Isolation and Characterization of the Gene (CLS1) Encoding Cardiolipin Synthase in Saccharomyces cerevisiae *

In eukaryotic cells, cardiolipin (CL) synthase catalyzes the final step in the synthesis of CL from phosphatidylglycerol and CDP-diacylglycerol. CL and its synthesis are localized predominantly to the mitochondrial inner membrane, and CL is generally thought to be an essential component of many mitochondrial processes. By using homology searches for genes potentially encoding phospholipid biosynthetic enzymes, we have cloned the gene (CLS1) encoding CL synthase in Saccharomyces cerevisiae. Overexpression of the CLS1 gene under its endogenous promoter or the inducible GAL1 promoter in yeast and expression of CLS1 in baculovirus-infected insect cells resulted in elevated CL synthase activity. Disruption of theCLS1 gene in a haploid yeast strain resulted in the loss of CL synthase activity, no detectable CL, a 5-fold elevation in phosphatidylglycerol levels, and lack of staining of mitochondria by a dye with high affinity for CL. Thecls1::TRP1 null mutant grew on both fermentable and non-fermentable carbon sources but more poorly on the latter. The level and activity of cytochrome c oxidase was normal, and a dye whose accumulation is dependent on membrane proton electrochemical potential effectively stained the mitochondria. These results definitively identify the gene encoding the CL synthase of yeast.

In eukaryotic cells, cardiolipin (CL) synthase catalyzes the final step in the synthesis of CL from phosphatidylglycerol and CDP-diacylglycerol. CL and its synthesis are localized predominantly to the mitochondrial inner membrane, and CL is generally thought to be an essential component of many mitochondrial processes. By using homology searches for genes potentially encoding phospholipid biosynthetic enzymes, we have cloned the gene (CLS1) encoding CL synthase in Saccharomyces cerevisiae. Overexpression of the CLS1 gene under its endogenous promoter or the inducible GAL1 promoter in yeast and expression of CLS1 in baculovirus-infected insect cells resulted in elevated CL synthase activity. Disruption of the CLS1 gene in a haploid yeast strain resulted in the loss of CL synthase activity, no detectable CL, a 5-fold elevation in phosphatidylglycerol levels, and lack of staining of mitochondria by a dye with high affinity for CL. The cls1::TRP1 null mutant grew on both fermentable and non-fermentable carbon sources but more poorly on the latter. The level and activity of cytochrome c oxidase was normal, and a dye whose accumulation is dependent on membrane proton electrochemical potential effectively stained the mitochondria. These results definitively identify the gene encoding the CL synthase of yeast.
In eukaryotic cells cardiolipin (CL) 1 and its synthesis are predominantly localized to the mitochondrial inner membrane (1)(2)(3). Due to its cellular distribution, CL is thought to be involved in critical cellular functions such as cytochrome c oxidase function (4), binding of matrix Ca 2ϩ (5), maintenance of mitochondrial membrane permeability (6,7), and mitochondrial protein import (8 -12). In addition to basic cell functions, CL has been postulated to play a role in disease processes such as lipid peroxidation observed in reperfusion after ischemic injury and in cellular aging (13,14). A vast amount of literature details biochemical approaches to determine the role of CL in cell processes (2,15); however, it has been difficult to correlate the in vitro experimental findings with in vivo cellular processes due to the absence of genetic evidence supporting CL involvement in these processes. The focus of the work in our laboratory is to characterize the genes and enzymes necessary for the biosynthesis of CL and the role this anionic lipid plays in mitochondrial function.
CL is synthesized by three sequential reactions ( Fig. 1) in all organisms and requires the biosynthetic intermediate, CDPdiacylglycerol (CDP-DAG). The committed and rate-limiting step in CL biosynthesis in Saccharomyces cerevisiae is catalyzed by phosphatidylglycerophosphate (PG-P) synthase (step 1) yielding PG-P which is rapidly dephosphorylated (16). In eukaryotic cells phosphatidylglycerol (PG) reacts in the final step (step 3) with another molecule of CDP-DAG to yield CL catalyzed by CL synthase (17)(18)(19). In Escherichia coli the synthesis of PG proceeds by an analogous set of reactions (20) except CL is synthesized by the condensation of two PG molecules with the release of glycerol (21).
