Title: Molecular rearrangement of waxy and normal maize starch granules during in vitro digestion

Transcription

1 Title: Molecular rearrangement of waxy and normal maize starch granules during in vitro digestion Author: Anju Teng Torsten Witt Kai Wang Ming Li Jovin Hasjim PII: S (15) DOI: Reference: CARP To appear in: Received date: Revised date: Accepted date: Please cite this article as: Teng, A., Witt, T., Wang, K., Li, M., and Hasjim, J.,Molecular rearrangement of waxy and normal maize starch granules during in vitro digestion, Carbohydrate Polymers (2015), This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

2 1 2 Molecular rearrangement of waxy and normal maize starch granules during in vitro digestion Anju Teng, Torsten Witt, Kai Wang, Ming Li, and Jovin Hasjim* The University of Queensland, Centre for Nutrition and Food Sciences, Queensland Alliance for Agriculture and Food Innovation, Brisbane, QLD 4072, Australia *Corresponding author. Current address: Roquette Management (Shanghai) Co., Ltd., Shanghai, , China. Tel: Fax: jovin.hasjim@roquette.com 1 Page 1 of 39

3 14 15 Abstract The objective of the present study is to understand the changes in starch structures during digestion and the structures contributing to slow digestion properties. The molecular, crystalline, and granular structures of native waxy maize, normal maize, high-amylose maize, and normal potato starch granules were monitored using SEC, XRD, DSC, and SEM. The amylose and amylopectin molecules of all four starches were hydrolyzed to smaller dextrins, with some having linear molecular structure. Neither the A- nor B-type crystallinity was resistant to enzyme hydrolysis. Starch crystallites with melting temperature above 120 ºC appeared in waxy and normal maize starches after digestion, suggesting that the linear dextrins retrograded into thermally stable crystalline structure. These crystallites were also observed for high-amylose maize starch before and after digestion, contributing to its low enzyme digestibility. On the contrary, the enzyme-resistant granular structure of native normal potato starch was responsible for its low susceptibility to enzyme hydrolysis. Keywords: starch granules, amylose, in vitro digestion, size-exclusion chromatography, X-ray diffractometry, scanning electron microscopy Abbreviations CLD, Chain length distribution; DP, Degree of polymerization; DSC, Differential scanning calorimetry/calorimeter; H, Enthalpy of starch gelatinization; RDS, Rapidly digestible starch; RS, Resistant starch; SDS, Slowly digestible starch; SEC, Size-exclusion chromatography; SEM, Scanning electron microscopy; T c, Conclusion temperature of starch gelatinization; T o, Onset temperature of starch gelatinization; T p, Peak temperature of starch gelatinization; XRD, X-ray diffractometry. 38 Introduction 2 Page 2 of 39

4 39 40 Starch is the major energy source in most staple foods. Its digestion rate and extent of digestion in the small intestine have a large impact on human health and nutrition According to Englyst et al. (1992), starch can be categorized into three groups, which are rapidly digestible starch (RDS), slowly digestible starch (SDS) and resistant starch (RS). It has been proposed that RDS is completely digested within 20 min, causing high postprandial glycemic and insulinemic responses and therefore increasing the risks of metabolic disorders such as insulin resistance, diabetes, and obesity (Byrnes et al., 1995; Ludwig, 2002; Willett et al., 2002). SDS is digested between 20 and 120 min after consumption and RS is the remnants that are not digested after 120 min. Both SDS and RS can lower postprandial glycemic and insulinemic responses with SDS prolonging the supply of glucose to the body, especially to the brain, which is particularly beneficial for maintaining cognitive functions. Similar to other dietary fibers, RS escapes the digestion in the small intestine and reaches the colon, where it serves as an important substrate for gut fermentation producing short-chain fatty acids with the potential to prevent the development of colon cancer cells (Ferguson et al., 2000; Zhao et al., 2011). The structures of native starch granules can be simplified into five levels (Dona et al., 2010; Tran et al., 2011). Linear branches of starch molecules (Level 1 structure) are made of glucose monomers linked together by α-(1 4) glycosidic linkages, and these linear branches are connected together by α-(1 6) glycosidic linkages forming individual fully branched molecules (Level 2 structure). These lowest two levels are molecular levels composed of mainly amylose and amylopectin. Amylose is primarily linear with a few long branches, whereas amylopectin is highly branched with a vast number of short branches and a molecular size two orders of magnitude larger than 3 Page 3 of 39

5 63 64 amylose. Some branches of amylopectin are arranged in a double helical conformation, packing into crystallites that form alternating crystalline and amorphous layers (semi-crystalline structure, Level 3). In contrast, amylose is typically amorphous or in a single helical complex with lipid molecules. The semi-crystalline structures are arranged into concentric growth rings, which alternate with amorphous growth rings around the amorphous hilum (Level 4 structure) in a single starch granule (Level 5 structure). The rate and extent of starch digestion are largely influenced by starch structures, including the branch-chain length of amylose and amylopectin molecules (Srichuwong et al., 2005; Syahariza et al., 2013), molecular size (Li et al., 2015b), amylose content (Li et al., 2008; Syahariza et al., 2013), amylose-lipid complexes (Ai et al., 2013; Hasjim et al., 2010b), crystalline structure (Hasjim & Jane, 2009; Jane et al., 2003; Sievert & Pomeranz, 1990), granule size and surface structure (Dhital et al., 2010; Hasjim et al., 2009). The changes in starch structure during in vitro and in vivo digestions have recently been investigated. The rearrangement of starch molecules during in vitro digestion of extruded high-amylose maize starches to form enzyme-resistant crystalline structures was reported by several authors (Htoon et al., 2009; Li et al., 2015b; Lopez-Rubio et al., 2008b; Shrestha et al., 2010). The dextrins produced after prolonged in vitro digestion of either extruded maize starches with various amylose contents or native normal maize starch granules were both found to be predominately linear with degree of polymerization (DP) X ~50 (Hasjim et al., a; Witt et al., 2010). Similarly, the digesta collected from the lower-half small intestine of the pigs fed with native normal maize starch granules contained substantial amounts of linear dextrins with DP X ~50 (Hasjim et al., 2010a). Combining the results from these studies, it can be assumed that the linear dextrins 4 Page 4 of 39

