The histone modification reader ZCWPW1 links histone methylation to PRDM9-induced double strand break repair

The histone modification writer PRDM9 has been shown to deposit H3K4me3 and H3K36me3 at future double-strand break (DSB) sites during the very early stages of meiosis, but the reader of these marks remains unclear. Here, we demonstrate that ZCWPW1 is an H3K4me3 reader that is required for DSB repair and synapsis in mouse testes. We generated H3K4me3 reader-dead ZCWPW1 mutant mice and found that their spermatocytes were arrested at the pachytene-like stage, which phenocopies the Zcwpw1 knock–out mice. Based on various ChIP-seq and immunofluorescence analyses using several mutants, we found that ZCWPW1’s occupancy on chromatin is strongly promoted by the histone-modification activity of PRDM9. ZCWPW1 localizes to DMC1-labelled hotspots in a largely PRDM9-dependent manner, where it facilitates completion of synapsis by mediating the DSB repair process. In sum, our study demonstrates the function of ZCWPW1 that acts as part of the selection system for epigenetics-based recombination hotspots in mammals.


147
Having established that ZCWPW1 facilitates the completion of synapsis during meiosis prophase 148 I in male mice, we observed that the ZCWPW1 W247I/E292R/W294P mutant mice exhibited the same 149 synapsis defect as Zcwpw1 knockout mice, suggesting that these residues are essential for the 150 recombination-related functions of ZCWPW1. We then performed immunofluorescence staining of 151 chromosome spreads to evaluate the recruitment of DMC1 and RAD51 to single-stranded overhang 152 sequences in WT and Zcwpw1 KI/KI mice ( Figure 1C and E). There were no differences in the numbers 153 of DMC1 or RAD51 foci in the leptotene or zygotene stages of the two genotypes. However, analysis 154 of WT pachytene and Zcwpw1 KI/KI pachytene-like spermatocytes revealed an obvious discrepancy.

155
Decreased numbers of DMC1 and RAD51 foci were seen in the pachytene WT spermatocytes,    To determine whether the H3K4me3 binding ability of ZCWPW1's CW-domain is necessary for its recruitment to chromatin in vivo, we conducted an additional ZCWPW1 ChIP-seq analysis of 10 ZCWPW1 peaks were detected in the Zcwpw1 −/− or Zcwpw1 KI/KI mice ( Figure 3A and D). These in 244 vivo results, viewed alongside the previous reports of ZCWPW1 function in the meiotic process 245 demonstrating that these specific mutations in the ZCWPW1 zf-CW domain affect the protein's ability 246 to read histone modifications (including H3K4me3), together indicate that the 247 ZCWPW1 W247I/E292R/W294P mutant is an H3K4me3 reader-dead variant of ZCWPW1. Furthermore, 248 these results suggest that the H3K4me3 reader function of this protein is essential for its ability to 249 bind to chromatin and to function in meiosis prophase I in male mice.

250
ZCWPW1 also has a PWWP domain which was found in multiple other proteins to specifically 251 bind to histone H3 containing an H3K36me3 mark (Qin and Min, 2014), so we next sought to better 252 understand the overlap between ZCWPW1 peaks and H3K36me3 peaks in our ChIP-seq dataset. We

256
As with our data from whole testes, the average H3K36me3 signal of ZCWPW1 peaks obtained by

259
In analyzing the correlation between ZCWPW1 binding sites and these two histone modification 260 marks, we found that 88.8% of the ZCWPW1 peaks overlapped with regions containing both 261 H3K4me3 and H3K36me3 marks, while only 9.1% and 1.3% of ZCWPW1 peaks overlapped with 262 H3K4me3 and H3K36me3 peaks individually ( Figure 3E). Furthermore, the ZCWPW1 peak intensity 263 was significantly higher for the dual overlapping regions than for regions containing either H3K4me3        Figure 4A and B). The high overlap between ZCWPW1 and PRDM9 291 peaks further suggested that ZCWPW1 occupancy occurs in a PRDM9-dependent manner.

292
To further explore this finding of high overlap between ZCWPW1 and PRDM9 peaks in our supplement 1D and E). Allowing for differences in the binding performance of different antibodies in 306 different ChIP-seq analyses, the fact that some but certainly not all of the ZCWPW1 peaks overlapped 307 with PRDM9 peaks suggests that it is the H3K4me3 and perhaps H3K36me3 epigenetic marks 308 deposited by PRDM9, rather than the PRDM9 protein per se, that can explain the observed overlap of 309 the ZCWPW1 and PRDM9 peaks.

310
To determine whether the activity of PRDM9 is necessary for ZCWPW1 recruitment to chromatin 311 in vivo, we conducted an additional ZCWPW1 and H3K4me3 ChIP-seq analysis of testes samples

320
Having established that ZCWPW1 binding to chromatin is strongly promoted by the histone 321 modification activity of PRDM9, we next examined changes in ZCWPW1 binding sites between WT 322 and Prdm9 −/− mutant testes. We found that although 94.7% of the ZCWPW1 peaks were apparently

361
However, it bears mentioning that we also detected 781 ZCWPW1 peaks in WT testes that did not 362 obviously overlap with DSB hotspots and we detected 652 ZCWPW1 peaks that did not obviously

396
In Saccharomyces cerevisiae, Spp1-whose PHD finger domain is known to read H3K4me3

430
While we clearly show that ZCWPW1 greatly facilitates PRDM9-dependent DSB repair, we do 431 not yet have strong evidence for the precise nature of its functional role. One possibility is that 432 ZCWPW1, upon binding to PRDM9-dependent histone modification hotspots, might serve as a DSB 433 mark, which can perhaps subsequently recruit other factors involved in DSB repair. Recent studies 434 have reported that PRDM9 binds on both the cut and uncut template chromosomes to promote meiotic interact with the SC machinery by using its SCP1-like domain to tether PRDM9-bound loops to the 437 SC in order to promote homologous DSB repair.

438
In summary, our study identifies ZCWPW1 as an H3K4me3 and H3K36me3 reader that promotes

562
For the H3K4me3 and H3K36me3 antibodies, the incubated Protein A beads were washed once 563 with RIPA buffer containing 250 mM NaCl, three times with RIPA buffer containing 500 mM NaCl, 564 and once with TE buffer (10 mM Tris-HCl pH 8.0, 1mM EDTA). For the ZCWPW1 antibody, the 565 incubated Protein A beads were washed twice with RIPA buffer containing 250 mM NaCl, once with 566 RIPA buffer containing 500mM NaCl, and once with TE buffer. Next, the beads were transferred to a 567 new 0.5ml tube and incubated in 100 μL ChIP elution buffer (10mM Tris-HCl pH8.0, 5mM EDTA, 568 300mM NaCl, 0.5% SDS) containing 5 µL proteinase K (Qiagen, 20mg/ml stock) at 55°C for 2 h and 569 then at 65°C for 4 h. The eluate was transferred to a 0.5 mL tube, and the enriched DNA was purified 570 by phenol-chloroform, followed by dissolution in 50 μL TE buffer.   identify DEGs from the raw counts produced by Salmon with the two conditions: P adjust < 0.05, and Two-tailed Wilcoxon rank sum tests were performed to obtain inferential statistical significance (p

660
All data generated or analyzed during this study are included in the manuscript and supporting files.

661
The raw sequencing data produced in this study (ChIP-seq data listed in Supplemental file1) and the     unit of Y axis is average fold change as described in the method. All ChIP-seq experiments were performed using PD14 mice with n = 4 for each genotype.   All experiments were performed on adult mice (6−8weeks old) with n = 3 for each genotype.