Investigation by site-directed mutagenesis of the role of cytochrome P450 2B4 non-active site residues in protein-ligand interactions based on crystal structures of the ligand-bound enzyme

SUMMARY

Residues located outside of the active site of cytochromes P450 2B have exhibited importance in ligand binding, structural stability, and drug metabolism. However, contributions of non-active site residues to the plasticity of these enzymes are not known. Thus, a systematic investigation was undertaken of unique residue-residue interactions found in crystal structures of P450 2B4 in complex with 4-(4-chlorophenyl)imidazole (4-CPI), a closed conformation, or in complex with bifonazole, an expanded conformation. Nineteen mutants distributed over eleven sites were constructed, expressed in E. coli, and purified. Most mutants showed significantly decreased expression, especially in the case of interactions found in the 4-CPI structure. Six mutants (H172A, H172F, H172Q, L437A, E474D, and E474Q) were chosen for detailed functional analysis. Among these, the K_s of H172F for bifonazole was ~20-times higher than wild type 2B4, and the K_s of L437A for 4-CPI was ~50-times higher than wild type, leading to significantly altered inhibitor selectivity. Enzyme function was tested with the substrates 7-ethoxy-4-(trifluoromethyl)coumarin (7-EFC), 7-methoxy-4-(trifluoromethyl)coumarin (7-MFC), and 7-benzyloxyresorufin (7-BR). H172F was inactive with all three substrates, and L437A did not turn over 7-BR. Furthermore, H172A, H172Q, E474D and E474Q showed large changes in k_cat/K_M for each of the three substrates, in some cases up to 50-fold. Concurrent molecular dynamics simulations yield distances between some of the residues in these putative interaction pairs that are not consistent with contact. The results indicate that small changes in the protein scaffold lead to large differences in solution behavior and enzyme function.

INTRODUCTION

Cytochrome P450 (P450)-dependent monooxygenases (EC 1.14.14.1) are members of a ubiquitous superfamily of heme containing proteins and are responsible for oxidation of a broad range of substrates in the biogenesis of sterols and hormones and metabolism of xenobiotic compounds [1, 2]. Some mammalian P450s can accept a wide variety of hydrophobic substrates of differing shapes and sizes and render them more hydrophilic for excretion or subsequent conjugation. Additionally, P450s mediate the conversion of prodrugs to the respective bioactive compounds [3-5].

While P450s accept a broad range of substrates, the protein fold of P450s, consisting of a large, primarily alpha helical single domain is well conserved across families [6, 7], and a high degree of flexibility in these enzymes has previously been observed [7-9]. P450s form compact structures around small ligands or ligand-free active sites, as seen in P450 2B4 complexed with 4-(4-chlorophenyl)imidazole (4-CPI) [10], the P450 2A subfamily [11, 12], and P450 2C5 and P450 2C9 [13, 14]. Some P450s also appear to be able to accommodate ligands of greater volume, as demonstrated with P450 2B4 with bifonazole [15] or P450 3A4 with erythromycin, ketoconazole or ritonavir [16].

Much of this conformational plasticity is exhibited in the P450 2B subfamily of enzymes. These proteins show low catalytic conservation across species, providing a good model system to examine structure-function relationships in P450s [8]. Studies of the P450 2B subfamily have yielded considerable biochemical and biophysical insight into substrate binding, protein-protein interactions, and the catalytic mechanisms of microsomal monooxygenases. Crystal structures of an engineered P450 2B4 show a remarkable amount of plasticity is possible while retaining enzyme function [8, 17]. Experiments utilizing hydrogen/deuterium exchange coupled to mass spectrometry (DXMS) reinforced the view concerning plasticity of P450 2B4 [18].

One member of this subfamily, P450 2B6, plays an important role in human drug metabolism [19, 20] and is highly polymorphic [5, 19, 21, 22]. Furthermore, some of these single nucleotide polymorphisms (SNPs) (Fig. 1) have been linked to differential metabolism by P450 2B6 of a variety of drugs [23-25]. Most of the amino acid changes are far from the active site of the enzyme [23]. Previous studies of subfamily 2B enzymes demonstrated that mutations located outside of the active site of P450 2B enzymes significantly alter enzyme function [17, 26-29].

Figure 1.

Ribbon diagram showing the location of known SNPs (yellow sphere and sticks) in 2B6. The heme is shown as red sticks. All the known coding sequence variants of 2B6 are located outside of the active site.

Given the demonstrated importance of non-active site residues in P450 enzymes, an investigation of the functional and structural role of non-active site residues possibly involved in plasticity through residue-residue interactions was undertaken. Residues of interest were selected from interactions unique to either the structure of 2B4 complexed with 4-CPI or with bifonazole [30]. Site-directed mutagenesis was used to interrupt four interactions from the 4-CPI structure and five interactions from the bifonazole complex. One of the interactions in the bifonazole complex involved a site known to be polymorphic in P450 2B6, H172, so another site known to have large effects on ligand binding and catalysis was also investigated, R262. Mutants were characterized by expression level, thermal stability, thermal inactivation, H₂O₂-mediated heme depletion, binding affinity for 4-CPI and bifonazole, steady-state kinetics with a variety of substrates, enzyme inhibition in the presence of 4-CPI or bifonazole, and molecular dynamics simulation.

