Investigation of the substrate range of CYP199A4: Modification of the partition between hydroxylation and desaturation activities by substrate and protein engineering

Article abstract of DOI:10.1002/chem.201202776

The cytochrome P450 enzyme CYP199A4, from Rhodopseudomonas palustris HaA2, can efficiently demethylate 4-methoxybenzoic acid. It is also capable of oxidising a range of other related substrates. By investigating substrates with different substituents and ring systems we have been able to show that the carboxylate group and the nature of the ring system and the substituent are all important for optimal substrate binding and activity. The structures of the veratric acid, 2-naphthoic acid and indole-6-carboxylic acid substrate-bound CYP199A4 complexes reveal the substrate binding modes and the side-chain conformational changes of the active site residues to accommodate these larger substrates. They also provide a rationale for the selectivity of product oxidation. The oxidation of alkyl substituted benzoic acids by CYP199A4 is more complex, with desaturation reactions competing with hydroxylation activity. The structure of 4-ethylbenzoic acid-bound CYP199A4 revealed that the substrate is held in a similar position to 4-methoxybenzoic acid, and that the C ^β C-H bonds of the ethyl group are closer to the heme iron than those of the C^α (3.5 vs. 4.8 A?). This observation, when coupled to the relative energies of the reaction intermediates, indicates that the positioning of the alkyl group relative to the heme iron may be critical in determining the amount of desaturation that is observed. By mutating a single residue in the active site of CYP199A4 (Phe185) we were able to convert the enzyme into a 4-ethylbenzoic acid desaturase. Engineering a P450 desaturase: The substrate range of CYP199A4 from Rhodopseudomonas palustris was investigated. The partition between the hydroxylation and desaturation activities of 4-ethylbenzoic acid was studied by changing the substrate and by mutation. The activity of CYP199A4 with 4-ethylbenzoic acid was changed to a desaturase by a single mutation at F185. Copyright

Full text of DOI:10.1002/chem.201202776

FULL PAPER
DOI: 10.1002/chem.201202776
Investigation of the Substrate Range of CYP199A4: Modification of the
Partition between Hydroxylation and Desaturation Activities by Substrate
and Protein Engineering
Stephen G. Bell,*[a, b] Ruimin Zhou,[c] Wen Yang,[c] Adrian B. H. Tan,[a]
Alexander S. Gentleman,[b] Luet-Lok Wong,*[a] and Weihong Zhou*[c]
Abstract: The cytochrome P450
enzyme CYP199A4, from Rhodopseu-
domonas palustris HaA2, can efficient-
ly demethylate 4-methoxybenzoic acid.
It is also capable of oxidising a range
of other related substrates. By investi-
gating substrates with different sub-
stituents and ring systems we have
been able to show that the carboxylate
group and the nature of the ring
system and the substituent are all im-
portant for optimal substrate binding
and activity. The structures of the vera-
tric acid, 2-naphthoic acid and indole-
strate binding modes and the side-
chain conformational changes of the
active site residues to accommodate
these larger substrates. They also pro-
vide a rationale for the selectivity of
product oxidation. The oxidation of
alkyl substituted benzoic acids by
CYP199A4 is more complex, with de-
saturation reactions competing with hy-
droxylation activity. The structure of 4-
ethylbenzoic acid-bound CYP199A4
revealed that the substrate is held in a
similar position to 4-methoxybenzoic
b
ꢀ
acid, and that the C C H bonds of the
ethyl group are closer to the heme iron
than those of the Ca (3.5 vs. 4.8 ꢀ).
This observation, when coupled to the
relative energies of the reaction inter-
mediates, indicates that the positioning
of the alkyl group relative to the heme
iron may be critical in determining the
amount of desaturation that is ob-
served. By mutating a single residue in
the active site of CYP199A4 (Phe185)
we were able to convert the enzyme
into a 4-ethylbenzoic acid desaturase.
Keywords: biocatalysis
·
cyto-
chromes · desaturation · hydroxyla-
tion · X-ray crystal structure
6-carboxylic
acid
substrate-bound
CYP199A4 complexes reveal the sub-
Introduction
cent atom can yield the dehydrogenation product.[2b] Other
ꢀ
types of activity are also known and include C C bond
ꢀ
Cytochrome P450 (CYP) enzymes are heme-dependent
monooxygenases with physiological roles that include hor-
mone and secondary metabolite biosynthesis, as well as
xenobiotic metabolism.[1] CYP enzymes primarily catalyze
the insertion of an oxygen atom from dioxygen into the
cleavage, oxidative dealkylation, C C bond formation and
various ring alterations.[1b,c,3] The stable, soluble nature of
bacterial CYP enzymes and the wide range of compounds
they oxidise are of interest for applications in synthesis in-
[4]
ꢀ
volving C H bond oxidation. The ability to engineer these
ꢀ
carbon hydrogen bonds of organic molecules. Mechanisti-
bacterial CYP enzymes by rational or random methods can
further enhance their activities and broaden their substrate
range.[4b,5] A detailed knowledge of the enzyme–substrate in-
teractions is critical in these efforts in order to maximise the
catalytic efficiency of the biocatalytic systems and to achieve
the turnover numbers required for synthetic applications.
Rhodopseudomonas palustris is a family of purple photo-
synthetic bacteria that are isolated in temperate, aquatic
sediments and possess extraordinary metabolic versatility.
Bacteria classified as Rhodopseudomonas spp. can grow
with or without light or oxygen, fix nitrogen and have a
wide range of biodegradation abilities.[6] Studies have shown
that different Rhodopseudomonas species share many char-
acteristics but different species have unique sets of function-
al genes to take advantage of the microenvironment in
which they are found.[7] We recently reported the properties
of the complete CYP199A2 and CYP199A4 class I P450 sys-
tems[8] from the R. palustris strains CGA009 and HaA2.[9]
Both CYP199A2 and CYP199A4 have a high affinity for 4-
cally, hydrogen abstraction by Compound I is followed by
oxygen rebound to give the hydroxylated compound.[2] Al-
ternatively, a second hydrogen atom abstraction at an adja-
[a] Dr. S. G. Bell,+ A. B. H. Tan, Dr. L.-L. Wong
Department of Chemistry, University of Oxford
South Parks Road, Oxford, OX1 3QR (UK)
E-mail: luet.wong@chem.ox.ac.uk
[b] Dr. S. G. Bell,+ A. S. Gentleman
School of Chemistry and Physics
University of Adelaide, SA 5005 (Australia)
E-mail: stephen.bell@adelaide.edu.au
[c] R. Zhou,+ Dr. W. Yang, Dr. W. Zhou
State Key Laboratory of Medicinal Chemical Biology
College of Life Sciences, Nankai University
Tianjin 300071 (P.R. China)
E-mail: zhouwh@nankai.edu.cn
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202776.
Chem. Eur. J. 2012, 18, 16677 – 16688
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
16677

