Mrc1 and Rad9 cooperate to regulate initiation and elongation of DNA replication in response to DNA damage (original) (raw)

Introduction

Genome integrity is continuously challenged by genotoxic insults that invoke repair mechanisms and arrest cell cycle progression. At the cellular level, this is achieved through the engagement of specific checkpoint pathways (Hartwell & Weinert, 1989). Checkpoints are sophisticated signaling cascades in which proteins interact, modify each other, and spread throughout the cell to translate the stress signal into coordinated cellular responses, with protective effects for the cell physiology (Nyberg et al, 2002; Bartek et al, 2004; Ciccia & Elledge, 2010). Checkpoint signaling depends on the coordinated action of sensor, mediator, and effector proteins. In budding yeast, the apical checkpoint kinase Mec1 (ATR in human) is activated during DNA replication through its binding to RPA‐coated single‐stranded DNA, a general stress signal that accumulates at abnormal DNA structures and stalled forks (Zou & Elledge, 2003; Hashimoto et al, 2010; Pardo et al, 2017). Mec1 acts in cis to phosphorylate multiple components of the replisome and prevent irreversible fork collapse (Cortez, 2015; Pardo et al, 2017). It also acts in trans to amplify and transmit the checkpoint signal throughout the cell via the effector kinase Rad53 (CHK1 in human; Gobbini et al, 2013; Pardo et al, 2017). The mechanisms by which Mec1 activates Rad53 defines two distinct branches of the S‐phase checkpoint, namely the DNA damage checkpoint (DDC) and the DNA replication checkpoint (DRC).

Activation of the DDC depends on the checkpoint mediator Rad9 and occurs throughout the cell cycle (Siede et al, 1993; Weinert, 1998). Rad9 is constitutively bound to chromatin (Andreadis et al, 2014; Dibitetto et al, 2016) but is recruited to damage sites, where it acts as a scaffold to promote the autophosphorylation of Rad53 (Emili, 1998; Gilbert et al, 2001; Sweeney et al, 2005; Wysocki et al, 2005; Grenon et al, 2007; Pfander & Diffley, 2011). Unlike Rad9, Mrc1 is always associated with the fork and activates the DRC exclusively during S phase in response to a variety of replication impediments (Alcasabas et al, 2001; Katou et al, 2003; Osborn & Elledge, 2003; Gambus et al, 2006). Mrc1 activates Rad53 through an unclear mechanism involving Sgs1 and RFCCtf18 (Bjergbaek et al, 2005; Chen & Zhou, 2009; Crabbé et al, 2010; Kubota et al, 2011; Hegnauer et al, 2012). The full activation of the Mrc1 pathway depends on the stalling of at least 70% of ongoing forks (Shimada et al, 2002; Tercero et al, 2003). Besides its checkpoint function, Mrc1 promotes fork progression by coupling the activity of DNA polymerase ε and the CMG helicase (Nedelcheva et al, 2005; Szyjka et al, 2005; Tourrière et al, 2005; Hodgson et al, 2007; Lou et al, 2008; Yeeles et al, 2017).

Since the discovery of Mrc1, the role of Rad9 in S phase has been largely disregarded (Alcasabas et al, 2001; Tanaka & Russell, 2001). In HU‐treated cells, Rad9 is only recruited to stalled forks when Mrc1 is absent (Alcasabas et al, 2001; Katou et al, 2003; Komata et al, 2009). In the presence of MMS, it has been recently proposed that Rad9 signals ssDNA gaps that are left in the wake of replication forks upon lesion bypass (García‐Rodríguez et al, 2018). Since Rad9 is essential for viability in mrc1Δ mutants, Rad9 is generally considered as a backup pathway for Mrc1 (Tourriere & Pasero, 2007; Branzei & Foiani, 2009). However, this view neglects earlier reports stressing the importance of Rad9 in S phase (Navas et al, 1996; Paulovich et al, 1997a,b) and the interplay between Rad9 and Mrc1 has remained largely unexplored.

Once activated by Mrc1 and/or Rad9, Rad53 acts in trans to regulate multiple cellular events, including the prevention of entry into mitosis, the upregulation of dNTP pools, the induction of DNA repair genes, and the protection of nascent DNA against nucleolytic degradation (Branzei & Foiani, 2009; Pardo et al, 2017). When cells are exposed to hydroxyurea (HU), a drug that slows down replication by preventing the expansion of dNTP pools (Koc et al, 2004), Rad53 also blocks the activation of late‐firing origins to arrest the replication program (Santocanale & Diffley, 1998; Shirahige et al, 1998; Zegerman & Diffley, 2010). This program has been extensively studied in budding yeast using a variety of genomic approaches and defined times of replication (Trep) have been assigned to all origins (Raghuraman et al, 2001; Yabuki et al, 2002; Katou et al, 2003; Knott et al, 2009; Müller et al, 2014). Since late origins fire within minutes after entry into S phase, their repression depends on the timely activation of Rad53 (Crabbé et al, 2010; Poli et al, 2012).

In HU‐treated cells, the repression of late origins depends on Mrc1, but not on Rad9 (Katou et al, 2003; Crabbé et al, 2010). Since the Rad9 pathway is active in S phase, this raises the question of how checkpoint mediators confer different functions to the same effector kinase. Interestingly, it has been reported that the pattern of Rad53 phosphorylation changes with the genotoxic agent used to activate it (Smolka et al, 2005; Sweeney et al, 2005). It is therefore tempting to speculate that checkpoint‐triggering signals modulate the substrate specificity of Rad53 by modifying its phosphorylation profile, in order to adapt the checkpoint response to the nature of the challenge (Yoo et al, 2006; Branzei & Foiani, 2009). Alternatively, Mrc1 and Rad9 could differentially regulate the timing and extent of checkpoint activation, to fine tune the execution of the DNA replication timing program, and elicit an optimal replication stress response. However, experimental evidence supporting one or the other model is currently lacking.

Another important question regarding the regulation of the replication program in response to DNA damage is whether fork progression is actively downregulated by checkpoint kinases or passively slowed down by the presence of DNA lesions. Indeed, replisomes encountering DNA lesions are phosphorylated in cis by Mec1/ATR to ensure fork protection and recovery (Smolka et al, 2007; Chen et al, 2010; Bastos de Oliveira et al, 2015; Cortez, 2015). But whether Rad53/CHK1 can also act in trans to slow down elongation at undamaged forks is currently unclear. This question has been extensively addressed during the past decades but work from different groups led to contrasted results (Tercero & Diffley, 2001; O'Neill et al, 2007; Seiler et al, 2007; Szyjka et al, 2008; Willis & Rhind, 2009; Iyer & Rhind, 2013, 2017a). The conundrum is that the drugs generally used to activate the checkpoint response also impede fork progression. Moreover, inactivation of checkpoint kinases leads to irreversible fork collapse, precluding the analysis of their role in fork progression. Consequently, the question of whether checkpoint kinases regulate elongation through damaged DNA has remained largely unanswered.

