Substrate Routes to the Buried Active Site May Vary among

James B.Y.H. Behrendorff , Weiliang Huang , Elizabeth M.J. Gillam ... Engineering of LadA for enhanced hexadecane oxidation using random- and site-dir...
0 downloads 0 Views 115KB Size
Download PDF
Chem. Res. Toxicol. 2002, 15, 1407-1413

1407

Substrate Routes to the Buried Active Site May Vary among Cytochromes P450: Mutagenesis of the F-G Region in P450 2B1† Emily E. Scott,* You Qun He, and James R. Halpert Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, Texas 77555 Received June 25, 2002

Until recently, all known structures of bacterial cytochromes P450 suggested that substrate access to the buried active site occurred via the F-G region, a surface loop distal to the heme cavity. However, the structure of P450 51 indicates a large opening from the protein surface along the I helix N-terminus, at right angles to the F-G channel. The single available microsomal P450 structure (2C5) does not obviously favor one potential access route over the other. To determine whether the F-G region forms part of the substrate access channel in the microsomal cytochrome P450 2B1, 11 residues between positions 208 and 230 were substituted with smaller and larger side chains in a highly expressed truncated form of the enzyme. Steadystate kinetic parameters were determined with the substrates testosterone, 7-ethoxy-4trifluoromethylcoumarin (7-EFC), and 7-benzyoxyresorufin (7-BR). The largest changes, 2-6fold increases in kcat with testosterone and 7-EFC, were observed for L209A, which also exhibits an altered testosterone metabolite profile and probably forms part of the active site roof. F219W demonstrated little or no activity with any of the three substrates examined, although the Ks value for benzphetamine binding was unaltered. S221F showed little activity with 7-BR. No significant changes were observed in Km(testosterone) or S50(7-EFC) values for any of the mutants, in stark contrast to the 10-fold and 100-fold changes in Km observed for mutants in this region of other cytochromes P450. The minimal changes in 2B1 do not support access via the F-G region of 2B1 and suggest the alternate access route identified in P450 51.

Introduction Cytochromes P450 are a large superfamily of monooxygenases involved in the catalysis of a wide variety of endogenous and xenobiotic substrates. An intriguing aspect of many P450 enzymes, including those of family 2, involves the differential regio- and stereoselectivity of highly structurally related proteins. Until recently, P450 research has focused on understanding the interplay of substrate and active site characteristics to explain selectivity. For P450 2B enzymes, steric constraints imposed by the enzymes have been shown to play a major role in substrate specificity (1). However, most structures of cytochromes P450 reveal that the heme group is buried deep within the protein matrix, indicating that protein motion is required for substrates to reach the active site and suggesting that substrate accessibility to the active site may also play a significant role in selectivity. Structural, mutagenesis, and computational evidence from several bacterial P450 enzymes indicates the possibility of an access route near a region between the Fand G-helices. In the structure of P450 101, the F-G region forms part of a small opening above the ligand binding site that initiates a continuous channel to the active site (2, 3) (Figure 1). Although comparison of P450 † This work was supported by National Research Service Award GM20674 (E.E.S.), Grant ES03619 (J.R.H.), and Center Grant ES06676. * To whom correspondence should be addressed at the Department of Pharmacology and Toxicology, University of Texas Medical Branch, 301 University Blvd., Galveston, TX 77555-1031. Tel: (409) 772-9670; Fax: (409) 772-9642; E-mail: [email protected].

Figure 1. Cartoon showing the general spatial relationship between the two potential substrate access channels to the heme active site revealed in the X-ray structures of P450 101/102 and P450 51 in a homology model of 2B1 (28) based on the 2C5 structure (14). The F- and G-helices are shown in dark gray ribbons at the top of the figure. The I-helix is shown in light gray ribbons just above the heme (stick representation). The B/C loop is represented as a thin medium gray backbone.

101 with and without bound camphor reveals little change in the protein structure, the B factors, which often correlate with important dynamic motions, are much higher for the F-G region in the substrate-free vs camphor-bound structure (4). Similarly, in the structure of P450 102, the heme is accessible via a hydrophobic channel partially formed by the B′- and F-helices. In this case, the largest changes between the substrate-free and substrate-bound forms occur in the F-G region, resulting in a narrowing of the channel when substrate is bound

10.1021/tx020057u CCC: $22.00 © 2002 American Chemical Society Published on Web 10/09/2002

