Role of trafficking protein particle complex 2 in medaka development

The skeletal dysplasia spondyloepiphyseal dysplasia tarda (SEDT) is caused by mutations in the TRAPPC2 gene, which encodes Sedlin, a component of the trafficking protein particle (TRAPP) complex that we have shown previously to be required for the export of type II collagen (Col2) from the endoplasmic reticulum. No vertebrate model for SEDT has been generated thus far. To address this gap, we generated a Sedlin knockout animal by mutating the orthologous TRAPPC2 gene (olSedl) of Oryzias latipes (medaka) fish. OlSedl deficiency leads to embryonic defects, short size, diminished skeletal ossification, and altered Col2 production and secretion, resembling human defects observed in SEDT patients. Moreover, SEDT knock-out animals display photoreceptor degeneration and gut morphogenesis defects, suggesting a key role for Sedlin in the development of these organs. Thus, by studying Sedlin function in vivo, we provide evidence for a mechanistic link between TRAPPC2-mediated membrane trafficking, Col2 export, and developmental disorders.

A distinct form of skeletal dysplasia called spondyloepiphyseal dysplasia tarda (SEDT, OMIM#313400) is caused by mutations in the TRAPPC2 gene.SEDT is less severe than SEDC, arises in late childhood, and is characterized by short stature, disproportionate short trunk and precocious mild osteoarthritis, which manifests in the early teens. 3The TRAPPC2 (or SEDL) gene is located on chromosome Xp22.2and encodes the ubiquitously expressed Sedlin/TRAPPC2 protein. 4Sedlin is a core component of the highly conserved oligomeric TRAnsport Protein Particle (TRAPP) complex that controls different segments of vesicular transport. 5The mammalian TRAPP complex exists in two different forms, TRAPPII and TRAPPIII.Both complexes, which control different stages of trafficking, have common core components but differ in their peripheral elements.Sedlin functions as a connector between the shared core and specific peripheral subunits.
We have shown that Sedlin, in addition to interacting with other TRAPP components, interacts with the small GTPase Sar1, modulating its GTP-GDP cycle by promoting GTP hydrolysis, a key event for the formation of ER-derived carriers. 6Due to this activity and the ability to interact with the procollagen export factor TANGO1 at ER-exit sites (ERES), Sedlin is required for efficient procollagen export, as shown in cultured chondrocytes and in SEDT patient fibroblasts. 6APPC2 is evolutionarily conserved, ubiquitously expressed in mammals and essential in yeast. 4,7The cartilage and bone restricted phenotype in SEDT patients seem contradictory but could be explained by the existence of an expressed TRAPPC2 pseudogene on chromosome 19.][10] Notably, the TRAPPC2 pseudogene is also present in mice (0610009B22Rik) and in rats (Gene ID: 100910318), and to date no functional studies have been performed in vivo to elucidate the role of Sedlin.We therefore decided to address the in vivo role of TRAPPC2 in a well-characterized teleost model for the study of osteochondral diseases, Japanese medaka (Oryzias latipes), 11 which has only one copy of the TRAPPC2/SEDL gene, hereon called olSedl.
We show that olSedl deletion induces a skeletal phenotype reminiscent of SEDT, but also extra-skeletal defects in the eye and gut.The marked defects in collagen deposition that we find in the Sedlin-deficient fish may provide the common basis for skeletal and extra-skeletal manifestations, as suggested by the co-occurrence of eye involvement in syndromic forms of osteodysplasia caused by COL2 mutations 1 and by the known role of the extracellular matrix (ECM), including collagen, in intestinal epithelium polarization. 12Our data suggest that Sedlin is required for vertebrate embryo development by regulating collagen production and secretion in different tissues.

