INTRODUCTION

Sphingolipids, essential components of eukaryotic plasma membranes, consist of a hydrophilic head attached to a ceramide. Ceramides contain a fatty acid attached to a sphingoid base through an amide linkage (Fig. 1). They can be classified according to their level of hydroxylation (1); both the sphingoid and the fatty acid moieties are found with different levels of hydroxylation (Fig. 1). In mammals, the sphingoid moiety is mostly sphingosine, which is desaturated at C-4,5; however, some is phytosphingosine that is hydroxylated at C-4, or dihydrosphingosine, which is neither desaturated at C-4,5 nor hydroxylated at C-4 (2, 3). In the yeast Saccharomyces cerevisiae, the C-4 is mostly hydroxylated (1). The fatty acid that is attached to the sphingoid base is either un-, mono-, or dihydroxylated (1). In yeast, the first hydroxylation of the fatty acid moiety occurs in the endoplasmic reticulum, and the second hydroxylation is in the Golgi apparatus (4) and requires Cu²⁺ and the Golgi copper transporter encoded by CCC2 (5).

[View Larger Version of this Image (19K GIF file)]

The physiological role of the different hydroxylation states is not known. Hydroxylation of ceramide and sphingolipids may alter their cellular location, their effect on the physical properties of membranes, and their interaction with proteins either as a substrate or regulator. Identification of the genes and proteins required for the hydroxylation reactions will facilitate the investigation of the function of the hydroxyl groups.

The S. cerevisiae protein Scs7p is required for the first hydroxylation of the ceramide fatty acid moiety (6). This enzyme belongs to a family of desaturase/hydroxylase enzymes that contain an oxo-diiron domain (Fe-O-Fe) (7, 8). This domain consists of four transmembrane segments. The loop between the second and third transmembrane segments has a histidine-containing motif (HX_3,4HX_8-31HX_2,3HH). Another histidine-rich motif (HX_2,3HH or HX_2,3HX_13-39HX_2,3HH) follows the fourth transmembrane segment. Sur2p also contains the oxo-diiron motif (9).

The SUR2 gene was initially identified in a screen for suppressors of rvs161 mutants (10). Rvs161p is required for endocytosis (11), correct actin localization (12), and viability upon nitrogen, carbon, or sulfur starvation (13). It is similar to amphiphysin, a neuronal protein found in synaptic vesicles that is the autoantigen in stiff-man syndrome (14, 15). The molecular function of Rvs161p and the basis of suppression by mutations in the SUR genes have not been identified. However, other SUR genes have been found to function in sphingolipid synthesis. SUR1 is allelic to CSG1, a gene required for mannosylation of sphingolipids (5). SUR4/ELO3 encodes a fatty acid elongase required for the synthesis of the very long chain fatty acids (VLCFA)¹ found in ceramide and sphingolipid (16). The genetic relationship between SUR2, SUR1, and SUR4 suggests that SUR2 may also be involved in sphingolipid synthesis.

Based on the homology of Sur2p with Scs7p, required for hydroxylation of the fatty acid of sphingolipids (6), and the genetic relationship between SUR2 and other sphingolipid synthesis genes, the possibility that SUR2 is required for hydroxylation of C-4 on the long chain base (LCB) found in sphingolipids was investigated.

EXPERIMENTAL PROCEDURES

The yeast strains used in this study were TDY2037 (Mat lys2 ura3-52 trp1 leu2), TDY2038 (Mat lys2 ura3-52 trp1 leu2 csg2::LEU2), 2037scs7 (Mat lys2 ura3-52 trp1 leu2 scs7::LEU2), and 6715b (Mat lys2 ura3-52 trp1 leu2 csg2::LEU2 scs7::LEU2). SUR2 was disrupted in these strains as described below. Media were prepared, and cells were grown using standard procedures (17). Phytosphingosine, dihydrosphingosine, and sphingosine were purchased from Sigma and added to the growth medium at 25 µM in 1% Tergitol.

