Selective Substrate Supply in the Regulation of Yeast de Novo Sphingolipid Synthesis*

The heat stress response of Saccharomyces cerevisiae is characterized by transient cell cycle arrest, altered gene expression, degradation of nutrient permeases, trehalose accumulation, and translation initiation of heat shock proteins. Importantly heat stress also induces de novo sphingolipid synthesis upon which many of these subprograms of the heat stress response depend. Despite extensive data addressing the roles for sphingolipids in heat stress, the mechanism(s) by which heat induces sphingolipid synthesis remains unknown. This study was undertaken to determine the events and/or factors required for heat stress-induced sphingolipid synthesis. Data presented indicate that heat does not directly alter the in vitro activity of serine palmitoyltransferase (SPT), the enzyme responsible for initiating de novo sphingolipid synthesis. Moreover deletion of the small peptide Tsc3p, which is thought to maximize SPT activity, specifically reduced production of C20 sphingolipid species by over 70% but did not significantly decrease overall sphingoid base production. In contrast, the fatty-acid synthase inhibitor cerulenin nearly completely blocked sphingoid base production after heat, indicating a requirement for endogenous fatty acids for heat-mediated sphingoid base synthesis. Consistent with this, genetic studies show that fatty acid import does not contribute to heat-induced de novo synthesis under normal conditions. Interestingly the absence of medium serine also ameliorated heat-induced sphingoid base production, indicating a requirement for exogenous serine for the response, and consistent with this finding, disruption of synthesis of endogenous serine did not affect heat-induced sphingolipid synthesis. Serine uptake assays indicated that heat increased serine uptake from medium by 100% during the first 10 min of heat stress. Moreover treatments that increase serine uptake in the absence of heat including acute medium acidification and glucose treatment also enhanced de novo sphingoid base synthesis equivalent to that induced by heat stress. These data agree with findings from mammalian systems that availability of substrates is a key determinant of flux through sphingolipid synthesis. Moreover data presented here indicate that SPT activity can be driven by several factors that increase serine uptake in the absence of heat. These findings may provide insights into the many systems in which de novo synthesis is increased in the absence of elevated in vitro SPT activity.

The heat stress response of Saccharomyces cerevisiae is characterized by transient cell cycle arrest, altered gene expression, degradation of nutrient permeases, trehalose accumulation, and translation initiation of heat shock proteins. Importantly heat stress also induces de novo sphingolipid synthesis upon which many of these subprograms of the heat stress response depend. Despite extensive data addressing the roles for sphingolipids in heat stress, the mechanism(s) by which heat induces sphingolipid synthesis remains unknown. This study was undertaken to determine the events and/or factors required for heat stress-induced sphingolipid synthesis. Data presented indicate that heat does not directly alter the in vitro activity of serine palmitoyltransferase (SPT), the enzyme responsible for initiating de novo sphingolipid synthesis. Moreover deletion of the small peptide Tsc3p, which is thought to maximize SPT activity, specifically reduced production of C 20 sphingolipid species by over 70% but did not significantly decrease overall sphingoid base production. In contrast, the fatty-acid synthase inhibitor cerulenin nearly completely blocked sphingoid base production after heat, indicating a requirement for endogenous fatty acids for heat-mediated sphingoid base synthesis. Consistent with this, genetic studies show that fatty acid import does not contribute to heat-induced de novo synthesis under normal conditions. Interestingly the absence of medium serine also ameliorated heat-induced sphingoid base production, indicating a requirement for exogenous serine for the response, and consistent with this finding, disruption of synthesis of endogenous serine did not affect heat-induced sphingolipid synthesis. Serine uptake assays indicated that heat increased serine uptake from medium by 100% during the first 10 min of heat stress. Moreover treatments that increase serine uptake in the absence of heat including acute medium acidification and glucose treatment also enhanced de novo sphingoid base synthesis equivalent to that induced by heat stress. These data agree with findings from mammalian systems that availability of substrates is a key determinant of flux through sphingolipid synthesis. Moreover data presented here indicate that SPT activity can be driven by several factors that increase serine uptake in the absence of heat. These findings may provide insights into the many systems in which de novo synthesis is increased in the absence of elevated in vitro SPT activity.
The contribution of sphingolipids to yeast biology has become increasingly appreciated because of recent findings that they regulate processes as diverse as endocytosis, cytoskeletal dynamics, ubiquitin-dependent proteolysis, the cell cycle, sporulation, translation, and post-translational protein assembly, modification, and routing (for reviews, see Refs. [1][2][3]. In particular, the role of sphingolipids in stress responses has generated significant interest due to the discovery that heat stress induces transient flux through de novo sphingolipid synthesis (4 -6), and importantly, the resulting sphingolipids regulate all major aspects of the yeast heat stress response, including protein degradation (7), translation of heat shock proteins (8), trehalose synthesis (5), cell cycle arrest (9), heat-induced changes in mRNA levels (10), and ultimately survival during heat stress (9). Although much research has addressed functions of sphingolipids in yeast, the mechanisms of regulation of sphingolipid metabolism during heat stress remains largely undetermined.
De novo sphingolipid metabolism commences with the condensation of serine with palmitoyl-CoA by serine palmitoyltransferase (SPT), 2 the rate-limiting step in de novo sphingolipid synthesis (11). This reaction generates 3-ketodihydrosphingosine, a short lived intermediate that is quickly converted to dihydrosphingosine. The Syr2 hydroxylase then catalyzes the conversion of dihydrosphingosine to phytosphingosine (12,13), and both dihydro-and phytosphingosine can undergo either phosphorylation (forming sphingoid base phosphates) or N-acylation (forming ceramides) (for a review, see Ref. 2). Shifting yeast cultures from normal growth temperature (i.e. 24 -30°C) to elevated temperature (i.e. 37-42°C) increases dihydro-and phytosphingosine by 2-10-fold over the first 5-15 min (5,6) and subsequently the downstream products of de novo synthesis, including sphingoid base phosphates and ceramides (4 -6, 14). These time frames are too short to involve transcriptional or translational regulation of the enzymes responsible, and furthermore numerous microarray studies have failed to find significant regulation of the genes encoding the subunits of SPT, LCB1 or LCB2, during heat stress (10,15,16). Furthermore the mechanism(s) of acute regulation of sphingolipid production has generated intense interest as numerous mammalian cellular processes including chemotherapy-mediated cell death require acute de novo sphingolipid synthesis analogous to that observed during the yeast heat stress response (17)(18)(19). * This work was supported in part by National Institutes of Health Grant GM63265 (to Y. A. H.) and a Veterans Affairs Merit Review Entry Program award (to L. A. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom correspondence should be addressed. E-mail: hannun@musc.edu.
