Effect of Pelleting Temperature on the Recovery and Efficacy of a Xylanase Enzyme in Wheat-Based Diets

Transcription

1 Effect of Pelleting Temperature on the Recovery and Efficacy of a Xylanase Enzyme in Wheat-Based Diets F. G. SILVERSIDES* and M. R. BEDFORD,1 *2705 Piercy Road, Denman Island, British Columbia, Canada V0R 1T0, and Finnfeeds International, Box 777, Marlborough, Wiltshire, United Kingdom SN8 1XN ABSTRACT Two trials were performed to test the stability and efficacy of a commercial enzyme preparation containing xylanase and protease activities to pelleting. In Trial 1, 576 male Cobb chicks were fed wheat-based diets to 21 d with or without enzyme and pelleted after conditioning for 55 or 140 s at 70, 75, 80, 85, 90, or 95 C. In Trial 2, 2,880 male Cobb chicks were fed wheat-based diets to 42 d with no enzyme, or with enzyme addition before or after pelleting with conditioning for 30 s at 70, 80, 90, or 95 C. Enzyme addition had a positive effect on BW gain and the feed to gain ratio (FG) in Trial 1 and on FG to 42 d in Trial 2. All treatment effects were significant for intestinal viscosity in both trials. Two- (BW) and three-way (BW, FG) interactions between enzyme, temperature, and time were seen in Trial 1. With enzyme supplementation, second degree polynomial equations of performance on conditioning temperature had high R2 values for BW and FG, with temperatures between 80 and 85 C being most favorable. Enzyme activity in unsupplemented diets measured in vitro was largely eliminated at temperatures above 80 C, whereas the decline in activity in supplemented diets was linear. In spite of the decline in enzyme activity measured in vitro, intestinal viscosity of chicks fed enzyme-supplemented diets remained low with high processing temperatures, whereas that of chicks fed unsupplemented diets increased dramatically. (Key words: xylanase enzyme, broiler, wheat, heat stability, intestinal viscosity) 1999 Poultry Science 78: INTRODUCTION The use of enzymes in poultry feed has increased dramatically in the last decade. Enzymes are used primarily with wheat and barley in mixed feeds, but may also be useful in diets based on corn and soybeans (Zanella et al., 1999). Wheat and barley contain variable levels of soluble nonstarch polysaccharides (NSP), the level depending on the cereal variety and growing conditions (Scott et al., 1998). It is known that a proportion of NSP is of high molecular weight, which dissolves in the intestine, leading to an increase in the viscosity of the gut contents. High viscosity leads to reduced growth rate, increased feed to gain ratio (FG), and sticky droppings (Choct and Annison, 1990; Annison and Choct, 1991). Enzymatic digestion of NSP improves performance and allows more efficient use of nutrients (Bedford, 1996; Bedford and Morgan, 1996). Most feed used for commercial industrial broilers is heated and compressed to form pellets. Pelleting is associated with positive effects for both feed handling and bird performance, including increased feed utilization and better growth rate (Leeson and Summers, 1991; Gibson, 1995). Pelleting can also be associated with negative effects, especially when the mixture is overheated. Overheating can result in resistance of protein (Araba and Dale 1990) and starch (Brown, 1996) to digestion, inactivation of vitamins (Pickford, 1992), increased intestinal viscosity (Nissinen, 1994), and inactivation of endogenous enzymes (Inborr and Bedford, 1994). Intestinal viscosity is determined by many factors that can vary significantly between batches of feed. As a result, enzyme activity in the feed must be maintained at a level that will provide optimum performance under the worst likely conditions of viscosity. The stability of added feed enzymes to the pelleting process is a major concern of feed manufacturers, because pelleting can significantly reduce the safety margins incorporated into the feed formulation. Most of the inactivation takes place during conditioning, when the feed is heated with steam, rather than during extrusion of the pellets (Eeckhout et al., 1995). Adding Received for publication November 23, Accepted for publication March 31, To whom correspondence should be addressed: Mike.Bedford@cultor.com Abbreviation Key: FG = feed to gain ratio; FI = feed intake; MSD = minimum significant difference; NSP = nonstarch polysaccharides. 1184

