Metabolism and regulation of glycerolipids in the yeast Saccharomyces cerevisiae.

Due to its genetic tractability and increasing wealth of accessible data, the yeast Saccharomyces cerevisiae is a model system of choice for the study of the genetics, biochemistry, and cell biology of eukaryotic lipid metabolism. Glycerolipids (e.g., phospholipids and triacylglycerol) and their precursors are synthesized and metabolized by enzymes associated with the cytosol and membranous organelles, including endoplasmic reticulum, mitochondria, and lipid droplets. Genetic and biochemical analyses have revealed that glycerolipids play important roles in cell signaling, membrane trafficking, and anchoring of membrane proteins in addition to membrane structure. The expression of glycerolipid enzymes is controlled by a variety of conditions including growth stage and nutrient availability. Much of this regulation occurs at the transcriptional level and involves the Ino2-Ino4 activation complex and the Opi1 repressor, which interacts with Ino2 to attenuate transcriptional activation of UAS(INO)-containing glycerolipid biosynthetic genes. Cellular levels of phosphatidic acid, precursor to all membrane phospholipids and the storage lipid triacylglycerol, regulates transcription of UAS(INO)-containing genes by tethering Opi1 to the nuclear/endoplasmic reticulum membrane and controlling its translocation into the nucleus, a mechanism largely controlled by inositol availability. The transcriptional activator Zap1 controls the expression of some phospholipid synthesis genes in response to zinc availability. Regulatory mechanisms also include control of catalytic activity of glycerolipid enzymes by water-soluble precursors, products and lipids, and covalent modification of phosphorylation, while in vivo function of some enzymes is governed by their subcellular location. Genome-wide genetic analysis indicates coordinate regulation between glycerolipid metabolism and a broad spectrum of metabolic pathways.

THE yeast, Saccharomyces cerevisiae, has emerged as a powerful model system for the elucidation of the metabolism, cell biology, and regulation of eukaryotic lipids. Due to the strong homology of yeast proteins, pathways, and regulatory networks with those in higher eukaryotes, yeast has provided numerous insights into the genetics and biochemistry of lipid-related diseases. As a system for the study of eukaryotic lipid metabolism, the advantages of yeast include its vast, well-curated, and electronically accessible archives of genetic data, including those detailing gene–enzyme relationships in the pathways for lipid synthesis and turnover. Another major advantage is the rapidly increasing understanding of the regulation and localization of enzymes and the movement of lipids within and among cellular membranes and compartments in yeast.

This article presents an overview of progress in elucidating gene–enzyme relationships, cellular localization, and regulatory mechanisms governing glycerolipid metabolism in yeast. The metabolism covered in this YeastBook chapter includes the regulation, synthesis, and turnover of phospholipids and triacylglycerol (TAG) and their precursors in the context of changing growth conditions and nutrient availability. All glycerophospholipids in yeast, including the major phospholipids, phosphatidylinositol (PI), phosphatidylserine (PS), phosphatidylethanolamine (PE), and phosphatidylcholine (PC), are derived from the precursor lipid, phosphatidic acid (PA) (Figures 1 and 2). A major topic of this review article is the tremendous recent progress in elucidating the complex regulatory mechanisms that control the connected and coordinated pathways involved in the synthesis of glycerophospholipids and TAG, which is also derived from PA (Figures 2, 3, and 4). The regulation of glycerolipid metabolism occurs at many levels and is a major topic discussed in this article. Adding complexity to analysis of the regulatory networks connected to lipid metabolism is the fact that critical signals controlling this regulation arise during the ongoing biosynthesis and turnover of the lipids themselves and involve precursors and metabolites embedded in the metabolism.

Figure 1 

Phospholipid structures. The diagram shows the structures of the phospholipid PA and the major phospholipids PI, PS, PE, and PC that are derived from PA. The hydrophilic head groups (H, inositol, serine, ethanolamine, and choline) that are attached to the basic phospholipid structure are shown in red. The four most abundant fatty acids esterified to the glycerol-3-phosphate backbone of the phospholipids are palmitic acid, palmitoleic acid, steric acid, and oleic acid. The type and position of the fatty acyl moieties in the phospholipids are arbitrarily drawn. The relative amounts of the phospholipids as well as their fatty acyl compositions vary depending on strain (e.g., mutation) and growth condition.

Figure 2 

Pathways for the synthesis of glycerolipids and their subcellular localization. Phospholipids and TAG share DAG and PA as common precursors. In the de novo synthesis of phospholipids, PA serves as the immediate precursor of CDP-DAG, precursor to PI, PG, and PS. PS is decarboxylated to form PE, which undergoes three sequential methylations resulting in PC. PA also serves as a precursor for PGP, PG, and ultimately CL, which undergoes acyl-chain remodeling to the mature lipid. Alternatively, PA is dephosphorylated, producing DAG, which serves as the precursor of PE and PC in the Kennedy pathway. DAG also serves as the precursor for TAG and can be phosphorylated, regenerating PA. The names of the enzymes that are discussed in detail in this YeastBook chapter are shown adjacent to the arrows of the metabolic conversions in which they are involved and the gene–enzyme relationships are shown in Tables 1–3. Lipids and intermediates are boxed, with the most abundant lipid classes boxed in boldface type. Enzyme names are indicated in boldface type. The abbreviations used are: TAG, triacylglycerols; PI, phosphatidylinositol; PA, phosphatidic acid; CDP-DAG, CDP-diacylglycerol; DAG, diacylglycerol; MAG, monoacylglycerol; Gro, glycerol; DHAP, dihydroxyacetone phosphate, PS, phosphatidylserine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PGP phosphatidylglycerol phosphate; CL* precursor cardiolipin; MLCL, monolyso-cardiolipin; CL, mature cardiolipin; PMME, phosphatidylmonomethylethanolamine; PDME, phosphatidyl-dimethylethanolamine; PC, phosphatidylcholine; FFA, free fatty acids; Cho, choline, Etn, ethanolamine, Ins, inositol; Cho-P, choline phosphate; CDP-Cho, CDP-choline; Etn-P, ethanolamine phosphate; CDP-Etn, CDP-ethanolamine; PI 3-P, phosphatidylinositol 3-phosphate; PI 4-P, phosphatidylinositol 4-phosphate; PI 4,5-P2, phosphatidylinositol 4,5-bisphosphate; PI 3,5-P2, phosphatidylinositol 3,5-bisphosphate. Nucl, nucleus; ER, endoplasmic reticulum; Mito, mitochondria; LD, lipid droplets; G/E/V, Golgi, endosomes, vacuole; Pex, peroxisomes; Cyt, cytoplasma; PM, plasma membrane. CL* indicates a precursor of cardiolipin (CL) with saturated acyl-chain that undergoes deacylation/reacylation to mature CL. See text for details.

Figure 3 

Model for PA-mediated regulation of phospholipid synthesis genes. (A) Growth conditions (e.g., exponential phase, inositol depletion, or zinc supplementation) under which the levels of PA are relatively high, the Opi1 repressor is tethered to the nuclear/ER membrane, and UASINO-containing genes are maximally expressed (boldface arrow) by the Ino2-Ino4 activator complex. (Inset) Localization of Opi1, fused with GFP at its C-terminal end and integrated into the chromosome, being expressed under its own promoter in live cells growing logarithmically in synthetic complete medium lacking inositol (−Ins) and analyzed by fluorescence microscopy. (B) Growth conditions (e.g., stationary phase, inositol supplementation, or zinc depletion) under which the levels of PA are reduced, Opi1 dissociates from the nuclear/ER membrane, and enters into the nucleus where it binds to Ino2 and attenuates (thin arrow) transcriptional activation by the Ino2–Ino4 complex. (Inset) Localization of Opi1, as described in A, except that the cells are growing logarithmically in medium containing 75 μM inositol. PA level decreases by the stimulation of PI synthesis in response to inositol (Ins) supplementation and by Zap1-mediated induction of PIS1, that results in an increase in PI synthesis in response to zinc depletion. The regulation in response to zinc depletion and stationary phase occurs without inositol supplementation. Pah1 and Dgk1 play important roles in controlling PA content and transcriptional regulation of UASINO-containing genes. The synthesis of TAG (which is stored in lipid droplets, LD) and phospholipids (with the exception of PE, which occurs in the mitochondria and Golgi) occurs in the ER. Fluorescence microscopy images of Opi1 localization courtesy of Yu-Fang Chang, Henry Laboratory, Department of Molecular Biology and Genetics, Cornell University, Ithaca, NY.

Figure 4 

Regulation of phospholipid synthesis by soluble lipid precursors and metabolites. The diagram shows the major steps in the synthesis of phospholipids. The enzymes that are biochemically regulated by phospholipid precursors and products are shown. The green arrow designates the stimulation of enzyme activity, whereas the red line designates the inhibition of enzyme activity. Details on the biochemical regulation of these enzymes are discussed in the section Water soluble precursors of phospholipids, metabolism, and regulatory roles and in Carman and Han (2009a). See Figure 2 for abbreviations.

Table 1 

Glycerolipid synthesis enzymes

Gene	Ino− or Opi− phenotype	Enzyme	Mol mass (kDa)	Isoelectric point	Molecules per cellb	Locationc	Transmembrane domains	Phosphorylation sitesd
SCT1 (GAT2)	Ino−	Glycerol-3-P /dihydroxyacetone-P acyltransferase	85.7	7.27	1050	ER	4	Few
GPT2 (GAT1)	—	Glycerol-3-P /dihydroxyacetone-P acyltransferase	83.6	10.3	3100	ER, lipid droplets	4	Several
AYR1	—	Acyl DHAP reductase	32.8	9.92	3670	ER, lipid droplets	None	None
SLC1	—	LysoPA/Acylglycerol-3-P acyltransferase	33.8	10.41	ND	ER, lipid droplets	1	None
ALE1 (SLC4,LPT1,LCA1)	—	LysoPA/Acylglycerol-3-P acyltransferase	72.2	10.3	ND	ER	7	Several
PSI1 (CST26)	—	LysoPI acyltransferase	45.5	10.15	2010	ER	4	None
TAZ1	—	LysoPC acyltransferase/ monolysoCL acyltransferase	44.2	9.38	1340	Mitochondria	None	None
CDS1 (CDG1)a	Opi−	CDP-DAG synthase	51.8	8.64	ND	ER, mitochondria	6	Few
CHO1 (PSS1)a	Opi−	PS synthase	30.8	6.23	ND	ER	2	Several
PSD1a	Opi−	PS decarboxylase	56.6	9.84	1080	Mitochondria	None	None
PSD2	—	PS decarboxylase	130	7.85	ND	Vacuole, endomembranes	None	Few
CHO2 (PEM1)a	Opi−	PE methyltransferase	101.2	8.56	1810	ER	8	Few
OPI3 (PEM2)a	Opi−	Phospholipid methyltransferase	23.1	9.6	5890	ER, mitochondria	None	None
PAH1 (SMP2)	—	PA phosphatase	95	4.68	3910	Cytoplasm, ER	None	Several
DGK1 (HSD1)	—	DAG kinase	32.8	9.48	784	ER	4	Few
EKI1a	—	Ethanolamine kinase	61.6	5.69	3420	Cytoplasm	None	Few
ECT1 (MUQ1)	—	Ethanolamine-P cytidylyltransferase	36.8	6.44	4700	Cytoplasm	None	None
EPT1a	—	Ethanolamine/choline phosphotransferase	44.5	6.5	ND	ER	7	None
CKI1a	—	Choline kinase	66.3	5.43	3930	Cytoplasm	None	Several
PCT1 (CCT1,BSR2)a	Opi−	Choline-P cytidylyltransferase	49.4	9.26	3050	Cytoplasm, nucleus	None	Several
CPT1a	—	Choline phosphotransferase	44.8	6.57	981	Membrane	8	None
PGS1 (PEL1)	—	PGP synthase	59.3	10.5	ND	Mitochondria	None	None
GEP4	—	PGP phosphatase	20.9	9.18	ND	Mitochondria	None	None
CRD1 (CLS1)	—	CL synthase	32	10.55	876	Mitochondria	3	None
PIS1	Ino−	PI synthase	24.8	8.92	3810	ER	3	None
LSB6	—	PI 4-kinase	70.2	6.68	57	Plasma membrane, vacuole membrane	None	None
STT4	Ino−	PI 4-kinase	214.6	7.44	846	Plasma membrane	None	Few
PIK1	—	PI 4-kinase	119.9	6.46	1600	Plasma membrane, nucleus, Golgi	None	Several
VPS34 (END12,PEP15,VPL7,VPT29,STT8,VPS7)	Opi−	PI 3-kinase	100.9	7.79	1080	Vacuole	None	None
MSS4	Ino−	PI 4-P 5-kinase	89.3	10.13	ND	Cytoplasm	None	Several
FAB1 (SVL7)	Ino−	PI 3-P 5-kinase	257.4	8.45	149	Vacuole	None	Several
DGA1a	—	Acyl-CoA diacylglycerol acyltransferase	47.7	10.39	907	ER, lipid droplets	1	Few
LRO1a	—	Phospholipid diacylglycerol acyltransferase	75.3	6.67	ND	ER	1	Few
ARE1 (SAT2)a	—	Acyl-CoA sterol acyltransferase	71.6	8.27	ND	ER	9	Several
ARE2 (SAT1)a	—	Acyl-CoA sterol acyltransferase	74.0	7.71	279	ER	9	Several
Phospholipid synthesis regulatory proteins
INO2 (DIE1, SCS1)a	Ino−	Transcriptional activator	34.2	6.23	784	Nucleus	None	None
INO4	Ino−	Transcriptional activator	17.4	10.21	521	Nucleus, cytoplasm	None	None
OPI1	Opi−	Transcriptional repressor	46	4.87	1280	Nuclear/ER membrane	None	Few

Much of the information in the table may be found in the Saccharomyces Genome Database. Ino−, inositol auxotrophy; Opi−, inositol excretion; ND, not determined.

