Fads2b Plays a Dominant Role in ∆6/∆5 Desaturation Activities Compared with Fads2a in Common Carp (Cyprinus carpio)

Highly unsaturated fatty acids (HUFAs) are essential for mammalian health, development and growth. However, most mammals, including humans, are incapable of synthesizing n-6 and n-3 HUFAs. Fish can convert C18 unsaturated fatty acids into n-6 and n-3 HUFAs via fatty acid desaturase (Fads), in which Fads2 is a key enzyme in HUFA biosynthesis. The allo-tetraploid common carp theoretically encode two duplicated fads2 genes. The expression patterns and desaturase functions of these two homologous genes are still unknown. In this study, the full length of the fads2a and fads2b were identified in common carp (Cyprinus carpio). Expression analyses indicate that both genes were mainly expressed in the liver and the expression of fads2b is higher than fads2a at different developmental stages in carp embryos. Heterogenous expression and 3D docking analyses suggested that Fads2b demonstrated stronger ∆6 and ∆5 desaturase activities than Fads2a. The core promotor regions of fads2a and fads2b were characterized and found to have different potential transcriptional binding sites. These results revealed the same desaturase functions, but different activities of two homologues of fasd2 genes in common carp. The data showed that fads2b played a more important role in HUFA synthesis through both expression and functional analyses.


Introduction
Highly unsaturated fatty acids (HUFAs) are fatty acids which contain three or more double bonds and longer than 20 carbon atoms. HUFAs have long been recognized as nutritional elements and essential dietary components in human health [1]. Nutritional deficiencies in n-6 and n-3 HUFAs may have adverse effects on the development of brain and neural systems, and may also lead to negative metabolic changes [2]. Dietary n-3 HUFA deficiency is associated with a potential increased risk for attention deficit hyperactivity disorder (ADHD) and other similar behavioral disorders [3]. As they are not endogenously produced by humans [4], fish have proven to be an important source of dietary sources of HUFAs [5]. The biosynthesis of HUFAs in fish involves the sequential desaturation and elongation of precursor C18 polyunsaturated fatty acids. Among fish species, there is still wide variation in the ability to synthesize HUFAs, which depends on their content of fatty acid desaturase (Fads) and elongase enzymes [6].
In fish, Fads controls the degree of unsaturation of HUFAs and catalyzes the first desaturation step [7]. More specifically, ∆6 and ∆5 desaturase catalyze the desaturation at the ∆6 and ∆5 positions in the carbon backbone during the synthesis of arachidonic

