Expression of mutant CHMP2B, an ESCRT-III component involved in frontotemporal dementia, causes eye deformities due to Notch misregulation in Drosophila

Abstract

Endosomal sorting complexes required for transport (ESCRTs) mediate sorting of ubiquitinated membrane proteins into multivesicular bodies en route to lysosomes for degradation. A mutation in CHMP2B (CHMP2B^Intron5, an ESCRT-III component) that is associated with a hereditary form of frontotemporal dementia (FTD3) disrupts the endosomal-lysosomal pathway and causes accumulation of autophagosomes and multilamellar structures. We previously demonstrated that expression of CHMP2B^Intron5 in the Drosophila eye using GMR-Gal4 causes misregulation of the Toll receptor pathway. Here, we show that ectopic expression of CHMP2B^Intron5 using eyeless-Gal4 (ey>CHMP2B^Intron5), a driver with different spatiotemporal expression attributes than GMR-Gal4 in the Drosophila eye, causes eye deformities when compared to expression of wild-type CHMP2B (CHMP2B^WT) and the Drosophila homologue of CHMP2B (CG4618). In addition, ey>CHMP2B^Intron5 flies showed defects in photoreceptor cell patterning and phototactic behavior. Furthermore, ey>CHMP2B^Intron5 flies showed accumulation of Notch in enlarged endosomes and up-regulation of Notch activity. Partial loss of Notch activity in ey>CHMP2B^Intron5 flies significantly rescued eye deformities, photoreceptor patterning defect, and phototactic behavior defect, indicating that these defects are primarily due to Notch misregulation. These results demonstrate that CHMP2B^Intron5 preferentially affects different receptor signaling pathways in a cellular and developmental context-dependent manner.—Cheruiyot, A., Lee, J-A., Gao, F-B., Ahmad, S. T. Expression of mutant CHMP2B, an ESCRT-III component involved in frontotemporal dementia, causes eye deformities due to Notch misregulation in Drosophila.

Eukaryotic cells degrade transmembrane proteins and receptors within lysosomes via a process that involves sorting of ubiquitinated transmembrane proteins into multivesicular bodies (MVBs), which then fuse with the lysosomes to degrade the cargo (1–3). The endosomal sorting complex required for transport (ESCRT) machinery, which primarily consists of 4 heteromeric subcomplexes (ESCRT-0, -I, -II, and -III), mediates MVB biogenesis through cargo sorting and endosomal membrane remodeling (1, 2). Recycling of cell surface receptors via MVBs allows homeostasis of receptors and their signaling pathways. In addition to MVB formation, ESCRTs are involved in macroautophagy, epithelial cell polarity, cell migration, cytokinesis, dendritic spine morphogenesis, and viral budding (2). Consequently, disruption of ESCRT function is associated with a number of diseases, including cancer, pathogenic infections, and neurodegenerative diseases, such as frontotemporal dementia (FTD), amyotrophic lateral sclerosis (ALS), Huntington's disease, paraparesis, and prion disease (2, 4, 5).

Mutations in CHMP2B (vps2 in yeast and CG4618 in Drosophila), an ESCRT-III subunit, are associated with familial cases of FTD linked to chromosome 3 (6) and ALS (7–9). One of the CHMP2B mutations in the splice acceptor site for the sixth exon (CHMP2B^Intron5), which was originally identified in a Danish family with FTD3, results in truncation of 35 aa at the carboxyl terminus of the protein (6, 10). Multiple cellular and physiological defects of CHMP2B^Intron5 have been observed in primary neurons, cell lines, Drosophila, and mice, including impaired endosomal-lysosomal fusion, accumulation of autophagosomes and ubiquitin- and p62-positive inclusions, morphological defects in dendritic arbor and spines and axonal swelling, and receptor misregulation (11–18). It has been proposed that CHMP2B^Intron5-mediated defects have a gain-of-function effect, since overexpression of CHMP2B^WT in mice and flies or CHMP2B-knockout mice results in either mild or no phenotypes (11, 14, 15).

