RNAs anchoring replication complex control initiation and firing of DNA replication

Abstract

Coordinated initiation of DNA replication is essential to ensure efficient and timely DNA synthesis. Yet the molecular determinants that confer origin selectivity in mammalian cells remain incompletely defined. Herein, we present data demonstrating a pivotal role for RNAs transcribed in the proximity of actively replicating gene loci. We show that RNAs aNChoring ORC1 (ANCORs) to the histone variant H2A.Z facilitate origin firing during DNA replication. This ANCOR-H2A.Z interaction appears to be essential for cells to duplicate their genetic material. Widespread and locus-specific perturbations of these transcripts correlate with anomalous replication patterns and a notable loss of the H2A.Z replicative marker at the origin site. Collectively, we present a previously undescribed RNA-mediated mechanism that is associated with the generation of active replication origins in mammalian cells. Our findings delineate a strategy to modulate the origins of replication in human cells at a local and global level, with potentially broad biomedical implications.

The molecular basis of DNA replication origin selectivity in mammalian cells is not fully understood. Here, the authors show that RNAs (ANCORs) anchoring the replication factor ORC1 to the histone variant H2A.Z facilitate DNA replication. Perturbation of this RNA-mediated mechanism alters DNA replication, revealing a strategy to modulate replication origins.

Introduction

In human cells, more than 6 billion base pairs of genomic DNA need to be replicated and packaged into chromatin at every cell division cycle. Disruptions to this process can lead to mutations and chromosomal aberrations with consequent effects on genomic integrity and the epigenome, potentially leading to genetic diseases and cancer¹. Bacteria have a unique, well-defined origin of replication, whereas eukaryotes have multiple origins of replication². In yeast, conserved autonomously replicating sequences (ARS) have been thoroughly characterized³. In contrast, mammalian replication origins are not conserved among cell types or tissues, and only a handful of them have been characterized^4–6. Thus, the molecular mechanisms controlling the distribution and activation of replication origins across diverse cell types and developmental stages are at the center of studies worldwide.

DNA replication timing (RT) defined as the temporal order of chromosomal segments replicated during the S-phase in the cell cycle is cell-type specific. Human cell differentiation involves reproducible changes in RT. In fact, RT is disrupted upon malignant transformation, and comparisons of RT profiles between normal and cancerous tissues led to the identification of replication defects at specific gene loci, e.g., c-MYC, FXN1, and ROR1, that would not be detected by standard transcriptome analyses^7–9.

Foundational studies in budding yeast and mammals have established that origin recognition is mediated by a hexameric protein complex, termed the Origin Recognition Complex (ORC)^10–15. More recent developments using click chemistry-based approaches have facilitated the identification of multiple replication origins and confirmed the pivotal role of Origin Recognition Complex Subunit 1 (ORC1) in origin selection and RT¹⁶. Further, CryoEM structural analyses have suggested the existence of a pre-licensing step wherein the interaction of ORC1 with the preassembled ORC2-5 complex promotes the opening of the ORC core, facilitating its DNA binding¹⁷. Yet, despite multiple efforts, no consensus sequence motif has emerged from the analyses of ORC1 DNA-binding sites.

Intriguingly, ORC1 appeared to be preferably bound at enhancers and promoters marked by activating histone chromatin modifications^18,19. It was further hypothesized that the extension of the ORC1’s intrinsically disordered region (IDR) may facilitate recognition of specific nucleosome elements by direct physical interactions^18–21. Following this initial ORC1-nucleosome engagement, the pre-replication complex (pre-RC) is assembled throughout the G1/S phase, serving as a platform to recruit additional origin-licensing factors²¹. The ORC1–6, later engage CDC6 and CDT1 proteins²², At the last step, MCM2–7 loading onto chromatin ensures bidirectional replication of the origin^23–26.

Importantly, disruption of the initial ORC1-nucleosome recognition appears to undermine the entire pre-RC assembly. Supporting this notion, H Long et al. showed that HeLa cells treated with siRNA to H2A.Z histone variant stall at the G1/S boundary^27,28. The authors suggested therefore that H2A.Z, the major H2A variant associated with gene activation, is also responsible for recruiting ORC1 to early replication origins in an H4K20me2-dependent manner²⁸.

This model, however, does not address the mechanisms leading to H2A.Z enrichment coordinated with ORC1 engagement at early G1/S phase replication origins, but rather suggests that the plasticity of mammalian DNA replication programs may result from specific epigenetic features that modulate the selection and activation of the replication origins.

In this study, we unveil the role of previously undetected participants, namely, RNAs aNChoring ORC1 (ANCORs) as facilitators of mammalian replication origins, and describe a mechanism by which ANCORs engage ORC1 genome-wide while enabling binding of the histone variant H2A.Z^29–31. Importantly, we show that perturbation of the ANCOR-H2A.Z interaction results in parallel changes in DNA replication origins activity.

Our findings reveal a mechanism mediated by non-coding RNAs that links genetic and epigenetic features of mammalian replication origins and define an uncharted territory to control RT in human cells.

Results

c-MYC G1/S phase transcribed RNAs are implicated in c-MYC origin formation

Herein, we investigated the contribution of G1/S phase transcribed RNAs in the initiation of early replication origins. Our rationale was twofold: (1) the coordination by cell type- and phase-specific RNAs would reconcile the multiple replication origins observed across different cells, and (2) the colocalization of H2A.Z and ORC1 at early replication origins would align with recent findings linking G1/S phase transcribed RNAs to the deposition and modification of the histone variant H2A.Z³². First, we surveyed the c-MYC locus, the origin of which has been extensively studied^4,6. To examine the enrichment of ORC1 and H2A.Z at the c-MYC origin⁶, hereafter referred as region 9–10 (Fig. 1a), we performed chromatin immunoprecipitation (ChIP) with validated antibodies, followed by quantitative qPCR (Fig. 1a and Supplementary Fig. 1a). We detected 4.7- and 14.5-fold enrichment within region 9–10 for ORC1 and H2A.Z, respectively, suggesting the presence of both proteins at the target locus (Fig. 1a). Next, we probed the expression of nuclear RNAs arising from the c-MYC promoter and/or encompassing its replication origin during the G1 phase and upon the release to the S phase: S1 (2 h), S2 (5 h) (Fig. 1b and Supplementary Fig. 1b, c)³³. The highest expression levels of c-MYC promoter and origin transcripts were detected in the G1 phase, followed by the early S phase (S1) and then a marked decline in later S phase stages (S2), as compared to the unsynchronized cells (Fig. 1b)^32,34.

Fig. 1. c-MYC G1/S phase RNAs interact with ORC1.

a Upper panel: Schematic of the c-MYC locus, primer sets indicated by numbers. The red bar indicates the c-MYC replication origin (region 9–10). Bottom panel: ORC1 and H2A.Z enrichment at c-MYC origin detected by ChIP - qPCR analysis. Bars indicate the mean ± SD of three independent biological replicates (n = 3), paired Student’s t test (two-tailed), **P value = 0.0085 (ORC1) and **P value = 0.0055 (H2A.Z). b RNAs originating within the promoter of c-MYC, measured for K562 synchronized in G1 and S phase: S1 (2 h), S2 (5 h). Bars show mean ± SD of three independent biological replicates (n = 3), unpaired Student’s t test (two-tailed), *P value < 0.0001 (G1 vs WT), P value < 0.0001 (S1 vs WT) and P value = 0.00189 (S1 vs S2). c RIP - qRT-PCR performed using two ORC1 antibodies (ab85830 and ab60). Bars indicate mean ± SD correspond to two technical replicates (n = 2). d–f In vitro RNA/DNA electrophoresis mobility shift assay and RNA pull-down assays. d RNA secondary structures of probes overlapping c-MYC origin (RM9) and unrelated origin coordinates (MYC1, MYC2, and MYC3) were calculated using RNAfold. e RNA EMSA showing the preferential binding of ORC1 to RM9. f RNA pull-down assay. Bound proteins were purified on streptavidin beads and immunoblotted with antibodies against ORC1 and H2A.Z. The experiment was repeated three times with similar results; Protein marker (M, KDa): Precision Plus Protein™ (Bio-Rad, 1610374). g Molecular dynamics simulations: frontal views of the ORC1a, ORC1b, and ORC1c structures in complex with either dRM9 (DNA) or RM9 (RNA). RNA and DNA structures were derived using a RMSD-based clustering approach in the last half of the simulation time. All molecules are color-coded as follows: orange: ORC1-BAH domain (residues 1–200); marine: ORC1 disordered region and AAA+ domain (residues 201–782); gray: orc1-WHD domain (residues 783–861); red and yellow: RM9 and dRM9, respectively. h The zoomed-in view of both the top and frontal perspectives, illustrating the preferred binding state of the ORC1-RM9 complex. i Zoomed-in view showing the ORC1 amino acid residues interacting with RM9. Source data are provided as a Source Data file.

The precise mechanisms guiding the localization of human hORC1 on chromatin and its replication function remain poorly understood³⁵. RNase A treatment impairs ORC1 recruitment to the Epstein–Barr virus plasmid replication origin in vitro³⁵, as well as the interaction between telomeric repeat-binding factor 2 (TRF2) and ORC1³⁶. Recent studies have further indicated preferential binding of ORC1 to G-rich RNA or single-stranded DNA sequences in vitro³⁷, indicating a potential contribution of RNA to ORC1 recruitment. Thus, we explored the possibility of an RNA-dependent mechanism coordinating the initiation and chromatin status at replication origins in eukaryotic cells. To determine whether ORC1 physically associates with the c-MYC G1/S phase transcripts, we performed RNA immunoprecipitation (RIP) (Fig. 1c) and combined with Cross-Linking and Immunoprecipitation (CLIP) as described in Guo et al. (Supplementary Fig. 1d)³⁸ using two different ORC1 antibodies. Enrichment of c-MYC RNAs in the ORC1-RNA precipitates was detected by both approaches. We termed those transcripts c-MYC ANCORs (RNAs aNChoring ORC1).

To further assess whether cell-cycle synchronization influences the observed interaction, we repeated the RIP experiment using both G1/S phase-synchronized and unsynchronized K562 cells. The results demonstrated that the enrichment was consistently maintained across both conditions, with a modest yet reproducible increase detected in synchronized cells (Supplementary Fig. 1e).

To study the molecular properties of the RNA-ORC1 interaction in vitro, we performed an RNA/DNA electrophoresis mobility shift assay (EMSA) and RNA pull-down assay with K562 cell extracts (Fig. 1d–f and Supplementary Figs. 1f–h and 2a, b). The RNA oligonucleotides were selected based upon: (1) the proximity to the c-MYC origin and higher CG content (RM9) as well as (2) the distance from the c-MYC origin (oligonucleotides 7_8 and G4) and the lower CG content (oligonucleotides MYC1,2,3 Supplementary Fig. 2c). A DNA oligonucleotide (dRM9) corresponding to the RM9 sequence was also analyzed to compare the RNA and DNA interactions with ORC1. RNA-ORC1 complex formation was observed with RM9 (Fig. 1e, lane 2 and Supplementary Fig. 1f, lane 2 of left and middle panels). By contrast, no binding with the homologous DNA oligonucleotide was detected (Fig. 1e, lane 6 and Supplementary Fig. 1f, lane 6 left and middle panels), suggesting a non-preferential interaction of ORC1 with DNA as compared to RNA. RNA oligonucleotides unrelated to the c-MYC origin failed to produce a shift by RNA EMSA (Fig. 1e, lane 4, Supplementary Figs. 1f, lane 4 left and middle panels and  1g, h) and did not produce ORC1 enrichment in the RNA pull-down assay (Fig. 1f and Supplementary Fig. 2a, b), suggesting binding specificity between RM9 and ORC1. We then confirmed the presence of H2A.Z in the RM9 precipitates by RNA pull-down, indicating the presence of both ORC1 and H2A.Z in the complex (Fig. 1f and Supplementary Fig. 2a, b). Consistently, the interaction between H2A.Z and RM9 but not MYC1 was also detected by EMSA (right panel of Supplementary Fig. 1f), supporting the dual role of RNA in coordinating the localization of ORC1 while enabling binding to the histone variant H2A.Z³².

Next, we analyzed in silico the interaction between hORC1 and c-MYC ANCOR with respect to the equivalent DNA sequence to determine the structural and dynamic features of ORC1-RNA complexing. The interaction between ORC1 and RNA (RM9) and the respective DNA sequence (dRM9) was modeled using computational approaches. Three apo ORC1 states (ORC1a, ORC1b, and ORC1c) were considered equally probable conformations targeted by the sequence (Fig. 1g and Supplementary Fig. 2d). These states were derived from three distinct molecular dynamics (MD) trajectories performed starting from the AlphaFold ORC1 full-length model with the application of diverse velocities to maximize the exploration of the protein conformational ensemble. The considered states mainly differ from the relative orientation between the N- and C- ORC1 regions and that of the bromo adjacent homology domain (BAH domain), consistent with the prevalence of disordered and flexible residue patches between them. Both RM9 and dRM9 were docked against each of the three ORC1 states, and the obtained complexes were subjected to extended MD simulations to assess the reliability of the interaction (Supplementary Fig. 2e). Five out of six of the ORC1-RNA complexes showed a deviation from the initial binding mode. However, the ORC1a-RM9 complex revealed a conserved protein-RNA interface and no notable structural rearrangements of either molecular entity. These findings indicate that RM9 efficiently targets ORC1 at the specific binding site exhibited by the ORC1a conformation. The simulation revealed the smallest perturbations from the initial state, as reflected by the lowest mean value of the root mean squared deviation (RMSD) profiles with the respect to the starting simulation state (Supplementary Fig. 2e). Indeed, RM9 fits within the pocket embraced by the disordered region of ORC1 and the AAA+ domain, otherwise inaccessible within the 1b and 1c conformation (Fig. 1g, h). Along with the simulation (Supplementary Tables 8-9), the ORC1-RM9 hydrogen indicates that RM9 mainly connects to ORC1a through the ARG297, SER450, GLU575, THR593, and ARG846 residues, with occurrences greater than the 70% among the frames of the last 250 ns of trajectory (Fig. 1i, Supplementary Fig. 2f, g, and Supplementary Table 9). No significant rearrangements of the local structure of the protein were induced by RM9 targeting. Nevertheless, during the simulation, BAH approached the other protein domains (Supplementary Fig. 2f), suggesting that RM9 likely plays a role in favoring a specific ORC1 conformational equilibrium. We did not observe an effective binding mode in the ORC1a-dRM9 simulation, as shown by RM9. Importantly, the simulations evidenced that the DNA binding is facilitated by the involvement of the BAH domain in the interaction. This is consistent with the established knowledge that the BAH domain interacts with DNA/chromatin^20,39,40. In contrast, the most efficient RM9 binding mode resulted from the exposure of a specific pocket in the ORC1a conformation, wherein the BAH domain was displaced from the interacting region (Fig. 1g, h and Supplementary Tables 8 and 9).

