Fig 1.
An RNAi screen of putative de novo genes identifies CG13541 as a major contributor to Drosophila melanogaster male fertility.
A) All RNAi lines that showed at least partial knockdown of the target gene were screened in group fertility assays (see Materials and Methods). Relative fertility was calculated by dividing the average number of progeny produced per female mated to knockdown males by the average number of progeny produced per female mated to control males in a contemporaneous experiment. Relative fertility measurements lack error bars because each gene was tested in only 1–2 replicates. Knockdown of goddard was used as a positive control. B) A single-mating, single-pair fertility assay confirms the observed defect when males are knocked down for CG13541, as knockdown males showed significantly reduced fertility (control fertility (mean ± SEM): 109.0 ± 5.3; knockdown fertility: 0.2 ± 0.1; two-sample t-test assuming unequal variances, p = 5.6 x 10−13).
Fig 2.
CRISPR-generated deletion and frameshift alleles of atlas confirm the gene’s requirement for male fertility.
A) Single-pair fertility assay for males homozygous for the null (Δatlas) allele or heterozygous controls (Δatlas/SM6). Flies homozygous for the deletion had significantly reduced fertility (control fertility: 82.9 ± 4.5; null fertility: 0.3 ± 0.6; two-sample t-test assuming unequal variances, p = 5.4 x 10−18). B) Single-pair fertility assays for males homozygous or heterozygous for three frameshift alleles of atlas generated by imprecise non-homologous end joining at a CRISPR/Cas9 target site just downstream of the start codon: atlas52 (control fertility: 104.7 ± 3.8, mutant fertility: 11.2 ± 4.6; two-sample t-test assuming unequal variances: p = 8 x 10−12), atlas62 (control fertility: 96.2 ± 5.7; mutant fertility: 8.4 ± 2.9; two-sample t-test assuming unequal variances: p = 6.1 x 10−10) and atlas86 (control fertility: 67.5 ± 6.6; mutant fertility: 17.3 ± 5.8; two-sample t-test assuming unequal variances: p = 9.5 x 10−6).
Fig 3.
Cytological investigations of the atlas mutant fertility defect.
A) Phase contrast microscopy of male reproductive tracts dissected from 7-day-old, unmated control (w1118) or atlas null males. Control males show the expected accumulation of sperm tails in the seminal vesicle (SV), which appear here as a darker brown shading, while null male have an aberrant accumulation of sperm tails at the basal end of the testis (T). B) Visualization of Mst35Bb-GFP in 4-day old control and atlas null testes. While Mst35Bb localizes to spermatid nuclei in the absence of atlas, the nuclei appear shorter and much less numerous in the outlined SV. C) Representative images from phalloidin staining of w1118 and atlas null testes used to assess the association of individualization complexes (ICs) with nuclear bundles and the progression of ICs down the length of sperm tails. D) At top, number of nuclear bundles with ICs associated in control and atlas null testes. Significantly more ICs were observed in control testes (control: N = 17, median = 14; mutant: N = 13, median = 7; Wilcoxon rank-sum test W = 34, p = 0.0014). At bottom, proportion of all observed ICs that were intact and that had progressed away from nuclear bundles. Three mutant testes with no observed ICs were excluded from the analysis. A significantly higher proportion of ICs progressed in control testes (control: N = 17, median = 0.27; control: N = 10, median = 0; Wilcoxon rank-sum test W = 28, p = 0.0038).
Fig 4.
atlas null males show aberrant nuclear shaping at and beyond the elongated stage of spermatid nuclear condensation.
Early and late canoe stages were distinguished by the absence or presence of Mst35Bb-GFP, respectively. Late canoe and elongated stages were distinguished by the absence or presence of GFP-positive puncta, respectively. Condensed nuclei were distinguished from elongated nuclei by size. As shown in S1 Table, nuclear bundles from atlas null testes consistently took on a curved shape after the canoe stage, though the degree of curvature was variable, as exemplified by the two examples of elongated nuclei from atlas null testes above.
Fig 5.
An atlas-GFP allele generated at the endogenous atlas locus is fully functional for male fertility and encodes a protein that localizes to condensing spermatid nuclei.
A) A single copy of the atlas-GFP allele is sufficient for full fertility when paired with the Δatlas allele, as compared to males heterozygous for the wild-type atlas allele (Δatlas/atlas-GFP fertility: 86.0 ± 3.2; Δatlas/+ fertility: 84.9 ± 4.1; two-sample t-test assuming unequal variances, p = 0.85). B) Atlas-GFP does not co-localize with histone H2Av-RFP, a marker of the initial stages of spermatid nuclear condensation. C) Visualization of Atlas-GFP in the basal portion of whole-mount testes from atlas-GFP homozygotes shows that the fusion protein co-localizes with a subset of condensing spermatid nuclear bundles. While actin associates with fully condensed nuclei at the basal testis, Atlas-GFP does not overlap and is also absent from the seminal vesicle. D) Atlas-GFP partially colocalizes with Mst35Bb-dsRed, a marker of the final stage of nuclear condensation, in the basal portion of whole-mount testes. Open arrow: example of co-localization. Filled arrowhead: example of Atlas-GFP that does not co-localize with Mst35Bb-dsRed. Collectively, these data suggest that atlas may serve as a transition protein involved in the final stages of nuclear condensation. The whole testes from which the basal portions are shown in panels C and D are shown in S7 Fig.
Fig 6.
Atlas-GFP is present in late canoe-stage spermatid nuclei and then appears to leave the nucleus in puncta.
Staging of condensing spermatid nuclei fixed in paraformaldehyde from atlas-GFP males stained with TO-PRO-3 DNA stain. Atlas-GFP is not detectable in (A) round stage or (B) early canoe stage nuclei. Atlas-GFP is nuclear localized in the late canoe stage (C). When nuclei become fully elongated (D), puncta of Atlas-GFP appear to be removed from the nucleus.
Fig 7.
Molecular evolution and gene expression of atlas across the Drosophila genus.
The gene structure of atlas in D. melanogaster is shown at top left. The predicted protein-coding sequence is contained entirely within exon 1, while exon 2 encodes the presumed 3’ UTR. The gene is located on chromosome 2R, equivalent to Muller element C. The phylogeny shows BLAST- and synteny-based detection of sequences orthologous to the protein-coding sequence and the 3’ UTR sequence across Drosophila species. Sex-specific adult RNA-seq data were used to assess male expression across species, with RT-PCR verification performed in species marked with asterisks. RNA-seq data for the syntenic region of the 3’ UTR in D. virilis were ambiguous; see S8 and S10 Figs.
Table 1.
PAML sites tests for positive selection acting on atlas in the melanogaster group.