Deletion of Elongator Protein 1 (Elp1) relieves heterochromatin defects in a Pol II mutant of Schizosaccharomyces pombe

Abstract

Heterochromatin is a repressive epigenetic state that suppresses transcription and safeguards genomic integrity. However, the full mechanism of its regulation remains elusive. Here, we focus on a previously described RNA polymerase II (Pol II) variant called m203 in Schizosaccharomyces pombe, which has a single substitution mutation within the Rpb2 subunit of Pol II (rpb2-N44Y) that reduces RNA interference (RNAi)-dependent heterochromatin at a pericentromeric reporter locus. Through CRISPR-Cas9 site-directed mutagenesis, we find that rpb2-N44Y is a gain-of-function mutation. Furthermore, the heterochromatin defects of the m203 variant require a subunit of the Elongator complex called Elongator Protein 1 (Elp1), a protein that canonically promotes mcm⁵s²U₃₄ tRNA modifications. While the single knockout of Elp1 in the m203 strain majorly restored heterochromatin formation, single knockouts of the Elp3 or the Elp5 subunits of Elongator showed only modest effects. Furthermore, mcm⁵s² U₃₄ tRNA modifications are dispensable for Elp1-dependent heterochromatin. In contrast, the heterochromatin required core factors that are critical for heterochromatin formation, including protein mediators of the RNAi pathway and H3K9 methylation. Overall, our study reveals two distinct Rpb2-centric pathways, via RNAi or Elp1 that can positively or negatively regulate heterochromatin, respectively. Furthermore, our findings reveal a chromatin function for Elp1 that does not rely on Elongator-dependent mcm⁵s²U₃₄ tRNA modifications. This work expands our understanding of how Elp1 can influence chromatin biology.

Introduction

Eukaryotic chromatin consists of DNA and histone proteins that maintain genetic information. Chromatin organization can determine how genetic information is used. Chromatin can adopt two distinct organizational states called euchromatin and heterochromatin. Euchromatin is a relaxed state that is accessible by transcription machinery and is often marked by transcription-associated histone modifications such as tri-methylation of histone H3 lysine-36 (H3K36me3) (Georgescu et al. 2020; Morrison and Thakur 2021). In contrast, heterochromatin is condensed and transcriptionally repressed. Molecular hallmarks of heterochromatin include the enrichment of di/tri-methylation of histone H3 lysine-9 (H3K9me2/3), H4K20me3, and histone hypoacetylation (Grewal and Jia 2007; Dambacher et al. 2013). The distribution of euchromatin and heterochromatin helps ensure overall genome integrity and proper gene regulation (Grunstein et al. 1995; Murakami 2013; Papamichos-Chronakis and Peterson 2013). Understanding how these states are assembled is important for determining how eukaryotes respond to developmental and environmental cues (Zhu et al. 2013).

The fission yeast Schizosaccharomyces pombe is a powerful model organism for studying heterochromatin because it utilizes conserved regulatory factors and mechanisms while exhibiting less genetic redundancy compared to higher eukaryotes. Furthermore, S. pombe has high genetic tractability, which facilitates the deduction of molecular pathways through genetic means (Hoffman et al. 2015). As in animals and plants, heterochromatin in S. pombe is characterized by gene repression, position-effect variegation, and the presence of specific chromatin modifications such as H3K9me2/3 (Grewal and Jia 2007). H3K9 methylations are molecular platforms that recruit chromodomain proteins such as Swi6 (the yeast homolog of mammalian Heterochromatin Protein 1) and the histone deacetylase Clr3 (the yeast homolog of mammalian HDAC6/10) to mediate transcriptional repression and maintenance of the heterochromatic state (Sugiyama et al. 2007; Ragunathan et al. 2015; Seman et al. 2023). The initial establishment of H3K9me2/3 can occur via DNA- or RNA-dependent mechanisms. DNA-binding proteins such as Atf1 or Taz1 can recruit the sole H3K9 methyltransferase, Clr4, in S. pombe (Jia et al. 2004; Kanoh et al. 2005; Zofall et al. 2016). RNA-dependent H3K9me2/3 relies on long regulatory RNAs or small interfering RNAs (siRNAs) (Faber and Vo 2022), which can trigger heterochromatin establishment in cis (Verdel et al. 2004; Cam et al. 2005; Zofall et al. 2012; Tashiro et al. 2013; Kowalik et al. 2015; Asanuma et al. 2022). Mechanisms that couple H3K9me2/3 with siRNAs require RNA interference (RNAi) proteins, which are conserved in S. pombe, Tetrahymena, plants, and animals (Verdel et al. 2004; Matzke and Birchler 2005; Shabalina and Koonin 2008; Andrews et al. 2014; Gutbrod and Martienssen 2020).

In S. pombe, the establishment of RNAi-dependent heterochromatin requires Pol II-mediated transcription to generate precursor transcripts, which are then processed into mature siRNAs (Grewal and Elgin 2007; Holoch and Moazed 2015). Previous studies have linked heterochromatin assembly with three Pol II subunits in S. pombe (Rpb1/2/7) (Djupedal et al. 2005; Kato et al. 2005; Inada et al. 2016; Kajitani et al. 2017). Phosphorylatable serine residues in the C-terminal domain (CTD) of Rpb1 (Ser2 and Ser7) are needed for heterochromatin at pericentromeres and subtelomeres (Inada et al. 2016; Kajitani et al. 2017), with Ser7 being crucial for the production of siRNAs at the site of transcription (Kajitani et al. 2017). The Rpb7 Pol II subunit, which forms a dissociable heterodimer with the Rpb4 subunit and facilitates transcription initiation (Edwards et al. 1991; Choder 2004; Meka et al. 2005), promotes Pol II-mediated transcription of precursor siRNAs that promote heterochromatin (Djupedal et al. 2005). In the Rpb2 subunit, a missense variant called m203, which encodes for Rpb2-N44Y, produces normal levels of precursor siRNAs but exhibits reduced mature siRNAs at pericentromeres (Kato et al. 2005). The affected amino acid is positioned on the outer surface of Pol II (Spåhr et al. 2009) and is believed to modulate Pol II interactions with RNAi proteins (Kato et al. 2005). Outside of S. pombe, it was found that auxin-induced depletion of Rpb2, but not Rpb1, reduced the abundance of H3K9me3 in mouse embryonic stem cells (Bao et al. 2024). Although these results suggest that Rpb2 might have distinctive roles in H3K9 methylation compared to Rpb1, variations in knockout kinetics during the 3 h auxin treatment window may have also contributed to the observed differences between the subunits (Macdonald et al. 2022). Nevertheless, these earlier studies leave open the question of whether the link between Pol II and heterochromatin formation is essentially a reflection of overall transcriptional activity or whether Pol II may have other roles through interactions with factors that regulate heterochromatin.

Pol II-associated protein complexes play crucial roles in modulating the chromatin landscape (Gerber and Shilatifard 2003; Cramer 2004; Kim et al. 2004). One such complex, known as Elongator (Otero et al. 1999; Kim et al. 2002), was previously linked to gene activation and heterochromatic gene silencing. Elongator is comprised of six protein subunits called Elongator Protein 1 (Elp1)/2/3/4/5/6, in which Elp3 is an enzyme with a radical S-adenosylmethionine domain and a lysine acetyltransferase domain (Wittschieben et al. 2000; Winkler et al. 2001; Paraskevopoulou et al. 2006; Jaciuk et al. 2023; Abbassi et al. 2024). It was initially thought that Elongator could directly acetylate nucleosomes to promote transcription (Wittschieben et al. 2000; Winkler et al. 2002). Later, the Elp3 subunit was reported to promote heterochromatic silencing at telomeres and pericentromeres (Li et al. 2009; Bauer et al. 2012; Muhammad et al. 2024). However, it was later proposed that some of those transcription and heterochromatin observations were likely indirect consequences of a tRNA modification function carried out by an intact Elongator (Chen et al. 2011; Bauer et al. 2012). It remains unclear whether Elongator and/or its individual protein subunits have bona fide roles in heterochromatin regulation.