Using a Light Shift Assay REMSA, we examined in greater detail the binding affinity of ORC1 for RM9 (Supplementary Fig. 3a–c). Specifically, a fixed amount of ORC1 (1 µg) was tested against increasing concentrations of RM9, ranging from 1 to 100 pmol. Construction of the binding curve and fitting with the Hill equation, under the assumption of non-cooperative binding (n = 1), yielded a dissociation constant (Kd) of 0.7441 µM. This Kd value is consistent with those reported for many other RNPs and transcription factors^41–46.

Subsequently, the relevance of interactions mediated by specific residues within the ORC1 lateral pocket and c-MYC ANCOR (RM9) for replication initiation was examined both in silico (Fig. 2a and Supplementary Fig. 3d, e) and in vitro (Fig. 2b, c). Using targeted mutagenesis, we generated two ORC1 variants: EYYE and AAAA. The EYYE variant was modeled by substituting ARG297, SER450, THR593, and ARG846 with ASP297, TYR450, TYR575, and ASP593, respectively. In contrast, the AAAA variant was generated through alanine substitutions at positions 297, 540, 593, and 846.

Fig. 2. Structural and functional characterization of ORC1-RM9 interactions and ORC1, RM9, H2A.Z in vitro reconstitution.

a RM9 interacting with ORC1a residues highlighted in pink. ARG297, SER450, GLU575, THR593, and ARG846 predominantly mediate the interaction during the simulation. The molecules are colored as follows: white: disordered region (residues 201–479), marine: AAA+ domain (residues 481–782), red and green: RM9 with native ORC1 and RM9 with mutated proteins, respectively. The mutated residues are displayed in sticks and colored in magenta. b Schematic representation of the structural domains of ORC1 (top panel). Zoomed-in view highlighting the ORC1 amino acid residues in the lateral pocket in interaction with RM9 (bottom panel). c Light Shift RNA EMSA with biotinylated RM9 (20 pmol), showing preferential binding of ORC1 to the Intrinsically Disordered Region (IDR; residues 201–782, lane 1), with no detectable binding to the BAH domain (residues 1–200, lane 2) or the Winged-Helix Domain (WHD; residues 783–861, lane 3). A mutant ORC1 protein (R297E, S450Y, T593Y, R846E) was also tested (lane 5), showing no binding. This REMSA experiment was performed three times. d RNA EMSA of in vitro reconstituted complexes containing ANCOR probe (RM9), ORC1, and recombinant nucleosomes containing H2A.Z (Active Motif, Cat. No. 81072). Free RNA probe (RM9) was loaded in control lanes 1 and 5. The RM9 was incubated with nucleosome containing H2A.Z alone, or with ORC1 alone, lanes 2 and 3, respectively. A specific complex formation is observed when H2A.Z, ORC1, and RM9 are together (lanes 4 and 6). Addition of anti-H2A.Z (lane 7) and anti-ORC1(lane 8) increases or decreases the intensity of the signal, consistent with antibody-dependent stabilization or perturbation of ORC1-RM9-H2A.Z assembly. e Western blot analyses used to validate the quality of the recombinant proteins employed for the RNA EMSA assay. The protein marker (M, KDa) is indicated in the figure and corresponds to Precision Plus Protein™ (Bio-Rad, 1610374). The ORC1 and recombinant nucleosomes containing H2A.Z were tested both separately and in combination to eliminate the possibility of mutual interference. Source data are provided as a Source Data file.

Although complete dissociation of RM9 was not observed within the molecular dynamic’s simulation timescale, the EYYE-RM9 and AAAA-RM9 complexes adopted distinct binding conformations, with the key anchoring interactions between the two molecules occurring at reduced frequency throughout the simulations (Supplementary Fig. 3d). When RM9 engages the mutated lateral pocket of ORC1, it fails to adopt its proper conformation and appears partially unfolded (Supplementary Fig. 3e). The introduced ORC1 mutations at specific residues appear to displace the RNA, disrupting its canonical stem-loop structure and compressing it toward one side of the binding pocket. The RM9 structure fails to maintain the functional stem-loop conformation, thereby compromising proper RNA-protein interactions^47,48 (Fig. 2a and Supplementary Fig. 3e).

This compelling result prompted us to directly assess the EYYE variant’s binding to RM9 in vitro. Strikingly, binding was abolished with the EYYE variant mutation, underscoring the high specificity of ORC1 for c-MYC ANCOR (Fig. 2b, c).

Further experiments were conducted to validate the role of ANCOR in mediating ORC1 anchoring to H2A.Z. Recombinant nucleosome containing H2A.Z and ORC1 proteins were first reconstituted and then incubated with RM9 (Fig. 2d, e). The rationale for using mononucleosome containing H2A.Z was to mimic in vitro chromatin contexts relevant to those present in living cells. A distinct complex was detected only when mononucleosome containing H2A.Z, ORC1, and RM9 were mixed together (Fig. 2d). Addition of anti-ORC1 antibody resulted in a reduction of this complex, whereas anti-H2A.Z antibody increased its intensity, consistent with antibody-dependent perturbation. Mass spectrometry analysis of gel-excised material confirmed enrichment of ORC1 and H2A.Z within this complex (Fig. 2d and Supplementary Fig. 3f). By adding this step in the protocol, we not only confirmed that the complex observed was the same in both conditions independently whether anti-ORC1 or anti-H2A.Z antibodies were used but were able to verify that ORC1 and H2A.Z are present within the same complex formed in vitro (Fig. 2d).

Altogether, these results support distinct RNA and DNA binding modalities of ORC1 and suggest that the ORC1 lateral pocket displays a defined spatial arrangement of functional groups that confers binding specificity to the GC-rich duplex within the RM9. As a result, RNA in its double-helical conformation can be accommodated within ORC1, thereby facilitating accurate origin site recognition and subsequent H2A.Z anchoring.

In summary, our findings denote colocalization of H2A.Z and ORC1 mediated by c-MYC ANCOR and suggest preferential binding of ORC1 to RNA over DNA both in vitro and in silico.

c-MYC ANCORs control the replication process at the c-MYC locus

To assess whether alterations in the transcriptional levels of c-MYC ANCORs could impact the integrity of the replication process at the c-MYC origins, we performed both CRISPR-mediated loss- and gain-of-function experiments. Loss-of-function was achieved by repressing c-MYC ANCORs using CRISPR interference (CRISPRi, CRISPR-dCAS9 system)⁴⁹ in the K562 cell line. Three single guide RNAs (gRNA1-2), targeting a region located 1.4 kb upstream of the c-MYC origin, were constitutively induced, along with the scramble control, in K562 cells stably expressing dCas9-mCherry (K562 dCas9-mCherry) (Supplementary Fig. 4a, b). The gRNAs were specifically designed to interfere with c-MYC ANCOR transcription without impairing the integrity of the replication origin. gRNA-2 led to lasting downregulation of c-MYC ANCOR, followed by a reduction in c-MYC mRNA levels (Supplementary Fig. 4c). Hence, gRNA-2 was selected for all downstream experimental analyses. To control the expression of the gRNAs, gRNA-2 and the respective scramble sequence were cloned under a Tet-inducible promoter and delivered into K562-dCas9-mCherry cells by lentiviral transduction (Fig. 3a, b). Upon treatment with 5 µM doxycycline, we followed the expression of gRNA-2 and the scramble control over multiple timepoints (Fig. 3c). The gRNA levels peaked at day three, which corresponded to the greatest decrease (more than twofold) in c-MYC ANCOR expression (Fig. 3d).

Fig. 3. c-MYC ANCOR levels regulate c-MYC origin activation.

a c-MYC locus schematic, guide RNA-2 located at 1.4 kilobases upstream of the replication origin. b Schematic of doxycycline-responsive promoter (TET-ON) for inducible expression of gRNA-2 in asynchronized K562 expressing dCas9; c qRT-PCR analyses of scramble and gRNA-2 expression after single 5 mM doxycycline addition. The mean ± SD of two technical replicates. d c-MYC ANCOR and c-MYC expression upon induction of gRNA-2 and scramble. Bars indicate mean ± SD of three biological replicates (n = 3). Paired Student’s t test (two-tailed), ***P = 0.0001 (c-MYC ANCOR) and **P = 0.0028 (c-MYC mRNA). e nasDNA abundance at c-MYC origin regions 7–8 and 9–10, respectively, upon c-MYC ANCOR downregulation. The mean ± SD of two technical replicates. f–h 15 µM ASOs targeting c-MYC ANCOR or non-targeting control delivered into synchronized G1/ early S phase K562 cells. f qRT-PCR analyses of c-MYC ANCOR at 72 h. The mean ± SD of two technical replicates. g A schematic outlining the c-MYC origin (9, 10). ChIP-qPCR analyses, H2A.Z (left panel) and ORC1(right panel) antibodies, confirming reduced occupancy after ASO delivery. The mean ± SD for three biological replicates (n = 3). A paired Student’s t test (two-tailed) for H2A.Z vs. IgG, shows **P = 0.0046 for the scramble and ***P = 0.0006 for the ASO targeting ANCOR. And ORC1 vs. IgG, **P = 0.0096 for the scramble and **P = 0.0011 for the ASO targeting ANCOR; h. nasDNA abundance decreases by ASO treatment at c-MYC origin (9–10). The mean ± SD for three biological replicates (n = 3). A paired Student’s t test (two-tailed) comparison of scramble vs ASO-treated cells, *P = 0.00193. i gRNA-2 in asynchronous HEK 293 expressing VP64-dCas9 upregulates c-MYC ANCOR and c-MYC mRNA. Bars indicate mean ± SD of three biological replicates. Paired Student’s t test (two-tailed) *P 0.0161 (c-MYC ANCOR) and **P 0.0015 (c-MYC mRNA). j c-MYC ANCOR activation increases nasDNA within regions 7–8 and 9–10. The mean ± SD of two technical replicates. Source data are provided as a Source Data file.

To define the impact of ANCOR depletion on DNA replication at the c-MYC origin, we quantified the nascent DNA (nasDNA) abundance within the origin and found a stark decrease in nasDNA generated near the origin (between 2.4- and 4.5-fold), (Fig. 3e). Importantly, no substantial changes in cell growth were observed as no differences in MYC protein levels between the control and gRNA-2 expressing cells were observed upon doxycycline treatment and after its suspension (Supplementary Fig. 4d). These findings suggest that the effect on nasDNA was independent of c-MYC levels^50,51. To rule out potential off-target effects of CRISPRi on the c-MYC expression, as well as the effects from replication roadblocks imposed by chromatin-bound dCas9^52,53, wild-type K562 cells were treated with FANA antisense oligonucleotides (ASOs) selectively targeting c-MYC ANCOR (Fig. 3f–h and Supplementary Fig. 4e). Depletion of c-MYC ANCOR led to a comparable reduction in c-MYC expression (Fig. 3f and Supplementary Fig. 4e). Consistent with the proposed mechanism, reduced occupancy of both H2A.Z and ORC1 was observed at the c-MYC replication origin and measurement of nascent DNA abundance revealed a concomitant decrease in nasDNA synthesis (Fig. 3g, h).

To complement the loss-of-function experiments and to ensure the consistency of the observed phenotype, gain-of-function experiments were then performed using the CRISPR activator (CRISPRa, CRISPR-VP64) system⁵⁴. To that end, gRNA-2 was delivered in stably expressing dCas9-VP64 HEK 293 cells, in which c-MYC is not actively expressed in contrast to K562 cells^55,56. This led to the robust upregulation of c-MYC ANCORs lasting up to five weeks, which was paralleled by an increase in c-MYC expression (Fig. 3i). nasDNA abundance was measured in dCas9-VP64 HEK 293 cells stably expressing gRNA-2 or the scramble control. We observed enrichment of newly synthesized DNA strands at the origin of replication 127 bp upstream of the c-MYC TSS upon expression of c-MYC ANCORs as compared to controls, hence supporting the hypothesis that c-MYC ANCORs could facilitate activation of the c-MYC replication origin (Fig. 3j). Interestingly, we also observed an increase in cell growth, likely due to the increased levels of c-MYC protein (Supplementary Fig. 4f).

Collectively, these results delineate the role of c-MYC ANCORs in the formation of the c-MYC replication origin and underscore how alterations in their levels may perturb the integrity of the replication process at the targeted locus.

