Dietary Lysine Requirements of Colossoma Macropomum (Cuvier, 1818) Based on Growth Performance, Hepatic and Intestinal Morphohistology and Plasma Biochemistry


 This study aimed to determine the dietary lysine requirements of juvenile Colossoma macropomum tambaqui based on growth performance. We also evaluated gut and hepatic histomorphometry as well as blood metabolites in accordance with the increased levels of dietary lysine. The juveniles (33.88 ± 2.47 g) were fed until apparent satiation with diets containing 6.60, 9.72, 12.84, 15.96, 19.08 and 22.20 g/kg of lysine. Fish were randomly distributed in groups of 10 fish per tank and assays were performed in triplicate, during 90 days. Tambaqui fed with 15.96 g/kg dietary lysine showed higher final weight (p = 0.001) and optimized feed conversion ratio (p = 0.001). Morphohistological modifications were present in livers of fish fed with low levels of lysine. In the proximal intestine, mucosa layer density was greater at the level of 15.96 g/kg (p = 0.001). In the middle intestine, height (p = 0.001) and perimeter (p = 0.001) of the villi were greater at low levels of lysine (respectively, 9.72 and 12.84 g/kg dietary lysine). Tambaqui fed with 15.96 g/kg of lysine achieved higher plasma protein concentrations (p = 0.01). Using the second-order polynomial regression analysis as support, and based on protein efficiency rate and body weight gain, dietary lysine requirement for juvenile tambaqui was calculated as 15.4–15.6 g/kg of diet (5.7–5.8% of dietary protein).


Introduction
Diets formulated with amino acids that meet the requirements of the sh are necessary in order to support biological development, health and the performance in sh farming (Mai et  Dietary lysine may affect the morpho-histological aspects of the liver (Hua et al. 2019). This is because lysine stimulates the synthesis of L-carnitine, which mobilizes long-chain fatty acids into the hepatocytes, where they are catabolized and provide energy for the maintenance and performance of the sh (Yang et according to the manufacturer's protocol and maintained between 0.00 and 1.0 mg/L (Schimittou 1993).
The juvenile tambaqui were maintained under a constant photoperiod of 12 h light: 12h dark.

Fish sampling
Prior to the feeding trials, nine juvenile tambaqui fed with the initial diet were randomly selected after 24 hours of feeding restriction. Similarly, at the end of the feeding trials, three sh were sampled from each experimental unit. These were anesthetized (benzocaine 100 mg/L) for blood collection by puncture of the caudal vessel using syringes (1 mL) with 15 μL of the anticoagulant heparin sodium 5,000 IU (150 IU/mL). Heparin was diluted in 0.65% saline (1:50) and blood samples were preserved at -4 °C from the time of the collection until centrifugation for obtaining the plasma (Oliveira et al. 2012).
Subsequently, the sh were euthanized with benzocaine (250 mg/L) and spinal cord dislocation was performed for extraction of the cell tissues (Underwood et al. 2013). After that, they were weighed (g, accuracy: 0.001g; Gehaka ® , São Paulo, Brazil) and measured (cm, ichthyometer). The muscle, which was analyzed for proximate composition, was extracted from the dorsolateral region using scissors and cutting blades. In the nal period of the feeding trial, the livers and intestines of the sh (n= 54) were weighed (0.001 g) for analysis of the hepatosomatic index and intestinal morphometric analysis.
Additionally, visceral fat from the sh was sampled. For morphohistological analyses, the livers and intestines selected in the initial and nal period (respectively: n= 9 and n= 54) were xed in a speci c solution.

Muscle proximate composition
In the analysis of the proximal composition of the muscles for crude protein and lipids, the extracted tissue was homogenized, dried in an oven at 105 °C and crushed, for veri cation of nutritional content (AOAC 2005). The crude protein was analyzed using the Microkjehldal method, with steps for digestion and distillation of the matter, followed by titration with acid. Total lipids (ether extract) were veri ed using the bligh and dyer method, and the total lipids and the ash content were veri ed by incinerating the organic matter in a mu e furnace at 550 ºC.

