(Oryza Glaberrima) Flour and Starch

Structural and Functional Characteristics of African Rice (Oryza glaberrima) Flour and Starch

Joseph Kwesi Gayin

A Thesis

presented to

The University of Guelph

In partial fulfillment of requirements

for the degree of

Doctor of Philosophy

Food Science

Guelph, Ontario, Canada

STRUCTURAL AND FUNCTIONAL CHARACTERISTICS OF AFRICAN RICE

(Oryza glaberrima) FLOUR AND STARCH

Joseph Kwesi Gayin Advisors:

University of Guelph, 2015 Dr. Massimo Marcone

Dr. ElSayed Abdelaal

African rice (Oryza glaberrima) and Asian rice (Oryza sativa) are two distinct types of domesticated rice with the former being less extensively studied. With the expansion of rice-based food products in Sub-Saharan Africa, characterization of African rice and their starches is essential for product development. Seven African Rice Accessions (ARAs) were evaluated for grain quality attributes, thermal and in vitro digestive properties.

Additionally, the physical, molecular and structural properties of starches isolated from the

ARAs were investigated. They were compared with two Asian Rice Varieties (ARVs) and

NERICA 4 (cross between sativa x glaberrima). The physical grain quality characteristics of the intermediate-high amylose ARAs were comparable to the low amylose ARVs however, the ARAs had significantly (p<0.05) lower cooking solids loss and harder texture while the ARVs were softer and more adhesive. The glycemic indices of both rice types were > 70. The ARV flours showed higher breakdown tendency than the ARAs. All the starches displayed 'A' type X-ray diffraction pattern and micrographs showed tightly packed polyhedral granules. The gelatinization transition temperatures: onset (To), peak

(Tp) and conclusion (Tc) were significantly (p<0.05) higher in ARA flours, and melting transition temperatures of retrograded flour gels showed significant differences (p<0.05) among samples. The energy (ΔH) required to gelatinize the starches or melt retrograded starch gels was greater in the ARAs. The relative molar concentration of B-chains in the amylopectins was approximately 50% in all samples, consequently, the ratio of A:B-chains was 1.0. The ARAs and NERICA 4 had greater amount of “fingerprint” Bfp-chains than the ARVs but the major group of short B-chains was similar in all samples. Positive correlations (p<0.05) were observed between external chain length (ECL) and To, Tp, Tc and ΔH. The average degree of polymerization (DP) of the ARA clusters were larger and the ARAs and NERICA 4 had significantly higher average DP of the branched building blocks (Bbl) than the ARVs due to abundance of large Bbl of group 6. This also resulted in higher average number of chains (NC), longer average chain length (CL) and higher degree of branching (DB) of the Bbl in the ARAs.

DEDICATION

This thesis is dedicated to JEHOVAH GOD in whom I live, move and have my being; my wife Susuana Gayin and kids- Christopher, Jason, Jeffrey and Amanda.

ACKNOWLEDGMENTS

My profound appreciation goes to the Almighty God for his immeasurable Grace, favor and protection for my family and I throughout my stay in Canada for my PhD programme in Food Science. He comforted me and showed me compassion; His steadfast love never ceased, it was new every morning!

I am grateful to my sponsors; GLOBAL RICE SCIENCE PARTNERSHIP (GRiSP) for the award of scholarship to pursue this PhD programme in Food Science tenable at the

University of Guelph. I wish to also show gratitude to the AFRICA RICE CENTER

(AfricaRice) for their immeasurable financial support and facilitating role in making this dream a reality. My special thanks go to the Management, the Training Coordinator and the Head of Grain Quality Laboratory Dr. John Manful who was so instrumental in many ways.

It is imperative to mention my Advisor Dr. Koushik Seetharaman and family (Debrah and

Samuel) who gave me an opportunity to work with him. My sincere appreciation to him for the priceless guidance and advice throughout the time I worked with him. It is difficult to find words to describe his personality, kindness, selflessness, benevolence and unique human relations that has arguably made him one of the cherished people in my life.

Unfortunately he did not live to see the completing of the great work he began. May his soul rest in peace! I wish to express many thanks to Dr. Rickey Yada who took over as

Adviser when Koushik left for the University of Minnesota. His inspiration and encouragement is worth mentioning. My heart goes for Dr. Massimo Marcone who succeeded Dr. Rickey Yada as Adviser when the latter left for the University of British

Columbia. His kindness and willingness to help at all times is amazing to say the least. It is noteworthy to mention Dr. ElSayed Abdelaal for his co-advisory role. Your valuable suggestions and encouragement right from the proposal stage through till the end of my programme is very much appreciated. Dr. John Manful did not allow distance to hinder his role as a Committee member. He ensured that AfricaRice met her financial obligations and was very instrumental in getting funds transferred timeously. I owe a great debt to Dr. Eric

Bertoft for his kind advice and valuable suggestions throughout the course of this research.

Dr. Sanaa Ragaee perfectly played the role of a mum to my family and I. She ensured our well-being and my kids never cease mentioning her name.

