A Functional Analysis of the Drosophila Gene hindsight: Evidence for Positive Regulation of EGFR Signaling.

We have investigated the relationship between the function of the gene hindsight (hnt), which is the Drosophila homolog of Ras Responsive Element Binding protein-1 (RREB-1), and the EGFR signaling pathway. We report that hnt mutant embryos are defective in EGFR signaling dependent processes, namely chordotonal organ recruitment and oenocyte specification. We also show the temperature sensitive hypomorphic allele hntpebbled is enhanced by the hypomorphic MAPK allele rolled (rl1 ). We find that hnt overexpression results in ectopic DPax2 expression within the embryonic peripheral nervous system, and we show that this effect is EGFR-dependent. Finally, we show that the canonical U-shaped embryonic lethal phenotype of hnt, which is associated with premature degeneration of the extraembyonic amnioserosa and a failure in germ band retraction, is rescued by expression of several components of the EGFR signaling pathway (sSpi, Ras85D V12 , pntP1 ) as well as the caspase inhibitor p35 Based on this collection of corroborating evidence, we suggest that an overarching function of hnt involves the positive regulation of EGFR signaling.

The gene hindsight (hnt), also known as (), was first identified in mutagenesis screens for embryonic lethal mutations performed in the early 1980s (Wieschaus ). The embryonic lethal phenotype of hnt was categorized as “U-shaped”, reflecting a failure to undergo or complete germ band retraction. hnt has since been identified as the Drosophila homolog of mammalian Ras Responsive Element Binding Protein -1 (RREB-1) (Melani ; Ming ), which strongly suggests a connection between hnt and the EGFR/Ras/MAPK signaling pathway (hereafter referred to as EGFR signaling). Interestingly, in Drosophila, hnt has been identified as a direct transcriptional target of the Notch signaling pathway (Krejci ; Terriente-Felix ). Mammalian RREB-1, on the other hand, has not been linked with Notch signaling but functions downstream of Ras/MAPK signaling and may either activate or repress certain Ras target genes (Liu ; Kent ). RREB-1 has also been implicated in a number of human pathologies, including pancreatic, prostate, thyroid, and colon cancer (Thiagalingam ; Mukhopadhyay ; Kent ; Franklin ).

The hnt gene encodes a transcription factor composed of 1893 amino acids containing 14 C2H2-type Zinc-fingers (Yip ). Based on genetic interaction studies, Hnt’s target genes are likely numerous and disparate with respect to function (Wilk ). Candidate direct target genes of Hnt identified using molecular methods include hnt itself, , and (Ming ; Oliva ). The gene encodes a Drosophila homolog of the human proto-oncogene ETO/MTG8, while encodes a conserved actin binding protein also known as filamen.

During development hnt is expressed in a broad range of tissues. In the embryo these include the amnioserosa (AS), anterior and posterior midgut primordia, the peripheral nervous system (PNS), the developing tracheal system, and the oenocytes (Yip ; Wilk ; Brodu ). During larval stages, in addition to the tracheal system, PNS, midgut, and oenocytes, hnt is expressed in the larval lymph gland, differentiated crystal cells, imaginal tracheoblasts, and the salivary glands of the third instar (Pitsouli and Perrimon 2010; Ming ; Terriente-Felix ). In pupae, the sensory organ precursors (SOPs) of developing micro- and macrochaetae, as well as myoblasts, and all photoreceptor cells (R cells) of the developing retina express hnt (Pickup ; Reeves and Posakony 2005; Krejci ; Buffin and Gho 2010). In the adult, hnt is expressed in the midgut (intestinal stem cells, enteroblasts, and enterocytes), developing egg chambers (follicle cells and the migratory border cells), spermathecae, and in mature neurons of the wing (Sun and Deng 2007; Melani ; Baechler ; Shen and Sun 2017; Farley ).

While hnt is expressed in many different tissues, its expression within a given tissue can be dynamic. For example, in the adult intestinal stem cell lineage there is an increase of Hnt during enteroblast-to-enterocyte differentiation, but a decrease during enteroblast-to-enteroendocrine cell differentiation (Baechler ). Hnt levels are particularly dynamic in the ovarian follicle cells, where Hnt is observed in stage 7-10A egg chambers as these cells initiate endoreduplication. A subset of follicle cells are subsequently devoid of Hnt through stages 10B to 13, and then display a strong increase in stage 14 egg chambers prior to follicle cell rupture and an ovulation-like event (Deady ).

