Pontin (Ruvbl1) and Reptin (Ruvbl2) are closely related AAA ATPases. They are components of the Ruvbl1-Ruvbl2-Tah1-Pih1 (R2TP) complexes that function as co-chaperones for the assembly of multiple macromolecular protein complexes. Here, we show that Pontin is essential for cilia motility in both zebrafish and mouse and that Pontin and Reptin function cooperatively in this process. Zebrafish pontin mutants display phenotypes tightly associated with cilia defects, and cilia motility is lost in a number of ciliated tissues along with a reduction in the number of outer and inner dynein arms. Pontin protein is enriched in cytosolic puncta in ciliated cells in zebrafish embryos. In mouse testis, Pontin is essential for the stabilization of axonemal dynein intermediate chain 1 (DNAI1) and DNAI2, the first appreciated step in axonemal dynein arm assembly. Strikingly, multiple dynein arm assembly factors show structural similarities to either Tah1 or Pih1, the other two components of the R2TP complex. Based on these results, we propose that Pontin and Reptin function to facilitate dynein arm assembly in cytosolic foci enriched with R2TP-like complexes.
The cilium is a microtubule-based structure that has been classified into two major categories: motile and primary cilia. Motile cilia beat to generate fluid flow across an epithelial surface or propel cell movement. Defective movement of motile cilia leads to primary (genetic) ciliary dyskinesia (PCD), a group of rare human genetic diseases characterized by recurrent infections of the respiratory system, male infertility and frequent laterality defects (for a recent review, see Horani et al., 2016). More than 30 PCD genes have been identified and together they account for ∼60-70% of PCD cases (Horani et al., 2016; Olcese et al., 2017; Paff et al., 2017), suggesting that a large number of PCD genes remain to be identified.
In a previous study, we found that a zebrafish loss-of-function mutant of reptin (ruvbl2, or reptin52//ino80j/tip48), which encodes an AAA family ATPase, displays phenotypes that can be attributed to cilia motility defects, including ventral body curvature and kidney cyst (Zhao et al., 2013). In the present study, we found that Pontin (Ruvbl1, or Pontin52/Ino80h/Tip49), a protein closely related to Reptin, is also essential for cilia motility in zebrafish and mouse.
Reptin and Pontin interact with each other and numerous other partners to regulate a host of cellular processes, including transcription (Shen et al., 2000; Jin et al., 2005; Wu et al., 2005; Doyon et al., 2004; Gstaiger, 2003). They may antagonize each other in these processes (Bauer, 2000; Diop et al., 2008; Rottbauer et al., 2002). Pontin and Reptin are also key components of the R2TP (Ruvbl1-Ruvbl2-Tah1-Pih1) complex, which functions as a co-chaperone for the assembly of multiple large protein complexes (Kakihara and Houry, 2012; Te et al., 2007; Zhao et al., 2005). In this study, we show that loss-of-function mutants of pontin display cilia motility defects and phenocopy loss-of-function mutants of reptin in zebrafish. In addition, pontin and reptin double mutants display similar dysfunctional cilia phenotypes as single mutants. Furthermore, the stability of Pontin and Reptin proteins is interdependent in vivo in zebrafish. Finally, by generating a conditional knockout mouse model of Pontin, we show that its role in cilia motility is conserved in mammals. Together, these results suggest that Pontin and Reptin function together, as opposed to antagonistically, in cilia motility in vertebrates.
To elucidate a potential mechanism for the cilia motility defects in pontin mutants, we examined for the presence of ciliary dynein arms. Cilia motility is powered by inner and outer dynein arms (IDAs and ODAs, respectively), large motor protein complexes anchored on the A-tubule of each of the nine peripheral microtubule doublets that comprise the ciliary axoneme. In this study, we found that the number of both ODAs and IDAs is significantly reduced in motile cilia of pontin mutants, similar to reptin mutants as shown in our previous study (Zhao et al., 2013). Moreover, Pontin is enriched in cytoplasmic puncta, granular structures that are similar to Reptin and Lrrc6 (Seahorse) puncta that we previously observed (Zhao et al., 2013). Furthermore, in the mouse testis, Pontin is required for the stability of both DNAI1 and DNAI2 and Reptin physically interacts with Hsp90. Based on these results, we propose that Pontin and Reptin function cooperatively to facilitate axonemal dynein arm assembly in cytoplasmic foci containing R2TP-like complexes.