6 88 89 produced during starch digestion are able to rearrange themselves into highly ordered crystalline structures that are less susceptible to enzyme hydrolysis. However, this relationship has not been explicitly proven in a single study before. The aim of this study is to understand the molecular rearrangement of starch during digestion and the structures contributing to slow digestion properties by investigating the changes in the molecular, crystalline, and granular structures at different digestion times. Native starch granules (Level 5 structure) were used as the substrates for an in vitro digestion. Although native starch granules are less common in human food than gelatinized starch, they are present in animal feed, fresh produce (fruits and vegetables), and low-moisture foods (such as biscuits and some breakfast cereals), and the lack of retrograded starch in native starch granules allows the distinct observation of starch molecular rearrangement (retrogradation) during digestion. The molecular structures (Levels 1 and 2) of the starch digesta collected during the in vitro digestion were characterized using size-exclusion chromatography (SEC), while their crystalline structures (Level 3) were characterized using X-ray diffractometry (XRD) and differential scanning calorimetry (DSC). The changes on the granular structure (Level 5) were identified using scanning electron microscopy (SEM). Materials and methods Materials Four types of native starch granules were used: waxy maize, normal maize, high-amylose maize (Gelose 50), and normal potato starches. The maize starches were obtained from Penford Australia Ltd. (Lane Cove, NSW, Australia). Normal potato starch, pancreatin from porcine pancreas, and LiBr (ReagentPlus) were obtained from Sigma Aldrich (Castle Hill, NSW, Australia). Amyloglucosidase from Aspergillus niger, isoamylase from Pseudomonas sp., and D-Glucose (GOPOD Format) kit were 5 Page 5 of 39

7 purchased from Megazyme International Ltd. (Bray, Co. Wicklow, Ireland). Dimethyl sulfoxide (DMSO, GR for analysis ACS) was from Merck and Co., Inc. (Kilsyth, VIC, Australia). Other chemicals were reagent grade and used as received. In vitro starch digestion Native starch granules (1.5 g, dry weight basis) were suspended in sodium acetate buffer (5.0 ml, 0.2 M, ph 6.0) containing 200 mm CaCl 2, 0.49 mm MgCl 2, and 0.02 % NaN 3. An enzyme solution (5.0 ml) containing 60 mg pancreatin and 300 µl amyloglucosidase in sodium acetate buffer solution was added to the starch suspension after being equilibrated at 37 C with stirring for 5 minutes. The mixture was incubated at 37 C with stirring. The digestion was stopped by adding 30 ml absolute ethanol to the mixture at 0, 2, 4, 8, and 16 h for waxy maize starch; 0, 2, 4, 8, 16, and 24 h for normal maize starch; and 0, 16, 24, and 48 h for high-amylose maize and normal potato starches. The starch digesta (precipitate) was collected by centrifugation at 4000 g for 10 min and dried in a fume hood at room temperature. The structure of starch digesta was characterized using SEC, DSC, XRD, and SEM. The amount of glucose in the supernatant was determined using the Megazyme D-Glucose kit following the instructions from the manufacturer and converted to the amount of digested starch using the factor of 0.9 (the conversion factor for glucose to anhydroglucose unit in starch). The degree of starch hydrolysis was calculated as follows: % hydrolysis = total weight of glucose in supernatant 0.9 dry weight of starch 100% The concentrations of pancreatin and amyloglucosidase (dry starch weight basis) were higher than previous studies (Hasjim et al., 2010a; Sopade & Gidley, 2009; Witt et al., 2010) in order to produce substantial digestion of the native high-amylose 6 Page 6 of 39

8 maize and normal potato starch granules, which contain large amounts of RS (Dhital et al., 2010; Jane et al., 2003; Tester et al., 2006). The in vitro digestion steps simulating the digestion in the mouth and the stomach were not used in the present study as the starch granules entering the small intestine from a previous in vivo study were not significantly digested in the mouth or the stomach (Hasjim et al., 2010a). The digestion times were selected based on the digestibility of the starches. Because the native high-amylose maize and normal potato starch granules are more resistant to enzyme hydrolysis than the native waxy and normal maize starch granules, the digestions of the native high-amylose maize and normal potato starch granules were carried out for longer times. Size-exclusion chromatography and amylose content Starch digesta (~2 mg) were dissolved in 1 ml DMSO solution containing 0.5% w/w LiBr for 24 h at 80 ºC and 350 rpm for the characterization of whole molecular structure (Level 2 structure) following the method of Syahariza et al. (2010). The preparation of the individual branches of starch molecules (Level 1 structure) for characterization involved dissolving ~6 mg starch digesta in 1 ml DMSO solution containing 0.5% w/w LiBr for 24 h at 80 ºC and 350 rpm, precipitated using 5 ml absolute ethanol, dissolved in 0.9 ml distilled water in a boiling water bath, and then debranched using 2.5 µl isoamylase at 37 C for 3 h (ph was adjusted using 0.1 ml 0.1 M acetate buffer of ph 3.5) following the method described elsewhere (Hasjim et al., 2010a; Tran et al., 2011). The debranched starch was neutralized to ph ~7 with µl 0.1 M NaOH, heated at 80 C for 2 h, and freeze dried. The dried debranched starch was redissolved in 1 ml DMSO solution containing 0.5% w/w LiBr at 80 ºC and 350 rpm for 24 h. 7 Page 7 of 39

