Cloning and characterization of the yeast HEM14 gene coding for protoporphyrinogen oxidase, the molecular target of diphenyl ether-type herbicides.

Protoporphyrinogen oxidase, which catalyzes the oxygen-dependent aromatization of protoporphyrinogen IX to protoporphyrin IX, is the molecular target of diphenyl ether type herbicides. The structural gene for the yeast protoporphyrinogen oxidase, HEM14, was isolated by functional complementation of a hem14-1 protoporphyrinogen oxidase-deficient yeast mutant, using a novel one-step colored screening procedure to identify heme-synthesizing cells. The hem14-1 mutation was genetically linked to URA3, a marker on chromosome V, and HEM14 was physically mapped on the right arm of this chromosome, between PRP22 and FAA2. Disruption of the HEM14 gene leads to protoporphyrinogen oxidase deficiency in vivo (heme deficiency and accumulation of heme precursors), and in vitro (lack of immunodetectable protein or enzyme activity). The HEM14 gene encodes a 539-amino acid protein (59,665 Da; pI 9.3) containing an ADP-βαβ-binding fold similar to those of several other flavoproteins. Yeast protoporphyrinogen oxidase was somewhat similar to the HemY gene product of Bacillus subtilis and to the human and mouse protoporphyrinogen oxidases. Studies on protoporphyrinogen oxidase overexpressed in yeast and purified as wild-type enzyme showed that (i) the NH-terminal mitochondrial targeting sequence of protoporphyrinogen oxidase is not cleaved during importation; (ii) the enzyme, as purified, had a typical flavin semiquinone absorption spectrum; and (iii) the enzyme was strongly inhibited by diphenyl ether-type herbicides and readily photolabeled by a diazoketone derivative of tritiated acifluorfen. The mutant allele hem14-1 contains two mutations, L422P and K424E, responsible for the inactive enzyme. Both mutations introduced independently in the wild-type HEM14 gene completely inactivated the protein when analyzed in an Escherichia coli expression system.

Protoporphyrinogen oxidase, which catalyzes the oxygen-dependent aromatization of protoporphyrinogen IX to protoporphyrin IX, is the molecular target of diphenyl ether type herbicides. The structural gene for the yeast protoporphyrinogen oxidase, HEM14, was isolated by functional complementation of a hem14-1 protoporphyrinogen oxidase-deficient yeast mutant, using a novel one-step colored screening procedure to identify heme-synthesizing cells. The hem14-1 mutation was genetically linked to URA3, a marker on chromosome V, and HEM14 was physically mapped on the right arm of this chromosome, between PRP22 and FAA2. Disruption of the HEM14 gene leads to protoporphyrinogen oxidase deficiency in vivo (heme deficiency and accumulation of heme precursors), and in vitro (lack of immunodetectable protein or enzyme activity). The HEM14 gene encodes a 539-amino acid protein (59,665 Da; pI 9.3) containing an ADP-␤␣␤-binding fold similar to those of several other flavoproteins. Yeast protoporphyrinogen oxidase was somewhat similar to the HemY gene product of Bacillus subtilis and to the human and mouse protoporphyrinogen oxidases. Studies on protoporphyrinogen oxidase overexpressed in yeast and purified as wild-type enzyme showed that (i) the NH 2 -terminal mitochondrial targeting sequence of protoporphyrinogen oxidase is not cleaved during importation; (ii) the enzyme, as purified, had a typical flavin semiquinone absorption spectrum; and (iii) the enzyme was strongly inhibited by diphenyl ether-type herbicides and readily photolabeled by a diazoketone derivative of tritiated acifluorfen. The mutant allele hem14-1 contains two mutations, L422P and K424E, responsible for the inactive enzyme. Both mutations introduced independently in the wild-type HEM14 gene completely inactivated the protein when analyzed in an Escherichia coli expression system.
Protoporphyrinogen oxidase (EC 1.3.3.4) is the penultimate enzyme of the heme biosynthetic pathway. This membranebound enzyme catalyzes the oxidative O 2 -dependent aromatization of the colorless protoporphyrinogen IX to the highly conjugated protoporphyrin IX. The existence of an enzyme catalyzing this oxidative step has long been controversial, since porphyrinogens are rapidly oxidized to their corresponding porphyrins in the presence of air via a light-sensitive, autocata-lytic reaction (1). However, biochemical and genetic evidences now indicates the enzymatic nature of protoporphyrinogen oxidation in living cells. Protoporphyrinogen oxidase deficiency has been described in bacteria (2)(3)(4), yeast (5), and man, where a lack of protoporphyrinogen oxidase activity is responsible for the inherited disease porphyria variegata (6). Studies on the structure and function of protoporphyrinogen oxidase were recently stimulated by the discovery that diphenyl ether-type herbicides are very potent inhibitors of the protoporphyrinogen oxidase activities of yeast, mammalian, and plant mitochondria and plant chloroplasts in vitro (7,8). The phytotoxicity of diphenyl ether-type herbicides is light-dependent and involves intracellular peroxidations promoted by protoporphyrin IX, the heme and chlorophyll precursor, leading to cell damage and lysis (9), mimicking the symptoms of human patients with porphyria variegata. Protoporphyrin IX toxicity is due to the nonenzymatic oxidation of accumulated protoporphyrinogen that diffused away from its site of synthesis and is thus not further metabolized to metalloporphyrins by iron or magnesium chelatases. However, the mechanism of protoporphyrin accumulation in diphenyl ether-treated plants may also involve some nonspecific porphyrinogen-oxidizing enzymes, such