Parental nutritional programming and a reminder during juvenile stage affect growth, lipid metabolism and utilisation in later developmental stages of a marine teleost, the gilthead sea bream (Sparus aurata)

Parental nutritional programming and a reminder during juvenile stage affect growth, lipid metabolism and utilisation in later developmental stages of a marine teleost, the gilthead sea bream ( Sparus aurata ) Serhat Turkmen*, Maria J. Zamorano, Hipólito Fernández-Palacios, Carmen M. Hernández-Cruz, Daniel Montero, Lidia Robaina and Marisol Izquierdo Aquaculture Research Group (GIA), Research Institute in Sustainable Aquaculture and Marine Conservation (IU-ECOAQUA), Universidad de Las Palmas de Gran Canaria, Crta. Taliarte s/n, 35214 Telde, Spain (Submitted 1 May 2017 –Final revision received 9 August 2017 –Accepted 11 August 2017) Abstract Nutrition during periconception and early development can modulate metabolic routes to prepare the offspring for adverse conditions through a process known as nutritional programming. In gilthead sea bream, replacement of fish oil (FO) with linseed oil (LO) in broodstock diets improves growth in the 4-month-old offspring challenged with low-FO and low-fishmeal (FM) diets for 1 month. The present study further investigated the effects of broodstock feeding on the same offspring when they were 16 months old and were challenged for a second time with the low-FM and low-FO diet for 2 months. The results showed that replacement of parental moderate-FO feeding with LO, combined with juvenile feeding at 4 months old with low-FM and low-FO diets, significantly (P<0·05) improved offspring growth and feed utilisation of low-FM/FO diets even when they were 16 months old: that is, when they were on the verge of their first reproductive season. Liver fatty acid composition was significantly affected by broodstock or reminder diets as well as by their interaction. Moreover, the reduction of long-chain PUFA and increase in α-linolenic acid and linoleic acid in broodstock diets lead to a significant down-regulation of hepatic lipoprotein lipase (P<0·001) and elongation of very long-chain fatty acids protein 6 (P<0·01). Besides, fatty acid desaturase 2 values were positively correlated to hepatic levels of 18 : 4n-3, 18 : 3n-6, 20 : 5n-3, 22 : 6n-3 and 22 : 5n-6. Thus, this study demonstrated the long-term nutritional programming of gilthead sea bream through broodstock feeding, the effect of feeding a ‘reminder’diet during juvenile stages to improve utilisation of low-FM/FO diets and fish growth as well as the regulation of gene expression along the fish’s life-cycle. Key words: Nutritional adaptation of offspring: Long-term effects of parental nutrition: Hepatic gene expression: Fatty acid desaturase: Epigenetics in aquaculture Aquaculture is the fastest-growing animal production sector accounting, at present, for more than 50 % of worldwide fish consumption (1) , but one major issue concerning such development is the over-dependence of fish feeds on capture fisheryderived raw materials such as fishmeal (FM) and fish oil (FO) (2) . The established beneficial roles of FO on human health (3) and the use of FO, albeit in small proportions, in other animal production systems, have led to an increase in the demand for this raw material, consequently, raising the prices. Despite great achievements in the reduction of FM in diets of marine fish species (4–8) , complete replacement of FO still remains a major challenge. Moreover, complete substitution of FO negatively affects the immune system and stress and disease resistance (9) as well as reduces the fillet content of long-chain n-3 PUFA (n-3 LC-PUFA, includes twenty or more carbon atoms), such as EPA and DHA, negatively affecting the nutritional value of fish flesh for humans (10–13) . FO is rich in n-3 LC-PUFA, whereas vegetable oil (VO) sources, except in some cases (14) , lack the essential fatty acids (EFA) for marine fish such as EPA and DHA but can have significant amounts of α-linolenic acid (ALA), 18 : 3n-3, and linoleic acid (LA), 18 : 2n-6, which are biological precursors of EFA. Yet, the bioconversion of 18-carbon PUFA to EPA and DHA depends on the elongation and desaturation capacity of the fish species (15,16) . Generally speaking, whereas freshwater fish possess the ability to convert ALA and LA into LC-PUFA (17,18) , marine fish do not possess the sufficient enzyme activity (19) . Nonetheless, LC-PUFA synthesis capacity also appears to differ among marine species (20–23) . The higher LC-PUFA biosynthesis capacity in freshwater fish in comparison with that of marine fish could be Abbreviations: ALA, α-linolenic acid; f, fish oil-based diet; elovl6, elongation of very long-chain fatty acids protein 6; fads2, fatty acid desaturase 2; FM, fishmeal; FO, fish oil; HL, 20 % fish oil–80 % linseed oil; LA, linoleic acid; LL, 40 % fish oil–60 % linseed oil; LO, linseed oil; lpl, lipoprotein lipase; v, reminder diet; VO, vegetable oil. *Corresponding author: S. Turkmen, email [email protected] British Journal of Nutrition (2017), 118, 500–512 doi:10.1017/S0007114517002434 © The Authors 2017 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at related to differences in the feeding habits and nutrient intake, with marine fish having a continuous access to LC-PUFA-rich sources throughout their lives (24,25) . Besides, these differences among fish species have been related to the diverse evolution of certain genes involved in lipid biosynthesis (16) . Recent evidence suggests that environmental factors experienced by the parents can have long-lasting effects in the offspring or in the later generations (26,27) . Thus, early environmental signs, such as available nutrients during reproduction, can modulate metabolic routes and offspring phenotype (27–29) . This type of metabolic regulation, known as ‘nutritional programming’, has been principally derived from mammalian models, because