Genome-wide Regulation of Pol II, FACT, and Spt6 Occupancies by RSC in Saccharomyces cerevisiae

Abstract

RSC (remodels the structure of chromatin) is an essential ATP-dependent chromatin remodeling complex in Saccharomyces cerevisiae. RSC utilizes its ATPase subunit, Sth1, to slide or remove nucleosomes. RSC has been shown to regulate the width of the nucleosome-depleted regions (NDRs) by sliding the flanking nucleosomes away from NDRs. As such, when RSC is depleted, nucleosomes encroach NDRs, leading to transcription initiation defects. In this study, we examined the effects of the catalytic-dead Sth1 on transcription and compared them to those observed during acute and rapid Sth1 depletion by auxin-induced degron strategy. We found that rapid depletion of Sth1 reduces recruitment of TBP and Pol II in highly transcribed genes, as would be expected considering its role in regulating chromatin structure at promoters. In contrast, cells harboring the catalytic-dead Sth1 (sth1-K501R) exhibited a severe reduction in TBP binding, but, surprisingly, also displayed a substantial accumulation in Pol II occupancies within coding regions. The Pol II occupancies further increased upon depleting endogenous Sth1 in the catalytic-dead mutant, suggesting that the inactive Sth1 contributes to Pol II accumulation in coding regions. Notwithstanding the Pol II increase, the ORF occupancies of histone chaperones, FACT and Spt6 were significantly reduced in the mutant. These results suggest a potential role for RSC in recruiting/retaining these chaperones in coding regions. Pol II accumulation despite substantial reductions in TBP, FACT, and Spt6 occupancies in the catalytic-dead mutant could indicate severe transcription elongation and termination defects. Such defects would be consistent with studies showing that RSC is recruited to coding regions in a transcription-dependent manner. Thus, these findings imply a role for RSC in transcription elongation and termination processes, in addition to its established role in transcription initiation.

INTRODUCTION

DNA in eukaryotes is packaged into chromatin using nucleosomes, which are formed by wrapping 147 base pairs of DNA around a histone octamer consisting of two copies each of histones H3, H4, H2A, and H2B (Luger, 2003). Nucleosomes are potent barriers and have been shown to prevent transcription initiation in vitro (Lorch et al., 1987). Furthermore, nucleosomes impede RNA polymerase II (Pol II) from accessing DNA during transcription elongation and termination (Govind et al., 2010; Formosa and Winston, 2020; Kujirai and Kurumizaka, 2020; Noe Gonzalez et al., 2021). The nucleosome-induced impediment can be relieved by ATP-dependent chromatin remodeling complexes (Clapier et al., 2017).

The remodels the structure of chromatin (RSC) complex is a member of the SWI/SNF family of ATP-dependent chromatin remodelers and is the only essential remodeler in budding yeast (Cairns et al., 1996). It has an ATPase subunit known as Sth1, which is essential for viability in Saccharomyces cerevisiae. The mutant form of Sth1 (Sth1-K501R) lacks ATPase activity (Du et al., 1998). RSC binds to the −1 and +1 nucleosomes flanking the nucleosome-depleted regions (NDRs), which are typically situated upstream of transcription start sites (TSSs) (Lee et al., 2007; Xu et al., 2009; Yen et al., 2012; Brahma and Henikoff, 2019). It plays an important role in maintaining NDRs width by evicting or sliding −1 and +1 nucleosomes away from NDRs (Hartley and Madhani, 2009; Yen et al., 2012; Rawal et al., 2018). Deficiencies in RSC function result in a narrowing of NDRs, caused by the inward movement of the −1 and +1 nucleosomes, which subsequently leads to reduced TBP binding at promoters (Kubik et al., 2018).

In addition to binding nucleosomes in promoters, the RSC complex also localizes to coding regions, which appears to be dependent on the transcriptional activity (Ganguli et al., 2014; Spain et al., 2014; Vinayachandran et al., 2018; Biernat et al., 2021). This observation has led to the idea that RSC could also play a role in Pol II elongation by remodeling nucleosomes within the coding sequences (CDSs) (Spain and Govind, 2011). In support of this idea, we have shown that RSC binds to the nucleosomes in the CDSs (Biernat et al., 2021). Remarkably, our data also revealed that the nucleosomes bound by RSC within transcribed CDSs exhibit greater sensitivity to MNase digestion than other non-RSC-bound nucleosomes in the same regions. These observations support the possibility of RSC contributing to increasing accessibility of nucleosomes within open reading frames (ORFs). Furthermore, we also noticed increased Pol II occupancy in genes displaying exceptionally high transcription levels when the Sth1 subunit of the RSC complex is depleted from the nucleus (Kubik et al., 2018; Biernat et al., 2021). Remarkably, the genes exhibiting increased Pol II occupancies upon depletion also displayed diminished TBP binding within their promoters. The increased Pol II levels in the coding regions could potentially be attributed to transcription elongation defects resulting from the inefficient remodeling of nucleosomes within ORFs. However, the effects of RSC depletion and the resulting buildup of Pol II on other factors involved in transcription elongation remain unclear.