1408

Chem. Res. Toxicol., Vol. 15, No. 11, 2002

(5). A Phe-42 f Ala substitution at the mouth of the substrate channel in P450 102 increases the Km 442-fold and 7-fold for two different substrates, indicating that this may be a gatekeeping residue controlling substrate entry into the binding pocket (6). In neither enzyme, however, is the channel of sufficient diameter to permit the entry of normal ligands without movement of the protein backbone. Molecular dynamics simulations suggest that for ligand entry, gatekeeper residues would move first, followed by rotation of the F-G domain to give dramatic opening of the access channel (7-11). Molecular dynamics simulations supplemented with an artificial randomly oriented force on the substrate have identified multiple potential substrate exit routes, with the most probable and consistent route including the F-G loop (9-12). The exit channel in this type of simulation is thought to be the same as the substrate access channel. Recently a very different substrate access channel has been described for a sterol 14R-demethylase (P450 51) from Mycobacterium tuberculosis (13). The most obvious access route originates between the B-C loop and the N-terminus of the I-helix and continues along the I-helix in the plane of the heme to the active site (Figure 1). The opening of this channel is very large, 20 Å by 10 Å. In P450 51, the channel described above in P450 101 and 102 is apparent but capped at the protein surface by interactions between the A′-helix and the F-G region, which includes a short F′-helix. The thermal factors of this region are also high, however, and it could potentially undergo the types of motions explored in the molecular dynamics simulations on P450 102. This exit route has also been identified from modified molecular dynamics simulations of P450 101. The probability of exit in this direction is approximately 0.1 that of exit along the F-G loop, but longer simulations might emphasize this exit route (11). The only mammalian P450 whose structure is known, an engineered form of rabbit 2C5, appears to be intermediate between the two morphologies described above. The F-G region fills the P450 101/102-like substrate access channel but could be capable of the motions necessary for substrate entry. The density for these residues is ambiguous, complicated by possible partial occupancy by a bound ligand or ligands and constraint by a crystal packing contact (14). A P450 51-like pathway is also apparent between the I-helix and the B-C loop. Substrate entry via this point would also require protein flexibility, consistent with the high-temperature factors observed in the B-C loop. The variety of potential ligand access channels surmised from crystallographic structures has given rise to the idea that different enzymes or classes of cytochromes P450 may have different routes of substrate entry. In addition, suggestions have been made that some cytochromes P450 may have multiple routes from solvent to the active site, each of which might accommodate different kinds of substrates or serve as differential substrate entry and product exit routes (13, 14). Site-directed mutagenesis studies of various mammalian P450s provide support for both channels. In P450 4A2 and 4A3 (15) and 2B6 (16) residues in the B′-helix that appear to be too far from the heme to be in direct contact with substrate contribute nonetheless to substrate specificity and/or regioselectivity, whereas F-G region residues in P450 2B5 (17), P450 2C19 (18), and

Scott et al.

P450 27A1 and 7A1 (19, 20) play a similar role. In P450 2C9, nonsubstrate contact residues in the I-helix contribute to substrate specificity (21), consistent with a P450 51-like channel. Thus, at present, the identity of the access channel in a given mammalian P450 and the role that surface recognition and access play relative to substrate binding near the heme iron remain unresolved. Therefore, to determine whether the F-G region indicated in P450 101 and 102 structures might form part of the substrate access channel in the microsomal cytochrome P450 2B1, 11 amino acids in this area of the protein were mutated to smaller and larger residues, and steady-state kinetics were determined with the typical 2B substrates testosterone, 7-ethoxy-4-trifluoromethylcoumarin (7-EFC),1 and 7-benzyloxyresorufin (7-BR).

Experimental Procedures Materials. Oligonucleotide primers were obtained from the University of Texas Medical Branch Molecular Biology Core Laboratory (Galveston, TX). 7-BR and the resorufin standard were obtained from Sigma. All other reagents were obtained from sources previously described (22) or from standard suppliers. Mutagenesis, Expression, and Purification. The truncated version of 2B1 that served as the background for all F-G mutations described in this study, 2B1dH, was generated using overlap extension PCR as described (22). Basically, residues 3-21 have been deleted, several mutations made at the new N-terminus, and a four histidine tag added at the C-terminus. All F-G mutants were generated using overlap extension PCR with the primers shown in Figure 2 and either the truncated construct (pKK2B1dH) or the full-length His-tagged construct as the template (pKK2B1wtH) except for F209A and S221P. These two mutants had been previously generated in the fulllength enzyme without the His tag (17, 23). Mutants originally made in the full-length versions of the 2B1 cDNA were subcloned into the truncated 2B1dH background using the unique restriction sites PstI and KpnI. All constructs were sequenced to confirm the desired mutation and verify the absence of unintended mutations (Protein Chemistry Laboratory, University of Texas Medical Branch, Galveston, TX). 2B1dH and all mutants were expressed in E. coli TOPP3 (Stratagene) and purified as described (22). Cytochrome P450 was quantitated using the reduced-CO difference spectrum (24). Specific contents for the purified proteins were 8.6 nmol of P450 (mg of protein)-1 for 2B1 full-length wild type, 20.9 nmol of P450 (mg of protein)-1 for 2B1dH, and 7.3-14.9 nmol of P450 (mg of protein)-1 for 2B1dH F-G mutants except for F219W which had a specific content of 4.6 nmol of P450 (mg of protein)-1. Enzymatic Assays. Testosterone hydroxylase activity was assayed essentially according to the protocol described by Ciaccio and Halpert (25). The final 100 µL reaction mixture contained 5 pmol of P450, 20 pmol of cytochrome P450 reductase, 10 pmol of cytochrome b5, and 0.5-250 µM [14C]testosterone [20 000 dpm (nmol)-1] in 50 mM HEPES, 15 mM MgCl2, and 0.1 mM EDTA, pH 7.6. The reaction was initiated by the addition of 1 mM NADPH, allowed to proceed for 5 min at 37 °C, and stopped with 50 µL of tetrahydrofuran. Metabolites were resolved on TLC plates by two cycles of chromatography in (4:1 v/v) dichloromethane/acetone, identified by autoradiography, and quantitated by scintillation counting. 7-EFC oxidation was measured in a final volume of 100 µL. The reaction mixture contained 2.5 pmol of P450, 10 pmol of cytochrome P450 reductase, 5 pmol of cytochrome b5, and 2.5160 µM EFC in 50 mM HEPES, 15 mM MgCl2, and 0.1 mM 1 Abbreviations: 7-BR, 7-benzyloxyresorufin; 7-EFC, 7-ethoxy-4-trifluoromethylcoumarin; DLPC, dilauryl-L-3-phosphatidylcholine; HEPES, N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid; PCR, polymerase chain reaction.

