Regulation of the Sphingoid Long-chain Base Kinase Lcb4p by Ergosterol and Heme

Sphingoid long-chain base 1-phosphates (LCBPs) are widely conserved, bioactive lipid molecules. In yeast, LCBPs are known to be involved in several cellular responses such as heat shock resistance and Ca2+ mobilization, although their target molecules and signaling pathways remain unclear. To identify genes involved in LCBP signaling and in regulation of LCBP synthesis, we performed transposon mutagenesis in cells lacking the LCBP lyase DPL1 and LCBP phosphatase LCB3 genes and screened for phytosphingosine-resistant clones. Further isolation and identification revealed eight genes (PBP1, HEM14, UFD4, HMG1, TPS1, KES1, WHI2, and ERG5), in addition to the previously characterized LCB4 and PDR5 genes, that are involved in phytosphingosine resistance. Of these eight, four are ergosterol-related genes (HEM14, HMG1, KES1, and ERG5). We also demonstrated that protein expression of the long-chain base kinase Lcb4p was reduced in Δhem14 and Δhmg1 cells, likely as a consequence of decreased activity of the heme-dependent transcription factor Hap1p. In addition, phosphorylation of Lcb4p was decreased in all the ergosterol-related mutants isolated and other ergosterol mutants constructed (Δerg2, Δerg3, and Δerg6). Finally, plasma membrane localization of Lcb4p was found to be reduced in Δerg6 cells. These results suggest that changes in sterol composition affect the phosphorylation of Lcb4p because of the altered localization. The other genes isolated (PBP1, UFD4, TPS1, and WHI2) may be involved in LCBP signaling.

Sphingolipids are major membrane components of eukaryotic cells. Ceramide, the backbone of sphingolipids, is formed by the amide linkage of a sphingoid long-chain base (LCB) 2 with a fatty acid. The major mammalian LCB is sphingosine, whereas in the yeast Saccharomyces cerevisiae, dihydrosphingosine (DHS) and phytosphingosine (PHS) serve as LCBs. LCBs are bioactive lipid molecules involved in apoptosis, endocytosis, and G 1 cell cycle arrest (1)(2)(3). The phosphorylation of LCBs at C-1 results in the production of long-chain base 1-phosphates (LCBPs), sphingosine 1-phosphate (S1P) in mammals and PHS 1-phosphate (PHS1P) and DHS 1-phosphate in yeast. Interestingly, LCBPs possess completely different biological functions compared with the original LCBs.
In mammalian cells, S1P is known to elicit various cellular responses via the cell-surface SIP/Edg (endothelial differentiation gene) receptors, such as proliferation, motility, differentiation, and immunity (4 -7). In addition to its role as an extracellular signal mediator, S1P also functions intracellularly. Intracellular S1P has been proposed to be involved in Ca 2ϩ mobilization, cell proliferation, G 1 /S cell cycle transition, and inhibition of apoptosis (4 -7), although its intracellular target molecules and signaling pathways remain unclear. Because yeast cells do not possess a cell-surface receptor for LCBP/ S1P, their LCBPs function only intracellularly (8 -11). Yeast LCBPs and mammalian intracellular S1P share some effects. For instance, both influence Ca 2ϩ mobilization. Similarly, LCBPs confer protection from environmental stresses, as in the inhibition of apoptosis in mammalian cells and heat resistance in yeast. Such shared functions imply evolutionarily conserved target molecules and signaling pathways for LCBPs.
The enzymes responsible for both the production and degradation of LCBPs are, in fact, completely conserved from yeast to mammals. In mammalian cells, two known sphingosine kinases, SPHK1 and SPHK2, produce S1P; in yeast cells, two LCB kinases, Lcb4p and Lcb5p, which share significant homology with SPHK1 and SPHK2, synthesize LCBPs. In turn, LCBPs are degraded either back to LCBs by the mammalian SPP1 and SPP2 or yeast Lcb3p and Ysr3p phosphatases or to fatty aldehyde and phosphoethanolamine by mammalian SPL or yeast Dpl1p lyase.