Genes containing the UASINO element and regulated by the Ino2-Ino4-Opi1 circuit. The names in parentheses are aliases.

Ghaemmaghami .

Habeler ; Kumar ; Huh ; Natter .

Li, X. ; Bodenmiller .

Table 3 

Glycerolipid turnover enzymes

Gene	Ino− or Opi−phenotype	Enzyme	Mol mass (kDa)	Isoelectric point	Molecules per cella	Locationbb	Transmembrane domains	Phosphorylation sitesc
CLD1	—	CL specific phospholipase A2	52	10.3	ND	Mitochondria	None	None
NTE1	Ino−	PC specific phospholipase B	187.1	8.19	521	ER	3	Several
PLB1	—	PC/PE specific phospholipase B	71.6	4.36	ND	Plasma membrane, ER, vesicles, extracellular	None	None
PLB2	—	Nonspecific phospholipase B	75.4	4.35	623	Plasma membrane, ER, vesicles	None	None
PLB3	—	Nonspecific phospholipase B	75	4.7	ND	Cytoplasm, vacuole, vesicles	None	None
PLC1	—	PI 4,5-P2 specific phospholipase C	110.5	9.84	ND	Cytoplasm	None	Few
PGC1	—	PG specific phospholipase C	37	8.93	3270	Mitochondria	None	None
SPO14 (PLD1)	—	PC specific phospholipase D	195.2	7.58	49	Cytoplasm	None	Several
SAC1 (RSD1)	Ino−	Nonspecific polyphosphoinositide phosphatase	71.1	7.75	48,000	ER, Golgi, vacuole	2	None
INP51 (SJL1)	—	PI 4,5-P2 phosphatase	108.4	6.7	98	Cytoplasm	None	None
INP52 (SJL2)	—	Nonspecific polyphosphoinositide phosphatase	133.3	8.97	ND	Actin	None	Several
INP53 (SJL3, SOP2)	—	Nonspecific polyphosphoinositide phosphatase	124.6	7.18	1520	Actin	None	Several
INP54	—	PI 4,5-P2 phosphatase	43.8	7.6	1200	ER	None	None
YRM1	—	PI 3-P phosphatase	91	7.19	125	Cytoplasm	None	Few
FIG4	—	PI 3,5-P2 phosphatase	101.7	5.52	339	Vacuole	None	None
DPP1 (ZRG1)	—	DGPP phosphatase/nonspecific lipid phosphate phosphatase	33.5	6.42	3040	Vacuole	5	Few
LPP1	—	Nonspecific lipid phosphate phosphatase	31.5	8.25	300	Golgi	4	None
PHM8	—	LysoPA phosphatase	37.7	5.14	195	Cytoplasm, nucleus	None	None
TGL1	—	Triacylglycerol lipase, sterylester hydrolase	63.0	6.83	1470	ER, lipid droplets	None	Several
TGL2	—	Acylglycerol lipase	37.5	8.41	ND	Mitochondria	None	None
TGL3	—	Triacylglycerol lipase, lysoPA acyltransferase	73.6	8.50	3210	Lipid droplets	1	Few
TGL4	—	Triacylglycerol lipase, Ca++ dependent phospholipase A2, lysoPA acyltransferase	102.7	8.05	195	Lipid droplets	None	Several
TGL5	—	Triacylglycerol lipase, lysoPA acyltransferase	84.7	9.84	358	Lipid droplets	1	Several
YJU3	—	Monoacylglycerol lipase	35.6	8.5	2140	Lipid droplet, ER	None	None

Much of the information in the table may be found in the Saccharomyces Genome Database. The names in parentheses are aliases. Ino−, inositol auxotrophy; ND, not determined.

Ghaemmaghami .

Habeler ; Kumar ; Huh ; Natter .

Li ; Bodenmiller

For example, PA plays a number of signaling roles vital to the regulation of lipid metabolism in yeast (Figure 3), in addition to its function as precursor to all phospholipids and TAG (Figure 2). PI synthesis is regulated in response to its precursor, inositol, on several levels (Figures 3 and 4) and PI also serves as precursor to both phosphatidylinositolphosphates and inositol-containing sphingolipids, both of which are implicated in a wide range of signaling and regulatory activities (Strahl and Thorner 2007; Dickson 2008, 2010), topics that will not be dealt with in detail in this YeastBook chapter. In addition, the enzymes controlling the metabolism of gycerolipids are localized to specific cellular compartments (Figure 2; Tables 1–3), while the lipids themselves are, for the most part, distributed to a much wider range of cellular compartments. Moreover, the regulation of TAG metabolism plays a major role in lipid droplet (LD) formation and depletion (Murphy and Vance 1999; Rajakumari ; Kurat ; Kohlwein 2010a), a topic that will also be addressed in detail in the chapter on Lipid Droplets and Peroxisomes. Thus, detailed knowledge of pathways, gene–enzyme relationships, and subcellular localization of enzymes and pools of lipids and metabolites involved in the synthesis and turnover of glycerolipids (Figures 2–4, Tables 1–3) is essential to the elucidation of the complex mechanisms responsible for their regulation.

Notably, in eukaryotes phospholipids play many vital roles in the biology of the cell that extend beyond lipid metabolism itself. These include roles in membrane trafficking and membrane identity (Vicinanza ) and anchoring of membrane proteins (Roth ; Pittet and Conzelmann 2007; Fujita and Jigami 2008), complex topics in their own right, which will be discussed only in brief in this YeastBook chapter. Phospholipids also serve as signaling molecules and as precursors of signaling molecules (Strahl and Thorner 2007). Thus, the advancements in yeast glycerolipid metabolism discussed in this review article also have enormous potential to contribute critical insights into these vital roles of lipids and lipid-mediated signaling in eukaryotic cells.

Pathways of glycerolipid metabolism

Major glycerolipids of S. cerevisiae include the phospholipids PC, PE, PI, PS (Figure 1), phosphatidylglycerol (PG), and cardiolipin (CL) (Rattray ; Henry 1982; Carman and Henry 1989; Paltauf ; Guan and Wenk 2006; Ejsing ). Minor phospholipids include intermediates such as PA, CDP-diacylglycerol (CDP-DAG), phosphatidylmonomethylethanolamine (PMME), phosphatidyldimethylethanolamine (PDME), the D-3, D-4, and D-5 polyphosphoinositides, and lysophospholipids (Rattray ; Oshiro ; Strahl and Thorner 2007). TAG and diacylglycerol (DAG) are the major neutral glycerolipids. The fatty acids that are commonly esterified to the glycerophosphate backbone of yeast glycerolipids include palmitic acid (C16:0), palmitoleic acid (C16:1), stearic acid (C18:0), and oleic acid (C18:1) (Rattray ; Henry 1982; Bossie and Martin 1989; McDonough ; Martin ). The pathways for the synthesis of phospholipids and TAG are shown in Figure 2. The enzymes and transporters of glycerolipid metabolism and the genes that encode them are listed in Tables 1–3. The gene–protein relationships shown in the tables have been confirmed by the analysis of gene mutations and/or by the biochemical characterization of the enzymes and transporters (Carman and Henry 1989; Greenberg and Lopes 1996; Henry and Patton-Vogt 1998; Carman and Henry 1999; Black and Dirusso 2007; Tehlivets ; Kohlwein 2010b; Carman and Han 2011).

Synthesis and turnover of phospholipids

In the de novo pathways (Figure 2, Table 1), all membrane phospholipids are synthesized from PA, which is derived from glycerol-3-P via lysoPA by two fatty acyl CoA-dependent reactions that are catalyzed in the endoplasmic reticulum (ER) by the - and -encoded glycerol-3-P acyltransferases and the - and -encoded lysoPA/lysophospholipid acyltransferases, respectively (Athenstaedt and Daum 1997; Athenstaedt ; Zheng and Zou 2001; Benghezal ; Chen ; Jain ; Riekhof ). The glycerol-3-P acyltransferase enzymes also utilize dihydroxyacetone-P as a substrate, and the product acyl dihydroxyacetone-P is converted to lysoPA by the lipid droplet (LD) and ER-associated -encoded reductase (Athenstaedt and Daum 2000). At this point in the pathway, PA is partitioned to CDP-DAG, catalyzed by -encoded CDP-DAG synthase (Carter and Kennedy 1966; Kelley and Carman 1987; Shen ) or to DAG, catalyzed by -encoded PA phosphatase (Han ) (Figure 1). CDP-DAG synthase activity has been detected in the ER and in mitochondria (Kuchler ), whereas PA phosphatase is a cytosolic enzyme that must associate with membranes to catalyze the dephosphorylation of PA to produce DAG (Han ; Carman and Han 2009a). CDP-DAG and DAG are used to synthesize PE and PC by two alternative routes, namely, the CDP-DAG and Kennedy pathways (Figure 2). In the CDP-DAG pathway, CDP-DAG is converted to PS by the ER localized -encoded PS synthase (Atkinson ; Letts ; Bae-Lee and Carman 1984; Kiyono ; Nikawa ). Yeast has two PS decarboxylases encoded, by the and genes. Psd1, localized to the inner mitochondrial membrane, accounts for the majority of the enzymatic activity in yeast, while the minor activity, Psd2, associates with Golgi/vacuole (Clancey ; Trotter , 1995; Voelker 2003). PE then undergoes three sequential methylation reactions in the ER (Gaynor and Carman 1990), the first of which is catalyzed by the -encoded PE methyltransferase, while the final two methylations are catalyzed by the -encoded phospholipid methyltransferase (Kodaki and Yamashita 1987; Summers ; Kodaki and Yamashita 1989; McGraw and Henry 1989). The CDP-DAG pathway is the major route for synthesis of PE and PC when wild-type cells are grown in the absence of ethanolamine and choline, and , , and mutants defective in this pathway have choline/ethanolamine auxotrophy phenotypes (Atkinson Summers ; McGraw and Henry 1989; Trotter and Voelker 1995; Trotter ). PE and PC synthesis in mutants defective in the CDP-DAG pathway can also be supported by exogenously supplied lysoPE, lysoPC, or PC with short acyl chains, which are transported into the cell. LysoPE and lysoPC are acylated by the -encoded lysophospholipid acyltransferase (Jain ; Tamaki ; Riekhof ,b). Short chain PE and PC are remodeled with C16 and C18 acyl chains prior to incorporation into the membrane (Tanaka ; Deng ).

In the Kennedy pathway (Hjelmstad and Bell 1990), PE and PC are synthesized, respectively, from ethanolamine and choline (Figure 2, Table1). Exogenous ethanolamine and choline are both transported into the cell by the -encoded choline/ethanolamine transporter (Nikawa ). The -encoded ethanolamine kinase (Kim ) and the -encoded choline kinase (Hosaka ) are both cytosolic enzymes, which, respectively, phosphorylate ethanolamine and choline with ATP to form ethanolamine-P and choline-P. Ethanolamine-P may also be derived from sphingolipids by dihydrosphingosine-1-P lyase, encoded by the gene (Saba ; Schuiki ). These intermediates are then activated with CTP to form CDP-ethanolamine and CDP-choline, respectively, by the -encoded ethanolamine-P cytidylyltransferase (Min-Seok ) and the -encoded choline-P cytidylyltransferase (Tsukagoshi ), which are associated with the nuclear/ER membrane (Huh ; Natter ). Finally, the -encoded ethanolamine phosphotransferase (Hjelmstad and Bell 1988, 1991) and the -encoded choline phosphotransferase (Hjelmstad and Bell 1987, 1990), respectively, catalyze the reactions of CDP-ethanolamine and CDP-choline with DAG provided by the -encoded PA phosphatase to form PE and PC. Ept1 will also catalyze the CDP-choline–dependent reaction (Hjelmstad and Bell 1988). Cpt1 and Ept1 have somewhat ambiguous patterns of localization to vesicular structures, but have also been described to localize to mitochondria or ER (Huh ; Natter ).

