Fission yeast decaprenyl diphosphate synthase consists of Dps1 and the newly characterized Dlp1 protein in a novel heterotetrameric structure Ryoichi Saiki, Ai Nagata, Naonori Uchida, Tomohiro Kainou, Hideyuki Matsuda and Makoto Kawamukai Department of Applied Bioscience and Biotechnology, Faculty of Life and Environmental Science, Shimane University, Matsue, Japan The analysis of the structure and function of long chain- producing polyprenyl diphosphate synthase, which syn- thesizes the side chain of ubiquinone, has largely focused on the prokaryotic enzymes, and little is known about the eukaryotic counterparts. Here we show that decaprenyl diphosphate synthase from Schizosaccharomyces pombe is comprised of a novel protein named Dlp1 acting in part- nership with Dps1. Dps1 is highly homologous to other prenyl diphosphate synthases but Dlp1 shares only weak homology with Dps1. We showed that the two proteins must be present simultaneously in Escherichia coli transformants before ubiquinone-10, which is produced by S. pombe but not by E. coli, is generated. Furthermore, the two proteins were shown to form a heterotetrameric complex. This is unlike the prokaryotic counterparts, which are homodimers. The deletion mutant of dlp1 lacked the enzymatic activity of decaprenyl diphosphate synthase, did not produce ubiqui- none-10 and had the typical ubiquinone-deficient S. pombe phenotypes, namely hypersensitivity to hydrogen peroxide, the need for antioxidants for growth on minimal medium and an elevated production of H 2 S. Both the dps1 (formerly dps)anddlp1 mutants could generate ubiquinone when they were transformed with a bacterial decaprenyl diphosphate synthase, which functions in its host as a homodimer. This indicates that both dps1 and dlp1 are required for the S. pombe enzymatic activity. Thus, decaprenyl diphosphate from a eukaryotic origin has a heterotetrameric structure that is not found in prokaryotes. Keywords: Schizosaccharomyces pombe; decaprenyl diphos- phate synthase; ubiquinone; CoenzymeQ. Ubiquinone was identified initially as an essential factor in aerobic growth and oxidative phosphorylation in the electron transport system. Recently, however, multiple additional functions of ubiquinone have been proposed. One such function is its apparent role as a lipid-soluble antioxidant that prevents the oxidative damage of lipids due to peroxidation [1–4]. Studies using ubiquinone-deficient yeast mutants also suggested that one of the in vivo functions of ubiquinone is to protect against oxidants [5,6]. There is also a proposed function that links between sulfide meta- bolism and ubiquinone. Sulfide-ubiquinone oxidoreductase, which was previously thought to occur mainly in photo- biosynthetic bacteria, and which acts as a component in an energy metabolic pathway, has now been shown to be present in S. pombe and other eukaryotic organisms [7]. In addition, the clk-1 mutant of Caenorhabditis elegans,which has perturbed ubiquinone biosynthesis, shows a prolonged life-span, suggesting a novel role of ubiquinone [8–11]. Furthermore, an elegant study showed that ubiquinone (or menaquinone) accepts electrons that are generated by the formation of protein disulfide in E. coli [12]. Thus, ubiqui- none appears to play multiple roles. The ubiquinone biosynthetic pathway is composed of 10 steps, including methylation, decarboxylation, hydroxy- lation and isoprenoid transfer. The elucidation of this pathway has mostly involved studying respiratory-defici- ent mutants of E. coli and S. cerevisiae [13,14]. The length of the isoprenoid side chain of ubiquinone varies between organisms. For example, S. cerevisiae has ubiquinone-6, E. coli has ubiquinone-8, rats and Arabidopsis thaliana have ubiquinone-9, and humans and S. pombe have ubiquinone-10. The length of the side chain appears to be determined by polyprenyl diphosphate synthase, but not by the 4-hydroxybenzoate-polyprenyl-diphosphate trans- ferases, which catalyze the condensation of 4-hydroxy- benzoate and polyprenyl diphosphate [15–17]. The basis of this notion is that the heterologous expression in E. coli and S. cerevisiae of polyprenyl diphosphate synthase genes from other sources generated the same type of ubiquinone as that expressed in the donor organisms. By this method, Okada et al. successfully produced different ubiquinone species (ubiquinone-5 to ubiquinone-10) in the S. cerevisiae COQ1 mutant, which lacks its own poly- prenyl diphosphate synthase [18]. Moreover, 4-hydroxy- benzoate (PHB) polyprenyl diphosphate transferase was confirmed to have a broad substrate specificity by the heterologous expression of COQ2 in the ppt1D S. pombe strain [19] and the ubiA-delta E. coli strain [20]. Correspondence to M. Kawamukai, Department of Applied Bioscience and Biotechnology, Faculty of Life and Environmental Science, Shimane University, 1060 Nishikawatsu, Matsue 690–8504, Japan. Fax: + 81 852 32 6092, Tel.: + 81 852 32 6587, E-mail: kawamuka@life.shimane-u.ac.jp Abbreviations: PHB, 4-hydroxybenzoate; GPP, geranyl diphosphate; FPP, farnesyl diphosphate; GGPP, geranylgeranyl diphosphate; IPP, isopentenyl diphosphate; GST, glutathione-S-transferase. (Received 26 May 2003, revised 17 July 2003, accepted 22 August 2003) Eur. J. Biochem. 