Hepatic morphohistology
Liver samples were collected from ve sh, per each diet type (n= 30) were packed in histological cassettes and xed for 24 hours in Davison's solution. These were dehydrated in increasing concentrations of alcohol and embedded in histological resin (Technovit ® , 71000, Kulzer Hanau, Germany), according to the manufacturer's protocol. The 3 µm-thick slides were made in triplicate using semiautomatic microtome (Slee ® , Cut 5062, Mainz, Germany) at the Functional Morphology Laboratory (FML) of the Federal University of Amazonas. Subsequently, slides from the liver were stained with hematoxylin and eosin (He) in order to analyze the hepatocytes. All slides were analyzed using an optical microscope (Leica®, DM 500, Wetzlar, Germany) with a 32 megapixel camera attached to capture images.

Intestinal morphohistology
Total intestinal weight (TIW), total intestinal length (TIL) and relative intestine length (RIL: intestine length/ sh length) were veri ed (n= 9 intestines per diet), prior to the confection of the slides (Rotta 2003). The intestines were sampled from ve sh per diet and segmented in the proximal (PI), middle (MI) and distal (DI) intestine, with cross sections in triplicate and then packed in histological cassettes (n= 310 samples). Subsequently, these were xed for 24 hours (Davison's solution), dehydrated in alcohol, embedded in histological resin and sectioned at a thickness of 3 µm using a semi-automatic microtome to create the slides. All slides were photographed and analyzed using an optical microscope (Leica ® , DM 500, Wetzlar, Germany) with a 32 megapixel camera attached to capture images.
The slides were stained with blue toluidine dye to quantify the fractional volume (based on the delesse principle) or density of the serosa, muscular, submucosa and mucosa layers. In each slide, four villi were selected according to the visibility of the layers (n= 60 villi per diet). Overlapping the digital image, equidistant points (144) were computed, disregarding empty spaces or content other than the layers and villi analyzed.
To detect the expression of collagen on the intestinal tissue, the sections on the slides were stained with marllory's trichrome to identify the green color of the dye reaction in the presence of the collagen. In parallel, the density of the goblet cells secreting acid mucins was quanti ed using alcian blue (Ab) dye pH 2.5 on the slides and analyzed using Stepanizer ® software (Tschanz et al. 2011). The slides were stained with he to analyze of the height and perimeter of the intestinal villi (n= 60 villi per diet) using Image J ® histological evaluation software (Scijava Consortium, Madison, USA). The height was veri ed by measuring from the base to the apex of each villus, and the perimeter was veri ed following the outline of the area of villus, as in Ferreira et al. (2014).

Plasma biochemistry
For the plasma biochemical analysis, plasma from the remaining material from the samples obtained after centrifugation (3,000 rpm) for 10 minutes was used. Total proteins (g/dL), glucose (mg/dL), cholesterol (mg/dL) and triglycerides (g/dL) were analyzed using commercial enzymatic-colorimetric assay kits and spectrophotometric readings (Thermo Fisher Scienti c ® , Waltham, USA). The analyses were performed at the Chemistry Laboratory of the Federal Institute of Espírito Santo, Campus Piúma, ES, according to the manufacturer's speci c recommendations (Labtest ® , Belo Horizonte, Brazil).

Statistical analysis
The feeding trial was organized using a randomized design, containing a control diet (initial period) and six formulated diets with three replicates. The data were submitted to One-way anova, followed by the Tukey test (p< 0.01 and p< 0.05) to compare results, which were shown as mean values ± standard deviation. Additionally, the dietary lysine requirement was estimated based on body weight gain and on protein e ciency rate, by second-order polynomial analysis model, given by the equation: Y = a + bX + cX 2 . The broken-line regression analysis [(Y1 = LU * (X -R); Y = IF (X < R, Y1, L); L = plateau, U = slope and R = breaking point] was used to estimate the requirement of lysine in the diet by feed e ciency rate. The values of X and Y in the equation correspond, respectively, to dietary lysine (g/kg) and the variable analyzed in the current study. All statistical analyses were performed using R ® statistical software (rproject, Auckland, New Zealand), version 3.5.3.

