Persistent acetylation of histone H3 lysine 56 compromises the activity of DNA replication origins

In Saccharomyces cerevisiae, newly synthesized histone H3 are acetylated on lysine 56 (H3 K56ac) by the Rtt109 acetyltransferase prior to their deposition on nascent DNA behind replication forks. Two deacetylases of the sirtuin family, Hst3 and Hst4, remove H3 K56ac from chromatin following S phase. hst3Δ hst4Δ cells present constitutive H3 K56ac, which sensitizes cells to replicative stress via mechanisms that remain unclear. We performed a screen to identify genes that influence cell fitness upon nicotinamide (NAM)-induced inhibition of sirtuins. The screen revealed that DBF4 heterozygosity causes NAM sensitivity. DBF4 and CDC7 encode subunits of the Dbf4-dependent kinase, which activates origins of DNA replication. We show that i) cells harboring the dbf4-1 or cdc7-4 hypomorphic alleles are sensitive to NAM, ii) Rif1, an inhibitor of Cdc7-dependent activation of origins, causes DNA damage and replication defects in NAM-treated cells and hst3Δ hst4Δ mutants, and iii) cdc7-4 hst3Δ hst4Δ cells display synthetic temperature sensitivity associated with delayed initiation of DNA replication. Such replication defects are not due to activation of the intra-S phase checkpoint but require Rtt109-dependent H3 K56ac. Overall, these results suggest that persistent H3 K56ac sensitizes cells to replicative stress in part by negatively influencing replication origin activity.

ABSTRACT 1 In Saccharomyces cerevisiae, newly synthesized histone H3 are acetylated on lysine 56 (H3 2 K56ac) by the Rtt109 acetyltransferase prior to their deposition on nascent DNA behind 3 replication forks. Two deacetylases of the sirtuin family, Hst3 and Hst4, remove H3 K56ac 4 from chromatin following S phase. hst3∆ hst4∆ cells present constitutive H3 K56ac, which 5 sensitizes cells to replicative stress via mechanisms that remain unclear. We performed a 6 screen to identify genes that influence cell fitness upon nicotinamide (NAM)-induced inhibition 7 of sirtuins. The screen revealed that DBF4 heterozygosity causes NAM sensitivity. DBF4 and 8 CDC7 encode subunits of the Dbf4-dependent kinase, which activates origins of DNA 9 replication. We show that i) cells harboring the dbf4-1 or cdc7-4 hypomorphic alleles are 10 sensitive to NAM, ii) Rif1, an inhibitor of Cdc7-dependent activation of origins, causes DNA 11 damage and replication defects in NAM-treated cells and hst3∆ hst4∆ mutants, and iii) cdc7-4 12 hst3∆ hst4∆ cells display synthetic temperature sensitivity associated with delayed initiation of 13 DNA replication. Such replication defects are not due to activation of the intra-S phase 14 checkpoint but require Rtt109-dependent H3 K56ac. Overall, these results suggest that 15 persistent H3 K56ac sensitizes cells to replicative stress in part by negatively influencing 16 replication origin activity. 17 18 19 INTRODUCTION 1 DNA replication initiates at multiple origins throughout chromosomes during the S phase of 2 the cell cycle (1). During G1, Cdt1 and Cdc6 load the MCM helicase complex on DNA at 3 origins of replication bound by the Origin Recognition Complex (ORC). At the beginning of S 4 phase, cyclin-dependent (CDK) and Dbf4-dependent (DDK) kinase activities promote the 5 recruitment of factors including Cdc45 and the GINS complex to replication origins as well as 6 the activation of the MCM helicase. Melting of origin DNA resulting from MCM helicase 7 activity allows the formation of two replication forks (RF) that travel in opposite directions 8 along chromosomal DNA. Depending on the timing of their activation in S phase, eukaryotic 9 origins are classified as early, mid, or late. Such sequential activation of origins has been 10 shown to result at least in part from the recycling of limiting replication initiation factors from 11 early to mid, and then to late replicating genomic regions (2,3). Such temporal organization of 12 DNA replication is evolutionarily conserved among eukaryotes; however, the repertoire of 13 cellular factors and molecular mechanisms modulating origin activation remains incompletely 14 characterized. 