MiR-92 Family Members Form a Cluster Required for Notochord Tubulogenesis in Urochordate Ciona savignyi

MicroRNAs are frequently clustered in the genome and polycistronically transcribed, regulating targeted genes in diverse signaling pathways. The miR-17-92 cluster is a typical miRNA cluster, playing crucial roles in the organogenesis and homeostasis of physiological processes in vertebrates. Here, we identified three miRNAs (csa-miR-92a, csa-miR-92b, and csa-miR-92c) that belonged to the miR-92 family and formed a miRNA cluster in the genome of a urochordate marine ascidian Ciona savignyi. Except for miR-92a and miR-92b, other homologs of the vertebrate miR-17-92 cluster members could not be identified in the Ciona genome. We further found that the mature sequences of urochordate miR-92 family members were highly conserved compared with the vertebrate species. The expression pattern revealed that three miR-92 family members had consistent expression levels in adult tissues and were predominantly expressed in heart and muscle tissue. We further showed that, at the embryonic and larval stages, csa-miR-92c was expressed in the notochord of embryos during 18–31 h post fertilization (hpf) by in situ hybridization. Knockout of csa-miR-92c resulted in the disorganization of notochord cells and the block of lumen coalescence in the notochord. Fibroblast growth factor (FGF), mitogen-activated protein kinase (MAPK), and wingless/integrated (Wnt)/planar cell polarity (PCP) signaling pathways might be involved in the regulatory processes, since a large number of core genes of these pathways were the predicted target genes of the miR-92 family. Taken together, we identified a miR-92 cluster in urochordate Ciona and revealed the expression patterns and the regulatory roles of its members in organogenesis. Our results provide expression and phylogenetic data on the understanding of the miR-92 miRNA cluster’s function during evolution.


Introduction
MicroRNAs (miRNAs), a class of small, endogenous, noncoding RNAs [1,2], play an essential role in multiple biological processes by regulating gene expression at the posttranscriptional level [3][4][5]. Approximately 37% of human miRNA genes localize as clusters in the human genome [6]. MiRNAs in a cluster are transcribed together with physically adjacent miRNAs and usually have similar functions or participate in regulating the same signaling pathway [7].
The miR-17-92 cluster is one of the best-studied miRNA clusters and was the first recognized as an oncogenic gene in human B-cell lymphomas [8]. This cluster localizes at chromosome 13q31 and comprises six mature miRNAs, including mir-17, mir-18a, mir-19a, The DIG-labeled locked nucleic acid (LNA) probes were synthesized by JIE LI BIOL-OGY. Ascidian embryos at 10,18,21,24,31, and 42 h postfertilization (hpf) were collected and fixed in 4% paraformaldehyde (PFA) overnight at 4 • C. The embryos were permeabilized by digesting with 12-µg/mL proteinase K at 37 • C for 15 min. Subsequently, in situ hybridization was performed at a temperature 20 • C below the probe melting temperature for 18 h using csa-miR-92c LNA probes in a hybridization solution. After hybridization, samples were washed in gradient saline-sodium citrate at the hybridization temperature. Hybridization signals were detected using alkaline phosphatase-conjugated digoxigenin antibody (Roche, Basel, Switzerland) at a 1:2000 dilution. Embryos were stained with the BCIP/NBT substrate system (Roche, Basel, Switzerland) and visualized under the microscope.

