An ultra-conserved poison exon in the Tra2b gene encoding a splicing activator is essential for male fertility and meiotic cell division

Abstract

The cellular concentrations of splicing factors (SFs) are critical for controlling alternative splicing. Most serine and arginine-enriched (SR) protein SFs regulate their own concentration via a homeostatic feedback mechanism that involves regulation of inclusion of non-coding 'poison exons' (PEs) that target transcripts for nonsense-mediated decay. The importance of SR protein PE splicing during animal development is largely unknown despite PE ultra-conservation across animal genomes. To address this, we used mouse genetics to disrupt an ultra-conserved PE in the Tra2b gene encoding the SR protein Tra2β. Focussing on germ cell development, we found that Tra2b PE deletion causes azoospermia due to catastrophic cell death during meiotic prophase. Failure to proceed through meiosis was associated with increased Tra2b expression sufficient to drive aberrant Tra2β protein hyper-responsive splice patterns. Although critical for meiotic prophase, Tra2b PE deletion spared earlier mitotically active germ cells, even though these still required Tra2b gene function. Our data indicate that PE splicing control prevents the accumulation of toxic levels of Tra2β protein that are incompatible with meiotic prophase. This unexpected connection with male fertility helps explain Tra2b PE ultra-conservation and indicates the importance of evaluating PE function in animal models.

Keywords: Alternative Splicing; Fertility; Poison Exon; Spermatogenesis; Ultraconserved Genome Sequence.

Figure 1. The Tra2b PE has an essential role in meiosis required for male fertility.

(A) Screenshot from the UCSC genome browser (http://genome.ucsc.edu) (Karolchik et al, 2014) showing the mouse Tra2b gene, with PhyloP representation of base-wise vertebrate genome conservation, and the positions of the LoxP sites used to make the conditional Tra2b-cPEko mouse line. (B) Schematic of germ cell development showing the cell types between primordial germ cells and sperm. The timeframe of cell type appearance in embryonic development from E15 and during the first wave of neonatal spermatogenesis is shown above (postnatal days P1–P35). (C) Anatomic levels of inclusion of mouse SR protein gene PEs. Percentage Splicing Inclusion (PSI) levels were monitored using RT-PCR amplification between primers in exons flanking the PE for each gene. Mean data are shown from n = 5 adult mice (8 week old, tissues from 3 females and 2 males plus one extra testis sample). (D) Testis morphologies of wild type and Tra2b-cPEko 19-week-old adult testes (scale shown in millimetres). (E) Testis:body weight ratios of different genotype adult mice (age range 9–25 weeks). Sample sizes n = 6 testes from Tra2b-cPEko, n = 11 testes from Tra2b-cPEhet mice and n = 8 testes from wild type mice). The mean is shown as a horizontal black line, and P values were calculated using a t test. (F) Litter sizes of Tra2b-cPEko males and wild type males after crossing with wild type female mice. Individual litter sizes from wild type mice are shown as black dots, and the mean as a horizontal line. No litters were obtained from Tra2b-cPEko mice (red dot). 3 breeding cages of each cross were maintained until each of the wild type cages had produced a litter (male mice mated aged 7–11 weeks). (G) Micrograph of wild type adult testis section stained with periodic acid Schiff (PAS). Slides analysed from n = 1 wild type testes. Black arrows point to examples of spermatogonia (abbreviated Spg), spermatocytes (abbreviated Spc) and round spermatids (abbreviated Rtd). Scale bar = 20 μm. (H, I) Micrographs of PAS-stained adult Tra2b-cPEko testis sections. Panel (H) shows 3 stage IV arrested tubules: red arrows point to apoptotic pachytene cells detected by PAS staining. Panel (I) shows rare tubule containing late surviving pachytene cells and some round spermatids. Scale bar = 20 μm. Abbreviations for cell types as in (B). Slides analysed from n = 2 Tra2b-cPEko testes. Scale bar = 20 μm. (J) Sperm counts from wild type and Tra2b-cPEko mice. Sample size n = 2 Tra2b-cPEko; n = 3 Tra2b-cPEhet mice; n = 5 wild type mice (age range 18–25 weeks). Graph shows mean +/− SD. Statistical significances were measured using unpaired t tests (Graphpad prism). Source data are available online for this figure.

Figure 2. Tra2β protein expression levels increase during the developmental window when germ cells die in the Tra2b-cPEko mice.

