C19ORF84 connects piRNA and DNA methylation machineries to defend the mammalian germ line

Abstract

In the male mouse germ line, PIWI-interacting RNAs (piRNAs), bound by the PIWI protein MIWI2 (PIWIL4), guide DNA methylation of young active transposons through SPOCD1. However, the underlying mechanisms of SPOCD1-mediated piRNA-directed transposon methylation and whether this pathway functions to protect the human germ line remain unknown. We identified loss-of-function variants in human SPOCD1 that cause defective transposon silencing and male infertility. Through the analysis of these pathogenic alleles, we discovered that the uncharacterized protein C19ORF84 interacts with SPOCD1. DNMT3C, the DNA methyltransferase responsible for transposon methylation, associates with SPOCD1 and C19ORF84 in fetal gonocytes. Furthermore, C19ORF84 is essential for piRNA-directed DNA methylation and male mouse fertility. Finally, C19ORF84 mediates the in vivo association of SPOCD1 with the de novo methylation machinery. In summary, we have discovered a conserved role for the human piRNA pathway in transposon silencing and C19ORF84, an uncharacterized protein essential for orchestrating piRNA-directed DNA methylation.

Keywords: DNA methylation; DNMT3C; PIWI; germline; human male fertility; piRNA; transposon.

Figure 1.. SPOCD1 is essential for human fertility and transposon silencing.

A, Table of clinical data for three identified infertile men carrying homozygous SPOCD1 variants. FSH, follicle stimulating hormone, LH, luteinizing hormone, T, testosterone, TV, testicular volume right/left, NA: not available. Reference values: FSH 1–7 IU/L, LH 2–10 IU/L, T > 12 nmol/L, TV > 12 mL. B, Schematic representation of human SPOCD1 variants identified in the three infertile men. C, Pedigree of individual M3457 and sequencing trace of SPOCD1 exon 15 harbouring the SPOCD1^c.2912T>G variation with substituted nucleotide highlighted. NOA, non-obstructive azoospermia. D, Pedigree of individual GEMINI-88 and sequencing trace of SPOCD1 exon 16 harbouring the SPOCD1^{c.3354_3355insA} variation with inserted nucleotide highlighted from GEMINI-88. E, PAS and haematoxylin-stained testis sections of GEMINI-88 and control individual. SCO, Sertoli cell only. Scale bars, 50 μm. F, Pedigree of individual M2021 and sequencing trace of part of SPOCD1 exon 8 harbouring the SPOCD1^{c.1991_1992del} variation from M2021 and brother; missing nucleotides are indicated. WT, canonical SPOCD1 allele, MUT, SPOCD1^{c.1991_1992del} allele. G, PAS and haematoxylin-stained testis sections of M2021. SPG, spermatogonia, SPC, spermatocytes, RS, round spermatids, MeiA, meiotic arrest, SCO, Sertoli cell only. Scale bars, 5 and 50 μm. G, Testis section of M2021 and control individual stained for LINE1 ORF1p (brown DAB stain). SPG, spermatogonia, Scale bars, 50 and 5 μm. See also Figure S1.

Figure 2.. Crystal structure of SPOCD1 SPOC domain.

A, Multiple sequence alignment of the SPOCD1 SPOC domains from indicated species. Sequence identity conservation is denoted by depth of colour. Red box highlights homologous position of human L971. B, Crystal structure of the mouse SPOCD1 SPOC domain, displayed from left to right as cartoon view indicating C- and N-terminus, surface conservation and surface charge. C, Surface charge of PHF3 SPOC domain with RNA polymerase II CTD diheptapeptide phosphorylated on Ser2 (2xS2PCTD). PDB: 6IC8. Structures in (B) and (C) are shown in the same orientation. D, SPOCD1 SPOC domain shown in alternative orientation with residue L792 highlighted in red in the cartoon ribbon display and circled in the surface conservation and charge displays. Below is a zoom in on region around L792 with interacting residues highlighted in grey and distances between residues indicated. A similar view shows modelling of the L792R mutation, indicating clashes with the surrounding residues. E, Western blot of HEK 293T cell expression of indicated full-length human and mouse SPOCD1. See also Figure S2.

Figure 3.. SPOCD1 interacts with C19ORF84 via a carboxy-terminal helix.

