Introduction

Canavalia ensiformis is a legume high in protein also known as Jack bean, Snake bean, Horse bean, White bean, Creole bean. It is scattered in the tropics and subtropics across the world, but it is not cultivated on a large scale.

It is a promising crop in Yucatán, México, since in an agronomical trial on stony soil (cultivated seasonally, without fertilization) a maximum yield of 1.9 tons / ha was obtained, which is more than the obtained with corn or Phaseolus vulgaris under the same conditions ^{(Herrera, 1983)}. This makes it an alternative as a feedstuff for farm animals.

The utilization of raw Canavalia, as any other legume, is risky since it as antinutritional components limiting its use as a feed ingredient for humans and livestock, particularly monogastrics. Besides having non absorbable phosphorous in form of phytic acid, Canavalia seeds have cyanogenic glycosides, tannins, proteases inhibitors, lectins, L-canavanine, urease, concanavalin A and B ^{(Rajaram and Janardhanan, 1992)}, among other toxic substances. When Canavalia is administrated raw, representing between 250 to 500 g / kg in the diet of broilers, serious tissue damage is observed in pancreas, lungs, intestines, liver and kidney ^{(Ologhobo et al., , 2003)}.

Some available methods to eliminate toxicity are: soaking and heating (3 h at 95° C) the broken kernels ^{(Udedibie and Carlini, 1998)}; cooking at 100°C for 50 min ^{(Alagbaoso et al., 2015)}; roasting at 190 °C or more during 16 min ^{(Mendez et al., 1998}; ^{Michelangeli et al., 2004)}; steaming at 120 psi during 45 min ^{(Akande et al., 2013)} among others. Another treatment using heat between 180-200 °C, eliminates effectively the antinutritional factors in raw Canavalia and therefore enhances the energy value of the seed to 2.4 Mcal ME/kg ^{(Síboli et al., 2004)}, but this amount of energy is considered to be low. The success of all this treatments is based on the fact that most of this antinutritional factors are thermolabile; but some are thermostable like saponins and may render inefficient heat treatments and could help to explain the lack of success in trials such as one carried out with fattening pigs ^{(Emenike et al., 2016)}. In most of this experiments no information on protein quality value was investigated. Since the use of high heat may reduce indirectly the protein digestibility of a feedstuff, as well as the metabolic energy content due to Maillard reactions (Payle and Gerrard, 2002), it would be expected that a detoxifying method using low heat or no heat, may be more suitable to preserve the nutritional value of Canavalia.

There are very few methods that use low temperature or do not use temperature at all. The use of low heat in an extraction using an acidified saline solution proved to be efficient removing most of the antinutritional components of C. ensiformis^{(Gamboa et al., 1994)}: this method is simple to achieve as well as inexpensive. Another method that does not use temperature and also eliminated toxic substances of Canavalia is the extraction of a protein concentrate ^{(Betancur-Ancona et al., 2008)}, but no information concerning the protein digestibility and the metabolic energy content was provided in these reports either. Therefore the objective of the present work was to reduce the antinutritional factors and to determine the possible change in protein digestibility and True metabolic energy content of Canavalia ensiformis after detoxification using low heat or obtaining a protein concentrate, in order to incorporate it into poultry diets.

Materials and methods

Fifty kg of Canavalia ensiformis obtained from X'pujil community located in the state of Campeche, were used. Beans were grinded and submitted to two detoxification methods: treated in an acidified saline solution (salting-in phenomena) submitted to low heat (ASOL-H); the extraction of a protein concentrate (PC). Two experiments were carried out:

Experiment 1

To obtain the ASOL-H flour, a dispersion of 5 kg of Canavalia was prepared in 25 L of 0.5 M NaCl solution at a 1: 5 (w / v) (salting-in). The pH was adjusted to 4.9 with 2N HCl, the mixture was heated and stirred at 80°C for 60 min. The suspension was cooled at room temperature and then centrifuged at 2500 rpm for 10 min at 25°C in a Mistral centrifuge model 3000i (Sanyo-Gallen Kamp, Leicester, UK). The supernatant was discarded and the precipitate dried at 60° C for 12 h. ASOL-H obtained was weighed and ground in a mill through a mesh 20 grid.

The PC was obtained using the method described by ^{Betancur et al., (2004)} which consists in a humid fractionation of grinded seeds, alkaline solubilization and precipitation of protein at isolectric point. After the elimination of the supernatant, the precipitate was dried. The method was as following: a dispersion of 10 kg of Canavalia flour in 70 L water (1: 7 w / v) was prepared. The pH was adjusted to 12 with 2N NaOH and allowed to stand 12 h at room temperature. Subsequently the suspension was sieved through 80 mesh grid (0.177 mm pore opening), to separate the solid with the fiber and the liquid with the protein and starch. The fiber residue was discarded. The supernatant was passed through a 150 mesh grid (0.106 mm pore opening) to remove the finest fiber. The suspension was allowed to stand for 2 h until complete sedimentation of starch. Solubilized protein in the supernatant was siphoned.

