The ciliopathy gene product Cep290 is required for primary cilium formation and microtubule network organization

Abstract

The mammalian centrosome/cilium complex is composed of the centrosome, the primary cilium, and the centriolar satellites, which together function in key cellular processes including signaling. Defective assembly, maintenance, and function of the centrosome/ cilium complex cause the human genetic diseases known as ciliopathies, which are characterized by a multitude of developmental syndromes including retinal degeneration and kidney cysts. The molecular mechanisms underlying pathogenesis in ciliopathies remain poorly understood, which requires structural and functional characterization of the mutated ciliopathy proteins at the cellular level. To this end, we elucidated the function and regulation of Cep290, which is the most frequently mutated gene in ciliopathies and importantly its functions remain poorly understood. First, we generated Cep290-null cells using the CRISPR/Cas9 genome editing approach. Using functional assays, we showed that Cep290-null cells do not ciliate and that they have defects in centriolar satellites dynamics and interphase microtubule organization. The centriolar satellites were tightly clustered around the centrosome in Cep290-null cells, and the interphase microtubule network lost its radial organization. Our results provide phenotypic insight into the disease mechanisms of Cep290 ciliopathy mutations and also the tools for studying genotype/phenotype relationships in ciliopathies.

1. Introduction

The centrosome is the main microtubule-organizing center of most animal cells and has emerged as an important regulator of many critical cellular processes including cell division, polarity, and signaling. In cycling cells, the centrosome, composed of two centrioles and the associated pericentriolar material, organizes the interphase microtubule network and the mitotic spindle (Azimzadeh and Bornens, 2007) . Importantly, in some cycling cells and most quiescent noncycling cells, one of the centrioles forms the basal body that nucleates the microtubule axoneme of the cilia, including the nonmotile primary cilium, motile cilia, and the flagellum (Nigg and Ra,f 2009) . The primary cilium is a nonmotile sensory organelle, which serves as the nexus for growth factor and mechano-sensing signaling pathways important in development and tissue homeostasis including the Hedgehog, Wnt, PDGF, and cyclic nucleotide pathways (Johnson and Leroux, 2010; Nachury, 2014) . In addition, specialized multiciliated epithelial cells have many motile cilia on their surface, required for polarized movement of liquid and particles along epithelial cells, and sperm cells have a single motile flagellum, required for sperm motility. Defects in the assembly and function of cilia are associated with a number of human diseases, including ciliopathies and primary ciliary dyskinesia (Bettencourt-Dias et al., 2011) .

The centrosome is composed of two evolutionarily conserved microtubule-based cylindrical structures termed centrioles, which recruit a matrix of associated pericentriolar material (PCM) that nucleates and organizes microtubules. All centrioles share a remarkable 9-fold symmetric structure, but they differ in structural details, age, and function. Due to the mechanism of centriole duplication, one of the centrioles is the mother centriole and the other is the daughter centriole. The mother centriole, but not the daughter centriole, bears distal appendages that interact with the apical plasma membrane (Sorokin, 1968; Vorobjev and Chentsov, 1982) and these appendages are critical structures in transforming the centriole to a basal body for the elongation of the axoneme during cilia formation (Kobayashi and Dynlacht, 2011). In addition to the centrioles and pericentriolar material, animal cells have an array of 70- to 100-nm granules that localize around the centrosomes called centriolar satellites. Centriolar satellites are scaffolded by the 228 kDa coiled-coil protein pericentriolar material 1 (PCM1) and they localize and move around the centrosome in a microtubule- and molecular-motor-dependent manner (Tollenaere et al., 2015) . Centriolar satellites have been implicated in regulating centrosome- and cilia-related cellular processes by sequestering and delivering proteins to or away from the centrosomes and cilia including the ciliopathy-associated proteins OFD1 and Cep290 (Lopes et al., 2011) .

Ciliopathies are human genetic diseases that are characterized by a broad spectrum of developmental anomalies affecting multiple organ systems including retinal degeneration, polydactyly, situs inversus, hydrocephaly, polycystic kidney disease, and neurocognitive defects (Van der Heiden et al., 2011) . Among the ciliopathies are Joubert syndrome, nephronophthisis, Bardet–Biedl syndrome (BBS), oral-facial-digital syndrome (OFD1), and Meckel–Gruber syndrome (MKS), which are characterized by a distinct but overlapping set of clinical symptoms. Importantly, a majority of the ciliopathy mutations identified to date have been mapped to the genes that encode components of the centrosome/cilium complex (Reiter and Leroux, 2017) . Although over 187 genes were shown to be mutated in ciliopathies, the molecular mechanisms underlying the pathogenesis of these diseases remain poorly understood, which requires structural and functional characterization of the affected proteins in cells (Reiter and Leroux, 2017) .

