Members of the cytochrome P450 monooxygenase family CYP268 are found across a broad range of Mycobacterium species including the pathogens Mycobacterium avium, M. colombiense, M. kansasii, and M. marinum. CYP268A2, from M. marinum, which is the first member of this family to be studied, was purified and characterised. CYP268A2 was found to bind a variety of substrates with high affinity, including branched and straight chain fatty acids (C10–C12), acetate esters, and aromatic compounds. The enzyme was also found to bind phenylimidazole inhibitors but not larger azoles, such as ketoconazole. The monooxygenase activity of CYP268A2 was efficiently reconstituted using heterologous electron transfer partner proteins. CYP268A2 hydroxylated geranyl acetate and trans-pseudoionone at a terminal methyl group to yield (2E,6E)-8-hydroxy-3,7-dimethylocta-2,6-dien-1-yl acetate and (3E,5E,9E)-11-hydroxy-6,10-dimethylundeca-3,5,9-trien-2-one, respectively. The X-ray crystal structure of CYP268A2 was solved to a resolution of 2.0 Å with trans-pseudoionone bound in the active site. The overall structure was similar to that of the related phytanic acid monooxygenase CYP124A1 enzyme from Mycobacterium tuberculosis, which shares 41% sequence identity. The active site is predominantly hydrophobic, but includes the Ser99 and Gln209 residues which form hydrogen bonds with the terminal carbonyl group of the pseudoionone. The structure provided an explanation on why CYP268A2 shows a preference for shorter substrates over the longer chain fatty acids which bind to CYP124A1 and the selective nature of the catalysed monooxygenase activity.

Introduction

Cytochrome P450 (CYP) enzymes are a family of haem monooxygenases, capable of catalysing the insertion of a single oxygen atom, derived from molecular oxygen, into an inert carbon–hydrogen bond of a wide range of organic substrates [1]. Cytochrome P450 enzymes are ubiquitous in nature, with genes in humans [2], other animals [3], plants, and many fungal and bacterial species. CYP enzymes perform both anabolic (building up metabolites) and catabolic (breaking them down) processes, making them key enzymes in secondary and xenobiotic metabolism [2] and targets for antibacterial drug design [4]. They have also been shown to perform a variety of reactions, most commonly hydroxylation but also further oxidation, C–C bond formation, desaturation, and epoxidation, using electrons ultimately sourced from NAD(P)H [reduced nicotinamide adenine dinucleotide (phosphate)] [5]. Individual CYPs often show high specificity for their substrate and selectivity in the site of the target molecule where the oxidation takes place [6,7]. The catalytic activity of CYP enzymes is dependent on the delivery of the two electrons to the haem in two separate, highly regulated steps. In bacterial species, this is most often achieved by the combination of two cytosolic electron transfer partners, a flavin adenine dinucleotide (FAD) containing ferredoxin reductase and an iron–sulphur ferredoxin, together known as a Class 1 electron transfer system [8]. CYPs are classified based on amino acid sequence similarity, where members of a family share >40% sequence homology and subfamily members >55% [9].

Upon sequencing of the Mycobacterium tuberculosis genome [10], the high number of CYP genes found (20) was unprecedented for a bacterial species at the time. (In contrast, Escherichia coli has none.) As a result, the CYPome of M. tuberculosis has been a target for inhibitory drug design in the years since [11]. Owing to increasing numbers of drug-resistant and multi-drug-resistant strains, and a disease profile that overlaps with that of HIV/AIDS, M. tuberculosis continues to be responsible for large-scale loss of life [12,13]. Members of the Mycobacterium family are widespread and range from soil bacteria to human pathogens. Mycobacterium marinum M, a pathogen of frogs and fish, is a close relative of both M. tuberculosis H37Rv (85% nucleotide identity) and also the pathogenic species Mycobacterium ulcerans Agy99 (97%) [14]. M. ulcerans is responsible for the Buruli ulcer (also referred to as the Daintree or Bairnsdale ulcer), which is a skin disease primarily found in tropical areas, most often in central and western Africa [15]. M. marinum has 47 individual CYP encoding genes in its genome, while M. ulcerans has 24. M. marinum is thought to resemble a common ancestor of the more pathogenic Mycobacterium species, with a genome that has not undergone the extensive reduction by gene deletion and pseudogene formation that characterises the genome of M. tuberculosis and, to a lesser extent, M. ulcerans [14,15]. M. marinum is a more versatile pathogen, primarily affecting aquatic life but also capable of surviving in a human host as the causal agent of aquarium granuloma, and, unlike the human pathogens, of persisting outside of its host entirely [14,16]. It has been shown to adapt to a variety of hosts, altering virulence mechanisms to suit, including the differential regulation of polyketide lipids, and sterol utilisation [16]. Thus, the larger genome of M. marinum provides both increased redundancy, with a smaller percentage of essential genes than M. tuberculosis, but also increased adaptability. This is part of a common trend, where the number of CYP genes in Mycobacterium species decreases as the organism environment changes from soil living mycobacteria (average of 50 CYPs) to a human pathogen (average of 19) [17].

Where there are direct counterparts for M. marinum CYPs in M. tuberculosis, the roles of the majority are still unknown. Several of the enzymes are reported as cholesterol oxidases, including some that are essential for viability in vitro. The deletion of the CYP125A1 enzyme together with CYP124A1 leads to a build-up of the intermediate cholest-4-en-3-one and the inhibition of the organism [18]. The cholesterol degradation activity of CYP125A1 has been linked to the density of the mycobacterial cell wall, increasing the mass of phthiocerol dimycocerosates (PDIM) [19]. The Mycobacterial cell envelope is distinguished by several features, most prominently the additional layer of long-chain fatty acids known as mycolic acid covalently bound to the peptidoglycan of the cell wall [20]. PDIM and other multiple methyl-branched long-chain fatty acids further increase the density and thickness of this layer. Lipid metabolism is another common role of bacterial CYPs, and in M. tuberculosis, CYP124A1 has been shown to hydroxylate phytanic acid and similar fatty acid compounds [21]. The majority of M. tuberculosis CYPs, however, have either resisted recombinant expression efforts, or have not shown activity when screened against libraries of common substrates, leaving uncertainty about the roles of the Mycobacterial CYPome as a whole. Even where the substrate and product is known, it is not well understood how these play into the metabolism of the organism. For example, while M. tuberculosis is known to have no sterol synthesis pathway, CYP51B1, a highly conserved sterol-α-demethylase, is present [22]. The current understanding of M. tuberculosis virulence points to specialised areas such as mycolic acid synthesis, other lipid metabolism, and cholesterol catabolism as critical [23], in the second two of which there is evidence for CYP involvement.

Analogues of many of these characterised CYP proteins are present in other species of Mycobacteria. For example, CYP51B1 is found in a highly conserved operon in Mycobacterium smegmatis MC2 155 and M. tuberculosis, containing a CYP123 enzyme, a ferredoxin, a TetR regulator, and an ORF of unknown function [24]. The same operon is conserved in M. marinum M and M. ulcerans Agy99. The conservation of various CYPs outside of the M. tuberculosis complex (MTBC) has been taken as evidence that they perform general or housekeeping roles, while conservation only in Mycobacterium bovis or mammalian pathogens has led to a suggested role in virulence or human infectivity [25]. The characterisation of CYP enzymes present only in non-MTBC species has been attempted for various reasons. Some were potential catalysts for reactions of biosynthetic interest, such as CYP153A16, from M. marinum, which was found like other members of the CYP153 family to oxidise medium-chain alkanes [7]. CYP151A1 in M. smegmatis MC2 155 and CYP151A2 from Mycobacterium sp. strain RP1 were identified in the effort to understand the ability of the respective organisms to utilise pyrrolidine and piperidine as sole carbon sources [26,27]. Both can oxidise secondary amines, catalysing ring opening and allowing further catabolism. In addition, CYP150 family members from Mycobacterium vanbaalenii PYR-1 have been hypothesised to oxidise polycyclic aromatic hydrocarbons, making them a catalyst of interest for environmental remediation [28].

As a result of its larger CYPome, many CYPs are found in M. marinum M that do not have direct counterparts in either of M. tuberculosis or M. ulcerans, and which may play a role in the different pathogenicity of the organism. The enzyme CYP268A2 is one instance of this, with only a truncated pseudogene (cyp268A2P) remaining in the M. ulcerans Agy99 genome, and no relative in M. tuberculosis. It is highly conserved across a broad range of other Mycobacterium species and other bacteria. We have performed preliminary structural and functional characterisation of this enzyme, with a view to elucidating its role in the metabolism of M. marinum and across the Mycobacteria (Mycobacteriaceae) as a whole.

Experimental

General

All organic substrates, derivatisation agents, and general laboratory reagents, except where otherwise noted, were purchased from Sigma–Aldrich, Alfa-Aesar, VWR International, or Tokyo Chemical Industry. Antibiotics, detergents, DTT (dithiolthreitol), and IPTG (isopropyl β-d-1-thiogalactopyranoside) were from Astral Scientific. Restriction enzymes used for cloning were purchased from New England Biolabs. KOD polymerase, used for the PCR (polymerase chain reaction) steps, and the expression vectors were from Merck-Millipore.

