Hippo signaling regulates dorsoventral patterning in Drosophila head and eye development

Summary

The integration of growth and patterning is essential for proper tissue development. In Drosophila, the Hippo pathway controls organ size, including formation of adult structures from the eye-antennal imaginal disc. Defective Proventriculus (Dve), a K-50 domain transcription factor, regulates dorsal-ventral patterning by suppressing eye fate and promoting head cuticle development, thereby influencing eye position and interocular distance. We investigated how the Hippo effector Yorkie (Yki) and Dve coordinate growth and patterning in the dorsal head region. Our results show that Yki regulates dorsal head size, while Dve suppresses Yki in this region. Yki also controls dorsal-ventral patterning genes, including pnr, wg, and mirr, to coordinate eye and head development. Because mammalian orthologs of yki, dve, pnr, and wg are linked to developmental disorders such as facial dysmorphia, our findings reveal conserved mechanisms connecting growth, patterning, and craniofacial development across species, with potential relevance to human developmental disease and tissue morphogenesis.

Graphical abstract

Highlights

• •Yki and Dve coordinate growth and patterning in the eye-antennal disc
• •Dve suppresses Yki in the dorsal head region to restrict growth
• •Yki controls DV patterning genes pnr, wg, and mirr
• •Yki-Dve studies reveal mechanisms linked to head and eye development and facial dysmorphia

Cell biology; Developmental biology; Genetics; Molecular biology

Introduction

The integration of growth with patterning is a complex developmental process in which cells within a developing field respond to intrinsic and extrinsic cues to regulate cell division, specification, and differentiation to generate tissues and organs in a stereotypical manner.^1,2 Growth is a fundamental process in all metazoans and is governed by intricate regulatory mechanisms that control cell proliferation and survival.^1,3 These processes are highly conserved across organisms due to the conservation of core developmental patterning and signaling pathways.^4,5 Drosophila melanogaster is a well-studied model for studying intricate developmental processes due to its conserved pathways and powerful genetic tools.⁶

The Drosophila eye-antennal imaginal disc develops into the adult eye, head and antenna; and serves as a robust model to study developmental patterning.^2,7,8,9 Eye antennal progenitors arise from the embryonic head ectoderm and develop into the eye-antennal epithelial imaginal discs that give rise to the compound eyes, antennae, ocelli, the head cuticle and palpus of the adult fly.^2,7,10 The growth and differentiation mechanisms involved in the transition of the monolayered eye-antennal imaginal discs to the adult structures are tightly controlled by interactions among morphogens, signaling pathways and transcriptional regulators.^{2,11,12,13,14,15} Although much is known about the mechanisms that regulate the differentiation of photoreceptor neurons in the wake of the MF that sweeps across the eye disc from posterior to anterior,^10,14,16 the regulation of other regions of the eye discs including the head cuticle domain remains poorly understood.

The Hippo pathway is an evolutionary conserved signaling pathway for regulation of patterning and growth to control organ size in both Drosophila and mammals.^1,3 The upstream regulators of the pathway impinge on a core-kinase cascade comprised the serine-threonine kinase Hippo (Hpo, ortholog of mammalian Sterile-20 like kinases 1 & 2, MST1/2) which partners with the adaptor protein Salvador (Sav, hSAV1) to activate the downstream serine-threonine kinase Warts (WTS) a.k.a. large tumor suppressor (LATS, hLATS1/2). LATS in turn, partners with the adaptor protein Mob as tumor suppressor (MATS, hMOB1A/B) that potentiates its catalytic activity. This complex phosphorylates and inactivates the transcriptional co-activator Yorkie (Yki), ortholog of human Yes-associated protein 1/2 (YAP1/2) and transcriptional coactivator with a PDZ-binding domain (TAZ). Activated Yki/YAP partners with TEA/ATTS family of transcriptions factors like Scalloped (Sd, TEAD1/2/3/4) or MEIS-family transcription factor Homothorax (Hth, hMEIS1) or zinc-finger transcription factor Tea-shirt (Tsh, hTSH3). The Hippo pathway is known to regulate overall organ size, and changes in pathway activity perturb developmentally regulated organ size.^1,3 Over-expression of Yki results in its nuclear translocation and association with the TEAD-family transcription factor Scalloped (Sd) or Meis-family transcription factor Homothorax (Hth).^17,18,19 These transcriptional complexes activate target genes that promote cell division and suppress cell death. Recently, multiple roles for the Hippo pathway in developmental patterning have been described in Drosophila and higher vertebrates.^20,21,22

During eye development, dorsal-ventral (DV) patterning plays a key role in defining the first lineage restriction by the expression of Pannier (Pnr) in the dorsal eye margin.^{23,24,25,26,27} We recently identified defective proventriculus (dve) as a Pnr-regulated dorsal fate selector gene.²⁸ Dve is a K50 homeodomain-containing transcription factor in Drosophila that is highly conserved from insects to humans, and homologous to the human Special AT-rich sequence-binding proteins (SATB1/2)^2,29. SATB1 regulates chromatin to provide docking sites for transcription factor and chromatin modifying enzymes, thereby controlling gene expression.²⁹ Initially, dve was shown to be required for embryonic gut development.^30,31 Subsequent studies showed that dve also regulates development of other organs including legs,³² wings,^33,34 and eyes.²⁸ In the eye discs in Drosophila (and other insect species), Dve is expressed in the dorsal head region, where it regulates the expression of wingless (wg), a secreted morphogen expressed along the anterolateral eye margins, and is involved in determining the eye versus head fate.^17,28 Dve and Wg suppress eye fate by downregulating retinal determination genes and promoting head cuticle development. Loss of function of dve results in dorsal eye enlargement and gain of function of dve causes eye suppression as well as reduced size of the dorsal head region derived from the eye-antennal imaginal disc in flies.²⁸ Overall, Dve plays a key role in DV patterning by controlling the positioning of eyes on the head and defining interocular distance during development.²⁸

However, how growth and patterning mechanisms are coordinated during development is not fully characterized. In particular, the regulation of the dorsal head region that determines the spacing between the two eyes remains largely unexplored. Here, we investigated a putative role of Hippo signaling and its effector yki and dve in determining the optimum growth of eye and dorsal head region for normal tissue homeostasis. We show that Hippo signaling affects size of the dorsal head region size and MF progression during eye development. We further show that Dve-a dorsal fate selector-suppresses Yki activity in the dorsal head region. Further, we found that dve itself is not transcriptionally regulated by Yki in the eye disc; however, pnr the upstream regulator of Dve and mirror (mirr), a gene from the Iroquois-complex (Iro-C) that acts downstream of Wg in the dorsal eye, as new transcriptional targets of Yki. Taken together, our studies support a model in which Dve controls dorsal head region size by antagonizing Yki within the Hippo pathway, and Yki functions upstream of pnr and transcriptionally regulates genes involved in DV patterning and growth regulation during head cuticle expansion and retinal specification. These studies provide a direct mechanistic link between growth control and axial patterning during head-eye morphogenesis, and given the conservation of genes and developmental pathways, our studies are relevant for understanding the etiology of conditions involving craniofacial malformations and dysmorphia in higher vertebrates including humans.

Results

Hippo signaling affects the size of the dorsal head region and MF progression during eye development

Loss of function of dve in the eye (using ey-FLP, FRT42D cl w+) resulted in dorsal eye enlargements and suppression of Wg expression.²⁸ Compared to wild-type eye-imaginal discs (Figures 1A and A′) and adults (Figure 1B), dve-GAL4 UAS-GFP (referred to as dve>GFP, Figures 1C and C′) shows Dve expression (GFP-positive cells in Figure 1C, gray in C′) in about 150–200 cells in the dorsal head region of the eye disc. These cells do not overlap with the differentiated photoreceptors marked by antibodies against ELAV, a pan-neural marker in Drosophila (Figure 1C red). Upregulation of Hippo pathway activity under dve-GAL4 by over-expression of Hpo (dve>GFP, Hpo Figures 1E and E′) or Wts (dve>GFP, Wts Figures 1G and G′) resulted in loss of Dve domain (Figures 1E′ and G′) and concomitant loss of antenna and dorsal head regions in the discs (Figures 1E and 1G).²⁸ In addition, these discs showed remarkable extension of eye by formation of ectopic or elongated furrows leading to defects in patterning in the entire disc (Figures 1E and 1G). These defects in differentiation and patterning can be further tracked post-metamorphosis in adults or pharates. Compared to wild-type (Figure 1B) or dve>GFP adult flies (Figure 1D), the dve>GFP, Hpo (Figure 1F) or dve>GFP, Wts (Figure 1H) animals fail to eclose, and pharate adults show defects in head eversion and/or head capsule formation. As a result, the differentiated eyes remain trapped within the thorax. Together, these data suggest that overexpression of Hippo pathway components in the Dve domain affects eye patterning and affects the size of the Dve expression domain. Therefore, we asked if the Hippo pathway modulates the head versus eye fate decisions in the developing eye antennal discs.

