The high cost of larvae production in hatcheries limits the development of marine fish aquaculture. Because there is no formulated diet suitable for marine fish larvae, they are fed live prey that are expensive to rear. Indeed, there is no formulated diet available that can be substituted for live prey during larval stages. Over the last two decades, several studies have been conducted to determine the nutritional requirements of marine fish larvae (for review see Watanabe and Kiron 1994). Determining only the optimal level and the nature of dietary lipids and proteins has proved insufficient for formulating a compound diet which is as effective as live prey for rearing larvae. Protein is the major diet component, and the amino acid requirement of fish larvae is met by diets containing fish meal (Kanazawa et al. 1989). However, the molecular size of the dietary protein fraction could play a major role in larval development. Indeed, the incorporation of casein hydrolysate in the diet led to increased survival of Carassius auratus (Szlaminska et al. 1991) and Dicentrarchus labrax (Cahu and Zambonino Infante 1995a), but no effect on growth was reported. On the other hand, Berge et al. (1993) observed only a slight growth improvement of salmon fry fed Concentré de Protéines Solubles de Poissons (CPSP). Even if no clear effect on larval growth has been reported in the literature, protein hydrolysate has long been supposed to be advantageous for larvae (Gabaudan et al. 1980). This product is incorporated into most larval diets, both for improving physical properties (Pigott et al. 1982) and nutritional value (Carvalho et al. 1995) of the diet.

Recent data demonstrating that the increased survival of larvae fed hydrolysate was paralleled by the enhanced development of certain digestive functions (Cahu and Zambonino Infante 1995b) have aroused a new interest in this field of investigation. In particular, substitution of casein hydrolysate for part of the fish meal induced an earlier and greater rise in enzyme activity of brush border membranes, though native casein is of lower nutritional value than native fish meal. It was concluded that the presence of hydrolysate in the diet is essential for larval development. Moreover, hydrolysate containing short peptides has been shown to be effective in stimulating enzyme activity in brush border membranes and in facilitating nutritional rehabilitation in mammals (Sasaki et al. 1989, Scheppach et al. 1994). These short peptides, particularly di- and tripeptides, are absorbed quickly and efficiently by the intestine without any prior pancreatic digestion.

Taking these data into account, we hypothesized that the incorporation in larval diet of a hydrolysate processed from a high quality protein and containing a high proportion of short peptides may be beneficial. The aim of this study was to test the effect of a fish meal hydrolysate characterized by 75% di- and tripeptides on the growth and survival of sea bass larvae, and to verify whether di- and tripeptides influenced the activity of pancreatic proteases and the development of intestinal enzymes.

MATERIALS AND METHODS

1

1

1

1

2

The larvae were fed live prey from mouth-opening until d 19 in the following sequence, expressed per larva per day: d 6 to d 9, 50-200 Brachionus plicatilis; d 10 to d 12, 200 Brachionus plicatilis and 30-60 Artemia nauplii; d 13 to d 19, 90-200 1-d-old Artemia. Then the larvae were divided into 3 groups (5 tanks per group) and fed for 21 d one of three experimental diets. These diets were formulated to be isoenergetic and isonitrogenous using fish meal as the protein source or fish meal with fish meal hydrolysate at a level of 20 and 40% of total nitrogen (designated as P0, P20 and P40, respectively) (Table 1). The fish meal hydrolysate was obtained from and processed by the Institut Univeritaire de Technologie de Limoges, using an alkaline bacterial serine proteinase, according to Bressollier et al. (1988). Resulting peptides were continuously extracted using a membrane ultrafiltration process (molecular weight cut off: 1000). Then, each peptide class, characterized by it size, was quantified on the basis of its amino acid content after separation by ligand exchange chromatography using Cu(II)-modified silica gel. The relative distribution was (mol/100 mol): single amino acids, 5; di- and tripeptides, 75; peptides with chain length < 6 residues, 20. The size of the microparticulate diets was 200-400 µm. Fish were continuously fed in large excess to ensure a constant and high level of suspended microparticles in the water column. Food was distributed 18 h/d using a belt feeder. Food ingestion was monitored by observing digestive tracts of larvae under a binocular microscope.

