The fork protection complex promotes parental histone recycling and epigenetic memory

Abstract

The inheritance of parental histones across the replication fork is thought to mediate epigenetic memory. Here, we reveal that fission yeast Mrc1 (CLASPIN in humans) binds H3-H4 tetramers and operates as a central coordinator of symmetric parental histone inheritance. Mrc1 mutants in a key connector domain disrupted segregation of parental histones to the lagging strand comparable to Mcm2 histone-binding mutants. Both mutants showed clonal and asymmetric loss of H3K9me-mediated gene silencing. AlphaFold predicted co-chaperoning of H3-H4 tetramers by Mrc1 and Mcm2, with the Mrc1 connector domain bridging histone and Mcm2 binding. Biochemical and functional analysis validated this model and revealed a duality in Mrc1 function: disabling histone binding in the connector domain disrupted lagging-strand recycling while another histone-binding mutation impaired leading strand recycling. We propose that Mrc1 toggles histones between the lagging and leading strand recycling pathways, in part by intra-replisome co-chaperoning, to ensure epigenetic transmission to both daughter cells.

Keywords: Claspin; DNA replication; H3K9 methylation; chromatin replication; epigenetic inheritance; epigenome maintenance; fission yeast; heterochromatin; histone chaperone; histone recycling; mouse embryonic stem cells.

Graphical abstract

Figure 1

The FPC is essential for the maintenance of heterochromatic gene silencing (A) Illustration of the FPC (Mrc1, Swi1/Tof1, Swi3/Csm3) at the replisome, showing the CMG helicase (MCM2-6, Cdc45), Mcl1/Ctf4, DNA primase (Polα), and polymerase epsilon (Polε). (B) Heterochromatic mating-type region between the IR-L and IR-R boundaries depicting the silent mating-type cassettes mat2-P and mat3-M, Kint2::YFP and (EcoRV)::mCherry fluorescent reporter genes, RNAi-dependent nucleation center cenH, Atf1 transcription factor binding sites, and REII and REIII silencing elements. (C) Establishment of mCherry silencing following clr4⁺ reintroduction in FPC mutants. (D) Cells with derepressed reporters in clonal cultures of FPC mutants (n = 6). Data are represented as mean ± SD. (E) Micrographs of FPC mutants expressing mCherry. Scale bar: 10 μm. (F) Histograms of mCherry fluorescence (mean, black; replicates n = 6, red). (G) Interpretation of loss-of-silencing events in (H). The mCherry locus asymmetrically loses heterochromatic silencing in S phase, and the protein production starts in G2 (top cell). The expressed (mCherry ON) and repressed (mCherry OFF) chromatids segregate to sister cells. In subsequent cell divisions, mCherry protein is produced in the ON lineage (blue arrows) but is diluted in the OFF lineage (yellow arrows). (H) Loss of silencing in the mrc1Δ mutant followed by time-lapse microscopy. White arrowheads point to cells experiencing a loss-of-silencing event. Blue arrowheads point to ON cell lineages and yellow arrows point to OFF lineages. Yellow arrows are omitted at 10 h for clarity. Scale bar: 10 μm.

Figure S1

Mrc1 maintains heterochromatin independently of its functions in genome stability and origin regulation, related to Figure 2 (A) H3K9me2 levels in mrc1ΔHBS mutants with an (EcoRV)::ura4⁺ reporter. (EcoRV)::ura4⁺ and (EcoRV)::mCherry differ only at their ORF. (EcoRV)::ura4⁺ was used to select for cells in the expressed state by propagating them in EMM2 medium lacking uracil. Statistics using unpaired Student’s t test. Average of three independent replicates. H3K9me2 was analyzed by ChIP-seq. The bar-diagram shows H3K9me2 levels quantified by ChIP-seq normalized to input and shown relative to signal at cnt3. (B and C) Histograms of mCherry cell fluorescence intensities used to generate Figure 2B B and Figure 2C C. (D) Histograms of mCherry cell fluorescence intensities used to generate Figure 2G and representative micrographs. Note that some hsk1-89 cells displayed autofluorescence rather than the expected nuclear signal for YFP expression. Scale bar: 10 μm. (E) Mrc1 functions redundantly with the REII element to repress mat2-P mating-type information, dependent on the HBS domain. The proportion of cells undergoing haploid meiosis is shown as in Figure 2H but for the REII⁺ control. Scale bar: 1 μm. (F) Mrc1 and HBS domain are necessary for the repression of the (XbaI)::ura4⁺ reporter gene near mat2-P, redundantly with the REII element. Ten-fold serial dilutions of cell suspensions were spotted onto the indicated media. Growth on AA-ura reflects ura4⁺ expression while growth on the toxigenic substrate FOA reflects ura4⁺ repression.

