Mutations affecting DNA-replication coupled nucleosome assembly exhibit a synthetic interaction with ctf4Δ

The HIR complex (formed by Hir1, Hir2, Hir3, and Hpc2) promotes replication-independent chromatin assembly [101]. This complex is important for normal growth and silencing in the absence of CAF-1, indicating functional overlap between HIR and CAF-1 complexes [102]. We found that HIR1 was dispensable for the growth of ctf4Δ cells (Figures S1B and S2J), suggesting that the replication-independent chromatin assembly function is not important for ctf4Δ cells. We next conducted an extensive genetic analysis to assess the importance of genes encoding histone chaperones and histone acetyltransferases in ctf4Δ mutant cells (Figure S2). We first confirm that RTT106 inactivation affected ctf4Δ cell growth less severely than CAC1 inactivation (Figure S2E). Among many candidates tested we found that the Spt16 subunit of the heterodimeric FACT complex (Spt16-Pob3), which binds histone H3-H4 and Cac2 subunit of CAF-1, and functions in DNA replication-coupled nucleosome assembly [103], exhibited negative genetic interactions with ctf4Δ (Figure S2H). In addition, we show that rfa1-A88P mutation, which exhibits attenuated nucleosome assembly on nascent chromatin [104] was lethal with ctf4Δ (Figure S2K). Finally, we observed that the lysine acetyltransferase Gcn5, which regulates the interaction between H3-H4 and CAF-1 to promote the deposition of newly-synthesized histones [105] was also important for the normal growth of ctf4Δ cells (Figure S2M). We noticed that asf1Δ as well as rtt109Δ only slightly affected ctf4Δ cells (Figure S2D and L). H3 and H4 N-terminal tail acetylation serve as an important regulator of nucleosome assembly [26] [105] [106] [107]. We therefore, determine the importance of different H3-H4 histone modifications involved in nucleosome assembly, in ctf4Δ cells.

We next crossed a strain expressing H3K9,14,18,23,27R or H4K5,8,12R from a centromeric plasmid as the sole source of histone H3 or H4 with the ctf4Δ strain and analyzed the spores after diploids sporulation. We observed that in the absence of CTF4, the mutation at H3 or H4 lysine residues, implicated in nucleosome assembly, became deleterious (Figure 1B). We further analyzed the consequences of the absence of Nap1, which promotes H2A-H2B tetramer assembly in nucleosomes, in ctf4Δ cells. We found that nap1Δ did not negatively affect the growth of ctf4Δ cells (Figure S1C). Altogether, these genetic analyses indicate that defects in chromatin assembly during replication related to H3-H4 histones are deleterious in absence of CTF4. These results further support the notion that CAF-1 has a crucial role in a process linked to the Ctf4 function.

Cac1 is important for genome integrity in absence of CTF4

To eliminate the possibility that cac1Δ ctf4Δ growth defects originate from meiotic events, we deleted CAC1 in the ctf4Δ strain background by gene targeting. We first confirmed the slow growth phenotype, observed for cac1Δ ctf4Δ, during the segregation analyses and observed that this phenotype is amplified at both 25°C and 35°C (Figure 2A). We next examined the sensitivity of cac1Δ ctf4Δ mutant to both ultraviolet (UV) light and chronic exposure to DNA-damaging agents. We found that the cac1Δ ctf4Δ double mutant was more sensitive to UV, camptothecin (CPT), hydroxyurea (HU), and methyl-methanesulfonate (MMS) compared to every single mutant (Figure 2B). Then, we assessed in ctf4Δ mutant, the importance of Cac1 in the absence of Rrm3 which facilitates the progression of replication forks through non-histone DNA-protein complexes [108] [109]. We found that Cac1 was crucial for rrm3Δ ctf4Δ cells (Figure S3). These results indicate that replication stress and replication-induced DNA damage are lethal for cac1Δ ctf4Δ mutant. We next evaluated if the absence of Cac1 in ctf4Δ cells cause a synthetic interaction with mutations affecting the S-phase checkpoint pathway. We found that the absence of Mec1 or Rad53 kinases, that activate both branches of the S-phase checkpoint pathway [110], strongly affect the viability of cac1Δ ctf4Δ cells (Figure 2C).

Open in a separate window

Figure 2

FIGURE 2: CAC1 is important for growth in the absence of CTF4.