| Generation of Sedlin knock-out medaka fish
The medaka genome harbors a single olSedl gene (http://www.ensembl.org/Oryzias_latipes;ENSORLG00000025160) that encodes a protein with 90% homology to human Sedlin (Figure S1A).Quantitative real-time PCR (qRT-PCR) analysis revealed high levels of olSedl expression from stage (st) 32 to 37 (Figure 1A) when organogenesis and bone tissue mineralization are occurring in medaka embryos. 13,14Notably, olSedl expression paralleled that of TRAPPC3, another core component of the TRAPP complex, and of Col2a1a (Figure 1A).Whole-mount in situ hybridization (ISH) using a DIGlabeled antisense RNA probe in stage 40 embryos showed the expression of olSedl in vertebrae with the strongest signal in the centrum and weaker expression in the neural spine and in the ventral rib (Figure S1B).

| OlSedl is required for skeletogenesis
No severe morphological abnormalities or defects in somatogenesis were observed in olSedl À/À compared to wild-type embryos during the initial stages of development.However, olSedl À/À larvae at stage 40 displayed a delay in mean hatching time, suggesting a severe reduction of prenatal motility.Indeed, olSedl À/À showed a dramatic reduction in larval movement and died within the first few hours after hatching (Movies S1 and S2).
olSedl À/À larvae also exhibited abnormalities in craniofacial morphogenesis, a shorter body, and a slight curvature of the trunk in the dorsal-ventral and medial-lateral planes, accompanied by highly penetrant bone defects (Figure 1B).Cartilaginous elements in the head, including the ceratobranchial pairs, ceratohyal and ethmoid plate, were almost absent in olSedl À/À fish, leading to a completely dysmorphic craniofacial skeleton compared to control fish (Figure 1C).As a consequence, olSedl À/À displayed an evident decrease in bone mineralization of the head, caudal fin rays and skeletal vertebrae (Figure 1D).
When analyzed by alizarin red staining, the number of calcified vertebrae at st40 was significantly lower than in WT animals (À18.5% ± 1.2 SD, Figure 1E).Furthermore, a reduction in the vertebrae thickness was detected in olSedl À/À (Figure 1F), in line with the spine phenotype observed in SEDT patients. 17 assess whether, as observed in cell systems, Sedlin controls collagen deposition also in vivo, we performed whole-mount immunofluorescence analysis for Col2A on larvae at st40.A notable reduction of Col2A in the ECM was found in olSedl À/À larvae (Figure 2A), which correlates with a reduction of protein levels (Figure S1E) and with diminished ECM thickness (Figure 2B).
Of note, fibroblasts derived from SEDT patients show ER dilation and acute Sedlin depletion in cell models leads to ER accumulation of Col2A and consequent ER expansion. 6 rule out possible TALEN off-target effects, 18 we injected WT fertilized eggs with a specific morpholino (MO) directed against the Sedlin ATG initiation codon within the 5 0 untranslated region (MO Sedl ) 19,20 (Figure 3A).MO Sedl display skeletal defects (70 ± 5% of 3000 injected embryos) including trunk curvature, reduction of cartilaginous elements, that is, vertebrae and caudal fin rays, and absence or reduction of the neural arch (Figure 3B-E).Moreover, MO Sedl shows a reduction of Col2A, phenocopying olSedl À/À and thus corroborating our findings (Figure 3F,G).This was accompanied by a reduction in WGA staining (Figure 3F) which could be correlated with a reduction in the expression or trafficking of proteoglycans to the ECM.Studying whether this additional indication of ECM disruption is due to a direct or indirect effect of Sedlin would be informative in delineating the role of TRAPPC2 in the trafficking of ECM components.
These data indicate that olSedl À/À recapitulates the major clinical signs observed in humans with TRAPPC2 mutations and represents a suitable model to study SEDT.

