Phosphatidylcholine and N-methylated phospholipids are nonessential in Saccharomyces cerevisiae.

Phosphatidylcholine (PtdCho) is the most abundant phospholipid in numerous eukaryotes and is generally thought to be essential for membrane structure and cellular function. We designed a specific test of this idea by using genetic and biochemical manipulation of yeast. Yeast mutants (pem1 pem2Delta) lacking the phosphatidylethanolamine (PtdEtn) methyltransferase enzymes require choline for growth and cannot make N-methylated phospholipids. When these strains are grown on a glucose carbon source supplemented with 20 mm propanolamine (Prn), the PtdCho level declines precipitously to the limits of detection (<0.6%), and the hexagonal phase-forming, primary amine-containing lipids, PtdEtn and PtdPrn, constitute approximately 60% of the total phospholipid content of the cell. When the lipids were analyzed by mass spectrometry, there was no compensatory shift in unsaturation of the PtdEtn and PtdPrn toward more bilayer-forming species. Thus the majority of the cellular amino phospholipids remained hexagonal phase-forming. The pem1 pem2Delta cells will also grow without choline, in the presence of Prn, on nonfermentable carbon sources (requiring functional mitochondria) and accumulate nearly 70% of their phospholipid as hexagonal phase-forming types. These data provide compelling evidence that the functions of PtdCho and N-methylated lipids in membranes are nonessential in Saccharomyces cerevisiae.

Etn and PtdCho are synthesized via the Kennedy pathways that use CDP-ethanolamine and CDP-choline intermediates. PtdEtn derived from the Kennedy pathway can also be methylated to form PtdCho (6). Phosphoethanolamine, an intermediate in the Kennedy pathway, can also be synthesized as a consequence of sphingolipid degradation (7,8). Monomethylethanolamine (Etn(Me)) and dimethylethanolamine (Etn(Me) 2 ) can also be assimilated into PtdEtn(Me) and Ptd-Etn(Me) 2 , respectively, via the Kennedy pathways. Yeast mutants (pem1 pem2⌬) lacking the methyltransferase enzymes cannot make N-methylated phospholipids by the de novo pathway. The growth defect of these mutants can be rescued by choline, Etn(Me) 2 , or Etn(Me) but not by Etn (4). These findings have led to the conclusions that some form of methylated PtdEtn is essential for yeast growth (1,4).
To test if PtdCho and N-methylated lipids are critical for cell viability, we examined whether propanolamine (Prn), an ethanolamine analogue, could rescue the growth defect of methylation-defective mutant strains. Previous research showed that yeast mutants (psd1⌬ psd2⌬) lacking PtdSer decarboxylases require Cho or Etn for growth but can also grow on Prn (8). Mutant strains supplemented with Prn accumulate a novel lipid, phosphatidylpropanolamine (PtdPrn). PtdPrn can comprise up to 40% of the total phospholipid content in Prn-supplemented cells, at the expense of PtdCho and PtdEtn. In Prn-supplemented cells with psd1⌬ psd2⌬ mutations, the Ptd-Cho content was reduced from 38 to 10% total phospholipid, and the PtdEtn content was reduced to 1% of total phospholipid. This result suggests that with appropriate genetic manipulation PtdPrn might be able to replace the entire PtdCho and PtdEtn pools. Under physiological conditions, unsaturated PtdCho, PtdSer, and PtdIns all form bilayers when hydrated, but PtdEtn and PtdPrn form a hexagonal phase in vitro (8 -10). These hexagonal phase structures have been proposed to be important for protein translocation, membrane fluidity, and membrane fusion events (11)(12)(13).
In this study we applied genetic and biochemical approaches to determine whether: 1) PtdCho and its N-methylated phospholipid precursors could be eliminated from cells without loss of viability; 2) PtdCho could also be eliminated for growth on both a fermentable carbon source (glucose) and a nonfermentable carbon source (requiring functional mitochondria); and 3) PtdCho depletion was accompanied by a shift to more saturated species of primary amine-containing phospholipids. The results demonstrate that PtdCho and Nmethylated phospholipids are nonessential in yeast and are not required for mitochondrial function.