9 The structures of whole and debranched starch molecules were characterized in duplicate using SEC, also known as gel permeation chromatography (GPC). The SEC system consisted of an Agilent 1100 Series (Agilent Technologies, Waldbronn, Germany) equipped with a refractive index detector (RID-10A, Shimadzu Corp., Kyoto, Japan) following the method described elsewhere (Cave et al., 2009; Liu et al., 2010; Witt et al., 2010). GRAM 30 and GRAM 3000 SEC columns (Polymer Standards Service GmbH, Mainz, Germany) were used to separate whole starch molecules, whereas GRAM 100 and GRAM 1000 SEC columns (Polymer Standards Service GmbH) were used to separate debranched starch molecules. DMSO with 0.5% w/w LiBr was used as the eluent with flow rates of 0.3 and 0.6 ml/min for whole and debranched starch molecules, respectively. Pullulan standards with peak molecular weights ranging from 342 to were used to obtain the universal calibration to convert elution volume to hydrodynamic volume (V h ) or corresponding hydrodynamic radius (R h ), where V h = 4/3π R 3 h. The size distributions of whole and debranched starch molecules are presented as the SEC weight distributions, w(logv h ), plotted against R h, and normalized to either the yield of digesta or the highest peak. The DP of debranched starch molecules was also included in the plot as the secondary X-axis. Although there is no unique relationship between hydrodynamic size and molecular weight for any branched polymer, V h or R h of linear polymers, such as those in debranched starch, can be converted to molecular weight or equivalent DP X using the Mark Houwink equation (Liu et al., 2010), as follows: V h = 2 K M1+α 5 N A with M = 162 (X 1) Page 8 of 39

10 Where N A is the Avogadro s constant, K and α are the Mark Houwink parameters ( ml/g and for linear starch, respectively), 162 is the molecular weight of anhydroglucose monomer in starch, and 180 is the molecular weight of glucose at the reducing end. The DP obtained using the Mark Houwink equation is, however, prone to error for small molecules and thus is only semi-quantitative. The amylose content was analyzed from the SEC weight distribution of debranched starch molecules from the digesta collected at 0 h in vitro digestion. It was calculated as the ratio of the area under the curve (AUC) of amylose branches to the AUC of the whole distribution (both amylopectin and amylose branches) following the method described elsewhere (Vilaplana et al., 2012). X-ray diffractometry XRD was performed using an XPert Pro multi-purpose diffractometer (PANanalytical, Almelo, The Netherlands). The instrument used a copper long fine focus tube with an incident beam divergence slit, a diffracted beam scatter slit fixed at 0.125º, and an X celerator high speed detector. The starch digesta were observed between the angular range of 4º and 35º with a step size of º and a count time of 220 s per point. Approximately 2 g of each starch digesta, which was obtained by combining several replicates from the in vitro starch digestion, was used to give a sample depth of 4 mm. The data were analyzed using PeakFit v (Systat, San Jose, CA, USA), and the percentages of A-, B- and V-type crystallinity were calculated following the method described elsewhere (Li et al., 2015a; Lopez-Rubio et al., a). Differential scanning calorimetry The thermal properties of native starch and starch digesta were analyzed in duplicate using a DSC (DSC 1, Mettler Toledo, Schwerzenbach, Switzerland). The 9 Page 9 of 39

11 thermal properties of native starch granules after being annealed at 37 C for 16 h (waxy and normal maize starches) or 48 h (high-amylose maize and normal potato starches) without digestive enzymes were also determined in duplicate to differentiate between the effects of incubation at elevated temperature and those of enzyme digestion on the thermal properties of starch granules. Precisely weighed starch (approximately 4 6 mg, dry weight basis) with excess water (three times the weight of starch) was placed in a high-pressure crucible pan with gold-plated copper sealing disk, and allowed to equilibrate for at least 1 h at room temperature. The heating profile consisted of holding the sample at 10 C for 1 min, heating from 10 to 180 C at a rate of 10 C/min, rapid cooling from 180 to 10 C at a rate of 40 C/min, and reheating from 10 to 180 C at a rate of 10 C/min. The reheating was used to determine the endotherm associated with the melting of amylose-lipid complexes, which reformed spontaneously during rapid cooling (Li et al., 2008). Indium was used for calibration, and an empty crucible pan with gold-plated copper sealing disk was used as the reference. Onset temperature (T o ), peak temperature (T p ), conclusion temperature (T c ), and enthalpy change ( H) were determined from each endotherm using the built-in software (STARe system, Mettler Toledo). Due to the overlapping of starch gelatinization and the melting of amylose-lipid complexes in the high-amylose maize starch granules, the H of starch gelatinization from the first heating was corrected by subtracting the H of the melting of amylose-lipid complexes obtained from the reheating step. Furthermore, the high-pressure crucible pans used in this study were made of stainless steel, which had poorer heat conductivity compared with those made of aluminum. Due to this limitation, it is difficult to obtain precise information from broad small endotherms, such as the melting of amylose-lipid complexes. Hence, only the endotherm from starch 10 Page 10 of 39

