Functional characterization of tektin-1 in motile cilia and evidence for TEKT1 as a new candidate gene for motile ciliopathies

Abstract

A child presenting with Mainzer-Saldino syndrome (MZSDS), characterized by renal, retinal and skeletal involvements, was also diagnosed with lung infections and airway ciliary dyskinesia. These manifestations suggested dysfunction of both primary and motile cilia, respectively. Targeted exome sequencing identified biallelic mutations in WDR19, encoding an IFT-A subunit previously associated with MZSDS-related chondrodysplasia, Jeune asphyxiating thoracic dysplasia and cranioectodermal dysplasia, linked to primary cilia dysfunction, and in TEKT1 which encodes tektin-1 an uncharacterized member of the tektin family, mutations of which may cause ciliary dyskinesia. Tektin-1 localizes at the centrosome in cycling cells, at basal bodies of both primary and motile cilia and to the axoneme of motile cilia in airway cells. The identified mutations impaired these localizations. In addition, airway cells from the affected individual showed severe motility defects without major ultrastructural changes. Knockdown of tekt1 in zebrafish resulted in phenotypes consistent with a function for tektin-1 in ciliary motility, which was confirmed by live imaging. Finally, experiments in the zebrafish also revealed a synergistic effect of tekt1 and wdr19. Altogether, our data show genetic interactions between WDR19 and TEKT1 likely contributing to the overall clinical phenotype observed in the affected individual and provide strong evidence for TEKT1 as a new candidate gene for primary ciliary dyskinesia.

Introduction

Cilia are evolutionarily conserved microtubule-based organelles which are classified into two main subgroups: (i) non-motile sensory or primary cilia, controlling key signaling pathways during development and tissue homeostasis and present individually on nearly all cell types and (ii) motile cilia, present on highly specialized cells for fluid displacement (oviduct, bronchial epithelial cells or ependymal cells) or to mediate cell movement (flagella). Primary and motile cilia can be distinguished by the structural organization of their axonemes.

All axonemes comprise nine pairs of peripheral microtubule doublets (‘9 + 0’ structure). While primary cilia do not present additional axonemal structures (‘9 + 0’ cilia), most motile cilia and flagella contain a ‘central’ pair of microtubules (‘9 + 2’ cilia) as well as additional extra-axonemal structures, including inner and outer dynein arms, directly required for motility, and radial spokes, which are believed to serve as sensors that control beating (1,2). Besides these ‘9 + 2’ motile cilia, another category of motile cilia has been described at the embryonic node in the mouse embryo and at the Kupffer’s vesicle (KV) in teleost fish. Those cilia show a 9 + 0 axoneme with dynein arms but no central pair, yet are still capable of the rotational movement required to establish directional fluid flow within these structures which govern laterality establishment in the body plan during development (3–6).

An additional structural component described in the axonemes of motile cilia and flagella is the ribbon, which consists of filaments assembled all along the lumen of the ‘A’ microtubule of each peripheral doublet. Ribbon filaments are composed of tektins, which can hetero/homo-dimerise thanks to their coiled-coil domains (7–10). In mammals, the tektin family is composed of five members (tektin-1–5). All tektins have been localized to sperm flagella, and knock-out mouse models for Tekt2, Tekt3 and Tekt4 have shown that they are critical for spermatozoa motility (11–15). Localization of tektin family members has not been investigated in the motile cilia of airway cells but their beating is impaired in Tekt2 knock-out mice (15). Tektin-1 remains poorly characterized and its role in ciliary/flagellar functions has not been directly investigated. Interestingly, the expression of tektins, including TEKT1, was shown to be controlled by the transcription factor FOXJ1, a master regulator of genes required for ciliary motility (16,17). In particular, TEKT1 expression is controlled by FOXJ1 in human and mouse bronchial epithelial ciliated cells as well as in the mouse node (18,19). In zebrafish, tekt1 expression is also controlled by foxj1a in the KV (17) as well as in the developing pronephros (20).

All cilia are assembled from basal bodies which correspond to the pre-existing mother centriole of the centrosome (primary cilia, sperm flagellum) or amplified centrioles (motile cilia). Assembly requires the intraflagellar transport (IFT) system, mediated by highly conserved machinery thought to select cargo (cilia components) in the cytoplasm and mediate their transport into the ciliary compartment along the forming axoneme. Anterograde transport from the base (basal body) to the distal ciliary tip is mediated by the IFT sub-complex B (IFT-B) in association with kinesin II, whereas retrograde transport, from the ciliary tip back to the base, is mediated by the IFT subcomplex A (IFT-A) in association with a dynein motor. Individual loss of most IFT subunits results in severe ciliogenesis defects in vitro as well as in vivo (21–23).

Dysfunction of either primary or motile cilia is associated with genetically heterogeneous disorders called ciliopathies, which can present a wide spectrum of clinical manifestations.

Defects in motile cilia are associated with primary ciliary dyskinesia (PCD) (OMIM #244400) characterized by neonatal respiratory distress and sinopulmonary syndrome (i.e. chronic rhinosinusitis and bronchitis with media otitis) that may be accompanied, depending on the mutated gene, by laterality defects [situs inversus (SI), ∼50% of patients] therefore corresponding to Kartagener syndrome (24–26). In most individuals with PCD, the ciliary defects are linked to mutations in the genes encoding either dynein arms, radial spoke proteins, or cytoplasmic chaperones responsible for stabilization, folding and pre-assembly of the dynein arm motors. However, some individuals presenting PCD do not have detectable structural axonemal defects (26,27).

Defects of primary cilia manifest as ‘ciliopathies’, complex multi-organ syndromes (retina, kidneys, liver, bones, etc.) linked to autosomal recessive mutations in genes encoding cilia-associated proteins (26,28). Ciliary chondrodysplasias are a ciliopathy sub-group including Short-Rib Polydactyly Syndrome (SRPS, forms 1–4), Ellis-van Creveld syndrome (EVC, MIM 225500), Jeune Asphyxiating Thoracic Dysplasia (JATD, MIM 208500), Sensenbrenner syndrome or Cranioectodermal Dysplasia (CED, MIM 225500), and Mainzer-Saldino Syndrome (MZSDS, MIM 266920). MZSDS is characterized by kidney disease (nephronophthisis; NPH), retinal dystrophy (RD), and skeletal abnormalities, of which the most characteristic is phalangeal cone-shaped epiphyses (29). Despite significant clinical overlap, and depending on the mutated genes, MZSDS cases generally have a milder rib phenotype with higher incidence of renal, retinal and liver involvement than the classical chondrodysplasia JATD (30–33).

Recently, our group and others have shown that ciliary chondrodysplasias are associated with mutations in genes encoding IFT-A complex subunits and associated dynein motor proteins (30,34,35). Specifically, mutations in WDR19, which encodes the IFT-A subunit IFT144, result in a large spectrum of ciliopathies including either isolated NPH, Senior-Løken (SLS: NPH and RD), NPH and liver involvement (Caroli disease or syndrome) as well as CED or JATD, depending on the severity of the mutations (29,36–41). Despite the fact that proteins encoded by primary ciliopathy genes are present in all types of cilia, clinical symptoms classically associated with PCD (e.g. bronchitis, rhinosinusitis, otitis and SI) have surprisingly rarely been described in patients presenting with primary cilia-related ciliopathies (26,28). This is particularly striking for IFT components which are required for the assembly of all types of cilia (42).

Notably, one individual presenting MZSDS also presented with additional manifestations not classically reported in other MZSDS conditions. These included severe airway infections associated with airway ciliary dyskinesia, classically displayed by PCD patients, and dilatation of brain ventricles, a phenotype rarely described in PCD but potentially linked to motile cilia dysfunction (43,44). We identified biallelic variations in two genes, WDR19 and TEKT1 which encode IFT144 and tektin-1, respectively. The aim of the present study was to investigate the contribution of each gene to the clinical manifestations observed in the affected individual. To do so we studied the possible function(s) of tektin-1 in cilia both in vitro and in vivo in the zebrafish. We also used the zebrafish model to evaluate the possible genetic interactions between these two genes.

