Prey quality impact on the feeding behavior and lipid composition of winter flounder(Pseudopleuronectes americanus)larvae

Maria Angelica Martinez-Silva,Céline Audet,Gesche Winkler,Réjean Tremblay

Institute de sciences de la mer à Rimouski,Université du Québec à Rimouski,310 Allée des Ursulines,Rimouski,QC G5L 2Z9,Canada

Keywords:Winter flounder Copepods Rotifers Nutrition Fatty acids

A B S T R A C T Copepods are the main natural food of many marine fish larvae.However,enriched rotifers are the most commonly used live prey in larval rearing.Impacts on the feeding behavior,growth,survival,and fatty acid(FA)composition of winter flounder larvae fed with copepods and rotifers were determined and compared to the FA composition of the two live prey.Nauplii of Eurytemora spp.and Acartia sp.,two of the main species of copepods present in the St.Lawrence estuary,showed no significant differences in their essential fatty acid profiles,suggesting similar nutritional quality.Thus,only Eurytemora herdmani was compared to enriched rotifers in this study.Copepod nauplii were characterized by higher levels of essential fatty acids,particularly docosahexaenoic acid(DHA)and eicosapentaenoic acid(EPA).The selective incorporation of essential fatty acids from diets in larval tissues(polar lipids)indicated that nauplii might better fulfill larval nutritional requirements for DHA than rotifers.Furthermore,larval behavior was modified according to the diet:larvae fed with nauplii spent more time swimming with no changes in the occurrence of hunting events.

1.Introduction

In nature,it has been demonstrated that marine fish larvae mainly feed on zooplankton,especially copepod nauplii and copepodites(Llopiz,2013;Robert,Murphy,Jenkins,&Fortier,2014).However,in marine larval fish rearing,which is the case with winter flounder(Pseudopleuronectes americanus),the most common live prey used as a proxy for copepods are rotifers(Fraboulet,Lambert,Tremblay,&Audet, 2011,2010).Rotifers are easy to produce in high abundance year-round,they are an edible size for many species of fish larvae,and their slow swimming speeds make them easy for fish larvae to capture(Øie,Reitan,Evjermo, Stottrup,&Olsen,2011).However,the nutritional value of rotifers is generally considered poor for marine fish larvae because of their low levels of essential fatty acids(EFA;Castell et al.,2003).

EFA include eicosapentaenoic acid(20:5n-3,EPA),docosahexaenoic acid(22:6n-3,DHA),and arachidonic acid(20:4n-6,AA),which have been reported to be essential for the optimal growth of several marine fish species whose biosynthetic production is insufficient to meet their nutritional requirements(Glencross,2009).EFA functions can be divided into two broad areas,with EPA and DHA involved in maintenance of the structural and functional integrity of biological membranes(Hazel,1995),while AA and EPA act as precursors of eicosanoids,a group of highly biologicallyactive hormones(Howard&Stanley,1999).Several studies have evaluated different ways to enrich rotifers to obtain the best proportions of essential fatty acids(Castell et al.,2003;Haché&Plante,2011;Haché,Plante,Forward,&Pernet,2016;Vagner,de Montgolfier,Sévigny,Tremblay,&Audet,2014).

The optimization of EFA composition in marine fish feed is complicated by the competitive interaction between EPA and DHA for phospholipid biosynthesis and between AA and EPA for the eicosanoid response(Sargent et al.,1999).Thus,the correct balance of AA,EPA,and DHA with total fatty acid content is important to avoid nutritional deficiency(Glencross,2009).Seychelles,Audet,Tremblay,Fournier,and Pernet(2009;2011)and Vagner,de Montgolfier,Sevigny,Tremblay,and Audet(2013;2014)used different rotifer enrichments with specific proportions of EFA to rear winter flounder larvae.They found that while low levels of highly unsaturated fatty acids(HUFA)did not affect growth performance orlipid reserves,an essential combination of EPA,AA,and DHA is required to sustain the up-regulation of growth hormone gene expression through out larval development and metamorphosis in winter flounder.These authors found that the best combination was an enrichment containing an EPA/DHA/AAratio of around 4/3/1.

Despite fast growth in winter flounder reared with enriched rotifers,a high larval mortality rate persists at metamorphosis(Audet&Tremblay,2011).Mercier et al.(2004)suggested that the use of another live prey more representative of the natural environment of larvae could be an interesting hypothesis to test for optimizing larval rearing of winter flounder.Therefore,a new alternative in aquaculture is the use of copepods as live feed,since they are natural prey(Ajiboye,Yakubu,Adams,Olaji,&Nwogu,2011;Hernandez-Molejon&Alvarez-Lajonchere,2003;Støttrup,2003;van der Meeren,Olsen,Hamre,&Fyhn,2008)and some copepods are exploited successfully in extensive systems in Taiwan and Denmark(Hansen,2017).

