Lipids are essential molecules in larval physiology supplying not only the energy required for growth but also participating in membrane formation and synthesis of signaling molecules. Triacylglycerol (TAG) is the main lipid class responsible for energy provision and fatty acid (FA) delivery to the cells (Mourente and Vazquez, 1996; Hamre et al., 2013; Hachero-Cruzado et al., 2014). As an energy supply molecule, dietary TAG levels are key effectors of larval development and tightly modulate FA and TAG biosynthesis pathways to provide sufficient energy for an optimal growth (Hachero-Cruzado et al., 2014). Other important lipid classes are the phospholipids (PL) with a major structural role in the assembly of cell membranes and tissue formation (Niu et al., 2008; Tocher et al., 2008; Saleh et al., 2013). Moreover, PL are an integral part of lipoproteins and play a key role in lipid transport (Chapman, 1980; Babin and Vernier, 1989) and occasionally act as a source of energy during embryogenesis and early larval development (Rainuzzo et al., 1992; Rønnestad et al., 1995; Mourente and Vazquez, 1996). Microdiets (MD) represent an experimental methodology suited for evaluation the effect of dietary compounds on fish larvae and avoid ambiguous results caused by the metabolic transformations that can occur if live preys are used as the vehicle (Tocher et al., 2008; Cahu et al., 2009; Hamre et al., 2013). This is especially relevant when the aim is to evaluate structural molecules such as PL and live preys tend to maintain their levels constant making difficult to establish a good gradient between the dietary treatments (Hachero-Cruzado et al., 2014; Roman-Padilla et al., 2017).

It is well-accepted that marine larvae require a supply of lipids from the diet for growth and development as lipid metabolism is still not fully-functional. Hence, a good understanding of the mechanisms involved in lipid transport and fat accumulation in the intestine under different lipid levels is required since these two factors are a constraint for nutrient trafficking in early fish stages (Fontagné et al., 1998; Morais et al., 2007; Gisbert et al., 2008). Several authors have reported that an optimal ratio between PL and neutral lipids is required to prevent the negative effects of lipid accumulation in the intestine and this requires more knowledge about the larval biosynthetic capacity and the endogenous capacity for lipoprotein formation and lipid transport that are highly dependent on intestine maturation and age (Deplano et al., 1991; Diaz et al., 1997; Gisbert et al., 2008; Carmona-Antoñanzas et al., 2015). In a previous study using Senegalese sole larvae we demonstrated that early larval stages modified cell regulatory pathways as a function of dietary TAG levels to mitigate excess lipid accumulation in the intestine (Hachero-Cruzado et al., 2014; Roman-Padilla et al., 2017). These experiments were carried out using dietary PL levels higher than those recommended for fish larvae (Tocher et al., 2008). Thus, new research is required in sole to understand the physiology of intestinal lipid transport under suboptimal dietary PL and to identify compensatory responses triggered by the larvae when the PL/TAG ratio is modified.

Phospholipid biosynthesis in fish larvae is highly dependent on intestine maturity and liver function. Diaz et al. (2002) reported that early fish larvae use biliary lipids to synthesize the first set of lipoproteins. In mammals, two phosphatidylcholine (PC) synthetic pathways are present in the intestine, (i) the remodeling pathway by acylation of dietary lysoPC (LPCAT) and (ii) de novo synthesis by CDP-choline-dependent (CPT) pathway. Moreover, the phosphatidylethanolamine N-methyltransferase (PEMT) pathway appears to be liver-specific (Mansbach and Gorelick, 2007). In fish, all the enzymes for the endogenous production of PC are present in the liver and intestine. However, the activity of the PEMT pathway in vivo using animals has not been tested (Oxley et al., 2007; Carmona-Antoñanzas et al., 2015). Interestingly, expression levels of the main enzymes involved in PL biosynthesis were lower in the intestine of salmon fry compared to par, supporting the idea that the developmental status of the intestine determines lipid transport. In this study, the effects on growth and the physiological and expression responses that occur in early larval stages of the flatfish Senegalese sole fed a MD with a low PL/TAG ratio with suboptimal total PL content compared with larvae fed with a control MD (high PL levels) were investigated. In a first trial, long-term effects of larvae fed with both MD on growth and survival were determined and results were validated using a diet based on live prey (rotifer). Since clear differences in growth associated with MD were found a second trial was performed to validate growth results and identify the specific physiological and gene expression responses associated with each MD. Changes in lipid classes, FA profile, histological characterization of the intestine and liver and gene expression patterns were determined. This multidisciplinary approach provided new clues about the role of PL levels in early larval stages of marine fish and the compensatory mechanisms triggered in the larvae in response to dietary PL levels.

