Impact of reference diet composition on apparent digestibility coefficients of two protein-rich ingredients in Nile tilapia

ABSTRACT Protein quality is related to amino acid composition and digestibility. Accurate evaluation of apparent digestibility coefficients (ADCs) of nutrients in commonly used feedstuffs is paramount for formulating efficient aquafeed. ADCs of soybean meal (SBM) and poultry by-product meal (PBM) were evaluated using reference diets formulated with two types of ingredients (semi-purified [SP] and practical [P]) for juvenile Nile tilapia ( Oreochromis niloticus, Linnaeus) of the GIFT strain. Groups of 20 juveniles (65.05 ± 12.37 g) were fed twice a day to apparent satiety with one of the four experimental diets (SBM-SP, SBM-P, PBM-SP, and PBM-P) in quadruplicate for 30 days. After the last feeding, feces were collected by siphoning hourly and the ADCs of dry matter, protein, and amino acids (AAs) were calculated. Nile tilapia exhibited a high capacity to digest SBM and PBM, with most ADCs exceeding 90 %. The type of reference diet affected the ADCs of protein and AAs on the test ingredients, with the SP reference diet providing the highest ADC, mainly in SBM. Digestibility data generated with a P-type reference diet demonstrated more practical relevance than those generated with an SP-type reference diet. They can be applied in digestibility studies for Nile tilapia.


Introduction
Accurate elucidation of apparent digestibility coefficients (ADCs) of commonly used protein-rich nutrient sources is vital to formulate nutrient-and cost-effective diets for commercial species (Fracalossi and Cyrino, 2013;Hardy, 2010;Lupatsch et al., 1997).In the evaluation of feed ingredient digestibility, the test ingredient is paired with a reference diet, usually at a ratio 30:70 (NRC, 2011).Highquality semi-purified (SP) ingredients with well-defined composition and without anti-nutritional factors are still used in some digestibility trials to compose the reference diet (Glencross et al., 2007;Lovell, 1998).Such ingredients predominantly provide one macronutrient and only trace amounts of vitamins and minerals; however, they can reduce palatability and feed intake (Hardy and Barrows, 2002) and do not represent the reality of commercial diets.In contrast, despite anti-nutritional factors, a reference diet composed of practical (P) ingredients has advantages, such as similar composition to commercial feeds, high feed consumption due to high palatability, and production of more fecal material (NRC, 2011).
Soybean meal (SBM) and poultry by-product meal (PBM) were the ingredients investigated in the present study because they are protein-rich sources widely used in commercial diets for Nile tilapia (Oreochromis niloticus, Linnaeus).SBM is abundantly available and a cost-effective source of digestible plant protein and essential amino acids (EAAs) (Gatlin et al., 2007;Nguyen et al., 2009).PBM is an animal source of highly digestible protein and a good alternative to fish meals because of its similar protein composition, high production volume, and relatively low cost (Cruz-Suárez et al., 2007;Hardy, 2010).
Nile tilapia is the second most-produced fish group in aquaculture worldwide, with a production of 4.4 million tons, and Brazil is the 4 th largest producer (FAO, 2022a, b).Data on nutrient digestibility of Nile tilapia are available; nevertheless, the methodology varies and lacks a standard regarding the reference diet composition (Borghesi et al., 2008;Cardoso et al., 2021;Davies et al., 2011;Maas et al., 2019).Therefore, in the present study, we compared the use of different reference diet compositions (P or SP) in two important protein-rich feedstuffs to verify possible variations in digestibility of protein and AAs by Nile tilapia.