Loss of function of Pontin leads to cilia-associated phenotypes in zebrafish
hi1055B/hi1055B was identified in a large-scale insertional mutagenesis screen in zebrafish together with multiple mutants of genes essential for cilia morphogenesis, such as ift172 and arl13b (Sun, 2004). Similar to these cilia morphogenesis mutants, hi1055B/hi1055B mutant embryos show ventral body curvature and kidney cyst formation (Fig. 1A). Transverse sections verified the presence of cysts in the glomerular neck region (Fig. 1B). As cilia also play an essential role in the establishment of the left-right (LR) axis of the vertebrate body plan, we inspected for lateral heart positioning in mutant embryos. However, no significant defect was detected (not shown). The lack of LR defects in hi1055B/hi1055B mutants is consistent with other known cilia mutants and is likely to be due to masking by maternal contribution of gene products.
The proviral insertion in hi1055B is located in the first intron of the pontin (ruvbl1) gene. RT-PCR analysis using oligos against cDNA regions both 5′ and 3′ to the proviral insertion shows that the expression of pontin is abolished in hi1055B/hi1055B mutants (Fig. 1C). Moreover, microinjection of in vitro synthesized pontin-eGFP mRNA into embryos from hi1055B/+ crosses reduced the frequency of embryos showing a mutant phenotype to an average of 2.4% with body curvature and 3.6% with kidney cysts, as compared with 27.1% and 25.6%, respectively, in embryos injected with eGFP mRNA (n=3 independent experiments; Fig. 1D,E), indicating a significant rescue. Together, these results demonstrate that hi1055B is a loss-of-function allele of pontin.
pontin and reptin double mutants show similar cilia-associated phenotypes to single mutants in zebrafish
In a previous study, liebeskummer (lik) was identified as a gain-of-function mutant of reptin (ruvbl2) (Rottbauer et al., 2002). Similar to pontinhi1055B/hi1055B mutants, lik mutants show a prominent ventral body curvature phenotype, raising the possibility that Pontin and Reptin functionally antagonize each other. To clarify the relationship between Pontin and Reptin in cilia-associated phenotypes, such as ventral body curvature and kidney cyst, we generated double mutants homozygous for hi1055B and hi2394 (Zhao et al., 2013), loss-of-function alleles of pontin and reptin, respectively. Whereas double-heterozygous carriers display no obvious phenotypes, double mutants show identical ventral body curvature and kidney cyst phenotypes as in either pontinhi1055B/hi1055B or reptinhi2394/hi2394 single mutants (Fig. 2A), suggesting that Pontin and Reptin do not antagonize the function of each other in these processes.
Pontin and Reptin require each other for protein stability
In tissue culture cells, it was shown that Pontin and Reptin are reliant on each other for protein stability (Venteicher et al., 2008). To investigate whether a similar interdependency exists in vivo in zebrafish embryos, we analyzed the level of Pontin protein in reptin mutants and vice versa. At 4 days post fertilization (dpf), reptinhi2394/hi2394 mutants were sorted from control siblings based on their body curvature phenotype. Immunoprecipitation (IP) followed by western blotting on total embryo lysates showed that, in reptinhi2394/hi2394 mutants, not only is the level of Reptin abolished, but Pontin is greatly reduced as well (Fig. 2B). Similarly, in pontinhi1055B/hi1055B mutants, the levels of both Pontin and Reptin are drastically reduced (Fig. 2B), suggesting that the two interacting proteins require each other for stability.
Pontin is enriched in cytoplasmic puncta in zebrafish ciliated tissues
To investigate the function of pontin in zebrafish development, we first analyzed its expression pattern by in situ hybridization. pontin transcript can be readily detected at the 4-cell stage (Fig. S1Aa). As zygotic transcription starts at the 500- to 1000-cell stage, this suggests that pontin mRNA is maternally deposited. At the bud stage, pontin is ubiquitously expressed (Fig. S1Ab). At 1 dpf, pontin mRNA is still widely distributed with some enrichment in the neural tube and pronephric tubule (Fig. S1Ac), which are regions enriched with cilia. The specificity of the signal is confirmed by its absence in the sense control (Fig. S1B). This expression pattern is similar to that of reptin, as shown in our previous study (Zhao et al., 2013).
To analyze the subcellular localization pattern of Pontin in zebrafish, we raised a polyclonal antibody against a recombinant protein spanning 80 amino acids of the N-terminus. Immunostaining of 2 dpf embryos showed that Pontin is enriched in cytoplasmic puncta in a subset of pronephric tubule cells (Fig. 3A), similar to the localization pattern of Reptin revealed in our previous study (Zhao et al., 2013). The specificity of the antibody was confirmed by the significantly reduced signal in pontinhi1055B/hi1055B mutants (Fig. 3A).