Figure 1.

Upregulation of Hippo pathway activity affects eye and head development

Comparison of dve expression domain size and its effects on eye and head development is shown for (A and B) wild-type, (C and D) dve-GAL4>UAS GFP (GFP, green), (E and F) dve-GAL4> UAS GFP, UAS hpo, and (G and H) dve-GAL4> UAS GFP, UAS wts. (A, C, E, and G) show confocal images of third instar eye imaginal discs stained for the pan-neuronal marker ELAV (red), and dve-GAL4 expression domain marked by UAS GFP (green in C, E, and G; gray in C′, E′, and G′). Panels (B, D, F, and H) show adults flies from corresponding genotypes. In (C′), wild type dveGAL4 expression domain is marked by UAS GFP and is outlined by yellow-dotted line. Scale bars, 100 μM.

Hippo activity affects dorsal head domain size

To test if the Dve domain is indeed affected by manipulating Hippo pathway function, we tested the effects of depleting Hippo pathway activity (Figure 2). To do so, we overexpressed the physiological- (UAS-Yki) and hyperactivated- (UAS-Yki^3SA) forms of Yki that are known to mimic the loss of function effects of the Hippo pathway.³⁵ Compared to the controls (dve>GFP) (Figures 2A and A″), over-expression of the physiological form of Yki in the Dve domain (dve>GFP, Yki, Figures 2B and B″) resulted in a significant increase in size of the Dve expression domain (Figure 2G), and weak suppression of the MF leading to formation of smaller eyes (Figure 2B″, marked with arrowheads). Interestingly, over-expression of hyperactivated form of Yki (dve>GFP, Yki^3SA) further enhanced these defects in the discs (Figures 2C, 2C″, and S1A) and adults (Figure S1B). Quantification of the domain size (area of GFP expression in each genotype) shows that overexpression of both forms of Yki caused significant expansion in the Dve expression domain (Figure 2G) and decrease in eye size due to suppression of the MF progression both in the dorsal and ventral margin (Elav expression in Figures 2B and 2C marked with arrowheads, quantified in Figure 2I). Interestingly, overexpression of Yki and Yki^3SA showed non-autonomous effects on overall disc growth as well (Figure 2H), where the disc area is larger for dve>Yki and dve>yki^3SA as compared to wild-type. However, the dve>Yki^3SA discs were highly folded, and their size was not as large as the hyperplastic Yki overexpressing discs (dve>GFP, Yki). Quantification of the ratio of GFP area to overall disc size showed that Yki^3SA (dve>GFP, Yki^3SA) has greater effect on dve domain size compared to Yki (dve>GFP, Yki, Figure 2H), indicating that the overall growth is higher for hyperactivated Yki.

Figure 2.

Modulating Hippo activity or Dve levels affects the domain size of the dorsal head region and eye

Effects of modulating Hippo pathway activity (A–D, quantified in G–I) and Dve expression (E and F, quantified in I) on Dve expression levels in the dorsal head region and eye size. Eye size is compared based on ELAV expression (red in A–F, gray in A″–F″) while the size of the dve domain, which affects dorsal head region size, was assessed by GFP expression (dveGal4>GFP, green in A–F, gray in A′–F′). Effects of modulating Hippo activity are shown in (A–D) and quantified in G–I. (A–D) Confocal images of third instar eye imaginal discs from (A) dve-GAL4>UAS GFP (dve>GFP) control, (B) dve-GAL4>UAS GFP, UAS yki (dve>GFP, yki), (C) dve-GAL4>UAS GFP, UAS yki^3SA (dve>GFP, yki^3SA), and (D) dve-GAL4>UAS GFP, UAS yki^RNAi (dve>GFP, yki^RNAi). ELAV (red in A–D, gray in A″–D′) expression marks differentiated photoreceptors and was used to measure eye size. The dve domain is marked by GFP expression (outlined by yellow dotted line in A′–D′). Graphs show quantification of GFP domain size (G), the GFP area/whole disc ratio to assess non-autonomous effects (H), and effect on overall eye size due to effects on MF progression (I).

(E–K) Effects of Dve modulation. Confocal images of third instar eye imaginal discs from (E) dve-GAL4>UAS GFP UAS dve (dve>GFP, dve) and (F) dve-GAL4>UAS GFP UAS dve^RNAi (dve>GFP, dve^RNAi) are shown. ELAV expression (red in E and F; and gray in E″ and F″) marks differentiated photoreceptors, and GFP (green in E and F, gray in E′ and F′ outlined in yellow dotted line) marks the Dve domain. Graph in (J) shows quantification of dorsal head region size (GFP), while graph in (K) shows the GFP area/whole disc ratio to assess non-autonomous effects. For CTCF quantification, n = 5, where n represents 5 different discs, error bars show SEM. Statistical analyses were performed using one-way ANOVA for graphs in (G, H, J, and K). Asterisks indicate significant differences based on differences based on p values: ns = p > 0.05, ∗ = p < 0.05, = p < 0.01, ∗∗∗ = p < 0.001. Scale bars, 100 μM.

Next, we tested the effect of downregulation of Yki on the Dve expression domain (Figures 2D and D″). Compared to wildtype control (dve>GFP, Figure 2A), downregulation of Yki (dve>GFP, Yki^RNAi, Figure 2D) showed no significant effect on the domain size (quantified in Figures 2G and 2H), and mild extension of the MF (Figure 2D″ marked with arrowhead, quantified in 2I). Consistent with this, loss of function clones of null alleles of hpo (FRT42D hpo^42-47/FRT82B cl w+ in Figure S1D) or wts (FRT82B wts^X1/FRT82B cl w+ in Figure S1E) showed expansion of the Dve domain (green, gray in Figures S1D′ and S1E′) and concomitant suppression of the MF at the margins of the eye disc (Figures S1D″ and S1E″ marked with arrowheads). Thus, Dve domain size is significantly affected by gain- or loss of Hippo signaling.

Next, we checked if modulation of Yki binding transcription factor Sd also affected head cuticle domain size or MF progression in the eye disc (Figure S1). Compared to wild-type (dve>GFP, Figure S1F), overexpression of Sd (dve>GFP, Sd, Figures S1G–S1H) showed significant reduction in the Dve domain size (quantified in Figure S1J) whereas RNAi mediated knockdown of Sd (dve>GFP, Sd^RNAi, Figure S1I) showed statistically insignificant differences in the Dve domain size (Figure S1J). In addition, overexpression of Sd showed interesting expansion of the MF at the dorsal (Figures S1G″ and S1H″) and ventral margins (Figure S1H″), a phenotype that is opposite to the effects of Yki overexpression where the MF was suppressed. Together, these data confirmed that the manipulation of the Yki/Sd complex affects both dorsal head region size and MF progression in eye discs.

To assess if changes in domain size are sensitive to Dve itself, we tested if over-expression of Dve (dve>GFP, dve, Figure 2E) or its downregulation (dve>GFP, dve^RNAi, Figure 2F) affects Dve domain size. We noticed an increase in the Dve domain size when we over-expressed Dve (Figure 2E, quantified in Figure 2I), which also resulted in reduced growth of eye imaginal discs which appeared smaller compared to wild type. The Dve overexpressing larvae (dve>GFP, dve) developed into adult flies with smaller heads and reduced or no eyes.²⁸ In contrast, down regulation of Dve (dve>GFP, dve^RNAi, Figure 2F, quantified in Figures 2J and 2K) showed no significant change in Dve domain size or MF progression as compared to wild type. Overall, these data suggest that the differentiation of the photoreceptor neurons and the overall dorsal head region controlled by Dve function in the developing eye imaginal discs are very sensitive to the activity of Hippo effector Yki and the DV patterning gene Dve.

