CYP105D1 has previously been expressed as an active recombinant cytosolic form in Escherichia coli using the IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible tac promoter (20). Selective permeability of E. coli to many substrates and products can cause problems when using whole-cell systems. For such reasons, cell wall mutants of Salmonella enterica serovar Typhimurium were developed for use in mutagenesis tests (10). One approach to overcome these problems could be to engineer CYPs that can be exported to the periplasm or the cell exterior. Previous studies with cytochrome b₅ have shown that cytosolic hemoproteins can be efficiently exported for their enhanced accumulation in the less hostile environments of the periplasm (see, for example, references 5 and 9). Based on these observations, we genetically manipulated an S. griseus CYP (CYP105D1) to attempt export to the periplasm.

Bacteria and plasmids. The E. coli DH5α strain was used for genetic manipulation, and the TB1 strain was subsequently employed for expression of the recombinant CYP105D1 cloned in the expression vector pLiQ. The vector pLiQ is a derivative of the previously described pAA-cyt (6), reengineered with appropriate restriction sites downstream of the alkaline phosphatase signal sequence. Heterologous expression was induced in E. coli grown in phosphate-limited (0.1 mM) MOPS (morpholinepropanesulfonic acid) medium containing trace elements and vitamins (18), 1 mM δ-aminolevulinic acid, and 100 μg of ampicillin per ml at 30°C for specified periods. Inocula consisted of a 10% (vol/vol) addition from saturated cultures grown on Luria-Bertani medium with ampicillin (100 μg/ml).

DNA manipulations. Standard procedures for molecular biology were performed as described by Sambrook et al. (15). The gene was amplified as a PCR fragment of 1,239 bp containing the engineered PstI and EcoRV sites at the 5′ and 3′ ends, respectively, using the forward primer 5′-AACTGCAGATGACGGAATCCACGACGGAC-3′ and the reverse primer 5′-ATGATATCTCACCAGGCCACGGGCAGGT-3′. The PCR conditions were as described previously (20). The resulting CYP105D1 DNA fragment was cloned into the pLiQ E. coli expression plasmid. The restriction and DNA-modifying enzymes were purchased from Promega (Southampton, United Kingdom) and used as recommended by the supplier.

E. coli subcellular fractionations. Bacteria (500 ml) were cultivated over time and harvested by centrifugation at 1,500 × g for 10 min. Periplasmic fractions were prepared by osmotic shock. Cells were plasmolyzed by suspension in 20 ml of 20% (wt/vol) sucrose–0.3 M Tris-HCl (pH 8)–1 mM EDTA (STE buffer) and incubation at 22°C for 10 min, harvested, and resuspended in residual STE buffer. Osmotic shock was performed by rapid immersion in 2 ml of ice-chilled 0.5 mM MgCl₂. After incubation on ice for 10 min, the periplasmic fraction was recovered by centrifugation at 10,000 × g for 10 min. The pellet was retained to provide the material for the preparation of cytoplasmic and membrane fractions as described previously (20).

Enzyme assays. CYP content was monitored by reduced carbon monoxide difference spectroscopy as described by Omura and Sato (13), using a Hitatchi U3010 scanning spectrophotometer. The protein content in bacterial fractions was estimated using the bicinchoninic acid (Sigma Chemicals) assay with bovine serum albumin as a standard. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed as described by Laemmli (8); approximately 75 μg of protein was loaded per lane. CYP105D1 enzyme activities were monitored in a 1-ml final reaction volume composed of 200 pmol of recombinant CYP105D1 contained in an appropriate volume of E. coli periplasmic fraction or cells. The reaction was initiated by addition of NADH (1 mM final concentration). For whole-cell studies, cells were harvested by centrifugation at 5,000 × g for 2 min and washed twice with 0.1 M potassium phosphate buffer (pH 7.4) containing 20% (vol/vol) glycerol prior to use. Erythromycin N-demethylation activity was determined as described previously (2). Benzo[a]pyrene 3-hydroxylation activity was measured fluorometrically using a Perkin-Elmer fluorescence spectrophotometer according to the method of Nebert and Gelboin (11).