P450 2B1 Substrate Access Channel

Chem. Res. Toxicol., Vol. 15, No. 11, 2002 1409 Table 1. Steady-State Kinetics of Testosterone, 7-EFC, and 7-BR Oxidation by P450 2B1 Wild Type vs 2B1dH testosterone P450 2B1 WT 2B1dH

kcata 16R

kcat 16β

8.0 6.9 (0.1)c (0.2) 4.6 3.5 (0.2) (0.2)

7-EFC

Kmb Km 16R 16β 24 (2) 27 (4)

kcat

n

24 9.8 1.3 (3) (0.5) (0.1) 29 6.2 1.3 (4) (0.7) (0.3)

7-BR S50b

kcat

26 (3) 20 (5)

7.0 2.5 (0.3) (0.3) 3.5 2.5 (0.3) (0.5)

Km

a k , min-1. b K , S , µM. c Standard error for fit to Michaeliscat m 50 Menten or Hill equation.

Figure 3. Alignment of the sequences from the crystallized 2C5 structure (2C5/3LVdH) and 2B1 in the region of the F-helix, F-G loop, and G-helix. 2B1 numbering is shown above. Helices (2C5) are boxed. Residues whose electron density is uncertain in the 2C5 structure are shaded. Boldface lettering denotes the position of mutations made in 2B1. Spectral Binding. Binding spectra were recorded on a Shimadzu-2600 spectrophotometer in the difference mode. A 1.0 mL solution containing 0.8 or 2 µM protein in 10 mM potassium phosphate, 10% glycerol, and 1 mM EDTA was divided into two quartz cuvettes and a baseline recorded between 350 and 500 nm. An aliquot of benzphetamine in water was added to the sample cuvette, and the same amount of water was added to the reference cuvette. The difference spectra were obtained after the system reached equilibrium at 37°C. ∆Amax and Ks values were determined by nonlinear regression using the hyperbolic equation: ∆A ) ∆AmaxS/Ks + S.

Results

Figure 2. Primers for the construction of F-G mutations using overlap extension PCR. Nucleotides changed from the original sequence, either to make the desired mutation or to engineer a restriction site without changing the amino acid, are shown in boldface type. The mutant codon is underlined. EDTA, pH 7.6. The reaction was initiated by the addition of NADPH (1 mM), incubated for 10 min at 37 °C, and terminated by the addition of 20% trichloroacetic acid. Metabolites were quantitated using fluorescence and a 7-hydroxy-4-trifluoromethylcoumarin standard as described (16). 7-BR O-debenzylation was measured essentially using the standard fluorometric assay as described (26). The 500 µL final reaction mixture contained 20 pmol of P450, 80 pmol of cytochrome P450 reductase, 40 pmol of cytochrome b5, and substrate (0.01-10 µM) in 50 mM HEPES, pH 7.6, 15 mM MgCl2, and 0.1 mM EDTA. The reaction was started by adding 0.5 mM NADPH, continued for 10 min at 37 °C, and terminated with the addition of 2 mL of methanol. Resorufin formed was monitored by comparison with a standard curve using a PerkinElmer model 3000 spectrofluorometer with excitation at 550 nm and emission at 585 nm. Km, kcat, S50, and n values were calculated using MichaelisMenten or Hill nonlinear regression analysis using SigmaPlot (Jandel Scientific, San Rafael, CA).

Comparison of 2B1 and 2B1dH Enzyme Kinetics. All F-G mutants were examined in the background of a truncated form of the 2B1 enzyme, termed 2B1dH. This version has been engineered for high expression and increased solubility (22). To verify the functional similarity of 2B1dH to the full-length enzyme, the steady-state kinetics of testosterone, 7-EFC, and 7-BR oxidation were compared. These substrates were chosen because they are all within the ranges of molecular weight, volume, pKa, and log P of typical 2B substrates (27). To maximize the activity of the truncated and full-length versions of the enzyme individually, there are two differences in the enzymatic assays, as previously described (22). First, the cytochrome P450:cytochrome P450 reductase:cytochrome b5 ratio was 1:4:2 for the truncated enzyme compared with 1:2:1 for the full-length enzyme. Second, DLPC was omitted from the assays using protein in the 2B1dH background, but present [0.03 mg of DLPC (mL of reaction)-1] in reconstitutions of the full-length enzyme. Both proteins show hyperbolic kinetics for testosterone and 7-BR oxidation but sigmoidal kinetics for 7-EFC. For all three substrates, the Km or S50 values for the two proteins are very similar, but the kcat for 2B1dH is about half of the wild type (Table 1). Selection of Mutants. Residues in the F-G region that might be likely to form part of a 2B1 substrate access channel were chosen using a homology model of P450 2B1 (28) based on the P450 2C5 structure. Cytochromes P450 2B1 and 2C5 share 50% amino acid identity. The published 2C5 structure is missing electron density for 11 residues in the F-G region (Figure 3). However, the 2B1 model based on unpublished experimental electron

1410

Chem. Res. Toxicol., Vol. 15, No. 11, 2002

Scott et al.