methoxybenzoic acid and use similar electron transfer
chains consisting of a [2Fe-2S] ferredoxin (Pux and HaPux)
and a flavin-dependent ferredoxin reductase (PuR and
HaPuR) to mediate heme reduction by NADH.[8–10]
In addition to 4-methoxybenzoic acid CYP199A4 and
CYP199A2 also bind and oxidise 4-ethylbenzoic acid and
veratric acid. Both 4-methoxybenzoic acid and veratric acid
are O-demethylated to 4-hydroxybenzoic acid and vanillic
acid, respectively. This could proceed via oxidation of a
Here we compare the binding and kinetic properties of
CYP199A4 with selected substrates in order to more fully
understand fundamental aspects of binding and catalysis and
to better determine the biocatalytic scope of this enzyme.
By modifying the substrate and the protein we have been
able to alter the partition between hydroxylation and desa-
turation activity with alkyl substituted compounds. In addi-
tion, we report the crystal structures of CYP199A4 bound
with 4-ethylbenzoic acid, veratric acid, indole-6-carboxylic
acid and 2-naphthoic acid. All the structures have a closed
conformation with a bound chloride ion in a similar position
to that observed in the 4-methoxybenzoic acid-bound struc-
tures of CYP199A4 and CYP199A2. The substrates are ori-
entated in the CYP199A4 active site in such a fashion that
the reported regioselectivity of oxidation can be rational-
ised. These structures also provide a basis for understanding
how substrates that contain a carboxylic acid functional
group are recognised and bound by CYP199A4 and how the
enzyme could be engineered for improved substrate binding
and modified product profiles.
ꢀ
methyl C H bond followed by spontaneous decomposition
of the resulting hemiacetal.[11] 4-Ethylbenzoic acid oxidation
gives a more complex mixture with two majority and three
minority metabolites being formed. The two majority prod-
ucts (>90%) were identified as 4-vinylbenzoic acid, formed
a
b
ꢀ
by desaturation of the C C bond (36–38%), and 4-(1-hy-
droxyethyl)benzoic acid from hydroxylation at the Ca ben-
zylic carbon (51–59%).[9,10b] The alkene desaturation prod-
uct could arise from initial Ca hydrogen atom abstraction
followed by a second hydrogen atom abstraction from Cb of
the radical intermediate. However, the substrate radical can
also undergo one-electron oxidation by the Compound II
species to a carbocation that loses a proton to form the alke-
ne.[2b] Alternatively, initial hydrogen abstraction could occur
at the terminal carbon (Cb) followed by electron transfer ox-
idation to form a primary carbocation that rearranges to the
more stable, benzylic carbocation, which undergoes a Cb
proton abstraction.[2b,10b,12]
Results and Discussion
Substrate binding studies on the CYP199A4 system: A
range of substrates was tested for binding with CYP199A4
(Figure 1). We had shown previously that both 4-methoxy-
and 4-ethylbenzoic acid induced a ꢁ95% type I spin state
shift (Table 1 and Figure S1 in the Supporting Information)
whereas veratric acid shifted the spin state to 70% high-
spin.[9] The binding of veratric acid was considerably weaker
(29.5 mm) compared to 4-ethyl- and 4-methoxybenzoic acid
(0.34 and 0.28 mm, respectively). The nature of the para-sub-
CYP199A2 has also been reported to be a biocatalyst for
the oxyfunctionalisation of various aromatic carboxylic
acids, including 2-naphthoic acid and indole-6-carboxylic
acid. Using a hybrid whole-cell oxidation system with non-
physiological electron transfer proteins Furuya and Kino
identified the products as 7- and 8-hydroxy-2-naphthoic acid
and 2-indolinone-6-carboxylic acid, respectively.[13]
The crystal structure of substrate-free CYP199A2 has
been determined at 2.0 ꢀ resolution (PDB code: 2FR7).[10b]
The B’ helix found in some P450 structures is missing in
CYP199A2 and the G helix is broken into two (forming G
and G’ helices). These G and G’ helices are bent back from
the extended BC loop and the I helix to open up a clearly
defined substrate access channel. The crystal structures of 4-
methoxybenzoic acid-bound CYP199A2 and CYP199A4
have also been reported (1.8 and 2.0 ꢀ; PDB codes: 4DNJ
and 4DO1, respectively).[14] In the substrate-bound forms of
the enzymes there are significant changes in the positions of
residue side chains in the substrate access channel. These
changes and the binding of an anion (chloride), which caps
the entrance to the channel, result in a closed conforma-
tion.[14] In both these structures the substrate carboxylate
group interacts with polar and basic residue side chains
whereas hydrophobic interactions with the benzene ring
help hold the substrate in position. In this arrangement the
4-methoxy substituent of the substrate is oriented towards
the heme iron (3.9 ꢀ), which is consistent with exclusive
attack by CYP199A2 and CYP199A4 at the methoxy
methyl group, leading to demethylation to form 4-hydroxy-
benzoic acid as the sole product.[10b,14]
Figure 1. Substrates tested for binding with CYP199A4.
16678
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16677 – 16688

The Substrate Range of CYP199A4
FULL PAPER
Table 1. Substrate binding parameters and catalytic turnover activity data for CYP199A4 with various substra-
tes.[a]
To appraise the importance
of the planar benzene ring,
trans-4-ethyl-, trans-4-isopropyl-
and trans-4-methylcyclohexane-
carboxylic acid were assayed
with CYP199A4 (Table 1). All
Substrate
HS [%] Kd [mm]
N[b]
PFR[c]
Coupling [%][d]
4-methoxybenzoic acid
4-ethylbenzoic acid
veratric acid
4-isopropylbenzoic acid
4-methylbenzoic acid
4-methoxybenzaldehyde
trans-4-methylcyclohexanecarboxylic acid
trans-4-ethylcyclohexanecarboxylic acid
trans-4-isopropylcyclohexanecarboxylic acid
l-perillic acid
ꢁ95
ꢁ95
70
0.28ꢂ0.01 2180ꢂ57 1970ꢂ47
0.34ꢂ0.02 1180ꢂ34 1030ꢂ52
91ꢂ2
88ꢂ2
96ꢂ3
89ꢂ2
89ꢂ4
17ꢂ1
13ꢂ4
75ꢂ9
n.d.
29.5ꢂ3.1
0.29ꢂ0.1
0.66ꢂ0.05
197ꢂ7
1100ꢂ59 1060ꢂ88
ꢁ95
321ꢂ24
444ꢂ8.0
15ꢂ2.0
19ꢂ3.0
80ꢂ1.0
287ꢂ18
397ꢂ22
3ꢂ0.1
three
compounds
induced
70
smaller spin state shifts (55, 50
and 0% respectively) and
showed weaker binding (172-
80
0
55
50
90
5
50
[e]
–
2ꢂ0.1
172ꢂ16
187ꢂ15
10.1ꢂ0.6
26.0ꢂ1.0
44.4ꢂ2.2
60ꢂ7.0
86ꢂ2.0 n.d.
A
216ꢂ13
70ꢂ3.0
696ꢂ21
197ꢂ20
91ꢂ4
[f]
[f]
[f]
(ꢂ15) mm for 4-isopropylcyclo-
2-naphthoic acid
indole-6-carboxylic acid
–
–
–
–
[f]
hexanecarboxylic acid) com-
pared with their benzoic acid
analogues (Table 1 and Figur-
es S1 and S2). The lack of any
discernible spin state shift with
[a] The data are given as mean ꢂ S.D. with nꢁ3. The reaction mixtures (50 mm Tris, pH 7.4) contained P450
(0.5 mm), HaPux (5 mm) and HaPuR (0.5 mm). Rates are given as nmol(nmolCYP)ꢀ1 minꢀ1. [b] NADH turnover
rate. [c] Product formation rate. [d] The percentage of NADH consumed in the reaction that led to the forma-
tion of products. [e] Not determined due to absence of spin state shift. [f] Not determined due to further reac-
tions of the products.
stituent of the substrate is important for the binding affinity
to CYP199A2, for example, benzoic acid induces a smaller
spin state shift of 35%.[10a] In order to further investigate
the effect of the size of the para substituent on substrate
binding to CYP199A4 we studied the binding of 4-methyl-
and 4-isopropylbenzoic acids. The spin state shift and bind-
ing of 4-isopropylbenzoic acid was comparable to that of 4-
ethylbenzoic acid (ꢁ95% and 0.29 mm, Figure S2 in the Sup-
porting Information) whereas 4-methylbenzoic acid binding
resulted in a smaller shift to the high-spin state (70%, Fig-
ure S1) and the binding was also weaker (0.66 mm, Table 1
and Figure S2). Based on the location of 4-methoxybenzoic
acid in the structure of substrate-bound CYP199A4 (PDB
code: 4DO1)[14] and of 4-ethylbenzoic acid (PDB code:
4EGM, vide infra) the smaller methyl substituent may not
Figure 2. The active site of 4-methoxybenzoic acid-bound CYP199A4.
The 2 mFoꢀDFc density of the 4-methoxybenzoate (yellow) is shown con-
toured at approximately 0.35 eꢀ3 (blue mesh). The heme and the active
site residues are shown in grey and green, respectively. Wat170 is shown
as a red sphere and hydrogen bonding interactions as dashed lines.
be close enough to the heme to induce complete dissocia-
tion of the water ligand from the heme iron centre.
In the 4-methyoxybenzoic acid-bound crystal structure of
CYP199A4 (PDB code: 4DO1), the substrate carboxylate
group forms hydrogen bonds to the side chains of Arg92,
Ser95 and Ser244 (Figure 2). In addition, Arg243, which sub-
stantially changes its location between the substrate-bound
and substrate-free forms, interacts with the substrate carbox-
ylate group through a bridging water molecule via hydrogen
bonds. All of these interactions help to position the para-
substituent over the heme iron, and account for the sub-
strate specificity and product selectivity. Mutation of these
residues dramatically weakened substrate binding and re-
duced enzyme activity.[10a,14] We have also found that the
substrate carboxylate group is important for substrate bind-
ing: whereas 4-methoxybenzaldehyde induced a large spin
state shift (80%) the binding was three orders of magnitude
weaker (197(ꢂ7) mm) compared to 4-methoxybenzoic acid
(Table 1 and Figure 3). Similar results were obtained with 4-
ethyl- and 4-isopropylbenzaldehyde (both induced 90% spin
state shift) whereas the corresponding alcohols hardly in-
duced any spin state shift (4-ethylbenzylalcohol, 10% and 4-
isopropylbenzylalcohol 0%, data not shown).
4-methylcyclohexanecarboxylic acid reinforces the tenet that
the para-substituent requires a minimum of two atoms in
order for it to approach close enough to the heme iron to
displace the sixth water ligand effectively. Increasing the ri-
gidity of the six membered ring by introducing a double
bond was tested by evaluating l-perillic acid binding. This
induced a larger spin state shift (90%) than trans-4-ethyl-
and trans-4-isopropylcyclohexanecarboxylic acid and the
binding was also significantly tighter (Kd =10.1 mm, Fig-
ure S2) though still 30-fold weaker than for 4-ethyl- and 4-
isopropylbenzoic acid (Table 1). The data suggest that the
planar benzene ring is important for tight binding.[14]
CYP199A2 has been reported to oxidise larger aromatic
carboxylic acids including 2-naphthoic acid and indole-6-car-
boxylic acid.[13] 2-Naphthoic acid elicited a small spin state
shift with CYP199A4 (<10%, Figure S1) and the binding
(Kd =26(ꢂ1) mm) was also weaker compared with the benzo-
Chem. Eur. J. 2012, 18, 16677 – 16688
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16679

