Several retinal dystrophies, including retinitis pigmentosa type 12 and Leber congenital amaurosis, are caused by a large variety of mutations in the CRB1 (Crumbs homologue 1) gene. This discovery led to an increased focus on the function of CRB1 and the Drosophila homologue Crumbs. In the present study, we review the current knowledge on Crumbs and its vertebrate homologues, their function in cell polarity and their pathogenicity in retinal degeneration.
Mutations in CRB1 (Crumbs homologue 1) cause different retinal dystrophies
Linkage analysis in a large family from a genetically isolated population in the Netherlands initially mapped autosomal-recessive RP (retinitis pigmentosa) with PPRPE (preserved para-arteriolar retinal pigment epithelium) RP12 (RP type 12) to chromosome 1q31-q32.1 . A novel, retina-specific cDNA clone mapping to this locus showed several causative mutations in a collection of RP12, early onset RP, RP with coats-like exudative vasculopathy and Leber congenital amaurosis patients [2,3]. This gene was denoted CRB1, due to its protein homology with the Drosophila melanogaster protein Crumbs (Crb). The different phenotypes caused by this variety of mutations in CRB1 might be partially explained by their different effects on the protein, since the frequency of null mutations seems to be higher in Leber congenital amaurosis, the most severe retinal dystrophy. However, statistically significant genotype–phenotype correlations could not be established, suggesting additional modifier genetic factors or environmental influences .
Homology between Drosophila Crumbs and human CRB1
The human CRB1 protein is a single transmembrane protein with a large extracellular part, containing three laminin G domains and 19 epidermal growth-factor-like domains. The 37-amino-acid intracellular part is very conserved and contains a C-terminal PDZ-binding motif (ERLI). Although the Drosophila Crumbs protein contains a larger extracellular part with four laminin G domains and 30 epidermal growth-factor-like domains, the overall architecture is conserved.
The Crumbs/CRB1 function has most extensively been studied in Drosophila. In Drosophila embryonic epithelia, Crb localizes to a region just apical of the ZA (zonula adherens), called the SAR (subapical region). Crb is also expressed in the PRC (photoreceptor cell), where the protein specifically localizes in the stalk membrane, apical to the ZA . Analysis of mutant flies revealed that Crb is required for the establishment and maintenance of the ZA and, in PRC, Crb also regulates the stalk membrane development .
Recently, large parts of the protein complexes at the ZA and SAR have been identified (reviewed in [5,6]). Their localization and function seems to be very well conserved (Figure 1). The first protein complex to form at the SAR of Drosophila epithelia is the Baz–Par6–aPKC complex. This complex is then antagonized by the Scrib–Lgl–Dlg complex, located basal to the ZA. Finally, the Crb–Sdt–Patj complex is recruited to the SAR . The mammalian homologues of these complexes are Par3–Par6–aPKC, Scrib–Lgl–Dlg1 and Crb1–Mpp5 (Pals1)–Patj (Figure 2) respectively. A direct interaction has been observed between Par6 and Mpp5, physically linking these two complexes at the SAR . In mouse PRC, Crb1 was also shown to form a complex with the membrane-associated guanylate kinase proteins Mpp5 and Mpp4 and the multiple PDZ protein Mupp1 . In addition, all three Crb family members have now been shown to co-localize at the SAR of mouse PRC  and Crb3 was also shown to interact with Par6 , MPP5  and the multiple PDZ protein Patj , albeit in epithelial cells. In summary, the transmembrane Crumbs/CRB1 protein is part of an evolutionary conserved scaffolding protein complex at the SAR of polarized cells.