12 gelatinization will be discussed quantitatively, whereas other smaller, broader peaks will be discussed qualitatively Scanning electron microscopy Native starches and starch digesta were dried in a freeze dryer (BenchTop 2K, VirTis, Gardiner, NY, USA) overnight, and then placed onto specimen stubs with double-sided carbon tape. Each starch sample was coated three times with gold to a thickness of 20 nm using a sputter coater (SPI-Module, SPI Supplies, West Chester, PA, USA). The direction of specimen was rotated to obtain an even layer on the surface. The granule morphology of starch samples was examined using an SEM (XL30, Philips, Eindoven, The Netherlands) with an accelerating voltage of 3 kv and a spot size of 6 nm. Statistical analysis The mean values of the thermal properties of starch were analyzed with Minitab 16 (Minitab Inc., State College, PA, USA) using the General Linear Model and Tukey s Pairwise Comparisons with a confidence level of 95.0% in performing an analysis of variance (ANOVA) test. Results In vitro digestion of native starch granules The digestograms of native waxy maize, normal maize, high-amylose maize, and normal potato starch granules are shown in Figure 1A, B, C, and D, respectively. The digestion of waxy and normal maize starch granules reached completion with ~100% of starch hydrolyzed at 8 and 24 h, respectively. Waxy and normal maize starch granules also displayed rapid digestion from 0 to 4 h followed by slower digestion after 4 h. On the other hand, the digestion of high-amylose maize and normal potato 11 Page 11 of 39

13 starch granules only reached 82% and 71% hydrolysis, respectively, after 48 h with neither digestogram showing a plateau Starch hydrolysis (%) Length of in vitro starch digestion (h) Figure 1. Digestograms of (A) waxy maize, (B) normal maize, (C) high-amylose maize, and (D) normal potato starch granules Molecular size distributions of whole (fully branched) starch The SEC weight distributions of whole (fully branched) starch molecules (Level 2 structure) from the digesta of waxy maize, normal maize, high-amylose maize, and 12 Page 12 of 39

14 normal potato starch granules at different in vitro digestion times are shown in Figure 2A, B, C, and D, respectively. The distributions were normalized to the yield of digesta as calculated from the degree of hydrolysis (yield of digesta = 100% % starch hydrolysis, Figure 1). Four distinct peaks could be observed in the SEC weight distributions of the starch samples, which were R h 1 4 nm, 3 10 nm, nm, and nm. Native whole (fully branched) amylopectin molecules have R h larger than 100 nm and native whole (fully branched) amylose molecules have R h smaller than 100 nm (Hasjim et al., 2010a; Vilaplana & Gilbert, 2010). During the in vitro digestion of normal maize, high-amylose maize, and normal potato starch granules, the ratios of amylopectin peak to amylose peak decreased with digestion time. Furthermore, the amount of molecules with the size of amylose, R h nm, became substantial in waxy maize starch digesta immediately after the addition of digestive enzymes although waxy maize starch is essentially devoid of amylose. This indicated that partially digested amylopectin molecules had a similar hydrodynamic size to amylose molecules, and hence co-eluted during SEC analysis, consistent with a previous study (Hasjim et al., 2010a). The molecules with R h 1 4 nm (peak maximum at R h 2 3 nm) that appeared after each starch was substantially hydrolyzed (at 0, 4, 16, and 24 h in vitro digestion of waxy maize, normal maize, high-amylose maize, and normal potato starch granules, respectively) are similar to the enzyme-resistant dextrins with peak R h ~2.5 nm observed after prolonged in vitro digestion of extruded starches (Witt et al., 2010) and prolonged in vitro and in vivo digestions of native normal maize starch granules (Hasjim et al., 2010a). The presence of small molecules (with peak maxima at 3 and 40 nm) in waxy maize starch granules immediately after the addition of the digestive enzymes (at 0 h) was likely due to the rapid digestion of waxy maize starch granules compared with other starches, 13 Page 13 of 39

15 especially at high concentrations of digestive enzymes. The additional peak with R h 4 10 nm (peak maximum at R h 5 7 nm) observed only in the waxy and normal maize starch digesta after prolonged in vitro digestion (Figure 2E and F, respectively, normalized to the highest peak) is associated with the digestive enzymes (Figure S1 in Supplementary Data), which became more concentrated as the starch substrate was greatly hydrolyzed. Most of the native molecules in the waxy and normal maize starch digesta disappeared after 16 and 24 h in vitro digestion (Figure 2E and F), respectively, whereas some of the native molecules in the high-amylose maize and normal potato starch digesta were still visible after 48 h in vitro digestion (Figure 2C and D, respectively). The results confirm that the native high-amylose maize and normal potato starch granules are more resistant to enzyme hydrolysis than the native waxy and normal maize starch granules, which is consistent with the results reported by others (Dhital et al., 2010; Jane et al., 2003; Tester et al., 2006). Furthermore, there was a higher amount of large branched molecules (R h > 100 nm, undigested native amylopectin) in the digesta of the normal potato starch than those of the high-amylose starch, but the latter produced more dextrins with R h 1 4 nm after in vitro digestion, suggesting different mechanisms of enzyme resistance in these two starches. 14 Page 14 of 39

16 w log (V h ) [arb. Unit) R h (nm) R h (nm) Figure 2. SEC weight distributions of whole (fully branched) starch molecules from (A and E) waxy maize, (B and F) normal maize, (C) high-amylose maize, and (D) normal potato starches after in vitro digestion at different times. A, B, C, and D were normalized to the yield of digesta (calculated from percentage of starch hydrolysis in Figure 1), whereas E and F were normalized to the highest peak because of their low digesta yield. Chain length distributions of starch molecules 15 Page 15 of 39

