Iterative oxidation by TET1 is required for reprogramming of imprinting control regions and patterning of mouse sperm hypomethylated regions

Abstract

Ten-eleven translocation (TET) enzymes iteratively oxidize 5-methylcytosine (5mC) to generate 5-hydroxymethylcytosine (5hmC), 5-formylcytosine, and 5-carboxylcytosine to facilitate active genome demethylation. Whether these bases are required to promote replication-coupled dilution or activate base excision repair during mammalian germline reprogramming remains unresolved due to the inability to decouple TET activities. Here, we generated two mouse lines expressing catalytically inactive TET1 (Tet1-HxD) and TET1 that stalls oxidation at 5hmC (Tet1-V). Tet1 knockout and catalytic mutant primordial germ cells (PGCs) fail to erase methylation at select imprinting control regions and promoters of meiosis-associated genes, validating the requirement for the iterative oxidation of 5mC for complete germline reprogramming. TET1^V and TET1^HxD rescue most hypermethylation of Tet1^-/- sperm, suggesting the role of TET1 beyond its oxidative capability. We additionally identify a broader class of hypermethylated regions in Tet1 mutant mouse sperm that depend on TET oxidation for reprogramming. Our study demonstrates the link between TET1-mediated germline reprogramming and sperm methylome patterning.

Keywords: DNA methylation; TET enzyme; epigenetic; germline reprogramming; imprinting; primordial germ cell establishment; reprogramming; sperm DNA methylation establishment; sperm methylome.

Figure 1.

Generation and validation of 5hmC stalling Tet1-V and catalytically inactive Tet1-HxD mouse lines. A) Schematic of WT and mutant TET1 proteins with the oxidative capabilities indicated. B) Sanger sequencing of Tet1^+/+, Tet1^V/V, and Tet1^HxD/HxD alleles. C) Western blot for full length TET1 protein in mutant testes with GAPDH as loading control. D) Expression of Tet isoforms (CPM – counts per million reads) in E14.5 XY PGCs as determined by RNAseq (n=3-4; FDR < 0.05; fold change > 1.5). E) Venn overlaps of DEGs in Tet1^−/−, Tet1^V/V, and Tet1^HxD/HxD PGCs compared to WT with enriched GO terms indicated for the commonly altered genes. F) Sparse BS-seq of Tet1^−/−, Tet1^V/V, and Tet1^HxD/HxD PGCs compared to WT at E12.5 and E14.5 (t-test vs WT, *p<0.05, ***p<0.0005). See also Figure S1, Figure S2 and Table S1.

Figure 2.

Loss of TET1 or TET1 catalytic activity leads to incomplete reprogramming of ICRs and meiosis-associated promoters. A) Methylation levels of representative ICRs as measured by pyrosequencing. Each data point is E12.5 XY PGC sample collected from one embryo. Time course analysis of methylation levels of KvDMR (B), Peg1 (C), and Peg3 (D) from E11.5 to E14.5 in XY PGCs. Theoretical passive dilution curve (maroon-dashed line) is a fitted one-phase logarithmic decay curve with PGC doubling time of 12.6 hours. Methylation levels of meiosis-associated promoters, Mael (E) and Sypc1 (F) at E13.5 and E14.5 in XY PGCs (mean methylation ± SEM; n=4-9, one-way ANOVA with Dunnett’s multiple comparisons test, distinct letters indicate statistical significance) . G) Expression of Mael and Sycp1 in E14.5 XY PGC RNAseq (n=3-4; FDR < 0.05; fold change > 1.5). See also Figure S3.

Figure 3.