A, Representative western blot protein expression analysis of n = 3 293T cell lysates after transfection with human SPOCD1-HA and SPOCD1-Q1119fs-HA. B, Conservation analysis and structure disorder prediction for human SPOCD1 is displayed together with protein schematic showing the region homologous to the amino acids mutated in human SPOCD1-Q1119fs. C, AlphaFold2 structure prediction of human SPOCD1 (Q6ZMY3) is shown and indicated domains are highlighted. D, Gene expression analysis from CAGE data of whole testes from mice of the indicated age is shown. E-J, Representative western blot analyses of n = 3 immunoprecipitations of the indicated mouse (E-H) or human (I-J) proteins expressed in 293T cells. *, Ig light chain eluting from beads. K, Volcano plot showing enrichment (log₂(mean LFQ ratio of n = 3 anti-C19ORF84 immunoprecipitates / immunoprecipitates of n = 3 rabbit serum controls) and statistical confidence (log₁₀(P-value of two-sided Student’s t-test)) of proteins co-purifying with C19ORF84 from E16.5 testes. Previously identified SPOCD1-associated proteins more than 4-fold enriched highlighted in blue. L, Representative Coomassie gel image of n = 3 co-precipitation experiments with indicated recombinant mouse C19ORF84 and SPOCD1 fragments. M, Analytical size exclusion chromatography of the C19ORF84 (blue), SPOCD1 (orange) and the complex (green). The top panel shows superposed representative chromatograms (n = 2) for each sample, with void (black; 7.2 ml) and peak elution volumes indicated. Below, separate Coomassie gels of each run are shown; with samples from the same set of fractions are used on each gel. Dashed line indicates the lowest volume that SPOCD1 elutes in on its own. See also Figure S3 and Table S1.

Figure 4.. C19ORF84 is a gonocyte-specific protein that associates with nuclear piRNA pathway and de novo DNA methylation factors.

A, C19ORF84 localisation in male germ cells at E16.5 with zoom-in of indicated gonocyte shown in insert. B, C19ORF84 protein expression in gonocytes for the indicated developmental stages. C, C19ORF84 (green) co-localisation with HA-SPOCD1 (red) in Spocd1^HA/+ foetal testes shown for the indicated time points. D, Left plot shows percentage of germ cells at E18.5 with at least one C19ORF84 focus. Right plot datapoints show number of foci in foci-containing cells with median foci count displayed as green bar. Data are mean and SD on left plot and median and datapoints on right for n = 3 biological replicates with datapoints offset and shaded differently for each replicate in right plot. E, Percentage of C19ORF84 foci overlapping with HA-SPOCD1 foci (n = 31 foci). F-K, C19ORF84 (green) co-localisation in gonocytes at E18.5 with in red (F) HA-MIWI2, (H) HA-DNMT3L and (J) HA-DNMT3C and quantification of C19ORF84 foci overlap with (G) HA-MIWI2 (n = 58 foci), (I) HA-DNMT3L (n = 58 foci) and (K) HA-DNMT3C (n = 27 foci) foci. White arrows denote localisation of C19ORF84 foci. Images in (A, B, C, F, H, J) are representative for n = 3 biological replicates of the indicated genotype or timepoint. Scale bars are 10 μm (A) and 2 μm (A (insert), B, C, F, H, J). L, Representative anti-HA western blot from Dnmt3c^HA/+; Dnmt3l^HA/+ and wildtype E16.5 testes lysates. Quantification of relative expression of DNMT3L to DNMT3C shown to the right. Data is mean and s.e.m of n = 6 biological replicates. M, Volcano plot showing enrichment (log₂(mean LFQ ratio of anti-HA immunoprecipitates from Dnmt3l^HA/+/wildtype)) and statistical confidence (log₁₀(P-value of two-sided Student’s t-test)) of proteins co-purifying with HA-DNMT3L from E16.5 foetal testes (n = 3). N, Volcano plot as in (M) showing proteins co-purifying with HA-DNMT3C in anti-HA IP-MS from E16.5 Dnmt3c^HA/+ foetal testes (n = 3). Proteins of interest highlighted in blue in (M) and (N). See also Figures S4–S5 and Tables S2–S3.

Figure 5.. C19ORF84 is required for male fertility and transposon silencing.

