True Metabolisable Energy and True Amino Acid Availability in Chinese Varieties of Dehulled and Hulled Soybean Meals Determined with Adult Roosters *

Studies were conducted with intact White Leghorn roosters to determine the true metabolisable energy (TME) and the true amino acid availability (TAAA) in five dehulled and hulled soybean meals produced in China. 60 roosters, kept in individual cages, were fasted for 48 h and then tube-fed 50 g of one of experimental feedstuffs and their excreta was then collected for the subsequent 48 h period. Two separate collection periods were used with each meal being fed to 12 roosters. The birds were given a 15-day recovery period between collection periods. An additional 12 roosters were either fasted or fed a protein-free diet in order to estimate the extent of endogenous losses of energy and amino acids in excreta. The average values per bird for nitrogen loss, endogenous energy losses (EEL) and endogenous energy losses corrected to zero-nitrogen balance (EELn) were found to be 0.74 g, 47.0 kJ and 21.5 kJ, respectively. It was found that the TME and TAAA values of dehulled soybean meal were higher than those of hulled soybean meal. The TME and nitrogen-corrected TME metabolisable energy values of dehulled soybean meal were 10.58 and 10.74 MJ/kg, respectively, while the corresponding values for hulled soybean meal were 10.03 and 10.27 MJ/kg, respectively. The average indispensable and dispensable amino acid availability of dehulled soybean meal was 92.1 and 93.5%, compared with 89.3 and 91.4% for hulled soybean meal. Dehulled soybean meal would therefore appear to be superior to hulled soybean meal as a source of protein and energy for use in poultry rations. (Asian-Aust. J. Anim. Sci. 2003. Vol 16, No. 10 : 1487-1494)


INTRODUCTION
developed the true metabolisable energy (TME) bioassay in which excreta energy from a test ingredient is corrected for endogenous energy excretion measured using an accompanying fasted bird.Many studies had been conducted to evaluate the accuracy of this bioassay and to measure the TME of different diets and feed ingredients using adult roosters (Askbrant, 1988;Farrell, 1981;Han and Parsons, 1990;McNab and Blair, 1988).
A series of experiments by Sibbald (1979a) demonstrated that the amino acid excretion associated with different feedstuffs was independent of energy intake.Thus, the basic design of the TME bioassay can be used with confidence for estimating the true amino acid availability (TAAA) in feedstuffs.Studies evaluating the TAAA of feed ingredients using cecetomized chickens or intact chickens are also numerous (Sibbald, 1979ab;Parsons, 1985;Johns et al., 1986;Green et al., 1987ab).
Dehulled soybean meal is widely utilized in animal diets because the content of crude protein in dehulled soybean meal is higher (48%) and its crude fiber content lower than hulled soybean meal (NRC, 1994).However, there is a lack of information describing the TME and TAAA of dehulled soybean meals produced in China.Therefore, the purpose of the present study was to determine the TME and TAAA of different varieties of Chinese soybean meals in order to compare the quality of dehulled soybean meal with that of hulled soybean meal.

Feedstuffs and diets
Five paired-samples of dehulled and hulled soybean meals were obtained from various commercial sources across China.For a given location, the dehulled and hulled meals were initially drawn from the same batch of meal and then a portion of the meal was dehulled by the normal method used at that location.Three of the meals came from the East Ocean Cereal and Oil Chemistry Company located in Jiangsu Province, one from the Jinzhou Liulu Oil Industry Company located in Liaoning Province and one from Jilin Deda Company Limited located in Jilin Province.

