Studies on the requirements of heavier rainbow trout for histidine, valine, leucine and isoleucine have not been published to date. Values reported for the lysine requirement vary from 13.0 to 28.7 g/kg diet (Ketola 1983, Kim et al. 1992, Lanari et al. 1991, Pfeffer et al. 1992, Walton et al. 1984b) and for tryptophan from 2.0 to 2.5 g/kg (Kim et al. 1987, Poston and Rumsey 1983, Walton et al. 1984a). Several factors such as size of the fish, dietary energy concentration, dietary ingredients used, feeding regimen, growth rate achieved, response criteria and mathematical model used may influence recommendations derived from dose-response experiments and thus make it difficult to compare or combine results from different laboratories. Thus far, nothing is known about the amount of essential amino acids required to cover maintenance requirements.

To overcome this lack of consistent requirement estimates, we developed a high energy, semipurified diet that fosters a high growth rate in rainbow trout (Rodehutscord et al. 1995c). This diet can be used for dose-response studies on requirements in trout; it has been used for experiments on requirements for almost all essential amino acids, with the exception of phenylalanine, in an attempt to standardize experimental conditions as far as possible. Results for methionine and cystine, arginine, and threonine have already been published (Rodehutscord et al. 1995a and 1995b). In this paper, results of the dose-response experiments on lysine, tryptophan, histidine, valine, leucine and isoleucine are presented.

Table 3. Parameters estimated by fitting the experimental data to an exponential curve1 [View Table]

Based on the results presented here in earlier papers, an attempt to estimate the maintenance requirements for individual amino acids based on protein retention data is made.

MATERIALS AND METHODS

Table 1. Composition of the basal diets used in six dose-response experiments with rainbow trout [View Table]

Table 2. Number of diets, range in concentration of individual amino acids, initial body mass of fish, experimental period and water temperature in six experiments with rainbow trout [View Table]

In each experiment, graded levels of the amino acid under test were added to the respective basal diet. A total of 24 diets were used in Experiment Lysine and twelve in each of the other experiments. L-Isomers of the amino acids replaced the corresponding amount of the mixture of nonessential amino acids (NEAA) in Experiment Lysine and of L-glutamic acid in each of the other experiments. The following concentrations of the amino acid under test were used in the experiments (in g/kg dry matter): lysine 4.5, 5.5, 7.0, 8.5, 10.0, 11.5, 13.0, 14.5, 16.0, 17.5, 19.0, 22.0, 25.0, 28.0, 31.0, 34.0, 37.0, 40.0, 43.0, 46.0, 49.0, 52.0, 55.0 and 58.0; tryptophan 1.3, 1.5, 1.8, 2.0, 2.3, 2.6, 2.8, 3.1, 3.7, 4.3, 5.0 and 5.6; histidine 2.6, 3.3, 4.1, 4.9, 5.7, 6.5, 7.5, 8.5, 9.5, 10.5, 12.0 and 13.5; valine 6.2, 7.2, 8.2, 10.2, 12.2, 14.2, 16.2, 18.2, 22.2, 26.2, 30.2 and 34.2; leucine 10.0, 11.0, 12.0, 14.0, 16.0, 18.0, 22.0, 26.0, 30.0, 34.0, 38.0 and 42.0; isoleucine 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 9.0, 10.0, 11.3, 13.3 and 15.3. Fish oil, sunflower oil and warmed beef tallow were stirred with the binder before being added to the other components in a drum mixer. Mixtures were moistened in a cutter to a sticky paste which was pelleted by forcing it through 3-mm screens of a mincer. The moist pellets were stored at 18°C until feeding.