There is a wealth of information regarding hnt mutant phenotypes and hnt expression, yet a general definition of Hnt function remains elusive. Given that Hnt is the Drosophila homolog of RREB-1, we present an examination of hnt mutant phenotypes as well as hnt overexpression with specific attention to EGFR signaling. With respect to loss-of function analysis, we report two new findings that link hnt and EGFR signaling: first, hnt mutant embryos are defective in the processes of chordotonal organ recruitment as well as oenocyte specification, both of which are EGFR signaling-dependent processes (Makki ); and second, we show that the temperature sensitive hnt allele hnt (hnt), which is associated with defective cone cell specification in the pupal retina (Pickup ), is enhanced by the hypomorphic MAPK allele (rl). In terms of hnt overexpression, we first show ectopic DPax2 expression in embryos overexpressing hnt. We show similar ectopic DPax2 expression in embryos in which EGFR signaling is abnormally increased through global expression of the active EGFR ligand secreted Spitz (sSpi). We subsequently demonstrate that loss-of-function mutants abrogate ectopic DPax2 expression in the context of hnt overexpression. Last, we show that the U-shaped phenotype of hnt mutants, which involves premature degeneration of the AS and a failure in the morphogenetic process of germ band retraction (GBR) - which is also a phenotype displayed by mutants (Clifford and Schupbach 1992) - can be rescued by expression of components of the EGFR signaling pathway (sSpi, , ) as well as the caspase inhibitor p35. Interestingly, expression of the isoform, which (unlike the isoform) requires activation by MAPK (O’neill ; Shwartz ), does not rescue hnt mutants. Given this collection of corroborating evidence, we suggest that a primary function of hnt involves the positive regulation of EGFR signaling.

Materials and Methods

Drosophila stocks

All cultures were raised on standard Drosophila medium at 25° under a 12 hr light/dark cycle, unless otherwise indicated. The hindsight (hnt) alleles used were hnt, hnt (Yip ; Wilk ), and hnt (this study). As previously described (Yip ), hnt is a strong hypomorphic embryonic lethal allele while hntpeb is a viable temperature sensitive hypomorphic allele associated with a rough eye phenotype at the restrictive temperature of 29°. The mutant alleles used were Egfr and Egfr as previously described (Shen ). The (rl) allele was provided by A. Hilliker. To drive ubiquitous expression throughout the early embryo we used daGAL4 as previously described (Reed ). The BO-GAL4 line was used to mark embryonic oenocytes (Gutierrez ) and was provided by A. Gould. Overexpression of hnt used UAS-GFP-hnt as previously described (Baechler ). The adherens junctions marker Ubi-DEcadherin-GFP was used to outline cell membranes as previously described (Cormier ). The reporter gene DPax2 was as previously described (Johnson ). UAS-sSpi was obtained from N. Harden. pebBAC was obtained from M. Freeman. All other transgenes used originated from stocks obtained from the Bloomington Drosophila Stock Center (UAS-CD8-GFP, UAS-GFP, UAS-p35, UAS-Ras85D, UAS-pnt, UAS-pnt)

Construction of DPax2-dsRed reporter lines

The DPax2 and DPax2 reporter lines were generated by standard P-element transgenic methods (Bachmann and Knust 2008) using the vector pRed H-Stinger (Barolo ) containing a previously described 3 KB DPax2 enhancer (Johnson ). Briefly, the 3 KB enhancer (position -3027 to +101 relative to the DPax2 transcription start site) was excised from the Bam HI sites of a DPaxB-pBluescript KS + plasmid. The insert was then cloned into the Bam HI site of pRed H-Stinger.

Crossing schemes for analysis of DPax2B2dsRed expression in Egfr mutants, and DPax2B1GFP expression in embryos with elevated EGFR signaling

In order to analyze DPax2 reporter construct expression in different backgrounds, the Ubi-DEcadherin-GFP (on second chromosome) was recombined with , UAS-GFP-hnt (on second chromosome) was recombined with , daGAL4 (on third chromosome) was recombined with DPax2, and daGAL4 (on third chromosome) was recombined with DPax2 creating the following stocks:

Stock 1: dp/ CyO

Stock 2: UAS-GFP-hnt Egfr/ CyO

Stock 3: daGAL4 DPax2

Stock 4: daGAL4 DPax2 / TM6C

To visualize DPax2 expression in /Egfr mutants, as well as /+ heterozygotes, the following approach was used. Non-balancer male progeny of Stock 1 x Stock 3 (dp/+; daGAL4 DPax2/+) were crossed to Stock 2. In embryos collected from this cross, /Egfr mutants were recognized as embryos expressing UAS-GFP-hnt, DPax2, and Ubi-DE-cadherin-GFP, while /+ heterozygotes also expressed UAS-GFP-hnt and DPax2, but lacked Ubi-DE-cadherin-GFP.