The Differences in Gene and Protein Structures of Common Carp fads2a and fads2b
To obtain the full-length sequences of two fads2 genes of common carp, the predicted CDS regions were amplified according to the zebrafish fads2 gene sequence, and then the 5 race and 3' race ends were cloned completely from the common carp liver cDNA sample using a rapid cDNA end cloning technique. Common carp fads2a (GenBank accession number: MK852165.1), located in the A25 chromosome of the common carp genome (from 2,753,781 bp to 2,765,247 bp) [19], and fads2b (GenBank accession number: MK852166.1), located in the B25 chromosome of the common carp genome (from 4,452,255 bp to 4,465,379 bp), have a total length of 1980 bp and 1943 bp, respectively. Both genes are composed of 12 introns and 12 exons, including a 1335 bp open reading frame (ORF) (Figure 1). Although fads2a and fads2b have the same number of introns and exons and both encode 444 amino acids, the global identity between these two mRNAs was 89.10%. The protein sequence similarity between common carp Fads2a/Fads2b and zebrafish Fads2 were 90.77% and 88.51%, respectively. The similarity between common carp Fads2a and Fads2b was 89.86%. Common carp Fads2a and Fads2b, as well as zebrafish Fads2 protein sequences all have a cytochrome b5-like heme/steroid binding domain and a fatty acid desaturase domain, which indicate that they all possess a fatty acid desaturase function ( Figure S1). using a rapid cDNA end cloning technique. Common carp fads2a (GenBank accession number: MK852165.1), located in the A25 chromosome of the common carp genome (from 2,753,781 bp to 2,765,247 bp) [19], and fads2b (GenBank accession number: MK852166.1), located in the B25 chromosome of the common carp genome (from 4,452,255 bp to 4,465,379 bp), have a total length of 1980 bp and 1943 bp, respectively. Both genes are composed of 12 introns and 12 exons, including a 1335 bp open reading frame (ORF) (Figure 1). Although fads2a and fads2b have the same number of introns and exons and both encode 444 amino acids, the global identity between these two mRNAs was 89.10%. The protein sequence similarity between common carp Fads2a/Fads2b and zebrafish Fads2 were 90.77% and 88.51%, respectively. The similarity between common carp Fads2a and Fads2b was 89.86%. Common carp Fads2a and Fads2b, as well as zebrafish Fads2 protein sequences all have a cytochrome b5-like heme/steroid binding domain and a fatty acid desaturase domain, which indicate that they all possess a fatty acid desaturase function ( Figure S1). Predicted 3D protein structure analysis revealed that the protein structures of common carp Fads2a and Fads2b are highly similar to each other, and both are also similar to zebrafish Fasd2 ( Figure S2). However, there are still 45 amino acid differences in Fads2a and Fads2b, including 6 differences in cytochrome b5-like heme/steroid binding domains and 30 differences in fatty acid desaturase domains. The protein secondary structure information showed that the Fads2a and Fads2b proteins both have 16 alpha helical structures but four and three beta pleated sheet structures, respectively ( Figure S1). Further analysis revealed that there are differences in amino acid sequences in the 6th, 8th and 10th alpha helix structures of Fads2a and Fads2b, which may lead the two genes to have diverse functions.
A phylogenetic tree comparing the deduced amino acid sequences of Fads2 in common carp with those of marine fish, other freshwater fish and mammals (Table S1) is shown in Figure 2. The analysis shows that Cyprinus carp Fads2 were clustered together with other two Cyprinidae family freshwater teleosts Carassius auratus and Sinocyclocheilus anshuiensis and were separate from marine teleosts like Salmo salar, which belongs to the Salmonidae family, in the tree. The common carp Fads2a/Fads2b protein sequences were most closely related to S.anshuiensis Fads2 (the similarity was 94.82% and 95.95%) and C. auratus Fads2 (the similarity was 95.50% and 93.47%), respectively. Combining the syntenic gene analysis between common carp and other Cyprininae species in our previous work [19], fads2 genes of Cyprinus carpio, Carassius auratus and Spinibarbus anshuiensis might be linearly homologous and may have evolved from the same ancestor, as they all have two fads2 genes which may be due to whole genome replication of the common ancestor. Predicted 3D protein structure analysis revealed that the protein structures of common carp Fads2a and Fads2b are highly similar to each other, and both are also similar to zebrafish Fasd2 ( Figure S2). However, there are still 45 amino acid differences in Fads2a and Fads2b, including 6 differences in cytochrome b5-like heme/steroid binding domains and 30 differences in fatty acid desaturase domains. The protein secondary structure information showed that the Fads2a and Fads2b proteins both have 16 alpha helical structures but four and three beta pleated sheet structures, respectively ( Figure S1). Further analysis revealed that there are differences in amino acid sequences in the 6th, 8th and 10th alpha helix structures of Fads2a and Fads2b, which may lead the two genes to have diverse functions.
A phylogenetic tree comparing the deduced amino acid sequences of Fads2 in common carp with those of marine fish, other freshwater fish and mammals (Table S1) is shown in Figure 2. The analysis shows that Cyprinus carp Fads2 were clustered together with other two Cyprinidae family freshwater teleosts Carassius auratus and Sinocyclocheilus anshuiensis and were separate from marine teleosts like Salmo salar, which belongs to the Salmonidae family, in the tree. The common carp Fads2a/Fads2b protein sequences were most closely related to S.anshuiensis Fads2 (the similarity was 94.82% and 95.95%) and C. auratus Fads2 (the similarity was 95.50% and 93.47%), respectively. Combining the syntenic gene analysis between common carp and other Cyprininae species in our previous work [19], fads2 genes of Cyprinus carpio, Carassius auratus and Spinibarbus anshuiensis might be linearly homologous and may have evolved from the same ancestor, as they all have two fads2 genes which may be due to whole genome replication of the common ancestor.

Divergent Spatial and Temporal Expression Patterns of fads2a and fads2b In Vivo
Tissue distributions of fads2a and fads2b were determined by real-time quantitative PCR (qPCR) performed on cDNAs reverse transcribed from different adult (one year old) common carp organs. The results showed that fads2a and fads2b genes were predominantly expressed in the liver compared to other tissues ( Figure 3a). Transcription levels of fads2b were significantly higher than fads2a in the liver (hepatopancreas) (p < 0.001), intestine (p < 0.001), brain (p < 0.001) and spleen (p < 0.05). However, the amount of expression of fads2a in the heart (p < 0.05) and gill (p < 0.001) were higher than fads2b. Although the expression patterns of fads2a and fads2b are diverse in different tissues, it can be inferred that fads2b plays a more dominant role in the desaturation process than fads2a because the synthesis of HUFAs mainly occurs in the liver.

