Isp7 Is a Novel Regulator of Amino Acid Uptake in the TOR Signaling Pathway

Abstract

TOR proteins reside in two distinct complexes, TOR complexes 1 and 2 (TORC1 and TORC2), that are central for the regulation of cellular growth, proliferation, and survival. TOR is also the target for the immunosuppressive and anticancer drug rapamycin. In Schizosaccharomyces pombe, disruption of the TSC complex, mutations in which can lead to the tuberous sclerosis syndrome in humans, results in a rapamycin-sensitive phenotype under poor nitrogen conditions. We show here that the sensitivity to rapamycin is mediated via inhibition of TORC1 and suppressed by overexpression of isp7⁺, a member of the family of 2-oxoglutarate-Fe(II)-dependent oxygenase genes. The transcript level of isp7⁺ is negatively regulated by TORC1 but positively regulated by TORC2. Yet we find extensive similarity between the transcriptome of cells disrupted for isp7⁺ and cells mutated in the catalytic subunit of TORC1. Moreover, Isp7 regulates amino acid permease expression in a fashion similar to that of TORC1 and opposite that of TORC2. Overexpression of isp7⁺ induces TORC1-dependent phosphorylation of ribosomal protein Rps6 while inhibiting TORC2-dependent phosphorylation and activation of the AGC-like kinase Gad8. Taken together, our findings suggest a central role for Isp7 in amino acid homeostasis and the presence of isp7⁺-dependent regulatory loops that affect both TORC1 and TORC2.

INTRODUCTION

TOR (target of rapamycin) is an atypical serine/threonine kinase that functions as a central regulator of growth (1, 2). TOR was originally identified in the budding yeast in a genetic screen for mutations that conferred resistance to the growth-inhibitory effect of the immunosuppressant and anticancer drug rapamycin (3). Rapamycin forms a complex with the highly conserved small protein FKBP12 and as part of such a complex can bind and inhibit TOR proteins. TOR proteins exist in two distinct complexes, TOR complex 1 (TORC1) and TORC2 (4, 5). In many different eukaryotes, TORC1 positively regulates cell growth in response to nutrients, growth factors, energy signals, and stress. TORC2 affects metabolism, cell survival, and proliferation, yet its cellular functions are less well understood than those of TORC1 (1, 2). The main target of rapamycin is TORC1, but TORC2 can also be inhibited by rapamycin following long exposures to the drug (6, 7). One of the main negative regulators of TOR signaling is the TSC1-TSC2 heterodimer. Mutations in TSC1 or TSC2 can lead to the tuberous sclerosis syndrome (8). The TSC complex converts the GTPase Rheb into its inactive form and thus prevents TORC1 activation (9–11). TORC1 is regulated by additional GTPase proteins called Rag that were found to mediate mostly amino acid signaling (12).

In the fission yeast, Schizosaccharomyces pombe, two TOR homologues exist, Tor1 and Tor2, which were named in the order of their discovery (13). Later, it was found that Tor1 is the catalytic subunit of TORC2 while Tor2 is the catalytic subunit of TORC1. TORC1 also contains the protein Mip1 (raptor in humans), and TORC2 also contains the proteins Ste20 (rictor in humans) and Sin1 (mSin1 in humans) (13–20). TORC1 is essential under normal growth conditions and plays major roles in the control of cellular growth, possibly in response to nitrogen availability (17, 20–25). Consistently, most of the genes that are upregulated in tor2^ts mutant cells are involved in the nitrogen starvation response (17). S. pombe contains homologues for the TSC1 and TSC2 genes, known as tsc1⁺ and tsc2⁺, which, similarly to human cells, negatively regulate TORC1 activity via inhibition of Rhb1 (Rheb in humans) (22, 26–28). Disruption of tsc1⁺ or tsc2⁺ leads to amino acid uptake defects and impaired sexual development or gene induction upon nitrogen starvation (29, 30). Recently, Rag homologues Gtr1 and Gtr2 were characterized in fission yeast. Gtr1 and Gtr2 function as a heterodimeric complex and were found to induce cellular growth and repress sexual differentiation by activating TORC1 in response to the presence of amino acids (23).

S. pombe TORC2 regulates cell survival under stress conditions and is required for starvation responses via the AGC protein kinase Gad8 (31). Disruption of TORC2 or gad8⁺ results in pleiotropic defects that include inability to initiate sexual development or acquire stationary-phase physiology, severe sensitivity to a variety of stresses, a delay in entrance into mitosis, decrease in amino acid uptake, sensitivity to DNA-damaging agents, and elongated telomeres (13, 15, 16, 18–20). Interestingly, TORC1 and TORC2 in S. pombe oppositely regulate starvation responses, including sexual development and transcription of nitrogen starvation-induced genes (25). Moreover, TORC1 and TORC2 play antagonistic roles with respect to mitosis (15) and longevity (32).

Rapamycin does not inhibit growth of wild-type S. pombe cells, indicating that under normal growth conditions the essential function of TORC1 is resistant to rapamycin. Nevertheless, rapamycin inhibits TORC1-dependent phosphorylation of the AGC kinase Sck1, Sck2, or Psk1 (33). A triple deletion mutant of these kinases is viable (33), indicating that additional TORC1 downstream effectors are yet to be discovered. More recently, it was shown that rapamycin can inhibit TORC1 in the presence of caffeine, which potentially lowers the kinase activity of TORC1 and induces a nitrogen starvation-like response (32, 34). Rapamycin inhibits sexual development and amino acid uptake (20, 24, 35) and following long exposure reduces the expression of nitrogen starvation-induced amino acid permeases (20). Under certain conditions, rapamycin can also induce sexual development (18), suggesting that rapamycin may involve inhibition of either TORC1 or TORC2, depending on the experimental conditions.

Previously, we showed that tsc1 or tsc2 mutant cells are highly sensitive to rapamycin on proline medium. No inhibition by rapamycin was observed when the nitrogen source was ammonium, which is normally used in the standard growth medium, or glutamate, which is considered a relatively good nitrogen source (25). The rapamycin-sensitive phenotype of tsc mutant cells is dependent on the presence of the S. pombe FKBP12 protein (25), indicating that TOR (either Tor1 or Tor2) is the target of rapamycin. Here we show that the sensitivity of tsc mutant cells to rapamycin is mediated by TORC1 and can be suppressed by overexpression of the 2-oxoglutarate-Fe(II)-dependent oxygenase, Isp7. We show that Isp7 is a novel regulator of amino acid uptake that acts via regulation of gene expression, both upstream and downstream of TOR signaling.

MATERIALS AND METHODS

Yeast techniques.

Table S4 in the supplemental material shows the list of S. pombe strains used in this work. Unless otherwise specified, S. pombe strains were grown at 30°C in Edinburgh minimal medium (EMM; 5 g/liter NH₄Cl) as described before (36). For proline medium, the ammonium chloride in the minimal medium was replaced with 10 mM proline. Rapamycin (R0395; Sigma) was used at a final concentration of 100 ng/ml unless otherwise specified. S. pombe cDNA library in pREP3X multicopy plasmid was used for genetic screens (37). The open reading frames (ORFs) were identified by comparison with the S. pombe gene database, GeneDB (http://www.genedb.org/genedb/pombe/).

