Identification and Characterization of Elf1, a Conserved Transcription Elongation Factor in Saccharomyces cerevisiae

Yeast strains, plasmids, and media.All S. cerevisiae strains (Table 1) are isogenic to a GAL2⁺ derivative of S288C (78). To epitope tag Elf1, DNA sequence encoding 13 copies of the Myc epitope was integrated at the 3′ end of the ELF1 coding region (37). The ELF1-13MYC gene is wild type with respect to an Spt phenotype, and it does not cause any synthetic defects in growth when crossed to any other mutation tested in this study. The Flag-tagged Spt4 construct was provided by G. Hartzog and has been previously described (36). The Ctr9-Myc construct has been previously described (25). The dst1Δ::KANMX allele was constructed by a one-step PCR-mediated disruption which replaces the entire open reading frame with the KANMX4 cassette (8). The elf1Δ::NATMX allele, referred to as elf1Δ throughout this paper, was constructed by a one-step PCR-mediated disruption which replaces the entire open reading frame with the NATMX4 cassette (20). The GAL1-FLO8-HIS3 reporter, to be described elsewhere, contains the HIS3 gene inserted out-of-frame into the 3′ coding region of FLO8 such that wild-type HIS3 product is only expressed when transcription initiates from a cryptic promoter within the FLO8 gene. In addition, the FLO8 promoter has been replaced by the GAL1 promoter. All deletion mutants used from the yeast deletion set (InVitrogen) were confirmed by PCR. Plasmid pKC13 (gift of Kristine Johnson and Caroline Kane) contains the DST1 coding region under GAL1 control in the YEp435 vector (URA3 ⁺, 2μm plasmid). Western analysis with TFIIS antisera (a gift of C. Kane) revealed a significant level of TFIIS protein expressed from this plasmid under repressive growth conditions (glucose) and a greatly elevated level under inducing conditions (galactose). Plasmid pDP1 was constructed by addition of BamHI and EcoRI restriction sights to a PCR fragment containing ELF1 (−487 to +840 bp, where +1 = ATG and +436 = TAA) and ligated into pRS416 (45). Double mutant yeast strains were constructed by standard genetic methods. An elf1Δ mutant containing plasmid pDP1 was crossed to strains containing mutations in genes encoding a number of elongation factors. After sporulation and tetrad dissection, double mutant strains were identified and tested for growth on medium containing 5-fluoroorotic acid (5-FOA). In all cases, at least 10 tetrads were dissected and analyzed for growth. In most cases, strains used were in the FY (S288C) strain background and are listed in Table 1. However, strains containing leo1Δ, rpb4Δ, hir2Δ, hir3Δ, and hpc2Δ were obtained from the yeast deletion set, also in the S288C background (InVitrogen).

Mating, transformation, sporulation, and tetrad analysis were performed by standard procedures previously described (58). Rich (yeast extract-peptone-dextrose [YPD]), minimal (SD), synthetic complete (SC), omission (SC-), 5-FOA, and sporulation media were prepared as previously described (58). For SGA analysis, the sporulation medium was 2% agar, 1% potassium acetate, 0.1% yeast extract, 0.05% glucose supplemented with uracil, histidine, and lysine. Selection for cells containing the KANMX4 cassette was conducted on YPD containing 200 μg/ml of G418 sulfate (Gibco), and selection for the deletion was on YPD containing 100 μg/ml clonNAT (Werner BioAgents).

Identification of synthetic lethal mutations.To identify novel elongation factors, we performed a genetic screen to identify mutations causing synthetic lethality with mutations in dst1Δ and spt6-14. First, strain FY2371 (dst1Δ spt6-14) was transformed with plasmid pKC13, a high-copy-number plasmid containing the wild-type DST1 ⁺ and the URA3 ⁺ genes. This strain was then transformed with NotI-generated linear DNA fragments from an mTn3 transposon insertion library (59), and transformants were selected on medium lacking leucine. Approximately 20,000 transformants were replica plated to medium lacking leucine and medium containing 5-FOA. Initially, 28 synthetic lethal candidates were identified due to the lack of growth on 5-FOA-containing medium. These colonies were purified and retested for growth on 5-FOA. Many of the insertions were eliminated as candidates due to the instability of the transposon insertion, weakness of phenotype upon rechecking, or lack of linkage between the phenotype and the transposon insertion. Of these candidates, we found one strong synthetic lethal candidate where the phenotype was linked to the transposon insertion. We identified the genomic location of this insertion by vectorette PCR (34, 56). This insertion was located in the YKL160W/ELF1 gene and predicts a fusion of the first 62 amino acids with the lacZ gene from the transposon insertion.

