Functional characterization of cytochromes P450 2B from the desert woodrat Neotoma lepida

Abstract

Mammalian detoxification processes have been the focus of intense research, but little is known about how wild herbivores process plant secondary compounds, many of which have medicinal value or are drugs. cDNA sequences that code for three enzymes of the cytochrome P450 (CYP) 2B subfamily, here termed 2B35, 2B36, and 2B37 have been recently identified from a wild rodent, the desert woodrat (Malenke et al., 2012). Two variant clones of each enzyme were engineered to increase protein solubility and to facilitate purification, as reported for CYP2B enzymes from multiple species. When expressed in E. coli each of the woodrat proteins gave the characteristic maximum at 450 nm in a reduced carbon monoxide difference spectrum but generally expressed at lower levels than rat CYP2B1. Two enzymes, 2B36 and 2B37, showed dealkylation activity with the model substrates 7-ethoxy-4-(trifluoromethyl)coumarin and 7-benzyloxyresorufin, whereas 2B35 was inactive. Binding of the monoterpene (+)-α-pinene produced a Type I shift in the absorbance spectrum of each enzyme. Mutation of 2B37 at residues 114, 262, or 480, key residues governing ligand interactions with other CYP2B enzymes, did not significantly change expression levels or produce the expected functional changes. In summary, two catalytic and one ligand-binding assay are sufficient to distinguish among CYP2B35, 2B36, and 2B37. Differences in functional profiles between 2B36 and 2B37 are partially explained by changes in substrate recognition site residue 114, but not 480. The results advance our understanding of the mechanisms of detoxification in wild mammalian herbivores and highlight the complexity of this system.

Introduction

Cytochrome P450 (CYP)-dependent monooxygenases are a crucial superfamily of enzymes involved in the metabolism of numerous endogenous and xenobiotic compounds, including many pharmaceuticals (Johnson and Stout, 2013). CYPs are found in all multi-cellular organisms. These enzymes carry out Phase I metabolism of hydrophobic substrates through a variety of mechanisms, thereby rendering the compounds more hydrophilic to facilitate excretion or subsequent conjugation (Al Omari and Murry, 2007; Guengerich and Isin, 2008). In mammals, particularly herbivorous ones, CYPs likely arose as a protective response to the rapid evolution of chemical diversity in plants (Gonzalez, 1988; Stamp, 2003; Weng et al., 2012). Along with the central role in drug clearance, activation of prodrugs via CYPs also contributes to the importance of these enzymes in the pharmaceutical and healthcare industries (Guengerich, 2001; Ingelman-Sundberg, 2004).

CYPs accept a broad array of compounds as substrates, utilizing a highly conserved protein fold consisting of a largely α-helical domain and a domain dominated by β-sheets (Poulos, 2005a; Anzenbacher et al., 2008). The broad substrate specificity is due in part to the high degree of flexibility previously observed in these enzymes (Poulos, 2005b; Gay et al., 2010; Otyepka et al., 2012; Wilderman and Halpert, 2012). In addition to exhibiting considerable conformational plasticity, members of the CYP2B subfamily also show relatively low conservation of catalytic activity across mammalian species, making a good system for exploration of structure-function relationships within and across mammalian species (Zhao and Halpert, 2007). Studies of CYP2B enzymes have yielded a substantial body of biochemical and biophysical information relating to protein-ligand and protein-protein interactions and the catalytic mechanisms of microsomal monooxygenases. Studies utilizing isothermal titration calorimetry (ITC) and hydrogen/deuterium exchange coupled to mass spectrometry (DXMS) reinforced our understanding of the ability of the rabbit 2B4 to bind to a wide range of ligands in solution (Muralidhara et al., 2006; Wilderman and Halpert, 2012).

Despite the wealth of information concerning detoxification enzymes in model systems, interactions between plants and herbivores are more complex, and knowledge of the contributions of CYP enzymes to the biotransformation of plant secondary compounds in wild mammalian species is generally lacking (Dearing et al., 2005; Al Omari and Murry, 2007). Plant secondary compounds are metabolites that are not essential for the life of a plant, but may play a role in defense and are usually byproducts of primary metabolism. Molecular characterization of enzymes from wild populations is necessary for elucidating biotransformation mechanisms in wild herbivores. Unfortunately, genetic information on CYPs is available for only a small number of such species. Recent studies document involvement of CYPs in detoxification processes in koalas (Phascolarctos cinereus) and in several species of woodrats (Neotoma lepida) (Haley et al., 2007b; Haley et al., 2007a; Skopec et al., 2007; El-Merhibi et al., 2008; Haley et al., 2008; Magnanou et al., 2009). Furthermore, members of the 3A subfamily of CYPs were recently characterized from koala and the Eastern gray kangaroo (Macropus giganteus) (El-Merhibi et al., 2011; El-Merhibi et al., 2012) and the 4A subfamily in the koala and common bush tail possum (Trichosurus vulpecula) (Ngo et al., 2003; Ngo et al., 2006).