Construction of tor1SE and tor2SE mutant strains.

A mutation at the FRB (FKBP12-rapamycin binding) domain of tor2⁺, converting serine to glutamic acid (S1837E), was created by site-directed mutagenesis using PCR overlap extension, as described previously (20). The resulting PCR fragment was cloned into the pREP81 plasmid. A plasmid carrying an equivalent mutation in the tor1⁺ gene, tor1SE, was previously described (13). The tor1SE and tor2SE mutations were integrated into their respective loci in the genome. First, a 1.5-kbp kanMX6 fragment was cloned into plasmids carrying tor1SE or tor2SE. Fragments containing the tor1SE or tor2SE and the kanMX6 cassette were then released from the plasmids by enzymatic restrictions and introduced into the genome by homologous recombination to replace the wild-type copies of the respective genes. The presence of the mutations was confirmed by sequencing.

Construction of the isp7-H276A plasmid.

Site-directed mutagenesis of pREP3X-isp7⁺ was carried out by PCR amplification with Accuzyme DNA Polymerase (Bioline), using the complementary primers 584 (5′-CGTCTAGGTGTTCAAGAGGCCACGGATGCTGATGCGCT-3′) and 585 (5′-AGCGCATCAGCATCCGTGGCCTCTTGAACACCTAGACG-3′), which contain the desired mutation (the underlined region). PCR products were digested with DpnI (NEB) to eliminate the methylated wild-type template and propagated in Escherichia coli XL1-Blue cells. Mutant clones were identified by DNA sequencing and checked for the absence of additional mutations in the isp7⁺ coding sequence. Their growth phenotype was tested by transformation into the tsc2 mutant strain.

Construction of isp7-HA.

The C terminus of Isp7 was tagged with triple hemagglutinin (HA). PCR was used to amplify the HA tag using the pFA6a-3HA-kanMX6 plasmid as a template (38) with primers 633 (5′-GAACCAATTGCTGTGGAAGACTTGTTACGCGATCATTTCCAAAACAGCTATACTTCACATACCACCTCATTGGAAGTTGCACGGATCCCCGGGTTAATTAA-3′) and primer 634 (5′-CAAGACAAATATACAATAATAATAGGACAACAAACAACAACGAGCAAGTCTAAATTGAAATTTTTTTTCCTCCAAGTTCAAGAATTCGAGCTCGTTTAAAC-3′), where underlining indicates the plasmid sequence. The PCR fragment was purified and transformed into wild-type strain 972. The DNA fragment was introduced into the genome by homologous recombination. Correct integration at the isp7⁺ loci was validated by PCR. The expression of tagged proteins was confirmed by Western blotting.

Construction of LacZ reporter gene.

The isp7⁺ promoter was fused into the promoterless β-galactosidase gene of the shuttle vector pSPE356 (39). The upstream sequence of isp7⁺ gene was amplified by PCR using primer 712 (5′-TATGGATCCTCAGTTTGGCATCTATAAAACAGGCA-3′) and primer 713 (5′-TATGTCGACTATGCAGAATGTGAATTAAGTAGAAAAGAAAAAAT-3′), which contained BamHI and SalI sites, respectively. The resulting PCR fragment containing the DNA sequence from positions −2550 to +1 of isp7⁺ was cloned into pSPE356.

Northern blotting.

Yeast RNA was extracted from logarithmic growing cells, and RNA was prepared using the hot phenol method and subjected to Northern blot analysis as described in reference 25. Gene-specific probes were labeled with [α-³²P]dCTP using the Random Primer DNA labeling kit (20-101-25A; Biological Industries).

Uptake assays.

We followed amino acid uptake assays as previously described (20). Briefly, cells were grown to mid-log phase in minimal or proline media. One-half milliliter of logarithmic cells was harvested and resuspended in 0.5 ml cold medium containing 100 μM arginine together with ³H-labeled arginine {2 μCi of [1-4,5-3H(N)]-arginine, 54.6 Ci/mmol; PerkinElmer} or 100 μM proline together with ³H-labeled proline {2 μCi of [1-4,5-3H(N)]-proline, 92.6 Ci/mmol; PerkinElmer}. Cells were incubated at 30°C, and samples were taken after 6 min and mixed with chilled medium containing 5 mM arginine or proline. Cells were then washed three times before being resuspended in water containing 0.5% SDS. ³H-labeled arginine or proline was measured by scintillation counting. Error bars represent standard deviations calculated from three independent cultures.

Microarray analysis.

RNA was prepared using the hot phenol method (25). RNA was hybridized to the Affymetrix Yeast Genome 2.0 microarray (http://media.affymetrix/support/technical/datasheets/yeast2_datasheet.pdf Affymetrix, Santa Clara, CA). Three independent biological replicates were performed. Microarray analysis was performed on CEL files using Partek Genomics Suite, version 6.5 (http://www.partek). Data were normalized and summarized with the robust multiaverage method, followed by analysis of variance (ANOVA). Gene expression data were sorted using cutoffs of P values of <0.05 under false discovery rate (FDR) control and a fold change cutoff of 1.5 or 2. Venny was used to cross between gene lists (http://bioinfogp.cnb.csic.es/tools/venny/index.html).

Western blotting.

Proteins were extracted with tricarboxylic acid (TCA) and resolved by SDS-PAGE using 12% acrylamide gels. Proteins were transferred to nitrocellulose membranes and blocked with 5% milk in Tris-buffered saline–Tween 20 (TBST) before immunoblotting. Detection was carried out using the ECL SuperSignal detection system (Thermo Scientific). Rps6 phosphorylation was detected using the anti-phospho(S/T)-Akt substrate (anti-PAS) antibody (Cell Signaling Technology), Rps6 by anti-Rps6 antibody (Abcam), tubulin by antitubulin antibody (Sigma), Isp7-HA by anti-HA antibody (Santa Cruz), Cdc2 by anti-PSTAIRE antibody (Santa Cruz), and actin by antiactin antibody (MP-Biomedicals). Antibodies against Gad8 phosphorylated at Ser546 and antibodies against the C terminus of the Gad8 were created using the peptides CRFANW-Ps-YQRPT and CKSDDINTIAPGSVIR, respectively (Bio Basic Canada Inc.).

β-Galactosidase assays.

Cells were grown to mid-log phase in minimal medium supplemented with adenine, leucine, and histidine before being harvested and subjected to β-galactosidase assays (40). Error bars represent standard deviations calculated from three independent cultures.

In vitro kinase assay.