Figure 2

Mrc1 maintains heterochromatin independently of its functions in genome stability and origin regulation (A) Mrc1 protein with annotated domains and residues. MRC1-like domain refers to the most conserved portion of Mrc1, Pfam domain PF09444. (B and C) Cells expressing mCherry (n = 6). (B) [ANOVA, F = 654, p = 2.22 × 10⁻¹⁶]; (C) [ANOVA, F = 99, p = 2.13 × 10⁻⁹]. Data are represented as mean ± SD. (D) Checkpoint proficiency of mrc1 mutants. 10-fold serial dilutions of cell suspensions were spotted onto HU-containing YES plates and incubated at 33°C for 4 days. (E) BrdU incorporation profiles in the mating-type region in cells released into S phase in the presence of HU. (F and G) Cells expressing mCherry (F, n = 3; G, n = 6). (G) Cells were propagated at the permissive temperature for hsk1-89, 37°C, which inherently weakens heterochromatin seen by slight loss of mCherry silencing even in WT cells. [ANOVA, F = 135, p = 1.88 × 10⁻¹³]. Data are represented as mean ± SD. (H) Haploid meioses in mrc1Δ and mrc1ΔHBS mutants lacking the REII silencing element visualized by bright field imaging and Hoechst staining of cells propagated on MSA sporulation medium. Scale bar: 1 μm. See also Figure S1.

Figure 3

Coordinated function of Mrc1 and Mcm2-HBD in heterochromatin maintenance (A) Aligned S. cerevisiae and S. pombe Mrc1 proteins showing contacts and cross-links to replisome components in S. cerevisiae^, and highly conserved residues in Mrc1 HBS (bottom). Boxed residues were mutated in this study. Magenta residues are identified in the cryo-EM structure. (B) Cells expressing mCherry (n = 6). Data are represented as mean ± SD. (C) Expression of ade6⁺ reporter in mrc1 mutants. Red color on YE medium indicates repression. (D) Mrc1 contacts with replisome components shown in magenta on PDB 8B9A with the expected positions of the Mcm2 HBD and Mrc1 HBS and DSE indicated. Cross-linked residues (from Baretić et al. [2020]²) are indicated by magenta circles and labeled by their position in S. cerevisiae Mrc1. The conserved F848 in KAF is in green. (E) Conserved tyrosine residues in the Mcm2 HBD mutated to alanine in mcm2-2A. (F) Cells expressing mCherry (n = 18 [left] [ANOVA, F = 73.9, p = 2.2 × 10⁻¹⁶]; n = 6 [right]). Data are represented as mean ± SD. (G) H3K9me2 in subtelomeric region Tel1R in mrc1ΔHBS and mcm2-2A mutants. Centromere 1 is shown for comparison. (H) Major heterochromatic regions of S. pombe. (I) Heatmap depicting H3K9me2 at heterochromatic regions. (J and K) Tethering Pob3 to Mrc1ΔHBS through a GFP-GBP interaction (J) restores silencing of mCherry reporter (K). (n = 6) [ANOVA, F = 45.9, p = 7.1 × 10⁻⁷]. Data are represented as mean ± SD. See also Figures S2 and S3.

Figure S2

Proficient checkpoint in mrc1-DSE and mrc1-KAF mutants and similarities of mrc1ΔHBS and mcm2-2A mutants, related to Figure 3 (A) Ten-fold serial dilutions of cell suspensions were spotted on HU-containing medium to estimate checkpoint proficiency. (B) Ten-fold serial dilutions of cell suspensions were spotted on rich YES medium and incubated at the indicated temperatures to check for bypass of the hsk1-89 mutation at the restrictive temperature of 30°C. (C) Asymmetric loss of heterochromatic gene silencing of (EcoRV)::mCherry reporter in mcm2-2A mutant visualized by time-lapse microscopy. Scale bar: 10 μm. (D) Similar temperature-dependency of heterochromatic silencing in mrc1ΔHBS and mcm2-2A mutants. In both (C) and (D), microcolonies were grown from single cells under a fluorescence microscope.