We recently isolated from S. cerevisiae and characterized the gene (PGS1) encoding the first enzyme in this pathway (22). In this paper we report the isolation and characterization of the gene (CLS1) encoding the last enzyme in the pathway for CL biosynthesis. The gene was cloned and expressed in both yeast and baculovirus-infected insect cells and shown to encode CL synthase activity. Interruption of the CLS1 gene resulted in no detectable CL or CL synthase activity and elevated levels of PG. Mitochondrial function was not grossly perturbed as evidenced by growth on non-fermentable carbon sources and maintenance of mitochondrial membrane potential in vivo. This is the first detailed report definitively establishing the isolation of a eukaryotic gene encoding CL synthase activity. A preliminary report of this work has appeared (23).
Phos TM 530 were purchased from Boehringer Mannheim. All media for yeast growth and selection were from Bio 101 and Difco, and all media for bacterial growth were from Difco. Radiochemicals were obtained from Amersham Pharmacia Biotech and American Radiochemicals. The BCA kit for protein determination was from Pierce. The lambda phage clone PM-5829 was purchased from American Type Culture Collection. PG derived from E. coli was purchased from Matreya Inc. CDP-DAG (dioleolyl) was synthesized (24) and generously provided by Dr. George Carman (Rutgers University). All other phospholipids were from Sigma.
Strains, Plasmids, and Growth Conditions-A list of strains and plasmids used in this work is given in Table I. Methods for E. coli growth and selection were described previously (25). Strain ADC90/ pDD72 was used to make radiolabeled PG. Strain DH5␣ was used to propagate plasmids and was grown in LB medium (1% Bacto-tryptone, 0.5% Bacto-yeast extract, 1% NaCl, pH 7.4) at 37°C. Ampicillin (200 g/ml) was added to cultures carrying the appropriate plasmids. E. coli strain LE392 was used to propagate lambda phage (25). Methods of yeast growth and selection were described previously (29,30). YPD medium consisted of 1% Bacto-yeast extract, 2% Bacto-peptone, and 2% dextrose. In YPGly or YPGal medium, 2% glycerol or galactose, respectively, replaced dextrose as the carbon source. Complete synthetic media (CSM) were constituted as described previously (30) and contained the indicated carbon source (dextrose, galactose, or glycerol) at 2% (w/v). Yeast selection media contained the components of CSM except those noted for selection purposes (i.e. ϪTrp and Ϫuracil). All the above media were supplemented with 2% agar (yeast) or 1.5% agar (E. coli) for growth on plates. Yeast strains were grown at 30 or 37°C, and bacterial strains were grown at 30, 37, or 42°C.
DNA Manipulations-Methods for plasmid and genomic DNA preparation, restriction enzyme digestion, DNA ligation, E. coli and yeast transformation, DNA sequencing, and sequence analysis were carried out as described previously (22). For both analytical and preparative purposes, PCR was performed after optimizing conditions (31). Amplification of the CLS1 gene from yeast chromosomal DNA employed primers based on the open reading frame YDL142c 2 as follows: primer 1, 5Ј-CGCAAGCTTACAATGATTCAAATGGTGC-3Ј (the added HindIII site is underlined and the start codon for the open reading frame is in boldface), and primer 2, 5Ј-CGTTGTCTAGTGAAAGAGAGAG-3Ј (priming beginning 500 bp 3Ј of the stop codon). For Southern hybridization analysis DNA samples were digested with restriction endonuclease and prepared for hybridization with a digoxigenin-dUTP-labeled probe as described previously (22). Template DNA for labeling by random priming using the Genius TM 1 Kit and use in the above procedure was generated by HindIII and BamHI digestion of plasmid pYCLS10-1 which produced a fragment (982 kb) containing the CLS1 gene.
Preparation of Mitochondria-enriched Membrane Fraction-All cell fractionation procedures were performed at 4°C. Mitochondria-enriched membrane fractions were prepared as described previously (22). Briefly, yeast cells were grown to late log phase in CSM with auxotrophic selection when appropriate and either 2% glucose or 2% galactose as carbon source. Cells were pelleted and washed with homogenization buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 0.3 M sucrose, 10 mM 2-mercaptoethanol) and frozen at Ϫ80°C for later use. Cells were disrupted at 4°C by mechanical shearing using a Mini-BeadBeater TM 8 and silica/zirconium beads (Biospec Products). A crude mitochondrial pellet was obtained by centrifugation at 27,000 ϫ g and suspended in the above homogenization buffer.
Expression of the Yeast CLS1 Gene in Insect Cells-The yeast CLS1 gene was subcloned into plasmid pVL1392 for baculovirus expression. The HindIII-XhoI (1046 kb) fragment of plasmid pYCLS10-1 containing the CLS1 gene was subcloned into the HindIII-XhoI sites of pBluescript II KS, and the resulting plasmid was digested with BamHI to generate a fragment containing the CLS1 gene flanked by BamHI sites. Next, the BamHI fragment was subcloned into plasmid pVL1392, and orientation was verified for proper expression behind the polyhedrin promoter (P PH ) for the resulting plasmid pVCLS1B (containing P PH -CLS1).
