AB/Tübingen (AB/Tue) zebrafish were maintained at 28°C on 14/10 h light/dark cycle. Embryos were raised at 28°C, collected and maintained in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl, 0.33 mM MgSO4, 10^–5% Methylene Blue). All used procedures were approved by the ethical committees for animal experimentation of the Consejo Superior de Investigaciones Científicas (CSIC) and Comunidad Autónoma de Madrid.

Fertilized chick embryos (Santa Isabel Farm, Cordoba, Spain) were incubated at 38°C in a humidified incubator until the desired stage, determined according to Hamburger and Hamilton (1992).

Total mRNA from AB/Tue zebrafish embryos was extracted using RNeasy Mini kit (Qiagen) according to manufacturer instructions. cDNA was synthesized using Super Script kit (Roche) following manufacturer instructions. PCR products, obtained from cDNA amplification using specific primers (Supplementary Table 1), were cloned in PCSA plasmid (Agilent Technologies), as described by the manufacturer. Plasmid DNA preparations were obtained using Genopure Plasmid Midi kit (Roche) following kit instructions. Digoxigenin-UTP-labeled antisense probes for in situ hybridization (ISH) were synthesized and purified using Super Script kit (Roche) following the manufacturer instructions. ISH was performed by standard procedures and visualized with NBT/BCIP (dark blue).

The PCSA-mosmoa_p1 plasmid was used as a template to amplify by PCR mosmoa and further add an hemagglutinin tag (HA) and restriction sites with the following primers: Fw 5′-aatCTCGAGCCTGAGATGGATAAACTC-3′. Rv 5′-ttaGAA TTCTCAAGCGTAATCTGGAACATCGTATGGGTAGCCAGG AAGACACACTTC-3′. The PCR product was cloned in PCSA plasmid (Agilent Technologies) as described by the manufacturer. The mosmoa-HA fragment was then excised with restriction enzymes and cloned in the pCIG vector (Megason and McMahon, 2002) for chick embryo electroporation and in pCS2 for cell transfection and synthesis of mRNA to be injected in zebrafish embryos.

The pCIG Mosmoa-HA plasmid (1 μg/μl) was co-injected with a pCAG-2A-Arl13b-tRFP [Arl13b-tRFP construct generated by Schmitz et al. (2017)] (1 μg/μl) into the neural tube ventricle of HH10 chick embryos followed by in ovo electroporation as previously described (Cardozo et al., 2014).

Human embryonic kidney (HEK) cells were cultured on glass coverslips in DMEM supplemented with 10% fetal calf serum and glutamine (2 mM). The pCS2-mosmoa-HA construct was transfected using lipotransfectin (Solmeglas) following the manufacture instructions. Cells were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer pH 7.2 (wt/vol) at 37°C and then washed in PBS containing 0.5% Triton-X-100. Chick embryos were fixed by immersion in 4% cold PFA overnight at 4°C, washed, incubated in a 15% sucrose-PBS solution (wt/vol), embedded and frozen in a 7.5% gelatine in 15% sucrose solution (wt/vol). Cryostat sections, whole embryos or cells samples were stained with Hoechst and α-HA antibody produced in rabbit (1:250. Sigma, H-6908) and Donkey anti-Rabbit Alexa Fluor 488 secondary antibody (Invitrogen, A-21206), following standard procedures.

Single guide RNAs (sgRNAs) targeting coding regions of mosmoa and mosmob for CRISPR/Cas9 deletion were designed using the tools provided by CHOPCHOP online service¹ searching for potential disruption of restriction enzyme sites (Labun et al., 2019). Oligos were designed as described in Varshney et al. (2016) and their sequence is reported in Supplementary Table 1. sgRNAs were transcribed and purified using Maxi Script T7 (NEB) following the manufacturer’s instructions. sgRNAs were microinjected together with Cas9 protein (300 ng/μL; EnGen^® Spy Cas9 NLS, New England Biolabs) in 1 to 2 cell stage AB/Tue using a Narishige microinjector. F0 embryos were let grown and outcrossed with wt AB/Tue fish. Genomic DNA from tail clips of F1 zebrafish embryos was amplified by PCR and digested to identify disruption of selected restriction sites. DNA from potential mutants were sequenced, those with a disrupted and truncated reading frame were selected to generate the fish lines.

