TOR1 and TOR2 Have Distinct Locations in Live Cells

Abstract

TOR is a structurally and functionally conserved Ser/Thr kinase found in two multiprotein complexes that regulate many cellular processes to control cell growth. Although extensively studied, the localization of TOR is still ambiguous, possibly because endogenous TOR in live cells has not been examined. Here, we examined the localization of green fluorescent protein (GFP) tagged, endogenous TOR1 and TOR2 in live S. cerevisiae cells. A DNA cassette encoding three copies of green fluorescent protein (3XGFP) was inserted in the TOR1 gene (at codon D330) or the TOR2 gene (at codon N321). The TORs were tagged internally because TOR1 or TOR2 tagged at the N or C terminus was not functional. The TOR1^D330-3XGFP strain was not hypersensitive to rapamycin, was not cold sensitive, and was not resistant to manganese toxicity caused by the loss of Pmr1, all indications that TOR1-3XGFP was expressed and functional. TOR2-3XGFP was functional, as TOR2 is an essential gene and TOR2^N321-3XGFP haploid cells were viable. Thus, TOR1 and TOR2 retain function after the insertion of 748 amino acids in a variable region of their noncatalytic domain. The localization patterns of TOR1-3XGFP and TOR2-3XGFP were documented by imaging of live cells. TOR1-3XGFP was diffusely cytoplasmic and concentrated near the vacuolar membrane. The TOR2-3XGFP signal was cytoplasmic but predominately in dots at the plasma membrane. Thus, TOR1 and TOR2 have distinct localization patterns, consistent with the regulation of cellular processes as part of two different complexes.

TOR (target of rapamycin) is a large (∼2,500 amino acids), highly conserved protein kinase that controls cell growth in response to nutrients. In Saccharomyces cerevisiae, there are two highly related TOR proteins, TOR1 and TOR2 (11). Both proteins associate with membranes and are partially resistant to extraction with Triton X-100 (3, 21). TOR is found in two structurally and functionally distinct multiprotein complexes, TORC1 and TORC2 (41). TORC1 contains TOR1 or TOR2 and Kog1, Tco89, and Lst8. TORC2 contains TOR2, Avo1, Avo2, Avo3, Bit61, and Lst8. Both TORC1 and TORC2 are essential, but only TORC1 is inhibited by rapamycin. TORC1 controls several growth-related processes, including transcription, translation, ribosome biogenesis, nutrient transport, and autophagy. TORC2 controls a different set of processes, including actin organization, endocytosis, and lipid biosynthesis (4). The TORCs, their components, and their key role in cell growth have been conserved from yeast to human. A major unanswered question is how TOR regulates so many processes, directly or indirectly.

One potential source for the diversity of biologic processes under TOR regulation is the subcellular localization of TOR. TOR signaling may depend on the colocalization of TOR with key substrates involved in different cellular functions at different times to define branches in TOR signaling. The localization of TOR1 and TOR2 in cells has been studied by biochemical fractionation and immunostaining. In an early study, TOR2 was localized to the surface of the vacuolar membrane by using antibody raised against a sequence in TOR2 (5). Subsequently, both [³⁵S]TOR1 and [³⁵S]TOR2 were found in P13 and P100 membrane fractions (21). The P13 fraction is enriched in the plasma membrane (PM), endoplasmic reticulum (ER), vacuoles, and mitochondria, whereas the P100 fraction is enriched in Golgi bodies, endosomes, and secretory vesicles. TOR1 and TOR2 from the P13 fraction further fractionated on equilibrium sucrose gradients similarly to Pma1, a PM marker. The TORs were also found in a distinct, unidentified membrane pool. As revealed by immunofluorescence, overexpressed hemagglutinin-tagged TOR1 and TOR2 localized to discrete sites or “dots” at or just beneath the PM (21). By immunogold electron microscopy, Wedaman et al. identified hemagglutinin-tagged TOR2 adjacent to the PM and along intracellular membranous tracks (40). In a recent biochemical study, TORC2 components Avo3 and TOR2 correlated best with an early endosome marker (Rsv5), whereas TORC1 components overlapped diffusely with trans-Golgi, ER, and vacuolar markers (3). Li et al. reported in an immunostaining study that TOR1 is predominantly nuclear and exported to the cytoplasm in response to rapamycin treatment or nutrient starvation (24). Thus, the localization of TOR is ambiguous, possibly because previous studies always relied on fixed or fractionated cells. Here, we visualize the localization of functional, internally green fluorescent protein (GFP)-tagged TOR1 and TOR2 in live cells. We find that the localization patterns of TOR1-3XGFP (TOR1 tagged with a DNA cassette encoding three copies of GFP) and TOR2-3XGFP (TOR2 with a DNA cassette encoding three copies of GFP) are distinct and complex, possibly accounting for the heretofore ambiguity in TOR localization. Furthermore, the various localization patterns of the TORs may underlie how they control many different cellular processes.

MATERIALS AND METHODS

Yeast media.

Rich medium (YPD) or synthetic medium (SD or SC) was prepared as described previously (35). Rapamycin was added to YPD medium from dilutions of a 1-mg/ml stock in 90% ethanol-10% Tween 20 just before plates were poured.

Strains with 3XGFP inserted into TOR1 or TOR2.

