Crystal Structure of a Putative Cytochrome P450 Alkane Hydroxylase (CYP153D17) from Sphingomonas sp. PAMC 26605 and Its Conformational Substrate Binding

Enzymatic alkane hydroxylation reactions are useful for producing pharmaceutical and agricultural chemical intermediates from hydrocarbons. Several cytochrome P450 enzymes catalyze the regio- and stereo-specific hydroxylation of alkanes. We evaluated the substrate binding of a putative CYP alkane hydroxylase (CYP153D17) from the bacterium Sphingomonas sp. PAMC 26605. Substrate affinities to C10–C12 n-alkanes and C10–C14 fatty acids with Kd values varied from 0.42 to 0.59 μM. A longer alkane (C12) bound more strongly than a shorter alkane (C10), while shorter fatty acids (C10, capric acid; C12, lauric acid) bound more strongly than a longer fatty acid (C14, myristic acid). These data displayed a broad substrate specificity of CYP153D17, hence it was named as a putative CYP alkane hydroxylase. Moreover, the crystal structure of CYP153D17 was determined at 3.1 Å resolution. This is the first study to provide structural information for the CYP153D family. Structural analysis showed that a co-purified alkane-like compound bound near the active-site heme group. The alkane-like substrate is in the hydrophobic pocket containing Thr74, Met90, Ala175, Ile240, Leu241, Val244, Leu292, Met295, and Phe393. Comparison with other CYP structures suggested that conformational changes in the β1–β2, α3–α4, and α6–α7 connecting loop are important for incorporating the long hydrophobic alkane-like substrate. These results improve the understanding of the catalytic mechanism of CYP153D17 and provide valuable information for future protein engineering studies.


Introduction
Cytochrome P450 enzymes (CYPs) are heme-containing proteins that catalyze the insertion of oxygen into a wide variety of substrates [1,2]. Although many different CYPs have been studied, alkane hydroxylases are among the most widely identified and characterized [3]. Hydroxylation of alkanes is important for their degradation or further metabolism, and chemical synthesis of hydroxylated alkane 2 of 12 molecules is very difficult and costly [4]. Recently, CYP alkane hydroxylases have gained attention because of their commercial applications [5].
The CYP2E, CYP4B, CYP52, CYP86, and CYP153 families are known to hydroxylate n-alkanes or fatty acids in both eukaryotes and prokaryotes [6][7][8][9][10]. Among them, CYP2E, CYP4B, CYP52, and CYP86 are eukaryotic CYPs; these enzymes are membrane-bound proteins, limiting their structural characterization and making it difficult to use them in engineered pathways. Several bacterial CYPs in the CYP153 family have been characterized [11,12]. In 2012, the crystal structure of CYP153A7 (P450pyr) was determined [13]. In addition, researchers have attempted to engineer enzymes from the CYP101 and CYP102 families for alkane hydroxylation. For example, P450 BM-3 from Bacillus megaterium, a well-characterized CYP102 enzyme that functions as a fatty acid monooxygenase, has been used for this purpose [14]. To create useful oxidation biocatalysts for alkane compounds, P450 BM-3 enzyme was engineered to hydroxylate linear alkanes by directed evolution with mutagenesis [15]. In 2016, the structure of CYP153A from Marinobacter aquaeolei (CYP153A M.aq ) was solved with and without its 12-hydroxydodecanoic acid ligand. The authors found that CYP153A M.aq has terminal hydroxylase activity and that the A231G single mutation induced three-fold higher activity by increasing the flexibility of the substrate binding region [16]. Previous studies also demonstrated that Sphingomonas sp. HXN-200 bacteria contain hydrocarbon-degrading enzymes [17,18]. The results of these studies suggest that Sphingomonas sp. possesses CYP alkane hydroxylases with high alkane hydroxylation activity.
In the current study, we evaluated CYP153D17 (NCBI Reference Sequence, WP_010183267.1) from the endosymbiotic bacterium Sphingomonas sp. strain PAMC 26605. This bacterium was isolated from Arctic lichens found on the Svalbard Islands, Norway [19]. In addition, we determined the crystal structure of CYP153D17 at 3.1 Å resolution in complex with a co-purified alkane-like compound. Comparison of homologous CYP structures revealed a unique rearrangement of the active-site loop regions for adaptation of the long hydrophobic alkane-like compound near the heme iron in the active site. Here, we describe the structural determination of CYP153D17 and comparison of this structure with structural and sequential counterpart CYPs to improve understanding of the structural basis of conformational changes due to substrate binding.

