DNA-damage dependent interaction of Orc6 to SMARCA1 in S-phase modulates chromatin remodeling

Abstract

The human Origin Recognition Complex subunit 6 has recently garnered significant attention. The loss of human Orc6 doesn’t impact MCM loading in vivo, while in vitro reconstitution experiments have demonstrated that MCM loading can occur without Orc6, though could be stimulated upon the addition of Orc6. We recently reported an unexpected role for hOrc6, in promoting S-phase progression post pre-RC assembly and DNA damage response (DDR). Using CUT&RUN to map the binding of Orc6 genome-wide, we find that Orc6 is present at replication origins and at sites of DNA damage during S-phase. Upon encountering oxidative damage in S-phase, Orc6 interacts with the chromatin remodelers, SMARCA1/SNF2L, also a component of the replication fork. Loss of Orc6 shows enhanced genome-wide association of SMARCA1 to Orc6-bound sites, concomitant with chromatin remodeling at these sites supporting the model that Orc6 is a negative regulator of SMARCA1 and constrains its localization. Upon DNA damage, enhanced association of Orc6 to SMARCA1 prevents chromatin remodeling activity to halt replication progression. We propose that during S-phase, Orc6 is a barrier factor, which limits the activity of chromatin remodelers.

Graphical Abstract

Graphical Abstract.

Introduction

Coordinating DNA replication and repair within the chromatin environment is a complex but crucial process for preserving genome integrity [1]. At the replication fork, many repair factors actively prevent errors by activating mechanisms, including fork reversal, translesion synthesis, template switching, etc. [2]. Similarly, replication factors at the fork play active roles in the DNA damage response. For example, the DNA polymerase clamp, Proliferating Cell Nuclear Antigen (PCNA), plays a vital role in tolerating and repairing DNA damage by recruiting diverse factors to the site of damage [3]. Equally important is the process of nucleosome assembly during replication that is coupled with DNA synthesis [4]. ATP-dependent chromatin remodelers are essential for replication and repair, as they regulate nucleosome positioning to ensure accessibility for the replication and repair machinery [5]. Recent work has highlighted the intrinsic ability of the replication machinery in nucleosome remodeling [6].

The initiation of DNA replication requires the assembly of the Origin Recognition Complex (ORC) at replication origins in all eukaryotes during the G1 phase [7]. ORC, consisting of six subunits Orc1-6, initiates the formation of the pre-replicative complex essential for licensing replication origins [7, 8]. Multiple pre-RC factors, including various ORC subunits, are phosphorylated in an ATM/ATR-dependent manner, likely to inhibit DNA replication when cells encounter DNA damage [9]. The smallest subunit of ORC, Orc6, is unique because of its varying roles in different organisms [10–14]. While Orc6 is required for replication licensing, the process of minichromosome maintenance (MCM) proteins' loading in yeast and Drosophila, recent work from our lab has shown the dispensability of Orc6 in licensing in human cell lines [15]. In metazoans, Orc6 also regulates cytokinesis [10–12]. Recent in vitro reconstitution work from several laboratories supported our initial observation, that unlike in yeast, the hOrc6 subunit is not essential for the initial MCM recruitment as the double hexamer assembles [16, 17]. The human Orc6, however, seems to contribute to an alternative human double hexamer assembly pathway [16, 17].

While we continue to understand the subtle nuances of Orc6’s role in MCM loading, our lab has established that hOrc6 is required for S-phase progression and DNA damage response [9, 15]. We reported that hOrc6 localized to the replication fork and is an integral component of the mismatch repair complex [15]. Further, hOrc6 was found to be phosphorylated upon DNA damage at Thr229 (T229), specifically during S-phase, and this was required for fork pausing and halting replication progression until the damage is rectified [9]. The hOrc6 was also phosphorylated at Thr195 (T195) during mitosis, which seemed critical for modulating ribosome biogenesis [18]. Yet, the molecular details of Orc6 at the fork during S-phase and DNA damage remain to be elucidated.

In this work, by performing genome-wide binding analyses of Orc6 during S-phase and upon oxidative damage, we observe that Orc6 is enriched at open chromatin regions, replication origins, and sites of DNA damage during the S-phase. Upon encountering oxidative damage in S-phase, Orc6 associates with ATP-dependent chromatin remodelers, including but not limited to SMARCA1/SNF2L, also a component of the replication fork [19, 20]. Chromatin remodelers like SMARCA1, belonging to the Imitation SWItch (ISWI) family, facilitate replication by modulating nucleosome positioning, enabling the replication apparatus to assemble and prevent fork collapse [19]. Loss of this factor also induces DNA damage, though the functional relevance remains to be elucidated.

Yeast ORC (Orc1 subunit) has recently been implicated as a chromatin remodeler that creates a chromatin environment conducive for origin activity [6]. We find that the loss of hOrc6 leads to enhanced genome-wide association of SMARCA1 to previously Orc6-bound sites, concomitant with chromatin remodeling, and increased accessibility at these sites. Our results suggest that upon DNA damage, Orc6 prevents extensive chromatin remodeling by SMARCA1 to halt replication progression. Our results provide novel mechanistic insights on how Orc6 binding to SMARCA1 modulates chromatin remodeling.

Materials and Methods

Cell culture, transfection, DNA damage induction, and cell-cycle synchronization

U2OS cell line was cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with high glucose and 5% fetal bovine serum (FBS). All plasmid DNAs were delivered by Lipofectamine 3000 (Invitrogen). All small interfering RNAs (siRNAs) were delivered by Lipofectamine RNAimax (Invitrogen). siRNAs targeting Orc6 has the following sequences: #1 5′-AGAUAGAUAAACGGAAUUGGAGCCA (sense), 5′-UGGCUCCAAUUCCGUUUAUCUAUCUAU (antisense); #2 5′-UGUACAGCCUAAGUUAAUAAAUGTT (sense), 5′-AACAUUUAUUAACUUAGGCUGUACAAU (antisense). Unless otherwise mentioned, sequence #1 was used. siRNA sequence targeting Orc2 is available in the reference [21]. Double thymidine block and release was used for G1/S-, S-, and G2-phase synchronization. G1-phase cells were acquired by Nocodazole synchronization at metaphase followed by 6h release.

Immunoblotting and immunoprecipitation

Immunoblot procedures began with SDS–PAGE of protein samples followed by transferring to polyvinylidene fluoride (PVDF) membranes. Membrane surfaces with proteins were then incubated with milk/BSA containing primary antibodies at 4°C overnight: Orc6 (1:400, Dr. Bruce Stillman lab pAb982), Orc2 (1:500, Dr. Bruce Stillman lab pAb 205), Orc1 (1:100, pAb S2518), Orc6-pT229 (1:200, pAb 1122), SMARCA1 (1:500, Cell Signaling #9450S), T7 (1:5000, Sigma–Aldrich #69 522), actin (1:500, Santa Cruz #sc-47778), geminin (1:100, Santa Cruz #sc-13015), Cyclin D1 (1:300, Cell Signaling #2922S), GST (1:500, Santa Cruz #sc-138), tubulin (<1:5000, Sigma–Aldrich #T5168), and HA (1:500, Cold Spring Harbor Lab, 12CA5). Membranes were washed three times before incubating with HRP-conjugated secondary antibodies for 1 h at RT. After that, membranes were washed another three times and subject to chemical luminescence development under X-ray machine.

For immunoprecipitation, cells were collected, washed with PBS and lysed in IP lysis buffer (50 mM Tris pH 7.4, 300 mM NaCl, 1 mM MgCl₂, 10% glycerol, and 0.2% NP-40) containing protease inhibitors. Lysates were then sonicated and treated with benzonase nuclease (Millipore Sigma) for 30 min at room temperature, then EDTA was added to a final concentration of 2 mM to stop the benzonase reaction. Centrifugation was done at 15 000 rpm for 10 min to remove insoluble debris. Next, lysates were pre-cleared with GammaBind Plus Sepharose (Cytiva) for 30 min at 4°C. Antibodies were then added into lysates and incubated at 4°C overnight. Proteins bound by antibodies were pulled down by GammaBind Plus Sepharose for 2 h at 4°C. After incubation, beads were washed in wash buffer (same as lysis buffer, with 150 mM NaCl), then captured proteins were eluted with loading dye and analyzed with immunoblotting.