EXPERIMENTAL PROCEDURES
Construction of Methylation-defective Strains-Deletion mutants for the PEM2 gene were generated by homologous recombination in the RYY52 strain (Mat␣, trp1, lys2, ura3, leu2, his3, suc2, psd1⌬-1::TRP1) (14) by transformation with PCR products harboring a KanMX sequence. These PCR products were made by using primers identical to a * This work was supported by National Institutes of Health Grants 2R37 GM32453 (to D. R. V.) and R24 GM067877 (to R. C. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
24-nucleotide sequence, 541 bp upstream of PEM2, and a 22-nucleotide sequence, 545 bp downstream of PEM2. Chromosomal DNA templates for the PCR were prepared from a pem2⌬ strain (Research Genetics, Inc.) in which the entire PEM2 coding sequence was deleted and replaced by KanMX. The resulting psd1⌬ pem2⌬ strain was isolated by selection for G418 resistance. Correct replacement of PEM2 by the KanMX sequence was confirmed by PCR analysis using primers outside of the target sequence and a primer specific to the KanMX sequence. Disruption of the PEM1 gene was accomplished by homologous recombination of the psd1⌬ pem2⌬ strain using transformation with PCR products containing a 3.2-kb pem1::LEU2 cassette. The pem1::LEU2 module was amplified by PCR from the pem1 disruption strain ESXI-1(4d) obtained from Dr. Susan Henry (Cornell University). The disruption of the PEM1 gene was verified by PCR using one primer within LEU2 and the other outside of the target locus. Strains with a functional Psd1p were generated from pem1 pem2⌬ psd1 mutant strains by transformation with a YCp50-PSD1 plasmid.
Growth Measurement of Lipid Methylation-defective Strains-The methylation-defective strains (pem1::LEU2 pem2⌬::KanMX or pem1::LEU2 pem2⌬::KanMX psd1⌬::TRP1) were pre-grown at 30°C on synthetic complete medium plates containing Prn (20 mM) to dilute out cellular PtdCho. The cells were next inoculated into synthetic complete lactate or glucose medium and washed twice with the same medium. Cells were re-inoculated at an A 600 of 0.005 in synthetic glucose or 0.02 in synthetic lactate medium plus or minus 20 mM Etn, 20 mM Prn, or 2 mM Cho. Cell growth at 30°C was monitored by culture absorbance at 600 nm.
Lipid Phosphorus Measurement-Lipid methylation-defective strains were grown to mid-log phase at 30°C in synthetic glucose or semi-synthetic lactate (2% lactate and 0.05% glucose) media plus Prn (20 mM) or Cho (2 mM). The cells were harvested by centrifugation and washed twice with water. Mitochondria were prepared as described previously (15). The lipids were extracted from whole cells and mitochondria as described previously (16). The phospholipids were separated by two-dimensional TLC on Silica 60 plates using chloroform/ methanol/ammonium hydroxide (65:35:5, v/v/v) followed by chloroform/ acetic acid/methanol/water (75:25:5:2.2, v/v/v). Lipids were visualized with iodine vapor and quantified by measuring phosphorus (17). The results are shown as the percentage of total lipid phosphorus in each phospholipid fraction. Data are means Ϯ S.D. for two or three independent experiments.
Mass Spectrometry-Lipids extracted from the cells grown in synthetic complete medium with glucose plus Prn (20 mM) or Cho (2 mM) were separated by two-dimensional TLC as described above. Lipid spots corresponding to PtdPrn, PtdEtn, or PtdCho were scraped into glass tubes and extracted from the silica gel by the method of Bligh and Dyer (18). Samples dissolved in methanol were analyzed with a PE Sciex API 3000 triple quadrupole mass spectrometer configured to perform multiple experiments during a single flow injection and by using electrospray ionization. Tandem mass spectrometry was performed in the positive ion mode with a first scan event corresponding to the detection of all precursor ions for m/z 184.3, an ion diagnostic for individual molecular species of PtdCho (19). The next scan event involved configuring the tandem mass spectrometer to detect the loss of 141.1 daltons, a neutral loss characteristic of the PtdEtn molecular species (20). The next scan event detected the neutral loss of 155.1 daltons indicative of PtdPrn molecular species, as the homologous ion to the constant neutral loss of 141 for PtdEtn. The instrument next acquired a full mass spectrum of positive ions from m/z 650 -850. The final scan event was in the negative ion mode scanning the mass range from m/z 650 -850. In separate experiments phospholipid classes were separated by TLC, eluted from the silica gel, and examined by full scan negative ion flow injection analysis. Anions present in the negative ion full scan were further analyzed in the subsequent injection of the sample for collisioninduced dissociation of [M Ϫ H] Ϫ or [M Ϫ 15] Ϫ in order to determine the individual fatty acyl components present in each molecular species, using conditions described previously (21). We also conducted experiments to rule out the possibility that the Prn used in these studies was contaminated by Cho. The Prn was analyzed by electrospray mass spectrometry that was sensitive for both the presence of Prn ([M ϩ H] ϩ , m/z 76) as well as Cho ([M ϩ H] ϩ , m/z 104). There was no evidence for the occurrence of Cho in the commercial Prn at a level less than 1 ppm as evidenced by a lack of any ion at m/z 104 when the signal for m/z 76 was 4.4 ϫ 10 5 intensity units. An equal quantity of choline analyzed after the propanolamine generated 3.2 ϫ 10 6 intensity units, thus having a higher detectability as expected for this quaternary amine.