Here, we have reinvestigated the role of Rad9 in the replication stress response and characterized its functional interplay with Mrc1. By combining genome‐wide approaches, single‐molecule analysis, 2D gel, and pulsed‐field gel electrophoreses, we show that Mrc1 and Rad9 play distinct but complementary functions in the replication stress response. As an integral component of the replisome, Mrc1 is suited for the fast activation of Rad53 at stalled forks and is responsible for the repression of late origins. However, this mode of Rad53 activation is transient and rapidly declines with the number of stalled forks. Sustained activation of Rad53 requires Rad9, which is also able to mediate the inhibition of late origins firing if activated early enough. Moreover, we show that Rad9 acts in trans to slow down replication forks in a Rad53‐dependent manner, unambiguously establishing a role for the S‐phase checkpoint in the elongation of DNA replication. Altogether, these data indicate that Mrc1 and Rad9 coordinate distinct checkpoint responses under replication stress conditions.

Results

Mrc1 and Rad9 repress late origins upon MMS exposure

Rad53 is phosphorylated in a Rad9‐dependent manner when mrc1Δ cells are exposed to HU, indicating that the DDC is functional in S phase (Alcasabas et al, 2001). Yet, Rad9 is unable to repress late origins under these conditions (Crabbé et al, 2010). Since Rad9 signals DNA damage and HU slows down replication without inducing DNA lesions, we reasoned that Rad9 could have a more prominent role in the signaling of damaged forks. To address this possibility, we monitored the repression of late origins in cells exposed to methyl methanesulfonate (MMS), a drug that impedes fork progression by alkylating DNA (Tercero & Diffley, 2001). Wild‐type, mrc1Δ, and rad9Δ cells were released synchronously from G1 in medium containing BrdU to label nascent DNA in the presence of 0.1% MMS (Fig 1A). Flow cytometry profiles showed that S phase progression was severely delayed in wild‐type cells exposed to 0.1% MMS (Fig 1B), which is consistent with earlier reports (Paulovich & Hartwell, 1995; Tercero & Diffley, 2001). A similar delay was observed in rad9Δ mutants. However, mrc1Δ cells progressed further into S phase than wild‐type and rad9Δ cells (Fig 1B), presumably because they were unable to repress late origins, as reported earlier for rad53 mutants (Tercero & Diffley, 2001; Tercero et al, 2003). To confirm this hypothesis, we monitored the ability of mrc1∆ and rad9∆ mutants to repress late origins, a very sensitive readout of Rad53 activation (Crabbé et al, 2010). Genomic DNA was extracted 45 min after release into S phase, and sites of BrdU incorporation were mapped in the yeast genome. Under these conditions, nearly identical patterns of early origin activation were detected in all strains. Late origins (marked by asterisks) were repressed in wild‐type and rad9Δ cells, but fired in mrc1Δ mutants (Fig 1C), as it is the case in HU‐treated cells (Crabbé et al, 2010). To quantify this difference, we determined the amount of BrdU incorporated within 1‐kb intervals centered on 384 individual origins and plotted this value relative to the time of replication (Trep) (Müller et al, 2014) for bins containing 96 origins each. In wild‐type and rad9Δ cells exposed to 0.1% MMS, this analysis confirmed that all early origins fired efficiently (positive signal log ratio), whereas late origins were globally repressed (Fig 1D). In contrast, most origins fired in mrc1Δ cells, regardless of their Trep (Fig 1D). We conclude that the repression of late origins in response to 0.1% MMS is primarily mediated by Mrc1.

Figure 1

Figure 1

The alternative text for this image may have been generated using AI.

Full size image

Genome‐wide analysis of BrdU incorporation in cells exposed to MMS

A. Wild‐type, rad9Δ, and mrc1Δ cells were arrested in G1 with α‐factor and were released synchronously into S phase with the addition of pronase, in medium containing BrdU and 0.1% or 0.033% MMS. Cells were collected at the indicated time points.

B. Analysis of DNA content by flow cytometry.

C. Analysis of BrdU incorporation by BrdU‐IP‐chip 45 min after release into S phase in cells treated with 0.1% MMS as described in (A). A representative region on chromosome VII is shown. The position of replication origins is indicated above the map. Asterisks: late origins firing in mrc1Δ cells.

D. Quantification of the BrdU incorporation (signal log ratio) at 384 early and late replication origins in wild‐type, rad9Δ, and mrc1Δ strains exposed to 0.1 or 0.033% MMS. Red lines indicate median values. Median replication times (Trep) are indicated for bins of 96 origins.

Since the high density of DNA lesions induced by 0.1% MMS severely impedes S‐phase progression (Fig 1B) and may completely block elongation, we monitored DNA replication in cells treated with a lower dose of MMS (0.033%). Under these conditions, all three strains progressed slowly through S phase (Fig 1B), but this progression was faster in mrc1∆ mutants, owing to the derepression of late origins (Figs 1B and D, and EV1A; asterisks). Intriguingly, rad9Δ mutants also progressed faster through S phase than wild‐type cells (Fig 1B), despite the fact that late origins were largely repressed by Mrc1 in these cells (Fig EV1A). However, a quantitative analysis of BrdU incorporation revealed that late origins are partially derepressed in rad9Δ cells compared to wild‐type cells upon exposure to 0.033% MMS (Figs 1D and EV1A; arrows). This result was confirmed by plotting the difference in BrdU signal between rad9Δ and wild‐type cells (Fig EV1B; arrows). Together, these data indicate that the repression of late origins in MMS‐treated cells depends primarily on Mrc1. Nevertheless, Rad9 could contribute to the repression if replication forks were let to progress through the damaged template DNA at a lower MMS concentration.

Rad9 and Mrc1 regulate the kinetics of Rad53 activation

To further characterize the role of Rad9 in the replication of damaged DNA, we next exposed wild‐type, mrc1Δ, and rad9Δ cells to 0.033% MMS as described in Fig 1A and monitored replication initiation at the level of individual DNA molecules by DNA combing (Fig 2A) as described previously (Bianco et al, 2012; Tourrière et al, 2017). In this assay, each BrdU track corresponds to two sister forks diverging from a single origin. The center‐to‐center distance between adjacent BrdU tracks is indicative of the rate of initiation, and the length of BrdU tracks reflects fork speed. The distance between BrdU tracks was significantly shorter in mrc1Δ cells (43 kb) than in wild‐type cells (62 kb), confirming that more origins fired in mrc1Δ cells relative to wild type (Fig 2B). This distance was also significantly shorter in rad9Δ mutants (54 kb), which is consistent with the fact that late origins were partially derepressed in rad9Δ cells (Fig EV1A). The length of BrdU tracks was shorter in mrc1Δ cells than in wild‐type and rad9Δ cells (Fig 2C), which reflects the role of Mrc1 in fork progression (Tourrière et al, 2005; Yeeles et al, 2017). In contrast, BrdU tracks were significantly larger in rad9Δ mutants (48 kb) than in wild‐type cells (35 kb) when analyzed 60 min after release from G1 (Fig 2D). Calculation of the fork rate between the 45 and 60 min time points revealed a sixfold difference between rad9Δ and control cells (1.2 and 0.2 kb/min, respectively), suggesting that forks progress faster in the absence of Rad9. Although longer BrdU tracks could also result from the fusion of adjacent replicons, these data suggest that Rad9 could have a role in the regulation of the elongation of DNA synthesis in response to DNA damage.