Ectopic expression of CHMP2B^Intron5 in the Drosophila eye by GMR-Gal4 misregulates Toll receptor and its signaling pathway (11). Since receptor expression and activity are spatiotemporally regulated, we investigated whether CHMP2B^Intron5-mediated disruption of ESCRT function during different stages of eye development will result in differential receptor misregulation and therefore different phenotypes. To accomplish this, we used the ey-Gal4 driver of the Gal4-upstrean activating sequence (UAS) expression system, which enables spatiotemporal control of transgene expression (19). ey-Gal4 expression is initiated during the embryonic stage and continues later in larva in all cells of the eye disc preceding the morphogenetic furrow, which is formed by a series of events defining the specification of Drosophila photoreceptor cell fate determination and differentiation (20, 21). GMR-Gal4 expression is initiated later than ey-Gal4 during eye development in cells posterior to the morphogenetic furrow (22, 23). Here, ey-Gal4 driven expression of UAS-Flag-CHMP2B^Intron5 (ey>CHMP2B^Intron5) was used to examine the effects of dysfunctional ESCRT-III caused by CHMP2B^Intron5 in the regulation of Notch signaling during eye morphogenesis.

MATERIALS AND METHODS

Fly stocks

UAS-Flag-CHMP2B^Intron5 transgenic flies were generated by including an N-terminal Flag tag using the previously described cloning and transgenic techniques (11, 15). To generate UAS-CG4618 flies, cDNA (LD36173; Drosophila Genomics Resource Center, Indiana University, Bloomington, IN, USA) was cloned into pUAST vector, which was followed by sequencing and microinjection into w¹¹¹⁸ flies to obtain the transgenic line. The primers used for cloning into pUAST vector were as follows: CG4618, 5′-GCGGCCGCGATGTTCAACAATA-3′ and 5′-TCTAGACTAAGAGGAGCGCAG-3′. Canton-S (CS), eyeless-Gal4 (ey-Gal4), and Notch (N^264-39) Drosophila stocks were obtained from the Bloomington Stock Center (Bloomington, IN, USA). Flies were reared on a yeast-supplemented cornmeal-based diet (Nutri-Fly Bloomington Formula, Genesee Scientific, San Diego, CA, USA) and kept at 25°C under a 12-h light-dark cycle. CS was used as the wild-type control. ey-Gal4 flies were recombined with UAS-Flag-CHMP2B^Intron5 flies.

Phenotype range quantification and eclosion assay

The eye phenotype was quantified by assigning each fly (100 flies/genotype) to one of the following categories: normal shape and size, small/ablated, enlarged with approximately normal shape, and deformed and enlarged.

To determine the eclosion percentages, flies of the appropriate genotype were placed in an embryo collection chamber. Next, 100 embryos in 4 batches of 25 embryos were transferred to a vial containing food, and the numbers of adults emerging in each vial were counted. One-way ANOVA with the post hoc Scheffé test was used to identify significant differences (P<0.05) among genotypes.

Western blotting

Adult fly heads of the appropriate genotype were homogenized in RIPA buffer (Thermo Scientific, Pittsburgh, PA, USA), after which Western blots were performed following standard procedures. The gels were then transferred to PVDF membranes, which were later detected and scanned using GelDoc (Bio-Rad, Hercules, CA, USA). Specifically, 1:1000 dilutions of anti-CHMP2B (a gift from F.-B.G.), anti-rhodopsin [4C5; Developmental Studies Hybridoma Bank (DSHB), University of Iowa, Iowa City, IA, USA], anti-tubulin (E7; DSHB), anti-Notch (C17.9C6, DSHB), anti-Numb (Santa Cruz Biotechnology, Dallas, TX, USA), and anti-Flag (Sigma-Aldrich, St. Louis, MO, USA) primary antibodies produced in mice were used to probe for Rh1 rhodopsin, tubulin, and Flag-CHMP2B^Intron5, respectively. In addition, 1:1000 dilutions of HRP-goat anti-mouse (Santa Cruz Biotechnology), HRP-goat anti-rabbit (Santa Cruz Biotechnology), and HRP-donkey anti-goat (Santa Cruz Biotechnology) were used as the secondary antibody. The data presented are based on three independent replicates.

