Efficacy of Different Commercial Phytase Sources and Development of a Phosphorus Release Curve

Two experiments used 184 pigs (PIC, 22.7 and 21.3 lb BW, respectively) to develop an available P (aP) release curve for commercial phytase products. In Exp. 1 and 2, pigs were fed a basal diet (0.06% aP) and 2 levels of added aP from inorganic P (monocalcium P) to develop a standard curve. In Exp. 1, 100, 175, 250, or 500 phytase units (FTU)/kg OptiPhos (Enzyvia LLC, Sheridan, IN) or 200, 350, 500 or 1,000 FTU/kg Phyzyme XP (Danisco Animal Nutrition, Marlborough, UK) was added to the basal diet. In Exp. 2, 250, 500, 750, or 1,000 FTU/kg OptiPhos; 500, 1,000, or 1,500 FTU/kg Phyzyme XP; or 1,850 or 3,700 phytase units (FYT)/kg Ronozyme P (DSM Nutritional Products, Basel, Switzerland), was added to the basal diet. Manufacturerguaranteed phytase levels were used in diet formulation. Diets were analyzed for phytase using both the Phytex and AOAC methods. Pigs were blocked by sex and weight and allotted to individual pens with 8 pens per treatment. Pigs were euthanized on d 21, and fibulas were analyzed for bone ash. In Exp. 1, pigs fed increasing monocalcium P had improved (linear; P = 0.01) ADG, G/F, and percentage bone ash. Similarly, pigs fed increasing monocalcium P in Exp. 2 tended to have improved (quadratic; P = 0.09) ADG in addition to significantly improved (linear; P ≤ 0.001) G/F and percentage bone ash. In Exp. 1, pigs fed increasing OptiPhos had increased (linear; P ≤ 0.02) ADG, G/F, and percentage bone ash. Likewise, pigs fed increasing OptiPhos in Exp. 2 had improved (linear; P ≤ 0.001) ADG and G/F, as well as increased (quadratic; P ≤ 0.001) percentage bone ash. In Exp. 1, pigs fed increasing Phyzyme XP had increased (linear; P ≤ 0.04) ADG and G/F and tended to have improved (linear; P = 0.06) percentage bone ash. Pigs fed increasing Phyzyme XP in Exp. 2 had increased (quadratic; P ≤ 0.001) G/F and percentage bone ash. In Exp. 2, pigs fed increasing Ronozyme P had improved (linear; P ≤ 0.001) ADG in addition to increased (quadratic; P ≤ 0.03) G/F and percentage bone ash. When AOAC analyzed values and bone ash are used as the response variable, aP release for up to 1,000 FTU/kg of Escherichia coli-derived phytases (OptiPhos and Phyzyme XP) can be predicted by the equation (y = -0.000000125x2 + 0.000236245x + 0.015482000), where x is the phytase level in the diet.


Introduction
Phosphorus is one of the most significant minerals in swine nutrition. It is essential for bone development, plays a key role in metabolic processes such as the formation of cellular membranes, and is vital for enzymatic systems involved in fat and carbohydrate metabolism.
In cereal grains and oilseed meals, a large amount of P is in the form of phytic acid (myo-inositol hexaphosphate). The P in phytic acid is largely unavailable to the pig. Thus, a phytase enzyme is added to diets to enhance the pig's ability to use P from phytic acid. Many trials have been conducted to evaluate different sources of the phytase enzyme, including some prominent versions of the enzyme obtained from Escherichia coli or Aspergillus oryzae.
Because manufacturers have their own individual analytical techniques, it is often confusing to compare phytase sources by a single analytical method. To avoid this confusion, the current study used inclusion rates as directed by the product labels, which gives field-applicable available P (aP) release values. To further clarify comparisons, the current industry standard analysis (AOAC) was also conducted on all phytase samples.
