Biomacromolecules 2010, 11, 3275–3289
3275
Effect of Enzymatic Hydrolysis on Native Starch Granule Structure Jaroslav Blazek and Elliot Paul Gilbert* Bragg Institute, Australian Nuclear Science and Technology Organization, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia Received June 16, 2010; Revised Manuscript Received September 29, 2010
Enzymatic digestion of six starches of different botanical origin was studied in real time by in situ time-resolved small-angle neutron scattering (SANS) and complemented by the analysis of native and digested material by X-ray diffraction, differential scanning calorimetry, small-angle X-ray scattering, and scanning electron microscopy with the aim of following changes in starch granule nanostructure during enzymatic digestion. This range of techniques enables coverage over five orders of length-scale, as is necessary for this hierarchically structured material. Starches studied varied in their digestibility and displayed structural differences in the course of enzymatic digestion. The use of time-resolved SANS showed that solvent-drying of digested residues does not induce any structural artifacts on the length scale followed by small-angle scattering. In the course of digestion, the lamellar peak intensity gradually decreased and low-q scattering increased. These trends were more substantial for A-type than for B-type starches. These observations were explained by preferential digestion of the amorphous growth rings. Hydrolysis of the semicrystalline growth rings was explained on the basis of a liquid-crystalline model for starch considering differences between A-type and B-type starches in the length and rigidity of amylopectin spacers and branches. As evidenced by differing morphologies of enzymatic attack among varieties, the existence of granular pores and channels and physical penetrability of the amorphous growth ring affect the accessibility of the enzyme to the substrate. The combined effects of the granule microstructure and the nanostructure of the growth rings influence the opportunity of the enzyme to access its substrate; as a consequence, these structures determine the enzymatic digestibility of granular starches more than the absolute physical densities of the amorphous growth rings and amorphous and crystalline regions of the semicrystalline growth rings.
Introduction Starch is not only the most important carbohydrate in the human diet, but is also a key ingredient in animal feed and has many industrial applications outside the food industry, such as in paper manufacture, packaging, and the production of biofuels. In the human diet, starch is generally consumed after processing. An excess of water and high temperature during processing results in starch gelatinization and destroys its granular structure. However, in several low moisture food products such as biscuits and muesli or in fruit and vegetables, the granular structure of starch can be retained. Unlike for human consumption, animal feed mostly contains native unprocessed starch. Reducing the energy requirements in the conversion of granular starch to glucose via improved molecular disassembly and depolymerization is relevant to the overall efficiency of bioconversion of starch to chemicals, ingredients, and fuels.1 Native starch is digested (i.e., hydrolyzed) slowly compared with processed (gelatinized) starch whose crystallinity has been largely destroyed and where the accessibility of substrate to enzymes is greater and not restricted by R-glucan associations such as double helices (especially in crystallites) or amylose-lipid complexes (in cereal starches). Therefore, understanding the factors that control the extent of R-amylolysis of native starches is essential to ensure fitness-for-purpose for such a diverse range of end uses. Bacteria, fungi, plants, animals, and humans produce R-amylase where the enzyme hydrolyses R-(1-4) bonds in a random, * To whom correspondence should be addressed. Tel.: (61)-2-9717-9470. Fax: (61)-2-9717-3606. E-mail:
[email protected]. 10.1021/bm101124t
endoacting fashion, thereby reducing the molecular weight of starch molecules (amylose and amylopectin). Previous studies have shown that the action of R-amylase on starches of different botanical origin results in varied digestion kinetics and degradation patterns.2-4 The features of native starch granules that control the site, rate, and extent of hydrolysis by R-amylase are interrelated and not easily definable. While the size and shape of granules is clearly a controlling factor, access of the enzyme to its substrate and the release of reaction products is influenced by many more factors including granule integrity, crystallinity, and porosity of granules, amylose to amylopectin ratio, structural inhomogeneities, phosphate content, proteins, and lipids on the surface of starch granules and hydrolysis products such as maltooligosaccharides that have been shown to inhibit amylolysis.1,5-9 The rate and degree of digestion is less influenced by the quantity of double helices or total crystallinity of the starch but, far more so, by the extent to which R-glucans are associated to facilitate or hinder the diffusion of amylase and starch hydrolysis products through the granules.8,10 The digestion of starch has been the subject of many investigations, mostly involving in vitro measurement of the susceptibility of starches to attack by different enzymes, rather than measuring actual digestibility in vivo. Changes of the lamellar architecture and crystalline structures during digestion have been studied previously, but most in vitro studies of granular starch digestion have been limited to samples for which aliquots have been removed from the reaction mixture at various time intervals and solvent- or freeze-dried prior to structural analyses.11,12 Using simultaneous small and wide-angle X-ray scattering (SAXS and WAXS, respectively), Muhr et al.13 and
Published 2010 by the American Chemical Society Published on Web 10/29/2010
3276
Biomacromolecules, Vol. 11, No. 12, 2010
Blazek and Gilbert
Figure 1. Schematic description of the in situ digestion chamber.