270, 4113–4121 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03804.x The genes encoding the short-chain polyprenyl diphos- phate synthases, including geranyl diphosphate (GPP, C 10 ) synthase, farnesyl diphosphate (FPP, C 15 )synthaseand geranylgeranyl diphosphate (GGPP, C 20 ) synthase, have been cloned from various organisms ranging from bacteria to mammals [21–24]. The mechanisms that determine the chain length synthesized by FPP synthase and GGPP synthase have been extensively studied [25,26]. All of these short-chain polyprenyl diphosphate synthases function as homodimers, except for the GPP synthase from spearmint, which functions as a heterotetramer [27,28]. In this latter GPP synthase, one subunit is similar to known prenyl diphosphate synthases but the other component does not contain the typical aspartate-rich motifs that are considered to be the substrate-binding sites. Unlike most of the short- chain polyprenyl diphosphate synthases, the medium-chain polyprenyl diphosphate synthases (C 30 and C 35 )thatare responsible for synthesizing the side chain of the menaqui- nones in Micrococcus luteus BP26, Bacillus stearothermo- philus,andBacillus subtilis are heterodimers [29,30]. In contrast, all the long-chain polyprenyl diphosphate synth- ases that synthesize the ubiquinone side-chains are thought to be homodimeric enzymes [14,21]. However, it has not been known what kind of enzyme component(s) the eukaryotic polyprenyl diphosphate synthase contains even though its essential gene (e.g. COQ1 or dps1) has been identified [6,31]. In this study, we characterized the decaprenyl diphos- phate synthase of S. pombe. We found that the prenyl diphosphate synthase homologue of S. pombe,Dps1,forms a heterotetramer with another component that we identified and denoted Dlp1. Dlp1 is needed to make a functional enzyme as the dlp1 disruptant produced no ubiquinone, is sensitive to H 2 O 2 and requires an antioxidant to grow on glucose-containing medium, which is typical of the ubiqui- none-deficient S. pombe strains [6,19]. This is the first molecular characterization of a long chain-producing poly- prenyl diphosphate synthase that synthesizes the side-chain of ubiquinone in a eukaryote. Materials and methods Materials Restriction enzymes and other DNA-modifying enzymes were purchased from Takara Shuzo Co. Ltd. and New England Biolabs, Inc. [1– 14 C]IPP (1.96 TBqÆmol )1 )was purchased from Amersham Pharmacia Biotech Ltd. IPP, all-E-farnesyl diphosphate, geranylgeraniol, solanesol (all- E-nonaprenol), and polyprenols (C 40 –C 60 )fromAilanthus altissima were purchased from Sigma Chemical Co. Kiesel gel 60 F 254 thin-layer plates were purchased from Merck. Reverse-phase LKC-18 thin-layer plates were purchased from Whatman Chemical Separation. Strains and plasmids E. coli strains DH10B and DH5a were used for the general construction of plasmids [32]. Plasmids pBluescript SK+/–, pT7Blue-T (Novagen), pSTV28 (Takara Shuzo), pQE31 (Qiagen), pGEX-KG (Amersham Pharmacia), pREP1 [33], pREP2 (the LEU2 marker of pREP1 was exchanged with the ura4 marker), pDS473 (ura4 marker), pDS472 (ura4 marker) and pSLF173 (LEU2 marker) were used as vectors [34,35]. The S. pombe homothallic haploid wild-type strain SP870 (h 90 leu1–32 ade6-M210 ura4-D18)andthe diploid strain SP826 (h + leu1–32 ade6-M210 ura4-D18/h + leu1–32 ade6-M216 ura4-D18) [36] were used to pro- duce Ddlp1::ura4 strains by homologous recombination. KS10 (h + leu1–32 ade6-M216 ura4-D18 Ddps1::ura4)and NU609 (h 90 leu1–32 ade6-M210 ura4-D18 Dppt1::ura4) have been described previously [6,19]. Yeast cells were grown in YE (0.5% yeast extract, 3% glucose) or PM minimal medium with appropriate supplements as described by Moreno et al. [37]. YEA and PMA contain 75 lgof adenine per ml in YE and PM, respectively. The concen- tration of the supplemented amino acids was 100 lgÆmL )1 . DNA manipulations Cloning, restriction enzyme analysis, and preparation of plasmid DNAs were performed essentially as described previously [32]. The dlp1 gene was cloned and expressed in E. coli and S. pombe as follows. Two oligonucleotides were used to amplify the 1.8 kb fragment containing the dlp1 gene and the surrounding region, namely, 19G12X, 5¢-TCGAATTCGATGAGCTTTCCGTTC-3¢ (creates an EcoRI site) and 19G12Y, 5¢-CATGGATATCGCATTC-3¢ (contains an EcoRV site). The resulting 1.8 kb fragment was digested with EcoRI and EcoRV, cloned into the EcoRI- SmaI sites of pSTV28 to yield pSTVDLP1 and the EcoRI- EcoRV sites of pBluescript SK+ to yield pBSDLP1. To make pDISDLP1, it was necessary to remove the BamHI site of pBluescript SK+. After pBSDLP1 was digested with NotIandSmaI, it was blunt-ended using the Klenow Fragment (Takara Shuzo) and self-ligated. The resulting plasmid was digested with BamHI and ligated with the ura4 cassette derived from pHSG398-ura4. Two other oligonu- cleotides were also used to amplify the dlp1 gene, namely, 5-dlp1, 5¢-TCGTCGACGAGCTTTCCGTTC-3¢ (creates a SalI site) and 3-dlp1, 5¢-TCCCCGGGATTACTTCG AAAC-3¢ (creates a SmaI site). The amplified fragment was cloned into the SalI-SmaI sites of pREP1 to yield pRDLP1. To construct pGSTDLP1, the 5-dlp1Fu oligonucleotide (5¢-TCGCGGCCGCATGAGCTTTCCG-3¢, which creates a NotI site) and 3-dlp1 were used to amplify the dlp1 gene. The amplified fragment was cloned into the NotIandSmaI sites of pDS473 to yield pGSTDLP1. To construct pGEXDLP1, pBSDLP1 was digested with SmaIandSalI and then cloned into the SmaIandSalI sites of pGEX-KG to yield pGEXDLP1. To construct pBSDPS1, the oligonucleotides 5-dps1 (5¢-TCCTGCAGCATGATTCAGTATGTA-3¢, which cre- ates a PstI site) and 3-dps1 (5¢-TCGTCGACTCACTTC TTTCTCGTTAT-3¢, which creates a SalI site) were used to amplify the dps1 gene. The plasmid pKS18 containing the dps1 region from the S. pombe cDNA library was used as a template for PCR. The amplified fragment was cloned into the PstIandSalI sites of pBluescript SK+ to yield pBSDPS1. To construct pHADPS1, the oligonucleotides 5-dps1Fu (5¢-TCGCGGCCGCATGATTCAGTAT-3¢,cre- ates a NotIsite)and3-dps1wereusedtoamplifythedps1 region from the S. pombe cDNA library. The amplified fragment was cloned into the NotIandSalIsitesof 4114 R. Saiki et al.(Eur. J. Biochem. 270) Ó FEBS 2003 pSLF173 to yield pHADPS1. To construct pSTVHI- SDPS1, the oligonucleotides 5-Sph I-dps1 (5¢-GCGCATG CGATGATTCAGTATGTA-3¢,createsaSphIsite)and 3-dps1 were used to amplify the dps1 gene. The plasmid pBSDPS1 was used as a template for PCR. The amplified fragment was cloned into the SphIandSalI sites of pQE31. The pQE31 plasmid harboring the dps1 gene was digested with XhoIandSalI and cloned into the Sal IsiteofpSTV28 to yield pSTVHISDPS1. The ddsA gene that encodes the decaprenyl diphosphate synthase of Gluconobacter suboxydans was endowed with the putative mitochondrial transit sequences of the gene encoding PHB polyprenyl diphosphate transferase (ppt1)as follows. The mitochondrial transit sequences were amplified by PCR using the oligonucleotides 5¢-AGGTCGACAGA TTAGCATGTAAATAG-3¢ (creates a SalIsite)and 5¢-CGAAGCTTGGGGGTTACAGAGTTTGA-3¢ (cre- ates a HindIII site). The PCR products were cloned into the SalIandHindIII sites of pBluescript SK+ to yield pBSTP. Two oligonucleotides, namely, 5-ddsA (5¢-GCA AGCTTAAAGCTGTGGTTCAGGGTGCAG-3¢,creates a HindIII site) and 3-ddsA (5¢-TAGCATGCTTAGCGGG CCCGATTC-3¢,createsaSphI site), were used to amplify the 1.0 kb fragment containing the ddsA gene and an additional 23 amino acid segment beyond the first methio- nine. The amplified fragment was then cloned into pT7Blue- T to yield pT7DDSA. pT7DDSA was then digested with HindIII and BamHI (this site was in pT7Blue-T) and cloned into the HindIII and BamHI sites of pBSTP to yield pBSTPDDSA. To construct pRDDSA, pBSTPDDSA was digested with SalIandBamHI, and the fragment was cloned into the SalIandBamHI sites of pREP1. Gene disruption The one-step gene disruption technique was performed according to the procedure of Rothstein [38]. The pDISDLP1 plasmid was linearized using the appropriate restriction enzymes, and the linear plasmid was used to transform SP826 for uracil prototrophy. Southern hybridi- zation was performed as described previously [32]. Ubiquinone extraction and measurement Ubiquinone was extracted as described previously [6,19]. The crude extract of ubiquinone was analyzed by normal- phase TLC with authentic ubiquinone-10 as the standard. Normal-phase TLC was carried out on a Kieselgel 60 F 254 plate with benzene/acetone (97 : 3, v/v). The band contain- ing ubiquinone was collected from the TLC plate following UV visualization and extracted with chloroform/methanol (1 : 1, v/v). Samples were dried and redissolved in ethanol. The purified ubiquinone was further analyzed by HPLC with ethanol as the solvent. Prenyl-diphosphate synthase assay and product analysis Prenyl diphosphate synthase activity was measured by the method described previously [15]. S. pombe cells were grown on the mid-to-late log phase in PMA-based medium. All subsequent steps were performed at 4 °C. Cells were harvested by centrifugation, suspended in buffer A (100 m M potassium phosphate, pH 7.4, 5 m M EDTA, and 1 m M 2-mercaptoethanol). The washed cells were ruptured by vigorous shaking with glass beads 14 times for 30 s at 60 s intervals in an ice bath. After centrifugation of the homogenate, the supernatant was used as a crude enzyme extract. The incubation mixture contained 2 m M MgCl 2 , 0.2% (w/v) Triton X-100, 50 m M potassium phosphate buffer (pH 7.4), 5 m M KF, 10 m M iodoacetamide, 20 l M [ 14 C]IPP (specific activity 0.92 MBqÆmol )1 ), 100 l M FPP, and 1.5 mgÆmL )1 protein of the enzyme in a final volume of 0.5mL.Thesamplemixtureswereincubatedfor60minat 30 °C. Reaction products, such as prenyl diphosphates, were extracted with 1-butanol-saturated water and hydro- lyzed with acid phosphatase. The hydrolysis products were extracted with hexane and analyzed by reverse-phase TLC with acetone/water (19 : 1, v/v). Radioactivity on the plate was detected with a BAS1500-Mac imaging analyzer (Fuji Film Co.). The plate was exposed to iodine vapor to detect the spots of the marker prenols. Immunoprecipitation of Dps1 protein with Dlp1 protein To test whether Dlp1 and Dps1 form a heterologous complex, pGSTDLP1, which produces the glutathione-S- transferase (GST) and Dlp1 fusion protein, GST-Dlp1, and pHADPS1, which produces HA-fused Dps1, were trans- formed into S. pombe strain