Over the years, a number of the affected proteins in ciliopathies have been studied for their structure and function in order to reveal the molecular defects underlying ciliopathies. Ciliopathy-linked proteins were shown to function in the assembly, maintenance, and function of the primary cilium and these processes are tightly regulated by selective traficking of macromolecules in and out of the primary cilium in healthy cells (Stephen et al., 2017; Nachury, 2018) . Given their strong link to ciliopathy phenotypes, dissecting the ciliary protein transport mechanisms has been widely studied. Traficking of proteins from the cytosol to the primary cilium and the moving of proteins along the ciliary axoneme is regulated in part by vesicular traficking (Stephen et al., 2017) , intraflagellar transport (IFT) (Taschner and Lorentzen, 2016) , the BBSome complex (Mourao et al., 2016; Nachury, 2018) , and the transition zone (Goncalves and Pelletier, 2017).

The transition zone is a selective barrier at the part of the primary cilium adjacent to the basal body that is composed of the transition bifers, ciliary necklace, and Y-links (Goncalves and Pelletier, 2017). The transition zone regulates ciliary compartmentalization through coupling the ciliary membrane to the microtubule axoneme and impeding diffusion of proteins and lipids, and it controls the entry and exit of proteins from the primary cilium. Over the years, the majority of the molecular components of the transition zone have been identified, including CC2D2A, MKS1, MKS3, MKS5, NPHP1, NPHP4, Tctn1, Tctn2, and Cep290 (Goncalves and Pelletier, 2017).

Cep290 is one of the most intriguing ciliopathy disease genes as mutations of Cep290 cause a wide variety of distinct phenotypes, ranging from isolated blindness, Joubert syndrome, and BBS to the lethal MKS (Valente et al., 2006; Baala et al., 2007; Brancati et al., 2007; Leitch et al., 2008) . The wide spectrum of diseases linked to Cep290 highlights the central and important roles of Cep290 in cilium-related processes. Importantly, over 100 unique Cep290 mutations were identified in ciliopathy patients, making this gene the most frequently mutated gene in ciliopathies (Coppieters et al., 2010) . However, genotype/ phenotype studies for these unique mutations are very limited and thus the molecular mechanisms underlying the pathogenesis of Cep290-linked ciliopathies remain poorly understood, limiting the predictive power of a Cep290-related genotype. In order to understand how the unique Cep290 mutations result in the various Cep290 disease phenotypes, it is essential to elucidate the function and regulation of Cep290 at the cellular and organismal level.

Cep290 forms the Y-links of the transition zone and functions as a gate keeper to control ciliary protein composition by restricting the entry of nonciliary proteins (Betleja and Cole, 2010) . In the green alga Chlamydomonas, Cep290 localizes to the flagellar transition zone and forms the Y-links of the flagellar transition zone, which associates microtubules with the flagellar membrane (Betleja and Cole, 2010) . The Chlamydomonas Cep290 mutant flagella has abnormal levels of IFT complex proteins and the BBSome component BBS4 (Craige et al., 2010) . Phenotypic characterization of Cep290-depleted cells by different groups together showed that Cep290 localizes to the centriolar satellites and the transition zone, and that Cep290 functions in regulating cilium formation and ciliary recruitment of the ciliogenic small GTPase Rab8 (Frank et al., 2008; Kim et al., 2008; Tsang et al., 2008; Stowe et al., 2012) . Finally, characterization of Cep290-/- mice showed that Cep290 loss in mice recapitulates ciliopathy symptoms including early vision loss and hydrocephalus and that Cep290 is required for formation of connecting cilia in photoreceptor cells and maturation of ciliated ventricular ependymal calls (Rachel et al., 2015) .

Here, we describe the phenotypic consequences of Cep290 ablation on cilium formation and function, centriolar satellite dynamics, and microtubule organization in hTERT-RPE1 retinal pigmental epithelial cells. To elucidate Cep290 function and regulation in mammalian cells, we first generated Cep290-null cells in the RPE1::p53-/- parental background (Cep290 KO) by the CRISPR/Cas9 genome editing approach. Using functional assays, we showed that Cep290 KO cells do not ciliate and that they have defects in centriolar satellite distribution and interphase microtubule organization. The centriolar satellites were tightly clustered around the centrosome in Cep290 KO cells, and the interphase microtubule network lost its radial organization. These findings together identify functions for Cep290 in regulating specific cilium, satellite, and microtubule-mediated activities. The results of this study will provide important insight in elucidating how the unique Cep290 mutations result in the wide spectrum of human diseases.