In an unrelated study we isolated a YCp50-based plasmid containing a fragment of yeast DNA that included the amino terminus (through to a Sau3A site at codon 120) of the SUR2 gene. A restriction fragment extending from the HindIII site 367 base pairs upstream of the start codon of the SUR2 gene to the SaII site in YCp50 was subcloned from this plasmid into pUC19. The resulting plasmid was linearized at the PstI site in codon 9 of the SUR2 gene, treated with Bal-31 to remove about 100 base pairs, and incubated with dNTPs, Klenow fragment, ligase, and XhoI linkers. A candidate plasmid with a XhoI linker at the deletion junction that was missing about 50 base pairs from each side of the original PstI site was used to construct the disrupting plasmid. A SalI fragment carrying the TRP1 gene was ligated into the XhoI site, the SUR2-disrupting fragment was cut out of the pUC19 plasmid with PvuII and used in a one-step gene replacement (18). The disruption of SUR2 was confirmed using a polymerase chain reaction.

Cells were grown in synthetic minimal medium containing 12 nM inositol and 1 µCi/ml [³H]myoinositol for several generations (from A₆₀₀ 0.01 to A₆₀₀ 1.0). Cells (about 5 A₆₀₀ units) were pelleted and washed with 4 mM sodium azide. Lipids were extracted into 600 µl of CHCl₃:MeOH (1:1) by vortexing with glass beads, removing the CHCl₃:MeOH to a fresh tube and washing the cell pellet and beads with 600 µl of CHCl₃:MeOH:H₂O (10:10:3). The pooled extract was dried, alkali-treated, and BuOH-desalted as described previously (5, 19, 20). The samples were analyzed by TLC on silica gel plates using CHCl₃:MeOH:AcOH:H₂O (16:6:4:1.6) as the developing solvent (4).

Cells were grown in synthetic medium at 26 °C, spun down, and washed once with H₂O. The cells were vortexed with glass beads in hexane:EtOH (95:5) at 40 A₆₀₀/ml. The supernatant was transferred to a fresh tube, the pellet and beads were washed with hexane:EtOH, and the pooled extract was dried. The lipids from 60 A₆₀₀ units of cells were alkali-treated by suspending in 1 ml of EtOH:H₂O:Et₂O:pyridine (15:15:5:1) and adding KOH to 0.1 M followed by incubation at 37 °C for 3 h (21). After neutralizing with 1 M AcOH, the sample was dried, BuOH-desalted (20), and dried again. Ceramides were analyzed by TLC on silica gel plates using CHCl₃:MeOH:AcOH (95:4.5:0.5) as the developing solvent (22). Plates were sprayed with 10% copper sulfate in 8% orthophosphoric acid and heated for 20 min at 180 °C to char the ceramides (22). The arsenite and borate treated silica gel plates were supplied by Analtech (Newark, DE).

Ceramides were purified and separated by TLC as described above. The ceramides were visualized by ultraviolet light after spraying the plates with 0.01% 8-anilino-1-napthalenesulfonic acid. The silica gel was scraped off the plate, and the ceramides were eluted by repeated sonication (five times for 10 min each) in 2 ml of CHCl₃:MeOH (1:1).

Sphingolipids were extracted from 600 A₆₀₀ units of cells by vortexing with glass beads in 100 ml of CHCl₃:MeOH (1:1). The extract was dried, alkali-treated, and BuOH-desalted as described previously (5, 19, 20). The sample was spotted in a line on a silica gel plate and developed using CHCl₃:MeOH:AcOH (95:4.5:0.5). In this system, fatty acids and ceramides migrate while sphingolipids remain at the origin. The material left at the origin was subjected to acid methanolysis for analysis of the fatty acid methyl esters (FAMEs) and LCBs as described below.

Ceramides and sphingolipids were purified by silica gel TLC as described above. The purified ceramides or sphingolipids were subjected to acid methanolysis by resuspending in 2 ml of HCl:MeOH:H₂O (3:29:4) and incubating at 78 °C for 18 h (23). The FAMEs were recovered by extracting 3 times with 2 ml of hexane (24). The extracts were pooled, dried, and subjected to TLC using petroleum ether:Et₂O (17:3) as the developing solvent (24). Plates were sprayed with 10% copper sulfate in 8% orthophosphoric acid and heated for 20 min at 180 °C to char the FAMEs.