SP826. Transformants were grown to the stationary phase in PMA medium and 0.5 mL of culture was then inoculated into 50 mL of the same medium. The cultures were grown to the mid-to-late log phase. After the cells were collected by centrifugation at 3000 g for 5 min, the pellets were suspended in 0.1 mL of buffer A. The suspended cells were ruptured by vigorous shaking with glass beads 14 times for 10 s at 60 s intervals in an ice bath. After the addition of 0.3 mL of buffer A kept at 4 °C, the homogenate was centrifuged at 2000 g for 5 min. The supernatant solution was then mixed with glutathione Sepharose 4B (Amersham Pharmacia Biotech) at 30 °Cfor 60 min. This mixture was washed five times with 140 m M NaCl, 2.7 m M KCl, 10 m M sodium phosphate and 1.8 m M potassium phosphate, and then once with 50 m M Tris/HCl (pH 8.0) and 10 m M glutathione (glutathione elution buffer) to elute the GST-Dlp1 protein. To detect Dps1-Dlp1 heterotetramers, a crude protein extract of E. coli harboring pGEXDLP1 and pSTVHISDPS1 was incubated with the water-insoluble cross linker disuccinimidyl suberate (Pierce) (1 m M in final concentration) and subjected to Western blotting. Results Cloning of the dlp1 gene and construction of the dlp1D strain We hypothesized that the decaprenyl diphosphate synthase of S. pombe might be a heteromeric enzyme because we observed that expression of the dps1 gene alone in E. coli did not give any enzymatic activity [6]. In contrast, when single bacterial genes encoding prenyl diphosphate synthases are expressed in E. coli, functional enzymes are produced [16,39,40]. We looked for a potential partner of Dps1 by searching the S. pombe genomic DNA sequence in the Ó FEBS 2003 Decaprenyl diphosphate synthase from fission yeast (Eur. J. Biochem. 270) 4115 National Center for Biotechnology Information database for a gene with homology to dps1. We found meaningful sequence similarity with an uncharacterized gene (SPAC19G12.12). The deduced amino acid sequence of the SPAC19G12.12 gene possessed the conserved domains I to VII of Dps1, but did not contain DDXXD sequence motifs that are typically found in all prenyl diphosphate synthases (Fig. 1). We designated this gene dlp1 (for D-less polyprenyl diphosphate synthase) and characterized it further. Fig. 1. Alignment of the amino acid sequences of long chain-producing polyprenyl diphosphate synthases. (1) Decaprenyl diphosphate synthase encoded by ddsA from Gluconobacter suboxydans (NCBI accession no. AB006850); (2) octaprenyl diphosphate synthase encoded by ispB from E. coli (accession no. NP417654); (3) component of decaprenyl diphosphate synthase encoded by dps1 from S. pombe (accession no. D84311); (4) hexaprenyl diphosphate synthase encoded by COQ1 from S. cerevisiae (accession no. J05547); (5) a novel component of deca- prenyl diphosphate synthase encoded by dlp1 from S. pombe (acces- sion no. AB118853). Residues conserved in more than two of the five sequences are boxed. Conserved regions (I–VII) are underlined. Numbers on the right indicate amino acid residue positions. Fig. 2. Plasmid constructs used in this study and Southern hybridization analysis of genomic DNAs from SP826, SP826Ddlp1 and RS312. (A) Plasmid constructs. pRDDSA and pRDLP1 contain the entire length of the ddsA and dlp1 genes, respectively. pHADPS1 and pGSTDLP1 express the HA–Dps1 and GST–Dlp1 fusion proteins, respectively. pRDDSA, pRDLP1, pHADPS1 and pGSTDLP1 are under the control of the strong nmt1 promoter. pGEXDLP1 contains the entire length of the dlp1 genefusedwiththeGST gene in pGEX-KG vector. pSTVHISDPS1 contains the entire length of the dps1 gene fused with His 6 -tag that allows to express the His–Dps1 fusion protein in E. coli. In pDISDLP1 the dlp1 gene was disrupted by the ura4 cassette on the vector pBluescript SK+. (B) Southern hybridization analysis. (I) Restriction map of the dlp1 and the dlp1::ura4 regions. Genomic DNAs of wild-type and dlp1 disruptants were prepared, separated on agarose gel, and probed with the ura4 gene (II) and the dlp1 gene from pSTVDLP1 (III). Arrows and the size calculated from the sequences in (I) matched with arrows indicated in (III). Lane 1, wild-type SP826 (diploid); lanes 2 and 3, SP826Ddlp1 (diploid); lane 4, RS312 (haploid). TP: Transit peptide from P p +1 [19]. Ba, BamHI; EI, EcoRI; EV, EcoRV; H, HindIII; N, NotI; P, PstI; Sa, SalI; Sp, SphI; Sm, SmaI; Xh, XhoI. 4116 R. Saiki et al.(Eur. J. Biochem. 270) Ó FEBS 2003 To assess the relevance of the dlp1 gene in ubiquinone biosynthesis, we constructed an S. pombe strain whose dlp1 gene has been disrupted (dlp1D). To do this, we constructed the plasmid pDISDLP1, in which the dlp1 gene is disrupted by the ura4 gene (Fig. 2A). This plasmid was then linearized by the appropriate restriction enzymes and the fragment was used to transform the S. pombe wild-type diploid strain SP826. About 20 Ura + transformant colonies could be picked and they were grown on YEA medium. After the stability of the ura4 + marker was examined by replica plating, two stable Ura + transformants were obtained. One of these strains, designated SP826Ddlp1, was allowed to make spores, and the germinated haploid cells were plated in replicates on plates containing