2 PELLETING TEMPERATURE AND ENZYME SUPPLEMENTATION 1185 the enzyme to the finished pellets as a liquid overcomes the problem of inactivation but can result in a lack of homogeneity and requires special equipment. The enzyme can be protected in substrate-bound and granulated enzyme preparations coated with hydrophobic compounds. Despite these possibilities, many enzymes are mixed in the feed as a dry powder prior to pelleting because of simplicity. Some compensation can be made for enzyme inactivation by using higher quantities, but this is not an option at higher processing temperatures. It has become clear that commercial enzyme preparations vary in their ability to withstand pelleting (Inborr and Bedford, 1994; Gibson, 1995; Petterson and Rasmussen, 1997), although the measurement of efficacy after processing is not straightforward. Direct measurement of enzyme activity in vitro has been used, but a standard test is not available (McCleary, 1995; Bedford et al., 1997). In addition, measuring the recovery of enzyme activity from feed is complex, and the relationship between recovered activity and performance has not been established. Furthermore, pelleting affects all nutrients, not just the enzyme. Measuring performance or in vivo intestinal viscosity may be more useful than direct tests of enzymatic recovery or activity. This paper reports on two studies that were performed to test the effect of pelleting temperature on the recovery and the in vivo efficacy of a commercial enzyme preparation containing xylanase and protease activity. MATERIALS AND METHODS The feed for Trial 1 was mixed at the Scottish Agricultural College2 then sent to The Flour Milling and Baking Research Association3 for pelleting before being returned to the Scottish Agricultural College for the chick trials. Diets were conditioned in a twin extruder barrel set to specific temperatures ± 1 C by use of an insulating water jacket. Feed rate through the barrel was between 300 to 500 g/min and screw speed varied between 71 and 128 rpm depending upon the residence time and the temperature of feed manufacture. For Trial 2, the feed was mixed in a Forberg4 400L batch mixer and pelleted, and the growth trial was performed at Roslin Institute.5 The feed production process at Roslin mimics a typical short-term, horizontal barrel conditioning and pelleting process in that temperatures are generated and held using live steam injected into the conditioner at between 1.5 and 2 bars of pressure. In 2Craibstone Estate, Aberdeen, Scotland U.K., AB2 9SX. 3Chorleywood, Herefordshire, U.K., WD3 SSH, now renamed as Camden and Chorleywood Food Research Association. 4Forberg AS, Hegdal, N-321 Larvik, Norway. 5Roslin, Midlothian, Scotland, U.K., EH25 9PS. 6Brookfield Engineering Laboratories, Stoughton, MA Finnfeeds International, Marlborough, Wiltshire, U.K., SN8 1XN. each case, the animal care protocols were accepted followed standards set in the United Kingdom. In each trial, BW gain to 21 d was measured, as was the feed intake (FI), FG, and mortality. Intestinal viscosity was measured using a Brookfield6 model DVII viscometer as described by Inborr and Bedford (1994). Enzyme additions were to the mixed diets so actual levels of ingredients should be factored accordingly. Enzyme activity in the enzyme preparations and in the feed after pelleting was measured by Finnfeeds International.7 Trial 1 For Trial 1, 576 male Cobb chicks were housed in 96 groups of six birds to give 24 treatments with four repetitions per treatment. The diet was based on soft wheat and soybean meal (Table 1) with or without Avizyme (analyzed at 2,860 U/g xylanase from Trichoderma and 940 U/g protease derived from Bacillus) at a rate of 1.0 g/kg. The diets were conditioned for 55 or 140 s at 70, 75, 80, 85, 90, or 95 C, at a moisture content of 18.5%. The viscosity of the contents of the proximal small intestine from the duodenal loop to Meckel s diverticulum was measured on eight birds per diet at the age of 21 d. Trial 2 Trial 2 used 2,880 male Cobb chicks distributed into 72 groups of 40 chicks each giving 6 repetitions of 12 treatments. Starter and finisher diets (Table 1) were based on wheat (variety Brigadier) and soybean meal. Three feed treatments were used: no enzyme, dry enzyme (Avizyme 1300, 1.0 g/kg analyzed at 2,626 U xylanase and 643 U protease/g) added before pelleting, and liquid enzyme (Avizyme 1310, g/kg analyzed at 4,428 U xylanase/g) added after pelleting. The pelleting process used 30 s conditioning at 70, 80, 90, or 95 C. Performance data for Trial 2 were collected until the birds were 42 d of age. Viscosity of the intestinal contents was measured at 21 d on one or two birds per pen. Enzyme Activity Determination Xylanase activity in feed samples was determined according to the method described by McCleary (1995) with ph adjusted to 4.3. Statistical Analysis Data were analyzed for each trial using the General Lineal Models procedure of SAS (Littell et al., 1991). The ANOVA for Trial 1 included the effects of enzyme addition, treatment temperature, treatment time, and the two- and three-way interactions among these effects. The ANOVA for Trial 2 included the effects of enzyme addition, treatment time, and the interaction between them. Minimum significant differences (MSD) in both

