The Fission Yeast TOR Homolog,tor1 +, Is Required for the Response to Starvation and Other Stresses via a Conserved Serine*

Targets of rapamycin (TORs) are conserved phosphatidylinositol kinase-related kinases that are involved in the coordination between nutritional or mitogenic signals and cell growth. Here we report the initial characterization of twoSchizosaccharomyces pombe TOR homologs,tor1 + and tor2 +.tor2 + is an essential gene, whereastor1 + is required only under starvation and other stress conditions. Specifically, Δtor1 cells fail to enter stationary phase or undergo sexual development and are sensitive to cold, osmotic stress, and oxidative stress. In complex with the prolyl isomerase FKBP12, the drug rapamycin binds a conserved domain in TORs, FRB, thus inhibiting some of the functions of TORs. Mutations at a conserved serine within the FRB domain ofSaccharomyces cerevisiae TOR proteins led to rapamycin resistance but did not otherwise affect the functions of the proteins. The S. pombe tor1 + exhibits different features; substitution of the conserved serine residue, Ser1834, with arginine compromises its functions and has no effect on the inhibition that rapamycin exerts on sexual development in S. pombe.

Targets of rapamycin (TORs) are conserved phosphatidylinositol kinase-related kinases that are involved in the coordination between nutritional or mitogenic signals and cell growth. Here we report the initial characterization of two Schizosaccharomyces pombe TOR homologs, tor1 ؉ and tor2 ؉ . tor2 ؉ is an essential gene, whereas tor1 ؉ is required only under starvation and other stress conditions. Specifically, ⌬tor1 cells fail to enter stationary phase or undergo sexual development and are sensitive to cold, osmotic stress, and oxidative stress. In complex with the prolyl isomerase FKBP12, the drug rapamycin binds a conserved domain in TORs, FRB, thus inhibiting some of the functions of TORs. Mutations at a conserved serine within the FRB domain of Saccharomyces cerevisiae TOR proteins led to rapamycin resistance but did not otherwise affect the functions of the proteins. The S. pombe tor1 ؉ exhibits different features; substitution of the conserved serine residue, Ser 1834 , with arginine compromises its functions and has no effect on the inhibition that rapamycin exerts on sexual development in S. pombe.
The target of rapamycin (TOR) 1 -mediated signaling pathway in the yeast Saccharomyces cerevisiae activates a cell growth program in response to nutrient availability (reviewed in Refs. 1 and 2). S. cerevisiae contains two TOR homologs, TOR1 and TOR2. TOR1 is a nonessential gene that shares a common function with TOR2 in controlling G 1 progression (3)(4)(5)(6). This common function is sensitive to the drug rapamycin. Thus, either the loss of function of both TOR1 and TOR2 or the inhibition of their products by rapamycin results in a G 1 arrest (6). The rapamycin G 1 arrested cells exhibit various characteristics of starved cells (6 -11), suggesting that the rapamycinsensitive activity of the two Tor proteins inhibits the program that leads to stationary phase. Tor2p also controls the polarization of the actin cytoskeleton (12,13), an essential function that is not shared by Tor1p and is not inhibited by rapamycin (4,5). Similar to the S. cerevisiae TORs, the mammalian TOR homolog FRAP/RAFT1/mTOR is also involved in signaling pathways that regulate transcription, translation, and G 1 progression (reviewed in Ref. 20).
All known TORs contain a conserved domain, the FKBP12rapamycin binding (FRB) domain, that lies adjacent to the phosphatidylinositol kinase domain in the C-terminal region of the protein (14,15). A conserved serine residue within this domain is crucial for the binding of the rapamycin-FKBP12 complex. Studies in mammals (14 -16), S. cerevisiae (3,5,(17)(18)(19), and Cryptococcus neoformans (20) have shown that a mutation at the conserved serine residue confers dominant rapamycin resistance by abolishing the binding to the FKBP12rapamycin complex. The importance of the conserved serine for this binding is reinforced by the atomic structure of the ternary complex FRB-rapamycin-FKBP12 (21). Despite the high conservation of this serine, TOR proteins that are expressed at normal levels and carry a mutated serine appear to retain wild type activities other than FKBP12-rapamycin binding (5,(22)(23)(24).