Several studies have established that copepods are naturally richer in EFA than rotifers(Avella,Olivotto,Gioacchini,Maradonna,&Carnevali,2007;Barroso,de Carvalho,Antoniassi,&Cerqueira,2013;Støttrup&Norsker,1997),with higher levels of phospholipids(＞50%)(McEvoy,Naess,Bell,&Lie,1998);this could improve rearing success because this lipid class seems to be more easily assimilated by various fish species (Gisbert, Villeneuve,Zambonino-Infante,Quazuguel,&Cahu,2005;Tocher,Bendiksen,Campbell,&Bell,2008).Another important advantage of copepods is that their size varies according to their ontogeny,providing prey of different size ranges for marine fish larvae and juveniles(Lee,O'Bryen,&Marcus,2005).Shaheen et al.(2001)evaluated the preference of winter flounder larvae for two copepod species(E.affinisandA.hudsonica)and demonstrated that behavior and morphology of the prey are key factors in selection by fish.However,the relationship between prey preferences and their nutritional value in the early stages of winter flounder remains unknown.Furthermore,knowledge of the biochemical and nutritional value of copepods for early stages of marine fish is fragmentary.

The objectives of this current study were(1)to characterize the composition,abundance,and fatty acid content of adult copepods from a natural environment in order to know their availability and their nutritional value as live prey for winter flounder rearing;(2)to compare the nutritional value of copepod nauplii and enriched rotifers with the larval rearing success of winter flounder;and(3)to analyze and compare the feeding behavior(swimming,resting,and hunting)of larvae fed with copepod nauplii or rotifers.We hypothesized that copepod nauplii would be more preyed upon and more appropriate to meet nutritional needs.This study highlights new knowledge of the nutritional needs and feeding behavior of winter flounder larvae.

2.Material and methods

2.1.Copepod sampling

Three stations were sampled in the St.Lawrence estuary(Riviˋere-du-Loup:47°84′N,69°57′W;Kamouraska:47°56′N,69°87′W;Riviˋere-Ouelle:47°48′N,70°02′W)on five occasions(1,16 June;1,29 July;14 August 2015).Seawater salinity and temperature were measured at the surface at the beginning and end of sampling at each site with a YSI professional plus probe.Zooplankton was obtained by 10 min horizontal surface tows from the dock using a ring net(0.5 m diameter,245μm mesh size)equipped with a flux meter in the middle of the net.For each station,one sample was preserved in ethanol(95%),and placed in the freezer(-20°C)until laboratory identification;after 24 h,ethanol was totally replaced.Zooplankton from the second sample were preserved in seawater at-80°C for lipid analyses.Three 10 ml subsamples from a 1 L sample were taken for zooplankton identification(minimum of 300 individuals per subsample)and abundance determination using a counting chamber under a stereomicroscope.

2.2.Copepod nauplii production

Adult copepods(Eurytemora herdmani)were collected in June 2016at the Riviˋere-du-Loup site and females with egg sacks were placed in ovitraps(100m mesh size)within 10 L tanks filled with 1μm filtered seawater at 18°C and salinity 27.Hatched nauplii were collected and counted daily,thenplaced in a 1L bottle at 10°C with 1μm filtered seawater at 18°C,salinity 27,and low aeration until use for larval feeding experiments.

2.3.Winter flounder larval production

All experiments were conducted at the aquaculture station of the Institute des Sciences de la Mer de Rimouski,Québec,Canada(48°27′N,68°32′W)from April to August 2016.Female winter flounder broodstock were collected in the Baie-des-Chaleurs,Québec,Canada(48°10′N,67°30′W)and larvae were obtained using the fertilization protocol described by Ben Khemis,de la Noue,and Audet(2000).The rearing protocol was similar to the one used by Vagner et al.(2014).Newly hatched larvae were briefly reared in 55 L cylindroconical polyethylene tanks at a rearing density of 1 larva·ml-1,supplied with filtered sea water(10 μm),and maintained at 10°C with a 12 L:12D photoperiod.Permanent upwelling was maintained in each tank by an aeration system.From the mouth-opening stage(4 days post-hatch[dph])until the beginning of experiments,larvae were fed daily with the rotiferBrachionus plicatilisenriched with Sparkle(Selco,INVE aquaculture Ltd.,Thailand)at a concentration of 12 individuals ml-1.In larval rearing tanks,the water supply was stopped each day for 12 h while a green water preparation(Nannochloropsis oculataat 1.6×106cells·L-1)was added to each tank.At the end of the day,water circulation resumed,allowing complete renewal of the tank water during the night.Tank bottoms were cleaned weekly.

