RICE CRC

FINAL RESEARCH REPORT

ISBN 1 876903 41 4

P3402FR09-05

Title of Project : Understanding amylose structure, what it controls and what controls it. Project Reference number : 3402

Research Organisation Name : Yanco Agricultural Institute, NSW Agriculture, and CSU Wagga Wagga Principal Investigator Details :

Name : Melissa Fitzgerald and Christopher Blanchard

Address : MF: YAI, PMB, Yanco, NSW 2703 CB: School of Food and Wine Science, CSU, PO Box 588, Wagga Wagga, 2678. Telephone contact : 69 512 656 (current 63 2 580 5600 exn 5680) 69 332 364

TABLE OF CONTENTS

1. Summary ...... 1 2. Background ...... 2 3. Objectives ...... 2 4. Introduction ...... 2 5. Methodology ...... 4 6. Results ...... 9 7. Discussion ...... 31 8. Conclusions ...... 39 9. References ...... 39

RICE CRC Final report – Project 3402

Understanding amylose structure, what it controls and what controls it

1. Summary

Starch accounts for at least 92% (dry weight) of a milled rice grain. Starch is comprised of two fractions, amylose and amylopectin. Amylose content can range from 0% (in waxy rice) to about 30%. Amylose is essentially a linear molecule ranging from about 800 degrees of polymerization (DP) to about 10 000 DP. It carries a few widely spaced chains. Amylose plays a significant role in almost all of the cooking qualities of rice. The process of cooking of rice begins with the softening of the starch granules, which is primarily a function of amylopectin. The next process, swelling, is greatly affected by amylose. As the starch granules swell, amylose leaches from the granules into the solution phase. Behaviour observed in the field of synthetic polymer science suggests that the linear amylose molecules surround the swelling granules and inhibit the swelling. After amylose leaches from granules, it joins the continuous phase and van der Waal forces inside the helices of chains cause double helices to form. The double helices aggregate into a gel; the more double helices, the firmer the gel. The early stages of gel formation would occur in the interval between removing from heat and eating the rice. Long chains of amylose have a higher viscosity than short chains, and this limits the mobility of the long chains. Thus, with long chains, the formation of double helices and aggregations is slower, leading to a softer gel. Therefore, amylose structure could explain why two varieties with the same amylose content differ in cooked texture. In the later stages of gel formation, typically occurring well after cooking the rice, and when the temperature falls below 25 ºC, short chains of amylose will form double helices and crystallites much more readily than long chains of amylose. Therefore, rice that contains short chains of amylose are likely to be hard when cooled after cooking. The knowledge and information that could be provided by developing a method to measure amylose structure will provide a tool allowing greater insights into the effect of amylose structure on different cooking properties, with the ultimate aim of developing the knowledge into a selection tool for rice breeders.

After developing a tool to measure amylose structure, it was applied to understanding a particular nutritional property of rice, namely resistant starch. Literature and early research indicated some link between resistant starch content and amylose content, however, detailed investigations of the structure of resistant starch, hypothesised to reveal more of the secrets of amylose, in actuality, revealed some of the secrets of amylopectin.

-1-

2. Background

A rice grain consists only of starch (93 - 95%), protein (5-7%), lipid (0.5 – 1%), and minute amounts of other components like aromatic compounds, yet there is enormous variability in quality. The variability must lie in the processes of grain-filling that lead to structural differences in the grain, which, in turn, lead to perceptible differences in the sensory properties of the rice. The quality of cooked rice includes important traits like texture after cooking, stickiness, increase in hardness when consumed several hours after cooking (retrogradation), and satiety. Amylose accounts for as much as 30% of the starch, and amylose content contributes to most, if not all, of the cooking properties of rice. However, two varieties with the same amylose content do not necessarily have the same cooking properties. Furthermore, when the same variety is grown in different environments, the cooked rice from the different environments will have different properties. Amylose structure has often been cited as a likely explanation for the differences in cooking properties. Techniques were available last century for the sophisticated measurement of amylopectin structure, but not of amylose structure. The knowledge and information that could be provided by developing a method to measure amylose structure, correlating amylose structure with the properties of cooked rice, and understanding the control and heritability of amylose structure will allow the rice breeders to select for a particular structure of amylose. Informed selection for that trait will be a powerful tool that will hasten the development of new varieties with known properties. Increasing our capacity to guarantee quality will improve the reputation of Australian rice, assist in increasing our market-share, and thereby contribute to the economic sustainability of the NSW rice industry.

3. Objectives

The over-arching objectives of the project are:

• Understand some of the factors that control amylose structure at the genetic level; • Develop a method to measure the structure of amylose; • Understand how amylose structure contributes to selected cooking properties of rice.

4. Introduction

A single gene, Granule Bound Starch Synthase (GBSS1) (waxy) is known to synthesise amylose. Two alleles have been identified at the waxy locus – Wxa and Wxb. The alleles differ by a single base at the 5’ end of the splice site of intron 1 and the substitution in the Wxb allele decreases the splicing efficiency at that site, resulting in less transcript, protein and amylose. Temperate japonica rice predominantly carries the Wxb allele, and so Australian rice varieties, being predominantly genetically temperate japonica, are expected to carry the Wxb allele and be of low amylose. Amylose content does not explain cooking properties, thus it is timely to move beyond the gene and to search for allelic variation within the Wxb allele and associate that with variation in structure and function.

During the course of the project, it was reported that the two waxy alleles in rice could be further divided by a microsatellite in the flanking region of exon 1 which correlated with amylose content. Subsequent collaboration with that group led to the classification of Australian rice breeding program on the basis of that microsatellite.

-2-

Little is known about the structure of amylose, and even less is known about the variability in structure across the species. Amylose is a polymer of glucose units, and in rice, can range in size from about degrees of polymerisation (DP) 800 to DP 10 000. Molecules of amylose are described as ‘essentially linear’. They consist of a long back-bone which is connected to a small number of distally spaced branches. The branches are thought to range from DP 200 – 1000. It is likely that an isoform of the branching enzyme adds the branches. Branching enzyme 1 mutants show no difference in amylopectin structure in wheat and maize, suggesting either no or a minimal role in amylopectin synthesis or complete compensation by the other isoforms. Moreover, BE1 has shown a preference for inserting branches on long chains, suggesting that it might contribute to amylose synthesis and structure. Figure 1 shows a diagram of the structure of amylose. Linear, or essentially linear, synthetic polymers form gels much more easily than branched polymers, thus the processes of gel formation and entanglement of molecules are likely to be strongly associated with the amylose fraction of the starch

Fig. 1 Schematic Amylose rather than the amylopectin fraction. Application of knowledge from synthetic polymer chemistry suggests that different lengths of linear polymers and different branching structures affect the entanglement capacity of molecules, the speed which chains move through gels and thus the properties of the gel. Therefore, if we apply that knowledge to amylose, a biological polymer (essentially a chain), it becomes clear that we must delve deeper than content and investigate amylose structure. If variability exists in the chain length distribution, the variability will certainly contribute to our understanding of cooking properties.

Amylose content significantly contributes to many of the cooking properties of the rice, but amylose content alone is not sufficient to predict the cooking properties of rice. Two varieties of the same amylose content can easily differ in cooking properties. For example, Koshihikari, Amaroo and Millin are all varieties of low amylose, but the texture of the cooked rice of each variety differs significantly, and the texture of the cooked and cooled rice (retrograded) also differs significantly. Retrogradation is an important parameter of quality for Australian rice and is likely to be influenced greatly by amylose, thus understanding the association between amylose structure and retrogradation of rice forms part of the second and third objectives of this project.

Quality encompasses physical, sensory and nutritional parameters. As the western world becomes aware of the lower rates of obesity and terminal illnesses in the eastern world where a rice-based diet prevails, understanding the effects of rice on physiological function, well- being, and health is becoming increasingly important. Amylose content is likely to influence retrogradation, which is an important sensory property, and retrogradation is linked to an important nutritional property – resistant starch. Resistant starch is that portion of starch which is not digested in the gut. Retrograded rice is recrystallised and resists digestion. Resistant starch moves into the colon where it is fermented by microflora, producing butyrate and other short-chain fatty acids which have beneficial effects on a number of physiological functions and bowel conditions.

-3-

Resistant starch has been classified into four categories: Type 1 is physically inaccessible to digestive enzymes; Type 2 is naturally-occurring granules; Type 3 is retrograded; and Type 4 is chemically modified starch. The third objective of this work is three-fold: to determine the relationship between amylose content, structure, GBSS allele, and resistant starch, to determine the molecule in rice that accounts for naturally occurring resistant starch, and to determine the resistant starch in several ‘nationalities’ of rice, of cultural and economic importance within the nations, and the effect of cooking and processing on the RS.

5. Methodology

Three varieties of rice were chosen for this study. They were Amaroo, Millin and Koshihikari. All are medium grain and temperate japonica. Amaroo and Millin were developed within the Australian Rice Improvement Program and Koshihikari is a Japanese variety grown in Australia. Amaroo is descended from Californian germplasm, Millin is a cross between a Californian variety and a traditional Japanese variety, and Koshihikari was selected on-farm from a traditional variety. These three rices vary greatly in their cooking properties.

Objective 1: Searching for allelic variation in the GBSS gene

Primers were designed for each exon of the GBSS 1 gene in the three varieties, Amaroo, Millin and Koshihikari. The exons were amplified by PCR and the product was sequenced. Sequences of each exon from each variety were compared to search for variation within the exons that could alter the amino acid sequence and subsequently functionality of the active sites of the enzyme. This work was carried out at Charles Sturt University. Primers were also designed for exons of branching enzyme (BE) 1 and the PCR products of those assays were analysed by gel electrophoresis and any differences in size were sequenced to determine the genetic difference.

For identification of the microsatellite in each variety, DNA was prepared from brown rice, PCR was carried out, polyacrylamide gel electrophoresis was used to separate the fragments and fluorescence was used to visualise the bands. In order to confirm that the three varieties Amaroo, Millin and Koshihikari were temperate japonica in terms of their waxy allele, ie carried the T at the splice site of exon 1, DNA was incubated with the restriction enzyme AccI1 and products were separated on an agarose gel.

Objective 2: A method to measure amylose structure

In order to measure amylose structure, it must be separated from amylopectin. It is well known that amylose leaches from starch granules in hot water (tsai). The method was optimised to maximise the amount of amylose and minimise the amount of amylopectin in the soluble fraction. Different methods of heating, different temperatures and different speeds of centrifugation were tested. The optimised method was used for all further work. The Rapid Visco Analyser was used to solubilise the amylose. Flour (250 mg) was mixed with water (25 ml) and the standard RVA profile was run. After the run, the contents of the canister were centrifuged (10000 g, 10 min). The supernatant contained the hot water soluble (HWS) starch. HWS starch was separated using size exclusion chromatography (SEC). SEC was carried out using a Waters system consisting of an Alliance (2695) and Differential Refractive Index Detector (Waters 2410). Waters software (Empower®) was used to control the pump, and to acquire and process the data. The eluant was ammonium acetate (0.05 M, pH 5.2) flowing at 0.5 ml min-1. Two columns, an Ultrahydrogel 250 (UH 250) and an Ultrahydrogel 500 (UH

-4-

500), both from Waters, were tested to determine the range of molecular weights separated. Each column was used independently, and columns were held at 60°C. To determine the range of molecular weights separated by each column, pullulan standards ranging in molecular weight from equivalent to short amylopectin to long amylose chains were injected into each column.

