Genetic and Physical Interactions between Tel2 and the Med15 Mediator Subunit in Saccharomyces cerevisiae

Background In budding yeast, the highly conserved Tel2 protein is part of several complexes and its main function is now believed to be in the biogenesis of phosphatidyl inositol 3-kinase related kinases. Principal Findings To uncover potentially novel functions of Tel2, we set out to isolate temperature-sensitive (ts) mutant alleles of TEL2 in order to perform genetic screenings. MED15/GAL11, a subunit of Mediator, a general regulator of transcription, was isolated as a suppressor of these mutants. The isolated tel2 mutants exhibited a short telomere phenotype that was partially rescued by MED15/GAL11 overexpression. The tel2-15mutant was markedly deficient in the transcription of EST2, coding for the catalytic subunit of telomerase, potentially explaining the short telomere phenotype of this mutant. In parallel, a two-hybrid screen identified an association between Tel2 and Rvb2, a highly conserved member of the AAA+ family of ATPases further found by in vivo co-immunoprecipitation to be tight and constitutive. Transiently overproduced Tel2 and Med15/Gal11 associated together, suggesting a potential role for Tel2 in transcription. Other Mediator subunits, as well as SUA7/TFIIB, also rescued the tel2-ts mutants. Significance Altogether, the present data suggest the existence of a novel role for Tel2, namely in transcription, possibly in cooperation with Rvb2 and involving the existence of physical interactions with the Med15/Gal11 Mediator subunit.


Introduction
TEL2 is a highly conserved gene that has been found in all eukaryotic organisms examined so far. Before describing the very important and fundamental roles of Tel2 known to date, it is worth briefly describing the chronology of Tel2's history. Tel2 was originally isolated as a potential regulator of telomere length in the budding yeast Saccharomyces cerevisiae, but became eventually, over twenty years later, highly suspected, in humans at least but also in the fission yeast Schizosaccharomyces pombe, of having nothing to do with telomeres [1]. In the meantime, several, sometimes contradictory, telomere studies on Tel2 from various organisms had accumulated. The Tel2 story begins in 1986, when Lustig and Petes isolated, on the basis of telomere tracts length, two S. cerevisiae mutant strains that had abnormally short telomeres and named them tel1 and tel2 [2]. Telomeres, specialized nucleoprotein complexes, represent the natural ends of linear chromosomes and their integrity is essential for genome stability. Telomeres protect against unwanted chromosome end-to-end fusions, against degradation by DNA modifying enzymes and prevent chromosome ends from being mistaken for DNA double-strand breaks [3]. The genes corresponding to the short telomere mutants isolated by Lustig and Petes [2], TEL1 and TEL2, were cloned around ten years later [4,5]. Tel1, as well as its human ortholog, ATM (Ataxia Telangiectasia Mutated), have been since extensively documented as they both play pivotal roles in genome stability as well as in the response of the cell to DNA damage, while Tel2 remained poorly documented for many years following its identification. In fact, Tel1 has a true role in telomere biology in budding yeast as it was recently found to localize at the ends of the shortest telomeres in the cell, an event which then favors telomerase recruitment and telomere re-elongation [6][7][8].
On the other hand, early experiments on S. cerevisiae Tel2 revealed that it could bind telomeric DNA, at least in vitro [9,10]. Not long later, probably led by the potential implication of S. cerevisiae Tel2 in telomeric functions, several studies on Tel2 from other organisms aiming at looking for telomeric functions were undertaken. Thus, human TEL2, also known as HCLK2, and Caenorhabditis elegans TEL2, also known as RAD-5 or CLK-2, were implicated in the control of telomere length [11][12][13]. Meanwhile, other studies investigated in other directions and, as a consequence, Tel2 was also implicated in the response to DNA damage, in C. elegans [12,14], S. pombe [15] and humans [16,17]. This was also the case in S. cerevisiae, as Tel2 was found to physically bind Tel1, an event that was needed to recruit Tel1 at DNA doublestrand breaks, as well as Mec1, an ATR ortholog [18,19]. Given the localization of Tel1 at short telomeres [6,7], it would be extremely important to know whether Tel2 is needed to recruit Tel1 at short telomeres, as it does at DNA double-strand breaks [18,19]. In mammals, Tel2 physically associated with -and was required for the stability-of all six mammalian phosphoinositide 3kinase related kinases (PIKKs), ATM, ATR, DNA-PKcs, mTOR, SMG1 and TRRAP [17]. These PIKKs complexes are important for basic transactions associated with DNA damage signaling/ repair and associated cell cycle control, nutrient sensing and cell growth control, degradation of mRNA and control of gene expression, principally [1,20]. Strikingly, however, the study on mammalian TEL2 failed to reveal any evidence for telomeric phenotypes associated with TEL2 genetic inactivation [17], in spite of the fact that ATM is pivotal in the response of telomeres to DNA damage [3]. In parallel, in S. pombe, Tel2 was found to bind to all three PIKKs, Rad3, Tel1 and Trapp1/2 [20,21]. The data in humans revealing a role of TEL2 in ATM stability, as seen above [17], might well explain the short telomere phenotype of the original S. cerevisiae tel2-1 mutant [2] as Tel1, the ortholog of ATM, might become unstable in that mutant. In fact, very recently, an S. cerevisiae temperature-sensitive tel2 mutant was found to exhibit depressed levels of Tel1 [22], thus potentially indicating instability of Tel1 as was the case for ATM following inactivation of TEL2 in mammalian cells [17].
