Replisome Function During Replicative Stress Is Modulated by Histone H3 Lysine 56 Acetylation Through Ctf4

Abstract

Histone H3 lysine 56 acetylation in Saccharomyces cerevisiae is required for the maintenance of genome stability under normal conditions and upon DNA replication stress. Here we show that in the absence of H3 lysine 56 acetylation replisome components become deleterious when replication forks collapse at natural replication block sites. This lethality is not a direct consequence of chromatin assembly defects during replication fork progression. Rather, our genetic analyses suggest that in the presence of replicative stress H3 lysine 56 acetylation uncouples the Cdc45–Mcm2-7–GINS DNA helicase complex and DNA polymerases through the replisome component Ctf4. In addition, we discovered that the N-terminal domain of Ctf4, necessary for the interaction of Ctf4 with Mms22, an adaptor protein of the Rtt101-Mms1 E3 ubiquitin ligase, is required for the function of the H3 lysine 56 acetylation pathway, suggesting that replicative stress promotes the interaction between Ctf4 and Mms22. Taken together, our results indicate that Ctf4 is an essential member of the H3 lysine 56 acetylation pathway and provide novel mechanistic insights into understanding the role of H3 lysine 56 acetylation in maintaining genome stability upon replication stress.

THE eukaryotic replisome consists of polymerases and an essential DNA helicase that are linked by a number of factors assembled during the initiation of chromosome replication. Progression of the replication fork depends on the activity of the replisome progression complex (RPC). This complex is uniquely present during S phase (Gambus et al. 2006) and remains associated with the replication fork until completion of DNA replication. In Saccharomyces cerevisiae, the RPC is made up of Mcm10, Mrc1, Tof1, Csm3, Ctf4, Top1, FACT (Spt16 and Pob3), and the CMG complex comprising Cdc45, Mcm2-7 (MCM), and the go ichi ni san (GINS) complex. The CMG constitutes the core replicative helicase responsible for the movement and activities of the replication fork (Pacek et al. 2006; Bochman and Schwacha 2009).

The link between helicase and polymerases is a crucial determinant for the regulation of the replisome. The leading-strand DNA polymerase-ɛ was recently shown to be integrated into the replisome via an interaction with the GINS complex (Sengupta et al. 2013). Furthermore, the DNA polymerase–α-primase complex, which initiates DNA synthesis at replication origins and continues to prime Okazaki fragments at the fork, remains associated with the RPC via the Ctf4 trimer, which simultaneously interacts with the GINS complex (Gambus et al. 2009; Tanaka et al. 2009; Gosnell and Christensen 2011; Simon et al. 2014).

Cells have evolved different mechanisms to maintain genome integrity under the conditions threatening replication progression (Jossen and Bermejo 2013; Leman and Noguchi 2013). The S-phase checkpoint mediated by MRC1 was initially characterized as a pathway activated by fork stalling and able to both stabilize the replisome and delay cell cycle progression (Elledge 1996; Sancar et al. 2004; Labib and De Piccoli 2011).

It was further shown that, during DNA replication stress, lack of either MRC1 or CTF4 leads to uncoupling between the replicative polymerases and RPC, as well as a dissociation of replisome components (Bando et al. 2009; Tanaka et al. 2009; Mimura et al. 2010). Unlike MRC1, CTF4 is not required for S-phase checkpoint activation.

Ctf4 was initially identified in S. cerevisiae as a chromosome transmission fidelity factor required for the maintenance of genome stability and sister-chromatid cohesion (Spencer et al. 1990; Jawad and Paoli 2002; Gambus et al. 2006; Lengronne et al. 2006). CTF4 is not essential for budding yeast viability (Miles and Formosa 1992), but its deletion greatly sensitizes cells to DNA replication drugs (Ogiwara et al. 2007). Mechanistically, Ctf4 is required for coordination between DNA unwinding and synthesis, and it also stabilizes polymerase-α at the replication forks (Gambus et al. 2009; Tanaka et al. 2009; Mimura et al. 2010). Among various partners, Ctf4 interacts with an F-box protein Dia2 involved in the regulation of DNA replication (Mimura et al. 2009) and with Mms22, an adaptor protein of the Cul4(Ddb1)-like E3 ubiquitin ligase complex (Gambus et al. 2009; Mimura et al. 2009, 2010). The latter also includes Mms1 and cullin Rtt101, both crucial for maintaining replisome integrity in hydroxyurea and therefore for efficient recovery from replication stress (Luke et al. 2006; Duro et al. 2008; Zaidi et al. 2008; Gambus et al. 2009; Mimura et al. 2010; Vaisica et al. 2011).

The Rrm3 helicase travels with the replication fork and facilitates the progression of replication forks through nonhistone protein–DNA complexes throughout the genome (Azvolinsky et al. 2009; Fachinetti et al. 2010). In the absence of RRM3, chromosome breaks occur at discrete fork pause sites at specific genomic locations (Ivessa et al. 2003). A number of studies indicate that the DNA breaks generated in rrm3∆ cells affect cell viability in the absence of the so-called “H3K56 acetylation pathway” that comprises ASF1, RTT109, RTT101, MMS1, and MMS22 (Tong et al. 2004; Luke et al. 2006; Pan et al. 2006; Collins et al. 2007; Duro et al. 2008; Roberts et al. 2008; Zaidi et al. 2008; Costanzo et al. 2010; Koh et al. 2010; Mimura et al. 2010).

In S. cerevisiae, H3K56 localizes at the DNA entry and exit points of a nucleosome (Masumoto et al. 2005; Ozdemir et al. 2005; Xu et al. 2005). H3K56 is transiently acetylated during the S phase of the cell cycle and after DNA damage and is rapidly de-acetylated by the action of the sirtuins Hst3 and Hst4, when cells enter the transition between G2 and M phases and after DNA repair (Masumoto et al. 2005; Xu et al. 2005). Asf1 binds to all newly synthesized H3 and presents the H3-H4 dimer to the Rtt109 lysine acetyltransferase for H3K56 acetylation (H3K56ac). Following acetylation of H3K56, ubiquitylation of H3 and H4 by the Rtt101–Mms1–Mms22 E3 ligase complex weakens the Asf1–H3–H4 interaction (Han et al. 2013) and facilitates the transfer of H3-H4 to other histone chaperones complexes, thereby coordinating nucleosome formation as well as stable progression of the replication fork (Li et al. 2008; Clemente-Ruiz et al. 2011; Han et al. 2013). In addition to its well-characterized function in replication-coupled chromatin assembly, H3K56ac is also required for a number of other processes such as transcription, DNA repair-coupled chromatin assembly, deactivation of the DNA damage checkpoint, and repair of DNA lesions that occur during DNA replication (Chen et al. 2008; Endo et al. 2010; Wurtele et al. 2011; Tanaka et al. 2012; Haber et al. 2013; Muñoz-Galván et al. 2013). Yet, despite its multiple roles, the mechanism by which the H3K56ac pathway sustains viability under replication stress and its targets remain unknown.

Here, we show that the H3K56ac pathway is essential for the viability of cells lacking RRM3. Strikingly, we discovered that, in cells devoid of RRM3, CTF4 mediates a deleterious effect in the absence of H3K56ac. This finding poises Ctf4 as a potential target of the H3K56ac pathway. Genetic analysis of the negative effect of CTF4 revealed that it is related to the interaction of Ctf4 with the GINS complex and DNA polymerase-α. Consistently, we found that destabilization of the catalytic subunit of polymerase-α rescues the viability of rrm3∆ cells in the absence of H3K56ac. Finally, our data strongly suggest that this effect is dependent upon an interaction between Ctf4 and Mms22. Similarly to ctf4∆, deletion of MRC1 induces uncoupling between helicase and polymerase (Tanaka et al. 2009; Mimura et al. 2010; Vaisica et al. 2011). In accord with this notion, we found that the replication function of MRC1 is also strongly deleterious for cells experiencing constitutive replicative damages in the absence of a functional H3K56ac pathway.