of their potential effects on development of metabolic disorders in humans in later life (30) . Therefore, better understanding of the outcomes of parental nutrition and the underlying mechanisms can contribute to the prevention of consequences in the offspring. Besides, nutritional programming may also have potential applications in animal production (31) . In aquaculture, one of the potential beneficial applications of nutritional programming may be the production of individuals better prepared to use some feedstuffs supplying or lacking in specific nutrients, such as VO and plant-protein sources. For instance, specific fat and dietary fatty acid supply during embryonic and offspring development may adjust fish metabolism for better utilisation of 18-carbon fatty acids later in life. Thus, parental nutritional interventions can modulate epigenetic mechanisms that control ‘metabolic decisions’that are meiostatically and mitotically stable through life in humans (32) , rodents (33,34) and cattle (35) .Infish, nutritional programming studies have mostly focused on early feeding (36–43) , whereas parental nutritional interventions are scarcer (44–46) . The gilthead sea bream (Sparus aurata), is a multi-batch spawner whose oligolecitic eggs largely depend on their continuous intake of nutrients during reproduction (47) . For this reason, egg nutrient content of the gilthead sea bream can be markedly affected by the parental diet even during the spawning season, in turn, affecting early embryonic development (48) . Our previous studies have demonstrated that feeding gilthead sea bream broodstock with high-linseed oil (LO) diets markedly affects fecundity, spawn quality and growth of 45-dafter-hatch larvae and 4-month-old juveniles (45) . Interestingly, when 4-month-old juveniles were challenged with a low-FM and low-FO diet, offspring from parents fed a replacement of 60 %-FO with LO showed a faster growth and better feed utilisation than those whose parents had been fed with FO (45) . However, the potential persistence of the effects of broodstock nutritional history on the offspring later in life is still unknown. Further, little is known on the physiological or molecular mechanisms involved in the effect of parental diets on the metabolic performance of the offspring. When n-3 LC-PUFA are limited and the 18-C atom fatty acids are available in gilthead sea bream diet, the gene for fatty acid desaturase 2 (fads2), the key-limiting enzyme for LC-PUFA synthesis, is up-regulated (49) . Long-chain fatty acid synthesis also involves chain-elongation catalysed by elongases (Elovl) with different substrate preferences (50) . Among them, Elovl6 is a key lipogenic enzyme that elongates long-chain SFA and MUFA of 12, 14 and 16 carbon atoms, which has received much attention because of its importance in metabolic disorders (51) . Besides, LC-PUFA may have a direct effect on the expression of other genes related to lipid or carbohydrate metabolism (52) . Lipoprotein lipase (lpl) facilitates the tissue uptake of circulating fatty acids (53) from lipoproteins and its expression in the liver can be regulated by n-3 LC-PUFA (54) . The provision of energy is accomplished by β-oxidation of free fatty acids transported into the mitochondria in the form of fatty acyl-carnitine esters by carnitine acyltransferases, such as carnitine palmitoyltransferases (15) . Replacement of FO with VO changes the fatty acid composition of liver and muscle, affecting the β-oxidation capacity and regulating the expression of cptI and cptII genes (17,55–57) . β-Oxidation also takes place in the peroxisome and is modulated by PPAR. A total of three different PPAR isoforms (α,β,γ) have been characterised in gilthead sea bream, pparαbeing the major form expressed in the liver (58) . PPAR are nuclear receptors that regulate differentiation, growth and metabolism and, in mammals, epigenetic mechanisms have been described to regulate these processes involving all the PPAR isoforms (59) . For instance, feeding pregnant rats a protein-restricted diet reduces methylation of the pparαpromoter in the offspring and the hypomethylation persists into adulthood (60) . Finally, another gene potentially regulated by LC-PUFA is cyclo-oxygenase-2 (cox2), a key enzyme in prostanoid biosynthesis (61) . The objective of the present study was to explore the potential persistence of nutritional programming through parental feeding in offspring later in life and to analyse the physiological or molecular mechanisms implied. For this purpose, the offspring of gilthead sea bream broodstock fed diets with different FO/LO levels were followed up for 18 months until the beginning of first gonad development and nutritionally challenged at 4 and 16 months with low FM and FO diets. The effects of both broodstock feeding and nutritional challenge on growth, chemical and fatty acid composition of muscle and liver as well as on the expression of selected genes in the liver were investigated. Methods Experimental animals All fish were obtained from spontaneous spawns of gilthead sea bream broodstock fed three diets with three levels of FO substitution with LO: 100 % FO, 40 % FO–60 % LO (LL) and 20 % FO–80 % LO (HL) (45) . Offspring from all groups were fed the same commercial diet during larval rearing, weaning and during the growing period until they reached 4 months of age (120 d) (45) . At this stage, triplicate groups of juveniles were nutritionally challenged for 1 month with either a high-FM/FO diet (f) or with a high-VM/LO diet (v) named as ‘reminder diet’ in the present study. Details of the broodstock feeding, juvenile nutritional challenge (reminder) at 4 months and feed formulation have been reported earlier (45) .Afterthisfirst nutritional challenge (reminder), fish were maintained separately in 1000 litre tanks and fed the same commercial diet until they were 16 months old for use in the present study. A schematic