2.4.Larval feeding experiments

E.herdmani nauplii and enriched rotifers used for larval feeding experiments were maintained in 0.2μm filtered and aerated seawater(salinity 27)and fed daily with a mix of microalgae(Nannochloropsis oculata,Isochrysis galbana,andPavlova lutheri,1:1:1).Two similar experiments were carried out with two different batches of winter flounder larvae of the same age(16 dph).Larval growth,survival,and fatty acid composition were evaluated in six 20 ml well plates after feeding withE.herdmaninauplii orBrachionus plicatilisfor six days.Two winter flounder larvae were placed in each well(total of 72 larvae,for 36 replicates)filled with 20 ml of filtered(1μm)green seawater(as already described)at 10°C and salinity 27.Larvae were transferred every day to a new well containing 24 prey(nauplii or rotifers).After six days,the surviving larvae were preserved at-80°C for lipid analyses.

Specific growth rate(SGR)was calculated according to the following equation(where wt.is weight and the time interval used was six days;Cook,Mcniven,Richardson,&Sutterlin,2000):

2.5.Larval behavior analyses

Larval feeding behavior in the six 20 ml well plates(three well plates per treatment)was recorded.Each well contained two winter flounder larvae 16 day post-hatch.Two trials were performed on two different days on larvae of a similar age.Two cameras(GoPro HERO4 Silver+LCD[CHDHY-401]),equipped with a macro polar pro filter were positioned over the three well plates per treatment to give a complete view of each well for video recordings.The behavioral responses of larvae were videotaped and analyzed during the first ten minutes after feeding using NOLDUS OBSERVER XT 9.0 software(Noldus Information Technology Inc.,Leesburg,VA,USA).Behavior was determined and quantified as the time spent(percentage of the total observation time)swimming and at rest(as state events)and hunting(point event).

2.6.Fatty acid(FA)composition

For adult copepods of Acartia sp.and E.herdmani,60 individuals per sampling period per site were kept on ice and randomly selected under stereomicroscope,for a total of 1840 individuals.Each individual was rinsed with 1μm filtered seawater to remove particles on their exoskeleton(Cabrol,Winkler,&Tremblay,2015).Half of the individuals were used for FA analyses and preserved in 4 ml amber glass vials with 3 ml of dichloromethane:methanol(2:1 v/v)under a nitrogen atmosphere at-80°C.The remaining individuals were used for dry mass estimation(60°C for 48h).

ForE.herdmaninauplii and rotifers used in feeding experiments,200 individuals were randomly selected and filtered on a precombusted(450°C for 2 h)and pre-weighed GF/C filter(25mm)for FA analysis and 200 others were used for dry mass determination.For winter flounder larvae,six larvae per treatment and eight replicates were used for FA composition analyses and six others for dry mass determination.

Lipids were extracted in dichloromethane-methanol using the modified Folch procedure(Folch,Lees,&Sloane-Stanlez,1957)described in Parrish(1987).Total fatty acid composition was analyzed for wild copepods and prey(nauplii and rotifers),but neutral and polar lipids were separated from larvae to determine the fatty acid composition of each fraction.The quantification of total fatty acids was made relative to wet mass.The neutral(including triacylglycerols,free fatty acids,and sterols)and polar(including mainly phospholipids)fractions were separated by silica gel(30×5 mm i.d.,packed with Kieselgel 60,70-230 mesh;Merck,Darmstadt,Germany)hydrated with 6%water,and eluted with 10 ml of chloroform:methanol(98:2 v/v)for neutral lipids followed by 20 ml of methanol for polar lipids.The neutral fraction was further eluted on an activated silica gel with hexane and diethyl ether to eliminate the free sterols(Budge,Iverson,&Koopman,2006;Parrish,1987).Each fraction was prepared as described in Lepage and Roy(1984)and analyzed using a multichannel Trace GC ultra(Thermo Scientific)gas chromatograph equipped with a model Triplus autosampler,PTV injector,and model ITQ900(Thermo Scientific)mass detector,and analyzed with Xcalibur v.2.1 software(Thermo Scientific,Mississauga,ON,CA).FAMEs were identified with known standards(Supelco 37 Component FAME Mix and menhaden oil;Supleco Inc.,Belfonte,PA,USA)after manual verification of the FA integration.

2.7.Statistical analyses

PRIMER software(version 7.0.11)was used for non-metric multi-dimensional scaling(n-MDS)and PERMANOVA analyses based on Bray-Curtis dissimilarities following validation of the assumptions of homoscedasticity using PERMDISP tests.When necessary,an arcsine square root transformation was used for percentage data(Sokal&Rohlf,1995).To visualize the similarity between the fatty acid profiles of each species from different dates,n-MDS was performed using a Bray-Curtis similarity matrix.Furthermore,PERMANOVA was used to compare the temporal variability( five sampling dates)of fatty acid composition between the two copepods species(Acartiasp.andE.herdmani).

Since the fatty acid composition ofAcartiasp.andE.herdmaniwere similar and showed no significant variation among sampling dates(species:pseudo-F1,45=1.78,p(perm)=0.13;date:pseudo-F4,45=1.26,p(perm)=0.24;species×date:pseudo-F3,45=1.70,p(perm)=0.08),we decided to use onlyE.herdmanifor feeding experiments(Fig.1,Annex 1).