Materials and Methods

Strain construction

All strains used in this study are listed in Supporting Information, Table S1. Null mutations were obtained after polymerase chain reaction amplification of a disruption cassette as described previously (Corda et al. 2005).

Cell cycle analysis

For synchronous cell cultures, yeast cells were grown at 25° or 30° in yeast extract peptone dextrose (YPD) to OD₆₀₀ = 0.6 and then arrested in G1 by the addition of 15 µg/ml of α-factor (GENEPEP SA). After 2 hr, α-factor was removed to allow cells to progress synchronously through the cell cycle either in the presence or absence of 40 µM camptothecin (CPT). Samples were taken every 10 min for fluorescence-activated cell sorting (FACS), Ctf4-Myc chromatin-binding assay, and H3K56ac measurement.

Protein chromatin-binding assay

The assays were performed as described previously (Liang and Stillman 1997). Briefly, cells were harvested and treated with sodium azide at the indicated time and spheroplasted by incubating them first in 3 ml of prespheroplasting buffer [100 mM PIPES Piperazine-1,4-bis(2-ethanesulfonic acid) (pH 9.4), 10 mM dithiothreitol (DTT)] for 10 min at room temperature and then in 2 ml of spheroplasting buffer [50 mM KH₂PO₄/K₂HPO₄ (pH 7.5), 0.6 M sorbitol, 10 mM DTT] containing 80 µl of 1 mg/ml of oxalyticase (Sigma) at 30° for 25 min with occasional shaking. Spheroplasts were washed with 1 ml of cold wash buffer [100 mM KCl, 50 mM HEPES–KOH (pH 7.5), 2.5 mM MgCl₂, and 0.4 M sorbitol], pelleted at 4000 × g for 1 min at 4°, and resuspended in 100 µl of extraction buffer (EB) [100 mM KCl, 50 mM HEPES–KOH (pH 7.5), 2.5 mM MgCl₂, 50 mM NaF, 5 mM Na₄P₂O₇, 0.1 mM NaVO₃] containing protease inhibitors [1 mM phenylmethylsulfonyl fluoride (PMSF), 20 µg/ml of leupeptin, 2 µg/ml of pepstatin, 2 mM benzamidine HCl, and 0.2 mg/ml of bacitracin]. The suspension was split into two tubes. Spheroplasts were lyzed by adding Triton X-100–0.25% and incubating on ice for 5 min with gentle mixing. The lysate was underlayered with 50% vol of 30% sucrose and spun at 12,000 × g for 10 min at 4°. Supernatants correspond to the soluble fractions. Pellets were washed with 25% EB containing 0.25% Triton X-100, spun again at 10,000 × g for 5 min at 4°, and resuspended in Laemmli buffer. Proteins were transferred onto a nitrocellulose membrane. Ctf4-Myc was detected with the 9E10 anti-Myc monoclonal antibody (Santa Cruz). We detected H3K56ac by using H3K56ac antibody (Active Motif).

Mass spectrometry

Cell culture and cross-link:

Cells were synchronyzed as described above. At time 0 min (G1 phase) and 20 min (S phase), cells were treated with 1% formaldehyde (Sigma-Aldrich) at room temperature for 15 min. Formaldehyde was quenched at room temperature with 0.125 M glycine for 5 min. Finally, cells were washed three times with 1× cold TBS, frozen in liquid nitrogen, and stored at −80°. Cell cycle progression was monitored by FACS.

Isolation of Ctf4-green fluorescent protein for spectrometry analysis from yeast cell extracts:

Purification of the Ctf4-green fluorescent protein (Ctf4-GFP) was performed as previously described (De Piccoli et al. 2012) with minor modifications. Pellets were resuspended in lysis buffer (mass/vol) containing 50 mM HEPES (pH 7.9), 140 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA) supplemented with a protease inhibitor cocktail (Roche) and 1 mM PMSF, 0.1 mM NaVO₃, 2 mM NaF, and 1 mM DTT. Cells were disrupted using a Fast Prep machine (MP Biomedicals, three runs of 30 sec at maximum speed). After addition of 0.25 vol of extraction buffer (50 mM HEPES pH 7.9, 140 mM NaCl, 1 mM EDTA, 5% Triton-X100, 0.5% sodium deoxycholate), the samples were sonicated in a bioruptor (Diagenode), and insoluble material was removed by centrifugation at 16,000 × g for 30 min. The cell extracts were then incubated with anti-GFP-coated magnetic beads (Meek et al. 2012) for 2 hr before washing three times with each buffer 1 [50 mM HEPES (pH 7.9), 500 mM NaCl, 1 mM EDTA, 1% Triton-X100, 0.1% Na deoxycholate] and buffer 2 [10 mM Tris–HCl (pH 8), 250 mM NaCl, 0.5% NP-40, 0.5% sodium deoxycholate, 1 mM EDTA]. The immunoprecipitated protein samples were incubated for 10 min at 65° with elution buffer [50 mM Tris–HCl (pH 8), 10 mM EDTA, 1% SDS] before addition of Laemmli buffer and incubation at 95° for 30 min.

Sample preparation:

Pulldown protein extracts were loaded on NuPAGE 4–12% Bis–Tris acrylamide gels in MOPS buffer according to the manufacturer’s instructions (Invitrogen, Life Technologies). Electrophoresis was stopped as soon as proteins were stacked as a single band. Protein-containing bands were stained with Thermo Scientific Imperial Blue, cut from the gel, and, following reduction and iodoacetamide alkylation, digested with high-sequencing grade trypsin (Promega, Madison, WI). Extracted peptides were concentrated before mass spectrometry (MS) analysis. MS analyses were carried out by liquid chromatography–tandem mass spectrometry using a LTQ-Velos-Orbitrap (Thermo Fisher Scientific) connected to a nanoLC Ultimate 3000 rapid separation liquid chromatography system (Dionex, Sunnyvale, CA). Five microliters corresponding to 20% of each sample was injected in triplicate into the system. After preconcentration and washing of the sample on a Dionex Acclaim PepMap 100 column (C18, 2 cm × 100 µm i.d. 100-Å pore size, 5-µm particle size), peptides were separated on a Dionex Acclaim PepMap RSLC column (C18, 15 cm × 75 µm i.d., 100-Å, 2-µm particle size) at a flow rate of 300 nl/min with a two-step linear gradient (4–20% acetonitrile/H20; 0.1% formic acid for 90 min and 20–45% acetonitrile/H₂O; 0.1% formic acid for 30 min). For peptide ionization in the nanospray source, spray voltage was set at 1.4 kV and the capillary temperature at 275°. All samples were measured in a data-dependent acquisition mode. Each analysis was preceded by a blank MS run to monitor system background. The peptide masses were measured in a survey full scan [scan range 300–1700 m/z, with 30-K full width at half maximum resolution at m/z = 400, target automatic gain control value of 1.00 × 106 and maximum injection time of 500 ms]. In parallel to the high-resolution full scan in the Orbitrap, the data-dependent collision-induced dissociation scans of the 10 most intense precursor ions were fragmented and measured in the linear ion trap (normalized collision energy of 35%, activation time of 10 ms, target automatic gain control value of 1.00 × 104, maximum injection time of 100 ms, isolation window of 2 Da). Parent masses obtained in the Orbitrap analyzer were automatically calibrated on the 445.120025 ion used as lock mass. The fragment ion masses were measured in the linear ion trap to have a maximum sensitivity and the maximum amount of MS/MS data. Dynamic exclusion was implemented with a repeat count of 1 and exclusion duration of 30 sec.

