Tsc1+ and tsc2+ Regulate Arginine Uptake and Metabolism in Schizosaccharomyces pombe*

Mutations in either TSC1 or TSC2 cause tuberous sclerosis complex, an autosomal dominant disorder characterized by seizures, mental retardation, and benign tumors of the skin, brain, heart, and kidneys. Homologs for the TSC1 and TSC2 genes have been identified in mouse, rat, Fugu, Drosophila, and in the yeast Schizosaccharomyces pombe. Here we show that S. pombe lacking tsc1+ or tsc2+ have similar phenotypes including decreased arginine uptake, decreased expression of three amino acid permeases, and low intracellular levels of four members of the arginine biosynthesis pathway. Recently, the small GTPase Rheb was identified as a target of the GTPase-activating domain of tuberin in mammalian cells and in Drosophila. We show that the defect in arginine uptake in cells lacking tsc2+ is rescued by the expression of a dominant negative form of rhb1+, the Rheb homolog in S. pombe, but not by expressing wild-type rhb1+. Expression of the tsc2+ gene with a patient-derived mutation within the GAP domain did not rescue the arginine uptake defect in tsc2+ mutant yeast. Taken together, these findings support a model in which arginine uptake is regulated through tsc1+, tsc2+, and rhb1+ in S. pombe and also suggest a role for the Tsc1 and Tsc2 proteins in amino acid biosynthesis and sensing.

Tuberous sclerosis complex (TSC) 1 is a tumor suppressor syndrome that is characterized by the development of a variety of benign tumors (hamartomas) and severe neurological problems including seizures, mental retardation, and autism. TSC is caused by mutations in either TSC1 (1) or TSC2 (2). Hamartin, the TSC1 gene product, and tuberin, the TSC2 gene product, are known to interact (3,4) and function in a complex. Tuberin has a highly conserved GTPase-activating protein (GAP) domain with activity for Rheb1 (Ras homolog enriched in brain) (5)(6)(7)(8)(9)(10)(11), a small GTPase that may be involved in nutrient signaling and cell cycle regulation (12).
Schizosaccharomyces pombe contains genes with significant similarity to TSC1 and TSC2, which were named tsc1 ϩ and tsc2 ϩ (24). The GAP domain of tuberin is particularly highly conserved with 39% identity. In addition to TSC1 and TSC2 homologs, S. pombe also has a Rheb homolog, rhb1 ϩ . Loss of rhb1 ϩ in S. pombe results in growth arrest, similar to that caused by nitrogen starvation (25), and loss of farnesylation of the Rhb1 protein has been postulated to regulate arginine uptake in S. pombe (26).
Recently, S. pombe strains lacking tsc1 ϩ (⌬tsc1) or tsc2 ϩ (⌬tsc2) were shown to have abnormal localization of the amino acid permease c359.03 ϩ , suggesting that the S. pombe Tsc1-Tsc2 protein complex regulates protein trafficking (24). Here we report for the first time that ⌬tsc1 and ⌬tsc2 have a defect in arginine uptake, which is regulated through rhb1 ϩ in S. pombe, providing evidence that TSC-Rheb signaling is conserved in S. pombe. A mutation in the GAP domain of tsc2 ϩ at a site corresponding to a patient-derived missense mutation could not rescue the uptake defects, strengthening the relationship of the S. pombe model to human TSC. The transcriptional expression profile and intracellular amino acid levels associated with ⌬tsc1 and ⌬tsc2 overlapped extensively, suggesting similar roles for tsc1 ϩ and tsc2 ϩ in S. pombe. These findings support S. pombe as a model for TSC and indicate that the S. pombe Tsc1 and Tsc2 proteins play central roles in amino acid biosynthesis and sensing.

