Elongator Complex Influences Telomeric Gene Silencing and DNA Damage Response by Its Role in Wobble Uridine tRNA Modification

Elongator complex is required for formation of the side chains at position 5 of modified nucleosides 5-carbamoylmethyluridine (ncm5U34), 5-methoxycarbonylmethyluridine (mcm5U34), and 5-methoxycarbonylmethyl-2-thiouridine (mcm5s2U34) at wobble position in tRNA. These modified nucleosides are important for efficient decoding during translation. In a recent publication, Elongator complex was implicated to participate in telomeric gene silencing and DNA damage response by interacting with proliferating cell nuclear antigen (PCNA). Here we show that elevated levels of tRNALys s2 UUU, tRNAGln s2 UUG, and tRNAGlu s2 UUC, which in a wild-type background contain the mcm5s2U nucleoside at position 34, suppress the defects in telomeric gene silencing and DNA damage response observed in the Elongator mutants. We also found that the reported differences in telomeric gene silencing and DNA damage response of various elp3 alleles correlated with the levels of modified nucleosides at U34. Defects in telomeric gene silencing and DNA damage response are also observed in strains with the tuc2Δ mutation, which abolish the formation of the 2-thio group of the mcm5s2U nucleoside in tRNALys mcm5s2UUU, tRNAGln mcm5s2UUG, and tRNAGlu mcm5s2UUC. These observations show that Elongator complex does not directly participate in telomeric gene silencing and DNA damage response, but rather that modified nucleosides at U34 are important for efficient expression of gene products involved in these processes. Consistent with this notion, we found that expression of Sir4, a silent information regulator required for assembly of silent chromatin at telomeres, was decreased in the elp3Δ mutants.


Introduction
Elongator complex, first identified in Saccharomyces cerevisiae, consists of a core complex, Elp1-Elp3 and a sub-complex, Elp4-Elp6 [1][2][3]. Orthologs of Elp1 to Elp4 has been identified in higher eukaryotes and a six-subunit Elongator complex has been purified from humans [4][5]. In yeast, Elongator mutants display pleiotropic phenotypes in multiple cellular processes including RNA polymerase II transcription and exocytosis [1][2][3][6][7][8][9]. A crucial observation in understanding the role of the yeast Elongator complex was the discovery of its requirement for formation of 5-carbamoylmethyl (ncm 5 ) and 5-methoxycarbonylmethyl (mcm 5 ) side chains of wobble uridines [10]. In yeast Elongator mutants, the formation of ncm 5 and mcm 5 side chains were abolished in the 11 tRNA species that normally contain one of these two side chains [10][11][12]. Elongator complex in C. elegans and A. thaliana is also required for formation of ncm 5 and mcm 5 side chains at wobble uridines [13][14]. When the ncm 5 and mcm 5 side chains were eliminated, the corresponding tRNA species acted less efficiently in translation [12]. Although lack of modifications at position 5 affects the decoding properties of many tRNAs, it appears that the pleiotropic phenotypes of Elongator mutants are predominantly due to decreased translational decoding by hypomodified tRNA Lys s 2 UUU and tRNA Gln s 2 UUG [15]. Simultaneous over-expression of hypomodified tRNA Lys and tRNA Gln s 2 UUG , which both have the mcm 5 s 2 U modification at wobble position U 34 in wild type strains, compensated all phenotypes observed in Elongator mutants including those in RNA polymerase II transcription and exocytosis without restoring formation of ncm 5 and mcm 5 side chains in tRNA [15]. These observations not only argue against a direct involvement of Elongator complex in other cellular processes than tRNA modification, but they also suggest that the mcm 5 side chain is important for efficient translation of mRNAs encoding gene products critical for the processes in which Elongator mutants generate phenotypes.
In eukaryotes, the whole genome is packed into a nucleoprotein complex known as chromatin through which the genetic material is processed to regulate cellular processes including transcription, cell division, DNA replication and DNA repair [16][17].