Chromatin fractionation

Cells were treated with 5 mM H₂O₂ and then lysed in a buffer containing 50 mM Tris–HCl (pH 7.4), 1 mM MgCl₂, 10% glycerol, 0.5% Triton X-100, and 2 mM EDTA, supplemented with protease and phosphatase inhibitors. The lysate was sonicated at high intensity for 3 min and then centrifuged at 15 000 rpm for 10 min at 4°C. The supernatant was collected as the cytoplasmic fraction, while the pellet was collected as the chromatin fraction. Samples were subjected to immunoblotting analysis.

Gel filtration

Cells were subjected to S-phase synchronization and H₂O₂ treatment prior to collection. Extraction of nuclei follows the procedure in [22]. Nuclear extracts were loaded on Superdex 200 Increase 10/300 GL column (GE Healthcare), which was pre-equilibrated by gel filtration buffer [25 mM HEPES, 150 mM NaCl, 10% glycerol, and 1 mM dithiothreitol (DTT), pH 7.5]. Fractions were collected, concentrated by TCA-precipitation procedure, and re-dissolved with loading dye, which were subjected to immunoblotting analysis.

CUT&RUN experiments

CUT&RUN was performed using a CUTANA ChIC/CUT&RUN kit (Epicypher, 14-1048) according to the manufacturer’s instructions. Briefly, 0.5 million cells were fixed in culture medium supplemented with 0.1% formaldehyde for 1 min before quenching with 0.125 M glycine. Cells were then permeabilized, bound to concanavalin A magnetic beads and incubated overnight at 4°C with primary antibodies using a 1:50 dilution. Primary antibodies used in this work included Orc6 [pAb 1123, targeting both non-phosphorylated Orc6 and Orc6 phosphorylated at T229; pAb 1121, targeting both non-phosphorylated Orc6 and Orc6 phosphorylated at T195 [9, 18]], SMARCA1 (Epicypher, #13-2005), OGG1 (Invitrogen, #PA1-31402), and non-targeting IgG control (Epicypher, #13-0042). After overnight incubation and washing to remove excess antibodies, protein A/G-MNase was added to the antibody-bound beads and activated with calcium at 4°C for 2 h to cleave target-bound genomic DNA. After quenching the cleavage reaction, the samples were reverse cross-linked under proteinase K treatment overnight at 55°C. CUT&RUN DNA was isolated with column chromatography, and libraries were prepared using an NEBNext Ultra II DNA Library Prep kit for Illumina (New England BioLabs). Libraries were sequenced on an Illumina NovaSeq 6000 with paired-end 150-base pair reads.

CUT&RUN analysis

Trimmed reads were aligned to reference genome UCSC hg38 using Bowtie2 with options: –local –very-sensitive-local –no-unal –no-mixed –no-discordant –phred33 -I 10. Also, reads were mapped to the Escherichia coli genome assembly to quantify E. coli spike-in DNA. Low quality reads (MAPQ < 10) were removed using SAMtools. SAMtools fixmate, positionsort, and markdup commands were used. BedGraph files were generated using Bedtools. Peak calling on individual replicates was performed using the normalized bedgraph files using following command SEACR_1.1.sh [target bedgraph] [control bedgraph] norm stringent [23]. Bedtools intersect (1 bp overlap) was used to find the high confidence overlapping peaks between two replicates. Signal tracks were generated using deepTools-bamCoverage, with parameter: –binSize 10 –normalizeUsing RPKM –effectiveGenomeSize hg38 -bl hg38.blacklist.bed –extendReads 200. Average signal files were generated from two replicates using bigWigAverage command. Intervene was used to generate Venn diagrams. Eulerr webtool was used to generate proportionate Venn diagrams (https://eunoia.bz/app/). Peak annotation was performed using ChIPseeker tool on Galaxy web server. Differential peaks were identified by DiffBind-DEseq2 using published methods [24], with an FDR < 0.05 and [Fold change] > 1.5 cutoff. Heatmaps and profile plots were generated using deepTools. Signal track was visualized in IGV.

Venn diagrams were generated using Intervene [25], which uses bedtools intersect to define overlaps at base-pair resolution. Peaks with any overlap are counted as shared, which may reduce the number of peaks in each set.

ATAC-seq and sequencing analysis

ATAC-seq sample processing and adaptor trimming were performed by the Roy J. Carver Biotechnology Center at the University of Illinois at Urbana-Champaign, while the sample processing procedures follow the protocol of commercial ATAC-seq kit (Active Motif, #53150). Trimmed reads were aligned to reference genome UCSC hg38 using bowtie2 with –very-sensitive setting and other default parameters. PCR duplicates were then removed using Picard. Low quality reads (MAPQ < 30) as well as reads from mitochondrial DNA (chrM) were removed using SAMtools. Signal tracks were generated using deepTools-bamCoverage, in which sequencing reads were normalized by RPKM and two replicates were averaged. Peaks were called from merged BAM files using Genrich ATAC-seq mode (-j -q 0.05). Differential peaks were identified by csaw-edgeR, with an FDR < 0.05 and (Fold change) > 1.5 cutoff [26]. Heatmaps and profile plots were generated in deepTools. Signal track was visualized in IGV.

Protein purification

GST-Flag-hOrc6 and GST were induced to overexpress in E. coli BL21 (DE3) strain. Cells were resuspended in GST resuspension buffer (25 mM HEPES, 0.1 mM EDTA, 150 mM NaCl, 1 mM DTT, and 5% glycerol, pH 7.9). Cells were lysed by supplementing with 0.5 mg/ml chicken white lysozyme, 0.1% Triton X-100, and 500 mM NaCl into the suspension, followed by sonication. The cell lysate was collected by 15 000 rpm centrifugation at 4°C for 30 min, filtered, and loaded onto GSTrap FF 1 ml column (GE Healthcare). After being washed with GST washing buffer (same as resuspension buffer but contains 500 mM NaCl) and GST resuspension buffer, the purified protein was eluted by GST elution buffer (resuspension buffer supplemented with 20 mM reduced glutathione, pH 7.9). The elution fractions containing GST-Flag-hOrc6 was concentrated with Amicon Ultra centrifugal filter units with 50 kDa cutoff (Millipore Sigma, 10 kDa cutoff for GST). The concentrated protein sample was subjected to size-exclusion chromatography on Superdex 75 10/300 GL column (GE Healthcare), which was pre-equilibrated by gel filtration buffer (25 mM HEPES, 150 mM NaCl, 10% glycerol, 1 mM DTT, pH 7.5). Fractions were collected and examined by SDS–PAGE, and fractions containing protein were pooled, aliquoted, and liquid-nitrogen-frozen for further use. The concentration of proteins was measured with protein BCA assay. The purified Flag-SMARCA1-C-strep protein is a gift from Dr. Andrea Cochran, and the purification procedure of SMARCA1 was done exactly as published [27].

In vitro immunoprecipitation assay

The Flag-SMARCA1-C-strep protein was purified as mentioned in the “Protein purification” section, while the gel filtration procedure was omitted. Eluted Flag-SMARCA1-C-strep was incubated with purified GST or GST-Flag-hOrc6 for 2 h at 4°C with rotation in IP buffer [20 mM HEPES pH 8, 100 mM NaCl, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and 0.1 mM DTT]. Then samples were co-incubated with Orc6 antibody (pAb 1123, used in CUT&RUN) at 4°C overnight. Protein–antibody complexes were pulled down using GammaBind Plus Sepharose (Cytiva) for 2 h at 4°C with rotation. Beads were washed in IP buffer and proteins were eluted by boiling in loading dye. The resulting samples were analyzed with immunoblotting.

