Protein Dispersibility Index as an Indicator of Adequately Processed Soybean Meal

Dispersibility Index as an Indicator of Adequately Processed Soybean Meal A. B. Batal, M. W. Douglas, A. E. Engram, and C. M. Parsons 1 Department of Animal Sciences, University of Illinois, Urbana, Illinois 61801 ABSTRACT Three chick assays (8 to 17 or 21 d) were times (1.65 to 2.4), and then decreased suddenly to 0.3 or conducted to evaluate protein dispersibility index (PDI) as an indicator of minimum adequate heat processing of soybean meal compared with the urease index and protein solubility in 0.2% KOH. Solvent-extracted soyflakes (SF) were subjected to various autoclaving times at 121 C and 105 kpa and were included in 23% CP dextrose- SF diets or 20% CP corn-sf-corn gluten meal diets. Autoclaving times in Chick Assays 1, 2, and 3 were 0 to 36 min, 0 to 30 min, and 0 to 12 min, respectively. Body weight gains and gain-to-feed ratios increased (P < 0.05) with increased SF heating time (0 to 18 min in Chick Assay 1, 0 to 10 min in Chick Assay 2, and 0 to 9 min in Chick Assay 3), with no additional improvement for longer autoclaving times. Urease index values (ph increase) were high initially and at the shorter autoclaving below as autoclaving time increased in two of the three chick assays. The KOH protein solubility values generally decreased as autoclaving time increased, but the responses were often inconsistent. dispersibility index displayed the most consistent responses to heating time: it decreased from above 70% to generally below 30% as autoclaving time increased from 0 to 30 or 36 min (mean r 2 from linear regression of PDI on autoclaving time was 0.92 for the three chick assays). The latter responses were particularly evident for the heating times, which yielded the greatest changes in chick growth performance. These results suggest that PDI is a more consistent and sensitive indicator of minimum adequate heat processing of soybean meal than urease index or protein solubility in KOH. (Key words: soybean meal, protein dispersibility, urease index, KOH solubility) 2000 Poultry Science 79:1592 1596 INTRODUCTION Soybean meal (SBM) is the most important source of dietary protein for poultry in the U.S. and much of the world. Variation in protein quality among samples of SBM can occur due to either insufficient heating (underprocessing) or excessive heating (overprocessing). The most common assay used to evaluate SBM processing is urease index (American Oil Chemists Society, 1980a). Urease is used as an indirect indicator of the presence of antinutritional factors, such as trypsin inhibitors, that indicate underprocessing of SBM. This assay is useful for detecting undercooking of SBM, but is of limited use for detecting overcooking (Araba and Dale, 1990a; Parsons et al., 1991). Moreover, the recommended maximum level of urease is controversial, with acceptable values varying from 0.2 or less (McNaughton and Reece, 1980) to 0.5 units of ph change (Waldroup et al., 1985). The fact that the urease index is not linear and that it rapidly falls from approximately 2.0 units of ph change to near zero as SBM is heated contributes to the difficulty in determining a precise maximum acceptable level of urease. solubility in 0.2% KOH has been shown to be a good indicator of in vivo protein quality for overprocessed SBM (Araba and Dale, 1990a; Parsons et al., 1991). The KOH solubility assay was also reported to be useful for detecting underprocessing of SBM (Araba and Dale, 1990b); however, Anderson-Haferman et al. (1992) later concluded that this assay was not very accurate for assessing underprocessing of SBM. Another method that is often used in ruminant and human nutrition to monitor optimum heat processing of soy products is protein dispersibility index (PDI; American Oil Chemists Society, 1980b). The PDI measures protein solubility in water with high-speed mixing. To our knowledge, PDI has not been evaluated as an indicator of SBM quality for poultry. Therefore, the present study was conducted to evaluate PDI as an indicator of minimum adequate heat processing of SBM for poultry and to determine whether PDI is more sensitive than the urease and KOH protein solubility assays. Received for publication January 26, 2000. Accepted for publication June 30, 2000. 1 To whom correspondence should be addressed: poultry@uiuc.edu. Abbreviation Key: PDI = protein dispersibility index; SBM = soybean meal; SF = soyflakes. 1592