The CDP-DAG and Kennedy pathways are both used by wild-type cells, regardless of whether or not ethanolamine and choline are present in the growth medium (Patton-Vogt ; Henry and Patton-Vogt 1998; Kim ). In the absence of exogenous choline, this Kennedy pathway precursor may be derived from hydrolysis of PC synthesized by the CDP-DAG pathway and subsequently hydrolyzed by phospholipase D (Patton-Vogt ; Xie ) encoded by the gene (Rose ; Waksman ). Choline for PC synthesis can also be derived from Nte1-catalyzed deacylation (Zaccheo ) followed by hydrolysis by Gde1 to produce choline and glycerophosphate (Fernandez-Murray and McMaster 2005a; Fisher ; Patton-Vogt 2007). The activity of a Ca++-dependent phospholipase D with preference for PS and PE has been detected (Mayr ), but the corresponding gene encoding a PE-specific phospholipase D has yet to be identified. Thus, proof that ethanolamine derived from PE is recycled for PE synthesis via the Kennedy pathway is still lacking. Kennedy pathway mutants (e.g., and ) defective in both the CDP-choline and CDP-ethanolamine branches can synthesize PC only by the CDP-DAG pathway (McMaster and Bell 1994; Morash ; Kim ). However, unlike the CDP-DAG pathway mutants, the Kennedy pathway mutants do not exhibit any auxotrophic requirements and have an essentially normal complement of phospholipids (Morash ; Kim ).

In the synthesis of the inositol-containing phospholipids (Figure 2, Table 1), CDP-DAG donates its phosphatidyl moiety to inositol to form PI (Paulus and Kennedy 1960; Fischl and Carman 1983) in a reaction that competes in the ER with PS synthase for their common substrate, CDP-DAG (Kelley ). While -encoded PI synthase is essential, a strain expressing a mutant form of Pis1 with a lower affinity for inositol has been isolated as an inositol auxotroph (Nikawa and Yamashita 1984; Nikawa ). The inositol used in PI synthesis is either synthesized de novo (discussed below) or obtained from the growth medium via the - and -encoded inositol transporters (Table 2) (Nikawa ). Once formed, PI may be converted to PI 3-P by the -encoded PI 3 kinase (Herman and Emr 1990; Schu ) or to PI 4-P by the PI 4 kinases encoded by (Han G-S et al. 2002; Shelton ), (Yoshida ), and (Flanagan ; Garcia-Bustos ). PI 4-P may be further phosphorylated to PI 4,5-P2 by the -encoded PI 4-P 5 kinase (Yoshida ), whereas PI 3-P may be phosphorylated to PI 3,5-P2 by the -encoded PI 3-P 5 kinase (Yamamoto ). The specific localization of these kinases and the lipids they produce play important roles in signaling, membrane identity, and membrane trafficking (Strahl and Thorner 2007). PI also serves as a precursor in the synthesis of the complex sphingolipids in yeast, which also play essential roles in signaling and membrane function (Dickson and Lester 2002; Dickson 2008), topics that are beyond the scope of this YeastBook chapter.

Table 2 

Glycerolipid precursor enzymes and transporters

Gene	Ino− or Opi− phenotype	Enzyme	Mol mass (kDa)	Isoelectric point	Molecules per cellb	Locationc	Transmembrane domains	Phosphorylation sitesd
ACC1 (ABP2,FAS3,MTR7)a	—	Acetyl CoA carboxylase	250.4	6.22	20,200	Cytoplasm	None	Several
HFA1	—	Acetyl CoA carboxylase	241.8	8.05	396	Mitochondria	None	Few
FAS1a	—	Fatty acid synthase (β subunit)	228.7	5.79	91,800	Cytoplasm	None	Several
FAS2a	—	Fatty acid synthase (α subunit)	206.9	5.21	17,000	Cytoplasm	None	Several
ETR1(MRF’)	Ino−	2-Enoyl thioester reductase	42	9.78	1560	Mitochondria	None	None
HTD2 (RMD12)	—	3-Hydroxyacyl-thioester dehydratase	33	9	799	Mitochondria	None	None
MCT1	—	Malonyl-CoA:ACP transferase	40.7	6.9	1360	Mitochondria	None	None
OAR1	—	3-Oxoacyl-ACP reductase	31.2	9.3	1760	Mitochondria	None	None
ACP1	—	Acyl carrier protein	13.9	4.64	60,500	Mitochondria	None	Few
PPT2	—	Phosphopantetheine:protein transferase	19.9	8.46	486	Mitochondria	None	None
CEM1	—	β-keto-acyl synthase	47.6	8.26	1660	Mitochondria	None	None
FAA1	—	Fatty acyl CoA synthetase	77.8	7.58	7470	ER, lipid droplets	None	None
FAA2 (FAM1)	—	Fatty acyl CoA synthetase	83.4	7.7	358	Peroxisomes	None	None
FAA3	—	Fatty acyl CoA synthetase	77.9	9.7	6440	Plasma membrane	None	None
FAA4	—	Fatty acyl CoA synthetase	77.2	6.52	31,200	ER, lipid droplets	None	None
FAT1	—	Fatty acid transporter and fatty acyl CoA synthetase	77.1	8.47	16,900	ER, lipid droplets	3	None
OLE1 (MDM2)	—	Δ9 Fatty acid desaturase	58.4	9.71	ND	ER	4	None
ELO1	—	FA elongase, condensing enzyme	36.2	10.2	937	ER	5	None
FEN1 (ELO2)	—	FA elongase, condensing enzyme	40.0	10.35	3510	ER	7	Several
SUR4 (ELO3)	—	FA elongase, condensing enzyme	39.5	10.13	ND	ER	6	None
IFA38 (YBR159w)	—	β-keto acyl-CoA reductase	38.7	10.28	41900	ER	1	None
PHS1	—	3-Hydroxy acyl-CoA dehydratase	24.5	10.84	ND	ER	6	None
TSC13	—	Enoyl-CoA reductase	36.8	10.38	23600	ER	4	None
INO1a (APR1)	Ino−	Inositol 3-P synthase	59.6	5.77	ND	Cytoplasm	None	None
INM1	—	Inositol 3-P phosphatase	32.8	5	2440	Cytoplasm, nucleus	None	None
URA7	Ino−	CTP synthetase	64.7	5.93	57,600	Cytoplasm	None	Several
URA8	—	CTP synthetase	63	6.02	5370	Cytoplasm	None	None
ITR1a	Opi−	Inositol transporter	63.5	6.51	ND	Plasma membrane	12	Several
ITR2 (HRB612)	—	Inositol transporter	66.7	8.25	468	Plasma membrane	12	None
HNM1 (CTR1)a	—	Choline transporter	62	6.83	ND	Plasma membrane	12	Few
GIT1	—	GroPIns/GroPCho transporter	57.3	8.64	ND	Plasma membrane	11	None

Much of the information in the table may be found in the Saccharomyces Genome Database. Ino−, inositol auxotrophy; Opi−, inositol excretion; ND, not determined.

Genes containing the UASINO element and are regulated by the Ino2-Ino4-Opi1 circuit, The names in parentheses are aliases.

Ghaemmaghami .

Habeler ; Kumar ; Huh ; Natter .

Li ; Bodenmiller .

In the synthesis of mitochondrial-specific phospholipids (Figure 2, Table1), CDP-DAG donates its phosphatidyl moiety to glycerol-3-P to form phosphatidylglycerophosphate (PGP) in the reaction catalyzed by the -encoded PGP synthase (Janitor and Subik 1993; Chang ). PGP is then dephosphorylated to PG by the -encoded PGP phosphatase (Osman ). The -encoded CL synthase (Jiang ; Chang ; Tuller ) catalyzes the reaction between PG and another molecule of CDP-DAG to generate CL. The -encoded cardiolipin-specific phospholipase and the -encoded monolyso cardiolipin acyltransferase, involved in establishing specific unsaturated cardiolipin species, are also specifically associated with the mitochondria (Beranek ; Brandner ).

The enzymes that catalyze the turnover of phospholipids include both phospholipases and lipid phosphatases (Table 3). -encoded phospholipase B (Zaccheo ; Fernandez-Murray and McMaster 2005b, 2007) is an integral ER membrane protein and removes both fatty acids from PC to produce glycerophosphocholine (GroPCho). GroPCho may be re-acylated by an uncharacterized acyltransferase to PC (Ståhlberg ). The phospholipase B enzymes encoded by , , and catalyze the same type of reaction, but they are not specific and their localization is ambiguous. Plb1 localizes to the ER, vesicles, the plasma membrane, and the extracellular space and primarily utilizes PC and PE as substrates, whereas Plb3 is found in vesicles, vacuoles, as well as the cytosol and primarily uses PI as a substrate (Lee ; Fyrst ; Merkel , 2005). The GroPCho and glycerophosphoinositol (GroPIns) produced by the phospholipase B enzymes may be excreted into the growth medium and then transported back into the cell by the -encoded GroPCho/GroPIns transporter localized in the plasma membrane (Patton-Vogt and Henry 1998; Fisher ). In turn, these molecules are hydrolyzed by a phosphodiesterase to produce the phospholipid precursor molecules choline and inositol, respectively (Patton-Vogt 2007). -encoded phospholipase C is cytosolic and specific for PI 4,5-P2 and produces DAG and inositol 1,4,5-trisphosphate (Flick and Thorner 1993; Yoko-O ), and the -encoded phospholipase C localizes to lipid droplets and mitochondria and is specific for PG and produces DAG and glycerol 3-P (Simockova ). The -encoded cytosolic phospholipase D is specific for PC and produces PA and choline (Rose ; Waksman ). Most phospholipids undergo rapid turnover and acyl-chain remodeling, which yields the typical complex pattern of lipid molecular species found in yeast (Schneiter ; Guan and Wenk 2006; Ejsing ). This remodeling is governed by specific acyltransferases, such as the -encoded acyltransferase in the ER that is involved in stearoyl-acylation of PI (Le Guedard ), or the -encoded cardiolipin-specific phospholipase A in mitochondria (Beranek ) and the -encoded acyltransferase (yeast tafazzin ortholog; Gu ; Testet ). Moreover, several enzymes of TAG synthesis and degradation have additional activities, suggesting that they also may play a role in phospholipid acyl-chain remodeling (Rajakumari ; Kohlwein 2010b; Rajakumari and Daum 2010a,b).

There are several phosphatase enzymes that catalyze the dephosphorylation of the polyphosphoinositides. Some of these enzymes are specific and some have broad substrate specificity. The -encoded (Taylor ) and -encoded (Gary ) phosphatases are specific for PI 3-P and PI 3,5-P2, respectively, whereas -encoded (Stolz et al. 1998b) and -encoded (Wiradjaja ) phosphatases are specific for PI 4,5-P2. The phosphatase encoded by will utilize PI 3-P, PI 4-P, and PI 3,5-P2 (Guo ), whereas the phosphatases encoded by and will utilize any polyphosphoinositide as a substrate (Stolz et al. 1998; Guo ).

The - and -encoded lipid phosphate phosphatase enzymes dephosphorylate a broad spectrum of substrates that include DAG pyrophosphate (DGPP), PA, lysoPA, sphingoid-base phosphates, and isoprenoid phosphates (Toke , 1999; Faulkner ). While these enzymes may utilize PA as a substrate, they are not involved in the de novo synthesis of phospholipids and TAG; the function of which is ascribed to the -encoded PA phosphatase (Han ; Carman and Han 2009a). The gene encodes a lipid phosphatase that is specific for lysoPA and yields monoacylglycerol (MAG) and Pi (Reddy ). N-acyl PE is a minor phospholipid species implicated in signaling processes (Merkel ) and is degraded by the -encoded phospholipase D to N-acylethanolamide (NAE), which is related to endocannabinoids (Muccioli et al. 2009). NAE may be catabolized by Yju3, the major MAG lipase in yeast (see below).

Synthesis and turnover of TAG

The pathways for the synthesis of TAG and phospholipids share the same initial steps (Figure 2; Table 1; Kohlwein 2010b). Indeed, TAG is derived from the phospholipid, PA. The -encoded PA phosphatase provides DAG, which is acylated by the - and -encoded acyl-CoA–dependent and phospholipid-dependent diacylglycerol acyltransferases, respectively, to TAG (Oelkers ; Sorger and Daum 2002; Kohlwein 2010b). The - and -encoded sterol acyltransferases utilize acyl-CoA but contribute only marginally to TAG synthesis from DAG (Yang, H. ). Notably, deletion of both and genes does not result in a readily detectable growth phenotype in wild-type cells, indicating that TAG synthesis is not essential (Oelkers ; Sorger and Daum 2002). Similarly, simultaneous deletion of the and genes does not impair cell growth (Yang ). Even the quadruple mutant that lacks both TAG and steryl esters, and has no lipid droplets (LD), exhibits only a slight extension of the lag phase after recovery from quiescence (Petschnigg ). The quadruple mutant, however, does exhibit a defect in sterol synthesis (Sorger ) and is highly sensitive to unsaturated fatty acid (FA) supplementation (Garbarino ; Petschnigg ). Strikingly, provision of exogenous unsaturated FA in the absence of TAG synthesis leads to respiration-dependent cell necrosis (Rockenfeller ), challenging the dogma that lipid-induced cell death is exclusively apoptotic. The quadruple mutant also displays an Ino– phenotype at elevated growth temperature, indicative of altered expression and/or defective PI synthesis (Gaspar ), ER stress induced by tunicamycin, which inhibits protein glycosylation and stimulates LD formation in wild-type cells. However, the quadruple mutant is no more sensitive to tunicamycin than wild type, indicating that LD formation is not protective against this form of stress (Fei ). Dga1 and Lro1 acyltransferase-dependent TAG synthesis, on the other hand, is required for growth at semipermissive temperatures in the absence of inositol in the mutant, defective in COPII vesicle budding from the ER. When shifted to higher growth temperature, the mutant channels PA and DAG precursors from phospholipid into TAG, which apparently provides a degree of protection from the secretory stress caused by a block in membrane trafficking (Gaspar ).