RESULTS

Identification of interactions of interest

Residue-residue interactions involving two or more amino acid side chains within 4 Å of each other were analyzed in structures of P450 2B4 in the open ligand-free conformation (PDB ID: 1PO5), in complex with 4-CPI (PDB ID: 1SUO), and in complex with bifonazole (PDB ID: 2BDM). Interactions were grouped as conserved (found in all three structures) and alternate (unique to the 4-CPI-bound or bifonazole-bound structure) [30]. Eight alternate interactions are unique to the 4-CPI complex, and seven are unique to the bifonazole complex (Table 1). To simplify interpretation of the effects of mutants, interactions consisting of two amino acid residues were chosen as initial candidates for interruption (Fig. 2). Interestingly, one of the observed interactions (H172-E301) occurs at a SNP (Q172H) in the closely related P450 2B6. Another SNP in P450 2B6 (K262R) involves a residue hypothesized to participate in a hydrogen bonding network in the complex of this enzyme with the small molecule 4-CPI [31]. P450 2B6 Q172H shows marked differences in metabolism of cancer and HIV chemotherapeutics, and P450 2B6 K262R shows altered in vitro binding of clopidogrel, itraconazole, raloxifine, and sertraline [19, 25]. The corresponding mutations (H172Q and R262K) were therefore introduced into P450 2B4.

Table 1.

Unique residue-residue interactions in crystal structures of 2B4.

4-CPI complex1		Bifonazole complex2
Interaction3	Mutants	Interaction3	Mutants
F115-F108-L290	None	*R48-N479	N479Q, N479R
R120-E282-N287	None	*K49-E474	E474D, E474Q
*F127-F283	F127A, F283A	*T131-F296	T131A, T131S
*M132-L437	M132A, L437A	*D166-S483	S483A, S483T
F184-F296-F202-L241	None	*H172-E301	H127A, H172F
*R197-E240	E240A, E240D	Y268-I289-Q286	None
Q204-S205-Q233-N237	None	E403-T404-N409-H412	None
*D275-S281	S281A, S281T

^*Interactions selected for mutational analysis

¹Crystal structure of 2B4 in complex with 4-CPI (1SUO)

²Crystal structure of 2B4 in complex with bifonazole (2BDM)

³Residues within 4 Å of each other in the crystal structure.

Figure 2.

Location of interactions interrupted by mutation. A) Interactions in the 2B4-4-CPI structure (1SUO). B) Interactions in the 2B4-bifonazole structure (2BDM). Residues in each interaction are shown as sticks with a red sphere for the alpha carbon. Positions for focused analysis from SNP sites (H172A/F/Q), the 2B4-4-CPI structure (L437A), and the 2B4-bifonazole structure (E474D/Q) are marked with an asterisk (*).

Expression and purification of 2B4 mutants

Mutants were first expressed and purified as previously described [17]. Each mutation resulted in lower expression levels than wild type enzyme (Table 2). Seven mutants (T131S, E240Q, S281T, F283A, L437A, E474Q, and S483A) showed greater than 50% of wild type protein expression levels, and seven more mutants (T131A, H172A, H172F, H172Q, E474D, N479Q, and S483T) expressed at 20-50% of wild type. E240D, S281A, and N479R showed expression levels less than 20% of wild type. F127A and M132A did not show measurable expression levels as measured by CO difference spectra. For those mutants that expressed, P420 content was measured as described in the Materials and Methods section, and most had levels between 5% less than and 10% more than wild type P420 values, which were ~9%. However, R262K expressed as 100% P420.

Table 2.

Expression levels of 2B4 mutants

4-CPI complex		Bifonazole complex
Mutant	Expression Level (nmoles P450/L)	Mutant	Expression Level (nmoles P450/L)
Wild type	530 ± 20	Wild type	522 ± 39
F127A	N.D.	T131A	164 ± 15
M132A	N.D.	T131S	277 ± 112
E240D	95 ± 15	#H172A	156 ± 51
E240Q	391 ± 65	#H172F	220 ± 39
S281A	15 ± 5	#H172Q	154 ± 20
S281T	351 ± 12	E474D	184 ± 35
F283A	420 ± 22	E474Q	324 ± 70
L437A	264 ± 23	N479Q	165 ± 70
		N479R	107 ± 20
		S483A	275 ± 15
#R262K	N.D.	S483T	236 ± 30

^#Location of genetic polymorphism in 2B6

Results are the mean ± standard deviation (n=3).

N.D. – not detectable

Enzyme stability

Mutations in P450s resulting in lower expression levels generally indicate less stable enzymes. Therefore, the thermal stability of the 14 mutants that expressed at greater than 20% of wild type levels was assessed with the exception of S483A, which was 100% P420 after purification and not studied further. The T_M for each of the other mutants except S483T was within ± 3 °C of P450 2B4 wild type (Fig. 3A). In addition, thermal inactivation showed greater deviation from wild type values, but no clear trends in effects of mutation at a specific residue arose (Fig. 3B).

Figure 3.

