Structure and biochemical properties of the alkene producing cytochrome P450 OleTJE

37 Pages·2014·2.13 MB·English

JBC Papers in Press. Published on January 18, 2014 as Manuscript M113.527325
The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M113.527325
Structure of an alkene producing P450
Structure and biochemical properties of the alkene producing cytochrome
P450 OleT (CYP152L1) from the Jeotgalicoccus sp. 8456 bacterium
JE

1James Belcher, 1Kirsty J. McLean, 1Sarah Matthews, 1Laura S. Woodward, 1Karl Fisher
1Stephen E. J. Rigby, 2David R. Nelson, 3Donna Potts, 3Michael T. Baynham, 4David A. Parker,
1David Leys and 1Andrew W. Munro*
1Manchester Institute of Biotechnology, Faculty of Life Sciences, University of Manchester, 131
Princess Street, Manchester M1 7DN, UK. 2Department of Microbiology, Immunology and
Biochemistry, University of Tennessee Health Science Center, 858 Madison Avenue Suite G01,
Memphis, TN 38163, USA. 3Agilent Technologies UK Ltd, Lakeside, Cheadle Royal Business Park,
Stockport, Cheshire SK8 3GR, UK. 4Westhollow Technology Center, 3333 Highway 6 South,
Houston, TX 77028-3101, USA.

Running title: Structure of an alkene producing P450
D
o
w
n
*To whom correspondence should be addressed: Andrew Munro Tel: [phone]; Fax: lo
a
d
[phone]; E-mail: [email protected] ed
fro
m
h
Bfaattcyk agcriodus ntod :p OroldeuTcJEe otexrimdaintiavle alylk deenceas.r boxylates aOdlveTanJEt aigne floorw p suarlitf icbautfifoenr, sciannc eb ree stoulurnbeildiz etdo ttp://w
w
Results: OleT is an efficient peroxide- OleT is fully active and extensively w
dependent lipiJdE decarboxylase, with high dissocJEiated from lipids. OleT binds avidly to .jbc
JE .o
affinity substrate-binding and the capacity to be a range of long chain fatty acids and rg
b/
resolubilized from precipitate in an active form. structures of both ligand-free and arachidic y
g
Conclusion: OleTJE has key differences in active acid-bound OleTJE reveal that the P450 active ues
site structure and substrate binding/mechanistic site is preformed for fatty acid binding. t o
n
properties to related CYP152 hydroxylases. OleT heme iron has an unusually positive A
JE p
Significance: OleTJE is an efficient and robust redox potential (-103 mV vs. NHE) which is ril 4
biocatalyst with applications in biofuel not significantly affected by substrate , 2
0
1
production. binding, despite extensive conversion of the 9
heme iron to a high-spin ferric state.
SUMMARY Terminal alkenes are produced from a range
The production of hydrocarbons in Nature of saturated fatty acids (C12-C20), and
has been documented for only a limited set of stopped-flow spectroscopy indicates a rapid
organisms, with many of the molecular reaction between peroxide and fatty acid-
components underpinning these processes bound OleT (167 s-1 at 200 µµµµM H O ).
JE 2 2
only recently identified. There is an obvious Surprisingly, the active site is highly similar
scope for application of these catalysts, and in structure to the related P450 , which
BSββββ
engineered variants thereof, in future catalyzes hydroxylation of fatty acids as
production of biofuels. Here we present opposed to decarboxylation. Our data
biochemical characterization and crystal provide new insights into structural and
structures of a cytochrome P450 fatty acid mechanistic properties of a robust P450 with
peroxygenase: the terminal alkene forming potential industrial applications.
OleT (CYP152L1) from Jeotgalicoccus sp.
JE The cytochromes P450 (P450s or CYPs) are
8456. OleT is stabilized at high ionic
JE oxidases that catalyze a vast array of oxidative
strength, but aggregation and precipitation of
reactions in nature (1). These hemoproteins are
1
Copyright 2014 by The American Society for Biochemistry and Molecular Biology, Inc.
Structure of an alkene producing P450
found in virtually all organisms from bacteria flavodoxins (5,11). However, other types of
and archaea through to man, and are responsible P450 redox partner systems exist (e.g. P450-
for several chemical transformations that are redox partner fusion enzymes, such as the
essential, for instance, in the microbial CYP116B family of P450:phthalate dioxygenase
biosynthesis of antibiotics (e.g. erythromycin in reductase fusions) (12). In addition, other P450s
Saccharopolyspora erythraea, and vancomycin catalyze isomerization (e.g. mammalian
in Amycolatopsis orientalis) (2,3), and in the thromboxane synthase, CYP5A1) and
mammalian formation of estrogens (estrone and dehydration (e.g. flax allene oxide synthase,
17β-estradiol) through the action of the CYP74A1) reactions that do not require an
aromatase P450 (CYP19A1) on androgen external source of electrons and which are
substrates (androstanedione and testosterone, completed entirely within the P450 active site
respectively) (4,5). The majority of (13,14). Further, through exploration of in vitro
characterized P450s are monooxygenases that routes to driving P450 catalysis, it is now well
interact with one or more redox partners to established that the addition of hydrogen
provide them with the two electrons (typically peroxide (H O ) or organic peroxides (e.g.