(A) Indirect immunofluorescent image of mouse testis seminiferous tubule at stage 9–10 stained for Tra2β and γH2AX. Representative leptotene cells (labelled L), pachytene cells (labelled P) and elongating spermatids (labelled Spd) are arrowed. Scale bar = 50 μm (n = 1 wild type testis). (B) Quantitation of individual nuclear expression levels of Tra2β and γH2AX from the image shown in (A). Nuclear intensities from cells identified as pachytene are labelled red, and from cells identified as leptotene are labelled green. The individual pachytene and leptotene nuclei were identified visually after quantification (n = 63 leptotene nuclei, and n = 61 pachytene nuclei). (C) Comparison of Tra2β protein immunofluorescence levels in the individual pachytene and leptotene nuclei (n = 63 leptotene nuclei, and n = 61 pachytene nuclei). Plots show mean and standard deviation, and statistical significance calculated using a t test (Graphpad prism). Source data are available online for this figure.

Figure 3. Tra2b-cPEko germ cells develop cellular and molecular defects.

(A) Immunofluorescent detection of γH2AX on sections of P12 wild type and Tra2b-cPEko mouse testes (n = 4 of each genotype were analysed). Seminiferous tubules are labelled according to their cell content (L/Z: seminiferous tubules containing leptotene/zygotene cells; P: seminiferous tubules containing pachytene cells). Abbreviations: sb, sex body; esb, elongated sex body. Scale bar = 20 μm. (B) MAplot showing gene expression levels in P12 testis transcriptomes and how they change between testes from wild type and Tra2b-cPEko mice (using RNAseq data from n = 4 testes from each genotype, analysed using DESeq2). Genes with an adjusted p value of less than 0.05 (calculated by using the DESEQ2 default settings, that use a Wald test) are shown as red dots (if upregulated) or blue dots (if downregulated). Significantly downregulated genes that have been identified in mouse genetic analyses with an essential role in meiosis are shown as black dots (with further details in Fig. S2). Expression of all other genes are shown as grey dots. (C) Violin plot of Tra2β and γH2AX protein expression quantitation using indirect immunofluorescence of P12 testis sections (n = 331,667 identified and measured nuclei). Sections from 3 Tra2b-cPEko (shown in red) versus 1 wild type and 2 Tra2b-PEhet mice (shown in blue, chosen since these genotypes are both fertile) were analysed. Statistical significance was assessed using a Mann–Whitney U Test, and the median values are shown as a horizontal line. (D) Confirmation of predicted expression changes of a panel of genes between P12 wild type and Tra2b-cPEko testes using RT-qPCR (n = 4 per genotype). Pairwise tests of statistical significance were done using t tests, and the median line is shown in the scatter plot. Source data are available online for this figure.

Figure 4. Tra2b-cPEko testes display aberrant splicing patterns of Tra2β target exons.

(A) Volcano plot showing results of differential splice isoforms detected using Leafcutter (Li et al, 2018). Splice isoform changes with a p.adjust equal or less than 0.05 and a ΔPSI of greater than 0.1 are shown in either red (upregulated) or blue (downregulated), with remaining splice isoforms shown in grey (calculated by default settings in Leafcutter, that uses a Likelihood ratio test). Note that Leafcutter outputs a number of datapoints for each gene corresponding to the ΔPSI of each mRNA isoform detected. Some individual genes with high amplitude splice isoform switches in response to PE deletion are labelled. (B) Experimental confirmation of splice isoforms activated within P12 Tra2b-cPEko testes using RT-PCR analysis. Lower panels: Representative capillary gel electrophoretograms. Upper panels: Data from 4 biological replicates of each genotype also used for RNAseq. The mean PSI is shown as a bar with SD as error bar, and p values were calculated using a t test. Note that Map7d2 uses an additional internal primer, so the lower band represents exon skipping. Independent confirmation of some of these splice changes within independent P14 wild type and Tra2b-cPEko testes are shown in Appendix Fig. S4. (C) Tra2b PE deletion changes splicing patterns of the Tra2a gene. UCSC genome browser screenshot showing P12 testis RNAseq reads and adult testis Tra2β iCLIP tags aligned to the Tra2a gene on the mouse genome (mm39). (D) Experimental confirmation of splice isoforms repressed within P12 Tra2b-cPEko testes using RT-PCR. Lower panels: Representative capillary gel electrophoretograms. Upper panels: Data from 4 biological replicates of each genotype. The mean PSI is shown as a bar with SD as error bar, and p values were calculated using a t test. Independent confirmation of some of these splice changes within independent P14 wild type and Tra2b-cPEko testes are shown in Appendix Fig. S4. Source data are available online for this figure.

Figure 5. Activated splice sites in the Tra2b-cPEko testes are hypersensitive to Tra2β protein concentrations.

