Targeted NGS gene panel identifies mutations in RSPH1 causing primary ciliary dyskinesia and a common mechanism for ciliary central pair agenesis due to radial spoke defects

Abstract

Primary ciliary dyskinesia (PCD) is an inherited chronic respiratory obstructive disease with randomized body laterality and infertility, resulting from cilia and sperm dysmotility. PCD is characterized by clinical variability and extensive genetic heterogeneity, associated with different cilia ultrastructural defects and mutations identified in >20 genes. Next generation sequencing (NGS) technologies therefore present a promising approach for genetic diagnosis which is not yet in routine use. We developed a targeted panel-based NGS pipeline to identify mutations by sequencing of selected candidate genes in 70 genetically undefined PCD patients. This detected loss-of-function RSPH1 mutations in four individuals with isolated central pair (CP) agenesis and normal body laterality, from two unrelated families. Ultrastructural analysis in RSPH1-mutated cilia revealed transposition of peripheral outer microtubules into the 'empty' CP space, accompanied by a distinctive intermittent loss of the central pair microtubules. We find that mutations in RSPH1, RSPH4A and RSPH9, which all encode homologs of components of the 'head' structure of ciliary radial spoke complexes identified in Chlamydomonas, cause clinical phenotypes that appear to be indistinguishable except at the gene level. By high-resolution immunofluorescence we identified a loss of RSPH4A and RSPH9 along with RSPH1 from RSPH1-mutated cilia, suggesting RSPH1 mutations may result in loss of the entire spoke head structure. CP loss is seen in up to 28% of PCD cases, in whom laterality determination specified by CP-less embryonic node cilia remains undisturbed. We propose this defect could arise from instability or agenesis of the ciliary central microtubules due to loss of their normal radial spoke head tethering.

Figure 1.

Sanger sequencing showing familial segregation of the RSPH1 mutations. (A) Pedigrees of families with RSPH1 mutations found in this study, demonstrating segregation consistent with autosomal recessive inheritance. (B) Sanger sequencing confirmed the segregation of three RSPH1 mutations as indicated in the pedigrees.

Figure 2.

Structural models of RSPH1, RSPH4A and RSPH9 showing location of the RSPH1 mutations. (A) The predicted structure of RSPH1 is shown with the position of the three identified mutations shown in relation to the gene and protein structure. RSPH1 has a low complexity region (yellow) at amino acid residues 238–251, and seven predicted MORN repeats (gray) according to Pfam and detailed in (B), consistent with the predicted structure of the mouse homolog (35). The Pfam consensus structure was chosen because Uniprot and SMART each predict six of these MORNs, but two are non-overlapping. (B) The three dimensional RSPH1 protein structure as predicted by I-TASSER is shown, with the seven MORN repeats highlighted at amino acid residues 26–43 (red), 44–66 (cyan), 67–89 (green), 90–112 (magenta), 113–133 (orange), 137–152 (purple) and 159–181 (yellow), in relation to the p.Trp94* and p.Glu29* nonsense mutations. (C) Shown in comparison to (A) are the protein structures of RSPH4A containing low complexity regions at residues 29–45, 53–66 and 80–100 (yellow) and a conserved region termed the ‘radial spoke domain’ at 207–697 (pink); and RSPH9 containing low complexity regions at residues 5–21 and 250–268. (D) Protein alignments of human (Hs) and C. reinhardtii (Cr) MORN repeat proteins of the radial spoke head show that the size of human RSPH1 (309 amino acids) is closer to CrRSP10 (216 amino acids) than CrRSP1 (814 amino acids). However, multiple sequence alignments and phylogenetic analyses support the genome annotation that RSPH1 closer resembles CrRSP1, which is 500 amino acids longer but has more similar sequences to RSPH1 outside of the MORN repeats (shown in pink). Human RSPH1 and CrRSP1 spoke head proteins therefore appear to have diverged in the sequences flanking their MORN motifs.

Figure 3.

Ciliary ultrastructural defects in individuals carrying RSPH1 mutations reveal distal central pair (CP)-deficient transposition defects sometimes retaining intermittent CP proximally. (A) Representative longitudinal TEM of affected individuals PCD-282 III:1 (left) and PCD-166 II:1 (right) shows the transposition event characteristic of the CP-deficient cilia. This occurs when a peripheral microtubule doublet is found in the central area of the cilium, in place of the central microtubule pair (white arrows) and this corresponds to an 8 + 1 pattern in cross-sections. Towards the cell body immediately beneath the transposition event the central area is visibly ‘empty’, lacking the CP completely, which corresponds to a 9 + 0 pattern in cross-sections. Further below, a partial CP is visible as a stump extending up from the base of the cilium, and this can sometimes be captured as having an intermittent appearance, apparently coming in and out of the plane of section (white arrow heads in PCD-282 III:1, left), however this ‘intermittent’ CP is not always visible in sections, as seen in individual PCD-166 II:1. Scale bar, 500 nm (B) Examples of transmission electron micrographs of cilia from the same affected individuals shown in cross section compared with control (9 + 2), showing they can have either a normal 9 + 2 arrangement, or 9 + 0, whilst more occasionally 8 + 1 sections are also seen. Scale bar, 100 nm.

Figure 4.

Individuals carrying RSPH1 mutations lack RSPH1 protein along the length of the axoneme. High-resolution immunofluorescence analysis in respiratory epithelial cells obtained by nasal biopsy is shown in unaffected controls compared with individuals carrying mutations in RSPH1 with anti-acetylated-α-tubulin used as a marker to stain the entire axoneme (red) compared with anti-RSPH1 (green), with nuclei DAPI-stained to show the DNA (blue). (A) RSPH1 protein is detected along the length of the cilia in a healthy individual. Individuals PCD-166 II:1 (B) and PCD-282 III:1 (C) have markedly reduced levels of RSPH1 protein. Scale bars represent 10 μm.

Figure 5.

Individuals carrying RSPH1 mutations lack other proteins of the radial spoke head suggesting the entire structure may be absent. High-resolution immunofluorescence analysis in respiratory epithelial cells obtained by nasal biopsy is shown in unaffected controls compared with individuals carrying mutations in RSPH1 with anti-acetylated-α-tubulin used as a marker to stain the entire axoneme (red) compared with two markers of the radial spoke head complex that are also implicated in PCD, anti-RSPH4A and anti-RSPH9 (both green). Nuclei are DAPI-stained to show the DNA (blue). (A) RSPH4A and (B) RSPH9 proteins decorate the entire cilia in healthy individuals whilst levels of both proteins in PCD-166 II:1 are severely reduced. Scale bars represent 10 μm.

Figure 6.

Individuals carrying RSPH1 mutations retain markers of the radial spoke stalk and the inner and outer dynein arms (ODA). High-resolution immunofluorescence analysis in respiratory epithelial cells obtained by nasal biopsy is shown in unaffected controls compared with individuals carrying mutations in RSPH1 with anti-acetylated-α-tubulin used as a marker to stain the entire axoneme (red) compared with a marker of the radial spoke stalk complex, anti-ROPN1L (green) and antisera against the ODA protein DNAH5 and the inner dynein arm (IDA) protein DNALI1 (green). Nuclei are DAPI-stained to show the DNA (blue). (A) ROPN1L, (B) DNAH5 and (C) DNALI1 immunostaining in PCD-166 II:1 is unaltered compared with healthy individuals. Scale bars represent 10 μm.