Current data from JBS United demonstrates that 0.12% aP can be replaced in a cornsoybean meal-based diet with 250 phytase units (FTU)/kg OptiPhos (Enzyvia LLC, Sheridan, IN). Recommendations for Phyzyme XP (Danisco Animal Nutrition, Marlborough, UK) and Ronozyme P (DSM Nutritional Products, Basel, Switzerland) are that 500 FTU/kg or 1,850 phytase units (FYT)/kg, respectively, should be used to replace 0.10% aP. Phytase may be added at levels less than that needed to replace the 0.12% or 0.10% P. However, more data is needed to determine a response curve for OptiPhos, Phyzyme XP, and Ronozyme P. The development of dose response curves for P release could allow the optimum use of the different sources of the enzyme at all levels.
Our objectives for these trials were to evaluate the effects of three different sources of commercially available phytase on late nursery pig performance and to develop a P release curve.

Procedures
In Exp. 1, a total of 88 barrows (initially 22.7 lb) were used in a 21-d growth trial. Pigs were blocked by weight and allotted to 1 of 11 dietary treatments. In Exp. 2, a total of 104 pigs (initially 21.3 lb) were used in a 21-d growth trial. Pigs were blocked by sex and weight and allotted to 1 of 13 dietary treatments. In both experiments, there was 1 pig per pen and 8 pens per treatment. Each pen (31.6 × 39 in.) contained a 2-hole, dry self-feeder and a nipple water to provide ad libitum access to feed and water. The study was conducted in 4 adjacent rooms in the Discovery Nursery at JBS United's Burton Russell Research Farm in Frankfurt, IN. Samples of phytase and inorganic phosphorus premixes and complete feed were taken at the time of diet preparation and analyzed for phytase.
A common starter diet (meal form) containing 0.06% aP was fed to pigs for 6 d prior to the experiment while pigs were being acclimated to the barn. In Exp. 1 and 2, pigs were fed a basal diet (0.06% aP) and 2 levels of added aP monocalcium P (0.075 and 0.15 for Exp. 1 and 0.07 and 0.14 for Exp. 2) to develop a standard curve. In Exp. 1,100,175,250,or 500 FTU/kg OptiPhos or 200,350,500, or 1,000 FTU/kg Phyzyme XP was added to the basal diet. In Exp. 2, 250, 500, 750, or 1,000 FTU/kg OptiPhos; 500, 1,000, or 1,500 FTU/kg Phyzyme XP; or 1,850 or 3,700 FYT/kg Ronozyme P was added to the basal diet.
In Exp. 1, all treatment diets were constructed from a single basal diet (Table 1) made in two batches at the Kansas State University (K-State) Animal Science Feed Mill. Each bag was marked with batch and bagging order. The first 3 and last 2 bags of each batch were not used in diet preparation. Individual treatments were mixed from the basal diet at the K-State Poultry Feed Mill. A total of 197.5 lb of each batch of the basal diet were used to create 395 lb of each treatment diet. Each of the 2 batches contributed 98.75 lb (a total of 197.5 lb) and was mixed for 2 min. Five pounds (2 lb phytase premix and 3 lb P premix) of premix was added to the mixer while the mixer hands were on the upside, and the diet was mixed for an additional 2 min. The additional 98.75 lb of each batch of the basal diet was added, and the diet was mixed for an additional 2 min. Approximately 30 lb of feed was removed from the mixer discharge and deposited back into the top of the mixer. The treatment was mixed for an additional 6 min, for a total treatment addition mixing time of 12 min. Treatments were bagged into 30-lb bags and tagged with labels including the K-State and JBS United protocol number and correlating treatment letter.
In Exp. 2, premixes were manufactured at K-State and shipped to Sheridan, IN, where they were added to a single basal diet (Table 1), which was made in 3 batches at the Burton Russell Research Farm Feed Mill in Frankfort, IN. Each bag was marked with batch and bagging order. The first and last 2 bags of each batch were not used in diet preparation trial. A total of 92, 152, and 150 lb of batches 1, 2, and 3 of the basal diet, respectively, were used to create 394 lb of each treatment diet. Half of each batch (a total of 197 lb) was added to the mixer and mixed for 2 min. Six pounds (2 lb phytase premix and 4 lb inorganic P premix) of premix was added to the mixer while the mixer hands were on the upside, and the diet was mixed for an additional 2 min. The remainder each batch of the basal diet was added, and the diet was mixed for an additional 2 min. Approximately 30 lb of feed was removed from the mixer discharge and deposited back into the top of the mixer. The treatment was mixed for an additional 2 min for a total treatment addition mixing time of 8 min. Treatments were bagged into 30-lb bags and tagged with labels including the K-State and JBS United protocol number and correlating treatment letter.