Constructions and Electroporation
Three target sequences of CRISPR/Cas9 against csa-miR-92c were designed by CRISPRdirect (http://crispr.dbcls.jp, accessed on 1 December 2019). The sequences of selected single guide RNAs (sgRNAs) are listed in Table S1, and the sequence of control sgRNA was designed according to the previous study [33]. Based on the target sequences of CRISPR/Cas9, all sgRNAs were synthesized and cloned into the vector of Cr-U6>sgRNA(F+E) (Addgene number: 59986) for expressing the sgRNA, respectively. The CRISPR/Cas9 system included the PCR product of sgRNA and the plasmid of Cs-Brachyury(3k)>NLS::Cas9::NLS::P2A::mCherry. Electroporation was performed according to the method described previously, with some modifications [34]. Briefly, the PCR product of sgRNA (50 µL), the plasmid of Cs-Brachyury(3k) >NLS::Cas9::NLS::P2A::mCherry (30 µg), the embryos (300 µL), and the electroporation buffer (420 µL) were electroporated into fertilized eggs at the following conditions: the capacitance was 1500 µF, the voltage was 50 V, and the time constant was within the range of 15-20 ms. After electroporation, the eggs were transferred to a fresh seawater agarose dish, and the embryos were allowed to develop at 18 • C to the desired stages (16 hpf, 21 hpf, and 31 hpf) for confocal observation.

Validation of the Efficiency of the gRNAs
The PCR product of sgRNA and the plasmid of Cr-EF1α>NLS::Cas9::NLS and Cr-EF1α>NLS::Cas9::NLS::mCherry were co-electroporated into the fertilized eggs. After electroporation, embryos were cultured for about 12 h at 18 • C. The embryos with red positive signals were collected and lysed by a direct PCR lysis buffer at 60 • C for 10 min. PCR was performed from the lysate using Phanta ® Max Super-Fidelity DNA Polymerase (Vazyme, Nanjing, China) as follows: 55 • C annealing temperature for 15 s and 72 • C extension temperature for 30 s, 35 cycles. PCR products were purified by the GeneJET Gel Extraction Kit (Thermo Fisher, Waltham, Massachusetts, USA). Two hundred-nanogram PCR products in the knockout (KO) group and 200-ng control PCR products were diluted to 17 µL by ddH 2 O and mixed with 2-µL T7 reaction buffer (Vazyme, Nanjing, China) in a 200-µL PCR tube and incubated with 1-µL T7 endonuclease I (Vazyme, Nanjing, China) at 37 • C for 30 min. The products were detected on a 2% agarose gel at 100 V for 30 min, and the image of the gel was analyzed by ImageJ (National Institutes of Health). Cleavage efficiency was determined according to the previously published protocol [33]. The PCR product of mutant gDNA was ligated into the pEasy Blunt-3 vector (Transgen, Beijing, China) for Sanger sequencing (GENEWIZ, Suzhou, China).

Immunofluorescences
The fixed embryos were washed three times in PBST and then incubated with 1/100 dilution of phalloidin overnight at 4 • C. After removing the antibody solution, the samples were washed three times in PBST at room temperature. DAPI staining was performed in the dark for 30 min. The samples were imaged using a confocal laser scanning microscope.

Bioinformatics Analysis
The precursor and mature sequence of miR-92 from different species were extracted from the miRNA database miRbase22.1 (http://www.mirbase.org/, accessed on 14 August 2020). The Ensembl database (http://asia.ensembl.org/, accessed on 14 August 2020) was used to analyze the distribution of the miR-92 family or cluster members in the genome. The secondary structure for "csa-mir-92" was predicted using miR-Deep2 and miReap (https://sourceforge.net/projects/mireap/, accessed on 14 August 2020). For the evolutionary analysis, the multiple sequence alignment was performed using CLC Main Workbench 5. The phylogenetic tree was constructed using the Maximum-Likelihood (ML) method by the Molecular Evolutionary Genetic Analysis of MEGA X.
For target gene prediction, 3 UTRs of C. savignyi mRNAs obtained from previous RNA-seq data [24] were used as potential target sequences. MiRNA-binding sites were predicted using miRanda and targetscan 7.1 (targetscan.org, accessed on 24, August, 2019). The underlying functions and vital pathways of the target gene candidates were analyzed by Gene Ontology (GO) functional annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses (www.kegg.jp/kegg/kegg1.html, accessed on 14 August 2020). The OmicShare tools were used to test the statistical enrichment of the target gene candidates in the GO and KEGG pathways. The calculated p-value went through False Discovery Rate (FDR) Correction, taking FDR ≤ 0.05 as a threshold.

