Introduction

Aquaculture feed production is increasingly challenged by the rising cost and limited production of fish meal, the main ingredient in feeds. In the shrimp farming industry, feed is the most expensive input in the variable costs of production when operating grow-out farms. Today a wide range of alternative feedstuffs have to be considered to reduce the amount of fish meal used in penaeid shrimp feed. These feedstuffs must be well characterized to understand their nutritional quality, which is particularly important in the formulation of efficient (highly digestible), ecologically sustainable (low-polluting), and economically competitive feeds. In particular, the apparent digestibility of nutrients in feedstuffs, namely amino acids, plays a very important role in the nutritional quality of feeds. The aim of the present study is to evaluate the apparent digestibility efficiency of dry matter, energy, proteins, and amino acids in a reference diet (commercial diet) and in 4 feedstuffs used in shrimp feed: Pilchard meal (Sardinops sagax), potato protein concentrate (Solanum tuberosum), brewer’s yeast (Saccharomyces cerevisiae), and crustacean meal (Heterocarpus reedi). Apparent digestibility coefficients were adjusted to account for nutrient losses (nutrient leaching) in seawater.

Materials and Methods

Feedstuffs were obtained from feed mills and/or their distributors. We obtained the fish meal (S. sagax) and potato protein concentrate (S. tuberosum) from Sonora (Mexico), the brewer’s yeast (S. cerevisiae) from Baja California (Mexico), and the crustacean meal (H. reedi) from Santiago (Chile). Feeds were prepared following the method described by ^{Villarreal-Cavazos et al. (2017)}; test ingredients were mixed in a KitchenAid mixer (Benton Harbor, Michigan) for 10 min while adding warm water (30%), and then for 15 min after water was incorporated. A Torrey meat grinder (Monterrey, Mexico) with 1.6-mm holes was used. Manufacturing time was 40 min per kilogram of feed, and manufacturing temperatures reached 75-80 ºC. The extruded feed was dried in a ventilated oven at 100 ºC for 8 min and kept at room temperature for one night before packing. Digestibility of feedstuffs was determined following the method described by ^{Cho and Slinger (1979)}. Experimental diets contained 30% of the test ingredient and 70% of the reference diet (^{Cruz-Suárez et al. 2009}).

The digestibility trial was carried out in a closed recirculating artificial seawater system at the Mariculture Program facilities (Faculty of Biological Sciences, Autonomous University of Nuevo León). Twenty-eight identical, fiberglass aquariums with 120-L capacity and regulated flow rate of 710 mL·min^-1 were used. Tanks were randomly distributed. Each diet was tested in 4 replicate tanks with 24 Peaneus vannamei juveniles (5.4 ± 0.1 g) each. Shrimp were fed ad libitum starting with 10% of total biomass in the tank. Fresh feces were collected (12 g) from each tank with a siphon, immediately after excretion, over a period of 7 d.

The bromatological composition of feedstuffs, diets, and feces was determined with the following methods: 930.15, 990.03, 942.05, and 962.09B for moisture, crude protein, ash, and fiber, respectively (^{AOAC International 1997}). Lipids were extracted using the Soxhlet method (^{Tecator 1983}), and the nitrogen free extract was calculated as the difference. Nutrient loss in seawater was determined according to the method reported by ^{Tapia-Salazar et al. (2012)}. Chrome content was determined according to the method described by ^{Bolin et al. (1952)} and modified by ^{Cruz-Suárez et al. (2009)}. Amino acid composition was determined according to ^{Llames and Fontaine (1994)} and ^{Fontaine (2003)}. The apparent digestibility coefficients (ADCs) for dry matter, protein, and amino acids in the diets were calculated using the following equation: %ADC_diet = 100 - [100(C_diet/N_diet) × (N_feces/ C_feces)], where C and N are, respectively, the chromium oxide and nutrient concentrations in the diets or feces (dry weight basis). The ADCs for feedstuffs were calculated according to ^{Bureau and Hua (2006)}. The ADCs for diets and feedstuffs were adjusted considering the percentage of nutrient loss in water due to leaching (^{Villarreal-Cavazos et al. 2014}). We used the leaching values that were obtained after 1-h immersions of feeds in seawater. To account for losses in dietary dry matter, protein, and amino acids before ingestion by the animal, the ADC for each nutrient was corrected for using the equations reported by ^{Nieto-López et al. (2011)}. Nutrient concentration values (feedstuffs, diets, leached diets, feces) and ADCs were subjected to a one-way analysis of variance and Duncan’s multiple range test to separate the treatment averages into normal and homogenous groups. The software used was SPSS v.22 for Windows.

