Dietary Natural Plant Extracts Modulate Growth Performance and Antioxidant Status of Senegalese Sole Postlarvae

Oxidative stress has a direct impact on sh production, affecting both growth and health status. Plant based extracts, such as those from green tea, curcumin and grape seeds, are known for their abundant content and diversity of polyphenols. These bioactive compounds have a high antioxidant capacity making such extracts good additives to include in sh diets potentially improving the oxidative status of sh, and therefore enhancing growth and stress resistance of farmed sh. A growth trial with Senegalese sole postlarvae (45 DAH) fed with four experimental diets, a control (CTRL) and three supplemented with natural antioxidants, namely curcumin (CC), green tea (GT) and grape seed (GS) extracts, was performed during 25 days to check if these supplements could improve growth performance and sh oxidative status. Moreover, a thermal stress was applied at the end of the growth trial, to assess the effect of these supplemented diets in stress resistance of sole juveniles. Sole growth was improved by the dietary inclusion of CC and GS compared to CTRL. Juveniles from CC and CTRL present the lowest values of oxidative damage (lipid peroxidation (LPO) and protein carbonylation (PC)). Moreover, the sh fed CC showed a decrease in stress related biomarkers (heat shock protein 70 (HSP70) and glutathione-S-transferase (GST)) compare to CTRL, which might be due to direct antioxidant capacity. In contrast the supplementation of GT and GS increased the content of oxidative damage in sole reared in standard conditions. However, after a thermal stress exposure both GT and GS treatment seem to prevent the increase of PC and decrease of antioxidant glutathione (GSH) in sole and depending on the time of exposure. In conclusion, dietary supplementation with these natural extracts modulates oxidative status and stress response after a short/long term temperature increase. The incorporation of curcumin seems to be a safe additive to include in sole diets, enhancing growth and oxidative status. The supplementation of the studied doses of green tea and grape seed extracts in sh diets needs further research as they may act as pro-oxidant depending of the culture conditions. on growth 16 . In Nile tilapia fed with ve experimental diets containing graded levels of green tea, only two doses, namely 0.25 and 0.50 g/kg, were able to improve sh nal body weight and specic growth rates 37 . The negative effects of green tea on rat’s growth have been related to the content of caffeine and the combination of polyphenols with polysaccharides that interact with the absorption of fatty acids and serum triglyceride levels 38 . All these results suggest that the dose of green tea supplementation able to promote growth performance in sh still needs to be ascertained and is probably species and dose dependent. (measured as MDA) compare to the control group 17 . While in hybrid sturgeon-sterlet (Huso huso × Acipenser ruthenus) LPO content and CAT and SOD activity decreased in sh fed dietary green tea supplementation 46 . The observed increase in PC content in juveniles feed with grape seeds has not been previously reported. Juveniles of rainbow trout fed with grape seed oil supplementation, at higher concentrations, also showed an improvement in the growth and a decreased activity of some antioxidant defenses (SOD, CAT and GST) 33 . Moreover, the use of grape pomace our, in juveniles of grass carp, did not signicantly affect serum and splenic CAT and SOD activities when compared to the control group 34 . In contrast, a reduction of oxidative damage of lipids (LPO) was observed in rainbow trout fed with grape seed extracts supplementation, whilst the TAC content was unaltered however, conrming our present observations in sole. An increase in the expression of endogenous antioxidants (kat, gpx1 and gst) was also observed in this species when fed at the lower dose 47 . The inclusion of green tea and grape seed extracts in sh diets can improve the oxidative status and consequently improve the endogenous antioxidant defences in sh species. However, the dose of the antioxidant supplementation and the duration of the administration period still need tuning. Moreover, it is also important to highlight that postlarvae have higher growth rates and feed intake than juveniles and adults, which might have contributed to increased concentration of antioxidants per body weight unit possibly explaining the pro-oxidant effect observed in this study. Nevertheless, these results also corroborate previous concepts, including that polyphenols can modulate oxidative status biomarkers by different mechanisms of action 6 . interfere in numerous routes by acting as direct antioxidant as chelating metals, reducing via electron transfers or hydrogen atoms transference, and indirectly by upregulating the expression of endogenous antioxidants. The CC and GS diets improved growth of sole postlarvae, although the mechanism by which these supplements act seem different. It seems that curcumin extract (CC diet) improved the oxidative status of the sole, showing low levels of protein and lipid degradation and a reduction on GST and HSP70 activity, allowing sh to invest more energy into growth rather than on endogenous oxidative defences. The GS diet, on the other hand, showed a negative impact in the oxidative status of the juveniles by increasing the protein damage, so the positive effect on growth may be reverted in the long-term. The mechanisms by which this antioxidant acts still requires further investigation. The inclusion of green tea extract (GT diet) did not improve growth performance nor the sole antioxidant status during the growth trial. Both green tea and grape seed extracts seem to act as pro-oxidants, by increasing the oxidative damage in sole juveniles, under standard rearing conditions. However, in response to a thermal stress, both GT and GS diets increased the sh capacity to cope with the new stressful event. However, these antioxidants appear to act in distinct time manners and exert different responses: Grape seed extract seems to have an immediate action and result in a short-term improvement on sh oxidative status, while green tea extract has a long-term effect on the antioxidant capacity of the sole. Hence, curcumin extract seems a good candidate for long-term supplementation of young sh diets, as it improves welfare and growth of Senegalese sole. The use of green tea and grape seed extracts in diets for young sh still requires further evaluation to identify the most adequate inclusion level, although the short-term use of the tested doses seems a feasible solution before highly stressful short periods (e.g., transportation, handling and temperature rises). These results suggest that dietary natural plant extracts can improve young sh robustness and even promote growth when supplemented at the optimal doses.