For each experiment, trout (mean initial body mass given in Table 2) were taken from homogenous populations either reared in our department (Experiments Lysine, Histidine, Valine and Leucine) or obtained from a commercial fish farmer (Fischzucht Mohnen, Stollberg, Germany) 3 wk before starting the experiments (Experiments Tryptophan and Isoleucine). Up to the beginning of the experiments, all trout had been fed a commercial diet (Aminoforte, Rheinkrone, Wesel, Germany). Groups of 20 trout each were placed in 24 (Experiment Lysine) or 12 (all other experiments) 250-L round plastic tanks that were part of a circulatory system and continuously supplied with water in parallel, with ~70% of the outflowing water recirculated after clarification and aeration. The water supply to each tank was ~4.5 L/min. Water could be heated during winter but could not be cooled during summer; therefore, the mean water temperature ranged among experiments from 15.5 to 17.0°C (Table 2), but within each experiment only by a maximum of ±1°C. Experiments were performed in a room illuminated for 16 h/d.

Body masses of each group of fish were recorded at the beginning of the experiments after the fish were anesthetized in a solution of 1,1,1-trichloro-2-methyl-2-propanol-hemihydrate, and at the end of the experiments after the fish had been killed in 4-ethyl-aminobenzoate. Feed was withheld 36 h before killing. One additional representative group of 20 trout was killed for base-line measurements at the beginning of each experiment.

After being frozen at 18°C, all fish (experimental and base-line groups) were cut into small pieces with a ribbon saw, forced repeatedly through a mincer, homogenized in a cutter and freeze-dried. The concentrations of protein and lipids in the weight gain of experimental groups were calculated from differences between the experimental and base-line groups as described previously (Rodehutscord et al. 1995c).

In each experiment, each group was fed one of the experimental diets for a period given in Table 2. Trout were fed to near satiation twice daily during the week and once daily on weekends. Thawed feeds were offered by hand until pellets were first seen to sink to the bottom of the tank. Thus, feed losses could be avoided almost completely. General care, handling and maintenance of trout followed the procedures approved by the Animal Welfare Commissioner of the University of Bonn in accordance with the German Animal Welfare Law.

In diets and body homogenates, dry matter (105°C), ash (550°C), total N and lipids (petroleum ether extract after HCl treatment) were determined according to the official methods (Naumann and Bassler 1976). Amino acid content of all basal diets was determined by ion exchange chromatography (Llames and Fontaine 1994). Tryptophan was determined by HPLC after alkaline hydrolysis. Supplemented levels of crystalline amino acids were checked separately after extraction with diluted HCl. Similarly, added tryptophan was determined by HPLC after hydrolysis with barium hydroxide solution and autoclaving under vacuum (J. Fontaine, Degussa AG, Germany, personal communication).

Exponential functions were calculated to fit the experimental data because the response to a limiting dietary component is nonlinear. For traits that increased with increasing dietary concentration of the amino acid under test, the equation employed for describing the respective response was Lipid concentration in weight gain, on the contrary, responded negatively to increasing dietary amino acid concentration in most of the experiments. Therefore, the equation employed was where x = dietary concentration of the amino acid under test (g/kg dry matter), a = plateau value of the respective curve, b = parameter characterizing the steepness of the curve, c = dietary concentration of the amino acid under test at y = 0, and d = maximum response to supplemented amino acid.

Model parameters were estimated by the least-squares principle (Rawlings 1988, Seber and Wild 1989). The resulting nonlinear equations were solved iteratively by the Levenberg-Marquard method (Press et al. 1992). Calculations were performed using the program BFIT (H. P. Helfrich, Seminar of Mathematics, Faculty of Agriculture, University of Bonn), which implements this method.

RESULTS

Parameters of all estimated regression equations are summarized in Table 3. Additionally, responses of trout in weight gain, dry matter intake and in composition of the weight gain to increasing dietary concentration of individual amino acids are shown in Figures 1 (Lysine), 2 (Tryptophan), 3 (Histidin), 4 (Valine), 5 (Leucine) and 6 (Isoleucine). Dry matter intake, weight gain, gain per feed ratio and protein concentration of weight gain increased with increasing dietary concentrations of the respective amino acid under test in all experiments, and the rate of increase decreased with increasing dietary concentration. Concentration of lipids in weight gain, however, did not respond as uniformly to the increase in dietary concentrations of individual amino acids. In Experiments Lysine, Tryptophan, Leucine and Isoleucine, lipid concentration in weight gain decreased with increasing concentration of the respective amino acid (Fig. 1, 2, 5 and 6) whereas it increased with increasing dietary concentration of valine (Fig. 4). No clear dose-response relationship was detectable between dietary concentration of histidine and lipid concentration in weight gain (Fig. 3).

