Introduction

SIR2 is the founding member of a family of NAD-dependent protein deacylases called sirtuins, which have important effects on transcription, replication, chromatin structure and metabolism in organisms from yeast to humans (Houtkooper et al. 2012; Shahgaldi and Kahmini 2021). SIR2 was initially identified in budding yeast in 1979 on the basis of its role in transcriptional repression at the silent mating type loci (Haber and George 1979; Klar et al. 1979; Rine et al. 1979), and it was therefore dubbed a “Silent Information Regulator”. The silent mating type loci are heterochromatic regions that contain non-transcribed copies of the small open reading frames that determine yeast mating type, and deletion of SIR2 not only activates their transcription but also abolishes their heterochromatic structure (Ivy et al. 1986; Rine and Herskowitz 1987; Weiss and Simpson 1998; Ravindra et al. 1999; Haber 2012). It was later discovered that, in addition to regulating transcription at the silent mating type loci, Sir2 also regulates DNA replication, and it does so not only at the silent mating type loci but also at the two other heterochromatic regions of the yeast genome, namely at telomeres and the rDNA (Pasero et al. 2002; Yoshida et al. 2014). It is this last activity of Sir2, namely its effect on replication at the rDNA, that is the focus of this report.

The rDNA, which encodes the structural RNA components of the ribosome, is composed of approximately 150 tandemly arranged 9.1 kb repeats (Woolford and Baserga 2013). At almost 1.4 megabases, the rDNA is too large to be replicated passively, and thus it must contain its own origins of replication. This is achieved by the organization shown in Figure 1A: The ribosomal RNAs are produced as two major species, known as the 35S and 5S, which are transcribed by RNA polI and RNA polIII, respectively (Fernandez-Pevida et al. 2015). Between the 5’ ends of these genes is an origin of replication, and immediately to the left of this origin is a short Sir2-repressed RNA PolII-transcribed non-coding transcript called C-pro (Kobayashi and Ganley 2005; Li et al. 2006; Vasiljeva et al. 2008), which we will discuss further below. In wild-type cells, rDNA origins fire in the second half of S phase, making the rDNA among the later replicating regions of the genome, while in sir2 mutants, they fire early in S phase (Pasero et al. 2002; Yoshida et al. 2014; Foss et al. 2017).

Figure 1.

An indication of the biological significance of maintaining the late replication status of the rDNA came from a genome wide association study aimed at identifying naturally occurring genetic variants that affect replicative lifespan (Kwan et al. 2013). Replicative lifespan in budding yeast is defined as the number of daughter cells that a mother cell is able to produce and is therefore a measure of a cell’s replicative capacity at birth. By measuring lifespan in segregants from a cross between two genetically diverse strains of yeast, Kwan et al. were able to identify a point mutation at the rARS that (1) decreased origin activity at the rDNA; (2) increased origin activity elsewhere in the genome; and (3) prolonged replicative lifespan. They interpreted these results as reflecting competition for replication resources between the origins in two distinct compartments of the nucleus, namely the heterochromatic repetitive nucleolus or rDNA and the non-repetitive sequences that make up the bulk of the yeast genome. This competition is exacerbated by the perturbation of nucleolar heterochromatin caused by the absence of SIR2 (Yoshida et al. 2014; Foss et al. 2017). Most replication outside of the rDNA initiates from approximately 300 replication origins, and thus the 150 origins at the rDNA make up approximately one third of total origins. Increased replication resources directed toward the rDNA might therefore be expected to create a significant strain on replication in the rest of the genome.

Replication origins are known to compete for limiting resources at the level of origin firing. Broadly speaking, replication origin activity is regulated at two levels, namely origin licensing and origin firing (Remus and Diffley 2009). Origin licensing refers to loading of the replicative helicase in G1, and only origins that have been licensed are capable of being used in the subsequent S phase. Licensing occurs when an Origin Recognition Complex, Orc1-6, that is bound to an 11-17 base pair so-called ARS consensus sequence (ACS) within the origin loads two hexameric helicases, each consisting of a ring of subunits, Mcm2-7, that encircles the DNA (Bell and Stillman 1992; Francis et al. 2009; Bell and Botchan 2013; Deegan and Diffley 2016; Deegan et al. 2016; Douglas et al. 2018; Lewis et al. 2022). The mechanism by which Orc recognizes origins and loads an Mcm double-hexamer has been elucidated in great detail (Bell and Labib 2016; Costa and Diffley 2022; Hu and Stillman 2023). DNA replication initiates from only a subset of these licensed origins, namely those where the helicase complex is activated, a process referred to as origin firing (Rhind and Gilbert 2013). Origin firing is triggered by activation of two kinase signaling pathways, DDK and CDK, which then cause each hexameric Mcm helicase to encircle a single strand of DNA and track in opposite directions in front of the replisome, thereby establishing bidirectional replication (Heller et al. 2011; Larasati and Duncker 2016; Li et al. 2023). Firing factors are not present in sufficient quantities to activate all licensed origins simultaneously, and thus origins are activated in waves, with firing factors being recycled until the entire genome has been replicated (Mantiero et al. 2011; Tanaka et al. 2011). Determining the criteria by which origins are prioritized for activation remains an area of active research.

Importantly, if the competition between the origins at the rDNA and those in the rest of the genome is tilted in favor of the rDNA, either by a point mutation in the rARS or by deletion of SIR2, replication gaps between distantly spaced origins outside of the rDNA persist late into S phase, and genome duplication may remain incomplete when the cell enters mitosis (Foss et al. 2017). Indeed, the observation that weakening the rARS by a point mutation can suppress the short replicative lifespan characteristic of sir2 mutants suggests that these replication gaps may be the ultimate cause of “replicative death”. Teleologically, one can rationalize the prioritization of replication of the unique areas of the genome over the repetitive and heterochromatic areas, because (1) there are mechanisms available for repair in repetitive regions, like single-strand annealing, that are not available in unique regions, and they may therefore be able to survive lesions created when mitosis precedes the completion of replication; and (2) the cell is able to tolerate variation in copy number at the rDNA, but not deletion of protein-coding sequences in unique regions. Consistent with this view of the utility of assigning a low priority to replication of repetitive sequence, heterochromatic regions are also late-replicating in mammalian cells, and disease states have been associated with disruption of this pattern (Donley and Thayer 2013).

