Fig 1.
Males from the ENU-induced mutant line rahu display meiotic defects.
(A) Breeding scheme used to isolate third-generation males with recessive defects in meiosis. Un-filled shapes represent animals that are wild-type for a mutation of interest, half-filled shapes are heterozygous carriers, and filled shapes are homozygotes. (B) Representative images showing the SYCP3 and γH2AX immunofluorescence patterns during meiotic prophase stages in squashed spermatocyte preparations from wild type and rahu mutants. Examples of cells with abnormal staining (nucleus-wide γH2AX along with longer tracks of SYCP3) are also shown. (C) Distribution of meiotic prophase stages in four G3 mutants obtained from the rahu line (a, b, c, d) and their phenotypically wild-type littermates (a′, b′, c′, d′). The number of SYCP3-positive spermatocytes counted from each animal is indicated.
Fig 2.
rahu and CRISPR/Cas9-targeted frameshift alleles of Gm14490 cause meiotic arrest and fail to complement each other.
(A) The ratios of testes weight to body weight for 5- to 9-week-old mice carrying the rahu and CRISPR/Cas9-targeted alleles (em). Half-filled and fully filled circles represent heterozygous and homozygous genotypes, respectively. (B) Representative PAS-stained testis sections from adult mice of the indicated genotypes. Arrows indicate post-meiotic germ cells (spermatids) and arrowheads point to spermatocytes with an apoptotic morphology (condensed and/or fragmented). (C) Representative TUNEL-stained testis sections from adult mice of the indicated genotypes. Arrowheads point to TUNEL-positive cells (stained dark brown). (D and E) Representative PAS-stained testis sections from adult mice of the indicated genotypes. Arrows indicate post-meiotic germ cells.
Fig 3.
rahu is an allele of the testis-expressed gene Gm14490.
(A) SNP genotypes of seven rahu mutants (b, c, d, e, f, g, h) obtained using the Illumina Medium Density Linkage Panel are shown on the left. The single 33.58-Mbp region of B6 SNP homozygosity shared between mutants is highlighted in yellow. A detailed view of variants within this region is shown on the right for two informative rahu mutants (c, f) and two informative phenotypically wild-type mice (i, j). The reference SNP ID numbers (rs ID) for known variants and the gene names of previously un-annotated novel variants (i.e., presumptive ENU-induced lesions; asterisks) are listed. Phenotypes were assayed as shown in Fig 1C or Fig 2A. (B) Schematic of Gm14490 (as predicted by Ensembl release 87) showing the locations of the ENU-induced lesion (red asterisk) and the gRNA used for CRISPR/Cas9-targeting. (C) Splice junctions that start or end within 5 bp of Gm14490 exon boundaries, from ENCODE long RNA-sequencing (release 3) from adult testis. The black asterisk indicates a predicted exon with no splice junctions detected within 10 bp of its exon boundaries. (D) Density of mapped ENCODE long RNA-sequencing reads (release 3) from adult testis within a window spanning from 500 bp upstream of Gm14490 to 500 bp downstream. The vertical viewing range of the displayed track is set at a minimum of 0 and maximum of 25; read densities exceeding this range are overlined in pink. (E) Gm14490 expression estimate (ENCODE relative averaged score) in adult tissues.
Fig 4.
DNMT3C is a putative DNA methyltransferase with similarity to DNMT3B.
(A) Schematics of DNMT3C and DNMT3B showing the location of conserved domains and the rahu mutation (asterisk). (B) Cladogram of Clustal Omega aligned human and mouse DNMT3 family sequences rooted with HhaI. (C) Motif IX in Clustal Omega aligned sequences showing the location of the rahu mutation (asterisk). DNMT3L proteins do not contain Motif IX. Amino acids identical to those in DNMT3C are shaded gray. Amino acid positions refer to DNMT3C. (D) Homology-based model of DNMT3C carboxy-terminal domain (cytosine methyltransferase domain and the preceding four amino acid residues) with E692 depicted in red. (E) Crystal structure of the DNMT3A carboxy-terminal domain dimer (PDB ID:2QRV), with monomers depicted in two shades of blue. The DNMT3A amino acid equivalent to the glutamic acid that is mutated in Dnmt3crahu mutants (DNMT3A E861) is shown in red.
Fig 5.
Dnmt3crahu mutants exhibit phenotypes consistent with a role in DNA methylation and transposon repression in the male germline.
(A) Quantitative RT-PCR analysis of whole-testis RNA samples from six littermates (represented as individual data points) aged 14 dpp. Asterisk represents p<0.05 and double-asterisk represents p<0.01 in one-sided Student’s t-test. (B) Immunofluorescence of retrotransposon-encoded proteins L1 ORF1p and IAP Gag in testis sections from adults and from juveniles at 14 dpp. Matched exposures are shown comparing heterozygotes with Dnmt3crahu mutants. Arrows indicate spermatogonia and arrowheads point to spermatocytes. (C) Southern blot analysis of DNA extracted from either the testes or the tails of three Dnmt3crahu mutants or a wild-type littermate at 15 dpp. DNA was digested with either the methylation-sensitive restriction enzyme HpaII, or its methylation-insensitive isoschizomer MspI. Arrowheads mark the positions expected for fully digested bands.
Fig 6.
Dnmt3crahu mutants up-regulate retrotransposons belonging to L1 and ERVK families.