17 The SEC weight distributions of debranched starch molecules (chain length distribution or CLD, Level 1 structure) from the digesta of waxy maize, normal maize, high-amylose maize, and normal potato starch granules after different in vitro digestion times are shown in Figure 3A, B, C, and D, respectively, normalized to the highest peak. The branches of native starch molecules obtained from complete enzyme debranching can be divided into short amylopectin branches, long amylopectin branches, and amylose branches (DP X 1 15, , and , respectively) (Hasjim et al., 2010a; Vilaplana & Gilbert, 2010; Witt et al., 2010). The normalization to the highest peak shows the amount of each branch population in relative to that of short amylopectin branches, which have the highest peak. The amylose contents of native waxy maize, normal maize, high-amylose maize, and normal potato starch granules were 2, 27, 32, and 26%, respectively. The CLDs of normal maize starch did not show apparent changes up to 8 h in vitro digestion. Similarly, the CLDs of normal potato starch only showed slight changes from 16 h to 48 h in vitro digestion, suggesting that amylopectin and amylose molecules were digested at similar rates for these two starches. Molecules with R h 4 20 nm (peak maximum at R h ~5 nm) became apparent in the CLDs of waxy and normal maize starches after 8 and 16 h in vitro digestion, respectively, which are associated with the digestive enzymes (Figure S1 in Supplementary Data). The proportions of long amylopectin branches (R h nm or DP X with peak maximum at R h ~2 nm or DP X ~30) increased after 16, 24, and 16 h in vitro digestion of waxy maize, normal maize, and normal potato starch granules (Figure 3A, B, and D), respectively. The opposite trend was, however, observed from the CLDs of high-amylose maize starch (Figure 3C). 16 Page 16 of 39

18 DP X w log (V h ) [arb. Unit] R h (nm) DP X R h (nm) DP X R h (nm) DP X R h (nm) Figure 3. Chain length distribution obtained by SEC analysis of debranched starch molecules from (A) waxy maize, (B) normal maize, (C) high-amylose maize, and (D) normal potato starches after in vitro digestion at different times, normalized to the 346 highest peak. 17 Page 17 of 39

19 Table 1. Polymorphism and degree of crystallinity of starch granules after in vitro digestion at different times. Treatment Polymorphi sm A/B-type crystallinity (%) V-type crystallinity (%) Waxy maize starch Digested for 0h A Digested for 2h A Digested for 4h A Digested for 8h A Normal maize starch Digested for 0h A and V Digested for 2h A and V Digested for 4h A and V Digested for 8h A and V Digested for 16h A and V High-amylose starch maize Digested for 0h B and V Digested for 16h B and V Digested for 24h B and V Digested for 48h B and V Normal potato starch Digested for 0h B and V Page 18 of 39

20 Digested for 16h B and V Digested for 24h B and V Crystalline structure of starch The waxy and normal maize starch granules displayed the A-type crystallinity, while the high-amylose maize and normal potato starch granules displayed the B-type crystallinity (Table 1). All starches containing amylose also displayed the V-type crystallinity (> 1%). The degree of the A- or B-type crystallinity of all starches decreased after the in vitro digestion. The normal and high-amylose maize starches exhibited similar behaviors, where the changes in the degree of crystallinity were less obvious at the midpoints of the in vitro digestion. The proportion of the V-type crystallinity in the normal maize starch was largely unchanged, while those in the high-amylose maize and normal potato starches decreased after the in vitro digestion. Thermal properties of starch In general, there were three endotherms in the DSC thermogram from the first heating and one from the reheating. Starch gelatinization normally occurs below 100 C for waxy and normal starches and slightly above 100 C for high-amylose maize starch (Jane et al., 1999), and the melting of retrograded amylose (or long linear branches) occurs above 120 C (Hasjim & Jane, 2009; Sievert & Pomeranz, 1990). The melting of amylose-lipid complexes occurs at C (form I) and above 110 C after annealing (form II) (Biliaderis & Galloway, 1989; Hasjim & Jane, 2009), 368 while only the endotherm of form I appears during reheating (Li et al., 2008). 19 Page 19 of 39

21 369 Table 2. Thermal properties of starch granules after in vitro digestion at different times and after annealing without digestion. a Treatment Starch gelatinization ( ºC) Melting of retrograded dextrins (> 125 ºC) b T o (ºC) T p (ºC) T c (ºC) H (J/g) T o (ºC) T p (ºC) T c (ºC) H (J/g) Waxy maize starch Native 67.7 ± 1.4 d 76.4 ± 1.8 d 86.5 ± 2.5 c ± 1.17 a ND c ND ND ND Digested 0h 74.3 ± 0.3 bc 80.6 ± 0.6 b-d 91.5 ± 0.6 a-c ± 0.06 ab ± ± ± ± 0.81 Digested 2h 78.2 ± 0.9 ab 83.8 ± 1.4 a-c 93.3 ± 1.9 a-c ± 0.93 ab ± ± ± ± 0.48 Digested 4h 78.9 ± 0.2 ab 84.2 ± 0.3 ab 92.8 ± 0.5 a-c ± 0.77 a ± ± ± ± 0.98 Digested 8h 79.8 ± 0.9 a 85.7 ± 1.5 a 93.9 ± 2.4 ab 9.21 ± 0.72 bc ± ± ± ± 0.03 Digested 16h 74.8 ± 2.7 a-c 86.2 ± 1.0 a 96.3 ± 1.6 a 5.80 ± 0.32 c ± ± ± ± 0.87 Annealed 16h 70.7 ± 0.2 cd 79.5 ± 0.7 cd 89.0 ± 1.4 bc ± 1.89 a ND ND ND ND Normal maize starch Native 67.3 ± 0.7 d 73.6 ± 1.1 d 82.0 ± 1.9 c ± 2.96 a ND ND ND ND Accepted Manuscrip Digested 0h 70.7 ± 0.0 c 76.2 ± 0.5 cd 83.9 ± 0.5 bc ± 1.63 ab ± ± ± ± 0.04 Digested 2h 73.4 ± 0.7 b 78.4 ± 0.7 bc 85.3 ± 1.1 bc 7.90 ± 0.66 ab ± ± ± ± 0.06 Digested 4h 74.5 ± 0.1 ab 79.4 ± 0.1 b 86.7 ± 0.5 ab 7.10 ± 0.80 ab ± ± ± ± 0.57 Digested 8h 74.4 ± 0.0 b 79.1 ± 0.3 b 85.9 ± 0.5 b 7.65 ± 0.24 ab ± ± ± ± Page 20 of 39

