Directed evolution reveals requisite sequence elements in the functional expression of P450 2F1 in Escherichia coli

Abstract

Cytochrome P450 2F1 (P450 2F1) is expressed exclusively in the human respiratory tract and is implicated in 3-methylindole (3MI)-induced pneumotoxicity via dehydrogenation of 3MI to a reactive electrophilic intermediate, 3-methyleneindolenine (3-MEI). Studies of P450 2F1 to date have been limited by the failure to express this enzyme in Escherichia coli. By contrast, P450 2F3, a caprine homologue that shares 84% sequence identity with P450 2F1 (86 amino acid differences), has been expressed in E. coli at yields greater than 250 nmol/L culture. We hypothesized that a limited number of sequence differences between P450s 2F1 and 2F3 could limit P450 2F1 expression in E. coli and that problematic P450 2F1 sequence elements could be identified by directed evolution. A library of P450 2F1/2F3 mutants was created by DNA family shuffling and screened for expression in E. coli. Three generations of DNA shuffling revealed a mutant (named JH_2F_F3_1_007) with 96.5% nucleotide sequence identity to P450 2F1 and which expressed 119 ± 40 pmol (n = 3, mean ± SD) hemoprotein in 1 mL microaerobic cultures. Across all three generations, two regions were observed where P450 2F3-derived sequence was consistently substituted for P450 2F1 sequence in expressing mutants, encoding nine amino acid differences between P450s 2F1 and 2F3: nucleotides 191-278 (amino acids 65-92) and 794-924 (amino acids 265-305). Chimeras constructed to specifically test the importance of these two regions confirmed that P450 2F3 sequence is essential in both regions for expression in E. coli but that other non-P450 2F1 sequence elements outside of these regions also improved the expression of mutant JH_2F_F3_1_007. Mutant JH_2F_F3_1_007 catalyzed the dehydrogenation of 3MI to 3-MEI as indicated by the observation of glutathione adducts after incubation in the presence of glutathione. The JH_2F_F3_1_007 protein differs from P450 2F1 at only 20 amino acids and should facilitate further studies of the structure-activity relationships of P450s of the 2F subfamily.

Figure 1. Expression of mutants from the first generation P450 2F1/2F3 shuffled library

The first generation shuffled library (JH_2F_F1_1) was screened for P450 hemoprotein (pmol) produced in 1 mL, 72 h microaerobic cultures as determined by Fe(II)]CO vs. Fe(II) difference spectroscopy. Inset: mutants that expressed greater than 50 pmol P450 per 1 mL culture. Data from the controls P450 2F1, P450 2F3, and hNPR are included in both graphs for reference, and are shown on the main plot in the same order as in the inset. Data from mutants were from single cultures but data from controls represents the mean ± SD of n = 4 independent cultures. The dashed line on the inset plot indicates the sensitivity limit of 50 pmol for detecting P450 hemoprotein.

Figure 2. Recombination patterns of shuffled P450 2F mutants

Sequences of mutants from A) first(generation library JH_2F_F1_1, B) second(generation library JH_2F_F2_1, and C) third(generation library JH_2F_F3_1, that were expressed successfully in E. coli as detectable hemoprotein were analyzed by alignment to the P450 2F1 and P450 2F3 parental sequences. Software described previously was used to identify which sequence elements were derived from each of the relevant parental sequences. P450 2F1- and P450 2F3-derived sequence is represented by blue and green dots respectively. Where no dot occurs, both parental sequences encode the same base. PCR(derived point mutations are indicated by a cross (+). Boxes indicate regions where P450 2F1 sequence does not occur in any successfully(expressed mutant.