Prn Supplementation of Methylation-defective Mutants Supports Growth on Glucose and Completely Depletes PtdCho-
Methylation-defective strains that have deletion/disruption in PEM1 and PEM2 genes cannot make PtdCho and N-methylated lipids (4). These mutant cells were grown in the presence of Prn to enable them to make PtdPrn via the Kennedy pathway as shown in Fig. 2. The mutant strains grew on minimal glucose agar media supplemented with Prn. However, when lipids extracted from these cultures were separated by twodimensional TLC and stained with iodine vapor, there was a minor spot in the PtdCho region. This minor component was identified as PtdCho by mass spectrometry. Quantification of this PtdCho demonstrated that it comprised 2-3% of total phospholipids (data not shown). Consistent with previous findings (8), this result demonstrates marked reduction of PtdCho can occur without affecting cell viability. These data also reveal that trace amounts of choline can be derived from agar plates. Because our goal was to deplete all PtdCho and N-methylated precursors from these yeast strains, all subsequent growth analyses were performed in liquid media.
In liquid medium, yeast mutants (pem1 pem2⌬) lacking the FIG. 3. Prn supplementation of methylation-defective mutants supports growth on glucose medium. The pem1 pem2⌬ mutant strains were inoculated into synthetic glucose medium and washed twice with the same medium. Cells were re-inoculated at an absorbance of 0.005 in synthetic glucose medium plus or minus 20 mM Etn, 20 mM Prn, or 2 mM Cho. Cell growth at 30°C was monitored by A 600 . The data are from a representative one of four experiments.

FIG. 4. Methylation-defective mutants supplemented with Prn accumulate PtdPrn with depletion of Ptd-Cho.
A, the pem1 pem2⌬ mutant strains were grown to mid-log phase at 30°C in synthetic glucose medium plus Prn (20 mM) or Cho (2 mM). Cells were harvested by centrifugation and washed twice with water. Lipids were extracted from whole cells and separated by two-dimensional TLC on Silica 60 plates. The lipids were visualized with iodine vapor. The arrow in the right panel indicates the location expected for PtdCho, which is not detectable. B, lipid spots from the TLC plates were quantified by measuring phosphorus. Results are expressed as the percentage of total lipid phosphorus in each phospholipid fraction. Values shown are means Ϯ S.D. for three experiments. The total amount of phospholipids per mg of wet cell weight was 8.00 Ϯ 0.75 nmol/mg for Prn-supplemented cells and 6.64 Ϯ 0.74 nmol/mg for Cho-supplemented cells, respectively. http://www.jbc.org/ plemented either with Etn, Prn, Cho, or no phospholipid precursor (Fig. 3). Unsupplemented cultures or those receiving Etn failed to support growth of the mutant strain. As expected, the strain grew normally when supplied with Cho. In the presence of Prn the cultures grew with a pronounced lag but reached nearly the same saturation level obtained with Cho supplementation. Lipids extracted from the cells grown on either Cho or Prn medium were separated by two-dimensional TLC (Fig. 4A). In the presence of Cho, PtdCho was recovered as the major phospholipid. In the absence of Cho but the presence of Prn, the lipid composition of the lipid methylation-defective mutant showed dramatic changes. The PtdCho spot disappeared, and a novel lipid, PtdPrn, was detected. Quantification of each lipid extracted from the TLC plates also confirmed the above observation (Fig. 4B). PtdCho was reduced to the limits of detection. The PtdIns pool more than doubled, and PtdPrn accounted for 30% of the total phospholipid. Furthermore, with Prn supplementation, the hexagonal phase lipids PtdEtn and PtdPrn comprised 58% of the total phospholipid.