Results

An individual presenting with a complex and severe ciliopathy phenotype

The affected individual (NPH1848) is the only daughter of healthy parents of French (Caucasian; mother) and Japanese (father) origin. The diagnosis of MZSDS was made based on the combination of her clinical features. She presented with congenital Leber amaurosis, cystic liver and pancreas, bone dysplasia, cone-shaped epiphyses (Fig. 1A) and thoracic distension (Fig. 1B). An ultrasound showed hyperechogenic kidneys with small cysts, and moderate dilation of the intra-renal cavities and left ureter. By 7 months of age the patient underwent end-stage renal disease, and a successful renal transplant was performed when she was 4 years old. Histological examination of the kidneys revealed a thickened basement membrane, associated with interstitial fibrosis and cysts throughout the renal parenchyma, phenotypes compatible with end-stage NPH.

Figure 1.

Identification of compound heterozygous mutations in two different genes, TEKT1 and WDR19, in an individual presenting with MZSDS and respiratory distress. (A) Left hand x-ray of individual NPH1848 showing cone-shaped epiphyses. (B) X-ray image of the thorax, showing cardiomegaly, thoracic distension, and interstitial infiltrates in both pulmonary fields (arrows). (C) MRI scan showing triventricular cerebral dilation and marked enlargement of the pericerebral space. (D) Abdominal contrast-MRI image showing hypodense parenchymal plaques and hyperdense patches of ‘ground glass’ opacity in the ventral culmen segment. (E) Variations, translational changes and predicted pathogenicity according to SIFT and Polyphen2 algorithms are indicated for both genes. (F) Inter-species sequence alignment showing conservation of the Lys311 (K311) residue. (G, H) Exon, cDNA and protein domain structure of human TEKT1 (G) and WDR19 (H). Positions of variations are indicated.

The index case also presents additional phenotypes that were not classically described in MZSDS. She presents intellectual disability associated with brain ventricle dilatation (Fig. 1C) as well as cardiomegaly (Fig. 1D). During her first year of life, the affected individual had recurrent airway infections with lung collapse, together with respiratory syncytial virus infections, and presented diffuse interstitial infiltrate in both pulmonary fields (Fig. 1B and D). Ciliary motility was first visually evaluated by optical microscopy on cells obtained from nasal brushing which led to the diagnosis of rarefaction of cilia and ciliary dyskinesia (4 months). A second analysis was done at 5 years of age by high-speed videomicroscopy, showing that cilia were sparse (see also below) and displayed aberrant beating patterns (slow or immotile) and heterogenous and non-synchronized movements (Supplementary Material, Video S1). Different populations of cilia were identified: 75% were immotile and 25% displayed dyskinetic movements with reduced amplitude and a very low ciliary beat frequency (4.7 Hertz) compared with control (Supplementary Material, Video S2). Such defects in cilia motility are frequently observed in PCD. However, despite the clinical evidence suggestive of PCD (brain ventricle dilations and ciliary dyskinesia with concomitant airway infections), the affected individual did not present sinusitis, media otitis nor situs inversus frequently observed in PCD (45). PCD is also characterized by lower levels of nasal nitric oxide, however, it was not possible to do such measurements in the NPH1848 individual due to her general health state.

In conclusion, the affected individual presented with both MZSDS and PCD-like airway disease. These manifestations are evocative of both primary and motile ciliary defects, and present an association of manifestations not described previously.

The affected individual carries compound heterozygous variations in two different genes

In order to identify the genetic cause of this complex ciliopathy spectrum, we applied exon-enriched NGS targeting up to 1, 205 ‘ciliary’ genes (‘ciliome’, Supplementary Material, Table S1) including ciliopathy genes, genes encoding known ciliary proteins and genes encoding paralogues or ciliary protein family members (30,31,46,47). We consequently identified biallelic variations in two genes (Fig. 1E).

Compound heterozygous variations were identified in TEKT1 encoding tektin-1, which has not previously been involved in ciliopathies but was considered a good candidate gene for motile cilia defects. One TEKT1 variation corresponds to a rare nonsense variant (c.730C > T [p.Arg244*]; 4.118e-05 in ExAC/3.658e-05 in gnomAD) inherited from the mother (Supplementary Material, Fig. S1A and B) which is predicted to lead to a truncated protein with the loss of the two last C-terminal coiled-coils (Fig. 1G). The other TEKT1 variation is a missense event (c.933G > T [p.Lys311Asn]; rs77092590; 0.0009721 in ExAC/0.0006713 in gnomAD) most frequent in the East Asian population (0.0119 in ExAc/0.0086 in gnomAD) and inherited from the father (Supplementary Material, Fig. S1A and B) who is of Japanese origin. This variation is predicted to be benign by Polyphen2 (0.286) and SIFT (0.15) but disease causing by Mutation taster (0.999), and the Lys311 residue is highly conserved across species and tektin family members (Fig. 1F). It is located in a well-conserved stretch of amino acids which corresponds to the linker domain between the third and fourth coiled-coils (Fig. 1G), thought to be important for the stabilisation of tektins (48).

In addition, we found a heterozygous missense variation (c.3533G > A [p.Arg1178Gln]) in exon 32 of WDR19 (Fig. 1H) which was inherited from the father (Supplementary Material, Fig. S1A and C). This variation was previously described as pathogenic in 10 affected individuals from 8 unrelated families presenting with NPH when either homozygous or associated with another mutation of WDR19 (37,39,49) (Table 1). Interestingly, among these individuals, all presented with NPH (or cystic dysplastic kidneys) and liver involvement (fibrosis or Caroli disease or syndrome), 6 with RD, 2 with mental disability associated with brain anomalies including hydrocephalus or subdural hygroma, 2 with polydactyly and 1 with cone shaped epiphyses. Furthermore and importantly, none of these individuals presented with airway manifestations.

Table 1.

Phenotypic spectrum associated with WDR19 mutations. Phenotypes of individuals presenting bi-allelic WDR19 mutations and heterozygous (He) or homozygous (Ho) for the p.R1178Q recurrent variation as identified in their respective cohorts including the NPH1848 individual described in this study, as well as individuals described in the two indicated published studies. The clinical features of an individual bearing a p.G495R homozygous variation were also included since she was presenting brain ventricle dilatations similar to NPH1848 individual

Cohorts	Individuals	Kidney	Retina	Liver	Pancreas	Skeleton	Cone- shaped epiphyses	Mental disability	Brain ventricles	Respiratory tract	Variations
this study	NPH1848	ESRD at 7 mo	x	x	x	thoracic distension	x	x	x	infections, ciliary dyskinesia	He: p.R1178Q
											He: del exons 1-4
Yoshikawa T 2017	Patient 1	ESRD at 3 mo	x	x		narrow thorax	x	x	subdural hygroma		He: p.R1178Q
						brachidactyly					He: N319Ifs*16
	Patient 2	ESRD at 11 mo	x	x	x	narrow thorax		x	x		He: p.R1178Q
						polydactyly					He: c.2645 + 1G
Lee JM 2015	I-1	ESRD at 6 y		x							He: p.R1178Q
											He: p.E1235K
	I-2	Proteinuria at 6 y		x							He: p.R1178Q
		ESRD at 9 y									He: p.E1235K
	I-3	Proteinuria at 8 y		x							He: p.R1178Q
											He: p.E1235K
	II-1	Proteinuria at 11 y	x	x							He: p.R1178Q
		ESRD at 15 y									He: p.G495C
	III-1	ESRD at 5 mo		x							He: p.R1178Q
											He: p.L618P
	IV-I	ESRD at 2 y	x	x							Ho: p.R1178Q
Halbritter J 2013	A4436-22	ESRD <1 y	x	x		polydactyly					Ho: p.R1178Q
	A3241-21	ESRD <1 y	x	x	x						He: p.R1178Q
											He: c.3565 + 1G>A
Fehrenbach 2014	1	Proteinuria at 6 y		x		scoliosis, hip dysplasia		x	x		Ho: p.G495R
		ESRD at 7 y

Though no other single nucleotide change in the exonic sequence of WDR19 could be detected, we observed that the number of reads for exons 1 to 4 was decreased by approximately ∼50% compared with the mean number of reads for these exons of individuals analysed in the same run (Supplementary Material, Fig. S2A), suggesting a deletion of the corresponding genomic region. The genomic deletion was further confirmed using SNPs and deletion boundaries were mapped within the 5’UTR region and intron 4 of WDR19 (Supplementary Material, Fig. S2B and C). RT-PCR and Sanger sequencing analyses of cDNA from individual NPH1848 showed that only the paternal allele (c.3533G > A [p.Arg1178Gln]) was expressed (Supplementary Material, Fig. S1D), confirming the absence of expression of the WDR19 maternal allele.