Despite its biological importance, the mechanism by which Sir2 suppresses replication initiation is poorly understood. Clues to Sir2’s role in replication may be gleaned from its role in transcription, although here, too, our understanding is limited. Sir2’s ability to repress transcription at the silent mating type loci has been postulated to follow from its role in chromatin compaction. Experiments demonstrating SIR2-dependent protection of heterochromatin from digestion with endonucleases or methylation by bacterial DNA methyltransferases support the notion that DNA in heterochromatin is generally less accessible (Gottschling 1992; Singh and Klar 1992; Loo and Rine 1994; Ansari and Gartenberg 1999). Thus, a widely publicized model attributes Sir2’s ability to inhibit transcription to blocking access of RNA PolII to the DNA, although a simple model of indiscriminate steric occlusion cannot account for this phenomenon (Chen and Widom 2005; Lynch and Rusche 2009; Kitada et al. 2012; Bondra and Rine 2023). Whether Sir2’s effect on replication is similarly the result of chromatin compaction limiting access to enzymes involved in DNA metabolism is not known.

A study examining the distribution of Mcm helicase complexes at the rDNA suggests that, rather than a global manifestation of lack of accessibility of the rDNA, the early activation of replication at the rDNA in the absence of Sir2 may reflect a more localized phenomenon (Foss et al. 2019). In this study, Mcm binding was assessed at nucleotide-level resolution by fusing a subunit of Mcm to micrococcal nuclease (MNase), a technique known as Chromatin Endogenous Cleavage, or “ChEC” (Figure 1B) (Schmid et al. 2004; Zentner et al. 2015). Permeabilization of cells with detergent followed by addition of calcium activates MNase, thereby generating approximately 65 base-pair fragments of DNA under Mcm double-hexamers that can be identified by high-throughput sequencing. This analysis revealed that, while deletion of SIR2 did not elicit an obvious change in the overall levels of licensing at the rDNA, it led to a dramatic redistribution of Mcm helicase complexes: In wild-type cells, the helicase complexes are located almost exclusively immediately adjacent to a high occupancy well-positioned nucleosome (+1 nucleosome in Figure 1C) approximately 130 base pairs to the left of the ACS (position A in Figure 1C). In contrast, the Mcm complexes in sir2 mutants are largely displaced to the right into an area with lower nucleosome occupancy (positions B and C in Figure 1C and Figure 1-figure supplement 1). This displacement appears to be caused by the de-repression of C-pro transcription in the absence of SIR2, which results in RNA PolII pushing the Mcm double-hexamer into the relatively nucleosome-free region. This observation is provocative in light of studies demonstrating that the arrangement of Mcm-proximal nucleosomes can have a dramatic effect on Mcm activation (Azmi et al. 2017). Specifically, by using chromatin remodeling enzymes (CREs) to create a variety of nucleosomal arrangements on an origin-containing fragment of DNA, Azmi et al. demonstrated that activation of an Mcm double-hexamer loaded at that origin could be inhibited by nucleosomes arranged by specific CREs. In another study (Chacin et al. 2023), a mutant version of Orc1 that retains the ability to load Mcm but alters phasing of Mcm adjacent nucleosomes was shown to inhibit DNA replication in vivo and in vitro, consistent with a critical role for nucleosome organization at origins in the initiation of replication.

Taken together, the observations above led us to hypothesize that (1) relocation of Mcm helicases from regions of high nucleosome occupancy to nearby regions of low nucleosome occupancy promotes early activation of replication origins in sir2; (2) this local difference in nucleosome occupancy is the result of modulation of nucleosomal architecture by one or more CREs; and therefore (3) screening CREs for the ability to suppress the early rDNA replication phenotype in sir2 mutants might allow us to identify the relevant CRE. In this report, we show that the displaced Mcm complexes fire before their non-displaced counterpart and that increased nucleosome occupancy near the displaced Mcms in the absence of Fun30 CRE suppresses their early activation. This is the first in vivo demonstration of the modulation of an origin activity by a specific chromatin remodeler in vivo, and has implications for the broader relationship between transcription and replication timing.

Results

Fun30 chromatin remodeling activity is required for early rDNA replication and for maintaining the rDNA array size in sir2 mutant

To identify CREs required for the early replication of the rDNA in sir2 mutants, we needed a method to determine replication timing, and for this we turned to S-seq. In this widely used technique, G1 and S phase populations of cells from a log phase culture are sorted using flow cytometry and their DNA subjected to genomic sequencing (Muller et al. 2014; Batrakou et al. 2020). Replication timing can then be inferred from read depths in the S phase fraction, because regions that replicate early in S phase will be present in two copies in most cells, while regions that replicate late will be present mostly as a single copy. S phase read depths are normalized to those in G1 to account for differences in DNA copy numbers, which is particularly important for repetitive loci whose copy number can vary, such as rDNA. As shown in Figure 2A, which depicts the S-seq profile of chromosome 12, relative replication timing (Trel) values using this approach typically range from 1.4 to 0.8 for early- and late-replicating regions, respectively, with a value of 1 indicating that a locus replicates at the same time as the genome-wide average (see Materials and Methods for specific calculation). The late replication of the rDNA, defined as coordinates 450,000 - 470,000 on chrXII, in wild type is reflected in a Trel value of 0.89, whereas this ratio increases to 1.2 upon deletion of SIR2 (Figure 2A,B). Using this method, we showed that a point mutation in the ACS of the rARS suppresses early replication of the rDNA in sir2, while promoting replication in the rest of the genome (Kwan et al. 2013; Foss et al. 2017). To identify potential chromatin remodelers required for early rDNA replication in sir2, we measured rDNA replication timing in double-mutants in sir2 and each of 10 non-essential CRE subunits (Chd1, Fun30, Htl1, Irc5, Isw1, Isw2, Nhp10, Swp82). Deletion of FUN30, but no other gene, delayed replication of the rDNA in a sir2 background, shifting the Trel value from 1.2 to 1.06, while having had no significant effect in SIR2 cells (0.89 for WT and 0.91 for fun30) (Figure 2B).

