Tryptophan-96 in cytochrome P450 BM3 plays a key role in enzyme survival

Abstract

Flavocytochrome P450 from Bacillus megaterium (P450_BM3) is a natural fusion protein containing reductase and heme domains. In the presence of NADPH and dioxygen the enzyme catalyzes the hydroxylation of long-chain fatty acids. Analysis of the P450_BM3 structure reveals chains of closely spaced tryptophan and tyrosine residues that might serve as pathways for high-potential oxidizing equivalents to escape from the heme active site when substrate oxidation is not possible. Our investigations of the total number of enzyme turnovers before deactivation have revealed that replacement of selected tryptophan and tyrosine residues with redox inactive groups leads to a twofold reduction in enzyme survival time. Tryptophan-96 is critical for prolonging enzyme activity, suggesting a key protective role for this residue.

Introduction

The cytochromes P450 (CYP) are a superfamily of heme enzymes found in archaea, bacteria, and eukaryotes that catalyze the aerobic oxidation of organic substrates [1, 2]. The CYP enzymes play critical roles both in biosynthesis and in the degradation of xenobiotic compounds. The name P450 derives from the position of the Soret absorption maximum in the ferrous-carbonmonoxy form of the enzyme. Ground and excited-state mixing of heme and cysteine thiolate orbitals accounts for the red-shifted Soret signature [3].

Decades of research have produced a consensus mechanism for CYP catalysis (Figure 1). The resting ferric enzyme is low-spin, six-coordinate, with a H₂O ligand occupying the distal axial site trans to the cysteine thiolate ([(Por)(^CysS)Fe^III(OH₂)]). Substrate (RH) binding in the distal heme pocket induces dissociation of the aquo ligand with a concomitant conversion to high-spin and increase in the Fe^III/II formal potential. The shift in potential enables electron transfer from NAD(P)H via a reductase, generating a ferroheme that binds dioxygen to produce a ferric-superoxide heme ([(Por)(^CysS)Fe^III(O₂⁻)]⁻). A second electron transfer, coupled with a proton transfer, leads to a ferric-hydroperoxide intermediate ([(Por)(^CysS)Fe^III(O₂H)]⁻). Heterolytic O─O bond cleavage produces the key intermediate, Compound I (Cmpd-I), a ferryl coupled to a ligand (porphyrin or thiolate) radical ([(Por^•)(^CysS)Fe^IV(O)], [(Por)(^CysS^•)Fe^IV(O)]). Cmpd-I abstracts a hydrogen atom from the substrate, generating transient Compound II ([(Por)(^CysS)Fe^IV(OH)], Cmpd-II) and a substrate radical (R^•). Hydroxyl rebound from Cmpd-II back to R^• produces the oxygenated product (ROH) and completes the catalytic cycle. Detours from this canonical pathway lead to uncoupling of O₂ consumption from ROH production. Loss of O₂⁻ from [(Por)(^CysS)Fe^III(O₂⁻)]⁻ (autoxidation shunt), loss of H₂O₂ from [(Por)(^CysS)Fe^III(O₂H)]⁻ (peroxide shunt), and electron transfer to Cmpd-I (oxidase shunt) can compete with ROH production.

Figure 1.

Consensus mechanism for cytochrome P450 catalysis. Dashed arrows indicate uncoupling pathways.

A longstanding curiosity about CYP enzymes is their capacity to oxygenate refractory substrates such as saturated hydrocarbons without inflicting damage on the protein or its cofactor [4, 5]. The Cmpd-I intermediate, in particular, is a powerful oxidant. If Cmpd-I forms in the absence of substrate, or in the presence of a substrate that cannot be oxygenated, how does the enzyme avoid damage? We have suggested that in this situation, electron transfer along a chain of tryptophan (Trp) and tyrosine (Tyr) residues can direct oxidizing equivalents (holes) away from the heme and toward the enzyme surface where they can be scavenged by cellular reductants [6-11].

