength at 460 nm under excitation at 380 nm. The concentration of
the product was calculated from the fluorescence intensity, referring to
the fluorescence intensity of known concentrations of standard samples.
For the reaction with diclofenac sodium, the substrate was added to
the cell suspension at a final concentration of 0.1 mM and the tube was
agitated at 120 rpm at 28 ꢀC for 1 h. The reaction was terminated by the
addition of 1 M HCl to pH 2–3 and extracted with 2 ml of ethyl acetate.
The organic layer (1 ml) was evaporated to dryness under a vacuum.
The residue was then dissolved in 200 ml, methanol and the products
were analyzed by the Agilent 1100 HPLC system under the following
conditions: column, J’sphere ODS-H80 (4 mm ID ꢁ 75 mm); flow rate,
1.0 ml/min; column temperature, 40 ꢀC; detection, 267 nm; eluent,
0.85% phosphoric acid (in water):methanol, 40:60. The concentration
of the product was calculated from the HPLC peak area, referring to
the peak area of known concentrations of standard samples.
Mapping of mutation on structure. A model of the mutant was built
˚
on the basis of the 2.8 A resolution structure of P450moxA (Protein
Date Bank [PDB] code 2z36).23) The positions of mutation were
mapped on the P450moxA structure using Pymol.24)
Results and Discussion
Directed evolution of actinomycete P450moxA
In this study, we used directed evolution to modify
class I P450s from actinomycetes in an attempt to
improve their properties or complement their drawbacks
further. The specific concerns of the study were: (i)
improvement of P450moxA activity against 7-ethoxy-
coumarin; (ii) to test activity against diclofenac and
naringenin to determine whether variants with improved
activity against 7-ethoxycoumarin retain intrinsic broad
substrate specificities; and (iii) to understand the
structure–function relationship on the basis of the
three-dimensional structure.
For the reaction with naringenin, 2 ml of the cell suspension were
mixed with 0.1 mM of the substrate and agitated at 120 rpm at 28 ꢀC for
3 h. The reaction was terminated by the addition of 1 M HCl to pH 2–3,
and 2 ml of methanol was then added. After brief centrifugation
(1;200 ꢁ g, 5 min), the supernatant was analyzed using the Agilent
1100 HPLC system under the following conditions: column, J’sphere
ODS-H80 (4 mm ID ꢁ 75 mm); flow rate, 1.0 ml/min; column temper-
ature, 40 ꢀC; detection, 290 nm; eluent, gradient time program with
methanol/0.1% acetic acid (in water) at methanol concentrations of
30% (0–8 min), 30 to 70% (8–9.5 min), 70% (9.5–10 min), and 70 to
30% (10–13.5 min). The concentration of the product was calculated
from the HPLC peak area of reversed-phase HPLC by the procedure as
described above.
Screening of the error-prone PCR library
A random mutagenesis library of P450moxA was
constructed by error-prone PCR. The mutation rate was
adjusted to 1–4 nucleotide substitutions per 1,000 bp per
gene copy. The mutated gene fragment was cloned back
into an expression plasmid, which co-expressed moxB
from N. recticatena and pdx and pdR from P. putida
under the control of the T7 promoter.11,13) High-
throughput screening of the library was performed using
7-ethoxycoumarin as the substrate, generating fluores-
cent 7-hydroxycoumarin as the product. M9-based
medium was used instead of rich media (e.g., LB or
TB) to minimize background fluorescence.
A total of 800 clones were screened. Approximately
40 variants surpassed the level of wild-type activity, of
which the top 17 clones were selected. DNA sequencing
of these clones revealed the presence of redundant
clones (Table 1), because the mutant plasmid library
was amplified once in E. coli DH5ꢀ prior to introduction
to an expression host of E. coli BL21(DE3). Of 11
independent variants, 25 different amino acid substitu-
For determination of the P450 concentration, lysozyme (4 mg/ml)
and BugBuster (Novagen) were added to the prepared 2-ml cell
suspension. After 20 min of incubation at 30 ꢀC, debris was removed by
centrifugation and the cell lysate (supernatant) was used for measure-
ment of the P450 concentration from the reduced CO difference
spectra.24) The cell lysate, reduced with sodium dithionite (Wako), was
bubbled with carbon monoxide for 30 s, and A450 and A490 were
measured with
a non-bubbled aliquot as the blank. The P450
concentration (nM) was calculated according to the formula
ðA450 ꢂ A490Þ=91. The enzyme activities were estimated from the
concentration of product, the reaction time, and the P450 concentration.