Sequence Analysis, Purification, and Spectral Analysis of CYP153D17
CYP153D17 encodes a protein of 409 amino acids with an overall G + C content of 63.41%. Sequence analysis revealed that CYP153D17 has 48.27% identity with P450pyr hydroxylase from Sphingopyxis macrogoltabida and that it contains conserved motifs including an oxygen-binding domain and heme-binding domains found in typical CYPs [20]. When over-expressed heterologously in Escherichia coli C41 (DE3), as previously reported by our group [21], high levels of folded CYP153D17 were obtained. After purification by Ni 2+ -affinity chromatography, pure CYP153D17 was concentrated to 215.29 mg/mL. The molecular mass of CYP153D17 (48 kDa) was confirmed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis and was consistent with the predicted molecular mass ( Figure 1A). The ferric to ferrous reduction of CYP153D17 was achieved by the addition of sodium dithionite, and carbon monoxide was gassed through the sample tube. A shift of approximately 30 nm (from 420 to 450 nm) was observed in CYP153D17, indicating that the distribution of electron density on the heme was significantly perturbed ( Figure 1B).

Substrate Binding Assay of CYP153D17
To determine the binding affinity of CYP153D17, spectral binding titrations were performed with the alkanes (decane and dodecane) and fatty acids (capric acid, lauric acid, and myristic acid). All tested substrates changed the spectrum of CYP153D17, with a shift of the Soret band at 420 to 390 nm, which is typical of the high-spin form (Figure 2), after which the apparent dissociation constants from the titration curves (Kd values) were obtained (Table 1). Dodecane, capric acid, and lauric acid binding to CYP153D17 induced a ≥90% shift in the heme spin state from low-spin (Soret maximum at 420 nm) to high-spin (Soret maximum at 390 nm), whereas decane and myristic acid induced a ≥80% shift in the same heme spin state. In addition, the Kd values were 0.59 ± 0.69, 0.42 ± 0.14, 0.46 ± 0.41, 0.49 ± 0.70, and 0.54 ± 0.33 μM for decane, dodecane, capric acid, lauric acid, and myristic acid, respectively. Interestingly, although CYP153D17 showed different substrate affinities against C10-C12 alkanes and C10-C14 fatty acids, C12 alkane (dodecane) strongly bound to CYP153D17 with similar Kd values of C10-C12 fatty acids (capric acid and lauric acid).

Substrate Binding Assay of CYP153D17
To determine the binding affinity of CYP153D17, spectral binding titrations were performed with the alkanes (decane and dodecane) and fatty acids (capric acid, lauric acid, and myristic acid). All tested substrates changed the spectrum of CYP153D17, with a shift of the Soret band at 420 to 390 nm, which is typical of the high-spin form (Figure 2), after which the apparent dissociation constants from the titration curves (K d values) were obtained (Table 1). Dodecane, capric acid, and lauric acid binding to CYP153D17 induced a ≥90% shift in the heme spin state from low-spin (Soret maximum at 420 nm) to high-spin (Soret maximum at 390 nm), whereas decane and myristic acid induced a ≥80% shift in the same heme spin state. In addition, the K d values were 0.59 ± 0.69, 0.42 ± 0.14, 0.46 ± 0.41, 0.49 ± 0.70, and 0.54 ± 0.33 µM for decane, dodecane, capric acid, lauric acid, and myristic acid, respectively. Interestingly, although CYP153D17 showed different substrate affinities against C10-C12 alkanes and C10-C14 fatty acids, C12 alkane (dodecane) strongly bound to CYP153D17 with similar K d values of C10-C12 fatty acids (capric acid and lauric acid).