Data analysis:

Raw files generated from mass spectrometry analysis were processed with Proteome Discoverer 3.1 (Thermo Fisher Scientific). This software was used to search data via an in-house Mascot server (version 2.4.1; Matrix Science Inc., London) against the S. cerevisiae database subset (7802 sequences) of the SwissProt database (version 2014-06). For the database search, the following settings were used: a maximum of two miscleavages, oxidation as a variable modification of methionine, carbamido-methylation as a fixed modification of cysteine, and trypsin as the enzyme. A peptide mass tolerance of 6 ppm and a fragment mass tolerance of 0.8 Da were allowed. Only peptides with a high-stringency Mascot score were used for protein identification. A peptide false discovery rate of <1% was used. The precursor ion area detector from Proteome Discoverer 3.1 was used for relative quantitation of proteins (sum of the area of the three more intense precursor ions used for protein identification). Protein lists obtained from cells with or without Ctf4-GFP were compared to discriminate between those proteins that specifically interact with Ctf4 from contaminant proteins.

Results

Loss of chaperone function of Asf1 causes rrm3∆ lethality

As previously suggested by genetic screens (Pan et al. 2006; Collins et al. 2007; Fiedler et al. 2009; Costanzo et al. 2010), we observed that RRM3 disruption strongly affects the growth of asf1∆ cells (see below). To characterize the genetic interaction between ASF1 and RRM3, one allele of ASF1 was deleted in an RRM3/rrm3∆ diploid. Dissection of meiotic tetrads shows that asf1∆ rrm3∆ spores do not form visible colonies after 3 days at 30° (Figure 1A, left). This synthetic lethality between asf1∆ and rrm3∆ cannot be attributed to spore germination defects because asf1∆ rrm3∆ spores formed visible microcolonies after 5 days at 30° (Figure 1A, right). Microscopic analysis indicated that asf1∆ rrm3∆ cells displayed aberrant morphologies, and that for almost all cells, the lethality can be attributed to death during mitosis (Figure 1B).

Figure 1.

Asf1 interaction with histone H3 is crucial for viability of yeast rrm3∆ cells. (A) Deletion of ASF1 is lethal in rrm3∆ cells. Tetrad dissection of the diploid strain asf1∆/ASF1 rrm3∆/RRM3. In this and subsequent figures, the four spores from a given tetrad are in a vertical line on a YPD plate. Four representative tetrads are shown after 3 days (left) and after 5 days (right) at 30°. Squares indicate the rrm3∆ single mutants. Circles indicate the asf1∆ single mutants. Assuming that 2:2 segregation of the marker allows one to identify asf1∆ rrm3∆ double mutants (indicated by dashed circles). (B) Wild-type and asf1∆ rrm3∆ cells (from microcolonies shown in A, right) analyzed by differential interference contrast and 4′,6-diamidino-2-phenylindole staining. Almost all asf1∆ rrm3∆ cells analyzed have no distinct nucleus compared to wild-type cells. (C) Effects of Asf1 interactions on the viability of rrm3∆ cells. Tetrad analysis of the meiotic progeny of asf1∆/ASF1 rrm3∆/RRM3 diploid cells expressing asf1-D37R E39R (i) or asf1-V94R (ii) mutated forms of ASF1 from plasmid pRS314. The presence of asf1∆ rrm3∆ spores is indicated by dashed circles. asf1∆ rrm3∆ spores expressing asf1-D37R E39R (i) or asf1-V94R (ii) are indicated by circles. A plus sign (+) indicates spores carrying the plasmid. wt, A1, R3, and A1R3 indicate wild-type, asf1∆, rrm3∆, and asf1∆ rrm3∆ spores, respectively. (iii) Tetrad analysis of the meiotic progeny of rad53-ALRR/RAD53 rrm3∆/RRM3 diploid cells. Circles indicate the rad53-ALRR mutant. Dashed circles indicate the rad53-ALRR rrm3∆ double mutant. wt, R3, ALRR, and ALRR R3 indicate wild-type, rrm3∆, rad53-ALRR, and rad53-ALRR rrm3∆ spores, respectively.

The histone chaperone Asf1 fulfils various chromatin-related functions in both replication-coupled and replication-independent fashion and through physical interactions with multiple partners (Mousson et al. 2007). Asf1 participates in the assembly and disassembly of chromatin through its H3 histone chaperone activity, in transcriptional silencing via an interaction with Hir1 and in several aspects of the cellular response to genotoxic stress through an interaction with H3 or Rad53 (Hu 2001; Sharp et al. 2001; Sutton et al. 2001; Celic et al. 2006; Recht et al. 2006; Takahata et al. 2009; Jiao et al. 2012; Burgess et al. 2014).

To understand which function of Asf1 is crucial for viability of rrm3∆ cells, we tested various separation-of-function mutants of Asf1. Plasmids expressing asf1 mutant proteins were introduced individually into an asf1∆/ASF1rrm3∆/RRM3 diploid strain. After sporulation of diploids, 100 tetrads were dissected and the genotypes of the viable spores were determined. To assess the replication-independent chromatin assembly function of Asf1, we first analyzed the double mutation D37R E39R (asf1-DE) that abolishes Asf1–Hir1 interaction but not histone H3 binding (Mousson et al. 2005). The HIR complex (formed by Hir1, Hir2, Hir3, and Hpc2) binds to Asf1 and promotes replication-independent chromatin assembly (Green et al. 2005). We found that the asf1-DE mutant protein expressed in the asf1∆ rrm3∆ cells complemented the lethality of the double mutant (Figure 1C, i). In the complementary experiment we observed that the HIR2 deletion did not affect rrm3∆ viability (Figure S1). Given that HIR2 is required for the integrity of the HIR complex and, consequently, for histone deposition activity of Asf1/HIR (Green et al. 2005; Silva et al. 2012), these observations demonstrate that the replication-independent chromatin assembly function of Asf1 is not required for rrm3∆ cell viability. We next evaluated the contribution of the single-residue substitution V94R, important for histone H3 interaction (Mousson et al. 2005; Jiao et al. 2012). We found that expression of the asf1-V94R allele was not able to complement asf1∆ rrm3∆ lethality (Figure 1C, ii). The checkpoint kinase Mec1 and the Ddc1–Mec3–Rad17 sliding clamp regulate the interaction between Asf1 and Rad53 (Hu 2001; Sharp et al. 2001; Burgess et al. 2014). Since the asf1-V94R mutation also affects Asf1–Rad53 interaction (Jiao et al. 2012; Dennehey et al. 2013), we tested the viability of the Rad53-A806R-L808R (rad53-ALRR) mutant, which is strongly affected in its interaction with Asf1 (Jiao et al. 2012) in the rrm3∆ background. We found that the rad53-ALRR rrm3∆ double mutant is viable (Figure 1C, iii). These data suggest that the crucial function of ASF1 required for rrm3∆ cells viability is related to its histone chaperone function.