SGA analysis.SGA analysis was performed as previously described (74) with minor modifications. The S. cerevisiae deletion set was transferred from frozen stocks in 96-well containers to large YPD plates and grown as spots for 2 days at 30°C. The strains were then mated to lawns of the query strain L1095 on fresh YPD plates, incubated for 1 day, and then replica plated to YPD plates containing G418 and clonNAT to select for diploids. The diploids were replica plated to sporulation medium and incubated at 22°C for 5 days. MAT a spores were selected by replica plating to SC-his-arg plates containing canavanine (50 mg/liter) for 2 days. Finally, strains were replica plated to haploid selection medium containing G418 and haploid selection medium containing G418 and clonNAT. Growth was scored for 3 days, and synthetic lethal candidates were identified by lack of growth on the medium containing G418 and clonNAT. From this analysis, 10 synthetic lethal candidates were identified. For nine of the candidates, we tested the possible synthetic lethality with elf1Δ by tetrad analysis. The tenth candidate (yml009w-BΔ) was not tested, because this deletion removes a small portion of the essential gene, SPT5, and some alleles of SPT5 were already known to be synthetically lethal with elf1Δ. Of the nine remaining candidates, only three were determined to be synthetically lethal with elf1Δ by tetrad analysis (cdc73Δ, leo1Δ, and rpb4Δ). Five candidates were not synthetically lethal (rtf1Δ, hir1Δ, hir2Δ, hir3Δ, and hpc2Δ), and one (bur2Δ) germinated too poorly to allow any conclusion to be made. As previously pointed out, SGA analysis is subject to the identification of both false-negative and false-positive interactions (38, 74, 75). Using SGA analysis, we identified an interaction with mutations in three of the five known components of the Paf1 complex (cdc73Δ, leo1Δ, and rtf1Δ). Lethality of elf1Δ with deletions of genes encoding the other two components of the Paf1 complex, paf1Δ and ctr9Δ, was not able to be determined by SGA due to the sickness of these strains in the deletion set. In addition, another mutation known to be synthetically lethal with elf1Δ, spt4Δ, was not identified in the SGA analysis, likely because the spt4Δ strain is not healthy in the deletion set and thus could not be scored by SGA. In addition to false negatives, we also identified five false positives (see Table 3, below) for unknown reasons. There are several differences in the SGA method when compared to standard tetrad analysis that could account for these differences, including the use of synthetic complete media for germination or the requirement for sporulation and germination among patches of other wild-type and single mutant yeast cells.

Chromatin immunoprecipitation.All chromatin immunoprecipitations were carried out as previously described with minor modifications (35). Antibodies used in these experiments included the 12CA5 mouse monoclonal antibody against the hemagglutinin (HA) epitope (0.5 μl), A14 anti-Myc rabbit polyclonal (5 μl; Santa Cruz), and M2 anti-FLAG monoclonal antibody (2 μl; Sigma). Immunoprecipitations were carried out in FA lysis buffer containing 150 mM NaCl and no sodium dodecyl sulfate (SDS). Dilutions of input DNA and immunoprecipitated DNA were subjected to quantitative PCR by the incorporation of [α-³²P]dATP. The products were separated on a 6% nondenaturing polyacrylamide gel, and quantification was carried out by phosphorimager analysis (Molecular Dynamics). The percent immunoprecipitated was calculated for each sample, and all samples were normalized to a control PCR product amplified in each reaction mixture. The primers for this control PCR product amplify a region of chromosome V that is not contained in any open reading frames (28). Relative positions of primer sites were as follows for GAL1: UAS, −536 to −276; TATA, −190 to +54; 5′, +590 to +877; 3′, +1330 to +1657; and 3′UTR, +1805 to +2079; For PMA1 they were as follows: UAS, −623 to −390; promoter, −370 to −90; 5′, +584 to +807; 3′, +2018 to +2290; 3′UTR, +529 to +742 (relative to stop codon). All values are relative to the translation start site (ATG = +1) except where noted. All experiments were performed at least three times, and standard errors are shown on each graph. By Western analysis, each protein examined by ChIP was present at the same level in the wild-type and mutant strains.

Northern analysis.RNA isolation and Northern hybridization experiments were carried out as described previously (4, 72). Northern hybridization analysis was conducted using probes to FLO8 and SCR1. The FLO8 probe corresponds to the +1515 to +2326 region of the FLO8 gene. The SCR1 probe corresponds to the −221 to +452 region of the SCR1 gene. For each experiment, Northern analysis was performed in duplicate with similar results between the two experiments. Viability of all strains used for Northern analysis was approximately 100% for all times and temperatures used in the experiments.