2 2
derived from NAD(P)H) required for oxidative cumene hydroperoxide) to P450s can facilitate
catalysis (6). The first electron reduces the P450 substrate oxidation by directly producing
cysteine thiolate-coordinated heme iron from compound 0, which is then protonated to D
o
ferric to ferrous, enabling dioxygen binding to generate compound I (15) (Figure 1). This w
n
the ferrous iron. The second electron reduces the “peroxide shunt” procedure is rarely an efficient loa
d
resulting ferric-superoxo complex to the ferric- means of driving P450s, since the peroxides ed
peroxo state. Two successive protonations oxidize heme and protein. However, a small fro
m
produce first the ferric-hydroperoxo species number of P450s that have evolved to exploit h
ttp
(compound 0) and then (following the loss of a the peroxide shunt are now known. Notably, the ://w
water molecule) the ferryl-oxo compound I (7) Bacillus subtilis CYP152A1 (P450 ) and the w
BSβ w
(Figure 1). The transient and highly reactive Sphingomonas paucimobilis CYP152B1 .jb
c
nature of compound I prevented its definitive (P450 ) naturally use H O to catalyze long .o
SPα 2 2 rg
characterization for many years, until Rittle and chain fatty acid hydroxylation, and are thus b/
y
Green produced compound I in large yield referred to as peroxygenases (16,17). P450SPα gue
following rapid mixing of CYP119 (from the catalyzes near-exclusively hydroxylation at the st o
thermoacidophilic crenarchaeon Sulfolobus alpha position, whereas P450BSβ catalyzes n A
achcildoroocpaeldrbaerniuzso)i c waciitdh, atnhde coonxfiidramnet d mits- hwyidthr otxhyel amtioanjo raitt ya lapt htah ea nbde tba eptao spitoisointi o(n~s6,0 :b4u0t pril 4, 2
identity using Mössbauer, EPR and UV-visible ratio) (16). 01
9
spectroscopy (8). Compound I is considered to
In recent studies, Rude et al. characterized a
be the major oxidizing species in the P450
novel enzyme from the bacterium
catalytic cycle, and to be responsible for the bulk
Jeotgalicoccus sp. ATCC 8456 (OleT ) that is
of oxidative reactions (e.g. hydroxylation, JE
41% identical in amino sequence to P450 , and
epoxidation, oxidative demethylation etc) BSβ
observed throughout the P450 superfamily (9, 37% identical to P450SPα. OleTJE was identified
10). as a P450 based on this sequence similarity, and
designated by the authors as a CYP152 P450
The vast majority of P450s use NAD(P)H-
family member (18). The Jeotgalicoccus ATCC
dependent redox systems consisting of either (i)
8456 host strain was shown to produce a number
an FAD-binding reductase that shuttles electrons
of C18-C20 linear and branched chain terminal
to the P450 via a ferredoxin (or a flavodoxin in a
alkenes, and other Jeotgalicoccus strains were
small number of cases), or (ii) an FAD- and
shown to generate a similar spectrum of terminal
FMN-binding cytochrome P450 reductase
alkenes in the C18-C21 range. A His-tagged
(CPR), the individual flavin-binding domains of
version of OleT was expressed in E. coli and
JE
which are evolutionarily related to NAD(P)H-
purified using Ni-NTA column chromatography,
binding ferredoxin oxidoreductases and
and shown to catalyze formation of n-1 alkenes
2
Structure of an alkene producing P450
through H O -dependent decarboxylation of
2 2
C14, C16, C18 and C20 saturated fatty acids
Expression and purification of OleT
(18).
The gene encoding OleT from
In view of the potential importance of the JE
Jeotgalicoccus sp. ATCC 8456 was codon
OleT enzyme as a producer of terminal alkenes
JE optimized (for expression in E. coli),
for exploitation in areas such as biofuels and
synthesized and cloned into the pET47b (Merck
fine chemical production, we have undertaken a
Millipore, Madison USA) vector by GenScript
study of the biochemical and biophysical
(New Jersey, USA). The E. coli strain C41
properties of the isolated OleT (CYP152L1)
JE (DE3) (Lucigen, Middleton USA) was used as
enzyme, and have determined its crystal
the expression host. Cells transformed with the
structure in complex with arachidic acid. These
pET47b-OleT plasmid were grown at 37°C
data reveal novel properties of this JE
with shaking at 200 rpm in total volumes of 500
biotechnologically important P450
ml to 3 l of 2YT broth containing kanamycin (30
peroxygenase. These include (i) OleT ’s high
JE µg/ml) supplemented with 500 µM δ-
catalytic efficiency and capacity to be
aminolevulinic acid. Expression of OleT was
resolubilized from a precipitated form as a fully JE
induced by addition of 100 µM IPTG when an
active enzyme; (ii) the extensive development of
optical density of 0.5 (at 600 nm) was reached, D
high-spin (HS) heme iron in OleT on binding o
JE at which point the incubation temperature was w
various long chain fatty acids (distinguishing it lowered to 25°C and the cells grown for a nloa
from related bacterial peroxygenases); and (iii) d
further 16 h. Cells were harvested by ed
its unusually positive heme iron reduction centrifugation at 6000 rpm, 4°C using a JLA- fro
potential, which is also negligibly affected by m
8.1000 rotor in an Avanti J-26 XP centrifuge. h
fatty acid binding despite the substrate inducing ttp
extensive HS ferric heme iron formation. Pellets were resuspended in a minimal volume ://w
of ice cold buffer A (100 mM potassium w
w
phosphate [KPi], pH 8.0), combined and .jb
c
centrifuged as before. The cell pellet was then .o
Experimental Procedures rg
frozen at -80°C until required. b/
y
Bioinformatics g
Cells were thawed at 4°C and resuspended in 3 u
e
s
The OleTJE sequence and additional members volumes of extraction buffer per gram of cell t on
of the CYP152 family, including all known pellet. The extraction buffer consisted of buffer A