(A) cac1Δ ctf4Δ growth is affected at various temperatures. Tenfold serial dilutions of wild-type, cac1Δ, ctf4Δ, and cac1Δ ctf4Δ cells were spotted onto YPD plates and incubated at 30° (left), 25° (middle), or 36° (right) for 3 days. (B) cac1Δ ctf4Δ growth is affected in presence of DNA damage. Tenfold serial dilutions of wild-type, cac1Δ, ctf4Δ, and cac1Δ ctf4Δ cells were assayed on normal growth media (YPD), after UV irradiation or not, and on media containing the indicated DNA-damaging agents, camptothecin (CPT), hydroxyurea (HU), and methyl-methanesulfonate (MMS). (C) The S-phase checkpoint is required for cac1Δ ctf4Δ mutant viability. The diploid strains mec1Δ/MEC1 sml1Δ/SML1 cac1Δ/CAC1 ctf4Δ/CTF4 (left) and rad53-K227A/RAD53 cac1Δ/CAC1 ctf4Δ/CTF4 (right) were sporulated and one hundred tetrads were dissected on YPD plates and incubated at 30° for 5 days. Four representative tetrads are shown for each dissection. mec1Δ sml1Δ and rad53-K227A mutations are lethal in cac1Δ ctf4Δ cells.

To evaluate if homologous recombination (HR) was important for cac1Δ ctf4Δ mutant, we investigated whether cac1Δ ctf4Δ leads to an increase in spontaneous Rad52 foci, which reflect HR proteins recruitment into repair foci [111] [112]. We found that cac1Δ ctf4Δ cells exhibited a higher frequency of Rad52-YFP foci compared to cac1Δ and ctf4Δ single mutants (Figure 3A, left). We also found that cac1Δ ctf4Δ exhibited an abnormally high frequency of spontaneous Rfa1 foci compared to that in cac1Δ and ctf4Δ single mutants (Figure 3A, right). These results suggest that the absence of Cac1 spontaneously creates chromosome breaks or ssDNA gaps during replication in ctf4Δ cells and that these damages, are repaired by HR.

Open in a separate window

Figure 3

FIGURE 3: Genomic integrity is affected in cac1Δ ctf4Δ cells.

(A) Left, Rad52 foci are increased in cac1Δ ctf4Δ cells. Wild-type, cac1Δ, ctf4Δ, and cac1Δ ctf4Δ cells encoding Rad52-YFP were analyzed with fluorescence microscopy. Right, Rfa1 foci are increased in cac1Δ ctf4Δ cells. Wild-type, cac1Δ, ctf4Δ, and cac1Δ ctf4Δ cells encoding Rfa1-CFP were analyzed with fluorescence microscopy. Numbers indicate the percentage of cells that contained Rad52-YFP (left) or Rfa1-CFP foci (right). DNA replication was monitored by FACS analysis of DNA content. At least 200 cells were analyzed for each strain from three independent experiments. Statistical significance was measured using the two-tailed Mann-Whitney test. (B) γH2AX is increased in the absence of Cac1 in ctf4Δ cells. Left, Western blot was used to detect phosphorylation of H2A serine 129 (γH2AX). Right, histogram shows for WT, cac1Δ, ctf4Δ, and cac1Δ ctf4Δ, γH2AX/H2A ratios calculated based on Western blots signal intensities. DNA replication was monitored by FACS analysis of DNA content. The experiment has been done in duplicate. (C) CAF-1 inactivation increases mutation rate in ctf4Δ cells at the Tus/Ter barrier. Top, Schematic representation of the unidirectional and site-specific Tus-Ter replication fork barrier. Ter sequence is integrated 3.5 kb downstream ARS305 on Chromosome III where Tus protein (colored rectangles) binds specifically to Ter sequence, causing replication fork pausing. Upstream to Tus/Ter replication fork barrier is the URA3 reporter gene, which permits the positive selection for ura3 mutations in presence of 5-FOA to measure mutation rate. Bottom, exponentially growing cells expressing Tus protein were plated for 3 days at 30° on YPGal plates and plated out on 5-FOA to select for ura3 mutation. Box-and-whisker plots, representing the upper and lower quartile with the median, show the mutation rate in WT, cac1Δ, ctf4Δ, and cac1Δ ctf4Δ cells. Statistical analyses were done on n=5 independent experiments using two-tailed Mann-Whitney test; *p<0.05; ** p< 0.005; **** p<0.0001; ns, not significant.

In S. cerevisiae, phosphorylation of histone H2A on serine 129 (γH2AX) is tightly associated to DNA damage, and it has been shown that DNA double-strand breaks levels can be obtained by measuring levels of γH2AX [113] [114]. To determine if the absence of Cac1 increased DNA damage in ctf4Δ cells, we analyzed by Western blot γH2AX levels, after a nocodazole-imposed mitotic arrest, (Figure 3B). In agreement with previously published results showing that cac1Δ mutant had only a weak requirement for replication or DNA-damage checkpoint proteins [99], we found a slight increase of H2A phosphorylation in absence of Cac1 compared to wild-type (WT) cells. In contrast, we observed a more consistent increase over WT cells in cac1Δ ctf4Δ mutant, indicating that the absence of Cac1 induces DNA damages in ctf4Δ cells.