PROTEIN DISPERSIBILITY INDEX 1593 MATERIALS AND METHODS Sample Preparation and Analyses Different batches of dehulled, solvent-extracted soyflakes (SF) (postexpander but prior to desolventizertoaster) were obtained from a commercial soybean plant. 2 The SF were ground to a consistent particle size, hydrated with 12% water, spread (2.5 cm thick) in aluminum pans, covered with aluminum foil, and autoclaved for various times at 121 C at 105 kpa (1 kpa = 0.145 pounds per square inch). For the hydration, 12% water was added to the SF on a total-weight basis and mixed in a bowl-type mixer. The hydrated SF were then screened to remove clumps and remixed. The SF samples (which were autoclaved at various times) were then allowed to air-dry at room temperature and were incorporated into experimental diets as described later. The original and autoclaved SF were analyzed in triplicate for urease index (ph increase) and protein solubility in 0.2% KOH and PDI. The KOH solubility was determined by the method specified by Araba and Dale (1990a) and Parsons et al. (1991). Urease index and PDI were determined by the methods described by the American Oil Chemists Society (1980a,b). Trypsin-inhibitor analyses were determined on triplicate SF samples in Chick Assay 1 by Ralston Analytical Laboratories (Ralston Purina Co., St. Louis, MO 63164) using the method of the American Association of Cereal Chemists (1995). Chick Assays All chick assays used 8-d-old New Hampshire Columbian male chicks. The chicks had been fed a nutritionally complete 23% CP corn-sbm diet from 0 to 7 d posthatching. After an overnight fast, the chicks were allotted to pens of five chicks as described by Sasse and Baker (1973) so that the average body weight per pen was approximately equal. Four pens of five chicks each (Experiment 1) or four pens of four chicks each (Experiments 2 and 3) were assigned to each dietary treatment. Chicks were housed in thermostatically controlled battery brooders with raised wire floors. Feed and water were provided ad libitum for 8 to 17 d in Assay 1, and 8 to 21 d in Assay 2 and 3. Body weight gains and feed efficiencies were then determined. In Chick Assay 1, a dextrose-sf diet (Table 1) containing SF as the sole source of protein was used. This diet was calculated to be adequate in all amino acids (National Research Council, 1994) except that it was slightly deficient in Met + Cys. Soyflake samples that had been autoclaved for 0, 6, 12, 18, 24, 30, or 36 min and commercial, dehulled SBM were evaluated. In Chick Assays 2 and 3, a corn-sf-corn gluten meal diet (Table 1) was formulated to supply 20% CP and 3,200 2 Soyflakes were obtained from Owensboro Grain Company, Owensboro, KY 42301. kcal Me n /kg diet. The diet was designed to be deficient in lysine (0.75%) but adequate in all other amino acids (National Research Council, 1994). A preliminary trial was conducted to determine whether the diet was deficient in lysine and would respond to lysine supplementation. This diet with no supplemental lysine (0.75% lysine) resulted in a weight gain of 253 g per chick, whereas the diet supplemented with 0.45% lysine resulted in a weight gain of 273 g per chick (P < 0.05) from 8 to 21 d of age (four pens of five chicks per diet). The lysine-deficient diet was used to increase the effects of autoclaving SF on chick performance. Soyflakes for Chick Assay 2 were autoclaved for 0, 10, 20, and 30 min. Because no growth differences were observed among autoclaving times greater than 10 min in Chick Assay 2, SF were autoclaved for 0, 3, 6, 9, and 12 min in Chick Assay 3 to produce more differences among treatments. TABLE 1. Composition and calculated composition of the diets Chick assay Item 1 2 and 3 (%) Ingredient Soyflakes 1 48.54 16.45 Corn 63.00 Dextrose 42.20 Corn gluten meal 10.10 Soybean oil 4.98 2.50 Ground limestone 1.00 1.39 Dicalcium phosphate 2.20 1.87 Iodized salt 0.40 0.40 Vitamin premix 2 0.20 0.20 Mineral premix 3 0.15 0.15 DL-methionine 0.20 0.18 L-threonine 0.11 L-isoleucine 0.02 L-arginine (free base) 0.25 Sodium bicarbonate 1.00 Choline chloride (60%) 0.10 0.20 Bambermycins (0.44%) 4 0.03 0.025 Solka floc 2.15 Total 100 100 Calculated composition 5 ME n, kcal/kg 3,100 3,200 CP 23 20 Lysine 1.40 0.75 Methionine + cystine 0.87 0.90 Calcium 1.00 1.00 Available phosphorus 0.51 0.45 1 Defatted soyflakes obtained after the expander and before desolventizing-toasting were subjected to different autoclaving times at 121 C and 105 kpa. 2 Provides the following per kilogram of diet: retinyl acetate, 4,400 IU; cholecalciferol, 1,000 IU; DL-alpha-tocopheryl acetate, 11 IU; niacin, 22 mg; D-Ca-pantothenate, 10 mg; riboflavin, 4.4 mg; vitamin B 12, 0.01 mg; menadione sodium bisulfite, 2.33 mg. 3 Provides the following per kilogram diet; Mn, 75 mg (MnO); Fe, 75 mg (FeSO 4 H 2 O); Zn, 75 mg (ZnO); Cu, 5 mg (CuSO 4 H 2 O); I, 0.75 mg (ethylene diamine dihydroiodide); Se, 0.1 mg (Na 2 SeO 3 ). 4 Flavomycin: Hoechst-Roussel Agricultural Veterinary Co., Somerville, NJ. 5 The analyzed CP content of the soyflakes in Assay 1 and Assays 2 and 3 was 47 and 46%, respectively (American Association of Analytical Chemists, 1980). All other values for the soyflakes were assumed to be the same as the NRC (1994) values for dehulled soybean meal. The ME n of dextrose was assumed to be 3,450. All values for other ingredients were taken from the NRC (1994).