Jenkins and Donald14 showed that the SAXS lamellar peak becomes less pronounced upon acid hydrolysis of native starch granules, whereas the WAXS peaks become more intense. Jenkins and Donald14 explained these observations by considering preferential destruction of the amorphous background (growth-ring) region during the hydrolysis process. Zhang et al.15 analyzed cereal starch residuals following different digestion times and proposed a mechanism of side-by-side enzymatic digestion of concentric layers of semicrystalline shells of native starch granules. In a subsequent study, Zhang et al.16 reported that the slow digestion property of native cereal starches is determined by the supramolecular A-type crystalline structure, including the distribution and perfection of crystalline regions (both crystalline and amorphous lamellae). The objective of this work was to follow the structural changes in native starch granules during enzymatic digestion and to examine the structural features of starch granules that control the susceptibility of starch to enzymatic hydrolysis. We specifically use the phrase ‘enzymatic hydrolysis’ to distinguish this in vitro process from physiological hydrolysis occurring in vivo. To achieve this objective, a broad range of techniques has been applied to interrogate the hierarchical structure of this system over five orders of magnitude in length-scale and for six commercial starches of different botanical origin. Furthermore, it remains unclear from existing studies whether sample preparation creates structural artifacts. To address the latter, we have studied the time-resolved structural changes with smallangle neutron scattering using an in situ digestion chamber allowing, for the first time, to follow the structural changes of starch in the course of digestion directly in the digestion mixture.
Experimental Section Materials. Six commercial starches were used: Eliane 100 waxy (amylose-free) potato starch and high-amylose maize starch Hylon VII from National Starch Pty Ltd. (Seven Hills, NSW, Australia) and regular wheat, tapioca, regular maize, and regular rice starches from Penford Australia Pty Ltd. (Lane Cove, NSW, Australia). Eliane amylopectin potato starch is GMO-free and contains more than 99% amylopectin (AVEBE, Netherlands). Hylon VII unmodified high amylose maize starch is reported to have an amylose content of approximately 70% (National Starch Pty Ltd.). The commercial wheat starch has 25% amylose and other properties as described by Tang and Copeland.17
Regular maize starch from Penford Australia has around 28% amylose, regular rice around 10%, and regular tapioca starch contains 17% amylose. In Vitro Starch Digestion. An in situ digestion chamber was designed as shown in Figure 1 composed of a peristaltic pump (variable speed C/L single channel peristaltic pump, MasterFlex, Cole-Palmer Instrument Company, Illinois, U.S.A.), a water bath with inbuilt magnetic stirrer (Daintree Scientific, St. Helens, Australia), Microbore pump tubing (Tygon Lab, 2.79 mm diameter, MasterFlex, Cole-Palmer Instrument Company, Illinois, U.S.A.), a digestion container, and an absorption QS quartz cell for flow-through measurement with a path length of 1 mm (Hellma GmbH & Co., Mu¨llheim, Germany). The digestion container consisted of a plastic container with diameter of approximately 50 mm, into which four rectangular holes were cut and subsequently lined with Spectra/Por standard dialysis sheet (regenerated cellulose, cutoff size 6-8 kDa; Spectrum Laboratories, Inc., Rancho Dominguez, California) held in place by rubber bands. Dialysis membrane was used to enable the digestion products (reducing sugars and glucose) to exit the reaction medium, whereas starch and enzymes were maintained inside the chamber. Previous experiments by the authors (unpublished results) showed that enzymatic activity is partly inhibited by the digestion products when the dialysis membrane was not used. The sample was maintained at a temperature of 37 °C throughout the course of the experiment. Granular starch (30 g, dry-basis) was digested in 70 mL of sodium acetate buffer (0.1 M, pH 6.0) with 1 g porcine pancreatic R-amylase (0.67 IU per 1 mg starch, A-3176, Sigma-Aldrich, St. Louis, MO, U.S.A.) and 50 µL of amyloglucosidase from Aspergillus niger (0.02 IU per 1 mg starch, A-3042, Sigma-Aldrich, St. Louis, MO, U.S.A.). Enzyme solutions were prepared fresh prior to each digestion. This solution was centrifuged to remove solids for 5 min at 5,000 rpm prior to use. The height of fluid in the surrounding water bath was the same as the height of fluid in the digestion chamber to ensure no water exchange between reservoirs. A lid was