Table 1. Composition of the experimental diets [View Table]

On d 26 and d 40, before morning food distribution, 50 larvae were collected from each tank. They were immediately stored at 80°C pending dissection and assays. Dissection under microscope was conducted on a glass maintained at 0°C. Individuals were cut into four parts as described by Cahu and Zambonino Infante (1994); head, pancreatic segment, intestinal segment and tail, in order to limit the assay of enzymes to specific segments. This dissection inevitably produced a crude mixture of organs in each segment. The pancreatic segment, besides pancreas, contained liver, heart, muscle and spine. Intestinal segment contained intestine, muscle and spine.

1

RESULTS

The results of statistical analyses of enzyme activities are presented in Table 2. The two proteolytic enzymes assayed in the pancreatic segment were differently affected by the presence of dietary peptides (Fig. 4). Trypsin activity (panel A) was greater in the group fed only native proteins than in groups fed diets containing peptides, whereas chymotrypsin activity (panel B) was enhanced by peptides, but this effect appeared clearly only at d 26. Trypsin activity decreased between d 26 and d 40 when there was a tendancy for chymotrypsin activity to increase.

Table 2. Summary of two-way ANOVA of specific activities of some pancreatic and intestinal enzymes [View Table]

Aminopeptidase, alkaline phosphatase and maltase were affected by diet. The presence of peptides in diet induced a lower activity of aminopeptidase (P < 0.005; Table 2), compared to the P0 group (Fig. 5A). This effect was more pronounced at the high peptide level. Alkaline phosphatase and maltase were only influenced by the dose of dietary peptides (Fig. 5B, C), significant differences being observed between enzymatic activities of P20 and P40 groups (P < 0.001; Table 2). Specific activity of aminopeptidase increased slightly but significantly with age (Table 2); a threefold enhancement was observed in specific activities of maltase between d 26 and d 40. The specific activity of alkaline phosphatase did not change with age (Table 2). The activity of -glutamyl transpeptidase was not detectable at d 26 (Fig. 6). At d 40, the one-way ANOVA showed that this enzymatic activity was higher in the P20 group than in the group fed native proteins, and did not significantly differ in the 2 groups fed short peptides.

Diet composition slightly affected the activity of leu-ala peptidase (Table 2; Fig. 7). Leu-ala activity exhibited a sharp decrease between d 26 and d 40 (P < 0.001). The activity level of this enzyme was modulated much more by the age of larvae than by the dietary peptide level.

Segmental activity ratios of the four brush border enzymes vs. leu-ala are reported in Table 3. The ratios calculated for d 26 were higher in the P20 group than in the other two groups for aminopeptidase, alkaline phosphatase and maltase. On the other hand, lowest ratios were observed in the P40 group for aminopeptidase and alkaline phosphatase. No difference was shown in the maltase to leu-ala peptidase ratio between the P40 and P0 groups. At d 40, the ratios of the three enzymes to leu-ala peptidase were not significantly different among the dietary groups. The GT to leu-ala peptidase ratio was significantly higher in P40 group than in the other two dietary groups.

Table 3. Segmental activity ratios of some brush border enzymes versus leucine-alanine peptidase in 26- and 40-d-old sea bass larvae fed diets in which fish meal proteins were substituted by peptides at a level of 0% (P0), 20% (P20) and 40% (P40)1 [View Table]

DISCUSSION

Secondly, di- and tripeptide diets enhanced larval survival; this effect was maximum for a moderate dose of peptides, i.e. 12 g/100 g of total ingredients. It is noteworthy that peptide chains of two or three amino acids induce improvement in survival as do the commercial casein hydrolysates (Szlaminska et al. 1991) which are mainly composed of peptide chains of 10 to 20 amino acids. On the other hand, free amino acids incorporated in diets failed to enhance larval survival (Cahu and Zambonino Infante 1995a).

Finally, the impressive reduction in skeletal malformation observed in groups fed peptides was an unexpected finding. Indeed, it is known, although rarely reported, that marine fish larvae fed formulated diets are affected by some malformations. This study showed, for the first time, that the molecular form of dietary nitrogen fraction can influence the malformation rate.

