Respiratory Deficiency Mediates the Regulation of CHO1-encoded Phosphatidylserine Synthase by mRNA Stability in Saccharomyces cerevisiae*

The CHO1-encoded phosphatidylserine synthase (CDP-diacylglycerol:l-serine O-phosphatidyltransferase, EC 2.7.8.8) is one of the most highly regulated phospholipid biosynthetic enzymes in the yeast Saccharomyces cerevisiae. CHO1 expression is regulated by nutrient availability through a regulatory circuit involving a UASINO cis-acting element in the CHO1 promoter, the positive transcription factors Ino2p and Ino4p, and the transcriptional repressor Opi1p. In this work, we examined the post-transcriptional regulation of CHO1 by mRNA stability. CHO1 mRNA was stabilized in mutants defective in deadenylation (ccr4Δ), mRNA decapping (dcp1), and the 5′–3′-exonuclease (xrn1), indicating that the CHO1 transcript is primarily degraded through the general 5′–3′ mRNA decay pathway. In respiratory-sufficient cells, the CHO1 transcript was moderately stable with a half-life of 12 min. However, the CHO1 transcript was stabilized to a half-life of >45 min in respiratory-deficient (rho– and rhoo) cells, the cox4Δ mutant defective in the cytochrome c oxidase, and wild type cells treated with KCN (a cytochrome c oxidase inhibitor). The increased CHO1 mRNA stability in response to respiratory deficiency caused increases in CHO1 mRNA abundance, phosphatidylserine synthase protein and activity, and the synthesis of phosphatidylserine in vivo. Respiratory deficiency also caused increases in the activities of CDP-diacylglycerol synthase, phosphatidylserine decarboxylase, and the phospholipid methyltransferases. Phosphatidylinositol synthase and choline kinase activities were not affected by respiratory deficiency. This work advances our understanding of phosphatidylserine synthase regulation and underscores the importance of mitochondrial respiration to the regulation of phospholipid synthesis in S. cerevisiae.

The CHO1-encoded (1-3) PS 2 synthase (CDP-diacylglycerol:L-serine O-phosphatidyltransferase, EC 2.7.8.8) 3 is one of the most highly regulated enzymes of phospholipid synthesis in the yeast Saccharomyces cerevisiae (4 -6). PS synthase is an integral membrane protein that is localized to the endoplasmic reticulum (7). It catalyzes the formation of PS by a Mn 2ϩ -dependent sequential reaction by displacing CMP from CDP-DAG with serine (8). The reaction product PS is a major structural component of yeast membranes accounting for 4 -18% of total phospholipids depending on growth conditions (9 -11). PS also serves as the precursor for the synthesis of the most abundant membrane phospholipids PE (20 -32%) and PC (35-55%) that are synthesized by the de novo CDP-DAG pathway ( Fig. 1) (5,9). cho1 mutants defective in PS synthase activity can still synthesize PE and PC if they are supplemented with ethanolamine or choline; indeed, cho1 mutants are ethanolamine/choline auxotrophs (12,13). Ethanolamine is used for PE synthesis via the CDP-ethanolamine branch of the Kennedy pathway (Fig. 1). The PE synthesized by the Kennedy pathway may be methylated to PC via the CDP-DAG pathway (Fig. 1). Choline is used for PC synthesis via the CDP-choline branch of the Kennedy pathway ( Fig. 1). In wild type cells, both the CDP-DAG and Kennedy pathways contribute to the synthesis of PC regardless of whether choline is supplemented to the growth medium (14 -19). If choline is not present in the growth medium, the choline required for the Kennedy pathway is derived from the phospholipase D-mediated turnover of PC synthesized by way of the CDP-DAG pathway (19,20). PS synthase is regulated by biochemical and genetic mechanisms, both of which have an impact on the synthesis of PC via the CDP-DAG and Kennedy pathways (4 -6, 21). The activity of PS synthase is modulated (i.e. inhibited or activated) by membrane phospholipids (e.g. PA, phosphatidylglycerol, and cardiolipin) (22)(23)(24) and is inhibited by inositol (25) and by the nucleotide CTP (17). In addition, the phosphorylation of the enzyme inhibits its activity, whereas dephosphorylation stimulates its activity (26,27). In general, the inhibition of PS synthase activity favors PC synthesis via the Kennedy pathway (4,5). The biochemical regulation of PS synthase activity also governs the partitioning of the substrate CDP-DAG between PS and PI; the inhibition of PS synthase activity favors PI synthesis ( Fig. 1) (4).
