The CXC Chemokine Receptors in Four-Eyed Sleeper (Bostrychus sinensis) and Their Involvement in Responding to Skin Injury

CXC Chemokine signaling plays an important role in wound healing. The four-eyed sleeper (Bostrychus sinensis) is a commercially important marine fish, which is prone to suffer skin ulceration at high temperature seasons, leading to mass mortality of fish in aquaculture farms. The genetic background related to skin ulceration and wound healing has remained unknown in this fish. Herein, we identified 10 differentially expressed Bostrychus sinensis CXC chemokine receptors (BsCXCRs) in skin ulcerated fish by de novo transcriptome sequencing. The transcripts of these BsCXCRs were classified in seven types, including BsCXCR1a/1b, BsCXCR2, BsCXCR3a1/3a2, BsCXCR4a/4b, and BsCXCR5-7, and BsCXCR6 was the first CXCR6 homologue experimentally identified in teleost fish. These BsCXCRs were further characterized in gene and protein structures, as well as phylogenetics, and the results revealed that BsCXCRs have expanded to divergent homologues. Our results showed that, in healthy fish, the BsCXCR transcripts was mainly distributed in the muscle and immune related organs, and that BsCXCR1a/1b proteins located in the cytomembrane, BsCXCR4a/4b/5/6 in the cytomembrane and perinuclear region, and BsCXCR3a1/3a2/7 in the cytomembrane, perinuclear region, and nuclear membrane, respectively. In skin injured fish, the transcripts of all BsCXCRs were transiently increased within one hour after injury, suggesting the involvement of BsCXCRs into the early inflammatory response to skin injury in the four-eyed sleeper. These results are valuable for understanding the evolutionary events of fish CXCR genes and provide insights into the roles of CXCR family in fish skin injury.


Introduction
Commercial aquaculture is one of the most important sources of food, especially of high-quality protein from farmed fish. However, fish culture has become increasingly vulnerable to various aquatic environmental stressors, including pathogens and pollution, in the intensive cultivation system. The skin is the first barrier against these factors [1,2], and skin damage of fish may result in pathogen infection, which eventually causes fish death. Most cutaneous injuries accompanied by vascular damages trigger an acute inflammatory response [3] and are followed by wound healing, which is a dynamic, interactive process consisting of three phases, inflammation, tissue formation, and tissue remodeling, with the involvement of various soluble mediators, extracellular matrix, resident cells, and leukocytes [4,5]. Inflammation is the first critical step of wound healing, which happens in the first several hours of injury for preventing ongoing blood and fluid losses, as well as identified 10 CXCR transcripts, and qPCR examination indicated the significant up-regulation of BsCXCR1a/1b/2/3a1/3a2/4a/6 and down-regulation of BsCXCR4b/5/7 in skin ulcerated fish compared to healthy fish.

Sequence Analysis of CXCRs
Several studies have reported that CXCR6 is absent in teleost fish [18][19][20]. However, of these 10 four-eyed sleeper CXCR transcripts, one was predicted as CXCR6. To obtain an explicit annotation, a phylogenetic tree was constructed based on previously identified amino acid sequence of vertebrate CXCR homologs. Overall, the phylogenetic analysis supported the identification because all BsCXCRs, except for BsCXCR6, were clustered together with corresponding teleost fish CXCR homologues. The BsCXCR6 was clustered together with cartilaginous fish CXCR6 homologues ( Figure 2). Therefore, BsCXCR6 is the first experimentally identified CXCR6 gene in teleost fish. Additionally, we found that many teleost fish, including four-eyed sleeper, had at least two copies of CXCR1/3/4 homologues.