The office staff at the Department of Food Science - Dr. Art Hill, Anne Ingram, Leona

Varga and Tricia Anderson deserve recognition. I would like to thank you for all your administrative support and providing me with laboratory supplies I needed for my research.

I would also like to acknowledge Dr. Lisa Duizer, Dr. Bruce Holbein and Dr. Kakuda.

In no particular order, I would like to acknowledge and appreciate the post doctorial fellows in the cereal and bake laboratories and my laboratory mates (both seniors and contemporaries) for making life in Guelph worthwhile and enjoyable: Gupreet Kaur, Jayne

Bock, Zhu Fan and Zhan-Hui Lu, Renuka Waduge, Komeine Nantanga, Shane Walker,

Sahar, Varatharajan Vamadevan, Simarata Dhillon, Danusha Kalinga, George and Lizzy

Annor, Anna Källman, Ronny, Bruce Manion, Matt MacSweeney, Avi Goldstein, Azadeh,

Allison, Ryan West, Hong Kai Ma, Pragyani Bora, Igor Guzar, Saba, Magid, Hanadi,

Seham, Ghazal Peyampour, Vivian Adams, and Enoch Quayson. My thanks also go to the

Anim-Somuah family for their help and support in many ways.

Additionally, I would also like to thank my family Susan, Christopher, Jason, Jeffrey,

Amanda, for their support in many ways. I am forever grateful to my mom Christiana

Sam, all my siblings- Gifty, Sarah and Emma

Finally I would like to again acknowledge the immense contributions of Dr. Koushik

Seetharaman, Dr. John Manful, Dr. Eric Bertoft, Dr. Elsayed Abdelaal and Dr. Massimo

Marcone. To God be the Glory!

vii

Table of Contents

DEDICATION………………………………………………………………………… iv

ACKNOWLEDGMENTS………………………………………………………...... v

Table of Contents………………………………………………………………………… viii

List of Tables……………………………………………………………………...... xv

List of Figures…………………………………………………………………………….. xvii

List of Abbreviations……………………………………………………………………… xxi

CHAPTER 1: INTRODUCTION………………………………………………………… 1

CHAPTER 2: LITERATURE REVIEW…………………………………………………. 5

2.1. Rice……………………………………………………………………………... 5

2.2 Proximate Composition of Starch………………………………………………. 7

2.3 Rice Starch……………………………………………………………………… 9

2.3.1 Morphology of Rice Starch Granule……………………………………………. 9

2.3.2 Starch Structures………………………………………………………………... 10

2.3.3 Cooking Characteristics………………………………………………………… 14

2.3.4 Thermal Properties……………………………………………………………… 15

2.3.5 Pasting Properties………………………………………………………………. 18

2.3.6 Swelling and Solubility Characteristics………………………………………… 20

2.4 In vitro Digestibility and Glycemic Attributes of Rice…………………………. 21

2.5 Conclusions …………………………………………………………………….. 23

2.6 Significance of project …………………………………………………………. 24

2.7 References (Chapters 1 and 2)………………………………………………….. 25

viii

CHAPTER 3: CLASSIFICATION OF RICE BASED ON STATISTICAL ANALYSIS

OF PASTING PROPERTIES AND APPARENT AMYLOSE CONTENT: THE CASE

OF Oryza glaberrima ACCESSIONS FROM AFRICA…………………………………. 42

3.1 Introduction……………………………………………………………………... 44

3.2 Materials and Methods………………………………………………………….. 45

3.3 Materials………………………………………………………………………... 45

3.3.1 Measurement of Pasting Properties ………..…………………………………... 46

3.3.2 Measurement of Apparent Amylose Content…………………………………... 46

3.4 Statistical Analysis……………………………………………………………… 47

3.5 Results and Discussion…………………………………………………………. 47

3.5.1 Pasting Properties of Rice Flours ……………………………………...... 47

3.5.2 Apparent Amylose Content of Rice Flours …………………………………….. 48

3.6 Classification of Rice Accessions………………………………………………. 50

3.7 Comparison of Clustering Pattern………………………………………………. 50

3.7.1 Cluster I…………………………………………………………………………. 50

3.7.2 Cluster II………………………………………………………………………... 51

3.7.3 Cluster III………………………………………………………………………. 51

3.7.4 Cluster IV……………………………………………………………………….. 52

3.7.5 Cluster V………………………………………………………………………... 53

3.8 Correlations……………………………………………………………………... 53

3.9 Conclusion……………………………………………………………………… 54

3.10 Acknowledgements……………………………………………………………... 55

3.11 Tables and Figures……………………………………………………………… 56

3.12 References………………………………………………………………………. 68

CHAPTER 4: PHYSICOCHEMICAL PROPERTIES, COOKING QUALITY AND 72

GLYCEMIC ATTRIBUTES OF AFRICAN RICE (Oryza glaberrima)