| Transcriptome changes induced by Sedlin depletion
To gain an unbiased view of the impact of Sedlin depletion on medaka fish development, we performed transcriptomics analysis on both morpholino-induced MO Sedl and on olSedl À/À embryos to mitigate the risk of transcriptional changes not directly or specifically due to the absence of functional Sedlin.QuantSeq 3 0 mRNA-Seq analysis was performed on olSedl À/À embryos while RNA-Seq was performed on MO Sedl embryos (Figure S2).QuantSeq 3 0 mRNA-Seq (GEO accession GSE143538) revealed 1052 differentially expressed genes (414 genes induced and 638 inhibited), while 4005 differentially expressed genes (1,870 genes induced and 2135 inhibited) were found by RNA-Seq (GEO accession GSE186769) (Figure S2A).Differentially expressed genes (DEGs) were converted into the corresponding human orthologues using the BioMart browser. 21 found 695 DEGs in common between the two datasets, of which 254 were induced and 399 were inhibited genes (Figure S2A).This comparison of the MO Sedl and olSedl À/À data increased the confidence that the changes were due specifically to the loss of Sedlin.
Among the most down-regulated gene classes were genes involved in extracellular matrix organization (GO:0030198), but also genes involved in visual perception (GO:0007601) and intestinal absorption (GO:0050892) (Figures S2B and S3).On the one hand, these findings are

| olSedl oversees chondrocyte differentiation
Collagens and collagen-related genes were among the most affected of the down-regulated genes (Figure 4A; Figure S2C) with the Col2a1 gene being one of the most inhibited genes in the olSedl À/À .The expression of type XI collagens (Col11a1 and Col11a2), which are part of the nucleating core of type II collagen fibrils, and of Col10a, a marker for chondrocyte maturation, 22 was also significantly reduced in olSedl À/À (Figure 4A).The impaired Col10 expression was validated in medaka lines expressing GFP under the control of the Col10a1 promoter. 23Here, GFP fluorescence was significantly reduced in olSedl +/À heterozygous fish and almost lost in olSedl À/À (Figure 4B).
One possibility for the reduction in Col2A mRNA expression was that transcription factors controlling its expression might be affected by Sedlin depletion.However, we found no significant changes in the expression levels of the main transcription factors upstream of col2a1, such as Sox9b, Sox5 and Sox6 24 in olSedl À/À .It has been reported that altered ECM due to Col6A2-KO in mice leads to abnormal chondrocyte differentiation and maturation associated with altered ECM-gene expression profiles 25 and that the absence of aggrecan affects the coexpression of genes encoding collagens II, X and XI. 26 Thus, alteration of one ECM component may lead to alterations in the expression of other ECM components that together contribute to the observed phenotype, but how this occurs mechanistically is still unclear.This appears to be the case here, where impairment of Col2A secretion due to the lack of Sedlin causes ECM remodeling and induces a feed-forward loop that negatively impacts on the expression of ECM genes.

| Sedlin is required for photoreceptor development
As mentioned above, our transcriptomics dataset revealed that Sedlin depletion strongly downregulated genes involved in visual perception (GO:0007601), detection of light stimulus (GO:0009583) and visible light (GO:0009584).For example, the expression of photoreceptor markers such as GNAT1, GNAT2 and GNGT1, encoding the Transducin 1,2 and gamma subunits which couple rhodopsin stimulation and cGMP-phosphodiesterase 27 (Figure 4C), and RPGRIP1 (logFC, À3.2), encoding a photoreceptor component of cone and rod photoreceptor cells, 28 was strongly impaired.
Similarly, the expression of genes involved in the maintenance of retinal integrity such as ROM1, a member of a photoreceptor-specific gene family that encodes Rod Outer Segment Protein, an integral membrane protein found in the photoreceptor disk rim of the eye 29 (Figure 4C), and RS1 (logFC, À2.915), encoding retinoschisin, an extracellular protein that plays a crucial role in the cellular organization of the retina, 30 was also reduced.
The reduced expression of genes involved in signal transduction and retinal integrity prompted us to explore the biological relevance of these results, evaluating whether olSedl ablation affected photoreceptor differentiation and maintenance during eye development.
We found that while the lamination of the retina was preserved, retinas from olSedl À/À displayed significant alterations in photoreceptor marker staining, which was more evident in ventral regions compared to WT larvae (Figure 5).Specifically, a strong reduction of rhodopsinpositive mature rod photoreceptors was accompanied by a decrease in Zpr-1-positive mature cone photoreceptors (Figure 5A,B).Notably, outer/inner segments of both cones and rods appeared significantly shorter and morphologically abnormal in olSedl À/À retina compared to WT siblings (Figure 5A,B).The disorganization of the photoreceptors detected at st40 was already visible at st36 (Figure S4A,B), supporting an alteration of photoreceptor development.Consistent with defective photoreceptor differentiation and structures, TEM analysis confirmed that photoreceptor outer segments were shorter and thinner in olSedl À/À retinas (Figure 5C,D).However, we did not detect any apparent morphological defects in the retinal pigment epithelium of olSedl À/À with respect to WT controls, as determined by TEM analysis.
In addition, we observed an increase of apoptotic cells in retinal tissue of olSedl À/À at Stage 34 (Figure S4C,D,G), when olSedl expression levels peak (Figure 1A) and the highest rate of photoreceptor differentiation in the retina of medaka fish occurs. 31Notably, at the same stage, olSedl À/À showed a reduction in proliferating cells, as determined by immunostaining for phosphorylated histone H3 (Figure S4E,F,H).
Altogether, these data indicate that olSedl is required for photoreceptor differentiation and maintenance.