as peroxidases (10 -12), that can oxidize protoporphyrinogen and also coproporphyrinogen or uroporphyrinogen to their corresponding porphyrins. These activities are insensitive to diphenyl ether inhibition, but are markedly inhibited by reducing agents such as dithiothreitol (11). These nonspecific protoporphyrinogen-oxidizing activities may give rise to some discrepancies in the biochemical characterization of protoporphyrinogen oxidase in various organisms that are quite surprising, since the biochemical and structural properties of the enzymes of the heme biosynthesis pathway are remarkably well conserved through evolution (for a revue, see Ref. 13). In contrast, protoporphyrinogen oxidases purified from various sources appear to be rather different in terms of molecular mass, cofactor, and kinetic properties (14 -17), electron acceptor (18 -22), and even subunit composition (20). This biochemical diversity seems to be also found at the genetic level. The HemG gene of Escherichia coli, located at 87 min on the genetic map of E. coli K12 (4), has been cloned and sequenced (23). This gene encodes a 181-amino acid protein (M r 21,200) and complements a mutation affecting the oxidation of protoporphyrinogen and heme synthesis in E. coli. It restores the protoporphyrinogen oxidase activity to 7-25 times that of a nontransformed wild-type strain when introduced into the original mutant on a high copy number plasmid. However, a second genetic locus, HemK, has been recently described as being involved in protoporphyrinogen oxidation in E. coli (24). This gene is part of a HemA-PrfA-HemK operon and encodes a 225-amino acid protein without homology to the HemG gene product, that may be a component of an alternative protoporphyrinogen oxidation pathway. The hemG mutant of E. coli was recently used to clone a human cDNA (25) and its murine homologue (26) that complement the mutation. These cDNAs encode a 477-amino acid protein with protoporphyrinogen oxidase activity and deduced sequences identical to that of peptides obtained from partially purified bovine protoporphyrinogen oxidase (26). These mammalian protoporphyrinogen oxidases have substantial domains that are similar to the HemY gene product of Bacillus subtilis. HemY was isolated by complementation of a mutation affecting heme synthesis in vivo by altering either coproporphyrinogen oxidase, or protoporphyrinogen oxidase activities, or both (27). This gene encodes a 472-amino acid extrinsic membrane-bound polypeptide that, when overexpressed in E. coli, is able to oxidize protoporphyrinogen IX to protoporphyrin IX in vitro and, more efficiently, coproporphyrinogen III to coproporphyrin III (28,29), unlike any eukaryotic protoporphyrinogen oxidase. The HemY gene product is not inhibited by diphenyl ether-type herbicides (29) and thus resembles more the nonspecific oxidases recently found in higher plants that may be involved in the rapid accumulation of protoporphyrin in diphenyl ethertreated plants.
As part of our efforts to understand the relationship between structure and function of eukaryotic protoporphyrinogen oxidase and eventually to elucidate the molecular basis of protoporphyrinogen oxidase and hence the mechanisms underlying the pathophysiology associated with the enzyme defect (human porphyria) or inhibition (herbicidal effects of diphenyl ethers), we recently purified protoporphyrinogen oxidase from the yeast Saccharomyces cerevisiae. This integral protein of the inner mitochondrial membrane is a 55-kDa cationic FAD-containing flavoprotein synthesized as a high molecular mass precursor (58 kDa) that is rapidly converted to the mature membranebound form in vivo (30). Antibodies raised against purified yeast protoporphyrinogen oxidase were used to characterize the enzyme in a heme-deficient mutant lacking protoporphyrinogen oxidase activity (5) carrying a single nuclear mutation hem14-1 (31). This mutant strain was shown to synthesize normal amounts of an inactive protein that did not bind a tritiated inhibitor (32). We therefore undertook the cloning of the yeast gene by functional complementation of the hem14-1 gene defect. Like any heme-deficient yeast strain, the mutant cannot grow aerobically on nonfermentative carbon sources (ethanol, glycerol, or lactate), but grows via fermentative metabolism, provided the growth medium contains Tween 80 and ergosterol as precursors of unsatured fatty acids and sterols, whose syntheses are dependent on specific hemoproteins. The present study describes the isolation and nucleotide sequence of the protoporphyrinogen oxidase gene from the yeast S. cerevisiae, the phenotypic effects of its disruption and further characterization of the protoporphyrinogen oxidase protein overexpressed in yeast cells. The mutant hem14-1 allele was also sequenced, and normal and mutated protoporphyrinogen oxidases were analyzed in a E. coli expression system.
Plasmid pools of partial Sau3A digests of yeast chromosomal DNA ligated at the BamHI site of the high copy number vector YEpEMBL23 (34) or partial HindIII digests of yeast chromosomal DNA ligated at the HindIII site of the low copy number vector pRS316 (35) were kindly provided by Dr. D. Thomas (Laboratoire d'Enzymologie, CNRS, Gif sur Yvette, France).
The E. coli strains used were (i) DH5␣ (from Life Technologies, Inc.) for cloning, maintenance, and propagation of plasmids; (ii) BL21(DE3) (from Novagen) for expression studies; and (iii) XLmutS (from Stratagene) for selection of altered plasmids in site-directed mutagenesis experiments. They were grown in LB medium containing 100 g/ml ampicillin when necessary to maintain all of the Amp R -based plasmids.