Divergent Spatial and Temporal Expression Patterns of fads2a and fads2b In Vivo
Tissue distributions of fads2a and fads2b were determined by real-time quantitative PCR (qPCR) performed on cDNAs reverse transcribed from different adult (one year old) common carp organs. The results showed that fads2a and fads2b genes were predominantly expressed in the liver compared to other tissues ( Figure 3a). Transcription levels of fads2b were significantly higher than fads2a in the liver (hepatopancreas) (p < 0.001), intestine (p < 0.001), brain (p < 0.001) and spleen (p < 0.05). However, the amount of expression of fads2a in the heart (p < 0.05) and gill (p < 0.001) were higher than fads2b. Although the expression patterns of fads2a and fads2b are diverse in different tissues, it can be inferred that fads2b plays a more dominant role in the desaturation process than fads2a because the synthesis of HUFAs mainly occurs in the liver.

Divergent Spatial and Temporal Expression Patterns of fads2a and fads2b In Vivo
Tissue distributions of fads2a and fads2b were determined by real-time quantitative PCR (qPCR) performed on cDNAs reverse transcribed from different adult (one year old) common carp organs. The results showed that fads2a and fads2b genes were predominantly expressed in the liver compared to other tissues ( Figure 3a). Transcription levels of fads2b were significantly higher than fads2a in the liver (hepatopancreas) (p < 0.001), intestine (p < 0.001), brain (p < 0.001) and spleen (p < 0.05). However, the amount of expression of fads2a in the heart (p < 0.05) and gill (p < 0.001) were higher than fads2b. Although the expression patterns of fads2a and fads2b are diverse in different tissues, it can be inferred that fads2b plays a more dominant role in the desaturation process than fads2a because the synthesis of HUFAs mainly occurs in the liver.   To quantify the expression patterns of fads2a and fads2b during the early developmental stages of common carp, we analyzed their temporal expression in carp embryos at 0, 24, 48, 72 and 120 hpf by qPCR. As seen in Figure 3b, the transcript levels of fads2b were significantly higher than that of fads2a at 0, 24, 72, 96 and 120 hpf (p < 0.001). The expression patterns of fads2a and fads2b were both higher before the hatching period (48-72 hpf) than after. Although fads2a was more highly expressed than fads2b at 48 hpf, its expression level remained low during the larva stage (after 72 hpf) and was significantly lower than that of fads2b (p < 0.001). These results indicated that fads2b might play a more important role than fads2a during the early developmental stages of common carp.
In order to further elucidate the spatial and temporal expression of fads2a and fads2b in various common carp tissues during the early developmental stages after hatching, carp embryos were collected for whole embryo in situ hybridization at 48, 72, 96 and 120 hpf, respectively (Figures 4 and 5). The results showed that the expression of fads2a and fads2b were mainly detected in the brain region at 48, 72 and 96 hpf. Notably, the transcript signals of fads2b ( Figure 4b) were stronger than that of fads2a (Figure 5b) in the brain at 72 hpf. The expression of both genes were detected in the liver and intestine at 96 hpf (Figures 4c and 5c) and significantly increased at 120 hpf (Figures 4d and 5d). In order to study the effect of carp fads2 genes on fatty acid content during embryonic development, the desaturase activity of these two genes was estimated by comparing the fatty acid level of C18 substrate (C18:3n-3 and C18:2n-6) with the fatty acid level of all potential desaturation products. During the whole process of embryo development, the total amount of C18 substrates decreased by nearly 50% and the content of the fads2 product showed a stable growth trend (Table S2). In summary, the areas of expression of fads2a and fads2b are consistent in the same period during the early developmental stages of common carp, mainly in the brain and liver.
To quantify the expression patterns of fads2a and fads2b during the early de mental stages of common carp, we analyzed their temporal expression in carp emb 0, 24, 48, 72 and 120 hpf by qPCR. As seen in Figure 3b, the transcript levels of fads2 significantly higher than that of fads2a at 0, 24, 72, 96 and 120 hpf (p < 0.001). The expr patterns of fads2a and fads2b were both higher before the hatching period (48-72 hp after. Although fads2a was more highly expressed than fads2b at 48 hpf, its expressio remained low during the larva stage (after 72 hpf) and was significantly lower th of fads2b (p < 0.001). These results indicated that fads2b might play a more importa than fads2a during the early developmental stages of common carp.
In order to further elucidate the spatial and temporal expression of fads2a and in various common carp tissues during the early developmental stages after ha carp embryos were collected for whole embryo in situ hybridization at 48, 72, 96 a hpf, respectively (Figures 4 and 5). The results showed that the expression of fads fads2b were mainly detected in the brain region at 48, 72 and 96 hpf. Notably, the tra signals of fads2b ( Figure 4b) were stronger than that of fads2a (Figure 5b) in the brai hpf. The expression of both genes were detected in the liver and intestine at 96 hpf (F 4c and 5c) and significantly increased at 120 hpf (Figures 4d and 5d). In order to stu effect of carp fads2 genes on fatty acid content during embryonic development, t saturase activity of these two genes was estimated by comparing the fatty acid l C18 substrate (C18:3n-3 and C18:2n-6) with the fatty acid level of all potential desatu products. During the whole process of embryo development, the total amount of C1 strates decreased by nearly 50% and the content of the fads2 product showed a growth trend (Table S2). In summary, the areas of expression of fads2a and fads2b a sistent in the same period during the early developmental stages of common carp, in the brain and liver.