| Sedlin is necessary for intestinal epithelium polarization and development
As noted above, genes belonging to the functional class intestinal absorption were significantly downregulated both in MO Sedl and olSedl À/À (Figure 4D; Figure S2C).These genes belonged mainly to the cell component class brush border and included solute and fat transporters (e.g.SLC26A6, SLC3A1, CD36, NPC1L1) and regulatory proteins (PDZK1) present on the apical membrane of intestinal cells, together with structural components of the microvilli (EZR, ESPN).
These findings prompted us to analyze the intestine of olSedl À/À fish.
To this end, semi-thin (1 μm) transverse sections at a comparable cranio-caudal axis level were obtained from both WT and olSedl À/À medaka embryos and stained with toluidine blue.Morphological analysis of posterior gut sections showed a disorganization of epithelial tissue with alterations of cell polarization and a reduction of epithelial cell layers in olSedl À/À relative to age-matched WT embryos (Figure 5E) in which the gut epithelium is a bilayer sheet and has a radially organized structure with basal-side nuclear localization.Furthermore, the elongated shape of the WT gut epithelium cells in the olSedl À/À embryos is reduced.
Moreover, the structural alteration of the epithelial tissue in olSedl À/À was associated with a significant reduction of the gut lumen (Figure 5E).These data suggest a central role for Sedlin in key phases of gut morphogenesis.

