Transcriptional responses to hyperplastic MRL signalling in Drosophila

Abstract

Recent work has implicated the actin cytoskeleton in tissue size control and tumourigenesis, but how changes in actin dynamics contribute to hyperplastic growth is still unclear. Overexpression of Pico, the only Drosophila Mig-10/RIAM/Lamellipodin adapter protein family member, has been linked to tissue overgrowth via its effect on the myocardin-related transcription factor (Mrtf), an F-actin sensor capable of activating serum response factor (SRF). Transcriptional changes induced by acute Mrtf/SRF signalling have been largely linked to actin biosynthesis and cytoskeletal regulation. However, by RNA profiling, we find that the common response to chronic mrtf and pico overexpression in wing discs was upregulation of ribosome protein and mitochondrial genes, which are conserved targets for Mrtf/SRF and are known growth drivers. Consistent with their ability to induce a common transcriptional response and activate SRF signalling in vitro, we found that both pico and mrtf stimulate expression of an SRF-responsive reporter gene in wing discs. In a functional genetic screen, we also identified deterin, which encodes Drosophila Survivin, as a putative Mrtf/SRF target that is necessary for pico-mediated tissue overgrowth by suppressing proliferation-associated cell death. Taken together, our findings raise the possibility that distinct targets of Mrtf/SRF may be transcriptionally induced depending on the duration of upstream signalling.

Keywords: Drosophila; MRL proteins; hyperplastic growth; serum response factor; wing development.

Figure 1.

Transcriptome analysis of wing discs overexpressing pico. (a) Pie chart showing relative abundance of gene ontology (GO) term enrichment in genes underexpressed and overexpressed in wing imaginal discs of MS1096>pico third instar larvae relative to abundance of GO terms for all genes in the genome as determined by DAVID. For each category of functional group (in bold), the most prominent biological function (in italic) has been annotated with the number of genes affected in that category, the total number of genes in that category, the statistical significance (p-value) of the match and fold enrichment (FE). The most prominent GO categories among those upregulated in response to pico overexpression are related to protein biosynthesis, including initiation of translation, ribosomal function and protein maturation. There is also an enrichment in proteins localized to mitochondria. (b) Predicted interacting network for genes over-represented in response to pico overexpression, visualized using STRING. Potential associations are indicated by the links in the graph and colour coded by type: co-citation from the abstract of scientific literature (green), proteins related in curated databases (blue) and physical protein-protein for interaction databases (pink).

Figure 2

Validation of RNA-Seq by qRT-PCR. Expression levels of selected genes from MS1096>pico wing discs from third instar larvae, relative to control, determined by qRT-PCR and by RNA-seq. Error bars represent the s.e.m. of at least three biological replicates. The GO categories to which the genes belong are shown at the top. Individual t-tests without adjustment for multiple comparisons showed a significant difference (p < 0.05) in each case between transcript levels in MS1096>pico and control discs.

Figure 3.

Distribution of an in vivo SRF-responsive reporter gene in wing discs. (a) Model for Mrtf/SRF activation by Pico overexpression. Increased F-actin formation leads to sequestration of G-actin, relieving inhibition of Mrtf, which translocates to the nucleus and complexes with SRF to drive transcription of genes containing SRE. SRF activation can be monitored using an SRE-mCherry reporter. (b) Confocal images of wing discs harbouring SRE-mCherry transgenic reporter, stained with anti-SRF antibody. The distribution of the SRE-mCherry reporter closely matches the distribution of SRF protein in presumptive intervein cells.

Figure 4.

Overexpression of pico induces SRE-mCherry expression in larval and pupal wing discs. Scatterplot shows measurements from different wing imaginal discs of the ratio of mean mCherry intensity in cells inside : outside the GFP-labelled posterior half of each disc (from z-stacks of at least four wing discs per genotype). Mean values ±s.e. for each genotype are indicated with a line. The genotype and developmental stage are as indicated (dashes correspond to hh-GAL4, UAS-GFP alone). The results of t-tests comparing mCherry levels between discs with or without overexpressed pico at each stage are indicated.

Figure 5.