Stability of 2B4 wild type and mutant proteins. A) Thermal melting temperature. B) Thermal inactivation rate constant. C) Hydrogen peroxide mediated heme depletion rate constant. Black bars represent P450 inactivation; gray bars represent P420 inactivation.

In the interests of focusing efforts, further characterization of mutants was limited to: H172A/F/Q, representative of a SNP in CYP 2B6; L437A, representing the 4-CPI structure mutants; and E474D/Q, representing the bifonazole structure mutants. Hydrogen peroxide mediated heme destruction was monitored as a measure of water access to the active site pocket. Each mutant showed at least a 2-fold decrease in the P450 k_inact, and most mutants at least a small decrease in P420 k_inact (Fig. 3C). Interestingly, substituting phenylalanine at residue 172 increased the P420 k_inact by 50%.

Ligand interaction

To investigate effects of these mutations on ligand binding, catalysis, and inhibition, we used 7-EFC, 7-MFC, and 7-BR as substrates and 4-CPI and bifonazole as inhibitors. Since mutants were created to interrupt interactions found either in the 2B4-4-CPI complex or the 2B4-bifonazole complex, spectral binding titrations were performed using these two small molecules. The spectral dissociation constant (K_s) values for both imidazole inhibitors generally increase with the introduction of a mutation (Fig. 4A-B). Comparing the ratio of K_s for 4-CPI to K_s for bifonazole for each mutant shows a large change for H172F and L437A (Fig. 4C). Furthermore, a difference in relative specificity for these two imidazole inhibitors was created relative to wild type 2B4, with H172F showing increased selectivity for 4-CPI and L437A increased selectivity for bifonazole (Fig. 4C).

Figure 4.

Spectral binding of imidazole ligands to 2B4 wild type and mutant proteins. A) K_s for 4-CPI. B) K_s for bifonazole. C) Ratio of constants for 4-CPI to bifonazole.

While most of the mutations caused less than 3-fold changes in K_s for the two imidazole molecules, there were greater effects on enzyme function. The substrates 7-EFC, 7-MFC and 7-BR were turned over by many of the mutants (Table 3). However, H172F was inactive with all three compounds, and L437A showed no activity with 7-BR. For 7-EFC, changes in k_cat/K_M for H172A and L437A appear to be due to a 2- to 3-fold lower k_cat. H172Q and E474Q showed a ~5- to ~7-fold lower k_cat, and E474D showed a >20-fold decrease in k_cat. These three mutants also showed a 2- to 3-fold increase in K_M. As with 7-EFC as a substrate, changes in k_cat/K_M for these mutants using 7-MFC as a substrate seem to be mainly due to changes in k_cat. With the exception of the inactive H172F, each mutant showed <2-fold change in K_M for 7-MFC. H172A and H172Q showed ~2-fold lower k_cat and L437A, E474D and E474Q a 4- to 5-fold decrease. Turnover of 7-BR is affected more by these mutations, as the k_cat/S₅₀ values change 6- to 40-fold, mainly due to changes in k_cat. The mutants that turned over 7-BR exhibited a 2- to 3-fold increase in S₅₀. E474D and E474Q displayed ~2.5- and ~4.5-fold changes in k_cat, while H172A and H172Q showed larger effects on k_cat, ~19- and ~15-fold changes, respectively.

Table 3.

Steady-state kinetics of 2B4 mutants

Protein	7-EFC			7-MFC			7-BR
	K M 1	k cat 2	kcat/KM	K M 1	k cat 2	kcat/KM	S 50 1	k cat 2	kcat/S50
WT	25 ± 3	9.6 ± 0.4	0.36	132 ± 6	11.3 ± 0.2	0.08	2.3 ± 0.5	7.6 ± 0.3	3.3
H172A	26 ± 4	4.2 ± 0.3	0.15	126 ± 11	5.2 ± 0.6	0.04	5.0 ± 0.8	0.4 ± 0.1	0.08
H172F	N.D.	0	0	N.D.	0	0	N.D.	0	0
H172Q	48 ± 6	1.7 ± 0.3	0.04	115 ± 15	7.4 ± 0.5	0.06	8.2 ± 1.0	0.5 ± 0.1	0.06
L437A	25 ± 4	2.7 ± 0.2	0.12	85 ± 8	1.8 ± 0.2	0.02	N.D.	0	0
E474D	77 ± 10	0.4 ± 0.1	0.01	185 ± 14	3.8 ± 0.2	0.02	6.2 ± 0.7	3.1 ± 0.2	0.5
E474Q	64 ± 8	1.2 ± 0.2	0.02	178 ± 17	2.4 ± 0.3	0.01	4.8 ± 0.5	1.7 ± 0.3	0.35

Standard errors for fit to the respective equations are shown. Results are representative of at least two independent determinations.

¹K_m and S₅₀ are measured in μM.

²k_cat is measured in min⁻¹.

N.D., not detectable

Inhibition of 7-EFC O-deethylation was also affected by mutation of non-active site residues (Table 4). Alteration of the non-active site residues generally increases the IC₅₀ for 4-CPI and causes small changes in the IC₅₀ of bifonazole. The ratio of the two values also increases, since the 4-CPI IC₅₀ generally increases more than the bifonazole IC₅₀ changes. L437A is unique in that this ratio changes drastically and with a significant decrease in the bifonazole IC₅₀. This single non-active site mutant thus shows a 100-fold decrease in selectivity of 4-CPI:bifonazole.

