Polycomb safeguards imaginal disc specification through control of the Vestigial-Scalloped complex

Abstract

A fundamental goal of developmental biology is to understand how cell and tissue fates are specified. The imaginal discs of Drosophila are excellent model systems for addressing this paradigm as their fate can be redirected when discs regenerate after injury or when key selector genes are misregulated. Here, we show that when Polycomb expression is reduced, the wing selector gene vestigial is ectopically activated. This leads to the inappropriate formation of the Vestigial-Scalloped complex, which forces the eye to transform into a wing. We further demonstrate that disrupting this complex does not simply block wing formation or restore eye development. Instead, immunohistochemistry and high-throughput genomic analysis show that the eye-antennal disc unexpectedly undergoes hyperplastic growth with multiple domains being organized into other imaginal discs and tissues. These findings provide insight into the complex developmental landscape that tissues must navigate before adopting their final fate.

Keywords: Drosophila; Epigenetics; Eye; Imaginal disc; Polycomb; Scalloped; Vestigial; Wing.

Fig. 1.

Antp is ectopically expressed when some PcG members are knocked down. (A-G) Third instar eye-antennal imaginal discs of WT or knockdown flies stained with anti-Antp (green), anti-Elav (blue) and phalloidin (red). (A′-G′) Antp channel alone from the same eye-antennal discs. (A″-G″) WT (A″) or flies lacking either Pho (C″), Sfmbt (D″) or Calypso (E″) can eclose and maintain mostly WT morphology. However, knocking down either Pc (B″), Sce (D″) or Scm (E″) is pharate lethal. Asterisks in B′ and B″ mark the eye-to-wing transformation in the disc or pharate fly, respectively. Dashed arrows in E′ and E″ show outgrowth of the antennal field. Arrows in F′,F″,G′,G″ mark the antenna-to-leg transformation and presence of a sex comb (F″), indicating that the Sce knockdown transforms into a T1 leg. (H) Bar graph representing the percentage of PcG knockdown discs that show ectopic Antp expression with three different GAL4 drivers (ey-GAL4: blue; DE-GAL4: green; c311-GAL4: orange). For all phenotypic scoring, n=30. Penetrance and severity of each phenotype are listed in Table S1. Scale bar: 50 µm.

Fig. 2.

Overexpression of Hox genes disrupts eye development. (A-H) Third instar eye-antennal imaginal discs stained with anti-Elav (green) and phalloidin (magenta). Whole-mounted disc–brain complexes are shown for ey>lab (E) and ey>pb (F). Asterisks denote where the missing eye-antennal discs should be in these complexes. (A-D) Hox genes with elevated transcript levels in Pc knockdown discs compared with WT eye-antennal discs (Antp, Scr, Ubx and Abd-B). (E-H) Hox genes with relatively unchanged transcript levels in Pc knockdown discs compared with WT eye-antennal discs (lab, pb, Dfd and abd-A). (A′-H′) Eclosed adult flies are eyeless, have a missing head capsule (G′,H′) and have reduced number of ocelli (B′-D′). Pharate lethal flies are headless (E′,F′). (A″-H″) Bar graphs representing DESeq2 normalized transcript counts of three replicates for the WT eye-antennal disc (blue), Pc knockdown (PcKD) disc (green) and WT wing disc (orange). Error bars represent standard deviation. For time point A (84 h WT versus 96 h PcKD) only WT EAD and PcKD data are shown. Key for time-point comparisons and disc type are given at the bottom. For all phenotypic scoring, n=30. Penetrance and severity of each phenotype are listed in Table S1. Scale bar: 50 µm.

Fig. 3.

Differential expression of genes in response to Pc reduction. (A-C) Heatmap of log2-transformed differential transcript expression between Pc knockdown eye-antennal discs and WT wing discs (left) or eye-antennal discs (right) throughout early (A; 108 h versus 96 h), mid (B; 120 h versus 108 h) and late (C; 144 h versus 120 h) third instar development. Known eye genes, wing genes and Hox genes are listed to the right. (D) Schematic representing the selection method for potential candidates behind the eye-to-wing transformation. Green arrow represents genes upregulated in wild-type wing disc and Pc knockdown eye-antennal disc, red arrow represents genes downregulated in eye-antennal disc. (E) Venn diagram showing the overlap of candidate genes throughout early (green, n=516), mid (orange, n=220) and late (blue, n=233) third instar development. Asterisk denotes candidate genes that are differentially expressed throughout the entire tested developmental window. (F) List of annotated candidates in each of the overlapping categories. Genes in bold have available UAS overexpression lines. The remaining genes for each category are listed in Table S3.

Fig. 4.

Pc reduction results in H3K27me3 profiles that resemble wing discs. (A,B) H3K27me3 (A) and H2AK119Ub (B) CUT&RUN heatmaps of WT eye-antennal discs (EAD, left), Pc knockdown eye-antennal discs (PcKD, center) and wing discs (WD, right). (C) CUT&RUN H3K27me3 (indigo), H2AK119Ub (blue) and transcript (green) peaks at the CG4382 (Group 1), CG7173 (Group 2) and Ir87a (Group 3) loci of each genotype. The corresponding histone mark IgG control track is overlaid in yellow.