In vitro nucleosome remodeling assay

The nucleosome sliding reaction was carried out at a final volume of 20 μl in the following buffer: 20 mM Tris pH 7.4, 50 mM KCl, 3 mM MgCl₂, 0.1 mg/ml BSA, 1 mM DTT, 50 U DpnII (NEB), and 5 mM ATP (100 mM stock solution, Sigma). The mononucleosome construct (EpiCypher) with a 147-bp Widom-601 nucleosome positioning sequence flanked by 4 bp upstream and 66 bp downstream (4-N-66) was used as the substrate. GST-Flag-Orc6 of indicated molarities (1.2, 2.4, and 3.6 pmol) were incubated with 0.4 pmol nucleosome on ice for 30 min before 1.2 pmol of Flag-SMARCA1-C-strep was added into indicated samples. DpnII and ATP were added at last to initiate the reaction. The remodeling reaction was carried out at 30°C for 30 min before stopped by 20 μl of quenching buffer (10 mM Tris, 0.6% SDS, 40 mM EDTA, 50 μg/ml Proteinase K) on 50°C for another 20 min. The resulting DNA contents were run on 8% native polyacrylamide gel in 1× TBE at room temperature at 150 V for about 3 h, stained by 0.2 μg/ml EtBr in water for 10 min, and visualized with Axygen Gel Documentation System.

Statistical analyses

Statistical analyses were conducted using Rstudio software. For comparation of the CUT&RUN and ATAC-seq profile curves , two-sided Komologrov–Smirnov test was used to generate the P-values. All other P-values were generated by two-sided non-parametric Wilcoxon rank sum test. All experiments were performed at least twice.

Data availability

All CUT&RUN sequencing data generated in this study have been deposited in the NIH GEO database under accession code GSE295487. All ATAC-seq sequencing data generated in this study are associated with the accession code GSE295485. NS-seq data and RNA-seq generated by Chetlangia et al. are deposited in GEO database under accession code GSE291923 and GSE291919, respectively.

Histone modification ChIP-seq data in K562 cells were retrieved from ENCODE with the following accession numbers: H3K4me1: GSM733692; H3K4me2: GSM733651; H3K4me3: GSM788087; H3K9ac: GSM788082; H3K27ac: GSM733656; H4K20me1: GSM733675; H3K9me3: GSM733776; H3K27me3: GSM733658; H3K36me3: GSM733714. Histone modification ChIP-seq data in U2OS cells were retrieved with the following accession numbers: H3K4me1: GSM8133822, GSM8133824, GSM8133826 (averaged); H3K4me3: GSM6600679, GSM6600681, GSM6600683 (averaged); H3K27ac: GSM7793922. ChIP-seq data of H2A.Z were retrieved from GEO database through GSE73742. G-quadraplex ChIP-seq data were retrieved from GEO database through GSE207567. DNase-seq data were retrieved from GEO database through GSE142160.

Results

Genome-wide distribution of Orc6 shows binding to replication origins during S-phase

The genome-wide binding of human Orc6 (hOrc6) has not been well elucidated. To map the genomic distribution of hOrc6 during the S-phase, we performed Cleavage Under Targets and Release Using Nuclease (CUT&RUN) in S-phase–synchronized U2OS cells with Orc6-antibody (Ab 1123, see Materials and methods for details) [28]. Orc6 peaks were identified for each replicate using the SEACR peak-calling algorithm [23], and only overlapping peaks between the two replicates were used for further analysis to ensure high confidence. By this approach, we identified 23 776 Orc6 binding sites. We confirmed the reproducibility of these sites by performing CUT&RUN using an independent Orc6 antibody (Ab1121) (Supplementary Fig. S1A). We next examined Orc6 enrichment at some known replication origins. Genome browser tracks revealed strong Orc6 binding at known replication origins, loci encoding MYC, MCM4, and TOP1, with prominent peaks observed upstream of each gene (Fig. 1A, top). In the HoxA cluster, we observed a high density of Orc6 peaks coinciding with regions of increased origin activity, as indicated by numerous small nascent strand sequencing (SNS-seq) peaks (Fig. 1A, bottom) [29]. To further quantify this association, we overlapped Orc6 peaks with SNS-seq data in U2OS cells (GSE291923) and found that approximately 26.9% (6393 peaks) of Orc6 binding sites co-localized with active replication origins (Fig. 1B). To characterize the genomic features of Orc6 binding sites, we annotated peaks using ChIPseeker [30]. We found that Orc6 peaks were enriched at promoter regions (>48%), as well as intronic (21%) and intergenic regions (22%) (Fig. 1C). Notably, among Orc6 peaks that overlapped with replication origins, 63% were located at promoters, compared to only 43% of non-origin-associated Orc6 peaks (Supplementary Fig. S1B). To assess whether Orc6 binding is associated with transcriptionally active chromatin, we plotted the average signal of Orc6 and H3K27ac, a marker of active transcription, around transcription start sites (TSSs) of the top and bottom 10% genes expressed in U2OS cells (GSE291919). As expected, H3K27ac showed strong enrichment near TSS regions. Similarly, Orc6 was highly enriched at TSSs of highly expressed genes (the top 10%), indicating that a substantial portion of Orc6 binding occurs within transcriptionally active chromatin (Fig. 1D). To further examine whether Orc6 binding sites are associated with specific histone modifications, we analyzed publicly available histone ChIP-seq datasets by plotting enrichment profiles centered on Orc6 peak summits. We included histone H3 and H4 modifications previously implicated in DNA replication, transcription, and chromatin accessibility [31–33]. Orc6 binding sites were strongly enriched for active histone marks, including H3K4me1, H3K4me2, H3K4me3, H3K9ac, and H3K27ac regardless of the cell type (K562 or U2OS cells) that those publicly available datasets were based on (Fig. 1E and Supplementary Fig. S1C, left). This is consistent with previous findings that ORC proteins preferentially associate with open chromatin bearing active histone modifications such as H3 acetylation and H3K4 methylation. It is interesting to note that upon the intersection of Orc6 peaks with H3K4me3 peaks (GSE96303), we found that 42% (2702 peaks out of 6452 Orc6-H3K4me3 co-bound peaks) co-localized with active origins. However, some Orc6 sites overlapped with active origins, which were not marked by the H3K4me3 signature, suggesting that two independent populations of Orc6 are present, likely with different functions. In contrast, no substantial enrichment was observed for H3K27me3, H4K20me1, H3K9me3, or H3K36me3 (Supplementary Fig. S1C, right).

Figure 1.

Genomic binding profile of Orc6. (A) Genome browser profiles displaying average Orc6 CUT&RUN signals (n = 2) at known DNA replication origins such as MYC, MCM4, and TOP1 (top), as well as a replication origin-enriched region, the HoxA cluster (bottom). Replication origins identified from SNS-Seq data are indicated by blue bars. (B) Venn diagram illustrating the overlap between Orc6 peaks and SNS-Seq peaks representing active replication origins. (C) Orc6 peak annotation. (D) Heatmap showing Orc6 and H3K27ac enrichment at TSSs of top and bottom 10% expressed genes in U2OS cells. H3K27ac ChIP-seq data are from U2OS cells (GSM7793922). Top/bottom 10% genes were determined by RNA-seq data from U2OS cells (GSE291919). (E) Heatmap analysis of ChIP-seq data for active histone marks from K562 cells around the summits of Orc6 peaks (scale is indicated in kilobases). (F) Heatmap of Orc6 signal distribution over H2A.Z peaks (scale is indicated in kilobases). (G) Genome browser profile of regions showing Orc6 co-localization with origin (SNS-Seq) and active histone marks. (H) Heatmaps of read distributions for DNase-seq around Orc6 peaks (scale is indicated in kilobases). Signal strengths of Orc6 CUT&RUN were normalized by RPKM.