In both experiments, treatment premixes were made at the K-State Swine Research Laboratory. The phytase premixes consisted of a phytase source (OptiPhos, Phyzyme XP, or Ronozyme P) and/or cornstarch. The same lot of each OptiPhos and Phyzyme XP were used to make both Exp. 1 and 2 premixes. Phytase was stored in a freezer for approximately 3 mo between experiments. The negative control and diets with monocalcium P were made with no phytase and 2 lb of cornstarch. In Exp. 1, a single batch of the 500 FTU/kg OptiPhos premix and the 1,000 FTU/kg Phyzyme XP premix was manufactured and analyzed for lysine, Ca, P, and phytase content (Table 2). Micro-ingredients were also analyzed for Ca (Table 3). In Exp. 2, a single batch of the 1,000 OptiPhos premix, 1,500 FTU/kg Phyzyme XP premix, and 3,700 FYT/kg Ronozyme P premix was made and analyzed for Ca, P, and phytase content. Cornstarch was added in increasing levels to the base mixes to dilute them to the various phytase levels used in the trials. In both experiments, the P premixes consisted of monocalcium phosphate (21% P) and/or sand of similar particle size. The negative control and diets containing phytase were made with no monocalcium P and 3 (Exp. 1) or 4 (Exp. 2) lb of sand. Premixes were analyzed for Ca and P, and phytase analysis was conducted according to the AOAC and Phytex methods (Table 4).
Treatment diets were fed in meal form for 21 d. Average daily gain, ADFI, and G/F were determined by weighing pigs and measuring feed disappearance on d 0 and 21 of the trial. Animals were euthanized via lethal injection with Euthanasia-III Solution (Exp. 1; Med-Pharmex) or Beuthanasia-D Special (Exp. 2; Schering-Plough) according to the K-State Institutional Animal Care and Use Committee standards. The right fibula was removed without cartilage caps from each animal, autoclaved, and boiled for 45 to 60 min. Fibulas were cleaned of adhering tissue, dried at 105°C for 24 h, and ashed in a muffle furnace at 600°C for 24 h. Total ash weight and percentage ash were measured.

Data Analysis
All values that were at least three SD away from the mean of each response criteria were considered outliers. In Exp. 1, 4 pigs with outliers for growth data (ADG, ADFI, or G/F) were removed from both the growth and bone (ash weight and percentage ash) results. Two pigs with outliers for percentage ash were removed from the ash weight and percentage ash results but were used for the calculation of growth data. One pig with an outlier for ash weight was removed from the ash weight results but was used in the calculation of percentage ash and growth data. Three fibulas were broken during analysis, preventing ash weight and percentage ash for these fibulas from being calculated. Growth data from these pigs were used. In Exp. 2, 1 pig was deemed an outlier for G/F and was removed from all data. One pig was considered an outlier for percentage ash and was removed from the ash weight and percentage ash results but was used for the calculation of growth data.
Data were analyzed as a randomized complete block design with pig as the experimental unit. Treatment was fixed, whereas pigs and room were randomly assigned. Analysis of variance was performed using the MIXED procedure of SAS (SAS Institute Inc., Cary, NC). Results were considered to be significant if their P-values were ≤ 0.05 and were considered to be a trend if their P-values were ≤ 0.10. Main effects from Exp. 1 showed that all treatments that included inorganic P remained in the linear portion of the quadratic curve of phytase release, and so all treatments were used for analysis. Conversely, main effects from Exp. 2 showed that the treatment supplemented with an additional 0.21% aP from inorganic P (0.27% total aP) was in the quadratic portion of the phytase curve. Because aP release curves must be generated from data that are only in the linear portion of this curve, the treatment was removed from all data analysis. For reference, adding 0.21% aP from monocalcium P (0.27% total aP) resulted in pigs with an ADG of 1.10 lb/d, an ADFI of 1.57 lb/d, a G/F of 0.70, a bone ash weight value of 775 mg, and a bone ash percentage of 41.9.
