RNA helicase D1PAS1 resolves R-loops and forms a complex for mouse pachytene piRNA biogenesis required for male fertility

Abstract

During meiosis, RNA polymerase II transcribes pachytene piRNA precursors with unusually long and unspliced transcripts from discrete autosomal loci in the mouse genome. Despite the importance of piRNA for male fertility and a well-defined maturation process, the transcriptional machinery remains poorly understood. Here, we document that D1PAS1, an ATP-dependent RNA helicase, is critical for pachytene piRNA expression from multiple genomic loci and subsequent translocation into the cytoplasm to ensure mature piRNA biogenesis. Depletion of D1PAS1 in gene-edited mice results in the accumulation of R-loops in pachytene spermatocytes, leading to DNA-damage-induced apoptosis, disruption of piRNA biogenesis, spermatogenic arrest, and male infertility. Transcriptome, genome-wide R-loop profiling, and proteomic analyses document that D1PAS1 regulates pachytene piRNA transcript elongation and termination. D1PAS1 subsequently forms a complex with nuclear export components to ensure pachytene piRNA precursor translocation from the nucleus to the cytoplasm for processing into small non-coding RNAs. Thus, our study defines D1PAS1 as a specific transcription activator that promotes R-loop unwinding and is a critical factor in pachytene piRNA biogenesis.

Graphical Abstract

Figure 1.

BTBD18 interacts with D1PAS1 in pachytene spermatocytes nucleus via the BTB domain independent of RNA. (A) Structure of BTBD18 and GST-BTB. Blue, corresponds to the BTB (aka POZ) domain (top). Silver-stained SDS-PAGE of GST pull-down using GST-BTB with testicular cell nuclear extracts and GST control (bottom). Indicated bands were identified by mass spectrometry. Input, mouse testicular cell nuclear extracts; G, GST or GST-BTB; E, eluate; B, beads. (B) D1PAS1 binds to BTBD18. Co-immunoprecipitation assay of BTBD18-mCherry and D1PAS1-FLAG interaction in co-transfected HEK293T cells. The anti-TUBA (α-tubulin) antibody was used as a load control. (C) D1PAS1 interacts with BTBD18 in an RNA-independent manner. Co-immunoprecipitation assay of BTBD18-mCherry and D1PAS1-FLAG interaction in co-transfected HEK293T cells untreated (left) or treated with RNase A (right). WCE, whole cell extracts; *, non-specific band. (D) Stage-specific D1PAS1 expression during spermatogenesis was examined by immunoblot (top). Protein samples from testicular cells (TC), testicular sperm (TS), and mature sperm (S) were immunoblotted with anti-D1PAS1 antibody (bottom). ACTB (β-Actin) antibody was used as a load control. *, non-specific band. (E) Pachytene spermatocyte nuclear co-localization of BTBD18^FLAG/mCherry and D1PAS1 in cross-sections of mouse seminiferous tubules from 8 wk/old Btbd18^FLAG/mCherry males after immunostaining with antibodies to mCherry (BTBD18) and D1PAS1. SG, spermatogonia; PS, pachytene spermatocytes; RS, round spermatids; scale bar, 50 μm. (F) Nuclear co-localization of BTBD18 and D1PAS1 in meiotic chromosome spreads of pachytene spermatocytes from 4 weeks old Btbd18^FLAG/mCherry mice. The white frame indicates an enlargement of the boxed region. (G) Generation of D1Pas1^Null and 3XFLAG knock-in (D1Pas1^FLAG) mice using CRISPR/Cas9. The original sequence is black, and the 3XFLAG sequence is blue. Red and blue arrows indicate the forward and reverse primers used for KO and KI mice genotyping, respectively. (H) Average litter size of D1Pas1^Null and littermate control (+/+) males (squares, n = 6). Each square represents the average litter size of one female. The box indicates the median ± interquartile range, the whiskers indicate the highest/lowest value and the midlines are median values. (I) Testes weight of adult testes (left), macroscopic appearance (middle, scale bar, 2 mm), and histological sections from 12 weeks old control (WT) and mutant (Null) mice testes (right) were stained with periodic acid-Schiff (PAS) and hematoxylin (H) (n = 3). Pl, preleptotene spermatocytes; PS, pachytene spermatocytes; RS, round spermatids; ES, elongating spermatids; scale bar, 50 μm.

