Comparsion of Activities of Fatty Acyl Desaturases and Elongases Among Six Teleosts With Different Feeding and Ecological Habits

Fatty acyl desaturases 2 (Fads2) and elongases of very-long-chain fatty acid 5 (Elovl5) are two key enzymes involved in the biosynthesis of long-chain polyunsaturated fatty acids (LC-PUFAs), and their activities determine the LC-PUFA biosynthetic ability of teleost. In order to investigate the relation of enzymic activities with fish’s feeding habits and ecological habits, the activities of Fads2 and Elovl5 were compared among six teleosts, namely, freshwater carnivorous mandarin fish (Siniperca chuatsi), freshwater herbivorous grass carp (Ctenopharyngodon idellus), marine carnivorous orange-spotted grouper (Epinephelus coioides), marine herbivorous rabbitfish (Siganus canaliculatus), anadromous Atlantic salmon (Salmo salar), and catadromous Japanese eel (Anguilla japonica). Among them, the enzymatic features of Fads2 and Elovl5 from the last five fish species have been characterized, whereas those of S. chuatsi were unknown. And thus, the coding sequences (CDSs) of S. chuatsi fads2 and elovl5 (elovl5a and elovl5b) were isolated, and their functions were further characterized by heterologous expression in yeast. The results showed that S. chuatsi Fads2 has a monofunctional Δ6 desaturase and that Elovl5a has a higher activity toward C18–C20 PUFAs than has Elovl5b, which showed a noteworthy activity toward C22 PUFAs. The comparison of enzymatic activities among the six teleosts showed that the Δ6 Fad and Elovl5 activities varied markedly among fish species; in particular, the activity of Δ6 Fad in C. idellus, S. canaliculatus, and A. japonica was significantly higher than that in S. chuatsi, S. salar, and E. coioides. For C18 PUFA substrates, A. japonica Elovl5 has a higher elongation than has the other tested fish, and it exhibits a higher activity toward the C20 PUFAs. The results suggest that the Δ6 Fad activity is influenced by both feeding habits and ecological habits, whereas the Elovl5 activity was more affected by the feeding habits. These data enrich our knowledge on LC-PUFA biosynthesis diversity of fatty acid desaturation and elongations in teleosts and provide guidance for the choice of dietary PUFA precursors for farmed fish.


INTRODUCTION
The health benefits of fish consumption are derived from n-3 long-chain polyunsaturated fatty acids (n-3 LC-PUFA), eicosapentaenoic acid (EPA; 20:5n-3) and docosahexaenoic acid (DHA; 22:6n-3). These above-mentioned bioactive molecules are involved in maintaining the normal development of the nervous system (Uauy et al., 2001) and improving in lipid metabolism, inflammatory response, and cardiovascular and neurological health (Delgado-Lista et al., 2012;Awada et al., 2013). Farmed fish are becoming an increasingly important source of LC-PUFA in the human diet, because the wild fishery stocks are declining (FAO, 2014). Fish oils (FOs), rich in digestible energy and n-3 LC-PUFA, are considered as the most important raw materials for aquafeeds (especially for carnivorous fish feed). During the recent two decades, the supplementation of C18 PUFA-rich plant oils [vegetable oils (VOs), devoid of n-3 LC-PUFA] was increased to replace FOs in fish formula feed, because of the limited and increasingly expensive FO resources (Turchini et al., 2011;Lu et al., 2017). Consequently, the increasing supplementation of VO in aquafeeds has resulted in decreasing the levels of n-3 LC-PUFA in farmed fish, especially in marine fish (Hossain, 2011;Sprague et al., 2016).
The extent to which fish species have the LC-PUFA biosynthetic capability varies with species and is associated with not only their complement of fads and elovl genes but also the activities of those key enzymes (Fonseca-Madrigal et al., 2014). 6 Fad and Elovl5 have been cloned and characterized from all fish so far investigated including marine carnivorous fishes (Tocher, 2010). However, it remained unclear whether the enzymatic activities of 6 Fad and Elovl5 are linked to the habitat, trophic level, and ecology of fish species. To this end, the coding sequences (CDSs) of putative 6 Fad and Elovl5 were isolated from the freshwater carnivorous fish, mandarin fish (Siniperca chuatsi), and functionally characterized by heterologous expression in yeast (Saccharomyces cerevisiae). The enzymic activities of 6 Fad and Elovl5 among six fish species with different habitats and trophic ecologies, namely, freshwater herbivorous fish grass carp (Ctenopharyngodon idellus), S. chuatsi, marine carnivorous fish orange-spotted grouper (Epinephelus coioides), marine herbivorous fish rabbitfish (Siganus canaliculatus), anadromous fish Atlantic salmon (Salmo salar), and catadromous fish Japanese eel (Anguilla japonica), were determined by heterologous expression in the yeast. The results were expected to identify the relationships between the enzymatic activities of 6 Fad and Elovl5 and the habitat, trophic level, and ecology of fish species and to increase our knowledge of the molecular basis of LC-PUFA biosynthesis and its regulation in teleosts.