Dve suppresses Yki activity in the dorsal head region

The Hippo pathway effector Yki transcriptionally controls several genes involved in the regulation of organ size, including expanded (ex), death-associated inhibitor of apoptosis 1 (diap1), bantam miRNA, and hth, which are well-characterized Yki/Sd-dependent target genes.¹⁹ First, we checked the expression of an ex-reporter (ex⁶⁹⁷-lacZ)³⁶ in the dorsal head region where Dve is expressed. Compared to control eye discs (dve>GFP, Figure 3A) that show normal expression of ex-lacZ (Figure 3A′), Dve overexpression (dve>GFP, dve) (Figure 3B) showed reduced but not significant differences in the ex-lacZ reporter expression (Figure 3B′) in the dorsal head region (GFP in Figure 3B, quantified in Figure 3D). As expected, overexpression of Yki (dve>GFP, Yki) caused significant upregulation of ex-lacZ expression (Figure 3C, quantified in D). Interestingly, we also found non-autonomous upregulation of ex-lacZ in dve>GFP, Yki (Figure 3C, quantified in E). Next, we tested if Dve exhibits any domain-specific constraints in its ability to modulate Yki activity in the developing eye. To do so, we ectopically expressed Dve using GMR-GAL4, which drives expression in the differentiating retinal neurons that do not normally express Dve during development (Figures 3F–3H). Compared to controls (GMR>, Figure 3F), ectopic expression of Dve did not significantly affect ex-lacZ expression (GMR>dve, Figure 3G, quantified in Figure 3H). Taken together, these data suggest that ex-lacZ expression is not significantly affected by dve overexpression in the eye discs.

Figure 3.

Yki target gene expanded is not significantly affected by Dve

(A–C) Confocal images of third instar eye discs showing expression of the Yki reporter gene, ex⁶⁹⁷-lacZ (red in A–E, gray in A′–E′) are shown: (A) ex⁶⁹⁷-lacZ dve-GAL4 UAS GFP control (ex-lacZ dve>GFP), (B) ex⁶⁹⁷-lacZ dve-GAL4 UAS GFP UAS dve (ex-lacZ dve>GFP, dve), (C) ex⁶⁹⁷-lacZ dve-GAL4>UAS GFP UAS yki (ex-lacZ dve>GFP, yki).

(D) Graph shows quantification of ex⁶⁹⁷-lacZ reporter expression (CTCF), and (E) shows quantification of ex⁶⁹⁷-lacZ reporter expression (CTCF) in the whole disc to assess non-autonomous effects of Yki overexpression.

(F–H) Eye discs from (F) ex⁶⁹⁷-lacZ GMR-GAL4 (ex-lacZ GMR>) and (G) ex⁶⁹⁷-lacZ GMR-GAL4>UAS GFP UAS dve (ex-lacZ GMR>GFP, dve).

(H) Quantification of the ex⁶⁹⁷-lacZ reporter expression (CTCF) is shown in graph in (G). For CTCF quantification, n = 5 where n represents 5 different discs, error bars show SEM. Statistical analyses were performed using one-way ANOVA for graphs (D and E), and an unpaired t test for graph (H). Asterisks indicate significant differences based on p values: ns = p > 0.05, ∗ = p < 0.05, ∗∗ = p < 0.01, ∗∗∗ = p < 0.001. For (A–F) scale bars, 100 μM, for (A′–F′) scale bars, 25 μM.

Next, we tested expression of DIAP1 an important cell survival factor transcriptionally regulated by the Hippo signaling pathway (dve>GFP,Yki, Figure S2B, quantified in E).³⁷ Compared to control (dve>GFP, Figure S2A, quantified in E), discs over-expressing dve (dve>GFP, dve) showed no significant difference in DIAP1 expression (Figure S2C, quantified in E). Taken together, our data suggests that within the dorsal head region (Dve domain), Dve does not significantly affect DIAP1 expression. However, in the photoreceptor cells downregulation of Dve can lead to expansion/enlargement of eyes,²⁸ suggesting that optimum dve and yki activity are required to make organs of proportionate size. Next, we tested Yki target genes that regulate cell death, for example, the microRNA bantam using the GFP-labelled bantam sensor (Figure S2). The bantam-sensor reports inverse profile of miRNA bantam activity such that high GFP levels correspond to low miRNA bantam expression, and vice versa.³⁸ Compared to wild-type eye discs (Figure S2F), overexpression of dve (dve>dve) (Figure S2G) resulted reduced eye size; however, no obvious effect on expression of the bantam sensor was seen (compare Figure S2F′ with Figure S2G′, quantified in Figure S2I). In contrast, knockdown of dve (dve>dve^RNAi) showed enlargement of the photoreceptor domain (Figure S2H) as compared to the control (Figure S2E), and a concomitant expansion of the bantam sensor expression (Figure S2H′). These results show that modulation of dve directly affects photoreceptor differentiation and eye size but does not specifically affect bantam sensor expression.

Hth is a transcription factor that binds with yki and regulates Sd-independent target genes in the eye disc.¹⁹ Hth plays multiple roles as a negative regulator of the eye fate and a positive regulator of proliferation of eye precursor cells.^11,27,39,40 Hth expression normally overlaps the dve-GAL4 domain in dorsal head region (dve>GFP, Figure 4A). Consistent with its role as target gene of Yki, Yki overexpression (dve>GFP, yki) induced Hth expression (Figure 4B, quantified in Figure 4E). Interestingly, Hth is suppressed in discs where Dve (Figure 4C, quantified in Figure 4E) or Dve and Yki are co-expressed (Figure 4D, quantified in Figure 4E). Taken together, these data suggests that Dve acts downstream of Yki and suppresses Yki activity in the dorsal head region.

Figure 4.

Dve suppresses Yki activity in the dorsal head region

Confocal images of third instar eye discs stained for Hth (red in A–D, gray in A″–D″) for the following genotypes: (A) dve-GAL4>UAS GFP (dve>GFP) control, (B) dve-GAL4>UAS GFP, UAS yki (dve>GFP, yki), (C) dve-GAL4>UAS GFP UAS dve (dve>GFP, dve), and (D) dve-GAL4>UAS GFP UAS dve UAS yki (dve>GFP, dve, Yki). The dve expression domain (dveGal4>GFP) marks dorsal head region and is shown in green (A–D) and gray (A′–D′).

(E) Quantification of Hth expression (CTCF) in the indicated genotypes is shown. For CTCF quantification, n = 5 for all quantification where n represents 5 different discs, error bars show SEM. Statistical analyses was performed using one-way ANOVA for graph (E). Asterisks indicate significant differences based on p values: ns = p > 0.05, ∗ = p < 0.05, ∗∗ = p < 0.01, ∗∗∗ = p < 0.001. For (A–D) scale bars, 100 μM, and for (A′–D′) scale bars, 50 μM.

dve is not transcriptionally regulated by Yki in the eye disc

RNAseq studies have reported that dve may be a target of Yki/Sd transcriptional complexes.⁴¹ We, therefore, tested if Yki directly regulates dve transcription in the eye discs. We overexpressed Yki (UAS-yki) and its hyperactivated form (UAS-yki^3SA) in the eye using GMR-GAL4 and bi-GAL4 and tested effects on dve-lacZ expression, a reporter for dve transcription (Figure S3). We mis-expressed yki in the differentiating photoreceptor cells using GMR-GAL4 (Figures S3A–S3C). Compared to controls (Figures S3A and S3B), over-expression of yki (GMR>yki, Figure S3C) did not trigger ectopic expression of dve-lacZ in the GMR domain. bi-GAL4 drives expression of UAS-tagged genes along the dorsal and ventral margins of the eye (Figure S3D, GFP). In the bi>GFP controls (Figure S3D), dve-lacZ expression is restricted to its normal domain in the dorsal head vertex region (Figure S3D′). In comparison, overexpression of Yki (bi>GFP, yki, Figure S3E) or its hyperactivated form (bi>GFP, yki^3SA, Figure S3F) showed the activation of Yki in dorsal or ventral margins (marked by GFP in Figures S3E‴ and S3F‴) does not induce dve-lacZ expression (Figures S3E′ and S3F′). Thus, unlike the wing discs,⁴¹ Yki may not directly regulate dve expression in the eye imaginal discs. Since transcriptional reporter may not fully reflect the in-vivo effects of Yki over-expression, we tested dve expression using RT-qPCR and found that in the eye discs Yki overexpression does not induce dve expression (Figures S3G and S3H). As expected, over-expression of Dve (Figures S3G and S3H) significantly induced dve transcript levels as compared to controls (dveGal4>UASGFP). Taken together, these results suggest that Yki does not transcriptionally regulate dve during eye imaginal disc development.