S. G. Bell, L.-L. Wong, W. Zhou et al.
Figure 3. Substrate binding constant analysis of: a) 4-methoxybenzalde-
hyde (Kd =197(ꢂ7) mm), and b) indole-6-carboxylic acid (Kd =44.4-
(ꢂ2.2) mm) with CYP199A4.
ic acids. Indole-6-carboxylic acid produced a larger spin
state shift (20%, Figure S1) but the binding affinity was
Figure 4. Substrates tested for product formation with CYP199A4. All
the products shown have been confirmed by GC co-elution experiments
with authentic standards and GC-MS, with the exception of the 4-isopro-
pylbenzoic acid products, which were rationalised on the basis of GC-MS
experiments and by analogy with the products of 4-ethylbenzoic acid.
The products of 2-naphthoic acid and indole-6-carboxylic acid are from
the work of others.[13]
lower (Kd =44.4
N
Activity and product formation assays: The oxidation of 4-
methoxybenzoic acid by CYP199A4 results in a single prod-
uct (4-hydroxybenzoic acid, Figure 4) with high NADH con-
sumption activity and tight coupling of the reducing equiva-
lents to product formation.[9] The NADH consumption rates
with veratric and 4-ethylbenzoic acids were lower but the re-
actions were still tightly coupled (Table 1). Veratric acid also
produced a single product (vanillic acid) from selective de-
methylation of the para-methoxy group whereas 4-ethylben-
zoic acid yielded two major products, 4-vinylbenzoic acid
(38%, TMS-derivatised product: m/z 221.0984 vs. 223.1064
ly due to a decrease in the NADH consumption rate with
coupling remaining high for both (Table 1 and Figure S4 in
the Supporting Information). Oxidation of 4-methylbenzoic
acid gave a single product, which was identified as 4-(hy-
droxymethyl)benzoic acid by GC co-elution with the TMS-
derivatised authentic compound and by GC-MS (TMS-deri-
vatised product: m/z 297.1221, Figure 4 and Figure S5 in the
Supporting Information). 4-Isopropylbenzoic acid oxidation
resulted in two major products (Figures 4 and 5). These
were assigned by GC-MS as the desaturation product 4-
(prop-1-en-2-yl)benzoic acid (54%, TMS-derivatised prod-
uct: m/z 235.1107 vs. 237.131083 for the starting material)
a
b
ꢀ
for the starting material) formed by C C bond desatura-
tion, and 4-(1-hydroxyethyl)benzoic acid (51%, TMS-deri-
vatised product: m/z 311.1473, Figure 4 and Figure S3 in the
Supporting Information).[9,10b]
The turnover activities with 4-methyl- and 4-isopropylben-
zoic acid were lower than 4-ethylbenzoic acid, predominant-
16680
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16677 – 16688

The Substrate Range of CYP199A4
FULL PAPER
NADH turnover activities with all the para-substituted ben-
zaldehydes and benzyl alcohols tested were very low and
little or no products were formed.
The turnover activity with the cyclohexanecarboxylic
acids was much lower than the benzoic acids. For example,
the product formation rate for trans-4-ethylcyclohexanecar-
boxylic acid was 60(ꢂ7) versus 1030(ꢂ52) nmol (nmol
G
CYP)ꢀ1 minꢀ1 with 4-ethylbenzoic acid (Table 1 and Fig-
ure S4). This reduction in activity predominantly arose from
a drop in the NADH consumption rate but the coupling of
the reducing equivalents to product formation was also
lower. The single product arising from trans-4-methylcyclo-
hexanecarboxylic acid oxidation was identified as trans-4-
(hydroxymethyl)cyclohexanecarboxylic acid by co-elution
with one of the peaks in a derivatised standard containing
cis- and trans-4-(hydroxymethyl)cyclohexanecarboxylic acid.
The oxidation of trans-4-ethylcyclohexanecarboxylic acid re-
sulted in two products. The TMS derivatives of both had
GC retention times that were far removed from the sub-
strate (tR =5.4 and 5.6 min vs. 3.2 min, Figure S6a in the
Supporting Information). Comparing the retention times of
these to those of the derivatised products of 4-ethylbenzoic
and trans-4-methylcyclohexanecarboxylic acid indicates that
both products are likely to be hydroxylation products rather
than desaturation products (Figure S6b). The GC-MS data
for both products indicated that this was indeed the case
(observed masses: 317.1673 and 317.1743, calcd: 317.1968
for [M+H]+, Figure S6c). It is also worth noting that the
oxidation of trans-4-isopropylcyclohexanecarboxylic acid re-
sulted in a single majority product at an extended retention
time with mass of approximately 331.2, indicating that this is
also a hydroxylation product (calcd: 331.212476 for [M+
H]+, Figure S7 in the Supporting Information). For these
compounds hydroxylation could potentially occur at Ca, Cb
or even the C4 and C3 carbons of the cyclohexyl ring and
further investigation is required in order to fully characterise
them.
The addition of a single double bond to the cyclohexane
ring in l-perillic acid resulted in an increase in the activity
versus trans-4-ethyl- and trans-4-isopropylcyclohexanecar-
boxylic acids (Table 1 and Figure S4). Four product peaks
were observed in the GC analysis—potentially due to the
more complex product profile that may arise from the prop-
1-en-2-yl substituent or the double bond on the cyclohexene
ring (Figure S8 in the Supporting Information). Due to the
multiple products we have as yet been unable to identify the
exact nature of each of the metabolites but the retention
times and GC-MS data indicate that the products are de-
rived from hydroxylation products (m/z of trimethylsilyl de-
rivatives 327.1778, calcd: 327.181176 for [M+H]+). We
were also able to estimate that the coupling of the reducing
electrons to product formation was high (91%).
Figure 5. a) Gas chromatography analysis of CYP199A4 turnovers with
4-ethylbenzoic acid (black, tR =3.6 min) and 4-isopropylbenzoic acid
(grey, substrate tR =4.3 min). The 9-fluorenol internal standard (tR =
6.3 min) is also shown. For clarity the chromatograms have been offset
along the y axis. MS analysis of: b) the TMS derivative of 4-isopropyl
benzoic acid, and the peaks assigned to: c) the TMS derivatives of 4-
(prop-1-en-2-yl)benzoic acid, and d) the hydroxylated product assigned
as 4-(1-methyl-1-hydroxyethyl)benzoic acid. The GC-MS analysis of the
4-ethylbenzoic acid turnover is provided in the Supporting Information
(Figure S2).
and 4-(1-methyl-1-hydroxyethyl)benzoic acid (30%, TMS-
derivatised product: m/z ~325.3, calcd: 325.165526 for [M+
H]+; Figure 5). The remaining unidentified products (total
16%) had longer GC retention times and may be other hy-
droxylation products, for example, 4-(1-methyl-2-hydroxye-
thyl)benzoic acid or further oxidation products.[9,10b] The
The activity of 2-naphthoic acid oxidation was low
(Table 1 and Figure S4) and it yielded two products, both of
which fluoresced under UV light and are presumably the 7-
and 8-hydroxynaphthoic acid products identified previously
(Figure 4 and Figure S9 in the Supporting Information).[13b]
Chem. Eur. J. 2012, 18, 16677 – 16688
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16681