1594 BATAL ET AL. TABLE 2. Effect of autoclaving soyflakes on chick performance, protein solubility, urease index, protein dispersibility index, and trypsin inhibitor (Chick Assay 1) 1 Weight Gain:feed Urease dispersibility Trypsin Item gain 2 ratio 3 solubility 4 index index 4 inhibitor 5 (g) (g:g) (%) (units of ph change) (%) (units/g) 0 178 c 0.578 bc 97 2.40 76 44.2 6 180 bc 0.557 c 93 2.20 63 31.0 12 189 b 0.599 b 93 2.10 63 26.8 18 204 a 0.671 a 94 1.80 47 12.3 24 207 a 0.685 a 81 0.20 30 3.4 30 205 a 0.678 a 81 0.30 32 4.5 36 210 a 0.682 a 78 0.10 24 2.6 Soybean meal 210 a 0.693 a Pooled SEM 3 0.010 1.0 2.0 2.08 a c Means in a column with no common superscript are significantly different (P 0.05). 1 Values for weight gain and gain:feed ratio are means of four pens of five chicks from 8 to 17 d of age, and values of protein solubility, urease index, protein dispersibility index, and trypsin inhibitor are means of duplicate analyses on the soyflakes. 2 Quadratic increase as a function of increased autoclaving time (P < 0.07). 3 Linear increase as a function of increased autoclaving time (P < 0.001). 4 Linear decrease as a function of increased autoclaving time (P < 0.001; r 2 = 0.94). 5 Quadratic decrease as a function of increased autoclaving time (P < 0.001). Using the Statistical Analysis System (SAS Institute Inc., 1990), the data from each chick assay were analyzed on a pen basis by a one-way ANOVA. Significant differences between treatment means were assessed using the least significance difference test (Steel and Torrie, 1980). Regression analysis for linear and quadratic effects was also used to further evaluate the responses to autoclaving times (Steel and Torrie, 1980). RESULTS As autoclaving time increased to 18 min in Chick Assay 1, chick weight gain and gain-to-feed ratio increased, with no further increase at longer heating times (Table 2). Urease index and KOH solubility remained high and relatively constant during the 0- to 18-min autoclaving times (1.8 to 2.4 ph units and 94 to 97%, respectively), and then decreased at 24 min and remained relatively constant through 36 min. Conversely, as the SF were autoclaved from 0 to 36 min, PDI decreased incrementally from 76 to 24%. Most importantly, there was a 29% unit decrease in PDI (76 to 47%) during the critical 0- to 18-min autoclaving range. Trypsin inhibitor levels in the SF decreased from 44 to 3 units per g as the SF were autoclaved from 0to36min. In Chick Assay 2, chick growth and gain-to-feed ratio increased when SF were autoclaved for 10 min, with no further significant changes at longer autoclaving times (Table 3). During the first 10 min of autoclaving, there was a small reduction in the urease index, but it still remained high at 1.65 units of ph change. Urease index then dropped to 0.02 at 20 min of heating and remained at that level at 30 min of autoclaving. The KOH solubility decreased from 89 to 79% as autoclaving time increased from 0 to 10 min, remained approximately the same at 20 min, and then decreased again at 30 min. dispersibility index displayed large incremental decreases from 71 to 14% as heating times increased from 0 to 30 min. A 26% unit decrease in PDI (71 to 45%) was observed between 0 and 10 min autoclaving, when the significant difference in growth performance was observed. Chick growth increased as the SF autoclaving time increased from 0 to 9 min in Chick Assay 3, with no further increase at 12 min (Table 4; significant contrast for 3 and 6 min versus 9 and 12 min; P < 0.05). Feed efficiency was increased by 3 min of autoclaving, with no further significant increase at longer times of 3 to 12 min when individual treatments were compared; however, there was a significant linear increase from 0 to 12 min (P < 0.05). Changes in KOH solubility with increased heating time were inconsistent, with a decrease between 3, 6, and 9 min, and then an increase at 12 min. The PDI decreased incrementally from 84% at 0 min to 40% at 12 min of autoclaving. Thus, decreases in PDI were again quite consistent and large. DISCUSSION The overall results of this study indicated that PDI demonstrated more consistent responses to heating of SF than did urease index or protein solubility in KOH. In agreement with earlier studies (McNaughton et al., 1981; Araba and Dale, 1990b; Anderson-Haferman et al., 1992), urease ph change generally remained very high during the shorter SF heating times and then usually decreased precipitously to levels of 0.02 to 0.3. Moreover, there often was little or no change in urease for the shorter autoclaving times that produced the largest changes in chick growth performance (i.e., 0 to 18 min in Assay 1 and 0 to 10 min in Assay 2). The inconsistent and nonlinear