Feed ingredients and diet preparation
Two reference diets were formulated (Table 1) with semi-purified (SP) or practical (P) ingredients to meet the nutritional requirements of Nile tilapia (Pezzato et al., 2010) or Oreochromis sp.(NRC, 2011).Each reference diet was used to estimate nutrient digestibility of two practical protein-rich ingredients (Table 2), PBM and SBM, at a 30 % level of inclusion.The proximate composition, AAs, and energy content of the feed ingredients (corn, fish meal, and soy protein concentrate) were determined before the experimental diet formulation.The two reference diets (SP and P) and the four test diets (SBM-SP, SBM-P, PBM-SP, and PBM-P) contained 0.1 % of yttrium oxide as an inert marker.
Ingredients were ground using a hammer mill (1.0 mm screen mesh) and then manually sieved (0.6 mm), weighed, and homogenized in a horizontal mixer.The Sci. Agric.v.80, e20220189, 2023 moisture of the ingredient mixture was adjusted to 25 % using water.Diet extrusion was performed in a singlescrew extruder (model MX-40; Inbramaq).The extrusion parameters were as follows: temperature 100 °C; thread speed 220 rpm; flow rate 20 % of rated capacity; width to diameter ratio 2.3:1; thread diameter 92.5 mm; and cylinder length 210 mm.After extrusion, pellets (4 mm) were dried in a forced-air circulation oven at 55 °C and then packaged and stored in air-tight containers at a constant temperature of 23 °C.The proximate and AA compositions of the two reference and the four experimental diets are shown in Tables 2 and 3, respectively.

Fish and experimental conditions
The digestibility trial was performed in Florianópolis, Santa Catarina State, Brazil (27°43'45" S, 48°30'31" W, altitude 3 m), following protocol 9377080618 approved by the Ethics Committee on Animal Use of the Federal University of Santa Catarina (CEUA, UFSC).Nile tilapia juveniles of the GIFT strain, sexually inverted to male, were obtained from a commercial fish farm (Piscicultura Pomerode).Before the digestibility trial, fish were acclimated to laboratory conditions for two weeks in three 1,000 L tanks connected to a recirculation system and equipped with biological and mechanical filtration, an air supply, and a heat exchanger.The temperature was set to 28 °C, and the photoperiod was adjusted to 12 h.27.96 ± 0.36 °C; dissolved oxygen 5.69 ± 0.41 mg L -1 ; pH 7.42 ± 0.24; salinity 1.97 ± 0.20 g L -1 ; alkalinity 59.82 ± 8.30 CaCO 3 mg L -1 ; total ammonia 0.55 ± 0.10 mg L -1 ; and 0.01 mg L -1 nitrite.All variables measured remained within the comfort range of Nile tilapia (Webster and Lim, 2006).The water inflow rate in each experimental unit was 25 mL s -1 .

Chemical analysis
The proximate analysis of the diet ingredients, diets, and feces followed procedures standardized by the "Association of Official Analytical Chemists" (AOAC, 1999): moisture (dried at 105 °C to a constant weight, method 950.01), total lipid (Soxhlet, method 920.39C), and ash (incineration at 550 °C, method 942.05).
According to the manufacturer's instructions, gross energy was determined using a calorimeter (PARR, model ASSY 6200).Crude protein, the amino acid content, crude fiber, and neutral detergent fiber of the diet ingredients (corn, fish meal, soy protein concentrate, poultry by-product, and soybean meal) were analyzed using near-infrared spectroscopy (NIRs) at the Animal Nutrition Laboratory (Evonik).Crude protein and amino acid from feed and feces were analyzed by wet chemistry at the Evonik Laboratory using ion-exchange chromatography with Following this period, fish with an average initial weight of 65.05 g ± 12.37 and a total length of 14.75 cm ± 0.86 (mean ± standard deviation) were transferred to 24 experimental units (115 L circular tanks), with biomass of approximately 1,500 g per unit (23 fish per unit).Tanks were connected to a closed freshwater recirculation system with aeration, mechanical and biological filtration, and temperature and photoperiod were adjusted to 28 °C and 12 h of light, respectively.The experiment was conducted in a completely randomized design, using four replications for each experimental diet.Fish were fed twice a day (10h00 and 16h00) until apparent satiation and fecal collection started 7 day after changing the experimental diets to allow the excretion of all previously ingested feed.Feces were collected within each experimental unit by siphoning for 30 day.One hour after the last feeding, the tanks were cleaned, and the total water volume was renewed to avoid contamination of newly voided feces with uneaten feed and old feces.After cleaning, the water flow was interrupted for one hour, and newly voided feces were collected by siphoning.The collected feces were lyophilized, homogenized, and stored at -20 °C until analysis.
The water quality indicators were measured weekly, except for temperature, dissolved oxygen, and the pH, which were monitored daily.The average values (± standard deviation) were as follows: temperature post-column derivatisation with ninhydrin.Amino acids (AAs) were oxidized with performic acid and neutralized with sodium metabisulfite25 (Commission Directive 1998).AAs were liberated from the protein by hydrolysis with 6N HCl for 24 h at 110 °C and quantified using the internal standard method by measuring the absorption of reaction products with ninhydrin at 570 nm.The inert marker yttrium was measured using inductively coupled plasma mass spectrometry (ICP-MS) at the Atomic Spectrometry Laboratory (UFSC).The samples were introduced using a pneumatic nebulizer.

