Synthetic lethal interaction of the mitochondrial phosphatidylethanolamine and cardiolipin biosynthetic pathways in Saccharomyces cerevisiae.

Saccharomyces cerevisiae mitochondria contain enzymes required for synthesis of the phospholipids cardiolipin (CL) and phosphatidylethanolamine (PE), which are enriched in mitochondrial membranes. Previous studies indicated that PE may compensate for the lack of CL, and vice versa. These data suggest that PE and CL have overlapping functions and that the absence of both lipids may be lethal. To address this hypothesis, we determined whether the crd1delta mutant, which lacks CL, was viable in genetic backgrounds in which PE synthesis was genetically blocked. Deletion of the mitochondrial PE pathway gene PSD1 was synthetically lethal with the crd1delta mutant, whereas deletion of the Golgi and endoplasmic reticulum pathway genes PSD2 and DPL1 did not result in synthetic lethality. A 20-fold reduction in phosphatidylcholine did not affect the growth of crd1delta cells. Supplementation with ethanolamine, which led to increased PE synthesis, or with propanolamine, which led to synthesis of the novel phospholipid phosphatidylpropanolamine, failed to rescue the synthetic lethality of the crd1delta psd1delta cells. These results suggest that mitochondrial biosynthesis of PE is essential for the viability of yeast mutants lacking CL.

The phospholipid composition of the mitochondrial membrane is unique in that it is highly enriched in cardiolipin (CL) 2 and phosphatidylethanolamine (PE) (1,2). CL is a dimeric glycerophospholipid that is synthesized exclusively in mitochondria and plays an important role in oxidative phosphorylation and mitochondrial membrane biogenesis (3). In contrast, PE biosynthesis occurs via multiple pathways (4). PE is commonly present in all subcellular membranes, although PE levels are highest in the mitochondrial membrane (2). PE and CL have similar physical properties in that they have a propensity toward the formation of nonbilayer, inverted hexagonal (H II ) phase structures (5,6). The local, transient formation of nonbilayer structures is thought to play an important role in vital cellular processes, such as vesicle formation, vesicle-mediated protein trafficking, and membrane fusion (7). In addition, nonbilayer lipids affect integration of proteins into the membrane, their lateral movement within the membrane, and the function and folding of certain integral membrane proteins (8).
In the yeast Saccharomyces cerevisiae, phospholipid biosynthesis is compartmentalized in various subcellular organelles, including the Golgi body, endoplasmic reticulum, and mitochondria (Fig. 1). CL biosynthesis occurs in three steps, all catalyzed by enzymes present in the mitochondria. The first step, catalyzed by phosphatidylglycerol phosphate (PGP) synthase, is the synthesis of PGP from CDP-diacylglycerol and glycerol 3-phosphate. PGP phosphatase dephosphorylates PGP to phosphatidylglycerol. In the final step, CL synthase catalyzes the formation of CL from phosphatidylglycerol and CDP-diacylglycerol (9,10). PGP synthase and CL synthase have been characterized in yeast, and the genes encoding these enzymes, PGS1 (11,12) and CRD1 (13)(14)(15), have been identified. The biosynthesis of aminoglycerophospholipids occurs in the endoplasmic reticulum, vacuole/Golgi body, and mitochondrial compartments (Fig. 1). PE biosynthesis is accomplished by two de novo pathways involving decarboxylation of phosphatidylserine (PS) by PS decarboxylase, including Psd1p-catalyzed PE synthesis in mitochondria and Psd2p-catalyzed PE formation in the vacuole/Golgi body (16,17). In addition, PE can be synthesized from ethanolamine (Etn) via the Kennedy pathway (Fig. 1). The Kennedy pathway is also linked to sphingolipid catabolism through a reaction catalyzed by dihydrosphingosine phosphate lyase (Dpl1p). This enzyme cleaves the phosphorylated sphingoid base to a long chain aldehyde and ethanolamine phosphate (18), which is then incorporated into PE via the Kennedy pathway (19). The presence of CL and PE biosynthetic machinery in mitochondria suggests that these phospholipids are required for mitochondrial function.