p
subfamilies, were BLAST searched against a set A containing 1 M NaCl, 20% glycerol, with a ril 4
of all the prokaryotic P450 sequences. Members CompleteTM EDTA-free protease inhibitor , 2
0
of the highest scoring CYP families from these cocktail tablet (Roche, Mannheim Germany) per 19
searches were used to build a tree. Sequence 50 ml of cell suspension, DNase I (100 µg/ml,
alignments were computed using ClustalW and bovine pancreas, Sigma-Aldrich, Poole UK) and
checked manually for consistent alignment of lysozyme (100 µg/ml, hen egg white, Sigma-
known CYP motifs. Neighbor-joining trees were Aldrich). The cells were disrupted by two passes
generated with the Phylip package (Felsenstein, through a French Press (Thermo Scientific,
J. [2005], PHYLIP – Phylogeny Inference Hemel Hempstead UK), and the homogenate
Package version 3.6, distributed by the author,
centrifuged at 20,000 rpm, 4°C for 90 min using
Department of Genome Sciences, University of a JA-25.50 rotor. Alternatively, cells were lysed
Washington, Seattle) using ProtDist (a program by sonication using a Bandelin Sonopuls
in Phylip) to compute difference matrices. Trees sonicator set to 45% amplitude with 30 x 30 s
were drawn and colored with FigTree version pulses, at 60 s intervals, with the cell suspension
1.3.1 (http://tree.bio.ed.ac.uk/software/figtree/)
kept on ice throughout. The homogenate was
and labeled in Adobe Illustrator CS version
then centrifuged as previously. The supernatant
11.0.0 (Adobe Systems Incorporated). was removed and the pH re-set to 8.0 as
Sequences producing long branches on the tree necessary. The sample was then incubated
were removed and the tree was recomputed. overnight with 10 ml (per 100 g cell pellet) of
3
Structure of an alkene producing P450
Ni-IDA chromatographic medium (Generon, was concentrated to >20 mg/ml using a Vivaspin
Maidenhead UK) on a rolling table at 4°C. The centrifugal concentrator (Generon), snap frozen
mixture was then poured into a column and the in liquid nitrogen and stored at -80°C.
collected bed of OleT -bound medium was
JE
washed with 10 column volumes (CV) of 100
mM KPi (pH 8.0) containing 750 mM NaCl, UV-visible spectroscopy
20% glycerol (buffer B) and 50 mM imidazole
Analysis of the UV-visible spectroscopic
to remove weakly bound contaminants. The
properties of OleT was done on a Cary 60 UV-
column was then washed with 2 CV of the JE
visible spectrophotometer (Varian UK). Spectra
buffer B containing 125 mM imidazole,
were recorded using ~4-10 µM OleT in 100
followed by 5 CV of buffer B plus 150 mM JE
mM KPi (pH 8.0) plus 750 mM NaCl (buffer
imidazole, which eluted the bulk of the OleT
JE D). Reduction of OleT was achieved by
protein. The partially purified OleT sample JE
JE addition of sodium dithionite to enzyme in
was dialyzed overnight against 15 l of buffer A
buffer D made anerobic by extensive bubbling
at 4°C, which caused OleT to precipitate. Post-
JE with oxygen-free nitrogen. The ferrous-CO
dialysis, precipitated protein was isolated by
complex of OleT was formed by slow bubbling
centrifugation at 4000 rpm, 4°C using an A-4-62 JE
of gas into anerobically reduced enzyme until no
rotor in an Eppendorf 5810 R centrifuge. The D
further absorbance change occurred. The NO o
pellet was washed gently with 50 ml of buffer A w
and centrifugation repeated. OleTJE was coof mNpOle ixn two aas sfaomrmpleed obfy f eardrdici tOiolne Tof 5i-n8 a bnuebrobbleics nload
resuspended in 5 ml of buffer A containing 1 M JE ed
NaCl and 10% glycerol, which produced OleT buffer. fro
JE m
at high purity (Method 1). For OleT destined h
JE ttp
for crystallographic studies, HRV 3C protease Fatty acid and inhibitor binding titrations with ://w
(Merck Millipore, Darmstadt Germany) was w
OleT w
incubated with OleTJE for ~16 h at 4°C (50:1 µg JE .jb
c
protein/U protease) to remove the N-terminal Spectral binding titrations of OleTJE with .org
poly-histidine tag. The proteolysed protein was saturated fatty acids (C12, C14, C16, C18 and b/
y
applied to 5 ml of pre-equilibrated Ni-Sepharose C20) were performed at 25°C in buffer D. Fatty g
u
resin (GE Healthcare, Little Chalfont UK) to acids were from Sigma-Aldrich. Substrates es
t o
bind the cleaved His-tag and the tagged HRV (typically 0.25 mg/ml) were dissolved in 70% n
A
3coCl.u mTnh eb yc lweaavsehdin gO wleiTthJE 1w00a sm eMlu tKedP i f(rpoHm 8t.h0e) (thve/v )s oEdtOiuHm (fsoarlt sC 1o8f, CC2102), oCr 1740 %an Md eCO1H6 (fwaittthy pril 4
, 2
plus 750 mM NaCl and 10% glycerol (buffer C). acids) and 30% (v/v) Triton X-100 (Sigma- 0
1
9
Aldrich). A parallel set of binding titrations was
In separate preparations (avoiding the OleT
JE also performed using fatty acids (1 mg/ml)
precipitation step, Method 2) the dialysis step
dissolved in 100% EtOH or MeOH without
post Ni-NDA chromatography was removed and
Triton X-100. Prior to titrations, OleT samples
the OleT eluate was instead diluted (5x) in JE
JE (~50 µl at >20 mg/ml) in buffer C were passed
buffer C and concentrated in an Amicon
through a Lipidex column of dimensions 5 x 1
ultrafiltration device. The OleT sample was
JE cm (Perkin Elmer, Cambridge UK) in order to
then centrifuged to clarify the sample (16000
remove any residual lipid retained during
rpm, 4°C using the JA-25.50 rotor) and the
purification of the protein from E. coli. OleT
supernatant was then applied again to a 5 ml Ni- JE
recovered from the column was in an
IDA column. The column was washed with 5
extensively low-spin (LS) ferric state, and was
CV of buffer C containing 50 mM imidazole and
used directly for titration at a final P450
then 10 CV of buffer C plus 100 mM imidazole.
concentration in the range from 5-10 µM.