Here, we show that m203 is a gain-of-function genetic mutation within Rpb2 that impairs the formation of RNAi-dependent heterochromatin. By focusing on a pericentromeric reporter locus, we find that deleting the elp1 gene alleviates this defect. Although Elp1 is primarily known as a scaffolding protein subunit of the Elongator complex, which promotes 5-methoxycarbonylmethyl-2-thiouridine (mcm⁵s²) modification of transfer RNAs (tRNAs) (Huang et al. 2005; Dauden et al. 2019), we found that neither mutations of other Elongator-associated factors nor global inhibition of translation could fully mimic the pronounced impact of Elp1 loss on heterochromatin. Furthermore, mcm⁵s² tRNA modifications were dispensable for the heterochromatin role of Elp1. These results suggest that Elp1 may have a moonlighting role in epigenetic gene regulation.

Materials and methods

Yeast culturing and manipulation

S. pombe yeast strains that were used in this study are indicated in Supplementary Table 1. Strains were generated by using standard genetic crossing, by PCR-based gene deletion with transformation by lithium acetate method (Krawchuk and Wahls 1999), or by SpEDIT CRISPR-Cas9 editing (Torres-Garcia et al. 2020). Schizosaccharomyces pombe gene sequences were obtained from the Pombase database (Rutherford et al. 2024). All yeast were cultured in medias that were based on the yeast extract rich medium with adenine supplement and incubated at 32 °C for 2 to 4 d. Liquid cultures were grown at 32 °C with 220 rpm shaking for aeration. As appropriate for our experiments, we have used YEA media with the following: 5-FOA (GoldBio, 850 mg/L), cycloheximide (Sigma, 20 mg/mL), G418 sulfate (GoldBio, 50 mg/L), hygromycin (GoldBio, 200 µg/mL), and nourseothricin (GoldBio, 100 mg/L).

Serial spotting assays

Yeast cells were pre-cultured at 32 °C for 220 rpm overnight until cells reached log growth phase. Next, cell optical densities (OD₆₀₀) were measured using a Denovix DS-11 + apparatus and cells were normalized to the equivalent OD₆₀₀ of 0.2 to 0.5. Finally, cells were serially diluted 10-fold and equal volumes of cells per spotted onto the indicated media plates. Spotted plates were incubated at 32 °C without shaking for 2 to 4 d. Images were collected using a BioRad Chemidoc imager using the Colorimeteric setting. Any image correction was uniformly applied to the entirety of the shown images.

Generation of tRNA overexpression plasmids

tRNA expression plasmids were generated similar to how it was described in Fernández-Vázquez et al. (2013) with the major modification being that the plasmids were designed to be selectable with nourseothricin. This allowed us to culture our cells uniformly in YEA-based media, minimizing phenotypic variation due to major differences in media composition. Briefly, the pRep41X plasmid (Addgene) was modified by replacing the LEU2 gene with the natMX resistance gene between the HindIII and NheI restriction sites. The resulting construct was designated pRep_Nat. Next, the SPBTRNALYS.06 (encoding for tRNA^Lys_UUU) and SPCTRNALYS.11 (encoding for tRNA^Lys_CUU genes, along with 500 bp upstream and downstream DNA sequences, were PCR-amplified from the S. pombe genome as previously described by Fernández-Vázquez et al. (2013). These PCR products were cloned into the NheI site of the pRep_Nat plasmid. Plasmids used in this study are listed in Supplementary Table 2.

Derived Cleaved Amplified Polymorphic Sequences

This is a PCR-based method to detect specific gene variants from genomic DNA, without requiring direct DNA sequencing (Kaundun and Windass 2006). Derived Cleaved Amplified Polymorphic Sequences (dCAPS) was used to verify the rpb2 variants that were analyzed in this study. Briefly, yeast cells were first grown on YEA-based media plate. Next, cells were treated with zymolyase (Zymo Research) at 30 °C for 45 min and lysed by heating to 95 °C for 10 min. The yeast lysate, containing the released genomic DNA, was used as template for PCR (with GoTaq DNA polymerase) to amplify part of the rpb2 gene to include the codon that encodes for the 44th amino acid residue. PCR oligos are listed in Supplementary Table 3. Next, the PCR products were directly subjected to restriction digestion with BstZ171-HF (NEB), which would only cleave if the products were amplified from the rpb2-m203 variant. PCR products lacking the variant would not be digested, yielding a higher molecular-weight band in a DNA gel. An aliquot of digestion enzyme-treated PCR products were resolved in a 1% agarose gel that was pre-stained with ethidium bromide.

Chromatin immunoprecipitation–quantitative PCR

S. pombe cells were inoculated into 50 mL of YEA medium and incubated overnight at 32 °C with shaking at 220 rpm to an OD₆₀₀ of 0.5 to 0.6. Cells were crosslinked with 1% formaldehyde for 20 min at room temperature, followed by quenching with 2.5 M glycine. Then, cell pellets were washed twice with 20 mL of cold 1× PBS. The cell pellet was resuspended in 350 µL of chromatin immunoprecipitation (ChIP) lysis buffer (50 mM HEPES pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, EDTA free 1× cOmplete) and disrupted with 1 mL of glass beads using bead-beating (1 min beating, 2 min on ice, repeated 3 times). Chromatin was sheared using a QSonica Q800R sonicator (20 s ON/40 s OFF cycles, 70% amplitude) for 10 min at 4 °C. Cell debris was removed by centrifugation at 4 °C, 1,500 × g for 5 min. The supernatant was transferred to fresh 1.5 mL tubes and diluted with ChIP lysis buffer. Lysates were precleared with 20 µL of protein A/G agarose beads (Santa Cruz, cat no. A3124) and incubated at 4 °C for 1 h. 50 µL of pre-cleared lysate was reserved as the 5% input control. The remaining lysate was incubated with the H3K9me3 (Active Motif, cat no. 39062) antibody and protein A/G agarose beads overnight at 4 °C. Beads were washed sequentially (twice each) with the following buffers: ChIP Buffer I (50 mM HEPES pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% deoxycholate), ChIP Buffer II (same as Buffer I, but with 500 mM NaCl), ChIP Buffer III (10 mM Tris-HCl pH 8.0, 250 mM LiCl, 0.5% NP-40, 0.5% deoxycholate, 1 mM EDTA), and 1 × Tris-EDTA buffer.

Immunoprecipitated chromatin was extracted by incubating the agarose beads with 100 µL of elution buffer (50 mM Tris-HCl pH 8.0, 10 mM EDTA) at 65 °C and agitating at 1,000 rpm for 30 min. This step was repeated to obtain a total volume of 200 µL eluant. To each ChIP sample, 4 µL of 5N NaCl and 1 µL RNase A (Thermo Fisher Scientific, 20 mg/mL) was added, and samples were incubated overnight at 65 °C in a water bath to reverse crosslinks. Similarly, input samples were adjusted to 200 µL with elution buffer, followed by the addition of 2.6 µL of 5N NaCl and 1 µL RNase A, and incubated overnight at 65 °C in a water bath to reverse crosslinks. After the sample tubes were cooled to room temperature, 1 µL of Proteinase K (Thermo Fisher Scientific, 20 mg/mL) was added to each tube and incubated at 50 °C for 1 h. DNA was purified by ethanol precipitation using 3 M sodium acetate, 1 µL glycoblue (Thermo Fisher Scientific, 15 mg/mL) and resuspended into 50 µL of molecular-grade water. For qPCR, 1 µL of ChIP DNA was used in each reaction with iTaq SYBR Green reagents (BioRad). PCR oligos used for qPCRs are listed in Supplementary Table 3. Antibodies used for ChIP are listed in Supplementary Table 4.