Because LCBPs are signaling molecules, their intracellular levels must be strictly regulated. However, knowledge about the regulatory mechanism(s) of their synthesis and degradation is still limited. The activity and/or localization of mammalian SPHK1 is known to be regulated by its association with other proteins (12)(13)(14)(15) or by its phosphorylation by ERK1/2 (extracellular signal-regulated kinase-1/2) (16). Likewise, we recently found that the stability of yeast Lcb4p is regulated through its phosphorylation by the cyclin-dependent protein kinase Pho85p via a ubiquitin-dependent pathway (17).
In this study, we investigated the LCBP signaling pathways as well as the regulatory mechanism(s) of LCBP synthesis by screening transposon-inserted mutants of a PHS-resistant yeast cell. Identification of the sites of transposon insertion revealed that, in addition to the previously described genes LCB4 and PDR5, eight other genes are involved in PHS resistance. In mutants of four ergosterol-related genes (HEM14, HMG1, KES1, and ERG5), the amount and/or phosphorylation of Lcb4p was reduced. Additional analyses suggested that dysfunction of heme synthesis and altered composition of ergosterol are related to the decreased amount and phosphorylation of Lcb4p, respectively.
Isolation of PHS-resistant Yeast Mutants-Mutagenesis by random insertion of the transposon mTn-lacZ/LEU2 was performed as described previously (19) using a yeast genomic library kindly provided by Dr. Michael Snyder (Yale University, New Haven, CT). The library was digested with NotI, and the resulting DNA fragments were transformed into the KHY389 (⌬dpl1 ⌬lcb3) cells. Pooled transposon-inserted mutants were plated at a density of 1 ϫ 10 5 cells/plate on YPD medium containing 7.5 M PHS and 0.015% Nonidet P-40 as a dispersant. The plates were then incubated at 30°C for 3 days. PHS-resistant yeast mutants were obtained at a frequency of ϳ1/600 mutants. We randomly selected 36 PHS-resistant mutants and subjected these to additional analyses. The sites of transposon insertion in the isolated mutants were identified according to the recommendations of the Yale Genome Analysis Center (available at ygac.med.yale.edu/).
Biochemical and Immunochemical Studies-Cellular uptake of [4, H]DHS (American Radiolabeled Chemicals, St. Louis, MO) was assessed by TLC as described previously (19). Radioactivity associated with DHS was quantified using NIH Image Version 1.62 software.
Quantification of LCBs and LCBPs-Lipid extraction and LCB/LCBP separation were performed by a method previously applied to sphingosine/S1P (21). We confirmed that the unlabeled LCBs and LCBPs were effectively extracted and separated by this method. Cells (1 A 600 nm unit) grown in YPD medium at 30°C were suspended in 100 l of water, and lipids were extracted by adding 600 l of 1:2 (v/v) chloroform/methanol, 16 l of 7 N NH 4 OH, 400 l of chloroform, and 400 l of 1 M KCl. After vigorous mixing, the phases were separated by centrifugation, and both the organic phase containing LCBs and the aqueous phase containing LCBPs were collected. To quantify the amount of LCBs, the dried organic phase was first suspended in 2 l of dimethyl sulfoxide, and then 187 l of buffer containing 20 mM Tris-HCl (pH 8.0) and 0.25% Triton X-100 was added. The samples were subjected to the LCB kinase assay described above using purified Lcb4p (10 ng) as the enzyme.
For quantification, LCBPs were first converted to LCBs by alkaline phosphatase and processed as described above. The aqueous phase was mixed with 95 l of buffer containing 500 mM Tris-HCl (pH 8.0) and 10 mM MgCl 2 and with 10 units of calf intestine alkaline phosphatase (Roche Diagnostics). After a 1-h incubation at 37°C, the reaction was terminated by adding 32 l of HCl and 600 l of chloroform. After vigorous mixing, the phases were separated by centrifugation, and the organic phase was collected and dried.

