Characterization of Tetratricopeptide Repeat-Containing Proteins Critical for Cilia Formation and Function

Cilia formation and function require a special set of trafficking machinery termed intraflagellar transport (IFT), consisting mainly of protein complexes IFT-A, IFT-B, BBSome, and microtubule-dependent molecular motors. Tetratricopeptide repeat-containing (TTC) proteins are widely involved in protein complex formation. Nine of them are known to serve as components of the IFT or BBSome complexes. How many TTC proteins are cilia-related and how they function, however, remain unclear. Here we show that twenty TTC genes were upregulated by at least 2-fold during the differentiation of cultured mouse tracheal epithelial cells (MTECs) into multiciliated cells. Our systematic screen in zebrafish identified four novel TTC genes, ttc4, -9c, -36, and -39c, that are critical for cilia formation and motility. Accordingly, their zebrafish morphants displayed typical ciliopathy-related phenotypes, including curved body, abnormal otolith, hydrocephalus, and defective left-right patterning. The morphants of ttc4 and ttc25, a known cilia-related gene, additionally showed pronephric cyst formation. Immunoprecipitation indicated associations of TTC4, -9c, -25, -36, and -39c with components or entire complexes of IFT-A, IFT-B, or BBSome, implying their participations in IFT or IFT-related activities. Our results provide a global view for the relationship between TTC proteins and cilia.


Introduction
Cilia are evolutionarily conserved, microtubule-based, hair-like organelles present in most animal cells. They play important roles in cell movement, environment sensing, and signal transduction [1][2][3]. Defects in cilia assembly and function result in severe disorders called ciliopathies, including infertility, retinal degeneration, hydrocephalus, polycystic kidney, and situs inversus [2,4,5]. Therefore, it is important to elucidate the mechanisms underlying ciliogenesis and ciliary functions for the understanding and the treatment of ciliopathies.

Plasmids and mRNAs
The full-length cDNAs of human TTC4 (NM_004623), human TTC9c (NM_173810), human TTC25 (NM_031421), mouse TTC36 (NM_138951) and mouse TTC39c (NM_028341) were amplified by RT-PCR from HEK293T cells or mouse testis and cloned into pcDNA3.1-FLAG to express FLAG-tagged proteins in mammalian cells. For antibody productions, the fulllength cDNAs were subcloned into pET28a to express polyhistidine (His)-tagged TTC proteins in E. coli to serve as antigens. Full-length cDNAs were also subcloned into pGEX-4T-1 to express GST-tagged TTC proteins in E. coli for antigen competition experiment. For GST pulldown assays, the full-length cDNA of mouse BBS7 (NM_027810) was amplified by RT-PCR from mouse testis and cloned into pGEX-4T-1.
The GFP reporters for MO efficiency analyses were constructed in pCS2 + . An approximately 200-bp fragment of each zebrafish TTC cDNA containing the MO targeting sequence was cloned both upstream of and in frame with the coding sequence of EGFP. The mRNAs for the GFP reporters were in-vitro transcribed with mMESSAGE mMACHINE SP6 kit (Ambion, AM1340).
Primers used are listed in S1 Table. All the constructs were verified by sequencing.

Morpholino oligonucleotides and microinjection
The control MO (ctrl-MO) is the Morpholino Standard Control from Gene Tools. The antisense MOs were designed and ordered from Gene Tools. Their sequences are as follows, with the underlined bases complementary to the initiation codon ATG: MOs were dissolved in nuclease-free water and injected into the yolk of one-cell-stage zebrafish embryos by using Narishige IM300 microinjector as described before [24].
In the initial screen, a dosage of 8 ng per embryo was used [26][27][28]. As the morphants of ttc25, -4, -9c, -36, and -39c, but not of the rest, displayed some embryonic death, we further examined the dose-dependent effects (S1 Fig) and optimized the dosage for each of the MOs to minimize the death effect. Accordingly, all the presented experimental data were achieved with the following dosages: 4 ng per embryo for the MO of ttc25, -9c, -36, or -39c; 6 ng per embryo for the MO of ttc4; 8 ng per embryo for the MOs of the remaining TTC genes; 6 or 8 ng per embryo for the control MO to match the maximal dosage of TTC MOs in the experiments.
To validate the knockdown efficiency, 100 pg of in-vitro transcribed GFP-reporter mRNA were coinjected with either the corresponding MO or ctrl-MO per zebrafish embryo. The GFP autofluorescent and bright-field images of the embryos were acquired at 12 hpf using a SZX16 stereomicroscope (Olympus, Tokyo, Japan).
In the rescue experiments, 400 pg of in-vitro transcribed mRNA per embryo were coinjected with MO. The phenotypes were assayed at 72 hpf.

