Mutations in human crumbs 1 (CRB1) are a major cause of retinal diseases that lead to blindness. CRB1 is a transmembrane protein found in the inner segment of photoreceptor cells (PRCs) and the apical membrane of Müller glia. The function of the extracellular region of CRB1 is poorly understood, although more than 80 disease-causing missense mutations have been mapped to it. We have recreated four of these mutations, affecting different extracellular domains, in Drosophila Crumbs (Crb). Crb regulates epithelial polarity and growth, and contributes to PRC differentiation and survival. The mutant Crb isoforms showed a remarkable diversity in protein abundance, subcellular distribution and ability to rescue the lack of endogenous Crb, elicit a gain-of-function phenotype or promote PRC degeneration. Interestingly, although expression of mutant isoforms led to a substantial rescue of the developmental defects seen in crb mutants, they accelerated PRC degeneration compared to that seen in retinas that lacked Crb, indicating that the function of Crb in cellular differentiation and cell survival depends on distinct molecular pathways. Several Crb mutant proteins accumulated abnormally in the rhabdomere and affected rhodopsin trafficking, suggesting that abnormal rhodopsin physiology contributes to Crb/CRB1-associated retinal degeneration.

Mutations in human crumbs 1 (CRB1) are responsible for retinal degeneration in an estimated 80,000 people worldwide. Pathologies associated with CRB1 mutations are characterized by a degeneration of the neural retina; they include Leber congenital amaurosis where blindness is observed around birth and retinitis pigmentosa where patients often develop severe visual impairments as young adults (den Hollander et al., 2008; Ehrenberg et al., 2013; Alves et al., 2014). Human and mouse genomes encode three CRB genes with multiple protein products. All three genes are expressed in the mouse retina, and protein is enriched in the apical membranes of Müller glia cells and the inner segment of the apical membrane of photoreceptor cells (PRCs; Fig. 1A) (Pellikka et al., 2002; Mehalow et al., 2003; van de Pavert et al., 2004; van Rossum et al., 2006). Loss of mouse Crb1 function causes a disorganization of the retinal layers and the formation of rosette-like structures, defects indicative of a loss of integrity in the retinal epithelium (Mehalow et al., 2003; van de Pavert et al., 2004). Similar defects were seen in zebrafish embryos that lacked the function of Crb2a (Omori and Malicki, 2006; Hsu and Jensen, 2010). Gross morphological defects are preceded by problems in the integrity of the zonula adherens (ZA; the outer limiting membrane of the vertebrate retina) that emerge during terminal differentiation of the retina, and a failure to form inner and/or outer segments of normal size. Crb1 Crb2 double mutants show more severe defects than Crb1 mutants and resemble human patients with Leber congenital amaurosis (Pellissier et al., 2013), suggesting that Crb2 also plays an important role in retinal development. Numerous disease-causing mutations in CRB1 have been reported in human patients (Fig. 1C; Table S1) (den Hollander et al., 2004; Bujakowska et al., 2012), which include many missense mutations and short in-frame deletions that affect specific protein domains. How these mutations, in particular in the large extracellular region of the CRB1 transmembrane protein, affect protein function and cause retinal disease is not known.

Human CRB1, CRB2 and CRB3 are orthologues of Drosophila melanogaster crumbs (crb; Tepass et al., 1990). Crb acts as an apical determinant and important regulator of epithelial polarity and ZA integrity (Bulgakova and Knust, 2009; Tepass, 2012). Drosophila Crb is also a key regulator of PRC morphogenesis and survival (Pellikka et al., 2002; Izaddoost et al., 2002; Johnson et al., 2002; Pocha et al., 2011; Chartier et al., 2012). Crb accumulates at the stalk membrane within the PRC apical membrane, which is subdivided into the central light-sensing rhabdomere (the equivalent of the outer segment of vertebrate PRCs) and the surrounding stalk membrane (which corresponds to the inner segment in vertebrates) (Fig. 1B) (Tepass and Harris, 2007). Loss and gain of Crb function leads to defects in Drosophila PRC structure that are very similar to the defects observed in mouse or zebrafish mutants. Loss of Crb causes a fragmentation of the ZA and a shortened stalk membrane (Izaddoost et al., 2002; Pellikka et al., 2002). In addition, loss of Crb results in light-dependent PRC degeneration and death, which may result from abnormal trafficking of the photopigment rhodopsin (Johnson et al., 2002; Pocha et al., 2011) and the excessive production of radical oxygen species resulting from the deregulation of a Rac1-NADPH oxidase pathway (Chartier et al., 2012).

Crb interacts with several cytoplasmic proteins to regulate epithelial apical-basal polarity, Hippo signaling and actomyosin organization (Bulgakova and Knust, 2009; Tepass, 2012). The short, highly conserved cytoplasmic tail of Crb contains a FERM domain-binding site and a PDZ domain-binding site that can interact with several known partners such as the FERM proteins Yurt or Expanded, and the PDZ proteins Stardust (Sdt) and Par6, respectively (Tepass, 2012). The loss of the Crb complex components Sdt, Veli and Patj all lead to light-dependent retinal degeneration (Richard et al., 2006; Berger et al., 2007; Bachmann et al., 2008). In contrast, the molecular and cellular function of the large extracellular region of Crb proteins (Fig. 1C) remains poorly understood, although several studies have highlighted the importance of the extracellular region for Crb function including roles in cell adhesion and Notch signaling (Tepass, 2012; Thompson et al., 2013; Nguyen et al., 2016; Nemetschke and Knust, 2016).

The identification of many missense mutations that change amino acids in the extracellular domain of human CRB1 (Fig. 1C; Table S1) also emphasizes that this region makes important contributions to Crb protein function (den Hollander et al., 2004; Bujakowska et al., 2012). However, it is not clear whether such missense mutations primarily act by reducing CRB1 surface abundance through, for example, endoplasmic reticulum (ER)-mediated decay of a misfolded protein (Griciuc et al., 2011; Brodsky, 2012), or whether these mutations have a specific impact on protein activity at the cell surface that would lead to retinal degeneration. To address this question and to generate animal models for individual human disease-causing CRB1 mutations, we recreated four missense mutations that are linked to retinal degeneration in humans in Drosophila Crb, and tested their protein distribution and function in PRCs.

CRB1 missense mutations preferentially target evolutionarily conserved amino acids

den Hollander and colleagues identified human CRB1 as the retinal disease locus RP12 (Retinitis Pigmentosa 12; den Hollander et al., 1999), and reported several missense mutations in CRB1. Four of these mutations affect amino acids conserved in Drosophila Crb (Fig. 1C). We recreated CRB1 missense mutations in Drosophila crb and tested protein function in a wild-type or a crb mutant background. Since the original report by den Hollander and colleagues, many more retinal disease-associated CRB1 mutations have been found. Virtually all missense mutations that are considered likely to cause retinal defects map to the extracellular region of CRB1 (Bujakowska et al., 2012); 50% of these missense mutations affect residues conserved in Drosophila Crb (Fig. 1C; Table S1), although CRB1 and Crb show only 32% overall identity. This suggests that changes in conserved amino acids are more likely to cause abnormal protein function. CRB1 missense mutations map along most of the protein but show a marked accumulation in the central region that comprises three LamG domains separated by either one or two EGF-like domains (den Hollander et al., 2004; Bujakowska et al., 2012) (Fig. 1C; Table S1). The domain organization of this central region of the protein is, next to the highly conserved cytoplasmic tail, the best conserved portion of Crumbs proteins across metazoans. In contrast, the clusters of EGF-like domains that flank the central part of the Crumbs extracellular region have undergone variation in domain numbers during evolution (Fig. 1C). Three of the four missense mutations we generated map to the central region, in the LamG3 (Crb-T1283M), EGF21 (Crb-N1486S) and EGF22 (Crb-C1540Y) domains, whereas Crb-C749W affects the EGF13 domain within the N-terminal cluster of EGF-like domains (Fig. 1C).

Fig. 1.

CRB1/Crb structure and mutations used in this study. (A) Schematic of human rod PRCs (R) and associated Müller glia (MG) cells. The predicted distribution of CRB1 protein in the human retina is marked in red. CRB1 is found in the basal aspects of the inner segment (IS) just apical to the outer limiting membrane/zonula adherens (OLM/ZA), and in the apical membranes of Müller glia cells. OS, outer segment. (B) Schematic of the PRCs (R1–R7) in a Drosophila ommatidium. Crb is found in the stalk membrane (red) that links the ZA to the light-sensing organelle, the rhabdomere (RB). (C) Protein structures of human CRB1 and Drosophila Crb were aligned with Clustal Omega. Retinal disease-causing missense mutations in CRB1 are indicated by small dots above the protein (see Table S1). Mutations that are affecting an amino acid that is identical in both proteins or that underwent a conservative exchange are indicated with red or orange dots, respectively. The red arrowheads point to the missense mutations recreated in Drosophila Crb. In addition to Crb-FL and the missense mutations, we tested Crb-extraTM::GFP and Crb-intra::GFP in this study. cbEGF, calcium-binding EGF.

Fig. 1.

CRB1/Crb structure and mutations used in this study. (A) Schematic of human rod PRCs (R) and associated Müller glia (MG) cells. The predicted distribution of CRB1 protein in the human retina is marked in red. CRB1 is found in the basal aspects of the inner segment (IS) just apical to the outer limiting membrane/zonula adherens (OLM/ZA), and in the apical membranes of Müller glia cells. OS, outer segment. (B) Schematic of the PRCs (R1–R7) in a Drosophila ommatidium. Crb is found in the stalk membrane (red) that links the ZA to the light-sensing organelle, the rhabdomere (RB). (C) Protein structures of human CRB1 and Drosophila Crb were aligned with Clustal Omega. Retinal disease-causing missense mutations in CRB1 are indicated by small dots above the protein (see Table S1). Mutations that are affecting an amino acid that is identical in both proteins or that underwent a conservative exchange are indicated with red or orange dots, respectively. The red arrowheads point to the missense mutations recreated in Drosophila Crb. In addition to Crb-FL and the missense mutations, we tested Crb-extraTM::GFP and Crb-intra::GFP in this study. cbEGF, calcium-binding EGF.