ORC1 engagement of H2A.Z mediated by RNAs transcribed at G1/S phase controls DNA replication

The identification of c-MYC ANCORs and the RNA-binding features retained by ORC1 and H2A.Z suggested the involvement of RNA in the formation of replication origins and their chromatin status at the c-MYC locus. Therefore, we sought to explore the extent of ORC1-S phase RNA association in other genomic loci with respect to the presence of DNA replication origins. To this end, we assessed the transcriptional profile of G1/S phase transcribed RNAs upon cell synchronization by performing nascent RNA sequencing (nasRNA-seq) at earlier (S1) and later (S2) timepoints after the release in the S phase over the course of 5 h³³ (Fig. 4a). This approach led to the identification of 10,087 genes expressed in the earlier S1 timepoint and 10,171 in the later S2 timepoint. Both included c-MYC within the group of genes displaying the highest expression (Fig. 4b and Supplementary Fig. 5a) and FXN among the group of genes displaying the lowest expression (Fig. 4b and Supplementary Fig. 5a). FXN is harbored in a mid-late replicating locus, and its replication profile has been well characterized owing to its role in Friedreich’s ataxia^9,57,58. Next, we examined the interaction between ORC1 and the RNAs transcribed in the first 3 h of S phase to determine whether we could establish a correlation between the presence of ORC1–RNA binding and early replication origin formation. We uncovered 81,931 predicted loci associated with ANCORs, thereby expanding on the potential global nature of transcripts showing features similar to those of c-MYC ANCORs. Since Long et al.^27,28 have proposed a function for H2A.Z in licensing DNA replication loci, we investigated the linkage between the genomic location of ANCORs and enrichment of H2A.Z in K562 cells (Fig. 4c). Over 56% of loci associated with ANCORs overlapped with H2A.Z-enriched genomic locations, with 46,360 (ANCORs/H2A.Z) predicted early replication origins (Fig. 4c, d). The ANCOR-defined origins were broadly distributed across the genome, including promoters/transcription start site (TSS), gene bodies (3’ UTR, 5’ UTR, exons, and introns), and intergenic regions. Specifically, when narrowing the window for origin localization from +/−100 kb to +/−0.5 kb with respect to TSS, the number of predicted early DNA replication origins only decreased by twofold, from 43,188 to 19,725. A significant proportion of these loci were located within the proximity of promoters and/or TSS of protein-coding genes, specifically 58% within 2 kb or 67% within 0.5 kb as compared to 28% within 100 kb (Fig. 4d and Supplementary Fig. 5b, c). Overall, these results confirm that most DNA replication origins tend to localize in closer proximity to the promoters and/or TSS of protein-coding genes, supporting earlier observations^2,59.

Fig. 4. ORC1 engagement of H2A.Z by ANCORs controls DNA replication.

a Schematic showing cell synchronization in the G1/S phase by double thymidine block. The growth medium was supplemented with EU at the indicated time points: S1 (2 h) and S2 (5 h). EU-labeled nascent RNA was bound to biotin by Click-it chemistry reaction, isolated by streptavidin beads, and sequenced. b Heatmap showing gene expression profiles of K562 at S1 and S2. Welch’s t test: P value = 0.02993. c Workflow employed to predict early replication origins through the integration of RNAs aNChoring ORC1 (ANCOR), detected by RNA immunoprecipitation with anti-ORC1 antibody, followed by sequencing (RIP-seq) and H2A.Z-enriched sites captured by Chromatin Immunoprecipitation sequencing (ChIP-seq). d Upper panel: Venn diagram showing early DNA replication origins as predicted by the integration of ANCOR and H2AZ-enriched loci. Overlap between the two genomic interval sets was assessed by interval intersection, and the significance of the overlap was evaluated using a Fisher’s exact test (hypergeometric framework). Lower panel: Annotation of the predicted early DNA replication origins located within +/−100 kb, +/−50 kb, +/−20 kb, +/−10 kb, +/−5 kb, and +/−2 kb of the transcription start site (TSS). A significant proportion of the predicted early DNA replication origins are within close proximity to promoters and/or TSS of coding genes. e Schematic showing K562 synchronization in G1/S phase cycle, followed by incorporation of BrdU into nasDNA and pharmacological inhibition of transcription using DRB (100 mM). Samples were collected 3 h after DRB treatment and every hour upon DRB reversal over the following 3 h. f Reversible downregulation of c-MYC ANCOR and c-MYC mRNA expression by qRT-PCR upon DRB treatment. Bars indicate mean ± SD for three independent biological replicates (n = 3). A paired Student’s t test (two-tailed), ***P value = 0.0007 (c-MYC ANCOR) and ***P value = 0.0006 (c-MYC mRNA). g Flow cytometry analysis showing nonreversible inhibition of nasDNA synthesis at the selected time points. h DNA abundance within the c-MYC locus decreases upon DRB treatment within the region encompassing the origin (7–8 and 9–10). Bars indicate mean ± SD of two technical replicates (n = 2). Source data are provided as a Source Data file.

To verify whether pharmacological inhibition of transcription could lead to alterations in S-phase cell cycle progression, K562 cells were synchronized by double thymidine block⁵⁶ and treated with the RNA Polymerase II (RNAPII) inhibitor 5,6-Dichlorobenzimidazole-1-β-D-ribofuranoside (DRB) upon release into S phase. Concurrently, 3.5 µM 5-bromo-2′-deoxyruridine (BrdU) was added to the cells to monitor the nascent DNA synthesis by flow cytometry following transcriptional inhibition. We took advantage of the reversibility features of DRB and examined the effect of global transcriptional inhibition 3 h upon DRB treatment and every hour for 3 h after DRB reversal (treatment outline is shown in Fig. 4e–g). We monitored transcriptional inhibition by measuring c-MYC ANCOR levels by qRT-PCR (Fig. 4f). Strong downregulation of c-MYC ANCOR and c-MYC mRNA (nearly three- and ninefold change, respectively) was noted 3 h after DRB treatment. However, these changes were transient and reversible, and the expression returned to baseline during DRB reversal (Fig. 4f). By contrast, nonreversible inhibition of cell cycle progression was observed at all selected time points upon DRB reversal, as evidenced by the sharp reduction to 38% of the S phase cell population in DRB-treated cells as compared to the 60% of the S phase population in DMSO-treated control cells (Fig. 4g). Consistently, this pattern was associated with a decrease in nasDNA synthesis at the c-MYC locus and loss of ORC1 and H2A.Z enrichment at the c-MYC origin upon transcriptional inhibition (Fig. 4h and Supplementary Fig. 5d).

The loss of H2A.Z upon DRB treatment is not attributable to a global decrease in protein levels. Western blot analysis performed in G1/S synchronized K562 cells 3 h after DRB treatment demonstrates persistence of the H2A.Z (Supplementary Fig. 5e, f). The local reduction of H2A.Z enrichment at origin sites indicates a loss of an origin-associated marker and impaired activation of ANCOR-regulated origins. Consistently, Long et al observed a similar G1/S arrest in HeLa cells following siRNA-mediated depletion of H2A.Z, revealing a stall at the G1/S boundary^27,28.

We, therefore, reasoned that if the formation of the ternary ORC1-ANCOR-H2A.Z complex is disrupted, the origin site recognition would be compromised, ultimately impairing efficient origin activation. To test this concept, we performed ChIP-qPCR to compare MCM complex loading at the c-MYC origin for both the DMSO control and DRB-treated K562 (Fig. 5a–c). Notably, MCM2 and MCM7, which normally occupy the c-MYC origin during DNA replication, were absent from this site in DRB-treated cells, where c-MYC ANCOR is inhibited, and a similar lack of MCM2 loading was observed at the FXN1 origin (Fig. 5a–c and Supplementary Fig. 5g). At the same time, c-MYC ANCOR does not directly interact with MCM2, as shown by RIP-qRT-PCR (Fig. 5d), nor with ORC2 (Supplementary Fig. 5h), supporting a model in which formation of the ORC1-ANCOR-H2A.Z complex is required for proper origin activation.

Fig. 5. ANCORs promote efficient origin activation and mark early replication origins.

a–c Investigation by ChIP-qPCR, of MCM2 at the c-MYC origin in DRB versus DMSO-treated cells revealed enrichment of MCM2 and MCM7, confirming loading of MCM2 complex in the origin (set primers 9–10) 3 h post-release in G1/S-synchronized K562 cells in control cells and its absence 3 h post-release in G1/S-synchronized K562 cells in DRB-treated cells. Bars indicate mean ± SD for three independent biological experiments (n = 3). Paired Student’s t test (two-tailed) ***P = 0.0003 (MCM2 vs. IgG) at the c-MYC origin in DMSO-treated K562 cells, and non-significant (n.s.): P = 0.4598 (MCM2 vs. IgG) at the c-MYC origin in DRB-treated K562 cells. No occupancy of MCM2 was detected upstream c-MYC origin in both control and DRB-treated cells. d RIP with anti-MCM2 shows no association between c-MYC ANCOR with MCM2. Bars indicate mean ± SD from three independent biological replicates (n = 3). A paired Student’s t test (two-tailed) indicated no significant difference (n.s.; P = 0.9949) between MCM2 and IgG at the c-MYC origin. e Heatmap of the K562 Repli-seq signal in G1/G1b, S1, S2, S3, S4, and G2 cell cycle fractions (GSE34399) centered at the ANCOR-H2A.Z predicted 46,360 early DNA replication origins in K562 show stronger signal intensity in the G1 and S1/S2. f Heatmap showing analysis of Repli-seq signal centered at the replication peaks (GSE34399) in K562 cells appears randomly distributed. Source data are provided as a Source Data file.

DRB treatment of another cell line, HL60, resulted in impairment of DNA synthesis, followed by a drastic reduction in the S phase cell population to nearly 30% in DRB-treated cells as compared to 50% in the mock control cells (Supplementary Fig. 5i).

Lastly, using a publicly available dataset (GSE34399) of nascent DNA sequencing (Repli-seq), obtained from K562 sorted at different cell cycle stages (G1/G1b, S1, S2, S3, S4, G2), we interrogated the distribution of early replication origins associated with ANCORs and enriched in H2.AZ against all the replication origins captured by Repli-seq (Fig. 5e, f). Briefly, the Repli-seq dataset was intersected with our dataset of origins initiated by ANCORs (Fig. 5e). Stronger signal intensity was detected in the G1 and S1/S2 cell fractions when the analysis was computed with respect to the ANCOR predicted 46,360 DNA replication origins, while it gradually declined in the later phases (Fig. 5e). In contrast, the signal appeared randomly distributed across the different cell-fractions when the signal was not centered across the ANCOR predicted replication origins but around the replication peaks⁶⁰ (Fig. 5f). These results are consistent with the hypothesis that ANCOR enriched loci mark early DNA replication origins.

Taken together, our data demonstrate a global interaction of RNAs transcribed in G1/S phase and ORC1, in association with H2A.Z that provides a tool for the detection of early replication origins across the genome. On a related note, these findings support the hypothesis that ANCORs are involved in the initiation of DNA replication.

ORC1 engagement and H2A.Z enrichment is lost upon inhibition of G1/S phase transcribed RNAs

The changes in cell cycle progression upon DRB treatment prompted us to compare ORC1 engagement with respect to H2A.Z distribution across the 46,360 predicted early DNA replication origins (Fig. 4c, d). We performed these experiments in synchronized cells and applied inhibition of transcription by DRB in the first 3 h after release from G1/S. Blocking early G1/S phase RNAs resulted in a significant reduction of ORC1-RNA interactions as compared to the DMSO control (P value < 2.2e^‒16) and corresponded to depletion of H2A.Z (P value = 4.473e^‒14) at the respective loci (Fig. 6a–c). Importantly, when inspecting the c-MYC locus, we confirmed that DRB-mediated inhibition of transcription resulted in depletion of ORC1-RNA interactions and H2A.Z enrichment at the c-MYC origin (Fig. 6d and Supplementary Fig. 6a, b). This also held true for the FXN locus that clustered with the lowest expressing group (Fig. 4b). Indeed, the expected decrease in ORC1-RNA interactions associated with reduced enrichment of H2A.Z at the FXN origin upon inhibition of transcription (Fig. 6c). Next, we stratified the K562 transcriptional profiles at the earlier (S1) and later (S2) timepoints after the release in S phase centering the analysis on the ANCOR-predicted replication origins over a distance of +/−10 kb from the TSS. Consistently, we detected higher transcriptional activity in the S1 timepoint as compared to the S2 timepoint (P value = 0.02993), corroborating the hypothesis that ANCOR-associated loci mark early replication origins (Fig. 6e).

Fig. 6. Impaired ORC1 engagement and loss of H2A.Z upon inhibition of ANCORs.

a Heatmaps showing a decrease in ORC1 enrichment by RIP-seq (a) and H2A.Z enrichment by ChIP-seq (b) at 46,360 predicted early DNA replication origins upon DRB treatment as compared to DMSO mock control in K562 cells. Welch’s t test: P < 2.2e^‒16; Welch’s t test, P: 4.473e^‒14. c Box plots representing normalized read counts calculated for (a, b) using bamliquidator. Differences in normalized read counts for ORC1-RIP (left panel, P < 2.2e^‒16) and H2A.Z-ChIP (right panel, P: 4.473e^‒14) at 46,360 predicted early DNA replication origins upon DRB treatment as compared to DMSO mock control in K562 cells were assessed by Welch two-sample t test, two-sided was calculated. Multiple-comparison adjustment: none. Box plots show the median (center line), 25th and 75th percentiles (box), and whiskers extending to the most extreme values within 1.5 × IQR; overlaid points represent individual observations. d Genomic snapshots of the c-MYC and FXN loci depicting the impact of pharmacological inhibition of transcription using DRB on ORC1 and H2A.Z enrichment at the respective origin sites. The transcriptional profile of the c-MYC and FXN loci for the S1 and S2 time points are presented. The rose gold bar indicates the well-documented c-MYC origin (4, 6), and the predicted origin by the ANCORs association for FXN. e Heatmap showing the transcriptional profiles of K562 at S1 (2 hr) and S2 (5 hr) centered on ANCOR-predicted replication origins within +/−10 kb of the transcription start site (TSS). The higher transcriptional activity in S1 supports the association of ANCOR-associated loci with early DNA replication origins. Significance of figures is indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and ns, not significant.

Based on this initial investigation, we extended the analysis to identify a subset of “high-confidence” replication origins where DRB had the most pronounced effect on both H2A.Z and ORC1 enrichments. A more detailed analysis was performed using the following filtering criteria: (1) for ORC1 depletion by DRB, we selected genomic sites by applying a fold change threshold of >1.25 (Control vs. DRB); (2) In parallel, sites showing reduced H2A.Z enrichment upon DRB treatment were retained using the same threshold (> 1.25). Only regions satisfying both criteria 1 and 2 simultaneously were retained and defined as high-confidence early DNA replication origins. This filtering strategy identified a total of 19,230 origin sites (Fig. 7a, b).