Dve acts downstream of Yki in regulating dorsal head domain size

To further explore the interactions between Dve and Yki, we performed genetic interaction experiments. We hypothesized that Yki and Dve function to regulate shared transcriptional targets like wg that control growth via Hippo signaling and DV patterning. In a simple epistatic model, Yki may act upstream of Dve, or vice versa. Using multiple GAL4 drivers like dve-GAL4, bi-GAL4, GMR-GAL4 and ey-GAL4 we investigated if different GAL4 drivers affect disc growth. Therefore, we investigated if genetic interactions between Yki and Dve using early (before differentiation) GAL4 drivers (e.g., ey-GAL4) show different effects on disc growth and differentiation as compared to later GAL4 drivers (e.g., dve-GAL4 and GMR-GAL4) (Figure 5).

Figure 5.

Genetic epistasis shows Dve acts downstream of Yki

(A–D) Third instar eye discs (A) ey-GAL4, (B) ey-GAL4 UAS yki-GFP (ey>yki-GFP), (C) ey-GAL4 UAS dve (ey>dve), and (D) ey-GAL4 UAS yki-GFP, UAS dve (ey>Yki-GFP, dve) showing expression of Wg (red in A–D, gray in A′–D′) and Yki tagged with GFP (green in A–D, gray in A″–D″). Images in (A′–D′) and (A″–D″) are higher magnification images of the box highlighted in (A–D).

(E) Graph shows quantification of Yki-GFP (CTCF) in the indicated genotypes.

(F–I) Expression of Wg (red in F–I, gray in F′–I′) in (F) dve GAL4 UAS GFP (dve>GFP), (G) dve GAL4 UAS GFP UAS yki (dve>GFP, yki), (H) dve GAL4 UAS GFP UAS dve (dve>GFP, dve) and (I) dve GAL4 UAS GFP UAS yki UAS dve (dve>GFP, yki, dve). Images in (F′–I′) are higher magnification images of the box highlighted in (F–I).

(J–M) Expression of Wg (red in J–M, gray in J′–M′) in (J) GMR GAL4 (GMR>), (K) GMR GAL4 UAS yki (dve>yki), (L) GMR GAL4 UAS dve (GMR>dve), and (M) GMR GAL4 UAS GFP UAS yki UAS dve (GMR>yki, dve). Images in (J′–M′) are higher magnification images of the box highlighted in (J–M).

(N and O) Graphs show quantification of wg expression from RT-qPCR when over-expressing Yki or Dve or both with dve-GAL4 and GMR-GAL4. For CTCF quantification, n = 5 for all where n represents 5 different discs, error bars show SEM. Statistical analyses were performed using an unpaired t test for graph in (E), and one-way ANOVA for graphs in (N and O). Asterisks indicate significant differences based on p values: ns = p > 0.05, ∗ = p < 0.05, ∗∗ = p < 0.01, ∗∗∗ = p < 0.001. For RT-qPCR, n = 3 represents three technical repeats (three independent cDNA samples prepared for each sample) independently collected from three biological replicates. Each sample was tested in triplicate for RT-qPCR measurements. For (A–D), (F–M) scale bars, 100 μM, and for (A′–D′), (F′–M′) scale bars, 50 μM.

The ey-GAL4 driver causes mis-expression of UAS-transgenes very early in the embryonic eye primordium when ey expression is turned on.⁴² At this early time point transgenes are expressed at the end of embryogenesis in the eye disc primordium prior to axial patterning and MF progression that later delineate patterning and differentiation events. Using ey-GAL4 (Figures 5A and 6A), we over-expressed a GFP-tagged version of Yki in ey-GAL4 UAS-yki-GFP (ey>yki-GFP in Figure 5B), or Dve in ey-GAL4 UAS-dve (ey>dve in Figure 5C) or both in ey-GAL4 UAS yki-GFP, UAS dve (ey>yki-GFP, dve in Figures 5D and 6A) eye discs, and tested effects on growth and differentiation in mature wandering third instar eye antennal imaginal discs. Compared to control (Figure 5A), over-expression of Yki-GFP caused over-growth of the eye disc along with suppression of MF in the dorsal and ventral regions resulting in delay in MF progression (Figure 5B, quantified in Figure 6A). Over-expression of Dve (Figure 5C), on the other hand, caused a significant decrease in eye growth and differentiation (quantified in Figure 6A). Several discs lacked ELAV positive photoreceptors suggesting a complete suppression of MF initiation in these eye discs (Figure 5C). Interestingly, co-expression of Yki-GFP and Dve (Figure 5D), showed similar phenotypes in which the growth of the disc was significantly reduced (quantified in Figure 6A), and no photoreceptor differentiation was observed. Further, Yki-GFP expression was downregulated (Figure 5D GFP, D″ gray, quantified in E). Thus, early in eye development, misexpression of Dve disrupts both the growth and differentiation of eye discs (Figure 6A) suggesting that Dve may function downstream of Yki.

Figure 6.

Effects on eye disc area and Wg expression in genetic epistasis interactions

Quantification of average eye disc area following overexpression of either Yki or Dve or both using (A) ey-GAL4, (B) dve-GAL4 UAS GFP, and (C) GMR-GAL4.

(D–F) Graphs showing quantification of wg expression (CTCF) upon over-expression of Yki or Dve or both with ey-GAL4, dve-GAL4 and GMR-GAL4. Graphs in (D–F) show average fold-changes from three independent experiments, with all samples tested in triplicate for each run. For quantification, bar scatterplots were generated (n = 5), and error bars represent SEM. Statistical analyses were performed using one-way ANOVA for graphs (A–F). Asterisks indicate significant differences based on p values: ns = p > 0.05, ∗ = p < 0.05, ∗∗ = p < 0.01, ∗∗∗ = p < 0.001.

Next, we tested if a similar effect was observed using dve-GAL4 which drives gene expression beginning in the second-instar larva stage, when DV patterning occurs. Using dve-GAL4, we over-expressed UAS-yki (dve>yki), UAS-dve (dve>dve) or both (dve>yki, dve) (Figures 5F–5I, quantified in Figure 6B) and tested effects on overall patterning and growth in mature third instar eye imaginal discs. We observed a high percentage of pupal-lethality in these experiments, for example, overexpression of yki showed 28% lethality whereas dve or yki, dve overexpression both showed about 38% pupal lethality. Compared to control (Figure 5F), we observed that over-expression of Yki caused expansion of the dve domain (head primordium) and mild suppression of MF progression along the dorsal and ventral margins as reported earlier (Figure 5G).⁴⁰ Over-expression of Dve on the other hand, caused a significant increase in the dve domain size and a concomitant decrease in eye growth (Figure 5H). Interestingly, co-expression of Yki and Dve (Figure 5I) produced phenotypes reminiscent of Dve-misexpression alone, and the overall Dve domain was reduced compared to Yki overexpression alone (Figure 5I, quantified in Figure 6B). Thus, Dve acts downstream of Yki to regulate growth and differentiation.

Next, we assessed the phenotypes using GMR-GAL4 which drives expression of UAS-linked transgenes from early third-instar larva stage, when the MF is formed, and photoreceptor differentiation begins. We over-expressed UAS-yki (GMR>yki), UAS-dve (GMR>dve) or both (GMR>yki, dve) (Figures 5J–5M) and tested effects in mature wandering third instar eye antennal imaginal discs. Compared to control (Figure 5J), over-expression of Yki caused increased disc size with mild suppression of MF progression in the dorsal and ventral eye disc margins (Figure 5K), and resulted in 100% viable adult flies at eclosion that show enlarged eyes as previously reported.⁴⁰ Overexpression of Dve caused reduction in imaginal disc growth and suppressed differentiation (Figure 5L) resulting in formation of smaller eyes with pigmentation defects in adults (quantified in Figure 6C). Overexpression of dve by GMR-GAL4 resulted in 87.75% lethality at late pupal stages, while 12.24% of flies that survived at eclosion showed smaller eyes. Interestingly, co-expression of Yki and Dve showed 53.38% pupal lethality and dve significantly suppress the enlarged eye phenotype caused by yki overexpression (Figure 5M). Taken together, these data support a model in which Dve acts downstream of Yki and negatively regulates Yki activity and its effects on disc growth. Further, using early and later GAL4 drivers, we found that the growth of disc in the early developmental time points (ey-GAL4) is weakly suppressed by Dve (Figures 6A–6C). This suggests that Yki is required for early disc growth and proliferation to generate a pool of cells that later undergo differentiation.