Essential roles for Crb extracellular and cytoplasmic domains in PRC development

As controls for Crb missense mutations, we first examined the activity of three transgenes: full-length Crb (Crb-FL), and two Crb isoforms in which either the extracellular region (except the signal peptide) was missing (Crb-intra or Crb-intra::GFP), or the cytoplasmic region was replaced by GFP (Crb-extraTM::GFP; Fig. 1C). Our transgenes were inserted at the same genomic site (attP2) in order to minimize differences in protein levels that arise from differences in levels of transcription. All constructs were expressed using the Gal4/UAS system. The impact of these three Crb constructs on PRC development has been described previously (Izaddoost et al., 2002; Pellikka et al., 2002; Richard et al., 2009; Muschalik and Knust, 2011). However, in those studies, the Crb constructs had been expressed from random genomic insertion sites, which may have led to significant differences in expression levels.

We first examined the consequences of overexpression of Crb-FL, Crb-extraTM::GFP and Crb-intra on PRCs. Expression was controlled by the Rh1-Gal4 driver, which is active in PRCs R1 to R6 after 78% of pupal development (PD) and in adults (Kumar and Ready, 1995; Tabuchi et al., 2000). PRCs have established a rhabdomere at 78% PD, although the rhabdomere is not yet fully matured (Longley and Ready, 1995). Flies that carry only Rh1-Gal4 do show defects in rhabdomere shape, with rhabdomeres displaying a less rounded and more square-shaped morphology compared to that seen in wild-type controls (Fig. 2A; Fig. S1). Our previous work with Crb constructs inserted at random sites in the genome suggested that Crb-FL and Crb-extraTM::GFP were equally potent in eliciting a dramatic overgrowth of the stalk membrane (Pellikka et al., 2002). In contrast, expression of transgenes out of attP2 resulted in a much stronger effect for Crb-extraTM::GFP than Crb-FL (Fig. 2B,C,E). Crb-FL showed only small amounts of spongy membranous material associated with the stalk, indicating a moderate overgrowth of the stalk membrane (Fig. 2C, arrows), whereas Crb-extraTM::GFP showed predominantly large cyst-like expansions of the interrhabdomeral space (Fig. 2E, asterisks), suggesting a dramatic overgrowth of the stalk membrane. Expression of Crb-extraTM::GFP also often led to a disruption or loss of rhabdomeres (Fig. 2E, arrows and arrowheads). Consistent with previous findings (Pellikka et al., 2002), overexpression of Crb-intra did not change stalk membrane size but resulted in a collapse of rhabdomere microvilli into the cytoplasm (Fig. 2D, arrows).

Fig. 2.

Overexpression of Crb isoforms causes defects in stalk membrane formation and rhabdomere degeneration. (A) Cross-section of wild-type ommatidium. Scale bar: 1 μm. (B,C) PRCs overexpressing Crb-FL under the control of Rh1-Gal4, which is active in PRCs R1–R6. A close-up of the area in the red box in B is shown in C to illustrate the spongy membranous material associated with the stalk not seen in wild-type (arrows). We also observed that rhabdomeres remain smaller with a more rectangular shape compared to those in wild-type, a defect that can be attributed to Rh1-Gal4 as it was also observed in control ommatidia that do not express a UAS-transgene (see Fig. S1). (D) Overexpression of Crb-intra with Rh1-Gal4 causes some rhabdomere degeneration with membranous material protruding into the cytoplasm of PRCs (arrows). Owing to the late onset of expression (78% PD), Crb-intra overexpression with Rh1-Gal4 does not cause the formation of multiple rhabdomeres per cell in contrast to overexpression with GMR-Gal4 (see Fig. 3D). (E) Overexpression of Crb-extraTM::GFP with Rh1-Gal4 causes cyst-like expansion of the stalk membrane (asterisks) and severe rhabdomere degeneration with rhabdomeres being small (arrows) or absent (arrowheads). (F) Overexpression of Crb-C749W with Rh1-Gal4 causes slight rhabdomere degeneration (arrowheads) and no apparent overgrowth of the stalk membrane. (G–I) Overexpression of Crb-T1283M (G), Crb-N1486S (H), or Crb-C1540Y (I) with Rh1-Gal4 caused PRC defects similar to those seen with overexpression of Crb-extraTM::GFP with cyst-like enlargements of the stalk membranes (asterisks) and severe rhabdomere degeneration evident for Crb-T1283M and Crb-N1486S. Rhabdomere degeneration is less severe with Crb-C1540Y. (J) Crb constructs were overexpressed with Rh1-Gal4 and retinas were stained at 95% PD for F-actin (phalloidin) and Crb. With the exception of Crb-C749W, all Crb isoforms are strongly overabundant. All Crb isoforms are inserted at attP2 except Crb-FL(P), which is a previously characterized P element-based insertion of UAS-Crb (Wodarz et al., 1995; Pellikka et al., 2002). Crb-FL(P) levels are higher than Crb-FL and comparable to those of Crb-extraTM::GFP. Asterisks mark examples of cyst-like expansions of the stalk membrane surrounding interrhabdomeral space. (K) Immunoblot detection (WB) and quantification of Crb protein in wild-type and Crb-overexpressing adult fly head lysates with anti-Crb antibody. Crb constructs were overexpressed through GMR-Gal4. β-Tubulin was used as a loading control. Expression levels of Crb constructs was tested in two or three independent experiments with similar results.

Fig. 2.

Overexpression of Crb isoforms causes defects in stalk membrane formation and rhabdomere degeneration. (A) Cross-section of wild-type ommatidium. Scale bar: 1 μm. (B,C) PRCs overexpressing Crb-FL under the control of Rh1-Gal4, which is active in PRCs R1–R6. A close-up of the area in the red box in B is shown in C to illustrate the spongy membranous material associated with the stalk not seen in wild-type (arrows). We also observed that rhabdomeres remain smaller with a more rectangular shape compared to those in wild-type, a defect that can be attributed to Rh1-Gal4 as it was also observed in control ommatidia that do not express a UAS-transgene (see Fig. S1). (D) Overexpression of Crb-intra with Rh1-Gal4 causes some rhabdomere degeneration with membranous material protruding into the cytoplasm of PRCs (arrows). Owing to the late onset of expression (78% PD), Crb-intra overexpression with Rh1-Gal4 does not cause the formation of multiple rhabdomeres per cell in contrast to overexpression with GMR-Gal4 (see Fig. 3D). (E) Overexpression of Crb-extraTM::GFP with Rh1-Gal4 causes cyst-like expansion of the stalk membrane (asterisks) and severe rhabdomere degeneration with rhabdomeres being small (arrows) or absent (arrowheads). (F) Overexpression of Crb-C749W with Rh1-Gal4 causes slight rhabdomere degeneration (arrowheads) and no apparent overgrowth of the stalk membrane. (G–I) Overexpression of Crb-T1283M (G), Crb-N1486S (H), or Crb-C1540Y (I) with Rh1-Gal4 caused PRC defects similar to those seen with overexpression of Crb-extraTM::GFP with cyst-like enlargements of the stalk membranes (asterisks) and severe rhabdomere degeneration evident for Crb-T1283M and Crb-N1486S. Rhabdomere degeneration is less severe with Crb-C1540Y. (J) Crb constructs were overexpressed with Rh1-Gal4 and retinas were stained at 95% PD for F-actin (phalloidin) and Crb. With the exception of Crb-C749W, all Crb isoforms are strongly overabundant. All Crb isoforms are inserted at attP2 except Crb-FL(P), which is a previously characterized P element-based insertion of UAS-Crb (Wodarz et al., 1995; Pellikka et al., 2002). Crb-FL(P) levels are higher than Crb-FL and comparable to those of Crb-extraTM::GFP. Asterisks mark examples of cyst-like expansions of the stalk membrane surrounding interrhabdomeral space. (K) Immunoblot detection (WB) and quantification of Crb protein in wild-type and Crb-overexpressing adult fly head lysates with anti-Crb antibody. Crb constructs were overexpressed through GMR-Gal4. β-Tubulin was used as a loading control. Expression levels of Crb constructs was tested in two or three independent experiments with similar results.

Loss of Crb in the retina causes severe defects in PRCs, including shorter stalk membranes, a fragmentation of the ZA and rhabdomeres shortened in the proximo-distal direction that, consequently, appear bulkier and are often not separated when viewed in cross-section profiles (Pellikka et al., 2002) (Fig. 3B, arrows). As Crb overexpression causes defects in PRCs, we tested different expression conditions to find an experimental setup in which expression of Crb-FL in a crb mutant retina (for the null allele crb11A22) did not cause gain-of-function defects but restored wild-type morphology to PRCs. We found that expressing Crb-FL with GMR-Gal4 in animals raised at 18°C restored normal PRC and ommatidial structure in 80% of PRCs (see Figs 3C and 5G for quantification). GMR-Gal4 is active in all cells of the developing retina posterior to the morphogenetic furrow, and expression continues into adult flies (Freeman, 1996). In contrast to Crb-FL, we observed no rescue of the crb mutant phenotype in PRCs upon expression of Crb-intra or Crb-extraTM::GFP (Fig. 3D,E). In fact, expression of these constructs in a crb mutant background led to more severe defects than were seen in crb mutants alone (Figs 3D and E, and see Fig. 5G) consistent with previous findings showing that expression of Crb-extraTM::GFP or Crb-intra in wild-type retinas caused structural defects in PRCs (Pellikka et al., 2002; Muschalik and Knust, 2011). Taken together, gain-of-function and rescue experiments with Crb-FL, Crb-intra and Crb-extraTM::GFP indicate that both the extracellular region and the cytoplasmic tail of Crb are essential for normal PRC development.

Fig. 3.