S. G. Bell, L.-L. Wong, W. Zhou et al.
Table 2. Data collection and structure refinement statistics.
Data set
4-Ethylbenzoic acid
PDB: 4EGM
Veratric acid
PDB: 4EGN
Indole-6-carboxylic acid
PDB: 4EGO
2-Naphthoic acid
PDB: 4EGP
wavelength [ꢀ]
space group
1.5418
P212121
1.0000
P212121
0.9796
P212121
1.5418
C2221
cell dimensions a/b/c [ꢀ]
resolution [ꢀ]
106.9/143.3/172.5
50–2.90 (2.95–2.90)
14.9 (3.6)
99.7 (99.6)
4.4 (4.3)
106.5/143.4/172.8
50–2.00 (2.07–2.00)
16.6 (3.1)
98.5 (98.5)
6.0 (5.7)
106.9/143.7/172.9
50.0–1.76 (1.82–1.76)
15.4 (2.4)
90.0 (66.5)
4.6 (3.7)
143.6/172.2/106.3
50.0–3.00 (3.11–3.00)
15.5 (2.6)
99.9 (100)
5.3 (5.3)
average I/s(I)[a]
completeness [%][a]
redundancy[a]
Rmerge [%][a,b]
14.1 (56.6)
10.0 (59.5)
6.9 (41.7)
13.8/74.9
structure refinement statistics
average B-factor [ꢀ2]
Rwork/Rfree [%][c]
r.m.s.d. bond lengths [ꢀ]
r.m.s.d. bond angles [8]
Ramachandran favoured [%]
Ramachandran outlier [%]
molprobity score
39.5
24.3
23.2
18.3/21.3
0.030
2.22
98.4
0
1.19
0.47
4.1
62.6
18.1/24.9
0.017
1.719
95.91
0.32
2.01
2.35
8.22
17.0/20.7
0.028
2.094
98.34
0.06
1.26
1.25
3.96
23.3/28.7
0.014
1.798
85.8
1.3
2.36
poor rotamers [%]
all atom clash scores
1.3
13.72
[a] Values in parentheses correspond to the highest-resolution shell. [b] Rmerge =Si jIiꢀhIij/ShIi, where Ii is an individual intensity measurement and hIi is
the average intensity for all the reflections. [c] Rwork/Rfree =Sj jFo jꢀjFc j j/SjFo j, where Fo and Fc are the observed and calculated structure factors, re-
spectively.
The monooxygenase activity with indole-6-carboxylic acid
was higher, with a NADH consumption rate of 696-
A
drophobic interactions hold the benzene ring and position
the substrate methyl group closest to the iron, consistent
with the exclusive formation of the demethylation product.
We have now solved the structures of CYP199A4 complexes
with four other substrates, 4-ethylbenzoic acid (PDB code:
4EGM), 3,4-dimethoxybenzoic (veratric) acid (PDB code:
4EGN), 2-naphthoic acid (PDB code: 4EGP) and indole-6-
carboxylic acid (PDB code: 4EGO) to provide further in-
sight into substrate binding, specificity and product selectivi-
ty of these enzymes. Data collection and refinement statis-
tics of the complexes are given in Table 2. The electron den-
sities of individual substrates in the active sites, and the resi-
dues around, are shown in Figure 6.
In the crystal structures of the new CYP199A4–substrate
complexes, there are two (2-naphthoic acid) or four mole-
cules (the others) in each asymmetric unit, which can be
traced from residues 17 to 409 for each molecule. The re-
finement of the second molecule of 2-naphthoic acid in the
asymmetric unit was poor although it could be traced from
residue 17 to the C-terminal residue. This substrate yielded
a thin, plate-like crystal form, which may not have been a
perfect single crystal. In addition radiation damage from the
longer data collection time from the home source X-ray
generator potentially led to lower quality diffraction data.
Nevertheless, geometry and chemical restraints evaluations
indicated that the structural parameters were in acceptable
regions (Table S1 in the Supporting Information).
(ꢂ21) nmol(nmolCYP)ꢀ1 minꢀ1 (Figure S4). No product
could be detected by GC after derivatisation but the turn-
over mixtures had a distinct blue tinge. The product 2-indo-
linone-6-carboxylic acid has been reported previously (Fig-
ure 4).[13c] In an attempt to identify the products of indole-6-
carboxylic acid and 2-naphthoic acid we used an in vivo
HaPuR/HaPux/CYP199A4 expression system.[9] Addition of
the indole-6-carboxylic acid or 2-naphthoic acid to this
whole-cell oxidation system resulted in a change in the
colour of the media to purple and red, respectively. These
could be pigments formed from the coupling together of
two molecules of monooxygenase product as is observed
with indole and other aromatic compounds.[12c,15] Preparative
scale reactions will be required for purification of the oxida-
tion products of 4-isopropylbenzoic acid, trans-4-ethyl- and
trans-4-isopropylcyclohexanecarboxylic acids, l-perillic acid,
indole-6-carboxylic acid and 2-naphthoic acid for full prod-
uct characterisation.
Structure of CYP199A4 with bound aromatic carboxylic
acid substrates: In previous studies we solved the crystal
structures of substrate-free CYP199A2 and the 4-methoxy-
benzoic acid-bound forms of both CYP199A2 and
CYP199A4.[10b,14] Significant conformational changes and
residue side-chain movements in the substrate access chan-
nel were found in the transition from the open structure of
the substrate-free CYP199A2 to a closed conformation for
the substrate-bound form, and a chloride ion is bound near
the top of the substrate pocket and closes off the active site
to the external solvent. The substrate carboxylate group in-
teracts with polar and basic residue side chains whereas hy-
The overall structure of these substrate complexes of
CYP199A4 is very similar to that of the 4-methoxybenzoic
acid-bound form. All show the closed conformation and a
chloride ion is bound close to the surface, which shields the
active site from the external solvent (Figure S10 in the Sup-
porting Information). The tertiary structures of the enzyme
16682
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16677 – 16688

The Substrate Range of CYP199A4
FULL PAPER
in the different complexes are superimposable on one an-
other and the r.m.s.d. for all resolved Ca between all mole-
cules is <0.34 ꢀ. In each instance the crystal structure can
be used to rationalise the product selectivity of CYP199
family members with the different substrates. Chain A of
each complex was used for further comparison and analysis.
There are two closely related substrate binding modes,
one is found with 4-methoxy- and 4-ethylbenzoic acid, vera-
tric acid and indole-6-carboxylic acid for which the benzene
rings and the carboxylate groups are at virtually identical lo-
cations. The active site water molecule (Wat170 in PDB
code: 4DO1), which bridges the substrate carboxylate group
and the Arg243 found in the 4-methoxybenzoic acid-bound
structure, is retained in the veratric acid and indole-6-car-
boxylic acid-bound structures. There is not a strong enough
region of electron density to assign to a water molecule at
this position in the 4-ethylbenzoic acid structure; this is pos-
sibly due the lower resolution and higher B-factor of the
structure (Figure 6a–c). However, the distance between the
substrate carboxylate group and Arg243 is similar to that
found in the 4-methoxybenzoic acid-bound structure. The
second binding mode is observed with 2-naphthoic in which
the carboxylate group is displaced and as a result there is no
water molecule in the active site pocket (Figure 6d and Fig-
ure S11 in the Supporting Information).
4-Ethylbenzoic acid is structurally the most closely related
to 4-methoxybenzoic acid; the enzyme–substrate contacts
are very similar and for both structures the terminal carbon
(Cb) of the para-substituent is closest to the heme iron (Fig-
ure 6a and Figure S11). There is a small rotation about the
a
ipso
ꢀ
C
C
bond in the ethyl substituted compound, leading to
the Cb of the ethyl group being closer to and more directly
over the heme iron than the CH3 of the methoxy group (3.5
ꢀ
ꢀ
vs. 4.1 ꢀ and C Fe SACTHNUGRTENUNG(Cys358) angle=1648 vs. 1558,
Table S2 in the Supporting Information). Veratric acid is
bound with its benzene ring, the carboxylate group and 4-
methoxy substituent being superimposable with 4-methoxy-
benzoic acid (Figure 6b and Figure S11). The 4-methoxy
carbon is 3.9 ꢀ from the heme, consistent with vanillic acid
being the only oxidation product. The 3-methoxy substituent
points away from the heme where it makes contact with the
benzene ring of Phe182 and Phe185. The Phe182 side chain
is displaced slightly from its position in the 4-methoxybenzo-
ic acid complex to accommodate the additional methoxy
group. This movement also affects the position of the side
chains of Leu396 and Asp251 and these steric clashes pre-
sumably contribute to the weaker binding of this substrate
(Figure 6b and Figure S11). Remarkably indole-6-carboxylic
acid is bound in such an orientation that its carboxylic acid
group and the benzene ring are superimposable on their
counterparts in 4-methoxybenzoic acid. The indole NH is in
close contact with the meso carbon and the adjacent carbons
of the C- and D-pyrrole rings of the porphyrin, with strong
Figure 6. The active site of: a) 4-ethylbenzoic acid-bound, b) veratric
acid-bound, c) indole-6-carboxylic acid-bound, and d) 2-naphthoic acid-
bound CYP199A4. The 2mFoꢀDFc density of the substrates are shown
contoured at approximately 0.35 eꢀ3 (blue mesh). The substrates, the
heme and the active site residues are shown in yellow, grey and green, re-
spectively. The active site water (where present) equivalent to Wat170 in
4-methoxybenzoic acid-bound CYP199A4 is shown as a red sphere. In
the 4-ethylbenzoic acid-bound structure, the density is too weak to assign
a water molecule in this region. For clarity the residues are only labelled
in (a).
ꢀ
C H–p-like interactions (Figure 6c and Figure S11). The
indole C2 is closest to the heme iron (4.2 ꢀ) with the indole
nitrogen being the next nearest atom (4.4 ꢀ); this is consis-
tent with attack at this part of the molecule to form 2-indoli-
Chem. Eur. J. 2012, 18, 16677 – 16688
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16683

