Intake, digestibility and nitrogen balance in sheep fed diets containing detoxified castor cake1

The effect of substituting soybean meal (SM) with four levels (0, 33, 67 and 100%) of detoxified castor cake (DCC) on intake, nutrient digestibility and nitrogen balance was evaluated in uncastrated, male, crossbred sheep in a randomised complete block design with four treatments and five replications. There was a decreasing linear effect from the levels of substitution on dry matter (DM), organic matter (OM) and crude protein (CP) intake, total carbohydrates (TC), nonfibrous carbohydrates (NFC) and total digestible nutrients (TDN), and on DM, OM and TC digestibility, for reductions of 2.04, 1.70, 0.35, 1.56, 1.89 and 1.67 g dia-1 and 0.34, 0.53 and 0.51 g kg-1 DM respectively, for each 1% SM substituted with DCC. DMI and NDFap expressed, in %BW and g kg-0.75, showed a quadratic response, with a maximum value of 3.85 and 1.73%BW, and 95.15 and 42.47 g kg-0.75, for 25.00, 65.00, 20.75 and 61.50% SM substituted with DCC respectively. There was a linear increase in the intake of acid detergent fibre and in ether extract digestibility. The nitrogen balance was not affected by the substitution of SM with DCC. Substituting soybean meal with castor cake detoxified by autoclaving in lamb diets alters nutrient intake and digestibility without affecting the nitrogen balance.


INTRODUCTION
In semi-arid regions, the exploitation of sheep represents one of the principal sources of animal protein for human consumption, and is an activity of worldwide socio-economic importance. However, historically, sheep herds in these regions have been subject to food deficiency, mainly due to the prolonged period of drought. This scenario is likely to worsen as climate change forecasts for the coming years lead to more intense and constant extremes of climate, such as droughts and floods.
The lack of planning on the part of producers, and the low use of available technology to circumvent these limitations, make animal farming more vulnerable to the extremes of climate expected in the future. Among existing alternatives, sheep confinement has aroused interest for intensifying the production system, with the aim of minimising the impact of forage shortages, leading to a reduction in the age at slaughter, improving carcass quality, and maintaining the meat supply throughout the year to serve the domestic market (MEDEIROS et al., 2009). However, the lack of bulk during the dry season raises the costs of confinement due to the higher demand for concentrate.
Castor cake, a by-product of castor-oil (Ricinus communis L.) extraction, has the potential for use as a protein ingredient in feeding ruminants, despite several studies (COBIANCHI et al., 2012;GOMES et al., 2017;NICORY et al., 2015) showing that this by-product presents a wide variation in crude protein content (268.7, 357.4 and 380.0 g kg -1 DM). In addition, the presence of ricin in the kernel, one of the most potent toxic proteins known in the plant kingdom, severely disturbs digestion, and can lead to death if ingested by sheep in an amount equal to 1.25 g kg -1 body weight, according to Tokarnia et al. (2012). Because of this, Anandan et al. (2005), carried out research comparing the effectiveness of different methods of ricin detoxification of castor bean meal. Among the methods under evaluation, autoclaving at 1.23 kgf cm -2 or 15 psi at 123 ºC for 60 minutes completely destroyed the ricin.
In this context, evaluating the nutritional value of diets containing detoxified castor cake is essential for the safe use of this ingredient in sheep feed. As such, it is important to determine intake, nutrient digestibility and nitrogen balance in sheep fed on diets containing this byproduct.
The the aim of this study was to evaluate the influence of substituting soybean meal with detoxified castor cake on intake, nutrient digestibility and nitrogen balance in sheep.

MATERIAL AND METHODS
The study was carried out at the Centre for Forage Studies (NEEF), of the Department for Animal Science at the Centre for Agricultural Sciences of the Federal University of Ceará (UFC) in Fortaleza, in the state of Ceará. The city of Fortaleza is located on the coast, at 3º43'02" S and 38º32'35" W and an altitude of 15.50 m.