2. Materials and methods

2.1. Cell culture and transfection

hTERT-immortalized retinal pigment epithelial (RPE-1, ATCC, CRL-4000) and RPE1:p53-/- cells were grown in DMEM/F12 50/50 (Invitrogen) supplemented with 10% fetal bovine serum (FBS; Atlanta Biologicals, Lawrenceville, GA, USA). RPE1::p53-/- cells were generated by CRISPRCas9 genome editing, kindly provided by Meng-Fu Bryan Tsou (Izquierdo et al., 2014). Human embryonic kidney cells (HEK293T, ATCC, CRL-3216) were grown in DMEM supplemented with 10% FBS. All cells were cultured at 37 °C and 5% CO2. All cell lines were authenticated by multiplex cell line authentication and were tested for Mycoplasma contamination by the MycoAlert Mycoplasma Detection Kit (Lonza). For ciliogenesis experiments, RPE1 cells were serum-starved in DMEM/F12 50/50 supplemented with 0.5% FBS for 24 or 48 h.

RPE1 cells were transfected with the plasmids using Lipofectamine LTX according to the manufacturer’s instructions (Invitrogen). HEK293T cells were transfected with the plasmids using 1 µg/µL polyethylenimine, MW 25 kDa (PEI, Sigma-Aldrich, St. Louis, MO, USA).

2.2. Nocodazole washout experiments

Cells were seeded on glass coverslips in 24-well plates and grown to 70%–80% confluence. Nocodazole (SigmaAldrich) was diluted to 10 µg/mL in culture medium from the 10 mM stock solution in DMSO. Cells were incubated with 10 µg/mL nocodazole for 1 h at 37 °C to achieve complete microtubule depolymerization. Cells were then washed 2 times with normal growth medium and incubated at 37 °C to allow regrowth of microtubules. During regrowth, cells were fixed at 0, 10, and 30 min in methanol and stained with the anti-α-tubulin and antiγ-tubulin antibodies, markers for the microtubules and the centrosome, respectively. Images shown correspond to a representative experiment from three independent experiments.

2.3. Lentivirus production and cell infection

HIV-derived recombinant lentivirus expressing the gRNA targeting Cep290 and Cas9 nuclease was made using the pLentiCRISPRv2g1 lentiviral transfer vector. Recombinant lentivirus was produced by cotransfection of HEK293T cells with the pCMVDR8.74 packaging vector and pMD2. VSVG envelope vector (Dull et al., 1998) using PEI. Six hours after transfection, media were replaced with complete medium containing 10% FBS, and lentiviral supernatant was harvested after 48 h and passed through a 0.45-µm filter. For infection, 1 × 10 5 RPE1 cells were seeded on 6-well tissue culture plates the day before infection, which were infected with 1 mL of viral supernatant the following day. Twenty-four hours after infection, the medium was replaced with complete medium, and 48 h after infection, cells were split and selected in the presence of 10 µg/mL puromycin for 5–7 days until all the control cells had died. LentiCRISPRv2 empty vector-transfected and puromycinselected cells were used as the control. After the Cep290 knockout efficiency was determined in the heterogeneous pool, cells were then trypsinized and serial dilutions were performed in normal growth medium. Colonies formed within 10–14 days were then trypsinized and expanded for screening for knockouts by immunoflourescence and western blotting.

2.4. Cloning and genome editing

The guide RNAs targeting the first exon of Cep290 were designed using the MIT CRISPR design tool (http://crispr. mit.edu). The highest scoring guide RNA was selected based on the predicted specificity to the target gene in order to reduce the of-target effects. Lentiviral transfer vector LentiCRISPRv2Cep290g1 was generated by cloning the following complementary oligos containing the Cep290 guide sequence and BsmBI (New England Biolabs, Ipswich, MA, USA) ligation adapters into LentiCRISPRv2 (gift of Feng Zhang, Addgene plasmid #52961). Oligos were purchased from Macrogen (Amsterdam, the Netherlands) and resuspended to a stock concentration of 100 µm in nuclease-free water and stored at –20 °C. The oligo sequences were as follows:

Oligo 1: CACCGTCTTGACGGGGCAGGTCATC, Oligo2: AAACGATGACCTGCCCCGTCAAGAC.