EXPERIMENTAL PROCEDURES
Yeast Strains, Media, and Growth Conditions-The yeast strains used in this study are listed in Table I. CHP428 and CHP429 were constructed by Charlie Hoffman (Boston College) and were a gift from Janet Leatherwood (University of New York, Stony Brook, NY). Wildtype strain 972 (27) and ura4-D18 (28) were gifts from J. Bä hler (Sanger Institute). S. pombe cells were grown in essential minimal medium (EMM, Qbiogene, Carlsbad, CA) or yeast extract complete medium with 50 g/ml uracil, histidine, adenine, and leucine (YES) at 30°C unless otherwise stated. Transformations were performed with Frozen-EZ yeast transformation II kit (Zymo Research, Orange, CA).
Construction of tsc1 ϩ and tsc2 ϩ -deficient Strains-Tsc1 ϩ and tsc2 ϩdeficient strains were constructed with the PCR one-step homologous recombination method (29). The kanamycin cassette was amplified from plasmid pFA6a-kanMX6 (a gift from J. Bähler) using primers with 75 extra bases corresponding to sequences immediately upstream of the start codon of the tsc genes and primers whose gene-specific portions correspond to sequences 75 bases downstream of the gene. For gene disruption of tsc1 ϩ , the entire open reading frame was deleted from the genome of the * This work was supported by a fellowship from the Polycystic Kidney Disease Foundation (to M. v. S.) and by the Department of Defense. 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.
haploid strain CHP429 (h Ϫ , leu1-32, ura4-D18, ade6 -216, his7-366) and replaced by the kanamycin cassette to create MVS5. We will refer to ⌬tsc1 with this genotype as F15⌬tsc1. ⌬tsc2 was constructed using identical strategy and resulted in MVS6, which will be referred to as F15⌬tsc2 in figures and text. Correct integration of the kanamycin cassette into the yeast genome was confirmed by PCR over the integration site, by Southern blotting, and by sequencing. Subsequently, F15⌬tsc1 and F15⌬tsc2 were crossed into the ura4-D18 strain to generate MVS3 (ura4⌬tsc1) and MVS4 (ura4⌬tsc2) and into wild-type 972 to generate MVS1 (972⌬tsc1) and MVS2 (972⌬tsc2) using random spore analysis on selective plates. , and F15⌬tsc2 strains were grown in rich liquid medium with supplements (YES) overnight to mid-log phase (A 595 ϭ 0.4 -0.6) and diluted to A 595 ϭ 0.2. Cells were grown for an additional 6 h, and the generation time was determined between 3 and 6 h. The F15⌬tsc1 and F15⌬tsc2 needed ϳ5 h to complete one generation compared with 3.5 h for wild-type strains. B, 40,000 F15, F15⌬tsc1 and F15⌬tsc2 cells were spotted on YES ϩ G418 (200 g/ml). F15⌬tsc1 and F15⌬tsc2 were G418-resistant as expected. The cells were also spotted on YES plates and grown at 25, 30, and 37°C for 3 days. No growth differences were seen among F15, F15⌬tsc1 and F15⌬tsc2. C, F15, F15⌬tsc1, and F15⌬tsc2 cells were spotted (40,000 cells) onto EMM plates with different amounts of adenine, leucine, histidine, and uracil. At regular amounts (50 g/ml), the F15⌬tsc1 and F15⌬tsc2 could not grow but growth was partially restored by increasing the amount of supplements to 1 mg/ml. D, Tsc1 and Tsc2 expression constructs in pREP4X (ura4 ϩ ) were transformed into F15⌬tsc1 and F15⌬tsc2 and plated onto EMM plates supplemented with 50 g/ml leucine, adenine, and histidine. pREP4X-Tsc1 expression rescued the growth of F15⌬tsc1, and pREP4X-Tsc2 expression rescued the growth of F15⌬tsc2. Construction of Plasmids-Tsc1 and Tsc2 expression constructs were generated by a PCR-cloning approach. The tsc1 ϩ and tsc2 ϩ genes were amplified from the cosmids c23F3 and c630C13 (a gift from J. Bä hler) using primers with SalI restriction sites and cloned into the pREP4X expression vector (ATCC). The Rhb1 expression construct was generated by PCR of full-length rhb1 ϩ from total cDNA using primers with SalI and XmaI and cloned into pREP4X. After sequence verification, tsc1 ϩ , tsc2 ϩ , and rhb1 ϩ were inserted in-frame into the HA-tagged pSLF173/273/373 series with different nmt (no message in thiamin) promoter strength (ATCC). The GAP (Tsc2-N1191K) and the Rhb dominant negative (Rhb-D60K) mutations were introduced into the pSLF373 constructs using site-directed mutagenesis (Stratagene, La Jolla, CA). All of the constructs were verified by sequencing.