S. G. Bell, L.-L. Wong, W. Zhou et al.
none-6-carboxylic acid.[13c] The close proximity of the indole
nitrogen may contribute to the lower type I spin state shift
observed with this substrate.
in a largely hydrophobic groove on the enzyme surface be-
tween the F and I helices and the long loop linking the two
b4 strands (Figure S12a and b in the Supporting Informa-
tion). The location of this second molecule of the substrate
is equivalent to the groove on the surface where a camphor
molecule has been reported to bind to CYP101A1 and
CYP101D2.[17] These two substrates adopt slightly different
orientations. Taking indole-6-carboxylic acid as an example,
it contacts Glu153, Leu175, Pro176, and Gly179 in the
F helix, Asp251 in the I helix, and Arg391 in the b4 loop
(Figure S12a). The CYP199A4 complex with indole-6-car-
boxylic acid has a third substrate molecule bound close to
the entrance to the substrate access channel, contacting the
BC loop residues but also Lys89. This molecule is at the in-
terface between two protein molecules in the asymmetric
unit, and contacts with Arg142 and Gln141 of another
CYP199A4 molecule (Figure S12c). The locations of the
substrate molecules in relation to other molecules in the
asymmetric suggest that these binding sites may not be sig-
nificant in the mechanism and function of the enzyme.
Although 2-naphthoic acid is similar in size to indole-6-
carboxylic acid its binding mode is significantly different. In-
stead of the benzene ring bearing the carboxylate group it is
5
8
ꢀ
the C C ring that is superimposable on the benzene ring
of 4-methoxybenzoic acid (Figure 6d and Figure S11). The
carboxylate group is thus located about 1.9 ꢀ further to-
wards the N terminus of the I helix, and the ordered water
molecule (Wat170) found in the 4-methoxybenzoic acid-
bound form is displaced. As a result one of the carboxylate
oxygen atoms of 2-naphthoic acid forms a hydrogen bond
directly with Nh1 (3.0 ꢀ) of Arg243 rather than through a
water molecule (Wat170 in the 4-methoxybenzoic acid-
bound complex). However, the hydrogen bonds between the
other carboxylate oxygen atom and the side chains of Ser95
and Ser244 are maintained (Figure 6d). In order to maintain
the hydrogen bond interaction between Ser95 and the sub-
strate carboxylate the alcohol group of the serine side chain
5
ꢀ
moves by 1.4–2.3 ꢀ. In this binding mode the substrate C
C8 ring is further away from the heme iron; the closest
carbon atoms are C7 (4.7 ꢀ) and C8 and C6 (both 5.4 ꢀ), in
agreement with the observation of lower spin state shift on
substrate binding and the products reported by others
(though C8, 62% was the majority product over C7,
38%).[13a]
Analysis of the desaturation versus hydroxylation partition:
Desaturation between Ca and Cb of the para-substituent in
substituted benzoic acids could be initiated by hydrogen
atom abstraction at either carbon centre (Scheme 1).[10b,12a,b,18]
Hydrogen abstraction, from the less activated Cb, followed
by abstraction from Ca could occur. Alternatively electron
transfer from the b-radical to Compound II yields a b-carbo-
cation with a tendency to rearrange to the more highly sub-
stituted benzylic a-cation. Loss of a proton from the Cb of
The crystal structures of the CYP199A4–substrate com-
plexes account for the substrate specificity of the enzyme.
The requirement of a carboxylate group is explained by the
binding pocket for this group comprising the side chains of
Arg92, Arg243, Ser95 and Ser244; neutral alcohol and alde-
hyde groups, for example, 4-methoxybenzaldehyde and 4-
ethylbenzyl alcohol, do not form sufficiently strong hydro-
gen bonds to induce the conformational changes that occur
on substrate binding, for example, relocation of Arg243 and
cleavage of the Arg92–Glu99 ion pair. The side chains of
Leu98 and Ala248 appear to function as a vice clamp to
hold the substrate aromatic ring in position via strong van
der Waals interactions. The alicyclic rings of cyclohexanecar-
boxylic acids and l-perillic acid would not fit well between
these two key residue side chains, hence the low activity for
their oxidation. It can be envisaged that mutations at these
residues will relax the specificity for aromatic acids and po-
tentially alter the product selectivity. Moreover, mutation of
the cluster of three aromatic residues, Phe182, Phe185 and
Phe298, will enable the binding of larger substrates and
compounds with a larger substituent on the aromatic rings.
During the preparation of this manuscript the importance of
the Phe182 residue in substrate binding was demonstrated
by mutations at the equivalent residue in CYP199A2
(Phe185). The mutation of this residue to smaller non-aro-
matic residues modified the substrate binding mode of 2-
naphthoic acid and improved the activity with cinnamic acid
derivatives.[14,16]
ꢀ
this cation forms the alkene desaturation product. The C H
bond of Ca, which is more activated for hydrogen atom ab-
straction, could be the initial site of attack followed by a
second abstraction of a hydrogen atom from Cb (or a proton
if the Ca radical transfers an electron to the Compound II
intermediate, Scheme 1).
We observe higher proportions of the desaturation prod-
uct with CYP199A4 and CYP199A2 with 4-ethylbenzoic
acid compared to P450BM3 (CYP102A1) with ethylbenzene
where Ca hydroxylation dominates.[12c] DFT calculations
were performed to explore why the turnover of 4-ethylben-
zoic acid by CYP199A4 is more prone to desaturation than
the equivalent reaction of ethylbenzene with CYP102A1
(Table S3 and Figure S13 in the Supporting Information).
The relative stabilities of the different intermediates in the
desaturation and hydroxylation pathways of these substrates
were assessed. We found that the energy differences be-
tween the intermediates for 4-ethylbenzoic acid and 4-ethyl-
benzene oxidation were not large enough to infer that the
electronic effects of the carboxylic acid group is a major
factor in the different reactivity profiles of these two com-
pounds with CYP199A4 and CYP102A1 (Table S3 and Fig-
ure S13).
The additional branching in 4-isopropylbenzoic acid re-
sults in an increase in the proportion of the desaturation
pathway compared to 4-ethylbenzoic acid. Increased desatu-
ration percentages were also observed for the oxidation of
The indole-6-carboxylic acid and 2-naphthoic acid struc-
tures all contain a second molecule of the substrate located
16684
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16677 – 16688