Rescue of crb mutant retinas by expression of Crb mutant isoforms. (A) Cross-section of wild-type ommatidium showing round rhabdomere profiles of PRCs R1–R7 separated by interrhabdomeral space. Scale bar: 1 μm. (B) crb11A22 mutant PRCs show irregular shaped, often broader, rhabdomeres that are in contact, displaying a ‘fused rhabdomere’ defect (arrows). (C) crb11A22 mutant PRCs rescued by the expression of Crb-FL expressed with GMR-Gal4. (D,E) crb11A22 mutant PRCs are not rescued by the expression of Crb-intra (D) or Crb-extraTM::GFP (E). Expression of Crb-intra elicits formation of two or more rhabdomeres in many PRCs (Muschalik and Knust, 2011). Rhabdomeres in individual PRCs are connected by red lines. Expression of Crb-intra or Crb-extraTM::GFP in crb11A22 mutant PRCs disrupts ommatidial organization more severely than seen in crb11A22 mutant PRCs alone. (F–I) Expression of Crb-C749W (F), Crb-T1283M (G), Crb-N1486S (H), or Crb-C1540Y (I) in crb11A22 mutant PRCs causes substantial rescue of crb mutant defects. For quantification of rescue, see Fig. 5G. All mutant constructs were expressed with GMR-Gal4.

Fig. 3.

Rescue of crb mutant retinas by expression of Crb mutant isoforms. (A) Cross-section of wild-type ommatidium showing round rhabdomere profiles of PRCs R1–R7 separated by interrhabdomeral space. Scale bar: 1 μm. (B) crb11A22 mutant PRCs show irregular shaped, often broader, rhabdomeres that are in contact, displaying a ‘fused rhabdomere’ defect (arrows). (C) crb11A22 mutant PRCs rescued by the expression of Crb-FL expressed with GMR-Gal4. (D,E) crb11A22 mutant PRCs are not rescued by the expression of Crb-intra (D) or Crb-extraTM::GFP (E). Expression of Crb-intra elicits formation of two or more rhabdomeres in many PRCs (Muschalik and Knust, 2011). Rhabdomeres in individual PRCs are connected by red lines. Expression of Crb-intra or Crb-extraTM::GFP in crb11A22 mutant PRCs disrupts ommatidial organization more severely than seen in crb11A22 mutant PRCs alone. (F–I) Expression of Crb-C749W (F), Crb-T1283M (G), Crb-N1486S (H), or Crb-C1540Y (I) in crb11A22 mutant PRCs causes substantial rescue of crb mutant defects. For quantification of rescue, see Fig. 5G. All mutant constructs were expressed with GMR-Gal4.

Extracellular and cytoplasmic domains of Crb are required to exclude Crb from the rhabdomere

To assess whether differences in protein levels or distribution correlated with the different effects on stalk membrane development upon overexpression of Crb-FL, Crb-extraTM::GFP, and Crb-intra, we examined levels and distribution of Crb. Protein levels were strongly increased for Crb-FL, and even more so for Crb-extraTM::GFP and a previously generated Crb-FL construct expressed from a random P element insertion site (Pellikka et al., 2002; Wodarz et al., 1995). The latter two proteins did not only decorate the apical membrane but were also found in the cytoplasm and the basolateral membrane (Fig. 2J). The strong overabundance of Crb-FL and Crb-extraTM::GFP when expressed with Rh1-Gal4 did not lend itself to sensitive comparison of protein levels. In contrast, Crb-intra::GFP showed only moderate protein levels by comparison, and was mislocalized to the rhabdomere as reported previously for Crb-intra (Pellikka et al., 2002) (Fig. S2B), indicating that the extracellular region of Crb is required for normal protein localization within the apical membrane of PRCs, and may also contribute to protein surface stability, as the levels of Crb-intra::GFP were low compared to that seen for Crb-FL and Crb-extraTM::GFP. We also confirmed that expression of Crb-intra::GFP and Crb-intra during PRC development led to the formation of ectopic rhabdomeres (Muschalik and Knust, 2011) (Fig. 3D; Fig. S2D). Finally, we noted that overexpression of Crb-intra caused the depletion of endogenous Crb from stalk membranes (Fig. S2C), possibly as a result of the sequestration of cytoplasmic Crb-binding partners, such as Sdt, which is predicted to destabilize Crb at the plasma membrane (Hong et al., 2003; Berger et al., 2007; Bulgakova and Knust, 2009; Lin et al., 2015).

To express Crb isoforms in crb mutant PRCs, we also employed the mosaic analysis with a repressible cell marker (MARCM) technique (Lee and Luo, 1999) (Fig. 4). MARCM generated crb mutant cell clones that were positively labeled with GFP, and at the same time promoted expression of UAS-controlled transgenes in those crb mutant cells. Transgenes were expressed with a Tub-Gal4 driver. Interestingly, we found that MARCM-based expression of Crb-FL in crb mutant PRCs restored normal Crb protein levels (Fig. 4E). In contrast, Crb-extraTM::GFP was strongly overabundant in the apical membrane and localized not only to the stalk but also to the rhabdomere (Fig. 4G). The striking overabundance of Crb-extraTM::GFP compared to that seen for Crb-FL was confirmed through immunoblot analysis (Fig. 2K). MARCM expression of Crb-intra::GFP generated cells with more than one rhabdomere (Fig. S2D). Taken together, our comparison of the impact of Crb-FL, Crb-extraTM::GFP and Crb-intra on PRC development when expressed at comparable transcript levels suggests that the extracellular and the cytoplasmic regions of Crb cooperate in confining Crb to the stalk membrane of PRCs. The strong enrichment of Crb-extraTM::GFP compared to Crb-FL suggests that the cytoplasmic tail of Crb negatively regulates surface abundance of Crb, potentially by facilitating Crb endocytosis (Lin et al., 2015).

Fig. 4.

Crb mutant isoforms show abnormal protein levels and distributions. Expression of Crb isoforms in crb11A22 mutant PRCs was facilitated by use of the MARCM system (see Materials and Methods) in which crb mutant PRCs are generated through mitotic recombination and marked by GFP, and Crb transgene expression under the control of Tub-Gal4 is restricted to GFP-positive cells. PRCs were stained for Crb and GFP at 95% PD (when PRCs are fully differentiated) and at 50% PD (when PRC apical membrane have initiated differentiation into rhabdomere and stalk). Dashed lines circle crb mutant PRCs. (A,B) Schematic illustration of PRCs at 95% and 50% PD. (C,D) crb mutant PRCs lack Crb protein. (E,F) crb mutant PRCs expressing Crb-FL show normal levels of Crb protein that is found at the stalk membrane. (G,H) Crb-extraTM::GFP expressed in crb mutant PRCs is abnormally enriched and distributed throughout the rhabdomere and stalk. (I,J) Protein levels of Crb-C749W expressed in crb mutant PRCs are very low. (K,L) Crb-T1283M expressed in crb mutant PRCs is abnormally enriched and found in the rhabdomere and stalk. (M,N) Crb-N1486S expressed in crb mutant PRCs is abnormally enriched and distributed throughout the rhabdomere and stalk. (O,P) Crb-C1540Y expressed in crb mutant PRCs is abnormally enriched and distributed throughout the rhabdomere and stalk at 50% PD but shows normal levels and distribution at 95% PD.

Fig. 4.

Crb mutant isoforms show abnormal protein levels and distributions. Expression of Crb isoforms in crb11A22 mutant PRCs was facilitated by use of the MARCM system (see Materials and Methods) in which crb mutant PRCs are generated through mitotic recombination and marked by GFP, and Crb transgene expression under the control of Tub-Gal4 is restricted to GFP-positive cells. PRCs were stained for Crb and GFP at 95% PD (when PRCs are fully differentiated) and at 50% PD (when PRC apical membrane have initiated differentiation into rhabdomere and stalk). Dashed lines circle crb mutant PRCs. (A,B) Schematic illustration of PRCs at 95% and 50% PD. (C,D) crb mutant PRCs lack Crb protein. (E,F) crb mutant PRCs expressing Crb-FL show normal levels of Crb protein that is found at the stalk membrane. (G,H) Crb-extraTM::GFP expressed in crb mutant PRCs is abnormally enriched and distributed throughout the rhabdomere and stalk. (I,J) Protein levels of Crb-C749W expressed in crb mutant PRCs are very low. (K,L) Crb-T1283M expressed in crb mutant PRCs is abnormally enriched and found in the rhabdomere and stalk. (M,N) Crb-N1486S expressed in crb mutant PRCs is abnormally enriched and distributed throughout the rhabdomere and stalk. (O,P) Crb-C1540Y expressed in crb mutant PRCs is abnormally enriched and distributed throughout the rhabdomere and stalk at 50% PD but shows normal levels and distribution at 95% PD.

Three out of four missense mutations cause abnormal Crb protein enrichment and misdistribution to the rhabdomere

Crb-C749W

EGF-like domains are characterized by six cysteine residues (denoted C1 to C6) that form three stereotypic cystine bridges (C1–C3, C2–C4 and C5–C6). Crb-C749W affects C5 and therefore compromises the third cystine bridge in EGF13 of the Crb protein (Fig. 1C). Expression of Crb-C749W in crb mutant retinas caused 42% of PRCs to display a normal morphology compared to the 11% seen in crb mutants and the 80% seen in crb mutants expressing Crb-FL under our experimental conditions (Figs 3F and 5G). MARCM analysis (Fig. 4I) indicated that Crb-C749W protein levels remained very low and were sometimes undetectable. Similarly, overexpression of Crb-C749W led to protein levels similar to that seen for Crb in wild-type retinas (Fig. 2J,K) and elicited no significant overgrowth of the stalk membrane (Fig. 2F). These findings suggest that Crb-C749W is at least a partially functional protein and may retain full function once at the cell surface because the lower performance of Crb-C749W in comparison to Crb-FL can be rationalized by the lower observed cell surface levels of Crb-C749W.

Crb-T1283M

Crb-T1283M affects the third LamG domain of Crb (LamG3) (Fig. 1C). Expression of Crb-T1283M in crb mutants restored 64% of PRCs to normal morphology (Figs 3G and 5G). Crb-T1283M was strongly overabundant when expressed with Rh1-Gal4 or GMR-Gal4 (Fig. 2J,K), and elicited a striking enlargement of the stalk membrane leading to the formation of large cysts similar to what is seen with Crb-extraTM::GFP (Fig. 2G). MARCM analysis indicated that Crb-T1283M is more abundant than Crb-FL and, interestingly, mislocalized to the rhabdomere, similar to what was seen for Crb-extraTM::GFP (Fig. 4K). Taken together, these observations suggest that Crb-T1283M is an almost fully functional protein given its strong rescue of the crb loss-of-function defects. However, Crb-T1283M was partly mislocalized to the rhabdomere indicating that LamG3 is required to regulate protein distribution in PRCs. Moreover, as with Crb-extraTM::GFP, the overabundance of Crb-T1283M correlates with the strong gain-of-function effect that massively enlarges the stalk membrane.