TAG hydrolysis (Figure 2, Table 3) clearly contributes lipid precursors that are essential to the resumption of growth upon exit from stationary phase (Kurat ; Kohlwein 2010b). TAG degradation provides substrates for the synthesis of phospholipids and sphingolipids (Rajakumari ), which are required for efficient cell cycle progression upon exit from quiescence (Kurat ). Degradation of TAG is catalyzed by -, -, -, and -encoded TAG lipases, all of which are localized to LD (Athenstaedt and Daum 2003, 2005; Jandrositz ; Kurat ; Rajakumari ; Kurat ; Kohlwein, 2010b). Tgl1 harbors the canonical lipase catalytic triad, consisting of serine201, aspartate369 and histidine396 (Jandrositz et al. 2005). In contrast, Tgl3, Tgl4, and Tgl5 (and also Nte1, see above) are members of the patatin domain-containing family of (phospho) lipases, characterized by a catalytic diad of serine and aspartate (Kurat ). The substrate specificities of the Tgl3, Tgl4, and Tgl5 TAG lipases differ. Tgl5 preferentially hydrolyses TAG molecular species harboring very long chain fatty acids (VLCFA) (Athenstaedt and Daum 2005), while Tgl3 also accepts DAG as a substrate in addition to TAG (Kurat ). Tgl3, Tgl4, and Tgl5 also possess lysoPA acyltransferase activities, and Tgl4 additionally functions as a Ca++-dependent phospholipase A2 and steryl ester hydrolase (Rajakumari and Daum 2010a,b). Although Tgl1 also contributes to TAG hydrolysis, its major activity is as a steryl ester hydrolase, in conjunction with the - and -encoded enzymes (Jandrositz ; Koffel ). encodes an acylglycerol lipase localized to mitochondria, but its role in TAG homeostasis has not been clarified yet (Ham ). encodes the major monoacylglycerol (MAG) lipase in yeast (Heier ), but Δ mutants do not display any detectable growth phenotype when tested under multiple conditions (Heier ).

Stationary phase cells shifted into fresh media rapidly initiate TAG and steryl ester breakdown, which leads to almost full depletion of cellular LD pools within 4–6 hr in lag phase. This initial phase of TAG breakdown is governed by the activity of - and -encoded lipases on LD and the mutant strain exhibits a delay in entering vegetative growth during exit from stationary phase (Kurat ). Tgl4 is constitutively present on LD and activation of Tgl4 by Cdc28p-dependent phosphorylation is involved in TAG lipolysis that contributes to bud formation exiting from stasis (Kurat ). Resumption of growth following stasis is also dependent on the -encoded DAG kinase. In comparison to wild-type cells, stationary phase cells fail to initiate growth when de novo FA synthesis is impaired. The mutant also fails to mobilize TAG under these conditions and accumulates TAG, phenotypes that are partially suppressed by the mutation or by channeling DAG into PC synthesis when choline is present (Fakas ).

TAG synthesis and breakdown are also coordinated with phospholipid metabolism during logarithmic growth. Mutants with defects in TAG hydrolysis exhibit decreased synthesis of inositol containing sphingolipids and decreased PC and PI content during active growth (Rajakumari ). Furthermore, mutants defective in synthesis or hydrolysis of TAG exhibit reduced capacity to restore cellular levels of PI when exogenous inositol is resupplied following an interval of inositol starvation during logarithmic growth. Under these conditions, alterations in the synthesis of inositol-containing sphingolipids are also observed in the strain (Gaspar ).

Glycerolipid precursors

Fatty acid synthesis and regulation:

The FA that are esterified to phospholipids and TAG are derived from de novo synthesis, the growth medium, and from lipid turnover (Tehlivets ). The spectrum of FA in yeast consists mainly of C16 and C18 FA that are either saturated or monounsaturated, harboring a single double bond between carbon atoms 9 and 10 (Δ9 desaturation). Whereas de novo FA synthesis mostly takes place in the cytosol, all the enzymes involved in FA desaturation and elongation are associated with the ER membrane (Table 2) (Tehlivets ).

Minor FA species are C12, C14, and very long chain FA, up to C26. However, FA compositions vary substantially between strains and are also dependent on cultivation conditions (Martin ). Two different FA synthesis pathways exist in yeast, as in mammalian cells (Tehlivets ): the major cytosolic pathway, which resembles the “eukaryotic” type I pathway and is responsible for the bulk synthesis of all major FA, and the mitochondrial (type II pathway), which is organized similarly to the bacterial FA biosynthetic pathway (Hiltunen ). The latter is believed to synthesize FA up to C8 carbons, which serve as precursors for the synthesis of lipoic acid. Cytosolic FA synthesis is initiated by the -encoded acetyl-CoA carboxylase, which synthesizes malonyl-CoA from acetyl-CoA (Al-Feel ; Hasslacher ). This reaction requires a covalently bound biotin prosthetic group, which is attached to lysine735 in the biotin-carrier domain of Acc1 by the -encoded biotin:apoprotein ligase. The FA synthase complex is composed of two subunits, encoded by (Fas1, β-subunit) and (Fas2, α-subunit) and is organized as an α6/β6 heterooligomeric complex (Kuziora ; Schweizer ; Chirala ). Fas2 carries a pantetheine prostetic group on the acyl-carrier protein (ACP) domain, which is a central element in a cycling series of reactions. In a first condensation step, malonyl-CoA is condensed with pantetheine-bound acetyl-CoA to form 3-keto-acyl-ACP, which is reduced to 3-hydroxyacyl-ACP, dehydrated to 2,3-trans-enoyl-ACP, and further reduced to acyl(C+2)-ACP. Both reduction steps require NADPH and, as a result, FA synthesis is a major consumer of this dinucleotide. This multistep process results in the addition of two carbon units to the growing FA chain and cycles up to seven times, resulting in acyl-chains typically of 16 carbon atoms. The newly formed FA is transferred to coenzyme A, to yield cytosolic long chain acyl-CoA (Tehlivets ). Acyl-CoA molecules are precursors for all acylation reactions involved in the synthesis of phospholipids, TAG, long chain bases, ceramide, and steryl esters, and also serve as precursors in protein acylation. The acyl-CoAs, whether derived from FA de novo synthesis or lipid recycling, are subject to elongation and desaturation, yielding the typical pattern of saturated and unsaturated FA species (see below). Yeast also contains an -encoded acyl-CoA binding protein, which plays an important regulatory function in delivering acyl-CoA into various pathways (Schjerling ; Gaigg ; Faergeman ; Rijken ). Mitochondrial FA synthesis involves enzymatic steps similar to those catalyzed by the cytosolic FAS complex. However, in contrast to cytosolic FA synthesis, but resembling archae- and eubacterial type II fatty acid synthase, the reactions of FA synthesis in mitochondria are catalyzed by individual polypeptides, encoded by separate genes (Hiltunen ).

Defects in the or genes encoding cytosolic FA synthase lead to FA auxotrophy and can typically be rescued by the addition of exogenous C14 or C16 FAs. However, defects in (Hasslacher ) and BPL2 are lethal and cannot be suppressed by the addition of long chain FA. The essential role of these genes, when exogenous C14 and C16 FAs are present, is most likely the requirement for malonyl-CoA in the synthesis of essential VLCFA, which are components of glycerophosphoinositol (GPI) anchors and sphingolipids (Pittet and Conzelmann 2007; Dickson 2008; Dickson 2010). The synthesis of VLCFAs is accomplished by an ER membrane-localized elongase complex that acts on long chain acyl-CoA, consisting of the condensing enzymes encoded by , , and (Oh ), the β-ketoacyl-CoA reductase encoded by gene (Han et al. 2002), the dehydratase encoded by (Denic and Weissman 2007; Kihara ), and the enoyl-CoA reductase, encoded by (Kohlwein ). Tsc13 accumulates within a specialized region of the ER in juxtaposition to the vacuole, termed the nucleus-vacuole junction (Pan ) through interaction with Osh1p and Nvj1p (Kvam ). The physiological role for a localization of just one component of the FA elongation complex to this subregion of the ER is currently unclear, but indicates a role for VLCFA in the formation of microautophagic vesicles involved in piecemeal autophagy of the nucleus (Kvam ). Triple mutations in all three condensing enzyme genes as well as mutations in or are lethal, further supporting the notion that VLCFAs are essential in yeast. Yeast also harbors a single FA Δ9-desaturase encoded by , which localizes to the ER (Stukey ) and mutants defective in Ole1 require exogenous C16:1 or C18:1 FA for growth. Since FA desaturation is an oxygen-dependent process (see below), growth of yeast under anaerobic conditions also requires the supplementation of these unsaturated FA.

In the absence of de novo FA synthesis or desaturation, as in , , or mutants, and in wild-type cells under anaerobic conditions, cellular growth is dependent on an exogenous supply of FA. Under these conditions, the activity of acyl-CoA synthetases (Table 2), encoded by , , , , and (Duronio ; Harington ; Johnson ,b; Watkins ; Choi and Martin 1999; Black and Dirusso 2007) is also required. The acyl-CoA synthetases activate free FA with coenzyme A in an ATP-dependent reaction, and are also believed to be required for FA uptake into yeast cells (Faergeman ; Black and Dirusso 2007; Obermeyer ). Acyl-CoA synthetases differ in their substrate specificity and subcellular localization and are present in the ER membrane, plasma membrane, peroxisome, and LD (Natter ; Black and Dirusso 2007). The acyl-CoA synthetase Faa2 and the enzymes for FA β-oxidation, a cyclic series of reactions breaking down FA into acetyl-CoA and generating FADH2 and NADH, are localized to the peroxisomes (planned YeastBook chapter “Lipid particles and peroxisomes” by Kohlwein and van der Klei). FA uptake into yeast cells is also mediated by an endocytotic mechanism that requires the activity of Ypk1, the yeast ortholog of the human serum- and glucocorticoid-induced kinase (Jacquier and Schneiter 2010). In the absence of acyl-CoA synthetases, FA released by lipid turnover (see above) cannot be activated and utilized for lipid synthesis. For this reason, mutants lacking these activities excrete FA (Scharnewski ).

FA synthesis is regulated at multiple levels (Tehlivets ). , , and expression is under control of an UASINO element and coregulated with genes involved in phospholipid synthesis (Chirala 1992; Schuller ) (see below). expression is also regulated by the SAGA protein complex and TFIID (Huisinga and Pugh 2004) and may therefore depend on histone acetylation. expression was also found to fluctuate during the cell division cycle, with a peak of expression in early G1 phase (Cho ). Compensatory changes in expression in mutants defective in Acc1 activity indicate an autoregulatory loop (Shirra ). Expression of the and genes is regulated by the transcription factors Gcr1, Reb1, Rap1, and Abf1 (Schuller ; Greenberg and Lopes 1996) and repressed by long chain FA (Chirala 1992). is also regulated by the SAGA complex and TFIID, and both genes show cell cycle-dependent expression, with a peak level at M/G1 (Spellman ; Huisinga and Pugh 2004).

Acc1 enzyme activity is greatly elevated in the Δ mutant and the data are consistent with Acc1 being a substrate of AMP-activated protein kinase, encoded by the (Woods ; Shirra ). Snf1 also regulates chromatin by phosphorylating histones among other proteins and, therefore, has multiple effects on expression of genes, including , , and . Notably, and promoter sequences differ, and the stoichiometry of the FAS complex is established both by the level of gene expression and Fas1 protein abundance (Schuller ; Wenz ; Tehlivets ). Excess of either protein is rapidly eliminated by degradation via vacuolar proteases (Fas1) or ubiquitination and proteasomal digestion (Fas2), respectively (Peng ).

Knowledge about the regulation of FA elongation leading to VLCFA synthesis is limited. Microarray experiments have shown that expression is upregulated in the presence of myristic acid (C14:0), consistent with the preference of the Elo1 protein for this FA (Toke and Martin 1996). Expression of , , and genes also responds to stationary phase, nitrogen limitation, glycosylation defects, or α-factor treatment, indicating a regulatory cross-talk between nutritional state and cell proliferation to VLCFA synthesis (Gasch ).

Ole1 is a nonheme iron-containing integral ER membrane protein (Table 2), which harbors an intrinsic cytochrome b5 domain as an electron carrier (Martin ). To subtract two electrons and two protons from the saturated acyl-chain and transfer them to oxygen, Ole1 activity is complemented by the activity of an NADH cytochrome b5 reductase (dehydrogenase) that is also localized to the ER. Desaturation of FA by Ole1 is a highly regulated, oxygen-requiring process (Martin ), which is discussed in detail in the section below on cell biology of lipids.