Expression Profiling-Yeast were grown overnight in EMM to early log phase (A 595 ϭ 0.2-0.3), and total RNA was isolated by phenol extraction and purified using RNeasy (Qiagen, Valencia, CA). The total RNA of three independent biological samples was pooled (10 g of each sample), reverse-transcribed into cDNA, and labeled with Cy3 and Cy5 (Amersham Biosciences). Hybridizations were carried out overnight at 42°C. The slides were scanned with a GMS 428 scanner (Affymetrix, Santa Clara, CA), and spot quantification was performed with the ImaGene software (BioDiscovery, Marina del Rey, CA). Each of the 4976 S. pombe genes was present in duplicate on each slide, and the experiments were repeated using opposite labels (dye-flip), resulting in a total of four measurements for each gene per sample. Genes were considered expressed when all four measurements exceeded a threshold of 3.5 times above the background. A linear regression normalization was applied to the data (30), and fold changes were calculated. Genes were grouped and annotated on the basis of predicted function in the Proteome Knowledge Library (Incyte, Beverly, MA).
Western Blot Analysis-Yeast were grown to mid-log phase (A 595 ϭ 0.4 -0.6) and washed once in ice-cold buffer containing 50 mM Tris-HCl, pH 7.5, 25 mM NaCl, and 0.1 mM phenylmethylsulfonyl fluoride. Cells were lysed in the same buffer with 0.5 mm-glass beads in a BeadBeater (Biospec Products, Bartlesville, OK). 20 g of each sample was run on 4 -20% SDS-PAGE gel (Bio-Rad) and transferred to nitrocellulose membranes using standard methods. The immobilized proteins were detected using enhanced chemiluminescence (Amersham Biosciences). Antibodies used were anti-HA (Roche Applied Science) and monoclonal TAT1 for S. pombe tubulin (gift from K. Gull, University of Manchester, United Kingdom).
Northern Blot Analysis-10 g of total RNA was run on a 1% formaldehyde gel at 60 V for 4 h and transferred to nylon membrane overnight in 20ϫ SSC. Probes for p7G5.06 ϩ , c869.10 ϩ , isp5 ϩ , and gpd3 ϩ were PCR-amplified from cDNA, cleaned over 0.8% agarose gel, and labeled with [␣-32 P]dCTP (PerkinElmer Life Sciences) using standard methods. Hybridizations were performed in rapid hybridization buffer (Amersham Biosciences).
Arginine Uptake Assays-Arginine uptake assays were performed in triplicate as described by Urano et al. (31) with minor modifications. Cells were grown in EMM minimal medium with no supplements to mid-log phase. 1 Ci of L-[ 3 H]arginine (40 -70 Ci/mmol) (PerkinElmer Life Sciences) and 100 M of non-radioactive arginine (Sigma) were added to 25,000 cells in 600 l of EMM. 200-l aliquots were removed at 0 and 10 min, injected into 5 ml of deionized water, and immediately subjected to vacuum manifold filtration. Cells were collected on Whatman glass microfiber filters, washed twice, and dried. [ 3 H]Arginine was measured by scintillation counting.