The pBluescript KS(ϩ) phagemid vector (Stratagene) was the cloning vehicle for all HEM14 containing fragments used for sequencing reactions. The pT7-5 vector was used for expression studies of HEM14 in E. coli and as the template for in vitro site-directed mutagenesis of the cloned HEM14 gene.
Cloning and Sequencing Manipulations-The yeast strain CC750-3D was grown overnight in YEPD-Te at 30°C. Cells were diluted to an A 600 of 0.1 and allowed to grow to an A 600 of 0.6. Cells were collected by centrifugation and processed for transformation. Yeast was transformed by lithium acetate treatment with single stranded DNA as carrier (36). The cells were plated on glucose plates minus uracil (to select for Ura ϩ transformants) containing 50 mM nitroprussiate (to select for heme-synthesis complementation). Plasmids were extracted from transformed yeast cells for back-transformation into E. coli by the single-step procedure of Ward (37).
The overlaps of the HEM14 clone with the PRP22 and ACS1 clones were recognized initially by sequencing the 5Ј and 3Ј ends of the 4.9-kb 1 insert of plasmid pBS19-1 that complemented the hem14-1 mutation. The restriction site patterns of the cloned insert matched those published and prompted us to delete a 1.9-kb SalI-XhoI fragment yielding the plasmid pBS19-1XS9 that complemented the hem14-1 mutation. This plasmid was used for expression studies of the cloned protoporphyrinogen oxidase in yeast. The 3-kb XhoI-SmaI fragment of pBS19-1 was cloned at the XhoI-SmaI sites of the phagemid pBluescript(KS ϩ ). This fragment was sequenced on subclones obtained by (i) internal deletions in the KS ϩ BS19-1XhSm plasmid taking into account unique restriction sites in the insert and adequate sites of the polylinker of pKS ϩ (NruI/SmaI, EcoR5/SmaI, EcoR1-Klenow/SmaI, XbaI/XbaI); and (ii) one-directional deletions generated by exonuclease III digestions (38) in plasmid KS ϩ BS19-1XhSm cut by SacII and SmaI using the Erase-a-Base kit (Promega, Madison, WI). The double-stranded plasmid DNA was sequenced by the chain termination method using the modified T7 DNA polymerase (39) from the Sequanase v. 2.0 sequencing kit (U. S. Biochemical Corp. and ␣-35 S-dATP (1000Ci/mmol, Amersham Corp.). The sequencing primers used were universal (Ϫ40) and reverse primers available for pBluescript and various 17-mer oligonucleotides complementary to the HEM14 sequence already determined. The plasmid DNA (2-3 g) was first denaturated in 0.4 N NaOH for 10 min at room temperature, precipitated with cold ethanol and annealed to the primer for 30 min at room temperature at a molar ratio of 1:1 when using pBluescript primers, and 1:20 when using synthetic HEM14specific oligonucleotides. The oligonucleotides used as sequencing primers were synthesized on an Applied Biosystems DNA synthesizer and were used without further purification. To sequence the mutant allele hem14-1, the HEM14 locus was amplified by PCR from genomic DNA of the strains CC750-3D and W303-1A using the primers 5Ј-CGCGGATC-CCTTTGTTGCTTCGGAGTCGC-3Ј (HindIII; (ϩ) strand) and 5Ј CCCAAGCTTATAGTTCCAACCCATCATCG-3Ј (BamHI; (Ϫ) strand). The PCR products were purified from agarose gels (JetSorb, Genomed Inc.) digested by HindIII and BamHI and the repurified DNA (JetSorb, Genomed Inc.) was ligated into the dephosphorylated phagemid BlusScript KSϩ linearized with HindIII and BamHI. The nucleotide sequence of the wild-type (control processed under the same conditions) and mutant alleles were obtained on DNAs pooled from seven independent clones. The nature of the mutations was confirmed on both strands. The DNA for PCR amplification of genomic sequences was prepared by the technique of Hoffman and Winston (40).
HEM14 Gene Disruption-The KS ϩ BS19 -1XSm plasmid was used for gene disruption experiments. The 439-bp NruI-EcoR5 fragment comprising the initiation ATG codon was deleted and replaced by the 820-bp BamHI TRP1 gene taken from pDW (41) to construct the plasmid KS ϩ ⌬hem14::TRP1. This plasmid was cut completely with BglI, and used to transform the strains W303-1B/D and S150-2B. Trpϩ transformants were analyzed for their protoporphyrinogen oxidase deficiency phenotypes. Integration of the deleted/disrupted allele hem14⌬::TRP1 at the HEM14 locus was confirmed by PCR analysis of genomic DNA (data not shown). The phenotype of the disrupted cells was identical when a 1493 NruI-EcoRI fragment including most of the HEM14 open reading frame was deleted and replaced by the 820-bp BamHI TRP1 gene taken from pDW. HEM14 Expression in E. coli and Site-directed Mutagenesis-For expression of the yeast protoporphyrinogen oxidase gene in E. coli the HEM14 locus was amplified by PCR from the cloned gene on pBS19-XS1 using the primers 5Ј-GACTGGATCCAAGGAGTGTATAATGT-TATTACCATTAACAAAG-3Ј (HindIII; ribosome binding site; (ϩ) strand) and 5Ј-CCCAAGCTTATAGTTCCAACCCATCATCG-3Ј (BamHI; (Ϫ) strand). The PCR products were purified from agarose gels (JetSorb, Genomed Inc.) digested by HindIII and BamHI, and the DNA was repurified (JetSorb, Genomed Inc.) and ligated into the dephosphorylated expression vector pT7-5 linearized with HindIII and BamHI. The nucleotide sequence of the gene was checked by analysis of the plasmid DNA of a clone that expressed an active protoporphyrinogen oxidase. This plasmid pT7HB-H14 was used as a template for sitedirected mutagenesis using the Stratagene Cameleon mutagenesis kit as recommended by the supplier.