As a substrate for the Gad8 kinase assay, the region of amino acid residues 291 (Gln) to 411 (Pro) of Fkh2 was expressed in Escherichia coli strain BL21 as a glutathione S-transferase (GST) fusion using the pGEX-4T1 expression vector and purified. Cells expressing Gad8-HA were grown in minimal or yeast extract (YE) medium and disrupted with glass beads in lysis buffer (20 mM Tris-HCl [pH 7.5], 0.5 mM EDTA, 1 mM dithiothreitol [DTT], 125 mM KAc, 0.5 mM EGTA, 0.1% Triton X-100, 12.5% glycerol). After centrifugation at 10,000 × g for 10 min at 4°C, supernatants were incubated overnight with anti-HA antibodies. Protein A and protein G mixtures were added, and immunoprecipitates were washed once with the lysis buffer, once with lysis buffer containing 0.5 M NaCl, and twice in buffer A (50 mM Tris-HCl [pH 7.5], 0.1 mM EGTA, 0.1% β-mercaptoethanol). The immunoprecipitations were incubated with 0.1 mg of GST-Fkh2 in kinase buffer (10 mM MgAc, 100 mM ATP, and phosphatase inhibitor cocktail; Sigma) for 10 min in 30°C. The reaction products were detected by Western blotting using anti-PAS antibody (Cell Signaling Technology). The level of Gad8 was detected by anti-HA antibody (Santa Cruz).

Microarray data accession number.

The microarray data determined in this study have been deposited in GEO under accession number GSE52759.

RESULTS

The sensitivity of tsc mutant cells to rapamycin is mediated by TORC1.

The growth of Δtsc1 or Δtsc2 mutant cells is strongly inhibited by rapamycin when the sole nitrogen source in the medium is proline but not ammonia (25) (Fig. 1A). We have previously shown that Δtsc1/2 mutant cells are unable to grow on proline medium in the absence of the tor1⁺ gene (25). Tor1 is found mainly as part of TORC2 (14). Deletion of the specific component of TORC2, sin1⁺, or the downstream kinase gene gad8⁺ also resulted in a similar growth inhibition in the genetic background of Δtsc1 or Δtsc2 (Fig. 1B). Thus, either TORC2-Gad8 or the TSC complex is required for growth on proline.

FIG 1.

TORC1 or TORC2 is essential for growth on proline in the absence of TSC. (A) tsc mutant cells are sensitive to rapamycin when the nitrogen source is proline. Serial dilutions of exponentially growing wild-type (WT) and Δtsc1 or Δtsc2 mutant cells were spotted on minimal (EMM) or proline medium, in the presence or absence of rapamycin (R). (B) TORC2-Gad8 is required for growth on proline in the absence of Tsc1 or Tsc2. Serial dilutions of the indicated strains were performed as described for panel A. (C) tor2SE but not tor1SE suppresses rapamycin sensitivity of tsc mutant cells. Serial dilutions of the indicated strains were performed as described for panel A.

In order to determine whether rapamycin inhibits the growth of tsc mutant cells via inhibition of Tor1 (TORC2) or Tor2 (TORC1), we used rapamycin-binding defective alleles of Tor1 (20) or Tor2 (see Materials and Methods). These mutated alleles contain a substitution of the conserved serine within the FRB domain of the TOR proteins. The tor1S1834E (tor1SE) or tor2S1837E (tor2SE) alleles were integrated into their respective genomic loci, replacing the wild-type tor1⁺ and tor2⁺ genes. In Δtsc1 or Δtsc2 mutant cells, only the tor2SE allele, but not tor1SE, conferred rapamycin resistance (Fig. 1C). This result indicates that the target for rapamycin is TORC1 and not TORC2. Thus, although TORC2 is required for growth on proline in the absence of tsc1/2, it is the inhibition of TORC1 that is critical for rapamycin sensitivity.

Overexpression of isp7⁺ rescues rapamycin sensitivity of tsc mutant cells.

We transformed a cDNA library into Δtsc2 cells and looked for high-copy-number suppressors of the rapamycin-sensitive phenotype on proline plates. We isolated five suppressors, including tsc2⁺:aap1⁺ (SPBC1652.02), encoding a predicted amino acid transporter; SPAC29B12.11c, a human WW domain binding protein 2 ortholog; SPAC1296.01c, a predicted phosphoacetylglucosamine mutase; and isp7⁺, encoding a predicted 2-oxoglutarate-Fe(II)-dependent oxygenase.

Overexpression of isp7⁺ suppressed rapamycin sensitivity either in Δtsc2 or in Δtsc1 mutant cells (Fig. 2A), indicating that the suppression activity is not specific for Δtsc2 but compensates for loss of activity of the TSC complex. The isp7⁺ gene was originally isolated during a screen for genes that show increased transcription under nitrogen starvation conditions (41). Δisp7 mutant cells are viable but lose viability following heat shock during the stationary phase (42). We found that while overexpression of isp7⁺ suppressed the sensitivity of Δtsc1/2 mutant cells to rapamycin, deletion of isp7⁺ in the background of Δtsc1 or Δtsc2 cells resulted in double mutant cells that were unable to grow on proline plates (Fig. 2B). These findings suggest that Isp7 and TSC work on separate, but possibly converging, pathways that affect growth on proline.

FIG 2.

The nitrogen starvation response gene isp7⁺ suppresses rapamycin sensitivity of tsc mutant cells. (A) Overexpression of isp7⁺ rescues rapamycin sensitivity of tsc mutant cells. Cells lacking either tsc1⁺ or tsc2⁺ were transformed with pREP3X-isp7⁺ and streaked on minimal medium (EMM) or proline medium with or without rapamycin (R). Δtsc1 or Δtsc2 mutant cells containing vector only were used as a control. (B) Synthetic lethality between Δtsc1/2 and Δisp7. Serial dilutions of the indicated strains on minimal (EMM) or proline medium. (C) Isp7 is preferentially transcribed upon shift to proline. Wild-type (WT) or Δtsc2 cells were grown to mid-log phase in minimal medium and shifted into proline or minimal medium with or without rapamycin. Samples were taken after 15 and 60 min. Total RNA was extracted and analyzed by Northern blotting using a probe against isp7⁺. rRNA was used as a loading control. (D) The protein level of Isp7 is transiently induced in response to poor nitrogen source. Western analysis of Isp7-HA in wild-type cells in minimal medium and after a shift to proline. Samples were taken after 15 and 60 min. Cdc2 was used as a loading control. (E) Northern blot analysis of isp7⁺ in wild-type versus Δtor1 cells. Cells were treated as described for panel C. (F) Expression of isp7-LacZ fusion constructs in TORC2 mutant cells. Wild-type and cells disrupted for specific subunits of TORC2, tor1⁺, sin1⁺, or ste20⁺, containing LacZ driven by the isp7⁺ promoter were grown to mid-log phase. β-Galactosidase activity was measured from protein extracts.