Histone chaperones FACT and Spt6 are cotranscriptionally recruited to coding regions (Pathak et al., 1988; Mason and Struhl, 2003; Yoh et al., 2008; Burugula et al., 2014; Martin et al., 2018; Formosa and Winston, 2020). FACT is a histone H2A-H2B chaperone composed of Spt16 and Pob3 (Formosa and Winston, 2020). FACT recruitment to the coding regions depends on the active transcription (Mason and Struhl, 2003). More recently, it was shown that FACT is recruited to partially unwrapped +1 nucleosomes by the chromatin remodeler Chd1 (Jeronimo et al., 2021). Furthermore, it is worth noting that histone modifications, particularly histone acetylation, have also been implicated in FACT-nucleosome interactions in vitro and in its recruitment to a few genes in vivo. (Stuwe et al., 2008; Pathak et al., 2018). Spt6 is an H3 chaperone that interacts with Pol II, and its recruitment is promoted by the phosphorylation of residues in the Pol II CTD (carboxy-terminal domain) and linker regions (Burugula et al., 2014; Sdano et al., 2017). Both FACT and Spt6 aid in the cotranscriptional reassembly of histones in the wake of Pol II elongation and are also needed for suppressing aberrant transcription within the coding sequences (Kaplan et al., 2003; Carrozza et al., 2005; Govind et al., 2010).

In this study, we determined the occupancies of Pol II, TBP, FACT, and Spt6 in the catalytically dead RSC (sth1-K501R) mutants and compared it to cells with acute RSC (Sth1) depletion (Du et al., 1998; Muñoz et al., 2019). We observed more significant reductions in TBP binding at the promoters of highly transcribed genes in cells with the inactive RSC compared to the Sth1-depleted cells, indicating more severe transcriptional defects in catalytic-dead mutants. In contrast to reduced TBP binding, we found a substantial accumulation of Pol II in coding regions throughout the genome, in cells containing catalytic-deadSth1. Interestingly, we also observed notable reductions in the occupancies of Spt16 and Spt6 in these cells. These reductions occurred despite increased Pol II occupancies, suggesting that RSC may help in recruiting or retaining Spt16 and Spt6 over transcribed regions. Overall, our study indicates that RSC plays a role in regulating multiple stages of transcription, including transcription initiation.

METHODS

Yeast Strains and Growth Conditions

All Saccharomyces cerevisiae strains used in this study were kindly provided by Dr. Frank Uhlmann of the Francis Crick Institute in London and are described in (Munoz et al., 2019). The experiments were performed using at least two independent cultures. The Sth1-AID strain (MATa trp1Δ1 can1Δ100 leu2Δ3-112 his3Δ11-15 ura3Δ0 GAL psi+ pADH1-OsTIR1-9myc::ADE2 URA2::pMET3::STH1-IAA17::KanMX6) and Sth1-AID/sth1-KR (contains catalytic-dead copy of Sth1; MATa trp1Δ can1Δ100 leu2Δ3-112 his3Δ11-15 ura3Δ0 GAL psi+ pADH1-OsTIR1-9myc::ADE2 URA2::pMET3::STH1-IAA17::KanMX6 STH1K501R-PK3::TRP1), or WT BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) cells were grown in Synthetic Complete media deficient for methionine (SC/Met-) media to an A₆₀₀ of 0.7 before being treated with methionine and 3-indole acetic acid (3-IAA, auxin) to repress transcription of the endogenous Sth1 copy and to deplete any endogenous Sth1 present due to leaky expression via auxin-induced degradation as described previously (Nishimura et al., 2009; Morawska and Ulrich, 2013; Munoz et al., 2019). Briefly, methionine and 3-IAA were added to final concentrations of 250 μM and 500 μM, respectively, and cultures were incubated at 30°C for 1 hour before crosslinking and harvesting. Spike-in was performed by growing Schizosaccharomyces pombe cells to an A₆₀₀ of 10, then adding S. pombe cells to the S. cerevisiae cultures so that the final A₆₀₀ of S. pombe added would be 10% of the A₆₀₀ of S. cerevisiae. Spike-in was performed after cell growth but prior to crosslinking and harvesting. Crosslinking was performed as described previously: crosslinking solution (50 mM HEPES-KOH [pH 7.5], 1 mM EDTA, 100 mM NaCl, 11% formaldehyde) was added to cultures to a final concentration of 1% formaldehyde, and the cultures were swirled to mix. Cultures were then incubated at room temperature with intermittent shaking for fifteen minutes, and cells were incubated for 5 minutes after adding glycine solution to a final concentration of 350 mM to quench crosslinking. Cells were then collected via centrifugation at 4000 rpm at 4°C for 5 mins, then washed twice with chilled 1X Tris-Buffered Saline solution before being stored at −70°C until ready for sonication.

Sonication of Crosslinked Chromatin

Crosslinked cell pellets consisting of ^~100 ml of crosslinked cell cultures were thawed on ice and resuspended in 500 μl of FA lysis buffer (50 mM HEPES-KOH [pH 7.5], 1 mM EDTA, 140 mM NaCl, 1% Triton X-100, 0.1% sodium deoxycholate) containing protease inhibitors (Pepstatin, Leupeptin and PMSF, added to final concentrations of 0.729 mM, 1 μg/ml, and 100 mM, respectively). 500 μl of acid-washed glass beads was added to the resuspended cells and cells were disrupted for 45 mins in a 4°C cold room. The cell extracts were then collected by centrifugation at 1000 rpm at 4°C for 5 mins, then glass beads were washed with 500 μl FA lysis buffer and the washes were pooled with the collected cell extracts. The pooled extracts were then centrifuged at 4000 rpm at 4°C for 5 mins to collect the pellet containing the chromatin. The pooled extracts were then sonicated at 4°C using a Branson Sonifier set at 60% Duty Cycle and an Output Control of 2 to shear the chromatin into smaller fragments. The chromatin was sonicated for twelve one-minute cycles, where each cycle consisted of active sonication on ice for 30 seconds, followed by a 30 second rest period. The chromatin was collected by centrifugation. Sheared chromatin with fragment sizes ranging from 250 – 400 bp was used for chromatin immunoprecipitation.

