Mutations in NR2E3 can cause dominant or recessive retinal degenerations in the same family

Abstract

NR2E3, a photoreceptor-specific nuclear receptor (PNR), represses cone-specific genes and activates several rod-specific genes. In humans, mutations in NR2E3 have been associated with the recessively-inherited enhanced short-wavelength sensitive S-cone syndrome (ESCS) and, recently, with autosomal dominant (ad) retinitis pigmentosa (RP) (adRP). In the present work, we describe two additional families affected by adRP that carry a heterozygous c.166G>A (p.G56R) mutation in the NR2E3 gene. Functional analysis determined the dominant negative activity of the p.G56R mutant protein as the molecular mechanism of adRP. Interestingly, in one pedigree, the most common causal variant for ESCS (p.R311Q) cosegregated with the adRP-linked p.G56R mutation, and the compound heterozygotes exhibited an ESCS-like phenotype, which in 1 of the 2 cases was strikingly "milder" than the patients carrying the p.G56R mutation alone. Impaired repression of cone-specific genes by the corepressors atrophin-1 (dentatorubral-pallidoluysian atrophy [DRPLA] gene product) and atrophin-2 (arginine-glutamic acid dipeptide repeat [RERE] protein) appeared to be a molecular mechanism mediating the beneficial effect of the p.R311Q mutation. Finally, the functional dominance of the p.R311Q variant to the p.G56R mutation is discussed.

Figure 1

Pedigrees of affected families of Swiss (A) and Jewish-American (B) origin. All patients with black symbols harbor the G56R heterozygous mutation and are affected with adRP, while the two patients with dark gray symbols are compound heterozygous G56R/R311Q and are affected with ESCS syndrome. Individuals with the heterozygous R311Q mutation (light gray symbols) have no clinical phenotype. Arrow denotes the proband of the American family. C) Lod scores for microsatellite markers obtained in the Swiss family.

Figure 2

Clinical data of the American family. A-F) Fundus photographs. The proband II-1 shows a few small spots of clumped pigmentation in her peripheral retina, mild atrophic changes of the retinal epithelium in the macula, but no attenuation of retinal vessels (A, B). Her sister II-3 shows attenuated retinal vessels and clumped intraretinal pigmentation in her mid-peripheral retina (C, D). Patients IV-1 (E) and IV-2 (F) show a retinal pigmentary clump (arrow) in the peripheral retina. G) ERGs of patients in the American family. The subjects were studied in the dark-adapted state using Wratten filters providing red (Wratten 29), yellow (Wratten 21), green (Wratten 61) and blue (Wratten 98) light (see Methods). The proband, her son and her sister were analyzed in 1980, when the two most severely affected individuals had detectable responses. The proband, who is the oldest of all affected, has much better responses than her sister II-3 and her son III-2. Patient II-3 response to blue light has greater amplitude than that of III-2 (arrows). The converse is true for red light (arrows). The calibration, lower right, indicates 7 microvolts vertically and 20 milliseconds horizontally.

Figure 3

NR2E3 protein structure. A) The NR2E3 DNA-binding domain (DBD) folds into two ‘zinc finger’-like structures. The α-helical DNA binding motif (HZnf1 and HZnf2) of each zinc finger lies perpendicular to the other. The p.G56R mutation is located at the basis of the α-helix of the first zinc finger. B) The NR2E3 ligand-binding domain (LBD) presents a canonical α-helical structure forming a hydrophobic ligand-binding pocket. R311 is located at the end of helix 7 (H7). Modelling of the mutant protein suggests a slight destabilization of the α-helical turn immediately C-terminal of R311 (orange dots). Ligand-binding involves residues of H7, but is mainly mediated by H3, H6 and HAF (AF-2: Activator Function-2). The dimerization interface is mainly located in H9 and H10. Modeling was performed with SWISSMODEL [Schwede et al., 2003]. C) Human (hs) NR2E1 and NR2E3 amino acid sequences were aligned with ClustalW 1.83 algorithm [Chenna et al., 2003]. A/B: N-terminal A/B domain; C: DBD; D: hinge region; E/F: LBD. Identical amino acids are denoted by arrows, conserved amino acids by dots. In the DBD, the two zinc fingers are indicated by dotted lines. In the LBD, α-helices (H) and β-sheets (s) are indicated by black boxes. A bold line underlines the C-terminal part of the NR2E1 ligand-binding domain sufficient for NR2E1/atrophin-1 interaction [Wang et al., 2006; Zhang et al., 2006]. The sites of p.G56R and p.R311Q mutations are indicated by red letters in the NR2E3 sequence.