Figure 2.

Because Fun30 can form a complex with the replication scaffold protein Dpb11, it was possible that Fun30 played a structural rather than an enzymatic role in regulating rDNA replication. However, we found that a point mutation that abolishes Fun30 ATPase activity, K603R (Neves-Costa et al. 2009; Eapen et al. 2012), also suppresses early rDNA replication (Figure 2B), arguing that Fun30 instead promotes rDNA replication through its enzymatic activity in nucleosome remodeling. Fun30 is also known to play a role in the DNA damage response; specifically, phosphorylation of Fun30 on S20 and S28 by CDK1 targets Fun30 to sites of DNA damage, where it promotes DNA resection (Chen et al. 2016; Bantele et al. 2017). To determine whether the replication phenotype that we observed might be a consequence of Fun30’s role in the DNA damage response, we tested non-phosphorylatable mutants for the ability to suppress early replication of the rDNA in sir2; these mutations had no effect on the replication phenotype (Figure 2B), arguing against a primary role for Fun30 in DNA damage repair that somehow manifests itself in replication.

We next wanted to determine whether FUN30 affects rDNA array size. The repetitive nature of the rDNA allows it to fluctuate in size, and mutations that impair DNA replication, either globally or in a manner specific to the rDNA, are associated with shrinkage of the rDNA. For example, the rDNA in a wild-type strain in the S288c background consists of 150 copies, but a point mutation at the ARS consensus sequence (ACS) that impairs Orc binding causes the rDNA to shrink to just 80 copies (Kwan et al. 2013). rDNA shrinkage in response to global replication defects is thought to reflect the fact that the rDNA, which typically comprises approximately 10% of genomic DNA, imposes a significant replicative burden on the cell that can be ameliorated by decreasing the number of repeats (Salim et al. 2017).

We determined rDNA size by assessing the fraction of DNA sequencing reads from a sonicated G1 population that arose from the rDNA. While deletion of FUN30 had no notable effect on rDNA size in a SIR2 strain, it caused a 50% reduction in a sir2 mutant, from 9.2% to 4.2% (Figure 2C). rDNA size reduction of a similar magnitude was observed in the ATPase-deficient Fun30K603R mutant, but not in the non-phosphorylatable Fun30S20AS28A mutant, mirroring the disparate effect of the two alleles on rDNA replication timing. The reduction of the rDNA size in sir2 fun30 mutants was confirmed by qPCR analysis of DNA isolated from log cultures, and by Southern blots (Figure 2-figure supplement 1 and Figure 3). Based on measurements in log-growing cells, we estimated the rDNA size in fun30 sir2 mutants to ∼40 repeats (Figure 2-figure supplement 1). This places it in a category comparable to the shortest rDNA arrays in a collection of DNA replication mutants noted for their contraction of the rDNA (Salim et al. 2017), thus emphasizing the severity of the replicative disadvantage conferred by the deletion of FUN30 when Sir2 function is absent.

Figure 3.

In the analysis above, we are assuming that the reduced rDNA size is a downstream effect of decreased rDNA origin activity. While there is ample support for this view of cause and effect (Kwan et al. 2013; Salim et al. 2017), rDNA shrinkage itself has been reported to increase rDNA origin activity (Kwan et al. 2023). Such an effect would be expected to counteract the rDNA replication suppression conferred by FUN30 deletion, and it therefore suggests an explanation for our observation that rDNA replication timing in sir2 fun30 cells is not supressed to the level observed in WT cells.

In summary, we conclude that Fun30 nucleosome remodeling activity promotes rDNA replication and is required for maintaining rDNA size in sir2 mutants. We next aimed to determine whether Fun30 exerts this effect by reducing Mcm loading or the ability of loaded MCMs to fire.

High level of origin licensing at rDNA origins in WT, sir2 and sir2 fun30 mutants

It has been estimated that ∼20% of potential rDNA origins are activated during any single S-phase in wild-type cells (Brewer and Fangman 1988; Linskens and Huberman 1988). Furthermore, single-molecule studies suggested that the fraction of active origins is increased in sir2 mutant cells (Pasero et al. 2002; Yoshida et al. 2014). To determine whether SIR2 and FUN30 modulate rDNA replication by changing origin licensing, i.e. the fraction of origins that are bound by Mcm helicase complexes prior to the beginning of S phase, we assessed licensing in WT, sir2, fun30, and sir2 fun30 cells by measuring DNA cleavage by Mcm2-MNase using Southern blots (Foss et al. 2021): Digestion with XmnI yields a 3.5 kb fragment that spans the rARS (Figure 1A), and, if the rARS is licensed, activation of MNase by addition of calcium will eliminate this fragment using a probe upstream of the rARS. As a control, we used a fragment that we know does not contain Mcm complexes. Activation of MNase in all 4 genotypes led to an ∼70-85% decrease in the XmnI fragment (Figure 3), therefore we conclude that (1) licensing is high, at ∼70-85%; and (2) a much higher fraction of origins are licensed than are activated, making it unlikely that regulation of origin activity occurs at the level of licensing. This is the first quantitative assessment of licensing at the rDNA in either wild type or sir2.

