High levels of docosahexaenoic acid are present in eight New World silversides (Pisces: Atherinopsidae)

Docosahexaenoic acid (DHA) is the most critical and least available omega-3 fatty acid in the Western human diet. Currently, the source of omega-3 long chain polyunsaturated fatty acids (LC-PUFA) is mainly dependent on wild fisheries, making this resource unsustainable in the foreseeable future. In recent years, a high rate of biosynthesis and accumulation of DHA has been discovered in a freshwater species (Chirostoma estor) belonging to the Atherinopsidae family. Interest in evaluating fatty acid composition in other members of the family has emerged, so this study compiles original data of flesh composition of eight atherinopsid species from freshwater and brackish environments, either wild or cultured. High levels of DHA (16 to 31%) were found in all analyzed members of the family, except in C. grandocule, independently of their habitat or origin. The analyzed species of the Jordani group (C. estor, C. promelas and C. humboldtianum) showed high DHA and low EPA levels (<0.5%) as previously reported for cultured C. estor. The low trophic niche of these atherinopsids and their fatty acid accumulation capabilities are factors that make these species noteworthy candidates for sustainable aquaculture.

In recent years, a high rate of biosynthesis and accumulation of DHA has been discovered in the pike silverside (Chirostoma estor Jordan, 1880) (Fonseca-Madrigal et al., 2012, 2014, a freshwater species from the neovolcanic plateau of Central Mexico that belongs to the Atherinopsidae family. This species together with C. promelas Jordan &Snyder, 1899 andC. humboldtianum (Valenciennes, 1835) have an ancestral regional cultural value associated to human health (Martínez-Palacios et al., 2006). These species are nearly extinct due to habitat decay and overfishing driven by their cultural value and high local market prices ($ 30-60 US dollars/kg).
In particular, the pike silverside C. estor has been the most studied of the family in terms of fatty acid composition and/or metabolism, showing high proportion of DHA in flesh (31% of total lipids), providing 15 mg of DHA per gram of fillet, and 3/11 scielo.br/ni | sbi.bio.br/ni concomitantly low levels of EPA (less than 0.5%) (Martínez-Palacios et al., 2006). The mechanisms behind this proportion have been recently elucidated, finding two active DHA biosynthetic pathways: the Sprecher and a novel delta 4 pathway (Fonseca-Madrigal et al., 2014;Oboh et al., 2017).
Studies on the South American pejerrey species Odontesthes bonariensis (Valenciennes, 1835), another member of this family, corroborated similar DHA levels ( Gómez-Requeni et al., 2013;Bertucci et al., 2018). Therefore, considering the need for more sustainable aquaculture species and the fact that New World atherinopsids sit low in the trophic chain as secondary consumers (Martínez-Palacios et al. 2019), we aimed to analyze if other Neotropical Atherinopsid species captured from freshwater and brackish water environments, either wild or cultured, present a similar fatty acid profile.