To test this hypothesis, we have investigated catalysis in cytochrome P450_BM3 (CYP102A1), a bacterial enzyme from Bacillus megaterium that is a natural fusion protein of reductase and heme domains. P450_BM3 substrates are long chain fatty acids which are hydroxylated at the ω-1, ω-2, and ω-3 positions by the enzyme [12]. We have used p-nitrophenoxydodecanoic acid (12-pNCA) as a substrate because ω-1 hydroxylation produces an unstable hemiacetal that releases spectrophotometrically detectable p-nitrophenolate [13]. Because P450_BM3 can catalyze several hundred turnovers of 12-pNCA, enzyme damage must be a relatively minor pathway compared to productive catalysis. Any protective effects of Trp/Tyr chains are, therefore, unlikely to be detectable in measurements of enzyme kinetics. An enzyme protected by hole transfer along a Trp/Tyr chain, however, will complete more substrate turnovers than one lacking this pathway. Consequently, the total number of enzyme turnovers before deactivation (TTN) should be an indicator of protective pathways in catalysis. We have tested the role of Trp/Tyr chains in protecting cytochrome P450_BM3 by determining TTN values in wild-type enzymes and mutants with disrupted hole-transfer chains.

Materials and Methods

Details of substrate (12-pNCA) synthesis and characterization are provided in Supporting Information. All other reagents were commercially available. Full-length wild-type and mutant cytochromes P450_BM3 (CYP102A1) were expressed in E. coli with a C-terminal (His)₆ tag for isolation and purification according to published procedures [14]. Full details of expression, isolation, purification, and characterization are provided in Supporting Information.

Enzyme activity kinetics were evaluated spectrophotometrically by measuring the initial rate of of p-nitrophenolate appearance immediately after addition of NADPH to a solution of enzyme and 12-pNCA. Complete details and results are in Supporting Information. TTN values were determined by allowing the enzymes to continue oxidizing substrate, periodically removing product and adding more NADPH, until p-nitrophenolate production ceased. Total amount of product formed was determined spectrophotometrically. Coupling efficiency was evaluated by spectrophotometric determination of NADPH consumption.

Kinetics modeling was performed using the Matlab (The MathWorks, Inc) programming and numerical computing platform (details are in Supporting Information). Trp/Tyr hole transfer networks were identified using graph visualization tools implemented in the Bioinformatics Toolbox of the Matlab platform.

Results and Discussion

Kinetics Modeling.

We have developed a simplified model of P450_BM3 catalysis (Figure 2) to predict the effects of Trp/Tyr chain disruption on TTN values. We combine into a single step (k_F) all of the processes involved in converting [(Por)(^CysS)Fe^III(OH₂)] into [(Por)(^CysS)Fe^III(O₂H)]⁻, implicitly assuming that the intermediate conversions are rapid and not substantially altered by Trp/Tyr mutations. We assume that two forms of Cmpd-I result from heterolytic O─O bond cleavage: the functional Cmpd-I forms with rate constant k_C-I; and a nonfunctional ferryl-ligand radical intermediate (Cmpd-I′) that forms with rate constant k_C-I′. Cmpd-I hydroxylates the substrate with rate constant k_P, but Cmpd-I′ fails to generate product, owing to the absence of substrate or the presence of a substrate that cannot be oxygenated. We propose that Cmpd-I′ leads to deactivated enzyme (k_D) but can be rescued by hole transfer away from the heme (k_R). We also assume that the rescue pathway can direct holes away from the functional Cmpd-I intermediate with the same rate constant (k_R). A peroxide shunt pathway also is included (k_H) to model the uncoupling reactions. Although this is an extremely simplified version of P450 catalysis, the rate law is still too complex to solve analytically. We can, however, develop an expression for TTN and coupling efficiency by considering the relative yields of the three branches originating from [(Por)(^CysS)Fe^III(O₂H)]⁻.

Figure 2.

Simplified kinetics model of cytochrome P450_BM3 catalysis, uncoupling, and inactivation.

$$\text{TTN}=\frac{{\varphi }_{\text{ROH}}}{{\varphi }_{\text{DEAD}}}=\frac{{\varphi }_{\mathrm{C}\text{−}\mathrm{I}}{\varphi }_{k_{\mathrm{P}}}}{{\varphi }_{\mathrm{C}\text{−}{\mathrm{I}}^{\prime }}{\varphi }_{\mathrm{D}}}=R_{\mathrm{C}\text{−}\mathrm{I}}\frac{(1+R_{\mathrm{D}})}{(1+R_{\mathrm{P}})}\approx R_{\mathrm{C}\text{−}\mathrm{I}}(1+R_{\mathrm{D}})$$ (1)