Fun30 promotes firing of rDNA origins and suppresses firing of non-rDNA origins in sir2 mutants

Our observation that FUN30 status had no obvious effect on rDNA origin licensing suggested that Fun30 might instead advance rDNA replication in sir2 mutants by promoting rDNA origin firing. To evaluate this possibility we measured the effect of FUN30 status on origin activity at the rDNA using 2D gels. In this technique, active origins generate replication bubble signals, whereas passive replication of an origin appears as a y arc, and origin activity can therefore be assessed as the ratio of bubble-to-y arcs. We grew cells to log phase, arrested them for 1.5 hours in alpha factor, and then released them into 200 mM hydroxyurea (HU). As seen previously, deletion of SIR2 advances replication at the rDNA (for example, compare the bubble-to-y-arc ratio in WT and sir2 at 30 minutes (Figure 4A), while deletion of FUN30 suppresses this effect. Deletion of FUN30 in a strain with a wild-type copy of SIR2 had no obvious effect. We conclude that deletion of FUN30 partially reverses the accelerated activation of the ribosomal origins caused by deletion of SIR2.

Figure 4.

We and others have previously shown that deletion of SIR2 not only advances replication at the rDNA but also retards replication elsewhere in the genome (Kwan et al. 2013; Yoshida et al. 2014). The absence of SIR2 has been linked to decreased activation of less robust early replication origins, likely due to their heightened sensitivity to overall levels of replication-initiating factors. To ascertain the specificity of the interaction between FUN30 and SIR2, we investigated whether deleting FUN30 also mitigates the delayed activation of weak early origins in a sir2 mutant background. We focused on the activation of ARS305, an origin we previously identified as delayed in a sir2 mutant using 2D gel analysis (Foss et al. 2019). As seen previously, deletion of SIR2 suppresses ARS305, while additional deletion of FUN30 restores the bubble-to-arc ratio to near that of wild type (Figure 4B). Deletion of FUN30 in a SIR2 background had no perceptible effect in this assay.

Delayed initiation of genomic origins following SIR2 deletion has been previously demonstrated on a global scale as a decline in the incorporation of a labeled uridine analogue, bromodeoxyuridine, at weak early origins in sir2 cells compared to wild-type cells upon their release from G1 into HU (Yoshida et al. 2014). We employed a similar approach to investigate whether the deletion of FUN30 mitigates this effect, utilizing a different labeled nucleotide, ethynyl-deoxy-uridine (EdU), which can be isolated through click chemistry rather than immunoprecipitation. Wild-type (WT), sir2, sir2 fun30, and fun30 cells engineered for the uptake and phosphorylation of the analogue (Viggiani and Aparicio 2006) were arrested in G1 using alpha factor and collected 60 minutes after release into medium containing HU and EdU. DNA extracted from these cells was sonicated and subjected to click chemistry with biotin azide to covalently link biotin to EdU incorporated into DNA. Biotinylated DNA was then purified and sequenced. The impact of SIR2 and FUN30 status on replication timing was assessed by graphing the EdU signal at 111 early origins in WT cells on the x-axis against the signal in sir2, fun30, and sir2 fun30 cells on the y-axis (Figure 4C). Similar to observations with BrdU, we noted reduced EdU incorporation at a subset of early origins in sir2 mutants, coupled with increased EdU incorporation at rDNA origins. Deletion of FUN30 in a sir2 background partially restored EdU incorporation at early origins, concomitant with reduced EdU incorporation at rDNA origins. However, FUN30 deletion in the WT background did not significantly impact EdU incorporation at either rDNA or genomic origins. These findings highlight that FUN30 deletion-induced suppression of rDNA origins is accompanied by the activation of genomic origins.

Early firing of displaced helicase complexes

While our results above demonstrate that deletion of SIR2 promotes firing of the rDNA origins and that additional deletion of FUN30 suppresses this phenotype, they lack the resolution to distinguish replication arising from the displaced versus the non-displaced Mcm complexes. Other methods for measuring origin activation, such as generation of single-stranded DNA or incorporation of nucleoside analogs, also suffer from this limitation. Distinguishing between firing of the displaced versus the non-displaced Mcm complexes is important to test our hypothesis that displacement of the helicase complex into a relatively nucleosome-free region is responsible for its early activation. We therefore sought to devise a method that can differentiate activation of Mcm complexes that are only ∼150 bp apart, i.e. the approximate distance between displaced and non-displaced Mcms. We reasoned that, because MCM-ChEC has the resolution to distinguish closely spaced helicase complexes, the differential disappearance of such closely spaced footprints might permit us to distinguish their relative activation times.

To explore the possibility that disappearance of Mcm-ChEC signal could be used to distinguish firing of the displaced versus non-displaced populations of Mcm, we first asked whether the pattern of disappearance of Mcm signal recapitulates that of origin firing. In particular, while late origins are inhibited by hydroxyurea (HU), early origins are not; therefore we asked whether this phenomenon was reflected in Mcm-ChEC. We arrested cells in G1 using alpha factor and then released them into medium containing hydroxyurea (HU), taking samples for ChEC analysis at 15, 30, 45 and 90 minutes (Figure 5A). HU inhibits ribonucleotide reductase and thereby dramatically slows S phase, making it easier to follow replication kinetics, but the relative order of firing of different origins remains unchanged. In Figure 5B, we plot the Mcm-ChEC signal at 111 early origins and 101 late origins, with G1 on the x axis and S phase on the Y axis. While we observed little difference in the signal abundance between early and late origins just 15 minutes after release from alpha factor, before origins have had a chance to fire, this difference increases over the time course, with early origins dropping more than late. We conclude that disappearance of Mcm-ChEC signal as cells progress through S phase can be used to assess origin firing.