Fig. 5.

PRC degeneration under light stress is accelerated by Crb missense mutations. (A–G) Wild-type (wt), crb11A22 mutant PRCs (crb−/−), or crb11A22 mutant PRCs in which Crb isoforms were expressed with GMR-Gal4 (as in Fig. 3), were exposed to normal light (NL) or continuous light (light stress, LS) for 7 days at 18°C after which retinas were examined by transmission electron microscopy. For quantification of defects, PRCs were scored as: normal (A); having abnormally shaped rhabdomeres, as seen in crb mutants (B); dead (C), and having degenerating rhabdomeres showing a split rhabdomere (D, arrows), a frayed rhabdomere base (E, arrows), or more severe signs of degeneration (F, arrows), including absent rhabdomeres. (G) Quantification of PRC defects. Light stress does not elicit PRC degeneration in wild-type flies. We carried out pair-wise comparisons of normal light versus light stress for each genotype taking all four categories of defects into account. Except for wild-type flies, differences are highly significant in each case (***P>0.0001). Light stress always increases the severity of the phenotype. We also assessed the ability of Crb mutant isoforms to rescue crb mutant defects in comparison to rescue achieved by Crb-FL. In all cases, significant differences were found with Crb-intra and Crb-extraTM::GFP worsening the phenotype, and Crb proteins carrying missense mutations showing substantial levels of rescue (***P>0.0001). Finally, we compared crb mutant PRCs with crb mutant PRCs expressing Crb isoforms when exposed to light stress. For this comparison, we only took into account the number of degenerating or dead PRCs for each genotype versus cells showing no signs of degeneration. Crb-FL expression showed a significant rescue whereas no significant differences were seen with Crb-intra. All other Crb isoforms caused a significant enhancement of PRC degeneration (**P=0.0015; ***P>0.0001). Statistical significance was assessed with two-tailed unpaired t-tests. n values are given in Table S2.

Fig. 5.

PRC degeneration under light stress is accelerated by Crb missense mutations. (A–G) Wild-type (wt), crb11A22 mutant PRCs (crb−/−), or crb11A22 mutant PRCs in which Crb isoforms were expressed with GMR-Gal4 (as in Fig. 3), were exposed to normal light (NL) or continuous light (light stress, LS) for 7 days at 18°C after which retinas were examined by transmission electron microscopy. For quantification of defects, PRCs were scored as: normal (A); having abnormally shaped rhabdomeres, as seen in crb mutants (B); dead (C), and having degenerating rhabdomeres showing a split rhabdomere (D, arrows), a frayed rhabdomere base (E, arrows), or more severe signs of degeneration (F, arrows), including absent rhabdomeres. (G) Quantification of PRC defects. Light stress does not elicit PRC degeneration in wild-type flies. We carried out pair-wise comparisons of normal light versus light stress for each genotype taking all four categories of defects into account. Except for wild-type flies, differences are highly significant in each case (***P>0.0001). Light stress always increases the severity of the phenotype. We also assessed the ability of Crb mutant isoforms to rescue crb mutant defects in comparison to rescue achieved by Crb-FL. In all cases, significant differences were found with Crb-intra and Crb-extraTM::GFP worsening the phenotype, and Crb proteins carrying missense mutations showing substantial levels of rescue (***P>0.0001). Finally, we compared crb mutant PRCs with crb mutant PRCs expressing Crb isoforms when exposed to light stress. For this comparison, we only took into account the number of degenerating or dead PRCs for each genotype versus cells showing no signs of degeneration. Crb-FL expression showed a significant rescue whereas no significant differences were seen with Crb-intra. All other Crb isoforms caused a significant enhancement of PRC degeneration (**P=0.0015; ***P>0.0001). Statistical significance was assessed with two-tailed unpaired t-tests. n values are given in Table S2.

Crb-N1486S

The Crb-N1486S mutation is located in EGF21, the first of two EGF-like domains separating LamG3 and LamG4 (Fig. 1C). When Crb-N1486S was expressed in crb mutant retinas, a rescue with 48% normal PRCs was observed, compared to 11% in crb mutants and 80% in crb mutants expressing Crb-FL (Figs 3H and 5G). Crb-N1486S levels in PRCs were strongly increased when overexpressed (Fig. 2J,K), and we observed a corresponding dramatic overgrowth of the stalk membrane (Fig. 2H). In MARCM clones, it was readily apparent that Crb-N1486S becomes dramatically enriched at the apical membrane, much more so than Crb-FL. Similar to Crb-extraTM::GFP and Crb-T1283M, Crb-N1486S was found to be distributed throughout the apical membrane in both the stalk and rhabdomere (Fig. 4M). Our findings suggest that Crb-N1486S has enhanced surface abundance correlating with an ability to strongly enlarge the stalk membrane, but a reduced function compared to Crb-FL. Similar to LamG3, EGF21 is also required to control the normal subcellular distribution of Crb and to exclude Crb from the rhabdomere.

Crb-C1540Y

EGF22 is mutated in Crb-C1540Y (Fig. 1C). EGF22 in Drosophila, its counterpart EGF14 in CRB1 and the equivalent EGF domains in other Crumbs proteins show a conserved loss of C5 and C6 and are therefore predicted to form only two cystine bridges. Because of the loss of the last two cysteine residues, protein domain recognition software programs, such as SMART (http://smart.embl-heidelberg.de/; Letunic et al., 2015) often fail to identify EGF22 or its homologous domains in other Crumbs proteins as an EGF-like domain. Crb-C1540Y affects the C4 residue of EGF22 and is therefore likely to severely compromise the structure of this truncated EGF-like domain. Overexpression of Crb-C1540Y through Rh1-Gal4 or GMR-Gal4 caused the protein to be overabundant in PRCs (Fig. 2J,K), with cells showing a corresponding elongation of the stalk membrane (Fig. 2I). Interestingly, MARCM analysis showed that the levels and distribution of Crb-C1540Y was indistinguishable from the Crb-FL control in fully differentiated PRCs (Fig. 4O). When Crb-C1540Y was expressed in the crb mutant retina, we found a level of rescue close to that for the Crb-FL control (71% vs 80%; Figs 3I and 5G). Thus, in contrast to other missense mutations, Crb-C1540Y behaved rather normally with respect to protein levels and protein distribution in fully formed PRCs, and the ability to substitute for Crb in promoting normal PRC development. However, it also can elicit a strong gain-of-function phenotype, in contrast to Crb-FL, when overexpressed under our experimental conditions.

Abnormal distributions of Crb mutant proteins originate early during PRC differentiation

To ask whether the defects in Crb protein distribution or levels observed in fully formed PRCs originated earlier during development, we examined PRCs at 50% PD. The differentiation of the PRC apical membranes is initiated at mid-pupal stages as the apical membranes separate to form the lumen of the interrhabdomeral space and the apical membranes segregate into rhabdomere and stalk (Longley and Ready, 1995). The distribution and qualitative levels of Crb isoforms at 50% PD were found to be similar to those in mature PRCs in most cases (Fig. 4). Crb-FL levels were normal (Fig. 4F) and Crb-C749W levels were very low (Fig. 4J). In contrast, Crb-extraTM::GFP (Fig. 4H), Crb-T1283M (Fig. 4L) and Crb-N1486S (Fig. 4N) were abnormally enriched and misdistributed to the rhabdomere. Notably, we found differences between 50% and 95% PD for Crb-C1540Y, which was overabundant at 50% PD and also mislocalized to the rhabdomere (Fig. 4P), compared to what was seen in fully formed PRCs, where its distribution and levels were normal (Fig. 4O). Taken together, all Crb mutant isoforms showed differences in protein distribution and levels early during PRC differentiation, which were corrected during further development only for Crb-C1540Y.

Crb interacts with several cytoplasmic binding partners, including Sdt, Patj and Myosin V (Pocha et al., 2011; Tepass, 2012). We asked whether Crb isoforms that distribute to rhabdomeres co-recruit binding partners. We found no evidence for mislocalization of Crb-binding partners to the rhabdomere in PRCs expressing Crb-extraTM::GFP, Crb-T1283M or Crb-N1486S instead of Crb (Fig. S4).

Crb missense mutations enhance retinal degeneration of crb-null mutants

One important question concerns the relationship between PRC morphogenesis and homeostasis. Vertebrate and fly Crumbs proteins are involved in both processes, raising the possibility that a loss of normal cell structure seen in crb mutants precipitates PRC degeneration. Alternatively, these two functions of Crumbs proteins may be distinct, and this might be revealed through the analysis of Crb mutations that have an impact on PRC homeostasis but not PRC morphogenesis. Flies with crb mutant retinas kept under a normal light–dark cycle (‘normal light’) show pervasive signs of retinal degeneration after 20 days (Pocha et al., 2011). PRC degeneration in response to light is progressive (Lee and Montell, 2004), and, in crb mutants, can be accelerated by exposing flies to continuous light (‘light stress’) (Johnson et al., 2002). We exposed wild-type controls, animals with crb mutant eyes, and crb mutants that expressed one of our constructs to continuous light for 7 days at 18°C to induce PRC degeneration and cell death. Control flies were maintained under normal light for 7 days. Quantification of PRC defects was based on examination of ultrathin sections. We distinguished four classes of PRC defects (Fig. 5A–F): (0) normal rhabdomeres (Fig. 5A); (1) abnormal rhabdomeres with broader cross-section profiles, as typically seen in crb-null mutants (Fig. 5B); (2) degenerating rhabdomeres, including split rhabdomeres and rhabdomere fragmentation, disruption of microvilli organization, collapse of microvilli material into the cell body, and absence of the rhabdomere (Fig. 5D–F); and (3) dead PRCs, as indicated by an electron-dense cytoplasm (Fig. 5C).