Many factors can in uence the antioxidant defence response in sh. Biotic factors such as age, species and feeding behaviour and abiotic conditions such as temperature, diet, dissolved oxygen, and toxins present in the water can modulate the antioxidant defences and as a consequence, the oxidative status of the animals 28,29 . It is crucial to assess the sh oxidative status, as it has a direct impact in health, esh quality and growth, in order to improve sh welfare and promote cost-effectiveness in aquaculture. Therefore, the aim of this work was to evaluate if dietary plant extracts supplementation (curcumin, green tea and grape seed) modi es growth performance and antioxidant status of Senegalese sole postlarvae, and help sh to cope with a stressful event.

Dietary antioxidant properties
Main features for Radical scavenging activity (RSA) of the methanol extracts made from the experimental diets are summarized in Table 3. Overall, the extracts had low to moderate RSA. Regarding the ABTS assay, the GS diet presented the highest activity (28.6% for ABTS), while in the DPPH method, the utmost activity (50.2%) was recorded for the CC diet. The TPC ranged from 1.3 to 6.4 mg GAE/g diet, and the three supplemented diets had higher levels than CTRL (P<0.001). The content of avonoids varied from 0.7 to 4.4 mg QE/g diet. The GS diet presented the highest amounts of avonoids, similar to the CC diet (P<0.001). As average, the total phenols and avonoids content in the supplemented diets was 4-fold higher than in the CTRL diet.

Key performance indicators
Postlarvae fed with CC and GS diets were heavier than CTRL diet at the end of the feeding trial (P<0.001) ( Table 4). All postlarvae fed with supplemented diets showed a higher length than CTRL group (P<0.001). Condition factor (K) was higher than 1 in all treatments, however, sh fed the supplemented diets had signi cantly lower values than those fed with CTRL. The relative growth rate (RGR) remained unaffected by the dietary treatments. The average survival rate was 98.3±0.6% and did not differ between dietary treatments (Table 4).

Oxidative status indicators
The CAT and TAC activities were not statistically different among the different dietary treatments (Table 5). All juveniles fed with the supplemented diets showed a lower GST activity than CTRL (P<0.01). The content of GSH was signi cantly higher in soles from dietary CC and CTRL treatments (P<0.001). When HSP70 proteins were analysed, soles fed with CC and GS had lower levels than CTRL and GT (P<0.001). The oxidative damage measured as LPO was not signi cantly different between supplemented diets and the CTRL, and only a signi cant reduction in CC with respect to the GT diet was found. The juveniles from CC and CTRL diets had the lowest values of oxidative damage observed at the protein level (PC) when compared with sh from GT and GS diets that presented the highest values (P<0.001) (Fig. 1).
Expression of most of the genes related to oxidative stress was not modi ed by dietary treatments except for cat and gpx1 gene (P<0.05). The postlarvae fed with GT had lower cat mRNA levels than GS. An up-regulation of gpx1 transcription was observed in sh fed with CC and GT compared to the other dietary treatments. There was also a tendency (but not statistically signi cant) for an up-regulation of hsp90aa in CC and to a lesser extent GS and GT treatments compared to CTRL.

