Elongator function in tRNA wobble uridine modification is conserved between yeast and plants

Based on studies in yeast and mammalian cells the Elongator complex has been implicated in functions as diverse as histone acetylation, polarized protein trafficking and tRNA modification. Here we show that Arabidopsis mutants lacking the Elongator subunit AtELP3/ELO3 have a defect in tRNA wobble uridine modification. Moreover, we demonstrate that yeast elp3 and elp1 mutants expressing the respective Arabidopsis Elongator homologues AtELP3/ELO3 and AtELP1/ELO2 assemble integer Elongator complexes indicating a high degree of structural conservation. Surprisingly, in vivo complementation studies based on Elongator-dependent tRNA nonsense suppression and zymocin tRNase toxin assays indicated that while AtELP1 rescued defects of a yeast elp1 mutant, the most conserved Elongator gene AtELP3, failed to complement an elp3 mutant. This lack of complementation is due to incompatibility with yeast ELP1 as coexpression of both plant genes in an elp1 elp3 yeast mutant restored Elongator's tRNA modification function in vivo. Similarly, AtELP1, not ScELP1 also supported partial complementation by yeast–plant Elp3 hybrids suggesting that AtElp1 has less stringent sequence requirements for Elp3 than ScElp1. We conclude that yeast and plant Elongator share tRNA modification roles and propose that this function might be conserved in Elongator from all eukaryotic kingdoms of life.


Introduction
Studies in fields as disparate as leaf development (Nelissen et al., 2003; and drought resistance (Chen et al., 2006) in plants, neurodegeneration in humans (Anderson et al., 2001;Slaugenhaupt et al., 2001), cytotoxicity of a fungal toxin (Frohloff et al., 2001) and transcription elongation  have surprisingly converged on a conserved protein complex termed Elongator. The protein complex is composed of two subcomplexes with subunits Elp1, Elp2, and Elp3 as well as Elp4, Elp5 and Elp6 forming the hexameric holo-Elongator (Petrakis et al., 2004). Elp3 has a GNAT-type histone acetyltransferase (HAT) domain and has also been named Kat9 K-acetyltransferase to conform with recent acetyltransferase nomenclature (Allis et al., 2007). The Elongator complex was initially identified in yeast to interact with hyperphosphorylated RNA polymerase II and implicated in chromatin modification during transcription elongation Wittschieben et al., 1999;Kristjuhan et al., 2002;Winkler et al., 2002;Gilbert et al., 2004;Kristjuhan and Svejstrup, 2004).
Surprisingly, it turned out that yeast Elongator mutants defective in any of the Elongator subunit genes (ELP1 to ELP6) are lacking tRNA modifications at wobble uridines or thiouridines in position 34 of the anticodon . One of these tRNA modifications, 5methoxycarbonyl-methyl-2-thiouridine (mcm 5 s 2 U), renders Saccharomyces cerevisiae sensitive to a toxin (zymocin) secreted by Kluyveromyces lactis (reviewed in Schaffrath and Breunig, 2000) and leads to the identification of ELP/TOT genes in a screen for zymocin-resistant target of toxin (tot) mutants (Frohloff et al., 2001). The g-subunit (g-toxin) of the zymocin is a tRNA endonuclease that cleaves a subset of tRNAs carrying the modified wobble uridine (mcm 5 s 2 U) Lu et al., 2005;Jablonowski et al., 2006). Genomic sequence analysis suggests that an Elongator complex exists in all eukaryotes. In plants, Elongator mutants have been isolated in a screen for mutants with aberrant leaf morphology. The Arabidopsis thaliana elongata (elo) class of leaf mutants consisting of four loci, elo1 to elo4, is characterized by narrow and elongated shape of leaves and deficiencies in organ growth. ELO1 identified the Arabidopsis homologue of yeast ELP4, ELO2 is homologous to ELP1, and ELO3 to ELP3 (Nelissen et al., 2005). Moreover, the DEFORMED ROOTS AND LEAVES (DRL1) gene is allelic to elo4 and shows sequence similarity to yeast KTI12, which encodes an Elongator partner protein (Fichtner et al., 2002a;Nelissen et al., 2003). The similar phenotypes of these mutants support the view that the gene products function in a complex and suggest that plant Elongator is similar to the one in yeast.