H3K56 acetylation is necessary for viability of rrm3∆ cells

Cells expressing the asf1-V94R mutant that cannot bind to histone H3-H4 (Mousson et al. 2005) lose H3K56ac (Recht et al. 2006), an important mark of all newly synthesized histone H3’s preceding the histone fold domain (Masumoto et al. 2005; Celic et al. 2006; Han et al. 2007b). H3K56ac is catalyzed by the histone acetyl transferase (HAT) Rtt109 upon presentation of the H3-H4 heterodimer by the histone chaperone Asf1 (Schneider et al. 2006; Driscoll et al. 2007; Han et al. 2007a; Tsubota et al. 2007; Dahlin et al. 2014). To ascertain the contribution of H3K56ac to rrm3∆ viability, we analyzed the consequences of deleting RTT109 in rrm3∆ cells. We found that similarly to ASF1, RTT109 is required in the absence of RRM3 (Figure S2). Because RTT109 also acetylates other H3 and H4 lysines (Fillingham et al. 2008; Abshiru et al. 2013) and has functions independent of its H3K56 HAT activity (Roberts et al. 2008), we crossed a strain expressing H3K56R from a centromeric plasmid as the sole source of histone H3 to an rrm3∆ strain and analyzed the spores after sporulation of the diploid. Results obtained from the dissection of 160 tetrads are presented in Table 1. We were unable to recover any rrm3∆ spores expressing wild-type H4 and the H3K56R mutant as the sole source of histone H3, showing that H3K56ac is vital in rrm3∆ cells (Figure 2A and Table 1).

Table 1. Histone H3-H4 dependence on rrm3∆ viability.

Histone genes	RRM3	rrm3Δ
HHT1-HHF1 HHT2-HHF2	58	44
HHT1-HHF1 hht2Δ-hhf2Δ	24	48
hht1Δ-hhf1Δ HHT2-HHF2	20	41
hht1Δ-hhf1Δ hht2Δ-hhf2Δ	0	0
hht1Δ-hhf1Δ hht2Δ-hhf2Δ+hht1-K56R-HHF1	20	0
HHT1-HHF1 hht2Δ-hhf2Δ+hht1-K56R-HHF1	37	27
hht1Δ-hhf1Δ HHT2-HHF2+hht1-K56R-HHF1	37	37
HHT1-HHF1 HHT2-HHF2+hht1-K56R-HHF1	75	31

A total of 160 tetrads from diploids for hht1∆−hhf1∆/HHT1-HHF1 hht2∆−hhf2∆/HHT2-HHF2 rrm3∆/RRM3 expressing HHF1 and hht1-K56R from a centromeric plasmid were dissected, and the genotype of the viable spores was determined. The number of viable spores carrying each deletion and/or plasmid is indicated. Among the rrm3∆ spores carrying the hht1-K56R allele, only spores also expressing the wild-type HHT1 allele are viable.

Figure 2.

Hyperacetylation and hypoacetylation of lysine 56 of histone H3 affect rrm3∆ cells differently. (A) H3K56R mutation is lethal in rrm3∆ cells. Tetrads from diploids for hht1∆-hhf1∆/HHT1-HHF1 hht2∆-hhf2∆/HHT2-HHF2 rrm3∆/RRM3 expressing HHF1 and hht1-K56R from a centromeric plasmid were dissected and analyzed for the presence of auxotrophic markers. Dashed circle indicates rrm3∆ spore expressing H3K56R as sole source of histone H3. (B) rrm3∆ cells are viable with constitutively acetylated H3K56. The hst3∆/HST3 hst4∆/HST4 rrm3∆/RRM3 sir2∆/SIR2 diploid strain was dissected. The presence of rrm3∆ hst3∆ hst4∆ and rrm3∆ sir2∆ hst3∆ hst4∆ mutants is indicated by a circle and by dashed circles, respectively.

The Sir2-related Hst3 and Hst4 histone deacetylases regulate histone H3K56ac both during the normal cell cycle and after DNA damage. Consequently, loss of HST3 and HST4 results in the constitutive hyperacetylation of H3K56 (Masumoto et al. 2005; Celic et al. 2006; Maas et al. 2006; Miller et al. 2006; Yang et al. 2008; Delgoshaie et al. 2014). SIR2 is also required to deacetylate H3K56 at specific heterochromatic sites (F. Xu et al. 2007). To further examine the impact of H3K56ac in the absence of RRM3, we deleted one allele of RRM3, HST3, HST4, and SIR2 in a diploid strain. By analyzing the spores derived from this diploid, we found that both the hst3∆ hst4∆ rrm3∆ triple and sir2∆ hst3∆ hst4∆ rrm3∆ quadruple mutants were viable (Figure 2B). Thus, in contrast to the absence of H3K56ac, constitutive H3K56ac is not deleterious in the absence of RRM3.

H3K56ac-dependent coordination between nucleosome assembly and stability of advancing replication forks is not required for viability in the absence of RRM3

We have shown that HIR-dependent replication-independent chromatin assembly is not required for the viability of rrm3∆ cells. We then asked if the replication-coupled chromatin assembly function of H3K56ac is important for rrm3∆ cell viability. H3K56ac facilitates replication-coupled chromatin assembly by increasing the association of new histone molecules with CAF-1 and Rtt106 (Li et al. 2008; Han et al. 2013). This pathway coordinates nucleosome assembly and stability of the advancing replication forks but is not required for H3K56ac-mediated protection against replicative DNA-damaging agents by DNA repair/tolerance mechanisms (Clemente-Ruiz et al. 2011). Consequently, during replication, similar defects arise in the asf1∆ cac1∆ rtt106∆ and the cac1∆ rtt106∆ mutants (Clemente-Ruiz et al. 2011; Prado and Clemente-Ruiz 2012). Thus, we analyzed the viability of rrm3∆ cac1∆ rtt106∆ cells by looking at the meiotic progeny of a diploid strain heterozygous for RRM3, CAC1, and RTT106 deletions. We found that simultaneous deletion of CAC1 and RTT106 in the rrm3∆ causes slow growth, but unlike asf1∆, is viable (Figure 3, A and B). This result indicates that a defective DNA repair/tolerance mechanism, rather than an alteration of the replication-coupled chromatin assembly per se, causes lethality of rrm3∆ cells in the absence of H3K56ac.

Figure 3.

A Cac1/Rtt106-independent function of H3K56ac is required for viability of rrm3∆ cells. (A) Defective Cac1/Rtt106-dependent chromatin assembly does not cause lethality in the absence of RRM3. Tetrads from the rtt106∆/RTT106 cac1∆/CAC1 asf1∆/ASF1 rrm3∆/RRM3 diploid strain were dissected. Diamonds indicate cac1∆ rtt106∆ mutants. Hexagon indicates the rtt106∆ rrm3∆ mutant. Square indicates the cac1∆ rrm3∆ mutant. Circle indicates the cac1∆ rtt106∆ rrm3∆ mutant. Dashed circles indicate asf1∆ rrm3∆ mutants. Triangle indicates asf1∆ cac1∆ mutant. (B) Effects of cac1∆ and rtt106∆ on viability of rrm3∆ cells. Yeast strains of indicated genotypes were streaked onto YPD plates and grown at 30° for 3 days.

CTF4 is harmful upon DNA damage in absence of a functional H3K56ac pathway

Both RTT107, which supposedly functions in the same genetic pathway as RTT101, MMS1, and MMS22 to maintain genome stability, and CTF4 interact with the Rtt101–Mms1 complex through the adaptor protein Mms22 (Pan et al. 2006; Collins et al. 2007; Gambus et al. 2009; Mimura et al. 2010). We have analyzed the consequences of RTT101, MMS1, MMS22, RTT107, and CTF4 deletions in rrm3∆ cells and found that all double mutants are lethal with the only exception of ctf4∆ rrm3∆ mutant (Figure S3). Based on the fact that ctf4∆ suppresses some negative phenotypes associated with constitutive H3K56ac (Celic et al. 2008) and on our observations, we hypothesized that Ctf4 may be a target of the H3K56ac pathway. To test this hypothesis, we evaluated the effects of deleting CTF4 when the RRM3 deletion was combined with mutations affecting different steps of the H3K56ac pathway. The results presented in Figure 4 show that, despite its negative effect in the single rrm3∆ mutant, the inactivation of CTF4 restores the growth of each of the double mutants (asf1∆ rrm3∆, rtt109∆ rrm3∆, rtt101∆ rrm3∆, mms1∆ rrm3∆, and mms22∆ rrm3∆). To ascertain whether the effect of ctf4∆ is linked to H3K56ac, we further examined the impact of the CTF4 deletion in the rrm3∆ H3K56R double mutant. As expected, we found that ctf4∆ restored the viability of rrm3∆ cells expressing H3K56R as the sole source of H3 (Figure S4). Strikingly, the growth of each triple mutant seemed similar to the one of the rrm3∆ ctf4∆ mutant with one exception: ctf4∆ mms22∆ rrm3∆ (Figure 4). The slower growth of the ctf4∆ mms22∆ rrm3∆ strain may be explained by the multiple roles of Mms22 in the response to DNA damage (Wurtele et al. 2011). Moreover, we have found that ctf4∆ is not able to modify the viability of rtt107∆ rrm3∆ cells (Figure S5), indicating that CTF4 and RTT107 act in distinct pathways to deal with replicative damage and suggesting that RTT107 does not belong to the H3K56ac pathway.