Chromatin properties can be altered by the posttranscriptional modifications of histones including acetylation, methylation, phosphorylation and ubiquitination [16]. The Elp3 protein of Elongator complex contains a tentative histone acetyltransferase (HAT) domain in the C-terminal region and the histone acetylation levels are decreased in elp3 mutants [7]. However, the reduced histone acetylation levels in the elp3 mutant were restored by increased expression of tRNA Lys s 2 UUU and tRNA Gln s 2 UUG , indicating that the involvement of Elongator complex in chromatin remodeling is indirect [15]. In addition to the HAT domain, Elp3 contains an N-terminal region with sequence similarity to the radical S-adenosylmethionine (SAM) enzymes [18]. A recent report showed that Elongator mutants have a partial loss of telomeric gene silencing and are sensitive to DNA damage agents [19]. It was also observed that strains with different point mutations in the ELP3 gene, resulting in amino acid substitutions in the radical SAM and HAT domains, displayed differences in telomeric gene silencing and DNA damage response [19]. The participation of Elongator complex in telomeric gene silencing and DNA damage response was linked to its interaction with proliferating cell nuclear antigen (PCNA), a protein involved in DNA replication and DNA repair [19].
In this report, we demonstrate that defects observed in DNA damage response and telomeric gene silencing of yeast Elongator mutants are caused by the absence of wobble uridine tRNA modifications. So far, all phenotypes observed in yeast Elongator mutants can be explained by their influence on tRNA modification. We conclude that the primary role of Elongator complex in yeast is in formation of ncm 5 and mcm 5 side chains at U 34 of tRNAs.

Results
Elevated levels of hypomodified tRNA Lys s 2 UUU , tRNA Gln s 2 UUG , and tRNA Glu s 2 UUC suppress defects in telomeric silencing and DNA damage response induced by Elongator mutants In a recent report, Elongator mutants were shown to have decreased telomeric gene silencing, which was investigated by using an ura3-1 strain with a wild-type copy of the URA3 gene inserted near the left telomere of chromosome VII [19]. Cells with increased expression of Ura3 show reduced growth on plates containing 5-fluoroorotic acid (5-FOA) since the nontoxic 5-FOA is converted to the toxic 5-flurouracil by the URA3 gene product. In such a strain, 30-50% of the cell population are resistant to 5-FOA [20]. The URA3 gene was expressed in a population of cells in both wild type and elp3D strains ( Figure 1A). However, the elp3D strain grew poorly on the 5-FOA containing plates compared to the wild type ( Figure 1A), suggesting that telomeric gene silencing was decreased in the elp3D strain. Since we earlier showed that the primary function of Elongator complex is in formation of wobble uridine tRNA modifications, we investigated whether increased levels of hypomodified tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC could suppress the defects in telomeric gene silencing of an elp3D strain. Over-expression of these tRNA species significantly improved the growth of the elp3D strain on 5-FOA plates ( Figure 1B). The telomeric gene silencing defect of Elongator mutants was also investigated by using a color assay with the ADE2 marker inserted near the telomeric region. The elp3 mutant forms white color colonies due to loss of silencing of ADE2, which could be rescued by increased expression of tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC (data not shown). This observation confirmed that Elongator mutants have a defect in telomeric gene silencing, which is caused by a translational dysfunction. The decreased telomeric silencing observed in other Elongator deletion mutants (elp1D, elp2D, elp4D, elp5D and elp6D) was also suppressed by elevated levels of tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC ( Figure 1C). Elongator mutants are also sensitive to DNA damaging agents, especially hydroxyurea (HU) [19] ( Figure 2). Similar to the defect in telomeric gene silencing, the HU sensitivity of Elongator mutants was suppressed by elevated levels of tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC (Figure 2). Collectively, these observations indicate that the reduced gene silencing in telomeric regions and the defect in DNA damage response of Elongator mutants is caused by inefficient translation due to lack of wobble uridine tRNA modifications.
To investigate which of the tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC species most efficiently suppressed the defects in telomeric silencing and DNA damage response of the elp3D strain, we introduced plasmids encoding these tRNAs independently or in various combinations into the mutant. Increased expression of tRNA Lys s 2 UUU alone could efficiently suppress the telomeric silencing defect and the HU-sensitivity of an elp3D strain ( Figure  S1). Simultaneous over-expression of tRNA Lys s 2 UUU , tRNA Gln and tRNA Glu s 2 UUC gave a minor improvement in suppression of the telomeric gene silencing defect compared to over-expression of tRNA Lys s 2 UUU alone ( Figure S1A). In the HU sensitivity assay, increased expression of tRNA Lys s 2 UUU together with tRNA Gln improved the suppression compared to that of tRNA Lys s 2 UUU and was as good as elevated levels of tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC ( Figure S1B). These results indicate that certain open reading frames, encoding gene products critical for telomeric gene silencing and DNA damage response, might be enriched in AAA, CAA and GAA codons. Of these three codons, translation of AAA codons by tRNA Lys s 2 UUU seems to be most affected by lack of the mcm 5 side chain.
