In vitro assessment of the starch digestibility of western Canadian wheat market classes and cultivars

Abstract: The objective of the study was to measure the effect of wheat market class and cultivar on starch digestibility using an in vitro model that mimics the chicken digestive tract and relate it to grain characteristics. The study evaluated 18 wheat cultivars from eight western Canadian wheat classes and, each cultivar was replicated four times. Samples were subjected to gastric and small intestine (SI) digestion phases and each sample was assayed in triplicate; glucose release was measured in SI phase. Starch granule distribution, amylose, total starch, crude protein (CP), ash, and non-starch polysaccharides (NSP) were analyzed in all wheat samples. Small intestinal phase times of 15, 60, and 120 min were chosen to approximate digestion in the terminal duodenum, jejunum, and ileum. Starch digestibility of wheat classes ranged as follows: 15 min — 33.1% to 49.1%, 60 min — 80.2% to 93.3%, and 120 min — 92.4% to 97.6%. Starch digestibility positively correlated with CP, ash, NSP, and proportion of large granules, whereas it negatively correlated with total starch, and proportion of small and medium granules. In conclusion, market class and cultivar of western Canadian wheat affects both rate and extent of starch digestibility and it is related to various grain characteristics.

acids that results in complex formation on the surface of starch granules (Svihus et al. 2005).
Starch digestibility is reduced with a higher proportion of long amylopectin chains because longer amylopectin chains form longer helices, and increase stabilization by hydrogen bonds. In wheat, starch granule diameter ranges from around 1 to 50 µm and there are different classifications of starch granules according to their size (Raeker et al. 1998;Ahuja et al. 2013). A higher proportion of small starch granules theoretically increases starch digestibility due to the increased starch granule surface area accessible by digestive enzymes (Svihus et al. 2005).
It was hypothesized that wheat market class and cultivar impact the rate and extent of in vitro starch digestibility because of differences in relevant grain characteristics. The objectives of this research were to determine the effect of wheat market class and cultivar on the rate and extent of in vitro starch digestibility, and to examine the relationship of these values to grain characteristics.