Fluorescence microscopy
MTECs were fixed and stained as described previously [29]. Briefly, MTECs grown on transwells were fixed with 4% paraformaldehyde in PBS for 10 min at room temperature. After fixation, cells were permeabilized with 0.5% Triton X-100 in PBS for 15 min and blocked with 1% BSA in PBS for 1 h. The incubations of primary and secondary antibodies were carried out at room temperature for 2 h and 1h, respectively. To validate the specificity of the homemade TTC antibodies, antigen competition experiments were performed as described [30,31]. Briefly, each antibody was pre-incubated with 15 μg of GST or GST-tagged antigen for 2 h at 4°C. After a 10-min full spin in a benchtop centrifuge, the supernatant was used for immunostaining.
Zebrafish embryos were manually dechorionated and fixed with 4% paraformaldehyde in PBS at room temperature for 4 h. After dehydration and rehydration, embryos were incubated with the antibody against acetylated tubulin (1:1,000) overnight at 4°C. After washing with 0.5% Triton X-100 in PBS, the embryos were incubated with Alexa Fluor-488-conjugated secondary antibody (1:1,000) for 2 h. After extensive washing, the embryos were embedded in 1% low-melting-point agarose.
Images were taken by using a Leica TCS SP5 MP confocal microscope. Optical sections were captured at 0.5-μm intervals and z-stack images were obtained by maximum intensity projections. Positions of the KV were determined through brightfield view of the 7-somitestage zebrafish embryos prior to the acquisition of the fluorescent images.

High-speed video microscopy
High-speed video microscopy for cilia motility was performed as described with some modifications [32]. Briefly, zebrafish embryos were treated with N-phenylthiourea (PTU, Sigma) at 24 hpf to inhibit pigmentation [33]. At 60 hpf, zebrafish were treated with 40 mmol/l BDM (2,3-butanedione monoxime, Sigma) and then embedded in 1% low melting agarose (Sigma) and the images of cilia beating were captured at 200 frames per second (fps) with IX71 inverted microscopy (Olympus) equipped with a 63 × oil-immersion objective lens and a Neo sCMOS camera (Andor). For each examined gene, cilia motilities of 10 randomly picked morphants were recorded.

In situ hybridization and histology
In situ hybridization was performed as described before [24]. The cmlc2 RNA probe was labeled using the DIG RNA labeling kit (Roche) and the probe concentration used was 2 μg/ml. Hybridization was performed overnight at 65°C.
For hematoxylin-eosin staining, embryos at 72 hours post fertilization (hpf) were fixed with 4% paraformaldehyde in PBS at room temperature for 4 h and embedded in paraffin after a gradual dehydration in ethanol. The embryos were cross-sectioned at a thickness of 5 μm using a rotary microtome (Leica RM 2235). Slides were then stained with hematoxylin and eosin.

Microarray hybridization and data analysis
To screen for genes upregulated during ciliogenesis, total RNAs of MTECs samples from ALI d0 to d5 were isolated and hybridized to a GeneChip Mouse Genome 430 2.0 array (Affymetrix) through a contracted service (CapitalBio). About 900 genes showed increased hybridization signals by at least 2 folds during multiciliation. The results for selected genes are shown in S2 Table. Co-immunoprecipitation and GST pull-down HEK293T cells transiently expressing FLAG-tagged proteins were lysed in the lysis buffer [20 mM Tris-Cl (pH 7.5), 100 mM KCl, 1% NP-40, 1 mM EDTA, 10% glycerol, 10 mM sodium pyrophosphate, 1 mM PMSF, 3 mM DTT, and protease inhibitor cocktail (Calbiochem)]. 3 ml of cell lysates from~2 × 10 7 cells were incubated with 3 ml of testis lysates from six 8-week mice for 2 h at 4°C. After the incubation, co-IP experiments were carried out with anti-FLAG resin (Sigma) as described previously [34].
Quantitative PCR 2 μg of total RNA was used for each reverse transcription reaction using the Superscript III First Strand Synthesis System with oligo-dT primers (Lifes). Quantitative PCR (qPCR) was performed by using an Applied Biosystems 7500 HT Sequence Detection System with the Power SYBR Green PCR Master Mix Kit (Applied Biosystems, Foster City, CA). GAPDH served as the control.