view of this nutritional programming history is shown in Fig. 1. Transgenerational effects of parental diets 501 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at In the present study, 16-month-old offspring sea breams of homogeneous weight were selected and distributed into 18 500litre light-grey fibreglass cylinder tanks (2·8 kg/m 3 ). Each tank contained thirty fish with mean initial body weight of 243·2 (SD12·7) g. Tanks were supplied with filtered seawater (37 parts per million (ppm) salinity), which entered from the tank surface and drained from the bottom at a rate of 250 litre/h to maintain a high water quality, which was tested daily and no deterioration was observed. O 2 level, water temperature and pH were monitored in real-time using Miranda aquaculture water quality monitoring system (Innovaqua). Water was continuously aerated (125 ml/min), attaining an average of 6·8( SD 0·8) ppm dissolved O 2 during the experimental period. The average water temperature and pH for the duration of the trial were 24·6±0·6°C and 7·89, respectively. Natural photoperiod was maintained during the whole experimental period (10h light). Experimental diet The experimental diet was formulated and produced by Biomar to be low in FO (3%) and FM (5 %). Thus, the diet was high in oleic acid (18 : 1n-9), LA (18 : 2n-6) and ALA (18 : 3n-3) (Table 1). Juveniles from each group were fed daily until apparent satiation for 60 d, three times a day at 09.00, 13.00 and 17.00 hours. The feed was supplied in small portions (<5–6 pellets at a time) to ensure that all feed was eaten. After each feeding, uneaten feed was collected, kept in aluminium oven trays, dried overnight at 105°C and weighted to calculate feed intake. Biochemical analyses Moisture, protein (63) and crude lipid (64) contents of the tissue samples and diets were analysed. Fatty acid methyl esters were obtained by trans-methylation of crude lipids as previously described (65) . Fatty acid methyl esters were separated using GLC (GC-14A; Shimadzu) following the conditions described previously (66) and identified by comparison with previously characterised standards and GLC-MS (Polaris QTRACE TM Ultra; Thermo Fisher Scientific). Molecular studies Liver samples from three fish per each tank (nine per group) were collected at the beginning (480-d-old fish) and at the end of the feeding challenge (540-d-old fish). Samples were collected on ice from fish kept unfed for 24 h, each tissue sample from one individual was assigned to a corresponding 1·5-ml Eppendorf tube and was snap frozen in liquid N 2 immediately after sampling. The samples were then stored at – 80°C until RNA extraction and analyses. RNA was extracted (Reminder diet) First nutritional challenge* Feeding Nutritional programming Hatching Rotifers Rotifers + Artemia Artemia Commercial enrichment Commercial enrichment Commercial enrichment Commercial Pellets OR Pellets Commercial Second nutritional challenge† 3 % Fish oil 5 % Fishmeal (v) 100 % Fish oil 100 % Fishmeal (f) 3 % Fish oil 5 % Fishmeal (V) Time Spawning –3 months Day 0 Day 15 Day 30 Day 45 Day 90 4 g Day 120 9 g Day 480 243 g *Izquierdo et al. (2015) Day 540 350 g Broodstock groups F F F LL LL LL HL HL HL Ff Fv Ff Fv Ff Fv LLf LLv LLf LLv LLf LLv HLf HLv HLf HLv HLf HLv Ff Fv Ff Fv Ff Fv LLf LLv LLf LLv LLf LLv HLf HLv HLf HLv HLf HLv 20 % Fish oil 80 % Linseed oil HL HL HL 40 % Fish oil 60 % Linseed oil LL LL LL 100 % Fish oil diet F F F Present study Final sampling 36 36 36 Fig. 1. Schematic view of the nutritional programming history of gilthead sea bream. F, 100 % fish oil; LL, 40 % fish oil–60 % linseed oil; and HL, 20 % fish oil–80 % linseed oil; f, fish oil-based diet; v, reminder diet. * Previous publication, Izquierdo et al. (2015) (45) .†Time of the present study. 502 S. Turkmen et al. https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at using Qiagen RNeasy Mini Kit (Qiagen). Before real-time PCR analysis two different potential housekeeping genes, β-actin (β-act) and ribosomal protein L27 (rpl27), were tested. Data from duplicate samples (n18) using the two candidate housekeeping genes were compared using an online program (http://leonxie.esy.es/RefFinder/?type=reference) (RefFinder) (67) and β-act was selected as the most suitable housekeeping gene for the present study (β-act, threshold cycle (C t )values:min= 18·9, max =20·6, mean =19·64, SD =0·5; rpl27:min=17·0max= 21·2, mean =19·1, SD =1·01). Real-time quantitative PCR were performed in an iQ5 Multicolor Real-Time PCR detection system (Bio-Rad) using β-act as the housekeeping gene in a final volume of 15 µl/reaction well and with 100 ng of total RNA reverse-transcribed to complementary DNA (cDNA). Samples, housekeeping gene, cDNA template and reaction blanks were analysed in duplicates (Table 2). Primer efficiency was tested with serial dilutions of a cDNA pool (1:5, 1:10, 1:100 and 1:1000). A single ninetysix-well PCR plate was used to analyse each gene, primer efficiency and blank samples (Multiplate; Bio-Rad). Melting-curve analysis was performed and amplification of a single product was confirmed after each run. Fold expression of each gene was determined by delta–delta C T method (2ΔΔCT) (66) .PCR efficiencies were similar and no efficiency correction was required (71,72) (Table 2). Fold expression was related to that of offspring obtained from FO diet-fed broodstock and fed commercial diets throughout their life (Ff group). Statistical analysis Data on growth and biochemical composition were statistically analysed using two-way ANOVA, using broodstock diet and reminder diet as fixed factors in IBM SPSS version 23.0.0.2 for Mac (IBM SPSS Inc.). Data were split into groups based on each fixed factor (broodstock and reminder diet) and compared with one-way ANOVA. Scheffe’spost hoc multiple comparisons for broodstock and reminder diet, separately, assessed differences between groups. Before the analysis of data, equality