Purification of Elf1 and mass spectrometry of associated proteins.Tandem affinity purification (TAP)-tagged Elf1 was purified on immunoglobulin G (IgG) and calmodulin columns from extracts of yeast cells (10 liters) grown in YPD medium to an optical density at 600 nm of 1.0 to 1.5 (19, 55). The cell pellets (7 to 10 g) were frozen in liquid nitrogen and lysed by grinding with dry ice in a Krups coffee grinder (model number 203-70). An equal volume of YEB (250 mM KCl, 100 mM HEPES-KOH, pH 7.9, 1 mM EDTA, 2.5 mM dithiothreitol [DTT]) was added and, following centrifugation in a Beckman Ti70 rotor at 4°C for 2 h at 34,000 rpm, the supernatant was dialyzed against IPP buffer (10 mM Tris-Cl, pH 7.9, 0.1% Triton X-100, 0.5 mM DTT, 0.2 mM EDTA, 20% glycerol, 100 mM NaCl). After dialysis, the extract was again centrifuged in a Ti70 rotor at 4°C for 30 min at 34,000 rpm and the supernatant was mixed for 3 h with 200 μl IgG-Sepharose (Pharmacia) equilibrated with IPP buffer. Following binding, the IgG-Sepharose was washed with 1 ml IPP buffer followed by 400 μl TEV protease cleavage buffer (50 mM Tris-Cl, pH 7.9, 1 mM DTT, 0.1% Triton X-100, 100 mM NaCl). The beads were then incubated overnight at 4°C with 100 U of TEV protease (Life Technologies) in 200 μl TEV cleavage buffer. The eluate was combined with a 200-μl wash with TEV cleavage buffer. To this was added 200 μl calmodulin binding buffer (10 mM Tris-Cl, pH 7.9, 10 mM β-mercaptoethanol, 2 mM CaCl₂, 0.1% Triton X-100, 100 mM NaCl) and 200 μl of calmodulin beads (Pharmacia) equilibrated with the same buffer. After binding for 1 to 2 h at 4°C, the calmodulin beads were washed with 200 μl calmodulin binding buffer and 200 μl calmodulin wash buffer (10 mM Tris-Cl, pH 7.9, 10 mM β-mercaptoethanol, 0.1 mM CaCl₂, 0.1% Triton X-100, 100 mM NaCl). The purified protein complexes were eluted from the calmodulin beads with 5× 100 μl calmodulin elution buffer (10 mM Tris-Cl, pH 7.9, 10 mM β-mercaptoethanol, 3 mM EGTA, pH 8.0, 0.1% Triton X-100, 100 mM NaCl). The purified proteins were either directly reduced, alkylated, digested with trypsin, and analyzed by liquid chromatography tandem mass spectrometry (LC-MS/MS) on a Finnigan LCQ Deca tandem mass spectrometer or else separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on gels containing 10% polyacrylamide and stained with silver. After SDS-PAGE, protein bands were reduced, alkylated, and digested with trypsin, and purified peptide samples were spotted onto a target plate with a matrix of α-cyano-4-hydroxycinnamic acid (Fluka). Matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry analysis was conducted utilizing a Reflex IV (Bruker Daltonics, Billerica, MA) instrument in positive ion reflectron mode (39, 63). For tandem mass spectrometry, the trichloroacetic acid-precipitated protein was resuspended in 100 mM NH₄HCO₃-1 mM CaCl₂ buffer, pH 8.5, and digested overnight at 37°C with 2 μl of immobilized Poros trypsin beads (PerSeptive). The entire digest was fractionated as described previously (17) on a 7.5-cm (inner diameter, 100 μm) reverse-phase C₁₈ capillary column attached in-line to a Finnigan LCQ-Deca ion trap mass spectrometer by ramping a linear gradient from 2 to 60% solvent B in 90 min. Solvent A consisted of 5% acetonitrile, 0.5% acetic acid, and 0.02% heptafluorobutanoic acid, and solvent B consisted of acetonitrile-water (80:20) containing 0.5% acetic acid and 0.02% heptafluorobutanoic acid. The flow rate at the tip of the needle was set to 300 nl/min by programming the high-performance liquid chromatography pump and using a split line. The mass spectrometer cycled through four scans as the gradient progressed. The first was a full mass scan followed by successive tandem mass scans of the three most intense ions. A dynamic exclusion list was used to limit collection of tandem mass spectra for peptides that eluted over a long period of time. All tandem mass spectra were searched by using the SEQUEST computer algorithm against the complete yeast protein database. Each high-scoring peptide sequence was manually compared with the corresponding tandem mass spectrum to ensure that the match was correct.

Identification of a mutation that causes lethality in combination with mutations in DST1 and SPT6.To identify previously unstudied genes involved in transcription elongation in S. cerevisiae, we performed a genetic screen for mutations that cause inviability when combined with mutations in genes known to encode elongation factors. Specifically, we screened for mutations that cause inviability when combined with both dst1Δ and spt6-14 (described in Materials and Methods). From this screen, we identified one mutation in YKL160W, a previously uncharacterized gene. This gene is hereafter referred to as ELF1, and the mutation is referred to as elf1-1::Tn3. To determine the phenotype of elf1-1::Tn3 when combined with either dst1Δ or spt6-14 individually, we constructed elf1-1::Tn3 dst1Δ and elf1-1::Tn3 spt6-14 double mutants. The elf1-1::Tn3 dst1Δ double mutant had very little defect in growth. In contrast, the elf1-1::Tn3 spt6-14 mutant was extremely sick. Thus, the elf1-1::Tn3 and spt6-14 mutations contribute the majority of the growth defect initially observed in our screen. Later in this paper, we show that elf1Δ spt6-14 double mutants are inviable and that elf1Δ dst1Δ double mutants are viable but have mutant phenotypes.