Siniperca chuatsi
In order to clone the fads2 and elovl5 CDSs of S. chuatsi, liver samples were collected from three S. chuatsi individuals (about 250 g) brought from the local fish market, after the fish were anesthetized with 0.01% 2-phenoxyethanol. Tissue samples were frozen in liquid nitrogen immediately after collection and stored at -80 • C until RNA extraction. The liver tissue of S. chuatsi was sampled, and the total RNA extracted using TRIzol reagent (Invitrogen, United States) was reverse transcribed into cDNA using random primers and an appropriate RT-PCR kit (Invitrogen, United States). The CDS-specific primers of 6 fad and elovl5 containing restriction sites BamHI and XbaI were designed on the basis of S. chuatsi 6 fad (EU683737) and elovl5 (EU683736) ( Table 1). PCR was performed using high-fidelity DNA polymerase (TianGen, Beijing, China). For all genes, PCR consisted of an initial denaturation at 94 • C for 4 min, followed by 30 cycles of denaturation at 94 • C for 30 s, annealing at 55 • C for 30 s, and extension at 72 • C for 1 min, followed by a final extension at 72 • C for 10 min. The resulting PCR fragments were sequenced (Sangon, Shanghai, China).

Ctenopharyngodon idellus, Siganus canaliculatus, Epinephelus coioides, Salmo salar, and Anguilla japonica
The recombinant plasmids containing 6 fad and elovl5 CDSs of Ctenopharyngodon idellus, Siganus canaliculatus, Epinephelus coioides, S. salar, and Anguilla japonica were constructed and kept in our laboratory . The 6 fad and elovl5 CDSs of C. idellus, S. canaliculatus, E. coioides, S. salar (elovl5a and elovl5b), and A. japonica were cloned from the corresponding recombinant plasmids with CDS-specific primers ( Table 1). For all genes, PCR was performed using high-fidelity DNA polymerase (TianGen, Beijing, China) under the following conditions: initial denaturation at 94 • C for 4 min, followed by 30 cycles of denaturation at 94 • C for 30 s, annealing at 55 • C for 30 s, and extension at 72 • C for 1 min, followed by a final extension at

Sequence and Phylogenetic Analysis of 6 Fad and Elovl5
The amino acid (aa) sequences of the cloned 6 Fad and Elovl5 were aligned among the six species by ClustalW2. 1 Alignments and similarity matrices were calculated using the EMBOSS Needle Pairwise Sequence Alignment tool. 2 Phylogenetic analysis was performed by constructing a tree using the neighborjoining method (Saitou and Nei, 1987). Confidence in the resulting phylogenetic tree branch topology was measured by bootstrapping through 1,000 iterations.