Recent studies revealed that budding yeast Tel2 was part of a novel complex, named ASTRA (ASsembly of Tel, Rvb and Atmlike kinase), which contained seven subunits: Tel2, Tti1 and Tti2, defining a PIKK biogenesis complex, the AAA+ ATPase complex Rvb1/Rvb2, the PIKK Tra1 and Asa1 [23]. In addition, other even more recent studies revealed the existence of functional interactions between the Tel2, Tti1 and Tti2 subunits (now known as the TTT complex) of the ASTRA complex and the Rvb1/2, Tah1, Pih1 subunits of the so-called R2TP complex (Rvb1/2 being also part of the ASTRA complex) [23][24][25][26][27][28][29]. In particular, it is interesting to note that, based on structural analyses, yeast Tel2 acted with the molecular chaperone Hsp90 (part of the R2TP complex) in the maturation of PIKK complexes [29]. It has been proposed that TEL2 might act as a scaffold to coordinate the activities of R2TP/prefoldin-like and HSP90 chaperone complexes during the assembly of the PIKKs [26].
In the present study, we report the isolation and initial characterization of tel2-ts mutants and their utilization as tools in genetic screenings. Unexpectedly, we have isolated MED15, a subunit of the Mediator complex, a general regulator of transcription, as a suppressor of these tel2-ts mutants.

Temperature-sensitive tel2 mutants
We first set out to isolate temperature-sensitive (ts) mutants of S. cerevisiae TEL2 (see Materials and Methods). We could isolate six such tel2 alleles. We could distinguish two classes of tel2 mutants, based on the severity of the temperature-associated growth defects and the capacity to form colonies at 34-37uC ( Figure 1A) and the nature of the mutations harbored ( Table 1). Death in these tel2 mutants was accompanied by cell lysis, which appeared to occur at any of the cell cycle stages. FACS analysis, performed on alpha factor-synchronized populations of tel2-15 and tel2-19 mutant cells, as well as examination in the light microscope, failed to reveal any obvious cell-cycle defect at restrictive temperatures for growth ( Figure 1B and data not shown).
The MED15 Mediator subunit is a suppressor of tel2-ts mutations All six tel2 alleles identified in the present study were used to perform genetic screens in order to try and isolate extragenic suppressors. Only one such gene, namely GAL11, was isolated in these genetic screens, besides TEL2 itself. GAL11 behaved both as a low-copy and a high-copy suppressor. Thus, GAL11 was contained in the YCp50#8 centromeric (CEN4, URA3) plasmid (from the YCp50 genomic library; [30]) isolated as a weak rescuer of tel2-15 at 31uC (see below), as well as in the YEp24#2 multicopy (episomal, 2 m, URA3) plasmid (from the YEp24 genomic library; [31]), strong rescuer of tel2-26 at 34uC. Restriction analysis identified the GAL11 gene and surrounding sequences as responsible for the rescue in YCp50#8 and YEp24#2, both around 4.2 kb in length (GAL11 ORF is 3243 base pairs in length). In addition, TEL2 itself expressed from a YEp24 library plasmid was isolated twice as a strong rescuer of tel2-19 and tel2-25 at 35 and 33uC, respectively (see below). Gal11, which according to a unified nomenclature should now be referred to as Med15 [32], is a subunit of the yeast Mediator, a large complex required for the recruitment of RNA polymerase II to activated promoters (see, for instance, [33]). We therefore considered the possibility that increased expression of MED15 might activate the transcription of TEL2, thus resulting in increased Tel2 protein levels being responsible for the observed suppression of tel2-ts growth defects by the YCp50#8 and YEp24#2 plasmids. By Western analysis, the levels of either HA 2 -Tel2 or Myc 2 -Tel2, both expressed under the control of native promoter from the tagged construct integrated at TEL2 locus, did not vary whether MED15 was expressed from either the centromeric or multi-copy plasmid, in addition to endogenous Med15, or expressed from endogenous locus only, in cells transformed with plasmid alone ( Figure 1D). These data therefore suggest that suppression of the thermosensitivity of tel-15 or tel2-26 by YCp50-MED15 or YEp24-MED15, respectively, is not due to increased amounts of TEL2 protein. In addition, we also verified that the suppression was not due to increased transcription of TEL2 ( Figure 1C).
TEL2 genetically interacts with other Mediator subunits, as well as with SUA7/TFIIB In parallel with the genetic screenings described above, we performed a two-hybrid screen using Tel2 as the bait. Only one clone scored positive, identifying S. cerevisiae RVB2 cDNA (see Appendix S1 in the Supporting information and Figures S1, S2).
Rvb1/2, two highly conserved members of the AAA+ family of ATPases, are parts of several distinct complexes, namely, the SWR1 and INO80 complexes of chromatin modification, the R2TP complex which, together with the Hsp90 chaperone, regulates the accumulation and stability of snoRNPs, and the ASTRA complex, the function of which is presently unknown [23][24][33][34][35]. In all of these multi-protein complexes, the presence of Rvb1 and that of Rvb2 were reported to be essential for their proper function. In addition, Rvb1/pontin and Rvb2/ reptin have been implicated in general transcription, not only in S. cerevisiae in which roughly 5% of all genes were found to be deregulated in rvb2 mutants [36,37] but also in Vertebrates [38]. In view of the solid and constitutive physical association between Rvb2 and Tel2 and of the genetic interaction between TEL2 and the MED15 Mediator subunit, described above, we hypothesized that Tel2 might also function as a general regulator of transcription.
Sua7, budding yeast TFIIB, an essential component of RNA polymerase II [39] has previously been reported to bind Tel2 by two-hybrid [40]. This putative physical interaction, obtained from high-throughput data, has not been retested and cannot therefore be used to draw any valid conclusion. However, this arose our curiosity and we set out to test the potential existence of genetic interactions between TEL2 and SUA7. Interestingly, continuous overexpression (on agar-based semi-solid medium) of SUA7 under the control of the GAL1-10 promoter resulted in the suppression of tel2-19 and tel2-25 growth defects at 35 and 32uC, respectively (data not shown). On the other hand, under the same conditions, overexpression of RVB2, also under GAL1-10 promoter control, did not rescue the growth defects in the thermo-sensitive tel2-19 and tel2-25 mutants and, in fact, aggravated these defects (data not shown).