Elf1 is a member of a conserved eukaryotic family of zinc finger proteins.The ELF1 gene of S. cerevisiae is predicted to encode a small protein of 145 amino acids. BLAST analysis (2) reveals that the predicted Elf1 protein is conserved in other sequenced eukaryotic genomes, including other yeasts, human, Drosophila melanogaster, Caenorhabditis elegans, and Arabidopsis thaliana (Fig. 1). The conservation of Elf1 occurs in the region corresponding to the N-terminal half of the S. cerevisiae protein. These hypothetical proteins are short (often less than 100 amino acids) and contain a C4 zinc-binding region. The four cysteine residues of the potential zinc-binding domain are completely conserved in all homologues, and this conserved domain has been termed DUF701 (domain of unknown function 701) (40). To date, only the C. elegans homologue has been functionally characterized; in RNA inhibition experiments, no significant growth defect was observed (24). Two genome-wide localization screens (22, 33) have suggested that Elf1 is localized to the nucleus, consistent with our localization analysis of an Elf1-green fluorescent protein fusion (data not shown). Interestingly, Elf1 was also identified in one of those studies to be associated with meiotic chromosomes (33). The conservation of Elf1 strongly suggests that it plays an important role throughout eukaryotes.

• Open in new tab
• Download powerpoint

FIG. 1.

Elf1 is a conserved member of a zinc finger family of uncharacterized proteins. Homologues of Elf1 were identified by standard protein BLAST analysis. Full-length homologous protein sequences from several eukaryotic organisms were then aligned using ClustalW. Identical residues are noted by an asterisk, conserved substitutions are noted by a colon, and semiconservative substitutions are noted by a period. The four conserved cysteines of the zinc finger domain are noted by arrows above the alignment. Abbreviations: YEAST, Saccharomyces cerevisiae; SCHPO, Schizosaccharomyces pombe; HUMAN, Homo sapiens; MUS, Mus musculus; CAEEL, Caenorhabditis elegans; DROME, Drosophila melanogaster; ARATH, Arabidopsis thaliana.

elf1Δ causes mild sensitivity to 6-AU and genetically interacts with dst1Δ.To further characterize ELF1, we constructed a complete deletion of this gene and tested it for mutant phenotypes. The elf1Δ mutation caused no significant defect in growth (Fig. 2). We also tested an elf1Δ mutant for sensitivity to 6-azauracil (6AU) and mycophenolic acid (MPA), phenotypes shared by many transcription elongation mutants due to the effect of these compounds on depleting nucleotide pools in vivo (12). Our results show that elf1Δ mutants exhibit a mild sensitivity to 6AU, but not to MPA (Fig. 2). We also tested elf1Δ dst1Δ double mutants and found that they are extremely sensitive to both 6AU and MPA (Fig. 2). In addition, elf1Δ dst1Δ double mutants fail to grow on galactose medium at 37°C (data not shown). Finally, we also tested whether elf1Δ caused an Spt⁻ (suppressor of Ty insertion) phenotype. This was tested using the insertion mutation lys2-128δ (67), and it was shown that elf1Δ causes a modest suppression, allowing growth on SC-lys plates after 3 days of incubation at 30°C (data not shown). Together, the 6AU phenotype, the Spt⁻ phenotype, and genetic interactions with dst1Δ suggest a potential role for Elf1 in transcription elongation.

• Open in new tab
• Download powerpoint

FIG. 2.

The elf1Δ mutation causes mild sensitivity to 6AU and genetically interacts with dst1Δ. Yeast strains were spotted onto plates in a series of 10-fold dilutions. Strains were grown on SC-ura or SC medium for 2 days, medium containing 6AU (50 μg/ml) for 5 days, or medium containing MPA (20 μg/ml) for 3 days.

elf1Δ genetically interacts with mutations in the genes encoding the Spt4, Spt5, Spt6, and Spt16 elongation factors.The screen by which we identified ELF1 strongly suggested that it is critical for growth in an spt6 mutant background. Therefore, we tested if elf1Δ also genetically interacts with mutations that impair several known elongation factors. These tests were performed by two approaches. In the first approach, described in this section and the following one, we constructed specific double mutants, combining elf1Δ with mutations in genes encoding other known elongation factors, some of which are essential for viability. The second approach, described later, used SGA analysis (74, 75) to examine the genetic interactions between elf1Δ and deletions in all other nonessential yeast genes.