Synergistic growth reduction and HU sensitivity of elp3D asf1D or elp3D rtt109D strains are compensated by increased expression of tRNA Lys s 2 UUU , tRNA Gln s 2 UUG , and tRNA Glu s 2 UUC Asf1 functions as a histone chaperone to direct the histone acetyltransferase Rtt109 in substrate selection and stimulate its acetyltransferase activity [21][22][23]. The combination of elp3D asf1D or elp3D rtt109D mutations causes synergistic phenotypes to the strains, such as a more pronounced reduction in growth and increased sensitivity to HU ( Figure 3 and Figure S2), which was suggested to be caused by loss of histone acetylation in the elp3D strain [19]. GCN5 encodes a histone acetyltransferase that

Author Summary
Elongator is a conserved protein complex in eukaryotes. Studies in yeast, worms, and plants have revealed that Elongator complex is required for formation of mcm 5 and ncm 5 side chains at wobble uridines in a subset of tRNA species. The primary function of Elongator complex in yeast is to modify U 34 in tRNAs. Lack of these tRNA modifications causes pleiotropic phenotypes in yeast Elongator mutants due to inefficient translation. In this report, we demonstrate that the defects in telomeric silencing and DNA damage response observed in yeast Elongator mutants are a consequence of a tRNA modification defect. We suggest that the requirement of Elongator complex in tRNA modification is conserved in all eukaryotes, and diseases linked to human Elongator mutations may involve impaired translation due to lack of tRNA modifications.
acetylate H2B and H3 [24][25]. Previously it was shown that the elp3D gcn5D mutations generate a synergistic growth reduction [26]. However, increased levels of hypomodified tRNAs suppressed the synergistic growth reduction caused by the elp3D gcn5D mutations, and restore the histone acetylation levels in the elp3D mutant but not in the gcn5D strain [15]. When we over-expressed tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC from a high copy vector in the elp3D asf1D or elp3D rtt109D double mutants, the growth reduction and HU sensitivity of the double mutants were similar to the defects observed in an asf1D or rtt109D strain, respectively ( Figure 3 and Figure S2). These observations support the earlier conclusion that Elp3 is not directly required for histone acetylation [15].
Wobble uridine tRNA modification levels correlate to phenotypic variations generated by different mutant alleles of the ELP3 gene Elp3 contains two conserved domains, a radical S-adenosylmethionine (SAM) domain in the N-terminal region and a putative histone acetyltransferase (HAT) domain located in C-terminal end ( Figure 4A). Most strains expressing Elp3 proteins with amino acid substitutions in these two domains showed a reduction in telomeric gene silencing and HU resistance [19] (Figure 4). The elp3-C103A and elp3-G168R mutations did not influence telomeric gene silencing and HU sensitivity ( Figure 4B and 4C) [19]. The elp3-Y540A and elp3-Y541A mutations partially reduced telomeric gene silencing and increased HU sensitivity but not as much as elp3D ( Figure 4B and 4C) [19]. The remaining strains were similar as an elp3D null strain in telomeric gene silencing and HU sensitivity ( Figure 4B and 4C) [19]. Moreover, all strains carrying individual mutations listed in Figure 4A except for elp3-C103A were resistant to Kluyveromyces lactis killer toxin (data not shown), indicating that these mutants have a defect in formation of wobble uridines tRNA modification [11].
To examine the status of wobble uridine tRNA modification in these elp3 mutants, total tRNAs from these mutants were isolated and analyzed by HPLC. The elp3-C103A and elp3-G168R mutants, which did not have defects in telomeric silencing and DNA damage response, had 96% and 51% mcm 5 s 2 U left, respectively ( Figure 5, Table 1). Mutations in the HAT domain did not completely eliminate the formation of wobble uridine modifications, both elp3-Y540A and elp3-Y541A have 2 or 6% mcm 5 s 2 U left compared to the wild type ( Figure 5, Table 1). In the rest of mutants, the mcm 5 side chain formation was entirely abolished ( Figure 5, Table 1). We conclude that phenotypes exhibited by elp3 mutants correlate with the levels of wobble uridine tRNA modification.