Briefly, oligos were phosphorylated using the T4 polynucleotide kinase (New England Biolabs), annealed, and ligated into the BsmBI-digested LentiCRISPRv2 plasmid. The construct was verified by DNA sequencing.

2.5. Antibodies

Anti-PCM1 antibody was generated and used for immunoflourescence as previously described (FiratKaralar et al., 2014). Other antibodies used for immunoflourescence in this study were goat anti-PCM1 (sc-50164; Santa Cruz Biotechnology) at 1:1000, mouse anti-γ-tubulin (GTU-88; T5326; Sigma-Aldrich) at 1:4000, rabbit anti-Cep290 (ab84870; Abcam) at 1:1000, and mouse anti-polyglutamylated tubulin (GT335; AG-20B0020-C100; Adipogen) at 1:500. Secondary antibodies used for immunoflourescence experiments were AlexaFluor 488-, 568- or 633-coupled (Life Technologies) and they were used at 1:2000. Antibodies used for western blotting were rabbit anti-Cep290 (ab84870; Abcam) at 1:1000 and rabbit anti-actin (CST 4970; Cell Signaling Technology) at 1:2000. Secondary antibodies used for western blotting experiments were IRDye 680- and IRDye 800-coupled and were used at 1:1000 (LI-COR Biosciences).

2.6. Immunouflorescence, microscopy, and quantification

For immunoflourescence experiments, cells were grown on coverslips and fixed in either methanol or 4% paraformaldehyde in PBS. After rehydration in PBS, cells were blocked in 3% BSA (Sigma-Aldrich) in PBS + 0.1% Triton X-100. Coverslips were incubated in primary antibodies diluted in blocking solution, and Alexa Fluor 488-, 594-, or 633-conjugated secondary antibodies were diluted at 1:2000 in blocking solution (Invitrogen). Coverslips of cells were imaged using LAS X software (Premium; Leica) on a scanning confocal microscope (SP8; Leica Microsystems) with Plan Apouflar 63X 1.4 NA objective.

Quantitative immunoflourescence for PCM1 was performed on cells by acquiring a z-stack of control and depleted cells using identical gain and exposure settings. The z-stacks were used to assemble maximum-intensity projections. The centrosome region was defined by g-tubulin staining in each cell and its fluorescence pixel intensity was measured by defining a circular area of 3 µm2 centered on the centrosome as the region of interest. Quantifications and image processing were performed using ImageJ (National Institutes of Health, Bethesda, MD, USA). The background was quantified by measuring the fluorescence intensity of a region of equal dimensions in the area neighboring the centrosome and the satellites. Primary cilium formation was assessed by counting the total number of cells and the number of cells with primary cilia, as detected by glutamylated tubulin staining. All data acquisition was done in a blinded manner.

2.7. Cell lysis and immunoblotting

For preparation of cell extracts, cells were lysed in 50 mM Tris (pH 7.6), 150 mM NaCI, 1% Triton X-100, and protease inhibitors for 30 min at 4 °C, briefly sonicated, and centrifuged at 15,000 × g for 20 min. The protein concentration of the resulting supernatants was determined with Bradford solution (Bio-Rad Laboratories, Hercules, CA, USA) using bovine serum albumin as a standard. For immunoblotting, cell extracts were first resolved on 10% acrylamide sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using the Mini-Protean system (Bio-Rad Laboratories). Resolved proteins were then transferred onto nitrocellulose membranes via the Mini Transblot wet transfer system (Bio-Rad Laboratories). After transfer, membranes were incubated in blocking solution (PBS + 5% milk, 0.1% Tween-20) for 1 h at room temperature with gentle rocking, followed by primary antibody incubation overnight at 4 °C, 3 washes with PBS + 0.1% Tween-20 (PBST), secondary antibody incubation for 1 h at room temperature, and 3 washes with PBST. Visualization of the blots was carried out with the LI-COR Odyssey scanner and software (LI-COR Biosciences). Blots were imaged using an Odyssey Infrared Imaging System with scan resolution at 169 µm.

2.8. Statistical analysis

Statistical significance and P-values were assessed by oneway analysis of variance and Student t-tests using Prism software (GraphPad Software, La Jolla, CA, USA). Error bars reflect SEM or SD. The following key is followed for asterisk placeholders for P-values in the figures: ***P < 0.001, **P < 0.01, *P < 0.05.