We additionally examined the localization of Pontin in multiple extra-renal tissues that have cilia. Similar cytoplasmic puncta were detected in Kupffer's vesicle, the neural tube (Fig. S2A), the olfactory placode and the lateral line organ (Fig. 3B). These findings support the proposal that Pontin is enriched in multiple ciliated tissues across development.
The punctate localization pattern of Pontin prompted us to investigate whether it localizes to or near basal bodies. However, colabeling of the pronephric tubule with anti-γ-tubulin, a basal body marker, revealed that Pontin puncta do not overlap with basal bodies (Fig. S2B).
Pontin is dispensable for cilia morphogenesis but essential for cilia motility in zebrafish
As pontinhi1055B/hi1055B zebrafish show morphological phenotypes similar to those observed in cilia biogenesis mutants, we analyzed cilia formation in the pontin mutants. Immunofluorescence staining using anti-acetylated tubulin as a cilia marker revealed that abundant cilia are able to form in single-ciliated and multiciliated cells in the pronephric tubule, the olfactory placode and the lateral line organ (Fig. 3C). Live Nomarski imaging also revealed abundant cilia in the olfactory placode (Fig. 3D). Together, these results suggest that cilia morphogenesis can proceed in pontinhi1055B/hi1055B mutants, similar to reptinhi2394/hi2394 mutants (Zhao et al., 2013).
As cilia motility defects can lead to similar kidney cyst and ventral body curvature phenotypes in zebrafish, we measured the movement of motile cilia in the pronephric tubule at 3 dpf using high-speed Nomarski video microscopy and kymographic analysis. In control siblings, cilia in multiciliated cells in the anterior portion of the pronephric tubule beat vigorously, display an oscillatory motion and are neatly bundled within the lumen of the tubule (Fig. 4A, Movie 1). Strikingly, cilia in the same region of pontinhi1055B/hi1055B mutants are paralyzed, fail to display an oscillatory beating patterning and are randomly oriented (Fig. 4A, Movie 2). We further investigated cilia motility in the olfactory placode, which is lined with motile cilia in zebrafish. Whereas these cilia beat in a coordinated fashion at 35 Hz in control siblings, they are paralyzed in pontinhi1055B/hi1055B mutants (Fig. 4B, Movies 3 and 4). Together, these results demonstrate that Pontin is essential for cilia motility in multiple organs, similar to Reptin (Zhao et al., 2013).
Both ODAs and IDAs are reduced in zebrafish pontinhi1055B/hi1055B mutants
To investigate the molecular mechanism underlying the observed cilia motility defects in pontinhi1055B/hi1055B mutants, we analyzed the ultrastructure of cilia in the pronephric tubule. 5 dpf embryos were fixed and subjected to transmission electron microscopy. In transverse sections of wild-type control cilia, we frequently observed both ODAs and IDAs on the A-tubule of peripheral microtubule doublets of the ciliary axoneme (Fig. 4C). By contrast, both IDAs and ODAs appeared reduced in mutant cilia (Fig. 4C). To quantify the difference, we counted ODAs and IDAs on 25 transverse sections of cilia from control and mutant embryos. Statistical analysis demonstrates that the number of ODAs, IDAs and total dynein arms were significantly reduced in pontinhi1055B/hi1055B mutants (Fig. 4D), similar to our previous observation in reptinhi2394/hi2394 mutants (Zhao et al., 2013).
Pontinflox/flox; Stra8-Cre male mice are sterile
We next asked whether the function of Pontin in cilia motility is conserved in mice. Consistent with the multitude of functions of Pontin, Pontin (Ruvbl1) null mice were reported to be lethal at an early embryonic stage (Bereshchenko et al., 2012). Since the sperm tail is a specialized motile cilium and male infertility is a major component of PCD, we generated Pontinflox/flox; Stra8-Cre mice to specifically delete Pontin in premeiotic male germ cells. Specifically, we acquired Pontintm1a/+ mice generated by the European Conditional Mouse Mutagenesis Program (EUCOMM). In the tm1a allele, exon 3 of Pontin is flanked by two LoxP sites and an Frt-lacZ-neo-Frt selection cassette is inserted into intron 2 (Fig. 5A). Pontintm1a/+ mice were crossed with actin-FLPe mice to generate Pontinflox/+ mice, which were further crossed with Stra8-Cre mice. Stra8-Cre directs recombination in male germ cells as early as postnatal day (P) 3 (Sadate-Ngatchou et al., 2008).