Directed Evolution of Actinomycete CYP105 MoxA
Table 1. Variants Showing Increased Activity against 7-Ethoxycou-
marin Identified in a Library Generated by Error-Prone PCR
1925
A
Relative activitya
-fold increase
Clones
Amino acid substitutions
Wildtype
HAM1
HAM2
HAM3
HAM4
HAM5
HAM6
HAM7
HAM8
HAM9
HAM10
HAM11
HAM12
HAM13
HAM14
HAM15
HAM16
HAM17
—
G294E
1
3.4
3.5
3.3
2.4
2.4
2.8
2.8
2.1
2.3
3.3
2.6
2.1
2.4
3.1
3.6
2.3
2.7
G294E
G294E
Q87W/D202N
Q87W/D202N
E55K/R191W/T206S/T301S
E55K/R191W/T206S/T301S
C20S/H345R
B
T229A/V304A/A335S/V396D
T115A/H132L
S44C/T179A/L197Q
L158Q
Q91L/A300T
D63G/M321I
T115A/H132L
Q220H/L403F
D63G/M321I
aRelative activities are given under-fold increase relative to 7-hydroxycou-
marin production for the wildtype.
C
tions were identified. Two variants contained single
amino acid substitutions and nine variants contained
multiple substitutions. To identify key mutations for
enhanced activity, we constructed variants containing
one of 25 mutations, either by site-directed mutagenesis
or the cut and ligation procedure. Each of the five
variants with a single mutation (Q87W, T115A, H132L,
R191W, and G294E) were identified, all of which
showed more than 2-fold activity compared to the
wild-type.
To determine whether the increase in activity was due
to an increase in specific activity or to the expression
level, the concentration of each expressed P450 was
determined based on the CO difference spectrum
described by Omura and Sato.25) The former four
variants showed no notable change in expression,
indicating that the increased levels of activity were
due to an increase in catalytic activity against 7-
ethoxycoumarin (data not shown). The peak at 420 nm,
which represents an inactive form of P450,26) was not
observed. In contrast, the peak at 450 nm for G294E
increased over approximately 5 min; the final spectrum
indicated a combination of P450 and P420, suggesting
that the G294E variant has the unstable form.
Fig. 2. Specific Activities of P450moxA Variants against 7-Ethox-
ycoumarin (A), Diclofenac (B), and Naringenin (C).
Bars represent the mean ꢃ standard deviation for four independ-
ent experiments.
Cumulative effects of single-point mutations
The specific activities of the five variants Q87W,
T115A, H132L, R191W, and G294D were determined
(Fig. 2A). Q87W exhibited a 7-fold increase, whereas
the other four variants, T115A, H132L, R191W, and
G294D, showed 2- to 4-fold increases. To determine
whether these mutations would additively enhance
activity, variants with multiple mutations were produced.
Activity increased in most combinations. Q87W and
T115A alone produced 7- and 2-fold increases respec-
tively, whereas the double mutant Q87W/T115A pro-
duced a 13-fold increase. Similarly, R191W and G294D
alone gave 2.5- and 4-fold increases respectively,
whereas R191W/G294D showed a 8.5-fold increase.
Yet further enhancement was observed in the triple
mutant H132L/R191W/G294D, with a 17-fold increase.
The quintet mutant containing all mutations in a single
protein (Q87W/T115A/H132L/R191W/G294D) showed
a 20-fold increase.