Sequence Analysis of CXCRs
Several studies have reported that CXCR6 is absent in teleost fish [18][19][20]. However, of these 10 four-eyed sleeper CXCR transcripts, one was predicted as CXCR6. To obtain an explicit annotation, a phylogenetic tree was constructed based on previously identified amino acid sequence of vertebrate CXCR homologs. Overall, the phylogenetic analysis supported the identification because all BsCXCRs, except for BsCXCR6, were clustered together with corresponding teleost fish CXCR homologues. The BsCXCR6 was clustered together with cartilaginous fish CXCR6 homologues ( Figure 2). Therefore, BsCXCR6 is the first experimentally identified CXCR6 gene in teleost fish. Additionally, we found that many teleost fish, including four-eyed sleeper, had at least two copies of CXCR1/3/4 homologues.
Subsequently, we characterized gene and protein structures of BsCXCRs family. The exon-intron arrangements are highly variable in BsCXCRs genes, even in duplicated paralogs ( Figure 3). Subsequently, we characterized gene and protein structures of BsCXCRs family. The exon-intron arrangements are highly variable in BsCXCRs genes, even in duplicated paralogs ( Figure 3). BsCXCR1a has two exons and one intron, whereas BsCXCR1b is intronless; BsCXCR4a has two exons and one intron, whereas BsCXCR4b has four exons and three introns. BsCXCR2/3a1/3a2 have two exons and one intron, and BsCXCR5/6/7 are all intronless. Multiple alignments indicated that all BsCXCR proteins have seven transmembrane domains (TM), three extracellular loops (ECL), three intracellular loops (ICL), and a DRY signature motif (Figure 4).  Subsequently, we characterized gene and protein structures of BsCXCRs family. The exon-intron arrangements are highly variable in BsCXCRs genes, even in duplicated paralogs ( Figure 3). BsCXCR1a has two exons and one intron, whereas BsCXCR1b is intronless; BsCXCR4a has two exons and one intron, whereas BsCXCR4b has four exons and three introns. BsCXCR2/3a1/3a2 have two exons and one intron, and BsCXCR5/6/7 are all intronless. Multiple alignments indicated that all BsCXCR proteins have seven transmembrane domains (TM), three extracellular loops (ECL), three intracellular loops (ICL), and a DRY signature motif (Figure 4). BsCXCR1a has two exons and one intron, whereas BsCXCR1b is intronless; BsCXCR4a has two exons and one intron, whereas BsCXCR4b has four exons and three introns. BsCXCR2/3a1/3a2 have two exons and one intron, and BsCXCR5/6/7 are all intronless. Multiple alignments indicated that all BsCXCR proteins have seven transmembrane domains (TM), three extracellular loops (ECL), three intracellular loops (ICL), and a DRY signature motif (Figure 4). In addition, the number of N-glycosylation sites was also variable in these BsCXCRs (Table 1).  In addition, the number of N-glycosylation sites was also variable in these BsCXCRs (Table 1). By using PyMol, the 3D-structures of BsCXCR proteins were predicted to reveal the structural similarities and differences. The prediction showed that all BsCXCRs had seven α-helices representing the seven TM domains and that BsCXCR3a2/4a/4b/6/7 had two additional β-sheets ( Figure 5E,F,I,J). BsCXCR1a/1b/2/3a1/3a2/4a/5/6 proteins were existed as monomer, whereas BsCXCR4b/7 proteins were homodimers, and BsCXCR1a/1b/2/3a1 and BsCXCR4b/7 had relatively highly identical structures, respectively ( Figure 5G,J). By using PyMol, the 3D-structures of BsCXCR proteins were predicted to reveal the structural similarities and differences. The prediction showed that all BsCXCRs had seven α-helices representing the seven TM domains and that BsCXCR3a2/4a/4b/6/7 had two additional β-sheets ( Figure 5E,F,I,J). BsCXCR1a/1b/2/3a1/3a2/4a/5/6 proteins were existed as monomer, whereas BsCXCR4b/7 proteins were homodimers, and BsCXCR1a/1b/2/3a1 and BsCXCR4b/7 had relatively highly identical structures, respectively ( Figure 5G,J).

Cellular Localizations and Tissue Distributions of BsCXCRs
To characterize the subcellular localization of BsCXCR proteins, GFP fusion plasmids of pEGFP-N3-BsCXCR were transiently transfected in HEK293T cells, and the cytomembrane was labeled by immunofluorescence staining of β-catenin. BsCXCRs' localizations were classified in three types. BsCXCR1a/1b were completely co-localized with β-catenin, demonstrating that they were specifically localized at the cytomembrane ( Figure 6A-F); BsCXCR3a1/3a2/7 were expressed at the cytomembrane, perinuclear region, and nuclear membrane ( Figure 6G-O); and BsCXCR4a/4b/5/6 were expressed in the cytomembrane and perinuclear region ( Figure 7).