Figure S3

Heterochromatin loss away from nucleation sites in Mrc1 and Mcm2 mutants, related to Figure 3 (A) H3K9me2 occupancy at Chromosome 1 (top) and Chromosome 2 (bottom) in histone-recycling mutants. siRNAs are depicted as red blocks across Tel1L/Tel2L (left), centromeres (middle), or Tel1R/Tel2R (right). (B) H3K9me2 occupancy at the mating-type region indicating siRNAs originating from cenH depicted in red and the two Atf1-binding sites in magenta. (C) Ten-fold serial dilutions of cell suspensions were spotted onto the indicated media. Growth on the toxigenic substrate FOA reflects ura4⁺ repression. (D) H3K9me2 occupancy at the h and ΔK mating-type regions. (E) Quantification of H3K9me2 occupancy at indicated regions within the mating-type region normalized to cnt3 [unpaired Student’s t test used]. (F) Schematic diagram of Pob3 with functional domains annotated and deletion mutations indicated. (G) Proportion of cells expressing mCherry in cells co-expressing Pob3-GBP (full-length and mutants) and Mrc1ΔHBS-GFP (mean ± SD, n = 6) [ANOVA, F = 41.6, P = 8.7 × 10⁻⁹].

Figure 4

Mrc1 cooperates with Mcm2 in recycling of parental histones to the lagging strand (A) Workflow of xSCAR-seq in fission yeast. (B) Partition of H3K36me3 (top) and H4K20me0 (bottom) at a genomic region. Replication origin centers are depicted as black lines with their respective firing efficiency score. (C) Average partitioning score across replication initiation centers (with score >20) for parental (H3K36me3, top) and newly synthesized (H4K20me0, bottom) histones. (D) Heatmap representing the partitioning score across all replication origin centers (score >20) for H3K36me3 (left) and H4K20me0 (right). Each row represents the partition score of xSCAR-seq sequence reads at one origin. Average of two independent replicates. See also Figure S4.

Figure S4

SCAR-seq density analysis reveals no overall loss or gain of parental and new histones in recycling mutants, related to Figure 4 (A and B) Density of parental histones (H3K36me3) and new histones (H4K20me0) on total replicated DNA (unstranded SCAR-seq/EdU Input) in ±2.5 kb bins centered on origins of replication. (C- and D) Density of parental histones (H3K36me3) and new histones (H4K20me0) on the leading and lagging strand (stranded SCAR-seq/EdU Input) in ±2.5 kb bins centered on origins of replication.

Figure 5

Mrc1 chaperones H3-H4 tetramers in a manner compatible with Mcm2 co-chaperoning (A) AF^, prediction of a complex comprising full-length S. pombe Mrc1 (pink) and Mcm2 (light blue) bound to a histone H3-H4 tetramer (gray). Histone tails and unstructured Mrc1 residues predicted with low confidence (residues 1–381, 475–710, 800–804, and 856–1,020) are not depicted for clarity. Closeups: (1) interaction of the Mrc1 KAF motif with a hydrophobic groove in Mcm2 and (2) interaction of the N-terminal alpha helix of the Mrc1 HBS domain with charged aa in histone H3 αN and α2 helices and histone H4 C terminus. (B) AF predictions of Mrc1 histone binding via the HBD, including beginning of the HBS domain (dark pink) and upstream region (710–800). Interaction between Mrc1 (pink) and the histones (gray) remains unchanged in the presence of Mcm2 (light blue). Closeups highlight residues subjected to mutational analysis. (C) Mcm2(34–165) and Mrc1(674–879) polypeptides used for pull-downs. (D) Pull-downs of full-length Mrc1-FLAG and H3-H4 with HA-Mcm2 HBD (34–165). Pulled-down proteins were detected by western blot (WB) or Coomassie staining (CBB). MW in KDa. (E) Pull-downs of H3-H4 dimers or tetramers with His₆-Mrc1(674–879). (F) Pull-downs of H3-H4 tetramers with His₆-Mrc1(674–879) WT and indicated mutants. See also Figure S5 and S6.