4.1 Introduction……………………………………………………………………... 74

4.2 Materials and Methods………………………………………………………….. 76

4.3 Materials………………………………………………………………………... 76

4.4 Methods………………………………………………………………………… 77

4.4.1 Physico-Chemical Properties…………………………………………………… 77

4.4.2 1000-Kernel Weight……………………………………………………………. 77

4.4.3 Bulk Density……………………………………………………………………. 77

4.4.4 Length-Width Ratio…………………………………………………………….. 78

4.4.5 Amylose Content……………………………………………………………….. 78

4.4.6 Nitrogen………………………………………………………………………… 78

4.4.7 Total Starch……………………………………………………………………... 78

4.4.8 Colour Measurement of Rice Kernels ………………………………………….. 79

4.4.9 Cooking Characteristics………………………………………………………… 79

4.4.10 Optimum Cooking Time………………………………………………………... 79

4.4.11 Textural Profile Analysis (TPA)………………………………………………... 80

4.4.12 Pasting Properties………………………………………………………………. 81

4.4.13 Differential Scanning Calorimetry (DSC)……………………………………… 82

4.4.14 In vitro Digestibility and Expected Glycemic Index of Optimally Cooked Rice 82

4.5 Statistical Analysis……………………………………………………………… 84

4.6 Results and Discussion…………………………………………………………. 84

4.7 Physicochemical Properties…………………………………………………….. 84

4.8 Cooking and Textural Properties……………………………………………….. 85

4.9 Pasting and Thermal Characteristics……………………………………………. 85

4.10 In vitro Digestibility of Cooked Rice Samples…………………………………. 87

4.11 Discussion………………………………………………………………………. 87

4.12 Conclusion…………………………………………………………………….... 91

4.13 Acknowledgements……………………………………………………………... 92

4.14 Tables and Figures……………………………………………………………… 93

4.15 References………………………………………………………………………. 106

CHAPTER 5: PHYSICAL AND MOLECULAR CHARACTERIZATION OF

STARCHES ISOLATED FROM Oryza glaberrima ACCESSIONS…………………..... 113

5.1 Introduction……………………………………………………………………... 115

5.2 Materials and Methods………………………………………………………….. 117

5.2.1 Samples…………………………………………………………………………. 117

5.2.2 Methods………………………………………………………………………… 117

5.2.3 Starch Isolation…………………………………………………………………. 117

5.2.4 Physical Characteristics………………………………………………………… 118

5.2.5 Starch Damage………………………………………………………………….. 118

5.2.6 Nitrogen………………………………………………………………………… 118

5.2.7 Particle Size Determination…………………………………………………….. 119

5.2.8 Morphology ……………………………………………………………………. 119

5.2.9 Scanning Electron Microscopy…………………………………………………. 119

5.2.10 Wide Angle X-ray Diffraction Scattering ………..……………………………. 120

5.2.11 Reﬂectance Spectra after Exposure of Starch to Iodine Vapor…………………. 121

5.2.12 Thermal Properties……………………………………………………………… 121

5.2.13 Rapid Visco Analyzer…………………………………………………………... 121

5.2.14 Differential Scanning Calorimetry ……...……………………………………… 122

5.2.15 Molecular Composition of Rice Starches………………………………………. 123

5.2.16 Chain Length and β-Limit Dextrins…………………………………………….. 123

5.2.17 Chain Length of Debranched Starches ………………………………………… 124

5.2.18 Amylose ………………………………………………………………………... 124

5.2.19 Statistical Analysis……………………………………………………………… 125

5.3 Results ………………………………………………………………………….. 125

5.3.1 Starch Yield and Composition ………………………………………………… 125

5.3.2 Morphology and Structure of Rice Starch Granules…………………………..... 125

5.3.3 Thermal Properties of Starch Granules…………………………………………. 127

5.3.4 Molecular Structure of Rice Starches…………………………………………... 128

5.4 Discussion………………………………………………………………………. 130

5.5 Conclusion……………………………………………………………………… 136

5.6 Acknowledgements……………………………………………………………... 137

5.7 Tables and Figures……………………………………………………………… 138

5.8 References………………………………………………………………………. 155

CHAPTER 6: UNIT AND INTERNAL CHAIN PROFILE OF AMYLOPECTIN

ISOLATED FROM Oryza glaberrima ACCESSIONS…………………………………... 162

6.1 Introduction……………………………………………………………………... 164

6.2 Materials and Methods………………………………………………………….. 167

xii

6.2.1 Materials………………………………………………………………………... 167

6.2.2 Methods………………………………………………………………………… 167

6.2.3 Starch isolation…………………………………………………………………. 167

6.2.4 Isolation of Amylopectin from Rice Starch…………………………………….. 167

6.2.5 Determination of Purity f Amylopectin by Gel Permeation Chromatography…. 169

6.2.6 Production of φ,β-Limit Dextrin of Amylopectin……………………………… 170

6.2.7 Analysis of Unit Chain Distribution …………………………………………… 171

6.2.8 Statistical Analysis ……………………………………………………………... 172

6.3 Results and Discussion…………………………………………………………. 172

6.3.1 Purity of Amylopectin …………………………………………………………. 172