of variances was tested using Levene’s test and distribution of data using Shapiro–Wilk tests. All data except cox2 gene expression showed normal distribution and equality of variances. Gene expression data (except for cox2) were analysed by means of two-way ANOVA using broodstock and reminder diet as fixed factors. Expression data were next analysed using Welch’s ANOVA and subsequently compared with the Games– Howell test for identification of differences between groups. The fixed factor for Welch’s ANOVA was experimental groups in gene expression data analysis. Pearson’s correlation test was performed using R (R Foundation for Statistical Computing). Gene expression figures were created using lattice package (version 0.20–33) (73) downloaded from the comprehensive R Archive Network library. The sample size for all analysed data was nine and data were expressed as means and standard deviations. Results Growth performance The low-FM/FO diet was well accepted and there were no significant differences in feed intake among fish groups (mean, 3·95 (SD 0·39) kg, P>0·05). At the beginning of the trial, there were no significant differences in fish body weight (mean, 243·2 (SD 12·7) g) among experimental groups (P>0·05) (Table 3). However, after 60 d of feeding the low-FM/FO diet, LLv fish (obtained from broodstock fed low LO and fed at 4 months with the v, high in VM and VO) showed the highest body weight, being significantly (P<0·05) higher than that of fish LLf, from the same broodstock but fed the diet f at 4 months (high in FM and FO) (Table 3). Besides, the growth of LLv fish was also significantly higher than that of fish from Fv or HLv that had been fed the same v, but came from broodstock fed FO or high LO. Thus, the two-way ANOVA analysis of final body weight showed a significant effect of the broodstock diet (P<0·05) and the interaction between broodstock and reminder diet (P<0·01). The specific growth rate of LLv fish was significantly higher than that of LLf, denoting the significant effect of the v, as well as higher than Fv and HLv (P<0·05). Thus, the two-way Table 1. Main ingredients*, energy, protein and % total fatty acids contents of diet for the nutritional challenge of gilthead sea bream juveniles obtained from broodstock fed diets 100 % fish oil (FO), 40 % FO–60 % linseed oil (LO) and 20 % FO–80 % LO during spawning Main ingredients (%) % Proximate composition (% DM) Fishmeal SA†68 super prime 5·00 Crude lipids 21·7 Fishmeal alternative protein sources‡ 54·50 Crude protein 45·1 Rapeseed meal cake 11·30 Moisture 9·0 Wheat 6·89 Ash 5·4 Fish oil SA†3·00 Vegetable oil mix§ 13·00 Gross energy (MJ/kg, as is) 22·5 % Total fatty acids % Total fatty acids 14 : 0 6·618:3n-3 11·8 14 : 1n-5 0·118:4n-3 0·4 15 : 0 0·118:4n-1 0·0 16 : 0iso 0·020:0 0·4 16 : 0 12·320:1n-9 0·0 16 : 1n-7 2·120:1n-7 1·0 16 : 1n-5 0·120:1n-5 0·1 16 : 2n-4 0·220:2n-9 0·0 17 : 0 0·320:2n-6 0·1 16 : 3n-4 0·120:3n-9 0·0 16 : 3n-3 0·020:3n-6 0·0 16 : 3n-1 0·020:4n-6 0·2 16 : 4n-3 0·420:3n-3 0·0 18 : 0 3·220:4n-3 0·1 18 : 1n-9 32·320:5n-3 2·5 18 : 1n-7 2·322:1n-11 0·1 18 : 1n-5 0·022:1n-9 0·3 18 : 2n-9 0·022:4n-6 0·0 18 : 2n-6 20·322:5n-6 0·1 18 : 2n-4 0·122:5n-3 0·3 18 : 3n-6 0·122:6n-3 1·7 18 : 3n-4 0·0 * Please see Torrecillas et al. (62) for the complete list of feed ingredients. †South American, Superprime (Feed Service). ‡Blood meal spray (Daka), soya protein concentrates 60 % (Svane Shipping), maize gluten 60 (Cargill), wheat gluten (Cargill). § Linseed (2·6 %) (Ch. Daudruy), rapeseed (5·2 %) (Emmelev) and palm oils (5·2%) (Cargill). Transgenerational effects of parental diets 503 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at ANOVA showed the significant effect of the reminder diet at 4 months as well as the interaction between broodstock and reminder diet (P<0·05). Regarding feed conversion, the best values were also obtained for fish in the LLv, Fv or HLv group. For fish coming from broodstock fed FO, the feed conversion ratio (FCR) was better when fish had been fed FO at 4 months (Ff) than when fed diet v (Fv). The two-way ANOVA showed a strong interaction between broodstock and reminder diets (P<0·01). Biochemical composition At the end of the study, protein, lipid and ash contents of liver or muscle were similar (P>0·05) (Table 4). However, liver fatty acid composition was significantly affected by broodstock or reminder diets as well as by their interaction (Table 5). For instance, LO increase in broodstock diet significantly reduced liver contents on 16 : 4n-3, a product of EPA β-oxidation, and increased 16 : 3n-1 or 18 : 0, whereas the interaction of broodstock and reminder diets affected the ratios 18:0:16: 0 and 18:1:16: 1, indicators of elovl6 activity, and the related ratio 16: 1:16: 0 (Table 5). The fads2 products 20 : 3n-6 and 20 : 4n-3 were significantly reduced by the reminder diet (P=0·018) and its interaction with the broodstock diet (P=0·015), respectively, whereas 20 : 4n-6, 20 : 5n-3 and 22 : 6n-3tendedtobehigherinoffspringfedthereminderdiet, but were not significantly different (P=0·17) (Table 5). Muscle fatty acid composition did not differ significantly among the different experimental groups (P>0·05) (Table 6). Gene expression Reduction of LC-PUFA and increase in ALA and LA in broodstock diets lead to a significant (P<0·001) down-regulation of hepatic lpl (Fig. 2), which was significantly (P<0·01) emphasised by feeding the 4-month-old juveniles the v diet, based on Table 2. Primers, RT-PCR reaction efficiencies, and GeneBank accession numbers and reference articles for sequences of target and housekeeping genes Genes Primer sequence 5'-3' (F) and 5'-3' (R) Efficiency (%) GenBank access no. References lpl CGT TGC CAA GTT TGT GAC CTG 98·0 AY495672 (68) AGG GTG TTC TGG TTG TCT GC ppara TCT CTT CAG CCC ACC ATC CC AY590299 (68) ATC CCA GCG TGT CGT CTC C 102·0 elovl6 GTG CTG CTC TAC TCC TGG TA JX975702 (68) ACG GCA TGG ACC AAG TAG T fads2 CGA