Figure S5

AlphaFold predicts an Mrc1-Mcm2 co-chaperone complex consistent with the Mcm2 HBD crystal structure and the Mcm2 cryo-EM structure at the replisome, related to Figure 5 (A and B) AF^, predictions of complexes comprising full-length Mrc1 and Mcm2 bound to a histone H3-H4 tetramer with Mrc1 (A) and Mcm2 (B) colored according to AF pLDDT (predicted local distance difference test) score to indicate prediction confidence. (C) AF predictions of a complex comprising full-length Mrc1 bound to an H3-H4 tetramer with Mrc1 colored according to AF pLDDT score. (D) The predicted interaction between Mcm2 (light blue) and Mrc1 (pink) mediated by Mrc1 helices (residues 829–845 and 406–437) is in agreement with Mcm2 (yellow)/Mrc1 (dark blue) positions previously experimentally determined by cryo-EM in the context of the replisome (PDB 8B9A). (E) The predicted interaction of the N-terminal domain of Mcm2 (light blue) with a histone H3-H4 tetramer is in agreement with an available crystal structure (yellow, PDB 5BNV⁵¹). (F) Amino acid sequence of the Mrc1(674–879) fragment used in in vitro pull downs with relevant domains and residues highlighted.

Figure S6

PAE plots for AF best scoring models, related to Figure 5 (A) Mrc1 with histone dimer, (B) Mrc1 with histone tetramer, and (C) Mrc1 and Mcm2 with histone tetramer. The predicted interaction interfaces of Mrc1 and Mcm2 with histones are highlighted in pink and light blue, respectively. Predicted aligned error (PAE) plots^, were visualized by ChimeraX 1.7. Tables in (D) show DockQ and QS-score validation scores for Mcm2 and Mrc1 (Figure 6C), and Mcm2 and H3/H4 histone (Figure S5C) interaction interface in Mrc1-Mcm2-H3/H4 tetramer AF model compared to selected experimental protein structures (PDB: 8B9A and 5BNV, respectively—see STAR Methods).

Figure S7

Characterization of Mrc1 histone-recycling mutants, related to Figure 6 (A) Ten-fold serial dilutions of cell suspensions were spotted on HU-containing medium to estimate checkpoint proficiency. (B) Immunoblot of Mrc1-FLAG and tubulin from whole cell extracts of S. pombe mrc1 mutants. (C) Immunoblot of Mrc1-Flag upon chromatin fractionation of S. pombe mrc1 mutants performed as in Shimmoto et al. (2009). Tubulin serves as soluble control. (D) Partition of H3K36me3 (top) and H4K20me0 (bottom) at a genomic region. Replication origin centers are depicted as black lines with their respective firing efficiency score.

Figure 6

Mrc1 coordinates parental histone inheritance within the replisome for heterochromatin maintenance (A) Overview of the Mrc1-HBD (710–800; based on Figure 5B) and the HBS (782–879) domains. Colored residues were mutated in this study. (B) Cells expressing mCherry (n = 6). WT, mrc1ΔHBS, and mrc1-KAF values are included for comparison from Figure 2B. Data are represented as mean ± SD. (C and E) Average partitioning score across replication initiation centers (with score >20) for parental (H3K36me3) histones (C) and newly synthesized (H4K20me0) histones (E). (D and F) Heatmap representing the partitioning score across all replication origin centers (score >20) for H3K36me3 (D) and H4K20me0 (F). See also Figure S7.

Figure 7

Histone recycling function of Mrc1 is conserved in mammalian cells (A) (top) Mouse CLASPIN with annotated domains (BP1 and BP2: basic patch 1 and 2, CKBD: Chk1 binding domain, AP: acidic patch) and alignment with Polα and Mcm2 HBDs. (Bottom) overview of mutation in CLASPINΔYY. (B) WB analysis of WT and CLASPIN mutant mESCs. (C) Cell-cycle distribution based on mean EdU intensity and total DAPI intensity of WT and CLASPIN mutant mESCs. (D) High-content microscopy of mean EdU intensity in WT and CLASPIN mutant mESCs. (E) Average SCAR-seq profile of H3K27me3 and H4K20me0 in WT and CLASPIN mutant mESCs. (C–E) n = 2 biological replicates. (F) Model: Mrc1 acts as a central coordinator of histone-based inheritance though its ability to bind and transfer H3-H4 tetramers to both leading and lagging strands, with the latter involving joint histone binding with Mcm2 to facilitate transfer to Polα and the lagging strand.