To visualize DPax2 expression in embryos with elevated EGFR signaling, Stock 4 was crossed to homozygous UAS-sSpi.

Immunostaining and Imaging

Immunostaining of embryos was carried out as described (Reed ). The following primary antibodies were used at the indicated dilutions: mouse monoclonal anti-Hindsight (Hnt) 27B8 1G9 (1:25; from H. Lipshitz, University of Toronto), mouse monoclonal anti-22C10 (1:500; Developmental Studies Hybridoma Bank (DSHB)), mouse monoclonal anti- Armadillo (1:100; DSHB), and rabbit polyclonal anti-DPax2 (1:2000; J. Kavaler, Colby College). The secondary antibodies used were: Alexa Fluor 488 goat anti-mouse and goat anti-rabbit (1:500; Cedarlane Labs), and TRITC goat anti-mouse (1:500; Cedarlane Labs). Staining embryos for f-actin using TRITC-phalloidin was performed as previously described (Reed ). Confocal microscopy and confocal image processing were performed as previously described (Cormier ). Preparation of embryos for live imaging was as previously described (Reed ).

Fluorescent in situ hybridization (FISH)

Whole mount fluorescent in situ hybridization used 3 hr embryo collections of wild-type or daGAL4 > UAS-GFP-hnt aged for 10 hr at 25°, giving embryos at stage 13-16. Embryo fixation followed protocols as described (Lecuyer ). cDNA clones were acquired from the Drosophila Genomics Resource Center (Indiana University), including the DPax2 clone IP01047.

Cone cell distribution quantification

48hr APF pupal eye discs were immunostained using anti-armadillo as described above in three genetic backgrounds (, , ). is a temperature sensitive recessive visible allele and was reared under permissive (25°) and restrictive (29°) conditions. and lines were reared at 25°. Five to six independent eye discs were examined for each genotype and condition ( 25°, 25°, 29°, and 25°). The average frequencies of cone cell within an ommatidium, ranging from 1-5, were calculated with the standard deviation then plotted onto a stacked bar graph.

Recovery of hntNP7278ex1

The viable and fertile GAL4 enhancer trap line NP7278, inserted 158 bp upstream of the hnt transcription start site (Thurmond ), was mobilized by crossing to Δ2-3 transposase. Progeny were crossed to FM7h, w B and lines were established from single virgin females that had lost the w marker of NP7278. Lethal lines (not producing B progeny) were subsequently selected and tested for GAL4 expression by crossing to UAS-GFP.

hntNP7278ex1 rescue experiments

The hnt stock was crossed into a background carrying second chromosome insertions UAS-GFP and Ubi-DE-cadherin-GFP. Virgin females of this resulting stock (y w hnt, w; UAS-GFP) were subsequently crossed to tub-GAL80 hsFLP FRT19A males (for control mutant) or to tub-GAL80 hsFLP FRT19A; UAS-X males for rescue experiments (where UAS-X was the homozygous 2 chromosome insertion UAS-p35, or one of the homozygous 3 chromosome insertions UAS-sSpi, UAS-Ras85D, or UAS-pnt). In the case of the 3 chromosome insertion UAS-pnt, which is not homozygous viable, male tub-GAL80 hsFLP FRT19A; UAS-pnt outcross progeny were used. Embryos between 12-14 hr old were collected from crosses of 30-40 females and males using an automated Drosophila egg collector (Flymax Scientific Ltd.) at room temperature (22°) and mounted for live imaging as previously described (Reed ). For each imaging session, non-mutant embryos were confirmed as having completed or being in the terminal stages of dorsal closure. Mutant embryos (hnt; UAS-GFP or hnt; UAS-GFP; UAS-X/+) were unambiguously identified by expression of UAS-GFP (Fig. S3). In the case of UAS-pnt, mutant embryos also expressing UAS-pnt were identified as those embryos having UAS-GFP expression while lacking UAS-Cherry expression. A control rescue was performed by crossing to y w hnt; pebBAC males (BAC insert is hnt). Images of mutant embryos were scored as one of three possible categories: 1) GBR failure (telson pointed anteriorly) with a small AS remnant; 2) GBR partial (telson pointed vertically or posteriorly but not at full posterior position) with an intact but distorted AS; 3) GBR complete (telson pointed posteriorly and located at normal posterior position) and with an intact but distorted or normal AS.

Data availability

Stocks used that are unique to this study are available upon request. Supplemental material has been uploaded to figshare. The image data sets and embryo scoring result used to evaluate hnt rescue (presented in Figure 5K) are available as supplemental material (Fig. S1). Other supplemental material includes the demonstration of reduced hnt expression in hnt mutant embryos (Fig. S2) and Punnett square diagrams detailing the genetic crosses used for the unambiguous identification of mutant and rescued hnt mutant embryos (Fig. S3). Supplemental material available at figshare: https://doi.org/10.25387/g3.9992405.

