Cytochrome P450 (CYP450) is a superfamily of membrane-bound, haem-containing monooxygenases that are responsible for the oxidative and reductive metabolism of many drugs, steroid hormones and fatty acids (Sakaki and Inouye, 2000). They are expressed in many animal tissues, although found in relatively high quantities in the liver. The principal function of CYP450 is to introduce an oxygen atom into the substrate, to increase the hydrophilicity of the product and hence the ease with which the product is eliminated from the animal body (Crespi and Miller, 1999). The catalytic cycle of microsomal CYP450s requires electron transfer from NADPH cytochrome P450 reductase (CPR), while the presence of cytochrome b5 stimulates the activities of certain CYP450 towards some substrates (Yamazaki et al., 1996; Backes and Kelley, 2003).
Different members of CYP450 play an important role in the biotransformation of exogenous compounds including therapeutic drugs, environmental pollutants and procarcinogens (Crespi and Miller, 1999; Sakaki and Inouye, 2000). It is generally believed that CYP450s are of importance in the mechanisms leading to cancer, and have as a result, attracted considerable attention in the pharmaceutical industries (Kan et al., 2002). Human populations exhibit considerable variabilities in CYP450-dependent activities because most CYP450 species are polymorphic, leading to different susceptibilities to carcinogenic compounds and variation in drug metabolism (Donato and Castell, 2003). CYP1A1 polymorphism in particular has been implicated in the development of lung cancer (Chen et al., 2001; Lee et al, 2001; Wang et al., 2002). Most mammalian CYP450s contain extremely homologous isozymes, such that the isolation and identification of CYP450 species by traditional protein purification procedures is almost impossible. The expression of human CYP450s from their cDNAs in heterologous systems has greatly reduced the need for human tissue and significantly advanced the knowledge of substrate specificities and catalytic capabilities of human P450 isoforms.
While bacterial expression of these enzymes is relatively easy and inexpensive, bacteria lack the endoplasmic reticulum to which mammalian microsomal CYP450s bind, the requisite electron transfer partner for microsomal CYP450 activity. The CYP450 activity of the protein expressed in Escherichia coli very often requires reconstitution, which is achieved by isolation and solubilization of host E. coli membranes, followed by the addition of CPR and removal of detergent by dialysis before activity can be measured (Barnes et al., 1991). This has necessitated the development of CYP450–CPR bicistronic or biplasmid E. coli expression systems (Dong and Porter, 1996; Parikh et al., 1997; Kranendonk et al., 1999). The other drawbacks of CYP450 expression in E. coli are that the expression levels are quite different among CYP450 species, and modification of 20–30 N-terminal amino acid residues is needed for the high-level expression of mammalian microsomal CYP450 proteins (Barnes et al., 1991; Iwata et al., 1998; Sakaki and Inouye, 2000). The high level expression of CYP450–CPR often results in the loss of bacterial cell viability (Kranendonk et al., 1999) and even reduced protein catalytic activities (Crespi and Miller, 1999).
Several mammalian cells have been used as hosts for the expression of CYP450 cDNAs (Crespi and Miller, 1999). Although a high expression level of CYP450 cannot be expected in a stable expression system, mammalian cells have the machinery for subcellular localization of CYP450 and electron transport chains. In addition, they can carry out further metabolism of the CYP450 products and appear very useful in the evaluation of cytotoxicity and mutagenicity tests of drugs (Sakaki and Inouye, 2000). Baculovirus systems can be used to co-express human CYP450 and CPR genes (Schwarz et al., 2001; Barnes et al., 1994), except for the noted deficiency in haem incorporation (Lee et al., 1995) and technical problems associated with the optimization of conditions for cell culture and appropriate times for microsome harvesting (Crespi and Miller, 1999). While the expression of human CYP450 in Schizosaccharomyces pombe (Ehmer et al., 2002) and Pichia pastoris (Andersen and Moller, 2002) has been reported, Saccharomyces cerevisiae proved to be the yeast most often used for the heterologous expression of human CYP450 proteins (Murakami et al., 1990; Sakaki et al., 1990; Shibata et al., 1990; Ching et al., 1991). The optimal production of CYP450 activity in S. cerevisiae has been achieved by the construction of hybrid cDNA encoding fused proteins of mammalian CYP450 and yeast CPR (Sakaki et al., 1990, 1996; Shibata et al., 1990).