Chromatin Immunoprecipitation (ChIP)

Chromatin immunoprecipitation was performed using a slightly modified protocol (Govind et al., 2012). Briefly, Dynabeads were washed twice with PBS containing 5 mg/ml BSA, resuspended in 150 μl PBS/BSA, and conjugated with the appropriate antibody for three hours at 4°C on a nutator. 50 μl of anti-mouse Dynabeads (Thermofisher Scientific, catalog # 11-041) and 3 μl of anti-Rpb3 antibody (Neoclone, catalog # W0012) per ChIP were used for Rpb3 (Pol II) ChIPs, whereas 40 μl of anti-rabbit Dynabeads (Thermofisher Scientific, catalog # 11-204-D) and 0.5 μl anti-TBP antibodies (gift from Dr. Joseph Reese, Penn State University) were used per ChIP for TBP ChIPs. For Spt16 and Spt6 ChIPs, 50 μl of anti-rabbit Dynabeads and 1 μl of either anti-Spt16 or anti-Spt6 antibodies (gift from Dr. Tim Formosa) were used per ChIP. After antibody conjugation, the beads were washed twice with PBS/BSA. 60 μl of PBS/BSA and 40 μl FA lysis buffer were added to the beads and either sonicated or MNase chromatin was added to the beads. For Spt16 or Spt6 ChIPs, 100 μl of chromatin per ChIP was used, 110 μl of chromatin per ChIP was used for Rpb3 ChIPs, and 130 μl of chromatin per ChIP was used for TBP ChIPs. The antibody-conjugated Dynabeads were then resuspended in the chromatin solution using a pipette and the conjugated beads were incubated with the chromatin on a nutator at 4°C for either 3.5 hours for Spt16, Spt6, and Rpb3 ChIPs, or overnight TBP ChIPs. After incubation, beads were washed once with PBS/BSA, then twice (once for Spt16, Spt6, or Rpb3 ChIPs) with the following buffers: FA lysis buffer, Wash Buffer II (50mM HEPES-KOH, 500 mM NaCl, 1 mM EDTA, 0.1% sodium deoxycholate, 1% Triton-X 100), and Wash Buffer III (10mM Tris-HCl [pH8.0], 250 mM LiCl, 1mM EDTA, 0.5% sodium deoxycholate, 0.5% NP-40 substitute). Beads were then washed once with Ultrapure Water before being resuspended in 100 μl of Elution Buffer I (50mM Tris-HCl, 10mM EDTA, 1% SDS) and incubated at 65°C for 15 mins. The eluate was collected and the beads were resuspended in 150 μl of Elution Buffer II (10mM Tris-HCl, 1mM EDTA, 0.67% SDS) before being incubated at 65°C for 10 mins to elute any remaining chromatin. The ChIP eluates were then combined and then incubated at 65°C for at least 4 hours (overnight for Spt16, Spt6 and Rpb3 ChIPs, 4 hours for TBP) to reverse the formaldehyde crosslinks. Proteinase K (Ambion, catalog # AM2546) was added to a final concentration of 0.4 mg/ml and the eluates were further incubated at 65°C for 2 hours to degrade any remaining protein. DNA was extracted using an equal volume of chloroform:isoamyl alcohol and ethanol-precipitated. DNA was then resuspended in 1X TE/RNase A and incubated at 37°C for 1 hour to degrade RNA. DNA was further purified using AMPure XP beads (Beckman Coulter, catalog # A63881) before library preparation.

Library Preparation for ChIP-seq

Library preparation was performed using the NEBNext Ultra II DNA Library Kit (New England Biolabs, catalog # E7645S) and a protocol modified from the manufacturer’s instructions for the kit. Briefly, 10 ng of ChIP DNA resuspended in 50 μl 0.1X TE was thawed on ice and then combined with 7 μl of End Prep Buffer and 3 μl End Prep enzyme before being incubated on the thermocycler according to the manufacturer’s instructions to clean up the ends of the DNA fragments for adapter ligation. 1 μl of diluted NEXTFlex DNA barcode (PerkinElmer, catalog # 514104, diluted to 0.24 μM using Ultrapure water) was then added to each prepped input or ChIP and the volume was made up to 62.5 μl using Ultrapure water. 30 μl of Ligation Master Mix and 1 μl of Ligation Enhancer were then added to the end-prepped DNA and the samples were incubated at 20°C for 15 mins to ligate the DNA barcodes to the end-prepped DNA. Bead cleanup of the adapter-ligated DNA was then performed using AMPure XP beads (Beckman Coulter, catalog # A63881), where the volume of beads added was 0.9X that of the input or ChIP volume to prevent size selection of the fragments during bead cleanup. The DNA was eluted in 22.5 μl 0.1X TE, then 2.5 μl of universal primers and 25 μl Q5 master mix were added to the DNA. The libraries were amplified according to the manufacturer instructions, where the ChIP libraries were amplified for 9-15 cycles. The amplified libraries underwent one more round of bead cleanup without size selection using AMPure XP beads before being eluted in 0.1X TE and loaded onto a 2% E-Gel EX (Thermofisher Scientific, catalog # G401002) and run according to the E-Gel manufacturer’s instructions. DNA for the ChIP and input libraries were then excised from the gel and the library DNA was extracted from the gel using the Qiaex II Gel Extraction Kit (Qiagen, catalog # 20051) according to the manufacturer’s instructions. Gel slices containing fragments ranging from ^~100 bp to 600 bp were excised for libraries from sonicated chromatin. The resulting library DNA samples were sequenced on the Illumina HiSeq platform in paired-end mode at GENEWIZ in South Plainfield, NJ, USA.