Isolation of PHS-resistant Yeast Mutants-Exogenous
PHS is known to cause growth arrest in yeast (22). Because PHS induces down-regulation of amino-acid permeases, auxotrophic cells are more sensitive to PHS compared with autotrophic cells (22). SEY6210 cells, which carry the leu2 trp1 mutation and so are tryptophan and leucine auxotrophs, were found to be sensitive to 12.5 M PHS. In contrast, SEY6210 cells harboring the pRS314 (TRP1 ϩ ) and pRS315 (LEU2 ϩ ) plasmids were resistant to as much as 15 M PHS (TABLE TWO).
Exogenous PHS is readily imported into yeast cells and converted to PHS1P by Lcb4p. In turn, PHS1P is degraded by Dpl1p to a long-chain fatty aldehyde and phosphoethanolamine or by Lcb3p to PHS. Because overaccumulated PHS1P has toxic effects (23)(24)(25), the toxicity of exogenously added PHS is determined not only by the PHS itself, but also by PHS1P. In growth studies, the ⌬dpl1 cells (KHY13/pRS315) were more sensitive to PHS compared with the wild-type cells (SEY6210/pRS314/ pRS315), whereas the ⌬dpl1 ⌬lcb4 double mutant cells (KHY22) were highly resistant (TABLE TWO). These results indicate that the effect of PHS1P is more prominent than that of PHS in cells autotrophic for tryptophan and leucine.
A ⌬lcb3 or ⌬dpl1 single mutation does not affect cell growth, but a double deletion mutation is lethal or impairs growth severely depending on the yeast background (9,24,25). In the yeast background used here (SEY6210), the ⌬dpl1 ⌬lcb3 cells (KHY388) were able to grow, but did so poorly (TABLE TWO). Moreover, when grown on YPD plates, their colony sizes were highly heterogeneous. When KHY388 (⌬dpl1 ⌬lcb3) cells were visualized by 4Ј,6-diamidino-2-phenylindole staining under an immunofluorescence microscope, most cells contained nuclei with aberrant morphology (Fig. 1, right panels). Some had fragmented or multiple nuclei, and even cells with no nucleus were observed. This morphology was quite different from that observed in the wild-type cells, which each possessed a single round nucleus ( Fig. 1, left panels). The heterogeneous growth of the cells might be the result of differences in the extent of the chromosomal integrity.
We isolated KHY389 cells from KHY388 (⌬dpl1 ⌬lcb3) cells as clones exhibiting homologous growth. Although KHY389 cells still grew more slowly than the wild-type cells (TABLE TWO), the morphology of their nuclei was indistinguishable from that of the wild-type cells (data not shown). The introduction of plasmids encoding LCB3 or DPL1 into KHY389 cells only slightly suppressed their slow growth (data not shown). These results suggest that KHY389 cells carry a suppressor mutation that is a disadvantage for growth but that confers tolerance to overaccumulated LCBPs. However, KHY389 cells still retained high sensitivity to exogenous PHS because they could not grow on YPD plates containing 7.5 M PHS (TABLE TWO). These results indicate that KHY389 cells are less responsive to LCBPs compared with the original KHY388 (⌬dpl1 ⌬lcb3) cells, but are still more sensitive compared with the wild-type or ⌬dpl1 cells.
a Cells exhibited heterogeneous growth.
In our previous study (19), we introduced transposon insertion mutations into KHY13 (⌬dpl1) cells and screened for mutants that were resistant to 15 M PHS. In addition to the expected LCB4 gene, we identified PDR5 and DHH1 as genes conferring PHS resistance (19). To discover additional genes, we repeated the screening using transposoninserted KHY389 cells and a lower concentration of PHS (7.5 M). We expected low concentrations of PHS to provide an advantage in isolating genes involved in LCBP signaling and in regulation of LCBP synthesis rather than those involved in LCB signaling. We obtained transposoninserted mutants that exhibited PHS resistance at a frequency of ϳ1/600 mutants. Of these, 36 mutants were randomly selected, and their sites of transposon insertion were determined. We identified eight genes (PBP1, HEM14, UFD4, HMG1, TPS1, KES1, WHI2, and ERG5) that were involved in PHS resistance, in addition to the previously obtained LCB4 and PDR5 genes (TABLE THREE). Interestingly, of these eight genes, four (HMG1, ERG5, HEM14, and KES1) are known to be involved in ergosterol synthesis or distribution, either directly or indirectly.