Concentrations of essential amino acids were determined in body homogenates of trout fed the diets containing either the lowest or the highest level of the respective amino acid in each experiment. Results are shown in Table 4.

Table 4. Essential amino acids in whole-body protein of trout fed the diets with the lowest or the highest level of individual supplemental amino acids, determined at the end of the experiments1 [View Table]

DISCUSSION

Each dietary treatment comprised only one group of 20 trout as one data point, and no repeated measurements were done for the individual dietary concentrations. This would be completely unsatisfactory if different individual treatments were to be compared. Evaluation of monofactorial dose-response experiments by regression analysis, however, always uses the complete set of data for deriving one function. Parameters of the function are estimated with the aim of minimizing the sum of squared deviations. Maximizing the number of finely graded dietary levels at the expense of replications on the levels chosen seems to us to be the preferred way of conducting these experiments. All of our conclusions concerning response to varying supply or to requirements are based on derived functions and never on individual measuring points. Values for r² and mean square error (MSE) calculated for the regressions in this study (Table 3) indicate that the responses to changes in dietary amino acid concentrations are well described by the calculated functions within the ranges of concentrations studied. In previous studies (Rodehutscord et al. 1995c), using four repeated measurements per treatment (with one tank of 20 trout as one replicate), we found that the pooled SEM in body weight gain was between 2 and 5% of the mean, depending on the experiment. This gives an indication of the between-tank variability in our tank system.

Choo et al. (1991) studied effects of dietary excesses of leucine on growth and body composition of trout fry initially weighing 2 g. Body protein concentration appeared to be reduced when dietary leucine was increased from 11 to 35 g/kg but was unaffected by further increases in dietary leucine up to 134 g/kg, whereas body concentration of lipids tended to increase with increasing dietary leucine. A similar response was observed in lake trout (Salvelinus namaycush) fry when dietary leucine concentration was increased from 5.2 to 16.0 g/kg (Hughes and Rumsey 1983). However, in the present experiment, increasing dietary leucine up to ~12 g/kg dry matter increased concentration of protein in gain and decreased concentration of lipids. At present, no clear explanation can be given for this discrepancy between our results and those quoted, but we assume that the composition of the experimental diets (in particular lipid concentration) may account at least for part of this. Choo et al. (1991) assume, referring to experiments of Tischler and Goldberg (1980), that dietary leucine was used to synthesize triglycerides in adipose tissue and that this may account for increased body lipid seen in fish fed high dietary leucine. We used a diet containing 280 g lipid/kg dry matter, whereas a lipid concentration of about 100 g/kg can be recalculated for the diets used by Choo et al. (1991). It has frequently been shown that lipid concentration in the fish body increases with increasing dietary lipid concentration (Rodehutscord et al. 1995c, Storebakken and Austreng 1987), and any effect of leucine on the synthesis of triglycerides in adipose tissue was probably overlapped by the high amount of fatty acids directly transferred to this tissue. As long as dietary leucine limits protein retention when dietary leucine concentration is increased, an increased proportion of ingested digestible energy can be used for body protein synthesis and does not have to be deposited in adipose tissue; this may explain the decrease in body lipid concentration observed when dietary leucine was elevated to 12 g/kg dry matter.

We cannot explain why body lipid concentration increased with dietary valine concentration whereas it decreased when dietary concentrations of all other amino acids studied here increased.