TTN will be given by the relative yields of ROH production and enzyme death (eq. 1). In this approximation, TTN depends on the ratio of Cmpd-I and Cmpd-I′ formation rate constants (R_C-I = k_C-I/k_C-I′), the ratio of rate constants for rescue and damage (R_D = k_R/k_D), and the ratio of rate constants for rescue and product formation (R_P = k_R/k_P). If the rescue pathway does not compete with product formation (i.e., R_P ≪ 1), then TTN is approximately equal to R_C-I(1 + R_D). The efficiency of coupling ROH production to O₂

$$\text{Coupling}=\frac{1}{\left(1+\frac{1}{R_{\mathrm{C}\text{−}\mathrm{I}}}+R_{\mathrm{H}}\right)(1+R_{\mathrm{P}})}\approx \frac{1}{\left(1+\frac{1}{R_{\mathrm{C}\text{−}\mathrm{I}}}+R_{\mathrm{H}}\right)}$$ (2)

consumption depends on the same parameters as well as the ratio of rate constants for H₂O₂ production and Cmpd-I formation (R_H = k_H/k_C-I) (eq. 2).

The critical parameter for enzyme protection in this model is k_R, the rate constant for delivery of holes away from Cmpd-I′. Using semiclassical electron transfer (ET) theory and the structure of the P450_BM3 heme domain (PDB ID 2IJ2) [15], we can estimate the survival time of a high-potential hole at the heme in wild-type and mutants of P450_BM3. Details of the rate calculations are in Supporting Information. Analysis of Tyr/Trp hole transfer pathways from the P450_BM3 heme with a 10-Å cutoff for maximum ET distance reveals a direct pathway from the heme through W96 to surface exposed residues W90 and Y334 (Figure 3A). Relaxing the distance constraint to 15 Å exposes the possible involvement of many more residues and pathways (Figure 3B). W96, the closest Trp or Tyr residue to the heme, is a likely gateway for hole transfer from the heme. Y115 is the next closest redox-active residue, but kinetics simulations do not predict a substantial enhancement of heme hole survival over the W96H mutant. Calculations indicate that residues W325 and Y256, each 12.8 Å from the heme, would accept holes if W96 were unavailable. We estimated heme hole survival times for wild-type and four mutants (W96H, W96H/W90F, W96H/W90F/Y334F, W96H/Y256F/W325F; see Supporting Information) (Figure 4). Removal of W96 from the hole transfer pathway is expected to protract the lifetime of a hole on the heme by about a factor of 8. Replacement of additional residues in the W96-W90-Y334 pathway has only a modest impact. An additional fourfold increase in heme hole lifetime is estimated for the W96H/Y256F/W325F triple mutant.

Figure 3.

Trp/Tyr hole transfer networks in P450_BM3. (A) 10-Å ET distance cutoff. (b) 15-Å ET distance cutoff. Asterisks denote surface-exposed residues.

Figure 4.

Predicted heme-hole survival kinetics in wild-type and mutant cytochromes P450_BM3.

Enzyme Activity.

We evaluated the activity of wild-type and mutant P450_BM3 by determining the initial rates of p-nitrophenolate production as functions of 12-pNCA concentration (2-128 μM). All enzymes exhibited rate saturation at [12-pNCA] > 60 μM (Supporting Information). Fits of the data to a Michaelis-Menten kinetics model produced k_cat and K_M values that varied by about a factor of 2.5 among the different enzymes, although the enzyme activities defined by k_cat/K_M remained relatively constant (Table 1). The mutations tended to lower the maximum turnover rate compared to the wild-type enzyme with the largest decrease appearing in the W96H/W90F mutant. The K_M values were lower in the mutants, suggesting somewhat tighter substrate binding. Care should be taken when interpreting these parameters, as the complex P450 mechanism (Figure 1) is a far cry from the ideal Michaelis-Menten model.

Table 1.

Parameters extracted from wild-type and mutant cytochrome P450_BM3 kinetics measurements.