Table 2. Steady-State Substrate Kinetics for F-G Mutants in the 2B1dH Background testosterone kcata P450

16R

2B1dH 4.6 (0.2)c L208A 4.0 (0.1) L209A 29 (0.6) S213A 4.4 (0.1) S213W 6.3 (0.3) F220A 6.4 (0.1) F220W 2.2 (0.1) S221P 7.7 (0.2) G222F 6.4 (0.2) F227A 2.8 (0.1) A230F 5.6 (0.2) a

b

kcat 16β

Km 16R

Km 16β

3.5 (0.2) 2.3 (0.1) 8.1 (0.2) 3.2 (0.1) 4.7 (0.2) 4.5 (0.1) 1.7 (0.1) 5.8 (0.2) 4.6 (0.2) 1.9 (0.1) 3.9 (0.1)

27 (4) 24 (2) 19 (2) 22 (2) 26 (4) 25 (2) 28 (3) 17 (2) 24 (3) 28 (3) 24 (3)

29 (4) 25 (2) 20 (2) 22 (2) 24 (3) 27 (2) 30 (5) 19 (2) 20 (3) 31 (8) 29 (4)

7-EFC kcat/Km kcat/Km 16R 16β 0.17 0.16 1.5 0.20 0.24 0.26 0.08 0.46 0.26 0.10 0.24

0.12 0.09 0.41 0.15 0.19 0.16 0.06 0.30 0.23 0.06 0.13

7-BR

kcat

n

S50b

kcat/S50

kcat

Km

kcat/Km

14 (0.4) 12 (0.5) 34 (0.9) 15 (0.8) 14 (0.8) 18 (0.5) 4.0 (0.4) 23 (0.6) 16 (0.7) 5.9 (0.6) 12 (0.8)

1.7 (0.2) 1.9 (0.3) 1.2 (0.1) 1.8 (0.2) 1.7 (0.2) 1.6 (0.1) 2.7 (1) 1.6 (0.1) 1.5 (0.2) 2.0 (0.5) 1.7 (0.3)

23 (2) 22 (0.1) 8.3 (0.6) 33 (3) 30.3 (3) 21 (1) 34 (6) 25 (2) 30 (3) 35 (6) 31 (4)

0.63 0.53 4.1 0.47 0.48 0.88 0.12 0.89 0.53 0.17 0.38

3.5 (0.3) 4.5 (0.3) 3.7 (0.2) 4.8 (0.3) 3.2 (0.2) 6.5 (0.6) 2.5 (0.2) 4.6 (0.4) 6.9 (0.5) 1.3 (0.1) 2.4 (0.2)

2.5 (0.5) 3.6 (0.6) 2.6 (0.4) 3.6 (0.6) 1.6 (0.3) 3.9 (0.8) 2.3 (0.4) 3.2 (0.7) 4.0 (0.7) 1.1 (0.3) 2.2 (0.6)

1.4 1.3 1.4 1.3 1.9 1.7 1.1 1.4 1.7 1.2 1.1

kcat, min-1. b Km, S50, µM. c Standard error for fit to Michaelis-Menten or Hill equation. Table 3. Steady-State Substrate Oxidation Parameters of F-G Mutants in the 2B1dH Background testosterone

P450

kcata 16R

kcat 16β

Kmb 16R

Km 16β

2B1dH F219A F219W S221A S221F F223A F223W

4.6 (0.2)c 3.7 (0.1) NDd 6.1 (0.1) 2.7 (0.1) 3.0 (0.1) 5.9 (0.2)

3.5 (0.2) 2.9 (0.1) ND 5.3 (0.1) 1.8 (0.1) 2.2 (0.1) 5.5 (0.1)

27 (4) 24 (2) ND 28 (2) 55 (5) 32 (3) 22 (2)

29 (4) 24 (2) ND 31 (2) 35 (8) 31 (4) 26 (2)

7-EFC kcat/Km kcat/Km 16R 16β 0.17 0.15

0.12 0.12

0.22 0.05 0.09 0.27

0.17 0.05 0.07 0.21

kcat

n

7-BR S50b

6.2 (0.7) 1.3 (0.3) 20 (5) 5.0 (0.2) 1.5 (0.2) 17 (2) ND ND ND 5.0 (0.4) 1.6 (0.3) 34 (6) 1.2 (0.4) 1.8 (1) 31 (20) 2.1 (0.3) 1.4 (0.3) 36 (10) 4.7 (0.6) 1.2 (0.3) 17 (6)

a k , min-1. b K , S , µM. c Standard error for fit to Michaelis-Menten or Hill equation. cat m 50 0.1 min-1 for testosterone hydroxylation and 0.08 min-1 for 7-EFC oxidation.