| DISCUSSION
Eight distinct genetic disorders are caused by mutations of eight different components of the TRAPP complex: four (TRAPPC2, TRAPPC2L, TRAPPC4, TRAPPC6B) belonging to the core TRAPP complex (common to TRAPPII and TRAPPIII), two subunits (TRAPPC9, TRAPPC10) specific for the TRAPPII complex and two (TRAPPC11 and TRAPPC12) specific for TRAPPIII. 5The clinical manifestations of TRAPPC2 mutations that cause SEDT are distinct from those of the remaining seven.SEDT has cartilage-restricted manifestations and the age-of-onset of clinical signs is in puberty while the other seven TRAPPopathies have neurological and muscular manifestations and a much earlier onset. 5This difference in the phenotypic manifestations in SEDT relative to the other TRAPPopathies would lend support to the idea that the TRAPPC2 pseudogene might mask other phenotypes caused by mutation of TRAPPC2.The cartilage defects are consistent with the role of Sedlin in procollagen transport that we have shown previously in cell systems 6 and confirmed here in vivo, and with the extent of collagen secretion during chondrogenesis largely exceeding that of other tissues, while the cartilage-restricted phenotype and later onset could be related to differential temporal/spatial expression of the pseudogene TRAPPC2B.
We have reported here the first animal model for SEDT in medaka fish, which has only one Sedlin-encoding gene.The skeletal phenotype (platyspondyly/reduced length) is highly reminiscent of that observed in SEDT patients, confirming that Sedlin plays a major role in skeletal development.The altered deposition of collagen and the dilation of the ER observed in patient tissues were reproduced in Sedlin KO fish.The olSedl À/À fish not only represent a reliable model for SEDT but have also revealed multiple aspects that help in understanding the role of Sedlin in development, and hence in the pathogenetic basis of the disease.Altered deposition of Col2A at the ECM in the olSedl À/À fish leads to ECM disorganization and affects expression not only of the collagen II gene but also other collagen genes such as collagens XI and X, the latter a marker for chondrocyte maturation.
Defective ECM deposition affects chondrocyte differentiation markers and the expression of ECM components 25,26,32 and cellmatrix interactions impact on chondrocyte differentiation and cartilage development through cell receptors and their signal transduction pathways. 33It is well documented that mechanotransduction (stretching or loading) leads to the upregulation of ECM genes via an integrated network of various signaling pathways.However, the details in terms of the mechanisms and individual players are limited, as are the receptors that relay the response, 34,35 perhaps the best characterized being collagen interaction with members of the integrin family. 36erefore, it is not unreasonable to suppose that alteration of the ECM due to defective collagen trafficking as in SEDT would alter signaling events that impact on the transcriptome.
In addition to confirming that Sedlin plays a major role in skeletal development, we also observed additional phenotypes in olSedl À/À that affect the eye and the gut.These may emerge in medaka because there is only one Sedlin-encoding gene and are not observed in humans due to the presence of the expressed pseudogene TRAPPC2B. 9,10The extraskeletal manifestations could be due to altered ECM deposition in the eye or in the gut, which is an important determinant for their correct development.The ECM plays a key role in retinal cell proliferation, homeostasis and survival. 37In addition, a remodeling of various ECM molecules has been associated with retinal neurodegeneration. 38,39ong the ECM components, collagen II plays a pivotal role in eye development as seen by the ocular manifestations that are associated with mutations of PCII, such as Stickler syndrome type I (membranous vitreous type, OMIM:#108300), Stickler syndrome type I (predominantly Ocular, OMIM:#609508) and vitreoretinopathy with phalangeal epiphyseal dysplasia (OMIM: #619248).
A non-mutually exclusive hypothesis, in line with eye phenotypes reported in COPII mutant zebrafish, 40 is that Sedlin may be required for rhodopsin trafficking, which occurs at an extremely high rate in photoreceptors.This possibility is supported by the mis-localization of rhodopsin that we observed in rod cells (Figure 5A; Figure S4A), where rhodopsin can be seen throughout the cell bodies rather than in the outer segments.In addition, the alteration of collagen gene expression in the olSedl À/À larvae was accompanied by defects in cell death and proliferation, indicating that Sedlin is dispensable for early steps of eye development but has an unanticipated role in photoreceptor differentiation and maintenance.
Finally, another unexpected role for Sedlin was found to be in gut morphogenesis.Although the basis of this defect will require further analysis, it has been reported that mutation of Sec13 in zebrafish, a subunit of the outer coat of the COPII complex, affects organogenesis of the digestive system, including disruption of the ER in chondrocytes. 41Furthermore, Sar1B loss of function results in chylomicron retention disease, a rare recessive condition with defects in fat malabsorption in hepatocytes and enterocytes. 42Together, this evidence supports a model in which finely-tuned regulation of the COPII machinery is critical for gut function, corroborating our data.
It should be kept in mind that there is no evidence for extraskeletal developmental phenotypes in SEDT patients so these may be specific for medaka and are not relevant for the disease phenotype.Nonetheless, they are still informative regarding the in vivo TRAPPC2 function and its role in collagen transport.Instead, the similarities in the skeletal phenotypes in humans and fish described earlier represent a suitable model to study SEDT.The fish model has been useful in highlighting the transcriptome changes that occur upon Sedlin KO that extends the alterations in the cell beyond simply trafficking of collagen from the ER, suggesting the existence of a deleterious vicious circle through which impaired collagen secretion inhibits the transcription of various collagens and other ECM components.A similar analysis in a mammalian system would be informative to extend these observations.The model could also address the issue of whether there is an impairment in the secretion of other ECM components as well as Col2A, or of other types of high load cargoes, such as rhodopsin, in other tissues.Finally, fish models lend themselves to high throughput automated drug screening 43 or, additionally, they could provide an in vivo assay to test the efficacy of drug hits isolated from screens on cellular models.
In summary, we generated and characterized the first vertebrate model for SEDT, which serves as a valuable tool for gaining deeper mechanistic insights into developmental disorders arising from abnormal Col2 transport and ECM deposition due to membrane trafficking defects.