Male Pontinflox/flox; Stra8-Cre mice are viable with no overt phenotypes. To test the fertility of mutant males, we paired individual male mice of 2 months of age with four wild-type female mice for an experimental duration of 3.5 months. Whereas on average each control sibling male sired 80±8.33 (s.d.) pups in 11±1 litters during the test period, none was produced by mutant males (three male mice of each genotype were analyzed). Notably, copulatory plugs were detected in females housed with mutant males, suggesting normal mating behavior by mutant male mice. However, plugged females paired with Pontinflox/flox; Stra8-Cre males never showed signs of pregnancy and failed to produce any progeny. Together, these results show that Pontin may be essential for male fertility in mice.
Pontin is essential for mouse sperm motility
To investigate the cause of male infertility in Pontinflox/flox; Stra8-Cre mice, we analyzed testis morphology and sperm motility. At 5.5 months of age, the testis of Pontinflox/flox; Stra8-Cre mice is significantly smaller and the testis/body weight ratio is reduced relative to control siblings (Fig. S3A,B). Histological analysis of transverse sections shows that in mutant testis the seminiferous tubule structure is severely disrupted and only residual spermatogonial cells close to the basement membrane remain (Fig. S3C). To analyze the impact of Pontin on sperm motility specifically, we focused our studies on 12-week-old mice. At this stage, the testis is smaller and the testis/body weight ratio is significantly reduced in mutants compared with sibling controls (Fig. 5B,C). However, on histological transverse sections, the structure of the seminiferous tubule appears relatively normal with abundant mature spermatids in the center (Fig. 5D). The size of the caudal epididymis is also normal in mutant mice compared with sibling controls (Fig. 6A).
We then isolated spermatozoa from the caudal epididymis to investigate the impact of Pontin deletion on the morphology and motility of sperm flagella. Sperm cells from Pontinflox/flox; Stra8-Cre males exhibit prominent flagella that appear morphologically normal by transmitted light microscopy (Fig. 6C). However, flagella motility defects were readily detected by high-speed video microscopy. Whereas 86.44% of sperm cells from wild-type mice swim vigorously (Fig. 6B,C, Movie 5), most sperm cells from mutant mice display complete immotility, with only 20.65% retaining weak motility (Fig. 6B,C, Movie 6).
Pontin is essential for DNAI1 and DNAI2 stability
To investigate the cause of the immotile sperm phenotype in Pontinflox/flox; Stra8-Cre mice, we examined the distribution of outer dynein arm components in sperm cells by immunofluorescence staining. This analysis showed that the signal of two specific ODA components, DNAI2 and DNAH9, is greatly reduced in the mutant sperm tail (Fig. 7A), thereby suggesting that Pontin plays an essential role in ODA assembly.
An early step during ODA assembly is complex formation between DNAI1 (DNAIC1) and DNAI2 (DNAIC2), which stabilizes both proteins (Fowkes and Mitchell, 1998). We analyzed the protein level of DNAI1 and DNAI2 in 12-week-old Pontinflox/flox; Stra8-Cre mutant mice by western blot analysis of total testis lysate. Strikingly, the level of both DNAI1 and DNAI2 was drastically reduced in mutant testes in comparison to wild-type siblings (Fig. 7B). Quantification of western blot data by densitometric analysis revealed that the level of DNAI1 and DNAI2 in mutant testes is reduced to 11.0±2.9% (P<0.05) and 3.6±2.1% (P<0.05), respectively, of that in sibling controls, suggesting that Pontin is required for stability of the two proteins and thus ODA assembly.
Further, the level of Reptin in Pontinflox/flox; Stra8-Cre mutant testis is also reduced to 8.0±0.5% (P<0.005) of that of control siblings (Fig. 7B), suggesting that the requirement of Pontin for Reptin stability is conserved in the mouse testis.
Reptin interacts with Hsp90 in the mouse testis
The stability of DNAI1 and DNAI2 depends on complex formation between the two proteins (Fowkes and Mitchell, 1998). Interestingly, Pontin and Reptin are known to form the R2TP complex with Hsp90 and they collectively function as a chaperone required for the maturation and assembly of numerous multisubunit complexes (Kakihara and Houry, 2012; Te et al., 2007; Zhao et al., 2005, 2008; Hořejší et al., 2010). To investigate whether Pontin and Reptin function similarly as a co-chaperone for the assembly of the DNAI1-DNAI2 complex, we first sought to address whether the Pontin-Reptin-Hsp90 complex exists in mouse testis. We performed co-immunoprecipitation (co-IP) using Reptin as the bait protein due to the robust performance of the anti-Reptin antibody in IP assays. Western blotting of IP samples using the anti-Reptin antibody showed that both Reptin and Hsp90 are pulled down from the testis lysate of 12-week-old wild-type mice. By contrast, both failed to be pulled down in Pontinflox/flox; Stra8-Cre mutant testis lysate (Fig. 7B). Together, these findings support the proposal that Pontin-Reptin interact with Hsp90 and function as co-chaperones for assembly of the DNAI1-DNA12 complex, the first step in axonemal dynein arm assembly.