Fig. 7. Gene expression dynamics and ORC1/H2A.Z redistribution at early replication origins following DRB treatment.

a, b Heatmaps showing gene expression profiles of K562 at S1 (2 h) and S2 (5 h) of genes with high-confidence predicted early DNA replication origins within +/− 10 kb from TSS (N = 6913). Genes were stratified into two subgroups based on log₂(Fold Change): a increased expression at S1 (log₂(Fold Change) S1 vs. S2 > 1; n = 3545). Welch two-sample t test, two-sided was employed, obtaining P: 9.739e^‒05 and b decreased expression at S1 (log₂(Fold Change) S1 vs. S2 < 1; n = 3368). Welch two-sample t test, two-sided was employed, obtaining P: 2.657e^‒12. Expression levels (log₂(FPKM) values) for each subgroup were plotted, illustrating transcriptional dynamics of early replication origin–associated genes. Heatmaps showing decreased ORC1 recruitment by RIP-seq (c) and H2A.Z enrichment by ChIP-seq (d) at high confidence 19,230 predicted DNA replication origins upon DRB treatment versus DMSO. Box plots representing normalized read counts calculated for (a, b) using bamliquidator. Differences in normalized read counts for ORC1-RIP (e), P: 0 and H2A.Z-ChIP (f), P: 0 at high confidence 19,230 predicted DNA replication origins upon DRB treatment versus DMSO mock control in K562 cells were assessed by Welch two-sample t test. Two-sided was calculated. Multiple-comparison adjustment: none. Box plots show the median (center line), 25th and 75th percentiles (box), and whiskers extending to the most extreme values within 1.5 × IQR; overlaid points represent individual observations. Significance of figures is indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and ns, not significant.

We then centered the ORC1 RIP-seq and H2A.Z ChIP-seq signals on these high-confidence early DNA replication origins, revealing the most pronounced differences in signal intensity between control and DRB-treated conditions (Fig. 7e, f).

We ultimately focused on characterizing the high-confidence early DNA replication origins, analyzing their genomic distribution relative to transcription start sites (TSSs) of protein-coding genes. The number of origins was quantified within incremental distance windows from TSSs: ±100 kb, ±50 kb, ±20 kb, ±10 kb, ±5 kb, ±2 kb, ±1 kb, and ±0.5 kb, with origins within each window annotated accordingly (Supplementary Fig. 6c, d). Notably, a significant fraction of these early replication origins is located in close proximity to gene promoters and TSSs, highlighting a strong association between early replication origin activity and transcriptional regulatory regions. Finally, we examined RNA-seq signal enrichment at the 19,230 high-confidence early DNA replication origins, extending transcriptional profiles ±5 kb from the origin centers. This analysis revealed a marked reduction in RNA-seq signal at the S2 time point relative to S1, indicating decreased transcriptional activity. The elevated transcriptional activity observed at S1 further supports the association of ANCOR-linked loci with early-replicating regions (Supplementary Fig. 6e, f). The newly identified ANCOR coordinates were used to interrogate the asynchronous K562 nascent RNA-seq dataset and ultimately confirmed that the ANCOR set is conserved while exhibiting variable peak intensities (Supplementary Fig. 7a–c).

These findings demonstrate that inhibition of transcription during the early S phase impairs both ORC1 engagement and H2A.Z enrichment, pointing to an RNA-mediated mechanism coordinating the formation of early replication origins.

Disruption of ANCORs perturbs DNA replication dynamics

Does the disruption of ANCOR’s transcription in G1/S phase destabilize the origin formation and the replication progression overall? To address this question, we monitored the replication fork dynamics genome-wide by leveraging DNA fiber analysis at single-molecule resolution^61,62 (Fig. 8a). Synchronized K562 cells were treated with DRB for 3 h upon release into S phase, while sequential additions of 5-iodo-2′-deoxyuridine (IdU) and 5-chloro-2′-deoxyuridine (CldU)⁶² were included in the medium to follow the progression of the replication forks during transcriptional inhibition (Fig. 8a). Upon ANCOR downregulation, nasDNA production was impaired and both DNA replication initiation and fork progressions were decreased in DRB-treated K562 cells as compared to DMSO-treated cells (Fig. 8b, c).

Fig. 8. DNA Replication dynamics is perturbed by the inhibition of ANCOR levels.

a Depicted the experimental design for DNA fiber analysis. K562 cells synchronized at G1/S phase by double thymidine block were treated with DRB for 3 h upon release into the S phase, and sequential additions of 5-iodo-2′-deoxyuridine (IdU) and 5-chloro−2′-deoxyuridine (CldU) were used to follow the progression of replication forks during transcriptional inhibition. b Quantification of DNA fibers at the “origin initiation” and “origin termination” in K562 DRB-treated cells as compared to DMSO; The mean ±  SD is calculated based on the number of labeled molecules for the replicating event, n = 47 technical replicates in the initiation and n = 33 technical replicates in the termination. c Quantification of DNA fibers during the two consecutive pulses with IdU and CldU, respectively, for K562 DRB-treated cells compared to DMSO. The mean ±  SD is calculated based on the number of labeled fibers collected, n = 99 technical replicates for both pulses. An unpaired Student’s t test (two-tailed), ****P value < 0.0001 (DMSO control vs DRB in the first pulse, IdU) and ****P value < 0.0001 (DMSO control vs DRB in the second pulse, CldU). d Single-molecule analysis of replicated DNA (SMARD) performed on FXN locus. Origin initiation, DNA replication fork progression (3’ to 5’ and 5’ to 3’), and origin termination were monitored upon DRB treatment and DMSO control. The percentage of molecules with IdU incorporation is indicated in the column chart below each respective condition. A two-tailed Student’s t test was performed on these data and confirmed the statistical significance of the observed differences between experimental conditions. Source data are provided as a Source Data file.

To elucidate the impact of ANCOR transcriptional inhibition on DNA replication progression within a specific gene locus, we performed single-molecule analysis of replicated DNA (SMARD) on FXN, a locus amenable to this approach, unlike c-MYC (Fig. 8b, d)^9,63. Strikingly, DNA replication fork progression at the FXN locus was disrupted by DRB-mediated transcription inhibition. The abnormal replication pattern exhibited a reduced number of initiation events, altered replication progression from 5’ to 3’, and increased terminating events when compared to the DMSO control (Fig. 8d), thus confirming that ANCORs play a critical role in maintaining the integrity of the replication process.

Our findings reveal that downregulation of ANCORs halts the DNA replication program, ultimately leading to a broad and deleterious perturbation of DNA replication dynamics.

In conclusion, these results show that enforced expression of ANCORs promotes origin complex assembly and/or its stabilization at specific genomic loci, which impacts subsequent origin activation and DNA replication dynamics. This knowledge could lead to a potential strategy to modulate the origins' firing at both local and global levels in mammalian cells.

Discussion

DNA replication initiation needs to be molecularly and temporally coordinated to ensure efficient DNA synthesis and to preserve cell-specific replication dynamics. Accumulating evidence indicates that RNA contributes to multiple stages of DNA replication including initiation, origin firing, nucleosome deposition onto nascent DNA strands, and the establishment of origin-associated chromatin markers^27,64–67.

Notably, the histone variant H2A.Z was shown to be essential for origin licensing²⁷. Its deposition at replication origins correlates with the presence of additional origin markers and with dynamic changes in local chromatin architecture, including H4K20 methylation^68–70 and H3K4 and H3K9 trimethylation^64,66.

Our investigation examined the molecular steps facilitating activation of early DNA replication origins. Specifically, we focused on the unexplored contribution of RNAs transcribed in the G1/early S phase of the cell cycle as the masterminds coordinating the engagement of ORC1 (the first subunit of the pre-replication complex) with H2A.Z, enabling DNA replication in a timely and selective fashion.

Through the identification of ANCORs, we delved into the role of RNA in maintaining replication origin activity and provided evidence that it governs the initial events of DNA replication. We demonstrate that changes in the expression of ANCORs result in perturbation of the replication profile. Additionally, we describe a tool to regulate site-specific gene replication in human cells by means of RNA.

Our data show that by linking the first subunit of the origin recognition complex ORC1 and the histone variant H2A.Z, ANCORs control gene replication at precise genomic sites, which may open avenues for future translational biomedical applications. Indeed, control of gene replication “in loco” holds promise for precise intervention in diseases associated with dysfunctional DNA replication.

The lack of understanding of how DNA replication is initiated at specific sites has represented a major roadblock to the development of successful therapeutics for these diseases. Herein, we present evidence that RNA can be harnessed as a tool to control locus-specific replication in human cells.

GC-rich elements have been identified as integral components of replication origins in human cells. These GC-rich sequences serve as binding sites for key replication factors, including the origin recognition complex (ORC)³⁷. In unraveling the mechanisms governing mammalian DNA replication, we have identified a distinctive RNA sequence (RM9) embedded within the c-MYC origin (region 9–10) that can bind ORC1. With unrelated and GC-poor RNA (i.e., MYC1-3) or homologous DNA (dRM9) sequences, only a weak or no interaction was detected, respectively, indicating that ORC1-RNA complex formation has both sequence and structural requirements (Fig. 1d–f).

By examining a range of ORC1 conformations in silico, mainly to vary the orientation between the N- and C-termini, we modeled ORC1-RNA recognition. The ORC1 N- and C-termini include the BAH domain (1–200 aa) and AAA+ domain (201–782 aa, ATPases associated with diverse cellular activities)/winged-helix domain (WHD, 783–861 aa), respectively. At present, the complete architecture of the ORC1 subunit structure remains vaguely characterized owing to the presence of flexible regions, referred as disordered domains, that connect the N- and the C-terminal regions. These disordered domains within ORC1 as well as the conformational dynamics of its AAA+ domain have been considered functional to its role in inducing the replication machinery¹³. Our theoretical model demonstrates that when the BAH domain is part of the binding region, dRM9 lies in close proximity to ORC1 during the ORC1b-dRM9 and ORC1c-dRM9 simulations. When the BAH domain is not included in the binding site in the ORC1a state, dRM9 deviates from its initial position. By contrast, RM9 loses its fold when protein complex formation involves the BAH domain and efficient targeting when the BAH domain is displaced away from the interacting region (Fig. 1g, h).

Taken together, these data suggest that the BAH domain is not involved in complex formation with RNA and provide a model wherein the BAH domain needs to be structurally free to interact with the histones. In this proposed model, the ORC1a conformation exposes a specific binding site that perfectly fits a RM9 stem-loop structure but cannot accommodate a similar DNA structure. Accordingly, RNA would promote a long-range effect on the ORC1 dynamic by affecting the conformational equilibrium of the protein and targeting the ORC1a conformation in a pocket formed by ORC1 disordered and AAA+ domains. This change in the conformational equilibrium would then facilitate the engagement of the BAH domain for subsequent chromatin binding.

A more detailed structural analysis using the ORC1 mutant (R297E, S450Y, T593Y, R846E) revealed critical insights into RNA interactions. RM9 contains two GC-rich inverted repeats (Supplementary Fig. 3g) This structural feature suggests that RM9 RNA may fold into Z-form helix, or form Z-form DNA-RNA hybrid at the origin^48,71–74. Alternatively, it can fold into A-type RNA helix or A-type DNA-RNA hybrid with the origin. Either of these conformations can be preferentially recognized by ORC1. This suggests a high degree of specificity in the interaction, as failure of RM9 to adopt such structure does prevent engagement with key residues in the ORC1 lateral pocket, ultimately compromising the active conformation of ORC1 required for origin recognition (Supplementary Fig. 3d, e).

Pharmacological inhibition of transcription by DRB treatment yielded compelling evidence of a significant reduction in ORC1-RNA interactions and H2A.Z enrichment globally, consistent with the decrease in DNA replication initiation and fork progression detected by fiber analysis. Compared with control treatment, DRB treatment also disrupted DNA replication fork progression (reduced initiating events, altered replication progression from 5’ to 3’, and increased terminating events) at the FXN locus as captured by SMARD analysis in K562 DRB-treated cells as compared to the control (Figs. 6–8).

Consistently, the presence of other Pre-RC factors at origin site, is not detected when ANCOR levels were reduced upon DRB transcriptional inhibition indicating that lack of ORC1 engagement by ANCORs at the replication origins alters the Pre-RC at the c-MYC and FXN origins (Fig. 5a–c and Supplementary Fig. 5g). Although CDC6 was not detectable by ChIP in either untreated cells or cells treated with DRB for 3 h (Supplementary Fig. 7d), this result does not exclude its functional involvement in origin licensing but rather reflects its transient chromatin association compared with the more stably loaded MCM complex. However, an additional factor contributing to the reduced replication fork progression observed in DRB-treated cells could be the occurrence of transcription-replication conflicts. Previous studies have shown that pharmacological inhibition of CDC7 suppresses origin activation and is associated with increased replication fork speed, highlighting the interplay between origin usage and fork dynamics⁷⁵. In this context, altered transcriptional regulation may influence both origin activity and replication progression.

We also showed that loss of H2A.Z observed upon DRB treatment is not attributable to a global decrease in protein levels (Supplementary Fig. 5e, f). Rather, the local reduction in H2A.Z enrichment at origin sites suggests a loss of origin identity and impaired activation of ANCOR-regulated replication origins (Supplementary Fig. 6a, b). Long et al. have previously demonstrated the functional significance of H2A.Z in cell-cycle progression using an auxin-inducible degron (AID) system. Six hours after release from G2/M arrest, H2A.Z-depleted cells exhibited features consistent with a cell-cycle block at the G1/S transition, supporting the notion that H2A.Z is required for proper progression into S phase^27,28. A comparable phenotype was observed in HeLa cells treated with siRNA against H2A.Z, where nascent DNA labeling with BrdU and subsequent FACS analysis revealed accumulation at the G1/S boundary. The same study demonstrated that H2A.Z orchestrates genome-wide deposition of H4K20me2 and promotes ORC1 recruitment, as shown by ChIP-seq analysis^27,28.