The L422P mutation was introduced using the oligonucleotide 5Ј-AACCCAAACGCCCCCAACAAATAAACAAAAGTG-3Ј. The K424E mutation was introduced using the oligonucleotide 5Ј-CGCCCTCAAC-GAATATACAAAAGTGACTGCGATG-3Ј. Enrichment in mutated plasmid was done by eliminating the single ScaI restriction site in the ampicillin resistance gene of pT7-5. The plasmids were then transformed into the BL21(DE3) strain of E. coli. The T7 RNA polymerase gene was chemically induced by adding 0.2 mM isopropyl-1-thio-␤-Dgalactopyranoside to the cell cultures grown to an OD 600 of 0.6 -1 and incubation for 3 h at 25°C. The cells were collected by centrifugation, resuspended in the lysis buffer (0.1 M potassium phosphate buffer, pH 7.2, containing 0.1 M KCl, 1% n-octyl glucoside, 1 mM EDTA, and 20 g/ml PMSF), sonicated three times for 5 s each, and centrifuged. The protoporphyrinogen oxidase activity in the resulting cell-free extracts was measured. The homogenates were diluted twice in Laemmli sample buffer and processed for electrophoresis and immunodecoration as described below.
Purification of Yeast Protoporphyrinogen Oxidase-Protoporphyrinogen oxidase was purified from the strain S150-2Bhem14⌬::TRP1 transformed with the plasmid pBS19-XS1 grown on YPEtOH medium (3 liters), harvested during the late stationary phase of growth (A 600 Ϸ 30; 20 g of cells, wet weight⅐liter Ϫ1 ). The purification procedure was modified from Gietz et al. (30) as follows. The membrane fraction enriched in mitochondrial membranes (42) was homogenized in a Potter-Elvejhem homogenizer in 0.1 M potassium buffer containing 1 mM EDTA, 1 mM DTT, and 70 g⅐ml Ϫ1 PMSF at a protein concentration of 40 mg⅐ml Ϫ1 . The membrane suspension was then diluted twice with the same buffer containing 2% n-octyl glucoside and left for 20 min at 4°C under magnetic stirring to solubilize. Solid ammonium sulfate (50% saturation, final concentration) was added, and the mixture was stirred for 30 min and centrifuged for 1 h at 150,000 ϫ g. Soluble protoporphyrinogen oxidase was recovered in the supernatant and was loaded onto phenyl-Sepharose equilibrated with 0.1 M potassium phosphate, pH 7.2 containing 20% glycerol, 1 M KCl, 0.1 mM EDTA, 1 mM DTT, and 20 g⅐ml Ϫ1 PMSF. The column was washed with the same buffer and then with 0.01 M potassium phosphate, pH 7.2, containing 20% glycerol, 0.1 mM EDTA, 1 mM DTT, and 20 g⅐ml Ϫ1 PMSF. The enzyme was eluted from the column with the same buffer containing 30% ethylene glycol and 1% n-octyl glucoside. The active fractions were concentrated on Amicon YM30 ultrafiltration membranes and loaded onto a DEAE-Sepharose column equilibrated in 0.01 M potassium phosphate, pH 7.8, containing 20% glycerol, 0.1 mM EDTA, 1 mM DTT, and 20 g⅐ml Ϫ1 PMSF. Protoporphyrinogen oxidase passed through the column. The active fractions were concentrated on Amicon YM30 ultrafiltration membranes. The resulting enzyme preparation was apparently homogeneous on SDSpolyacrylamide gel electrophoresis. The specific activity of the purified protein was 40,000 nmol of protoporphyrinogen oxidized⅐h Ϫ1 ⅐mg Ϫ1 protein at 30°C. Absorption spectra of the protein were recorded with a Uvikon 860 spectrophotometer. Interferences due to detergents were avoided by determining the NH 2 -terminal peptide sequence of purified protoporphyrinogen oxidase after SDS-polyacrylamide gel electrophore-sis (43), and electrotransfer of the protein to an Immobilon P membrane (Millipore SA) with 10 mM Tris borate, pH 8.8, as buffer. The protein was stained on Immobilon with Amido Black 0.01% in methanol and destained with water. The stained band was cut out and processed for automatic Edman degradation on an Applied Biosystems 490A peptide sequencer using standard procedures for 20 cycles.
Miscellaneous-Photoaffinity of purified protoporphyrinogen oxidase was carried out as described previously (32). Published procedures were used to prepare extracts from trichloroacetic acid-treated cells (44), for SDS-polyacrylamide gel electrophoresis (43), and for electrophoretic transfer of the proteins to nitrocellulose sheets (45). Preparations were incubated with the antiserum (IgG fraction) and visualized with alkaline phosphatase-conjugated anti-rabbit IgG-secondary antibodies, as recommended by the manufacturer (Promega).

