Starch/Starke 2011, 63, 395405

DOI 10.1002/star.201000150

395

REVIEW

Human a-amylase and starch digestion: An interesting marriage

Peter J. Butterworth, Frederick J. Warren and Peter R. Ellis
Kings College London, School of Medicine, Diabetes and Nutritional Sciences Division, Biopolymers Group, London, UK

a-Amylase catalyses the rst step in the digestion of starch, a main source of carbohydrate in the human diet. Amylase present in human saliva was one of the rst enzymes ever to be recognised but many puzzles remain about the molecular mechanisms involved in amylolysis of starch and even of the physiological role of the salivary amylase itself. Native starch granules represent a formidable challenge for attack from an enzyme in solution. Moreover the frequently reported differences in the susceptibility to amylolysis of starches from various botanical species, plus the changes that occur in starch structure and properties during domestic and commercial food processing means that studies of the enzymology of starch digestion can be challenging. We review the molecular properties of mammalian a-amylase including its genetics, and speculate on the role of salivary amylase in digestion of dietary starch. Also considered are enzyme kinetic approaches that have been used in vitro studies of amylase digestion of starches. Such studies can result in better understanding of reasons for the differences in glycaemic responses of humans following ingestion of different starch containing foods. Keywords: a-Amylase / Genetics / Kinetics / Salivary / Starch digestion

Received: December 6, 2010 Accepted: January 5, 2011

Introduction

In 1831, Erhard Leuchs reported that starch is broken down when mixed with human saliva and used the named ptyalin to describe the agent in saliva that was responsible for the chemical reaction. This seems to have been one of the earliest accounts of an enzyme experiment and soon after, in 1833, Payen and Persoz isolated an enzyme from barley that broke down starch and named it diastase. Thus enzymes now named amylases have a long history and most readers will have memories from school biology

Correspondence: Dr. Peter J. Butterworth, Kings College London, School of Medicine, Diabetes and Nutritional Sciences Division, Biopolymers Group, Franklin Wilkins Building, 150 Stamford Street, London SE1 9NH. E-mail: peter.butterworth@kcl.ac.uk Fax: 44-207-848-4170 Abbreviations: AM, amylose; AP, amylopectin; CE, catalytic efficiency; RDS, rapidly digested starch; RS, resistant starch.

classes of spitting into test tubes containing starch suspensions stained dark blue with iodine and observing the disappearance of the colour as the starch is hydrolysed in the presence of salivary amylase. Despite this long history and apparent familiarity with amylase properties, there are still many aspects of amylase and starch biochemistry that continue to puzzle investigators including the full physiological signicance of that familiar amylase in saliva. Starch is normally the main source of digestible carbohydrate in the human diet and as such, is the major source of glucose that appears at relatively high concentrations in the blood circulation following intestinal digestion of a starch-containing meal. The rst stage in the metabolism of starch is catalysed by a-amylase which progressively brings about hydrolysis of the polysaccharide resulting in the production of maltose, maltotriose and limit dextrins as the main products [1]. Considerable differences, however, can occur in the postprandial blood glucose and the corresponding insulin response, to the ingestion of different foods containing identical amounts of starch [2]. That such differences www.starch-journal.com

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

396

P J. Butterworth et al. .

Starch/Starke 2011, 63, 395405 gland [13]. In humans and other mammals, two distinct but closely related genes code for amylase. AMY1, the salivary gene, produces the enzyme present in saliva and mammary gland and AMY2 produces the enzyme synthesised in the pancreas that is secreted into the duodenum in digestive juices [13]. The expression of AMY1 can also occur in certain tumour tissues, e.g. lung. In humans both AMY1 and AMY2 are located on the short arm of chromosome 1 [1416]. The coding regions of AMY1 and AMY2 are very similar but the 50 untranslated region of AMY2 is shorter than the equivalent region in AMY1 [12]. The genes code for enzymes with MWs of about 56 000 and the amino acid sequences of the salivary and pancreatic enzymes differ by only 3%. Both AMY1 and AMY2 are polymorphic and alloenzyme forms of the salivary and pancreatic amylases have been detected [12, 13]. The two pancreatic forms are exclusive to the pancreas and are expressed at nearly equal levels in the gland [12]. The secreted salivary amylase has the N-terminal blocked by a pyroglutamic acid residue [17] that may confer some resistance to proteolysis. The organisation of the amylase gene cluster shown below is based on published data [12]. The genes are arranged in a head to tail orientation except for AMY1B which is arranged in the reverse order. AMY1P is a pseudogene that lacks the rst three exons of AMY1 [13].
AMY2____AMY2A____AMY1A____AMY1B____AMYP1____AMY1C

occur is evidence of large variations in the rate and extent of starch digestion in the gastrointestinal tract [36]. Attenuating the uctuations in postprandial glycaemia and insulinaemia is important in the prevention and treatment of life-style associated diseases, notably diabetes mellitus and cardiovascular disease, and also has implications for obesity management [710]. The search for full explanations for the differences in the rates of digestion of starch-containing foods continues to be a main focus for research, but inter-alia, the differences are mainly attributed to inherent dissimilarities in the structure of starch granules of different botanical species and from structural alterations subsequent to domestic and commercial processing of foods [11]. Native starch is packaged in granules that are semicrystalline and which are very large compared with the size of the a-amylase molecule. This contrasts with the vast majority of intracellular metabolic enzymes where the enzyme molecules are usually huge in size compared with the metabolites that they act upon. The starch granule size seemingly presents a very favourable target for attack by amylase with many potential sites for binding of the enzyme. In spite of this apparent binding advantage, the complete breakdown of starch within an intact granule is a fairly slow process. Crystalline areas tend to be unfavourable for enzyme attack and, in addition, the granules may contain small but variable amounts of proteins and lipids that can also hinder starchamylase interaction. Most of the starch consumed by humans will have been cooked and/or subjected to various processes during food production that disrupts the granules to a greater or lesser extent, but raw starch is consumed in many animal feeds. Processing that disrupts general granule integrity and reduces the degree of crystallinity, increases the susceptibility to amylase. In recent years, considerable advances have been made in knowledge of the structures of starch and mammalian a-amylases and much has also been learned about the genes coding for a-amylase and their tissue expression. It seems timely, therefore, to review relevant aspects of what is known about starch and the enzyme to see whether certain uncertainties in the starch-amylase story can now begin to be claried and also perhaps point up aspects of this subject that still need to be resolved by further experimentation.