Different mcm 5 modification levels correlate with ochre stop codon read through by a suppressor tRNA
Our observations suggest that phenotypes of Elongator mutants are caused by an inefficient translation due to lack of tRNA modification. If our model is correct, reduction in modification levels in elp3 mutants should result in decreased translation efficiency. To analyze whether the modification levels of different elp3 mutants listed in Table 1 influence translation The wild type (UMY2584) and elp3D (UMY3790) strains were 10-fold diluted, spotted on SC, SC-Ura and SC+5-FOA plates, and incubated at 30uC for 2 days. (B) The elp3D strain (UMY3790) with plasmids, pRS315-ELP3, pRS425-tK-tQ-tE or pRS425, were 10-fold diluted, spotted on SC-Leu, SC-Leu-Ura and SC-Leu+5-FOA plates, and incubated at 30uC for 2 days. (C) The wild type (UMY2584), elp1D (UMY3788), elp2D (UMY3789), elp4D (UMY3791), elp5D (UMY3792) and elp6D (UMY3793) with plasmids pRS425-tK-tQ-tE or pRS425 were treated as described in (B). Abbreviations for the tRNA genes encoding tRNA Lys UUU , tRNA Gln UUG and tRNA Glu UUC are tK, tQ and tE, respectively. doi:10.1371/journal.pgen.1002258.g001 efficiency, we used a dual-luciferase reporter system ( Figure 6A) [27] to measure the ochre stop codon read through by a suppressor tRNA encoded by the SUP4 allele. The SUP4 allele encodes a tRNA Tyr suppressor with a G 34 to U 34 substitution in its anticodon. The U 34 of this suppressor tRNA is modified at position 5 with a mcm side chain [10]. Presence of this modification improves the ability of the suppressor tRNA to read UAA ochre stop codons [10,12].
In the dual-luciferase construct, the Renilla and firefly luciferase genes are separated by an UAA ochre stop codon [27]. Read through of the ochre stop codon was determined by calculating the ratio of firefly luciferase activity to Renilla luciferase activity. This ratio was compared to the value obtained from a control construct in which a CAA codon replaces the UAA stop codon ( Figure 6A). Due to lack of mcm 5 side chain in the SUP4 tRNA, the stop codon read through in the elp3D strain is reduced to 46% of wild type (t-test, p = 0.001), supporting that the mcm 5 side chain is important for efficient decoding ( Figure 6B). In the elp3-G168R mutant, in which the mcm 5 side chain is reduced to 51%, the level of read through was significantly decreased compared to that in wild type (t-test, p = 0.008), but is higher than that observed in strains carrying the elp3-Y540A, elp3-Y541A or elp3D alleles (t-test, p = 0.04 and 0.03 respectively) ( Figure 6B). In the elp3-Y540A and elp3-Y541A mutants, a small fraction of total tRNA was modified (2-6%) ( Figure 5, Table 1), which contributed to an improvement of stop codon read through by the SUP4 suppressor tRNA compared to the elp3D strain (t-test, p = 0.004 and 0.006 respectively) ( Figure 6B). In mutant alleles eliminating formation of the mcm 5 side chain, no differences were observed in stop codon read through by the SUP4-encoded suppressor tRNA compared to the elp3 null mutant ( Figure S3). These data show that reduced mcm 5 modification levels correlate with decreased translational efficiency.  , and tRNA Glu s 2 UUC bypass the phenotypes of asf1D elp3D double mutants. (A) The asf1D elp3D strain (UMY3805) was transformed with pRS315-ELP3, pRS425-tK-tQ-tE or pRS425. Transformants were streaked on SC-Leu plates and incubated at 30uC or 37uC for 2 days. (B) Ten fold dilutions of elp3D (UMY2843), asf1D (UMY3800) and asf1D elp3D (UMY3805) strains carrying either pRS425-tK-tQ-tE or pRS425 were spotted on SC-Leu and SC-Leu+50 mM HU plates, and incubated 4 days at 30uC. The elp3D (UMY2843) transformed with pRS315-ELP3 was used as control. Abbreviations for the tRNA genes encoding tRNA Lys UUU , tRNA Gln UUG , and tRNA Glu UUC are tK, tQ and tE, respectively. doi:10.1371/journal.pgen.1002258.g003 Defects in telomeric silencing and DNA damage response are also observed in strains unable to form the s 2 group of mcm 5 s 2 U Our findings that the defects in telomeric silencing and DNA damage response in Elongator mutants were bypassed by elevated levels of tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC indicated