Transcriptome Sequencing
Ascidian embryos and larvae were collected at 18, 21, and 42 hpf and were then immediately frozen in liquid nitrogen. Total RNA was extracted with RNAiso Reagent (Takara, Beijing, China). in total, 3 µg of total RNA per sample was used as the input material for the construction of small RNA libraries. Libraries were generated using NEBNext ® Multiplex Small RNA Library Prep Set for Illumina ® (NEB, Beijing, China) following the manufacturer's instructions and were sequenced on an Illumina Hiseq 2500/2000 platform at Novogene, China.

Expression Profiles of the csa-miR-92 in C. savignyi
To understand the expression patterns of the csa-miR-92 family members, we validated the expression profiles of three csa-miR-92 members at different embryonic stages (14 hpf, 18 hpf, 21 hpf, 24 hpf, 28 hpf, 32 hpf, and 42 hpf) and various tissues (gonad, intestines, heart, blood, muscle, and stomach) using qPCR. The results showed that three csa-miR-92 members were expressed in all the examined developmental stages. Their expression levels in 42 hpf larvae dramatically increased (p < 0.01) compared with those at

Expression Profiles of the csa-miR-92 in C. savignyi
To understand the expression patterns of the csa-miR-92 family members, we validated the expression profiles of three csa-miR-92 members at different embryonic stages (14 hpf, 18 hpf, 21 hpf, 24 hpf, 28 hpf, 32 hpf, and 42 hpf) and various tissues (gonad, intestines, heart, blood, muscle, and stomach) using qPCR. The results showed that three csa-miR-92 members were expressed in all the examined developmental stages. Their expression levels in 42 hpf larvae dramatically increased (p < 0.01) compared with those at the earlier stages (14-32 hpf) ( Figure 4A), indicating that csa-miR-92 family members play vital roles in larval tail regression and metamorphosis. Further examinations of the expression profiles in various tissues showed three csa-miR-92 family members that were expressed in all the examined adult tissues and predominantly (p < 0.01) in the heart and muscle ( Figure 4B).
Genes 2021, 12, x FOR PEER REVIEW 9 of 16 the earlier stages (14-32 hpf) ( Figure 4A), indicating that csa-miR-92 family members play vital roles in larval tail regression and metamorphosis. Further examinations of the expression profiles in various tissues showed three csa-miR-92 family members that were expressed in all the examined adult tissues and predominantly (p < 0.01) in the heart and muscle ( Figure 4B). The spatial expression of csa-miR-92c at different larval stages was examined by in situ hybridization ( Figure 5 and Figure S1). The results showed that the expression of csa-miR-92c was distributed throughout the embryo at the early stage (10 hpf) and larval tail regression stage (42 hpf), which is consistent with the results of the qPCR. At the larvae stages 18 hpf, 21 hpf, 24 hpf, and 31 hpf, stronger expression signals of csa-miR-92c were showed in the trunk. They were expressed in the epithelial cells and notochord cells as well ( Figure 5). With the elongation of the tail, the positive signals in notochord cells turned to be weak after the lumen connection ( Figure 5). These results suggested that csa-miR-92c was involved in the regulation of Ciona embryogenesis and larval development, especially in metamorphosis and notochord tubulogenesis. Relative expression of csa-miR-92a, csa-miR-92b, and csa-miR-92c in various tissues (gonad, intestines, heart, blood, muscle, and stomach). The relative expression of each miRNA was calculated using the 2 −ΔΔCt method and normalized to the U6 gene expression. Results represented the mean of triplicate assays with the standard error (mean ± SD). Significance (p < 0.05) was indicated by a, b, c, d, e, f, g, h, i, j, and k. Relative expression of csa-miR-92a, csa-miR-92b, and csa-miR-92c in various tissues (gonad, intestines, heart, blood, muscle, and stomach). The relative expression of each miRNA was calculated using the 2 −∆∆Ct method and normalized to the U6 gene expression. Results represented the mean of triplicate assays with the standard error (mean ± SD). Significance (p < 0.05) was indicated by a, b, c, d, e, f, g, h, i, j, and k.
The spatial expression of csa-miR-92c at different larval stages was examined by in situ hybridization ( Figure 5 and Figure S1). The results showed that the expression of csa-miR-92c was distributed throughout the embryo at the early stage (10 hpf) and larval tail regression stage (42 hpf), which is consistent with the results of the qPCR. At the larvae stages 18 hpf, 21 hpf, 24 hpf, and 31 hpf, stronger expression signals of csa-miR-92c were showed in the trunk. They were expressed in the epithelial cells and notochord cells as well ( Figure 5). With the elongation of the tail, the positive signals in notochord cells turned to be weak after the lumen connection ( Figure 5). These results suggested that csa-miR-92c was involved in the regulation of Ciona embryogenesis and larval development, especially in metamorphosis and notochord tubulogenesis.