secured onto the water bath to prevent water evaporation from the chamber during digestion. After digestion, the digestion mixture was centrifuged for 8 min at 6000 rpm, and the resultant pellet was resuspended in 50 mL of 100% acetone. The mixture was then centrifuged for 8 min at 6000 rpm and the pellet was air-dried at room temperature in a fumehood. Native and digested starches were equilibrated in a desiccator containing silica gel for a week to ensure identical moisture content (approximately 10%) for X-ray diffraction and differential scanning calorimetry analyses. The extent of starch hydrolysis was determined by measuring the weight of undigested starch with respect to the starting weight. As a control,
Hydrolysis of the Native Starch Granule
Biomacromolecules, Vol. 11, No. 12, 2010
a wheat starch sample was prepared in the similar manner but in the absence of enzyme addition to the buffer. The kinetics of starch digestion were determined by measuring the release of glucose from starch during the course of digestion. Glucose concentration in the digestion chamber and in the surrounding solution was measured using the Accu-Check Performa Blood Glucose Meter based on the method described by Sopade and Gidley.18 The percentage of digested starch (%DS) was calculated from the measured glucose concentrations in mmol/L within the known volumes in the digestion chamber and in the surrounding solution in the water bath using the following equation:
%DS )
(
180 × GG2 × V2 100 × 0.9 180 × GG1 × V1 + ms 1000 1000
) (1)
where GG1 and GG2 are glucometer readings (mmol/L) in the digestion chamber and in the surrounding solution, respectively; V1 and V2 are volumes of the digestion chamber and the surrounding solution, respectively; ms is the weight of native starch (30 g, dry-basis) and 0.9 is a stoichiometric constant for starch from glucose contents. The rate of starch hydrolysis is principally controlled by porcine pancreatic R-amylase, while the function of amyloglucosidase in the assay is to convert amylase degradation products to glucose and to prevent inhibition of R-amylase activity; the direct hydrolysis of raw starch granules by amyloglucosidase is considerably limited. Small-Angle Neutron Scattering (SANS). SANS experiments were performed on the 40 m Quokka instrument at the OPAL reactor; this instrument has been described previously.19 A wavelength, λ, of 5.0 Å and 10% wavelength resolution was used with source and sample aperture diameters of 50 and 7.5 mm, respectively. Because the sample scattering continuously varied with digestion time, a single instrument configuration was selected that includes the full q range associated with the lamellar region of native starches, namely, with a source-to-sample distance (SSD) of 4.008 m and sample-to-detector distance (SDD) of 4.520 m providing a q range of 0.01-0.3 Å-1, where q is the magnitude of the scattering vector defined as
q)
4π sin θ λ
(2)
and where 2θ is the scattering angle. The scattering in this configuration was collected continuously in 5 min intervals throughout the 18 h of in vitro digestion. In addition, two further instrument configurations were used to access both lower q (SSD ) 16.050 m, SDD ) 20.145 m) and higher q (SSD ) 4.008 m, SDD ) 1.247 m) to provide continuous q coverage from 0.003 to 0.7 Å-1. The instrumental resolution, ∆q, at the main lamellar peak (located at 0.06-0.07 Å-1 and discussed below) is 0.003 Å-1. The final digested samples (i.e., after 18 h of digestion) were measured over all three configurations (total of 6 h), after which the reaction was stopped and remaining starch was solvent-dried for subsequent analyses (i.e., total digestion time of 24 h). Native starch samples were prepared separately (i.e., in the absence of enzyme addition to the buffer) and measured over all three configurations while being circulated through the in situ digestion chamber. Enzyme dissolved in sodium acetate buffer of the same concentration as used during the digestion was used as a background. Scattering from the control wheat starch sample was collected over 18 h in the same manner as for the digested samples. Small-Angle X-ray Scattering (SAXS). SAXS measurements were obtained on native and final digested samples (i.e., after 24 h of digestion) using a Bruker NanoStar SAXS instrument at a wavelength of 1.5418 Å and described previously.20 The SSD was 700 mm, which provided a q range from 0.014 to 0.430 Å-1. The instrumental resolution, ∆q, at the main lamellar peak was calculated to be 0.001 Å-1. In contrast to the SANS studies in which the starch was
3277