Müller et al. (1998) compared several yeasts, including S. cerevisiae and Y. lipolytica, for their efficiency as hosts in heterologous protein production. Y. lipolytica was revealed as the most suitable host for heterologous production of proteins. Moreover, inducible vectors for expression and gene amplification by multiple integrations into the Y. lipolytica genome have been constructed (Juretzek et al., 2000, 2001; Nicaud et al., 2002). We report in this study the expression of human CYP1A1 in Y. lipolytica in the presence or absence of additional copies of yeast NADPH CPR gene under the control of Y. lipolytica acyl-CoA oxidase 2 (POX2) and isocitrate lyase (ICL) promoters.
Materials and methods
Strains, plasmids and culture conditions
E. coli DH5α strain cells (GIBCO/BRL, Life Technologies) were used as host for plasmid DNA propagation. E. coli cells carrying plasmids were grown in Luria–Bertani (LB) medium at 37°C. Standard microbial and recombinant techniques are as described by Sambrook et al. (1989). Standard media were supplemented with kanamycin (40 µg/ml) or ampicillin (100 µg/ml). Plasmid DNA was isolated from bacteria with the Qiagen or Promega Plasmid Purification kits. The DNA fragments for subcloning were recovered and purified from agarose gel using either the QIAEXII gel extraction kit or Promega Wizard PCR Preps kit.
The Y. lipolytica strains used in this study were PO1d derivatives (MATaura3-302, leu2-270, xpr2-322). The culture media and techniques used to grow and handle Y. lipolytica were described by Barth and Gaillardin (1996). The yeast media YPD, YNBD and YNBcas have been described (Wang et al., 1999). The Y. lipolytica strains were transformed using the lithium acetate method (Barth and Gaillardin, 1996) and the transformants were selected on YNBD (Leu+) or YNBcas (Ura+). The medium YPDH, for the induction of the Y. lipolytica POX2 and ICL promoters, contained (per litre) yeast extract (10 g), bactotryptone (20 g), glucose (10 g), phosphate buffer, pH 6.8 (50 mM) and olive oil (2.5% v/v). Cell growth and density for the Y. lipolytica strains was monitored by measuring light scattering at 600 nm. The solid media for culturing yeast or bacterial cells contained 15 g/l agar.
Construction of NADPH CPR expression vectors
Plasmid JMP21–pICL–CPR was constructed by digestion of plasmid p67RYL (Mauersberger et al., unpublished), which contains Y. lipolytica cytochrome P450 reductase CPR gene expressed under the ICL promoter (pICL–CPR), with MluI and KpnI to release the pICL–CPR fragment. This fragment was ligated into JMP21 (Nicaud et al., unpublished) at the corresponding sites (Figure 1).
Plasmid JMP61-pPOX2-CPR was constructed in two steps as described in Figure 2. First, the CPR gene was amplified by PCR, using p67RYL as template, with primers YLCPR-F (5′-GGGCCCAAGCTTATGCCTCTACTCGACTCTCTCGACTTTATT-3′) and YLCPR-R (5′-GTTCATCCTAGGTTACTACCACACATCTTCCTGGTAGACGTTCTG-3′) as forward and reverse primers, respectively. The underlined bold sequences represent the HindIII and AvrII restriction enzymes recognition sequences. The PCR was carried out using Ex Taq polymerase (Takara Biomedicals) and cloned Pfu (Stratagene) under the following reaction conditions: 1 denaturation cycle of 2 min, 25 cycles of amplification (95°C, 30 s, 60°C, 1 min; 72°C, 2 min) followed by a final extension at 72°C for 10 min. The amplified fragment was digested with HindIII and AvrII, followed by ligation into the JMP61 vector digested with similar enzymes, giving rise to plasmid JMP61–pPOX2–CPR. In the second step, the 3.3 kb ClaI–EcoRI fragment carrying pPOX2–CPR was isolated from JMP61–pPOX2–CPR and ligated at the corresponding sites of JMP21, giving rise to JMP21–pPOX2–CPR (Figure 2).