Molecular Docking Analyses of Fads2a and Fads2b
To study the catalytic behavior of Fads2a and Fads2b at the molecular level during the desaturation process, the binding pockets of both Fads proteins and ten fatty acid small molecules, including the ∆6 and ∆5 substrates, were analyzed using molecular docking ( Figure S4A). We calculated the cavity volumes of Fads2a and Fads2b binding pockets, which were 398.2 Å 3 and 411.7 Å 3 , respectively ( Figure S4B). To investigate the interaction activity between Fads2a/Fads2b and ∆6/∆5 substrates, we specifically estimated the docking positions of the four fatty acids, including C18:2n-6, C18:3n-3, C20:3n-6 and C20:4n-3, in both Fads peptide structures ( Figure S5), respectively, and the interaction energies between them ( Table 2). The results of the molecular docking analysis showed that Fads2b had a lower affinity energy of interaction with C18:2n-6 and C20:3n-6 than Fads2a, which indicated that Fads2b was more active in the desaturation process than Fads2a due to its greater preference for interaction with the ∆6 and ∆5 substrates fatty acids.

Promoter Activity Analysis of fads2a and fads2b
In order to investigate the molecular regulation of fads2a and fads2b gene expression in common carp, sequences of different length upstream of the 5 UTR of each gene were selected to be cloned as their candidate core promoter regions. A potential promoter sequence approximately 2500 bp in length, as well as four truncated sequences of fads2a or fads2b were recombined into a pGL3-basic vector, respectively, then subsequently transfected into carp epithelioma cells (EPC). The transcriptional activity of the fads2a and fads2b gene promoters were analyzed by performing a dual luciferase test. As shown in Figure 8, five different candidate promoters for the fads2a (Figure 8a) and fads2b (Figure 8b) genes showed no significant differences in luciferase activity, which leads to the conclusion that the core active promoters of the fads2a and fads2b genes were 622 bp and 812 bp upstream of 5 UTR of each gene, respectively. The core promoter activities of fads2b (−812 bp to −1 bp) and fads2a (−622 bp to −46 bp) were also compared. The result indicated that the fads2b promoter showed significantly higher activity than the fads2a promoter (p < 0.001) (Figure 8c). 8b) genes showed no significant differences in luciferase activity, which leads to the conclusion that the core active promoters of the fads2a and fads2b genes were 622 bp and 812 bp upstream of 5′UTR of each gene, respectively. The core promoter activities of fads2b (−812 bp to −1 bp) and fads2a (−622 bp to −46 bp) were also compared. The result indicated that the fads2b promoter showed significantly higher activity than the fads2a promoter (p < 0.001) (Figure 8c). Finally, we predicted the potential transcription factor binding sites in the core promoters of common carp fads2a and fads2b. The results showed that TBP, NF-1 and AP-1 may regulate fads2a, while SP1, NF-1 and AP-1 may regulate fads2b (Figure 8d). The promoter regions of both genes also have the same binding site of transcription factor NF-1. It is speculated that the promoter activity of the fads2a and fads2b genes may be different due to the differences in promoter region sequences and regulatory factors. Finally, we predicted the potential transcription factor binding sites in the core promoters of common carp fads2a and fads2b. The results showed that TBP, NF-1 and AP-1 may regulate fads2a, while SP1, NF-1 and AP-1 may regulate fads2b (Figure 8d). The promoter regions of both genes also have the same binding site of transcription factor NF-1. It is speculated that the promoter activity of the fads2a and fads2b genes may be different due to the differences in promoter region sequences and regulatory factors.