Thermal stress -acute exposure
All physiological indicators, except LPO, were signi cantly modi ed after the acute thermal stress. Fish CAT and GST activities, GSH levels and HSP70 contents were reduced, and TAC and PC levels were signi cantly increased after the acute thermal stress (Table 6).
When the effects of dietary treatments were evaluated, statistically signi cant differences were found for GST, GSH, HSP70, TAC and PC; a signi cant interaction between dietary treatment and response to thermal stress was found for CAT, GST, GSH, HSP70 and PC (Diet*Stress P<0.05) ( Table 6).
Unlike CTRL and GT diets, CC and GS diets did not modify the juveniles' GST activity after the acute stress. Moreover, the CTRL and CC soles were unable to maintain high levels of GSH (P<0.001). When the shift in GSH levels was compared after an acute stress, all dietary treatments revealed a reduction in GSH activity, but the lowest decrease was observed in GS and GT treatments (P<0.001) (Fig. 2). The levels of HSP70 were similar after the acute stress conditions in soles fed with CTRL, CC and GS diets, however, in GT a 10-fold increase of HSP70 levels was observed after acute stress (P<0.001; Fig. 2). A slight increase was also observed in GS juveniles compared to CTRL group that presented a decrease in the delta variation between acute stress/standard conditions of the HSP70 levels (P<0.001) (Fig. 2). The total antioxidant content (TAC) was in uenced by diet and stress independently without a signi cant interaction (Diet*Stress >0.05). TAC increased in postlarvae after acute stress conditions (P<0.001). The juveniles fed with CC have a signi cant higher TAC than juveniles fed GS diet (P<0.05). The PC content was in uenced by diet and stress with a signi cant interaction (P <0.001). The PC was higher in soles fed with CTRL and CC diets after the acute stress (P<0.001). On the other hand, the sh fed with GS were able to decrease the PC content under an acute stress condition.
The LPO content in sole juveniles was affected only by the dietary treatments (P<0.01) and not by the acute stress conditions (P>0.05). Fish from CC treatment had the lowest LPO when compared to those from GT treatment and did not differ from the remaining treatments.
Expression of most genes (cat, gpx3, gsr, hsp90b and hsp70) related to oxidative stress were affected by the acute stress (P<0.05) ( Table 6). Juveniles exposed to a thermal acute stress reduced gpx3 and gsr mRNA levels and increased hsp90b and hsp70 transcript amounts independently of the dietary treatment (P>0.05). On the other hand, the expression levels of cat decreased after acute stress, however, the sh fed with GS up-regulated the expression of cat compared to CTRL and GT (P<0.05). The expression of sod3 and gpx1 were not modi ed by the thermal stress and dietary treatments (P>0.05). A signi cant interaction between stress and the diets (P<0.001) was observed in the hsp90aa expression showing that the dietary treatments modulated differently sole response to a stressful situation. Both CC and GT soles downregulated hsp90aa mRNA levels after the acute stress contrarily to CTRL and GS sh that up-regulated the expression of this gene.

Thermal stress -chronic exposure
All biomarkers, except PC, modi ed their levels after chronic stress, and four of them (GST, GSH, LPO and PC) responded to the dietary treatment; a signi cant interaction was found for GSH, HSP70 and PC (P<0.001) ( Table 7). After chronic stress the juveniles from GS diet presented the highest activity of GSH compared to CTRL and CC groups (P<0.001). The variation of the activity of this tripeptide, under standard rearing and chronic stress conditions, decreased in all treatments, however in GT and GS sole this decline was smaller than in the remaining diets (P<0.001) (Fig. 3). After chronic stress, juveniles from all dietary treatments showed similar levels of HSP70, however, only CC and GS sh were able to improve the activity of this chaperone when compared to standard rearing conditions. The PC content decreased in sh fed with GT after the chronic exposure; on the contrary, sh fed with GS presented the highest amounts. The sh activity of CAT and TAC was higher after exposure to a chronic thermal stress (P<0.001), regardless of the dietary treatment. Sole decreased the activity of GST and LPO content after chronic stress exposure (P<0.05). Fish from CC, GS and GT diets presented lower GST than the CTRL fed sh (Diet<0.001). The LPO content was lower in CTRL and CC fed sh that in sh fed GT diet (Diet<0.001). The variation from chronic stress and standard condition showed that CC, GS and CTRL sole had improved PC content when compared to GT sh (Fig. 3).
The expression of gpx3, gsr, hsp90b and hsp70 was affected by the long exposure to higher temperature (P<0.05), but not by the dietary treatments (P>0.05). After 7 days all juveniles up-regulated the expression of hsp70 and gpx3 and down-regulated gsr. Moreover, the cat mRNA levels increased after thermal chronic stress (P<0.001) and the sh fed with GS presented a signi cantly higher expression than those fed with GT (P<0.05). The expression of sod1 was not affected by the dietary treatments or thermal exposure. (P>0.05) The expression of gpx1 and hsp90aa presented an interaction between chronic stress conditions and the dietary treatments (P<0.05). The long exposure to high temperature up-regulated gpx1 mRNA levels only in CTRL and GT postlarvae. After chronic temperature exposure, the transcription of hsp90aa in CC sole signi cantly decreased to values similar of those observed in the remaining treatments.

Integrative analysis
An integrative analysis of all the data collected in the present study is showed in Figure 4. The performance indicators and physiological and gene expression biomarkers of oxidative status of the postlarvae fed the supplemented diets (CC, GT, and GS) were normalised in relation to CTRL sole at standard conditions ( Fig. 4). At the end of the growth trial an increase of 12-17% in sole growth was observed in CC and GS fed sh compare to CTRL sole. However, an increase in protein oxidation and decrease in some endogenous antioxidant defences were also recorded in GT and GS sh. After acute stress exposure the endogenous antioxidant defences decreased and PC content increased, although in GS sole the values were similar to the CTRL sh at standard conditions. Overall, after a chronic exposure an increase in the expression and activity of several antioxidant defences was observed regardless of the dietary treatment.