A sequence for 3XGFP with the S65G mutation and optimized for expression in yeast was amplified from pBS-3XGFP-TRP1 (22) with primers MT101 (5′-CCCGATATCGGAGGATCC ATGTCTAAAGGT-3′) and MT102 (5′-GGACTAGTTTTGTACAATT CATCCATACCAT-3′) and cloned into the EcoRV and SpeI sites of pUG6 after restriction, creating pOM3. We constructed 3XGFP strains as described previously using pOM3, with minor modifications (15). The strategy uses a kanMX6 resistance marker to select recombinants that is later removed by the Cre recombinase, restoring the open reading frame (ORF) for the gene containing the repeated tag. The specific TOR primers used to amplify the 3.8-kb cassette from pOM3 are given in Table 1. Portions of the PCRs were used to transform haploid TB50a (for TOR1) and diploid TB50a/α (for TOR2) strains by a high-efficiency lithium acetate-polyethylene glycol method. Transformants were allowed to recover for 4 to 6 h in YPD before being plated on YPD plates containing G418. Single colonies were streaked on YPD plates and allowed to grow at 30°C, and plates were stored at 4°C as the primary isolates.

TABLE 1.

Primers for 3XGFP tagging of TOR1 and TOR2 by using pOM3

Primer	Target	Sequence (5′ to 3′)a
TOR1-N-Fwd	N terminus	GGTAAAGTGAAACATACATCAACCGGCTAGCAGGTTTGCATTGATATG TGC AGG TCG ACA ACC CTT AAT
TOR1-N-Rev	N terminus	CGCTTTCAAAAGTTTACTCTTCCAAATCTGCTCCTCATGCGGTTC GCG GCC GCA TAG GCC ACT
TOR1-D67-Fwd	D67	TGACTTCTAGTAGGTTTGATGGAGTGGTGATTGGCAGTAATGGG TGC AGG TCG ACA ACC CTT AAT
TOR1-D67-Rev	D67	GGTTAATTCGCGGAAAATTTTCTCCAAAATGGGCTTAAAATTTAC GCG GCC GCA TAG GCC ACT
TOR1-D330-Fwd	D330	ATCCATGCAAGTTTGTTGGTTTATAAGGAAATCTTGTTTTTGAAG TGC AGG TCG ACA ACC CTT AAT
TOR1-D330-Rev	D330	TATGCAATTTAGACACATTTGGTCGAACACTTGATTCAAAAAGGG GCG GCC GCA TAG GCC ACT
TOR2-N-Fwd	N terminus	TCTTTCTCAAAGAGATTTCTGATCTTTACTTTCCCCATATGAAAAATG TGC AGG TCG ACA ACC CTT AAT
TOR2-N-Rev	N terminus	AGACAATAAGTTAGGTGGCGTGGTGTATTTGTTAATGTATTTATT GCG GCC GCA TAG GCC ACT
TOR2-N321-Fwd	N321	CAAAGATTATTTCAAGGTTGTACACATGGCTTAAGTCTCAATACG TGC AGG TCG ACA ACC CTT AAT
TOR2-N321-Rev	N321	GCTGAGTAATTCTCGAAATACCAACAGAGTAGCATGCACTGAATC GCG GCC GCA TAG GCC ACT

^aThe underlined portion is the sequence in TOR1 or TOR2. The italicized portion is the ATG codon added to the primer to replace the initiating ATG codon. Sequence that is neither italicized nor underlined is the sequence in the pOM3 template.

The TOR1 ORF is disrupted by the stop codon of the kanMX6 marker until removal by the Cre recombinase (15). The primers used in the characterization of strains are listed in Table 2. Colonies were tested for integration at the correct site by a colony PCR before the Cre step, using a reverse primer (kanMX6/Rev) in the nonrepeated kanMX6 ORF. With TOR1/Fwd/−500, the expected products were observed in 9/15 colonies tested for N-terminal tagging (741 nucleotides [nt]), 8/15 tested for D67 tagging (942 nt), and 11/16 tested for D330 tagging (1,731 nt), and with TOR2/Fwd/−486, the expected product (727 nt) was observed in 2/2 colonies.

TABLE 2.

Primers used in characterization of strains

Primer	Locationa	Sequence
TOR1/Fwd/−500	Promoter, −500 to −480	5′-AGACAGCAAACCACCTGTTGC-3′
TOR1/Fwd/−72	Promoter, −72 to −53	5′-CCATAGTAGCTTCACGAGAG-3′
TOR1/Fwd/+801	ORF, 801 to 821	5′-GGTGATCAGAATTGATGCGTC-3′
TOR1/Rev/+280	ORF, 258 to 280	5′-TACTGGCCAATTTTCGTTCCTCC-3′
TOR1/Rev/+1101	ORF, 1080 to 1101	5′-TGCTAATAGGGGAACAATCTGG-3′
TOR2/Fwd/−486	Promoter, −486 to −464	5′-CGTCATTTTTTACATACTGCTGC-3′
TOR2/Fwd/+711	ORF, 711 to 730	5′-ATACAGGAGACATGCTGCGC-3′
TOR2/Fwd/+931	ORF, 931 to 951	5′-TTTCAAGGTTGTACACATGGC-3′
TOR2/Rev/+990	ORF, 970 to 990	5′-CAGAGTAGCATGCACTGAATC-3′
kanMX6/Rev	3XGFP cassette	5′-CCGTGCGGCCATCAAAATG-3′
GFP/Rev	3XGFP cassette, 165 to 185 (after Cre)	5′-GCATCACCTTCACCTTCACCG-3′

^aLocation is nt from ATG start codon where specified.