Calculation of apparent digestibility coefficient
Protein, AAs, and dry matter apparent digestibility coefficients (ADCs) were estimated using the following equations: For the diets (NRC, 2011): For the test ingredients (Bureau et al., 1999): where nutr = nutrient, ing = ingredient, td = test diet, and ref = reference diet.

Statistical analysis
All data are reported as mean ± standard deviation.To test differences between the two types of reference diets, ADC data were first tested for normality and homoscedasticity and then subjected to the Student's t-test.The same procedures were applied to test the differences between the ADCs of the tested ingredients (SBM and PBM) within each different reference diet (SP and P).Statistical analyses were performed using Statistica 10.0 software, adopting a 5 % confidence level.

Results
The ADCs of dry matter, protein, energy, and AAs in the reference diets used in this study are presented in Table 4.The ADCs of dry matter and crude protein of the P reference diet were higher than those of the SP reference diet.The ADCs of most AAs exceeded 90 % and were similar between the reference diets.However, for the EAAs arginine and histidine and the non-essential amino acids (NEAA) aspartic acid and cysteine, the type of reference diet affected the ADC values, which were higher for the P reference diet.In turn, the ADC of leucine was higher in the SP reference diet.In both reference diets, the AAs arginine, lysine, and glutamic acid presented the highest ADC values (92.76 -96.11 %), whereas isoleucine presented the lowest values (86.09 -86.66 %).
The ADCs of selected nutrients in SBM tested using the two types of reference diets are presented in Table 5.The type of ingredients used in the reference diets did not affect the ADCs of dry matter, while ADC of protein was higher in the SBM tested using the SP reference diet than the P reference diet.Similarly, the ADCs for all AAs were higher in SBM when tested using the SP versus the P-type reference diet, except for cysteine, but with no difference.The EAAs arginine, histidine, and lysine, and NEAAs aspartic acid and glutamic acid exhibited the highest ADC values (95.21 -103.41 %).In contrast, the EAA isoleucine exhibited the lowest ADC values (89.11 and 94.21 %, respectively, for the P and SP diets).
The ADCs of the selected nutrients of PBM tested using two reference diets are presented in Table 5.The ADCs of dry matter, protein, and most AAs were not affected by the type of reference diet.However, the ADCs of histidine, glycine, and proline were higher in the PBM-SP diet, whereas only the ADC of cysteine was substantially higher in the PBM-P diet than in the PBM-SP diet.For the latter test ingredient, the EAAs arginine, histidine, and lysine, and the NEAAs alanine, glycine, and glutamic acid exhibited the highest values of ADC (93.28 -98.84 %), regardless of the type of reference diet.