Four levels (0, 33, 67 and 100%) of soybean meal substitution with detoxified castor cake in lamb diets were evaluated in a randomised complete block design with five replications, using Tifton 85 hay (Cynodon sp.) as bulk. The castor cake was obtained from the Fazenda Normal Farm via the Institute for Sustainable Development and Renewable Energy (IDER), in the municipality of Quixeramobim, Ceará, by mechanical extraction (pressing) of the seed oil at temperatures of between 90 and 100 ºC, and detoxified at Embrapa -Agroindústria Tropical, by autoclaving (Sercon model HAE23 autoclave) at a pressure of 1.23 kgf cm -2 or 15 psi at 123 ºC for 60 minutes, as per Anandan et al. (2005). As the castor cake was found to be contaminated with the husk during seed processing, eight samples of the crude material (seed and husks) were randomly taken, fractionated and weighed to obtain the proportion of husk in the castor cake, before the seed oil was mechanically extracted.
The experimental diets were formulated based on the recommendations of the National Research Council (2007), and contained 135.2 g kg -1 DM crude protein and 659.2 g kg -1 DM total digestible nutrients, with a bulk to concentrate ratio of 50:50. The Tifton 85 hay was prepared at approximately 50 days of age.
Twenty crossbred sheep were used (½ Morada Nova red x ½ of no defined racial pattern), male, uncastrated, with a mean weight of 37.9 ± 4.4 kg and age of 11.5 months. The animals were housed in individual cages equipped with faeces and urine collectors and separators, as well as troughs for food, and drinking fountains with water available at will. The animals were weighed at the beginning and end of the experimental period and received an injectable supplement of vitamins A, D and E before the start of the experiment.
The experiment lasted 21 days: fourteen days for the animals to adapt to the diets and experimental environment, and seven days for collecting the supplied food, leftovers, faeces and urine. The experimental diet was provided daily in two meals, at 0800 and 1600, with the leftovers collected the following day. These were weighed and monitored, to remain at around 15%. The chemical composition of the ingredients is shown in Table 1, and the proportion of ingredients and chemical composition of the total diets are shown in Table 2. Intake, digestibility and nitrogen balance in sheep fed with rations containing detoxified castor cake  Each morning, the food, leftovers and faeces were weighed and sampled. The samples were placed in marked plastic bags and stored at -10 ºC. At the end of the experiment, the samples from each animal were defrosted and homogenised, and a sample of approximately 300 g was removed and taken to the Animal Nutrition Laboratory of UFC, weighed and placed in a forced ventilation oven at 55 ºC to constant weight. The levels of dry matter (DM), crude protein (CP), neutral detergent fibre corrected for ash and protein (NDFap), acid detergent fibre (ADF), ether extract (EE) and ash in the samples were then determined as per techniques described in Silva and Queiroz (2002).
The total carbohydrates (TC) were obtained with the equation TC = 100 -(%CP + %EE + %ash), as per Sniffen et al. (1992). The non-fibrous carbohydrates were estimated from the expression NFC = 100 -(%CP + %EE + %NDFap + %ash). The levels of neutral detergent insoluble nitrogenous compounds (NDIN) and acid detergent insoluble nitrogenous compounds (ADIN) were estimated in the residue obtained from the NDF and ADF using the micro Kjeldahl procedure. The total digestible nutrient content (TDN) of the experimental diets was estimated from the equation TDN = DCP + DNDFap + (DEE x 2.25) + DNFC, representing respectively, digestible crude protein, digestible ether extract, digestible neutral detergent fibre corrected for ash and protein, and digestible non-fibrous carbohydrates (SNIFFEN et al., 1992).
From weighing and determining the DM and nutrient content of the food, leftovers and faeces, the nutrient intake and digestibility were obtained, and the digestibility coefficients of the DM, OM, CP, EE, NDFap, ADF, TC and NFC determined with the formula [(Nutrient intake in grams -amount of nutrient in the faeces in grams)/Nutrient intake in grams] x 100.