(A) UCSC genome browser screenshot of the mouse genome (mm39), showing P12 testis RNAseq reads and Tra2β iCLIP tags aligned to the Ptbp2 gene, including the position of the cryptic splice site within intron 8. (B) Detection of cryptic splice selection using RT-PCR (lower panel: representative capillary gel electrophoretogram; upper panel: data from 4 biological replicate mice from each genotype showing mean percentage splicing inclusion +/− SEM). (C) Model predicting behaviour of Tra2β hyper-responsive exons in Tra2b-cPEko germ cells. (D) Design of Ptbp2 intron 8 minigene to test behaviour of a candidate Tra2β hyper-responsive splice site. The relative position of primers (F, R1 and R2) used for the RT-PCR analysis are shown. (E) Representative capillary electrophoretogram showing detection of cryptic splice isoforms in RNA isolated from HEK293 cells after transfection with the Ptbp2 intron 8 minigene and expression constructs encoding either GFP or Tra2β-GFP-fusion proteins (from n = 4 biological replicates). (F) Corresponding Western blot analysis probed for GFP of protein extracted from HEK293 cells after transfection of expression constructs and minigenes (from n = 4 biological replicates). (G) Graph showing mean PSI levels of the Tra2β hyper-responsive splice site within each group of transfected HEK293 cells. In each case, data were generated from 4 biological replicates and error bars represent SD. P values were calculated using t tests. Source data are available online for this figure.

Figure 6. Tra2b gene function is required for mitotic proliferation within the germline.

(A) Schematic of the Tra2b exon 4 conditional allele showing the position of the LoxP sites used to create the Tra2b-cko mouse line. (B) Morphologies of adult testes from different genotype mice (scale shown beside testes is in millimetres). (C) Adult testis:body weight ratios of different genotype mice. Individual values are shown, with the mean and SD indicated. P values were calculated using a t test (n = 46 Tra2b-cko testes, n = 46 Tra2b-het testes, n = 28 wild type testes, [age range 4–33 weeks]). (D) Epididymal sperm counts from different genotype mice (n = 8 Tra2b-cko, n = 4 Tra2b-het, n = 3 wild type, age range 10–18 weeks). P values were calculated using a t test. The error bar represents the SEM. (E) Micrographs of hematoxylin-stained adult testis sections from different germline genotypes (from n = 3 per genotype). Scale bar = 20 μm. Abbreviations for cell types shown in Fig. 1B, with the addition of SC (Sertoli Cell). (F) Histological analysis of testis sections from Tra2b-cko (P0–P3), and wild type (P0–P2) or Tra2b-het (P3) neonatal mice. Testes were harvested between the day of birth (P0) and postnatal day 3 (P3), stained for RBMY protein and counterstained with hematoxylin (Abbreviations for cell types are shown in Fig. 1B). Sample numbers: P0 (n = 2 Tra2b-cko, n = 2 wild type); P1 (n = 2 Tra2b-cko, n = 2 wild type); P2 (n = 2 Tra2b-cko, n = 3 wild type); P3 (n = 2 Tra2b-cko, n = 1 Tra2b-het, n = 1 wild type). Scale bar = 20 μm. (G) Immunostained P1 testis sections stained for Tra2β protein, and counterstained with hematoxylin (from n = 2 Tra2b-cko and n = 2 wild type). Abbreviations for cell types shown in Fig. 1B, with the addition of SC (Sertoli Cell). Scale bar = 20 μm. Source data are available online for this figure.

Figure EV1. Generation of mice with germ cell-specific deletion of the Tra2b PE.

(A) The Tra2b PE was flanked with LoxP sites. The blue primers bind only to the WT allele as the forward primer sits in the region that is removed and replaced by the insertion of the 5′ LoxP site. Thus, they will only amplify a non-floxed or non-recombined allele. The black reverse primer sits on the 3′ LoxP cassette and the forward in the floxed region so will only amplify floxed animals. Neither primer pair will amplify in cre-recombined animals. (B) Breeding scheme. Vasa-Cre was used to excise the Tra2b PE. Male Vasa-Cre transgenic mice were mated with female Tra2bPE^fl/fl mice to obtain Tra2bPE^fl/+;Vasa-cre mice. We then mated male Tra2bPE ^fl/+;Vasa-Cre mice (the floxed allele will be deleted in mature sperm by the Vasa-Cre to generate a deletion allele) with female Tra2bPE^fl/fl mice to generate four possible genotypes. Actual mouse numbers born of each genotype are shown, along with actual and expected frequencies. (C) Example of genotyping result by agarose gel electrophoresis. Lane M marker, lanes 1, 2 presence of the Vasa-Cre transgene, lanes 3, 4, 5 different Tra2b PE alleles as shown in part (A). (D) Litter sizes of Tra2b-cPEko females and wild type females, after crossing with wild type male mice. Individual litter sizes from wild type female mice are shown as black dots, and the mean as a horizontal line. No litters were obtained from Tra2b-cPEko female mice (red dot). 3 breeding cages of each cross were maintained until each of the wild type cages had produced a litter.