Construction of CYP1A1 expression vectors
The 1.5 kb BamHI–KpnI fragment carrying the human CYP1A1 gene was rescued from the YeDP 1A1 plasmid (Gautier et al., 1993) by digestion with BamHI and KpnI. The human CYP1A1 was cloned into the expression vectors JMP62 and JMP64, which contain, respectively, the non-defective marker ura3d1 for single integration and the defective marker ura3d4 for integration in multiple copies into the genome of Y. lipolytica (Nicaud et al., 2002). To clone CYP1A1 into expression vectors, plasmids JMP62 and JMP64 were separately digested with ClaI and KpnI to release a large backbone fragment of about 4420 bp, which contains the Y. lipolytica selection marker, the zeta elements for genome integration (Mauersberger et al., 2001) and the sequences required for plasmid replication and selection in the E. coli host. The 1017 bp fragment containing the POX2 promoter was obtained by digestion of JMP62 plasmid with ClaI and BamHI. The appropriate backbone DNA, POX2 promoter and CYP1A1-containing fragments were ligated together to create JMP62–CYP1A1 and JMP64–CYP1A1 expression vectors (Figure 3).
Transformation of E. coli and Y. lipolytica strains
The ligation mixtures were used to transform CaCl2-competent E. coli cells. The recombinant plasmids were identified by the isolation of the plasmid from the E. coli host, followed by digestion with appropriate restriction enzymes. Vectors JMP62–CYP1A1 and JMP64–CYP1A1 were digested with Not1 and used to transform Y. lipolytica Po1d strain. Y. lipolytica transformants containing single integrants of the JMP62–CYP1A1 vector (strains JMY331–JMY334) or multiple integrants of the JMP64–CYP1A1 vector (strains JMY334–JMY347) were obtained upon selection for Ura+ on YNBCas medium (Table 1). The JMP5 vector (Mauersberger et al., 2001) was used to transform Y. lipolytica Po1d strain to Ura+ phenotype and served as a negative control for CYP P450 1A1 activity (strain JMY330). The transformed strains were grown on oleic acid induction medium and assayed for CYP1A1 activity. The monocopy transformant JMY331 and the multicopy transformant JMY339 presenting the highest CYP1A1 activity were selected for further studies, together with the JMY330 strain representing CYP1A1 negative control strain (Table 1).
Table 1. Comparative analysis of the cytochrome P450 CYP1A1 activity co-expressed with the CPR under the ICL or POX2 promoters
CYP activity (pM/min/dw)
Ampl. fold (average)
The Y. lipolytica transformed strains were harvested after 48 h of growth in 50 ml induction medium and the activity determined on whole cells using the ethoxyresorufin bioconversion method. A, CYP activity; B, average CYP activity; 1, CYP activity amplification factor compared to the multicopies CYP 1A1 parental strain JMY339; 2, CYP activity amplification factor compared to the monocopy CYP 1A1 parental strain JMY331. No. indicates the number of the strain presented in Figure 5.
The JMP21–pICL–CPR and JMP21–pPOX2–CPR expression vectors were digested to completion with Not1. The digested JMP21–pICL–CPR plasmid was used to transform Y. lipolytica strains containing zero copy (strain JMY330), monocopy (strain JMY331) or multiple copies (strain JMY339) of human CYP1A1 (Table 1). The transformed strains were selected for Leu+ phenotype on YNBD medium. Independent colonies were selected from each transformed strain to represent strains harbouring the CPR gene within the genomes of JMY330 (strains JMY859 and JMY860), JMY331 (strains JMY856–JMY858) and JMY339 (strains JMY846–JMY849). The independent strains obtained from the integrative transformation of JMY330, JMY331 and JMY339 with the JMP21–pPOX2–CPR plasmid were referred to as strains JMY1063–JMY1065 (JMY330), JMY1059–JMY1061 (JMY331) and JMY1055–JMY1057 (JMY339) (Table 1). The strain for the CYP1A1-negative control (JMY330) was also transformed with the empty JMP21 (strain JMY854).
CYP450 1A1 and CPR NADPH CPR induction
Yeast transformants were grown overnight in YPD medium. Exponentially growing cells were collected by centrifugation, washed twice in distilled water and resuspended in the induction medium. The cells were used to inoculate 50 ml induction medium YPDH to a cell density of 0.5 at OD600. The cultures were incubated at 28°C at 250 rpm in baffled flasks and samples were collected at different times of growth for CYP1A1 assays.
CYP450 CYP1A1 activity measurements
The sample culture (1ml) was centrifuged and re-suspended in 1 ml 50 mM Tris–HCl, pH 7.5, containing 1 mM EDTA. An aliquot of the whole-cell suspension was added into 1 ml resuspension buffer containing 2 µl ethoxyresorufin (EOR), prepared by dissolving to saturation in methanol. Cytochrome P450 CYP1A1 activity (cm/min/OD) of the cell suspension was measured with the fluorescence assay at the excitation and emission wavelengths of 530 and 580 nm, respectively (Pompon et al., 1996). The activity was converted to pM/min/dw by extrapolation with hydroxyresorufin, the product of cytochrome P450 CYP1A1 activity on ethoxyresorufin, as the standard.