Discussion
In the last decade there has been a growing interest in dietary inclusion of natural extracts with high biological activity as feed additives to enhance growth performance and welfare of farmed sh species. In this work, three experimental diets supplemented with antioxidants (extracts from turmeric, green tea and grape seed) were tested. To gain insight into the antioxidative potential of the dietary supplements, two radical scavenging assays (ABTS and DPPH) and the TPA and avonoid contents were assessed in methanol extracts of the experimental diets. The extract from curcumin (CC) and grape seed (GS) diets exhibited the highest capacity to scavenge free radicals. As expected, the supplementation of these antioxidants increased the TPA and avonoid contents when compared to CTRL diet. Phenolic compounds contain an aromatic ring bearing one or most hydroxy substituents, including functional derivatives (e.g., esters, methyl ethers, glycoside) and their content is usually associated with the antioxidant capacity of plant extracts, as they can act as hydrogen or electron donator in the presence of oxidants 30,31 .
The present results showed that diet supplementation with curcumin (CC diet) and grape seed (GS diet) extracts promote larger sole postlarvae. In fact, previous data showed that these supplemented diets improve sole growth performance through a modulation in the expression of genes related to muscle growth and development 21 . The supplementation of curcumin started recently in diets for freshwater sh juveniles, such as in crucian carp 32 , rainbow trout 14 , common carp 10 and Nile tilapia 13,15 . Fewer information is available regarding grape seed extract and results are not always congruent regarding growth performance. In one experiment using juveniles of rainbow trout, an improvement of growth was registered 20 . In another experiment on rainbow trout, only juveniles fed the highest doses of grape seed improved the nal weight 33 . On contrary, in grass carp, the supplementation was unable to show any effect on growth 34 . Therefore, antioxidants supplementation may present different outputs depending on the compound dose and the species. In this study, CC and GS diets improved sole growth, but the same was not observed for green tea extract supplementation (GT diet). Similarly, in rainbow trout 17,35 , grass carp 16 and olive founder juveniles 36 fed with green tea supplemented diets no positive effect on growth performance ( nal body weight) or feed conversion ratio could be observed. In the case of grass carp juveniles, the green tea supplementation had a detrimental effect on growth 16 . In Nile tilapia fed with ve experimental diets containing graded levels of green tea, only two doses, namely 0.25 and 0.50 g/kg, were able to improve sh nal body weight and speci c growth rates 37 . The negative effects of green tea on rat's growth have been related to the content of caffeine and the combination of polyphenols with polysaccharides that interact with the absorption of fatty acids and serum triglyceride levels 38 . All these results suggest that the dose of green tea supplementation able to promote growth performance in sh still needs to be ascertained and is probably species and dose dependent.
The K factor decreased in sole fed with supplemented diets with respect to the CTRL. A high K is sometimes linked to a better nutritional state of the sh. However, this does not seem to be the case in our data. The decrease observed in sh supplemented with the dietary antioxidants does not mean that these sh were in a worse physical condition than those fed with the CTRL diet. In fact, this result indicates that sh had a faster growth in length than body mass provoking a shift in body shape (more elliptic). A similar study performed in tilapia fed with different curcumin supplementations also showed that the sh from the treatment with nal higher body weight and growth had a signi cant lower K 13 .
Growth and development are well orchestrated processes that depend on the balance of cellular proliferation, differentiation, and apoptosis. The cell fate depends on a variety of molecular pathways, gene expression and protein function that are sensitive to the cellular reduction potential. Measuring the levels of antioxidant defences and oxidative damage is highly relevant to assess tness, because high oxidative stress levels may compromise survival and growth 39,40 . The results from the present study suggest that CC diet supplementation improved oxidative status in Senegalese sole juveniles. Contradictory results were observed with the GT and GS diets that showed an increase in the protein oxidative damage content at the end of the growth trial.
Both sh fed with CC and CTRL diet showed the lowest levels of lipid and protein oxidation. However, in order to maintain these similar levels of LPO and PC the juveniles from CTRL needed to increase endogenous antioxidant defence (GST and HSP70), which are linked to the biotransformation process of metabolites and to the protection of the cell during stress. Increasing antioxidant defences is an energy-consuming process and may de ect energy or nutrients that might be needed for other physiological functions. So, curcumin extract supplementation in sole diets may spare antioxidant defences by acting directly as a ROS scavenger and consequently mitigating cellular oxidative stress, thereby allowing juveniles to invest the energy in other functions like growth.
Similarly, was reported in tilapia 41 and common carp 10 fed a lower dose of curcumin supplementation, an improvement in nal weight. However, no differences were registered in the content of lipid peroxidation (measured as MDA) and activity of endogenous antioxidants defences when compared with the control treatment. In a test performed with climbing perch (Anabas testudineus) fed with two doses of curcumin for 8 weeks, unaltered levels of endogenous enzymes (GR, CAT and GPx) and oxidative damage LPO were observed in treated sh compared to the control, also suggesting that curcumin have a direct free radical scavenging activity 42 . In juveniles of tilapia the growth improvement observed in the juveniles fed the higher dose of curcumin was accomplish by a decrease of LPO, but with no differences in the activity of CAT and GSH 15 . Nevertheless, in a different study the supplementation of curcumin also reported enhanced growth of tilapia and trout juveniles but through a decrease in the lipid peroxidation and improvement in the activity of antioxidant defences 13,43 . In fact, the impact of curcumin in the antioxidant status of sh highly depends on the doses 10,15 and duration of the experimental trial 42 .
It appears that the supplementation of both green tea (GT diet) and grape seed (GS diet) might have a pro-oxidant effect in sole postlarvae. Although there are just a few studies regarding the pro-oxidant effects of antioxidants, a variety of plant extracts showed to have both pro-oxidative and antioxidative properties, depending on their main bioactive molecules characteristics (e.g. metal-reducing potential, chelating behaviour), concentration and the microenvironment (e.g. pH and presence of metal ions and redox status) 44,45 . As far we know there are no previous studies regarding the pro-oxidant effect of either green tea or grape seeds extracts in the oxidative status of sh. The inclusion of Epigallocatechin-3-gallate (catechin present in green tea) and green tea in diets for rainbow trout juveniles 35 and grass carp 16 , respectively, did not affect the oxidative status parameters analysed, corroborating the present results. On the other hand, rainbow trout juveniles presented a higher SOD activity and lower LPO content (measured as MDA) compare to the control group 17 . While in hybrid sturgeon-sterlet (Huso huso × Acipenser ruthenus) LPO content and CAT and SOD activity decreased in sh fed dietary green tea supplementation 46 .