Western blot

For protein extraction, 2 OD₆₀₀ units of yeast cells were harvested by centrifugation into 2 mL screw-cap tubes. The cell pellet was resuspended in 200 µL of cold 20% trichloroacetic acid (TCA). Approximately 400 µL of acid-washed glass beads were added, and the cells were lysed using mechanical disruption with a bead beater (20 s beating × 3 cycles with 45 s intervals on ice). Following lysis, each tube was punctured, placed in a 5 ml polystyrene collection tube, and the lysate was collected by centrifugation at 1,000 rpm for 30 s. The remaining beads were washed with 400 µL of cold 5% TCA, spun again, and the wash was pooled with the initial lysate. The combined lysate (∼600 µL total) was transferred to a 1.5 mL microcentrifuge tube and centrifuged at 13,000 rpm for 5 min at 4 °C to pellet the proteins. Supernatant was discarded and centrifuged 2 to 3 times to remove residual TCA. The resulting protein pellet was resuspended in 50 to 100 µL of SDS-PAGE 1 × sample loading buffer (Thermo Fisher Scientific, cat no. NP0007-2988243). Pellets were often difficult to dissolve and required heating to fully solubilize. Samples were then boiled at 95 °C and loaded on SDS-PAGE 4% to 12% Bis-Tris gels (Thermo Fisher Scientific, cat no. 25010870). Electrophoresis was performed in an Invitrogen mini tank apparatus using NuPage 1 × running buffer at room temperature at 200 volts for 45 min. Proteins were transferred onto PVDF membranes (pre-activated with methanol) using wet transfer (Nupage 1 × transfer buffer (NP0006-2988243) at 20 volts for 60 to 90 min at room temperature.

Following the transfer, membranes were blocked for 1 h at room temperature in 1 × TBST (Tris-buffered saline with 0.1% Tween-20) containing 5% non-fat dry milk. Membranes were then incubated overnight at 4 °C with 1:2,000 mouse primary antibody that was diluted in blocking buffer. After washing three times with 1 × TBST (10 min each), membranes were incubated with an HRP-conjugated secondary anti-mouse (Jackson laboratory, cat no. 115-035-146) for 1 h at room temperature. Following three additional 1 × TBST washes, signal detection was performed using an enhanced chemiluminescence (Cytiva, cat no. RPN2232-18206643) substrate and visualized using the BioRad chemiluminescence imaging system. Antibodies used are listed in Supplementary Table 4.

Isolation of total RNA

Total RNA was isolated from 10 mL of S. pombe cells that were grown in liquid YEA media at 32 °C to OD₆₀₀ between 0.5 and 1. Before starting the isolation procedure, all centrifuges were set to 6 °C, and 100% isopropanol was kept at −20 °C. A water bath was set to 65 °C with enough water to submerge half of a 15 mL Falcon tube. Heavy phase lock tubes (Qiagen) were centrifuged and placed on ice. The cell pellet was thawed on ice (if frozen), and any excess media was removed. 1 mL of AES buffer (50 mM sodium acetate pH 5.3, 10 mM EDTA, 1% SDS) was added to the cell pellet and mixed gently using pipetting. In a fume hood, 1 mL of acidic phenol (pH 4.5) was added to the cell suspension and gently mixed by inverting the tube 10 times. Next, the sample was vortexed for 10 s and incubated in the 65 °C water bath for 45 min. This incubation was repeated five times without interruption. The resulting cell-AES-phenol mixture was transferred into heavy phase lock tubes and centrifuged at 4,000 × g for 5 min at 6 °C, taking care not to disturb the two layers. Next, 1 mL of phenol:chloroform:isoamyl alcohol (25:24:1) (VWR) was added to the sample, and the tube was inverted gently 10 times to mix, followed by centrifugation at 4,000 g for 5 min at 6 °C. In the fume hood, 1 mL of chloroform was added to each sample, gently mixed by inversion (10 times), and centrifuged again at 4,000 g for 4 min at 6 °C. Carefully, 700 µL of the top clear aqueous phase was pipetted into a 1.5 mL microcentrifuge tube. To this, 70 µL of 3 M NaOAc and 1 µL of glycoblue (Thermo Fisher Scientific, 15 mg/mL) were added and mixed thoroughly by inverting the tube 10 times. Ice-cold 100% isopropanol was then added, and the tube was inverted 10 times to mix. The sample was centrifuged at maximum speed for 20 min at 6 °C. A moderate RNA pellet was observed, the supernatant was discarded, and the pellet was washed with 900 µL of 75% ethanol by inverting the tube 10 times, followed by centrifugation at maximum speed for 5 min at 6 °C. The supernatant was removed using a P1000 pipette, and any residual liquid was removed using a P200 or P20 pipette. The pellet was air-dried for 10 min before adding 100 µL of DEPC-treated water. The sample was incubated at room temperature for 5 min to dissolve the pellet and mixed gently by pipetting. RNA concentration was measured using a DS-11 + spectrophotometer. Then, 10 µg of isolated RNA was then treated with Turbo DNase (Thermo Fisher Scientific, cat no. AM2238) in 50 ul of total volume to remove genomic DNA. Then, the volume was brought to 100 µL by adding DEPC water. 100 µL of phenol:chloroform:isopropyl alcohol was added. The sample was gently inverted 10 times to mix and centrifuged at 20,000 × g for 5 min at 6 °C to separate the organic (bottom) and aqueous (top) layers. 70 µL of the top aqueous layer was carefully transferred to a fresh microcentrifuge tube. To this, 7 µL of 3 M sodium acetate and 1 µL of glycogen were added and mixed by gentle tapping. Then, 235 µL of 200-proof ethanol was added and mixed by inverting the tube 40 times. The tubes were incubated overnight at −20 °C to precipitate the RNA. The sample was centrifuged at maximum speed (20,000 × g) for 30 min at 6 °C to pellet the purified RNA. After centrifugation, the supernatant was carefully discarded using a P1000 pipette, leaving behind the RNA pellet. To wash the pellet, 900 µL of 75% ethanol was added, and the tube was inverted 10 times to mix. The sample was then centrifuged again at 20,000 × g for 5 min at 6 °C. As much of the supernatant as possible was removed, followed by a brief spin to collect residual liquid, which was removed using a P10 pipette. The pellet was air-dried for 5 min to allow the remaining ethanol to evaporate. Finally, 20 µL of DEPC-treated water was added to dissolve the RNA. The tube was incubated at room temperature for 10 min, mixed gently by pipetting, and the RNA concentration was measured. Successful DNase treatment was validated by synthesizing cDNA using the Invitrogen SuperScript III Reverse Transcriptase kit, followed by PCR to confirm the absence of genomic DNA contamination. Tapestation analysis at the MSU Genomics core facility was performed to confirm the quality of RNA based on RIN values.

Small RNA sequencing

Total RNA (free from contaminating genomic DNA) was used to prepare small RNA libraries with 18 to 40 bp inserts. Briefly, 3′ and 5′ Illumina sequencing adapters were ligated to the end of RNAs. Next, cDNA was generated and subsequently used for PCR amplification. Next, libraries were resolved using gel electrophoresis and were size-selected. Sequencing of the small RNA-seq libraries was performed on an Illumina NovaSeq with single-end reads of 50 nt. Approximately 10 to 14 million raw reads were generated per sequenced sample.

The raw sequencing data was first assessed by FastQC to ensure high-quality data. Next, Cutadapt was used to remove poor quality reads, reads with lengths below 15 nt, and adapter sequences. The remaining reads (74% to 80% of total raw reads) were aligned to the S. pombe genome (May 2012 version) using Bowtie. Samtools was used to keep alignments from reads of lengths 21 to 24 nt. After, only uniquely mapped reads (-q 40) were used for coverage calculations. Coverage was calculated using the per million scaling factor. Additionally, total coverage was determined as the sum of coverage on the positive strand (having positive coverage) and on the negative strand (having negative coverage). siRNAs that are generated through bidirectional transcription and/or the activity of RNA-dependent RNA polymerase typically show total coverage that spans both positive and negative strands (Yamanaka et al. 2013).

