OUP user menu

Accounting for strain-specific differences during RTG target gene regulation in Saccharomyces cerevisiae

Ivanka Dilova, Ted Powers
DOI: http://dx.doi.org/10.1111/j.1567-1364.2005.00008.x 112-119 First published online: 1 January 2006

Abstract

Mitochondrial dysfunction results in the expression, via the retrograde response pathway, of a concise set of genes (RTG target genes) that encode enzymes involved in the anapleurotic production of α-ketoglutarate. Inhibiting the rapamycin-sensitive TOR kinases, important regulators of cell growth, similarly results in RTG target gene expression under rich nutrient conditions. Retrograde and TOR-dependent regulation of RTG target genes requires a number of shared components, including the heterodimeric bZip/HLH transcription factors Rtg1p and Rtg3p, as well as their upstream regulator Mks1p. Two unresolved discrepancies exist with regard to the mechanism of RTG target gene control: (1) deletion of MKS1 results in constitutive expression of RTG target genes in most but not all strain backgrounds; and (2) RTG target gene expression has been correlated with both decreased as well as increased Rtg3p phosphorylation. Here we have addressed both of these issues. First, we demonstrate that the mks1 deletion strain used in a previous study by Shamji and coworkers contains a nonsense mutation within codon Ser 231 in RTG3 that likely accounts for the inactivity of the RTG system in this strain. Second, we confirm results by Butow and coworkers that Rtg3p is dephosphorylated as a primary response to induction of the pathway. Hyper-phosphorylation of this protein appears to be a secondary consequence of rapamycin treatment and is influenced both by strain background as well as by specific supplied nutrients. That hyper-phosphorylation of Rtg3p is also caused by heat shock suggests that it may reflect a more generalized response to cell stress. Together these results contribute toward a uniform view of RTG target gene regulation.

Keywords
  • rapamycin
  • TOR kinase
  • signal transduction
  • mitochondria
  • Rtg3p
  • Mks1p

Introduction

The rapamycin-sensitive Target of Rapamycin (TOR) kinases are central components of a conserved signaling network that is proposed to couple cell growth to nutrient-derived cues within all eukaryotic organisms examined to date, from yeast to humans (Shamji, 2003; Rohde & Cardenas, 2004; Martin & Hall, 2005). Saccharomyces cerevisiae contains two highly similar TOR proteins, Tor1p and Tor2p, both of which are inhibited by rapamycin (Martin & Hall, 2005). Results from several studies reveal that TOR controls the expression of a wide scope of genes in yeast, including genes involved in carbon and nitrogen metabolism (Beck & Hall, 1999; Cardenas, 1999; Hardwick, 1999; Komeili, 2000; Shamji, 2000; Crespo & Hall, 2003; Butow & Avadhani, 2004; Powers, 2004; Rohde & Cardenas, 2004). One such cluster of genes, termed RTG target genes, encodes mitochondrial, peroxisomal and cytoplasmic proteins whose expression is required for de novo biosynthesis of the amino acids glutamate and glutamine (Komeili, 2000; Butow & Avadhani, 2004; Powers, 2004). These genes have originally been identified as being under the control of a mitochondria-to-nucleus signaling pathway, or retrograde response pathway, that modulates their expression according to the respiratory state of the cell (Liao & Butow, 1993; Butow & Avadhani, 2004). Both TOR and retrograde-dependent control of the RTG system converge upon a heterodimer transcription factor complex composed of the bZip/HLH proteins Rtg1p and Rtg3p (Komeili, 2000; Sekito, 2000). Thus, in response to impaired mitochondrial function or following rapamycin treatment of cells grown in rich medium, the Rtg1p/Rtg3p complex translocates from the cytoplasm into the nucleus and activates their target genes (Komeili, 2000; Sekito, 2000). Regulated RTG target gene expression requires several additional proteins, including the cytoplasmic protein Mks1p, which controls the nucleocytoplasmic disposition of Rtg1p/Rtg3p in response to TOR- and/or retrograde-derived signals (Dilova, 2002; Sekito, 2002; Tate, 2002).