Figure 2

Figure 2

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Differential regulation of DNA replication and Rad53 activation by Rad9 and Mrc1 in MMS‐treated cells

A Wild‐type, mrc1Δ, and rad9Δ cells were released into S phase in the presence of BrdU and 0.033% MMS, as described in Fig 1A. DNA combing analysis was performed for samples collected 45 and 60 min after release into S phase. Representative DNA fibers are shown. Red: DNA, green: BrdU. Bar is 20 kb.

B Distribution of center‐to‐center distances in cells released for 45 min into S phase.

C, D Distribution of BrdU track lengths in cells released for 45 or 60 min into S phase, respectively.

E, F Wild‐type, mrc1Δ, and rad9Δ cells were synchronized in G1 with α‐factor prior to release with pronase in the presence of 0.033% MMS. Cells were collected at the indicated times, and Rad53 phosphorylation was monitored by Western blot (E) or with an in situ Autophosphorylation assay—ISA (F). The asterisk indicates a non‐specific band detected by the anti‐Rad53 antibody.

Data information: In panels (B–D), median values are indicated in red. For each sample, at least 100 tracks were measured in two biological replicates. Asterisks indicate difference with wild‐type cells. ****P < 0.0001, ***P < 0.001, **P < 0.01, ns: non‐significant, Mann–Whitney rank sum test.

This prompted us to ask whether Rad9 is recruited to nascent chromatin in MMS‐treated cells. To address this possibility, cells were released synchronously into S phase for 30 min in the presence of 0.033% or 0.1% MMS, as described in Fig 1A, and the binding of Rad9‐PK6 was monitored by ChIP‐qPCR at a 22‐kb region located downstream of the early origin ARS453. Changes in DNA copy number in the input fraction revealed that forks progressed up 11 kb away from ARS453 in the presence of 0.033% MMS (Fig EV2A and B), which is consistent with DNA combing analyses (Fig 2C), but was restricted to ~5 kb in the presence of 0.1% MMS. Interestingly, Rad9 was reproducibly detected behind forks confronted to 0.1% MMS, but not at unreplicated regions (Fig EV2C), in agreement with a replication‐dependent signaling of MMS lesions by the Rad9 pathway.

To determine whether the different replication profiles observed in MMS‐treated wild‐type, mrc1Δ, and rad9Δ cells reflect differences in Rad53 activity, we next analyzed the kinetics of Rad53 phosphorylation by Western blot (WB) and in situ autophosphorylation assay (ISA). To this end, cells were released synchronously into S phase in the presence 0.033% MMS and the electrophoretic mobility shift of Rad53 was monitored by Western blot. In wild‐type cells, Rad53 was rapidly phosphorylated after release from the G1 arrest and remained phosphorylated throughout the experiment (Fig 2E). This perfectly matched the autophosphorylation activity of Rad53, as measured by ISA (Fig 2F). In mrc1Δ cells, Rad53 activation was delayed by 15 min, as seen both by mobility shift and ISA (Fig 2E and F), which accounts for the derepression of late origins in this mutant (Fig 1C). In rad9Δ cells, Rad53 was phosphorylated upon S phase entry with kinetics similar to the wild type, but never reached the same level of activation observed in wild‐type and mrc1Δ cells (Fig 2E and F). These results likely account for the partial derepression of late origins observed in these cells (Figs 1C and EV1A and B) and indicate that Rad9 is needed to achieve a complete and sustained Rad53 activation in MMS‐treated cells. Together, these data show that Rad9 and Mrc1 cooperate to regulate the extent and the kinetics of Rad53 activation in the presence of DNA damage.

Rad9 represses late origins when activated early in S phase

Since late origins fire within minutes after early origins, we reasoned that the Rad9‐dependent activation of Rad53 in MMS‐exposed mrc1∆ cells could be too slow to repress late origins. To address this possibility, we designed an experiment in which the Rad9 pathway could be artificially activated in early S phase with the addition of the radiomimetic agent Zeocin (Fig 3A). To this end, wild‐type and rad9Δ cells were released synchronously from G1 into S phase in medium containing HU to repress late origins in an Mrc1‐dependent manner. Zeocin was added 60 min later to create DSBs and activate the Rad9 pathway. Then, cells were released from the HU arrest in fresh medium still containing Zeocin to allow fork restart in the presence of an active Rad9 pathway. Under the conditions used, we estimated that Zeocin generates on average 11 DSBs per haploid genome after 45 min and 48 DSBs after 90 min (Fig EV3A; see Materials and Methods for details). This density of DSBs is sufficient to activate Rad53 but does not directly interfere with fork progression. In wild‐type cells, Rad53 phosphorylation reached a maximum 45–60 min after release from the HU arrest, as shown by Western blot and ISA (Figs 3B and EV3B and C). This activation persisted after 90 min, which is consistent with the inability of cells to complete S phase (Fig 3C). In contrast, rad9Δ cells showed a much weaker Rad53 activation (Figs 3B and EV3B and C). Moreover, the level of Rad53 phosphorylation induced by HU rapidly decreased upon HU removal (time 0), explaining why cells rapidly progressed through S phase despite the presence of DSBs (Fig 3C).

Figure 3

Figure 3

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Pre‐activation of the DDC allows Rad9‐dependent repression of late origins

A. Wild‐type and rad9Δ cells were synchronized in G1 and released into S phase with the addition of pronase in medium containing IdU and 200 mM HU to activate the DRC. Then, DSBs were induced by the addition of 100 μg/ml Zeocin to activate the DDC. HU was washed away 45 min later to inactivate the DRC, and cells were resuspended in fresh medium containing CldU and 100 μg/ml Zeocin, in order to keep the DDC active. Samples were collected at the indicated times.

B. Analysis of Rad53 phosphorylation by Western blot.

C. Analysis of DNA content by flow cytometry.

D. Analysis of DNA replication by DNA combing in cells released for 45 and 60 min into S phase in the presence of 100 μg/ml Zeocin. IdU tracks (red) correspond to initiation events occurring during the HU treatment. Elongation events after HU removal appear as CldU tracks (yellow) flanking IdU tracks. New initiation events are detected as isolated CldU tracks (asterisks). Bar is 20 kb.

E. Number of new initiation events per DNA fibers detected at 60 min in the experiment shown in (D). For each sample, at least 65 individual DNA fibers larger than 200 kb were analyzed, for a total of > 17 Mb.