The observed increase in PC content in juveniles feed with grape seeds has not been previously reported. Juveniles of rainbow trout fed with grape seed oil supplementation, at higher concentrations, also showed an improvement in the growth and a decreased activity of some antioxidant defenses (SOD, CAT and GST) 33 . Moreover, the use of grape pomace our, in juveniles of grass carp, did not signi cantly affect serum and splenic CAT and SOD activities when compared to the control group 34 . In contrast, a reduction of oxidative damage of lipids (LPO) was observed in rainbow trout fed with grape seed extracts supplementation, whilst the TAC content was unaltered however, con rming our present observations in sole. An increase in the expression of endogenous antioxidants (kat, gpx1 and gst) was also observed in this species when fed at the lower dose 47 . The inclusion of green tea and grape seed extracts in sh diets can improve the oxidative status and consequently improve the endogenous antioxidant defences in sh species. However, the dose of the antioxidant supplementation and the duration of the administration period still need tuning. Moreover, it is also important to highlight that postlarvae have higher growth rates and feed intake than juveniles and adults, which might have contributed to increased concentration of antioxidants per body weight unit possibly explaining the pro-oxidant effect observed in this study. Nevertheless, these results also corroborate previous concepts, including that polyphenols can modulate oxidative status biomarkers by different mechanisms of action 6 .
High water temperatures increase oxygen consumption and mitochondrial respiratory rates leading to an increment of proton leakage rates that could lead to oxygen incomplete reduction and ROS formation [64]. Our data suggest that sole fed with CC and CTRL diets were unable to maintain the lower PC and the highest GSH content after an acute increase of temperature. Moreover, sh fed with CTRL could not maintain the levels of GST after 24 h of thermal stress. On the other hand, sole fed with GS were able to maintain the activity of GST and GSH after the acute stress and were even able to decrease the protein carbonylation (PC). It seems that the pro-oxidant effect under standard rearing condition, has changed to antioxidant under acute challenging conditions. In fact, dietary pro or antioxidant can have different impacts on oxidative status of sh depending on normal/stressful rearing condition 48 . This may suggest that the supplementation dose needs to be tuned between promoting growth in standard rearing conditions and to help sh coping with environmental changes. In other sh species, a more evident effect of curcumin supplementation in response to a stressful situation was observed. For example, both rainbow trout 14 and tilapia 41 fed diets supplemented with CC showed a decrease in LPO content. Moreover, the expression of Nrf2 and antioxidant defences activity (SOD and GPx) were also promoted in these sh species, respectively.
In abalone fed a grape seed extract supplementation diet under a thermal stress (increase of water temperature) the expression of cat was up-regulated compared to control diet, corroborating our study and evidencing the protective effect of grape extracts on oxidative induced stress. Tilapia fed diets with different doses of resveratrol were able to counteract some negative effects of intraperitoneal injection of H 2 O 2 in the oxidative status of the sh, by decreasing LPO and enhancing CAT activity and TAC content [66]. Juveniles of grass carp infected with pseudmonas aeruginosa and fed grape pomace our supplementation had an increased CAT activity in serum and splenic tissues 34 .
After one week of thermal stress (chronical exposure), the sh might be adapted to the new conditions and restore the homeostatic values or reach an allostatic equilibrium. LPO values of the sh decreased independently of the dietary treatment and the values of PC of the juveniles from GS raised again. On the other hand, the values of protein oxidation (PC) in the GT sole were not affected by the thermal chronic expose, which may indicate that a high content of HSP70, as a response to an acute stress, might have a long-term effect in the protection of proteins to oxidative stress. Similar adjustments of oxidative status and adaptive responses of antioxidant defences to stressful events were also reported in other sh species fed with green tea extracts. The antioxidant role of the green tea was reported to remedy the toxic effects of 4-nonylphenol (4-NP) in African cat sh Clarias gariepinus 49 and the toxic effect of oxidised sh oil in sturgeon 46 . In both studies, sh were able to decrease LPO content and increase the levels of some antioxidant biomarkers.
Overall, sh fed GS diet, after an acute thermal stress exposure, did not decrease the activity of GST and GSH, and were even able to decrease the content of PC, however, after a week of exposure to increased water temperature the PC content returned to higher values compared to sh from the other treatments. In sole fed GT, after 24 h of increased water temperature, the sh maintained the values of GSH and increased the content of HSP70, which suggests that after a chronic exposure these sh were able to protect protein oxidation which is supported by the decrease of PC content. The CC treatment did not show any signi cant improvement in the oxidative status of the sh after an acute/chronic increased of the water temperature.
In conclusion, the dietary supplementation of natural antioxidants was an effective way of modulating the oxidative status, antioxidant response and growth of Senegalese sole juveniles. These antioxidants are known to interfere in numerous routes by acting as direct antioxidant as chelating metals, reducing via electron transfers or hydrogen atoms transference, and indirectly by upregulating the expression of endogenous antioxidants. The CC and GS diets improved growth of sole postlarvae, although the mechanism by which these supplements act seem different. It seems that curcumin extract (CC diet) improved the oxidative status of the sole, showing low levels of protein and lipid degradation and a reduction on GST and HSP70 activity, allowing sh to invest more energy into growth rather than on endogenous oxidative defences. The GS diet, on the other hand, showed a negative impact in the oxidative status of the juveniles by increasing the protein damage, so the positive effect on growth may be reverted in the long-term. The mechanisms by which this antioxidant acts still requires further investigation. The inclusion of green tea extract (GT diet) did not improve growth performance nor the sole antioxidant status during the growth trial. Both green tea and grape seed extracts seem to act as pro-oxidants, by increasing the oxidative damage in sole juveniles, under standard rearing conditions. However, in response to a thermal stress, both GT and GS diets increased the sh capacity to cope with the new stressful event. However, these antioxidants appear to act in distinct time manners and exert different responses: Grape seed extract seems to have an immediate action and result in a shortterm improvement on sh oxidative status, while green tea extract has a long-term effect on the antioxidant capacity of the sole. Hence, curcumin extract seems a good candidate for long-term supplementation of young sh diets, as it improves welfare and growth of Senegalese sole. The use of green tea and grape seed extracts in diets for young sh still requires further evaluation to identify the most adequate inclusion level, although the short-term use of the tested doses seems a feasible solution before highly stressful short periods (e.g., transportation, handling and temperature rises). These results suggest that dietary natural plant extracts can improve young sh robustness and even promote growth when supplemented at the optimal doses.