Although there has been, in general, substantial agreement among different investigators who have studied the mechanism of RTG target gene regulation, a number of unresolved discrepancies remain. For example, Butow and coworkers have demonstrated that Rtg3p is a multiply phosphorylated protein and that retrograde-dependent activation of the RTG system correlates with dephosphorylation of this protein (Sekito, 2000). In collaboration with the O'Shea laboratory, we demonstrated independently that Rtg3p is a phosphoprotein, yet we observed that rapamycin-induced RTG target gene expression correlated instead with an increase in Rtg3p phosphorylation (Komeili, 2000). Initially, we suggested that these differences might reflect distinct mechanisms for retrograde- vs. TOR-based control of Rtg3p function (Komeili, 2000). However, Butow and coworkers subsequently observed that rapamycin also induces Rtg3p dephosphorylation (Sekito, 2002), indicating that differences in strain background and/or experimental approach are likely to be involved. A second discrepancy is with respect to the role of Mks1p in RTG target gene regulation. Three independent studies had observed that deletion of MKS1 results in constitutive activation of the pathway, leading to the conclusion that Mks1p is a negative regulator of the pathway (Dilova, 2002; Sekito, 2002; Tate, 2002). By contrast, a fourth study (Shamji, 2000) reported that RTG target gene expression was impaired in an mks1Δ strain following rapamycin treatment, suggesting instead that Mks1p is an important positive regulator of the system. Based on our own analysis of this same mks1Δ strain isolate used by Shamji. (2000) (here termed Σ1278SSmks1Δ), we suggested that this strain was likely to have acquired one or more additional mutations that blocked activation of the pathway, even in the absence of the repressive effects of Mks1p (Dilova, 2002). Here we report that this mks1Δ strain contains a nonsense mutation in RTG3, which we conclude accounts for its aberrant behavior in the study by Shamji. (2000). We also present data that demonstrate that the phosphorylation state of Rtg3p is remarkably variable in its response to rapamycin treatment, and depends on strain background as well as on precise media conditions. Nevertheless, our results are consistent with dephosphorylation of Rtg3p representing an important step for RTG target gene activation.

Materials and methods

Strains, media and general methods

All strains of Saccharomyces cerevisiae used in this study are listed in Table 1. The following culture media were used: YPD [1% yeast extract (YE), 2% bactopeptone (BP) and 2% dextrose]; YPD+Gln (same as YPD, but additionally contains 0.2% glutamine); MDN (0.8% yeast nitrogen base without amino acids, 2% dextrose, pH 5.5); MDN+Gln (same as MDN, but additionally contains 0.2% glutamine); MDN+Glu (same as MDN, but additionally contains 0.2% glutamate); SCD (0.7% yeast nitrogen base without amino acids, 2% dextrose, 0.006% adenine sulfate, 0.003% uracil, 0.008% tryptophan, 0.006% histidine, 0.002% arginine, 0.002% methionine, 0.003% tyrosine, 0.008% leucine, 0.003% isoleucine, 0.003% lysine, 0.005% phenylalanine, 0.010% glutamic acid, 0.010% aspartic acid, 0.015% valine, 0.020% threonine and 0.020% serine); and SCD minus Leu (same as SCD, but without leucine). Concentrations of additional amino acids and nucleotides added to media are as specified in the legends to the figures and tables. Rapamycin (Sigma-Aldrich, St Louis, MO) was dissolved in dimethyl sulfoxide (DMSO) and added to a final concentration of 0.2 μg mL−1.

View this table:
1

Strains of Saccharomyces cerevisiae used in this study

StrainDescriptionReference/source
W303aMATα ade2-1 trp1-1 can1-100 leu2-3 112 his3-11 15 ura3 GAL+Nasmyth. (1990)
W303αMATα ade2-1 trp1-1 can1-100 leu2-3 112 his3-11 15 ura3 GAL+Nasmyth. (1990)
PLY190Same as W303a, except mks1::LEU2Dilova. (2002)
PLY186Same as W303a, except rtg3::TRP1Dilova. (2002)
PLY188Same as W303α, except rtg2::TRP1 mks1::LEU2Dilova. (2002)
Σ1278SSMatα ura2 leu2::HISGShamji. (2000)
Σ1278SSmks1ΔSame asΣ1278SS, except mks1::G418Shamji. (2000)
PLY481Same as W303a, except RTG3-MYC:His3MX6This study
PSY142Mata ura3-52 leu2-2,112 lys2-801Liao & Butow (1993)