Table 5. Dietary amino acid concentrations required to reach 95% of plateau value of performance traits in six experiments with rainbow trout [View Table]

Dietary excess of leucine may depress growth and feed conversion when concentrations > 92 g/kg are used in rainbow trout (Choo et al. 1991) or >50 g/kg in lake trout (Hughes et al. 1984). Accordingly, in the present study, increasing dietary leucine to a concentration that was about threefold higher than that required for high protein deposition (42 g/kg dry matter) did not influence either the amount or the composition of weight gain. However, studying the effects of dietary excesses of amino acids has not been the objective of our experiments.

For each of the traits evaluated with exponential functions, plateau values that describe a theoretical maximum or minimum that can be approached by continuously increasing the dietary concentration of the individual amino acids under test were calculated (parameter a in Table 3). The exponential functions describe the changes in efficiency of supplemented amino acids with increasing dietary concentrations of the respective amino acid and, therefore, leave space for interpretation regarding the "optimum" concentration required in the diet. With the following transformations of Equations (1) and (2), the dietary concentration of an amino acid x_z (g/kg dry matter) required to reach z% of the respective plateau value can be calculated as follows: for traits evaluated with Equation (1) and for traits evaluated with Equation (2)where a, b, c and d are the respective parameters summarized in Table 3. Table 5, as an example, summarizes those dietary concentrations of all essential amino acids studied here that were required to reach 95% of the calculated plateau values for each trait. We decided more or less arbitrarily to use this 95% level for further conclusions, but data given in Table 5 can be recalculated for any level differing from 95% using the original data given in Table 3. Data shown in Table 5 indicate that, among the traits monitored, protein deposition is the most sensitive indicator of a suboptimal supply of an amino acid. This conclusion could also be drawn from results found previously for methionine, arginine and threonine (Rodehutscord et al. 1995a and 1995b), and it shows that at least part of the differences found in recommendations for amino acid supply in the literature must be due to different response criteria chosen in individual studies. Our recommendations concerning necessary amino acid concentrations in trout diets are based on protein deposition data.

Lanari et al. (1991) studied the lysine requirement using trout similar in size to those in this study. From their results in body weight gain, they recommended 21.6 g lysine/kg dry matter, which is close to that concentration required for 95% of maximum body weight gain in the present experiment (23.2 g/kg dry matter).

Results presented in this paper and those published previously (Rodehutscord et al. 1995a and 1995b) form the basis to derive recommendations for the dietary supply with nine out of ten essential amino acids, based on a uniform method and almost identical basal diets. These recommendations are summarized in Table 6. Concentration of digestible energy grossly determines the gain/feed ratio and, therefore, the amount of amino acids ingested per unit of gain. Because DE concentrations may vary greatly between diets, it is preferable to relate concentrations of nutrients to MJ DE rather than to kilograms of feed. Based on the 20.1 MJ DE/kg dry matter of the diets used here, the recommendations are expressed in Table 6. Recommendations of NRC (1993) were recalculated in a similar way. Recommendations derived from our data exceed those of NRC (1993) for lysine and isoleucine by 16 and 13%, respectively, but they are lower by 3-4% for valine and threonine and by 23-42% for the remaining amino acids. It must be pointed out, however, that the 1993 NRC recommendations are based on a literature survey considering a large number of papers dealing with requirement studies on trout and salmon, most of which were performed with juvenile fish and widely different methodological approaches. Obviously, different response criteria lead to different recommendations (see Table 5), as do differences in growth rate (Cowey 1994), probably because of an increase in the proportion of maintenance requirements with decreasing growth rate. Requirement for any amino acid comprises both retention of the amino acid and maintenance requirement. Assuming that the maintenance requirement depends on the body size of a fish, the proportion of maintenance requirement on total amino acid requirement decreases with increasing retention of the amino acid. When the daily growth coefficient is calculated for the present experiments as a way of standardizing growth rates according to the suggestion of Cowey (1992), the figures which result (at least 2.9) are relatively high (Cowey 1992). This might explain why most of the requirement figures derived from our own results are lower than those collected by NRC (1993).