Recent studies have shown that nucleosomes containing the histone variant H2A.Z facilitate ORC binding and origin licensing [34]. To explore this further, we examined H2A.Z ChIP-seq data [GSE73742; [35]] and observed a strong co-localization between Orc6 and H2A.Z (Fig. 1F). Approximately 38% of H2A.Z peaks overlapped with Orc6 binding sites, and among these, 87% were located within promoter regions (Supplementary Fig. S1D). Collectively, the heatmaps, profile plots, and genome browser tracks demonstrate that Orc6 predominantly localizes to open chromatin regions enriched with replication origin-associated marks, including H2A.Z, H3K4me3, and H3K27ac, while showing minimal overlap with marks such as H3K9me3, H3K27me3, H3K36me3, and H4K20me1 (Fig. 1G, and Supplementary Figs. S1E and S1F). To further increase rigor, we performed the same analysis on the 10 416 peaks co-identified by two independent Orc6 antibodies (Supplementary Fig. S1A), which exhibited a similar enrichment at active histone marks (Supplementary Fig. S1G).

We then examined the distribution of G-quadruplex (G4) DNA structures, which have been previously mapped to replication origins and implicated in replication initiation [36]. Using ChIP-seq data generated with a G4-specific nanobody (SG4) in U2OS cells [GSE207567; [37]], we observed SG4 enrichment at a subset of Orc6 binding sites, indicating potential co-localization (Supplementary Fig. S1H). Further to assess the nature of chromatin around the binding sites for Orc6, we analyzed DNase-seq data from U2OS cells [GSE142160; [38]]. Orc6 peak summits were frequently positioned between two closely spaced DNase hypersensitive sites, indicative of open chromatin regions (Fig. 1H). Collectively, these results show that Orc6 is associated with a reproducible set of binding sites in the human genome. Orc6 peaks are enriched at promoter regions (48%) and are also found within intronic (21%) and intergenic regions (22%), including TSSs. Notably, 26.9% of Orc6 peaks co-localized with active replication origins identified by SNS-seq (Fig. 1B). Our results demonstrate that Orc6 associates with open chromatin regions and is found at nucleosome-free regions.

Oxidative DNA damage promotes increased recruitment of Orc6 to damage sites

We previously reported that Orc6 is upregulated during the S-phase in response to oxidative stress [9]. We performed gel-filtration of Orc6 in S-phase-enriched U2OS cells with and without H₂O₂ treatment. Upon H₂O₂ treatment (5 mM for 30 min), Orc6 was found to be in a higher molecular-weight complex, implying the redistribution of Orc6 into certain protein complexes (Fig. 2A). This incorporation into a higher molecular-weight protein complex was not observed in G1-phase-synchronized cells (Fig. 2A). To assess the genomic binding profile of Orc6 under oxidative stress, we treated S-phase–synchronized U2OS cells with H₂O₂ and performed CUT&RUN. It is important to note that the Orc6 final bleed Ab 1123 recognizes both unphosphorylated and threonine 229 (T229) phosphorylated Orc6. Using Ab 1123 for CUT&RUN, we assessed the binding status of phosphorylated Orc6 (at position 229) in both undamaged conditions (representing endogenous damage) and after exposure to external oxidative stress. We observed a ∼1.5-fold increase in the number of Orc6-bound peaks in H₂O₂-treated cells (37 665 peaks) compared to untreated control cells (23 776 peaks), indicating enhanced chromatin association of Orc6 upon oxidative DNA damage (Fig. 2B). Significantly, the antibody that detected T229 phosphorylation, a modification specific to DNA damage, showed several newer Orc6 binding sites upon DNA damage. Genome browser visualization of representative loci, including MYC, MCM4, TOP1, and the HoxA gene cluster, showed continued association of Orc6 to origins following H₂O₂ treatment (Supplementary Fig. S2A). We observed more Orc6 peaks co-localized with origins (9646 out of 37 665) following H₂O₂ treatment. However, the overall fraction of Orc6 co-localized with origins remained the same (Supplementary Fig. S2B). Next, we sought to characterize peaks for Orc6 upon H₂O₂ treatment. Annotation of Orc6 peaks in H₂O₂-treated samples revealed a redistribution of binding under oxidative stress, characterized by a reduction in promoter-associated peaks (34%) and a concomitant increase in peaks within intronic (29%) and intergenic (25%) regions, relative to untreated cells (Supplementary Fig. S2C). We then quantified the total normalized read counts per peak across the genome in untreated and H₂O₂-treated conditions. H₂O₂ treatment did not lead to an increase of Orc6 read count on Orc6 peaks suggesting similar overall binding under the two conditions (Supplementary Fig. S2D). Note that 26.21 and 23.90 are the median RPKM values of Orc6 read counts in untreated condition and with H₂O₂ [plotted on y-axis as log2(RPKM + 1)]. However, we observed increased peak width for Orc6 upon H₂O₂ treatment, indicating broader binding at certain loci (Supplementary Fig. S2E). We have previously demonstrated that Orc6 is phosphorylated at T229 in response to oxidative damage during S-phase [9]. To investigate whether this modification correlates with chromatin association, we performed chromatin fractionation followed by immunoblotting. Orc6 levels in the chromatin-bound fraction increased upon H₂O₂ exposure, along with phosphorylated Orc6 (pT229) levels (Fig. 2C), supporting a role for this modification in the DNA damage response.

Figure 2.

Orc6 upon oxidative damage. (A) Gel-filtration of Orc6 in S- and G1-phase with presence or absence of oxidative damage. The numbers below represent different fractions. Red box represents gel-filtration fractions containing >200 kDa protein complexes. (B) Venn diagram showing overlap between Orc6 CUT&RUN peakset under non-treated and H₂O₂-treated conditions. (C) Chromatin fractioning of Orc6 and Orc6-pT229 upon oxidative damage. S, soluble. C, chromatic. The middle band in the Orc6 blot represents Orc6 (labeled by the triangle pointer). Bar graph shows the relative protein level (normalized to NT soluble). N = 3 biological replicates were done, and error bars represent standard deviations. (D) Overlap between Orc6 and OGG1 CUT&RUN peaksets under non-treated and H₂O₂-treated conditions. (E) Genome browser tracks of average signals for Orc6 CUT&RUN (n = 2) at OGG1 peaks without or with oxidative damage. Orc6 peaks are shown as blue bars and OGG1 peaks as red bars. Note that the Venn diagrams were generated using Intervene, a bedtool that intersects by overlapping at a base-pair resolution. Peaks with any overlap are counted as shared, resulting in reduced number of peaks in each set.

Oxidative stress induces the oxidation of guanine residues in DNA, resulting in the formation of 8-oxo-7,8-dihydroguanine (8-oxo-dG). This lesion is recognized and excised by the base excision repair (BER) pathway, which is initiated by DNA glycosylases such as OGG1 [39]. To identify genomic sites of oxidative DNA damage following H₂O₂ treatment, we performed CUT&RUN for OGG1 in both untreated and H₂O₂-treated S-phase-synchronized U2OS cells. In untreated cells, approximately 43% of Orc6 peaks overlapped with OGG1, likely reflecting basal levels of endogenous oxidative damage (Fig. 2D, left). Upon H₂O₂ treatment, this overlap increased substantially, with ∼87% of Orc6 peaks overlapping OGG1 peaks, indicating Orc6 accumulation at sites of oxidative DNA damage (Fig. 2D, right). Genome browser profiles further illustrated enhanced Orc6 binding following H₂O₂ exposure, characterized by broader peaks co-localized with increased OGG1 signal (Fig. 2E and Supplementary Fig. S2F). Notably, analysis of newly gained Orc6 peaks upon oxidative stress revealed that ∼80% of these sites overlapped with OGG1 peaks (Supplementary Fig. S2G), further supporting Orc6 enrichment at DNA damage loci. Similar to Orc6, OGG1 showed enrichment at TSSs, while its signal becomes more broadly distributed across the genome upon oxidative damage (Supplementary Fig. S2H). Finally, peak annotation revealed that ∼50% of the Orc6–OGG1 co-bound peaks are associated with promoters (Supplementary Fig. S2I). These findings demonstrate that Orc6 exhibits widespread chromatin binding in response to oxidative stress, preferentially localizing to regions marked by oxidative DNA damage.