Having found that Cac1 was required for the viability of ctf4Δ mutant in absence of Rrm3 helicase (Figure S3), which helps replication fork traverse protein-DNA complexes [108] and assists fork progression across TERs [109], we sought to evaluate whether loss of CAF-1 function affected genome stability in ctf4Δ cells. For this end, we used the natural Escherichia coli Tus/Ter barrier system known to induce an unidirectionally and site-specific replication fork stalling in S. cerevisiae. Interestingly, this system, which reflects the natural protein-bound DNA barriers arising in yeast cells, represents one of the most physiological replicative stresses that yeast cells could encounter [115] [116] [117]. The Tus/Ter barrier system, composed by a 21-bp DNA sequence (Ter) which is bound by the Tus terminator protein, is coupled with the genetic URA3 gene reporter located immediately upstream the Tus/Ter barrier (Figure 3C, top), allowing us to quantify mutagenic outcomes as previously described [118] (Ghaddar et al 2023 Nat Commun, in press). In correlation with our previous data showing an increased Rad52 and Rfa1 foci formation and an increased level of γH2AX in cac1Δ ctf4Δ double mutant, we observed an increase in mutation rate in cac1Δ ctf4Δ cells compared to cac1Δ and ctf4Δ cells expressing Tus protein (Figure 3C, bottom).

Finally, we found that the elimination of Mad2 spindle checkpoint, which causes impairment of microtubule-kinetochore attachment and incomplete sister chromatid cohesion, negatively affects the growth of cac1Δ ctf4Δ cells (Figure S4). Overall, these results reflect the requirement of both DNA damage and spindle checkpoint for the viability of cac1Δ ctf4Δ cells and that HR is required for cell growth in the absence of both CAC1 and CTF4. They show that loss of CAF-1 function causes DNA damages and leads to spontaneous mutation in ctf4Δ cells, revealing that CAF-1 prevents DNA damage formation and maintains genome stability.

Dissecting the importance of Cac1 in functions mediated by Ctf4

Ctf4 performs different functions in DNA metabolism. To explore the significance of the negative genetic interaction between CAC1 and CTF4, we sought to investigate which function of Ctf4 is responsible for this negative interaction. To determine if the growth defect observed in cac1Δ ctf4Δ mutant is a direct effect of uncoupling between helicase and DNA-polymerase α (Polα) [82] [83] [119], we analyzed the consequences of inactivating CAC1 in the thermosensitive (ts) mutant cdc17-1 encoding the catalytic subunit of Polα. We found that cac1Δ did not affect the viability of cdc17-1 cells (Figure S5A), which indicates that the growth defect of cac1Δ ctf4Δ was not related to the ctf4Δ cell's inability to incorporate Polα into the replisome. This result shows that replication fork architecture defects, due to uncoupling arising in ctf4Δ mutant, are not responsible for the synthetic fitness defects observed in cac1Δ ctf4Δ cells. Moreover, the loss of Ctf4 affects DNA damage tolerance function due to faulty MCM-uncoupled Polα/Primase activity [93]. Thus, this result also indicates that the severe growth defect observed in cac1Δ ctf4Δ was not a consequence of defective DNA damage tolerance. We next searched for synthetic growth defects caused by combining cac1Δ mutation with the mcm2-3A mutation known to affect the Ctf4-Mcm2-Polα-dependent transfer of (H3-H4) parental histones [85] [120]. We found that affecting the parental histones transfer onto the lagging strand did not cause growth defects in absence of CAF-1 function (Figure S5B).

Another major function assigned to Ctf4 is associated with its role in cohesion establishment. Genetic analyses have defined two pathways for cohesion establishment at the replication fork, one containing CTF4, CHL1, CSM3, TOF1, and the second containing MRC1, CTF18, CTF8, DDC1 [69]. We addressed if any of these non-essential replisome proteins, previously implicated in cohesion establishment, were required for growth in cac1Δ cells. We found that only CTF4 inactivation became deleterious in absence of CAC1 (Figure S5C-J).

These data suggest that the important growth defect observed in the ctf4Δ cac1Δ mutant was unrelated to the Ctf4 functions mentioned above.