Several previous studies in mammalian cells demonstrated dependence of de novo sphingolipid synthesis on substrate concentration. Specifically in mouse LM cells, intracellular serine concentration was found to be directly proportional to that of extracellular serine, and radiolabeled serine was incorporated into sphingoid bases at rates consistent with in vitro kinetic parameters for SPT, suggesting that SPT activity was limiting for extracellular serine-induced sphingolipid synthesis (20). In a subsequent study, it was determined that an acute increase in extracellular serine (accomplished by medium change) caused a "burst" of de novo sphingolipid synthesis (21). Importantly, however, regulatory factors for serine import remain undefined in these systems.
This study aimed to determine the mechanism(s) by which heat stress induces de novo sphingolipid synthesis in yeast. In brief, the data indicate that during heat stress, the sphingolipid fatty acid component is derived from fatty-acid synthase (FAS), whereas the serine is derived from exogenous sources. Moreover heat stimulated serine uptake from the medium, and importantly, other external stimuli that increase amino acid uptake were determined to also increase de novo sphingolipid synthesis in the absence of heat. These findings indicate that increased amino acid import is sufficient to promote a flux through de novo sphingolipid synthesis pathways. Furthermore these data suggest that heat-stimulated serine uptake may be a primary signal initiating the yeast heat stress response.

MATERIALS AND METHODS
Yeast Strains and Culture Conditions-The yeast strains used in this study are shown in Table 1  Culture and Heat Stress-For routine culturing, cells were seeded from overnight 5-ml cultures in yeast proteose dextrose medium (YPD) consisting of 1% yeast extract, 2% proteose peptone, and 2% dextrose and grown in a water bath at 30°C with shaking at 200 -250 rpm. For heat stress experiments, unless otherwise indicated, cultures were aliquoted into 50-ml Falcon tubes and either returned to the 30°C water bath or transferred to an identical water bath set at 39°C with shaking at 200 -250 rpm. For some experiments, cells were aliquoted into Falcon tubes, harvested, washed in water, and then resuspended in either fresh YPD, synthetic complete medium (SC), synthetic serine-free medium (SC Ser Ϫ ), or a non-buffered 10 mM serine, 20 mM KCl solution with or without 50 mM glucose. SC consisted of yeast nitrogen base, 2% dextrose, and complete supplemental mixture containing amino acids and uracil. SC Ser Ϫ was prepared similarly, but rather than using the complete supplemental mixture, amino acids except serine and uracil were added separately.
In Vitro SPT Measurements-In vitro SPT activity was measured in yeast microsomes prepared by the glass bead procedure. Equal amounts of protein were used to assay in vitro the production of sphingoid bases by the microsomal fraction according to a yeast protocol described previously (22). Scintillation detection of [ 3 H]serine incorporated into Bligh-Dyer extracts (23) was used to calculate enzyme activity as described previously (22).
Mutagenesis-The tsc3⌬ mutant was generated by short flanking homology using primers with 5Ј regions complimentary to non-coding portions of the TSC3 gene and 3Ј regions complimentary to the G418/neomycin resistance gene. This resulted in a recombinant DNA fragment containing the G418/ neomycin resistance cassette framed by 5Ј and 3Ј non-coding genomic sequences for TSC3. The ser3⌬/ser33⌬ double mutant was constructed in a similar fashion. The ser3⌬ deletion strain in the BY4742 background was obtained from Invitrogen. This library was constructed by systematic replacement of coding regions with the G418 resistance gene. Therefore, to delete SER33 in this mutant, short flanking homology recombination was used as above with primers having 5Ј homology and 3Ј homology to non-coding areas of SER33 and the internal sequence of the URA3 gene. In each case, appropriate recombination was confirmed by PCR using a primer from upstream of the 5Ј flanking region of the recombinant construct with a downstream primer complimentary to the G418 cassette or the URA3 cassette for tsc3⌬ and ser3⌬/ser33⌬, respectively, for PCR of genomic DNA of putative recombinant clones. The ser3⌬/ser33⌬ double mutant was also confirmed by its inability to grow in the absence of exogenously added serine (not shown).
Sphingoid Base Measurement-After heat treatments as described, cells were collected by centrifugation at 3000 ϫ g at room temperature for 2-3 min followed by suspension in 3 ml of 2:1 chloroform:methanol. Lipids were extracted by the method of Bligh and Dyer (23). One-third of the sample was reserved for total lipid phosphate determination as described previously (24). The remaining two-thirds were subjected to mild alkaline hydrolysis. Sphingoid bases were resolved by HPLC after conversion to fluorescent derivatives with orthophthalaldehyde as described previously (25). Measurement of fluorescent bases occurred by in-line fluorescence detection.
[ 3 H]Serine Uptake Assay-[ 3 H]Serine uptake was measured according to a method described previously (26). In brief, cells were grown to midlog phase, washed in assay buffer consisting of 50 mM sodium citrate, pH 5.5, containing 2% D-glucose, and suspended in assay buffer at a final density of 1-4 ϫ 10 7 cells/ ml. After 1-2 min of temperature equilibration in a shaking water bath set at 30 or 39°C, [ 3 H]serine was added to a final concentration of 8.7 M containing 23 dpm/pmol specific activity. Aliquots of 100 l were taken at indicated times and immediately pipetted into 1 ml of ice-cold assay buffer on ice in Eppendorf tubes. At the conclusion of the time course, samples were harvested by benchtop microcentrifugation for 30 s at 13,000 rpm. Pellets were washed three times in ice-cold assay buffer, resuspended in 100 l of buffer, and quantified by liquid scintillation.