The His-tagged OleT was then eluted with 150
JE Titrations were performed by stepwise additions
mM imidazole in the same buffer. All
of aliquots (0.1-1 µl) of the fatty acids to the
procedures generated highly purified OleT
JE
OleT sample (substrate additions to < 1% of
protein. In both cases, the pure OleT protein JE
JE
total volume). Spectra (800-300 nm) were
4
Structure of an alkene producing P450
recorded for the ligand-free OleT and Photophysics). The observed reaction rate
JE
following each addition of substrate using a constants (k values) were plotted versus the
obs
Cary 60 UV-visible spectrophotometer. relevant H O concentrations, and the resultant
2 2
Difference spectra at each stage in the titration data plot fitted using a linear function to obtain
were computed by subtracting the spectrum of the 2nd order rate constant reporting on H O -
2 2
ligand-free OleT from each successive fatty dependent decarboxylation of substrate and the
JE
acid-bound spectrum collected during the consequent heme iron spin-state conversion.
titration. A pair of wavelengths were identified Entire spectral acquisition (750-280 nm) was
that defined the absorbance maximum (A ) also done using the PDA detector for the same
peak
and minimum (A ) in the difference spectra set of stopped-flow reactions analyzed in single
trough
from each titration set. The overall absorbance wavelength mode.
change (A ) at each substrate concentration
max
point was calculated as A minus A , and
peak trough
A was plotted versus [substrate]. These data Redox potentiometry
max
were fitted using either a hyperbolic (Michaelis-
To determine the midpoint potential for the
Menten) function, the Morrison equation for
OleT Fe3+/Fe2+ couple, redox titrations were
tight binding ligands, or the Hill function (where JE
performed at 25°C in an anerobic glove-box
sigmoidal behaviour was observed) in order to D
(Belle Technology, Weymouth UK) under a o
determine dissociation constants (K values), as w
d nitrogen atmosphere with O levels maintained n
described previously (19,20). Titrations and data at less than 2 ppm. A2ll solutions were load
fitting for OleT with dithiothreitol (DTT), ed
imidazole and cyJaEnide (sodium salt) inhibitors deoxygenated by sparging with nitrogen gas. For fro
substrate-free OleT , the titration was done m
were done in the same way as for the fatty acids, JE h
with ligands dissolved in buffer D. pulsuins g1 09.%3 µglMyc eOrolelT. JFE oirn s1u0b0st rmatMe- bKouPni d( pOHl e7T.0), ttp://w
JE w
the titration was done under the same conditions, w
following addition of arachidic acid (from a 32 .jbc
Stopped-flow analysis of substrate turnover .o
mM stock in 80% EtOH, 20% Triton X-100) rg
Stopped-flow absorption measurements were until no further conversion of the heme iron to by/
g
made using an Applied Photophysics SX18 MR the HS heme state was observed (ca 12 µM ue
s
stopped-flow spectrophotometer (Leatherhead, arachidic acid). Mediators were added to t o
n
UK). Stopped-flow spectral accumulation was expedite electronic equilibration in the system (2 A
p
done using a photodiode array (PDA) detector µM phenazine methosulfate, 7 µM 2-hydroxy- ril 4
on the same instrument. Fatty acid substrate- 1,4-naphthoquinone, 0.3 µM methyl viologen , 2
0
bound OleTJE was mixed versus different and 1 µM benzyl viologen to mediate in the 19
concentrations of H O in 100 mM KPi (pH 8.0)
2 2 range from +100 to -480 mV versus NHE) and
containing 750 mM NaCl at 25°C. OleT (9.2
JE data fitting (using the Nernst equation) and
µM) was converted to an extensively HS heme
analysis was done as described in previous
iron form by mixing with arachidic acid (12 µM) publications (21-23).
from a concentrated stock prepared in 80%

EtOH/20% Triton X-100. Reactions were
initiated by mixing the arachidic acid bound EPR analysis of OleT
JE
OleT (4.6 µM final concentration) with H O
JE 2 2 Continuous wave X-band electron
(3.29 – 200 µM final concentration). Stopped-
paramagnetic resonance EPR spectra of OleT
flow traces at single wavelengths reporting on JE
were obtained at 10 K using a Bruker ELEXSYS
the conversion of HS OleT heme iron towards
JE E500 EPR spectrometer equipped with an ER
LS were collected over periods of up to 30 s to
4122SHQ Super High Q cavity. Temperature
follow depletion of HS (390 nm) and formation
control was effected using an Oxford
of LS heme iron (418 nm). Data were analyzed
Instruments ESR900 cryostat connected to an
and fitted using a single exponential function
ITC 503 temperature controller. Microwave
with the Pro-Data SX software suite (Applied
power was 0.5 mW, modulation frequency was
5
Structure of an alkene producing P450
100 KHz and the modulation amplitude was 5 G. Analysis of products formed by OleT in
JE
EPR spectra were collected for OleT (305 µM) reactions with H O and fatty acids
JE 2 2
in the substrate-free form, and for OleT (205
JE OleT reactions with long chain saturated
JE
µM) bound to arachidic (C20:0) acid (at a
fatty acids (C12 to C20) were set up as follows.
saturating concentration).