To advance knowledge of CYP enzymes in herbivores, we characterized members of the CYP2B subfamily of enzymes from desert woodrats (Neotoma lepida). This species was selected because it has a wide distribution across the southwestern USA, and individuals ingest a wide variety of secondary compounds such as terpenes and phenols depending on the availability of plant species in the habitat (Mangione et al., 2000). Sequence analysis of a large number of cDNA sequences from individual animals produced four major CYP2B clades (A, F, G, and H) that have different numbers of unique clones depending on home habitat and laboratory diet (Malenke et al., 2012). To initiate our investigation of the molecular mechanisms of plant-animal interactions mediated by detoxification enzymes we modified cDNA clones from clades A, F, and H to produce engineered proteins with a truncated and modified N-terminal region and a C-terminal tetra-His tag. These proteins were expressed, and the enzymes were characterized utilizing 7-ethoxy-4-trifluoromethyl coumarin (7-EFC) and 7-benzyloxyresorufin (7-BR) as substrates and the monoterpene (+)-α-pinene as a ligand probe. Results from these assays were compared with those from rat 2B1. The effects of mutations at residues corresponding to known polymorphisms in the human 2B6 (262) or known to alter substrate specificity and/or product profile in other 2B enzymes (114 and 480) were also explored. Two catalytic assays and one binding assay yielded unique results for each of the woodrat 2B enzymes, and subset of residues previously identified as governing substrate specificity, product production, or ligand binding affinity showed smaller changes in catalytic efficiency than expected. These findings highlight the overall complexity of the detoxification mechanisms utilized by wild mammalian herbivore.

Methods

Materials

All reagents were of the highest grade commercially available and were used without further purification. Macroprep CM-Sepharose cation exchange resin was from Bio-Rad Laboratories (Hercules, CA). TOPP3 cells were from Stratagene (La Jolla, CA). Primers were synthesized by Integrated DNA Technologies (Coralville, IA).

Clone Selection

The clones chosen for functional characterization were part of a large dataset of CYP2B cDNAs compiled from individuals of two desert woodrat populations from either the Mojave Desert or Great Basin (Malenke et al. 2012). Prior to genetic sampling, individuals were fed a diet including plant secondary compounds from either juniper (Juniperus osteosperma) or creosote bush (Larrea tridentata). These two plants are representative of the wild diets of the past and present diets of woodrats in the Mojave Desert, respectively. For additional information regarding the diets and feeding trial, see Malenke et al. (2012).

Woodrat 2B cDNA identification was performed as previously described (Malenke et al., 2012). Briefly, primers were designed from a search of Genbank for deposited 2B genes. Full open reading frames were acquired using rapid amplification of cDNA ends (First Choice RLM-RACE kit, Life Technologies, Carlsbad, CA) following the manufacturers’ protocol. Sequences were compared to deposited sequences from rat (Rattus norvegicus, CYP2B1 and CYP2B3), mouse (Mus musculus, CYP2B9 and CYP2B10), and human (Homo sapiens, CYP2B6). Variants were categorized into clades using phylogenetic analysis, and the different clades displayed variation at known substrate recognition site (SRS) residues (Malenke et al., 2012; Wilderman and Halpert, 2012). Two clones were chosen from each of three clades: A – 2B35v1/2, F – 2B36v1/2, and H – 2B37v1/2 (GenBank: JN105889 – 2B35v1, JN105900 – 2B35v2, JN105951 – 2B36v1, JN105955 – 2B36v2, JN105945 – 2B37v1 and JN105939 – 2B37v2). It is unknown whether the variant proteins in each clade represent unique alleles or different loci in the CYP2B subfamily. Full length clones were in either the pCR®2.1-TOPO® plasmid (2B35v2, 2B36v1/2, and 2B37v1) or the pCR™4-TOPO® plasmid (2B35v1 and 2B37v2) from Life Technologies (Grand Island, NY).

Sequence Alignment, Protein Engineering, and Site-Directed Mutagenesis

Sequence alignments and identity calculations were performed with Geneious v. 6.1.3 from Biomatters Ltd. (Auckland, New Zealand), using standard parameters. 2B1 was the reference sequence in all cases. Engineering efforts were performed as previously described (Scott et al., 2001). Sequences for all primers used in protein engineering efforts are contained in Table S1.

For 2B35v1, both variants of 2B36, and 2B37v1, the engineered gene was created using two PCR reactions. For 2B35v1 and 2B36v2, the first reaction contained the T7 primer as a forward primer and a reverse primer containing the native C-terminal end of the enzyme, four His codons, the stop codon, and an EcoRV recognition site (NSpdH R2 – 2B35v1; NSpdH R – 2B36v2). For 2B36v1 and 2B37v1, the first reaction contained the M13 reverse primer as a forward primer and the NSpdH R primer as the reverse primer. Prior to engineering 2B37v1, an internal NcoI recognition sequence was removed by introduction of a silent mutation (NcoI site removal primers in Table S1). Site directed mutagenesis was performed using primers for the desired sequence modifications and an appropriate PCR protocol. For these constructs, the second reaction contained a forward primer containing the 5’-truncation and modifications along with an NcoI recognition site (NSpdH F) the reverse primers used in the first reaction. A single PCR reaction was used to engineer 2B35v2 and 2B37v2 using NSpdH F as the forward primer and NSpdH R2 (2B35v2) or NSpdH R (2B37v2) as the reverse primer. The product of these reactions was purified using Macherey-Nagel NucleoSpin® Gel and PCR Clean-Up from Clontech (Mountain View, CA). Following clean up the PCR product for both variants of 2B35 and 2B37 were digested with NcoI and EcoRV and ligated into a similarly cut pKK2B4dH plasmid. Successful insertion of the engineered 2B36 variants into this vector required use of the In-Fusion® HD Cloning Kit from Clontech using the standard protocol from the kit. Construct sequences were verified at Retrogen, Inc. (San Diego, CA).

For site-directed mutagenesis, single mutants were created using as a template the plasmid that containing the protein of interest, along with appropriate forward and reverse primers (Table S2). Mutants were generated by PCR using Phusion High-Fidelity DNA Polymerase and a standard site-directed mutagenesis protocol. Constructs were sequenced at Retrogen, Inc.