We are faced with the question as to why protein hydrolysates appear to be beneficial for fish larvae but do not affect or in some cases depress juvenile growth. The differences in digestive physiology between larvae and juveniles could partially explain this paradox. Indeed, the relative length of the intestine in larvae is short compared to juveniles (Segner et al. 1994). This can be compared to partially intestinectomized mammals, for which short peptides represent the best nitrogen supply (Cosnes et al. 1992), while these peptides adversely affect the nitrogen balance in healthy subjects (Grimble et al. 1987). Moreover, digestion in marine fish larvae shows some specificities compared to juveniles: the synthesis of some pancreatic enzymes during larval stages is quite different from that observed in juveniles (Dabrowski 1984), and in intestine, larvae exhibited poorly differentiated brush border membranes and a high level of cytosolic digestion (Gawlicka et al. 1995).

In our experiment, we showed that the activities of two pancreatic enzymes, trypsin and chymotrypsin, were strongly linked to the age of larvae. Moreover, the activities of the two enzymes were modulated by the dietary protein content. Trypsin activity was enhanced by the native protein, whereas chymotrypsin activity was enhanced by the diets containing di- and tripeptides. Similarly, chymotrypsin activity in rat was enhanced to a greater extent than trypsin by a modification of the dietary protein, as reported by Lhoste et al. (1994). We assume that the total proteolytic capacity of pancreas of young larvae was enhanced by the incorporation of short peptides in diet. The better growth rate observed in groups fed P20 and P40 could be in part the result of a greater proteolytic capacity of the pancreas. The relationship between elevated proteolitic activity of pancreas and improved growth of pigs has already been suggested by Owsley et al. (1986).

During our experiment, the decrease in the activity of leu-ala peptidase, an enzyme mainly located in the cytosol, was observed concurrent with the increase (or onset) of some brush border enzymes of the enterocytes, aminopeptidase, maltase or GT. The decline of cytosolic digestion coinciding with the rise of membranous digestion illustrates the normal development of intestinal digestion processes, as described in mammals by Henning (1987) and more recently in fish by Cahu and Zambonino Infante (1995a).

The protein fraction of the diet influenced the activities of aminopeptidase and GT. Di- and tripeptides are absorbed into the enterocytes without any hydrolysis by microvillous peptidases. GT, which is involved in peptide transport (Griffith and Meister 1980), was stimulated in fish fed diets containing peptides rather than native protein. On the contrary, aminopeptidase, for which only a hydrolytic function has been described, had lower activity in groups fed diets containing peptides compared to those fed native protein. These findings are in agreement with the observation of Rouanet et al. (1990) who reported a greater stimulation of aminopeptidase by native protein rather than small peptides and, inversely, a stimulation of GT by small peptides in rats fed isonitrogenous diets. Hydrolases that were not involved in peptide digestion, maltase and alkaline phosphatase, were not affected by the diets.

Fish larvae undergo their ontogenesis, and particularly development of digestive function, during the first two months of life. As it has been shown extensively in mammal pups (Henning 1987), the changes in hydrolytic activities, which are genetically determined, can in part be influenced by the diet in fish larvae (Cahu and Zambonino 1995b). The level of development of intestinal digestion can be evaluated by considering the segmental activity ratio of brush border enzymes vs. leu-ala peptidase, which reflects the relative importance of the brush border membrane digestion compared to intracellular digestion (Cahu and Zambonino Infante 1995a). At d 26, the highest ratios were obtained for all the enzymes in the P20 group, revealing an earlier maturation of the enterocytes in this group compared to the others. This high ratio resulted from a sharp increase in activities of brush border enzyme. At d 40, ratios revealed a similar intestinal maturation for the three dietary groups. The high of GT:leu-ala peptidase ratio in the P40 group did not reveal a better intestinal maturation, but was the consequence of a stimulation of GT by short peptides as previously discussed.

In conclusion, this study shows that the incorporation of short peptides in the diet improves the development and the nutritional status of sea bass larvae, by stimulating the acquisition of the adult mode of digestion.