Table 4.

Inhibition of non-active site mutants by imidazole inhibitors.

Protein	4-CPI IC50 (μM)	Bifonazole IC50 (μM)	Ratio 4-CPI IC50/Bifonazole IC50
WT	0.04	0.90	0.04
H172A	0.31	0.58	0.53
H172Q	0.22	0.76	0.29
L437A	0.97	0.05	3.88
E474D	0.24	0.62	0.39
E474Q	0.50	0.81	0.62

All experiments were completed using 50 μM 7-EFC as substrate.

Results are the mean of at least two independent determinations.

Molecular dynamics simulation

With these findings, a molecular dynamics simulation was employed to explore whether the interactions taken from crystal structures actually occur in a dynamic setting. Using the 4-CPI-bound 2B4 structure (PDB ID: 1PO5) or the bifonazole bound 2B4 structure (PDB ID: 2BDM) as a starting structure for a 2 ns simulation using GROMACS, the backbone of 2B4 appears to breathe in a similar manner to previous simulations (Fig. 5A) [18]. In the interest of examining interaction partners, the closest heteroatoms in each pair from the respective crystal structures were used to measure the distance between residues during the course of the simulation (Fig. 5B and Supplementary Fig. S1). With a 0.6 nm cutoff, four interactions are close enough to persist for the entirety of the simulation (K49-E474, D166-S483 from the 2BDM and F127-F283, M132-L437 from 1SUO), and the remaining interactions are transient to varying degrees (R48-N479, T131-F296, H172-E301 in 2BDM and R197-E240, D275-S281 in 1SUO). Most interaction pairs are within a fairly close range of distances from each other; however R48-N479, R197-E240, and D275-S281 have wide swings in distance between the interaction partners.

Figure 5.

Molecular dynamics simulation of 2B4 wild type. A) Representative change in backbone RMSD over 2 ns simulation for wild type protein using the 4-CPI structure (1PO5) as the starting structure; this change is representative of changes found in simulations of mutant proteins or in the simulation using the bifonazole structure (2BDM) as the starting structure. B) Distance between closest atoms for each interaction found in the crystal structures of 2B4 complexed with bifonazole or 4-CPI for mutants interrupting interactions from the respective structure. The five interactions from the bifonazole structure are listed at the top, and the four interactions from the 4-CPI structure are on the bottom. C) Distance between residues found in the proposed hydrogen bonding network in cytochromes 2B from Gay et al. [31].

The interaction of residues in the proposed hydrogen bonding network around R262 were also examined. The inactive R262K mutant does not exhibit a change in the distance between H252 and D263 (Fig. 5C). However, the T255-R262 and R262-D266 interactions show increased distance between the interaction partner and residue 262 when the latter is Lys (not shown). This likely leads to a disruption of any stabilizing benefit of the hydrogen bonding network provided by R262.

X-ray crystal structure

Because of the marked changes in sensitivity to inhibition by 4-CPI vs. bifonazole, we solved an X-ray crystal structure of the 2B4 L437A mutant in complex with 4-CPI (PDB ID: 3TK3) (Table 5). Comparison with the X-ray crystal structure of wild type 2B4 in complex with 4-CPI (PDB ID: 1SUO) yields an RMSD of 0.26 Å, showing the structures are virtually superimposable (Fig. 6A). The orientation of active site residues is maintained in the mutant; however, the density for E301 indicates split occupancy, where one orientation points into the active site to hydrogen-bond with the azole nitrogen of 4-CPI, while the other orientation flips out to interact with H172 (Fig. 6B). The C_β carbons of residue 437 for the wild type and L437A proteins in complex with 4-CPI are in the same orientation.

Table 5.

X-ray Data Collection & Refinement Statistics

Crystal Data
Space Group	P3
Unit Cell
a, b, c (Å)	232.9, 232.9, 57.0
α, β, γ (°)	90, 90, 120
Molecules per AU	4
Data Collection
X-ray source	SSRL BL 11-1
Wavelength (Å)	0.98
Resolution Range (Å)	76.2-2.80
Total Observations	354,184
Unique Observations (F>0)	81,306
Completeness (%)a	99.4 (99.4)
Redundancya	4.4 (4.2)
I/σa	5.1 (1.6)
Rmerge (%)a,b	20.1 (85.7)
Refinement Statistics
R-factor (%)c	19.30
Rfree (%)c	24.08
RMS bond lengths (Å)	0.016
RMS bond angles (°)	1.356
Number of Atoms
Proteind	14,727 (52.6)
Hemed	172 (41.4)
4-CPI	48 (50.4)
Waterd	290 (41.9)

^aValues for the highest resolution shell are in parentheses.

^bR_merge = [Σ_hΣ_i|I_h-I_hi|/Σ_hΣ_iI_hi] where I_h is the mean of I_hi observations of reflection h.

