Adult Ciona robusta was collected from Rongcheng and Qingdao in Shandong Province, China. Following collection, the animals were temporarily housed in laboratory artificial seawater tanks, which were carefully maintained at a temperature range of 17°C to 19°C and a salinity level of 30-32‰. Sperm and eggs were obtained separately from different adult Ciona individuals. The sperm and eggs were mixed in sea water over 5 minutes for fertilization. Then, the fertilized eggs underwent a dechorionation process before proceeding with electroporation. The electroporation method for Ciona embryos was performed based on previously reported techniques (Stolfi and Christiaen, 2012). 40 µg of each plasmid was dissolved in sterile ddH₂O, reaching a total volume of 80 µl. This plasmid solution was then mixed with 420 µl of electroporation buffer. The resultant mixture was transferred into a 4 mm electroporation cuvette. Subsequently, 300 µl of sea water together with dechorionated Ciona fertilized eggs were added to the cuvette. Electroporation was carried out using a pulse generator, employing a voltage of 50V and a capacitance of 1500/2000 µF. Following the electroporation step, the embryos were placed into a constant temperature incubator set at 16°C for further incubation.

In the promoter analysis experiment, the upstream 3 kb DNA sequence of Nas15 was amplified by PCR using genomic DNA from C. robusta as a template with primers Nas15 (3 k) -F and Nas15 (3k) -R. The PCR product was then inserted into the pEGFP-N1 vector to construct the Cr-Nas15 (3 kb) > GFP fusion construct.

The Ciona notochord specific expressed vector was constructed via replacing the CMV promoter of pEGFP-N1 with Ciona savignyi (Cs) brachyury upsteam 3kb promoter. Using C. robusta cDNA as a template, full-length coding sequences (CDS) of Nas15 and PPP2CA were amplified via PCR with primers Bra > Nas15-F and Bra > Nas15-R and Bra>PPP2CA-F and Bra>PPP2CA-R, respectively. The PCR products were then inserted into Brachyury (3 kb) > EGFP-N1 vector, leading to the construction of the expression plasmids Brachyury (3 kb) > Nas15-EGFP and Brachyury (3 kb) > PPP2CA-EGFP, and the construction of mutant plasmids were generated introducing mutated bases for overexpression experiments. Further, the PCR product was cloned into the linearized vectors CMV > GFP and CMV > HA, generating the expression plasmids CMV > Nas15-GFP and CMV > PPP2CA-HA for cell experiments. Additionally, the PCR product was cloned into the linearized vectors pGBKT7 and pGADT7 to construct the expression plasmids pGBKT7-Nas15 and pGADT7-PPP2CA for yeast two-hybrid assays. Finally, the CDS of Nas15 was ligated to the pET-30a vector for prokaryotic expression of Nas15 protein.

The ClonExpress II one-step cloning kit (Vazyme, Nanjing, China) was utilized to generate all constructs, and their accuracy was verified through sequencing analysis. All the PCR primers are listed in the Supplementary Table S1 .

The qPCR was performed based on the previously reported techniques (Livak and Schmittgen, 2001). Nas15 qPCR primers were designed using Beacon Designer 7 software. Total RNA was extracted from various developmental stages (8, 12, 13, 15, 16, 17, 20, 22, and 24 hours post fertilization, hpf) of C. robusta embryos/larvae using RNAiso plus (TAKARA, Japan). The extracted RNA was reverse transcribed into cDNA using the HiScript II Q RT SuperMix for qPCR kit (Novogene, Nanjing, China). Quantitative PCR amplification was performed using the ChamQTM SYBR Color qPCR Master Mix kit (Novogene, Nanjing, China). Data analysis was carried out using the 2^−ΔΔCt method, and graphical representation was created using Excel software. The primers are listed in the Supplementary Table S1 .

23 hpf (late tailbud stage) Ciona embryos were collected and lysed in loading buffer at 100°C for 10 minutes to denature the proteins. The samples were then loaded onto an SDS-PAGE gel for electrophoresis. After electrophoresis, the protein was transferred to a polyvinylidene fluoride (PVDF) membrane. Following membrane transfer, the PVDF membrane was blocked with 5% skimmed milk at room temperature for 2 hours. Primary antibody (Nas15 1:300) was diluted in 10% goat serum and incubated with PVDF membrane for overnight at 4°C. After completion of the primary antibody incubation, the PVDF membrane was washed three times with 0.05% TBST solution for 15 minutes each time. Secondary antibody Goat Anti-Mouse IgG, HRP Conjugate (TransGen, 1:2000, HS201–01) was diluted in 10% goat serum and incubated with PVDF membrane at room temperature for 2 hours. After the secondary antibody incubation, the PVDF membrane was washed three times with 0.05% TBST solution for 15 minutes each time. Once the washing was completed, the membrane was subjected to staining and photographed.