A regression equation was calculated for ADG, G/F, ash weight, and percentage ash to predict the percentage aP released from the E. coli-derived phytases, given each response criteria. First, the total intake of aP from the diet was calculated and termed to be the dosage of aP administered to each pig through its diet. Dosage for pigs fed the negative control, OptiPhos, Phyzyme XP, and Ronozyme P diets was the product of 0.06 and individual grams of feed intake. In Exp. 1, dosage for pigs fed the negative control diet plus 0.075% aP from the monocalcium P diet was the product of 0.135 and individual grams of feed intake. Dosage for pigs fed the negative control plus 0.15% aP from the monocalcium P diet was the product of 0.21 and individual grams of feed intake. In Exp. 2, dosage for pigs fed the negative control diet plus 0.07% aP from the monocalcium P diet was the product of 0.13 and individual grams of feed intake. Dosage for pigs fed the negative control plus 0.14 aP from the monocalcium P diet was the product of 0.20 and individual grams of feed intake.
Using these aP dosages, regression was used to determine the aP release from each phytase source for a given aP dosage (intercept) and the aP release from each response variable for a given aP dosage (slope). The percentage aP released from each phytase source (Y) was then calculated by adding the value of aP release from each phytase source for a given aP dosage to the product of the value of aP release from each response variable for a given aP dosage and the value of the response variable (X).

Results
In Exp. 1, lysine and P analysis of the diets resulted in concentrations similar to those used in diet formulation (Table 2). However, Ca levels were higher than expected because of higher than anticipated Ca levels in the microingredients. The high Ca levels resulted in high Ca to total P ratios (2.04 to 2.20) for the negative control and all phytase diets. As previous research suggests, these ratios likely decreased ADG and G/F. However, these ratios did not appear to affect percentage bone ash or the aP release levels calculated from percentage bone ash. Lower Ca:P ratios were used in Exp. 2, in which analysis of the diets resulted in concentrations similar to those used in diet formulation.
According to the AOAC analysis, the phytase concentration in OptiPhos was nearly 3.1 and 2.5 times the concentration listed on the label by the manufacturer for Exp. 1 and 2, respectively (Tables 5 and 6). The phytase level in Phyzyme XP was at the concentration listed on the label by the manufacturer in Exp. 1 and 0.7 times the listed concentration in Exp. 2. Ronozyme P was used and analyzed only in Exp. 2, in which the analyzed values were similar to levels reported on the label by the manufacturer.
Results of the AOAC analysis in both experiments indicated that, as expected, phytase levels increased linearly as more phytase premix was added to the diet. Phytase analysis with the Phytex assay found much lower phytase levels for all premixes and diets. Results from the Phytex analysis assay were not as consistent with added dietary levels as the AOAC assays; however, the Phytex assay was conducted only by one laboratory, whereas the AOAC assay was an average of results from three (Exp. 1) or two (Exp. 2) laboratories. Within laboratory, the Phytex assay was less consistent with our calculated values than any single AOAC assay.
Percentage aP released from each phytase source varied depending on the response criteria used to calculate the value (Table 9). The lowest aP release value for both phytase sources was calculated with ADG as the response criteria. The aP release values calculated with G/F as the response criteria were nearly identical for all levels of Opti-Phos, whereas levels generally increased with increasing Phyzyme XP to an overall release value that was similar for both phytase sources. The aP release values calculated from bone ash weight were similar for all levels of Phyzyme XP, with the exception of 500 FTU/kg. However, the calculated aP release values were not as consistent for OptiPhos, as evidenced by the second lowest phytase dose releasing the highest percentage aP. The clearest response to percentage aP release was calculated with percentage bone ash as the response criteria. As both OptiPhos and Phyzyme XP levels increased, calculated aP increased in a quadratic fashion to the highest phytase dose.
