Mutations in multiple components of the nuclear pore complex cause nephrotic syndrome

Abstract

Steroid-resistant nephrotic syndrome (SRNS) almost invariably progresses to end-stage renal disease. Although more than 50 monogenic causes of SRNS have been described, a large proportion of SRNS remains unexplained. Recently, it was discovered that mutations of NUP93 and NUP205, encoding 2 proteins of the inner ring subunit of the nuclear pore complex (NPC), cause SRNS. Here, we describe mutations in genes encoding 4 components of the outer rings of the NPC, namely NUP107, NUP85, NUP133, and NUP160, in 13 families with SRNS. Using coimmunoprecipitation experiments, we showed that certain pathogenic alleles weakened the interaction between neighboring NPC subunits. We demonstrated that morpholino knockdown of nup107, nup85, or nup133 in Xenopus disrupted glomerulogenesis. Re-expression of WT mRNA, but not of mRNA reflecting mutations from SRNS patients, mitigated this phenotype. We furthermore found that CRISPR/Cas9 knockout of NUP107, NUP85, or NUP133 in podocytes activated Cdc42, an important effector of SRNS pathogenesis. CRISPR/Cas9 knockout of nup107 or nup85 in zebrafish caused developmental anomalies and early lethality. In contrast, an in-frame mutation of nup107 did not affect survival, thus mimicking the allelic effects seen in humans. In conclusion, we discovered here that mutations in 4 genes encoding components of the outer ring subunits of the NPC cause SRNS and thereby provide further evidence that specific hypomorphic mutations in these essential genes cause a distinct, organ-specific phenotype.

Keywords: Genetics; Monogenic diseases; Nephrology.

Figure 1. Homozygosity mapping and whole exome sequencing identify recessive mutations of NUP107, NUP85, and NUP133 in 12 families with steroid-resistant nephrotic syndrome.

(A) Renal histology of individual A3825-21 (NUP107 mutation) shows diffuse mesangial sclerosis on light microscopy. (B, F, and I) Exon structure of human cDNAs. Positions of start codons and of stop codon are indicated. For protein domain structures, α-helices are depicted as red zigzag lines and β-turns as purple arrows. Arrows indicate positions of pathogenic mutations detected in families with SRNS. H, homozygous; h, heterozygous. (B) Exon structure, protein domain structure, and human mutations of NUP107. (C) Homozygosity mapping identifies 3 recessive candidate loci (red circles) in patient A4649-21. Nonparametric lod (NPL) scores and SNP positions (Affymetrix 250K StyI array) are plotted on human chromosomes concatenated from p-ter (left) to q-ter (right). Genetic distance is given in centimorgans (cM). Whole exome sequencing identifies a homozygous mutation of NUP107 (p.Met101Ile) that is positioned within the maximum NPL peak on chromosome 12 (arrowhead). (D, G, and K) Evolutionary conservation of amino acid residues that are altered in patients with SRNS. (D) Altered amino acid residues of NUP107 (p.Met101Ile, p.Tyr889Cys). (E) Renal histology of A3259-21 (NUP85 mutation) showing podocyte foot process effacement on transmission electron microscopy (TEM) (arrowheads). (F) Exon structure, protein domain structure, and human mutations of NUP85. (G) Altered amino acid residues of NUP85 (p.Ala477Val, p.Ala581Pro, p.Arg645Trp). (H) Renal histology of individual F797-21 (NUP133 mutation) shows podocyte foot process effacement on TEM (arrowheads). (I) Exon structure, protein domain structure, and human mutations of NUP133. (J) Homozygosity mapping in individual F797-21 identifies regions of homozygosity as recessive candidate loci. Within the maximum NPL peak on chromosome 1 (arrowhead), we identified a homozygous mutation in NUP133 (p.Ser974Arg). (K) Altered amino acid residues of NUP133 (p.Arg231Gly, p.Ser974Arg, p.Leu1055Ser).

Figure 2. Mutations of NUP107, NUP85, and NUP133 weaken protein-protein interactions within the NPC.