In contrast to wild-type PRCs, which showed no sign of degeneration or cell death under normal or light stress conditions, crb mutants displayed 8% degenerating PRCs under normal light conditions, and under light stress showed 13% degenerating rhabdomeres and 11% dead PRCs (Fig. 5G). Expression of Crb-FL in crb mutant PRCs limited rhabdomere degeneration to ∼3% under normal light and light stress. Expression of Crb-intra in crb mutant PRCs resulted in similar degeneration/death values in normal light and light stress to that seen with crb mutants alone. In contrast, expression of Crb-extraTM::GFP enhanced the phenotype of crb mutants, with 22% degenerating PRCs under normal light and 15% degenerating PRCs and 27% dead PRCs under light stress (Fig. 5G).

The behavior of the four Crb missense mutations in our light stress test was striking (Fig. 5G). Whereas expression of these mutations in crb mutant PRCs did not enhance the number of degenerating cells significantly in normal light, we found that, under light stress, each mutation enhanced degeneration and death of PRCs compared to the crb mutant control. The effect was particularly strong for Crb-T1283M, which showed a combined 89% degenerated (21%) and dead (68%) PRCs (Fig. 5G). We also noticed that light stress led to an increase in the fraction of PRCs with abnormally shaped rhabdomeres in some cases, such as upon expression of Crb-FL and Crb-C1540Y, suggesting that abnormal PRC shape may be an early sign of PRC degeneration. Taken together, expression of Crb missense mutations or Crb-extraTM::GFP does not simply fail to rescue the crb loss-of-function defects but has a negative impact on PRC survival. Moreover, although Crb missense mutations accelerate PRC degeneration, they significantly rescue the morphological defects of crb mutants, suggesting that Crb function in PRC survival is largely independent of its role in PRC differentiation.

Crb-T1283M displaces rhodopsin from the rhabdomere

How could expression of a Crb missense mutation enhance PRC degeneration to a higher degree than is observed in PRCs that lack Crb? Four of the mutant Crb proteins showed ectopic distribution to the rhabdomere, suggesting that they might directly interfere with rhabdomere structure or physiology. Mutations in the rhabdomere protein rhodopsin are the major cause for PRC degeneration in both flies and humans, and even small abnormalities in biosynthetic or endosomal trafficking of rhodopsin cause PRC degeneration (Xiong and Bellen, 2013). We therefore examined Rhodopsin 1 (Rh1, also known as NinaE in flies) distribution in crb mutant PRCs expressing our transgenes. Rh1 is often seen in large endosomal vesicles in wild-type or crb mutant PRCs (Fig. 6A). The number of endosomal Rh1-positive vesicles and staining intensity increases with light-induced endocytosis of Rh1 (Satoh and Ready, 2005). In contrast, Crb was not detected in vesicular compartments in normal PRCs. Crb-FL, Crb-intra, Crb-C749W and Crb-C1540Y expression in crb mutant PRCs did not lead to any obvious enrichment of Rh1 in endosomes or elsewhere in the cytoplasm (Fig. 6B; Fig. S3). Interestingly, Crb-extraTM::GFP and Crb-N1486S, which strongly accumulate in the rhabdomere when expressed in crb mutant PRCs, colocalize with Rh1 in vesicular compartments (Fig. 6C,D). Rh1 staining appears more intense in the vesicular compartments, suggesting that Crb-extraTM::GFP and Crb-N1486S expression may enhance Rh1 endocytosis. We detected colocalization of Crb-extraTM::GFP with Vps26, a component of the retromer complex (Wang et al., 2014) (Fig. S5), confirming the endosomal distribution of Crb-extraTM::GFP. In contrast, no colocalization of Crb-N1486S with either Vps26 or Hrs (a component of the ESCRT-0 complex) was observed (Fig. S5).

Fig. 6.

Crb-T1283M displaces rhodopsin from rhabdomeres. Crb isoforms were expressed in crb11A22 mutant PRCs by use of MARCM (see Materials and Methods) in which crb mutant PRCs are marked by GFP, and Crb transgene expression under the control of Tub-Gal4 is restricted to GFP-positive cells. crb mutant PRCs were stained for Crb, Rhodopsin 1 (Rh1), F-actin (phalloidin) and GFP at 95% PD. Dashed outlines mark crb mutant GFP-positive cells. (A,B) crb mutant PRCs (A) and crb mutant PRCs in which Crb-FL is expressed show normal distribution of Rh1. In wild-type and mutant PRCs, Rh1 is often seen in endosomes (arrows). (C,D) crb mutant PRCs expressing Crb-extraTM::GFP or Crb-N1486S show no apparent reduction in the level of Rh1 in rhabdomeres but show a larger number of Rh1-positive endosomes that also contain Crb (arrows). (E,F) crb mutant PRCs expressing Crb-T1283M show reduced levels of Rh1 in rhadomeres and endosomes (arrows).

Fig. 6.

Crb-T1283M displaces rhodopsin from rhabdomeres. Crb isoforms were expressed in crb11A22 mutant PRCs by use of MARCM (see Materials and Methods) in which crb mutant PRCs are marked by GFP, and Crb transgene expression under the control of Tub-Gal4 is restricted to GFP-positive cells. crb mutant PRCs were stained for Crb, Rhodopsin 1 (Rh1), F-actin (phalloidin) and GFP at 95% PD. Dashed outlines mark crb mutant GFP-positive cells. (A,B) crb mutant PRCs (A) and crb mutant PRCs in which Crb-FL is expressed show normal distribution of Rh1. In wild-type and mutant PRCs, Rh1 is often seen in endosomes (arrows). (C,D) crb mutant PRCs expressing Crb-extraTM::GFP or Crb-N1486S show no apparent reduction in the level of Rh1 in rhabdomeres but show a larger number of Rh1-positive endosomes that also contain Crb (arrows). (E,F) crb mutant PRCs expressing Crb-T1283M show reduced levels of Rh1 in rhadomeres and endosomes (arrows).

Interestingly, we noticed a striking loss of Rh1 from the rhabdomeres with expression of Crb-T1283M (Fig. 6E,F). The level of Crb-T1283M protein in the rhabdomere is significantly lower than that of Crb-extraTM::GFP or Crb-N1486S, which do not noticeably reduce Rh1, suggesting that Crb-T1283M does not simply displace Rh1 in the rhabdomere through mass action, but that the loss of Rh1 is a specific consequence of the Crb-T1283M missense mutation. Vesicular Crb-T1283M colocalized with low levels of Rh1 and with Hrs in endosomes (Fig. 6E,F; Fig. S5). Taken together, our findings suggest that Crb isoforms that are mislocalized to the rhabdomere cause enhanced endocytosis or abnormal endocytotic trafficking of Rh1, which could contribute to PRC degeneration.

The Human Gene Mutation Database currently lists more than 200,000 genetic variants associated with human disease, of which ∼56% are missense mutations (Stenson et al., 2014). Uncovering the specific consequences of missense mutations on protein function is a major challenge in understanding disease mechanisms. One possibility is that missense mutations that disrupt protein function cause defects in protein structure that lead to instability and degradation. Such mutations should behave as loss-of-function alleles and could in principle be compensated for by providing normal gene product through gene therapy approaches. However, recent analysis of a large number of missense mutations linked to Mendelian disorders suggests that the majority of such mutations do not show abnormal folding and instability but result in specific defects in protein function, such as the disruption of protein–protein interaction (Sahni et al., 2015). Analyzing missense mutations in suitable models will be essential to identifying the molecular and cellular consequences of such mutations.

Table 1 summarizes our data on the different Crb mutant isoforms. Of the four retinal disease-associated CRB1 mutations that we recreated in Drosophila crb, only Crb-C749W showed reduced protein levels. A mouse model for human CRB1-C250W, Crb1-C249W (corresponding to Drosophila Crb-C749W) was generated through a knock-in approach (van de Pavert et al., 2007). Mouse Crb1-C249W showed normal protein distribution, similar to what is seen for Crb-C749W in Drosophila, but in contrast to the Drosophila protein, also showed normal protein levels. The differences in protein levels may result from differences in how the ER-associated degradation machinery handles misfolded proteins in flies and mice. Like Crb-C749W, Crb1-C249W-expressing retinas showed weaker defects than retinas lacking Crb1 (van de Pavert et al., 2007), indicating that both mutant Crumbs proteins are not null alleles. As in Drosophila, defects in Crb1-C249W-expressing mice could be enhanced by exposure to light (van de Pavert et al., 2007). This comparison suggests that Crb-C749W and Crb1-C249W elicit similar responses in fly and mouse models, respectively. However, a noticeable difference is that Crb1-C249W mice, even with light stress, showed much weaker retinal degeneration pathologies than Crb1−/− mice, whereas Crb-C749W-expressing crb mutant retinas showed levels of degeneration stronger than those seen in crb mutants. Human patients expressing CRB1-C250W show a range of defects, including the most severe forms of CRB1-associated phenotypes (Bujakowska et al., 2012). The large phenotypic variation associated with CRB1/Crb1 mutations in humans and mice was attributed to genetic background effects (Mehalow et al., 2003; Bujakowska et al., 2012), and correlate with observations that the crb mutant phenotype in flies can be modified through, for example, mutations in other polarity proteins (e.g. Tanentzapf and Tepass, 2003; Laprise et al., 2009) or interactions with other proteins required for PRC survival (Gurudev et al., 2014).

Table 1.