| Medaka fish stocks
Medaka fish (O.latipes) from the Cab inbred strain were used throughout the study and maintained following standard conditions (12 h/12 h dark/light at 27 C).Embryos were staged according to the method proposed by Iwamatsu. 13All studies on fish were conducted in strict accordance with the institutional guidelines for animal research and approved by the Italian Ministry of Health, Department of Public Health, Animal Health, Nutrition and Food Safety in accordance with the law on animal experimentation (D.Lgs.26/2014).Furthermore, all fish treatments were reviewed and approved in advance by the Ethics Committee at the TIGEM Institute (Pozzuoli (NA), Italy).

| Whole-mount in situ hybridization
Whole-mount RNA in situ hybridization was performed and photographed as previously described. 44A digoxigenin-labeled antisense riboprobe for olSedlin or a sense riboprobe as negative control were used.Probes were generated from cDNA amplified from st40 larvae using specific primers (Table S1) and cloned into the Topo-TA-vector (Invitrogen).Larvae were sectioned using a Vibratome (Leica) at 25 μm and images were acquired using Leica DM6000 microscopy.

| Real-time qRT-PCR
Transcriptional levels of olSedl, olBet3 and olCol2a1 were analyzed by quantitative real-time RT-PCR.RNAs were obtained from whole mount larvae at the stages indicated in the legend to Figure 1A using a RNeasy Mini Kit (Qiagen, UK), according to the manufacturer's protocol.1 μg of total mRNA from each sample was retrotranscribed using QuantiTech reverse Transcription Kit (Qiagen).Real-Time PCR was performed using SYBR Green Master Mix (Bio-Rad) using the primers listed in Table S1.
Each reaction was performed in triplicate using 25 ng of cDNA in 20 μL.
The results were normalized against an internal control (olHprt).

| Genomic analysis and generation of olSedl À/À medaka
The medaka olSedlin genomic sequence was retrieved from public databases (http://www.ensembl.org/Oryzias_latipes;ENSORLG00000025160). Custom-designed transcription activatorlike effector nucleases (TALEN) were used to induce targeted mutagenesis in the exon 3 of the olSedl gene.Potential TALEN target sites were identified by using the TALEN Targeter program at (https://tale-nt.cac.cornell.edu/node/add/talen). Custom-designed TALEN vectors were assembled by Zgenebio (Zgenebio Biotech Inc, Taiwan).The left (L) and right (R) recognition sequences and the spacer sequence on the olSedl gene are reported in Table S1.TALEN RNA was synthesized by transcription using a commercial kit (mMESSAGE mMACHINE SP6 Transcription Kit, Ambion, Life Technologies) and injected into fertilized eggs at the 1-2 cell stage, a concentration of 100 ng/μL.After hatching, 28 larvae were sacrificed and lysed for genomic DNA extraction.The target region on the olSedl gene was amplified by PCR using the primers indicated in Table S1, and then sequenced to determine the presence of TALEN-induced mosaicism.G0 generation fish from the TALEN-injected embryos and G1 heterozygous fish were selected to generate G2 populations.G2 and G3 populations were genotyped through PCR amplification on genomic DNA extracted from the caudal fin (primers are listed in Table S1).Once amplified, samples were incubated with the restriction enzyme Xmn1 (NEB, USA), which is able to cut only the WT DNA, as the sequence recognized by the enzyme is lost after TALEN-induced mutagenesis (Figure S5).For cartilage staining, the fixed medaka larvae were incubated in a freshly prepared cartilage staining solution containing 0.02% alcian blue (Sigma-Aldrich), 200 mM MgCl 2 , and 70% ethanol.Samples were thoroughly washed in 70% ethanol.After the background coloration was removed, the stained medaka larvae were moved through a graded series of glycerol (50% to 90%) and then kept in 100% glycerol.Images were acquired using a Leica DM6000 microscope.
For staining with alcian blue and alizarin red, the alcian blue stained larvae were washed briefly with a solution containing 20% ethylene glycol and 1% KOH, then incubated in a freshly prepared bone staining solution containing 0.05% alizarin red (Sigma-Aldrich), 20% ethylene glycol, and 1% KOH at RT for 30 min with gentle agitation.Samples were thoroughly washed in a prewarmed clearing solution containing 20% polyoxyethylene, 20 0.1% Tween 20 and 1% KOH at 42 C for 3 h or more with agitation.