So far the molecular function of the complex is not well defined. Apparently, its acetyltransferase activity is essential but the substrate spectrum may be wider than initially anticipated (Svejstrup, 2007). Phenotypes associated with Elongator deficiency range from impaired zygotic paternal genome demethylation (Okada et al., 2010) to altered microtubule dynamics in human and Caenorhabditis elegans (Creppe et al., 2009;Solinger et al., 2010). Recently, mammalian Elongator was linked to acetylation of a-tubulin and shown to affect the migration and differentiation of cortical neurons (Creppe et al., 2009). In C. elegans Elongator deficiency was also associated with neurological and developmental defects and tRNA modification defects (Chen et al., 2009). C. elegans mutants still contain acetylated a-tubulin although the level may be reduced (Chen et al., 2009;Solinger et al., 2010). It has also been questioned whether histones are primary substrates for the acetyltransfer reaction because transcription-related phenotypes in yeast Elongator mutants could be suppressed by tRNA overexpression (Esberg et al., 2006). Whether protein acetylation is the primary and only biochemical activity of Elongator in eukaryotes is unknown. The pleiotropic phenotypes of yeast Elongator mutants in processes as diverse as translation, exocytosis, filamentous growth and transcriptional silencing (Rahl et al., 2005;Johansson et al., 2008;Abdullah and Cullen, 2009;Li et al., 2009) might be explained by consequential effects of improper tRNA modification or by multiple acetylation substrates for Elongator, but there is also evidence for additional biochemical activities Li et al., 2009;Lipardi and Paterson, 2009).
Here we have addressed the question whether the tRNA modification function is conserved between yeast and plants. We show that in an A. thaliana elp3 mutant, tRNA wobble uridine modifications including mcm 5 s 2 U are compromised. By complementing yeast elp mutants with A. thaliana ELO/AtELP genes we demonstrate that the yeast subunits can assemble with plant polypeptides to form hybrid Elongator complexes indicating high structural similarity between yeast and plant Elongator. Strikingly, despite the fact that AtElp3/ELO3, the most conserved subunit, could structurally replace yeast Elp3 functional complementation with AtELP3/ELO3 was not observed unless ELP1 was simultaneously replaced by AtELP1/ELO2. Taken together the data strongly support the view that the tRNA modification function of Elongator is conserved between yeast and plants and most likely among all eukaryotes.

Formation of a yeast Elongator complex containing AtELP3
The A. thaliana genome contains one and only one homologue for each of the yeast ELP genes and evidence for a similar hexameric complex, composed of two subcomplexes, was recently obtained by tandem affinity purification (Nelissen et al., 2010). We tried to complement yeast elp mutant strains with the corresponding plant cDNAs fused to a yeast promoter. Sensitivity to g-toxin, the active component of the K. lactis killer toxin zymocin provided a sensitive assay for Elongator function. If the heterologous protein integrated into the yeast Elongator complex and functioned in restoring tRNA modification, we expected reversion of the toxin resistance phenotype of the Elongator mutant. Because the AtELP3 subunit is most similar to its yeast homologue, we first tried to complement the elp3 mutant. A c-myc-tagged version of the AtELP3 protein could be produced at levels comparable to those of yeast Elp3-c-myc (see below), but the toxin resistance of the yeast elp3D mutant was unaffected by the AtELP3-c-myc gene (not shown) or an untagged AtELP3 allele (Fig. 1A). Likewise, thermosensitivity and hypersensitivity to caffeine, additional phenotypes of Elongator mutants, were not altered by the plant gene (Fig. 1B). Reintroduction of the yeast ELP3 gene into the elp3D reporter, however, fully complemented all three phenotypes ( Fig. 1A and B).
The failure of plant AtELP3 to substitute for yeast Elp3 function was not due to instability of the protein as AtELP3c-myc was precipitated from total yeast protein extracts at levels comparable to Elp3-c-myc (Fig. 1C, top panel). c-myc-tagged Elp3 and AtElp3 both could co-precipitate the largest Elongator subunit Elp1 indicating interaction between AtELP3 and Elp1 (Fig. 1C, bottom panel).
To analyse whether AtELP3-Elp1 interaction occurred in the context of the Elongator complex we made use of the fact that interaction between the subunits Elp5 and Elp2 depends on the structural integrity of the complex and the presence of Elp3 (Frohloff et al., 2003;Petrakis et al., 2004). We constructed elp3D reporter strains expressing Plant Elongator function in tRNA modification 1083 a c-myc-tagged version of Elp2 and an HA-tagged version of Elp5. As expected, co-immunoprecipitation of Elp2-cmyc with Elp5-HA was not observed when Elp3 was lacking ( Fig. 2A, lane 4) but was found when the elp3D mutant was complemented with the yeast ELP3 gene on a plasmid (Fig. 2A, lane 3). When AtELP3 or AtELP3-cmyc alleles were introduced instead of ELP3, pull-down of Elp2-c-myc by Elp5-HA was less efficient ( Fig. 2A, lanes 5 and 6), but significantly higher than in the empty vector control ( Fig. 2A, lane 4). Thus, plant AtELP3 promotes Elp5-Elp2 interaction in a yeast elp3D mutant background.