PHS Sensitivity in the Mutants of Identified Genes-We next investigated the PHS sensitivity of each isolated yeast mutant. Because all the transposon-inserted mutants contain the LEU2 ϩ marker within the transposon, we created control KHY406 cells by introducing the LEU2 ϩ gene into KHY389 cells, thereby providing uniform auxotrophic conditions. As shown in TABLE FOUR, all the isolated PHS-resistant yeast mutants were able to grow in the presence of 7.5 M PHS, in contrast to control KHY406 cells, although the degree of PHS sensitivity varied. The lcb4 mutants exhibited the strongest PHS resistance, although the hmg1, pdr5, and ufd4 mutants were also quite resistant. In contrast, the whi2 and hem14 mutants exhibited only modest PHS resistance. The growth of the hem14 mutants was slower than that of control KHY406 cells in the absence of PHS and was inhibited slightly further by 7.5 M PHS. The PHS sensitivity of the tps1 mutants differed completely from that of the other mutants. Their growth in the absence of PHS was extremely slow, yet PHS did not inhibit growth further, but rather promoted it, so that in the presence of 10 M PHS, the tps1 mutants grew as slowly as the whi2 mutants. The pbp1, kes1, and erg5 mutants displayed the weakest resistance to PHS. Additionally, the PHS sensitivities of all these transposon-inserted mutants were identical to those of their respective deletion mutants, which we constructed using KHY406 (⌬dpl1 ⌬lcb3 sup) cells as the parental strain (data not shown). These results confirm that the genes identified here by transposon mutagenesis are indeed responsible for PHS resistance.
The parental strain used for transposon mutagenesis (KHY389) may carry an unidentified suppressor mutation as described above. To exclude any indirect effects of such a suppressor, we introduced the deletion mutation of each PHS-resistant gene into the suppressor-less ⌬dpl1 single mutation-containing KHY360 cells and examined their PHS sensitivities. The growth of KHY360 (⌬dpl1) cells was inhibited by PHS in a dose-dependent manner, so that at 12.5 M PHS, the cells grew only weakly (TABLE FIVE). Most of the mutations introduced here conferred apparent PHS resistance, and the order of the strength of this resistance was similar to that observed in the transposon-inserted mutants presented in TABLE FOUR. However, the PHS resistance of the ⌬hem14 and ⌬erg5 cells was rather unclear (TABLE FIVE). The growth rate of the ⌬hem14 cells was similar to that of control KHY360 cells at 10 M PHS and even weaker at 12.5 M PHS. However, considering that the growth of the ⌬hem14 cells was slow in the absence of PHS and that 10 M PHS did not inhibit this already slow growth further, we concluded that the ⌬hem14 mutation did confer PHS resistance. On the other hand, the ⌬erg5 cells exhibited PHS sensitivity similar to that of control KHY360 cells at all concentrations of PHS tested. The erg5 mutation also exhibited the weakest PHS resistance in the KHY389 (⌬dpl1 ⌬lcb3 sup) background (TABLE FOUR), suggesting that the effect of this mutation was not detectable in the KHY360 (⌬dpl1) background.