The LCBs were recovered from the hydrolyzed ceramides or sphingolipids by adjusting the pH of the acid hydrolysate (after extraction of FAMEs) to 11.5 using 1 M NaOH and extracting three times with 2 ml of Et₂O (24). The pooled extracts were dried, and the LCBs were separated by silica gel TLC using CHCl₃, MeOH, 2.5 M NH₄OH (40:10:1) as developing solvent (24). Plates were sprayed with 0.2% ninhydrin in ethanol and incubated at 100 °C for 5-10 min to visualize the amine-containing LCBs.

RESULTS

Sphingolipid Synthesis Is Altered in sur2 Mutant Cells

The main purpose of this study was to determine if SUR2 encodes the enzyme that hydroxylates C-4 of the sphingoid moiety of sphingolipids. The sur2 mutant was constructed as described under "Experimental Procedures." Sphingolipid synthesis in a sur2 mutant was compared with that of wild-type. Cells were grown for several generations in synthetic minimal medium containing 12 nM inositol with 1 µCi/ml [³H]inositol. [³H]Inositol was incorporated into phosphatidylinositol, inositolphosphorylceramide (IPC), mannosylinositolphosphorylceramide (MIPC) and mannosyldiinositolphosphorylceramide (M(IP)₂C). These lipids were extracted out of the cell, alkali-treated to remove phosphatidylinositol, and separated by TLC (Fig. 2). The sphingolipid composition of the sur2 mutant cells differs from that of wild-type cells (Fig. 2, lane 1 and 3). The predominant sphingolipid in wild-type cells is MIPC-C (lane 1, MC) along with its precursor IPC-C (C). IPC-D (D) containing dihydroxyl fatty acid is also observed (1). The major sphingolipids in sur2 mutants (lane 3) differ from those of wild-type cells. These sphingolipids are named MIPC-A and MIPC-B (lane 3, MA and MB) for reasons discussed below.

Analysis of sphingolipids from wild-type (lane 1), csg2 (lane 2), sur2 (lane 3), scs7 (lane 4), sur2scs7 (lane 5) csg2sur2 (lane 6), csg2scs7 (lane 7), and csg2sur2scs7 (lane 8) cells.

[View Larger Version of this Image (38K GIF file)]

Sphingolipid synthesis in a sur2csg2 double mutant (lane 6) was compared with that in a csg2 mutant (lane 2). The csg2 mutants are defective in mannosylation of inositolphosphorylceramide (17), therefore the sphingolipids in a csg2 mutant are IPC-C and IPC-D (lane 2, C and D). The sur2csg2 double mutant accumulates two sphingolipid species IPC-A and IPC-B (lane 6, A and B) that are not found in wild-type cells. IPC-A and IPC-B are more hydrophobic than is IPC-C (lane 6). The two sphingolipids found in a sur2 single mutant, MIPC-A and MIPC-B, (lane 3, MA and MB) are the mannosylated forms of the sphingolipid species seen in a sur2csg2 double mutant, IPC-A and IPC-B (lane 6, A and B).

The LCB species present in the inositolphosphorylceramides from sur2csg2 mutants (IPC-A and IPC-B) was determined. Polar lipids were extracted and alkali-treated to hydrolyze the glycerol-based phospholipids, and the sphingolipids were isolated by preparative TLC as described under "Experimental Procedures." The sphingolipids were subjected to acid methanolysis to hydrolyze both the phosphodiester bond between the inositol and the ceramide, and the amide bond between the sphingoid moiety and the VLCFA. The LCB was extracted and analyzed by TLC (Fig. 3A). The LCB in the sphingolipids from the sur2 mutant cells (lanes 3 and 5) is exclusively dihydrosphingosine, while the sphingolipids from cells with a wild-type SUR2 gene (lanes 1, 2, and 4) contain primarily phytosphingosine.