By crosses, we first combined elf1Δ with spt4, spt5, and spt6 mutations. Genetic and biochemical results have implicated the Spt4, Spt5, and Spt6 proteins functioning through effects on transcription elongation and chromatin structure (82). By crosses, we tested elf1Δ combined with a complete deletion of the nonessential SPT4 gene and with several different missense mutations in the essential SPT5 and SPT6 genes. Our results demonstrate that an elf1Δ spt4Δ double mutant is inviable (Fig. 3A; Table 2). Consistent with these results, an elf1 mutation was independently identified in a screen for mutations that cause lethality when combined with spt4Δ (G. Hartzog and J. Speer, personal communication). We tested elf1Δ with three different spt5 mutations and observed inviability for two of the combinations (Table 2). The third combination is described below. We also tested four different spt6 alleles: three combinations resulted in synthetic lethality, and the fourth double mutant grew extremely poorly (Fig. 3A; Table 2). From these results, we conclude that, in general, elf1Δ causes lethality when combined with spt4, spt5, and spt6 mutations.

• Open in new tab
• Download powerpoint

FIG. 3.

elf1Δ genetically interacts with mutations in the genes encoding the transcription elongation factors Spt4, Spt5, Spt6, and the Paf1 complex. (A) Deletion of elf1Δ is lethal in combination with mutations in SPT4, SPT6, or Paf1 complex genes. elf1Δ strains were transformed with plasmid pDP1 containing ELF1 ⁺ on a CEN URA3 plasmid and mated to spt4Δ, spt6-1004, cdc73Δ, or paf1Δ strains. Diploids were sporulated and dissected, and representative spores from the resulting complete tetrads were patched onto YPD. Patches were replica plated to medium lacking uracil for 1 day and medium containing 5FOA for 2 days. (B) Deletion of elf1Δ suppresses spt5-242 phenotypes. spt5-242 is a Cs⁻ Spt⁻ mutation that can be suppressed by mutations in Pol II or other elongation factors that impair elongation (21). elf1Δ strongly suppresses the Cs⁻ phenotype and modestly suppresses the Spt⁻ phenotype of spt5-242. The Spt⁻ phenotype was scored by observing growth on medium lacking lysine and using the lys2-128δ allele. Yeast strains were spotted in a series of 10-fold serial dilutions onto YPD for 2 days at 30°C, YPD at 14°C for 9 days, SC for 2 days, or SC-lys for 4 days.

One interesting exception to the results described above came from the combination of elf1Δ with the spt5-242 mutation. Previous analysis demonstrated that the spt5-242 mutation is distinct from other spt5 mutant alleles with respect to its mutant phenotypes, including genetic interactions with other transcription mutants (21). In addition, spt5-242 confers a cold-sensitive (Cs⁻) growth phenotype (21). When combined with spt5-242, elf1Δ suppressed the Cs⁻ phenotype (Fig. 3B), similar to results previously obtained for mutations in the genes encoding the Rpb1 and Rpb2 subunits of Pol II (21), the Paf1 complex (70), and Chd1, a chromodomain-containing protein that has also been implicated in elongation (68). Additionally, elf1Δ also modestly suppressed the Spt⁻ phenotype of the spt5-242 allele (Fig. 3B).

We next tested to see whether an elf1Δ mutation interacted with an spt16 mutation. We observed no significant growth defect when elf1Δ was combined with the spt16-197 mutation, although there was a slight synthetic Spt⁻ phenotype (Table 2 and data not shown). Thus, mutations in elf1Δ interact strongly with mutations in SPT4, SPT5, and SPT6 but only cause a mild phenotype when combined with a mutation in SPT16. However, it is possible that other alleles of spt16 might cause more severe defects than the spt16-197 mutation.

elf1Δ genetically interacts with mutations in the genes encoding the Paf1 complex.We next tested for genetic interactions between elf1Δ and null mutations in the genes encoding components of the Paf1 complex. We found that an elf1Δ mutation causes synthetic lethality when combined with deletion mutations in four of the five Paf1 complex genes, paf1Δ, cdc73Δ, ctr9Δ, and leo1Δ. In agreement with this result, elf1Δ was identified in a large-scale SGA screen to genetically interact with a mutation in the gene encoding the Cdc73 component of the Paf1 complex (75). In contrast, when elf1Δ was combined with rtf1Δ, the double mutant was viable, although it displayed both Spt⁻ and Ts⁻ phenotypes (Table 2). Together, the strong genetic interactions of elf1Δ with mutations in the genes encoding the elongation factors Spt4, Spt5, Spt6, and the Paf1 complex further support an important role for Elf1 in transcription elongation.

SGA analysis identifies additional mutations that genetically interact with elf1Δ.To examine the genetic interactions of elf1Δ with deletions of all of the nonessential genes of S. cerevisiae, we performed SGA analysis (74). In this approach, elf1Δ was combined with the approximately 4,700 deletions contained within the yeast deletion set. This analysis led to the identification of 10 candidate mutations that were initially scored as causing synthetic lethality with elf1Δ (Table 3). Of these 10 candidates, 9 merited further investigation (see Materials and Methods) and were tested for synthetic lethality with elf1Δ by standard tetrad analysis. These crosses demonstrated that only three of the nine mutations, cdc73Δ, leo1Δ, and rpb4Δ, actually caused synthetic lethality with elf1Δ, while some of the others, while viable, caused other mutant phenotypes (Table 3). As previously pointed out by others (38, 74, 75), analysis of genetic interactions by SGA often leads to false positives (see Materials and Methods). Of the three cases of synthetic lethality, two of them, cdc73Δ and leo1Δ, had been identified by the direct tests described in an earlier section. Consistent with our results, elf1Δ was also identified in a large-scale SGA screen to genetically interact with cdc73Δ (75).