3. Results

3.1. Generation and validation of RPE1::p53-/- Cep290/- cell lines

To elucidate the function of Cep290 in ciliopathyassociated phenotypes, we first generated Cep290 null hTERT-immortalized retinal pigment epithelial cells null for p53 (hereafter RPE1::p53-/-) by employing the clustered regularly interspaced short palindromic repeats CRISPR/Cas9 technology to disrupt both copies of the Cep290 locus. Cells ablated for most centrosome proteins are not viable in the presence of p53 as the lack of centrosome proteins results in cell cycle arrest (Bazzi and Anderson, 2014) . Therefore, we used RPE1::p53-/cells for generating Cep290 nulls, which is a widely used approach for generating cell lines null for genes encoding centrosome proteins (Izquierdo et al., 2014). RPE1::p53-/cells were transfected with the pLentiCRISPRv2 plasmid harboring the Cep290 guide RNA (gRNA) designed to target exon 2 (protein-coding exon 1) or with control pLentiCRISPRv2 plasmid, sorted, and plated as single cells. We screened 10 single cell clones, detecting Cep290 first with immunoflourescence experiments and later with immunoblotting experiments for the clones that we validated as null by microscopic analysis. The two clones we picked for further phenotypic characterization of RPE1::p53-/- Cep290-/- (hereafter Cep290 KO1 and KO2) lacked detectable Cep290 signals while the signals for the centrosome protein γ-tubulin were detectable at similar levels in control and Cep290 KO cells (Figure 1A). In agreement with the immunoflourescence data, as shown in Figure 1B, the rabbit polyclonal Cep290 antibody recognized a major band above 250 kDa in control cells in immunoblot experiments, which disappeared in the Cep290 KO 1 clone, validating the Cep290 null state. Characterization of the Cep290 KO clones by immunoflourescence and immunoblotting experiments together confirmed that our strategy generated a clone that is null for Cep290 and we used this clone in further experiments to elucidate the cellular functions of Cep290.

Figure 1.

Generation and validation of Cep290 KO cells. A) Immunofluorescence analysis of control and Cep290 KO1 and KO2 cells using the Cep290 antibody (green) and γ-tubulin (red) antibody, which marks the pericentriolar material. DNA was stained with DAPI. Scale bar = 5 μm. The yellow boxes in the merged images were selected to zoom in in order to emphasize the lack of Cep290 signal in Cep290 KO cells relative to control cells that are positive for Cep290 signal at the centrosome and centriolar satellites; scale bar for the zoomed image is 1 μm. Images represent cells from the same coverslip taken with the same camera settings. Cep290 KO cells lack detectable Cep290 signal. B) Immunoblots of cell extracts of the control and Cep290 KO1 cells with anti-Cep290 antibody raised against the C-terminus of Cep290 and anti-actin antibody, which was used as a loading control. The 290 kDa band for the full-length Cep290 is absent in the Cep290 KO1 cells.

3.2. Cep290 KO cell lines are defective in primary cilium formation

Given that Cep290 is frequently mutated in ciliopathies and that its RNAi depletion resulted in about twofold decrease in the ciliogenesis efficiency of mammalian cells (Kim et al., 2008; Stowe et al., 2012) , we set out to determine the ability of Cep290 KO clones to form a primary cilium using quantitative immunoflourescence experiments. To test whether Cep290 is required or essential for efficient ciliogenesis, we serum-starved control and Cep290 KO1 and KO2 clones over 48 h to induce ciliogenesis and counted the percentage of cells that formed a primary cilium using glutamylated-tubulin antibody as the cilium marker (Figure 2A). Importantly, both Cep290 KO1 and KO2 clones did not ciliate after 48 h of serum starvation, indicating that Cep290 activity is essential for primary cilia formation (Figure 2B and 2C). Previous studies identified regulatory roles for Cep290 in cilium formation using RNAi experiments, where Cep290-depleted cells still ciliated to 30%–40% relative to the 80% ciliating control population (Kim et al., 2008; Stowe et al., 2012) . The derivation of Cep290 KO clones allowed us to assay ciliogenesis in a clean background, which is very different from the phenotypic analysis done using cells transiently depleted for Cep290 using RNAi experiments. Our results are significant in identifying Cep290 as an essential and critical component of the cilium biogenesis pathway and are in agreement with the cellular and organismal phenotypes reported in the Cep290-/- mouse (Rachel et al., 2015) .

Figure 2.