Schizosaccharomyces pombe is genetically tractable yeast that is highly divergent from S. cerevisiae. These two yeasts often have distinct differences in carrying out the same cellular functions, which makes their comparative study especially revealing (25). Upon starvation, S. pombe cells enter either the stationary phase or the sexual development pathway (reviewed in Ref. 26). We previously reported that rapamycin has a different effect on S. pombe compared with its effect on S. cerevisiae. Rapamycin does not inhibit the growth of S. pombe but specifically inhibits sexual development in response to starvation (27). As a first step toward understanding the response to rapamycin in S. pombe, we cloned and initiated a functional analysis of the S. pombe TOR homologs, named tor1 ϩ and tor2 ϩ . We show here that at least some of the functions of each of these TORs are distinct. Thus, tor2 ϩ is essential for growth, whereas tor1 ϩ is required only under starvation and other stress conditions. We also demonstrate that the conserved serine residue within the FRB domain of Tor1 is important for the protein function and does not play a detectable role in the response of starved S. pombe to rapamycin.

EXPERIMENTAL PROCEDURES
Yeast Strains, Media, and General Techniques-Yeast strains used in this paper are described in Table I. All media used in this study are based on those described previously (28). EMM-N contains no nitrogen; EMM lowG contains 0.1% glucose. Rapamycin was used as described previously (27). Transformation of S. pombe cells was performed by electroporation (29). Assays for mating or sporulation efficiency were carried out as described in (27).
Fluorescence-activated Cell Sorter Analysis-Cells were stained with the DNA fluorochrome propidium iodide and analyzed by a Becton Dickinson FACSort as described by (30). Data were analyzed by Cell Quest software for Macintosh.
Disruption of S. pombe tor1 ϩ -A fragment containing 5.84 kbp of the C-terminal region of tor1 ϩ gene was amplified by PCR using the Expand Long Template PCR System (Roche Molecular Biochemicals) with a genomic S. pombe DNA preparation as a template and the primers 104 (5Ј-TTGAAGAATCTGCAGCAATAAATATTC) and 105 (5Ј-AA-GATTTGATCGGCATTTGGCAC). The resulting PCR fragment was subcloned into a pGEM-T vector (Promega) to give pGEMtor1. A 3.63kbp HindIII fragment of pGEMtor1 containing the kinase and FRB-like domains was replaced with an HindIII fragment containing the entire ura4 ϩ , resulting in the plasmid ptor1::ura4. NotI and SacI were used to release the 4-kbp tor1::ura4 disruption fragment, and this was gel purified and transformed into the diploid TA07. Stable Ura4 ϩ diploids were selected, and their DNA was extracted and subjected to PCR analysis with primers 105 and 137 (5Ј-TTGTAAATAGGATAGCCAG-CACC), which lies 100 bp upstream of the 5Ј end of the disruption construct.
Disruption of S. pombe tor2 ϩ -0.4-and 0.5-kbp fragments containing the very 5Ј and 3Ј ends of the tor2 ϩ open reading frame, respectively, were amplified by PCR, using genomic DNA as the template. The primers were: 152 (5Ј-ATAAGAGTCGACTCACAAGTGTTGTGAACTT-GGTGG); 154 (5Ј-GGGGTACCGAGCTCATTCCTGCTTTTCAACCCA-GG); 153 (5Ј-ATAAGAGTCGACAGATACGTGAAGAGGGGTGGTGA-C); and 155 (5Ј-CGGGATCCGTTTCTGGTAGGTGACAGTCCC). The resulting 0.4-and 0.5-kbp PCR fragments were digested with KpnI and BamHI, respectively (sites are underlined in the sequence of the primers) and ligated with the 1.8-kbp kpnI-BamHI fragment containing the entire ura4 ϩ gene. The fragment containing the ura4 ϩ gene flanked by tor2 ϩ sequences was amplified by PCR using 152 and 153 and used to transform the diploid strain TA07. The correct disruption of tor2 ϩ was verified by PCR analysis using the primer 165 (5Ј-CGGGATCCCCAT-TTAATAGAGAAAGGGATATTAGC) that lies 32 bp upstream of the 5Ј end of the disruption construct in combination with primer 135 (5Ј-G-TTATAAACATTGGTGTTGGAACAG) that lies in the ura4 ϩ gene.