F. 2D gel analysis of the initiation of DNA replication at the early origin ARS305 and the late origin ARS1212. The experiment was performed as in (A) for the indicated times. Arrows indicate late origin firing.

G. Length distribution of CldU tracks 45 min after release into S phase in the DNA combing experiment shown in (D). Median values are indicated in red. For each sample, at least 250 tracks were analyzed in two biological replicates. ****P < 0.0001, Mann–Whitney rank sum test.

To characterize the mechanism by which Rad9 delays DNA synthesis, wild‐type and rad9Δ cells were labeled with IdU during the HU arrest and with CldU after release from the HU arrest in medium containing Zeocin (Fig 3A). Genomic DNA was extracted in agarose plugs, and DNA fibers were analyzed by DNA combing (Tourrière et al, 2017). Short IdU tracks were detected in HU‐arrested cells, corresponding to the activation of early origins (Fig 3D, HU). Upon HU removal, CldU tracks flanking IdU tracks (yellow‐red‐yellow signals) were detected in both strains, reflecting replication restart. However, new initiation events (CldU only; asterisks) were three times more frequent in the rad9Δ mutant than in control cells (Fig 3D and E). To confirm this Rad9‐dependent regulation of origin activity, the firing of the early origin ARS305 and the late origin ARS1212 was monitored by neutral/neutral 2D gel electrophoresis in wild‐type and rad9Δ cells. As reported earlier (Crabbé et al, 2010), ARS305 fired and ARS1212 was repressed in the presence of HU (Fig 3F). When cells were released from HU in fresh medium containing Zeocin, this repression was maintained in wild‐type cells but was relieved in the rad9Δ mutant, confirming our DNA combing results. Altogether, these data indicate that Rad9 can efficiently repress late origins firing if activated early enough in S phase.

Rad9 slows down fork progression in response to DNA damage

DNA combing experiments revealed that 45 min after release from the HU arrest, CldU tracks were significantly longer in rad9Δ cells (13 kb) than in wild‐type cells (9 kb; P < 0.0001; Fig 3G). This difference further increased 60 min after release but could not be quantified due to the fusion of adjacent replicons (Fig 3D, 60 min). These results suggest that Rad9 slows down fork progression in response to DNA damage. Alternatively, longer CldU tracks in rad9Δ mutants could be due to the merging between advancing forks and new initiation events (Fig 3E) and/or to larger dNTP pools, which are known to increase fork rates (Poli et al, 2012). To discriminate between these three possibilities, we first monitored dNTP levels in wild‐type and rad9Δ cells released synchronously from G1 into S phase in the presence of Zeocin. In G1, wild‐type and rad9Δ cells showed identical dNTP pools (Fig EV3D). Upon addition of Zeocin, dNTP levels markedly increased in wild‐type cells, but to much lower levels in rad9Δ mutants (Fig EV3D), which is consistent with the role of Rad9 in the upregulation of RNR activity in response to DNA damage (Zhao et al, 2001). We can therefore conclude that the larger BrdU tracks observed in rad9Δ mutants are not due to increased dNTP pools.

Next, we monitored the progression of individual forks by 2D gel analysis following the same experimental design as above (Fig 4A). In HU‐treated wild‐type and rad9Δ cells, the early origin ARS719 fired with the same efficiency, as indicated by the relative intensity of Y‐ and bubble arcs (Fig 4B). In both strains, the intensity of the bubble arc decreased 20 min after release from the HU arrest in the constant presence of Zeocin (probe #1), indicating that replication resumption occurred with the same kinetics. We then monitored fork passage 7 kb downstream of this origin (probe #2). At this location, forks can be exclusively assigned to events initiating at ARS719, for the next origin (ARS718) is located 55 kb away. The Y‐arc signal was detected earlier and was more intense in the rad9Δ mutant than in wild‐type cells (arrows), suggesting that forks progress faster in the absence of Rad9 (Fig 4B).

Figure 4

Figure 4

The alternative text for this image may have been generated using AI.

Full size image

Rad9 slows down fork progression in response to DSBs

A. Cells synchronized in G1 were released into S phase by pronase addition in the presence of 200 mM HU, to activate the DRC. 45 min later, 100 μg/ml Zeocin was added to activate the DDC. Still 45 min later, HU was washed away, Zeocin kept, and different time points taken for analyses.

B. 2D gel analysis of initiation and elongation in wild‐type and rad9Δ cells at the early origin ARS719 (probe 1) and at a locus 7 kb away of the origin (probe 2). Time in Zeocin after the HU wash is indicated. Arrows point to Y arcs (fork signals) with the strongest intensity.

C. 2D gel analysis of fork progression in wild‐type and rad9Δ cells through a 70‐kb region of chromosome VI for which all active origins downstream of ARS607 have been deleted. The position of probes and the time after release from HU are indicated. Arrows point to Y arcs (fork signals) with the strongest intensity.

To prove this further, we next used an engineered strain in which all the active origins downstream of the early origin ARS607 are deleted (Tercero & Diffley, 2001). This allows the unambiguous analysis of single forks progressing over 70 kb toward the telomere (TEL‐6R). In HU, ARS607 fired efficiently in both wild‐type and rad9Δ cells (probe #3) and fork restart occurred with the same kinetics after release from the drug (Fig 4C). In the rad9Δ mutant, forks were detected 20 kb away from the origin (probe #4) 40 min after release and reached a maximum at 60 min (arrow). They were also detected 20 kb farther at 60 min (probe #5). In contrast, forks were hardly detectable with probe #4 at 60 and 80 min in wild‐type cells and never reached the probe #5 location (Figs 4C and EV4A). These data indicate that in rad9Δ cells, replication forks progress at a rate of 0.7–1 kb/min after release from HU, which is consistent with our DNA combing analyses (Fig 3D). Altogether, these data indicate that Rad9 restrains elongation in response to DSBs, independently of dNTP pools and of the derepression of late origins.

Rad9 regulates fork progression independently of Mrc1

In the experiments described above, cells were released from an HU arrest, which means that Rad53 was first activated by Mrc1 and then by Rad9. To rule out the possibility that this pre‐activation by Mrc1 interferes with the subsequent activation by Rad9, we next introduced the _cdc7_‐4 mutation in our three strains to arrest cells at the G1/S transition without activating the Mrc1 pathway. To this end, cells were grown at the permissive temperature for the cdc7‐4 mutation (23°C) and were arrested in G1 with α‐factor. Then, the culture was shifted to 37°C and cells were released from the α‐factor arrest with the addition of pronase. Under these conditions, cells accumulate at the G1/S transition due to the lack of DDK activity (Bousset & Diffley, 1998; Donaldson et al, 1998). Zeocin was added to the culture to induce DSBs, and cells were released from the cdc7 arrest 45 min later by shifting the temperature back to 23°C (Fig EV4B). Interestingly, we observed that cdc7‐4 cells were able to progress through S phase in the absence of Rad9, but not in control cells or mrc1Δ mutants (Fig EV4C). Together with the data presented above, these results indicate that Rad9 restrains DNA synthesis under DNA damage, independently of Mrc1.