The following were used as media for cell growth and maintenance: LB (lysogeny broth); tryptone (10 g), yeast extract (5 g), and NaCl (10 g) per litre; 2×YT contains tryptone (16 g), yeast extract (10 g), and NaCl (5 g) per litre; SOC (Super Optimal broth with Catabolite repression); tryptone (20 g), yeast extract (5 g), NaCl (0.5 g), KCl (0.2 g), MgCl2 (1 g), and 5 ml of 40% glucose per litre; EMM (E. coli minimal media); K2HPO4 (7 g), KH2PO4 (3 g), Na3citrate (0.5 g), (NH4)2SO4 (1 g), MgSO4 (0.1 g), 20% glucose (20 ml), and glycerol (1%, v/v) in 1 l; trace elements: 0.74 g CaCl2·H2O, 0.18 g ZnSO4·7H2O, 0.132 g MnSO4·4H2O, 20.1 g Na2EDTA, 16.7 g FeCl3·6H2O, 0.10 g CuSO4·5H2O, 0.25 g CoCl2·6H2O. Antibiotics were added to the following working concentrations: ampicillin, 100 µg ml−1 and kanamycin, 30 µg ml−1.

UV–visible spectra were recorded on a Varian Cary 5000 at 30 ± 0.5°C. GC–MS analysis was performed using a Shimadzu GC-17A equipped with a QP5050A GC–MS detector and DB-5 MS-fused silica column (30 m × 0.25 mm, 0.25 µm) or a Shimadzu GC-2010 equipped with a QP2010S GC–MS detector, an AOC- 20i autoinjector, an AOC-20s autosampler, and a DB-5 MS-fused silica column (30 m × 0.25 mm, 0.25 µm). The injector was held at 250°C and the interface at 280°C. Column flow was set at 1.5 ml min−1 and the split ratio was 24. Solvent cut time was set to 3 min. For ionone and acetate substrates, the oven was held at 80°C for 3 min followed by an increase of 10°C min−1 up to 220°C and a final hold for 3 min. For 4-phenyltoluene, the initial temperature and hold time were the same but the rate of increase was 12°C min−1 to a maximum of 230°C, where it was held for 5 min. Preparative HPLC analysis was carried out on a Shimadzu system equipped with a DGU-20A5R degasser, 2× LC-20AR pumps, SIL-20AC HT autosampler, an SPD-M20A photodiode array detector, and a CT0-20AC column oven. Separation was performed using an Ascentis Si HPLC column (25 cm × 10 mm × 5 µm; Sigma–Aldrich). NMR was performed using an Agilent DD2 spectrometer at 500 MHz for 1H and 126 MHz for 13C.

Sequence alignment performed by ClustalW. Phylogenetic tree drawn using the maximum-likelihood method based on the Jones–Taylor–Thornton (JTT) model with complete deletion of missing data. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model and then selecting the topology with a superior log-likelihood value.

Recombinant protein expression and purification

The CYP268A2 gene (Mmar_3761) was amplified by PCR using oligonucleotide primers (vide supra). The gene was amplified by 30 cycles of strand separation at 95°C for 45 s followed by annealing at 55°C for 30 s and extension at 68°C for 80 s. The genes were cloned into the pET26 vector using the appropriate NdeI and HindIII restriction enzymes, and the correct insertion was checked by Sanger sequencing performed by Australian Genome Research Facility Ltd (AGRF). The plasmid was then transformed into E. coli BL21(DE3). The transformed E. coli cells were grown on an LBkan plate and incubated in 2YTkan (1.2 l total, in six flasks) at 37°C for 5 h at 160 rpm. Following this, the temperature was reduced to 18°C, the speed to 90 rpm. Benzyl alcohol (0.02%, v/v), ethanol (2%, v/v) and, after 30 min, IPTG (0.1 mM) were added to induce protein expression [29]. The growths were continued for a further 16 h before harvesting of the cell pellet by centrifugation (5000 g, 15 min). The cells were then resuspended in 50 mM Tris (pH 7.4) with 1 mM DTT (henceforth Buffer T), with β-mercaptoethanol (1 ml), TWEEN (1 ml), and glycerol (20%, v/v), to a total volume of 200 ml. The resuspended cells were then lysed by sonication (25 cycles of 20 : 40 s on/off, 70%, 19 mm probe, Sonics Vibra-Cell) while kept on ice. The supernatant was isolated from the cell debris by centrifugation (40 000 g, 30 min) and then loaded onto a DEAE Sepharose column (XK50, 200 mm × 40 mm, GE Healthcare) and eluted with a linear gradient of 100–300 mM KCl in Buffer T. The fractions containing the desired P450 were identified by their red colour, and pooled, concentrated using a Vivacell 100 (Sartorius Stedim, 10 KD membrane), and desalted using a Sephadex G-25 medium grain column (250 mm × 40 mm) with elution using Buffer T. The protein sample was then loaded onto a further Source-Q anion exchange column (XK26, 80 mm × 30 mm, GE Healthcare) and eluted with a gradient of 0–1 M KCl in Buffer T. Collected fractions were concentrated again by ultrafiltration and stored at −20°C in an equal volume of glycerol, after filtration with a 0.22 µM syringe filter. The final A420/A280 ratio was 1.6.

Before use, the glycerol was removed from CYP268A2 by gel filtration, using a 5 ml PD-10 column (GE Healthcare) and elution with Buffer T (without DTT). The extinction coefficient for CYP268A2 was determined by CO difference spectra using ε450 = 91 mM−1 cm−1 for the reduced CO-bound form [30]. The CYP268A2 concentration was determined using ε419 = 108 ± 7 mM−1 cm−1.

Spin-state shifts and substrate-binding titrations

To determine the extent of substrate binding, CYP268A2 was diluted to ∼1 µM in 50 mM Tris (pH 7.4) to a volume of 500 µl, and various substrates (100 mM, EtOH or DMSO) were added. The absorbance between 600 and 250 nm was recorded on the UV spectrophotometer until no further spectral change was observed. The high spin percentage was estimated (±5%) by comparison to a set of spectra, generated by the sum of substrate-free (>95% low spin, 418 nm Soret maximum) and camphor-bound (>95% high spin, 392 nm Soret maximum) CYP101A1 to the appropriate percentages.

For substrate-binding titrations, CYP268A2 was diluted to 2 µM in 50 mM Tris (pH 7.4) to a volume of 2.5 ml, and 1–3 µl of substrate was added via a Hamilton syringe from either 1, 10, and 100 mM (EtOH or DMSO) stock solution, starting from the lowest concentration. The peak-to-trough difference in absorbance, between 600 and 250 nm, was recorded until additional aliquots caused no further spectral change in the Soret band. The dissociation constant for that substrate was obtained by fitting the difference in absorbance against the substrate concentration to the hyperbolic function:

 
formula

where Kd is the binding constant, [S] is the substrate concentration, ΔA the peak-to-trough ratio, and ΔAmax the maximum peak-to-trough absorbance. In the instances where the substrate exhibited tight binding (Kd < 10 µM, less than five times the concentration of the enzyme), the data were instead fitted to the tight-binding quadratic equation:

 
formula

where Kd is the binding constant, [S] is the substrate concentration, ΔA the peak-to-trough ratio, ΔAmax the maximum peak-to-trough absorbance, and [E] is the enzyme concentration [31].

Construction of in vivo systems to support product formation

To construct whole-cell turnover systems, CYP268A2 was cloned into a pRSFDuet vector using the NdeI and KpnI sites introduced by PCR. ArR (a ferredoxin reductase from Novosphingobium aromaticivorans) has been cloned into pETDuet previously [32]. The terpredoxin gene was purchased (gblock; Integrated DNA Technologies) and cloned into pETDuetArR using NcoI and HindIII sites (Supplementary Material). The two vectors, pRSFDuet-containing CYP268A2, and pETDuet, containing Tdx [terpredoxin (a ferredoxin from Pseudomonas sp.)] and ArR, were transformed into E. coli BL21 cells. The transformed cells were grown overnight on an LBamp/kan plate. A colony was picked and grown in 50 ml LBamp/kan for 6 h at 37°C and 110 rpm. It was then cooled to 18°C and slowed to 90 rpm, with the addition of benzyl alcohol (0.02%, v/v) and ethanol (2%, v/v), followed after 30 min by IPTG (0.1 mM). The culture was then left overnight. The cell pellet was harvested by centrifugation (5000 g, 10 min) before being resuspended in EMMamp/kan (100 ml). The substrate was then added to 1 mM final concentration before shaking at 150 rpm at 30°C. After 24 h, the turnover was then centrifuged (15 min, 5000 g) and the supernatant was isolated. Samples (1 ml) of the turnover were taken for initial testing via GC–MS at various time points. The samples were extracted into ethyl acetate, dried over MgSO4 before resuspension in anhydrous acetonitrile (200 µl). Where GC–MS showed product formation, larger scale growths by the same method (200 ml EMM) were performed and extracted. The extract was dissolved in 5% isopropanol : hexane and purified by semi-preparative HPLC using an elution gradient of 5–10% isopropanol. The chromatogram was monitored at 254 nm. Relevant peaks (as confirmed by GC–MS) were collected, pooled, and resuspended in deuterated chloroform to allow analysis by NMR.

In vitro NADH activity assays

Purified CYP268A2 (0.5 µM) together with the ferredoxin Tdx (5 µM) and ferredoxin reductase ArR (0.5 µM) (the purification method of these are reported elsewhere [32,33]) were mixed to a total volume of 1.2 ml in oxygenated 50 mM Tris (pH 7.4), with added catalase (120 µg). The mixture was equilibrated for 2 min at 30°C before the addition of 320 µM NADH (A340 ≈ 2.0). Substrate was then added to a concentration of 0.25 mM. The reaction was monitored at 340 nm for the duration. The rate of NADH turnover was calculated by plotting the A340 against time, using ε340 = 6.22 µM. Once the reaction was at completion, 1 ml of the turnover was extracted into ethyl acetate and analysed by GC–MS as above.