Although numerous and diverse molecular functions have been described for Pontin and Reptin, their function in animal development remains poorly characterized. Biochemically, these two AAA-ATPases form single- or double-hexameric rings (Puri et al., 2007; Gribun et al., 2008; Cheung et al., 2010a,b; Torreira et al., 2008). Reptin was found to inhibit the transcriptional activity of β-catenin, whereas Pontin stimulates it (Bauer, 2000; Rottbauer et al., 2002). It is thought that this counterbalancing act of Pontin and Reptin, whereby one antagonizes the function of the other, regulates heart growth during zebrafish development (Rottbauer et al., 2002). However, in this study, we found that Pontin and Reptin interact and function cooperatively to establish cilia motility in both zebrafish and mice. Zebrafish loss-of-function reptin and pontin mutants show almost identical cilia-associated morphological phenotypes, including ventral body curvature and kidney cyst formation. On a cellular and molecular level, both Pontin and Reptin are required for cilia motility through the assembly of ciliary ODAs and IDAs. Moreover, double mutants for pontin and reptin show similar phenotypes to single mutants, as opposed to ameliorated or exacerbated phenotypes. Further, we found that Pontin and Reptin share a unique interdependence on each other for protein stability in vivo in zebrafish embryos and mouse testes, which provides a molecular mechanism for the nearly identical phenotypes observed in these mutants.
Our study shows that Pontin is required for cilia motility in multiple tissues in zebrafish, similar to Reptin (Zhao et al., 2013). By generating a conditional mouse knockout model, we additionally show that Pontin is essential for male fertility and more specifically sperm motility in mice, suggesting that its function in cilia motility is conserved in mammals.
In search of a mechanism for the cilia motility defects in pontin or reptin mutants, we detected a significant reduction in the density of ODAs and IDAs in both mutants. Axonemal dynein arms are large intricate machineries that power cilia motility. Studies performed in the green alga Chlamydomonas suggest that dynein arm components are folded and preassembled into discrete subunits in the cytosol (Fowkes and Mitchell, 1998; Fok et al., 1994) by dynein arm assembly factors (DNAAFs) before being transported into cilia/flagellum. During early stages of ODA assembly, DNAI1 and DNAI2 form a complex, which is essential for the stability of both proteins. Heavy chain subunits are attached to the DNAI1-DNAI2 complex in subsequent steps (Fowkes and Mitchell, 1998). In recent years, human genetics and model organisms have provided crucial insights into the complex process of axonemal dynein arm assembly. Multiple PCD genes, namely DNAAF1/LRRC50 (Duquesnoy et al., 2009; Loges et al., 2009), DNAAF2/KTU (Omran et al., 2008), DNAAF3/PF22 (Mitchison et al., 2012), DNAAF4/DYX1C1 (Tarkar et al., 2013), DNAAF5/HEATR2 (Horani et al., 2012; Diggle et al., 2014), LRRC6 (Kott et al., 2012; Horani et al., 2013), ZMYND10 (Moore et al., 2013; Zariwala et al., 2013), SPAG1 (Knowles et al., 2013), C21ORF59 (Austin-Tse et al., 2013) and PIH1D3 (Paff et al., 2017; Olcese et al., 2017), have been shown to be involved in the cytoplasmic assembly of dynein arm subunits and mutations in these factors lead to loss of both ODAs and IDAs. Notably, DNAAF2 and DNAAF4, respectively, show structural similarities to Pih1 and Tah1, two components of the R2TP complex (Omran et al., 2008; Tarkar et al., 2013). The R2TP complex is known to interact with Hsp90 and functions as a co-chaperone for the assembly of a growing number of macromolecular complexes (Kakihara and Houry, 2012; Te et al., 2007; Zhao et al., 2005, 2008; Hořejší et al., 2010). Here, we show that Reptin interacts with Hsp90 in the mouse testis. Combined, these results support a model in which Reptin and Pontin function as co-chaperones in R2TP-like complexes to facilitate axonemal dynein arm assembly in the cytosol (Fig. 7C).