Figure 5

GBR and premature amnioserosa death of hnt is rescued by caspase suppression and by activation of EGFR signaling. (A) Anti-Hnt immunostained showing AS expression prior to onset of GBR. (B) Live confocal image of hnt/+; UAS-GFP/+ embryo showing AS expression associated with hnt prior to onset of GBR. (C) Same embryo shown in B imaged 67 min later during initiation of GBR. The AS is folded over the extended tail and lamellopodia-type extensions contact the epidermis (white arrowheads. (D) Live confocal image of hnt/Y; UAS-GFP/+ mutant embryo at onset of GBR showing a failure of AS to maintain the fold over the posterior tail. AS apoptotic corpses are also present (white arrowheads). (E) Terminal GBR failure phenotype of hnt/Y; UAS-GFP/+ mutant embryo showing tail-up phenotype and AS remnant (white arrowhead). (F) Control rescue embryo: hnt or hnt/hnt mutant with UAS-GFPnls Ubi-DEcadherin showing rescue by pebBAC. (G) GBR complete rescue of hnt by UAS-sSpi. (H) GBR complete rescue of hnt by UAS-p35. (I) GBR complete rescue of hnt by UAS-Ras85D. (J) GBR complete rescue of hnt. (K) Stacked bar graph showing the frequency of GBR defects in hnt mutants and rescue backgrounds.

Results

PNS, chordotonal organ and oenocyte specification are disrupted in hnt loss-of-function mutants

In order to determine if phenotypes associated with reduced EGFR signaling are present in hnt mutants, we first examined the development of the PNS in hnt mutant embryos using anti-Futsch/22C10 (hereafter referred to as 22C10), which labels all neurons of the PNS as well as some neurons of the central nervous system (CNS) (Hummel ). hnt mutant embryos lack sensory neurons (Figure 1A, B). The absence of sensory neurons is most evident in the abdominal segments. Each embryonic abdominal hemisegment normally contains eight internal stretch receptors known as chordotonal organs, arranged as a single dorsal lateral organ (v’ch1), a lateral cluster of five (lch5), and two single ventral lateral organs (vchB, and vchA) (Brewster and Bodmer 1995). 22C10 immunostaining shows the neurons of the lch5 clusters are frequently reduced from five to three in number in hnt mutants (blue arrowheads, Figure 1A, B and Figure 1A’, B’). TRITC-phalloidin staining of f-actin confirms the reduction of the lch5 clusters from five to three (asterisks, Figure 1C and Figure 1D), and reveals a complete absence of the single chordotonal organs in hnt mutants (arrowheads in Figure 1C).

Figure 1

The embryonic hnt mutant phenotype includes hallmarks of reduced EGFR signaling. (A) Wild-type stage 15 embryo immunostained using the neuronal marker 22C10 showing typical development of the PNS, including clusters of ventral neurons in the second and third thoracic segments (yellow arrowheads) and five neurons associated with lateral chordotonal organ clusters in the abdominal segments (blue with white outline arrowheads and inset A’). (B) 22C10 immunostained hnt mutant embryo showing the absence of neurons (arrowheads cf. panel A) including two of the five neurons of each lateral chordotonal cluster (blue with white outline arrowheads and inset B’). (C) TRITC-phalloidin stained stage 15 wild-type embryo showing the f-actin rich structure of the lateral chordotonal lch5 organ clusters (asterisks) and the dorsolateral chordotonal organ lch1 (arrowheads). (D) TRITC-phalloidin stained hnt mutant embryo showing differentiated lateral chordotonal organs that are reduced in number (asterisks) and the absence of the dorsolateral chordotonal lch1 organ. (E) Wild-type embryo showing UAS-GFP expression using the oenocyte-specific driver BO-GAL4. (F) hnt mutant embryo showing reduced number of GFP-positive oenocytes (BO-GAL4 > UAS-GFP) and failure to form oenocyte clusters. Scale bars represent 20 microns (C,D).