Results and discussion
The overproduction of CYP450 proteins has both biotechnological and therapeutic significance. The heterologous expression of active CYP450 protein complexes is an essential step towards investigating their role in xenobiotic and therapeutic drug degradation pathways and their utilization in biocatalysis and synthesis of hydroxylated compounds. For these purposes the expression of CYP450 with efficient electron transport systems is important. Studies have demonstrated that the ratio CYP450:CPR is important, although this ratio is dependent on the type of CYP450 being investigated (Truan et al., 1993).
We wanted to develop an expression system that would be suitable for the expression of different species of CYP450 proteins. We took advantage of the recently developed genetic tools for heterologous expression of proteins in the yeast Y. lipolytica. The POX2 and ICL promoters are induced in media that contain hydrophobic substrates, such as oleic acid and olive oil (Juretzek et al., 2000). In order to express CYP1A1 in Y. lipolytica, the coding region was cloned under the POX2 promoter in the new vector JMP62 for single-copy integration and into JMP64 for multiple-copy integrations (Figure 3). The Y. lipolytica strain PO1d, transformed with the plasmid JMP62–CYP1A1, exhibited CYP1A1 activity upon culture on a medium containing olive oil (YPDH). CYP1A1 activity started to accumulate after about 25 h of culture (Figure 4), the time which coincides with the exhaustion of glucose in the media (data not shown). It has been shown that POX2 promoter remains repressed in the presence of glucose in the culture medium (Juretzek et al., 2000). Maximum production of CYP450 was achieved at about 70 h of culture, based on CYP1A1 activity on whole yeast cells using ethoxyresorufin as the substrate (Figure 4).
As shown in Table 1 and Figure 5, no activity could be detected in JMY330 strains transformed with the empty vector JMP21 (JMY854), even in the presence of the CPR gene under the ICL promoter (JMY852 and JMY853) or the POX2 promoter (JMY1063 and JMY1064). The JMY331 strains (JMY859 and JMY860), which contain a single copy of CYP1A1 under the POX2 promoter, presented the average activity of about 32.17 pM/min/dw. In contrast, strains transformed with the defective vector (JMY850 and JMY851) showed a four-fold increase in activity as compared to the monocopy integrants (128.7 pM/min/dw). This indicated that the CYP1A1 expression cassette is integrated in multiple copies within the Y. lipolytica genome.
Since it has been shown that the CYP450:CPR ratio is important for optimal CYP1A1 activity, and that Y. lipolytica CPR may be a limiting factor (Truan et al., 1993), we co-expressed the Y. lipolytica CPR under two strong promoters, the promoter of the isocitrate lyase gene (pICL) and that of the acyl-CoA oxidase POX2 gene (pPOX2), which are both induced by oleic acid (Juretzek et al., 2000). The average CYP1A1 activity of the JMY331 monocopy strains containing the CPR under the ICL promoter is 48.6 pM/min/dw, while an average activity of 65.8 pM/min/dw was obtained with the CPR gene under the POX2 promoter (Table 1, Figure 5). There was no marked difference in CYP1A1 activity when the CPR was expressed under the ICL or the POX2 promoter. The co-transformation of CPR into CYP1A1 monocopy integrants resulted in a 1.5-fold increase in activity with pICL–CPR and a two-fold increase with pPOX2–CPR. The results indicated that in monocopy CYP1A1 integrants, although CPR is required for optimal activity, it is not a strong limiting factor for CYP1A1 activity.