RESULTS
Identification and Deletion of the HEM14 Gene-A new direct colorimetric assay on plates was used to identify transformants complemented in the hem14 mutation. The assay depends on the capacity of heme synthesizing cells to form blue colonies when plated on YEPD medium containing the dye sodium nitroprusside (YEPD-NP); heme-deficient mutants and mutants deficient in plasma membrane ferri-reductase activity grow as white colonies on YEPD-NP (47). A hem14-1 strain, CC750-3D was transformed with the DNA of genomic libraries cloned in either single-copy (pRS316) or multiple-copy (pEM-BLY23) cloning vectors. The cells were plated on a synthetic medium containing sodium nitroprusside. Heme-sufficient cells appeared as blue colonies on a background of white hemedeficient transformed colonies. Cells from blue clones were subsequently tested for their ability to grow on glycerol and ethanol (Gly ϩ phenotype), and for recovery of protoporphyrinogen oxidase activity. One blue colony out of 50,000 Ura ϩ transformants was obtained from the pRS316 library. The insert carrying the complementing function was Ͼ15 kb long and restored protoporphyrinogen oxidase activity to the level of the wild-type cell (1.5-2 nmol⅐h Ϫ1 ⅐mg Ϫ1 protein of cell free extract). Seven blue colonies out of 120,000 Ura ϩ transformants were obtained from the pEMBL23 library. The inserts carrying the complementing function were 5-7 kb long with common restriction fragments and restored protoporphyrinogen oxidase activity to 10 -25 times the level of a wild-type cell (30 -50 nmol⅐h Ϫ1 ⅐mg Ϫ1 protein of cell free extract). The plasmid carrying the shortest insert (4.9 kb) was further analyzed. Subcloning and deletion analysis of the insert indicated that the complementing activity was restricted to a 2.5-kb fragment (Fig. 1). This fragment was sequenced (see "Experimental Procedures") and found to contain a 1.6-kb open reading frame encoding a 539-amino acid protein (Fig. 2). The phenotype of a cell lacking this gene was determined by removing a 439-bp NruI-EcoRV fragment containing the initiation codon and the 120 amino-terminal residues from the cloned insert and replacing it with a 820-bp fragment containing the TRP1 gene. A 2.7-kb linear BglI fragment, now containing the TRP1 substitution, was transformed into haploid trp1 cells. Tryptophan prototroph transformants were selected on a synthetic medium supplemented with Tween 80, ergosterol, and hemin. The transformants had a phenotype closely related to that of the original hem14-1 mutants. They were respiratory deficient (no growth on glycerol or ethanol as sole source of carbon), they accumulated porphyrins (Fig. 3) and had no detectable protoporphyrinogen oxidase activity in vitro. However, they were strictly auxotroph for Tween 80 and ergosterol (Te), while the original mutant could sustain some growth without Te. The cells disrupted for the HEM14 gene also lacked any immunodetectable protoporphyrinogen oxidase protein (Fig. 4). Similar phenotypes were obtained by deleting most of the open reading frame by replacing the 1493 bp NruI-EcoRI fragment of HEM14 (Fig. 2) by the 820-bp fragment containing the TRP1 gene.
Mapping of the HEM14 Gene-Initial homology searches of several data banks using the E-mail Blast server at NCBI with the blastp and tblastn algorithms (48) revealed that the HEM14 gene product is identical to that of the open reading frame between the gene PRP22 coding for a RNA helicase-like protein (49) and a gene coding for an acyl-CoA synthase, FAA2 (50) on yeast chromosome V, sequenced as part of the yeast genome sequencing project (GeneBank TM SCE9537). This result is consistent with the genetic analysis of the protoporphyrinogen oxidase deficiency that revealed a linkage of the hem14-1 locus to its centromere (31) and to URA3, a known marker of chromosome V (approximately 17 cM distal of URA3, 16 tetrades analyzed). The physical location of HEM14 on the chromosome V was confirmed by in vitro amplification of a HEM14-specific sequence on selected lambda clones (5-6A3, 5-3D5, 5S1A2 and 5-2H3) of the ordered library of yeast chromosome V (51) and control DNA. Only clones 5S1A2 and 5-2H3 allowed the amplification of HEM14-specific sequences. These results allowed us to map the cluster of genes PRP22-HEM14-FAA2 on the right arm of chromosome V, between the centromere and the GPA2 locus (52).
Nucleotide Sequence of HEM14 and Features of the Deduced HEM14 Protein Sequence-We have determined the nucleotide sequence of both strands of a 2400-bp fragment overlapping the NruI and XbaI restriction sites (Fig. 2). The polypeptide predicted from the HEM14 gene is 539-amino acid long (molecular mass of 59,665 Da) with an ATG at position 462 and a TAA stop codon at position 2078. A search for a putative "TATA" box showed that one (TATAATA) is present at position 268. The 3Ј-flanking region of the HEM14 gene contains several T and A rich stretches and the TGA . . . TAGT . . . TTT characteristic of transcription termination and polyadenylation regions (53). The codon bias of the 540 codons is low (0.125), suggesting that the HEM14 gene is expressed at a low level. The amino-terminal sequence of the predicted amino terminus of protoporphyrinogen oxidase is consistent with that of proteins targeted to the mitochondria. It is rich in positively charged and hydroxylated amino acids and contains a potential amphiphilic helix. The distribution of positively and negatively charged amino acids in the protein sequence is shown in Fig. 5B. The calculated isoelectric point of the protein was 9.3. The COOH terminus of the protein appears to be more acidic than the NH 2 terminus (pI 6.5 for the 350 carboxyl-terminal fragment of the protein). The hydropathy profile of the protein determined according to Kyte and Doolittle (54) (Fig. 5A) revealed that protoporphyrinogen oxidase is a moderately hydrophobic protein with a single potential membrane-spanning segment (residues 13-33). However, this domain is very like the ADP-␤␣␤binding fold of several FAD-containing enzymes, such as amino acid oxidases, monoamine oxidases, fumarate reductase, lipoamide dehydrogenase, and succinate dehydrogenase (55). This domain is shown in Fig. 6 where the 11 highly conserved residues fingerprinting this structure are labeled with the symbols defined by Wierenga et al. (55). The only deviation to the fingerprint is in its position 7 (threonine instead of a small or hydrophobic residue) as found in pig and human glyceraldehyde phosphate dehydrogenase. This domain is therefore probably not a trans-membrane domain. Data base searches for sequences homologous to yeast protoporphyrinogen oxidase show that the yeast enzyme is significantly similar to the mammalian protoporphyrinogen oxidases and the HemY gene product of Bacillus subtilis. Fig. 6 shows a tentative alignment of the yeast, human and bacterial protoporphyrinogen oxidases using the Clustal w (v. 1.5) multiple sequence analysis software. These proteins all have the ADP-␤␣␤-binding fold.