Several published studies are consistent with the possibility that CL and PE have redundant functions and that each can compensate for the loss of the other. The growth phenotype of crd1⌬ cells, which lack CL synthase and CL, is characterized by temperature sensitivity at 37°C, reduced growth in nonfermentable media, and increased frequency of petite formation (20). Depletion of PE in psd1⌬ cells produces a similar phenotype (16). Furthermore, the Escherichia coli AD93 strain, which lacks PE, has increased levels of CL (21), and the absence of CL in yeast crd1⌬ mutant cells results in increased PE (22). The null mutants pgs1⌬, crd1⌬, psd1⌬, psd2⌬, pem1⌬, pem2⌬, and dpl1⌬ are viable in complex and synthetic media in the absence of external supplementation with Etn or choline, consistent with the redundancy of phospholipid functions in S. cerevisiae.
Based on the observed similarities in the phenotypes of crd1⌬ and psd1⌬ cells, the physical properties of PE and CL, and the compensatory increases of PE in CL-deficient mutants and CL in PE-deficient mutants, we hypothesized that cells lacking both PE and CL are inviable. In support of this hypothesis, we determined that crd1⌬ is synthetically lethal with psd1⌬ but not with mutants in the nonmitochondrial pathways for PE synthesis. Supplementation with ethanolamine or propanolamine failed to rescue the lethality of the crd1⌬psd1⌬ double mutant. These data indicate that the combined loss of synthesis of CL and PE from mitochondria is lethal to yeast cells.

Materials
Glucose, yeast extract, and peptone were purchased from Difco. Galactose, raffinose, mannitol, G418, and Etn were from Sigma. Propanolamine (Prn) was purchased from Aldrich. 32 P was from PerkinElmer Life Sciences. Zymolyase was obtained from ICN Biomedicals, Inc. (Aurora, OH). TLC plates (LK5 silica gel 150 A and K6 silica gel 60 A) were purchased from Whatman, Inc. (Clifton, NJ). Chloroform, ethanol, methanol, glycerol, boric acid, and triethylamine were from Fisher. Alkaline phosphatase substrate kits, precast polyacrylamide gels (4 -20%), polyvinylidene difluoride membranes, and protein assay reagent were obtained from Bio-Rad. Prestained protein molecular weight markers were from Fermentas (Hanover, MD); protease inhibitor mixture, Complete Mini, was from Roche Applied Science; mouse anti-V5 antibody, blasticidin, and the pYES6/CT yeast expression system were purchased from Invitrogen. Wizard Miniprep DNA purification system and alkaline phosphatase-conjugated anti-mouse IgG were purchased from Promega (Madison, WI). All other chemicals used were reagent grade or better.

Methods
Yeast Strains and Growth Media-The yeast S. cerevisiae strains and plasmids used in this work are listed in TABLE ONE. Yeast strains were grown aerobically at 30°C. Complex medium (YPD) contained yeast extract (1% w/v), peptone (2% w/v), and glucose (2% w/v). Complete synthetic medium (CSM) and sporulation medium were prepared as described (23). Etn and Prn were prepared as 0.5 M stocks (pH 6 -7) that were filter-sterilized. The expression of V5-tagged Crd1p was induced from pYES6/CRD1 by growth in YP induction (YPI) medium, which contained raffinose (1%) and galactose (2%) as carbon sources. The crd1⌬ cells transformed with pYES6 plasmids were grown in the presence of 50 g/ml blasticidin to prevent loss of the plasmids. Liquid cultures were inoculated to an A 550 of 0.1 from precultures grown to the stationary phase unless otherwise indicated.