Plasmid construction

Plasmids expressing RTG1 or RTG3 were constructed by PCR to amplify each gene using Σ2000 (Microbia, Cambridge, MA), genomic DNA as a template and Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA). Primers for amplification of RTG1 were designed such that the entire open reading frame, as well as 550 bp upstream and 70 bp downstream, was amplified. Primers for amplification of RTG3 were designed such that the entire open reading frame, as well as 520 bp upstream and 200 bp downstream, was amplified. PCR fragments were cloned in pCR-Blunt II-TOPO plasmid (Invitrogen, Carlsbad, CA). In a final step, RTG1 and RTG3 genes were excised and subcloned into pRS315 (Sikorski & Heiter, 1989) to generate plasmids pPL135 and pPL136, respectively. The functionality of these plasmid-expressed genes was confirmed by their ability to suppress the glutamine auxotrophy of appropriate rtg1Δ and rtg3Δ strains.

Construction of yeast strains

A derivative of W303a (PLY481) that expresses Rtg3p with multiple copies of the Myc epitope at its carboxy terminus was created by a one-step PCR-based gene tagging technique, using plasmid pFA6a-13Myc-His3MX6 as a template for the Myc gene sequence (Longtine, 1998). This strain grew as well as its untagged parent strain on MDN agar plates that lacked both glutamate and glutamine, demonstrating that the presence of the epitope tag did not interfere with Rtg3p function.

Northern blotting

Northern blot analysis was carried out as described previously (Powers & Walter, 1999; Komeili, 2000; Dilova, 2004). Hybridization to actin mRNA (ACT1) was used as a loading control and for normalization of CIT2 expression levels.

Western blotting

Western blot analysis was carried out essentially as described previously (Sekito, 2000; Dilova, 2004): 10 mL of cells grown to mid-logarithmic phase (OD600=0.5), were harvested by centrifugation [Allegra centrifuge (Beckman Coulter, Fullerton, CA), 3000 r.p.m.] and total protein extraction was performed using an NaOH extraction procedure described by Butow and coworkers (Sekito, 2000). Specificially, cells were lysed in 0.255 M NaOH/1% 2-mercaptoethanol and proteins were precipitated by treatment with trichloroacetic acid (TCA) at a final concentration of 6.1%. The protein pellet was then washed with 1 M Tris-HCl (pH 6.8) and resuspended in SDS PAGE sample buffer [65 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 100 mM DTT, 0.002% bromophenol blue]. Proteins from equivalent amounts of cell extract were resolved on 7.5% SDS-PAGE gels and transferred to nitrocellulose membranes (Protran, Intermountain Scientific, Kaysville, UT). Membranes were probed using either polyclonal antibody raised against Rtg3p or monoclonal anti-Myc antibody (Covance, Princeton, NJ) and visualized by chemiluminescence (PerkinElmer, Boston, MA). Equal loading of protein samples was confirmed by Ponceau staining of membranes prior to incubation with antibody.

DNA sequence analysis

The sequence of the RTG3 gene was determined for strains Σ1278SSmks1Δ and Σ1278SS by PCR amplification of the promoter, ORF and downstream regions, followed by DNA sequence analysis using a 3730 Capillary Electrophoresis Genetic Analyzer (Applied Biosystems, Foster City, CA).

Results

Identification of an RTG3 mutation in strain Σ1278SS mks1Δ

The strain Σ1278SSmks1Δ is unable to grow on minimal dextrose (MDN)+ ammonia agar plates unless supplemented with glutamine (or glutamate), a hallmark of a deficient RTG pathway (Fig. 1a). This is in contrast to an mks1Δ deletion strain made in the W303 background, which grows as well as either the isogenic wild-type W303 strain or the wild-type Σ1278SS strain on MDN agar plates (Fig. 1a). To gain insight into the genetic basis for the glutamine auxotrophy of Σ1278SSmks1Δ, we crossed this strain with the W303-derived mks1Δ strain to form a homozygous mks1Δ/mks1Δ diploid that was of mixed genetic background. This strain grew as well as the haploid W303 mks1Δ strain on MDN agar plates, demonstrating that the glutamine auxotrophy of Σ1278SSmks1Δ was recessive (Table 2). Sporulation of the diploid, followed by tetrad dissection, revealed that the glutamine auxotrophy segregated with a 2 : 2 ratio (eight out of eight tetrads examined; data not shown). We conclude from these results that the glutamine auxotrophy of Σ1278SSmks1Δ is caused by the presence of a single recessive mutation.