that the mcm 5 side chain in tRNA is critical for the expression of gene products in these two processes ( Figure 1 and Figure 2). In addition to the mcm side chain at position 5 of U 34 , these three tRNAs also contain a 2-thio group forming mcm 5 s 2 U. Since the s 2 group is also important for decoding [12,15,28], we hypothesized that strains deficient in formation of the 2-thio group might also display defects in telomeric silencing and DNA damage response as Elongator mutants. Tuc2 in yeast is required for the formation of the 2-thio group of the mcm 5 s 2 U nucleoside [15]. In a tuc2D strain, the formation of s 2 group is abolished. As expected, telomeric gene silencing was decreased in the tuc2D strain ( Figure 7A). This strain was also sensitive to 50 mM HU nearly to the same extent as observed in Elongator mutants (Figure 2 and Figure 7B). The defects in telomeric gene silencing and DNA damage response were completely suppressed by increased levels of tRNA Lys mcm 5 UUU , tRNA Gln mcm 5 UUG and tRNA Glu mcm 5 UUC (Figure 7). The phenotypes of Elongator and tuc2D mutants demonstrates that a translational dysfunction due to lack of U 34 modifications in tRNA Lys UUU , tRNA Gln UUG and tRNA Glu UUC causes the defects in telomeric gene silencing and DNA damage response.

Sir4 expression is decreased in an elp3D strain
Among the three tRNA species responsible for the suppression of elp3D induced phenotypes, increased expression of tRNA Lys gives the best suppression of the defect in telomeric gene silencing ( Figure S1). Since tRNA Lys mcm 5 s 2 UUU decodes AAA codons, elimination of the mcm 5 side chain from tRNA Lys mcm 5 s 2 UUU in the elp3D strain could influence the decoding efficiency of AAA codons. Therefore, we searched for open reading frames highly enriched in AAA codons (unpublished results). This analysis lead to the identification of SIR4, encoding a silent information regulator in yeast. Based on this observation, we hypothesized that the telomeric gene silencing defect of the elp3D mutant might be caused by decreased Sir4 expression. Accordingly, the Sir4 protein levels in the elp3D mutant were decreased to 34% of wild type ( Figure 8A). The decreased Sir4 levels were restored to 80% of wild-type by increased expression of tRNA Lys s 2 UUU , tRNA Gln and tRNA Glu s 2 UUC , and to 74% of wild-type by elevated levels of tRNA Lys s 2 UUU alone ( Figure 8A and data not shown). We also observed that SIR4 mRNA levels were reduced to 76% of wildtype ( Figure 8B), which cannot account for the decreased Sir4 protein levels. In addition, introducing the SIR4 gene on a high copy vector significantly suppressed the telomeric gene silencing defect of the elp3D strain, confirming that this defect seems to be caused by decreased Sir4 expression ( Figure 8C). However, we do not exclude the possibility that there might be other open reading frames enriched in AAA codons whose translation is also affected and which might weaken silencing, directly or indirectly.

Discussion
Elongator complex was initially identified by its apparent association with the elongating form of RNA polymerase II, implicating a role in PolII transcription [1]. However, its requirement in transcription was controversial based on its cytoplasmic localization and failure to detect this complex on actively transcribed genes [8,[29][30]. We discovered that Elongator complex was required for formation of mcm 5 and ncm 5 side chains at wobble uridines of tRNA [10]. The participation of Elongator complex in PolII transcription and exocytosis was indirect as elevated expression of hypomodified tRNA Lys s 2 UUU and tRNA Gln s 2 UUG could suppress previously reported phenotypes of Elongator mutants without restoring tRNA modification [15]. Recently, it was reported that Elongator complex modulates telomeric gene silencing and DNA damage response by its interaction with PCNA and its requirement for histone acetylation [19]. Since the histone acetylation defect of the elp3D mutant could be completely suppressed by increased expression of tRNA Lys s 2 UUU and tRNA Gln s 2 UUG [15], we assumed that Elongator complex indirectly participated in telomeric gene silencing and DNA damage response.