Saturation mutagenesis of the 294th amino acid
Although G294E was identified as one of the most
effective mutations, that mutation led to a loss of
stability. Hence, we applied site-saturation mutagene-
sis27) to further modify the 294th amino acid to obtain a
stable variant. Through the screening 200 clones, a
variant was identified that had comparable activity to
G294E. DNA sequencing revealed that that variant
contained an Asp residue at the 294th residue. Despite
the apparently similar activity to G294E, the G294D
variant lost the peak at 420 nm, indicating stable ligand
formation around the heme and uniform, active con-
formation. In addition to the four remaining variants, we
used this G294D variant in further studies.
Activity against diclofenac and naringenin
P450moxA accepts an array of organic compounds as
substrates.1,28) Diclofenac sodium and naringenin are
structurally different from 7-ethoxycoumarin, and yet are

1926
H. KABUMOTO et al.
the preferred substrates of this enzyme (Fig. 1). We
tested to determine whether the improved P450moxA
variants would also show improved activity with these
structurally distinct substrates. Wild-type P450moxA
catalyzes hydroxylation at C-40 of diclofenac sodium. All
variants with single mutations produced the same
product at increased efficiencies (Q87W, 2.4-fold;
T115A, 2.0-fold; H132L, 2.1-fold; R191W, 1.4-fold;
G294D, 2.9-fold; Fig. 2B). Thus all the single mutations
resulted in increased activity against diclofenac sodium,
although not to the level of activity against 7-ethox-
ycoumarin. As for the variants with multiple mutations,
Q87W/T115A showed a 5.2-fold increase and H132L/
R191W/G294D a 3.4-fold increase. The quintet mutant
Q87W/T115A/H132L/R191W/G294D showed a 7.7-
fold increase. Similar results were obtained using
naringenin as the substrate (Fig. 2C). A slight increase
in activity was observed (1.2- to 1.8-fold) for all variants.
As for the variants with multiple mutations, the quintet
mutant Q87W/T115A/H132L/R191W/G294D showed
the greatest increase (4.3-fold). By re-plotting the results
of these improvements (Fig. 3), we found that the degree
of improvement depended on the site of the muta-
tions. Q87W, R191W, and G294D showed preferences
in the order 7-ethoxycoumarin > diclofenac sodium ꢄ
naringenin, whereas T115A and H132L enhanced activ-
ity equally well, independently of the substrate. To date,
it is not certain what determines substrate dependency.
Structural basis of the increased activities
To gain further insight into the mechanism of
enhanced activity, we attempted to purify C-terminal
His-tagged variants, but failed to obtain large enough
amounts of proteins to perform detailed kinetic analysis.
Although it is difficult to discern the structure-function
relationships of enzymes in the absence of detailed
experimental evidence, we found some general tenden-
cies between the mutation site and substrate specificities
as follows.
Fig. 3. Substrate Dependence of Enhanced-Activity P450moxA
Variants with 7-Ethoxycoumarin (black bar), Diclofenac (shaded
bar), and Naringenin (open bar).
Fig. 5. Close-up View of Gly294.
The figure was produced using Pymol.6)
The specific activities of the wildtype are shown as 1.
Fig. 4. Stereo View of Mutation Sites Mapped on the Three-Dimensional Structure of P450moxA.
Mutation sites are shown in green; pale blue, helices; magenta, sheets; tint, loops; light green, heme; yellow, MES. The figure was produced
using Pymol.6)

Directed Evolution of Actinomycete CYP105 MoxA
1927
The overall structure of P450moxA exhibited a
References
typical P450 fold, as illustrated in Fig. 4, consisting of
14 ꢀ-helices (ꢀA-ꢀL, ꢀB0, and ꢀK0) and six ꢁ-strands
(ꢁ1–ꢁ6). Three of the five amino acids, at positions 87,
191, and 294, were located at the substrate recognition
site, whereas the amino acids at positions 115 and 132
were located distantly from the active site.