Intracellular Free Serine Determinations-Cells were grown in YPD to midlog phase, harvested by centrifugation, and lysed using glass beads with 5 min of continuous agitation in a Bead Beater apparatus at 4°C. Beads were removed by centrifugation, and lysates were sent to Scientific Research Consortium, Inc. (St. Paul, MN) for analysis of serine concentration. For all experiments, statistics were performed using Microsoft Excel, and graphs were prepared using SigmaPlot.

RESULTS
Effects of Heat on in Vitro SPT Activity-Shifting yeast growth temperature from 30°C to the heat stress temperature of 39°C induces a flux through de novo sphingolipid biosynthesis resulting in transient elevation of the sphingoid bases phytoand dihydrosphingosine (6). Although this synthesis mediates the acquisition of heat tolerance in yeast, the mechanism(s) for the increased sphingolipid production remains unknown. A direct stimulation of SPT by heat could possibly explain this observation. To test this hypothesis, measurements of microsomal SPT activity in vitro at typical growth temperature (30°C) and at heat stress temperature (39°C) were obtained. At 30°C, SPT activity measurements indicated the formation of 0.318 pmol of product (ketodihydrosphingosine)/g of protein/min; however, increasing the temperature to 39°C did not affect SPT activity (0.297 pmol/g of protein/min), indicating that heat does not directly increase SPT activity in vitro. These values agree with those found in a previous study; although those assays were performed at 37°C, the results are consistent with the current finding that heat did not change SPT activity in vitro (27). Thus, it seems unlikely that increasing enzyme reaction temperature in vivo stimulates de novo sphingolipid synthesis.
Heat Does Not Induce Stimulating Modifications to SPT-An alternative hypothesis is that heat stress causes post-translational modification of SPT resulting in increased enzymatic activity (although such a modification has not been described previously). If such a modification were to occur (such as phos-phorylation or acylation) and were preserved during isolation of microsomes, one would expect increased in vitro SPT activity from heat-treated cells. Therefore, to test this hypothesis, microsomes isolated from yeast at normal growth temperature as well as from heat-stressed yeast were assayed for SPT activity. Interestingly in vitro activity assays of microsomes isolated from heat-stressed yeast showed no increase in SPT activity relative to those isolated from untreated yeast. In fact, SPT activity from these treated cells significantly decreased relative to untreated controls (from 0.32 to 0.24 pmol/g of protein/ min). This might be explained by heat-mediated denaturation of the enzyme during heat stress, which occurs for many proteins. The results from these microsomal studies suggested that neither a direct effect of heat on SPT nor heat-mediated stable post-translational modifications mediate the observed heatstimulated sphingoid base synthesis.
On the other hand, several possibilities exist that could not be ruled out based on these data; for example, heat might induce SPT to associate with other proteins in such a manner that would not be preserved during isolation of microsomes. One such interaction may involve the recently identified yeast homologue of Serinc, a protein thought to mediate supply of endogenously produced serine to SPT (28). In this study, several Serinc subtypes were identified as expressed in rat brain and found to share 24 -30% amino acid identity with the yeast protein encoded by TMS1. Moreover expression of Serinc1 in COS cells enhanced both phosphatidylserine biosynthesis and de novo sphingolipid synthesis, and Serinc1 colocalized with an red fluorescent protein-tagged Lcb1 SPT subunit and demonstrated interaction with Lcb1 in two-hybrid experiments. Additionally a tms1⌬ deletion strain was found to have ϳ50% microsomal SPT activity as compared with the background BY4743 strain. These data provided compelling evidence that Tms1p may facilitate serine accessibility by yeast SPT during heat stress. To test this hypothesis, heat-induced sphingolipid synthesis was measured in the parental BY4742 and tms1⌬ deletion strains. The heat stress in the parental strain increased dihydrosphingosine and phytosphingosine 2.7-and 6.2-fold, respectively; and in the tms1⌬ strain those increases were similar at 2.5-and 4.9-fold, respectively. Moreover basal levels of dihydrosphingosine and phytosphingosine were 1.4-and 1.6fold higher in the tms1⌬ strain, respectively, possibly suggesting that the strain may compensate for reduced in vitro SPT activity determined in the previous study. Importantly, however, these data indicated that Tms1p played only a minor role in heat-induced de novo sphingolipid synthesis.
The Role of Tsc3p in Regulation of SPT during Heat Stress-A previous report identified the small peptide encoded by the TSC3 gene as essential for full SPT activation (29). This peptide was demonstrated to interact specifically with the Lcb2 subunit of SPT (30). Data presented in that report suggested that interaction of Tsc3p with the Lcb1/Lcb2 heterodimer may regulate its activity. The product of SPT, 3-ketodihydrosphingosine, is rapidly converted to dihydrosphingosine, which in turn is converted to phytosphingosine. When derived from palmitoyl-CoA, these species contain 18 carbon atoms, and when derived from stearoyl-CoA, they contain 20 carbons. Each of these species (C 18 and C 20 phyto-and dihydrosphingosine) accumulates transiently during heat stress, and although the C 20 species are present at much lower mass levels, they increase by higher -fold changes (6). To determine the role of Tsc3p in heat-induced sphingoid base production, a tsc3⌬ deletion strain was constructed in the JK93d␣ background, and the mutant and its parental strain were heat-stressed followed by extraction of their sphingoid bases and their conversion to fluorescent derivatives. Resolution of these derivatives by HPLC as described previously (25) showed that heat stress induced the formation of C 18 chain length sphingoid bases from around 1 to 1.5-2.5 pmol/nmol of phosphate in both the wild type and tsc3⌬ strains (Fig. 1). However, deletion of TSC3 resulted in a significant decrease in specifically the C 20 chain length species relative to the wild type both basally and with heat stress. Whereas the wild type yeast contained C 20 species at around 0.2 pmol/nmol of phosphate at normal temperature, in mutant yeast, these species occurred near the low limits of detection (Fig. 1). Furthermore after heat stress, the wild type yeast increased their C 20 sphingoid base content to around 0.8 -1.0 pmol/nmol of phosphate; however, the mutant strain failed to generate more than 0.25 pmol/nmol of phosphate, indicating significant impairment of C 20 , but not C 18 , sphingoid base production in the mutant strain. This indicates a role for Tsc3p in product determination; however, deletion of TSC3 did not compromise the ability of yeast to initiate sphingolipid synthesis upon heat stress. Thus, it seems unlikely that Tsc3p mediates regulation of overall heat-induced sphingoid base production.