In this report, we show that the defects in telomeric gene silencing and DNA damage response in Elongator mutants were also suppressed by increased expression of hypomodified tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC (Figure 1, Figure 2, and Figure S1). Thus, all phenotypes exhibited by Elongator mutants except the tRNA modification defect are overcome by elevated tRNA levels, indicating that the major function of this complex, at least in yeast, is in the formation of mcm 5   and tRNA Glu s 2 UUC were over-expressed in Elongator mutants, the HU sensitivity phenotype, but not the defect in telomeric gene silencing, was fully suppressed (Figure 1 and Figure 2). Since Elongator mutants affect the mcm 5 and ncm 5 side chain formation in 11 tRNA species, it is possible that poor translation of codons decoded by any of the other 8 hypo-modified tRNA species contributes to the defect in telomeric gene silencing, but not the HU sensitivity. In addition to the mcm side chain at position 5, U 34 of tRNA Lys mcm 5 s 2 UUU , tRNA Gln mcm 5 s 2 UUG and tRNA Glu mcm 5 s 2 UUC are also thiolated at position 2. If our model is correct that the phenotypes observed in Elongator mutants are a consequence of inefficient translation, strains lacking the 2-thio group in tRNA Lys mcm 5 s 2 UUU , tRNA Gln mcm 5 s 2 UUG and tRNA Glu mcm 5 s 2 UUC will have similar phenotypes as Elongator mutants. We observed that the failure to form the 2-thio group in the tuc2D mutant resulted in defects in telomeric gene silencing and DNA damage response (Figure 7). These defects of the tuc2D mutant were completely suppressed by increased expression of tRNA Lys mcm 5 UUU , tRNA Gln mcm 5 UUG and tRNA Glu mcm 5 UUC . In addition, lack of the methyl ester in mcm 5 side chain at wobble uridines in a trm9D strain has been linked to the defect of DNA damage response [31]. Thus, both mcm 5 and s 2 side chains of mcm 5 s 2 U containing tRNAs are important for efficient expression of gene products required for telomeric gene silencing and DNA damage response. These observations strongly suggest that Elongator complex influence these two processes by promoting efficient translation.  Since increased expression of tRNA Lys s 2 UUU gives the best suppression of the telomeric gene silencing defect in Elongator mutants, we assumed genes encoding products important for this process are enriched in AAA codons. One such gene is SIR4. We demonstrate that Elongator mutants influence telomeric gene silencing by impairing efficient expression of SIR4. Even though we observed a slight reduction in SIR4 mRNA levels in the elp3D mutant, it cannot fully explain the decrease in Sir4 protein levels, and it is unclear if this reduction is caused by reduced transcription or increased decay of the poorly translated mRNA.
Recently, it was discovered that Elongator complex in C. elegans and A. thaliana is also required for formation of mcm 5 and ncm 5 side chains at wobble uridines of tRNA [13][14], indicating that this function of Elongator complex might be conserved in eukaryotes. In multicellular organisms, Elongator complex has also been linked to multiple processes including transcription, cytoplasmic kinase signaling and development [32][33][34]. Two recent articles suggested that Elongator complex was also required for a-tubulin acetylation and played a role in neurological processes in both mice and C. elegans [35][36]. In early Figure 7. The tuc2D strain is deficient in telomeric gene silencing and show increased HU sensitivity. (A) The wild type strain (UMY2584) harboring plasmid pRS425 and the tuc2D mutant (UMY3804) harboring plasmids pRS425-tK-tQ-tE or pRS425 were assayed as described in Figure 1. (B) The wild type strain (UMY2067) harboring plasmid pRS425 and the tuc2D mutant (UMY3442) harboring plasmids pRS425-tK-tQ-tE or pRS425 were assayed as described in Figure 2. Abbreviations for the tRNA genes encoding tRNA Lys UUU , tRNA Gln UUG and tRNA Glu UUC are tK, tQ and tE, respectively. doi:10.1371/journal.pgen.1002258.g007 . Western blot analysis of Sir4-Myc protein levels in the