The N-Terminal Membrane Domain of Yeast NADPH-Cytochrome P450 (CYP) Oxidoreductase Is Not Required for Catalytic Activity in Sterol Biosynthesis or in Reconstitution of CYP Activity*

The disruption of Saccharomyces cerevisiae NADPH-cytochrome P450 oxidoreductase ( CPR ) gene resulted in a viable strain accumulating approximately 25% of the ergosterol observed in a sterol wild-type parent. The associated phenotypes could be reversed in transformants after expression of native CPR and a mutant lack-ing the N-terminal 33 amino acids, which localized in the cytosol. This indicated availability of the CPR in each case to function with the monooxygenases squalene epoxidase, CYP51, and CYP61 in the ergosterol biosynthesis pathway. Purification of the cytosolic mutant CPR indicated properties identical to native CPR and an ability to reconstitute ergosterol biosynthesis when added to a cell-free system, as well as to allow reconstitution of activity with purified CYP61, sterol 22-desatu-rase. This was also observed for purified Candida albicans and human CYP51 in reconstituted systems. The ability of the yeast enzyme to function in a soluble form differed from human CPR, which is shown to be inactive in reconstituting CYP activity. strain, the primers and conditions mentioned above (the cells were heated in a microwave for 30 s and quickly cooled on ice for lysis before being subjected to PCR), and ketoconazole susceptibility tests. Sterol Isolation and Analysis— Yeast cells harvested from 100 ml of culture were resuspended in 3 ml of methanol, 2 ml of 60% (w/v) KOH in water, and 2 ml of 0.5% (w/v) pyrogallol dissolved in methanol and saponified by heating at 90 °C for 1 h. Nonsaponifiable lipids (sterols) were extracted from the saponified mixture three times with 5 ml of hexane each time, pooled, and dried under nitrogen. The sterols were suspended in 100 m l of toluene and heated at 60 °C for one hour for silylation after adding 20 m l of bis(trimethylsilyl) trifluoride. The silyl sterols were analyzed by gas chromatography/mass spectrometry (VG 12–250; VG BIOTECH) by using split injections with a split ratio of 20:1. Sterol identification was by reference to relative retention times and mass spectra reported previously (19, 20). Immunoblot Analysis— SDS-10% polyacrylamide gel electrophoresis, nitrocellulose filter transfer, and immunodetection of protein were per- formed as described previously (21, 22). Yeast CPR antiserum was kindly provided by J. C. Loper of Cincinnati University. Anti-CPR IgG was purified from the CPR antiserum as described (23). ratio was calculated from the areas of the two peaks, and the turnover (nmol of product formed/min/nmol of P450) was obtained from the amount of substrate added and the conversion ratio. Spectrophotometric Measurements— A Philips PU8800 UV/VIS scanning spectrophotometer was used for all spectral studies. Cytochrome c reductase activity was measured as described previously (29). The absorption spectra of the purified yeast CPR ( D 33) was obtained by scanning from 300 to 700 nm (4). The ergosterol content in the nonsaponifiable lipids was calculated on basis of its molar extinction coeffi- cient at 282 nm ( E m 5 11,900).

NADPH-cytochrome P450 oxidoreductase (CPR) 1 (EC 1.6.2.4) is required for microsomal eukaryotic cytochrome P450 (CYP) monooxygenase activity, transferring either both electrons or (sometimes) the first electron for these reactions (1). The CYP enzymes are involved in the metabolism of foreign compounds, such as lipophilic pollutants, pesticides, and drugs, as well as in many biosynthetic reactions (for instance, in steroid, alkaloid, and terpenoid biosynthesis). Although in plants there is CPR diversity, in animal and fungal systems only one CPR has been identified; it functions with the many members of the microsomal P450 superfamily in a particular organism (2). The soluble domain of the enzyme can be generated for structural studies via x-ray crystallography (3,4), but it was previously reported to be unable to support mammalian CYP activity where the N-terminal membrane anchor was required (5). We are interested in the yeast CPR, and we developed and characterized systems to examine the function of the soluble domain of the enzyme with comparison to human CPR.