Using the same approach we also tested whether the chimeric complex was able to interact with Kti12, a protein that associates with the Elongator complex (Fichtner et al., 2002a;Petrakis et al., 2005). As with Elp5-HA, Kti12-HA was able to pull-down Elp2-c-myc in the presence but not in the absence of yeast Elp3 (Fig. 2B, lanes 3 and 4) and again AtELP3 could substitute for yeast Elp3 in this interaction assay (Fig. 2B, lanes 5 and 6). The latter finding is particularly intriguing because Elp2-Kti12 interaction relies on the association between the two Elongator subcomplexes Elp1-Elp2-Elp3 and Elp4-Elp5-Elp6 (Fichtner et al., 2002b;Frohloff et al., 2003). We conclude A. To test for functional complementation the elp3D mutant (CMY135) was transformed with plasmids containing ELP3 (pFF9), AtELP3 (YEpA4) and vector control (YEplac195) and subsequently with the GAL1 promoter driven g-toxin expression plasmid pHMS14. Transformants were spotted in replica onto glucose-repressing (glc) or galactose-inducing (gal) media and grown for 4 days at 30°C. Growth on galactose indicates g-toxin resistance (Tox R ) and no growth corresponds to g-toxin sensitivity (Tox S ). B. To test for thermosensitivity and hypersensitivity to caffeine strains were serially diluted and replica spotted on YPD plates lacking (control) or containing 7.5 mM caffeine (right) and incubated for 4 days at 30°C and 37°C (middle). C. Anti-c-myc immunoprecipitates (IP) of strains containing chromosomally tagged ELP3-c-myc (FFY3t), AtELP3-c-myc on a plasmid in an elp3D background (CMY307 + yatELP3M) and wild-type parent without epitope (FY1679-8A) were analysed on Western blots probed with anti-c-myc or anti-Elp1 antibodies. Positions of the respective proteins are indicated by arrows. Protein extracts from cells without epitope-tag served as negative control. The asterisk denotes an N-terminally truncated Elp1 form . The relative abundance of this truncated form varies between experiments but is probably not dependent on which Elp3 peptide is present.

Fig. 2. Reconstitution of Elongator subunit interactions by AtELP3
demonstrating AtELP3 integration in the yeast Elongator complex. A. Protein extracts from cells expressing HA-tagged Elp5 together with c-myc-tagged Elp2 in an elp3D background (strains CMY304 and CMY301, respectively) and transformed with plasmids containing ELP3 (pFF9), AtELP3 (YEpA4), AtELP3-c-myc (yatELP3M) or vector control (YEplac195) were immunoprecipitated using anti-HA antibody and analysed by Western blotting. B. Same as (A) but with strain CMY301, which expresses HA-tagged Kti12 instead of Elp5-HA. The anti-HA antibody was used to detect Elp5-HA or Kti12-HA and anti-c-myc antibodies recognized Elp2 and AtELP3. Protein extracts from cells without epitope tag and cells expressing only AtELP3-c-myc served as negative controls. The pre-IPs served as loading control (bottom panels). that AtELP3 can structurally replace Elp3 in the yeast Elongator complex but apparently the complex is not functional.

Plant AtELP1 restores Elongator subunit interactions in an elp1D yeast mutant
Incorporation of AtElp3 into yeast Elongator requires interaction with much less conserved components. Elp1 and AtELP1 display only 19% amino acid identity compared with 67% between Elp3 and AtELP3. To analyse whether AtELP1 could also structurally replace yeast Elp1 in the complex, an elp1D strain that contained the epitopetagged Elongator subunits Elp3-HA and Elp2-c-myc was transformed with an AtELP1 cDNA clone. Consistent with previous reports (Frohloff et al., 2003;Petrakis et al., 2004), we found that the Elp1 subunit is not only required for Elp3-Elp2 interaction (Fig. 3A, lane 2) but also for stability of yeast Elp3. Hence, an HA-tagged version of Elp3 was not detectable in total protein extracts of the elp1D mutant ( Again, as in the AtELP3-expressing elp3 mutant, the efficiency of subunit interactions was somewhat reduced compared with that of the ELP1 transformants. Nonetheless, our data show that AtELP1 like AtELP3 are assembled into complexes where they are able to structurally replace the respective yeast Elongator subunits.