	Wild-type	W96H	W96H/W90F	W96H/W90F/Y334F	W96H/W256F/Y325F
kcat (s−1)	1.8±0.3	1.7±0.1	0.7±0.1	1.45±0.3	1.2±0.2
KM (μM)	22±9	17.5±4.0	9±2	14.±4	9±4
kcat/KM (M−1s−1)	(9.1±2)×104	(1.0±0.2)×105	(8.2±2)×104	(1.1±0.2)×105	(1.4±0.4)×105

Total turnover numbers.

We allowed the wild-type and mutant enzymes to react with substrate, periodically removing product and replenishing NADPH and substrate, until consumption of NADPH ceased. At the termination of catalysis, the amount of p-nitrophenolate produced was determined spectrophotometrically and divided by the amount of enzyme to give TTN values (Figure 5, Table 2). On average, each wild-type enzyme converts (3.4±0.2)×10² substrate molecules before deactivation. The mutant protein TTN values are roughly comparable to one another, in the range of 48-70% of wild-type TTN.

Figure 5.

TTN values for wild-type and W96H, W96H/W90F, W98H/W90F/Y334F, and W96H/Y256F/W325F mutants of cytochrome P450_BM3.

Table 2.

TTN values and coupling efficiencies for wild-type and mutant cytochromes P450_BM3.

	Wild-type	W96H	W96H/W90F	W96H/W90F/Y334F	W96H/W256F/Y325F
TTN	(3.4±0.2)×102	(2.25±0.03)×102	(1.8±0.25)×102	(1.6±0.2)×102	(2.4±0.1)×102
Coupling (%)	12.1±0.1	8.9±2.2	9.1±0.9	9.9±1.2	8.9±1.3

These observations are broadly consistent with predictions of the kinetics modeling. The W96H mutation is expected to have the largest impact on k_R, lowering the R_D ratio, and reducing the TTN value. The W90F and Y334F mutations are not predicted to lower k_R substantially, and we observe only modest further reductions in TTN. The W96H/Y256F/W325F enzyme is expected to have a fourfold smaller k_R value, but its TTN value is close to that of the W96F mutant. This observation suggests that the decrease in k_R resulting from the W96H mutation leads to R_D ≪ 1, so that further reductions in k_R have little impact on TTN.

The kinetics model suggests that the coupling efficiency should not be greatly affected by changes in k_R (eq. 2). Our observation that coupling efficiencies decreased in the mutant proteins (Table 1) may indicate that the kinetics of the peroxide shunt (k_H) and/or the autoxidation pathways (not included explicitly in our model) may not be the same for all P450s in our study.

Concluding Remarks

Replacement of selected Trp and Tyr residues in the heme domain of cytochrome P450_BM3 negatively impacts enzyme survival. A possible explanation for this behavior is that the Trp and Tyr residues are members of a hole transfer chain that deliver potentially damaging oxidizing equivalents away from the heme when substrate oxidation is prevented. Our simplified model of P450_BM3 catalysis and protection is consistent with our observations of total turnover numbers in wild-type and mutant enzymes. W96 appears to be the gateway residue for the primary Trp/Tyr protection pathway. Sequence alignments suggest that this position is more commonly occupied by a histidine residue in bacterial enzymes, but tryptophan is found at this position in more than 80% of mammalian P450s [16]. High enzyme turnover rates have been associated with increased energy demand and reduced productivity in microbes and plants [17, 18]. In the case of P450_BM3 it is difficult to assess whether the twofold increase in enzyme turnover resulting from the placement of a tryptophan residue at position 96 offers a substantial selective advantage to the organism. Full-length cytochrome P450_BM3 is relatively large and a conservative estimate suggests that more than 4700 ATP molecules are required to degrade and resynthesize one enzyme [19]. This energetic cost may be great enough to provide a selective advantage for a wild-type enzyme that functions two times longer than a single site variant (e.g., W96H).

Abbreviations:

P450_BM3

flavocytochrome P450 from Bacillus megaterium

Cmpd-I

cytochrome P450 Compound I

Cmpd-II

cytochrome P450 Compound II

12-pNCA

p-nitrophenoxydodecanoic acid

Trp

tryptophan

Tyr

tyrosine

TTN

total turnover number

Data Availability

The data that support the findings of this study are available in Table 1, Table 2, Figure 4, Figure 5, and the supplementary material of this article.