6.3.2 Unit Chain Length Distributions of Amylopectins……………………………... 173

6.3.3 Unit Chain Length Distributions of φ,β-Limit Dextrins………………………... 173

6.4 Correlations between Chain Length Classifications and Some Functional

Properties……………………………………………………………………… 177

6.5 Conclusion……………………………………………………………………… 179

6.6 Acknowledgements……………………………………………………………... 179

6.7 Tables and Figures……………………………………………………………… 180

6.8 References………………………………………………………………………. 188

CHAPTER 7: STRUCTURE OF CLUSTERS AND BUILDING BLOCKS IN

AMYLOPECTIN FROM AFRICAN RICE ACCESSIONS……………………………... 193

7.1 Introduction……………………………………………………………………... 196

7.2 Materials and Methods………………………………………………………….. 198

7.2.1 Materials………………………………………………………………………... 198

xiii

7.2.2 Enzymes………………………………………………………………………… 198

7.2.3 Methods………………………………………………………………………… 199

7.2.4 Starch Isolation…………………………………………………………………. 199

7.2.5 Isolation of Amylopectin from Rice Starch and Determination of Purity……… 199

7.2.6 Cluster preparation - Time Curse for α-Amylolysis of Amylopectin ………... 199

7.2.7 Production of Clusters and Their Corresponding φ,β-Limit Dextrins………….. 200

7.2.8 Chain Length Distribution of Clusters………………………………………….. 201

7.2.9 Preparation of Building Blocks and Determination of Structure……………….. 202

7.2.10 Statistical Analysis……………………………………………………………… 202

7.3 Results and Discussion…………………………………………………………. 203

7.3.1 Time Course Analysis of α-Amylosis of Amylopectin from Rice Starch……… 203

7.3.2 Characterization of φ,β-Limit Dextrins Produced from Clusters………………. 204

7.3.3 Unit Chains in Clusters…………………………………………………………. 205

7.3.4 Characterization of Building Blocks in Clusters……………………………….. 208

7.3.5 Correlations……………………………………………………………………... 211

7.3.6 Conclusions……………………………………………………………………... 211

7.3.7 Acknowledgements……………………………………………………………... 212

7.3.8 Tables and Figures……………………………………………………………… 213

7.3.9 References………………………………………………………………………. 225

CHAPTER 8: CONCLUSIONS AND FUTURE WORK………………………………... 229

xiv

List of Tables

Table 2.1 Nutritional Composition of Milled Rice……………………………………… 8

Table 3.1 Descriptive Statistics of Entire Population of 1,020 O. glaberrima

Accessions……………………………………………………………………. 56

Table 3.2 Distribution of 1020 O. glaberrima Rice Accessions Obtained through

Cluster Analysis Using SPSS……………………………………………… 57

Table 3.3 Correlations between Pasting Parameters and Apparent Amylose Content

Within the 5 Clusters of 1,020 O. glaberrima Accessions…………………... 59

Table 4.1 Physicochemical Characteristics of African and Asian Rice Kernels and

flours………………………………………………………………………….. 93

Table 4.2 Cooking and Textural Properties of African and Asian Rice Kernels………... 94

Table 4.3 Pasting Properties of Flours from African and Asian Rice Kernels………….. 95

Table 4.4 DSC Gelatinization Temperatures and Enthalpy of Flours from African and

Asian Rice Kernels………….………………………………………………... 96

Table 4.5 Transition Parameters of Stored African and Asian Rice Flour Gels……….. 97

Table 4.6 Nutritional Starch Fractions During Hydrolysis of Cooked African and Asian

Rice…..………………………………………………………………………. 98

Table 4.7 Kinetic Constant, Hydrolysis Index and Expected Glycemic Index of Cooked

African and Asian Rice……………………………………………………….. 99

Table 4.8 Correlations between Amylose; Protein and Cooking and Textural Properties 100

Table 4.9 Correlations between Amylose and Thermal and Pasting Properties………… 101

Table 5.1 Characteristics and Composition of Starches Isolated from African Rice

Accession and Asian Rice Varieties………………………………………….. 138

Table 5.2 Pasting Properties of Native Starches Isolated from African Rice Accessions

and Asian Rice Varieties…………………………………………………….. 139

Table 5.3 Gelatinization Parameters of Native Starches Isolated from African Rice

Accessions and Asian Rice Varieties………………………………………… 140

Table 5.4 Melting Parameters of Retrograded Starch Gels of African Rice Accessions... 141

Table 5.5 Correlations between Amylose Concentration and Some Measured

Parameters…………………………………………………………………….. 142

Table 6.1 Chain Length Classifications and φ, β-Limit Values of Rice Amylopectin as

obtained by HPAEC Analyses……………………………………………… 180

Table 6.2 Relative Molar Amounts (%) of Chain Groupings in Amylopectins Isolated

from Rice Starches…………………………………………………………... 181

Table 6.3 Molar Ratios of Different Chain Groupings of Rice Amylopectins and Their