YEA and PMA+Leu. While all cells grew well on YEA medium, some grew only very slowly on the PMA+Leu plate. One of these haploid strains, designated RS312, and the parental diploid SP826Ddlp1 strain were subjected to Southern hybridization analysis to confirm the proper disruption of dlp1 by ura4 (Fig. 2B). RS312 (dlp1::ura4) was then examined for ubiquinone synthesis as described in the Materials and methods. No ubiquinone was detected in RS312, although the RS312 strain that harbored the plasmid expressing dlp1 did show ubiquinone synthesis (Fig. 3). This encouraging result indicates that the dlp1 gene is involved in ubiquinone biosynthesis. Phenotypes of the dlp1 disruptant It was reported previously that KS10 (Ddps1::ura4), a strain of S. pombe whose decaprenyl diphosphate synthase- encoding dps1 gene has been disrupted, and NU609 Fig. 3. Detection of ubiquinone-10 in S. pombe strains. Ubiquinone was extracted from Wild-type SP870, KS10 (Ddps1::ura4), RS312 (Ddlp1::ura4), KS10 harboring pRDDSA or pRDPS1, and RS312 harboring pRDDSA or pRDLP1. Ubiquinone was first separated by thin-layer chromatography and then further analyzed by high-pressure liquid chromatography. Fig. 4. Recovery of RS312 growth on minimal medium by adding cys- teine or glutathione. (A) Wild-type SP870 harboring pREP2 (ura4 marker), NU609 (Dppt1::ura4), RS312 (Ddlp1::ura4) and RS312 har- boring pRDLP1 were grown on PM medium supplemented with 75 lgÆmL )1 adenine and 100 lgÆmL )1 leucine. (B) The same strains were grown on PM medium supplemented with adenine, leucine and 200 lgÆmL )1 glutathione. (C) The same strains were grown as in B except that cysteine was used instead of glutathione. NU609 and RS312 could not grow on PM medium (A) unless it was supplemented with glutathione (B) or cysteine (C). RS312 harboring pRDLP1 could grow on PM medium lacking glutathione or cysteine (A). Ó FEBS 2003 Decaprenyl diphosphate synthase from fission yeast (Eur. J. Biochem. 270) 4117 (Dppt1::ura4), a strain of S. pombe whose PHB polyprenyl diphosphate transferase-encoding ppt1 gene has been disrupted, are unable to produce ubiquinone and have some notable additional phenotypes [6,19]. These include H 2 O 2 and Cu 2+ sensitivity and a requirement of cysteine, glutathione or a-tocopherol for growth on minimal medium [19]. We thus tested RS312 for these phenotypes. RS312 (Ddlp1::ura4) was first grown on PM-based medium with and without supplementation with 200 lgÆmL )1 of cysteine or glutathione. RS312 cells did not grow on the minimal medium but the addition of cysteine or glutathione effect- ively caused their growth to recover (Fig. 4). RS312 cell growth also recovered when they were grown on minimal medium containing 1 m M a-tocopherol, a well-known lipid antioxidant (data not shown). In addition, the good growth of RS312 on supplemented medium was severely inhibited when 1 m M H 2 O 2 or 1 m M Cu 2+ was added (Fig. 5). Thus, RS312 cells bear the same phenotypes as the ubiquinone- nonproducers KS10 and NU609. RS312 was similar to KS10 and NU609 in another phenotype. We previously found that the S. pombe strains that were deficient in either dps1 or ppt1 produced H 2 S but the wild-type cells did not [6,19]. When we tested for the presence of H 2 S by assaying for its chemical reaction with lead acetate (which produces PbS), the RS312 culture was found to produce H 2 S (data not shown). In addition, when we measured the amount of acid-labile sulfide present in the cells, we found that the ubiquinone-less mutants of RS312 (Ddlp1) and NU609 (Dppt1) produced 12-fold-higher amounts of S 2– (1064.9 and 1110.1 nmol per 10 9 cells, respectively) than the wild-type strain SP870 (83.3 nmol per 10 9 cells). All these phenotypes of RS312 are identical to those of the two ubiquinone-less mutants that had been constructed earlier [6,19]. Thus, dlp1 is essential for ubi- quinone synthesis in S. pombe. Complementation of dlp1 -disrupted cells by expressing G. suboxydans ddsA To determine whether dlp1 is directly involved in the activity of decaprenyl diphosphate synthase, we expres- sed ddsA in the dlp1 and dps1 disruptants. The ddsA gene encodes the decaprenyl diphosphate synthase of G. suboxydans.ThisddsA gene is known to be completely functional when it is expressed in E. coli [40], which suggests that DdsA is a homomeric enzyme. A 45 amino- acid mitochondrial transfer signal from Ppt1, which is known to locate in the mitochondria [19], was first added to the ddsA gene product so that it would be located in mitochondria. The resulting ppt1-ddsA fusion gene suc- cessfully complemented both the dps1 and dlp1 disruptants as both cell types were able to produce ubiquinone (Fig. 3). Thus, the lack of ubiquinone synthesis in the dlp1 mutant,aswellasinthedps1 mutant, is due to the lack of decaprenyl diphosphate synthase activity. We further confirmed that the Ddlp1 strain, as well as the dps1D strain, is defective in decaprenyl diphosphate synthase activity by measuring the in vitro activity of this enzyme as described in the materials and methods. Both the Ddlp1 and Ddps1 strains did not produce decaprenol, while the Dppt1 strain and wild-type retained their activities (Fig. 6). However, when the Ddlp1 