Crystallography, data collection, data processing, and structural determination

CYP268A2 was further purified by size-exclusion chromatography (Enrich SEC Column, 650 10 × 300 mm, 1 ml min−1 flow rate) before being concentrated to 30–35 mg ml−1 in Buffer T. The substrate pseudoionone (100 mM EtOH, mixture of isomers) was added to the protein to a final concentration of 1 mM immediately before crystallisation. The initial screening of crystallisation conditions was performed by the sitting-drop method in 96-well plates with commercially available screening conditions (Hampton Research) using 1 µl of both the protein solution and the reservoir solution. Crystal conditions were refined using the hanging drop vapour diffusion method, and again with 1 µl of both the protein solution and the reservoir solution, equilibrated with a 500 µl reservoir. Diffraction-quality crystals (plates with dimensions ∼150 × 140 × 20 µM, Supplementary Figure S4) were obtained after 2 weeks at 16°C from the condition containing 0.96 M ammonium phosphate and 0.3 M sodium citrate (pH 7.0). They were harvested with a Micromount (MiTeGen) and cryo-protected by immersion in Parabar 10312 (Paratone-N, Hampton Research) before flash cooling in liquid N2. Data were collected by X-ray diffraction at 100 K on the Australian Synchrotron MX1 beamline (360 exposures using 1° oscillations at a wavelength of 0.9537 Å) [34]. The data were processed in the space group C2 using XDS [35,36]. Molecular replacement phasing was carried out using the modified search model (residues 1–23, 145–152, 322–337, and 244–253 were removed from the surface to improve the search model clashes) from PDB entry 2WM4 [21] (CYP124A1 with phytanic acid bound, found by searching the FFAS server) with the MRSAD (Molecular Replacement Single Wavelength Anomalous Diffraction) protocol of Auto-Rickshaw [37,38] using the native dataset (however, SAD failed and phasing were obtained by MR only). Within the pipeline, various programmes from the CCP4 program suite [39] were used and the model phases were improved by model refinement using CNS [40,41] and REFMAC5 [42], density modification using PIRATE [43], and rebuilding of the model using SHELXE [44], RESOLVE [45], and Buccaneer [46], finally refinement of the resulting model using Phenix [47] and REFMAC5. The structure was further rebuilt using Coot [48] based on the initial electron density maps, with multiple structural refinement iterations using phenix.refine. Composite omit maps were generated using the Composite Omit Maps program in Phenix.

Results/discussion

Phylogenetic and sequence analysis

Alongside CYP268A2, genes encoding members of the cytochrome P450 monooxygenase CYP268 family are primarily found in other Mycobacterium species. These include CYP268A1 from M. avium subsp. paratuberculosis, A3 and B2 from M. smegmatis, and B1 and C1 from M. vanbaalenii (accession numbers and similarities are listed in Supplementary Table S1). A single member, CYP268A4 is found in Streptomyces bingchengensis. A BLAST search found a large number of proteins with high sequence identity (250 results >70%), almost exclusively from other Mycobacterium species (none of the first 250 results were from non-Mycobacteria species). CYP268A2 is conserved in Mycobacterium liflandii (98%) and as a pseudogene in M. ulcerans Agy99 (95%, truncated after 195 residues). A similar gene without truncation is found in M. ulcerans subsp. shinshuense (97%), a clinical isolate from Japan [49]. Close analogues are also found in M. avium (78%), M. colombiense (77%), and M. kansasii (85%), among others. A M. tuberculosis strain TKK-01-0051 contained an analogous protein with high sequence similarity (78%), although there is some suggestion that this strain may be misclassified and is better referred to as M. colombiense [50]. M. colombiense is a member of the M. avium complex and is known to opportunistically infect HIV-positive immuno-compromised patients [51].

The CYP268 family also shares high sequence similarity with the CYP124, CYP125, and CYP142 families, of which members from M. tuberculosis, M. marinum, M. smegmatis, and M. vanbaalenii are known (Supplementary Table S1). In particular, CYP268A2 has a sequence similarity above 40% with many members of the CYP124 family. The CYP142B subfamily members cluster together (Figure 1), more closely to the 124 and 268 families than to the 142A subfamily, suggesting a high degree of overlap between the three families. CYP124A1 from M. tuberculosis (with 41% similarity to CYP268A2) has been characterised as a lipid hydroxylase [21] and additionally can oxidise vitamin D3 and other analogues [52]. CYP142A1, also from M. tuberculosis (33% similarity to CYP268A2), can oxidise cholesterol esters [53], and CYP125A1 (40%) is similarly a cholesterol hydroxylase [54].

Phylogenetic tree of the CYP268 family.

Figure 1.
Phylogenetic tree of the CYP268 family.

Included are some members of the CYP124 and CYP142 families, both of which have been predominately found in Mycobacterium species. The grouping shows that CYP268A2 is closely related to members of the 268 family, with significant sequence similarity to the 124 family (all above 40%) and the 142B subfamily (43% to 142B1). CYP124G1, which does not cluster with the remainder of the 124 family, has 41% similarity to CYP268A2. Percentage identities can be found in Supplementary Table S1. The tree was drawn to scale, with branch lengths and scale measured in the number of substitutions per site.

Figure 1.
Phylogenetic tree of the CYP268 family.

Included are some members of the CYP124 and CYP142 families, both of which have been predominately found in Mycobacterium species. The grouping shows that CYP268A2 is closely related to members of the 268 family, with significant sequence similarity to the 124 family (all above 40%) and the 142B subfamily (43% to 142B1). CYP124G1, which does not cluster with the remainder of the 124 family, has 41% similarity to CYP268A2. Percentage identities can be found in Supplementary Table S1. The tree was drawn to scale, with branch lengths and scale measured in the number of substitutions per site.

A sequence alignment (Supplementary Figure S1) revealed that of the cytochrome P450 commonly conserved sequence elements, CYP268A2 retains the glutamate and arginine pair (Glu300, Arg303), as well as the phenylalanine residue (Phe364) in the K helix, seven residues before the conserved proximal cysteine at Cys372 [55]. The acid alcohol pair in the enzyme is an aspartate (Asp263) and threonine (Thr264). The CYP124A1 family members contain a glutamate (Glu270 in CYP124A1) rather than the aspartate found in the CYP268 enzymes (Supplementary Figure S1).

The area surrounding putative cyp268 genes is marked by the presence of a highly conserved operon (Supplementary Figure S2), containing a GTPase, a large ribonuclease, two regulatory proteins (AcrR and Sir2-like), and a downstream NAD synthetase. The cyp268A2 gene is flanked on both sides by a PE-PGRS (proline glutamate-polymorphic GC-rich sequence proteins) gene (glycine-rich proteins detected across the Mycobacteria with possible roles as antigens [56] or fibronectin-binding [57]). The environments of the cyp124 genes (M. marinum and M. tuberculosis) do not have any of these genes, but share some similarity with each other (Supplementary Figure S2).

Characterisation of CYP268A2 and its substrate range

The CYP268A2 enzyme was produced in E. coli and purified by two ion exchange chromatography steps. The protein was tested for the characteristic Soret absorbance, occurring when the ferrous form of the CYP enzyme binds CO [58]. CYP268A2 after reduction by sodium dithionite and gentle bubbling with CO shifted almost completely from the resting state absorbance at 419–450 nm, with only a small shoulder (<5%) at 420 nm (Figure 2). The extinction coefficient for the Soret absorbance [58], ε450 = 91 mM cm−1, was used to determine the extinction coefficient for CYP268A2 ε419 = 108 ± 7 mM cm−1, which was henceforth used to determine the concentration of the enzyme.

Ferric CYP268A2 (black, A419), the reduced ferrous form (blue, A412), and the ferrous form bound with CO (red, A449), showing the characteristic absorbance at ∼450 nm.

Figure 2.
Ferric CYP268A2 (black, A419), the reduced ferrous form (blue, A412), and the ferrous form bound with CO (red, A449), showing the characteristic absorbance at ∼450 nm.

The shoulder at 420 nm in the ferrous–CO form comprises ∼5% of the total area.

Figure 2.
Ferric CYP268A2 (black, A419), the reduced ferrous form (blue, A412), and the ferrous form bound with CO (red, A449), showing the characteristic absorbance at ∼450 nm.

The shoulder at 420 nm in the ferrous–CO form comprises ∼5% of the total area.