GAG CCA CAG CAG CAG GGA 109·2 AY055749 (40) CGG CCT GCG CCT GAG CAG TT cox2 GAG TAC TGG AAG CCG AGC AC 107·0 AM296029 (69) GAT ATC ACT GCC GCC TGA GT cpt1b CCA CCA GCC AGA CTC CAC AG 98·0 DQ866821 (70) CAC CAC CAG CAC CCA CAT ATT TAG β-Act TCT GTC TGG ATC GGA GGC TC 100·7* X89920 – AAG CAT TTG CGG TGG ACG lpl, lipoprotein lipase; elovl6, elongation of very long-chain fatty acids protein 6; fads2, fatty acid desaturase 2; cox2, cyclo-oxygenase-2; cpt1, carnitine palmitoyltransferase I; β-act,β-actin. * The average efficiency of housekeeping gene from six RT-PCR runs. Table 3. Growth performance parameters after 2 months’feeding of very low-fishmeal (5 %) and very low-fish oil (FO) (3 %) diet in 16-month-old gilthead sea bream (Sparus aurata) originated from broodstock fed linseed oil (LO) as a replacement for FO –0 % (100 % FO (F)), 60 % (40 % FO–60 % LO (LL)), 80 % (20 % FO–80 % LO (HL)) –and fed either a fishmealand FO-based diet (f) or a very low-fishmeal (5 %) and very low-FO (3 %) ‘reminder’diet (v) for 1 month at 4 months of age (Mean values and standard deviations; n3) Initial body weight (g) Final body weight (g) SGR (%/d)* FCR† Groups Mean SD Mean SD Mean SD Mean SD Ff 243·01·9337·43·30·70·06 1·5 A 0·1 LLf 244·10·9336·3 B 2·60·6 B 0·03 1·7 B 0·1 HLf 242·52·0339·67·20·70·05 1·60·1 Fv 243·60·9319·5 b 5·50·5 b 0·07 1·8 B,b 0·1 LLv 246·40·4351·3 A,a 1·20·7 A,a 0·03 1·4 A,a 0·2 HLv 242·31·3321·3 b 4·00·6 b 0·07 1·9 b 0·2 Two-way ANOVA Broodstock diet (B)ND P<0·05 ND ND Reminder diet (R)ND ND P<0·05 ND B×RND P<0·01 P<0·05 P<0·01 SGR, specific growth rate; FCR, feed conversion ratio; ND, no difference. A,B Mean values with unlike superscript letters were significantly different between fish fed f or v diets during the first nutritional challenge (reminder) coming from the same parental feeding. a,b Mean values with unlike superscript letters were significantly different between fish coming from different parental feeding and fed the same diet during the first nutritional challenge (reminder) (P<0·05). * SGR (%/d) =(Ln (final weight (g)) −Ln (initial weight (g)))/(number of days) × 100. †FCR =(total weight of consumed feed (g))/(weight gain (g)). 504 S. Turkmen et al. https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at plant ingredients and with low LC-PUFA and high ALA and LA contents. Thus, the lowest relative expression of lpl was found in HLv and HLf fish. Similarly, the origin of the fish based on different broodstock diets significantly (P<0·01) downregulated hepatic elovl6 (Fig. 2), with the lowest relative expression of elovl6 found in LLv and HLf fish. Besides, elovl6 expression was significantly correlated to liver contents of 18 : 1:16 : 1 (r0·89) and 18 : 0:16 : 0 (r0·89), ratios of product: substrate of elovl6 activity. There was no significant effect of either the broodstock or the reminder diet on hepatic expression of fads2 (Fig. 2), but their values were positively correlated to hepatic levels of 18 : 4n-3 (r0·86) and 18 : 3n-6 (r0·8), products of the fads2 activity, as well as to the end desaturation products 20 : 5n-3 (r0·98), 22 : 6n-3 (r0·95) and 22 : 5n-6 (r0·95). Besides, fads2 expression values were negatively correlated (r−0·52) to elovl6. Regarding fatty acid catabolism biomarkers, reduction of LC-PUFA and increase in ALA and LA acids in broodstock diets lead to a significant (P<0·001) downregulation of hepatic cpt1b (Fig. 2), which was significantly (P<0·05) emphasised by the reminder diet. Moreover, the relative expression of cpt1b was highly correlated to 18 : 1n-9 (r0·82) and negatively correlated to 20 : 5n-3 (r–0·62). No significant differences were found in the relative expression of ppara or cox2, which were negatively correlated (r–0·57 and r–0·73, respectively) to cpt1b expression. The overall response showed similar trends for lpl,cpt1b and elovl6 expressions, whose values showed a high correlation in their relative gene expression between lpl and elovl6 (r0·52), cpt1b and elovl6 (r0·72) and lpl and cpt1b (r0·74). Discussion In animal production, nutritional programming can be useful to improve offspring adaptation to farm conditions (74,75) . As the limited availability of FM and FO is the main constraint in fish production, modulation of offspring phenotype through parental feeding for an improved utilisation of low-FM and low-FO diets can have important advantages (31) . Previous studies in gilthead sea bream have demonstrated that it is possible to improve low-FM and low-FO feed utilisation in the offspring coming from broodstock fed with increased substitution of FO with LO (45) . This adaptation included the regulation of expression of genes for key metabolic enzymes in the liver such as fads2 (45) or glucocorticoid receptor (gr) (S Turkmen et al., unpublished results). However, the persistence of these phenotypic or metabolic changes later in life had not been studied yet. The present study shows that replacement of parental feeding with moderate-FO with LO combined with juvenile feeding with low-FM and low-FO diets improves offspring growth and feed utilisation of low-FM/FO diets even when they are 16 months old: that is, when they are on the verge of their first reproductive season. Thus, among fish fed the low-FM/FO diet during the juvenile stages, those obtained from parents fed moderate LO levels showed the highest growth, denoting the persistent effect of parental nutrition. However, higher LO levels (80 % replacement of FO) in broodstock diets did not improve the growth of 16-month-old offspring, in agreement with previous studies (45) . Thus, feeding broodstock with this high-LO diet markedly reduced spawning quality, larval survival and larval and juvenile growth (45) , as a consequence of the deleterious effects of very low n-3 HUFA levels in broodstock diets (47) . The present study demonstrated the persistence of these negative effects of early EFA deficiencies during offspring life. On the