Construction of Deletion Strains and Plasmid-The crd1⌬ deletion was generated by homologous recombination in the BY4742 wild-type strain via transformation with a linearized pUC19Ј plasmid containing a crd1⌬::URA3 disruption sequence (13). Transformants were selected on Ura Ϫ plates. The crd1⌬::URA3 disruption was confirmed by PCR, Southern blot analysis, and phospholipid profile of the transformants.
The primers for PCR were forward primer (5Ј-AAAAGCTTCTGG-TAGCATAGTTTGGTCC-3Ј) and reverse primer (5Ј-ACGGATC-CTGATGATATTGCATTCAGCCG-3Ј) (13). Yeast deletion strains including BY4741 psd1⌬::KanMX4, BY4741 psd2⌬::KanMX4, BY4741 dpl1⌬::KanMX4, BY4741 pem1⌬::KanMX4, and BY4741 pem2⌬::KanMX4 were obtained from the yeast deletion collection (Invitrogen). The genotypes of these strains were confirmed by growth phenotypes and phospholipid profiles. Double mutant strains were created by sporulation of the diploids followed by tetrad dissection on either YPD medium or glucose-containing CSM medium with 2 mM or 20 mM Prn or Etn supplementation. The identities of all viable double mutant strains were also confirmed by their genotypes and characterization of their phospholipid profiles. The pYES6/CRD1 plasmid was constructed by inserting the CRD1 open reading frame sequence into pYES6/CT at the BamHI/XbaI site, upstream of the V5 tag and downstream from the GAL1 promoter. Primers were designed based on the SGD CRD1 gene sequence: forward primer (5Ј-AATTGGATCCAT-GATTCAAATGGTGCCC-3Ј) and reverse primer (5Ј-CTTTTCTA-GAAGGATCGCAATTATACAATT-3Ј). Restriction endonuclease sites BamHI and XbaI as indicated by the underlines, were added to the forward and reverse primers to facilitate subsequent DNA manipulations. The CRD1 stop codon was deleted in the reverse primer. The insertion was confirmed by sequencing, and functionality of V5-tagged Crd1p was checked by transforming the crd1⌬ mutant with the pYES6/ CRD1 plasmid. Growth of crd1⌬ ϩ pYES6/CRD1 in inducing conditions resulted in expression of the Crd1-V5 protein, which is ϳ30 kDa in size. V5-tagged Crd1p was enriched in mitochondria and restored wildtype CL levels in the CL-deficient crd1⌬ mutant.
Steady State Total Cellular Phospholipid Determination-Phospholipid extraction, separation, and quantification were carried out as described (24). Briefly, cells were grown at 30°C in YPD or CSM media from a starting A 550 of 0.1. Immediately after inoculation, cultures were supplemented with 10 Ci 32 P ml Ϫ1 and allowed to grow for several generations to achieve steady state labeling. For Etn and Prn incorporation experiments, YPD or CSM medium was supplemented with either 2 or 20 mM Etn or Prn prior to inoculation. Radiolabeled phospholipids were separated by one-dimensional TLC (24) or two-dimensional TLC (25). 32 P in individual phospholipids was visualized by phosphorimaging and quantified by ImageQuant software (Amersham Biosciences). Incorporation of radiolabel ( 32 P) into individual phospholipid is expressed as a percentage of radiolabel incorporated into total phospholipids.
Mitochondrial Phospholipid Determination-Cells were grown in CSM in the absence or presence of 2 mM Prn. Mitochondria were isolated from 5-6 g of cell pellet as described by Daum et al. (26). Total mitochondrial phospholipids were extracted and purified as described (27). Isolated total phospholipids were applied to silica gel 60 plates and separated by two-dimensional TLC (25). The developed TLC plates were dried and exposed to iodine vapor to visualize phospholipid spots.