Marine larvae require high levels of dietary PL for an optimal growth and survival (Cahu et al., 2009; Daprà et al., 2011; Azarm et al., 2013; Saleh et al., 2013; Zhao et al., 2013; De Santis et al., 2015; Taylor et al., 2015; Cai et al., 2017; Feng et al., 2017). Studies of PL requirements in fish larvae have been limited by the fact that live prey (LP) do not accumulate these structural compounds making it difficult to establish their nutritional requirements. The use of MD represents an attractive approach to overcome this limitation since their formulation can be easily modified and pelagic sole larvae accept these inert diets from mouth opening. Currently, main protocols adopted for sole larval rearing are based on MD and live prey co-feeding from mouth-opening until weaning. In contrast, MD supplied alone for a long term have low survival rates (Engrola et al., 2007, 2009; Pinto et al., 2018). In this study, we investigated how different dietary PL levels affect pelagic larval stages of sole, a short period of high energy demand due to the imminent metamorphosis (Yufera et al., 1999; Cañavate et al., 2006). The main effects of PL levels were evaluated at 6 and 8 dph to avoid interference from yolk sac reserves, which were observed to be almost fully depleted by 3 dph as previously reported (Roman-Padilla et al., 2016a, b, 2018). As expected, larvae supplied with live prey grew faster than those fed with MD. However, survival remained unaffected indicating that MD could be useful to investigate the effect of micronutrients when supplied for short-periods in pelagic sole larvae as previously suggested (Gamboa-Delgado et al., 2011; Jimenez-Fernandez et al., 2015). Sole larvae require an adaptation period to inert diets that is inversely proportional to size and hence co-feeding protocols appear as the most adequate strategy in long-term procedures and for the industry (Engrola et al., 2007; Pinto et al., 2018). However, the use of MD for short periods provides a suitable experimental window to investigate nutritional requirements and larval physiology in this species.

Larvae fed with High-PL grew faster than low-PL MD in our two independent trials demonstrating the importance of PL for larval growth. Interestingly, the replacement of MD by Artemia after 8 dph triggered compensatory growth and by the end of the trial there were no differences in survival or mass compared to a reference diet based on LP indicating physiological compensation of the larvae and the full capacity of development when optimal conditions were restored. The shift to an optimal diet enhanced energy accumulation so that the Senegalese sole underwent successful metamorphosis and recovered mass and were no different from the LP diet at the end of the trial. Jimenez-Fernandez et al. (2015) observed similar effects on growth and development during the period that we supplied MD in our trial. Overall, our data demonstrate that pelagic sole larvae can deal with suboptimal levels of dietary PL by reducing transiently growth but recovering full performance if a shift to an optimal composition occurs. This further supports the suitability of our approach to evaluate the physiological effect of dietary treatments.

One major effect of PL dietary treatments was the presence of vacuoles in the intestine. According to our previous study, using oil red staining of cryostat sections, larvae tend to accumulate neutral lipids in the anterior intestine within large vacuoles (Hachero-Cruzado et al., 2014). Others authors using haematoxylin and eosin staining have also reported intestinal vacuolization linked to neutral lipid accumulation (Ribeiro et al., 1999; Gisbert et al., 2005; Morais et al., 2006; Daprà et al., 2011). Lipid vacuoles were far more abundant in the intestine of larvae fed with Low-PL than High-PL in our study (18.9% in larvae fed LPL vs. 3.05% for HPL larvae) and this observation coincides with the higher TAG, CHO and cetoleic acid (CA, 22:1n-11) detected in LPL larvae, suggesting accumulation of dietary neutral lipids in the intestine. These findings may be explained by ineffective lipid mobilization and transport from the intestine to internal tissues when dietary PL/TAG ratio is low. The results reinforce the importance of balanced dietary PL for efficient lipid transport (Fontagné et al., 1998; Olsen et al., 1999; Salhi et al., 1999; Gisbert et al., 2005; Cahu et al., 2009; Carmona-Antoñanzas et al., 2015).

Gene expression data indicated a major effect of development from day 3 to 8 dph that is in alignment with the maturation of the digestive system and metabolism (Manchado et al., 2008; Gamboa-Delgado et al., 2011; Hachero-Cruzado et al., 2014). However, the effects of the dietary treatments were still observable at 8 dph with a wide set of apolipoproteins-encoding genes differentially regulated. A previous study in early larval stages of sole indicated that animals responded to dietary TAG levels under adequate levels of PL by activating the transcription of several apolipoproteins to facilitate lipid mobilization in the intestine (Hachero-Cruzado et al., 2014; Roman-Padilla et al., 2017). In this study, larvae fed with Low-PL had higher expression of the 8 genes encoding for main canonical apolipoproteins (types apoA-IV, apoB, apoC and apoE) than the High-PL group. This seems contradictory in the context of the accumulation of dietary lipids in vacuoles in the intestine of the Low-PL larvae. However, De Santis et al. (2015) reported that salmon fry fed a low PL diets also had up-regulated apoa4 transcription without modifying intestinal lipid transport. These authors hypothesized that in salmon fry dietary lipids were absorbed into the enterocytes but not exported to the blood stream and this was associated with transcription of genes involved in chylomicron formation. Our results suggest a similar response to Low-PL in the intestine of the sole larvae, with a coordinated transcriptional activation of apolipoproteins induced by dietary lipids unable to promote an effective intestinal lipid transport probably due to the limited amounts of PL and failed lipoprotein assembly. In this sense, it is noteworthy the down-regulation of apolipoprotein apob2 in larvae fed with Low-PL. This inverse pattern of regulation in relation to other apolipoproteins was not detected in our previous study (Hachero-Cruzado et al., 2014). Moreover, Daprà et al. (2011) found reduced apob expression in rainbow trout fry fed a PL-free diet compared to those fed PL-supplemented diets. However, they did not detect differences in expression for apolipoproteins apoa1 and apoa4. Thus, the PL levels seem to positively regulate apob2 expression in the intestine of sole larvae and this may be linked to chylomicron biosynthesis particularly since ApoB is an important structural component of chylomicrons (Iqbal and Hussain, 2009).