Mass Spectrometry Analysis Reveals That There Is No Marked Shift in the Degree of Unsaturation in PtdEtn and
PtdPrn-The dramatic increase of total hexagonal phase lipids (PtdEtn and PtdPrn) in the mutants supplemented with Prn might result in fatty acid compositional changes to make a compensatory shift in the degree of saturation of these lipids toward more bilayer-forming species. To test this idea the individual molecular species of PtdEtn, PtdPrn, and PtdCho were analyzed by negative ion mass spectrometry. A relatively simple mixture of molecular species was found in yeast grown in the presence of choline, because only four major molecular species of PtdEtn were observed at m/z 686, 688, 714, and 716 (Fig. 5A). The identity of these molecular species was determined after collisional activation of each of these ions, which yielded the carboxylate anions corresponding to each fatty acyl substituent (Fig. 5A, inset).  (Fig. 5B). A slightly more abundant species was observed, corresponding to 16:0/16: 1-PtdCho relative to the corresponding molecular species in the PtdEtn class. When cells were incubated with propanolamine, there was no change in the relative molecular species distribution within the phosphatidylethanolamine species either at the level of [M Ϫ H] Ϫ ion abundances or their collision-induced decomposition products (Fig. 5C). New PtdPrn molecular species were now observed at m/z 700, 702, 728, and 730 which corresponded to 16:1/16:1-PtdPrn, 16:0/16:1-PtdPrn, 16:1/18:1-PtdPrn, and 16:1/18:0-PtdPrn plus 16:0/18:1-PtdPrn, respectively (Fig. 5D). The identities of these were confirmed again by tandem mass spectrometry (inset, Fig. 5D). This direct determination of the individual molecular species and fatty acyl substituents (Fig. 5) demonstrated that the saturated fatty acyl species were not increased in the PtdPrn and PtdEtn classes of the mutant supplemented with Prn.

Prn Supplementation Supports Growth of Methylation-defective Mutants and Eliminates the PtdCho Pool Under
Respiratory Conditions-Functional mitochondria are required when yeast cells grow on nonglycolytic carbon sources and must use respiration for ATP generation. We analyzed cell growth under respiratory conditions to determine whether PtdPrn could support mitochondrial function. For this analysis, the methylationdefective mutants were grown on minimal lactate media either without supplementation or with Etn, Prn, or Cho (Fig. 6). Both unsupplemented cultures and those provided with Etn failed to support the growth of methylation-defective strains. In contrast, the mutants grew normally when supplied with Cho. In addition, after a significant lag, the mutants also grew when supplemented with Prn.
Lipids extracted from the cells grown on minimal lactate medium supplemented either with Cho or Prn were separated by two-dimensional thin layer chromatography. As with glucose grown cells, the mutants grown in minimal lactate medium supplemented with Prn showed dramatic alterations in lipid composition. Quantification of phospholipids resolved on TLC plates, by phosphorus assay (Fig. 7), revealed that PtdCho levels declined precipitously, to the limits of detection. The loss of PtdCho was accompanied by a modest increase in PtdIns and the appearance of PtdPrn as greater than 20% of the total phospholipid. Under these conditions the hexagonal phase lipids PtdEtn and PtdPrn accounted for more than 60% of the phospholipid pool.