The expression of the PS synthase (CHO1) gene is regulated by the supplementation of water-soluble phospholipid precursors (e.g. inositol) (28 -31), zinc deprivation (32), and by growth phase (33,34). CHO1 is maximally expressed in exponential phase cells when grown in the absence of inositol (28 -31) and grown in the presence of zinc (32). CHO1 is repressed when inositol is supplemented to the growth medium (28 -31) or when zinc is depleted from the growth medium (32). The zincmediated regulation of CHO1 occurs in the absence of inositol supplementation (32). Repression of CHO1 also occurs when cells enter the stationary phase of growth (33,34). These forms of regulation are dependent on the UAS INO cis-acting element in the promoter of the CHO1 gene (21). The derepression of CHO1 is mediated by a heterodimer complex of the positive transcription factors Ino2p and Ino4p that bind to a UAS INO cis-acting element to drive transcription (5,21,35,36). Repression of CHO1 is mediated by the repressor Opi1p, which interacts with Ino2p to attenuate transcription (5,21,35,36). Opi1p repressor function is regulated by the cellular concentration of PA, which helps anchor the repressor to the nuclear/endoplasmic reticulum membrane apart from the Ino2p-Ino4p complex bound to the UAS INO element (37). PA concentration and Opi1p repressor function is mediated in part by the PAH1encoded Mg 2ϩ -dependent PA phosphatase enzyme (38 -40). Data (41) are consistent with a model (40) whereby activation of Mg 2ϩ -dependent PA phosphatase activity results in a decrease in PA concentration followed by the translocation of Opi1p into the nucleus for interaction with Ino2p to repress CHO1 transcription. As discussed above for the biochemical regulation of PS synthase activity, the repression of CHO1 favors PI synthesis and the Kennedy pathway for PC synthesis (5).
Decay analysis of CHO1 mRNA in a cki1⌬ eki1⌬ mutant defective in the synthesis of phospholipids via the Kennedy pathway ( Fig. 1) has revealed a novel mechanism by which CHO1 expression is regulated independent of the UAS INO element in the CHO1 promoter (42). In wild type cells, CHO1 mRNA is moderately stable with a half-life of 12 min when compared with other S. cerevisiae mRNAs that have half-lives ranging from 1 to 60 min (43). However, CHO1 mRNA is greatly stabilized with a half-life Ͼ45 min in the cki1⌬ eki1⌬ (KS106) mutant (42). This results in increased levels of the PS synthase protein and its associated activity (42). The objective of this work was to identify an intermediate or end product of the Kennedy pathway that was responsible for regulation of CHO1 mRNA stability. During the course of this work, we dis-covered that the stabilization of CHO1 mRNA was not mediated by components of the Kennedy pathway, but instead it was mediated by a defect in mitochondrial respiration. This work underscores the importance of respiration to the regulation of phospholipid synthesis and advances our understanding of PS synthase regulation in yeast.

EXPERIMENTAL PROCEDURES
Materials-All chemicals were reagent grade. Growth medium supplies were purchased from Difco. The plasmid DNA purification and DNA gel extraction kits were from Qiagen. NEBlot kit, restriction endonucleases, recombinant Vent DNA polymerase, and modifying enzymes were purchased from New England Biolabs. RNA size markers were purchased from Promega. Yeast transformation kit was obtained from Clontech. Enhanced chemifluorescence Western blotting detection kit, polyvinylidene difluoride membranes, and Probe-Quant G-50 columns were purchased from GE Healthcare. Bio-Rad was the supplier of Zeta Probe blotting membranes, protein assay reagents, electrophoretic reagents, acrylamide solutions, immunochemical reagents, the DNA size ladder used for agarose gel electrophoresis, and protein molecular mass standards for SDS-PAGE. AdoMet, ampicillin, aprotinin, benzamidine, bovine serum albumin, choline, CTP, leupeptin, N-ethylmaleimide, pepstatin, phenylmethylsulfonyl fluoride, L-serine, and Triton X-100 were purchased from Sigma. Phospholipids were purchased from Avanti Polar Lipids. TLC plates were from EM Science. Radiochemicals and scintillation counting supplies were from PerkinElmer Life Sciences and National Diagnostics, respectively. Thiolutin was a gift from Pfizer.