A, Western blot analysis of C19ORF84 protein abundance in E16.5 testes from C19orf84^+/− and C19orf84^−/− mice. B, C19ORF84 staining of E16.5 foetal testis sections from C19orf84^+/− and C19orf84^−/− mice. C, Number of embryos per plug fathered by studs with the indicated genotype mated to wildtype females. Data are mean and s.e.m. from n = 3 C19orf84^+/− (10 plugs total) and n = 4 C19orf84^−/− studs (15 plugs total), ***P<0.001. D, Testis weight of adult mice with the indicated genotype. Data are mean and s.e.m. from n = 6 C19orf84^+/− and n = 4 C19orf84^−/− mice, ***P<0.001. Insert shows a representative image of testes from C19orf84^+/− (left) and C19orf84^−/− (right) mice. E-F, Representative images of PAS and haematoxylin-stained epididymis (E) and testes (F) sections of two stages of the seminiferous cycle, indicating a germ cell arrest at the early pachytene stage. VII, stage 7, PL, pre-leptotene, P, pachytene, RS, round spermatids, eS(16), step 16 elongated spermatids, SC, Sertoli cell, Z, zygotene, m2, secondary meiocytes. G-H, Adult testes from wildtype and C19orf84^−/− mice stained for LINE1 ORF1p (G) or IAP GAG (H) protein. I, RNA-seq heat map analysis showing fold-change of expression (relative to wildtype) of the 10 most upregulated LINE1 and ERV transposons in P20 testes from C19orf84^−/−, Spocd1^−/− and Miwi2^−/− mice. n = 3 for all genotypes. J-K, Adult testis sections stained for DNA damage marker γH2AX (J) or for apoptotic cells by TUNEL assay (K) from wildtype and C19orf84^−/− mice. Scale bars, 5 μm (B, F), 20 μm (E), 50 μm (G, H, J, K). Germ cell nuclei are indicated with a dashed line in (B). Images shown in (B, E-F, G-H, J-K) are representative of data from n = 3 mice per genotype. See also Figure S6.

Figure 6.. C19ORF84 is essential for piRNA-directed DNA methylation.

A-E, Analysis of whole genome CpG methylation of P14 spermatogonia from n = 3 C19orf84^−/−, and wildtype (A-E) as well as Spocd1^−/− and Miwi2^−/− (C, E) mice. A, B, Percentage of CpG methylation levels for the indicated genomic features and transposon families (genic, promoter, intergenic and CpG islands (CGI) being defined as non-overlapping with transposons) shown as box plots. The horizontal line represents the median, boxes the 25^th to 75^th interquartile range, and dots datapoints outside that range. C, Metaplots of mean CpG methylation over the consensus sequence for the indicated LINE1 and IAP families and adjacent 2 kb. D, Correlation analysis for individual elements of the specified LINE1 and IAP families of mean CpG methylation loss in C19orf84^−/− spermatogonia (relative to wildtype) in relation to the element’s sequence divergence from the consensus. E, Mean CpG methylation of imprinted loci is presented as a heatmap. The imprinted control region (ICR) of Rasgrf1 is shown in detail on the right.

Figure 7.. C19ORF84 connects the piRNA and DNA methylation machineries in vivo.

A-C, piRNA analysis of small RNA-seq data from n = 3 C19orf84^+/− and C19orf84^−/− E16.5 testes, showing nucleotide (nt) length distribution of small RNAs in (A) (no significant differences were observed in Bonferroni-adjusted Student’s t-test – P≈0.99), annotation of piRNA targets from merged replicates in (B) and piRNA ping-pong analysis presented as relative frequency of the distance between 5’ ends of complementary piRNAs mapping to the LINE1 L1Md_T consensus in (C). D, MIWI2 staining of foetal testis sections from n = 3 wildtype and C19orf84^−/− E16.5 mice. Germ cell nuclei are indicated with dashed line. Scale bars, 2 μm. E, HA-SPOCD1 staining of foetal testis sections from n = 3 Spocd1^HA/+; C19orf84^+/− and Spocd1^HA/+; C19orf84^−/− mice. Scale bars, 2 μm. F-G, E18.5 C19orf84^+/+ and C19orf84^−/− gonocytes stained for SPOCD1 (F) and quantification of germ cell percentage presenting with at least one SPOCD1 foci (G), **P≌0.0035. H-I, E18.5 C19orf84^+/+; Dnmt3c^HA/HA and C19orf84^−/−; Dnmt3c^HA/HA gonocytes stained for HA-DNMT3C (H) and quantification of germ cell percentage presenting with at least one HA-DNMT3C foci (I), *P≌0.022. Representative images of gonocytes in (F, H) and data (mean and s.e.m) in (G, I) from n = 3 biological replicates per genotype. Scale bars, 2 μm (F, H). J, Volcano plot showing enrichment (log₂(mean LFQ ratio of SPOCD1-HA immunoprecipitates from n = 2 Spocd1^HA/+; C19orf84^+/− / immunoprecipitates from n = 3 Spocd1^HA/+; C19orf84^−/− foetal testes) and statistical confidence (log₁₀(P-value of two-sided Student’s t-test)) of proteins co-purifying with HA-SPOCD1 from E16.5 testes. Previously identified SPOCD1-associated proteins highlighted in blue. See also Figure S7 and Table S4.