Histology

Cryosectioning and immunohistochemistry were performed as described previously, with minor modifications (24). Briefly, 12-μm cryosections of adult fly heads embedded in TissueTek optimal cutting temperature (OCT) compound (Ted Pella, Redding, CA, USA) were probed with 1:50 anti-Rh1 rhodopsin (4C5), anti-Notch (C17.9C6), anti-Rab5 (a gift from F.-B.G.), and Rab7 (a gift from F.-B.G.) and 1:50 dilutions of Alexa Fluor 488 anti-rabbit (Invitrogen, Grand Island, NY, USA), and Alexa Fluor 568 anti-mouse (Invitrogen) were used as secondary antibodies. The slides from 3 independent replicates were then mounted with Fluoro Gel II with DAPI (Electron Microscopy Sciences, Hatfield, PA, USA) and imaged using an AxioCam MRm camera on an Imager A2 fluorescent microscope (Zeiss, Thornwood, NY, USA).

Adult phototactic assay

An adult phototactic assay was conducted according to a previously described countercurrent distribution method with minor modifications (25). Briefly, the phototactic apparatus consisted of two polypropylene narrow fly vials (Genesee Scientific) lined with black electrical tape inside and out, except for the bottom of one vial, and a 20-W electric bulb was positioned immediately above the vial at the clear end. Adult flies (25–30) were acclimated to the apparatus for 10 min, after which they were subjected to 3 trials to measure whether they selected the light or dark end of the apparatus. One-minute rests were provided between trials. Independent replicates (3–5) consisted of previously untested flies. One-way ANOVA with a post hoc Scheffé test was used to identify significant differences (P<0.05) among genotypes.

Real-time polymerase chain reaction (PCR)

Total RNA (with DNase treatment) from 10–15 adult fly heads of the appropriate genotype was isolated using an RNeasy miniprep kit (Qiagen, Valencia, CA, USA). RNA quality and yield were then measured using a Nanodrop spectrometer (Thermo Scientific). Quantitative reverse transcription PCR (qRT-PCR) was performed in duplicate with 50–100 ng total RNA using a 1-step Quantifast SYBR Green RT-PCR kit (Qiagen) and a StepONE Real-Time PCR system (Applied Biosystems, Grand Island, NY, USA). The relative normalized transcript level was determined by the ΔΔC_t method based on 3 independent replicates. RP49 was used as the normalizing gene. The primers used for qRT-PCR were as follows: enhanced green fluorescent protein (eGFP), 5′-AGTCCGCCCTGAGCAAAGA-3′ and 5′-TCCAGCAGGACCATGTGATC-3′; RP49, 5′-CGGTTACGGATCGAACAAGC-3′ and 5′-CTTGCGCTTCTTGGAGGAGA-3′.

RESULTS

ey>CHMP2B^Intron5 flies exhibit eye deformities and a phototactic defect

To identify the effects of CHMP2B^Intron5 expression, we used ey-Gal4 to express UAS-Flag-CHMP2B^Intron5 (ey>CHMP2B^Intron5) during early eye specification and morphogenesis in flies. In contrast to wild-type (Fig. 1A, A′), ey>CHMP2B^Intron5 flies showed severe defects in eye morphology, including enlargement, rippling, protuberances, and reduction/ablation, whereas ey-Gal4 and UAS-CHMP2B^Intron5 eyes appeared like wild-type eyes (Fig. 1B–D′ and Supplemental Fig. S1). The enlarged and deformed eye phenotypes were observed more frequently than reduced/ablated eye phenotype in ey>CHMP2B^Intron5 flies (Table 1). Expression of untagged CHMP2B^Intron5 also showed phenotypes similar to Flag-CHMP2B^Intron5 (Supplemental Fig. S1 and Table 1). ey-Gal4-mediated ectopic expression of CHMP2B^WT (Fig. 1E, E′) or CG4618 (Drosophila homologue of CHMP2B; Fig. 1F, F′) showed no obvious defects in eye morphology. ey>CHMP2B^Intron5 flies also showed a significantly lower adult eclosion rate (i.e., higher mortality) before adulthood relative to wild-type and ey>CHMP2B^WT flies (wild-type, 94±1.0%; ey>CHMP2B^WT, 92±1.6%; ey>CHMP2B^Intron5, 76±5.4%, ANOVA, Scheffé, P < 0.01). We previously demonstrated that UAS-CHMP2B^Intron5 and UAS-CHMP2B^WT transgenes had comparable expression levels by GMR-Gal4 (11). However, ey-Gal4 driver revealed variations in the expression levels of the transgenes. Flag-CHMP2B^Intron5 showed more robust expression than UAS-CHMP2B^Intron5 and UAS-CHMP2B^WT (Supplemental Fig. S2). Taken together, these findings indicate that CHMP2B^Intron5 causes eye deformities.