Plasmid pVCLS1B was cotransfected, using the calcium phosphate technique, into the insect cell line, Sf9, with linearized Autographa californica genomic DNA (BaculoGold, PharMingen). The transfected cells were cultured in TNMFH medium supplemented with 10% fetal bovine serum for 5 days. Culture supernatants were harvested and used as a source of mixed population virus for the following transfection. Sf9 cells were infected with recombinant virus at a multiplicity of infection of 10 for 1 h and then incubated with TNMFH medium supplemented with 10% fetal bovine serum for 48 h at 27°C. After 48 h the cells were harvested, washed twice with 137 mM NaCl, 2.6 mM KCl, 8 mM Na 2 HPO 4 , 15 mM KH 2 PO 4 , pH 7.4, and resuspended in 5 mM Tris, pH 7.4, 150 mM NaCl. Control Sf9 cells without viral infection were prepared as above. Cells were frozen for subsequent CL synthase assays. After thawing on ice, cell suspensions were adjusted to 0.5-1 mg of protein per ml in a final concentration of 10 mM Tris-HCl, pH 9, containing 0.5 mM phenylmethanesulfonyl fluoride and lysed by sonication for 30 s. Cell debris was removed by centrifugation at 1500 ϫ g for 10 min. The supernatant was assayed for CL synthase activity as described below.
Synthesis of [ 32 P]PG-[ 32 P]PG was synthesized in vivo using E. coli strain ADC90/pDD72 which is deficient in phosphatidylethanolamine synthesis after a shift to growth at 42°C and only makes trace amounts of CL. From a saturated culture grown at 30°C, cells were inoculated at an 1:1000 dilution into 10 ml of M-56 medium (100 mM Tris, pH 7.4, 10 mM KCl, 0.4% NH 4 SO 4 w/v, 20 mM MgSO 4 , 0.3 mM K 2 HPO 4 , 0.4% vitamin-free casamino acids, 0.4% glucose) plus 400 mM sucrose and [ 32 P]orthophosphate (final concentration of 20 Ci/ml); the culture was incubated at 42°C overnight. The phospholipids were extracted as described previously (32) and spotted on a Silica Gel 60 thin layer chromatography plate along with 5 g of unlabeled CL. A one-dimensional thin layer chromatogram was developed in chloroform/methanol/ acetic acid (65:25:8, v/v). The 32 P-labeled PG was scraped from the plate, the silica gel extracted to isolate the lipids (33), and the extracted lipid was subjected to a second purification using the above one-dimensional system.
CL and PG-P Synthase Activities-CL synthase activity was measured by the conversion of [ 32 P]PG to CL as described previously (19). The assay was performed at 45°C and in the presence of 100 mM Tris-HCl, pH 9.0, 20 mM MgCl 2 , 0.1 mM CDP-DAG, 0.5 mM [ 32 P]PG (specific activity of 5,000 -10,000 cpm/nmol adjusted with E. coli PG as carrier), and 0.6 mM Triton X-100 in a total volume of 100 l. The reaction was terminated with 0.5 ml of 0.1 N HCl in methanol, and the lipids were extracted with 1 ml of chloroform and 1.5 ml of 1 M MgCl 2 . A 0.7-ml aliquot of the chloroform phase was removed and taken to dryness. The labeled phospholipids were dissolved in chloroform, separated in one dimension on boric acid-impregnated Silica Gel 60 thin layer plates using chloroform/methanol/water/ammonium hydroxide (60:37:5:3:1, v/v) as solvent (34), and quantified using a Packard Instant Imager. A unit of enzymatic activity is defined as the amount of enzyme that catalyzes the formation of 1 nmol of product per min under the assay conditions described above. Specific activity is defined as units/mg protein (determined using the BCA protein assay). PG-P synthase activity was measured as described previously (22).