Analysis of the LCB moiety (panel A) and the VLCFA (panel B) of sphingolipids extracted from wild-type (lane 1), csg2 (lane 2), csg2sur2 (lane 3), csg2scs7 (lane 4), and csg2sur2scs7 (lane 5) mutants.

[View Larger Version of this Image (23K GIF file)]

The FAMEs released by methanolysis were also analyzed using a TLC system which resolves unhydroxylated very long chain fatty acid methyl esters (NVLCFAME) and hydroxylated very long chain fatty acid methyl esters (HVLCFAME). The sphingolipids from sur2csg2 mutant cells (lane 3) contain both hydroxylated and unhydroxylated fatty acids (Fig. 3B). Because sur2csg2 mutant cells synthesize two sphingolipids, IPC-A and IPC-B, these results indicate that IPC-A contains unhydroxylated fatty acids, whereas IPC-B contains hydroxylated fatty acids.

The hydroxylation of the VLCFA is dependent on Scs7p (Fig. 3B). Cells lacking both Scs7p and Csg2p synthesize an IPC-B species (Fig. 2, lane 7) (6, 19), which contains mostly phytosphingosine as the LCB (Fig. 3A, lane 4) and an unhydroxylated VLCFAME (Fig. 3B, lane 4). Like IPC-B, IPC-B can be mannosylated if Csg2p is present (Fig. 2, lane 4, MB) (6, 19).

In a sur2scs7csg2 triple mutant where hydroxylation of C-4 of the LCB is blocked by deletion of SUR2, hydroxylation of the VLCFA is blocked by deletion of SCS7, and mannosylation is blocked by deletion of CSG2, the only sphingolipid synthesized is the very hydrophobic IPC-A (Fig. 2, lane 8, A). The IPC-A species can be mannosylated if Csg2p is present (Fig. 2, lane 5, MA). The sur2 mutants accumulate some IPC-A or MIPC-A even when Scs7p is present (Fig. 2, lanes 3 and 6, MA and A) suggesting that phytosphingosine-containing substrates are preferred by Scs7p over dihydrosphingosine-containing substrates. These data (summarized in the model shown in Fig. 1) support the proposal that Sur2p is the hydroxylase that converts dihydroceramide to phytoceramide.

S. cerevisiae cells can incorporate exogenous phytosphingosine into sphingolipids (25). As would be predicted if Sur2p is required for hydroxylation of C-4 on the LCB, exogenous phytosphingosine, but not dihydrosphingosine, restores synthesis of IPC-C to a sur2csg2 mutant (Fig. 4, lanes 8 and 9). A sur2scs7csg2 triple mutant, that normally makes IPC-A, makes IPC-B in the presence of phytosphingosine (data not shown). These observations provide further evidence that the altered sphingolipids that accumulate in the sur2csg2 and sur2csg2scs7 mutants (IPC-B and IPC-A, respectively) differ from the sphingolipids in csg2 and csg2scs7 mutants (IPC-C and IPC-B, respectively) in the LCB. Addition of phytosphingosine to the scs7csg2 mutant does not result in any IPC-C synthesis (Fig. 4, lane 11), because the IPC-B that accumulates in the scs7 mutant arises from failure to hydroxylate the VLCFA (6).

Exogenous phytosphingosine restores synthesis of IPC-C to the csg2sur2 double mutant.

[View Larger Version of this Image (50K GIF file)]

The IPCs are synthesized by the transfer of phosphoinositol from phosphatidylinositol to ceramide. Comparison of ceramides isolated from the sur2csg2, scs7csg2 and the sur2scs7csg2 mutants with ceramides from wild-type or csg2 mutant cells demonstrates that the altered mobility of the sphingolipids arises from differences in the ceramide moiety of the sphingolipids. Yeast ceramides were analyzed by TLC (Fig. 5). The predominant ceramide in wild-type and in csg2 mutant cells (lanes 3 and 4), labeled "C" (for C-ceramide) because it is the ceramide of IPC-C, migrates slower in this TLC system than the bovine hydroxylated ceramide standard (Sigma type IV, lane 2) which consists of sphingosine and a hydroxylated fatty acid. Analysis of the LCB and VLCFA of the C-ceramide by TLC following methanolysis confirms that it contains phytosphingosine (Fig. 6A, lanes 5 and 6) and a HVLCFA (Fig. 6B, lanes 6 and 7) (1, 26). The slower mobility of yeast ceramide compared with hydroxylated bovine ceramides is expected, since it contains an additional hydroxyl group (phytosphingosine versus sphingosine).