CAC1 is required to sustain the viability of cells affected in essential cohesion pathways

Because Ctf4 is involved in multiple cohesion establishment pathways, its absence negatively influences sister chromatid cohesion in different ways [69] [77] [93] [121]. Among the non-essential proteins associated with replisomes involved in the two parallel pathways for cohesion establishment at the replication fork [69], ctf4Δ mutant is the only one that causes lethality in the absence of the S-phase acetyltransferase ECO1 [121], a protein that locks sister chromatid entrapment by acetylating Smc3 both during S phase and in response to DNA damage [54] [55] [123] [124]. Therefore, we wondered whether affecting cohesion, in a more severe way than deleting non-essential cohesion genes, might reveal the importance of CAC1 for the growth of cells affected in cohesion. We first investigated the relationship between CAC1 and the essential cohesion genes ECO1. Since the inactivation of RAD61/WPL1, which counteracts cohesion-establishing reaction, suppresses eco1Δ lethality [54] [55], we deleted one allele of ECO1 and one allele of RAD61 in a cac1Δ/CAC1 diploid strain. We found that cac1Δ rad61Δ cells grew normally and that cac1Δ cells exhibited a synthetic sick phenotype with eco1Δ rad61Δ mutations at 30°C (Figure 4A, left). RAD61 inactivation impacts cohesion, chromatin structure, and intra-chromosomal loops organized by cohesins [125] [126] [127]. Thus, we also analyzed the consequence on cac1Δ cells of the temperature-sensitive eco1-1 mutation which confers severe cohesion defects at 37°C, and to a lesser extent, at lower temperatures [53]. We found that the eco1-1 mutant was strongly affected by CAC1 inactivation (Figure 4A, right), which confirmed that Cac1 was required for efficient growth in eco1 mutants. These data suggest that Cac1 could act, at least in part, in a cohesion pathway parallel to the Smc3ac function.

Open in a separate window

Figure 4

FIGURE 4: Essential sister chromatid cohesion genes are required for the viability of CAF-1 deficient cells.

(A) Genetic interaction of CAC1 with ECO1. Left, fifty tetrads from eco1Δ/ECO1 cac1Δ/CAC1 rad61Δ/RAD61 diploid strain were dissected. Tetrads were grown at 30° for 5 days. Dashed circles indicate eco1Δ cac1Δ rad61Δ mutant. Right, top, tenfold serial dilutions of indicated genotypes were spotted onto YPD plates and incubated at 30° for 3 days. Right, down, tenfold serial dilutions of wild-type, cac1Δ, eco1-1, and eco1-1 cac1Δ cells were spotted onto YPD plates and incubated at 25° (left), or 36°C (right) for 3 days. (B) Genetic interaction of CAC1 with SCC2. Tenfold serial dilutions of exponentially growing cells were spotted onto YPD plates and incubated at 25° (left), or 32° (right) for 3 days. (C) Genetic interaction of CAC1 with SCC1. Tenfold serial dilutions of exponentially growing cells were spotted onto YPD plates and grown at 25° (left), or 32° (right) for 3 days.