φ,β-Limit Dextrins…………………………………………………………… 182

Table 6.4 Correlation between Chain Groupings of Rice Amylopectins and Some

Function Properties of Starch..……………………………………………….. 183

Table 7.1 Characterization of φ,β Limit Dextrins of Clusters from Amylopectin ……… 213

Table 7.2 Relative Molar Distribution (%) of Different Chain Categories in Clusters

Isolated from Amylopectin…………………………………………………. 214

Table 7.3 Structure of Branched Building Blocks in Clusters of Amylopectin from

African Rice and Asian Rice………………………………………………… 215

Table 7.4 Building Block Structures of Clusters of Amylopectin from African Rice

Accessions and Asian Rice Varieties………………………………………… 216

xvi

List of Figures

Figure 2.1 Structure of amylose and amylopectin…………………………………….. 11

Figure 3.1 Graphical representation of distribution of 1,020 African rice accessions

into 5 clusters and 23 sub clusters………………………………………… 61

Figure 3.2 Distribution of entire population of 1, 020 O. glaberrima accessions based

on apparent amylose content and pasting properties………………………. 62

Figure 3.3 Distribution of O. glaberrima accessions in cluster I population based on

apparent amylose content and pasting properties………………………….. 63

Figure 3.4 Distribution of O. glaberrima accessions in cluster II population based on

apparent amylose content and pasting properties………………………….. 64

Figure 3.5 Distribution of O. glaberrima accessions in cluster III population based

on apparent amylose content and pasting properties………………………. 65

Figure 3.6 Distribution of O. glaberrima accessions in cluster IV population based

on apparent amylose content and pasting properties………………………. 66

Figure 3.7 Distribution of O. glaberrima accessions in cluster V population based on

apparent amylose content and pasting properties………………………….. 67

Figure.4.1 Pasting profiles of African and Asian rice flours………………………….. 102

Figure.4.2 DSC thermograms of African and Asian rice flours………………………. 103

Figure.4.3 Hydrolysis kinetics of cooked and freeze-dried African and Asian rice

kernels……………………………………………………………………... 104

Figure 4.4 Multiple linear regression models to describe the relationship between

amylose, OCT and cooking and textural parameters……………………… 105

Figure 5.1 SEM micrographs of isolated starch from African rice accessions and

xvii

Asian rice varieties………………………………………………………… 143

Figure 5.2 Particle size distribution of starches isolated from African rice accessions

and Asian rice varieties……………………………………………………. 144

Figure 5.3 Wide angle x-ray diffractograms of starches isolated from African rice

accessions and Asian rice varieties………………………………………... 145

Figure 5.4 Wide angle x-ray diffractograms of starch and iodinated starches from

African and Asian rice…………………………………………………….. 146

Figure 5.5 K/S Spectra of starches isolated from African rice accessions and Asian

rice varieties……………………………………………………………….. 147

Figure 5.6 RVA pasting profiles of starches isolated from African rice accessions

and Asian rice……………………………………………………………… 148

Figure 5.7 Thermal properties of starches isolated from African rice accessions and

Asian rice varieties………………………………………………………… 149

Figure 5.8 Thermal properties of retrograded starch gels fromAfrican rice accessions

and Asian rice varieties……………………………………………………. 150

Figure 5.9 Gel-Permeation Chromatograms of debranched rice starches fractionated

on Sepharose CL-6B………………………………………………………. 151

Figure 5.10 Sepharose CL-2B Gel-Permeation Chromatograms of whole rice starches

and their corresponding β-LDs…………………………………………….. 152

Figure 5.11 Correlation between amylose concentration and pasting parameters……... 153

Figure 5.12 Multiple linear regression models to describe the relationship between

amylose and thermal parameters of isolated starches……………………... 154

Figure 6.1 Gel permeation chromatograms (Sepharose CL-6B) of debranched whole

xviii

starch and purified amylopectin from rice starch (Koshihikari)…………... 184

Figure 6.2 Unit chain outline of debranched rice amylopectins obtained by High

Performance Anion-Exchange Chromatography (HPAEC)………………. 185

Figure 6.3 Unit chain outline of debranched φ,β-limit dextrins of rice amylopectins

obtained by High-Performance Anion-Exchange Chromatography

(HPAEC)…………………………………………………………………... 186

Figure 6.4 Multiple linear regression models to describe the relationship between