Heterologous Expression of 6 Fad and Elovl5 Coding Sequences in Yeast
To maximize the reliability of results, the comparison of Fads2 and Elovl5 conversion rates among the six species was performed in the same experimental conditions. Briefly, the recombinant plasmids (0.8 µg) containing 6 fad and elovl5 CDS of C. idellus, S. canaliculatus, E. coioides, S. salar (elovl5a and elovl5b), and A. japonica were sequenced and transformed into yeast (strain INVSc1, Invitrogen) using the S.C. Easy Comp Transformation kit (Invitrogen, United States). Recombinant yeast solution (100 µl) was coated on the SCMM uraci plate and cultured for 3 days; a single colony was picked and propagated in the SCMM uracil culture. For determining the 6 Fad activity, the recombinant yeast was cultured in SCMM uracil and supplemented with C18 PUFA substrates, LNA or LA. For testing the Elovl5 activity, the transgenic yeast was supplemented with one of the PUFA substrates from among the following: 18:3n-6, 18:4n-3, ARA, and EPA. The PUFA substrates were added at final concentrations of 0.5 (C18) and 0.75 (C20) mM. After 2 days (the OD 600 of bacterium solution reached 10), recombinant yeast cells were harvested and washed as described previously (Li et al., 2010;Xie et al., 2014).

Lipid Extraction and Fatty Acid Analysis
The total lipid of yeast samples was extracted by homogenization in chloroform/methanol (2:1, v/v) containing 0.01% BHT (Sigma, United States) (Folch et al., 1957). Fatty acid methyl esters (FAMEs) from yeast total lipids were prepared by transesterification with boron trifluoride etherate (ca. 48%, Acros Organics, Morris Township, NJ, United States) as described previously (Li et al., 2010). FAMEs were determined by the gas chromatograph GC2010-plus (Shimadzu, Japan). The parameters were the same as we used before (Li et al., 2010). The conversion rates of genes were calculated as follows: 100 × [product areas/(product area + substrate area)] (Li et al., 2010).

Statistical Analysis
Genes conversion rates were presented as means ± standard error of the mean (n = 3). Differences among the six fish species were tested using one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison. Differences were considered significant at P < 0.05. All analyses were conducted using SPSS v17.0 (SPSS, Inc., Chicago, IL, United States).

Sequence and Phylogenetic Analysis of 6 Fad Coding Sequences
The  Figure 1). The 6 Fad polypeptides of C. idellus, S. chuatsi, E. coioides, and A. japonica deduced from their CDS obtained in the present study showed 1, 3, 1, and 1 aa differences, respectively, compared with previously published results (Li et al., 2010(Li et al., , 2014Du et al., 2011;Wang et al., 2014). The polypeptides of S. canaliculatus and S. salar 6 Fad obtained in this study showed the same sequences as previously published (Li et al., 2010;Monroig et al., 2010).
The 6 Fad polypeptide of C. idellus had 65. 86, 68.99, 69.21, 69.59, and 71.91% sequence identities with other 6 FIGURE 2 | Phylogeny of deduced amino acid (aa) sequences of Fads2 of different fish species. The tree was constructed using the neighbor-joining (N-J) method (Saitou and Nei, 1987) using CLUSTALX and MEGA 7.0. The horizontal branch length is proportional aa substitution rate per site. The numbers represent the frequencies with which the tree topology presented here was replicated after 1,000 bootstrap iterations.
Phylogenetic analysis compared all the deduced amino acid sequences of Fads2 along with a variety of 4 Fad, 5 Fad, and 6 Fad desaturases from different fish species (Figure 2). The result showed that teleost Fads2 sequences cluster according to accepted habitat ecology as displayed in the phylogenetic tree, with all the marine fish Fads2 sequences clustered together and more closely related to migration fish Fads2 sequences than freshwater fish species sequences. The salmonid sequences clustered together with the other diadromous fish species, A. japonica. Interestingly, the Nile tilapia and S. chuatsi sequence clustered closer to the marine fish than to the other freshwater fish, clarias leather (Clarias gariepinus), carp, and zebrafish (Danio rerio), which clustered together.
A neighbor-joining phylogenic tree was constructed based on the deduced Elovl5 aa sequences from different species (Figure 4). Our results showed that the phylogenic tree of Elovl5 homologs is almost in accordance with the affinity of those fish. The phylogenic tree of Elovl5 homologs does not really cluster FIGURE 4 | Phylogeny of deduced amino acid (aa) sequences of Elovl5 of different fish species. The tree was constructed using the neighbor-joining (N-J) method (Saitou and Nei, 1987) using CLUSTALX and MEGA 7.0. The horizontal branch length is proportional aa substitution rate per site. The numbers represent the frequencies with which the tree topology presented here was replicated after 1,000 bootstrap iterations. according to habitat ecology as displayed in the phylogenetic tree of teleosts Fads2. The salmonid sequences clustered together with a freshwater fish, Esox lucius, but not with the other diadromous fish species, A. japonica. Interestingly, Oreochromis niloticus and S. chuatsi Elovl5 sequences clustered more closely to the marine fish than to the other freshwater fish.