In SEC, molecules are separated either on the basis of their hydrodynamic volume (Vh), which is defined by the shape of each molecule in solution, or (if they are all linear chains) on the basis of the molecular weight distribution (MWD) of the chains.

Linear Branched

Fig 2: Two molecules of the same volume clearly differ in molecular weight

In SEC the elution time of a molecule, whether it be linear or branched, is a direct function of its size in the eluant, and not its molecular weight. Thus each elution slice of the trace could contain molecules of the same shape, but not of the same MW (Figure 2). In order to understand the relationship between molecules of the same hydrodynamic volume but different structure, HWS starch from each variety was fractionated by eluting with ammonium acetate (0.05 M) from a column (20 x 5 cm) packed with Sepharose 6B. Each fraction was collected for 2 min and the presence of starch in each fraction was determined by addition of a drop of iodine solution to each fraction. Collection was terminated when fractions no longer continued to bind iodine. Fractions were analysed by SEC on the UH 500 column (as described earlier). Linear chains were determined in each fraction by digesting each fraction with iso-amylase (Megazyme), which hydrolyses the α1-6 linkages specifically. This enzyme was used to debranch starch in each fraction. For digestion with isoamylase, aliquots of HWS starch were mixed with sodium acetate buffer (pH 4, 0.5 mM), and then isoamylase was added (7 ul). Samples were incubated (50ºC, 2h). After incubation, samples were boiled and centrifuged as described above. Samples from both enzyme digestions were desalted using mixed bed resin (Biorad) for one hour. Fractions were analysed by SEC (as described above) before and after incubation with each enzyme. After digestion with beta amylase, if no chains remain in a particular fraction, then all molecules in that fraction were linear. If debranching does not increase the elution time of a particular fraction, then all chains in that fraction are similar and of similar length to the starch before debranching. .

-5-

SEC with a mass-sensitive detector, like refractive index, does not give the number of chains directly, but the weight average of the chains (w(log M) (Castro et al. 2005) – the bigger the molecule, the more the detector detects it, thus signal does not equate to molecular weight. The number of chains (P) of a particular molecular weight (M) can be determined by detecting each chain once, with ultraviolet or fluorescence detection (Castro et al. 2005). In order to use UV or fluorescence to detect the number of chains, it is necessary to label each chain only once with either a UV absorbing chromophore or a fluorophore. Chains of amylopectin have been successfully labelled and quantified using reductive amination with the fluorophore 8-amino 1, 3, 6, trisulphonic acid (APTS) (O’Shea and Morell), but not chains of amylose. Takeda et al. have attempted to measure amylose with other fluorophores, but problems are cost of the fluorophore, toxicity, co-elution of the label and the starch, and confidence with labelling efficiency being independent of molecular weight. In this project several chromophores and fluorophores were analysed for their capacity to measure the number of chains of amylose of each molecular weight. Attempts were made to separate the label and the starch to overcome the problem of co-elution. Various ion exchange resins and spin columns were used to bind the ions, and the best was incubation for 1h with mixed bed resin (Biorad). SEC conditions were also manipulated to separate the peak of excess label from peaks of interest. Labelling efficiency was tested by utilising the simple relationship of w(log M) = M2P(M) by labelling and analysing a range of standards of known molecular weight (M) using inline mass–sensitive detection (w(log M)) and fluorescence/UV detection (P(M)). The signal obtained by the UV/fluorescence detector was then expressed as w(log M) by multiplication of the UV/fluorescence signal by M2 and the mathematically treated signal compared with the signal from the mass sensitive detection. Fluorophores and chromophores were all sulphonic acids because these can be attached to the reducing end of the starch molecule by reductive amination – the only way to label starch.

Reductive amination was carried out as described by O’Shea and Morell (1998). SEC was carried out as described above, but with the additional detector, either fluorescence or UV, before the RI detector.

Objective 3: Relationship between amylose and resistant starch

* Resistant Starch (RS)

The objective of this work is three-fold. First, is to determine the relationship between amylose and resistant starch, second is to determine the molecule in rice that accounts for naturally occurring resistant starch, and third to determine the resistant starch in several ‘nationalities’ of rice, of cultural and economic importance within those nations, and the effect of cooking and processing on RS.

Three sample sets were used for this work. The first was Australian rice, the second was rice from the US, and the third was rice from Asia. The Asian collection included the amylose extender mutant of IR36. The amylose extender mutant is that used commercially in RS applications. Table 2 shows the varieties used and the origin of them. Australian rices were received freshly milled from SunRice, NSW Australia, the US rice was grown in buffer rows at the Beaumont Experiment Station of the USDA in Beaumont, Texas US, and the Asian rice was grown either at the International Rice Research Institute (IRRI) in the Philippines, or at PhilRice, also in the Philippines. Rice grown in Australia, or at IRRI or PhilRice was

-6-

dehulled with a Satake dehuller (THU35A 250V 50Hz Test Husker, Satake), and milled for 60 sec (Magill No.2). Rice grown in the US was received as milled.

Resistant Starch (RS) was measured on the different samples of freshly cooked or processed rice using the Resistant Starch Assay (Megazyme) (AACC Method 32-40). Figure 3 shows the steps of the method. Briefly, sample (0.5 g) was incubated with amyloglucosidase (AMG) (4 ml, xUml-1) and pancreatic alpha-amylase (10 mgml-1) for 16 hours at 37°C to reduce digestible starch to glucose. The reaction was terminated with 4 mL ethanol and the RS pellet was recovered by centrifugation (5000 g, 10 min). The supernatant was decanted and the washing and decanting procedure was repeated with ethanol (50% v/v). The pellet was solubilised in KOH (2 ml, 2 M) in an ice bath, neutralised with sodium acetate (8 ml, 1.2 M) and the RS hydrolyzed to glucose with of AMG (0.1 ml, 3300 Uml-1, 50°C). The glucose oxidase/peroxidase reaction (GOPOD) was used to measure glucose. Absorbance was read (510 nm) (GBC UV/VIS918) after a 20 minute incubation period at 50°C. RS was calculated as a percentage of dry weight of sample. The efficiency of the method was determined by analysing the amount of starch in both supernatants and both pellets.

The relationship between resistant starch, amylose content, solubility and amylose allele were also determined. Amylose content was measured by iodine binding (Blakeney et al 1994). Solubility was measured as described earlier, proportion of long chains was determined by debranching gelatinised flour and determining the proportion of chains greater than DP 100 by SEC (described earlier), and genetic variation at the GBSS locus was measured by determining polymorphism in the microsatellite in the flanking region of GBSS using the simplified method described by Bergman et al. (2000). In brief, DNA was extracted in NaOH and the neutralised DNA subjected to PCR using published primers (W484 and W485), and the PCR products were separated with 8% acrylamide gel, and visualised by staining with GelStar and visualising on a Dark Reader.

In order to determine the structure of RS, the RS was solubilised and neutralised (for all varieties) according to the RS assay, but it was not digested to glucose. Instead, it was incubated with mixed bed resin to absorb the acetate ions (Biorad AG 501-X8(D) Resin) and an aliquot of the sample (40 µl) was injected directly into the UH 500 column for SEC analysis (conditions as described earlier). Another aliquot was debranched (as described earlier) and the debranched RS was analysed by SEC. The debranched material was also labelled and the chain length distribution analysed by Capillary Electrophoresis (CE). Labelling was done by reductive amination of the reducing end and attaching a molecule of 8 amino 1,3 6, trisulphonic acid to each reducing end. Labelled chains were separated by CE as described by O’Shea and Morell.

The effect of several processing methods on resistant starch was measured using only the variety Doongara as an example of the way cooking and processing affected RS. Doongara was chosen because of its intermediate amylose, a trait common to the 3 sets of rice. Resistant Starch was measured on freshly cooked brown and white rice, rice cooked by rapid boil or absorption, under- and over-cooked rice, mashed or sieved cooked rice (to simulate chewing a little or chewing a lot), storing (retrogradation) for 1, 2 or 8 days, and freshly cooked and retrograded gels.

The three sets of rice were not analysed simultaneously, and so different data sets were collected for each set, though some data sets are common to all three.

-7-

Fig. 3. Resistant-Starch Determination by Megazyme Method

Weigh 0.5 g + 4 ml AMG-Pancreatin α -Amylase

Milled Rice Boiling for Stand for 16 Hours in In Water 35 minutes 10 minutes 37°C Water Bath

Terminate Reaction with 4 ml 99% EtOH

Centrifugation

1 ml + GOPOD Pellet 1 (RS)

Supernatant 1 (NRS) Residue 100 ml volumetric flask Add 2 ml 2M KOH

Wash with 50% EtOH Stir for 20 2X and minutes Centrifuge Read Absorbance at 510 nm For Non-Resistant Starch Neutralize with 8 ml 1.2M NaOAc

1 ml + GOPOD Centrifuge

Supernatant 2 Hydrolyze to glucose Incubate 50°C for 20 minutes with AMG in 50°C 100 ml volumetric flask for 30 minutes

Read Absorbance at 510 nm For Resistant Amylose Pellet 2 Starch Analysis

-8-

6. Results

Fig. 4. CT repeat in Aust pedigrees

Objective 1: Searching for allelic variation in the GBSS1 gene

Primers were designed around each of the exons of GBSS1. Exons of GBSS1 were amplified and sequenced from Millin, Koshihikari and Amaroo. No difference in the DNA sequence was found for any of the exons for these three rices. During planning of the next stage, a group in the US, led by Dr Bill Park, published data describing a polymorphic microsatellite in the flanking region of GBSS1 showing how it was inherited throughout the US germplasm and how it segregates with amylose content. Collaboration was developed with the group which led to determining the microsatellite variability in the Australian germplasm. Figure 4 shows polymorphism of the microsatellite throughout the Australian germplasm. Figure 4 shows that the 3 varieties chosen for the difference in their cooking properties, Amaroo, Millin and Koshihikari, all carry a different polymorph-ism of the gene (microsatellite). Amaroo carries the microsatellite that descends from the Californian stock (CT19), Millin from the low-quality, but quite cold-resistant, traditional variety Somewake (CT18), and Koshihikari has CT17, for which the origin is unknown. Efforts examining expression patterns of the genes showed no differences. No difference was found between the three varieties for the sequence in the exons of the Branching Enzyme 1.

-9-

Objective 2a: A method to measure amylose structure

Separation of amylose from amylopectin was achieved by using a low 0.7 High AM control concentration of flour in water and Millin heating with RVA. The low 0.6 Amaroo Koshi concentration ensured that the sample 0.5 Waxy was not subjected to any shear. The 0.4 RVA was stopped at several temperatures between 50 and 97ºC to 0.3 determine the temperature at which 0.2 amylose was released from the starch

Absorbance of iodine complex complex iodine of Absorbance granules. Amylose was determined by 0.1 iodine binding. Figure 5 shows that 0 amylose was not released until 95ºC. A 50 60 70 80 90 100 waxy variety shows that a small amount of amylopectin was released after Temperature (ºC) Fig. 5: Absorbance of ioding-starch complex in the gelatinisation and again after 95ºC supernatant at different temperatures (Figure 5). However, much higher absorbance was noted for the non- waxy varieties, and higher for Doongara (Figure 5), which was included as a positive control due to its higher amylose content.

12 a 4 b

10

3 8

6 2

4

1 2

0 0 10 11 12 13 14 15 16 17 10 12 14 16 18 20 22

-2 Time (min) Time (min)

Fig. 6: Separation of pullulan standards on the UH 250 column (a) and the UH500 column (b). Grey curves are either unseparated in the exclusion zone or too small to be within the accepted range of molecular weight for amylose.