Figure 2.

ATP-dependent RNA helicase, R-loop resolving activity of D1PAS1 is critical for genome integrity. (A) BTBD18 and D1PAS1 protein abundance in adult testes from control (D1Pas1^WT), D1Pas1^Null and Btbd18^Null mice. *, Non-specific band. (B) BTBD18^FLAG/mCherry and D1PAS1 abundance in adult testes from control (D1Pas1^WT), Btbd18^FLAG/mCherry; D1Pas1^WT, and Btbd18^FLAG/mCherry; D1Pas1^Nullmice. The anti-ACTB (β-Actin) antibody was used as a load control. *, Non-specific band. (C) Macroscopic appearance and histological sections of testes from 12 wk/old Btbd18^FLAG/mCherry; D1Pas1^WT (left) and Btbd18^FLAG/mCherry; D1Pas1^Null mice (right). Scale bar, 2 mm. Histological sections were stained with periodic acid-Schiff (PAS) and hematoxylin (H) (n = 3). PS, pachytene spermatocytes; RS, round spermatids; ES, elongating spermatids; scale bar, 100 μm. (D) Structure of D1PAS1. Blue, DEAD box domain. Green, helicase C domain (top). Coomassie Blue-stained gel of recombinant D1PAS1-6XHis (bottom). (E) Representative image of helicase reactions performed by incubating duplex RNA (dsRNA) substrate with increasing protein concentrations (top) and increasing time (0–60 min) at 37 °C (bottom). (F) R-loop unwinding assay in the presence of increasing D1PAS1. A representative image of R-loop unwinding activity assay by incubating increasing D1PAS1 protein with biotinylated RNA hybrids of RNA:DNA (top). The bar graph (bottom) depicts quantification. The average is expressed as a percentage unwinding (n = 3). (G) The depletion of D1PAS1 resulted in accumulated R-loops in D1Pas1^Null pachytene spermatocytes (PS). A representative immunohistochemistry (IHC) analysis image with anti-S9.6 antibody. PS, pachytene spermatocytes; RS, round spermatid; ES, elongating spermatid; n = 3; scale bar, 50 μm. (H) Representative IHC with anti-S9.6 and TUNEL assay images of testis sections from 12 weeks old control (WT) and mutant (Null) mice (left, n = 3). Enlarged insets (bottom) correspond to the dashed box. PS, pachytene spermatocytes; scale bar, 50 μm. Quantifying R-loop-induced DNA damage-associated apoptosis at the pachytene stage in control (WT) and mutant (null) mice (right). The box indicates the median interquartile range, the whiskers indicate the highest/lowest values, and the midlines are median values: n = 3, ***P < 0.001 and ****P < 0.0001.

Figure 3.

The abundance of mRNA and piRNA transcripts in D1Pas1^Null mice. (A) Timeline for the first post-natal wave of mouse spermatogenesis. P, post-natal day. (B) MA plot of mRNA and piRNA transcripts in control (WT) compared to mutant (Null) determined by RNA-seq at P18. The y-axis is the log₂-fold change, and the x-axis is the average expression in both genotypes. Each point represents a transcript. Transcripts with P-adjusted values <0.05 are colored in green (mRNAs) and red (piRNA precursors). (C) Enriched gene ontology of differentially expressed genes after comparison of control (WT) and mutant (Null) mice. See Supplementary Table S2. (D) Heat map of differentially expressed mRNA and piRNA precursors in P18 testes. The most abundant down-regulated transcript clusters (blue) in the mutant (Null) mice are pachytene piRNA precursors. (E) Heat map of piRNA levels in control (WT) and mutant (Null) testes at P18. (F) The percentage of annotated 214 A-MYB-regulated piRNAs and 64 downregulated piRNAs in D1Pas1^Null testes at P18. Numbers represent annotated piRNAs in each group. See Supplementary Table S1. (G) Venn diagram depicting the overlap of downregulated pachytene piRNAs between Btbd18^Null and D1Pas1^Null of RNA-seq data at P18.