Functional Characterization
The functional characteristics of putative 6 Fad and Elovl5 isolated from S. chuatsi were determined by heterologous expression in yeast S. cerevisiae grown in the presence of potential FA substrates. The FA profiles of control yeast transformed with the empty pYES2 vector was 16:0 and 16:1 isomers (16:1n-9 and 16:1n-7), 18:0 and 18:1n-9, and any exogenously added PUFA substrate (data not shown).
The FA profile of yeast transformed with S. chuatsi putative 6 Fad cDNA additionally showed extra peaks when grown in the presence of 18:2n-6 and 18:3n-3, which corresponded to 18:3n-6 and 18:4n-3, respectively, whereas 5 and 4 activities were not detected ( Figure 5 and Table 2). These data show clearly that the cloned S. chuatsi putative Fads2 had 6 Fad specificities. When the S. chuatsi putative Elovl5a/b cDNA was expressed in the yeast cells, evidence of elongation of all fatty acids was observed (Figures 6, 7 and Table 2). Generally, n-3 PUFAs were elongated to a greater extent compared with their corresponding n-6 isomers. Moreover, S. chuatsi Elovl5a had an apparent preference for C18 and C20 over C22 FA substrates.
Interestingly, the Elovl5b more effectively converted C22 PUFA substrates than did C18 and C20 substrates.
For the six teleosts, the conversion rates of Elovl5 toward fatty acid substrates are shown in Table 4 and Supplementary  Figures 5-8. Generally, high elongations were obtained with n-3 compared with n-6 PUFAs. For C18 PUFA substrates (Supplementary Figures 5, 6), A. japonica Elovl5 had higher elongations than those in other fish species but exhibited lower activities toward the C20 PUFAs (Supplementary Figures 7, 8).
but the conversion function of S. salar Elovl5a and Elovl5b showed no difference.