Chip-Seq Data Analysis

Sequences were trimmed to remove adapters using the Cutadapt package (parameter: -m 20) and were aligned to the S. cerevisiae genome (SacCer3) using Bowtie2 (parameters: -X 1000 –very-sensitive –no- mix –no-unal). Samtools was used to sort, index, and remove PCR duplicates from the resulting bam files. The average ORF occupancies for Pol II, Spt6, Spt16, and TBP occupancies at the TATA box or TATA-like sequences were calculated using modified scripts (Zheng et al., 2023). The TBP occupancies were normalized by count per million (CPM) +/− 100 bp centered on the TATA or TATA-like elements. Pol II, Spt16, and Spt6 occupancies were calculated using the spike-in normalization method. The spike-in factor was calculated by dividing the average of S. pombe reads for all the samples by the average S. pombe reads for the replicate samples. Metagene profiles and scatterplots were generated in JMP (https://www.jmp.com/en_us/software/predictive-analytics-software.html), whereas boxplots were generated using the BoxplotR website (http://shiny.chemgrid.org/boxplotr/).

RESULTS

TBP is Reduced Genome-Wide in Cells Depleted of Sth1

We had previously observed that Sth1 nuclear depletion through anchor away strategy resulted in increased Pol II occupancies in the ORFs of the top 10% of transcribed genes, by reanalyzing published data (Kubik et al., 2017; Biernat et al., 2021). Transcription regulation involves the coordinated action of multiple chromatin remodelers, including RSC. Under RSC-depleted conditions, other remodelers could remodel nucleosomes and compensate for the loss of RSC functions, which could mask or confound the direct effects of RSC depletion on transcription. Therefore, in this study, we compared the effects of the catalytic-dead RSC (STH1^K501R; sth1-KR henceforth, (Du et al., 1998)) to Sth1-depletion on transcription. Since sth1-KR cells are inviable, we used the strategy employed previously (Haruki et al., 2008; Muñoz et al., 2019). In this strategy, the endogenous Sth1 is tagged with AID (auxin-induced degron) and also harbors a copy of catalytic-deadsth1-KR in the genome. Thus, the impact of the catalytic-dead RSC can be studied by depleting the endogenous Sth1 after auxin treatment. The Sth1-AID cells were treated with auxin for one hour to achieve depletion. We noted that the WT cells and the Sth1-AID cells showed very similar levels of Sth1, as previously seen (Muñoz et al., 2019). One hour of auxin treatment resulted in a substantial loss of Sth1 from the Sth1-AID cells (Figure 1A).

Figure 1: The presence of catalytic-dead RSC severely reduces TBP binding.

A) Western blot showing Sth1 levels in WT cells and in Sth1-AID cells without auxin and with treatment with auxin for one hour. Sth1 was detected using anti-Sth1 antibodies, and Rpb3 was included as a loading control. The molecular weights (Mol. Wt.) are shown in kilodaltons (kDa).

B) Boxplots depicting average TBP occupancies +/− 100 bp of the TATA box for the Q1 genes in the untreated and auxin-treated Sth1-AID and Sth1-AID/sth1-KR cells. The reads are shown as reads/million (RPM). The p-values were calculated using the Welch student t-test.

C-D) Heatmaps depicting TBP binding at +/− 1000 bp from the center of the TBP binding sites for untreated and auxin-treated Sth1-AID (C) and Sth1-AID/sth1-KR (D) for the Q1 genes. Genes are centered at the TATA site and are sorted by decreasing TBP levels.

E-F) TBP binding profiles for Q1 (E) and D1 (F) are shown for the untreated and auxin-treated Sth1-AID and Sth1-AID/sth1-KR cells. The reads are shown as reads/million (RPM).

Next, we determined the TATA-binding protein (TBP) occupancies by ChIP-seq in cells depleted for Sth1. TBP is critical for preinitiation complex (PIC) assembly and thus serves as a measure of transcription initiation. We examined the impact of Sth1 deficiency on TBP occupancies in transcribed genes by analyzing the top 25% of genes with the highest Pol II occupancies in their coding regions (Q1-genes; based on the spiked-in normalized Pol II occupancies, see below). We sorted these genes based on the average TBP occupancies +/− 100 bp of the TATA box (or TATA-like elements) for each gene. Depleting Sth1 from the Sth1-AID cells only had a small but statistically significant reduction in the average TBP occupancies compared to untreated cells (Figure 1B and 1C; p-value=1.7 x 10⁻⁶). Metagene plots confirm that Sth1 deficient cells experience a slight decrease in TBP occupancies from promoters of the Q1-genes (Figure 1E; compare solid and dashed blue traces). A similar reduction was seen for the top 10% of genes with the highest Pol II occupancies (D1-genes) (Figure 1F). These findings align with previous studies using Sth1 anchor-away cells (Kubik et al., 2017).

Notably, the untreated Sth1-AID/sth1-KR cells showed TBP occupancies nearly identical to the auxin-treated Sth1-AID cells for both Q1 and D1 genes (Figures 1E and 1F; compare dotted blue and red traces). These results point to a potential PIC defect in the cells containing inactive Sth1 even in the presence of endogenous Sth1. Moreover, the depletion of endogenous Sth1 from the Sth1-AID/sth1-KR cells led to a striking reduction in TBP binding from Q1-genes (p-value =3.7 x 10⁻²⁶) and D1-genes (p-value =3.1 x 10⁻¹⁸) (Figure 1B and Figures 1E–1F). Thus, we observed more significant reductions in TBP binding for the catalytic-dead RSC cells than the Sth1-depleted cells. The TBP reduction in RSC mutants is consistent with the role of RSC in remodeling promoter nucleosomes (Ramachandran et al., 2017; Kubik et al., 2018; Biernat et al., 2021).