Analysis of ceramides from wild-type, csg2, csg2sur2, csg2scs7, and csg2sur2scs7 cells.

[View Larger Version of this Image (111K GIF file)]

The LCBs and the FAMEs from A-, B-, B-, and C-ceramides were analyzed by silica gel TLC.

[View Larger Version of this Image (64K GIF file)]

The major ceramide from sur2 mutant cells (B-ceramide) has a mobility similar to the hydroxylated bovine ceramide standard. Ceramides having dihydrosphingosine might be expected to have similar hydrophobicity to those having sphingosine. The LCB from the B-ceramide is dihydrosphingosine (Fig. 6A, lane 7) and the VLCFA is hydroxylated (Fig. 6B, lane 8).

The B-ceramide that accumulates in the scs7 mutant cells contains phytosphingosine (Fig. 6A, lane 8) and unhydroxylated VLCFA (Fig. 6B, lane 9). The mobility of the ceramide in scs7 mutant cells is quite distinct from that in sur2 mutant cells (Fig. 5, lanes 5 and 6). Either hydroxylation of the VLCFA increases the hydrophilicity of the ceramide less than does the hydroxylation of the LCB, or these species interact differently with the silica gel matrix.

The A-ceramide that is present in the sur2scs7 double mutant (Fig. 5, lane 7) migrates with the unhydroxylated bovine ceramide standards as would be expected if it lacks hydroxyl groups on both C4 of the LCB and on the VLCFA (Fig. 5, lanes 1 and 7). The LCB of the A-ceramide is dihydrosphingosine (Fig. 6A, lane 9) and the VLCFA is unhydroxylated (Fig. 6B, lane 10). The absence of vicinal hydroxyl groups (C-3 and C-4 of phytosphingosine) on ceramide from sur2 mutants is also indicated by the effect of the glycol-complexing ions arsenite and borate on the chromatographic behavior of the B- and A-ceramides (Fig. 7). The complex between vicinal hydroxyl groups with arsenite increases their mobility on silica gel, while the borate complex decreases their mobility (27, 28). The mobility of C-ceramide (from wild-type and csg2 mutant cells) and B-ceramide (from csg2scs7 mutant cells) is greatly increased by addition of NaAsO₂ to the silica gel (compare Fig. 7, B to A, lanes 1, 2, and 4) and reduced by addition of Na₂B₄O₇ (Fig. 7C, lanes 1, 2, and 4), indicating that these ceramides contain the C-3,4 vicinal hydroxyl groups of phytoceramide. The mobility of the B-ceramide (from csg2sur2 mutant cells) and A-ceramide (from csg2sur2scs7 mutant cells) is much less affected by arsenite or borate, consistent with the conclusion that they have dihydrosphingosine instead of phytosphingosine as the LCB.

Effect of arsenite and borate on the relative chromatographic mobilities of the C (lanes 1 and 2)-, B (lane 3)-, B (lane 4)-, and A (lane 5)-ceramides. The isolated ceramides used in the experiment described in Fig. 6 were analyzed by TLC on silica gel plates without (panel A) or with either 1% sodium meta arsenite (panel B) or 1% sodium borate (panel C) as described by Karlsson and Pascher (27). The borate plate and the untreated plate were run once in CHCl₃:CH₃OH (95:5), while the arsenite plate was run twice in CHCl₃:CH₃OH:AcOH (95:4.5:0.5).