Growth performance
The results show that supplementation of dietary lysine signi cantly affected the performance for nal weight (FW) in juvenile tambaqui. The FW of tambaqui fed 15.96 g/kg and 22.20 g/kg of lysine was comparable though higher (p= 0.010) than the FW of tambaqui fed with other diets ( Table 2). The dietary lysine requirement for tambaqui, based on body weight gain (Figure 1a) was estimated according to the second-order polynomial analysis at 15.661 g/kg of lysine (5.800% of the diet), calculated using the equation: y= 20.43378 + 13.00819x -0.415286x²; r²= 0.200.
The lysine intake increased (p= 0.001) in parallel to the addition of levels of the dietary lysine, analyzed by second-order polynomial regression, and described by equation: y= 0.04395736 + 0.01223681x + 6.99e-06x², r²= 0.901. In contrast, the feed intake of tambaqui was unaffected by dietary lysine levels (p= 0.118).
Tambaqui fed 15.96 g/kg of dietary lysine showed lower apparent feed conversion (AFC: 1.45 kg/kg; p= 0.001) than those fed with different levels of dietary lysine ( Table 2). The levels of dietary lysine 9.72 g/kg and 12.84 g/kg signi cantly affected (p= 0.049) the hepatosomatic index (HSI) of the tambaqui (Table 2). However, the visceral fat index (VFI) was not signi cantly affected (p= 0.242) by the lysine levels in the diet ( Table 2).
Lysine retention e ciency (LRE) gradually reduced (p= 0.001) with increasing dietary lysine. The feed e ciency rate (FER) of tambaqui did not differ (p= 0.056) ( Table 2) (Table 3) showed that an increase dietary lysine did not increase crude protein percentage (p= 0.051) and lipids (p= 0.061). Fish fed with 22.20 g/kg of dietary lysine demonstrated a higher percentage of ash in the muscle compared to those fed with 6.60, 12.84 and 19.08 g/kg of lysine (p= 0.010), but were comparable to sh fed with 9.72 and 15.96 g/kg of lysine in the diet. In parallel, sh fed with 22.20 g/kg of dietary lysine showed a higher percentage of moisture (p= 0.001) in the muscle compared to those fed with other diets (Table 3).

Hepatic morphohistology
In the macroscopic analysis of liver tissue in the tambaqui, the livers showed reddish-brown and slightly paler brown tones. Microscopically, the tambaqui parenchyma showed hepatocytes with a polyhedral, irregular round and hexagonal shape, with a centralized nucleus (Figure 2a).
The lipids were visible in the hepatocytes of all sh fed the elaborated experimental diets (Figure 2a). Fish fed diets that were elaborated with higher levels of lysine (15.96, 19.08 and 22.20 g/kg) presented livers with lipid molecules in small quantities. These molecules were generally present close to arteries, arterioles, veins and sinusoids, in isolated or grouped forms (Figure 2a and 2b).
Fish fed diets with 9.72 g/kg of lysine had livers with a higher amount of lipids, showing steatosis hepatodystrophy in low intensity. Additionally, these livers showed edematous degeneration, which is con gured by cellular edema with architectural cortical breakdown, clearing of the cytoplasm and centralized permanence of the nucleus (Figure 2c).

Intestinal morphohistology
The averages of the total intestine weight (TIW), of the total intestine lengths (TIL) and of the relative intestinal lengths (RLI) of the tambaqui did not showed a signi cant diference (respectively p= 0.221, p= 0.163, p= 0.590) in response of increasing dietary lysine ( Table 4).
The serosa layers of the proximal (p= 0.273), middle (p= 0.254) and distal intestine (p= 0.172) of tambaqui ( Figure 3) from the nal period and the initial period of the experiment did not differ signi cantly. However, one exception was the tambaqui juveniles fed 12.84 g/kg of dietary lysine, which presented a proximal muscle layer of the intestine equivalent to that of the sh from the initial period, and the proximal muscle layer of the tambaqui from the initial period was greater (p= 0.001) than those of the nal period (Table 5). In the middle portion of the intestine (p= 0.350) and the distal intestine (p= 0.290), the percentage of muscle layer was not in uenced by the different lysine levels in the diet ( Table 5).
The submucosa layer (Figure 3) of the proximal intestine (PI) showed signi cant differences (p= 0.001) between the tambaqui fed with the diet of the nal period and those fed with the diet of the initial period (Table 5). In the middle intestine, the submucosa layer was smaller (p= 0.010) in sh fed 9.72 g/kg of lysine in the diet when compared to those fed diets with 15.96 and 22.20 g/kg of lysine. The mucosa membrane of the pi of sh fed with 15.96 g/kg dietary lysine was statistically different (p= 0.001) from those that received 19.08 kg/g of lysine and these showed the greatest density of this layer in the nal period (Table 5).
Morphohistological images depicting the presence of collagen protein in the cell layers (serosa, muscular, submucosa and mucosa) that form the intestinal villi where the goblet cells are located are shown in the  (Table 6). The height of the villi of the proximal intestine (PI) was not affected (p= 0.100) by increasing dietary lysine (Table 6). In the middle intestine (MI), the lowest and greatest height of the intestinal villi (p= 0.010) of the tambaqui were for those that received 19.08 and 9.72 g/kg of dietary lysine, respectively (Table 6).