Ethical statement
Animal handling and subsequent procedures complied with European laws (2010/63/EU) and Portuguese legislation for the use of laboratory animals (DL nº113/2013, 7 August). This study was approved by the ORBEA Animal Welfare Committee of CCMAR. This experiment was performed by trained scientists

Experimental diets
Four diets were tested in this study, including a commercial diet (WINFlat, SPAROS Lda., Portugal) that was used as the control (CTRL diet). This diet contains ingredients such as krill meal, squid meal, wheat gluten, sh meal, shrimp meal, sh hydrolysate, pea protein concentrate, sh gelatine, sh oil, lecithin and a micronutrient premix comprising vitamins, minerals, and other additives. Moreover, three experimental diets were prepared by supplementing the CTRL diet with an extract of either curcumin (CC diet) at 46 g/kg of the micronutrient premix, green tea (GT diet) at 12g/kg of the micronutrient premix or grape seed (GS diet) at 12 g/kg of the micronutrient premix. These selected doses of each antioxidant extract are under a patent pending application (PCT/IB2020/056001) and were chosen based on preliminary trials conducted in the Centre of Marine Science of Algarve (CCMAR, unpublished data). All diets were prepared by  (Table 1), and they only changed in the supplementation with extracts, and this supplementation did not exceed 1% of the diets.