strain was transfectedwiththeplasmidcontainingdlp1, the enzy- matic activity was restored. Thus, both dlp1 and dps1 are essential for decaprenyl diphosphate synthase activity in S. pombe. Fig. 5. Sensitivity of the ubiquinone-less mutant to oxygen radical pro- ducers. Wild-type (circles) and RS312 (Ddlp1::ura4) (triangles) strains werepregrownandthenplacedinfreshYEAmediumwith1 m M H 2 O 2 (A), 1 m M Cu 2+ (B) or neither. Cell numbers were counted at 4-h intervals. Fig. 6. Thin-layer chromatogram of the product of decaprenyl diphos- phate synthase. The decaprenyl diphosphate synthase reactions of KS10 (Ddps1::ura4, lane 1), RS312 (Ddlp1::ura4, lane 2), SP870 (Wild- type, lane 3), NU609 (Dppt1::ura4, lane 4) and RS312 harboring plasmid pRDLP1 (lane 5) were measured using [ 14 C]IPPandFPPas substrates. The products were hydrolyzed with phosphatase and the resulting alcohols were analyzed by reverse-phase thin-layer chroma- tography. The same amounts of radiolabeled products (5000 d.p.m.) were applied to the TLC plate. The arrowhead indicates the position of the synthesized decaprenols. The positions of standard alcohols are indicated on the right: GGOH, all-E-geranylgeraniol; SOH, all-E-solanesol; Ori., origin; S.F., solvent front. 4118 R. Saiki et al.(Eur. J. Biochem. 270) Ó FEBS 2003 Expression of dps1 and dlp1 in E. coli We speculated that if the dps1 and dlp1 are sufficient for decaprenyl diphosphate synthase activity, cotransformation of E. coli DH5a with the pSTVHISDPS1 plasmid that expresses the dps1 gene and the pGEXDLP1 plasmid that expresses the dlp1 gene may result in the formation of a functional enzyme that generates the ubiquinone-10 species that is produced by S. pombe. We found that while the cells that were transformed only with pSTVHISDPS1 or pGEXDLP1 produced only ubiquinone-8, which is syn- thesized by the endogenous E. coli octaprenyl diphosphate synthase, the pSTVHISDPS1 and pGEXDLP1 double transformant produced a small amount of ubiquinone-10 Fig. 7. Detection of ubiquinone in E. coli transfected with constructs expressing dlp1 and/ or dps1. Ubiquinone was extracted from wild-type DH5a and DH5a harboring pGEXDLP1 and/or pSTVHISDPS1. Fig. 8. Tetrameric formation of Dlp1 and Dps1. (A,B) Crude proteins were extracted from SP826 harboring pHADPS1 and pGSTDLP1 (lanes 1 and 3) or pHADPS1 and pDS472 (lanes 2 and 4). The crude proteins were incubated in buffer A at 30 °Cfor1h and then purified by a GST column (lanes 1 and 2). Western blot analysis was performed using an anti-HA (A) or anti-GST Ig (B). Arrows indicate the positions of HA-Dps1 (A) andGST-Dlp1(B)protein.TheasteriskinB indicates the GST protein. (C,D) Crude proteins were extracted from E. coli BL21 harboring pGEXDLP1 and pSTVHISDPS1 (lane 2). The crude proteins were incubated with a cross-linker, disuccinimidyl suberate, at 30 °C for 30 min and then purified by a GST column (lane 1). Western blot analysis was performed using an anti-GST (C) or anti-His Ig (D). Ó FEBS 2003 Decaprenyl diphosphate synthase from fission yeast (Eur. J. Biochem. 270) 4119 (Fig. 7). This result indicates Dps1 and Dlp1 together form a heteromeric decaprenyl diphosphate synthase. Heterotetrameric formation of Dps1 and Dlp1 That dps1 and dlp1 are both required for the decaprenyl diphosphate synthase activity suggests that Dps1 and Dlp1 might form a complex. Consequently, we tested whether Dps1 and Dlp1 interact with each other to form a heteromer. HA-linked Dps1 and GST-linked Dlp1 were expressed together in S. pombe and subjected to pull-down assays. When the GST–Dlp1 fusion was pulled down by the GST-column, HA–Dps1 was coprecipitated, although when HA–Dps1 was expressed with GST alone, HA–Dps1 was not pulled down by the GST-column (Fig. 8A,B). Thus, Dlp1andDps1bindtoeachotherinS. pombe. To analyze the size of the Dps1–Dlp1 heteromer, we prepared the crude proteins of E. coli expressing the GST– Dlp1 and His–Dps1 fusion proteins and incubated them with the protein crosslinker disuccinimidyl suberate. The crosslinked proteins were then purified by a GST column and subjected to SDS electrophoresis. Western blotting with anti-GST and anti-His Igs detected a band with a molecular mass near the marker of 175 kDa, which can be considered to be a heterotetrameric molecule because the calculated molecular mass of heterotetromeric GST–Dlp1 and His– Dps1 is 188 kDa (Fig. 8C,D). Thus, wild-type Dlp1 and Dps1 form a heterotetramer in E. coli. Discussion We identified a novel gene named dlp1 that encodes a partner of Dps1, which together constitute the active decaprenyl diphosphate synthase in S. pombe. The Dlp1 protein is weakly similar in sequence to Dps1 but lacks the conserved regions of domains II and VI that are likely to be the prenyl diphosphate synthase substrate-binding sites. Dlp1 and Dps1 form a heterotetrameric complex in S. pombe and together reconstitute ubiquinone-10-gener- ating enzymatic activity in E. coli. The heterotetrameric structure of Dps1 and Dlp1 is novel since other long chain-producing polyprenyl diphosphate synthases that synthesize the side chain of ubiquinone appear to exist as homodimers [14,39]. That the S. pombe prenyl diphosphate synthase exists as a heteromer, unlike its known