Figure 4.

Genome-wide profiling of D1PAS1-dependent R-loops. (A) Schematic of the BisMapR workflow and experimental outline. R-loops were detected by modified MNase and subjected to non-denaturing bisulfite conversion. Bisulfite-converted products were processed for second-strand synthesis and strand-specifically was aligned based on sequence (top). Coomassie Blue-stained gel containing purified recombinant GST-RHΔ-MNase protein for the BisMapR (bottom). (B) Correlation plots of normalized read densities between forward and reverse strands in BisMapR data from each genotype. Transcription starts site (TSS) regions with high specificity toward forward (teal, 669 genes for D1Pas1^FLAG and 1140 for D1Pas1^Null) or reverse (orange, 662 genes for D1Pas1^FLAG and 1223 for D1Pas1^Null). Forward and reverse strand signals in BisMapR were defined as a log₂ ratio of at least 1.5 in either direction between forward or reverse read densities for all the high-specificity genes. r, Pearson correlation coefficients between forward and reverse read densities. (C) BisMapR signals in each genotype and overlapping R-loop peaks. (D) Distribution of aggregated R-loop peak numbers around TSS for D1Pas1^FLAG (left) and 1223 for D1Pas1^Null (right). (E) Genomic element distribution of R-loop peaks comparing piRNAs and other genes for D1Pas1^FLAG (top 3 bars of elements) and D1Pas1^Null (bottom 3 bars of elements). (F) Genome browser view of the selected pachytene piRNAs peaks showing BisMapR signal (reads per million, RPM) separated into forward (teal) and reverse (orange) strands. D1Pas1^Null RNA-seq data at P18 is shown in red, and control (D1Pas1^WT) data is in blue. Detected R-loop peaks in D1Pas1^Null, indicated by red bars.

Figure 5.

D1PAS1 forms a complex with transcription elongation and nuclear export factors. (A) Identification of D1PAS1 interactome in mouse testicular cells. LC-MS/MS analyzed proteins immunoprecipitated by anti-D1PAS1 antibody. (B) Interaction between D1PAS1 and binding molecules in mouse testis. Immunoprecipitants with an anti-D1PAS1 antibody were analyzed by immunoblotting with anti-SRSF1, anti-PRMT1, and anti-THOC2 antibodies. IgG was used as a negative control. IP, immunoprecipitation; IgG, immunoglobulin G from normal rabbit serum. *, Non-specific band. (C) The reciprocal interaction between D1PAS1 and SRSF1. Anti-SRSF1 immunoprecipitants were analyzed by anti-PRMT1 and anti-THOC2 antibodies. (D) RNase A treatment deprecates the D1PAS1 and THOC2 interaction. D1PAS1 interactome was immunoprecipitated using lysates from mouse testicular cells untreated (left) or treated with RNase A (right). *, Non-specific band. (E) Immunoblot analysis of mouse testis lysates from control (WT) and mutant (Null) mice. Anti-ACTB (β-actin) antibody was used as a load control.

Figure 6.