In this study, we found that Pontin is enriched in cytoplasmic foci – structures very similar to the Reptin and Lrrc6/Reptin puncta observed in our previous study (Zhao et al., 2013) – raising the intriguing possibility of the existence of novel dynein arm assembly foci in the cytosol.
Mutations in PONTIN (RUVBL1) or REPTIN (RUVBL2) have not been associated with PCD in humans to date. As Pontin and Reptin are involved in a large number of essential cellular processes, it is possible that mutations in these two genes are too deleterious to recover in human populations. Nonetheless, interactors of Pontin and Reptin could serve as candidate proteins for PCD. One such example is Lrrc6, a DNAAF found to directly interact with Reptin and regulate cilia motility in zebrafish in our previous study (Zhao et al., 2013). Many more Pontin/Reptin interactors have been identified (Kim et al., 2007; Izumi et al., 2010; Ikura et al., 2000; Ewing et al., 2007), providing a rich source of candidate proteins involved in cilia motility. It might also be interesting to perform additional assays to uncover Pontin/Reptin binding partners specifically in cells with motile cilia.
In addition to their role in cilia motility, Pontin and Reptin are known to function in multiple signaling pathways. Pontin and Reptin interact with all six mammalian phosphatidylinositol-3 kinase-related protein kinases, including ATM, ATR, PKA-PKCs, mTOR, SMG1 and TRRAP (Izumi et al., 2010). Further, Pontin and Reptin have been linked to Wnt signaling (Bauer, 2000), the repair of DNA double-strand breaks (Raymond et al., 2015) and the activity of the mTOR complex (David-Morrison et al., 2016). Similar to multiple genes involved in cilia morphogenesis or function, the transcription of Reptin and Pontin is upregulated after deflagellation in the green alga Chlamydomonas (Marshall, 2004). Thus, in addition to their direct involvement in axonemal dynein arm assembly, Pontin-Reptin could mediate cellular responses to ciliary defects through secondary responses. Aberrant DNA damage response, mTOR and Wnt signaling have been observed in several ciliary mutants or ciliopathies (Simons et al., 2005; Shillingford et al., 2006; Kishimoto et al., 2008; Gerdes et al., 2007; Dibella et al., 2009; Corbit et al., 2008; Choi et al., 2013; Chaki et al., 2012; Boehlke et al., 2010). Future studies may address whether Pontin and Reptin mediate crosstalk between these pathways and their potential association with defects observed in ciliopathies.
In our study, we detected Pontin protein in the cytoplasm of tissues with primarily motile cilia. However, we also found Pontin protein in the cytoplasm of the zebrafish lateral line organ, which contains only immotile cilia in the form of kinocilia. Notably, the zebrafish lateral line kinocilium could potentially harbor dynein arms despite being immotile. In the teleost burbot, the kinocilium of the lateral line is also immotile but displays ultrastructural features that have historically defined motile cilia: a ʻ9+2' axonemal configuration and prominent dynein arms on the outer microtubule doublets (Flock and Duvall, 1965). Currently, the existence and distribution of immotile primary cilia in early zebrafish embryos are incompletely characterized and the potential function of Pontin in primary cilia remains an open question.
MATERIALS AND METHODS
Animal care ethics
Zebrafish and mouse experiments were conducted in accordance with the guidelines of Yale University Institutional Animal Care and Use Committee. Zebrafish were of the Tu/AB background. Pontintm1a/+ from EUCOMM (Bradley et al., 2012) and Stra8-Cre mice (Sadate-Ngatchou et al., 2008) were on the C57BL/6J background, and FLPe mice (Rodriguez et al., 2000) on the 129/C57BL/6J background.
Zebrafish were maintained according to standard protocols (Westerfield, 2000). Embryos were obtained through natural spawning.
Zebrafish embryos were fixed in Bouin's fixative (EM Science, L7831) overnight at room temperature, washed five times with PBS containing 0.1% Tween 20 (PBT), and embedded using a JB-4 Kit (Polysciences) following the manufacturer's instructions. Blocks were cut into 4 µm sections and stained with Hematoxylin and Eosin (H&E). Mouse tissue was fixed by heart perfusion with 4% paraformaldehyde (PFA) followed by immersion in 4% PFA at 4°C overnight. Samples were washed three times with PBT, embedded with OCT (Tissue-Tek, 4583) and blocks were cut into 5 µm sections and stained with H&E. Sections were visualized under a Nikon Eclipse E800 microscope driven by the NIS-Elements software.