15 RF progression can be halted upon encountering DNA lesions induced by any among 16 a multitude of environmental or endogenous genotoxins (4). This engenders a state of 17 replicative stress which can prevent completion of chromosomal duplication, thereby causing 18 genomic instability. Stalled RF activate Mec1 (ATR in humans), the apical kinase of the intra- 19 S phase checkpoint response in yeast (4). In turn, Mec1 promotes activation of the kinase 20 Rad53 via one of two pathways that depend upon either the RF component Mrc1 or the 21 adaptor protein Rad9 (5). Activated Mec1 and Rad53 phosphorylate a plethora of substrates 22 to i) promote the stability of stalled RF, and ii) to prevent further activation of replication 23 origins (6). In yeast, this latter effect has been shown to depend on Rad53-dependent 24 phosphorylation of the key replication factors Dbf4 and Sld3, which prevents activation of 25 MCM helicase complexes at origins that have not yet been fired (7). Intra-S phase 26 checkpoint-dependent inhibition of origin activity is important to prevent inordinate 27 accumulation and eventual collapse of stalled RF during periods of genotoxic stress (8). 28 Histone post-translational modifications are critical determinants of DNA replication 29 dynamics and origin activity (9,10). Among those modifications, histone lysine acetylation can 30 either promote and inhibit origin activity depending on the identity of the modified residue 31 and/or chromosomal context (11). The sirtuin family of histone deacetylases is well-conserved 1 throughout evolution, and several of its members have been shown to influence DNA 2 replication and repair (12). The yeast Saccharomyces cerevisiae possesses 5 sirtuins: the 3 founding member, Sir2, and Homologues of Sir Two 1 through 4 (Hst1-4) (12). Sir2 targets 4 histone H4 lysine 16 acetylation (H4 K16ac), which regulates origins at the rDNA locus and 5 telomeres (13, 14). Hst1, which can also target H4 K16ac, forms a complex with Sum1 and 6 Rfm1 and modulates the activity of a subset of origins genome-wide (15,16). While the 7 impact of Hst2 on DNA replication has not been directly assessed, at least some of the 8 functions of this sirtuin are known to be partially redundant with those of Sir2, as 9 overexpressed Hst2 rescues gene silencing defects caused by sir2∆ (17,18). 10 The only known histone substrate of the redundant sirtuins Hst3 and Hst4 is acetylated 11 H3 lysine 56 (H3 K56ac) (19). This modification is catalyzed by the acetyltransferase Rtt109 12 on newly synthesized histones H3 prior to their deposition onto nascent DNA during S phase 13 (20,21). After S phase, Hst3 and Hst4 remove H3 K56ac chromosome-wide such that a large 14 majority of nucleosomes do not harbor H3 K56ac at the start of the next cell cycle. While 15 under normal circumstances the bulk of H3 K56ac is removed by Hst3 during G2, Hst4 can 16 compensate for its absence. As such, the stoichiometry of H3 K56ac approaches 100% 17 throughout the cell cycle in hst3∆ hst4∆ double mutants (19,22). While constitutive H3 K56ac 18 has been shown to cause spontaneous DNA damage, thermosensitivity, and increased 19 sensitivity to genotoxins that cause replicative stress (19, 23, 24), the molecular mechanisms 20 underlying such striking phenotypes remain poorly understood. 21 Nicotinamide (NAM) is a non-competitive pan-inhibitor of sirtuins (25). Our group 22 previously performed genetic screens in S. cerevisiae with the goal of identifying genes 23 whose homozygous deletion (i.e. complete loss-of-function) confers either fitness defect or 24 advantage in response to NAM-induced sirtuin inhibition and consequent H3 K56 25 hyperacetylation (26). These screens revealed that several genes encoding regulators of the 26 DNA replication stress response promote resistance to NAM-induced elevation in H3 K56ac 27 caused by inhibition of Hst3 and Hst4 (26, 27). Previously published data also indicate that 28 cells lacking HST3 are defective in the maintenance of artificial chromosomes harboring a 29 reduced number of DNA replication origins (28), further linking H3 K56ac with the regulation 30 of DNA replication dynamics. Here, we present the results of a genome-wide screen aimed at 31 identifying genes whose haploinsufficiency modulates cell fitness in response to NAM. 32 Overall, we found that i) appropriate dosage of genes involved in various cellular pathways 1 influence cell fitness in response to NAM, ii) factors promoting DNA replication origin 2 activation are critical for survival in the absence of Hst3 and Hst4 activity, and ii) abnormal 3 persistence of the acetylation of new histones H3 on lysine 56 throughout the cell cycle 4 compromises the activity of replication origins. 5

6