Quantification and statistical analysis
Cilia length was measured using Image J (Media Cybernetics). Statistical results, obtained from three independent experiments in a blind manner, were presented as mean values and standard deviations.
Cilia beat frequencies (CBFs) were calculated by playing the movies frame by frame for five complete beat cycles. For each group of zebrafish morphants, 40 cilia from 4 randomly picked embryos were used for statistical analysis. Only cilia that were clearly traceable in the timelapse movies were used for the analysis.

Multiple TTC genes are upregulated during MTEC differentiation
Following the establishment of an air-liquid interface (ALI), MTECs cultured in vitro simultaneously generate hundreds of centrioles, which serve as the basal bodies of cilia, followed by the formation of dense motile cilia ( Fig 1A) [23,29]. Accordingly, genes important for centriole biogenesis and cilia formation are upregulated. In our previous studies, we have demonstrated that during MTEC differentiation the centriole amplification peaks at ALI day 3 (d3) and multiciliation reaches plateau at around ALI d5 [24,29]. In the presence of DAPT, approximate 90% of the cells became multiciliated at ALI d5 ( Fig 1A) [24,29]. Therefore, we performed cDNA microarray analyses with MTECs samples from ALI d0 to d5 to identify ciliarelated genes.
Our microarray analyses identified approximately 900 upregulated genes, i.e., genes whose microarray signals increased by at least 2-fold, during the multiciliation of MTECs. As expected, the mRNA level of genes important for centriole biogenesis, such as DEUP1 (Fig 1B  and S2 Table) [29], peaked at ALI d2-d3, whereas genes related to cilia elongation, such as the components of BBSome and IFT complexes, reached their highest expression levels from ALI d4-d5 ( Fig 1B and S2 Table). Interestingly, twenty of the upregulated genes encoded TTC proteins. Compared to ALI d0, their microarray signals increased by 2 to 121 folds in later ALI days (Fig 1B and 1C and S2 Table). Half of them, including the TTC components of IFT-A,-B, and BBSomes and TTC25 and TTC26, are known to function in cilia ( Fig 1B) [21,22,35]. The remaining ten, however, did not have documented cilia-related functions (Fig 1C).
We then validated the microarray results with qPCR. Similar to the known cilia-related genes, BBS8, TTC25, and TTC26, the mRNA levels of all the ten novel TTC genes increased during ALI d0 and d5 (Fig 1D and S2 Fig), correlated with ciliogenesis and cilia function.
Zebrafish morphants of ttc4, -9c, -36, and -39c display abnormal body curvatures Zebrafish (Danio rerio) is an excellent vertebrate model animal because cilia-associated defects result in ciliopathy-related phenotypes, such as curved bodies, abnormal otoliths (ear stones), kidney cysts, hydrocephalus, and left-right patterning defects, during the embryonic development [24,32,36]. To determine whether the ten novel TTC genes had cilia-related functions, we repressed their protein expressions in zebrafish embryos with antisense MOs. Compared to the control fish, the ttc25 morphants exhibited severe curved body phenotype at 72 hpf (95.1% vs. 1.3%) as expected (Fig 2A). In contrast to the morphants of ttc5, -12, -16, -18, -29, and -39a, whose abnormality incidence was below 17%, more than 84% of the morphants of ttc4, -9c, -36, and -39c developed curved bodies [36,37] (Fig 2A and 2B). To validate the efficiency and specificity of the MOs against ttc4, -9c, -36, and -39c, both GFP-reporter assays and rescue experiments were carried out. In contrast to the control MO, all the ttc MOs silenced their corresponding GFP-reporters (S3 Fig). More importantly, the body curvature phenotypes of the morphants were rescued upon coinjection of the corresponding MO-resistant mRNA of zebrafish ttc4, -9c, -36, or -39c, but not of the mRNA of EGFP (Fig 2C and 2D). These results suggest that ttc4, -9c, -36, and -39c could be cilia-related genes. We thus performed further analyses on them. Zebrafish morphants of ttc4, -9c, -36, and -39c develop multiple ciliopathy-related phenotypes Next, we analyzed other ciliopathy-related phenotypes. There are two otoliths of unequal sizes in the inner ear of zebrafish (Fig 3A) [38]. Since proper otolith formation requires cilia motility in the inner ear, defects in the inner ear motile cilia often lead to abnormal otoliths, which are fused, mispositioned, or of wrong sizes or numbers [39]. We found that, similar to the ttc25 morphants, more than 75% of the morphants of ttc4, -9c, -36, and -39c developed abnormal otoliths at 72 hpf, whereas in the control morphants the incidence was only 0.4% (Fig 3A and  3B).
We found that more than 52% of the ttc (including ttc25) morphants at 72 hpf exhibited bulged head phenotype resembling hydrocephalus, compared to the value (2.7%) for the control animals (Fig 3C and 3D). To confirm that the phenotype was indeed due to hydrocephalus, we performed histological analyses on brain cross-sections of the affected ttc morphants and confirmed that such morphants indeed contained dilated ventricles of varying severities compared to control (Fig 3C).
The incidences of pronephric cysts were relatively low. Obvious pronephric cysts were observed in 17.0% of ttc25 and 18.3% of ttc4 morphants at 72 hpf, but none in the other three ttc morphants and control animals (Fig 3E and 3F).
Last, we examined the left-right patterning of the internal organs. When the heart tube was labeled through in situ hybridization of cardiac myosin light chain 2 (cmlc2) at 30 hpf [24,40], its leftward orientation was observed in 95.4% of control morphants (Fig 3G and 3H). By contrast, the ttc morphants exhibited varying extent of defects in the left-right asymmetry. More than 64% of the morphants of ttc25, -36, and -39c showed misoriented heart tubes toward either the right or the middle side, indicating a randomized heart tube orientation (Fig 3G and  3H). The morphants of ttc4 and ttc9c showed milder defects: 31.1% and 35.8% of them had misoriented heart tubes, respectively (Fig 3G and 3H).
ttc4, -9c, -36, and -39c are critical for cilia formation and motility We then investigated whether the ciliopathy-related defects of the zebrafish morphants were indeed attributed to cilia defects. Kupffer's vesicle (KV), a monociliated organ functionally equivalent to mouse embryonic node, is essential for establishment of the left-right asymmetry during early stages of embryonic development [32,41]. We thus examined KV cilia in embryos at the 7-somite stage and found that both the number and length of the KV cilia were significantly reduced in the ttc morphants (Fig 4A-4C). Furthermore, when the motile cilia in the pronephric duct were examined for zebrafish embryos at 24 hpf [24,32], we found that they were apparently disorganized in the ttc morphants as compared to the control ones (Fig 4D), suggesting defects in cilia motility. Quantification for the cilia number and length, however, was difficult due to intertwining of the cilia.
TTC4, -9c, -25, and -36 localize to multicilia in MTECs All these TTC genes are conserved in the evolution and their human proteins shared more than 50% sequence identities with their zebrafish homologs (S4 Fig). To further study functions of TTC25 and the four novel TTC proteins in mammals, we generated their polyclonal antibodies. Consistent with the expression profiles (Fig 1C and 1D), the levels of all these TTC proteins increased following ALI (Fig 5A). GFP-tagged Xenopus TTC25 is known to locate at both cilia base and axonemes [22]. Similar localizations were observed for endogenous TTC25 in multiciliated MTECs (Fig 5B). TTC4, TTC9c, and TTC36 were also markedly enriched in the ciliary axonemes in the MTECs (Fig 5B). Antigen competition experiments also confirmed that the immunofluorescent signals were specific (S5 Fig). As to TTC39c, we were not sure of its subcellular localization because the antibody did not work for immunostaining.