-10-

Two SEC columns were tested to determine the suitability of use for separating amylose.Both were tested by injection of pullulan standards ranging in molecular weight. A pullulan standard of the same molecular weight as starch will not have exactly the same hydrodynamic volume, so will not elute at the same time as the starch of equivalent MW. However, since pullulan is glucose linked with similar bonding as starch, it was expected that elution times would be close to those of starch of equivalent MW.

Table 1 shows the molecular weight of the pullulan standards, and the equivalent degrees of polymerisation if the standard was starch. The separation of the pullulan standards on two different columns. The first column is of small pore size and the higher molecular weight standards, those corresponding closely to the range of molecular weights found for amylose, are not separated (Figure 6a dark grey). Chains less than DP 1000 elute in the separating phase of the column, and chains of the order of amylopectin chains (the lowest standard) elute towards the end of the separating phase of the column. Figure 6b shows the standards separated on a column with larger pores (UH 500). On this column, the higher molecular weight standards are nicely separated, but the two smallest standards are less well separated. Figure 7 is a calibration curve of elution time vs log M and shows that the relationship is linear, and that the slope alters for the last two, less well separated, standards. Since the UH500 seemed better able to separate chains of the order of amylose (Figure 6b and Table 1), that column was used to determine whether or not amylopectin contributed to the HWS starch and if amylose contributes to the HW insoluble starch.

10000000

MW DP equivalent 1000000 std of starch

100000 5900 36.4 11800 72.8 10000 22800 140.7 47300 292.0

112000 691.4 1000 212000 1308.6

404000 2493.8 100

788999 4864.2 10 12 14 16 18 20

Time (min)

Table 1: MW of standards and DP of starch Fig.7: Calibration curve for pullulan standards on the UH500 column showing linearity. at each MW. The first two are too small to equate to amylose Change in slope (arrow) suggests poorer separation

-11-

30 10 a b a b

25

8

20

6

) 15 M g lo ( 4 W 10

2 5

0 0

810121416182022 81012141618202224

Time (min) Time (min) -2 Fig. 8: Debranched HWSS (blue) and pellet (red) of a non-waxy variety (a) and debranched flour of a waxy variety (b) run on the UH500 showing separation of amylose and amylopectin chains, and complete solubilisation of amylose

Figure 8a shows debranched HWS starch and HW insoluble starch of a non-waxy variety and Figure 8b shows debranched gelatinised flour of a waxy variety for reference for elution of amylopectin chains. Figure 8b, the waxy variety, shows that amylopectin chains are not separated on the column and that they elute after 18 minutes, after total permeation of the column. No chains were detected before 18 minutes. Amylopectin chains all eluted in the region where the two smallest pullulan standards eluted (Figure 6b) on the UH500. The debranched HWS starch (Figure 8a) contains many chains that eluted after 18 minutes, as well as chains that eluted between 10.2 and 18 minutes - the entire separation range of the UH 500, and over the same range as the largest six pullulan standards. However, coelution with pullulan standards does not mean that the amylose chains span the same MW range as the pullulan standards. The pullulan standards contain both α1-4 and α1,6 linkages whereas the amylose chains are linked only by α1, 4 bonds. Thus we can imagine that the amylose chains form much tighter coils than the pullulan chains, and so for the same hydrodynamic volume, it is likely that the MW of an amylose chain would be somewhat higher than that of a pullulan chain. The HW insoluble fraction did not contain any chains that eluted before 18 minutes and all chains in that fraction eluted after 18 minutes (Figure 8b).

-12-

Figure 9 shows the HWSS injected into the UH 500 before and after debranching. Very few chains eluted after 18 minutes 20 in the sample that was not debranched, but there was a large peak in the

15 exclusion volume of the column, before 10.5 minutes. The peak in the exclusion volume is material that could not 10 penetrate the pores of the column, so is, by definition, of very high molecular weight. Upon debranching of the 5 HWSS, the peak in the exclusion volume disappeared, and a new peak at the end of

0 the chromatogram appeared. Further, 810121416182022there is a significant difference in the Time (min) signal between 11 and 18 min between Fig. 9: HWS starch from a non-waxy variety (blue) and the HWSS and the debranched HWSS; debranched HWS starch from the same sample. the signal was lower for the debranched

HWSS.

Different speeds and different lengths of time of centrifugation were tested to attempt to minimise the contribution of amylopectin to the HWS fraction. Figure 10a shows HWS starch from a waxy variety at three different spinning speeds and Figure 10b shows HWS starch from a non-waxy rice. Speed of spinning or length of time of spinning did not give any difference in the size of the peak in the exclusion volume (Figure 10) or the amount of amylopectin in the soluble fraction of the waxy variety (Figure 10a).

Fig. 10: Three different centrifugation conditions did not remove or change the amount of the HWS amylopectin from either a waxy variety (a) or a non-waxy variety (b).

-13-

HWS starch for all further work was collected from flour using 250 mg flour in 25 ml water with heating and mixing in the RVA, followed by centrifugation at 10000 g.

Determining the number of molecules in the amylose of HWSS was attempted using several fluorophores and chromophores. We were unable to reproduce data shown in the literature using the same chromophores (2AP and APTS) for amylose labelling (data not shown). Further, 2AP and APTS are expensive, so very dilute samples were used to try to maximise the utility of the label. Chromatography was poor with these labels since the signal from a large peak of unattached fluorophore swamped the chromatogram. Several methods were used to clean up the sample before chromatography, and all were abandoned because the cost per sample was approaching $15 with label and clean-up. Another sulphonic acid, the chromophore ANDS, was discovered towards the end of the project. ANDS is negatively charged, contains sulphonic acid groups, so is suitable as a label to be attached to the reducing end of amylose chains using the same chemistry as the fluorophores. It required no clean-up and each sample cost only 4c to label. Efficiency of labelling across the MW range was measured by labelling the pullulan standards and analysing them with both RI and UV.

Figure 11 shows the chromatogram of these and shows that the UV trace is not equivalent to the RI trace. By using the simple relationship described earlier, the UV data can be mathematically transformed and plotted as if it was RI data. Figure 12 shows the replotted UV trace as RI data.

50 0.6

40 0.5

0.4 30

0.3 20 0.2

10 0.1

0 0 911131517911131517 Time Time Fig. 11: RI trace (a) of labelled pullulan standards, where the detector is mass- sensitive not label-sensitive, and (b) the UV trace of the same labelled standards where only the label, not mass, is detected.

-14-

5000

4000

3000

2000

1000

0 911131517

Time (min) Fig.12: Mathematical transformation of the UV trace in Figure 11b to make it equivalent to the RI trace in Figure 11 a.

Fig.13: Separation of HWS starch (a) from Amaroo, Millin and Koshihikari, and debranched HWS starch (b) showing slight differences between varieties.

Objective 2b: Amylose structure of three different varieties of the same amylose content

Figure 13 shows the SEC traces for HWS starch and debranched HWS starch (Figure 13b) for Millin, Amaroo and Koshihikari. There are differences in the distribution of the HWS starch and the MWD of the debranched amylose chains in the debranched starch. These differences were further investigated by fractionating the HWS starch. Figure 14 shows the iodine staining of each fraction (Amaroo). The intensity of the blue of the iodine staining shows clearly that fractions 7 – 24 contained amylose, and the green fractions 25 – 30 contained just a small amount of starch. Figure 15 shows the hydrodynamic volume distribution of the fractions of Amaroo. In Figure 15, fractions are grouped on the basis of elution pattern and show a group of early eluting fractions (red) and a group of fractions eluting in the middle of the run (blue), and a group that eluted towards the end of the run (green) molecules. For the

-15-

three varieties, each fraction was analysed in terms of its hydrodynamic volume and the proportion of linear chains. Figure 16 shows the hydrodynamic volume distribution of the representative fractions of the three varieties. Figure 17 shows the molecular weight distribution of the debranched chains of each of the fractions for the three varieties.

Fig.14: Iodine staining of fractions of HWS starch from Amaroo, each fraction collected for 2 min. Iodine binding, indicating the presence of amylose, commenced by fraction 7 and disappeared by fraction 28.

Figure 16 shows the hydrodynamic volume distribution for the fractions from each variety. For each group of fractions, early, mid and late, each variety shows a different elution profile. Basically there are two peaks: before 10 min (peak 1) and after 10 min (peak 2). The early fractions of Millin show much more starch in Peak 1 than the early fractions of the other two varieties. The early fractions (first column in Figure 16) of Amaroo show a wider distribution in Peak 2 than the other two varieties. The mid fractions (middle column Figure 16) of Koshihikari and Millin show a bimodal distribution for Peak 2, a shoulder at about 15 minutes, whereas the mid-fractions of Amaroo show a normal distribution for Peak 2 with the peak eluting between 10 and 20 min, 2 - 4 minutes longer than for the other two varieties. The late fractions (third column of Figure 16) of Amaroo and Millin show similar distributions, with most starch in the last shoulder of the peak (15 – 20 min), but those for Koshihikari showed least starch in the last shoulder of the peak (Figure 16, column 3).

20

AP& AM

15 • Long mostly linear

• Long mostly branched 10 • Short mostly linear

5

0 81012141618

Time (min) Fig.15: SEC traces of each fraction on the UH500 showing a group of high molecular weight chains (red), intermediate molecular weight chains (blue), and low molecular weight chains (green).

-16-

20 15 3 11 -14 20, 21 Amaroo 12, 14 Amaroo 22, 23 Amaroo 15 -18 24, 25 15 10 19 2

10

5 1

5

0 0 0 510152025 510152025 510152025

-5 -5 -1

45 12 3 6 11 19, 20 Millin 7 10 Millin 12, 13 Millin 21, 22 35 8 14-18 23, 24 9 8 2 25 6

15 4 1

2 5

0 0 -5 510152025 510152025 510152025

20 10 3 Koshi Koshi Koshi 8 15 2

6 10 1 4

5 2 0 510152025 0 0 510152025 510152025-1

Fig.16 showing the SEC traces of each group of fractions for each variety on the UH500. High molecular weight fractions are column 1, intermediate molecular weight fractions are column 2 and low molecular weight fractions are column 3. Comparison between Figures 16 (HWS) and 17 (debranched) shows that for every fraction, for each variety, the starch eluting before 10 min, Peak 1, in Fig. 16 was always almost fully eliminated after debranching with isoamylase, and a new, very big peak begins at about 17 min for each (Figure 17). The large peak beginning at 17 min (Figure 17) is not shown within the scale of the figure. Figure 17 shows shows that for the early fractions, peak 2 elutes between 10 and 15 minutes, for the mid fractions, closer to 15 minutes and for the late fractions, after 15 minutes. Figure 17 shows that Peak 2 for Millin is twice as large (compare y axis) as it is for the other two varieties for the early, mid and late fractions.

Figure 18 shows the grains of the three varieties, and shows only small differences between the three, that are probably not distinguishable.

Individual amylose chains were not labelled successfully during the time that this portion of the work was carried out.

-17-

20 5 3 Amaroo Amaroo Amaroo

15 4 2 3 10

2 1 5 1

0 0 510152025 5101520250 510152025

-5 -1 -1

60 10 5 Millin Millin Millin 9 50 8 4

7 40 6 3

30 5 4 2 20 3

2 1 10 1

0 0 0 510152025510152025510152025

10 4 10 Koshi Koshi Koshi

8 8 3

6 6 2

4 4

1 2 2

0 0 0 510152025510152025510152025

Fig.17 showing the SEC traces of each group of fractions for each variety on the UH500 after debranching. High molecular weight fractions are column 1, intermediate molecular weight fractions are column 2 and low molecular weight fractions are column 3.