contrary, 60 %-FO substitution with LO in sea bream broodstock diets did not negatively affect spawning quality or larval growth and produced 4-month-old juveniles with a better ability to utilise low-FM/FO diets (45) , in agreement with the present study. The above-mentioned growth improvement in 16-month-old fish obtained from broodstock fed moderate LO levels and the reminder low-FM/FO diet at 4 months of age was also Table 4. Biochemical composition of liver and muscle tissue after 2 months’feeding with a very low-fishmeal (5 %) and very low-fish oil (FO) (3 %) diet in 16-month-old gilthead sea bream (Sparus aurata) originated from broodstock fed linseed oil (LO) as a replacement for FO –0 % (100 % FO (F)), 60 % (40 % FO–60 % LO (LL)), 80 % (20 % FO–80 % LO (HL)) –and fed either a fishmealand FO-based diet (f) or a very low-fishmeal (5 %) and very low-FO (3 %) ‘reminder’diet (v) for 1 month at 4 months of age (Mean values and standard deviations; n3) Moisture (%) Protein (%) Lipids (%) Ash (%) Groups Mean SD Mean SD Mean SD Mean SD Liver* Ff 64·22·011·10·314·93·02·50·5 LLf 66·65·910·12·412·95·92·71·2 HLf 66·50·112·10·411·71·82·90·3 Fv 66·64·211·21·311·02·22·90·4 LLv 68·42·812·21·410·94·92·70·8 HLv 67·14·110·71·312·74·42·40 ·1 Muscle* Ff 72·30·121·60·15·10·51·40·1 LLf 72·70·520·60·95·20·71·40·1 HLf 72·50·821·30·25·41·11·40·1 Fv 73·00·121·30·34·70·21·50·1 LLv 72·51·021·30·45·20·51·50·1 HLv 72·90·121·40·24·90·51 ·40·1 * No significant differences were found for broodstock diet (P>0·05), reminder diet (P>0·05) and interaction of these two factors (P>0·05) using the two-way ANOVA analysis. Transgenerational effects of parental diets 505 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at Table 5. % Total fatty acids (FA) of livers after 2 months’feeding a very low-fishmeal (5 %) and very low-fish oil (FO) (3 %) diet in 16-month-old gilthead sea bream (Sparus aurata) originated from broodstock fed linseed oil (LO) as a replacement for FO –0 % (100 % FO (F)), 60 % (40 % FO–60 % LO (LL)), 80 % (20 % FO–80 % LO (HL)) –and fed either a fishmealand FO-based diet (f) or a very low-fishmeal (5 %) and very low-FO (3 %) ‘reminder’diet (v) for 1 month at 4 months of age (Mean values and standard deviations; n3) Groups Ff LLf HLf Fv LLv HLv Two-way ANOVA FA (%) Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD BRB×R 14 : 0 4·17 0·53 4·68 0·33 5·29 0·72 4·86 0·59 4·89 0·11 4·68 0·90 0·403 0·732 0·201 14 : 1n-7 0·05 0·01 0·05 0·00 0·05 0·00 0·05 0·01 0·05 0·00 0·04 0·00 0·745 0·477 0·480 14 : 1n-5 0·06 0·00 0·07 0·02 0·05 0·01 0·07 0·02 0·06 0·01 0·06 0·01 0·581 0·485 0·206 15 : 0 0·16 0·01 0·20 0·05 0·17 0·03 0·18 0·03 0·18 0·02 0·19 0·00 0·506 0·409 0·333 15 : 1n-5 0·03 0·00 0·03 0·00 0·03 0·00 0·03 0·00 0·03 0·00 0·03 0·00 0·625 0·220 0·654 16 : 0iso 0·03 0·00 0·04 0·01 0·03 0·00 0·03 0·00 0·03 0·00 0·04 0·01 0·888 0·851 0·241 16 : 0 12·61 0·93 14·33 1·96 14·81 0·83 13·92 0·71 14·77 0·30 13·63 0·64 0·141 0·717 0·200 16 : 1n-7 4·51 0·15 4·89 0·84 4·00 0·07 4·76 0·54 4·37 0·10 4·70 0·69 0·557 0·545 0·148 16 : 1n-5 0·10 0·01 0·11 0·01 0·09 0·00 0·10 0·01 0·10 0·00 0·11 0·02 0·476 0·401 0·038 16 : 2n-6 0·00 0·00 0·01 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·01 0·00 0·230 0·735 0·040 16 : 2n-4 0·21 0·00 0·24 0·05 0·17 0·02 0·23 0·05 0·21 0·02 0·23 0·00 0·542 0·329 0·071 17 : 0 0·19 0·00 0·20 0·01 0·15 0·01 0·20 0·04 0·18 0·01 0·20 0·01 0·259 0·165 0·112 16 : 3n-4 0·18 0·01 0·18 0·01 0·16 0·00 0·17 0·00 0·16 0·00 0·17 0·03 0·158 0·208 0·058 16 : 3n-3 0·05 0·00 0·06 0·02 0·05 0·00 0·05 0·00 0·05 0·00 0·06 0·01 0·805 0·655 0·287 16 : 3n-1 0·01 0·00 0·02 0·00 0·02 0·00 0·01 0·00 0·02 0·00 0·01 0·00 0·015 0·787 0·600 16 : 4n-3 0·12 0·01 0·11 0·01 0·07 0·01 0·12 0·04 0·11 0·01 0·09 0·00 0·037 0·609 0·437 16 : 4n-1 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·441 0·377 0·441 18 : 0 4·28 0·26 4·15 0·30 4·97 0·22 3·88 0·37 4·50 0·32 4·41 0·42 0·026 0·206 0·068 18 : 1n-9 31·93 1·02 29·93 2·77 32·06 0·48 30·14 3·42 29·60 2·52 30·10 2·38 0·571 0·259 0·812 18 : 1n-7 2·72 0·06 2·85 0·18 2·66 0·06 2·68 0·15 2·64 0·07 2·88 0·26 0·671 0·921 0·065 18 : 1n-5 0·12 0·01 0·13 0·03 0·09 0·01 0·12 0·03 0·10 0·00 0·13 0·02 0·762 0·911 0·045 18 : 2n-9 0·61 0·13 0·57 0·28 0·44 0·02 0·62 0·07 0·53 0·19 0·55 0·27 0·546 0·791 0·793 18 : 2n-6 13·10 0·47 12·92 1·26 13·87 0·81 12·93 1·02 12·86 0·18 12·93 0·95 0·617 0·375 0·672 18 : 2n-4 0·15 0·01 0·16 0·01 0·13 0·00 0·14 0·02 0·14 0·01 0·16 0·02 0·943 0·653 0·047 18 : 3n-6 1·08 0·21 1·04 0·45 0·78 0·06 1·08 0·08 0·91 0·21 0·93 0·31 0·375 0·944 0·665 18 : 3n-4 0·15 0·00 0·15 0·01 0·12 0·00 0·15 0·02 0·13 0·01 0·14 0·02 0·063 0·743 0·216 18 : 3n-3 5·32 0·36 5·03 1·19 5·74 0·28 5·15 0·78 5·31 0·06 5·03 1·16 0·891 0·577 0·556 18 : 3n-1 0·01 0·00 0·01 0·00 0·00 0·00 0·01 0·00 0·01 0·00 0·01 0·00 0·392 0·491 0·173 18 : 4n-3 0·72 0·07 0·75 0·25 0·55 0·03 0·77 0·13 0·66 0·08 0·66 0·02 0·222 0·765 0·444 18 : 4n-1 0·14 0·01 0·14 0·01 0·11 0·01 0·12 0·02 0·12 0·00 0·14 0·00 0·323 0·671 0·039 20 : 0 0·12 0·01 0·13 0·03 0·13 0·02 0·12 0·01 0·14 0·01 0·14 0·02 0·308 0·764 0·970 20 : 1n-9 0·16 0·02 0·18 0·06 0·11 0·02 0·18 0·07 0·15 0·02 0·19 0·05 0·720 0·352 0·141 20 : 1n-7 1·37 0·11 1·41 0·42 1·28 0·07 1·28 0·01 1·33 0·12 1·40 0·04 0·903 0·869 0·629 20 : 1n-5 0·13 0·01 0·14 0·03 0·11 0·01 0·12 0·01 0·12 0·01 0·14 0·02 0·803 0·835 0·232 20 : 2n-9 0·78 0·07 0·53 0·09 0·57 0·08 0·55 0·18 0·55 0·06 0·48 0·10 0·098 0·068 0·160 20 : 2n-6 0·59 0·08 0·54 0·04 0·63 0·03 0·49 0·10 0·55 0·06 0·56 0·06 0·385 0·156 0·314 20 : 3n-9 0·02 0·02 0·02 0·01 0·02 0·01 0·02 0·01 0·02 0·00 0·02 0·01 0·709 0·533 0·409 20 : 3n-6 0·52 0·08 0·39 0·06 0·37 0·04 0·34 0·07 0·35 0·02 0·37 0·04 0·144 0·018* 0·048* 20 : 4n-6 0·41 0·06 0·50 0·15 0·42 0·05 0·47 0·15 0·57 0·18 0·49 0·07 0·450 0·312 0·998 20 : 3n-3 0·45 0·05 0·39 0·03 0·47 0·03 0·37 0·09 0·42 0·04 0·43 0·09 0·448 0·277 0·308 20 : 4n-3 0·72 0·08 0·64 0·06 0·51 0·04 0·59 0·09 0·56 0·05 0·64 0·06 0·136 0·421 