homodi- meric counterparts, may relate to the fact that the latter are prokaryotic enzymes. In other words, eukaryotes may have evolved heteromeric prenyl diphosphate synthases from the homodimeric prokaryotic synthases. Supporting this notion is our study with the homodimeric long chain-producing polyprenyl diphosphate synthase from E. coli, namely, the IspB octaprenyl diphosphate synthase [39]. Our group previously showed that when E. coli is transfected with a construct encoding a functionally inactive IspB molecule due to a mutation, an active enzyme is nonetheless formed when the mutant is paired with the wild-type enzyme [39]. This observation suggests that the components of the homodimeric enzyme could be subjected to evolutionary alteration wherein they act in a heteromeric form with another molecule. At present, it is not clear whether this heteromeric enzyme form occurs commonly in eukaryotes, but our preliminary data do suggest that the human decaprenyl diphosphate synthase is not a homomeric enzyme (data not shown). Heteromeric prenyl diphosphate synthases may be widely spread than it was thought as the medium chain-producing heptaprenyl diphosphate synth- ases from B. subtilis and M. luteus, which synthesize the side-chain of menaquinone [29,30], do occur as hetero- dimers and it was recently observed that the short chain- producing geranyl diphosphate synthase from spearmint forms a heterotetramer [27,28]. To date, we have obtained three ubiquinone-less mutants of S. pombe due to the disruption of dps1, ppt1 and dlp1. All three mutants displayed essentially the same phenotypes, namely sensitivity to H 2 O 2 and Cu 2+ ,the need for an antioxidant such as glutathione for growth on minimal medium, and the production of H 2 S. The former two phenotypes reflect the role ubiquinone plays as an antioxidant, while the latter phenotype supports the notion that ubiquinone acts as a sulfide oxidant. Unlike S. pombe, ubiquinone-less S. cerevisiae do not produce H 2 S. The lack of H 2 S production may be due to the fact that, unlike S. pombe and other eukaryotes, S. cerevisiae does not possess the genes encoding sulfide ubiquinone reductase, which oxidizes sulfide by ubiquinone. Thus, ubiquinone may play an common and important role in eukaryotes by participating in a sulfide detoxication pathway. Acknowledgements This work is supported by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan. We thank Asuka Honma for technical assistance. References 1. Frei, B., Kim, M.C. & Ames, B.N. (1990) Ubiquinol-10 is an effective lipid-soluble antioxidant at physiological concentrations. Proc.NatlAcad.Sci.USA87, 4879–4883. 2. Witting, P., Pettersson, K., Letters, J. & Stocker, R. (2000) Anti- atherogenic effect of coenzyme Q10 in apolipoprotein E gene knockout mice. Free Radic. Biol. Med. 29, 295–305. 3. Ernster, L. & Dallner, G. (1995) Biochemical, physiological and medical aspects of ubiquinone function. Biochim. Biophys. Acta 1271, 195–204. 4. Crane, F.L. (2001) Biochemical functions of coenzyme Q10. J. Am. Coll. Nutr. 20, 591–598. 5. Do, T.Q., Schultz, J.R. & Clarke, C.F. (1996) Enhanced sensitivity of ubiquinone-deficient mutants of Saccharomyces cerevisiae to products of autoxidized polyunsaturated fatty acids. Proc. Natl Acad. Sci. USA 93, 7534–7539. 6. Suzuki, K., Okada, K., Kamiya, Y., Zhu, X., Tanaka, K., Nakagawa, T., Kawamukai, M. & Matsuda, H. (1997) Analysis of the decaprenyl diphoshate synthase (dps) gene in fission yeast suggests a role of ubiquinone as an antioxidant. J. Biochem. (Tokyo) 121, 496–505. 7. Vande Weghe, J.G. & Ow, D.W. (1999) A fission yeast gene for mitochondrial sulfide oxidation. J. Biol. Chem. 274, 13250–13257. 8. Ewbank, J.J., Barnes, T.M., Lakowski, B., Lussier, M., Bussey, H. & Hekimi, S. (1997) Structural and functional conservation of the Caenorhabditis elegans timing gene clk-1. Science 275, 980–983. 9. Jonassen, T., Larsen, P.L. & Clarke, C.F. (2001) A dietary source of coenzyme Q is essential for growth of long-lived Caenorhabditis elegans clk-1 mutants. Proc. Natl Acad. Sci. USA 98, 421–426. 4120 R. Saiki et al.(Eur. J. Biochem. 270) Ó FEBS 2003 10. Branicky, R., Benard, C. & Hekimi, S. (2000) clk-1, mitochondria, and physiological rates. Bioessays 22, 48–56. 11. Larsen, P.L. & Clarke, C.F. (2002) Extension of life-span in Caenorhabditis elegans by a diet lacking coenzyme Q. Science 295, 120–123. 12. Bader, M., Muse, W., Ballou, D.P., Gassner, C. & Bardwell, J.C.A. (1999) Oxidative protein folding is driven by the electron transport system. Cell 98, 217–227. 13. Tzagoloff, A. & Dieckmann, C.L. (1990) PET genes of Saccharo- myces cerevisiae. Microbiol. Rev. 54, 211–225. 14. Kawamukai, M. (2002) Biosynthesis, bioproduction and novel roles of ubiquinone. J. Biosci. Bioeng. 94, 511–517. 15. Okada, K., Suzuki, K., Kamiya, Y., Zhu, X., Fujisaki, S., Nishimura, Y., Nishino, T., Nakagawa, T., Kawamukai, M. & Matsuda, H. (1996) Polyprenyl diphosphate synthase essentially defines the length of the side chain of ubiquinone. Biochim. Bio- phys. Acta 1302, 217–223. 16. Okada, K., Kamiya, Y., Zhu, X., Suzuki, K., Tanaka, K., Nakagawa, T., Matsuda, H. & Kawamukai, M. (1997) Cloning of the sdsA gene encoding solanesyl diphosphate synthase from Rhodobacter capsulatus and its functional expression in Escher- ichia coli and Saccharomyces cerevisiae. J. Bacteriol. 