^cR-factor & R_free = Σ∥F_obs|-|F_calc∥/Σ|F_obs| × 100 for 95% of the recorded data (R-factor) and 5% of data (R_free).

^dAverage B-factors (Ų) are in parentheses.

Figure 6.

Overlay of wild type and mutant 2B4 in complex with 4-CPI. A) Overlay of C_α backbone of wild type 2B4 (yellow, PDB ID: 1SUO) and the L437A mutant (blue, PDB ID: 3TK3). B) The active site and heme support of wild type 2B4 and the L437A mutant.

DISCUSSION

Adapting mammalian P450s for biomedical and biotechnological applications via improved activity and/or stability has been of considerable recent interest [32-34]. Previous studies focused on active site and substrate access channel residues [8, 35-37]. However, the highly plastic nature of P450s suggests that non-active site residues stabilize different conformations through residue-residue contacts. Recent studies of the P450 2B enzymes have demonstrated the importance of non-active site residues in enzyme inhibition, catalysis, stability, and expression [17, 26-29, 38]. Furthermore, most of the known SNPs in human 2B6 are distal from the active site [23]. In light of this information, this study focused on the role of non-active site residues in protein-ligand interactions involving P450 2B4. Introduction of mutants outside of the enzyme active site led in general to small changes in binding affinity but much larger affects on substrate catalysis and enzyme inhibition.

Comparison of interactions of two residues from each of three crystal structures of 2B4 identified four unique interactions in the 4-CPI-bound structure (1SUO) and five in the bifonazole-bound structure (2BDM). Not surprisingly, interrupting these interactions by site directed mutagenesis changed expression levels of the protein to varying degrees. The mutants that did express did not have significantly altered thermal stability or protection from H₂O₂ heme degradation. In general, interrupting interactions found in the bifonazole-bound structure produced more mutants that yielded significant levels of P450 than interruptions of interactions found in the 4-CPI-bound structure. These results suggest that some of the interactions seen exclusively in the 4-CPI-bound complex also exist in solution in the absence of ligand and may be important for enzyme stability.

Conversely, changes in protein-ligand interactions were more pronounced. Small perturbations in binding affinities of the 4-CPI and bifonazole were seen in most of the six mutants tested. Interestingly, two mutants (H172F and L437A) had large changes in binding affinity for one or both inhibitors. H172F showed increased affinity for 4-CPI, while L437A showed increased affinity for bifonazole. Both mutants had greater than 9-fold changes in the ratio of 4-CPI K_s to bifonazole K_s. While the remaining mutants showed less than 3-fold changes in this ratio, larger changes were seen in steady-state kinetic assays using 7-EFC, 7-MFC, and 7-BR. The two mutants with greatest change in imidazole binding were inactive with one (L437A) or all (H172F) substrates tested. Moreover, the changes in imidazole binding affinity of E474D and E474Q (~5-fold) are accompanied by changes in enzyme activity (>30-fold). While small changes in binding affinity for either 4-CPI or bifonazole were seen for most of the mutants (<2-fold), much greater changes were seen in IC₅₀ values for each of the compounds.

While little structural rearrangement is seen in the crystal structure, replacement of Leu 437 with Ala may create a cavity allowing for water access to or change in redox potential of the heme leading to a decreased k_cat; this cavity could also allow for greater porphyrin mobility and resultant affects on kinetics of imidazole binding. Interestingly, profound effects on ligand interaction were recently observed upon mutation of the P450 BM3 residue just C-terminal from the heme cysteine [39], which corresponds to residue 437 in 2B4. Specifically, the BM3 I401P mutant is highly active towards non-natural substrates, and the crystal structure of the ligand-free mutant resembles substrate-bound forms of the wild type enzyme.

Interaction behavior of the residue pairs was accessed using MD simulations. A small number of interaction pairs remained within 6 Å of each other for the entire simulation: F127-F283, D166-S483, and K49-E474. Except for K49-E474, these interactions involve residues in the conserved regions of the protein, which does not show large conformational changes between crystal structures (residues 58-100, 141-176, 189-202, 299-473, and 481-491) [8], the I-helix or very close to the heme cysteine. Portions of the protein showing large conformational changes among crystal structures were previously termed plastic regions [8]. Interestingly, the K49-E474 interaction from the bifonazole structure involves residues in plastic region 1 (residues 39-57) and plastic region 5 (residues 474-480). Another subset of residue pairs (T131-F296, M132-L437, and H172-E301) was transiently within 6 Å of each other in a narrow range of distances over the course of the simulation. These residues are in the conserved region, plastic region 2 (residues 101-140) and plastic region 4 (residues 203-298). The remaining interactions, R48-N479, R197-E240, and D275-S281, have average distances greater than 6 Å and have a wide variance, and they involve residues in the conserved region, plastic region 1, plastic region 4, and plastic region 5. Interestingly, the interactions in the 4-CPI bound structure involving residues in plastic region 4 are highly transient (R197-E240 and D275-S281). R197-E240 involves residues in the F-G cassette of 2B4, which provides variable volume to the active site through coordinated movement around the E-F and G-H loops [40]. D275-S281 is between residues in the H-I loop or transiently in the I-helix. These regions and plastic region 2, consisting of the B`-C loop and the C-helix, show the greatest amount of movement between various crystal structures of the enzyme [8, 40] and have the greatest difference in solvent protection in DXMS experiments. Furthermore, the S281 and F283 are in the H-I loop or the N-terminal end of the I-helix; previous work with 2B1 showed that mutating residues in these regions has detrimental consequences for expression, stability, and kinetic fidelity [41].