Results

Crude protein content ranged from 42.5% to 89.3%. Lipids oscillated between 1.4% and 10.8% and crude fiber between 0.1% and 3.4%. The range for ash content was 5.3-18.1%. Amino acid (AA) proximate composition and AA contents varied greatly between feedstuffs (Table 1). The composition of experimental feeds showed values that were very close to the expected values, according to calculations based on the nutrient concentrations of feedstuffs in the formulated feeds (Table 2). Feeds containing crustacean meal and the fish meal showed the highest nutrient losses (20% and 24%, respectively) in seawater (Table 3). The AAs with the highest losses in seawater were lysine, methionine, and arginine (27%, 24%, and 20%, respectively) (Table 3).

The ADC varied between 79% and 91% for dry matter, between 89% and 95% for energy, and between 78% and 93% for crude protein. The ADC for total AAs varied between 75% and 97% (Table 4). AAs in the crustacean meal feed proved to be highly digestible (94.6% mean digestibility). Digestibility of AAs in the brewer’s yeast meal was good (87.1%). Mean AA digestibility was lowest in the potato protein concentrate (75.2%). In general, mean digestibility of AAs in the experimental feedstuffs was more than 90%, except in the case of cystine (80.6%). The AAs with digestibility values greater than 95% were methionine, isoleucine, lysine, glycine, and proline. Cystine was the least digestible AA. Corrections for nutrient leaching tended to reduce standard ADC values for most of the experimental feedstuffs (Table 5).

Discussion

The AA proximate composition and AA contents found for the different feedstuffs were similar to what has been previously reported (^{NRC 1983}, ^{Novus 1996}, ^{Hess et al. 2006}). In addition, in terms of dry matter, energy, crude protein, and total AAs, all feedstuffs (fish meal, crustacean meal, brewer’s yeast, and potato protein concentrate) were more digestible than the reference diet. ^{Goytortúa-Bores et al. (2006)} analyzed the crustacean meal (red crab, Pleuroncodes planipes) used to feed Pacific white shrimp and found that ADC values for crude protein were lower than the value observed in the present study (82-84% vs 92.8%). This value coincides with that reported by ^{Tibbets et al. (2006)} for whole krill fed to cod (96%). ^{Ozório et al. (2012)} evaluated the increasing levels of brewer’s yeast in tilapia feed and found that apparent digestibility of crude protein decreased (82.3%, 81.3%, 78.6%, 78.1%, and 79.4%) as inclusion levels of brewer’s yeast in the diet increased (0%, 10%, 15%, 20%, and 30%, respectively). In the present study, for Pacific white shrimp, brewer’s yeast was a highly digestible feedstuff (91.8%) in terms of crude protein at an inclusion level of 30%. ^{Terrazas-Fierro et al. (2010)} reported ADC values for essential AAs in a red crab (P. planipes) meal that were much lower (82.1%) than the values found for crustacean meal in the present study (95%). In the present study, experimental diets were formulated with 30% of the test ingredient. The Commonwealth Scientific and Industrial Research Organisation (Australia) work group indicates that the higher the inclusion level of the feedstuff in the experimental diet, the higher the precision in the determination of feedstuff digestibility because differences with the reference diet become clearer and more accurate (^{Cruz-Suárez et al. 2008}). Correction for nutrient leaching tended to increase ADC values for brewer’s yeast, whereas the rest of the feedstuffs showed the opposite effect. ^{Tusche et al. (2011)} reported low growth rates for rainbow trout when feed contained 50% of potato protein concentrate. ^{Xie and Jokumsen (1997)} reported that growth rates for rainbow trout decreased as inclusion levels of potato protein concentrate increased, which could be associated with low AA digestibility, as shown in the present study for shrimp. The present study showed that the digestive efficiency of essential AAs in the experimental fish meal was high (97%), much higher than the values reported by ^{Yang et al. (2009)} (87%), ^{Terrazas-Fierro et al. (2010)} (89%), and ^{Carvalho et al. 2016} (84%). This differences could be associated with different factors, such as fish species, raw material (whole fish, by-products resulting from the filleting process, mixture of viscera from other species, etc.), or the elaboration process (drying by direct flame or by steam jacket). However, the previously mentioned studies do not specify on the raw material or elaboration process used.

English translation by Claudia Michel-Villalobos