In conclusion, the affected individual carries bi-allelic variations in two genes: two compound heterozygous mutations in WDR19 including a previously reported pathogenic missense mutation (c.3533G > A [p.Arg1178Gln]) and a large deletion of the 5’ region encompassing the 4 first exons of the gene, as well as two compound heterozygous variations in TEKT1. We therefore investigated the functional effect of variations/mutations in both genes.

WDR19 mutations impair IFT144 functions at primary cilia

In agreement with a monoallelic expression of WDR19, expression of IFT144 was decreased in fibroblasts from the affected individual (Fig. 2A). Ciliogenesis was analyzed by immunofluorescence using acetylated α-tubulin (AcTub) as a marker of primary cilia (Fig. 2B), and fibroblasts from the affected individual showed reduced ciliogenesis (percentage of ciliated cells; Fig. 2C) and shorter cilia (Fig. 2D). Similar defects in ciliogenesis and/or cilium maintenance have been previously described in fibroblasts of affected individuals with WDR19 mutations (36). Moreover, staining with the ciliary membrane marker Arl13b showed that the ciliary tip adopted a spatulate shape in affected cells (Fig. 2E, arrows), with accumulation of the IFT-B subunit IFT46 at the ciliary tip (Supplementary Material, Fig. S3, arrows), indicative of retrograde transport defects (50). In control fibroblasts, IFT144 accumulated at the cilium base (Fig. 2E, arrowheads) and at a single spot at the ciliary distal tip (Fig. 2E, arrows), as previously described (36). In cells from the affected individual, IFT144 was absent from the tip (Fig. 2E, arrows). Finally, we analysed the ciliary localization of adenylate cyclase III (ACIII), since it was reported to be absent from cilia in mouse models with hypomorphic mutations in Wdr19 (51) and in fibroblasts from WDR19 mutated patients (31,46). ACIII was present in most cilia in control fibroblasts (>80%, Fig. 2F) whereas only a small proportion of cilia (<30%) were positive for ACIII in cells from individual NPH1848 (Fig. 2F, arrows).

Figure 2.

Fibroblasts from the affected individual show phenotypes typical of IFT-A subunit defects. (A) cell lysates from fibroblasts of two unrelated controls and from the affected individual (NPH1848) were analyzed by western blot using antibodies against IFT144 and α-tubulin as a loading control. (B) Fibroblasts from control and affected individuals were starved for 48 h to induce ciliogenesis and stained with acetylated α-tubulin (AcTub, red) in order to quantify ciliated cells (C) and measure the length of cilia (D) using ImageJ as indicated in materials and methods. Insets show higher magnifications of representative cilia indicated by white squares in the corresponding images. (E) Fibroblasts from control and affected individuals were starved for 48 h to induce ciliogenesis and stained for Arl13b (ciliary membrane, red) and IFT144 (green). Arrows indicate distal ciliary tips, arrowheads indicate basal body region. (F) Cilia in control and affected individual fibroblasts were stained for acetylated tubulin (AcTub, red) and for Adenylyl Cyclase III (ACIII, green). Insets show higher magnifications of representative cilia indicated by arrows. ***P < 0.001 from n = 3 independent experiments, t-test for ciliogenesis (C) and ACIII staining (F) and Dunn’s Multiple Comparison Test following the Kruskal-Wallis ANOVA test for cilia length (D).

In conclusion, the mutations in WDR19 present in the affected individual result in decreased expression of IFT144 and its mislocalization from the ciliary tip, likely resulting in defects in retrograde IFT transport as well as transport of ciliary membrane proteins.

Characterization of tektin-1 distribution at primary cilia and impact of identified variations

As the role of tektin-1 in cilia has not previously been investigated, we thus aimed to characterize the expression and distribution of tektin-1 in ciliated cells.

In western blot experiments two commercial antibodies against tektin-1 detected a band of 48 kDa corresponding to the size of the endogenous protein and additional bands corresponding to the expected sizes of transiently expressing GFP fusion forms of TEKT1 (GFP-TEKT1) in lysates from RPE1 cells (Supplementary Material, Fig. S4A and B). Immunofluorescence experiments also demonstrated specific detection of GFP-TEKT1 in transiently transfected cells (Supplementary Material, Fig. S4C). Altogether, these results indicate that both antibodies are specific (see also Fig. 5).

We then investigated the distribution of tektin-1 relative to the primary cilium in various cell lines (IMCD3, RPE1, fibroblasts) as well as in kidney biopsies. Tektin-1 staining was diffuse in the cytoplasm and nucleus (Fig. 3A and Supplementary Material, Fig. S4D and E). No co-localization could be observed with AcTub (Fig. 3A and Supplementary Material, Fig. 4D and E; arrowheads), indicating that tektin-1 does not localize to the axoneme of primary cilia. However, bright spots were observed at the base of cilia in all tested cell lines which co-localized with γ-tubulin (Fig. 3A and B; arrows). We then analysed the distribution of a tektin-1 GFP fusion (GFP-TEKT1) upon transient expression in RPE1 cells (Fig. 3C–E). In low-expressing cells, two bright spots of GFP-TEKT1 were present close to the nucleus. Those spots were ringed by pericentrin (Fig. 3C) and present between centrin and rootletin stainings (Fig. 3D), indicating that GFP-TEKT1 localizes to the proximal region of centrioles, in agreement with co-localization of endogenous Tektin-1 with γ-tubulin (Fig. 3B). In addition, GFP-TEKT1 could not be detected in cilia in transfected ciliated RPE1 or IMCD3 cells (Supplementary Material, Fig. S5), similar to the endogenous protein (see above). Altogether, our data confirm that tektin-1 is present at the centrosome in cycling cells as previously reported (52,53) and show that it is present at the basal body in ciliated cells but absent from the axoneme of primary cilia.

Figure 3.

The localization of tektin-1 at the centrosome is affected by the identified variations. (A,B) Cycling fibroblasts were processed for immunofluoresence with the Proteintech antibody against tektin-1 (TEKT1, green) and acetylated α-tubulin (AcTub, red, (A) or γ-tubulin (red, (B). Higher magnifications of representative centrosomes are shown in the right panels where AcTub (A) or γTub staining (B) is indicated by arrows. (C,D) RPE1 cells were transfected with wild-type (WT) GFP-TEKT1 encoding constructs, fixed in methanol and stained with pericentrin (peric, red; (C) or centrin (blue, (D) and rootletin (red, D). Higher magnifications of the centrosome-associated stainings are shown in the right panels and a schematic representation of the centrosome structural organization is indicated below with the localization of the different centriolar subdomain markers. (E). RPE1 cells were transiently transfected with plasmids encoding for WT, p.Arg244*(p.R244X), p.Lys311Asn (p.K311N) or p.Arg232Gln (p.R232Q) TEKT1 variants, fixed in methanol and stained for pericentrin (peric, red). Higher magnifications of representative centrosomes are shown in the right panels where centrioles stained with pericentrin are indicated by arrows.