Detection of tRNA modifications by liquid chromatography mass spectrometry

Detection of cellular tRNA modifications was performed according to ArrayStar, Inc. (Maryland, USA), based on a previously described procedure (Su et al. 2014). Briefly, tRNAs were isolated from ≥5 μg of total RNA (free from contaminating genomic DNA) using urea-PAGE electrophoresis to extract 70 to 90 nt small RNAs. RNAs were excised from the gel and purified by ethanol precipitation. Once isolated, the tRNAs underwent enzymatic digestion to produce dephosphorylated, single nucleosides. Nucleosides were deprotonized using Satorius 10,000-Da MWCO spin filter. These purified nucleosides were subsequently analyzed by LC-MS/MS, on an Agilent 6460 Triple Quadrupole mass spectrometer with an Agilent 1260 HPLC system, to identify modified nucleosides. Multi reaction monitoring (MRM) mode was performed to ensure high selectivity and sensitivity while working with parent-to-product ion transitions. LC-MS data was acquired using Agilent Qualitative Analysis software. Signal intensities were normalized by the amount of total tRNAs that were loaded. Finally, the resulting data were processed and normalized using Agilent Qualitative Analysis software, where MRM peaks for each modified nucleoside were extracted.

Phylogenetic analyses

Protein sequences of DNA-dependent RNA polymerase II subunit beta from 45 different organisms were downloaded from the UniProt database in FASTA format. The organisms were selected to broadly represent different eukaryotic kingdoms. The sequences were aligned using Clustal Omega (with the ClustalW output format) (Madeira et al. 2024). The alignment was viewed using Jalview (version 2.11.5.0). A conservation graph was generated for each aligned position based on the physio-chemical properties of the queried amino acids, according to the previously described Analysis of Multiply Aligned Sequences method of multiple sequence alignment analysis (Livingstone and Barton 1993), using S. pombe Rpb2 as the reference. For each bar within the graph, a value of “*” indicates perfect amino acid identity, a value of “+” indicates all properties of the amino acids at the homologous positions are conserved, and lower values indicate varying levels of conservation. Additionally, conserved residues were colored in Jalview according to the BLOSUM64 alignment scoring matrix (with conservation ≥30%) (Eddy 2004). Highlights of dark blue indicate identity with the consensus residue, light blue denotes residues similar to the consensus, and white represents non-conserved residues.

Supplementary Table 5 lists the Uniprot ID and amino acid sequences of all analyzed Rpb2 homologs. Additionally, it lists the NCBI Taxonomy Identifier and the kingdom of each organism that was analyzed.

Results

Gain-of-function Rpb2 mutations promote heterochromatic defects

The m203 variant is defined by an asparagine (Asn, N) to tyrosine (Tyr, Y) substitution at amino acid residue 44 of Rpb2 (N44Y), the second largest subunit of Pol II (Kato et al. 2005; Spåhr et al. 2009). The residue position is located on the outer surface of Pol II (Kato et al. 2005). Cells of this genotype lose heterochromatic silencing of a ura4⁺ reporter gene, which has been inserted into the pericentromere of chromosome 1 (otr1R::ura4⁺) (Kato et al. 2005) (Fig. 1a). While prior analyses have indicated that m203 leads to heterochromatin defects through the disruption of the RNAi pathway (Kato et al. 2005), it remains unclear how this occurs. We first determined whether m203 is a loss- or gain-of-function mutation by performing CRISPR-Cas9 site-directed mutagenesis to substitute Asn⁴⁴ with other amino acid residues and indirectly measured their impacts on the repression of the otr1R::ura4⁺ reporter (Fig. 1a). In wild-type cells with Rpb2-Asn⁴⁴ (WT), heterochromatin forms at the otr1R::ura4⁺ reporter which leads to transcriptional repression of ura4⁺ (Allshire et al. 1994). Accordingly, wild-type cells can readily form yeast colonies in the presence of the drug 5-fluoroorotic acid (5-FOA) (Fig. 1b). Cells that do not express ura4⁺ will not convert 5-FOA into a toxic metabolite and will proliferate (Verdel et al. 2004; Kato et al. 2005; Cam and Whitehall 2016). Mutant cells that relieve heterochromatin at the otr1R::ura4⁺ reporter, such as m203 cells (Kato et al. 2005), express ura4 + that promotes metabolism of 5-FOA into a toxic product that inhibits growth (Fig. 1b). We tested the hypothesis that the similarity of amino acid side-chains would predict the growth patterns of cells on media with 5-FOA. Cells in which Rpb2-Asn⁴⁴ was substituted with glutamine (Glu, Q), whose side-chain is most similar to that of Asn, grew similarly to wild-type in the presence of 5-FOA (Fig. 1b). Cells where Rpb2-Asn⁴⁴ was substituted with phenylalanine (Phe, F), which has a bulky side-chain similar to Tyr, grew as poorly as rpb2-N44Y (m203) cells on media with 5-FOA (Fig. 1b). We further validated by chromatin immunoprecipitation with quantitative PCR (ChIP-qPCR) that both rpb2-N44Y and rpb2-N44F mutants decreased the levels of heterochromatic H3K9me3 at the otr1R::ura4⁺ reporter (Fig. 1c). Together, our results support the notion that the side-chain of the 44th amino acid of S. pombe Rpb2 is important for pericentromeric heterochromatin.

Fig. 1.

The 44th amino acid residue of Rpb2 is a critical determinant of heterochromatin status at the pericentromere. a) Schematic illustration of the pericentromeric otr1R::ura4⁺ reporter that is used in this work. b) Serial spotting assay to assess the impact of rpb2 substitution mutations on yeast growth on media −/+ 5-FOA. The wild-type strain is denoted by WT. The rpb2-N44Y (2005) yeast strain was previously used in Kato et al. (2005). The other rpb2-N44Y yeast strain was generated by our lab using CRISPR-Cas9. c) ChIP enrichment of H3K9me3 at otr1R::ura4⁺ in yeast strains of various rpb2 genotypes. Enrichments were normalized relative to the euchromatic leu1 gene and a 5% input control. Error bars indicate standard deviation based on n ≥ 2.

Next, we substituted Rpb2-Asn⁴⁴ with alanine or serine (rpb2-N44A or rpb2-N44S), amino acids that are highly dissimilar to the other amino acids tested. We found that cells with rpb2-N44A and rpb2-N44S variants proliferated similarly to wild-type cells in the presence of 5-FOA (Fig. 1b). Furthermore, there was still H3K9me3 enrichment at otr1R::ura4⁺ in the rpb2-N44S variant (Fig. 1c). This suggests that the dissimilar variants rpb2-N44, -N44A, and -N44S can all support heterochromatic repression of the otr1R::ura4⁺ reporter. We propose that rpb2-N44Y and rpb2-N44F are gain-of-function mutations that can disrupt the assembly of H3K9me3-marked heterochromatin. Supporting this conclusion, substitution of N44 with tryptophan, lysine, methionine, or glutamic acid resulted in wild-type growth phenotypes on media with 5-FOA (Supplementary Fig. 1a), suggesting that only specific substitutions at the 44th position of Rpb2 lead to loss of heterochromatic repression.