Discussion
Fish have proven to be a unique and rich source of HUFAs [5]. Common carp, rich in HUFAs [20,21], is the most abundant global aquatic product and accounts for 10% of all freshwater production [22,23]. As a polyploid fish, possessing twice as many genes as other fish, common carp is an ideal model to study the evolution of the vertebrate polyploid genome [24]. Previous studies have successfully cloned one carp fads gene [10]. However, whole genome sequencing has revealed that carp have two fads genes rather than one. This discovery highlights the need for more in-depth research into the study of desaturase functions in the HUFA's synthesis pathways of common carp.
In contrast to mammals, which have distinct fads1 and fads2 genes with separate specificities of ∆5 and ∆6 activities [25], the fads1 gene has been lost during evolution in fish. This phenomenon may be caused by a comprehensive process of gene replication, gene loss, and functional diversity [26,27]. As a result, most desaturation steps in the HUFA biosynthetic pathway are catalyzed by Fads2 enzymes in fish. In this study, the full-length sequences of the common carp fads2a and fads2b genes were successfully cloned for the first time. The gene structure of two fads2 genes were fundamentally the same, including 12 introns and 12 exons, and both encoded 444 amino acids. However, the amino acid sequence similarity of the Fads2a and Fads2b genes is 89.86%, which shows a higher degree of difference than the gene structure. Three-dimensional structure analysis showed differences in positions and lengths of alpha helix and beta pleated sheet structures, which indicated that the fads2a and fads2b genes might also have different desaturation activities, functions and regulation mechanisms in the process of unsaturated fatty acid synthesis.
Expression studies of the fads2a and fads2b genes in different adult common carp tissues revealed both genes were expressed in all experimental tissues, while expression levels in various tissues were different. Both the fads2a and fads2b genes exhibited the highest expression in liver and brain tissues, which are the main effector organs of desaturase. However, the expression of fads2b is higher than that of fads2a in those tissues. Based on these results, it is speculated that fads2b plays a stronger role and is more efficient in the process of HUFAs synthesis than fads2a.
Temporal and spatial expression analysis of fads2a and fads2b by whole embryo in situ hybridization revealed that the transcripts of fads2 and fads2b were maternally expressed, as well as appearing in different expression patterns over time. The Fads2b gene was expressed at 24 hpf and showed an increased expression trend in the subsequent development period, which led to speculation that it might play an important role throughout the entire embryonic development stage in common carp. Comparatively, the expression of the fads2a gene remained at low levels before 72 hpf and increased significantly at 96 hpf and 120 hpf. This result indicated that the fads2a gene mainly regulates the metabolism of fatty acids and participates in the synthesis of HUFA when carp embryos develop into fry after membrane emergence.
The study of spatial expression of fads2a and fads2b in various tissues during common carp embryonic development demonstrated that the transcripts of both genes had positive signals in the brain regions at a very early stage. Comparatively positive signals develop in the early embryonic stage. In comparison, the expression of fads2a increased sharply and gained functionality after the carp embryo emerged from the membrane. In adult common carp, the expression of fads2b in most tissues is higher than that of fads2a, suggesting that the fads2b gene plays a major role in the HUFA synthesis pathway.
The desaturation activities of common carp Fads2a and Fads2b were investigated in transgenic yeast in this study. The results showed that both Fads2a and Fads2b had ∆5 and ∆6 activities. However, the activities of ∆4 and ∆8 were not verified. Experimentally, transfection of Saccharomyces cerevisiae functioned to verify gene expression in vitro, but also required secondary verification in vivo. Unlike freshwater fish like common carp, ∆4 desaturase activity is mostly found in seawater fish, such as Siganus canaliculatus [28], Trachinotus ovatus [29], Leuresthes sardina, etc. The addition of exogenous fatty acids can inhibit the metabolic pathway of ∆6 activity. Studies have shown that Atlantic salmon fed with fish oils rich in EPA and DHA have significantly lower HUFA biosynthesis pathway activity in the liver and intestines compared to those fed with fish oil that did not contain EPA and DHA [30], indicating that the level of n-3 series unsaturated fatty acids directly affects HUFAs that were found in the liver after 96 hpf. These results suggest that the fads2a and fads2b gene are mainly expressed in the liver after hatching.
In the process of common carp embryonic development, the total amount of C18 fatty acids substrate (C18:3n-3 and C18:2n-6) decreased by nearly 50%. DHA content increased steadily during the stage from 0 hpf to 120 hpf. Combined with the results of temporal and spatial expression of fads2a and fads2b, it is speculated that fads2b, which is expressed in the head of carp embryo, would synthesize DHA in the brain and contribute to brain development. The Fads gene combined with ∆6 activity plays a certain regulatory role in the synthesis of HUFAs. The expression of ∆6 fads mRNA is lower in the liver of Atlantic salmon fed with fish oil than those fed with the vegetable oil control [31].
We defined the active promoter regions of common carp fads2a and fads2b in this study. We then screened the predicted transcription factors, TBP and NF-1 for fads2a, NF-1, Sp1 and AP-1 for fads2b. Sp1 is particularly important in constitutive expression of the Atlantic salmon ∆6 fads gene. ∆6 fads has lower promoter activity and tissue expression levels in Atlantic cod than Atlantic salmon and other seawater fish [32]. This difference may be caused by the absence of Sp1 binding sites in the promoter region of Atlantic cod ∆ 6 fads [12]. As a ubiquitous universal transcription factor, Sp1 has also been confirmed in the activation of the human ∆6 fads gene promoter [33]. Compared with the fads2b promoter, the promoter of fads2a shows lower activity and does not contain Sp1 binding sites. The results of the promoter activity comparison were consistent with the results of the yeast transfection experiment in vitro. The conversion rate of Fads2b to different products was also higher than that of Fads2a. Therefore, it can be inferred from this study that the difference between common carp fads2a and fads2b may be due to the existence of SP1 binding sites in the fads2b gene promoter, which are necessary for the complete expression of ∆6 activity of the common carp fads gene.
Our study successfully cloned two fads2 genes of common carp and revealed that both fads2a and fads2b had ∆6/∆5 desaturation activities and had a similar expression pattern. However, the expression level, desaturation activities and promoter activities of fads2a were generally higher than fads2b.