(A) Volcano plot of differential RNA-seq values for various classes of retrotransposons. RNA-seq was performed on testis RNA samples from six 14-dpp animals from a single litter: three Dnmt3crahu mutants and three heterozygotes (same mice analyzed in Fig 5A). Q-value is the Benjamini-Hochberg-adjusted p-value from DESeq2. Retrotransposons with expression fold change of >2 (up or down) and q < 0.01 are depicted as large, colored circles. (B) Heatmap showing the z-score of differentially expressed retrotransposon families (with expression fold change >2 and q < 0.01) in individual Dnmt3crahu mutants (x, y, z) and their heterozygous littermates (x′, y′, z′). Labels on rows indicate the retrotransposon family, followed by superfamily, followed by class, and then in parentheses the log2 fold change of median expression in mutant versus heterozygote. Rows with greater than two-fold change in median expression (up or down) are in bold. The log2 fold changes are also provided in the bar graph at left (greater than two-fold change shown as black bars). (C) Correlation between differentially expressed retrotransposon families in 14-dpp Dnmt3crahu mutants and 20-dpp Dnmt3l mutants. The regression line is shown and r is the Pearson correlation coefficient.
Fig 7.
Retrotransposons belonging to the L1, ERVK, and ERV1 families are hypomethylated in Dnmt3crahu mutants.
(A) Proportions of differentially methylated CpGs within genomic elements, as determined by WGBS. WGBS was performed on whole-testis DNA samples from six 12-dpp animals (two Dnmt3crahu mutant and two wild-type mice from one litter, and one Dnmt3crahu mutant and one wild-type mice from a second litter). CpGs with >25% differential methylation (up or down) and with methylKit Sliding Linear Model-adjusted p-value < 0.01 were considered differentially methylated. “Remaining intergenic” refers to genomic regions that do not overlap with LINE, LTR, SINE, genic, or CpG island annotations. (B) Meta-element plot showing the average difference in methylation levels between individual Dnmt3crahu mutants and their wild-type littermates (mice u, v, u′, v′ from one litter; w, w′ from a second litter), across LINE, LTR, and SINE retrotransposons (including 5,000 bp of flanking sequence on each side) with minimum 95% coverage of the consensus sequence. (C) Heatmap showing the mean methylation levels of significantly changed retrotransposon families in individual Dnmt3crahu mutants and their wild-type littermates (p < 0.01, two-sided Student’s t-test). Labels on rows indicate the retrotransposon family and differentially expressed retrotransposon families (families with greater than two-fold increase in median expression in Dnmt3crahu mutants; rows in bold in Fig 6B) are in bold. Three differentially expressed retrotransposon families with changed methylation levels using a less stringent cutoff (p < 0.05, two-sided Student’s t-test) are included and are indicated with an asterisk. The median methylation difference is provided in the bar graph at left.
Fig 8.
Dnmt3c arose by tandem duplication of Dnmt3b in rodents.
(A) Triangular dot plot of DNA sequence identities within a 156,377-bp region encompassing Dnmt3b, Dnmt3c, and flanking genes in Mus musculus. Each dot on the plot represents 100% identity within a 20-bp window. Direct repeats appear as horizontal lines. The yellow-tinted square shows the region within the plot that compares Dnmt3b to Dnmt3c, and the blue-tinted squares within it reflect regions with identical sequences within 20-bp windows. Immediately below the plot are gene models with shaded boxes representing coding sequences. (B) Dot-plot comparison of M. musculus Dnmt3c (including 3,500 bp of flanking sequence on each side) and Dnmt3b. Each black dot on the plot represents 100% identity within a 20-bp window, and blue-tinted rectangles highlight these regions. Each orange dot represents 100% identity within a 13-bp window, and orange-tinted rectangles highlight such regions when they are exonic or lie along the diagonal axis. Gene models annotated with exons encoding conserved domains are shown schematically along the axes. (C) Dot-plot comparisons of the M. musculus 156,377-bp region shown in (A) with its homologous region in other rodents (Rattus norvegicus, Norway rat; Cricetulus griseus, Chinese hamster; Microtus ochrogaster, prairie vole; Nannospalax galili, Upper Galilee Mountains blind mole rat; Jaculus jaculus, lesser Egyptian jerboa; Dipodomys ordii, Ord's kangaroo rat; Castor canadensis, American beaver; Ictidomys tridecemlineatus, thirteen-lined ground squirrel; Cavia porcellus, domestic guinea pig), a lagomorph (Oryctolagus cuniculus, rabbit), and human (Homo sapiens). Each dot on the plot represents 100% identity within a 15-bp window. Yellow-tinted rectangles highlight M. musculus Dnmt3b and Dnmt3c, as well as Dnmt3b in rat and human. Gene models are shown for M. musculus, R. norvegicus, and H. sapiens. The putative Dnmt3c gene location in R. norvegicus is depicted by the gray dashed line above the dot plot. Segments of contiguous inter-species sequence identity between Dnmt3b and Dnmt3c appear as off-center partial diagonals (arrows) for those species that harbor the Dnmt3b and Dnmt3c pair, or as two offset diagonals (arrowheads) for species that lack the duplication. (D) Cladogram showing the evolutionary relationship of species analyzed (UCSC Genome Browser; [60]). Species that showed evidence of harboring Dnmt3c are highlighted in blue.