´
1) Bentley SD, Chater KF, Cerden˜o-Tarraga AM, Challis GL,
Thomson NR, James KD, Harris DE, Quail MA, Kieser H,
Harper D, Bateman A, Brown S, Chandra G, Chen CW, Collins
M, Cronin A, Fraser A, Goble A, Hidalgo J, Hornsby T,
Howarth S, Huang CH, Kieser T, Larke L, Murphy L, Oliver K,
O’Neil S, Rabbinowitsch E, Rajandream MA, Rutherford K,
Rutter S, Seeger K, Saunders D, Sharp S, Squares R, Squares S,
Taylor K, Warren T, Wietzorrek A, Woodward J, Barrell BG,
Parkhill J, and Hopwood DA, Nature, 417, 141–147 (2002).
2) Ikeda H, Ishikawa J, Hanamoto A, Shinose M, Kikuchi H, Shiba
T, Sakaki Y, Hattori M, and Omura S, Nat. Biotechnol., 21,
526–531 (2003).
Positions 87 (on ꢀB0) and 191 (on F-G loop) were
located at the entrance of the tunnel through which the
substrate enters the heme site. Both regions have been
suggested to play a key role in substrate recognition and
uptake.29) In the site-directed mutagenesis experiment,
the mutation of Y96 on ꢀB0 of P450cam showed a 6-fold
enhanced oxidation of styrene.30) Recent studies of
P450cam and P450BM3 also indicate that small changes
in backbone conformation at the entrance to the active
site allow substrates access to the active site.31) Because
Q87 and R191 are exposed to the surface, the tryptophan
substitutions might have facilitated smoother uptake of
substrates through ꢂ-ꢂ stacking between the aromatic
side chains and substrates. Gly294 was located close to
the heme and was in contact with the hydrophobic
cluster formed by Trp50, Phe78, and Ile318 (Fig. 5).
The mutation of Gly294 to aspartic acid can affect the
hydrophobic cluster by introducing steric hindrance and
negative charge. The loop carrying the Asp294 would
then shift away from the hydrophobic cluster. Mutation
to Glu would be expected to further shift the loop,
leading to disruption of the local environment around
the heme, as observed in the appearance of a peak at
420 nm in the CO difference spectrum.
Thr115 and His132 do not appear to be involved in
substrate binding, and it is unclear why we observed
enhanced activity. However, considering that these mu-
tations increased activity independently of the substrates,
the mutations may have accelerated the rate of electron
transfer rather than affecting the substrate interaction. A
similar phenomenon has been observed for P450cam.
Nagano et al.28) reported that either Arg109 or Arg112 in
P450cam is involved in the interaction with the ferre-
doxin subunit. The corresponding residue in P450moxA
is assumed to be Arg107 or Arg108. Residue-115 in
P450moxA is in proximity to these residues. Therefore,
the conversion of Thr115 to alanine may have affected
the interaction of Arg107 or Arg108 with putidaredoxin,
leading to acceleration of electron transfer.
3) Chapman FK, P450 Search on World Wide Web, http:
//www.chem.ed.ac.uk/chapman/P450 (2006).
4) Lomovskaya N, Otten SL, Doi-Katayama Y, Fonstein L, Liu
XC, Takatsu T, Inventi-Solari A, Filippini S, Torti F, Colombo
AL, and Hutchinson CR, J. Bacteriol., 181, 305–318 (1999).
5) Chung L, Liu L, Patel S, Carney JR, and Reeves CD,
J. Antibiot., 54, 250–256 (2001).
´
6) Molnar I, Aparicio JF, Haydock SF, Khaw LE, Schwecke T,
Konig A, Staunton J, and Leadlay PF, Gene, 169, 1–7 (1996).
´
7) Fouces R, Mellado E, Dıez B, and Barredo JL, Microbiology,
¨
145, 855–868 (1999).
8) Ikeda H, Nonomiya T, Usami M, Ohta T, and Omura S, Proc.
Natl. Acad. Sci. USA, 96, 9509–9514 (1999).