0·015* 20 : 5n-3 2·85 0·25 3·01 0·28 2·16 0·23 3·09 1·22 3·11 0·48 3·10 0·02 0·489 0·168 0·491 22 : 1n-11 0·48 0·10 0·61 0·33 0·33 0·09 0·52 0·11 0·52 0·10 0·57 0·01 0·525 0·479 0·301 22 : 1n-9 0·40 0·01 0·43 0·10 0·40 0·04 0·38 0·00 0·41 0·03 0·43 0·01 0·576 0·943 0·618 22 : 4n-6 0·08 0·01 0·08 0·02 0·06 0·00 0·07 0·02 0·07 0·01 0·09 0·01 0·875 0·301 0·118 22 : 5n-6 0·07 0·01 0·08 0·01 0·06 0·01 0·07 0·03 0·08 0·01 0·08 0·00 0·587 0·217 0·514 22 : 5n-3 2·25 0·47 1·98 0·21 1·36 0·24 2·06 0·91 1·79 0·21 2·25 0·28 0·457 0·487 0·152 22 : 6n-3 5·78 1·09 5·90 1·15 4·35 0·66 6·67 3·09 6·55 1·37 6·33 0·35 0·612 0·168 0·785 ∑SFA†21·55 1·23 23·73 2·49 25·55 0·81 23·21 1·48 24·70 0·62 23·28 1·97 0·088 0·876 0·135 ∑MUFA‡42·06 0·91 40·83 0·77 41·26 0·71 40·44 2·58 39·47 2·28 40·78 3·51 0·592 0·243 0·882 ∑n-6§ 15·84 0·32 15·55 1·54 16·19 0·77 15·47 0·92 15·39 0·19 15·45 0·71 0·806 0·343 0·859 ∑n-3‖18·28 1·56 17·91 0·39 15·27 1·25 18·88 4·61 18·58 1·98 18·60 1·25 0·510 0·209 0·582 c16 : 1:c16 0·36 0·03 0·34 0·01 0·27 0·01 0·34 0·05 0·30 0·01 0·35 0·07 0·122 0·748 0·036* 18 : 0:16 : 0 0·34 0·04 0·29 0·02 0·34 0·02 0·28 0·02 0·30 0·02 0·32 0·02 0·204 0·089 0·064 18 : 1:16 : 1 1·18 0·06 1·20 0·11 1·02 0·13 1·19 0·22 0·94 0·05 0·92 0·02 0·275 0·633 0·047* B, broodstock; R, reminder; B×R, interaction of broodstock and reminder. *Pvalues under 0·05. †∑SFA include 14 : 0, 15 : 0, 16 : 0, 17 : 0, 18 : 0 and 20 : 0. ‡∑MUFA include 14 : 1n-7, 14 : 1n-5, 15 : 1n-5, 16 : 1n-5, 18 : 1n-9, 18 : 1n-7, 18 : 1n-5, 20 : 1n-9, 20 : 1n-7, 20 : 1n-5, 22 : 1n-11 and 22 : 1n-9. §∑n-6: n-6 series PUFA include 16 : 2n-6, 18 : 2n-6, 18 : 3n-6, 20 : 2n-6, 20 : 3n-6, 20 : 4n-6, 22 : 4n-6, 22 : 5n-6. ‖∑n-3:n-3 series PUFA include 16 : 3n-3, 16 : 4n-3, 18 : 3n-3, 18 : 4n-3, 20 : 3n-3, 20 : 4n-3, 20 : 5n-3, 22 : 5n-3, 22 : 6n-3. 506 S. Turkmen et al. https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at accompanied by an enhanced utilisation of the low-FM/FO diet, as denoted by the better FCR, which could be related to the modulation of lipid and carbohydrate metabolism. There were no large differences in the proximate composition and fatty acid profiles of sea bream liver and muscle, reflecting the profound effect of the diet, regardless of the nutritional history of the different fish groups. Nevertheless, both broodstock diet and reminder diet had a significant effect on some major fatty acids in the liver such as 18 : 0, a terminal product of lipogenesis, and in the ratios 18 : 0:16 : 0 and, particularly, 18 : 1:16 : 1. Both 16 : 0 and 16: 1 are substrates for elovl6, a key rate-limiting enzyme in the long-chain fatty acid elongation cycle and, therefore, the ratios 18: 0:16: 0 and 18:1:16:1, are indicators of the activity of this enzyme. These results are in agreement with the hepatic elovl6 expression, which was down-regulated by the increase in LO in the broodstock diets and was correlated inversely to the 16:0 contents in the liver and directly to the 18 : 0:16 : 0 and 18 : 1:16 : 1, denoting a significant post-transcriptional effect. Besides, LO increase in broodstock diets also increased the hepatic 18 : 0:18 : 1 ratios in the 16-month-old offspring. These results are in agreement with the 16 : 0 reduction and 18 : 0:18 : 1 increase in mice models with Elovl6 disruption, which showed protection against a high-SFA diet-induced insulin resistance that lead to hepatosteatosis similar to that of Table 6. % Total fatty acids of muscle after 2 months’feeding of a very low-fishmeal (5 %) and very low-fish oil (FO) (3 %) diet in 16-month-old gilthead sea bream (Sparus aurata) originated from broodstock fed linseed oil (LO) as a replacement for fish oil –0 % (100 % FO (F)), 60 % (40 % FO–60 % LO (LL)), 80 % (20 % FO–80 % LO (HL)) –and fed either a fishmealand FO-based diet (f) or a very low-fishmeal (5 %) and very low-FO (3 %) ‘reminder’diet (v) for 1 month at 4 months of age (Mean values and standard deviations; n3) Groups Ff LLf HLf Fv LLv HLv FA (%)* Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD 14 : 0 4·83 0·15 4·90 0·15 5·17 0·15 4·99 0·33 4·98 0·03 4·97 0·35 14 : 1n-7 0·08 0·00 0·08 0·00 0·08 0·00 0·08 0·00 0·07 0·00 0·08 0·00 14 : 1n-5 0·08 0·00 0·09 0·03 0·08 0·02 0·09 0·01 0·08 0·00 0·08 0·01 15 : 0 0·21 0·01 0·23 0·01 0·21 0·01 0·22 0·02 0·22 0·00 0·22 0·02 15 : 1n-5 0·02 0·00 0·03 0·00 0·03 0·00 0·02 0·01 0·03 0·00 0·03 0·00 16 : 0iso 0·03 0·00 0·03 0·00 0·03 0·00 0·03 0·00 0·03 0·00 0·03 0·01 16 : 0 14·69 0·42 15·03 0·41 14·52 0·21 15·03 0·12 15·18 0·31 14·88 0·48 16 : 1n-7 4·89 0·31 5·12 0·18 4·51 0·14 4·92 0·64 4·89 0·36 4·87 0·72 16 : 1n-5 0·08 0·00 0·09 0·00 0·08 0·00 0·08 0·01 0·08 0·01 0·08 0·01 16 : 2n-6 0·05 0·04 0·01 0·00 0·01 0·00 0·01 0·00 0·01 0·00 0·01 0·00 16 : 2n-4 0·36 0·07 0·33 0·02 0·29 0·01 0·31 0·05 0·31 0·01 0·31 0·04 17 : 0 0·28 0·03 0·30 0·01 0·27 0·00 0·29 0·04 0·28 0·01 0·29 0·03 16 : 3n-4 0·14 0·00 0·14 0·00 0·14 0·01 0·14 0·01 0·14 0·00 0·14 0·01 16 : 3n-3 0·05 0·00 0·05 0·01 0·04 0·00 0·05 0·02 0·04 0·01 0·05 0·01 16 : 3n-1 0·05 0·02 0·07 0·01 0·07 0·01 0·09 0·01 0·07 0·00 0·08 0·02 16 : 4n-3 0·35 0·11 0·34 0·03 0·31 0·01 0·33 0·06 0·34 0·01 0·30 0·09 16 : 4n-1 0·01 0·00 0·01 0·00 0·01 0·00 0·01 0·00 0·01 0·00 0·01 0·01 18 : 0 3·50 0·04 3·45 0·11 3·55 0·11 3·55 0·07 3·53 0·00 3·61 0·10 18 : 1n-9 27·27 0·21 27·23 0·92 28·47 0·99 27·23 1·37 27·77 0·28 27·66 1·00 18 : 1n-7 2·85 0·06 2·90 0·01 2·74 0·04 2·84 0·16 2·82 0·08 2·84 0·14 18 : 1n-5 0·39 0·45 0·13 0·01 0·11 0·01 0·13 0·02 0·12 0·01 0·13 0·02 18 : 2n-9 0·22 0·02 0·25 0·06 0·22 0·01 0·24 0·02 0·24 0·02 0·22 0·02 18 : 2n-6 11·52 0·40 11·46 0·55 12·77 0·49 11·62 1·36 11·74 0·62 11·79 1·59 18 : 2n-4 0·13 0·01 0·14 0·00 0·13 0·01 0·13 0·01 0·13 0·01 0·14 0·01 18 : 3n-6 0·35 0·05 0·36 0·05 0·33 0·01 0·35 0·03 0·34 0·02 0·32 0·00 18 : 3n-4 0·15 0·01 0·14 0·00 0·13 0·00 0·15 0·01 0·13 0·01 0·14 0·01 18 : 3n-3 5·11 1·11 4·53 0·51 5·48 0·47 4·57 1·16 4·76 0·40 4·75 1·18 18 : 3n-1 0·18 0·30 0·01 