Accumulation of LCBs in PHS-resistant Yeast Mutants-Reduced accumulation of exogenous PHS caused by a decrease in its uptake or an

PHS sensitivity in PHS-resistant yeast mutants
The isolated transposon-inserted mutants and the control cells (⌬ dpl1 ⌬ lcb3 sup) were cultured at 30°C for 2 days on YPD plates containing 7.5 or 10 M PHS as indicated. The PHS sensitivity was evaluated in seven grades (Ϫ, ϩ/Ϫ, ϩ, ϩϩ, ϩϩϩ, ϩϩϩϩ, ϩϩϩϩϩ) according to colony size.  (Fig. 2). This was similar to our previous results for pdr5 deletion mutants, which exhibit an enhanced efflux of LCBs attributable to an increase in the LCB-specific translocase/transporter Rsb1p (19). In contrast, the accumulation of [ 3 H]DHS was markedly increased in the hem14 and hmg1 mutants. Other mutations did not affect the accumulation of [ 3 H]DHS significantly, although slight increases were observed in the pbp1, kes1, and erg5 mutants. These results suggest that the PHS resistance of the mutants, other than the pdr5 mutants, is not caused by a reduction in [ 3 H]DHS accumulation. Reduced Levels of Lcb4p in the Ergosterol-related Mutants-Decreased production of PHS1P from imported PHS is another possible mechanism for acquiring PHS resistance. Therefore, we next examined whether any of the PHS-resistant mutations affected the protein levels of the major LCB kinase Lcb4p. Cell lysates were prepared from each deletion mutant in the KHY360 (⌬dpl1) background and subjected to immunoblotting with anti-Lcb4p antibodies. Two protein bands corresponding to phosphorylated and non-phosphorylated forms of Lcb4p (17) were detected in the control cells, and the upper, phosphorylated band was predominant (ϳ70% of total Lcb4p) (Fig. 3, A and B). The ⌬pbp1, ⌬ufd4, ⌬tps1, ⌬pdr5, and ⌬whi2 cells had equivalent levels of Lcb4p with similar phosphorylation profiles. On the other hand, the amount and/or phosphorylation of Lcb4p was reduced in the ergosterol-related mutants, i.e. the ⌬hem14, ⌬hmg1, ⌬kes1, and ⌬erg5 cells (Fig.  3, A and B). In the ⌬hem14 cells, the amount of Lcb4p was greatly reduced to ϳ20% of the level found in the control cells, and the phosphorylated form of Lcb4p was scarcely detected. In the ⌬hmg1 cells, the total amount of Lcb4p (phosphorylated and non-phosphorylated) was slightly decreased to ϳ80% of the total in the control cells, with the phosphorylated form declining to ϳ50%. In the ⌬kes1 and ⌬erg5 cells, the total amounts of Lcb4p were similar to the levels observed in the control cells; however, a decrease in the phosphorylated form was observed (Fig. 3, A and B). In particular, the non-phosphorylated form was more prevalent than the phosphorylated form in the ⌬kes1 cells, with a slight but reproducible increase apparent in the non-phosphorylated form in the ⌬erg5 cells (Fig. 3, A and B; see also Fig. 4B).

Strain
We also examined the in vitro LCB kinase activity of the ergosterolrelated mutants. As shown in Fig. 3C, the LCB kinase activities in the ⌬hem14 and ⌬hmg1 cell lysates were reduced to ϳ30 and 85%, respectively, of the activity observed in the control cells. In contrast, the LCB kinase activities in the ⌬kes1 and ⌬erg5 cells were similar to the activity in the control cells. These results suggest that the PHS resistance of the ergosterol-related mutants is caused by changes in the expression and/or phosphorylation of Lcb4p.
Effect of Sterol Composition on the Phosphorylation of Lcb4p-Because mutations in the ergosterol-related genes (HEM14, HMG1, KES1, and ERG5) resulted in a reduction in the phosphorylation of Lcb4p, we considered that the sterol composition might influence the phosphorylation of this protein. Genes involved in the early stages of ergosterol synthesis are essential for cell growth; however, genes involved in the later stages of synthesis (ERG6, ERG2, ERG3, ERG5, and ERG4) (Fig. 4A) can be deleted without interrupting cell growth, and disruption of any is known to cause changes in the sterol composition (26,27). Using the ⌬erg6, ⌬erg2, ⌬erg3, and ⌬erg5 cells in the KHY360 (⌬dpl1) background, we investigated the roles of ergosterol in the phosphorylation of Lcb4p by immunoblotting. The phosphorylation of Lcb4p was reduced in all the ergosterol mutants, although the degree of reduction varied (Fig. 4B). The ⌬erg6 cells exhibited the greatest reduction in phosphorylation, and the ⌬erg3 and ⌬erg2 mutations also significantly affected the phosphorylation of Lcb4p. Although a single deletion of the ERG2, ERG3, or ERG5 gene had a moderate or weak effect on the phosphorylation of Lcb4p, double deletions induced striking changes in phosphorylation (Fig. 4B). These results indicate that changes in the sterol composition can indeed cause a reduction in the phosphorylation of Lcb4p. On the other hand, the levels of Lcb4p appear not to be affected.