We show here, using whole transformants and cell-free systems, that the soluble yeast enzyme can support CYP activity, in contrast to human CPR, which is shown to be inactive.
Within the CYP superfamily, CYP51 is the only family found in animal, plant, and fungal kingdoms, and it represents an ancient metabolic role for CYP in sterol biosynthesis, undertaking C14 demethylation via three sequential hydroxylations (6). The fungal enzyme is the target for azole antifungals, which are selective in their inhibition over the human and plant orthologues, are central to antifungal chemotherapy, and represent about one-third of the agricultural fungicides used. Two other enzymes of fungal ergosterol biosynthesis require CPR, CYP61 (a sterol 22-desaturase (7-9)), and a non-CYP monooxygenase, squalene epoxidase (10). The former may also be present in plants, in which 22-desaturation is observed, unlike in animals, but the latter is present in all organisms producing sterols (11).
As expected for an antifungal target, gene disruption of CYP51 was observed to be lethal, but the strain could be rescued by providing an ergosterol supplement that could be taken up only anaerobically (12). In contrast, gene disruption of yeast CPR produced viable mutants (13) and ergosterol was present. 2 No additional CPR genes could be detected or were subsequently revealed within the yeast genome. In reconstituted assays with purified CYP, enzyme cytochrome b 5 can act as alternative donor for the second electron required for monooxygenase activity (14) and may have been supporting catalytic activity in the disruptant. Supporting this concept was the observation that the gene encoding cytochrome b 5 can act as a suppressor of a CPR gene disruption phenotype, noted by Sutter and Loper (13), namely hypersensitivity to the CYP51 inhibitor ketoconazole (15). The hypersensitivity suggested that a large reduction in ergosterol biosynthesis might have occurred, making the strain more sensitive, with reduction in ergosterol levels below the critical level occurring at a lower concentration of drug.
Here, we present the first full biochemical characterization of a yeast strain disrupted in the gene encoding CPR, including measurement of activity, enzyme localization, and qualitative and quantitative investigation of sterol profiles. It is shown that an absence of yeast CPR does not result in a dramatic reduction in ergosterol synthesis in the cell. Furthermore, the ability of an N-terminal truncated soluble yeast enzyme to complement a strain containing no CPR and reconstitute CYP activity is demonstrated. This is in direct contrast to genetically engineered soluble human CPR, which is shown to be * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  inactive, thus revealing a fundamental structural difference between the yeast and mammalian CPR forms.
Construction of Expression Vectors-The expression vectors containing yeast and human CPR, and the 5Ј-end truncated CPR forms were constructed in YEp51, a galactose inducible yeast expression plasmid. Yeast CPR and CPR (⌬33) genes were isolated by PCR using pFBY4 (16) containing yeast CPR as template. The 5Ј-sense oligonucleotide primers were 5Ј-CCCGTCGACATCATGCCGTTTGGAATAGACAAC-3Ј for yeast CPR and 5Ј-CCCGTCGACATCATGTCCGATGACGGAGATA-TC-3Ј for yeast CPR (⌬33) The latter corresponds to the sequence starting from methionine codon of the 33 amino acid from the N terminus and contained a SalI site at the 5Ј-end. The 3Ј-antisense oligonucleotide primers (5Ј-CCCAAGCTTTTACCAGACATCTTCTTGGTA-3Ј for yeast CPR and CPR (⌬33)) encoded a HindIII site at the 3Ј-end. The reaction conditions were as follows: 94°C (denaturation) for 1 min, 45°C (annealing) for 1 min and 72°C (extension) for 5 min with 2 min ramp for first 5 cycles and 94°C for 1 min and 72°C for 5 min for the remaining cycles in a 30-cycle reaction. PCR was carried out using Pfu polymerase (Stratagene) and a Perkin-Elmer DNA thermal cycler. Construction of expression vectors for human CPR, a truncated human CPR equivalent to yeast CPR (⌬33) (human (⌬50)), and human CPR digested with trypsin (⌬55) were produced in a similar way using human CPR as a template. The 5Ј-sense oligonucleotide primers (each encoding a SalI restriction site) were 5Ј-CCCGTCGACATCATGGGAG-ACTCCCACGTGGAC-3Ј for human CPR, 5Ј-CCCGTCGACATCATGC-CCGAGTTCACCAAAATTC-3Ј for human CPR (⌬50), and 5Ј-CCCGTC-GACATCATGATTCAGACATTGACCTCCT-3Ј for human CPR (⌬55). The 3Ј-antisense oligonucleotide primer for human CPR, CPR (⌬50), and CPR (⌬55), encoding a HindIII site following the stop codon, was 5Ј-CCCAAGCTTCTAGCTCCACACGTCCAGGGA-3Ј. PCR conditions for generation of the human CPR fragments were 1 min at 94°C (denaturation), 2 min at 45°C (annealing), and 1.5 min at 72°C (extension) in a 30-cycle reaction using Pfu DNA polymerase. The target fragments were gel purified, digested with SalI and HindIII, and cloned into YEp51. Transformants were screened by restriction digestion and confirmed by sequencing. All DNA manipulations and transformations were done using standard protocols (17).