Microinjection and rescue assay
Microinjection in zebrafish embryos was performed as described (Yuan and Sun, 2009). Capped mRNA was transcribed in vitro using the mMESSAGE mMACHINE Kit (Ambion, SP6 AM1340) following the manufacturer's instructions. Synthesized mRNA was injected into zebrafish embryos at the 1- to 2-cell stage. Injected embryos were examined under a dissecting scope for morphological phenotypes and scored at 3 dpf.
Zebrafish embryos were individually lysed in 20 µl 50 mM NaOH at 95°C for 20 min. Samples were then neutralized with 4 µl 1 M Tris-HCl and subjected to PCR analysis. Primers for genotyping hi2394 have been published (Zhao et al., 2013). Primers for hi1055B are listed in Table S1.
For mouse, 1-2 mm of the tail tip was cut from mice younger than P21 and similarly lysed in 100 µl 50 mM NaOH and neutralized with 20 µl 1 M Tris-HCl. After clearance by centrifugation at 12,000 g for 5 min, 0.5 µl lysate was used in 10 µl PCRs for genotyping using the primers listed in Table S1.
Zebrafish embryos at different stages were fixed in Dent's fixative (80% methanol/20% DMSO), except for phalloidin-stained samples, which were fixed in a 1:2.7 dilution of formalin in PBT. Whole-mount immunostaining was performed as previously described (Drummond et al., 1998; Dent et al., 1989). Stained samples were flat mounted using Fluoro-Gel with DABCO (Electron Microscopy Sciences, 17985-01). Images were taken under a Nikon Eclipse E800 microscope driven by NIS-Elements software. For immunostaining, sperm was collected from the caudal epididymis of 12-week-old mice, fixed by incubation with 4% PFA for 1 h at 4°C, transferred to slides, dried and stored at −80°C. After rehydration, samples were permeabilized by incubation with methanol at −20°C for 10 min. Slides were then incubated sequentially with 10% fetal bovine serum for 1 h at room temperature, primary antibody at 4°C overnight and secondary antibody for 1 h at room temperature.
Total RNA was extracted from zebrafish embryos at 5 dpf using Trizol Reagent (Ambion, 15596018) according to the manufacturer's instructions. cDNA was synthesized using SuperScript II Reverse Transcriptase (Invitrogen, 100004925) according to the manufacturer's instructions and used as template for PCR using Taq DNA polymerase (Invitrogen, 100021276) on a Bio-Rad C1000 Touch thermal cycler with the primers listed in Table S1.
Whole-mount in situ hybridization
Whole-mount in situ hybridization was performed as described (Duldulao et al., 2009; Hauptmann and Gerster, 2000). Briefly, zebrafish embryos were fixed in a fresh 1:2.7 dilution of formalin in PBT at 4°C overnight. After prehybridization at 68°C for 1 h, samples were hybridized with DIG-UTP-labeled probes at 68°C overnight, followed by incubation with alkaline phosphatase-coupled anti-DIG antibody (Roche, 11 093 274 910; 1:3000). BM Purple (Roche, 11442074001), a precipitating substrate of alkaline phosphatase, was used to develop color.
Zebrafish pontinhi1055B mutants and wild-type siblings at 5 dpf were fixed in Karnovsky fixative, washed and post-fixed with Palade's osmium. They were then in-block stained with Kellenburger's uranyl acetate stain (28.6 mM sodium acetate trihydrate, 39.2 mM sodium barbiturate, 28 mM HCl, pH adjusted to 6.0 with 0.1 M HCl prior to adding uranyl acetate to 12.8 mM) at room temperature for 1 h, dehydrated through a 70% to 100% ethanol series and infiltrated with propylene oxide. Samples were incubated in 1:1 propylene oxide/Embed 812 resin (Electron Microscopy Sciences, EMS#14120) for a minimum of 2 h to overnight, followed by two changes of 100% Embed 812 for 1 h each and then embedded in Embed 812 overnight at 60°C. Sections were cut at 70 nm thickness on a Leica Ultracut UTC and stained with uranyl acetate and Reynold's lead citrate. Samples were observed on a Tecnai Biotwin electron microscope.