9) Matsuoka T, Miyakoshi S, Tanzawa K, Nakahara K, Hosobuchi
M, and Serizawa N, Eur. J. Biochem., 184, 707–713 (1989).
10) Serizawa N and Matsuoka T, Biochim. Biophys. Acta, 1084, 35–
40 (1991).
11) Agematu H, Matsumoto N, Fujii Y, Kabumoto H, Doi S,
Machida K, Ishikawa J, and Arisawa A, Biosci. Biotechnol.
Biochem., 70, 307–311 (2006).
12) Fujii Y, Hirosue S, Fujii T, Matsumoto N, Agematu H, and
Arisawa A, Biosci. Biotechnol. Biochem., 70, 2299–2302
(2006).
13) Arisawa A and Agematu H, ‘‘Modern Biooxidation Enzymes,
Reactions and Applications,’’ eds. Schmid RD and Urlacher VB,
Wiley-VCH Verlag, Weinheim, pp. 177–192 (2007).
14) Bell SG, Stevenson JA, Boyd HD, Campbell S, Riddle AD,
Orton EL, and Wong LL, Chem. Commun., 7, 490–491 (2002).
15) Harford-Cross CF, Carmichael AB, Allan FK, England PA,
Rouch DA, and Wong LL, Protein Eng., 13, 121–128 (2000).
16) Jones JP, O’Hare EJ, and Wong LL, Eur. J. Biochem., 268,
1460–1467 (2001).
17) Loida PJ and Sligar SG, Biochemistry, 32, 11530–11538 (1993).
18) Kim D and Guengerich FP, Arch. Biochem. Biophys., 432, 102–
108 (2004).
19) Kim D and Guengerich FP, Biochemistry, 43, 981–988 (2004).
20) Kumar S, Chen CS, Waxman DJ, and Halpert JR, J. Biol.
Chem., 280, 19569–19575 (2005).
21) Farinas ET, Schwaneberg U, Glieder A, and Arnold FH, Adv.
Synth. Catal., 343, 601–606 (2001).
22) Glieder A, Farinas ET, and Arnold FH, Nat. Biotechnol., 20,
1135–1139 (2002).
Conclusion
23) Yasutake Y, Imoto N, Fujii Y, Fujii T, Arisawa A, and Tamura
T, Biochem. Biophys. Res. Commun., 361, 876–882 (2007).
24) DeLano WL, The PyMOL Molecular graphics system on World
Wide Web, http://www.pymol.org. (2002).
We improved the activity of class I P450moxA from
the actinomycete N. recticatena NBRC 14525 by di-
rected evolution. These mutations additively enhanced
activity without compromising broad substrate specific-
ity. Because the chief interest of actinomycete P450s,
particularly CYP105, lies in their broad substrate
specificity, our results show the potential of these
enzymes in related biotechnological applications.
25) Omura T and Sato R, J. Biol. Chem., 239, 2370–2378 (1964).
26) Perera R, Sono M, Sigman JA, Pfister TD, Lu Y, and Dawson
JH, Proc. Natl. Acad. Sci. USA, 100, 3641–3646 (2003).
27) Miyazaki K and Arnold FH, J. Mol. Evol., 49, 716–720 (1999).
28) Nagano S, Shimada H, Tarumi A, Hishiki T, Kimata-Ariga Y,
Egawa T, Suematsu M, Park SY, Adachi S, Shiro Y, and
Ishimura Y, Biochemistry, 42, 14507–14514 (2003).
29) Graham-Lorence S and Peterson JA, FASEB J., 10, 206–214
(1996).
Acknowledgments
30) Bell SG, Harford-Cross CF, and Wong LL, Protein Eng., 14,
797–802 (2001).
The authors thank Dr. Yoshiaki Yasutake and Dr.
Tomohiro Tamura, both of AIST, for helpful discussion
and comments on the manuscript.
31) Winn PJ, Ludemann SK, Gauges R, Lounnas V, and Wade RC,
¨
Proc. Natl. Acad. Sci. USA, 99, 5361–5366 (2002).