Animals and Ethics Statement
Common carp studied in this research were cultivated for one year in the breeding base at the Chinese Academy of Fishery Sciences (Beijing, China), with the same commercial diet. Common carp samples were collected for cloning and quantitative expression analysis of the fads2a and fads2b genes. The experimental common carp were one year old, and the body weight was 500 g on average. Common carp embryos studied in this research were obtained by artificial fertilization and incubated in 25 • C fresh water with an oxygen supply. This research was approved by the committee for the ethics of animal experiments of the Chinese Academy of Fishery Science.

Sequence and Structure Analysis of Two Common Carp fads2 Genes
To identify the common carp fads2 gene, the zebrafish fads2 mRNA sequence (GenBank accession number: NM_131645.2) was aligned to the recent common carp genome (GenBank accession number: GCA_018340385.1) [19]. Two homologous genome locations which overlapped with the whole zebrafish fads2 cDNA sequences were selected as candidate common carp fads2 genes. The protein sequences and complete open reading frames (ORFs) of these two candidate genes were predicted by FGENESH+ web server (http:// www.softberry.com (accessed on 28 October 2019)) by taking the zebrafish Fasd2 sequence (GenBank accession number: AAH49438.1) as the template. To verify the ORF sequences of the two common carp fads2 genes, specific primers were designed for 5 and 3 rapid amplification of cDNA ends (RACE) to obtain full-length transcripts (Tables S3 and S4). To compare the three-dimensional structures of the common carp and zebrafish Fads2 proteins, their amino acid sequences were used for ab initio modeling with the I-TASSER server [34]. The modeled protein structures were refined through molecular dynamics simulations with AMBER16 software [35].

Phylogenetic Analysis
The protein sequences of common carp Fads2a and Fads2b aligned with selected mammalian and other defined teleost ∆6 desaturase proteins (Table S1) were used for constructing the phylogenetic tree by the neighbor-joining method of MEGA7 [36,37]. Confidence of tree branch topology was obtained by a bootstrapping measurement of 1000 iterations [38]. Finally, the phylogenetic tree was embellished using the iTOL v6.6 web server (https://itol.embl.de/itol.cgi (accessed on 24 October 2022)) [39].

GC-MS Analysis of Common Carp Embryos at Different Developmental Stages
Common carp embryos at 0, 24, 48, 72, 96 and 120 hpf were collected during development and freeze-dried after grinding into a powder while adding liquid nitrogen into the mortar. Fatty acids methyl esters (FAMEs) were prepared with 2 mL of 0.5 M NaOH/MeOH at 100 • C for 30 min after ultrasonic concussion for 10 min, followed by the addition of 2 mL 14% BF3·MeOH (wt/vol) at 100 • C for 1 h. FAMEs were extracted with 1 mL of hexane and 5 mL of a saturated solution of NaCl by vigorous mixing then centrifugation at 1000× g for 5 min. The aqueous top phase was transferred to a clean glass tube and the lower phase was used to extract FAMEs a second time by adding 1 mL hexane. The two FAME extracts were combined and dried in a stream of oxygen-free nitrogen. The FAMEs were then resuspended in 1 mL of hexane and filtered through a nylon syringe filter SCAA-104 (ANPEL, Shanghai, China). Fatty acid analyses were performed using a 7890A GC System (Agilent Technologies, Santa Clara, CA, USA) equipped with a flame ionization detector. The gas chromatograph was equipped with a capillary column (60 m × 0.25 mm i.d., Agilent Technologies, Santa Clara, CA, USA) hydrogen was used as the carrier gas. Samples were applied when the temperature gradient was from 60 to 150 • C at 10