22 Digested 16h 76.7 ± 1.1 a 82.1 ± 1.1 a 89.8 ± 0.9 a 5.03 ± 1.77 b ± ± ± ± 0.08 Annealed 16h 70.7 ± 0.0 c 77.2 ± 0.3 bc 85.2 ± 0.0 bc ± 0.04 a ND ND ND ND a Means ± standard deviations. T o = onset temperature, T p = peak temperature, T c = conclusion temperature, and H = enthalpy change during transition. Different letters in the same column represent significant difference at p < 0.05 among the samples from the same starch source. b Due to the limitation of the DSC, the melting of retrograded dextrins was only regarded as qualitative and the differences were not statistically analyzed. c Not detected. d The H of starch gelatinization was corrected by subtracting the H of the melting of amylose-lipid complexes from the rescan. Accepted Manuscrip 21 Page 21 of 39

23 376 Table 2. (Continued) Treatment Starch gelatinization Melting of retrograded dextrins T o (ºC) T p (ºC) T c (ºC) H (J/g) T o (ºC) T p (ºC) T c (ºC) H (J/g) High-amylose maize starch d Native 71.9 ± 0.5 b bc 81.0 ± 1.4 Digested 0h 75.8 ± ± 0.6 a ab Digested 16h 76.5 ± ± 0.2 a a Digested 24h 76.5 ± ± 0.6 Digested 48h 76.2 ± 0.9 Annealed 48h 69.1 ± 1.4 a 4.8 a ± Accepted Manuscrip a ± ± ± ± ± 0.04 a ± ± ± ± ± 0.12 ± a a a ± ± ± ± ± ± 0.09 ab ± ± ± ± ± ± 0.21 ab 0.7 a ab ± ± 7.67 ± 1.94 b ± ± ± 0.29 ± 0.40 ab 8.6 a ± ± ± ± ± ± ± 0.12 b c 1.8 a ab Normal potato starch 22 Page 22 of 39

24 377 Native 65.4 ± ± 0.1 ab a Digested 0h 66.5 ± 0.3 a a Digested 16h 66.8 ± 0.4 a a 71.9 ± ± a 0.3 a 0.9 a 80.6 ± 80.5 ± 82.2 ± ± 0.02 Accepted Manuscrip a ab ab ± ± 0.69 Digested 24h 66.7 ± ± ± ± 0.35 a a Digested 48h 66.5 ± 1.3 a Annealed 48h 63.4 ± 1.2 b a a 72.2 ± ± a 1.7 a 3.0 a 80.9 ± 78.0 ± ab b ab ± ± 1.81 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 23 Page 23 of 39

25 The endotherm associated with the melting of amylose-lipid complexes was not observed from waxy maize starch granules because of the absence of amylose (Figure A). In addition, it was not observed from normal potato starch granules because only very small amount of lipids is present in potato starch granules although the V-type crystallinity was visible from XRD diffractograms (Table 1), probably due the higher sensitivity of XRD technique or the presence of non-complexing single helices (Dhital et al., 2011; McPherson & Jane, 1999). The melting temperatures of amylose-lipid complexes of normal and high-amylose maize starch granules did not seem to be altered by the in vitro digestion (Table S1 in Supplementary Data), and hence are not further discussed. However, as expected, annealing of normal maize starch granules increased the melting temperature of amylose-lipid complexes, which was attributed to the conversion of form I complex to form II complex. The gelatinization temperatures of the waxy and normal maize starch granules, in general, increased with the in vitro digestion time (Table 2). The increase in the gelatinization temperature was not observed as strongly in the high-amylose maize starch granules, with only the native starch displaying a lower T o, whereas the normal potato starch granules did not show significant changes in the gelatinization temperature after the in vitro digestion. The gelatinization temperatures of waxy and normal maize starch granules after 16 h in vitro digestion as well as the T o of high-amylose maize and normal potato starch granules after 48 h in vitro digestion were higher than their counterparts after being annealed at the same temperature and time without the addition of the digestive enzymes. The H of gelatinization of all starches did not significantly change at the beginning of the in vitro digestion and then significantly decreased after prolonged in vitro digestion. Furthermore, the H of gelatinization of waxy and normal maize starch granules after 16 h in vitro digestion 24 Page 24 of 39

