Defective transfer of parental histone decreases frequency of homologous recombination by increasing free histone pools in budding yeast

Abstract

Recycling of parental histones is an important step in epigenetic inheritance. During DNA replication, DNA polymerase epsilon subunit DPB3/DPB4 and DNA replication helicase subunit MCM2 are involved in the transfer of parental histones to the leading and lagging strands, respectively. Single Dpb3 deletion (dpb3Δ) or Mcm2 mutation (mcm2-3A), which each disrupts one parental histone transfer pathway, leads to the other's predominance. However, the biological impact of the two histone transfer pathways on chromatin structure and DNA repair remains elusive. In this study, we used budding yeast Saccharomyces cerevisiae to determine the genetic and epigenetic outcomes from disruption of parental histone H3-H4 tetramer transfer. We found that a dpb3Δ mcm2-3A double mutant did not exhibit the asymmetric parental histone patterns caused by a single dpb3Δ or mcm2-3A mutation, suggesting that the processes by which parental histones are transferred to the leading and lagging strands are independent. Surprisingly, the frequency of homologous recombination was significantly lower in dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A mutants, likely due to the elevated levels of free histones detected in the mutant cells. Together, these findings indicate that proper transfer of parental histones during DNA replication is essential for maintaining chromatin structure and that lower homologous recombination activity due to parental histone transfer defects is detrimental to cells.

Graphical Abstract

Figure 1.

Combination of dpb3Δ and mcm2-3A mutations neutralizes single dpb3Δ or mcm2-3A mutants’ strand bias in transferring parental histone H3–H4 tetramers during DNA replication. (A) Procedure for monitoring the deposition of parental (H3K4me3) and newly synthesized (H3K56ac) histone H3 at early replication origins. (B) Snapshot of parental histone H3 (H3K4me3) eSPAN reads enrichment at leading and lagging strands at the early replication origin ARS1309 for wild-type (WT), dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains. The sequence reads were mapped to both the Watson strand (red) and the Crick strand (green) of the reference genome. (C–F) Top: heatmaps representing the bias ratio of parental histone H3 (H3K4me3) eSPAN peaks for WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains at each of the 10 individual nucleosomes surrounding each of the 134 early DNA replication origins. Individual nucleosomes are represented by the circles at the top of the heatmaps, and their positions are indicated relative to the origin (−10 to + 10). Each row represents the average log₂ Watson/Crick ratio of H3K4me3 eSPAN sequence reads at one origin. Bottom: average bias ratio of parental histone H3 (H3K4me3) eSPAN peaks for WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains at each of the 10 nucleosomes surrounding the 134 early replication origins.

Figure 2.

The chromatin-structure changes by dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A mutation at early S phase. (A) Procedure used for micrococcal nuclease (MNase) chromatin accessibility assay and MNase-seq, to characterize chromatin accessibility and nucleosome positioning in strains (WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A). Cells were arrested at G1-phase with alpha factor and then released into fresh YPD medium. After 80 min, the samples were collected. The following sample treatment process is described in Materials and method section. (B) Results from a micrococcal nuclease (MNase) chromatin accessibility assay for WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains, shown on a 2% agarose gel. Chromatin sensitivity assays were performed using digestion with various concentrations of MNase for 10 min followed by quenching with stop solution and DNA extraction. MNase amount from lane 1 to lane 6: 10; 0.5; 0.25; 0.13; 0.06; 0 in Unit. High MNase mount led to smaller fragment and nucleosome bands (poly, tri, di and mononuclesomes) means strong nucleosome positioning. The lane 2 DNA was used for mononulclesome/undigested chromatin fragment calculation in (C). The Lane 1 DNA was used for sequencing library preparation and sequencing data analysis in (D) and (E). (C) Calculated relative ratio of mononulclesome/undigested chromatin fragment fraction. The ImageJ software was used to quantify the band intensity. (D) MNase-seq profiles of mean nucleosome occupancy around transcription start sites (TSS) for the WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains. The y-axis is standardized read counts. Read coverage was calculated at each nucleosome position using reads of length 149–170 bp. For each region surrounding the origins of replication (D) and transcription start sites (E), the coverage of the nucleosome was collected and standardized to a mean of 0 and unit variance. Read counts at each position were plotted and can be used to infer nucleosome positioning and relative binding. (E) MNase-seq profiles of mean nucleosome occupancy around early replication origin sites for the WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains. A complete independent repeat was provided in Supplementary Figure S6.