Together with the significance of PL regulating the synthesis and secretion of lipoproteins, it is assumed that fish larvae have a limited PL biosynthesis capacity and a major dependence on dietary supply (Tocher et al., 2008; Olsen et al., 2014; Carmona-Antoñanzas et al., 2015; De Santis et al., 2015; Feng et al., 2017). Results of the present study indicate that PL biosynthesis pathways were not fully developed at this early life stages. Sole larvae did not increase the transcription of key genes involved in de novo PL biosynthesis (ethanolamine phosphotransferase, ept; choline/ethanolamine phosphotransferase; cpt; phosphatidylethanolamine methyltransferase, pemt; or CDP-diacylglycerol inositol 3-phosphatidyltransferase, cdipt), despite differences in dietary PL ingested (0.73 vs. 4.97% DM). Only those transcripts involved in phosphatidylserine (PS) synthesis (CDP-diacylglycerol synthase (cds2) and transmembrane protein 269 (tmem269) responded to low dietary PL level. PS is a quantitatively minor PL, and constitute approximately 2–10% of total PL, differing between tissues (Vance and Steenbergen, 2005). About PS, it is interesting to note the clear dependence of gene expression of both transcripts involved in PS synthesis (cds2 and ptdss1a) on sole larvae development (see Supplementary Table S2). So, we cannot discard that larval development was related with differences detected between dietary treatments. About the most pathways involved in PL synthesis, our results match with those reported for other fish species (Daprà et al., 2011; De Santis et al., 2015; Feng et al., 2017), which neither showed substantial changes in larval PL biosynthesis pathway with dietary PL level. However, we found changes in expression levels of other genes indirectly related with the synthesis of intermediate products of PL biosynthesis pathways (phosphatidic acid and DAG; gpat3, agpat1, pnpla2, lipg and mogat2) or with PL remodeling pathway (pla2g12a, pla2g12b1, pla2g12b2, pla2g3, and lpcat1). Phosphatidic acid is the precursor for CDP-DAG that is used in the synthesis of phosphatidylinositol (PI) and PS, and is dephosphorylated to diacylglycerol (DAG) for the synthesis of PC and phosphatidylethanolamine (PE) as well as TAG. Regarding PL remodeling, this pathway allows altered FA composition of individual PL by FA hydrolysis and re-esterification reactions. These combinations and permutations of FA attached to the sn-1 and sn-2 positions result in thousands of different PL species (Ridgway, 2016). These findings indicate that early sole larvae modulate the synthesis of PL intermediates and activate PL remodeling pathways when dietary PL levels are suboptimal but they fail in the regulation of key genes involved in de novo PL biosynthesis.

In summary, the present study demonstrates that dietary PL levels play a key role in growth of pelagic larvae in sole and they can deal with different dietary PL levels triggering an adaptive response to mobilize dietary lipids. While a low supply of PL delays growth, the larvae maintain viability and competence to trigger a compensatory growth when a shift in the diet occurs without affecting metamorphosis success or survival rates. The lipid accumulation in the intestine and the lack of efficient lipid mobilization mechanisms in spite of a coordinated activation of the gene expression related to PL intermediate synthesis, PL remodeling as well as several apolipoprotein transcripts are mainly responsible for the negative effects on growth. The lack of full functionality for PL biosynthesis pathways limits the lipid dietary uptake in these early sole larvae. These results are of relevance to optimize nutrition in sole larvae and provide a new marker (apob2) for monitoring specific responses to dietary PL.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by Bioethics and Animal Welfare Committee of IFAPA. Written informed consent was obtained from the owners for the participation of their animals in this study.

IH-C and MM conceived and designed the experiments and wrote the manuscript. IT, AR-R, IH-C, and JR-P performed the experiments. IH-C, MM, and AR-R analyzed the data. All authors contributed to reagents, materials, and analysis tools. AR-R, IT, and JR-P contributed to additional writing and discussion.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.