179, 5992– 5998. 17. Okada, K., Minehira, M., Zhu, X., Suzuki, K., Nakagawa, T., Matsuda, H. & Kawamukai, M. (1997) The ispB gene encoding octaprenyl diphosphate synthase is essential for growth of Escherichia coli. J. Bacteriol. 179, 3058–3060. 18. Okada, K., Kainou, T., Matsuda, H. & Kawamukai, M. (1998) Biological significance of the side chain length of ubiquinone in Saccharomyces cerevisiae. FEBS Lett. 431, 241–244. 19. Uchida, N., Suzuki, K., Saiki, R., Kainou, T., Tanaka, K., Matsuda, H. & Kawamukai, M. (2000) Phenotypes of fission yeast defective in ubiquinone production due to disruption of the gene for p-hydroxybenzoate polyprenyl diphosphate transferase. J. Bacteriol. 182, 6933–6939. 20. Suzuki, K., Ueda, M., Yuasa, M., Nakagawa, T., Kawamukai, M. & Matsuda, H. (1994) Evidence that E. coli ubiA product is a functional homolog of yeast COQ2 and the regulation of the ubiA gene expression. Biosci. Biotechnol. Biochem. 58, 1814– 1819. 21. Koyama, T. (1999) Molecular analysis of prenyl chain elongating enzymes. Biosci. Biotechnol. Biochem. 63, 1671–1676. 22. Liang, P.H., Ko, T.P. & Wang, A.H. (2002) Structure, mechanism and function of prenyltransferases. Eur. J. Biochem. 269, 3339– 3354. 23. Kainou, T., Kawamura, K., Tanaka, K., Matsuda, H. & Kawa- mukai, M. (1999) Identification of the GGPS1 genes encoding geranylgeranyl diphosphate synthases from mouse and human. Biochim. Biophys. Acta 1437, 333–340. 24. Okada, K., Saito, T., Nakagawa, T., Kawamukai, M. & Kamiya, Y. (2000) Five geranylgeranyl diphosphate synthases expressed in different organs are localized into three subcellular compartments in Arabidopsis. Plant Physiol. 122, 1045–1056. 25. Ohnuma, S., Hirooka, K., Tsuruoka, N., Yano, M., Ohto, C., Nakane, H. & Nishino, T. (1998) A pathway where polyprenyl diphosphate elongates in prenyltransferase. Insight into a common mechanism of chain length determination of prenyltransferases. J. Biol. Chem. 273, 26705–26713. 26. Wang, K.C. & Ohnuma, S. (1999) Isoprenyl diphosphate syn- thases. Trends Biochem. Sci. 24, 445–451. 27. Burke, C.C., Wildung, M.R. & Croteau, R. (1999) Geranyl diphosphate synthase: cloning, expression, and characterization of this prenyltransferase as a heterodimer. Proc. Natl Acad. Sci. USA 96, 13062–13067. 28. Burke, C.C. & Croteau, R. (2002) Interaction with the small subunit of geranyl diphosphate synthase modifies the chain length specificity of geranylgeranyl diphosphate synthase to produce geranyl diphosphate. J. Biol. Chem. 277, 3141–3149. 29. Shimizu, N., Koyama, T. & Ogura, K. (1998) Molecular cloning, expression, and characterization of the genes encoding the two essential protein components of Micrococcus luteus B-P 26 hexa- prenyl diphosphate synthase. J. Bacteriol. 180, 1578–1581. 30. Zhang, Y W., Koyama, T. & Ogura, K. (1997) Two cistrons of the gerC operon of Bacillus subtilis encode the two subunits of heptaprenyl diphosphate synthase. J. Bacteriol. 179, 1417– 1419. 31. Ashby, M.N. & Edwards, P.A. (1990) Elucidation of the defi- ciency in two yeast coenzyme Q mutants. Characterization of the structural gene encoding hexaprenyl pyrophosphate synthetase. J. Biol. Chem. 265, 13157–13164. 32. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 33. Maundrell, K. (1993) Thiamine-repressible expression vectors pREP and pRIP for fission yeast. Gene 123, 127–130. 34. Forsburg, S.L. & Sherman, D.A. (1997) General purpose tagging vectors for fission yeast. Gene 191, 191–195. 35. Ozoe, F., Kurokawa, R., Kobayashi, Y., Jeong, H., Tanaka, K., Sen, K., Nakagawa, T., Matsuda, H. & Kawamukai, M. (2002) The 14–3)3 proteins Rad24 and Rad25 negatively regulate Byr2 by affecting its localization in Schizosaccharomyces pombe. Mol. Cell. Biol. 22, 7105–7109. 36. Kawamukai, M., Gerst, J., Field, J., Riggs, M., Rodgers, L., Wigler, M. & Young, D. (1992) Genetic and biochemical analysis of the adenylyl cyclase-associated protein, cap. Schizosaccharo- myces pombe. Mol. Biol. Cell. 3, 167–180. 37. Moreno, S., Klar, A. & Nurse, P. (1991) Molecular genetic ana- lysis of fission yeast Schizosaccharomyces pombe. Methods Enzymol. 194, 795–823. 38. Rothstein, R.J. (1983) One-step gene disruption in yeast. Methods Enzymol. 101, 202–211. 39. Kainou, T., Okada, K., Suzuki, K., Nakagawa, T., Matsuda, H. & Kawamukai, M. (2001) Dimer formation of octaprenyl diphos- phate synthase (IspB) is essential for chain length determination of ubiquinone. J. Biol. Chem. 276, 7876–7883. 40. Okada, K., Kainou, T., Tanaka, K., Nakagawa, T., Matsuda, H. & Kawamukai, M. (1998) Molecular cloning and mutational analysis of the ddsA gene encoding decaprenyl diphosphate syn- thase from Gluconobacter suboxydans. Eur. J. Biochem. 255, 52–59. Ó FEBS 2003 Decaprenyl diphosphate synthase from fission yeast (Eur. J. Biochem. 270) 4121 . Fission yeast decaprenyl diphosphate synthase consists of Dps1 and the newly characterized Dlp1 protein in a novel heterotetrameric structure Ryoichi. T., Tanaka, K., Nakagawa, T., Matsuda, H. & Kawamukai, M. (1998) Molecular cloning and mutational analysis of the ddsA gene encoding decaprenyl diphosphate