The third mutation identified by SGA to cause synthetic lethality with elf1Δ was rpb4Δ. RPB4 encodes a nonessential subunit of Pol II (79). Based on this interaction, we also tested whether elf1Δ confers synthetic phenotypes when combined with three different alleles in the essential Pol II genes RPB1 and RPB2 that confer a 6AU^S phenotype. We found that these three mutations, rpb1-221, rpb2-7, and rpb2-10 (21, 52), cause increased sickness and sensitivity to MPA when combined with an elf1Δ mutation (data not shown). Thus, 6AU^S mutations in the genes encoding subunits of the Pol II complex can lead to an increased requirement for the Elf1 elongation factor.

Another class of mutations identified by the elf1Δ SGA analysis included the HIR/HPC genes. These four genes, HIR1, HIR2, HIR3/HPC1, and HPC2, were initially identified due to their roles in the regulation of histone gene transcription (49, 81). More recently, the HIR genes have been implicated in transcription elongation due to synthetic lethal interactions with mutations in spt16 and genes encoding members of the Paf1 complex (16). Although SGA analysis suggested that elf1Δ is synthetically lethal with mutations in the HIR/HPC genes, tetrad analysis demonstrated that the double mutants between elf1Δ and deletions in the HIR genes were all viable and healthy (Table 3). However, when we tested the elf1Δ hir/hpc double mutants, we found several synthetic phenotypes, including Ts⁻ and Spt⁻ phenotypes. Thus, there are genetic interactions present between elf1Δ and mutations in the HIR genes, although not lethality as implied by SGA analysis. Taken together, our results from SGA analysis have identified only genetic interactions between elf1Δ and mutations in genes implicated as playing a role in transcription elongation, further highlighting an important and specific role for Elf1 in transcription elongation.

Elf1 is localized to regions of active transcription.The genetic interactions of elf1Δ suggested that Elf1 might play a direct role in transcription elongation. To test if Elf1 is associated with transcribed regions in vivo, we performed ChIP experiments using a functional, 13xMyc-tagged version of Elf1 (see Materials and Methods). In these experiments, we tested the physical association of Elf1 with PMA1 and GAL1, two genes commonly used for such studies (25, 27, 28). PMA1 is a highly transcribed gene, and GAL1 is regulated by carbon source, allowing us to test whether Elf1 ChIP is dependent upon ongoing transcription. Our results show that Elf1 is localized at high levels over the coding region of PMA1 and of GAL1 when it is induced (Fig. 4A and B). Furthermore, Elf1 is present at reduced levels over the TATA regions and at low levels over the UAS regions of these genes (Fig. 4A and B). Comparison of the distribution of Elf1 to that of Pol II shows that the level of Elf1 association is lower than Pol II over the promoter region compared to the coding regions, suggesting that Elf1 is recruited to genes subsequent to Pol II recruitment (Fig. 4A and B). This pattern is similar to that observed for Spt16 (41) and Spt6 (25, 27). The localization of Elf1 to the GAL1 coding region is dependent on transcription, as we only observed the association of Elf1 with GAL1 upon the induction of transcription with galactose (Fig. 4B). Additionally, Elf1 appeared to be present at similar levels to Pol II over the 3′ UTR of both GAL1 and PMA1, similar to what has previously been observed for the elongation factors Spt4, Spt5, Spt6, and Spt16 (25, 27) and in contrast to other elongation factors, including the Paf1 and the TREX complexes, that show low levels of association over the 3′ UTR (25, 27). Taken together, these results suggest that Elf1 is preferentially localized to regions of active transcription.

• Open in new tab
• Download powerpoint

FIG. 4.