The promoter regions upstream of all the amylase genes contain a g-actin pseudogene and this nding has been interpreted to indicate that all the amylase genes are derived from a single ancestral gene [18]. In all except AMY2B, the g-actin pseudogene is interrupted by a retroviral-like sequence and this element probably serves a regulatory role in the expression of the salivary enzymes [12]. Meisler and Ting [12] suggest that an ancestral mammalian pancreatic gene gave rise to an ancestral AMY2 gene which in turn, after g-actin and retroviral insertion, gave rise to the present AMY1 and AMY2 forms. A remarkable feature is that the copy number of the amylase genes is variable [12, 19]. For the AMY1 (salivary) gene, variations in copy number from 2 to 15 have been detected from the screening of 50 individuals [20].

3 2 Human a-amylase genes

a-Amylase structures and properties

The genes for a-amylase have been cloned by several groups and the article by Meiser and Ting [12] provides an excellent review of the subject. Salivary glands and the pancreas express the enzyme in abundance but amylase expression can also be found in the jejunum and mammary 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

X-ray crystallography has been used to obtain 3D structures of the a-amylase enzymes from human pancreas [21] and saliva [17] and from porcine pancreas [22]. The structures of these enzymes are all very closely related. Mammalian amylases are composed of three structural domains, the largest of which (domain A) contains important active site residues that include two aspartate and one www.starch-journal.com

 Starch/Starke 2011, 63, 395405 glutamate residue. Domain A also contains a bound Cl ion. The activation of amylase by chloride has long been known [23, 24]. Domain B borders the active site region and contains a bound Ca2 ion that is probably important in maintaining protein conformation in the region of the active site. Domain C at the N-terminal region of the molecule appears unlikely to play a direct role in the catalytic mechanism. It is less rmly associated with domains A and B [21, 25]. The catalytic event in amylases is an example of a double displacement mechanism and carboxyl groups of aspartate and glutamate residues function as acid/base catalysts and as a nucleophilic reactant for the formation of a covalent intermediate during the catalytic sequence. The activation by a Cl ion may come about by facilitating the protonation of a catalytically important carboxyl group through a charge relay-system [2628]. a-Amylase is an endoenzyme that carries out multiple attack on linear portions of AM and AP with maltose and maltotriose as the principal short chain products [1, 29]. Only minor amounts of glucose are produced [1, 30]. Kinetic studies suggested that the active site of porcine pancreatic enzyme can accommodate up to ve glucose residues, i.e. there are ve sub-sites at which glucose residues can become bound [1]. The existence of multiple sites was conrmed by the production of 3D structures from X-ray crystallography, although a report of the structure of pig pancreatic amylase complexed with the potent carbohydrate inhibitor acarbose provided evidence of a sixth site [22]. The glucosidic bond susceptible to attack is that linking residues 3 and 4 (Fig. 1) and the catalytically important aspartate and glutamate residues of the active site are suitably located for attack on the susceptible link. Kinetic studies using malto-oligosccharides as substrates [31] were used to determine the binding free energy (DG) associated with each of the ve sites. DG values become more negative (from approx 5 to 16 kJ/mol), i.e. binding becomes stronger, as the sub-sites 15 become

397 lled with glucose residues in a substrate [3032]. The binding energy of site 3 is 17.6 kJ/mol and this unfavourable condition probably arises from forced distortion of the glucan ring on binding. The binding energies are similar to those observed for the sites on lysozyme [33]. The total free energy of binding accruing from occupancy of all the subsites is of consequence when considering the action of amylase on native and hydrothermally treated starch (see below).

4 Starch structure and implications for a-amylase action

Descriptions and discussion of starch structure and properties are covered only briey in this review. The information presented here is that which seems most pertinent to understanding the role of amylase in starch digestion. The literature contains many excellent general accounts of starch structure both in its native state and of the changes that occur in structure resulting from various methods of processing and readers are referred to these for more detailed information [3441]. Native starch granules are semi-crystalline with amorphous regions largely composed of AM and the branch points in AP molecules, and crystallites formed from longer chains of AP that twist into H-bonded double helices. Crystalline regions are of two polymorphic types, A and B [4245], that differ in the packing of the helices. The helices in A are orthogonally arranged whereas there is hexagonal packing in B starches [43, 44]. The number of associated water molecules is greater in the B crystals [43, 44]. Cereal starches have A type crystallinity and tuber starches such as potato possess the B polymorph, whereas legume starches have a so-called C pattern of crystallinity that is known to arise from the presence of both A and B polymorphs in different regions of the granule [39]. AM is enriched towards the periphery of the granules and this peripheral material has shorter chain lengths than AM located more centrally within the granule [46]. Granules from cereal starches, e.g. maize, wheat and barley, have surface pores that open onto channels with diameters that range in size from 5 to 400 nm [46]. Native starches with a high B polymorph content, e.g. potato, are attacked slowly by amylase but can be digested more rapidly after hydrothermal processing [47]. Native A type starches tend to be more favourable substrates [47]. During domestic and commercial food processing, granules can be subjected to varying degrees of hydrothermal treatments which cause granules to swell by absorption of water and crystalline structures are disrupted with an increase in amorphous material [41]. The increase of amorphous material makes the starch more digestible by a-amylase [11]. Continued heating in aqueous www.starch-journal.com

Figure 1. Subsites at the active centre of a-amylase. Bond breakage occurs between residues 3 and 4. Carboxyl groups of key aspartate and glutamate residues function as acid-base and nucleophilic catalysts [22]. 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

398

P J. Butterworth et al. .