Western Blot Analyses-Protein extraction was carried out as per the Invitrogen product manual. The extraction buffer contained 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5% deoxycholate, 0.1% SDS, 1% Nonidet P-40, and protein inhibitor mixture at the concentration of 1 tablet/10 ml of buffer. Protein quantification was carried out using the Bio-Rad protein assay reagent. Protein extracts (50 g) were separated by SDS-PAGE using precast (4 -20%) gradient polyacrylamide gels. Protein from the separating gel was electroblotted onto a polyvinylidene difluoride membrane. The membrane was incubated with the monoclonal anti-V5 antibodies (1:3000 dilution; Invitrogen) and then washed and incubated with alkaline phosphatase-linked secondary antibodies. The protein bands were detected using a 5-bromo-4-chloro-3-indolyl phosphate-nitro blue tetrazolium color development kit.
Other Procedures-Standard molecular techniques were used in this study. Plasmid purification from E. coli was performed using the Wizard Miniprep DNA Purification system. Transformations of E. coli cells were performed using a BTX electroporation system. Yeast transformations were carried out by the lithium acetate method (28).

RESULTS
The crd1⌬ and psd1⌬ Mutants Are Synthetically Lethal-To test the hypothesis that crd1⌬ and psd1⌬ are synthetically lethal, we constructed a crd1⌬::URA3 strain in MAT␣ background and mated it with a MATa psd1⌬::KanMX4 strain. The resulting heterozygous diploids were sporulated, and meiotic tetrad analysis was carried out. In 63 tetrads dissected, 54 tetrads contained less than four viable spores, and no viable crd1⌬psd1⌬ spores (Ura ϩ Kan R ) were recovered (TABLE TWO), indicating that the crd1⌬psd1⌬ double mutant is synthetically lethal. The lethality was not due to defective spore germination, since microscopic analysis indicated that tiny colonies of no more than a few hundred cells formed from the spores that were deduced to be crd1⌬psd1⌬ (not shown). In order to demonstrate a conditional synthetic lethal relationship in the crd1⌬psd1⌬ double mutant, the heterozygous diploid BY4741/2 (CRD1/crd1⌬::URA3 PSD1/psd1⌬::KanMX4) was transformed with the pYES6/CRD1 plasmid, from which the expression of V5-tagged CRD1 is controlled by the GAL1 promoter, and meiotic tetrad analysis of the sporulated diploid was carried out. We chose the pYES6/CT expression system, because it allowed controlled expression of epitope-tagged Crd1p, and thus growth phenotype could be correlated with expression of the Crd1p protein. Viable crd1⌬psd1⌬ spores were recovered on YPI medium in which Crd1p is induced (TABLE  TWO). These cells exhibited normal growth on YPI ( Fig. 2A) and reduced, but discernible, growth on YPD (Fig. 2B). Growth on YPD medium could be attributed to low level expression of Crd1-V5 from the leaky GAL1 promoter (Fig. 2C), which was sufficient to synthesize  almost 70% of wild-type levels of CL (Fig. 2D). However, the growth phenotype in liquid YPD medium of the crd1⌬psd1⌬ cells containing the pYES6/CRD1 plasmid was similar to crd1⌬psd1⌬ cells containing the plasmid that were cultured in liquid YPI medium (Fig. 2E). Phenotypic differences between culturing in solid or liquid medium have been previously observed in the psd1⌬ and crd1⌬ mutants. The temperaturesensitive phenotype of psd1⌬ cells is apparent at 37°C in liquid medium but is not observed in solid medium at 37°C (25), whereas the reverse is true for the crd1⌬ mutant (29). Disruption of Nonmitochondrial Pathways for PE Biosynthesis Does Not Affect the Growth of crd1⌬ Cells-Although mitochondrial PSD1 is the major source of PE (4), it can be synthesized in significant amounts via multiple nonmitochondrial pathways. This was confirmed by analysis of the total cellular phospholipid composition of psd2⌬ and dpl1⌬ mutants in YPD and CSM. Phospholipid extraction was carried out from cells grown to the late logarithmic phase in glucose-containing medium in order to simulate the growth conditions of spores obtained from tetrad dissection on glucose-containing medium. As shown in TABLES THREE and FOUR, the decrease in total PE levels in psd2⌬ cells was similar to the decrease in PE observed in psd1⌬ cells, suggesting that the decrease in cellular PE per se does not result in inviability of the crd1⌬psd1⌬ cells. This was confirmed by the recovery of viable crd1⌬psd2⌬ colonies (TABLE TWO). The dpl1⌬ mutation resulted in only a minor decrease in PE (TABLE FOUR), and crd1⌬dpl1⌬ double mutants were also recovered (TABLE TWO). Mitochondrial membranes are highly enriched in PC, comprising almost 40% of total mitochondrial phospholipids (2). However, disruption of PC synthesis did not lead to loss of viability of crd1⌬ cells, since viable crd1⌬pem1⌬ and crd1⌬pem2⌬ double mutants were recovered (TABLE TWO). Growth of these double mutants was similar to that of single mutant or wild-type cells in YPD or CSM (Fig. 3), even with up to a 20-fold reduction in PC in the crd1⌬pem2⌬ cells (TABLE THREE).