Next, to see whether tel2-ts rescue by MED15 might reflect a general property of Mediator, we constructed plasmids to express one of several additional Mediator subunits under the control of the strong, inducible GAL1-10 promoter in a multi-copy (2 m) plasmid. Among the MED1, MED2, MED5, MED12, MED16, . Endogenous Tel2, tagged with either Myc 2 or HA 2 in its N-terminus, was immunoprecipitated from crude extracts with the corresponding monoclonal antibody and Western blotting with the same antibody. Therefore, overexpression of MED15 does not rescue the tel2-ts mutants by increasing either TEL2 transcription or even indirectly by increasing Tel2 protein levels. doi:10.1371/journal.pone.0030451.g001 MED17, MED18 and CDK8 genes chosen for this experiment, we found that overexpression of MED16, MED18 and CDK8, in addition to that of MED15, as described above, resulted in a modest rescue of tel2-19-ts growth defects at 35uC, all slightly less efficient than that by MED15 ( Figure 2A).

The Med15 Mediator subunit has physical affinities with Tel2
The genetic interaction between MED15 and TEL2, described above, prompted us to look for possible physical interactions between the two proteins, interactions that are frequently suspected of taking place under similar circumstances. Therefore, we assessed possible Tel2-Med15 physical interactions under conditions of increased expression by using a GAL1-10 promotercontrolled inducible system. Such a system in which the duration of induction of the promoter in galactose-based liquid culture medium is controlled is currently used in genome-wide analyses because it presents the advantage compared to other systems of assessing in vivo interactions between proteins in their native configuration. Upon induction of the GAL1-10 promoter for 2 hr at 29uC, we could readily detect in vivo physical interactions between Tel2-Myc 3 and Med15-HA 2 ( Figure 2). Thus, full length Tel2-Myc 3 was found to specifically bind the first half of Med15 (Med15 1-540 -HA 2 ), but not the second half of Med15 (Med15 541-1081 -HA 2 ) ( Figure 2B, C). Additional restriction of the expressed proteins showed that the first quarter of Med15 (Med15 1-270 -HA 2 ) efficiently bound the first half of Tel2 (Tel2 1-343 -HA 2 ) ( Figure 2D). Therefore, given the genetic suppression of tel2-ts mutants by MED15, as well as the physical affinity between Med15 and Tel2, it is tempting to speculate that the Rvb2-Tel2 module might play an important role during transcription. The previously established physical interaction between Tel2 and Sua7/TFIIB [40] supports this view.
Finally, under transient overexpression conditions, Med15 1-540 -HA 2 still physically associated with the temperature-sensitive Tel2-15-Myc 3 protein at either 29uC or 34uC ( Figure 2E). Tel2-15 was chosen for these experiments because it confers a tighter ts phenotype than Tel2-19. Note that all five mutations in Tel2-15 ( Table 1) lie in the part of Tel2 (amino acids 1-343) that is relevant for interaction with Med15 ( Figure 2D). One can therefore reasonably conclude that these mutations do not affect association with Med15.

Short telomeres of tel2 mutants rescued by MED15 overexpression
Prior to isolation of the temperature-sensitive tel2 mutants described above, we had been searching for tel2 mutants with altered telomere length (see Materials and Methods), and found one of them (out of 200 hundreds screened mutants), tel2-12, that exhibited shortened telomeres. The mutations in tel2-12 were later found to be identical to those in the tel2-19 mutant ( Table 1) and later found to confer temperature sensitivity. Upon further analysis of these mutants, we found that both the tel2-15 and tel2-19 mutants exhibited telomere shortening, just like the tel2-12 mutant, when grown at either permissive, 24uC, ( Figure 3A) or semitemperature for growth of 29uC ( Figure 3B). To further document the telomere length deregulation occurring in the tel2ts mutants, we additionally constructed tel2-ts-based double mutants. Importantly, we observed the absence of an additional effect on telomere length when the tel1 null and tel2-19 mutations were combined (tel1 null also confers telomere shortening; [2]; Figure 3B), as expected from the previous observation that Tel1 and Tel2 function in the same pathway of telomere length regulation [2]. We also observed that telomeres in the tel1D yku70D double mutants were shorter than in the two corresponding single mutants ( Figure S3). This indicated that tel1D cells have telomeres that can get even shorter when they are in combination with an additional mutation, thus serving as a positive control for the present situation in which the tel2 mutation fails to further shorten telomeres in the tel1 mutant background ( Figure 3B). This point had been previously established by a study reporting that the tel1 hdf1/yku70 null double mutant had telomeres shorter than either of the two single tel1 and hdf1 single mutants and that the double mutant lost telomeres at an accelerated rate compared with the single mutants [41]. In contrast, combining the tel2-19 (or tel2-15) and yku70 null mutations (yku70 null also confers telomere shortening; [42] provoked an additive effect on telomere shortening ( Figure 3C). This finding was expected since previous data showed that Tel1 and Yku70 are in separate genetic pathways for telomere length regulation [41] while Tel1 and Tel2 are in the same pathway [2].
Interestingly, overexpression of MED15 not only partially rescued the temperature sensitivity of the tel2-15 and tel2-19 mutants ( Figure 4A), but also partially rescued their short telomere phenotype at semi-permissive temperatures for growth ( Figure 4B; see also Figure 4C for the wild-type control). At the permissive temperature of 24uC, the telomere length defect of the tel2-ts mutants, which is already present ( Figure 3A), was also partially rescued by overexpression of MED15 (data not shown). We noted that overexpression of MED15 caused growth defects at 24uC in the tel2-ts mutants (Figures 2A, 4A) and at all temperatures tested in the wild type ( Figure 4A). While this did not prevent rescue of the tel2-ts at higher temperatures, this may be important to keep in mind. We also note that overexpression of other Mediator subunits did not result in such a toxic effect (Figure2A). Therefore, manipulating the levels of Med15 in wild- Table 1. Sequence analysis of the amino acid changes in the tel2 mutant proteins.  Overexpression of the indicated genes, under the control of the GAL1-10 promoter, was continuously induced by growing on solid media containing galactose as the sole carbon source. Cells transformed with the indicated plasmid were first grown in liquid culture in selective minimal medium containing glucose as the carbon source before being re-streaked on agar-based selective minimal medium containing galactose as the carbon source. Growth was assessed after 3 days at the indicated temperature in the serially diluted cells (ten-fold dilutions from left to right in each condition). At 35uC, MED16, CDK8 and MED18 overexpression rescued tel2-19, although slightly less efficiently than MED15. (B) Tel2 and Med15 physically interact in vivo. Putative physical interactions between Tel2-Myc 3 and Med15-HA 2 were measured upon immunoprecipitation (IP) with anti-type cells may cause toxic effects by potentially titrating out other essential components. We do not have at the moment the exact explanation for this phenomenon and its elucidation will require future analysis.