Measurements of Intracellular Amino Acid Pools-Protein extracts were prepared as described under Western blot analysis, quantified using the Bradford assay (Bio-Rad), and diluted to 1 g/l. Proteins were precipitated by treatment of 100 l of sample with 100 l of 10% 5-sulfosalicylic acid at 4°C for 1 h. The pH value of the supernatant was adjusted to 2.2 with 3 M LiOH. 100 l of sample was injected into the Biochrom 30 amino acid analyzer (Biochrom, Cambridge, United Kingdom) including a standard amino acid mixture of 10 nM (Sigma).

F15⌬tsc1 and F15⌬tsc2
Have Growth Defect-As a first step toward understanding the physiological functions of tsc1 ϩ and tsc2 ϩ , we disrupted tsc1 ϩ and tsc2 ϩ in the S. pombe genome by one-step gene replacement. To initiate phenotypic analysis, the F15⌬tsc1 strain was mated with the CHP428 strain (h ϩ , leu1-32, ura4-D18, ade6 -210, his7-366) and spores were analyzed on yeast extract medium supplemented with 50 g/ml leucine, uracil, adenine, and histidine (YES). Dissection of asci from het-FIG. 2. The 972⌬tsc1 and 972⌬tsc2 have a decreased uptake of arginine, which can be restored in the 972⌬tsc2 by expressing a dominant negative Rhb1 mutation. A, wild-type 972, 972⌬tsc1, and 972⌬tsc2 were grown in EMM without supplements overnight to midlog phase, and cells were diluted to A 595 ϭ 0.4. Three different dilutions (40,000 -4000-400 cells) were spotted on EMM plates with and without 60 g/ml canavanine. Plates were incubated at 30°C for 3 days. Canavanine killed wild-type cells, but 972⌬tsc1 and 972⌬tsc2 were resistant to 60 g/ml canavanine. B, cells were grown until mid-log phase (A 595 ϭ 0.5) in EMM, and 25,000 cells were resuspended in 100 M L-arginine with 1 Ci of L-[ 3 H]arginine (40 -70 Ci/mmol) in 600 l. Aliquots of 200 l were injected into 5 ml of EMM in a vacuum manifold at 0 -10 min, washed twice with 5 ml of deionized water, and assayed for L-[ 3 H]arginine uptake. The arginine uptake was 3.5-fold less in the 972⌬tsc1 and 972⌬tsc2 strains compared with wild-type 972. Experiments were done in triplicate, and similar results were seen in two independent experiments. C, the arginine uptake was measured for the ura4⌬tsc2 strain expressing empty vector, Tsc2, Rhb1, and dominant negative Rhb-D60K from a plasmid with ura4 ϩ . Expression of either Tsc2 or Rhb1-D60K restored the arginine uptake. erozygous diploid cells showed that two of four colonies were smaller in size (data not shown). These smaller colonies were found by PCR to be ⌬tsc1. Similar results were obtained for the F15⌬tsc2 strain. The slower growth phenotype on YES medium was quantified in exponential liquid-growing cultures. The generation time (time required for cell population to double) of F15⌬tsc1 and F15⌬tsc2 was ϳ5 h compared with 3.5 h for the F15 strain (Fig. 1A). To test whether growth was further affected by temperature stress, F15⌬tsc1 and F15⌬tsc2 were plated on YES plates and incubated at 25 and 37°C. No temperatureinduced growth defect was observed in the F15⌬tsc1 and F15⌬tsc2 strains (Fig. 1B).