A genetic screen to identify genes modulating cellular fitness in response to NAM 7 We performed a screen using the pooled yeast strains of the heterozygote diploid collection to 8 identify haploinsufficient genes that influence cell fitness upon NAM exposure (Table S1). 9 Using a Z-score cut-off of +/-2.58 (99% cumulative percentage), the screen identified 131 10 and 58 genes whose heterozygosity caused reduced or increased fitness, respectively, during 11 propagation for 20 generations in YPD medium containing 41 mM NAM ( Figure 1A). This list 12 of genes presents only modest overlap with that obtained from our previously published 13 screen using the homozygote deletion strain collection ( Figure 1B), suggesting that most of 14 the genes identified in the latter screen are not haploinsufficient with regards to NAM 15 sensitivity. We note that such limited overlap between screens performed on the homozygous 16 and heterozygous deletion collections has also been observed in several other chemogenetic 17 screens (29). Gene Ontology (GO) term analysis of genes whose heterozygosity sensitizes 18 cells to NAM revealed an obvious enrichment in DNA replication and DNA damage response 19 pathway, whereas terms reflecting proteasome-related and catabolic processes were 20 associated with mutations that enhanced fitness in NAM ( Figure 1C and Table S2-S3). 21 We next sought to validate individual heterozygous mutations representing the main 22 categories of "hits'' identified in the screen. WT diploid and heterozygote mutant strains of 23 interest (yfg1Δ::KanMX/YFG1) were mixed in a 1:1 ratio and incubated for 20 generations in 24 YPD +/-NAM. Appropriate dilutions of cells were then plated on YPD-agar +/-G418, and the 25 ratio of the number of heterozygous mutant (G418-resistant) vs WT (G418-sensitive) colonies 26 was calculated ( Figure 1D). These competition assays confirmed the expected impact of 27 heterozygous mutations causing diminished cell fitness in NAM-containing medium, thereby 28 validating our screen results. While significant improvement in growth was observed for 29 individual heterozygous mutants expected to promote fitness in NAM, we note that 30 heterozygous mutations causing improved fitness in response to NAM displayed generally 1 lower absolute Z-scores than those reducing fitness ( Figure 1A, Table S1). 2 Reduced activity DNA replication origins activity sensitizes cells to NAM 3 As mentioned previously, cells lacking Hst3 have previously been shown to present defects in 4 the maintenance of an artificial chromosome harboring reduced number of DNA replication 5 origins, revealing a potential link between this sirtuin and origin activity (28). Nevertheless, the 6 mechanistic basis explaining the effect of Hst3 on origins remains unknown. Interestingly, 6 of 7 the 11 essential DNA replication genes identified in the screen as promoting NAM resistance 8 are members of the pre-replicative complex (ORC3 and MCM4) or involved in various steps 9 of origin activation (DBF4, SLD2, SLD3, PSF2). We therefore decided to further investigate 10 the possible relationship between NAM sensitivity and origin activity. To this end, we focused 11 on DBF4 and its associated kinase Dbf4-dependent kinase (DDK), a complex formed by Dbf4 12 and the Cdc7 kinase. Even though CDC7 was not identified as being haploinsufficient with 13 regards to NAM resistance in our screen, haploid cells expressing hypomorphic temperature 14 sensitive alleles of DBF4 (dbf4-1) and CDC7 (cdc7-4) were found to be NAM-sensitive at 15 30°C, a semi-permissive temperature for these alleles (Figure 2A). Our data also indicate that 16 bob1-1 cdc7∆ cells, which harbor a mutation in MCM5 that bypasses DDK-dependent 17 phosphorylation of the MCM complex that is necessary for origin activation (30), are not 18 sensitive to NAM ( Figure 2B). This argues that the MCM complex is likely to be the relevant 19 target of DDK in this context, and suggests that impaired activation of replication origins 20 sensitizes cells to NAM. 21 Rap1-Interacting Factor 1 (Rif1) acts in concert with the phosphatase Glc7 to reverse 22 DDK-dependent MCM phosphorylation, thereby inhibiting origin activation (31-34). Moreover, 23 a previous screen performed by our group identified RIF1 among the few yeast genes whose 24 homozygous deletion improved cell fitness in NAM-containing medium (26). We found that 25 NAM-induced growth defects and accumulation of cells in S phase is rescued by deletion of 26 RIF1 ( Figure 2C, 2G). Moreover, N-terminal truncation or mutations in the Glc7-interacting 27 motif (rif1-RVxF/SILK) of Rif1, both of which were previously shown to partially suppress the 28 temperature sensitivity of cdc7-4 mutants by eliminating Rif1 binding to Glc7 (33, 35), also 29 suppressed the NAM sensitivity of cdc7-4 cells ( Figure 2C). Overall, these data indicate that 30 Rif1/Glc7-dependent dephosphorylation of MCM influences NAM sensitivity. 