DNA from embryos or adult fish was amplified by PCR using the primers listed in Supplementary Table 1. PCR products were digested with selected enzymes at 37°C for 2 h to distinguish among wt, heterozygous, and mutant mosmoa and/or mosmob fish.

Cartilage staining of zebrafish larvae and adult fish bones was performed with Alcian Blue and Alizarin Red, as, respectively described in Schilling et al. (1996), Sakata-Haga et al. (2018).

Embryos were immersed in 75% glycerol and whole-body images were obtained using a Leica CTR5000 stereomicroscope connected to a Leica DFC500 digital camera operated by Leica software. The adult fish stained with the Alizarin Red chromogen, which also emits red fluorescence, were photographed under fluorescent light stimulation using a Leica CTR5000 stereomicroscope connected to a Leica DFC350 FX digital camera operated by Leica software. The drawings in Figure 4G were traced in Adobe Illustrator using representative photographs of wt and mosmoa^–/–;mosmob^–/– adult mutants. LSM710 confocal laser scanning coupled to an AxioObserver inverted microscope (Zeiss) was used to obtain digital images of cryostat sections or cells samples. ImageJ (Fiji) software was used to process and analyze images.

The ImageJ (Fiji) software was employed to obtain quantifications reported in Figure 3. Adult fish were anesthetized with tricaine and photographed in lateral views and the distance from the eye to the tip of the preorbital region was measured and normalized to the eye size in each one of the analyzed genotypes. Data were analyzed using GraphPad Prism 7 statistic software. One-way ANOVA test was used owing to the parametric distribution of the data, followed by Tukey’s multiple comparisons test to determine differences among groups.

The zebrafish genome carries two different paralogs of the mosmo gene: mosmoa and mosmob (ZFIN:ZDB-GENE-101203-6; ZFIN:ZDB-GENE-060929-1030). To determine their expression pattern during embryonic and larval development, we generated two different specific ISH probes for each one of the two paralogs (Figure 1A). Both mosmoa and mosmob were detected at gastrulation and bud stages as well as during somitogenesis (Figures 1H,I,O,P). At this stage shha, one of the ligands of the pathway, is expressed along the midline of the entire ventral neural tube (Figures 1B,C), from which it diffuses to pattern the adjacent cells with a mechanism highly conserved across vertebrates (Martí et al., 1995; Roelink et al., 1995). During somitogenesis, mosmoa and mosmob were also found localized along the length of the ventral neural tube (Figures 1J,Q) with an overlapping distribution that, however, was more dorsally extended than that of shha (Figures 1C–G). Specific expression was also observed in the mesenchyme surrounding the neural tube (Figures 1D–G) and in the optic vesicles (Figures 1K,R). At 2 dpf and larval stages, mosmoa and mosmob were no longer detected in the neural tube but strongly localized in the head mesenchyme (Figures 1L,M,S,T), surrounding, among others, the ethmoid plate (Figures 1N,N′,U,U′).

The reported patterns were consistently observed with both of the probes generated for each one of the paralogs (Figure 1), validating the reported distribution.

Attempts to determine the subcellular localization of the protein showed that, at least in NIH-3T3 cells, Mosmo localizes at the plasma membrane and endosome (Pusapati et al., 2018). To verify if this is the subcellular distribution in the developing embryo, we generated a human influenza hemagglutinin (HA) tagged version of mosmoa (mosmoa-HA). We focused on this paralog because its amino acid (aa) sequence is 100% identical to that of its human ortholog, whereas mosmob aa sequence has a lower homology (89.8% identity). We first verified the efficiency of our construct by transfecting HEK cells with the mosmoa-HA containing plasmid followed by immunostaining for HA. As reported for NIH-3T3 cells (Pusapati et al., 2018), the tagged protein was detected at the plasma membrane and endosomes (Figure 2A). When mosmoa-HA mRNA was injected in zebrafish, HA immunosignal was similarly localized at the blastomers’ plasma membrane and endosomes (Figure 2B).