Substrate Binding Assay of CYP153D17
To determine the binding affinity of CYP153D17, spectral binding titrations were performed with the alkanes (decane and dodecane) and fatty acids (capric acid, lauric acid, and myristic acid). All tested substrates changed the spectrum of CYP153D17, with a shift of the Soret band at 420 to 390 nm, which is typical of the high-spin form (Figure 2), after which the apparent dissociation constants from the titration curves (Kd values) were obtained (Table 1). Dodecane, capric acid, and lauric acid binding to CYP153D17 induced a ≥90% shift in the heme spin state from low-spin (Soret maximum at 420 nm) to high-spin (Soret maximum at 390 nm), whereas decane and myristic acid induced a ≥80% shift in the same heme spin state. In addition, the Kd values were 0.59 ± 0.69, 0.42 ± 0.14, 0.46 ± 0.41, 0.49 ± 0.70, and 0.54 ± 0.33 μM for decane, dodecane, capric acid, lauric acid, and myristic acid, respectively. Interestingly, although CYP153D17 showed different substrate affinities against C10-C12 alkanes and C10-C14 fatty acids, C12 alkane (dodecane) strongly bound to CYP153D17 with similar Kd values of C10-C12 fatty acids (capric acid and lauric acid).

Overall Structure of CYP153D17
The crystal structure of CYP153D17 from Sphingomonas sp. strain PAMC 26605 was determined by molecular replacement (MR) performed using the P450pyr hydroxylase structure from S. macrogoltabida (Protein Data Bank code, 3RWL) as a search model. The structure was refined to a resolution of 3.1 Å ( Table 2) and had a typical CYP fold, comprising a central heme molecule surrounded by 13 α-helices and 7 β-strands ( Figure 3). The putative substrate-binding pocket of CYP153D17 was found to be shaped similar to a cone funnel with a length and outer diameter of approximately 17 and 10 Å, respectively ( Figure 4A). Notably, an additional electron density map appeared within the substrate-binding pocket during refinement. Because exogenous ligands were not added to the crystallization solutions, the bound molecule may have co-purified with the enzyme during the production of recombinant protein. The resulting electron density map revealed the presence of a long fatty acid or alkane-like compound. Substrate binding assay data suggested that the ligand was an alkane-like compound and the electron density was modeled as dodecane. Although dodecane was modeled as a putative ligand bound to CYP153D17, the actual substrate was not determined. The co-purified ligand was bound to the hydrophobic substrate-binding pocket created by the Met90, Ala175, Ile240, Leu241, Val244, Leu292, Met295, and Phe393 residues ( Figure 4B).

Overall Structure of CYP153D17
The crystal structure of CYP153D17 from Sphingomonas sp. strain PAMC 26605 was determined by molecular replacement (MR) performed using the P450pyr hydroxylase structure from S. macrogoltabida (Protein Data Bank code, 3RWL) as a search model. The structure was refined to a resolution of 3.1 Å ( Table 2) and had a typical CYP fold, comprising a central heme molecule surrounded by 13 α-helices and 7 β-strands ( Figure 3). The putative substrate-binding pocket of CYP153D17 was found to be shaped similar to a cone funnel with a length and outer diameter of approximately 17 and 10 Å, respectively ( Figure 4A). Notably, an additional electron density map appeared within the substrate-binding pocket during refinement. Because exogenous ligands were not added to the crystallization solutions, the bound molecule may have co-purified with the enzyme during the production of recombinant protein. The resulting electron density map revealed the presence of a long fatty acid or alkane-like compound. Substrate binding assay data suggested that the ligand was an alkane-like compound and the electron density was modeled as dodecane. Although dodecane was modeled as a putative ligand bound to CYP153D17, the actual substrate was not determined. The co-purified ligand was bound to the hydrophobic substrate-binding pocket created by the Met90, Ala175, Ile240, Leu241, Val244, Leu292, Met295, and Phe393 residues ( Figure 4B).    The residues involved in binding of the co-purified ligand in CYP153D17 are indicated by a black circle above the sequence alignment, and those involved in phytanic acid binding in the CYP124 structure, by a black triangle below the sequence alignment. Multiple sequence alignment was performed using ClustalX (version 1.81) [22] and edited using GeneDoc (Ver 2.5.000) [23]. . Highly conserved residues are shaded in black and similar residues are in gray. The secondary structural elements in the crystal structure of CYP153D17 are shown above the multiple sequence alignment. The residues involved in binding of the co-purified ligand in CYP153D17 are indicated by a black circle above the sequence alignment, and those involved in phytanic acid binding in the CYP124 structure, by a black triangle below the sequence alignment. Multiple sequence alignment was performed using ClustalX (version 1.81) [22] and edited using GeneDoc (Ver 2.5.000) [23]. Next, a structural homology search was performed using the DALI server [24]. The top six unique hits were identified and are shown in Table 3. The results showed that CYP153D17 had the highest degree of structural similarity with P450pyr (PDB code, 3RWL) [13] and that CYP153D17 shared structural similarity with CYP124 (PDB code, 2WM4) [25], CYP142 (PDB code, 2XKR) [26], P450terp (PDB code, 1CPT) [27], CYP125 (PDB code, 2X5W) [28], and CYP108D1 (PDB code, 3TKT) [29]. Sequence alignments of these proteins revealed that several residues constituting the hydrophobic substrate-binding pocket of CYP153D17 were highly conserved. Specifically, Ile240, Leu292, and Phe393 were found to be highly conserved, while other residues (Met90, Ala175, Leu241, Val244, and Met295) were less conserved ( Figure 3B).