Together, plant AtELP1 and AtELP3 support tRNA modification in yeast
Because Elp3 requires Elp1 for stability, its function may depend on specific contacts between these two proteins explaining the failure of AtELP3 to functionally complement the elp3D mutant in combination with yeast Elp1. To address this possibility we tested whether AtELP3 might function in yeast when supplied with AtELP1. FLAG-AtELP1 was expressed together with the AtELP3-c-myc gene in a yeast elp1D elp3D double mutant. FLAG-AtELP1 could be detected in the anti-c-myc precipitates together with AtELP3-c-myc (Fig. 4A). Like in the single mutants, interaction between the epitope-tagged subunits Elp5-HA and Elp2-c-myc (Fig. 4B) as well as Kti12-HA and Elp2-c-myc (Fig. 4C) could be restored in the double mutant by the simultaneous expression of untagged (lanes 5) or tagged (lanes 6) versions of AtELP1 and AtELP3. Elp2-c-myc (lanes 3, 5 and 6) and AtElp3-c-myc (lane 6) could be detected in the anti-HA precipitates of strains containing yeast (lane 3) or A. thaliana Elp1 and Elp3. We conclude that a chimeric yeast-plant Elongator complex can form under these conditions. Intriguingly, AtELP1 and AtELP3 together were able to restore g-toxin sensitivity in the yeast elp1 elp3 double mutant indicating that the chimeric Elongator complex was functional (Fig. 5A). Because toxin sensitivity requires the Elongator-dependent tRNA wobble uridine modification mcm 5 s 2 U , the A. thaliana subunits can reasonably be assumed to carry out a function in tRNA A. AtElp1 is able to replace yeast Elp1 and to support Elp2-Elp3 interaction and Elp3 protein stability. Protein extracts from cells expressing chromosomally c-myc-tagged Elp2 together with HA-tagged Elp3 in an elp1D background (FFY2/3-dt-1d) and transformed with plasmids containing ELP1 (pFF13), AtELP1 (pDJ98) or vector control (YEplac181) were immunoprecipitated using anti-c-myc antibodies. Western blots of the precipitates were probed with anti-c-myc or anti-HA antibodies to detect Elp2 and Elp3 respectively (arrows). Instability of Elp3 in the elp1D strain is revealed in pre-IPs (bottom panel, lane 3). B. AtElp1 is able to restore Kti12-Elp2 interaction. Protein extracts from cells expressing c-myc-tagged Elp2 together with HA-tagged Kti12 in an elp1D background (CMY300), and transformed with plasmids containing ELP1 (pFF13), AtELP1 (pDJ98) or vector control (YEplac181) were immunoprecipitated and analysed as in (A). modification that is equivalent to that of the yeast homologues.
To support this conclusion, we used a second assay monitoring Elongator-dependent tRNA modification that is based on the SUP4 suppressor tRNA gene . The SUP4 allele codes for a UAA ochre suppressor tRNA Tyr with a G34-to-U34 exchange in the anticodon. Suppression of ochre nonsense codons by SUP4 requires Elongator-dependent wobble uridine modification . In an elp1D elp3D double deletion strain (CMY160) carrying the ochre mutation ade2-1 the SUP4 tRNA suppressor was inactive resulting in an Adephenotype (Fig. 5B, vector control). Transformation with two plasmids encoding either yeast Elp1 and Elp3 or AtElp1 and AtElp3 restored SUP4 tRNA suppression, allowing for ade2-1 readthrough and growth on adenine-lacking medium. Although growth in the latter case was weak we conclude that the AtElp1-AtElp3 containing chimeric Elongator can functionally replace yeast Elongator and promote ochre suppression by SUP4. Yeast Elp1 with AtElp3 was inactive in the adenine growth assay (data not shown) as in the g-toxin assay (Fig. 5A).