[View Larger Version of this Image (34K GIF file)]

Cells lacking the CSG2 gene are defective in mannosylation of inositolphosphorylceramides and therefore accumulate the inositolphosphorylceramide, IPC-C (Fig. 2, lane 2) (5, 19). Overaccumulation of IPC-C or a related metabolite confers Ca²⁺-sensitivity. Mutations in a variety of genes required for the synthesis of IPC-C suppress the Ca²⁺-sensitive phenotype of the csg2 mutants. For example, deletion of SCS7, which encodes the enzyme that hydroxylates the VLCFA suppresses the Ca²⁺ sensitivity of the csg2 mutant (Fig. 8) (6, 19). Therefore, the effect of deletion of SUR2 on the Ca²⁺ sensitivity of the csg2 mutant was investigated. As shown in Fig. 8, deletion of the SUR2 gene reverses the Ca²⁺ sensitivity of a csg2 mutant.

Suppression of the Ca²⁺-sensitive phenotype of csg2 mutants by deletion of SUR2 or SCS7.

[View Larger Version of this Image (55K GIF file)]

DISCUSSION

The effect of deleting SUR2 on the hydroxylation of C-4 on the LCB of ceramide and sphingolipid and the sequence similarity between Sur2p and a family of desaturases/hydroxylases indicate that Sur2p catalyzes the hydroxylation of C-4. The LCB of ceramides and sphingolipids in sur2 mutants is dihydrosphingosine instead of the phytosphingosine predominantly found in wild-type cells. Exogenous phytosphingosine restores synthesis of sphingolipids with a phytosphingosine LCB in sur2 mutants.

The substrate (dihydrosphingosine or dihydroceramide) for Sur2p has not been identified. Since hydroxylation of C-4 is not required for ceramide or sphingolipid synthesis, either dihydrosphingosine or phytosphingosine can serve as substrate for ceramide synthase, and either dihydroceramide or phytoceramide can serve as substrate for IPC synthase. S. cerevisiae cells contain both dihydrosphingosine and phytosphingosine, and inhibition of ceramide synthase by fumonisin B₁ causes the accumulation of both LCBs (29). However, it is not known whether phytosphingosine comes from de novo synthesis or from turnover of ceramide and sphingolipid.

Scs7p is also a member of this family of hydroxylases/desaturases, but it is responsible for hydroxylation of the VLCFA rather than the LCB. Failure to hydroxylate C4 of the LCB decreases the Scs7p-catalyzed hydroxylation of the VLCFA, indicating hydroxylation occurs subsequent to ceramide formation. Furthermore, the ceramides from SCS7⁺ cells are hydroxylated, while those from scs7 mutant cells are not suggesting that the substrate for Scs7p is ceramide. However, it is not yet known whether most of the free ceramides in the cell arise from de novo synthesis or from turnover of sphingolipid, so it remains to be determined whether the substrate for Scs7p is free ceramide or inositolphosphorylceramide.

Martin and co-workers (16) recently reported that cells lacking the elongase encoded by ELO3/SUR4 accumulate relatively high levels of hydroxylated C16 fatty acids. We have found that elo3 mutant cells incorporate fatty acids with shorter than normal chain lengths into ceramide.² Therefore, it will be interesting to determine whether the hydroxylated C16 fatty acids in the mutants arise from Scs7p-catalyzed hydroxylation of the (shorter than normal) fatty acids on the ceramide.

Ceramide and IPC-C are synthesized in the endoplasmic reticulum (4) which appears to be the location of Scs7p and Sur2p as well. Both Scs7p and Sur2p contain C-terminal sequences (KMKYE and VKKEK), matching a consensus sequence specifying retention in the endoplasmic reticulum (6, 30). In S. cerevisiae, all five proteins that are members of the oxo-diiron family appear to reside in the endoplasmic reticulum. Along with Sur2p and Scs7p, these are -9 fatty acid desaturase (Ole1p), C-4 sterol methyl oxidase (Erg25p), and C-5 sterol desaturase (Erg3p). The oxo-diiron centers in these enzymes are believed to receive electrons from either cytochrome b₅ or a cytochrome b₅-like domain. Scs7p and Ole1p contain cytochrome b₅-like domains at their N and C termini respectively (6, 31). Cytochrome b₅ may function to transfer electrons to Sur2p and the other two enzymes. Cytochrome b₅ reductase may catalyze the reduction of both cytochrome b₅ and the cytochrome b₅-like domains on Scs7p and Ole1p.