Catalytic-Dead RSC Increases Pol II in Coding Sequences

Given the robust TBP reductions in the catalytic-dead mutant, we sought to determine its impact on Pol II occupancies genome-wide by ChIP-seq using S. pombe cells as spiked-in control. Chromatin from both untreated and auxin-treated cells was subjected to Pol II ChIP-seq using antibodies against the core Pol II subunit Rpb3. In order to quantify Pol II occupancies, we calculated their spiked-in average occupancies for individual ORFs. For analyses reported below, we removed all genes with ORFs shorter than 300 bp. Since RSC also binds to nucleosomes in coding sequences, the changes in Pol II occupancies upon depleting RSC subunits might represent a combination of transcription initiation, elongation, and termination defects (Spain et al., 2014; Kubik et al., 2018; Ocampo et al., 2019; Biernat et al., 2021).

First, we compared Pol II occupancies in untreated Sth1-AID and Sth1/sth1-KR cells. Since untreated Sth1-AID/sth1-KR and auxin-treated Sth1-AID showed similar TBP occupancies, we expected to see lower Pol II occupancies in the sth1-KR-containing cells. However, contrary to our expectations, we observed an approximately 2.7-fold increase in Pol II ORF occupancies in the untreated Sth1-AID/sth1-KR cells compared to the untreated Sth1-AID cells, genome-wide (Figure 2A). While there was a slight reduction in average Pol II occupancies in 23 genes, the vast majority of genes showed increased Pol II in ORFs in cells containing Sth1-AID/sth1-KR. Thus, the presence of a catalytic-inactive RSC leads to Pol II accumulation in ORFs, genomewide.

Figure 2: Cells containing catalytic-dead RSC show increased Pol II occupancies in the ORFs.

A) Scatterplot showing spiked-in normalized average Rpb3 ChIP-seq reads in each ORF for the untreated Sth1-AID and Sth1-AID/sth1-KR. The diagonal line in black indicates that there is no variation in the Pol II ChIP-seq signal.

B) The profiles of Pol II occupancies at RPGs and at the top 25% (Q1) and the top 10% (D1) Pol II-occupied genes were generated for the indicated untreated cells.

Next, we wanted to determine whether the increase in Pol II occupancies in Sth1-AID/sth1-KR was linked to transcription levels. Therefore, we analyzed Pol II occupancies at highly transcribed genes: Q1-genes, D1-genes, and ribosomal protein genes (RPGs). All three subsets of transcribed genes showed higher Pol II occupancies throughout their ORFs in the untreated Sth1-AID/sth1-KR compared to Sth1-AID cells (Figure 2B; compare the dotted traces to solid traces). Interestingly, a significantly higher increase was seen in the D1-genes compared to the Q1-genes, suggesting that the increase in Pol II in ORFs might be liked to their transcription levels. Interestingly, while D1-genes and RPGs displayed similar Pol II levels in their ORFs in Sth1-AID cells, RPGs showed a smaller increase than the top decile. One possible explanation for this is that RPGs have a unique chromatin structure with fewer nucleosomes in ORFs, which could make them less sensitive to the presence of inactive Sth1. Additionally, variations in transcription initiation and regulation of promoter nucleosomes at RPGs may also play a role in the observed differences.

Pol II Occupancies Increased in the RSC Catalytic-Dead Mutant After Depleting Endogenous Sth1

We next examined the changes in Pol II occupancies after depleting Sth1 from Sth1-AID and Sth1-AID/sth1-KR cells. The results showed that depletion of Sth1 from Sth1-AID cells significantly reduced Pol II ORF occupancies compared to the untreated cells (p-value = 1.6 x 10⁻⁴⁵; N=5461) (Figure 3A). However, consistent with the results shown above (Figure 2), depletion of Sth1 from Sth1-AID/sth1-KR cells resulted in increased Pol II ORF occupancies (p-value = 1.4 x 10⁻¹⁴) (Figure 3A). Thus, depleting Sth1 in Sth1-AID cells reduces Pol II occupancies from ORFs and by contrast, the presence of inactive RSC (sth1-KR) increases Pol II ORF occupancies, regardless of the presence or absence of endogenous Sth1 (Figures 1 and 2).

Figure 3: Pol II occupancies are reduced in ORFs after depleting Sth1 from Sth1-AID, but are increased in catalytic-dead mutant.

A) Boxplots showing the average ORF Pol II occupancies for all genes (N=5461) in untreated and auxin-treated cells. Pol II decreased in auxin-treated Sth1-AID cells but increased in the treated Sth1-AID/sth1-KR cells. The p-values were calculated using the Welch student t-test.

B) Box plots showing Pol II occupancy at Q1-genes, D1-genes, and RPGs in untreated cells and auxin-treated Sth1-AID and Sth1-AID/sth1-KR. The p-values were calculated using the Welch student t-test.

C-D) The metagene profiles showing Pol II profiles at Q1-genes, D1-genes, and RPGs for auxin-treated and untreated Sth1-AID cells (C) and Sth1-AID/sth1-KR cells (D).