To investigate whether the 44th position of S. pombe Rpb2 is possibly under evolutionary constraints, we performed multiple-sequence alignment of S. pombe Rpb2 and its homologs across 44 other organisms (see Materials and methods). These organisms span broadly across eukaryotes to include metazoans, plants, fungi, amoeba, and ciliates (Supplementary Table 5). By analyzing the first 70 residues at the N-terminus of S. pombe Rpb2, we found that the 44th position lies within a well-conserved region of the protein (Supplementary Fig. 1b). Furthermore, we observed that certain fungi, ciliates, and amoebas have asparagine (N) in positions that are homologous to the 44th position of S. pombe Rpb2 (Supplementary Fig. 1b). Interestingly, some of those ciliates, such as Paramecium tetraurelia, are known to have RNAi mechanisms (Marker et al. 2014). In contrast, tyrosine (Y) is used in other fungi and certain ciliates such as Stylonychia lemnae (Supplementary Fig. 1b). A previous analysis of Rpb2 homologs from ciliates also revealed a preference for tyrosine or phenylalanine in other ciliate organisms such as Oxytricha trifallax (Khurana et al. 2014). On the other hand, most metazoans use methionine (M) instead (Supplementary Fig. 1b). Our analyses show that the 44th position of S. pombe Rpb2 lies within a conserved region of the protein and that there may be evolutionary pressure on which specific amino acid exists at this position. These findings support the notion that the 44th position might be functionally important.

Heterochromatin loss in rpb2-N44Y and -N44F mutants require the elp1 gene

To identify the factors that mediate heterochromatin loss in the rpb2-N44Y (m203) and rpb2-N44F mutant cells, we sought suppressor mutations that could restore heterochromatic repression of our otr1R::ura4⁺ reporter. We deleted individual genes that are associated with transcription elongation to test whether any single deletion could repress our otr1R::ura4⁺ reporter in m203 cells (Fig. 2a). We selected the tfs1 gene (encodes for a homolog of mammalian TFIIS) because this mutant suppresses heterochromatic defects in S. pombe cells that lack the RNAi gene called ago1 (Reyes-Turcu et al. 2011). We also selected the rpb9 and elp1 genes because they have been implicated in the regulation of transcription elongation (Otero et al. 1999; Hemming et al. 2000). We discovered that loss of elp1 (elp1Δ) in m203 cells enabled strong growth in the presence of 5-FOA, similar to wild-type cells (Fig. 2a and Supplementary Fig. 2a). The growth phenotype of m203 elp1Δ cells on 5-FOA media persisted through genetic crosses, indicating that it is dependent on the rpb2 and elp1 genotypes (Fig. 2b). Furthermore, ChIP-qPCR confirmed that elp1Δ increased heterochromatic H3K9me3 and decreased euchromatic H3K36me3 at our reporter locus in m203 cells (Fig. 2c and d). The changes in H3K9me3 likely affect the expression state of our otr1R::ura4⁺ reporter because m203 elp1Δ cells reverted to hypersensitivity on 5-FOA-containing media after the loss of the clr4 gene (which encodes for the sole H3K9 methyltransferase) (Fig. 2e). Lastly, we confirmed that elp1Δ in rpb2-N44F mutant cells also robustly grew on 5-FOA media (Supplementary Fig. 2b). Collectively, our results suggest that the rpb2-N44Y and rpb2-N44F mutants promote an elp1-dependent process that reduces the heterochromatic repression of our pericentromeric reporter.

Fig. 2.

Loss of elp1 restores heterochromatic repression of the otr1R::ura4⁺ reporter in m203 (rpb2-N44Y) cells. a) Serial spotting assay to determine gene deletions that can restore 5-FOA resistance in m203 cells. b) Tetrad analysis reveals that the segregation of the 5-FOA resistance phenotype follows the elp1Δ genotype. Shown are the growth phenotypes for yeast derived from a representative tetrad. Cells were grown on non-selective (n/s) YEA media, media lacking uracil, or media containing 5-FOA. c) ChIP enrichment of heterochromatic H3K9me3 at otr1R::ura4⁺ in the indicated yeast strains. d) ChIP enrichment of euchromatic H3K36me3 at otr1R::ura4⁺ in the indicated yeast strains. e) Spotting assay showing that loss of the H3K9 methyltransferase clr4 reverts the 5-FOA sensitivity of m203 elp1Δ cells to mirror the sensitivity of m203 cells. ChIP enrichments were normalized relative to the euchromatic leu1 gene and a 5% input control. Error bars indicate standard deviation and experiments were based on n ≥ 2.

We considered the possibility that Elp1 could regulate the ura4⁺ gene outside of the heterochromatin context. To test this, we examined the impact of elp1 loss in cells that expressed ura4⁺ from its native euchromatic locus. Similar to wild-type cells with ura4⁺ in euchromatin, elp1Δ cells robustly grew on media lacking uracil but not on media containing 5-FOA (Supplementary Fig. 2c). Furthermore, by measuring ura4 transcript levels using reverse transcriptase PCR (RT-PCR), we confirmed that steady-state ura4 RNA levels were not reduced in elp1Δ cells relative to wild-type (Supplementary Fig. 2d). This suggests that loss of elp1 does not intrinsically decrease ura4 RNA expression. Elp1 regulation of otr1R::ura4⁺ expression is likely based on the chromatin state.

Loss of elp1 enhances DNA damage sensitivity (Li et al. 2009; Chen et al. 2021). Therefore, we tested the possibility that elp1Δ could revert the rpb2-N44Y mutation of m203 cells to the wild-type allele (rpb2-N44) to restore heterochromatin. Towards this end, we isolated genomic DNA from m203 elp1Δ double mutant cells and performed dCAPS (Kaundun and Windass 2006) to probe for the wild-type and m203 alleles of rpb2. We detected rpb2-N44Y as the dominant allele in m203 elp1Δ cells (Supplementary Fig. 2e), suggesting that elp1 loss does not restore heterochromatic H3K9me3 by majorly promoting the correction of the rpb2 mutant allele.

Elp1 can also regulate some native heterochromatin domains

There are native RNAi-dependent heterochromatin regions in S. pombe that are known as heterochromatin domains (HOODs) (Yamanaka et al. 2013). Unlike pericentromeric heterochromatin, which is constitutive, HOODs are facultative regions that can exist as euchromatin or heterochromatin (Cam et al. 2005; Yamanaka et al. 2013). Upon the loss of the RNA exosome subunit Rrp6, RNAi promotes heterochromatic di-methylation of H3K9 (H3K9me2) at HOODs that repress developmentally or environmentally responsive genes (Yamanaka et al. 2013; Shah et al. 2014). We found that m203 reduced H3K9me2 at one HOOD known as HOOD-23, which encompasses the pho1 gene (Fig. 3a) (Yamanaka et al. 2013). Interestingly, elp1 loss restored H3K9me2 at HOOD-23 in m203 cells. We also tested whether elp1Δ can restore H3K9me2 at another HOOD known as HOOD-10 (which encompasses the Tf2 retrotransposon element tf2-5) (Yamanaka et al. 2013). However, we did not find H3K9me2 enhancement in m203 rrp6Δ elp1Δ cells (Fig. 3b). This suggests that the role of Elp1 varies depending on the heterochromatin region. Overall, our results suggest that Elp1 could negatively regulate H3K9 methylation at diverse native and reporter loci.

Fig. 3.

Loss of elp1 restores heterochromatic H3K9me2 at certain HOODs. a) ChIP enrichment of heterochromatic H3K9me2 at HOOD-23, which encompasses the annotated pho1 gene, in the indicated yeast strains. b) ChIP enrichment of heterochromatic H3K9me2 at HOOD-10, which encompasses the annotated tf2-5 gene, in the indicated yeast strains. At both HOODs, H3K9me2 is known to be induced upon the loss of the rrp6, which encodes for a catalytic subunit of the nuclear exosome (Yamanaka et al. 2013). ChIP enrichments were normalized relative to the euchromatic leu1 gene and a 5% input control. Error bars indicate standard deviation based on n ≥ 2.