elp3D strain transformed with plasmids pRS315-ELP3, pRS425-tK-tQ-tE or pRS425. The ratios of Sir4-Myc to Actin signals were calculated. The values are shown relative to elp3D pRS315-ELP3 strain, which was arbitrarily set to 1, and are the average of two independent experiments. (B). Northern blot analysis of SIR4 mRNA. The elp3D strain was transformed with plasmids pRS315-ELP3, pRS425-tK-tQ-tE or pRS425. Signals of SIR4 mRNAs were normalized to the non-coding SCR1 transcript. The values are shown relative to elp3D pRS315-ELP3 strain, which was arbitrarily set to 1, and are the average of two independent experiments. (C). The wild type strain (UMY2584) transformed with plasmids pRS424-SIR4 or pRS424, and the elp3D mutant (UMY3790) with plasmids pRS424-SIR4 or pRS424 were assayed as described in Figure 1. For A, B  developmental stages, C. elegans Elongator mutants have a decreased a-tubulin acetylation [36]. However, in adult Elongator mutant worms, normal levels of a-tubulin acetylation were observed, suggesting that Elongator complex is not absolutely required for acetylation of a-tubulin [13,36]. Elongator mutants in C. elegans were also resistant to the acetylcholinesterase inhibitor aldicarb, indicating a reduced efficiency of synaptic exocytosis [13,36]. However, a mutant allele of mec-12, which is completely missing a-tubulin acetylation, was not resistant to aldicarb, suggesting that the defect in synaptic exocytosis of Elongator mutants was not caused by reduced levels of a-tubulin acetylation [13]. Furthermore, mec-17 was discovered to be the a-tubulin acetylase in in Tetrahymena cells, C. elegans, zebrafish and mammalian cells, suggesting that Elongator might indirectly influence a-tubulin acetylation by modulating the expression of a-tubulin acetylase [37]. Based on these observations, it is tempting to speculate that the primary function of Elongator complex in multicellular organism is, as in yeast, in formation of wobble uridine tRNA modifications.
The Elp3 subunit in yeast has an N-terminal radical Sadenosylmethionine (SAM) domain and a C-terminal histone acetyltransferase (HAT) domain. In Methanocaldococcus jannaschii, the radical SAM domain of mjElp3 contains an iron sulfur cluster region and a region that binds SAM [38]. Cysteine residues at positions 96, 101 and 104 are critical for the FeS cluster formation in M. jannaschii [38]. When these corresponding cysteines at position 108, 118 and 121 in the yeast Elp3 were substituted with alanines, it eliminated the activity of yeast Elongator in formation of modified nucleosides at U 34 . In vitro, SAM can bind to M. jannaschii Elp3, but the binding of SAM to Elp3 from S. cerevisiae has not been detected [38][39]. However, when the conserved SAM binding sites (G180R G181R) in the radical SAM domain were mutated in yeast ELP3, a defect in formation of modified nucleosides was observed ( Figure 5, Table 1). This observation shows that the FeS cluster and the SAM binding regions of the radical SAM domain of Elp3 are critical for the tRNA modification reaction. Substitution of glycine at position 168 to arginine, another conserved site located in the SAM binding region, reduced the wobble uridine tRNA modification to 51% of wild type ( Figure 5, Table 1). In telomeric gene silencing and HU sensitivity assays, the elp3-G168R mutant displays the same phenotypes as a wild type strain suggesting that a 49% reduction in the levels of modified nucleosides do not cause phenotypes in telomeric gene silencing and DNA damage response. Two mutations in the HAT domain (Y540A and Y541A) of Elp3 did not entirely eliminate the formation of modified nucleosides at U 34 ; 2 and 6% of mcm 5 s 2 U was detected in each mutant ( Table 1). The residual level of modified nucleosides significantly improves the decoding capacity of the SUP4 encoded suppressor tRNA compared to the unmodified tRNA in the elp3 null mutant ( Figure 6). This observation explains why the elp3-Y540A and elp3-Y541A mutants had increased telomeric silencing and reduced HU sensitivity compared to the elp3D strain ( Figure 4).