MATERIAL AND METHODS
Fish sampling. Different species of atherinopsids were sampled, both from the natural environment and from farming systems of different regions. Within this family, the genus Chirostoma synonymous with Menidia (Miller et al., 2005), with species from the Mexican highlands, is divided into two large groups. The first, called Jordani, is composed of large fish, and in the second group, called Arge, is composed of small bait size fish, such as "charals" (Miller et al., 2005). All organisms were identified with dichotomous keys (Alvarez del Villar, 1970;Miller et al., 2005). Specifically, all species from Chirostoma genus were classified according to Miller et al. (2005) and Bloom et al. (2009). From the Jordani group, cultivated specimens (Freshwater/Cultured; FW/C) of C. estor and C. promelas were sampled from domesticated stocks held at Universidad Michoacana, Mexico. On the other hand, wild specimens (Freshwater/Wild; FW/W) of C. estor estor, C. estor copandaro and C. humboldtianum were sampled from the Mexican lakes of Pátzcuaro, Zirahúen and Zacapu, respectively. From the Arge group, wild brackish water (Brackish water/Wild; BW/W) Menidia beryllina (Cope, 1867) were caught in Sarasota Bay, Florida, USA and wild organisms (FW/W) were captured from Lake Pátzcuaro, Mexico: C. attenuatum (Meek, 1902) and C. grandocule (Steindachner, 1894 Fatty acid analysis. Analysis of saturated, unsaturated, polyunsaturated fatty acids (PUFA) and long-chain polyunsaturated fatty acids (LC-PUFA) such as docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA) and arachidonic acid (ARA), were carried out in all species. In the case of wild O. bonariensis, from Chasicó Lake (brackish water), Argentina, fatty acid values were taken from Kopprio et al. (2015).
Samples of all species were transferred on ice to the laboratory where they were subsequently frozen at -80°C until processing. Samples from Japan and USA were freeze-dried after their collection and transferred to the laboratory where they were analyzed following the same analytical protocol than all the other species.
Fatty acid composition was analyzed from fish muscle as previously described by Fonseca-Madrigal et al. (2012). Lipid fraction of the different samples was extracted using chloroform:methanol (2:1, v/v) containing 0.01% butylated hydroxytoluene (BHT). Fatty acid methyl esters (FAME) were prepared from extracted lipids by acid-catalyzed trans-esterification with 14% boron trifluoride in methanol at 90°C for 45 min and extracted using hexane. FAME were separated and quantified by gas chromatography with on-column injection using an Agilent Technologies GC 6850 equipped with a DB-23 silica column (30 m x, 0.25 µm x 0.25 mm, Agilent), flame ionization detector with helium as gas carrier (0.7 cm 3 min -1 ) and a temperature ramp (110 °C -220 °C). Individual methyl esters were identified by comparison with known standards (SIGMA, ALDRICH) and by reference to published data.
After normality and homogeneity of variance were analyzed, a Student´s t-test was applied between same species to find differences between individual fatty acids of wild and cultured fish (SIGMA PLOT 11.0).

RESULTS
All atherinopsid species analyzed contained between 24 and 37% of saturated fatty acids. On the other hand, monounsaturated fatty acids levels between 12.8 and 27.5% were observed for O. bonariensis (wild freshwater) and C. grandocule, respectively. In all cases, oleic acid (18:1n-9) was the most abundant fatty acid contributing to monounsaturated fatty acids levels, which had higher amounts (at least two times, compared to the wild organisms) in all cultured species analyzed (Tab. 1).
Eicosapentaenoic acid (EPA). Docosahexaenoic acid (DHA). * Indicate statistical differences (p<0.05). Student´s t-test was applied between same species to find differences between wild and cultured fish. +data published by Kopprio et al., 2015. Regarding total omega-3 PUFAs, concentrations ranging from 21 to 42% were observed. It should be noted that linolenic acid (18:3n-3) was found in high percentages in wild Chirostoma spp., compared to the cultured specimens. The same pattern was found in brackish water wild O. bonariensis, where cultured specimens were significantly lower. Freshwater O. bonariensis, together with charals C. attenuata and C. grandocule, presented the higher percentages of linolenic acid of all other analyzed species (Tab. 1).
The most important feature found in the studied atherinopsids was their high flesh docosahexaenoic acid (DHA) content (12.5-31% of total fatty acids). In general, wild specimens presented higher DHA levels than cultured ones, having C. estor estor the highest levels.