Starch/Starke 2011, 63, 395405 of uncertainty still remains with regard to the physiological signicance of salivary amylase. A signicant paper was published in 2007 [20] that reported on the relationship between the copy number of the human amylase gene and dietary patterns in a number of human populations. The authors reported that individuals from populations that consume diets with relatively high starch content have a higher number of AMY1 gene copies than populations who consume less starch. They also reported that amylase protein levels correlated to a considerable degree, with the gene copy number but the presence of some discrepancy between copy number and protein expression is evidence of other factors that control the expression of the AMY1 genes. Thus it was concluded that the evolution of salivary amylase was associated with increases in the quantity of starch within the diet. Unfortunately, their estimates of amylase protein level in saliva are so much higher than other data in the literature that there are reasons for questioning the validity of their ndings (see below). An extension of the study to include other primates revealed that chimpanzees possess fewer copies of AMY1 and Perry et al. [20] point out that salivary amylase protein levels are much higher in humans than in chimpanzees. The diet of chimpanzees probably contains less starch and therefore the greater copy number of AMY1 in humans may reect a response to increased evolutionary pressure and that salivary amylase brings survival advantages. Presumably, early hominids would have consumed considerable amounts of uncooked and minimally processed (i.e. milled) food that would have contained native starch granules. Native starch is digested much more slowly than hydrothermally treated material (see below) and perhaps higher copy numbers of the genes were required in order to produce the enzyme in quantity so that the relatively non-reactive starch was broken down sufciently fast enough to contribute usefully to an individuals energy requirement. Given therefore, that the enzyme brings advantages to human survival, what is its exact role? Amylase activity from the AMY1 gene can be detected in saliva and gastric aspirates of prematurely born infants and the level increases with gestational age from 26 to 42 wk [51]. The expression of pancreatic amylase is minimal until weaning introduces non-milk feeds [50, 51] and the pancreatic gene is probably under partial control by corticosteroids and other hormones [52]. Comparing estimates of salivary amylase activity quoted in the literature is complicated by the fact that authors do not always dene their units for expressing catalytic activity and different assay temperatures may have been used. Some authors express the enzyme concentration in terms of mg/mL. If activity has been measured in international units (IU) per mL, i.e. the production of 1 mmol of product www.starch-journal.com

conditions results in solubilisation and gelatinisation of the starch. On cooling, the disrupted material can recrystallise into a form that is resistant to amylase action [48, 49]. This material is known as retrograded starch. In crystalline starch, the packing of the double helices are stabilised by hydrogen bonding between glucan residues. For optimal amylase action, a run of glucan residues needs to ll the ve subsites. Each glucan residue H bonds to various amino acid side chains that line the active site. Intuitively therefore, it can be expected that portions of polysaccharide chain that are intimately associated in forming crystallites are unlikely to bind securely to the active site. Estimations of the binding energies associated with the binding of a glucan residue into each of the subsites (see above) reveal the importance of lling of all sites for optimal catalytic action. It is likely however, that crystalline regions contain a number of areas where the crystals are imperfect to give amorphous zones where the chains are capable of becoming H-bonded to active site residues of amylase. There would seem to be a high probability of damage to crystalline regions during industrial processing such as milling. Also, it can be envisaged that if damaged areas are relatively few in number, enzyme binding to these amorphous zones is unlikely unless the concentration of amylase is high. Provided that sufcient water molecules are available for the hydrolytic reaction, amylase can catalyse chain breakage. Introducing breaks in polysaccharide chains at these amorphous foci could lead to further local destabilisation of crystallinity and favour continued amylolysis. The rates of reaction are still likely to be relatively slow however, because after bond breakage, the enzyme has to shuttle along the glycan chain so that the subsites that span the residues where bond breakage occurs are newly occupied. Alternatively, amylase can associate with another section of chain. Before either of these actions is possible, the product from the rst step must be expelled from subsites 4 and 5 of the active site and diffuse away. The constraints imposed by localised crystallinity and relatively low water availability are likely to be reected in a low rate of hydrolysis.

Salivary amylase

An editorial comment that appeared in a 1987 issue of Digestive Diseases and Sciences described how the role of salivary amylase in the digestion of starch was controversial because of the assumption that most of the digestion of starch occurs in the small intestine in the presence of a very active pancreatic enzyme [50]. Since that article appeared, considerably more has been learned about starch structure and the molecular biology and metabolomics of amylases, including that from saliva, but a degree 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

 Starch/Starke 2011, 63, 395405 per min, comparison is easier but again, assay temperatures of 20, 30 and 378C have all been popular choices. The commercial supplier SigmaAldrich1 list their amylase activities in terms of milligram maltose produced in 3 min at 208C. This corresponds closely to 1 IU [53]. If the turnover number for human salivary amylase is similar to that of porcine pancreatic enzyme (i.e. approximately 105/ min), a concentration of 1 mg/mL would represent approximately 20 103 IU/mL of saliva. Sigma1 market a puried preparation of human salivary amylase with a specic activity of 1.5 103 per mg that accords well with the estimate given above that is based on the properties of porcine pancreatic enzyme. Using this assumption, the comparison of AMY1gene copy number with amylase content of saliva (see [20] and above) indicates that saliva would contain up to 2 104 IU/mL. Activities obtained by direct measurements and reported in the literature give values that range from 10 to 160 IU/mL, dependent on the degree of stimulation of saliva ow and the method of collection [5457]. Thus the expected amylase activity of secreted saliva based on measurement of protein [20] is far in excess of the activities reported from direct measurements [5457]. There is an obvious disparity therefore, between calculations based on enzyme assay and those based on protein content. Mastication and physical characteristics of food inuence the secretion rate of saliva and its amylase activity [58]. The total protein content of saliva has been reported to range from 0.1 to 1.0 mg/mL [58] and therefore the report of concentrations of up to 7 mg/mL of amylase per mL saliva [20] must be in error. This is a problem that needs resolving. In vitro experiments have demonstrated that the salivary enzyme can be protected by starch and oligosaccharides against inactivation by the gastric acidic environment [59, 60] and thus it can be assumed that it could reach the duodenum and continue to act upon dietary starch. The AMY1 gene is also expressed in the lactating mammary gland of humans and many other mammals including cows and secreted a-amylase is present in human and cows milk [61, 62]. Thus although amylase from saliva and from milk can reach the duodenum in an active condition it begs the question of what, prior to weaning and the introduction of some starch to the diet, does it act upon when it gets there. Perhaps the enzyme protein possesses important functional properties that have yet to be recognised. Human milk contains non-starch oligosaccharides that are largely formed from galactose but these oligosaccharides are not acted upon by a-amylase. The undigested oligosaccharides can reach the colon and provide a source of nutrient for developing colonies of bacteria [63]. From examination of faecal samples from infants, it was found that bacteria capable of metabolising lactose and plant polysaccharides appear at a stage prior to the introduction 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