Cep290 is essential for primary cilium formation. A) Effect of Cep290 KO on cilium formation. Representative images of RPE1 control and Cep290 KO1 and KO2 cells fixed 48 h after serum starvation and stained with anti-glutamylated antibody (red), which marks the cilium, and the anti-PCM1 antibody (green), which marks the centriolar satellites. DNA was stained with DAPI. Scale bar = 10 μm. B) Quantification of percentage of cells with cilia in control and Cep290 KO1 and KO2 cells after 48 h of serum starvation. Cells were fixed, stained for cilia, and quantified. n = 500 cells for each sample, from three independent experiments. Data represent mean ± SD from three independent experiments, t-test was used for statistical analysis, **P < 0.01.

3.3. Centriolar satellites are redistributed in Cep290 KO cell lines

Centriolar satellites localize around the centrosome in a microtubule- and molecular-motor-dependent manner, which results in a majority of them clustering around the centrosome and the remaining being homogeneously distributed throughout the cytoplasm. Depletion of other satellite proteins including CCDC11 (Silva et al., 2016) and CEP72 (Stowe et al., 2012) were shown to reduce the ciliogenesis ability of the depleted cells by dispersing satellites throughout the cytoplasm or tightly clustering satellites around the centrosome, respectively. Given that centriolar satellite distribution was shown to play important roles during primary cilium formation and that Cep290 also localizes both to the centriolar satellites and the centrosome, we hypothesized that the primary cilium formation defect in Cep290 KO1 cells can be due to changes in the cellular distribution of centriolar satellites. To test this hypothesis and to test whether Cep290 is required for satellite dynamics in general, we used quantitative immunoflourescence experiments to determine the amount of pericentrosomal PCM1 in control and Cep290 KO1 cells. PCM1 is the molecular scaffold for centriolar satellites and quantification of its levels within an area of 3 µm2 centered on the centrosome is a standard assay used to assay satellite distribution phenotypes in cells (Stowe et al., 2012; Gupta et al., 2015) . Relative to control cells, Cep290 KO1 cells had a significant increase in the PCM1 intensity around the centrosome (P < 0.01) and a corresponding reduction in cytoplasmic centriolar satellite distribution (Figures 3A and 3B), identifying a tight clustering phenotype for satellites in Cep290 KO1 cells. These results show that Cep290 plays a role in the regulation of the regulation of dynamic satellite distribution in cells, possibly through regulating the satellite dynamics by modulating the affinity of the satellites to the molecular motors or microtubules. These results are in agreement with the previous reports that showed redistribution of centriolar satellites associated with an increase in their concentration around the centrosome upon RNAi-mediated depletion of Cep290 in mammalian cells (Kim et al., 2008) and thus further corroborate the function of Cep290 in regulation of satellite dynamics.

Figure 3.

Cep290 regulates centriolar satellite distribution. A) Effect of Cep290 KO on centriolar satellite organization. Control and Cep290 KO1 cells were fixed and stained with anti-PCM1 antibody (green), which marks centriolar satellites, and anti-Cep290 (red), which marks the centrosome and the centriolar satellites. DNA was stained with DAPI. Images represent cells from the same coverslip taken with the same camera settings. Scale bar = 10 μm. B) Effect of Cep290 KO1 on the pericentrosomal level of PCM1. Pericentrosomal PCM1 fluorescence intensity was measured for control and Cep290 cells in an area of 3 μm2 around the centrosome and levels are represented as arbitrary fluorescence intensity values. n = 50 cells for each sample, from two independent experiments. t-test was used for statistical analysis. Data represent mean ± SD from two independent experiments, **P < 0.01.

3.4. Microtubule network organization is disrupted in Cep290 KO cell lines

The pericentrosomal distribution of centriolar satellites is in part dependent on a centrosomally focused interphase microtubule radial array. Given that Cep290 was previously implicated to function in the formation and maintenance of a radial microtubule array anchored at the centrosome in interphase (Kim et al., 2008), we hypothesized that Cep290 might be regulating satellite distribution through its function during microtubule network organization. To determine whether redistribution of centriolar satellites in Cep290 KO1 cells is a direct effect of Cep290 or an indirect effect through loss of microtubule organization, we assayed the organization of microtubules in Cep290 KO1 and control cells using nocodazole washout experiments. Control and Cep290 KO1 cells were treated for 1 h with 10 µg/mL nocodazole at 37 °C to depolymerize all microtubules, which we confirmed by staining for microtubules in cells fixed after nocodazole treatment indicated as 0 min in Figure 4A. Following nocodazole treatment, cells were washed to remove nocodazole and to allow regrowth of microtubules at 37 °C. We visualized the microtubule network in control and Cep290 KO1 cells by staining with a monoclonal antibody against α-tubulin (Figure 4A). Ten minutes after nocodazole washout, both control and Cep290 KO1 cells nucleated microtubule asters of similar sizes around their centrosomes, indicating that Cep290 does not result in microtubule nucleation defects (Figure 4A). However, 30 min after nocodazole washout, control cells had significantly (P < 0.05) higher percentages of centrosomally organized distinct radial microtubule arrays (78.5 ± 4.9%, n = 200) than Cep290 KO1 cells (52 ± 4.2%, n = 200) (Figure 4B). This result indicates that Cep290 is required for microtubule organization and that might in part explain the centriolar satellite distribution and ciliogenesis phenotypes of Cep290 KO1 cells.