Overexpression of mrtf or pico induces SRE-mCherry expression. (a) Shown are representative images of an apical view of wing discs overexpressing mrtf or pico in the posterior compartment of third instar wing imaginal discs (marked with GFP) using hh-GAL4. GFP labels hh-GAL4 expressing cells (in green), anti-RFP antibody staining reveals SRE-mCherry distribution (in red), TOPRO-3 staining reveals DNA (in blue). For clarity, a dotted line in the images showing SRE-mCherry alone indicates the position of the anterior/posterior boundary. (b) mrtf and pico induce SRE-mCherry expression in vivo. Shown are graphs of the distribution of SRE-mCherry signal intensities inside (IN) or outside (OUT) GFP-labelled compartments (from z-stacks of at least 10 wing discs). Levels of SRE-mCherry were noticeably elevated in the posterior half of discs expressing mrtf or pico.

Figure 6.

Divergent and common responses to mrtf and pico overexpression. (a) Principal component analysis showing divergent (PC1) and common (PC2) response to overexpressed pico and mrtf, respectively, which together explain approximately 70% of the variance in gene expression. Data points are four independent biological repeats for each condition (MS1096>pico, red; MS1096>mrtf, yellow, MS1096-GAL4 alone, blue). (b,c) GO enrichment for the top (red) and bottom (blue) 10% of loadings from divergent and common responses. The five most significant GO categories for each grouping are shown. Both pico and mrtf overexpression stimulate ribosome protein genes belonging to the GO category ‘ribosome subunit’.

Figure 7.

Genes induced by pico overexpression are required for pico-mediated tissue growth. (a) Venn diagram showing overlap between genes significantly overexpressed by pico (identified in this study) and genes identified for their role in the cell cycle or growth control in high-throughput functional genetic studies [–28]. (b) Results of genetic screen showing effect of inverted repeat (IR) constructs for 12 genes on adult male wing size, with or without overexpressed pico, expressed as a percentage of control (first column shows MS1096-GAL4 alone). All strains contained MS1096-GAL4 to drive expression in the developing wing disc. Data are shown as scatterplot and mean values indicated with a line (n ≥ 30). Overexpression of pico induced approximately a 10% increase in wing size compared with the control (red data points). Most IR constructs had little effect although Gadd45^IR, diap2^IR and bruce^IR had a modest effect on wing size in both the presence and absence of overexpressed pico. Notably, UAS-RpS9^IR and UAS-det^IR significantly suppressed pico-mediated overgrowth but had little effect in an otherwise wild-type background (n.s., not significant, p > 0.05 t-test). (c) Knockdown of deterin (det) suppressed pico-mediated tissue overgrowth, but not an alteration to wing shape. Flies carrying a UAS-inverted repeat for deterin under the control of MS1096-GAL4 (MS1096>det^IR) resembled wild-type wings (not shown).

Figure 8.

Deterin is a Mrtf/SRF target and suppresses apoptosis during pico-mediated tissue overgrowth. (a) A site 5′ of the transcription start site (TSS) of deterin binds FLAG-SRF. Chromatin immunoprecipitation (ChIP) analyses of three sites at the 5′ end of deterin containing a potential CArG box. Position of the CArG box relative to the TSS is indicated at the bottom. ChIP from third instar larval da>Flag-SRF and control larvae (da-GAL4 alone) was performed using monoclonal anti-FLAG and mouse IgM antibodies. Immunoprecipitated DNA was quantified by qPCR. For each genotype, percentage input is the amount of precipitated DNA relative to input DNA. Results are mean ± s.e.m. from three independent experiments. One-way ANOVA: *p < 0.05; n.s., not significant. Distance of sites from the TSS is indicated on the x-axis. (b) det^IR induces cell death in wing discs co-overexpressing pico. Discs overexpressing Venus-pico (in yellow) are overgrown and show little cleaved Caspase-3-staining (in red); coexpression of detIR reduced tissue size and induced cleaved Caspase-3 in the centre of the wing pouch (arrow). (c) Quantitation of number of Caspase-3-positive foci in wing discs expressing Venus-pico or det^IR, alone or in combination (mean ± s.e.m. from z-stacks of at least three wing discs). t-test: *p < 0.05, n.s., not significant.