Structural Comparisons of CYP153D17 with Other CYPs
The overall protein structure of P450pyr (best hit in the DALI search) was very similar to that of CYP153D17 (r.m.s. deviation, 0.96 Å for 307 Cα atoms), but P450pyr contained no ligand and the loop region (residues 88-96 in P450pyr) between the α3 and α4-helix was missing in the P450pyr structure ( Figure 5). This loop may be more flexible in the absence of substrate in P450pyr. Structural comparison of CYP124 (second hit in the DALI search) with CYP153D17 showed overall structural similarity in the core region structure (r.m.s. deviation, 1.24 Å for 327 Cα atoms), whereas considerable local conformational differences were observed in the β1-β2, α3-α4, and α6-α7 connecting loop regions (in secondary structure annotation of CYP153D17). The CYP124 structure was complexed with phytanic acid, with hydrophobic interactions involving Leu60, Phe63, Asn97, Phe107, Leu198, Phe212, Val266, and Val315 within 4 Å distance. The mapping of interaction residues onto the sequence alignment showed that the substrate-binding residues differed considerably between CYP153D17 and CYP124 ( Figure 3B). Structural superposition revealed that the binding mode of phytanic acid was similar to that of the co-purified ligand of CYP153D17 as an extended conformation. This finding suggests mechanical-reaction-related similarities between CYP153D17 Next, a structural homology search was performed using the DALI server [24]. The top six unique hits were identified and are shown in Table 3. The results showed that CYP153D17 had the highest degree of structural similarity with P450pyr (PDB code, 3RWL) [13] and that CYP153D17 shared structural similarity with CYP124 (PDB code, 2WM4) [25], CYP142 (PDB code, 2XKR) [26], P450terp (PDB code, 1CPT) [27], CYP125 (PDB code, 2X5W) [28], and CYP108D1 (PDB code, 3TKT) [29]. Sequence alignments of these proteins revealed that several residues constituting the hydrophobic substrate-binding pocket of CYP153D17 were highly conserved. Specifically, Ile240, Leu292, and Phe393 were found to be highly conserved, while other residues (Met90, Ala175, Leu241, Val244, and Met295) were less conserved ( Figure 3B).