Plasma biochemistry
The plasma protein levels of tambaqui fed with the initial diet was higher than the plasma protein levels of those fed in the nal period with 15.96 g/kg of dietary lysine (p= 0.010), as described in Table 7. The blood glucose of the tambaqui was not affected by the dietary lysine (p= 0.120).
The blood of tambaqui fed with the initial diet showed lower cholesterol than that of tambaqui sampled in the nal period that were fed with 9.72, 15.96 and 22.20 g/kg of dietary lysine (p= 0.001). In contrast, tambaqui sampled in the nal period and fed 6.60, 12.84 and 19.08 g/kg of dietary lysine showed lower plasma cholesterol (Table 7). Triglycerides showed no difference among treatments with dietary lysine (p= 0.072).

Discussion
Growth performance In the present study, the 15.98 g/kg level showed the best results in weight gain in tambaqui juveniles.

Proximate composition of muscle tissue
The proximal composition of the sh muscle varies between 15 and 24% for proteins, and between 1 and 2% for Ash. These values may be different between species and within the same species, since it is in uenced by age, growing conditions, body portion and feed intake (Arbeláez-Rojas et al. 2002). No statistical variation was observed in the muscle composition of tambaqui fed with different lysine levels in the current study in relation to the percentage of proteins and lipids, however, it was expressed in the ash and moisture levels.
Ash complexed with amino acids tends to increase the absorption of the nutrient in the intestine, providing transport through mucosa membranes. In parallel to this, the formation of insoluble compounds with possible anti-nutritional factors in the diet are avoided ( Barros et al. 2004).
In the muscles of sh that received diets containing 22.20 g/kg of lysine, the percentage of moisture

Hepatic morphohistology
The formats of the hepatocytes that were veri ed corroborate the structure characterized for tambaqui by Costa et al. (2012) and for pacu by Fujimoto et al. (2008). Microscopically, the tambaqui parenchyma showed hepatocytes with a polyhedral, irregular round and hexagonal shape, with a centralized nucleus (Figure 2a). The reddish brown color presented by the liver tissues occurs due to the abundant vascularization of the organ and indicates normal tissue. However, the brown color in a slightly paler tone suggests hydropic degeneration. It is caused by ionic and homeostatic imbalance, and can generate pallor, turgidity and weight increase in the tissue, as evidenced in the analysis of the HSI of the tambaqui treated with 9.72 g/kg, though it is reversible. Rocha et al. (2010) analyzed the liver of the bream Brachyplatystoma rousseauxii (Castelnau 1855) and detected the hepatic condition of steatosis, characterized by small drops of lipids, isolated or nonisolated, located close to blood vessels, and the same pattern was seen in the present study. Insu cient levels of proteins and amino acids in the diet can promote hepatic steatosis, since proteins act in the transport and uptake of lipids in the liver, and amino acids, such as lysine, synthesize metabolites that oxidize them (Furuya et al. 2013). Juvenile tambaqui fed with lower levels of dietary lysine (9.72 and 12.84 g/kg of lysine) had lipids stored in greater concentration in hepatocytes, had greater liver damage, according to the analysis of performance. The low lipid concentration in the hepatocytes of sh fed with 15.96, 19.08 and 22.20 g/kg of dietary lysine did not affect the growth performance of the juvenile tambaqui.