DISCUSSION
It is well known that in general, freshwater fish have a greater ability to biosynthesize LC-PUFA, such as DHA, than marine species. However, not all freshwater species studied to date accumulate abundant amounts of these fatty acids, as it depends on their environment, feeding habits and their de novo biosynthesizing capacities (Sargent et al., 2003).
A comparison of commercial fish species known to be a good source of omega-3 LC-PUFA have shown to contain high DHA levels ranging from 15 to 40% (Tab.2; Mohanty et al., 2016). Interestingly, these species are mostly marine carnivores, and thus can bio accumulate DHA directly from the food chain, independently of their biosynthesis capacity. On the other hand, C. estor a freshwater fish can accumulate similar levels of DHA directly by biosynthetic activity (Fonseca-Madrigal et al., 2012). In this work we further explore other commercially important atherinopsid species to find out if they share similar fatty acid profile. In fact, high levels of DHA (16 to 31%) were found in seven of the eight analyzed members of the family, independently of their habitat or origin. However, only the Jordani group (C. estor, C. promelas and C. humboldtianum) maintained the previously reported high levels of DHA with a concomitant low EPA levels (<0.5%), providing these species with a distinctive fatty acid profile (Tabs. 1-2).
Differences between captive and wild fish fatty acid profiles were found. Linolenic acid levels (18:3n-3) in cultured C. estor were lower than in wild fish, while the opposite was found in O. bonariensis, which had lower values in wild than in cultured fish. These contrasting findings could be explained by several factors: 1) The dietary contribution of linolenic acid by the inclusion of vegetable oils in cultured fish; 2) Differences in salinity of the environment; wild pejerrey were sampled from brackish water, which could promote DHA biosynthesis activity from linolenic acid as shown in C. estor 7/11 scielo.br/ni | sbi.bio.br/ni  (Fonseca-Madrigal et al., 2012), in contrast to cultured pejerrey in freshwater (Kopprio et al., 2015); 3) Different metabolic capacities between species to transform linolenic acid to DHA. Further studies are required to confirm the potential contribution of these or other factors. Oleic (18:1n-9) and linoleic (18:2n-6) fatty acids levels are higher in cultured than in wild fish, which could be reflecting the fatty acid content of the offered feeds, as observed in other fish species (Bell et al., 2003;Tocher, 2003). All species with the exception of charals (C. attenuata and C. grandocule) and cultured pejerrey O. bonariensis (FW/C), have low linolenic acid levels (<3.6%), suggesting that this fatty acid is being utilized as the precursor for de novo DHA biosynthetic pathway, as already described for C. estor (Fonseca-Madrigal et al., 2012). The high levels of linolenic acid found in O. bonariensis (FW/C) are probably due to their formulated diet.
Currently, aquaculture provides more than half of the total production of fish and seafood worldwide (F.A.O., 2018). However, fish culture is based mainly on the production of carnivore fish species (salmon, trout, tuna, etc.), which are on the top of the food chain and greatly depend on fisheries to provide fish oil and fishmeal as ingredients for their diets (Tacon, Metian, 2013), making it unsustainable in the foreseeable future. Efforts are currently being made to reduce costs and environmental impact by replacing marine fish sources with vegetable products. However, this strategy has led to a reduction of omega-3 levels in some aquaculture species (Shepherd et al., 2017;Sprague et al., 2017), thus reducing their nutritional value. Other species with low requirements of fishmeal and fish oil are also produced (such as tilapia and carp), however they do not contain high omega-3 concentrations, unless they are supplied in their diet, increasing production costs and reducing profitability.
For these reasons, FAO and several authors have recognized the need of selecting new aquaculture candidate species with more sustainable characteristics, such as low trophic level species, which translate in lower requirement of fishmeal and fish oil. In this respect, all species analyzed in this study are secondary consumers with filter-8/11 scielo.br/ni | sbi.bio.br/ni feeding characteristics and therefore low in the food chain (Horn et al., 2006;Ross et al., 2006;Martínez-Palacios et al., 2019). Thus, future large-scale production of some of these species may provide more sustainable alternative sources of DHA. This could also have the added benefit of avoiding pollutants normally acquired via bioaccumulation commonly present in marine fishmeal and fish oil (Tacon, Metian, 2013). Furthermore, regional and economical importance of atherinopsids together with their potential nutraceutical properties (i.e. salmon; Hasler, 2002) make them adequate species for aquaculture.