D1PAS1 directly targets pachytene piRNA primary transcript. (A) Schematic of enhanced CLIP (eCLIP) workflow. See the Methods for details. (B) Immunoblot validation of D1PAS1-RNA complexes prepared by eCLIP in adult D1PAS1^FLAG testes (left). Odyssey CLx infrared scan of 3′ RNA-IR800 adapter-ligated D1PAS1^FLAG-RNA complexes (right). IP, immunoprecipitation; C, UV-cross-linked; NC, non-cross-linked; Red arrow, immunoprecipitated D1PAS1^FLAG; Red bar, D1PAS1^FLAG-RNA complexes. (C) Genome-wide overlap of D1PAS1 target transcripts from three biological replicates of D1PAS1^FLAG (≥8-fold-enriched, P ≤ 10⁻⁵ above SMInput). (D) Distribution of D1PAS1^FLAG peaks in mRNA transcript regions. CDS, coding sequence. (E) UCSC Genome browser screenshot of BisMapR-seq signals, eCLIP signals, and piRNA profiles at the genomic region of pi7 (7-qD2-24830 and 7-qD2-11976, left) and 4-qD2-2182 (right), eCLIP signals are displayed as RPM (reads per million) values, and D1PAS1 binding sites are indicated by green vertical bars. D1Pas1^Null RNA-seq data at P18 is shown in red, and control (D1Pas1^WT) data is in blue. Detected R-loop peaks in D1Pas1^Null, indicated by red bars.

Figure 7.

D1PAS1 is required for pachytene piRNA biogenesis. (A) Genomic annotation of total piRNA from P18 control (WT) and mutant (null) testes. piRNA clusters: 214 piRNA clusters previously defined (9,10). The reads were aligned to 5 sets of sequences sequentially: 214 piRNA clusters, coding RNAs (RefSeq coding gene mRNAs), non-coding RNAs (RefSeq non-coding gene mRNAs), repeats (LINE, SINE, LTR, DNA, low_complexity, satellite, Simple_repeat) and intron (genic regions for RefSeq genes). Sequence reads not mapping to the above 5 sets of sequences were classified as other. (B) Relative abundance of total piRNA annotation from P18 control and mutant testes normalized by miRNA. Non-cluster includes five categories in (A) and is defined as all reads not mapping to the 214 piRNA clusters. (C) Size distribution of small RNA libraries from P18 control and mutant (Null) testes. Data normalized by miRNA reads (21–23 nt). (D) Nucleotide composition of the first nucleotide in control and mutant testes. The piRNAs in D1Pas1^Null exhibited a 5′ end U bias at position 1. (E) Scatter plot compares the change in primary transcript and piRNA abundance for P18 D1Pas1^Null compared to D1Pas1^WT testes. Grey: unaltered transcript and piRNA abundance in mutant mice; red: piRNA clusters whose transcript and piRNA abundance were altered in mutant (Null) mice; black: pachytene piRNA-producing genes in (F and Supplementary Figure S7). See Supplementary Table S5. (F) UCSC genome browser views of pachytene piRNA-producing genes, 8-qE1-3748 (left) and 12-qE-23911 (right). BisMapR signal separated into forward (teal) and reverse (orange) strands. Red bar is the detected R-loop signal. eCLIP signals are displayed as RPM (reads per million) values, and D1PAS1 binding sites are indicated by the green vertical bar. D1Pas1^Null RNA-seq data at P18 is shown in red, and control (D1Pas1^WT) data is in blue. P18 D1Pas1^Null small RNA-seq data is shown in pink, and control (D1Pas1^WT) data is in blue. (G) Schematic model illustrates dual function of D1PAS1 in pachytene piRNA biogenesis. Upon its expression and co-localization in mid and late-pachytene spermatocytes nucleus with BTBD18, D1PAS1 regulates BTBD18-dependent and -independent pachytene piRNA gene transcription elongation. D1PAS1 is involved in regulating pachytene piRNA-producing gene transcription by unwinding R-loops in the gene body or distal regions. D1PAS1 plays a role in SRSF1 nuclear retention by PRMT1-mediated SRSF1 arginine methylation during meiosis and forms a complex with THOC2 in an RNA-dependent manner. This latter process ensures the nuclear export of pachytene piRNA primary transcript to inter-mitochondrial cement (IMC) in late pachytene and diplotene spermatocytes.