Still, since Cdc7 has been implicated in the activation of Rad53 (Dohrmann et al, 1999; Ogi et al, 2008), we next checked that the observed phenomenon was not due to the lack of Cdc7 activity by inducing DSBs with Zeocin directly in G1‐arrested cells. In principle, induction of DNA damage in G1 cells activates the DDC and blocks the G1/S transition (Siede et al, 1993). However, we have previously shown that there is a window of time during which it is possible to activate Rad53 in late G1 without preventing cells entry into S phase (Alabert et al, 2009). Wild‐type, mrc1Δ, and rad9Δ cells were synchronized in G1 for two hours with α‐factor, and Zeocin was added 45 min before release into S phase (Fig 5A). Cells were collected at the indicated time points, and DNA content was monitored by flow cytometry. Again, rad9Δ cells were able to complete S phase after release from the G1 arrest, whereas wild‐type and mrc1Δ cells remained arrested with a 1C DNA content (Fig EV5A).

Figure 5

Figure 5

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Rad9 regulates S‐phase progression independently of Mrc1

A. Wild‐type, mrc1Δ, mrc1 AQ, and rad9Δ cells synchronized in G1 with α‐factor were incubated in the presence of BrdU and 100 μg/ml Zeocin for 45 min prior to release into S phase. Cells were collected at indicated time points.

B. Analysis of DNA content by flow cytometry.

C. Analysis of Rad53 phosphorylation in G1 prior to Zeocin addition (G1) and at indicated times. The asterisk indicates a non‐specific band detected by the anti‐Rad53 antibody.

D. DNA combing analysis of replication profiles 60 min after release into S phase in the presence of Zeocin. Representative DNA fibers are shown. Red: DNA, green: BrdU. The green channel is also shown separately. Bar: 20 kb.

E. Distribution of BrdU tracks length in wild‐type, mrc1Δ, and mrc1 AQ cells. DNA fibers are fully labeled in the case of rad9Δ cells. Median values are indicated in red and correspond to the analysis of at least 150 tracks from two biological replicates. ns: non‐significant Mann–Whitney rank sum test.

To ensure that wild‐type and mrc1Δ cells were not blocked in G1 due to the activation of the G1 checkpoint, we monitored the mobility of their chromosomes by pulsed‐field gel electrophoresis (PFGE). In G1 and G2/M, yeast chromosomes can be sorted according to their size by PFGE. However, when cells enter S phase, the chromosomes are trapped in the well due to the presence of non‐linear replication intermediates (RIs). This is illustrated for chromosome VII after Southern blot analysis using a specific probe (Fig EV5B). In G1, this chromosome migrated as a single band of 1.05 Mb. After release from G1, chromosome VII was progressively retained in the well in the three strains, showing S‐phase entry. This demonstrates that none of the strains were blocked in G1 after Zeocin treatment. In rad9Δ cells, chromosome VII signal reached a plateau 30–45 min after release, corresponding to the time when cells progressed through and exited S phase, as determined by flow cytometry (Fig EV5A). Later on, the signal could be detected again in the gel at 60 and 90 min, as replication was completed (Fig EV5B and C). In contrast, chromosomes from wild‐type and mrc1Δ cells failed to complete DNA synthesis (Fig EV5B), confirming thereby that Rad9 delays S‐phase progression in response to DSBs.

Since fork progression is intrinsically slower in mrc1Δ mutants (Tourrière et al, 2005), we next monitored S‐phase delay and Rad53 activation in mrc1 AQ mutants, in which only the checkpoint function of Mrc1 is compromised. We found that, alike wild‐type and mrc1Δ cells, mrc1 AQ mutants remained arrested with a 1C DNA content after Zeocin exposure (Fig 5B). Moreover, Rad53 activation was detected at 30 and 60 min post‐release in wild‐type, mrc1Δ, and mrc1 AQ cells, but was strongly delayed in rad9Δ mutants (Fig 5C). To further confirm this result, BrdU was added to the cultures before addition of pronase (Fig 5A) and the length of BrdU tracks was monitored by DNA combing 60 min after release into S phase. This analysis revealed that replication origins fired in all four strains but forks did not progress much in wild‐type, mrc1Δ, and mrc1 AQ cells (Fig 5D and E). In contrast, chromosomes were almost fully replicated in rad9Δ cells. We therefore conclude that Rad9‐ but not Mrc1‐dependent signaling slows down fork progression in response to DNA damage.

Rad9 slows down replication progression in a Rad53‐dependent manner

Lastly, we addressed the mechanism by which Rad9 restrains fork progression in response to DNA damage and in particular, whether it depends on the effector kinase Rad53. To this end, S‐phase progression was monitored in rad53‐11 mutants, using the same experimental setup as described in Fig 5A. This mutant was chosen because it is completely checkpoint‐deficient yet does not require SML1 deletion for viability, thus has normal dNTP pools. Flow cytometry analysis revealed that rad53‐11 cells progressed faster through S phase than wild‐type and mrc1 mutants despite the presence of Zeocin (Fig 6A), as it is the case for rad9Δ mutants. PFGE analysis confirmed that at least 40% of rad9Δ and rad53‐11 cells managed to complete DNA synthesis within 60 min, whereas chromosomes from wild‐type and mrc1 cells remained partially replicated or failed to initiate replication (Fig 6B). We also directly assessed fork progression by DNA combing and found that replication tracks were significantly longer in rad9Δ and rad53‐11 mutants compared to wild‐type cells at the early time point of 30 min (Fig 6C) and chromosomes were almost fully replicated in the mutants after 60 min, in contrast with the wild type (Fig EV6). Importantly, both S‐phase progression and replication completion were identical in rad9Δ, rad53‐11, and in the rad9Δ rad53‐11 double mutant (Fig 6A and B). Measurements of BrdU tracks 30 min after release into S phase to avoid fusions between adjacent replicons also showed similar fork speed (Fig 6C and D). Finally, all three strains showed similar dNTP pools before and after Zeocin exposure (Fig EV3D). In light of these data, we conclude that the mutations rad53‐11 and rad9Δ are epistatic with regard to the regulation of replication elongation and thus that, most likely, Rad9 exerts this function through Rad53.

Figure 6

Figure 6

The alternative text for this image may have been generated using AI.

Full size image

Source data are available online for this figure.

Rad9 controls fork progression via Rad53

A. Cells from the indicated strains were released synchronously into S phase in the presence of Zeocin as described in Fig 5A, and DNA content was analyzed by flow cytometry.

B. Analysis of chromosome mobility by pulse field gel electrophoresis (PFGE) after staining of DNA with ethidium bromide. In G1, the chromosomes are linear and are sorted according to their length. Upon entry into S phase (30 min), they cannot enter the gel due to the presence of replication intermediates. Quantitation of chromosomes re‐entering the gel shows completion of DNA replication in 40% of rad9Δ, rad53‐11, and rad9Δ rad53‐11 mutants, but not in wild‐type cells and mrc1 mutants as determined for a representative experiment. Error bars indicate SD between the signal intensity determined for the three larger chromosomes (arrowheads).