Urine samples were taken on the last day of collection from the animals housed in metabolic cages using collection funnels, which fed the urine to plastic bottles containing 20 mL of 1:1 hydrochloric acid solution (HCL). After collection, the containers containing the urine were properly weighed to determine the total volume produced. Aliquots of approximately 10% of the total volume were then removed, properly identified and stored at -5 °C for the nitrogenous compounds to be later quantified.
The nitrogen contained in the faeces (faecal N) and urine (urine N) was determined following a methodology described in Silva and Queiroz (2002). The nitrogen balance was obtained by subtracting the faecal N and urine N from the ingested nitrogen (N ingested ).
The data were submitted to analysis of variance and regression. The models were chosen based on the significance of the linear or quadratic coefficients using Student's t-test at 5% probability and on the coefficient of determination. The GLM procedure of the SAS software (SAS INSTITUTE, 2003) was used as an aid to the statistical analysis.

RESULTS AND DISCUSSION
A linear decreasing effect (P<0.05) was seen from the levels of soybean meal (SM) substitution with detoxified castor cake (DCC), estimated at 1482 and 1278 g day -1 for dry matter intake (DMI) and 1252 and 1082 g day -1 for organic matter intake (OMI) at substitution levels of 0 and 100% respectively ( Table  3). The decrease in DMI and OMI in the animals fed at the higher substitution levels was due to the lower digestibility of the dry matter and the probable lower palatability of the DCC compared to the SM, results that were corroborated by visual evaluation of the animals during confinement, which, at the highest replacement levels, left the concentrate in the trough, preferring to consume the bulk. In addition, DCC contains a large amount of ricinoleic fatty acid, which can reach 89% of the total fatty-acid composition of the cake (BOMFIM; SILVA; SANTOS, 2009), leading to a reduction in nutrient intake (SANTOS et al., 2011). Agreeing with this statement, Maia et al. (2010) noted a reduction in DMI when adding 5% castor oil to the diet of lactating crossbred goats. Furthermore, the higher levels of neutral detergent fibre corrected for ash and protein (NDFap) in the castor cake (453 g kg -1 DM) and the high percentage of crude protein linked to the NDF (157 g kg -1 DM) also contributed to this result, thereby affecting nutrient consumption.
When expressed as a percentage of body weight (%BW) and grams per unit of metabolic size (g kg-0.75), the DMI presented a quadratic response, with a maximum value of 3.85%BW and 95.15 g kg-0.75 for 25.00 and 20.75% DCC respectively. Considering the DMI for maintaining adult sheep recommended by the National Research Council (2007), of 53.2 g kg -0.75 , each of the diets met the requirements, and as such, there were no physical limitations on intake.
Crude protein intake (CPI) decreased linearly with the substitution of SM by DCC, with values estimated at 255.10 and 220.10 g day -1 at substitution levels of 0 and 100%. Despite the diets having been formulated to be isoproteic and isoenergetic, there were reductions in CPI in response to the lower DMI for the levels of DCC in the experimental diets, corroborated by the high correlation between DMI and CPI (r=0.98, p<0.0001). The preference of the animals for consuming bulk over concentrate at Intake, digestibility and nitrogen balance in sheep fed with rations containing detoxified castor cake the higher levels of substitution also contributed to this result.
Ether extract intake (EEI) in g day -1 was not influenced by the levels of SM substitution, with a mean value of 46.55g day -1 . This result is attributed to the higher EE content of DCC (61.0 g kg -1 DM) in relation to the SM (18.0 g kg -1 DM), which compensated for the lower DMI (g day -1 ). The DCC was obtained by mechanical extraction, which is less efficient than the solvent extraction used to obtain SM, and explains the greater EE content of DCC compared to SM. However, the greater EE content of the DCC was not enough to hinder digestion of the fibre, since each diet presented an EE content below the 6.0% limit, which, according to Van Soest (1994), causes a reduction in fibre digestion due to intoxication of fibrolytic ruminal microorganisms.