CPR Gene Disruption in JL20 -In the JL20 strain, the chromosomal CPR gene was disrupted by inserting a URA3 to generate the JL20 (CPR::URA3) strain (13). The 0.7-kilobase BamHI internal fragment of the CPR gene in the CPR:YEp51 construct was replaced by the 1.1kilobase URA3-containing HindIII fragment isolated from plasmid pJJ244 (18) by filling in and blunt-ended ligation to obtain a CPR::URA3:YEp51 construct. Chromosomal CPR in the haploid JL20 yeast strain was then inactivated by transplacement using the SalI-HindIII fragment containing CPR::URA3, and the disruption was confirmed by PCR, using the cells of URA3 colonies of the transformed JL20 strain, the primers and conditions mentioned above (the cells were heated in a microwave for 30 s and quickly cooled on ice for lysis before being subjected to PCR), and ketoconazole susceptibility tests.
Sterol Isolation and Analysis-Yeast cells harvested from 100 ml of culture were resuspended in 3 ml of methanol, 2 ml of 60% (w/v) KOH in water, and 2 ml of 0.5% (w/v) pyrogallol dissolved in methanol and saponified by heating at 90°C for 1 h. Nonsaponifiable lipids (sterols) were extracted from the saponified mixture three times with 5 ml of hexane each time, pooled, and dried under nitrogen. The sterols were suspended in 100 l of toluene and heated at 60°C for one hour for silylation after adding 20 l of bis(trimethylsilyl) trifluoride. The silyl sterols were analyzed by gas chromatography/mass spectrometry (VG 12-250; VG BIOTECH) by using split injections with a split ratio of 20:1. Sterol identification was by reference to relative retention times and mass spectra reported previously (19,20).
Immunoblot Analysis-SDS-10% polyacrylamide gel electrophoresis, nitrocellulose filter transfer, and immunodetection of protein were performed as described previously (21,22). Yeast CPR antiserum was kindly provided by J. C. Loper of Cincinnati University. Anti-CPR IgG was purified from the CPR antiserum as described (23).
Heterologous Expression in the JL20 Strain-The JL20 transformants carrying native and truncated yeast and human CPR expression vectors were grown in yeast minimal medium containing Difco yeast nitrogen base without amino acids (1.34%, w/v), 20 g/ml L-histidine, and glucose (2%, w/v) at 30°C until glucose was completely consumed, and then heterologous expression was induced with galactose (3%, w/v) for 20 h (24).