Water soluble precursors of phospholipids, metabolism, and regulatory roles:

A number of water-soluble molecules used in phospholipid synthesis, including inositol, choline, ethanolamine, serine, CTP, and S-adenosyl methionine (AdoMet) and the enzymes involved in these reactions are largely cytosolic (Table 2). However, choline and ethanolamine are produced in yeast in the context of ongoing synthesis and turnover of PE and PC in the membranes as discussed above. Considerable attention has been paid to inositol and CTP, which have major regulatory roles in phospholipid metabolism. The effects of inositol on transcriptional regulation UASINO-containing phospholipid biosynthetic genes controlled by the Opi1 repressor are shown in Figure 3 and will be discussed in a subsequent section. The regulatory effects of the soluble metabolites inositol, CTP, and S-adenosyl homocysteine (AdoHcy) on enzyme activity and metabolic flux in the CDP-DAG pathway (Figure 4) are discussed here.

Inositol is the precursor to PI (Figure 2), which is essential to the synthesis of polyphosphoinositides (Strahl and Thorner 2007), sphingolipids (Cowart and Obeid 2007; Dickson 2008, 2010), and GPI anchors (Pittet and Conzelmann 2007). Inositol is derived from glucose-6-P via the reactions catalyzed by the cytosolic -encoded inositol-3-P synthase (Donahue and Henry 1981; Klig and Henry 1984; Dean-Johnson and Henry 1989) and the -encoded inositol-3-P phosphatase (Murray and Greenberg 2000). Exogenously supplied inositol stimulates PI synthase and inhibits PS synthase activity, alleviating the competition with PI synthase for their common substrate, CDP-DAG (Figure 4). The presence of exogenous inositol leads to increased PI synthesis and reduced levels of both CDP-DAG and its precursor PA in wild-type cells (Kelley ; Loewen ). TAG, which is derived from PA by the action of Pah1 (Figure 2), is decreased in the presence of inositol and increases in its absence (Gaspar , 2011). In addition, the levels of the complex sphingolipids, which are derived from PI, are reduced when cells are deprived of inositol and increase when inositol is supplied (Becker and Lester 1977; Hanson and Lester 1980; Jesch ). The changes that occur in phospholipid synthesis and composition in response to exogenous serine (Kelley ) and choline (Gaspar ) are much less dramatic than the effects of exogenous inositol. However, when inositol is present in the growth medium, serine (Homann ), ethanolamine, and choline (Poole et al. 1986) result in the reduction of the activities of PS synthase and CDP-DAG synthase at the transcriptional level (see below). Exogenous choline also results in a dramatic change in the rate and mechanism of PC turnover, leading to a switch from a phospholipase D to a phospholipase B-mediated route (Dowd et al. 2001). The phospholipase responsible for turnover of PC under these conditions is Nte1 (Zaccheo ).

CTP is derived from UTP by the cytosolic - and -encoded CTP synthetase enzymes (Ozier-Kalogeropoulos , 1994).

The nucleotide CTP plays a critical role in phospholipid synthesis as the direct precursor of the high-energy intermediates CDP-DAG, CDP-choline, and CDP-ethanolamine that are used in the CDP-DAG and Kennedy pathways (Figure 2) (Chang and Carman 2008). It is also used as the phosphate donor for the synthesis of PA by DAG kinase (Han ). CTP synthetase (Ozier-Kalogeropoulos , 1994) that produces CTP is allosterically inhibited by the product (Yang ; Nadkarni ), and this regulation ultimately determines the intracellular concentration of CTP (Yang ; McDonough ). CTP inhibits the CTP synthetase activity by increasing the positive cooperativity of the enzyme for UTP, and at the same time by decreasing the affinity for UTP (Yang ; Nadkarni ). However, CTP synthetases containing an E161K mutation are less sensitive to CTP product inhibition (Ostrander ), and cells expressing the mutant enzymes exhibit a 6- to 15-fold increase in the CTP level (Ostrander ). They also show alterations in the synthesis of membrane phospholipids, which include a decrease in the amounts of PS and increases in the amounts of PC, PE, and PA (Ostrander ). The decrease in the amount of PS results from direct inhibition of PS synthase activity by CTP (McDonough ), and this inhibition favors the synthesis of phospholipids by the Kennedy pathway (Figure 4). The increase in PC synthesis is ascribed to a higher utilization of the CDP-choline branch of the Kennedy pathway due to the stimulation of choline P cytidylyltransferase activity (McDonough ; Ostrander ) by increased substrate availability of CTP (McDonough ; Kent and Carman 1999). Likewise, the increase in PE synthesis could be attributed to stimulation of ethanolamine-P cytidylyltransferase activity. The increase in PA content may result from the stimulation of DAG kinase activity by increased availability of its substrate CTP (Han ). CTP also inhibits the activity of Pah1 (Wu and Carman 1994), another factor that contributes to an elevation of PA content. The cells expressing the E161K mutant CTP synthase excrete inositol (Ostrander ), a characteristic phenotype that typifies the misregulation of UASINO-containing phospholipid synthesis genes (see below) in cells that accumulate an excess of PA (Carman and Henry 2007). It is unclear whether UASINO-containing genes in the CDP-DAG and Kennedy pathways are derepressed in CTP overproducing cells, but the overriding regulation that governs the utilization of the two pathways appears to be biochemical in nature.

AdoHcy is a product of the AdoMet-dependent methylation reactions that are catalyzed by Cho2 and Opi3 in the CDP-DAG pathway (Figure 2). -encoded AdoHcy hydrolase (Malanovic ) requires NADH for the hydrolytic breakdown of AdoHcy to adenosine and homocysteine (Takata ). AdoHcy (Figure 4) is a competitive inhibitor of the phospholipid methyltransferase enzymes (Gaynor and Carman 1990). Thus, down-regulation of the AdoHcy hydrolase causes the accumulation of AdoHcy and the inhibition of PC synthesis, which leads to an increase in PA content and consequently, as described below, to the derepression of UASINO-containing genes (Malanovic ). Although the effects of AdoHcy on phospholipid composition have not been addressed, its accumulation causes an increase in TAG synthesis and LD content (Malanovic ), further underscoring the metabolic interconnection between phospholipid and TAG homeostasis discussed above.

Cell Biology of Lipids

In yeast, most subcellular membranes are composed of similar glycerophospholipid classes but the quantitative phospholipid composition of subcellular organelles differs substantially (Schneiter ). Moreover, as described above, multiple cellular organelles and compartments contribute to glycerolipid metabolism (Figure 2, Tables 1–3). While multiple compartments contribute to lipid synthesis in yeast (Zinser ), most reactions are confined to single membrane compartments. Thus, extensive regulated flux of lipids, across and among individual membranes and organelles, is required to enable the cell to adjust lipid composition in specific compartments under changing growth conditions. A number of mechanisms involved in these complex interactions have been identified.

For example, a set of membrane-bound flippases are known to govern transmembrane movement of phospholipids (Pomorski ; Holthuis and Levine 2005). Intermembrane transfer of lipids is facilitated by a family of proteins, termed oxysterol-binding protein (OSBP) related proteins (ORP), of which seven members, Osh1–7, with overlapping functions, exist in yeast (Beh and Rine 2004). Some of the yeast Osh proteins have been localized to membrane contacts sites (Levine and Loewen 2006; Raychaudhuri and Prinz 2010). While none of these genes individually are essential, simultaneous deletion of all seven OSH genes is lethal and conditional alleles, following a shift to restrictive conditions, result in pleiotropic effects on vesicular trafficking and phospholipid and sterol composition of membranes, as do combinations of individual deletions (Beh and Rine 2004; Fei ). The six members of the Sec14 superfamily constitute another major class of proteins involved in sensing and regulating membrane lipid composition (Griac 2007; Bankaitis ). The five Sec14 homologs (Sfh1–5) share 28–76% similarity to Sec14 (Griac 2007) and localize to multiple subcellular organelles (Schnabl ). Sec14, which performs an essential function at the Golgi, was originally identified as a PI/PC transfer protein, but its role in establishing phospholipid homeostasis is complex (Bankaitis ).

In some instances, movement of lipids between membrane compartments is required to carry out a sequence of reactions. For example, PS is synthesized in the ER, but the major PS decarboxylase, Psd1, is located in the mitochondria (Figure 2, Table 1). Subsequently, the PE produced by Psd1 must be transported back to the ER to undergo methylation leading to PC. PS is transported into the mitochondria by an ubiquitin-regulated process that is insensitive to disruption of vesicular trafficking and involves specialized regions of apposition of mitochondria and ER known as the mitochondria-associated ER or MAM (Clancey ; Trotter ; Trotter and Voelker 1995; Achleitner et al. 1999; Voelker 2003; Osman ). An ER–mitochondria tethering complex potentially involved in movement of phospholipids between the ER and mitochondria has been described (Kornmann ).

The complex relationship between phospholipid metabolism in the ER and the synthesis and breakdown of cytosolic LD represents another example of the intricate interaction among cellular compartments that occurs in the course of lipid metabolism. LD are not merely storage depots for TAG and steryl esters. They are metabolically active; harboring multiple enzymes involved in lipid metabolism (Athenstaedt ; Rajakumari ; Goodmann 2009; Kohlwein 2010b) (See also chapter: Lipid Droplets and Peroxisomes). Interaction of LD with the ER, mitochondria, and peroxisomes has been reported (Goodmann 2009). Pah1 clearly plays a critical role in TAG synthesis, LD formation, and the balance between nuclear/ER membrane proliferation and LD formation (Carman and Han 2009a; Adeyo ). Several enzymes involved in TAG synthesis show a dual localization to the ER and LD (Figure 2, Table 1), including acyl-DHAP reductase (Ayr1), glycerol-3-P acyltransferase (Gpt2), and acyl-CoA–dependent DAG acyltransferase (Dga1). Formation of LD is believed to occur between the leaflets of the ER membrane, but alternative models such as vesicular trafficking have also been suggested (Thiele and Spandl 2008; Farese and Walther 2009; Guo ). The close interaction of LD with mitochondria and peroxisomes also underscores the important role of LD in metabolism (Goodmann 2009). TAG levels fluctuate drastically during cellular growth and division, and multiple conditions contribute to the level of TAG storage in LD in stationary phase cells (Kurat ; Czabany ; Daum ; Rajakumari ; Zanghellini ). As cells exit stationary phase, TAG hydrolysis supplies precursors for membrane lipid synthesis. As active cellular growth progresses, de novo FA synthesis increases, satisfying cellular requirements for lipid synthesis but even while TAG turnover is still in progress, de novo formation of TAG and LD is initiated (Kurat ; Zanghellini ). Methods for imaging-based screening of mutant libraries have been designed to identify factors involved in LD content, morphology, and mobilization (Wolinski and Kohlwein 2008; Wolinski ) and screens using such methods have identified >200 genes that potentially influence these processes (Szymanski ; Fei ). However, there is surprisingly little overlap between the mutants identified in the published screens, indicating that current screens are far from being saturated.

Roles of lipids in organelle function and morphology

Due to the ubiquitous presence of most phospholipids in all subcellular membranes it can be difficult to assess specific phospholipid functions, in particular organelles. However, quite specific functions have been attributed to several lipids in the mitochondrion. For example, PE synthesized in the mitochondria by Psd1 plays an important role in stabilizing mitochondrial protein complexes (Birner ; Osman ). Consequently, the mutant is nonviable in conjunction with defects in components of the prohibitin complex, which functions as a protein and lipid scaffold and ensures the integrity and functionality of the mitochondrial inner membrane (Osman ; Potting ; van Gestel ; Osman ). In another example, CL, which is present only in the mitochondrion (Li, G. ), plays an important role in mitochondrial genome stabilization and higher order organization of components of the mitochondrial respiratory chain (Koshkin and Greenberg 2002; Zhang ; Zhong ; Mileykovskaya ; Joshi ). However, CL function reaches beyond the mitochondria, indirectly regulating morphology and acidification of the vacuole through the retrograde () signaling pathway and the -encoded vacuolar Na+/H+ exchanger (Chen ). Lack of CL in mutants is in part compensated by an upregulation of the PE content, which is dependent on the mitochondrial synthesis through Psd1 (Gohil ). Defects in Crd1 are also, in part, compensated by an increase in the precursor lipid PGP. Accordingly, mutation of the gene encoding PGP synthase leads to more pronounced mitochondrial defects and a petite negative phenotype (Janitor ; Chang ).

Alterations in organelle morphology and/or function observed in response to changes in lipid composition due to mutation or pharmacological inhibition of enzymes (Schneiter and Kohlwein 1997) provide tantalizing clues as to the roles of specific lipids. Notably, the mutant exhibits the expansion of the nuclear/ER membrane and up-regulation of phospholipid biosynthesis genes (Santos-Rosa ). As discussed previously, dephosphorylation of PA by Pah1 provides DAG for the synthesis of TAG, as well as PE and PC in the Kennedy pathway. Consequently, the mutant has elevated PA levels, which result in up-regulation of phospholipid biosynthesis (see below) and membrane proliferation (Han ; Carman and Henry 2007; Carman and Han 2009b). A number of mutants defective in FA metabolism also exhibit morphological abnormalities. For example, conditional mutants defective in FA desaturation () show defective mitochondrial morphology and distribution upon cell division (Stewart and Yaffe 1991; Tatzer ). The finding that several dozen proteins species are required to sustain cell viability in the presence of unsaturated FA suggests an enormously complex network of processes and interactions to maintain membrane homeostasis and function (Lockshon ). In another example, a conditional mutant defective in FA de novo synthesis has impaired morphology of the vacuole due to reduced acylation of Vac8 that is involved in stabilizing the vacuolar membrane and vacuole inheritance (Schneiter ). Mutants defective in FA elongation show similar fragmentation of the vacuole due to impaired synthesis of sphingolipids, which are important for the stability of the vacuolar ATPase complex (Chung ).