399 of solids into the infant diet. The milk oligosaccharides could therefore have played a priming role. It seems that in some populations from undeveloped areas of the world, nursing mothers chew solid food materials containing starch, e.g. roots and tubers, and then pass the pre-masticated food to their infants to supplement the main supply of nutrients which the infants receive in milk [64]. The amylase present in the saliva of the mother and the infant will begin to digest any starch present and this process will continue when the food bolus reaches the duodenum, particularly if the amylase activity is boosted further by a contribution from that present in a recent feed of milk. In more developed nations, milk supplements containing starch are used frequently and if these are introduced into feeds at an early stage, digestion of the starch will require the action of salivary amylase until the pancreatic enzyme becomes expressed in response to dietary and/or hormonal signals [52]. Archeologists have detected starch granules in dental calculus taken from human skeletal remains [65] and presumably, there are potentially benecial effects of salivary enzyme that can come from attacks on starch lodged between the teeth and thus aid removal of the obstruction. The oral bacterium Streptococcal gordonii that is associated with dental plaque formation can bind to salivary amylase [66]. The bacterial-bound enzyme retains most of its activity and is thus able to catalyse the hydrolysis of starch in the mouth. S. gordonni can convert the maltose released from starch to glucose and subsequently to lactic acid. A localised concentration of lactic acid in the dental plaque area can damage the enamel on neighbouring teeth [66] and lead to dental caries in circumstances of poor oral hygiene. The acquisition in S. gordonii of a specic amylase binding mechanism seems to represent a developmental response to a ready supply of nutrient albeit at increasing the risk of tooth decay in the host. The potential benet to humans that could arise from amylasecatalysed removal of starch lodged between the teeth seems hardly to be realised, therefore. Recently it has become apparent that genes for taste receptors are expressed in stomach and other areas of the gastrointestinal tract. Sweetness receptors in upper sections of the tract could detect any maltose formed from starch by salivary amylase and if coupled to the release of a number of signalling peptides such as GLP-1 and PYY inuence CNS, mediated control of gastric emptying, insulin secretion and appetite [67]. In the case of maltose formed in the mouth, there could also be a feed-forward effect in that after detection by receptors in the gastrointestinal tract, the signals could be interpreted as a warning that a starch containing meal has been consumed and so lead to increased secretion of pancreatic amylase to digest the expected starch load. Perhaps not unrelated to this property, is the nding that salivary amylase decreases the perception, by mouth taste www.starch-journal.com

400

P J. Butterworth et al. .

Starch/Starke 2011, 63, 395405

sensors, of the saltiness of savoury liquids when the latter are thickened with starch [68].

6 Application of enzyme kinetic models to analyse amylase action on starch in vitro

A recently published review contains a full account of various kinetic approaches that have been taken by authors in studies of a-amylase catalysed breakdown of starch [69]. The methods include digestibility measurements made over relatively long time periods and ts to the MichaelisMenten equation in both its standard and integrated forms. This present article will mainly concentrate on the usefulness of MichaelisMenten kinetics because this is an area which our laboratory has concentrated upon for several years to yield fruitful information about starchamylase interactions. Rather surprisingly, it has been stated that the MichaelisMenten model cannot be applied to in vitro studies of a-amylase action because of high enzyme activities, uncertainties in structural changes occurring in the substrate (starch) and the presence of product inhibition [70]. Obviously, differences of opinion exist among workers in the amylase area. Part of the controversy may arise because of a lack of familiarity with the kind of information that can be gleaned from studies of initial rates of catalysis as opposed to studies of starch digestion made over relatively long reaction times with very high concentrations of enzyme. These conditions of time and amylase concentration are probably selected due to the insensitivity of the commonly used reductometric methods for the detection of reducing sugar products of amylolysis. Digestion curves are popular with many investigators. The curves are usually tted to a 1st order rate equation rst proposed by Goni et al. [71]: C C1 1ekt (1)

Figure 2. Digestibility curves for amylase acting on Sorghum grains. Particles of different size were produced by a controlled milling process. The experimental points are fitted by a first-order kinetic equation (reproduced, with permission, from Mahasuhkonthachat et al. [70]). equation to determine values for the apparent 1st order constant, k [70]. These authors have also shown that 1/k is proportional to the surface area of the granule calculated from estimates of the diameter of the granules that are assumed to be spherical. They state that if the ratedetermining step is related to the diffusion of amylase onto the surface of the granule particle, a diffusion coefcient can be calculated from the equation 1 X2 k 6D (2)

where C is the concentration of digested starch at time t, C1 the digested concentration at the end of the reaction and k is a rst order rate constant. Various botanical starches generate different values for k that reect their relative susceptibilities to digestion. Values typically reported for k are in the range of 105103 min1. C1 is somewhat misleadingly referred to as the equilibrium concentration [71]. There is no measurable back reaction under the conditions pertaining in these experiments however, and so reference to an equilibrium is not correct. More realistically, C1 represents the concentration of starch digested when the reaction has reached a stage where no more product is formed, i.e. when all starch that can be hydrolysed has been exhausted and only so-called RS remains. Figure 2 shows good examples, taken from the literature, of digestibility curves and ts to the 1st order 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

where X2 is the granule specic surface area (m2/g) of a spherical particle and D is the diffusion coefcient. The values obtained for D turn out to be extremely small, i.e. approximately 1 1010 cm2/s. Thus it takes about 100 s to diffuse through 1 mm2. The granuleamylase collision rate in water is probably at least as high as 108 s1 [72]. We have recently studied the rates of binding of amylase to starch granules at 08C and determined apparent rate constants for binding (Warren et al. manuscript in preparation). Binding involves collision with the granule and then capture. If the number of suitable binding areas on the granule surface is very low, then the expected capture rate will be low relative to the collision rate. Direct binding studies provide some information on capture rates. In fact we found that the binding constant kon is about 106 M1/s. This value seems reasonable given that the collision rate is likely to be about 108 s1 [72] and the turnover number for amylase is approximately 104 s1 [47, 53]. Our results indicate that the binding rate is sufciently fast to allow a conclusion that diffusion of the enzyme to the surface of granules must be much faster than the estimates derived from digestibility constants. Since the latter are obtained www.starch-journal.com

