Pseudouridylation defect due to DKC1 and NOP10 mutations causes nephrotic syndrome with cataracts, hearing impairment, and enterocolitis

Abstract

RNA modifications play a fundamental role in cellular function. Pseudouridylation, the most abundant RNA modification, is catalyzed by the H/ACA small ribonucleoprotein (snoRNP) complex that shares four core proteins, dyskerin (DKC1), NOP10, NHP2, and GAR1. Mutations in DKC1, NOP10, or NHP2 cause dyskeratosis congenita (DC), a disorder characterized by telomere attrition. Here, we report a phenotype comprising nephrotic syndrome, cataracts, sensorineural deafness, enterocolitis, and early lethality in two pedigrees: males with DKC1 p.Glu206Lys and two children with homozygous NOP10 p.Thr16Met. Females with heterozygous DKC1 p.Glu206Lys developed cataracts and sensorineural deafness, but nephrotic syndrome in only one case of skewed X-inactivation. We found telomere attrition in both pedigrees, but no mucocutaneous abnormalities suggestive of DC. Both mutations fall at the dyskerin-NOP10 binding interface in a region distinct from those implicated in DC, impair the dyskerin-NOP10 interaction, and disrupt the catalytic pseudouridylation site. Accordingly, we found reduced pseudouridine levels in the ribosomal RNA (rRNA) of the patients. Zebrafish dkc1 mutants recapitulate the human phenotype and show reduced 18S pseudouridylation, ribosomal dysregulation, and a cell-cycle defect in the absence of telomere attrition. We therefore propose that this human disorder is the consequence of defective snoRNP pseudouridylation and ribosomal dysfunction.

Keywords: H/ACA snoRNP; pediatrics; pseudouridylation; rRNA; telomere.

Fig. 1.

Phenotype and genetic identification of the two affected families. (A–C) Affected males (n = 6) (A) and females (n = 9) (B and C) in FamA had no dysmorphic features apart from maxillary and mandibular hypoplasia in adult females. (A, M–R, and V–X) Affected males in FamA (A, V) and the two affected females (n = 2) from the consanguineous FamB (X) developed nephrotic syndrome with focal segmental glomerulosclerosis (M, O), diffuse podocyte foot process effacement (N), and enterocolitis with extensive chronic nonspecific inflammation (P–R). (I and L) In FamB, patient V:2, developed progressive hypomyelination (I) and cerebellar hypoplasia (L); both IV:3 and V:2 passed away in infancy. (D–H, J, K, and S–U) Cataracts (D and E) and sensorineural hearing impairment with a modiolus (S, black arrow; and U), cochlea (T, white arrow; and U) and cochlear nerve (S and T, white arrowheads) of normal morphology were present, even in the affected females of FamA, who had a normal lifespan and developed pigmentary retinopathy (G and H), microphthalmia (F), and pineal hypoplasia (J and K; SI Appendix, Fig. S2). (V and W) We found linkage to a 5.1-Mb region at Xq28 in the descendants of II:3 (W), assuming germline mosaicism in I:2 (V) with a LOD score of 3.01. (V) Through targeted sequencing in an unaffected and an affected but haploidentical male (II:9 and IV:14), we found a single de novo mutation in DKC1 (c.616 G>A, p.Glu206Lys), which segregated among the six haploidentical members in the generation II with disease occurrence. (X and Y) Homozygosity mapping in FamB (X) showed a haploidentical homozygous region of 2.6 Mb at 15q14 with an LOD score of 3.03 (Y). Exome sequencing of patient V:2 and targeted segregation analysis of IV:3 and V:2, parents, and siblings revealed a homozygous missense mutation in NOP10 (c.47C>T, p.Thr16Met). Scale bar in N: 3.5 μm.

Fig. 2.

Absent or subtle symptoms of DC despite telomere shortening. (A and B) We found telomere shortening in the affected individuals by Southern blot (heterozygous individuals in FamA and affected individuals in FamB are in bold) (A), showing a significant difference between the severely affected children (FamA, IV:4 and IV:14; FamB IV:3 and V:2) and six age-matched controls (P = 0.038) (B). (C–F) Telomere attrition was also shown by MM-qPCR as telomere copy number/standard copy number (T/S) ratio (C) and flow-FISH (D–F). (G) The severely affected index female (FamA, IV:4) showed a highly skewed X-inactivation in the fibroblasts and skin, with the ratio of her PBMCs expressing the mutant DKC1 allele decreasing with age. FB, fibroblast. Blue, E206K; red, WT DKC1 mRNA. (H) Despite telomere shortening, no nail dystrophy or leukoplakia was observed, with one heterozygous female from FamA (III:3) and one female from FamB (V:2) being diagnosed with mild dyskeratosis after genetic diagnosis. ys, years.

Fig. 3.