In general, mutants lacking lateral chordotonal organs do not form oenocytes, and EGFR signaling has been implicated in oenocyte induction (Elstob ). We, therefore, used the oenocyte specific BO-GAL4 to drive expression of nuclear-GFP in wild-type and hnt mutants to evaluate oenocyte specification (Figure 1E,F). In addition to hnt mutants having reduced numbers of BO-GAL4-positive cells, these cells are not organized into clusters as in wild-type, but are scattered throughout the mutant embryos. This newly reported phenotype of hnt mutants, that of missing chordotonal organs and a failure in oenocyte differentiation, is a hallmark of reduced EGFR signaling (Makki ).

hntpeb is enhanced by reduced MAPK

Given the above findings, we were next interested in determining if a genetic background of reduced EGFR signaling would enhance a hnt mutant phenotype. Using anti-Armadillo (Arm) immunostaining, we evaluated the pupal ommatidial structure of the temperature sensitive hypomorphic hnt allele (hnt) as well as a viable hypomorphic mutant of the EGFR downstream effector MAPK, also known as (rl). At the permissive temperature of 25°, 87% of ommatidia in hntpeb mutants resemble wild-type and contain four cone cells (Figure 2A,B cf. 2C; Figure 2G). Likewise, 90% of ommatidia of rl mutants raised at 25° are normal (Figure 2D,G). The number of ommatidia showing a normal cone cell number is reduced to 28% in mutants raised at the restrictive temperature of 29° (Figure 2E,G) while peb; rl double mutants raised at the permissive temperature (25°) display a distinct enhancement of the mutant phenotype, having only 22% of ommatidia with the correct cone cell number (Figure 2F,G). These observations demonstrate a novel genetic interaction between hnt and MAPK, showing that rl behaves as an enhancer of the cone cell specification defect of Hnt. Interestingly, Hnt is not expressed in cone cells, but is expressed in photoreceptor precursor cells (R cells) where it is required for induction and expression within cone cells of the determinant DPax2 (Pickup ).

Figure 2

The viable temperature sensitive hypomorphic hnt allele (hnt) is enhanced by the viable hypomorphic MAPK allele (rl). (A) Anti-Arm immunostained wild-type pupal retina 48h APF showing the normal organization of ommatidial units. (B) Cartoon of wild-type ommatidial structure showing four cone cells (red - c), two primary pigment cells (yellow - 1°), and the secondary (white - 2°) and tertiary pigment cells (white - 3°) of the interommatidial lattice. Also depicted as a part of the lattice are the interommatidial bristles (dark green). (C) Anti-Arm immunostained pupal retina (48h APF) of mutant raised at the permissive temperature (25°C) showing normal ommatidial organization. (D) Anti-Arm immunostained pupal retina (48h APF) of mutant raised at 25°C showing normal ommatidial organization. (E) Anti-Arm immunostained pupal retina (48h APF) of mutant raised at the restrictive temperature (29°C) showing a disruption in ommatidial organization. (F) Anti-Arm immunostained pupal retina (48h APF) of peb; rl double mutant raised at the permissive temperature of 25°C showing disrupted ommatidial organization, indicating a genetic enhancement of under what is normally the permissive condition. (G) Stacked bar graph showing the average frequency of observed cone cells per ommatidium (1-5 CC) for 25°C, 25°C, 29°C, and peb; rl 25°C.

Overexpression of hnt during embryogenesis results in ectopic DPax2 expression

Using a candidate gene approach, we examined stage 13-16 embryos in which UAS-GFP-hnt was globally expressed using the daGAL4 driver. Among candidate genes tested, DPax2 (CG11049, also known as () or sparkling (spa)) was found to show a striking transcriptional upregulation in embryos overexpressing hnt compared to control embryos (Figure 3A,B). The upregulation of DPax2 in embryos overexpressing hnt was confirmed at the level of protein expression by anti-DPax2 immunostaining (Figure 3C,D) as well as by reporter gene construct expression (Figure 3E,F). Interestingly, hnt mutants do not abolish or reduce DPax2 expression (Figure 3G), suggesting that while hnt overexpression can result in DPax2 overexpression, Hnt is not required for endogenous DPax2 expression throughout the embryonic PNS.

Figure 3

Global overexpression of hnt results in ectopic DPax2 expression. (A) Wild-type embryo showing DPax2 mRNA distribution expression using FISH (green) (B) Embryo overexpressing hnt (daGAL4 > UAS-GFP-hnt) showing ectopic and increased levels of DPax2 mRNA using FISH (green). (C) Wild-type embryo showing DPax2 expression using anti-DPax2 immunostaining (blue). (D) Embryo overexpressing hnt immunostained for DPax2 (blue) showing ectopic DPax2 in large regions of lateral ectoderm. (E) Wild-type embryo showing expression of the reporter gene construct DPax2 (blue) as faithful to endogenous DPax2 expression throughout the developing PNS. (F) Embryo overexpressing hnt showing ectopic DPax2 expression using the DPax2 reporter gene. (G) Embryo immunostained for DPax2 (blue) and Hnt (yellow) showing that this embryo is a hnt mutant (absence of Hnt signal) and DPax2 throughout the PNS.