 Starch/Starke 2011, 63, 395405 from complete digestion of susceptible starch over relatively long times, the very small diffusion coefcients probably represent gradual penetration of amylase into the granule as digestion proceeds. Dhital et al. [73] came partly to this conclusion from studies of the effects of particle size on digestibility rate constants. In spite of demonstrations that digestibility proceeds by a single rst order reaction process, many authors employ a terminology introduced by Englyst et al. [74] whereby digestibility curves are divided into phases of rapidly digested (RDS) and slowly digested starch (SDS). The RDS fraction is calculated from estimates of the amount of products released within the rst 20 min of digestion and SDS from measurements of product formation after periods of up to 2 h [74]. The percentage of RDS material is said to correlate well with the glycaemic response measured in vivo following ingestion of starch. Unfortunately the use of RDS and SDS promulgates the false idea that granules contain fractions of starch that are digested at different rates. If the digestion reaction can be described by a single rst order reaction with only one observed rate constant k, there are no inherent differences in the potential rate of reaction of any digestible starch fraction. The measured rate of digestion will be greater at earlier stages of reaction than in later periods because the fall in available starch substrate concentration as reaction proceeds, inevitably results in a slower rate. There is a need to provide a straightforward description of the relative rates of digestibility of starches from different plant varieties and food sources, and a good way of expressing this for a particular starch source might be to quote the time required for digestion of 50% of available starch under agreed standard conditions of amylase concentration and temperature etc. This is calculated from the value of ln2/k. Digestibility curves can also be used to calculate hydrolysis indices (HI) from determination of the area under the curves (AUC) obtained by integration of the 1st order rate equation with the bounds of t0t. C1 t C1 ekt AUC k   (3)

401 recognition that sharp excursions (peaks) in blood glucose concentrations occurring relatively early in postprandial periods may provide valuable patient information for dieticians and clinicians [76]. The rate of reaction (v) during the early stages of amylase reaction on starch is linear with time and can be related to the initial starch concentration (S) by the familiar MichaelisMenten equation: v Vmax S Km S (4)

where Vmax is the maximum velocity and Km is the Michaelis constant. Vmax kcat E0 (5)

Hydrolysis indice values are an alternative to GI values for starch foods. GI is dened as the AUC of a postprandial blood glucose response curve following consumption of a test meal expressed as a percentage of the AUC measured with a reference food, e.g. glucose or white wheat bread [75]. GI determinations require estimations of blood glucose levels at various time points after the meal and not all laboratories possess suitable facilities. Hence, HI measurements may offer an easier alternative. GI measurements seem to be becoming less popular, perhaps because of the laboratory difculties already mentioned and of high cost implications [76]. There is also 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

where kcat is the catalytic rate constant (also called turnover number) and E0 is the total enzyme concentration. Reliable estimates of the kinetic parameters can be obtained if rate measurements are made within the early stages of the reaction before the onset of complications arising from a combination of product inhibition and substrate exhaustion. Reaction rates should be determined from measurements of product release with time and checked for linearity. Digestibility studies are often performed with very large concentrations of amylase that commonly far exceed the concentrations of enzyme that are present in duodenal uid or saliva. Initial rate measurements are better performed (in order to maintain linear product release with time) using reaction mixtures containing enzyme concentrations that approximate those found in vivo. From published data we estimate this to be approx 50450 IU/mL and 10160 IU/mL for pancreatic juice and saliva, respectively [54, 77]. We routinely use approximately 100 IU/mL of pancreatic amylase for kinetic work. For any starch sample, the Km value is the concentration of available (digestible) starch (As) that will support an initial rate of reaction of Vmax/2. Digestible starch is the fraction that can be readily acted upon by a-amylase. If the total starch concentration is S, the ratio S/As will be large and hence the measured Km value will also be large because it is calculated from total starch concentrations added to the reaction mixture. Any treatment of starch to increase the amount of As, will cause a fall in the S/As ratio and the measured Km value will reect this [78]. The Km measurements can provide a direct indication of the fraction of starch that can be readily digested and thus monitor changes in starch structure resulting from hydrothermal treatments and to study structure activity relationships in general. The proportion of digestible polysaccharide at a given temperature Tx relative to that present at the lowest measured temperature T0 can be obtained from the fracT T T T tion [Kmx =Km0 ], where Kmx and Km0 are the experimental Km values determined for starches pre-treated at temperawww.starch-journal.com

402

P J. Butterworth et al. .

Starch/Starke 2011, 63, 395405 comparing amylase reacting with starches from different botanical sources or foods and/or starches that have been pre-treated to mimic the effects on starch structure of food processing, e.g. hydrothermal treatments [80]. The ratio provides an extremely valuable guide to the effectiveness of the important early stages of a-amylase action on starch. For potato starch, CE increases 117-fold after pre-treatment at 608C [78]. The various kinetic parameters can be compared with the gelatinisation enthalpies of starch granules and with granule particle size. Figure 3 shows data from a recent study [54]. Both CE and initial hydrolysis rate are linearly related with granule surface area. Km increases with DH (i.e. increased order of a-glucan chains) but CE and initial hydrolysis rate decrease with DH (Fig. 3). These ndings add weight to the assumption that kinetic data can provide sensitive monitors for changes in starch structure and substrate availability induced by physical treatments [80].

