Animal Feed Science and Technology 162 (2010) 58–65
Contents lists available at ScienceDirect
Animal Feed Science and Technology journal homepage: www.elsevier.com/locate/anifeedsci
Effects of citric acid, alpha-galactosidase and protease inclusion on in vitro nutrient release from soybean meal and trypsin inhibitor content in raw whole soybeans夽 T. Ao, A.H. Cantor ∗ , A.J. Pescatore, J.L. Pierce, K.A. Dawson Alltech-University of Kentucky Nutrition Research Alliance, Lexington, KY 40546, USA
a r t i c l e
i n f o
Article history: Received 5 July 2009 Received in revised form 24 August 2010 Accepted 31 August 2010 Keywords: ␣-Galactosidase Protease Citric acid Reducing sugar ␣-Amino nitrogen
a b s t r a c t An experiment was conducted to evaluate the effects of citric acid (CA) and commercial preparations of ␣-galactosidase (␣-GAL) and protease on in vitro nutrient release from soybean meal (SBM) and trypsin inhibitor (TI) content in raw defatted whole soybeans. An in vitro model was used to simulate the chicken’s digestive process in the crop, the stomach (proventriculus and gizzard) and the small intestine. Soybean meal and ground whole soybeans were used as substrates. Graded levels of either ␣-GAL (0–13,792 units/kg) or protease (0–888 units/kg) and 0 or 20 g CA per kg were added to the substrates in a factorial arrangement. Reducing sugars (RS) and ␣-amino N were measured at the end of the crop phase, the stomach phase, and the whole phase (crop through small intestine). Trypsin inhibitor content was measured at the end of the stomach phase. Increasing ␣-GAL levels linearly (P<0.01) increased the release of RS in both the crop and the whole phases. Addition of CA with ␣-GAL further increased enzyme activity, resulting in a significant interaction (P<0.01). Linear increases (P<0.05) in ␣-amino N occurred with increasing doses of protease at the crop, the stomach and the whole phases. Inclusion of CA in the incubation mixture also increased (P<0.05) ␣-amino N at all three phases. An interactive effect (P<0.01) between protease and CA was observed only in the crop phase. Neither protease nor CA had an effect on TI activity in raw soybeans. The results suggest that ␣-GAL and protease may be useful in hydrolyzing the carbohydrates and proteins, respectively, in SBM and the effects of these two enzymes may be enhanced by acidifying the diet. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Alpha-galactosidase (␣-d-galactoside galactohydrolase, EC 3.2.1.22) and proteases are useful enzymes in improving the nutritional value of SBM. Several studies (Gitzelmann and Auricchio, 1965; Marsman et al., 1997; Graham et al., 2002) showed that ␣-GAL can help monogastric animals, such as poultry and pigs, to digest galacto-oligosaccharides (e.g., stachyose and raffinose) contained in SBM that cannot be digested by these animals due to absence of endogenous enzymes with ␣-(1,6)GAL activity. Treatment of SBM with exogenous protease to inactivate proteinaceous anti-nutritional factors, such as TIs is
Abbreviations: ␣-GAL, ␣-galactosidase; CA, citric acid; N, nitrogen; SBM, soybean meal; SB, soybean; TI, trypsin inhibitor(s); RS, reducing sugar; GIT, gastrointestinal tract. 夽 Paper Number 09-07-070 of the Kentucky Agricultural Experiment Station. ∗ Corresponding author. Tel.: +1 859 2577531; fax: +1 859 3231027. E-mail address:
[email protected] (A.H. Cantor). 0377-8401/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.anifeedsci.2010.08.014