To further test whether in the absence of CAC1 cohesion defects are deleterious, we combined cac1Δ with mutations in genes that affect the cohesin loader (Scc2-Scc4) and the cohesin ring complex (Scc1). We first analyzed the consequences of CAC1 deletion in scc2-4 temperature-sensitive mutant defective in cohesion establishment. Whilst at permissive temperature (25°C), scc2-4 cac1Δ cells were indistinguishable from scc2-4 and cac1Δ, the double mutant cells were strongly affected at the semi-permissive temperature of 30°C (Figure 4B). We next focused on Scc1, which is involved in both establishment of cohesion in S phase and the maintenance of cohesion in G2/M. Using the temperature-sensitive scc1-73 allele [128], we found that impairment of cohesin function strongly affects cell viability in the absence of any of CAF-1 subunits, Cac1, Cac2, or Cac3 (Figure 4C, and Figure S6A). Interestingly, we observed a slight difference in the severity of cac1Δ, cac2Δ, and cac3Δ mutant phenotypes, with the cac3Δ mutant showing a weaker phenotype (Figure S6A). Again, this could be explained by previous observations showing that CAC3 inactivation has a minor effect on nucleosome formation compared to CAC1 or CAC2 inactivation [100]. The similar weaker phenotype observed for cac3Δ compared to cac1Δ and cac2Δ mutants in both scc1-73 (Figure S6A) and ctf4Δ cells (Figure 1 and Figure S2), suggests that the requirement for Cac1 in ctf4Δ cells is related to important cohesion defects linked to the absence of Ctf4. cac1-F233L and cac1-F233A,F234G mutations were previously shown to alter Cac1 binding to PCNA as well as Cac1 DNA-replication-linked nucleosome assembly function [17]. To investigate the importance of Cac1 recruitment to the replication forks through its interaction with PCNA, we determined the impact of cac1-F233L and cac1-F233A, F234G mutations in scc1-73 cells. We found that both mutations caused growth defects for scc1-73 cells at permissive temperature (Figure S6B). These data indicate that the PCNA-dependent CAF-1 recruitment to chromatin is required when cohesin function is affected. These results confirm that CAF-1 complex is required for cell viability when cohesion is affected. On the other hand, CAF-1 and Rtt106 function in a coordinated manner in nucleosome assembly [107]. Henceforth, we tested whether RTT106 deletion affects the growth of the scc1-73 mutant. A significant growth defect was observed in rtt106Δ scc1-73 cells at the semi-permissive temperature compared to each rtt106Δ and scc1-73 single mutant (Figure S6C). In agreement with the fact that RTT106 exerts a minor role compared to CAF-1 in new histones deposition during DNA synthesis [26], we found that the growth defect of the rtt106Δ scc1-73 mutant was much weaker than that of the cac1Δ scc1-73 mutant. Furthermore, cac1Δ rtt106Δ scc1-73 triple mutant exhibited a more dramatic effect compared to the rtt106Δ scc1-73 and cac1Δ scc1-73 double mutants (Figure S6C). Taken together these results indicate that CAF-1, and more generally nucleosome assembly during replication, is crucial in maintaining genome stability when SCC is compromised. These results, in association with our previous findings showing that cac1Δ ctf4Δ cells are more affected than rtt106Δ ctf4Δ cells and that RTT106 inactivation amplified the growth defect of cac1Δ ctf4Δ cells (Figure 1A), strongly suggest that the requirement for CAF-1 in ctf4Δ mutant was related to important cohesion defects linked to the absence of CTF4.

Ctf4 interacts with Mms22 [82] [84] [90], an adaptor protein of the Rtt101-Mms1 E3 ubiquitin ligase complex to tether the Rtt101-Mms1-Mms22 E3 ubiquitin ligase to active replisome during S-phase [86]. Previous studies showed that Rtt101-Mms1-Mms22 inactivation is epistatic with ctf4Δ regarding cohesion defects [37]. We first assessed the importance of the ubiquitin ligase in ctf4Δ cells and found that its absence did not significantly affect growth (Figure S7). We next addressed the consequences of deleting RTT101, MMS1, or MMS22, in cac1Δ, scc1-73, and cac1Δ scc1-73 cells. We found that at 30°C, the growth of cac1Δ rtt101Δ, cac1Δ mms1Δ, and cac1Δ mms22 double mutants was not strongly affected compared to one of the rtt101Δ, mms1Δ, and mm22Δ single mutants (Figure 5A, ​,B,B, ​,CC, left). These results strongly suggest that sister chromatid cohesion established through the interaction between Rtt101-Mms1-Mms22 E3 ubiquitin ligase and Ctf4 has little or no involvement in the growth defect observed for cac1Δ ctf4Δ cells. Moreover, we found that impairment of cohesin function dramatically affected cac1Δ viability compared to rtt101Δ or mms1Δ (Figure 5A, ​,BB, left). We observed a higher impact for MMS22 inactivation compared to RTT101 or MMS1 inactivation. This may be explained by the multiple roles of Mms22 in response to DNA damage [31] [92] [96] and/or by an Mms22 role in promoting cohesion through its direct interaction with Eco1 [37] [39]. Interestingly, at the permissive temperature for the scc1-73 mutant (25°C, right), we observed that scc1-73 cac1Δ rtt101Δ, scc1-73 cac1Δ mms1Δ, and scc1-73 cac1Δ mms22Δ cells are more affected than each of the double mutants, suggesting that CAF-1 could act in parallel with the Scc1 and Rtt101-Mms1-Mms22 cohesion pathways. Taken together these genetic analyses indicate that the growth defect of cac1Δ ctf4Δ cells is not the consequence of cohesion defects caused by a deficiency in the Rtt101-Mms1-Mms22-Ctf4 replication-coupled sister chromatid cohesion pathway, but rely on another pathway independent of the Rtt101-Mms1-Mms22 E3 ubiquitin ligase complex.

Open in a separate window

Figure 5

FIGURE 5: Cac1 functions in a different pathway from that of Rtt101-Mms1-Mms22 in cohesion. Genetic interactions among RTT101, MMS1, MMS22, SCC1, and CAC1.