C. DNA combing analysis of replication profiles 30 min after release into S phase in the presence of Zeocin. Representative images are shown. Red: DNA, green: BrdU. Bar is 20 kb.

D. Length distribution of BrdU tracks measured at that time for wild‐type, rad9Δ, rad53‐11, and rad53‐11 rad9Δ cells. Median lengths are indicated in red. For each sample, at least 200 tracks were measured from two biological replicates. ****P < 0.0001. Mann–Whitney rank sum test.

Discussion

The prevalent concept of the S‐phase checkpoint places Mrc1 as the central mediator of the replication stress response, Rad9 acting as a backup pathway (Tourriere & Pasero, 2007; Branzei & Foiani, 2009; Finn et al, 2012). We have previously reported that Rad9 cannot compensate for the absence of Mrc1 to repress late origins in HU‐treated cells (Crabbé et al, 2010), presumably because HU slows down fork progression without generating DNA lesions. Here, we show that Rad9 is also unable to repress late origins in mrc1∆ cells treated with a high dose of MMS (0.1%), even though template DNA is heavily damaged. Still, these lesions are efficiently signaled by the Mrc1 pathway, as indicated by the repression of late‐firing origins. The reason why Rad9 is unable to mediate a timely activation of Rad53 in response to high doses of MMS is currently unclear, but it could relate to the mechanism by which Rad9 is recruited to stressed forks (Fig 7A and B). As a general rule, MMS does not generate DSBs in vivo (Lundin et al, 2005), even if clustered lesions in opposed DNA strands could theoretically lead to DSBs, but induces fork‐blocking lesions that can be bypassed by damage avoidance mechanisms, leaving single‐stranded DNA gaps behind replication forks (Hashimoto et al, 2010; García‐Rodríguez et al, 2018).

Figure 7

Figure 7

The alternative text for this image may have been generated using AI.

Full size image

Model of the interplay between Mrc1 and Rad9 during DNA replication through damaged DNA

A. Mrc1 is an integral component of the replisome that activates Rad53 when replication forks slow down or pause at damaged templates. In contrast, Rad9 is a mediator of DNA damage that signals post‐replicative DNA damage.

B. High doses of MMS (0.1%) induce a rapid stalling of replication forks and a robust induction of the Mrc1 pathway, leading to the efficient repression of late origins, as it is the case in HU‐arrested cells. In cells exposed to a lower dose of MMS, forks progress further and replicons merge, leading to the extinction of the Mrc1 pathway. Maintenance of Rad53 activation depends on the induction of the Rad9 pathway by post‐replicative DNA lesions. Induction of a limited number of DSBs (as in the case of Zeocin) activates the Rad9 pathway and represses both initiation and elongation at otherwise unchallenged forks.

Rad9 is recruited to DNA damage sites through its interaction with Dpb11 and the 9‐1‐1 complex (Puddu et al, 2008; Pfander & Diffley, 2011). The 9‐1‐1 complex is preferentially loaded at the 5′ junctions between double‐stranded DNA and single‐stranded DNA (Majka & Burgers, 2003), which is consistent with Rad9 signaling the presence of single‐stranded DNA gaps in MMS‐treated cells (García‐Rodríguez et al, 2018). Rad9 binding also depends on its phosphorylation by Cdc28CDK1, whose activity is low at the G1/S transition and increases toward G2/M (Mendenhall & Hodge, 1998). Moreover, Rad9 interacts with histone H3 methylated on lysine 79 by Dot1 (Wysocki et al, 2005; Grenon et al, 2007) and with histone H2A phosphorylated on serine 129 (γ‐H2A) by Mec1/Tel1 (Toh et al, 2006; Hammet et al, 2007). Here, we have shown that Rad9 is enriched on newly replicated chromatin but not on unreplicated DNA in cells exposed to 0.1% MMS. Since the activation of Rad53 by Rad9 depends on the formation of ssDNA gaps behind forks (García‐Rodríguez et al, 2018) and on the modification of histones by Mec1 and Dot1, it is not surprising that this pathway mediates a slow but sustained response to replication stress. In contrast, Mrc1 is an integral component of the replisome and is therefore ideally located to engage an immediate response to fork impediments (Osborn & Elledge, 2003). However, this pathway depends on the presence of a large number of stressed forks (Shimada et al, 2002; Tercero et al, 2003). Our data indicate that forks progress further in cells exposed to 0.033% MMS than to 0.1% MMS (this study) or 200 mM HU (Crabbé et al, 2010). This increased fork progression favors the merging of adjacent replicons, leading to the extinction of the Mrc1 pathway (Fig 7A and B). In agreement with this model, we found that Rad53 is rapidly but transiently activated by Mrc1 in rad9Δ cells exposed to 0.033% MMS and is slowly but continuously activated by Rad9 in mrc1Δ mutants. Together, our data indicate that Rad9 and Mrc1 signal different structures related to stressed forks and cooperate to promote a rapid and prolonged activation of Rad53 in response to a variety of replication impediments.

The results discussed above suggest that the ability of Mrc1 or Rad9 to repress late origins depends primarily on the kinetics of Rad53 activation. However, it has been reported that the profiles of Rad53 phosphorylation vary with the drugs or the mediators used to activate it (Alcasabas et al, 2001; Smolka et al, 2005; Sweeney et al, 2005; Crabbé et al, 2010). Here, we also observed clear differences in the mobility and in the autophosphorylation capacity of Rad53 between MMS‐treated rad9Δ and mrc1Δ cells, independently of the kinetics of Rad53 activation. Hence, the possibility remains that the mode of Rad53 activation determines its substrate specificity (Yoo et al, 2006; Branzei & Foiani, 2009). To determine whether Rad9 can mediate the repression of late origins if activated early enough, we have induced DSBs in cells arrested in early S phase with HU and monitored origin firing after release from the HU arrest. Under these conditions, Rad53 was fully proficient to repress late origins, indicating that the DNA damage checkpoint can control the replication program, provided that it is activated early enough. Whether this is also the case for other Rad53 functions is an important question that remains to be addressed.