Orc6 interacts with ISWI chromatin remodeling enzymes

To elucidate the potential mechanisms of Orc6 in DNA damage response during S-phase, we previously performed immunoprecipitation followed by mass spectrometry (IP-MS) in T7-Orc6-overexpressed U2OS cells with the presence or absence of oxidative damage (5 mM H₂O₂ for 30 min, raw data in [15]). We identified increased interaction between Orc6 and a series of chromatin remodeling enzymes upon H₂O₂ treatment (Supplementary Fig. S3A). An independent IP-MS assay showed SMARCA1 and SMARCA5 to be high confidence interactors with Orc6, especially upon H₂O₂ treatment (Fig. 3A). SMARCA1 and SMARCA5 are two major ATPase cores of human Imitation SWItch (ISWI) chromatin remodeling complexes, which can slide nucleosomes on their own or with the regulation of other non-catalytic units but are not able to evict nucleosomes or exchange histone subunits [5, 27]. Some non-ATPase subunits of ISWI chromatin remodelers were also identified in the IP-MS in Supplementary Fig. S3A, indicating high confidence in identifying the ISWI-family protein complex as an Orc6 interactor (proteins underlined in red in Supplementary Fig. S3B). It is important to note that the ISWI complex is also present in the higher molecular weight fraction where Orc6 is relocalized upon H₂O₂ treatment (Supplementary Fig. S3C). The interaction between Orc6 and SMARCA1 upon oxidative damage was validated by Co-IP (Fig. 3B). Methyl methanesulfonate (MMS), another damaging agent that causes DNA alkylation and oxidative stress, but not hydroxyurea (HU) or camptothecin (CPT), was also found to cause an increased Orc6–SMARCA1 interaction (Fig. 3C). Co-IP in U2OS cells synchronized at different cell cycle stages suggested that the interaction predominantly occurred in S-phase (Fig. 3D). Further, the phosphorylation status of Orc6 did not impact the ability to interact with SMARCA1, as confirmed by co-IPs of phosphorylation-mimic mutation (T229E) or the phosphorylation-dead mutation (T229A) of Orc6 with SMARCA1 (Supplementary Fig. S3D). This suggested that the oxidative damage-specific phosphorylation of Orc6 is dispensable for this interaction. This conclusion was corroborated by the AlphaFold Multimer predictions of Orc6–SMARCA1 interaction, which showed that Orc6–CTD was not involved in the interaction with SMARCA1 (Supplementary Fig. S3E). Instead, the N-terminal and middle TFIIB-like domain [40] was found to have a higher potential to be involved in interaction with SMARCA1. Co-IP with truncated Orc6 showed that amino acids 1–200 of Orc6 were sufficient for interaction with SMARCA1 (Supplementary Fig. S3F and S3G). These indicated that both of the TFIIB-like domains were needed for efficient interaction. We conclude that Orc6 interacts with SMARCA1 predominantly in S-phase and shows an enhanced interaction upon encountering oxidative DNA damage.

Figure 3.

Orc6 interacts with chromatin remodeler SMARCA1. (A) T7-Orc6 IP-mass-spectrometry (IP-MS) in S-phase synchronized cells with DNA damage showing high-confidence T7-Orc6 interactors upon H₂O₂ treatment (in contrast with expression of T7-tag only) with a [Fold change]>2 and FDR < 0.05 cutoff. ORC components and ISWI remodelers are labeled. Sample number n = 3 for each condition. (B) Immunoprecipitation (IP) between T7-Orc6 and Flag-SMARCA1. (C) IP between T7-Orc6 and Flag-SMARCA1 under different DNA damage treatments. H₂O₂, 5 mM. CPT, camptothecin (20 μM). HU, hydroxyurea (5 mM). MMS, methyl-methanesulfonate (5 mM). (D) IP between T7-Orc6 and Flag-SMARCA1 at G1/S checkpoint, mid-S-phase and G2-phase. Cell cycle profiles are determined by propidium iodide (PI) staining and flow cytometry.

Orc6 Acts as a negative regulator of SMARCA1 by constraining its localization to specific genomic loci

To gain insights into the functional significance of Orc6 interaction with SMARCA1, we performed SMARCA1 CUT&RUN in unperturbed and H₂O₂-treated S-phase–synchronized U2OS cells. We observed that H₂O₂ treatment led to an increase in total SMARCA1 peaks compared to untreated conditions, suggesting enhanced chromatin binding of SMARCA1 (Fig. 4A, top left). It is important to note that Orc6 interacts with SMARCA1 during the S-phase but shows enhanced association upon oxidative DNA damage (Supplementary Fig. S3G, compare lanes in T7-IP, FL-NT, and FL-H₂O₂).

Figure 4.

Orc6 negatively regulates SMARCA1 genomic binding. (A) Venn diagram depicting the overlap between SMARCA1 CUT&RUN peak sets under different conditions. NT, non-treated. (B) Venn diagram showing overlap between SMARCA1 CUT&RUN peakset (Ctrl-si) and Orc6 peakset under non-treated and H₂O₂-treated conditions (top). Enrichment profiles of SMARCA1 CUT&RUN signal (RPKM normalized) on Orc6 peak summits under non-treated and H₂O₂-treated conditions (middle, scale is indicated in kilobases), with P-value generated by Kolmogorov–Smirnov test to compare the similarity of two samples. Genome browser signal tracks of Orc6 (gray) and SMARCA1 (green) CUT&RUN without or with oxidative damage (bottom). (C) Venn diagrams showing overlap between SMARCA1 and OGG1 CUT&RUN peaksets under different conditions. Percentages of overlapped peaks relative to SMARCA1 peak numbers are labeled. (D) Venn diagram showing overlap between SMARCA1 CUT&RUN peakset (Orc6-si) and Orc6 peakset without or with oxidative damage (top left). Heatmaps of SMARCA1 CUT&RUN signal (non-treated and H₂O₂-treated conditions, RPKM normalized) on summits of Orc6 peaks (scale is indicated in kilobases) (top right). Genome browser signal tracks of Orc6 (gray) and SMARCA1 (green) CUT&RUN in control and Orc6-depleted cells under non-treated conditions (bottom). (E) Left: Violin plots of normalized read counts per peak for SMARCA1 (Ctrl-si and Orc6-si) over Orc6 peakset without or with oxidative damage. Medians are shown as lines and mean values are indicated. Paired Wilcoxon test was performed (P < 2e-16). Right: Distribution of log2-scaled SMARCA1 peak width in each condition. Medians are indicated with lines and values (in bp). (F) Heatmaps of SMARCA1 CUT&RUN signal (RPKM normalized) (Ctrl-si and Orc6-si) on origins (SNS-Seq) summits under non-treated and H₂O₂-treated conditions (scale is indicated in kilobases). Note that the Venn diagrams were generated using Intervene, a bedtool that intersects by overlapping at a base-pair resolution. Peaks with any overlap are counted as shared, resulting in reduced number of peaks in each set.