F15⌬tsc1 and F15⌬tsc2 Are Conditionally Lethal-We found a more severe growth defect in the F15⌬tsc1 and F15⌬tsc2 Tubulin is shown as a loading control. C, constructs expressing wild-type Tsc2 and Tsc2-N1191K were transformed into the F15⌬tsc2 strain and plated onto EMM plates supplemented with 50 g/ml leucine, adenine, and histidine and without uracil. Tsc2 expression rescued growth, whereas the Tsc2-N1191K mutation did not. D, canavanine sensitivity was measured in ura4⌬tsc2 expressing empty vector, Tsc2, or Tsc2-N1191K from a plasmid with ura4 ϩ . Tsc2 expression restored the canavanine sensitivity in ⌬tsc2, whereas Tsc2-N1191K expression did not. E, arginine uptake was measured in ura4⌬tsc2 expressing empty vector, Tsc2, or Tsc2-N1191K. The uptake defect was rescued by expressing Tsc2 but not by Tsc2-N1191K. strains when they were grown on EMM plates. F15⌬tsc1 and F15⌬tsc2 yeast could not grow on EMM supplemented with normal amounts (50 g/ml) of uracil, histidine, adenine, and leucine, but increasing supplements to 1000 g/ml partially restored growth (Fig. 1C). These results are in agreement with the previously reported defect in uptake of leucine, adenine, and histidine (24). To verify that the growth defect was attributed to deletion of the tsc1 ϩ and tsc2 ϩ genes, Tsc1 and Tsc2, expressed from a plasmid with ura4 ϩ , were transformed into the F15⌬tsc1 and F15⌬tsc2 strains and were plated on EMM without uracil. Expression of Tsc1 restored growth in F15⌬tsc1 yeast but failed to rescue F15⌬tsc2, whereas Tsc2 expressed restored growth in F15⌬tsc2 but not in F15⌬tsc1 (Fig. 1D).
972⌬tsc1 and 972⌬tsc2 Have a Defect in Arginine Uptake-Previously, Rhb1 was shown to regulate arginine uptake in S. pombe (26), prompting us to determine whether Tsc1 and Tsc2 also regulate arginine uptake. Because F15⌬tsc1 and F15⌬tsc2 have a growth defect, we crossed ⌬tsc1 and ⌬tsc2 into the 972 FIG. 4. 972⌬tsc1 and 972⌬tsc2 show a significant overlap in expression profile. A, correlation plot of the average gene expression ratios from the ⌬tsc1 dye-flip experiment. B, expression profiles were compared among 972, 972⌬tsc, and 972⌬tsc2. At fold change of Ͼ1.5, there were 14 down-regulated and 26 up-regulated genes in common. C, expression of three permease genes, p7G5.06 ϩ , c869.10 ϩ , and isp5 ϩ was determined by Northern blots. Fold changes were determined by densitometry. All three genes were down-regulated in 972⌬tsc1 and 972⌬tsc2, consistent with the array result.
background. This strain does not require amino acid supplements, and 972⌬tsc1 and 972⌬tsc2 did not show a growth defect on EMM. We found that 972⌬tsc1 and 972⌬tsc2 are resistant to 60 g/ml canavanine, a toxic analog of arginine ( Fig. 2A). This dose of canavanine was toxic to the wild-type 972 strain. To determine whether the canavanine resistance was the result of decreased uptake, the uptake of [ 3 H]arginine was measured. After 10 min, arginine uptake was ϳ3.5-fold less in the 972⌬tsc1 and 972⌬tsc1 strains compared with wild-type 972 (Fig. 2B), indicating that the canavanine resistance is the result of decreased uptake.
Dominant Negative Rhb1 Can Rescue the Arginine Uptake in ura4⌬tsc2-A recent screen in S. pombe identified a dominant negative Rhb1 mutation, Rhb1-D60K, that is unable to bind GTP or GDP (32). We generated this mutation in the pSLF373-ura4 ϩ expression vector and crossed ⌬tsc2 into the ura4-D18 strain to allow selection for cells expressing from the pSLF373-ura4 ϩ plasmid. We found that the decreased arginine uptake in ⌬tsc2 was restored by expression of Rhb1-D60K but not by wild-type Rhb1 (Fig. 2C), suggesting that arginine uptake is regulated through Tsc1, Tsc2, and Rhb1 in S. pombe.