In these experiments, we also quantified the expression of Wg, a transcriptional target of both Dve and Yki.^28,40 Interestingly, Yki moderately induced Wg expression when driven by ey-GAL4 or dve-GAL4 or GMR-GAL4 (Figures 5B, 5G, and 5K, quantified in Figures 6D–6F) whereas, Wg levels were strongly induced by Dve overexpression in all experiments (Figures 5C, 5H, and 5L, quantified in Figures 6D–6F). Consistent with the phenotypes described above, when compared to wild-type Wg expression was significantly elevated in all samples in which Yki and Dve were co-expressed (Figures 5D, 5I, and 5M), further suggesting that Dve may act downstream of Yki in this context. Regulation of Wg was further validated by RT-qPCR (Figures 5N and 5O). Compared to control (Figures 5N and 5O), overexpression of Yki by dve-GAL4 (Figure 5N) or GMR-GAL4 (Figure 5O) showed significant induction of wg expression. Similarly, overexpression of Dve in its native domain (dve>dve) caused significant upregulation of wg (Figure 5N). In contrast, overexpression of Dve outside of its native domain using GMR-GAL4 did not significantly induce wg (Figure 5O) suggesting a spatially restricted effect of Yki-mediated and Dve-mediated wg regulation. Together these data indicate that Dve acts downstream of or parallel to Yki and affects MF progression, dorsal head region versus eye fate by regulating dorsal head size, and overall patterning of the eye-antennal disc.

Transcriptional control of DV and Hippo targets during Dve-Yki interactions

The data so far suggest a model where Yki is required for growth and early proliferation of cells in the eye disc, while Dve is expressed in a small group of cells in the dorsal head region and affects both the head cuticle fate and Wg expression. Through these activities, Dve regulates MF progression and the overall number of cells differentiating into photoreceptors. In the eye disc, Yki is ubiquitously expressed and partners with transcription factors such as Sd or Hth/Tsh, whereas Dve is a K50-domain transcription factor that regulates genes in the DV patterning pathway. Given the interactions we observed, we asked if Yki could regulate expression of other genes in the DV patterning pathway, and if Dve could regulate other Yki targets besides Wg. We selected pannier (pnr), an upstream regulator of Dve, and mirror (mirr) a member of the Iro-C complex that acts downstream of Wg in DV patterning in the eye.^2,24 For assessing Yki targets, we selected diap1, an ubiquitously expressed Yki-target gene.³ Using RT-qPCR approach, we compared expression levels of these transcriptional targets when Dve or Yki or both were over-expressed in the dorsal head region defined by dve-GAL4, or in the differentiating photoreceptor neurons defined by GMR-GAL4 (Figure 7). We found no significant differences in diap1 expression levels (Figures 7A and 7B) when using dve-GAL4 we overexpressed Yki (dve>GFP, yki) or Dve (dve>GFP, dve) alone or together (dve>GFP, yki, dve) (Figure 7C). In contrast, using GMR-GAL4, diap1 levels were significantly induced by Yki (GMR>yki) but not by Dve (GMR>dve). In this domain, Dve did not affect Yki-mediated diap1 expression (GMR>yki, dve) (Figure 7D). These results indicate that Yki activity is differentially regulated by Dve in a domain-dependent manner.

Figure 7.

pannier and mirror are Yki-regulated genes during eye development

RT-qPCR analysis was performed to examine relative changes in expression of (A and B) diap1, (C and D) pannier and (E and F) mirror upon overexpression of Yki or Dve or both, under dve-GAL4 (A, C, and E) or GMR-GAL4 (B, D, and F). Graphs show average fold-changes from three independent experiments, with all samples tested in triplicate for each run. For quantification, bar scatterplots were generated (n = 5) and error bars show SEM. Statistical analyses were performed using one-way ANOVA for graphs A–F. Asterisks indicate significant differences based on p values: ns = p > 0.05, ∗ = p < 0.05, ∗∗ = p < 0.01, ∗∗∗ = p < 0.001.

We next tested expression of genes in the DV patterning cascade.² We tested the expression of mirror (mirr) in samples from the dve-GAL4 set. Compared to control (dve>GFP), over-expression of Yki (dve>GFP, yki) had no effect on mirr expression, whereas over-expression of Dve (dve>GFP, dve) caused a significant upregulation of mirr expression (Figure 7E). Expression of mirr remained significantly high when Dve and Yki were co-expressed (dve>GFP, yki, dve) (Figure 7E). Thus, in the Dve domain, Dve but not Yki can induce mirr expression (Figure 7E). In contrast using GMR-GAL4, mirr expression was induced in all combinations, suggesting that outside of the Dve domain Yki can upregulate mirr (Figure 7F). In DV patterning, pnr is expressed in the dorsal eye in the peripodial membrane in a pattern that partially overlaps Dve, and acts upstream of Dve.²⁸ Interestingly, Yki overexpression robustly induced pnr expression using either dve-GAL4 (dve>GFP, yki, Figure 7C) or GMR-GAL4 (GMR>Yki, Figure 7D). Overexpression of Dve did not affect pnr expression in both experimental settings (Figures 7C and 7D), and co-expression of Dve with Yki did not modify Yki-mediated upregulation of pnr (Figures 7C and 7D). This result suggests that pnr is a new Yki target gene; however, chromatin-level evidence (e.g., ChIP-seq based assays) would be required to confirm it as a direct target of Yki. Taken together, these data indicate the Yki can transcriptionally induce multiple genes like pnr, wg, and mirr in the DV patterning hierarchy including pnr, wg and mirr.

To test if pnr in turn affects Yki-Dve interactions, we tested dominant genetic interactions between Yki and pnr. Compared to control (dve>GFP) (Figure 8A, quantified in Figure 8E), loss of one copy of pnr (dve>GFP; pnr^VX6/+) did not modify the size of the Dve-GAL4 expression domain (Figure 8B, quantified in Figure 8E). However, the increase in dve domain size by Yki overexpression (dve>GFP, yki, Figure 8C, quantified in Figure 8E) is significantly reduced by loss of one copy of pnr (dve>GFP,yki; pnr^VX6/+) (Figure 8D, quantified in Figure 8E). These results suggest that Yki and Pnr genetically interact, and Yki activity is differentially regulated the dorsal head cuticle domain versus in the photoreceptor cells. Further, these findings indicate that maintaining appropriate Pnr levels is important for dorsal identity, DV patterning, and proper growth of the eye disc.^43,44 Taken together, rescue of Yki gain of function by pnr loss of function supports an epistatic relation where pnr acts downstream of Yki and regulates Dve expression/function.

Figure 8.

Yki genetically interacts with pnr, a DV patterning genes

Confocal images of third instar eye imaginal discs stained for ELAV (red in A–D, gray in A′–D′) from larvae of the following genotype: (A) dve-GAL4 UAS GFP (dve>GFP) control; (B) dve-GAL4 UAS GFP UAS yki (dve>GFP, yki), (C) dve-GAL4 UAS GFP UAS yki FRT82B pnr^VX6/+ (dve>GFP, yki, pnr^VX6/+) and (D) dve-GAL4 UAS GFP FRT82B pnr^VX6/+ (dve>GFP, FRT82B pnr^VX6/+).

(E) Quantification of dorsal head region size based on GFP expression is shown in the graph.

(F) Model proposed for the interactions of Hippo and DV patterning pathways. Independent and shared functions of the two pathways are shown. During development and patterning of the eye imaginal disc, the Hippo pathway effector Yki interacts with pnr and dve in the DV patterning pathway to promote dorsal head region domain size. For quantification, bar scatterplot was generated (n = 5), and error bars represent SEM. Statistical analyses was performed using one-way ANOVA for graph (E). Asterisks indicate significant differences based on p values: ns = p > 0.05, ∗ = p < 0.05, ∗∗ = p < 0.01, ∗∗∗ = p < 0.001. Scale bars, 100 μM.

Discussion

Regulation of organ size is controlled by coordination of cell proliferation, survival and growth. The Drosophila eye imaginal disc is an elegant model for studying the coordination of growth with differentiation during development due to the well-documented regulation of cell proliferation, developmental apoptosis and patterning from embryogenesis to adulthood.^2,8,10,45,46 Eye discs are specified in late embryogenesis and expand during larval stages, reaching ∼50,000 cells by the mature third instar, when photoreceptors differentiate posterior to the MF and anterior cells give rise to the head cuticle, antenna, and associated structures.^7,11,14,46 During pupal development, extensive tissue remodeling—including head eversion and formation of adult head structures—further refines these domains.⁷ Notably, both the mature third instar and early-to-mid pupal stages are especially sensitive to developmental perturbations that can manifest as overt morphological abnormalities in the adult.