Embryos were fixed with 4% paraformaldehyde at room temperature for 2 hours, followed by permeabilization with PBS solution containing 0.1% Triton X-100 (permeabilization was performed three times for a total of 8 hours). Blocking was carried out with 10% goat serum at room temperature for 1 hour. Incubation with the primary antibody (Nas15 1:100) was performed overnight at 4°C. Washing was performed three times with PBS solution containing 0.1% Triton X-100 (a total of 8 hours). Incubation with Alexa Fluor 488-conjugated phalloidin (Invitrogen) and the secondary antibody Goat Anti-Mouse IgG, HRP Conjugate (TransGen, 1:2000, HS201–01) was carried out overnight at 4°C. Washing was performed three times with PBS solution containing 0.1% Triton X-100 (a total of 8 hours). Finally, the embryos were mounted with DAPI for confocal microscopy observation.

14 hpf (early tailbud stage) Ciona embryos was collected and cultured into the 48-well plate (about 200 embryos per well with 2 ml filtered seawater until 16 hpf (mid tailbud stage). Then, Actinonin dissolved in DMSO (MedChemExpress) was added to a final concentration of 1.2 μm in the inhibitor treated group, and the control group was added an equal amount of DMSO. Three repeated experiments were set for each group. Then, the embryos were cultured in 18°C until 23 hpf (late tailbud stage) and collected respectively. The embryos were fixed with 4% paraformaldehyde. Following the staining with phalloidin, the embryos were mounted with DAPI for confocal microscopy observation and imaging.

The genome editing in Ciona embryos by CRISPR/Cas9 was performed based on the previously reported techniques (Stolfi et al., 2014). Using the CRISPRdi-rect (http://crispr.dbcls.jp), three Nas15 sgRNA sequences were designed and subsequently sent to Shanghai Biotechnology Company for synthesis. Then, two single strands of sgRNAs were mixed respectively and subjected to a 5-minute treatment at 100°C to ensure complete denaturing, followed by a cooling to room temperature for annealing. Subsequently, the annealed primers were ligated to the linearized Cr-U6 > sgRNA (F+E) (Addgene number: 59 986) for sgRNAs transcription in vivo. Then the constructed sgRNA expression transcription vector was then used as a template for PCR amplification of the U6>sgRNA fragment. After that, the Cr-EF1α > NLS:: Cas9::NLS::P2A-mCherry plasmid and the sgRNA PCR product was co-electroporated into Ciona embyros. Then, the embryos were cultured in 18°C until 18hpf, and the embryos with red fluorescence were sorted and collected. Genomic DNA extracted from the collected embryos was used for efficiency test using T7 endonuclease I digestion. Additionally, the Cs-Brachyury (1 kb) > NLS::Cas9::NLS::P2A::mCherry and the sgRNA PCR product was co-electroporated into Ciona embryos for phenotype observation. The sequences of sgRNAs were listed in the Supplementary Table S2 .

The yeast two-hybrid experiment was performed following the protocol provided by Matchmaker GAL4-based two-hybrid system (Clontech) (Hu et al., 2021).

HEK293T cells were cultured in DMEM medium supplemented with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin. The cells were maintained at 37°C in a 5% CO2 atmosphere. After reaching 80-90% confluency, the transfection was carried out using Lipofectamine 3000 Invitrogen (Thermo Fisher) according to the manufacturer’s instructions. 48 hours post-transfection, cells were collected and lysed in a buffer containing 50 mM Tris-HCl (pH 8), 75 mM NaCl, 1 mM MgCl₂, 0.05% NP-40, 100 mM sucrose, 1 mM DTT, and 1x Protease Cocktail inhibitors (Roche) at 4°C for 30 minutes. Centrifuge at 12000 rpm for 10 minutes to remove cellular debris. Afterward, the supernatant was subjected to overnight incubation with GFP-tap beads (gta 20, chromotek, Germany) on a rotator at 4°C. The immunoprecipitated samples were washed three times with lysis buffer, and then directly boiled at 95°C for 10 minutes in 2×SDS loading buffer. Subsequently, these samples were loaded into 10% SDS-PAGE gels and separated via electrophoresis.