Figure 5.

We then turned our attention to the Mcm signals at the rDNA origin, where the presence of the replication fork block approximately 1.5 kb from the sites of Mcm binding provides a fortuitous means of confirming our interpretation, since loss of Mcm-MNase signal at the origin should appear as a reciprocal increase where the replisome stalls. Mcms in WT cells remained at their initial loading though the duration of experiment (blue box in Figure 5C, corresponding to position A in Figure 1C). Like in wild type, this non-displaced population of Mcm in sir2 remains relatively constant (blue box); however, in striking contrast, the displaced population of Mcm in sir2 (orange box, corresponding to positions B and C in Figure 1C) shows a continual decrease over the course of the experiment, and there is a concomitant increase in signal in the region upstream of RFB that is absent in wild type. Furthermore, both the decrease in the displaced signal (orange box) and its corresponding increase upstream of RFB were suppressed by deletion of FUN30. The changes in the relative abundance of displaced and non-displaced Mcms during the time course, as well as repositioning of the Mcm signal to the region upstream of RFB across different geneotypes were highly reproducible (Figure 5-figure supplement 1). Taken together, these observations provide strong support for the hypothesis that the displaced population of Mcm in sir2 cells fires more readily than its non-displaced counterpart and that the increased firing of the displaced Mcms is suppressed by FUN30 deletion.

It is notable that, although a small fraction of the total G1 signal in wild type arises from the displaced location, this minor signal likewise disappears more rapidly than its non-displaced counterpart (Figure 5-figure supplement 2). This suggests that some feature of the immediate chromatin environment may be responsible for the difference in firing times at the two locations, and that this difference is present in both wild type and sir2; in other words, the fact that the displaced helicase complex is activated before the non-displaced complex regardless of SIR2 status is consistent with a model in which sir2 mutants replicate their rDNA early by virtue of displacing much of their loaded helicases to a region where they are more apt to fire, rather than by altering the chromatin environment more globally to affect all replicative helicases in the nucleolus.

Finally, although deletion of FUN30 could suppress replication initation at the rDNA either by inhibiting firing of the active (displaced) Mcm complex, or by preventing its displacement to the “activating location” in the first place, our results make clear that suppression occurs by the former mechanism rather than the latter. Indeed, deletion of FUN30 actually skews the distribution to favor the displaced location over the non-displaced one, and this skewing can be seen in both sir2 and SIR2. Because, as we have previously shown, rightward displacement of Mcms is accomplished by derepression of C-pro transcription and consequent pushing by RNA PolII, this suggests that deletion of FUN30 does not act via repression of C-pro transcription. Consistent with both this expectation and a previous report demonstrating that fun30 mutants are deficient in transcriptional silencing (Neves-Costa et al. 2009), C-pro RNA levels which were ∼2 times higher in sir2 fun30 cells compared to those in sir2 cells when adjusted for differences in rDNA size (Figure 5-figure supplement 3). Taken together, our findings indicate that FUN30 deletion specifically hampers the activity of repositioned Mcms when silencing mechanisms are compromised.

Fun30 maintains low nucleosome occupancy near repositioned Mcms

Our findings demonstrate that Mcms, when situated at their loading site adjacent to a high-occupancy +1 nucleosome, exhibit reduced firing propensity. In contrast, when these Mcms are displaced to a neighboring region with lower nucleosome occupancy, their firing propensity significantly increases. This heightened firing tendency is dependent on Fun30 CRE and its remodeling activity. This suggests that the absence of FUN30 could elevate nucleosome occupancy around the relocated Mcms, consequently inhibiting their ability to fire.

To evaluate this possibility, we examined the effect of FUN30 deletion on nucleosome occupancy at rDNA origins in WT and sir2 background using MNase-seq in G1 arrested cells. Chromatin treatment with MNase digests DNA that is not bound by proteins, such as nucleosomes and transcription factors. Nucleosomes, which are wrapped by approximately 150 base pairs of DNA, tend to protect fragments in the 150-200 base pair range, and thus profiling fragments in this size range is widely used to assess nucleosome distribution. Plotting nucleosome midpoints at the rARS in G1 cells (Figure 6A) reveals that nucleosome occupancy at the +2 (green box) and +3 (red box) positions is higher in the sir2 fun30 double mutant than in the sir2 single mutant. Furthermore, this effect of Fun30 is not limited to the sir2 background, since nucleosome occupancy is also modestly increased in fun30 cells compared to wild type cells. These observations support the notion that it is the relatively nucleosome-free nature of the region into which Mcms are displaced in sir2 that make the displaced complexes more prone to activation. Furthermore, this suggests that it could be particularly informative to assess nucleosome occupancy specifically adjacent to loaded Mcm complexes, as described below.

Figure 6.

It has previously been observed, both with Mcm and with various transcription factors, that the ChEC technique can reveal not only the binding site of the protein to which the MNase is fused, but also, if present, that of the adjacent nucleosome (Zentner et al. 2015; Foss et al. 2019). This is due to the fact that the entrance and exit sites for DNA wrapped around a nucleosome are in close physical proximity, despite the fact that they are separated by approximately 150 base pairs, and thus the MNase fusion protein adjacent to a nucleosome often cleaves both sites. Mcm-ChEC, therefore, provides a powerful tool for measuring nucleosome occupancy immediately adjacent to the Mcm helicase complex exclusively in those rDNA repeats where those complexes have been loaded. By plotting the midpoints of 150-200 bp fragments released by Mcm2-Mnase, we were able to determine that nucleosome occupancy at both the +2 and +3 nucleosomes, i.e. those adjacent to the displaced Mcm complexes, are elevated in sir2 fun30 as compared to sir2 (Figure 6B). This indicates that Fun30 is required for maintaining the low nucleosome occupancy in the region into which the preponderance of Mcm complexes have been displaced in sir2 mutant cells.