During the course of these studies we observed that various background strains consistently showed different levels of sphingoid bases both constitutively and after heat stress. Specifically experiments performed in the BY4742 strain yielded overall lower sphingoid base content than experiments performed in the JK93d␣ strain or the YB332 strain (below). Therefore, as with any yeast study, background strain must be taken into account prior to interpretation of data; however, -fold changes in general were consistent from strain to strain and thus may be more useful when comparing data obtained in different background strains.

The Respective Roles of Medium Fatty Acids Versus Endogenous Fatty Acids in Heat-induced Synthesis of Sphingoid Bases-Be-
cause the data suggested that heat did not directly activate the SPT complex and furthermore neither Tms1p nor the small peptide Tsc3p played a major role in regulating heat-mediated sphingoid base synthesis, it became possible that substrate availability might play some role in regulating heat-induced formation of sphingoid bases. Indeed previous studies in mammalian systems suggested that availability of SPT substrates determined levels of SPT products in cell culture (20,21). Therefore, to investigate whether substrate composition of medium could influence synthesis of sphingoid bases, heat-induced sphingoid bases were measured in BY4742 cells grown in rich medium (YPD) and then either resuspended in YPD or transferred to minimal medium (SC) immediately prior to heat stress. Indeed a significant reduction of heat stress-induced sphingoid base formation occurred in the SC ( Fig. 2A). In particular, although total sphingoid bases were similar in each medium at basal conditions (0.44 and 0.53 pmol of total bases/nmol of phosphate in YPD and SC, respectively), heat increased total sphingoid bases to 2 pmol/nmol of phosphate in YPD; however, in SC, total bases were only increased to 0.89 pmol/nmol of phosphate. This corresponds to a 4.5-fold increase in total sphingoid bases induced by heat in YPD and only a 1.7-fold increase in SC, suggesting that SC may be at least partially deficient in a factor required for heat-mediated sphingoid base production.
SPT requires both serine and palmitoyl-CoA as substrates; therefore, medium may supply serine and/or fatty acid for heatinduced production of sphingoid bases. Importantly rich medium contains a variety of lipids (at trace levels) and high concentrations of amino acids, whereas minimal medium contains amino acids at concentrations sufficient to support yeast growth but do not contain any lipid components. Therefore, heat-induced production of sphingoid bases may utilize exogenous fatty acids. Transport of exogenous fatty acids across the plasma membrane occurs through the Fat1p-Faa1p-Faa4p lipid transport complex, and the imported fatty acids subsequently undergo esterification to their CoA derivatives by a process termed "vectorial acylation" that couples transport to esterification; this creates a gradient for the fatty acids (31). Only two sources exist for fatty acids in yeast; they can come from medium by import through the Fat1p-Faa1p-Faa4p complex, or they are synthesized endogenously by FAS (31). Blocking both of these pathways by deletion of the fatty acid import complex along with addition of the FAS inhibitor cerulenin renders cells inviable (31). Moreover cerulenin-treated cells require exogenous fatty acids, and the fatty acid import complex deletion strain does not survive in the presence of cerulenin (31).
To determine whether exogenous fatty acids present in rich medium might supply SPT during heat stress, heat-induced sphingoid base production was measured in a yeast mutant deleted for the fatty acid transport complex comprised of Fat1p, Faa1p, and Faa4p as well as in its parental strain YB332. Interestingly the data indicated equivalent sphingoid base formation in this strain as compared with its parental wild type strain (Fig.  2B), suggesting that fatty acid import from the medium does not necessarily supply substrate for heat-induced de novo sphingolipid synthesis. It logically follows then that the fatty acid utilized in heat-stimulated SPT activity is derived from endogenous synthesis.
To test this hypothesis, 25 M cerulenin was added to midlog phase JK93d␣ cells to inhibit the endogenous FAS 15 min prior to heat stress. Consistent with the hypothesis, cerulenin treatment nearly completely blocked heat-induced de novo synthesis (Fig. 2C), reducing basal sphingoid bases to less than 50% of that detected in vehicle-treated cultures and, importantly, reducing the heat-induced increase in sphingoid bases from over 2-fold in the wild type (from ϳ2pmol/nmol of phosphate basally to ϳ4.5 pmol/nmol of phosphate after heat) to less than 0.5-fold after cerulenin treatment (from ϳ1 pmol/nmol of phosphate basally to barely 1.5 pmol/nmol of phosphate after heat) (Fig. 2C). The finding that impaired fatty acid import did not affect heat-induced sphingolipid synthesis coupled with the observation that FAS inhibition significantly attenuated the heat-induced synthesis indicates that endogenous fatty acid synthesis through FAS probably supplies substrate for SPT activity during heat stress (in control experiments, in vitro SPT activity was not inhibited by 25 M cerulenin).
An alternate possibility is that there are not sufficient levels of fatty acid in YPD to support the heat-induced increase in SPT activity. Consistent with this hypothesis, treatment of cells with cerulenin followed by heat stress in YPD supplemented with 100 M palmitic acid did allow heat-induced de novo synthesis of sphingolipids (not shown); however, the relevance of these conditions to normal yeast culture is questionable because YPD is rarely supplemented with fatty acids, and to our knowledge, all studies of heat-induced sphingoid base synthesis in yeast have been conducted in medium without fatty acid supplementation. Therefore, our data suggest that FAS-derived fatty acids are sufficient to supply SPT during the yeast heat stress response under typical laboratory conditions.