Dyskerin p.Glu206Lys and NOP10 p.Thr16Met mutations alter the pseudouridylation pocket of the H/ACA snoRNP complex. (A) Homology modeling of human H/ACA snoRNP dyskerin (amino acids 60 to 380), gray; NOP10, green; GAR1, red; NHP2, yellow; guide snoRNA, dark blue; and substrate RNA, light blue. The Glu206 (dyskerin) and Thr16 (NOP10) are in CPK format (carbon atoms are shown in the color of the backbone, oxygen in red, and nitrogen in blue). The C-alpha atoms of residues associated with DC and HH syndrome are shown in spheres. (B) Coimmunoprecipitation of native dyskerin from HEK293 cells transfected with green fluorescent protein (GFP)-tagged WT and T16M NOP10 and reciprocal immunoprecipitation of native NOP10 from HEK293 cells transfected with Flag-tagged WT and E206K dyskerin. Immunoblots show that both mutant proteins immunoprecipitate with their native counterpart (see also SI Appendix, Fig. S4). CNTL, control. (C and D) Pressure-tuning fluorescence spectra of WT and mutant dyskerin–NOP10 complexes, where mutant complexes show an altered stability. Arb. u., arbitrary units; barg, bar in gauge. (E) Structural heterogeneity indicates significant structural difference between mutants and WT (Welch test: DKC1, P = 1.8 × 10⁻²; NOP10, P = 4.04 × 10⁻¹¹). (F) Disassociation constants of WT and mutant complexes show both mutations cause a parallel change in pKd (Left; Welch test: DKC1, P = 9.54 × 10⁻³; NOP10, P = 1.55 × 10⁻⁷) and the interaction surface (Right). (G and H) Conformational changes induced by the mutations (G) WT dyskerin Glu206 interacts with the NOP10 Thr16–Leu17–Lys18 segment forming H-bonds (Left), these H-bonds are disrupted by both the dyskerin E206K (Center) and NOP10 T16M (Right) mutations. (H) The WT interaction (Left) between the substrate uridine and the catalytic D125 of dyskerin is uncoupled by both dyskerin E206K (Center) and NOP10 T16M (Right). 1 °, primary antibody; CNTL, cells transfected with empty expression plasmid, expressing GFP-Flag, size of 28 kDa; HEK293, nontransfected cells; IP, immunoprecipitated protein; S, supernatant. Statistical significance is denoted by asterisks. *P < 0.05; **P < 0.01; ****P < 0.0001.

Fig. 4.

The phenotype of dkc1^elu1/elu1 larvae recapitulates the human phenotype. (A) Histological analysis of dkc1^elu1/elu1 mutant larvae shows microphthalmia and cataracts. Both the eyes and the optic tectum of the mutants are abnormal and contain a high prevalence of cells with neuroepithelial character. Expression of cell-cycle markers ccnD1 and PH3 in the retinae and the tecta of 2-dpf and 3-dpf larvae, respectively, can be observed throughout these tissues instead of being restricted to the proliferative regions of the ciliary marginal zone and the mediolateral edges, suggesting a defective cell cycle. (All pictures show coronal sections.) (B) Further histological analysis shows 1) deformed semicircular canals, 2) undifferentiated gut, and 3) hypoplastic pronephros with a reduced number of Wt1-positive podocytes in the mutant animals. (Scale bars: 10 μm.) (C) When Dkc1 function is abrogated in Tg(foxd3:EGFP) animals using a synthetic MO oligo, parapineal migration is impaired, and the pineal–parapineal complex appears immature at 3 dpf. (White arrows denote the parapineal.) (D) Markers of tissue differentiation demonstrate a lack of differentiation in the intestines (ifbp), pancreas (try), and the major blood lineages (gata1 and rag1). (Black arrows denote area of expression.) (E) Injection of 1) human WT DKC1 mRNA resulted in phenotypic rescue of the mutant larvae, as shown by the genotyping of larvae showing a WT phenotype. In contrast, injection of 2) human Glu206Lys DKC1 mRNA elicited a much milder rescue, demonstrating the hypomorphic nature of this allele.

Fig. 5.

Ribosomal dysfunction in dkc1 zebrafish mutants due to defective pseudouridylation. (A) Telomere length is normal in dkc1^elu1/elu1 larvae at 4 dpf as measured by flow-FISH (n = 3 pooled samples of 10 larvae each, P = 0.7). Arb. u., arbitrary units; ns, not significant. (B) The 28S/18S rRNA ratio is increased in 4-dpf dkc1^elu1/elu1 larvae, suggesting impaired 18S rRNA processing (**P = 0.0033). (C and D) Immuno-Northern blot demonstrates a reduced pseudouridylation of 18S rRNA in dkc1^elu1/elu1, dkc1^elu8/elu8 4-dpf larvae (+/? vs. elu1/elu1: *P = 0.016, +/? vs. elu8/elu8: ***P = 0.00058) (C) and in the leukocytes of patient FamB IV:3 (D). (+/?: heterozygous or homozygous WT fish.) (E) The female with skewed X-inactivation shows a decreased pseudouridine (pseU)/U ratio in the leukocytes, as determined by HPLC-MS. (F) Gene Ontology analysis of differentially regulated genes from 36 hpf dkc1^elu1/elu1 larvae demonstrates an up-regulation of genes associated with ribosome assembly and function. Size of the circles indicates the number of genes associated with certain terms, and color indicates the level of enrichment: red indicates high enrichment, and blue indicates low. P.adjust, adjusted P. (G) Western blot suggests the stabilization of Tp53 in the affected cells. (H) Transcriptomic analysis shows that the truncated, antiapoptotic tp53 isoform (Δ 113p53) is up-regulated in mutants, while the canonical, full-length, proapoptotic isoform shows decreased expression; measured as fragments per kilobase of exon model per million reads mapped. (I) The phenotype of the dkc1^elu1/elu1 zebrafish mutants is Tp53-independent, as it is not rescued on a tp53⁻ background. (J) Homozygous carriers of the missense (c.567_568insGTG) hypomorphic allele (dkc1^elu2/elu2) are viable, but show significant growth retardation compared with their siblings (n = 130) (+/+ vs. elu2/elu2: ****P = 1.9 × 10⁻⁹; +/elu2 vs. elu2/elu2: ****P = 1.6 × 10⁻⁹). *P < 0.05;**P < 0.01; ***P < 0.001;****P < 0.0001.