Suppression of m203 by elp1Δ requires the RNAi pathway

The requirement of the RNAi pathway for heterochromatic repression of the otr1R::ura4+ reporter (Verdel et al. 2004) prompted us to investigate whether Elp1 may affect RNAi. First, we performed small RNA sequencing and mapped 21 to 24 nucleotide (nt) siRNAs to the S. pombe genome. We analyzed siRNAs that are derived from pericentromeric dh repeats because they promote heterochromatin at our reporter (Verdel et al. 2004). We found that elp1Δ increases pericentromeric siRNAs in m203 cells (Fig. 4a). Next, we indirectly tested whether repression in m203 elp1Δ cells required the RNAi pathway by deleting ago1, dcr1, or rdp1 in this strain (Fig. 4b). The ago1 and dcr1 genes encode for homologs of metazoan RNAi factors Argonaute or Dicer proteins, respectively (Tabara et al. 1999; Zamore et al. 2000; Bernstein et al. 2001; Verdel et al. 2004). The rdp1 gene encodes for an RNA-dependent RNA polymerase that is required for siRNA amplification and heterochromatin formation in S. pombe (Sugiyama et al. 2005). We found that the loss of ago1, dcr1, or rdp1 reverted cells to the 5-FOA-sensitive state, suggesting that elp1Δ depends on the RNAi pathway to repress the otr1R::ura4⁺ reporter (Fig. 4b). In elp1Δ cells with the wild-type rpb2 allele, RNAi-dependent heterochromatin represses the otr1R::ura4⁺ reporter (Verdel et al. 2004) and knockout of elp1 has no apparent effect (Supplementary Fig. 3). This suggests that rpb2 may regulate the contribution of Elp1 and RNAi towards negative and positive regulation of heterochromatin, respectively.

Fig. 4.

Heterochromatin regulation by elp1 requires RNAi. a) Coverage of mapped 21 to 24 nucleotide Illumina reads from small RNA-seq for the indicated yeast strains at the native dh repeats within the otr1R pericentromeric locus. The coverage was calculated as the sum across both positive and negative strands. b) Spotting assays showing that loss of the RNAi-related genes ago1, dcr1, or rdp1 enhances 5-FOA sensitivity in m203 elp1Δ cells.

The Elp1 subunit of the Elongator complex distinctly contributes to heterochromatin

Elp1 is a core subunit of the Elongator complex, which is comprised of six proteins (Elp1-6) (Otero et al. 1999; Winkler et al. 2001; Jaciuk et al. 2023). Within this complex, Elp1 interacts with the other subunits and tRNAs to promote Elongator-mediated tRNA modifications (Dauden et al. 2019; Jaciuk et al. 2023). We tested whether ablating the other subunits of Elongator in m203 cells would phenocopy the deletion of elp1 and restore heterochromatin (Fig. 2a). To do this, we knocked out the elp3 or elp5 gene (elp3Δ or elp5Δ) in m203 cells that carried our otr1R::ura4⁺ reporter. While m203 elp1Δ robustly grew on 5-FOA media (Fig. 2a), m203 cells lacking elp3 or elp5 only had modest growth (Fig. 5a). These phenotypes correlated with partial increases in H3K9me3 at the otr1R::ura4 + reporter in m203 cells lacking elp3, compared to a higher H3K9me3 restoration in m203 elp1Δ cells (Fig. 5b), compared to m203 cells. Interestingly, knocking out both Elp1 and Elp3 in m203 cells (m203 elp1Δ elp3Δ) reduced H3K9me3 enrichment at the otr1R::ura4⁺ reporter, relative to m203 elp1Δ cells (Fig. 5b). Our results suggest that Elp3 partially promotes heterochromatic restoration in m203 elp1Δ cells.

Fig. 5.

Elp1 and Elp3 have distinct contribution to heterochromatin regulation. a) Spotting assays depicting 5-FOA sensitivities of the indicated yeast strains. For the m203 elp3Δ and m203 elp5Δ strains, two independent isolates are shown. b) ChIP enrichment of heterochromatic H3K9me3 at otr1R::ura4⁺ in the indicated yeast strains. Enrichments were normalized relative to the euchromatic leu1 gene and a 5% input control. Error bars indicate standard deviation based on n ≥ 2. Asterix (*) denotes statistical significance as determined by Student's t-test.

Elp3-mediated tRNA modification is dispensable for heterochromatin loss in m203 cells

From yeast to humans, Elongator promotes the mcm⁵s² modification of wobble uridine (U₃₄) of tRNAs that recognize AA/AG-ending mRNA codons (Huang et al. 2005; Karlsborn et al. 2014b; Abbassi et al. 2020). This modification regulates codon-anticodon base pairing to influence translation and polypeptide folding (Laguesse et al. 2015; Nedialkova and Leidel 2015; Abbassi et al. 2020). Since AA/AG-ending codons are pervasive throughout the transcriptome (Bauer et al. 2012; Goffena et al. 2018), it is thought that the loss of mcm⁵s²-modified tRNAs in Elongator mutants is responsible for diverse defects, including transcriptional dysregulation, altered exocytosis, and cytoskeletal reorganization (Esberg et al. 2006; Svejstrup 2007; Johansson et al. 2018).

To determine whether defects in mcm⁵s² tRNA modifications could explain the epigenetic role of Elp1 in m203 cells, we introduced mutations within the Elp3 catalytic subunit of Elongator. By aligning the protein sequences of Elp3 homologs from M. jannaschii, S. pombe, and S. cerevisiae, we identified a pair of cysteine residues at S. pombe Elp3 positions 106 and 109 that may participate in an iron-sulfur cluster (Fe₄S₄) (Fig. 6a). In M. jannaschii and S. cerevisiae, the Elp3 Fe₄S₄ cluster is required for the mcm⁵s²U₃₄ modification function of Elongator (Paraskevopoulou et al. 2006; Abbassi et al. 2024). Using CRISPR-Cas9, we substituted those cysteines for alanines to generate an elp3-(C106A,C109A) mutant strain that we called elp3-CC. We first confirmed that Elp1 protein was expressed in elp3-CC mutant cells (Supplementary Fig. 4a). Next, we performed mass spectrometry of total tRNAs to confirm that elp3-CC cells completely lost mcm⁵s² tRNA modifications (Fig. 6b), which is also in elp1Δ and elp3Δ cells (Huang et al. 2005; Karlsborn et al. 2014b). Due to the lack of tRNA modification capability, the elp3-CC mutant was hypersensitive to H₂O₂-induced oxidative stress (Supplementary Fig. 4b), as previously reported for the elp1Δ and elp3Δ mutants (Fernández-Vázquez et al. 2013). These results demonstrate that our Elp3-CC mutant completely ablates mcm⁵s² tRNA modifications while permitting Elp1 protein expression.

Fig. 6.

Loss of mcm⁵s²U₃₄ tRNA modifications does not restore heterochromatic repression of otr1R::ura4⁺ in m203 cells. a) Alignment of homologous Elp3 protein sequences from the thermophilic archaea Methanocaldococcus jannaschii, the budding yeast Saccharomyces cerevisiae, and the fission yeast Schizosaccharomyces pombe. A partial view of the alignment that is focused on the conserved cysteine residues of the iron-sulfur cluster (underlined) is shown. Cysteine residues at positions 106 and 109 of S. pombe Elp3 were substituted for alanines to generate the Elp3-CC mutant. b) Peak area of detected 5-methoxycarbonylmethyl-2-thiouridine (mcm⁵s²U) in the total tRNAs that were derived from the indicated yeast strains. c) Spotting assays depicting 5-FOA sensitivities of the indicated yeast strains. Four-fold serially diluted cells were spotted. d) ChIP enrichment of heterochromatic H3K9me3 at otr1R::ura4⁺ in the indicated yeast strains. e) Spotting assays depicting 5-FOA sensitivities of WT, m203, ctu2Δ, and m203 ctu2Δ yeast strains. f) ChIP enrichment of heterochromatic H3K9me3 at otr1R::ura4⁺ for the yeast strains shown in panel E. ChIP enrichments were normalized relative to the euchromatic leu1 gene and a 5% input control. Error bars indicate standard deviation based on n ≥ 2. For panels D and F, the enrichments for WT and m203 strains are reused because they were performed in the same experiment.