Proximal analysis in raw, ASOL-H and PC were carried out using ^{A.O.A.C. (2012)} methods. Antinutritional compounds were also estimated. A synthetic description of the methodologies utilized is described.

Total phytates ^{(Makkar et al., 2007)}. Extraction was carried out with 3.5% (w/v) HCl and further purified through an AG1-X8 chloride anion exchange column. The pink color of Wade reagent, previously prepared, was due to the reaction between ferric ion and sulfosalicylic acid with an absorbance maximum at 500 nm. In the presence of phytate, the iron became bound to the phosphate ester and was unavailable to react with sulfosalicylic acid, resulting in a decrease in pink color intensity

Cyanogenic glycosides ^{(A.O.A.C., 1990)}. Cyanide was distilled from a chloroform and water solution into KOH solution, forming KCN which was titrated with silver nitrate; 2KCN+AgNO₃ = KCN·AgCN+KNO₃. An excess of AgNO₃ produced insoluble AgCN, which was the end point of the titration: Ag(CN)₂K + AgNO₃ = 2AgCN+KNO₃

Tannins ^{(ISO-9648, 1988)}. The method consists in their extraction by shaking with dimethylformamide, then after centrifuging, addition of ferric ammonium citrate and ammonia to an aliquot part of the supernatant liquid and spectrometric determination, at 525 nm, of the absorbance of the solution thus obtained, and determining of the tannic content using a calibration curve prepared using tannic acid. This technique measures total phenols and tannins.

Trypsin inhibitors activity ^{(Kakade et al., 1974)}. It was measured using BAPNA (Benzoyl-L-arginine-D-p-nitroanilide) as a substrate and the results reported as units of trypsin inhibited (UTI). A trypsin unit (UT) is defined as the increase in 0.01 absorbance units at 420 nm per 5 ml reaction mixture under the experimental conditions.

Lectins ^{(Jaffe et al., 1974)}. Hamster red blood cells were used, washed and senzibilized with trypsin using a microdilutor, to take into account the number of times that the samples were diluted.

Canavanine ^{(Makkar et al., 2007)}. The guanidoxy group, -O NH C (:NH) NH₂, in canavanine reacted with trisodium pentacyanoammonioferrate in aqueous solution at pH 7.0 and formed a magenta-colored chromophore. The color formed was measured at 520 to 530 nm using as a reference a standard curve with canavanine (reagent grade) in order to calibrate the spectrophotometer.

Experiment 2

Protein digestibility was evaluate in ASOL-H and PC using the multienzyme method described by ^{Hsu et al., (1977)}.

Finally, a bioassay was carried out to estimate the true metabolic energy value (TME) of Canavalia using nine Leghorn roosters ^{(Sibbald, 1976)}. The TME was estimated subtracting the excreted energy to the consumed energy. Treatments were as follow: experimental diet with ASOL-H, incorporated in the diet at 25, 30 and 35% of the protein requirements of the birds ^{(NRC, 1994)}. Experimental diet with PC, incorporated the same way as the previous one. Control diet. All diets were made with ground corn, soybean meal, wheat bran, vegetable oil and they were added with premix of vitamins and minerals, calcium carbonate, Ca orthophosphate, lysine, methionine and salt, to meet the nutritional requirements of growing broilers ^{(NRC, 1994)}.

Statistical analysis

Means and standard deviation of data were estimated in both experiments. One-way ANOVA method was applied to verify for significant differences between the treatments where applicable, followed by Tukey’s Test ^{(SAS, 1988)}.

TME value was analyzed according to a completely randomized design by means of two Latin Squares 3x3, one for ASOL-H and another for PC, each one with three replicates. The model was:

Y_ij(k) = µ + Row_i (animals) + Col_j (periods) + t_(k) (treatments) + E_ij(k)(random error)

Analyses were performed with the Means, ANOVA and GLM routines of the SAS statistical software ^{(SAS, 1988)}.

Results

Experiment 1

Analysis of chemical composition of raw and treated Canavalia seeds are shown in Table 1. Treated ASOL-H had a lower amount of protein (P<0.05) compared to raw Canavalia since some part of it (albumin) is water soluble. PC obviously had the highest amount of protein (P<0.01). Changes in all other constituents fluctuated as expected.

Table 2 shows the analysis of antinutritional factors in raw and treated Canavalia. A significant reduction in all the antinutritional factors was obtained with the treatments used. Maximum reduction (P < 0.01) was observed with cyanogenic glycosides, trypsin inhibitors, Canavanine and lectins. A minor effect, but also important was obtained with phytates and tannins.

Experiment 2

Results on the protein quality and energy value of Canavalia are shown in Table 3. Protein digestibility was improved in PC (P < 0.05) compared to raw Canavalia; no changes were observed in ASOL-H compared to raw Canavalia. No statistical difference was detected in the energy content between raw vs treated seeds (P > 0.05).