Besides initiation, another important question is whether the S‐phase checkpoint controls the rate of fork progression in response to DNA damage. Here, flow cytometry analyses revealed that rad9Δ cells progress faster through S phase than wild‐type or mrc1Δ cells in the presence of 0.033% MMS, regardless of the repression of late origins. Moreover, single‐molecule analyses confirmed that forks progress faster in rad9Δ cells exposed to MMS than in control cells independently of dNTP pools, suggesting that Rad9 represses replication elongation in response to DNA damage. This finding contrasts with other reports indicating that MMS slows down elongation because alkylated DNA physically impedes fork progression, in a checkpoint‐independent manner (Tercero & Diffley, 2001; Iyer & Rhind, 2017b). However, it is worth noting that these experiments were performed with rad53∆ (cds1∆ in fission yeast) or mec1∆ mutants, which completely lack checkpoint signaling and are therefore unable to protect stalled forks. As pointed out by others (Szyjka et al, 2008), replication forks are rapidly destabilized in these cells, which would mask any positive effect on elongation. Here, we have monitored fork progression in MMS‐treated rad9Δ cells, in which the Mrc1 pathway is functional, and that are thus largely proficient to repress late origins. In these cells, we showed that Rad9 slows down fork progression in a Rad53‐dependent manner. These data are consistent with the fact that Rad53 dephosphorylation is required to resume DNA synthesis after an MMS exposure (O'Neill et al, 2007; Szyjka et al, 2008). Because the conditions used to activate the S‐phase checkpoint also directly interfere with fork progression (Branzei & Foiani, 2009), attempts have been made to modulate the density of fork‐blocking lesions to confirm (Seiler et al, 2007) or infirm (Iyer & Rhind, 2017a) the existence of an elongation checkpoint. Here, we have induced a limited number of DSBs (11–48 breaks per haploid genome) and found that Rad9 and Rad53 acts in trans to slow down elongation at otherwise undamaged forks. Importantly, this role of Rad9 in fork elongation was observed both in MMS‐ and Zeocin‐treated cells with at least four different assays, namely BrdU‐IP‐chip, DNA combing, PFGE, and 2D gel electrophoresis. Whether this downregulation of fork speed is important for viability in response to DNA damage is currently unclear, but it is worth noting that Rad9 is more important for viability than Mrc1 when cells are exposed to UV or MMS (Osborn & Elledge, 2003; Fong et al, 2013).

The mechanism by which Rad53 negatively controls elongation in face of damaged DNA is currently unknown. Proteomic studies have identified several Mec1 and Rad53 targets that are phosphorylated after replication stress (Smolka et al, 2007; Chen et al, 2010; Bastos de Oliveira et al, 2015). Among them, a limited number of Rad53 targets are associated with the replisome and could be relevant for the regulation of elongation in MMS‐treated cells. These include Mrc1 and Tof1, two factors required for normal fork progression (Tourrière et al, 2005) and Pol1, the catalytic subunit of DNA polymerase α primase (Smolka et al, 2007; Chen et al, 2010; Bastos de Oliveira et al, 2015). Future studies will be required to determine whether their phosphorylation modulates replication elongation.

Another important question that remains to be addressed is to what extent this model applies to metazoan. Claspin is the vertebrate homolog of Mrc1 (Kumagai & Dunphy, 2000), and Rad9 has several functional orthologues, 53BP1, BRCA1, and MDC1, which share overlapping functions (Canman, 2003). Claspin is essential to activate CHK1 in response to fork stalling and to regulate replication origin firing (Chini & Chen, 2004; Masai et al, 2017). CHK1 also regulate elongation in response to DNA damage (Seiler et al, 2007), but its mechanism of action has been difficult to address as it is also involved in the regulation of dNTP pools, which directly affect the speed of replication forks (Técher et al, 2016). In light of our data, it would be interesting to test whether induction of a limited number of DSBs in CHK1‐proficient human cells is sufficient to actively slow down replication forks, independently of the regulation of dNTP pools.

In conclusion, our data show that the checkpoint mediators Mrc1 and Rad9 play distinct but complementary functions in response to DNA damage. Mrc1 promotes the rapid activation of Rad53 at stalled forks and the repression of late origins firing. Mrc1‐mediated Rad53 signaling decreases upon fusion of adjacent replicons but is taken over by the Rad9 pathway to regulate other aspects of the replication stress response, among which the slowing down of fork progression. This integrated regulation of initiation and elongation would ensure an optimal completion of DNA replication in response to DNA damage.

Materials and Methods

Yeast strains

All strains used in this study are listed in the following table. They were derived from W303 and corrected for the rad5‐535 mutation (RAD5).

Cells were grown at 25°C in YEP medium supplemented with 2% glucose. Cells were arrested in G1 for 170 min with 8 μg/ml α‐factor (Biotem, France) and released from G1 by the addition of pronase (75 μg/ml) into medium containing 0.033% or 0.1% MMS or 0.2 M HU. Zeocin was added at 100 μg/ml either 45 min before G1 release or at time 60 min post‐G1 release in the presence of 0.2 M HU. For genome‐wide BrdU‐IP‐chip experiments, 400 μg/ml BrdU were added 20 min before G1 release. CldU and IdU were added, when indicated, at a 400 μg/ml and a 200 μg/ml concentration, respectively. Thermosensitive cdc7‐4 cells were arrested at the G1/S transition by shifting the culture to 37°C 45 min before release by pronase addition. Cell cycle progression was monitored by flow cytometry.

Genome‐wide analysis of BrdU incorporation sites

Immunoprecipitation of BrdU‐labeled DNA and hybridization on Affymetrix tiling arrays was performed as described (Crabbé et al, 2010). Replication profiles were displayed with Integrated Genome Browser (IGB) as a signal log ratio of BrdU‐IP DNA sample against the Input DNA. Quantification of the BrdU incorporation for a 1‐kb window is expressed as the sum of the signal detected within this window for individual replication origins and is plotted relative to their mean replication time during a normal S phase (Müller et al, 2014). Statistical analyses were performed with GraphPad Prism 7.

DNA combing

DNA combing was done as described previously (Bianco et al, 2012; Tourrière et al, 2017). ssDNA was detected using anti‐mouse MAB 3034 (Millipore) and goat anti‐mouse coupled to Alexa 546 (red) or Alexa 647 (far red); BrdU or CldU was detected with a rat monoclonal antibody (Abcys, clone BU1/75) followed by a goat anti‐rat coupled to Alexa 488 (Invitrogen). IdU was detected with a mouse anti‐BrdU (BD Biosciences 347580). Images were recorded on a Leica DM6000 or on a Zeiss Apotome microscope and were processed as described earlier by extracting DNA fibers from the background (Pasero et al, 2002). Statistical analyses were performed with GraphPad Prism 7. To estimate the number of breaks induced by Zeocin, we have measured the length of DNA fibers by DNA combing, before and after Zeocin exposure. The total length of the 16 yeast chromosomes is 13,891 kb (12,071 kb of unique sequences plus 1,820 kb of rDNA). The mean length of DNA fibers (m) is therefore 13,891/16 = 868 kb. If the genome is broken once, the mean length of DNA fibers should be 13,891/(16 + 1) = 817 kb. It is therefore possible to derive the number of breaks (k) from the mean length of DNA fibers (m) using the following equation: m = 13,891/(16 + k) or k = (13,891−16*m)/m. For untreated cells (m = 321 kb), the number of breaks (k) is 27. Since the maximal length of DNA fibers that can be measured in a field of view is 400 kb, 19 of these breaks correspond to truncated molecules, the rest being caused by mechanical shearing of DNA fibers during the DNA combing procedure. The estimated number of breaks in wt cells exposed to Zeocin is 38−27 = 11 after 45 min and 75−27 = 48 after 90 min.