Fig.18: Grains of Koshihikari, Millin and Amaroo showing the K M A similiarity of each.

See Part B – commencing Objective 3: Resistant Starch

-18- Part B: Final Report Project 3402

Objective 3: Resistant Starch (RS)

This is the first report of the use of the Megazyme method for the analysis of resistant starch (RS) in rice, so investigations were carried out to validate the method before it was used for routine determinations of RS. Figure 3 shows the method. The first incubation produces supernatant 1 which contains the digested and non-resistant starch (NRS), and pellet 1 which contains the RS. The second step produces supernatant 2 which contains the resistant starch and pellet 2, which is discarded. Table 3 shows the amount of NRS (supernatant 1 (discarded), RS (supernatant 2) and the amount of amylose in pellet 2 of the Australian rices and of one from the Philippines (PSBRc 18). When protein and moisture is accounted for, the total of the starches in the different fractions almost accounts for the starch in the milled rice.

Variety RS NRS AM P2 TOTAL

Amaroo 1.26 67.8 1.50 70.6 Basmati 2.67 69.5 1.64 73.8 Doongara 2.15 70.6 1.36 74.1 Koshihikari 1.30 66.9 2.66 70.9 Koshi Japan 0.97 71.8 1.91 74.7

Kyeema 0.89 62.2 1.35 64.4 Langi 1.24 66.0 1.93 69.17

Opus 1.17 73.1 1.17 75.44 1.88 69.2 1.83 72.91 PSB Rc10

Table 3. Resistant starch (RS) non-resistant starch (NRS), amylose in pellet 2, and total starch. Allowing for moisture and protein, the Megazyme method accounts for almost all of the starch in a milled grain.

RS from freshly cooked rice

Table 4 shows the varieties, their origin, their amylose content, the amylose allele for each and the RS content of freshly cooked rice of each. Figure 19 shows the relationship between RS in freshly cooked rice and amylose content (Figure 19a), amylose allele (Figure 19b), and proportion of long chains (Figure 19c). The figures show some relationship to amylose content with a correlation coefficient of about 0.7, and to long chains, but little relationship with the amylose allele.

Figure 20 shows the HWS starch of a selection of varieties where RS in freshly cooked rice is low (red), intermediate (blue) or high (green). There is no clear relationship between the amount of HWS starch and the amount of RS in freshly cooked rice. Figure 21 shows chromatograms of the same varieties after debranching and again shows very little difference in MWD of the chains and RS content. These results were unexpected so the structure of the molecule that accounts for the RS reading was investigated.

-19- Figure 22 shows HWS starch from Doongara and the molecule that accounts for the reading of RS. Clearly, the molecule that accounts for the RS reading is very low in hydrodynamic volume, and there is only a very small amount relative to the HWS starch.

Variety Origin Amylose (%) CTn of GBSS RS (%)

Amaroo Sunrice, Australia 19 19 1.26 Basmati Sunrice, Australia 24 17 2.67 Doongara Sunrice, Australia 25 14 2.15 Koshihikari Sunrice, Australia 19 17 1.30 Koshihikari- Sunrice, Australia 18 17 0.97 Kyeema Sunrice, Australia 20 18 0.89 Langi Sunrice, Australia 19 19 1.24 Opus Sunrice, Australia 19 17 1.17 IR60 IRRI 26 10 2.99 IR8 IRRI 26 11 3.40 IR5 IRRI 27 17 3.38

IR64 IRRI 23 17 2.54 IR24 IRRI 18 18 0.83 AE IRRI 34 11 6.24 PSB Rc10 PhilRice,Philippines 25 11 1.88 PSB Rc98 PhilRice,Philippines 30 11 3.09 PSB Rc12 PhilRice,Philippines 24 20 2.47 PSB Rc16 PhilRice,Philippines 28 20 2.57 PSB Rc18 PhilRice,Philippines 23 20 1.84 Dawn YAI 22 14 1.60 Newbonnet USDA 22 14 1.80 Tebonnet USDA 22 14 2.00 Kaybonnet USDA 23 14 2.00 Jodon USDA 26 20 2.40 Rexmont USDA 24 11 2.70 Dixiebelle USDA 27 11 2.70

Bolivar USDA 25 11 2.80 Cocoderie USDA 26 20 2.90 L205 USDA 25 11 3.00 TX3042 USDA 27 11 3.20 Te Qing USDA 26 11 3.30 TX 4175 USDA 27 11 3.60 Sierra USDA 25 11 3.70

Table 4: Varieties used, their origin , amylose content, CT number of amylose allele, and RS content of freshly cooked rice.

-20-

33 C 25 16 31 ) 14 %

( 29 20 2 R = 0.3098 12 27 25 15 10 23 8 CT R2 = 0.6279 21 10 6 lo se co n ten t y 19 4

5 (%) chains Long Am 17 2 15 0 0 02468 02468 02468 Resistant starch freshly cooked (%) Resistant starch freshly cooked (%) Resistant starch freshly cooked (%)

Fig.19: RS in freshly cooked rice correlates reasonably with amylose content, poorly with CT and reasonably with % long chains ( ie chains > DP100).

15

13

11

9

7

5

Detector Response Detector 3

1

-1 810121416182022

Time (mins) Figure 20: HWSS starch of selected varieties with high RS (green), intermediate RS (blue) and low RS (red), showing no relationship between HWSS and RS.

-21-

20

18

16

14 12

10 8

RI response (MV) response RI 6

4

2

0 810121416182022 Elution time (min)

Figure 21: Debranched HWSS starch of selected varieties with high RS (green), intermediate RS (blue) and low RS (red), showing no relationship between MWD of long chains and RS.

10

5

(mV) response RI

0 8121620

Time (min)

Figure 22: Amount and elution profile of HWSS of Doongara and the of molecule that accounts for the RS reading from Doongara

-22- The data in the previous figures suggests that the molecule that accounts for the reading of RS is much smaller than amylose molecules. This was not expected. A selection of varieties was made from the set of Australian and Asian rices, spanning the range of RS, to explore the structure of RS further. Figure 22 suggests that RS molecules are small and so they were re- analysed using the UH 250 column, which separates smaller molecules better than the UH 500. Figure 23 shows the chromatograms of whole RS molecules on the UH 250 column for the sub-set of low (red), intermediate (blue) and high (green).

4

3

2

1

0 910111213141516171819

Time (min) -1

Figure 23: Whole molecules of RS of Australian and Asian varieties showing good grouping of high (blue), intermediate (green) and low (red) RS measured by the Megazyme method. Separation is SEC on the UH250.

Figure 23 shows that the high, intermediate and low RS values measured by the Megazyme assay relate well to the relative amounts of RS measured as detector response on the SEC. Figure 24 is a typical chromatogram of full flour showing the separation of amylose and amylopectin chains, for reference when comparing the elution time of RS molecule relative to the amylose and amylopectin chains. Note though, the RS molecule is a complete molecule, so its elution pattern relates to its hydrodynamic volume (size in the eluant) and not to its molecular weight, whereas the amylose and amylopectin chains in Figure 24 can be interpreted in terms of the MWD. Amylose chains elute between 11 and 12 minutes, and amylopectin chains elute between 15 and 19 minutes. Figure 23 shows that RS molecules elute between 14 and 19 minutes, so the size of the molecule in the eluent is the same as the size of the linear chains of amylopectin that elute between 14 and 19 min. There is a small amount of material that elutes at the same time as the chains of amylose elute (by reference to Figure 24).

-23-

35

30

25

20

15

10 Detector Response

5

0 10 12 14 16 18 20 Time(mins)

Figure 24: Typical chromatogram of debranched flour on the UH250 showing amylose chains eluting between 9 and 14 min and amylopectin chains eluting between 15 and 18 min.

The UH250 column has recently been calibrated for the analysis of the absolute MWD of linear starch chains. In order to determine the nature of the RS molecule, its MW was analysed using a light scattering detector (DAWN) and compared with the expected MW for that elution time, for linear chains. Figure 25 shows the expected MW and the actual MW of the RS at each elution time. Figure 25 shows that the MW of linear chains is proportional to elution time, and shows that at the elution time of RS, the molecular weight of a linear chain would be of the order of 1000 - 10000, whereas all the samples presented show that the average molecular weight of the molecules at each elution time for all the samples of RS is higher in MW than a linear chain at the corresponding elution time, and furthermore, there is some variation in the molecular weight of the RS molecules.

1.E+07

1.E+06

1.E+05

1.E+04 Log Molar Mass Molar Log

1.E+03

1.E+02 10 12 14 16 18 Elu t io n t im e ( m in ) Figure 25: Expected MW vs Elution time for linear chains fo starch (Castro et al. 2005) (blue diamonds) and actual MW of molecules of RS from different varieties (colours).

-24-

In order to test whether or not the structure of RS resembled digested amylose or digested amylopectin, the RS molecules were debranched with isoamylase and the debranched RS analysed on the UH250 (Figure 26).

5 a 1.0 b

4

3 0.5

2

1

0.0 0 8101214161820 8101214161820

-1

Fig. 26: Whole molecules of RS (a) from freshly cooked rice of Australian (red) and Asian (blue) origin, and (b) debranched RS from the same varieties. Note the difference in the scales.

Figure 26 shows chromatograms of the whole (Figure 26a) and debranched (Figure 26b) RS for the Asian and Australian varieties. Figure 26b shows that the debranched molecules are of similar hydrodynamic volume to the whole RS molecules, and for all, there is a significant increase in the proportion of very short chains in the debranched RS. Further, there is a very small peak eluting in the area of amylose. The short chains of the RS molecules (debranched) were analysed further with CE and the amylose-like molecules by SEC.

6 0.3 a b

5

4 0.2

3

2 0.1

1

0 0 10 11 12 13 14 10 11 12 13 14

Figure 27 showing (a) amylose chains from flour of 3 high, 2 intermediate and 6 low amylose varieties and (b) showing the amylose in the RS from the same varieties. Note, the scale of b is 1/20th of a..

-25-

Figures 27a and b show the amylose chains in debranched starch and in RS for the varieties. Concentration is taken into account and the amount of chains in the debranched RS and full flour can be directly compared. Note the differences in the scale of the y axis. For all the varieties, very little amylose remains in the RS relative to that in the full flour and the amount of amylose remaining in the RS is independent of the amount in the full flour. The remaining amylose is between 1 and 2% of the original. Figure 28 shows a CE trace of the chain length distribution (debranched) of amylopectin of Doongara for comparison with the chain length distribution of the debranched RS from the Australian varieties (Figure 29) (this project is still current so data is unavailable for the Asian rices).

8

6

4 Mol%

2

0 1591317212529333741454953576165697377818589

Fig. 28: CE trace of debranched flour showingDP baseline separation for each chain length (DP). The internal standard is maltose (DP 2) and is shown in pink.

-26- 2 2 Doongar a Bas m ati

1.5 1.5

1 1

Mol% Mol%

0.5 0.5

0 0 19172533414957657381 DP 191725334149576573 DP

Kos hihik ar i NSW 2 2 Kos hihik ar i Japan Koshihikari NSW

1.5 1.5

1 1 Mol% Mol% 0.5 0.5

0 0 171319253137434955616773 1917253341495765 DP DP

2 Kye e m a 2 Opus

1.5 1.5

1 1 Mol % Mol %

0.5 0.5

0 0 1611162126313641465156 191725334149576573 DP DP

Figure 29: Chain length distribution of molecules of RS from Australian rices.