31 We and others previously showed that NAM treatment causes replicative stress and 1 DNA damage in yeast (19,26,27). Since lack of MCM phosphorylation by DDK causes 2 sensitivity to replicative stress-inducing drugs (36, 37), we tested the impact of RIF1 deletion 3 on NAM-induced DNA damage. Compared to WT, rif1∆ cells presented reduced NAM-4 induced histone H2A S129 phosphorylation and Rad52-YFP foci formation ( Figure 2D-E), two 5 well-known markers of replicative stress-induced DNA damage (38, 39). Importantly, lack of 6 Rif1 did not compromise the formation of ionizing radiation (IR)-induced Rad52 foci, which are 7 not primarily caused by replication-associated DNA lesions. We note that, in addition to its 8 role in regulating DNA replication, Rif1 is known to limit telomere length by inhibiting 9 telomerase activity (40). Moreover we previously showed that cells with short telomeres are 10 sensitive to NAM-induced sirtuin inhibition (27); we therefore considered the possibility that 11 abnormal telomere elongation in rif1∆ cells might favor NAM resistance. Contrary to this 12 notion, deletion of RIF1 suppressed NAM-induced growth and S phase progression defects in 13 telomerase-defective est2∆ cells ( Figure 2F-G), indicating that the role of Rif1 in modulating 14 NAM sensitivity is independent of its influence on telomerase activity. 15 Rif1 and Cdc7 influence the phenotypes of hst3Δ hst4Δ cells 16 We previously showed that NAM-induced DNA damage is attributable in large part to 17 inhibition of Hst3 and Hst4, leading to elevated H3 K56ac (26). We found that deletion of RIF1 18 rescued the temperature sensitivity of hst3∆ hst4∆ cells as well as the synthetic lethality of 19 hst3∆ hst4∆ sir2∆ without noticeably affecting H3 K56ac levels ( Figure 3A-B). We note that for 20 unknown reasons hst3∆ hst4∆ cells are temperature sensitive in S288C-derived genetic 21 backgrounds but not in W303 (our unpublished observations; e.g., compare Figure 3A Figure 3D), which might be due to the fact that Rif acts to stabilize stalled RF (41) in addition 3 to its role in regulating origin activity. We also found that rif1∆ reduced spontaneous formation 4 of Rad52 foci and histone H2A S129 phosphorylation in hst3∆ hst4∆ cells ( Figure 3E-F), 5 indicating that, in the absence of exogenous replicative stress-inducing genotoxins, Rif1 6 activity causes DNA damage in Hst3/Hst4-deficient cells. The above data, combined with 7 those linking Rif1 to NAM sensitivity, support the notion that Rif1/Glc7-mediated reversal of 8 DDK-dependent phosphorylation, and consequent inhibition of origins of DNA replication, 9 contributes to the phenotypes of cells lacking Hst3 and Hst4. Consistently, deletion of HST3 10 and HST4 considerably exacerbated the temperature sensitivity of cdc7-4 mutant cells in a 11 Rif1-dependent manner ( Figure 3G). Deletion of either HST3 or HST4 alone did not increase 12 the temperature sensitivity of cdc7-4 ( Figure 3H), in accord with the known functional 13 redundancy of Hst3/Hst4 with respect to histone deacetylation (19, 22). Altogether, these 14 results suggest that reduction in replication origin activity is detrimental to cells lacking 15 Hst3/Hst4. We sought to further explore DNA replication dynamics in cdc7-4 cells lacking Hst3/Hst4. To 19 this end, we synchronized cdc7-4 hst3∆ hst4∆ and appropriate control cells in G1 at the 20 permissive temperature of 25°C using alpha factor, followed by release toward S phase at the 21 semi-permissive temperature (for cdc7-4) of 25°C. Strikingly, we found that cdc7-4 hst3∆ 22 hst4∆ cells displayed strong inhibition of S phase progression when released from G1 toward 23 S at 30°C compared to either hst3Δ hst4Δ or cdc7-4 ( Figure 4A). Such S phase progression 24 defect was not observed at the permissive temperature of 25°C, indicating that the impact of 25 reduced Cdc7 activity (due to incubation at the semi-permissive temperature of 30°C for cdc7-26 4) on DNA replication is strongly exacerbated by deletion of HST3 and HST4 ( Figure 4B). As 27 expected, this phenotype was rescued by deletion of RIF1 ( Figure 4A) or expression of 28 plasmid-borne copy of HST3 ( Figure 4C). The observed DNA replication defect does not 29 appear to result from compromised release from alpha factor-mediated G1 arrest, since 30 asynchronously growing cdc7-4 hst3∆ hst4∆ cells were also found to accumulate in early S 31 when incubated at 30°C ( Figure 4D). Moreover, the budding index of cdc7-4 hst3∆ hst4∆ cells 32 released from alpha factor-mediated G1 block toward S phase at 30°C was comparable to 1 that of cdc7-4 cells (≈ 50-60% of cells with detectable buds; Figure 4E) even though the 2 former cells present barely detectable S phase progression in these conditions ( Figure 4A). 