A phylogenetic tree for Nas15 was constructed using MEGA11 software (Tamura et al., 2021). The Conserved Domains Database tool provided by the NCBI website was utilized for protein domain prediction. Multiple sequence alignment was conducted using DNAMAN software (Lynnon Biosoft Bioinformatic Solutions, San Ramon, CA, USA), and the resulting alignments were visually represented. The diameter of the lumen was measured using the software ImageJ. All Significance statistical analysis was preformed using unpaired two-tailed t test. * represents p<0.05. ** represents p<0.01. *** represents p<0.001. **** represents p<0.0001. The colocalization of Nas15::tdTomato and PPP2CA::eGFP was analyzed with ImageJ. The phosphorylation sites in Nas15 were predicted using NetPhos - 3.1 (https://services.healthtech.dtu.dk/) website. The single cell sequencing data in different tissues and stages of Cr-Nas15 were downloaded from Single Cell Portal (SCP) database (https://singlecell.broadinstitute.org/single_cell/study/SCP454/comprehensive-single-cell-transcriptome-lineages-of-a-proto-vertebrate#study-summary ) (Tarhan et al., 2023). The in situ hybridization result of Cr-Nas15 were identified on Ghost database(http://ghost.zool.kyoto-u.ac.jp/cgi-bin/photogetkh.cgi?inkey=CLSTR07818&source=kh2013) (Kusakabe et al., 2002; Satou et al., 2005).

The sequence of Cr-Nas15 was downloaded from the Ghost database. The Cr-Nas15 is a 44.2 KDa protein and the in situ hybridization data demonstrated its expression in the larval notochord of Ciona (Kusakabe et al., 2002; Satou et al., 2005; Tassy et al., 2010). Meanwhile, the single-cell transcriptome sequencing data showed that the Cr-Nas15 was recognized in larval notochord cells at the lumen formation stage (Tarhan et al., 2023). To explore the evolution of Cr-Nas15, a group of Cr-Nas15 homologous sequences in other species was identified and then a phylogenetic tree constructed using full-length sequences revealed that Cr-Nas15 clustered with proteins from vertebrates, such as Homo sapiens MeprinA and Mus musculus MeprinA, suggesting its evolutionary conservation from invertebrates to chordates ( Figure 1A ). We also compared the Cr-Nas15 domain structure with homologous proteins in other species using the online tool NCBI Conserved Domain Search. The results showed that they both had conserved ZnMc-astacin-like domain and Zinc-binding metalloprotease motif ( Figure 1B ). Because of the key function of Zinc-binding metalloprotease motif, we analyzed the primary structure of this motif. The results showed that the Cr-Nas15 together with the homologous proteins in other species had a conserved motif HEXXHXXGXXH, which was the core structure for Zinc-binding metalloprotease catalytic activity (Bond and Beynon, 1995; Trevisan-Silva et al., 2010) ( Figure 1C ). In summary, the Cr-Nas15 sequence structure is evolutionary conserved including a typical ZnMc-astacin-like domain and Zinc-binding metalloprotease motif, suggesting that the Cr-Nas15 may has the similar catalytic activity with that in other species.