Suppressors of the Ca²⁺-sensitive Phenotype of csg2 Mutants, as Well as Suppressors of the Pleiotropic Phenotypes of rvs161 Mutants, Identify Sphingolipid Synthesis Genes

The Ca²⁺ sensitivity of csg2 mutants is suppressed by deletion of SUR2. Other mutations in sphingolipid biosynthetic genes (subunits of serine palmitoyltransferase, LCB1, SCS1/LCB2; ceramide hydroxylase, SCS7; fatty acid elongases, ELO2/SUR5/FEN1, ELO3/SUR4; and fatty acid synthetase, FAS2) also suppress the Ca²⁺ sensitivity of csg mutants. These mutations either decrease the rate of sphingolipid synthesis or alter the sphingolipids that are synthesized. The CSG2 and CSG1 genes are required for mannosylation of IPC to form MIPC. In the absence of mannosylation, IPC-C overaccumulation is observed. It appears that decreasing the accumulation of IPC-C or a related metabolite or altering its structure (to IPC-B, IPC-B, or IPC-A?) reverses the Ca²⁺ sensitivity. It is hoped that continued analysis of suppressor mutants will identify more genes that function in sphingolipid synthesis and identify the Ca²⁺ target that triggers cell death.

The genetic relationship between suppressors of the csg2 mutant and suppressors of the rvs161 mutant suggest a role for sphingolipid in some Rvs161p-dependent process. Three genes (SUR1, SUR2, and SUR4) (12) that mutate to suppress rvs161 mutants (10) are related to CSG1 and CSG2 or to genes that mutate to suppress the Ca²⁺-sensitive phenotype of csg1 and csg2 mutants. The sur1, sur2, and sur4 mutants have altered phospholipid compositions and abnormal morphologies in stationary phase (10). SUR1, which is allelic to CSG1 (5), is a high copy suppressor of csg2 mutants (5, 32). Both SUR1/CSG1 and CSG2 are required for mannosylation of IPC (5, 19). SUR2 encodes the enzyme that hydroxylates C-4 of dihydroceramide. SUR4/ELO3 and SUR5/ELO2/FEN1 encode fatty acid elongases required for the synthesis of the C26 fatty acids found in ceramide and sphingolipids (16).

The physiological function of Sur2p- and Scs7p-mediated hydroxylation is not known. Growth of S. cerevisiae cells does not depend on hydroxylation of either the C-4 of the LCB or the VLCFA moieties of ceramides and sphingolipids. Cells lacking serine palmitoyltransferase can utilize exogenous dihydrosphingosine or phytosphingosine but not sphingosine, the main long chain base in mammals (25). Since sur2 mutants do not synthesize phytosphingosine, it is not the lack of a C-4 hydroxyl group that precludes sphingosine from substituting as the LCB in ceramide synthesis.

Deletion of SUR2 greatly increases the resistance of cells to the Pseudomonas syringae cyclic lipodepsipeptide syringomycin (33) and to the morpholine fungicide fenpropimorph,² an inhibitor of several enzymes in the ergosterol synthesis pathway. As discussed above, deletion of SUR2 also suppresses the Ca²⁺-sensitive phenotype of csg2 mutants and the pleiotropic effects of rvs161 mutations. Elucidation of the mechanism by which blocking hydroxylation of the LCB C-4 increases the ability of cells to tolerate syringomycin, fenpropimorph, and high Ca²⁺ concentrations after CSG2 deletion, and RVS161 deletion may provide clues as to how C-4 hydroxylation affects the functional properties of the LCB, ceramide, and sphingolipids. In addition, the sur2 mutant can be used to identify genetically related genes that encode proteins whose functional properties are effected by C-4 hydroxylation.