Cellular Localizations and Tissue Distributions of BsCXCRs
To characterize the subcellular localization of BsCXCR proteins, GFP fusion plasmids of pEGFP-N3-BsCXCR were transiently transfected in HEK293T cells, and the cytomembrane was labeled by immunofluorescence staining of β-catenin. BsCXCRs' localizations were classified in three types. BsCXCR1a/1b were completely co-localized with β-catenin, demonstrating that they were specifically localized at the cytomembrane ( Figure 6A-F); BsCXCR3a1/3a2/7 were expressed at the cytomembrane, perinuclear region, and nuclear membrane ( Figure 6G   The distributions of BsCXCR transcripts in adult four-eyed sleeper tissues were examined by qPCR. The results indicated that BsCXCR transcripts were abundant in the spleen, liver, muscle, and brain ( Figure 8). Even though the duplicated paralogs of BsCXCR1/3/4 had similar tissue distributions, the expression exhibited significantly biased in adult tissues. The expression levels of BsCXCR1b and BsCXCR3a2 were significantly higher (2-7-folds) compared to BsCXCR1a and BsCXCR3a1, respectively ( Figure  8A,C); and the expression level of BsCXCR4b was remarkably higher (33-685-folds) compared to BsCXCR4a ( Figure 8D). The distributions of BsCXCR transcripts in adult four-eyed sleeper tissues were examined by qPCR. The results indicated that BsCXCR transcripts were abundant in the spleen, liver, muscle, and brain ( Figure 8). Even though the duplicated paralogs of BsCXCR1/3/4 had similar tissue distributions, the expression exhibited significantly biased in adult tissues. The expression levels of BsCXCR1b and BsCXCR3a2 were significantly higher (2-7-folds) compared to BsCXCR1a and BsCXCR3a1, respectively ( Figure 8A,C); and the expression level of BsCXCR4b was remarkably higher (33-685-folds) compared to BsCXCR4a ( Figure 8D).

Dynamic Expression of BsCXCR Transcripts in Response to Skin Injury
Because CXCR transcripts were enriched in immune related organs and muscle, and they were differentially expressed in healthy and skin ulcerated fish (Figures 1 and 8), we speculated that CXCR participated in skin injury in four-eyed sleeper. To investigate the involvement of BsCXCR in skin injury, we introduced mechanical skin injury and analyzed their expression before or shortly after skin injury ( Figure 9A). The transcript levels of all BsCXCRs were significantly increased as early as 1 h after injury, reaching a maximal increase within 6-12 h, and then decreased at 24 h ( Figure 9B-K). Of these BsCXCR transcripts, BsCXCR4a and BsCXCR6 had a maximum of 60-folds increase, suggesting that BsCXCR4a and BsCXCR6 might play an important role in skin injury. The asterisks indicate the significant differences between two groups (* p < 0.05; ** p < 0.01; *** p < 0.001).

Dynamic Expression of BsCXCR Transcripts in Response to Skin Injury
Because CXCR transcripts were enriched in immune related organs and muscle, and they were differentially expressed in healthy and skin ulcerated fish (Figures 1 and 8), we speculated that CXCR participated in skin injury in four-eyed sleeper. To investigate the involvement of BsCXCR in skin injury, we introduced mechanical skin injury and analyzed their expression before or shortly after skin injury ( Figure 9A). The transcript levels of all BsCXCRs were significantly increased as early as 1 h after injury, reaching a maximal increase within 6-12 h, and then decreased at 24 h ( Figure 9B-K). Of these BsCXCR transcripts, BsCXCR4a and BsCXCR6 had a maximum of 60-folds increase, suggesting that BsCXCR4a and BsCXCR6 might play an important role in skin injury.