Preparation of Cell Extracts, Cytosol, and Microsomes-Cells harvested from the cultures by centrifugation were resuspended in Buffer A (100 mM potassium phosphate containing 20% glycerol, 1 mM reduced glutathione, and 0.5 mM EDTA) and homogenized with glass beads (0.45-0.5 mm in diameter) in a Braun disintegrator (Braun GmbH, Mesungen, Germany) operating at 1500 ϫ g with 30-s bursts and carbon dioxide cooling. Cell extract was obtained as a supernatant by centrifuging cell homogenate at 1500 ϫ g for 10 min. The extract was centrifuged at 10000 ϫ g for 15 min to remove mitochondria as a pellet, and the resulting supernatant was centrifuged at 100,000 ϫ g for 90 min to obtain microsomes as a pellet and cytosol as a supernatant. The microsomal pellet was resuspended in Buffer A using a Potter-Elvehjem homogenizer. Protein content in cell extract and microsomes was measured using the BCA protein estimation kit (Sigma) and bovine serum albumin as a standard (19,24).
Purification of Soluble Yeast and Human CPR-Soluble yeast and human CPR purifications were carried out at 4°C and a flow rate of 1 ml/min as described previously, with few modifications, and omitting detergent in the purification steps (25). The cytosol of JL20 expressing soluble CPR was precipitated with ammonium sulfate, and the precipitate obtained with 40 -65% saturated ammonium sulfate was redis- Ci of 53 mCi/mmol) was incubated at 37°C and 150 rpm. After 2 h of incubation, the reaction was stopped by adding 1 ml of freshly prepared saponification reagent (15% (w/v) KOH in 90% (v/v) ethanol), and the mixture was saponified by heating at 90°C for 1 h. The nonsaponified sterols were extracted from the mixture twice with 3 ml of petroleum ether (boiling point 40 -60°C). The extracts were pooled, dried under nitrogen gas, and redissolved in 100 l of petroleum ether. The nonsaponifiable sterols were applied to silica gel thin layer chromatography plates (ART 573, Merck) and developed using toluene:diethyl ether at a ratio of 9:1 (v/v). Radioactive sterols were localized by autoradiogram and excised for scintillation counting (27).
Reconstitution of CYP Activity-The standard reaction mixtures contained purified S. cerevisiae CYP61, C. albicans and human CYP51 (0.5 nmol), 1 unit of either native yeast and human CPR or the N-terminal truncated soluble CPR forms, 13 nmol of the respective substrate (ergosta-5,7-dienol for S. cerevisiae CYP61 and dihydrolanosterol or 24methylene-24, 25-dihydrolanosterol for human and C. albicans CYP51, respectively) dispersed in 80 nmol dilauroylphosphatidylcholine; the reaction volume was adjusted to 950 l with 100 mM potassium phosphate buffer, pH 7.2. NADPH was added at a concentration of 1 mM to the mixture to start the reaction. C. albicans CYP51 was purified as described previously (28), and the human CYP51 was purified after GAL10 expression in yeast. 3 All reactions were incubated at 37°C for 20 min in a shaking water bath. Reactions were stopped by the addition of 3 ml of methanol, and the sterols were extracted and analyzed by gas chromatography/mass spectrometry as described above. Trimethylsilylated derivatives of sterol substrates and metabolites were clearly separated as two distinct peaks (for example, see Ref. 9). The conversion ratio was calculated from the areas of the two peaks, and the turnover (nmol of product formed/min/nmol of P450) was obtained from the amount of substrate added and the conversion ratio.
Spectrophotometric Measurements-A Philips PU8800 UV/VIS scanning spectrophotometer was used for all spectral studies. Cytochrome c reductase activity was measured as described previously (29). The absorption spectra of the purified yeast CPR (⌬33) was obtained by scanning from 300 to 700 nm (4). The ergosterol content in the nonsaponifiable lipids was calculated on basis of its molar extinction coefficient at 282 nm (E m ϭ 11,900).

Complementation Studies Using a cpr Ϫ Gene Disruptant-
The gene-disrupted strain (CPR::URA3/JL20) was generated as described above and had the phenotype described previously for a disruption using LEU2 in the same way (13). Deletion of the CPR gene led to a decrease in the levels of ergosterol. The ergosterol content in the cpr-disrupted strain was 4-fold lower than that of the undisrupted strain (JL20). However, ergosterol was the major sterol in both the strains and accounted for about 90% of the total sterols. In the cpr host strain, the ergosterol levels were restored to the levels in the CPR strain JL20 upon yeast CPR and CPR (⌬33) expression (Table I). Expression of these proteins in JL20 (CPR) did not alter the ergosterol content (data not shown).