Tet1^V/V and Tet1^HxD/HxD males exhibit methylation defects at ICRs that are inherited by offspring. A) Methylation levels at representative maternally methylated ICRs measured by pyrosequencing. Each data point is a sperm sample from one adult mouse. H19/Igf2 ICR is a paternally methylated ICR that exhibits full methylation in sperm (mean methylation ± SEM; n=3-5). The number of live embryos (B) and resorbed embryos (C) per litter at E10.5 (mean number of pups/litter ± SEM, n=3-5 litters). The number of live pups (D) and dead pups (E) per litter at PND0 (mean number of pups/litter ± SEM, n=5-6 litters). For panel A-E: one-way ANOVA with Dunnett’s multiple comparisons test, distinct letters indicate statistical significance. F) Heatmap representation of DNA methylation levels measured by pyrosequencing at ICRs of all E10.5 embryos from Tet1^+/+, Tet1^V/V and Tet1^HxD/HxD males. Each row is an individual embryo of the indicated paternal genotype. The same offspring are depicted by locus for Peg1 (G) and Peg3 (H) ICRs. pWT n=22 embryos (3 litters), pVV n=31 embryos (4 litters), pHxD n=37 embryos (4 litters). Fisher’s exact test for frequency of hypermethylated embryo; *p<0.05, **p<0.01; shaded bars indicate average methylation of pWT embryos ± 1 STDEV. See also Figure S4.

Figure 4.

Global methylation analysis using Mouse Infinium Methylation BeadChip shows distinct methylome defects in catalytic mutant (Tet1^V/V, Tet1^HxD/HxD) sperm compared to Tet1^−/− sperm. A) Flow chart showing differential methylation analysis of each Tet1 mutant sperm sample compared to Tet1^+/+ (WT), n=8-10. A DMR is a probe with FDR < 0.05 with minimum change in average methylation of greater than 10%. B) Venn overlap of significantly hypermethylated (top) and hypomethylated (bottom) DMRs in Tet1 mutant sperm vs. WT. Volcano plots comparing the methylation status of Tet1^−/− sperm DMRs to Tet1^V/V sperm (C) or Tet1^−/− sperm DMRs to Tet1^HxD/HxD sperm (D). KYCG analysis to identify enrichment for probe design groups of rescued (E) and not-rescued (F) DMRs. See also Figure S5, Table S2, and Table S3.

Figure 5.

Identification of TET1-dependent sperm-specific hypomethylated regions. A) Distribution of DMRs classified as related to imprinting biology in the array annotation (*p-value < 0.05; two-sided Bernoulli distribution test vs. all array probes). B) Schematic of a sperm hypomethylated region (sperm HMR) that is excluded from de novo methylation through enrichment of H3K4me3. C) Distribution of DMRs that overlap sperm HMRs (DNMTools function “hmr” of sperm WGBS GEO: GSE56697). D, F) Representative regions that are commonly hypermethylated in Tet1^−/−, Tet1^V/V, Tet1^HxD/HxD sperm, overlapping sperm HMRs (grey bars) and H3K4me3 enrichment during de novo methylation in E17.5 prospermatogonia. E) Heatmaps and metaplots of E17.5 prospermatogonia H3K4me3 enrichment centered on DMRs for each genotype or shared among three mutants, measured in counts per million (CPM). G) Genomic distribution of DMRs that overlap with H3K4me3 enrichment in E17.5 prospermatogonia. H) Methylation analysis of oxC base-dependent sperm HMRs in demethylating WT and Tet1 mutant E14.5 PGCs using targeted bisulfite sequencing (mean methylation ± SEM; n=3-4, one-way ANOVA with Dunnett’s multiple comparisons test, distinct letters indicate statistical significance). See also Figure S6 and Table S4.

Figure 6.

DMR-associated genes are expressed throughout spermatogenesis. A) Comparison of DMR-associated gene expression and all genes expressed throughout spermatogenesis based on normalized gene expression from publicly available scRNAseq (GEO: GSE112393). Fat1 (B,C) and Dyrk2 (D,E) 5’ RACE analyses of Tet1^+/+ and Tet1^−/− testes cDNA and the corresponding BLAT mapping of alternative (alt) transcripts to TET1 DMRs. “Main” indicates predicted main product and “alt” indicates identified alternative product.

Figure 7.

Summary of findings of TET1 iterative oxidative roles during germline reprogramming. TET1 generation of 5hmC and higher ordered oxidized bases 5fC/5caC are required for demethylation of select ICRs, meiosis-associated gene promoters, and some sperm HMRs, which are generated via exclusion of DNMT3 during de novo methylation in prospermatogonia. 5hmC generation by TET1^V is insufficient for these loci to achieve a hypomethylated state prior to mitotic arrest at E14.5. Due to incomplete methylation erasure during reprogramming, a subset of sperm HMRs in Tet1 mutant sperm become hypermethylated, similar to the rest of the sperm genome.