Figure 1.

ey>CHMP2B^Intron5 causes eye deformities. A, A′) Representative images of wild-type (CS) fly heads with normal eye size and shape. B–C′) ey>CHMP2B^Intron5 caused severe eye deformation, including deformities (B, B′), enlargement (C, C′), and small/ablated (D, D′) eyes. E–F′) ey>CHMP2B^WT (E, E′) or ey>CG4618 (fly homologue of CHMP2B; F, F′) contained no remarkable defects in eye morphology.

Table 1.

Distribution of eye deformity phenotype among genotypes

Genotype	Eye phenotype (%)
	Deformed	Enlarged	Small/Ablated	Normal
ey>CHMP2BIntron5	64	26	6	4
ey>Flag-CHMP2BIntron5	61	33	4	2

We further characterized the structural and functional aspects of the ey>CHMP2B^Intron5 phenotype. Longitudinal sections of the wild-type eyes showed normal curvature of the eye (Fig. 2A) and precise arrangement of the photoreceptor cell nuclei in the retina (Fig. 2A′). ey>CHMP2B^Intron5 flies showed defects in the eye curvature (Fig. 2B) and patterning of the photoreceptor cell nuclei (Fig. 2B′). However, the enlarged eye with additional eye units (ommatidia) appeared to contain normal photoreceptor cells, as indicated by a comparable rhodopsin localization pattern and content in ey>CHMP2B^Intron5 flies to wild-type flies (Fig. 2A″, B″, D). Furthermore, the ey>CHMP2B^Intron5 retina appeared better than that of the GMR>CHMP2B^Intron5 retina, which showed retinal degeneration due to defects in the depth of the retinal layer (Fig. 2C) and the density of photoreceptor cell nuclei (Fig. 2C′) as well as reduced rhodopsin content (Fig. 2C″, D).

Figure 2.

ey>CHMP2B^Intron5 disrupts eye curvature, photoreceptor patterning, and phototactic behavior. A–A″, D) Representative images of longitudinal sections of a wild-type (CS) eye with normal curvature (A), patterning of photoreceptor cell nuclei (DAPI stain; A′, arrowhead and inset), rhodopsin localization (A″), and rhodopsin level (D). B–B″, D) ey>CHMP2B^Intron5 showed perturbation of the eye curvature (B) and patterning of photoreceptor cell nuclei (B′, arrowhead and inset), indicating defects in the retina. However, rhodopsin localization (B″) and level (D) were not affected relative to wild type, indicating no retinal degeneration. C–C″, D) GMR>CHMP2B^Intron5 showed no perturbation of the eye curvature, but did show defects in the depth of the retinal layer (C), density of photoreceptor cell nuclei (C′), and reduced rhodopsin level (C″, D), indicating retinal degeneration. Western blots were probed using anti-Flag antibody (for Flag-CHMP2B^Intron5). ninaE^I18 (Rh1 rhodopsin null) was used as a negative control for rhodopsin blot, and tubulin was used as the loading control. E) Phototactic assay of wild-type (CS) and ey>CHMP2B^WT flies showed a strong preference for light (96 and 46% of flies making at least 2/3 light decisions, respectively). ey>CHMP2B^Intron5 and GMR>CHMP2B^Intron5 flies showed significant defect in preference for the illuminated end of the apparatus, indicating defective phototactic behavior. Columns represent means ± sem of 3–6 independent replicates. *P < 0.05, ***P < 0.001; ANOVA, post hoc Scheffe.