Discussion

In this study, we have attempted to shed light on the mechanism by which Sir2 represses replication origins at the rDNA. Sir2’s activity has been widely ascribed to the intuitive, but mechanistically imprecise, notion that “closed chromatin” is less accessible to both transcription and replication factors. Consistent with this model, deletion of SIR2 abolishes heterochromatic compaction and simultaneously increases the ability of methylases and restriction endonucleases to act on those sequences (Gottschling 1992; Singh and Klar 1992; Loo and Rine 1994; Weiss and Simpson 1998; Ansari and Gartenberg 1999; Ravindra et al. 1999). However, at least in its simplest form, non-specific steric exclusion of DNA metabolic enzymes cannot fully account for the suppression of transcription and replication at the rDNA. Particularly strikingly in its inconsistency with such a model is the fact that the rDNA is the most highly transcribed region of the genome, with transcription by PolI and PolIII accounting for approximately 80% of cellular RNA (Warner 1999; Woolford and Baserga 2013). These two polymerase complexes are similar in size to the PolII complex, thus size alone is not sufficient to selectively occlude PolII. Furthermore, while deletion of SIR2 causes a 1.5-fold increase in accessibility of restriction enzymes to heterochromatin at the HML locus, it increases PolII transcription by >3000 fold (Holland 2002), thus the magnitude of Sir2’s ability to sterically exclude macromolecular complexes from heterochromatin is insufficient to explain its effect on PolII transcription. Similar measurements in eukaryotic cells found little difference in accessibility of restriction enzymes and micrococcal nuclease between euchromatin and heterochromatin (Chereji et al. 2019).

In the current report, we argue that Sir2’s role in suppressing replication at the rDNA reflects local differences in the distribution of the Mcm replicative helicase complexes rather than more global occlusion of these complexes from the rDNA. In particular, our results suggest that the key difference between wild type and sir2, with regard to replication initiation, is the location of the Mcm complex, with the preponderance of that complex in wild type abutting a high-occupancy nucleosome while, in sir2, it is mostly in a relatively nucleosome-free region (Figure 7).

Figure 7.

Furthermore, we show that, even though this nucleosome-free region is less than 200 base pairs from Mcm’s original location, the displaced helicase complex is significantly more prone to activation than its non-displaced counterpart. Moreover, the chromatin-remodeling enzyme Fun30 is required to maintain the nucleosome-free character around the displaced Mcm, and its deletion both increases nucleosome occupancy and decreases the ability of the displaced Mcm to fire. Finally, as discussed below, we suggest that the increased proclivity of the displaced Mcm to fire is not an anomaly specific to the physiology of sir2 cells, but is an integral feature of replication of the rDNA during normal S phase.

Our model for the manner in which Sir2 regulates rDNA replication depends critically on the notion that the local nucleosome landscape can influence the activation of Mcm helicase complexes, and in vitro experiments suggest that this is, indeed, the case (Azmi et al. 2017): Azmi et al. monitored both licensing and firing in vitro, using a 3.8 kb nucleosomal template containing the ARS1 origin. They found that, while licensing was equivalent whether the template had been populated with nucleosomes using Swi/Snf or Rsc remodelers, on the one hand, or Isw1a, on the other, the loaded Mcm complexes in the latter templates were more prone to activation. Furthermore, Mcm complexes on Swi/Snf- or Rsc-remodeled templates could be made more prone to activation if subsequently treated with Isw1a, although the converse was not true. The salient difference between the two types of nucleosomal landscapes, however, was not obvious, as the distance between the Mcm complex and the nearest nucleosome was not notably affected, although it is possible that the composition (e.g. H2A/H2B content) differs. It was also unclear how the nucleosome landscape affected activation of Mcm complexes, though the relevant step in activation was determined to be the formation of the CMG complex with Cdc45 and GINS rather than the prior phosphorylation of the complex by DDK. In summary, while it is clear that the local nucleosome landscape can affect origin activity (Eaton et al. 2010; Belsky et al. 2015), the underlying mechanism remains murky.

Regardless of the reason for the increased proclivity of the displaced Mcm to fire, the observation has implications for replication at the rDNA not only in sir2 mutants but also in wild type. We have previously shown that the rightward displacement of Mcm double hexamers in sir2 is due to derepression of C-pro transcription (Foss et al. 2019), while our current results demonstrate that the displaced Mcm is more prone to activation. Combining these two observations, we suggest that the rDNA replicates earlier in sir2 than in wild type simply because more of the Mcm complex is in a location that favors its activation; in other words, we favor the view that the region of relatively low nucleosome occupancy to the right of Mcm’s normal location is more amenable to Mcm activation, regardless of SIR2 status. In support of this notion, close analysis of our Mcm-ChEC data reveals that, while Mcm complexes at the displaced location are much rarer in wild type than they are in sir2, this particular subpopulation appears equally susceptible to activation in both genotypes. This has the important implication that the manner in which the rDNA origin fires in sir2 mutants may be simply an exaggerated instance of its normal behavior, and that even in wild type cells, it is predominantly from Mcms that are displaced, likely due to low level of C-pro transcription, that replication initiates.

Classic studies that employed EM and two-dimensional gel electrophoresis showed that only a fraction of rDNA origins, estimated at one in five, are activated in any single cell cycle in yeast (Brewer and Fangman 1988; Linskens and Huberman 1988). These estimates were later corroborated by single molecule analysis, which also observed that the fraction of active origins increases in sir2 mutants (Pasero et al. 2002). This low fraction of active rDNA origins could be due to insufficient origins licensing, suppression of activation or licensed origins or a combination of the two mechanisms. Our measurements of rDNA licensing showed that MCM are loaded in up to 90% of origins in WT cells and allowed us to conclude that regulation or rDNA replication initiation occurs primarily at the level of origin firing.