Protein Expression and Purification

CYP2B enzymes were expressed in TOPP3 cells as previously described (Scott et al., 2001) and purified by the protocol used by Shah and coworkers (Shah et al., 2012). The pellet was resuspended in 10% of the original culture volume in buffer containing 20 mM potassium phosphate (pH 7.4 at 4 °C), 20% (v/v) glycerol, 10 mM 2-mercaptoethanol (BME), and 0.5 mM phenylmethanesulfonyl fluoride (PMSF). The resuspended cells were further treated with lysozyme (0.3 mg/mL) and stirred for 30 min, followed by centrifugation for 30 min at 7000 rpm in a JA-14 rotor in a Beckman Coulter Avanti J-26 XPI Centrifuge. After decanting the supernatant, spheroplasts were resuspended in 5% of the original culture volume in buffer containing 500 mM potassium phosphate (pH 7.4 at 4 °C), 20% (v/v) glycerol, 10 mM BME, and 0.5 mM PMSF and were sonicated for 3 × 45 s on ice. 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) was added to the sample at a final concentration of 0.8%, and this was allowed to stir for 30 min at 4 °C prior to ultracentrifugation for 45 min at 41000 rpm using a fixed-angle Ti 50.2 rotor in a Beckman Coulter Optima L-80 XP Ultracentrifuge. The CYP concentration was measured using the reduced CO difference spectra from the resulting supernatant (Omura and Sato, 1964; Omura et al., 1965).

The supernatant was applied to a Ni²⁺-NTA column. The column was washed with buffer containing 100 mM potassium phosphate (pH 7.4 at 4 °C), 100 mM NaCl, 20% (v/v) glycerol, 10 mM BME, 0.5 mM PMSF, 0.5% CHAPS and 5 mM histidine, and the protein was eluted using buffer containing 10 mM potassium phosphate (pH 7.4 at 4 °C), 100 mM NaCl, 20% (v/v) glycerol, 10 mM BME, 0.5 mM PMSF, 0.5% CHAPS, and 50 mM histidine. Protein fractions containing protein of the highest quality as measured by the A₄₁₇/A₂₈₀ ratios were pooled, and the CYP concentration in the eluted fractions was measured using the reduced CO difference spectra. The CYP containing fractions were pooled and loaded onto a Macroprep CM-Sepharose column. The cation exchange column was washed using low salt buffer containing 5 mM potassium phosphate (pH 7.4 at 4 °C), 20% (v/v) glycerol, 1 mM EDTA, and 0.2 mM DTT. The protein was eluted with high salt buffer containing 50 mM potassium phosphate (pH 7.4 at 4 °C), 500 mM NaCl, 20% (v/v) glycerol, 1 mM EDTA and 0.2 mM DTT. Protein fractions containing protein of the highest quality as measured by the A₄₁₇/A₂₈₀ ratios were pooled, and the CYP fractions were pooled and the concentration was measured using the reduced CO-difference spectra.

Enzymatic Assays

The standard NADPH-dependent assay for 7-EFC or 7-BR O-dealkylation by 2B enzymes was carried out as described previously (He et al., 1995; Kumar et al., 2005). The reconstituted system contained the following recombinant proteins at a molar ratio of 1:4:2 cytochrome P450, rat NADPH-cytochrome P450 reductase (Harlow and Halpert, 1997), and rat cytochrome b₅ (Holmans et al., 1994) as previously optimized (Scott et al., 2001).

Reactions for using 7-EFC were carried out using the reconstituted system in a 100 μl final volume and substrate at varying concentrations (0–150 μM). Assays were performed in buffer containing 50 mM HEPES (pH 7.4) and 15 mM MgCl₂, initiated by addition of NADPH to a final concentration of 1 mM and incubated for 10 min at 37 °C. Ice cold acetonitrile was added to quench the reaction. An aliquot of the reaction was then transferred to a glass tube containing 0.1 M Tris (pH 9.0), and fluorescence was determined with excitation at 410 nm and emission at 510 nm for 7-HFC using a Cary Eclipse Fluorimeter (Agilent, Santa Clara, CA).

Reactions for the 7-BR assay were carried out using the reconstituted system in a 500 μl final volume and substrate at varying concentrations (0–10 μM). Assays were performed in buffer containing 50 mM HEPES (pH 7.4) and 15 mM MgCl₂, initiated by addition of NADPH to a final concentration of 1 mM and incubated for 10 min at 37 °C. Methanol (2 mL) was added to quench the reaction. Fluorescence was determined with excitation at 550 nm and emission at 580 nm for resorufin using a Hitachi F2000 fluorescence spectrophotometer (Tokyo, Japan).

In both assays calibration was performed at the end of each day by measuring the intensity of fluorescence in a series of four to five samples of the same reaction mixture containing HFC or resorufin at concentrations from 0 to 2 μM. Steady state kinetic parameters were determined by regression analysis using SPECTRALAB (Davydov et al., 1995). Values of catalytic constants were determined using the Michaelis-Menten equation. Kinetic experiments included wild-type and mutant enzymes for more accurate comparison of the data.

Spectral Binding Titrations

The absorbance spectra were measured with an MC2000-2 multi-channel CCD rapid scanning spectrometer (Ocean Optics, Dunedin, FL, USA) equipped with one absorbance and one fluorescence channel, a pulsed Xe-lamp PX-2 light source, and a custom thermostated cell chamber with a magnetic stirrer. A semi-micro quartz cell with a stirring compartment (10 × 4 mm light path) from Hellma GmbH (Müllheim, Germany) was used in the titration experiments. All titration experiments were carried out at 25 °C with continuous stirring in buffer containing 100 mM HEPES (pH 7.4 at 4 °C), 20 % glycerol, and 100 mM KCl. A baseline was recorded between 340 and 700 nm using this buffer. A spectrum was recorded after the addition of protein to the buffer. Spectra were recorded after the addition of aliquots of ligand to the sample cuvette. The spectral dissociation constants (K_S) were obtained by fitting the data to the equation for ligand binding for high affinity ligands 2ΔA = (ΔA_max/[E₀])((K_S + [L₀] + [E₀]) – (K_S + [L₀] + [E₀])² – 4[E₀][L₀])^1/2) where ΔA_max is the maximum change in the percent high-spin P450 (HS_P450), E₀ is total enzyme concentration, and L₀ is total ligand concentration when data are normalized to total CYP concentration using SPECTRALAB (Davydov et al., 1995). The mean differences in the change in % HS_P450 values between mutant and wild-type enzymes were compared using an independent sample t-test (Microsoft Excel 2011, Redmond, WA, USA).