We used a PERMANOVA analysis to compare fatty acid composition between nauplii and rotifers used as prey during the two larval feeding experiments.Another PERMANOVA was carried out to compare fatty acid profiles of the neutral lipid fraction of larvae on day 0 and after six days of feeding on nauplii or rotifers(three levels)in each experiment(two levels).The SIMPER procedure was performed to identify FA explaining the most important dissimilarity between treatments.These PERMANOVA tests were done after verification of the assumptions of homoscedasticity with PERMDISP.Furthermore,to find differences between treatments,ANOVAs were carried out using SYSTAT software(version 12.02).

Length and weight growth(difference between day 6 and day 0 of the experiment),survival rates at the end of feeding trials,and specific growth rates(rotifers and nauplii)were compared with Student t-tests.Two-way ANOVAs were carried out to compare prey from each experiment(2 prey×2 experiments)with each essential fatty acid(AA,EPA,DHA in%of total fatty acids)and total EFA concentration(g mg-1)of prey used for feeding trials as the dependent variables.Two-way ANOVAs were performed on each EFA of larval neutral lipids to compare the impact of feeding trials(larvae on day 0 and after six days of feeding on nauplii or rotifers)for each experiment(three feeding levels×two experiments).A ratio between polar FA of larvae on day 6 and prey FA was also calculated for each EFA to explore potential nutritional needs in the diet;these were compared with two-way ANOVAs with experiment(two trials)and diet(nauplii and rotifers)as factors.Similar two-way ANOVAs were also used for behavior analyses of larval time spent(%)resting and swimming and the number of hunting events per minute.Homogeneity of variance and normality were verified using Levene and Kolmogorov-Smirnov tests,respectively.When differences were detected,a posteriori comparison using a Tukey pairwise test was used to determine which means were significantly different.If necessary,data were log+1 or square-root(%data)transformed to achieve homogeneity of variances.ANOVAs were carried out using SYSTAT software(version 12.02).

Fig.1.Non-metric multidimensional scaling(n-MDS)plot based on a Bray-Curtis dissimilarity matrix calculated on square-root-transformed fatty acid composition data for Acartia sp.and Eurytemora herdmani.

3.Results

3.1.Characterization of natural potential prey

The St.Lawrence estuary holds different zooplankton species,including bivalve larvae,polychaetes,and gastropods.In our summer 2015 samples,more than 80%of the zooplankton species composition consisted of copepods,particularly calanoid copepods such asAcartiasp.andEurytemoraspp.(mainlyE.herdmaniand someE.affinis).Abundances showed important spatio-temporal variations,with maximal values obtained in July for both taxa(Fig.2).

On the other hand,there were significant differences in the fatty acid composition of live prey(rotifers andE.herdmaninauplii)used during feeding trials(pseudo-F1,23=5.03,p(perm)＜0.001),withE.herdmaninauplii being richer in EPA and DHA compared to rotifers(Fig.3).In general,both experiments showed similar results for prey fatty acid composition (pseudo-F1,23=2.18,p(perm)=0.061).SIMPER analysis on each factor indicated that the three essential fatty acids(DHA,EPA,and AA)explained more than 25%of variability in the data.

Fig.2.Abundances of Acartia sp.and Eurytemora spp.at the three sampling sites in the St.Lawrence estuary and their relationship with temperature and salinity at the surface.

Fig.3.Mean proportions of essential fatty acids(EFA)of prey(rotifers Brachionus plicatilis and Eurytemora herdmani nauplii)used for each feeding trial.Error bars indicate standard errors.Different letters above bars indicate significant differences(p＜0.005).

For AA,we observed a significant difference between prey and experiments(F1,20=7.28,p=0.014),with higher values forE.herdmaninauplii from experiment 2 compared to rotifers from experiment 1(Fig.3).A significantly higher EPA value was observed inE.herdmaninauplii compared to rotifers(F1,20=25.75,p＜0.001),with no differences between experiments(F1,20=1.01,p=0.327)or interaction between factors,i.e.,prey×experiments(F1,20=2.732,p=0.114)(Fig.3).The DHA level inE.herdmaninauplii was more than twice as high as that in rotifers(F1,20=53.8,p＜0.001;Fig.3),with no differences between experiments(F1,20=2.025,p=0.17)or interactions(F1,20=0.076,p=0.786).Total fatty acid concentration(TFA,μg mg-1)was similar between experiments(F1,20=0.334,p=0.57)and prey(F1,20=0.25,p=0.623),and nointeraction was observed (F1,20=3.347,p=0.072).The mean value was 202.8±53.5「g mg-1dry mass of total fatty acids in prey.