Next, we measured the intracellular amounts of LCBs and LCBPs in the deletion mutants of the PHS-resistant genes and the ergosterolsynthesizing genes, all in the KHY360 (⌬dpl1) background (Fig. 5). As  expected, in the ⌬lcb4 cells, LCBPs were hardly detected, and concomitantly, their substrate LCBs were greatly increased. Both LCBs and LCBPs were reduced in the ⌬tps1 mutants, possibly an indirect result of the slow growth. Consistent with the low levels of Lcb4p, the amounts of LCBPs were significantly reduced in the ⌬hem14 cells. Unexpectedly, both LCBs and LCBPs were increased in the ergosterol-synthesizing mutants ⌬hmg1, ⌬erg2, ⌬erg3, ⌬erg6, ⌬erg2 ⌬erg5, and ⌬erg3 ⌬erg5. On the other hand, the LCB/LCBP levels were only slightly affected or unchanged in the other mutants. Involvement of the Heme-dependent Transcription Factor Hap1p in the Expression of Lcb4p-Although changes in the sterol composition seemed not to affect the Lcb4p levels (Fig. 4B), the protein levels were greatly and slightly reduced in the ⌬hem14 and ⌬hmg1 cells, respectively (Fig. 3A). To investigate whether the reduction was caused by reduced synthesis or increased degradation of the protein, we performed a pulse-chase experiment in the ⌬hem14 cells. As shown in Fig.  6 (A and B), the synthesis of Lcb4p was lower in the ⌬hem14 cells than in the wild-type cells by ϳ2-fold. Degradation of Lcb4p in the ⌬hem14 cells was not accelerated, but instead was delayed, indicating that the reduction in Lcb4p was caused solely by decreased synthesis.
Hem14p is protoporphyrinogen oxidase that catalyzes the late step of heme synthesis (28,29). Hmg1p is involved in the synthesis of farnesyl diphosphate (30), which is also required for synthesis of certain hemes (hemes O and A) (31). Some genes involved in ergosterol biosynthesis, including HMG1 and ERG11, are regulated by heme at the transcriptional level (32,33), and the transcription factor Hap1p is known to be involved in this regulation (34).
We investigated whether Hap1p might be involved in the transcriptional regulation of Lcb4p by examining Lcb4p expression in ⌬hap1 cells. Lcb4p levels were reduced in the ⌬hap1 cells to about half those found in the wild-type cells (Fig. 6C), not as drastic a reduction as in the ⌬hem14 cells. Furthermore, the amount of phosphorylated Lcb4p was also reduced in the ⌬hap1 cells. This reduction may be an indirect result of reduced transcription of the ergosterol-synthesizing genes HMG1 and ERG11. These data suggest that defects in heme synthesis in the ⌬hem14 and ⌬hmg1 cells affect the activity of Hap1p, resulting in a reduction in the expression of Lcb4p.
Effect of the erg Mutations on the Localization of Lcb4p-The cause of the reduced phosphorylation of Lcb4p in the erg mutants was not clear. One possible mechanism would be the altered cellular localization of Lcb4p in these mutants. Thus, we performed a sucrose gradient fractionation. In the wild-type cells, Lcb4p exhibited two peaks: a major peak in fractions 3-5 and a minor peak in fraction 10 (Fig. 7A). The major peak may represent endoplasmic reticulum (ER)-resident Lcb4p because the ER membrane protein Dpm1p was also highest in these fractions. The minor peak coincided with the plasma membrane marker Pma1p. These results suggest that Lcb4p is localized mainly in the ER, but that a small amount of Lcb4p also resides in the plasma membrane. We also examined fractions from the ⌬erg6 cells, which exhibited a great reduction in the phosphorylation of Lcb4p (Fig. 4B). In these cells, plasma membrane-localized Lcb4p was hardly detected, and most of Lcb4p was localized in the ER fractions (Fig. 7B). These results indicate that altered ergosterol composition prevents Lcb4p from localizing in the plasma membrane.