Labeling and Analysis of Phospholipids-Cells grown in CSM plus 2% glucose with auxotrophic selection when appropriate were labeled with [ 32 P]orthophosphate for six generations, the phospholipids extracted in the presence of carrier lipid (30 g of equal ratio of phospholipids including phosphatidylcholine, phosphatidylinositol, phosphatidylethanolamine, phosphatidylserine, phosphatidic acid, PG, and CL) as described previously (22). Isolated radiolabeled phospholipids (ϳ10,000 cpm) were analyzed on HPTLC Silica Gel plates by twodimensional thin layer chromatography (35) with the first dimension developed from right to left (see Fig. 4) with chloroform/methanol/acetic acid (65:25:10, v/v) and the second dimension developed from bottom to top with chloroform/methanol/88% formic acid (65:25:10, v/v). The position of each spot was determined both by authentic standards (visu- alized by iodine staining and ninhydrin spray) as well as comparison with previous reports (35). The results were visualized by exposure to x-ray film, and for quantitative results the radioactivity was counted using a Packard Instant Imager. Alternatively, one-dimensional thin layer chromatography was used (34). Phospholipid content is expressed as mol % of total phospholipid based on the radiolabel in each spot and its respective molar phosphate content.
Determination of Cytochrome c Oxidase Levels-Activity of cytochrome c oxidase was measured at 22°C with an oxygen electrode in 0.25 M sucrose, 50 mM Tris-HCl buffer, pH 7.4, 50 mM KCl, 1 mM MgCl 2 , 5 mM ascorbate, 0.5 mM N,N,NЈ,NЈ-tetramethyl-p-phenylenediamine, 5 M FCCP (carbonyl cyanide-p-(trifluoromethoxy)phenylhydrazone), 2 g of antimycin A, and mitochondria (0.05 mg of protein per ml) isolated as described above. To determine cytochrome aa 3 content optical spectroscopy was performed with Hitachi U-3000 spectrophotometer at 22°C using mitochondria (0.1 to 0.15 mg of protein per ml) in 0.25 M sucrose, 50 mM Tris-HCl buffer, pH 7.4, 1 mM MgCl 2 . Either solid dithionite or 10 M ferricyanide was added to either reduce or oxidize the cytochromes, respectively, and the visible difference spectrum (reduced versus oxidized) was determined. A ⌬⑀ red-ox ϭ 27 mM Ϫ1 cm Ϫ1 (36) for the peak at 605 nm (minus the differential absorbency at 630 nm) was used to calculate the cytochrome aa 3 content.
Fluorescence Microscopy-Exponentially growing strains DL1 and YCD4 in YPD or strains YPH98 and YCD2 in the CSM with 2% galactose were prepared for fluorescence microscopy by incubation with either 50 nM 10-N-nonyl acridine orange for 15 min (37) or 20 M DASPMI for 20 min (38) at room temperature; in the latter case addition of 50 M FCCP prior to addition of DASPMI was used to collapse the mitochondrial membrane potential. Similarly, strains DL1 and YCD4 grown in YPD were treated with DAPI at a final concentration of 1 g per ml. Cells were viewed with an Olympus BX60 epifluorescence microscope equipped with a 100-watt HBO lamp, standard fluorescein isothiocyanate filter set, and a 100 ϫ fluorite oil immersion objective. Images were captured with an Optronics DEI-750 video camera and manipulated in Adobe Photoshop version 3.0.

RESULTS
Isolation of the CLS1 Gene-Based on the deduced amino acid sequence of the products of pgsA genes (encoding PG-P synthase) from E. coli (39) and Rhodobacter sphaeroides (40), a computer-based homology search of the S. cerevisiae genome data base yielded open reading frame YDL142c on Chromosome IV 2 with a derived amino acid sequence showing 26 and 29% identity and 53 and 54% similarity, respectively, to the bacterial enzymes (Fig. 2). The CLS1 gene is defined by an 849-bp open reading frame encoding a protein of 283 amino acids with a predicted molecular mass of 31,999 Da. Although sequence comparison suggested the gene encodes a PG-P syn-thase, subsequent biochemical and genetic analyses established the gene to encode CL synthase (see below). The CLS1 gene product also showed weaker homology (20% or less identity and less than 45% similarity) with other phosphatidyltransferases from bacteria and yeast. The characteristic CDPalcohol/phosphatidyltransferase signature (41) is present between positions 106 and 136 of the CL synthase. This motif is present in the PG-P synthases shown in Fig. 2 and is also found in phosphatidylserine synthases from Bacillus subtilis and S. cerevisiae, in phosphatidylinositol synthase from S. cerevisiae, and in numerous other bacterial PG-P synthases (40); this motif is not found in either the phosphatidylserine synthase (42) or CL synthase (43) of E. coli.