Figure 3. Expression of mutants from the second generation P450 2F1/2F3 shuffled library

The second generation shuffled library (JH_2F_F2_1) was screened for P450 hemoprotein (pmol) produced in 1 mL, 72 h microaerobic cultures as determined by Fe(II)]CO vs. Fe(II) difference spectroscopy. Inset: the twenty mutants with greatest P450 expression. Data from the forms used as parents for the F2 library, P450 2F1, JH_2F_F1_1_064 (abbreviated as F1_1_064 on this plot), JH_2F_F1_1_207 (F1_1_207), and JH_2F_F1_1_217 (F1_1_217), plus controls P450 2F3 and hNPR are included in both graphs for reference (shown in the main plot in the same order as in the inset). F2 mutants were expressed as single cultures. However, results for controls and F1 mutants used as parents for the F2 library are shown as the mean ± SD of n = 4 independent cultures. The dashed line on the inset plot indicates the sensitivity limit of 50 pmol for detecting folded P450.

Figure 4. Alignment of P450 2F regions of interest with other P450 2 family P450s that have been expressed in E. coli

The amino acid sequences for P450s 2F1, 2F3, and shuffled mutant JH_2F_F3_1_007 were aligned with the sequences of all P450 2 family P450s for which crystal structures had been published to date. The two regions of interest (amino acid residues 65-92 and 266-301) are marked by boxes. Amino acid positions that differ between P450s 2F1 and 2F3 within the regions of interest are marked in red with bold type.

Figure 5. P450 2F1/2F3 chimeras constructed to test the importance of specific sequence regions

P450 2F1- and P450 2F3-derived sequence is represented in blue and green respectively. Chimeras 2FcAexp, 2FcA, and 2FcC contain P450 2F3 sequence between nucleotides 191 and 285. Chimeras 2FcAexp and 2FcBexp contain P450 2F3 sequence between nucleotides 794 and 940. Chimeras 2FcA and 2FcB contain P450 2F3 sequence between nucleotides 848 and 940. The remaining sequence in all chimeras is from P450 2F1.

Figure 6. SDS PAGE of subcellular fractions from E. coli expressing P450 2F1, P450 2F3, and shuffled mutant JH_2F_F3_1_007

Subcellular fractions were electrophoresed on a 7.5 % polyacrylamide gel and protein was visualized by silver staining. First panel: Insoluble fractions from cultures expressing empty pCW vector (lane 1), pCW/2F1(nat)/hNPR (lane 2), pCW/2F3/hNPR (lane 3), and pCW/JH_2F_F3_1_007/hNPR (lane 4). Second panel: Membrane fractions from cultures expressing the empty pCW vector (lane 5), pCW/2F1(nat)/hNPR (lane 6), pCW/2F3/hNPR (lane 7), pCW/JH_2F_F3_007/hNPR (lane 8), and pCW/hNPR. Third panel: Cytosolic fractions from cultures expressing empty pCW vector (lane 9), pCW/2F1(nat)/hNPR (lane 10), pCW/2F3/hNPR (lane 11), and pCW/JH_2F_F3_1_007/hNPR (lane 12). Protein bands corresponding to P450 from the P450 2F subfamily were identified in lanes 3, 4, 7, 8, 11, 12 and 13, by comparison to negative controls and their position is indicated by the arrowheads. The cell pellet and membrane samples shown were derived from 4 Zg wet mass of harvested cells, while cytosolic samples were derived from 2 Zg wet mass of cells.

Figure 7. Homology modelling of two P450 2F3 sequence regions associated with successful expression of shuffled mutants

A) A model structure of shuffled mutant JH_2F_F3_1_007 was created using the Robetta Full Chain Protein Prediction server. Two regions where P450 2F3 sequence occurred in all expressing shuffled mutants (region 1: amino acids 65-92, and region 2: amino acids 266-305) are highlighted in green, with amino acids that differ between P450 2F1 and P450 2F3 within these regions marked with orange spheres. B) The structure shown in panel A rotated 90° through the x(axis.

Figure 8. Metabolism of 3MI by shuffled mutant JH_2F_F3_1_007

A) Metabolites detected by LC/MS following incubation of mutant JH_2F_F3_1_007 in E. coli membrane preparations with 3MI and glutathione. Products were identified by comparison to authentic standards. B) Mass spectra of identified products.