Summary of Crb isoform analysis

Summary of Crb isoform analysis
Summary of Crb isoform analysis

Protein levels of missense mutant isoforms other than Crb-C749W were dramatically increased, and we observed either a transient (Crb-C1540Y) or permanent (Crb-T1283M and Crb-N1486S) mislocalization of Crb mutant proteins to the rhabdomere. Crb-T1283M, Crb-N1486S and Crb-C1540Y affect three adjacent domains (LamG3, EGF21 and EGF22, respectively) in the highly conserved central region of the Crb protein. The common behavior of all three mutant isoforms suggests that this part of the extracellular region of Crb regulates protein abundance and subcellular distribution in PRCs. Moreover, similar defects in protein abundance and distribution were found for Crb-extraTM::GFP. In contrast to previous work, we were able to clearly detect differences between Crb-FL and Crb-extraTM::GFP because we normalized transcription through expression of transgenes from the same genomic site in combination with a MARCM expression strategy that generated normal protein levels and distribution for Crb-FL. Taken together, we conclude that both the cytoplasmic domain of Crb and the LamG3/EGF21/EGF22 domains of the extracellular region are crucial for limiting Crb surface abundance, and for excluding Crb from the rhabdomere and restricting it to the stalk membrane. Recently, it was shown that LamG4 of Crb also controls Crb protein abundance in the Drosophila wing imaginal disc (Nguyen et al., 2016). The region in CRB1 that corresponds to LamG3/EGF21/EGF22/LamG4 in Drosophila is a hotspot for retinal disease-causing mutations (Fig. 1; den Hollander et al., 2004; Bujakowska et al., 2012), suggesting that upregulation and/or misdistribution of CRB1 protein levels is an important disease mechanisms.

Early during PRC development when apical membranes of PRCs are attached to each other, Crb is found throughout the apical membrane. However, as the lumen of the interhabdomeral space begins to form, Crb is removed from the emerging rhabdomere and becomes confined to the stalk membrane by ∼50% PD (Izaddoost et al., 2002; Pellikka et al., 2002). One possibility is that endocytotic removal of mutant Crb proteins from the developing rhabdomere is compromised. However, mutant Crb proteins that are enriched in the rhabdomere also colocalize with Rh1 in endosomes indicating that Crb proteins are endocytosed. Crb endocytosis depends on the AP-2 clathrin adaptor complex, which directly interacts with the Crb cytoplasmic tail to facilitate endocytosis (Fletcher et al., 2012; Lin et al., 2015), suggesting that at least Crb-extraTM::GFP shows higher steady-state levels in the rhabdomere as a result of reduced endocytosis.

One of the established causes of retinal degeneration is abnormal trafficking of rhodopsin in either the biosynthetic or endocytotic pathway (Wang and Montell, 2007; Xiong and Bellen, 2013; Wang et al., 2014). Retinal degeneration of crb mutant PRCs also appears to depend on abnormal rhodopsin physiology, as the depletion of rhodopsin from crb mutant PRCs through raising animals on a vitamin-A-poor diet, which reduces rhodopsin to 3% of normal levels, rescues PRC degeneration (Johnson et al., 2002). Moreover, Crb has been directly implicated in rhodopsin biosynthetic transport through its physical interaction and stabilization of Myosin V, an atypical myosin engaged in vesicle transport (Pocha et al., 2011). Crb mutant isoforms that accumulated in the rhabdomere (Crb-extraTM::GFP and Crb-N1486S) colocalized with rhodopsin in endosomes and caused a more intense rhodopsin labeling of endosomes consistent with enhanced endocytosis or defective endocytotic processing. The most notable rhodopsin defect was seen in PRCs that express Crb-T1283M instead of endogenous protein. Here, rhodopsin was strongly depleted from rhabdomeres. This dramatic impact on rhodopsin levels correlated with the most severe retinal degeneration we observed for any of the Crb mutant isoforms we tested.

In conclusion, our findings suggest that three domains within the Crb extracellular region, which are affected by numerous missense mutations in CRB1 associated with retinal disease, are required to limit Crb surface abundance and control its subcellular distribution. The examined mutations also accelerate retinal degeneration more so than is observed in crb-null retinas, suggesting novel detrimental activities of these mutant proteins. Our results are consistent with the hypothesis that defects in rhodopsin trafficking are crucial downstream effectors of at least some Crb missense mutations that cause retinal degeneration. Crb-dependent PRC degeneration mediated through its effect on rhodopsin physiology would also explain why Crb missense mutations could accelerate PRC degeneration while substantially rescuing the morphological integrity of PRCs of crb mutants.

Generation of Crb constructs

The pUAST-crb plasmid (Wodarz et al., 1993, 1995), which contains the full-length crb mini-gene, was modified by inserting an attB site-specific recombination target sequence into the NsiI site of pUAST, resulting in pUAST-Crb-FL-attB. The crb missense mutations Crb-C749W, Crb-T1283M, Crb-N1486S and Crb-C1540Y were made using the QuikChange site-directed mutagenesis kit (Stratagene). Crb-intra::GFP was generated by modification of pUAST-Crbintra-myc (Wodarz et al., 1995) to insert EGFP at residue 110 of Crbintra-myc, between the Myc tag and the transmembrane domain, by using a PCR-based strategy, resulting in pUAST-Crb-intra::GFP. attB was then inserted at NsiI into pUAST-Crbintra-myc and pUAST-Crb-intra::GFP. The pUAST-Crb-extraTM::GFP plasmid (Pellikka et al., 2002) was also modified by insertion of attB at NsiI. To generate transgenic lines, all constructs were inserted into the attP2 recombination target site located on the left arm of the third chromosome (Groth et al., 2004).

Overexpression and clonal analysis

For overexpression analysis (Fig. 2, Figs S1 and S2B,C), we drove UAS transgenes with Rh1-Gal4 (Tabuchi et al., 2000) at 29°C to maximize Gal4 activity in PRCs.

To express UAS transgenes in crb mutant retinas (Fig. 3), we induced mitotic recombination with eyeless promoter-controlled flip recombinase (ey-FLP) as described previously (Pellikka et al., 2002). Mitotic recombination in crb heterozygous animals is induced in virtually all cells of the eye with subsequent removal of homozygous crb+/+ cells through a recessive lethal mutation, while crb−/− cells survive and populate the retina. In this background, we expressed UAS transgenes with GMR-Gal4 at 18°C. The full genotype for these experiments was: y w ey-FLP; GMR-Gal4/+; UAS-Crb-XXX FRT82B crb11A22/FRT82B P{w+}90E l(3)cl-R3 where UAS-Crb-XXX represents any of our transgenes inserted at attP2.

We used the MARCM system (Lee and Luo, 1999) to generate small crb mutant cell clones in the retina that are positively marked by mCD8::GFP and express Crb transgenes (Figs 4 and 6; Figs S2D, S3, S4 and S5). FRT-mediated mitotic recombination was induced by heat-shock promoter-controlled Flip recombinase (hs-FLP). Flies were heat shocked for 2 h at 37°C during the first and second larval instar. Pupae were dissected at 50% and 95% PD to retrieve retinas for antibody staining. The y w hs-FLP; tub-Gal4 UAS-GFP; FRT82B tub-Gal80 stock used here was a gift from Norbert Perrimon, Harvard Medical School, Boston, MA. The full genotype for these experiments was: y w hs-FLP; tub-Gal4 UAS-mCD8::GFP; UAS-Crb-XXX FRT82B crb11A22/FRT82B tub-Gal80 where UAS-Crb-XXX represents any of our transgenes inserted at attP2.

Antibody staining

Pupal retinas were dissected in phosphate buffer (pH 7.4), and the cornea was removed with a tungsten needle. The retinal tissue was fixed for 15 min in phosphate buffer (pH 7.4) containing 4% formaldehyde. Antibody staining was performed following standard procedures. The following primary antibodies were used: rat polyclonal antibody (pAb) anti-Crb (F1, 1:500; Pellikka et al., 2002); mouse monoclonal antibody (mAb) anti-Rhodopsin1 (4C5, 1:50; Developmental Studies Hybridoma Bank, DSHB); guinea pig pAb Hrs (Lloyd et al., 2002); guinea pig pAb anti-Vps26 (1:1000; Wang et al., 2014); rabbit pAb against Stardust (1:4000; gift from Elizabeth Knust, Max-Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany); rabbit affinity-purified anti-MyoV pAb (1:1000; Li et al., 2007) rabbit pAb anti-Patj (1:1000; Pellikka et al., 2002); rabbit pAb anti-GFP::Alexa Fluor 488 (A-21311, 1:1000; Molecular Probes/Thermo Fisher Scientific). Secondary antibodies conjugated to Alexa Fluor dyes (Molecular Probes/Thermo Fisher Scientific) were used at a dilution of 1:400. F-actin was detected with Alexa Fluor 488-conjugated Phalloidin (A-12379, 1:20; Molecular Probes/Thermo Fisher Scientific) at a dilution of 1:25 and was added together with the secondary antibodies. Confocal images were obtained with a Zeiss LSM510 laser-scanning confocal microscope using a Plan-Apochromat 100× NA 1.40 oil lens.

Immunoblotting

We used the GMR-Gal4 driver to assess the total amount of Crb protein expression in fly heads carrying UAS-Crb transgenes in immunoblots. Flies were grown at 25°C and heads were dissected 1 day after hatching, and homogenized in 1× Laemmli Sample buffer (Bio-Rad). The homogenates were centrifuged at 16,000 g for 15 min at 4°C, and supernatants were quantified by using the Bradford method. Samples (10 μg/lane) were boiled for 5 min, briefly centrifuged, and subjected to SDS-PAGE in Bolt 4-20% Bis-Tris Plus separating gels (Novex). Proteins were transferred to nitrocellulose membranes (iBolt2 NC), and immunoblotted with one of the following antibodies: rat pAb anti-Crb (1:500; Pellikka et al., 2002), mouse mAb anti-β-tubulin (E7, 1:1000; Developmental Studies Hybridoma Bank, DSHB). The appropriate anti-rat IRDye680LT (926-68029) and anti-mouse IRDye800CW (926-32212)-conjugated secondary antibody (1:20,000; LI-COR) was applied, and signal was detected by IR luminescence (Li-COR, Model 2800).

Electron microscopy

For transmission electron microscopy, adult flies were immersed in ice-cold fixative (2.5% glutaraldehyde, 2% formaldehyde, 0.1 M sodium cacodylate, pH 7.4) and heads were bisected. After fixation at 4°C for 3 days, eyes were treated with 1% osmium tetroxide in 0.1 M sodium cacodylate (pH 7.4) and 0.1 M sorbitol for 1 h. Eyes were dehydrated and embedded in Spurr's resin. Ultrathin sections were post-stained with uranyl acetate and lead citrate. Sections were examined using a Hitachi H7000 transmission electron microscope (TEM) operated at 75 kV and images were obtained using an AMT XR-111 digital camera with AMT Capture Engine software (version 5.03).