No sequence in the 5Ј-untranslated region matched the UA-S INO (upstream activation sequence responsive to repression by inositol), 5Ј-CATGTGAAAT-3Ј (44,45). This is consistent with absence of regulation of the level of CL synthase activity by inositol in the growth media (19,46). Analysis of the predicted amino acid sequence by PSORT (47) indicated targeting of this protein to the inner mitochondrial membrane with a certainty of 0.76 versus the 0.56 for peroxisomes and 0.44 for plasma membranes. A M r of 28,000 based on sodium dodecyl sulfate-polyacrylamide gel electrophoresis has been reported for the CL synthase (48) which would be consistent with a mitochondrial targeting sequence of about 36 amino acids and place the beginning of the putative mature protein very close to the beginning of its homology with bacterial PG-P synthases.
The CLS1 gene was obtained from DNA isolated from strain DL1 using PCR and primers which added a 5Ј HindIII site. The PCR product was digested with HindIII and NsiI (110 bp 3Ј of the stop codon) and subcloned into the HindIII and PstI sites of the plasmid pUC19. Subsequently, the PCR-generated CLS1 gene was released from plasmid pUC19 by digestion with Hin-dIII and BamHI (vector site) and transferred to the HindIII and BamHI sites of plasmid pYES2 generating plasmid pY-CLS10-1 which placed the CLS1 gene under the regulation of the GAL1 promoter (P GAL1 -CLS1). The PCR product was transferred to plasmid pBluescript II KS in the same manner, and its DNA sequence was confirmed by automated DNA sequencing. The resulting DNA sequence matched that of open reading frame YDL142c at all nucleotides with the exception of the third position of codon 198 of the CLS1 gene which was not  (27). The N-terminal HindIII-EcoRI fragment (Ϫ667 bp to ϩ574 bp) from plasmid pYCLS-1 was subcloned into plasmid pUC19. Next, the SnaBI-BamHI fragment (1,608 bp) of plasmid pRS304 containing the TRP1 gene was ligated between the HpaI (ϩ89 bp) and BclI (ϩ421 bp) sites, respectively, within the N-terminal half of the CLS1 gene. The cls1::TRP1 gene was released from the above plasmid with HindIII and EcoRI (both outside of the CLS1 gene) digestion and transformed into the haploid yeast strain YPH98. Transformants were selected on CSM-Trp plus 2% glucose medium resulting in haploid null mutant strain YCD2 (cls1::TRP1).
Genomic DNA from wild type strain YPH98 and cls1 null mutant YCD2 was isolated and either subjected to PCR amplification using primers 1 and 2 or digested with NcoI and BamHI for Southern hybridization analysis. The products of PCR amplification (Fig. 3A) showed the predicted size for the wild type CLS1 gene-containing fragment (1384 bp) of strain YPH98 and the predicted size (2992 bp) for the interrupted gene-containing fragment of strain YCD2. A labeled probe specific for the CLS1 gene was generated using the HindIII-BamHI insert of the plasmid pYCLS-1 as template. In strain YPH98 (Fig. 3B), a band consistent with the predicted 3,903-bp fragment for the wild type allele hybridized with the probe, and in strain YCD2, a band consistent with the predicted 5,188-bp fragment for the disrupted allele hybridized with the probe.
Gross phenotypic characterization of strain YCD2 showed only small differences in growth properties compared with wild type cells. Overnight cultures of strains YCD2 and YPH98 were inoculated into rich YP medium, CSM or CSM-Trp with either 2% glucose or 2% glucose plus 0.2% glycerol or 0.2% galactose as carbon source. To induce the appropriate metabolic enzymes, the cells were allowed to grow for 4 h and subsequently with serial dilutions plated on the following agar plates: YPG; YPGal; YPGly; CSM or CMS-Trp with either 2% glucose, 2% glycerol, or 2% galactose as carbon source. The cells were grown at both 30 and 37°C. Strain YCD2 grew well on all carbon sources at all temperatures tested and therefore was not respiratory-deficient. In order to test whether the null cls1 mutant exhibited the petite lethal phenotype (49) (i.e. lack of growth in the presence of ethidium bromide which generates a high frequency of deletions in mitochondrial DNA), a duplicate set of plates (YPD and CSM or CSM-Trp with 2% glucose) containing 25 mM ethidium bromide was seeded. A high level of colony formation by both the mutant and wild type parental strain in the presence of 25 mM ethidium bromide ruled out incompatibility with mutations in mitochondrial DNA (petite lethal phenotype); pgs1 mutants which lack both PG and CL display the petite lethal phenotype (22,49). In liquid culture strain YCD2 grew slower and reached stationary phase at consistently lower cell densities when grown on a non-fermentable carbon source as compared with strain YPH98 but did not show other difference in growth properties compared with its parental strain YPH98.