Due to their important roles in protein modification, myristic and palmitic acid affect the membrane association of numerous proteins and consequently influence signaling, membrane function and organelle morphology (Dietrich and Ungermann 2004). encodes the essential N-myristoyl transferase that attaches coenzyme A-activated myristic acid (C14-CoA) to a glycine residue close to the N-terminus of target proteins, resulting in cleavage of the peptide bond and N-terminal glycine acylation (Towler ). Although the myristoyl residue is shorter than the typical C16 or C18 chain length found in membrane phospholipids and, therefore may not fully interdigitate into a membrane leaflet, it serves an essential function in promoting membrane association of proteins (Duronio ). Palmitoylation of internal cysteine residues within the peptide chain plays an important role in post-translational modification of some 50 proteins in yeast (Roth ), modulating membrane association, folding and activity. Protein palmitoylation is catalyzed by seven members of the “Asp-His-His-Cys-cystein-rich“ (DHHC-CRD) domain family of proteins, and includes the Erf2/Shr5 complex, Akr1, Akr2, Pfa3, Pfa4, Pfa5, and Swf1. Among the substrates for palmitoylation are Vac8, SNAREs, and Ras2, indicating important regulatory functions for this FA modification of proteins in multiple pathways involved in membrane trafficking and signaling.

Mechanisms of compartmentalization and localization of enzymes of lipid metabolism

Many enzymes involved in glycerolipid metabolism have transmembrane domains (Tables 1–3), which anchor the proteins in their specific membrane environment and promote access to hydrophobic lipid substrates. However, several membrane-associated enzymes lack defined transmembrane spanning domains and membrane interaction is mediated, perhaps in a regulated manner, with specific membrane-resident anchor proteins. For example, -encoded PI 4-kinase binds to the membrane through Frq1 and the -encoded PI 3-kinase binds through Vps15 (Strahl and Thorner 2007). In addition, enzymes may bind to the membrane through specific interaction with lipids in the membrane, such as PI 3-P, PI 4-P, PI 3,5-P2 and PI 4,5-P2 through PH, PX, FYVE or ENTH domains (Strahl and Thorner 2007). A number of enzymes involved in lipid biosynthesis utilize water soluble lipid precursors, including the choline and ethanolamine kinases and the enzymes of the FAS complex, involved in long chain FA synthesis, and are localized to the cytosol. Acyl-CoA, generated by the FAS complex, has a highly amphiphilic structure and easily associates with membrane surfaces, facilitating the subsequent incorporation of the acyl chains into lipids by membrane-bound enzymes. However, Pah1 is an example of an enzyme that is associated with both cytosolic and membrane compartments. PA, the substrate for Pah1, is present in the membrane, but the largest pool of Pah1 is cytosolic, highly phosphorylated and inactive. Pah1 must be dephosphorylated by the Nem1-Spo7 complex in order to be functional in vivo (Santos-Rosa ; Han ; Carman and Han 2009a; Choi ). Dephosphorylation of Pah1 by the Nem1-Spo7 complex leads to its membrane association and its activation (Choi et al. 2011), leading to the production of DAG and TAG and the lowering of PA levels, which in turn affects regulation of phospholipid biosynthetic genes (see below) (Carman and Henry 2007; Carman and Han 2009b).

Regulation of lipid metabolism in response to membrane function

Several regulatory circuits have been uncovered that govern, by different mechanisms, the cross-talk between membrane function and lipid synthesis. Proteins involved in this regulation may sense specific changes in lipid composition, membrane charge, fluidity, or curvature. One such mechanism controls the expression of the fatty acid desaturase, Ole1, by the ER membrane-bound transcription factors Mga2 and Spt23, which are cleaved and released from the membrane in response to changes in membrane fluidity or permeability (Hoppe ). Both proteins are synthesized as larger, ER membrane-bound proteins that are ubiquitinated and cleaved by the activity of the 26S proteasome under low oxygen conditions, in conjunction with the ubiquitin-selective chaperone . This cleavage detaches the soluble fragments from their membrane anchors, and allows them to enter the nucleus to control transcription of and several other genes through the low oxygen response (LORE) promoter element. Proteasomal cleavage of Mga2 and Spt23 requires the essential E3 ubiquitin ligase, Rsp5, and lack of this activity results in a cellular requirement for unsaturated FA. Cleavage of Spt23 and Mga2 in the ER is believed to be regulated by the membrane environment, adjusting membrane lipid composition through the control of FA desaturation. Under normoxic conditions, expression is activated by the oxygen-responsive transcription factor Hap1. Addition of unsaturated FA, including linoleic acid (C18:2) or arachidonic acid (C20:4), which are normally not present in yeast, results in a drastic and rapid reduction of mRNA expression and stability. However, low oxygen tension leads to a massive induction of expression through the LORE element present in its promoter. This induction is believed to increase efficiency of FA desaturation to support cellular growth under oxygen limiting (hypoxic) conditions (Martin ; see also the chapter on Lipid Droplets and Peroxisomes for more discussion of fatty acid metabolism and regulation). Mga2 (but not Spt23) was also shown to be responsible for the rapid, but transient, up-regulation of and other LORE-containing genes observed when inositol is added to cultures of wild-type yeast growing logarithmically in the absence of inositol (Jesch ). Regulation of the phospholipid biosynthetic genes containing the UASINO promoter element also occurs in response to changes in the lipid composition in the ER membrane, through the binding of the Opi1 repressor to PA (see below). The ER is also the locus of complex regulatory cross-talk involving membrane expansion, the secretory pathway, the unfolded protein response (UPR) pathway, and phospholipid metabolism (Cox ; Travers ; Block-Alper ; Chang , 2004; Hyde ; Brickner and Walter 2004; Jesch ; Gaspar ; Schuck ).

Another example of regulation of lipid metabolism in yeast in response to changing membrane conditions was discovered in the course of analysis of a mutant strain completely defective in PC formation via the PE methylation pathway (Boumann ). When this mutant is deprived of choline, PC levels decline and as this process progresses, concomitant changes in acyl-chain distribution occur in PC and other phospholipids, especially PE. However, similar changes were not observed in the neutral lipid fraction. The shortening and increased saturation of PE acyl chains decreases the bilayer forming potential of PE and Boumann suggest that phospholipid remodeling under these conditions may provide a mechanism for maintaining intrinsic membrane curvature. The nature of such a mechanism has not yet been defined, but yeast has a number of proteins harboring a BAR domain, including Rvs161 and Rvs167, which bind to liposomes in a curvature-dependent manner and promote tubule formation in vitro. In vivo studies in yeast indicate inappropriate regulation of phosphoinositide and sphingolipid metabolism impinges on Rsv161 and Rsv167 function (Ren ; Youn ). The Rim101 pathway, which is involved in regulating cellular pH in response to alkaline conditions and cell wall biogenesis, also appears to be involved in sensing membrane curvature (Mattiazzi ).

Genetic Regulation of Glycerolipid Synthesis

In wild-type cells, both glycerolipid composition and the expression of lipid biosynthetic genes are influenced by a wide variety of growth conditions, including growth phase, temperature, pH, and the availability of nutrients such as carbon, nitrogen, phosphate, zinc, and lipid precursors. Transcript levels of certain lipid biosynthetic genes in yeast are also regulated by message stability. A number of viable mutants defective in genes encoding lipid metabolic enzymes, transporters, and regulatory factors exhibit changes in lipid composition. These complex and interrelated topics have been the subjects of numerous reports and reviews (Henry 1982; Carman and Henry 1989; Paltauf ; Henry and Patton-Vogt 1998; Carman and Henry 1999; Gardocki ; Jesch and Henry 2005b; Santos-Rosa ; Carman and Henry 2007; Chen ; Daum ; Gaspar ; Li, G. ; Patton-Vogt 2007; Gaspar ; Schuiki ; Young ; Carman and Han 2011), as discussed below with a focus on the transcriptional regulation of lipid biosynthetic genes in response to PA.

Regulation of transcript abundance by mRNA degradation

The transcript abundance of some glycerolipid metabolic genes is regulated at the level of mRNA degradation. Genes that exhibit this level of regulation include (Choi and Carman 2007) and (Gonzalez and Martin 1996; Vemula ; Martin ). In wild-type cells, 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 (Herrick ). However, mRNA is greatly stabilized with a half-life >45 min in respiratory mutants (Choi and Carman 2007). This results in increased levels of the PS synthase protein and its associated activity (Choi and Carman 2007). Given that mRNA decays by the primary 5′–3′ decay pathway when cells are respiratory sufficient (Parker and Song 2004), it is reasonable to predict that the rate of deadenylation and/or decapping may be reduced when respiration is blocked. The transcript is destabilized when cells are supplemented with unsaturated FA (Gonzalez and Martin 1996; Vemula ; Martin ). This FA-regulated decay of mRNA occurs through both the 5′–3′ general pathway and via exosomal 3′–5′ degradation activities (Martin ).

Transcriptional regulation by inositol and choline

Expression of the gene (Figure 3, Table 2) is repressed by ≥100-fold in the presence of inositol. Exogenous choline results in an additional severalfold reduction in transcript level but only if inositol is also present. However, in the absence of exogenous inositol, choline has little or no effect on expression (Hirsch and Henry 1986; Griac ; Jesch ). Other genes encoding enzymes of lipid metabolism, including many involved in PC biosynthesis, show similar patterns of regulation in response to inositol and choline (Carman and Henry 1989, 1999, 2007; Paltauf ; Henry and Patton-Vogt 1998; Jesch and Henry 2005; Jesch ; Chen ; Carman and Han 2011). However, no other gene in the yeast genome shows as great a repression ratio in the presence of inositol and choline as (Santiago and Mamoun 2003; Jesch et al. 2005, 2006).

The regulation (Figure 3) controlling the expression of and coregulated genes of lipid metabolism has been the subject of a number of recent comprehensive reviews (Carman and Henry 2007; Chen ; Carman and Han 2011). Expression of these genes is controlled by the cis-acting element, consensus 5′CATGTGAAT3′ (Bachhawat ; Greenberg and Lopes 1996), known as the inositol sensitive upstream activating element (UASINO), or alternatively as the inositol/choline responsive element (ICRE) (Schuller et al. 1992). The Ino2-Ino4 heterodimer, which is required for activation of transcription of UASINO-containing genes, binds to this site (Ambroziak and Henry 1994). Genome-wide location analysis confirmed the sequence of the UASINO element and the binding of the Ino2-Ino4 heterodimer to the element across the genome (Lee ; Harbison ). Since both Ino2 and Ino4 are required for activation of expression of , , and mutants exhibit inositol auxotrophy (Ino− phenotype). However, since many genes involved in lipid metabolism in addition to contain the UASINO promoter element (Greenberg and Lopes 1996), the and mutants exhibit altered phospholipid compositions including reduced PC content, even when growing in the presence of inositol (Loewy and Henry 1984). The fact that the regulatory gene also contains a UASINO element in its promoter, and is auto-regulated in response to inositol, adds an additional layer of complexity to this regulatory mechanism (Ashburner and Lopes 1995, Chen ). The locus has also recently been reported to localize at the nuclear periphery when it is being actively transcribed (Brickner and Walter 2004), through a chromatin-mediated mechanism that persists for several generations following addition of inositol (Brickner ).

Repression of UASINO-containing genes in response to the presence of exogenous inositol involves the Opi1 repressor (Hirsch and Henry 1986; White ) (Figure 3), which interacts with transcriptional activation domains in Ino2 (Wagner ). Opi1 has also been reported to interact with a number of other regulatory proteins (Chen ), including the pleiotropic corepressors, Sin3 and Ssn6 (Jaschke et al. 2011). Deletion of the gene renders the cell incapable of repressing UASINO genes in response to inositol, and leads to expression of UASINO-containing genes at levels that exceed their normal derepressed levels by severalfold, whether inositol or choline are present or not (White ; Santiago and Mamoun 2003; Jesch ). Such high levels of expression lead to over-production of inositol (the Opi− phenotype for which the mutants are named) and the excess inositol excreted from Opi− mutants can be detected in a plate assay (Greenberg ,b). In addition, even when growing in the absence of inositol, cells have elevated levels of PI and levels of other lipids that resemble those of wild-type cells growing in the presence of inositol (Klig ).