To express tor1 ϩ in S. cerevisiae we used pIRT2-tor1 ϩ as a template in a PCR reaction with primers 143 (5Ј-CGGGATCCGTCTATCGTT-TCACTCGCTCTC) (BamHI, for subsequent cloning is underlined) and primer 141. This reaction resulted in the amplification of a 7.4-kbp fragment that contains the tor1 ϩ ORF, 30 and 300 bp upstream and downstream of the open reading frame, respectively. This fragment was cloned into the pCM190 S. cerevisiae expression vector (32), resulting in pCM190-tor1 ϩ .
The mutation at the conserved serine in the Tor1 FRB domain, S1843R, was created by site-directed mutagenesis by overlap extension (33). pIRT2-tor1 ϩ served as a template in a PCR reaction with primers 145 (5Ј-TGCATCCTCAGGCATTGGTGTATTC) and 146 (5Ј-GAAAAAT-AAGCCTGACGAGCTTCCTCTAATC; mutations are in bold type). The resulting PCR product (220 bp) was gel purified and used as a primer for a second-round of PCR with primer 144 (5Ј-AATAGATCTCTCGTT-GAGTCCTTCG). The resulting 1.58-kbp PCR product was cleaved with BglII and Bsu36I, whose restriction sites reside near the fragment ends. This fragment was used to replace the corresponding fragment in pIRT-2-tor1 ϩ and pCM190-tor1 ϩ . We verified that only the desired mutation had occurred during PCR or cloning of the BglII/Bsu36I fragment by DNA sequence analysis. In addition, to further ascertain that the only defect of the tor1 S1834R resided in Ser 1834 , the plasmid carrying tor1 S1834R was cut with BglII/Bsu36I, and the 1.58-kbp fragment containing the site of the mutation was recovered. This fragment was used as the template in a PCR reaction performed to replace Arg 1834 back with serine. The fragment containing Ser 1834 was ligated back to the BglII/Bsu36I cut plasmid to reconstruct a wild type tor1 ϩ . This tor1 ϩ , derived from tor1 S1834R , fully restored tor1 ϩ function.
Epitope Tagging and Western Blot Analysis of Tor1 and Tor1 S1834R -The wild type and mutant tor1 genes were amplified from the plasmids pIRT2-tor1 ϩ and pIRT2-tor1 S1834R using the primers 141 (containing a BamHI site, see above) and 199 (ATAAGAATGCGGCCGCATGGAG-TATTTTAGTGATCTAAAAAAC) (the NotI site is underlined). PCR products were digested with BamHI/NotI and cloned into the BglII/NotI sites of the thiamine-repressible vector pSLF273 (34), downstream and in frame with the plasmid sequences encoding the triple hemagglutinin (HA) epitope domain. The resulting plasmids carrying the N-terminal HA-tagged wild type and mutated proteins were transformed into a ⌬tor1 strain, TA163. Total protein extracts were prepared from cells growing under conditions that allow protein expression (in the absence of thiamine) following the method described by (28). Aliquots of whole cell extracts containing 60 g of protein were fractionated by SDSpolyacrylamide gels and transferred to membrane filters. The immobilized proteins were detected using the PerkinElmer Life Sciences enhanced chemiluminescence system. The HA-tagged proteins were detected with monoclonal antibody HA.11 (Berkley Antibody Co.). Polyclonal antibodies raised against S. cerevisiae FKBP12 (a kind gift of J. Heitman, Duke University Medical Center) were used to detect the S. pombe FKBP12 proteins as an internal loading control.