Figure 1

Determining the molecular basis for the glutamine auxotrophy of Σ1278SSmks1Δ. (a) Serial dilutions of wild-type (WT) and mks1Δ strains in both the W303a and Σ1278SS strain backgrounds on MDN or MDN+Gln agar plates. Agar plates contained the following auxotrophic supplements: 0.006% adenine sulfate, 0.003% uracil, 0.008% leucine, 0.006% histidine and 0.005% tryptophan. (b) Σ1278SS WT and Σ1278SSmks1Δ were transformed with a control vector (pRS315) or with plasmids that expressed either RTG1 (pPL135) or RTG3 (pPL136). Cells were grown in SCD media lacking leucine, and serial dilutions were plated onto MDN or MDN+Gln agar plates. Agar plates contained the following auxotrophic supplements: 0.006% adenine sulfate, 0.003% uracil, 0.006% histidine and 0.008% tryptophan. (c) Northern blot analysis of Σ1278SS and Σ1278SSmks1Δ transformed with plasmids described in (b). Cells were grown to mid-log phase (0.5 OD600 mL−1) in SCD media lacking leucine and were treated with rapamycin (0.2 μg mL−1 final concentration) for 30 min, where indicated. Cells were harvested and processed for Northern blot analysis, probing for the specified mRNAs. (d) Sequence analysis of RTG3 in Σ1278SS WT vs. Σ1278SSmks1Δ reveals the presence of a nonsense mutation at codon Ser 231 in Σ1278SSmks1Δ. (e) Western blot analysis of endogenous Rtg3p in strains Σ1278SS and Σ1278SSmks1Δ and, for comparison, an rtg3Δ strain (PLY186). Cells were grown to mid-log phase (0.5 OD600 mL−1) in YPD media followed by western blot analysis, using polyclonal anti-Rtg3p antiserum.

View this table:
2

Testing strains for growth in the presence or absence of glutamine

StrainGrowth on MDN+GlnGrowth on MDN
W303aYesYes
PLY190 (mks1Δ)YesYes
PLY186 (rtg3Δ)YesNo
Σ1278SSYesYes
Σ1278SSmks1ΔYesNo
Σ1278SS crossed with PLY190YesYes
Σ1278SSmks1Δ× PLY190YesYes
  • * Indicated strains were plated onto solid medium containing MDN or MDN+Gln, incubated for 3 days at 30°C and then scored for growth. MDN media used for this experiment contained the following auxotrophic supplements: 0.006% adenine sulfate, 0.008% tryptophan, 0.008% leucine, 0.006% histidine and 0.003% uracil.

We took a candidate-based approach and tested whether the glutamine auxotrophy of Σ1278SSmks1Δ was due to a mutation in either RTG1 or RTG3. Both genes were cloned from an unrelated wild-type strain and introduced into Σ1278SS and Σ1278SSmks1Δon low-copy (CEN/ARS) plasmids, under control of their native promoters, as described in the Materials and methods section. Control experiments demonstrated that these RTG1 and RTG3 plasmids could rescue the glutamine auxotrophies of rtg1 or rtg3 deletion strains, respectively, demonstrating that the plasmids produced functional Rtg1p and Rtg3p proteins (data not shown). We observed that introduction of plasmid-borne RTG3 but not RTG1 rescued the glutamine auxotrophy of Σ1278SSmks1Δ (Fig. 1b). Consistent with this observation, Northern blot analysis demonstrated that constitutive expression of the representative RTG target gene CIT2 was specifically restored when the Σ1278SSmks1Δ strain carried the RTG3 but not the RTG1 plasmid (Fig. 1c). From these results, we conclude that impaired RTG3 function is responsible for the glutamine auxotrophy of Σ1278SSmks1Δ.