Figures

All protein model figures were generated using PyMOL (DeLano, 2009). All chemical structures were created using MarvinSketch v. 5.2.6, 2009, ChemAxon (http://www.chemaxon.com). Figures depicting spectral data were created using Igor Pro 6.2 (Wavemetrics Inc., Lake Oswego, OR, USA).

Results

Sequence Alignment and Protein Engineering

In the interest of expanding the understanding of structure-function relationships of CYP2B enzymes, six previously isolated cDNAs from desert woodrats (Neotoma lepida) were engineered in the same manner as 2B1 (Scott et al., 2001; Malenke et al., 2012). Extensive prior results have documented that the N-terminal truncation and modification and the addition of a tetra-His tag to CYP2B enzymes increase protein expression levels, facilitate solubilization using high salt conditions, and result in high purity after chromatography with minimal impact on function. Figure 1A shows the N-terminal and C-terminal regions of the engineered 2B1dH and full-length 2B35v1, 2B35v2, 2B36v1, 2B36v2, 2B37v1, and 2B37v2. The cDNA sequences of the woodrat enzymes were among those reported by Dearing and coworkers (Malenke et al., 2012).

Figure 1. Comparison of engineered rat 2B1 and full-length woodrat 2B35v1/2, 2B36v1/2, and 2B37v1/2.

A) The N-terminal and C-terminal regions of engineered 2B1dH and the full-length woodrat enzymes. Amino acid numbering corresponds to the full-length position. Amino acid differences in the woodrat enzyme variants are shown underlined and bold for B) 2B35, C) 2B36, and D) 2B37. The N-terminal truncation and modification and C-terminal tetra-His tag found in 2B1dH were engineered into the woodrat 2B enzymes, so the first difference for each variant pair is not present in the engineered protein.

An alignment of the engineered 2B1dH and the full-length enzymes from woodrat is in Supplementary Figure S1. Using the previous notation 2B35v1/2 are from Clade A, 2B36v1/2 are from Clade F, and 2B37v1/2 are from Clade H. Variants of each enzyme display ≥98 % identity to each other across 491 amino acids. 2B35, 2B36, and 2B37 show roughly 90-95% amino acid identity between enzymes when compared pairwise. The Clade A and Clade F variant enzymes contain two amino acid differences apiece (2B35v1/2 – G2E and F408L; 2B36v1/2 – L8P and Q45R); in both pairs of enzymes only one of the differences is found in the engineered enzyme (Figure 1B-C). The variants from Clade H contain six differences (E2G, K262R, L431I, A452T, M463V, and V465P); of these five are present in the engineered enzymes (Figure 1D). Amino acid differences present in the engineered enzymes are highlighted on homology models in Figure 2A-C. Models were generated using the homology modeling program MODELLER (Eswar et al., 2008), the complete amino acid sequence of the engineered protein, and the CYP2B6-(+)-α-pinene complex (PDB: 4I91, (Wilderman et al., 2013)) with ligands and waters removed as the template. The heme is colored red in all panels. The amino acid sequences of the woodrat enzymes also show at least 87 % identity to full-length rat 2B1.

Figure 2. Homology models showing the location of amino acid differences and mutants in woodrat 2B35, 2B36, and 2B37.

A) 2B35v1 model in blue highlighting residue 408 in red. B) 2B36v1 model in magenta highlighting residue 45 in red. C) 2B37v2 model in cyan highlighting residues 262, 431, 452, 463, and 465 in red. D) 2B37v2 model showing the location of residues mutated for further analysis in this manuscript (114, 262, and 480).

Significant differences in expression levels were seen among engineered 2B35, 2B36, and 2B37 (Table 1). Each variant of 2B35 and 2B36 also showed different expression levels from its partner. 2B35v1 expressed in E. coli at ~35 % of the level of 2B1. 2B36v1 had expression levels about 20 % higher than 2B1. For 2B35 and 2B36 variant one expressed at roughly 3.5-times the level of variant two. Both 2B37 variants expressed at similar levels to each other, ~55-60 % of the level of the engineered rat enzyme. Following engineering, 2B35v1, 2B36v1, and both variants of 2B37 were purified to homogeneity using a Ni²⁺-NTA column followed by a Macroprep CM-Sepharose column. Characterization of 2B35v2 was not continued due to protein instability after elution from the Ni²⁺-NTA column, and separation of 2B36v2 from RNA contamination during the purification process was not possible.

Table 1.