Cells taken from cultures in liquid YPD were transformed with pSH47 with lithium acetate-polyethylene glycol by a lower-efficiency, simplified method, and the transformants were selected on SD-uracil (SD-ura) plates (secondary colonies). Plasmid pSH47 introduces Gal-inducible Cre on a 5-fluoroorotic acid removable plasmid. Tiny amounts of single colonies were inoculated into 4 ml of SC-ura containing 1% galactose and 1% raffinose in 15-ml disposable tubes. After induction for 5 h (30°C), the cells were collected by centrifugation, resuspended in a small residual volume, and streaked so as to obtain single colonies on YPD plates. These tertiary colonies were identified, circled and labeled, and streaked on YPD plates and YPD plates containing G418 to test for excision of kanMX6 by Gal-induced Cre. A convention, for example, TOR1N-15-6-1, was helpful to indicate the gene and site followed by three single colony numbers (for primary [G418], secondary [SD-ura], and tertiary [YPD] colonies after the introduction of Cre). The tertiary colonies that had lost the kanMX6 marker were studied further.

Validation of TOR1-3XGFP strains.

TOR2-3XGFP strains were verified as described later in the text. TOR1-3XGFP strains were verified as follows. Recombination to integrate the full 3XGFP cassette was confirmed by colony PCR with primers to TOR1 sequences outside the cassette. PCR with primers TOR1/Fwd/−72 and TOR1/Rev/+280 (Table 2) gave the 2,599-nt band diagnostic for 3XGFP for colony TOR1/D67-1-1-1, which became strain VA38 (Table 3). PCR with the same set gave the expected 2,602-nt band expected for 3XGFP for colony TOR1N-15-6-1, strain VA41 (Table 3). PCR with TOR1/Fwd/+801 and TOR1/Rev/+1101 gave the expected 2,548-nt band diagnostic for 3XGFP with TOR1/D330-14-1-6 and 14-1-3, strains VA34 and VA35, respectively, but a larger ∼3.4-kb band for TOR1/D330-3-1-2, strain VA33.

TABLE 3.

Yeast strains and plasmids

Strain or plasmid	Genotype or description	Reference or source
Strains
TB50a	MATaleu2-3,112 ura3-52 rme1 trp1 his3 HMLa	Hall collection
TB50a/α	MATa/α leu2-3,112/leu2-3,112 ura3-52/ura3-52 rme1/rme1 trp1/trp1 his3/his3 HMLa/HMLa	Hall collection
JK9-3da/α	MATa/α leu2-3,112/leu2-3,112 ura3-52/ura3-52 rme1/rme1 trp1/trp1 his4/his4 HMLa/HMLa	Hall collection
AN9-2a	MATaleu2-3,112 ura3-52 rme1 trp1 his3 tor1::kanMX6 HMLa
LJ25-1A	MATaleu2-3,112 ura3-52 rme1 trp1 his3 pmr1Δ::kanMX4 HMLa	10
LJ25-2C	MATα leu2-3,112 ura3-52 rme1 trp1 his3 pmr1Δ::kanMX4 HMLa	10
VA19	MATaleu2-3,112 ura3-52 rme1 trp1 his3 KOG18XGFP::URA3 HMLa	This study
VA33	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1D330-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA34	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1D330-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA35	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1D330-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA36	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1D67-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA38	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1D67-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA41	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1N-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA43	MATa/α leu2-3,112/leu2-3,112 ura3-52/ura3-52 rme1/rme1 trp1/trp1 his3/his3 TOR2N321-3XGFP/ TOR2 HMLa/HMLa	This study
VA66	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1D330-3XGFPHMLa	This study
VA68-9c	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR1D330-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA68-9c	MATaleu2-3,112 ura3-52 rme1 trp1 his3 pmr1Δ::kanMX4 TOR1D330-3XGFPHMLa	This study
YGD25	MATaleu2-3,112 ura3-52 rme1 trp1 his3 tor1Δ::LEU2-4 pmr1Δ::kanMX4 HMLa	9
VA86	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR2N321-3XGFPHMLa/pSH47 [CEN URA3]	This study
VA102	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR2N321-3XGFPHMLa	This study
VA109	MATaleu2-3,112 ura3-52 rme1 trp1 his3 OG18XGFP::URA3 HMLa	This study
VA121	MATaleu2-3,112 ura3-52 rme1 trp1 his3 TOR2N321-3XGFPKOG18XGFP::URA3 HMLa	This study
YGD25	MATaleu2-3,112 ura3-52 rme1 trp1 his3 tor1Δ::LEU2-4 pmr1Δ::kanMX4 HMLa	9
Plasmids
pSH47	pGAL-cre [CEN URA3]	16
pTA19	Integrating plasmid KOG1 (nt 3000-4671)-8XGFP:URA3	2
pTPQ128	pADH1-Sec7-dsRed [CEN LEU2]	31
pTPQ127	pGPD1-FYVE-dsRed [CEN LEU2]	31

As a second proof, we repeated the PCR with a primer specific for the GFP in pOM3. Using TOR1/Fwd/−72 and GFP/Rev, three bands corresponding to the expected 456-, 1,176-, and 1,896-nt products for 3XGFP insertion were all clearly present with strain VA38 (Table 3). VA34 was confirmed by observation of the expected 372-, 1,092-, and 1,896-products by using TOR1/Fwd/+801 and GFP/Rev. A PCR for VA38 with TOR1/Fwd/−72 and GFP/Rev gave the expected 260- and 980-nt products but not the 1,700-nt product in VA38, due to preferential amplification of smaller products.