Optimization of the Overexpression of the HEM14 Gene Product-The XhoI-SmaI fragment of pBS19-1 cloned in the multicopy vector pEMBL23 (pBS19XS9) was used for studies on the synthesis of the HEM14 gene product. This fragment contains the entire promotor of HEM14 and the transcription termination signals. A strain lacking an intact HEM14 (S150-2B,⌬hem14::TRP1) gene was transformed using this plasmid. The transformants were grown on synthetic or complete media with glucose, galactose or ethanol as carbon source. The increase in protoporphyrinogen oxidase activity was correlated with a similar increase in immunologically detectable protoporphyrinogen oxidase protein (Fig. 4). Protoporphyrinogen oxidase specific activity was maximum in the stationary phase of growth (Fig. 7). The activities in cells grown on ethanol and galactose were comparable; they were two to three times higher during the exponential phase of growth than those of cells grown on glucose. Transformants grown on ethanol and harvested during the stationary phase of growth produced 100times more protoporphyrinogen oxidase 100 times than the control strain (S150-2B) grown under identical conditions. The presence of the multicopy plasmid carrying HEM14 gene did not alter the generation time or yield of biomass of transformants. Over 95% of the protoporphyrinogen oxidase was bound to the membrane fraction in cell-free extracts.
propoxyphenyl)-1,3,4-oxadiazol-2-one) at concentrations comparable that inhibit wild-type yeast; the IC 50 were 2 and 30 nM, respectively (Fig. 8) (30). Protoporphyrinogen oxidase was purified to homogeneity (200-fold, 60% recovery) from the mitochondrial membranes of the transformed cells. The specific activity of the enzyme was 43,600 nmol of protoporphyrinogen oxidized⅐h Ϫ1 ⅐mg Ϫ1 protein. The Michaelis constant for protoporphyrinogen was 0.02 M. The enzyme was specific for protoporphyrinogen IX and did not oxidize coproporphyrinogen III to coproporphyrin III. The absorption spectrum of the enzyme as purified (Fig. 9) was typical of a flavoprotein with a flavin semiquinone. Preliminary electron spin resonance spectroscopy experiments did not detect the stable free radical, ex-pected in such a redox state of the flavin (data not shown). The reactivity of purified protoporphyrinogen oxidase toward a photoaffinity probe (32) was identical to that of the wild-type enzyme. The protoporphyrinogen oxidase polypeptide appeared to be specifically labeled by diazo-[ 3 H]acifluorfen (Fig. 10, lane  B). The labeling of the purified protein was completely blocked by 10 M acifluorfen-methyl, an efficient inhibitor of protoporphyrinogen oxidase (Fig. 10, lane C). The NH 2 -terminal peptide sequence of the protein was MLLPLTKLKPRAKVAVV. This sequence is identical to that deduced from the first ATG (Met) of the open reading frame of the HEM14 gene.
Sequencing of the hem14-1 Mutant Allele-The molecular basis for the defect leading to the synthesis of an inactive protein in a hem14-1 strain, was determined by sequencing the mutant allele in seven pools of genomic DNA amplified by PCR and cloned in the E. coli Bluescript KSϩ phagemid. The wildtype sequence was amplified and sequenced in a similar way as a control. Two point mutations were determined in the hem14-1 allele, T1726C, and A1731G, leading to two changes in the amino acid sequence L422P and K424E.