To evaluate the impact of the identified variations in TEKT1 (p.R244* and p.K311N), we tested their effect on the subcellular localization of tektin-1 (Fig. 3E). We first tested their impact on the distribution of transfected tektin-1 GFP fusion in RPE1 cells. As described above, while WT GFP-TEKT1 colocalized with pericentrin at the centrosome in low expressing cells, both the p.R244* and the p.K311N variants remained diffuse in the cytoplasm with little or no centrosomal accumulation (arrowheads, Fig. 3E). As an internal control, we analysed the effect of the p.R232Q variation (p.Arg232Gln), a variation previously recorded (c.805C/T; rs115985064; 0.002159 in ExAc/0.002411 in gnomAD), predicted as damaging by polyphen2 and SIFT (0, 837/0.03), which is enriched in the African population (0.01211 in ExAc/0.01307 in gnomAD) and was identified in several individuals from our cohort. The resulting GFP fusion (p.R232Q) showed a distribution similar to WT GFP-TEKT1, with two bright spots colocalizing with pericentrin (Fig. 3E), showing that the defect in centrosomal localization is specific to the variations identified in the NPH1848 individual.

The consequences of TEKT1 variations were then analysed in fibroblasts from the affected individual. Expression of endogenous tektin-1 was decreased by approximately 50% in patient cells compared with controls (Fig. 4A). Interestingly, the truncated protein resulting from the p.R244* variant was not detected (expected size 22 kDa), possibly due to RNA decay and/or protein instability. Furthermore, the corresponding GFP fusion showed very low expression upon transient transfection in RPE1 cells (Supplementary Material, Fig. S4B). Then, as expected, the immunofluorescence tektin-1 staining was globally fainter in NPH1848 fibroblasts, including at the centrosome where the intensity of tektin-1 was significantly decreased compared with control cells (Fig. 4B and C). Altogether, these data show that the variations in TEKT1 result in decreased expression of tektin-1 as well as in decreased localization to the centrosome.

Figure 4.

Centrosomal localization of tektin-1 is affected in fibroblasts from the affected individual. (A) Cell lysates from fibroblasts of two unrelated controls and from the affected individual (NPH1848) were analyzed by western blot using antibodies against tektin-1 (Sigma) and GADPH as a loading control. (B) Cycling fibroblasts from control and affected individuals were stained for γ-Tubulin (γTub, red) and tektin-1 (proteintech, green). Insets show higher magnification of representative centrosomes where centrioles stained with γTub are indicated by arrows. Tektin-1 staining intensity at the centrioles was quantified by ImageJ (normalized arbitrary units, one representative experiment out of 2) ***P < 0.0001, t-test.

Because of the impact of WDR19 mutations on ciliogenesis, it was not possible to investigate the impact of tektin-1 variants on ciliogenesis in fibroblasts from the affected individual. Therefore, we used a siRNA-based approach to transiently knock-down expression of tektin-1 (siTEKT1) in RPE1 cells. As shown in Figure 5A, global expression of tektin-1 was strongly decreased in siTEKT1-treated cells compared with cells transfected with a control luciferase targeting sequence (siLUC). Treatment with siTEKT1 resulted in decreased intensity of tektin-1 staining at the basal body region (Fig. 5B and C). In addition, whereas knockdown of tektin-1 expression did not affect the formation of cilia (Fig. 5D), cilia were significantly shorter in tektin-1-depleted cells (2.2±0.67 µm vs 3.1±0.7 µm in controls; Fig. 5E). These results indicate that tektin-1 acts as a positive regulator of cilium length and therefore the identified variations in TEKT1 may contribute to the observed ciliogenesis defects in cells from the affected individual.

Figure 5.

Tektin-1 is a positive regulator of cilium length. (A) RPE1 cells treated with control luciferase-targeting siRNA (siLUC) and a pool of 4 different siRNAs targeting TEKT1 were analyzed by western blot using antibodies against tektin-1 (TEKT1; proteintech) and GADPH as a loading control. (B) RPE1-cells treated with the indicated siRNA for 48h were serum starved for 24h and stained for the ciliary marker acetylated-tubulin (AcTub, red) and tektin-1 (proteintech, green). Higher magnifications of representative cilia are shown in the right panels where tektin-1 signal at the basal body region is indicated by white arrows. (C) tektin-1 intensity staining at the basal body was quantified by ImageJ from images obtained as in (B) (normalized arbitrary units, one representative experiment out of 2). (D,E) The percentage of ciliated cells and length of cilia in siRNA-treated cells was quantified based on AcTub staining using ImageJ as in Figure 2. (C–E) ***P <0.0001, Mann-Whitney test following t-test.

Altogether, these results show that tektin-1 localises at the basal body but not at the axoneme of primary cilia and may act as a positive regulator of cilium length. The two variations in TEKT1 carried by the affected individual result in decreased expression of tektin-1 and negatively impact its localization.

Distribution and characterisation of tektin-1 at motile cilia and impact of TEKT1 variations

As tektin family members have previously been implicated in cilia motility, the distribution of tektin-1 was analysed in bronchial multi-ciliated epithelial cells. In lung sections from a control foetus, strong TEKT1 staining was observed at the apical region, which partially co-localized with AcTub (Supplementary Material, Fig. S6A), showing that tektin-1 is present at the axoneme of motile cilia. The distribution of tektin-1 was then analyzed in ciliated cells collected by nasal brushing. In contrast to control cells, tektin-1 was completely absent from cilia in cells from the affected individual (Fig. 6A and Supplementary Material, Fig. S6B). To test whether the lack of tektin-1 in cilia affected the subcellular localization of other axonemal proteins, we investigated the distribution of RSPH1 and DNALI1, two markers of radial spokes and inner dynein arms, respectively (25,54). Both markers similarly co-localized in cilia in cells from control and affected individuals (Fig. 6B), indicating that loss of tektin-1 does not result in global defects in ciliary composition. Therefore, to better understand the structural basis of the motility defect observed in individual NPH1848 (Supplementary Material, Video S1), transmission electron microscopy (TEM) was performed on bronchial cells to analyse the structure of the axoneme. No major defect could be detected in transverse sections of cilia (n = 41) which presented apparently normal organization of the axonemal microtubules (doublets and central pairs) and axonemal structures (dynein arms and radial spokes; Fig. 6C). However, in multiciliated cells from the affected individual, cilia appeared sparse and misoriented (Fig. 6D, arrows). In agreement with the rarefaction of cilia (see above), some basal bodies were found deeper in the cytoplasm, rather than docked at the apical plasma membrane (Fig. 6D, arrowheads).

Figure 6.

Affected individual multiciliated cells show alterations in tektin-1 ciliary composition. (A, B) Ciliated cells collected by nasal brushing from both affected and control individuals were labeled either for tektin-1 (proteintech, green) and acetylated tubulin (AcTub, red) (A), or with RSPH1 (green) and DNALI1 (red) (B). Higher magnifications of ciliary tufts are shown in the right panels. (C, D) Transmission electron microscopy was performed on ciliated cells from the affected individual. Cross-sections of a representative cilium (C) and longitudinal section of the apical region of a ciliated cell (D) are shown. Mispositioned basal-bodies and a misoriented cilium are indicated by red arrowheads and arrow, respectively.

In conclusion, our data show that ciliary motility defects in patient cells are associated with a specific loss of tektin-1 in cilia with no detectable ultrastructural defect of the axoneme.

Characterization of tektin-1 in zebrafish

The expression pattern of tekt1 in zebrafish has previously been examined by in situ hybridization, which has shown that tektin-1 is expressed in ciliated tissues, including the KV, pronephros, nasal pit and otic vesicle (20,55).