A Missense Mutation in the GAP Domain of tsc2 ϩ Does Not Rescue the Conditional Lethality or Arginine Uptake in ⌬tsc2-The Tsc2 GAP domain in S. pombe is 39% identical to the GAP domain in human tuberin, and the conserved residues include the sites of three patient-derived TSC2 missense mutations (Fig. 3A). To determine whether these residues are crucial for the function of Tsc2 in S. pombe, we constructed one of them, Tsc2-N1191K, which corresponds with N1643K in human, in the HA-tagged pSLF373-ura4 ϩ expression vector. Western blot analysis showed protein expression for both Tsc2 and Tsc2-N1191K (Fig. 3B). The Tsc2 and Tsc2-N1191K expression constructs were transformed next into F15⌬tsc2, and cells were plated on EMM plates without uracil but with 50 g/ml leucine, adenine, and histidine. The wild-type Tsc2 expression construct restored growth, but no growth was detected when the Tsc2-N1191K mutation was expressed (Fig. 3C). We next asked whether re-introducing Tsc2-N1191K could revert the canavanine resistance in the ura4⌬tsc2 strain. Wild-type Tsc2 restored the canavanine sensitivity, whereas Tsc2-N1191K did not (Fig. 3D). The decreased arginine uptake was similarly rescued by wild-type Tsc2 but not by Tsc2-N1191K (Fig. 3E). These results suggest that the function of Tsc2 in regulating arginine uptake requires the GAP domain and support the use of S. pombe as a model system for human TSC.
⌬tsc1 and ⌬tsc2 Show a Significant Overlap in Expression Profile-To elucidate the mechanism through which tsc1 ϩ and tsc2 ϩ regulate amino acid uptake, we compared the expression profile of 972⌬tsc1 and 972⌬tsc2. Total RNA was isolated from 972, 972⌬tsc1, and 972⌬tsc2 yeast, labeled, and hybridized to cDNA arrays (Eurogentec, Liege, Belgium). The expression profile of 972⌬tsc1 was compared with 972 on two separate arrays including a dye-flip experiment. The 972-972⌬tsc2 comparison was completed using the same design. Because all four arrays showed a linear relation between Cy3 and Cy5, a linear regression normalization was applied to the data (30). In addition, as shown in Fig. 4A, the dye-flip experiment for ⌬tsc1 was highly correlated. The expression data were also validated by the absence of tsc1 ϩ expression in the ⌬tsc1 and absence of tsc2 ϩ in the ⌬tsc2 arrays, serving as internal controls.
There was a high degree of overlap in expression profile between 972⌬tsc1 and 972⌬tsc2. In total, 14 genes were downregulated at least 1.5-fold and 26 genes were up-regulated at least 1.5-fold both in 972⌬tsc1 and 972⌬tsc2 (Fig. 4B). Table II lists the genes that were up-regulated at least 1.5-fold in both ⌬tsc1 and ⌬tsc2. Many of the up-regulated genes have predicted roles in iron transport and amino acid metabolism including the arginase gene, car1 ϩ . Table III lists the genes that were down-regulated at least 1.5-fold in both ⌬tsc1 and ⌬tsc2. Many of the down-regulated genes were putative transporters including three amino acid permeases, two oligopeptide transporters, two polyamine transporters, and one with homology to vitamin/cofactor transporters. Interestingly, the three down-regulated amino acid permeases had homology to the Gap1p (general amino acid permease) in Saccharomyces cerevisiae. The expression change for these three permeases was confirmed by Northern blotting (Fig. 4C). The fold change on the Northern blot was determined by densitometry and was in each case slightly greater than the fold change on the array, further validating the array result. These data support that Tsc1 and Tsc2 function in the same pathway in S. pombe and suggest that they have a central role in the regulation of the biosynthesis and uptake of amino acids, oligopeptides, and polyamines.  