During organogenesis, successful patterning of the eyes, ocelli, antennae, and maxillary palps requires precise integration of multiple signaling pathways, including Wg, Decapentaplegic (Dpp), Hedgehog (Hh), the retinal determination network, and growth-regulatory mechanisms such as the Hippo pathway. Specifically, Yki functions in early retinal differentiation,⁴⁰ and as shown in this study, in regulating dorsal head domain size via Dve interactions. Our findings thus illustrate how a single transcriptional co-activator can drive distinct outcomes by engaging with different molecular partners. In the context of eye development, DV patterning is the first lineage restriction event controlled by the expression of Pannier (Pnr) which lies upstream of the DV patterning hierarchy. We recently showed that Pnr acts upstream of and regulates the expression of the K-50 domain transcription factor Dve.²⁸ Thus, dve and pnr can have some overlapping functions in dorsal fate determination but Dve mediated Wg regulation is different from Pnr.²⁸ Further, Pnr driven Wg regulation in dorsal eye and head region does not represent the complete spectrum of its function.²⁸ Specifically, the current study reveals that Yki interacts with Pnr to regulate dorsal head cuticle development, whereas it influences the MF progression and retinal differentiation by interacting with Wg. Loss of dve or other dorsal genes (e.g., pnr or Iro-C complex) in the DV patterning pathway cause dorsal eye enlargements due to increased growth.^{2,23,24,43,44} Gain of function of Dve or Pnr results in reduced growth and smaller head/eye domains. Similar effects on organ size are well-documented for the Hippo pathway, where loss of function of Hippo signaling results in enlargement of structure due to over-growth caused by increased proliferation and reduced apoptosis; while gain-of-function of Hippo signaling results in formation of smaller structures due to increased apoptosis.³ Here, we report that in the context of the eye imaginal discs, the Hippo pathway regulates the head cuticle domain (Figures 1, 2, and S1) in addition to its previously reported role in regulating MF progression.^17,40 Further, we show that loss of function of Hippo pathway results in expansion of Dve expression and the head cuticle domain associated with the distinct “large-headed” phenotype (Figures S1C–S1E) identified through mosaic genetic screens that uncovered upstream components of the Hippo pathway.^3,47 Thus, both the DV patterning and Hippo pathway genes play a significant role in organogenesis of the head.

In Drosophila, eye fate is antagonistic to the fates of other domains in the eye antennal disc during patterning, as seen by antagonistic interactions between eye and antenna fate,⁴⁸ and Pnr-regulated dorsal eye margin and eye fate.^26,27 Similar antagonistic interactions are evolutionarily conserved in mammalian models. For example, antagonistic interactions between Wnt and Hh pathway, and TGFb and BMP pathways regulate retinal development and differentiation. Similarly Srbep2 and Lrp2 regulate eye size, and Six3 and Prox1 regulate lens development by acting antagonistically.^49,50,51 Together, these examples illustrate that antagonistic interactions between developmental signaling pathways are evolutionarily conserved mechanisms for head and eye development across species.

In addition to the developmental patterning pathways, imbalance in the activity of growth regulatory pathways also differentially affects eye versus head cuticle fate. Activation of Yki either through over-expression of yki^3SA, or loss of hpo or wts, leads to expansion of the head cuticle domain and suppression of eye fate (Figures 2 and S1). The suppression of eye phenotype is due to Yki mediated suppression of retinal differentiation by blocking MF progression.⁴⁰ Interestingly, this effect is suppressed by overexpression of Dve, suggesting an antagonistic interaction between Yki and Dve (Figure 1). Over-expression of Dve suppresses Yki activity as seen by downregulation of DIAP1 (Figure S2), and depletion of Yki does not directly affect Dve domain size (Figure 2). Further, Sd also shows antagonism to Dve regulated head cuticle domain (Figure S1). Overexpression of Sd causes reduction in Dve domain size (Figure S1), a phenotype stronger than Yki^RNAi (Figure 2) and opposite of Yki over-expression (Figure 2) likely due to its roles as a default repressor. Further, loss of function of sd (hemizygous mutants) reduces Yki mediated growth⁵² suggesting its role in essential survival. Taken together, genetic interactions studies place Dve downstream of Yki, functioning as a suppressor of Yki activity.

Furthermore, the role of Dve in delimiting spatial domains (eye versus antenna or eye versus head) produces interesting phenotypes that extend well beyond the Dve expressing dorsal head vertex domain. For example, we found that Dve exerts differential effects on Yki activity, depending on the spatial context. Over-expression of Dve within its endogenous domain resulted in suppression of Yki activity as seen by altered expression of Hth (Figure 4), whereas expression of other Yki targets, like ex or DIAP1, was not significantly affected (Figures 3 and S2). These data suggest that boundary-dependent or antagonistic interactions determine a balance between activation and suppression of Hippo pathway coactivator in dve-expressing versus non-expressing cells. This spatially restricted regulation results in differential Dve-mediated regulation of Yki activity. In contrast, depletion of Dve (dve>dve^RNAi) showed no significant effects on Dve domain size or MF progression (Figure 2). This could be due to the weak RNAi effect, as we have previously shown that loss-of-function clones of a strong hypomorphic alleles dve¹ do not survive in the presence of wild-type twin clones, and show dorsal eye enlargements only when the wild-type sister-clones are eliminated by cell-lethal mutations,²⁸ suggesting a role of dve for essential survival. Together, these interactions play a major role in regulating optimal eye and head size during development.

In a model of wound induced regeneration in wing discs, dve is upregulated by Yki overexpression and downregulated by yki knockdown, suggesting that dve is a target of Yki.⁴¹ Consistently, RNAseq data identified dve as a potential Yki target in wing discs.⁴¹ Yki associates with some of its transcriptional targets through WW-PPxY interactions, for example, the PPxY domain of Ex associates with the WW-domain of Yki.^1,3 However, Dve lacks a PPxY motif suggesting that it does not directly interact with Yki via this mechanism. As a transcriptional co-activator Yki lacks a DNA-binding domain and instead associates with sequence specific DNA-binding transcription factors to regulate downstream targets. In the eye-antennal disc, Yki partners with transcription factors like Sd which is expressed in the photoreceptor domain,^52,53 and Hth or Tsh, which are expressed anterior to the MF and in the antenna, ptilinum, and ocellus. These factors define the boundary of the eye field and the head capsule in the eye disc.^11,13,15,39 Using a dve-lacZ reporter³¹ and RT-qPCR analysis, we found that dve is not directly regulated by Yki at the transcription level during eye development (Figure S3). These data suggest that dve regulation of Yki is tissue- and context-specific, while dve is transcriptionally regulated by Yki during wing disc regeneration, dve is not regulated by Yki during patterning and growth of eye antennal disc during development.

The development of head cuticle is a critical process for survival after metamorphosis. DV patterning requires sequential and regulated expression of several genes in cells that differentiate into the head cuticle, ocelli and antennae. Several signaling pathways like Hedgehog, Notch, and DV (Pnr/Dve) patterning during larval stages,^2,3,12 and Osiris gene family during pupal stages,⁵⁴ play essential roles in ensuring normal head development. These pathways act in coordination with growth-regulatory pathways throughout development. Yki partners with Sd or Hth or Tsh to regulate growth-promoting activities by controlling the expression of target genes involved in organ growth, cell proliferation (e.g., cyclin E, cyclin B, others), and cell survival (diap1, bantam miRNA). Hth maintains proliferation of uncommitted retinal progenitor cells in the eye disc,¹³ and together with transcriptional partners like Yki regulates miRNA bantam and dpp expression.^17,18,55 Yki can also partner with other transcription factors to regulate diverse biological processes, for example, Foxo in response to oxidative stress,⁵⁶ and GAGA factor (GAF) and Brahma (BRM) complex for chromatin remodeling.⁵⁷ Further, transcriptional repressor proteins like E2F⁴¹ or Tondu-domain-containing growth inhibitor (Tgi)¹ can compete with Yki for binding with Sd thereby modulating Yki transcriptional activity. Mammalian YAP and TAZ also bind to several transcription factors (e.g., TEAD1-4, p73, Runx2, β-catenin, SMADs) to regulate growth-promoting and oncogenic activities.^22,58 The large repertoire of genes regulated by Yki/YAP in multiple biological contexts led us to ask if additional genes within the DV patterning hierarchy might be regulated by Yki. Using genetic interactions (Figures 5 and 8) and RT-qPCR assays (Figure 7), we found that pnr and mirr (member of the Iro-C complex) are transcriptionally regulated by Yki. In addition, wg is a shared target of both Yki and Dve during the growth and patterning of the eye discs (Figures 5 and 6). Interestingly, GMR>yki strongly induces pnr which acts upstream of Dve (Figure 7D) and mirr which acts downstream of Dve in the DV patterning pathway (Figure 7F). However, dve>yki shows strong induction of pnr, but not mirr (Figure 7E). This is likely because in dve>yki, Dve suppresses Yki activity causing statistically insignificant induction of mirr (Figure 7E). Taken together, our data show multiple regulatory interactions between the Hippo pathway effector Yki and components of the DV patterning pathway that are essential for coordinated growth and patterning of the eye and head in flies.