T. Ao et al. / Animal Feed Science and Technology 162 (2010) 58–65
59
a potential method for improving the nutritional value of SBM (Classen et al., 1993; Verstegen and Voragen, 1997). Protease can also help young animals to digest the large storage protein molecules in SBM (Huisman and Tolman, 1992; Sheppy, 2001). The optimal pH level of fungal ␣-GAL is around 4.5–5.0 (Ademark et al., 2001; Ao et al., 2008) and that of protease is 3.0 (Committee on Food Chemicals Codex, 1996). The pH of a typical maize-SBM-based broiler diet and of the digesta in the crop is approximately 6.0–6.5 (Ao et al., 2008). Moran (1982), citing Herpol and Van Grembergen (1967), indicated that the average pH values of luminal contents in the GIT of chickens (ranges shown in parenthesis) were 6.3 (4.0–7.8) in the crop, 1.8 (0.3–4.1) in the proventriculus, 2.5 (0.4–5.4) in the gizzard, 6.4 (5.2–7.6) in the duodenum, 6.6 (5.5–7.7) in the jejunum and 7.2 (5.7–8.2) in the ileum. Thus, the pH levels in the diet and in the crop and small intestine are higher than that needed for optimal activity of these enzymes. The addition of organic acids is known to lower dietary pH (Giesting and Easter, 1985). The objectives of this experiment were: (1) to evaluate the effects of acidification of in vitro digestion mixtures on the activities of ␣-GAL and protease using SBM or raw defatted soybeans as substrates and (2) to study the effect of protease on the activity of TI in the raw soybeans. 2. Materials and methods 2.1. Materials Commercial preparations of ␣-GAL (Experiments 1 and 2) and protease (Aspergillus niger, Alltech Inc., Nicholasville, KY) (Experiments 3–5) were used in these studies. Pepsin (P 7012) and pancreatin (P 3292) were purchased from Sigma Chemical Co., St Louis, MO. All other chemicals used were of analytical grade. Soybean meal containing 480 g crude protein per kg and raw whole soybeans were obtained from commercial suppliers. The soybeans were defatted using petroleum ether before being used in the in vitro digestion process. To do so, the soybeans were ground to a fine texture (200 mesh) and put in a filter bag. The bag was sealed and soaked into petroleum ether. The ether was changed every 3 or 4 h until it was clear. Then, the excess ether was removed and the bags were spread out in a fume hood for 48 h allowing ether to evaporate. Food-grade anhydrous CA was obtained from Roche Vitamins Inc., Parsippany, NJ. 2.2. Measurement of enzyme activity Activity of ␣-GAL was determined based on the method described by Ratto and Poutanen (1988). One GAL unit (GalU) was defined as the quantity of the enzyme that liberates -nitrophenol at the rate of 1 mol/min under the conditions of the assay. The analysed activity of ␣-GAL used in this study was 1724 GalU/g. Activity of protease was assayed according to the method from the Committee on Food Chemicals Codex (1996). One spectrophotometric acid protease unit is the amount of enzyme that liberates 1 mol of tyrosine per min from casein under the conditions specified. The analysed activity of protease was 111 units/g. 2.3. Experimental treatments Substrates, experimental factors, in vitro measurements, phases of digestion and statistical designs performed in Experiments 1–5 are presented in Table 1. Substrates included SBM (Experiments 1–4) and ground raw whole soybeans (Experiment 5). 