Intracellular Amino Acid Concentrations Are Decreased in 972⌬tsc1 and 972⌬tsc2-The down-regulation of permease expression and decreased uptake of amino acids in the 972⌬tsc1 and 972⌬tsc2 strains could represent an appropriate response to high intracellular amino acid concentrations. However, we found that the intracellular levels of multiple amino acids were low in 972⌬tsc1 and 972⌬tsc2 compared with 972 wild-type yeast (Fig. 5A). Ornithine, which is a product of both glutamate and arginine metabolism, showed the largest relative decrease from ϳ15 nM in wild-type 972 to nearly undetectable levels in 972⌬tsc1 and 972⌬tsc, whereas lysine was not changed (Fig. 5B). A decrease of at least 40% was detected for alanine, asparagine, histidine, glutamine, ornithine, citrulline, and arginine. Interestingly, the latter four are linked to arginine biosynthesis (Fig. 5C). The low intracellular amino acid levels, combined with the low amino acid permease expression levels and the decreased arginine uptake, strongly suggest that yeast lacking tsc1 ϩ or tsc2 ϩ have an intrinsic defect in amino acid sensing. DISCUSSION We report here that S. pombe lacking tsc1 ϩ or tsc2 ϩ have defects in amino acid transport, involving not only the permease localization reported previously (24) but also decreased expression of amino acid permeases, decreased uptake of arginine, and low intracellular amino acid levels.
The decreased uptake of arginine in the ⌬tsc2 cells could be restored by expressing wild-type tsc2 ϩ but not by expressing the tsc2 ϩ gene carrying a mutation in the highly conserved GAP domain. This mutation is homologous to the patient-derived N1643K. Interestingly, the small GTPase Rheb was recently identified as the key target of the GAP domain of the TSC2 gene product, tuberin, in mammals and Drosophila (5)(6)(7)9). From previous studies, it was known that Rhb1 regulates arginine uptake in S. pombe (25) as well as in S. cerevisiae (31). We found that the arginine uptake defect in the ⌬tsc2 yeast was rescued by expression of a dominant negative form of S. pombe Rhb1, D60K. The rescue by Rhb1-D60K suggests that Rhb1 is downstream of Tsc2 in S. pombe as well as in other species and further strengthens the relevance of the S. pombe model to human TSC.
Previously, the mislocalization of an amino acid permease c359.03 ϩ (GenBank TM accession number CAB91572) in ⌬tsc1 and ⌬tsc2 was postulated to be the result of aberrant protein trafficking (24). However, we found that the expression of three other amino acid permeases with high homology to the general amino acid permease (Gap1p) in S. cerevisiae were down-regulated both in ⌬tsc1 and ⌬tsc2 cells. The permease c359.03 ϩ FIG. 5. Intracellular amino acid levels are low in the 972⌬tsc1 and 972⌬tsc2. A, intracellular amino acid levels in 972⌬tsc1 and 972⌬tsc2 were compared with 972 wild-type yeast. A decrease of at least 40% was detected for alanine, asparagine, histidine, glutamine, ornithine, citrulline, and arginine in 972⌬tsc1 and 972⌬tsc2. Two biological replicates were run for each sample, and similar results were seen in two independent experiments. B, ornithine and lysine amino acid profile in wild-type 972, 972⌬tsc1, and 972⌬tsc2. Ornithine levels were greatly decreased in 972⌬tsc1 and 972⌬tsc2, whereas lysine levels were similar for wild-type 972, 972⌬tsc1, and 972⌬tsc2. C, arginine metabolism in S. pombe. Enzymes are in italic. Arginase converts arginine into ornithine, a precursor of polyamines.
was not down-regulated in ⌬tsc1 or ⌬tsc2, suggesting that permeases are regulated at both the transcriptional and posttranslational levels in ⌬tsc1 and ⌬tsc2. In S. cerevisiae, decreased expression of GAP1 and sorting of Gap1p from the plasma membrane to the vacuole are the appropriate response to high levels of intracellular amino acids (33). In contrast, the decreased permease expression in ⌬tsc1 or ⌬tsc2 yeast was associated with low intracellular amino acids including alanine, asparagine, histidine, glutamine, ornithine, citrulline, and arginine. The inability to respond appropriately to low amino acid levels suggests that Tsc1 and Tsc2 play a role in amino acid sensing and would argue that the expression levels of permeases as well as their localization are crucial in coordinating sensing and growth in S. pombe.