Strains, Plasmids, and Growth Conditions-The bacterial and yeast strains used in this work are listed in Table 1. Methods for the growth of bacteria and yeast were described previously (44,45). Cells were grown at 30°C in complete synthetic medium without inositol (46) using either 2% glucose or 2% glycerol as the carbon source. For selection of cells bearing plasmids, appropriate nutrients were omitted from synthetic complete medium. Glucose-grown cells were also cultured in the presence of 1 mM hydrogen peroxide to induce oxidative stress (47,48). Cells in liquid media were grown to the exponential phase (1-2 ϫ 10 7 cells/ml), and cell numbers were determined spectrophotometrically at an absorbance of 600 nm. Plasmids were maintained and amplified in Escherichia coli strain DH5␣, which was grown in LB medium (1% tryptone, 0.5% yeast extract, 1% NaCl, pH 7.4) at 37°C. The plasmids used in this work are listed in Table 2. Ampicillin (100 mg/ml) was added to cultures of DH5␣ carrying plasmids. For growth on plates, yeast and bacterial media were supplemented with 2 and 1.5% agar, respectively. Respiratory sufficiency was scored by growth on YPG (1% yeast extract, 2% peptone, 2% glycerol) and YPD (1% yeast extract, 2% peptone, 2% glucose) media plates (44,45).
4Ј,6-Diamidino-2-phenylindole Staining of Mitochondria-Mitochondrial DNA of S. cerevisiae cells was examined by 4Ј,6-diamidino-2-phenylindole staining (49) using a Nikon Eclipse E800 fluorescence microscope equipped with a Hamamatsu Orca digital camera. Images were captured in monochrome and processed using Improvision Openlab software. DNA Isolation and Manipulations-Plasmid and genomic DNA preparation, restriction enzyme digestion, and DNA ligations were performed according to standard protocols (45). Transformations of yeast (50) and E. coli (45) were performed as described previously. Plasmid maintenance and amplifications were performed in E. coli strain DH5␣. Conditions for the amplification of DNA by PCR were optimized as described previously (51).
Analysis of CHO1 mRNA Decay-Total RNA was isolated from cells (43,56), resolved overnight at 22 V on a 1.1% formaldehyde gel (57), and then transferred to Zeta Probe membrane by vacuum blotting. The CHO1 (17) and PGK1 (58) probes were labeled with [␣-32 P]dTTP using the NEBlot random primer labeling kit. Unincorporated nucleotides were removed using ProbeQuant G-50 columns. Pre-hybridization, hybridization with the probes, and washes to remove nonspecific binding were carried out according to manufacturer's instructions. Images of radiolabeled mRNAs were acquired by phosphorimaging analysis. The half-life of CHO1 mRNA was determined after the arrest of transcription with thiolutin (15 g/ml) as described by Gonzalez and Martin (58).
Immunoblotting-SDS-PAGE (59) using 12% slab gels and transfer of proteins to polyvinylidene difluoride membranes (60) were performed as described previously. The membrane was probed with a 1:500 dilution of anti-PS synthase antibodies (42). Goat anti-rabbit IgG alkaline phosphatase conjugate was used as a secondary antibody at a dilution of 1:5000. The PS synthase protein was detected using the enhanced chemifluorescence Western blotting detection kit, and the protein signals were acquired by fluoroimaging. The relative density of the protein was analyzed using Image-Quant software. Immunoblot signals were in the linear range of detectability.
Preparation of Yeast Cell Extract and the Total Membrane Fraction-The cell extract and total membrane fraction were prepared as described previously (61). Cell pellets were homogenized with glass beads at 4°C in 50 mM Tris maleate buffer (pH   7.0) containing 1 mM EDTA, 0.3 M sucrose, 10 mM 2-mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, and 5 g/ml each of aprotinin, leupeptin, and pepstatin. The cell extract was obtained by centrifugation of the homogenate at 1,500 ϫ g for 10 min. The cell extract was centrifuged at 100,000 ϫ g for 1 h to obtain total membranes, which were then resuspended in 50 mM Tris maleate (pH 7.0) buffer containing 10 mM MgCl 2 , 10 mM 2-mercaptoethanol, 20% glycerol (w/v), and 0.5 mM phenylmethylsulfonyl fluoride. Protein concentration was estimated by the method of Bradford (62) using bovine serum albumin as the standard.  (69). All assays were linear with time and protein concentration. The average standard deviation of the assays was Ϯ5%. A unit of phospholipid enzymatic activity was defined as the amount of enzyme that catalyzed the formation of 1 nmol of product per min.