We also examined the consequences of the methyltransferase mutations and Prn supplementation upon the lipid composition of isolated mitochondria. The results of this analysis are shown in Fig. 8 and demonstrate that the mitochondrial phospholipid composition closely parallels that found for the whole cell. The findings clearly demonstrate that PtdCho is not essential for mitochondrial function and that the organelle can tolerate a hexagonal phase lipid content of greater than 60% of its total phospholipid.  (8,22) has demonstrated that mitochondrial PtdEtn generated by Psd1p is critical for the respiratory growth of yeast strains. We performed experiments to test if mitochondrial PtdEtn as well as PtdCho can be replaced by PtdPrn (Fig. 9). We used triple mutant strains defective in PEM1 and PEM2 for phospholipid methylation and defective in PSD1 for PtdEtn synthesis in mitochondria. The mutation in the PSD1 gene results in a decrease in the PtdEtn pool in mitochondria. The triple mutants were grown on minimal glucose media either lacking supplements or with Etn, Prn, or Cho (Fig. 10). Cultures of the triple mutant without supplementation or with Etn failed to grow. In contrast, the strain grew normally when supplied with Cho. The strain also grew with Prn supplementation, showing a pronounced lag. The results clearly show that Prn-supplemented mutants still can grow despite genetic elimination of PtdCho synthesis and mitochondrial PtdEtn synthesis.

Cells Lacking PtdCho and the Ability to Synthesize Mitochondrial PtdEtn Remain Viable with Prn Supplementation on Glucose-containing Medium-Previous research
Lipids extracted from the cells grown on either Cho-or Prncontaining media were separated by two-dimensional TLC. When the lipids were visualized by iodine staining the triple mutant strain grown on Prn lacked detectable PtdCho and accumulated PtdPrn (Fig. 11A). In addition, the PtdIns spot appeared much more prominent in the triple mutant supplemented with Prn when compared with this strain supplemented with choline. Quantification of the phospholipids (Fig.  11B) demonstrated the elimination of PtdCho and a 4-fold increase in PtdIns, and an accumulation of PtdPrn to 33% of the total phospholipid pool. Unlike Cho-supplemented cells, the Prn-supplemented cells did not accumulate relatively high levels of PtdSer.
We also isolated mitochondria from glucose-grown cells harboring pem1 pem2⌬ psd1⌬ mutations and quantified the phospholipids of these organelles. The results shown in Fig. 11C demonstrate that choline supplementation of the triple mutant produces mitochondria in which PtdCho is the most abundant phospholipid (44.3%), and PtdIns, PtdEtn, and PtdSer are nearly equivalent, ranging from 12 to 17% of the total phospholipids. As expected for a strain lacking Psd1p, the PtdEtn content of the organelle is reduced by 49% when compared with the level of this phospholipid in total cell membranes (see Fig.  11B). Analysis of the mitochondria derived from the triple mutant supplemented with 2 mM Prn demonstrates the or- ganelles are devoid of PtdCho and contain PtdPrn as the most abundant phospholipid (35.2%). These mitochondria also contain relatively high levels of PtdIns (30.1%) and lower levels of PtdEtn (18.7%) and PtdSer (8%).
We next sought to determine whether PtdPrn could support the growth of the triple mutant under respiratory conditions. The results of these experiments are shown in Fig. 12. Cells lacking both mitochondrial PtdCho and the pool of PtdEtn generated by Psd1p fail to grow with Prn supplementation under conditions requiring functional mitochondria. These findings demonstrate that PtdPrn is unable to fulfill both Ptd-Cho and PtdEtn requirements for mitochondrial function during respiration.

DISCUSSION
In experiments described in this report we tested the essentiality of PtdCho and other methylated forms of PtdEtn for the growth and mitochondrial function of S. cerevisiae. In many well described eukaryotes typically used for biochemical experiments, PtdCho is the most abundant phospholipid and is often assumed to be essential for membrane structure and function and cell growth (1,3,11). The results of our studies provide clear evidence that neither PtdCho, PtdEtn(Me), nor Ptd-Etn(Me) 2 is required for either cell viability or mitochondrial function in yeast.
Our general approach to examining the role of methylated forms of PtdEtn, including PtdCho, in supporting cell growth was to use the Etn analog Prn. Prn, like Etn, contains a primary amine. Previous studies have shown that Prn is readily incorporated into phospholipid in both Saccharomyces and Tetrahymena species (8,23). The physical properties of PtdPrn are similar to those of PtdEtn. Most notably, PtdPrn forms hexagonal H II phase structures. NMR studies reveal that PtdPrn has a higher propensity to form hexagonal phase structures than PtdEtn (8). Although PtdPrn forms hexagonal phase structures, it cannot fully substitute for the cellular requirements for PtdEtn in yeast. In addition, in earlier studies examining PtdPrn metabolism, we demonstrated that this novel lipid replaced much of the PtdCho present in cells and simultaneously enabled cells to survive with drastically reduced levels of PtdEtn, created by mutations in PtdSer decarboxylases (8). These observations led to the initial experiments described in this study designed to test whether PtdPrn formation would be permissive for cell growth in the absence of PtdCho synthesis.