To determine whether CHMP2B^Intron5-mediated structural defects in the eye were associated with physiological defects, we conducted a countercurrent distribution phototactic assay (25). Consistent with previous studies, wild-type flies showed a strong preference for light, with 96% favoring the illuminated end of the apparatus ≥2 times in 3 trials (25). GMR>CHMP2B^Intron5 flies, which showed retinal degeneration, showed severe defects in phototactic index (19%). ey>CHMP2B^Intron5 flies showed significant defects in phototactic behavior relative to wild-type and ey>CHMP2B^WT flies (Fig. 2E; wild-type, 94%; ey>CHMP2B^WT, 81%; ey>CHMP2B^Intron5, 46%, ANOVA, Scheffé P < 0.001). Therefore, these findings demonstrate that CHMP2B^Intron5 expression results in structural and physiological defects in the eyes (Figs. 1 and 2).

CHMP2B^Intron5 causes accumulation of Notch in endosomes and elevated signaling

The ey>CHMP2B^Intron5 eye phenotype was similar to the eye tumorigenesis observed in ey-Gal4 driven constitutive active Notch (ey>N^act; ref. 26) and flies with mutations in ESCRT-I (vps23/erupted/tsg101, vps28), ESCRT-II (vps22 and vps25) and ESCRT-III (vps2 and vps20) components due to Notch accumulation (27–32). Therefore, we investigated whether eye deformities in ey>CHMP2B^Intron5 flies was due to Notch up-regulation. We found that Notch level was significantly elevated in ey>CHMP2B^Intron5 heads relative to wild-type heads (Fig. 3A, B).

Figure 3.

ey>CHMP2B^Intron5 causes Notch accumulation. Representative Western blot of 3-d-old ey>CHMP2B^Intron5 head homogenate (A) and densitometric analysis of bands (B) showed elevated Notch level relative to wild-type (CS) flies. Western blots were also probed with anti-Flag antibody (for Flag-CHMP2B^Intron5) to show transgene expression. Tubulin was used as the loading control. Columns represent means ± sem of 3 independent replicates. *P < 0.05; ANOVA, post hoc Scheffe.

Immunolocalization in longitudinal sections of ey>CHMP2B^Intron5 eyes revealed accumulation of Notch in enlarged punctae relative to ey>CHMP2B^WT eyes (Fig. 4). To determine the vesicular identity of the enlarged Notch punctae, we colocalized Notch with Rab5 and Rab7 as markers of early endosomes and late endosomes/MVBs respectively (32–35). ey>CHMP2B^Intron5 eyes showed enlarged Rab5 and Rab7 positive endosomes, which showed partial colocalization with Notch positive puncta relative to ey>CHMP2B^WT eyes (Fig. 4).

Figure 4.

ey>CHMP2B^Intron5 causes Notch accumulation in enlarged endosomes. Representative images of immunostaining of longitudinal sections of ey>CHMP2B^Intron5 eyes showed accumulation of Notch in enlarged puncta (A, C) with partial colocalization (A″, C″, arrowheads) with enlarged Rab5 positive early endosomes (A′) and enlarged Rab7 positive late endosomes/MVBs (C′) relative to ey>CHMP2B^WT flies (B–B″, D–D″). Scale bars = 5 μm.

We next investigated whether the buildup of Notch in endosomes results in up-regulation of Notch signaling, which eventually leads to activation of genes containing Notch response element (NRE) in their promoter (36, 37). To accomplish this, we utilized transgenic flies containing NRE upstream of the eGFP open reading frame (36). Flies expressing ey>CHMP2B^Intron5 in the NRE-GFP background showed significantly elevated eGFP transcription when compared to sibling controls containing only the NRE-GFP element (Fig. 5A).

Figure 5.

ey>CHMP2B^Intron5 causes elevation in Notch signaling. A) Quantitative real-time PCR of ey>CHMP2B^Intron5 head total RNA in NRE-GFP background showed significantly increased eGFP transcript relative to sibling controls without ey>CHMP2B^Intron5, indicating ey>CHMP2B^Intron5 mediated elevation of Notch signaling. Columns represent means ± sem of 3 independent replicates. *P < 0.05; ANOVA, post hoc Scheffe. B) Representative Western blot of 3-d-old ey>CHMP2B^Intron5 head homogenate showed no apparent difference in Numb level when compared to wild-type (CS) flies.

To determine whether elevation in Notch signaling is influenced by changes in repressors of Notch in ey>CHMP2B^Intron5 flies, we tested the levels of Numb, a negative regulator of Notch signaling (38). Numb level in ey>CHMP2B^Intron5 flies did not appear to differ significantly from levels in wild-type and ey>CHMP2B^WT flies (Fig. 5B). Overall, these findings indicate that CHMP2B^Intron5 expression results in Notch accumulation in enlarged endosomes and elevated signaling.