Conclusions

Figure 3. Relationship between catalytic parameters and gelatinisation enthalpies for various starches. (A) CE (mM/min (mg/mL)S1 and DHgel (J/g). (B) Hydrolysis rate (mM/min) and DHgel (reproduced, with permission, from ref. [53]). tures Tx and T0, respectively. For example, for potato starch 25 60 the Km =Km ratio is 32.5 [78]. The turnover number kcat, gives the number of molecules of product formed per molecule of enzyme in unit time. The dimensions of kcat are time1. For amylase action on starch, the measured kcat value will reach the maximum possible when access to susceptible glycosidic linkages is not hindered by structural constraints, water molecules for the hydrolytic reaction are abundant and the reaction products are freely expelled from the catalytic site. These conditions are likely to be met when the enzyme acts upon gelatinised starch and the value for kcat should be the same for all starches. We have found this to be broadly true for a number of different botanical starches where the value ranges from (1.0 to 1.6) 105/min [47, 53]. The parameter represented by kcat/Km is called the specicity constant [79] or the catalytic efciency (CE). When an enzyme can react with more than one substrate, CE is a handy way of comparing the relative reactivity of the different substrates. Hence it can be very useful for 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Knowledge of the structures and properties of starch and a-amylase has increased signicantly in recent years. Such information is helping our understanding, at the molecular level, of the reasons for variations in the rates at which starches from different botanical origins and/or from different starch foods are digested by amylase. A number of uncertainties still remain, however, particularly with regard to the full physiological importance of the products of amylase gene AMY1 and the evolutionary signicance of possession of multiple copies of amylase genes. The choice of kinetic models for studying starch amylolysis in vitro is also a subject of some controversy. The authors thank the BBSRC, Danone Research, Paris, Unilever, Colworth and RHM Technology (Premier Foods), High Wycombe for their various nancial support of the work of our group over several years.

The authors have declared no conict of interest.

8 References
[1] Robyt, J. F in: Fraser-Ried, B. O., Tatsuta, K., Thiem, J., ., Cote, G. L., et al. (Eds.), Glycoscience, Springer-Verlag, Berlin, Heidelberg, Germany 2008, pp. 14371472. [2] Crapo, P A., Reaven, G., Olefsky, J., Postprandial plasma . glucose and insulin responses to different complex carbohydrates. Diabetes 1977, 26, 11781183. [3] Noah, L., Guillon, F Bouchet, B., Buleon, A., et al., Digestion ., of carbohydrate from white beans (Phaseolus viulgaris L.). J. Nutr. 1998, 128, 977985.

www.starch-journal.com

 Starch/Starke 2011, 63, 395405

[4] Jenkins, D. J. A., Kendall, C. W. C., Glycemic index: Overview of implications in health and disease. Am. J. Clin. Nutr. 2002, 76 (suppl), S266S273. [5] Ells, L. J., Seal, C. J., Kettlitz, B., Bal, W., Mathers, J. C., Postprandial glycaemic, lipaemic and haemostatic responses to ingestion of rapidly and slowly digested starches in healthy young women. Br. J. Nutr. 2005, 94, 948955. [6] Judd, P A., Ellis, P R., in: Souvamyanath, A. (Ed.), Traditional . . Medicine for Modern Times, CRC Press, Taylor and Francis Group, Florida, USA 2006, pp. 257272. [7] Jenkins, D. J. A., Kendall, C. W. C., Augustinson, L. S. A., Martini, M. C., et al., Effect of wheat bran on glycemic control and risk factors for cardiovascular disease in type 2 diabetes. Diab. Care 2002, 25, 15221528. [8] Salmeron, J., Manson, J. E., Stampfer, M. J., Colditz, G. A., et al., Dietary ber, glycemic load and risk of NIDDM in women. J. Am. Med. Assn. 1997, 277, 472477. [9] Frost, G., Leeds, A., Trew, G., Margava, R., Dornhorst, A., Insulin sensitivity in women at risk of coronary heart disease and the effect of low glycaemic diet. Metabolism 1998, 47, 12451251. [10] Mann, J., Dietary carbohydrate: Relationship to cardiovascular disease and disorders of carbohydrate metabolism. Eur. J. Clin. Nutr. 2007, 61 (Suppl 1), S100 S111. [11] Gallant, D. J., Bouchet, B., Buleon, A., Perez, S., Physical characteristics of starch granules and susceptibility to enzymatic degradation. Eur. J. Clin. Nutr. 1992, 46 (Suppl 20), S3S16. [12] Meiser, M. H., Ting, C.-N., The remarkable history of the human amylase genes. Crit. Rev. Oral Biol. Med. 1993, 4, 503509. [13] Groot, P C., Bleeker, M. J., Pronk, J. C., Arwert, F et al., . ., The human a-amylase multigene family consists of haplotypes with variable numbers of genes. Genomics 1989, 5, 2942. [14] Zabel, B. U., Naylor, S. D. L., Sakaguchi, A. Y., Bell, G. I., et al., High resolution chromosomal localization of human genes for amylase, proopiomelanocortin, somatostatin and a DNA fragment (D3S1) by in situ hybridization. Proc. Natl. Acad. Sci. USA 1983, 80, 69326936. [15] Muke, M., Lidgren, V., Martinville, B., Francke, U., Comparative analysis of mouse-human hybrids with rearranged chromosomes 1 by in situ hydridization and Southern blotting: High resolution of NRAS, NFGB and AMY on human chromosome 1. Somatic Cell Mol. Genet. 1984, 10, 589599. [16] Tricoli, J. V., Shows, T. B., Regional assignment of human amylase (AMY) to p22-p21 of chromosome 1. Somatic Cell Mol. Genet. 1984, 10, 205210. [17] Ramasubbu, N., Paloth, V., Luo, Y., Brayer, G. D., et al., Structure of human salivary a-amylase at 1.6 A resolution: Implications for its role in the oral cavity. Acta Cryst. 1996, 52, 435446. [18] Samuelson, L. C., Wiebauer, K., Snow, C. M., Meisler, M. H., Retroviral and pseudogene insertion sites reveal the lineage of human salivary and pancreatic a-amylase genes from a single gene during primate evolution. Mol. Cell. Biol. 1990, 10, 25132520. [19] Iofrate, A. J., Feuk, L., Rivera, M. N., Listewnik, M. I., et al., Detection of large scale variation in the human genome. Nat. Genet. 2004, 36, 949953.