The experiments described in Figs. 3 and 6 demonstrate that Prn supplementation enables strains harboring deletions in PEM1 and PEM2 genes to grow, albeit at a slower initial rate than the same cells supplemented with Cho. Analysis of the lipid composition of these cells reveals several interesting findings. Unexpectedly, cells grown on minimal agar plates lacking choline, but supplemented with Prn, are able to synthesize low levels of PtdCho. This result indicates that agar contains trace FIG. 9. Strategy for elimination of PtdCho and reduction of mitochondrial PtdEtn using Prn supplementation. In strains harboring pem1 pem2⌬ psd1⌬ mutations, PtdEtn methylation in the endoplasmic reticulum is blocked, and PtdEtn formation in the mitochondria is blocked. Prn enters the Kennedy pathways and is converted to PtdPrn.
FIG. 10. Cells lacking PtdCho and with reduced mitochondrial PtdEtn remain viable with Prn supplementation in glucose medium. The pem1 pem2⌬ psd1 mutant strains were pregrown at 30°C on synthetic lactate plates containing Prn (20 mM) to dilute out cellular PtdCho. Cells were inoculated into synthetic glucose medium and washed twice with the same medium. Cells were re-inoculated at an absorbance of 0.005 in synthetic glucose either without supplements or plus 20 mM Etn, 20 mM Prn, or 2 mM Cho. Cell growth at 30°C was monitored by A 600 . The data are from a representative one of four experiments. levels of Cho that the cells are able to successfully scavenge. With growth in liquid medium, no PtdCho was formed above the background levels of our assay system. These data clearly demonstrate that yeast does not require PtdCho or the monoor di-methylated forms of PtdEtn for growth. Previous work by Summers et al. (4) demonstrated that yeast cells could grow moderately well with PtdEtn(Me) 2 replacing PtdCho and poorly with PtdEtn(Me) replacing PtdCho but implicated a specific need for at least some methylation of PtdEtn as essential for growth. Additional work in mammalian cells also provided evidence that PtdEtn(Me) 2 could support growth of mammalian cells containing greatly diminished levels of PtdCho   FIG. 11. Methylation-defective mutants, also lacking mitochondrial Psd1p activity, supplemented with Prn accumulate PtdPrn with depletion of PtdCho. A, the pem1 pem2⌬ psd1 mutant strains were grown to mid-log phase at 30°C in synthetic glucose media plus Prn (20 mM) or Cho (2 mM). Cells were harvested by centrifugation and washed twice with water. Lipids were extracted and separated by two-dimensional TLC on Silica 60 plates. Lipids were visualized with iodine vapor. The arrow in the right panel indicates the location expected for PtdCho, which is not detectable. B, lipid spots were recovered from TLC plates described in A and quantified by measuring phosphorus. Results are the percentage of total lipid phosphorus in each phospholipid fraction. Values shown are means Ϯ S.D. for two experiments. The total amount of phospholipids per mg of wet cell weight was 9.42 Ϯ 0.03 nmol/mg for Prn-supplemented cells and 7.74 Ϯ 0.43 nmol/mg for Cho-supplemented cells, respectively. C, the mitochondrial fractions were isolated from the pem1 pem2⌬ psd1 mutant strains grown to mid-log phase at 30°C in synthetic glucose plus Prn (20 mM) or Cho (2 mM). Lipids were extracted from mitochondria and separated by two-dimensional TLC on Silica 60 plates. Lipid classes were visualized with iodine vapor and quantified by measuring phosphorus. Results are the percentage of total lipid phosphorus in each phospholipid fraction. Values shown are means Ϯ S.D. for three experiments. The total amount of phospholipids per mg of protein was 281.54 Ϯ 55.76 nmol/mg for Prn-supplemented cells and 302.07 Ϯ 88.51 nmol/mg for Cho-supplemented cells, respectively. (24). However, the complete replacement of PtdCho and all N-methylated forms of PtdEtn by the primary amine-containing and hexagonal phase-forming PtdPrn was quite surprising.