Synthetic Lethality of crd1⌬psd1⌬ Is Not Rescued by Ethanolamine or Propanolamine-Storey et al. (25) reported an increase in PE levels upon Etn supplementation and synthesis of a novel phospholipid, phosphatidylpropanolamine, upon Prn supplementation in CSM. We wished to determine whether synthetic lethality of crd1⌬psd1⌬ could be rescued by increasing the concentration of Etn or Prn, which rescue the temperature sensitivity of psd1⌬ cells. Steady-state phospholipid analysis of psd1⌬ cells in YPD medium showed that PE is not significantly increased with Etn supplementation, even at a concentration of  Wild-type, crd1⌬, psd1⌬, and crd1⌬psd1⌬ cells transformed with the pYES6/ CRD1 plasmid were plated on galactose-containing induction medium (YPI) (A) or glucose-containing repression medium (YPD) (B) and incubated for 3-4 days at 30°C. C, expression of Crd1-V5 from the GAL1 promoter is leaky. The crd1⌬ cells transformed with pYES6/CRD1 were grown in YPI or YPD medium to early stationary growth phase. Protein extraction, SDS-PAGE, and Western blot analysis were carried out as described under "Experimental Procedures." The Crd1-V5 fusion protein appears as a pair of closely migrating bands with a molecular mass of ϳ30 kDa. D, phospholipid analysis of wildtype and crd1⌬ cells transformed with pYES6/CRD1 or the control vector pYES6/CT. The wild-type ϩ pYES6 plasmids and crd1⌬ ϩ pYES6 plasmids were grown in YPI or YPD medium in the presence of 32 P to the early stationary growth phase. Cells were harvested by centrifugation, and phospholipid extraction, separation, and quantification were as described under "Experimental Procedures." Data are expressed as a percentage of radiolabel 32 P incorporated into total phospholipids. The mean value of two independent measurements is shown. N.D., not detectable. E, growth of WT ϩ pYES6/CRD1 and crd1⌬psd1⌬ ϩ pYES6/CRD1 cells in liquid YPD and YPI media is similar. Cells were precultured in liquid YPI medium for 24 h. Liquid YPD or YPI cultures were inoculated to an A 550 of 0.5 from preculture, and cell growth was monitored by measuring A 550 at the indicated times.