We next investigated the effects of Sth1 depletion on Q1-genes, D1-genes, and RPGs. The results show that the depletion of Sth1 in Sth1-AID caused significant reductions in Pol II occupancies for Q1-genes (p-value = 1.7 x 10⁻¹⁷), D1-genes (p-value = 1.8 x 10⁻⁹), and for RPGs (p-value = 1x 10⁻²) (Figure 3B). Thus, the reduced levels of RSC resulted in diminished Pol II occupancies across all three gene subsets analyzed. These significant reductions in Pol II occupancies from the ORFs were associated with the modest reductions in TBP binding upon depleting Sth1 for these genes (compare Figures 1E and 1F with Figures 3B and 3C). Moreover, the reduction in Pol II occupancy was observed uniformly throughout coding regions after Sth1 depletion (Figure 3C, compare solid traces). This is consistent with studies reporting higher RSC occupancies in ORFs of transcribed genes than in non-transcribed genes (Spain et al., 2014; Biernat et al., 2021).

In contrast to the Sth1 depletion in Sth1-AID cells, Sth1 depletion in Sth1-AID/sth1-KR background displayed significantly higher Pol II in the ORFs Q1-genes (p-value = 8.3x 10⁻⁵), D1-genes (p-value = 1 x 10⁻²), and RPGs (p-value = 2.9 x 10⁻¹¹) (Figure 3B). As expected, a greater increase was seen for D1-genes than those in Q1-genes, and the greatest increase was seen for the RPGs (Figure 3B and 3D). Since RSC localizes to both promoters and transcribed coding sequences, possible defects in transcription initiation, elongation, and termination could account for the Pol II reductions in Sth1-depleted cells. However, the contrasting effect of the catalytic-dead Sth1 vs. Sth1-depletion suggests that inactive RSC influences Pol II dynamics differently than the mere depletion of Sth1. We considered three possibilities to explain why inactive RSC leads to increased Pol II in coding regions of the catalytic-dead mutant. First, the two copies of RSC in the Sth1-AID/sth1-KR strain could have a greater contribution in stabilizing Pol II in ORFs than the single copy in the Sth1-AID cells, and hence higher Pol II occupancies were observed in the sth1-KR-containing strain. This explanation is supported by studies showing that RSC interacts with Pol II (Soutourina et al., 2006). However, if this was the primary cause for increased Pol II, then depletion of endogenous Sth1 from the Sth1-AID/sth1-KR cells should have reduced Pol II in ORFs compared to the levels seen in the undepleted cells. Instead, we saw an increase after depletion (Figures 3A, 3B, and 3D). Alternately, elongation defects in the Sth1-AID/sth1-KR could lead to an apparent increase in Pol II occupancy. We have shown that RSC-bound nucleosomes are more accessible to micrococcal nuclease (MNase) digestion than non-RSC-bound nucleosomes in transcribed regions (Biernat et al., 2021). Furthermore, nucleosomal accessibility is increased upon transcription induction and is reduced upon RSC depletion (Klein-Brill et al., 2019; Biernat et al., 2021). The presence of catalytic-inactive RSC in ORFs could hinder DNA access to elongating polymerases, causing a rise in Pol II in coding regions. Finally, RSC depletion (Rsc8 subunit) also leads to transcription termination defects; therefore, defective termination could increase Pol II level in ORFs (Ocampo et al., 2019). The observed increase in Pol II coupled with significant reductions in TBP (TATA-binding protein) levels in the catalytic inactive RSC mutant suggests the presence of transcriptional defects occurring at multiple steps.

Sth1 Depletion in The Sth1-AID/sth1-KR Mutant Leads to Severe Reductions in Spt16 And Spt6 Occupancies in ORFs

Considering that Pol II occupancies increased in the ORFs of the Sth1-AID/sth1-KR mutant upon Sth1 depletion, we investigated the impact of the catalytic mutant in the absence of the endogenous Sth1 on occupancies of elongation factors FACT and Spt6, which are cotranscriptionally recruited to transcribed ORFs and whose occupancies are linked to active transcription. Towards this end, we performed Spt16 and Spt6 Spiked-in ChIP-seq using the chromatin prepared from untreated and auxin-treated Sth1-AID/sth1-KR cells.

Spt16 and Spt6 occupancies in Sth1-AID/sth1-KR strongly correlated with the Pol II occupancies in the untreated Sth1-AID cells (Figure 4A). These data show that highly transcribed genes have higher levels of Spt16 and Spt6 in the untreated Sth1-AID/sth1-KR cells, similar to what is observed in previous studies (Burugula et al., 2014; Pathak et al., 2018). We next calculated the changes in Spt6 and Spt16 occupancies after depleting the endogenous Sth1 from Sth1-AID/sth1-KR cells. The ORF occupancies of both Spt16 and Spt6 were significantly reduced after Sth1 depletion (Fig. 4B). It was interesting to note that while Spt6 occupancies positively correlated with Pol II occupancies in the Sth1 undepleted cells (Figure 4A), the changes in Spt6 occupancies did not correlate with Pol II occupancies (ρ = 0.09; Figure 4B). In other words, while there were overall global reductions in Spt6 occupancies, this reduction was not greater at highly transcribed genes. The accumulating Pol II may offset the decrease in Spt6 occupancies since Spt6 is known to interact with elongating polymerases (Close et al., 2011; Burugula et al., 2014). In contrast, the changes in Spt16 occupancies negatively correlated with the Pol II occupancies (ρ = −0.34; Figure 4B).

Figure 4: RSC depletion in the catalytic-dead RSC mutant results in reduced Spt16 and Spt6 occupancies.

A) Scatterplot showing spiked-in Spt16 (blue dots) and Spt6 (red dots) ORF occupancies in untreated Sth1-AID/sth1-KR cells against Pol II occupancies in untreated Sth1-AID cells.