ECL, Enthalpy, Final and Setback viscosities…………………………….. 187

Figure 7.1 Time course for α-amylolysis of rice amylopectin by gel-permeation

chromatography on Sepharose CL-6B…………………………………….. 217

Figure 7.2 Development of relative molar amounts of branched dextrins in the time

course for α-amylolysis of rice amylopectin for the production of clusters 218

Figure 7.3 Size distribution of φ,β-limit dextrins of clusters isolated from rice

amylopectin by GPC on Sepharose CL-6B………………………………... 219

Figure 7.4 HPAEC chromatogram of unit chain distribution of cluster from TOG

12440 (African rice)……………………………………………………….. 220

Figure 7.5 Internal chain length distribution of cluster from TOG 12440 (African

rice)………………………………………………………………………... 221

Figure 7.6 Size distribution obtained by GPC on Superdex 30 of building blocks in

clusters of rice amylopectin ………………………………………………. 222

Figure 7.7a HPAEC chromatogram of unit chain distribution of building blocks from

TOG 12440 (African rice). ……………………….……………………….. 223

Figure 7.7b HPAEC chromatogram of unit chain distribution of building blocks from

xix

TOG 12440 (African rice) ……………………….………………………... 223

Figure 7.7c Example of unit chain compositions in building blocks of TOG 12440

(African rice) on weight basis……………………………………………... 224

List of Abbreviations

AAC Apparent Amylose Content

AACC The American Association of Cereal Chemists

Acrystal Chains Calculated as all A-chains less Afp

Afp Fingerprint A-Chains (DP 2)

ARA African Rice accession

Bfp Fingerprint B-Chains (DP 3 - DP 7)

BDV Breakdown Viscosity

BL Long B-Chains (DP >26)

BL-CLLD Chain Length of Long B-chains in Limit Dextrins

BS-CLLD Chain Length of Short B-chains in Limit Dextrins

BS Short B-Chains

BSmajor Major Group of Short B-Chains (DP 8-25)

C Cluster

CL Average Chain Length

CLLD Average Chain Length of φ, β-Limit Dextrin

xxi

DB Degree of Branching

DBbl Density of Building Blocks

DMSO Dimethyl Sulfoxide

DP Degree of Polymerization

DSC Differential Scanning Calorimetry

ECL External Chain Length eGI Expected Glycemic Index

FV Final Viscosity

GI Glycemic Index

GL Glycemic Load

GPC Gel-Permeation Chromatography

HPAEC- High Performance Anion-Exchange Chromatography with Pulsed

PAD Amperometric Detection

IB-CL Inter-Block Chain Length

ICL Internal Chain Length

L Long Chains in Whole Amylopectin (DP ≥ 38)

xxii

LCAM Long Chain Amylose

LCL Chain Length of Long Chains

LER Length Expansion Ratio

MR Melting Range

NBbl Number of Branched Building Blocks

NC Number of Chains

NERICA New rice for Africa

OCT Optimum Cooking Time

PKT Peak Time

PT Pasting Temperature

PV Peak Viscosity

RDS Rapidly Digestible Starch

RS Residual Starch

RT% Percentage retrogradation

RVA Rapid Visco Analyzer

RVU Rapid Visco Units

xxiii

S Short Chains in Whole Amylopectin (DP ≤ 37)

SB Setback Viscosity

SC Sub Cluster

SDS Sodium Dodecyl Sulfate

SEM Scanning Electron Microscopy

SCAM Short Chain Amylose

SCL Chain Length of Short Chains

SL Solids Loss

Tc Conclusion of Gelatinisation Temperature

To Onset of Gelatinisation Temperature

Tp Peak of Gelatinisation Temperature

TICL Total Internal Chain Length

TPA Texture Profile Analysis

TV Trough Viscosity

VER Volume Expansion Ratio

WAXS Wide Angle X-ray Diffraction Scattering

xxiv

WUR Water Uptake Ratio

β-LDs Beta Limit Dextrins

ΔH Enthalpy of Gelatinization

φ,β-LD Phi Beta Limit Dextrins

λmax Wavelength at the Absorbance Maxima

xxv

CHAPTER 1: INTRODUCTION

Rice is one of the major cereal crops cultivated worldwide and its starch is one of the key ingredients of various food products (Lawal et al. 2011). The Food and Agricultural

Organization (FAO) estimates that approximately 496.6 million tonnes of milled rice

(paddy that has its husk, germ and bran layers removed) was produced in 2013 while the global demand for rice as a food is about 410.4 million tonnes (FAO 2014). The production of milled rice in Africa in 2013 was 17.5 million tonnes and the current estimate for rice imports stands at 13.9 million tonnes (FAO 2014). The two widely cultivated rice species are Oryza sativa L. (Asian rice) and Oryza glaberrima Steud. (African rice) and both species can grow in a wide range of water-soil regimes (Luh 1991). Rice is consumed largely in the cooked form, however, products such as snacks, cakes and noodles can be made from rice (Marshall et al. 1990). Starch is the predominant component (90% dry basis) of milled rice and its content of amylose and physico-chemical properties (i.e. physical and chemical), including gelatinization temperature, gel consistency and pasting viscosity, determine the cooking and eating qualities of rice. Consequently these properties are important since there is increased consumer demand for the improvement of grain quality to meet diverse preferences (Bao et al. 2008).