Elp3 contains two protein motifs that are found in other functional contexts, the C-terminal GNAT-like HAT domain related to Gcn5 histone acteyltransferase (Wittschieben et al., 1999) and a central domain found in radical S-adenosyl-L-methionine enzymes (radical SAM domain) (Chinenov, 2002;Paraskevopoulou et al., 2006). To test which part of AtElp3 is not compatible with ScElp1 we constructed two hybrid variants with the HAT domain of yeast Elp3 replaced by that of AtELP3 and vice versa (Fig. 5C). Like AtELP3, both chimeric Elp3 variants were unable to complement the elp1 elp3 double mutant in the presence of yeast Elp1 (data not shown). However, in combination with AtELP1, AtELP3 and both hybrid proteins allowed for weak growth on adenine-deficient A. AtELP1 and AtELP3 interact with each other in yeast. Anti-c-myc immunoprecipitates of elp1D elp3D strains (CMY134) containing FLAG-AtELP1 (pJET13) and AtELP3-c-myc were probed with anti-c-myc antibodies to detect AtELP3-c-myc and with anti-FLAG-antibodies to detect FLAG-AtELP1. Immunoprecipitates with anti-c-myc antibodies from elp1D elp3D cells containing pFLAG-AtELP1 (lane 2) or no epitope-tag (lane1) served as negative controls. B. AtELP1 and AtELP3 restore Elp2-Elp5 interaction in the elp1D elp3D mutant. Protein extracts from cells expressing HA-tagged Elp5 together with c-myc-tagged Elp2 in an elp1D elp3D double mutant background (CMY305) and transformed with plasmids containing ELP1 (pFF13), ELP3 (pFF9), AtELP1 (pDJ98), AtELP3 (YEpA4), FLAG-AtELP1 (pJET13), AtELP3-c-myc (pELO3-myc) and vector controls (YEplac181, YEplac195) were immunoprecipitated using anti-HA antibodies. The antibodies used to detect the indicated proteins by Western blotting are marked on the right. Protein extracts from cells without epitope-tag and cells expressing only AtELP3-c-myc served as negative controls. Pre-IPs served as loading control (bottom panel). C. AtELP1 and AtELP3 restore Elp2-Kti12 interaction. Protein extracts from strain CMY302 expressing HA-tagged Kti12 together with c-myc-tagged Elp2 in an elp1D elp3D background and transformed (see B) were analysed as described in (B). medium indicative of SUP4-dependent ochre suppression of the ade2-1 mutation (Fig. 5B). Apparently, AtELP1 has a less stringent requirement with respect to Elp3 sequence than ScElp1. The weak Ade + phenotype in combination with AtELP1 but not with yeast ELP3 was highly reproducible; however, we could not confirm reconstitution of tRNA modification by the hybrid Elp3 variants using the g-toxin assay (data not shown). Probably partial complementation results in hypomodification of tRNAs and gives zymocin resistance tRNA modification is affected in Elongator-deficient plant cells Restoration of ochre suppression and g-toxin sensitivity of the yeast elp1D elp3D double mutant by AtELP1 and AtELP3 suggested that plant Elongator may also affect tRNA modification in plant cells. For all but one tRNA modification genes of yeast (TRM genes) homologues can be found in plant genomes suggesting that plant tRNAs are modified in a similar way. To test whether Elongator deficiency in A. thaliana had any impact on tRNA modification RNA preparations enriched for small stable RNA were compared between wild-type and homozygous elo3 mutants. RNA was isolated from leaves of several independent plants, degraded to nucleosides and analysed by HPLC. The elution profiles were very similar for both samples except for the parts shown in Fig. 6, where 5-carbamoyl-methyluridine (ncm 5 U) ( Fig. 6A and B) and mcm 5 s 2 U ( Fig. 6C and D) eluted. Both modified uridine nucleosides (compare Fig. 6E) were lacking in the sample from the elo3 mutant. These data clearly show that the elo3 mutation prevented the ncm 5 and mcm 5 s 2uridine modifications from being generated supporting the view that plant Elongator is involved in tRNA modification in a way similar to its yeast homologue. This indicated that the role of Elongator in tRNA modification might be conserved among lower and higher eukaryotes.