3 1186 SILVERSIDES AND BEDFORD TABLE 1. Composition of the diets in two trials 1 Trial 2 Ingredients and analysis Trial 1 0 to 21 d 21 to 42 d (%) Ingredients Wheat (soft) SBM 45% CP Soya oil Tallow NaCl DL-Methionine Lysine Threonine Limestone Dicalcium phosphate Vitamin and mineral premix Calculated analysis AME, kcal/kg 3,050 3,031 3,155 Crude protein Calcium Available phosphorus Fat Fiber Methionine Cystine Methionine + cystine Lysine Enzyme addition and activity is shown in the text. 2The premix contained per kilogram of diet: vitamin A, 10,500 IU; cholecalciferol, 2,100 IU; vitamin E, 22 mg; riboflavin, 4.4 mg; niacin amide, 40 mg; pantothenic acid, 14 mg; choline chloride, 560 mg; vitamin B 12,16mg; vitamin K 1, 5 mg; folic acid, 0.80 mg; biotin, 0.1 mg; selenium, 0.1 mg; iodine, 2 mg; cobalt, 0.2 m; copper, 4 mg; iron, 30 mg; zinc, 40 mg; manganese, 68 mg; sulfur, 125 mg. trials were determined for the interaction means using the Tukey test with a = Linear and second degree polynomial regression equations were calculated for the effect of processing time on performance, intestinal viscosity, and feed enzyme activity in Trial 1. RESULTS In Trial 1, enzyme addition, processing temperature, and processing time all had significant effects on intestinal viscosity, as did the interaction between enzyme addition and processing temperature (Table 2). Processing for a longer period produced a slightly lower intestinal viscosity. In vitro enzyme activity was detected in all samples whether or not enzyme was added, indicating the presence of endogenous enzyme activity. This activity was largely destroyed by heating above 80 C for 55 s and above 75 C for 140 s. Broiler performance in Trial 1 is also shown in Table 2. Mortality was significantly higher for chickens fed diets with added enzyme, but this percentage was based on a very small number of birds that died. None of the treatments affected FI. The addition of enzyme to the diet had a positive effect on BW gain and FG. The main effects of temperature and time of conditioning were not significant for BW gain or FG. Significant two- (FG) and three-way interactions (BW and FG) show that the positive effect of enzyme addition was not consistent over treatment time or temperature. The interaction of enzyme addition with the period and temperature of treatment was investigated further by regression analysis (Table 3). Broiler performance in this trial suggested that the processing period was less important than the temperature. Hence, to simplify the presentation, regression equations are shown only for feed that was processed for 55 s. Linear and seconddegree polynomial regressions were calculated of the response variables on the processing temperature. Without exogenous enzyme, the processing temperature had essentially no effect on BW gain or FG (R2 < 0.05). However, when enzyme was added before pelleting, the polynomial regression lines of BW gain and FG on temperature produced R2 values (representing the percentage of variation in BW and FG explained by the temperature) of 0.84 and 0.98, respectively. Fitted regression lines (not shown) reached a maximum BW gain and minimum FG at a processing temperature between 80 and 85 C. Endogenous enzyme activity measured in vitro was poorly related to the processing temperature, likely because of the low levels present. In feeds containing exogenous enzyme, the R2 value for the linear regression was not improved by calculating a polynomial regression, showing that the decrease in activity was linear. The increase in intestinal viscosity in both enzyme-supplemented and unsupplemented feeds was also linear, and was four times greater for unsup-