Knockout (KO) of csa-miR-92c by CRISPR-Cas9 Resulted in Disorganization of the Notochord Cells and the Failure of Lumen Formation
Among three members of miRNA92, csa-miR-92c was expressed at a relatively higher level compared with the other two members ( Figure 4A). We therefore picked it up to explore its developmental roles in Ciona organogenesis. We designed three gRNAs at three locations of csa-miR-92c, including the mature sequence, the active site of Dicer, and the active site of Drosha, to induce mutation by the CRISPR-Cas9 approach ( Figure  6A and Table S1). After validation of the efficiency of three gRNAs ( Figure S2), we coelectroporated the Cs-Brachyury>NLS::Cas9::NLS::P2A::mCherry and the PCR products of three sgRNAs mixture into the fertilized eggs and examined the phenotypes in gRNAsinjected embryos. The results showed that the arrangement of notochord cells in the gRNA-injected groups was disrupted compared with that in the control groups. In 16 hpfstaged embryos, the Cas9-positive notochord cells in csa-miR-92c gRNA-injected embryos

Knockout (KO) of csa-miR-92c by CRISPR-Cas9 Resulted in Disorganization of the Notochord Cells and the Failure of Lumen Formation
Among three members of miRNA92, csa-miR-92c was expressed at a relatively higher level compared with the other two members ( Figure 4A). We therefore picked it up to explore its developmental roles in Ciona organogenesis. We designed three gRNAs at three locations of csa-miR-92c, including the mature sequence, the active site of Dicer, and the active site of Drosha, to induce mutation by the CRISPR-Cas9 approach ( Figure 6A and Table S1). After validation of the efficiency of three gRNAs ( Figure S2), we co-electroporated the Cs-Brachyury>NLS::Cas9::NLS::P2A::mCherry and the PCR products of three sgRNAs mixture into the fertilized eggs and examined the phenotypes in gRNAs-injected embryos. The results showed that the arrangement of notochord cells in the gRNA-injected groups was disrupted compared with that in the control groups. In 16 hpf-staged embryos, the Cas9-positive notochord cells in csa-miR-92c gRNA-injected embryos became narrow and were squeezed together. The yellow dashed box indicated the disordered notochord cells ( Figure 6B). These cells either failed to complete the cell intercalation process or had the problems with the arrangement of notochord cells after the intercalation stage. With this development, the cas9-positive notochord cells were further failures during lumen formation ( Figure 6B). The yellow-dashed box showed that the lumen connection and tube formation were uncompleted in cas9-positive embryos compared with the control groups. The number and percentage of the csa-miR-92c-KO embryos were shown in Table S3. Overall, these results suggest that csa-miR-92c plays an essential role in the arrangement of ascidian notochord and lumen formation and coalescence in notochord tubulogenesis.