density suggests that this region may encompass a short helix (28), as is seen in P450 51 (Figure 1). Throughout the F-G region, individual residues were selected for their proximity to the putative F-G channel in the 2B1 homology model, the potential for changes in the size of the side chain, and conservation among 2B enzymes. Mutants were designed to investigate changes in side chain bulk by substituting a smaller Ala where possible and/or a larger Phe/Trp, depending on the size and identity of the native residue. Special attention was focused on the stretch of residues surrounding Ser-221, which is a surface residue. In 2B5, the P221S mutant shows reduced progesterone hydroxylase activity but an unaltered product profile (17). In 2C9, the corresponding Ser f Pro substitution helps to confer omeprazole hydroxylase activity (18). These effects of mutagenesis on substrate specificity and overall activity are consistent with the role of potential substrate access channel residues in the F-G region. Effects of F-G Mutations on Metabolism of Testosterone, 7-EFC, and 7-BR. Twenty mutants at 11 different positions throughout the F-G region were constructed (L208A/F, L209A/F, S210G/A, S213A/W, F219A/W, F220A/W, S221A/P/F, G222F, F223A/W, F227A, and A230F). A subset of the mutants constructed did not express holoprotein in E. coli, although significant amounts of apoprotein were detected by Western blot analysis (L208F, L209F, S210G, and S210A). This result suggests that residues 208-210 may be particularly important for protein folding and/or stability. This is in noted contrast to 2C5 where a Ser-210 to Gly substitution is easily accommodated, suggesting differences in the local packing of the F-G region. The remaining mutants expressed at 300-600 nmol of P450 (L of E. coli culture)-1 with the exceptions of S213W, F223A/W, and F227A [100-300

d

kcat/S50

kcat

Km

kcat/Km

0.31 0.29

3.5 (0.3) 3.2 (0.3) 0.30 (0.0) 2.8 (0.2) 0.06 (0.0) 1.4 (0.1) 3.8 (0.2)

2.5 (0.5) 1.7 (0.4) 0.11(0.0) 1.6 (0.4) 0.21(0.0) 0.45 (0.1) 1.3 (0.2)

1.4 1.9 2.7 1.8 0.30 3.2 2.9

0.15 0.04 0.06 0.27

ND, product was below the limits of detection,

nmol of P450 (L of culture)-1] and S213A [80 nmol (L of culture)-1]. The 16 mutant proteins were tested for testosterone, 7-EFC, and 7-BR kinetics. The results of an initial panel of mutants including residues 208, 209, 213, 220, 222, 227, and 227 are shown in Table 2. Based on the results from this survey, six additional mutants were examined at or near positions 220 and 221 (F219A/W, S221A/F, and F223A/W), as shown in Table 3. It should be noted that the 7-EFC results presented were unavoidably obtained with different lots of the substrate. As documented both by the manufacturer (Gentest) and by experience in our laboratory, this can lead to different kcat but not S50 values. Thus, for the 7-EFC assay, mutants should only be compared with the 2B1dH parameters presented in the same table and which were obtained with the same lot of substrate. Overall, the most striking result is the remarkable lack of kinetic effects for most of the mutants examined. No significant changes are observed for the Km with testosterone, or for the S50 or n for 7-EFC with the exception of L209A. The few mutants that show a greater than 2-fold change in one or more of the kinetic parameters (Tables 2 and 3) are L209A, F219W, F220W, S221F, F223A, and F227A. Of these, the most striking mutants are L209A, F219W, and S221F. L209A shows 2-6-fold increases in the kcat for both testosterone and 7-EFC and the largest increases observed in the kcat/Km and kcat/S50 values, 3-8-fold. In addition, the ratio of testosterone 16R-hydroxylation to 16β-hydroxylation is increased from 1.3 for 2B1dH to 3.6 for the L209A mutant. No other mutant exhibited an altered testosterone metabolite profile. F219W has no detectable testosterone or 7-EFC activity but very low activity with 7-BR such that the kcat/Km is increased by barely 2-fold. S221F has a very

P450 2B1 Substrate Access Channel

Figure 4. Spectral titrations of 2B1dH and the mutant F219W with benzphetamine. The ligand concentrations for 2B1dH are 5, 10, 15, 20, 25, 50, 100, and 200 µM. Because of the low ∆A for F219A, data from two separate titrations with ligand concentrations of 5, 10, 15, 20, 25, 50, 100, and 200 µM and 10, 20, 30, 40, 60, 100, and 200 µM were combined. The ∆A values shown are normalized to 1 µM P450. Ks values were 40 and 37 µM, and ∆Amax values were 0.038 and 0.012 µM-1 P450 for 2B1dH and F219W, respectively.