Within the context of ANCOR-regulated origins, additional histone variants or nucleosomal components could also contribute to the local regulation of H2A.Z deposition⁷⁶. A complementary hypothesis proposes that ORC-mediated nucleosome positioning facilitates replication initiation⁶⁶. Indeed, studies in S. cerevisiae indicate that nucleosome distribution and higher-order chromatin structure are integral to replication origin function. The proximity of ORC binding sites to the nucleosome-free region at ARS1 implies that ORC may influence the arrangement of adjacent nucleosomes. Similarly, impaired ORC1 engagement at human origins may perturb H2A.Z deposition, thereby altering local chromatin architecture and compromising interactions with replication licensing factors^66,67.

Remarkably, locus-specific modulation of ANCORs via CRISPRi, CRISPRa, and ASOs confirmed corresponding changes in nasDNA synthesis (Fig. 3).

Our group further hypothesizes that the absence of H2A.Z enrichment at origin sites may reflect distinct molecular mechanisms operating during nucleosome deposition in the G1 and S phases. Global inhibition of transcription with DRB disrupts interactions between ORC1-RNA and consequently H2A.Z. As a result the native distribution of H2A.Z is lost and epigenetic memory is not preserved during DNA replication, as recently shown⁶⁵.

These results shed light upon the interplay between G1/S phase transcription, chromatin dynamics, and establishment of DNA replication origins, supporting the role of RNA in the formation of active replication origins.

Mas et al. have reported an interaction between RNAs transcribed at active origins and ORC1. They linked this interaction to ORC1’s dynamic association with chromatin and concluded that disruption of ORC1-RNA binding “has an in impact on replication origin activity”⁶⁴. However, there are substantial differences between our study and the one by Mas et al. First and foremost, the authors suggest a mechanism by which “RNA binding favors ORC1 chromatin release”⁶⁴. By contrast, we identify ANCORs as the recruiters of ORC1 and facilitators of replication origins in human cells. We delineate the RNA-mediated molecular events responsible for the activation of early replication origins. In addition, we reveal a longstanding question on how the distribution and activation of replication origins are controlled in mammalian cells. We also show that replication can be initiated in a locus-specific manner by induction of ad hoc RNA (ANCORs), with ANCOR-like transcripts reverting genetic abnormalities caused by dysfunctional DNA replication. Conversely, Mas et al. describe RNAs as behaving as a facilitator for the release of ORC1 from the chromatin. We instead propose that ANCORs act as a facilitator of the replication initiation. In simple terms, we posit that ANCORs control ORC1 engagement with H2A.Z at defined genomic loci, thereby contributing to the establishment of an origin-competent state at early replication origins. In this way, ANCORs facilitate proper organization of the origin-associated complex, supporting efficient replication initiation. At the subsequent stages, other RNAs could favor ORC1 release from the chromatin as already described in a previous study⁶⁴.

The molecular steps involved in the formation of DNA origins, as well as their modulation, support a hypothetical model wherein ANCORs coordinate the engagement of ORC1 with H2A.Z across the genome, thereby facilitating activation of early replication origins (Supplementary Fig. 7e). Despite the fact that, ANCOR-induced conformational changes in ORC1 depicted in the figure remain hypothetical and require further structural investigation to be validated, our model has significant therapeutic implications, i.e., an RNA-guided strategy to control DNA replication. Indeed, ANCOR-specific downregulation by CRISPRi as well as the potential use of ASO treatment severely impacted the activity of the c-MYC origin, whereas locus-specific upregulation of ANCOR by CRISPRa led to increased production of nasDNA, and a schematic of the proposed model is depicted in Supplementary Fig. 7e.

In summary, our study identifies ANCORs as facilitators in the formation of early DNA replication origins and alludes to the broader involvement of ANCORs in DNA replication initiation. These findings reveal an RNA-centered mechanism to control and modulate DNA replication in eukaryotic cells at specific sites, thereby providing a tool to treat diseases caused by DNA replication abnormalities.

Methods

List of cultured cell lines

The chronic myeloid leukemia K562 and acute myeloid leukemia HL-60 cell lines were procured from ATCC and cultured in RPMI-1640 media (Corning, Cat. No. 10-040-CM) supplemented with 10% fetal bovine serum (FBS) without antibiotics at 37 °C in a humidified atmosphere with 5% CO₂. HEK 293 cells obtained from ATCC were cultured in DMEM (Corning, Cat. No. 10-013-CV) with 10% FBS without antibiotics at 37 °C in a humidified atmosphere with 5% CO₂.

RNA and DNA isolation from human cell lines

Total RNA isolation was conducted following a previously described protocol⁷⁷.

Subsequently, all RNA samples utilized in this study were subjected to DNase I treatment (10 U of DNase I per 3 mg of total RNA) at 37 °C for 1 h in the presence of RNase inhibitor to remove any contaminating DNA. To ensure complete DNA removal, the RNA samples were further purified using acidic phenol (pH 4.3). For cDNA synthesis, either Random Primers (Invitrogen) or gene-specific primers were employed in conjunction with Reverse Transcriptase (Roche Applied Science), following the manufacturer’s guidelines. The resulting cDNA was purified using the High Pure PCR Product Purification Kit (Roche Applied Science)⁷⁷.

Nuclear RNA was isolated using a modified method based on the nuclei isolation protocols⁷⁸. Approximately 50 million viable cells were washed with ice-cold PBS containing 5 mM vanadyl complex and 1 mM PMSF. The cells were then resuspended in ice-cold lysis buffer composed of 1× Buffer A (10 mM HEPES-NaOH pH 7.6, 25 mM KCl, 0.15 mM spermine, 0.5 mM spermidine, 1 mM EDTA, 2 mM Na butyrate), 1.25 M sucrose, 10% glycerol, 5 mg/mL BSA, 0.5% NP-40, and supplemented with various protease inhibitors, vanadyl complex, and RNase inhibitor (RNAguard; Amersham Biosciences). The samples were incubated at 0 °C for ~10 min and homogenized through multiple strokes in a Dounce homogenizer. The pelleted nuclei were further processed and centrifuged, and the resulting nuclear RNAs were extracted. All RNA samples, including total, cytoplasmic, and nuclear RNA, were treated with DNase I to remove any contaminating DNA and then extracted with acidic phenol (pH 4.3). Polyadenylated and non-polyadenylated RNA fractions were separated using the MicroPoly(A)PuristTM purification kit (Ambion). Subsequently, cDNA synthesis was performed with Random Primers and Transcriptor Reverse Transcriptase following the manufacturer’s recommendations. The cDNA was purified using a High Pure PCR Product Purification Kit.

Genomic DNA isolation was performed following a previously described method³⁴.

Briefly, cell pellets were resuspended in lysis buffer containing 0.5% SDS, 25 mM EDTA (pH 8), 10 mM TRIS (pH 8), and 200 mM NaCl. After treatment with RNase A (Epicenter) at 37 °C for 20 min and Proteinase K (Roche) overnight at 65 °C, high-quality genomic DNA was extracted using Phenol-Chloroform (Sigma, pH 8) and precipitated overnight with isopropanol. The genomic DNA was then resuspended in TE buffer (pH 8) and stored at 4 °C.

Nascent RNA/DNA capture was performed using a Click-iT® Nascent RNA Capture Kit (ThermoFisher) according to the manufacturer’s instructions with minor modifications. Briefly, (1) Labeling the cells with 5-ethynyl uridine (EU)/Ethynyl 2-deoxyuridine (EdU): 200 mM EU or 30 mM EdU stock solutions were added to the cells to a final concentration 0.5 mM or 30 µM, respectively. (2) Incubation: The cells were incubated for 1 or 2 h. (3) RNA/DNA isolation: The cells were harvested, and the RNA/DNA were isolated and dissolved in 14 μL of H₂O. (4) Biotinylation of RNA/DNA by Click reaction. Click-iT® reaction cocktail (50 μL per reaction) was prepared according to the manufacturer’s instructions. A mixture containing 1× Click-iT EU buffer; 2 mM CuSO₄; 1 mM Biotin azide; 13.25 μL of the isolated RNA; 10 mM Click-iT® reaction buffer additive 1; and 12 mM Click-iT® reaction buffer additive 2 was prepared. After adding each component, the reaction cocktail was gently mixed. The addition of the Click-iT® reaction buffer additive 1 stock initiated the click reaction between the EU-RNA/EdU-DNA and biotin azide. Subsequently, the Click-iT® reaction buffer additive 2 was added and incubated for 30 min with gentle vortexing. (5) RNA/DNA precipitation: 1 μL of UltraPure™ Glycogen, 55 μL of 7 M ammonium acetate, and 750 μL of chilled 100% ethanol were added to the click reaction and incubated at –70 °C for at least 30 min. After centrifugation, the pellet was dissolved in 125 μL of H₂O. (6) Binding biotinylated RNA/DNA to Dynabeads® MyOne™ Streptavidin T1 magnetic beads (ThermoFisher): The RNA/DNA binding reaction mixture included: 125 μL 2xClick-iT® RNA binding buffer; 2 μL Ribonuclease Inhibitor or 2 μL of water for DNA; and 125 μL of the isolated biotinylated RNA/DNA. The RNA binding reaction mixture was heated at 68–70 °C for 5 min, and 50 μL of bead suspension was added to the heated RNA binding reaction mixture. The tube containing the RNA/DNA binding reaction was incubated at room temperature for 30 min while gently vortexed to prevent the beads from settling. The beads were immobilized using a magnet and washed five times with 500 μL of Click-iT® reaction wash buffer 1 and five times with 500 μL of Click-iT® reaction wash buffer 2. Finally, the beads were resuspended in 50 μL of Click-iT® reaction wash buffer 2, and the captured RNA was immediately reverse transcribed to cDNA. The captured DNA was released into 50 μL of boiling water and used in qPCR analyses. The primers used to determine nasDNA abundance on the c-MYC locus are listed in Supplementary Table 1.

qRT-PCR analysis performed to quantify gene expression levels and nasDNA abundance

Sybr Green analysis

iQ Sybr Green supermix (Biorad, Hercules, CA) was employed with the following thermal cycling parameters: an initial denaturation step at 95 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 15 s, annealing at 60 °C for 1 min, and extension at 72 °C for 1 min.

TaqMan analysis

Hotstart Probe One-step qRT-PCR master mix (USB) was used with the following temperature profile: an initial step at 50 °C for 10 min, a denaturation step at 95 °C for 2 min, followed by 40 cycles of denaturation at 95 °C for 15 seconds and annealing/extension at 60 °C for 60 s.

Primers and TaqMan probes used in real-time PCR

Eukaryotic 18S rRNA Endogenous Control: ABI Cat. # 4310893E, Human GAPDH Endogenous Control: ABI Cat. # 4310884E, Human MYC mRNA: ABI Cat. # Hs00153408_m1, c-MYC ANCOR custom design (FRW-Primer: TCTGGGTGGAAGGTATCCAA, TaqMan probe: CCAACAAATGCAATGGGAGT, REW-Primer: TTGGAGAGCGCGTTATGAAT).

Primers used for strand-specific real-time RT-PCR (Sybr Green)

Reverse Transcriptase primer for c-MYC ANCOR: 5’- AAC CGC ATC CTT GTC CTG TGA GTA-3’; qRT-PCR primers: Forward: 5’-ACA GGC AGA CAC ATC TCA GGG CTA-3’, Reverse: 5’-ATA GGG AGG AAT GAT AGA GGC ATA-3’.

Antibodies

Rabbit polyclonal anti-ORC1 (Abcam Catalog No. ab85830); goat anti-rabbit (Invitrogen, Catalog No. 32460); rabbit polyclonal human Histone H2A.Z – ChIP Grade (Abcam Catalog No. ab4174); mouse monoclonal beta Actin (Santa Cruz, sc-81178); goat anti-mouse (Invitrogen 32430); rabbit monoclonal c-MYC (Cell Signaling Technology, D84C12); anti-CRISPR-Cas9 antibody [7A9-3A3] (Abcam, Ab191468), MCM2 (D7G11) XP® (Rabbit mAb Cell Signaling Catalog No. 3619S); MCM7 (Thermo Fisher Catalog No. PA579651); Anti-Cdc6 N-terminal (abcam Catalog No. ab188423); ORC2 (3G6) (Santa Cruz Biotechnology Catalog No. sc-32734).

Recombinant proteins

Recombinant human origin recognition complex subunit 1 (ORC1) (Cusabio Catalog No. CSB-BP621667HU(A4)); Human ORC1 (AA 1–200, Q13415 Cusabio Catalog No. CSB-BP621667HU1); Human ORC1 (AA 201–782, Q13415 Cusabio Catalog No. CSB-BP621667HU2); Human ORC1 (AA 783–861, Q13415 Cusabio Catalog No. CSB-BP621667HU3); Recombinant human origin recognition complex subunit 1 (ORC1, R297E, S450Y, T593Y, R846E) (Cusabio Catalog No. CSB-BP621667HU(A4)(M)); Recombinant mononucleosome H2A.Z (Active Motif, Cat. No. 81072); Recombinant Histone H2A.Z (Active Motif, Cat No. 31293);

RNA pull-down assay (RNAP)

The RNA probes were custom-designed to investigate ORC1 binding to RNA structures transcribed within the c-MYC locus, including the c-MYC promoter, c-MYC gene, and c-MYC origin. The specific set of synthetic biotinylated RNA oligonucleotides used in the RNAP is listed in Supplementary Table 2.

RNAP was conducted in triplicate following a modified protocol adapted from Pierce™ Magnetic RNA-Protein Pull-Down Kit (Thermo Scientific™, No.: 20164). Briefly, 3 µg of biotinylated RNA was denatured at 90 °C for 2 min and then allowed to fold into proper RNA structures at 25 °C for 20 min. In total, 3 × 10⁶ cells were resuspended in 1 ml of RIP Buffer containing 150 mM KCl, 25 mM Tris pH 7.4, 0.5 mM DTT, 0.5% NP40, 1 mM PMSF, and protease inhibitor (Roche Complete Protease Inhibitor Cocktail Tablets). Cells were homogenized with a homogenizer for 20 strokes, and debris was pelleted and discarded by centrifugation at 15,871×g (13,000 rpm) for 10 min. 0.1 µg of biotinylated RNA was incubated with 1 mg of cell lysate at room temperature for 1 h. In total, 60 µl of streptavidin agarose beads (Invitrogen) were added to the mix and incubated at room temperature on a shaker. The beads were then washed ten times with RIPA buffer and boiled at 100 °C in 30 μl of SDS sample buffer. Finally, the beads were removed using a magnetic field, and the samples were analyzed by western blot.