To assess whether the increased SMARCA1 levels upon H₂O₂ treatment correlate with Orc6, we analyzed the overlap between SMARCA1 and Orc6 peak sets in both control and H₂O₂-treated samples. We observed an increased overlap between Orc6 and SMARCA1 binding sites following H₂O₂ treatment, consistent with our co-immunoprecipitation results (Fig. 4B, top; see also Fig. 3B). In H₂O₂-treated cells, approximately 54% of Orc6 peaks overlapped with SMARCA1, compared to only ∼20% in control cells (Fig. 4B, top right compared to top left). Moreover, the total number of SMARCA1 peaks overlapping with Orc6 increased nearly 4-fold upon oxidative stress, indicating enhanced SMARCA1 association with Orc6. To further assess SMARCA1 occupancy at Orc6 sites, we plotted normalized SMARCA1 CUT&RUN signal centered on Orc6 peak summits. Surprisingly, we observed a ∼50% reduction in SMARCA1 signal at these regions in H₂O₂-treated cells relative to control (Fig. 4B, middle). The significance of this difference was confirmed by Kolmogorov–Smirnov test (P < 2e-16). Genome browser visualization corroborated these findings, showing increased Orc6 enrichment and reduced SMARCA1 binding at shared sites, particularly within promoter regions (Fig. 4B, bottom). This indicates that SMARCA1 exhibits enhanced association with Orc6 following oxidative stress; however, Orc6 may concurrently restrict SMARCA1 occupancy at specific genomic loci. In vitro immunoprecipitation using purified proteins showed that Orc6 can directly interact with a fraction of SMARCA1, independent of chromatin, suggesting that the increased association of Orc6 to SMARCA1 during DNA damage could also be occurring outside of chromatin in the nucleoplasm (Supplementary Fig. S4A).

To investigate how Orc6 loss affects SMARCA1 genomic binding, we performed SMARCA1 CUT&RUN in Orc6-depleted (Orc6-si) cells under untreated and H₂O₂-treated conditions. H₂O₂ treatment led to increased SMARCA1 chromatin binding genome-wide, which was unaltered by Orc6 depletion (Fig. 4A, top right). In unperturbed cells, Orc6 depletion resulted in a twofold increase in SMARCA1 peaks. A similar trend was observed following H₂O₂ treatment, with Orc6 depletion leading to increased SMARCA1 binding, though to a somewhat lesser extent than untreated conditions (Fig. 4A, lower left and lower right). To quantify global SMARCA1 occupancy, we calculated the total normalized read counts per peak across the genome. SMARCA1 signals were significantly elevated in Orc6-depleted cells, both in the absence and presence of H₂O₂ (Supplementary Fig. S4B). Paired Wilcoxon tests confirmed these differences were highly significant (P < 2e-16). The increased SMARCA1 binding in Orc6-depleted cells was not attributable to changes in SMARCA1 protein levels, as its abundance did not increase upon H₂O₂ treatment or following Orc6 depletion (Supplementary Fig. S4C). Moreover, SMARCA1 protein levels remained stable throughout the cell cycle, indicating that the observed redistribution is not attributable to the cell cycle defects caused by Orc6 depletion ([15], Supplementary Fig. S4D). Next, we assessed SMARCA1’s association with DNA damage sites by overlapping SMARCA1 peaks with OGG1, the marker for oxidative DNA damage. Upon H₂O₂ treatment, the overlap between SMARCA1 and OGG1 increased markedly from ∼48% in control cells to ∼91%, indicating SMARCA1 enrichment at damage sites (Fig. 4C, top). This overlap was further enhanced upon Orc6 depletion under untreated and H₂O₂-treated conditions (Fig. 4C, bottom). Annotation of SMARCA1 peaks revealed that in control conditions, binding was relatively evenly distributed across promoters, gene bodies, and distal intergenic regions. However, Orc6 depletion resulted in an increased SMARCA1 binding at promoter regions (Supplementary Fig. S4E). Together, these findings suggest that the loss of Orc6 promotes SMARCA1 recruitment to open chromatin and sites of oxidative DNA damage, particularly within promoter regions, also enriched for replication origins.

To determine whether Orc6 loss affects SMARCA1 binding at genomic sites originally occupied by Orc6, we overlapped SMARCA1 peaks with Orc6 peak sets in untreated and H₂O₂-treated cells. In the absence of treatment, Orc6 depletion led to a notable increase in SMARCA1 binding at original Orc6-bound sites, with approximately 33% of Orc6 peaks now overlapping with SMARCA1 (Fig. 4D, top left). In contrast, SMARCA1 occupancy at these sites decreased under H₂O₂-treated conditions upon Orc6 loss (Fig. 4D, bottom left). We then examined global SMARCA1 signal intensity at Orc6 peak summits in control and Orc6-depleted cells. Interestingly, SMARCA1 enrichment at original Orc6 sites increased following Orc6 depletion in both untreated and H₂O₂-treated cells (Fig. 4D, top right), with a more pronounced effect observed without oxidative stress. Consistently, genome browser profiles demonstrated enhanced SMARCA1 binding at Orc6 sites upon Orc6 depletion in both conditions (Fig. 4D, bottom; Supplementary Fig. S4F). The depletion of Orc2, a core-ORC component, did not impact SMARCA1 distribution, showing that enrichment of SMARCA1 to Orc6-bound sites is a specific outcome of Orc6 loss and not observed upon the loss of the core ORC (Fig. 4D bottom, Supplementary Fig. S4G). Also, genomic binding of the other ISWI chromatin remodeler, SMARCA5, did not display any change or reorganization upon the loss of Orc6 (Supplementary Fig. S4H).

To quantify the enrichment of SMARCA1, we calculated total normalized SMARCA1 read counts at Orc6 peak sets across control and Orc6-depleted samples. The results showed a significant increase in SMARCA1 signal at Orc6 sites in both untreated and H₂O₂-treated cells upon Orc6 depletion (Fig. 4E, left), with statistical significance confirmed by paired Wilcoxon tests (P < 2e-16). We also observed increased peak width for SMARCA1 upon Orc6 loss in both control and treated conditions, suggesting more focused or constrained binding of SMARCA1 in the presence of Orc6 (Fig. 4E, right).

To evaluate the impact of Orc6 depletion and/or H₂O₂ treatment on SMARCA1 genomic binding, we performed differential peak analysis of SMARCA1 CUT&RUN data using DiffBind with DESeq2. Differential peaks were defined using a cutoff of [Fold change] > 1.5 and FDR < 0.05. Orc6 depletion, both in untreated and H₂O₂-treated cells, resulted in a strong enrichment of SMARCA1 binding at specific genomic loci (Supplementary Fig. S4I), with 7062 enriched peaks identified in unperturbed conditions and 12 384 following H₂O₂ treatment. In contrast, H₂O₂ treatment alone did not lead to significant local enrichment of SMARCA1 at defined loci but instead increased the total number of SMARCA1 peaks genome-wide, indicating broader distribution without focal enrichment. These results again suggest that Orc6 may function to restrict SMARCA1 binding, particularly at its own occupied genomic sites.

Next, to investigate whether the observed SMARCA1 enrichment upon Orc6 depletion coincides with replication origin activity, we examined SMARCA1 distribution over SNS-seq peaks, which mark active replication origins. SMARCA1 was preferentially enriched at active origins in the absence of Orc6 under both untreated and H₂O₂-treated conditions (Fig. 4F). The genome browser profile further illustrates increased SMARCA1 binding at the HoxA cluster upon Orc6 depletion, supporting that SMARCA1 binding is enhanced at active replication origins when Orc6 is absent (Supplementary Fig. 4J).