Similar to depletion of nitrogen, isp7⁺ is also strongly upregulated upon a shift to growth medium that contains proline as the nitrogen source, either in wild-type or in Δtsc2 genetic backgrounds (Fig. 2C). The addition of rapamycin to Δtsc2 mutant cells shifted into proline medium did not affect the transcript level of isp7⁺ (Fig. 2C). Thus, the lethality of Δtsc2 on proline plates in the presence of rapamycin is unlikely the result of a reduction of isp7⁺ transcript under such conditions. The transcript of isp7⁺ was previously shown to be transiently induced upon nitrogen starvation (43). We examined the possibility that isp7⁺ is also transiently induced in response to a shift to proline. Although we did not observe a significant decrease in the transcript level of isp7⁺ following 1 hour of incubation in proline (Fig. 2C), we observed that the level of the Isp7 protein was upregulated following a 15-min shift to proline but decreased after 1 h of incubation in proline (Fig. 2D). This transient characteristic of expression of Isp7 may be relevant for the functional analysis of Isp7 as described below.

Genome-wide transcriptional analyses suggested that isp7⁺ is upregulated in tor2 temperature-sensitive (tor2-ts) mutant cells (17, 44) but downregulated in Δtor1 cells (19). Using Northern blot analysis, we confirmed that the transcript of isp7⁺ is downregulated in Δtor1 cells (Fig. 2E). In addition, we cloned the bacterial lacZ gene under the promoter of isp7⁺ (see Materials and Methods) and found a sharp decrease in lacZ-dependent activity in cells deleted either for tor1⁺ or any of the specific components of TORC2, sin1⁺ and ste20⁺ (Fig. 2F). In addition, we confirmed that the transcript level of isp7⁺ is upregulated in tor2-ts (Fig. 3A, top panel). The opposite regulation of isp7⁺ by TORC1 and TORC2 is consistent with the opposite effects of TORC1 and TORC2 on many nitrogen starvation-induced genes (25).

FIG 3.

Isp7 negatively regulates transcription of the nitrogen starvation-induced amino acid permeases per1⁺, put4⁺, and isp5⁺ but positively regulates cat1⁺. (A to D) Northern blot analysis of the transcription levels of isp7⁺, per1⁺, put4⁺, isp5⁺, and cat1⁺. rRNA was used as a loading control in all samples. (A and C) The indicated strains were grown to mid-log phase at 25°C (0 h), shifted to 32°C, and incubated for 4 h (4 h) (restrictive conditions) before extraction of RNA. For the blot presented in panel C, RNA levels were quantified relative to the rRNA using Gelquant software. All the indicated strains were grown on minimal medium.

Isp7 is a negative regulator of nitrogen starvation-induced amino acid permeases but a positive regulator of the basic amino acid permease gene cat1⁺.

Genome-wide transcriptional analysis suggested that S. pombe TORC1 negatively regulates nitrogen starvation-induced permease genes (17). This includes per1⁺ and put4⁺, which are related to the Saccharomyces cerevisiae general amino acid permease gene GAP1 and the proline permease gene PUT4, respectively (45). We confirmed by Northern blotting that per1⁺ and put4⁺ are induced upon inactivation of Tor2 (TORC1) (Fig. 3A). As described before (30), loss of the TSC complex results in a slight downregulation of per1⁺ and put4⁺ (Fig. 3B, lane 2). The negative regulation of put4⁺ and per1⁺ by S. pombe TORC1 is reminiscent of the negative regulation of PUT4 and GAP1 by S. cerevisiae TORC1 (45) and suggests that TORC1 has a conserved role in negatively regulating nitrogen starvation-induced amino acid permeases.

Unexpectedly, although the transcript level of isp7⁺ is negatively regulated by TORC1 (Fig. 3A), we found a strong induction of the transcripts of per1⁺ or put4⁺ in Δisp7 cells (Fig. 3B). The isp5⁺ gene, another amino acid permease gene that is induced upon inactivation of tor2⁺ (17, 44) or nitrogen starvation (43), is also strongly induced in Δisp7 cells (Fig. 3B). Thus, although the isp7⁺ transcript is negatively regulated by TORC1, Isp7, like TORC1, is a negative regulator of nitrogen starvation-induced amino acid permeases.

In contrast with put4⁺ and per1⁺, we found that cat1⁺, the main transporter for arginine (46), is strongly downregulated upon inactivation of Tor2 (TORC1) (Fig. 3A). The cat1⁺ transcript is slightly downregulated under poor nitrogen conditions (see Fig. S1 in the supplemental material). Opposite effects of TORC1 on general amino acid permeases versus specific amino acid permeases have also been described in S. cerevisiae: while GAP1 and PUT4 are negatively regulated by TORC1, the tryptophan-specific permease gene TAT2 is positively regulated by TORC1 (47). Similar to disruption of TORC1, Δisp7 also resulted in downregulation of the transcript level of cat1⁺ (Fig. 3B). Thus, TORC1 and Isp7 are similarly involved in positive and negative regulation of transcription of amino acid permeases depending on the nature of the amino acid permease.

In order to examine the genetic relationships between tor2⁺ and isp7⁺, we constructed Δisp7 tor2-51 double mutant strains. We observed no additive effects in the double mutant cells with respect to transcriptional regulation (Fig. 3C), supporting a model in which Isp7 and Tor2 (TORC1) act on the same pathway. The level of cat1⁺ in Δisp7, tor2-51, or Δisp7 tor2-51 cells is below detectable levels and thus cannot be used for epistasis analysis (Fig. 3C).

TORC2 acts oppositely to TORC1 in regulating nitrogen starvation-induced amino acid permeases. Accordingly, per1⁺, put4⁺, and isp5⁺ are downregulated in Δtor1 mutant cells (20) (Fig. 3D, lane 2). We found that the transcription of cat1⁺ is induced in Δtor1 cells (Fig. 3D, lane 2). Thus, TORC1 and TORC2 also regulate cat1⁺ in opposite directions. Δisp7 Δtor1 double mutant cells show expression levels of per1⁺ or cat1⁺ similar to those of the single Δtor1 mutant cells (Fig. 3D, compare lanes 2, 3, and 5), indicating that cells deficient in isp7⁺ require Tor1 (TORC2) activity to affect gene transcription and suggesting that isp7⁺ may lie upstream of TORC2 and negatively regulates its activity. In contrast, the upregulation of per1⁺ or downregulation of cat1⁺ in Δisp7 cells does not require tsc2⁺ (Fig. 3D, lane 4), suggesting that Isp7 does not regulate transcription of amino acid permeases in a TSC-dependent manner.

We conclude that although the expression of isp7⁺ is negatively regulated by TORC1 and positively regulated by TORC2, isp7⁺ regulates the expression of amino acid permeases in a fashion similar to that of TORC1 and opposite to that of TORC2. These unexpected findings fit a model in which Isp7 lies both upstream and downstream of TOR-dependent signaling (see Fig. 10). Indeed, our data presented below support such a feedback loop model.

FIG 10.

Working model. TORC1 and TORC2 regulate amino acid permease transcription and amino acid uptake via isp7⁺-dependent regulatory loops. TORC1 and TORC2 oppositely regulate the transcription of isp7⁺. Overexpression of isp7⁺ induces phosphorylation of Rps6 in a TORC1-dependent manner but inhibits TORC2-dependent phosphorylation and activation of Gad8.