The interactions among yki, dve, pnr and wg are particularly significant given the conserved roles of their mammalian orthologs YAP1/2, SATB1/2, GATA3/4/5 and Wnt5 in development and disease.^{59,60,61,62,63,64} YAP–TEAD–GATA3 signaling is essential for trophoblast and trophectoderm specification across species, and aberrant YAP activity is associated with defects in early embryonic development.^65,66 SATB1 and SATB2 regulate multiple developmental processes, including neurogenesis and craniofacial patterning, and mutations in these genes result in severe neurodevelopmental and craniofacial abnormalities like dysmorphic facial features including brachycephaly (a flat head shape), cleft palate, prominent forehead, hypertelorism (widely spaced eyes), and low nasal bridge.^{67,68,69,70,71,72,73,74} Both factors are also implicated in tumor progression, where elevated SATB1/2 expression correlates with aggressive behavior and metastasis.⁷⁵ Moreover, YAP-SATB2 interactions have been documented in cortical neuron development and in renal cell carcinoma, where YAP/TEAD4 drives SATB2-dependent proliferation.^74,75,76,77 Together, these conserved relationships highlight the relevance of our findings and provide a framework for exploring how growth and patterning pathways integrate to regulate head development.

Limitations of the study

Our findings from genetic interactions studies support a model where Dve and Yki interact to regulate dorsal head region. However, despite these findings, several limitations remain. First, do Dve and Yki act in a linear model where Dve acts to suppress Yki activity or if Dve acts in parallel to refine the dorsal head region needs to be confirmed with additional molecular studies. Second, our findings do not distinguish whether Dve directly represses Yki-dependent transcription, indirectly alters tissue identity, or changes the cellular context in which Yki acts. In this context, suppression of the ey>Yki phenotype by Dve co-expression may simply reflect reduced expression of the Yki-GFP transgene, or Yki-GFP protein may become unstable under double-overexpression conditions, rather than physiological inhibition of Yki activity by Dve. The current data do not clearly distinguish between these possibilities. It will be interesting to elucidate the molecular mechanisms, for example, Yki/Sd or Yki/Hth target occupancy for transcriptional activity in the Dve domain to confirm direct molecular suppression mechanism or indirect effects via changes in tissue context or cell identity. Third, pnr is proposed as a new Yki target gene; however, chromatin-level evidence (e.g., ChIP-seq based assays) would be required to confirm it as a direct target of Yki. Fourth, sex-associated effects were not formally assessed. Further elucidation of the molecular mechanisms underlying these interactions will refine our understanding of coordinated tissue growth across species.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Madhuri Kango-Singh (mkangosingh1@udayton.edu).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact without restriction.

Data and code availability

• •All data reported in this study will be shared by the lead contact upon request.
• •This study does not report original code.
• •Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request.

Acknowledgments

The authors would like to thank the members of the Kango-Singh and Singh laboratory for their support and helpful discussions. For monoclonal antibodies used in this study, we thank the Developmental Studied Hybridoma Bank (DSHB), created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA 52242. Stocks obtained from the Bloomington Drosophila Stock Center (NIH P40OD018537) were used in this study. Work in MKS lab and AS labs is supported by the National Institutes of Health (1R01EY032959-01, MPIs AS and MKS). MKS lab is also supported by funds from the Schuellein Chair in Biological Sciences, and work in AS lab is supported by Leonard Mann Chair in Natural Sciences. AR and NG were supported by the Graduate Teaching Assistantship from University of Dayton.

Author contributions

Conceptualization, M.K.-S. and A.S.; methodology, M.K.-S., A.S., B.N.R., and N.G.; investigation, B.N.R. and N.G.; validation, B.N.R. and N.G.; formal analysis, B.N.R., A.R., and N.G.; validation, B.N.R. and N.G.; funding acquisition, M.K.-S. and A.S.; project administration and supervision, M.K.-S.; writing – original draft, B.N.R., N.G., A.R., A.S., and M.K.-S.; writing – review and editing, M.K.-S.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
rat anti-Elav (1:100)	Developmental Studies Hybridoma Bank, Iowa University, Iowa USA	7E8A10; RRID: AB_528218
mouse anti-Wg (1:100)	Developmental Studies Hybridoma Bank, Iowa University, Iowa USA	4D4; RRID: AB_528512
mouse anti-βGAL (1:100)	Developmental Studies Hybridoma Bank, Iowa University, Iowa USA	40-1a; RRID: AB_528100
mouse anti-DIAP1 (1:200)	Dr. Bruce Hay	N/A
rabbit anti-Dve (1:1000)	ThermoFisher	This study
rabbit anti-Hth (1:250)	ThermoFisher	This study
mouse anti-FITC (1:250)	Jackson Immuno-Research Labs, West Grove, PA, USA	715–096-150; RRID: AB_2340795
mouse anti-Cy3 (1:250)	Jackson Immuno-Research Labs, West Grove, PA, USA	715–166-151; RRID: AB_2340817
rabbit anti-Cy5 (1:250)	Jackson Immuno-Research Labs, West Grove, PA, USA	711–605-152; RRID: AB_2492288
mouse anti-Cy5 (1:250)	Jackson Immuno-Research Labs, West Grove, PA, USA	715–606-151; RRID: AB_2340866
rat anti-Cy5 (1:250)	Jackson ImmunoResearch Labs, West Grove, PA, USA	712–606-153; RRID: AB_2340696
Biological samples
Drosophila melanogaster mutants and transgenes	N/A	N/A
Chemicals, peptides, and recombinant proteins
10X PBS Stock (made with the following chemicals):
Sodium Chloride	Fisher Scientific	CAS 7647–14-5
Potassium Chloride	Ricca Chemical Company	RDCP0380
Potassium Phosphate, Monobasic	EMD Chemicals	PX1565
Sodium Phosphate, Dibasic	Fisher Scientific	CAS 7558–79-4
16% Paraformaldehyde	Electron Microscopy Sciences	Cat No.# 15710
Normal Donkey Serum	Jackson Immuno Research	Cat No.# 017–000-121; RRID: AB_2337258
Vectashield	Vector Labs	Cat No.# H-1000; RRID: AB_2336789
TRIzol reagent	Ambion Life Technologies	Cat No.# 15596026; RRID: SCR_15596026
Critical commercial assays
ZYMO RNA Clean and Concentrator Kit	ZYMO Research	Cat No.#R1050
First Strand cDNA synthesis kit	Cytiva, Danaher Corporation, Washington, DC, USA	Cat No.#27926101
iQ-SYBER Green Supermix	Bio-Rad laboratories, Hercules, CA, USA	Cat No.# 1708882
Experimental models: Organisms/strains
Drosophila melanogaster/w, bi-GAL4	Lecuit et al., 199678	NA
Drosophila melanogaster/w; dve-GAL4	Nakagoshi et al.31	NA
Drosophila melanogaster/w; ey-GAL4	Hazlett et al., 199842	NA
Drosophila melanogaster/w; GMR-GAL4	Moses and Rubin16	NA
Drosophila melanogaster/yw; FRT42D dve1/CyO	Nakagoshi et al.31	NA
Drosophila melanogaster/yw; FRT42D hpo42-47/CyO	Wu et al., 200379	NA
Drosophila melanogaster/eyFLP; FRT42D cl w+/CyO, y+	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC # 5617; RRID:BDSC_5617
Drosophila melanogaster/w; +; FRT82B wtsX1/TM6B	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC# 44251; RRID:BDSC_44251
Drosophila melanogaster/yw eyFLP; FRT82B cl w+/TM6B	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC# 43657; RRID:BDSC_43657
Drosophila melanogaster/yw; FRT82B pnrVX6/TM6B	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC #6334; RRID:BDSC_6334
Drosophila melanogaster/w; UAS GFP	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC# 4775: RRID:BDSC_4775
Drosophila melanogaster/w; UAS dve	Nakagoshi et al.31	NA
Drosophila melanogaster/w; UAS dveRNAi	Nakagoshi et al., 199831	NA
Drosophila melanogaster/w; UAS hpo	Udan et al., 200380	NA
Drosophila melanogaster/w; UAS wts13F	Kwon et al., 201581	NA
Drosophila melanogaster/w; UAS yki	Huang et al.37	NA
Drosophila melanogaster/w; UAS yki-GFP	Oh and Irvine35	NA
Drosophila melanogaster/w; UAS ykiRNAi	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC# 31965; RRID: BDSC_31965
Drosophila melanogaster/w; UAS sd	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC# 9373; RRID: BDSC_9373
Drosophila melanogaster/w; UAS sdRNAi	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC # 29352; RRID: BDSC_29352
Drosophila melanogaster/w; dve1-lacZ/CyO	Bloomington Drosophila Stock Center, Bloomington (IN)	BDSC# 11073; RRID: BDSC_11073
Drosophila melanogaster/w; ex697-lacZ/CyO	Boedigheimer and Laughon36	RRID: BDSC_44248
Drosophila melanogaster/w; bantam-sensor GFP	Thompson and Cohen38	NA
Oligonucleotides
wg Forward primer: 5′ CGT CAG GGA CGC AAG CAT A 3′	Integrated DNA technology	Waghmare et al. #1
wg reverse primer: 5′ ATT GTG CGG GTT CAG TTG G 3′	Integrated DNA technology	Waghmare et al. #1
diap1 Forward primer: 5′ CGC ACT ACC AAC AAT GTG C 3′	Integrated DNA technology	NCBI Primer-BLAST
diap1Reverse primer: 5′ GTA GGC ATG CTC CCT CAT CT 3′	Integrated DNA technology	NCBI Primer-BLAST
pnr Forward primer: 5′CCA TCC AAC CGA CAG TAC AA 3′	Integrated DNA technology	NCBI Primer-BLAST
pnr Reverse primer: 5′GGC GTT TGG CGA GTA GAA 3′	Integrated DNA technology	NCBI Primer-BLAST
mirr Forward primer: 5′TAA CGA TCT GTT AGA CGC CAA G 3′	Integrated DNA technology	NCBI Primer-BLAST
mirr Reverse primer: 5′CAC TCG CTG TTG TTG CTT TC 3′	Integrated DNA technology	NCBI Primer-BLAST
dve Forward primer: 5′CAC CGC TGG ATT GTC CGA TA 3′	Integrated DNA technology	NCBI Primer-BLAST
dve reverse primer: 5′GGA GTT GGG AAC CAT AGG AAA G 3′	Integrated DNA technology	NCBI Primer-BLAST
GAPDH Forward primer: 5′ TAA ATT CGA CTC GAC TCA CGG T 3′	Integrated DNA technology	Waghmare et al. #1
GAPDH Reverse primer: 5′ CTC CAC CAC ATA CTC GGC TC 3′	Integrated DNA technology	Waghmare et ak #1
Software and algorithms
GraphPad Prism V9.5	Graph Pad	RRID: SCR_002798
Olympus FV3000 with FV10-ASW 4.0 software	Olympus	RRID: SCR_017015; RRID: SCR_024433
Axioimager.Z1 Zeiss Apotome with MrC5 color camera	Zeiss
Adobe Photoshop 2025	Adobe	RRID: SCR_014199
Fiji/ImageJ	NIH	RRID: SCR_002285/RRID: SCR_003070