CLS1 Synthase Activity-Mitochondria-enriched fractions were prepared as described under "Experimental Procedures" of late log phase cultures of the strains YPH98 and YCD2 grown in CSM or strain YCD2 transformed with either plasmid pYCLS-1 or pYCLS10-1 grown in CSM-uracil; either glucose or galactose was used as carbon source. Expression of the CLS1 gene from plasmid pYCLS10 -1 should be induced by galactose and repressed by glucose due to transcription under GAL1 promoter control.
CL synthase specific activity of extracts made from strain YPH98 grown in either galactose (1.3 Ϯ 0.2 units/mg) or glucose (1.0 Ϯ 0.2 units/mg) as carbon source was about the same. Extracts of strain YCD2, irrespective of carbon source, showed no detectable activity even when assayed at twice the protein concentration of wild type strains; the lower limit of detection was on the order of 0.05 to 0.1 units/mg. CL synthase activity was elevated about 9-fold (8.8 Ϯ 0.8 units/mg) over wild type levels in extracts from strain YCD2 grown in glucose and carrying multiple copies (plasmid pYCLS-1) of the CLS1 gene under control of its native promoter. Expression of CLS1 under GAL1 regulation from a multicopy plasmid (pYCLS10-1) showed the expected overexpression of CL synthase activity (82 Ϯ 3 units/mg) when cells were grown in galactose as compared with cells grown in glucose (1.0 Ϯ 0.1 units/mg). These results are consistent with the CLS1 gene encoding over 90% if not all of the CL synthase of yeast.
To rule out the possibility that the mutation resulted in a general lack of anionic phospholipid biosynthetic capability, the level of PG-P synthase activity was measured. The specific activity of PG-P synthase in glucose grown cells was reduced in the cls1 null strain (0.17 Ϯ 0.02 units/mg) by about 35% relative to the wild type parent strain (0.26 Ϯ 0.02 units/mg). This difference is well within the range of specific activities for this enzyme as a function of growth conditions and stage of growth on glucose (22).
Expression of the Yeast CLS1 Gene Using Baculovirus-Plasmid pVL1392 (P PH -CLS1) was introduced into insect cell line Sf9 by transfection as described under "Experimental Procedures." The transfected cell lysates were greatly enriched in CL synthase activity (13 Ϯ 2 units/mg) as compared with the uninfected cell lysates (0.8 Ϯ 0.1 units/mg). This level of overexpressed synthase activity compares favorably with the specific activity of the synthase in the enriched mitochondrial fraction from cells in which CL synthase was amplified by multiple copies of the CLS1 gene (see above). This amplified activity in Sf9 cells is definitive proof that the CLS1 gene encodes a CL synthase and does not encode a regulator of either gene expression or mitochondrial biogenesis.
Phospholipid Composition-Strains YPH98 and YCD2 were labeled with [ 32 P]orthophosphate and analyzed for phospholipid composition as described under "Experimental Procedures" (Fig. 4 and Table II). The wild type strain showed the expected complement of phospholipids including easily detectable CL with trace amounts of PG (Fig. 4, A, C, and E). The cls1 null mutant lacked detectable CL and had elevated levels (about 5-fold over wild type) of PG (Fig. 4, B, D, and E). In addition the mutant showed a new radiolabeled component near the front of all solvent systems (arrow in Fig. 4, D and E). This material which constituted about 0.2% of the total radiolabeled material may be a higher order acylated derivative of PG which only becomes detectable when PG accumulates to abnormal levels. There were no large changes in the levels of other phospholipids in the mutant as compared with the wild type strain (Table II). YCD2 grown in galactose in which there is no catabolite repression of mitochondrial functions showed no detectable CL and a similar accumulation of PG (data not shown); introduction of plasmid pYCLS-1 into strain YCD2 returned the phospholipid composition to wild type (data not shown). These results further confirm the identity of the gene product of open reading YDL142c as the CL synthase.
10-N-Nonyl acridine orange is a lipophilic dye that has been shown to have an affinity for CL about 30-fold higher and a resulting red fluorescence emission yield 5-10-fold higher than for other anionic phospholipids such as phosphatidylserine or phosphatidylinositol (37). It has been used to specifically visualize mitochondria in living cells using fluorescence microscopy and for quantification of CL in mitochondria of whole cells (1). Wild type cells when treated with this dye clearly show fluorescent structures (Fig. 5, B and F) which have the morphology of mitochondria. Similar structures were barely detectable in the cls1 null mutant (Fig. 5D) consistent with the lack of CL in the mitochondria. Cells grown in either glucose (not shown) or galactose showed similar results. In the pgs1 null mutant (YCD4), which lacks both PG and CL (22), no distinct struc- tures could be seen (Fig. 5H), suggesting that the faint fluorescence seen in the cls1 null mutant was due to the high PG content of the mitochondria. These results are consistent with the predicted specificity of this dye for CL over PG and provide the first in vivo evidence that this dye can be used to identify the distribution of CL in vivo with a high level of confidence.