MATERIALS AND METHODS
An experiment was conducted to determine the rate and extent of starch digestion using an in vitro model of the chicken digestive tract. The study used 18 spring wheat cultivars, consisting of four independent samples for each cultivar obtained from the Crop Development Centre at the University of Saskatchewan. The cultivars were grown on fallow land in a Bradwell clay loam soil type at the University of Saskatchewan's North Seed Farm at Saskatoon, SK in 2012. The four samples of each cultivar were grown on different plots. The wheat cultivars tested and the market classes that they belong to are shown in Table 1. In vitro starch digestion was studied using a procedure that approximates the chicken gastric and SI digestion phases (Ebsim, 2013). The gastric phase contributes to sample mixing and moistening as well as exposure of samples to hydrochloric acid (HCl) and pepsin, which may increase digestive enzyme access to starch in the SI phase. In the SI phase, starch is hydrolyzed to glucose by the action of amylase (derived from pancreatin), amyloglucosidase and invertase enzyme activities. Protease and lipase activity derived from pancreatin may also benefit starch digestion by hydrolyzing lipid and protein blocking amylase access to starch. The released glucose is measured at different incubation times after the start of the SI phase using a glucose oxidase method. Digested starch is calculated based on released glucose, and starch digestibility is estimated based on the digested starch content in relationship to the total starch content of each wheat sample.
The in vitro starch digestibility method used in this research was primarily based on previously published in vitro methods (Englyst et al. 1992;Bedford and Classen 1993) with modifications according to Ebsim (2013). Englyst et al. (1992) established an in vitro method to measure starch digestibility in the SI phase of humans, while Bedford and Classen (1993) designed an in vitro digestion method to predict the intestinal viscosity of broiler chickens fed rye based diets with different dietary pentosanase levels. The gastric phase conditions from the latter research were used for the current in vitro starch digestibility assay. To more accurately reflect in vivo digestive tract conditions in the chicken, an incubation temperature of 41°C was used instead of 37°C, a SI buffer pH of 5.6 was used instead of 5.2 (Ebsim, 2013) and SI enzyme levels were increased to increase the rate of starch digestion to more closely match in vivo Louis, MO. USA) to 9 centrifuge tubes followed by 20 ml of distilled water. The solution was stirred magnetically for 10 min, and centrifuged for 10 min at 3000 rpm. Fourteen millilitres of supernatant from each tube were then added to a beaker (total 126 ml was adjusted to 5.6 using acetic acid and the volume adjusted to 1.0 l with distilled water. Finally 4 ml of 1 M CaCl 2 was added to 1.0 l of the buffer. The glucose oxidase peroxidase determination (GOPOD) reagent from Megazyme (D-Glucose Assay Procedure-GOPOD format, K-GLUC 07/11, Megazyme International Ireland, Bray, Co. Wicklow, Ireland) was used for the glucose oxidase method. Distilled water was added into the glucose reagent buffer (50 ml) until it reaches 1.0 l, and then GOPOD reagent was dissolved in the buffer (Ebsim, 2013).
Samples were fine ground using a Retsch laboratory mill (Retsch ZM 200,Germany) using a screen-hole size of 0.5 mm; fine grinding was used to mimic the impact of the chicken's gizzard. Three replications of approximately 700 mg of each wheat sample were weighed, and added into 50 ml polypropylene centrifuge tubes, and 50 mg of guar gum powder was also added to each tube to standardize the viscosity. A blank tube containing 50 mg of guar gum powder was used to correct glucose content in the amyloglucosidase solution, and was used as the blank sample. A starch standard was prepared by adding regular maize starch and guar gum powder into a tube. In vitro starch digestion was completed on a set of 9 wheat samples at a time.
Initially, 1.5 ml of enzyme solution І (2000 U/ml pepsin-HCl solution) was added to each centrifuge tube. Then tubes were capped, mixed on a vortex mixer and placed horizontally in a water bath (41°C) for 30 min. The enzyme solution ІІ was prepared during this time period.
Tubes were taken out of the water bath after 30 min, and three glass balls (1.5 cm diameter) were added to each tube. Then 20 ml of sodium acetate buffer (41°C) was added to each sample, standard and blank tube, capped and vortexed. For the SI phase 5 ml of enzyme solution ІІ was added to each tube, and then the tubes were capped, vortexed and immediately securely placed in a shaking water bath (41°C). The shaking water bath was set at a stroke length of 35 mm and 160 strokes per min. Timing was started immediately after adding enzyme solution to the first tube.
In this phase, starch is digested into maltose, isomaltose and dextrin by α-amylase and further hydrolyzed into glucose by amyloglucosidase. Sucrose present in wheat is hydrolyzed into glucose and fructose by the action of invertase enzyme. Aliquots (0.5 ml) were taken from each tube at 15,30,45,60,90,120,180 and 240 min of the SI phase and added to 50 ml polypropylene centrifuge tubes containing 20 ml of absolute ethanol (stop the enzyme reaction).
During aliquot removal, tubes were individually removed from the water bath, mixed before taking aliquots, and immediately returned to the water bath (30s for each tube to undergo this procedure).
Ethanol tubes which contained aliquots were centrifuged at 1500 rpm for 2 min to obtain a clear supernatant. The amount of released glucose was measured colourimetrically according to a glucose oxidase method of a Megazyme kit (D-Glucose Assay Procedure-GOPOD format, K-GLUC 07/11, Megazyme International Ireland, Bray, Co. Wicklow, Ireland).
Starch digestibility was calculated using the following formula.

Starch digestibility (%) = (TS in-vitro / TS) × 100
where TS in-vitro is the digested starch at a particular SI incubation time and TS is the total starch of the wheat sample.
All wheat samples were analyzed in duplicate for total starch, CP, ash, soluble and insoluble non-starch polysaccharides (NSP), soluble and total arabinoxylans (AX), amylose and software was used to estimate starch granule size distribution by volume.
The experiment was a Complete Randomized Design and the wheat cultivars were nested within wheat market class. Wheat class and cultivar were random effects. Four different samples (replications) were used from each cultivar. All data were analyzed using Proc Mixed in SAS (SAS 9.4, Carey, N.C. 2008) and Tukey's studentized range test was used for mean separation of treatments when there was a significant difference. Differences were considered significant when P ≤ 0.05. Correlations of in vitro starch digestibility with each grain characteristic and correlations among grain characteristics were determined using Proc Corr in SAS (SAS 9.4, Carey, N.C. 2008). Further, stepwise regression with forward selection was done using Proc Reg to determine the factors most affecting in vitro starch digestibility for each SI incubation time and, the prediction equations were developed for starch digestibility at each SI phase incubation time.