5 ml reactions were done in buffer D, with 250
µM dodecanoic acid (sodium salt), palmitic acid
or arachidic acid, 500 µM hydrogen peroxide
Crystallography of OleT
JE and 0.6 µM OleT . The final reaction mixtures
JE
Crystallization trials for OleT were were incubated for periods up to 30 minutes at
JE
performed using 400 nl (200 nl protein plus 200 room temperature. 1 ml of the reaction mixture
nl precipitant) sitting drops in Art Robbins 96- was then extracted (at different reaction times)
well plates, using Molecular Dimensions 96- with an equal volume of HPLC-grade heptane,
deep well crystallisation screens [Clear Strategy and the sample centrifuged at 14000 rpm for 20
Screen I (CSS1), Clear Strategy Screen II, minutes. The top layer was then analyzed by
PACT premier, JCSG-plus and Morpheus] and a GC/MS. Analysis was done using a Thermo
Mosquito nanoliter pipetting robot (TTP Fisher DSQ II GC/MS instrument with a 30 m x
Labtech, Melbourn UK). Crystals formed 0.25 mm x 0.25 µm ZB5MS GC column
D
between 2 days and 1 month at 4°C in several (Phenomenex). Injection was cold on-column. ow
conditions. The crystals giving best diffraction The oven program was set so that an initial nlo
a
were formed under the following conditions: 35 temperature of 50°C was ramped at 10°C/min to de
d
mg/ml OleTJE in 0.1 M Tris (pH 8.5) containing 300°C post-injection. Electronic ionization was fro
0.2 M MgCl2 and 25% (w/v) polyethylene used, and ions in the range of 40-640 m/z hm
glycol 2K monomethyl ether (substrate-free scanned at two scans per second. ttp
OleT ); and 43 mg/ml OleT incubated with ://w
JE JE w
235 µM arachidic acid in 0.1 M Tris (pH 8.5) w
.jb
containing 0.2 M MgCl , 10% (w/v) Results c
2 .o
polyethylene glycol 8K and 10% (w/v) rg
polyethylene glycol 1K (substrate-bound Classification of OleTJE as CYP152L1 by/
g
OleTJE). There are currently 21,039 named cytochrome ues
P450 sequences t o
For preparation of substrate-bound OleT , n
JE (drnelson.uthsc.edu/P450.stats.Aug2013.png). A
Pul4t5ra0f iltrastaiomnp laensd a wsteorcek solcuotniocne notrfa taerda chidbiyc Approximately 6% are bacterial (1254 pril 4
sequences) and an additional 48 are from , 2
acid (32 mM) dissolved in 100% EtOH was 0
1
archaea. Initial BLAST searches with OleT 9
added to a final concentration of 235 µM. The JE
showed that it was less than 40% identical to
concentration of EtOH did not exceed 1% of the
most known CYP152 sequences and barely over
total volume. The mother liquor was
the 40% recommended cutoff for CYP family
supplemented with 10% PEG 200 where an
membership to two CYP152 sequences (41% to
additional cryo-protectant was required and
CYP152A1 from Bacillus subtilis and 40% to
crystals were flash-cooled in liquid nitrogen
CYP152A2 from Clostridium acetobutylicum).
prior to data collection. Data were collected at
The location of the OleT sequence in a
Diamond synchrotron beamlines and reduced JE
phylogenetic tree (as CYP152L1) strongly
and scaled using XDS (24). Structures were
argues for inclusion in the distinct CYP152
solved by molecular replacement with the
clade. The same logic applies to the renamed
previously solved P450 BS crystal structure
β CYP152M1 from Enterococcus faecium that has
(PDB 2ZQJ) using PHASER (25). Structures
a long branch in the tree. This sequence was
were refined using Refmac5 (25) and Coot (26).
previously named CYP241A1, but that
Final refinement statistics are given in Table 1.
nomenclature has been changed based on its
inclusion within the CYP152 clade. A second
sequence, CYP152L2 from Staphylococcus
6
Structure of an alkene producing P450
massiliensis S46, is 64% identical to CYP152L1 equilibrated in buffer C. By washing the column
(Figure 2). with increasing concentrations of imidazole in
buffer C, His-tagged OleT was eluted at 150
JE
mM imidazole in a highly pure form
Expression and purification of OleT (purification Method 2). A typical yield of
JE
purified OleT was ~20 mg per liter of E. coli
The OleT gene was codon optimized for JE
JE cell culture using either Method 1 or Method 2
expression in E. coli, and preliminary studies
for protein purification.
revealed that the enzyme was expressed well in
a number of E. coli strains. The C41 (DE3)
strain (Lucigen) was selected for protein
UV-visible absorption properties of OleT
production with the gene cloned into pET47b via JE
the BamHI and EcoRI restriction sites with a 6- Rude et al. inferred the cytochrome P450
His N-terminal tag, and transcribed using the nature of OleT from amino acid sequence
JE
T7-lac RNA polymerase/promoter system. similarities to peroxygenase members of the
Expression cell extracts were red in color, CYP152 family of P450s, and demonstrated in
indicative of the production of a heme protein. vitro that cell extracts of Jeogalicoccus sp.