Heterologous Expression of fads2 ORFs in Yeast
ORF fragments of two common carp fads2 cDNAs were amplified from common carp liver cDNA with primers containing restriction sites for Hind III and Xho I (Table S5). The amplified DNA products were constructed into a pYES2 vector (Invitrogen, Waltham, MA, USA). The constructed plasmids, pYES2-Fads2a and pYES2-Fads2b, were extracted using an AxyPrep Plasmid Miniprep Kit (Axygen, Union City, CA, USA) and transformed into S. cerevisiae (strain INVSc1, a gift form Nanhai institution, Chinese Academy of Fishery Sciences) by using the S. c. EasyComp™ Transformation Kit (Invitrogen, Waltham, MA, USA), respectively.
The transformed yeast were grown in S. cerevisiae minimal medium-uracil overnight at 30 • C, then the cultures were diluted to an OD600 of 0.4 for further growth. When the yeast cultures reached an OD600 of 1, 2% (wt/vol) galactose was used for protein induction and supplemented with different fatty acid (FA) substrates at final concentrations of 0.5 mM (C18), 0.75 mM (C20) and 1 mM (C22), respectively. After 48 h, the yeast was centrifuged and washed twice, then dried by vacuum freeze-drying for 12 h. FAME preparation and FA analysis of yeast samples were performed as described above for common carp embryos. The conversion rate of potential substrates was calculated by the proportion of substrate FA converted to desaturated products as (product area/(product area + substrate area)) × 1000.

Expression of fads2 Genes during Common Carp Ontogeny
The fish were anaesthetized using a eugenol solution at a concentration of 40 mg/L. Twelve tissues including the liver, intestine, brain, muscle, eye, heart, spleen, skin, pancreas, gill, blood and kidney were sampled from each fish and immediately collected in 1.5 mL RNase-free centrifuge tubes, frozen in liquid nitrogen then stored in a −80 • C freezer until being further processed. Embryos at 0, 24, 48, 72, 96 and 120 hpf were sampled for RNA isolation. Expression of the common carp fads genes was measured by qPCR. The qPCR analysis was performed on a 7500 Real Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA) using SYBR Green Realtime PCR Master Mix (TOYOBO, Japan) according to a previously described procedure [40]. The relative expression of the target gene was normalized with β-actin expression calculated by the 2 −∆∆Ct method. Primers used for qPCR are shown in Table S6.

Whole-Mount In Situ Hybridization (WISH)
Digoxygenin-labelled RNA probes of common carp fads2a and fads2b were synthesized in vitro. First, approximate 400 bp fads2a and fads2b DNA templates were cloned from the common carp liver cDNA using EasyTaq PCR Supermix kit (TransGen Biotech, Beijing, China), 1 µL common carp liver cDNA, 25 µL 2× EasyTaq Supermix, 1 µL WISH-F/WISH-R primers (Table S7) and 22 µL ddH 2 O. They were processed in a PCR reaction program as follows: initial denaturation at 94 • C for 4 min, 35 cycles of denaturation at 94 • C for 20s-primer annealing at 60 • C for 20s-extension at 72 • C for 30s, and final extension at 72 • C for 5 min. Second, the products were inserted into the pEASY-T3 cloning vector (TransGen Biotech, Beijing, China) by setting up the following cloning reaction program: 0.5 µL PCR products, 1 µL pEASY-T3 cloning vector and 3.5 µL ddH 2 O were incubated at 25 • C for 5 min. Third, positive clones were identified by colony PCR using M13 forward and reverse primers and inoculating positive clones in Amp+ LB liquid medium. Then, plasmid DNA was isolated using a plasmid DNA MiniPrep Kit (Thermo Scientific, Waltham, MA, USA). The identified DNA templates of fads2a and fads2b RNA probes were analyzed by sequencing with M13 forward and reverse primers. Finally, RNA probes were transcribed the from 1µg purified DNA template by using the MAXIscript™ T7 transcription kit (Thermo Scientific, Waltham, MA, USA), 2 µL 10× transcription buffer, 2 µL DIG-Labling mix, 1 µL RNase inhibitor, and 1 µL RNA polymerase mix with up to 20 µL DEPC water were incubated for 3 h at 37 • C.
The 48, 72, 96 and 120 hpf common carp embryos were collected and fixed in 4% paraformaldehyde (PFA). Pigmentation of embryos was removed in 3% H 2 O 2 /0.5% KOH and dehydrated in an ascending MeOH row as described previously [41]. In brief, embryos were re-hydrated in a descending row of MeOH, washed in PBST and permeabilized using 10 µg/mL proteinase K at room temperature (20 min for 48 hpf embryos, 50min for 72 hpf embryos, 70 min for 96 hpf embryos and 90 min for 120 hpf embryos). Digoxigenin (DIG)labeled RNA probes were generated using DIG RNA Labeling Mix (Roche, Germany) with a MAXIscript kit (Invitrogen, Waltham, MA, USA) as described above. Whole-mount in situ hybridization (WISH) was performed as described previously [41].