26 were significantly lower than their counterparts annealed at 37 C without the addition of digestive enzymes. The results indicated that the changes in the gelatinization properties of starch granules, which related to the crystalline structure, after the in vitro digestion were not solely due to the effects of annealing. The starch crystallites with a high melting temperature (T o > 120 C), normally assigned as retrograded amylose (Hasjim & Jane, 2009; Sievert & Pomeranz, 1990), were detected in the digesta of waxy maize starch granules, indicating that these crystallites were not solely originated from amylose chains. Although the peak was small, it appeared consistently in all digesta of waxy maize starch granules at similar temperature range and was not detected in the native and annealed waxy maize starch granules. Normal maize starch granules also displayed a high-temperature endotherm after the in vitro digestion in similar temperature range to that of the digesta from waxy maize starch granules, with neither the native nor annealed counterparts displaying this endotherm. High-amylose maize starch granules displayed similar high-temperature endotherm in all samples including the native and annealed starches, indicating that the highly ordered structure formed during the in vitro digestion of waxy and normal maize starch granules overlaps with the amylose crystallites present in the native high-amylose maize starch granules. The normal potato starch granules did not show the high-temperature endotherm before and after the in vitro digestion. There was no clear trend between the H from the melting of the high-temperature endotherm and the in vitro starch digestion time; the inability to observe any trends in the H may be due to the small amount of the retrograded amylose/linear chains present in ungelatinized starch granules, leading to the difficulties in accurately measuring the H. 426 Granule morphology of starch 25 Page 25 of 39

27 The SEM structure of native starch granules from waxy maize, normal maize, high-amylose maize, and normal potato are shown in Figure 4A, B, C, and D, respectively. The granules of waxy and normal maize starch digesta displayed large pores on granule surface after the in vitro digestion (Figure 4E and F, respectively). The pores on the surface of high-amylose maize and normal potato starch granules after 24 h in vitro digestion were less apparent (Figure 4G and H, respectively), but the granule surface was evidently rougher compared with that of the native starch granule counterparts, which was more obvious for high-amylose maize starch digesta than normal potato starch digesta. It is well documented that the surface of the B-type starch granules (such as high-amylose maize and normal potato starches) is more resistant to enzyme hydrolysis than that of the A-type starch granules (such as waxy and normal maize starches), preventing the enzymes to hydrolyze the inner part of the granules (Dhital et al, 2010). The rough granule surface of the B-type starch digesta was the result of enzyme hydrolysis as the enzymes were not able to penetrate into the starch granules. Furthermore, the granules of high-amylose maize starch digesta were evenly hydrolyzed and agglomerated, whereas those of normal potato starch digesta were not evenly hydrolyzed, showing the presence of both intact individual granules and severely digested few. 26 Page 26 of 39

28 A E B C D F G H Figure 4. SEM images of native (A) waxy maize, (B) normal maize, (C) high-amylose maize, and (D) normal potato starch granules as well as (E) waxy and (F) normal maize starch granules after 4 hour in vitro digestion and (G) high-amylose maize and (H) normal potato starch granules after 24 hour in vitro digestion. Discussion 27 Page 27 of 39

29 The molecular, crystalline, and granular structures of digesta from waxy maize, normal maize, high-amylose maize, and normal potato starches after an in vitro digestion were monitored using SEC, XRD, DSC, and SEM. Native starch granules were used instead of cooked or gelatinized starch so that rearrangement of starch molecules during the in vitro digestion could be easily observed. Furthermore, the pre-existing native semi-crystalline structure is an obstacle for starch to undergo molecular rearrangement during the in vitro digestion and thus, if such was observed, it is very likely that molecular rearrangement happens during the digestion of cooked or gelatinized starch. The SEC weight distributions of branched molecules (Figure 2) and CLDs (Figure 3) showed that both amylose and amylopectin molecules were hydrolyzed by digestive enzymes. The CLDs of normal maize and normal potato starches revealed the similar ratios of amylose branches to amylopectin branches during the in vitro digestion although the amylopectin peaks of their SEC weight distributions of branched molecules reduced at faster rates than the amylose peaks, reinforcing the idea that some partially digested amylopectin molecules co-eluted with amylose molecules. The similar ratios of amylose branches to amylopectin branches also indicated the similar digestion rates between amylose and amylopectin for normal maize and normal potato starches. Similar observations were reported for normal maize starch granules in a previous study (Hasjim et al., 2010a). The degrees of crystallinity of all four starches were slightly reduced or not significantly changed at the beginning of the in vitro digestion as revealed by XRD and DSC (Tables 1 and 2, respectively). The results are consistent with those reported by Zhang et al. (2006) for normal maize starch granules, indicating that the crystalline structures of the A- and B-type starch granules are not more resistant to enzyme 28 Page 28 of 39

30 hydrolysis than their amorphous structures. The significantly lower H of gelatinization after prolonged in vitro digestion was likely due to the accumulation of non-starch components, such as the digestive enzymes as shown by the SEC weight distributions of whole starch molecules (Figure 2E and F), and these digesta appeared dark brown in color instead of white as observed from the native starches and other digesta. Thus, neither the amylose molecules nor the native crystalline structure is the dominant factor governing the digestibility of native starch granules. Linear dextrins with peak R h ~2.5 nm or DP X ~50 were observed after prolonged in vitro digestion of extruded waxy, normal, and high-amylose maize starches (Witt et al., 2010) and prolonged in vitro and in vivo digestions of native normal maize starch granules (Hasjim et al., 2010a). Similarly, small molecules with R h 1 4 nm were observed from both SEC weight distributions of branched molecules (Figure 2) and CLDs (Figure 3) from all four starches after prolonged in vitro digestion times in the present study although the latter have been enzymatically debranched, indicating that these small molecules/dextrins were mainly linear. The presence of a high-temperature endotherm (T o > 120 C) in all digesta from waxy and normal maize starch granules, which was not observed in the native and annealed starch granule counterparts, suggested that the linear dextrins produced during the in vitro digestion of starch granules could rearrange into thermally stable crystalline structure. This crystalline structure is similar to that of retrograded amylose, which is thermally stable and highly resistant to enzyme hydrolysis (Gidley et al., 1995; Sievert & Pomeranz, 1990). It was also reported that enzyme-resistant retrograded amylose had DP X ~50 after its amorphous fraction was removed by α-amylase (Jane & Robyt, 1984), similar to the size of linear dextrins after prolonged in vitro digestions of waxy 500 and normal maize starch granules. 29 Page 29 of 39