Discussion
Biochemical characterizations of yeast and human Elongator and genomic sequences have indicated that the subunit composition of this complex is conserved among eukaryotes (reviewed in (Svejstrup, 2007). The data presented in this paper reinforce this view and provide evi- A. Restoration of toxin sensitivity. The elp1D elp3D (CMY134) strain was transformed with plasmids containing ELP1 (pFF13), ELP3 (pFF9), AtELP1 (pCM26), AtELP3 (YEpA4) and vector controls (YEplac181 and YEplac195) in the indicated combinations. The g-toxin was expressed from pHMS14 (Frohloff et al., 2001) on galactose medium. Growth on galactose after 4 days at 30°C indicates g-toxin resistance (Tox R ) and no growth equals g-toxin sensitivity (Tox S ). B. Restoration of SUP4 suppressor tRNA function. SUP4 elp1D elp3D (CMY160) cells were transformed with plasmids containing wild-type ELP1 and ELP3 (pFF13 and p424TDH-ScElp3myc), the respective empty vectors (YEplac181 and p424TDH), the plant homologues AtELP1 and AtELP3 (pCM26 and p424TDH-AtElp3myc) or AtELP1 and the swapped HAT domain variants pElp3-AtHAT and pAtElp3-ScHAT. To check for suppression of the ade2-1 (UAA) ochre mutation  serial dilutions of the transformants were spotted on SC and SC-Ade plates and cultivated for 4 days at 30°C. SUP4 suppressor tRNA function results in growth on SC-Ade plates (Ade + phenotype). C. Schematic representation of the Elp3 domain structure and domain swapped hybrids generated between AtELP3 and ScElp3 used as in (B). dence for a remarkable structural conservation between plant and yeast Elongator complexes. We further demonstrate that an A. thaliana Elongator mutant is lacking the very same tRNA modifications previously shown in yeast to be Elongator-dependent . Together with similar deficiencies reported for C. elegans Elongator mutants (Chen et al., 2009), these data strongly support the view that tRNA modification might be a function of Elongator that is conserved in animals, fungi and plants.
A crucial question is how Elongator affects tRNA modification. The acetyltransferase activity of the Elp3 subunit has been shown to be required . One possibility is that biosynthesis or activity of the enzymes generating the ncm 5 or mcm 5 side chains on the tRNA wobble uridines depend on acetylation. Alternatively, Elongator may directly participate in the tRNA modification reaction. The Elongator subunits Elp1 and Elp3 could be specifically cross-linked to tRNA Glu but not tRNA Met  supporting the latter model. A radical mechanism based on the putative radical SAM domain flanking the HAT motif in Elp3 (Chinenov, 2002;Paraskevopoulou et al., 2006) might be involved.  .
We have used yeast as a tool to analyse whether AtELP3 was able to function in the context of a heterologous Elongator complex. Two sensitive assays that allow to monitor the Elongator-dependent tRNA modification were used. The first one detects mcm 5 s 2 U-tRNA Glu and, to a lesser extent, mcm 5 s 2 U-tRNA Lys and mcm 5 s 2 U-tRNA Gln . These tRNAs are the primary targets of the tRNA endonuclease activity associated with the K. lactis zymocin. Defects in formation of the modified uridine and typical of Elongator mutants cause zymocin resistance Jablonowski et al., 2006;Studte et al., 2008). Restoration of zymocin sensitivity in resistant mutants indicates efficient formation of the modified tRNAs. The second assay makes use of the SUP4 ochre suppressor tRNA, which also depends on the mcm 5 modification function of Elongator . Both assays are sensitive to a point mutation in the HAT domain of yeast Elp3 (Frohloff et al., 2001;Huang et al., 2005). Under conditions of partial complementation the two assays are expected to differ in sensitivity. Hypomodified tRNAs would confer (some degree of) resistance in the zymocin assay (i.e. the mutant phenotype) by escaping endonucleolytic cleavage but might support some growth on selective plates (the wild-type phenotype) in the nonsense suppression assay.
We have shown here that a double mutant lacking Elp3 and Elp1 can be complemented by the Arabidopsis homologues AtELP1 and AtELP3/ELO3, which together restore zymocin sensitivity and SUP4-mediated nonsense suppression. Thus, these two subunits cooperate to functionally replace yeast Elp1 and Elp3 indicating that they have a similar role in the plant Elongator complex.
Surprisingly, despite 67% sequence identity between ScElp3 and AtELP3, the latter was unable to complement the yeast elp3 mutation. Only in combination with AtELP1 could complementation of the Elongator mutant phenotypes by the AtELP3 genes be achieved. AtELP3-c-myc alone was incorporated into the yeast complex as shown by its ability to restore association between the large and small subcomplexes and between Elongator and its partner protein Kti12. Both types of interactions require Elp3 and the structural integrity of the complex (Frohloff et al., 2001;Fichtner et al., 2002a;Fichtner et al., 2003;Petrakis et al., 2005). We conclude that, although an Elongator complex in which AtELP3 replaces Elp3 appears structurally intact, the AtELP3 containing chimeric complex is unable to provide the activity required for tRNA modification.
The lack of functional complementation by AtELP3 is probably not due to the reduced efficiency of Elp2-Elp5 or Elp2-Kti12 interactions as the same reduction was observed in the AtELP3 and AtELP1 containing chimeric complex, which is functional. AtELP3 appears stable in yeast in the presence of ScElp1 excluding the possibility that lack of function is due to protein degradation, which is observed for ScElp3 in the absence of ScElp1 (Fig. 3).