Figure 4.

Cep290 is required for the interphase microtubule organization. A) Control and Cep290 KO1 cells were treated with 10 μg/mL nocodazole for 1 h and microtubules were visualized right after nocodazole treatment (t = 0 min) and 10 and 30 min after nocodazole washout. Representative images of RPE1 control and Cep290 KO1 cells fixed throughout the nocodazole washout experiment at the indicated time points and stained with α-tubulin antibody (green), which marks the microtubules, and the anti-Cep290 antibody (green), which marks the centrosome and the centriolar satellites. DNA was stained with DAPI. Scale bar = 10 μm. B) The microtubule staining of the same images of control and Cep290 KO1 cells 30 min after nocodazole washout in A is presented in inverted form in order to emphasize the defects in interphase microtubule organization.

4. Discussion

Ciliopathies are diseases that result from defective centrosome/cilium-complex assembly and function and are characterized by a shared set of developmental symptoms including retinal degeneration, polydactyly, situs inversus, hydrocephaly, polycystic kidney disease, and neurocognitive defects. While genetic studies identified at least 187 genes mutated in ciliopathies (Reiter and Leroux, 2017) , the molecular mechanisms underlying pathogenesis in the ciliopathies remain poorly understood due to the lack of information on the functional and structural consequences of the ciliopathy mutations in affected genes. An essential first step to dissecting the molecular pathways underlying ciliopathy phenotypes is the structural and functional characterization of the ciliopathy genes at the cellular level as the results of such characterization will provide insight into which cellular processes are defective in ciliopathies. To address this longstanding problem, we dissected the centrosome/cilium-complex-related functions of the Cep290 ciliopathy gene product using the Cep290 KO clones we generated in this study.

The centrosome and centriolar satellite protein Cep290 is the most mutated gene in ciliopathies and we chose Cep290 for further characterization of its cellular functions for two main reasons. First, over 100 unique mutations are mapped to the Cep290 gene, which cause a wide spectrum of disorders including Joubert syndrome, MKS, and BBS (Valente et al., 2006; Travaglini et al., 2009) , making Cep290 the most mutated gene in ciliopathies. Second, Cep290 is the gatekeeper for regulating transport of proteins in and out of the cilia by forming the Y-links of the ciliary transition zone (Betleja and Cole, 2010) . In this study, we sought to determine the molecular defects underlying Cep290-linked ciliopathies by generating Cep290 null cells and phenotypically characterizing the phenotypic consequences of the Cep290 null state in cells. Our results suggest that Cep290 is essential for cilium formation and regulates centriolar satellite dynamics and organization of intact interphase microtubule arrays. These results highlight the multitude of cellular functions associated with Cep290 and provide insight in understanding the wide range of phenotypic defects underlying ciliopathies and in the phenotypic heterogeneity associated with Cep290-mutated ciliopathies.

Here, we show that CRISPR/Cas9-mediated ablation of Cep290 inhibited ciliogenesis and disrupted organization of the centriolar satellites and the interphase microtubule network. Overall, these results are consistent with the previous RNAi and short hairpin RNA (shRNA) studies that showed that transient knock-down of mammalian Cep290 leads to reduction in ciliogenesis, defects in radial microtubule organization, and clustering of centriolar satellites around the centrosome (Frank et al., 2008; Kim et al., 2008; Tsang et al., 2008; Stowe et al., 2012) . One major discrepancy between these CRISPR and RNAi studies is that Cep290 KO cells do not ciliate but Cep290-depleted cells still ciliate despite the significant reduction in efficiency. This discrepancy is likely due to the incomplete acute/transient depletion of Cep290 in siRNA and shRNA experiments. Our results identify Cep290 as an essential component of the cilium assembly pathway.