Cytochrome c Oxidase Activity and Mitochondrial Function-Null mutants in pgs1 lacking both PG and CL cannot grow on non-fermentable carbon sources (22) and have only trace levels of cytochrome c oxidase content or activity but appear to be normal in the b cytochromes (50). In contrast, both cytochrome c oxidase activity (1120 Ϯ 79 and 1120 Ϯ 5 nanogram atoms of oxygen per min per mg of protein, respectively) or cytochrome aa 3 content (0.37 Ϯ 0.04 and 0.33 Ϯ 0.03 nmol per mg of protein, respectively) were the same in wild type and cls1 null mutant strains (grown in YPGal) consistent with the ability of the mutant to grow on non-fermentable carbon sources. Mitochondria have a high proton electrochemical potential (positive outward) and therefore accumulate the fluorescent dye DASPMI that has been used as a specific marker for mitochondria in living cells (38). Both wild type cells (Fig. 6, B and C) and the cls1 null mutant (Fig. 6, E and F) accumulate this dye in a similar manner suggesting that the potential across the mitochondrial membrane of the mutant is not grossly compromised; no fluorescence was observed with either strain if the uncoupler FCCP was added to cells prior to addition of DASPMI. With this dye the transfer of filamentous mitochondrial structures from mother to daughter cells could also be observed (Fig. 7) suggesting normal mitochondrial segregation during cell division.
On the other hand no distinct structures were seen with the  4. Phospholipid analysis of the wild type and cls1 null strain YCD2. Strains YPH98 (CLS1) (A and C) and YCD2 (cls1::TRP1) (B and D) were grown to late log phase in the presence of [ 32 P]orthophosphate. Total membrane phospholipid was extracted and separated on plates that were exposed to x-ray film as described under "Experimental Procedures" for either two-dimensional (A-D) or one-dimensional (E) thin layer chromatography. A and B were short exposures (2 h) and C and D were long exposures (12 h) of the same plates. The arrows in C-E denote a unique spot in strain YCD2 that is absent in the wild type strain YPH98. Standard phospholipids were used to assign the identity of each spot: CL, cardiolipin; PA, phosphatidic acid; PG, phosphatidylglycerol; PE, phosphatidylethanolamine; PS, phosphatidylserine; PI, phosphatidylinositol; PC, phosphatidylcholine. pgs1 null mutant (not shown) but only a dispersed DASPMI fluorescence that was not stable and disappeared rapidly. When cells were stained with DAPI for DNA (Fig. 8), wild type cells showed a characteristic focused fluorescence for the nucleus and strings of beads for the mitochondria. The pgs1 mutant strain only exhibited what appears to be a single condensed structure and no evidence of elongated mitochondria. Clearly this mutant still has mitochondrial DNA because its mutant characteristics can be complemented by introduction of a plasmid-borne copy of the PGS1 gene under the control of its own promoter (49). 3 These results would indicate a very low mitochondrial membrane potential and/or highly irregular mitochondrial structure in cells lacking both PG and CL consistent with gross defects in mitochondrial function in this mutant. DISCUSSION CL is an anionic phospholipid thought to be involved in numerous cellular processes. To gain further insight into its role in the cell, we undertook the isolation and characterization of genes involved in CL biosynthesis in S. cerevisiae. We have presented definitive evidence that open reading frame YDL142c (CLS1) encodes CL synthase. Overexpression of the gene in yeast resulted in the expected increase in CL synthase activity as did expression of the gene in a heterologous insect cell system. The null mutant had no detectable CL or CL synthase activity and elevated levels of PG.
The yeast CLS1 gene product shows significant homology with the PG-P synthases of prokaryotic organisms (40) and little or no homology with the CL synthase of E. coli (43). Although initially surprising, this result is still consistent with the enzymatic reactions catalyzed by PG-P and CL synthases (51)(52)(53)(54) which are both CDP-DAG-dependent phosphatidyltransferases, whereas the prokaryotic CL synthase does not utilize this liponucleotide. The PGS1 gene encoding the yeast PG-P synthase has been identified (22), and its product shows little homology with either bacterial PG-P synthases or any CL synthase. The PGS1 gene was originally named PEL1 based on 3   the petite lethal phenotype of point and null mutants in this gene (49,50). Subsequently, it also was named PSS2 based on the sequence homology of its gene product with the phosphatidylserine synthase (pssA gene product) of E. coli (55,56). The definitive identification of the PGS1 gene product as a PG-P synthase was established based on biochemical characterization of overexpression of the isolated gene in both yeast and E. coli and a more detailed biochemical characterization of mutants null in PGS1 (22). A cautionary note resulting from these unexpected homology relationships is the unreliability of assigning gene product function solely on the basis of sequence homology or even the phenotype of mutants without biochemical verification of a direct gene product relationship. The lack of homology at the protein level and at the mechanistic level between prokaryotic and eukaryotic organisms in the PG/CL biosynthetic pathway might make this required pathway in bacteria an important target for development of antibiotics.