Altered intracellular amino acid levels have not been detected in mammalian cells lacking tuberin or hamartin, although to our knowledge only limited studies have looked into this phenomenon. The only study published so far measured the levels of valine, leucine, phenylalanine, and lysine in Drosophila S2 cells treated with TSC2 small interference RNA (17). A difference in intracellular levels was not detected, but those four amino acids were not changed in S. pombe lacking tsc1 ϩ and tsc2 ϩ .
The expression profile of ⌬tsc1 and ⌬tsc2 cells showed extensive overlap, consistent with similar phenotypes of TSC1 and TSC2 mutations in humans, rodents, Drosophila, and S. pombe. In addition to the down-regulated amino acid permeases, two enzymes linked to arginine biosynthetic pathways were differentially expressed: arginase and 5-oxoprolinase. Two genes with homology to mammalian 5-oxoprolinase were down-regulated in both ⌬tsc1 and ⌬tsc2. 5-Oxoprolinase hydrolyzed pyroglutamic acid to glutamate and was down-regulated in some human tumors (34). Arginase was up-regulated in both ⌬tsc1 and ⌬tsc2 despite the low intracellular arginine levels. Arginase plays an important role in the production of ornithine (35), so the increase in arginase mRNA could be a response to the drop in ornithine levels from 15 nM in wild type to nearly undetectable levels in the ⌬tsc1 and ⌬tsc2 strains. Ornithine is the precursor of polyamines including spermidine. Spermidine is essential for growth and cell cycle progression in S. pombe (36). Polyamines are also critical to the growth and differentiation of mammalian cells and are elevated in many human cancers (37,38). Finally, arginase is important in mammalian cells because it competes with nitric-oxide synthetase for arginine, which is the substrate for both arginase and nitric-oxide synthetase. In mammalian cells, nitric oxide is a key second messenger regulating many processes including neuronal signaling (39). It will clearly be important to determine whether expression of permeases, arginase, or 5-oxoprolinase is regulated by mammalian TSC1 and TSC2.
Mice with conditional inactivation of Tsc1 in brain astrocytes develop seizures (40). Seizures are a major clinical problem in TSC, affecting 80% of patients, and are often refractory to treatment. Interestingly, the Tsc1 Ϫ/Ϫ astrocytes have decreased uptake of the excitatory neurotransmitter glutamate and decreased expression of two glutamate transporters (40). It is postulated that reduced astrocyte clearance of glutamate from the synaptic cleft slows the decay of excitatory stimuli, lowering the seizure threshold. If the decreased glutamate uptake is mechanistically related to the decreased amino acid uptake in S. pombe, the yeast model could provide a novel system for the study of epilepsy.
In conclusion, our data show for the first time that Tsc1 and Tsc2 regulate arginine uptake and arginine biosynthesis in S. pombe. Rescue of the arginine uptake defect by a dominant negative form of Rhb1 suggests that Rhb1 is downstream of Tsc2 in S. pombe as well as in other species. The complexity of the amino acid phenotype is suggestive of an intrinsic defect in amino acid sensing, involving amino acids and enzymes closely linked to ornithine and arginine. If similar pathways are affected in mammalian cells lacking TSC1 or TSC2, defects in polyamines and/or nitric oxide levels could be pathogenically linked to the clinical manifestations of TSC, including refractory seizures.