Quantitative expression of, model substrate metabolism by, and (+)-α-pinene binding to rat 2B1 and wild-type woodrat 2B enzymes

Protein (Clade)	Quantitative Expressiona (nmol/L culture)	7-EFC O-deethylationa			7-BR O-debenzylation			(+)-α-pinene spectral binding
		kcat (min−1)	KM(μM)	kcat/KM (min−1μM−1)	kcat (min−1)	KM(μM)	kcat/KM (min−1μM−1)	KS (μM)a	%HSP450b
									Max Δ Initial
2B1	800 ± 50	6.9 ± 1.2	32.4 ± 4.9	0.21	4.6 ± 0.8	2.0 ± 0.6	2.3	0.87 ± 0.08	48.7 ± 1.3
									32.6 ± 1.5
2B35v1 (A)	280 ± 30	NDc	--	--	NDc	--	--	1.27 ± 0.49	13.1 ± 1.6
									7.0 ± 0.3
2B35v2 (A)d	55 ± 20	--	--	--	--	--	--	--	--
2B36v1 (F)	945 ± 80	3.3 ± 1.0	18.3 ± 7.0	0.18	0.10 ± 0.03	1.7 ± 0.3	0.06	0.58 ± 0.04	42.8 ± 7.2
									12.3 ± 1.0
2B36v2 (F)e	270 ± 10	--	--	--	--	--	--	--	--
2B37v1 (H)	590 ± 30	10.0 ± 0.3	23.5 ± 2.0	0.43	0.16 ± 0.04	7.2 ± 0.6	0.02	9.33 ± 0.76	25.3 ± 0.5
									28.6 ± 1.7
2B37v2 (H)	450 ± 30	14.5 ± 8.3	20.0 ± 4.2	0.73	0.25 ± 0.06	9.5 ± 0.5	0.03	0.91 ± 0.07	38.4 ± 0.7
									25.8 ± 1.4

^aValues are represented as the mean ± standard deviation of at least three independent experiments.

^bHS_P450, high-spin P450

^cND, not detected.

^d2B35v2 was not stable during the Ni²⁺-NTA purification step.

^eSeparation of 2B36v2 from RNA during purification was not possible.

7-EFC and 7-BR Metabolism

Two model substrates for which CYP2B enzymes from other species have high selectivity are 7-ethoxy-4-trifluoromethylcoumarin (7-EFC) and 7-benzyloxyresorufin (7-BR). Steady-state kinetic measurements show unique profiles for each of the woodrat 2B enzymes (Table 1). With 2B35v1 O-dealkylation products were not detected for either 7-EFC or 7-BR. With 2B36v1 catalytic efficiency was similar to that of rat 2B1 for 7-EFC O-deethylation with k_cat about half that of 2B1 and K_M about two thirds that of 2B1. However the efficiency for 7-BR O-debenzylation by 2B36v1 was 1/40^th that of 2B1. This result is almost entirely due to a much lower k_cat.

Using 7-EFC as a substrate, 2B37v1 displayed a k_cat that was ~1.4-fold higher than that for 2B1, and K_M was ~25 % lower; the resulting catalytic efficiency for this substrate was ~1.8-fold higher than for 2B1. With 2B37v2, catalytic efficiency was ~2.5-fold greater than 2B1dH, which was almost entirely due to the higher k_cat compared with 2B1. Catalytic efficiency for metabolism of 7-BR by both variants of 2B37 was similar to that of 2B36v1. The decreased of the 2B37 variants compared with 2B1 is due to a combination of a markedly lower k_cat and a K_M roughly 3.5 – 4.5 times higher than the rat enzyme.

Spectral Binding Titrations

Spectral binding titrations were used to assess the interactions of a typical monoterpene that has been shown previously to be a potent inhibitor of CYP2B1 in rat liver microsomes. This class of compounds is found at high concentrations in juniper (Adams et al., 1981; De-Oliveira et al., 1997a; De-Oliveira et al., 1999). (+)-α-pinene (Figure 3A) induced a Type I difference spectra with a peak at ~388 nm and a trough at ~420 nm, as represented in Figure 3B with 2B1. Plotting the change in the % HS_P450 for each enzyme produced a curve similar to that displayed in Figure 3C. Spectral binding affinities (K_S) for (+)-α-pinene were very similar among 2B1 and most of the woodrat enzymes (Table 1). The K_S for (+)-α-pinene was at least an order of magnitude weaker for 2B37v1 (~9 μM) than for the other enzymes. In addition to the high affinity of this ligand for CYP2B enzymes, the magnitude of the spin shift seen upon ligand binding is remarkable (Table 1). 2B1 showed the greatest maximum percent change in (HS_P450) content upon (+)-α-pinene binding, followed by 2B36v1, 2B37v2, 2B37v1, and 2B35v1.

Figure 3. CYP2B interaction with (+)-α-pinene.

A) Chemical structure of (+)-α-pinene, a common monoterpene from plants, particularly juniper species. B) Absolute absorbance spectra from a titration of 2B1 with (+)-α-pinene. A distinct decrease in absorbance at 417 nm (low-spin state, LS_P450) and increase at 390 nm (high-spin state, HS_P450) is seen in these results. C) Plot of fraction HS_P450 as a function of ligand concentration. Fitting to the ligand binding equation is shown.

Residue 262 in 2B37

In human 2B6, residue 262 is the location of a single nucleotide polymorphism (SNP). The replacement of lysine with arginine at this residue increases expression levels and protein solubility, and alters susceptibility to inhibition by various drugs (Talakad et al., 2009). Accordingly, the reciprocal mutants at position 262 (Figure 2D) were made in the engineered 2B37v1 and 2B37v2. Effects on expression were small (Table 2). Introduction of arginine at residue 262 reduced the catalytic efficiency of 2B37v1 by ~35 %, mainly due to a ~45 % decrease in k_cat. Replacement of R262 with lysine in 2B37v2 had a much greater effect on oxidation of 7-EFC, yielding an 85 % decrease in the catalytic efficiency. Much smaller decreases were seen in 7-BR O-dealkylation by these mutants, and virtually no change was seen in the catalytic efficiencies of this reaction. The K_S for (+)-α-pinene was four times lower for 2B37v1 K262R than for the wild-type enzyme, but the 2B37v2 R262K substitution had little effect on (+)-α-pinene K_S. The relative increase in high-spin hemoprotein content of wild-type 2B37v1 is ~90 % with an initial content of 28.6 % and an increase of 25.3 %; this increase is ~145 % for 2B37v2 (initial – 25.8 %, increase – 38.4 %). 2B37v1 K262R displays a relative increase of ~190 % (initial – 30.6 %, increase – 58.1 %), and for 2B37v2 R262K the relative increase remains ~145% (initial – 33.8 %, increase – 49.2 %). Analysis of these changes using an independent sample t-test yielded P-values <<0.001.