Discussion
The correct assessment of nutrient use for a species of great economic importance, such as Nile tilapia, is essential to formulate efficient diets.Several strategies can be applied to assess the nutritional quality of ingredients; however, the choice of these strategies can affect data interpretation (Glencross, 2020).The composition of the reference diet varies greatly in digestibility studies for Nile tilapia from those using only practical (Cardoso et al., 2021;Guimarães et al., 2008;Hernandéz et al., 2010;Köprücü and Özdemir, 2005;Magalhães et al., 2018;Schneider et al., 2004;Vidal et al., 2015Vidal et al., , 2017) ) or SP type ingredients (Borghesi et al., 2008;Furuya et al., 2001;Rodrigues et al., 2012) to those using a mixture of both (Davies et al., 2011;Xavier et al., 2014).Nevertheless, data on the ADCs of nutrients within reference diets are challenging to interpret or discuss.Our findings showed that the type of ingredients in the reference diet could affect the ADCs of dry matter, protein, and selected AAs.
For both types of reference diets, digestibility of protein and AAs was similar, with minor variations in ADCs, except for cysteine.Four AAs (arginine, histidine, aspartic acid, and cysteine) were more digestible in the P reference diet.Leucine was the only AA with a high ADC in the SP reference diet.Therefore, these differences suggested that the digestibility of a particular AA depends on the protein type that composes the P or SP reference diet and the capacity of Nile tilapia to digest each particular AA, considering their chemical characteristics.Arginine and histidine are positively charged (basic) AAs, whereas aspartic acid is a negatively charged (acidic) AA under physiological conditions, with both types mostly exposed to the protein surface.These charged R-groups are more hydrophilic, facilitating enzyme-catalyzed reactions by functioning as proton donors and acceptors (Nelson and Cox, 2017).Considering our results and that a reference diet with P ingredients is more palatable than the SP type (NRC, 2011), we speculated that the highest feed intake, promoted by the greatest diet palatability, increases the activity of digestive enzymes (Moraes and Almeida, 2020).Although our findings exhibited similarities between the protein ADC and the average AA coefficients, individual AA ADCs are variable and can be higher or lower than the coefficient value for protein digestibility (NRC, 2011;Storebakken et al., 2000), as also observed in our study.A possible reason for the higher ADC values in the SBM obtained when feeding the SP diet is the endogenous loss that occurs during digestion.The primary sources of N endogenous AA losses in animals are proteins that are endogenously synthesized and secreted in the digestive tract but are not digested and re-absorbed (Nyachoti et al., 1997).Endogenous losses can be induced by ingestion of a diet with a particular composition, such as protein level and fibre type (Cowieson and Ravindran, 2007;Stein et al., 1999).SP and P ingredients have distinct nutritional compositions and properties that directly affect the final characteristics of experimental diets.Although the nutrient contents of the reference diets were similar, the protein type (casein and gelatin versus fish meal and soy protein concentrate) and the fibre type (cellulose versus corn) have distinct characteristics.They can affect endogenous loss by changing viscosity and ingesta transit speed, which can affect mucin secretion and epithelial cell turnover (Parsons et al., 1983;Sauer et al., 1991).The highest ADC of AAs in the SBM-SP diet suggests that the ingredients used in the SP reference diet reduced endogenous AA losses.This may occur during the digestive process because of the absence of anti-nutritional factors in the SP ingredients and the interaction between dietary nutrients, such as fiber and protein.
Notably, regardless of the type of reference diet tested, the AAs arginine, histidine, lysine, aspartic acid, and glutamic acid presented the highest ADC values (> 95 %) in SBM.This is probably because they are charged AAs, highly hydrophilic, and more susceptible to enzyme reactions.Possible explanations for coefficients above 100 % include analytical errors for nutrients and markers, sampling, improper diet mixing, or interactions between diet ingredients (Glencross et al., 2007).We suggest that the SP ingredients, in addition to not having anti-nutritional factors and reducing endogenous AA losses, may interact with the nutrients contained in the SBM leading to overestimations of the final ADC values of some AAs.The digestibility of all EAAs was high in both types of reference diet, confirming the applicability of this ingredient as a vegetable source of digestible EAAs in aquafeeds, as previously reported by Elesho et al. (2021) and Vidal et al. (2017).