Ectopic DPax2 expression in the context of hnt overexpression is EGFR dependent

DPax2 encodes a paired domain transcription factor and is expressed in the developing PNS, including the embryonic PNS, pupal eye, and micro- and macrochaetes (Fu ). We next wished to determine if DPax2 expression in embryos overexpressing hnt is dependent on EGFR signaling. Compared to the overexpression control (Figure 4A-A’’), we found that reduced EGFR (Egfr) suppresses ectopic DPax2 expression (Figure 4B-B’’). We also observed that DPax2 overexpression associated with hnt overexpression is sensitive to dosage as Egfr/+ heterozygous embryos show reduced DPax2 expression relative to the overexpression control (Figure 4C-C’’). To further corroborate DPax2 ectopic expression as EGFR-dependent, we examined DPax2 reporter gene expression in embryos globally expressing the activated EGFR ligand secreted Spitz (sSpi). Such embryos also show ectopic DPax2 expression, suggesting that ectopic DPax2 expression is elicited through increased EGFR signaling (Figure 4 D,E). In addition, we found that the same mutant (Egfr) does show expression of the DPax2 reporter. Although the total number of DPax2 expressing cells is reduced relative to wildtype, this indicates that mutants are capable of producing cells that express DPax2 (Figure 4F). Taken together, these data are consistent with the interpretation that DPax2 is not a direct target of hnt, that ectopic DPax2 expression is a consequence of excessive EGFR signaling, and that hnt overexpression may result in DPax2 overexpression through excessive EGFR signaling. Moreover, these results raise the possibility that hnt loss-of-function mutants could possibly be rescued by ectopic activation of Egfr signaling.

Figure 4

Ectopic DPax2 expression associated with hnt overexpression requires EGFR signaling. (A-A’’) Immunostained pan-GFP-hnt embryo (daGAL4 > UAS-GFP-hnt) showing Hnt (yellow, A’) and associated ectopic DPax2 (Blue, A’’). (B-B’’) Pan-GFP-hnt embryo that carries the loss-of-function allelic combination / , showing absence of ectopic DPax2 expression using the DPax2 reporter. (C-C’’) Pan-GFP-hnt embryo heterozygous for the allele showing reduced ectopic expression of the DPax2 reporter. (D) Wild-type stage 15 embryo showing that expression of the DPax2 reporter gene is consistent with endogeneous DPax2 (cf. Fig. 3C). (E) Embryo expressing the DPax2 reporter gene in the background of globally activated EGFR signaling (daGAL4 > UAS-sSpi) showing ectopic DPax2 expression. (F) The loss-of-function allelic combination / in the absence of hnt overexpression, showing DPax2 expression using the DPax2 reporter.

The embryonic U-shaped terminal mutant phenotype of hntNP7278ex1 is rescued by activation of EGFR signaling