The Effect of Medium Serine Depletion on Heat-induced de Novo Sphingolipid Synthesis-Because minimal medium decreased sphingoid base synthesis in response to heat ( Fig. 2A) yet the fatty acid was derived from endogenous synthesis (Fig. 2, B and C), the hypothesis arose that yeast require medium serine for sphingolipid synthesis during heat stress. Standard SC contains 4 mM serine, whereas rich medium contains 5-10 mM serine (depending on components used to prepare them). The intracellular serine concentration in Saccharomyces cerevisiae has been shown to fluctuate between 1 and 5 mM (32). The K m of yeast microsomal SPT for serine has been shown to be around 4 mM (22), so these seemingly high concentrations are potentially required for appreciable SPT activity; furthermore because intracellular serine concentrations hover around the SPT K m , the enzyme activity likely depends on small changes in serine concentration. To test whether exogenous serine from medium supplies heat-stimulated SPT, yeast were grown in YPD and then transferred to either fresh YPD or SC Ser Ϫ immediately prior to heat stress. Interestingly even at basal temperature, sphingoid base levels were decreased in SC Ser Ϫ , particularly for the C 18 dihydrosphingosine species (from ϳ2.3 pmol/nmol of phosphate in rich medium to ϳ1.2 pmol/ nmol of phosphate in Ser Ϫ medium) (Fig. 3A), indicating that, FIGURE 2. Requirement for endogenous fatty acids for heat-mediated de novo sphingolipid synthesis. Yeast were cultured in rich medium at 30°C to midlog phase and either kept at 30°C or shifted to 39°C for 5 min. After collecting cells by centrifugation, lipids were extracted and quantified by HPLC as described under "Materials and Methods." A, immediately prior to heat stress, yeast were collected by centrifugation, washed in water, and resuspended in either fresh YPD or SC. B, a fatty acid importdefective mutant (fat1⌬/faa1⌬/faa4⌬) and its wild type background strain were evaluated for heat-mediated sphingoid base production as described above. C, wild type cells were treated for 15 min with 25 mM cerulenin or ethanol vehicle prior to heat stress as above. Data are presented as the mean Ϯ S.E. of two independent experiments performed in triplicate. Ctl, control; HS, heat-stressed. even in wild type cells at basal temperature, endogenous serine cannot support maximum SPT activity. Moreover the shift to medium with no serine severely compromised sphingoid base formation after heat stress (Fig. 3A) from a 1.5-4-fold increase in each species measured in YPD to less than a 1.3fold increase for the C 18 species and less than a 2-fold increase in the C 20 species (Fig. 3A) in SC Ser Ϫ . On the other hand, supplementation of SC Ser Ϫ with increasing concentrations of serine partially restored sphingoid base levels during heat stress in a dose-dependent manner (Fig. 3B, open circles) indicating a direct correlation between the medium serine available and the level of sphingoid bases produced upon heat stress. These findings confirm the requirement for exogenous serine for SPT activity during heat stress.
Because it was possible that increased medium serine might stimulate sphingoid base synthesis independently of heat, a similar experiment was performed at 30°C. Interestingly addition of increasing serine concentrations in the absence of heat did not cause a significant increase in sphingoid bases as compared with heat-treated cells (Fig. 3B, closed circles). Importantly this finding indicates that, although extracellular serine is required for the heat-mediated SPT activity, heat plays some role in allowing the enzyme to incorporate exogenous serine into sphingolipids. Moreover these data suggest that heat regulates the uptake and/or utilization of extracellular serine for sphingolipid production during heat stress.
To determine the selective requirement for exogenous serine over endogenously synthesized serine in heat-induced synthesis of sphingoid bases, a double deletion mutant was constructed in the BY4742 background strain with deletions in the SER3 and SER33 genes. These genes encode phosphoglycerate dehydrogenases that catalyze the first committed step of serine biosynthesis (33). Although other sources of endogenous serine potentially exist (e.g. proteolysis or conversion from other related amino acids including threonine and glycine), previous work demonstrated that deletion of both of these genes rendered the yeast auxotrophic for serine (33,34). Additionally growth curves in SC Ser Ϫ (not shown) indicated a requirement for Ser3 and/or Ser33 to allow survival under these conditions, indicating that other residual sources of serine in this deletion mutant were insufficient to support growth. To determine the contribution of endogenous serine biosynthesis to heat-mediated production of sphingolipids, this mutant was heat-stressed in YPD, and its sphingoid bases were quantified as described above. As shown in Fig. 3C, the serine auxotrophic strain showed no difference compared with its parental strain in heat-mediated sphingoid base synthesis. This observation, coupled with the finding of the requirement for medium serine for de novo sphingolipid synthesis during heat stress, suggests that serine utilized for this reaction is derived exclusively from exogenous sources.
The Effects of Heat on Serine Uptake-Because extracellular serine allowed the observed increased flux through SPT after heat stimulation and heat was required to maximize sphingoid base production in the presence of exogenous serine (Fig. 3B), it seemed feasible that heat may stimulate the transport of extracellular serine across the plasma membrane. A [ 3 H]serine uptake assay was utilized to measure the effects of temperature . Effects of extracellular serine concentration on heat-mediated de novo sphingolipid synthesis. Yeast were cultured in YPD at 30°C to midlog phase and harvested by centrifugation. A, pellets were resuspended in either fresh YPD or SC Ser Ϫ and cultured at 30°C or subjected to heat stress at 39°C for 5 min. Sphingoid bases were quantified as described under "Materials and Methods." Data are presented as the mean Ϯ S.E. of two experiments performed in triplicate. B, yeast were cultured to midlog phase in YPD, collected by centrifugation, and resuspended in SC Ser Ϫ with serine supplemented at increasing concentrations. The cultures were then either maintained at 30°C or subjected to heat stress at 39°C for 5 min. Sphingoid bases were measured as described under "Materials and Methods," and data are presented as total sphingoid bases. Data shown are the mean Ϯ S.E. of triplicates. C, heat-induced sphingoid base formation was measured as described above in a serine auxotroph (ser3⌬/ser33⌬) and its parental strain. Data are presented as the mean of duplicates with their range. Ctl, control; HS, heat-stressed; wt, wild type. on the rate of serine import at normal (30°C) and elevated (39°C) temperature. At 30°C, uptake of serine from medium proceeded at an average rate of about 200 pmol/10 6 cells/min over the first 10 min (Fig. 4A); however, at 39°C, the rate of uptake significantly increased to an average of about 400 pmol/ 10 6 cells/min over the time course (Fig. 4A). Moreover thinlayer chromatography of lipid extracts from 3 H-labeled cultures demonstrated incorporation of label into sphingoid bases, complex sphingolipids, and phosphatidylserine (not shown). Importantly analysis of three independent experiments performed in the BY4742 background (the same used for the serine uptake assays) indicated that heat stress induced an accumulation of 111 Ϯ 15 pmol of total sphingoid bases/10 8 cells after 5 min. This translates to ϳ0.2 pmol of sphingoid bases/10 6 cells/ min, which is at least 3 orders of magnitude less than the values of serine uptake observed. Although the fate of the serine not incorporated into sphingolipids is unknown, it may become incorporated into phosphatidylserine and/or proteins, go to the vacuole, or enter the one-carbon pool through serine hydroxymethyltransferase (34). These findings indicate that heat stimulates serine uptake from medium well above the molar ratio required to support sphingoid base synthesis, suggesting that heat-induced serine uptake could fuel the observed increase in sphingoid bases.