4 PELLETING TEMPERATURE AND ENZYME SUPPLEMENTATION 1187 plemented feeds than it was for feeds supplemented with enzyme. Trial 2 compared diets without enzyme with those containing an enzyme preparation added before pelleting and an enzyme preparation added after pelleting (Table 4). Differences in BW and FI due to the enzyme addition and processing temperature were not significant. Enzyme addition before and after pelleting both improved the FG to 42 d over that of the control diet by an approximately equal amount (by and 0.067, respectively). Although intestinal viscosity was relatively low even in the control diets, the effects of the enzyme addition, the temperature, and the interaction between the two were all significant. As was seen for Trial 1, analysis of enzyme activity in the feed showed that the diets without exogenous enzyme had low levels of activity. Enzyme activity in diets supplemented with enzyme before pelleting declined in a linear manner with increasing processing temperature (R2 = 0.91). As expected, the activity in diets supplemented after pelleting was not dramatically altered with increasing processing temperature. DISCUSSION Trial 1 demonstrated that addition of a commercial enzyme with xylanase and protease activities to wheatbased diets was effective in reducing intestinal viscosity and improving broiler performance. These effects of enzyme addition have been shown in a number of experiments (Pettersson and Åman, 1988; Bedford and Morgan, 1996) and are well known commercially, as over 80% of wheat-based broiler feeds now contain enzyme. However, the temperature used for condition- TABLE 2. Chick performance, digesta viscosity, and in vitro enzyme activity in Trial 1 1 BW gain Feed:gain Feed intake Intestinal In feed Treatment (0 to 21 d) (0 to 21 d) (0 to 21 d) Mortality viscosity xylanase 2 (kg) (g:g) (g) (%) (cps) (U/g) No enzyme, 55 s , Enzyme, 55 s No enzyme, 140 s Enzyme, 140 s MSD Enzyme ** ** NS * ** Temperature NS (0.09) NS (0.07) NS NS (0.06) ** Time NS NS NS NS * Enzyme temperature NS * NS (0.06) NS ** Enzyme time NS NS NS NS NS Temperature time NS NS NS NS (0.06) NS Enzyme temp time * ** NS * NS 1Each mean represents four pen measurements. Only one measure of enzyme activity was made per treatment. 2Activity of the enzyme added was analyzed at 2,860 U/g, providing 2.86 U/g feed. 3Minimum significant difference. *P < **P < 0.01.

5 1188 SILVERSIDES AND BEDFORD TABLE 3. The effect of processing temperature (for 55 s) on BW gain, feed:gain ratio, in vitro enzyme activity, and in vivo intestinal viscosity in Trial 1 Response variable Enzyme 1 Regression equation R 2 Body weight gain gain = temperature 0.03 gain = temperature 0.02 temperature gain = temperature 0.17 gain = temperature 0.19 temperature Feed:gain ratio Feed:gain = temperature Feed:gain = temperature temperature Feed:gain = temperature 0.01 Feed:gain = temperature temperature In vitro enzyme activity Activity = temperature 0.21 Activity = temperature temperature Activity = temperature 0.97 Activity = temperature temperature In vivo viscosity Viscosity = temperature 0.97 Viscosity = temperature temperature Viscosity = temperature 0.58 Activity = temperature temperature = without enzyme; + = with enzyme. ing before pelleting was important for broiler performance. Moderate heating results in gelatinization of starch and cell wall destruction, both of which improve the availability of nutrients (Pickford, 1992). More extreme heating inactivates vitamins and enzymes (Pickford, 1992) and decreases starch (Brown, 1996) and protein utilization (Batterham et al., 1993; 1994). Heating of wheat-based rations has been shown to increase intestinal viscosity (Nissinen, 1994), which actually increases the likely response from enzyme addition. Because the effects of increasing processing temperature change from positive to negative, and because enzymes alter the intestinal viscosity, it is not surprising that the effect of enzyme addition on performance changed as TABLE 4. Chick performance, digesta viscosity, and in vitro enzyme activity in Trial 2 1 In feed BW gain Feed:gain Intestinal xylanase 3 Treatment 0 to 21 d 0 to 42 d 0 to 21 d 0 to 42 d Mortality 2 viscosity Starter Grower (g) (g:g) (%) (cps) (U/g) No enzyme , < , <0.10 < , <0.10 < , <0.10 <0.10 Avizyme , , , , Avizyme , , , , MSD Enzyme NS NS NS ** NS ** Temperature NS NS NS NS NS * Enzyme temperature NS NS NS NS NS * 1Each mean represents six pen measurements. 2Includes one or two birds per pen selected for in vivo viscosity. 3Activity of Avizyme 1300 was analyzed at 2,626 U/g and that of Avizyme 1310 was analyzed at 4,428, providing and U/g feed respectively. 4Minimum significant difference.

