Regulation of the transport and protein levels of the inositol phosphorylceramide mannosyltransferases Csg1 and Csh1 by the Ca2+ binding protein Csg2

Complex sphingolipids in yeast are known to function in cellular adaptation to environmental changes. One of the yeast complex sphingolipids, mannosylinositol phosphorylceramide (MIPC), is produced by the redundant inositol phosphorylceramide (IPC) mannosyltransferases Csg1 and Csh1. The Ca2+-binding protein Csg2 can form a complex with either Csg1 or Csh1 and is considered to act as a regulatory subunit. However, the role of Csg2 in MIPC synthesis has remained unclear. In this study, we found that Csg1 and Csh1 are N-glycosylated with core-type and mannan-type structures, respectively. Further identification of the glycosylated residues suggests that both Csg1 and Csh1 exhibit membrane topology with their C termini in the cytosol and their mannosyltransferase domains in the lumen. After complexing with Csg2, both Csg1 and Csh1 function in the Golgi, and then are delivered to the vacuole for degradation. However, uncomplexed Csh1 cannot exit from the endoplasmic reticulum. We also demonstrated that Ca2+ stimulates IPC-to-MIPC conversion, because of a Csg2-dependent increase in Csg1 levels. Thus, Csg2 has several regulatory functions for Csg1 and Csh1, including stability, transport, and gene expression.

phosphoinositol. There are five IPCs (IPC-A, -B, -B', -C, and -D), which differ in the position and/or number of hydroxyl groups within the ceramide moiety (5)(6)(7). The other two yeast sphingolipids are mannose-containing GSLs that are not required for normal cell growth, though their loss results in altered sensitivities to several drugs (8)(9)(10)(11). MIPC synthesis is catalyzed by either of two homologous IPC mannosyltransferases, Csg1 and Csh1, which prefer different IPC species as substrates (12,13). Cells carrying a double mutation, Δcsg1 Δcsh1, are sensitive to external Ca 2+ (5,12,14), suggesting a role for the GSLs in cellular stress response. We recently demonstrated that the Ca 2+ -binding protein Csg2 (14,15), interacts with both Csg1 and Csh1 (12). Our Δcsg1 Δcsg2 cells had no MIPC synthase activity, and introduction of a Δcsg2 mutation into the Δcsh1 cells also resulted in a significant reduction in activity (12). Thus, interaction with Csg2 is essential for the activity of Csh1 and is quite important for that of Csg1.
The existence of the Ca 2+ -binding regulatory subunit Csg2 and two different catalytic subunits, Csg1 and Csh1, implies that IPC-to-MIPC conversion is a highly regulated process. However, the exact functions of Csg2 and Ca 2+ in MIPC synthesis remain unclear. In the present study, we report that Csg2 functions in the production, stability, and transport of Csg1 and/or Csh1. In addition, we found that Ca 2+ treatment increases Csg1 levels in a Csg2-dependent manner and enhances MIPC synthesis.

Experimental Procedures
Yeast Strains and Media --Saccharomyces cerevisiae strains used are listed in Table I. The Δcsg2::URA3 cells were constructed by replacing the 0.41 kb EcoRI-HincII region in the CSG2 gene with URA3 marker. The Δpep4::LEU2 cells were constructed as described elsewhere (16).   Immunoblotting --Immunoblotting was performed as described previously (18  µM forward primer, and 0.9 µM reverse primer.
The reactions were incubated in a 96-well plate at 95˚C for 10 min, followed by 40 cycles at 95˚C for 15 s, and 60˚C for 1 min. All reactions were run in triplicate.

In Vivo [ 3 H]Dihydrosphingosine (DHS)
Labeling --[ 3 H]DHS labeling assay was performed as described previously (12). Therefore, we next examined the effects of Csg2 on the glycosylation status of Csg1 and Csh1.