As CYP268A2 is the first member of its P450 family to be studied, previously characterised family members are not available to give an indication as to the role of the enzyme. Based on the similarity to members of the CYP124 lipid-hydroxylase family [21], the initial substrates tested on CYP268A2 were branched and straight chain fatty acids and esters (Figure 3). Many of these were successful in shifting the majority of the enzyme into the high spin form, indicating that they were accommodated by the active site of the enzyme (Supplementary Figure S5). Geranyl acetate (80% HS) and farnesyl acetate (75% HS) both induced higher spin-state shifts in CYP268A2 than undecanoic acid (70% HS), which was the best performing acid substrate (Table 1). Additionally, it was discovered that a range of aromatic compounds could bind to the enzyme. 4-Phenyltoluene (55% HS), phenyl acetate (50% HS), and phenylcyclohexane (50% HS) all successfully induced a Type I Soret shift. The binding affinity of CYP268A2 for the substrates that gave the highest spin-state shifts was then assessed (Figure 4). Farnesol (75% HS) bound to the enzyme with high affinity, Kd, 0.8 ± 0.2 µM. Undecanoic acid gave Kd, 1.1 ± 0.5 µM. Pseudoionone, a linear ionone precursor, gave 80% HS and Kd, 3.6 ± 0.6 µM (as a mixture of isomers). 4-Phenyltoluene also bound tightly, Kd, 13 ± 2.4 µM. Notably, the longer substrates preferred by CYP124A1 such as phytanic acid failed to induce a significant spin-state shift in CYP268A2 (Supplementary Table S2).

Structures of compounds tested as potential CYP268A2 substrates.

Dissociation constant analysis of CYP268A2 with selected substrates.

Figure 4.
Dissociation constant analysis of CYP268A2 with selected substrates.

(a) Pseudoionone, (b) geranyl acetate, (c) 4-phenyltoluene, (d) undecanoic acid, (e) 1-phenylimidazole, and (f) 4-phenylimidazole. The inset shows the spectral change upon titration with substrate. The peak-to-trough difference in absorbance was measured from 420 to 390 nm, except with 1-phenylimidazole (peak-to-trough 413–435 nm) and 4-phenylimidazole (414–434 nm), where Type II spectral shifts were recorded.

Figure 4.
Dissociation constant analysis of CYP268A2 with selected substrates.

(a) Pseudoionone, (b) geranyl acetate, (c) 4-phenyltoluene, (d) undecanoic acid, (e) 1-phenylimidazole, and (f) 4-phenylimidazole. The inset shows the spectral change upon titration with substrate. The peak-to-trough difference in absorbance was measured from 420 to 390 nm, except with 1-phenylimidazole (peak-to-trough 413–435 nm) and 4-phenylimidazole (414–434 nm), where Type II spectral shifts were recorded.

Table 1
Spin-state shift and dissociation constants of CYP268A2 with a variety of substrates, presented in descending order of magnitude of the type I spin-state shift.

The spin-state shifts of a range of additional substrates are listed in Supplementary Table S2.

CYP268A2 substrates Spin-state shift (% HS) Kd (µM) 
Geranyl acetate 80 8.5 ± 1.9 
Pseudoionone 80 3.6 ± 0.6 
Farnesol 75 0.8 ± 0.2 
Farnesyl acetate 75 5.1 ± 1.9 
10-Undecenoic acid 70 1.6 ± 0.2 
Undecanoic acid 70 1.1 ± 0.5 
Capric acid 65 1.1 ± 0.5 
4-Phenyltoluene 55 13 ± 2.4 
Geraniol 55 7.9 ± 1.6 
Lauric acid 55 270 ± 36 
Linalyl acetate 55 110 ± 25 
Neryl acetate 30 106 ± 29 
CYP268A2 substrates Spin-state shift (% HS) Kd (µM) 
Geranyl acetate 80 8.5 ± 1.9 
Pseudoionone 80 3.6 ± 0.6 
Farnesol 75 0.8 ± 0.2 
Farnesyl acetate 75 5.1 ± 1.9 
10-Undecenoic acid 70 1.6 ± 0.2 
Undecanoic acid 70 1.1 ± 0.5 
Capric acid 65 1.1 ± 0.5 
4-Phenyltoluene 55 13 ± 2.4 
Geraniol 55 7.9 ± 1.6 
Lauric acid 55 270 ± 36 
Linalyl acetate 55 110 ± 25 
Neryl acetate 30 106 ± 29 

The enzyme did not appear to have the strict requirement for branching methyl groups at the terminus of the substrate as shown by CYP124A1, with lauric acid (55% HS) outperforming both 11-methyllauric acid (30% HS) and 10-methyllauric acid (20% HS), although 15-methylhexadecanoic acid (15%) was more effective than palmitic acid (0%). The binding data suggested that the enzyme active site could better accommodate straight chain substrates over those with longer bent chains, preferring the trans-isomer, geranyl acetate (Kd, 8.5 ± 1.9 µM), over the cis-form, neryl acetate (30% HS, Kd, 106 ± 29 µM, Supplementary Figure S7). CYP124A1 has additionally been characterised as having cholesterol and vitamin D binding activity [52] (as do CYP125 and CYP142 family members, [54,59]). As a result, CYP268A2 was tested with a variety of cholesterol and vitamin D analogues, but those gave no or very little (0–5% high spin form) indication of binding. The substrate-binding data demonstrate that the active site of CYP268A2 is versatile and can accommodate a range of linear and aromatic hydrocarbons, terpenes, and fatty acids. Some substrates such as farnesol that have been reported for CYP124A1 bind well with CYP268A2, but it appears to support a wider range of substrate binding.

Many azole drugs have been reported to bind CYPs as competitive inhibitors, coordinating directly to the Fe atom of the haem and generating Type II spectral shifts, shifting the Soret band to a higher wavelength rather than to 390 nm which is characteristic of displacement of the distal water ligand [60,61]. These have been proposed as methods for inhibiting bacterial growth, particularly in species where CYPs play essential roles [62]. Many possible azole inhibitors were tested with CYP268A2 (Table 2). CYP268A2 gave Type II shifts with two of these, 1-phenylimidazole (shifted the Soret maximum to 421 nm) and 4-phenylimidazole (shifted the Soret maximum to 423 nm). The binding affinity of these to CYP268A2 was assessed by determining the dissociation constant (Figure 4). 1-Phenylimidzole bound more tightly (Kd, 0.9 ± 0.3 µM) than 4-phenylimidazole (Kd, 4.5 ± 0.6 µM). The addition of 2-phenylimidazole and other larger azole inhibitors yielded a small Type I shift.

Table 2
Crystal refinement data for CYP268A2 from M. marinum (PDB code: 6BLD).
Data collection statistics1 
 Wavelength 0.95370 
 Unit cell a = 154.76, b = 44.726, c = 57.727
α = 90, β = 100.84, γ = 90 
 Space group C
 Number of molecules in asymmetric unit 
 Resolution 2.00–42.91 (2.00–2.05)2 
 Number of unique reflections 26 452 (1743) 
 Completeness 99.2 (89.4) 
 Redundancy 7.4 (7.0) 
 (I)/[σ(I)] 9.7 (2.0) 
Rmerge (all I+ and I−) 0.130 (0.742) 
Rpim (all I+ and I−) 0.071 (0.411) 
CC(1/2) 0.997 (0.768) 
Rwork 19.96% 
Rfree 24.79% 
 % solvent 42.89 
 Number of residues modelled 413 
RMS deviation from restraint values 
 Bond lengths 0.002 
 Bond angles 0.551 
Ramachandran analysis 
 Most favoured 98.06% 
 Additionally allowed 1.94% 
Data collection statistics1 
 Wavelength 0.95370 
 Unit cell a = 154.76, b = 44.726, c = 57.727
α = 90, β = 100.84, γ = 90 
 Space group C
 Number of molecules in asymmetric unit 
 Resolution 2.00–42.91 (2.00–2.05)2 
 Number of unique reflections 26 452 (1743) 
 Completeness 99.2 (89.4) 
 Redundancy 7.4 (7.0) 
 (I)/[σ(I)] 9.7 (2.0) 
Rmerge (all I+ and I−) 0.130 (0.742) 
Rpim (all I+ and I−) 0.071 (0.411) 
CC(1/2) 0.997 (0.768) 
Rwork 19.96% 
Rfree 24.79% 
 % solvent 42.89 
 Number of residues modelled 413 
RMS deviation from restraint values 
 Bond lengths 0.002 
 Bond angles 0.551 
Ramachandran analysis 
 Most favoured 98.06% 
 Additionally allowed 1.94% 
1

Data collected from one crystal.

2

Values in parenthesis are for highest resolution shell.

Product characterisation

As CYP268A2 has no closely located ferredoxin gene in the M. marinum genome, it was expressed in E. coli with a variety of electron transfer partners from different organisms (Supplementary Table S3). The pair terpredoxin (2Fe–2S ferredoxin, from a Pseudomonas sp.) and an FAD-containing ferredoxin reductase ArR (from N. aromaticivorans) were selected (based on levels of product formation with geranyl acetate, Supplementary Material) to enable characterisation of the products of the enzyme. In vitro turnovers were used to analyse the product formation rate and where product was formed, substrates were scaled up with in vivo turnovers to allow characterisation.

The addition of geranyl acetate to CYP268A2 in vitro gave an NADH consumption rate of ∼60 nmol nmolP450−1 min−1 and when extracted and analysed by GC–MS, it showed a major product (9.5 min, Figure 5a) with a mass peak of 152.15 (expected mass of hydroxylation product is 214.16, 152.12 with loss of the acetate) after 24 h. To characterise the product, large-scale in vivo turnovers were performed (Figure 5a). The in vivo major product had a slightly different retention time at 9.4 min and the mass peaks (154.25, 136.20, and others) were two mass units higher than those of the in vitro product. The pattern of products in the GC trace of the whole-cell turnovers over time combined with mass spectrum peaks suggests that the major product in vivo was also hydrogenated in addition to the CYP-mediated hydroxylation (Supplementary Figure S8). HPLC purification of the major product followed by NMR analysis identified the metabolite as the terminal hydroxylation product (E)-8-hydroxy-3,7-dimethyloct-2-en-1-yl acetate (Figure 6 and Supplementary Figure S15). The in vitro turnover experiment with geranyl acetate confirms that the hydroxylation is CYP-mediated while it is likely that the hydrogenation was performed by an endogenous E. coli ene reductase [63,64].