Figure 4.

CTF4 deletion suppresses rrm3∆ lethality in different genetic contexts affecting the H3K56ac pathway. (A) CTF4 deletion rescues asf1∆ rrm3∆ lethality. Tetrads from diploids heterozygous for ctf4∆, rrm3∆, and asf1∆ were dissected and analyzed after 3 days at 30°. Circles indicate ctf4∆ asf1∆ rrm3∆ mutants. Dashed circle indicates asf1∆ rrm3∆ mutants. In A–E, diamonds indicate ctf4∆ rrm3∆ mutants. (B) CTF4 deletion rescues rtt109∆ rrm3∆ lethality. Tetrads from diploids heterozygous for ctf4∆, rrm3∆, and rtt109∆ were dissected and analyzed as in A. Circles indicate ctf4∆ rtt109∆ rrm3∆ mutants. Dashed circles indicate rtt109∆ rrm3∆ mutants. (C) CTF4 deletion rescues rtt101∆ rrm3∆ lethality. Tetrads from diploids heterozygous for ctf4∆, rrm3∆, and rtt101∆ were dissected and analyzed as in A. Circle indicates ctf4∆ rtt101∆ rrm3∆ mutant. Dashed circles indicate rtt101∆ rrm3∆ mutants. (D) CTF4 deletion rescues mms1∆ rrm3∆ lethality. Tetrads from diploids heterozygous for ctf4∆, rrm3∆, and mms1∆ were dissected and analyzed as in A. Circle indicates ctf4∆ mms1∆ rrm3∆ mutants. Dashed circles indicate mms1∆ rrm3∆ mutants. (E) CTF4 deletion partially rescues mms22∆ rrm3∆ lethality. Tetrads from diploids heterozygous for ctf4∆, rrm3∆, and mms22∆ were dissected and analyzed as in A. Circles indicate ctf4∆ mms22∆ rrm3∆ mutants. Dashed circles indicate mms22∆ rrm3∆ mutants.

We next investigated the deleterious effect of CTF4 when cells affected in the H3K56ac pathway were treated with the methylating agent methyl methanesulfonate (MMS). Deletion of CTF4 increased the viability of asf1∆, rtt109∆, mms1∆, and mms22∆ single mutants. These results corroborate our findings obtained in the settings of the RRM3 deletion and further support the notion that H3K56ac modulates replisome function during replicative stress through CTF4 (Figure 5). The only exception is the double mutant ctf4∆ rtt101∆, which is more sensitive compared to the rtt101∆ single mutant. A previous report showed that, in addition to its function in the H3K56ac pathway, RTT101 exerts an MMS1- and MMS22-independent function during replication through its interaction with histone chaperones (Han et al. 2010). Interestingly, the ctf4∆ rtt101∆ double mutant is less sensitive to MMS as compared to ctf4∆, suggesting that the CTF4 deletion is beneficial in the presence of MMS in the rtt101∆ mutant.

Figure 5.

In presence of MMS, CTF4 becomes harmful for cells affected in the H3K56 acetylation pathway. Fivefold serial dilutions of exponentially growing cells were spotted onto a YPD plate or 0.015% MMS plate and incubated at 30° for 3 days.

Ctf4-mediated uncoupling of DNA polymerase-α and GINS is crucial for the viability of rrm3∆ cells experiencing DNA damage

Ctf4 couples the CMG helicase to polymerase-α in the replisome by interacting simultaneously with both polymerase-α and the GINS complex (Gambus et al. 2009; Tanaka et al. 2009; Simon et al. 2014). To determine the importance of the Ctf4-bridging function, we analyzed a Ctf4-(1-383) truncated mutant (Ctf4-NT) that is unable to bind the GINS and polymerase-α together with the Ctf4-(1-383) mutant (Ctf4-∆NT) that has kept its ability to bind GINS and polymerase-α (Gambus et al. 2009). The ctf4-NT construct restores the viability of asf1∆ rrm3∆ (Figure 6A), as does the complete absence of CTF4. In contrast, analysis of >100 tetrads did not yield a single ctf4-∆NT asf1∆ rrm3∆ triple mutant (Figure 6B). Similar results were observed when the double mutants ctf4-NT asf1∆ and ctf4-∆NT asf1∆ were exposed to CPT (Figure 6C). Based on these results we suggest that the bridging function of Ctf4 is deleterious in the absence of H3K56ac following DNA damage.

Figure 6.

Uncoupling of MCM helicase and DNA polymerase-α favors cell viability during replicative stress in the absence of H3K56 acetylation. (A) The inability of Ctf4 to bind GINS and DNA polymerase-α restores the viability of the asf1∆ rrm3∆ mutant. Tetrad dissection from the ctf4-NT/CTF4 asf1∆/ASF1 rrm3∆/RRM3 diploid strain. Diamond, circle, and dashed circle indicate ctf4-NT rrm3∆, ctf4-NT asf1∆ rrm3∆, and asf1∆ rrm3∆ mutants, respectively. (B) The ability of Ctf4 to bind GINS and DNA polymerase-α is lethal in the asf1∆ rrm3∆ mutant. Diamond, circles, and dashed circle indicate ctf4-∆NT rrm3∆, ctf4-∆NT asf1∆ rrm3∆, and asf1∆ rrm3∆ mutants, respectively. (C) Sensitivity to CPT of the combination of the mutants ctf4∆, ctf4-NT, and ctf4-∆NT with asf1∆. Fivefold serials dilutions of exponentially growing cells were spotted onto YPD and 4µM CPT plates and incubated at 30° for 3 days. (D) Affecting the stability of the catalytic subunit of the DNA polymerase-α (Cdc17) restores the viability of asf1∆ rrm3∆ mutant. Tetrad dissection from cdc17-1/cdc17-1 asf1∆/ASF1 rrm3∆/RRM3 diploid strain. Circles indicate the asf1∆ rrm3∆ cdc17-1 mutants.

To confirm that the restoration of the viability of the asf1∆ rrm3∆ mutant by the CTF4 deletion is associated with the uncoupling of the helicase and polymerase-α, we used the thermosensitive (ts) mutant cdc17-1 encoding the catalytic subunit of the DNA polymerase-α. In agreement with our hypothesis, we found that the cdc17-1 mutation rescues asf1∆ rrm3∆ cell viability at the semirestrictive temperature (30°), but not at the permissive temperature (25°) (Figure 6D).