Role of PA in regulation of UASINO-containing genes

Opi− phenotypes are also associated with mutants (, , and ) defective in the synthesis of PC via the CDP-DAG pathway (Greenberg ; Letts ; Summers ; McGraw and Henry 1989) (Figure 2). Similar to the mutant, and mutants overexpress and fail to repress and other UASINO-containing genes in response to inositol. However, in contrast to mutants, the constitutive overexpression of and Opi− phenotypes of and mutants are suppressed when they are grown in the presence of exogenous soluble precursors that restore PC biosynthesis by entering the Kennedy pathway downstream of their respective genetic lesions in PE methylation (Figure 2). Thus, the Opi− phenotype is eliminated and PC biosynthesis is restored in mutants grown in the presence of monomethylethanolamine (MME), dimethylethanolamine (DME), or choline (Summers ), while these phenotypes are corrected only by DME and choline in mutants (McGraw and Henry 1989). Such observations led to the hypothesis that the mechanism by which the yeast cell senses the presence of exogenous inositol, choline, and other phospholipid precursors, such as serine, ethanolamine, MME, and DME, involves their effect on phospholipid synthesis (Carman and Henry 1989).

Further support for this hypothesis came from studies involving strains carrying temperature-sensitive mutations of the essential PI/PC transport protein, Sec14. While a full discussion of the biology of Sec14 is beyond the scope of this review article, the role it has played in discovery that PA is the signaling molecule controlling expression will be covered in brief here. The growth and secretory phenotypes of mutants are suppressed by mutations in any one of the three genes encoding enzymes in the Kennedy pathway for PC biosynthesis (i.e., , , and ) (Cleves ). Such strains (for example: ) have Opi− phenotypes and excrete choline (Opc− phenotype) (Patton-Vogt ). These phenotypes are due to increased PA levels resulting from elevated turnover of PC, catalyzed by the -encoded phospholipase D (Sreenivas ; Xie ). These observations led to the identification of PA as the probable signaling lipid responsible for derepression of and coregulated genes in the absence of exogenous inositol (Griac ). This model is consistent with the Opi− phenotypes of mutants defective in reactions in the CDP-DAG pathway for PC synthesis and the fact that these phenotypes are suppressed by supplying these mutants with precursor that enters the Kennedy pathway downstream of the defect in the CDP-DAG pathway (Henry and Patton-Vogt 1998). Interruption of any of the reactions downstream of PA via the CDP-DAG pathway leads to a buildup of PA and overexpression of , which raises endogenous inositol production. This leads to excretion of inositol and renders the cell insensitive to exogenous inositol (Henry and Patton-Vogt 1998). Metabolites that enter the Kennedy pathway downstream of the specific metabolic lesion counteract PA build up by increasing the use of DAG in the Kennedy pathway (Figure 2). Also consistent with this model is the Opi− phenotype observed in cells engineered to have reduced expression of , thereby slowing PI synthesis and raising PA levels (Jani and Lopes 2009).

The mechanism by which PA regulates expression was elucidated when Loewen showed that Opi1 contains PA binding domains that facilitate its interaction with PA in the perinuclear ER in cells growing in the absence of inositol (Figure 3). Opi1 also contains a motif known as FFAT (two phenylalanines in an acidic tract) that binds to Scs2, an integral ER membrane protein (Loewen ). To remain localized to the ER membrane in the absence of inositol, Opi1 must interact with both Scs2 and PA. In wild-type yeast cells, addition of inositol to the growth medium leads to a dramatic increase in the rate of synthesis of PI, resulting in consumption of both PA and CDP-DAG. As PA levels drop, Opi1 is released from the ER and enters the nucleus where it represses expression of and other UASINO-containing genes (Loewen ). Consistant with this model, mutations in Opi1 that interfere with its interaction with either PA or Scs2 result in Ino− phenotypes (Loewen ) and mutants also exhibit an Ino− phenotype (Nikawa et al. 1995). Increased PC synthesis via the Kennedy pathway is also characteristic of mutants, a phenotype that has not been fully explored, and mutations blocking PC synthesis by this route suppress both the Ino− phenotype and increased synthesis of PC in the mutant (Kagiwada and Zen 2003). Since, as described above, blocking the reuse of choline via the Kennedy pathway should reduce the draw on PA and thus enhance expression, these phenotypes are consistent with the model of Loewen concerning Opi1 function (Figure 3).

Regulation of gene expression by zinc

The expression of several phospholipid synthesis genes is also regulated in coordination with mechanisms that control zinc homeostasis (Figure 3) (Carman and Han 2007; Eide 2009). Zinc is an essential nutrient required for the growth and metabolism of S. cerevisiae (Eide 2009). It is a cofactor for key metabolic enzymes and a structural component of a diverse set of proteins that include chaperons, lipid binding proteins, and transcription factors (Vallee and Falchuk 1993; Schwabe and Klug 1994; Ellis ; Eide 2009). Cells grown in medium lacking zinc exhibit the induced expression of plasma membrane and vacuolar membrane zinc transporters (encoded by , , , and ) to maintain the cytosolic levels of zinc (Guerinot and Eide 1999; Eide 2003, 2009). At the same time, a reduction in zinc causes changes in membrane phospholipid composition that are brought about by changes in the expression of phospholipid synthesis enzyme activities (Han , 2005; Iwanyshyn ; Kersting and Carman 2006; Carman and Han 2007). It is postulated that changes in phospholipid composition may govern the insertion, topology, and/or function of the membrane-associated zinc transporters (Carman and Han 2007).

The regulation of phospholipid synthesis genes by zinc availability (Figure 3) involves the control of PA content through the activation of PI synthase function. This regulation occurs in the absence of inositol supplementation and is mediated by the zinc-sensing and zinc-inducible transcriptional activator Zap1 and the zinc-responsive cis-acting element (UASZRE) (Carman and Han 2007). Zinc depletion causes an increase in the synthesis of PI through increased expression of PI synthase (Iwanyshyn ; Han ). This regulation is controlled by the interaction of Zap1 with a UASZRE in the promoter (Iwanyshyn ; Han ). As discussed above, increased PI synthesis causes a decrease in PA content, which results in the Opi1-mediated repression of UASINO-containing genes and a decrease in the activities of the CDP-DAG pathway enzymes (Iwanyshyn ). The major effects of zinc depletion on phospholipid composition include an increase in PI and a decrease in PE (Iwanyshyn ). Although levels of enzymes in the CDP-DAG pathway are reduced by zinc depletion, the amount of PC is not significantly affected (Iwanyshyn ). This is attributed to the Zap1-mediated inductions of choline kinase and ethanolamine kinase for PC synthesis via the Kennedy pathway (Kersting and Carman 2006; Soto and Carman 2008). Like , the and genes contain a UASZRE in their promoters that interact with Zap1 for gene activation (Kersting and Carman 2006; Soto and Carman 2008). Opi1-mediated regulation of and are overcome by their derepression via Zap1 (Kersting and Carman 2006; Soto and Carman 2008).

Among the genes in lipid metabolism that contain the UASZRE, is the most highly regulated by zinc availability (Lyons ; Carman and Han 2007; Eide 2009). The gene encodes DGPP phosphatase, an enzyme associated with the vacuole membrane (Wu ; Toke ; Han ). This enzyme catalyzes the removal of the β-phosphate from DGPP, a minor phospholipid in yeast, to form PA, followed by the dephosphorylation of PA to form DAG (Wu ). The zinc-mediated regulation of expression correlates with the metabolism of DGPP and PA in the vacuole membrane (Han ). In zinc-replete medium, DGPP and PA account for 0.6 mol% and 1.4 mol% of the total phospholipids in vacuole membranes, but in zinc-depleted medium, the amounts of DGPP and PA are decreased to an undetectable level and 0.3 mol%, respectively (Han ). The function of -encoded DGPP phosphatase is still unclear. However, it is speculated that the enzyme controls the levels of DGPP and PA in vacuolar membranes, which in turn mediates the cellular functions occurring in response to zinc depletion (Carman and Han 2007).

Regulation by phosphorylation

Phosphorylation is a major covalent post-translational modification by which the activity of an enzyme or a transcription factor is regulated (Karin and Hunter 1995; Calkhoven and Ab 1996; Hung ; Kaffman ; Komeili and O'Shea 1999; Liu ). Global analyses of protein phosphorylation indicate that several enzymes and transporters of glycerolipid metabolism are subject to phosphorylation (Tables 1–3). Some proteins have many sites of phosphorylation, whereas others have only a few. It is also telling that many enzymes and transporters have no sites of phosphorylation, and thus the function of these proteins might be regulated by other mechanisms (e.g., substrate availability). The identity of the protein kinases involved and the physiological consequences of their phosphorylations have only been determined for a few proteins in glycerolipid metabolism (Carman and Han 2011). The protein kinases known to regulate the function of catalytic and regulatory proteins in glycerolipid metabolism include AMP-activated protein kinase, protein kinases A and C, casein kinase II, and cyclin-dependent kinase. Glycerolipid enzymes known to be bona fide substrates of protein kinases and regulated by phosphorylation include CTP synthetase, choline kinase, PS synthase, PA phosphatase, and TAG lipase. The repressor Opi1 is also regulated by phosphorylation. A discussion of these phosphorylations may be found in a recent review by Carman and Han (2011).

Regulation of PA phosphatase

Pah1 is one of the most highly regulated enzymes in lipid metabolism. Its activity is governed by several of the biochemical mechanisms discussed above including phosphorylation, enzyme location, and modulation by components of lipid metabolism. As discussed above, the DAG generated in the Pah1 reaction is used for the synthesis of TAG (Han ) and for the synthesis of PE and PC via the Kennedy pathway (Carman and Han 2006, 2008). The enzyme also plays a major role in controlling the cellular concentration of its substrate PA (Figure 3) (Han ), the precursor of phospholipids that are synthesized via the CDP-DAG pathway (Carman and Zeimetz 1996; Carman and Henry 1999; Carman and Han 2008). In addition, the substrate PA plays a signaling role (see above) in the transcriptional regulation of phospholipid synthesis genes (Carman and Henry 2007). Δ mutants exhibit a >90% reduction in TAG content, as well as derepression of phospholipid synthesis genes and massive expansion of the nuclear/ER membrane (Santos-Rosa ; Han ). Thus, the regulation of Pah1 activity governs the synthesis of TAG, the pathways by which phospholipids are synthesized, PA signaling, and the growth of the nuclear/ER membrane (Carman and Han 2008). Pah1 is associated with the cytosolic and membrane fractions of the cell, and its association with the membrane is peripheral in nature (Han ). Chromatin immunoprecipitation analysis indicates that Pah1 may also be localized in the nucleus (Santos-Rosa ).

The association of Pah1 with the membrane where its substrate PA resides is essential to its function in vivo, and membrane association is largely governed by the phosphorylation state of the enzyme (Karanasios ; Choi ). Phosphorylation favors a cytosolic association, whereas dephosphorylation favors a membrane association (Choi ). The important sites of phosphorylation that govern this regulation include the seven Ser/Thr-Pro targets for -encoded and -encoded cyclin-dependent kinases. Pah1 is dephosphorylated by the Nem1-Spo7 phosphatase complex that is associated with the nuclear/ER membrane (Siniossoglou ; Santos-Rosa ). In the absence of Nem1-Spo7, wild-type Pah1 is enriched in the cytosol where it is physiologically inactive, whereas a nonphosphorylatable mutant of Pah1 with alanine substitutions of Cdc28 and Pho85 target sites is enriched in the membrane and is physiologically active (Choi ). The requirement of Nem1-Spo7 indicates that Pah1 is recruited to the membrane for its physiological function. Indeed, the Nem1-Spo7–dependent membrane localization of Pah1 is enhanced by elevated levels of PA (Karanasios ). Once the enzyme is dephosphorylated, it anchors onto the nuclear/ER membrane via a short N-terminal amphipathic helix allowing for the production of DAG for TAG synthesis (Karanasios ).

Under normal physiological conditions (i.e., presence of the Nem1-Spo7 complex), the level of wild-type Pah1 detected on the membrane is very low. In fact, microscopic analysis of live S. cerevisiae cells expressing Pah1-GFP show a cytosolic localization without a detectable fluorescence signal associated with the nuclear/ER membrane unless PA levels are elevated (Karanasios ). Yet we know that Pah1 is physiologically active with respect to lipid metabolism throughout cell growth (Han ). Purified Pah1 has a relatively high catalytic efficiency when compared with other enzymes of lipid metabolism (Lin and Carman 1989), and the lethal phenotype of cells that overexpress Nem1-Spo7 (Santos-Rosa ) indicates that an excess of Pah1 function is detrimental to cell physiology. Indeed, suppression of this phenotype by overexpression of Pah1 (Santos-Rosa ) led to the discovery of Pah1 function (Han ). Thus, under normal physiological conditions, the amount of Pah1 associated with membranes must be small for its physiological function to be controlled, and this regulation must be mediated by the amount of the Nem1-Spo7 on the membrane. In support of this hypothesis, the expression level of Nem1 is 10-fold lower when compared with Pah1 (Ghaemmaghami ).