2.4. In vitro digestion procedure and measurement of RS and ˛-amino N A modified in vitro digestion procedure described by Tervila-Wilo et al. (1996) and Zyla et al. (1999) was used in all of the experiments. The procedure is outlined in Fig. 1. After grinding through a 1 mm screen, triplicate samples of substrates weighing 2.5 ± 0.0001 g were placed in 50 ml centrifuge tubes. The samples were hydrated with 6 ml distilled water and 1 ml of enzyme solution (10 mg/ml). No adjustment of pH was made at this step, and, therefore, the resulting pH was that of the substrate. The tubes were vortexed, sealed with flexible plastic film, and then incubated without agitation in a water bath at 40 ◦ C for 30 min. The first step simulated the digestion in the crop. If only the crop phase digestion was considered (Experiments 1 and 3), 5 ml ice-cold distilled water was added to the tube to stop the hydrolysis. In Experiments 2, 4 and 5, the pH was adjusted to 3.0 by adding 1.5–2.0 ml of 1 M HCl plus sufficient water to keep the volumes equal in all tubes. Then 7500 units (0.5 ml) of pepsin solution were added to the tubes, which were then vortexed, sealed and reincubated for 45 min at the same temperature. This step simulated digestion in the proventriculus and gizzard of chickens. If only the crop through gizzard digestion was considered (Experiment 5), 3 ml ice-cold distilled water was added to the tube to stop the reaction. In Experiments 2 and 4, 3 ml NaHCO3 containing 4.63 mg pancreatin was added drop wise with constant stirring to each tube. The tubes were vortexed, sealed and further incubated for another 60 min. This step simulated digestion in small intestine. Once the digestion process was finished, the tubes were immediately put in an ice-cold water bath for 15 min to stop the further hydrolysis. Then, the tubes were centrifuged at 16,240 × g for 20 min. The supernatants (0.5 ml) were withdrawn and treated with 0.3 N barium hydroxide and 0.3 N zinc sulfate solutions to precipitate protein as described by Sonnenwirth and Jarett (1980). The deproteinization was completed by centrifuging at 7676 × g for 10 min at 20 ◦ C. Deproteinized samples
60
Experiment
Substrate
Experimental factors Citric acid, g/kg
1 2 3 4 5 a
SBM SBM SBM SBM SB
0 and 20 0 and 20 0 and 20 0 and 20 0 and 20
From crop through small intestine.
Digestion phase
In vitro analysis
Experimental design
Crop GITa Crop GIT Crop and gizzard
RS RS ␣-amino N ␣-amino N ␣-amino N and TI
2 × 6 factorial 2 × 7 factorial 2 × 5 factorial 2 × 5 factorial 2 × 4 factorial
Enzyme Type
Dose
␣-GAL ␣-GAL Protease Protease Protease
0, 1724, 3448, 5172, 6896 and 8620 units/kg 0, 5172, 6896, 8620, 10,344, 12,068 and 13,792 units/kg 0, 222, 444, 666 and 888 units/kg 0, 222, 333, 444 and 555 units/kg 0, 222, 333 and 444 units/kg
T. Ao et al. / Animal Feed Science and Technology 162 (2010) 58–65
Table 1 Experimental factors, in vitro analysis, phase of digestion and statistical design for Experiments 1–5.
T. Ao et al. / Animal Feed Science and Technology 162 (2010) 58–65
61
Substrate (2.5 ± 0.0001g)
Crop
+ 6 ml dislled water + 1 ml enzyme soluon 30 min, 40 ºC
Gizzard (pH = 3.0)
+ 7500 U of pepsin
+ ice-cold
+ 1 M HCl (1.5-2.0 ml)
water
45 min, 40 ºC Small intesne
+ 1 M NaHCO3 (3 ml)
( pH = 6.5) + 4.63 mg pancrean 60 min, 40 ºC
Centrifugaon and deproteinizaon
Reducing sugars
Whole phase
α-Amino nitrogen
Crop phase
Gizzard phase
Fig. 1. Flow chart of in vitro digestion procedure designed for the determination of reducing sugars and ␣-amino nitrogen for Experiments 1–4.
were used either to measure RS using method outlined by Miller (1959) or to measure ␣-amino N with the procedure described by Moore and Stein (1954).