Intestinal morphohistology
Data from the intestinal morphometric analysis of the tambaqui in this study showed that they were not affected by the lysine levels in the diet. However, the relative intestinal length was 1.20 to 1.37 cm/cm, which is in accordance with 0.6 to 8.0 cm/cm as recommended for omnivorous sh (Rotta 2003;Ferreira et al. 2014).
The absorption of amino acids, monosaccharides and fatty acids is performed by the proximal intestine (PI), while the absorption of macromolecules by pinocytosis occurs in the distal intestine (DI). This characteristic corroborates what was observed in this study. Mucosa, submucosa and muscle layers of the villi of the pi varied among sh fed with initial diets and fed with the diets of the nal period. Tambaqui fed with 15.96 kg/kg of lysine showed statistically greater cell density among the sh fed with the experimental diets, indicating a greater ability to absorb dietary lysine. The data from this study validate the information presented in the performance for feed conversion, which expressed the best response with 15.96 g/kg of lysine. The middle intestine (MI) showed alterations in the submucosa, but it was not possible to associate this variation with the diet.
In the intestinal segments of the tambaqui, the collagen shown in the cell layers corroborates the description by Honorato et al. (2013) for the intestine of nile tilapia. Collagen is a synthesized protein and essentially composed of the amino acids glycine, proline and lysine. Thus, diets with high levels of lysine tend to increase the synthesis of this protein, which in insu cient quantities can limit animal growth. The mucin protein secreted by goblet cells was more active in the pi of sh fed 22.20 g/kg of lysine in the diet.
Acid mucins are in uenced by the type of diet and form barriers against bacteria and agents that limit absorption (Rocha et al. 2016).
The height and perimeter of the villi of the PI did not vary according to diet, in contrast, the villi of the DI showed great variability, which made it impossible to associate these aspects with the levels of dietary lysine. In the MI, the greater villus height shown in sh fed with 9.72 g/kg dietary lysine was equivalent to that of sh fed with 6.60 g/kg dietary lysine. This suggests a strategy for increasing nutrient uptake in diets with insu cient levels of lysine. In the MI, the greater villus height shown in sh fed with 9.72 g/kg dietary lysine was equivalent to that of sh fed with 6.60 g/kg dietary lysine.
This suggests a strategy for increase the uptake of nutrients in diets with insu cient levels of lysine. Diets with 6.60 g/kg of lysine were prepared without the inclusion of l-lysine. diets with 9.72 g/kg of lysine have 4.00 g/kg of L-lysine in total lysine. Competition at the absorption sites between intact lysine and L-lysine, or the action of adverse variables on sh metabolism, may have caused a greater imbalance at the level of 9.72 g/kg, which generates greater urgency in the capture of nutrients. According to Rotta et al. (2003), crystalline amino acids are absorbed more slowly. However, Nguyen and Davis (2016) found no difference in the performance of channel cat sh, or american cat sh Ictalurus punctatus (Ra nesque 1818) and nile tilapia Oreochromis niloticus when fed with L-lysine and with intact lysine.

Plasma biochemistry
The protein content in the blood tissue of tropical sh in sh farming is between 2.3 and 8.2 g/dL. Thus it possible to assume the values shown in this study to be normal (Tavares-Dias and Moraes, 2003; Tavares-Dias and Mataqueiro 2004). The experimental diets were rich in amino acids, which are structural components of proteins, thus we expected an increase in plasma protein concomitant with the increase in dietary lysine levels. However, blood data showed a higher protein concentration in sh fed with 15.96 g/kg of dietary lysine, with an increase of 56.8% in relation to tambaqui not fed with experimental diets (p= 0.010). Adesola et al. (2017) investigated the lysine requirement for juvenile african dusky kob Argyrosomus japonicus (Gri ths and Heemstra 1995), and observed no differences for total proteins, triglycerides, glucose and cholesterol. Similarly, the same occurred in the current study in regards to triglycerides and glucose, which showed no difference, however, a difference was observed among plasma cholesterol concentrations in the different groups.        a-c. Photomicrograph of the hepatopancreas of tambaqui (Colossoma macropomum) fed with 9.72 g/kg of dietary lysine: a) multifocal presence of lipid droplets (black arrowhead) in hepatocytes near blood vessels. hepatocytes with polyhedral shape, irregular round and hexagonal, with centralized nucleus. b) lipids in the portal space (black arrowhead) and pancreas (white arrowhead). c) hydropic degeneration characterized by cell edema, cord breakdown and hepatocytes with clear cytoplasm (black arrowhead).