DISCUSSION
In this study, we screened for PHS-resistant genes in yeast using transposon mutagenesis. In addition to the previously identified LCB4 and PDR5 genes, we identified eight other genes (PBP1, HEM14, UFD4, HMG1, TPS1, KES1, WHI2, and ERG5) that appear to be involved in PHS resistance (TABLE THREE). Of these eight genes, four (HMG1, ERG5, HEM14, and KES1) were identified as ergosterol-related genes, and two of these (HMG1 and ERG5) are directly involved in ergosterol synthesis. Hmg1p and its homolog Hmg2p redundantly catalyze the FIGURE 3. The amount and/or phosphorylation of Lcb4p is reduced in ⌬hem14, ⌬hmg1, ⌬kes1, and ⌬erg5 cells. A and B, total proteins (10 g) prepared from TS14 (⌬lcb4), KHY479 (⌬pbp1), FKY47 (⌬hem14), KHY481 (⌬ufd4), TS33 (⌬hmg1), TS34 (⌬tps1), KHY364 (⌬pdr5), TS45 (⌬kes1), KHY480 (⌬whi2), FKY48 (⌬erg5), and KHY360 (wild-type) cells were separated by SDS-PAGE and transferred to membranes for immunoblotting. Proteins were detected with anti-Lcb4p antibodies or, to demonstrate uniform protein loading, with anti-Pgk1p antibodies (A). The intensity of phosphorylated (p) and nonphosphorylated Lcb4p was quantified using NIH Image Version 1.62 software (B). Each value is presented as a percentage of the amount of phosphorylated Lcb4p relative to that of total Lcb4p and represents the mean Ϯ S.D. from three independent experiments. Statistically significant differences between values for the wild-type and mutant cells are indicated. *, p Ͻ 0.001. C, cell lysates were prepared from FKY47 (⌬hem14), TS33 (⌬hmg1), TS45 (⌬kes1), FKY48 (⌬erg5), and KHY360 (wild-type) cells. Samples (20 g of protein) were subjected to an in vitro LCB kinase activity assay using a mixture of [␥-32 P]ATP, 1 mM unlabeled ATP, and 50 M sphingosine and incubated at 37°C for 15 min. Lipids were extracted and separated by TLC. Radioactivity associated with the resulting S1P was quantified using a Fuji Photo Film BAS-2500 bioimaging analyzer. Each value is presented as a percentage relative to the LCB kinase activity associated with the control cells and represents the mean Ϯ S.D. from three independent experiments. Statistically significant differences between values for the wild-type and mutant cells are indicated. *, p Ͻ 0.001; **, p Ͻ 0.01. NOVEMBER 4, 2005 • VOLUME 280 • NUMBER 44 rate-limiting step, the conversion of hydroxymethylglutaryl-CoA to mevalonic acid, and ⌬hmg1 cells exhibit largely reduced enzyme activity (35). Erg5p is the enzyme that desaturates C-22 of the side chain in the late step of ergosterol synthesis (36). Hem14p converts protoporphyrinogen IX to protoporphyrin IX in the late step of heme synthesis (28,29), so in ⌬hem14 cells, the function of heme-containing proteins such as the ergosterol synthesis-related enzymes Erg5p and Erg11p (36,37) would be reduced. Moreover, the expression of HMG1 and ERG11, regulated by heme at the transcriptional level (32,33), would be also decreased. Finally, Kes1p, an oxysterol-binding protein homolog, is thought to be involved in ergosterol transport from the Golgi to the plasma membrane, and although total ergosterol levels in ⌬kes1 cells are unchanged, the levels on the cell surface are reduced (38).