| MO and mRNA injections in medaka embryos
Morpholinos (Gene Tools, LLC Philomath, OR, USA) were designed against the ATG translational start-site on the olSedl gene.Sequences are reported in Table S1.
Fifty picolitres of Mo-Sedlin solution (approximately 1/10 of the cell volume) were injected at a 0.09 mM concentration into one blastomere at the one/two-cell stage as described previously. 45A morpholino designed against the O. latipes p53 gene (olp53) gene was used to control the specificity and inhibitory efficiencies of the morpholinos and the absence of off-targeting effect due to activation of p53, as previously described. 45

| Whole mount immunostaining
Embryos from Stage 30 onwards were fixed in 4% paraformaldehyde, 2Â PBS and 0.1% Tween-20.Fixed embryos were detached from the chorion and washed with PTW 1Â.Embryos were digested 20 min with 1m μg/mL proteinase K and washed twice with 2 mg/mL glycine/PTW 1Â.

| Image analysis
The images were processed with Fiji (ImageJ, National Institutes of Health (NIH)) software.Brightness and contrast were adjusted with Adobe Photoshop, and figure panels were assembled with Adobe Illustrator.Images in Figures 2A and 3A and Figure S1C were created using BioRender.

| Detection of apoptotic cell death
The extent and distribution of apoptotic cell death was determined by TUNEL, using the In Situ Cell Death Detection Kit, POD (Roche), following the manufacturer's protocol.TUNEL assay was performed on 20 μm medaka retina cryosections.As negative control to evaluate possible non-specific effects, fixed and permeabilized retina sections were incubated with the reaction mix without TUNEL reaction enzyme.Sections were observed with a Leica DM-6000 microscope and then confocal images were acquired using the LSM710 Zeiss Confocal Microscopy system.

| Electron microscopy analysis
Medaka larvae at stage 40 were fixed using a mixture of 2% paraformaldehyde and 1% glutaraldehyde prepared in 0.2 M HEPES buffer (pH 7.4) for 24 h at 4 C. Larvae were then post-fixed.After dehydration the specimens were embedded in epoxy resin and polymerized at 60 C for 72 h.Thin 60 nm sections were cut on a Leica EM UC7 microtome.EM images were acquired from thin sections using a FEI Tecnai-12 electron microscope equipped with a VELETTA CCD digital camera (FEI).To assess the gut epithelium tissue organization, toluidine blue staining (1%) was performed on semi-thin (1 μm) transverse sections at a comparable cranio-caudal axis level from both WT and olSedl À/À plastic-embedded medaka larvae and the sections were examined by light microscopy.