Cell synchronization and drug treatment

To enrich cells at the G1/S boundary and early S phase, the double Thymidine block method was employed as previously described^{34,56,79–81}. Briefly, K562 and HL-60 cells were grown overnight to 70–80% confluence, washed twice with 1x PBS, and then cultured in DMEM (10% FBS) supplemented with 2.5 mM thymidine for 18 h (first block). Subsequently, thymidine was washed out with 1×PBS, and the cells were cultured in DMEM (10% FBS). After 8 h, the cells were treated with thymidine again for 18 h (second block). Cells were then released from the block as previously described^{34,56,79–81}.

K562’s cell cycle synchronization in G1 phase was achieved by incubating the cells for 24 h with Nocodazole (15 mM) as previously described⁸².

G1 synchrony was assessed 2–3 h after release from the G2/M block by flow cytometry analysis after propidium iodide 1 mg/mL (Sigma)- or 7AAD vs BrdU cell staining, using a LSRII flow cytometer (BD Biosciences) at the Harvard Stem Cell Institute/Beth Israel Deaconess Medical Center flow cytometry facility.

DRB treatments were carried out as previously described³⁴. Briefly, after release from double thymidine block, cells were treated with 100 μM of DRB (Sigma-Aldrich).

EU (5-ethynyl uridine, to label nasRNA) and or EdU (Ethynyl 2-deoxyuridine to label nasDNA) were purchased from Click Chemistry Tool (CCT) and added to cell cultures at a final concentration of 0.5 mM (EU) or 30 µM (EdU). Cells were cultured for 1 h with EU and 40 min with EdU to achieve maximum labeling of nasRNA and nasDNA.

BrdU staining and flow cytometric analysis of the G1/S phase cycle

Minor modifications were made to the manufacturer’s instructions provided in a BD Pharmingen BrdU Flow Kit (Catalog No. 559619). Briefly, K562 cells were cultured in cell culture medium and incubated with BrdU at a final concentration of 10 µM for a pulsing period of 45 min. After the pulsing period, cells were resuspended in 100 µL of BD Cytofix/Cytoperm Buffer and washed with 1 mL of 1× BD Perm/Wash Buffer. The cells were then incubated with BD Cytoperm Permeabilization Buffer at all steps, keeping the samples on ice to facilitate fixation and permeabilization of the cell membranes. Following permeabilization, the cells were resuspended in 100 µL of diluted DNase, which was prepared at a concentration of 300 µg/mL in DPBS. This resulted in a dilution of 30 µg of DNase per 106 cells. The cells were incubated with DNase for 1 h at 37 °C. After the incubation, a wash step with 1 mL of 1× BD Perm/Wash Buffer was performed. To detect BrdU and total DNA content, the cells were stained with a Fluorochrome-conjugated anti-BrdU Antibody and 7-aminoactinomycin D (7-AAD) fluorescent dye following the manufacturer’s instructions of BD Pharmingen BrdU Flow Kit (Catalog No. 559619), for recommended concentrations and dilutions of the antibodies. The staining was carried out for 20 min at room temperature. The stained cells were then analyzed using a flow cytometer to determine the incorporation of BrdU and DNA content. Cell acquisition and analysis were performed on BD LSRFortessa (BD Biosciences, Franklin Lakes, NJ, USA) using BD FACSDivaTM software (BD Bioscience, Franklin Lakes, NJ, USA). Analysis was performed using Flowjo software (Flowjo LLC, Ashland, OR, USA).

Protein expression analysis by western blotting

ORC1

In total, 0.2 × 10⁶ cells were lysed in SDS Sample buffer, separated on 8% SDS-PAGE gels, and transferred to a nitrocellulose membrane. Immunoblots were stained overnight at 4 °C with primary anti-ORC1 (1 µg, 1:1000 TBST/5% milk; Abcam, ab85830). Immunoblots were then incubated in secondary goat anti-rabbit (0.001 µg, 1:1000 TBST/5% milk; Invitrogen, 32460) for 1 h at room temperature. Immunoreactive proteins were detected using the Pierce^® ECL system (Thermo Scientific #32106).

H2A.Z

Whole-cell lysates from ~0.2 × 10⁶ cells per sample were separated on 13% SDS-PAGE gels and transferred to a nitrocellulose membrane. The membrane was blocked with TBST/5% BSA for 4 h. Immunoblots were stained overnight at 4 °C with the primary anti-H2A.Z (1 µg, 1:2000; Abcam ab4174). The immunoblots were then incubated with secondary goat anti-rabbit (0.001 µg 1:4000; Invitrogen 32460) for 1 h at room temperature with TBST/5% BSA. Immunoreactive proteins were detected using the Pierce^® ECL system (Thermo Scientific #32106).

c-MYC

Cells (0.2 × 10⁶) were lysed using SDS Sample buffer, and the lysates were then subjected to separation on an 8% SDS-PAGE gel, followed by transfer to a nitrocellulose membrane. The membrane was incubated overnight at 4 °C with the primary antibody mAb c-MYC (1 µg, dilution 1:1000 in TBST/5% milk; Cell Signaling Technology, D84C12). Subsequently, the membrane was incubated with secondary goat anti-rabbit antibodies (0.001 µg, dilution 1:1000 in TBST/5% milk; Invitrogen, 32460) for 1 h at room temperature. Immunoreactive proteins were visualized using the Pierce® ECL system (Thermo Scientific #32106).

Cas9

In total, 0.2 × 10⁶ cells were lysed in SDS sample buffer, and the resulting lysates were separated on an 7% SDS-PAGE gel, followed by transfer onto a nitrocellulose membrane. The membrane was incubated overnight at 4 °C with the anti-CRISPR-Cas9 antibody [7A9-3A3] (1 µg, 1:300 dilution in TBST with 3% milk; Abcam, Ab191468). It was then incubated for 1 h at room temperature with a goat anti-Rabbit IgG H&L (HRP) secondary antibody (0.001 µg, 1:5000 dilution in TBST with 3% milk, Abcam, ab6721); Immunoreactive proteins were detected using the Pierce® ECL system (Thermo Scientific, #32106).

β-actin

Stripped membranes were stained with β-actin antibody (ACTBD11B7) (1 µg, 1:3000 dilution in TBST with 5% milk, Santa Cruz sc-81178). Membranes were then incubated for 1 hour at room temperature with a secondary goat anti-mouse IgG H&L (HRP) antibody (0.001 µg, 1:5000 dilution in TBST with 5% milk, Abcam, Ab6789).

Electrophoretic gel mobility shift assays (EMSAs) to investigate ORC1–RNA and H2A.Z–RNA interactions

DNA and RNA oligonucleotides were end-labeled with [γ-³²P] ATP (Perkin Elmer) and T4 polynucleotide kinase (New England Biolabs). Reactions were incubated at 37 °C for 1 h and then passed through G-25 spin columns (GE Healthcare) according to the manufacturer’s instructions to remove unincorporated radioactivity. Labeled samples were gel-purified on 10% polyacrylamide gels. Binding reactions were carried out in 10 µl volumes in the following buffer: HEPES-NaOH 25 mM, Magnesium Acetate 10 mM, Sodium Glutamate 100 mM, EDTA 0,1 mM, DTT 1 mM, ATP 3 mM, RNase inhibitor 20U, Glycerol 5%³⁷. Full-length purified ORC1 (Cusabio Catalog N. CSB-BP621667HU(A4)) or Recombinant Histone H2A.Z (Active Motif, Cat. No. 81072) proteins were incubated with 1.1 nM of ³²P-labeled single-stranded RNAs. All reactions were assembled on ice and then incubated at room temperature for 30 min. Samples were loaded onto 6% native polyacrylamide gels (0.5×TBE) at 4 °C and ran for 3 h at 140 V. All gels were fixed (20% methanol, 10% acetic acid), dried, and exposed to X-ray film. RNA oligonucleotides are listed in Supplementary Table 3.

LightShift chemiluminescent RNA EMSA

All series LightShift® Chemiluminescent RNA EMSA assays were performed according to the manufacturer’s instructions (Thermo Fisher Scientific, Cat. No. 20158).

This kit enables chemiluminescent detection with sensitivity comparable to that of radioactive methods. Custom-designed RNA probes were generated to assess ORC1 binding to RNA structures transcribed within the c-MYC locus, including the c-MYC origin (RM9) and the c-MYC gene region (MYC1). The specific biotinylated synthetic RNA oligonucleotides used in the Chemiluminescent RNA EMSA are in Supplementary Table 4. In addition, Chemiluminescent RNA EMSA assays were employed to assess the binding affinity and kinetics of native ORC1, as well as to analyze the structural contribution of individual ORC1 domains in interaction with RM9. These assays were also used to examine the loss of RNA-binding ability in the ORC1 mutant (R297E, S450Y, T593Y, R846E).

ORC1, ANCOR, H2A.Z in vitro reconstitution and ternary complex formation analysis by EMSAs

The c-MYC ANCOR (RM9) oligonucleotide was end-labeled with [γ-³²P] ATP (PerkinElmer) using T4 polynucleotide kinase (New England Biolabs). The labeling reaction was incubated at 37°C for 1 h and subsequently passed through G‑25 spin columns (GE Healthcare) according to the manufacturer’s instructions to remove unincorporated radioactivity material. Labeled samples were gel-purified on 10% polyacrylamide gels. Binding reactions were carried out in 15 µl volumes in the following binding buffer: HEPES-NaOH 25 mM, Magnesium Acetate 10 mM, Sodium Glutamate 100 mM, EDTA 0,1 mM, DTT 1 mM, ATP 3 mM, RNase inhibitor 20U, Glycerol 5%)³⁷. In all, 1 µg of full-length purified ORC1 (Cusabio Catalog N. CSB-BP621667HU(A4)) and H2A.Z (Active Motif, Cat. No. 81072) proteins were incubated with ~1.1 nM of ³²P-labeled single-stranded RNAs. A second incubation was carried out using: rabbit polyclonal human Histone H2A.Z (2 µg, Abcam Catalog No. ab4174) and rabbit polyclonal anti-ORC1 (2 µg, Abcam Catalog No. ab85830). Samples were loaded onto 6% native polyacrylamide gels (0.5xTBE) at 4 °C and ran for 3 h at 140 V. All gels were fixed (20% Methanol, 10% Acetic acid), dried, and exposed to phosphorimaging plates between 24 and 48 h. The signal was acquired using Amersham™ Typhoon™ biomolecular imagers (#06230697). RNA oligonucleotides are listed in Supplementary Table 3. The upper and lower bands were excised from the gel for subsequent mass spectrometry analysis to verify the co-presence of ORC1 and H2A.Z within the complex (Supplementary Fig. 1f).

Sample preparation and ERLIC separation for Mass spectrometry analysis

Each sample was submitted for a single LC-MS/MS experiment that was performed on Orbitrap QE240 (Thermo Fisher) equipped with Waters (Milford, MA) NanoAcquity HPLC pump. Peptides were separated onto a 100-µm inner diameter microcapillary trapping column packed first with 5 cm of C18 Reprosil resin (5 µm, 100 Å, Dr. Maisch GmbH, Germany) followed by analytical column 15 cm of PepSep (Bruker, Germany). Separation was achieved through applying a 5–27% ACN gradient in 0.1% formic acid over 60 min at 400 nl min⁻¹. The Orbitrap QE240 was operated in data-dependent mode for mass spectrometry methods. The mass spectrometry survey scan was performed in the Orbitrap within the 400–800 m/z range at a resolution of 6 × 10⁴, followed by the selection of the twenty most intense ions (TOP20) for MS2 fragmentation in the Ion trap, using a precursor isolation width window of 2 m/z, AGC setting of 10,000, and a maximum ion accumulation of 100 ms. Singly charged ion species were not subjected to fragmentation. Normalized collision energy was set to 30 V and an activation time of 10 ms. Ions within ±10 ppm of selected MS2 precursor ions were dynamically excluded from fragmentation for 60 s.

Mass spectrometry analysis

Raw data were submitted for analysis in PEAKS 13 studio (BSI, Canada) software. Assignment of MS/MS spectra was performed using the Sequest HT algorithm by searching the data against a protein sequence database including all entries from the Human Uniprot database (SwissProt 16,768) and other known contaminants such as human keratins and common lab contaminants. Searched were performed using a 10 ppm precursor ion tolerance and requiring each peptides N-/C termini to adhere with Trypsin protease specificity, while allowing up to two missed cleavages. Methionine oxidation ( + 15.99492 Da) was set as a variable modification. A MS2 spectra assignment false discovery rate (FDR) of 1% on the protein level was achieved by applying the target-decoy database search. Filtering was performed using a Percolator 64 bit version⁸³.

Chromatin immunoprecipitation (ChIP) and nascent chromatin immunoprecipitation (nasChIP) to investigate the Origin firing

ChIP was performed as follows. Cells were crosslinked with 1% formaldehyde for 10 min at room temperature. Pellets of 1 × 10⁶ cells were used for immunoprecipitation as previously described^32,34 and lysed for 10 min on ice, and chromatin was fragmented using a Bioruptor Diagenode sonicator for 35 cycles (1 pulse each 30 s). Each ChIP was performed with 4 µg of antibody, incubated overnight at 4 °C. A 50/50 slurry of protein A and protein G Dynabeads was used to capture enriched chromatin, which was then washed before reverse-crosslinking and proteinase K digestion at 65 °C. Immunoprecipitated DNA was analyzed by ChIP-PCR analysis and ChIP-seq. The following antibodies were used for ChIP: ORC1 (4 µg, Abcam, ab85830), H2A.Z (4 µg, Abcam ab4174, lot GR3176820-1), IgG (4 µg, Abcam ab171870), MCM2 (D7G11) (4 µg, Cell Signaling #3619), MCM7 (4 µg, Thermo Fisher, PA5-79651), and Cdc6 - N-terminal (4 µg, Abcam, ab188423). Fold enrichment was calculated using the formula:

1. $2\left(-\mathrm{\Delta }\mathrm{\Delta }Ct\left(ChIP/non-immuneserum\right)\right)$

Primer sets used for ChIP are listed in Supplementary Table 5.