GC chromatograms of CYP268A2 catalysed substrate turnovers.

Figure 5.
GC chromatograms of CYP268A2 catalysed substrate turnovers.

(a) GC chromatogram of the in vivo and in vitro turnovers of CYP268A2 with geranyl acetate. The retention times are as follows: geranyl acetate 6.3 min, in vivo product 9.4 min, in vitro product 9.5 min. (b) In vivo turnover with pseudoionone monitored over 24 h. Identification of isomer peaks was done by retention time comparison with neryl acetate (cis) and geranyl acetate (trans), where the cis form had the shorter retention time. The retention times are: cis-pseudoionone 12.0 min, trans-psuedoionone 12.7 min, 4 h products 15.9 min and 16.1 min, 24 h product 15.9 min.

Figure 5.
GC chromatograms of CYP268A2 catalysed substrate turnovers.

(a) GC chromatogram of the in vivo and in vitro turnovers of CYP268A2 with geranyl acetate. The retention times are as follows: geranyl acetate 6.3 min, in vivo product 9.4 min, in vitro product 9.5 min. (b) In vivo turnover with pseudoionone monitored over 24 h. Identification of isomer peaks was done by retention time comparison with neryl acetate (cis) and geranyl acetate (trans), where the cis form had the shorter retention time. The retention times are: cis-pseudoionone 12.0 min, trans-psuedoionone 12.7 min, 4 h products 15.9 min and 16.1 min, 24 h product 15.9 min.

Hydroxylation products of CYP268A2.

Figure 6.
Hydroxylation products of CYP268A2.

*In each, the hydroxylation at ω-terminus would generate a stereo-centre at the ω-1 carbon. However, in vivo the hydrogenation occurred after the hydroxylation, and as a result, the stereo-selectivity of the isolated product would be dependent on the E. coli ene reductase enzyme.

Figure 6.
Hydroxylation products of CYP268A2.

*In each, the hydroxylation at ω-terminus would generate a stereo-centre at the ω-1 carbon. However, in vivo the hydrogenation occurred after the hydroxylation, and as a result, the stereo-selectivity of the isolated product would be dependent on the E. coli ene reductase enzyme.

The in vitro turnover of pseudoionone by CYP268A2 had an NADH consumption rate of 48 nmol nmolP450−1 min−1 and generated a single major product at 16.1 min with a mass peak at 208.1 (expected hydroxylation product at 208.15). The addition of pseudoionone to the in vivo CYP/Tdx/ArR system generated a single major product after 24 h with a retention time of 15.9 min (Figure 5b). Mass peaks of the in vivo product (210.2, 195.20, and others) were again two mass units higher than the in vitro product (Supplementary Figure S9). The major product from the whole-cell turnover was isolated by HPLC and characterised by NMR to be (3E,5E)-11-hydroxy-6,10-dimethylundeca-3,5-dien-2-one (Figure 6 and Supplementary Figure S16). Similarly to geranyl acetate, the in vivo product has been hydrogenated at the ω-2 alkene in addition to ω hydroxylation. The pseudoionone was a mixture of isomers, and there are two substrate peaks at 12.0 and 12.7 min (192.15 and 192.20 m/z compared with expected molecular ion peak at 192.15), which are the cis (3E,5Z) and trans (3E,5E) isomers, respectively. In both the in vitro and in vivo turnovers, the enzyme showed a preference for the trans-pseudoionone isomer, which was consumed over the course of the turnover, whereas the cis was not.

A product from the CYP268A2 catalysed oxidation of 4-phenyltoluene was also identified by GC–MS (Supplementary Figure S10). This was assigned as 4-biphenylmethanol by MS comparison and GC co-elution experiment with other P450 turnovers [65]. Other substrates, including undecanoic acid, neryl acetate, farnesol, farnesyl acetate, and geraniol, were tried but no product was detected. The selectivity for trans- over cis-pseudoionone and geranyl over neryl acetate, which aligns with the binding data, indicates that the enzyme prefers the straight chain isomer over the bent form. The absence of product for geraniol suggests that substrates with the acetate ester moiety are favoured. The catalytic turnover rate reported is expected to be limited by the use of non-native electron transfer partners, which generally support lower levels of oxygenase activity [66]. However, the ability of Tdx to support CYP268A2 activity is a strong indication that the physiological electron transfer partner may be a [2Fe–2S] ferredoxin. Ideally, the native transfer partners of CYP268A2 from M. marinum would be identified. The availability of known substrates of CYP268A2 which have demonstrated product formation will facilitate this process.

Crystal structure of pseudoionone-bound CYP268A2

Crystallisation of both substrate-bound and substrate-free CYP268A2 was attempted. When pseudoionone was added before crystallisation, CYP268A2 formed sharp-edged single crystals after 2 weeks and diffraction data were collected to 2.0 Å at the Australian Synchrotron. No suitable crystals of the substrate-free form were obtained. The solved structure consists of a single polymer chain in the asymmetric unit and the trans (3E,5E) form of the substrate pseudoionone in the active site (PDB code: 6BLD). All residues were modelled except the first five of the N-terminus. Refinement statistics for the structure are located in Table 2. The overall fold conforms closely to the canonical P450 fold (Figure 7b). A loop of residues (Pro46 to Phe61) between two β-sheet regions forms an active site ‘cap’ similar to that of CYP124A1 [21]. The electron density map shows that the pseudoionone is arranged with one of the ω methyl groups held directly over the haem (4.3 Å away from the Fe atom). The carbonyl group of the substrate interacts with residues from the G helix and the B–C loop, towards the apparent exit of the active site (Figure 7a). The residues of the active site that interact directly with the carbonyl of pseudoionone, Gln209 and Ser99, sit close enough at 2.8 and 3.4 Å, respectively, to form hydrogen bonds (Figure 7c). This arrangement provides structural justification for the substrate-binding preference for oxygen-containing groups at one end of the substrate (acetates, acids, or alcohols). Geranyl acetate, undecanoic acid, and other similar substrates would, in theory, have a functional group that could also interact in a similar manner and position.

X-ray crystal structure of CYP268A2.

Figure 7.
X-ray crystal structure of CYP268A2.

(a) The overall structure of CYP268A2 (green) from M. marinum with pseudoionone bound (yellow), showing the composite omit 2Fo − Fc electron density at σ = 1.5 (blue mesh) of the pseudoionone in the active site and (b) CYP268A2 (green) with pseudoionone (yellow) bound overlaid with M. tb CYP124A1 (blue) with phytanic acid (orange) bound, showing the conserved P450 helices [55]. (c) The active site region of CYP268A2 showing side chains of amino acids within 5 Å of the pseudoionone molecule. There are 12 residues within 4 Å: S99, Q100, L103, L187, L205, Q209, L256, V259, A260, T264, F311, and F408. The terminal carbonyl of the pseudoionone has polar contacts (red) with the amine group of Q209 and the hydroxyl group of S99. (d) An overlay of the active site regions of CYP268A2 and CYP124A1 showing side chains of amino acids from both enzymes within 5 Å of the pseudoionone and phytanic acid molecules (CYP268A2 labels in black and CYP124A1 blue).

Figure 7.
X-ray crystal structure of CYP268A2.

(a) The overall structure of CYP268A2 (green) from M. marinum with pseudoionone bound (yellow), showing the composite omit 2Fo − Fc electron density at σ = 1.5 (blue mesh) of the pseudoionone in the active site and (b) CYP268A2 (green) with pseudoionone (yellow) bound overlaid with M. tb CYP124A1 (blue) with phytanic acid (orange) bound, showing the conserved P450 helices [55]. (c) The active site region of CYP268A2 showing side chains of amino acids within 5 Å of the pseudoionone molecule. There are 12 residues within 4 Å: S99, Q100, L103, L187, L205, Q209, L256, V259, A260, T264, F311, and F408. The terminal carbonyl of the pseudoionone has polar contacts (red) with the amine group of Q209 and the hydroxyl group of S99. (d) An overlay of the active site regions of CYP268A2 and CYP124A1 showing side chains of amino acids from both enzymes within 5 Å of the pseudoionone and phytanic acid molecules (CYP268A2 labels in black and CYP124A1 blue).

The pseudoionone molecule appears to be completely enclosed in the active site (Figure 8), suggesting that CYP268A2 has crystallised in the ‘closed form’ of the P450 [67]. Indeed, the enzyme completely encloses the substrate, showing no access to the substrate channel from the surface (Supplementary Figure S11a). The active site residues (12 within 4 Å of the substrate, Figure 7c) together create a linear substrate-binding pocket which seems likely to preferentially exclude the cis form of pseudoionone, which was also present in the crystallisation conditions but not observed binding in the solved structure. This further supports the product formation data, which have shown that the enzyme preferentially hydroxylates the trans isomer of pseudoionone at the ω terminus from the carbonyl group. The only significant vacant space in the active site is near the haem, where there is room to accommodate a larger group which would rationalise the binding of phenyltoluene-like molecules and the phenylimidazole inhibitors.

CYP268A2 (green) with the inner surface cavity of the enzyme active site (transparent grey) which encloses the pseudoionone substrate (yellow).