Multiple copies CYP1A1 integrants showed only a four-fold increase in CYP1A1 activity (average 129 pM/min/dw) in the absence of the CPR gene. The introduction of the CPR gene into the multiple copy integrant JMY 339 with the CPR gene under the ICL promoter showed CYP1A1 activity in the range 815–1845 pM/min/dw, while the same strain with the CPR under the POX2 promoter showed activity in the range 803–1645 pM/min/dw. This marked increase in CYP450 activity upon the introduction of the CPR gene could be explained by the sudden availability of sufficient quantities of the CPR protein for optimal transfer of electrons to an increased number of CYP1A1 proteins. Moreover, the multicopy integrant strains transformed with the CPR gene showed a huge range in CYP1A1 activity, indicating that CPR level was limiting in the multicopy transformant (Figure 5). When we compared the average activity for JMY848 (815.1 pM/min/dw) with that of JMY1055 and JMY1056 (829.4 pM/min/dw) we observed 6.3- and 6.4-fold increases of activity, respectively. This is a higher magnitude compared to the two-fold CYP1A1 activity obtained with the monocopy strains. In addition, in some strains, two copies of the CPR gene were integrated, as for JMY846 and JMY847, which presented a 12.3-fold increase in CYP1A1 activity and JMY1057 which presented a 12.8-fold increase in activity as compared to the multicopy integrant JMY339 (Table 1, Figure 5). No significant differences could be observed for CYP1A1 activity when CPR was expressed under the ICL or the POX2 promoters, although these results indicate that CPR is limiting in monocopy integrants (two-fold increased) and is absolutely required in multicopy integrants (six- and 12-fold increased). Data obtained with our Y. lipolytica system were compared with those determined with CYP1A1 expressed in S. cerevisiae. We obtained 11.2 pM/min/dw/107 cells in Y. lipolytica for intact cells presenting CYP1A1 in monocopy. In comparison, we previously reported 2 pM/min/dw/107 cells in S. cerevisiae strain W303 expressing CYP1A1 alone (Urban et al., 1990). Co-expression in Y. lipolytica, of Y. lipolytica CPR and CYP1A1 in multicopies, results in a 50-fold increase to 575 pM/min/dw/107 cells. Co-expression of Y. lipolytica reductase results in a 6.4-fold increase of CYP1A1 activity for the multicopy transformants, which is similar to the increase observed by over expression of S. cerevisiae reductase in W303. Indeed, Urban et al. (1993) observed a 5.5-fold increase in W(R) strain (11 pM/min/dw/107 cells). Therefore, Y. lipolytica presents about a 50-fold higher CYP1A1 activity of intact cells compared to the highest activity for intact cells in S. cerevisiae.
In this study we have demonstrated the efficiency of the Y. lipolytica yeast in the expression of human CYP450 protein. We have also demonstrated the in vivo efficiency of the Y. lipolytica CPR in shuttling electrons to the human CYP1A1 protein. The human CYP1A1 activity increased nearly 50-fold by co-induction of the CPR gene in the strain containing multiple integrants of the CYP1A1. Unlike other studies, in which bicistronic or biplasmid constructs of CYP450 and CPR were developed, we were able to integrate the respective genes into the yeast genome independently of one another.
Studies have shown that CYP450s form a 1 : 1 functional CPR : CYP450 complex (Miwa and Lu, 1984). There are, however, multiple forms of CYP450s in the endoplasmic reticulum and the ratio of CYP450:CPR has been suggested to lie between 10 : 1 and 20 : 1 (Backes and Kelley, 2003), implying that the CPR is capable of supplying electrons to each of many different CYP450 enzymes. This expression system could therefore serve as a major tool in studies aimed at understanding the effects of alteration of the CYP450:CPR ratio on CYP450 reduction and activity. The JMY339 strains containing multiple integrations of the CYP1A1 could not produce optimal activity in the presence of endogenous activities of the yeast CPR activity alone. Amplification of the CPR activity, however, resulted in up to 12-fold increase in strains containing multiple copies of CYP1A1 genes within their genomes. The expression system is therefore a potential tool for in vivo studies of the behaviour of CYP450 proteins at limiting levels of CPR activity. These studies are essential in understanding how the CYP450 species with low affinity for the association with the CPR protein have evolved to remain metabolically functional (Backes and Kelley, 2003).
This expression system has potential for the expression of other CYP450 proteins that require different levels of CPR activity. Although the Y. lipolytica yeast contains CYP450 of its own, no activity was observed in the absence of the CYP1A1 gene. The expression system did not require the isolation and solubilization of host membranous systems for CYP450 activity reconstitution. This has biotechnological significance where substrates could be added in the growing yeast cultures expressing the CYP450 proteins, and the bio-converted products recovered from the culture medium, facilitating the downstream processing of the required products.
The authors would like to thank Stephan Mauersberger for the kind gift of the p67RYL plasmid. This work was supported by the Institut National de la Recherche Agronomique, and by the Centre National de la Recherche Scientifique. M.B.N. was the recipient of a grant from the Ministère de la Recherche.