Experimental model and study participant details

This study was done using Drosophila melanogaster (mutants and transgenic flies) as the model system. The flies (controls and experimental) were maintained on standard cornmeal, agar, molasses-based medium at 25°C unless otherwise specified. Both males and female flies (larvae and adults) were analyzed in the study, and no significant differences were found between sexes for the genetic interactions described in this manuscript.

Method details

Drosophila genetic interaction studies

Appropriate genetic crosses were done to assess genetic interactions using the GAL4-UAS system. Eye imaginal discs were dissected from wandering third-instar larvae to assess changes in dorsal head region (using dve>GFP) or changes in gene expression using immunohistochemistry or qRT-PCR.

Immunohistochemistry

Third instar eye imaginal discs were dissected in PBS and fixed in 4% paraformaldehyde fixative for 20 min. Fixed discs were washed in 0.3%PBTX for 10min, three times and blocked in 2% PBTN (0.3% PBTX and 2% Normal Donkey Serum). Samples were then incubated overnight at 4°C in the primary antibodies, washed with 0.3% PBTX 20min three times, and incubated with appropriate secondary antibodies at room temperature for 2 h. The samples were subsequently washed in 0.3%PBTX for 10min, three times and mounted in Vectashield. The secondary antibodies were donkey Fab fragments conjugated to FITC, Cy3 or Cy5 against rabbit, mouse, or rat hosts (Jackson ImmunoResearch Labs, West Grove, PA, USA).

Image acquisition and analysis

Confocal microscopy

The eye antennal discs were imaged with a confocal laser scanning microscope (Olympus FV3000) with FV10-ASW 4.0 software, at 1024 × 1024 pixels per image, using Plan-Apochromat 20× and 40× objectives.

Bright field microscopy

Adult eyes were imaged at 10× using MrC5 color camera mounted on a Axioimager.Z1 Zeiss Apotome using the z-sectioning function of Axiovision software 4.6.3. Adult flies were frozen at −20°C for ∼4h and mounted on a needle for imaging.

All imaginal discs, and adult flies are imaged at identical magnification, and IHC stains imaged under identical confocal settings. The images were further processed and analyzed using Adobe Photoshop 2025 and Fiji/ImageJ (NIH). The disc orientation is dorsal up and anterior to the right in all figures.

Image analysis

The Z stacks obtained from confocal imaging were analyzed by Fiji/ImageJ software to quantify area, and fluorescence intensity. Area of selected regions of interest (ROI) were outlined by referring to GFP-positive regions to mark Dve expressing cells and Elav positive cells to mark differentiating photoreceptor cells across the entire eye disc. Area and fluorescence measurements were obtained by using the ‘Measure’ tool. Total fluorescence was calculated using the formula in the ImageJ program, CTCF (Corrected Total Cell Fluorescence) = (Integrated Density of ROI) – (Area of ROI X Mean fluorescence of background readings) (https://theolb.readthedocs.io/en/latest/imaging/measuring-cell-fluorescence-using-imagej.html).⁸² Integrated density was calculated as Mean Intensity X Area of the selected ROI.

qRT-PCR

Eye imaginal discs were dissected from control and experimental groups, stored in TRIzol at −80°C, and total RNA was extracted using the ZYMO RNA Clean and Concentrator Kit (Cat No.#R1050). A total of 500 ng RNA was used for cDNA synthesis using the First Strand cDNA synthesis kit (Cytiva, Danaher Corporation, Washington, DC, USA, Cat No.#27926101) following the manufacturer’s guidelines. qRT-PCR was performed using iQ-SYBER Green Supermix (Bio-Rad laboratories, Hercules, CA, USA; Cat No.# 1708882) with each sample in triplicate on the iCycler iQ Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). Each primer set (see key resources table) was tested in triplicate and relative gene expression was calculated using the ΔΔCT method (2−ΔΔCT method).

Quantification and statistical analysis

Statistical analyses were performed using GraphPad Prism (v9.5) software. One-way ANOVA was applied, followed by post-hoc analyses using Tukey’s test for comparisons among multiple groups (Figures 2G, 2H, 2J, 2K, 3D, 3E, 4E, 6A–6F, 8E, S1J, S2E, and S2I), and a t test was used for comparisons between two groups (Figures 3H and 5E). Graphs are presented as bar scatterplots showing mean ± SEM for each control and experimental group from IHC experiments. For imaging-based quantifications, biological replicates refer to independently dissected eye-antennal imaginal discs obtained from distinct third-instar larvae of the indicated genotype (e.g., n = 5 represents discs from five different larvae). Quantification was performed independently for each disc. Asterisks indicate significant differences based on p-values: ns = p > 0.05, ∗ = p < 0.05, ∗∗ = p < 0.01, ∗∗∗ = p < 0.001.

qRT-PCR quantification

For RT-qPCR experiments, three biological replicates (three independent cDNA samples prepared for each sample) independently collected from separate groups of larvae, whereas technical replicates refer to repeated qPCR measurements performed on the same biological sample (performed in triplicate for each run). Statistical analyses of the qRT-PCR data were performed using GraphPad Prism (v9.5) using the Ordinary One-way ANOVA method (Figures 5N, 5O, 7A–7F, S3G, and S3H), followed by correction for multiple comparisons using the Tukey’s post-hoc test assuming statistical significance at p ≤ 0.05. Graphs were generated using GraphPad Prism (v9.5).