3 We note that the small size of buds at 45 and 60 minutes post-release from alpha factor 4 rendered precise assessment of budding index challenging. To further confirm our results, we 5 performed an experiment in which cdc7-4 and cdc7-4 hst3Δ hst4Δ cells were released from 6 alpha factor-induced G1 arrest toward S at the non-permissive temperature of 39°C for 3h, 7 thereby allowing time for buds to become larger while preventing Cdc7 activity and, 8 consequently, initiation of DNA synthesis at origins ( Figure 4F). After monitoring the budding 9 index, the temperature of the culture was decreased to 30°C for 30 minutes to evaluate S 10 phase progression ( Figure 4F). While for unknown reasons the fraction of cdc7-4 hst3Δ hst4Δ 11 and cdc7-4 mutants with detectable buds did not reach more than 60 to 80%, respectively, in 12 these conditions ( Figure 4G), S phase progression remained completely blocked in cdc7-4 13 hst3Δ hst4Δ, but not cdc7-4 cells, after incubation at 30°C ( Figure 4F). Overall, the data 14 indicate that even though cdc7-4 hst3∆ hst4∆ cells enter S phase at the semi-permissive 15 temperature of 30°C, DNA replication progression is strongly inhibited in these conditions. 16 Given the known role of Dbf4-Cdc7 in activating MCM helicase complexes at origins, 17 we suspected that cdc7-4 hst3Δ hst4Δ cells might present synthetic defects in the initiation of 18 DNA replication. Formation of RF at origins prevents the entry of yeast chromosomes in 19 Pulsed-Field Gel Electrophoresis (PFGE) gels (42). We found that PFGE signals, reflecting 20 entry of chromosomes in the gel, were significantly stronger in cdc7-4 hst3∆ hst4∆ cells at 45 21 and 60 minutes after release from alpha factor compared to WT, hst3Δ hst4Δ and cdc7-4 22 ( Figure 5A -B). This is consistent with the notion that a reduced proportion of cells activated 23 origins throughout chromosomes in cdc7-4 hst3∆ hst4∆ cells compared to control strains. We 24 next used alkaline gel electrophoresis and Southern blotting to detect formation of low 25 molecular weight nascent DNA at the efficient early origin ARS305, as described in (6). The 26 results indicate a strong reduction in the amount of low molecular weight DNA formed at this 27 origin within 60 minutes of release from alpha factor arrest at 30°C in cdc7-4 hst3∆ hst4∆ cells 28 compared to cdc7-4 mutants ( Figure 5C). 29 To further compare origin initiations in cdc7-4 hst3∆ hst4∆ vs cdc7-4 cells, we released 30 G1-arrested cells toward S phase at the non-permissive temperature of 37°C in the presence 31 of the nucleoside analog BrdU for 60 minutes, and then switched the temperature of the 32 cultures to 30°C for 30 minutes. BrdU-IP followed by quantitative PCR (qPCR) was then used 1 to quantify incorporation of BrdU in genomic DNA at three early origins (ARS305, ARS315 2 and ARS1211). This analysis revealed that BrdU incorporation into nascent DNA is 3 significantly reduced in cdc7-4 hst3∆ hst4∆ compared to cdc7-4 cells at these early/efficient 4 origins of replication ( Figure 5D-F). Consistently, qPCR analysis on total genomic DNA 5 showed that compared to cdc7-4 cells, duplication of DNA at these origins was inhibited in 6 cdc7-4 hst3∆ hst4∆ mutants 30 minutes after release from alpha factor arrest at 30°C (Figure   7 4G). We note that cdc7-4 hst3∆ hst4∆ cells eventually initiated DNA replication and 8 completed S phase when incubated for extended periods at 30°C (240 minutes post-release 9 from alpha factor arrest; Figure 4G-H). Taken together, the results indicate that the hst3∆ 10 hst4∆ mutations cause synthetic defects in the activation of replication origins when combined 11 with cdc7-4, thereby strongly delaying S phase progression at 30°C. 12

Inhibition of origin activity in cdc7-4 hst3∆ hst4∆ cells is not due to activation of Rad53
13 in early S phase 14 One of the key roles of intra-S phase checkpoint signaling is to limit the activation of origins in 15 response to replication stress (6,43). In yeast, this has been shown to occur via Rad53-  27 We next tried to delete RAD53 and SML1 in cdc7-4 hst3∆ hst4∆ cells to directly assess 28 the role of intra-S phase checkpoint signaling on the phenotypes of these mutants; SML1 29 deletion is necessary to permit viability of rad53Δ mutants (44). However, even though hst3∆ 30 hst4∆ rad53Δ sml1Δ cells are viable (23), we failed to generate a cdc7-4 hst3∆ hst4∆ rad53Δ 31 sml1Δ, suggesting that for unknown reason this combination of mutations causes synthetic 1 lethality. To circumvent this, we engineered cdc7-4 hst3∆ hst4∆ strains harboring mutations in 2 MRC1 and RAD9, two key mediators of the activation of the intra-S phase checkpoint (5). We 3 found that deletion of RAD9 in cdc7-4 hst3∆ hst4∆ cells led to modest improvement in S 4 phase progression, but only after 60 minutes of release from alpha factor arrest ( Figure 6C). 