Table 2.

The effect of altering residue 262, a known site of genetic polymorphism in human 2B6, on Clade H enzymes

Protein	Quantitative Expressiona (nmol/L culture)	7-EFC O-deethylationa			7-BR O-debenzylation			(+)-α-pinene spectral binding
		kcat (min−1)	KM(μM)	kcat/KM (min−1μM−1)	kcat (min−1)	KM(μM)	kcat/KM (min−1μM−1)	KS (μM)a	%HSP450b
									Max Δ Initial
2B37v1	590 ± 30	10.0 ± 0.3	23.5 ± 2.0	0.43	0.16 ± 0.04	7.2 ± 0.6	0.02	9.33 ± 0.76	25.3 ± 0.5
									28.6 ± 1.7
2B37v1 K262R	460 ± 40	5.4 ± 0.2	19.0 ± 2.3	0.28	0.15 ± 0.07	8.1 ± 0.7	0.02	2.04 ± 0.23	58.1 ± 1.2
									30.6 ± 0.9
2B37v2	450 ± 30	14.5 ± 8.3	20.0 ± 4.2	0.73	0.25 ± 0.06	9.5 ± 0.5	0.03	0.91 ± 0.07	38.4 ± 0.7
									25.8 ± 1.4
2B37v2 R262K	510 ± 60	2.0 ± 0.1	20.3 ± 2.8	0.10	0.18 ± 0.05	9.2 ± 0.7	0.02	1.06 ± 0.07	49.2 ± 0.4
									33.8 ± 1.3

^aValues are represented as the mean ± standard deviation of at least three independent experiments.

^bHS_P450, high-spin P450

Effects of altering residues 114 and 480

Residues 114 and 480 (Figure 2D) play roles in substrate recognition and/or product profiles in various 2B enzymes (Zhao and Halpert, 2007). Dearing and coworkers noted these residues and others in previously identified substrate recognition sites (SRS) show similarity within and variation across the phylogenetic clades identified from the sequence data (Malenke, et al., 2012). Thus, V114I, V480L, and V114I/V480L were created in 2B37v2 with the aim of engineering 2B36 metabolism profiles into 2B37. Each of these mutants showed less than 25 % change in expression levels when compared with the wild-type enzyme (Table 3). These mutants displayed altered catalytic and binding activities from the wild-type 2B37v2, but none of these profiles were comparable to that of 2B36v1. With 7-EFC as a substrate the V114I mutant showed much lower catalytic efficiency than 2B37v2 due to a lower k_cat and much higher K_M, but V480L was roughly as efficient as the wild-type enzyme with a higher k_cat and lower K_M for 7-EFC. The double mutant V114I/V480L behaved in a manner more similar to the V480L mutant and wild-type 2B37v2 than the V114I mutant. Small changes in catalytic efficiency were seen with 7-BR with both single mutants. V114I had a lower k_cat and lower K_M for 7-BR than 2B37v2, and V480L had higher k_cat and lower K_M for this ligand. The double mutant, V114I/V480L, showed a low k_cat, but the K_M was greater than the solubility limit of the ligand in aqueous solutions (~10 μM) preventing determination of catalytic constants. The K_S of the mutants for (+)-α-pinene was lower than wild-type 2B37v2 by roughly a factor of three or more. These affinities were greater than that of 2B36v1. The effects on the K_S for (+)-α-pinene binding to each of the single mutants seems to be somewhat additive in the case of the double mutant. These mutations in 2B37v2 did not produce a pseudo-2B36v1, and each of the enzymes and mutants had a unique profile across the 7-EFC, 7-BR, and (+)-α-pinene assays.

Table 3.

Changes at residues 114 and 480 alter enzyme activity in 2B37v2, but do not create a pseudo-clade F enzyme

Protein (Clade)	Quantitative Expressiona (nmol/L culture)	7-EFC O-deethylationa			7-BR O-debenzylation			(+)-α-pinene spectral binding
		kcat (min−1)	KM(μM)	kcat/KM (min−1μM−1)	kcat (min−1)	KM(μM)	kcat/KM (min−1μM−1)	KS (μM)a	%HSP450b
									Max Δ Initial
2B36v1 (F)	945 ± 80	3.3 ± 1.0	18.3 ± 7.0	0.18	0.10 ± 0.03	1.7 ± 0.3	0.06	0.58 ± 0.04	42.8 ± 7.2
									12.3 ± 1.0
2B37v2 (H)	450 ± 30	14.5 ± 8.3	20.0 ± 4.2	0.73	0.25 ± 0.06	9.5 ± 0.5	0.03	0.91 ± 0.07	38.4 ± 0.7
									25.8 ± 1.4
2B37v2 V114I	400 ± 30	6.3 ± 1.9	120 ± 2	0.05	0.13 ± 0.01	5.7 ± 1.0	0.02	0.36 ± 0.05	30.4 ± 3.5
									47.5 ± 1.8
2B37v2 V480L	550 ± 70	22.3 ± 2.0	23.9 ± 3.6	0.93	0.47 ± 0.08	4.5 ± 0.7	0.10	0.18 ± 0.10	31.6 ± 1.5
									36.7 ± 1.2
2B37v2 V114I/V480L	470 ± 50	20.8 ± 2.8	32.5 ± 4.4	0.64	NDc	> 10d	NDc	0.25 ± 0.15	26.0 ± 3.5
									49.0 ± 2.1

^aValues are represented as the mean ± standard deviation of at least three independent experiments.