Figure EV2. Analysis of the effect of Tra2b PE deletion at days P12 and P14 of mouse testis development.

(A) Micrographs showing immunohistochemical detection of Tra2β protein in sections made from wild type and Tra2b-cPEko adult mouse testes (n = 2). Segments of seminiferous tubules at different stages are shown to enable all the major stages of postnatal mouse spermatogenesis to be visualised. Abbreviations for cell types shown in Fig. 1B, with the addition of SC (Sertoli Cell). Scale bar = 20 μm. (B, C) Micrographs of Haematoxylin-stained histological sections of (B) P12 mouse testes and (C) P14 mouse testes of different genotypes. Scale bar = 20 μM. Red arrows indicate sloughing germ cells. Scale bars = 20 µm. Sample numbers for each time point, n = 4 wild type and n = 4 Tra2b-cPEko.

Figure EV3. iCLIP Identification of endogenous binding sites for Tra2β in the adult mouse testis.

(A) After cross-linking, endogenous Tra2β protein was immunoprecipitated from wild type mouse testis. Autoradiograph shows 32P-labelled RNA cross-linked to endogenous Tra2β from adult mouse testis during one of three replicate iCLIP experiments. At high RNase concentrations a single radiolabelled RNA-protein adduct of ~40 kDa was detected (arrowed), corresponding to the approximate molecular weight of uncross-linked endogenous Tra2β protein (37 kDa). Tra2β iCLIP tags were recovered at lower RNase concentrations (region used highlighted in red box) in biological triplicate, and these tags were used to map endogenous Tra2β binding sites across the mouse testis transcriptome. (B) Analysis of most enriched 5-mers detected close to cross-linking sites within exons (data from one iCLIP replicate shown, similar data were obtained in each of three independent iCLIPs using biological replicate testes). The most frequently occurring pentamers within the iCLIP tags were highly enriched in AGAA nucleotide sequences, exactly corresponding to the known Tra2β binding site (Cléry et al, ; Tsuda et al, 2011). (C) Genomic distribution of Tra2β binding sites. Average percentage of cross-links from triplicate iCLIP experiments within each genomic region are shown.

Figure EV4. Increased inclusion of Tra2β target exons in the Tra2bcPEko mouse testis.

RNAseq reads are merged tracks from testes of 4 P12 mice of different genotypes. Tra2β iCLIP tags are pooled from 3 biological replicate iCLIP experiments using adult wild type mice. (A) UCSC mouse genome (mm39) browser screenshot of the Ggnbp2 gene locus. This screenshot contains the fully expanded iCLIP track, showing an accumulation of iCLIP tags mapping to exon 6. Increased inclusion of Ggnbp2 exon 6 is detected in Tra2b-cPEko mouse testes compared to wild type. (B) Detail of exon 6, indicating the positions of experimentally mapped iCLIP tags, and consensus Tra2β protein-RNA binding sites (GAA-containing) sequences. (C) UCSC genome browser screenshot showing RNAseq and iCLIP reads aligned to the mouse genome (mm39) at the Rpgr locus. Increased inclusion of Rpgr exon 14 was detected within the Tra2b-cPEko testes compared to wild type.

Figure EV5. Generation of mice with germ cell-specific deletion of Tra2b exon 4.

(A) An existing conditional Tra2b^fl allele in which Tra2b exon 4 was flanked with LoxP sites was used to inactivate Tra2b gene function. (B) Experimental crosses. Male Vasa-Cre transgenic mice were mated with female Tra2b^fl/fl mice to obtain Tra2b^fl/+;Vasa-cre mice. We then mated male Tra2b^fl/+;Vasa-Cre mice (the floxed allele will be deleted in mature sperm by the Vasa-Cre to generate a Tra2b knockout allele) with female Tra2b^fl/fl mice to generate four possible genotypes. Actual mouse numbers born of each genotype are shown, along with actual and expected frequencies. (C) Example of agarose gel electrophoresis genotyping result. Lane M marker, lanes 1, 2, 3 distinguish the different Tra2b alleles shown in part (A). (D) Litter sizes of Tra2b-cko females, Tra2b-het and wild type females, after crossing with wild type male mice. Individual litter sizes from wild type female mice are shown as black dots, and the mean as a horizontal line. No litters were obtained from Tra2b-cko female mice (red dot). Breeding cages of each cross were maintained until each of the wild type cages had produced a litter.