By analyzing the Ctf4 truncations, we observed that the ctf4-∆NT allele was lethal by itself when combined with rrm3∆ (Figure 6B). Based on the results from K. Labib’s and T. Kamura’s laboratories showing that the N-terminal region of Ctf4 interacts with Mms22 (Morohashi et al. 2009; Mimura et al. 2010) and because the double ctf4∆ rrm3∆ was viable, we hypothesized that the lethality of the ctf4-∆NT rrm3∆ mutant could be due to the inability of Ctf4-∆NT to interact with Mms22 and consequently to modulate Ctf4 interaction with GINS and polymerase-α. If this hypothesis is valid, we expect that the Ctf4–Mms22 interaction functions in the H3K56ac pathway and that a defect in this interaction phenocopies a lack of H3K56ac. In such a case, the viability of hst3∆ hst4∆ rrm3∆ should be also compromised by ctf4-∆NT despite the constitutive hyperacetylation of H3K56. We have shown that the hyperacetylation of H3K56 resulting from the simultaneous deletion of HST3 and HST4 is not lethal in rrm3∆ cells (Figure 2B). We now show that the viability of hst3∆ hst4∆ rrm3∆ is compromised by ctf4-∆NT (Figure S6). These observations reinforce the idea that the ctf4-∆NT mutant is insensitive to H3K56ac and strongly suggest that the role of Ctf4 in the H3K56ac pathway relies on its interaction with Mms22. These results indicate that the key function of Ctf4 in coupling DNA polymerase-α to the CMG helicase in wild-type cells becomes deleterious in rrm3∆ cells lacking H3K56ac. Taken together, these genetic analyses suggest that one function of H3K56ac could be to modulate the replisome in the presence of DNA damage, probably through an interaction between Ctf4 and Mms22.

Ctf4 association to chromatin is not notably affected in rrm3∆ cells

To further examine the relationships between CTF4 and H3K56ac, we analyzed the chromatin-bound Ctf4 together with H3K56ac (see Materials and Methods). Consistent with published data (Masumoto et al. 2005; Wang et al. 2010), we observed that Ctf4 was bound to chromatin mainly during S phase and dissociated in late S or G2 phase (Figure 7A). Interestingly, chromatin dissociation of Ctf4 occurred concomitantly with the increase in H3K56ac (Figure 7A). In the absence of RRM3, H3K56ac levels were not significantly affected, but the timing of Ctf4 dissociation from chromatin was delayed (Figure 7B). We assumed that the persistence of Ctf4 association to chromatin in the rrm3∆ mutant probably reflected the fact that rrm3∆ cells took longer to traverse from late S phase into the next cell cycle, as a consequence of the accumulation of DNA lesions during S phase in the absence of RRM3. In agreement with this, we found that Ctf4 persists in chromatin also in the presence of CPT (Figure 7C). These results indicate that the uncoupling observed upon DNA damage is not mediated by Ctf4 degradation. Finally, we analyzed the global level of Ctf4 in the wild-type control strain CTF4-MYC and in H3K56R CTF4-MYC mutant cells. Consistent with our genetic analysis and with the fact that CTF4 function is detrimental to yeast cells lacking H3K56ac (Pan et al. 2006), we repeatedly observed that the amount of Ctf4 was reduced in the H3K56R mutant (Figure 7D), suggesting that one adaptation of H3K56R cells is to reduce their levels of Ctf4 to promote their growth.

Figure 7.

Ctf4 chromatin association is not affected in presence of replicative damages. (A) Cell cycle chromatin association of Ctf4. Ctf4-Myc cells were synchronized in G1 with α-factor and released into fresh medium at 25°. Samples were collected every 10 min, crude chromatin was prepared and analyzed by Western blot with 9E10 antibody for Ctf4-Myc (upper) and H3K56ac antibody for H3K56 acetylation (lower) using the same blot. Cell cycle progression was followed by FACS analysis (right). (B) Cell cycle chromatin association of Ctf4 in rrm3∆ cells. Ctf4-Myc rrm3∆ cells were treated and analyzed as in A. (C) Cell cycle chromatin association of Ctf4 in the presence of CPT. Ctf4-Myc cells were synchronized in G1 with α-factor and released in a new cell cycle at 25° in the presence of 40 µM of CPT. Samples were collected, prepared, and analyzed as in A. (D) The level of Ctf4 is reduced in hht1-K56R cells. ctf4-myc and ctf4-myc hht1∆-hhf1∆ hht2∆-hhf2∆ +phht1-K56R-HHF1 cells were synchronized in G1 with α-factor and released in a fresh medium at 30°. Samples were collected every 10 min and analyzed by Western blot with 9E10 antibody for Ctf4-Myc detection (upper). Anti-Rfa1 antibody was used as a loading control. Cell cycle progression was followed by FACS analysis (right).

Interaction between Mms22 and the replisome is promoted in rrm3∆ cells

Next, we examined whether the interaction between the replisome and Mms22 is modulated by RRM3 deletion using a quantitative mass spectrometry approach. Mass spectrometry analyses conducted in wild-type and rrm3∆ cells after immunoprecipitation of Ctf4-GFP protein during S phase allowed us to identify Mms22, Mcm2-7 helicase, the catalytic subunit of DNA polymerase-α (Cdc17), and other components of the replisome-progressing complex and factors that will be described elsewhere (Figure 8A and Figure S7). Consistent with our previous finding that Ctf4 is not degraded in rrm3∆ cells (Figure 7), we found similar Ctf4 levels in either the presence or absence of RRM3. Interestingly, despite lower Mms22 levels in rrm3∆ cells (Figure 8B, right), we repeatedly observed, after immunoprecipitation of Ctf4-GFP, an enrichment of Mms22 without any increase of Ctf4, Mcm2-7, or Cdc17 levels in the absence of RRM3 (Figure 8A). These results, showing a specific enrichment of Mms22 at forks in the absence of RRM3, are consistent with our genetic observations obtained with the Ctf4 mutants. They reinforce our conclusion that the interaction between Ctf4 and Mms22 is regulated during replicative damage and is crucial in dealing with replicative stress.

Figure 8.

Replicative stress induced by the absence of RRM3 increases Mms22 association with the replisome. (A) Pulldown protein extracts were loaded on Bis–Tris acrylamide gels in MOPS buffer and staked as a single band before trypsin digestion followed by mass spectrometry analysis. (Left) Mass spectroscopy data obtained after immunoprecipitation of Ctf4-GFP during S phase for Mms22, the catalytic subunit of the DNA polymerase-α (Cdc17), and MCM helicase. Spectral counts show the total number of identified peptide sequences for the indicated protein in each sample (RRM3 CTF4-GFP, rrm3∆ CTF4-GFP, and control CTF4). (Right) Relative quantitation of Mms22 protein compared to Ctf4 protein measured by the ratio of the sum of the areas of the three more intense precursor ions used for each protein identification. The averages of several independent experiments are shown. (B) The level of Mms22 is decreased in rrm3∆ cells. Protein extracts from mms22∆ + pG16adh-TAP-MMS22 and mms22∆ rrm3∆ + pG16adh-TAP-MMS22 strains were prepared from S-phase-synchronized cells and analyzed by Western blot with a protein-A antibody (right). Total proteins on the membrane were stained with Ponceau S as a loading control (left).