Figure 28 shows two populations of chains in amylopectin of full flour, one population between DP 6 and DP 33 and the other between DP 38 and DP 61. For each variety, there are clear differences between the chain length distribution of amylopectin and that of RS. Figure 29 shows that there are far fewer chains in the RS than in the amylopectin (amounts can be directly compared), and for all, there are far fewer chains above about DP 25, and there are no chains beyond about DP 40. Further, the chain length distribution of all the RS samples shows a much higher proportion of chains of DP 2 - 6 than the chain length distribution of amylopectin shows and the RS samples contain large peaks of glucose and maltose (DP 1 and 2), which the amylopectin does not show . Figure 29 also shows that there is not a consistent chain length distribution for RS from the different varieties.

-27- Resistant starch in processed rice

Analysis of the effect of different cooking and processing techniques was carried out using the Australian variety Doongara. The structure of RS in retrograded rice was analysed using a range of Australian and Asian rices.

Resistant starch was measured on freshly cooked samples of both brown and milled rice of Doongara, and was 3.7% in freshly cooked brown rice and 2.1% in freshly cooked milled rice.

The effect of different cooking times and methods was compared using milled rice of Doongara. Different cooking times showed little difference in resistant starch content. When rice was under-cooked, RS was 1.9%, when it was over-cooked, RS was 2.2%. Rice cooked by the absorption method retained more resistant starch than rice cooked by the rapid-boil method (2.1% versus 1.3% respectively). The cooking water from the rapid boil method was retained and the water evaporated. The RS content of the dried slurry from the rapid boil methods was 5.3%. Cooking rice and then mashing or sieving it did not alter the RS content relative to freshly cooked milled rice.

The white rice was ground to a flour and gels prepared (8% flour). RS was determined on fresh gels and on gels that were stored for a week at 4 ºC. The RS content of fresh gels was 0.1% and of stored gels was 0.5%.

The effects of retrogradation were observed when cooked rice was chilled for 24 hours. The RS value increased significantly after 24 hours at 4°C (from 2.2% to 3.6%), but the value did not change with storage beyond 7 days.

The structure of RS in retrograded rice

RS in retrograded rice was explored further using the set of Australian and Asian rices. Table 5 shows the RS content of retrograded rice for the set of Australian and Asian rices.

Variety Origin Amylose (%) CTn of GBSS RS (%) Amaroo Sunrice, Australia 19 19 1.67 Basmati Sunrice, Australia 24 17 3.70 Doongara Sunrice, Australia 25 14 3.17 Koshihikari Sunrice, Australia 19 17 1.73 Koshi Japan Sunrice, Australia 18 17 0.92 Kyeema Sunrice, Australia 20 18 1.60 Sunrice, Australia 19 19 1.50 Langi Opus Sunrice, Australia 19 17 1.41 IR60 IRRI 26 10 4.50 IR8 IRRI 26 11 3.59 IR5 IRRI 27 17 4.85 IR64 IRRI 23 17 2.57 IR24 IRRI 18 18 1.20 AE IRRI 34 11 11.97 PSB Rc10 PhilRice,Philippines 25 11 4.02 PSB Rc98 PhilRice,Philippines 30 11 5.90 PSB Rc12 PhilRice,Philippines 24 20 2.63 PSB Rc16 PhilRice,Philippines 28 20 4.07 PSB Rc18 PhilRice,Philippines 23 20 3.92

Table 5: Varieties used, their origin, amylose content, CT number of amylose allele, and RS content of retrograded rice.

-28-

Figure 30 shows the relationship between RS in retrograded rice and amylose content (Figure 30a), RS in retrograded rice and long chains of starch (Figure 30b), RS in retrograded rice and amylose allele (Figure 30c), and RS in retrograded and freshly cooked rice (Figure 29d). Amylose content correlates better with RS in retrograded rice (Figure 30a) than it does with RS in freshly cooked rice (Figure 19a), however, in a contradictory sense, the proportion of long chains shows no correlation with RS in retrograded rice (Figure 30b) or with amylose allele (Figure 30d). However, the amount of RS in freshly cooked rice correlates nicely with the amount of RS in retrograded rice (Figure 30c).

35 a 16 b a b 30 14 12 25 10 20 R2 = 0.7466 8 15 6

10 %Long Chains content(%) Amylose 4 5 2

0 0 02468101214 02468101214 Resistant Starch Retrograded (%) Resistant Starch Retrograded (%)

b C 7 c 22 d 21 6 20 19 5 18 17 4 16 15 2 R2 = 0.8758 CT R = 0.3128 3 14 13 2 12 11 1 10 9 Resistant Starch freshly cooked( %) cooked( freshly Starch Resistant 0 8 02468101214 02468101214 Resistant Starch Retrograded (%) Resistant Starch Retrograded (%)

Fig.30: Correlation between RS from retrograded rice and amylose content (a), % long chains (b), resistant starch from freshly cooked rice (c) and CT number of GBSS1 (d).

The structure of RS in retrograded rice was determined by SEC of the RS molecules and of the debranched molecule. Figure 31a shows the chromatograms of the whole RS molecules from retrograded rice, and Figure 31b shows the chromatograms of the debranched RS from retrograded rice. Figure 31a shows that no chains elute where amylose chains elute, but the molecules elute in the region of amylopectin chains. Debranched RS from retrograded rice

-29- shows that the molecules consist of chains of the order of amylopectin. The ae variety shows more RS when retrograded and its chains differ in distribution from the wild-type rices and Koshihikari grown in Australia shows much more RS when retrograded than does Koshihikari grown in Japan.

5 5 a b 4 4

3 3

2 2

1 1

0 0 10 12 14 16 18 20 10 12 14 16 18 20

-1 -1

Figure 31: showing the SEC traces of RS from retrograded rice from Australian (red) and Asian rices (blue). Australian Koshi is green and Japanese Koshi is pink, and (b) debranched RS from retrograded rice. The trace from the ae is in grey in (b) and the highest in (a). Amylose elutes around 12 min

Figure 32 amalgamates the data from Figure 31b and from Figure 26b and shows the debranched RS molecules from freshly cooked rice and from retrograded rice. Figure 32 shows that there is more RS in retrograded rice, the RS molecules from retrograded rice elute earlier than those from freshly cooked rice, there is no difference between the amount of amylose chains (eluting at 12 min) in RS from freshly cooked and from retrograded, but there is a peak representing short chains (eluting at 17.5 – 18 min) in the RS molecules from all the retrograded samples but there is not a corresponding peak in the RS molecules from freshly cooked rice. Studies are ongoing investigating the chain length distribution of RS in retrograded rice using CE.

3

2

1

0 14 15 16 17 18 19 Figure 32: Debranched RS from Australian and Asian rices from freshly cooked (blue) and retrograded (red), showing a different distribution of chains that comprise each type of RS.

See Part C – Discussion etc.

-30- Part C: Final Report Project 3402

7. Discussion

Amylose content is one of the most important factors affecting the cooking and sensory properties of rice. However, rice quality focuses mostly on amylose content and has not addressed amylose structure. It is well known that the cooking properties of different rices that have the same amylose content, can differ enormously1-4. An example of this in the Australian program are the three varieties Amaroo, Koshihikari and Millin. These have exactly the same amylose content but very noticeable differences in cooking properties, especially to Japanese palates. Amylose structure has been cited as the cause of differences between varieties of different sensory properties but identical amylose content5-10, but no definitive answer is available. Here, we attempt to develop a method to investigate amylose structure using those three varieties of rice, Amaroo, Millin and Koshihikari and then to use that information to investigate the relationship between amylose structure and cooking properties. The cooking property chosen was resistant starch. Resistant starch is thought to be related to amylose content much more strongly than to amylopectin11 because of the rapid way that amylose chains aggregate and resist attack by digestive enzymes. Further, as the world becomes familiar with nutritional requirements other than minerals and vitamins, and the requirement for starch in our diets for satiety and health12-15, research on resistant starch is timely. The ultimate aim of this project is to deliver the outcomes in a form that that will benefit the rice improvement program.

Objective 1: Searching for allelic variation in the GBSS gene

The first objective towards finding differences in amylose structure was to seek differences in the sequence of the gene considered responsible for amylose synthesis, namely granule bound starch synthase (GBSS1). Sequence variation was sought by determining the sequence of the 8 exons of GBSS1. No differences were found during this project. However, simultaneously, a group in the US published two genetic differences in GBSS that explained more than 80% of the variation in amylose content of rice16. The first of these was a single nucleotide polymorphism at the splice site of intron 1. This SNP had been reported previously17, but the functional significance and relationship with amylose content was reported by Ayres et al. (1997). The particular base determined whether the exon spliced correctly, which determined how much GBSS protein was translated. Many workers have concluded that this SNP defines the Wxa and Wxb allele18-22. The Wxa allele is found in the indica and tropical japonica varieties and the Wxb allele is found in temperate varieties. Amaroo, Koshihikari and Millin are all varieties with the temperate japonica class, so there would be no sequence difference at the splice site of intron 1 between these varieties. The second genetic variation was found in the flanking region of the gene. It is a simple sequence repeat (SSR) of alternating C and T 16, 23. The number of dinucleotide repeats related well with amylose content. In the work reported here, the flanking region was not sequenced, thus the SSR was not detected. However, collaboration was generated with the American group, and the SSR was determined for a sub-set of the Australian collection. This is shown in Figure 4. Interestingly, the three varieties chosen all carry different CTn. Koshihikari is CT17, Millin is CT18 and Amaroo is CT 19. This was the only sequence variation found within the GBSS gene for these 3 varieties.

-31- Amylose molecules are lightly branched24, 25, thus a branching process is required to obtain that structure. Sequence variation was sought in Branching Enzyme 1 since several studies implicate BE1 in the synthesis of amylose. The earliest studies linking BE1 and GBSS are in vitro studies showing that BE I is much more likely to branch long chains like amylose rather than short chains like amylopectin 25, 26. More recent studies using mutants show that the absence of BE1 does not affect the chain length distribution of amylopectin in rice, maize or wheat27-29, suggesting either that BE1 is not involved in amylopectin synthesis, or that the other isoforms of BE can carry out the function of BE1 in its absence. BE1 has been linked genetically to GBSS30 by the discovery of some interaction between the transcription of BE1 and GBSS, leading to the proposal that the expression of the BE1 and GBSS is coordinated30. These studies all suggest that BE1 does not play a crucial role in amylopectin synthesis, but that its kinetics and genetics could link it with the synthesis of amylose. However, no variation was found among the exons of that gene for the three varieties.

Thus, the outcome of the first objective was finding the variation in the SSR for the three varieties, as well as the rest of the Australian germplasm, in collaboration with Drs Bill Park and Nicola Ayres at Texas A & M University, College Station, Texas. The microsatellite is now being used successfully in the Australian Rice Improvement program.

Objective 2: A method to measure amylose structure

In order to try to relate the variation in SSR with amylose structure, and in order to obtain some way to obtain information on the structure of amylose, a method to measure amylose structure was developed by using tools of chromatography. Size exclusion chromatography (SEC) is commonly used to analyse two properties of polymers. These properties are the hydrodynamic volume of the molecules in particular eluants and the molecular weight distributions of linear chains using standards of the same chemical structure as the polymer, or by using the principle of universal calibration to calibrate the column by using Mark Houwink parameters to relate the elution profiles of standards of a similar composition to the polymer of interest to the elution profile of the polymer of interest31.