low kcat for 7-BR compared to 2B1dH but 19% and ∼50% of the 2B1dH kcat with 7-EFC and testosterone, respectively. With testosterone and 7-BR, the kcat/Km for S221F is decreased by 2-4-fold while the kcat/S50 for 7-EFC is decreased by 7-fold, the largest decreases observed. F220W, F223A, and F227A demonstrated more moderate changes in the individual kinetic parameters and kcat/ Km. F220W exhibits barely a 2-fold decrease in the kcat for testosterone and a 4-fold decrease in the kcat for 7-EFC, but no change in either kinetic parameter for 7-BR oxidation. The results correspond to 2-fold and 5-fold changes in kcat/Km or kcat/S50 for testosterone and 7-EFC, respectively. Both F223A and F227A show 2-3fold decreases in kcat for both 7-EFC and 7-BR as well as a decrease in the Km for 7-BR. The cumulative results are 4-5-fold decreases in kcat/S50 for both mutants (7EFC) and a 2-fold increase in kcat/Km for F223A (7-BR). Finally, approximately 2-fold changes were observed in the kcat/Km for S221P (testosterone) and F223W (7-BR). Spectral Titration of F219W. To determine whether the low activity of F219W was due to impaired substrate binding as opposed to turnover, spectral titrations of 2B1dH and F219W were performed. Since none of the three substrates examined in the steady-state kinetics demonstrate significant spectral changes upon binding to the enzyme, benzphetamine was used as the ligand. The ∆Amax value for F219W is 3-fold lower than for 2B1dH (Figure 4), but the Ks values are the same. In addition benzphetamine N-demethylation by F219W was unaltered (data not shown). In combination with the very low kcat values for F219W with the other substrates, these data suggest that the substitution at 219 affects oxidation within the active site, rather than substrate entry.

Discussion Substrate interactions with cytochromes P450 have been proposed to occur in three stages: (1) recognition of the substrate by surface residues; (2) entry into the buried active site through a hydrophobic access channel; (3) substrate orientation in the active site to allow

Chem. Res. Toxicol., Vol. 15, No. 11, 2002 1411

catalysis (29, 30). For a given P450, active site residues that control substrate positioning in the active site have been predicted with reasonable accuracy from structurebased sequence alignments (31) with cytochromes P450 of known structure and from analysis of allelic variants with differential activities. The effects of active site residues on substrate metabolism and inhibition have been extensively investigated using site-directed mutagenesis, especially in the family 2 enzymes [see review in (1)]. However, the tertiary arrangement of secondary structure elements can vary among P450 enzymes, particularly at sites more distant from the heme. These factors make prediction of substrate entry routes much more tenuous, particularly in the absence of a variety of high-resolution mammalian P450 structures. The distinction between an active site residue and a substrate access-channel residue may not be definitive in every instance. Both types of residues can affect kinetic parameters, dissociation constants, and ligand binding rates. Since substrates have different sizes and shapes and bind in different orientations, the residues that form the functionally relevant active site may vary for different compounds. For example, a residue that functions as a part of the active site for a large substrate may function as an access channel residue for a smaller substrate. To address this concern, we are have tested mutants with a variety of substrates. In addition, we have tried to weigh the sites selected for mutation in favor of surface residues that may form the substrate entry point rather than residues that would be more interior, closer to the heme, and presumably have a higher probability of functioning as active site residues. However, residue L209, whose substitution yielded the largest observed effects in this study, probably constitutes part of the active site. L209A showed the largest increases in the kcat for both testosterone and 7-EFC, yielding 3-8-fold increases in kcat/Km or kcat/S50 over that of 2B1dH for both testosterone and 7-EFC. In addition, this mutation alone among those examined had a significant effect on the testosterone metabolite profile, altering the 16R-OH:16β-OH ratio from 1.3 to 3.6. In the 2B1 homology model generated from the mammalian 2C5 structure, this residue is the closest one examined to the active site, forming part of the roof of the substrate binding cavity. Altogether, the probable location of the residue, the effects of its substitution on the metabolite profile, and the changes in kcat for two of the three substrates examined are most consistent with a residue more significant in substrate oxidation rather than substrate entry. The lack of significant effects from this mutation on the kinetic parameters of 7-BR oxidation may indicate that 7-BR is oriented in the active site such that the residue at position 209 is not involved. S221F is the only mutant examined that exhibited greater than 2-fold decreases in the kcat/Km or kcat/S50 values for all three substrates and the second largest changes in kcat/S50 overall (7-fold decrease). In addition, S221F shows only very low activity with 7-BR, another potential indication that 7-BR is oriented in the active site differently than testosterone and 7-EFC. It is noteworthy that the log P values of testosterone (3.32) and 7-EFC (3.3) are similar and much lower than that of 7-BR (5.1). The lack of activity for F219W is most likely due to defects in oxidation within the active site, possibly as the result of excessive substrate mobility, since the unaltered Ks value for benzphetamine binding suggests