RNA immunoprecipitation (RIP) to investigate RNA-binding protein interactions

The RIP protocol was conducted following the established procedure³⁴ with slight modifications. In summary, 6 × 10⁶ K562 cells were crosslinked with 1% freshly made formaldehyde solution (50 mM HEPES-KOH; 100 mM NaCl; 1 mM EDTA; 0.5 mM EGTA; 11% formaldehyde) for 10 min at room temperature. Crosslinking was stopped by adding 1/10th volume of 2.66 M Glycine, kept for 5 min at room temperature and 10 min on ice. Cell pellets were then washed twice with ice-cold PBS (freshly supplemented with 1 mM PMSF). Cell pellets were resuspended in cell lysis buffer (volume = 4 mL), 1× Buffer (10 mM Tris pH 7.4; 10 mM NaCl; 0.5% NP-40), freshly supplemented with protease inhibitors (protease inhibitor cocktail: Roche Applied Science, Cat. No. 1836153, 1 tablet/375 μL H₂O; add as ×100), 0.1 mM PMSF, and 0.2 mM vanadyl complex (NEB). Cells were incubated at 0 °C for 10–15 min and homogenized in a Dounce (10 strokes pestle A and 40 strokes pestle B). Nuclei recovered by centrifugation at 375×g (2000 rpm) for 10 min at 4 °C. Nuclei were resuspended in 3 ml of 1× Resuspension Buffer (50 mM HEPES-NaOH, pH 7.4; 10 mM MgCl₂) supplemented with 0.1 mM PMSF and 0.2 mM vanadyl complex. 8. DNase I treatment (250 U/ml) was performed for 30 min at 37 °C, and EDTA (final concentration 20 mM) was added to stop the reaction. Resuspended nuclei were sonicated once for 30 s (1 pulse of 30 s) using Bioruptor Diagenode.

Immunoprecipitation was performed as follows: Before preclearing, the sample was adjusted with 1% Triton X-100; 0.1% sodium deoxycholate; 0.01% SDS; 140 mM NaCl; Protease inhibitors; 0.2 mM vanadyl complex; 0.1 mM PMSF. Preclearing step: ~50 μL magnetic beads (Protein A or G Magnetic Beads; #S1425S or #S1430S NEB) were added to the sample, and incubation was carried out for 1 h on a rocking platform at 4 °C. Beads were removed in the magnetic field. The sample was then divided into aliquots according to different conditions and antibodies of interest: (1) H2A.Z (5.5 µg, Abcam ab4174, lot GR3176820-1), ORC1 (5.5 µg, Abcam, ab85830), ORC1 (5.5 µg, Abcam, ab60), ORC2 (3G6) (5.5 µg, Santa Cruz Biotechnology, sc-32734),  (2) preimmune serum: IgG (5.5 µg, Abcam ab171870); (3) no antibody, no serum (input). Antibody or preimmune serum was added to the respective aliquot, and incubation was performed on a rocking platform overnight at 4 °C. Input was stored at −20 °C after the addition of SDS to 2% final concentration. Day II. 200 μL of Protein A-coated super-paramagnetic beads (enough to bind 8 μg IgG) were added to the samples and incubated on a rocking platform for 1 h at 4 °C. Six washes of beads in the magnetic field were performed with immunoprecipitation buffer (150 mM NaCl; 10 mM Tris-HCl, pH 7.4; 1 mM EDTA; 1 mM EGTA pH 8.0; 1% Triton X-100; 0.5% NP-40 freshly supplemented with 0.2 mM vanadyl complex and 0.2 mM PMSF) in a magnetic field. Proteinase K treatment to release DNA/RNA into solution and to reverse the crosslinking was performed in 200 μL of: 100 mM Tris-HCl, pH 7.4; 0.5% SDS for the immunoprecipitated samples and in parallel for the input using 500 μg/mL of Proteinase K at 56 °C overnight. Day III. Beads were removed in the magnetic field. Immunoprecipitated RNAs were further purified by extraction with Phenol (pH 4.3) after addition of NaCl (0.2 M final concentration). Ethanol precipitation (in the presence of glycogen); 3 h at −20 °C. The pellet was dissolved in 180 μL H₂O, heated at 72 °C for 2 min, and immediately chilled on ice. Samples were treated with DNase I (20 units) in the presence of RNase inhibitor at 300 U/ml in ×1 buffer # 2 (NEB) at 37 °C for 30 min. Phenol (pH 4.3) extraction and EtOH precipitation were repeated. The RNA pellet was dissolved in 50 μL H₂O and used in qRT-PCR (Primer sets used for RIP-PCR are listed in Supplementary Table 5.

Downregulation of c-MYC ANCOR by orthogonal approaches

dCAS9 -mediated stable and inducible downregulation

K562 and HEK 293 cell lines were maintained in appropriate culture media supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin before lentiviral transduction. The guide RNA (gRNA) targeting the c-MYC ANCOR in a region located 1.4 kb upstream of the c-MYC origin was designed using the online tool provided by the Broad Institute (https://portals.broadinstitute.org/gpp/public/analysis-tools/sgrna-design), and an initial guide screening was conducted. To implement CRISPR interference (CRISPRi), the lentiviral vector pdCAS9-humanized (code 44246) with an engineered mCherry fluorescent signal was acquired from Addgene and transduced into K562 cells, following the protocol outlined in ref. ⁴⁹. Four lentiviral vectors (Cellecta-pRSGT16-sgRNA containing guide RNAs and no-targeting control guide (gRNA-1, gRNA-2, gRNA-3, ntRNA)) were purchased from Collecta, packaged into second-generation lentiviral particles, and transduced into K562 cells stably expressing dCas9 using Lipofectamine 3000 transfection reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. In the Tet-inducible system, all guide RNAs were cloned under the control of a Tet-inducible promoter. In parallel, a non-targeting RNA control was also transfected. To induce the expression of gRNAs, cells were treated with a single addition of 5 µM doxycycline. The list of specific guides RNA tested is provided in Supplementary Table 6. For assessing guide RNA-2 expression in cells, custom primers and TaqMan probe were designed a list is provided in Supplementary Table 7.

ASO/FANA-ASO antisense oligonucleotide-mediated downregulation

ASOs targeting c-MYC ANCOR and no-targeting scrambled control were purchased from FANA AUM biotech (proprietary sequence). They were added to cell cultures at 10–20 µM concentration. Downregulation was monitored, and cell samples were collected at 48, 72, and 96 h after ASO treatment for subsequent qPCR analysis. Nascent DNA capture was carried out using the Click-iT® Nascent RNA Capture Kit (Thermo Fisher Scientific) following the manufacturer’s instructions with minor modifications, as previously described. Nascent DNA abundance was quantified 72 hours after ASO treatment by qPCR at the c-MYC origin (Primer set mychip9 and mychip10 in Supplementary Table 5).

dCAS9_VP64 -mediated upregulation of c-MYC ANCOR

A lentiviral vector (pLenti-U6-sgRNA-PGK-Neo) encoding the designed gRNA was synthesized by Applied Biological Materials Inc. pMD2.G, pCMV delta R8.2, and the lentiviral vectors were transfected into 10 million 293 T cells using TransIT-LT1 reagent (Mirus). Supernatants were harvested at 48 and 72 h after transfection. Virus was concentrated 100 times by Lenti-X Concentrator (Takara) after filtering through a 0.45 µm syringe filter and resuspended in DMEM and stored at –80 °C. dCas9-VP64 (Addgene, plasmid #61425) lentiviral transduction was performed in the presence of hexadimethrine bromide (final concentration 8 µg/ml) in HEK 293 cells. Blasticidin (10 µg/ml) was added to the cultures 2 days after infection. Stably expressing dCas9-VP64 cells were subsequently transduced with c-MYC ANCOR gRNA2 or a scrambled control. G418 (500 µg/ml) was added to the cultures 2 days after infection. Resistant clones were selected and screened to assess DNA replication and c-MYC ANCOR levels. The specific gRNA sequences used for the c-MYC ANCOR upregulation are provided in Supplementary Table 6.

DNA fiber analysis after blocking transcription by cell treatment with DRB

K562 cells were cultured and maintained in appropriate growth media until reaching a cell number of 15 million. Synchronization in the S phase of the cell cycle was achieved using the Double Thymidine block method, following established procedures⁵⁶. In addition, DRB treatments were employed to induce an early S phase block in transcription, as described in the literature³⁴. The DNA fiber assay was performed as previously described⁸⁴. Briefly, cells were sequentially pulse labeled with 25 µM CldU followed by 250 µM IdU for 30 min at 37 °C. Untreated cells were used as controls. Cells were washed with ice-cold PBS, trypsinized, and resuspended at 2.5 × 10⁶ cells/mL concentration in cold PBS. In all, 2 µl of cell suspension was gently placed on a glass slide, and after 2–3 min of incubation, lysis buffer (200 mM Tris-HCl; pH 7.5, 50 mM EDTA and 0.5% SDS) was added, followed by 2 min of incubation at room temperature. The slides were tilted at an angle (15–45°) to allow the fibers to spread along the slide. After drying, the slides were fixed in methanol-acetic acid (3:1) for 15 min at −20 °C, denatured in 2.5 M HCl for 30 min, and blocked in 5% BSA in 1× PBS for 20 min (blocking buffer). Anti-BrdU (Alexa Fluor® 488 Anti-BrdU antibody [BU1/75 (ICR1)], Abcam, ab220074) to recognize CldU (1:15, 1 µg) and IdU (1:15, 1 µg) were applied for 1 h at room temperature, followed by PBS wash. Slides were then stained with secondary antibodies, goat anti-rat Alexa Fluor 488 (2 µg, 1:15, Thermo Fisher, A-11006) and anti-mouse Alexa Fluor 568 (2 µg, 1:15, Thermo Fisher, A-11004), for 1 h at room temperature. Slides were mounted with ProLong Antifade Mountant (Thermo Fisher Scientific #P10144) and imaged on a Carl Zeiss Axio Imager M2 microscope. Labeled replication structures were identified, and the track length was measured using ImageJ. The replication event initiation and termination were also recorded and quantified. We used Student’s t test with a two-tailed distribution for P value calculation.

Single-molecule analysis of replicated DNA (SMARD) analysis after blocking transcription by cell treatment with DRB

When the cells reached 70–80% confluence, they were synchronized in S phase using the Double Thymidine block method⁵⁶, and DRB treatments were employed to induce block of early S phase transcriptional activities as described in the literature³⁴. The cells were then incubated at 37 °C for 4 h in the presence of 25 μM IdU (Sigma-Aldrich, St. Louis, MO). After washing cells with PBS, hESC medium with 25 μM CldU (Sigma-Aldrich, St. Louis, MO) was added to the cultures, and the cells were incubated for an additional 4 h. The cells were harvested with Accutase for hESC or Trypsin for mammalian epithelial cells. Following centrifugation, the cells were resuspended at 3 ×  10⁷ cells per ml in PBS. 1% InCert agarose (Lonza Rockland, Inc, FMC, lot 221291) in PBS was added to an equal volume of cells at 42 °C. The cell suspension was pipetted into a chilled plastic mold with 0.5- by 0.2-cm wells with a depth of 0.9 cm for preparing DNA gel plugs. The gel plugs were allowed to solidify on ice for 30 min. Cells were lysed in buffer containing 1% n-laroylsarcosine (Sigma-Aldrich), 0.5 M EDTA, and 20 mg/ml proteinase K. The gel plugs remained at 50 °C for 64 h and were treated with 20 mg/ml proteinase K every 24 h. Gel plugs were then rinsed several times with Tris-EDTA (TE) and once with phenylmethanesulfonyl fluoride. The plugs were washed with 10 mM MgCl₂ and 10 mM Tris-HCl (pH 8.0). The genomic DNA in the gel plugs was digested with 40 units of PmeI (at 37 °C overnight. The digested gel plugs were rinsed with TE and cast into a 0.7% SeaPlaque GTG agarose gel (Lonza Rockland, Inc.). A gel lambda ladder PFG marker, and yeast chromosome PFG marker (both from New England BioLabs, Inc.) were cast next to the gel plugs. PCR was performed to determine the appropriate positions of DNA on the pulsed-field electrophoresis gel, and the gel piece was cut and melted at 72 °C for 20 min. β-Agarase enzyme (1 unit per 50 μl of agarose suspension) was carefully added to digest the agarose and incubated at 45 °C for a minimum of 2 h. The resulting DNA solutions were stretched on 3-aminopropyltriethoxysilane-coated glass slides. The DNA was pipetted along one side of a coverslip that had been placed on top of a silane-treated glass slide and allowed to enter by capillary action. The DNA was denatured with sodium hydroxide in ethanol and then fixed with glutaraldehyde. The slides were hybridized overnight with a biotinylated probe (the blue bars diagrammed on the maps indicate the positions of the probes used). The following day, the slides were rinsed in 2× SSC buffer with 1% SDS and washed in 40% formamide solution containing 2× SSC at 45 °C for 5 min and rinsed in 2× SSC-0.1% IGEPAL CA-630. Following several detergent rinses (4 times in 4× SSC-0.1% IGEPAL CA-630), the slides were blocked with 1% BSA for at least 20 min and treated with Avidin Alexa Fluor 350 conjugate (1 µg, Invitrogen Molecular Probes, A11236) for 20 min. The slides were rinsed with PBS containing 0.03% IGEPAL CA-630, treated with biotinylated anti-avidin D (Vector Laboratories) for 20 min, and rinsed again. The slides were then treated with Avidin Alexa Fluor 350 for 20 min and rinsed again, as in the previous step. The slides were incubated with the IdU antibody, mouse anti-bromodeoxyuridine mouse anti-BrdU (1 µg, BD Horizon™ BV510 BD Bioscience, 563445), the CldU antibody, monoclonal rat anti-bromodeoxyuridine (anti-BrdU) Anti-BrdU antibody [IIB5] (1 µg, Abcam, Ab8955), and biotinylated anti-avidin D (Abcam ab34824) for 1 h. This was followed by incubation with Avidin Alexa Fluor 350 (1 µg, Invitrogen Molecular Probes, A11236) and secondary antibodies, Alexa Fluor 568 goat anti-mouse IgG (H + L) (1 µg, A-11004) and Alexa Fluor 488 goat anti-rat IgG (H + L) (1 µg, A-11006), both from Invitrogen Molecular Probes, for 1 h. After a final PBS/ IGEPAL CA-630 rinse, the coverslips were mounted with ProLong Gold Antifade Mountant. A fluorescent microscope (Axioscop 2 M2 with Plan Apochromat 63×/1.4 NA oil differential interference contrast objective; Carl Zeiss) with a camera (CoolSNAP HQ; Photometrics) was used to detect the nucleoside incorporation into the DNA molecules. Images were processed with Photoshop CS5 software. Image processing involves alignment of the DNA molecules according to the FISH probes, adjusting the contrast, and removing unspecific background signals. The molecules presented are the complete dataset.

