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).

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 breaks, 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 be ∼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.

Alterations of rDNA size are promoted by the replication fork barrier (RFB), which imposes a unidirectional replication fork arrest to prevent head-to-head collisions between DNA polymerase and RNA polymerase I while transcribing 35S units. Replication forks that are stalled at the RFB are prone to breakage, which can lead to unequal sister chromatid exchange and fluctuations in rDNA array size. These rDNA size fluctuations are exacerbated in sir2 mutants, which exhibit an increased rate of recombination, resulting in an unstable rDNA array size. Deletion of FOB1, which abolishes RFB function, stabilizes rDNA size by reducing,though not eliminating,the frequency of recombination-initiating double-strand breaks.

We have interpreted the shrinkage of the rDNA caused by deletion of SIR2 and FUN30 to be a specific example of the general phenomenon demonstrated by Salim et al. (Salim et al. 2017), i.e. that mutations that impair replication generate selective pressure that favors rDNA shrinkage. According to this interpretation, while deletion of FOB1 is expected to retard the transition to reduced array sizes, it should not prevent it, and thus rDNA shrinkage should not be FOB1-dependent.

To determine whether the reduction of rDNA size in sir2 fun30 mutants was dependent on RFB activity and FOB1, we deleted FUN30 in a sir2 fob1 strain with 160 repeats. Using this strategy, we obtained fob1 sir2 fun30 transformants with over 150 rDNA repeats.. If rDNA shrinkage does not require Fob1, it should occur, albeit slowly, in fob1 mutants. To assess this, we continuously grew several independent fob1 sir2 fun30 colonies, along with control fob1 sir2, and fob1 fun30 strains for 120 generations. We found that, unlike the control fob1 strains, which maintained a stable rDNA size, all four fob1 sir2 fun30 cultures reduced their rDNA arrays size to fewer than 100 copies by 120 division (Figure 2-figure supplement 2). These results demonstrate that, although FOB1 affects the kinetics of rDNA size reduction in sir2 fun30 strains, the reduced rDNA array size upon FUN30 deletion size does not depend on FOB1. Thus, suppressed rDNA replicaton drives the selection for reduced rDNA array size in both RFB-proficient and RFB-deficient fun30 sir2 cells. Consistent with this notion, our S-seq results showed delayed rDNA replication timing in fob1 sir2 fun30 cells with 150 rDNA repeats, compared to their fob1 sir2 counterparts (Figure 2-figure supplement 3). 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.

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 3), and, if the rARS is licensed, activation of MNase by addition of calcium will eliminate this fragment, thus making it undetectable when using a probe upstream of the rARS. As a control, we used a fragment that, based on our MCM-ChEC data, should not contain MCM complexes. Activation of MNase in all 4 genotypes led to an ∼60-85% decrease in the XmnI fragment (Figure 3 and Figure 3 figure supplement 1), therefore we conclude that (1) licensing is high, at ∼60-85%; and (2) a much higher fraction of origins are licensed than are activated, making it unlikely that significant 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.

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. We grew cells to log phase, arrested them for 1.5 hours in alpha factor and collected them at different time points after release into 200 mM hydroxyurea. Prior to separation on 2D gels, DNA was digested with NheI, which releases a 4.7 kb rARS-containing linear DNA fragment at the internal rDNA repeats (1N) and a much larger, 24.4 kb single-rARS-containing fragment originating from the rightmost repeat (Figure 4A). In 2D gels, active origins generate replication bubble arc signals, whereas passive replication of an origin appears as a y-arc (Figures 4B). Having a signal emanating from a single-rARS-containing fragment (Figures 4A and 4B) simplifies the comparison of rDNA origin activity in strains with different numbers of rDNA repeats, such as in sir2 vs sir2 fun30 mutants. Origin activity is expressed as a ratio of the bubble to the single-ARS signal, effectively measuring the number of active rDNA origins per cell at a given time point.