Figure 8.
CYP268A2 (green) with the inner surface cavity of the enzyme active site (transparent grey) which encloses the pseudoionone substrate (yellow).

There is a small empty cavity near the haem.

Figure 8.
CYP268A2 (green) with the inner surface cavity of the enzyme active site (transparent grey) which encloses the pseudoionone substrate (yellow).

There is a small empty cavity near the haem.

Two obvious water channels from the surface to the haem are observed, and these are similar to those found in other CYPs. One, referred to as the ‘water channel’, approaches the coordinating cysteine from the proximal face of the enzyme, beginning from the base of the B–C loop (Supplementary Figure S12), while the other, on the distal side between the E, F, and I helices, is the ‘solvent channel’ [68]. Residues of the I helix, including the acid Asp263, participate in the hydrogen-bonding network, by interacting directly with the water molecules that comprise the solvent channel (Supplementary Figure S13). The water channel is thought to be involved in active site solvation, and the solvent channel to be responsible for proton relay [68].

Comparison of the structure of the enzyme to the close relative CYP124A1 was made (Figure 7b). The active site of CYP124A1 accommodates larger substrates (∼C16), and a phytanic acid-bound structure is available (PDB code: 2WM4) [21]. This structure shows the substrate bound in a similar manner to that of pseudoionone close to the haem, but the carboxylate end of the phytanic acid curves into a pocket that is not available in the CYP268A2 enzyme, as a result of the presence of a tryptophan residue, Trp90 (Figure 7d). Both the Gln209 and Ser99 residues are not present in CYP124, replaced by a serine group (Ser216, which is shifted further away from the active site) and a phenylalanine (Phe107), respectively, neither of which interact electrostatically with the longer phytanic acid. Furthermore, the hydrophilic Gln100 is not present in CYP124A1 (replaced by Gly108) and the Thr264 (of the acid–alcohol pair) is flipped (Thr271 in CYP124A1). The position of Thr264 also affects the commonly conserved hydrogen bond between the alcohol and the nearby alanine carbonyl (Ala260 in CYP268A2 and Ala267 in CYP124A1), with the bond distance in CYP268A2 2.7 Å compared with 3.7 Å in CYP124A1 (Supplementary Figure S14). The position of the threonine in CYP268A2 matches the orientation of the equivalent residue in camphor-bound P450cam, which has a distance of 2.5 Å [69]. The tightness of this hydrogen bond is hypothesised to affect solvent access to the active site of the enzyme: in the closed position, solvent access is restricted, in the open form allowed [70].

The crystal structure of pseudoionone-bound CYP268A2 provides important insights into the substrate selectivity of these enzymes. The comparison with related CYP124A1 from M. tuberculosis provides an understanding on how these enzymes have evolved to modify their substrate selectivity. M. marinum also contains a gene encoding a CYP124A1 protein (84% similarity to CYP124A1 from M. tuberculosis), and both could have evolved from a common ancestor. At some point in time, the progenitors of the CYP268A2 and CYP124A1 enzymes may have had an overlapping function, providing genetic redundancy, and hence CYP268A2 may only be maintained in the larger genome of the less immuno-challenged organism. However, our data suggest that CYP268A2 would support a more varied range of substrate hydroxylation in the native system than CYP124A1. Along with other protein-encoding genes, the broad substrate range of the CYP268A2 system potentially aids M. marinum in surviving in more diverse environments.

Conclusion

The larger CYP complement of M. marinum contains several CYP families, including CYP268, which are widespread across Mycobacterium species but are absent in M. tuberculosis. The characterisation of CYP268A2 demonstrates that the selectivity and versatility of these enzymes can vary significantly even where there is structural and sequence similarity in the active site. CYP268A2 is the first member of this family to be characterised, and the substrate range of the enzyme reported here is broad with several molecules binding with high affinity. The successful hydroxylation of a long-chain branched acetate and ketone demonstrates that activity can be efficiently reconstituted with non-native electron transfer proteins including a [2Fe–2S] ferredoxin. This will assist with future attempts to identify the native electron transfer proteins. The substrate binding and turnover data presented here may be only the first of a broader range of molecules that can be oxidised by CYP268 family members. Furthermore, the crystallisation and structural analysis rationalises the observed catalytic activity and forms the basis for any future attempt to improve it. Finally, potential azole inhibitors were identified for the enzyme, which bind to the haem iron. As CYP268 family members are present across a range of Mycobacterium species including human pathogens, this could form the basis of future inhibitor design against bacterial infection.

Abbreviations

     
  • ArR

    a ferredoxin reductase from Novosphingobium aromaticivorans

  •  
  • CYP

    Cytochrome P450 enzyme

  •  
  • DTT

    dithiolthreitol

  •  
  • EMM

    E. coli minimal media

  •  
  • FAD

    flavin adenine dinucleotide

  •  
  • IPTG

    Isopropyl β-D-1-thiogalactopyranoside

  •  
  • JTT

    Jones–Taylor–Thornton

  •  
  • LB

    Lysogeny broth (also known as Luria or Lennox Broth)

  •  
  • MAC

    Mycobacterium avium complex

  •  
  • MTBC

    Mycobacterium tuberculosis complex

  •  
  • Tdx

    terpredoxin (a ferredoxin from Pseudomonas sp.)

  •  
  • PCR

    polymerase chain reaction

  •  
  • PDIM

    phthiocerol dimycocerosates

  •  
  • PE-PGRS

    proline glutamate-polymorphic GC-rich sequence proteins

Author Contribution

S.A.C., E.F.N., and S.G.B. performed molecular biology experiments. S.A.C. and J.B.B. determined the crystal structure. S.A.C. and S.G.B. wrote the manuscript.

Funding

This work was supported by the Australian Research Council through a Future Fellowship [FT140100355 to S.G.B.]. The authors also acknowledge the award of University of Adelaide Faculty of Sciences Divisional Scholarship (PhD to S.A.C.). X-ray diffraction data collection was undertaken on the MX1 beamline at the Australian Synchrotron, part of the Australian Nuclear Science and Technology Organisation. We acknowledge financial support from the Australian Synchrotron in the form of MXCAP9674.