Tel2 controls the transcription of EST2/telomerase
The tel2-15 mutant described above has been recently reported to also affect telomere length at semi-permissive temperature for growth [22]. In the tel2-15 mutant, the levels of Tel1 were dramatically depressed [22], thus providing a rational explanation for the fact that this mutant exhibits short telomeres ( [22]; present data), tel1 null mutant also exhibiting short telomeres [2].
However, based on the data above, the short telomere phenotype of the tel2-15 and tel2-19 mutants might additionally be due to defects in telomerase transcription potentially resulting from alterations of functional interactions between Tel2 and Med15 and/or between Tel2 and Rvb2 in these mutants. These observations next prompted us to measure telomerase transcription in these tel2 mutants. First, we set out to measure levels of TLC1 RNA, the template subunit for S. cerevisiae telomerase, in the tel2-15 and tel2-19 mutants. No significant difference could be seen in TLC1 RNA levels in these mutants when compared with the wild type ( Figure 5A). In contrast to the situation with TLC1, the transcription of EST2, coding for the protein subunit of telomerase, was clearly diminished in the tel2-15 mutant grown at semi-permissive temperature and even at permissive temperature for growth and only modestly diminished in the tel2-19 mutant ( Figure 5C), consistent with the fact that the tel2-15 ts mutant is tighter than the tel2-19 ts mutant (Figures 1A). In agreement with the data on EST2 transcription reported above, we observed that the level of endogenous Myc 18 -Est2 was depressed in strains harboring the tel2-15 mutation compared with the TEL2 + strain ( Figure 5B). We noted that, inconsistently, in this particular experiment the level of Myc 18 -Est2 at the permissive temperature of 24uC was even higher in the tel2-15 mutant than in the wild-type strain ( Figure 5B). To understand this phenomenon, we examined all available data on the subject. In fact, in other experiments, one of which using the tel2-19 mutant is shown in Figure S4, the level of Myc 18 -Est2 was found to be similar to that in the wild type at 24uC. In another experiment, the level of Myc 18 -Est2 in tel2-15 at 24uC was lower than that in the wild type (data not shown). Therefore, we do not know yet whether there is a mode of regulation of Est2 particular to the tel2-15 mutant, perhaps, which we do not understand yet and which might be an indirect effect. The tel2-15 mutant is more severe than the tel2-19 mutant and may be deregulated in a more unpredictable manner at permissive temperatures. Nevertheless, it is clear that Myc 18 -Est2 levels were greatly perturbed in both mutants at 29uC (Figures 5B, S4). Incidentally, we found that the tel2-15 and tel2-19 mutations did not provoke a reduction in TEL1 trancript levels (data not shown). Therefore, the depression of Tel1 protein levels observed in the tel2-15 mutant [22] seems to be due to uncorrect assembly of the Tel1 PIKK kinase [17] rather than to repressed transcription, as previously suggested [22].
As mentioned above, Rvb2 has been found to control the expression of around 5% of the S. cerevisiae genes [36]. If Tel2 controlled general transcription in association with Rvb2, mutations in either one should result in a similar gene expression signature. To test this possibility, we selected two genes, MAE1 and MEP2, the expression of which have been previously shown to be up-regulated and down-regulated by Rvb2, respectively [36,37] and measured their levels of expression in the tel2-15 mutant. Interestingly, the same trend in the expression of MAE1 and MEP2 was found in the tel2-15 mutant compared with that in the tih2-160 mutant, a temperature-sensitive mutant of RVB2 [37]. Indeed, MAE1 mRNA levels were severely depressed in the tel2-15 mutant, while, on the opposite, MEP2 mRNA levels were dramatically increased in that same mutant ( Figure 5D), a figure similar to that found in the tih2-160/rvb2 mutant [37].
These observations next prompted us to evaluate the impact of overexpressing MED15 on EST2 transcription. Interestingly, overexpressing the MED15 subunit of Mediator under the control of its natural promoter resulted in a dramatic increase in EST2/ telomerase transcription ( Figure 5E).
We presumed that est2 null mutant cells were not temperaturesensitive during the period of time, around 75 generations, they were still alive prior to reaching telomeric senescence [43], as this has never been reported in the literature. However, to make sure that this was the case, we directly tested it and found that, indeed, the est2 null mutant cells prior to senescence were only moderately temperature-sensitive ( Figure S5) unlike the tel2-15 and tel2-19 mutants which were markedly temperature-sensitive, as seen above. Moreover, the growth defects in the est2D mutant could be explained by the damage generated by telomere erosion (see the legend to Figure S5). These observations strongly suggested that the tel2-ts phenotype was not simply due the diminution of EST2/ telomerase transcription. In addition, we note that the tel2-ts mutants arrest at various stages during the cell cycle, as seen above, while the telomerase/est2D mutants arrest at the G2/M border [43]. In fact, recent genome wide data suggest that the tel2ts mutants are likely to harbor cumulative defects in the transcription of several genes, thereby leading, directly or indirectly, to the depressed levels of endogenous Est2 observed here. Indeed, one of the tel2-ts mutants uncovered here, tel2-7 (see mutations, Table 1) exhibited genetic interactions, under the form of synthetic growth defects, with numerous and various mutations ( [44]; see also the BioGRID, Toronto, at http:// thebiogrid.org; the tel2-7 mutant was provided by us to these authors). We therefore conclude that the short telomere phenotype of tel2-ts mutants is not solely due to a diminution in telomerase transcription, as proposed above and in agreement with previous observations [22].