Chromatin remodeling enzymes have been proposed to assist DNA replication by modulation of nucleosome phasing in the vicinity of replication origins, but it has proven difficult to assign specific roles to remodelers in genome replication in vivo (Cutler et al. 2018). One of the reasons for this difficulty is the redundancy among different remodelers on one hand, and the multiple essential roles CRE play in DNA transactions in addition to replication. For example, in vitro studies have identified four different remodelers that in combination with ORC can phase nucleosome at 320 replication origins (Chacin et al. 2023). However, a deletion of any single of these remodelers had no discernable effect on nucleosome phasing in vivo and a quadruple mutant was required to abolish phasing. While the characterization of this quadruple mutant suggested impaired replication initiation, it was difficult to exclude indirect effects that might have been caused by the effects on transcription. In contrast, the effect of a chromatin remodeler on DNA replication we describe in this work is remarkably specific: we found that the activity of a single remodeler, Fun30, is required to activation not only of a specific origin, but that its action is specific to nucleosomes adjacent to repositioned Mcm complexes at the rDNA. The suppression of rDNA origin activity in sir2 mutants, which contained a large fraction of repositioned Mcms, by FUN30 deletion is so severe that it reduces the rDNA size ∼3-5 fold, from 160 to 35-50 copies.

In conclusion, our findings contribute to a growing body of evidence highlighting the importance of positioning of the Mcm replicative helicase within the local nucleosome environment in determining replication timing (Azmi et al. 2017; Rodriguez et al. 2017). While it is clear that small differences in this positioning can have significant effects on the timing of DNA replication, it is unlikely that a simple rule exists to directly infer origin timing solely based on the distance between the Mcm complex and the nearest nucleosome. Nonetheless, our results suggest that this parameter could be a crucial piece of the puzzle in understanding why some origins fire early in S phase, while others fire late. Further research is needed to unravel the precise molecular mechanisms governing the interplay between the Mcm complex and nucleosomes and how these dynamics influence replication timing across different genomic loci. Such insights will deepen our understanding of DNA replication regulation in the context of disease states associated with epigenetic drift, such as aging and cancer.

Material and methods

Yeast Strains

Strains used in the study were derived from BY4741 and W303 and are provided in Supplemental Table 1. Cells were grown in standard yeast peptone 2% dextrose media except for hydroxyurea time course study which employed synthetic media.

Strains containing point mutations in FUN30 were made using a plasmid that contains CAS9 and gRNA targeting the PAM sites adjacent to the desired mutation sites in FUN30 as described. Oligos containing FUN30-targeting gRNAs were cloned into pML104. pML104 was a gift from John Wyrick (Addgene plasmid # 67638)(Laughery et al. 2015). The plasmids, which induce double strand breaks within FUN30, were co-transformed with DNA blocks containing the repair templates with the desired mutations and altered PAM sites, synthetically synthesized by IDT. FUN30 DNA was PCR amplified from the resulting transformants and sequenced to verify that the desired mutations have been introduced.

Edu-incorporating yeast strains were derived from W303 and contain four BrdU vectors that express HSV-TK and hENT1 (Viggiani and Aparicio 2006). Plasmids, containing BrdU cassette obtained from Addgene (Addgene plasmid # 71789-71792), were integrated at four different auxotrophic loci (HIS3, TRP1, LEU2, and URA3). Single integration of each BrdU cassette was confirmed via PCR using plasmid-specific primer sets as described (Viggiani and Aparicio 2006), which assured that the strains contain exactly four copies of the BrdU cassette.

Measurements of rDNA replication timing using S-seq

S-seq experiments were carried out as previously described (Foss et al. 2017). Cells were grown to log phase in YEPD media, fixed using 70% (vol/vol) ethanol, subjected to proteinase K digestion, and their DNA stained with Sytox Green as described (Foss 2001). Cells were sorted according to DNA content on a BD Biosciences FACSAria II cell sorter into G1 and S fractions. DNA from a minimum of 1 e-6 cells from each fraction was isolated using the YeaStar Genomic DNA Kit (Zymo Research). DNA was fragmented by sonication and sequenced.

MNase seq

We carried out MNase-Seq as previously described (Foss 2019). Briefly, cells grown to log phase in rich medium, Yeast Peptone Agar with 2% glucose (YEPD), from an overnight 25 mL culture were synchronized with 3 μM alpha-factor for 1.5 hrs. at 30 ° C. Arrested cells were crosslinked with 1% formaldehyde for 30 min at room temperature water bath with shaking. Formaldehyde was quenched with 125 mM glycine and cells were centrifuged at 3000 rpm for 5 min. Cells were washed twice with water and resuspended in 1.5 mL Buffer Z (1 M sorbitol, 50 mM Tris-HCl pH 7.4) with 1 mM beta-mercaptoethanol (1.1 μL of 14.3 M beta-mercaptoethanol diluted 1:10 in Buffer Z) per 25 mL culture. Cells were treated with 100 μL 20 mg/mL zymolyase at 30 ° C for 20–30 min. Spheroplasts were centrifuged at 5000 rpm for 10 min and resuspended in 5 mL NP buffer (1 M sorbitol, 50 mM NaCl, 10 mM Tris pH 7.4, 5 mM MgCl₂, 1 mM CaCl₂) supplemented with 500 μM spermidine, 1 mM beta-mercaptoethanol and 0.075% NP-40. Nuclei were aliquoted in tubes with varying concentrations of micrococcal nuclease (Worthington), mixed via tube inversion, and incubated at room temperature for 20 mins. Chromatin digested with 1.9 U– 7.5 U micrococcal nuclease per 1/5th of spheroplasts from a 25 mL culture yielded appropriate mono-, di-, tri-nucleosome protected fragments for next-generation sequencing. Digestion was stopped with freshly made 5x stop buffer (5% SDS, 50 mM EDTA) and proteinase K was added (0.2 mg/ml final concentration) for an overnight incubation at 65 ° C to reverse crosslinking. DNA was extracted with phenol/chloroform and precipitated with ethanol. Micrococcal nuclease digestion was analyzed via gel electrophoresis prior to proceeding to library preparation. Sequencing libraries were prepared as described below for ChEC.