DISCUSSION
The present study provided evidence for the existence of both fads2 and elovl5 encoding cDNAs and demonstrated their role in the LC-PUFA biosynthesis of the freshwater carnivorous Siniperca chuatsi. The aa sequence of S. chuatsi Fads2 possesses all the main structural features common for Fads protein family members, including N-terminal cytochrome-b5-like domains, the heme-binding motif HPGG, three histidine boxes, and four predicted transmembrane domains (Hashimoto et al., 2008). Phylogenetic analysis revealed that S. chuatsi Fads2 shares high homology with other teleosts, and the Fads2 of representatives basically cluster together by their habitats. However, S. chuatsi Fads2 is not closely related to that of the other freshwater fish species but rather to the marine carnivorous fish species (Figure 2). Consistently, the S. chuatsi Fads2, like the majority of Fads2 isolated from marine carnivorous fish (Tocher et al., 2006;Morais et al., 2012), has a monofunctional 6 desaturase and does not appear to possess 5 or 4 activities as in striped snakehead (Kuah et al., 2015). Attempts have been made to search another fads2 gene from S. chuatsi genome data, but these have so far been unsuccessful. Although the S. chuatsi seems more like a marine fish without LC-PUFA biosynthesis ability, a feeding trial demonstrated that total replacement of dietary FOs with alternative VO has no negative impact on the growth performance and health of mandarin fish juvenile, which indirectly suggested this species could bioconvert C18 PUFA to their corresponding LC-PUFA (Sankian et al., 2019). Collectively, these findings highlight that teleosts have an adaptive plasticity and diversity of LC-PUFA biosynthesis mechanism (Fonseca-Madrigal et al., 2014).
In the present study, two elovl5 CDSs (elovl5a and elovl5b) were identified in S. chuatsi, the likes of which have been seen in Atlantic salmon and common carp (Cyprinus carpio var. Jian) (Morais et al., 2009;Ren et al., 2012). Analysis of the deduced aa sequences of S. chuatsi Elovl5a and Elovl5b showed that they both have all the typical characteristic features of the predicted transmembrane domains, the histidine box, and the canonical C-terminal ER retrieval signal (Jakobsson et al., 2006). Phylogenetic analysis showed that the Elovl5 homologs are in accordance with the order of fish but not with their feeding habits and habitat. All the Elovl5 sequences of Pereiformes, Salmoniformes, and Cypriniformes clustered together (Figure 4). A similar phylogenetic grouping was observed previously for the sequences of elongases from a range of teleosts (Agaba et al., 2005). The results of the functional characterization revealed that the capability of S. chuatsi Elovl5a and Elovl5b, similar to the other Elovl5 homologs, exhibits an effective ability to elongate both C18 and C20 PUFA and displays a preference to elongate n-3 PUFA substrates compared with n-6 PUFA substrates (Monroig et al., 2011;Kuah et al., 2015;Xie et al., 2016;Janaranjani et al., 2018;Ferraz et al., 2019). Interestingly, S. chuatsi Elovl5a has a higher activity toward C18-C20 FAs than has Elovl5b, whereas Elovl5b showed a noteworthy activity toward C22 FAs (60.47% conversion of 22:5n-3) as the Elovl2 does (Morais et al., 2009;Gregory and James, 2014;Oboh et al., 2016). As for most fish species, S. chuatsi elovl2 cDNA has not been isolated successfully (data not shown), whereas in these teleost, the conversion of C22 FAs by other elongases, such as Elovl5 and Elovl4, potentially compensates for the absence of Elovl2 in DHA biosynthesis (Morais et al., 2009;Wang et al., 2014;Xie et al., 2016).
To date, fads2 cDNAs have been identified in numerous fish species, and all tested Fads2 enzymes showed the ability to operate as 6 Fad, which is likely the primary function of Fads2 for teleosts (Castro et al., 2016;Janaranjani et al., 2018;Ferraz et al., 2019). In the present study, the 6 Fad activity of six tested fish species was performed in the same yeast expression system, which showed high efficiency, with conversion rates of 18:2n-6 and 18:3n-3 ranging at 30.38-60.6 and 52.61-79.23%, respectively. However, the activities of desaturases varied markedly among these species. The conversion    Values (mean ± SEM of three replicates) in each row with different superscript letters are significantly different (P < 0.05). efficiency of 6 Fad in marine carnivorous fish Epinephelus coioides is significantly lower than that in other fish species. Previous studies reported that the conversion rates of 6 Fad in grouper, rabbitfish, salmon, and eel were 4.4-9.78% (Li et al., 2014), 35-59% (Li et al., 2010), 25-47% , and 20.7-60.8% , respectively. Although these data of genes conversion rates in the present study are somewhat different from those in the previous studies; the order of genes conversion rates in grouper, rabbitfish, salmon, and eel is consistent in the present study and previous works. Consistently, nutritional trials have shown that VO can satisfy the essential fatty acid (EFA) requirements of Ctenopharyngodon idellus, Siganus canaliculatus, Anguilla japonica, S. chuatsi, and Salmo salar; and their 6 fad gene expression was increased by dietary VO (rich in C18 PUFA) (Takeuchi et al., 1980;Monroig et al., 2011;Lei et al., 2017;Xie et al., 2018;Sankian et al., 2019), whereas E. coioides 6 Fad had a low enzymatic activity in converting LNA and LA because of the deficiency of binding site for the stimulatory protein 1 (Sp1) in its promoter (Li et al., 2014;Xie et al., 2018). Among the freshwater species, herbivorous C. idellus has a higher 6 Fad activity than has carnivorous S. chuatsi, and the similar effect of feeding habit on the desaturase's activity was also exhibited between the marine species. Among the carnivorous fish species, the highest 6 Fad activity was detected in the catadromous A. japonica, followed by freshwater and anadromous species, and the lowest in marine species, whereas the influence of habitats on the 6 Fad activity was little in the herbivorous fish species. Those results suggested that the 6 Fad activity of fish is under the influence of both feeding habits and habitats. Furthermore, the 8 activity of Fads2 varied notably among the different fish species, and a higher 8 capability was detected in marine fish compared with freshwater/diadromous species (Monroig et al., 2011). On the other hand, those results confirmed that the functions and capabilities of teleost Fads2 have diversified remarkably as a result of environmental factors including habitat, trophic level, and ecology (Castro et al., 2016). Besides Fads2, the above-mentioned environmental factors have also been suggested as potential drivers modulating the elongation capabilities of teleosts (Agaba et al., 2005;Carmona-Antoñanzas et al., 2013;Wang et al., 2014;Janaranjani et al., 2018). In a comparison of Elovl5 activities on C18-C22 PUFA substrates among seven fish species, Elovl5 activities were more likely to elongate n-3 substrates than n-6 substrates, with the exception of the Atlantic cod (Gadus morhua, Gadiformes) elongase, which was more active toward the n-6 homologs (Agaba et al., 2005). Simiarly, Elovl5 activities of all fish from different ecological backgrounds were high toward n-3 PUFAs, whereas the S. chuatsi Elovl5b exhibited a preference for n-6 substrates. The pattern of activities on different PUFAs substrates showed that the Elovl5 from A. japonica and S. salar exhibited a rank order of C18 > C20, which was very similar to those of most fish species (Castro et al., 2016;Xie et al., 2016;Janaranjani et al., 2018;Ferraz et al., 2019). The Elovl5 from S. chuatsi and E. coioides showed a similar activity with C18 and C20 PUFAs. Interestingly, herbivorous C. idellus and S. canaliculatus elongases were less active toward C18 substrates than were the other elongases, but they displayed higher activities toward the C20 PUFAs , which might be linked to the abundance of C18 PUFAs and limited C20 PUFAs in their food web. In general, the trophic level and ecology have a bigger impact on the Elovl5 activity of teleosts than the habitat.