As seen previously, deletion of SIR2 increased the number of activated rDNA origins, while deletion of FUN30 suppressed this effect (Figure 4D). When analyzed in aggregate at 20, 30, 60 and 90 minutes following release into HU, the average number of activate rDNA origins in sir2 mutant was increased 6.3-fold compared to those in WT (5.0±2.3 in sir2 vs 0.8±0.4 in wt, p<0.05 by two-tailed t-test), and this increased number was reduced upon FUN30 deletion (1.3±0.7 in sir2 fun30, p<0.05 by 2 tailed t-test vs sir2, NS for comparison to WT). Deletion of FUN30 in a strain with a wild-type copy of SIR2 did not have asignificant effect. We conclude that deletion of FUN30 partially reverses the accelerated activation of ribosomal origins caused by deletion of SIR2.

The overall impact of FUN30 deletion on rDNA origin activity in a sir2 background is expected to be a composite of two opposing effects: a suppression of rDNA origin activation and increased rDNA origin activation due to reduced rDNA size. To evaluate the effect of FUN30 on rDNA origin activation independently of rDNA size, we generated an isogenic set of strains in a fob1 background, all of which contain ∼30 copies of the rDNA repeats. (Deletion of FOB1 is necessary to stabilize rDNA copy number.) Comparing rDNA origin activity in sir2 versus sir2 fun30 genotypes, we observed a robust and reproducible reduction in rDNA origin activity upon FUN30 deletion (Figure 4C). This finding confirms that the FUN30 suppresses rDNA origin firing in sir2 backgrpound independently of both rDNA size and FOB1 status.

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. Overall, the comparison of origin activity at ARS305, measured as bubble arc to 1N signal, with that at rDNA demonstrates a reciprocial relationship between activity at ARS305 and rDNA across the four genotypes (Figure 4E). As seen previously, deletion of SIR2 suppresses ARS305, whereas additional deletion of FUN30 increased activity close to that of wild type. Deletion of FUN30 in a SIR2 background also showed a trend of increased origin activation, though this difference was not statistically significant.

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 4F). 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. In particular, the median value of log₁₀ of read depths at 111 early origins, as the data are shown in Figure 4F, dropped from 6.5 for wild type to 6.2 for sir2 but then returned almost to wild type levels (6.4) in sir2 fun30. The p value obtained by Student’s t test, comparing the drop in 111 origins from wild type to sir2 with that from wild type to sir2 fun30 was highly significant (<< 10^-16) In contrast, FUN30 deletion in the WT background did not reduce EdU incorporation at genomic origins (median 6.6). These findings highlight that FUN30 deletion-induced suppression of rDNA origins in sir2 is accompanied by the activation of genomic origins.

While our results above demonstrate that deletion of SIR2 promotes firing of rDNA origins and that additional deletion of FUN30 suppresses this phenotype, they lack the resolution to distinguish replication arising from the displaced versus the nondisplaced 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 nondisplaced 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 nondisplaced 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 the disappearance of the MCM-ChEC signal could be used to distinguish the firing of the displaced versus nondisplaced populations of MCM, we first asked whether the pattern of disappearance of the 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. To do this, 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, 60 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. MCM signal quantification demonstrates that the differences in MCM abundance between early and late origins increase over the time course in HU (Figure 5-figure supplement 1). Using chromatin immunoprecipitation followed by sequencing (ChIP-seq) for FLAG-tagged MCM2 in cells arrested in alpha factor and those released into HU for 60 minutes, we observed a similar reduction of the MCM signal at early compared to later origins, confirming our MCM-ChEC results (Figure 5-figure supplement 2). We conclude that the disappearance of the MCM-ChEC signal as cells progress through S phase can be used to assess origin firing.

We then turned our attention to the MCM signals at the rDNA origin; here 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 the loss of the MCM-MNase signal at the origin should appear as a reciprocal increase where the replisome stalls. We first used MCM2-Flag ChIP in WT and sir2 cells after release into HU to confirm this prediction. In both WT and sir2 cells, the MCM can be found at the rDNA origin in alpha factor-arrested cells as expected (Figure 5 - figure supplement 3). However, at 60 minutes following release into HU, unlike in WT cells where the MCM signal remains at its loading site, in sir2 mutants, we can see that a large fraction of the signal has now redistributed from its loading site and accumulated left of the RFB, consistent with early activation of rDNA origins in sir2 cells and the arrest of the replicaiton forks at the RFB. While these results demonstrate that ChIP can be used to monitor the repositioning of the MCM signal at a kb scale and origin activation, its resolution is inadequate for small-scale changes in MCM signal distribution, which is needed to monitor the proposed differences in activation of repositioned and non-repositioned MCM complexes. We therefore turned to ChEC to address this question.