6 PELLETING TEMPERATURE AND ENZYME SUPPLEMENTATION 1189 well. Broiler performance with enzyme supplementation was best between 80 and 85 C in both trials, confirming the results of an earlier trial (Bedford, unpublished data). Above 85 C, the negative effects of high temperature, not just on enzyme but on other feed ingredients as well, outweigh the positive ones. The in vitro activity of endogenous enzyme (activity measured in feeds that were not supplemented with enzyme) was dramatically reduced with heating above 80 C, as has been found by Carre et al. (1994). The activity of exogenous enzyme (activity measured in feeds that were supplemented with enzyme) decreased in a linear manner with increasing processing temperature, confirming the results of Eeckhout et al. (1995), Bedford et al. (1997), and those of an earlier trial (Bedford, unpublished data). In supplemented feeds, some enzyme activity was present even at the highest processing temperature. It is unclear why endogenous activity is lost within a narrow temperature range whereas the decrease in exogenous activity is linear. Trial 1 showed that heating of diets that were not supplemented with enzyme resulted in greatly increased intestinal viscosity, although this increase was not observed in Trial 2, possibly because of a lower level of NSP in the wheat used. Increased intestinal viscosity with higher processing temperatures was also seen by Nissinen (1994) and is likely due to the increased solubility of NSP as well as the destruction of endogenous enzyme. In feeds supplemented with enzyme, intestinal viscosity was relatively low even with higher processing temperatures, meaning that the enzyme had a greater positive effect with higher temperatures. Control of intestinal viscosity is the best method for determining whether or not the enzyme has been destroyed. Arabinoxylan in wheat is a substrate for xylanase, hence inactivation of the enzyme allows increased arabinoxylan concentration, leading to higher viscosity. However, an increase in viscosity was not evident even at the highest temperature encountered, suggesting that xylanase activity was not limiting in any of the supplemented diets. It seems at first contradictory that higher processing temperatures decrease in vitro enzyme activity and increase the enzyme response measured as the difference in intestinal viscosity between control and supplemented diets. Trial 2 showed that enzyme addition before pelleting, which destroyed the majority of the enzyme activity measured in vitro, produced only marginally higher intestinal viscosity than addition after pelleting, when processing temperature should be irrelevant. Increased enzyme activity at high processing temperatures could be due to increased substrate being made available for enzyme degradation or to enzymatic degradation of NSP taking place during the conditioning process (Bedford et al., 1997). In the first trial, the specific mechanical energy, which is the energy required for pelleting, was determined (results not shown) and was seen to decrease by an average of 20% with enzyme addition, suggesting that the enzyme is active in the conditioner phase of manufacture. In vitro enzyme assay determination showed destruction of enzyme at high processing temperatures, whereas enzyme addition produced the best performance response and the best viscosity reduction at the highest temperatures studied. This work therefore shows that measuring in vitro enzyme activity after pelleting may not be as good an indicator of enzyme activity or utility as performance or in vivo intestinal viscosity. ACKNOWLEDGMENTS The authors would like to thank Robin Guy of the Camden and Chorleywood Food Research Association, Ian Wallis of the Scottish Agricultural