Expression of Yeast and Human CPR Proteins in the Transformed Strains-For
Western blot analysis of the expression and localization of both yeast CPR and CPR (⌬33), the cytosolic and microsomal fractions of the cpr transformants expressing these proteins were probed with anti-CPR (Fig. 1). The mobilities of the expressed yeast proteins were consistent with their molecular masses (75 kDa for CPR (⌬33) and 78 kDa for CPR). The Western blot data suggested that GAL10-mediated expressed CPR (⌬33) and CPR proteins were localized in cytosolic and microsomal fractions, respectively. Heterologously expressed human native CPR and truncated human CPR (⌬50) (the equivalent deletion to yeast CPR (⌬33)) and (⌬55) were located in the microsomal and cytosolic fractions as determined spectrophotometrically. The yeast CPR contents were determined by using the cytochrome c reduction assay and compared with the expressed human CPR, CPR (⌬50), and CPR (⌬55) forms. Yeast and human CPR contents in the cytosolic and microsomal fractions of the transformants containing different expression vectors are shown in Table II. The specific content and yield of human CPR compared with yeast CPR was lower following heterologous expression in S. cerevisiae. Furthermore, it was observed that protein expression in the transformants expressing both soluble yeast and human CPR were lower than those expressing full-length CPR. No CPR was detected in the cytosolic fractions of the strain expressing either yeast or human native CPR protein. However, in transformants expressing soluble yeast and human CPR in the CPR host JL20, the enzyme was observed in the cytosol as well as in the microsomes. The amount of CPR present in the soluble CPR-expressing microsomes was quite similar to that of micro-somes from transformants containing the empty vector, YEp51, representing the basal endogenous CPR level in the strain JL20. In addition, no CPR was detected in the microsomal fractions of the cpr strain transformed with either YEp51 or CPR (⌬33):YEp51 after induction of expression, but it was observed in the cytosolic fractions. The CPR level found in the cytosol was the same as that observed in the cytosol of CPR strain expressing CPR (⌬33):YEp51 (data not shown). These results clearly indicated that the N-terminal truncated yeast and human CPR localized in cytosol after expression similar to the previous reports following trypsin cleavage experiments on mammalian CPR (3,4).
Purification of Soluble Truncated CPR-Purification of yeast CPR (⌬33) to homogeneity by standard CPR methodology of ammonium sulfate fractionation, hydroxyapatite, and 2Ј,5Ј-adenosine diphosphate agarose (affinity) chromatography yielded a 75-kDa protein. The purified yeast CPR (⌬33) reduced cytochrome c at a rate of 24.3 mol/min/mg of protein, compared with 36 mol/min/mg of protein for human CPR (⌬50) and CPR (⌬55). From this activity, the CPR activity was estimated as 8 and 12 nmol/mg of protein for yeast and human truncated CPR by assuming 1 nmol CPR reduces 3 mol of cytochrome c/min (26). The oxidized and reduced spectra of purified yeast CPR (⌬33) protein were recorded to confirm the presence of flavin cofactors (Fig. 2). The oxidized spectrum of yeast CPR (⌬33) had peaks at 380 and 455 nm. The spectrum obtained with the NADPH reduced CPR (⌬33) showed a decrease in absorbance at 455 nm, an increase in absorbance at 380 nm, and a broad absorption band between 550 and 650 nm characteristic of the air-stable semiquinone form. These spectra were identical to those reported for native microsomal yeast CPR (4).