Chromatin Endogenous Cleavage (ChEC)

ChEC-seq was carried out as previously described (Foss et al. 2019; Foss et al. 2021). Briefly, MNase was activated by addition of CaCl2 to cells that were permeabilized with digitonin. DNA was extracted from cells using phenol and chloroform, precipitated using salt and ethanol and used to construct sequencing libraries without size fractionation. Cells were centrifuged at 1,500 x g for 2 mins, and washed twice in cold Buffer A (15 mM Tris pH 7.5, 80 mM KCl, 0.1 mM EGTA) without additives. Washed cells were carefully resuspended in 570 μL Buffer A with additives (0.2 mM spermidine, 0.5 mM spermine, 1 mM PMSF, ½ cOmplete ULTRA protease inhibitors tablet, Roche, per 5 mL Buffer A) and permeabilized with 0.1% digitonin in 30 ° C water bath for 5 min. Permeabilized cells were cooled at room temperature for 1 min and 1/5th of cells were transferred in a tube with freshly made 2x stop buffer (400 mM NaCl, 20 mM EDTA, 4 mM EGTA)/1% SDS solution for undigested control. Micrococcal nuclease was activated with 5.5 μL of 200 mM CaCl₂ at various times (5 mins and 10 mins) and the reaction stopped with 2x stop buffer/1% SDS. Once all time points were collected, proteinase K was added to each of the collected time points and incubated at 55 ° C water bath for 30 mins. DNA was extracted using phenol/chloroform and precipitated with ethanol. Micrococcal nuclease digestion was analyzed via gel electrophoresis prior to proceeding to library preparation. Library was prepared as previously described used using total DNA, without any fragment size selection (Foss et al. 2019; Foss et al. 2021).

Quantifying replication using the thymidine analogue Edu

Cells were grown to log phase in rich YEPD medium from a 50 mL culture and synchronized in G1 using 3 μM alpha factor for 2 hours at 30°C. Using 500 μg/mL pronase, arrested cells were released into rich media containing 200 mM hydroxyurea and 130 μM Edu (5-Ethynyl-2’-deoxyuridine) (Biosearch Technologies). Cells were harvested in G1 and after 60 minutes release into HU. DNA was extracted using YeaStar Genomic DNA Kit (Zymo Research) and sonicated. Separate sequencing primers were ligated to each of the samples. One hundred ng of each sample was pooled together, and subjected to click chemistry reaction in 10 mM Tris pH 7.4, 50 mM NaCl, 20 μM biotin-TEG azide (Vector Laboratories), 5 mM THPTA (tris-hydroxypropyltriazolylmethylamine) (Vector Laboratories), 1 mM copper sulfate, and 15 mM sodium ascorbate (total volume 200 μL). Click reaction proceeded for 30 minutes at room temperature prior to DNA ethanol precipitation. Streptavidin magnetic beads (New England Biolabs) were washed 3 times with B&W buffer I (100 mM NaCl, 10 mM Tris-HCl pH 8.0, 1 mM EDTA, 0.05% Tween20, 0.5% SDS) before adding 100 μL of bead suspension to 180 μL DNA plus B&W buffer I. After a 30-minute incubation at room temperature while gently inverting tubes, bead bound DNA was washed with 200 μL B&W buffer I followed by 100 μL Stringency wash buffer (0.1% SDS, 0.1% SSC buffer) and 200 μL Wash Buffer II (100 mM NaCl, 10 mM Tris-HCl pH 8.0, 1 mM EDTA, 0.05% Tween 20). DNA was eluted off the beads by denaturation at 98°C for 5 minutes, followed by cooling at 25°C. Edu-labeled DNA was PCR amplified post-pulldown and subjected to sequencing.

Hydoxyurea ChEC time course experiment

Cells grown in synthetic media to logarithmic phase were arrested in G1 using alpha factor for 90 minutes, centrifuged, washed in water to remove alpha factor, and released into media containing 200 mM Hydroxyurea. Cells were harvested in G1 and at indicated time points following release into HU-containing media, chilled at 4°C and analyzed by ChEC as described above.

qRT-PCR analysis of c-pro transcript and rDNA copy number

RNA was extracted from logarithmically growing cells after spheroplasting using an RNA-Easy column. The sequences of the primers used in qRT-PCR for C-pro and PDA1 mRNA are provided in Supplemental Table 2. rDNA size was measured by qPCR using DNA that was extracted by YeaStar Genomic DNA Kit (Zymo Research) using the primers listed in Supplemental Table 2. As an internal standard, qPCR was done in parallel on the DNA extracted from fob1 strains with 180 and 35 rDNA copies whose rDNA sizes had been verified by Pulsed Field Gel Electrophoresis (Kwan et al. 2023).

ChEC Southern blots for measuring origin licensing

Cells carrying Mcm2-Mnase were subjected to ChEC as described above, except that the SDS was omitted from the stop buffer to allow subsequent spheroplasting. DNA was extracted from spheroplasts using the YeaStar Genomic DNA Kit (Zymo Research), digested with XmnI, separated on 1.75% agarose gels and analyzed by standard Southern blotting technique. The primers used to make the rDNA and PIK1 control probe are listed in Supplemental Table 2.

Acknowledgements

We thank Toshi Tsukiyama, Bonny Brewer and M.K. Raghuraman for yeast strains. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number R01GM117446.