N-glycosylation in
Neither the gel mobility nor protein levels of Csg1-3xFLAG differed in the Δcsg2 cells compared to those in cells bearing the wild-type CSG2 ( Fig. 2A). In contrast, the broad bands of 55-70 kDa representing Csh1-3xFLAG were completely converted in the Δcsg2 mutant to the 49 kDa band ( Fig. 2A), which may have the endoplasmic reticulum (ER) "high mannose-type" glycan as discussed below.
We also investigated the glycosylation status of overproduced Csg1 and Csh1. Both proteins were expressed as C-terminally As shown in Fig. 3B, the gel mobility of Csg1-His 6 -Myc was unchanged in the Δvan1 cells. In contrast, the broad 55-70 kDa bands of Csh1-His 6 -Myc observed in the wild-type cells were diminished in the Δvan1 mutants, but a 51 kDa band appeared, which may carry an incomplete mannan-type structure (Fig. 3C).
These results suggest that Csh1 is N-glycosylated with a 'mannan-type' structure, whereas Csg1 may be modified with 'core-type' glycosylation.

Determination of N-glycosylation sites in
To investigate whether the increases in protein levels were due to enhanced gene expression, we performed real-time PCR analyses. The mRNA levels of CSG1 were increased upon treatment with Ca 2+ , in a dose-dependent manner (Fig. 7B)  oxidative stress, and drug treatment (5, 8-12, 14, 26). Therefore, IPC-to-MIPC conversion may be a key regulatory step for determining cellular sphingolipid composition. Indeed, we found that this step is stimulated in the presence of Ca 2+ (Fig. 8). As a result, IPC levels were reduced, and MIPC/M(IP) 2 C levels were increased. The enhanced IPC-to-MIPC conversion in the presence of Ca 2+ was nicely explained by increases in the mRNA and protein levels of CSG1 but not of CSH1 (Fig. 7 A and B).
Csg2 was found to be involved in the Ca 2+ -dependent Csg1 increase (Fig. 7A). In addition, overproduction of Csg2 also resulted in an increase in Csg1 level (Fig. 2B). Thus, a signaling pathway that transduces signal from Csg2 to the CSG1 gene expression may exist.
Binding of Ca 2+ to Csg2 or Csg2 overproduction might enhance the signal, leading to increases in CSG1 mRNA.
Secreted proteins and those localized in the ER, Golgi, endosomes, plasma membrane, and vacuole are synthesized by the ER-bound ribosome. Only properly folded proteins can exit from the ER, while proteins that are misfolded due to mutation or misassembly into a proper complex are removed by an ER quality control system, the so-called ERAD (27)(28)(29). In the present study, we determined that Csh1 is unable to exit from the ER unless it forms a complex with Csg2 (Figs. 2, 4, and 5B); apparently free Csh1 is removed by ERAD. In contrast, free Csg1 was transported to the vacuole via the Golgi, in a similar fashion to Csg1 complexed with Csg2 ( Figs. 2A and 5A). This suggests that Csg1 can fold into its proper conformation to some extent in the absence of Csg2. Consistent with these results, Csh1 had no activity in the Δcsg2 cells, whereas Csg1 retained activity, albeit weak activity (12). GTPase and Tor pathways (31)(32)(33)(34)(35). These signaling events are mediated by PtdIns(4,5)P 2 -binding proteins (33)(34)(35). A recent study found that the plasma membrane localization of Mss4 is disturbed in Δcsg2 cells (36), suggesting that accumulation of IPC-C or loss of MIPC causes mislocalization of Mss4.
IPC-C contains three hydroxyl groups in the ceramide moiety and, thus, is predicted to interact, via hydrogen bonds, with other IPC-C molecules or with other lipid molecules that contain hydroxyl groups and/or amino groups, and thereby generate lipid microdomains. We speculate that Ca 2+ binds to the phosphate group of IPCs and alters the lipid microdomain, leading to changes in the signaling pathways of Mss4 and Tor2.
Yeast microdomains are known to be composed of ergosterol and sphingolipids (2).
We recently reported that a shift in the sphingoid base of the yeast sphingolipids from phytosphingosine/DHS to sphingosine disrupts the lipid microdomain (37). In that study we also demonstrated that yeast was highly sensitive to Ca 2+ , suggesting a link between Ca 2+ -sensitivity and lipid microdomain formation. However, further work will be required to confirm any  Cells were grown at 30˚C in SC medium lacking histidine and leucine. The fractions of integral membrane proteins were prepared and subjected to immunoprecipitation using anti-Myc antibodies.