20 mM (Fig. 4A). Similarly, there was minimal synthesis of phosphatidylpropanolamine upon Prn supplementation in YPD medium (Fig.  4A). However, consistent with the findings of Storey et al. (25), supplementation with Etn in CSM medium resulted in a 2-fold increase in PE levels (Fig. 4B). Prn supplementation resulted in the synthesis of phosphatidylpropanolamine, comprising up to 30% of total phospholipid levels (Fig. 4B). The synthesis of PE and phosphatidylpropanolamine upon Etn and Prn supplementation occurs in the endoplasmic reticulum. To determine whether these phospholipids could be imported into mitochondria of mutant cells, we analyzed the phospholipid composi-tion of mitochondria from cells grown in the presence of Prn. Consistent with the findings of Storey et al. (25), we observed the presence of phosphatidylpropanolamine in crd1⌬ and psd1⌬ mitochondria (data not shown). However, crd1⌬psd1⌬ colonies could not be recovered on medium supplemented with either Etn or Prn (TABLE FIVE). The tiny colonies from spores deduced to be crd1⌬psd1⌬ did not have a greater number of cells than those germinated in the absence of Etn and Prn (data not shown). These results clearly demonstrate that the absence of both mitochondrial PE and CL biosynthesis is lethal.

DISCUSSION
In this study, we demonstrated a synthetic lethal interaction between the crd1⌬ mutant and the psd1⌬ mutant, which cannot synthesize PE in the mitochondria. Disruption of nonmitochondrial pathways of PE biosynthesis in crd1⌬ cells did not result in synthetic lethality. Increasing cellular PE and synthesis of phosphatidylpropanolamine also did not rescue the synthetic lethality of the crd1⌬psd1⌬ double mutant. These data indicate that cells cannot survive if synthesis of both PE and CL in the mitochondria is eliminated. This suggests that CL and PE have overlapping functions in the mitochondria and that loss of both is lethal.
The reduction in total cellular PE per se was not responsible for synthetic lethality in the crd1⌬ background, since the decrease in total cellular PE in psd2⌬ cells was similar to the decrease observed in psd1⌬ cells (TABLES THREE and FOUR). Up to a 20-fold decrease in PC, the most abundant mitochondrial phospholipid, due to a mutation in PEM2 (TABLE THREE), did not result in synthetic lethality, as evidenced by the recovery of viable crd1⌬pem2⌬ colonies (TABLE TWO). Consistent with the specific requirement of PE in the crd1⌬ background, we observed severely diminished growth of the crd1⌬psd2⌬ mutant in CSM compared with double mutants of the PC pathway (Fig. 3). These results pointed to a specific requirement for either mitochondrial PE biosynthesis or the mitochondrial PE biosynthetic enzyme Psd1p for the viability of CL-deficient cells. The presence of Psd1p itself in mitochondria may not be sufficient to allow survival of crd1⌬ cells, since Janitor et al. (30) have shown that disruption of the CHO1 gene in the CL-deficient pgs1⌬ background results in synthetic lethality. Cho1p catalyzes the synthesis of PS, which is the substrate for Psd1p. Therefore, the absence of PS results in an inability to synthesize mitochondrial PE via Psd1p. Thus, pgs1⌬ and cho1⌬ are synthetically lethal despite the presence of Psd1p.
It is well documented that in the yeast S. cerevisiae, mitochondrial biogenesis and respiration is triggered in nonfermentable medium or in  late logarithmic growth phase of cells grown in glucose-containing (fermentable) medium. The CRD1 gene expression, CL biosynthesis, and CL levels are higher in nonfermentable medium and upon entering the stationary growth phase in glucose-containing medium (29,31,32). Mitochondrial PE biosynthesis is essential for growth in nonfermentable medium (4). These results point to the requirement of CL and PE for respiratory functions. Thus, our results showing crd1⌬psd1⌬ synthetic lethality in fermentable medium are surprising. This suggests that continuous synthesis of mitochondrial PE or CL is required for essential processes other than mitochondrial respiration. A recent report has shown that Psd1p is the major source of mitochondrial PE (33). The same authors also observed a small but significant level of import of microsomal PE into mitochondria. Others (25) have shown that Etn supplementation could rescue the temperaturesensitive phenotype of psd1⌬ cells, consistent with import of PE into the mitochondria. Thus, we tried to rescue the crd1⌬psd1⌬ synthetic lethality by increasing PE levels via Etn supplementation. However, a 2-fold increase in PE due to Etn supplementation (Fig. 4B) failed to rescue the crd1⌬psd1⌬ synthetic lethal phenotype. Phosphatidylpropanolamine, a novel phospholipid not normally present in yeast, rescues the growth defects of PE-deficient psd1⌬ and psd1⌬psd2⌬ cells and has been shown to efficiently incorporate into mitochondrial membranes (25,34). However, we did not recover viable crd1⌬psd1⌬ cells even under conditions in which phosphatidylpropanolamine formed 30% of total phospholipids (Fig. 4B). The possibility that import of phosphatidylpropanolamine may be defective in CL-deficient mitochondria was ruled out, since we observed a significant amount of phosphatidylpropanolamine in crd1⌬ and psd1⌬ mitochondria (data not shown), suggesting that CL is not essential for the mitochondrial import of nonbi-layer phospholipids. Taken together, these data indicate that mitochondrial synthesis of PE is critical for survival of crd1⌬ cells.