5 This suggests that Rad9 might contribute to the long-term maintenance, rather than the 6 establishment, of S phase progression defects in cdc7-4 hst3∆ hst4∆ cells. In contrast, 7 expression of a mutated allele of Mrc1 (mrc1-AQ) which compromises its role in activating 8 intra-S phase checkpoint kinases (45) did not have any influence on S phase progression in 9 cdc7-4 hst3∆ hst4∆ cells ( Figure 6D). We further found that while inhibition of the apical 10 kinase of the intra-S phase checkpoint Mec1 using caffeine (46, 47) completely abrogated 11 Rad53 phosphorylation, as expected, such treatment did not rescue the strong inhibition of 12 DNA replication progression of cdc7-4 hst3∆ hst4∆ mutants ( Figure 6E-F). We also note that 13 caffeine did not prevent cdc7-4 cells from completing DNA replication at 30°C ( Figure 6F). 14 Expression of Dbf4 and Sld3 variants that cannot be phosphorylated by Rad53 was 15 previously shown to abrogate intra-S phase checkpoint-dependent inhibition of origin activity 16 in yeast (7). We found that introducing such mutated alleles of DBF4 and SLD3 in cdc7-4 17 hst3∆ hst4∆ cells does not alleviate their S phase progression defects at 30°C; moreover, 18 mutations of DBF4 and SLD3 did not prevent cdc7-4 cells from completing S phase in these 19 conditions ( Figure 6G). We conclude that the incapacity of cdc7-4 hst3∆ hst4∆ mutants to 20 initiate DNA replication in a timely manner at the beginning of S phase when released from 21 G1 at 30°C is not due to Rad53-dependent phosphorylation of Sld3 and Dbf4 and consequent 22 inhibition of early origins of replication. 24 Constitutive acetylation of histone H3 lysine 56 (H3 K56ac) causes most of the phenotypes 25 associated with hst3∆ hst4∆ mutants, including their temperature and DNA damage sensitivity 26 (19, 22, 23). While H3 K56ac strictly depends on the Rtt109 histone acetyltransferase (48), 27 this acetyltransferase also acetylates other residues in the N-terminal tail of histone H3 (21,28 49, 50), although there are currently no evidence that link the acetylation of these residues 29 with the phenotypes of hst3∆ hst4∆ mutants. We found that deletion of RTT109 significantly 30 rescued DNA replication progression and growth of cdc7-4 hst3∆ hst4∆ cells at semi-31 permissive temperatures for cdc7-4 (between 28°C and 30°C; Figure 7A-C). Moreover, 32 replacing histone H3 by a H3 K56Q variant to mimic constitutive H3 K56ac in cdc7-4 cells 1 caused strong synthetic temperature sensitivity ( Figure 7D). 2 We next sought to engineer a cdc7-4 hst3∆ hst4∆ strain lacking H3 K56ac via 3 expression of a histone H3 variant in which lysine 56 is replaced by a non-acetylable arginine 4 residue (H3 K56R). To this end, both copies of the endogenous genes encoding histone H3 5 (HHT1 and HHT2) were deleted while one copy of the HHT1 gene +/-K56R mutation was 6 integrated at the TRP1 locus. We failed to generate the cdc7-4 hst3∆ hst4∆ cells expressing 7 either H3 WT or H3 K56R using this standard strategy, suggesting that abnormal histone 8 gene dosage due to deletion of HHT2 may be lethal in this context. To circumvent this issue 9 and reduce H3 K56ac levels without changing histone gene dosage, we replaced HHT1 by assembly factors CAF1 and Rtt106 (55). While this "pre-deposition" function of H3 K56ac is 2 well-established, several observations suggest that this mark also plays important biological 3 roles following its incorporation into chromatin (56). Constitutive nucleosomal H3 K56ac 4 causes spontaneous DNA damage and extreme sensitivity to replication-blocking drugs in 5 hst3∆ hst4∆ cells, suggesting that chromatin-associated H3 K56ac influences the cellular 6 response to replicative stress (22, 23). Moreover, cells have evolved molecular mechanisms 7 to degrade Hst3 in response to replicative stress (57, 58), which raises the possibility that the 8 ensuing persistence of H3 K56ac may somehow contribute to the DNA damage response. 9 Nevertheless, while a multitude of cellular pathways have been associated with nucleosomal 10 H3 K56ac (23, 24, 26, 27, 51), the molecular basis of the sensitivity of hst3∆ hst4∆ mutants to 11 replicative stress, as well as the role of H3 K56ac persistence after DNA damage, are poorly 12 understood. 13 We previously showed that NAM-induced inhibition of Hst3 and Hst4 causes replicative 14 stress by elevating H3 K56ac (26, 27). In accord with this, our current and previously 15 published screens (26) revealed that several genes conferring NAM resistance participate in 16 DNA replication and repair ( Figure 1C-D). Interestingly, the current screen also revealed that 17 lack of TAF5 and TAF12, which encodes proteins shared by TFIID and the SAGA 18 acetyltransferase complex, strongly sensitizes cells to NAM. Since the SAGA complex 19 modulates the expression of stress-responsive genes (59), we speculate that the presence of 20 subunits of this complex among the "hits" of our screen might reflect transcriptional activation 21 of critical cellular stress responses pathways during NAM treatment. Our screen also 22 identified genes involved in proteasome regulation and ubiquitin-dependent processes as 23 modulators of NAM sensitivity ( Figure 1C). As mentioned previously, the Rtt101-Mms1-24 Mms22 ubiquitin ligase complex displays clear genetic links with H3 K56ac in the context of 25 the response to replicative stress (51-54). It is therefore possible that heterozygosity in genes 26 involved in ubiquitin-and proteasome-related processes elevate cell fitness upon NAM-27 induced inhibition of Hst3 and Hst4 by influencing Rtt101-Mms1-Mms22-related processes. 