The intake of neutral detergent fibre corrected for ash and protein (NDFapI) was not influenced by the levels of DCC when expressed in g day -1 . Despite the reduction in DMI with the levels of DCC, there was an increase in the NDFap content, which resulted in there being no effect on the NDFapI. The increase in the NDFap content of the diets at the higher levels of SM substitution with DCC is due to the high NDFap content of DCC (453 g kg -1 DM) compared to that of SM (120 g kg -1 DM). When expressed in %BW and g kg -0.75 , a quadratic effect from the substitution levels on NDFapI was seen, with a maximum of 1.73 %BW and 42.47 g kg -0.75 when 65 and 61.5% of the SM was substituted with DCC respectively (Table 4). Furtado et al. (2012), testing different methods of detoxifying castor cake, found a NDFapI of 1.96 %BW in growing sheep consuming diets containing 67% autoclaved DCC.
Acid detergent fibre intake (ADFI) increased linearly as the percentage of DCC in the total diet increased. Each percentage increase in DCC increased the ADFI by 0.557 g kg -1 DM. In this study, the greatest NDFap and ADF intake was seen in diets with a higher DCC content, and can be explained by the higher levels of these nutrients in DCC compared to SM, a result of the 15% castor bean husk present in DCC. Similar results to those obtained in the present study were reported by Ferreira et al. (2009), who found an NDFI greater than 2% body weight (BW) using diets containing agro-industrial by-products. Cunha et al. (2008) reported an NDFI equal to 2% BW when adding whole cottonseed cake to sheep diets. Therefore, despite the NDFap content of the diets having increased due to the higher percentage of DCC, other factors related to the characteristics of the DCC, such as palatability, may have influenced the behaviour of the animals, leading to a reduction in DMI as DCC was added to the diets.
The intake of total carbohydrates (TCI), nonfibrous carbohydrates (NFCI) and total digestible nutrients (TDNI) decreased linearly (P<0.05) as the SM in the diets was substituted with DCC. Each percentage point of DCC added to the diets reduced the TCI, NFCI and TDNI by 1.558, 1.890 and 1.668 g kg -1 DM. The lower NFC content and intake in diets containing a greater proportion of DCC are responsible for the reduction in TCI and TDNI.
There was a linear decreasing effect (P<0.05) for the levels of SM substitution with DCC on dry matter digestibility (DMD) and organic matter digestibility (DMD), with reductions of 0.34 and 0.53 g kg -1 DM for DMD and OMD with the addition of each percentage point of DCC to the diet. This result is attributed to the greater lignin content and the reduction in NFC content (high digestibility) of diets with higher levels of DCC. This relationship is corroborated by the results presented by Diniz et al. (2011), who, working with growing cattle, found no effect on DMD or OMD from substituting SM in the diet with detoxified castor meal, attributing this result to the similarity between the fibrous and non-fibrous carbohydrate content of the diets.
No effect was seen from the levels of SM substitution with DCC on crude protein digestibility (CPD), with a mean value of 766.33 g kg -1 DM. This was not expected, as the acid detergent insoluble nitrogen (ADIN) content of the SM was 37 g kg -1 TN, while the DCC showed an ADIN content of 133 g kg -1 TN, which explains the lower CPD at the higher levels of DCC. However, it is possible that the presence of a greater proportion of urea in diets with a higher percentage of DCC was enough to stimulate microbial action on the ingested diets, compensating for the limitations on CPD of the high levels of ADIN in the DCC.
An increasing linear effect (P<0.05) was seen for the levels of SM substitution with DCC on ether extract digestibility (EED), with estimated values of 838.6 and 968.8 g kg -1 at substitution levels of 0 and 100% respectively. This result is attributed to the larger EED of the DCC, which contains a large proportion of polyunsaturated fatty acids, especially ricinoleic acid, which can represent up to 90% of the total fatty acids present in castor cake compared to SM. This statement is corroborated by Borja et al. (2017), who also found an increase in EED when including castor bean meal, detoxified by a mixed combination of calcium oxide and autoclaving, in the diet of growing sheep.
