Fisheries and aquaculture provide important sources of food and income, and their importance continues to grow as wild fish stocks decline (Jennings et al., 2016; Belton et al., 2018). Body color is an essential phenotypic trait that has implications for predator avoidance and thermoregulation in animals (Leclercq et al., 2010). In fish, body color can be an indicator of the quality and economic value of fish, as fish with brighter, more vibrant colors are generally preferred by consumers (Vissio et al., 2021). Several genes, miRNAs, and metabolites that affect body color have been identified in previous studies (Curran et al., 2009; Zhu et al., 2021; Hao et al., 2022a). For example, Forkhead box D3 (Foxd3) controls melanophore specification in the zebrafish neural crest by regulation of microphthalmia transcription factor (Mitf) which upstream of multiple genes necessary for melanin production (Curran et al., 2009). Zhu et al. (2021) reported that low-density lipoproteins receptor adapter 1, guanosine triphosphatase, and regulator of G-protein signaling genes and metabolites including tyrosine, lecithin, prostacyclin may contribute to skin color differences in P. leopardus. A total of 158 differentially expressed miRNAs were identified in the skin color differentiation of red tilapia and among them, miR-138-5p and miR-722 were predicted to play important roles in regulating the pigmentation process (Wang et al., 2018a). However, the molecular regulatory mechanisms of these genes have not yet been clarified (Bertolini et al., 2020).

Circular RNAs (circRNAs) are a large class of non-coding RNAs with covalent bonds that link the 3´ and 5´ ends by back splicing; they are involved in the regulation of various biological processes in eukaryotic cells (Zhang et al., 2014). CircRNAs are highly stable and resistant to degradation by ribonucleases because they lack free 3′ or 5´ ends (Lasda and Parker, 2016). CircRNAs perform their functions in various biological processes by binding to their source genes via the formation of an R-loop, which results in the cessation of transcription and even changes in phenotype (Xu et al., 2020). CircRNAs might also function as microRNA (miRNA) sponges and hinder miRNA-mediated gene inhibition or silencing through competing endogenous RNA (ceRNA) networks (Salmena et al., 2011; Memczak et al., 2013). Ribonucleoprotein complexes can be formed by the binding of circRNAs to transcription factors and RNA-binding proteins (Aktas et al., 2017; Quan et al., 2021). Several studies have reported that circRNAs function in muscle development (Kotb et al., 2015), transcriptional regulation (Zhang et al., 2013), as well as cellular communication and signal transduction (Li et al., 2015). Studies of circRNAs in fish have mainly examined the roles of circRNAs in the pathogenesis of various diseases in grass carp (Ctenopharyngodon idellus) (He et al., 2017; Liu et al., 2019), flounder (Paralichthys olivaceus) (Xiu et al., 2019), and Nile tilapia (Oreochromis niloticus) (Fan et al., 2019). However, few studies have examined the function of circRNAs in body color formation.

Several miRNAs involved in body color formation have been documented in fish (Wang et al., 2018a; Dong et al., 2012; Kennell et al., 2012; Dong et al., 2020). Multiple miRNAs in red tilapia and common carp have been reported to regulate the expression of body color formation related genes (Wang et al., 2018a; Yan et al., 2013; Luo et al., 2018). Recent studies have shown that messenger RNAs (mRNAs) are not the only molecules regulated by miRNAs (Poliseno et al., 2010; Quinn and Chang, 2016). CircRNAs can mediate decreases in the expression of miRNAs via ceRNA networks by acting as miRNA sponges, and this can have downstream regulatory effects on the expression of target mRNAs (Jiang et al., 2020).

P. leopardus is an important tropical and subtropical fish that varies in body color (Zhu et al., 2021). Several body color-related genes of P. leopardus have been identified in previous transcriptome analyses (Zhu et al., 2021; Hao et al., 2022b; Yang et al., 2020a). However, the specific regulatory mechanism of these genes has not yet been clarified. Here, we identified several circRNAs in black and red P. leopardus to clarify the molecular mechanism by which circRNAs, via their corresponding ceRNAs, regulate the expression of mRNAs involved in body color formation in P. leopardus.

Body color is an essential phenotypic trait that affects the adaptation of P. leopardus to its environment (Chen et al., 2019). An increasing number of studies have examined the regulatory roles of circRNAs in body color formation (Zhu et al., 2020). However, the special mechanism by which circRNAs regulate body color formation in marine animals has not been extensively studied. We identified circRNAs involved in body color formation and constructed circRNA-miRNA-mRNA networks in P. leopardus. Pathway enrichment analyses were clarified to show the functions of circRNAs in body color formation; the results of these analyses provide data that could be used in future studies of the regulatory mechanism by which circRNAs participate in marine animal body color formation. The potential for miRNAs and mRNAs to be used to alter the body color of fish is low given that their half-lives in cells and tissues are short. The potential for circRNAs to be used for this purpose through targeted miRNA silencing is higher than that of linear RNAs given that circRNAs are more chemically stable than linear RNAs (Holdt et al., 2018; Rbbani et al., 2021).

We identified circRNAs in individuals of P. leopardus varying in body color; these circRNAs were classified as exonic circRNAs (one_exon, annot_exon, and exon_intron), intronic circRNAs, and intergenic circRNAs. Exonic circRNAs were the most common, and these findings are consistent with the results of previous studies of tilapia and mice (Jeck et al., 2013; Fan et al., 2019). In tilapia, multiple circRNAs can be derived from a single gene (Fan et al., 2019). In our study, one source gene could generate multiple circRNA isoforms in P. leopardus, and this was associated with alternative back splicing.