Structural Comparisons of CYP153D17 with Other CYPs
The overall protein structure of P450pyr (best hit in the DALI search) was very similar to that of CYP153D17 (r.m.s. deviation, 0.96 Å for 307 Cα atoms), but P450pyr contained no ligand and the loop region (residues 88-96 in P450pyr) between the α3 and α4-helix was missing in the P450pyr structure ( Figure 5). This loop may be more flexible in the absence of substrate in P450pyr. Structural comparison of CYP124 (second hit in the DALI search) with CYP153D17 showed overall structural similarity in the core region structure (r.m.s. deviation, 1.24 Å for 327 Cα atoms), whereas considerable local conformational differences were observed in the β1-β2, α3-α4, and α6-α7 connecting loop regions (in secondary structure annotation of CYP153D17). The CYP124 structure was complexed with phytanic acid, with hydrophobic interactions involving Leu60, Phe63, Asn97, Phe107, Leu198, Phe212, Val266, and Val315 within 4 Å distance. The mapping of interaction residues onto the sequence alignment showed that the substrate-binding residues differed considerably between CYP153D17 and CYP124 ( Figure 3B). Structural superposition revealed that the binding mode of phytanic acid was similar to that of the co-purified ligand of CYP153D17 as an extended conformation. This finding suggests mechanical-reaction-related similarities between CYP153D17 and CYP124. CYP124 catalyzes ω-oxidation of phytanic acid; that is, the terminal methyl group of phytanic acid is specifically hydroxylated. Therefore, it is expected that for CYP153D17, hydroxylation may also be performed at the end of alkane molecules, such as in CYP124-mediated ω-oxidation. and CYP124. CYP124 catalyzes ω-oxidation of phytanic acid; that is, the terminal methyl group of phytanic acid is specifically hydroxylated. Therefore, it is expected that for CYP153D17, hydroxylation may also be performed at the end of alkane molecules, such as in CYP124-mediated ω-oxidation.
For overexpression, E. coli transformants harboring pN153D17 were cultured in LB (Luria-Bertani) medium with an appropriate amount of kanamycin, and the culture was grown at 37 °C until cell density reached 0.6-0.8 at OD600. Next, 1 mM 5-α-aminolevulinic acid and 0.5 mM FeCl3 were added to the culture. After 30-min incubation at 20 °C, protein overexpression was induced by adding 0.5 mM IPTG (isopropyl-β-D-1-thiogalactopyranoside) followed by incubation for 70 h at 20°C. Escherichia coli cultures were harvested and washed twice with 50 mM Tris-HCl buffer with 10% glycerol (pH 7.4), and finally lysed by an ultrasonicator. After 12,000-rpm centrifugation for 25 min at 4 °C, the obtained soluble proteins were purified by metal affinity with Ni 2+ -NTA resin (Qiagen, Hilden, Germany). The eluted CYP153D17 protein was concentrated using Amicon Ultra centrifugal filters (Ultracel-3K; Millipore, Billerica, MA, USA) and confirmed by 15% SDS-PAGE analysis. After the poly-His tag was cleaved by thrombin by overnight incubation at 4 °C, CYP153D17 was loaded onto a Superdex-200 column (GE Healthcare, Little Chalfont, UK) that had been equilibrated with 50 mM potassium phosphate and 100 mM sodium chloride, pH 7.4. The fractions containing red-colored CYP153D17 were collected and concentrated to 215.29 mg/mL by centrifugal filtration.
For overexpression, E. coli transformants harboring pN153D17 were cultured in LB (Luria-Bertani) medium with an appropriate amount of kanamycin, and the culture was grown at 37 • C until cell density reached 0.6-0.8 at OD 600 . Next, 1 mM 5-α-aminolevulinic acid and 0.5 mM FeCl 3 were added to the culture. After 30-min incubation at 20 • C, protein overexpression was induced by adding 0.5 mM IPTG (isopropyl-β-D-1-thiogalactopyranoside) followed by incubation for 70 h at 20 • C. Escherichia coli cultures were harvested and washed twice with 50 mM Tris-HCl buffer with 10% glycerol (pH 7.4), and finally lysed by an ultrasonicator. After 12,000-rpm centrifugation for 25 min at 4 • C, the obtained soluble proteins were purified by metal affinity with Ni 2+ -NTA resin (Qiagen, Hilden, Germany). The eluted CYP153D17 protein was concentrated using Amicon Ultra centrifugal filters (Ultracel-3K; Millipore, Billerica, MA, USA) and confirmed by 15% SDS-PAGE analysis. After the poly-His tag was cleaved by thrombin by overnight incubation at 4 • C, CYP153D17 was loaded onto a Superdex-200 column (GE Healthcare, Little Chalfont, UK) that had been equilibrated with 50 mM potassium phosphate and 100 mM sodium chloride, pH 7.4. The fractions containing red-colored CYP153D17 were collected and concentrated to 215.29 mg/mL by centrifugal filtration.