Labeling and Analysis of Phospholipids-Labeling of phospholipids with [ 14 C]serine was performed as described previously (12,13). Phospholipid synthesis was followed by labeling exponential phase cells for 30 min. Phospholipids were extracted from labeled cells by the method of Bligh and Dyer (70). Two-dimension TLC with chloroform/methanol/ammonium hydroxide/H 2 O (90:50:4:6) (dimension 1) and chloroform/methanol/acetic acid/H 2 O (63:8:10:2) (dimension 2) was used to separate phospholipids (42). Radiolabeled phospholipids on the chromatography plates were visualized by phosphorimaging, and their identities were confirmed by comparison with standard phospholipids after exposure to iodine vapor. The amount of each 14 C-labeled phospholipid was determined by liquid scintillation counting.
Analyses of Data-Statistical analyses were performed with SigmaPlot software. Statistical significance was determined by performing the Student's t test. p values Ͻ0.05 were taken as a significant difference.

RESULTS
CHO1 mRNA Decay Involves Deadenylation, Decapping, and 5Ј-3Ј-Exonuclease Cleavage-Many mRNAs in S. cerevisiae are degraded by a general deadenylation-dependent pathway involving the deadenylation of the poly(A) tail by Ccr4p, the removal of the 5Ј cap by the Dcp1p-Dcp2p complex, and the 5Ј-3Ј-exonuclease cleavage by Xrn1p (71). To examine whether CHO1 mRNA is degraded by this general pathway, a decay analysis was performed in mutants defective in each of the three steps. The rate of CHO1 mRNA decay was reduced in the ccr4⌬ mutant defective in deadenylation, in the dcp1⌬ mutant defective in decapping, and in the xrn1⌬ mutant defective in 5Ј-3Ј-exonuclease cleavage ( Fig. 2 and Table 3). That CHO1 mRNA was stabilized in these mutants indicated that the CHO1 transcript is degraded by the general decay pathway (71).
Effects of the cki1⌬, eki1⌬, and cki1⌬ eki1⌬ Mutations on CHO1 mRNA Decay-Previous studies have shown that the cki1⌬ eki1⌬ mutations have a stabilizing effect on the abundance of CHO1 mRNA (42). It is unclear, however, whether the signal responsible for increased CHO1 mRNA stability in the cki1⌬ eki1⌬ (KS106) mutant results from a defect in the CDPcholine branch, the CDP-ethanolamine branch, or from both branches of the Kennedy pathway. To address this question, CHO1 mRNA decay was examined in the cki1⌬ (KS105) mutant (53) defective in the first step of the CDP-choline pathway and in the eki1⌬ (KS101) mutant (53) defective in the first step of the CDP-ethanolamine pathway (Fig. 1). CHO1 mRNA decay was also reexamined in the cki1⌬ eki1⌬ (KS106) mutant. The half-life of CHO1 mRNA decay in the cki1⌬ mutant was similar to that observed in the wild type control (Fig. 3 and Table 3). In contrast, CHO1 mRNA was stabilized in the eki1⌬ (KS101) mutant ( Fig. 3 and Table 3). Thus, the increase in CHO1 mRNA stability in the cki1⌬ eki1⌬ (KS106) mutant was because of the eki1⌬ mutation (i.e. defect in the CDP-ethanolamine branch).
Effects of CDP-ethanolamine Pathway Mutations on CHO1 mRNA Decay-Mutations in the second (ect1⌬) and third (ept1⌬) steps of the CDP-ethanolamine pathway (i.e. phosphoethanolamine cytidylyltransferase and phosphocholine cytidylyltransferase, respectively) were constructed and examined for their effects on CHO1 mRNA decay. Of the three mutants, eki1⌬ (KS101) was the only mutant in which CHO1 mRNA was stabilized ( Fig. 4 and Table 3). The effects of the three CDP-ethanolamine pathway mutations on the abundance of CHO1 mRNA and PS synthase protein (Cho1p) were also examined (Fig. 5). The levels of CHO1 mRNA and PS synthase protein were elevated by about 2-fold in the eki1⌬ (KS101) mutant when compared with the wild type control (Fig. 5). As described previously (42), CHO1 mRNA and Cho1p were elevated in the cki1⌬ eki1⌬ (KS106) mutant. These results correlated with the increase in CHO1 mRNA stability observed in the eki1⌬ and cki1⌬ eki1⌬ mutants. On the other hand, the CHO1 gene products were not affected in the ect1⌬ and ept1⌬ mutants (Fig. 5).