(A) Effect associated with RTT101 inactivation on viability of cac1Δ, scc1-73, and cac1Δ scc1-73 cells. (B) Effect associated with MMS1 inactivation on viability of cac1Δ, scc1-73, and cac1Δ scc1-73 cells cells. (C) Effect associated with MMS22 inactivation on the viability of cac1Δ, scc1-73, and cac1Δ scc1-73 cells. Genetic interactions were assessed by spotting a tenfold series dilution of cells of the indicated genotype onto YPD. Cells were grown at 30° for 3 days (left), or at 25°C (right) for 4 days.

Cac1 is required for sister chromatid cohesion in ctf4Δ cells

In contrast to ctf4Δ cells, which present important cohesion defects [129], cac1Δ cells exhibit almost none or only moderate cohesion defects according to previous studies [37] [121]. To address whether the synthetic growth defect of cac1Δ ctf4Δ cells could be due to an additive effect on sister chromatid cohesion, we compared cohesion in metaphase-arrested cac1Δ and ctf4Δ cells to cohesion in cac1Δ ctf4Δ double mutant. To evaluate sister chromatid cohesion, we used a haploid strain containing Lac operator tandem repeats integrated at a site near the centromere of chromosome III and expressing a GFP-Lac repressor fusion protein. We found that ctf4Δ cells exhibited cohesion defects and that cohesion was not significantly affected by the absence of CAF-1, suggesting that CAF-1 is not important for cohesion in a wild-type context. However, cohesion defects in the cac1Δ ctf4Δ double mutant were significantly more severe than that in ctf4Δ single mutant (Figure 6). While we do not know the reason behind this increased cohesion defect, this result together with our genetic analyses, strongly suggests that, when cohesion is severely affected, CAF-1 exerts an important role required for cohesion maintenance and cell viability.

Open in a separate window

Figure 6

FIGURE 6: Cac1 is important to maintain cohesion in ctf4Δ cells.

Sister chromatid cohesion was analyzed by monitoring the tagged centromere of chromosome III. Top, experimental design. DNA replication was monitored by FACS analysis of DNA content. Over 100 cells were counted for each experiment. The results represent the average of five independent experiments. Statistical significance was measured using Two Tailed Mann-Whitney test.

Cohesin association to chromatin is increased in the absence of Cac1 in WT and ctf4Δ cells

We next investigated if the level of Scc1, at known cohesin binding sites on chromatin, was affected in the absence of CAF-1 function by performing ChIP experiments in WT and cac1Δ cells. To our surprise, we found that deletion of CAC1 markedly increased cohesin levels both at centromeres and chromosome arms (Figure 7A). Cohesin enrichment is enhanced genome-wide in response to DSB induction and this enrichment at undamaged sites globally tight-ens sister chromatid cohesion [124] [130] [131]. Taking into account that CAF-1 plays multiple roles in maintaining genome stability [3] [6] [7] [11] [13] [96] [98] [122] and that the absence of Cac1 in both wild-type and ctf4Δ cells increased Rad52 foci (Figure 3A, left), Rfa1 foci (Figure 3A, right) [7], and γH2AX (Figure 3B), a possible explanation could be that DNA-damage-induced cohesion establishment was the source of the high cohesin level arising in absence of Cac1. Because Chk1, which mediates the DNA damage response in parallel with RAD53, is a key component of the damage-induced cohesion establishment pathway required for the generation of damage-induced cohesion [134], we analyzed cohesin levels in absence of Chk1 in WT and cac1Δ cells. We observed that the cohesin levels were indistinguishable between chk1Δ and wild-type cells and found that deletion of CHK1 in cac1Δ mutant reduced the association of Scc1 at both centromeres (CEN3, CEN9) and chromosome arms (PAO1) (Figure 7B). This indicates that DNA damage-induced cohesion establishment is the source of the increased Scc1 level observed in absence of CAC1.

Open in a separate window

Figure 7

FIGURE 7: CAC1 inactivation increases cohesin level at chromatin in WT and ctf4Δ cells.

Top, experimental design. In all experiments, DNA replication was monitored by FACS analysis of DNA content. (A) CAC1 inactivation increases cohesin level at centromere and chromosome arm. Cells of the indicated genotype were synchronized in G1 and were released into nocodazole imposed mitotic arrest for 120 minutes. Scc1 levels at two centromeres (CEN3 and CEN9), a chromosome arm cohesin binding site (POA1), and a negative control-binding site (GLT1) were measured by ChIP, followed by real-time qPCR. SEM shown represents four independent experiments. (B) The increased cohesin level at chromatin observed in absence of CAC1 is due to DNA damage. ChIP-qPCR analyses of Scc1 level at centromere and chromosome arm in WT, cac1Δ, chk1Δ, and chk1Δ cac1Δ. Same experimental conditions as in (A). (C) ChIP-qPCR analyses of Scc1 level at centromere and chromosome arm in WT, scc2-4, cac1Δ, and cac1Δ scc2-4. (D) CAC1 inactivation increases cohesin level at centromere and chromosome arm in absence of CTF4.