Some unique properties of rice starch are flavor carrying capability, hypoallergenicity, bland taste, smoothness, creaminess and spreadability. These properties permit rice starch and flour to be applied in a number of value added products like gluten free bread, beverages, puddings and low fat sauces (Lawal et al. 2011; Bryant et al. 2001).

Additionally, the functional properties of rice have significant effects on the quality of 1

starch-based products (Bhattacharyya et al. 2004). Physico-chemical properties of rice starch depend largely on the rice variety, isolation procedure, climatic and soil conditions during development of the rice grain (Bao et al. 2004; Singh et al. 2005).

The functional properties of starch such as solubility, swelling, water absorption capacity, gelatinization and pasting provide useful insights into diverse prospective applications of starches for domestic and industrial purposes. The gelatinization and pasting properties of rice starch give indications of the ease of its cooking while rheological characteristics offer information about the viscoelastic properties of the starches.

The two main components of starch are amylose and amylopectin and the proportions and structures of these two components are the main factors that affect the cooking quality of rice. Previous studies have shown that amylose has a large impact on the cooking quality of a rice variety (Cuevas and Fitzgerald 2007) and its content in rice has been reported to vary from 0 to 33%. Milled rice cultivars may be generally classified on the basis of their amylose into waxy (1–2%), very low (2–12%), low (12–20%), intermediate (20–25%), or high (>25%) (Bao 2012). Rice grains with low amylose levels are associated with tender, cohesive, glossy grains when cooked, while grains with higher amylose levels appear dry, fluffy and separated as cooked rice (Preecha 2008).

The textural properties of rice ﬂours are mainly contributed by starch due to its gelatinization, retrogradation and rheological behavior. Starch gelatinization involves the disruption of molecular/crystalline order within the starch granule and the process includes changes in the physical, chemical and nutritional properties of starch (Tsai et al. 1997).

Additional changes include water and heat diffusivity, rheological behavior, swelling and 2

deformation of original shape of starchy products and susceptibility to enzymatic digestion

(Bhattacharyya et al. 2004).

Rice has become a popular food in Africa over the years because of urbanization, population growth, taste, ease of cooking and changes in consumer habits. This rice diet shift has been supported by income growth which has shifted urban consumers’ preferences in favour of products that can be easily cooked, at the expense of other types of cereals and tuber crops that require more time to prepare (Hirsch 1998; Diagana et al. 1999). However, the volume of rice produced in West Africa is insufficient to meet the growing consumer demand in the sub-region (FAO 2012).

Asian Rice (O. sativa) has become popular with farmers compared to African rice (O. glaberrima) because of comparative advantages of higher grain yield, better grain qualities and wide adaptability to rice production ecologies and agronomic practices. However, in spite of its predisposition to lodging and shattering, O. glaberrima is generally adapted to a number of rice biotic and abiotic stresses and has other interesting traits that include a preferred taste locally, early maturity, better weed competitiveness and ceremonial and cultural importance (Maji 2010; Falade et al. 2014). This has necessitated the systematic characterization of a collection of more than 2,500 accessions of Oryza glaberrima by

Africa Rice Center (AfricaRice) because it is thought that there may be a valuable resource of potentially very important gene pool that may be of benefit in rice breeding programs in this region. A special effort is being made to screen for resistance against major diseases and environmental stresses such as acidity, iron toxicity, cold and salinity. To contribute to the broader objective of this characterization process, this study seeks to explore the

structural and functional characteristics of African rice flour and starches. Samples were selected from a population of 1,020 African rice accessions from West Africa based on data of their pasting characteristics and apparent amylose content using hierarchical cluster analysis technique.

The goal of this research is to ultimately promote and add value to African rice and to identify African rice accessions exhibiting promising traits for further improvement. The study is aimed to investigate the following hypotheses:

a) Organization of glucan chains in African rice starches is related to their in vitro

digestibility.

b) Organization of glucan chains in African rice starches is related to their thermal and

pasting properties.

To investigate the above hypotheses the following objectives were studied: a) Determine the cooking quality and glycemic attributes of selected African rice

accessions b) Establish the physical, thermal and molecular characteristics of starches isolated from

African rice c) Investigate the structure of African rice starches and relate it to their thermal and

pasting properties

CHAPTER 2: LITERATURE REVIEW

2.1. Rice

Rice is the seed of the monocot plant of the genus Oryza and of the grass family

Gramineae. Other crops that belong to this grass family include wheat, corn, rye, oats and barley (Khush 1997). Rice cultivation has a long history and it is believed to have been first domesticated nearly 9000 years ago by inhabitants in the region of the Yangtze River valley in China (Fuller et al. 2010; Liu et al. 2007). Rice is now grown worldwide and it is a major source of food for more than half of the world’s population, particularly those living in some of the most populous countries such as China and India. Rice is cultivated in three main ecologies namely; rain-fed lowland, rain-fed upland and irrigated plains

(Sarla and Swanmy 2005) and it ranks fifth in terms of area under cultivation after maize, sorghum, millets and cassava (Nayar 2010).