RESULTS
The nature of starch digestion pattern for wheat samples in the in vitro assay was as  Table 2). The range in digestibility for each time period tended to decrease with increasing digestion time with a maximum range of 21% at 30 min and 3.8% at 240 min. Based on incubation time in the SI phase and Ebsim (2013) data, 15, 60 and 120 min were assumed to be representative of in vivo starch digestibility in the terminal duodenum, jejunum and ileum, respectively. These values were considered important in assessing rate of starch digestibility, and will be described in more detail. Starch digestibility of wheat classes at 15 min ranged from 33.1 (Spelt) to 49.1% (CWAD) with an overall difference between the minimum and maximum values of 16%. At 60 min, a 13.1% difference was found between the minimum value of 80.2% (CWRS) and the maximum value of 93.3% (CWAD). At 120 min, the range was from 92.4 (CWRS) to 97.6% (CWES), and the difference was 5.2%. At 15 min, the CWAD class resulted in the highest digestibility, followed by CWES and CWGP, and the remainder of the classes being lowest and statistically equal. Canadian Western Amber Durum maintained the highest digestibility at 60 min, followed by CWGP, which was not higher than CWES, but was higher than the remaining classes. Starting with CWES the digestibility ranking for the remaining cultivars was CWES, CWSWS, CPS, CWHWS, Spelt and CWRS (see Table 2 for statistical separation of means). At 120 min, CWES and CWAD demonstrated the highest digestibility followed by, but not different than, CWSWS and CPS. The numerical ranking from high to low digestibility for the remaining cultivars was CWGP, CWHWS, Spelt and CWRS (see Table 2 for statistical interpretation). Table   3. Similarly to class, cultivar affected starch digestibility at all time periods. In vitro starch digestibility (%) of wheat cultivars at 15 min ranged from 32.6 (CDC Zorba) to 51.6% (Transcend) with a maximum difference of 19.0%. At 60 min, digestibility ranged from 77.3 (Glenn) to 94.8% (CDC Verona) resulting in a maximum difference of 17.5%. At 120 min, the range was from 91.0 (CDC Origin) to 100.0% (Transcend) with a difference between these means of 9.0%. Similarity of cultivars within a class can be estimated based on separation of cultivar means. When this is done, differences among cultivars within a class were found for CPS (120 min), CWAD (15,30,90,120,180    Stepwise regression analysis revealed the grain characteristics that explained the most variation in starch digestibility at different SI phase incubation times of the in vitro assay (Table 9) and, grain characteristics were used to develop the prediction equations for in vitro starch digestibility (Table 10). Regression coefficient values were cumulative for each of the time periods.

Examination of variation in in vitro starch digestibility among cultivars is shown in
Correlation analyses among grain characteristics are presented in Table 11.  were affected by wheat market class and cultivar. Total starch content ranged from 52.2 to 58.8%, which is less than previously published values ranging from 68.6 to 69.8% (Hucl and Chibbar 1996). In contrast, crude protein values of the wheat classes ranged from 15.0 to 22.9% and, were higher than the 12.8 to 17.0% range reported by Hucl and Chibbar (1996). Appropriate standards and repeat analyses of samples confirmed the original analysis suggesting that analytical errors were not responsible for the variation in starch and protein levels from expected values. Grain growing conditions can have an important impact on nutrient content, and this may have been the case for these samples. Samples originating from research plots tend to have higher nitrogen fertilization rates than commercial production and this may have been a reason    With the exception of soluble AX, grain fibre content as estimated by measurement of NSP and AX was positively correlated with starch digestibility with the relationships stronger earlier in the SI phase. Stepwise regression similarly showed a strong association of total NSP, soluble NSP, total AX and insoluble AX with in vitro starch digestibility mostly in the earlier SI phase incubation times. One or more of the above fibre fractions were able to predict in vitro starch digestibility in those earlier incubation times of the SI phase. However, soluble AX was not associated with starch digestion. Relatively strong correlations among fibre fractions preclude assigning responsibility to a specific fibre fraction. In general, a positive relationship is opposite to a generally accepted negative association between soluble fibre and digestibility (Classen 1996). Soluble NSP in wheat, mainly soluble AX, increase viscosity of digesta, decrease digesta passage rate, reduce the interaction between digestive enzymes and substrates  conditions of the in vitro assay. The dry matter content of the in vitro model is much less than the digesta dry matter content in the middle to distal portion of the SI in chickens. The viscosity of a solution is strongly affected by its moisture content (Scott 2002), and therefore viscosity is much lower inside the centrifuge tubes in the in vitro assay compared to digesta in the chicken GI tract.

DISCUSSION
As a result, digesta viscosity is less likely play a negative role in in vitro starch digestibility.  Although starch digestion rate is not an indication of complete starch digestion, it provides valuable information on where digestion occurs and as a consequence effects related to that site.
For instance, starch digestion rate may impact post-prandial metabolism, nutrient utilization and also GI tract health in animals (Regmi et al. 2011 a, b;Yin et al. 2011). Therefore, the effects of class and cultivar on the rate and extent of starch digestion both have value in selecting superior wheat cultivars for poultry feed to improve production and health in poultry.
The in vitro starch digestion model is a repeatable assay and was able to demonstrate genotypic differences in both estimated rate and extent of starch digestion. In addition, it requires less time and cost in comparison with in vivo broiler chicken experiments and therefore, it can be used to test a large number of starch containing feed ingredients. Limitations of the in vitro assay are that it can't exactly mimic the dynamic and adaptable nature of the chicken digestive tract. Therefore, it serves as a procedure to select promising effects and differences for later in vivo