Discussion
The present study does not bring much more information concerning the possible telomeric functions of S. cerevisiae Tel2, with the exception of a possible general role in EST2/telomerase transcription (but also in the transcription of other genes).
HA or anti-Myc monoclonal antibodies, followed by Western blotting (West.), from crude extracts from strains having transiently overexpressed the indicated constructs after incubation in galactose-based liquid selective media, for 2 hr at 29uC (same conditions for panels B-E). All experiments were conducted in parallel in strains expressing both constructs and in strains with the single constructs to take into account possible background signals. Full length Tel2-Myc 3 specifically binds the first half of Med15 (Med15 1-540 -HA 2 ) in both directions, that is to say that a positive signal was detected whether one or the other of the two proteins was immunoprecipitated. (C) In contrast, the second half of Med15 (Med15 541-1081 -HA 2 ) did not associate in vivo with full length Tel2-Myc 3 . (D) Finally, the first quarter of Med15 (Med15 1-270 -HA 2 ) efficiently bound the first half of Tel2 (Tel2 1-343 -HA 2 ). (E) At the maximal permissive temperature for growth for tel2-15 of 29uC, the Tel2-15-Myc 3 protein still physically interacted with Med15 1-540 -HA 2 , as was also the case at 34uC (data not shown). doi:10.1371/journal.pone.0030451.g002 However, unexpectedly, we find that the MED15 subunit of Mediator (a general regulator of transcription) is a genetic suppressor of the tel2-ts mutations isolated in the present study. This genetic interaction was further documented by the finding of an in vivo interaction between Tel2 and Med15. Overexpression of other Mediator subunits, as well as of SUA7/TFIIB, also rescued the tel2-ts growth defects. We also report here the existence of a physical association between Tel2 and Rvb2. Potentially, based on the already known role of Rvb2 in controlling transcription, the Tel2-Rvb2 interaction might represent a necessary basis for the functional interactions between Tel2 and Med15.
In addition to the recently discovered role of human RUVBL1/2 in mRNA stability (at least those of the PIKKs) via the SMG1-mediated nonsense-mediated mRNA decay (NMD) pathway [28], there is overwhelming evidence that they regulate transcription when functioning in the SRCAP, TIP60 and INO80 chromatin-modifying complexes [34,38]. Most likely, S. cerevisiae Rvb1/2 perform similar tasks in regulating transcription when part of the respective homologous complexes, SWR1, NuA4 and INO80 [34]. In view of the present results, it is interesting to speculate that the S. cerevisiae Rvb2-Tel2 module might play a major role in the expression of multiple genes, as already demonstrated for Rvb2 [34][35][36][37], including of EST2/ telomerase, as shown here. This is further supported by the observation that Med15-HA 2 physically interacted with Rvb2-Myc 3 when both were transiently overproduced (NG, MC, unpublished data). Interestingly, genetic inactivation of pontin/ hRUVBL1 or of reptin/hRUVBL2 was recently found to result in decreases in mRNA levels of hTERT, the catalytic subunit of human telomerase, effect partly exerted through binding of reptin to hTERT proximal promoter [45]. Moreover, overexpression of reptin was observed in primary gastric cancer specimens [45]. In a separate study, on colon cancer cells, RUVBL2 was also found to regulate hTERT transcription [46]. Regarding EST2/telomerase expression in S. cerevisiae, it is worth mentioning that its levels of mRNA, as well as those of other genes coding for telomeric proteins, namely Est1, Est3, Stn1 and Ten1, were found to be regulated by the NMD pathway [47], thus establishing a potential link between the Rvb2-Tel2 module and the NMD pathway, as shown for human RUVBL1/2 [28].
The putative role of Tel2 in transcription might be parallel and distinct from its role in the biogenesis of the PIKKs complexes. Because MED15 overexpression rescued the growth defects of the tel2-ts mutants, and Med15 can physically associate with Tel2, we further speculate that the function of Tel2 in transcription uncovered here probably depends on physical interactions with the Mediator via its Med15 subunit. A possible clue stems from the observation of genetic interactions between Tel2 and Sua7, uncovered here, as well as between Sua7 and Med15/Gal11 [48]. SUA7 encodes yeast TFIIB, a general transcription factor required for the initiation of transcription by RNA polymerase II [39]. Mediator's general function is to recruit RNA polymerase II at sites of active transcription [32,33]. Med15/Gal11 was found to partially suppress deletions of GAL4 and GCN4, encoding two general transcriptional activators [48]. These authors proposed that transcriptional activators work by raising the local concentration of the limiting factor Med15/Gal11, and that Med15/ Gal11 works by recruiting Mediator and Taf14-containing transcription factors like TFIID and SWI/SNF and by competing general repressors like Ssn6-Tup1 off the target promoters [48]. Potentially, transcriptional activators might recruit the Mediator, via Med15, to the chosen site of transcription, while simultaneously recruiting Sua7 via Tel2. The role of Tel2 in these transactions might merely be to assist in chromatin remodeling in cooperation with Rvb1/2 in order to facilitate physical interactions between Sua7 and the Mediator.