What function of Psd1p-synthesized PE is required in the absence of CL? One of the most likely roles of PE is to meet a requirement for nonbilayer formation in the inner mitochondrial membrane in the absence of CL. Whereas externally synthesized PE or phosphatidylpropanolamine (data not shown) can be transported to mitochondria, these nonbilayer phospholipids may not be transported efficiently to the inner mitochondrial membrane to compensate for the lack of both PE and CL, and thus supplementation does not rescue the synthetic lethality of crd1⌬psd1⌬ cells. In support of this hypothesis, Burgemeister et al. (33) reported a 3-fold reduction in inner mitochondrial membrane PE levels in the psd1⌬ mutant as compared with wild type cells, indicating that PE synthesized in the microsomes or via Psd2p catalysis was not efficiently incorporated into the inner mitochondrial membrane. Previous studies in E. coli have shown that bacteria need a critical amount of nonbilayerforming phospholipid for survival (21). This observation is supported by the recent report showing that incorporation of a foreign nonbilayerforming glycolipid, ␣-monoglucosyldiacylglycerol, restores cellular function in PE-deficient E. coli cells (35). PE is also required for the formation of glycosylphosphatidylinositol-linked proteins (4,36), which play an essential role in cell viability by maintaining structural integrity of the cell wall. A recent report from our laboratory indicated an important role of CL and its precursor, phosphatidylglycerol, in maintaining cell wall integrity (37). Based on these results, we speculated that loss of viability of crd1⌬psd1⌬ cells could be the result of compromised cell wall stability. However, 1 M sorbitol, which rescues the growth defects of the pgs1⌬ mutant by stabilizing the cell wall, did not rescue the lethality of crd1⌬psd1⌬ (data not shown), indicating that synthetic lethality is

Supplementation of Etn and Prn does not rescue synthetic lethality of crd1⌬psd1⌬ cells
The BY4741/2 (CRD1/crd1⌬::URA3 PSD1/psd1⌬::KanMX4) heterozygous diploid was pregrown with the indicated supplementation. Diploids were sporulated, and tetrads were dissected on plates containing the same concentration of supplement. not due to loss of cell wall integrity. Genetic screens designed to identify suppressors of crd1⌬psd1⌬ synthetic lethality, currently in progress, will elucidate the potential function(s) of mitochondrial nonbilayer phospholipids.
The large scale synthetic lethal screen of Tong et al. (38) did not report crd1⌬psd1⌬ synthetic lethality, because it focused on genes involved in cell polarity, cell wall biogenesis, chromosome segregation, and DNA synthesis and repair. However, their work provided the framework to identify synthetic lethal interactions between genes of interest. Their study indicated a high likelihood of synthetic lethal interactions between genes with the same mutant phenotypes and subcellular localization. Identification of a synthetic lethal interaction between crd1⌬ and psd1⌬ in this study supports their prediction.