Identification of TOR Homologs in the S. pombe Genome-
The DNA sequence of most of the S. pombe genome has been determined through the coordination of the Sanger Center. Based on sequence comparisons, we identified two TOR homologs in chromosome II. We named one of these, on cosmid SPBC30D10, tor1 ϩ and the second, contained on two overlapping cosmids, SPBC216 and SPBC646, tor2 ϩ . The open reading This study a Strains TA99, TA100, and TA120 are haploid segregants of the diploid TA82. Strains TA99 and TA120 are the parents of TA132. TA157 is a haploid segregant from a cross between TA99 and TA16. TA163 is derived from TA99 and is the result of transformation of TA99 with the marker swap construct, ura4⌬ϻhis1 ϩ , a kind gift of P. Fantes (Edinburgh University, UK). frames of tor1 ϩ and tor2 ϩ encode for 2335-and 2337-amino acid proteins, respectively. No introns are found in either of the open reading frames. The two S. pombe TOR homologs share 52% overall identity. A slightly lower level of overall identity (42-44%) is revealed when the amino acids encoded by tor1 ϩ or tor2 ϩ are aligned with the human TOR, the S. cerevisiae Tor1p or Tor2p, or the C. neoformans Tor1. The C-terminal regions of the S. pombe Tor1 and Tor2 proteins contain the FRB and the phosphatidylinositol kinase-like domains and are the most conserved regions of the proteins. The C-terminal regions of Tor1 and Tor2, residues 1814 -2171 and 1817-2174, respectively, show 64 -65% identity with the corresponding amino acids of other TOR homologs in the data bases. Amino acid comparisons of these C-terminal regions indicate that both Tor1 and Tor2 in S. pombe contain the conserved structural features that characterize the TOR family of proteins. This includes the conserved serine at position 1834 or 1837 in Tor1 or Tor2, respectively (Fig. 1).
The S. pombe tor2 ϩ , but Not tor1 ϩ , Is Required for Growth-As a first step toward understanding the physiological function(s) of the TORs in S. pombe, we disrupted the chromosomal copies of tor1 ϩ or tor2 ϩ by one-step gene replacement ("Experimental Procedures"). Fig. 2 shows a tetrad analysis of a heterozygous diploid strain (TA137) in which one of the copies of tor2 ϩ was replaced with the selective marker ura4 ϩ . A total of 25 tetrads were dissected, and all yielded only two or one viable spores, none of which cosegregated with the ura4 ϩ marker, indicating that the tor2 disruption was lethal. Microscopic examination revealed that in the nonviable segregants, the spores produced single cells that did not undergo cell divisions (data not shown). This lethal phenotype was rescued by reintroduction of tor2 ϩ . Strain TA137 transformed with plasmid pIRT2-tor2 ϩ (tor2 ϩ expressed under the regulation of its endogenous promoter) was sporulated and dissected. Progeny disrupted in the chromosomal copy of tor2 ϩ were obtained, but in all cases they carried the plasmid-borne wild type tor2 ϩ gene.