Acknowledgments

The authors thank Santosh Panjikar for his help with onsite phasing at the Australian Synchrotron and Associate Professor Tara Pukala (University of Adelaide) for assistance with mass spectrometry. We acknowledge and are grateful to Professor Tim Stinear from the University of Melbourne and Professor Lalita Ramakrishnan of the University of Cambridge (formerly University of Washington) for providing the genomic DNA of M. marinum.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Ortiz de Montellano
,
P.R.
(
2015
) Substrate oxidation by cytochrome P450 enzymes. In
Cytochrome P450: Structure, Mechanism, and Biochemistry
(
Ortiz de Montellano
,
P.R.
, ed.), pp.
111
176
.
Springer International Publishing
2
Guengerich
,
F.P.
(
2015
) Human cytochrome P450 enzymes. In
Cytochrome P450: Structure, Mechanism, and Biochemistry
(
Ortiz de Montellano
,
P.R.
, ed.), pp.
523
786
.
Springer International Publishing
3
Nelson
,
D.R.
,
Zeldin
,
D.C.
,
Hoffman
,
S.M.G.
,
Maltais
,
L.J.
,
Wain
,
H.M.
and
Nebert
,
D.W.
(
2004
)
Comparison of cytochrome P450 (CYP) genes from the mouse and human genomes, including nomenclature recommendations for genes, pseudogenes and alternative-splice variants
.
Pharmacogenetics
14
,
1
18
4
Munro
,
A.W.
,
McLean
,
K.J.
,
Marshall
,
K.R.
,
Warman
,
A.J.
,
Lewis
,
G.
,
Roitel
,
O.
et al. 
(
2003
)
Cytochromes P450: novel drug targets in the war against multidrug-resistant Mycobacterium tuberculosis
.
Biochem. Soc. Trans.
31
,
625
630
5
Guengerich
,
F.P.
(
2001
)
Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity
.
Chem. Res. Toxicol.
14
,
611
650
6
Khatri
,
Y.
,
Girhard
,
M.
,
Romankiewicz
,
A.
,
Ringle
,
M.
,
Hannemann
,
F.
,
Urlacher
,
V.B.
et al. 
(
2010
)
Regioselective hydroxylation of norisoprenoids by CYP109D1 from Sorangium cellulosum So ce56
.
Appl. Microbiol. Biotechnol.
88
,
485
495
7
Scheps
,
D.
,
Malca
,
S.H.
,
Hoffmann
,
H.
,
Nestl
,
B.M.
and
Hauer
,
B.
(
2011
)
Regioselective ω-hydroxylation of medium-chain n-alkanes and primary alcohols by CYP153 enzymes from Mycobacterium marinum and Polaromonas sp. strain JS666
.
Org. Biomol. Chem.
9
,
6727
6733
8
Hannemann
,
F.
,
Bichet
,
A.
,
Ewen
,
K.M.
and
Bernhardt
,
R.
(
2007
)
Cytochrome P450 systems — biological variations of electron transport chains
.
Biochim. Biophys. Acta, Gen. Subj.
1770
,
330
344
9
Nelson
,
D.R.
,
Koymans
,
L.
,
Kamataki
,
T.
,
Stegeman
,
J.J.
,
Feyereisen
,
R.
,
Waxman
,
D.J.
et al. 
(
1996
)
P450 superfamily: update on new sequences, gene mapping, accession numbers and nomenclature
.
Pharmacogenetics
6
,
1
42
10
Cole
,
S.T.
,
Brosch
,
R.
,
Parkhill
,
J.
,
Garnier
,
T.
,
Churcher
,
C.
,
Harris
,
D.
et al. 
(
1998
)
Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence
.
Nature
393
,
537
544
11
Hudson
,
S.A.
,
McLean
,
K.J.
,
Munro
,
A.W.
and
Abell
,
C.
(
2012
)
Mycobacterium tuberculosis cytochrome P450 enzymes: a cohort of novel TB drug targets
.
Biochem. Soc. Trans.
40
,
573
579
12
Young
,
D.B.
and
Cole
,
S.T.
(
1993
)
Leprosy, tuberculosis, and the new genetics
.
J. Bacteriol.
175
,
1
6
13
Bloom
,
B.R.
and
Murray
,
C.J.L.
(
1992
)
Tuberculosis: commentary on a reemergent killer
.
Science
257
,
1055
1064
14
Stinear
,
T.P.
,
Seemann
,
T.
,
Harrison
,
P.F.
,
Jenkin
,
G.A.
,
Davies
,
J.K.
,
Johnson
,
P.D.R.
et al. 
(
2008
)
Insights from the complete genome sequence of Mycobacterium marinum on the evolution of Mycobacterium tuberculosis
.
Genome Res.
18
,
729
741
15
Demangel
,
C.
,
Stinear
,
T.P.
and
Cole
,
S.T.
(
2009
)
Buruli ulcer: reductive evolution enhances pathogenicity of Mycobacterium ulcerans
.
Nat. Rev. Microbiol.
7
,
50
60
16
Weerdenburg
,
E.M.
,
Abdallah
,
A.M.
,
Rangkuti
,
F.
,
AbdEl Ghany
,
M.
,
Otto
,
T.D.
,
Adroub
,
S.A.
et al. 
(
2015
)
Genome-wide transposon mutagenesis indicates that Mycobacterium marinum customizes its virulence mechanisms for survival and replication in different hosts
.
Infect. Immun.
83
,
1778
1788
17
Parvez
,
M.
,
Qhanya
,
L.B.
,
Mthakathi
,
N.T.
,
Kgosiemang
,
I.K.R.
,
Bamal
,
H.D.
,
Pagadala
,
N.S.
et al. 
(
2016
)
Molecular evolutionary dynamics of cytochrome P450 monooxygenases across kingdoms: special focus on mycobacterial P450s
.
Sci. Rep.
6
,
33099
18
Frank
,
D.J.
,
Zhao
,
Y.
,
Wong
,
S.H.
,
Basudhar
,
D.
,
De Voss
,
J.J.
and
Ortiz de Montellano
,
P.R.
(
2016
)
Cholesterol analogs with degradation-resistant alkyl side chains are effective Mycobacterium tuberculosis growth inhibitors
.
J. Biol. Chem.
291
,
7325
7333
19
Ouellet
,
H.
,
Chow
,
E.D.
,
Guan
,
S.
,
Cox
,
J.S.
,
Burlingame
,
A.L.
and
Ortiz de Montellano
,
P.R.
(
2013
)
Genetic and mass spectrometric tools for elucidating the physiological function(s) of cytochrome P450 enzymes from Mycobacterium tuberculosis
.
Methods Mol. Biol.
987
,
79
94
20
Barry
, III,
C.E.
,
Lee
,
R.E.
,
Mdluli
,
K.
,
Sampson
,
A.E.
,
Schroeder
,
B.G.
,
Slayden
,
R.A.
et al. 
(
1998
)
Mycolic acids: structure, biosynthesis and physiological functions
.
Prog. Lipid Res.
37
,
143
179
21
Johnston
,
J.B.
,
Kells
,
P.M.
,
Podust
,
L.M.
and
Ortiz de Montellano
,
P.R.
(
2009
)
Biochemical and structural characterization of CYP124: a methyl-branched lipid ω-hydroxylase from Mycobacterium tuberculosis
.
Proc. Natl Acad. Sci. U.S.A.
106
,
20687
20692
22
McLean
,
K.J.
and
Munro
,
A.W.
(
2008
)
Structural biology and biochemistry of cytochrome P450 systems in Mycobacterium tuberculosis
.
Drug Metab. Rev.
40
,
427
446
23
Forrellad
,
M.A.
,
Klepp
,
L.I.
,
Gioffre
,
A.
,
Sabio y Garcia
,
J.
,
Morbidoni
,
H.R.
,
de la Paz Santangelo
,
M.
et al. 
(
2013
)
Virulence factors of the Mycobacterium tuberculosis complex
.
Virulence
4
,
3
66
24
Jackson
,
C.J.
,
Lamb
,
D.C.
,
Marczylo
,
T.H.
,
Parker
,
J.E.
,
Manning
,
N.L.
,
Kelly
,
D.E.
et al. 
(
2003
)
Conservation and cloning of CYP51: a sterol 14α-demethylase from Mycobacterium smegmatis
.
Biochem. Biophys. Res. Commun.
301
,
558
563
25
Ouellet
,
H.
,
Johnston
,
J.B.
and
Ortiz de Montellano
,
P.R.
(
2010
)
The Mycobacterium tuberculosis cytochrome P450 system
.
Arch. Biochem. Biophys.
493
,
82
95
26
Poupin
,
P.
,
Ducrocq
,
V.
,
Hallier-Soulier
,
S.
and
Truffaut
,
N.
(
1999
)
Cloning and characterization of the genes encoding a cytochrome P450 (PipA) involved in piperidine and pyrrolidine utilization and its regulatory protein (PipR) in Mycobacterium smegmatis mc2155
.
J. Bacteriol.
181
,
3419
3426
PMID:
[PubMed]
27
Trigui
,
M.
,
Pulvin
,
S.
,
Truffaut
,
N.
,
Thomas
,
D.
and
Poupin
,
P.
(
2004
)
Molecular cloning, nucleotide sequencing and expression of genes encoding a cytochrome P450 system involved in secondary amine utilization in Mycobacterium sp. strain RP1
.
Res. Microbiol.
155
,
1
9
28
Brezna
,
B.
,
Kweon
,
O.
,
Stingley
,
R.L.
,
Freeman
,
J.P.
,
Khan
,
A.A.
,
Polek
,
B.
et al. 
(
2006
)
Molecular characterization of cytochrome P450 genes in the polycyclic aromatic hydrocarbon degrading Mycobacterium vanbaalenii PYR-1
.
Appl. Microbiol. Biotechnol.
71
,
522
532
29
de Marco
,
A.
,
Vigh
,
L.
,
Diamant
,
S.
and
Goloubinoff
,
P.
(
2005
)
Native folding of aggregation-prone recombinant proteins in Escherichia coli by osmolytes, plasmid- or benzyl alcohol-overexpressed molecular chaperones
.
Cell Stress Chaperones
10
,
329
339
30
Omura
,
T.
and
Sato
,
R.
(
1964
)
The carbon monoxide-binding pigment of liver microsomes. I. Evidence for its hemoprotein nature
.
J. Biol. Chem.
239
,
2370
2378
PMID:
[PubMed]
31
Pirotte
,
B.
,
Benoit
,
P.
,
Dogné
,
J.-M.
,
Masereel
,
B.
,
Rolin
,
S.
and
de Leval
,
X.
(
2002
)
Therapeutic potential of thromboxane inhibitors in asthma
.
Expert Opin. Investig. Drugs
11
,
275
281
32
Bell
,
S.G.
,
Dale
,
A.
,
Rees
,
N.H.
and
Wong
,
L.-L.
(
2010
)
A cytochrome P450 class I electron transfer system from Novosphingobium aromaticivorans
.
Appl. Microbiol. Biotechnol.
86
,
163
175
33
Bell
,
S.G.
,
Yang
,
W.
,
Yorke
,
J.A.
,
Zhou
,
W.
,
Wang
,
H.
,
Harmer
,
J.
et al. 
(
2012
)
Structure and function of CYP108D1 from Novosphingobium aromaticivorans DSM12444: an aromatic hydrocarbon-binding P450 enzyme