To confirm the temporal expression of Cr-Nas15 from single-cell transcriptome sequencing dataset, we performed a qRT-polymerase chain reaction (PCR) analysis of Cr-Nas15 using different developmental-staged embryos. The result shows that the expression level is increasing from 8 hpf (mid gastrula stage) to 24 hpf (hatching larva stage), and highly expressed during 20 hpf (late tailbud stage) to 24 hpf ( Figure 2A ), which is consistent with the single-cell transcriptome sequencing data. To further explore the expression level of Cr-Nas15 in different tissues, we built a UMAP plot based on the single cell sequencing data of Ciona early embryos. The result indicated that Cr-Nas15 was expressed in many different tissues, such as the notochord, epidermis, muscle, and endoderm (Cao et al., 2019; Tarhan et al., 2023) ( Figure 2B ). Together with the qPCR ( Figure 2A ) and in situ hybridization data (Kusakabe et al., 2002; Satou et al., 2005), we inferred that Cr-Nas15 was expressed in notochord in lumen expansion stages. To confirm this spatial and temporal overlap between Cr-Nas15 expression and notochord lumen formation, we investigated the Cr-Nas15 expression level in notochord during entire lumen formation process by the promoter analysis. At 19 hpf (mid gastrula stage), an extracellular lumen initial formed between two adjacent notochord cells, while at 23 hpf and 24 hpf, the lumen undergoes continuous expansion and tilts ( Figure 2C ). The upstream 3,000 bps were cloned and fused with eGFP and electroporated into the Ciona fertilized eggs. Before the lumen formation (19 hpf), we did not observe any fluorescent signal in the notochord. But at 21 hpf (late tailbud stage) and 23 hpf, corresponding to the lumen formation and expansion phase, the eGFP signal was observed in the notochord ( Figure 2D ). Furthermore, to explore the subcellular localization of Cr-Nas15, we prepared a Murine polyclonal antibody with the full length of Cr-Nas15 as antigen. To validate the specificity of the antibody, the 23 hpf staged Ciona larvae were collected to preform total protein extraction. The western blot with Cr-Nas15 antibody showed a band at expected size over 40 KDa, suggesting the high efficacy of self-prepared Cr-Nas15 antibody ( Figure 2E ). We thus performed the Cr-Nas15 immunofluorescence experiment with Ciona larva at different developmental stages. In 19 hpf embryos before the extracellular lumen formation, the Cr-Nas15 has a low signal density in notochord, and localized at the pre-apical membrane of each notochord cells. Then, the Cr-Nas15 signal was increased and localized at the apical domain with the lumen formation in 23 hpf. In 24 hpf, in which stage the notochord cells starting a bi-directional migration, Cr-Nas15 also enriched at the apical membrane of each notochord cells ( Figure 2F ). The Cr-Nas15 immunofluorescence results confirmed that the Cr-Nas15 expression is increasing with the lumen formation, and indicated a special localization at the apical domain near the extracellular lumen. Together with the results of RT-qPCR, promoter analysis and immunofluorescence, we described the Cr-Nas15 spatio-temporal expression profile. The Cr-Nas15 was expressed at the notochord in lumen formation stages from 19 to 24 hpf, and localized at the apical domain, spatial closing to the extracellular lumen, suggesting a potential function of Cr-Nas15 in lumen formation.

To explore the function of Cr-Nas15, actinonin, which binds at the catalytic site of astacin metalloproteinase as a competitive inhibitor (Talantikite et al., 2018). was used to block the function of all astacin metalloproteinase. In the actinonin-treated group, the embryos were cultured in sea water containing 1.2 μM DMSO dissolved actinonin from 16 hpf, while the control group was cultured in sea water with the same dose of DMSO. Then, the embryos were collected and fixed at 21 and 23 hpf stages, respectively. After dying with phalloidin to visualize the cell boundary, we observed the shape of extracellular lumen. In both 21 and 23 hpf, the lumen size in the actinonin-treated embryos obviously smaller than the control group ( Figure 3A ). So, we measured the diameter of the lumen in 23 hpf. Compared with the control group, the lumen diameter in actinonin-treated group was significantly smaller, both in 21hpf and 23 hpf ( Figure 3B ). Besides, the lumen shape was also abnormal in actinonin -treated embryos. Compared with the control group, the lumen in actinonin-treated embryos was not homogeneous growing, but expanded bialy towards one neighbor notochord cell ( Figure 3A ). These results indicate that the astacin metalloproteinase malfunction not only influence the Ciona notochord lumen formation, but also inhibit the lumen expansion process.

To further confirm the function of Cr-Nas15 involved in the notochord lumen formation, we constructed several dominant negative (DN) versions of Cr-Nas15 driven by brachyury promoter, which drives the constructs to specifically express in notochord tissue in Ciona larvae. According to previous studies, the Cr-Nas15-H234D, H238Y and H244Y lead to Nas15 loss of function by blocking the Zn binding, and the E235A lead to the Nas15 catalytic activity malfunction without influencing the Zn and substrate binding (Franco et al., 2005). The Cr-Nas15-DN and wild type (WT) expression plasmids were electroporated into the Ciona embryo, respectively. The results showed that the Cr-Nas15-WT was enriched at the notochord apical domain. However, some signals also appeared in the cytoplasm and around the contractile ring, which may be the mis-localization because of the overexpression. The notochord lumen in Cr-Nas15-WT overexpression group expanded normally, but all four DN versions led to abnormal lumen formation ( Figure 3C ). We measured the lumen diameter in all four DN version groups and the Cr-Nas15-WT group. The statistical results showed that the volume of lumen in each DN groups were significantly smaller than which in Cr-Nas15-WT group, phenocopying the inhibitor treatment ( Figure 3D ). Interestingly, the lumen shape in DN overexpression groups also different from the WT group. All four DN versions led to an asymmetrical lumen expansion. Due to the mosaic expression of Ciona, the Cr-Nas15-DN only expressed in several notochord cells. The lumen between a Cr-Nas15-DN overexpression positive cell and a negative cell always asymmetrically expanded into the negative cell ( Figure 3C ). These results reveal the essential function of Cr-Nas15 in Ciona notochord lumen expansion.