During the preparation of this manuscript two reports (57,58) were published that putatively assigned CL synthase as the product of the CLS1 gene based solely on characterization of cls1 null mutants. This assignment is consistent with lack of detectable CL synthase activity (57,58), lack of detectable CL (57,58), and accumulation of PG in mutants null in CLS1 (58). However, no definitive evidence was presented to rule out a regulatory role rather than a biosynthetic role for the CLS1 gene product in CL synthesis or the possibility that mutants in CLS1 are defective in mitochondrial processes necessary to support CL synthase function and/or synthesis. For instance mutants in cytochrome c oxidase subunits show up to a 60% reduction in CL synthase activity (48). Fortunately, the gene locus assigned by these papers as encoding CL synthase was correct as established by the results reported here.
Disruption of the CLS1 gene resulted in an absence of CL and a 5-fold increase in total PG content. This is the expected result based on the pathway for CL synthesis. The more interesting result was that the absence of CL was neither lethal to the cell nor did it result in gross mitochondrial dysfunction particularly in oxygen-dependent energy metabolism, although qualitatively cells grew poorer on non-fermentable carbon sources as was previously reported for this mutant (57,58). In addition gross mitochondrial morphology and the migration of mitochondria to daughter cells were not perturbed as indicated by staining with fluorescent dyes. Previous results had shown that cytochrome c, aa 3 , and b levels were normal in the null mutant (58) consistent with our findings which also showed that in vivo mitochondrial membrane potential was not significantly perturbed based on accumulation of DASPMI by mitochondria. Cells lacking mitochondrial DNA and hence completely devoid of functions necessary for oxidative phosphorylation can also tolerate the cls1 null allele, and therefore this mutant does not exhibit a petite lethal phenotype. These properties are in marked contrast to those of pgs1 null mutants that lack both PG and CL (22). These latter mutants cannot grow on non-fermentable carbon sources (22,50) and have very low levels of cytochrome c oxidase (50). These properties are consistent with the lack of staining in this mutant by both DASPMI and 10-N-nonyl acridine orange. In addition pgs1 null mutants exhibit the petite lethal phenotype (22,49) that would indicate a more extensive mitochondrial dysfunction than simply loss of the ability to carry out oxidative phosphorylation. The lack of detection of filamentous mitochondrial structures with DAPI also suggests a disruption of mitochondrial structure. Chinese hamster ovary cell mutants with reduced levels of both PG and CL are similar in their properties to the yeast pgs1 mutants and display significant alterations in mitochondrial morphology (59,60). These results taken to-gether strongly suggest that PG can substitute for CL in all essential functions in the mitochondria, but complete loss of both of these anionic phospholipids results in severe mitochondrial dysfunction.
Yeast cls1 mutants are somewhat compromised metabolically when grown on non-fermentable carbon sources, and in some genetic backgrounds have been reported to be temperature-sensitive for growth (57) as are pgs1 null mutants (22). It is likely that CL does not play an important role under well supported conditions, but in a nutrient-poor environment or under conditions of cell stress, CL may indeed be necessary for mitochondrial function or cell vitality. This is similar to the finding in E. coli where the absence of PG but not CL is lethal (32,61). However, E. coli lacking CL are reduced in their ability to survive under harsh conditions such as cycles of freeze-thaw or in long term arrest in stationary phase (62).
The availability of mutants in PG and CL biosynthesis in E. coli and construction of strains in which the levels of these lipids can be regulated have been effectively used to define specific roles for anionic lipids in membrane structure, DNA replication, and translocation of proteins across membranes (20). PG and/or CL may also be involved in protein import into the mitochondria (8 -12). In particular cytochrome c oxidase, which is very low in pgs1 mutants, may require either PG or CL for assembly and/or stability. Now that the genes encoding the enzymes for the synthesis of these two lipids in yeast have been identified, the requirements for PG and CL in mitochondrial function at the molecular level can be more precisely defined.