The MCM-ChEC signal in wild-type cells remained at their initial loading throughout the duration of the HU experiment (blue box in Figure 5C) as we have seen by ChIP in WT cells at 60 minutes. (Note that 200 mM HU causes an approximately 20-fold slowing of S phase, therfore a 60 minute time point is equivalent to <5 minutes without HU, i.e. prior to the time when the rDNA origin fires in wild-type cells). Like in wild type, this nondisplaced population of MCM in sir2 remains relatively constant (blue box); however, in striking contrast, the displaced population of MCM in sir2 (orange box) 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 the RFB were suppressed by deletion of FUN30. The changes in the relative abundance of displaced and nondisplaced MCMs during the time course, as well as the repositioning of the MCM signal to the region upstream of RFB across different genotypes, were highly reproducible (Figure 5-figure supplement 4).

Taken together, these observations provide strong support for the hypothesis that the displaced population of MCM in sir2 cells fires more readily than its nondisplaced 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 nondisplaced counterpart (Figure 5-figure supplement 5). 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 nondisplaced 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 initiation at the rDNA either by inhibiting the firing of the active (displaced) MCM complex or by preventing its displacement to the "activating location" in the first place, our results make it clear that suppression occurs via the former mechanism rather than the latter. Indeed, deletion of FUN30 actually skews the distribution to favor the displaced location over the nondisplaced one, and this skewing can be observed in both sir2 and SIR2. As we have previously shown, rightward displacement of MCMs is accomplished by the derepression of C-pro transcription and the consequent pushing by RNA polII, suggesting 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 were approximately twice as high in sir2 fun30 cells compared to those in sir2 cells when adjusted for differences in rDNA size (Figure 5-figure supplement 6). A combination of increased MCM repositioning and reduction of MCM signal at the RFB observed in sir2 fun30 mutants also confirms that the accumulation at the RFB is a reflection of replication fork stalling rather than RNA PolII-mediated repositioning of the MCMs. Taken together, our findings indicate that FUN30 deletion specifically hampers the activity of repositioned MCMs when silencing mechanisms are compromised.

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 and Figure 6-figure supplement 1) 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.

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 and Figure 6-figure supplement 1). This finding 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.

Strains used in the study were derived from BY4741 and W303 and are provided in Suplemental 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 cassete 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

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.

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.

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).

Cells were grown to log phase in rich medium, YEPD (1% Yeast Extract, 2% Peptone, 2% glucose), synchronized with 3 μM alpha-factor for 1.5 hrs. at 30°C. G1 arrested cells were released into 200 mM hydroxyurea (HU) YEPD and grown for 1hr. Cells were crosslinked with 1% formaldehyde for 30 min at room temperature, quenched with 125 mM glycine for 5 minutes, and centrifuged at 3000 rpm for 5 minutes. Cell pellets were washed three times with TBS (20 mM Tris-HCl, pH 7.6, 150 mM NaCl). Cell pellets were flash frozen.

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). Cells were harvested in G1 and after 60 minutes released into HU. DNA was extracted using YeaStar Genomic DNA Kit (Zymo Research) and sonicated. Separate sequencing primers were ligated to each of the samples. Samples were pooled together, and subjected to click chemistry reaction in 10 mM Tris pH 7.4, 50mM NaCl, 20 μM biotin-TEG azide (Vector Laboratories), 5mM THPTA (tris-hydroxypropyltriazolylmethylamine) (Vector Laboratories), 1mM 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.

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 degrees and analyzed by ChEC as described above.

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 phenol-chloroform extraction and ethanol precipitation using the primers listed in 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).

Cells carrying MCM2-Mnase were subjected to ChEC as described above, except that the stop 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.