We extended these studies to more restrictive growth conditions by examining methyltransferase-defective strains under respiratory growth conditions. As with cells grown on glucose, PtdPrn formation permitted growth in the absence of PtdCho and other N-methylated phospholipids when functional mitochondria were required. These results permit a number of additional important conclusions. Most importantly, mitochondrial maturation and function does not require PtdCho or methylated PtdEtn. In addition, the analysis of mitochondrial phospholipids shown in Fig. 8 demonstrates that PtdPrn is readily incorporated into this organelle as well as other membranes. We currently have preliminary data with choline kinase and ethanolamine kinase mutants that implicate the formation of phospho-Prn and CDP-Prn in the formation of PtdPrn. These latter findings are consistent with choline phosphotransferase (25) and ethanolaminephosphotransferase (26) acting as the terminal transferases in the synthesis of PtdPrn in the endoplasmic reticulum. Thus, it is highly likely that PtdPrn must be synthesized in the endoplasmic reticulum and Golgi (27) and be transported to the mitochondria. Consequently, the phospholipid transport systems that exist must be insensitive to the structural anomaly of PtdPrn. Another important finding from these experiments is that the mitochondria remain fully functional with as much as 70% of their complement of phospholipid containing primary amines.
Since earlier studies (8) demonstrated that PtdPrn synthesis allowed for dramatic reduction of cellular PtdEtn content, we also tested whether PtdPrn could eliminate and replace the PtdEtn pool of mitochondria. Cells harboring the triple mutations pem1 pem2⌬ psd1⌬ are unable to synthesize PtdCho by methyltransferase action and mitochondrial PtdEtn by Psd1p action. In glucose medium these mutant strains grow with a pronounced lag but reach normal levels of saturation when supplemented with Prn. As expected, PtdPrn replaced PtdCho, and the presence of the psd1⌬ mutation had little effect upon total cellular PtdEtn content. In psd1⌬ strains grown without Etn, PtdEtn formation depends mostly upon the activity of Psd2p and to a lesser extent upon the breakdown of sphingolipids and the action of Dpl1p (7,8). However, PtdEtn formation and content in the mitochondria is also dependent upon the catalysis performed by Psd1p. Shifting the triple mutant cells to lactate as the carbon source requires that they possess functional mitochondria for growth. The growth curve shown in Fig. 12 reveals that the triple mutant cannot grow with Prn supplementation but can grow with Cho supplementation. Thus PtdPrn can replace mitochondrial PtdCho only when there is adequate PtdEtn present in the organelle. The PtdPrn appears unable to replace both PtdCho and PtdEtn pools within the mitochondria. These findings are consistent with other observations indicating that PtdEtn plays an essential role in mitochondrial membrane structure and function during respiration (8,22).
In all conditions described in this report, the supplementation of cells with Prn resulted in elevated levels of PtdIns when compared with cells supplemented with Cho. This consistent finding was most pronounced when analyzing the phospholipids of the pem1 pem2⌬ psd1⌬ triple mutant. The elevation of membrane PtdIns content may reflect the biophysical need to replace some of the eliminated PtdCho with the bilayer forming PtdIns to counteract the propensity of PtdPrn and PtdEtn to form nonbilayer structures (8).
In summary, this work demonstrates that it is possible to alter dramatically the phospholipid composition of yeast and its mitochondria with a combination of genetic and biochemical manipulations. PtdCho, the major lipid of many eukaryotes, can be eliminated along with other N-methylated forms of PtdEtn, indicating that these phospholipids are nonessential. The dispensability of PtdCho is true under both respiratory and nonrespiratory growth conditions. The pem1 pem2⌬ psd1 mutant strains were pre-grown at 30°C on synthetic lactate plates containing Prn (20 mM) to dilute out cellular PtdCho. Cells were inoculated into synthetic lactate medium and washed twice with the same medium. Cells were re-inoculated at A 600 of 0.02 in synthetic lactate medium either without supplementation or 20 mM Etn, 20 mM Prn, or 2 mM Cho. Cell growth at 30°C was monitored by A 600 . The data are from a representative one of three experiments.

Saccharomyces cerevisiae
-Methylated Phospholipids Are Nonessential in N Phosphatidylcholine and