To identify the specific mutation(s) in RTG3 in Σ1278SSmks1Δ, we sequenced (1) the entire RTG3 open reading frame, (2) 500 nucleotides upstream from the translational start site, and (3) 200 nucleotides downstream from the stop codon, both from this strain and, for comparison, from wild-type Σ1278SS. A single nucleotide difference was observed, namely a C to A transversion within codon 231 that is predicted to result in a premature stop codon within the N-terminus of Rtg3p (Fig. 1d). Consistent with this result, western blot analysis of extracts prepared from both strains revealed the presence of full-length Rtg3p in wild-type Σ1278SS but not in Σ1278SSmks1Δ cells (Fig. 1e). We note that no shorter protein species was recognized by anti-Rtg3p antibody in Σ1278SSmks1Δ cells (Fig. 1e). Thus, we conclude that Σ1278SSmks1Δ is likely to produce a truncated and possibly unstable form of Rtg3p and that this accounts for the deficient RTG phenotype exhibited by this strain.

Complexity in Rtg3p phosphorylation and its response to rapamycin

We next sought to account for the differences between our finding that rapamycin treatment results in hyperphosphorylation of Rtg3p and observations made by Butow and coworkers that this protein becomes dephosphorylated following drug treatment (Komeili, 2000; Sekito, 2000). We considered a number of possible sources for this apparent discrepancy, including differences in strain background and media composition, as well as methods of protein extract preparation and/or protein detection. To address these possibilities, we compared directly the effect of rapamycin treatment on Rtg3p phosphorylation in both wild-type strains used in these prior studies, W303 and PSY142, where changes in the mobility of endogenous Rtg3p on SDS-PAGE gels were monitored using anti-Rtg3p polyclonal antiserum (generously provided by R. Butow). We examined both types of media used in the previous studies, namely rich (YPD) and minimal (MDN+Glu), and otherwise employed conditions used by Butow and coworkers (namely, 30 min of rapamycin treatment and a rapid NaOH method of protein extract preparation; Sekito, 2000). Results of this analysis are presented in Fig. 2a.

Figure 2

Comparison of Rtg3p phosphorylation in strains W303 and PSY142. (a) and (b) Western-blot analysis of Rtg3p (left-hand panels) and Northern blot analysis of ACT1, RPL30 and CIT2 (right-hand panels) in W303a vs. PSY142. In (a), strains were grown to mid-log phase (0.5 OD600 mL−1) in either YPD or MDN+Glu media and, where indicated, treated with rapamycin (0.2 μg mL−1 final concentration) for 30 min. In (b), strains were grown to mid-log phase (0.5 OD600 mL−1) in either MDN or MDN+Glu media and, where indicated, treated with rapamycin (0.2 μg mL−1 final concentration) for 30 min. (c) Western blot analysis of Rtg3p (left-hand panel) and Northern blot analysis of ACT1, RPL30 and CIT2 (right-hand panels) in PSY142 grown under different media conditions. Cells were grown to mid-log phase (0.5 OD600 mL−1) in the specified media and, where indicated, treated with rapamycin (0.2 μg mL−1 final concentration) for 30 min. YE denotes addition of 1% yeast extract and BP denotes addition of 2% bactopeptone. In (a–c), MDN and MDN+Glu media contained the following auxotrophic supplements: 0.006% adenine sulfate, 0.003% uracil, 0.008% leucine, 0.006% histidine, 0.003% lysine and 0.005% tryptophan.

Rapamycin treatment of W303 cells grown in YPD resulted in partial hyperphosphorylation of Rtg3p (Fig. 2a, left-hand panel, lanes 3–4), in agreement with our previously published findings (Komeili, 2000). Unexpectedly, Rtg3p was also partially hyperphosphorylated in PSY142 when this strain was grown in YPD (Fig. 2a, left-hand panel, lanes 1–2). By contrast, we observed that these strains behaved differently from one another when grown in MDN+Glu, media used by Butow and coworkers (Sekito, 2000), where drug treatment continued to result in partial hyperphosphorylation of Rtg3p in W303 but instead resulted in partial dephosphorylation in PSY142 (Fig. 2a, left-hand panel, lanes 5–8). Northern blot analysis confirmed that rapamycin treatment caused similar levels of CIT2 expression in both strains under each media condition tested (Fig. 2a, right-hand panels). An independent experiment demonstrated that these strain-specific differences between PSY142 and W303 persisted in MDN media that lacked glutamate, a condition that is less repressive for CIT2 expression (Fig. 2b: compare lanes 1–2 and 5–6). From these results, we conclude that the phosphorylation state of Rtg3p responds to rapamycin treatment in a reproducible manner that is influenced both by strain background as well as by the composition of growth media.