Senegalese sole husbandry and experimental set-up
Senegalese sole postlarvae were reared for 25 days, starting at 45 days after hatching (DAH), in a recirculation aquaculture system at CCMAR (Portugal), under optimized environmental and zootechnical conditions. Sole postlarvae were kept in at-bottom tanks (21 L), each tank stocking 630 individuals (corresponding to a 3,000 ind m -2 ). The dietary treatments (CTRL, CC, GT and GS) were randomly assigned to replicate tanks (n=3 tanks per treatment). Abiotic parameters were measured, and mortality was recorded daily. Dead sh were removed, and the rearing units were carefully cleaned with minimal disturbance.
Dissolved oxygen in water was maintained at 96.6 ± 7.2% of saturation, temperature at 19.6 ± 0.5 °C and salinity at 35.4 ± 0.7 g.L -1 . A 10:14 h light:dark photoperiod was maintained, and the light intensity was 400 lx at the water surface. Inert diet was delivered semi-continuously with automatic feeders. The amount of feed distributed to each tank was based on predicted maximum growth and daily adjustments were done based on visual inspection to avoid a large excess of uneaten food 50 .

Thermal stress test
A thermal stress test was conducted at the end of the experimental period (70 DAH) to assess how the dietary treatments modify the animal physiological responses. Two challenging periods were analysed in sh response: an acute stress (at 72 DAH) and a chronic stress (at 78 DAH). The seawater temperature of the rearing system was raised from 19.6 ºC until 24.0 ± 0.5 ºC (~5 ºC over the experimental temperature) during a 24h period. For the acute stress challenge sh were sampled after remaining at this temperature during 24h (72 DAH), whist for the chronic stress test sh were maintained at this temperature for one week (78 DAH). At the end of each thermal stress sh from the different dietary treatments were sample for biomarkers response assessment.

Key performance indicators
At the beginning (45 DAH, n = 60) and at the end of the experiment (70 DAH, n = 120) sh were killed by using an overdose of anaesthetic 2-phenoxyethanol (1000 ppm; Prolabo) then individually collected for dry weight (DW, mg) and body standard length (SL, mm) determination. These juveniles were frozen at -80ºC, photographed for measuring SL using Axio Vision L.E. 4.8.2.0 (Carl Zeiss Micro Imaging GmbH) and freeze-dried for DW determination (Denver Instrument, 0.001 mg precision). Survival rate (%) was calculated as the percentage of sh counted at the end of the trial relative to their initial number in each replicate. Growth, expressed as relative growth rate (RGR, %/day), was calculated, at the end of the experiment, using the formula: (e g−1 )×100, with g=[(ln nal weight−ln initial weight)/time] 51 . The condition factor (K) was calculated as ( sh wet weight/total length 3 ) x 100.

Preparation of diet extracts for antioxidant capacity assessment
Methanol extracts were prepared from the four diets CTRL, CC, GT and GS. For that purpose, the diets were freeze-dried, mixed with methanol (1:40, w/v) and maintained in an ultrasonic bath for 30 min. Then, samples were extracted overnight, with stirring, at room temperature (RT, approximately 20 ºC) 52 . The extracts were then ltered (Whatman no 4) to remove solid debris, and methanol was removed by using a rotary evaporator (60 ºC; 337 mbar). The obtained dried extracts were weighed, dissolved in methanol at 50 mg/mL and stored at -20 °C. 4.6.1. 2,2-Diphenyl-1-picrylhydrazyl (DPPH) and 2,2′-Azino-bis-(3-Ethylbenzothiazoline-6-Sulfonic Acid) (ABTS) Methanol extracts from diet samples were tested for radical scavenging activity against the DPPH and ABTS radicals at concentrations of 50 mg/mL, as described previously 52 . Ascorbic acid was used as a positive control at the same concentrations of the samples. Results were expressed as a percentage of inhibition, relative to a control containing methanol in place of the sample.

Total phenolic (TPC) and Flavonoids (TFC) content
The TPC and TFC were determined in the methanol extracts at the concentration of 50 mg/mL and absorbance was measured in a microplate reader (Biotek Synergy 4). The TPC was assessed by the Folin-Ciocalteu assay and TFC was estimated by the aluminium chloride colorimetric method adapted to 96-well microplates. Results were expressed respectively as gallic acid equivalents (GAE) and quercetin equivalents (QE) in milligrams per gram of diet (dry weight, DW). All methods were performed as previous described 52 .