403
[20] Perry, G. W., Dominy, N. J., Claw, K. G., Lee, A. S., et al., Diet and the evolution of human amylase gene copy number variation. Nat. Genet. 2007, 39, 12561260. [21] Brayer, G. D., Luo, Y., Withers, S. G., The structure of human pancreatic a-amylase at 1.8 A resolution and comparisons with related enzymes. Protein Sci. 1995, 4, 17301742. [22] Gilles, C., Astier, J.-P Marchis-Mouren, G., Cambillau C., et ., al., Crystal structure of pig pancreatic a-amylase isoenzyme II, in complex with the carbohydrate inhibitor, acarbose. Eur. J. Biochem. 1996, 238, 561569. [23] Levitski, A., Steer, M. L., The allosteric activation of mammalian a-amylase by chloride. Eur. J. Biochem. 1974, 41, 171180. [24] Lifshitz, R., Levitski, A., Identity and properties of the chloride effector binding site in hog pancreatic a-amylase. Biochemistry 1976, 15, 19871993. [25] MacGregor, E. A., Janacek, S., Svensson, B., Relationship of sequence and structure to specicity in the a-amylase family of enzymes. Biochim. Biophys. Acta 2001, 1546, 120. [26] Agarjahi, N., Feller, G., Gerday, C., Haser, R., Structural basis of a-amylase activation by chloride. Protein Sci. 2002, 11, 14351441. [27] Qian, M., Haser, R., Buisson, G., Duee, E., Payan, F The ., active centre of a mammalian a-amylase. Structure of the complex of a pancreatic a-amylase with a carbohydrate inhibitor rened to 2.2 A resolution. Biochemistry 1994, 33, 62846294. [28] Feller, G., le Bussy, O., Houssier, C., Gerday, C., Structural and functional aspects of chloride binding to Alteromonas haloplanctis a-amylase. J. Biol. Chem. 1996, 271, 23836 23841. [29] Robyt, J. F French, D., Multiple attack hypothesis of a., amylase action: Action of porcine pancreatic, human salivary and Aspergillus oryzae a-amylases. Arch. Biochem. Biophys. 1967, 122, 816. [30] Seigner, C., Proganov, E., Marchis-Mouren, G., The determination of subsite binding energies of porcine pancreatic a-amylase by comparing hydrolytic activity towards substrates. Biochim. Biophys. Acta 1987, 913, 200209. [31] Prodanov, E., Seigner, C., Marchis-Mouren, G., Subsite proles of the active center of porcine pancreatic a-amylase. Kinetic studies using maltooligosaccharides as substrates. Biochem. Biophys. Res. Com. 1984, 122, 7581. [32] Seigner, C., Prodanov, E., Marchis-Mouren, G., On porcine pancreatic a-amylase action: Kinetic evidence for the binding of two maltooligosaccharide molecules (maltose, maltotriose and o-nitrophenylmaltoside) by inhibition studies. Eur. J. Biochem. 1985, 148, 161168. [33] Fersht, A., Enzyme Structure and Mechanism, 2nd Edn., W. H. Freeman and Company, New York, USA 1985, pp. 435436. [34] Gallant, D. J., Bouchet, B., Baldwin, P M., Microscopy of . starch: Evidence of new level of granule organisation. Carbohydr. Polym. 1997, 32, 177191. [35] Buleon, A., Colonna, P Planchot, V., Ball, S., Starch ., granules: Structure and biosynthesis. Int. J. Biol. Macromol. 1998, 23, 85112. [36] Wang, T. L., Bogracheva, T. Y., Hedley, C. L., Starch as simple as A, B, C? J. Exp. Biol. 1998, 49, 481502. [37] Cooke, D. F Gidley, M. J., Loss of crystalline and molecular ., order during starch gelatinisation: Origin of the enthalpic transition. Carbohydr. Res. 1992, 227, 103112.

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.starch-journal.com

404

P J. Butterworth et al. .

Starch/Starke 2011, 63, 395405

[57] Panchbhai, A. S., Degwekar, S. S., Bhowte, R. R., Estimation of salivary glucose, salivary amylase, salivary total protein and salivary ow rate in diabetics in India. J. Oral Sci. 2010, 52, 359368. [58] Mackie, D. A., Pangborn, R. M., Mastication and its inuence on human salivary ow and a-amylase secretion. Physiol. Behav. 1990, 47, 593595. [59] Fried, M., Abramson, S., Meyer, J. H., Passage of salivary amylase through the stomach in humans. Dig. Dis. Sci. 1987, 32, 10971103. [60] Rosenblum, J. L., Irwin, C. L., Alpers, D. H., Starch and glucose oligosaccharides protect salivary-type amylase activity at acid pH. Am. J. Physiol. 1988, 254, G775780. [61] Lindberg, T., Skude, G., Amylase in milk. Pediatrics 1982, 70, 235238. [62] Fox, P F Kelly, A. L., Indigenous enzymes in milk: Overview . ., and historical perspectives. Int. Dairy J. 2006, 16, 500516. [63] Koenig, J. E., Spor, A., Scalfone, N., Frisker, A. D., et al., Succession of microbial consortia in the developing gut microbiome. Proc. Natl. Acad. Sci. USA 2010, DOI: 10.1073/pnas.1000081107. [64] Pelto, G. H., Zhang, Y., Habnicht, J.-P Premastication: The ., second arm of infant and young child feeding for health and survival. Matern. Child Nutr. 2010, 6, 418. [65] Hardy, K., Blakeney, T., Copeland, L., Kirkham, J., et al., Starch granules, dental calculus and new perspectives on ancient diet. Archeol. Sci. 2009, 36, 248255. [66] Scannapieco, F A., Bhandary, K., Ramasubbu, N., Levine, . M. J., Structural relationship between the enzymatic and streptococcal binding sites of human salivary amylase. Biochem. Biophys. Res. Comm. 1990, 173, 11091115. [67] Little, T. J., Gupta, N., Maynard-Case, R., Thompson, D. G., McLaughlin, J. T., Sweetness and bitterness taste of meals per se does not mediate gastric emptying in humans. Am. J. Physiol. 2009, 297, R632R639. [68] Ferry, A.-L., Hort, J., Mitchell, J. R., Cook, D. J., et al. Viscosity and avour perception: Why is starch different from hydrocolloids? Food Hydrocolloids 2006, 20, 855862. [69] Dona, A. C., Pages, G., Gilbert, R. G., Kuchel, P W., . Digestion of starch: In vivo and in vitro kinetic models used to characterise oligosaccharide or glucose release. Carbohydr. Polym. 2010, 80, 599617. [70] Mahasukhonthachat, K., Sopade, P A., Gidley, M. J., Kinetics . of starch digestion in sorghum as affected by particle size. J. Food Eng. 2010, 96, 1828. [71] Goni, I., Garcia-Alonso, A., Saura-Calixto, F A starch ., hydrolysis procedure to estimate glycemic index. Nutr. Res. 1997, 17, 427437. [72] Fersht, A., Structure and Mechanism in Protein Science, W. H. Freeman and Company, New York, USA 1999, pp. 158159. [73] Dhital, S., Shrestha, A. K., Gidley, M. J., Relationship between granule size and in vitro digestibility of maize and potato starches. Carbohydr. Polym. 2010, 82, 480 488. [74] Englyst, H. N., Kingman, S. M., Cummings, J. H., Classication and measurement of nutritionally important starch fractions. Eur. J. Clin. Nutr. 1992, 46, 3350. [75] Jenkins, D. J. A., Wolever, T. M. S., Taylor, R. H., Barker, H., et al., Index of foods. A physiological basis for carbohydrate exchange. Am. J. Clin. Nutr. 1981, 34, 362366.