As mentioned above, intracellular serine concentrations are maintained between 1 and 5 mM (32), and the K m for yeast SPT is around 4 mM. Therefore, the increase in intracellular serine due to increased uptake may serve as a mechanism whereby SPT activity would increase. In support of this notion, as demonstrated in Fig. 4A, serine taken up by cells after 1 min of heat stress was increased from ϳ150 to around 400 pmol/10 6 cells (although this may not represent only free serine). Using cell volume determined at 2.38 ml/g of cell dry weight (35) and determinations of cell dry weight per unit optical density (32), 10 6 cells have a volume of ϳ0.21 l. An uptake of 400 pmol/0.21 l indicates a potential heat-mediated concentration increase of ϳ2 mM intracellularly, although this value represents total serine rather than serine available as substrate to SPT; however, the magnitude of increase in cell serine concentrations due to heat-mediated serine uptake could potentially drive intracellular concentrations to or above the K m concentration for the enzyme. To determine the feasibility of this hypothesis, cells grown in YPD to midlog phase were harvested, washed, and snap frozen in an ethanol/dry ice bath. Cells were then lysed using glass beads, and beads were removed by centrifugation. Cell lysates were then analyzed for free serine concentration. Indeed in agreement with previously published findings, intracellular serine varied from 2 to 4.5 mM. Moreover yeast isolated after heat stress demonstrated a transient increase in free serine from 30 s to 1 min after heat stress ranging from 13 to 25% (not shown).
The Effect of Extracellular Serine Concentration on Heat-dependent Serine Uptake-Previous studies in mammalian systems demonstrated that medium serine concentration directly correlates with intracellular serine concentration and levels of de novo sphingolipids (20). Therefore, although heat stimulated uptake of exogenous serine (Fig. 4A), the amount of serine taken up may correlate with medium serine concentration. To . The effects of heat on [ 3 H]serine uptake from medium. A, wild type cells were grown to midlog phase, and then [ 3 H]serine uptake activity was measured at 30 or 39°C as described under "Materials and Methods." Data are means Ϯ S.E. from three independent experiments. B, serine uptake over 5 min was quantified as in A except that extracellular serine concentration was varied as indicated. Measurements are represented as mean Ϯ S.E. from two independent experiments. C, heat-induced sphingoid base synthesis was measured in wild type and various nutrient permease deletion strains as described under "Materials and Methods" and represented as -fold change of total sphingoid bases (C 18 and C 20 dihydro-and phytosphingosine) after 5 min of heat stress. Data shown are averages Ϯ S.E. of at least two experiments performed in triplicate. Basal values showed little variation between mutants at 1.8, 1.7, 1,9, 2.1, 2.3, 2.1, and 2.3 pmol/nmol of phosphate respective to position on the graph. wt, wild type. test this hypothesis, serine uptake assays were performed as described above over a range of serine concentrations from 1 to 20 mM for 5 min. Indeed as shown in Fig. 4B, heat stimulated serine uptake ϳ2-fold over 5 min, and moreover the amount of serine taken up increased as extracellular concentrations increased from 1 to 10 mM. Importantly during heat stress, the greatest increase in uptake occurred from 5 to 10 mM, which is consistent with the finding that synthetic complete medium (containing 4 mM serine) did not support maximum sphingoid base production observed in YPD (containing 5-10 mM serine). Also of interest is that no significant increase in uptake was observed at either 30 or 39°C above an extracellular concentration of 10 mM.
Based on the graph in Fig. 4B, half-maximum serine uptake for cells at 30°C is ϳ750 pmol/10 6 cells, whereas at 39°C it is 1250 pmol/10 6 cells. Interestingly extracellular concentrations that correspond to these values are around 3-5 mM, indicating that heat may affect the K m for serine transport minimally, although V max increased over 50%.
The Effect of Modulating Amino Acid Transport on Heatstimulated de Novo Sphingolipid Synthesis-In S. cerevisiae, amino acid transport across the plasma membrane requires proton symport mediated by members of the APC family of transporters (36); there are no other known mechanisms of amino acid uptake by yeast cells. This family has Ͼ20 members, and the majority of the amino acid permeases overlap in amino acid specificity. Therefore, amino acid permeases previously described to import serine were identified and screened for their ability to generate sphingoid bases during heat stress. Although strains deleted for these permeases, which included Gap1p, Gnp1p, Dip5p, Agp1p, Agp2p, and Agp3, contained similar levels of sphingoid bases under basal conditions (legend to Fig. 4C), each showed a partial decrease in heat-stimulated formation of sphingoid bases ranging from around 20 to 50% (Fig. 4C) with the agp2⌬ deletion strain showing the greatest decrease at slightly greater than 50%. These data lend further support to the role of serine uptake in mediating the heat stress-induced sphingolipid synthesis.