Antioxidant biomarkers assessment
Protein concentration of PMS was determined according to the Bradford method 53 , using bovine γ-globulin as a standard. Catalase (CAT) activity was determined in PMS by measuring the decomposition of the substrate H 2 O 2 at 240 nm 54 . Glutathione-S-transferase (GST) activity was determined following the conjugation of GSH with 1-chloro-2,4-dinitrobenzene (CDNB) at 340 nm 55 . Total glutathione (GSH) content was determined at 412 nm using a recycling reaction of reduced glutathione (GSH) with 5,5'-dithiobis-(2-nitrobenzoic acid) (DTNB) in the presence of glutathione reductase (GR) excess 56,57 . GSH content was calculated as the rate of TNB 2formation with an extinction coe cient of DTNB chromophore formed, ε = 14.1 x 103M -1 cm -157, 58 . LPO was determined by measuring thiobarbituric acid-reactive substances (TBARS) at 535 nm 59 . Protein carbonylation was measured by the quanti cation of carbonyl groups based on the reaction of 2,4-dinitrophenylhydrazine (DNPH) with carbonyl groups, according to the DNPH alkaline method 60  were removed by washing the microplates again, which were then incubate for 2 h at RT with 1 μg ml −1 of the secondary antibody, anti-mouse IgC (2º Antimouse IgC (fab speci c) Sigma). After another wash, the substrate p-nitrophenyl phosphate was added and incubated for 30 min at RT. Then, the stop solution (3 mol l −1 NaOH) was added to each well and the absorbance was read at 405 nm, using as a standard the puri ed HSP70 active protein (HSP70 protein Millipore).

Reverse transcription-quantitative real-time PCR (qPCR)
Gene expression analysis was performed in 70, 72 and 78 DAH soles from each dietary treatment. Fish were kept at -80ºC until analysis. A region from the operculum cavity until the end of the visceral cavity was select in each larva (n=4 per dietary treatment). This process was realised in each larva without thawing. Selected genes for oxidative stress defences and cellular stress proteins are described in detail in Table 2.
Samples were homogenised using a Fast-prep FG120 instrument (Bio101) and Lysing Matrix D (Q-Bio-Gene) with 1 ml Tri Reagent (Sigma-Aldrich) for 60s at speed setting 6. Chloroform (0.2 ml) was added to each sample before centrifuging at 14.000 rpm for 15 min. The supernatant content was transferred to columns of the Isolate II RNA Mini Kit (Bioline) and total RNA was treated twice for 30 min with DNase I following the manufacturer's protocols. Total RNA quality was checked by agarose gel electrophoresis and a Nanodrop ND-8000 (Thermo Scienti c) was used to determine its concentration. One μg of total RNA was reverse-transcribed using the iScript TM cDNA Synthesis kit (Bio-Rad) according to the manufacturer's protocol.
The qPCR assays were performed in duplicate in a 10 μL volume containing cDNA generated from 10 ng of the original RNA template, 300 nM of each speci c forward and reverse primers, and 10 µl of iQ™ SYBR ® Green Supermix (Bio-Rad). The genes analysed were involved in the oxidative stress defences: catalase (cat), glutathione peroxidase 1 (gpx1), glutathione peroxidase 3 (gpx3), superoxide dismutase [Cu-Zn] (sod3)) and cellular stress proteins (heat shock protein 70 (hsp70), heat shock protein 90 alpha (hsp90aa), heat shock protein 90 beta (hsp90ab). Primers for Senegalese sole hsp70, hsp90aa and hsp90ab were previously published 63,64 and new species-speci c primers for qPCR were designed for remaining genes ( Table 2). The qPCR ampli cation protocol was as follows: 7 min for denaturation and enzyme activation at 95 ºC followed by 40 cycles of 30 s at 95 ºC and 1 min at 60 ºC. Expression data were normalised using the geometric mean of two reference genes, ubiquitin (ubi) and glyceraldehyde-3-phosphate dehydrogenase 2 (gadph2) 65 and the relative mRNA expression calculated using the comparative Ct method 66 .

Data analyses
All data were tested for normality using a Kolmogorov-Smirnov (whenever n > 30) or Shapiro-Wilk (whenever n < 30) test and for homogeneity of variance using a Levene's test using STATISTICA v13 (StatSoft). Data were log transformed when required and percentages were arcsin transformed prior analysis.
Comparisons between groups fed different diets were made using one-way ANOVA followed by a Tukey post-hoc test for growth performance and oxidative stress biomarkers at the end of the growth trail. To assess the response of oxidative stress biomarkers to thermal stress exposure by each treatment group a two-way ANOVA was made and the analysis of the delta variation between pre-and post-stress indicators was performed by a one-way ANOVA, both followed by a Tukey post-hoc test. Signi cance levels were set at P < 0.05.

Declarations
Data availability Tables   Table 1  Proximate     Values are presented means ± SD. Different superscript letters indicate signi cant differences between the dietary treatments. Table 6 The response of several oxidative stress-related biomarkers of Senegalese sole fed with different diets at the end of the growth trial (standard) and after therm (acute Different subscription letters indicate statistical differences in stress (p, q), treatment (x, y, z) and interaction of treatment and stress (a, b, c, d, e) in the expression of antioxidant capacity biomarker.  Delta variation (%) of the response of oxidative stress-related biomarkers from standard to acute stress conditions. Values are presented means ± SD.