1412

Chem. Res. Toxicol., Vol. 15, No. 11, 2002

that substrate access has not been impeded. In the mammalian 2C5 structure, the electron density of the residues corresponding to positions 219 and 221 in 2B1 is ambiguous and has been omitted from the final model. In the 2B1 homology model, these residues lie at the end of a putative F′-helix. At both positions, an Ala substitution results in enzymatic activities near those of 2B1dH for all three substrates. The remaining mutants demonstrate smaller effects on the kinetic oxidation parameters for residues in the F-G region of the protein. Of all the mutants examined, none exhibit significant changes in the testosterone Km or 7-EFC S50. The changes observed in the Km for 7-BR are moderate except for F219W and S221F, which have very low activity. In contrast, mutational studies in this region of truncated P450 7A1, which metabolizes cholesterol, revealed 5-12-fold changes in the Km value. Even more impressively, a Phe-42 f Ala substitution at the mouth of the substrate channel in P450 102 increases the Km 442-fold for laurate and 7-fold for arachidonate. If the F-G region is the substrate access channel in cytochrome P450 2B1, then similar large and more generalized changes would be expected to occur in the kinetic parameters upon mutation of the residues in this area. One possible explanation for these differences is a negative correlation between the presence of an F′-helix and the likelihood of substrate entry in the F-G region. The P450 51 crystallographic structure shows an F′-helix, while the 2C5 structure is suggestive of such a secondary element. None of the other soluble P450 enzymes whose structure has been determined have an F′-helix but instead suggest a substrate access channel in this region. One prospect is that those P450 enzymes without an F′helix may allow substrate entry in the F-G region, while those with an F′-helix have a different route for substrate access to the active site. 2B1 would be expected to be most similar in structure to 2C5, and the putative F′-helix may indicate substrate access from another direction. The differences in the ability to express holoprotein with mutations at position 210 in P450 2C5 and P450 2B1 may also indicate differences in the local secondary structure of this region. In conclusion, mutational analysis of the F-G region in P450 2B1 suggests that substrate access may occur via a route that does not involve the residues of the F-G region. In addition, the very intriguing findings of Wade and collaborators suggest that even for bacterial P450 enzymes that share a similar pathway including the F-G region, the precise mechanism of channel opening may be different, reflecting different physicochemical properties of the preferred substrates (12). These results encourage further investigation of the pathways and mechanisms by which substrates gain access to the active site of different mammalian P450 enzymes.

Acknowledgment. We thank Margit Spatzenegger for assistance with the spectral titrations.

References (1) Domanski, T. L., and Halpert, J. R. (2001) Analysis of mammalian cytochrome P450 structure and function by site-directed mutagenesis. Curr. Drug Metab. 2, 117-137. (2) Poulos, T. L., Cupp-Vickery, J., and Li, H. (1995) in Cytochrome P450: Structure, Mechanism, and Biochemistry (Ortiz de Montellano, P. R., Ed.) pp 125-150, Plenum Press, New York.