3.2.Larval feeding experiments

After six days of feeding trials,growth was similar between winter flounder larvae fed withE.herdmaninauplii and those fed with rotifers in both experiments(Table 1).Specific growth rate(SGR)did not show significant differences in winter flounder larvae fed withE.herdmaninauplii(0.24±0.06%day-1)or with rotifers(0.20±0.04%day-1;t=1.317,p=0.209).Larval survival was also similar between diet treatments,with values of 36.94±5.69%and 35.35±16.78%for theE.herdmaninauplii and rotifer treatments,respectively,and no differences were found between experiments(t=1.0,p=0.931).

To evaluate the accumulation of lipid reserves in winter flounder larvae,the fatty acid composition from the neutral fractions was compared before(day 0)and after the feeding trials(day 6)using rotifers orE.herdmaninauplii.The fatty acid composition of neutral lipids from larvae showed a significant difference between experiments and diet treatments(pseudo-F2,34=3.016,p(perm)=0.022).SIMPER analysis indicated that the three essential fatty acids(DHA,EPA,and AA)explained 15%-20%of the dissimilarity between treatments.ANOVAs on each EFA from the neutral fraction of larvaeindicated a significant interaction between experiments and diet treatments for AA(F1,30=6.29,p=0.005),being high on day 6 in larvae fed on rotifers in the second experiment.A similar interaction between experiments and diet treatments was observed for EPA(F1,30=6.67,p=0.004)and DHA(F1,30=6.32,p=0.005),with high values on day 6 for larvae fed on rotifers(Fig.4).

Table 1 Daily growth of winter flounder larvae fed for 6 days with rotifers(Brachionus plicatilis)or Eurytemora herdmani nauplii.Data are mean±standard deviation of two different experiments.

The total fatty acid concentration in the neutral lipid fraction also showed a significant difference between experiments and diet treatments(F1,29=6.98,p=0.003)corresponding to higher values in larvae fed withE.herdmaninauplii for six days;however,this was only true for the first experiment(Fig.5).

To obtain an indicator of the prey's nutritional quality,we used the ratio of the polar fraction of fattyacids present in larvae aftersix days of feeding and the total fatty acids present in prey.This ratio indicates the selective incorporation into the cell membrane(polar lipids)of larvae for physiological needs of a given dietary essential fatty acid(from nauplii or rotifers).The same ratios were calculated for AA,EPA,and DHA.The observed ratio for AA was slightly above 1,and only larvae in the second experiment fed with rotifers showed a significantly higher value of 3.01±0.24(interaction term experiment×prey;F1,16=10.08,p=0.006).The ratio of EPA was significantly lower(＜1)in experiment 1 compared to experiment 2(F1,16=22.36,p＜0.001).The EPA ratio was lower in larvae fed with nauplii(＜1)than in larvae fed with rotifers(F1,16=25.99,p＜0.001)and no significant interaction(F1,16=0.41,p=0.531)was found.The DHA ratio was mostly＞1 for larvae fed with rotifers,particularly during the second experiment.Significantly lower ratios(closer to 1)were observed for larvae fed with nauplii(interaction termF1,16=7.03,p＜0.017;Fig.6).

3.3.Larval behavior experiments

The use of different prey for winter flounder larval rearing modified their behavior in a significant way.Larvae fed with nauplii were 32%more active(spent more time swimming)than larvae fed with rotifers(F3,136=37.013,p＜0.001;Fig.7).During the swimming period,we observed a trend toward higher hunting rates in larvae fed with nauplii(1.4±0.8 events·min-1)than with rotifers(1.1±0.7 events·min-1),although this difference was not significant(F1,68=4.996,p=0.29).

Fig.4.Percentage of essential fatty acids in the neutral lipid fraction of larvae at day 0 and day 6 fed with rotifers(Brachionus plicatilis)or Eurytemora herdmani nauplii.Error bars indicate standard errors.Different letters above bars indicate significant differences(p＜0.005).

Fig.5.Concentration(g mg-1)of neutral fatty acids of larvae at day 0 and day 6 fed with rotifers(Brachionus plicatilis)or Eurytemora herdmani nauplii.Error bars indicate standard deviation.Different letters above bars indicate significant differences(p＜0.005).

4.Discussion

The results confirm the hypothesis thatE.herdmaninauplii used as live prey have a more suitable fatty acid composition for winter flounder larvae compared to that of rotifers,which are generally used in hatcheries.In addition,we found that the nature of the prey affects larval feeding behavior.These results suggest that larval feeding could be improved,particularly since copepods seemed to be a more appropriate food to meet the DHA needs of winter flounder larvae.

Marine larvae can be successfully reared on a combination of live feed such asArtemiaand rotifers(Randazzo et al.,2018),including winter flounder(Audet&Tremblay,2011).However,it is generally considered that copepods better satisfy the nutritional requirements of fish larvae(Evjemo,Reitan,&Olsen,2003).Most fish and crustacean species largely depend on copepod species throughout their early life stages(Llopiz,2013;Robert et al.,2014)and some species feed exclusively on copepods during their entire life cycle(Perumal,Rajkumar,&Santhanam,2009). Thus,it is likely that the natural prey for winter flounder larvae of the St.Lawrence estuary are calanoid copepods(Perumal et al.,2009).In the middle estuary,our results indicate thatEurytemoraspp.is more abundant early in the summer whileAcartiasp.is more important later in the summer.The two genera are not present at the same time in the same quantities andEurytemorawould be more available as potential prey when winter flounder are in the larval stage.