To investigate the role of tektin-1 in cilia in vivo, the expression of tekt1 was knocked down in zebrafish using a morpholino oligonucleotide targeting the exon2-intron2 splice site. tekt1 morphants showed very mild body curvature (Fig. 7A), a classical ciliopathy associated phenotype thought to be linked to convergence-extension defects (56), which was only observed in 22.72% of the morphants. However, a larger fraction of embryos did present with cysts of the pronephros (40.90%; Fig. 7B, arrow) and heart defects such as failure of the heart to loop, situs inversus, midline positioning or ‘heart string’ (42.42%; Fig. 7C). These phenotypes are classically observed in ciliopathy models ((56), Supplementary Material, Table S2) and were not present in control morphants. The observed heart phenotypes suggested laterality defects. Whole mount in situ hybridization experiments were performed to analyse the expression pattern of the L-R asymmetry marker southpaw (spaw). Normal ‘left-sided’ expression of spaw was detected in approximately 80% of both WT control and control morpholino-injected embryos (Fig. 7D), whereas spaw expression was aberrant in 50% of tekt1-MO embryos (right-sided, bilateral, absent or localised only to a discrete region within the developing tail; Fig. 7D). Interestingly, tekt1 morphants also presented incorrect formation and positioning of the otoliths in the otic vesicle which requires motile cilia function (57). Indeed, 54% of tekt1 morphants presented fused (arrows, Fig. 7E), misshapen or supernumerary otoliths (arrowheads, Fig. 7E). These results show that loss of function of tekt1 is associated with both classical ciliopathy-associated phenotypes and specific motile cilia dysfunction (Supplementary Material, Table S2).

Figure 7.

Ciliopathy-associated phenotypes in tekt1 morphant embryos. (A–D) Representative pictures of relevant phenotypes observed in zebrafish embryos injected with control and tekt1 targeting morpholinos. (A) Gross phenotype of injected embryos at 48 hpf. >80% of tekt1 morphant embryos showed normal body curvature. (B) Live imaging of pronephros of Tg(wt1b: GFP) embryos injected with tekt1 control and morpholino (dorsal view, anterior to the top). >40% of tekt1 morphants exhibited pronephric cysts (white arrow). (C) Heart morphology at 48 hpf in control and tekt1 morphants (ventral view, anterior to the top). White dotted lines show heart outline. tekt1 morphants show situs inversus, midline positioning or heart string. (D) Whole-mount in situ hybridization showing expression of laterality marker southpaw (spaw) at the 18–20 somite stage. Expression was mislocalized in tekt1 morphants compared with controls. (E) Otic vesicles imaged at 48 hpf. 55% of morphant embryos showed impaired otolith formation, with supernumerary (arrowheads) and/or fused otoliths (arrows).

The observed loss of cilia in multiciliated nasal cells from the affected individual and ciliogenesis defects in siRNA RPE1 cells (Figs 5 and 6) suggested that tektin-1 might be involved in ciliogenesis of both motile and primary cilia. We did not observe significant loss of cilia in the pronephric duct and nasal pits in tekt1 morphants (Supplementary Material, Fig. S7A and B). The role of tektin-1 in ciliogenesis and ciliary motility was investigated in the KV (Fig. 8A and B) and the pronephros (Fig. 8D and E) using Tg(βactin: arl13B-GFP) transgenic fish (58). The total number of motile cilia per KV was significantly reduced (∼50%) in tekt1 morphants, though the number of immotile cilia was unaffected (Fig. 8A and B). Importantly, KV shape and size appeared unaffected in the morphants (Fig. 8A). The angle of motility of the remaining motile cilia in the KV in tekt1 morphants was significantly reduced compared with controls (Fig. 8A and C), indicating decreased motility in the remaining motile cilia. Both the decreased number of motile cilia and their reduced motility are in agreement with the laterality defects observed in the morphants. Motility of cilia was also analysed in the distal pronephros/cloaca region. An increased proportion of immotile cilia was observed in tekt1 morphants compared with controls (Supplementary Material, Videos S3 and 4). Reduced motility of cilia in the morphants was further confirmed by measuring the angle of motility formed by beating cilia as for the KV (Fig. 8D and E).

Figure 8.

Ciliogenesis and ciliary beating are affected by loss of tekt1. (A–D) Live confocal imaging of cilia in Tg(βactin: arl13b: GFP) zebrafish injected with control or tekt1 morpholino. (A) Cilia in the Kupffers vesicle at 12–14 somites. (B) Quantification of motile/non-motile cilia in the K-V. tekt1 morphants show a significant reduction in the number of motile cilia in the K-V. (C) Quantification of angle of movement of K-V cilia. tekt1 morphants show a significant reduction in the angle of movement compared with controls (n = 4, *P < 0.05, Student’s t-test). (D) High speed confocal imaging of distal and cloaca regions of 48 hpf control & tekt1 morphant embryos (41 frames per second). Time-lapse shows 10 consecutive frames over a 240 millisecond period. (E) Quantification of angle of movement of pronephric cilia. Morphant embryos show reduced motility compared with control embryos.

Altogether, the results obtained in the zebrafish model show that tektin-1 is required for both fully efficient motility of cilia in various organs but also in ciliogenesis or maintenance of motile cilia in the KV where it is highly and specifically expressed in 1–4 somite embryos (55).

Genetic interactions between WDR19 and tekt1 in zebrafish

As WDR19 is known to be involved in skeletal and renal ciliopathies, it is conceivable that mutations in this gene could explain most aspects of the defects observed in the patient whereas the identified TEKT1 variations might participate to the motile cilia phenotypes in the patient. We thus used the zebrafish model to further study the genetic interactions between tekt1 and wdr19.

We first established a zebrafish morphant model for wdr19. A translation-blocking morpholino for wdr19 and a specific 5 bp mismatch control were used. Loss of wdr19 resulted in a ciliopathy phenotype with shortened and curved body axis, as well as microphthalmia (arrows) and moderate (∼50%) to severe (∼30%) hydrocephalus (arrowheads; Fig. 9A and B), two phenotypes absent in tekt1 morphants. On the contrary, situs inversus and cysts of the pronephros (two phenotypes commonly observed in tekt1 morphants) were observed in only a very small proportion of embryos (<10%), and the otoliths were unaffected (Fig. 9B). In addition, wdr19 knockdown embryos showed a striking loss of cilia in some tissues (Supplementary Material, Fig. S7C), as expected from its key function in intra-flagellar transport and in agreement with results obtained for other IFT genes (Supplementary Material, Table S2).

Figure 9.

Double knockdown of wdr19 and tekt1 results in a synergic phenotype. (A) Representative pictures of gross phenotype observed for 48 hpf zebrafish embryos injected with control or wdr19 morpholino. (B) Hydrocephalus phenotypes (arrowheads) and microphthalmia phenotypes (arrows) observed in wdr19 morphant embryos at 48 hpf. (C) Representative images of embryos injected with control morpholinos (combined ‘low dose’ (LD) of tekt1 control at 2 ng + wdr19 control at 1 ng), tekt1 morpholino at ‘low dose’ (LD, 2 ng), wdr19 morpholino at ‘low dose’ (LD, 1 ng), and combined wdr19 + tekt1 morpholinos at ‘low dose’ (LD, tekt1 at 2 ng + wdr19 at 1 ng). Hydrocephalus and otolith phenotypes are shown in enlarged panels (middle and right respectively). Hydrocephalus phenotypes are indicated with a black arrowhead. Otolith phenotypes are indicated with white arrows (supernumerary and fused otoliths). Microphthalmia is indicated by a white arrowhead. (D) Percentage of phenotypes observed for high dose (‘HD’) tekt1, high dose (‘HD’) wdr19, and combined low dose (‘LD’) tekt1 + wdr19 morpholino injections.