The Substrate Range of CYP199A4
FULL PAPER
Scheme 1. a) The desaturation and hydroxylation pathways in the oxidation para-substituted benzoic acids. B) The heme species that are generated from
Compound I during the various desaturation and hydroxylation pathways.
substituted benzenes by wild-type (WT) CYP102A1, for ex-
ample, 8.9% for isopropylbenzene versus 0% for ethylben-
zene.[12c] It is tempting to conclude from these observations
and the absence of desaturation products in the oxidation of
trans-4-ethylcyclohexanecarboxylic acid by CYP199A4, that
the aromatic ring stabilises the intermediates of the desatu-
ration pathway and is required for this activity. However,
occur from Cb. Radical rebound could be slowed down by
a
rotation about the C
ipso bond, a radical rearrangement or
ꢀ
C
oxidation to the carbocation which undergoes rearrange-
ment followed by a second abstraction could ensue and lead
to desaturation. Alternatively, once the Cb has rotated away
from the heme a second abstraction can occur at Ca. By
comparison, CYP102A1 has a preference for hydrogen ab-
straction at subterminal locations for fatty acids, alkanes
and alkyl benzenes.[5b,20] Therefore, the proximity of the Cb
of 4-ethylbenzoic acid to the heme iron in the CYP199A4
structure may provide a more complete explanation of the
higher levels of desaturation observed in this system. By
contrast, the absence of desaturation products in 4-isoprop-
yl- and trans-4-ethylcyclohexanecarboxylic acid oxidation
may arise from these substrates being bound with their iso-
propyl and ethyl substituents located differently relative to
the heme iron. As noted in the analysis of the binding inter-
actions of the aromatic ring it would be of great interest to
examine the effect of mutations at the pair of vice clamp
residues Leu98 and Ala248 on substrate and product specif-
icity. Overall the structural data suggest that the positions of
Ca and Cb relative to the heme iron are significant in deter-
mining the relative amount of desaturation that is likely to
be observed for 4-ethylbenzoic acid and 4-ethylbenzene.
ꢀ
the first observations of C C bond desaturation were in the
oxidation of the aliphatic compounds valproic and lauric
acid.[12b,19] We propose that the major factors for the obser-
vation of desaturation is slower radical rebound to form the
normal, hydroxylation product and the accessibility of more
than one carbon centre to the ferryl, Compound I intermedi-
ate. The radical could then undergo rearrangement or elec-
tron transfer oxidation to form the carbocation that in turn
can rearrange. The inductive and resonance stabilisation ef-
fects of other parts of the substrate, saturated versus aro-
matic rings, tertiary versus secondary carbons, etc., will play
a role in these rearrangements and the relative populations
of intermediates that collapse to form the products. Radical
rebound would be slowed down if the substrate and/or the
group bearing the radical centre is mobile and can rapidly
adopt a different location away from the heme after initial
hydrogen atom abstraction. The partition between desatura-
tion and hydroxylation will depend on the accessibility and
ꢀ
reactivity of C H bonds compared to simple rebound.
The conversion of CYP199A4 into a 4-ethylbenzoic acid de-
saturase: We have previously shown that replacement of
Phe185 with valine or leucine residues results in a lowering
in the spin state shift, the affinity of binding and the activity
with 4-methoxybenzoic acid. These mutations of the rigid
phenyl side chain of phenylalanine with smaller and more
flexible alkyl side chains, such as isoleucine or valine, could
modify the position of the benzene ring or the para-substitu-
In the crystal structure of 4-ethylbenzoic acid-bound
CYP199A4 the Cb of the ethyl substituent is located over
the heme iron (3.5–3.7 ꢀ) with Ca being 1.3–1.5 ꢀ further
removed. Therefore, with the caveat that the position of the
substrate might be altered by oxygen binding and that the
a
b
ꢀ
C C H bond is more reactive than those of C , we would
expect a significant proportion of C H bond extraction to
ꢀ
Chem. Eur. J. 2012, 18, 16677 – 16688
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16685

S. G. Bell, L.-L. Wong, W. Zhou et al.
ent in the active site or allow increased substrate mobility
both of which could potentially alter the product selectivity.
If the location of Ca and Cb is altered relative to the heme
iron the level of desaturation may change. Therefore, we an-
alysed 4-ethylbenzoic acid binding and turnover with these
mutants.
As was observed with 4-methoxybenzoic acid the spin
shift and substrate binding affinities were lower with 4-ethyl-
benzoic acid and the Phe185Ile and Phe185Val mutants
compared to WT CYP199A4. This is consistent with the
substrate being positioned further away from the heme re-
sulting in weaker binding and allowing increased water
access to the ferric ion (Table 3). Both mutations also had a
Figure 7. Gas chromatography analysis of the turnover of 4-ethylbenzoic
acid (tR =5.6 min) with WT (black), Phe185Val (grey) and Phe185Ile
(dashed line) CYP199A4. TMS derivatised 4-vinylbenzoic acid is at tR =
5.7 min with the hydroxylated product at 7.3 min. For clarity the chroma-
tograms have been offset along the x and y axes.
Table 3. Substrate binding and catalytic parameters for CYP199A4 var-
iants with 4-ethyl- and 4-isopropylbenzoic acid.[a]
CYP199A4 HS
[%]
Kd [mm]
N[b]
PFR[c]
Coupling
[%][d]
4-ethylbenzoic acid
WT
Phe185Ile
Phe185Val
ꢁ95
0.34ꢂ0.02 1180ꢂ34 1030ꢂ52 88ꢂ2
70
10.9ꢂ0.4
7.0ꢂ0.3
19ꢂ2.0
421ꢂ28
3ꢂ0.1 16ꢂ2
In conclusion, the crystal structures of substrate-bound
forms of CYP199A4 have provided insight into the specifici-
ty of the CYP199 family for substituted benzoic acids and
rationalisation for observed product distributions. Studies on
80
393ꢂ30 93ꢂ1
4-isopropylbenzoic acid
WT
ꢁ95
0.29ꢂ0.1
1.2ꢂ0.1
321ꢂ24
20ꢂ1.0
287ꢂ18 89ꢂ2
3ꢂ0.5 13ꢂ2
ꢀ
the occurrence of C C bond desaturation led to the propos-
Phe185Ile
90
al that this unusual activity of P450 enzymes requires not
only for adjacent carbon centres to be close to the heme
iron but depends mainly on retardation of the radical re-
bound step. Although the factors involved are far from fully
understood it might be envisaged that the binding of a sub-
strate and the relative rates of the radical rebound and a
second hydrogen atom/proton abstraction might be manipu-
[a] The reaction mixtures (50 mm Tris-HCl, pH 7.4) contained P450
(0.5 mm), HaPux (5 mm) and HaPuR (0.5 mm). Rates are given as nmol(n-
molCYP)ꢀ1 minꢀ1
.
The data are given as mean ꢂS.D. with nꢁ3.
[b] NADH turnover rate; [c] product formation rate. [d] The percentage
of NADH consumed in the reaction that led to the formation of prod-
ucts.
ꢀ
deleterious effect on the activity of CYP199A4. The extra
carbon unit in the side chain of isoleucine over valine re-
duced the NADH consumption activity and the coupling
and hence product formation rate to a much greater degree
(Table 3).
lated to increasingly favour C C desaturation, leading to
novel strategies in functionalisation of chemically inert parts
of complex molecules.
Product analysis of the Phe185Val turnover demonstrated
an increased level of desaturation (68 vs. 54%). More dra-
matic was the effect of the Phe185Ile mutation, which shift-
ed the product selectivity and resulted in the generation of
100% of the desaturation product, 4-vinylbenzoic acid, with
no evidence of hydroxylated product (Figure 7). When the
same variant was used to turnover 4-isopropylbenzoic acid,
4-(prop-1-en-2-yl)benzoic acid was the only product ob-
served. The NADH consumption rate and coupling are also
reduced for 4-isopropylbenzoic acid (Table 3).
Experimental Section
General: General reagents, organic substrates and N,O-bis(trimethylsi-
lyl)trifluoroacetamide with trimethylchlorosilane (BSTFA/TMCS, 99:1)
were from Sigma–Aldrich, TCI or Merck (UK). Buffer components were
from Anachem (UK) or the Beijing Chemical Company (China). NADH
was from Roche Diagnostics (UK), and growth media and isopropyl-b-d-
thiogalactopyranoside (IPTG) were from Melford Laboratories (UK) or
Invitrogen (USA).
UV/Vis spectra and spectroscopic activity assays were recorded at 30-
(ꢂ0.5)8C on a Varian CARY-50 or CARY-1E spectrophotometer. Gas
G
chromatography analysis was performed on a ThermoFinnegan TRACE
instrument equipped with a flame ionisation detector (FID) and either a
DB-1 fused silica column (7 mꢂ0.32 mm) or a CP-SIL 8CB fused silica
column (15 mꢂ0.32 mm) by using helium as the carrier gas. Both the in-
jector and the FID were held at 2508C. GC-MS was performed on an
Agilent-6890 GC coupled to a Micromass GCT (Waters) by using a
HP6890 GC 15 m column with helium as the carrier gas and Cl+ as the
ionisation mode for the mass spectrometer.
The reduced activity and coupling of these Phe185 mu-
tants likely arose from increased substrate mobility and
modification of the substrate binding position relative to the
heme iron. These alterations also promoted the desaturation
a
b
ꢀ
pathway. It is likely that the C H bond of C and C are fur-
ther away from the heme iron in the mutants so the b-radi-
cal is at a far enough distance away from the heme to slow
down radical rebound and enable the desaturation pathway
to compete more effectively and indeed to dominate in the
Phe185Ile mutant.
Enzymes and molecular biology: General DNA and microbiological ex-
periments were carried out by standard methods.[21] The expression and
purification of CYP199A4, HaPux, and HaPuR have been described else-
where.[10a,b] Proteins were stored at ꢀ208C in Tris (pH 7.4, 50 mm) con-
16686
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16677 – 16688