0·00 0·01 0·00 0·01 0·00 0·01 0·00 0·01 0·00 18 : 4n-3 0·69 0·04 0·73 0·03 0·64 0·03 0·69 0·07 0·70 0·05 0·67 0·06 18 : 4n-1 0·22 0·13 0·14 0·00 0·14 0·01 0·13 0·01 0·14 0·01 0·15 0·01 20 : 0 0·24 0·01 0·24 0·01 0·24 0·01 0·24 0·00 0·24 0·01 0·24 0·00 20 : 1n-9 0·28 0·02 0·29 0·03 0·25 0·03 0·31 0·09 0·28 0·03 0·28 0·07 20 : 1n-7 2·36 0·13 2·44 0·24 2·19 0·10 2·43 0·38 2·37 0·11 2·35 0·34 20 : 1n-5 0·16 0·01 0·17 0·01 0·16 0·01 0·17 0·02 0·16 0·01 0·16 0·02 20 : 2n-9 0·31 0·01 0·31 0·02 0·31 0·03 0·30 0·02 0·30 0·01 0·28 0·01 20 : 2n-6 0·60 0·29 0·42 0·02 0·45 0·00 0·42 0·02 0·42 0·00 0·43 0·02 20 : 3n-9 0·03 0·01 0·03 0·01 0·05 0·01 0·04 0·02 0·04 0·03 0·06 0·03 20 : 3n-6 0·23 0·00 0·22 0·01 0·24 0·02 0·22 0·01 0·21 0·01 0·21 0·03 20 : 4n-6 0·46 0·04 0·46 0·04 0·42 0·06 0·46 0·06 0·44 0·00 0·46 0·05 20 : 3n-3 0·33 0·08 0·28 0·00 0·32 0·00 0·28 0·04 0·28 0·01 0·30 0·04 20 : 4n-3 0·58 0·00 0·58 0·01 0·54 0·02 0·55 0·05 0·55 0·03 0·57 0·05 20 : 5n-3 4·36 0·21 4·42 0·19 3·88 0·27 4·15 0·62 4·22 0·17 4·26 0·61 22 : 1n-11 1·42 0·16 1·53 0·24 1·21 0·12 1·52 0·40 1·45 0·15 1·41 0·36 22 : 1n-9 0·56 0·02 0·57 0·06 0·53 0·01 0·59 0·07 0·56 0·01 0·55 0·05 22 : 4n-6 0·10 0·01 0·10 0·01 0·10 0·01 0·11 0·01 0·10 0·00 0·11 0·01 22 : 5n-6 0·12 0·01 0·13 0·00 0·11 0·01 0·13 0·01 0·12 0·01 0·13 0·00 22 : 5n-3 2·07 0·15 2·10 0·20 1·88 0·15 2·08 0·27 1·99 0·10 2·09 0·26 22 : 6n-3 6·99 0·73 7·41 0·66 6·53 0·89 7·64 1·13 7·03 0·12 7·27 0·81 * No significant differences were found for broodstock diet (P>0·05), reminder diet (P>0·05) and interaction of these two factors (P>0·05) using the two-way ANOVA analysis. Transgenerational effects of parental diets 507 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at wild-type mice (51) . This protection was related to the restoration of hepatic insulin receptor substrate-2, suppression of hepatic protein kinase-C ɛand restoration of Akt phosphorylation (51) , overall, indicating a better utilisation of dietary carbohydrates under conditions of high-fat diet-induced insulin resistance. Indeed, insulin stimulates acetyl-CoA carboxylase that produces malonyl-CoA, which inhibits CPTI activity and affects utilisation of fatty acids and glucose as substrates (76) . Thus, genes related to fatty acid oxidation, such as CptI, are down-regulated in mice with Elovl6 disruption (51) , whereas up-regulation of CptI expression caused by intra-uterine growth restriction increases the risk of type 2 diabetes in adulthood (59) . In agreement, in the present study, down-regulation of elovl6 was correlated to down-regulation of cptI, a rate-limiting enzyme for fatty acid oxidation in mitochondria (77) . Moreover, increased LO in broodstock diet and increased plant-protein and lipid sources in the 4-month-old reminder diet induced the down-regulation of cptI in the gilthead sea bream offspring, evidencing a long-term nutritional programming effect. A down-regulation of cptIb gene expression was also found in the liver of juvenile rainbow trout by vitamin supplementation at first feeding showing that nutritional interventions during developmental plasticity (larval period) may provoke longer-term effects later in life (78) .CptI expression in the liver of fish is down-regulated by the reduction of dietary PUFA, particularly, LC-PUFA (57,79) . Accordingly, cptII expression is also down-regulated in Atlantic salmon when dietary FO is substituted with VO (80,81) .In vitro studies in rainbow trout hepatocytes showed that pparαand cptI are up-regulated by MUFA and down-regulated by EPA, among other fatty acids (80,81) . In the present study cpt1b expression was highly correlated to 18 : 1n-9 and negatively correlated to 20 : 5n-3 in the liver. Besides, LO increase in broodstock diets significantly reduced the liver contents of 16 : 4n-3, an intermediate product of β-oxidation of EPA. In gilthead sea bream offspring, cptI expression in the liver is negatively correlated to ppara, suggesting that nutritional programing by LO reduces β-oxidation in the mitochondria but not in the peroxisomes. In mammals, parental feeding with a high-lipid diet lead to hypomethylation of four specific CpG dinucleotides in PPARa and the modification of the mRNA transcript in juvenile offspring (60) . In gilthead sea bream fed a low-FM/FO diet, hepatic ppara is reportedly down-regulated, in association with retarded growth (82) . In the present study, ppara in the offspring of HLv fed a low-FM/FO diet was not down-regulated and Groups Relative fold expression 2 4 6 Ff LLf HLf Fv LLv HLv Ff LLf HLf Fv LLv HLv 1.5 2.0 2.5 3.0 3.5 4.0 0.4 0.6 0.8 1.0 fads2 0.5 1.0 1.5 2.0 cox2 elovl6 2 3 4 ppara lpl 2 4 6 8 cpt1 aa,b b a a,b b B<0 .001 R<0 .01 B×R=NS B<0 .01 R<0 .05 B×R=NS a b b,c c a,b a,b a a,b b a,b ba,b B<0 .01 R=NS B×R=NS Fig. 2. Box and whisker plots of relative fold expression (groups v. control sample) of six different genes after challenging 16-month-old gilthead sea bream individuals with high-vegetable oil and high-meal feeds for 2 months. lpl, lipoprotein lipase; elovl6, elongation of very long-chain fatty acids protein 6; fads2, fatty acid desaturase 2; cox2, cyclo-oxygenase-2; cpt1, carnitine palmitoyltransferase I; , maximum and minimum fold expression; , upper and lower quartiles; , median; ,Pvalues of two-way ANOVA; B, broodstock diet; R, reminder diet; B×R, interaction of these two parameters, ns, P>0·05; F, 100 % fish oil; LL, 40 % fish oil–60 % linseed oil; and HL, 20 % fish oil–80 % linseed oil; f, fish oil-based diet; v, reminder diet. n3 for all the groups and genes. a,b,c Mean values with unlike letters were significantly different between each group, no indications mean no significant difference (P>0·05). 508 S. Turkmen et al. https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114517002434 Downloaded from https://www.cambridge.org/core. ULPGC. Biblioteca Universitaria, on 20 Oct 2017 at 09:54:40, subject to the Cambridge Core terms of use, available at