Notch heterozygosity rescues ey>CHMP2B^Intron5 eye deformities and phototactic defects

To examine the relative contribution of Notch activation in the ey>CHMP2B^Intron5 eye phenotype, we expressed CHMP2B^Intron5 with ey-Gal4 in a reduced Notch gene dose background, i.e., the Notch (N^264-39) heterozygote. N^264-39; ey>CHMP2B^Intron5 flies showed remarkable recovery from eye deformities (Fig. 6A–C), eye curvature (Fig. 6A′–C′), patterning of the photoreceptor nuclei layer (Fig. 6A″–C″), and phototactic behavior (Fig. 6D). These findings demonstrate that ey>CHMP2B^Intron5 eye phenotypes are primarily due to Notch up-regulation.

Figure 6.

Notch heterozygosity ameliorates eye deformities and phototactic defects in ey>CHMP2B^Intron5 flies. A–A″, D) Wild-type (CS) flies with normal eye morphology (A), normal rounded eye curvature and patterning of the photoreceptor cell nuclei (DAPI staining of longitudinal sections; A′, A″), as well as strong preference for light in a phototactic assay (D). B–B″, D) ey>CHMP2B^Intron5 flies showed deformed eye morphology (B), defective eye curvature and abnormal patterning of the photoreceptor cell nuclei (B′, B″), and no preference for light (D). C–C″, D) ey>CHMP2B^Intron5 flies with a Notch heterozygote background (N^264-39) showed significant recovery of defective eye morphology (C), improvement in eye curvature and patterning of the photoreceptor cell nuclei (C′, C″), and no preference for light when compared to ey>CHMP2B^Intron5 flies, indicating a role of Notch up-regulation in ey>CHMP2B^Intron5 morphological and phototactic phenotypes (D). Columns represent means ± se of 3–5 independent replicates. **P < 0.01, ***P < 0.001; ANOVA, post hoc Scheffe.

DISCUSSION

CHMP2B^Intron5 mutation, originally identified in a Danish kindred with FTD associated with chromosome 3 (FTD-3; ref. 6), is a gain-of-function mutation that disrupts the endosomal-lysosomal pathway, resulting in accumulation of cell membrane receptors in cell culture and mouse models of FTD-3 (14, 15). We previously demonstrated that GMR-Gal4 driven ectopic expression of CHMP2B^Intron5 caused melanization due to misregulation of the Toll receptor-signaling pathway in Drosophila eye (11). Drosophila eye development is a well-studied developmental process in which multiple signaling pathways including the Notch receptor play key roles (39). Here, we demonstrated that ectopic expression of CHMP2B^Intron5 using eyeless-Gal4 (ey>CHMP2B^Intron5) caused changes in eye ranging from severe deformities and enlargement to ablation. ESCRT mutants were previously shown to have enlarged or smaller than normal eye disc sizes (28, 40). Although Flag-tagged CHMP2B^Intron5 showed higher expression level than the untagged CHMP2B^Intron5, their comparable eye phenotypes indicated the N-terminal Flag tag has no remarkable effect. In addition, ey>CHMP2B^WT flies showed no apparent phenotype. Lack of a phenotype in ey>CHMP2B^WT flies could be partially caused by a significantly lower expression levels relative to ey>CHMP2B^Intron5 flies. However, using GMR-Gal4, we previously demonstrated that GMR>CHMP2B^WT flies had a mild phenotype relative to severe melanotic phenotype in GMR>CHMP2B^Intron5 eyes, despite comparable expression levels (11). The differences in eye phenotypes between ey-Gal4 and GMR-Gal4 driven expression of CHMP2B^Intron5 were likely due to differential spatiotemporal expression by the GMR- and ey-Gal4 drivers. Previously, ey-Gal4 and GMR-Gal4 driven expression of a transgene were also shown to have different phenotypes (26). Specifically, expression of constitutively active Notch using ey-Gal4 (ey>N^act) caused severe eye hyperplasia, while only a mild phenotype was observed with GMR-Gal4 (26).