Many groups use SEC with pullulan standards to measure the ‘MWD’ of amylose. However, pullulan standards are not the same composition of starch, so their configuration in the eluant differs from that of a linear chain of starch, and consequently, the elution profile of a pullulan standard significantly under-estimates the MW of the starch molecule at the same elution volume31. This principle is not widely recognised in the field of starch chemistry, but it was observed in this work.

The first step towards measuring amylose structure is to separate the amylose from the amylopectin. It is well known that amylose exists as a non-crystalline polymer within the starch granule, and upon heating in water, it leaches, possibly though channels32, from the granule to join the liquid phase of the cooking pot33-35. This is the first study where the amylose was leached directly from flour, rather than after purifying the starch. Purification of rice starch is not trivial, and the most commonly used method is the alkaline precipitation method. However, other studies in the laboratory have found that the alkaline precipitation method damages the long chains of amylose36. Moreover, more recent work in our laboratory shows that purification of starch by any method increases the solubility of amylopectin (Willoughy unpublished), thus separating amylose from flour should minimise the contribution from amylopectin. Figure 5

-32- shows that amylose begins to leach above 90ºC, which is well above the gelatinisation temperature, thus other studies, relying on gelatinisation temperature, are not likely to collect all the amylose into the hot water soluble fraction. Further, the absorbance of the leached components of the intermediate amylose variety was higher than of the three low amylose varieties. The standard technique of Rapid Visco Analysis easily heats to that temperature, so the RVA was used to give a consistent set of heating/cooking conditions in order to investigate reproducibility of leaching of amylose without shear. The method was very reproducible, both within days and for different days. However, Figure 5 also shows that some leaching occurs from a waxy variety. This material could either be fragile outer layers of the starch granule that begin to dissolve37, or it could be less crystalline molecules of amylopectin that dissolve more readily than semi-crystalline lamellae. The material in the supernatant, the hot water soluble (HWS) fraction, was analysed to determine the proportion of amylopectin molecules in it.

Two SEC columns were tested for measuring amylose and amylopectin molecules and the chains that comprise the molecules. Pullulan standards underestimate the MW of amylose (Ward et al. submitted), but they provide a broad guide to the separation capacity of the columns. Through the use of pullulan standards (Figures 6 and 7), the UH 250 was chosen for separating the chains of amylose from the chains of amylopectin, though this column required that the starch was debranched to linear chains38, and the UH 500 was chosen for separating molecules in the HWS fraction of the starch (obtained from the modified RVA method). Thus the contribution of amylopectin to the HWS fraction could be determined.

Figure 8b shows that a waxy variety (amylopectin only) contains no chains in the HWS fraction that elute before 18 min, thus any chains eluting before 17.5 min on the UH500 could be attributed to amylose. Figure 8a shows the chains in the supernatant and the pellet of a non-waxy variety and shows that all the long chains (eluting before 18 min) are in the supernatant, and none are retained in the pellet. In other studies where a supernatant was obtained by leaching33-35, some of the amylose always remained in the pellet.

In this study, all the amylose (long chains) was recovered in the HWS fraction, however, Figures 8b and 9 show that amylopectin contributes to the HWS fraction. Figure 8b, shows this conclusively since the figure shows the starch (debranched) in the HWS fraction of a waxy rice. Figure 9 shows the distribution of the whole molecules within the HWS fraction and of the chains (after debranching). The trace of the whole molecules shows a large peak in the exclusion zone (before 11 min), indicating elution of very large molecules. The separated linear chains do not show the large peak in the exclusion zone, and show a large peak beyond 18 min instead. This indicates that the process of debranching effectively deconstructs the large molecules in the exclusion zone to their constituent small branches, which elute beyond 18 minutes. Such large molecules made of small branches are highly likely to be amylopectin molecules, thus it must be concluded that amylopectin contributes to the HWS fraction. Furthermore, many amylopectin molecules elute in the separating phase of the column, along with the amylose molecules. These amylopectin molecules must be much smaller than those that elute in the exclusion zone. Different speeds and times of centrifugation did not cause the amylopectin molecules to precipitate (Figure 10), thus it was not possible to exclude amylopectin from the fraction without altering the fraction. However, the reproducibility of the method indicates that the contribution of amylopectin to the HWS fraction is not random, but depends on the variety. In terms of processing, cooking and sensory properties, the solubility of starch is likely to be a key

-33- parameter, thus a contribution from amylopectin to the HWS fraction was accepted for the remainder of the work, and the amount was accounted for in any interpretation.

A deeper understanding of amylose structure can be gained by knowing the branching frequency, the length of the backbones and branches, and the number of amylose molecules. Snippets of this information can be obtained by labelling of the amylose molecule and obtaining the number average and the weight average degree of polymerisation, and then obtaining the same information for the amylose chains. Two conditions must be met in order to draw valid conclusions from such a study, (i) each amylose molecule or chain must be labelled and visible to the detector, and (ii) each chain must carry only one label.

Reductive amination of the reducing end of starch molecules has been used previously to introduce a label onto a starch chain8, 39-42. The chemistry of this method is well established and involves attaching a primary amine, carrying a fluorophore, onto the reducing end of the starch molecule. There is only one reducing end on each starch molecule, thus only one label is introduced per molecule and the first condition is met. Efficiency of labelling across a range of chain lengths has been tested up to DP 135 for one of the fluorophores (APTS)41, and in that study, efficiency decreased slightly, although reproducibly, as the chain length increased. That fluorophore has been optimised for fluorophore assisted capillary electrophoresis, but the method is useful only for amylopectin chains. In another study, starch was chemically purified, and amylose collected from that and labelled with 2 AP24. Labelling efficiency was tested and was reasonable. However, all our attempts to use that technique on amylose were unsuccessful. Attempts to use other fluorophores were made, but the cost, time and irreproducibility rendered the technique impractical. The major difference between the work here and the other study24, is that we used amylose that had not been collected by passage through a number of chemical reactions. The methods used in that study to purify starch have been found to alter the structure and the solubility of amylose36; this could explain why we were unsuccessful with the technique.

Late in the term of this project, labelling of amylose structure was revisited. A new label was discovered - a chromophore with simpler labelling, better chemistry and lower cost relative to the fluorescent options. Since the chemistry of the new label – ANDS – is essentially based on the same principles as the fluorescent labels described earlier, in that only the reducing end is labelled, we can be sure that each chain is labelled only once. In order to measure labelling efficiency, we can take advantage of pullulan standards, of known molecular weight, and the simple mathematical relationship between the response of the UV detector and the response of the Refractive Index (RI) detector31. The UV detector is a number sensitive detector, and the signal represents P(M), ie the proportion (P) of chains of a particular molecular weight (M), but the RI detector is mass-sensitive, and for most polymers, the signal represents M 2 P(M)31. The two signals can therefore be related by mathematically transforming the signal from one detector to the other so long as the molecular weight of the polymer is known. Figure 11 shows the pullulan standards after labelling with ANDS, and shows the RI and UV signals. The UV signal certainly decreases as the molecular weight increases (high molecular weight elutes earlier that low) (Figure 11). The same mass of each standard was labelled, thus for the higher molecular weight pullulans, the number of molecules is fewer than for lower molecular weight pullulans. If labelling efficiency is independent of molecular weight, we expect that the transformed UV signal will be equivalent to the RI signal. If labelling efficiency decreases as molecular weight increases, we expect the transformed UV signal to be lower than the actual RI signal. Figure 12 shows the UV trace expressed as a signal from the RI, and shows that the signals are essentially

-34- equivalent. Thus we can be confident that labelling efficiency is maintained up to a molecular weight of about 800 000.

Now that separation and measurement of amylose is validated and understood, we can investigate the structure of amylose in the three varieties, Amaroo, Millin and Koshihikari. Figure 13 shows the elution profile of the hot water soluble fraction of each. There is very little difference between the varieties, but as shown earlier, a significant amount of amylopectin co-elutes with the amylose, obscuring any differences between the varieties for the distribution of amylose. In SEC, the elution time of a molecule does not depend on its molecular weight, but on its hydrodynamic volume. This means that a long amylose molecule and a small amylopectin molecule could occupy the same volume in the eluent, and thus elute at the same time. In order to determine the amount of amylose and amylopectin in each fraction, a long SEC column was used to fractionate the samples. Figure 14 shows the colour of each fraction when mixed with iodine and shows which fractions contain amylose. Most fractions show blue staining, which is the expected colour of the amylose iodine complex (figure 14). Figure 15 shows the separation of each fraction on the UH 500, and shows that there are basically three groups. The earliest fractions have most starch eluting in the exclusion zone (before 10 min), and in the separating phase of the column, the starch in the earliest fraction elutes at about 12 min. The middle fractions also contain starch in the exclusion volume, but in the separating phase, the peak is at about 14 min, representing decrease in size of the molecules. The latest fractions also contain starch that elutes in the exclusion zone, and the molecules that elute in the separating phase are smaller again. Figure 15 shows the elution profiles of the fractions for each variety.

Figure 15 shows that the early fractions of Millin contain much more starch than the early fractions of the other two varieties (note the scale), and for Millin, most of it elutes in the void volume. Comparison with the elution profile of the debranched fractions (Figure 16), most of the starch in the exclusion zone is amylopectin. The starch that elutes in the separating phase of the column is long linear chains. Millin contains more long linear chains of amylose than do the other two varieties. For the intermediate fractions, Figure 16 shows that for Millin and Koshihikari, there are 3 populations: the exclusion zone (amylopectin), and two peaks in the separating phase. However, upon debranching there is a normal distribution of chains. The intermediate fractions contain chains of lower molecular weight than the early fractions, and again, Millin has the most. However, for the last few fractions, with shorter linear chains (Figure 17), Koshihikari seems to have the most. Recently, the UH 500 column was calibrated for determining the molecular weight distribution of linear chains of amylose (Ward et al.). Extrapolating from that work, for the chains in the early fraction (Figure 17) the peak maximum is about DP 5000, in the intermediate fractions the peak maximum is about DP 2000 and in the late fractions they are about DP 500.

The differences in distribution of amylose chains amongst the varieties suggest different functional properties of the rice. During cooking, the starch granule gelatinizes and the amylose leaches out of the granules35, 43 into a continuous phase44. For a variety like Millin, which has twice as many long chains as Koshihikari (Figure 17), the concentration of the continuous phase would be much higher than it is for Koshihikari, and the proportion of long chains in that phase would be higher than for Koshihikari. Aqueous solutions of amylose (as would be found in cooked rice) are inherently unstable, and thus the amylose is prone to forming a viscoelastic paste or a gel45. Whether a paste forms or a gel forms relates to the concentration of the amylose in the solution and the chain length of the amylose45. The rate of aggregation of amylose, is strongly

-35- related to chain length in that chains of DP 80 aggregate rapidly and chains of DP 2000 aggregate slowly, and the structure of the resultant aggregate differs45. In a study using synthetic amylose, it was found that for long chains of amylose, gelation was favoured over precipitation, and on standing, the gels exuded some water. For short chains (DP 500), some gelation occurred at higher concentrations, but precipitation predominated45. Thus, the higher concentration of long chains in the amylose of Millin would suggest that the amylose of Millin is more prone to forming a gel than the amylose of Koshihikari, and on cooling, the gel would harden for Millin. The amylose in Koshihikari would precipitate, leaving essentially a visco-elastic paste, that would retrograde very slowly46. Thus, investigating amylose structure can provide a significantly better insight into cooking properties than can the measure of amylose content. Figure 18 shows that the three rice varieties are similar on the outside, but clearly they differ on the inside.