Similar to the elp3Δ mutant, elp3-CC did not strongly restore growth of m203 cells on 5-FOA media (Fig. 6c). Validating this finding, H3K9me3 ChIP-qPCR analysis showed that elp3-CC only slightly restored H3K9me3 at otr1R::ura4 + in m203 cells (Fig. 6d). We also deleted the ctu1 gene (ctu1Δ), which encodes for the Ctu2 tRNA thiolase that is required for adding the sulfur group (s²) within the mcm⁵s² tRNA modification (Dewez et al. 2008), and found weak growth of m203 ctu1Δ cells on 5-FOA media (Fig. 6e). There was also minor restoration of H3K9me3 at the reporter locus in cells m203 ctu1Δ (Fig. 6f). Altogether, these data suggest that mcm⁵s² tRNA modifications play minor roles in the epigenetic function of Elp1 at our pericentromeric reporter.

Elongator-mediated tRNA modifications promote translation by preventing ribosomal stalling (Nedialkova and Leidel 2015). We tested whether partial inhibition of translation would be sufficient to restore heterochromatic repression at otr1R::ura4⁺ in m203 cells. Using 20 μM of cycloheximide, there was no restoration of m203 cell growth on 5-FOA media (Supplementary Fig. 5). Along with our elp3-CC results, this suggests that the epigenetic role of Elp1 involves processes beyond mcm⁵s² tRNA modifications and global translation.

Overexpression of tRNA^Lys_UUU that lack mcm⁵s² modification suppresses elp1Δ

One tRNA that Elongator modifies is tRNA^Lys_UUU (Huang et al. 2005; Karlsborn et al. 2014b). Although tRNA^Lys_UUU and tRNA^Lys_CUU are both charged with lysine, the latter tRNA cannot be modified by Elongator because its wobble nucleoside is cytosine. It was previously shown in S. cerevisiae and S. pombe yeast that overexpressing tRNA^Lys_UUU, but not tRNA^Lys_CUU, in elp3Δ cells suppresses Elongator-associated defects in transcription, cell cycle, and stress response (Esberg et al. 2006; Bauer et al. 2012; Fernández-Vázquez et al. 2013). We aimed to test whether altering tRNA levels also affects the epigenetic function of Elp1.

We cloned the SPBTRNALYS.06 and SPCTRNALYS.11 genes, which encode for tRNA^Lys_UUU and tRNA^Lys_CUU, respectively, along with 500 bp of their upstream and downstream regions (to include promoter and terminator elements), into high-copy yeast plasmids that contained the NatMX selectable marker gene. Overexpression of tRNA^Lys_UUU enhanced the colony sizes of m203 elp1Δ cells compared to vector-only and tRNA^Lys_CUU overexpression controls (Supplementary Fig. 6), agreeing with previous findings that tRNA^Lys_UUU overexpression can rescue growth defects in S. cerevisiae and S. pombe Elongator mutant cells (Esberg et al. 2006; Villahermosa and Fleck 2017). We next tested whether tRNA^Lys_UUU overexpression in m203 elp1Δ cells would affect yeast growth on 5-FOA media. We found that tRNA^Lys_UUU overexpression re-sensitized m203 elp1Δ cells to 5-FOA (Fig. 7a). In contrast, the empty vector or overexpression of tRNA^Lys_CUU showed far less suppressive effects (Fig. 7a). By H3K9me3 ChIP analyses, we also found that overexpression of tRNA^Lys_UUU, but not tRNA^Lys_CUU, in m203 elp1Δ cells dramatically reduced H3K9me3 at the otr1R::ura4⁺ locus (Fig. 7b). Taken together, these data indicate that tRNA^Lys_UUU overexpression can suppress the epigenetic function of Elp1.

Fig. 7.

Overexpression of tRNA^Lys_UUU in m203 elp1Δ cells reverts the otr1R::ura4⁺ reporter back to the expressed state. a) Spotting assays showing 5-FOA sensitivity of the indicated yeast strains, which carry empty vector or expression plasmids with tRNA^Lys_CUU (SPCTRNALYS.11) or tRNA^Lys_UUU (SPBTRNALYS.06). All yeast strains were grown on media containing nourseothricin antibiotic to select for cells with the plasmids. b) ChIP enrichment of heterochromatic H3K9me3 at otr1R::ura4⁺ in m203 cells (with wild-type elp1⁺ or elp1Δ) that carried tRNA-non-expressing or -expressing plasmid. c) Peak area of detected mcm⁵s²U in the total tRNAs that were derived from WT or elp1Δ yeast strains. Depicted are the peak area intensities from two biological replicates, per yeast strain (denoted as rep 1 or 2). ChIP enrichments were normalized relative to the euchromatic leu1 gene and a 5% input control. Error bars indicate standard deviation based on n ≥ 2. Asterix (*) denotes statistical significance as determined by Student's t-test.

We wondered whether elp1Δ cells could regain mcm⁵s² tRNA modifications when they carried the tRNA^Lys_UUU plasmid. After isolating total tRNAs and subjecting them to mass spectrometry analyses, we failed to detect mcm⁵s² tRNA modifications in cells lacking elp1 (Fig. 7c), indicating that the overexpressed tRNAs do not have the modification. Our finding agrees with a previous study in S. cerevisiae that overexpressed tRNAs in Elongator mutant cells failed to regain mcm⁵s² (Esberg et al. 2006). We conclude that Elp1-dependent loss of heterochromatin in m203 cells is influenced by tRNA^Lys_UUU but not by the mcm⁵s² modification, per se.

Discussion

Our study reveals that the highly conserved Elp1 protein possesses a novel function in regulating RNAi-dependent heterochromatin. This epigenetic role is independent of the mcm⁵s² tRNA modification. Furthermore, Elp3 and Elp5 are largely dispensable when the elp1 gene is intact. However, Elp3 is genetically involved when Elp1 is absent. These observations differ from the current model of how Elp1/3/5 cooperate to ensure proper tRNA modifications (Huang et al. 2005; Karlsborn et al. 2014b; Dauden et al. 2019; Jaciuk et al. 2023; Abbassi et al. 2024) and microtubule stability (Planelles-Herrero et al. 2022, 2025). We speculate that Elp3 may have cryptic epigenetic functions when Elp1 is lacking in cells. This work expands our understanding of Elp1 functional capabilities beyond the current paradigms of tRNA modifications (Dauden et al. 2019) and microtubule assembly (Planelles-Herrero et al. 2022, 2025).

In this study, an epigenetic role for Elp1 was revealed in the context of a tester S. pombe strain called m203. Because this unique strain disrupts the RNAi pathway while maintaining transcription at the heterochromatin reporter locus otr1R::ura4⁺ (Kato et al. 2005), it provides an elegant system to identify novel factors that mediate RNAi disruption. Interestingly, loss of elp1 not only restored heterochromatin in m203 cells but also in rpb2-N44F cells (Supplementary Fig. 2b), suggesting that there might be several contexts where Elp1 could have an epigenetic role. Supporting this idea, a recent genome-wide study reported that Elp1 can regulate heterochromatic repression at subtelomeres and the mating-type locus (Muhammad et al. 2024). Furthermore, we found that elp1 could regulate a subset of native RNAi-dependent heterochromatin domains known as HOODs (Yamanaka et al. 2013) (Fig. 3a). The fact that not all HOODs equally respond to the loss of elp1 suggests that distinct mechanisms may regulate HOODs across the genome (Yamanaka et al. 2013).