Isp7 is a positive regulator of arginine uptake.

The cat1⁺ gene is the main transporter for arginine (46). In accord with our findings showing strong downregulation of the cat1⁺ transcript in Δisp7 cells (Fig. 3B), these mutant cells are resistant to canavanine or thialysine, the toxic analogs of the basic amino acid arginine or lysine, respectively (Fig. 4A). Δtsc1/2 mutant cells are also resistant to canavanine or thialysine (30). Δtsc2 Δisp7 double mutant cells are further resistant to the toxic analogs compared to any of the single parent strains (Fig. 4A). Consistently, Δtsc2 Δisp7 double mutant cells show additive reduction in the uptake of radiolabeled arginine (Fig. 4B) and synthetic lethality when grown on arginine as the sole nitrogen source (Fig. 4C). These findings strongly support a model in which Isp7 and TSC work independently. In contrast, Δtor1 mutant cells and the double mutant Δtor1 Δisp7 cells are sensitive to canavanine or thialysine (Fig. 4A), in accord with the expression level of cat1⁺ and consistent with the possibility that Tor1 lies downstream from Isp7. The rhb1GS allele was isolated as a partial-loss-of-function allele in a screen for suppressors of amino acid uptake defect of Δtsc1 mutant cells (28). The rhb1 GS allele reversed the canavanine resistance of Δtsc1 (30) (Fig. 4D); however, rhb1GS did not restore canavanine sensitivity to Δisp7 mutant cells (Fig. 4D). Thus, unlike TSC, Isp7 does not act via Rhb1.

FIG 4.

Isp7 positively regulates arginine uptake. (A) Serial dilutions of the indicated strains on minimal medium (EMM) plates with or without 60 μg/ml canavanine or 200 μg/ml thialysine. (B) Arginine uptake assays. The indicated strains were grown to mid-log phase on minimal medium. (C) Serial dilutions of the indicated strains on EMM or plates containing 1.14 mM arginine as the sole nitrogen source. (D) The mutant allele rhb1 GS does not restore canavanine sensitivity in Δisp7 cells. Strains were spotted on minimal medium with or without 60 μg/ml canavanine. (E and F) Overexpression of isp7⁺ compensates for the lack of tsc2⁺. Cells lacking tsc2⁺ were transformed with the isp7⁺ plasmid. Overexpression of isp7⁺ induced sensitivity to canavanine (60 μg/ml) (E) and restored uptake of radioactively labeled arginine (F).

While Δisp7 and Δtsc1/2 showed additive effects with respect to arginine uptake, overexpression of Isp7 reversed the canavanine resistance of Δtsc2 mutant cells (Fig. 4E) and induced arginine uptake in Δtsc2 mutant cells (Fig. 4F). Thus, consistent with the isolation of isp7⁺ as a multicopy suppressor of the rapamycin sensitivity of tsc1/2 mutant cells, overexpression of isp7⁺ can compensate for loss of TSC activity in arginine uptake.

Isp7 is a negative regulator of proline uptake.

The isp7⁺ gene is expressed at basal levels during growth in the presence of a good nitrogen source (NH₄Cl; EMM) and is induced in response to a shift to proline medium (Fig. 2C). Yet, our Northern blot analyses indicate that isp7⁺ is a negative regulator of nitrogen starvation-induced amino acid permeases. Since such a mode of regulation is unexpected, we examined uptake of radiolabeled proline in Δisp7 mutant cells. In agreement with our transcriptional analysis, Δisp7 cells showed a strong upregulation of proline uptake (Fig. 5A) while overexpression of isp7⁺ caused a decrease in proline uptake (Fig. 5B, compare columns 1 and 2). Thus, although isp7⁺ is induced upon a shift to proline, Isp7 is a negative regulator of proline uptake. We speculate that isp7⁺ may be transiently induced upon shift to proline and subsequent reduction allows the upregulation of amino acid permeases required for proline uptake. Indeed, the transient increase in the level of Isp7 upon a shift to proline may support such a mode of action (Fig. 2D).

FIG 5.

Isp7 negatively regulates proline uptake. (A to D) Proline or arginine uptake in cells grown in the presence of NH4⁺ (A, C, and D) or in the presence of proline (B). When rapamycin was used (B), cells were shifted to proline with or without rapamycin (R) for 1 h. Uptake of arginine or proline was measured at 25°C and 4 h after a shift to 32°C (restrictive temperature) (C and D).

The uptake of proline is very low in Δtsc2 or Δtor1 mutant cells (Fig. 5A), in agreement with the low transcription levels of general amino acid permeases (Fig. 3B and D). Rapamycin also reduced uptake of proline (Fig. 5B, compare columns 1 and 5). Since isp7⁺ was identified as a multicopy suppressor of rapamycin sensitivity of tsc mutant cells on proline, we compared the level of proline uptake in Δtsc2 and Δtsc2 cells overexpressing isp7⁺ in the presence of rapamycin (Fig. 5B, columns 4 and 8, respectively). However, under such conditions proline uptake is below detectable levels. Thus, our amino acid uptake assays cannot account for the suppression activity of isp7⁺ in the presence of rapamycin in proline.

We observed a decrease in arginine uptake in TORC1 mutant cells (tor2-51) (Fig. 5C), which correlates with the decrease in cat1⁺ level (Fig. 3A) and with previous observations that demonstrated canavanine resistance of tor2 temperature-sensitive alleles (48). A slight additive defect in the uptake of arginine is observed in Δisp7 tor2-51 mutant cells, under either semipermissive or restrictive conditions (Fig. 5C), indicating that Isp7 acts at least partially independently of Tor2 (possibly via Tor1). Somewhat surprisingly, although the transcript level of per1⁺ and put4⁺ is upregulated in tor2-51 mutant cells, we did not detect much effect on proline uptake in single tor2-51 mutant cells (Fig. 5D, lanes 1, 3, 5, and 7). Moreover, the dramatic increase in proline uptake in Δisp7 cells is diminished by tor2-51 (Fig. 5D, columns 6 and 8), suggesting that TORC1 plays a positive role in proline uptake. TORC1 may regulate proline uptake via a transcription-independent mechanism, for example, via regulation of localization of the permeases. Interestingly, deletion of the TSC complex results in a decrease in proline uptake (Fig. 5A) and arginine uptake (Fig. 4B), suggesting that TSC does not control amino acid uptake via inhibition of TORC1.

We conclude that Isp7 is a major regulator of amino acid uptake. It positively regulates arginine uptake while negatively regulating proline uptake. This mode of action is opposite to that of TORC2 and is consistent with our transcriptional analyses. TORC1 and Isp7 have similar effects on transcriptional regulation of amino acid permeases; however, TORC1 is required for the uptake of either arginine or proline, possibly due to posttranscriptional effects.

Extensive overlap between genes upregulated by loss of Isp7 or loss of TORC1.