Light stress treatment

For ‘light stress’ treatment, animals were exposed to constant light at 18°C for 7 days. We used a compact fluorescent light (Sylvania CF15EL/830, 15 W, 3000 K, 850 lumens), which was placed ∼20 cm from the flies. For normal light conditions, flies were maintained in an incubator at 18°C with a 12-h-dark–12-h-light cycle.

Quantification and statistical analysis

Rhabdomere and PRC counts shown in Fig. 5G were done on TEM images using ImageJ software. Statistical analysis was performed through a Student's t-test [Prism software (GraphPad) and Excel (Microsoft Office)]. For the purpose of quantification, PRC defects were divided into four categories: normal morphology (0), abnormal morphology (1), degenerating (2), and dead (3). All four categories were taken into account when making pair-wise comparisons between normal light and light stress conditions for flies of the same genotype. In order to evaluate retinal degeneration as a result of light stress treatment, we grouped together PRCs with normally and abnormally shaped rhabdomeres (0) and PRCs with degenerating rhabdomeres and dead PRCs (1). Finally, we compared the PRCs of normal morphology in crb mutant retinas to crb mutant retinas expressing a Crb construct to assess the significance of phenotypic rescue as a result of construct expression under normal light conditions.

We would like to thank Michelle Li for technical assistance, Kenana Al Kakouni for providing the Crb-intra::GFP construct, Norbert Perimon and Hugo Bellen for reagents, and the Imaging Facility of the Department of Cell and Systems Biology at the University of Toronto for technical support. We thank Dorothea Godt for critical reading of the manuscript and discussion.

Author contributions

Conceptualization: U.T.; Methodology: U.T., M.P.; Validation: U.T., M.P.; Formal analysis: M.P.; Investigation: U.T., M.P.; Resources: U.T., M.P.; Data curation: M.P.; Writing - original draft: U.T.; Writing - review & editing: U.T., M.P.; Visualization: U.T., M.P.; Supervision: U.T.; Project administration: U.T.; Funding acquisition: U.T.

Funding

The project was funded by an operating grant from the Canadian Institutes of Health Research (MOP-14372) to U.T.