Given the above results showing phenotypes related to reduced EGFR signaling in hnt mutants, the genetic enhancement between hnt and rl, in addition to the EGFR-dependence of ectopic DPax2 expression associated with hnt overexpression, we wished to test if hnt loss-of-function phenotypes can be rescued by activation of Egfr signaling. As is the case for mutants, hnt mutants fail to undergo or complete GBR and are associated with premature AS degeneration and death (Frank and Rushlow 1996; Goldman-Levi ; Lamka and Lipshitz 1999). We conducted rescue experiments using a newly recovered hnt allele, hnt (see Materials and Methods). The hnt allele is a GAL4 enhancer trap insertion that is embryonic lethal, fails to complement hnt, shows premature AS degeneration, has GBR defects (Figure 5D,E,K), and is rescued by pebBAC (Figure 5F, K). Very similar to the previously described allele hnt (Reed ), hnt shows reduced anti-Hnt immunostaining (Fig. S2). hnt is, therefore, best characterized as a strong hypomorphic allele. Interestingly, the hnt mutant retains GAL4 expression in a pattern faithful to endogenous hnt expression, including early (prior to onset of GBR) expression in the AS (Figure 5A,B). The hnt mutant phenotype, however, does not disrupt oenocyte specification or the lch5 cluster of chordotonal organs as we described for hnt. We, therefore, chose to test for rescue of premature AS death and GBR failure. We were able to use hnt in combination with an X-linked tub-GAL80 insertion to unambiguously identify hemizygous hnt mutant embryos that also express an autosomal UAS transgene (see Materials and Methods, and Fig. S3). We found that 72.4% (n = 58) of control hnt embryos show a strong U-shaped phenotype in which the AS is reduced to a small remnant, indicative of GBR failure and premature AS degeneration, respectively (Figure 5E,K). The AS degeneration and GBR phenotype of hnt mutants was rescued by expression of the baculovirus caspase inhibitor UAS-p35 (5.9% GBR failure; n= 34; Figure 5F,I), the activated EGFR ligand UAS-sSpi (0% GBR failure; n = 27, Figure 5H,K), constitutively active RAS (8.3% GBR failure; n= 36; Figure 5I,K). We also tested for rescue of hnt by expression of two isoforms of the ETS transcription factor effector encoded by (), which is a downstream effector of the EGFR/Ras/MAPK pathway. The isoform PntP2 requires activation through phosphorylation by MAPK, whereas the PntP1 isoform, which is transcriptionally activated by the activated form of PntP2, is constitutively active without activation by MAPK (O’neill ; Shwartz ). Expression of the constitutively active isoform via UAS-Pnt resulted in rescue (9.1% GBR failure; n= 31; Figure 5J,K). Interestingly, expression the other isoform via UAS-Pnt did not rescue hnt (72.0% GBR failure, n= 25; Figure 5K). All image data sets and scoring annotations used to generate Figure 5K are presented as supplemental material (Fig. S1). Rescue by UAS-p35 confirms that premature AS degeneration in hnt mutants is associated with caspase activation. Furthermore, rescue of hnt mutants by expression of components of the EGFR signaling pathway is consistent with hnt operating either upstream or in parallel to this pathway. Rescue was not complete in that AS morphology was abnormal, and rescued embryos failed to complete dorsal closure likely due to the abnormal persistence of the rescued AS. Interestingly, the failure to rescue AS death and GBR defects by expression of the Pnt isoform, which requires activation through phosphorylation by MAPK (O’neill ; Shwartz ), is consistent with reduced MAPK activity within the AS of hnt mutants.

Discussion

Hnt loss-of-function and Hnt overexpression phenotypes are consistent with perturbations in EGFR signaling

The development of chordotonal organs and oenocyte specification are both disrupted in hnt mutants and these phenotypes are hallmarks of reduced EGFR signaling. As an overview, each embryonic abdominal hemisegment normally develops eight chordotonal organs, organized into three single organs (v’ch1, vchB, and vchA), and a cluster of five organs (lch5). The embryonic specification and differentiation of chordotonal organs initiates with the delamination of chordotonal precursor cells (COPs) from the ectoderm (reviewed in (Gould )). Briefly, chordotonal organs arise from five primary COPs (C1-C5), where C1-C3 give rise to the five organs of lch5, C4 is a precursor of v’ch1, and C5 is the precursor for vchB and vchA. The secretion of the active EGFR ligand Spitz by C3 and C5 expands the number of COPs from five to eight. Further EGFR signaling elicited by the C1 COP is also required for the induction of oenocytes (reviewed in (Makki )). In the absence of Egfr signaling, C1 fails to recruit oenocytes, and C3 fails to recruit secondary COPs to complete the five lateral chordotonal organs of the lch5 cluster (Gould ). Mutant phenotypes of genes belonging to what has been called the Spitz group (which encode components of the EGFR signaling pathway and include , , , and ), as well as the expression of dominant-negative EGFR, all display an absence of oenocytes and the formation of only three lateral chordotonal organs within the lch5 cluster (Bier ; Elstob ; Rusten ). Based on our analysis of hnt mutant embryos, we suggest that hnt can be aptly described as a previously unrecognized member of the Spitz group of mutants. Overall, however, our findings represent additions to the list of phenotypic similarities between hnt and mutants, including germ band retraction and dorsal closure failure, as well as the loss of tracheal epithelial integrity (Clifford and Schupbach 1992; Cela and Llimargas 2006; Shen ).

We found hnt overexpression in the embryo results in increased and ectopic expression of DPax2, and we found this effect to be unequivocally Egfr-dependent. We also found that global activation of Egfr signaling via expression of the Egfr ligand sSpi also causes DPax2 overexpression. Our results are consistent with previous work showing that Hnt is required in the developing eye imaginal disc for cone cell induction; here, it was also shown that reduced hnt expression resulted in reduced DPax2, that hnt overexpression resulted in increased DPax2, and that these effects were non-autonomous (Pickup ). The suggested model was that Hnt is required within the R1/R6 photoreceptor precursor cells to achieve a level of Delta sufficient for cone cell induction. While our suggestion that Hnt promotes Egfr signaling is not mutually exclusive with a role in promoting expression, it is noteworthy that the expression of within R-precursor cells is elevated by the activation of EGFR signaling in these cells (Tsuda ). The observation of reduced Delta associated with reduced hnt expression could, therefore, be attributed to reduced Hnt-dependent EGFR signaling within the R-precursor cells.