To investigate potential genetic interaction between the two genes, we performed a double knockdown of wdr19 and tekt1, having first validated that low doses (LD) of morpholino induced no effect on ciliated organ development and ciliogenesis when used individually (1 ng for wdr19 and 2 ng for tekt1, Fig. 9C). Double LD morphants exhibited hallmarks of both of the individual high dose (HD) morpholino models, including body curvature, microphthalmia, hydrocephalus, laterality and otolith defects (Fig. 9C). These phenotypes were associated with a dramatic loss of cilia (Supplementary Material, Fig. S7C), a phenotype not present in embryos treated with either of the single LD morpholinos.

Altogether the phenotypes observed in the double morphants indicate a synergic effect between tekt1 and wdr19, resulting in both primary and motile cilia defects, and mimicking the observed complex ciliopathy phenotype identified in the affected individual.

Discussion

We hereby characterize the subcellular localization of tektin-1 and its functions at both motile and primary cilia, in vitro and in vivo, and report for the first time damaging mutations in TEKT1 making it a new candidate gene for either motile or primary cilia ciliopathies. Interestingly, the affected individual described here carries compound heterozygous mutations in two genes, TEKT1 and WDR19, and though mutations in WDR19 have previously been associated with ‘primary cilia’ ciliopathies, it has never been associated with airway manifestations evocative of PCD. The latter manifestation is likely linked to motile cilia dysfunction, and so may be explained by the identified mutations in TEKT1.

Firstly, we have identified biallelic compound heterozygous mutations in WDR19; the p.R1178Q mutation which has been previously linked to a wide spectrum of NPH-associated ciliopathies [Table 1, (37,39,49)] and a loss of function allele resulting from a large deletion of the 5’ region of WDR19 including the 5’ UTR to intron 4. Compared with all other individuals bearing the p.R1178Q variation, individual NPH1848 presents with the most complex clinical features. The clinical presentation of the NPHP1848 individual is similar to that of two recently described cases (49), also carrying the p.R1178Q variant in association with a loss of function allele (Table 1). However, bronchiectasis and ciliary dyskinesia were not observed in those individuals.

The p.R1178Q mutation is quite frequent among ciliopathy cases carrying mutations in WDR19 and it was found in all the described cases with WDR19 mutations from Japan (49) and Korea (37) as well as in the NPH1848 individual allele inherited from her Japanese father. Indeed, this rare allele (0.0001304 in Exac; 6.802e-5 in gnomAD) is slightly enriched in the east Asian population (0.0005805 inExac; 0.0002211 in gnomAD), likely due to a founder effect similar to that which we reported for the p.P209L mutation in TTC21B/IFT139 (59).

Cognitive impairment was previously reported in several individuals with other WDR19 mutations (38,40,49) and was also associated with dilation of the brain ventricles (38,49). Interestingly, hydrocephaly was observed in wdr19 but not in tekt1 zebrafish morphants indicating that dilation of the brain ventricles observed in the NPH1848 individual is likely linked to mutations in WDR19 and must be included in their phenotypic spectrum. In addition, despite the fact that X-ray analyses were conducted in several patients (36,38), this is only the second report of cone-shaped epiphyses in the context of WDR19 mutations (49) allowing a clear diagnosis of MZSDS. It is thus very unlikely that TEKT1 mutations participate in this phenotype. Our data then also help to extend the spectrum of WDR19 mutations to MZSDS, as for IFT140 and TTC21B/IFT139 (30,32,60).

Mutations in WDR19 in the affected individual are associated with decreased expression of IFT144 and its mislocalization from the ciliary tip and mild ciliogenesis defects, as previously described in animal models (51) or for mutations in other individuals presenting with mutations in WDR19 or in other genes encoding IFT-A subunits (34,36). In addition, decreased ciliary ACIII was previously reported in the Wdr19 mouse model (50) and in individuals presenting with mutations in IFT-B encoding genes (31,46), in agreement with observations linking ACIII to IFT transport (21). Altogether, these data indicate that the primary cilia defects observed in fibroblasts from individual NPH1848 are likely predominantly due to the biallelic mutations in WDR19.

Though murine models have previously been described for this gene (50,51) there were no extant data on wdr19 in zebrafish. Morpholino-mediated loss of wdr19 results in hydrocephalus, body curvature, small eyes and a dramatic loss of cilia in most tissues examined. These phenotypes share some overlap with previously described zebrafish models for IFT-A components (60–62). Our wdr19 knockdown model does not display renal cysts, a phenotype that would have been expected for an IFT knock-down model (Supplementary Material, Table S2). This may be linked to the fact that the low non-toxic dose of morpholino used is not sufficient to generate an efficient knockdown of wdr19 in all tissues due to the differential maternal contribution.

In addition to mutations in WDR19, the NPH1848 individual carries variant alleles of TEKT1, a gene not previously identified in human ciliopathies. While there was no doubt about the pathogenicity of the rare p.R244* truncating mutation, the impact of the p.K311N variant remained to be demonstrated. Although this variant is predicted as benign by SIFT and Polyphen2, and is frequent in the Japanese population (∼1%, 3 homozygous individuals in gnomAD), we provide data supporting its deleterious effect on TEKT1 function. Firstly, the K311 residue is highly conserved across species and among tektin family members, and is predicted as disease causing using Mutation Taster. Secondly, the p.K311N protein is not correctly targeted to the centrosome in RPE1 cells. Thirdly, TEKT1 staining is completely lost at the axoneme of motile cilia in airway cells of the NPH1848 individual showing that the p.K311N variation is also defective for ciliary targeting. Although we cannot exclude that the 3 Asian individuals homozygous for p. K311N in gnomAD present mild airway disease, the association of the p.K311N variant with a loss of function mutation is expected to lead to a severe phenotype, as described for other allelic association of WDR19 in renal ciliopathies (36), or NPHS2 in nephrotic syndrome (63). Finally, the fact that none of the reported individuals with the p.R1178Q WDR19 mutation present with PCD-like manifestations further indicates that such clinical manifestations observed in the NPH1848 individual are likely linked to the mutations in TEKT1.

The ciliary functions of tektin-1 have not been studied previously. Our analyses show that neither endogenous nor transiently expressed tektin-1 were located at the axoneme of primary cilia. However, it localized to the axoneme of motile cilia of multiciliated nasal cells, in agreement with proteomic (64) and mRNA expression (11,18,48) studies. Interestingly, previously published work in the zebrafish has shown that transgenic ectopically expressed GFP-tagged human tektin-1 was found at the basal body of primary cilia whereas it was recruited to the axoneme only in cells ectopically expressing foxj1a (17), showing that the transport of tektin-1 into the cilium requires the expression of a specific transport machinery or partner(s) controlled by this master regulator of motile ciliogenesis.

Our investigations also indicate a positive role for tektin-1 in the length of primary cilia in vitro, in agreement with the results of a siRNA screen in the same cell line (65). In contrast, in vivo, zebrafish tekt1 single morphants do not show clear phenotypes linked to primary cilia which appeared normal in most tissues. However, the zebrafish model enabled us to show a synergic effect of tekt1 and wdr19 as demonstrated by the severe ciliopathy phenotypes and defective primary ciliogenesis in most tissues in embryos treated with suboptimal doses of both morpholinos. These results in the zebrafish might suggest that TEKT1 mutations could contribute to primary cilia dysfunction as well as the severity of the symptoms in the NPH1848 individual. However, the absence of airway-related features in other WDR19 individuals [Table 1; (49)] rather suggests that the effect of mutations in both genes is additive in humans.