^bHS_P450, high-spin P450

^cND, not detected.

^d7-BR is not soluble in an aqueous solution at concentrations over ~10 μM.

Discussion

Despite the intense focus in recent years on the ligand recognition and binding step(s) in the CYP catalytic cycle, there is still uncertainty as to how a single enzyme handles a diverse assortment of xenobiotic compounds (Wade et al., 2005; Davydov and Halpert, 2008; Isin and Guengerich, 2008; Halpert, 2011). Data from a host of complementary techniques including NMR, molecular dynamics simulations, and X-ray crystallography indicate an integral role of protein dynamics in ligand binding and catalysis for many enzymes, including CYPs (Hays et al., 2004; Henzler-Wildman et al., 2007; Bakan and Bahar, 2009; Pochapsky et al., 2009; Savino et al., 2009). Furthermore, multiple recent studies of CYP2B enzymes focused on the role of both active site residues and non-active site residues in function and protein stability (Kumar et al., 2007; Talakad et al., 2010; Wilderman et al., 2012). The techniques used in such studies should be able to shed light on the selective process whereby wild herbivores adapt to changes in the plant secondary compound (PSC) profiles of their diets. However, systematic analysis of enzymes from wild mammalian herbivores has only recently begun (Malenke et al., 2012; Malenke et al., 2013). The results presented here describe the characterization of CYP2B enzymes from individuals drawn from wild non-primate populations. Protein-ligand interactions were compared between rat 2B1 and the woodrat CYP2B enzymes. The results underscore the importance of small differences in amino acid sequences in the function of 2B enzymes. Furthermore, previously identified residues that determine protein-ligand interactions in rat, rabbit, human, and dog CYP2B enzymes, including residues 114, 262, and 480, fail to account for functional differences among the various woodrat enzymes.

Characterization of the woodrat CYP2B enzymes was initiated by examination of steady-state kinetics of the oxidation of the model CYP2B substrates 7-EFC and 7-BR. Both variants of 2B37 were studied, whereas 2B35v2 and 2B36v2 could not be purified with our standard chromatographic procedure. Of the purified enzymes, 2B35v1 did not show dealkylation activity in either assay. 2B36v1 turned over 7-EFC about as efficiently as 2B1, but had a much lower k_cat for 7-BR O-debenzylation than 2B1 (~0.10 vs. ~4.5). In comparison to rat 2B1, the 2B37 variants had similar K_M values for 7-EFC. The k_cat for 7-EFC-O-deethylation by 2B37v1 and 2B37v2 was ~50 % and 2-fold greater than that of 2B1, respectively. Similar to 2B36v1, the 2B37 variants were much less efficient at catalyzing O-debenzylation of 7-BR than 2B1. The lower catalytic efficiency in the 7-BR assay with 2B36v1 and the 2B37 variants does not rule out more efficient dealkylation of other resorufin derivatives by these enzymes. Similarly, 2B35v1 might show activity with other coumarin or resorufin derivatives. Furthermore, the possibility of oxidation events other than 7-dealkylation occurring is not excluded, as was seen in mutants of 2B1 and human 3A4 probed using 7-alkoxycoumarins of varying chain length (Kobayashi et al., 1998; Khan and Halpert, 2000).

In addition to catalytic studies, CYP2B interactions with a natural product encountered in a wild diet were examined. The ligand probe used here, (+)-α-pinene, is a monoterpene that occurs in the diets of woodrats (Mabry et al. 1977) and is also a component of the diets and the environments of other herbivorous and omnivorous species (rabbit, rat, human). CYP2B enzymes had previously been linked to interactions with a range of monoterpenoids including (+)-α-pinene (Figure 2A) (De-Oliveira et al., 1997a; De-Oliveira et al., 1999; Miyazawa et al., 2003; Miyazawa and Gyoubu, 2007; Wilderman et al., 2013). Interestingly, (+)-α-pinene is a pure hydrocarbon and produces a marked increase in HS_P450 content when each of the enzymes is saturated with the ligand (Table 3, Figure 2B-C). For reference, the drugs clopidogrel, sertraline, and ticlopidine bind with low micromolar K_S values and induce maximal increases in HS_P450 content upon binding to human 2B6 of only 2.6 %, 0.8 %, and 2.3%, respectively (Talakad et al., 2009). Even the relatively lower spin shift upon (+)-α-pinene binding to 2B35v1, is more than 5-times higher than that of clopidogrel with 2B6. (+)-α-pinene binds with very high affinity to 2B1 and three of the woodrat enzymes tested, 2B35v1, 2B36v1, and 2B37v2 but shows significantly lower binding affinity for 2B37v1 (Table 3). Since only five residues in the engineered 2B37v1 and 2B37v2 are different, one or more of these amino acid differences should explain these large differences in spectral binding affinity between the two enzymes.