Objective 3: Relationship between amylose and resistant starch

Resistant starch is that portion of the starch that is not digested in the gut and moves through the digestive tract to the large intestine47. It acts in a similar matter to dietary fibre and ferments in the colon. It is believed to contribute to good bowel health, acting as a substrate for colonic bacteria to produce butyrate. Butyrate and other short-chain fatty acids, such as propionate may reduce the risk of colorectal cancer and aid in management of inflammatory bowel conditions48. Several studies show that resistant starch can act as a prebiotic, an agent that can carry bacteria, such as lactobacillus, to the colon, enabling this beneficial bacteria to colonise the colon more easily (refs), and assist in the treatment of conditions, such as infectious diaorrhea12, 15, 49-53.

There are four types of resistant starch54: Type 1 is physically unavailable for digestion, such as brown rice where the bran layer might restrict access for digestive enzymes; Type 2 is naturally- occurring resistant starch granules, for example in green bananas,; Type 3 is retrograded starch which becomes difficult to digest because of the crystals formed; and Type 4 is chemically modified.

Resistant starch (RS) is rather difficult to measure. However, in vitro methods to measure resistant starch have recently been developed55. The method developed by McCleary et al. (2002) mimics the processes of digestion in vivo. Figure 3 shows the method, and it was tested with a number of varieties of rice, ranging in amylose content, to ensure that all the starch is accounted for. Table 3 shows that all the starch was accounted for, when moisture content is factored in. Despite the fact that all the starch is accounted for, it must be stressed that this is an in vitro method, using a limited number of enzymes, in conditions that probably are not the same as those in the gut. In this work, we determine the structure of resistant starch from freshly cooked rice and from retrograded rice.

The propensity of amylose to form gels and aggregations has led to a general assumption that amylose relates strongly to resistant starch, and a previous measurement of RS from rice also suggest that the molecule is amylose56. Table 4 shows the varieties used, their origin, the amylose content, CTn and RS content. The lowest RS is 0.9% and the highest, excluding the amylose extender, is 3.7%. Figure 19 shows that RS does relate reasonably well to amylose content and reasonably well to the proportion of long chains (as measured by SEC). The hot water soluble fraction contains all the amylose and a small amount of the amylopectin, so potentially was an ideal sample for measuring the structure of RS, but Figure 20 shows that there is very little relationship between the amount of hot water soluble starch and the RS content. In order to

-36- characterise the structure of the RS molecules, RS was collected during the process of measuring it at the stage just prior to the final hydrolysis. The RS molecules are much smaller than expected (Figure 22), eluting at the end of the separating phase of the UH 500 column, and necessitating the use of the UH 250 column (Figure 23). RS molecules elute after amylose (11 min) and before and with amylopectin. The signal obtained by SEC is consistent with the RS content obtained spectrophotometrically (Figure 23).

The UH 250 column has recently been calibrated for the molecular weight analysis of linear chains of starch31. Amylose molecules are essentially linear, and if the RS is actually portions of amylose chains, the molecular weight of the RS should agree with the predicted molecular weight given by the calibration for that particular elution time. Figure 25 shows the calibration and shows that the molecular weight of the RS is higher than it would be if it was linear chains. Thus, it is likely that at least some of the RS is not linear chains.

The RS from freshly cooked rice of some Australian and Asian rices was examined by SEC for whole molecules and the chains obtained after debranching. Figure 26 shows that the Asian rices are generally higher in RS than the Australian rices (except for Doongara), though they are also higher in amylose content and also are indica varieties. Comparing Figure 26b with Figure 24 shows that the amylopectin chains for full flour elute between 16 and 19 min, with most short chains eluting just before 18 min. However, the elution profile of the chains comprising the RS molecules differs from that of amylopectin chains in that it is not bimodal, but a normal distribution. Further, the longer chains of amylopectin elute between 15.5 and 17 min, and this is exactly the region that the chains of the RS molecules elute (Figure 26b). Figure 27 shows the SEC trace of amylose in the full flour and in the RS for a number of varieties, and shows that very little amylose remains. Either the amylose is easily digested by the enzymes, contributing to the non-resistant starch, or the amylose chains are partially digested and small chains remain. In order to examine the chain length distribution of the RS molecules more closely, capillary electrophoresis (CE) was used. CE gives a series of well separated peaks, each corresponding to a particular chain length of the starch41. It is useful in the range of DP 1 – 80. Figure 28 shows the CE trace of chains of amylopectin from full flour. The shortest chains are DP 6 and the longest detected are DP 87. Most chains are in the range of DP 10 – 15. Figure 28 shows the CE traces of six Australian rices and shows a large peak of DP 1 and 2 (glucose and maltose) and shows proportionally much fewer chains of DP 9 – 15, indicating that these were digested and formed part of the non-resistant starch. Further, the chain length distribution of the RS from freshly cooked rice is unique to each variety, suggesting that the tertiary architecture of the molecule affects the capacity of the digestive enzymes to access the amylopectin. Thus, we conclude that RS from freshly cooked rice is a mixture of shortened amylose chains and the remainder of digested amylopectin. The structure of the amylopectin determines how much of the molecule the enzymes can access.

Naturally occurring RS in rice is quite low, but different ways of handling and cooking rice can augment it. For example, parboiling increase it57 as does different baking methods58. The first type of RS resists digestion because it is physically inaccessible to the enzymes. A study with hamsters fed cooked white rice, cooked brown rice, and cooked white rice plus bran shows that classic responses linked with RS are much higher for the cooked brown rice13. Cooked white rice with added bran elicited the same responses as cooked white rice alone13, suggesting that the bran layer protected the rice from digestion. In this study, we found that the RS content of brown rice was almost twice as high as from white rice, consistent with the findings of Kahlon et al. (2000),

-37- and further establishing the health benefits of brown rice. Cooking rice by the rapid-boil method caused a decrease in RS, suggesting that the molecules that form the RS are leached. Fresh gels showed almost no RS, but after one week, staled gels showed a 5-fold increase in RS. Over a week, the amylose and amylopectin molecules retrograde59, and the retrograded starch would be inaccessible to the enzymes.

Resistant starch was examined in cooked and retrograded rice. Tables 5 and 6 shows the values. Table 6 shows the values for RS in freshly cooked and in retrograded, and shows the difference for each variety upon retrogradation. Koshihikari grown in Japan, IR 64 and PSBRc12 did not retrograde at all, and these are varieties known, at least anecdotally, not to retrograde. Koshihikari grown in NSW did retrograde, indicating an environmental effect on propensity to retrograde. Varieties known to retrograde, anecdotally, contained twice as much RS in the retrograded state than in freshly cooked (highlighted Table 6). These varieties are Kyeema, PSBRc 10, 98 and 18. We therefore suggest that the increase in RS content from freshly cooked to retrograded provides a quantifiable measure of the susceptibility of rice to retrogradation, and importantly, it is a test conducted on cooked rice grains, not on flour. However, it is not a method for routine screening, but perhaps for pre-release of late generation lines.

Variety RS fresh RS retro % difference

Amaroo 1.26 1.67 32.54 Basmati 2.67 3.7 38.58 Doongara 2.15 3.17 47.44 Koshihikari 1.30 1.73 33.08 Koshihikari- 0.97 0.99 2.06 Japan Kyeema 0.89 1.6 79.78

Langi 1.24 1.5 20.97

Opus 1.17 1.41 20.51 IR60 2.99 4.5 50.50 IR8 3.40 3.9 14.71 IR5 3.38 4.85 43.49 IR64 2.54 2.57 1.18 IR24 0.83 1.2 44.58 AE 6.24 11.97 91.83 PSB Rc10 1.88 4.2 123.40 PSB Rc98 3.09 5.9 90.94 PSB Rc12 2.47 2.6 5.26 PSB Rc16 2.57 4.07 58.37 Table 6: RS content of fresh and retrograded rice, and difference between the values (%) PSB Rc18 1.84 3.92 113.04

Figure 30 shows that the RS content of retrograded rice relates quite well to amylose content and to the RS content of freshly cooked rice, but as shown in the Table above, the fact that it doesn’t relate perfectly is an important finding for the quantification of retrogradation.

The structure of RS molecules from retrograded rice (Figure 31a) shows no detectable difference from that of freshly cooked rice. Notable is that Koshi from NSW showed more RS than Koshi from Japan, indicating an environmental effect on RS or retrogradation. Debranched RS from retrograded rice does show a difference in structure from fresh rice (Figure 32) in that the debranched RS from retrograded rice contains a population of chains of about DP 7-12 that is not

-38- present in the RS from freshly cooked rice. Retrogradation of amylopectin would occur in the samples, which would involve recrystallisation of amylopectin chains 60, and in the crystallised state, they would be protected from enzymic digestion.

8. Conclusions

A method to measure the structure of amylose has been developed and will be useful in the future to understand the relationship between amylose structure and cooking properties. Interpreting the chain length distribution of Millin and Koshi in terms of polymer science indicates that amylose from Millin probably forms a gel, which is capable of hardening over time, whereas the amylose from Koshi is more likely to form a visco-elastic paste, which is not capable of hardening to the same degree as Millin. Amylose appears to be related to the amount of resistant starch that occurs both naturally and in retrograded rice, but the main remaining molecules in a sample of resistant starch are the dextrins of amylopectin. Retrogradation is not simple to measure, but a quantifiable measure of retrogradation of cooked rice appears to be the increase in RS content between freshly cooked and retrograded rice. However, the method is not optimal for routine screening, but could conceivably be used for late-generation screening.

9. References

1. Bergman, C. J.; Delgado, J. T.; McClung, A. M.; Fjellstrom, R. G., An improved method for using a microsatellite in the rice waxy gene to determine amylose class. Cereal Chemistry 2001, 78, (3), 257-260. 2. Fitzgerald, M. A., Starch. In Rice Chemistry and Technology, 3rd ed.; Champagne, E. T., 'Ed.'^'Eds.' AACC: St Paul, Minnesota, 2004; 'Vol.' p^pp. 3. Juliano, B. O., Varietal impact on rice quality. Cereals Food World 1998, 43, (4), 207-222. 4. Ong, M. H.; Jumel, K.; Tokarczuk, P. F.; Blanshard, J. M. V.; Harding, S. E., Simultaneous determinations of the molecular weight distributions of amyloses and the fine structures of amylopectins of native starches. Carbohydrate Research 1994, 260, (1), 99-117. 5. Lansky, S.; Kooi, M.; Schoch, T. J., Properties of the fractions and linear subfractions from various starches. Journal of the American Chemical Society 1949, 71, 4066-4075. 6. Takeda, Y.; Hizukuri, S., Structures of branched molecules of amyloses of various origins, and molar fractions of branched and unbranched molecules. Carbohydrate Research 1987, 165, 139-145. 7. Takeda, Y.; Hizukuri, S.; Juliano, B. O., Purification and structure of amylose from rice starch. Carbohydrate Research 1986, 148, 299-308. 8. Takeda, Y.; Maruta, N.; Hizukuri, S., Examining the structure of amylose by tritium labelling of the reducing terminal. Carbohydrate Research 1992, 227, (113-120). 9. Takeda, Y.; Maruta, N.; Hizukuri, S., Structure of amylose subfractions with different molecular sizes. Carbohydrate Research 1992, 226, 279-285. 10. Takeda, Y.; Shitaozono, T.; Hizukuri, S., Structures of sub-fractions of corn amylose. Carbohydrate Research 1990, 199, 207-214. 11. Haralampu, S. G., Resistant Starch - a Review of the Physical Properties and Biological Impact of Rs3. Carbohydrate Polymers 2000, 41, (3), 285-292. 12. Henningsson, A. M.; Margareta, E.; Nyman, G. L.; Bjorck, I. M. E., Influences of Dietary Adaptation and Source of Resistant Starch on Short-Chain Fatty Acids in the Hindgut of Rats. British Journal of Nutrition 2003, 89, (3), 319-327. 13. Kahlon, T. S.; Chow, F. I., Lipidemic Response of Hamsters to Rice Bran, Uncooked or Processed White and Brown Rice, and Processed Corn Starch. Cereal Chemistry 2000, 77, (5), 673-678. 14. Mangala, S. L.; Malleshi, N. G.; Mahadevamma; Tharanathan, R. N., Resistant starch from differently processed rice and ragi (finger millet). Zeitschrift fur Lebensmittel-Untersuchung und-Forschung A-Food Research & Technology 1999, 209, (1), 32-37. 15. Park, O. J.; Kang, N. E.; Chang, M. J.; Kim, W. K., Resistant Starch Supplementation Influences Blood Lipid Concentrations and Glucose Control in Overweight Subjects. Journal of Nutritional Science and Vitaminology 2004, 50, (2), 93-99.