Schematic representation of vertebrate epithelia and photoreceptor polarity
Crb1 in the vertebrate retina
Drosophila Crumbs prevents light-induced photoreceptor degeneration
The influence of light on the Crb pathway was recently examined in several crb mutant flies . When flies are raised under low light conditions, crb null mosaic eyes show a mild phenotype consisting of thicker and shorter rhabdomeres and a reduction in length of the stalk membrane . However, crb null mosaic flies raised under constant light show a progressive and massive retinal degeneration . This degeneration was demonstrated to act through programmed cell death, due to the accumulation of internalized m-rhodopsin–arrestin complexes. Furthermore, expression of the membrane-bound cytoplasmic Crb domain in the crb null mosaic retinas could not inhibit light-induced photoreceptor degeneration, whereas flies carrying a deletion of the cytoplasmic, C-terminal 23 amino acids of Crb did not show any light-induced photoreceptor degeneration. This indicates that, in contrast with the intracellular domain required for the correct formation of adherens junctions and proper morphogenesis, the extracellular domain of Crb is crucial to suppress light-induced retinal degeneration .
Is Crb1 also functionally homologous to Crumbs?
The rd8 (retinal degeneration 8) mouse is a spontaneous mutant with a single base deletion in Crb1 (c.3481del) causing a frameshift with early stop (p.Arg1161GlyfsX48). The predicted protein is hypothesized to be secreted, lacking the transmembrane and cytoplasmic domains. The homozygous Crbrd8 retinas show irregular white spots in the inferior nasal quadrant from the age of 3 weeks. Morphological analysis shows disturbances in the ZA in the outer limiting membrane, disorganized inner and outer segments and a 25% shortening of the inner segments. The white retinal spotting and retinal dysplasia in homozygous Crbrd8 mice are highly variable when crossed on to different backgrounds, suggesting modifier genes .
The second model is a Crb1 knockout mouse, generated by the deletion of the Crb1 promoter and first exon. The Crb1−/− mice develop foci of retinal degeneration, where the outer limiting membrane integrity is compromised. Initially, PRC bodies protrude out of, or ingress down from, the outer nuclear layer, leading to double PRC layers or half rosettes and ultimately to cell death. Exposure to light significantly aggravates this phenotype, increasing the amount and size of degeneration foci so that they become visible on fundus photography as white retinal spots. This led to the hypothesis that Crb1 is essential for the maintenance but not the establishment of the PRC adherens junctions, especially during light exposure. Correct localization of Crb2 and Crb3 to the SAR does not seem to be sufficient to fulfil this function .
Crumbs of the future
The mammalian Crb1 and the Drosophila Crumbs are conserved to function in the maintenance of the ZA, as well as the protection against light-induced retinal degeneration. However, it is unclear whether the mouse Crb1 also has a role in the formation of the ZA and the inner segment morphogenesis. Furthermore, it is striking that all three Crb proteins co-localize at the SAR of mammalian PRC. Since their cytoplasmic parts are very homologous (Figure 2), it is not too surprising that Crb3 could also interact with Patj, Par6 and Pals1/Mpp5. This leads to the hypothesis that Crb1 and Crb3 (and possibly even Crb2) are interchangeable. Given the divergence of their extracellular domains, this interchangeability would have a major influence on the extracellular interactions and could therefore influence, for example, adhesion or signalling. It is therefore crucial to characterize further the protein complex in which the different Crb proteins are involved, especially in determining their extracellular interaction partners. This might help to unravel the mechanism by which Crb1 inhibits light-induced photoreceptor degeneration and could be instrumental in the development of a therapy for the retinal degenerations caused by mutations in CRB1.
Signalling Outwards and Inwards: A Focus Topic at BioScience2004, held at SECC Glasgow, U.K., 18–22 July 2004. Edited by J. Challiss (Leicester, U.K.), A. Harwood (University College London, U.K.), M. Humphries (Manchester, U.K.), C. Isacke (Institute of Cancer Research, London, U.K.), R. Liddington (Burnham Institute, La Jolla, CA, U.S.A.), T. Palmer (Glasgow, U.K.), K. Siddle (Cambridge, U.K.), C. Sutherland (Dundee, U.K.), H. Wallace (Aberdeen, U.K.) and M. Welham (Bath, U.K.).
Owing to space limitations, we are restricted to citing reviews and we therefore apologize to all original authors. J.W. is supported by the EC grant QLG3-CT-2002-01266.