Among the elp3 mutants described in Table 1, the elp3-G168R mutant, having 51% of modified nucleoside left ( Figure 5 and Table 1), has the same phenotype as a wild type strain with respect to phenotypes in telomeric gene silencing and DNA damage response ( Figure 4). However, this strain is resistant to killer toxin (data not shown), a phenotype tightly connected to wobble uridine tRNA modification [11]. The c subunit of killer toxin is a tRNA endonuclease which cleaves tRNA at the anticodon region [11]. The mcm 5 side chain at U 34 of tRNA is important for the substrate recognition by c toxin. In the elp3-G168R mutant, a fraction of the U 34 tRNAs are missing the mcm 5 side chain and the mutant is resistant to c toxin (data not shown). However, the modified tRNAs in the elp3-G168R support the efficient expression of gene products required for telomeric gene silencing and DNA damage response. Thus, strains with tRNAs partially modified at U 34 show weaker or no phenotypes compared to Elongator deficient strains.
In summary, the major function of Elongator complex in yeast is in formation of wobble uridine tRNA modifications and this function is probably conserved in eukaryotes. We suggest that when new phenotypes of Elongator mutants are discovered in yeast, an important first step is to investigate whether the phenotypes can be suppressed by over-expressing tRNA Lys s 2 UUU , tRNA Gln s 2 UUG and tRNA Glu s 2 UUC .
A two-step gene replacement procedure was used to obtain strains with different mutant alleles of ELP3. Plasmids pABY1672 (elp3-C103A), pABY1673 (elp3-C108A), pABY1676 (elp3-C118A), pABY1677 (elp3-C121A), pABY1984 (elp3-G168R) and pABY1985 (elp3-G180R G181R) were digested with EcoRI and the linearized fragments were transformed into the UMY2894. Transformants were selected on SC-Ura plates and streaked on YEPD plates. Eight independent colonies on YEPD plates were picked and streaked on 5-FOA containing plates. The strains with elp3 mutant alleles except for elp3-C103A were identified by their resistance to killer toxin and confirmed by sequencing. In order to identify the elp3-C103A mutant, DNA isolated from several candidates were sequenced.

tRNA isolation and HPLC analysis
Cells were grown at 30uC in 100 ml YEPD and harvested at OD600 = 1.5,2. The cell pellet was resuspended in 3 ml 0.9% NaCl. The cell suspension was vortexed at room temperature for 30 minutes in the presence of 8 ml water-saturated phenol and vortexed for another 15 minutes after adding 0.4 ml chloroform. Centrifugation was carried out at 12000 g for 20 minutes. The water phase was collected and re-extracted with phenol. The final water phase was collected, mixed with 2.5 volume 99.5% ethanol and kept at 220uC for at least 3 hours. Total RNA was pelleted at 12000 g for 20 minutes. The RNA pellet was dissolved in 5 ml DE52 binding buffer (0.1 M Tris.HCl pH 7.4 and 0.1 M NaCl) and loaded onto the DE52 cellulose column. The column was washed twice with 7 ml DE52 binding buffer and the tRNA was eluted with 7 ml elution buffer (0.1 M Tris.HCl pH 7.4 and 1 M NaCl). The tRNA was precipitated with 0.7 volume of isopropanol at 220uC for at least 3 hours and pelleted by centrifugation at 12000 g for 20 minutes. The pellet was washed once with 70% ethanol and dissolved in 50 ml MQ. Purified tRNA was digested with Nuclease P1 for 16 hrs at 37uC and treated with bacterial alkaline phosphatase for 2 hours at 37uC. The hydrolysate was analyzed by high pressure liquid chromatography with a Develosil C-30 reverse-phase column as described [41].

Telomeric gene silencing and DNA damage response assays
To investigate the defect in telomeric gene silencing of Elongator mutants, 10-fold dilutions of freshly cultivated yeast cells were spotted on 5-FOA containing plates and control plates. Plates were incubated at 30uC for 2 days. To analyze the DNA damage response, 10 fold dilutions of freshly cultivated yeast cells were spotted on the plates containing 50 mM HU and control plates. The results were scored after 2 days of incubation at 30uC.

Dual-luciferase reporter assay
The luciferase activities were measured by GloMax 20/20 luminometer (Promega) and the dual-luciferase reporter assay system (Promega). Cells were grown to 0.5 OD 600 and diluted 10 fold before use. 20 ml of diluted cell culture was mixed with 100 ml passive lysis buffer, vortexed for 12 seconds and 20 ml of cell lysate was used to determine the luciferase activity. Each culture was measured 3 times and 3 independent experiments were performed.