In conclusion, mutations to certain non-active site residues changed the relative selectivity of 4-CPI and bifonazole and, in some cases, caused profound functional changes. Interestingly, changes to active site residues produced much smaller changes, generally ~2-fold or less, in previous work [42]. A likely mechanism is disruption of residue-residue interactions that stabilize an active conformation of the enzyme. A similar explanation may account for the altered function of several genetic variants of human P450 2B6 [31, 43]. Moreover, future studies involving mutations based on crystal structures of cytochromes P450 should utilize molecular dynamics simulations to assess the likelihood in solution of residue-residue interactions inferred from the structure.

MATERIALS AND METHODS

Materials

7-hydroxy-4-(trifluoromethyl)coumarin (7-HFC), 7-methoxy-4-(trifluoromethyl)coumarin (7-MFC), and 7-ethoxy-4-(trifluoromethyl)coumarin (7-EFC) were purchased from Invitrogen (Carlsbad, CA). Sodium hydrosulfite, β-mercaptoethanol (βME), phenylmethylsulphonyl fluoride (PMFS), and NADPH were obtained from Sigma (St. Louis, MO). Recombinant NADPH-cytochrome P450 reductase (CPR) and cytochrome b₅ (b₅) from rat liver were prepared as described previously [44]. Oligonucleotide primers for PCR were obtained from Sigma (St. Louis, MO). Phusion High-Fidelity DNA Polymerase was purchased from New England Biolabs (Ipswich, MA). Nickel-nitrilotriacetic acid (Ni²⁺-NTA) affinity resin was purchased from Qiagen (Valencia, CA). All other chemicals were of the highest grade available and were used without further purification.

Cytochrome P450 and P420 quantification

During protein purification, the P450 concentration of the supernatant was measured using the reduced CO difference spectrum and a molar extinction coefficient of 91 mM⁻¹ cm⁻¹ [45, 46]. For crystallization trials, the P450 concentration was measured using the oxidized P450 spectrum and a molar extinction coefficient of 106 mM⁻¹ cm⁻¹. For P420 content, thermal stability studies, and H₂O₂-supported P450 heme depletion, total concentration of the P450 was gauged by nonlinear least-squares approximation of the spectra [47] using a linear combination of spectral standards of P450 2B4 low-spin P450, high-spin P450, and P420 states [48, 49] using the previously described SPECTRALAB software package [50] or Igor Pro version 6.1 (Wavemetrics, Inc., Lake Oswego, OR).

Site directed mutagenesis, protein expression, and purification

Single mutants in 2B4 were created using the 2B4dH plasmid as the template and appropriate forward and reverse primers (Supplementary Table S1). Constructs were sequenced at Retrogen, Inc. (San Diego, CA). Mutants were generated by polymerase chain reaction (PCR) using Phusion High-Fidelity DNA Polymerase and a standard site directed mutagenesis protocol.

For biophysical assays, 2B4 was expressed in TOPP3 cells as described previously [51] and purified by the same protocol. 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 βME, and 0.5 mM PMSF. The resuspended cells were further treated with lysozyme (0.3 mg/ml) and stirred for 30 min, followed by a brief centrifugation for 30 min at 7000 rpm in a JA-14 rotor using a Beckman Coulter Avanti J-26 XPI centrifuge. After decanting of 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 βME, and 0.5 mM PMSF and were sonicated for three times for 45 s on ice. The membrane pellet was separated by centrifugation for 10 min at 7000 rpm, and sodium cholate was added to the supernatant at a final concentration of 0.5% (w/v). This was allowed to stir for 30 min at 4 °C prior to ultracentrifugation for 45 min at 41,000 rpm using a fixed-angle Ti 50.2 rotor in a Beckman Coulter Optima L-80 XP ultracentrifuge.

The supernatant was applied to a nickel-nitrilotriacetic acid column. The column was washed with lysis buffer containing 100 mM potassium phosphate (pH 7.4 at 4 °C), 100 mM NaCl, 20% (v/v) glycerol, 10 mM βME, 0.5 mM PMSF, 0.5% (w/v) sodium cholate, 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 βME, 0.5 mM PMSF, 0.5% (w/v) sodium cholate, and 60 mM histidine. The P450-containing fractions were pooled and dialyzed against buffer containing 10 mM potassium phosphate (pH 7.4 at 4 °C), 10% (v/v) glycerol, and 1 mM ethylenediaminetetraacetic acid (EDTA) with two changes.

For crystallization trials, protein was expressed, harvested, and resuspended as described above. After lysosome treatment, centrifugation, and sonication, 3-[(3-cholaminopropyl)dimethylammonio]-1-propane sulfonate (CHAPS) was added to the solution at a concentration of 0.8% (w/v). This was allowed to stir for 90 min at 4 °C prior to ultracentrifugation for 45 min at 45,000 rpm using a fixed-angle Ti 50.2 rotor in a Beckman Coulter Optima L-80 XP ultracentrifuge.