B) Scatterplot showing changes (Δ) in Spt16 and Spt6 occupancies after depletion of Sth1 in Sth1-AID/sth1-KR cells (y-axis) and Pol II occupancies in untreated Sth1-AID cells (x-axis).

C-D) Boxplots showing Spt16 in untreated and auxin-treated Sth1-AID/sth1-KR cells at Q1, D1 and RPGs (C). The metagene profiles for the same genes are shown (D). The p-values were calculated using the Welch student t-test.

E-F) Boxplots showing Spt6 in untreated and auxin-treated Sth1-AID/sth1-KR cells at Q1, D1 and RPGs (E). The metagene profiles for the same genes are shown (F). The p-values were calculated using the Welch student t-test.

We next examined their occupancies in the highly transcribed gene subsets. All three subsets (Q1, D1, and RPGs) showed severe reductions in Spt6 and Spt16 occupancies upon depleting the endogenous Sth1 (Figures 4C-4F). Both Q1 and RPGs showed very similar declines in Spt16 and Spt6 occupancies from their ORFs. The median changes in the Log2 Spt16 and Spt6 occupancies were very similar in all three gene sets after depleting Sth1 in the Sth1-AID/sth1-KR mutant (Figure 4C and 4E). The metagene profiles reveal that the most robust reductions in their occupancies occurred for the top decile of transcribed genes (Figures 4D and 4F), and the reductions were uniform across the coding regions. Considering that Spt16 and Spt6 occupancies in ORFs are linked to transcription, these data would suggest severe transcription defects in cells containing the inactive RSC complex. These reductions contrast the increase in Pol II occupancies we saw in the Sth1-AID/sth1-KR mutant even after depleting Sth1 (Figures 2 and 3). The reductions in Spt6 and Spt16 occupancies suggest a potential role for RSC in the recruitment/retention of these factors to transcribed regions. As discussed above, the reduction in the occupancies could also be due to impaired transcription in the catalytic-dead mutant. Nonetheless, our data show that the presence of catalytic inactive RSC has a contrasting impact on the occupancies of Pol II and elongation factors, Spt16 and Spt6.

Collectively, our findings provide compelling evidence that the inactivity of RSC results in a profound decrease in TBP binding at promoters, indicating a defective PIC assembly. Furthermore, our observations revealed decreased occupancies of FACT and Spt6 within the ORFs, which implies a potential involvement of RSC in the recruitment or retention of these factors over transcribed regions. Surprisingly, these reductions coincided with a substantial accumulation of Pol II in transcribed ORFs, suggesting potential impairments in transcription elongation and termination. Consequently, these data strongly suggest that RSC plays a crucial role in multiple stages of the transcription process.

DISCUSSION

In this study, we have examined the impact of depleting RSC ATPase Sth1 and the presence of catalytic inactive Sth1 on transcription. Depleting Sth1 led to a reduction in TBP binding at the promoters and Pol II within the ORFs. By contrast, the catalytic-dead mutant exhibited significantly higher Pol II occupancies despite significantly reducing TBP occupancies at promoters. The accumulation of Pol II within the ORFs in the mutant could be a result of widespread transcription elongation and termination defects. Consistent with this, we also saw severely reduced FACT and Spt6 occupancies in ORF in the mutant.

The reduction in TBP binding was more pronounced in the catalytic inactive mutant with the endogenous Sth1 compared to the Sth1-depleted condition (Figure 1). The weaker effect of Sth1 depletion on TBP binding compared to the catalytic inactive RSC suggests that depletion might be incomplete and that the residual Sth1 after auxin treatment was able to promote TBP binding at promoters more effectively than the inactive RSC. Nonetheless, the reduced TBP binding suggests a strong role for RSC in promoting transcription initiation, considering that TBP binding is one of the initial steps of the preinitiation complex assembly (Jackson-Fisher et al., 1999). Consistent with TBP results, depletion of Sth1 reduced Pol II occupancies genome-wide, with greater reductions seen in the highly transcribed genes. These findings are consistent with previous studies which found a significant relationship between alterations in TBP and Pol II levels in Sth1 anchor-away cells (Kubik et al., 2018). While the auxin-induced degradation depletes a particular protein/subunit, the anchor-away strategy could potentially shuttle the entire RSC complex from the nucleus. Therefore, the functional loss of one subunit vs. the entire complex, or a better nuclear depletion of Sth1 could account for the differences in TBP binding defects seen in this study vs the previous study (Kubik et al., 2018). Reduced TBP binding in cells with compromised RSC function aligned with its role in evicting histones from promoters and sliding +1 nucleosome downstream to expose TATA biding sites or TSSs (Hartley and Madhani, 2009; Kubik et al., 2018; Rawal et al., 2018; Klein-Brill et al., 2019).

Increased Pol II occupancies were observed in the first decile of the Pol II-occupied genes upon Sth1 anchor-away (Biernat et al., 2021). By contrast, the Pol II ChIP-seq using spiked-in control in this study reveals that Pol II occupancies are significantly higher across the genome, and not limited to a small subset of transcribed genes (Figure 2). Although higher Pol II occupancies are generally associated with higher transcription, it is counterintuitive that cells harboring catalytic inactive RSC have higher transcription than those without, especially considering that auxin-treated Sth1-AID/sth1-KR cells showed dramatic reductions in TBP binding. One plausible explanation for such an observation is that RSC may promote transcription elongation, in addition to transcription initiation, as proposed previously (Ganguli et al., 2014; Spain et al., 2014; Biernat et al., 2021). In support of this possibility, RSC is recruited to the ORFs of highly transcribed genes, potentially to remodel ORF nucleosomes (Spain et al., 2014; Vinayachandran et al., 2018; Biernat et al., 2021). Recently, we showed that the DNA in the RSC-bound nucleosomes in coding regions is more accessible since it is digested to a greater extent by micrococcal nuclease (MNase) than in the non-RSC-bound nucleosomes in the same region (Biernat et al., 2021), suggesting that RSC may contribute to increasing DNA accessibility to elongating RNA polymerases. Therefore, loss of RSC function could lead to an apparent increase in Pol II occupancies in ORFs.