Fig.6.Ratio between proportions of polar fatty acids in larvae and those of total fatty acids in prey for larvae fed with rotifers(Brachionus plicatilis)or Eurytemora herdmani nauplii.Error bars indicate standard errors.Different letters above bars indicate significant differences(p＜0.005).

Fig.7.Percentage of time spent swimming and resting by winter flounder larvae during the 10 min following feeding.Error bars indicate standard errors.Different letters above bars indicate significant differences(p＜0.005).

The link between fattyacid profiles and the seasonal variation of zooplankton species in the field is still poorly known.Although some studies in other environments have demonstrated differences in the fatty acid compositions among species of copepods(Drillet,Jørgensen,Sørensen, Ramløv,&Hansen, 2006;Evjemo&Olsen,1997;Evjermo et al.,2003),no differences in the fatty acid compositions ofAcartiasp.andEurytemoraspp.were observed during the summer(June to August)sampling period of our study.This result suggests that,regardless of the copepod species,prey with similar fatty acid composition are always available in the St.Lawrence estuary during summer.This absence of differentiation is probably related to the small-scale sampling done only during summer,the period of winter flounder larval development.

Contrary to freshwater fish(Vargas et al.,2018),marine fish species have limited activities of elongases and desaturases involved in the biosynthesis of essential fatty acids,so they require DHA,EPA,and AA from food for growth and survival(Glencross,2009;Sargent et al.,1999).Furthermore,lipid reserves in fish larvae are very limited and their survival strongly depends on the continuous availability of high-quality food(Izquierdo&Koven,2011).

Most marine fish hatcheries have relied on readily available cultures of rotifers(Brachionus plicatilis)as first-feeding organisms.However,rotifers are naturally poor in nutrients,especially EFA(Bell,McEvoy,Estevez,Shields,&Sargent,2003).Thus,they need to be properly enriched before their use as live food for larvae(Avella et al.,2007).According to several authors,rotifers contain a high proportion of essential amino acids,but they are characterized by lowlevels of important fatty acids like DHA(Drillet et al.,2006).We determined the fatty acid profiles of enriched rotifers and copepod nauplii used during feeding experiments and observed that the nauplii ofE.herdmanihad higher values of EPA and DHA compared torotifers,as already observed in the literature(Barroso et al.,2013;Bell et al.,2003;Drillet et al.,2006;Evjemo&Olsen,1997).In general,copepods have a nutritional composition matching the nutritional requirements of marine fin fish larvae(Mæhre,Hamre,&Elvevoll,2013;van der Meeren et al.,2008),particularly of fatty acids(Watanabe,1978,1993).Several coastal copepods have a high content of total fatty acids,normally near 60%(Evjemo&Olsen,1997).In general,the predominant fatty acids of copepods are DHA,EPA,and the saturated fatty acid 16:0(Evjemo et al.,2003).Four fatty acids(16:0 palmitic,16:1 palmitoleic,18:0 stearic and 18:1oleic)are the most conspicuous in zooplankton in which they account for 40-80%of the total fatty acids(Lavaniegos&López-Cortés,1997),which is consistent with our lipid results.Hagemann,Oie,Attramadal, Veiseth,and Ovjemo(2013)and Cabrol et al.(2015)found similar results in the lipid composition ofEurytemorasp.andAcartiasp.,which are characterized by high levels of EPA,DHA,AA,18:0,18:1n9,16:1n7,16:0.According to Evjermo et al.(2003),the total lipid contents ofEurytemorasp.varies from 7 to 14%of dry weight(DW),that of DHA from 26 to 42%,EPA from 15 to 24%,and 16:0 from 8 to 12%,and the sum of n-3 HUFA is 55-62%,which is also consistent with our results.In adultAcartia tonsa,Hagemann et al.(2013)reported 18%of DW,and the sum of DHA and EPA was close to 25%of total FA.In newly hatched nauplii,the lipid content was 14%of DW while the DHA and EPA content was 16%and 11%of total FA,respectively.Similar values were found by Perumal et al.(2009)forAcartiasp.(12-17%of total FA).These results are also similar to our own.