The supernatant was applied to a nickel-nitrilotriacetic acid column. The column was washed with lysis buffer containing 100 mM potassium phosphate (pH 7.4 at 4 °C), 100 mM NaCl, 20% (v/v) glycerol, 10 mM βME, 0.5 mM PMSF, 0.5% (w/v) sodium cholate, 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 βME, 0.5 mM PMSF, 0.5% (w/v) sodium cholate, and 60 mM histidine. The P450-containing fractions were pooled and diluted 10-fold in buffer with 5 mM potassium phosphate (pH 7.4 at 4°C), 20% (v/v) glycerol, 1 mM EDTA, 0.2 mM DTT, 0.5 mM PMSF, and 0.5% (w/v) CHAPS before loading onto a Macroprep CM cation exchange column. The column was washed using 5 mM potassium phosphate (pH 7.4 at 4°C), 20 mM NaCl, 20% (v/v) glycerol, 1 mM EDTA, and 0.2 mM DTT and 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 with the highest A₄₁₇/A₂₈₀ ratios were pooled.

Spectral studies of ligand binding

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 home-made 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 50 mM potassium phosphate (pH 7.4 at 4 °C), 500 mM NaCl, 1 mM EDTA, and 0.2 mM DTT. A baseline was recorded between 350 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 a series of 5 μl aliquots of inhibitor (100 μM) to the sample cuvette. The spectral dissociation constants (K_s) were obtained by fitting the data to the equation for “tight binding” 2ΔA = (ΔA_max / [E₀]) ((K_D + [I₀] + [E₀]) + ((K_D + [I₀] + [E₀]))² - 4[E₀][I₀])_1/2) for high affinity ligands Igor Pro version 6.1.

Enzyme assay

The standard NADPH-dependent assay for 7-MFC, 7-EFC, or 7-BR O-dealkylation by 2B4 was carried out as described previously [26, 52]. The reconstituted system contained the following recombinant proteins at a molar ratio of 1:4:2: 2B4, rat NADPH-cytochrome P450 reductase [44], and rat cytochrome b₅ [53]. Reactions were carried out using the reconstituted system in a 100 μL final volume and substrate at varying concentrations (0–150 μM for 7-MFC/7-EFC; 0-10 μM for 7-BR). Assays were performed in buffer containing 50 mM HEPES (pH 7.4) and 15 mM MgCl₂, initiated by addition of NADPH (1 mM) and incubated for 10 min at 37 °C. Aqueous 20 % (v/v) trichloroacetic acid 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-EFC and 7-MFC and with excitation at 530 nm and emission at 585 nm for 7-BR using a Cary Eclipse Fluorimeter (Agilent, Santa Clara, CA). A blank was run for each sample, and the final activity was calculated by comparison to a standard curve for the respective substrate. Steady state kinetic parameters were determined by regression analysis using Igor Pro version 6.1. The k_cat and K_M values were determined using the Michaelis-Menten equation, and the S₅₀ and n values were determined using the Hill equation. Kinetic experiments included wild type and mutant enzymes for more accurate comparison of the data.

Thermal stability studies

Inactivation of P450 was monitored as described earlier [27]. The reaction mixture contained 1 μM protein in 100 mM NaOH-HEPES buffer (pH 7.4). Absorbance spectra were measured using a Shimadzu-2600 spectrophotometer (Shimadzu, Kyoto, Japan). Thermal inactivation was carried out by measuring a series of absorbance spectra in the 340- to 700-nm range as a function of temperature between 25 and 70 °C with 2.5–5 °C intervals and a 2 min equilibration at each temperature. For inactivation kinetics, the samples were treated at 45 °C, and the spectra (340–700 nm) were recorded at different time intervals. All data treatment and fitting of the titration curves were performed with Igor Pro version 6.1. Fitting of the temperature profile and time-dependent inactivation curves was performed by regression analysis using Igor Pro version 6.1. The inactivation profiles were fit to a two-state model to obtain the mid-point of the thermal transition temperature (T_m); a simple pseudo–first-order equation was used to determine the k_inact values [27].

H₂O₂-supported P450 heme depletion

Determination of the kinetics of P450 2B4 heme depletion in the presence of H₂O₂ was conducted under conditions similar to those described previously for P450 2B1dH [27]. The reaction was carried out at 25°C in 100 mM NaOH-HEPES buffer (pH 7.4) in a 1 mL semi-micro spectrophotometric cell with constant stirring using a MC2000-2 multi-channel CCD rapid scanning spectrometer from Ocean Optics. The reaction mixture contained 1 μM protein and 60 mM H₂O₂. A series of absorbance spectra in the 340- to 700-nm range over 5 minutes were recorded following addition of H₂O₂ to the hemoprotein buffer mixture. Determination of the total concentration of the heme protein and data fitting to determine the rate constants were performed as described under “Cytochrome P450 and P420 quantification” and “Thermal stability studies”, respectively.