Intriguingly, the other Elongator complex subunits Elp2/3/4 have been reported to promote RNAi-dependent heterochromatin (Muhammad et al. 2024). Additionally, an earlier study also reported that Elp3 promotes heterochromatin at the RNAi-dependent otr1R::ura4⁺ reporter (Bauer et al. 2012). It was suggested that Elp3 regulates heterochromatin by facilitating the translation of RNAi factors, including Ago1 and Rdp1, via its mcm⁵s² tRNA modification function (Bauer et al. 2012). In our study, we did notice that elp3Δ and elp5Δ single mutants had mild defects on H3K9me3 at the pericentromeric otr1R::ura4⁺ reporter (Fig. 5a and b). This partial defect could be due to a global reduction in RNAi protein abundance because S. pombe cells lacking Elp3 or Ctu1 showed nearly 50% reduction in Ago1 and Rdp1 protein levels (Bauer et al. 2012). Nonetheless, we found that the heterochromatic rescue phenotype of the elp1Δ mutant was more prominent compared to the single deletion of elp3, suggesting that Elp1 may regulate RNAi. How Elp1 regulates RNAi may depend on Rpb2, since the enhancement of pericentromeric dh siRNAs by elp1 deletion was more prominent in m203 background cells compared to wild-type (Fig. 4a). As a gain-of-function mutant, m203 might evoke an Elp1-dependent mechanism to antagonize pericentromeric RNAi. Such a mechanism would not be observed in wild-type cells under standard, rich media laboratory conditions. Our results reveal a novel genetic relationship between Elp1 and RNAi that requires further exploration in future studies that employ separation-of-function constructs, such as our elp3-CC mutant.

Overexpression of tRNA_UUU disrupts heterochromatin in m203 elp1Δ cells, but inhibition of mcm⁵s² tRNA modifications only had minor effects (Figs. 5a and 6c). This suggests that the epigenetic function of Elp1 may involve tRNA levels and/or function. It was previously shown in S. cerevisiae that deleting elp3 does not affect tRNA levels (Johansson et al. 2008), but it remains unclear whether elp1 deletion would have the same effect, especially in the m203 S. pombe background. A related model is that the m203 variant of Pol II may compete with tRNA_UUU for binding with Elp3. Elp1 normally promotes Elp3 binding with tRNAs (Dauden et al. 2019). Elp3 may bind with m203 Pol II (Otero et al. 1999), in the absence of Elp1, to promote transcription that promotes RNAi-mediated heterochromatin. However, overexpressed tRNA_UUU may outcompete m203 Pol II, thus reducing the transcription-based mechanism. Alternatively, Elp1 can regulate heterochromatin through Elongator-dependent and -independent mechanisms. We found that the overexpressed tRNA_UUU is not mcm⁵s²-modified in Elongator mutants (Fig. 7d). High levels of hypomodified tRNA_UUU help correct defects in translational decoding of AAA_Lys and AAG_Lys codons (Nedialkova and Leidel 2015). However, tRNA overexpression could also lead to widespread reprogramming within cells in poorly understood ways (Pavon-Eternod et al. 2013). It is possible that tRNA_UUU overexpression perturbs our heterochromatin system through pleiotropic mechanisms. Future exploration by overexpression of different tRNAs may provide clues into how Elp1 performs its epigenetic role.

The mechanism of Elp1-dependent siRNA regulation may involve serine-7 (Ser7) of the Pol II CTD, which was previously implicated in RNAi-dependent heterochromatin formation (Kajitani et al. 2017). By promoting the interaction of Pol II with RNA-induced transcriptional silencing (RITS) complex and of nascent RNA-chromatin associations, Ser7 promotes localized siRNA generation and heterochromatin establishment (Hall et al. 2002; Volpe et al. 2002). Interestingly, m203 shows similar heterochromatic silencing defects as an rpb1-ctd(S7A) mutant in which Ser7 is mutated into alanine (Kajitani et al. 2017). Also, Elp1 was reported to directly interact with the elongation form of Pol II from S. cerevisiae and human cells (Otero et al. 1999; Kim et al. 2002). It is conceivable that m203 disrupts Ser7-mediated heterochromatin formation by promoting Elp1-dependent Pol II elongation that leads to the release of the nascent RNAs from chromatin, thereby reducing the ability of RITS to generate siRNAs. A similar mechanism has been proposed for transcription elongation factor Paf1C (Kowalik et al. 2015). Alternatively, Elp1 may directly compete with the RITS complex for binding with Pol II.

It is unclear where Elp1 is positioned within the heterochromatin assembly pathway. We demonstrated that elp1 deletion can restore heterochromatin at the otr1R::ura4⁺ reporter in m203 cells if the H3K9 methyltransferase Clr4 and the RNAi proteins Dcr1, Rdp1, and Ago1 remain intact (Figs. 2e and 4b). Hence, Elp1 is likely a regulator of heterochromatin or RNAi factors. It is challenging to determine how Elp1 antagonizes heterochromatin in m203 cells because the RNAi-dependent heterochromatin pathway involves a self-reinforcing loop whereby RNAi promotes H3K9 methylation and the recruitment heterochromatin factors (eg Swi6/HP1), which promotes heterochromatin-localized RNAi (Noma et al. 2004; Sugiyama et al. 2005; Hayashi et al. 2012; Rougemaille et al. 2012; Yu et al. 2018). Based on previous reports of the diverse roles that Elongator may perform, it is possible that Elp1 regulates an early step preceding siRNA biogenesis, such as transcription (Otero et al. 1999) or pre-siRNA processing (Lipardi and Paterson 2010). Alternatively, Elp1 knockout may restore heterochromatin by promoting the putative in vivo histone acetyltransferase activity of Elp3 (Wittschieben et al. 2000; Winkler et al. 2002). Such histone acetylation could promote histone turnover (Aygün et al. 2013; Sahu et al. 2024) and enhance transcription of pre-siRNAs, akin to the proposed mechanism of the anti-silencing factor Epe1 recruiting the histone acetyltransferase complex, SAGA (Trewick et al. 2007; Bao et al. 2019).

The homolog of S. pombe Elp1 in humans is ELP1 (previously known as IKAP). Intronic mutations that prevent proper splicing of the ELP1 mRNA cause Riley-Day syndrome, a recessive autosomal condition characterized by progressive neurodevelopmental and neurodegenerative aberrations (Anderson et al. 2001; Slaugenhaupt et al. 2001). Although it is clear that ELP1 is critically important for neuronal health, the molecular basis behind how human ELP1 functions within neurons remains elusive. It is known that mis-splicing of ELP1 reduces protein expression and lowers mcm⁵s² tRNA modifications (Karlsborn et al. 2014a), however, it is unclear the extent to which these defects are pathogenic. Furthermore, there are no reports of mutations in the other Elongator subunits (ELP2/3/4/5/6) that cause Riley-Day syndrome, suggesting that neuronal ELP1 might be important for Elongator-independent processes or that subtleties in the process of Elongator-mediated tRNA modification help distinguish Elongator-associated diseases and disorders (Gaik et al. 2023). Future molecular studies using yeast and animal models will be important for investigating these roles of Elp1^ELP1 and their impact on human health.

Data availability

The strains and plasmids that were generated for this study are available from the corresponding author. The raw small RNA-seq data that support the findings of this study are openly available at the European Nucleotide Archive (ENA) at project accession number PRJEB102217.

Supplemental material available at GENETICS online.

Funding

The work was supported by the U.S. National Science Foundation under Grant No. 2422223 to T.V.Vo. and start-up funds from Michigan State University to T.V.Vo.

Author contributions

T.V.Vo. and M.B.N. conceived and designed the project. M.B.N., M.E.P., C.T.L., J.M.F., K.S.D., and T.V.Vo. performed experiments. M.B.N. and T.V.Vo. performed most of the experiments and their analyses. M.B.N. performed the phylogenetic analyses. M.B.N. prepared raw figures and tables, and wrote a draft of the Materials and methods. T.V.Vo. prepared the remainder of the manuscript with input from all authors.