Replisome destabilization allows cells lacking a functional H3K56ac pathway to survive under replicative stress

To better understand the role of H3K56ac at forks, we extended our analysis to another crucial component of the RPC. In an unperturbed S phase, Mrc1 senses and regulates replisome integrity by interacting with multiple components of the replisome, such as Tof1/Csm3 and CMG complexes, Dia2, Ctf4, and the DNA polymerase-ɛ (Kanemaki et al. 2003; Katou et al. 2003; Nedelcheva et al. 2005; H. Xu et al. 2007; Lou et al. 2008; Komata et al. 2009; Mimura et al. 2009; Morohashi et al. 2009; Naylor et al. 2009; Uzunova et al. 2014). Interestingly, as shown for ctf4∆ cells, replication forks progress more slowly in mrc1∆ cells (Szyjka et al. 2005; Tourrière et al. 2005; Hodgson et al. 2007). Considering our findings with respect to CTF4, and the results of previous genetic screens (Collins et al. 2007; Fiedler et al. 2009; Haber et al. 2013), it is possible that the replicative function of MRC1 could also be deleterious in the presence of DNA damage, when H3K56ac is compromised. To test this possibility, because mrc1∆ is synthetic lethal with rrm3∆ (Naylor et al. 2009), we first examined the effects of the MRC1 deletion in asf1∆ cells subjected to replicative stress. We evaluated the viability of the asf1∆ mrc1∆ cells in the presence of DNA damage induced by CPT and MMS. We observed that deleting MRC1 strongly suppressed the CPT and MMS sensitivity of asf1∆ cells without affecting asf1∆ thermosensitivity (Figure 9A). In addition to its role in replication, MRC1 is also required for checkpoint activation upon DNA replication stress. To further determine which function of MRC1 is harmful to asf1∆ cells in the presence of DNA replication stress, we tested two separation-of-function mutants of MRC1. We found that the mrc1-C14 mutant that is compromised for its replication function (Naylor et al. 2009) behaved as mrc1∆, whereas the checkpoint defective mrc1-AQ mutant had no effect (Figure 9B). We obtained similar results with mms1∆ cells (Figure S8). These results point out the importance of the MRC1 replication function in the sensitivity to DNA replication stress of cells deficient for the H3K56ac pathway.

Figure 9.

The replicative function of MRC1 is deleterious in asf1∆ cells experiencing replicative damages. (A) Effects of mrc1∆ on viability of the asf1∆ cells. Tenfold serial dilutions of exponentially growing cells were spotted onto YPD plates incubated at 30° or 38°, onto 5-µM CPT, and 0.01%-MMS plates incubated at 30° for 3 days. (B) Effects associated with the replicative and checkpoint functions of MRC1 on viability of the asf1∆ cells. Tenfold serial dilutions of exponentially growing cells were spotted onto YPD, 5-µM CPT, and 0.005%-MMS plates and incubated at 30° for 3 days.

Discussion

In this study, we first show that, in the absence of ASF1, the rrm3∆ growth defect is a direct consequence of the lack of H3K56ac. Our detailed genetic analyses indicate that, upon DNA replication stress, the lack of H3K56ac mainly affects DNA repair and/or DNA damage tolerance mechanisms implicated in the response to replicative DNA damage.

Both ASF1 and RTT109 are required for H3K56R and function together with the E3 ubiquitin ligase complex Rtt101–Mms1–Mms22 (which is itself dispensable for H3K56ac) in the H3K56ac pathway (Collins et al. 2007). We have discovered that inactivating CTF4 restores the viability of asf1∆ rrm3∆, rtt109∆ rrm3∆, mms1∆ rrm3∆, and mms22∆ rrm3∆ mutants. Our results show that the loss of CTF4 allows the cell to survive when the fork encounters obstacles in the absence of a functional H3K56ac pathway. According to the Ctf4 role in coupling the MCM helicase to polymerase-α during normal replication (Gambus et al. 2009; Tanaka et al. 2009), we propose that the regulation of this function is crucial in the presence of replicative damages. The importance of uncoupling the helicase from polymerase-α was further strengthened by our observations showing that the ctf4-NT mutation that affects Ctf4 interactions with GINS and polymerase-α (Gambus et al. 2009), and consequently its coupling ability between the helicase and polymerase, also restores the viability of asf1∆ cells upon replicative damage in contrast to the ctf4-∆NT mutation that preserves the Ctf4 coupling function (Gambus et al. 2009). Similarly, we have shown that affecting the stability of the large subunit of the DNA polymerase-α also restores the viability of asf1∆ rrm3∆ cells devoid of H3K56ac (see below). We therefore concluded that in the presence of replicative stress a coordinated progression of helicase and DNA polymerase-α is harmful in the absence of a functional H3K56ac pathway. We propose that one consequence of H3K56ac upon DNA replication stress would be to modulate replisome integrity by weakening the coupling between the MCM helicase and polymerase-α through Ctf4.

We also point out the importance of the MRC1 replication function in the sensitivity to replication stress of cells deficient for the H3K56ac pathway. Indeed, during normal replication, a Ctf4 trimer interacts with multiple replisome components including Mrc1 and polymerase-ɛ (Simon et al. 2014). Moreover, loss of Ctf4 affects the association of both leading and lagging replication proteins with the fork (Tanaka et al. 2009), suggesting that Ctf4 plays a role in the coordinated progression of the MCM helicase with both leading- and lagging-strand synthesis. Therefore, we propose that, in response to DNA replication stress, Ctf4 could modulate the coordinated progression of MCM helicase with both lagging- and leading-strand polymerases in a H3K56ac-dependent way. Our findings showing that CTF4 negatively affects the growth of rrm3∆ cells in the absence of a functional H3K56ac pathway (Figure 4) is consistent with the idea that Ctf4 is the major, if not only, target of the H3K56ac pathway during DNA replication stress.

Among the ctf4∆, ctf4-NT, and ctf4-∆NT mutants we found that only ctf4-∆NT, which retains the ability to bind GINS and DNA polymerase-α, is synthetic-lethal with rrm3∆ cells despite the presence of normal levels of H3K56ac. This negative genetic interaction indicates that deficient replication fork progression arising in rrm3∆ cells requires a CTF4-dependent regulation of the replisome function. Ctf4 has been shown to interact with the Rtt101–Mms1–Mms22 complex through an interaction with Mms22 (Mimura et al. 2010). Since Ctf4 needs its N-terminal portion to interact with Mms22 (Gambus et al. 2009; Mimura et al. 2010), we speculate that the inability of Ctf4-∆NT to interact with Mms22 affects the H3K56ac-dependent uncoupling between helicase and polymerase, required to cope with DNA replication stress, and causes rrm3∆ lethality. In the absence of Ctf4 or its C-terminal part, this interaction is probably no longer required for rrm3∆ cells viability because of the uncoupling between MCM helicase and polymerase-α. Taken together, these results reinforce the notion that CTF4 belongs to the H3K56ac pathway.

How could Ctf4 action be regulated by the H3K56ac pathway in the presence of DNA replicative damage? Post-translational modifications of replisome components play a key role in regulating fork progression (Zech and Dalgaard 2014). A recent study has shown that Schizosaccharomyces pombe cells elicit a program to degrade replisome upon DNA replication stress through the action of the ubiquitin–proteasome system SCF^Pof3, a homolog of the budding yeast SCF^Dia2 ubiquitin–proteasome system (Roseaulin et al. 2013a,b). Because Ctf4 physically interacts with Dia2, which drives CMG helicase and replisome disassembly at the end of DNA replication (Maric et al. 2014; Moreno et al. 2014) and is ubiquitylated by SCF^Dia2 (Ho et al. 2002; Collins et al. 2007; Mimura et al. 2009; Morohashi et al. 2009), one possibility could be that Ctf4 is targeted by Dia2 in response to replicative stress in a way dependent on H3K56ac. However, because the ctf4-∆NT mutation that preserves the Ctf4 interaction with Dia2 (Mimura et al. 2009; Morohashi et al. 2009) does not restore the viability of asf1∆ rrm3∆ cells and is lethal in the rrm3∆ mutant (in contrast to the ctf4-NT mutation that loses the Ctf4–Dia2 interaction), we exclude the possibility that Ctf4 action could be regulated during replicative stress by H3K56ac through an action mediated by the SCF^Dia2.