Cohesin association with centromeres, promoters, DSBs, and stalled replication forks depends on the Scc2-Scc4 complex [41] [45] [135] [136]. To determine whether the Scc2-Scc4 complex was involved in the higher level of cohesins observed in cac1Δ cells, we used the scc2-4 thermosensitive mutant and compared the level of Scc1 at centromeres and chromosome arms in wild-type, scc2-4, cac1Δ and cac1Δ scc2-4 cells. To this end, we synchronized cells by α-factor pheromone block, released them into S-phase at 25°C, and subsequently cultivate them at the restrictive temperature for the scc2-4 mutation in the presence of nocodazole to arrest cells in G2/M. As expected, at restrictive temperatures, cohesin levels were strongly decreased in scc2-4 mutant compared to wild-type cells (Figure 7C). Interestingly, our chromatin immunoprecipitation experiments revealed similar cohesin levels both at centromeres and chromosome arms of scc2-4 and scc2-4 cac1Δ cells, revealing that the increased level of cohesin association observed in cac1Δ cells depends on the cohesin loader Scc2 (Figure 7C).

We further compared the level of Scc1 on chromatin in WT, cac1Δ, ctf4Δ, and cac1Δ ctf4Δ cells. As expected we found that Scc1 binding at chromatin was greatly reduced in ctf4Δ cells [77] [121]. Compared to cac1Δ cells, cac1Δ ctf4Δ cells exhibited a slight reduction in Scc1 binding, which interestingly, remained much higher than in ctf4Δ mutant and WT (Figure 7D). The fact that Scc1 occupancy is higher in cac1Δ ctf4Δ double mutant compared to ctf4Δ single mutant while the double mutant exhibits a stronger cohesion defect, suggests that cohesins are not fully functional in cac1Δ ctf4Δ cells, possibly because they are not able to reach an efficient cohesive state when ctf4Δ mutation is combined with CAC1 gene deletion.

Spore viability and growth

Diploid strains were sporulated at 25°C during 3 days on solid sporulation media and treated with 3 μl of 1 mg/ml zymolyase 20T (Seikagaku Biobusiness, Japan) during 10 minutes in water before tetrads were dissected on rich media (YPD plate). We used a MSN400 micromanipulator from SINGER Instruments. Viable colonies were scored 3 days or 5 days later. An average of 50 tetrads were dissected. The number of tetrads analyzed is denoted in the figure legends. Quantification of the spore growth was done by image analysis of area of growth using Image J.

Fluorescence microscopy

Microscopy analyses were carried out in liquid media supplemented in adenine using a Nikon Eclipse Ti microscope with a 100x objective. We used a Neo sCMOS camera (Andor) to collect the images. The exposure time was DIC: 500 ms; CFP: 75 ms, YFP: 75 ms, and GFP 75 ms. We used ImageJ to analyze the images on 2D-maximum projections from 11-Z-stacks spaced 0.5 μ each. All the cells analyzed were prepared by growing the cells at 30°C in YPD media supplemented in adenine.

Mutation rate analysis

Yeast cells were grown overnight in liquid YPD medium at 30°C. Cells were then diluted to 0.2 OD_600nm in 1 ml of water. Tenfold serial dilutions were done in water and each dilution was plated on YPGal plates to induced TUS expression and obtain separated colonies after 3 days growth at 30°C. YPGal plates were then replica plated twice onto 5-FOA plates to confirm the 5-FOA resistant phenotype of the growing colonies. Mutation rates were measured by fluctuation analysis [156] [157]. Statistical analyses were done on n=5 independent experiments using two-tailed Mann-Whitney test.

Chromatin Immunoprecipitation (ChIP)