PROTEIN DISPERSIBILITY INDEX 1595 TABLE 3. Effect of autoclaving soyflakes on chick performance, protein solubility, urease index, and protein dispersibility index (Chick Assay 2) 1 Weight Gain:feed Urease dispersibility Item gain 2 ratio 3 solubility 4 index index 4 (g) (g:g) (%) (units of ph change) (%) 0 203 b 0.483 b 89 2.03 71 10 250 a 0.535 a 79 1.65 45 20 251 a 0.531 a 76 0.02 23 30 259 a 0.528 a 67 0.02 14 Pooled SEM 8.3 0.009 1.2 0.8 a b Means in a column with no common superscript differ significantly (P 0.05). 1 Values for weight gain and gain:feed ratio are means of four pens of four chicks from 8 to 17 d of age, and values of protein solubility, urease index, and protein dispersibility index are means of duplicate analyses on the soyflakes. 2 Quadratic increase as a function of increased autoclaving time (P < 0.07). 3 Quadratic increase as a function of increased autoclaving time (P < 0.01). 4 Linear decrease as a function of increased autoclaving time (P < 0.05; r 2 = 0.86). nature of the urease index to heating of SF contributes to the inconsistency among research studies for results on maximum acceptable levels of urease (ph change) in soybean meal. Although the usual recommended urease ph change is 0.05 to 0.2 (Balloun, 1980), Waldroup et al. (1985) reported that a ph rise of up to 0.5 units was acceptable for SBM processed for broiler chickens. Mian and Garlich (1995) reported that in one experiment, the growth performance of young turkeys fed an SBM containing a ph rise of 1.76 was similar to that of turkeys fed an SBM with a ph rise of 0.02. We found that SF with urease ph change values of 1.65 to 1.8 yielded maximum chick growth performance in two of the three chick assays conducted in the current study. solubility in KOH also often did not change consistently as SF were heated. As observed for urease index, the inconsistent response for KOH was particularly evident for the shorter heating times, at which the largest changes in chick growth performance occurred. These results agree with earlier conclusions in our laboratory that KOH protein solubility is a better indicator of overprocessing than underprocessing of soybeans or SBM (Parsons et al., 1991; Anderson-Haferman et al., 1992). The PDI generally produced large and consistent decreases as SF were heated, particularly for the shorter SF heating times that most affected chick performance. These results suggest that PDI is a better indicator of minimum adequate heating of SBM than is urease or KOH solubility. Our results generally indicate that SBM containing a PDI of 45% or lower is adequately heat processed. This value is somewhat higher than the range of 15 to 30% recommended by the National Soybean Processors Association (Balloun, 1980). Of course, further work is needed to determine the optimum range and maximum level for PDI in commercially-processed SBM, because all of our heating treatments were done in a laboratory autoclave. Con- TABLE 4. Effect of autoclaving soyflakes on chick performance, protein solubility, urease index, and protein dispersibility index (Chick Assay 3) 1 Weight Gain:feed Urease dispersibility Item gain 2,3 ratio 4 solubility 5 index index 5 (g) (g:g) (%) (units of ph change) (%) 0 219 b 0.472 b 90 1.76 84 3 257 a 0.496 ab 85 0.86 72 6 259 a 0.505 a 84 0.48 69 9 275 a 0.512 a 72 0.14 55 12 277 a 0.521 a 79 0.02 40 Pooled SEM 7 0.01 1.0 3.3 a b Means in a column with no common superscript differ significantly (P 0.05). 1 Values for weight gain and gain:feed ratio are means of four pens of four chicks from 8 to 17 d of age, and values of protein solubility, urease index, and protein dispersibility index are means of duplicate analyses on the soyflakes. 2 Using single-degree of freedom contrast, the weight gain for 3 and 6 min autoclaving time versus 9 and 12 min autoclaving time is significantly different (P < 0.05). 3 Quadratic increase as a function of increased autoclaving time (P < 0.05). 4 Linear increase as a function of increased autoclaving time (P < 0.01). 5 Linear decrease as a function of increased autoclaving time (P < 0.001; r 2 = 0.96).