Loss of the EKI1 Gene Is Not Responsible for Increased CHO1 mRNA Stability in the eki1⌬ (KS101) Mutant-To confirm that the deletion of the EKI1 gene was responsible for the increased stability of CHO1 mRNA, the eki1⌬ (KS101) mutant was transformed with a single copy plasmid bearing the EKI1 gene. Unexpectedly, the introduction of the EKI1 gene into the eki1⌬ (KS101) mutant did not reverse the stable CHO1 mRNA phe-notype of the mutant (Fig. 6 and Table 3). This result indicated that a second mutation in the eki1⌬ mutant was responsible for increasing CHO1 mRNA stability. Accordingly, a new eki1⌬ (HCY5) mutant was constructed by the same method (53) used  Table 3.  Table 3. to construct the original eki1⌬ (KS101) mutant. The analysis of CHO1 mRNA decay in the new eki1⌬ mutant confirmed that the loss of the EKI1 gene was not responsible for the increased stability of CHO1 mRNA ( Fig. 6 and Table 3). Moreover, a new cki1⌬ eki1⌬ (HCY7) mutant was constructed, and it exhibited the wild type rate of CHO1 mRNA decay. The eki1⌬ (KS101) and cki1⌬ eki1⌬ (KS106) Mutants Are Respiratory-deficient-Unlike wild type (W303-1B) cells that formed pink colonies on agar plates, the eki1⌬ (KS101) and cki1⌬ eki1⌬ (KS106) mutant colonies were white and relatively small (i.e. petite) in appearance. These characteristics were first thought to be due to the eki1⌬ mutation. However, the colonies produced by the new eki1⌬ (HCY5) and cki1⌬ eki1⌬ (HCY7) mutants were similar in appearance (i.e. normal size and pink) to their wild type parent W303-1B. Petite white colonies are characteristics of respiratory-deficient mutants (44). Respiratory-deficient mutants are defective in mitochondrial function and cannot grow on nonfermentable carbon sources (44,72). With this in mind, the respiratory sufficiency of the original eki1⌬ (KS101) and cki1⌬ eki1⌬ (KS106) mutants was examined by growth on agar plates containing glucose (fermentable) or glycerol (nonfermentable) as the carbon source. Like known respiratory mutants (MGY100 and W303 [rho o ]), the original eki1⌬ (KS101) and cki1⌬ eki1⌬ (KS106) mutants only grew on glucose agar plates, whereas the new eki1⌬ (HCY5) and cki1⌬ eki1⌬ (HCY7) mutants grew on both glucose and glycerol agar plates (Table 4). These data indicated that the original mutants were respiratory-deficient. The white colony phenotype can be attributed to respiratory deficiency because the development of a pink color in the ade2 genetic background found in the eki1⌬ (KS101) and cki1⌬ eki1⌬ (KS106) mutants and the parent W303-1B strain would be dependent on oxidative phosphoryl-ation (73). Another characteristic common to respiratory-deficient mutants is the inability to grow at elevated temperatures (74). Indeed, the original eki1⌬ (KS101) and cki1⌬ eki1⌬  Table 3.  cki1⌬ eki1⌬ (KS106), ect1⌬ (HCY3), and ept1⌬ (HCY4) mutant cells were grown to the exponential phase (1 ϫ 10 7 cells/ml) of growth. A, abundance of CHO1 mRNA was determined with 10 g of total RNA by Northern blot analysis. The relative amounts of CHO1 and PGK1 mRNAs from wild type and mutant cells were determined by ImageQuant analysis of the data. The relative amount of CHO1 to PGK1 mRNA in wild type cells was arbitrarily set at 1. B, total membrane fraction (12.5 g of protein) was subjected to immunoblot analysis using a 1:500 dilution of anti-PS synthase antibodies. The relative amounts of the PS synthase protein from wild type and mutant cells were determined by ImageQuant analysis of thedata.TheamountofPSsynthaseproteinfoundinwildtypecellswasarbitrarilyset at 1. The data shown in A and B are the average of three experiments Ϯ S.D.
(KS106) mutants were temperature-sensitive for growth at 37°C, whereas the new mutants were not temperature-sensitive for growth (Table 4).
Interestingly, the new eki1⌬ (HCY5) mutant had a tendency to give rise to spontaneous petite colonies that were white in color. Likewise, the wild type parent W303-1B also gave rise to petite white colonies but at a lower frequency. One petite colony (strain HCY6) that was derived from the new eki1⌬ (HCY5) mutant and one petite colony (strain HCY8) that was derived from the parent strain W303-1B were examined for their growth on glucose and glycerol, and for their growth on glucose at 37°C. Strains HCY6 and HCY8 exhibited the phenotypes of respiratory deficiency (Table 4).