As a third and final proof, correct integration was also confirmed by sequencing. Genomic DNA was purified after cell breakage by a phenol-chloroform method and amplified by PCR with TOR1/Fwd/+801 (for VA34 and VA35), TOR1/Fwd/−72 (for VA34 and VA41), and GFP/Rev. The band from priming to the first GFP was purified and sequenced. VA33 and VA34/VA35 are from independent primary colonies.

Isolation of proteins and Western blot analysis.

A crude supernatant protein (1,500 × g for 1 min) was isolated by a method using glass beads for the breakage of cells. The lysis buffer was ice-cold phosphate-buffered saline, 10% glycerol, and 0.5% Tween 20 with inhibitors (1.25 μg/ml leupeptin, 0.75 μg/ml antipain, 0.25 μg/ml chymotrypsin, 0.25 μg/ml elastinol, 5 μg/ml pepstatin, 1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA). Protein (40 μg), in sodium dodecyl sulfate sample buffer, was loaded on a 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel and used for Western blot analysis (6). A mixture of two mouse monoclonal antibodies (clones 7.1 and 13.1) to GFP (Roche Applied Science) was used at a dilution of 1:1,000. The antibody detection system (ECL kit) was from Amersham.

Kog1-8XGFP strains.

Plasmid pTA19 contains a portion (3,000 to 4,671 nt) of the KOG1/LAS24 ORF at the HindIII and XbaI sites of p8XGFPIU (2). TB50a was transformed with SalI-treated pTA19, generating strain VA19 after colony selection on SD-ura (Table 3). To generate strains with both Kog1-GFP and TOR1-3XGFP, VA109 (constructed similarly to VA19 but in TB50α) was crossed with VA66 and the desired segregant (VA121) was chosen.

Spottings for growth assays.

Strains were grown overnight in liquid YPD (5-ml culture), collected by centrifugation, and washed once with water, and 0.1 ml was used to inoculate SD (5 ml) overnight. Cells were adjusted in concentration to an optical density at 600 nm of 1.0, then serially diluted 1:10 in a 96-well plate, and transferred to plates with a pinning device.

FM4-64 staining.

Cells were grown in YPD to an A₆₀₀ of 0.8. Cells (20 to 40 A₆₀₀ units ml⁻¹ in YPD medium) were incubated on ice for 30 min with 30 μmol liter⁻¹ FM4-64 dye (Molecular Probes Inc.), washed once with YPD, and incubated for 60 min for steady-state experiments as described previously (38).

Microscopy.

Cells were imaged while in log phase, after immobilization on slides coated beforehand with concanavalin A. Microscopy of 3XGFP strains was performed with a Zeiss Axioplan 2 microscope equipped with an MRm camera (Carl Zeiss, Aalen Oberkochen, Germany) and a Plan Apochromat 63×/1.40-numerical-aperture objective for oil. Zeiss Axiovision 3.1 software was used to control filters and to acquire images once a field was chosen in differential interference contrast (DIC). Exposure settings for GFP in different figures varied between 800 and 2,000 ms except where noted; comparisons for intensity are valid within figures but not between figures. Images were analyzed with Axiovision software (Zeiss), with equal adjustments for all images (control and 3XGFP strains) taken in each experiment. Axiovision files were exported in TIFF format and cropped in Photoshop without further modification of brightness or contrast. For confocal imaging, we used an Olympus 1X81 spinning-disk microscope and a Plan Apochromat 60×/1.42-numerical-aperture oil objective, and we used IQ Andor software to collect and deconvolute Z-stack images. Deconvoluted images were exported into ImageJ and then to Photoshop.

RESULTS

Identification of a region in TOR1 for internal 3XGFP tagging.

Visualization of low-abundance proteins, such as TORs, often requires the incorporation of multiple tags. The incorporation of multiple copies can be either at the N terminus or at the C terminus of the protein in question. However, in our experience (data not shown), neither TOR1 nor TOR2 can tolerate the fusion of a tag to either the N or C terminus. Consistent with these previous observations, TOR1 or TOR2 with 3XGFP at the N terminus (encoded by TOR1^N-3XGFP or TOR2^N-3XGFP, respectively) was nonfunctional (see Fig. 5 and 8). TOR contains multiple subdomains, some of which have not been fully characterized but are important for interactions with conserved components of TOR complexes. We reasoned that an internal region permissive for the insertion of 3XGFP might be revealed by a comparison of evolutionary divergent TOR proteins. To identify a candidate region for the insertion of 3XGFP, we aligned six TOR proteins (viz., human mTOR [mammalian TOR], Saccharomyces cerevisiae TOR1 and TOR2, Schizosaccharomyces pombe Tor2, Caenorhabditis elegans TOR, and Cryptococcus neoformans TOR) (Fig. 1). We found a variable region near D330 in the noncatalytic domain of S. cerevisiae TOR1. This region in some of the TORs contains natural insertions, and D330 in TOR1 was thus viewed as potentially permissive for the insertion of 3XGFP. To test this hypothesis, we generated three independent strains (VA33, VA34, and VA35), all of which contain an in-frame cassette encoding 3XGFP replacing D330 in TOR1 (see Materials and Methods and Table 3). The 3XGFP insertion is shown by the alignments in Fig. S1 in the supplemental material. We used three different phenotypic tests to determine the functionality of the new TOR1^D330-3XGFP allele encoding TOR1-3XGFP, as described directly below.