Molecular dynamics simulations

Molecular dynamics (MD) simulations were performed with starting coordinates of the 4-CPI-bound 2B4 structure (PDB ID: 1PO5) or the bifonazole bound 2B4 structure (PDB ID: 2BDM) using the molecular dynamics software package GROningen MAchine for Chemical Simulation (GROMACS) Version 4.07 [54]. Residues not found in the 4-CPI complex or the bifonazole complex (residues 20-27 and 493-495) were added using the homology modeling program Modeller and the complete amino acid sequence of the protein [55]. The topology files used in the energy minimization and MD simulation were modified to reflect the cysteinyl ligation to heme [56, 57]. The 2B4 structure was immersed in a simulated water box with 120 Å sides and containing ~70,000-80,000 waters, corresponding to twice the length of the longest diagonal of the protein (~65 Å). The structure was energy minimized by the method of steepest descent to remove Van der Waals contact between overlapping waters and the amino acids of the protein. Simulations were run with Berendsen temperature and pressure coupling (a.k.a. “bath”) [58] at a simulated temperature of 300 K using the GROMOS 53a6 force field [59] and periodic boundary conditions in all directions. Electrostatics of the system were measured using the particle-mesh Ewald method [60]. Then, simulations were conducted using a Linux server at the University of California, San Diego. During the first 250 ps of the MD simulation, the protein was position restrained to allow the waters to fill in the cavities. After that, the MD simulation was continued without restraints for another 2000 ps.

Crystallization and data collection

Pooled L437A protein was diluted to 18 μM in 50 mM potassium phosphate (pH 7.4 at 4 °C), 500 mM NaCl, 500 mM sucrose, 1 mM EDTA, and 0.2 mM DTT. 4-CPI was added to this solution at a concentration of 180 μM and allowed to bind overnight at 4 °C. The protein-ligand complex was then concentrated to 550 μM and supplemented with an additional 1.0 mM of 4-CPI. Cymal-5 and 225-chol were added to this solution to final concentrations of 4.8 mM and 0.063% (w/v), respectively. The protein and detergents were allowed to equilibrate for approximately 20 min before mixing with crystallization reagents. Screening for crystallization conditions was performed by sitting drop vapor diffusion using the Wizard I high throughput kit from Emerald Biosystems (Bainbridge Island, WA) by mixing equal volumes of protein solution and well solution. Drops were equilibrated against the well solutions at 18 °C. Rod shaped crystals suitable for X-ray diffraction appeared over the course of approximately 2 weeks in drops containing 0.1 M MES (pH 6.0), 20% (w/v) PEG 8000, and 0.2 M calcium acetate. Crystals were briefly transferred to a solution of mother liquor supplemented with 335 mM sucrose before flash freezing in liquid nitrogen. 140° of data were collected using 1° oscillations and 20 s exposures at 100 K on a Quantum CCD detector (Area Detector Systems Corp., Poway, CA) at BL 11-1 of the Stanford Synchrotron Radiation Lightsource (SSRL) (Stanford, CA). The data were processed to 2.80 Å using iMOSFLM [61] and SCALA [62].

Structure determination and refinement

Phases were obtained by molecular replacement using the previously determined wild type 2B4−4CPI complex (PDB ID: 1SUO) (with the inhibitor molecular removed from the coordinates) in Phaser [63]. The structure solution was found in space group P3 containing 70.2% solvent, assuming four molecules per asymmetric unit. The initial model was first subjected to a simulated annealing step followed by restrained refinement in PHENIX [64] using non-crystallographic symmetry restraints to remove model bias. Model building was performed in Coot [65] using both 2Fo-Fc and Fo-Fc electron density maps contoured to 1-σ and 3-σ, respectively. The model was modified to fit the electron density and refined in an iterative manner until a final R-factor of 19.3% and an R_free of 24.1% were reached. Non-crystallographic symmetry restraints were slowly released during the refinement process. Structure refinement statistics are summarized in Table 5.

Protein figures

All protein figures were generated using PyMol [66].

Abbreviations and textual footnotes

P450

cytochrome P450

4-CPI

4-(4-chlorophenyl)imidazole

BIF

bifonazole

2B4dH

an N-terminal truncated and modified and C-terminal 4-His-Tagged form of cytochrome P450 2B4

pdb

protein data bank

7-BR

7-benzyloxyresorufin

SNP

single nucleotide polymorphism

DXMS

hydrogen/deuterium exchange coupled to mass spectrometry

7-HFC

7-hydroxy-4-(trifluoromethyl)coumarin

7-MFC

7-methoxy-4-(trifluoromethyl)coumarin

7-EFC

7-ethoxy-4-(trifluoromethyl)coumarin

Ni²⁺-NTA

nickel-nitrilotriacetic acid

ITC

isothermal titration calorimetry

βME

β-mercaptoethanol

PMSF

phenylmethylsulphonyl fluoride

NADPH

nicotinamide adenine dinucleotide phosphate

CPR

recombinant NADPH-cytochrome P450 reductase

b₅

cytochrome b₅

PCR

polymerase chain reaction

EDTA

ethylenediaminetetraacetic acid

CO

carbon monoxide

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

CHAPS

3-[(3-cholaminopropyl)dimethylammonio]-1-propane sulfonate