To further assess the cellular roles of Isp7, we carried out a genome-wide expression profiling of Δisp7 cells. This analysis revealed 280 and 253 genes that are at least 1.5-fold upregulated or downregulated, respectively (see Data Set S1 in the supplemental material). Using gene ontology (GO) analysis for the top differentially expressed genes, we found a significant enrichment for genes involved in stress response, transport, or oxidoreductase activity (see Table S1 in the supplemental material). According to the gene ontology analysis, there are about 30 predicted amino acid permeases in the fission yeast genome. Half of these (49) are affected by Δisp7, either positively (3) or negatively (50) (see Tables S2 and S3 in the supplemental material). Interestingly, the amino acid permease aap1⁺, which we found as a multicopy suppressor of the sensitivity to rapamycin in tsc mutant cells, is downregulated in Δisp7 cells.

Significantly, the transcriptome of Δisp7 mutant cells strongly overlaps with the transcriptome of the temperature-sensitive allele tor2-ts6 (17) or starvation-induced genes (43) (Fig. 6). About half of the genes that are upregulated in tor2-ts6 (17) are also upregulated in Δisp7 mutant cells (Fig. 6A). Also, most of the transporters and permeases that are upregulated in tor2-ts6 are also upregulated in Δisp7 mutant cells (see Table S2 in the supplemental material). Since overexpression of isp7⁺ suppresses a TORC1-dependent rapamycin-sensitive phenotype, we speculate that overexpression of isp7⁺ may compensate for loss of a TORC1-dependent function by inducing a transcriptional program that is normally maintained by fully active TORC1.

FIG 6.

Extensive overlap between genes upregulated by loss of Isp7, loss of Tor2, or nitrogen starvation. Venn diagrams presenting overlaps between genes that are upregulated at least 1.5-fold in Δisp7 cells and genes upregulated in tor2-ts6 (A) or under nitrogen starvation (−N) (B).

Isp7 is a positive regulator of TORC1 kinase activity toward Rps6.

Rps6 in fission yeast, encoded by two genes, rps601⁺ and rps602⁺, is phosphorylated in a Tor2 (TORC1)-dependent manner (13, 15, 16, 18–20, 33, 51). We used an anti-phospho(S/T)-Akt substrate (anti-PAS) antibody to detect the phosphorylated form of Rps6 protein. As previously shown, we found that Rps6 becomes dephosphorylated in response to nitrogen starvation (−N) (Fig. 7A), consistent with the idea that TORC1 becomes inactive under such conditions. Similar to the effect of nitrogen starvation, we found that the phosphorylation of Rps6 is reduced in response to a shift to proline (Fig. 7A), suggesting that a low quality of the nitrogen source also results in a transient or partial inactivation of TORC1. Strikingly, overexpression of isp7⁺ induced the phosphorylation of Rps6, in particular in no-nitrogen or proline medium (Fig. 7A). Overexpression of isp7⁺ also induced the phosphorylation of Rps6 in Δtsc2 mutant cells (Fig. 7B), indicating that the effect of isp7⁺ on Rps6 phosphorylation is TSC independent. While overexpression of isp7⁺ induced Rps6 phosphorylation, deletion of isp7⁺ resulted in a slight reduction in the phosphorylation of Rps6 (Fig. 7C), suggesting that Isp7 acts in redundancy with other proteins.

FIG 7.

Overexpression of isp7⁺ induces Rps6 phosphorylation toward TORC1. Wild-type cells (A) or cells lacking tsc2⁺ (B) transformed with the isp7⁺ plasmid or vector only were grown to mid-log phase in minimal medium (M) and shifted into medium without nitrogen (−N) or with proline instead of ammonia (P). Samples were taken after 15 or 30 min. (C) Wild-type cells or cells lacking isp7⁺ were grown to mid-log phase in minimal medium (M) and shifted to proline for 15 min (P). (D) Wild-type cells or cells lacking tsc2⁺ transformed with the isp7⁺ plasmid were grown to mid-log phase in minimal medium (M) and shifted to proline for 30 min (P). Cells were collected 30 min following addition of rapamycin (R). (E) Wild-type, tor1SE, or tor2SE cells transformed with isp7⁺ were treated as above. Rapamycin was used at a final concentration of 200 ng/ml. The double mutant strain Δrps601 Δrps602SSAA (rps601/2) was used as a negative control. Phosphorylation of Rps6 (Rps6-P) was detected by immunoblotting with the anti-PAS antibody. Tubulin and Rps6 were used as loading controls.

The addition of rapamycin strongly reduced phosphorylation of Rps6, an effect that is reversed by introducing a tor2 rapamycin-binding-defective allele (51). We found that addition of rapamycin abolished Rps6 phosphorylation in the presence of overexpression of isp7⁺, either in wild-type or Δtsc2 background (Fig. 7D). Introduction of the tor2SE allele, which is defective in rapamycin binding, restored phosphorylation of Rps6 in the presence of overexpression of isp7⁺ (Fig. 7E). Recently, TORC2 has been suggested to affect Rps6 phosphorylation (52). However, introduction of the tor1SE allele did not restore Rps6 phosphorylation (Fig. 7E). Thus, overexpression of isp7⁺ induced Rps6 phosphorylation in a TORC1-dependent manner (Fig. 7E).

Since isp7⁺-induced phosphorylation of Rps6 is rapamycin sensitive, it is unlikely that Rps6 phosphorylation plays a role in the suppression activity of isp7⁺ in tsc mutant cells, as further discussed below.

Isp7 does not require the Gtr1/2 complex to induce phosphorylation of Rps6.

The Gtr1/2 complex in S. pombe was recently suggested to induce TORC1, in a mechanism conserved in evolution (23). However, the relationship between Gtr1/2 and TORC1 is complex. Thus, for example tor2-ts cells are canavanine resistant (48), while Δgtr1 and Δgtr2 cells are canavanine sensitive (53). Recently, it was suggested that Tor2 has opposing effects on Cat1 activity; thus, while the transcript of cat1⁺ is positively regulated by Tor2, the localization of Cat1 at the plasma membrane is negatively regulated through endocytosis (53). Cells disrupted for isp7⁺ are resistant to canavanine (Fig. 8A). The double mutant Δisp7 Δgtr1 or Δisp7 Δgtr2 cells are sensitive to canavanine, similar to single mutant Δgtr1 or Δgtr2 cells (Fig. 8A). These results suggest that Isp7 may regulate arginine uptake via the Gtr1/2 complex, or via a parallel pathway. In this respect, it is interesting that while Δtsc2 cells show synthetic lethality with either Δisp7 or Δgtr1/2 on proline plates (Fig. 2B and 8A), the double mutant Δisp7 Δgtr1 or Δisp7 Δgtr2 cells grew well on proline, supporting a model in which Isp7 and the Gtr1/2 complex work, at least in part, on the same pathway.

FIG 8.