When Hh signaling is active, Smo localizes at the primary cilium of the targeted cells. Notably, Mosmo was also observed in the primary cilia of NIH-3T3 cells (Pusapati et al., 2018). We thus asked if this localization could be observed also in vivo. The primary cilium can be easily detected in the chick neural tube as this organelle protrudes in the rather wide ventricle of chicken embryos. We thus co-electroporated two plasmids carrying mosmoa-HA and arl13b-RFP, respectively. The latter is a primary cilium specific protein, widely used to visualize this structure (Schmitz et al., 2017). Indeed, 24 h after electroporation, at HH14, HA, and RFP fluorescent signals co-localized in the cilium of a subset of the electroporated cells (Figures 2C–D′).

Taken together these data indicate that in vivo Mosmoa localizes at the plasma membrane, endosomes, and the primary cilia, suggesting that it may favors Smo translocation to this organelle, thus influencing signaling activation.

To explore the possible roles of mosmoa and mosmob, we inactivated the two genes using CRISPR-Cas9 technology. We selected founders at the F1 generation that carried frameshift mutations in either mosmoa or mosmob gene and generated stable mosmoa^–/– and mosmob^–/– mutant lines (Figure 3A). Mosmoa^–/– and mosmob^–/– mutant embryos show no gross morphological defects and grew to adulthood without evident defects. This was perhaps not surprising given that both genes share expression pattern and their respective proteins present a high degree of sequence homology, suggesting that the two paralogs may compensate each other activity. To overcome this possible compensatory effect, we intercrossed the mosmoa^–/– and mosmob^–/– mutant lines obtaining a mosmoa^–/–;mosmob^–/– double mutant fish. At first glance, double mutant embryos showed no major gross alterations or histological defects along the neural tube and their size was similar to that of their sibling (not shown).

Other than in the neural tube, the two mosmo paralogs are expressed with a largely overlapping pattern also in different regions of the larva head. We thus used Alcian blue staining to label the cranio-facial cartilage of the larva. There were no obvious differences in the cartilaginous elements when wt, mosmoa^–/–;mosmob^–/+ and mosmoa^–/–;mosmob^–/– were compared (Figures 3B–G), although the rostral tip of the head appeared flatter at least in part of the double mutants (Figure 3F, arrowhead). To determine if this abnormality was only transient, we analyzed the morphology of the head in the adult fish. Mosmoa^–/–;mosmob^–/– double mutants consistently exhibited a significantly shorter frontonasal region (Figures 3H,I), which was not observed in their sibling of other genotypes (Figure 3J). Furthermore, the operculum was reduced in size, leaving the gills exposed (Figure 3I).

Hedgehog signaling is essential for the development of the anterior neurocranium (Wada et al., 2005) and disruption of smo activity in zebrafish affects the craniofacial skeleton (Eberhart et al., 2006; Swartz et al., 2012). Thus, the frontonasal hypoplasia observed in the double mutants could be the consequence of alterations in the osseous components of the craniofacial skeleton. To determine this possibility, we stained the skeleton of wt, mosmoa^±;mosmob^± and mosmoa^–/–;mosmob^–/– adult fish with Alizarin red (Figures 4A–F). The bones of the frontonasal region, especially the maxillary and premaxillary bones of mosmoa^–/–;mosmob^–/– double mutants were altered as compared to those of wt or heterozygous fish (Figures 4A–F) as highlighted in the schematic drawings reporting the phenotypes (Figure 4G).

Taken together these data indicate that mosmoa and mosmob have an overlapping function, which is required for the acquisition of a proper craniofacial structure in zebrafish.