Target Prediction of csa-miR-92 and Gene Ontology and KEGG Enrichment Analysis
MiRNAs function through targeting complementary sequences in mRNA molecules to regulate diverse biological processes. To further know about the molecular mechanisms and biological roles of csa-miR-92 family members, we predicted their target genes by using miRanda and targetscan software. A total of 984 putative target genes were identified. Moreover, we screened the target genes by their expression and collected the downregulated genes ( Figure S3) from 18 hpf to 42 hpf. GO analysis revealed that the target genes were enriched in the biological processes, molecular functions, and cell components ( Figures 7A and S4). The target genes found in the membrane-enclosed lumen or cell junction indicated the possibility that csa-miR-92 modulated the notochord lumen formation via these molecular targets. To explore the signal pathways that csa-miR-92 target genes were involved in, we collected 550 target genes that were annotated with the KEGG database for further analysis. The results showed that these genes were categorized into six parts, including organismal systems, cellular processes, genetic information process, metabolism, human diseases, and the environmental information process (Figures 7B and S5). Furthermore, we found that many important signaling pathways were involved in the putative target genes of csa-miR-92, such as the fibroblast growth factor (FGF), transforming growth factor beta (TGFβ), mitogen-activated protein kinase (MAPK), wingless/integrated (Wnt), nuclear factor kappa-B (NF-κB), nodal growth differentiation factor (Nodal), and planar cell polarity (PCP) signaling pathways ( Figure 7C). These pathway components are likely to represent the indirect targets of Brachyury-induced signaling in the notochord.
Genes 2021, 12, x FOR PEER REVIEW became narrow and were squeezed together. The yellow dashed box indicated th dered notochord cells ( Figure 6B). These cells either failed to complete the cell in tion process or had the problems with the arrangement of notochord cells after th calation stage. With this development, the cas9-positive notochord cells were furt ures during lumen formation ( Figure 6B). The yellow-dashed box showed that th connection and tube formation were uncompleted in cas9-positive embryos co with the control groups. The number and percentage of the csa-miR-92c-KO e were shown in Table S3. Overall, these results suggest that csa-miR-92c plays an e role in the arrangement of ascidian notochord and lumen formation and coalesc notochord tubulogenesis.    Figure S5). Furthermore, we found that many important signaling pathways were involved in the putative target genes of csa-miR-92, such as the fibroblast growth factor (FGF), transforming growth factor beta (TGFβ), mitogen-activated protein kinase (MAPK), wingless/integrated (Wnt), nuclear factor kappa-B (NF-κB), nodal growth differentiation factor (Nodal), and planar cell polarity (PCP) signaling pathways ( Figure 7C). These pathway components are likely to represent the indirect targets of Brachyury-induced signaling in the notochord. The cellular processes in the KEGG pathway annotation. The abscissa described five cellular processes, including transport and catabolism, cell growth and death, cellular community eukaryotes, cell motility, and cellular community prokaryotes. The ordinate indicated the number of genes in the cellular processes. (C) The signaling pathways related to the putative target genes of csa-miR-92, including fibroblast growth factor (FGF), transforming growth factor beta (TGFβ), mitogen-activated protein kinase (MAPK), wingless/integrated (Wnt), nuclear factor kappa-B (NF-κB), nodal growth differentiation factor (Nodal), and planar cell polarity (PCP) signaling. Red rectangles represented the predicted target genes of csa-miR-92; black rectangles represented the genes in a pathway. Csa-miR-92 was shaded by rectangles filled with blue color; black arrows, black line bar, and dashed arrow denoted positive regulation, negative regulation, and the biological regulation process, respectively.