[38] Donald, A. M., Waigh, T. A., Jenkins, P J., Gidley, M. J., et al., . in: Frazier, P J., Donald, A. M., Richmond, P (Eds.), Starch: . . Structure and Functionality, Royal Society of Chemistry, Cambridge, UK 1997. [39] Bogracheva, T. Y., Morris, V. J., Ring, S. G., Hedley, C. L., The granular structure of C-type pea starch and its role in gelatinisation. Biopolymers 1998, 45, 323332. [40] Svihus, B., Uhlen, A. K., Harstad, O. M., Effect of starch granule structure, associated components and processing on nutritive value of cereal starches. Animal Feed Sci. Technol. 2005, 122, 303320. [41] Tester, R. F Morrison, W. R., Swelling and gelatinisation of ., cereal starches. I Effects of amylopectin, amylose and lipids. Cereal Chem. 1990, 67, 551557. [42] Gidley, M. J., Bociek, S. M., Molecular organization in starches: A carbon-13 CP/MAS NMR-study. J. Am. Chem. Soc. 1985, 107, 70407044. [43] Imberty, A., Perez, S., A revisit to the three dimensional structure of B-type starch. Biopolymers 1988, 27, 12051221. [44] Imberty, A., Chanzy, H., Perez, S., Buleon, A., Tran, V., The double helical nature of the crystalline part of A starch. J. Mol. Biol. 1988, 201, 365378. [45] Manners, D. J., Recent developments in our understanding of amylopectin structure. Carbohydr. Polym. 1989, 11, 87112. [46] Perez, S., Baldwin, P Gallant, D. J., in: BeMiller, J. N., ., Whistler, R. L. (Eds.), Starch: Chemistry and Technology, 3rd edn. Academic Press, London, UK, New York, USA 2009, pp. 149192. [47] Slaughter, S. L., Ellis, P R., Butterworth, P J., An investigation . . of the action of porcine pancreatic a-amylase on native and gelatinised starches. Biochim. Biophys. Acta 2001, 1525, 2936. [48] Gidley, M. J., Cooke, D., Darke, A. H., Hoffman, P et al., ., Molecular order and structure in enzyme resistant retrograded starch. Carbohydr. Polym. 1995, 28, 2331. [49] Hoover, R., Starch retrogradation. Food Rev. Int. 1995, 11, 331346. [50] Lebenthal, E., Role of salivary amylase in gastric and intestinal digestion of starch. Dig. Dis. Sci. 1987, 32, 11551157. [51] Hodge, C., Lebenthal, E., Lee, P C., Topper, W., Amylase in . the saliva and in the gastric aspirates of premature infants: Its potential role in glucose polymer hydrolysis. Pediatr. Res. 1983, 17, 9981001. [52] Swanson, K. C., Matthews, J. C., Matthews, A. D., Howell, J. A., et al., Dietary carbohydrate source and energy intake inuence the expression of pancreatic a-amylase in lambs. J. Nutr. 2000, 130, 21572165. [53] Tahir, R., Ellis, P R., Butterworth, P J., The relation of physical . . properties of native starch granules to the kinetics of amylolysis catalysed by porcine pancreatic a-amylase. Carbohydr. Polym. 2010, 81, 5762. [54] Froehlich, D. A., Pangborn, R. M., Whitaker, J. R., The effect of oral stimulation on human parotid salivary ow rate and amylase secretion. Physiol. Behav. 1987, 41, 209217. [55] Harthoorn, L. F Salivary a-amylase: A measure associated ., with satiety and subsequent food intake in humans. Int. Dairy J. 2008, 18, 879883. [56] Beltzer, E. K., Fortunato, C. K., Guaderrama, M. M., Peckins, M. K., Salivary ow and a-amylase: Collection technique, duration and oral uid type. Physiol. Behav. 2010, 101, 289 296.

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.starch-journal.com

 Starch/Starke 2011, 63, 395405

[76] Seal, C. J., Daly, M. E., Thomas, L. C., Bal, W., et al., Postprandial carbohydrate metabolism in healthy subjects and those with type 2 diabetes fed starches with slow and rapid hydrolysis rates determined in vitro. Br. J. Nutr. 2003, 90, 853864. [77] Dahlqvist, A., Borgstrom, B., Digestion and absorption of disaccharides in man. Biochem. J. 1961, 81, 411418. [78] Egan, D., Biochemical mechanisms of starch digestion: Enzyme kinetic studies of alpha-amylase action on starch.

405
Ph.D Thesis, Kings College London, University of London 2008. [79] Fersht, A., Structure and Mechanism in Protein Science, W. H. Freeman and Company, New York, USA 1999, pp. 110111. [80] Roder, N., Gerard, C., Verel, A., Bogracheva, T. Y., et al., Factors affecting the action of a-amylase on wheat starch: Effects of water availability. An enzymic and structural study. Food Chem. 2009, 113, 471478.

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.starch-journal.com