FIG. 5.

TOR1^D67-3XGFP is partially functional. (A) Growth on YPD containing 1 nM rapamycin. The indicated strains (Table 3) were streaked, and after 3 days of growth, the plate was scanned. Growth patterns on YPD from streaks in the same experiment were similar (data not shown). (B) Growth at 15°C (15 deg) on YPD or YPD plus 1 nM rapamycin (Rap). The indicated strains were assayed by a 10-fold dilution and pinning; TOR1^D67-3XGFP (VA38) was used. The plates were scanned after 4 days.

FIG. 8.

TOR2 is functional if N321 is replaced by 1X-, 2X-, or 3XGFP. (A) A 441-nt product (asterisk) establishes the insertion of GFP in germinated spores with 2:2 segregation (see the text). Primers TOR2/+711 and GFP/Rev and spores 2A to 2D (template) were used for colony no. 1 (TB50a/α background). (B) Results of PCR consistent with 2XGFP replacing N321 after the recombination event in JK9 colony no. 1, 1XGFP in TB50 colony no. 2, and 3XGFP in TB50 colony no. 1 (see the text). Primers TOR2/Fwd/+931 and TOR2/Rev/+990 were used. The templates used were as follows: 2XGFP-JK9 lanes, 2A, 2B, none, and 20C; 1XGFP-TB50 lanes, 18A, 18B, 20A, and 20B (colony no. 2); and 3XGFP-TB50 lanes, 2A, 2B, 7C, and 7D (colony no. 1). Arrows indicate diagnostic bands for 1X-, 2X-, and 3XGFP (see the text). (C) Western blot with anti-GFP of 40 μg of lysate protein (see Materials and Methods) from the control (Ctl), TB50a, and VA102 strains.

FIG. 1.

The alignment of evolutionary divergent species identifies a region of TOR for the insertion of 3XGFP. Proteins were aligned by ClustalW (7). Only the relevant portion is shown. Residues D67 and D330 in TOR1 and N321 in TOR2 are indicated by asterisks. Arrows indicate sites where 3XGFP was inserted at D330 (TOR1) or N321 (TOR2) with retention of function. Humans, Caenorhabditis elegans, and Cryptococcus neoformans have only one TOR. Default colors and alignment parameters are from the implementation of ClustalW (7) at http://npsa-pbil.ibcp.fr/.

TOR1-3XGFP encoded by TOR1^D330-3XGFP is functional.

TOR1-3XGFP was expressed as an intact protein with an apparent mass greater than 250 kDa (Fig. 2A). A phenotype for the loss of TOR1 is rapamycin hypersensitivity. Rapamycin hypersensitivity of tor1Δ cells is due to a specific loss of TOR1 function because point mutations that inactivate TOR1 kinase activity cause hypersensitivity to a low concentration (1 nM) of rapamycin (33). The three independent strains of TOR1^D330-3XGFP were not hypersensitive to 1 nM or 2 nM rapamycin (Fig. 2B and C) and grew equivalently in the absence of rapamycin (Fig. 2A), indicating that TOR1-3XGFP was expressed and functional. A higher concentration of rapamycin (5 nM) nearly completely inhibited the growth of the TOR1^D330-3XGFP and TOR1 strains as expected (data not shown).

FIG. 2.

TOR2-3XGFP is expressed, and TOR1^D330-3XGFP strains have normal sensitivity to rapamycin. (A) Western blot with anti-GFP antibodies of 40 μg of total protein (see Materials and Methods) of the control (Ctl), TB50a, and VA34 strains. The indicated strains were streaked onto YPD (B), YPD plus 1 nM rapamycin (Rap) (C), and YPD plus 2 nM rapamycin (D). Cells were grown for 2 days at 30°C after streaking and then scanned.

The loss of TOR1 causes a cold-sensitive growth defect (Fig. 3). Adaptation to cold stress is not rescued by a kinase-defective TOR1 (data not shown). TOR1^D330-3XGFP strains were not cold sensitive (Fig. 3A). In fact, TOR1^D330-3XGFP strains grew somewhat better at 15°C than the wild type. This enhanced growth potential of TOR1^D330-3XGFP strains at 15°C was also observed in the presence of 2 nM rapamycin (Fig. 3B). This result shows as well that Tor1^D330-3XGFP is expressed and functional because adaptation to cold stress requires active TOR1 kinase activity.

FIG. 3.

TOR1^D330-3XGFP strains adapt to cold stress. The indicated strains were streaked onto YPD (A) or YPD plus 2 nM rapamycin (Rap) (B) Cells were grown for 4 days at 15°C (15 deg C) after streaking and then scanned.