2.5. In vitro digestion and measurement of activity of TI (Experiment 5) After grinding through a 1 mm screen triplicate samples of defatted soybeans weighing 1.0 ± 0.0001 g were placed in 50 ml centrifuge tubes. The samples were hydrated with 2 ml distilled water and 1 ml enzyme solution (10 mg/ml). The tubes were then vortexed, sealed and incubated in a water bath at 40 ◦ C for 30 min. The first step simulated the digestion in the crop. Then, 0.4–0.5 ml of 1.0 M HCl solution was used to adjust the pH to 3.0, sufficient water was added to keep the volumes equal in all tubes and 3000 units (0.5 ml) of pepsin solution were added to each tube. The tubes were vortexed, sealed and reincubated for 45 min at the same temperature. This step simulated digestion in the proventriculus and the gizzard of chickens. Once the digestion process was finished, the tubes were immediately put in an ice-cold water bath for 15 min to stop the further hydrolysis. Then, 10–15 ml Na2 HPO4 (0.5 M) was added to each tube to adjust the pH of the solution to 7.6. Distilled water (1.5–6.5 ml) was also added to each tube to adjust the total volume of the solution to 20.5 ml. The tubes were vortexed, sealed with parafilm and continuously shaken in a mechanical shaker for 1 h. Then, the tubes were centrifuged at 16,240 × g for 20 min. The supernatants were withdrawn and used to analyse the TI content according to a procedure proposed by Kunitz (1947), as modified by Kakade et al. (1969). One trypsin unit (TU) is arbitrarily defined as an increase of 0.01 absorbance units at 280 nm in 20 min per 10 ml of the reaction mixture under the assay conditions. Trypsin inhibitor activity is defined as the number of trypsin units inhibited.
62
T. Ao et al. / Animal Feed Science and Technology 162 (2010) 58–65
Table 2 Effects of ␣-GAL and CA on release of RS in SBM after in vitro digestion in crop phase (Experiment 1).a Factors ␣-GAL, units/kg
CA, g/kg
RS, g/kg
0 1724 3448 5172 6896 8620 0 1724 3448 5172 6896 8620 SEMb
0 0 0 0 0 0 20 20 20 20 20 20 0.65
0.0 6.3 9.4 13.9 17.1 19.1 0.0 7.0 13.4 17.0 23.1 26.9
Source of variation
P value
␣-GAL CA ␣-GAL × CA
<0.01 <0.01 <0.01
a b
Values are means of three samples, analysed in duplicate. Standard error of the mean.
2.6. Statistical analysis Data were collected from three replicates and statistically analysed using Statistix V.8 (2003) (Analytical Software, Tallahassee, FL). Following analysis of variance for a factorial arrangement of treatments, mean differences were determined using Fisher’s least significant difference test. Statistical significance was accepted at P<0.05. Polynomial contrasts were used to determine linear and quadratic effects of enzymes. 3. Results The effects of ␣-GAL and CA on RS release from SBM in the crop phase (Experiment 1) are presented in Table 2. The addition of ␣-GAL linearly (P<0.01, R2 = 0.99) increased the release of RS. The inclusion of CA further increased (P<0.01) the enzyme activity. A significant interaction between ␣-GAL and CA was found. At the same enzyme level, more RS were released for the treatments with CA than those without CA. The effects of ␣-GAL and CA on RS release from SBM after in vitro digestion in GIT phase (Experiment 2) are presented in Table 3. Similar results to those observed in the crop phase digestion were found in the whole phase. The effects of protease and CA on the release of ␣-amino N from SBM during simulated digestion through the crop and GIT phases are shown in Tables 4 and 5. In the crop phase (Experiment 3, Table 4), the enzyme effect was observed only when the CA was included in the substrate, resulting in a significant interactive effect between enzyme and CA. In the GIT phase digestion (Experiment 4, Table 5), a linear effect (P<0.01) of protease supplementation on ␣-amino N concentration was detected. Addition of CA further increased (P<0.05) the ␣-amino N concentration. No interactive effect