Molecular Docking
The crystal structure for common carp Fads2a and Fads2b were unavailable in the PDB databank, thus the homology models of these two proteins were ab initio constructed using I-TASSAR software [42]. The best three-dimensional (3D) structure of Fads2a/Fads2b by homology modeling was further optimized using AMBER16 software [43]. The structures of four fatty acid substrates (C18:2n-6, C18:3n-3, C20:3n-6 and C20:4n-3) used for molecular docking in this research were obtained from Pubchem database (https://pubchem.ncbi.nlm. nih.gov (accessed on 19 September 2019)). Binding of fatty acid substrates to Fads2a/Fads2b were investigated by molecular docking using the AutoDock (version 4.2), and the docking complexes with highest binding energy were used in the following molecular dynamics (MD) simulations using AMBER16 software.

Cloning of fads2a/fads2b Promoters and Construction of Expression Plasmids
The fads2a/fads2b candidate promoter region sequences were obtained from genomic sequencing data of common carp. To identify the core promoters of common carp fads2a/fads2b, different length (576 bp, 971 bp, 1484 bp, 1895 bp, 2479 bp length of fads2a upstream sequences and 811 bp, 1257 bp, 1786 bp, 2227 bp, 2675 bp length of fads2b upstream sequences) candidate promoter regions from −2525 to −46 bp upstream fads2a gene and −2676 to −1 bp upstream fads2b gene was amplified from common carp genomic DNA by PCR. The forward primers with the Xho I restriction endonuclease site and the reverse primers with the Hind III restriction endonuclease site are shown in Table S8. Then, different length products were purified by AxyPrepTM PCR Cleanup Kit (Axygen Biosciences, Union City, CA, USA) and constructed into a linearized pGL3-Basic vector (Promega, Madison, WI, USA) by T4 DNA ligase (Takara, Shiga, Japan) overnight at room temperature.

Cell Culture, Transfection and Luciferase Assay
Transient transfections were performed through lipofection using Lipofectamine 2000™ reagent (Invitrogen, Waltham, MA, USA) in the presence of reduced serum medium (Invitrogen, Waltham, MA, USA) according to the manufacturer's instructions when EPC cells reached 90-95% confluency, 0.8 µg of pGL3 luciferase reporter vectors (Promega Corporation, Madison, WI, USA) were used for a co-transfection with 0.08 µg of the pRL-CMV vector (Promega Corporation, Madison, WI, USA) for luciferase analysis normalization. Transfected cells were incubated at 28 • C without CO 2 .
At 48 h post transfection, EPC cell extracts were collected and firefly and Renilla luciferase activities were measured using a Dual Glo Luciferase Assay System (Promega Corporation, Madison, WI, USA). Briefly, 75 mL of the remaining DMEM serum-free medium was mixed with 75 mL of Dual-Glo™ Luciferase Reagent (Promega Corporation, Madison, WI, USA) and incubated for 1 min. The lysates were measured for firefly luciferase activity in a 96-well microplate-reading luminometer (Veritas™ Microplate Luminometer, Promega Corporation, Madison, WI, USA). Each sample was normalized to the absorbance of Renilla luciferase to correct for variations in transfection efficiency using 75 mL of Stop & Glo ® Reagent (Promega Corporation, Madison, WI, USA) added to the same well and incubated for 10 min before detecting. Experiments were performed in duplicate.

Statistical Analysis
All data were analyzed using GraphPad Prism software (GraphPad, San Diego, CA, USA). We performed an independent t-test to examine whether the expressions of fads2a and fads2b genes were significantly different in each tissue or each hour post-fertilization we measured in the qPCR analyses. Asterisks denote statistical significance (* p < 0.05; ** p < 0.01; *** p < 0.001). All data are reported as mean ± SD. Means and standard deviations are from at least four independent experiments.