The above results also suggested that factors that drive amino acid uptake at a level upstream from any individual transporter may also regulate sphingolipid synthesis. Current thinking suggests that amino acid uptake in yeast proceeds by proton co-transport according to potential across the plasma membrane, i.e. between external and internal proton concentration (37,38). To first test whether proton gradients can drive sphingolipid synthesis, yeast were grown overnight to midlog phase (A ϭ 0.5-0.8) in YPD. At this time, pH determinations of culture medium indicated a pH of 6.0 Ϯ 0.2. A small volume (Ͻ0.1% by volume) of H 3 PO 4 was added to medium to decrease pH to 4.0 Ϯ 0.1, thereby generating a gradient of excess extracellular protons. After 5 min, sphingoid base determination indicated a 1.7-and 2.5-fold increase in phytosphingosine and dihydrosphingosine, respectively, relative to non-acidified cultures, indicating that extracellular protons and/or medium acidification can drive sphingoid base synthesis (Fig. 5A). FIGURE 5. Factors that induce serine uptake increase sphingoid base synthesis in the absence of heat. A, wild type cells grown to midlog phase were treated with H 3 PO 4 to acidify medium to pH 4. After 5 min at control or acid pH, sphingoid bases were extracted and quantified as described under "Materials and Methods." Glucose treatment increased serine uptake similarly to that observed with heat stress. B, wild type cells were grown to midlog phase, washed in water three times, suspended in a non-buffered solution containing 10 mM [ 3 H]serine and 20 mM KCl, and then treated with either vehicle or 50 mM glucose. [ 3 H]Serine uptake was quantified as described under "Materials and Methods." C, cells were grown to midlog phase, resuspended in a non-buffered solution containing 10 mM serine and 20 mM KCl, and treated with either heat stress or 50 mM glucose as indicated. After 5 min cells were harvested by centrifugation, and sphingoid bases were extracted and quantified as described under "Materials and Methods." All data are presented as the mean Ϯ S.E. of two independent experiments performed in duplicate.
Under normal conditions in yeast, plasma membrane potential is maintained and/or modified by the ATP-dependent proton pump Pma1p (for a review, see Ref. 39). According to several previous studies, acute glucose treatment activates Pma1p by phosphorylation of its carboxyl terminus at several sites by kinases including Ptk2 (40,41) and subsequently strongly drives amino acid uptake (38). To test whether glucose treatment would stimulate uptake of serine, cells were grown to midlog phase, washed three times in water, and suspended in a solution of KCl containing 10 mM [ 3 H]serine. As shown in Fig. 5B, addition of glucose to the assay solution strongly increased uptake of [ 3 H]serine over 10 min, which was similar in time frame to that stimulated by heat (Fig. 4A). Although total serine taken up was significantly less after glucose treatment than heat treatment (3800 versus 1100 cpm/10 6 cells for heat versus glucose, respectively) as discussed above, this amount of serine remains orders of magnitude over the sphingoid bases measured and thus should be sufficient to mediate sphingoid base production similarly to heat stress.
Thus, for a more rigorous test of the notion that amino acid uptake is sufficient to drive de novo sphingolipid synthesis, yeast cells were grown in YPD to midlog phase, washed in sterile water three times to remove all residual medium, and suspended in a solution of KCl also containing 10 mM serine because of the previous findings described above that 10 mM serine was required in medium to allow maximum uptake of exogenous serine (Fig. 4B) and also maximum sphingoid base production (Fig. 3B). Determination of sphingoid base levels in control samples maintained in this solution for 5 min at the non-heat stress temperature of 30°C indicated phyto-and dihydrosphingosine levels of around 2 pmol/nmol of phosphate. Consistent with the hypothesis, however, treatment of samples with either heat stress or 50 mM glucose caused a 2-3fold increase in sphingoid bases (Fig. 5C). Importantly the heattreated and glucose-treated samples sustained a similar degree of sphingoid base increase within the same time frame. This result along with the result shown in Fig. 5A indicated that factors other than heat that drive amino acid uptake can also drive sphingolipid synthesis similar to heat stress.

DISCUSSION
Despite the ever increasing amounts of data elucidating sphingolipid functions in yeast and mammalian cells, very little is known about how de novo sphingolipid synthesis is regulated. In particular, sphingolipid synthesis mediates yeast adaptation to heat (6,9); however, the mechanism by which heat increases de novo sphingolipid synthesis remains unknown. The current studies were performed to identify potential mechanisms by which heat stimulates de novo sphingolipid synthesis. Data presented herein indicate that neither direct nor indirect heatmediated SPT activation likely causes the observed increase in sphingolipid synthesis. Furthermore because medium composition (i.e. medium serine concentration), FAS activity, and amino acid transport strongly affect the amount of sphingoid bases formed during heat stress, substrate supply appears to regulate de novo SPT activity during heat stress. Quite interestingly, the results from this study point to differential sources of the serine and fatty acid substrates from endogenous versus exogenous sources (Fig. 6). Finally the observation that heat stimulates serine transport across the plasma membrane and the finding that stimulation of amino acid uptake by changes in medium pH or by acute addition of glucose also stimulate de novo sphingolipid synthesis strengthen this hypothesis and provide mechanistic links between heat and de novo synthesis of sphingolipids.
Importantly deletion of the small subunit of SPT, Tsc3p, did not affect bulk de novo synthesis in response to heat, but it affected primarily substrate utilization (i.e. decreased utilization of stearoyl-CoA) of SPT. This interesting finding raises intriguing possibilities as to how Tsc3p may regulate substrate availability/specificity of SPT. This also may explain observations from this and previous studies (5, 6) that heat increases much larger -fold changes in C 20 than in the C 18 sphingoid bases. However, this difference might also be explained if metabolic activities downstream from sphingoid bases, i.e. sphingoid base kinase and ceramide synthase, preferentially utilize the C 18 species. The mechanism for this specificity and potential roles for C 18 versus C 20 sphingoid bases remain to be determined.
The finding that medium affected sphingolipid production during heat stress raised the possibility that medium may supply substrates to SPT, in particular free fatty acids. However, a yeast mutant defective in uptake of exogenous fatty acids showed normal sphingolipid synthesis during heat treatment. On the other hand, inhibition of fatty acid synthesis by cerulenin markedly decreased sphingoid base synthesis. Together these findings indicate that under normal laboratory conditions endogenous fatty acid produced by the FAS complex supplies the fatty acid component for heat-induced production of sphingoid bases (Fig. 6). Moreover under these conditions, exogenous fatty acid was not able to compensate for loss of FAS activity, and thus SPT selectively uses endogenous fatty acid.