Effects of the Gtr1/2 complex or the Isp7 oxygenase domain on Isp7-dependent activities. (A) Genetic interactions between Δisp7 and Δgtr1/2. Serial dilutions of the indicated strains on rich (YE), minimal (EMM), or proline medium with or without 60 μg/ml canavanine. (B) Overexpression of isp7⁺ induced Rps6 phosphorylation independent of Gtr1/2. Wild-type cells or cells lacking gtr1⁺ transformed with the isp7⁺ plasmid or vector only were grown to mid-log phase in minimal medium (M) and shifted into proline medium (P) for 15 min. Tubulin and Rps6 were used as loading controls. (C) The oxygenase domain of isp7⁺ is required to induce rapamycin resistance. Cells lacking tsc2⁺ were transformed with multicopy plasmids containing tsc2⁺, isp7⁺, isp7^H276A, or vector only. Serial dilutions were performed on minimal plates (EMM) with or without 60 μg/ml canavanine or proline plates with rapamycin (R). (D) Overexpression of isp7^H276A induces Rps6 phosphorylation. Cells lacking tsc2⁺ transformed with plasmids containing isp7⁺, isp7^H276A, or vector only were grown to mid-log phase in minimal medium (M) and shifted into proline medium (P) for 15 min. The double mutant Δrps601 Δrps602SSAA (rps601/2) strain was used as a negative control. Tubulin was used as a loading control.

To further explore the relationship between Isp7 and Gtr1/2, we asked whether the induction of Rps6 phosphorylation on proline medium requires Gtr1/2. Previously, it was shown that Rps6 phosphorylation is slightly reduced in Δgtr1 or Δgtr2 mutant cells (23). However, overexpression of isp7⁺ induced Rps6 phosphorylation in Δgtr1 mutant cells in a fashion similar to that of wild-type cells (Fig. 8B). Thus, while arginine uptake may be regulated by Isp7 via Gtr1/2, overexpression of isp7⁺ induces Rps6 phosphorylation in a Gtr1/2-independent mechanism.

The oxygenase domain of Isp7 is required to induce rapamycin resistance in tsc mutant cells.

Crystallographic studies of the family of 2-oxoglutarate-Fe(II)-dependent oxygenases revealed a double-stranded β-helix fold that supports a highly conserved Fe(II)-binding motif, H-X-E/D…H (54). This motif is also conserved in Isp7: His276-Thr277-Asp278…His334. It was previously shown that substitution of the first histidine in the consensus motif to alanine disrupted the enzymatic activity of Ofd2, another member of this family of oxoglutarate-Fe(II)-dependent oxygenases in fission yeast (55).

We introduced an equivalent mutation into Isp7, H276A. The mutated isp7^H276A gene was expressed on a multicopy plasmid in tsc2 mutant cells, but unlike isp7⁺, it failed to suppress rapamycin sensitivity on proline plates (Fig. 8C). In contrast, similar to isp7⁺, isp7^H276A reversed canavanine resistance of tsc2 mutant cells (Fig. 8C) and induced Rps6 phosphorylation following a shift to proline (Fig. 8D). These results suggest that the oxygenase domain of Isp7 is required to induce rapamycin resistance but not for arginine uptake or induction of Rps6 phosphorylation. Significantly, these findings fit our suggestion that the function of isp7⁺ that is relevant for rapamycin resistance is distinct from its function of inducing Rps6 phosphorylation.

isp7⁺ is inhibitory to TORC2-Gad8 signaling.

TORC2 positively regulates isp7⁺ at the transcriptional level, but the effect of isp7⁺ on amino acid permease gene transcription fits a model in which isp7⁺ inhibits TORC2. To examine the possibility that Isp7 affects TORC2-Gad8 activity, we developed a simple, nonradioactive in vitro kinase assay for Gad8. The Fkh2 protein was previously used in radioactive Gad8 in vitro kinase assays (31, 56). We used a fragment of Fkh2 (Gln291-Pro370) that contains three potential AKT (Gad8) phosphorylation consensus sites as the substrates. We immunopurified Gad8-HA expressed from its genomic locus under its own promoter and detected phosphorylation of the Fkh2 peptide using antibodies that recognize only the phosphorylated form of the AKT consensus sequence (anti-PAS). We observed a time course-dependent phosphorylation of the Fkh2 peptide in cells expressing Gad8-HA but not in Δtor1 (disruption of TORC2) cells or in cells expressing a kinase-dead Gad8-HA allele, Gad8^K259D (31) (Fig. 9A). Significantly, overexpression of isp7⁺ dramatically reduced Gad8-dependent kinase activity (Fig. 9B). Δisp7 cells showed a mild but repeatable induction of Gad8-dependent kinase activity (Fig. 9B). These results support a mechanism in which Isp7 inhibits the kinase activity of Gad8.

FIG 9.

Isp7 inhibits TORC2-Gad8 activity in vitro and in vivo. (A) Gad8 shows a time-dependent, Tor1-dependent, and kinase active site-dependent phosphorylation toward Fkh2. Protein extracts were immunoprecipitated with anti-HA antibody, and the kinase assay was performed using Fkh2-GST as the substrate. The kinase reaction was stopped after 4 or 15 min of incubation. Phosphorylation of Fkh2-GST was detected with anti-PAS antibody. Gad8-KD (Gad8-K259D) is a kinase-dead allele. A nontagged strain was used as a negative control (No tag). (B) Gad8-dependent kinase activity is affected by the levels of isp7⁺. Gad8 was immunoprecipitated from cells lacking isp7⁺ (Δisp7) or overexpressing isp7⁺ (WT+isp7), and kinase assays were performed as described for panel A. (C) TORC2-dependent phosphorylation of Gad8-Ser456 is reduced by overexpression of isp7⁺. Wild-type, Δisp7, Δtor1, or Δgad8 cells or cells overexpressing isp7⁺ (WT+isp7⁺) or an empty vector (WT+Vector) were grown to mid-log phase, and their crude cell lysates were analyzed by immunoblotting with anti-phospho-Ser546 and anti-Gad8 antibodies. Actin was used as a loading control.

We also raised antibodies that recognize the TORC2-dependent phosphorylation of Ser546 within the hydrophobic motif of the Gad8 kinase. This TORC2-dependent phosphorylation is essential for activation of Gad8 (31, 57). Phosphorylation of Gad8 was significantly reduced in cells overexpressing isp7⁺ (Fig. 9C), consistent with the reduction in kinase activity of Gad8 (Fig. 9B). We did not observe an increase in the phosphorylation levels of Gad8 in Δisp7 cells (Fig. 9C). This is in contrast to the elevated levels of Gad8 kinase activity in these cells. It is possible that our in vitro kinase assay is more sensitive for monitoring changes in Gad8 activity. Also, it is possible that another protein acts in redundancy with isp7⁺ and thus that the effects of overexpression of isp7⁺ are more pronounced than with the deletion of the gene.