To determine whether sequence divergence in the HAT domain was responsible for the lack of complementation, domain swapped variants with ScHAT replaced by AtHAT and vice versa were tested. Like AtELP3, both variants were inactive in the presence of ScELP1. This failure of the AtHAT to replace the corresponding yeast domain contrasts with the ability of the HsELP3-HAT to do so (Li et al., 2005). However surprisingly, in the SUP4 tRNA suppression assay (but not in the zymocin assay), both hybrid ELP3 genes were able to partially complement the yeast elp3 mutation in combination with AtELP1. This indicates weak U34-tRNA modification mediated by AtELP1 in combination with multiple structural variants of Elp3 including ScElp3 (Chen et al., 2006). AtELP1 appears more tolerant towards sequence variations in the Elp3 protein. The crucial role of ScElp1 phosphorylation/ dephosphorylation (Mehlgarten et al., 2009) might be important in this context. Alternatively but less likely, AtELP1 may contribute to complementation in yeast via an additional function not found in ScElp1, like the RNAdependent RNA polymerase activity recently reported for Drosophila Elp1 (Lipardi and Paterson, 2009).
A current challenge is to understand the biological function(s) of the Elongator complexes in the various organisms and their possible divergence during evolution. If, as proposed here and supported by studies in C. elegans (Chen et al., 2009), tRNA modification defects are a general consequence of Elongator deficiency in eukaryotes it will be essential to determine whether the mutant phenotypes are causally related to this defect. Studies in yeast support such a view. Pleiotropic phenotypes of yeast Elongator mutants including hypoacetylation of histones, delayed adaptation of transcription to changing environmental conditions as well as delocalization of the secretory protein Sec2 could all be suppressed by overexpression of two tRNAs, tRNA Lys and tRNA Gln , both of which in wild-type cells undergo Elongator-dependent anticodon modifications (Esberg et al., 2006). Based on these data it has been proposed that the effects of Elongator mutants on transcription and secretion are indirect consequences of influences on translation caused by inappropriate tRNA modification. Apparently, U34 modifications in the anticodon increase the decoding efficiency of A-and G-ending codons (Yokoyama et al., 1985;Lim, 1994;Durant et al., 2005;Esberg et al., 2006;Begley et al., 2007;Johansson et al., 2008). Thus, the translation of individual proteins might be differentially affected depending on their amino acid composition and codon usage. Moreover, yeast elp mutants are synthetic lethal with mutations affecting thiolation of U34 (Esberg et al., 2006;Johansson et al., 2008) emphasizing the importance of the wobble uridine modifications.

Plant Elongator function in tRNA modification 1089
A causal relationship between tRNA hypomodification and Elongator phenotypes does not exclude that Elongator has additional activities or multiple targets for the acetyltransferase reaction. In plants, yeast and human cells the complex is also found in the nucleus and its depletion affects transcription (Winkler et al., 2002;Close et al., 2006;Svejstrup, 2007;Nelissen et al., 2010). Transcriptome analysis in those organisms revealed that only a small set of genes is affected by Elongator depletion with little apparent overlap of functional categories (Krogan and Greenblatt, 2001;Close et al., 2006;Nelissen et al., 2010). This argues against a general role of Elongator in transcription elongation.
In A. thaliana several auxin-related genes were found to be differentially expressed in wild-type and elo mutants, and at least two of these genes, SHY2/IAA3 and LAX2, were also affected in histone H3K14 acetylation, which is a predominant target of the ScELP3-HAT (Nelissen et al., 2010). The data were taken as evidence for a role of Elongator in transcription elongation of specific genes, which may explain the observed influences on cell proliferation and development in elo mutants. How specificity for these genes might be achieved is currently unknown. Because auxin signalling is specific to plants whereas neuron development is specific to animals studies on the role of Elongator in these diverse processes might reveal how a protein complex with (a) conserved molecular function(s) has evolved to fulfil these kingdom specific functions. We believe that the answer to these questions can be found at the cellular level.
Taken together, our data indicate that the Elongator complexes are structurally and functionally highly con-served between yeast and plants, such that yeast subunits can be replaced by their plant orthologues. This offers the opportunity to take advantage of the yeast model system in structure-function studies of Elongator complexes from other kingdoms. It is likely that eukaryotic Elongator complexes have a conserved biochemical activity. Based on reconstitution of U34 tRNA modification in yeast by plant Elongator subunits and the requirement of AtElp3/ELO3 for this modification in plants, we favour the view that this conserved activity is directly related to tRNA modification.