Substrate Binding Assay
To determine whether CYP153D17 can bind alkanes and fatty acids, a UV-Vis (Ultraviolet-Visible) spectral binding titration assay was performed. Binding of the substrate to CYP153D17 induced a shift of the Soret band of the CYP absorption spectrum from A 420 to A 390 [30]. Alkanes (decane and dodecane) and fatty acids (capric acid, lauric acid, and myristic acid) were dissolved in ethanol to prepare stock solutions of different concentrations (0.4-5 mM) for each substrate. Binding of substrate to CYP153D17 was carried out by the consecutive addition of 1-µL aliquots from the stock solutions to 1 mL of 1 µM CYP153D17 solution, and the UV-Vis spectra of the mixture were recorded. To calculate K d , the difference between adsorption at A 390 and A 420 was calculated, corrected with the A 390 -to-A 420 value at 0 M substrate, and plotted against the substrate concentrations. The data were fitted to equations as described by Funhoff et al. [31]. To obtain optimal crystals, the crystallization condition of 25% (w/v) PEG 3350 (SG-1 #E2) was varied according to the precipitant concentration by using the hanging-drop vapor-diffusion method in 24-well crystallization plates (Molecular Dimensions) at 293 K. The drop volume was increased to 1 µL protein and 1 µL reservoir solution against 500 µL reservoir solution. An optimal single crystal was obtained using 22% (w/v) PEG 3350 and had a rectangular stick shape. The single crystal was picked using a loop and mounted without cryoprotectant. Diffraction data were collected using a synchrotron-radiation source at the BL-7A beam line, Pohang Accelerator Laboratory, Pohang, Korea.

Structure Determination and Refinement
The 3.1 Å resolution data set contained 100 images with a detector-to-crystal distance of 300 mm, exposure time of 1 s, and oscillation range of 1 • per image. The data set was indexed, integrated, and scaled using HKL-2000 software [32]. The data collection statistics are summarized in Table 2. For structure determination of CYP153D17, MR was performed using the program MOLREP in the CCP4 suite [33,34]. The coordinate of P450pyr hydroxylase from S. macrogoltabida (PDB code, 3RWL) was used for MR searching. The structure model of CYP153D17 was iteratively rebuilt and refined by manually fitting the residues and backbone into the electron density map by using the WinCoot and REFMACK programs from the CCP4 suite [35,36]. The quality of the refined structure was validated using the MOLPROBITY server and SFCHECK with an R work of 26.25% and R free of 34.07% [37,38]. The statistical data for structure refinement are provided in Table 2.

Conclusions
As major components of natural gas and petroleum, alkanes play important roles as energy and sole carbon sources in modern human life. However, because of their solid viscosity and inertness in the soil and water bodies resulting from accidental oil spills and leakage, many scientists have recently considered alkane-degrading microorganisms with various uptake mechanisms and enzyme systems to be a solution to these issues [4,5]. One of these enzyme systems, alkane hydroxylase, is quite efficient in oxidizing various substrates in industrial applications. In this context, the chemistry of bacterial alkane hydroxylase is interesting and rare. Among the structurally available bacterial alkane hydroxylases, P450pyr is the only member of the CYP153A family with alkane function [13], and this is the first study to determine structural information for the CYP153D family of alkane hydroxylases. In this study, we identified CYP153D17 from the endosymbiotic bacterium Sphingomonas sp. PAMC 26605 as a putative alkane-hydroxylating cytochrome P450 enzyme based on our substrate binding assay and structural data. However, we cannot exclude the possibility that CYP153D17 may have also fatty-acid hydroxylation activity considering its broad substrate specificity. Thus, further investigations are needed to identify true intrinsic ligands of CYP153D17. Nevertheless, our substrate binding assay and combined structural studies provide several insights into the molecular basis of substrate binding by CYP153D17 and its regio-and stereo-selectivity towards various substrates containing alkanes and fatty acids. In addition, our results will facilitate future studies identifying active-site residues and investigating their roles in different substrate (alkane and fatty acid) specificities. Our results may also provide a structural basis for protein engineering.