Discussion
Numerous studies have demonstrated that mammal CXC chemokines regulate the proliferation and differentiation of normal keratinocytes in inflammatory response to skin injury [3]. Although CXCR homologues have been cloned and characterized in several fish species [10,19,[21][22][23][24], it remains a lack of a global characterization of CXCR family and their involvement in skin injury in gobies. In this study, we identified 10 differentially expressed CXCR transcripts from skin ulcerated four-eyed sleeper (Figure 1), which were classified in seven types. CXCR is one of the evolutionary innovations in vertebrate [8]. Zou et al. have classified the fish CXCRs in three distinct groups: Agnatha (jawless fish), Chondrichthyes (cartilaginous fish), and Osteichthyes (teleost fish) [18]. The jawless fish sea lamprey has four CXCR genes (CXCR4, CXCR7a/7b, and one CXCR related to CXCR1/2) [20,25]. After second round genome duplication, cartilaginous fish, such as elephant shark, have developed more CXCR genes (CXCR1/2/4/5/6) to adapt the emergence of RAG-mediated adaptive immune system [26]. Due to the teleost specific third round genome duplication, many teleost fish species have duplicated more than two copies of CXCR genes, such as CXCR1, CXCR3, and CXCR4 [27]. Therefore, cartilaginous fish have all the known CXCR homologues (CXCR1-6) found in mammals, and teleost fish possess two or more copies of CXCR1/3/4 genes. Notably, CXCR6 is absent in teleost fish, to date. In this study, we identified a four-eyed sleeper CXCR6 gene, which is clustered with cartilaginous fish CXCR6 and possesses the traditional domains of CXCR proteins ( Figures  2, 4 and 5). To the best of our knowledge, BsCXCR6 is the first experimentally-identified CXCR6 gene in teleost fish.
Similar to other teleost fish, the exon-intron arrangement is variable in CXCR genes, even in duplicated paralogs (Figure 3). Asian swamp eel, rock bream, and four-eyed sleeper CXCR1a genes are comprised of two exons and one intron [10,28], whereas foureyed sleeper and grouper CXCR1b genes are intronless [19]. The different exon-intron arrangements of CXCR4a and CXCR4b are conserved in teleost fish, including four-eyed