For the FXN gene locus, biotinylated probes were prepared from Fosmids G248P83100G6 and G248P85473G3 that were selected using the public gene database (http://genome.ucsc.edu/).

To analyze the replication fork progression, the % molecules with IdU incorporation moving across the FXN locus was quantified. The segment was divided into 5 kb fragments, and the replication fork per segment was counted. The percentage of molecules with IdU incorporation was calculated for the 5 kb interval.

Modeling and molecular dynamics simulations of ORC1 in complex with RNA oligonucleotides

The RM9 and dRM9 models were built by using the MC-sym/MC-fold and the 3dnus procedures⁸⁵. The coordinates of the Orc1 protein as deposited in the alphaFold protein structure database⁸⁶ were downloaded and used as a starting model of the entire protein because none of the experimentally solved structures available in the PDB data bank contain the full-length protein coordinates. The created theoretical models of the interacting systems were individually subjected to molecular dynamic simulations to refine the structures and to explore the conformational dynamics, following validated protocols^87,88. In detail, 100 ns of simulation time was used for RM9 and dRM9 systems, assessing the validity of their initial conformations. For the Orc1 alphaFold predicted structure, we applied three diverse velocity simulations to refine the model and better explore the Orc1 conformational ensembles, which contain a high level of flexibility. Indeed, the high prevalence of disordered regions, particularly between the N- (residues 1–479) and the C- (residues 480–861) segments suggest that the protein can explore various equally probable states. No experimental data were available to define the relative orientations between these domains. We thus extracted from each of the three trajectories a representative Orc1 conformation (namely orc-1a, orc-1b, and orc-1c) and used each as a target to dock the RM9 and dRM9 sequences. The pdb 7JPS¹⁷ structures were used as a template to accommodate the RNA sequence. In detail, orc1a orc1b and orc1c states were structurally overlapped to the Orc1 subunit of 7jJPS (chain A) pdb, and the RNA sequence models were positioned to replace the region occupied by the DNA molecules of the same pdb that was used as a template for the docking. The Amber14sbOL15 force field⁸⁹ parameters were used in GROMACS 2020/5.0.5 package⁹⁰ to parametrize the molecules. The systems were solvated in an octahedron box with TIP3P water model in 150-mM KCl concentration and additional ions to neutralize. Electrostatic interactions were treated using the particle mesh Ewald method, and Berendsen algorithms were used for the control of temperature and pressure as previously applied on proteins^91,92, nucleic acids, and derivatives⁹³. The gromos method within the algorithm described by Daura et al.⁹⁴ was applied to perform clustering analyses to extract representative conformations along the trajectories based on RMSD criteria. VMD⁹⁵ and Pymol packages (https://www.pymol.org/) were utilized for analyses and image creation.

The models of EYYE and AAAA variants were obtained using as a template the conformation of the Orc1a extracted from the corresponding simulation in complex with RM9 (Orc1a-RM9) and by the use of MODELLER 9.22 software⁹⁶. The sequences used for the molecular modeling were that of the native Orc1a (uniprot code: Q13415) with ARG297, SER450, THR593, and ARG846 replaced in ASP297, TYR450, TYR575, ASP593 to obtain EYYE and in ALA297, ALA540, ALA593, ALA846 to model the AAAA variants, respectively. EYYE and AAAA in complex with RM9 were then subjected to 500 ns of MD simulations using the conditions described above.

Bioinformatic analysis adopted to investigate ANCORs distribution and Origin firing

RNA sequencing (RNA-seq)

Strand-specific transcriptome libraries were constructed on BGI Genomics platform (DNBSEQ). Pipeline of the experiment: (1) mRNA and non-coding RNAs were enriched by removing rRNA; (2) by using the fragmentation buffer, the mRNA and non-coding RNAs were fragmented into short fragments (about 200–700 bp); (3) the first-strand cDNA was synthesized by random hexamer-primer using the fragments as templates; (4) buffer, dNTPs, RNase H and DNA polymerase I were added to synthesize the second strand cDNA; (5) the double-stranded cDNA was purified with a QiaQuick PCR extraction kit and then used for end-polishing; (6) sequencing adapters were ligated to the fragments, and then the second strand was degraded using Uracil-N-Glyosylase); and (7) the fragments were purified by Agarose gel electrophoresis and enriched by PCR amplification. Libraries were paired-end sequenced on Hi-Seq-4000 Illumina. Raw reads were analyzed using a pipeline for RNA-seq analysis, the Visualization Pipeline for RNA-seq analysis (VIPER) based on workflow management system Snakemake (https://github.com/hanfeisun/viper-rnaseq). The read alignment to hg19 reference genome was performed using STAR aligner (2.7.0 f) with default parameters. Gene expression (FPKM values) was quantitated with Cufflinks (v2.2.1). Bedtools (v2.27.1) genomecov and bedGraphToBigWig v4 were used to generate bigwig files.

RIP-sequencing (RNA-seq) analysis

RNAs were processed for sequencing as described by Di Ruscio et al. with minor adaptations³⁴. RNA samples were depleted of ribosomal RNA with Ribo-ZeroTM Magnetic Gold Kit (cat. # MRZG126 Epicentre). Libraries were constructed on the BGI Genomics platform (DNBseq, RIPseq). Pipeline of the experiment: (1) RNA was enriched; (2) the first-strand cDNA was synthesized by random hexamer-primer using the fragments as templates; (3) second-strand synthesis and ER/A-tailing were performed; (4) ligation was accomplished with an overhanging T at the 3’ end of the bubble adapter; (5) the ligated products were amplified by PCR with specific primers; (6) single-strand separation was achieved by thermal denaturation and a bridge primer was used for circularization of the single-stranded DNA to obtain a single-stranded circular DNA library; (7) for sequencing of amplified products, DNA Nanoballs (DNBs) were used for paired-end sequencing on Hi-Seq-4000 Illumina. Raw data with adapter sequences or low-quality sequences were filtered. This step was completed by the SOAPnuke software developed by BGI. SOAPnuke software filter parameters: -n 0.001 -l 20 -q 0.4 -A 0.25 --cutAdaptor -Q 2 -G --minLen 100. Raw reads were analyzed using the pipeline, CHromatin enrIchment ProcesSor Pipeline (CHIPS), based on the workflow management system Snakemake (https://github.com/liulab-dfci/CHIPS) where the read alignment to the hg19 reference genome was performed using bwa mem aligner (0.7.15-r1140) with default parameters. The MACS2 algorithm (2.2.7.1) was used to call peaks, and both MACS2 (2.2.7.1) and bedGraphToBigWig (v 4) were used to generate bigwig files. Read counts for downstream analyses were quantified using the bioliquidator/bamliquidator tool from Docker image (https://hub.docker.com/r/bioliquidator/bamliquidator). Genomic regions were annotated using the Hypergeometric Optimization of Motif EnRichment tool (HOMER v3.12).

ChIP sequencing (ChIP-seq) analysis

ChIP libraries were constructed on the BGI Genomics platform (DNBseq, CHIPseq). ChIP library construction pipeline: (1) end repair and adapter ligation (filling the ends of IP/Input DNA followed by 5’ phosphorylation. A 3’ sticky-end with an overhanging A was formed, which could ligate with an overhanging T at the 3’ end of the bubble adapter); (2) PCR amplification (PCR amplification of the ligated products with specific primers); (3) single-strand separation and circularization (separation of PCR products into single-strands by thermal denaturation. Then circularization of the single-stranded DNA through a bridge primer to obtain single-stranded circular DNA library); (4) for sequencing of amplified products, DNA Nanoballs (DNBs) were used for paired-end sequencing on Hi-Seq-4000 Illumina. Raw data with adapter sequences or low-quality sequences were filtered. We first went through a series of data processing to remove contamination and obtain valid data. This step was completed by SOAPnuke software developed by BGI. SOAPnuke software filter parameters: -n 0.01 -l 20 -q 0.4 --cutAdaptor -Q 2 -G --polyX 50 --minLen 100. Raw reads were analyzed using a pipeline CHromatin enrIchment ProcesSor Pipeline (CHIPS) based on the workflow management system Snakemake (https://github.com/liulab-dfci/CHIPS) where the read alignment to the hg19 reference genome was performed using bwa mem aligner (0.7.15-r1140) with default parameters. MACS2 algorithm (2.2.7.1) was used to call peaks, and both MACS2 (2.2.7.1) and bedGraphToBigWig (v 4) were used to generate bigwig files. Read counts for downstream analyses were quantified using the bioliquidator/bamliquidator tool from Docker image (https://hub.docker.com/r/bioliquidator/bamliquidator). Genomic regions were annotated using the Hypergeometric Optimization of Motif EnRichment tool (HOMER v3.12).

Adopted methods for graphics

Genomic heatmaps were generated using computeMatrix and plotHeatmap scripts from deeptools (v3.5.1). All Boxplots were generated using the “ggplot2” R package (https://ggplot2.tidyverse.org). ScatterDensity plots were generated using the smoothScatter R function (https://www.rdocumentation.org/packages/graphics/versions/3.6.2). Genome browser pictures were generated using the “WashU Epigenome Browser” (https://github.com/lidaof/eg-react/) and IGV 2.11.1.

Statistical analysis

For the statistical evaluation of the results obtained in each experiment, we applied a paired or unpaired Student’s t test for comparison between two conditions. All statistical analyses are reported for at least three biological replicates, indicated in the figure or legend as n = 3. GraphPad Prism Software was used to generate a scatter plot with bars. Values of P ≤ 0.05 were considered statistically significant and are indicated in the figure or figure legend as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and n.s. (not significant).

All downstream analyses and statistical tests were performed in R. Comparisons of ChIP-seq, RIP-seq, and RNA-seq signal enrichment or gene expression between two conditions, when displayed as box plots, were evaluated using two-sided Welch’s two-sample t tests without multiple-comparison adjustment. Box plots show the median (center line), the 25th and 75th percentiles (box), and whiskers extending to the most extreme values within 1.5 × IQR; overlaid points represent individual observations. Overlap between genomic interval sets was determined by interval intersection, and the significance of the overlap was evaluated using Fisher’s exact test in a hypergeometric framework. Exact P values are provided in the text and figure legends. On figures, significance is indicated as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and n.s. (not significant).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

A.D.R., D.G.T., and S.M.M. supervised the project. A.D.R. and S.U. conceived and designed the study and wrote the manuscript; S.U., A.K.E., Y.Z., B.Q.T., L.R., B.B., and N.H.M. performed experiments; I.A. carried out molecular dynamics; L.P., J.D.L., and M.A.B. provided bioinformatic and statistical analysis of the data. T.P., M.D., and J.G. performed the DNA fiber analysis and SMARD. S.M.M. contributed to study design and manuscript revision. B.Q.T., S.M.M., S.P.B., R.F., and A.K.E. read the manuscript and provided suggestions. D.G.T., S.M.M., A.D.R., and A.K.E. critically reviewed the manuscript.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

A.D.R. discloses support for the research of this work from the National Institutes of Health, NIDDK [R01DK136116], NCI [R00CA188595]; the American Cancer Research Scholar Award RSG-23-1036643-01-DMC; the Department of Defense Bone Marrow Failure Research Program (DoD BMFRP) Idea Development Award [W81XWH-20-1-0518]; Harvard Initiative RNA Medicine (HIRM) Pilot Award 2021; Italian Association for Cancer Research (AIRC) Start-Up grant [2014-15347] and Giovanni Armenise Harvard Foundation. I.A. discloses support for the research of this work from SEE LIFE-StrEngthEning the ItaLIan InFrastructure of Euro-bioimaging, area ESFRI “Health and Food” [IR0000023]. D.G.T. discloses support for the research of this work from the award of National Institutes of Health, NCI [R35CA197697], [R01DK103858], [W81XWH-15-1-0161], and [P01HL131477-01A1]; AKE discloses support for the research of this work from National Institutes of Health [R50 CA211304]; S.U. discloses support for the research of this work from National Institutes of Health, National Research Service Award (NRSA) [5T32HL007917-24]. B.Q.T. discloses support for the research of this work from the National Institutes of Health, NCI [K01CA222707], NCI [R21 CA270067], and the American Cancer Society [Grant # 134088-IRG-19-143-33-IRG]. S.M.M. discloses support for the research of this work from the National Institute of General Medical Sciences [R35GM130322] and the National Science Foundation [2153071].

Data availability

The sequencing data have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE241684. The mass spectrometry raw data have been deposited to the ProteomeXchange Consortium via the MassIVE partner repository with the dataset identifier PXD077743. Computational simulations, including predictions of molecular structures, have been deposited in ModelArchive (https://www.modelarchive.org/) under the following project numbers ma-58rip (Project: orc1a-dRM9); ma-mrb6z (Project: orc1a-RM9); ma-33nek (Project: orc1b-dRM9); ma-pggyf (Project: orc1b-RM9); ma-drb5v (Project: orc1c-dRM9) and ma-59js0 (Project: orc1c-RM9). Source data are provided with this paper.

Competing interests

A.D.R. is a founder and scientific advisor of Aptadir Therapeutics; all interests are outside the submitted work. D.G.T. is a founder and scientific advisor of Aptadir Therapeutics; all interests are outside the submitted work. The remaining authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-73478-2.