The Substrate Range of CYP199A4
FULL PAPER
taining glycerol (50% v/v). Glycerol was removed immediately before
use by gel filtration on a 5 mL PD-10 column (GE Healthcare, UK) by
eluting with Tris (pH 7.4, 50 mm).
(100 mm) in ethanol to a final concentration of 1 mm. NADH was added
to about 320 mm (final A340 =2.00) and the absorbance at 340 nm was
monitored. The rate of NADH consumption was calculated by using
e340 =6.22 mmꢀ1 cmꢀ1
.
The production of CYP199A4 for protein crystallography was carried
out by using Escherichia coli BL21ACHTUNGTRNEUNG
(DE3) as described elsewhere.[14]
Substrate binding; spin state determination and binding titrations: The
high-spin heme content was estimated (to approximately ꢂ5%) by com-
parison with a set of spectra generated from the sum of the appropriate
percentages of the spectra of the substrate-free (>95% low-spin, Soret
maximum at 418 nm) and camphor-bound (>95% high-spin, Soret maxi-
mum at 392 nm) forms of WT CYP101A1.
Briefly the cells were grown at 378C in LB medium (1 L) containing ka-
namycin (50 mgmLꢀ1) to an OD600 of 0.6–0.8. Protein production was in-
duced with IPTG (0.5 mm) and the cells were grown for a further 20 h at
258C. The cell pellet obtained by centrifugation (4000 g) was resuspend-
ed in buffer A (20 mm HEPES, pH 7.4, 10 mm b-mercaptoethanol) and
lysed by sonication at 48C. The crude extracts were centrifuged at
27000 g for 30 min at 48C and the supernatant was loaded onto a Q Fast-
flow Sepharose column (GE Healthcare). The target protein was eluted
by using a linear salt gradient of KCl (0–1m) in buffer A, and the pooled
fractions were concentrated and then buffer-exchanged into buffer A.
Further purification was carried out on a Resource Q column (GE
Healthcare) by using a linear gradient of KCl (0–1m) in buffer A. Gel fil-
tration chromatography on a Superdex-75 column (GE Healthcare) was
used for further purification, eluting with HEPES (pH 7.4, 20 mm), KCl
(150 mm), b-mercaptoethanol (10 mm). The purity of the protein was esti-
mated to be greater than 95% by SDS-PAGE analysis. The CYP199A4
For binding constant determination the CYP199A4 mutants were diluted
to 0.5–2.0 mm by using Tris (pH 7.4, 50 mm) in a total volume of 2.5 mL,
and aliquots (0.5–2 mL) of the substrate were added with a Hamilton sy-
ringe from 1, 10 or 100 mm stock solutions in ethanol. The peak-to-
trough difference (DA) in absorbance between 700 and 250 nm was re-
corded. Further aliquots of substrate were added until the peak-to-trough
difference did not change. The dissociation constants, Kd, were obtained
by fitting DA against total substrate concentration [S] to a hyperbolic
function:
DAmaxx½Sꢃ
DA ¼
[14]
protein concentration was calculated by using e419 =119 mmꢀ1 cmꢀ1
.
Kd þ ½Sꢃ
Crystallisation: For co-crystallisation of CYP199A4 with 4-ethylbenzoic
acid, veratric acid, 2-naphthoic acid and indole-6-carboxylic acid, the pu-
rified protein was concentrated to approximately 40 mgmLꢀ1 in HEPES
(pH 7.4, 20 mm), KCl (150 mm), b-mercaptoethanol (10 mm) and a satu-
rated concentration of the substrate. The mixtures were kept on ice
before setting up the crystallisation trials. Crystals were obtained by
using the hanging-drop vapour-diffusion method at 208C with protein sol-
ution (1 mL) mixed with reservoir solution (1 mL) and equilibrated with
reservoir solution (200 mL). Crystal screening was carried out with Hamp-
ton Research Crystal Screen I, II and Index kits. Good quality crystals of
the complexes were obtained after 2–4 weeks from optimisation around
the conditions of Bis-Tris (pH 5.5, 0.1m), ammonium sulfate (1.45m),
sodium chloride (0.1m) by using different additives, such as NaF, NaI,
CoCl2 and 1,5-diaminopentane·2HCl. The final buffer conditions for each
set of crystals are detailed in Table S4 in the Supporting Information.
where DAmax is the maximum absorbance difference. Several substrates
exhibited tight binding, with Kd <1 mm. In these instances the data were
fitted to the tight binding quadratic equation:[28]
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
DA
DAmax
ð½Eꢃ þ ½Sꢃ þ KdÞ ꢀ fð½Eꢃ þ ½Sꢃ þ KdÞ ꢀ 4½Eꢃ½Sꢃg
2½Eꢃ
¼
where DAmax is the maximum absorbance difference and [E] is the
enzyme concentration.
Analysis of metabolites: After the NADH had been consumed in sub-
strate oxidation incubations, 990 mL of the reaction mixture was mixed
with 10 mL of an internal standard solution (25 mm 9-hydroxyfluorene in
ethanol) and 2 mL of concentrated HCl. The mixture was extracted three
times with ethyl acetate (400 mL) and the organic extracts were combined
and dried over MgSO4. Solvent was evaporated under a stream of dini-
trogen and the sample was dissolved in acetonitrile (200 mL). Excess
(25 mL) BSTFA/TMCS (99:1) was added and the mixture was left for at
least 120 min to produce the trimethylsilyl ester of the carboxylic acid
group and trimethylsilyl ether of the alcohol, if formed. The reaction
mixtures were used directly for GC analysis. The oven temperature was
held at 1008C for 1 min and then increased at 158Cminꢀ1 up to 2208C.
The retention times for the trimethylsilyl (TMS) derivatives are given in
the Supporting Information. Products were calibrated against derivatised
product samples of 4-vinylbenzoic acid, 4-hydroxybenzoic acid, vanillic
acid and 4-(hydroxymethyl)benzoic acid.[10b] When authentic samples
were not available the coupling was estimated based on the closest avail-
able compound from above. The same extractions of reaction mixtures
were also used for GC-MS analysis.
Data collection and structure determination: X-ray diffraction data for
CYP199A4 in complex with veratric acid were collected at ꢀ1738C on
beamline BL17U1 of the Shanghai Synchrotron Radiation Facility
(SSRF). The data for the CYP199A4 indole-6-carboxylic acid complex
were collected on beamline BL-17A of the Photon Factory (KEK) in
Japan. Diffraction data of CYP199A4 in complex with 4-ethylbenzoic
acid and 2-naphthoic acid were collected in-house on a Rigaku R-AXIS
HTC image plate by using CuKa radiation (l=1.5418 ꢀ) from a Rigaku
MicroMax-007 rotating anode X-ray generator operating at 40 kV and
30 mA. All crystals were cryoprotected by the addition of glycerol (20%
v/v) to the crystallisation solutions. All diffraction data were indexed, in-
tegrated and scaled with the HKL2000 package.[22] Complete data collec-
tion statistics are summarised in Table 2.
The structures of the complexes were solved by using the molecular re-
placement method with the program Phaser[23] in the CCP4 suite[24] with
the native structure of CYP199A4 (PDB code: 4DNZ) as the initial
search model. The substrates were manually built and adjusted under the
guidance of FoꢀFc difference maps by using the program COOT.[25] Struc-
tural refinements were carried out with Refmac5.[26] The stereochemical
quality of the refined structures were checked with the program MolPro-
bity.[27] A detailed summary of the refinement statistics is provided in
Table 2. The coordinates for the crystal structures of the complexes have
been deposited in the Protein Data Bank with the accession codes
4EGM, 4EGN, 4EGO and 4EGP for the 4-ethylbenzoic acid, veratric
acid, indole-6-carboxylic acid and 2-naphthoic acid-bound forms of
CYP199A4, respectively.
Computational studies on desaturation: All geometry optimisations were
performed under gas-phase conditions at 298 K by using the B3LYP[29]
density functional and the 6-31+GACHTUNGTRNEUNG
(d,p)[30] basis set on all atoms. Upon
attaining all optimised structures, harmonic vibrational frequency calcula-
tions at the same level of theory were subsequently performed to ascer-
tain whether they were local minima (0 imaginary frequencies) or local
transition state maxima (one imaginary frequency). Wave-function stabil-
ity tests (Stable=Opt) were also performed in order to ensure that there
were no instabilities in the optimised electronic wave-functions. All rela-
tive gas-phase energy values discussed and displayed are Gibbs free
energy corrected and are in kJmolꢀ1 unless otherwise specified. All cal-
culations were carried out by using the Gaussian09 suite of programs.[31]
Activity assays: NADH turnover rate assays were performed with mix-
tures (1.2 mL) containing Tris (pH 7.4, 50 mm), CYP199A4 (0.5 mm),
HaPux (5 mm), HaPuR (0.5 mm) and bovine liver catalase (100 mgmLꢀ1).
The buffer solution was oxygenated before use and the mixtures were
equilibrated at 308C for 2 min. Substrates were added as a stock solution
Chem. Eur. J. 2012, 18, 16677 – 16688
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16687