28 Published reports indicate that hst3Δ cells display H3 K56ac-dependent defects in the 29 maintenance of a chromosome harboring a reduced number of replication origins (28, 60), 30 which suggests that elevated H3 K56ac might negatively influence the completion of 31 chromosomal DNA replication in situations where the number of active origins is limited. In 32 accord with this, our data indicate that i) cells harboring hypomorphic alleles of the critical 1 origin activation genes CDC7 and DBF4 display strong growth defects in the presence of 2 NAM, and ii) firing of early/efficient origins of DNA replication is compromised in cdc7-4 hst3∆ 3 hst4∆ cells released from G1 toward S at the semi-permissive temperature for cdc7-4. 4 Deletion of RIF1 was found to alleviate these phenotypes, in agreement with the known role 5 of Rif1 in promoting Glc7-dependent dephosphorylation of MCM complexes leading to 6 inhibition of origin activation (33), and with the fact that homozygous deletion of RIF1 was 7 found to improve cell fitness in response to NAM in our previously published screen (26). 8 Importantly, data presented here also show that lack of Rif1 suppresses several phenotypes 9 of hst3∆ hst4∆ cells. While these results can be considered surprising in light of the fact that 10 Rif1 has also been reported to promote the stability of stalled RF (41), it is possible that the 11 elevation of origin activity caused by rif1∆ overrides the negative impact on replicative stress 12 responses caused by this mutation in hst3∆ hst4∆ cells. 13 RF stalling activates Rad53, which then phosphorylates the replication proteins Dbf4 14 and Sld3 to inhibit the firing of replication origins that have not yet been activated (7). In 15 apparent contrast to our results, a previously published report revealed that hst3∆ hst4∆ cells 16 present elevated activation of late origins in cells released from G1 toward S phase in 17 medium containing the replication-blocking ribonucleotide reductase inhibitor HU (11). 18 However, such effect was not specific to hst3∆ hst4∆ mutants; indeed, this was found to be 19 an indirect consequence of elevated spontaneous DNA damage and constitutive Rad53 20 activity in various replicative stress response mutants, leading to Rad53-dependent elevation 21 of dNTP pools and consequent HU-resistant DNA synthesis (61 Rad53 do not improve DNA replication in cdc7-4 hst3∆ hst4∆ cells. Taken together, the above 30 data argue that the impact of constitutive H3 K56ac on origin activity in early S phase does 31 not require prior RF stalling and ensuing elevation in Rad53 activity. Consistently, we also 32 showed that the activation of several early/efficient origins is strongly delayed in cdc7-4 hst3∆ 1 hst4∆ after release from G1 at the semi permissive temperature for cdc7-4, which presumably 2 explains why HU exposure does not cause Rad53 activation in these conditions. 3 Our data and those of others (28, 60) argue that elevated H3 K56ac strongly inhibits 4 DNA replication only under conditions of reduced DDK activity or in situations where the 5 number of active origins is limited. Such conditions are met in cdc7-4 or dbf4-1 mutants at the 6 semi-permissive temperature, or in cells harboring an artificial chromosome engineered to 7 have a low number of origin. We emphasize that as mentioned earlier RF stalling leading to 8 Rad53 activation and consequent phosphorylation of Dbf4 also diminishes DDK activity (7). It 9 is therefore possible that genotoxin-induced RF stalling and consequent Rad53 activation 10 might synergize with constitutive H3 K56ac in inhibiting late origin activity in hst3∆ hst4∆ cells 11 that harbor a wild-type allele of CDC7. In turn, this would be expected to compromise the  (12, 19, 23, 24). 25 The mechanism linking H3 K56ac to origin activity is currently unclear. As mentioned 26 previously, Mec1 activation promotes the degradation of Hst3, which causes inordinate 27 persistence of H3 K56ac in chromatin in cells exposed to genotoxic drugs (57, 58). Such The authors declare no competing interests exist.