-39- 16. Ayres, N. M.; McClung, A. M.; Larkin, P. D.; Bligh, H. F. J.; Jones, C. A.; Park, W. D., Microsatellites and a single-nucleotide polymorphism differentiate apparent amylose classes in an extended pedigree of US rice germ plasm. Theoretical & Applied Genetics 1997, 94, (6-7), 773-781. 17. Wang, Z.-Y.; Zheng, F.-Q.; Shen, G.-Z.; Gao, J.-P.; Snustad, D. P.; Li, M.-G.; Zhang, J.-L.; Hong, M.-M., The amylose content in rice endosperm is related to the post-transcriptional regulation of the waxy gene. The Plant Journal 1995, 7, (4), 613-622. 18. Hirano, H. Y.; Eiguchi, M.; Sano, Y., A single base change altered the regulation of the waxy gene at the posttranscriptional level during the domestication of rice. Molecular Biology & Evolution 1998, 15, (8), 978-987. 19. Hirano, H. Y.; Sano, Y., Comparison of Waxy gene regulation in the endosperm and pollen in Oryza sativa L. Genes & Genetic Systems 2000, 75, (5), 245-249. 20. Isshiki, M.; Morino, K.; Nakajima, M.; Okagaki, R. J.; Wessler, S. R.; Izawa, T.; Shimamoto, K., A naturally occurring functional allele of the rice waxy locus has a GT to TT mutation at the 5' splice site of the first intron. Plant Journal 1998, 15, (1), 133-138. 21. Mikami, I.; Dung, L. V.; Hirano, H. Y.; Sano, Y., Effects of the two most common Wx alleles on different genetic backgrounds in rice. Plant Breeding 2000, 119, (6), 505-508. 22. Olsen, K. M.; Purugganan, M. D., Molecular evidence on the origin and evolution of glutinous rice. Genetics 2002, 162, (2), 941-950. 23. Bligh, H. F. J.; Till, R. I.; Jones, C. A., A microsatellite sequence closely linked to the Waxy gene of Oryza sativa. Euphytica 1995, 86, (2), 83-85. 24. Hanashiro, I.; Takeda, Y., Examination of number-average degree of polymerization and molar-based distribution of amylose by fluorescent labeling with 2-aminopyridine. Carbohydrate Research 1998, 306, (3), 421- 426. 25. Takeda, Y.; Guan, H. P.; Preiss, J., Branching of amylose by the branching isoenzymes of maize endosperm. Carbohydrate Research 1993, 240, 253-263. 26. Guan, H. P.; Preiss, J., Differentiation of the properties of the branching isozymes from Maize (Zea mays). Plant Physiology 1993, 102, (4), 1269-1273. 27. Morell, M. K.; Blennow, A.; Kosarhashemi, B.; Samuel, M. S., Differntial expression and properties of starch branching enzyme isoforms in developing wheat endosperm. Plant Physiology 1997, 113, (1), 201-208. 28. Nakamura, Y.; Takeichi, T.; Kawaguchi, K.; Yamanouchi, H., Purification of two forms of starch branching enzyme (Q-enzyme) from developing rice endosperm. Physiologia Plantarum 1992, 84, (3), 329-335. 29. Rahman, S.; Regina, A.; Li, Z. Y.; Mukai, Y.; Yamamoto, M.; Kosar-Hashemi, B.; Abrahams, S.; Morell, M. K., Comparison of starch-branching enzyme genes reveals evolutionary relationships among isoforms. Characterization of a gene for starch-branching enzyme IIa from the wheat D genome donor Aegilops tauschii. Plant Physiology 2001, 125, (3), 1314-1324. 30. Cai, Y.; Xie, D. L.; Wang, Z. Y.; Hong, M. M., Interaction of rice bZIP protein REB with the 5 '-upstream region of both rice sbe1 gene and waxy gene. Chinese Science Bulletin 2002, 47, (4), 310-314. 31. Castro, J. V.; Ward, R. M.; Gilbert, R. G.; Fitzgerald, M. A., Measurement of absolute molecular weight distributions of starches. Biomacromolecules 2005, 6, (4), 2260-70. 32. Kossmann, J.; Lloyd, J., Understanding and influencing starch biochemistry. Critical Reviews in Plant Sciences 2000, 19, (3), 171-226. 33. Bhattacharya, K. R.; Sowbhagya, C. M.; Indudhara Swamy, Y. M., Importance of insoluble amylose as a determinant of rice quality. Journal of the Science of Food and Agriculture 1978, 29, 359-364. 34. Mizukami, H.; Takeda, Y.; Hizukuri, S., The structure of the hot-water soluble components in the starch granules of new Japanese rice cultivars. Carbohydrate Polymers 1999, 38, (4), 329-335. 35. Tsai, M. L.; Lii, C. Y., Effect of hot-water-soluble components on the rheological properties of rice starch. Starch-Starke 2000, 52, (2-3), 44-53. 36. Chiou, H.; Martin, M.; Fitzgerald, M. A., Effect of purification methods on rice starch structure. Starke 2002, 54, 415-420. 37. Davey, M. J.; Landman, K. A.; McGuinness, M. J.; Jin, H. N., Mathematical modeling of rice cooking and dissolution in beer production. AIChE Journal 2002, 48, (8), 1811-26. 38. Batey, I. L.; Curtin, B. M., Measurement of amylose/amylopectin ratio by high-performance liquid chromatography. Starch-Starke 1996, 48, (9), 338-344. 39. Hanashiro, I.; Tagawa, M.; Shibahara, S.; Iwata, K.; Takeda, Y., Examination of molar-based distribution of A, B and C chains of amylopectin by fluorescent labeling with 2-aminopyridine. Carbohydrate Research 2002, 337, (13), 1211-1215. 40. Morell, M. K.; Samuel, M. S.; O'Shea, M. G., Analysis of starch structure using fluorophore-assisted carbohydrate electrophoresis. Electrophoresis 1998, 19, (15), 2603-2611.

-40- 41. O'Shea, M. G.; Samuel, M. S.; Konik, C. M.; Morell, M. K., Fluorophore-assisted carbohydrate electrophoresis (face) of oligosaccharides - efficiency of labelling and high-resolution separation. Carbohydrate Research 1998, 307, (1-2), 1-12. 42. Takeda, Y.; Shibahara, S.; Hanashiro, I., Examination of the structure of amylopectin molecules by fluorescent labeling. Carbohydrate Research 2003, 338, (5), 471-475. 43. Banks, W.; Greenwood, C. T., Starch and its components. ed.; Edinborough University Press: Edinborough, 1975; 'Vol.' p. 44. Nguyen, Q. D.; Jensen, C. T. B.; Kristensen, P. G., Experimental and modelling studies of the flow properties of maize and waxy maize starch pastes. Chemical Engineering Journal 1998, 70, (2), 165-171. 45. Gidley, M. J., Molecular mechanisms underlying amylose aggregation and gelation. Macromolecules 1989, 22, 351-358. 46. Jacobs, H.; Delcour, J. A., Hydrothermal modifications of granular starch, with retention of the granular structure - a review [Review}. Journal of Agricultural & Food Chemistry 1998, 46, (8), 2895-2905. 47. Topping, D. L.; Fukushima, M.; Bird, A. R., Resistant starch as a prebiotic and synbiotic: state of the art. Proceedings of the Nutrition Society 2003, 62, (1), 171-176. 48. Brouns, F.; Kettlitz, B.; Arrigoni, E., Resistant starch and "the butyrate revolution". Trends in Food Science & Technology 2002, 13, (8), 251-261. 49. Bird, A. R.; Hayakawa, T.; Marsono, Y.; Gooden, J. M.; Record, I. R.; Correll, R. L.; Topping, D. L., Coarse Brown Rice Increases Fecal and Large Bowel Short-Chain Fatty Acids and Starch but Lowers Calcium in the Large Bowel of Pigs. Journal of Nutrition 2000, 130, (7), 1780-1787. 50. Cheng, H. H.; Lai, M. H., Fermentation of Resistant Rice Starch Produces Propionate Reducing Serum and Hepatic Cholesterol in Rats. Journal of Nutrition 2000, 130, (8), 1991-1995. 51. Croghan, M., Resistant Starch as Functional Component of Foods. Ernahrungs-Umschau 2003, 50, (2), 65- 67. 52. Han, K. H.; Fukushima, M.; Shimizu, K.; Kojima, M.; Ohba, K.; Tanaka, A.; Shimada, K. I.; Sekikawa, M.; Nakano, M., Resistant Starches of Beans Reduce the Serum Cholesterol Concentration in Rats. Journal of Nutritional Science and Vitaminology 2003, 49, (4), 281-286. 53. Wisker, E., Physiological Effects of Resistant Starch Part 2: Effect in the Colon, and Conclusion. Ernahrungs-Umschau 2000, 47, (2), 49-+. 54. Brown, I. L.; McNaught, K. J.; Moloney, E., Hi-maize™: new directions in starch technology and nutrition. ed.; Food Australia: 1995; 'Vol.' p. 55. McCleary, B. V.; Monaghan, D. A., Measurement of resistant starch. Journal of AOAC International 2002, 85, (3), 665-675. 56. Mangala, S. L.; Tharanathan, R. N., Structural Studies of Resistant Starch Derived From Processed (Autoclaved) Rice. Zeitschrift Fur Lebensmittel-Untersuchung Und-Forschung a-Food Research and Technology 1999, 209, (1), 38-42. 57. Devi, K.; Geervani, P., Rice Processing - Effect on Dietary Fibre Components and in Vitro Starch Digestibility. Journal of Food Science and Technology-Mysore 2000, 37, (3), 315-318. 58. Akerberg, A.; Liljeberg, H.; Bjorck, I., Effects of Amylose/Amylopectin Ratio and Baking Conditions on Resistant Starch Formation and Glycaemic Indices. Journal of Cereal Science 1998, 28, (1), 71-80. 59. Gidley, M. J.; Bulpin, P. V., Aggregation of amylose in aqueous systems: the effect of chain length on phase behavior and aggregation kinetics. Macromolecules 1989, 22, 341-346. 60. Slade, L.; Levine, H., A food polymer science approach to selected aspects of starch gelatinization and retrogradation. Frontiers in Carbohydrate Research 1989, 215-270.

-41-