DISCUSSION

Rapamycin does not inhibit growth of wild-type cells in S. pombe (35). More recently, it was shown that the rapamycin-FKBP12 complex can inhibit TORC1 kinase activity in vitro (32, 34) or TORC1-dependent phosphorylation in vivo (33, 51). Here we show that S. pombe cells become sensitive to rapamycin in tsc1 or tsc2 mutant cells due to inhibition of TORC1. Accordingly, only a mutation in the catalytic subunit of TORC1 that is defective in rapamycin binding (but not the equivalent mutation in the catalytic subunit of TORC2) can confer resistance to rapamycin. We show that tsc mutant cells are incapable of growing on proline medium either in the presence of rapamycin or in the absence of TORC2-Gad8. Thus, the two TOR complexes are involved in regulating growth on proline in the absence of the TSC complex. The sensitivity to rapamycin of cells disrupted for the TSC complex presents an apparent paradox. As a negative regulator of TORC1, deletion of TSC is expected to result in high activity of TORC1 (17, 27, 50) and thus relative resistance to rapamycin. Our findings suggest that rapamycin inhibits a specific function of TORC1 that becomes crucial in the absence of TSC. Intriguingly, mutations in TSC1 render tumors in cancer patients more responsive to inhibition by everolimus (rapamycin) (58). Thus, deregulation of TORC1 by loss of the inhibitory effect of the TSC complex may lead to rapamycin sensitivity through a mechanism conserved in evolution. The details of the molecular mechanism are yet to be discovered.

The sensitivity of tsc mutant cells to rapamycin is suppressed by overexpression of isp7⁺, a putative 2-oxoglutarate (2OG)-Fe(II)-dependent oxygenase gene. Isp7 was originally isolated during a screen for genes that are preferentially transcribed under nitrogen starvation conditions (41). 2OG-Fe(II)-dependent oxygenases are widespread in eukaryotes and bacteria. They catalyze a variety of reactions, usually by oxidation of an organic substrate, and participate in a variety of cellular processes, including fatty acid metabolism, hypoxic sensing, and histone and nucleic acid demethylations (59). Previous sequence analysis classified isp7⁺ into a family called small-molecule dioxygenase genes (54). Here, we show that the conserved histidine His276, within the oxygenase domain, is required for the suppression activity of isp7⁺ in the presence of rapamycin. In contrast, Isp7-dependent arginine uptake or induction of Rps6 phosphorylation does not require His276, suggesting that the oxygenase domain is not required for these activities. Except for the oxygenase domain, Isp7 contains a highly conserved nonheme dioxygenase domain (DIOX_N) at the N terminus. Further work will determine the contribution of each domain to the cellular activities of Isp7.

We speculate that isp7⁺ suppresses the rapamycin sensitivity observed in tsc1/2 mutant cells in a mechanism that involves TORC1 activation but is independent of Rps6 phosphorylation. Consistent with this suggestion, disruption of isp7⁺ or disruption of TORC1 results in similar transcriptional profiles. Overexpression of isp7⁺ results in induction of TORC1-dependent Rps6 phosphorylation, suggesting that Isp7 is a positive regulator of TORC1. However, since Rps6 is dephosphorylated by rapamycin, the suppression activity of Isp7 is independent of phosphorylation of Rps6. These findings may suggest that the overexpression of isp7⁺ activates TORC1 toward substrates that are relatively resistant to rapamycin. Indeed, it was suggested that in human cells rapamycin inhibits the phosphorylation of certain substrates of TORC1 while leaving the phosphorylation of other substrates far less sensitive to rapamycin (49).

We identify Isp7 as a novel regulator of amino acid uptake via regulation of amino acid permease gene transcription, in a mechanism that involves both TORC1 and TORC2 (Fig. 10). Isp7 negatively regulates the transcription of nitrogen starvation-induced amino acid permease genes per1⁺, put4⁺, and isp5⁺ and the uptake of proline. In contrast, Isp7 positively regulates cat1⁺ and arginine uptake. At the level of transcription, Isp7 regulates amino acid permeases in a fashion similar to that of TORC1 and opposite to that of TORC2. Amino acid uptake is also oppositely regulated by Isp7 and TORC2. The regulation of amino acid uptake by TORC1 is more complex. We observed a decrease in the uptake of arginine in tor2-ts (TORC1) mutants, which correlates with the decrease seen in the level of cat1⁺ transcript. However, although TORC1 negatively regulates per1⁺ and put4⁺, TORC1 plays a positive role in proline uptake in the absence of Isp7. TSC (Tsc2) is also required for proline uptake. Thus, with respect to amino acid uptake, TORC1 and TSC play a positive role irrespective of the nature of the amino acid, suggesting that TORC1 and TSC regulate amino acid uptake at a posttranscriptional and/or subcellular localization level, in accord with recent findings (53, 60).

The role of TORC1 in the control of amino acid permease transcription and uptake appears to be conserved in evolution. In S. cerevisiae, a well-known mechanism that is regulated by TOR is nitrogen catabolite repression (NCR). ScTORC1 negatively regulates GLN3, the gene for main transcription factor that is responsible for transcription of NCR genes, including the general amino acid permease gene GAP1 (61) and the proline permease gene PUT4. However, ScTORC1 also positively regulates the specific and high-affinity tryptophan permease gene TAT2 (61). Our findings support a conserved and differential mode of regulation for constitutive versus induced amino acid permeases by TORC1. Our findings show that TORC2 also controls amino acid uptake. Whether this cellular function is conserved in evolution is yet to be determined.

Overexpression of isp7⁺ induced TORC1-dependent Rps6 phosphorylation and decreased TORC2-dependent Gad8 phosphorylation and kinase activity (Fig. 7 and 9). Thus, Isp7 acts as a positive regulator of TORC1 and a negative regulator of TORC2. At least with respect to the induction of Rps6 phosphorylation, Isp7 does not act via TSC or Gtr1/2, the two complexes that negatively and positively regulate TORC1, respectively (23, 26, 30) (Fig. 7 and 8). Genetic epistasis analysis suggests that Isp7 may regulate growth on proline or arginine uptake via Gtr1/2 and independently of TSC.

Since Isp7 reduces the uptake of the relatively poor nitrogen source proline but induces the uptake of arginine, a relatively good nitrogen source, it is possible that Isp7 transmits a nitrogen or amino acid sufficiency signal. Such a scenario is consistent with a role of Isp7 in activation of TORC1. Surprisingly, the transcript level of isp7⁺ is strongly induced following a shift to proline, conditions under which TORC1 activity is downregulated. Moreover, our functional analysis suggests that Isp7 inhibits proline uptake. The transient expression of Isp7 (Fig. 2D) and the presence of feedback loop mechanisms may explain this paradoxical behavior of isp7⁺.

Finally, disruption of isp7⁺ results in an inability to acquire normal stationary-phase physiology, which is manifested by the inability of the cells to withstand heat shock under stationary (nondividing/G₀-like) conditions (42, 62). The long-term viability of nondividing cells is also known as chronological longevity (63). Key players that determine chronological longevity are the TOR signaling pathway and amino acid homeostasis (32, 64). Our findings suggest a novel link between amino acid uptake and TOR signaling via the 2OG-Fe(II) oxygenase Isp7. It will be interesting to address in future studies the possibility that Isp7 plays a role in regulating chronological longevity in a TOR-dependent manner.