We also used the ura4 ϩ selective marker to disrupt tor1 ϩ ("Experimental Procedures"). In contrast to tor2 disruption, sporulation of diploid heterozygous for tor1 ϩ disruption (TA82) yielded viable Ura4 ϩ haploids. PCR analysis confirmed that the Ura4 ϩ haploids carried a disputed allele of tor1 (data not shown). Thus, whereas tor2 ϩ is an essential gene, tor1 ϩ is a nonessential gene. tor1 ϩ Is Required for Entrance into Stationary Phase and for Sexual Development-We further analyzed two independently isolated ⌬tor1 haploid clones. Under optimal growth conditions, the growth rate and cell morphology of ⌬tor1 cells were indistinguishable from that of wild type cells (Fig. 3A). However, ⌬tor1 cells exited the logarithmic phase at a lower cell density, were abnormally long, and lost viability rapidly (Fig.  3). The loss of viability of ⌬tor1 depends on the growth medium. Whereas ⌬tor1 cells died when they reached saturation in rich medium, they maintained viability comparable with wild type cells when grown to saturation in minimal medium (Fig. 3B). This suggests that ⌬tor1 cells are defective in the response to particular sets of conditions rather than in the actual cellular processes that allow cells to acquire the stationary phase physiology.
We also noted that ⌬tor1 cells failed to cross with wild type strains. Microscopic examination suggested a defect at an early stage of sexual development, before conjugation had occurred. Media limiting for either nitrogen or carbon sources are conventionally used for a quantitative analysis of conjugation (mating). Under these conditions, haploid cells of the opposite mating types can conjugate to form a diploid zygote, which rapidly undergoes meiosis and sporulation and produces an ascus (reviewed in Ref. 26). Under either nitrogen or carbon starvation, ϳ60% of wild type cells underwent mating compared with less than 1% of ⌬tor1 cells. The sterile phenotype of ⌬tor1 was efficiently suppressed when we reintroduced tor1 ϩ (Fig. 4A).
Nitrogen or carbon starvation is also a signal for diploid cells to enter meiosis (reviewed in Ref. 26). Under these conditions most of the wild type diploid cells, Ͼ60%, underwent sporulation, whereas Ͻ1% of homozygous ⌬tor1 diploids sporulated (results not shown). Thus, in addition to its role in mating, tor1 ϩ is also required for meiosis/sporulation. Analysis of the DNA content of growing and starved ⌬tor1 cells also indicates that these cells are defective in their response to starvation. In the absence of a mating partner, starved S. pombe cells become arrested in either G 1 or G 2 , depending on the growth medium; nitrogen starvation arrests cells mainly at the G 1 phase, and carbon starvation arrests cells mainly at the G 2 phase (35). The DNA profile of ⌬tor1 cells under optimal growth conditions shows a major 2n DNA peak, characteristic of growing wild type cells (see Refs. 36 and 37 and Fig. 4B). However, under nitrogen starvation, ⌬tor1 cells show an abnormal DNA profile as cells failed to arrest in G 1 (Fig. 4B). Because G 1 arrest is a prerequisite for mating, the failure of ⌬tor1 cells to arrest their growth in G 1 may be associated with their inability to undergo sexual development. tor1 ϩ Is Required for Growth under Osmotic or Oxidative Stress Conditions-We noted that the phenotype of ⌬tor1 cells was particularly similar to that of cells disrupted for atf1 ϩ , a gene that encodes a bZIP (basic leucine zipper) transcription factor (37). Under starvation conditions, both ⌬tor1 and ⌬atf1 cannot arrest in G 1 , exhibit an abnormal elongated morphology, lose viability in rich but not minimal medium, and are sterile. atf1 ϩ has also been implicated in regulating the cellular response to a variety of stress conditions, such as cold, osmotic stress, and oxidative stress (37-39). We found that tor1 ϩ is also required under these stress conditions; unlike wild type cells,  FIG. 2. tor2 ؉ is an essential gene. A diploid tor2 ϩ /⌬tor2 was sporulated and dissected on a nonselective medium. All viable spores were Ura Ϫ , indicating that tor2 ϩ is an essential gene.
⌬tor1 could not form colonies on medium containing 0.5 M NaCl or 1 M KCl (Fig. 4C) or on medium containing 5 mM H 2 O 2 (Fig.  4D) or below 20°C (results not shown). Taken together, our findings reveal a striking similarity between the phenotypes of ⌬tor1 and ⌬atf1 cells. It remains to be determined whether tor1 ϩ and atf1 ϩ are involved in the same signaling pathway.