Catalytic Activities of Yeast CPR in Sterol Biosynthesis-In vitro ergosterol biosynthesis was carried out by using the cell extracts of the various transformants of the cpr host (Table III), an assay applied for assessment of ergosterol biosynthesis inhibitors. The amount of ergosterol synthesized in vitro was reduced by about 3-fold upon CPR disruption when compared  1. Immunodetection of localization of CPR (⌬33) and CPR protein expressed in JL20 (CPR::URA3). Cytosolic (C) and microsomal (M) proteins were separated by SDS-polyacrylamide gel electrophoresis, transferred onto a nitrocellulose sheet, and probed with polyclonal CPR anti-IgG. with the parental strain. However, these changes were reverted when yeast CPR and CPR (⌬33) proteins were expressed in the cpr strain or by adding purified yeast CPR (⌬33) to the cell extracts of cpr strain. The presence of the YEp51 expression plasmids in the cpr strain did not alter the in vitro ergosterol biosynthesis. To investigate further, reconstituted activity was examined using yeast CPR (⌬33) protein together with different purified CYP isozymes and comparison to the soluble human CPR forms. Reconstitution was observed for the soluble yeast CPR (⌬33) when added to samples containing S. cerevisiae CYP61, C. albicans, and human CYP51 (Table IV) Reconstituted CYP activities for the soluble yeast CPR (⌬33) were not significantly different from identical reactions using native yeast CPR. In contrast, only native human CPR reconstituted CYP61, C. albicans CYP51, and human CYP51 activity. Soluble human CPR (⌬50) (corresponding to yeast (⌬33)) and soluble human CPR(⌬55) (lacking the N terminus removed on tryptic digestion) both failed to produce detectable activities.

DISCUSSION
One of the earliest roles of CPR, as for the CYP superfamily, may have been in sterol biosynthesis. CPR also functions in the squalene epoxidation reaction (10), which precedes sterol 14␣demethylation undertaken by CYP51 (6), as well as sterol 22-desaturation undertaken by CYP61, one of the last steps in the ergosterol biosynthesis pathway (30). The previous finding of hypersensitivity to ketoconazole on disruption of the gene encoding CPR was in many ways consistent with the concept of inefficient sterol biosynthesis, so that only a low dose of CYP51 inhibitor would arrest growth.
We confirmed the observation of Sutter and Loper (13) of a 200-fold increased sensitivity in such a strain to ketoconazole, 2 but we observed only a relatively small reduction in ergosterol synthesized to about 25% of the parent strain when compared by dry weight of cells. Analysis of sterols by gas chromatography and thin layer chromatography did not reveal accumulation of intermediates, such as lanosterol or ergosta-5,7-dienol, that are indicative of a block to enzyme activity at the steps in which CPR participates. This indicated that the electron donor system remaining in the cells is efficient, delivering both electrons for CYP activity.
Previously, the N terminus of mammalian CPR has been indicated to be the membrane anchor for the protein because trypsin cleavage was observed to release a soluble domain (5). However, this was not catalytically active in reconstituted systems, suggesting a role for the N terminus in interactions with CYP. Our studies indicate that the N-terminal 33 amino acids are important for membrane anchoring and targeting for yeast CPR, which contains a similar hydrophobic sequence. In contrast to previous studies on mammalian enzyme, this cytosolic protein is catalytically active and complements cpr-disrupted sterol biosynthesis on both expression and addition of purified enzyme to the in vitro system. This observation was confirmed in reconstitution studies with CYP61, in which normal activity was observed, as well as studies with C. albicans and human CYP51. This difference from mammalian studies was confirmed for the human enzyme and reflects a fundamental difference between the CPRs from different kingdoms. We can conclude that the yeast CPR membrane anchor is not required for supporting P450 activity in vivo or in vitro. Recently, the structure of rat CPR was resolved at 2.6 Å (31), and similar resolution of the yeast enzyme structure may help to elucidate the fundamental differences between the enzymes. For crystallization, the rat protein was produced in its soluble form after tryptic digestion, and the manner in which the hydrophobic domain allows interaction with CYP enzymes remains unclear. Presumably, the hydrophobic domain is important for the correct spatial interaction to allow electron transfer to CYP enzymes. Such interactions appear unimportant for yeast CPR in its interaction with fungal and mammalian CYP forms, and we are currently undertaking crystallization trials on soluble yeast CPR to resolve this and other issues.    a Turnover numbers are recorded as nmol product formed/min/nmol of P450 and are shown Ϯ S.D. The products that were detected were ergosterol.
b ND, no detectable metabolism observed.