Rescue of the hnt U-shaped mutant phenotype

The AS, which is programmed to die during and following the process of dorsal closure, is possibly required for mechanical as well as signaling events that are critical for the morphogenetic processes of GBR and dorsal closure. Premature AS death may, therefore, lead to U-shaped or dorsal closure phenotypes. In support of this view, AS-specific cell abalation disrupts dorsal closure (Scuderi and Letsou 2005), and other U-shaped mutants display premature AS death, including (), tail-up (), (), and () (Frank and Rushlow 1996; Goldman-Levi ; Reed ).

AS programmed cell death normally occurs through an upregulation of autophagy in combination with caspase activation (Mohseni ; Cormier ). AS death can be prevented, resulting in a persistent AS phenotype, in a number of backgrounds. These include expression of the caspase inhibitor p35, RNAi knockdown of the proapoptotic gene , expression of activated Insulin receptor (dInR), dominant negative ecdysone receptor (EcR), active EGFR ligand secreted Spitz (sSpi), constitutively active RAS (Ras85D), as well as over expression of -GFP (Mohseni ; Shen ). In addition, embryos homozygous for Df(3L)H99, which deletes the pro-apoptotic gene cluster /hid/, also present a persistent AS phenotype (Mohseni ; Cormier ). During normal development, Hnt is no longer detectable by immunostaining within the AS as it begins to degenerate following dorsal closure (Reed ; Mohseni ). Thus, it is likely that hnt downregulation is required for normal AS degeneration, and that the mutant phenotype of hnt is the result of a premature activation of the normal death process. In support of this, we have demonstrated that several backgrounds associated with a persistent AS phenotype are able to rescue GBR failure and AS death in hnt mutants.

In the context of programmed cell death within the embryonic CNS, MAPK dependent phosphorylation has been show to inhibit the pro-apoptotic activity of the Hid protein (Bergmann ). We suggest that Egfr signaling within the AS could also represent a survival signal, leading to MAPK activation and Hid inhibition. Several observations are consistent with this model, including AS expression of several components of the Egfr signaling pathway. For example, within the AS anlage there is robust expression of () (Francois ), which encodes a intramembrane serine protease required for the activation of EGFR ligands; see (Shilo 2005). In addition, prior to the onset of GBR, there is pronounced AS expression of (), which encodes an additional EGFR ligand (Schnepp ). Vein is a weaker EGFR ligand, but it is produced in an active form and is not subject to inhibition by the EGFR antagonist Argos (Aos); see (Golembo ; Shilo 2005). At about the same stage, expression of a downstream EGFR effector () is found in the AS, as is , which is also expressed in the apoptotic AS (see Berkeley Drosophila Genome Project; https://insitu.fruitfly.org/cgi-bin/ex/insitu.pl).

Potential Hnt target genes and EGFR signaling

As a model for normal AS death, we suggest that a downregulation of hnt expression could lead to reduced EGFR AS signaling, thereby decreasing MAPK inhibitory phosphorylation of the pro-apoptotic protein Hid. According to this model, AS death and subsequent GBR failure in hnt mutants would be attributed to reduced EGFR signaling, lower MAPK activity, and pro-apoptotic activity of unphosphorylated Hid. But how might hnt expression promote Egfr signaling and maintain high MAPK activity?

A recent genetic screen for genes involved in the regulation of Wallerian degeneration (the fragmentation and clearance of severed axons) identified hnt as being required for this process. As part of this work, the authors performed ChIP-seq analysis of a GM2 Drosophila cell line expressing a tagged version of Hnt. This resulted in the identification of 80 potential direct targets of Hnt (Farley ). Interestingly, several of these putative Hnt target genes are also known targets of the EGFR signaling pathway, including (Zhang ), (Xiang ), bantam (Herranz ), (Tsuda ), and (Shirai ); while others have been implicated in the regulation of EGFR signaling and include (Qian ), (Campbell ), (Huang and Rubin 2000), (Singh ), and (Lim ). Also, and of particular interest, among the genes identified are known target genes of EGFR signaling that are also regulators or effectors of EGFR signaling. These include the gene , which encodes an ETS transcriptional activator - a key component for the transcriptional output of EGFR signaling that can also create a positive feedback loop through the transcription of (Golembo ; Paul ; Cruz ), and (Mitogen-activated protein kinase), which is a negative regulator of EGFR signaling (Gabay ; Kim ; Butchar ). Further investigations will be required to determine if the phenotypes associated with hnt overexpression, as well as hnt loss-of-function, can be attributable (in whole or in part) to changes in expression of any of these potential target genes.