Another possible clue to explain the events observed here is related to telomere position effect (TPE), a phenomenon that reversibly silences genes situated close to telomeres [49]. Interestingly, in S. cerevisiae, both Tel2 and Med15 have been reported to affect telomere position effect. Thus, the tel2-1 mutation reduced TPE but had no detectable effect on silencing of HMLa or HMRa, the silent mating type cassettes [5]. However, to our knowledge, the reasons for the implications of Tel2 in TPE have not been uncovered since, except for the proposal that Tel2 may be required for chromatin assembly at telomeres and elsewhere in the genome [5]. On the other hand, a mutation in GAL11/MED15 or, on the contrary, its overexpression, both affected TPE and telomere structure [50]. Moreover, a mutation in GAL11/SDS4/MED15 was found to strongly suppress a rap1-induced silencing defect at the telomeres and the HMR locus [51]. In fact, very recently, it was established that the Mediator directly associated with heterochromatin at telomeres, thereby influencing the exact boundary between active and inactive chromatin [52]. Mutations in Mediator subunits also resulted in increased levels of H4K16 acetylation near telomeres and in desilencing of subtelomeric genes, an effect that appeared to be distinct from the role of Mediator as a co-trancriptional activator [52]. Therefore, the interactions between Tel2 and Med15 described in the present work might theoretically have something to do with their effect on telomere structure. Further work will be required to test these hypotheses, which might be complex and difficult to experimentally approach. For the moment, we have ruled out a role for Tel2 in telomere end protection, as our tel2-ts mutants did not activate the DNA damage checkpoint, unlike mutants defective in telomere protection such as cdc13-1, for instance (NG, MC, unpublished data). However, more subtle defects of telomere structure resulting from TEL2 inactivation, such as those discovered in early studies [5], might be responsible for the functional interactions with MED15 uncovered here.
As stressed above, Tel2 has long been an enigmatic protein, implicated in a myriad of apparently unrelated biological processes in various organisms. However, recent studies have uncovered what appears to be the main function of Tel2, namely in PIKK biogenesis, and other more recent studies have begun to unravel these mechanisms at the molecular level. However, a recent analysis showed that C. elegans CLK-2/TEL2 depletion did not phenocopy PIKK kinase depletion and, in addition, Figure 3. The temperature-sensitive tel2-15 and tel2-19 mutants exhibit short telomeres. (A) tel2-15 confers a more severe telomere length defect than tel2-19, compared here with the previously described tel2-1 mutant [2], all grown at 24uC. (B) The shortened telomeres in the tel2-19 mutant were not furthered shortened when combined with a tel1 null mutation. These strains were grown at 29uC. (C). In contrast, both the tel2-19 yku70 null and tel2-15 yku70 null double mutants, grown here at 24uC, exhibited shorter telomeres than either one of the corresponding single mutants. Average telomere length (corresponding to the DNA smear migrating at around 1.3 kb) of tel2 mutant strains with the indicated relevant genotype, grown for at least 22 days to allow telomere length to attain a stable value, was detected by Southern blotting with a TG 1-3 telomeric P 32labeled probe. The dashed horizontal lines, drawn between the mean length values of two or more wild-type (wt) or mutant strains, as indicated in the margin, allow to get a better appreciation of the variations in telomere length in the mutants. doi:10.1371/journal.pone.0030451.g003 implicated CLK-2/TEL2 in multiple developmental and cell cycle related processes [53]. The genetic approach developed here has provided new clues to further understand Tel2 functions and, unexpectedly, orientate the research on Tel2 towards directions that had not been suspected before. The present data will hopefully serve as a starting point for further exploring the role of this pivotal protein, Tel2, in potentially novel mechanisms of regulation of transcription.

Yeast strains, plasmids and inducible overexpression by galactose
Yeast strains used in this study were derivatives of BF264-15Daub (ade1 his2 leu2-3,112 trp1-1a ura3Dns), as described previously [54]. Yeast cultures were grown at the indicated temperatures in YEP (1% yeast extract, 2% bacto-peptone, Figure 4. MED15, coding for a subunit of Mediator, found in the present study to be an extragenic suppressor of tel2-ts mutations, also partially rescues their telomere length defect. (A) The temperature-sensitive tel2-15 and tel2-19 mutants, as well as a wild-type (wt) strain,were re-transformed, here, with the MED15 or TEL2 plasmids, either low-copy (1 or 2 copies) CEN (YCplac) or multi-copy 2 m (YEplac), both originally isolated from the genetic screens performed in the present study, or with the corresponding empty plasmids (YCp or YEp), and their growth characteristics evaluated after spotting ten-fold dilutions (from left to right in each condition) on selective, -Ura, minimal medium at the indicated temperatures. YCp-MED15 rescued tel2-15 at up to 29uC, while YEp-MED15 rescued tel2-19 at up to 34uC. (B) Overexpression of MED15, from the plasmids described above, in A, partially rescued the short telomere phenotype of the tel2-15 and tel2-19 mutant strains. Compare the extent of the rescue with that provided by TEL2 overexpression. Telomere length was measured as described in the legend to Figure 3 after the strains had been propagated at 27.5uC (tel2-15) or 29uC (tel2-19) on minimal medium for 30 days. (C) Overexpression of MED15 had no elongation effect on telomere length in wild-type cells, grown here at 29uC. In fact, a slight shortening of telomeres was observed following MED15 overexpression compared with expression of vector alone. doi:10.1371/journal.pone.0030451.g004 Figure 5. Tel2 controls transcription of EST2/telomerase (but not that of the telomerase RNA subunit TLC1) and is regulated by the Med15 Mediator subunit. (A) Northern blot analysis of endogenous TLC1 RNA levels indicates an absence of deregulation in the indicated tel2 mutants. rvb2 mutants that exhibit slight and stable telomere shortening (NG, MC, submitted for publication) were also used in that experiment. A Tel2 and the Mediator PLoS ONE | www.plosone.org 0.005% adenine, 0.005% uracile) supplemented with 2% glucose (YEPD), sucrose or galactose, or in selective minimal medium. All strains were made isogenic by back crossing at least five times against our genetic background. Strain origins, prior to back crossing, were as follows. The rvb2::KanMX4/RVB2, tel2::KanMX4/ TEL2, est2::KanMX4/EST2, tel1::KanMX4 and yku70::KanMX4 strains were purchased at Euroscarf (Frankfurt, Germany). The tlc1::TRP1 strain was from the Gottschling laboratory.
Two-hybrid experiments using the pACT2 and pAS2 vectors and pACT1 cDNA library were performed as described previously [55].