However, our initial studies revealed that the ATCC 8456 could decarboxylate the saturated
D
OleT protein precipitated on dialysis following fatty acids arachidic acid (C20) and stearic acid o
JE w
elution from a Ni-IDA protein in the first (C18) to their respective n-1 terminal alkenes (1- n
lo
chromatographic purification step. Previous nonadecene and 1-heptadecene, respectively). A ad
e
d
studies by Rude et al. used high salt (NaCl) His-tagged OleTJE isolated from E. coli was also fro
concentration in several purification buffers shown to catalyze stearic acid decarboxylation m
h
(n1a8tu),r ea onfd thien hvoisetw b aocft ertihuims a(Jnedo ttghael ichoaclcoupsh islpic. iUnV a-v iHsi2bOle2 -adbespoernpdteionnt freeaatcutrieosn ty(p1i8c)a.l oHfo aw Pev4e5r0, ttp://w
w
ATCC 8456) we considered that the protein enzyme were not presented in this earlier study. w
might be stabilized in solution at high ionic .jbc
Figure 4 shows characteristic absorption .o
strength. This proved to be the case, and it was rg
found that the precipitation of OleTJE could be sFpee3+c)t raa nfdo r spoudrieu mO ledTitJhEi oinni tiet-sr eodxuicdeizde d( fe(frerorruics,, by g/
used to advantage, since resolubilization of the u
Fe2+) forms; and for the ferrous-carbon es
centrifuged protein pellet in buffer A containing monoxide (Fe2+-CO) and ferric-nitric oxide t on
1 M NaCl and 10% glycerol produced an OleT A
JE (Fe3+-NO) species. The resting (ferric) form of p
sample with a P450-like heme spectrum (Amax at OleT shows a heme spectrum typical of a P450 ril 4
~418 nm). SDS-PAGE at this stage also JE , 2
enzyme with its ferric heme iron in a LS state. 0
indicated the protein to be extensively purified 19
The major absorption feature (the Soret band) is
(purification Method 1). Specifically for
at 418 nm, with the smaller alpha and beta bands
crystallization, the OleT His-tag was removed
JE in the visible region at ~566 nm and 535 nm,
by incubation with HRV 3C protease, and the
respectively. These values are similar to those of
mixture loaded onto a Ni-Sepharose column.
other LS bacterial P450s (e.g. the Bacillus
Washing the column in buffer C (100 mM KPi
megaterium P450 BM3 [CYP102A1] heme
(pH 8.0) plus 750 mM NaCl and 10% glycerol)
domain with maxima at 418, 534 and 568 nm;
resulted in elution of a highly purified tag-free
and the Mycobacterium tuberculosis CYP121A1
OleT protein (Figure 3), and the retention of
JE at 416.5, 538 and 568 nm) (27,28). The two
the cleaved His-tag and the tagged protease on
methods of preparing OleT (i.e. with or
the column. JE
without a protein precipitation step) produced
Having identified the issues with propensity of identical oxidized OleT spectra. Any residual
JE
OleT to aggregate at low ionic strength, an imidazole ligand from nickel column
JE
alternative strategy was developed to avoid its chromatography (in both cases) was extensively
precipitation – by eluting OleT from Ni-IDA in depleted by ultrafiltration used to concentrate
JE
the high salt buffer C, centrifuging the sample the proteins and thus did not produce any
and then re-applying to Ni-IDA resin imidazole-ligated OleT heme iron.
JE
7
Structure of an alkene producing P450
Reduction of OleT with sodium dithionite heme iron developed were improved in all cases
JE
produced a ferrous hemoprotein with the Soret in presence of the detergent, although Triton X-
band diminished in intensity and shifted to 414 100 alone induces no spin-state change (e.g.
nm. In the visible (heme Q-band) region, a 0.67 ± 0.03 µM versus 6.20 ± 0.26 µM for
single, slightly asymmetric feature is seen at palmitic acid). The extent of spin-state change
~540 nm. The blue shift of the Soret spectrum induced varied according to chain length, with
on reduction indicates substantial retention of the longer chain fatty acids (C18:0 and C20:0)
cysteine thiolate proximal coordination in the inducing a more complete conversion to the HS
OleT ferrous state, and the spectral maxima are ferric state than observed for the C12:0 to C16:0
JE
similar to those features seen for e.g. the well fatty acids). For a titration using an arachidic
characterized Pseudomonas putida camphor acid (C20:0) stock including Triton X-100, the
hydroxylase P450cam (CYP101A1, 411 and 540 HS conversion was almost complete (estimated
nm) and for the explosive degrading P450 XplA at ≥95%), as shown in Figure 5A. In contrast,
from Rhodococcus rhodochrous strain 11Y lauric acid (C12:0) produced ~52% HS at
(CYP177A1, 408 and 542 nm) (23,29). Addition saturation (Figure 5B). For studies with non-
of carbon monoxide to anerobically reduced precipitated OleT (prepared using Method 2) in
JE
OleT produced a characteristic P450 heme the presence of Triton X-100, tight binding of
JE
spectrum with the Soret band red-shifted to 449 fatty acids was again observed (e.g. K values of D
d o
nm and a Q-band feature at 551 nm. A small 1.54 ± 0.19 µM for arachidic acid and 12.7 ± 0.3 wn
shoulder on the Soret feature at ~423 nm likely µM for lauric acid) (Table 2). However, the K loa
i(nlidkieclayt ecsy as tmeinineo rt hpirool-pcoorotirodnin (a~te5d%) ) foofr mth eo fP 4t2h0e vmaalugnesit uidnec rfeoars ea llb fya tatyp parcoixdism teastetelyd caonm opradreerd otofd ded from
OleTJE Fe2+-CO complex. The NO-bound ferric those for OleTJE prepared by Method 1 (Table http
OleTJE spectrum is also typical of other P450- 2). Thus, contrary to what may have been ://w
NO adducts, with an asymmetric Soret feature expected, the resolubilized OleT shows higher w
JE w
(~427 nm) and distinctive, enhanced intensity affinity than the non-precipitated form for the .jb
c
alpha and beta bands at ~573 and 540 nm (30). panel of fatty acid substrates tested. .org
Using the method of Berry and Trumpower, an b/
extinction coefficient of ε = 91.5 mM-1 cm-1 Binding of cyanide and imidazole to OleTJE y g
418 produced typical type II P450 heme absorption ue
was established for the LS ferric form of OleTJE shifts to longer wavelength. Soret shifts to 433 st o
(31,32). n A
nµmM )( Kwde r>e1 0o bmseMrv)e da nfdo r4 2cy4a nnmid e( Kand d= im19id3a ±zo 1le1, pril 4
Analysis of substrate and inhibitor binding to respectively. The binding of DTT to OleTJE was , 201
OleT also analyzed in view of the report from Rude et 9
JE
al., which indicated that DTT could support
The binding of substrates to P450s is often
OleT fatty acid decarboxylase activity by
JE
associated with alteration of the spin-state of
producing H O under aerobic conditions in the
2 2
their ferric heme iron, usually through displacing
presence of the P450 heme iron (18,35).