Orc6 Inhibits SMARCA1-mediated chromatin remodeling by regulating its binding to DNA

Since SMARCA1 is a chromatin remodeler, we hypothesized that the enriched SMARCA1 upon Orc6 depletion alters the chromatin architecture at Orc6-bound sites. To verify the functional relevance of SMARCA1 enrichment on Orc6 sites, we performed ATAC-seq analysis in unperturbed and Orc6-depleted cells under both untreated and H₂O₂-treated conditions. Globally, the depletion of Orc6 led to higher chromatin accessibility in both unperturbed and H₂O₂-treated cells, with a more pronounced effect observed in the absence of oxidative stress. However, H₂O₂ treatment reduced chromatin accessibility, indicating global chromatin compaction (Supplementary Fig. S5A). Further, to determine whether Orc6 loss affects chromatin accessibility at SMARCA1-bound sites, we examined ATAC-seq data over SMARCA1 peaks in both control (Ctrl-si) and Orc6-depleted (Orc6-si) cells under both untreated and H₂O₂-treated conditions. We observed higher chromatin accessibility at SMARCA1-bound sites upon Orc6 depletion in unperturbed cells (Fig. 5A). However, the chromatin accessibility changes were not noticeable in control and Orc6-depleted cells following H₂O₂ treatment (Fig. 5A). These findings supported the idea that SMARCA1 can create open chromatin in the regions in which it binds, while H₂O₂ treatment reduced the chromatin remodeling activity of SMARCA1. Further, to determine how change in chromatin accessibility at enriched SMARCA1 sites correlates with Orc6 loss, we examined ATAC signal at differential-SMARCA1 peaks (Fig. 5B, differential SMARCA1 peaks shown in Supplementary Fig. 4I). We observed that increased SMARCA1 binding upon Orc6 depletion correlated with higher chromatin accessibility at significantly enriched SMARCA1 sites in unperturbed cells (Fig. 5B, column 1). However, such correlation was not observed at sites with non-significant changes in SMARCA1 binding upon Orc6 depletion (Fig. 5B, column 2). We also observed a similar trend in H₂O₂-treated cells but to a somewhat lesser extent compared to untreated conditions (Fig. 5B, columns 3 and 4). Another recent study showed that the SMARCA1 KO cell line exhibited decreased chromatin accessibility at TSSs and SMARCA1-bound sites, consistent with our observations [41].

Figure 5.

Orc6 depletion-related SMARCA1 enrichment leads to chromatin accessibility increase. (A) Heatmaps of chromatin accessibility (ATAC-seq) profile within ±5 kb from SMARCA1 peak centers. (B) Profile plots of SMARCA1 CUT&RUN (upper) and ATAC-seq (lower) on SMARCA1 significantly enriched peaks (columns 1 and 3, 7062 in NT, 12 384 in H₂O₂, see also Supplementary Fig. S4E). SMARCA1 binding loci with no significant change (columns 2 and 4) were used for contrast. Profiles were plotted within ±5 kb range from centers of genomic regions. Signal strengths are normalized in RPKM. P-values are generated by Kolmogorov–Smirnov test. (C) Heatmaps of ATAC-seq and SMARCA1 CUT&RUN normalized signal (RPKM) on ATAC-seq peaks upregulated by Orc6 depletion in non-treated condition (7517). (D) Heatmaps of ATAC-seq and SMARCA1 CUT&RUN normalized signal (RPKM) on ATAC-seq peaks upregulated by Orc6 depletion in H₂O₂-treated condition (1143). (E) Genome browser tracks of average signals for ATAC-seq (blue) and SMARCA1 CUT&RUN (green) in unperturbed and H₂O₂-treated cells. (F) In vitro nucleosome sliding assay using mononucleosome construct 4N66. The upper band represents non-remodeled nucleosome, and the lower band represents remodeled nucleosome that exposes the DpnII restriction site. DpnII digestion lasted 30 min. Quantification of lanes 2–6 (corresponding to letters A–E) shows the percentages of remodeled nucleosomes that are normalized to positive control (B). N = 2 biological replicates were performed and used for quantification. No statistical significance was observed through Wilcoxon test.

Next, we performed a differential analysis of ATAC-seq data in the presence and absence of Orc6 under untreated and H₂O₂-treated conditions. Differential ATAC peaks were identified by csaw-edgeR ([Fold change] > 1.5, FDR < 0.05, Supplementary Fig. S5B, see also Fig. 5C and D, left panels). Orc6 depletion under unperturbed conditions resulted in 7517 enriched peaks identified but only 1143 following H₂O₂ treatment (Supplementary Fig. S5B). Further, we observed 5698 downregulated ATAC peaks in unperturbed and 1464 in H₂O₂-treated cells upon Orc6 depletion. This indicates that Orc6 loss resulted in an overall increased chromatin accessibility under unperturbed conditions. However, H₂O₂ treatment caused a global decrease in chromatin accessibility (with or without Orc6), as evident by the low numbers of loci with a significant change in ATAC-seq signal (Supplementary Fig. S5B). Next, we examined the SMARCA1 signal at differential ATAC peaks in both control and Orc6-depleted cells in the absence and presence of H₂O₂ treatment. We observed a strong SMARCA1 enrichment at genomic loci with increased chromatin accessibility (n = 7517, Fig. 5C and Supplementary Fig. S5C), but not at loci with decreased chromatin accessibility upon Orc6 depletion in unperturbed cells (n = 5698, Supplementary Fig. S5C). We also calculated the total normalized SMARCA1 read counts at differential ATAC peaks across control and Orc6-depleted samples and observed that SMARCA1 signals were significantly elevated upon Orc6 loss at significantly upregulated ATAC peaks but much less significantly at the downregulated peaks (Supplementary Fig. S5C). We did observe a similar correlation between increased SMARCA1 binding and chromatin accessibility upon H₂O₂ treatment, but the change appeared to be less significant (Fig. 5D and Supplementary Fig. S5C).

Next, we quantified the distance from ATAC-seq-enriched sites to their closest SMARCA1-enriched sites. We found that SMARCA1 binding sites enriched upon Orc6 loss showed enhanced chromatin accessibility within a range of 1 kb in unperturbed cells (Supplementary Fig. S5D, dashed line and arrow). Combining the ATAC-seq data with SMARCA1 CUT&RUN data, we infer that the depletion of Orc6 promotes SMARCA1 enrichment over original Orc6 sites, which leads to increased local chromatin accessibility. The genome browser profile further illustrated that enriched SMARCA1 binding correlates with increased chromatin accessibility upon Orc6 loss in unperturbed but not in H₂O₂-treated cells (Fig. 5E). This is likely to be a result of H₂O₂-induced chromatin condensation [42–45].

Since SMARCA1 is known to slide nucleosomes, we investigated how Orc6 impacts nucleosome remodeling. We especially wanted to know whether and how the increased level of Orc6 impacts chromatin accessibility and if Orc6 plays any role in regulating the SMARCA1 binding/activity. To answer this question, we performed an in vitro assay in which purified GST-Flag-hOrc6 were pre-incubated with edge-positioned mononucleosome (4-N-66) before purified SMARCA1 protein was added. When SMARCA1 slides the nucleosome to the middle of DNA, the DpnII restriction site is exposed. We quantified the fraction of DpnII-digested DNA as a sign of nucleosome sliding rate and normalized it to the positive control with only SMARCA1 present. We observed that the remodeling activity was increased when GST-Flag-hOrc6 was present at low molarity (1:1 relative to SMARCA1) but decreased with higher Orc6 concentration (2:1 or 3:1) (Fig. 5F). Orc6 expression level increases upon H₂O₂ treatment [9], so our results suggest that Orc6, at a high level, negatively regulates nucleosome remodeling either in a direct way by sequestering SMARCA1 or by limiting the binding of SMARCA1 on nucleosomes. This is consistent with our observations that Orc6 can directly bind to SMARCA1, suggesting that Orc6 binding to SMARCA1 may titrate SMARCA1 away to inhibit promiscuous chromatin remodeling.

Discussion

In this study, we mapped the genome-wide binding of hOrc6 during S-phase and found enrichment of Orc6 at open chromatin sites and replication origins. Detailed analyses of the various genomic features at the Orc6-bound sites supported the model that Orc6 shows excellent correlation with H3K4me2, H3K9ac, and H3K27ac, marks implicated in DNA replication and/or regulating replication timing [46–49]. Previous genomic analysis has suggested that early replication origins are associated with active chromatin marks [31]. The SNS-seq detects active origins, and our data demonstrating the enrichment of Orc6 on the SNS-seq peaks indicates the presence of Orc6 at active origins. We have previously shown that the loss of Orc6 results in defects in S-phase progression, likely due to defects in helicase activity. H3K4 methylation has been reported to be important for regulating Cdc45 binding to chromatin [49], suggesting that the association of Orc6 to certain genomic features may facilitate helicase loading.