| Protein isolation and Western blot analysis
Total protein extracts were obtained from a pool of 5 WT (control), and 5 olSedl À/À larvae for each experiment.Larvae were sacrificed and immediately processed with a pestle in SDS buffer (10 mM Tris-HCl pH 8.0, 0.2% SDS and protease inhibitor cocktail).Protein concentration was determined using the Bio-Rad protein assay (Bio-Rad).A total of 35 μg protein from each sample was resolved by SDS-PAGE and transferred to a nitrocellulose membrane (GE Healthcare).The following antibodies were used at the indicated dilutions: rabbit anti-collagen type II 1:1000 (Rockland, 600-401-104), rabbit anti-actin 1:1000 (Sigma-Aldrich, A2066), rabbit anti-sedlin 1:1000. 6Proteins were detected with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG antibody (1:8000, Merck Millipore) and visualized with the LiteAblot Plus substrate (Euroclone), according to the manufacturer's protocol.Images were acquired using the Chemidoc-lt imaging system (Uvitec Cambridge).

| Library preparation and deep sequencing
For RNA-seq analysis, total RNA was extracted from a pool of 3 Mo-Sedl stage 40 embryos and 3 control (WT) stage 40 embryos.
Total RNA (500 ng) from each sample was prepared using TruSeq RNA sample prep reagents (Illumina) according to the manufacturer's instructions.Quality control of library templates was performed using a High Sensitivity DNA Assay kit (Agilent Technologies) on a Bioanalyzer (Agilent Technologies).The Qubit quantification platform (Qubit 2.0 Fluorometer, Life Technologies) was used to normalize samples for the library preparation.Using multiplexing, up to six samples were combined into a single lane to yield sufficient coverage.The amplified fragmented cDNA of $200 bp in size were sequenced in paired-end mode using the HiSeq 1000 (Illumina) with a read length of 2 Â 100 bp.Each library was loaded at a concentration of 8 pM, which was previously established as optimal.An average yield of $4.5 Mb was obtained per sample.
Col2 expression and secretion is compromised in olSedl À/À fish.(A) Representative immunostaining of Meckel's cartilage (boxed area in the graphic representation on the left) of WT and olSedl À/À fish showing a reduction in type II collagen (red) levels.Nuclei are counterstained with DAPI (blue) [(cb1-4) ceratobranchial pairs 1 to 4, (ch) ceratohyal, (et) ethmoid plate, (mk) Meckel's cartilage, (po) posterior limit, (e) eye].(B) Quantification of ECM thickness in WT and olSedl À/À fish.(C) Representative electron microscopy images (16Â magnification) of a vertebral section of stage 40 WT and olSedl À/À fish.The ER is pseudocolored in green and the ER cisternae are indicated by black arrows.The dashed line and punctate pattern in the violin plot in (B) show the median and quartiles, respectively.(B): n > 90 measurements.***p < 0.001, two-tailed unpaired t-test with Welch's correction.Scale bar in (A): 30 μm, (C): 500 nm.consistent with the observed skeletal phenotype (by highlighting a dysregulation of ECM components) while on the other they point to a possible role of Sedlin in controlling eye and intestine development.
morpholino fish manifest a SEDT-like phenotype.(A) Schematic representation of the morpholino targeting site at the ATG in exon 1 of olSedl.(B) Representative brightfield images of WT and MO Sedl fish.(C) Representative whole-mount alcian blue-alizarin red staining images showing that MO Sedl lack cartilaginous intermediates in the caudal fin and have a reduced number of vertebrae.(D) High magnification alcian blue-alizarin red staining from the first to tenth vertebra showing altered morphology.(E) Quantification of the observed phenotypes.(F) Immunofluorescence analysis of WT and MO Sedl sections of frontal vertebra showing a significant alteration in size and morphology.Nuclei are counterstained with DAPI (blue), connective tissue is visualized with fluorescent Wheat Germ Agglutinin (WGA, green), while type II collagen (Col2) is shown in red.(G) Col2 fluorescence intensity measurement of WT and MO Sedl sections of frontal vertebra; N = 3, median is indicated.**p = 0.001, two-tailed unpaired t-test.Scale bar in (B) and (C): 1 mm; (F): 100 μm.