Respiratory-deficient mutants are differentiated by the extent of mitochondrial DNA that they lack. Cells that have a reduced amount of mitochondrial DNA are designated rho Ϫ , whereas cells that lack mitochondrial DNA are desig-nated rho o (44,72). Fluorescence microscopy of cells stained with 4Ј,6-diamidino-2-phenylindole (44) indicated that KS101, KS106, and HCY6 were rho o , whereas HCY8 was rho Ϫ .
Respiratory Deficiency Is Responsible for Increased CHO1 mRNA Stability-CHO1 mRNA decay was examined in two respiratory-deficient mutants. These mutants included a known rho Ϫ mutant (MGY100) and a known rho o mutant (W303-1A [rho o ]). The two respiratory mutants exhibited an increase in the half-life of CHO1 mRNA when compared with that of the respiratory-sufficient control ( Fig. 7 and Table 3). Thus, respiratory deficiency, whether because of a partial or total lack of the mitochondrial genome, gave rise to increased CHO1 mRNA stability. These data also confirmed that the basis for the increased stability of CHO1 mRNA in the original eki1⌬ (KS101) mutant was respiratory deficiency.
Respiration is a major function of the mitochondrion, which takes place at the inner mitochondria by five enzymatic complexes (75). KCN, a specific inhibitor of cytochrome c oxidase (complex IV), is commonly used to halt the electron transport chain and mitochondrial respiration (75,76). To determine the effects of inhibiting respiration on CHO1 mRNA stability, respiratory-sufficient cells (W303-1B) were grown in the absence and presence of KCN. The addition of KCN to the growth medium resulted in a dose-dependent increase in the half-life of CHO1 mRNA (Fig. 8 and Table 3). To further confirm that a block in respiration was responsible for the increase in CHO1 mRNA stability, the decay of the CHO1 transcript was analyzed in the cox4⌬ mutant. COX4 is a nuclear gene that encodes an indispensable subunit of cytochrome c oxidase (77). The decay rate of CHO1 mRNA was much longer in the cox4⌬ mutant  Table 3.  Table 3.

Relevant genotype (strain)
Growth a

Glucose, 37°C
Wild type (W303-1B) The indicated cells were grown in YPD medium to stationary phase at 30°C. Cells were diluted and spotted onto YPG and YPD media plates. The plates were incubated at the indicated temperatures, and colony growth was scored after 5 days. ϩ, growth; Ϫ, no growth. OCTOBER 26, 2007 • VOLUME 282 • NUMBER 43

Respiratory Deficiency Mediates CHO1 mRNA Stability
when compared with the wild type control ( Fig. 9 and Table 3). Taken together, these data provided strong evidence that respiratory deficiency mediates CHO1 mRNA stability. Effects of Respiratory Deficiency on the Abundance of CHO1 mRNA, PS Synthase Protein, and PS Synthase Activity and on the Synthesis of PS in Vivo-Northern blot and Western blot analyses for the CHO1 gene products were performed with a respiratory-deficient mutant, and the results are presented in Fig. 10, A and B, respectively. The levels of both CHO1 mRNA and PS synthase protein (Cho1p) were about 2-fold higher in the respiratory-deficient mutant when compared with the levels in the wild type control. The increased levels of CHO1 mRNA and PS synthase protein in the respiratory mutant correlated with an increased level of PS synthase activity (Fig. 10C). Elevated PS synthase activity was also characteristic of respiratory-deficient eki1⌬ (KS101) and cki1⌬ eki1⌬ (KS106) mutants but not respiratory-sufficient eki1⌬ (HCY5) and cki1⌬ eki1⌬ (HCY7) mutants. To examine the effect of respiratory deficiency on the synthesis of PS in vivo, exponential phase cells were labeled with [ 14 C]serine for 30 min followed by the extraction and analysis of phospholipids. [ 14 C]Serine is directly incorporated into PS via the PS synthase enzyme, followed by the incorporation of the label into PE and PC via the reactions catalyzed by the PS decarboxylase and phospholipid methyltransferase enzymes (i.e. CDP-DAG pathway) (5,12). The respiratory-deficient mutant showed a 1.9-fold increase in the incorporation of [ 14 C]serine into PS after the 30-min labeling period (Fig. 10D). The effects of respiratory deficiency on the incorporation of the label into PE and PC were less dramatic  Table 3.  