Pmr1 is a P-type ATPase that localizes predominantly to Golgi bodies and transports Ca²⁺ or Mn²⁺ into the lumen from the cytoplasm (13). Pmr1 function is important for Ca²⁺ regulation in the secretory pathway and for the removal of toxic levels of Mn²⁺. TOR1 has a genetic interaction with pmr1Δ (9). A pmr1Δ strain does not grow in the presence of 2 mM Mn²⁺, and growth is rescued by the loss of TOR1. We found that a TOR1^{D330-3XGFP pmr1}Δ strain is as sensitive to Mn²⁺ as a TOR1 pmr1Δ strain (Fig. 4). Thus, as assayed by three separate phenotypic tests, TOR1-3XGFP encoded by TOR1^D330-3XGFP is functional like wild-type TOR1.

FIG. 4.

TOR1^D330-3XGFP, like TOR1, is sensitive to Mn²⁺ in a pmr1Δ strain. The indicated strains were streaked onto YPD (A) or YPD containing 2 mM Mn²⁺ (final concentration) (B). Cells were grown for 2 days at 30°C, and then plates were scanned. The TOR1 PMR1 (TB50a), tor1Δ PMR1 (AN9-2a), tor1Δ pmr1Δ (YGD25), TOR1 pmr1Δ (LJ25-1A), TOR1^{D330-3XGFP pmr1}Δ (VA68-9c), and TOR1^{D330-3XGFP PMR1} (VA34) strains were used (Table 3).

For comparison to the insertion at D330, we also constructed 3XGFP strains that targeted residue D67, nearer to the N terminus of TOR1, where there is significantly more divergence in sequence than at D330 (Fig. 1). Notably, TOR1 is functional with 2XMyc but not higher numbers of Myc proteins, placed between residues 86 and 87 in this N-terminal segment (27; A. Lorberg, personal communication). We found that TOR1^D67-3XGFP was only partially functional because it conferred only partial resistance to 1 nM rapamycin (Fig. 5A). The function of TOR1^D67-3XGFP was greater than that of TOR1^N-3XGFP, as reflected in the rapamycin or cold sensitivity of the corresponding strains. The TOR1^D330-3XGFP strain was more functional for growth in the presence of rapamycin or at 15°C than either the TOR1^N-3XGFP or TOR1^D67-3XGFP strain (Fig. 5B).

Kog1 is an essential protein that binds TOR1 or TOR2 in TORC1 (27). Strains containing Kog1 with 3XGFP or 8XGFP incorporated at the C terminus as genomic tags are viable (2, 37). We confirmed that KOG1^C-8XGFP cells (VA19) were not hypersensitive to 1 nM rapamycin (Fig. 5B).

TOR1-3XGFP is diffusely cytoplasmic and concentrated at discrete sites near vacuolar membranes.

TOR1-3XGFP encoded by TOR1^D330-3XGFP was visualized in live cells. In cells grown in rich medium (YPD), TOR1-3XGFP was dispersed throughout the cytoplasm and was concentrated near the vacuolar membrane, sometimes as a dot (Fig. 6). TOR1-3XGFP was also observed, although rarely, in dots near the plasma membrane (data not shown). The vacuole (Fig. 6A) was identified by autofluorescence of its contents at red wavelengths and by morphology in DIC. Vacuolar autofluorescence was more prominent when cells were grown in rich medium (YPD). We confirmed the perivacuolar localization of the TOR1-3XGFP signal by comparing it to that of FM4-64, which specifically stains the vacuolar membrane at steady state (Fig. 6B) (38). A portion of TOR1-3XGFP overlapped with the expected ring-like staining of FM4-64 marking the vacuolar membrane.

FIG. 6.

TOR1^D330-3XGFP is predominantly cytoplasmic and concentrated as dots near the vacuolar membrane. (A) Localization of TOR1-3XGFP in cells grown in YDP. Merged GFP (green; GFP channel) and autofluorescence images for TB50a (control) lacking a GFP cassette and TOR1-3XGFP (VA34) strains are shown. Strains were grown overnight in YPD, diluted, and imaged while still in log phase after centrifugation and suspension in synthetic medium. The arrow indicates a dot near the vacuole, and the feathered arrow indicates a dot near the vacuolar membrane. (B) The TOR1-3XGFP signal overlaps FM4-64 staining (see Materials and Methods) of the vacuolar membrane. Control, strain TB50a lacking a GFP cassette; TOR1-3XGFP, strain VA102. The exposure settings used were as follows: DIC, 300 ms; GFP, 10,000 ms; and Fm4-64 (Cy3), 10,000 ms.

The predominant localization of KOG1-GFP to the vacuolar membrane has been reported previously (2, 37). We compared TOR1-3XGFP and Kog1-GFP strains to a strain containing both GFP fusions (VA121) (see Fig. S2 in the supplemental material). The perivacuolar signal for GFP in the double-GFP strain was dramatically increased (see Fig. S2 in the supplemental material) compared to that in either single-GFP strain. TOR2-3XGFP, studied for comparison, did not show perivacuolar localization (see Fig. S2 in the supplemental material).

The localization of TOR1-3XGFP was compared to that of Sec7-dsRed or FYVE-dsRed (Fig. 7) (31). Cells were grown in selective synthetic medium to maintain the plasmid encoding the dsRed marker. Sec7 is a high-molecular-weight protein that contains a guanine-nucleotide exchange activity for Arf proteins involved in Golgi function (8). Sec7-dsRed is a marker for the trans-Golgi (28). The localization of TOR1-3XGFP was qualitatively different from that of Sec7-dsRed. First, the Sec7-dsRed vesicles were more numerous than the dots of TOR1-3XGFP, and second, the punctate signals for TOR1-3XGFP (Fig. 7) did not exactly correspond. A closer correspondence was observed between TOR1-3XGFP and the FYVE-dsRed marker.