While (+)-α-pinene has been linked to CYP2B enzymes, terpenoids in general have been linked to multiple CYP enzymes. In addition to CYP2B enzymes from rat, rabbit, and human, (+)-α-pinene is also metabolized by CYP101A1 (P450_cam) (Bell et al., 2002). A mixture of terpenoids, including 1,8-cineole and (+)-α-pinene, induced expression of CYP enzymes that reacted with antibodies to human CYP2E1 and rat CYP2C6 and CYP2C11 in experiments with common brushtail possoms (Pass et al., 1999). However, 1,8-cineole was primarily turned over by CYP3A4 in experiments using human liver microsomes and recombinant CYP enzymes (Duisken et al., 2005). The monoterpene β-myrcene induces expression of CYP2B enzymes in rat, and multiple monoterpenoids inhibit CYP2B1 function in rat liver microsomes (De-Oliveiraet al., 1997a; De-Oliveira et al., 1997b). Furthermore, CYP2B enzymes oxidize (-)-α-pinene in rabbit and human (Noma and Asakawa, 2009). Limonene is metabolized by CYP2C11 in untreated rats and by CYP2B1 in phenobarbital treated rats but by CYP2C9 and CYP2C19 in humans (Miyazawa et al., 2001; Miyazawa et al., 2002). The ketones (-)-fenchone and (-)-verbenone are metabolized by CYP2A6 and 2B6 in humans, but (-)-verbenone is metabolized by CYP2C11 and CYP2B1 in rats (Miyazawa et al., 2003; Miyazawa and Gyoubu, 2007). Based on this evidence, herbivore response to monoterpenoid challenge is likely species specific, and while the possibility exists that another CYP enzyme might be the primary route of processing (+)-α-pinene in Neotoma species, the chemical properties of the molecule match the general description of a CYP2B substrate in terms of lipophilicity (log P = 4.44), molecular shape, and relative molecular mass (M_r = 136.23) (Sridhar et al., 2012).

The topic of residues important for governing substrate specificity, product production, and individual variability in humans has been well studied in CYP2B enzymes (Domanski and Halpert, 2001; Ingelman-Sundberg, 2004; Zhao and Halpert, 2007; Mo et al., 2009; Zanger and Schwab, 2013). In 2B37, mutation of residue 262 had small effects on K_M for the substrate 7-EFC, but large changes were seen in k_cat and catalytic efficiency. Two expected results of reduced k_cat are a lower rate of clearance of potential toxins and a diminished ability to deal with increases in the dietary levels of PSCs. Interestingly, mutation of this residue in 2B6 causes little difference in k_cat for 7-MFC oxidation, but there are significant differences in protein-ligand interactions, specifically competitive inhibition, for multiple important drugs including ticlopidine (Talakad et al., 2009). Mutation of residues 114 and/or 480 in identified substrate recognition sequences did not change the function of 2B37v2 to that of 2B36v1. In rat 2B1 residue 114 is a major determinant of substrate specificity, and the importance of residue 480 was shown in dog 2B11 (Hasler et al., 1994; Liu et al., 1996). Since residues previously implicated in substrate recognition did not markedly alter the preferences of 2B37v2, other amino acid differences yet to be identified and investigated between 2B36v1 and 2B37v2 must govern the differences in ligand recognition.

The functional differences among 2B35, 2B36, and 2B37 are of interest in the context of geographically distinct woodrat populations. Animals from the Mojave Desert regularly consume large amounts of creosote bush (Larrea tridentata), which contains resin dominated by phenolic compounds, while the diet of animals from the Great Basin population has a high percentage of juniper foliage, a plant rich in terpenes. Juniper is not currently found in the Mojave but is thought to be a part of the ancient flora displaced as a result of the natural climate change that occurred ~18,000 years ago (Hunter et al., 2001). Malenke and coworkers noticed that 2B35 was present in both woodrat populations, but the other enzymes were predominantly isolated from either the Great Basin (2B36) or the Mojave Desert (2B37) (Malenke et al., 2012). Based on this information we propose that 2B35 may be a general detoxification mechanism for woodrats, and 2B36 and 2B37 have evolved to deal with the unique PSC profiles from their diets. Future work characterizing the function of these and other woodrat CYP2B enzymes may reveal additional novel functions related to specific endogenous compounds, PSCs, or to the wide breadth of compounds encountered by these generalist herbivores in their daily diet.

The genus Neotoma has been used as an experimental system to study plant-herbivore interactions because its members have disparate plant diets that differ across species and habitats. Most Neotoma species consume plants with high levels of secondary compounds, which present a considerable toxic challenge to these herbivores. In response, woodrats have evolved a variety of mechanisms for dealing with these toxins including caching behaviors, efflux transporters in the gut, bacterial endosymbioses, and liver enzyme biotransformation (Sorensen and Dearing, 2003; Haley et al., 2008; Torregrossa and Dearing, 2009; Kohl and Dearing, 2012). The importance of the liver in biotransformation has been demonstrated multiple times within Neotoma, but the question remains as to how liver enzymes have adapted to the challenges of ingesting diets containing complex mixtures of potential toxins (Haley et al., 2007b; Haley et al., 2008). The comparative data presented here on woodrat CYP2B enzymes begin to answer this question by showing further evidence that small changes in protein sequence have large effects on metabolism of substrates and that ecologically relevant biochemical assays (such as the monoterpene binding assay) may reveal hitherto unknown function.

Highlights.

• Three CYP2B enzymes from Neotoma lepida were cloned, engineered, and expressed.

• A mix of catalytic and binding assays yields unique results for each enzyme.

• Substrate recognition in these enzymes remains to be clarified as demonstrated by functional analysis of mutants at key residues in other enzymes.

• Reported CYP2B gene sequences from Neotoma lepida allow for larger scale analyses of the structure-function relationships for these detoxification enzymes.

Abbreviations

CYP

Cytochrome P450-dependent monooxygenases

ITC

isothermal titration calorimetry

DXMS

hydrogen-deuterium exchange coupled to mass spectrometry

7-EFC

7-ethoxy-4-(trifluoromethyl)coumarin

7-BR

7-benzyloxyresorufin

2BXdH

an N-terminally truncated and modified 2B enzyme with a C-terminal tetra-His tag to facilitate purification

BME

2-mercaptoethanol

PMSF

phenylmethanesulfonyl fluoride

CHAPS

3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate

EDTA

ethylenediaminetetraacetic acid

DTT

dithiotheitol

PSC

plant secondary compound or plant specialized compound

HS_P450

high spin P450