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
68
,
277
291
34
McPhillips
,
T.M.
,
McPhillips
,
S.E.
,
Chiu
,
H.J.
,
Cohen
,
A.E.
,
Deacon
,
A.M.
,
Ellis
,
P.J.
et al. 
(
2002
)
Blu-ice and the Distributed Control System: software for data acquisition and instrument control at macromolecular crystallography beamlines
.
J. Synchrotron Radiat.
9
,
401
406
35
Kabsch
,
W.
(
2010
)
XDS.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
125
132
36
Kabsch
,
W.
(
2010
)
Integration, scaling, space-group assignment and post-refinement
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
133
144
37
Panjikar
,
S.
,
Parthasarathy
,
V.
,
Lamzin
,
V.S.
,
Weiss
,
M.S.
and
Tucker
,
P.A.
(
2005
)
Auto-rickshaw: an automated crystal structure determination platform as an efficient tool for the validation of an X-ray diffraction experiment
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
61
,
449
457
38
Panjikar
,
S.
,
Parthasarathy
,
V.
,
Lamzin
,
V.S.
,
Weiss
,
M.S.
and
Tucker
,
P.A.
(
2009
)
On the combination of molecular replacement and single-wavelength anomalous diffraction phasing for automated structure determination
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
65
,
1089
1097
39
Winn
,
M.D.
,
Ballard
,
C.C.
,
Cowtan
,
K.D.
,
Dodson
,
E.J.
,
Emsley
,
P.
,
Evans
,
P.R.
et al. 
(
2011
)
Overview of the CCP4 suite and current developments
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
67
,
235
242
40
Brünger
,
A.T.
,
Adams
,
P.D.
,
Clore
,
G.M.
,
DeLano
,
W.L.
,
Gros
,
P.
,
Grosse-Kunstleve
,
R.W.
et al. 
(
1998
)
Crystallography and NMR system: a new software suite for macromolecular structure determination
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
54
,
905
921
41
Brunger
,
A.T.
(
2007
)
Version 1.2 of the crystallography and NMR system
.
Nat. Protoc.
2
,
2728
2733
42
Murshudov
,
G.N.
,
Skubák
,
P.
,
Lebedev
,
A.A.
,
Pannu
,
N.S.
,
Steiner
,
R.A.
,
Nicholls
,
R.A.
et al. 
(
2011
)
REFMAC5 for the refinement of macromolecular crystal structures
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
67
,
355
367
43
Cowtan
,
K.
(
2000
)
General quadratic functions in real and reciprocal space and their application to likelihood phasing
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
56
,
1612
1621
44
Sheldrick
,
G.M.
(
2002
)
Macromolecular phasing with SHELXE
.
Z. Kristallogr.
217
,
644
650
45
Terwilliger
,
T.C.
(
2000
)
Maximum-likelihood density modification
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
56
,
965
972
46
Cowtan
,
K.
(
2006
)
The Buccaneer software for automated model building. 1. Tracing protein chains
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
62
,
1002
1011
47
Adams
,
P.D.
,
Afonine
,
P.V.
,
Bunkóczi
,
G.
,
Chen
,
V.B.
,
Davis
,
I.W.
,
Echols
,
N.
et al. 
(
2010
)
PHENIX: a comprehensive Python-based system for macromolecular structure solution
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
213
221
48
Emsley
,
P.
,
Lohkamp
,
B.
,
Scott
,
W.G.
and
Cowtan
,
K.
(
2010
)
Features and development of Coot
.
Acta Crystallogr., Sect. D: Biol. Crystallogr.
66
,
486
501
49
Yoshida
,
M.
,
Nakanaga
,
K.
,
Ogura
,
Y.
,
Toyoda
,
A.
,
Ooka
,
T.
,
Kazumi
,
Y.
et al. 
(
2016
)
Complete genome sequence of Mycobacterium ulcerans subsp. shinshuense
.
Genome Announcements
4
,
e01050-16
50
Riojas
,
M.A.
,
McGough
,
K.
and
Hazbon
,
M.H.
NGS-Based Phylogenomic Analysis Supports Reclassification of All Species Within the Mycobacterium tuberculosis Complex as Mycobacterium Tuberculosis
.
Bei Resources
. https://www.beiresources.org/Portals/2/PDFS/NGS-Based%20Phylogenomic%20Analysis%20of%20MTB%20Complex.pdf
51
Maya-Hoyos
,
M.
,
Leguizamón
,
J.
,
Mariño-Ramírez
,
L.
and
Soto
,
C.Y.
(
2015
)
Sliding motility, biofilm formation, and glycopeptidolipid production in Mycobacterium colombiense strains
.
BioMed Res. Int.
2015
,
1
11
52
Vasilevskaya
,
A.V.
,
Yantsevich
,
A.V.
,
Sergeev
,
G.V.
,
Lemish
,
A.P.
,
Usanov
,
S.A.
and
Gilep
,
A.A.
(
2017
)
Identification of Mycobacterium tuberculosis enzyme involved in vitamin D and 7-dehydrocholesterol metabolism
.
J. Steroid Biochem. Mol. Biol.
169
,
202
209
53
Frank
,
D.J.
,
Madrona
,
Y.
and
Ortizde Montellano
,
P. R.
(
2014
)
Cholesterol ester oxidation by mycobacterial cytochrome P450
.
J. Biol. Chem.
289
,
30417
30425
54
Capyk
,
J.K.
,
Kalscheuer
,
R.
,
Stewart
,
G.R.
,
Liu
,
J.
,
Kwon
,
H.
,
Zhao
,
R.
et al. 
(
2009
)
Mycobacterial cytochrome p450 125 (cyp125) catalyzes the terminal hydroxylation of c27 steroids
.
J. Biol. Chem.
284
,
35534
35542
55
Poulos
,
T.L.
,
Finzel
,
B.C.
,
Gunsalus
,
I.C.
,
Wagner
,
G.C.
and
Kraut
,
J.
(
1985
)
The 2.6-Å crystal structure of Pseudomonas putida cytochrome P-450
.
J. Biol. Chem.
260
,
16122
16130
PMID:
[PubMed]
56
Banu
,
S.
,
Honoré
,
N.
,
Saint-Joanis
,
B.
,
Philpott
,
D.
,
Prévost
,
M.-C.
and
Cole
,
S.T.
(
2002
)
Are the PE-PGRS proteins of Mycobacterium tuberculosis variable surface antigens
?
Mol. Microbiol.
44
,
9
19
57
Espitia
,
C.
,
Laclette
,
J.P.
,
Mondragόn-Palomino
,
M.
,
Amador
,
A.
,
Campuzano
,
J.
,
Martens
,
A.
et al. 
(
1999
)
The PE-PGRS glycine-rich proteins of Mycobacterium tuberculosis: a new family of fibronectin-binding proteins?
Microbiology
145
(
Pt 12
),
3487
3495
58
Omura
,
T.
and
Sato
,
R.
(
1964
)
The carbon monoxide-binding pigment of liver microsomes. Ii. Solubilization, purification, and properties
.
J. Biol. Chem.
239
,
2379
2385
PMID:
[PubMed]
59
Driscoll
,
M.D.
,
McLean
,
K.J.
,
Levy
,
C.
,
Mast
,
N.
,
Pikuleva
,
I.A.
,
Lafite
,
P.
et al. 
(
2010
)
Structural and biochemical characterization of Mycobacterium tuberculosis CYP142: evidence for multiple cholesterol 27-hydroxylase activities in a human pathogen
.
J. Biol. Chem.
285
,
38270
38282
60
Frank
,
D.J.
,
Waddling
,
C.A.
,
La
,
M.
and
Ortizde de Montellano
,
P.R.
(
2015
)
Cytochrome P450 125A4, the third cholesterol C-26 hydroxylase from Mycobacterium smegmatis
.
Biochemistry
54
,
6909
6916
61
Ouellet
,
H.
,
Podust
,
L.M.
and
Ortiz de Montellano
,
P.R.
(
2008
)
Mycobacterium tuberculosis CYP130: crystal structure, biophysical characterization, and interactions with antifungal azole drugs
.
J. Biol. Chem.
283
,
5069
5080
62
McLean
,
K.J.
,
Dunford
,
A.J.
,
Neeli
,
R.
,
Driscoll
,
M.D.
and
Munro
,
A.W.
(
2007
)
Structure, function and drug targeting in Mycobacterium tuberculosis cytochrome P450 systems
.
Arch. Biochem. Biophys.
464
,
228
240
63
Mueller
,
N.J.
,
Stueckler
,
C.
,
Hauer
,
B.
,
Baudendistel
,
N.
,
Housden
,
H.
,
Bruce
,
N.C.
et al. 
(
2010
)
The substrate spectra of pentaerythritol tetranitrate reductase, morphinone reductase, N-ethylmaleimide reductase and estrogen-binding protein in the asymmetric bioreduction of activated alkenes
.
Adv. Synth. Catal.
352
,
387
394
64
Grau
,
M.M.
,
van der Toorn
,
J.C.
,
Otten
,
L.G.
,
Macheroux
,
P.
,
Taglieber
,
A.
,
Zilly
,
F.E.
et al. 
(
2009
)
Photoenzymatic reduction of C=C double bonds
.
Adv. Synth. Catal.
351
,
3279
3286
65
Hall
,
E.A.
,
Sarkar
,
M.R.
and
Bell
,
S.G.
(
2017
)
The selective oxidation of substituted aromatic hydrocarbons and the observation of uncoupling via redox cycling during naphthalene oxidation by the CYP101B1 system
.
Catal. Sci. Technol.
7
,
1537
1548
66
Yang
,
W.
,
Bell
,
S.G.
,
Wang
,
H.
,
Zhou
,
W.
,
Hoskins
,
N.
,
Dale
,
A.
et al. 
(
2010
)
Molecular characterization of a class I P450 electron transfer system from Novosphingobium aromaticivorans DSM12444
.
J. Biol. Chem.
285
,
27372
27384
67
Poulos
,
T.L.
,
Finzel
,
B.C.
and
Howard
,
A.J.
(
1986
)
Crystal structure of substrate-free Pseudomonas putida cytochrome P-450
.
Biochemistry
25
,
5314
5322
68
Cojocaru
,
V.
,
Winn
,
P.J.
and
Wade
,
R.C.
(
2007
)
The ins and outs of cytochrome P450s
.
Biochim. Biophys. Acta, Gen. Subj.
1770
,
390
401
69
Poulos
,
T.L.
,
Finzel
,
B.C.
and
Howard
,
A.J.
(
1987
)
High-resolution crystal structure of cytochrome P450cam
.
J. Mol. Biol.
195
,
687
700
70
Poulos
,
T.L.
(
2014
)
Heme enzyme structure and function
.
Chem. Rev.
114
,
3919
3962