FIG. 7.

TOR1^D330-3XGFP localization is distinct from that of Sec7-dsRed and similar to that of FYVE-dsRed. We performed microscopy with the following strains in SD or SD-leu (top to bottom): TB50a (control strain, in SD), TOR1^D330-3XGFP (V66, in SD), TOR1^D330-3XGFP plus FYVE-dsRed (VA66 transformed with pTPQ127, in SD-leu), and TOR1^D330-3XGFP plus Sec7-dsRed (VA66 transformed with pTPQ128, in SD-leu). The arrow shows an example of colocalization of TOR1^D330-3XGFP with an FYVE domain marker in a punctate structure. The exposure settings used were as follows: DIC, 100 ms; GFP, 2,000 ms; and dsRed, 500 ms.

The FYVE-dsRed fusion protein binds phosphatidylinositol-3-phosphate (for a review, see reference 23) and is generally a marker for early endosomes but is also found near the vacuole. FYVE-dsRed localizes “to punctate structures adjacent to the vacuole, weakly on the vacuole-limiting membrane, and in some cases within the vacuole” (18). We emphasize this description because TOR1-3XGFP localization was similar to this description with regard to the vacuole. TOR1-3XGFP also appeared to localize very near to, if not within, the FYVE-dsRed punctate structures near the vacuole. However, TOR1-3XGFP localization was distinct from FYVE-dsRed localization in that TOR1-3XGFP was also diffusely cytoplasmic and not all the dot-like concentrations of TOR1-3XGFP were found near the vacuole.

TOR2 is functional when 3XGFP is inserted to replace N321.

Our success with TOR1^D330-3XGFP encouraged us to generate GFP fusion alleles of TOR2 (Fig. 8). We targeted the N terminus and residue N321 of TOR2. Like D330 in TOR1, N321 in TOR2 corresponds to a variable region in the noncatalytic domain of TOR2. The TOR2-targeted 3XGFP cassette was introduced into diploid strains because TOR2 is essential. To assess the functionality of the GFP fusion proteins, diploids containing the desired TOR2^N-3XGFP or TOR2^N321-3XGFP allele were sporulated, dissected, and germinated. The TOR2^N-3XGFP allele was nonfunctional because only two spores were viable in 13/13 tetrads dissected (see Fig. S3 in the supplemental material). In contrast, TOR2^N321-3XGFP was functional (see Fig. S3 in the supplemental material). Nineteen of 23 dissected tetrads for TOR2^N321-3XGFP produced four viable spores. An expected diagnostic PCR product (441 nt) was observed with 2:2 segregation for all tetrads analyzed, and the product was absent in control cells (Fig. 8A).

The 441-nt product is that expected from the reverse primer amplifying at the closest site in the first GFP sequence. To further assess this, we performed a colony PCR with primers chosen close to and flanking the N321 site (Fig. 8B). One candidate had the complete 3XGFP cassette because the principal product was an ∼2.3-kb band and was chosen for further study (becoming VA102). The other candidates were 2XGFP or 1XGFP. TOR2-3XGFP in VA102 was expressed as a >250-kDa protein by Western blotting with anti-GFP antibodies (Fig. 8C). These results together demonstrate that TOR2 remains functional despite the insertion of 3XGFP at N321 whereas TOR2 with 3XGFP at the N terminus is nonfunctional.

TOR2-3XGFP localizes to punctate structures near the plasma membrane.

The localization of TOR2-3XGFP (encoded by TOR2^N321-3XGFP in VA102) was compared with that of Sec7-dsRed or FYVE-dsRed under conditions similar to those used for TOR1-3XGFP (Fig. 9). TOR2-3XGFP did not colocalize with either Sec7-dsRed or FYVE-dsRed, and TOR2-3XGFP was not found concentrated near the vacuolar membrane. Instead, TOR2-3XGFP was detectable above the background in punctate structures. These structures were most apparent beneath the plasma membrane.

FIG. 9.

TOR2^N321-3XGFP localizes to punctate structures near the plasma membrane. We performed microscopy with the following strains in SD or SD-leu (top to bottom): TB50a (control strain, in SD), TOR1-3XGFP (V102, in SD), TOR1-3XGFP plus FYVE-dsRed (VA102 transformed with pTPQ127, in SD-leu), and TOR2-3XGFP plus Sec7-dsRed (VA102 transformed with pTPQ128, in SD-leu). The VA102 strain lost the pSH47 (URA3) plasmid by 5-fluoroorotic acid treatment (Table 3).

To better resolve these structures, we used confocal microscopy (Fig. 10; see also Fig. S4 in the supplemental material). The TOR2-3XGFP signal was cytoplasmic and concentrated in punctate structures at or very near the plasma membrane. TOR1-3XGFP was cytoplasmic and concentrated near the vacuole. The identities of the TOR structures near the plasma membrane are unknown, although they resemble eisosomes in location and appearance but are less numerous than eisosomes. Eisosomes have recently been defined and characterized as punctate structures at the plasma membrane involved in the early steps of endocytosis (36, 39).

FIG. 10.