Percentage aP released from each phytase source and level again varied depending on the response criteria used to calculate the value (Table 12). The lowest aP release value for 250 FTU/kg of OptiPhos was calculated from ADG. The lowest aP release values for 500, 750, and 1,000 FTU/kg of OptiPhos was calculated from bone ash weight. In contrast, the highest aP release level for all OptiPhos levels was calculated from bone ash percentage. The lowest aP release level for 500 FTU/kg of Phyzyme XP was calculated from bone ash percentage, whereas the lowest levels for 1,000 and 1,500 FTU/kg of Phyzyme XP were calculated from ADG. The highest aP release level for 500 FTU/ kg of Phyzyme XP was calculated from G/F, whereas the highest levels for 1,000 and 1,500 FTU/kg of Phyzyme XP were calculated from bone ash percentage. Finally, the lowest aP release level for 1,850 and 3,700 FTU/kg of Ronozyme P was calculated from bone ash weight and G/F, respectively. The highest aP release level for both Ronozyme P levels was calculated from bone ash percentage.

Experiments 1 and 2
By using the average values of the AOAC phytase assays from both E. coli phytase sources, the response to various criteria were plotted against the analyzed phytase level. Approximately 77% of the variation in response in percentage bone ash was explained by the analyzed phytase level in the diet (Figure 1). Similarly, by plotting the aP released for each phytase level against the analyzed AOAC phytase level, a P release curve was calculated. With percentage bone ash as the response criteria, approximately 73% of the variation in aP release was explained by the analyzed phytase level in the diet (Figure 2). When AOAC analyzed values and bone ash are used as the response variable, aP release for up to 1,000 FTU/kg of E. coli-derived phytases (OptiPhos and Phyzyme XP) can be predicted by the equation (y = -0.000000125x 2 + 0.000236245x + 0.015482000), where x is the phytase level in the diet.
Previous K-State recommendations, based on Kornegay (1996) P release curves 5 , agree well with the phytase release suggested by the aP curve developed from percentage bone ash ( Figure 3). The curve previously used by K-State was valid only to 700 FTU/kg, whereas the new curve suggested by this research is valid to 1,000 FTU/kg.

Discussion
Higher phytase concentrations in the AOAC analysis compared with the Phytex analysis were expected because of the key differences between the Phytex assay used by the manufacturer of OptiPhos and the AOAC method. The Phytex assay extracts P with a 0.2M sodium citrate buffer, whereas the AOAC assay uses a 0.2M sodium acetate buffer, Tween 20, and bovine serum albumin. The Phytex assay incubation time is 15 min; the AOAC assay incubation time is 60 min. Additionally, the color reagent used to measure the P released from phytic acid has a wavelength of 820 nm in the Phytex assay and 415 nm in the AOAC assay. Finally, the Phytex assay diafiltrates feed samples to remove high background P levels from monocalcium or dicalcium P before they are assayed; the AOAC assay does not.
The influence of E. coli-derived phytase source on level of percentage bone ash follows the typical quadratic response for aP release that has been shown in previous research. The 77% of variation in percentage bone ash that was explained by analyzed phytase value was the highest of any of the measured variables (63, 36, and 39 for ADG, GF, and bone ash weight, respectively). This reinforces that percentage bone ash was the best variable to use to predict aP release. The predicted aP release values from trials in which analyzed AOAC values were used agree largely with Kornegay's summary for E. coli-derived phytase levels, suggesting that we can predict aP release levels from E. coli-derived phytases when their AOAC assayed value is less than 1,000 FTU/kg. More research needs to be conducted to further evaluate release values for higher phytase levels.
In summary, when percentage bone ash was used as the response criteria, the aP release for these phytase sources was similar to the manufacturers' recommendations when the products were used according to label phytase levels (0.12% for 250 FTU/kg of OptiPhos, 0.10% for 500 FTU/kg of Phyzyme XP, and 0.10% for 1,850 FTU/kg of Ronozyme P). When analyzed on an AOAC basis, the aP release curves for the E. coli phytases had similar release curves, at least up to 1,000 FTU/kg.