Figure 3.

Parental histone chaperone mutants exhibit a slight increase in DNA lesions during S-phase. (A) Representative fluorescence images of Rad52-YFP expression, which represent sites of DNA lesion repair, in WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains. The Rad52-YFP foci are marked with white arrows. (B) Average percentage of cells with Rad52-YFP foci in WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains. A one-way ANOVA analysis was used for comparing WT and each mutant. Asterisks indicate statistical significance between two strains. **P < 0.01, ***P  < 0.001, ****P  < 0.0001. Error bars depict standard error of the mean. Over 500 cells were counted.

Figure 4.

Parental histone chaperone mutants do not show any additional checkpoint kinase Rad53 activation when combining with new histone chaperones under normal growth condition. Immunoblot analysis of Rad53 phosphorylation levels, a marker of the DNA damage response in yeast, before and after treatment with hydroxyurea (HU; a replication fork block agent in different combination of dpb3Δ, mcm2-3A, dpb3Δ mcm2-3A with asf1Δ, cac1 Δ and rtt106Δ. The whole cell lysate with log-phase cells without (left panel) or with 200 mM-HU-treated (1 h) (right panel) were used for these assays. Rad53 antibody (ab104232, Abcam) was used for Rad53 detection. Hyperphosphorylated forms of Rad53 (Rad53-P) will migrate slower in gel analysis. All HU-treated samples give the Rad53-P band compared with untreated samples.

Figure 5.

Parental histone chaperone mutations increase free histone levels in the cell. (A) Procedure to monitor free histone levels inside WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains. (B) Immunoblot of H3 and H3K4me3 (a marker for chromatin-associated proteins) and PGK1 (a marker for soluble proteins) in the whole cell extract, soluble fraction, and chromatin fraction of WT, dpb3Δ, mcm2-3A, dpb3Δ mcm2-3A and rad53Δ strains. H3K4me3 (ab8580 Abcam); PGK1(ab113687 Abcam) and H3-HA (12CA5 Sigma) were used for western blot. We did not observe any obvious H3 and H3K4me3 level difference in whole cell extract. (C) Quantitation of soluble H3. The data shown in (C) comes from four independent experiments. Error bars depict standard error of the mean. The signals obtained for soluble histones were normalized to the signals obtained for soluble PGK1 on western blots. Error bars depict standard error of the mean. A one-way ANOVA analysis was used for comparing between two strains. *P  < 0.05, **P  < 0.01, ***P  < 0.001. (D) Quantitation of soluble H3K4me3. The data shown in (D) comes from three independent experiments. Error bars depict standard error of the mean. A one-way ANOVA analysis was used for comparing between two strains.

Figure 6.

Parental histone chaperone mutations decrease the frequency of homologous recombination. (A) Procedure for quantifying HR efficiency in WT, dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A strains transformed with the Neo gene flanked by the 750-bp 5′ UTR and 550-bp 3′ UTR of the URA3 gene. (B) Relative HR frequency of dpb3Δ, mcm2-3A and dpb3Δ mcm2-3A mutants to the WT strain. (C) Procedure for quantifying HR efficiency using the LU system in different strains. (D) Percentage of WT, dpb3Δ, mcm2-3A, and dpb3Δ mcm2-3A cells undergoing HR, as detected using the LU system. (E) Deletion of hht2-hhf2 elevates HR rates of parental histone transfer defective mutants. (F) Deletion of hht2-hhf2 promotes the resistance of dpb3Δ mcm2-3A to bleomycin and high concentration of hydroxyurea. In this figure, a one-way ANOVA analysis was used for comparing two different mutants. *P <0.05, **P  < 0.01, ***P  < 0.001, ****P <0.0001. Error bars depict standard error of the mean in this figure.

Figure 7.

A proposed model showing that defective parental histone transfer impairs homologous recombination. In the wild type cells (left panel), parental histone H3–H4 (dark blue) can be efficiently transfer to newly synthesized DNA mediated through Polymerase ϵ subunit Dpb3/Dpb4 and Mcm2-Ctf4-Polα axis following DNA replication fork. The new histone H3–H4 (light blue) is deposited following the empty space left by the parental histone. In the dpb3Δ mcm2-3A mutant (right panel), a small fraction of parental histone H3–H4 may leave chromatin and release into free H3–H4. As a consequence, the higher free histone level may inhibit HR through inhibiting ssDNA resections and host HR factors (52).