Table 3. Wild type (WT, W303-1B) and rho Ϫ (HCY8) mutant cells were grown to the exponential phase (1 ϫ 10 7 cells/ml) of growth. A, abundance of CHO1 mRNA was determined with 10 g of total RNA by Northern blot analysis. The relative amounts of CHO1 and PGK1 mRNAs from wild type and mutant cells were determined by ImageQuant analysis of the data. The relative amount of CHO1 to PGK1 mRNA in wild type cells was arbitrarily set at 1. B, total membrane fraction (12.5 g of protein) was subjected to immunoblot analysis using a 1:500 dilution of anti-PS synthase antibodies. The relative amounts of the PS synthase protein from wild type and mutant cells were determined by Image-Quant analysis of the data. The amount of PS synthase protein found in wild type cells was arbitrarily set at 1. C, total membrane fraction was isolated and used for the assay of PS synthase activity. D, cells were incubated with [ 14 C]serine for 30 min. Phospholipids were extracted and analyzed by twodimensional TLC. The data shown in A-D are the average of three experiments Ϯ S.D. (Fig. 10D). Respiratory deficiency did not have a significant effect on the steady state composition of phospholipids labeled Effects of Respiratory Deficiency on the Levels of Phospholipid Synthesis Enzyme Activities-In addition to PS synthase, the activities of other CDP-DAG pathway enzymes are elevated in the original cki1⌬ eki1⌬ (KS106) mutant (42). To address whether this regulation was because of the Kennedy pathway mutations or because of respiratory deficiency, the enzyme activity analyses were performed with the rho Ϫ (HCY8) mutant. The levels of CDP-DAG synthase (29%), PS decarboxylase (26%), PE methyltransferase (60%), and phospholipid methyltransferase (34%) activities were elevated in respiratorydeficient cells when compared with the control (Fig. 11). On the other hand, the activities of PI synthase and choline kinase were not affected by respiratory deficiency (Fig. 11). The analyses of these activities in the new cki1⌬ eki1⌬ mutant (HCY7) showed that the Kennedy pathway mutations did not affect the activity levels of the CDP-DAG pathway enzymes.
Effects of Oxidative Stress and Respiratory Growth on CHO1 mRNA Decay/Stability-We questioned whether oxidative stress affected the rate of CHO1 mRNA decay. For this experiment, glucose-grown wild type cells were incubated with hydrogen peroxide, a treatment commonly used to induce oxidative stress (47,48). This treatment did not affect the rate of CHO1 mRNA decay. We also questioned whether respiratory growth affected the decay of the CHO1 transcript. The rate of CHO1 mRNA decay of wild type cells grown with glycerol was not significantly different from cells grown with glucose.

DISCUSSION
In this work, we showed that the increased stability of CHO1 mRNA in the original cki1⌬ eki1⌬ (KS106) mutant was caused by respiratory deficiency associated with the loss of ethanol-amine kinase (i.e. eki1⌬ mutation) and not from the Kennedy pathway mutations per se. Decay analysis of CHO1 mRNA using respiration-deficient mutants (rho Ϫ and rho o ), the cox4⌬ mutant defective in the cytochrome c oxidase, and wild type cells treated with KCN (a cytochrome c oxidase inhibitor) confirmed that respiratory deficiency was responsible for the increase in the half-life of the CHO1 transcript. Moreover, the increased CHO1 mRNA stability in response to respiratory deficiency was responsible for increased PS synthase protein and activity and the synthesis of PS in vivo.
It was not too surprising that the eki1⌬ mutation gave rise to respiratory-deficient cells. Phospholipid synthesis mutants (e.g. cho1, psd1⌬, opi3) commonly give rise to respiratory-deficient petite colonies (13,78,79). Although it is not yet clear how the eki1⌬ mutation caused respiratory deficiency, the reason for this general phenomenon may be attributed to the important roles that phospholipids (e.g. cardiolipin and PE) play in both the composition and function of mitochondrial membranes (78, 80 -88).