Construction and expression of CYP105D1. We have developed an expression system for efficient targeting of CYP to the periplasmic space of E. coli. Previous studies for the successful expression of CYPs in E. coli have usually required modification at the 5′ region or installing a leader sequence (pelB or ompA) (1, 7, 14). Here CYP105D1 was fused with the alkaline phosphatase signal sequence and placed under the tight transcriptional control of the phoA promoter (18). Whole lysates derived from E. coli after 20 h under expression conditions displayed the defined spectral maximum of this CYP at 448 nm. The progressive increase in the Soret absorbance peak from cell extracts over time was accompanied by an intensifying red color of the bacteria and the periplasmic fractions.

CYP105D1 is targeted to the periplasm of E. coli A time course profile of the production of CYP105D1 in the isolated periplasmic fractions (Fig. 1) revealed that a detectable periplasmic buildup of CYP105D1 occurred at around 8 h and increased to a peak point at around 25 h, reaching a peak value exceeding 600 nmol per liter of culture. Optimally, >1,000 nmol of CYP105D1 has been obtained from this expression system, comparing well with previous studies on this and other CYPs. Previous studies reporting the heterologous expression of CYP105D1 under the control of the strongly inducible tac promoter obtained levels of CYP105D1 of just over 400 nmol of cytosolic CYP per liter of culture (20). The present results indicated that heme could be successfully incorporated into CYP during folding of matured protein translocated into the periplasmic space of E. coli. In order to further investigate the subcellular localization of the recombinant CYP, bacteria induced to express the protein were subfractionated into periplasmic, cytoplasmic, and membrane fractions. That effective subcellular fractionation had occurred was indicated by >90% enrichment of the marker enzyme activities associated with the isolated subcellular fraction, i.e., alkaline phosphatase (periplasm), malate dehydrogenase (membranes), and fumarase (cytosol). More than 80% of the total cellular CYP105D1 content was found localized in the periplasmic space of E. coli. Further substantiation of the cellular location of CYP forms was sought by subjecting the periplasmic protein fractions derived from E. coli TB1 expressing CYP105D1 to SDS-PAGE. The results (Fig. 2) show induction of CYP at the expected size of ~45,000 Da. A new protein at 45,406 Da was observed by electron spray analysis, indicating a processed product lacking a signal sequence. Additionally, the periplasmic fraction showed dramatic changes in the overall composition compared with control samples. Most significant is the intense cooverproduction of a 30-kDa protein whose identity was confirmed as β-lactamase through N-terminal protein sequence analysis of the protein band.

FIG. 1 Time course for production of CYP105D1 in E. coli periplasm.

FIG. 2 SDS-PAGE of periplasmic fractions of E. coli expressing CYP105D1. Each lane contains 75 μg of periplasmic protein stained with Coomassie brilliant blue R-250. Lane 1, molecular mass markers as shown on the left of the gel; lane 2, periplasmic (more ...)

Periplasmically targeted CYP105D1 is enzymatically active in vivo and in vitro The catalytic activities of recombinant CYP105D1 were measured using a whole-cell biotransformation procedure and with isolated periplasmic fractions expressing CYP105D1. No activity was seen from E. coli TB1 harboring the empty plasmid pLiQ. The ability of the isolated periplasmic fraction to sustain the xenobiotic transformation in E. coli (Table 1) indicates that this compartment contains suitable electron donor proteins for P450 in the periplasmic space of E. coli. Similar activities were observed for whole cells, indicating that the substrate had good access to the CYP biotransforming system. CYP105D1 requires a ferredoxin and a ferredoxin reductase for normal activity and can achieve substrate turnovers approaching unity (20). While it appears that the E. coli periplasmic redox partners are present in sufficient quantity to allow the CYP to have activity, smaller amounts of reductase system may be present or inefficient coupling may be occurring. However, here we do validate this approach as a suitable one for use in order to obtain heterologous CYP105D1 at a high concentration and as an active protein in an advantageous, accessible cellular location.

CYP105D1 activity	Turnovera (pmol of product formed/min/nmol of CYP) in:
	Periplasm of E. coli TB1 carrying:		Intact E. coli TB1 carrying:
	pLiQ (control)	pLiCYP105D1	pLiQ (control)	pLiCYP105D1
Erythromycin N-demethylation	NDb	67±24	ND	58±16
Benzo[a]pyrene 3-hydroxylation	ND	72±25	ND	56±18

We are grateful to the Biotechnology and Biological Science Research Council and the University of Wales, Aberystwyth, for partial support.