Yeast strains, media and general methods
Yeast strains used are listed in Table 1. Yeast was grown in rich media containing yeast extract, peptone and 2% dextrose (YPD) or 2% galactose (YPG) or on synthetic complete medium (SC) (Sherman, 1991).
Thermosensitivity was assayed on YPD medium at 30 or 38°C for 2-3 days. Drug sensitivity was assayed at 30°C with 5 mM caffeine. ade2-1 ochre stop codon suppression by the SUP4 gene was tested as described . Yeast was transformed with plasmid DNA or polymerase chain reaction (PCR) products according to a previous protocol (Schiestl and Gietz, 1989). Primers are listed in Table S1.

Arabidopsis thaliana strains and growth conditions
An A. thaliana homozygous T-DNA insertion mutant in At5g50230 (AtELP1) derived from GABI-Kat Line 636H02

Plasmid constructions
pFF9, a YEplac195-based plasmid carrying ELP3 has been described (Frohloff et al., 2001). A. thaliana AtELP3/ ELO3 was amplified from cDNA clone RAFL0811J12 (http://www.brc.riken.jp) with primers cDNAfw and cDNA2re inserting flanking SalI restriction sites. The SalI fragment was fused to the yeast ADH1 promoter and integrated into YEplac195 to give YEpA4. The AtELP3 open reading frame was verified by DNA sequencing. In yatELP3 the ADH1 promoter and 5′ end of the AtELP3 gene from YEpA4 was replaced after cleavage with HindIII and AgeI by the yeast ELP3 promoter and 5′ end using primers E3′-fw and E3′-re and pFF9 as template. yatELP3M is a triple c-myc-tagged variant of yatELP3. pFF13 is a YEplac181-based plasmid, carrying the yeast ELP1 gene (Fichtner et al., 2002a). pDJ98 is a multicopy plasmid for galactose-regulated expression of the AtELP1 gene carrying the selection markers URA3 and leu2d (Chen et al., 2006). pJET13, carrying FLAG-AtELP1 under control of the GAL1 promoter was obtained by PCR amplification of the coding sequence from pFLAG-ELO2 (Z. Gong, unpublished) and cloned into SalI-cleaved pCM22. In the latter, the GAL1 promoter was introduced on a EcoRI-BamHI fragment from plasmid pRB1438 (a kind gift from Mike Stark, University of Dundee, UK) into YEplac181 (LEU2). pCM22 served also as destination vector to clone untagged AtELP1 from pDJ98 via SalI restriction sites to obtain pCM26 for the SUP4 suppression assay. The centromeric plasmid p424TDH (Mumberg et al., 1995) served as the destination vector for ELP3 variants in double complementation studies. p424TDH-ScElp3myc encodes a C-terminal triple c-myc-tagged variant of ScELP3 (kindly provided by O. Onuma, Univ. Halle), p424TDH-AtElp3myc a C-terminal triple c-myctagged variant of AtELP3. PCR-mediated domain swap of the AtELP3 HAT domain (amino acids 436-565) into p424TDH-ScElp3myc (1-426) and of ScElp3-HAT (382-557) into p424TDH-AtElp3myc (1-390) resulted in plasmids pELP3-AtHAT and pAtELP3-ScHAT respectively.

Immunological techniques
Detection of tagged proteins used anti-c-myc (A-14) and anti-HA (F-7) (Santa Cruz) antibodies. Anti-Elp1 antibodies Wittschieben et al., 1999) were kindly provided by Dr J. Svejstrup (London Research Institute, Cancer Research, UK). An anti-FLAG antibody (Sigma-Aldrich) was used for detection of FLAG-ELO2. Protein concentrations were determined by the method of Bradford (Bradford, 1976). Antibody cross-linking to protein A-Sepharose, preparation of protein extract and co-immunoprecipitations were performed as described previously (Zachariae et al., 1996;Frohloff et al., 2001).

tRNA isolation and HPLC analysis
tRNA was prepared from total RNA preparations as described (Bjork et al., 2001). Purified tRNA was digested with Nuclease P1 for 16 h at 37°C and then treated with bacterial alkaline phosphatase for 2 h at 37°C. The hydrolysate was analysed by high-pressure liquid chromatography with a Develosil C-30 reverse-phase column as described (Gehrke et al., 1982;Gehrke and Kuo, 1990).