CONCLUSION
The present study demonstrated S. chuatsi Fads2 with 6 Fad capabilities, and its Elovl5a showed a preference toward n-3 C18-C20 PUFAs, whereas the Elovl5b showed substrate specificity toward C22 PUFAs. Furthermore, the desaturation and elongation capabilities of 6 Fad and Elovl5 were compared among six fish species from different ecological backgrounds, which indicated that the 6 Fad activity of fish is under the double influence of feeding habits and habitats, whereas the Elovl5 activity of teleosts was affected more by the trophic level and ecology. Those differences in the functional competences of the 6 Fad and Elovl5 from different fish species may contribute to the different LC-PUFA biosynthesis abilities of the species. These results increase our knowledge of the molecular basis of LC-PUFA biosynthesis and its regulation in teleosts and provide guidance on choosing suitable dietary PUFA precursors for those farmed fish species.
This study has a drawback. The present study compared the conversion rates of Fads2 and Elovl5 among the six teleosts under the same in vitro conditions, which is not enough to explain the difference of Fads2 and Elovl5 activities in vivo. The enzymatic activities are related to the enzymatic kinetics, such as the Michaelis constant (Km) and maximum velocity of the reaction (Vmax). Therefore, more studies are needed to fully investigate the enzymatic kinetics of Fads2 and Elovl5.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by the statement to confirm that all experimental protocols were approved by the Guangdong Provincial Department of Science and Technology on the use and care of experimental animals. The study was reviewed and approved by the Ethics Committee of Animal Experiments of South China Agricultural University.

AUTHOR CONTRIBUTIONS
DX and YL wrote the manuscript. SW, CY, and YL designed the study. JY and ML conducted the experiments. All authors read and approved the final version of the manuscript.

FUNDING
This work was financially supported by the National Natural Science Foundation of China (31873040) and National Key R&D Program of China (2018YFD0900400).