Discussion and Conclusions
In this study, we found three miRNAs (csa-miR-92a, csa-miR-92b, and csa-miR-92c) that formed a miRNA cluster in the genome of C. savignyi and were highly conserved across species. The miR-92 cluster was predominantly expressed in the notochord cells in Ciona embryos during 18-31 hpf. Moreover, the knockout of the csa-miR-92c resulted in the disarrangement of the notochord cells, which might be regulated through the Wnt/PCP or NF-κB signaling pathways. These data together indicated that the miR-92 cluster plays vital roles in Ciona notochord development during embryogenesis. miRNAs are small noncoding RNAs. Many of them are clustered in the genomes of animals, which can be independently or simultaneously transcribed into a single polycistronic transcription [6]. The miR-17-92 cluster is a typical highly conserved polycistronic miRNA gene [36]. In humans, the miR-17/92 cluster together with its two paralogs comprises miRNAs: miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, miR-92a-1, miR-106b, miR-93, miR-25, miR-106a, miR-18b, miR-20b, miR-19b-2, miR-92a-2, and miR-363 [9]. However, miR-92c or miR-92d were identified in the invertebrate genome but not in vertebrates. We found three miRNAs (csa-miR-92a, csa-miR-92b, and csa-miR-92c) that formed a miRNA cluster in the Ciona genome, while other homologs of vertebrate miR-17-92 cluster members could not be identified. Other members, except for miR-92a/b, of the vertebrate miR-17-92 cluster also disappeared in many invertebrates. A previous study showed that, as the species diversity increased during evolutionary history, the varieties of distribution of miRNA genes also increased, catering to the individual and interactive effects of multiple complex evolutionary forces [35]. The low conservation of the miR-17-92 clusters between vertebrates and invertebrates might result from genome duplication in vertebrates.
Since the Ciona genome does not have the intact members of miR-17-92 cluster as vertebrates, we explored whether three members of miR-92 cluster identified in Ciona could form a cluster and effectively perform similar functions as the vertebrate miR-17-92 clusters do. Our results showed that the csa-miR-92 was ubiquitously expressed in different tissues and at different stages of embryo, identical to the miR-17-92-cluster, which was highly expressed from the embryonic cells to adulthood [37]. Noteworthy, three Ciona miR-92 family members had consistent expression levels in adult tissues or at different embryonic stages, indicating that they either were regulated by the same upstream signals or processed similar stability. In a word, the Ciona miR-92 cluster members identical conserved seed sequences and similar expression levels. Thus, we speculated that three miR-92 cluster members cooperated in the regulation of the same signaling pathways so that their functions were more effective in controlling orderly biological activity. However, it is worth noting that the six miRNA members in the vertebrate miR-17-92 cluster were often expressed at different levels. The transcriptional level of miR-92a was the most highly expressed members compared to that of the other miRNA members [38,39]. The cooperation of the miR-17-92 cluster members occurred in modulating a specific pathway, such as TGF-β signaling [40,41]. Thus miR-17-92 cluster had overlapping or complementary functions. We identified 984 putative target genes of csa-miR-92 in this study, which involved multiple biological processes. By comparison, between predicted target genes of csa-miR-92 family and reported targets of the mammalian miR-92 cluster, we found that the seed sequence of csa-miR-92 was conserved, but its target genes were diverse from vertebrates. However, csa-miR-92s might sometimes maintain their regulation of a conserved signal pathway by targeting different genes, even if they lost the interaction with a conserved target in vertebrates, which suggested that conservation of miRNA mediated transcriptional regulation system might exceed sequence level and live up to progress level.
Ciona notochord is made of 40 cells, arranged in a single row in the midline of the larval tail. In the early development of notochord, cells intercalate into a column during convergent extension. Then, extracellular pocket lumen forms, expands, and merges into a single lumen within the notochord [30,42,43]. Here, we observed that csa-miR-92c was expressed in the notochord cells at different embryonic stages. Furthermore, the knockout of csa-miR-92c by CRISPR/Cas9 resulted in a disarrangement of the notochord cells and failure of lumen formation. To detect the molecular mechanism of this phenotype, we screened the target genes of miR-92. Notably, some molecular networks underlying notochord intercalation and extension movements were identified as the target candidates, including FGF [44][45][46], Wnt/PCP [47], and NF-κB pathways and solute carrier family 26 (SLC26) [48]. These results suggested that the csa-miR-92 cluster was involved in the regulation of notochord development via multiple pathways. In general, our study revealed the phylogenesis, expression, and the roles of the miR-92 cluster in Ciona organogenesis. The underlying mechanisms and various additional roles need to be further explored in the future.

Institutional Review Board Statement:
The animals used in this study are marine invertebrates, which is not applicable to the institutional review board statement in our university.