There was no effect from the levels of SM substitution with DCC on the digestibility of NDFap or ADF, with a mean value of 554.70 and 507.35 g kg -1 DM respectively. This result was unexpected, given that the NDF and ADF content of the DCC were 47.99 and 40.23% respectively, against the 13.28 and 9.73% of the SM. However, the addition of urea may have optimised the amount of RDP and RNDP in the diet, favouring the growth of fibrolytic microflora, improving the digestion of digestible NDF, and increasing the passage rate of indigestible NDF, with a possible improvement in the digestibility of these nutrients (TEDESCHI; FOX; RUSSEL, 2000).
A linear reduction (P<0.05) was seen for the levels of SM substitution with DCC on total carbohydrate digestibility (TCD), estimated at 689.94 and 638.46 g kg -1 at substitution levels of 0 and 100% respectively, which Nitrogen intake (N intake ) expressed in grams per day (g day -1 ) and grams per kg metabolic weight (g kg -0.75 ) decreased linearly with the substitution of SM with DCC (Table 5). Each percentage point of SM substitution reduced N intake by 0.056 g day -1 and 0.0034 g kg -0.75 respectively. This result is due to the reductions in DMI at the higher levels of DCC, and is confirmed by the high positive correlation between DMI and N ingested expressed in g day -1 (r=0.98, p<0.0001) and g kg -0.75 (r=0.96, p<0.0001). Similarly, Palmieri et al. (2016) evaluated SM substitution with DCC, detoxified with calcium oxide, in growing goats, and saw a reduction in N ingested resulting from the effect of the DCC in reducing DMI and CPD.
Faecal N expressed in g day -1 and percentage of N intake (%NI) was not influenced by the substitution of SM with DCC, with mean values of 8.89 g day -1 and 23.37 %NI. When expressed in g kg -0.75 , there was quadratic response from the levels of DCC on faecal N, with a maximum value of 0.63 g kg -0.75 for 38.33% DCC. Furtado et al. (2014), evaluating castor cake not treated by detoxification or treated by different methods, in diets for growing sheep, also found no changes in faecal N content, with a mean value of 33.06 %NI.
Urine N decreased linearly with increasing levels of DCC in the diet, with variations of 13.15 and 9.15 g day -1 between the levels of 0 and 100% SM substitution with DCC. Despite the lower intake of N ingested at the higher DCC levels, there was no change in urine N expressed as %NI. This indicates that there was no increase in the loss of nitrogenous compounds via the urinary tract when SM was replaced with DCC. A positive correlation between N intake and urine N (r=0.64, p=0.0024) was found in the present study, and explains the probable decrease in urine N as a result of N intake , which was also reduced at the higher levels of DCC. Losses of urine N were higher than of faecal N, with mean values of 29.67 and 23.37 %NI respectively. The intake of dietary nitrogen was probably higher than that required to meet the demands of adult sheep, resulting in an excess of ammonia in the rumen environment and its later absorption by the rumen mucosa, which increased N excretion by the urine in the form of urea (LAVEZZO; LAVEZZO;BURINI, 1996).
The nitrogen balance was not influenced by the diets containing different levels of SM substitution, with mean values of 17.75 g day -1 , 1.17 g kg -0.75 and 46.96 %NI. It is worth pointing out that despite the reduction in N intake at the highest levels of DCC, the nitrogen balance was positive for all levels of substitution, showing that the nitrogen intake met the requirements of the animals for nitrogenous compounds (FURTADO et al., 2014) and caused no changes in the balance between non-protein nitrogen and true protein in the diet (SILVA et al., 2016).

CONCLUSIONS
Castor cake, detoxified by autoclaving, as a substitute for soybean meal in sheep diets, alters nutrient intake and digestibility without affecting the nitrogen balance.