The relative proportions of each essential fatty acid in polar lipids(fatty acids incorporated into cell membranes)in larvae vs.prey(E.herdmaninauplli and rotifers)were used to evaluate the nutritional quality of diets.This ratio indicates whether the DHA,EPA,and/or AA are selectively incorporated by larvae(Copeman,Parrish,Brown,&Harel,2002).If the relative proportion of the fatty acid in larvae/diet is equal to or below one,then the fatty acid requirement of a larva is presumably satisfied.In contrast,if the relative proportion of the fatty acid in larvae/diet is higher than 1,then a fatty acid is selectively incorporated by larvae,suggesting a potential dietary deficiency(Gendron et al.,2013;Pernet&Tremblay,2003).The EPA and AA ratios for both diets were close to one,suggesting that the EPA and AA contents of enriched rotifers and copepod nauplii seemed to be sufficient for winter flounder larval development.Thus,we assume that a similar fatty acid level in winter flounder cell membranes and in the prey(relative proportion≤1)confirms that the nutritional requirement for these fatty acids is satisfied.

The results for DHA were very different,showing a strong selective incorporation when winter flounder larvae were fed with enriched rotifers.This selective retention of DHA has been also reported by Vagner et al.(2014)in winter flounder larvae fed with enriched rotifers.Their results suggest that the availabilities of EFA from the different algae used to enrich rotifers were below the physiological needs of early settled and post-settled larvae.Thus,to optimize winter flounder larval rearing,a new rotifer enrichment method needs to be used to increase DHA.

In the same way,winter flounder larvae fed with rotifers containing higher AA levels and higher DHA:EPA ratios showed better growth and the lowest bacterial colonization of the intestinal lumen compared to larvae fed with rotifers containing lower AA levels and lower DHA:EPA ratios(Seychelles,Audet,Tremblay,Lemarchand,&Pernet,2011).AA,EPA,and DHA available from the algal cocktail diet used by Vagner et al.(2014)for winter flounder larvae seemed to be sufficient for larval and post-settled larval development,since the ratios between organismal FA content and dietary FA content were always below one for that treatment.However,a strong selective retention for DHA was found.In addition,these authors did not find a significant effect of diet or interaction between developmental stage and diet on total length or maximum width.

Thus,when DHA is deficient in food,its concentration decreases in the membranes,and this affects larval development(Watanabe,1978).However,Piccinetti et al.(2015)reported a good recovery of Dover sole(Solea solea)after starvation when favorable feeding conditions were restored.It has been demonstrated that the addition of DHA in diet of marine fish larvae improved growth and feeding efficiency,thereby reducing mortality(Watanabe,1993)more effectively than the addition of EPA or AA(Koven,Kolkovski,Tandler,Kissil,&Sklan, 1993;Wu,Ting,&Chen,2002).In fact,DHA is preferentially conserved during food deprivation(Koven,Kissil,&Tandler,1989)and is selectively assimilated into tissues and membranes,suggesting a metabolic priority for this fatty acid(Izquierdo,1996;Rainuzzo,Reitan,&Olsen,1997).We suggest that the DHA deficit in the rotifer diet might explain the trend of differential growth in the first experiment in winter flounder larvae fed withEurytemoraspp.nauplii,since DHA is one of the key biochemical components that determines the success of the development and the growth rate of larvae and young fish(Sargent et al.,1999).

In our experiment,rotifers were enriched with a cocktail of three microalgal species,Nannochloropsis oculata,Isochrysis galbana,andPavlova lutheri.These species were selected for their fatty acid profiles and their facility to enrich rotifers(Seychelles et al.,2009).For example,Seychelles et al.(2009)demonstrated that the high DHA content ofI.galbanaenriched DHA levels by more than 6%in rotifers.However,in our experiment,rotifers showed DHAvalues below 2%.We used a similar rotifer rearing method,but we fed rotifers with a mix of microalgal species,while Seychelles et al.(2009)fed rotifers only one microalgal species.Thus,it seems that theI.galbanaconcentration in the rotifer tanks of that experiment was too low or that rotifers preferentially ingestedC.gracilisoverI.galbana.Clearly,our enrichment method was not appropriate for producing rotifers meeting the DHA needs of winter flounder larvae.

The DHA deficiency in rotifer diets was not reflected by lower survival rates of winter flounderlarvae compared to those fed DHA-rich nauplii diets.The six-day periods of the feeding trials were sufficient to affect their feeding behavior(increasing swimming time in nauplii diet)and modify the lipid composition of cell membranes in larvae(in the first experiment),but probably too short to be reflected in growth or selective mortality.After six days of feeding,DHA deficiency in food was reflected in the polar lipids of winter flounder larvae.At the same time,only weak accumulations(between 0.5%and 6%)of DHA,EPA,and AA in the neutral lipid fraction were observed in the winter flounder larvae,independent of the diet.Combined with the high metabolism of larvae,it is not surprising to observe a DHA deficiency in the polar lipids of larvae after six days.

