A common allele in RPGRIP1L is a modifier of retinal degeneration in ciliopathies

Abstract

Despite rapid advances in the identification of genes involved in disease, the predictive power of the genotype remains limited, in part owing to poorly understood effects of second-site modifiers. Here we demonstrate that a polymorphic coding variant of RPGRIP1L (retinitis pigmentosa GTPase regulator-interacting protein-1 like), a ciliary gene mutated in Meckel-Gruber (MKS) and Joubert (JBTS) syndromes, is associated with the development of retinal degeneration in individuals with ciliopathies caused by mutations in other genes. As part of our resequencing efforts of the ciliary proteome, we identified several putative loss-of-function RPGRIP1L mutations, including one common variant, A229T. Multiple genetic lines of evidence showed this allele to be associated with photoreceptor loss in ciliopathies. Moreover, we show that RPGRIP1L interacts biochemically with RPGR, loss of which causes retinal degeneration, and that the Thr229-encoded protein significantly compromises this interaction. Our data represent an example of modification of a discrete phenotype of syndromic disease and highlight the importance of a multifaceted approach for the discovery of modifier alleles of intermediate frequency and effect.

Figure 1. Functional assessment of the RPGRIP1L A229T variant in vivo

a. Zebrafish embryos injected with rpgrip1l splice-blocking morpholino (MO) display gastrulation defects at the eight-nine somite stage; lateral and dorsal views of live embryos are shown. Embryos were categorized phenotypically based on the presence of a shortened body axis (anterior and posterior ends indicated with blue triangles) and small head/eyes only (magenta lines indicate eye size) (Class I; mild); or shortened body axis and small head/eyes in addition to at least one of the following: broad and thin somites (white line indicates somite span), notochord kinking, and tail extension defects (black asterisk) (Class II; severe). b. Live scoring of embryos co-injected with MO and human RPGRIP1L mRNA indicates that A229T is pathogenic in our assay as indicated by the comparison of WT rescue efficiency to mutant rescue efficiency. MO, morpholino alone; WT, wild-type human RPGRIP1L mRNA. Known JBTS mutations T615P and A695P are positive controls; neutral variant G1025S (rs2111119) is a negative control. Statistical significance (mutant rescue vs. WT rescue) is depicted as (***), p<0.0001 (χ²); see Suppl Table 2 for all χ² and p-values. c. Representative examples of whole-mount embryos hybridized in situ with pax2, krox20, and myoD riboprobes. d. Two-dimensional morphometric quantification of the efficiency of human RPGRIP1L mRNA to rescue rpgrip1l MO phenotypes. Whole embryos with eight-nine appreciable somites (n=8-20/injection) labeled with pax2, krox20, and myoD riboprobes were flat-mounted, photographed and measured in two dimensions. The width/length ratio was calculated as the width (w) spanning between the distal ends of the 5^th anterior somites (horizontal arrow in panel c) vs. the length (l) of the notochord as indicated by the staining of adaxial cells (vertical arrow in panel c). Statistical significance (mutant vs. WT rescue) is depicted as (*), p<0.05 and (***), p<0.0001 (two-tailed student's t-test); see Suppl Table 3 for all p-values. e. rpgrip1l morphant larvae display tail development defects at 5 days post-fertilization (dpf). Morphant larva (MO) with tail curvature and posterior structural disorganization can be rescued efficiently with WT human RPGRIP1L mRNA, however RPGRIP1L message harboring the 229T encoding allele does not completely rescue the tail phenotype. f. Quantification of the efficiency of tail phenotype rescue in 5dpf larvae. Whereas WT RPGRIP1L message completely rescued the tail phenotypes, A229T as well as positive controls T615P and A695P result in the persistence of affected tail structures in injected larva. G1025S was used as a negative control. Statistical significance (mutant rescue vs. WT rescue) is depicted as (*), p<0.05 (χ²).

Figure 2. RPGRIP1L interacts with RPGR and missense variant A229T abrogates this interaction

a. Known-bait and known-prey analysis. Yeast two-hybrid analysis using indicated RPGR bait-encoding constructs and RPGRIP1L [carboxyl-terminal (C) or full-length (FL)] prey encoding constructs revealed interaction of RPGR exons 1-11 with carboxyl-terminal or full-length RPGRIP1L. Laminin was used as negative control, whereas p53/T-antigen (T-Ag) interaction served as a positive control. b. RPGR-RPGRIP1L interaction was validated further by examining the activation of the LacZ gene in the β-galactosidase assay corresponding to (a). c. Interaction of RPGR and RPGRIP1L in vivo. Bovine retinal extracts were subjected to immunoprecipitation (IP) using anti-RPGR, RPGRIP1, RPGRIP1L antibodies, normal immunoglobulin (IgG), or pre-immune serum (Pre-Imm). Precipitated proteins were analyzed by SDS-PAGE and immunoblotting (IB) using anti-RPGRIP1L (left) or RPGR-ORF15^CP antibody (right). Arrow (left) indicates RPGRIP1L-imunoreactive band and the straight line in the right panel depicts the RPGR isoforms pulled down by RPGRIP1L. Asterisk (*) marks a non-specific immunoreactive band. d. RPGR-RPGRIP1L interaction in transfected cells. COS-1 cells were transiently transfected with constructs encoding Xpress-tagged RPGR, GFP-RPGRIP1L or Xpress-tag alone. Proteins were extracted and subjected to immunoprecipitation (IP) using the anti-GFP antibody. Precipitated proteins were analyzed by SDS-PAGE and immunoblotting using anti-Xpress antibody. Input lane contains 20% of the protein extract used for IP. Arrow indicates the Xpress-immunoreactive band. e. Interaction of RPGR with wild-type (WT) and mutant RPGRIP1L. COS-1 cells were transiently transfected with constructs encoding WT or mutant RPGRIP1L fused to GFP and Xpress-tagged RPGR. Panels ‘a’ and ‘b’ indicate the 20% of the GFP-RPGRIP1L and Xpress-RPGR input protein extract used for IP. Precipitation of GFP-RPGRIP1L using the GFP antibody was used as a positive control (panel ‘c’). Association of wt and mutant GFP-RPGRIP1L with Xpress-RPGR is represented in Panel ‘d’. T677I represents a bona fide pathogenic allele, but which likely does not contribute to retinal degeneration. f. Quantitative analysis of the band intensities are represented as percent of mutant RPGRIP1L immunoprecipitated by RPGR. Association of wt GFP-RPGRIP1L with RPGR was considered as 100%. The results represent an average of three independent experiments.