continuously circulated in the in situ digestion chamber, for SAXS, the raw and partly digested (after 24 h) samples were presented in 2 mm sealed quartz capillaries as suspensions containing excess water above the sedimented sample and the scattering was measured for 60 min. The two scattering techniques are complementary with SAXS providing greater intensity and instrumental resolution but at a higher minimum q (smaller maximum spatial dimension) than the SANS instrument used here. We note that an improved minimum q for X-rays would be achievable using a synchrotron-based system but with the caveat that considerable effort is required to minimize radiation damage to this sensitive biomolecule.21 Reduction and Fitting of the Scattering Data. SAXS data sets were radially averaged using Bruker AXS software 4.1.30. SAXS curves were normalized to sample transmission and background-subtracted using Igor software (Wavemetrics, Lake Oswego, Oregon, U.S.A.). SANS data sets were reduced, normalized, and radially averaged using a package of macros in Igor software originally written by Steve Kline22 and modified to accept HDF5 data files from Quokka. Scattering curves are plotted as a function of relative (SAXS) or absolute (SANS) intensity, I, versus q. SANS data collected over 5 min intervals throughout the course of enzymatic digestion, as well as the SANS and SAXS curves of native and final digested samples were fitted in Igor. Curve fitting was carried out iteratively; the fitting coefficients were refined for each iteration to minimize chi-squared via a nonlinear, least-squares fitting procedure using a Levenberg-Marquardt algorithm, as described in the Supporting Information. Differential Scanning Calorimetry (DSC). DSC measurements were made using a DSC 1 STARe System (Mettler Toledo). Starch (8 mg, dry-basis) and deionized water (30 mg) were weighed directly into a 120 µL stainless steel medium pressure crucible, and the crucible was hermetically sealed with a Viton O-ring and a stainless steel lid. A pan containing 30 mg of water was used as a reference. The pans were heated from 10 to 160 °C with a scan rate of 5 °C/min. The instrument was calibrated using indium as a standard. Melting temperatures were determined from the thermograms by means of the STARe software provided by the manufacturer. The calorimetric enthalpy (∆H) was determined by numerical integration of the area under the peak of the thermal transition above the extrapolated baseline. The average values of the thermodynamic parameters were determined using duplicate measurements and normalized per gram of starch. X-ray Diffraction (XRD). XRD measurements of samples were made with a Panalytical X’Pert Pro diffractometer. The instrument was equipped with a copper X-ray generator (λ ) 1.54 Å), programmable incident beam divergence slit and diffracted beam scatter slit, and an X’celerator high speed detector. X-ray diffraction patterns were acquired at room temperature over the 2θ range of 5-40° with a step size of 0.0330° and a count time of 400 s per step. Scanning Electron Microscopy (SEM). The dried starch powder was thinly spread onto circular metal stubs coated with double-sided adhesive carbon tape. The powder on the stubs was gold coated in an Eiko IB-5 Sputter Coater (at 6 mA, 5 min for medium coating) in an argon gas environment, yielding approximately 15 nm coating thickness. Images of the starch granules were acquired with a Philips XL30 scanning electron microscope under an accelerating voltage of 10 kV. Multiple micrographs of each sample were examined at multiple magnifications and typical representative images are shown.
Results Enzymatic Digestion of Starch by r-Amylase. The digestion of starch was monitored by measuring the release of glucose from starch by the collective action of R-amylase and amyloglucosidase at specific intervals during the reaction. The amount of starch hydrolyzed from the granules after 24 h of digestion ranged between 18 and 51% (Table 1). The most resistant starch granules were waxy potato (18%) and high-amylose maize (21%), whereas the other regular starches were more readily
3278
Biomacromolecules, Vol. 11, No. 12, 2010
Blazek and Gilbert
Table 1. Overview of the Starches Used in this Studya
AM content (%) typical granule size (µm) crystallinity type extent of digestion (%) DSC normalized δH (J/g) native onset T (°C) starches peak T (°C) DSC normalized δH (J/g) digested onset T (°C) starches peak T (°C)
regular wheat
regular rice
regular tapioca
regular maize
high-amylose maize
waxy potato
26 5-40 A 48 17.1 54.6 61.4 20.3 58.1 63.1
∼10 2-5 A 51 21.9 60.6 66.4 18.3 62.5 68.4
∼17 5-25 A 43 21.5 60.1 66.0 24.5 61.3 67.1
∼28 5-15 A 50 19.8 61.0 66.4 18.2 62.5 67.8
∼70 5-25 B 21 19.3 60.5 90.7 18.7 69.9 85.5