The nutrient digestibility values for PBM revealed less influence from the reference diet composition.Data showed high digestibility of dry matter (99.73 % and 99.74 %), protein (94.63 and 95.75 %), and most of the AAs (> 93 %).The mean AA ADCs were 94.10 % and 93.43 % for PBM-SP and PBM-P, respectively, similar to the mean 91.20 % reported by Guimarães et al. (2008).PBM is considered one of the most promising alternatives to replace fish meal because of its high protein content and quality, AA profile, essential fatty acids, vitamins, minerals, and good palatability (Cruz-Suárez et al., 2007;Gunben et al., 2014).The nutrient ADC of PBM confirmed the applicability of this ingredient in aquafeeds for Nile tilapia and other omnivorous species, such as African SBM exhibited very high digestibility for protein (93.94 % and 99.64 %), with most AA ADCs exceeding 93 %, regardless of the type of reference diet used.These findings were similar to those previously reported for Nile tilapia: 92.72 % (Furuya et al., 2001), 91.56 % (Pezzato et al., 2002), and 92.74 % (Guimarães et al., 2008), but higher than that reported by Ribeiro et al. (2011) (86.01 %) that used the dissection method for fecal collection.A comparison of the ADCs of SBM using different reference diets showed that SBM-SP resulted in a higher ADC of dry matter (80.13 %) and protein (99.64 %) than SBM-P (78.40 % and 93.94 %, respectively).The AA digestibility values followed the same pattern, with mean ADCs of 100.19 % and 93.66 % for SBM-SP and SBM-P diets, respectively.A similar result for the average AA ADC (92.30 %) was reported by Guimarães et al. (2008), while a lower value (87.10 %) was reported by Köprücü and Özdemir (2005), who used the same combination of SBM and a P-type reference diet.
The variation in our study regarding protein and AA digestibility of SBM could be explained by the differences in diet palatability and subsequent feed intake, anti-nutritional factors, and endogenous nitrogen (N) losses.The lower palatability of the SP ingredients compared to the P ingredients could partially explain such discrepancies.Moreover, SBM contains antinutritional factors that negatively affect its palatability to fish compared to animal sources, such as fish meal or PBM (Gatlin et al., 2007;NRC, 2011).The SP reference diet and SBM combination resulted in reduced feed intake, as evident in our trial.Fish were fed twice a day to apparent satiation, and during daily feeding (although feed intake was not recorded), fish fed with the P diet combinations exhibited greater voracity.
Another important factor in nutrient use is the speed at which the feed passes through the digestive tract because transit time has been reported to affect nutrient utilization efficiency (Elesho et al., 2021;Henken et al., 1985;NRC, 2011;Riche et al., 2004).When fish consumes less feed, as evident from the SBM-SP diet, the transit time is lowered, and the feed is subjected to more prolonged digestion exposure, resulting in increased absorption (Moraes and Almeida, 2020).In addition to palatability, the absence of anti-nutritional factors in SP ingredients made them more digestible to fish, thereby reducing the need for high feed intake.Phytate, an anti-nutritional present in plant feedstuffs such as soybean meal, cannot be digested by fish (Kumar et al., 2012;Oliva-Teles et al., 1998;Rodehutscord and Pfeffer, 1995).Besides, the deleterious effect of phytate on phosphorous availability also inhibits proteases, such as trypsin and pepsin (Liu et al., 2009), decreasing protein and AA digestibility (Lima et al., 2021;Spinelli et al., 1983).The presence of phytate in plant ingredients composing the P reference diet could explain the lower ADC of protein and AAs in the SMB-P diet compared to the SBPM-SP diet.catfish (Clarias gariepinus) and Pacific white shrimp (Litopenaeus vannamei), as reported by Elesho et al. (2021) and Cruz-Suárez et al. (2007), respectively.We did not observe a reduction in voracity in Nile tilapia when PBM was tested in an SP reference diet.The good palatability of animal protein sources could have minimized the effects of a non-attractive SP reference diet as registered with the plant ingredient SBM.Thus, the potential effects of an SP reference diet to increase the ADC values, verified when testing SBM, were observed in minor proportions when testing PBM because only three of the 16 AAs evaluated presented higher ADCs.

Conclusions
As a typical omnivorous species, Nile tilapia exhibits a high capacity to digest different protein-rich ingredients (SBM and PBM), with most ADCs exceeding 90 %.Our findings demonstrate that the type of ingredients used in the reference diet (SP or P) affects the plant protein source SBM more significantly compared to the animal protein source PBM.Thus, using practical ingredients in the reference diet has more relevance and can be applied in digestibility studies for Nile tilapia considering, diet palatability, amount of feces collected, costs, and availability.

Table 1 -
Formulation of reference diets.

Table 2 -
Analyzed nutritional composition of the reference diets and test ingredients.

Table 3 -
Analyzed nutritional composition of the test diets.

Table 4 -
Apparent digestibility coefficients of dry matter, protein, energy, and amino acids of the two reference diets for Nile tilapia.
a, b Values followed by different superscripts within the same row are different.Reference diets and digestibility Sci.Agric.v.80, e20220189, 2023

Table 5 -
Apparent digestibility coefficients of dry matter, protein, energy, and amino acids in soybean meal and poultry by-product meal for Nile tilapia.
a, b For soybean meal: within the same row, values followed by different letters are different; A, B For poultry by-products, within the same row, values followed by different letters are different.