The mobility of Rtg3p in untreated PSY142 cells grown in YPD vs. MDN+Glu showed subtle differences, in that there was a shift toward the dephosphorylated form of the protein in YPD (Fig. 2a, left-hand panel: compare lanes 1 and 5). To gain further insight into the basis for these differences, we examined the mobility of Rtg3p in PSY142 when the two different major components of YPD, yeast extract and bactopeptone, were added to MDN+Glu (Fig. 2c, left-hand panel). The results showed that addition of bactopeptone produced a significant effect in terms of shifting Rtg3p toward the dephosphorylated state (Fig. 2c, left-hand panel: compare lanes 1 and 3). Addition of yeast extract had a more subtle effect on Rtg3p phosphorylation (Fig. 2c, left-hand panel: compare lanes 1 and 5). By contrast, addition of either bactopeptone or yeast extract to MDN+Glu resulted in increased basal CIT2 expression in the absence of rapamycin, to a level comparable to what was observed in YPD (Fig. 2c, right-hand panels: compare lane 1 to lanes 3, 5, 7 and 9). These results emphasize that the phosphorylation state of Rtg3p does not correlate directly with CIT2 expression. Bactopeptone contains relatively high levels of small peptides as well as amino acids. In this regard, it is relevant that recent studies have linked RTG target gene expression to the SPS amino acid sensing system (Liu, 2001). Thus, one possibility is that the SPS system is involved in modulating the phosphorylation state of Rtg3p.

We next examined the phosphorylation state of Rtg3p in the W303 background when the RTG pathway was induced by means distinct from rapamycin treatment. Specifically, we examined Rtg3p phosphorylation in both mks1Δ as well as mks1Δrtg2Δ cells grown in YPD, conditions where Rtg3p is localized within the nucleus and the pathway is constitutively active (Dilova, 2002; Sekito, 2002; Tate, 2002; Fig. 3a). We observed that Rtg3p was dephosphorylated in the absence of rapamycin treatment in both mks1Δ and mks1Δrtg2Δ cells compared with wild-type cells (Fig. 3a, lower panels: compare lane 1 with lanes 3 and 5). These results are consistent with what Butow and coworkers have described for mks1Δ and mks1Δrtg2Δ mutants in the PSY142 background and, moreover, are in agreement with their conclusion that dephosphorylation of Rtg3p represents an important step for RTG target gene expression (Sekito, 2002). Interestingly, treatment of either mks1Δ or mks1Δrtg2Δ cells with rapamycin resulted in hyperphosphorylation of Rtg3p (Fig. 3a, lower panels, lanes 4 and 6). Because no significant difference in CIT2 expression was observed following rapamycin treatment of these mutants (Fig. 3a, upper panels, lanes 3–6), these results provide further evidence that the phosphorylation state of Rtg3p can be uncoupled from RTG target gene expression.

Figure 3

Further characterization of Rtg3p phosphorylation in wild-type and mutant W303 strains. (a) Northern blot analysis of ACT1 and CIT2 (upper panels) and western blot analysis of Rtg3p (lower panels) in WT, mks1Δand mks1Δrtg2Δ, all in the W303 background. Cells were grown to mid-log phase (0.5 OD600 mL−1) in YPD and, where indicated, treated with rapamycin (0.2 μg mL−1 final concentration) for 30 min. (b) Northern blot analysis of ACT1 and CIT2 (upper panels) and western blot analysis of Myc epitope-tagged Rtg3p (lower panel) during a time course of rapamycin treatment. Cells were grown to mid-log phase (0.6 OD600 mL−1) in YPD+Gln and treated with rapamycin for the time indicated. (c) Northern blot analysis of ACT1 and CIT2 (upper panels) and western blot analysis of Myc epitope-tagged Rtg3p (lower panel) before and after 15 min of heat shock. Cells were grown to mid-log phase (0.6 OD600 mL−1) in YPD+Gln at 25°C and either harvested immediately or combined with prewarmed media and incubated at 37°C for 15 min prior to harvesting. In (a–c), following harvesting, cells were prepared for Northern or western blot analysis, as indicated.