Determination of True Metabolisable Energy (TME)
TME assays were conducted according to the method of Sibbald (1976) with several modifications.Sixty White Leghorn roosters (1.98±0.20 kg BW) were kept in individual cages (50.3×45.0×37.2cm 3 ) in a temperaturecontrolled room (18°C) under a daily light period of 16 h and with free access to water.
In order to test the ten soybean meal samples, two separate collection periods were used.During each collection, five samples of soybean meal were tested using 12 roosters per treatment.The five samples tested in the first collection period were randomly chosen from the batch of ten and the remaining five samples were tested during the second collection period.At the completion of each collection period, the birds were given a 15-day recovery period during which they were fed ad libitum a diet composed of 74.7% corn, 20.3% soybean meal and 5% of a vitamin-mineral premix.
Immediately before the start of a collection period, the birds were fasted for 48 h and then tube-fed exactly 50 g of one of the 10 soybean meals.During the subsequent 48 h, birds had free access to water and their excreta were collected.
The fecal collection apparatus consisted of an 80 mL rigid plastic container with a threaded cap.A 3 cm hole was cut through the center of the cap, and six equally spaced slots were cut through the lid around the periphery.Three silk threads of equal length were inserted through the slots.The silk threads were then taped onto the skin around the cloacae.In this way, the hole in the center of the cap could be located directly over the bird's vent.The collection cup was weighed before each collection and then threaded firmly onto the apparatus.
The collected feces were transported to the laboratory and stored frozen at -20°C until needed for analysis.The frozen excreta from each bird was freeze-dried for 24 h, allowed to come to equilibrium with atmospheric moisture for a further 24 h period, weighed and then ground to pass a 40-mesh sieve.Excreta from two birds from the same treatment were then pooled in order to obtain sufficient sample for analysis.The extent of replication was therefore six replicates per treatment.
Endogenous energy losses (EEL) and endogenous energy losses corrected to zero-nitrogen balance (EEL n ) were determined in a separate collection using 12 roosters with an average body weight of 2.1 kg.The individual birds were fasted for four consecutive days and excreta were collected daily during the last 48 h (Sibbald, 1981b) using the collection procedures previously described.The roosters had free access to water and received 50 g of glucose/bird/day via the water supply.
The AME and AMEn values of the soybean samples were calculated as follows: The TME and TME n values of the soybean samples were calculated as follows: Where GE=gross energy of meal (kJ/g), FI=feed intake (g), EE=excreta energy of fed birds (kJ), EEL=excreta energy of fasted bird (kJ), RN=N balance of fed birds (g)× 34.4 kJ, EELn=EEL+(-RN 1 ×34.4 kJ) and RN 1 =N balance of fasted birds (g).

Determination of True Amino Acid Availability (TAAA)
TAAA assays were conducted according to the method of Sibbald (1976).Sixty White Leghorn roosters (average body weight 2.1 kg) were fed and housed in the same manner as those used for the determination of TME.The TAAA of the ten samples of soybean meal were determined using two separate collection periods.Five samples of soybean meal were tested during each collection period using 12 roosters per treatment.At the completion of each collection period, the birds were given a 15-day recovery Crude fiber (%) 6.0 5.9 5.8 6.2 6.1 6.0 6.2 6.4 5.8 5.9 period during which they were fed ad libitum the same diet as that which was used in the recovery period for the TME determination.
Ten corn starch-based diets were formulated to contain approximately 16% crude protein using one of the ten test soybean meals as the sole source of crude protein and amino acids (Table 3).Vitamins and minerals were supplemented according to NRC (1994) standards.The diets were fed in meal form.
Immediately before the start of a collection period, the birds were fasted for 48 h.The birds were then tube-fed 50 grams of one of the soybean meal containing diets.During the subsequent 48 h, birds had free access to water and their excreta were collected using the same procedures as for the determination of TME.Again, excreta from two birds per treatment were pooled in order to have sufficient sample for analysis leading to a total of six replicates per treatment.
An additional twelve roosters (2.0 kg) were selected to determine the endogenous amino acid secretion using a highly digestible protein-free diet.The diet consisted of 45.5% cornstarch, 45.5% glucose, 5.0% fiber and 4.0% vitamin-mineral premix and contained 93.9% dry matter, 0.2% crude protein and provided a gross energy of 15.8 kJ per g air-dry.
TAAA values for the soybean meals were calculated as follows: Where AAC=amino acid consumed (g/day), AAE=amino acid excreted (g/day), EAA=endogenous amino acid excreted (g/day).

Chemical analysis
Analyses for dry matter, crude protein, ash, ether extract, calcium, and total phosphorus were assayed according to the methods of the AOAC (1990).Neutral detergent fiber was analyzed according to the principles outlined by Goering and Van Soest (1970) using an Ankom 220 Fiber Analyzer (Ankom Technology Corporation, USA).Gross energy was determined using an adiabatic oxygen bomb calorimeter (Gallenkamp).
For AA analyses, with the exception of the sulfurcontaining amino acids and tryptophan, the samples were hydrolyzed with 6 N HCl at 110°C for 24 h, then analyzed using an Automatic Amino Acid Analyzer (Hitachi L-8800, Japan).Methionine and cysteine were determined as methionine sulfone and cysteine acid after oxidation with performic acid.The oxidation process was carried out according to the AOAC (1990) method.The oxidized samples were then hydrolyzed and analyzed in the same manner as in the acid hydrolysis.Tryptophan analysis was carried out using high performance liquid chromatography (Shimadzu LC-10A, Japan) according method GB/T18246-2000 of the Chinese Quality Supervisory Bureau (2000).Protein solubility and urease activity were determined by the methods of Yang (1993).