Tektin-1 appears to have specific functions in motile ciliogenesis. Multiciliated nasal cells from the affected individual showed a rarefaction of cilia with undocked cytoplasmic basal bodies and a specific loss of motile cilia was strikingly observed in the KV of tekt1 morphants. Mutations in PCD genes are rarely associated with loss of motile cilia except in the case of CCNO and MCIDAS which control amplification of centrioles in multiciliated cells (66,67). Our results indicate that TEKT1 function in ciliogenesis is not similar to those of CCNO and MCIDAS since TEKT1 mutations do not seem to affect centriolar amplification in multiciliated cells in humans and in the zebrafish tekt1 knockdown only affects ciliogenesis in monociliated cells in the KV. Interestingly, FOXJ1, the upstream regulator of tekt1, is required for ciliogenesis in the KV (16,17) and in the node (68) as well as for the docking of amplified centrioles in the mouse trachea (68). Tektin-1 is therefore likely one of the key downstream effectors of FOXJ1 during these processes. Its precise function(s) however remain to be determined.

Altogether, these data suggest that the function of tektin-1 might be restricted to specific cell types and/or compensated in others due to the expression of other tektins. Alternatively, depletion of tektin-1, or the presence of abnormal tektin-1 protein may result in the instability of cilia (and/or anchoring of basal bodies), thus explaining their rarefaction.

In addition to its role in ciliogenesis, our live imaging analyses show that the remaining cilia in nasal multiciliated cells of the affected individual as well as those of the tekt1 morphants present reduced motility, further indicating that tektin-1 also participates in motility. Accordingly, morphant embryos did show otolith formation and laterality defects, two robust phenotypes which can be linked to specific motility dysfunction (69–71), and coherent with the previously demonstrated strong expression of tekt1 in these organs (55,57,71,72).

It is noteworthy that the affected individual lacks certain symptoms classically observed in PCD, including sinusitis, media otitis and situs inversus. However, situs inversus and media otitis are not observed in all PCD cases, notably for radial spoke head genes with individuals presenting RSPH1 mutations showing milder manifestations (25,54,73). In the case of TEKT1, it may be linked to a functional redundancy of tektins, or to the tissue-specific expression pattern of each tektin family member. It is also important to note that situs inversus was observed in the tekt1 zebrafish model and that it is present in only 50% of cases which may explain why it was not detected in a single individual. Our study would thus further benefit from the identification of additional cases with mutations in TEKT1, however, we failed to find additional variations of TEKT1 in a cohort of 80 individuals with ciliary dyskinesia and normal axoneme organization. We cannot exclude that the phenotypic consequences of TEKT1 mutations in the NPH1848 individual were exacerbated/potentialized by the WDR19 background. In this case, TEKT1 mutations in the absence of a similar ciliopathy genetic background might not lead to a clearly detectable PCD-like phenotype.

In conclusion, our data reveal the ciliary functions of tektin-1 and TEKT1 as a new candidate gene for PCD. This indicates that the complex ciliary phenotype observed in the NPH1848 individual compared with other individuals harboring WDR19 mutations likely results from an additive and/or synergic effect of mutations in both TEKT1 and WDR19 as demonstrated in the zebrafish.

Materials and Methods

Patients mutation screening using ‘ciliome’ sequencing

Written informed consent was obtained from participants or their parents, and the study was approved by the Comité de Protection des Personnes ‘Ile-De-France II’.

Ciliary exome–targeted sequencing and bioinformatics filtering was conducted using custom SureSelect capture kits (Agilent Technologies) targeting 4.5 Mb of 20, 168 exons (1221 ciliary candidate genes). In brief, SureSelect libraries were prepared from 3 µg of 300 genomic DNA samples sheared with an S2 Ultrasonicator, according to the manufacturers’ instructions. Precapture SOLiD libraries were prepared without any barcode. The SOLiD molecular barcodes for traceable ID of samples were added at the end of the capture step. The Ovation Ultralow System (NuGEN Technologies) was used to prepare HiSeq2500 precapture barcoded libraries. The ciliome capture by hybridization was performed on a pool of 10–16 barcoded precapture libraries. Sequencing was performed on pools of barcoded ciliome libraries (64 barcoded ciliome libraries per SOLiD FlowChip,\and 16 ciliome libraries per HiSeq FlowCell lane) using SOLiD5500XL (Life Technologies) and HiSeq2500. Paired-end reads were generated (75 + 35 for SOLiD and 100 + 100 for HiSeq) and mapped on the human genome reference (NCBI build37/hg19 version) using Burrows-Wheeler Aligner or mapread (SOLiD). Downstream processing was performed with the Genome Analysis Toolkit, SAMtools, and Picard Tools, following documented best practices (74). All variants were annotated using a software system developed by the Paris Descartes University Bioinformatics platform. The mean depth of coverage obtained was >90× and >89% of the exome was covered at least 15×. Different filters were applied to exclude all variants located in nonexonic regions, pseudogenes, UTRs, or known polymorphic variants with a frequency ∼1% (i.e. present in databases such as dbSNP and 1000 Genome Projects, and all variants identified by in-house exome sequencing of 5150 exomes and 1020 ciliomes). The functional consequence of missense variants was predicted using the SIFT (http://sift.jcvi.org/www/SIFT_enst_submit.html), PolyPhen2 (http://genetics.bwh.harvard.edu/pph2/) and MutationTaster (www.mutationtaster.org/) programs.

To validate the variations identified by ciliome sequencing and their segregation within the family, Sanger sequencing was performed on genomic DNA and cDNA using Big Dye Terminator V3.1 kit (Applied Biosystems) on a ABI Prism 3500x1 Genetic Analyser. Extended haplotype analysis using the four microsatellite markers indicated in WDR19 sequence was performed using Sanger sequencing.

Cell culture

Telomerase-immortalized retinal Pigment Epithelial cells (RPE1, ATCC) and Inner Medullary Collecting Duct cells (IMCD3, ATCC) were cultured in Dulbecco’s Modified Eagle Medium(DMEM)/F: 12, (Gibco^®, Life Technologies) supplemented with 10% Foetal Bovine Serum (FBS), glutamine and penicillin/streptomycin. Fibroblasts from affected and control individuals were cultured in OptiMEM (Gibco^®, Life Technologies) supplemented with 10% Foetal Bovine Serum (FBS), uridine, glutamine, fungizone and penicillin/streptomycin. To induce ciliogenesis in RPE1 or fibroblasts, cells were grown to confluence in basic cell culture conditions and then transferred in media without serum for 48 h. For IMCD3 cells, ciliogenesis was induced by growing the cells two days postconfluence in the presence of serum. Hela cells were cultured in Dulbecco’s Modified Eagle Medium(DMEM) (Gibco^®, Life Technologies) supplemented with 10% Foetal Bovine Serum (FBS), glutamine and penicillin/streptomycin.

RNA extraction and RT-PCR

mRNA was isolated from affected individual fibroblasts using Qiagen Extraction Kit and then treated with DNase I. Five micrograms of total RNA was reverse-transcribed using Superscript II (Life Technologies).

Antibodies

Mouse monoclonal antibodies (mAbs) against α-tubulin (T5168) and acetylated-tubulin (clone 6–11B-1), GAPDH (MAB374) and IFT144 (H00057728) and DNALI1 (H00007802-B01P) were from Sigma and Millipore and Abnova, respectively. Goat polyclonal antibodies against γ-tubulin (sc-7396) and rootletin (sc-67824) were from Santa Cruz. Rabbit polyclonal antibodies against tektin-1 were from either Sigma (HPA044444) or Proteintech (18968–1-AP) and the ones against Arl13b (17711–1-AP), ACIII (sc-588), pericentrin (ab4448) and RSPH1 (HPA017382) were from Proteintech, Santa Cruz, Abcam and Sigma, respectively. Alexa Fluor-conjugated secondary antibodies (Alexa488, Alexa546 and Alexa647) were from Life Technologies (Molecular Probes).

Plasmids, siRNA and transfections

A plasmid encoding wild-type tektin-1 GFP fusion (GFP-TEKT1) was purchased from gene Copoeia (clone n° EX-T3486-lv103). The identified variations were introduced by site directed mutagenesis using Pfu Turbo RNA polymerase (Agilent). The presence of the mutations was verified by Sanger sequencing.

