Co-expression of active human cytochrome P450 1A2 and cytochrome P450 reductase on the cell surface of Escherichia coli

Abstract

Background: Human cytochrome P450 (CYP) enzymes mediate the first step in the breakdown of most drugs and are strongly involved in drug-drug interactions, drug clearance and activation of prodrugs. Their biocatalytic behavior is a key parameter during drug development which requires preparative synthesis of CYP related drug metabolites. However, recombinant expression of CYP enzymes is a challenging bottleneck for drug metabolite biosynthesis. Therefore, we developed a novel approach by displaying human cytochrome P450 1A2 (CYP1A2) and cytochrome P450 reductase (CPR) on the surface of Escherichia coli.

Results: To present human CYP1A2 and CPR on the surface, we employed autodisplay. Both enzymes were displayed on the surface which was demonstrated by protease and antibody accessibility tests. CPR activity was first confirmed with the protein substrate cytochrome c. Cells co-expressing CYP1A2 and CPR were capable of catalyzing the conversion of the known CYP1A2 substrates 7-ethoxyresorufin, phenacetin and the artificial substrate luciferin-MultiCYP, which would not have been possible without interaction of both enzymes. Biocatalytic activity was strongly influenced by the composition of the growth medium. Addition of 5-aminolevulinic acid was necessary to obtain a fully active whole cell biocatalyst and was superior to the addition of heme.

Conclusion: We demonstrated that CYP1A2 and CPR can be co-expressed catalytically active on the cell surface of E. coli. It is a promising step towards pharmaceutical applications such as the synthesis of drug metabolites.

Fig. 1

Illustration of the biocatalysis by CYP1A2 and CPR on the cell surface. Two electrons are shuttled by the outer membrane (OM) anchored CPR from NADPH via the cofactors FAD and FMN in single-electron steps to the heme group of surface displayed CYP1A2. The monooxygenase catalyzes the oxidative breakdown of a substrate by inserting one oxygen atom into the chemical compound while reducing the other one to water. IM inner membrane

Fig. 2

Schematic depictions of the expression vectors for the CPR (a) and CYP1A2 (b) autotransporter fusion proteins. The expression cassettes consist of the rhamnose inducible promoter (RhaP), the CtxB signal peptide (SP), passenger (CPR: orange, CYP1A2: red), a connecting region (grey) and the C-terminal parts of the autotransporter AIDA-I or EhaA comprising the β₁-domain (β₁), α-helix (α) and β-barrel. The connecting region between the CPR and the EhaA is composed of TEV, factor Xa and OmpT protease cleavage sites, a PEYFK epitope and a 15 amino acid long flexible glycine/serine (G₄S)₃ part (pPQ29). For flow cytometry analysis the myc epitope (Myc) was inserted in between the passengers and autotransporter unit yielding pPQ61 (CPR) and pPQ62 (CYP1A2). The N-termini of the fusion proteins are expanded as white arrows to show the DNA- and amino acid sequences. Plasmids for CYP1A2 autotransporter fusion protein expression (b) contained the pBR322 ori and β-lactamase gene as selection marker (bla) and plasmids for CPR autotransporter fusion protein expression (a) the P15A ori and kanamycin resistance (KanR)

Fig. 3

Surface accessibility of CYP1A2 and CPR. a SDS-PAGE of outer membrane protein isolations and investigation of protease accessibility. M: Protein marker, apparent molecular weights are indicated on the left. Lane 1–2 sample from E. coli BL21(DE3) host cells, lane 3 sample from cells with induced expression of CPR fusion protein, lane 4 sample from cells with induced expression of CYP1A2 fusion protein, lane 5–6 samples from cells with induced co-expression of both fusion proteins. Samples in lane 2 and 6 were treated with proteinase K prior to outer membrane protein isolation. b Flow cytometry analyses of immunolabelled cells. Cell samples were treated with a primary monoclonal anti-myc antibody and a secondary Dylight488 conjugated anti-IgG antibody, washed and then analyzed via flow cytometry. Grey E. coli BL21(DE3) control cells, red cells expressing CYP1A2, orange cells expressing CPR

Fig. 4

Cytochrome c activity. Cytochrome c reduction by whole cells (OD_578nm 0.25) was measured at 550 nm. Reaction was started by addition of 100 µmol L⁻¹ NADPH. (1) Assay without cells, (2) E. coli BL21(DE3) cells, (3) cells expressing the CYP1A2, (4) cells expressing CPR

Fig. 5

Conversion of known CYP1A2 substrates. a HPLC chromatogram of samples from the 7-ethoxyresorufin-O-deethylation assay. Samples were diluted 1:1 with methanol prior to HPLC analysis. (1) Resorufin standard (10 nmol L⁻¹), (2) co-expression sample without substrate in the assay, (3) host cells, (4) cells displaying CPR, (5) cells displaying CYP1A2, (6) cells expressing CYP1A2 and CPR. b Resorufin formation in the whole assay determined by HPLC following expression at 23 or 30 °C. c HPLC chromatogram of samples of the phenacetin-O-deethylation assay. (1) Paracetamol standard (10 µmol L⁻¹) with 8.0 min retention time, (2) 3-acetamidophenol internal standard (10 µmol L⁻¹) with 10.5 min retention time, (3) control sample with co-expression cells without substrate, (4) host cells, (5) cells displaying CPR, (6) cells displaying CYP1A2, (7) cells displaying CYP1A2 and CPR. d MultiCYP activity measured by luciferin formation. Blank: Control sample without cells

Fig. 6

Effect of growth medium components on 7-ethoxyresorufin-O-deethylation. a Effect of 50 mmol L⁻¹ phosphate buffer in LB medium with 500 µmol L⁻¹ 5-ALA and 4 µmol L⁻¹ heme. NaPO₄: sodium phosphate buffer, Na/KPO₄: mixed sodium/potassium phosphate buffer (37 mmol L⁻¹ Na⁺/12 mmol L⁻¹ K⁺), KPO₄: potassium phosphate buffer. b Variation of NaCl content in 50 mmol L⁻¹ potassium phosphate buffered LB supplemented with 500 µmol L⁻¹ 5-ALA and 4 µmol L⁻¹ heme. c Effect of CaCl₂ and MgCl₂ (each 1 mmol L⁻¹) in potassium phosphate buffered LB medium supplemented with 500 µmol L⁻¹ 5-ALA and 4 µmol L⁻¹ heme. d Effect of heme and 5-ALA supplementation. Protein expression was conducted in potassium phosphate buffered LB medium with either 4 µmol L⁻¹ heme, 62.5 µmol L⁻¹ heme, 500 µmol L⁻¹ 5-ALA or 500 µmol L⁻¹ 5-ALA and 4 µmol L⁻¹ heme

Fig. 7

Time dependent characterization of 7-ethoxyresorufin-O-deethylation activity. The whole cell assay was performed with 1 µmol L⁻¹ 7-ethoxyresorufin and NADPH regeneration in a 100 mL shake flash with 8 mL starting volume. Concentration of the product resorufin was determined in the sample supernatants. Data points are the mean of at least two independent biological replica and the fluctuation for each data point was for all cases <15 %