Expression of Yeast Protoporphyrinogen Oxidase in E. coli and Site-directed Mutagenesis-The wild-type HEM14 gene was amplified by PCR using synthetic oligonucleotides as primers, that included single restriction sites for subsequent cloning and a canonical bacterial ribosome binding site in an expression vector pT7-5 under the T7 RNA polymerase promoter. A recombinant plasmid was confirmed by sequencing the entire open reading frame and was transformed into the E. coli strain BL21(DE3). The synthesis of yeast protoporphyrinogen oxidase was induced by adding isopropyl-1-thio-␤-D-galactopyranoside to cells in exponential growth. The recombinant protein (1% of total protein after 3 h of induction at 25°C) was active and fully inhibited by a typical diphenyl ether type herbicide, acifluorfen-methyl, with an IC 50 of 2 nM. This cloned HEM14 gene was used as a template for site-directed mutagenesis. Two point mutations were introduced independently to obtain the L422P and K424E mutations alone. The two mutated proteins were produced in concentrations similar to the wild-type protein, and both were totally inactive. DISCUSSION We have isolated the HEM14 gene that encodes yeast protoporphyrinogen oxidase by functional complementation of a heme-deficient hem14-1 strain. It has been recently demonstrated that several oxidases are able to oxidize protoporphyrinogen to protoporphyrin in vitro (10 -12). Such oxidases may complement our mutant, depending on their intracellular location and concentration. It was therefore essential to establish nonambiguously that HEM14 is the structural gene of yeast protoporphyrinogen oxidase. This was done by genetic and biochemical studies. First, the hem14-1 mutation was complemented by the cloned gene on both single and multiple-copy plasmids. Second, yeast strains with a disrupted HEM14 open reading frame were constructed. These strains had the phenotypes of a protoporphyrinogen deficiency, producing a respiratory deficiency, accumulation of porphyrins and lack of immunodetectable protoporphyrinogen oxidase protein. Third, the HEM14 gene product (characterized by the NH 2 -terminal protein sequence) was overexpressed in yeast. The product was recognized by anti-yeast protoporphyrinogen oxidase antibodies and was purified as authentic protoporphyrinogen oxidase. The protein was fully inhibited by a variety of diphenyl ethertype herbicides and labeled by photoaffinity using a highly specific probe. Fourth, two mutations in the hem14-1 allele were characterized, and the mutations were further analyzed by site-directed mutagenesis of the wild-type allele in an expression system for the HEM14 gene in E. coli.
The mapping of the HEM14 gene on chromosome V was demonstrated by both genetic linkage to URA3, a known marker of this chromosome and sequence specific in vitro amplification of the HEM14 locus on selected fragments of an ordered library of chromosome V (51). The genes flanking HEM14 are PRP22 at the 5Ј end of HEM14 and FAA2 already mapped on chromosome V (50), at the 3Ј end of HEM14. The approximate genetic distance between URA3 and HEM14 (17 cm, 16 tetrades analyzed) is compatible with the physical distance calculated from Isono's and Rile's maps (60 kb) and from the complete sequence of a set of ordered cosmids covering the entire chromosome V, as the contribution led by Dr. D. Bostein to the Yeast Genome Sequencing Project. Our results allow the "putative 59.5-kDa open reading frame" between PRP22 and FAA2 to be assigned to HEM14. HEM14 lies in a quite compact region of chromosome V, since the distance between the TAA stop codon of PRP22 and the ATG initiation codon of HEM14 is only 322 bp, and the distance between the TAA stop codon of HEM14 and the ATG initiation codon of FAA2 is only 321 bp. A single plasmid complementing the hem14 mutation was obtained by screening a yeast DNA library generated by partial HindIII digestion of genomic DNA. The insert was very long (Ͼ15 kb). This result is consistent with the fact that the PRP22-HEM14-FAA2 locus lies in a Ϸ20-kb HindIII fragment.
Several features of yeast protoporphyrinogen oxidase were revealed by the amino acid sequence of the protein deduced from the nucleotide sequence of the HEM14 gene which were not predicted by the biochemical characteristics of the enzyme. The predicted molecular mass of the HEM14 gene product is Ϸ59 kDa, close to the 58-kDa determined for the precursor form of the protein, with an isoelectric point of 9.3 comparable to that of the purified protein (pI Ͼ8.5). Yeast protoporphyrinogen oxidase was described as an integral protein of the inner mitochondrial membrane (30). The predicted sequence of protoporphyrinogen oxidase should therefore contain those amino acid sequences that target the protein to the mitochondria and anchor it to the lipid bilayer. The first 13 amino acids of the open reading frame of the HEM14 gene contains all the information typical of a protein imported to the mitochondria, without acidic residues, but with high contents of Leu, Arg, Lys and Thr, and a propensity to form an amphiphilic helix. This sequence is immediately followed by a putative ␤␣␤-dinucleotide binding fold, suggesting that the 13 amino-terminal residues may act as a cleaved presequence. We therefore determined the NH 2 -terminal sequence of purified protoporphyrinogen oxidase. The terminal 17 residues are identical to the deduced NH 2 -terminal sequence of the HEM14 gene product, and starts with the initiation methionine, as expected from the known specificity of the yeast methionine aminopeptidase that do not cleave the Met-Leu bond (56). The lack of cleavage of the mitochondrial targeting sequence of the protein during the importation of the protein into the mitochondria is not usual for an inner membrane-bound protein, but a sequence as short as 11 residues has been characterized as the uncleaved targeting sequence of the mitochondrial isoform of the tRNA N 6 -adenosine isopentyl transferase in yeast (57). Although the sequence of protoporphyrinogen oxidase contains several hydrophobic regions, none of those is longer than 15 uncharged residues and they are therefore unlikely to form membrane spanning segments. However, shorter helical domains could be responsible for insertion of the protoporphyrinogen oxidase into the inner mitochondrial membrane, as described for prostaglandin H 2 synthase-1, a monotopic membrane-bound protein of the endoplasmic reticulum (58). Another possibility, compatible with a post-translational modification of the protein leading to the shift in electrophoretic mobility initially attributed to the proteolytic cleavage of a putative presequence previously described (30), is that protoporphyrinogen oxidase is anchored to the inner mitochondrial membrane by a different mechanism, such as acylation. Preliminary experiments tend to support this hypothesis. 2 One of our goals is to obtain enough yeast protoporphyrinogen oxidase for structural studies on the protein. Commercially available yeast cells contain slightly more protoporphyrinogen oxidase activity and protein than laboratory grown cells. The industrial strains are usually grown on molasses and harvested in the late stationary phase of growth. This indicated that protoporphyrinogen oxidase seems to be fairly stable under such conditions, but the enzyme accounts for less than 0.001% of the membrane-bound proteins. We further optimized the production of the enzyme in our transformed laboratory strains by measuring protoporphyrinogen oxidase activity during the aerobic growth of yeast cells in media containing various carbon sources. Protoporphyrinogen oxidase activity was maximum during the stationary phase of growth of cells up to 8000 units (0.2 mg)/g (wet weight) of transformed cell in all cases. The activity during the exponential phase of growth was 2-3 times higher in cells grown on galactose or ethanol than in cells grown on glucose. Detailed analysis of the promotor of HEM14 will be needed to find out more about the mechanism and physiological relevance of HEM14 expression regulation in yeast. The overexpressed protein is associated with the membrane fraction. Both the membrane-bound and the purified enzyme are inhibited by diphenyl ether type herbicides and specifically photolabeled with an affinity probe, as is the enzyme from a wild-type strain.