Plasmids

Transient transfections of plasmids were performed using FuGENE^® HD Transfection (Promega) and 0.5µg of plasmid on cells grown on coverslips. Cells were transfected and immediately transferred in low serum conditions for 24 h for RPE1 cells or grown for additional 24 or 48 h in basic conditions for IMCD3 cells.

siRNA

Pool of 4 siRNA targeting Human TEKT1 and control non-targeting pools were purchased from Dharmacon (GE Healthcare, #83659). Transfections were performed using Lipofectamine RNAi MAX (Invitrogen, Life Technologies) with 80nM of siRNA. Cells were cultured for 48 h in DMEM F: 12/10% FBS, without antibiotics. After which fresh DMEM F: 12 without FBS nor antibiotics was added for 48 h to induce ciliogenesis before being fixed and processed for immunofluorescence.

Immunoblotting

Cells were lyzed in buffer containing 50 mM Hepes, 150 mM NaCl, 10 mM MgCl2, 0.5% NP40, 1% Triton X-100 and 10 mM protease inhibitor cocktail (Sigma), for 30 mins at 4 °C. After centrifugation, cleared lysates were separated by polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride transfer membranes (PVDF, GE Healthcare). Immunoblotting was performed using the indicated primary antibodies and revealed using the ECL⁺ Detection Kit (GE Healthcare), as described previously (46).

Immunofluroscence and immunohistochemistry

Cells grown on coverslips were fixed on ice either in methanol (MeOH, −20 °C) for 5 min, or in 4% paraformaldehyde (PFA) for 20 min. Fixed cells were incubated with primary antibodies in PBS containing 0.1% Triton-X-100 (Sigma) and 1 mg/ml bovine serum albumin (BSA, sigma) for 1h at room temperature. Then cells were incubated for 30 min at room temperature in PBS-BSA containing secondary antibodies. Nuclei were stained with Hoechst (#33342, Sigma) at room temperature during 5 min. Finally coverslips were mounted on microscope slides in Mowiol (Sigma).

Respiratory epithelial cells obtained by nasal brushing were spread onto glass slides and fixed in acetone for 10 min. Sections from human tissues (control adult kidneys and foetal lung) embedded in OCT and fixed respiratory epithelial cells were treated with PBS–0.1% Tween 20–3% BSA–10% donkey serum for 1h and incubated in the same buffer overnight at 4 °C with primary antibodies. Secondary antibodies were incubating for 1h at room temperature, Hoechst (#33342, Sigma) was used to label the nuclei and slides were mounted with FLUOPREP (BioMerieux sa, Marcy L’Etoile, France). Human kidney biopsies and respiratory cells were obtained from Necker-Enfants Malades and Trousseau hospitals (Paris, France), respectively. Guidelines to the declaration of Helsinki were followed.

Samples were examined with an epi-illumination microscope (DMI 6000, Leica) with a cooled charge-coupled device (CCD) camera (MicroMax, Princeton Instruments). Images were acquired Leica software and processed with ImageJ and Photoshop CS2 (Adobe Systems Inc., San Jose, CA, USA).

Image and statistical analyses

The percentage of ciliated cells, the length of primary cilium and staining intensities were quantified using fluorescence images and ImageJ as previously described (47). Statistical analyses were performed by using either a t-test for two-group comparisons or a One-way ANOVA test followed by Mann-Whitney post-test for three-group comparisons with GraphPad Prism software. Statistical significance was set at P < 0.05.

Transmission electron microscopy

Nasal biopsies were taken from the middle turbinate. The sample was fixed in 2.5% glutaraldehyde, washed with 1.3% osmium tetroxide and embedded in 1, 2-epoxypropan-epon mixture (1: 1) at 4 °C overnight, as previously described (75). After polymerization, several sections were picked out onto copper grids. The sections were stained with Reynold’s lead citrate. TEM was performed with the Philips CM10.

Zebrafish

Breeding and embryo collection

Adult zebrafish were maintained in system water at 28 °C, pH 7 and conductivity of 500µS on a 14 h light/10 h dark cycle. They were bred by natural crosses, and embryo were collected and staged as in Kimmel CB et al 1995. Embryos were maintained at 28 °C in embryo medium (0.1 g/L Instant Ocean Sea Salts, 0.1 g/L sodium bicarbonate, 0.19 g/L calcium sulphate, 0.2 mg/L methylene blue, H₂O) until the desired developmental stage was reached. Embryos were manually dechorionated. Transgenic lines Tg(bactin: arl13bGFP)^hsc5 (58) and Tg(wt1b: GFP)^li1 (76) were used to visualise the cilia and pronephros, respectively.

Morpholinos

To knockdown the expression of tektin-1, the tekt1 splice blocking morpholino (GGTTGGTTTGTTTTACATACCCAGT) was synthesized targeting the exon2-intron2 boundary, and a 5-base pair mismatch morpholino (GcTTGcTTTGTTTTAaATAaCCAcT) was designed as a negative control. Knock-down of wdr19 was performed using an ATG translation-blocking morpholino (GAAAACGTTACACTTCCCAGTGTGC) with a 5 base-pair mismatch control morpholino (GAAAtCcTTAgACTTCCgAcTGTGC). Dose-response experiments were performed to determine the optimum concentration of morpholino which produced no toxic effects. A 4 ng/embryo dose of the tekt1 splice morpholino yielded a reproducible phenotype, whereas injection of the 5 base-pair mismatch control did not result in any detectable phenotype at the same dose. A dose of 2.5 ng/embryo was determined for the wdr19 translation-blocking morpholino. The morpholinos were injected into embryos at the 1–2 cell stage with phenol red as a vehicle to visualise injections. Morpholino sequences were designed by and ordered from GeneTools, LCC. After microinjection, embryos were maintained as described above. Embryo development was evaluated at 24 hpf, 48 hpf, and 72 hpf.

Imaging

Phenotypes were analysed and imaged using a Leica M165FC stereoscope. Confocal images were taken using a LEICA SP8 microscope. High-speed videomicroscopy of ciliary beating in vivo was performed using a Zeiss Axio Observer Z1 confocal microscope equipped with Yokogawa Spinning Disk technology. Images were analysed with ImageJ.

Whole-mount in situ hybridisation

T7 was used for spaw1 transcription and digoxigenin labelling. For whole-mount in situ hybridization, embryos were fixed in 4% paraformaldehyde overnight at 4 °C and processed as described in Thisse and Thisse, 2008.

Supplementary Material

Supplementary Material is available at HMG online.

Acknowledgements

We are grateful to the patient and her family for their participation. We greatly acknowledge N. Goudin (Necker cell imaging facility) for providing expert knowledge on confocal microscopy, and C. Faucon for electron microscopy (CHIC, Créteil) and the ‘Plateforme d'Imagerie PICS’ (UMPC, Paris). We thank J-F. Papon (Service d’ORL, APHP, Hôpital Kremlin Bicêtre) for the performing and analyzing the nasal biopsy, and B. Louis (UMR_S955, UPEC) for videomicroscopy. We acknowledge the Imagine Institute for the purchase of Leica SP8 STED and Zeiss Spinning Disk microscopes, and the Fondation ARC for the purchase of the LEICA SP8 confocal microscope.

Conflict of Interest statement. None declared.

Funding

Fondation pour la Recherche Médicale (DEQ20130326532 to SS and RR), and a fellowship from the ‘Région Ile de France, CORDDIM’ (RPH12173KKA MF), Imagine Institute [ANR grant (ANR-A0-IAHU-01)], Fondation ARC (EML20110602384).

Author notes

Rebecca Ryan and Marion Failler authors contributed equally to this work.

Alexandre Benmerah and Sophie Saunier authors contributed equally to this work.