Discussion

Raw Canavalia seeds had a high proportion of crude protein (31.4 %, Table 1), similar to grains analyzed by ^{Bramiro (1994)} (29.7 %). Such an elevated protein content of the grains used in this experiments was also observed in the PC (79.0%). When the protein recovery out of the total protein content in the grain is above 70%, the extraction method is considered to be efficient ^{(Cantoral et al., 1995)}, therefore offers a possible economic success if the method is escalated. The percentage of nitrogen free extract had the exact opposite tendency to the one observed with crude protein. All other constituents did not show important changes, except crude fiber, which as expected, was elevated in raw Canavalia. Some authors have found a potential use of these fiber residues in the food industry as a functional ingredient in reduced-calorie food ^{(Segura et al., 2014)}.

The reduction in antinutrients as a percentage showed better efficiency in PC compared to ASOL-H (Table 2). Reduction was very efficient (P < 0.01) in canavanine (99 % and 98 %), trypsin inhibitors (92 % and 84 %) and cyanogenic glycosides (74 % and 100 %) for PC and ASOL-H respectively. Also efficient, but in a lower degree (P <0 .05) were pythates (48 % and 43 %) and tannins (38 % and 27 %) for PC and ASOL-H respectively. Finally lectins were only reduced significantly (P < 0.01) in PC treatment.

Roasting, boiling and fermentation are other alternative methods used to reduce antinutritional factors of Canavalia, providing different efficiencies. According to ^{Ajeigbe et al., (2012)} mild roasting even increases phytate concentration but boiling and fermentation reduce drastically cyanide (87 % and 88 %) and tannins (93 % and 92 %) respectively. Comparing this results with the present study, ASOL-H and PC treatments where more efficient to eliminate phytates (34% and 48%), cyanide (100 % and 74 %) but less efficient to reduce tannins (27 % and 38 %) respectively.

The use of soaking and autoclaving of raw Canavalia^{(Sasipriya and Sidhuraju, 2013)} is less efficient to reduce the toxic non proteic α-aminoacid canavanine (61%), compared to ASOL-H and PC, where reductions of 98% and 99%, respectively were reached. Trypsin inhibitors were efficiently removed by all three methods: PC, soaking and autoclaving: 92 %, 84 % and 99 % respectively.

Previous reports indicate that toasting Canavalia at 190°C during 16 min reduced significantly the amount of Concanavaline A (Con-A), a lectin found in Canavalia; broiler chicks may consume up to 100 mg of Con-A daily with no deleterious effects on growth ^{(Mendez et al., 1998)}. Lectin reduction in the present study, was only efficient in PC (88%) therefore this treatment may be suitable to be use when Canavalia is considered to be an ingredient to poultry diets. Nevertheless lectins nowadays are considered important because of their bio-functionally due to involvement in immune mechanisms, antibacterial and anticancer mediation where their presence has allowed the explanation of mechanisms of action at a molecular level ^{(Singh and Sarathi, 2012)}. Therefore, a better way to comprehend the use of this molecule is necessary in order to incorporate it into poultry diets.

Protein digestibility was increased (P < 0.05) in the PC in comparisons to raw Canavalia where ASOL-H did not showed any deleterious effect (P>0.05) (Table 3). Similar results have been found by other authors, as for example ^{Ly et al, (2015)} who reported that soaking Canavalia in water during 24 h and then autoclaving it during 60 min increased the in vitro N digestibility from 72.1% to 89.1%. Preserving protein quality is very important, since it may have a positive impact on animal performance. The use of toasting at increasing temperatures of more than 200°C, reduced protein solubility of Canavalia as it may be expected when high temperatures are used ^{(Pizzani et al., 2006)}.

The TME value of Canavalia was not affected by treatments (P > 0.05) (Table 3), but this statistical approach must be taken carefully. There was an arithmetical difference of 0.245 Mcal between Raw Canavalia and ASOL-H; such a difference was only about half (0.141 Mcal) between Raw Canavalia and PC. This difference of 0.245 Mcal is important in poultry diets since it may improve substantially the productivity in growth rate ^{(NRC, 1994)}. Energy content was found similar or higher in PC (2.999 Mcal) compared to other conventional ingredients of poultry diets, such as fish menhaden (2.977) ^{(NRC, 1994)}, or peanuts kernels (2.462) ^{(NRC, 1994)}; but not as elevated as in corn (3.470) ^{(NRC, 1994)} a classical energetic source for poultry. Both treatments tested, ASOL-H and PC, rendered higher energy values (2.895 and 2.0999 Mcal respectively) than the one obtained by ^{Siboli et al., (2004)} with birds (2.536 Mcal). Other authors ^{(Fagbenro et al., 2004)} have found no change in the energy value of Canavalia fed to Nile tilapia by means of the elimination of antinutrients cooking cracked seeds at 100°C for one hour. Therefore animal specie may be important to be considered when Canavalia is detoxified.

Conclusions

Detoxification of Canavalia ensiformis may be partially achieved treating it with a solution of salt and acid with low heat; most antinutritional factors were reduced 27 % or more, except lectins. The extraction of a Canavalia protein concentrate eliminates more efficiently all these toxic agents. Protein digestibility was improved in the protein concentrate and true metabolic energy value was not negatively affected by the use of both methods of detoxification. The reduction of antinutritional factors and the preservation of protein and energy value using either of both treatments, may facilitate the incorporation of Canavalia into poultry diets.