31 Different from the waxy and normal maize starch granules, the high-amylose maize starch granules displayed the high-temperature endotherm in the native state and after the in vitro digestion. Amylose double helices were reported to be present in the native high-amylose maize starch granules, responsible for their high RS contents (Jiang, Srichuwong, Campbell, Jane, 2010). Furthermore, the ratio of dextrins with R h 1 4 nm to the native molecules (R h > 4 nm) in the SEC weight distributions of branched molecules for the high-amylose maize starch was evidently increased with the in vitro digestion time (Figure 2C), indicating that the native molecules were hydrolyzed to form enzyme-resistant linear dextrins with R h 1 4 nm. The increase of the linear dextrins with R h 1 4 nm was not observed from the CLDs of the high-amylose maize starch (Figure 3C), which could be due to the rapid hydrolysis of long amylopectin branches to similar size of the short amylopectin branches. Therefore, the rearrangement of linear dextrins with R h 1 4 nm produced during the in vitro digestion into enzyme-resistant crystalline structure is mainly responsible for the low digestibility of native high-amylose maize starch granules. Despite having higher enzyme resistance than high-amylose maize starch granules (Figure 1C and D), normal potato starch granules did not display the high-temperature endotherm before and after the in vitro digestion (Table 2). The digesta of normal potato starch also contained a smaller amount of dextrins with R h 1 4 nm, but a higher amount of large branched molecules (R h > 100 nm, undigested native amylopectin) than those of high-amylose maize starch (Figure 2D and C, respectively) Furthermore, the CLDs showed that the ratio of amylose branches to amylopectin branches of normal potato starch granules did not change much during the in vitro digestion (Figure 3D), whereas that of high-amylose maize starch granules changed with short amylopectin branches dominating after 48 h in vitro digestion (Figure 3C). 30 Page 30 of 39

32 The high-amylose maize starch granules had rougher surface structure and were more agglomerated than the normal potato starch granules after 24 h in vitro digestion (Figure 4G and H, respectively), suggesting that the normal potato starch granules were more resistant to enzyme hydrolysis. Thus, the presence of large amylopectin in normal potato starch digesta could be attributed to the granular structure protecting the molecules from being hydrolyzed by enzymes, reducing the amount of linear dextrins with R h 1 4 nm in the digesta. Differences in the susceptibility of granules from different starches to enzyme hydrolysis are well established, with potato starch granules particularly resistant as demonstrated by the lack of surface pores and internal void structure (Dhital et al., 2010; Jane et al., 2003; Tester et al., 2006). The presence of intact granular structure in the normal potato starch digesta could also create a restriction for the linear dextrins to form thermally stable crystalline structure. This is also the main reason for the difference between the digestion profiles of normal maize and normal potato starch granules although they have similar amylose contents (Figure 1B and D, respectively). In general, the gelatinization temperatures of waxy, normal, and high-amylose maize starch granules increased during the in vitro digestion although the extents depended on the types of starch (Table 2). The increase could be due to the more rapid hydrolysis of less perfectly aligned crystallites compared with more perfectly aligned crystallites, which the latter had higher thermal stability. The hydrolysis could also reduce the steric restriction for the molecules to rearrange into more perfectly aligned crystallites. The gelatinization temperature of normal potato starch granules, however, remained similar during the in vitro digestion as the highly enzyme-resistant granular structure of normal potato starch protected the less perfectly aligned crystallites from 550 being hydrolyzed by the enzymes. 31 Page 31 of 39

33 Conclusions All starches showed that amylose and amylopectin molecules in native starch granules were hydrolyzed to smaller dextrins during the in vitro digestion, and some small dextrins had linear structure. The degrees of crystallinity of all starches were slightly reduced or not significantly changed after digestion. However, the gelatinization temperatures of the waxy, normal, and high-amylose maize starches increased after the in vitro digestion, indicating that the heat labile crystallites were more susceptible to enzyme hydrolysis and/or the hydrolyzed starch molecules could rearrange into more perfectly aligned crystalline structure. A high-temperature endotherm appeared in waxy and normal maize starches after the in vitro digestion as well as in high-amylose maize starch before and after the in vitro digestion, which was similar to that of retrograded amylose, suggesting that the linear dextrins could form highly ordered crystalline structure with slow digestion properties and were mainly responsible for the high enzyme resistance of native high-amylose maize starch granules. On the other hand, normal potato starch granules did not show the high-temperature endotherm before and after digestion although they had higher enzyme resistance than high-amylose maize starch granules. The presence of intact individual granules lacking of surface pores in the normal potato starch digesta indicated that the surface structure of normal potato starch granules protected the molecules inside the granules from being hydrolyzed by the enzymes Acknowledgements The authors thank the Australian Research Council (DP and LP ) for funding. 32 Page 32 of 39