Statistical analysis
Statistical analysis was performed by SPSS 6.0 software (SPSS Inc., USA).The TME and TAAA values of feedstuffs were subjected to Analysis of Variance (ANOVA).When ANOVA indicated a significant P value, means were separated using Duncan`s multiple range test (Duncan, 1955).Paired-sample tests were used to compare the metabolisable energy and true amino acid availability of dehulled soybean meal with hulled soybean meal.

RESULTS
The chemical composition of the ten soybean samples is shown in Table 1.The crude protein content of the hulled soybean meals averaged 45.82% (range 43.13 to 47.58%) while the protein content of the dehulled soybean meals averaged 48.64% (range 46.70 to 49.49%).As expected, the content of neutral detergent fiber was lower for the dehulled The amino acid content of the ten soybean meals is presented in Table 2.In general, the amino acid content of the dehulled soybean meals was higher than that for the hulled soybean meals.Of primary interest to nutritionists would be the contents of lysine, threonine and tryptophan.The lysine content of the hulled meals averaged 2.97% (range 2.88 to 3.06%) compared with 3.08% (range 3.05 to 3.14%) for dehulled meal.The threonine content of the hulled meals averaged 1.88% (range 1.77 to 1.95%) compared with 1.93% (range 1.87 to 1.97%) for dehulled meal.The tryptophan content of the hulled meals averaged 0.59% (range 0.54 to 0.63%) compared with 0.60% (range 0.55 to 0.67%) for dehulled meal.
Endogenous losses of energy and nitrogen for 12 fasted White Leghorn roosters are shown in Table 4.The weight loss of roosters during excreta collection (48 h) following 48 h of starvation ranged from 56 to 105 g with a mean of 72.7±7.0 g.Nitrogen loss of roosters was 0.74±0.17g.The EEL value of birds was 47.0±5.2kJ/bird/48 h.
The endogenous outputs of amino acids by roosters fed on the protein-free diet are also listed in Table 4. Indispensable amino acid loss averaged 19.0±4.0 mg, 22.3±6.9mg and 13.3±3.6mg for lysine, threonine and tryptophan respectively.
The nitrogen and energy balances of the roosters fed the ten soybean meals are shown in Table 5.These values were then used in calculating the AME, AMEn, TME, and TMEn  as well as the energy digestibilities presented in Table 6.
Because nitrogen retention was negative, the AMEn and TMEn values of feedstuffs were higher than the corresponding AME and TME values.The extent of fecal excretion of amino acids is presented in Table 7.These values were then used in the calculation of TAAA shown in Table 8.Table 9 summarizes the differences in AME, AMEn, TME, and TMEn, between the dehulled and hulled soybean meals.In general, the energy values for the dehulled meals were higher than for the hulled soybean meals.The AME and the respective AMEn values for the dehulled soybean meals were 9.36 and 9.80 MJ/kg while the corresponding values for the hulled soybean meals were 9.09 and 9.64 MJ/kg.The TME and the respective TMEn values for the dehulled soybean meals were 10.58 and 10.74 MJ/kg while the corresponding values for the hulled soybean meals were 10.03 and 10.27 MJ/kg, respectively.
The TAAA values for dehulled soybean meal were only modestly higher than those for hulled soybean meals (p>0.05).The indispensable and dispensable amino acid availability of the dehulled soybean meals averaged 92.1 and 93.5% while the corresponding values for hulled soybean meal averaged 89.3 and 91.4%.