The TTC proteins associate with components of IFT complex or BBSome
In addition to the abundance of TTC proteins in the IFT complexes and BBSome [18][19][20], TTC26, its trypanosome ortholog PIFTC3, and Chlamydomonas ortholog DYF13 also associate with IFT-B [21,44]. Therefore, for further insights into the roles of the TTC proteins, we Identification of Ciliary TPR Containing Proteins investigated whether they could interact with the IFT complexes and BBSome as well. We mixed HEK293T cell lysates expressing FLAG-tagged TTC proteins with mouse testis lysates, which were abundant in proteins of IFT complex and BBSome (Fig 6A and 6B). Before the coimmunoprecipitation with anti-FLAG resin, the mixtures were pre-incubated for two hours to allow for the complex formation [45]. Immunoblotting indicated that TTC25 exhibited strong interaction with the tested components of both IFT-A and IFT-B complexes, suggesting an association with both complexes (Fig 6A). TTC36 mainly interacted with IFT-A complex, whereas TTC39c showed weak association with IFT144 and IFT139, two subunits of IFT-A. By contrast, although TTC4 and TTC9c hardly associated with the IFT complexes (Fig 6A), they interacted specifically with BBS7 ( Fig 6B). Reciprocal GST-pull down experiments further confirmed that the interactions between BBS7 and TTC4 or TTC9c were direct (S6 Fig).