Alves
,
C. H.
,
Pellissier
,
L. P.
,
Vos
,
R. M.
,
Garcia Garrido
,
M.
,
Sothilingam
,
V.
,
Seide
,
C.
,
Beck
,
S. C.
,
Klooster
,
J.
,
Furukawa
,
T.
,
Flannery
,
J. G.
, et al. 
(
2014
).
Targeted ablation of Crb2 in photoreceptor cells induces retinitis pigmentosa
.
Hum. Mol. Genet.
23
,
3384
-
3401
.
Bachmann
,
A.
,
Draga
,
M.
,
Grawe
,
F.
and
Knust
,
E.
(
2008
).
On the role of the MAGUK proteins encoded by Drosophila varicose during embryonic and postembryonic development
.
BMC Dev. Biol.
8
,
55
.
Berger
,
S.
,
Bulgakova
,
N. A.
,
Grawe
,
F.
,
Johnson
,
K.
and
Knust
,
E.
(
2007
).
Unraveling the genetic complexity of Drosophila stardust during photoreceptor morphogenesis and prevention of light-induced degeneration
.
Genetics
176
,
2189
-
2200
.
Brodsky
,
J. L.
(
2012
).
Cleaning up: ER-associated degradation to the rescue
.
Cell
151
,
1163
-
1167
.
Bujakowska
,
K.
,
Audo
,
I.
,
Mohand-Saïd
,
S.
,
Lancelot
,
M.-E.
,
Antonio
,
A.
,
Germain
,
A.
,
Léveillard
,
T.
,
Letexier
,
M.
,
Saraiva
,
J.-P.
,
Lonjou
,
C.
, et al. 
(
2012
).
CRB1 mutations in inherited retinal dystrophies
.
Hum. Mutat.
33
,
306
-
315
.
Bulgakova
,
N. A.
and
Knust
,
E.
(
2009
).
The Crumbs complex: from epithelial-cell polarity to retinal degeneration
.
J. Cell Sci.
122
,
2587
-
2596
.
Chartier
,
F. J.-M.
,
Hardy
,
E. J.-L.
and
Laprise
,
P.
(
2012
).
Crumbs limits oxidase-dependent signaling to maintain epithelial integrity and prevent photoreceptor cell death
.
J. Cell Biol.
198
,
991
-
998
.
den Hollander
,
A. I.
,
ten Brink
,
J. B.
,
de Kok
,
Y. J
,
van Soest
,
S.
,
van den Born
,
L. I.
,
van Driel
,
M. A.
,
van de Pol
,
D. J.
,
Payne
,
A. M.
,
Bhattacharya
,
S. S.
,
Kellner
,
U.
,
Hoyng
,
C. B.
, et al. 
(
1999
).
Mutations in a human homologue of Drosophila crumbs cause retinitis pigmentosa (RP12)
.
Nat. Genet.
23
,
217
-
221
.
den Hollander
,
A. I.
,
Davis
,
J.
,
van der Velde-Visser
,
S. D.
,
Zonneveld
,
M. N.
,
Pierrottet
,
C. O.
,
Koenekoop
,
R. K.
,
Kellner
,
U.
,
van den Born
,
L. I.
,
Heckenlively
,
J. R.
,
Hoyng
,
C. B.
, et al. 
(
2004
).
CRB1 mutation spectrum in inherited retinal dystrophies
.
Hum. Mutat.
24
,
355
-
369
.
den Hollander
,
A. I.
,
Roepman
,
R.
,
Koenekoop
,
R. K.
and
Cremers
,
F. P. M.
(
2008
).
Leber congenital amaurosis: Genes, proteins and disease mechanisms
.
Prog. Ret. Eye Res.
27
,
391
-
419
.
Ehrenberg
,
M.
,
Pierce
,
E. A.
,
Cox
,
G. F.
and
Fulton
,
A. B.
(
2013
).
CRB1: one gene, many phenotypes
.
Semin. Ophthalmol.
28
,
397
-
405
.
Fletcher
,
G. C.
,
Lucas
,
E. P.
,
Brain
,
R.
,
Tournier
,
A.
and
Thompson
,
B. J.
(
2012
).
Positive feedback and mutual antagonism combine to polarize crumbs in the Drosophila follicle cell epithelium
.
Curr. Biol.
22
,
1116
-
1122
.
Freeman
,
M.
(
1996
).
Reiterative use of the EGF receptor triggers differentiation of all cell types in the Drosophila eye
.
Cell
87
,
651
-
660
.
Griciuc
,
A.
,
Aron
,
L.
and
Ueffing
,
M.
(
2011
).
ER stress in retinal degeneration: a target for rational therapy?
Trends Mol. Med.
17
,
442
-
451
.
Groth
,
A. C.
,
Fish
,
M.
,
Nusse
,
R.
and
Calos
,
M. P.
(
2004
).
Construction of transgenic Drosophila by using the site-specific integrase from phage phiC31
.
Genetics
166
,
1775
-
1782
.
Gurudev
,
N.
,
Yuan
,
M.
and
Knust
,
E.
(
2014
).
chaoptin, prominin, eyes shut and crumbs form a genetic network controlling the apical compartment of Drosophila photoreceptor cells
.
Biol. Open
3
,
332
-
341
.
Hong
,
Y.
,
Ackerman
,
L.
,
Jan
,
L. Y.
and
Jan
,
Y.-N.
(
2003
).
Distinct roles of Bazooka and Stardust in the specification of Drosophila photoreceptor membrane architecture
.
Proc. Natl. Acad. Sci. USA
100
,
12712
-
12717
.
Hsu
,
Y.-C.
and
Jensen
,
A. M.
(
2010
).
Multiple domains in the Crumbs Homolog 2a (Crb2a) protein are required for regulating rod photoreceptor size
.
BMC Cell Biol.
11
,
60
.
Izaddoost
,
S.
,
Nam
,
S.-C.
,
Bhat
,
M. A.
,
Bellen
,
H. J.
and
Choi
,
K.-W.
(
2002
).
Drosophila Crumbs is a positional cue in photoreceptor adherens junctions and rhabdomeres
.
Nature
416
,
178
-
183
.
Johnson
,
K.
,
Grawe
,
F.
,
Grzeschik
,
N.
and
Knust
,
E.
(
2002
).
Drosophila crumbs is required to inhibit light-induced photoreceptor degeneration
.
Curr. Biol.
12
,
1675
-
1680
.
Kumar
,
J. P.
and
Ready
,
D. F.
(
1995
).
Rhodopsin plays an essential structural role in Drosophila photoreceptor development
.
Development
121
,
4359
-
4370
.
Laprise
,
P.
,
Lau
,
K. M.
,
Harris
,
K. P.
,
Silva-Gagliardi
,
N. F.
,
Paul
,
S. M.
,
Beronja
,
S.
,
Beitel
,
G. J.
,
McGlade
,
C. J.
and
Tepass
,
U.
(
2009
).
Yurt, Coracle, Neurexin IV and the Na+,K+-ATPase form a novel group of epithelial polarity proteins
.
Nature
459
,
1141
-
1145
.
Lee
,
T.
and
Luo
,
L.
(
1999
).
Mosaic Analysis with a Repressible Cell Marker for Studies of Gene Function in Neuronal Morphogenesis
.
Neuron
22
,
451
-
461
.
Lee
,
S.-J.
and
Montell
,
C.
(
2004
).
Light-dependent translocation of visual arrestin regulated by the NINAC myosin III
.
Neuron
43
,
95
-
103
.
Letunic
,
I.
,
Doerks
,
T.
and
Bork
,
P.
(
2015
).
SMART: recent updates, new developments and status in 2015
.
Nuc. Acids Res.
43
,
D257
-
D260
.
Li
,
B. X.
,
Satoh
,
A. K.
and
Ready
,
D. F.
(
2007
).
Myosin V, Rab11, and dRip11 direct apical secretion and cellular morphogenesis in developing Drosophila photoreceptors
.
J. Cell Biol.
177
,
659
-
669
.
Lin
,
Y.-H.
,
Currinn
,
H.
,
Pocha
,
S. M.
,
Rothnie
,
A.
,
Wassmer
,
T.
and
Knust
,
E.
(
2015
).
AP-2-complex-mediated endocytosis of Drosophila Crumbs regulates polarity by antagonizing Stardust
.
J. Cell Sci.
128
,
4538
-
4549
.
Lloyd
,
T. E.
,
Atkinson
,
R.
,
Wu
,
M. N.
,
Zhou
,
Y.
,
Pennetta
,
G.
and
Bellen
,
H. J.
(
2002
).
Hrs regulates endosome membrane invagination and tyrosine kinase receptor signaling in Drosophila
.
Cell
108
,
261
-
269
.
Longley
,
R. L.
and
Ready
,
D. F.
(
1995
).
Integrins and the development of three-dimensional structure in the Drosophila compound eye
.
Dev. Biol.
171
,
415
-
433
.
Mehalow
,
A. K.
,
Kameya
,
S.
,
Smith
,
R. S.
,
Hawes
,
N. L.
,
Denegre
,
J. M.
,
Young
,
J. A.
,
Bechtold
,
L.
,
Haider
,
N. B.
,
Tepass
,
U.
,
Heckenlively
,
J. R.
, et al. 
(
2003
).
CRB1 is essential for external limiting membrane integrity and photoreceptor morphogenesis in the mammalian retina
.
Hum. Mol. Gen.
12
,
2179
-
2189
.
Muschalik
,
N.
and
Knust
,
E.
(
2011
).
Increased levels of the cytoplasmic domain of Crumbs repolarise developing Drosophila photoreceptors
.
J. Cell Sci.
124
,
3715
-
3725
.
Nemetschke
,
L.
and
Knust
,
E.
(
2016
).
Drosophila Crumbs prevents ectopic Notch activation in developing wings by inhibiting ligand-independent endocytosis
.
Development
143
,
4543
-
4553
.
Nguyen
,
M. B.
,
Vuong
,
L. T.
and
Choi
,
K.-W.
(
2016
).
Ebi modulates wing growth by ubiquitin-dependent downregulation of Crumbs in Drosophila
.
Development
143
,
3506
-
3513
.
Omori
,
Y.
and
Malicki
,
J.
(
2006
).
oko meduzy and related crumbs genes are determinants of apical cell features in the vertebrate embryo
.
Curr. Biol.
16
,
945
-
957
.
Pellikka
,
M.
,
Tanentzapf
,
G.
,
Pinto
,
M.
,
Smith
,
C.
,
McGlade
,
C. J.
,
Ready
,
D. F.
and
Tepass
,
U.
(
2002
).
Crumbs, the Drosophila homologue of human CRB1/RP12, is essential for photoreceptor morphogenesis
.
Nature
416
,
143
-
149
.
Pellissier
,
L. P.
,
Alves
,
C. H.
,
Quinn
,
P. M.
,
Vos
,
R. M.
,
Tanimoto
,
N.
,
Lundvig
,
D. M. S.
,
Dudok
,
J. J.
,
Hooibrink
,
B.
,
Richard
,
F.
,
Beck
,
S. C.
, et al. 
(
2013
).
Targeted ablation of Crb1 and Crb2 in retinal progenitor cells mimics leber congenital amaurosis
.
PLoS Genet.
9
,
e1003976
.
Pocha
,
S. M.
,
Shevchenko
,
A.
and
Knust
,
E.
(
2011
).
Crumbs regulates rhodopsin transport by interacting with and stabilizing myosin V
.
J. Cell Biol.
195
,
827
-
838
.
Richard
,
M.
,
Grawe
,
F.
and
Knust
,
E.
(
2006
).
DPATJ plays a role in retinal morphogenesis and protects against light-dependent degeneration of photoreceptor cells in the Drosophila eye
.
Dev. Dyn.
235
,
895
-
907
.
Richard
,
M.
,
Muschalik
,
N.
,
Grawe
,
F.
,
Özüyaman
,
S.
and
Knust
,
E.
(
2009
).
A role for the extracellular domain of Crumbs in morphogenesis of Drosophila photoreceptor cells
.
Eur. J. Cell Biol.
88
,
765
-
777
.
Sahni
,
N.
,
Yi
,
S.
,
Taipale
,
M.
,
Bass
,
J. I. F.
,
Coulombe-Huntington
,
J.
,
Yang
,
F.
,
Peng
,
J.
,
Weile
,
J.
,
Karras
,
G. I.
,
Wang
,
Y.
, et al. 
(
2015
).
Widespread macromolecular interaction perturbations in human genetic disorders
.
Cell
161
,
647
-
660
.
Satoh
,
A. K.
and
Ready
,
D. F.
(
2005
).
Arrestin1 mediates light-dependent rhodopsin endocytosis and cell survival
.
Curr. Biol.
15
,
1722
-
1733
.
Stenson
,
P. D.
,
Mort
,
M.
,
Ball
,
E. V.
,
Shaw
,
K.
,
Phillips
,
A. D.
and
Cooper
,
D. N.
(
2014
).
The Human Gene Mutation Database: building a comprehensive mutation repository for clinical and molecular genetics, diagnostic testing and personalized genomic medicine
.
Hum. Genet.
133
,
1
-
9
.
Tabuchi
,
K.
,
Sawamoto
,
K.
,
Suzuki
,
E.
,
Ozaki
,
K.
,
Sone
,
M.
,
Hama
,
C.
,
Tanifuji-Morimoto
,
T.
,
Yuasa
,
Y.
,
Yoshihara
,
Y.
,
Nose
,
A.
, et al. 
(
2000
).
GAL4/UAS-WGA system as a powerful tool for tracing Drosophila transsynaptic neural pathways
.
J. Neurosci. Res.
59
,
94
-
99
.
Tanentzapf
,
G.
and
Tepass
,
U.
(
2003
).
Interactions between the crumbs, lethal giant larvae and bazooka pathways in epithelial polarization
.
Nat. Cell Biol.
5
,
46
-
52
.
Tepass
,
U.
(
2012
).
The apical polarity protein network in Drosophila epithelial cells: regulation of polarity, junctions, morphogenesis, cell growth, and survival
.
Annu. Rev. Cell Dev. Biol.
28
,
655
-
685
.
Tepass
,
U.
and
Harris
,
K. P.
(
2007
).
Adherens junctions in Drosophila retinal morphogenesis
.
Trends Cell Biol.
17
,
26
-
35
.
Tepass
,
U.
,
Theres
,
C.
and
Knust
,
E.
(
1990
).
crumbs encodes an EGF-like protein expressed on apical membranes of Drosophila epithelial cells and required for organization of epithelia
.
Cell
61
,
787
-
799
.
Thompson
,
B. J.
,
Pichaud
,
F.
and
Röper
,
K.
(
2013
).
Sticking together the Crumbs — an unexpected function for an old friend
.
Nat. Rev. Mol. Cell Biol.
14
,
307
-
314
.
van de Pavert
,
S. A.
,
Kantardzhieva
,
A.
,
Malysheva
,
A.
,
Meuleman
,
J.
,
Versteeg
,
I.
,
Levelt
,
C.
,
Klooster
,
J.
,
Geiger
,
S.
,
Seeliger
,
M. W.
,
Rashbass
,
P.
, et al. 
(
2004
).
Crumbs homologue 1 is required for maintenance of photoreceptor cell polarization and adhesion during light exposure
.
J. Cell Sci.
117
,
4169
-
4177
.
van de Pavert
,
S. A.
,
Meuleman
,
J.
,
Malysheva
,
A.
,
Aartsen
,
W. M.
,
Versteeg
,
I.
,
Tonagel
,
F.
,
Kamphuis
,
W.
,
McCabe
,
C. J.
,
Seeliger
,
M. W.
and
Wijnholds
,
J.
(
2007
).
A single amino acid substitution (Cys249Trp) in Crb1 causes retinal degeneration and deregulates expression of pituitary tumor transforming gene Pttg1
.
J. Neurosci.
27
,
564
-
573
.
van Rossum
,
A. G. S. H.
,
Aartsen
,
W. M.
,
Meuleman
,
J.
,
Klooster
,
J.
,
Malysheva
,
A.
,
Versteeg
,
I.
,
Arsanto
,
J.-P.
,
Le Bivic
,
A.
and
Wijnholds
,
J.
(
2006
).
Pals1/Mpp5 is required for correct localization of Crb1 at the subapical region in polarized Muller glia cells
.
Hum. Mol. Genet.
15
,
2659
-
2672
.
Wang
,
T.
and
Montell
,
C.
(
2007
).
Phototransduction and retinal degeneration in Drosophila
.
Pflugers Arch.
454
,
821
-
847
.
Wang
,
S. A.
,
Tan
,
K. L.
,
Agosto
,
M. A.
,
Xiong
,
B.
,
Yamamoto
,
S.
,
Sandoval
,
H.
,
Jaiswal
,
M.
,
Bayat
,
V.
,
Zhang
,
K.
,
Charng
,
W. L.
, et al. 
(
2014
).
The retromer complex is required for rhodopsin recycling and its loss leads to photoreceptor degeneration
.
PLoS Biol.
13
,
2
.
Wodarz
,
A.
,
Grawe
,
F.
and
Knust
,
E.
(
1993
).
CRUMBS is involved in the control of apical protein targeting during Drosophila epithelial development
.
Mech. Dev.
44
,
175
-
187
.
Wodarz
,
A.
,
Hinz
,
U.
,
Engelbert
,
M.
and
Knust
,
E.
(
1995
).
Expression of crumbs confers apical character on plasma membrane domains of ectodermal epithelia of Drosophila
.
Cell
82
,
67
-
76
.
Xiong
,
B.
and
Bellen
,
H. J.
(
2013
).
Rhodopsin homeostasis and retinal degeneration: lessons from the fly
.
Trends Neurosci.
36
,
652
-
660
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information