2D and pulsed‐field gel electrophoresis

Two‐dimensional gel analysis of replication intermediates was done as described (Tourrière et al, 2005). DNA extraction is based on the isolation of yeast nuclei and utilizes a modified version of the Qiagen Genomic DNA Handbook protocol of the Qiagen DNA Isolation Kit (Valencia, CA). Pulsed‐field gel electrophoresis was performed as described (Lengronne et al, 2001). Chromosomes were separated at 13°C in a 0.9% agarose gel in TBE 0.5× using a Rotaphor apparatus (Biometra) using the following parameters: interval from 100 to 10 s (logarithmic), angle from 120 to 110° (linear), voltage 200 to 150 V (logarithmic). The gel was subsequently stained with ethidium bromide and transferred to Hybond XL (GE Healthcare). Quantification of chromosome intensity was performed after Southern blotting and hybridization using a radioactive probe specific for chromosome VII (ADH4 gene), using a PhosphorImager (Typhoon Trio, GE).

Protein extraction and Western blot

Approximately 5 × 108 cells were collected at each relevant time point and were washed with 20% trichloroacetic acid to prevent proteolysis, then resuspended in 200 μl of 20% trichloroacetic acid at room temperature. The same volume of glass beads was added, and cells were disrupted by vortexing for 10 min. The resulting extract was spun for 10 min at 1,000 g also at room temperature and the resulting pellet resuspended in 200 μl of Laemmli buffer. Whenever the resulting extract was yellow‐colored, the minimum necessary volume of 1 M Tris base (non‐corrected pH) was added till blue color was restored. Then, water was added till a final volume of 300 μl was reached. These extracts were boiled for 10 min and clarified by centrifugation as before; 10–15 μl of this supernatant was loaded onto a 3–8% acrylamide gradient Invitrogen gel and migrated 70 min at 150 V to separate Rad53 isoforms, then proteins transferred to a nitrocellulose membrane. Detection by immunoblotting was accomplished with anti‐Rad53 antibody, a kind gift from Dr. C. Santocanale.

In situ phosphorylation assay (ISA)

The exact same amount of protein extract described in the previous section was loaded in a 7.5% acrylamide Bio‐Rad gel and migrated 70 min at 150 V, then transferred to a PVDF membrane. The rest of the protocol was done strictly as described in Pellicioli et al (2001).

Chromatin immunoprecipitation

Chromatin immunoprecipitation was performed as in Lengronne et al (2004) with minor modifications. 1 × 109 cells were crosslinked for 15 min with 1% formaldehyde (Sigma F8775) at RT under agitation. Fixation was quenched by addition of 0.25 M glycine (Sigma G8898) for 5 min under agitation. Cells were washed three times with cold 1× TBS (4°C). Dry pellets were frozen and conserved at −20°C. Cell pellets were resuspended in lysis buffer (50 mM HEPES‐KOH pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X‐100, 0.1% Na‐deoxycholate) supplemented with 1 mM PMSF and anti‐protease (cOmplete Tablet, Roche, 505649001) and lysed by beads‐beat method (MB400 U, Yasui Kikai, Osaka). Recovered lysate (WCE, Whole Cell Extract) was completed to 3 ml with cold lysis buffer and sonicated with a Q500 sonicator (Qsonica) (12 cycles: 15 s ON, 45 s OFF, amplitude 50). Dynabeads were washed three times and resuspended in 1 ml of PBS, 0.5% BSA, 0.1% Tween and incubated with antibodies on a rotating wheel for two hours at 4°C. 40 μl of anti‐PK (Anti‐V5 tag, AbD Serotec, MCA1360G) with 180 μl Dynabeads Prot. A (DPA) 25 μl of WCE was kept for the Input sample and 25 μl was collected for Western blotting (WB). Antibody‐coupled Dynabeads were washed three times with 1 ml of PBS, 0.5% BSA, 0.1% Tween and added to 2.7 ml of WCE for 4 h on a rotating wheel at 4°C. Beads were then collected on a magnetic rack. 25 μl of the supernatant was collected for WB analysis (Flow‐Through sample), and beads were washed on ice with cold solutions two times with Lysis buffer (50 mM HEPES‐KOH pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X‐100, 0.1% Na‐deoxycholate), twice with Lysis buffer + 0.36 M NaCl (50 mM HEPES‐KOH pH 7.5, 360 mM NaCl, 1 mM EDTA, 1% Triton X‐100, 0.1% Na‐deoxycholate), twice with Wash buffer (10 mM Tris–HCl pH 8, 0.25 M LiCl, 0.5% IGEPAL, 1 mM EDTA, 0.1% Na‐deoxycholate), and once with TE (10 mM Tris–HCl pH 8, 1 mM EDTA). Antibodies were uncoupled from beads with 40 μl of Elution Buffer (50 mM Tris–HCl pH 8, 10 mM EDTA, 1% SDS) for 10 min at 65°C. 5 μl of eluates was collected for WB (IP sample), and 35 μl was incubated with 120 μl of TE, 0.1% SDS for de‐crosslinking at 65°C for 6 h. 130 μl of TE containing 60 μg RNase A (Sigma, R65‐13) was added to the samples and incubated for 2 h at 37°C. Proteins were digested by addition of 20 μl of Proteinase K (Sigma, P6556) at 20 mg/ml and incubated for 2 h at 37°C. 50 μl of 5 M LiCl was added to DNA before purification by two rounds of Phenol:Chloroform:Isoamyl Alcohol 25:24:1 (Sigma, P2069) extractions and precipitation by addition of 100 mM Sodium Acetate (Sigma, S2889), 26 μg/ml of Glycogen (Roche, 10901393001), and two volumes of 100% ethanol overnight at −20°C. Samples were centrifuged for 45 min at 16,000 RCF at 4°C, washed with cold 70% ethanol, and centrifuged again 15 min at 16,000 RCF at 4°C. DNA pellets were dried and resuspended in 300 μl of H2O prior to qPCR products. qPCR product was performed with LightCycler480 (Roche). IP/Input ratio was calculated, and qPCR results were normalized on a late‐replicated zone (GLT1 on chromosome IV). Primers used for qPCR are listed in Table EV1.

dNTPs measurement

35 ml of logarithmically growing yeast cells at 1 × 107 cells/ml was harvested on nitrocellulose membrane and suspended immediately in an ice‐cold mixture of 12% TCA and 15 mM MgCl2. Cells were vortex‐mixed for 15 min at 4°C and then centrifuged at 16,200 g for 1 min at 4°C. The supernatant was neutralized with a freon–trioctylamine mix and processed as described previously (Jia et al, 2015).

Data accessibility

Microarray data were deposited at NCBI Gene Expression Omnibus (GEO). Accession number: GSE109654.

References

Download references