Scott et al. (3) Poulos, T. L., Finzel, B. C., Gunsalus, I. C., Wagner, G. C., and Kraut, J. (1985) The 2.6-Å crystal structure of Pseudomonas putida cytochrome P-450. J. Biol. Chem. 260, 16122-16130. (4) Poulos, T. L., Finzel, B. C., and Howard, A. J. (1986) Crystal structure of substrate-free Pseudomonas putida cytochrome P-450. Biochemistry 25, 5314-5322. (5) Li, H., and Poulos, T. L. (1997) The structure of the cytochrome P450BM-3 haem domain complexed with the fatty acid substrate, palmitoleic acid. Nat. Struct. Biol. 4, 140-146. (6) Noble, M. A., Miles, C. S., Chapman, S. K., Lysek, D. A., Mackeay, A. C., Reid, G. A., Hanzlik, R. P., and Munro, A. W. (1999) Roles of key active-site residues in flavocytochrome P450 BM3. Biochem. J. 339, 371-379. (7) Chang, Y. T., and Loew, G. H. (1999) Molecular dynamics simulations of P450 BM3sexamination of substrate-induced conformational change. J. Biomol. Struct. Dyn. 16, 1189-1203. (8) Paulsen, M. D., and Ornstein, R. L. (1995) Dramatic differences in the motions of the mouth of open and closed cytochrome P450 BM3 by molecular dynamics simulations. Proteins: Struct., Funct., Genet. 21, 237-243. (9) Ludemann, S. K., Gabdoulline, R. R., Lounnas, V., and Wade, R. C. (2001) Substrate access to cytochrome P450cam investigated by molecular dynamics simulations: An interactive look at the underlying mechanisms. Internet J. Chem. 4, 6. (10) Ludemann, S. K., Lounnas, V., and Wade, R. C. (2000) How do substrates enter and products exit the buried active site of cytochrome P450cam? 1. Random expulsion molecular dynamics investigation of ligand access channels and mechanisms. J. Mol. Biol. 303, 797-811. (11) Ludemann, S. K., Lounnas, V., and Wade, R. C. (2000) How do substrates enter and products exit the buried active site of cytochrome P450cam? 2. Steered molecular dynamics and adiabatic mapping of substrate pathways. J. Mol. Biol. 303, 813830. (12) Winn, P. J., Ludemann, S. K., Gauges, R., Lounnas, V., and Wade, R. C. (2002) Comparison of the dynamics of substrate access channels in three cytochrome P450s reveals different opening mechanisms and a novel functional role for a buried arginine. Proc. Natl. Acad. Sci. U.S.A. 99, 5361-5366. (13) Podust, L. M., Poulos, T. L., and Waterman, M. R. (2001) Crystal structure of cytochrome P450 14R-sterol demethylase (CYP51) from Mycobacterium tuberculosis in complex with azole inhibitors. Proc. Natl. Acad. Sci. U.S.A. 98, 3068-3073. (14) Williams, P. A., Cosme, J., Sridhar, V., Johnson, E. F., and McRee, D. E. (2000) Mammalian microsomal cytochrome P450 monooxygenase: Structural adaptations for membrane binding and functional diversity. Mol. Cell 5, 121-131. (15) Hoch, U., Falck, J. R., and Ortiz de Montellano, P. R. (2000) Molecular basis for the ω-regiospecificity of the CYP4A2 and CYP4A3 fatty acid hydroxylases. J. Biol. Chem. 275, 2695226958. (16) Domanski, T. L., Schultz, K. M., Roussel, F., Stevens, J. C., and Halpert, J. R. (1999) Structure-function analysis of human cytochrome P-450 2B6 using a novel substrate, site-directed mutagenesis, and molecular modeling. J. Pharmacol. Exp. Ther. 290, 1141-1147. (17) He, Y. Q., Harlow, G. R., Szklarz, G. D., and Halpert, J. R. (1998) Structural determinants of progesterone hydroxylation by cytochrome P450 2B5: The role of nonsubstrate recognition site residues. Arch. Biochem. Biophys. 350, 333-339. (18) Ibeanu, G. C., Ghanayem, B. I., Linko, P., Li, L., Pederson, L. G., and Goldstein, J. A. (1996) Identification of residues 99, 220, and 221 of human cytochrome P450 2C19 as key determinants of omeprazole hydroxylase activity. J. Biol. Chem. 271, 1249612501. (19) Pikuleva, I., Puchkaev, A., and Bjorkhem, I. (2001) Putative helix F contributes to regioselectivity of hydroxylation in mitochondrial cytochrome P450 27A1. Biochemistry 40, 7621-7629. (20) Nakayama, K., Puchkaev, A., and Pikuleva, I. (2001) Membrane binding and substrate access merge in cytochrome P450 7A1, a key enzyme in the degradation of cholesterol. J. Biol. Chem. 276, 31459-31465. (21) Tsao, C. C., Wester, M. R., Ghanayem, B. I., Coulter, S. J., Chanas, B., Johnson, E. F., and Goldstein, J. A. (2001) Identification of human CYP2C19 residues that confer S-mephenytoin 4′-hydroxylation activity to CYP2C9. Biochemistry 40, 1937-1944. (22) Scott, E. E., Spatzenegger, M., and Halpert, J. R. (2001) A truncation of 2B subfamily cytochromes P450 yields increased expression levels, increased solubility, and decreased aggregation while retaining function. Arch. Biochem. Biophys. 395, 57-68. (23) Szklarz, G. D., He, Y. A., and Halpert, J. R. (1995) Site-directed mutagenesis as a tool for molecular modeling of cytochrome P450 2B1. Biochemistry 34, 14312-14322.

P450 2B1 Substrate Access Channel (24) Omura, T., and Sato, R. (1964) The carbon monoxide-binding pigment of liver microsomes. J. Biol. Chem. 239, 2379-2387. (25) Ciaccio, P. J., and Halpert, J. R. (1989) Characterization of a phenobarbital-inducible dog liver cytochrome P450 structurally related to rat and human enzymes of the P450IIIA (steroidinducible) gene subfamily. Arch. Biochem. Biophys. 271, 284299. (26) He, Y. Q., He, Y. A., and Halpert, J. R. (1995) Escherichia coli expression of site-directed mutants of cytochrome P450 2B1 from six substrate recognition sites: Substrate specificity and inhibitor selectivity studies. Chem. Res. Toxicol. 8, 574-579. (27) Lewis, D. F., Dickins, M., Eddershaw, P. J., Tarbit, M. H., and Goldfarb, P. S. (1999) Cytochrome P450 substrate specificities, substrate structural templates and enzyme active site geometries. Drug Metab. Dispos. 15, 1-49.

Chem. Res. Toxicol., Vol. 15, No. 11, 2002 1413 (28) Spatzenegger, M., Wang, Q., He, Y. Q., Wester, M. R., Johnson, E. F., and Halpert, J. R. (2001) Amino acid residues critical for differential inhibition of CYP2B4, CYP2B5, and CYP2B1 by phenylimidazoles. Mol. Pharmacol. 59, 475-485. (29) Graham-Lorence, S., and Peterson, J. A. (1996) P450s: structural similarities and functional differences. FASEB J. 10, 206-214. (30) Waller, C. L., Evans, M. V., and McKinney, J. D. (1996) Modeling the cytochrome P450-mediated metabolism of chlorinated volatile organic compounds. Drug Metab. Dispos. 2, 203-210. (31) Hasemann, C. A., Kurumbail, R. G., Boddupalli, S. S., Peterson, J. A., and Diesenhofer, J. (1995) Structure and function of cytochromes P450: a comparative analysis of three crystal structures. Structure 3, 41-62.

TX020057U