between enzyme and acid was found. The effects of protease and CA supplementation on ␣-amino N release from raw whole soybeans and TI activity in the substrate after in vitro digestion (Experiment 5) are presented in Table 6. The addition of protease linearly increased (P<0.01) ␣-amino N release. The addition of CA further increased (P<0.01) ␣-amino N release during digestion. The TI activity in the digesta was similar for all treatments after incubation with or without either protease or CA. 4. Discussion The galacto-oligosaccharides found in SBM (approximately 60 g/kg) are not digested by nonruminant endogenous enzymes. However, they can be hydrolyzed either in vitro or in vivo when ␣-GAL is added to the diet (Gitzelmann and Auricchio, 1965; Trugo et al., 1995). Graham et al. (2002) reported that the ␣-GAL treatment of SBM hydrolyzed 0.69of the raffinose and 0.54 of the stachyose content. The results in the present study showed increased release of RS from SBM or whole soybeans from treatment with ␣-GAL indicating that galacto-oligosaccharides were hydrolyzed by the enzyme. More RS were released by the combination of both ␣-GAL and CA than by ␣-GAL alone indicating that the pH level in the incubation mixture is a limiting factor for maximizing the activity of ␣-GAL. Based on the analytical results, the pH levels for the SBM with or without 2 g/kg CA addition were 5.0 and 6.5, respectively. Ademark et al. (2001) reported that the optimal pH of
T. Ao et al. / Animal Feed Science and Technology 162 (2010) 58–65
63
Table 3 Effects of ␣-GAL and CA on release of RSs in SBM after in vitro digestion in GIT phase (Experiment 2).a Factors ␣-GAL, units/kg
CA, g/kg
RS, g/kg
0 5172 6896 8620 10,344 12,068 13,792 0 5172 6896 8620 10,344 12,068 13,792 SEMb
0 0 0 0 0 0 0 20 20 20 20 20 20 20 1.68
7.8 20.7 24.8 29.7 34.6 36.0 39.4 7.6 27.0 31.8 38.8 44.9 49.3 61.1
Source of variation
P value
␣-GAL CA ␣-GAL × CA
<0.01 <0.01 <0.01
a b
Values are means of three samples, analysed in duplicate. Standard error of the mean.
fungal ␣-GAL is about 4.5. Ao et al. (2008) indicated that the activity of fungal ␣-GAL was reduced by 50% when the pH was higher than 6.0. At the same enzyme level, more RS were released in the whole phase than in the crop phase. This means that the low pH (3.0) level found in gizzard and host proteolytic enzymes (pepsin and pancreatin) are not detrimental to fungal ␣-GAL. Chesson (1993) pointed out that fun polysaccharidases are not attacked in vitro by the major porcine digestive proteases used singly or in combination. Ao et al. (2008) showed that the activity of ␣-GAL at pH 3.0 was only 10% of the activity observed at the optimum pH of 5.0. Baas and Thacker (1996) studied the capability of different -glucanase products to recover activity after pre-incubation for 15, 30, 60 and 120 min at low pH levels. They found that pre-incubation at pH 3.5 caused loss of activity, with greater loss at pH 2.5. However, some enzyme activity was recovered upon return to pH 5.5. In a similar study, they found that pentosanases treated at pH 3.5 only partially recovered activity when returned to pH 5.5 and all enzymes exhibited a serious loss of activity when incubated at pH 2.5 and then returned to pH 5.5. After in vitro digestion in the present study, the release of ␣-amino N from the substrates was linearly increased by including graded levels of protease. This result is consistent with the data reported by Rook et al. (1998). They found that Table 4 Effects of protease and CA on release of ␣-amino N in SBM after in vitro digestion in crop phase (Experiment 3).a Factors Protease, units/kg
CA, g/kg
␣-Amino N, mg/kg
0 222 333 444 555 0 222 333 444 555 SEMb
0 0 0 0 0 20 20 20 20 20
218 247 222 275 227 381 407 379 405 510 12.1
Source of variation
P value
Protease CA Protease × CA
<0.01 <0.01 <0.01
a b
Values are means of three samples, analysed in duplicate. Standard error of the mean.