Regulation of Lcb4p by Ergosterol and Heme
One possible mechanism responsible for the PHS resistance of these mutants would be a reduction in LCB accumulation; however, DHS accumulation was decreased only in the pdr5 mutants (Fig. 2). We recently demonstrated that the LCB-releasing activity of pdr5 mutants is increased due to up-regulation of the LCB-specific translocase/transporter Rsb1p (19). In contrast, accumulation of [ 3 H]DHS was somewhat increased in the hem14 and hmg1 mutants (Fig. 2). Either mutation (hem14 or hmg1) affects ergosterol synthesis, so increased accumulation would be consistent with a previous finding that the uptake of certain compounds is increased in ergosterol mutants because of enhanced membrane permeability (39).
⌬hem14 and ⌬hmg1 cells also exhibited reduced phosphorylation of Lcb4p (Fig. 3, A and B), yet cells carrying mutations specifically affecting ergosterol synthesis (⌬erg6, ⌬erg2, ⌬erg3, and ⌬erg5) exhibited only reduced phosphorylation (Fig. 4B). We found that Lcb4p expression was partially regulated by the heme-dependent transcription factor Hap1p (Fig. 6C). Therefore, it is likely that the reduction in the Lcb4p levels in the hem14 and hmg1 mutants is caused by an impairment in heme synthesis and not by one in ergosterol synthesis. However, the amount of Lcb4p in the ⌬hem14 cells was still lower than in the ⌬hap1 cells (Fig. 6C). One possible explanation for this is that other transcription factors affected by the ⌬hem14 mutation are also involved in the expression of Lcb4p. Alternatively, the decreased expression of Lcb4p in the ⌬hem14 cells might be due solely to the reduced activity of Hap1p. However, the difference in the amount of Lcb4p between the ⌬hap1 and ⌬hem14 cells might be caused by a difference in their growth rates because the growth rate of the ⌬hem14 cells was significantly slow (TABLES FOUR and FIVE).
In our sucrose density gradient experiment (Fig. 7) and an immunofluorescence microscopic analysis 3 of the wild-type cells, Lcb4p was found to be localized mainly in the ER, with some presence in the plasma membrane, although this pattern differs from that found in other studies using a C-terminal hemagglutinin tag, which may have interfered with localization (40,41). In the mutant screening, PHS was added exogenously, so perhaps plasma membrane-localized Lcb4p is important for conversion to PHS1P of only imported PHS (i.e. that which is localized in the plasma membrane). This would appear to be the reason that the hmg1 and erg5 mutants were isolated by our screening, although they can normally produce LCBPs from endogenous LCBs (Fig. 5), which are localized mainly in the ER. We also speculate that Lcb4p is phosphorylated at the plasma membrane and that, in the erg mutants, reduced localization at the plasma membrane results in decreased phosphorylation. Additionally, we recently reported that Lcb4p is anchored to the membranes through palmitoylation (42). Because palmitoylation often facilitates recruitment into sterol/sphingolipid-rich microdomains (43,44), it is possible that the palmitic acid moiety of Lcb4p interacts with ergosterol, which is abundant in the plasma membrane but not in the ER and that this interaction causes the retention of Lcb4p in the membrane.
Interestingly, LCB/LCBP levels were increased in the erg mutants (Fig. 5). Previous studies have shown that ergosterol levels affect the amounts of certain sphingolipid species (45). Moreover, ⌬arv1 cells, in which intracellular sterol distribution is altered, also harbor defects in sphingolipid synthesis (46). Thus, there seems to be unknown mechanisms that regulate sphingolipid synthesis in response to cellular ergosterol status.
The mutations of the TPS1, WHI2, PBP1, and UFD4 genes affected neither LCB accumulation nor Lcb4p expression/phosphorylation (Figs. 3 and 4). Intracellular LCBPs were also nearly unchanged in these mutants, except for the tps1 mutant, which exhibited reduced LCB/ LCBP levels (Fig. 5), probably because of an indirect growth effect. Nevertheless, it is possible that these genes are involved in LCBP signaling. LCBPs are known to be involved in the heat stress response, and Tps1p and Whi2p indeed also function in this process (47,48). Future studies are required to reveal the link between these genes and LCBP signals.