Protein extraction, immunoprecipitation, and western blot
Zebrafish embryos at 4 dpf were deyolked and homogenized in Tris lysis buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% NP40 and freshly added 0.1 mM DTT) with a cocktail of protease inhibitors (Roche, 11836170001) and phosphatase inhibitors (Roche, 04906837001). Testis from 12-week-old mice was homogenized with the same buffer. Lysates were cleared by centrifugation at 12,000 g at 4°C for 5 min. For IP, lysates were incubated with respective primary antibodies at 4°C overnight, followed by incubation with A/G agarose (Pierce, 20421) at 4°C for 4 h. Samples were subjected to SDS-PAGE and transferred to PVDF membrane. Membranes were incubated with primary antibodies followed by HRP-conjugated secondary antibodies (Jackson ImmunoResearch; 1:5000 in 5% skimmed milk). Signal was detected using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, 34080). Western blot signals from two groups of mouse testis were quantified using Fiji Integrated Density software.
High-speed video microscopy of zebrafish cilia and mouse sperm
Mounting of live zebrafish embryos and video microscopy of motile cilia were performed according to a published protocol (Zhao et al., 2013; Yuan et al., 2012). Cilia beating dynamics were analyzed via kymograph using Fiji software. For mouse sperm, the caudal epididymis of adult wild type or Pontin mutant was immersed in 50 µl gCPA (18% D+ raffinose pentahydrate, 3% dehydrated skimmed milk, 100 mM L-glutamine dissolved in cell culture grade water), minced and incubated at 37°C for 5 min to extract sperm. Then, 2 µl samples were incubated with 120 µl Research Vitro Fertilization Medium (COOK Medical, K-RVFE-50) at 37°C for 5 min. Sperm motility was monitored under a Zeiss Observer Z1 inverted microscope with an X-Cite 120 LED illumination source and recorded by a Hamamatsu ORCA-Flash4.0 sCMOS camera driven by ZEN Pro software (Zeiss) at 1.004 ms exposure and 2048×2048 pixel resolution. Videos were processed with Fiji. Movement of individual sperm was traced with the built-in tracking function of Fiji.
Antibodies and fluorescent dyes
The following commercial antibodies were used: rabbit anti-Reptin (Abcam, ab91462, lot GR24126-5; 1:2000 in western blot), mouse anti-β-tubulin (Sigma, T4026; 1:8000 in western blot), rabbit anti-Hsp90 (Santa Cruz, sc-7947; 1:3000 in western blot), rabbit anti-DNAI1 (Sigma, HPA021843, lot R09108; 1:1000 in western blot), mouse anti-DNAI2 (Abnova, H00064446-M01, lot GB071-1C8; 1:500 in western blot, 1:100 in immunostaining), rabbit anti-GAPDH (GeneTex, GTX100118, lot 39408; 1:10,000 in western blot), mouse monoclonal anti-acetylated tubulin antibody (Sigma-Aldrich, clone 6-11B-1; 1:5000 in immunostaining), mouse anti-DNAH9 (Abcam, clone 440.4; 1:100 in immunostaining) and mouse anti-γ-tubulin (Sigma, clone GTU-88; 1:1000 in immunostaining).
Chicken anti-cadherin 17 custom antibodies were described previously (Zhao et al., 2013; Kishimoto et al., 2008). For rabbit anti-Pontin antibody, amino acids 1-80 of zebrafish Pontin fused with GST was produced in E. coli and purified using the pGEX system (GE Healthcare Life Sciences) according to the manufacturer's instructions. Rabbit polyclonal serum against the recombinant protein was produced by Covance. IgG was purified using CM Affi-Gel Blue Gel (Bio-Rad, 153-7304) following the manufacturer's protocol. Purified anti-Pontin IgG was used at 1:1000 in immunostaining and 1:2000 in western blotting. Rhodamine-phalloidin (Invitrogen, R415; 1:1000) was used to label F-actin in formalin-fixed samples, and DAPI (10 µg/ml) was used to label nuclei.
Student's t-test (two-tailed, paired) was performed in Microsoft Excel.
We thank N. Semanchik for zebrafish husbandry, D. Shao and S. Mentone for histology assistance, E. Cheng in the H. Lin lab for providing the Stra8-Cre mouse, and X. Tian in the S. Somlo lab for providing the FLPe mouse.
Conceptualization: Z.S.; Methodology: Y.L., S.Y.; Investigation: Y.L., L.Z., S.Y., J.Z.; Writing - original draft: Y.L., Z.S.; Writing - review & editing: S.Y., Z.S.; Supervision: Z.S.; Project administration: Z.S.
This work was supported by research grant 191G14a from the PKD Foundation and by National Institutes of Health grants R01 DK092808, R01 HL125885 (to Z.S.) and K99 HD086274 (to S.Y.). Deposited in PMC for release after 12 months.
The authors declare no competing or financial interests.