An intriguing question arises as to why Pol II occupancies decrease upon Sth1 depletion and increase in the catalytic RSC mutant (Figure 2). One possibility is that in Sth1-deficient cells, the absence of RSC-mediated chromatin remodeling might trigger compensatory mechanisms involving other factors responsible for increasing chromatin accessibility. Notably, previous studies have indicated that SWI/SNF is localized within transcribed coding sequences (Dutta et al., 2017; Rawal et al., 2018). Chd1 is also recruited to ORFs and has been shown to promote FACT recruitment to the +1 nucleosomes by enhancing the chromatin accessibility (Tran et al., 2000; Simic et al., 2003; Warner et al., 2007; Ocampo et al., 2019; Jeronimo et al., 2021). Therefore, SWI/SNF, Chd1, or other yet unidentified factors compensate for the absence of Sth1 to regulate chromatin dynamics and help Pol II navigate the nucleosomes. Therefore, the lower Pol II occupancies in the Sth1-depleted cells likely arise from the dampened PIC assembly, as evidenced by lower TBP occupancies. By contrast, the presence of an inactive form of RSC within the ORFs might impede the recruitment of other factors that can remodel nucleosomes and otherwise compensate for the loss of RSC functions. Consequently, the lack of remodeling by other factors in the catalytic-dead mutant compromises transcriptional elongation, leading to Pol II accumulation in coding regions. In addition, Pol II accumulation can also be driven by transcription termination defects caused by the loss of RSC function, as shown previously (Ocampo et al., 2019). Thus, a combination of elongation and termination defects could explain the substantial Pol II accumulation in ORFs despite striking reductions in TBP binding in the catalytic inactive RSC mutant.

Spt16 and Spt6 interact with histones and Pol II (Bortvin and Winston, 1996; Mason and Struhl, 2003; Hondele et al., 2013; Sdano et al., 2017) and are known to be cotranscriptionally recruited (Mason and Struhl, 2003; Burugula et al., 2014; Pathak et al., 2018). Interestingly, we find that despite increased Pol II in Sth1-AID/sth1-KR cells, occupancies of both Spt16 and Spt6 are reduced upon auxin treatment (Figure 4). The disconnect between higher Pol II and lower FACT and Spt6 occupancies can be explained by the idea that the impaired RSC function could interfere with the recruitment or retention of FACT and Spt6 in ORFs. This disruption could result in a reduced ability of these factors to associate with the elongating Pol II complex, leading to their decreased occupancies. Another explanation could be that the altered chromatin structure in the absence of functional RSC affects the ability of FACT and Spt6 to interact with the chromatin, thereby diminishing their association with transcribed regions. Both Spt16 and Spt6 are shown to oscillate on nucleosomes suggesting that they dynamically interact with the chromatin, and are not continually associated with Pol II (Fischl et al., 2017). If Spt16 and Spt6 were continuously associated with elongating Pol II, we would have expected to see an increase in both factors along with Pol II in cells with catalytically inactive RSC.

FACT recruitment to the +1 nucleosome is enhanced by the chromatin remodeler Chd1 (Jeronimo et al., 2021). Although RSC is not directly implicated in FACT recruitment, RSC does remodel the +1 nucleosomes (Yen et al., 2012; Ramachandran and Henikoff, 2016; Ramachandran et al., 2017; Brahma and Henikoff, 2019; Biernat et al., 2021). Therefore, inefficient remodeling of the +1 nucleosome in cells with inactive RSC could contribute to the observed reductions in Spt16 occupancies. Similarly, Spt6 association with chromatin may also depend on the remodeling state of the ORF nucleosomes. In addition to histones, Spt6 is also associated with phosphorylated Pol II (Dengl et al., 2009; Mayer et al., 2010; Close et al., 2011; Burugula et al., 2014; Dronamraju and Strahl, 2014). Thus, it is possible that increased Pol II occupancies in the Sth1-AID/sth1-KR cells might help in retaining some Spt6 in ORFs and thereby masking true reduction in the catalytic-dead RSC mutant. Further investigation is required to elucidate the precise mechanisms underlying the observed changes in FACT and Spt6 occupancies in RSC mutants and their impact on transcriptional regulation.

• Catalytic-inactive RSC remodeling complex leads to transcription elongation defects

• RSC stimulates PIC assembly

• RSC promotes recruitment of histone chaperones FACT and Spt6 genome-wide

Abbreviation list:

3-IAA

3-Indoleacetic Acid, auxin

AID

Auxin-Inducible Degradation

bp

base pair

CDS

Coding Sequences

ChIP

Chromatin Immunoprecipitation

ChIP-seq

Chromatin Immunoprecipitation followed by Sequencing

MNase

Micrococcal Nuclease

NDR

Nucleosome-Depleted Region

RPG

Ribosomal Protein Gene

Sth1

Snf2 Homolog 1, catalytic subunit of RSC

TBP

TATA-Binding Protein

TSS

Transcription Start Site

RSC

Remodels Structure of Chromatin

DATA AVAILABILITY

The ChIP-seq data discussed in this manuscript has been deposited in NCBI’s Gene Expression Omnibus (GEO) under GSE245521