Fig. 5.

Vestigial is a candidate for the Pc eye-to-wing transformation. (A) WT adult head. (B,C) WT third instar wing disc (B) and eye-antennal disc (C) stained with anti-Vg (green) and phalloidin (magenta). (D) Overexpression of vg in the eye-antennal disc is sufficient to drive an eye-to-wing transformation. Arrow indicates wing outgrowth from the eye. (E) ey>Pc RNAi-transformed discs stained with anti-Vg (green) and phalloidin (magenta). Vg is ectopically activated in the pouch of the transformed disc (arrow). (F) Bar graphs representing DESeq2 normalized transcript counts of three replicates for the WT eye-antennal disc (blue), Pc knockdown disc (green) and WT wing disc (orange). Error bars represent standard deviation. (G) CUT&RUN H3K27me3 (top), H2AK119Ub (middle), and forward transcript (bottom) peaks at the vg locus in the WT eye-antennal disc, Pc knockdown disc and WT wing disc. The corresponding histone mark IgG control track is overlaid in yellow. Asterisks denote the two PREs identified by Ahmad and Spens (2019). Scale bar: 50 µm.

Fig. 6.

The Pc-dependent eye-to-wing transformation relies on the Vg–Sd complex. (A-E) Third instar eye-antennal imaginal discs stained with anti-Antp (green), anti-Elav (blue) and phalloidin (red). (A′-E′) Eclosed or pharate lethal adult heads of the corresponding genotype. Arrows denote the indicated phenotype. (A) Overexpression of vg disrupts photoreceptor development and results in disc folds in the anterior eye field. (A′) Eclosed adults develop wing outgrowths on every axis of the eye. (B-C′) When vg is overexpressed while sd is knocked down (B,B′), the phenotype mimics sd knockdown animals – malformations in the morphology of the anterior eye field (C) and adults with abnormal, cone-shaped eye outgrowths (C′) – rather than vg overexpression animals (A′). (D,E) Knocking down Pc in a vg¹ (D) or sd¹ (E) mutant background disrupts the eye-to-wing transformation and instead leads to hyperproliferative discs. (D′,E′) The combined loss of vg and Pc (D′) is pupal lethal, whereas the combined loss of sd and Pc (E′) is pharate lethal and results in flies with diminished, ‘tumorous’ heads. For all phenotypic scoring, n=30. Penetrance and severity of each phenotype are listed in Table S1. Scale bar: 50 µm.

Fig. 7.

Fate transformations occur in hyperplastic tissue. (A) WT third instar eye-antennal, wing, leg and haltere imaginal discs stained with anti-Wingless (Wg). (B,C) When Pc is lost in vg¹ (B) or sd¹ (C) mutants, the eye-to-wing transformation is disrupted, and the tissue instead shows evidence of specification of other imaginal fates (arrows). (D) WT third instar eye-antennal, wing, leg and haltere imaginal discs stained with anti-Dachshund (Dac). (E,F) When Pc is lost in vg¹ (E) or sd¹ (F) mutants, discs specify other imaginal fates (arrows). (G) The loss of vg and Pc results in high instances of wing specification with additional antennal or leg transformations in the eye and antennal fields. (H) The combined loss of sd and Pc results in similar novel fate specification, with unique instances of haltere specification in both the eye and antennal fields. (I) Venn diagram comparing the differentially expressed genes of ey>Pc RNAi (PcKD), vg¹; DE>Pc RNAi (vgPcKD) and sd¹; ey>Pc RNAi (sdPcKD) eye-antennal discs. Green text represents the number of enriched genes and black represents total number of differentially expressed genes. (J) Enriched GO analysis of overlapping enriched genes between vgPcKD, sdPcKD and PcKD discs (asterisk in I). For all phenotypic scoring, n=30. Penetrance and severity of each phenotype are listed in Table S1. Scale bar: 50 µm.

Fig. 8.

Model for Pc regulation of eye-antennal disc fate. (A) We propose that Pc maintains the fate of the WT eye-antennal disc by controlling the expression of genes that encode growth factors, Hox, and selector genes for other imaginal discs. The repression of Vg allows for Sd to form a growth-promoting complex with Yki instead of directing wing fate. (B) When Pc is knocked down, vg expression is activated along with several Hox genes and growth factors. Formation of the Vg–Sd complex transforms the eye into a wing. (C) If the Pc knockdown is combined with a vg mutation, then the disc undergoes hyperplastic growth and additional imaginal disc transformations (brown, blue and light green) and possibly non-imaginal disc fates (gray). (D) If the Pc knockdown is combined with an sd mutation, the disc undergoes hyperplastic growth and additional imaginal disc transformations (brown, blue, purple and light green) and possibly non-imaginal disc fates (gray). The eye and antenna can be transformed into haltere only in this genetic background.