Discussion
Numerous studies have demonstrated that mammal CXC chemokines regulate the proliferation and differentiation of normal keratinocytes in inflammatory response to skin injury [3]. Although CXCR homologues have been cloned and characterized in several fish species [10,19,[21][22][23][24], it remains a lack of a global characterization of CXCR family and their involvement in skin injury in gobies. In this study, we identified 10 differentially expressed CXCR transcripts from skin ulcerated four-eyed sleeper (Figure 1), which were classified in seven types. CXCR is one of the evolutionary innovations in vertebrate [8]. Zou et al. have classified the fish CXCRs in three distinct groups: Agnatha (jawless fish), Chondrichthyes (cartilaginous fish), and Osteichthyes (teleost fish) [18]. The jawless fish sea lamprey has four CXCR genes (CXCR4, CXCR7a/7b, and one CXCR related to CXCR1/2) [20,25]. After second round genome duplication, cartilaginous fish, such as elephant shark, have developed more CXCR genes (CXCR1/2/4/5/6) to adapt the emergence of RAG-mediated adaptive immune system [26]. Due to the teleost specific third round genome duplication, many teleost fish species have duplicated more than two copies of CXCR genes, such as CXCR1, CXCR3, and CXCR4 [27]. Therefore, cartilaginous fish have all the known CXCR homologues (CXCR1-6) found in mammals, and teleost fish possess two or more copies of CXCR1/3/4 genes. Notably, CXCR6 is absent in teleost fish, to date. In this study, we identified a four-eyed sleeper CXCR6 gene, which is clustered with cartilaginous fish CXCR6 and possesses the traditional domains of CXCR proteins (Figures 2, 4 and 5). To the best of our knowledge, BsCXCR6 is the first experimentally-identified CXCR6 gene in teleost fish.
Similar to other teleost fish, the exon-intron arrangement is variable in CXCR genes, even in duplicated paralogs (Figure 3). Asian swamp eel, rock bream, and four-eyed sleeper CXCR1a genes are comprised of two exons and one intron [10,28], whereas foureyed sleeper and grouper CXCR1b genes are intronless [19]. The different exon-intron arrangements of CXCR4a and CXCR4b are conserved in teleost fish, including four-eyed sleeper [9,22,23,29]. The genomic divergences of duplicated genes might be due to the expansion of chemokine receptors during teleost specific genome duplication.
Because chemokines have high affinity to chemokine receptors, several chemokines can bind to one or two chemokine receptors. For example, the interactions of CXCL8-CXCR1, CXCL8-CXCR2, CXCL12-CXCR4, and CXCL12-CXCR7 have been demonstrated to be critical for directing immune response, wound repair and cell migration in mammals [9,20,30]. However, the evidence of chemokine and receptor binding in fish remains limited. In this study, we predicted BsCXCRs' structures and found that BsCXCR1a/1b/2/3a1 and BsCXCR4b/7, respectively, have similar structures, suggesting that they might bind to the same chemokine, such as CXCL8 and CXCL12. In mammals, CXCR6 is the sole chemokine receptor of CXCL16 ligand; however, we failed to identify CXCL16 in either genome or transcriptome data in four-eyed sleeper, and BsCXCR6's structure was different to other BsCXCRs. Therefore, to which chemokine ligand BsCXCR6 binds needs to be further identified. The localization of the chemokine receptors is intrinsically linked to the role in the structural and functional organization of the cell. However, the subcellular localization of most CXCR members is largely unknown in fish. In this study, we analyzed subcellular localization of BsCXCRs in four-eyed sleeper, and BsCXCRs' localizations were defined in three groups in the absence of binding to chemokines. We found that BsCXCR1a/1b/3 and BsCXCR4b/7 proteins were predicted to share similar 3D-structures ( Figure 5), but their subcellular localizations are different (Figures 6 and 7), suggesting that different CXCRs might cooperatively respond to chemokine signaling.
Paralogs usually show a dominant, biased, or silencing expression pattern [31][32][33]. Biased expression of CXCR paralogs has been reported in several fish species. The expression levels of Asian swamp eel CXCR1b and CXCR3a are significantly higher than their duplicated paralogs of CXCR1a and CXCR3b, respectively [10]. The biased expression of CXCR4a/4b has been observed in zebrafish embryos [11], Asian swamp eel [10], and grouper tissues [9]. Herein, we also observed that the expression levels of BsCXCR1b/3a2/4b were significantly higher than their corresponding paralogs of BsCXCR1a/3a1/4a, respectively (Figure 8), suggesting potentially functional divergence in duplicated BsCXCR paralogs.
CXCRs are required for early inflammation during wound repair in mammals. CXCR1 and CXCR2 play a critical role in the recruitment of leukocytes at the site of inflammation shortly after injury [2,34,35]. Loss of CXCR2 delays the epithelialization and neovascularization during skin repair [36]. The CXCL12-CXCR4 signaling is critical for healing the fracture, as well as neural and cutaneous injury [7,37,38]. Medaka CXCR3a is the hallmark of mononuclear phagocytic cells, and the CXCR3a expressing cells are accumulated in the wounds after fin injury [6]. Our results showed that the expression of all BsCXCRs was significantly increased in 1 h after skin injury, and it peaked within 6-12 h (Figure 9). The acute and transient increase demonstrated that BsCXCR should participate in early inflammation in response to skin injury. However, the exact role of BsCXCR in skin injury would be valuable for further investigation.
In summary, we identified 10 CXCR genes especially of the first experimentally-identified teleost CXCR6 gene BsCXCR6, and globally analyzed their similarities and divergences in genome and protein structure, tissue distribution, and subcellular localization. To the best of our knowledge, it is the first study globally characterizing the structure, expression, and localization of CXCR family in gobies and intertidal fishes. Furthermore, we observed the differential expression of BsCXCR transcripts in healthy and skin ulcerated fish and the significant and transient increase of BsCXCRs shortly after skin injury, demonstrating the involvement of CXCRs in inflammatory response to skin injury. However, which kind of immune cells BsCXCRs express in, and their potential roles, as well as by which mechanism CXCRs use to, in skin injury remain limited. These will be further investigated in our future studies. Ultimately, the present study provides essential data for study of the role and mechanism of CXCR in skin injury in four-eyed sleeper.

Ethical Procedures
All experimental procedures with four-eyed sleeper were approved by the Ethics Committee of Sun Yat-sen University (2020A1515010358), and the methods were in accordance with the approved guidelines.

Fish, Cells and Sample Collection
Four-eyed sleeper was purchased from the Longhai Ruiquan Aquaculture Cooperation Group (Fujian, China). Thirty 4-month-old individuals with average body length of 9.5 ± 0.2 cm and body weight of 14.1 ± 3.23 g were maintained in circle tanks (150 cm in height and 120 cm in diameter) with 500 l of filtrated seawater (salinity = 12 ppt, oxygen level = 3 mmg/L) at 28 • C under a 14 h/10 h light-dark photoperiod. The aquaculture water was replaced every two days. Fish were fed with fish manure twice a day at 8:30 and 18:00, respectively.
For RNA sequencing, the spleen tissues were collected independently from three healthy and three skin ulcerated fish, respectively. For tissue distribution analysis, eight tissues (heart, liver, spleen, kidney, brain, muscle, ovary, and testis) were collected for three healthy individuals.