its weakly bound 6th ligand water molecule and
However, in previous studies we showed that
inducing a shift towards the HS form (e.g.
DTT coordinated the heme iron in the explosive
33,34). For OleT , we investigated the binding
JE degrading XplA P450 (23). DTT is known to
of a series of saturated fatty acids (C12-C20),
bind P450 heme iron and ligation is feasible in
and found that in all cases the lipids induced a
both DTT thiol and thiolate forms (36,37).
LS to HS transition, with the Soret band shifting
Figure 5C shows data from a spectral titration of
from 418 nm towards 394 nm. Table 2 shows
OleT with DTT in buffer D. The DTT-bound
JE
fatty acid binding K data for OleT (purified
d JE form has three distinct absorption features in the
using Method 1) and using fatty acid stocks
Soret region, with peaks at 372 nm and 423 nm,
dissolved in alcohol, or in alcohol containing
and a strong absorbance shoulder at ~460 nm.
30% v/v Triton X-100 (see Experimental
The central band is the most intense. The 423
Procedures). The K values and the extent of HS
d nm peak arises from distal ligation of DTT thiol
8
Structure of an alkene producing P450
to OleT heme iron, whereas the outer peaks reactive iron-oxo species (initially the ferric-
JE
result from a split (hyperporphyrin) Soret hydroperoxo compound 0, which is likely
spectrum in which DTT thiolate ligates the iron transformed to the ferryl-oxo compound I), and
(36,37). Comparable spectral maxima are at 374, its positive potential is likely a consequence of
423.5 and 453.5 nm for XplA (23). In XplA, the the environment of the heme and its cysteine
intensities of the three absorbance bands are thiolate ligand. The fact that the OleT heme
JE
quite similar, but in OleT the outer bands are potential is effectively unchanged in the HS
JE
much weaker than the 423 nm feature, substrate-bound form may be a consequence of
suggesting that DTT favors heme ligation in the the proximity of a negatively charged substrate
thiol state under the conditions used. The Figure carboxylate group to the heme iron in the
5C inset shows fitting of DTT-induced heme arachidic acid bound form.
absorption change for OleT , leading to a K of
JE d Another notable feature in the spectra for the
159 ± 7 µM. In the Rude et al. study, DTT at
reduced forms of substrate-free and arachidic
200 µM was used to support OleT catalysis
JE acid-bound OleT is that neither form a unique
JE
(18). However, our data indicate that substantial
spectral species that could be assigned to a
inhibition of OleT likely occurs under such
JE cysteine thiolate-coordinated ferrous P450 heme
conditions.
iron. As shown in Figure 6, the UV-visible
D
spectrum for OleTJE immediately following ow
Determination of the heme iron redox potentials reduction has its Soret feature at 414 nm, with a nloa
small shoulder at ~423 nm – indicative of a d
e
of substrate-free and substrate-bound OleT d
JE mixture of Cys thiolate-coordinated (major fro
Fatty acid binding to OleT induces species) and thiol-coordinated (minor species) m
JE h
substantial shifts in heme iron spin-state forms. In the redox titration for substrate-free ttp
equilibrium towards HS (e.g. Figure 5A), and OleT (Figure 6A) the Soret peak for the ://w
JE w
such shifts in spin-state equilibrium are often reduced P450 is split into two components, with w
associated with the heme iron developing a more a peak at 406 nm and a shoulder at ~425 nm. .jbc
.o
positive potential and becoming easier to reduce The former likely represents thiolate- rg
(e.g. 22,33). Spectroelectrochemical titrations coordinated ferrous OleTJE, and the latter the by g/
were done for both substrate-free and arachidic thiol-coordinated form (39). A similar ue
s
acid-bound forms of OleT to determine the phenomenon is seen for the arachidic acid- t o
JE n
midpoint potentials for the heme iron Fe3+/Fe2+ bound OleT (Figure 6B), although in this case A
JE p
couples (versus the normal hydrogen electrode, the main peak is at 420 nm with a shoulder at ril 4
NHE). Despite the extensive HS heme content in ~400 nm, suggesting a higher proportion of the , 2
0
the arachidic acid-bound OleT , its heme thiol-coordinated ferrous form in the substrate- 19
JE
potential (-105 ± 6 mV) is not significantly bound OleTJE. For both substrate-free and
different from that of the substrate-free form (- arachidic acid-bound OleTJE redox titrations, it
103 ± 6 mV) (Figure 6). In both cases, the heme is evident that there is a single set of isosbestic
iron potentials are quite positive compared to points throughout the titrations, indicating that
many bacterial P450s which rely on NAD(P)H- the equilibrium between thiol- and thiolate-
dependent electron transfer from protein redox coordinated ferrous forms remains constant as
partner systems. Examples include the camphor the concentration of ferrous OleTJE accumulates.