All constructs were made by using Polymerase Chain Reaction (PCR) to adapt the relevant restriction sites to the sequence of the genes and details of the constructs will be made available upon request. To confirm the two-hybrid interaction between Tel2 and Rvb2 and delineate the domains of interaction, we opted for a transient overexpression system, currently used in genome-wide analyses, which presents the advantage of measuring in vivo interactions of proteins in their native configuration. To have accurate control of the extent of expression, we used the galactoseinducible GAL1-10 promoter and activated it only in a transient manner to avoid possible deleterious effects of heavy overexpression. RVB2 was tagged with a 2 HA-6 His (hereafter referred to as HA 2 because an anti-HA monoclonal antibody was used throughout) and TEL2 with a 3 Myc (Myc 3 ) epitope tag, both at their 39 end. Induction of genes placed under the control of the GAL1-10 promoter was done by transferring cells growing in liquid culture in glucose-containing medium into galactose-containing liquid medium after several washes by centrifugation. In experiments involving expression of protein parts under GAL1-10 promoter control, a supplementary methionine residue was added in front of the truncated sequence (if starting from places other than the natural initiating ATG) to initiate efficient translation.

Western blotting and immunoprecipitation
Techniques for block and release experiments, flow cytometry analysis (FACS), cell extract preparation, immunoprecipitation and immunoblotting (analyzed using an Odyssey Imager) have been described previously [54,55]. Mouse monoclonal anti-HA raw ascites fluid 16B12 (BabCO) and mouse monoclonal anti-HA 12CA5 antibody (Roche Diagnostics) were used for immunoprecipitation and imunoblotting, respectively. Myc-tagged proteins were visualized after immunoprecipitation and Western blotting with monoclonal anti-Myc antibody 9E10 (Roche Diagnostics). Anti-actin antibody, clone C4, was from MP Biomedicals.

Telomere length measurement
To analyze telomere length, genomic DNAs were prepared, separated in a 0.9% agarose gel (in TBE) run in TBE buffer overnight and, after denaturation, transferred and hybridized with a 270 base pair TG 1-3 32 P-labeled telomeric probe as described previously [55]. Following digestion of genomic DNA with XhoI, to cut within the Y9 regions of chromosomes, telomere tracts of wildtype cells appear as a broad band of ,1.1-1.3 kb which represents the average length of most chromosomes. Results were analyzed using an FLA-5100 Fuji phosphoimager and the ImageGauge software.

Northern blotting and analysis of transcription by RT-PCR
For analysis of TLC1 endogenous levels, total RNA was first isolated from yeast cell cultures according to standard procedures. Northern blot analysis was conducted according to classical techniques using a P 32 -labeled probe specific for TLC1 sequences. A tlc1 null strain was used in all experiments to attest for the specificity of the detected signals.
The relative quantification of mRNA was performed with a quantitative RT-PCR assay. Gene-specific primers were designed using Universal Probe Library Assay Design Center (Roche Applied Science) as primer software. Two micrograms of RNA were reverse transcribed using the first strand cDNA synthesis kit from Fermentas, with gene specific primer for EST2 and random hexamers as primers for the other genes. cDNAs were then diluted to a final concentration of 10 ng/ml in sterile H 2 O and amplified using a BioRad Opticon instrument. Amplification was performed in 20 ml of reaction mix, containing 5 ml of diluted cDNA and 15 ml Mesa Green qPCR Master SYBR Green I (Eurogentec), together with forward and reverse primers. PCR was performed according to a two-step protocol: 3 min at 95uC, followed by 45 cycles comprising each 10 s at 95uC (denaturation) and 30 s at 60uC (annealing/extension). Quantitative data of the samples were obtained using the BioRad CFX Manager software. All cDNA samples were assayed in triplicates. We chose the ACT1 housekeeping gene as the endogenous normalizer because its expression was constant.

Mutagenesis of TEL2
Mutagenesis of TEL2 to isolate temperature-sensitive was performed by mutagenic PCR coupled to the so-called gap repair method to generate in vivo the plasmids having copied the in vitrogenerated mutant alleles, as described below. To mutagenize TEL2, the 2064 base pair (bp)-long TEL2 ORF plus ,215 bp upstream of the ATG and ,270 bp downstream of the stop codon was amplified by error-prone PCR under the following conditions. The concentration of dNTPs was either kept as in standard conditions (200 mM each) or one of the four dNTP concentration changed to 0.5-1.0 mM, those of the other three being kept at 200 mM, and, in both cases, the concentration of MgCl 2 was changed from 1.5 to 3.0 or 4.0 mM and 0.5 mM MnCl 2 added to the reaction. Standard Taq polymerase and PCR buffer (Promega) were used. Following a 30-cycle amplification, the mutated fragments were transformed into a tel2::KanMX4 strain in which tlc1 null strain was used to ascertain that the highlighted band indeed corresponds to TLC1 RNA. ACT1 and SCR1 RNA levels were measured to serve as loading controls. (B) Immunoprecipitation-Western experiments aiming at assessing Myc 18 -Est2 levels (construct integrated at EST2 genomic locus, under the control of native promoter) indicate that EST2/telomerase levels are depressed in the tel2-15 ts mutant strain grown at the semi-permissive temperature for growth of 29uC for 2 hr or overnight (ON). The rvb2-2 and rvb2-24 mutants (NG, MC, submitted for publication) were used as controls. (C) Levels of EST2 mRNA, coding for the protein subunit of budding yeast telomerase, were measured relative to those of ACT1, in wild-type cells (wt, normalized to 1.0) and in the temperature-sensitive tel2- 15  the deletion was complemented by wild-type TEL2 borne on a CEN-URA3 (YClacp33) plasmid, together with a centromeric, LEU2-based (YCplac111), plasmid carrying TEL2 ORF plus the same flanking regions and made linear by digestion with NruI and StuI, endogenous sites located ,50 base pairs post-initiating ATG and ,140 base pairs downstream of the stop codon, respectively. After shuffling out the wild-type TEL2 allele on 5-FOA medium (which counter-selects for URA3 in the plasmid), colonies growing at 24uC were replica plated at 33-37uC to identify thermosensitive colonies. We could isolate six such tel2 alleles. All six alleles were sub-cloned from the initial centromeric vector into a single copy vector (LEU2-based, YIp128) that was subsequently integrated at the LEU2 locus and re-transformed into the original tel2 null YCplac33-TEL2 strain, followed by 5-FOA counter-selection.