However, because of the side effect of dominant negative overexpression, we decided to perform a Cr-Nas15 knock out (KO) using the CRISPR/Cas9 system. Due to the important function of Zinc-binding metalloprotease motif, we designed three guiding RNA around this motif area ( Figure 3E ). To improve the KO efficacy, all the three guiding RNA plasmids were mixed to electroporate together with Cas9 expression plasmid. We also electroporated the control guiding RNA plasmid together with Cas9 expression plasmid as a control group. Then, the genomic DNA of KO group and control group, respectively were extracted for T7 Endonuclease I efficacy test. The knockout efficiency was determined to be approximately 16.37% by calculating the stripe grayscale values, confirming the efficiency of Cr-Nas15 gene KO ( Figure 3F ). So, we collected the Cr-Nas15 notochord specific KO embryos and control embryos. After the phalloidin staining, we observed the phenotype of notochord lumen formation. In Cr-Nas15 gene KO, 23 of 26 embryos have an abnormal lumen formation, while in control group, there are only 3 of 15 embryos have an abnormal lumen formation ( Figure 3G ). We also measured the diameter of the lumen in both Cr-Nas15 gene KO group and control group. The statistical analysis showed that the lumen in Cr-Nas15 gene KO group were significantly smaller than control group, phenocopying the inhibitor treatment and dominant negative overexpression results ( Figure 3H ). This result confirms that the Cr-Nas15 is an important factor regulating the Ciona notochord lumen formation.

To further find out the working mechanism of Cr-Nas15, we performed a yeast two-hybrid (Y2H) screening assay. The Protein Phosphatase 2 Catalytic Subunit Alpha (PPP2CA) was found as a candidate protein interacting with Cr-Nas15 ( Figure 4A ). To confirm the interaction between Cr-Nas15 and the PPP2CA, we carried out the CO-IP experiments. When the Nas15-GFP and PPP2CA-HA both existed, a band could be found in the IP group ( Figure 4B ). These results indicate the interaction between Cr-Nas15 and PPP2CA, suggesting the Cr-Nas15 and PPP2CA may form a complex functioning together. To confirm the Cr-Nas15 and PPP2CA formed a complex intracellular, we co-expressed the PPP2CA-eGFP and Cr-Nas15-tdTomato in Ciona notochord. The imaging data showed that the PPP2CA-eGFP and Cr-Nas15-tdTomato significantly co-localized together at the apical domain of notochord cell ( Figure 4C ). Together with these data, the PPP2CA and Cr-Nas15 formed a complex around the notochord apical domain. Then, we performed a co-electroporation of PPP2CA-C201R or PPP2CA WT with Cr-Nas15 fusion protein. The Cr-Nas15 showed a lower enrichment on apical membrane in PPP2CA-C201R group, suggesting the PPP2CA involved in Nas15 apical localization ( Figures 4D, E ). To test if the interaction between the PPP2CA and Cr-Nas15 is essential for the lumen formation, we constructed the PPP2CA-C201R as a dominant negative and overexpressed it in Ciona notochord, to observe if it can lead to same phenotype with Cr-Nas15 loss-of-function. The result showed that the lumen in PPP2CA-C201R overexpressed group is significantly smaller than the control group ( Figure 4F ), phenocopying the Cr-Nas15 loss-of-function. The measurement of lumen diameter also confirmed the smaller lumen phenotype of PPP2CA-C201R overexpression ( Figure 4G ). The identification of multiple phosphorylation sites within Nas15 implies a potential regulatory role of PPP2CA in modulating the phosphorylation levels of Nas15 ( Figure 4H ). In conclusion, the PPP2CA is an interaction protein of Cr-Nas15, which is essential for Ciona notochord lumen formation.