Discussion
Our study provides a global view for the relationship between TTC genes and cilia functions. Proteomic analysis, cDNA microarray, and high throughput in situ hybridization have previously been used to identify cilia-related proteins or genes from isolated cilia, MTECs, or Xenopus or zebrafish embryos [22,[46][47][48][49]. FoxJ1, the master transcription factor of motile cilia [50,51], is shown to drive the expression of several TTC genes, including IFT88, IFT140, IFT172, and TTC26 [48], whereas TTC25 is found critical for multicilia formation in Xenopus embryonic skin cells [22]. We performed cDNA microarray using DAPT-treated MTECs and found that, in addition to ten known cilia-related TTC genes, ten novel TTC genes also exhibited correlated expression patterns during multiciliogenesis (Fig 1). Consistently, all the ten novel TTC genes, except TTC5 and TTC39c, have been found to be upregulated by at least two- fold in the previous cDNA microarray using FoxJ1-positive MTECs [46]. We characterized these ten novel genes using zebrafish embryos and found that ttc4, -9c, -36, and -39c are critical for motile cilia functions. The remaining six, ttc5, -12, -16, -18, -29, and -39a, were precluded from our further studies because their zebrafish morphants did not show striking body curvatures (Fig 2). Nevertheless, since we did not check the knockdown efficiency of the MOs, whether these genes are indeed unrelated to cilia still needs further clarification.
Our results indicate that, TTC4, -9c, -36, -39c, and -25 are critical for normal body shape, otolith formation, pronephric cilia function, and left-right asymmetry during zebrafish embryonic development (Figs 2, 3 and 4). They are also required for proper fluid flow of the brain ventricles (Fig 3) [32]. Kidney cysts are mainly attributed to overproliferation [52][53][54]. In zebrafish, defects in motile cilia can cause pronephric cysts but detailed mechanisms are still poorly understood [32,52,54]. Although the ttc morphants frequently exhibited enlarged pronephric ducts (S2-S6 Videa vs. S1 Video), only some morphants of ttc4 and ttc25 contained visible pronephric cysts (Fig 3). Nonetheless, there might exist tiny cysts that were not visible in our examinations. Detailed studies in the future are still required to clarify whether or not ttc9c, ttc36, and ttc39c are involved in pronephric cyst formation.
Although all these TTC proteins are important for KV cilia formation (Fig 4A-4C), they have different impact on the motility of pronephric cilia (Fig 4E and 4F and S2, S3, S4, S5 and S6 Videos). Compared to the control morphants, pronephric cilia motilities were strikingly repressed in the morphants of ttc9c, ttc36, and ttc39c. The average CBF was reduced by more than 81% (Fig 4E and 4F). By contrast, in the morphants of ttc25 and ttc4, the reduction in average CBF was less than 40% (Fig 4E and 4F). This suggests that ttc9c, ttc36, and ttc39c may have a role in the assembly or activity of axonemal dynein arms.
It is both surprising and interesting to find that all the cilia-related TTC proteins are implicated in IFT. Nine TTC proteins, including TTC26, have previously been shown as components of IFT-A or-B complex or BBSome [18][19][20][21] (Fig 1B). TTC25 is important for ciliogenesis and hedgehog signaling in Xenopus [22]. We further showed that mammalian TTC25 is strongly associated with both IFT-A and-B complexes but not BBSome (Fig 6). Therefore, TTC25 might function in bridging IFT-A and IFT-B during IFT. TTC36 appeared to associate more strongly with IFT-A complex than with IFT-B complex, whereas TTC39c, -4, and -9c associated with certain subcomplex or subunits of IFT-A or BBSome (Fig 6). Since TTC26/ DYF13 is involved in the transport of motility-related proteins into cilia/flagella through association with IFT-B [21], the association of TTC36 with IFT-A might also function in transport of certain cilia components. TTC36 (also named HBP21) and TTC4 have been proposed as cochaperons implicated in tumorigenesis [55,56] and might help protein folding as well. The punctate staining patterns of these TTC proteins (Fig 5B) are also consistent with their potential roles in IFT. Nevertheless, how these novel cilia-related TTC proteins function at the molecular level still needs future investigations.