binding-induced shift in heme iron potential The Soret isosbestic point is at 408 nm for the
from -300 mV to -170 mV (vs. NHE) in arachidic acid-bound form, and at 410 nm for
P450cam (enabling electron transfer from the substrate-free OleTJE. Thus, under the same
ferredoxin partner at -240 mV) (33,38); and the redox titration conditions, arachidic acid
arachidonic acid-induced shift in potential from substrate binding seems to push the ferrous
-429 mV to -289 mV (vs. NHE) in P450 BM3 heme cysteine thiolate/thiol equilibrium slightly
(22). However, unlike the aforementioned further towards the thiol-coordinated state.
P450s, OleT is evolutionarily adapted to
JE
interact directly with H O in order to form
2 2
9
Structure of an alkene producing P450
Stopped-flow analysis of OleT turnover ferric heme with a thiolate proximal ligand to
JE
kinetics the iron and a distal ligating water molecule
(Figure 8A). Several such LS forms with
In order to determine the kinetics of H O -
2 2 rhombic anisotropy are evident from the
dependent fatty acid oxidation, we exploited the
multiplicity of lines observed and the resolvable
fact that turnover of bound substrate is
contributions at g show g-values ranging from
accompanied by a reconversion of OleT heme z
JE those typical for LS ferric P450s (2.43, 2.48)
iron spin-state from HS to LS as the substrate is
(e.g. 39-42) to those associated with
decarboxylated. The two states of the P450 have
chloroperoxidases and the fatty acid hydroxylase
considerably different heme spectra, and thus we
P450 (CYP152B1) (2.55, 2.61 and possibly
used stopped-flow absorbance spectroscopy to SPα
2.70) (17,43). Overall the EPR spectrum
measure the rate constants for LS OleT heme
JE suggests a large, water filled site with multiple
formation at 417 nm across a range of H O
2 2 coordination geometries and hydrogen bonding
concentrations up to 200 µM. Reaction kinetics
partners available to the distal water ligand. The
are 2nd order with respect to [H O ], with
2 2 addition of substrate, arachidic acid, produces a
observed rate constants (k ) for arachidic acid
obs very different EPR spectrum dominated by two
oxidation and concomitant LS heme recovery up
S = 5/2 rhombic HS ferric thiolate-ligated heme
to 167 s-1 at the highest [H O ] tested (200 µM)
2 2 signals having five-coordinate iron. The g- D
(Figure 7A). The k versus [H O ] data were o
obs 2 2 values are 7.76, 3.84 and 1.75 for one signal, w
fitted using a linear equation – giving a 2nd order and 7.76, 3.67 and 1.71 for the second (the nloa
rate constant (kon) of (8.0 ± 0.2) x 105 M-1 s-1 to signal at g = 4.26 arises from non-specifically ded
describe the catalytic process. The apparent kobs bound non-heme iron) (Figure 8B). The fro
m
value at the y-axis intercept (zero [H2O2]) is 8.32 differences between these two forms reflect h
± 1.96 s-1, giving an estimate for the H O k ttp
rate constant. The k /k ratio thus gi2ve2s aonff small differences in the ligand field geometry ://w
off on and as such are likely to be a result of distortions w
estimate of the apparent K for H O as 10.40 ± w
d 2 2 of the heme group and the thiolate ligand rather .jb
2.71 µM. Figure 7B shows overlaid spectra than any change in the ligation of the heme iron. c.o
captured during the reaction of arachidic acid- rg
The observation of high spin heme is in contrast b/
bound OleT with H O at a final concentration y
JE 2 2 to P450 that shows no spin state change on g
SPα u
of 7.58 µM. The spectral overlay describes a e
substrate binding (17), and where x-ray s
smooth transition from the substrate-bound, crystallography has shown that the heme retains t on
A
teoxwteanrsdisv ethlye sHubSs trfaotrem-f reoef LSO lfeoTrmJE aat t4 1389 4n mn ams tAhpep wroaxteimr astiexltyh lig1a5n%d whoef n stuhbes traptreo ties inb oun(ads. pril 4
the oxidation reaction occurs and the product determined by relative integration of the low , 20
1
leaves the heme environment and a water ligand 9
spin forms, accounting for differences in
binds to the heme iron. A series of isosbestic
concentration and subtraction of baselines to
points are observed in the overlaid spectra
account for underlying high spin species) is
(notably in the Soret region at 410 nm) that
converted to a new LS species with g-values of
indicate no significant accumulation of any
g = 2.46, g = 2.25 and g = 1.89, and which is
z y x
intermediate species in the reaction. The Figure
not present in the substrate free enzyme. It is
7B inset shows the accompanying stopped-flow
likely that this minor LS species is in
data for this reaction at 417 nm and 7.58 µM equilibrium with the HS form.
H O , with data fitted accurately using a single
2 2

exponential function to give a k of 12.50 ±
obs
1.16 s-1. OleT -catalyzed substrate turnover
JE
OleT turnover assays were done using H O
JE 2 2
and with a range of saturated fatty acids (C12-
EPR analysis of OleT
JE
C20), as described in the Experimental
The continuous wave X-band EPR spectrum of Procedures. As reported, by Rude et al. (18),
substrate-free OleTJE (prepared using Method 1) products were identified and characterized as
displays features attributable to the S = ½ LS
10