Recent work has implicated H2A.Z nucleosomes localize to in vivo origin sites, especially where yeast ORC binds [34]. ORC was predicted to play a role in remodeling the nucleosomes to expand the nucleosome-free region to eventually enable MCM DH loading during G1 [6]. Chromatin remodelers, including the SWI/SNF or RSC, have been shown to be needed for DNA replication [50]. More recent biochemical evidence has demonstrated the requirement of chromatin remodelers in the maturation of MCM DH to CMG during S-phase [51], primarily attributed to the bigger size of the CMG helicase. An exciting recent discovery showed ORC as a master regulator of nucleosome reorganization at replication origins during G1 [52].

Extensive chromatin remodeling is required during DNA replication and repair. ATP-dependent chromatin remodelers can be classified into four categories based on their domain structure and function: SWI/SNF (switch/sucrose non-fermentable), ISWI (imitation switch), INO80 (inositol-requiring mutant), and CHD (chromodomain helicase DNA-binding) [53]. Multiple chromatin remodelers are present at the origins of replication, to regulate DNA replication and DNA repair [20]. ISWI remodelers have been shown to have increased residence times at replication and repair sites [54]. ISWI component, SMARCA1/SNF2L, travels with the replication fork and regulates nucleosome assembly to suppress post-replicative ssDNA gap accumulation [19]. Other ISWI components (including SMARCA1/SNF2L, SMARCA5/SNF2H, and ACF1) are also recruited to DNA damage sites for re-establishment of the nucleosome organization, and it is predicted to be an early DNA damage response [55]. We have observed the association of Orc6 to multiple chromatin remodeler complexes during DNA damage, including ISWI, CHD, and SWI/SNF. In this work, we have primarily focused on the association of Orc6 with SMARCA1, a chromatin remodeler known for generating ordered nucleosomal arrays. SMARCA1 is found at the replication fork, and its depletion results in increased DNA damage signaling [20, 56, 57]. Recent work has shown that nascent chromatin is more accessible to MNase upon the loss of SMARCA1, supporting the model that SMARCA1 organizes newly deposited chromatin in order to prevent gaps. When gaps are generated during DNA replication, it results in fork reversal to pause DNA synthesis.

Using immunoprecipitation experiments, we validated the interaction of Orc6 and SMARCA1 in human cells. Further, using genome-wide mapping using CUT&RUN, we also observe co-occupancy of Orc6 and SMARCA1 at replication sites. This association is further enhanced by DNA damage, which is evident in IP experiments. We observed increased co-occupancies of Orc6 and SMARCA1 upon DNA damage reflected by increased overlap between Orc6 and SMARCA1 peaksets. However, the peak width of SMARCA1-bound sites is much narrower, becoming more focused and restricted during DNA damage, while Orc6 binding is enhanced at those sites (Fig. 4D and E right). This redistribution of SMARCA1 could indicate increased specificity to DNA damage sites, resulting in a more targeted nucleosome positioning. The reduced enrichment at these narrow peaks suggests reduced overall chromatin engagement, likely due to chromatin reorganization during repair, evident by increased chromatin compaction upon DNA damage. It is also likely that SMARCA1 remains bound for a shorter duration or with lower affinity, providing clues to a transient role in the repair process. The loss of well-positioning of nucleosomes and enhanced chromatin compaction at replication forks may reduce replication fork progression upon DNA damage.

Moreover, the loss of Orc6 resulted in significant enrichment of SMARCA1 to the “Orc6-bound” sites, suggesting that Orc6 modulates the levels and extent of SMARCA1 binding to chromatin. Finally, chromatin remodeling experiments demonstrate that the levels of Orc6 are critical in regulating the chromatin remodeling activity of SMARCA1. We propose that Orc6 and SMARCA1 at the unperturbed replication forks are important for accurate chromatin remodeling. Enhanced Orc6 titrates the excess SMARCA1 from the replication sites upon DNA damage to prevent fork progression. Concomitantly, increased chromatin compaction is observed upon DNA damage, likely at sites adjacent to damage sites [58]. In the absence of Orc6, SMARCA1 seems to be overloaded at genomic sites, leading to promiscuous/altered chromatin remodeling and increased nucleosome spacing (Graphical Abstract).

Multiple possibilities could explain why the loss of Orc6 leads to enhanced binding of SMARCA1 at the sites originally occupied by Orc6. One possibility could be that Orc6 and SMARCA1 compete for the same binding sites and that Orc6 usually restricts the binding of SMARCA1 to specific chromatin sites. Orc6 doesn’t seem to have a role in chromatin compaction; therefore, we do not believe that its loss results in regions becoming more accessible to enable the binding of SMARCA1. There may be a functional redundancy between Orc6 and SMARCA1; therefore, the loss of Orc6 triggers a need for SMARCA1 to take over the Orc6 function to maintain the replication origin function or stabilize stalled forks. It is also possible that the changes in the chromatin marks upon the loss of Orc6 create a chromatin environment conducive to SMARCA1 binding. Equally likely is that Orc6 regulates SMARCA1 to prevent excessive remodeling at replication origins, and upon the loss of Orc6, SMARCA1 shows enhanced recruitment, leading to chromatin structure alteration. Since Orc6 interacts robustly with SMARCA1 during DNA damage, losing Orc6 may lead to aberrant DNA damage response, as evidenced by increased OGG1, leading to excessive recruitment of SMARCA1 to replication origins.

Our study provides a novel perspective on the role of Orc6 during DNA replication and damage during S-phase. Work in budding yeast has provided evidence that ORC is also a master regulator of nucleosome organization over replication origins by modulating the activity of chromatin remodelers INO80, ISW1a, ISW2, and Chd1 [52]. These remodelers can generate nucleosome arrays in vivo with ORC, supporting the model that ORC serves as a barrier factor in cooperation with chromatin remodelers in generating regularly spaced nucleosomes. It is known that the four chromatin remodelers act redundantly at origins, and therefore, it is not surprising that Orc6 is associated with multiple remodelers. The proper positioning of nucleosomes around origins is predicted to be essential for replication. Prior work has shown the importance of the Orc1-BAH and the ORCA-WD domains in interacting with nucleosomes. Chromatin defects associated with yOrc1 mutations that do not affect MCM loading suggest that proper nucleosome organization plays a critical role in DNA replication. This function of ORC to help generate NFR arrays doesn’t require MCM loading and happens in G1. In mammalian cells, Orc1 is degraded at the end of G1, and our results support that Orc6 interaction with several chromatin remodelers plays a key role in nucleosome organization during DNA replication and repair in post-G1 cells. Although the primary role of chromatin remodelers is to phase regularly spaced nucleosome arrays at barrier binding sites [59, 60], the mechanisms by which these barrier factors operate remain unclear. We suggest that the DNA-binding protein Orc6 is a barrier factor, which limits the activity of chromatin remodelers at replication and repair sites. This is consistent with wider SMARCA1-bound regions in the absence of Orc6 (Fig. 4E, right), supporting the conclusion that Orc6 binding to SMARCA1 is critical for modulating chromatin remodeling.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

The authors declare no competing interests.

Funding

This work was supported by NIH R01GM132458; ARPA-H to K.V.P., NSF-STC to KVP and SGP (NSF DBI 2243257) and NSF (2225464), NIH (R01GM125196, 1R35GM152450) to SGP.

Data availability

All data needed to evaluate and reproduce the conclusions in the paper are present in the paper and/or the Supplementary Materials. All materials generated are described in the Materials and Methods section.

Significance statement

Origin Recognition Complex 6 (Orc6) plays a fundamental role in genome surveillance during S-phase. Upon DNA damage, hOrc6 associates with ISWI chromatin remodeler SMARCA1 in S-phase and we provide evidence that Orc6 binding to SMARCA1 prevents chromatin remodeling likely to halt replication fork progression.