Elf1 is preferentially localized to transcribed regions of active genes. (A) Elf1 is localized to the transcribed region of PMA1. Yeast cells containing the Elf1-13Myc and Rpb3-HA epitope-tagged proteins were grown at 30°C to mid-log in YPD medium. Cells were cross-linked with 1% formaldehyde and lysed, and the chromatin was sheared by sonication to an average size of 400 bp. Elf1-13Myc was immunoprecipitated using the A14 anti-Myc antibody, and Rpb3-HA was immunoprecipitated using the 12CA5 anti-HA antibody. Radiolabeled PCR products were obtained with primers specific to the PMA1 UAS, TATA, 5′ ORF, 3′ ORF, or 3′ UTRs and were visualized on 6% polyacrylamide gels. As a control for DNA binding specificity, each PCR contained primers to amplify a nontranscribed region of chromosome V that is not contained in any open reading frames (28). ChIP values are represented as the fold enrichment of the specific PCR product compared to this control region. The locations of primer sets used for ChIP analysis are represented as lines below the schematic of the PMA1 gene. Results were obtained for at least three independent experiments, and the average and standard error are plotted on the graph. (B) Elf1 is localized across the GAL1 transcribed region upon transcriptional induction. Yeast cells containing Elf1-13Myc and Rpb3-HA epitope-tagged proteins were grown at 30°C to mid-log in medium containing 2% raffinose. Transcription of GAL1 was induced by the addition of 2% galactose for 2 hours. Chromatin immunoprecipitations were performed as described for panel A in at least three independent experiments, with the resulting averages and standard errors plotted on the graph.

Elf1 localization is partially dependent on Spt4 and Spt6.To test if the association of Elf1 with transcribed regions is dependent upon other elongation factors, we analyzed the recruitment of Elf1 in several elongation mutants. We found that recruitment of Elf1 to the coding region of PMA1 was decreased in spt4Δ and spt6-1004 mutants, while there were no strong effects on Elf1 localization in rtf1Δ, paf1Δ, or spt16-197 mutants (Fig. 5A). Although Elf1 binding is decreased slightly in the paf1Δ mutant, this is likely due to the decreased levels of Pol II binding seen in this mutant (Fig. 5A). These data suggest that Elf1 association with transcriptionally active regions is partially dependent on Spt4 and Spt6 and independent of the Paf1 complex and Spt16.

• Open in new tab
• Download powerpoint

FIG. 5.

Elf1 localization to PMA1 is partially dependent on the Spt4 and Spt6 elongation factors. (A) Elf1 localization at PMA1 in several elongation mutants. ChIP experiments were performed as described for Fig. 4A for wild-type, spt4Δ, rtf1Δ, paf1Δ, spt6-1004, or spt16-197 strains at 30°C. ChIP results for Elf1-13Myc (top panel) and Rpb3-HA (bottom panel) are shown. The results from at least three independent experiments are shown. (B) Localization of Spt4 and Ctr9 to PMA1 is not affected in an elf1Δ mutant. Wild-type or elf1Δ strains containing Spt4-Flag, Ctr9-Myc, and Rpb3-HA were subjected to ChIP analysis as described for Fig. 4A. Ctr9-Myc was immunoprecipitated using the A14 anti-Myc antibody, and Spt4-Flag was immunoprecipitated using the M2 anti-Flag antibody. The results from at least three independent experiments are shown.

To examine whether Elf1 is necessary for the association of other elongation factors with transcribed regions, we tested the recruitment of Spt4 and the Ctr9 subunit of the Paf1 complex to PMA1 in an elf1Δ mutant. In an elf1Δ mutant, the recruitment of Spt4 and Ctr9 was slightly reduced at PMA1, corresponding to a similar small reduction in the level of Pol II associated at PMA1. There was no effect on the overall distribution of Pol II in the elf1Δ mutant, as has been noted for other elongation factors (42, 43). These results strongly suggest that Spt4 and Ctr9 are recruited to transcribed regions independently of Elf1.

Evidence that Elf1 affects chromatin structure in actively transcribed regions.The genetic interactions of elf1Δ with elongation factors and its localization to actively transcribed regions are consistent with a role for Elf1 in transcription elongation. Based on the synthetic lethality observed between elf1Δ and mutations in SPT4 and SPT6 and the fact that the localization of Elf1 is partially dependent on Spt4 and Spt6, we hypothesized that Elf1 is performing a role related to these proteins in vivo. Recent studies have suggested a role for Spt6 in controlling chromatin structure over transcribed sequences, as spt6 mutations were found to cause defects in chromatin structure in actively transcribed regions. These defects allow transcription initiation from cryptic promoters within some coding regions (26).

To determine if Elf1 plays related roles, we employed a phenotypic assay based on the observation that spt6 mutations cause transcription initiation to start internally within the 3′ coding region of the FLO8 gene (26). We made use of a reporter construct in which the 3′ end of FLO8 has been replaced with the HIS3 gene (GAL1-FLO8-HIS3) (see Materials and Methods). In this construct, HIS3 can be expressed only upon transcription initiation from the FLO8 internal start site. When we analyzed an elf1Δ strain for internal initiation, by testing for growth on medium lacking histidine, we found that it was able to grow when GAL1-FLO8-HIS3 transcription levels are high (growth on galactose) but not when they are low (growth on glucose) (Fig. 6A). This result suggests that cells lacking elf1Δ are defective in maintaining proper chromatin structure at the GAL1-FLO8-HIS3 gene, thus allowing transcription to occur from a normally cryptic promoter within the coding region.

• Open in new tab
• Download powerpoint

FIG. 6.