College (Craibstone Estate) and Anne Knox of The Roslin Institute for carrying out the experimental part of this research. This work was supported financially by Finnfeeds International. REFERENCES Annison, G., and M. Choct, Anti-nutritive activities of cereal non-starch polysaccharides in broiler diets and strategies minimizing their effects. World s Poult. Sci. J. 47: Araba, M., and N. M. Dale, Evaluation of protein solubility as an indicator of overprocessing soybean meal. Poultry Sci. 69: Batterham, E. S., L. M. Anderson, and D. R. Baigent, Utilization of ileal digestible amino acids by growing pigs: Methionine. Br. J. Nutr. 70: Batterham, E. S., L. M. Anderson, and D. R. Baigent, Utilization of ileal digestible amino acids by growing pigs: Tryptophan. Br. J. Nutr. 71: Bedford, M. R., Interaction between ingested feed and the digestive system in poultry. J. Appl. Poult. Res. 5: Bedford, M. R., and A. J. Morgan, The use of enzymes in poultry diets. World s Poult. Sci. J. 52: Bedford M. R., M. Pack, and C. L. Wyatt, Relevance of in feed analysis of enzyme activity for prediction of bird performance in wheat based diets. Poultry Sci. 76(Suppl. 1):39. (Abstr.) Brown, I., Complex carbohydrates and resistant starch. Nut. Rev. 54:S115 S119. Carre, B., J. Gomez, J. P. Melcion, and B. G. Giboulot, La viscosité des aliments destinés à l aviculture. Utilisation pour predire la consommation et l excretion d eau. INRA Prod. Anim. 7: Choct, M., and G. Annison, Anti-nutritive activity of wheat pentosans in poultry diets. Br. Poult. Sci. 31: Eeckhout, M., M. DeSchrijver, and E. Vanderbeke, The influence of process parameters on the stability of feed enzymes during steam pelleting. Pages in: Proceedings of the 2nd European Symposium on Feed Enzymes. Noordwijkerhout, The Netherlands.

7 1190 SILVERSIDES AND BEDFORD Gibson, K., The pelleting stability of animal feed enzymes. Pages in: Proceedings of the 2nd European Symposium on Feed Enzymes. Noordwijkerhout, NL. Inborr, J., and M. R. Bedford, Stability of feed enzymes to steam pelleting during feed processing. Anim. Feed Sci. Technol. 46: Leeson, S., and J. D. Summers, Commercial Poultry Nutrition. University Books, Guelph, ON, Canada. Littell, R. C., R. J. Freund, and P. C. Spector, SAS System for Linear Models. 3rd ed. SAS Series in Statistical Applications. SAS Institute Inc., Cary, NC. McCleary, B. V., Problems in the measurement of betaxylanase, beta-glucanase and alpha-amylase in feed enzymes and animal feeds. Pages in: Proceedings of the 2nd European Symposium on Feed Enzymes, Noordwijkerhout, The Netherlands. Nissinen, V., Enzymes and processing: The effects and interactions of enzymes and hydrothermal pre-treatments and their contribution to feeding value. Int. Milling Flour and Feed, May Pettersson, D., and P. Åman, Effects of enzyme supplementation of diets based on wheat, rye or triticale on their productive value for broiler diets. Anim. Feed Sci. Technol. 20: Pettersson, D., and P. B. Rasmussen, Improved heat stability of xylanases. Pages in: Proceedings of the Australian Poultry Science Symposium, Sydney, Australia. Pickford, J. R., Effects of processing on the stability of heat labile nutrients in animal feeds. Pages in: Recent Advances in Animal Nutrition. P. C. Garnsworthy, W. Haresign, D.J.A. Cole, ed. Butterworth-Heinemann, Oxford, U.K. Scott, T. A., F. G. Silversides, H. L. Classen, M. L. Swift, and M. R. Bedford, Effect of cultivar and environment on the feeding value of Western Canadian wheat and barley samples with and without enzyme supplementation. Can. J. Anim. Sci. 78: Zanella, I., N. K. Sakomura, F. G. Silversides, A. Fiqueiredo, and M. Pack, Effect of enzyme supplementation of broiler diets based on corn and soybeans. Poultry Sci. 78: