A ribosomal protein S5 isoform is essential for oogenesis and interacts with distinct RNAs in Drosophila melanogaster

Abstract

In Drosophila melanogaster there are two genes encoding ribosomal protein S5, RpS5a and RpS5b. Here, we demonstrate that RpS5b is required for oogenesis. Females lacking RpS5b produce ovaries with numerous developmental defects that undergo widespread apoptosis in mid-oogenesis. Females lacking germline RpS5a are fully fertile, but germline expression of interfering RNA targeting germline RpS5a in an RpS5b mutant background worsened the RpS5b phenotype and blocked oogenesis before egg chambers form. A broad spectrum of mRNAs co-purified in immunoprecipitations with RpS5a, while RpS5b-associated mRNAs were specifically enriched for GO terms related to mitochondrial electron transport and cellular metabolic processes. Consistent with this, RpS5b mitochondrial fractions are depleted for proteins linked to oxidative phosphorylation and mitochondrial respiration, and RpS5b mitochondria tended to form large clusters and had more heterogeneous morphology than those from controls. We conclude that RpS5b-containing ribosomes preferentially associate with particular mRNAs and serve an essential function in oogenesis.

Figure 1

RpS5b is a germline ribosomal protein and RpS5a is expressed mostly in somatic cells. (a) Protein sequence alignment comparing RpS5a and RpS5b to each other and to RpS5 orthologues in yeast and human. These proteins are highly homologous but diverge in their N-terminal regions. (b) Co-immunoprecipitation experiments. Immunoprecipitations were conducted with antisera recognizing RpS5a, RpS5b, and non-immune IgG (top captions), blotted, and probed with antibodies as indicated on the right. (c) Western blot showing that RpS5b is detected in the ovary (O) and testis (T) but not in female or male carcasses (f.carc and m.carc), while RpS5a is ubiquitously expressed. (d) Western blot showing that RpS5b is highly abundant relative to RpS5a in 0–2 h embryos (E2h), but both paralogs have relatively equal abundance in 0–16 h embryos (E16h). (e) Immunostaining experiments showing that RpS5b is primarily expressed in the germline cells, while RpS5a is primarily, but not entirely, expressed in follicle cells. RpS6 is equally abundant in both tissues. Scale bars, 20 μm.

Figure 2

RpS5b mutant ovaries have numerous developmental defects that can be rescued by germline expression of RpS5a or RpS5b. (a) Western blot demonstrating that RpS5b is undetectable in RpS5b homozygotes and present at reduced levels in RpS5b heterozygotes, confirming the loss-of-function nature of the mutation. (b,c) Oogenesis does not proceed beyond stage 8 in RpS5b ovaries, at which point apoptosis is induced, as measured by increased levels of activated caspase-3 (C3). α-Orb is used to label the oocyte. (d–h) Various defects observed in RpS5b ovaries: (d) an extra round of germ cell division; (e) a compound egg chamber partially separated by follicle cells, (f) oocyte duplication in a single egg chamber; (g) mis-localized oocyte; (h) multiple layers of follicle cells at the posterior of the egg chamber (white arrow). (i–l) Alterations in the microtubule cytoskeleton in RpS5b oocytes, as measured by immunostaining against (i,j) α-Tubulin, or (k,l) α -Dynein heavy chain. Note the aberrant accumulation of α-Tubulin around the oocyte in (j), and the focus of Dynein in the centre of the oocyte in (l). (m,n) Distribution of Osk and Grk in (m) wildtype and (n) RpS5b oocytes, showing that deployment of these proteins is disrupted in the mutant. (o,p) Analysis of RpS5b germline clones, showing similar defects as found in the mutant, but more extreme overproliferation of follicle cells. (q) Western blot showing RpS5b expression in RpS5b mutant and germline clones. The residual expression in the germline clones is somatic, as is also apparent in (o,p). (r) Western blot comparing RpS5a expression in 0–2 h embryos collected from wildtype females and those expressing shRNA targeting RpS5a driven by the germline-specific promoter nos, showing the efficacy of knockdown. (s) Analysis of ovaries from females expressing shRNA targeting RpS5a driven by the germline-specific promoter nos, showing normal patterning. (t) Brightfield images of whole ovaries showing that RpS5a germline knockdown produces no phenotype but worsens the RpS5b mutant phenotype. (u–w) RpS5b mutant ovaries (u) without a transgene as control or expressing transgenic (v) RpS5b or (w) RpS5a under the control of the nos promoter. Normal oogenesis is restored in both cases. (x) Western blot of lysates from 0–2 h embryos collected from wildtype (WT), nos > RpS5a; RpS5b (NG4–5a; S5b) and nos > Rps5b; RpS5b (NG4-5b; S5b) females, confirming high-level expression from the transgenes. (y) Graph showing hatching rates of embryos from females of the genotypes indicated, demonstrating that either RpS5a or RpS5b can fully rescue the fertility of RpS5b females when expressed in germline.

Figure 3

Analysis of RNA populations recruited by FLAG-HA (FH)-RpS5a and RpS5b. (a) MA plot of RNA immunoprecipitations from the ovaries with germline overexpression of either FH-RpS5a or FH-RpS5b in the wildtype background with α-FLAG compared to input. Statistically enriched (>2 fold, padj <0.01) and depleted (<2 fold, padj <0.01) are highlighted in red and blue respectively. M = log2(pulldown) −log2(input), A = 0.5 * (log2(pulldown) + log2(input)). Fold changes and adjusted p-values (padj) calculated by DESeq2. The Venn diagram (http://bioinfogp.cnb.csic.es/tools/venny/) shows limited overlap between RNAs enriched in populations recruited by FH-RpS5a (FH-5a) and RpS5b (FH-5b). (b,c) Heat map representing biological process gene ontology (GO) terms of RNAs enriched in populations recruited by (b) FH-RpS5b (Statistical overrepresentation test on Pantherdb.org). The most highly significant matches are in red. The fold enrichment of each GO term is plotted in the bar chart. (c) Box plot of the length distribution, in nucleotides, of the 5′UTR, coding sequence (CDS) and 3′UTR for RNAs enriched in populations recruited by FH-RpS5a (FH-RpS5a_IP) and FH-RpS5b (FH-RpS5b_IP). (d,e) Heat maps representing the biological process GO terms associated with the proteins (d) enriched or (e) depleted in the mitochondrial fractions from Rps5b ovaries as compared with wild-type. The most highly significant matches are in red.

Figure 4

RpS5b ovaries have mitochondria with aberrant distribution and morphology, and elevated ROS. (a) Immunostaining for α-ATP5a, a subunit of mitochondrial ATP synthase, reveals a much more densely clustered distribution of mitochondria in nurse cells from RpS5b, or RpS5b germline clones (RpS5b_GC) than from wildtype (WT). (b) Transmission electron micrographs (TEM) of thin sections from nurse cells show that RpS5b mutant or RpS5b germline clones have mitochondria with aberrant cristae morphology and irregular shapes (red arrows). (c) TEM images of mitochondria labeled with colloidal-gold conjugated α-ATP5a, illustrating morphological changes in RpS5b mitochondria (red arrow). (d) Images of live ovaries from wildtype (WT) and RpS5b females treated with CellROX, a sensor for reactive oxygen species (ROS), showing elevated levels in RpS5b.