DISCUSSION
In our studies, the average endogenous energy loss during the 48 h collection period was 47.4 kJ for the White Leghorn roosters, which agrees with the results of Yalcin and Onol (1994), Shires et al. (1980) and McNab and Blair (1988).In one study, endogenous energy loss in 48 h showed considerable variation from 32.6 to 82.0 kJ (Farrell, 1981) while others have shown less variation with endogenous energy excretion varying between 25.0 and 69.0 kJ (Sibbald and Morse, 1983;Sibbald and Wolynetz, 1984).
Endogenous nitrogen loss was extremely variable and averaged 0.74 g during the 48 h collection period (range 0.46 to 1.68 g).The average N loss, for the 2.1 kg birds used in the present study, was converted to a body weight based and was found to be 0.35 g/kg BW.This nitrogen loss  was lower than the values of 0.45 and 0.61 g N/kg BW obtained using starved roosters by Sibbald and Morse (1983) and Sibbald and Wolynetz (1984), respectively.
In the present study, because nitrogen correction was negative, the AMEn and TMEn values for the soybean meals were higher than the corresponding AME and TME values (Table 6).This emphasizes the importance of nitrogen correction of the obtained endogenous energy loss estimate.Nitrogen correction has been shown to have a profound effect on the amount and variability of the energy voided by starved birds and also on the variability of the AME and TME estimates (Sibbald and Morse, 1983;Sibbald, 1982).In the present study, nitrogen correction resulted in a 2.15-4.36%increase in the TME values of the soybean meals.
Our study showed that the ME values of dehulled soybean meal were higher than those of hulled soybean meal.The main reason for this difference is likely the variation in the chemical composition of these feedstuffs, especially in neutral detergent fiber.ME has been shown to be strongly correlated with the content of indigestible materials, such as fiber (Tenesaca and Sell, 1978;Farrell, 1981).The present study supports the conclusion that excreted energy output can be increased with an increase in the neutral detergent fiber content in the feed, thereby reducing the finial ME estimates.At the same time, hulls of soybean contain higher fiber and energy provided by hulls is not starch but availability fiber.However, hulls dilute the energy content of the test material and can inhibit digestive enzymes (Longstaff and McNab, 1991).ME values of hulled soybean meal were decreased because of this cause.
The endogenous outputs of amino acids in the present study were determined with adult intact roosters fed on protein-free diet.Some studies have shown that quantities of the endogenous outputs of amino acids excreted after feeding a protein-free diet significantly differed between intact and caecectomised birds (Kessler and Thomas, 1979;Johns et al., 1986;Green and Kiener, 1989;Parsons, 1992;Hou et al., 1997).Cecectomized roosters excreted higher amounts of amino acids and tended to have a lower amino acid digestibility than intact birds (Parsons, 1985;John et al., 1986;Green et al., 1987a).However, other studies had shown that the microbes in the cecum have little direct effect on the amino acid availability of high digestible feedstuffs when determined with either cecectomized or intact birds (Sibbald, 1979a;Green et al., 1989;Han et al., 1990;Hou et al., 1997;Ragland et al., 1999).The TAAA values for hulled soybean meal obtained in the present study were similar to values obtained using intact roosters (Ragland et al., 1999), but higher than values obtained using cecectomized roosters (Ragland et al., 1999).
Compared with hulled soybean meal, the amino acid availability of dehulled soybean meal was generally higher for most of amino acids.However, the improvement was not as dramatic as was the case for energy.Again the principle reason for the higher TAAA in the dehulled meals can be attributed to a reduction in fiber content of these meals.
In conclusion, the present study determined that the metabolisable energy and true amino acid availability content of dehulled soybean meals were higher than hulled soybean meal.As a result, dehulled soybean meal is a higher quality feedstuff for use in poultry feeds.

Table 3 .
Ingredient composition and chemical analysis of experimental diets used to measure true amino acid availability of dehulled

Table 1 .
Chemical composition (as fed) of dehulled and hulled soybean meals produced at various locations throughout China soybean meals that averaged 9.04% (range 7.93 to 9.92%) compared with 10.10% (range 9.07 to 11.92%) for hulled soybean meal.Protein solubility was higher for the dehulled soybean meals, which averaged 79.22% (range 72.04 to 83.31%), than for the hulled soybean meals, which averaged 77.34% (range 63.90 to 81.92%).

Table 2 .
Amino acid content (as fed) of dehulled and hulled soybean meals produced at various locations throughout China 1 Values based on duplicate amino acid analysis.

Table 4 .
Endogenous 2 Chemical analysis conducted in duplicate.

Table 5 .
Nitrogen and energy balance of roosters fed dehulled and hulled soybean meals produced at various locations throughout China a-c Means in each row with no common superscript differ significantly (p≤ 0.05). 1 Based on 6 observations during excreta collection (48 h). 2 All values are presented on an air-dried basis.

Table 6 .
Apparent and true metabolisable energy values (MJ/kg) of dehulled and hulled soybean meals produced at various locations a-c Means in each row with no common superscript differ significantly (p≤0.05).

Table 7 .
Feces weight and amino acid content from roosters fed diets containing dehulled and hulled soybean meals produced

Table 8 .
True amino acid availability of dehulled and hulled soybean meal produced at various locations throughout China

Table 9 .
Comparison of metabolisable energy, nutrient digestibility and true amino acid availability of dehulled and hulled soybean