Table 6. Recommended dietary amino acid concentrations [View Table]

In this type of growth study, the total amino acid requirement for both maintenance and retention of protein usually is investigated. In fact, little is known about the proportion of the total amino acid requirement that has to be spent to cover the maintenance requirement of trout. Based on the data presented here and in previous studies from our laboratory, an attempt was undertaken to estimate maintenance requirement figures for each of the amino acids studied. When maintenance requirement is defined as the amount of an amino acid to be ingested by the fish to maintain its body protein pool in an equilibrium, which means that no net synthesis or net breakdown of body protein takes place, this amount can be obtained by extrapolating the dose-response curves for protein retention to y = 0. However, the equations given in Table 3 cannot be used immediately for this because they are based on the dietary concentration of amino acids as an independent variable and not on the quantitative amino acid intake. Therefore, protein retention data were recalculated using Equation (1), with intake of the respective amino acid (mg/fish) as an independent variable. Results are shown in Table 7. The parameter c (the intersection of the curves with the x-axis) in this case describes the amount of the amino acid required to be ingested to avoid any net change in the amount of body protein during the experimental phase. In Table 8, these values are presented together with the total amounts required, which were calculated by multiplying the dietary concentration necessary for 95% of the plateau value in protein deposition (from Table 5) and the dry matter intake achieved at this concentration (from equations for dry matter intake given in Table 3). It is apparent from Table 8 that the maintenance requirement and the proportion of ingested amino acid spent to cover the maintenance requirement differ between individual amino acids. The proportion of maintenance requirement, calculated in this way, was lowest for lysine and isoleucine (4% of total requirement for protein deposition) but could make up as much as one third of the total requirement for tryptophan and leucine (30 and 32%, respectively). The high value for tryptophan may result from the low requirement for growth purposes resulting from the low tryptophan concentration in body protein. The value for leucine, however, appears surprisingly high. Even expressed in absolute terms, the maintenance requirement for leucine is more than fourfold higher than the average value of all other amino acids.

Table 7. Parameters of the exponential functions relating protein deposition (g/fish) to intake of the respective amino acid under study (mg/fish)1 [View Table]

Table 8. Calculation of the maintenance requirement for amino acids in rainbow trout and the proportion of ingested amino acids spent to cover the maintenance requirement [View Table]

In the past, different groups have worked on the optimum amino acid pattern required in the diet to achieve high N retention in different animal species (e.g., Kirchgessner et al. 1995, Wang and Fuller 1989). This might be helpful in generalizing results in so far as amino acid pattern rather than amino acid requirements should remain largely unaffected by factors such as body weight, growth rate, dietary energy concentration or environmental factors. The authors determined these patterns by deleting individual amino acids from a mixture of all amino acids, assuming a linear dose-response relation between the dietary supply of the limiting amino acid and the N retention of the animal. On the basis of the experiments presented here, the optimum dietary amino acid pattern can be derived via an alternative method, i.e., by calculating the amino acids required for a certain level of protein retention directly from the dose-response curves. Depending on the level of protein retention chosen, the derived amino acid patterns are different, as shown in Table 9 for arbitrarily chosen levels of 95, 80 and 65% of the plateau in protein deposition. The lower the desired level of protein retention, the closer the amino acid pattern moves toward that determined for maintenance requirement. This suggests that the assumption of a linear relationship between the supply of a limiting amino acid and retention of body protein is not necessarily valid, at least not in trout. From Table 9 it is also obvious that the amino acid pattern determined in body protein cannot be used as a dietary pattern. Doing so would mean ignoring both differences in the proportion of maintenance requirement and differences in utilization of individual amino acids.

Table 9. Dietary amino acid patterns (relative to lysine) derived for different levels of protein retention and for maintenance and pattern of body protein in rainbow trout [View Table]

A conclusion from Tables 8 and 9 must be that estimates of amino acid requirements based solely on the amino acid pattern of body protein or based on the assumption that the maintenance requirement makes up only a negligible proportion of total requirement (Ogino 1980) or even is equal for all amino acids, most likely lead to inadequate interpretations.