Pah1 activity is also modulated by cytosolic- and membrane-associated factors. The nucleotides ATP and CTP, which are precursors for the synthesis of phospholipids (Carman and Henry 1999), are inhibitors of Pah1 activity (Wu and Carman 1994). Indeed high levels of ATP favor elevated PA content and phospholipid synthesis, whereas low levels of ATP favor reduced PA content and an increase in the synthesis of TAG (Wu and Carman 1994). As discussed elsewhere, elevated CTP content favors an increase in PA content and derepression of UASINO-containing phospholipid synthesis genes (Ostrander ). The patterns of regulation of PA phosphatase activity by ATP and by CTP are consistent with the regulation of lipid synthesis observed in cells that have fluctuations in ATP and CTP (Wu and Carman 1994; Ostrander ).

Membrane lipids modulate Pah1 activity. For example, CDP-DAG, PI, and CL enhance activity (Wu and Carman 1996), whereas the sphingoid bases phytosphingosine and sphinganine inhibit activity (Wu ). The major effect of the lipid activators is to decrease the Km of Pah1 for PA. Sphinganine antagonizes the activation of PA phosphatase activity by CL and PI, whereas it causes an increase in the cooperativity of CL activation (Wu and Carman 1996). Conversely, sphinganine has little effect on the cooperativity of PI activation, but causes an increase in the activation constant for PI (Wu and Carman 1996). On the basis of the activation/inhibitor constants and cellular concentrations for these lipid effector molecules, their regulatory roles on Pah1 should be physiologically relevant (Wu ; Wu and Carman 1996).

DAG kinase counteracts PA phosphatase in regulating PA levels

Dgk1 is a unique CTP-dependent nuclear/ER membrane-associated enzyme that catalyzes the formation of PA from DAG (Han ,b). Dgk1 counteracts Pah1 in controlling PA content and consequently, transcriptional regulation of UASINO-containing genes (Han ,b). overexpression causes an increase in PA content, derepression of UASINO-containing genes, and abnormal nuclear/ER membrane expansion (Han ) like those that occur in the Δ mutant (Santos-Rosa ; Han ). Also consistent with the regulatory role of PA, overexpression suppresses the inositol auxotrophy caused by overexpression (Han ), whereas the Δ mutation suppresses the phenotypes caused by the Δ mutation (Han ,b).

Ino− and Opi− phenotypes are associated with mutations affecting many functions

To date, the Ino− and Opi− phenotypes of mutants affecting lipid metabolism, which have been examined in detail for correlation of changes in lipid metabolism with changes in expression, have been consistent with predictions based on the role of PA as discussed above (Carman and Henry 2007). However, mutants defective in a wide variety of other cellular functions have also been reported to have these phenotypes (Henry and Patton-Vogt 1998) and several genome-wide screens for mutants with Opi− or Ino− phenotypes have been conducted (Hancock ; Young ; Villa-Garciá et al. 2011).

The screen of the Matα viable yeast deletion collection conducted by Hancock , using the Opi− plate test, as well as an -LacZ reporter construct, identified 89 Opi− mutants. Among these mutants were , , , , , , and , which had previously been shown to affect expression (Henry and Patton-Vogt 1998; Shirra and Arndt 1999; Ouyang ; Elkhaimi ). Hancock also reported Opi− phenotypes for the first time in mutants affecting functions such as the NuA4 histone acetyl transferase, vacuolar protein sorting, and other aspects of membrane trafficking and the UPR.

Young identified 231 mutants from the yeast haploid deletion collection that showed measurable growth reduction at 37°, when grown in medium lacking inositol. The major focus of this study was on the discovery of pH sensitive Ino− phenotypes associated with mutants defective in all of the subunits of the vacuolar adenosine triphosphatase (V-ATPase) complex, as well as factors in the ER responsible for the assembly of this complex. The mutant, defective in the essential P-type ATPase of the plasma membrane, which regulates cellular pH, was also found to have an Ino− phenotype, as was , defective in a K+ transporter that activates Pma1. The Ino− phenotypes of these mutants were observed at pH 3 but were suppressed at pH 4 and 5 and this pH sensitivity was attributed to pH dependence of binding of Opi1 to PA. The binding of PA to Spo20, a yeast SNARE (Nakanishi ), was also shown to be pH sensitive, suggesting that binding of proteins to PA is affected by cellular pH. Changes in the ionization state of PA affect binding of Opi1 to PA in vitro, supporting the electrostatic/hydrogen bond switch model of Kooijman for interaction of proteins with PA. Repression of in response to glucose starvation also correlates to cellular pH, but not to PA levels. Young have concluded that the Reg1-mediated mechanism that regulates expression via the glucose response pathway (Ouyang ; Shirra and Arndt 1999) is most likely not dependent on the absolute level of PA, but rather on the effect of pH on the ionization state of PA.

The screen for Ino− mutants of the diploid viable yeast deletion collection conducted by Villa-Garciá et al. (2011) has resulted in the identification of 419 genes, which when deleted confer the Ino− phenotype under one or more growth conditions. This screen involves comparing growth in the presence and absence of inositol at two growth temperatures, 30° and 37°, in the presence and absence of choline. Choline sensitive Ino− phenotypes and the strengthening of weak Ino− phenotypes by choline are potentially attributable to reduction in PA levels due to the consumption of DAG produced from PA when PC is synthesized via the Kennedy pathway (see above) using exogenous choline (Carman and Henry 2007). Growth temperature has also been shown to influence lipid metabolism in wild-type and mutant strains. Growth at 37° results in an increased rate of synthesis of PI in wild-type cells growing in the presence of inositol (Gaspar ), as well as an increase in PC synthesis and turnover (Dowd ). Among the gene ontology (GO) categories enriched among the Ino− mutants identified by Villa-Garciá et al. (2011) are: response to stress, protein modification, chromosome organization, response to chemical stimulus, cellular carbohydrate metabolism, psuedohyphal growth, and transcription. Mutants in the category response to stress include a number with defects in stress response pathways that had previously been reported to be associated with Ino− phenotypes, including the protein kinase C-cell wall integrity (PKC-CWI) (Nunez ; Jesch ), the UPR (Cox ; Chang ) and the glucose response pathway (Shirra , 2005). However, Ino− phenotypes have also been observed in mutants affecting signaling pathways not previously reported to be associated with such phenotypes including: target of rapamycin (TOR), high osmolarity glycerol (HOG), cAMP-dependent protein kinase, calcineurin and filamentous growth and cell cycle regulation (Villa-Garciá et al. 2011).

Causes of Ino− phenotypes

It is often assumed that an Ino− phenotype provides evidence that a mutant has impaired expression or, alternatively, that a signaling pathway and/or a transcription factor defective in the mutant in question, is directly involved in regulating transcription. For example, the Ino− phenotypes of the and mutants, defective in UPR activation, have been attributed to a role for the Hac1 transcription factor in activating expression of (Cox ). The evidence for this interpretation includes the observations that and mutants fail to sustain expression of when shifted to medium lacking inositol and that the mutation supresseses their Ino− phenotypes (Cox ). Cox also reported that growth in the absence of inositol activates the UPR in wild-type cells and that expression is activated in wild-type cells treated with tunicamycin in the presence of normally repressing levels of inositol. However, activation of expression in response to tunicamycin in the presence of inositol was not observed by Chang and is repressed with normal kinetics in response to addition of inositol in cells in which the UPR was constitutively induced by constitutive expression of activated Hac1 (Jesch ). The reasons for these divergent results with respect to the role of Hac1 in transcription are not clear.

A number of mutants in the PKC-CWI pathway have Ino− phenotypes that are intensified at 37° and in the presence of choline (Nunez ; Fernandez-Murray ; Jesch ; Villa-Garciá et al. 2011). Futhermore, the Ino− phenotypes of these mutants are suppressed by . It is clear, however, that the PKC pathway has no direct role in regulating expression. The PKC pathway mutant, , exhibits normal derepression in response to a shift to inositol medium even when choline is present at 37°, a condition under which growth ceases and the mutant begins to lose viability within 4–5 hr (Nunez ). The and mutants, defective in the PI kinases that are responsible for producing the PI 4-P and PI 4,5-P2 pools on the plasma membrane and are essential for PKC-CWI signaling during heat stress (Audhya and Emr 2002; Audhya and Emr 2003), also exhibit Ino− phenotypes at semipermissive growth temperatures (Jesch ). However, undergoes normal derepression immediately following a shift to medium lacking inositol in the and mutants, even at the temperature at which they exhibit an Ino− phenotype (Jesch ). Significantly, the PKC-CWI pathway is activated in wild-type cells grown in the absence of inositol (Nunez ) in response to reduction in synthesis of inositol-containing sphingolipids (Jesch ). Thus, it is clear that the Ino− phenotypes of PKC-CWI mutants are not attributable to a defect in expression, but rather to failure to mount a stress response that is essential for survival of stress caused by growth in the absence of inositol (Nunez ; Jesch ; Villa et al. 2011).

Supression of an Ino− phenotype by is also often cited as evidence that the mutant in question is defective in a functon that regulates . However, the mutation also suppresses the Ino− phenotypes of and other PKC pathway mutants (Nunez ), which, as discussed above, are not attributable to failure to express . The explanation for this observation lies in the fact that the mutation does not simply restore expression to the level seen in wild-type cells growing in the absence of inositol. Instead, the deletion of the gene completely eliminates the ability of the cell sense inositol, whether endogenously produced or exogenously supplied, leading expression of at levels as high as 5- to 6-fold higher than normal derepressed levels (Bachhwat et al. 1995). As a consequence of the resulting overproduction of inositol, cells have levels of PI and other lipids resembling the lipid composition of wild-type cells grown in the presence of high levels of exogenous inositol (Klig ). Indeed, mutants excrete so much inositol that they support growth of any Ino− mutant in their vicinity, the characteristic used in the original bioassay used in the isolation of Opi− mutants (Greenberg et al. 1982a). Microarray experiments showed that, in contrast to wild-type cells, cells do not exhibit activation of UPR target genes when grown in the absence of inositol (Jesch ). This is presumably due to the fact that overexpression of the gene in cells produces sufficient inositol to repress expression of all genes that are normally activated in absence of inositol, except of course the UASINO-containing genes, which are directly dependent on Opi1p for repression in the presence of inositol (Jesch and Henry 2005). Thus, suppression by of an Ino− phenotype conferred by mutation of a given gene does not consitute sufficient evidence on its own that the gene product in question is involved in regulation of .

Microarray studies have revealed that the transcript levels of hundreds of genes in yeast are affected by the availability of inositol and/or inositol plus choline in the medium (Santiago and Mamoun 2003; Jesch , 2006; Nunez ). Most of these “inositol-regulated” genes are not under the control of Opi1 and/or involved in lipid metabolism and a number of these genes are known targets of the stress response pathways that are associated with Ino− phenotypes, including the UPR and PKC pathways (Jesch , 2006; Nunez ; Jesch ; Villa et al. 2011). As discussed above, expression of at the level observed in wild-type cells growing in the absence of inositol does not suffice to support the rate of PI synthesis observed in the same cells growing in the presence of exogenous inositol (Kelley ; Gaspar ). Moreover, in wild-type cells, the changes in lipid metabolism that occur in response to removal of inositol and/or its resupply have been shown to affect signaling associated with PA (Loewen ; Young ), and with inositol containing sphingolipids and the pools of PI 4-P and PI 4,5-P2 in the plasma membrane (Jesch ).

As the exploration of the regulation and cell biology of lipid metabolism in yeast accelerates in response to genomic approaches, additional examples of lipid-mediated signaling influenced by the availability of lipid precursors are likely to be discovered. A major advantage of yeast in such studies is the fact, as described in this YeastBook chapter, that its lipid metabolism can readily be manipulated by changing the supply of exogenous precursors and/or by introducing mutations affecting specific steps in lipid metabolism.

Perspectives

Monumental advancements have been made in determining gene enzyme relationships and elucidating the fundamental biochemistry, regulatory mechanisms, and cellular roles of lipids in the 20 years since the publication of the last edition of the Molecular Biology of the Yeast Saccharomyces: Gene Expression in 1992. Yet while the identities of many of the genes encoding enzymes, transporters, and regulatory factors controlling lipid metabolism in yeast have been discovered, described, and analyzed, there remain many gaps in our knowledge. The functions of many genes, genome-wide, have yet to be clarified and among them are certainly a considerable number of “missing links” in yeast lipid metabolism and its regulation. For example, the molecular organization of lipid biosynthetic pathways that ensures metabolic channeling of critical compounds, enzyme topology, and the integration of soluble and membrane-bound intermediates in these pathways is only beginning to be addressed. Furthermore, we do not have complete knowledge of the mechanisms by which lipids synthesized in one compartment are trafficked to other locations within the cell or the mechanisms by which the precise lipid compositions of specific membrane compartments are established and maintained. High-resolution lipidomic analyses that have uncovered some 100 different molecular species of glycero- and sphingolipids have added another level of complexity to this question. Moreover, our understanding of the regulatory networks that link lipid metabolism to other aspects of cell biology and metabolism, including energy metabolism, membrane biogenesis, and trafficking, signaling, and cell division, remains limited. However, the increasing availability of archived genetic and biochemical data on a wide variety of eukaryotic organisms and the extensive homologies that exist among eukaryotes are opening new avenues for pursuing these important challenges. Due to its genetic tractability and comparably “simple” lipid complement, yeast continues to be among the most attractive and important model systems for such research.