CircRNAs are synthesized by back splicing, which differs from the traditional splicing mechanism of linear mRNAs; study of the function of the source mRNAs can provide insights into the functions of circRNAs. A total of 24 DECs were identified via analysis of the expression of DECs between black and red P. leopardus. CircRNAs can be co-expressed with their source genes in the same locus (Huang et al., 2017), and they can play a role in regulating the expression of their source genes by mediating decreases in the quantity of pre-mRNA for canonical splicing (Salzman et al., 2012). In addition, the linear transcripts of the source gene might regulate the expression of circRNAs. The novel_circ_001296 DEC was transcribed from the foxp1b (Dxb_GLEAN_10021390) gene. Previous studies have shown that miR-429 silencing can lead to an increase in Foxd3 expression in vivo and inhibit MITF expression, which leads to decreases in the expression of downstream genes, including the melanin synthesis-related genes tyrosinase (TYR), TYRP1, and TYRP2 (Yan et al., 2013). The novel_circ_000160 DEC was transcribed from the SLC9A7 (Dxb_GLEAN_10021160) gene. Solute carriers (SLCs) play a role in skin pigmentation or body color formation, and many SLC genes are differentially expressed in chicken, tilapia, and leopard grouper individuals varying in color (Hao et al., 2022b; Hoglund et al., 2011; Zhu et al., 2016; Chen et al., 2020). The source genes of the DECs are also involved in protein ubiquitination. E3 ubiquitin-protein ligase CBL (Dxb_GLEAN_10018739) is involved in the process of ubiquitin-mediated protein degradation (Hochrainer et al., 2005), and the up-regulation of the gene encoding this protein in black fish relative to red fish suggests that protein degradation has occurred. However, additional studies are needed to clarify the relationships between source genes and circRNAs.

The activity and expression of miRNAs can be inhibited by circRNAs, and this can alter the expression of the target genes of miRNAs (Hansen et al., 2011; Hansen et al., 2013). Genetics plays a role in trait inheritances and the SNP detection of trait-related genes or regulatory molecules to obtain potential markers for phenotype in breeding has been widely used in aquaculture (Fan et al., 2017; Lei et al., 2019; Yang et al., 2020b). In the present study, differential expression of circRNAs and genes in the two phenotypes is evident, which may suggest a direction for the SNPs detection which were associated with body color formation. Color phenotype is determined by the combination of genetics and environmental factors, and changes in environmental factors can also trigger variations in hormone and neuron secretion, resulting in different skin color responses (Luo et al., 2021). For example, background color signals are transmitted to the brain via vision, and are processed by the central nervous system to act on skin pigment cells to adjust the body color to the background color (Zimmermann et al., 2018). Dietary Cys and Tyr could affect the melanin synthesis pathway in the red tilapia and participate in the skin color differentiation (Wang et al., 2018b). Therefore, it is reasonable to hypothesize that body color changes triggered by environmental factor is partially related to the expression level changes in molecules including circRNAs, miRNAs and mRNAs through participating in or regulating pigmentation related pathway.

GO analysis of DEC source genes revealed that genes related to body color variation were involved in Golgi apparatus part, Golgi membrane, and Golgi apparatus. Tyrosinase family proteins involved in melanin biosynthesis might be trafficked through the Golgi apparatus and then to type II melanosomes, where they promote melanin deposition (Dooley et al., 2013). Therefore, novel_circ_001363 might regulate the expression of its source gene to modify the transport of TYR and TYRP1 through the Golgi apparatus (Hu et al., 2022). Black body color is highly associated with melanin synthesis (Henning et al., 2013); in our study, DEGs in the ceRNA network were enriched in melanin synthesis-related pathways, including melanogenesis and tyrosine metabolism (Kelsh, 2004).

RNAs with the same miRNA-binding site may compete with each other for miRNA binding and function as ceRNAs. Previous studies have shown that circRNAs play a regulatory role in determining the color of the hair on the skin of mice via the circRNA-miRNA-mRNA ceRNA network (Zhu et al., 2018; Zhu et al., 2020). In our study, the ceRNA network was constructed, and related DEGs were enriched in several immune pathways, indicating that circRNAs might regulate the immunity of fish varying in body color. Higher levels of melanin promote immune defense and wound healing in Amphiprion percula (Smith et al., 2018). Darker eumelanic individuals show higher immune activity than lighter individuals (McGraw, 2005; Ducrest et al., 2008). However, additional studies are necessary to determine whether these genes are targeted by circRNAs or whether they affect the body color of P. leopardus.

A total of 1,424 novel circRNAs were identified in black and red P. leopardus. A total of 24 circRNAs were detected as DECs. Integrated analysis of DECs, DEMs, and DEGs yielded a ceRNA network with 19 significant miRNA-circRNA and miRNA-mRNA pairs. Pathway enrichment of ceRNA network DEGs revealed that they may participate in the regulation of the immune response and melanin-related metabolism. The results indicated the potential functions of circRNAs and the ceRNA network in body color formation and will aid future efforts to breed P. leopardus individuals with colors.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

The animal study was reviewed and approved by the respective Animal Care and Use Committee of Guangdong Ocean University.

CZ designed and supervised the research. RH conducted the research, analyzed the data and wrote the manuscript. XZ, CT, YH nd GL contributed to the final writing of the manuscript. All authors contributed to the article and approved the submitted version.