Mms22 is recruited to chromatin at stalled replication forks (Dovey et al. 2009; Ben-Aroya et al. 2010; Vaisica et al. 2011). An attractive hypothesis would be that Ctf4 is a substrate of the Rtt101–Mms1–Mms22 complex and is further degraded at stalled replication forks under replicative stress. However, we failed to see any decrease in Ctf4 level in the presence of replicative damage (Figure 7, A and B; Figure 8A). Although we cannot exclude that Ctf4 degradation could occur locally and transiently, we presume that Ctf4 is not degraded. We infer that, if the Rtt101–Mms1–Mms22 complex ubiquitylates Ctf4, this ubiquitylation mediates a certain function of Ctf4, rather than Ctf4 degradation. Another possibility could be that Ctf4 is released from forks at damaged chromatin. However, we also failed to observe a significant difference in the amount of Ctf4 associated with bulk chromatin in the absence or presence of replicative damages, suggesting that the H3K56ac pathway did not affect Ctf4 association in chromatin globally. Thus, putative ubiquitylation of Ctf4 would neither affect Ctf4 stability nor Ctf4 chromatin association but would regulate Ctf4 action at the replisome. Interestingly, we found that a H3K56R mutant known to contain intrinsic DNA damage (Hyland et al. 2005; Masumoto et al. 2005; Pan et al. 2006; Recht et al. 2006; Celic et al. 2008; Wurtele et al. 2011) exhibited a lower level of Ctf4, suggesting that cells lacking H3K56ac increased their fitness by reducing the amount of Ctf4 and thereby the Ctf4 action.

Upon DNA replication stress, H3K56ac is maintained at high levels behind the replication fork in a checkpoint-dependent manner (Masumoto et al. 2005; Maas et al. 2006). The accumulation of H3K56ac at forks is thought to increase DNA accessibility (Masumoto et al. 2005; Driscoll et al. 2007; Yang et al. 2008) and to create a unique chromatin environment allowing cells to deal with replicative stress. Cells with constitutively high levels of H3K56ac are extremely sensitive to subtle perturbations in DNA replication, and deletion of CTF4 partially suppresses some of their phenotypes (Celic et al. 2008). We found that rrm3∆ cells, known to accumulate DNA replication stress at specific regions throughout the genome, were viable in this context, confirming that the chromatin environment induced by acetylation of H3K56 behind replication forks is crucial. Because the viability of ctf4-∆NT rrm3∆ hst3∆ hst4∆ cells experiencing constitutively high levels of H3K56ac is compromised, we assume that an interaction between Mms22 and Ctf4 promotes H3K56ac-, Rtt101-Mms1-Mms22-, and Ctf4-dependent molecular events required to rescue replication fork damages during replicative stress. This hypothesis is clearly reinforced by our findings showing that RRM3 deletion leads to an increased enrichment of Mms22. In good agreement with the fact that Mms22 is recruited to chromatin in a DNA damage-, RTT109-, and RTT101-dependent manner (Dovey et al. 2009; Ben-Aroya et al. 2010; Vaisica et al. 2011) and, in the light of recent work showing that the Rtt101–Mms1–Mms22 complex binds H3K56ac-H4 preferentially over unmodified H3-H4 (Han et al. 2013), we foresee that H3K56ac signals DNA damage and directly recruits the Rtt101–Mms1–Mms22 complex behind the fork to ubiquitylate H3K56 acetylated histones, Ctf4, and various substrates leading to an increased DNA accessibility at damage sites and improved DNA repair. Alternatively, Ctf4 could also recruit the Rtt101-Mms1 at the DNA damage fork through its interaction with Mms22, and its function at the replisome could be subsequently regulated by the complex. Consistent with the deleterious effect of CTF4 in rrm3∆ cells lacking H3K56ac, the coupling function of Ctf4, and its ability to act as a platform for multivalent interactions (Simon et al. 2014), we propose that upon DNA replication stress Ctf4 acts downstream of H3K56ac and the Rtt101–Mms1–Mms22 complex to modulate the integrity of the replisome by affecting both lagging and leading strands at the forks through an interaction with Mms22 (Figure 10).

Figure 10.

H3K56 acetylation prevents genomic instability by affecting Ctf4 function. During S phase, in an unperturbed cell cycle (left), Asf1 functions cooperatively with Rtt109 to acetylate new histone H3 at lysine 56 (solid circles). The Rtt101–Mms1–Mms22 complex binds and ubiquitylates new H3K56ac histones (shaded circles) and promotes an efficient progression of the replication fork and nucleosome assembly by favoring H3-H4 transfer from Asf1 to other histone chaperones. At the end of S-phase, Hst3 and Hst4 deacetylases remove H3K56ac. How ubiquitylation is removed is still unknown. In response to DNA-damaging agents and in rrm3∆ cells (right), replication fork progression is affected, leading to checkpoint activation and subsequently to transcriptional repression of HST3 and HST4 and degradation of Hst3 and Hst4, allowing H3K56ac to persist. The unique chromatin environment created by H3K56ac accumulation behind the fork triggers a crucial interaction between Ctf4 and the Rtt101–Mms1–Mms22 complex through the N-terminal domain of Ctf4. This interaction modulates replisome structure by uncoupling MCM helicase and DNA polymerases and increases genome accessibility at replication defective forks, thus preserving genome integrity. It is currently unclear whether the interaction between Ctf4 and the Rtt101–Mms1–Mms22 complex is required to ubiquitylate Ctf4 itself and modulate its function or to recruit the Rtt101–Mms1–Mms22 complex to the site of DNA damage to ubiquitylate other replisome components. The function of the H3K56ac pathway is totally abolished in the absence of the N-terminal domain of Ctf4 required for its interaction with Mms22, positioning Ctf4 in the H3K56ac pathway.

DDB1, which binds CUL4A in humans to form the CUL4A^DDB1 E3 ubiquitin ligase, shares sequence homology with Mms1 (Lee and Zhou 2007; Jackson and Xiong 2009; Havens and Walter 2011). The CUL4A^DDB1 complex ubiquitylates various proteins in response to DNA damage (Nouspikel 2011), regulates both the replication-coupled and replication-independent nucleosome assembly (Han et al. 2013), and is critical for new histone H3.3 accumulation at sites of DNA damage (Adam et al. 2013). H3K56ac is conserved in humans and seems to also be regulated in a DNA damage-dependent way (Das et al. 2009; Tjeertes et al. 2009; Yuan et al. 2009). Increased levels of H3K56ac have been observed in cancer cells in a manner that is proportional to tumor grade (Das et al. 2009). In mammals, H3K56 is acetylated by Gcn5 (Tjeertes et al. 2009; Burgess and Zhang 2010). Interestingly, the Ctf4 human homolog And-1 interacts with H3 and Gcn5, impairs the interaction between Gcn5 and an E3 ligase complex, and modulates the H3K56ac level (Li et al. 2011, 2012). Similarly to Ctf4, And-1 also plays an important role in bridging between CMG helicase and DNA polymerases, suggesting that And-1 coordinates DNA unwinding and polymerase activities in humans (Zhu et al. 2007; Aze et al. 2013; Kang et al. 2013). Our results raise the possibility that the H3K56ac pathway is conserved between yeast and humans and that And-1 regulates the human replisome in an H3K56ac-dependent manner through the action of the CUL4A^DDB1 complex, which is thought to increase DNA accessibility at damage sites (Nouspikel 2011). An attractive possibility that needs to be tested in yeast would be that Ctf4 directly interacts with histones and/or histone chaperones behind the fork and, by sensing the nature and the level of histone modifications, together with the Rtt101–Mms1–Mms22 complex, regulates replisome integrity and function in response to replicative stress.