Total RNA Isolation and Illumine Sequencing
Total RNA was isolated by using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instruction. High-quality RNA (28S/18S ≥ 1.8, OD260/280 ≥ 1.8, and RIN value ≥ 7) was subjected to construct cDNA library for de novo transcriptome sequencing. The de novo transcriptome sequencing of the spleens of three healthy and three skin injured fish was performed on the MGIseq2000 platform (GrandOmics, Wuhan, China).

Bioinformatic Analysis of BsCXCR Family
Multiple alignments were analyzed by the Bioedit (version 7.0). The domains of BsCXCRs' amino acids were predicted by Simple Modular Architecture Research Tool (http://smart.embl-heidelberg.de/, version 9.0, 17 August 2021). The structures of BsCXCR proteins were predicted by PyMOL (version 2.0.6). The N-glycosylation site, functional cysteine and signature motif were predicted by ExPASy (https://prosite.expasy.org/, version 20.0, 17 August 2021). The phylogenetic tree was constructed by using the Maximum-like methods with MEGA 6.0, and the tree was bootstrapped for 1000 replications.

Quantitative PCR (qPCR)
The isolated total RNA was reverse-transcribed into single strand cDNA with Prime-script™ First Strand cDNA Synthesis Kit (Takara, Kyoto, Japan). qPCR was performed on the LightCycle 480 II (Roche, Basel, Switzerland) using SYBR ® Premix Ex Taq™ II (Takara, Kyoto, Japan). The cycling conditions were: 94 • C for 5 min (min), followed by 40 cycles of 95 • C for 15 s (s), 58 • C for 15 s, 72 • C for 15 s [43]. The expression level was normalized by the 2 −∆CT methods using β-actin as an internal control. The primers are listed in Table A1.

Plasmid Construction
The open reading frames (ORFs) of each BsCXCR were cloned into the pEGFP-N3 vector to generate GFP fusion proteins. The ORFs of BsCXCR1a/b, BsCXCR3a1/2I, or BsCXCR7 were inserted into EcoR I/BamH I sites of the pEGFP-N3 vector; meanwhile, the ORFs of BsCXCR4a/4b, BsCXCR5, or BsCXCR6 were inserted into Nhe I/EcoR I sites of the pEGFP-N3 vector. All plasmids were verified by DNA sequencing. The primers are listed in Table A1.

Immunofluorescence
HEK293T cells were seeded on a glass coverslip in 24-well culture plates. At 80% confluence, the cells were transfected with different pEFP-N3-BsCXCR plasmids using Lipofectamine 8000 (Beyotime, Shanghai, China). At 24 h post-transfection, cells were fixed in 4% paraformaldehyde for 20 min at room temperature (RT), washed in PBS three times, permeabilized in PBS with 0.2% Triton X-100 for 15 min, blocked with 5% milk, and incubated in anti-β-catenin antibodies with a dilution of 1:400 at 4 • C overnight. After three washes in PBST (0.1% Tween-20 diluted in PBS), and followed by three washes in PBS, the cells were incubated with Alexa Fluor ® 555 donkey-anti-mouse antibodies IgG H&L (Invitrogen, Carlsbad, CA, USA) at a dilution of 1:400 for 1 h at RT. Then, the cells were washed three times with PBS, and the nuclei were stained with DAPI (10 µg/L diluted in PBS solution) for 10 min. Finally, the cells were visualized and documented with confocal microscope (LSM800, Zeiss, Oberkohen, Germany).

Mechanical Skin Injury
A total of 30 four-eyed sleeper individuals weighing 15-20 g were used for mechanical skin injury experiments. After anesthetized with 1% MS222 (Sigma, Fullerton, CA, USA), mechanical skin injury wound (about 5 mm × 5 mm × 0.3 mm) was induced using surgical scissors on one side of posterior belly of fish, as previously described [35]. Then, the fish were cultured in sterile water. At 1 h, 3 h, 6 h, 12 h, and 24 h after injury, three fish per time point were anesthetized, and the skin tissue around the wounds were collected for RNA extraction. Each experiment has been repeated for three times.

Statistics Analysis
The statistics was calculated using SPSS version 20. Differences between control and treatment groups were assessed by one-way ANOVA. p < 0.05 was considered as statistically significance.    Table A1. Sequences of primer used in this paper.