Genome-wide analyses of gene expression in S. cerevisiae have shown that the levels of CHO1 mRNA decrease when cells transit from glucose-based fermentative to glycerol-based respiratory growth (89) and that CHO1 mRNA levels increase in response to oxygen deprivation (90). These findings indicate that a lack of mitochondrial respiration is a physiological condition that requires an increase in the abundance of the CHO1 transcript, supporting the role of mitochondrial respiration in the control of CHO1 expression. We showed in this study that the rate of CHO1 mRNA decay was not affected in glycerolgrown cells. Thus, the regulation that occurs when cells transit from fermentative to nonfermentative growth occurs at the level of transcription. Because respiratory deficiency may lead to oxidative stress, we questioned whether incubation with hydrogen peroxide affected the rate CHO1 mRNA stability. This stress condition did not affect the decay rate of CHO1 mRNA.
CHO1 mRNA was stabilized in mutants defective in deadenylation (ccr4⌬), mRNA decapping (dcp1), and the 5Ј-3Ј-exonuclease (xrn1). These results indicated that the CHO1 transcript is primarily degraded through the general 5Ј-3Ј mRNA decay pathway (71). The specific mechanisms and regulators that control the stabilization of CHO1 mRNA in response to respiratory deficiency are unknown. Given that CHO1 mRNA decays by the primary 5Ј-3Ј decay pathway when cells are respiratory-sufficient, it is reasonable to predict that the rate of deadenylation and/or decapping may be reduced when respiration is blocked. Extensive studies will be required to identify cis-acting elements in the transcript as well as components of the signal transduction system that lead from the respiratory defect to the mRNA stabilization response.
The activities of other CDP-DAG pathway enzymes (e.g. CDP-DAG synthase, PS decarboxylase, and the phospholipid methyltransferase enzymes) were also elevated in respiratorydeficient cells. These results were similar to that found in the respiratory-deficient cki1⌬ eki1⌬ (KS106) mutant (42). A reanalysis of these enzyme activities in the respiratory-sufficient cki1⌬ eki1⌬ (HCY7) mutant confirmed that the regulation was solely due to the rho o mutation. The regulation of FIGURE 11. Effects of the rho ؊ mutation on the levels of phospholipid synthesis enzyme activities. Wild type (WT, W303-1B) and rho Ϫ (HCY8) mutant cells were grown to the exponential phase (1 ϫ 10 7 cells/ml) of growth. The total membrane fraction was isolated and used for the assay of CDP-DAG synthase (CDS), PS synthase (PSS), PS decarboxylase (PSD), PE methyltransferase (PEMT), phospholipid methyltransferase (PLMT), and PI synthase (PIS). The cell extract was used for the assay of choline kinase (CK) activity. The specific activities (nmol/min/mg) of these enzymes from wild type cells were 0.92 Ϯ 0.04, 2.2 Ϯ 0.03, 0.41 Ϯ 0.02, 0.4 Ϯ 0.05, 0.64 Ϯ 0.01, 2.5 Ϯ 0.14, and 4.5 Ϯ 0.14. Each data point represents the average of triplicate enzyme determinations from two independent experiments Ϯ S.D. OCTOBER 26, 2007 • VOLUME 282 • NUMBER 43 phospholipid synthesis enzymes in response to respiratory deficiency was not a universal response as the PI synthase and choline kinase activities were not affected by respiratory deficiency.

Respiratory Deficiency Mediates CHO1 mRNA Stability
OLE1, which encodes the ⌬-9 fatty acid desaturase, is another yeast gene of lipid metabolism whose expression is independently regulated at the levels of transcription and by mRNA stability (58,91,92). The OLE1 transcript is destabilized when cells are supplemented with unsaturated fatty acids (58,91,92). This fatty acid-regulated decay of OLE1 mRNA occurs through both the 5Ј-3Ј general pathway and via exosomal 3Ј-5Ј degradation activities (92). Interestingly, like CHO1 mRNA, OLE1 transcript levels are more abundant under oxygen-deprived conditions (92). It is not clear, however, whether this regulation is because of an increased rate of transcription, a decreased rate of mRNA decay, or contributions from both regulatory systems.
In summary, we showed that the CHO1 transcript is degraded by the general mRNA decay pathway, and we clarified that a block in the Kennedy pathway for phospholipid synthesis was not responsible for the regulation of PS synthase by mRNA stability. The rate of mRNA decay plays an important role in the control of gene expression (93). The half-life of an mRNA governs the number of times a transcript can be translated, which in turn governs the amount of protein that can be produced at a given rate of transcription (93). We hypothesize that S. cerevisiae cells compensated for the stress of respiratory deficiency by conserving the CHO1 transcript for translation to the PS synthase enzyme for the synthesis of PS. This work advances our understanding of PS synthase regulation and underscores the importance of mitochondrial respiration to the regulation of phospholipid synthesis in S. cerevisiae.