Figure 4

The adRP-linked mutation p.G56R results in a dominant negative NR2E3 mutant protein. A) The p.G56R mutant protein is defective in DNA-binding. Binding of NR2E3 wild-type and mutant proteins to a radiolabeled response element was tested in EMSA by nonspecific (n. sp.) and specific (spec.) competition (comp.). Fold excess of cold probe is indicated. A specific anti-NR2E3 antibody (Ab) recognizes the NR2E3 dimer, inducing the formation of a slower migrating complex (‘supershift’ Ab/NR2E3). Contrast for lanes 1 to 13 was increased to show absence of binding of the p.G56R mutant protein and presence of the ‘supershift’ in lane 8. B) Dominant negative activity of p.G56R towards NR2E3 wild-type and p.R311Q proteins. Amounts of reticulocyte lysates are indicated in μl. C) Western Blot with anti-NR2E3 antisera on 10 μl of in vitro translated proteins. A major band at the expected size of 48 kDa for His-tagged NR2E3 is detected. The higher translation efficiency of pcDNA3.1-NR2E3R311Q contributes to a stronger signal detected by EMSA.

Figure 5

Impaired regulation of opsin genes by the p.G56R mutant protein. 293T cells were transiently transfected with rhodopsin, S-opsin and M-opsin luciferase reporter constructs. CRX and NRL expression vectors (30 ng each) were present in all wells. Activity of NR2E3 wild-type (WT) or p.R311Q mutant protein (30 ng each) was tested in presence of increasing amounts of the p.G56R mutant protein (+:15 ng; ++:30 ng; +++:45 ng). Three independent experiments were performed in duplicates in 12-well plates using the luciferase assay system (Promega). Luciferase activity was normalized to β-galactosidase activity. Normalized luciferase activities of cells transfected with CRX and NRL alone were set to 1. Error bars represent S.E.M.

Figure 6

Atrophins are corepressors of NR2E3. a) Atrophins enhanced NR2E3-mediated repression of the M-opsin promoter in presence of NR2E3 wild-type, but not NR2E3 p.R311Q protein. 293T cells were transiently transfected with the M-opsin reporter construct, in presence of NRL, CRX, and, respectively, NR2E3 wild-type or p.R311Q (30 ng each), along with increasing amounts of atrophin-1, atrophin-2 and NCoR (33 ng, 66 ng or 100 ng). At least three independent experiments were performed in duplicates in 12-well plates using the luciferase assay system (Promega). Luciferase activity was normalized to β-galactosidase activity. Normalized luciferase activities of cells transfected without corepressors were set to 1. Error bars represent S.E.M; *: p<0.05; **: p<0.01. B) In situ hybridization for atrophin-1 mRNAs on retinal sections of 2-months-old Bl6/C57. RPE: retinal pigment epithelium; POS: photoreceptor outer segments; ONL: outer nuclear layer; OPL: outer plexiform layer; INL: inner nuclear layer; IPL: inner plexiform layer; GCL: ganglion cell layer. Note the absence of signal in presence of the control sense probe. Scale bar: 50 μm. C) Atrophin-1 associates with the M-opsin promoter in vivo. Chromatin-immunoprecipitation for the M-opsin promoter was performed with anti-atrophin-1 antisera (α-atn-1) on retinas of 2-months-old C57/BL6 mice. As controls, immunoprecipitation without antisera (no Ab) or with the buffer alone (mock) were used. PCR amplifications on total input chromatin (Total input) are shown in the lower panel.