Ser 1834 in the FRB Domain of Tor1 Is Required for Tor1 Activity-We previously reported that rapamycin specifically inhibits sexual development in S. pombe (27). As indicated above, ⌬tor1 cells are unable to undergo sexual development. Because rapamycin is known to inhibit the TOR proteins in mammals, S. cerevisiae, and C. neoformans, we considered the possibility that rapamycin exerts its effect by inhibiting the function of the S. pombe Tor1 during sexual development. A conserved serine residue in TORs has been identified as the site for missense mutations (serine substituted with arginine, isoleucine, or glutamic acid) conferring dominant rapamycin resistance (see the Introduction). We mutated the equivalent serine residue in S. pombe Tor1, Ser 1834 , into arginine ("Experimental Procedures"). The tor1 ϩ and tor1 S1834R genes, cloned into the S. pombe expression vector pIRT2 ("Experimental Procedures") were transformed into ⌬tor1 strains TA99 or TA157. TA99 and TA157 are isogenic except that TA157 is a homothallic strain (cells can switch their mating types between h ϩ and h Ϫ every other generation), whereas TA99 is a heterothallic strain composed of h Ϫ cells only. Surprisingly, we found that the mutation at Ser 1834 diminished the activity of Tor1; the mating efficiency was extremely low when TA157 cells carrying tor1 S1834R were induced to undergo mating (0.9%, Fig. 5A). In crosses between wild type and ⌬tor1 cells (TA99) carrying tor1 S1834R , we observed that the wild type cells partially suppressed the sterility of ⌬tor1 cells carrying tor1 S1834R (Fig. 5A). The mutation S1834R also diminished the activity of Tor1 under osmotic stress conditions (Fig. 5C) or in acquisition of stationary phase physiology (data not shown). Because mating is very inefficient in tor1 S1834R mutants, it was no surprise that this mutated allele did not confer dominant resistance to rapamycin in wild type TA16 transformants (Fig. 5B).
Our finding that Ser 1834 is critical for the function of Tor1 is surprising given that equivalent mutations did not affect TOR function in S. cerevisiae. To ascertain that the only defect of the tor1 S1834R resided in S1834, the plasmid carrying tor1 S1834R was used as the template in a PCR reaction performed to replace Arg 1834 back with serine (see "Experimental Proce-  4. Phenotypes of ⌬tor1 cells under starvation, osmotic, and oxidative stress. A, heterothallic ⌬tor1 strain (TA99) was transformed with the S. pombe vector pIRT2 or pIRT2-tor1 ϩ . The transformants were mixed with wild type (WT) cells of the opposite mating type (TA02) and induced to undergo sexual development in EMM-N medium. Wild type is TA16 induced to undergo sexual development in EMM-N. B, exponentially growing wild type (TA100) and ⌬tor1 (TA99) cells in minimal medium (log) were collected, washed and resuspended in nitrogen free (EMM-N), low glucose (EMMlowG), or EMM (Stationary) medium, and incubated at 25°C for 3 days. Aliquots were removed from growing and starved cells, and the DNA content of individual cells was measured by fluorescence-activated cell sorter. C, ⌬tor1 cells transformed with vector only or plasmid carrying tor1 ϩ were streaked onto YE agar plates supplemented with 1 M KCl or 0.5 M NaCl and incubated for 5 days at 32°C. D, ⌬tor1 cells transformed with vector only or plasmid carrying tor1 ϩ were streaked onto YE agar plates supplemented with 0 or 5 mM H 2 O 2 and incubated for 5 days at 32°C. 5. Expression of tor1 S1834R gene in S. pombe cells. A, heterothallic ⌬tor1 strain (TA99) or homothallic ⌬tor1 strain (TA157) were transformed with the S. pombe tor1 ϩ or tor1 S1834R . TA99 transformants were mixed with wild type (WT) cells of the opposite mating type (TA02) and induced to undergo sexual development in EMM-N medium. TA157 transformants were also induced to undergo sexual development in EMM-N medium. B, wild type homothallic cells (TA16) transformed with the S. pombe tor1 ϩ or tor1 S1834R were induced to undergo sexual development in EMM-N in the presence or absence of rapamycin. C, ⌬tor1 cells (TA99) transformed with vector only, plasmid carrying tor1 ϩ , or tor1 S1834R were streaked onto YE agar plates supplemented with 0.8 M KCl and incubated at 32°C. Photographs were taken after 4 days of incubation.