Fish larval behavior,mainly the costly activities of swimming and hunting(Hunt von Herbing,Gallager,&Halteman,2001),are influenced by the food's lipid composition.Bransden,Cobcroft,and Battaglene(2005)found that striped trumpeter fish larvae,Latis lineata,fed with low dietary DHA showed a modified swimming pattern.We observed similar results,with rotifer-fed larvae showing evidence of DHA deficits being 32%less active than larvae fed withE.herdmaninauplii.Although changes in swimming activity levels were significant,there was no significant change in hunting rates between diets.Nevertheless,a non-significant tendency was observed with rotifer-fed larvae,which had hunting rates that were 22%lower.It is possible that a longer feeding trial period leading to a higher DHA deficit could have led to a statistically significant trend for the hunting rate.

Larval feeding behavior results from the interaction between complex processes that include several pre-and post-consumption steps-search,detection,attack,capture,ingestion,digestion,and evacuation-and each of these steps has a specific pattern that changes throughout development(Yufera,2011).In the same way,the size and swimming behavior of zooplankton prey appear to influence prey selection in cultured fish larvae.Zooplankton,such asArtemianauplii,rotifers(B.rotundiformis),and copepods(A.tonsa),exhibit a wide range of swimming velocities and behavioral patterns that affect the ability of fish larvae to capture them(Turingan,Beck,Krebs,&Licamele,2005).In fact,copepod nauplii trigger the appropriate hunting behavior in fish larvae due totheir zig-zag swimming patterns,which the larvae find attractive(Hansen,2017).

Regarding attack behavior,Shaheen et al.(2001)demonstrated that significantly more attacks were made onE.affinisthanAcartia hudsonica.The consumption of femaleE.affiniscould be advantageous because they are more energetically valuable and egg sacs increase visibility and limit the escape(Sandström,1980).Furthermore,prey swimming behavior has also been considered sinceE.affinismovements are much quicker and more erratic thanA.hudsonica,which was slower and more rhythmic(Shaheen et al.,2001).This slowand rhythmic behavior seems similar toswimming pattern of rotifers used in our studies.Lubzens,Tandler,and Minkoff(1989)described rotifers as slow swimmers that maintain their position in the water column.

The selection of food by fish larvae not only depends on prey accessibility,but also on organoleptic preferences(Yufera,2011).Some metabolites from planktonic organisms(amino acids and nucleotides)have been identified as olfactory and taste stimuli for feeding.The ability to discriminate chemical cues by olfaction and gustation is present at first feeding and improves during larval development(Kolkovski,Koven,&Tandler,1997).Understanding predator-prey interactions is key to successful feeding and nutrition in larviculture,because the feeding mechanism of the developing larvae influences their performance in a number of ways(Turingan et al.,2005).

In general,our results demonstrate that copepods have a better fatty acid composition,particularly the high DHA content,than enriched rotifers.However,for commercial aquaculture purposes that seem to be the future of live prey in larval rearing,there are still some limitations to achieve successful copepod cultures(even at an experimental scale).Recently efforts in copepods culture were made in different species of calanoids and cyclopoids(Hansen,Hansen,Nielsen,&Jepsen,2017;Pan et al.,2017;Rayner, Hwang,&Hansen,2017b,2017a;Selander,Heuschele,&Larsson,2017).Considering that our results do not show differences in fatty acids content betweenAcartiaspp.andEurytemoraspp.there are more chances to move towards the culture ofAcartiaspp.,which are the better studied species.

However,if a culture of copepods is not yet available or achievable in a hatchery,another cost-effective solution is to continue the development of better rotifer enrichment.In the case of winter flounder,the rotifer enrichment protocol,which proved to be effective in AA and EPA,could be improved by the addition of DHA.However,even when rotifers are enriched,they are not a perfect replacement for copepods;for example,their swimming behavior could affect the feeding behavior of larval fish.

5.Conclusion

Based on our results,there was no difference in fatty acid composition among sampling dates or stations between the two species of copepod nauplii examined in our study(Acartiasp.andEurytemoraspp.).Thus,regardless of the copepod species,prey with similar fatty acid compositions were always available to fish larvae throughout their larval development period(summer in the St.Lawrence estuary).Larval rearing performance of winter flounder in terms of fatty acid content,in particular DHA,was higher in larvae fed withE.herdmaninauplii than with enriched rotifers,which are traditionally used in aquaculture.Finally,E.herdmaniimpacted the feeding behavior of winter larvae by stimulating swimming activity.Considering these results,it could be thinking about developing a copepod culture,which seems to be the future of live prey to fish larval rearing or improve the way to enrich the rotifers by the addition of more DHA.

Declaration of interest statement

All individuals listed qualify as authors and have approved the submitted version.Furthermore,we declare that this work is original and is not under consideration by any other journal and we have permission to reproduce any previously published material.Finally,all confirmation and conflict of interest questions were answered.

Acknowledgements

The authors wish to thank Renée Gagné and Nathalie Gauthier for their technical support at the aquaculture station and Jean-Bruno Nadalini for lipid extraction.This work was supported by the FQRNT(Fonds Québecois pour la Recherche,Nature et Technologies;project 2016-PR-190063)and by Ressources Aquatiques Québec(RAQ).