28
Yeast strains and growth conditions. Yeast strains used in this study are listed in Table 1 29 and were generated and propagated using standard yeast genetics methods. Yeast strains 30 used in Tables S1-S2-S3 were taken from the heterozygote yeast deletion collection 1 (ThermoFisher). For nicotinamide (NAM) treatments, asynchronously growing cells were 2 centrifuged and resuspended at 0.01-0.1 OD/mL in YPD or synthetic (SC) medium containing 3 20 mM NAM (Sigma-Aldrich). Cells were incubated on a shaker for indicated time. Cells 4 synchronization in G1 was performed by incubating MATa yeasts in medium containing 2 5 µg/mL alpha-factor for 90 minutes followed by the addition of a second dose of 2 µg/mL of 6 alpha-factor for another 75 minutes. Cells were then washed once in YPD or SC medium and 7 released in S phase in medium supplemented with 5 µg/mL pronase (Protease from 8 Streptomyces griseus, Sigma-Aldrich). For ionizing irradiation, exponentially growing cells 9 were exposed to 40 Gy followed by a 60-minute incubation at 30°C prior to sample collection. 10 For methyl methane sulfonate (MMS) treatment, cells were first synchronized in G1 using 11 alpha factor, then incubated in YPD containing 0.01% MMS (Sigma -Aldrich) and 5 µg/mL 12 pronase at a density of 1 OD 630 /mL for 60 minutes. After treatment, cells were washed twice 13 with YPD containing 2.5% sodium thiosulfate (Bioshop), followed by incubation in YPD.
14 Caffeine (Sigma-Aldrich) was used at a concentration of 0.15%. OD 630 were taken and cells were diluted appropriately to prevent saturation of the culture. 17 After 20 generations, 0.01 OD 630 of cells was spread on YPD-agar +/-G418 plates. Plates  Immunoblotting. 4 OD of cells were pelleted and frozen at -80°C prior whole-cell extraction. 25 Cells were extracted using 0.1M NaOH for 5 minutes at room temperature as described 26 before (72) or using standard tri-chloroacetic acid (TCA) and glass beads method (73). 27 Protein extracts were quantified using bicinchoninic acid (BCA) protein assay kit according to 28 the manufacturer's protocol (Pierce). SDS-PAGE and transfer were performed using standard 29 methods. Anti-H3 (Abcam; cat: ab1791) and anti-Rad53 (Abcam; cat: ab104232) were purchased from Abcam. Anti-H3 K56ac (AV105) and anti-H2A-S129-P (AV137) antibodies 1 were generously provided by Dr. Alain Verreault (Université de Montréal, Canada). Goat anti-2 rabbit (BioRad; cat: 1705046), goat anti-mouse (Bio Rad; cat: 1705047) and goat anti-rat 3 (Abcam; cat: ab97057) were used as secondary antibodies. Protein visualization was realized 4 by chemiluminescence using Pierce ECL Western Blotting Substrate. Images were captured 5 using an Azure c600 chemiluminescence Imaging System. 6 Fluorescence microscopy. Cells were fixed in 0.1M of potassium phosphate buffer pH 6.4 7 containing 3.7% formaldehyde (Sigma-Aldrich) and slides were prepared as described (52). 8 Images were taken by fluorescence microscopy using a 60X objective (numerical aperture 9 [NA], 1.42) on DeltaVision instrument (GE Healthcare). Images analysis was performed using 10 SoftWoRx 7 software and FIJI 1.53.  Table   23 1. BrdU incorporation quantification were performed using the standard percent of the input  Table 1. Briefly, qPCR signal for a given 31 origin was first normalized to the signal obtained from the NegV locus (ChrV: 532538-532516) 32 (62). This region is located ≈ 12 Kb from ARS521, an origin which has not been detected to be 1 active in several studies according to OriDB (http://cerevisiae.oridb.org/) and ≈ 18 kb from 2 ARS522, a subtelomeric origin of replication activated in late S. As such, the NegV locus is 3 expected to be replicated in late S, and therefore to generally remain unreplicated in a 4 majority of cdc7-4 and cdc7-4 hst3∆ hst4∆ cells 30 minutes post-release from G1 arrest 5 toward S phase. The NegV-normalized S phase signal was divided by the NegV-normalized 6 signal obtained from alpha factor arrested (G1) cells. Complete replication of an origin is 7 therefore expected to result in a ratio of S phase over G1 signal of 2. 8