(A) The published structure of the Y complex of the NPC (11) was used to determine the localization of amino acid residues that we found altered in individuals with SRNS. Left: The 3D structure of the Y complex (pdb: 5A9Q): NUP133 (dark blue), NUP107 (light green), NUP96 (light blue), SEC13 (orange), SEH1L (violet), NUP85 (light gray), NUP43 (dark green), NUP160 (yellow), and NUP37 (red). Inset on right: The 3D structure of the interface region between NUP107 and NUP133 (pdb: 3CQC). The C-terminal part of NUP85 could not be fully resolved experimentally; a gray area indicates its predicted position. Residues Ala477, Ala581, and Arg645 of NUP85 (gray) and Glu803 of NUP160 (yellow) are located within incompletely resolved areas, and their positions are estimated. Note that amino acid residues Tyr889 of NUP107 (light green) and Ser974 of NUP133 (dark blue) point toward the interaction interface. (B) N-terminally FLAG-tagged NUP107 wild-type (WT) or mutant cDNA was overexpressed in HEK293T cells. Coimmunoprecipitation (coIP) demonstrates that the missense mutation Tyr889Cys weakens the interaction with endogenous NUP133. (C) A coIP experiment using N-terminally FLAG-tagged WT or mutant NUP133 cDNA demonstrates that the missense mutation Ser974Arg weakens the interaction with endogenous NUP107. As expected based on structural data, the 2 other missense mutations (Arg231Gly and Leu1055Ser) do not alter the NUP107-NUP133 interaction. (D) N-terminally Myc-tagged WT or mutant NUP85 cDNA was overexpressed in HEK293T cells. CoIP using an antibody against endogenous NUP160 shows that the missense mutations Ala581Pro and Arg645Trp of NUP85 weaken the interaction between the 2 proteins. FL, full-length; MOCK, empty vector. CoIP experiments in B–D were confirmed in 3 independent experiments.

Figure 3. Morpholino knockdown of nup85, nup107, or nup133 in Xenopus embryos causes defects in glomerulogenesis.

Xenopus embryos were injected with morpholino oligonucleotides (MOs) targeting nup85, nup107, or nup133 at the 2-cell stage. Abnormalities in pronephric development, specifically improper formation of the convoluted pronephric duct, were scored at stages 35–37. The pronephros was detected using whole-mount in situ hybridization and atp1a1 as a marker. (A) Schematic of the experimental setup, in which injection of MO into 1 cell of a 2-cell embryo allows for 1 side of the embryo to develop normally, while the other half serves as an internal control for developmental phenotypes. (B) Based on the severity of the phenotype, morphants were sorted into 4 groups: normal, mild phenotype (delayed or decreased convolution of the pronephric duct, top panel), moderate phenotype (loss of the characteristic pronephric architecture, middle panel), and severe phenotype (pronephros entirely absent, bottom panel). Shown here are examples for each category. (C–J) Left panels display the uninjected control side. Right panels display the injected side. Scale bars: 200 μm. (C and D) Control embryo (injected with nontargeting MO) displaying appropriate pronephric morphology for this stage (see B). (C) Uninjected. (D) Injected. (E and F) MO knockdown of nup85 results in abnormalities of pronephric development. (E) Uninjected. (F) Injected. (G and H) MO knockdown of nup107 causes developmental defects of the pronephros. Note the developmental delay and simplification of the convoluted pronephric duct. (G) Uninjected. (H) Injected. (I and J) Morpholino knockdown of nup133 causes pronephric developmental abnormalities. (I) Uninjected. (J) Injected. (K) Phenotypes in nup85, nup107, nup133, or nup155 morphants were categorized into the 4 groups. Note that as compared with uninjected controls (UC, n = 32), nup85 (65% abnormal, n = 34) and nup133 (86% abnormal, n = 28) morphants more frequently display severe phenotypes, while nup107 (38% abnormal, n = 34) morphants tend to show milder phenotypes. All experiments were performed at least twice.

Figure 4. CRISPR/Cas9–mediated knockout of NUP107, NUP85, or NUP133 induces filopodia formation and increases active Cdc42 in human podocytes.

Immortalized human podocytes underwent lentiviral transduction with a plasmid expressing a Cas9-GFP fusion construct under the control of a doxycycline-inducible promoter and a single gRNA. For each gene, 2 different cell lines were generated expressing gRNAs against NUP107 (targeting exon 4 or 11), NUP85 (targeting exon 1 or 15), or NUP133 (targeting exon 1 or 5), respectively. Experiments were performed 72 hours after induction of Cas9 expression with doxycycline (1 μg/ml). (A) We stained immortalized human podocytes expressing empty vector (MOCK) or individual gRNAs targeting NUP107, NUP85, or NUP133 with phalloidin to detect F-actin fibers. Podocytes that had either 3 actin-based protrusions or 1 filamentous protrusion of more than one-quarter of the cell body were quantified as “filopodia positive.” Representative images showing “filopodia-negative” control cells and NUP107-, NUP85-, or NUP133-knockout podocytes that exhibited filopodia (arrowheads). Scale bars: 25 μm. The result was confirmed in 3 independent experiments. (B) Quantification of approximately 50 cells for each condition resulted in 22% of MOCK-expressing cells with filopodia (11/50), in contrast to 58% (29/50) for NUP107 gRNA exon 4 (ex4) and 46% (23/50) for NUP107 gRNA ex11; 50% (25/50) for NUP85 gRNA ex1 and 54% (30/56) for NUP85 gRNA ex15; and 44% (22/50) for NUP133 gRNA ex1 and 38% (19/50) for NUP133 gRNA ex5. Note that knockout podocytes show increased filopodia formation. (C) Using the colorimetric G-Lisa Cdc42 Activation Assay Biochem Kit (Cytoskeleton), we demonstrate an increase in the active state of Cdc42 following CRISPR/Cas9–mediated knockout of NUP107, NUP85, or NUP133 in human podocytes. Data points represent 3 independent experiments (highlighted in different colors) and are displayed with mean and SD. P values calculated by 1-way ANOVA are indicated in the figure as *P < 0.05; **P < 0.01.