The ciliogenesis defects we observed in human RPE1 cells are in part in agreement with the phenotypic characterization of Cep290-/- mouse photoreceptor and ependymal cells derived from Cep290-/- mice that identified severe phenotypes in formation and/ or maturation of cilia in these cells (Rachel et al., 2015) . While photoreceptor Cep290-/- photoreceptor cells lacked the connecting cilium and were defective in inner and outer segment formation, Cep290-/- ependymal cells still had primary cilia, but these cilia showed progressive developmental defects. The variance in the severity of the ciliogenesis phenotypes in different tissues in mice and in Cep290 KO RPE1 cells suggests that Cep290 have tissue-specific functions during ciliogenesis and that these functions might contribute to the multitude of phenotypes in Cep290-linked ciliopathies.

Our characterization of Cep290 nulls for cilium assembly defects has two important consequences. First, it highlights the importance of functional follow-up studies of the RNAi data using genetic ablation experiments and also cautions against the direct interpretation of the RNAi phenotypes to explain the in vivo phenotypes in model organisms and the disease phenotypes. Second, it identifies an essential role for Cep290 in primary cilium assembly instead of the regulatory role identified by RNAi experiments. The underlying mechanism of this essential role must be elucidated in future experiments in order to determine where Cep290 functions in the ciliogenesis pathway. Once the exact molecular mechanism is known, these results will lead to understanding the disease mechanisms of Cep290-mutations.

Genetic ablation of Cep290 results in centrosomal clustering of centriolar satellites, which normally have a cytoplasmically distributed pool in mammalian cells. This result is in agreement with the reported phenotypes in RNAi studies and identifies a regulatory role for Cep290 in the dynamic behavior of centriolar satellites, which localize and move around the centrosome/cilium-complex in a microtubule- and molecular-motor-dependent manner (Tollenaere et al., 2015) . The traficking of the centriolar satellites to the centrosome requires dynein/dynactin complex activity, which was shown by dynein inhibition experiments, and this requirement was studied at the molecular level (Kubo et al., 1999; Dammermann and Merdes, 2002) . The traficking of centriolar satellites away from the centrosome possibly involves kinesin motors, which has not been experimentally tested before. Given the clustering phenotype we observed in Cep290-null cells, we suggest that Cep290 is either a positive regulator of dyneinmediated transport or a negative regulator of kinesinmediated transport. Proteomics studies identified putative and validated interactions of Cep290 with the dynactin components and the kinesin motor Kif4a, supporting our proposed model for Cep290 function in centriolar satellite traficking through modulating molecular motor interactions or activity. Future work is required to test the predictions of this model.

The results of this work will be significant in correlating the phenotypes of the Cep290 null cells with the phenotypic consequences of the Cep290 ciliopathy mutations that result in early stop codons in the Cep290 gene. Such analysis will contribute to determining genotype/phenotype relationships for Cep290 mutations that cause a frameshift mutation introducing an early premature stop codon that results in complete loss of protein. Our results suggest that Cep290 ciliopathy mutations might effect regulation of cilium assembly and centriolar satellite distribution. As an indication of the potential broader involvement of Cep290 in human disease, Cep290 has been reported to interact with other disease-associated proteins, including CCDC66, which causes retinal degeneration when mutated (Dekomien et al., 2010; Gerding et al., 2011; Conkar et al., 2017) .

In addition to characterization of Cep290 function in a clean null genetic background, the Cep290-null cells generated in this study will allow us and others to address another important question: what are the phenotypic consequences of the unique Cep290 mutations identified in ciliopathies? For example, experiments where Cep290 gene fragments harboring the disease mutations or Cep290 truncation mutants that mimic the truncated proteins described in patients with Cep290 mutations could be expressed in the Cep290 null to determine whether their expression rescues the satellite and microtubule organization defects of these cells. These experiments will reveal the genotype and phenotype relationships for different Cep290 mutations. While we are still a long way away from understanding how Cep290 dysfunction leads to disease and what causes the diversity of Cep290 mutations discovered in patients, our study contributes to this longstanding problem by identifying the precise functions of Cep290 in cilium-related cellular processes.

Acknowledgments

I acknowledge Kübra Hazal Zırhlıoğlu and Ezgi Odabaşı for insightful discussions regarding this work. RPE1::p53-/- cells were a kind gift from Meng-Fu Bryan Tsou (Memorial Sloan-Kettering Cancer Center, New York, NY, USA). This work was supported by the ERC StG679140 Grant, an EMBO Installation Grant, TÜBİTAK Grant 115Z521, a FABED Eser Tümen Research Award, and Turkish Academy of Sciences Distinguished Young Scientist Award.