Given the results above, we considered the possibility that rapamycin-induced hyperphosphorylation of Rtg3p might represent a secondary consequence of inhibiting TOR function that follows activation of the RTG pathway. To test this, we examined Rtg3p phosphorylation after a shorter period (15 min) of rapamycin treatment of W303 cells grown in YPD+Gln (here glutamine was added to further repress the RTG pathway). Significant expression of CIT2 was observed yet was now accompanied by dephosphorylation of Rtg3p (Fig. 3b: compare lanes 1 and 2). As expected, Rtg3p became hyperphosphorylated following longer periods of drug treatment (Fig. 3b, lanes 3 and 4). Thus, inhibiting TOR has two distinct effects on Rtg3p phosphorylation: first, dephosphorylation that correlates with RTG target gene expression, followed by hyperphosphorylation. Interestingly, we found that 15 min of heat shock of W303 cells grown in YPD+Gln also caused modest hyperphosphorylation of Rtg3p without resulting in significant CIT2 expression (Fig. 3c). This latter result suggests that hyperphosphorylation of Rtg3p may reflect, at least in part, a more generalized response to cell stress.

Discussion

Our results presented here contribute toward a uniform view on the mechanism of RTG target gene regulation, wherein (1) MKS1 may be viewed as a strict negative regulator of the pathway and (2) dephosphorylation of Rtg3p represents an important step for RTG target gene activation. We have found that whereas a short period of rapamycin treatment results in dephosphorylation of Rtg3p, longer periods of treatment result in hyperphosphorylation of this protein in a manner that does not correlate in a simple way with activation of the pathway and, moreover, is influenced by strain- and nutrient-specific conditions. These latter observations underscore the complexity that has been revealed by recent studies into the relationship between TOR, the RTG system and the availability of specific nitrogen sources, in particular glutamate, glutamine and ammonia (Liu, 2001; Tate & Cooper, 2003; Butow & Avadhani, 2004; Dilova, 2004).

Our observation that heat shock also results in Rtg3p hyperphosphorylation is consistent with previous conclusions by Butow and coworkers that Rtg3p phosphorylation is subject to stress responses (Sekito, 2000). This conclusion was based on their analysis of a feedback control mechanism wherein the phosphorylation state of Rtg3p responds to the presence of functional Rtg3p protein within cells (Sekito, 2000). As part of this study, it was concluded that both phosphorylation and dephosphorylation of Rtg3p are likely to be cytoplasmic events. However, we conclude that rapamycin-induced hyperphosphorylation of Rtg3p may occur within the nucleus as well. This conclusion is based on previous observations that the bulk of the Rtg1p/Rtg3p complex is localized constitutively within the nucleus in mks1Δ cells (Dilova, 2002; Sekito, 2002). As we have shown here, rapamycin treatment of this strain nevertheless results in hyperphosphorylation of Rtg3p (Fig. 3a).

An important question is what precise mechanistic role is played by Rtg3p phosphorylation. Butow and coworkers have suggested that dephosphorylation of Rtg3p is required for nuclear entry of the Rtg1p/Rtg3p complex (Sekito, 2000; Butow & Avadhani, 2004). We have shown here, however, that active CIT2 expression correlates with Rtg3p existing in either a hypophosphorylated or hyperphosphorylated state, depending upon strain and environmental factors. According to one scenario, Rtg3p may become dephosphorylated within the cytoplasm, allowing for nuclear entry and RTG target gene activation, but is subsequently phosphorylated (or hyperphosphorylated) within the nucleus. Given that Rtg3p is a multiply phosphorylated protein (Sekito, 2000), it is likely that distinct sites of phosphorylation contribute to regulatory events that occur in both the cytoplasm and the nucleus.

At present, the identity of the kinases and phosphatases that act upon Rtg3p are unknown. The identification of these proteins, as well as the understanding of their relationship to TOR, represents critical next steps in our understanding of the role(s) of Rtg3p phosphorylation. An additional question remains concerning the relationship between TOR signaling and mitochondrial retrograde-dependent control of the RTG system. Our results presented here are important as they affirm that, although this system is complex, both pathways converge upon a common mechanism that involves Mks1p and that regulates the phosphorylation state of Rtg3p.

Acknowledgements

We thank Ron Butow for important conversations and for providing polyclonal anti-Rtg3p antiserum. We thank Shine (Cindy) Kikawa for technical support during molecular genetic analyses of the Σ1278SSmks1Δ strain. This work was supported by NSF grant MCB-0131221 (to T.P.).

Footnotes

  • Editor: Terrance Cooper

References

View Abstract