Figure 5. A truncating mutation but not a hypomorphic mutation of nup107 causes early lethality and developmental defects in zebrafish.

Zebrafish lines with mutations of nup107 were generated using CRISPR/Cas9 technology. Lethality following het × het in-crossing was monitored twice daily over the indicated periods. Genotyping was performed in all fish and was compatible with Mendelian ratios. (A) Kaplan-Meier survival curves of 86 larvae demonstrate that homozygous (hom) larvae carrying the frameshift mutation p.Thr81Argfs*74 of nup107 died before 5 dpf, contrary to heterozygous (het) and wild-type (WT) controls (n = 26 hom, 39 het, 21 WT). (B) Kaplan-Meier survival curves of a zebrafish line carrying a hypomorphic mutation of nup107 (p.Ala46delAla). Note that, contrary to the truncating allele, this in-frame deletion of nup107 does not impair survival of homozygous larvae compared with WT fish or heterozygous clutch mates (n = 14 hom, 27 het, 15 WT). (C–G) Phenotypes of homozygous nup107-knockout larvae (p.Thr81Argfs*74) on day 4 dpf. Specifically, the phenotype included small eyes, ventral body axis curvature, and peripheral as well as periorbital edema. (C and D) Yellow circumferences drawn around the pigmented area of the eyes of knockout fish (C) versus heterozygous clutch mates (D) assess eye size using ImageJ. (E) Quantification of eye size measurements (see C and D) demonstrates significantly smaller eyes in homozygous fish compared with heterozygous or WT clutch mates. One-way ANOVA with a standard confidence interval of 95% results in F(2, 84) = 84.72; P < 0.0001. Two-tailed P values (Šidák’s multiple-comparisons test) are shown in the figure (****P < 0.001). (F) Representative image showing ventral body axis curvature in a homozygous knockout fish. (G) Representative image displaying body and periorbital edema in a homozygous knockout fish. For quantification of F and G, see Table 2. Scale bars in C, D, F, and G: 500 μm.

Figure 6. Impact of a NUP37 mutation on the composition of the NPC and on nuclear structure.

A primary fibroblast cell line from patient PN-2 with the homozygous truncating mutation p.Arg306* of NUP37 (mutant) was compared with control fibroblasts (WT). DAPI stains DNA (blue). (A–C) Confocal microscopy of immunostaining for NUP37, NUP107, and NUP160 in mutant versus control fibroblasts. (D) Immunoblotting of NUP37, NUP160, and NUP107 in mutant versus control fibroblasts. α-Tubulin serves as a loading control. Note that protein levels are reduced in mutant fibroblasts. (E) Immunostaining with an antibody against several FG-repeat nucleoporins (mAb414) in control (left) versus mutant (right) fibroblasts. (F) Quantification of E demonstrates a significant reduction in the number of NPCs per square micrometer in mutant versus control cells. Data were obtained for 100 cells from 3 different experiments. Error bars denote SEM. P = 0.0048 (Student’s t test). (G) Immunostaining of HP1β (green), labeling heterochromatin, demonstrates an altered pattern in mutant versus control fibroblasts. (H) Fibrillarin (green) was used to stain nucleoli. Note that fibrillarin staining was more dispersed in mutant fibroblasts. (I) Quantification of 150 cells from 3 independent experiments demonstrates a significantly increased percentage of nuclei with abnormal nucleoli in mutant versus control cells. Error bars represent SEM. P = 0.0018 (Student’s t test). Scale bars in A–C, E, G, and H: 5 μm. (J–M) TEM images of control versus mutant fibroblasts. In control (WT) cells, a regular nuclear envelope (arrow outside the nucleus in J) and well-arranged heterochromatin in the proximity of the nuclear envelope (arrow inside the nucleus in J) can be seen. Note that the nuclear architecture of mutant fibroblasts is altered; specifically, (a) there is abnormal arrangement of heterochromatin and nucleoli (arrowheads, J vs. K); (b) the perinuclear space is widened and irregular (arrows, L vs. M); and (c) bulbous invasions of the nuclear envelope were observed (star, M). Scale bars are defined in each image. Immunofluorescence experiments (A–C, E, G, and H) and immunoblotting results (D) were confirmed in 3 independent experiments. **P < 0.01.