ey>CHMP2B^Intron5 flies also showed patterning defects in the photoreceptor cells and defective phototactic behavior response. Even though ey>CHMP2B^Intron5 flies showed rhodopsin content similar to normal, patterning defects of the photoreceptor cells could affect the innervation of photoreceptor neurons into the optic lobe. ey>CHMP2B^Intron5 retina also appeared to respond to light in an electroretinogram (Dr. J. E. O'Tousa, University of Notre Dame, Notre Dame, IN, USA, personal communication February 27, 2012). Previously, flies lacking neural dopamine were shown to have normal light response in photoreceptor cells indicative of normal rhodopsin function, but defective phototaxis (41). Moreover, enlarged eyes in ey>N^act flies were found to contain differentiated photoreceptor neurons (26).

The morphological and functional defects in ey>CHMP2B^Intron5 flies were associated with accumulation of Notch levels and up-regulation of Notch-mediated signaling. Moreover, ey>CHMP2B^Intron5 eyes in Notch heterozygous background showed robust improvement of deformities and phototactic defects. The ey>CHMP2B^Intron5 eye phenotype was similar to the eye tumorigenesis observed due to ectopic expression of constitutively active Notch in ey>N^act flies (26) and Notch accumulation in response to mutations in ESCRT-I (vps23/erupted/tsg101, vps28), ESCRT-II (vps22 and vps25), and ESCRT-III (vps2 and vps20) components (27–32). In addition, Notch is a well-characterized positive regulator of eye disc proliferation during eye development (42, 43).

Cell surface receptors such as Notch normally initiate signaling on ligand binding when embedded in the plasma membrane. Incorporation of receptors into endosomes is generally considered a mechanism that limits signaling of receptors by negating their access to extracellular ligands. However, Notch signaling has been shown to persist when Notch is located in the abnormally enlarged endocytic vesicles in ESCRT mutants (44, 45). Previous studies with ESCRT component mutants (e.g., vps23 and vps25), mostly using ESCRT component mutant mosaics generated in the eyeless expression pattern (ey-FLP system), showed Notch misregulation-associated overproliferation in both normal (nonautonomous) and mutant cells (autonomous) followed by apoptosis in only mutant cells (27–30, 40, 46). A recent study of ESCRT-II mutants using eye disc mosaics containing predominantly mutant cells revealed autonomous neoplastic proliferation followed by apoptosis and mortality before adulthood (40). In this study, we used the ey-Gal4-UAS expression paradigm; therefore, all eye disc cells expressed CHMP2B^Intron5 prior to differentiation into photoreceptor cells because of the ey-Gal4 expression pattern (21). ey>CHMP2B^Intron5 flies showed substantial mortality before adulthood, and enlarged eye phenotype was more frequent than ablated eye phenotype in the surviving adults. Therefore, the pathway mediating enlargement may have been stronger than the pathway mediating early mortality and ablation in ey>CHMP2B^Intron5 flies that survived to adulthood.

In summary, our findings support the notion that CHMP2B^Intron5 mutation disrupts the endosomal-lysosomal pathway, resulting in accumulation of cell membrane receptors. Depending on the spatiotemporal attributes of CHMP2B^Intron5 expression, different receptors and their signaling are predominantly misregulated. Similar to ESCRT mutants, CHMP2B^Intron5-mediated blockage of the endosomal-lysosomal pathway may lead to the buildup of many transmembrane proteins (27, 29, 30, 46). Interestingly, tumorigenic phenotype was not observed in affected individuals of the Danish kindred with CHMP2B^Intron5 mutation (7–9). A mouse model of CHMP2B^Intron5 also showed no remarkable cell overproliferation (14). Nevertheless, the results of the present study contribute to the overall understanding of how molecular and cellular defects can manifest in the presence of CHMP2B^Intron5. Here and in our previous study (11), we focused on the most prominent phenotype due to CHMP2B^Intron5 expression in fly eyes, which primarily occurs as a result of misregulation of one receptor and its signaling pathway. The strongest phenotype can likely be attributed to the most active receptor and/or the most sensitive receptor to misregulation during CHMP2B^Intron5 expression. It will be interesting to identify additional receptors misregulated by CHMP2B^Intron5 in eyes and central nervous system neurons.