1596 BATAL ET AL. sequently, our results may not be directly extrapolated to commercial SBM samples. Perhaps combining the PDI test with the urease test would be useful to soybean processors and poultry nutritionists for better monitoring of SBM quality. For example, an SBM containing low urease (0.3 or below) and high PDI (40 to 45%) may indicate that the sample is definitely high quality because it has been adequately heat processed, but not overprocessed. ACKNOWLEDGMENT Appreciation is expressed to Owensboro Grain Company (Owensboro, KY 42302) for providing the SF used herein. REFERENCES American Association of Cereal Chemists, 1995. Approved Method for Trypsin Inhibitor, Method 71-10. 9th ed. American Association of Cereal Chemists, St. Paul, MN. American Oil Chemists Society, 1980a. Urease Activity. Official Method Ba 9-58. American Oil Chemists Society, Champaign, IL. American Oil Chemists Society, 1980b. Dispersibility Index. Official Method Ba 10-65. American Oil Chemists Society, Champaign, IL. Anderson-Haferman, J. C., Y. Zhang, C. M. Parsons, and T. Hymowitz, 1992. Effect of heating on the nutritional quality of Kunitz-trypsin-inhibitor-free and conventional soybeans for chicks. Poultry Sci. 71:1700 1709. Araba, M., and N. M. Dale, 1990a. Evaluation of KOH solubility as an indicator of overprocessing soybean meal. Poultry Sci. 69:76 83. Araba, M., and N. M. Dale, 1990b. Evaluation of protein solubility as an indicator of underprocessing of soybean meal. Poultry Sci. 69:1749 1752. Association of Official Analytical Chemists, 1980. Official Methods of Analysis. 13th ed. Association of Official Analytical Chemists, Washington, DC. Balloun, S. L., 1980. Soybean Meal in Poultry Nutrition. American Soybean Association, St. Louis, MO. McNaughton, J. M., and F. N. Reece, 1980. Effect of moisture content and cooking time on soybean meal urease index, trypsin inhibitor content, and broiler growth. Poultry Sci. 59:2300 2306. McNaughton, J. M., F. N. Reece, and J. W. Deaton, 1981. Relationships between color, trypsin inhibitor contents, urease index of soybean meal and effects on broiler performance. Poultry Sci. 60:393 400. Mian, M. A., and J. D. Garlich, 1995. Tolerance of turkeys to diets high in trypsin inhibitor activity from undertoasted soybean meals. Poultry Sci. 74:1126 1133. National Research Council, 1994. Nutrient Requirements of Poultry. 9th rev. ed. National Academy Press, Washington, DC. Parsons, C. M., K. Hashimoto, K. J. Wedekind, and D. H. Baker, 1991. Soybean KOH solubility in potassium hydroxide: An in vitro test of in vivo protein quality. J. Anim. Sci. 69:2918 2924. SAS Institute Inc., 1990. SAS Users Guide: Statistics. Version 6, 4th ed. SAS Institute Inc., Cary, NC. Sasse, C. E., and D. H. Baker, 1973. Availability of sulfur amino acids in corn and corn gluten meal for growing chicks. J. Anim. Sci. 37:1351 1355. Steel, R.G.D., and J. H. Torrie, 1980. Principle and Procedure of Statistics, A Biometrical Approach. 2nd ed. McGraw-Hill Book Co. Inc., New York, NY. Waldroup, P. W., B. E. Ramsey, H. M. Helwig, and N. K. Smith, 1985. Optimum processing for soybean meal used in broiler diets. Poultry Sci. 64:2314 2320.