64
T. Ao et al. / Animal Feed Science and Technology 162 (2010) 58–65
Table 5 Effects of protease and CA on release of ␣-amino N in SBM after in vitro digestion in GIT phase (Experiment 4).a Factors Protease, units/kg
CA, g/kg
␣-Amino N, g/kg
0 222 444 666 888 0 222 444 666 888 SEMb
0 0 0 0 0 20 20 20 20 20 0.052
1.36 1.48 1.44 1.54 1.57 1.46 1.51 1.59 1.57 1.62
Source of variation
P value
Protease CA Protease x CA
<0.05 <0.05 >0.05
a b
Values are means of three samples, analysed in duplicate. Standard error of the mean.
Table 6 Effects of protease and CA on release of ␣-amino N and trypsin units inhibited (TUI/ml) in raw soybean after in vitro digestion (Experiment 5).a Factors Protease, units/kg
CA, g/kg
␣-Amino N, g/kg
TUI/ml
0 222 333 444 0 222 333 444 SEMb
0 0 0 0 20 20 20 20 0.048
1.36 1.37 1.40 1.51 1.62 1.82 1.86 1.85 0.51
19.7 20.5 20.6 21.1 20.4 19.7 18.2 21.1
Source of variation
P value
Protease CA Protease × CA
<0.01 <0.01 >0.05
a b
>0.05 >0.05 >0.05
Values are means of three samples, analysed in duplicate. Standard error of the mean.
the pretreatment of SBM with proteases changed the composition, and increased the soluble ␣-amino N concentration of the SBM. Although increasing levels of protease resulted in a linear increase in the release of ␣-amino N from raw soybean, it did not reduce TI activity. Huo et al. (1993) reported that bacterial proteases appeared to be more effective breaking down TI than fungal proteases. 5. Conclusions The enzymes ␣-GAL and protease may be useful in hydrolyzing the carbohydrates and proteins in SBM. The activity of these enzymes can be enhanced by acidification with CA. However, treating defatted whole soybeans with fungal protease does not appear to decrease TI activity. References Ademark, P., Larsson, M., Tjerneld, F., Stalbrand, H., 2001. Multiple ␣-GALs from Aspergillus niger: purification, characterization and substrate specifications. Enzyme Microb. Technol. 29, 441–448. Ao, T., Cantor, A.H., Pescatore, A.J., Pierce, J.L., 2008. In vitro evaluation of feed-grade enzyme activity at pH levels simulating various parts of the avian digestive tract. Anim. Feed Sci. Technol. 140 (3–4), 462–468. Baas, T.C., Thacker, P.A., 1996. Impact of gastric pH on dietary enzyme activity and survivability in swine fed -glucanase supplemented diets. Can. J. Anim. Sci. 76, 245–252. Chesson, A., 1993. Feed enzymes. Anim. Feed Sci. Technol. 45, 65–79.
T. Ao et al. / Animal Feed Science and Technology 162 (2010) 58–65
65
Classen, H.L., Balnave, D., Bedford, M.R., 1993. Reduction of legume antinutritional factors using biotechnological techniques. In: Van Der Poel, A.F.B., Huisman, J., Saini, H.S. (Eds.), Recent Advances of Research in Antinutritional Factors in Legume Seeds. Proceedings of the Second International Workshop on ‘Antinutritional factors (ANFs) in Legume seeds’. Wageningen Pers, Wageningen, The Netherlands, pp. 501–516. Committee on Food Chemicals Codex, 1996. Food Chemicals Codex, Fourth ed. National Academy Press, Washington, DC. Giesting, D.W., Easter, R.A., 1985. Response of starter pigs to supplementation of corn-soybean meal diets with organic acids. J. Anim. Sci. 60, 1288–1294. Gitzelmann, R., Auricchio, S., 1965. The handling of soy ␣-galactosidase by a normal and galactosemic child. Pediatrics 36, 231–232. Graham, K.K., Kerley, M.S., Firman, J.D., Allee, G.L., 2002. The effect of enzyme treatment of soybean meal on oligosaccharide disappearance and chick growth performance. Poult. Sci. 81, 1014–1019. Herpol, C., Van Grembergen, G., 1967. La signification du pH dans le tube digestif de gallus domesticus. Ann. Biol. Anim. Biochim. Biophys. 7, 33–38. Huisman, J., Tolman, G.H., 1992. Antinutritional factors in the plant protein of diets for non-ruminants. In: Garnsworthy, P.C., Haresign, W., Cole, D.J.A. (Eds.), Recent Advances in Animal Nutrition. Butterworth-Heinemann, Oxford, UK, pp. 3–32. Huo, G.C., Fowler, V.R., Bedford, M., 1993. The use of enzymes to denature antinutritive factors in soybean. In: Van Der Poel, A.F.B., Huisman, J., Saini, H.S. (Eds.), Recent Advances of Research in Antinutritional Factors in Legume Seeds. Proceedings of the Second International Workshop on ‘Antinutritional Factors (ANFs) in Legume Seeds’. Wageningen Pers, Wageningen, The Netherlands, pp. 517–521. Kakade, M.L., Simons, N., Liener, I.E., 1969. An evaluation of natural vs. synthetic substrates for measuring the antitryptic activity of soybean samples. Cereal Chem. 46, 518–526. Kunitz, M., 1947. Crystalline soybean trypsin inhibitor. II. General properties. J. Gen. Physiol. 30, 291–310. Marsman, G.J.P., Gruppen, H., Van Der Poel, A.F.B., Kwakkel, R.P., Verstegen, M.W.A., Voragen, A.G.J., 1997. The effect of thermal processing and enzyme treatments of soybean meal on growth performance, ileal nutrient digestibility, and chime characteristics in broiler chicks. Poult. Sci. 76, 864–872. Miller, G.L., 1959. Use of dinitrosalisalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31, 426–428. Moran Jr., E.T., 1982. Comparative Nutrition of Fowl and Swine: The Gastrointestinal Systems. University of Guelph, Guelph, Ontario, Canada. Moore, S., Stein, W., 1954. A modified ninhydrin reagent for the photometric determination of amino acids and related compounds. J. Biol. Chem. 211, 907–913. Ratto, M., Poutanen, K., 1988. Production of mannan-degrading enzymes. Biotechnol. Lett. 10, 661–664. Rook, J.A., Slessor, M., Fraser, H., Thomson, J.R., 1998. Growth performance and gut function of piglets weaned at four weeks of age and fed protease-treated soybean meal. Anim. Feed Sci. Technol. 70, 175–190. Sheppy, C., 2001. The current feed enzyme market and likely trends. In: Bedford, M.R., Partridge, G.G. (Eds.), Enzymes in Farm Animal Nutrition. CABI Publishing, Wallingford, UK, pp. 1–10. Sonnenwirth, A., Jarett, L., 1980. Clinical Laboratory Methods and Diagnosis. C.V. Mosby Co., St. Louis, MO, USA, pp. 149–192. Tervila-Wilo, A., Parkkonen, T., Morgan, A., Hopeakoski-Nurminen, M., Poutanen, K., Heikkinen, P., Autio, K., 1996. In vitro digestion of wheat microstructure with xylanase and cellulase from trichoderma reesei. J. Cereal Sci. 24, 215–225. Trugo, L.C., Farah, A., Cabral, L., 1995. Oligosaccharide distribution in Brazilian soya been cultivars. Food Chem. 52, 385–387. Verstegen, M.W.A., Voragen, A.G.J., 1997. The effect of thermal processing and enzyme treatments of soybean meal on growth performance, ileal nutrient digestibility, and chime characteristics in broiler chicks. Poult. Sci. 76, 864–872. Zyla, K., Gogol, D., Koreleski, J., Swiatkiewicz, S., Ledoux, D.R., 1999. Simultaneous application of phytase and xylanase to broiler feeds based on wheat: in vitro measurements of phosphorus and pentose release from wheat and wheat-based feeds. J. Sci. Food Agr. 79, 1832–1840.