The primary sequences of yeast protoporphyrinogen oxidase and related proteins (mammalian protoporphyrinogen oxidases and HemYp) are not easy to compare because of the overall low level of similarity of the proteins; the best conserved domain is the ␤␣␤-ADP binding fold in the NH 2 terminus of these proteins. If the targeting sequence of yeast protoporphyrinogen oxidase is short and not cleaved, the mammalian protoporphyrinogen oxidases are synthesized without such mitochondrial targeting sequences and therefore have a NH 2 terminus starting with the ␤␣␤-ADP binding fold as do monoamine oxidases, proteins associated with the outer membrane of mitochondria. As mentioned previously, this motif is found in the NH 2 -terminal domain of several well characterized flavoproteins (monoamine oxidases, amino-acids oxidases, and so forth) and, in this area, of some proteins of unknown function. A search in the S. cerevisiae genome data base (at www.genome-stanford.edu) with "protoporphyrinogen" as a keyword identified the YHR009c open reading frame. Careful examination of this sequence showed that this gene product has a NH 2 -terminal sequence very similar to that of HemYp (and therefore to the yeast protoporphyrinogen oxidase) but has no other domain similar to either HemYp or HEM14p. This unknown protein may, however, be a flavoprotein. The central question is to determine whether the conserved residues in the alignment of the yeast mammalian and bacterial proteins are involved in the substrate specificity of the enzymes and/or the catalytic activity. The B. subtilis enzyme has a poor substrate specificity, while the eukaryotic enzymes are more specific for protoporphyrinogen IX. However, the mammalian protoporphyrinogen oxidases are more closely related to HemYp than to the yeast 2 S. Arnould and J.-M. Camadro, unpublished results. enzyme, especially in their COOH-terminal domain. The two similar blocks in the mammalian and yeast enzymes are in poorly structured domains. These domains may, however, be important for the specificity of the enzyme.
The characterization of mutations affecting protoporphyrinogen oxidase activity should help determine the structure/function relationships. The short deletion affecting the HemYp function recently described (28) is located close to the end of the ␤␣␤-ADP binding fold. The present paper, describes the first mutations affecting an eukaryotic protoporphyrinogen oxidase. The hem14-1 mutant allele carries two point mutations that introduce two changes in amino acids in a relatively COOHterminal part of the protein, without significant similarity to the mammalian or bacterial protoporphyrinogen oxidases. The two mutations, L422P and K424E, are located in a domain close to a cysteine residue (Cys-435) in a Gly-Gly-Cys motif that may, by analogy to bovine monoamine oxidase (59), be involved in the covalent binding of FAD to the enzyme. The two mutations are rather different in nature. A proline may drastically affect secondary structure in the protein, while the lysine to glutamic acid change may affect the catalytic properties of the enzyme. One model of how the enzyme acts involves the removal of hydrogen from protoporphyrinogen as at least one hydride involving the flavin, and one proton involving a basic residue in the protein. It is thus tempting to speculate that Lys-422 may be this basic residue. Alternatively, the lysine may help stabilize the flavin in an anionic semiquinone form necessary for activity, as in other oxidases (for a review, see Ref. 60). In a preliminary attempt to better characterize each mutation, we introduced the two mutations independently into the wild-type gene cloned in an expression system in E. coli. Both mutations led to the production (under standard conditions) of an inactive protein in E. coli. These mutated genes will now be introduced into the yeast strain lacking a functional HEM14 gene to detect residual activity through functional complementation or the involvement of some post-translational modification affecting the functioning of the mutated enzymes in vivo and in vitro.
The availability of transformed cells overexpressing protoporphyrinogen oxidase (yeast and E. coli) overcomes the basic barriers to investigating the structural properties of the enzyme and the topology of its active site, due to the small amounts of the protein in wild-type yeast strains. This approach will also help the characterization of the molecular changes involved in the mutation in the protoporphyrinogen oxidase-deficient yeast strain that provided the basis for sitedirected mutagenesis of the yeast protoporphyrinogen oxidase gene. Our present results provide evidence that yeast protoporphyrinogen oxidase is an adequate model to elucidate the molecular basis of protoporphyrinogen oxidase function, and hence the mechanisms underlying the pathophysiology associated with the enzyme defect (human porphyria) and inhibition (herbicidal effects of diphenyl esters).