To isolate tel2 mutants potentially deregulated in telomere length control, the initial processes were similar to those exposed above. However, at the step next after shuffling out of the wild type TEL2 allele on 5-FOA medium, transformants were picked out randomly from the plates, propagated for ,20 days at 24uC to attain telomere length equilibrium and selected by Southern blot analysis, as described above.

Cell viability assays
The viability of cells previously grown in liquid was determined by performing and analyzing the so-called ''drop tests''. To do this, cells from exponential growth cultures were counted with a hematocytometer and the cultures were then serially diluted by 1/ 5 th or 1/10 th and spotted onto either selective plates or YEPD non-selective plates, as required, and incubated at the desired temperature for 2-3 days before being photographed. Figure S1 (see Appendix S1): Physical interactions between Tel2 and Rvb2. (A) Endogenous Rvb2 and Tel2 physically associate even when telomerase access to the telomeres is restricted or even totally prevented. Immunoprecipitation-Western (IP-West.) blotting experiments on endogenous HA 2 -RVB2 and Myc 2 -TEL2 in an otherwise wild type background (2 nd lane) or in strains containing either an est1 (3 rd lane), a yku70 (4 th lane) or est1 yku70 null mutation (5 th lane) to restrict telomerase access to the telomeres either in the G1 and early S phases (yku70 null), late S, G2 and M phases (est1 null), or in all cell cycle phases (est1 null yku70 null) [2]. The 1 st lane is a control with no Myc 2 -TEL2 expression to assess for HA 2 -Rvb2 background. (B) Overproduced full length Tel2-Myc 3 physically interacts with full length Rvb2-HA 2 , in both directions. Tel2-Myc 3 was specifically detected in Rvb2-HA 2 immunoprecipitate (left panel, top gel) and reciprocally (right panel, bottom gel) over the background; compare, in each panel, lanes 3 (strain harboring both constructs) with lanes 1 and 2 (strains harboring the single constructs). Overexpression of the constructs, here and below in C, under the control of the inducible GAL1-10 promoter was for 2 hr at 29uC. (C) Overproduced Tel2 and Rvb2 still physically associate in the absence of Tel1 (which has been previously reported to physically associate with Tel2; [3]), also creating a situation in which telomerase access to the telomeres is restricted. The association between Rvb2-HA 2 and Tel2-Myc 3 could be detected whether the immunoprecipitation (IP) was directed against Tel2-Myc 3 (left panel) or, on the opposite, against Rvb2-HA 2 (right panel), and the Western blotting (West.) realized with anti-HA and anti-Myc monoclonal antibody, respectively, as indicated below each gel. (D) Tel2 did not physically associate with amino acids 420 to 740 of Est2, comprising the entire catalytic domain when they were overproduced, for 2 hr at 29uC, under the control of the inducible GAL1-10 promoter in a high-copy vector. These experiments were performed using strains expressing either full length Tel2 (Tel2-Myc 3 , left panel), or Tel2-Myc 3 first half (middle panel) or Tel2-Myc 3 second half (right panel), as well as Est2's catalytic domain (Est2 420-740 -HA 2 ). (EPS) Figure S2 (see Appendix S1): Delineating the Tel2 and Rvb2 fragments needed for physical interaction. (A) The first half of Tel2, Tel2-(1-343)-Myc 3 , but not its second half, Tel2-(344-688)-Myc 3 , had affinity with full length Rvb2-HA 2 . The two panels show the same thing in two distinct experiments. Note, in bottom gels, that the first 343 amino acids of Tel2 migrated at an apparent molecular weight of around 33 kD, while roughly the same number of amino acids in the second part, the last 345 ones, migrated at an apparent molecular weight of around 48 kD.  Figure S3 The tel1 null yku70 null double mutant exhibited shorter telomeres than the two single mutants tel1D and yku70D. The mutants were grown (at 29uC) sufficiently long to harbor telomeres having attained length equilibrium. See the legend to Figure 3 for more technical detail. (EPS) Figure S4 Immunoprecipitation-Western experiments, performed as described in the legends to Fig. 5B and S1, aiming at assessing Myc 18 -Est2 levels (construct integrated at EST2 genomic locus, under the control of native promoter) indicate that EST2/telomerase levels are depressed in the tel2-19 ts mutant strain grown at the semi-permissive temperatures for growth of 296C for 2 hr or restrictive temperature of 346C for 4 hr, but not at the permissive temperature of 246C, as indicated. (EPS) Figure S5 (A) est2 null mutants prior to telomeric senescence are only moderately temperature-sensitive. Therefore, the strong ts phenotype of the tel2-15 and tel2-19 mutants cannot be due to the diminution of EST2 transcript levels in these mutants. As soon as EST2 has been genetically inactivated, telomeric DNA damage accumulates, thus presumably provoking the slight growth defect seen at all temperatures tested in these cells compared with the wild type. Note that at elevated temperatures, e. g. 36uC, cells divide faster thus accelerating accumulation of damage and progression through senescence and provoking these growth defects. Finally, the absence of telomerase provokes a general destabilization of the telomeres that will eventually activate the DNA damage checkpoint and slow down cell cycle progression. Ten-fold serial dilutions (from left to right in each condition) of cultures of the indicated relevant genotype were grown for 3 days on YEPD agar at the indicated temperature and photographed. (B) Southern blot analysis of telomere length in the tel2-ts and pre-senescing est2D mutants grown at 29uC.