TOR1 and TOR2 have distinct localization patterns by confocal microscopy. The control (TB50), TOR1-3XGFP (VA66), and TOR2-3XGFP strains grown in YPD were imaged by confocal microscopy (see Materials and Methods). The exposure settings used were as follows: Tor1, 400 ms GFP and 50% laser intensity; Tor2, 800 ms GFP and 70% laser intensity; and TB50, 800 ms GFP and 70% laser intensity.

DISCUSSION

We constructed functional TOR1-3XGFP and TOR2-3XGFP proteins to localize TOR in living cells. Importantly, this required the insertion of GFP within the TOR1 and TOR2 ORFs rather than at the N or C terminus. TOR1-3XGFP was cytoplasmic and concentrated at a prevacuolar compartment and at the vacuolar membrane. TOR2-3XGFP was also cytoplasmic and, most strikingly, concentrated at the plasma membrane. We did not detectably observe TOR1-3XGFP in the nucleus, and the nucleus qualitatively often correlated with decreased staining versus the surrounding cytoplasm (data not shown). More studies will be required to address this. Imaging functional tagged proteins should be complementary to imaging proteins in fixed cells with antibodies. Fixation may cause a loss of vesicular localization of proteins, for example, Rheb (34). The various localization patterns of TORs may provide a molecular basis for the large number of processes controlled by TOR and may also explain why different studies have observed TOR in different cellular locations. The localization of TOR in mammals has also been ambiguous, possibly because mTOR (also known in literature as FRAP [FKBP12-rapamycin-associated protein[) is also found in different locations. mTOR has been reported to be principally cytoplasmic and to shuttle between the cytoplasm and nucleus (20), to be principally nuclear in some cancer cell lines (42), to be cytoplasmic and associated with Golgi bodies and the ER (26), or to be associated with a regulatory subunit of protein kinase A (RIα) on late endosomes and autophagosomes (30). Only the last study used GFP-tagged mTOR, but GFP-tagged mTOR was overexpressed and N terminally tagged with 1XGFP and its functionality remains to be established.

The concentration of TOR1-3XGFP near the vacuolar membrane is consistent with some reports in the literature. The loss of TOR1 is synthetically lethal with loss of class C VPS (vacuolar protein sorting) genes (43). This genetic interaction is likely specific to TOR1 because the overexpression of TOR2 fails to rescue the synthetic lethality. As a second correlation, the TOR1 interactors Kog1 and Tco89 were detected near the vacuolar membrane by imaging of Kog1-GFP in live cells or by immunogold staining for Tco89 in fixed cells (2, 32, 37). Because strain differences can affect TOR1-related phenotypes (9, 32), we confirmed that Kog1-8XGFP was also concentrated at the vacuolar membrane in TB50 (see the supplemental material). As a third correlation, TOR1 interacts genetically and biochemically with Gtr2 in the Ego complex, found near the vacuolar membrane (12). Finally, the phosphorylation of Sch9 by TORC1 may occur near the vacuole because Sch9 localizes near this organelle and an artificial Sch9 substrate tethered to the vacuolar membrane is phosphorylated in a TORC1-dependent manner (37). If there is a connection of TORC1 to the vacuole, it may derive from the role of the vacuole in nutrient supply or regulation of autophagy (19). The physiological relevance of TOR1 in the cytoplasm or at the plasma membrane also remains to be determined.

The localization of TOR2-3XGFP at discrete sites near the plasma membrane is similar to that of eisosomes, recently described as protein complexes important for endocytosis (39). Interestingly, TOR2 is found mainly in TORC2, which is implicated in endocytosis and actin dynamics (14). Furthermore, TOR2 activates the AGC family kinase Ypk2 (17). Ypk2 phosphorylates Pil1 and Lsp1, proteins involved in eisosome formation and function (29). Moreover, TORC2 is required for sphingolipid biosynthesis (4) and Ypk2 is also activated by sphingolipids (for a review, see reference 25). These findings make eisosomes an interesting candidate for the location of TOR2 at the plasma membrane. We attempted to address this by introducing an mCherry tag at the C terminus of LSP1 or SUR7 in the TOR2-3XGFP strain and found that (i) the Lsp1 and Sur7 mCherry signals are very much brighter than the TOR2-3XGFP signal and (ii) the TOR2-3XGFP signal showed partial colocalization with these markers (data not shown). Avo3 shows partial colocalization with Pil1 (an eisosome marker) by immunofluorescence of fixed cells (R. Shioda, unpublished data). The specific localization of TOR2-3XGFP to eisosomes remains to be defined.

Our findings provide insight into the structure of TOR. It is remarkable that TOR, a strongly conserved protein of ∼2,500 amino acids, retains at least partial function after the insertion of 748 amino acids (3XGFP cassette) in the noncatalytic domain. The functionality of the internally tagged TOR1-3XGFP and TOR2-3XGFP proteins may be due to the placement of 3XGFP between subdomains. The majority of the N terminus of TOR consists of repeated HEAT motifs (21). D330 and N321 are in a gap between HEAT repeats (see Fig. S5 in the supplemental material). Furthermore, a recent electron microscopy structure of TOR1 suggests that the noncatalytic domain forms an N-terminal head, a turn, and an arm (1). We predict that D330 and N321 are near a gap between two of these subdomains. The preservation of function with internal tagging of TOR with GFP may also be due to the fact that the N terminus and the C terminus of GFP are near each other (see Protein Data Bank entry 1EMM), minimizing displacement at the point of insertion.