We used the following steps to prepare chromatin samples: We first crosslinked the cells for 15 minutes with formaldehyde (1%) and used glycine (125 mM) to quench the reaction for 5 minutes. Cells were then lysed in 50 mM HEPES-KOH [pH7.5], 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate (FA lysis buffer) supplemented with cocktail of protease inhibitors by vortexing cells with glass beads (6 cycles × 20 s, with cooling between the cycles). After a centrifugation to remove debris, we then used a Bioruptor Pico sonicator to share the chromatin to around 200 bp. Insoluble materials were removed by centrifugation for 10 minutes (14,000 rpm, 4 °C). We mixed 500 μg of chromatin and the recommended amount of anti-V5 and conducted the immunoprecipitation overnight. We next added 25 μl of FA lysis buffer containing Protein G-Sepharose beads (3 hours incubation at 4 °C). Protein G-Sepharose beads were next washed successively once with FA buffer and twice with a FA buffer containing 500 mM NaCl. We next washed the protein G-Sepharose beads two times using a wash buffer composed of 10 mM Tris-HCl [pH 8.0], 0.25 M LiCl, 1 mM EDTA, 0.5% NP-40, 0.5% Na-Deoxycholate and one time with a TE buffer (10 mM Tris-HCl [pH8.0], 1 mM EDTA). The precipitated materials were eluted after incubation at 65 °C (10 minutes) in a buffer containing 50 mM Tris-HCl [pH7.5], 10 mM EDTA and 1% SDS. The decrosslinking step was performed overnight at 65 °C. We used the Invisorb Fragment CleanUp Kit to purify the DNA fragments.

Quantitative PCR analysis (qPCR)

All the qPCR experiments realized for individual gene analysis were conducted with the BioRad CFX384 qPCR machine using the following parameters: Five minutes at 95°C followed by forty cycles (15 s at 95 °C + 15 s at 50 °C + 40 s at 72 °C), followed by ten minutes at 95°C. The oligonucleotides used for the qPCR reactions are listed in Table S2.

Western blotting

Cells were grown in YPD and blocked in G2 as described previously. 10 ml cultures were collected and crosslinked with 1% formaldehyde for 10 minutes followed by quenching using 1.2M glycine. Pellets were resuspended in 5% TCA (5ml) and left on ice for 10 minutes, the pellets were snap frozen in liquid nitrogen and resuspended in 1 ml acetone at room temperature. The pellets are left to dry for at least 3 hours, and then resuspended in 100μl lysis buffer (50 mM Tris pH7.5, 1 mM EDTA pH7.5) 2.75 μl; 1M DTT, 20 μl 50x protease inhibitors). Glass beads were added to break cells in fast prep 3x45secs. 50 μl of 3xSDS sample buffer was added to the lysate followed by immediate heating at 95°C for 5 min, cooled and centrifuged before loading onto SDS-PAGE gels (8–10%). PAGE was carried out using a Bio-Rad Mini Trans-Blot System (Bio-Rad) in SDS running buffer (25 mM Tris, 190 mM glycine, 0.01% SDS). SDS-PAGE gels were transferred onto nitrocellulose membrane (0.45 μM, Amersham-GE Healthcare, Amersham, UK) in transfer buffer (25 mM Tris, 1.5% glycine, 0.02% SDS, 20% EtOH) in a Bio-Rad Mini Trans-Blot system. Membranes were blocked in 5% milk in PBS with 0.05% Tween20 (PBST) for at least 1 hr at room temperature before incubating in primary antibody in 2% milk/PBST overnight at 4°C. Membranes were washed in PBST three times for 15 min, incubated in secondary antibody in 2% milk/PBST for overnight at 4°C, and washed in PBST three times. Signals were detected with Amersham ECL detection reagents (RPN2105; Cytiva) and images were directly acquired with a ChemiDoc MP Imaging System (Bio-Rad). Primary antibodies used were rabbit anti-Smc3-K112,113Ac (a kind gift from Pr Adele Marston), mouse anti-V5 (Invitrogen), and rat anti-tubulin (Santa-Cruz). The ECL signals were quantified using Image Lab 6.0 (Bio-Rad). The quantification of the relative levels of proteins were calculated by normalizing the ratio signals of Smc3K112,113ac to tubulin. The statistical analyses were performed using one-tailed Mann-Whitney test.

We gratefully acknowledge Adele Marston for anti-Smc3-K112,113Ac antibody. We thank Armelle Lengronne for providing PP1418 and PP1628 strains, Jennifer Gerton for providing eco1-1 mutant strain, Qing Li for rfa1-A88P mutant, Sue Biggins for SBY885 strain, and Zhiguo Zhang for ZGY1069 and ZGY1077 strains. We acknowledge Paul D. Kaufman for the pRS414-CAC1, the pRS414-cac1-F233L, and the pRS414-cac1-F233A, F234G plasmids. We thank Hocine Mankouri and Ian Hickson for providing the Tus/Ter barrier system. Thanks to Michel-Hervé Moimême for helpful discussions and permanent support. V.G. laboratory is supported by the "Ligue Nationale Contre le Cancer" (LNCC) (Equipe labellisée). N.G. is supported by “Bourse du Ministère de l'Enseignement Supérieur, et de la Recherche (MESR)” and by "Fondation pour la Recherche Médicale (FRM)".