S. G. Bell, L.-L. Wong, W. Zhou et al.
C. Wiese, T. Fujishiro, C. Shirataki, B. Wunsch, Y. Watanabe, J. Biol.
Inorg. Chem. 2010, 15, 1109–1115.
Acknowledgements
[16] T. Furuya, Y. Arai, K. Kino, Appl. Environ. Microbiol. 2012, 78,
6087–6094.
[17] a) W. Yang, S. G. Bell, H. Wang, W. Zhou, M. Bartlam, L. L. Wong,
Z. Rao, Biochem. J. 2011, 433, 85–93; b) H. Yao, C. R. McCullough,
A. D. Costache, P. K. Pullela, D. S. Sem, Proteins Struct. Funct.
Bioinf. 2007, 69, 125–138.
[18] R. S. Obach, Drug Metab. Dispos. 2001, 29, 1599–1607.
[19] X. Guan, M. B. Fisher, D. H. Lang, Y. M. Zheng, D. R. Koop, A. E.
Rettie, Chem.-Biol. Interact. 1998, 110, 103–121.
[20] a) M. J. Cryle, R. D. Espinoza, S. J. Smith, N. J. Matovic, J. J.
De Voss, Chem. Commun. 2006, 2353–2355; b) M. J. Cryle, J. M.
Stuthe, P. R. Ortiz de Montellano, J. J. De Voss, Chem. Commun.
2004, 512–513; c) C. J. Whitehouse, S. G. Bell, H. G. Tufton, R. J.
Kenny, L. C. Ogilvie, L. L. Wong, Chem. Commun. 2008, 966–968;
d) Y. Miura, A. J. Fulco, J. Biol. Chem. 1974, 249, 1880–1888; e) Y.
Miura, A. J. Fulco, Biochim. Biophys. Acta Lipids Lipid Metab.
1975, 388, 305–317.
[21] J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular Cloning: A Labo-
ratory Manual, Cold Spring Harbor Laboratory Press, New York,
1989.
[22] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276, 307–326.
[23] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn,
L. C. Storoni, R. J. Read, J. Appl. Crystallogr. 2007, 40, 658–674.
[24] C. C. Project, Acta Crystallogr. Sect. D 1994, 50, 760–763.
[25] P. Emsley, K. Cowtan, Acta Crystallogr. Sect. D 2004, 60, 2126–2132.
[26] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. Sect.
D 1997, 53, 240–255.
This work was supported by the Ministry of Science and Technology of
China Project 973 grant 2010CB530100 (to W.Z.), and the Natural Sci-
ence Foundation of China grant 31170684 (to W.Z.).
[1] a) F. P. Guengerich, Chem. Res. Toxicol. 2008, 21, 70–83; b) P. R.
Ortiz de Montellano, in Cytochrome P450: Structure, Mechanism,
and Biochemistry, Kluwer Academic/Plenum, New York, 2005; c) A.
Sigel, H. Sigel, R. Sigel, in The Ubiquitous Roles of Cytochrome
P450 Proteins, Vol. 3, Wiley-VCH, Weinheim, 2007.
[2] a) J. Rittle, M. T. Green, Science 2010, 330, 933–937; b) P. R. Ortiz
de Montellano, Chem. Rev. 2010, 110, 932–948.
[3] a) E. M. Isin, F. P. Guengerich, Biochim. Biophys. Acta Gen. Subj.
2007, 1770, 314–329; b) M. J. Cryle, J. E. Stok, J. J. De Voss, Aust. J.
Chem. 2003, 56, 749–762.
[4] a) G. Grogan, Curr. Opin. Chem. Biol. 2011, 15, 241–248; b) D.
Monti, G. Ottolina, G. Carrea, S. Riva, Chem. Rev. 2011, 111, 4111–
4140; c) V. B. Urlacher, M. Girhard, Trends Biotechnol. 2012, 30,
26–36.
[5] a) S. G. Bell, N. Hoskins, C. J. C. Whitehouse, L. L. Wong, Design
and Engineering of Cytochrome P450 Systems, Wiley, New York,
2007, p.652; b) C. J. Whitehouse, S. G. Bell, L. L. Wong, Chem. Soc.
Rev. 2012, 41, 1218–1260.
[6] F. W. Larimer, P. Chain, L. Hauser, J. Lamerdin, S. Malfatti, L. Do,
M. L. Land, D. A. Pelletier, J. T. Beatty, A. S. Lang, F. R. Tabita,
J. L. Gibson, T. E. Hanson, C. Bobst, J. L. Torres y Torres, C. Peres,
F. H. Harrison, J. Gibson, C. S. Harwood, Nat. Biotechnol. 2004, 22,
55–61.
[7] Y. Oda, F. W. Larimer, P. S. Chain, S. Malfatti, M. V. Shin, L. M.
Vergez, L. Hauser, M. L. Land, S. Braatsch, J. T. Beatty, D. A. Pel-
letier, A. L. Schaefer, C. S. Harwood, Proc. Natl. Acad. Sci. USA
2008, 105, 18543–18548.
[27] V. B. Chen, W. B. Arendall 3rd, J. J. Headd, D. A. Keedy, R. M. Im-
mormino, G. J. Kapral, L. W. Murray, J. S. Richardson, D. C. Ri-
chardson, Acta Crystallogr. Sect. D 2010, 66, 12–21.
[28] J. W. Williams, J. F. Morrison, Methods Enzymol. 1979, 63, 437–467.
[29] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100; b) A. D. Becke, J.
Chem. Phys. 1993, 98, 5648–5652; c) C. Lee, W. Yang, R. G. Parr,
Phys. Rev. B Condens. Matter. 1988, 37, 785–789.
[8] F. Hannemann, A. Bichet, K. M. Ewen, R. Bernhardt, Biochim. Bio-
phys. Acta Gen. Subj. 2007, 1770, 330–344.
[30] a) T. Clark, J. Chandrasekhar, G. W. Spitznagel, P. V. Schleyer, J.
Comput. Chem. 1983, 4, 294–301; b) P. C. Hariharan, J. A. Pople,
Theor. Chim. Acta 1973, 28, 213–222; c) W. J. Hehre, R. Ditchfie,
J. A. Pople, J. Chem. Phys. 1972, 56, 2257; d) G. W. Spitznagel, T.
Clark, J. Chandrasekhar, P. V. Schleyer, J. Comput. Chem. 1982, 3,
363–371.
[9] S. G. Bell, A. B. Tan, E. O. Johnson, L. L. Wong, Mol. Biosyst. 2010,
6, 206–214.
[10] a) S. G. Bell, N. Hoskins, F. Xu, D. Caprotti, Z. Rao, L. L. Wong, Bi-
ochem. Biophys. Res. Commun. 2006, 342, 191–196; b) S. G. Bell, F.
Xu, I. Forward, M. Bartlam, Z. Rao, L.-L. Wong, J. Mol. Biol. 2008,
383, 561–574; c) S. G. Bell, F. Xu, E. O. Johnson, I. M. Forward, M.
Bartlam, Z. Rao, L. L. Wong, J. Biol. Inorg. Chem. 2010, 15, 315–
328; d) F. Xu, S. G. Bell, Y. Peng, E. O. Johnson, M. Bartlam, Z.
Rao, L. L. Wong, Proteins Struct. Funct. Bioinf. 2009, 77, 867–880.
[11] F. P. Guengerich, Chem. Res. Toxicol. 2001, 14, 611–650.
[12] a) C. A. Reilly, G. S. Yost, Drug Metab. Dispos. 2005, 33, 530–536;
b) A. E. Rettie, M. Boberg, A. W. Rettenmeier, T. A. Baillie, J. Biol.
Chem. 1988, 263, 13733–13738; c) C. J. Whitehouse, S. G. Bell, L. L.
Wong, Chem. Eur. J. 2008, 14, 10905–10908.
[13] a) T. Furuya, K. Kino, Biosci. Biotechnol. Biochem. 2009, 73, 2796–
2799; b) T. Furuya, K. Kino, ChemSusChem 2009, 2, 645–649; c) T.
Furuya, K. Kino, Appl. Microbiol. Biotechnol. 2010, 85, 1861–1868.
[14] S. G. Bell, W. Yang, A. B. Tan, R. Zhou, E. O. Johnson, A. Zhang,
W. Zhou, Z. Rao, L. L. Wong, Dalton Trans. 2012, 41, 8703–8714.
[15] a) S. G. Bell, C. F. Harford-Cross, L. L. Wong, Protein Eng. 2001, 14,
797–802; b) E. M. Gillam, A. M. Aguinaldo, L. M. Notley, D. Kim,
R. G. Mundkowski, A. A. Volkov, F. H. Arnold, P. Soucek, J. J.
De Voss, F. P. Guengerich, Biochem. Biophys. Res. Commun. 1999,
265, 469–472; c) E. M. Gillam, L. M. Notley, H. Cai, J. J. De Voss,
F. P. Guengerich, Biochemistry 2000, 39, 13817–13824; d) O. Shoji,
[31] Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V.
Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J.
Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A.
Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, ꢃ.
Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian,
Inc. Wallingford CT, 2009.
Received: August 2, 2012
Published online: November 7, 2012
16688
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16677 – 16688