FIG.
dures"). This tor1 ϩ , derived from tor1 S1834R , fully restored tor1 ϩ , demonstrating that tor1 S1834R carries no mutations other than S1834R. We also examined whether the mutation at S1834 affected protein stability. To this aim, the wild type and mutated Tor1 proteins were tagged with the HA epitope (see "Experimental Procedures"). Analysis of ⌬tor1 cells expressing these proteins demonstrated that the HA tagging did not affect the activity of the proteins in vivo (data not shown). Western blot analysis using antibody raised against the HA epitope demonstrated that the level of Tor1 and Tor1 S1834R are comparable (Fig. 6). Therefore, we conclude that the S1834R mutation did not affect protein stability in vivo, but rather the S1834 residue is required for the function of Tor1 under starvation or osmotic stress conditions. Because the S. pombe tor1 ϩ shows a significant level of homology with the S. cerevisiae TOR genes (about 43% identity), we examined the functions of the S. pombe tor1 ϩ and tor1 S1834R in S. cerevisiae. Plasmids carrying tor1 ϩ or tor1 S1834R , expressed under the control of the S. cerevisiae ADH1 promoter, were introduced into wild type S. cerevisiae cells. Fig. 7 shows that tor1 S1834R but not tor1 ϩ conferred dominant rapamycin resistance in S. cerevisiae. This is similar to the effect of equivalent point mutations in the S. cerevisiae TOR proteins (3,5,19), except that tor1 S1834R did not confer rapamycin resistance at a high concentration of rapamycin (100 ng/ml, results not shown). The dominant resistance exhibited by the S. pombe tor1 S1834R indicates that the S. pombe TOR homolog can complement the function of the S. cerevisiae TORs. It also implies that the conserved serine in tor1 ϩ is critical for the binding of FKBP12-rapamycin when expressed in S. cerevisiae cells. In conclusion, it appears that although S1834 is required for Tor1 function in S. pombe, it is not required for the rapamycin-sensitive TOR function of S. cerevisiae. DISCUSSION We report here the identification and initial characterization of the S. pombe homologs of the TOR genes. We found the S. pombe tor2 ϩ gene is essential for growth, whereas tor1 ϩ is required only under starvation, osmotic stress, and oxidative stress.
The inability of ⌬tor1 cells to respond appropriately to starvation conditions may suggest that the S. pombe tor1 ϩ , like its S. cerevisiae TOR homologs (2,40), participates in signal transduction pathways that are involved in nutrient sensing. However, there are significant differences between the functions of the S. cerevisiae and S. pombe TORs. First, the S. pombe tor1 ϩ has a positive role in the sexual development pathway and entry into stationary phase, whereas the activity of the S. cerevisiae TORs is required to repress meiosis (41) and entry into stationary phase (6 -10, 42). Second, the S. pombe tor1 ϩ is required for the appropriate response to a variety of stress conditions, whereas there is no evidence that the S. cerevisiae TOR homologs are involved in the response to stresses other than starvation. Third, each of the two S. pombe TOR homologs carries out a distinct function that is not shared by the other homolog. In contrast, the S. cerevisiae TOR1 is a nonessential gene, and its function in regulating growth in response to nutrient availability is shared with TOR2 (3)(4)(5).