Elf1 is necessary for maintaining proper chromatin structure in actively transcribed regions. (A) elf1Δ mutations allow use of a cryptic TATA element within the GAL1-FLO8-HIS3 reporter. Wild-type and elf1Δ strains containing the GAL1-FLO8-HIS3 reporter construct were patched onto YPD medium. Strains were then replica plated to synthetic complete medium, medium lacking histidine and containing glucose, and medium lacking histidine and containing galactose as the carbon source. Strains were scored for growth at 30°C for either 1 or 3 days. (B) Mutations in elf1Δ are Ts⁻ in combination with rtf1Δ or hir1Δ mutations. elf1Δ strains were crossed to rtf1Δ and hir1Δ mutants. Diploids were sporulated and dissected, and representative spores from the resulting complete tetrads were spotted onto YPD medium at 30°C or 39°C for 2 days. (C) elf1Δ hir1Δ mutants allow transcription to initiate from within the endogenous FLO8 coding region. Wild-type, elf1Δ, hir1Δ, elf11Δ hir1Δ, rtf1Δ, and elf1Δ rtf1Δ strains were grown in YPD to mid-log at 30°C. Strains were shifted to 39°C by addition of an equal volume of YPD medium prewarmed to 48°C. Strains were then incubated at 39°C, and cells were collected at 2 and 4 h after the temperature shift. Northern analysis was performed using probes to the 3′ region of FLO8 (+1515 to +2326) and also to the SCR1 loading control.

To look directly at this transcriptional effect, we also performed Northern analysis to examine the wild-type FLO8 gene for cryptic initiation in an elf1Δ mutant. In this case, the production of short transcripts was not detectable in the elf1Δ mutant (Fig. 6C), presumably due to the lower levels of transcription driven by the FLO8 promoter compared to the GAL1 promoter used for GAL1-FLO8-HIS3, and to the more modest phenotypes caused by elf1Δ compared to spt6 mutations. We also examined two double mutants for effects at FLO8, including elf1Δ rtf1Δ and elf1Δ hir1Δ. As described earlier, these double mutants possess synthetic Spt⁻ and synthetic Ts⁻ phenotypes when compared to the single mutants (Fig. 6B and data not shown). Analysis of the elf1Δ hir1Δ double mutant revealed the production of the FLO8 short transcript when shifted to the nonpermissive temperature of 39°C (Fig. 6C). The size of this short transcript is similar to that observed in the spt6-1004 mutant (data not shown). Although the elf1Δ rtf1Δ mutant also fails to grow at elevated temperatures, it does not produce FLO8 short transcripts when shifted to the nonpermissive temperature. This difference between the elf1Δ hir1Δ and the elf1Δ rtf1Δ double mutants suggest that these factors may facilitate elongation through distinct mechanisms. Together, these results strongly suggest that the role of Elf1 in transcription elongation involves effects on chromatin structure.

Elf1 associates with casein kinase II.In order to further characterize S. cerevisiae Elf1, we purified it using a C-terminal TAP tag (55) containing a calmodulin-binding peptide and Staphylococcus aureus protein A, separated by a TEV protease cleavage site. The tagged protein was then purified on IgG and calmodulin columns and analyzed by SDS-PAGE followed by staining with silver. Protein bands absent from a control preparation and corresponding to the tagged protein and any associated proteins were then identified by MALDI-TOF mass spectrometry (Fig. 7). Elf1-TAP copurified substoichiometrically with subunits of casein kinase II (CKII). We were able to detect the two largest components of CKII (Cka1 and Ckb1) using MALDI-TOF mass spectrometry on trypsin-digested silver-stained gel bands, but not its two smaller subunits (Cka2 and Ckb2), which coelectrophoresed with the much more abundant Elf1 polypeptide. Cka2 and Ckb2 were, however, detected by trypsin digestion and liquid chromatography/tandem mass spectrometry (LC-MS/MS) analysis of the entire purified protein preparation. Consistent with this association, Elf1 contains seven potential CKII sites in its carboxy-terminal end at positions S86, S95, S101, T109, S117, S124, and S142.

• Open in new tab
• Download powerpoint

FIG. 7.

Purification of Elf1. Purification of Elf1 was carried out with strains containing either an untagged protein or a TAP-tagged version of Elf1. The protein complex was purified and was then analyzed by SDS-PAGE and silver staining. Elf1 and the two largest subunits of CKII (Cka1 and Ckb1) were identified by trypsin digestion and MALDI-TOF. In the mass spectrum, there were 6 (35% coverage), 16 (43% coverage), and 10 (39% coverage) peptides used to identify Elf1, Cka1, and Ckb1, respectively. The two smaller subunits of CKII, which comigrate with Elf1-TAP, were identified (along with Cka1, Ckb1, and Elf1) by using gel-free LC-MS/MS. The + indicates the position of TEV protease. Other detectable bands are present in both lanes, including those at around 70 kDa that are likely the heat shock proteins Ssa1 and Ssb1, present in most purifications (N. Krogan and J. Greenblatt, unpublished results).