Because SPT requires fatty acyl-CoA and serine as substrates, it seemed likely that medium supplied serine to SPT during heat stress. Indeed this was demonstrated by performing heat stress in medium devoid of serine in which yeast did not produce sphingoid bases upon heat stimulation. However, replacing serine in the medium allowed sphingoid base synthesis in a dosedependent and heat-dependent manner. On the other hand and quite interestingly, disruption of genes required to synthesize serine intracellularly did not impair sphingoid base synthesis during heat stress. Thus, SPT appears to selectively utilize exogenous serine during heat stress, and similar to the situation with fatty acid, endogenous serine is unable to compensate for loss of medium serine. This is consistent with the finding that deletion of the yeast homologue of Serinc (TMS1), which encodes a protein thought to facilitate delivery of endogenously derived serine to SPT, had little to no effect on heat-induced sphingoid bases.
These findings led us to propose a model (Fig. 5) whereby heat stimulates serine uptake from medium that can be blocked by depletion of medium serine, and FAS supplies SPT with fatty acid (which becomes esterified to CoA prior to utilization), which cerulenin inhibits. Without either of the sources of substrates, sphingoid bases cannot be produced from alternate substrate sources (i.e. exogenous fatty acids or endogenous serine). The reason for the apparent selectivity remains unknown. One possibility is that the localization of fatty acids after import or serine after biosynthesis precludes their incorporation into sphingolipids. Another possibility is that the rate of fatty acid import and/or serine synthesis is not sufficient to supply SPT to generate the acute 2-10-fold increase in sphingoid bases.
Because heat stimulated serine uptake from medium and sphingoid bases failed to increase during heat stress in the absence of extracellular serine, it was possible that heat-mediated amino acid uptake drives heat-mediated sphingoid base synthesis. If this hypothesis is true, then any factors that drive amino acid uptake in the absence of heat must also drive de novo sphingoid base synthesis. Indeed two conditions that drive amino acid uptake by different mechanisms, namely acid treatment, which increases the proton gradient from extracellular to intracellular spaces and drives protoncoupled amino acid import (37), and glucose treatment, which indirectly stimulates amino acid transport by activation of the Pma1 proton pump (41), were assayed for their ability to cause de novo sphingolipid synthesis in the absence of heat. Indeed each of these conditions caused increases in de novo synthesis of -fold change and time course similar to those increases observed during heat stress. These findings support the hypothesis that during heat stress heat-induced serine uptake is a key determinant and/or the driving force behind the observed increase in de novo sphingolipids. The plausibility of this scenario gains credence from the relatively high K m of SPT for serine (4 mM) (22,27) and reports that intracellular serine con-centrations fluctuate in this range (1-5 mM under normal conditions) (32). This would leave SPT acutely sensitive to changes in intracellular serine. Therefore, increased serine uptake seems to be a possible regulatory mechanism for SPT activation, and moreover, data presented here demonstrate that increased intracellular serine is both necessary and sufficient to cause the increased flux through de novo synthesis observed during heat stress. Yeast survive heat stress only by initiation of a pleiotropic response including cell cycle arrest, translation of heat shock proteins, transcriptional reprogramming, trehalose synthesis, and degradation of misfolded proteins (42). Despite the fundamental biological importance of these events, the primary signal that initiates this response upon temperature increase has remained elusive, although some event occurring at or near the plasma membrane may serve this function (43). Previous data indicate that heat-induced sphingoid bases mediate, at least in part, each of these necessary responses, and thus events that initiate sphingolipid synthesis may very well lie immediately proximal to the elusive primary signal (5, 8 -10, 44). Because heat stimulates serine uptake to drive sphingolipid synthesis, the event that connects heat to serine uptake may initiate the entire response. Several possibilities seem likely, including that heat stimulates Pma1, generating a proton gradient to drive serine uptake. This hypothesis is further supported by the previous observation that maximum activation of Pma1p occurs at the heat stress temperature of 42°C (45). On the other hand, heat-induced changes in membrane fluidity may stimulate amino acid uptake directly. Thus, studies are currently underway to identify the putative primary signals that initiate the heat stress response. These and other studies may allow placement of heat-stimulated synthesis of sphingoid bases within the broader cellular context of coupling alterations in membrane proton gradients to nutrient import.
Previous studies have shown that sphingoid bases generated during the heat stress response mediate the ubiquitin-dependent degradation of nutrient permeases of the same family as Agp2p, including Fur4p and Gap1p, and significantly reduce import of nutrients including uracil, histidine, tryptophan, and leucine (7,44). Although sphingoid base-dependent degradation of serine transporters or inhibition of serine uptake was not specifically tested in these studies, data shown here present the opportunity for a negative feedback regulation of nutrient permeases, i.e. heat stress stimulates permeases to increase amino acid uptake and incorporation of serine into sphingoid bases, which leads to degradation of nutrient permeases, possibly including those responsible for heat-induced serine transport.
The potential for substrate availability to regulate de novo sphingolipid synthesis may provide an explanation for observations in mammalian systems where, although heat stimulated sphingolipid production, no in vitro activation of SPT was detected (46). The hypothesis that serine can mediate de novo sphingolipid synthesis in mammalian systems as well as yeast is further supported by data from a previous study demonstrating a burst of de novo sphingolipid synthesis upon medium change in J774A.1 cells that correlated directly to the serine concentration in medium (21). Interestingly an increased supply of exogenous free fatty acid to cultured cells can also stimulate de novo sphingolipid synthesis in skeletal muscle and cultured myo-tubes (47,48), pancreas (49), and several other tissues and/or tissue culture model systems (50 -52) often with pathophysiological consequences. Whether oversupply of serine can mediate similar effects is also under current investigation. In general, however, supply of either substrate may drive de novo synthesis in a multitude of experimental situations; if so, the concept that metabolic flux through a biochemical pathway important for signaling depends on substrate supply is intriguing.
In conclusion, although several studies have shown increased de novo sphingolipid synthesis correlating to substrate supply, few studies, if any, have identified regulatory mechanisms to explain the observed increases in SPT substrates and thus activity. To our knowledge this is the first study in yeast to demonstrate that acute increases in substrate supply drive de novo synthesis. The regulation of SPT substrate supply and availability may thus emerge as a key regulator of de novo sphingolipid synthesis.