Mutational analysis of tor1 ϩ revealed that the conserved serine residue within the FRB domain of Tor1 plays a critical role in the protein cellular function. Thus, although the mutation at Ser 1834 did not affect the level of protein expression (Fig.  6), tor1 S1834R can only partially complement the defects observed in ⌬tor1 cells (Fig. 5). Although Ser 1834 is required for Tor1 function, its equivalent serine residues in the S. cerevisiae TOR proteins (3,5,18) and possibly the human TOR protein (22)(23)(24) do not appear to play an essential role in the studied functions of the proteins. However, given the conserved nature of this serine residue, its role in other TOR homologs may become evident under yet unidentified conditions.
None of the functions of the S. pombe TORs described in this work appear to be inhibited by rapamycin. Rapamycin specifically inhibits sexual development in S. pombe, at an early stage, before mating (27). Cells disrupted for tor1 ϩ are deficient in their ability to undergo mating. However, the effects of Tor1 and rapamycin on sexual development appear to be unrelated. The inability of ⌬tor1 cells to enter sexual development seems to be part of a general defect in responding to nutritional deprivation. Thus, ⌬tor1 cells fail to enter stationary phase, arrest in G 1 in response to starvation, or undergo meiosis/ sporulation. In contrast, rapamycin specifically inhibits the sexual development pathway and does not interfere with other responses to starvation. Thus, cells treated with rapamycin can enter stationary phase properly, are only slightly defective in meiosis/sporulation (27), and can arrest their growth in G 1 under starvation conditions. 2 Our finding that tor1 S1834R could not alleviate the inhibitory effect of rapamycin on sexual development (Fig. 5B) is consistent with our suggestion that rapamycin does not exert its effect in S. pombe by forming a toxic FKBP12-rapamycin complex that inhibits the Tor1 function.
If neither Tor1 nor Tor2 is the protein target for rapamycin, than what is the target for the action of rapamycin in S. pombe? Recent findings in our lab show that cells disrupted for the S. pombe FKBP12 homolog exhibit a phenotype highly similar to treatment with rapamycin. Thus, it is most probable that rapamycin inhibits sexual development by inhibiting the cellular 2 R. Weisman and M. Choder, unpublished results.
FIG. 6. Wild type and mutant HA-TOR1 fusion proteins are stably expressed. Top panel, HA-Tor1 and HA-Tor1 S1834R fusion proteins were expressed in strain TA163 and detected by Western blot with antibodies against the HA epitope. Bottom panel, the same extracts as in the top panel were detected using ␣ FKBP12 antiserum raised against the S. cerevisiae FKBP12 homolog. This antibody cross-reacts with the S. pombe FKBP12 protein and is used here as a control for protein loading.
FIG. 7. The S. pombe tor1 S1834R can confer rapamycin resistance in S. cerevisiae. S. cerevisiae wild type cells, JK9 -3d (19) transformed with plasmids containing the S. pombe tor1 ϩ or tor1 S1834R genes, were streaked on minimal plates containing 0 or 10 ng/ml rapamycin. function of FKBP12 in sexual development 3 and not by inhibiting TOR-related function.
Why rapamycin does not inhibit neither the function of Tor1 or Tor2? Our functional analysis of tor1 ϩ and tor1 S1834R in S. cerevisiae cells (Fig. 5A) suggests that Tor1 can bind the FKBP12-rapamycin complex, at least in S. cerevisiae cells. It is possible that Tor1 interacts with FKBP12-rapamycin complexes in S. pombe, but this interaction does not inhibit the studied functions of Tor1 or Tor2. By analogy, the function of the S. cerevisiae Tor2p in the control of the actin cytoskeleton organization is not inhibited by rapamycin (5).
The features of S. pombe TOR proteins, together with other studies of TOR functions, indicate that these proteins are involved in many distinct cellular functions. Given that the Cterminal region containing the FRB and the kinase domains of the TORs is highly conserved, the differences between the TORs might reside in the less conserved N-terminal region. This is an intriguing possibility yet to be explored.