There are two distinct types of domesticated rice; Oryza sativa (Asian rice) and Oryza glaberrima (African rice) (Ma and Bennetzen 2004). O. glaberrima is reported to have been domesticated more than 3,500 years ago from wild ancestors along the Niger River in

Mali, West Africa (Carney 2001). O. sativa is distributed globally but highly concentrated in Asia making it more diverse than O. glaberrima. The two major cultivated subspecies of O. sativa are the sticky, short grained japonica or sinica variety and the non-sticky, long-grained indica variety. O. glaberrima is grown in West Africa which is the most dominant rice growing region in Africa (Sweeney and McCouch 2007; Oko and Ogwu

2010; Nayar 2010). The growing areas span eighteen countries, namely; Benin, Burkina

Faso, Cameroon, Cape Verde, Chad, Cote d’Ivoire, Gambia, Ghana, Guinea, Guinea-

Bissau, Liberia, Mali, Mauritania, Niger, Nigeria, Senegal, Sierra Leone and Togo (Vernet 5

2002). The two major types of this species reported are the photosensitive type that grows in deep water and the early erect type that grows both in lowland and upland ecologies,

(Ghesquiere et al. 1997). African rice has some useful qualities, which include its ability to compete with weeds for nutrients (Jones et al. 1997; ), tolerance to drought conditions, acidity and iron toxicity (Maji et al. 2001; Sano et al. 1984) and the ability to respond to low input conditions (Sarla and Swanmy 2005). Additionally, African rice is resistant to the rice yellow mottle virus (Ndjiondjop et al. 1999), nematodes (PlowRight et al. 1999) and stem borer (Sauphanor 1985).

Climatic and field conditions, cultural management methods and genetics of the crop influence the physiological processes that occur within the plant. The above-mentioned factors also affect the duration of the major developmental phases namely; germination, vegetative, reproductive and grain filling stages. The composition and dimensions of the grain that eventually develop are also affected (Champagne 2004). Rice cultivars are typically different in many physical characteristics and these include kernel size, length:width ratio, kernel shape and kernel hardness (Pomeranz and Webb 1985). The rice grain (rough rice or paddy) consists of an outer protective covering (husk) and inner edible part (brown, cargo, dehulled or dehusked rice) (Juliano and Bechtel 1985). Rice is marked as paddy, brown rice or polished rice. The hull constitutes 16-28% of the rough rice, while the pericarp constitutes 1-2%, aleurone, nucellus and seed-coat 4-6%, germ 1%, scutellum

2%, and endosperm 90 to 91% of the brown rice (Juliano 1972). The predominant type of rice consumed worldwide is the polished or white rice (Nanri et al. 2010) and it is obtained by subjecting paddy rice through a series of mechanized unit operations (rice milling) during when the embryo, pericarp, seed-coat, testa and aleurone layer are removed. 6

Consequently, there is considerable loss of nutrients such as protein, fibre, ash, vitamins, fat, and polyphenols, leaving the endosperm to be largely starch (Shobana et al. 2011).

White rice has lower total dietary fiber content (0.7-2%) than brown rice (3-4%) due to the removal of the bran layer (Juliano 1985). Using apparent amylose content as a basis, milled rice cultivars may be generally classified into waxy (0-2%), very low (2-12%), low

(12-20%), intermediate (20-25%) or high (25-33%) (Lawal et al. 2011). The International

Standards organization on the basis of length:width ratio, classifies milled rice kernels into slender (>3.0), medium (2.1-3.0), bold (1.1-2.0) or round (≤1.0).

2.2 Proximate Composition of Rice

The crude protein content of milled rice has been reported to be approximately 7% and it is considered to be a major factor in determining the texture, pasting capacity and sensory characteristics of rice. Rice protein consists of glutelin (80%), globulin (12%), albumin

(5%) and prolamin (3%) (Juliano 1985). The lipid content of rice, which can be up to 15%, is mainly in the bran layer, however, about 1.5 to 1.7 % is present in milled rice (Juliano and Goddard 1986). The B vitamins, mainly thiamine, riboflavin, niacin, folate and B6 are concentrated in the bran layers and no vitamins A, D or C are present in the rice grain

(FAO 1954). Minerals such as calcium, magnesium and phosphorus are present along with some traces of iron, copper, zinc and manganese (Yousaf 1992). The nutrient data for white rice (medium-grain, raw unenriched) and brown rice (medium-grain, raw) is presented in Table 2.1.

Table 2.1. Nutritional Composition of milled rice Value/100g White rice Brown rice (medium-grain, raw, (medium-grain, Nutrient Unit unenriched) raw) Water g 12.89 12.37 Energy kcal 360 362 Protein g 6.61 7.5 Total lipid (fat) g 0.58 2.68 Carbohydrate, by difference g 79.34 76.17 Calcium, Ca mg 9 33 Iron, Fe mg 0.8 1.8 Magnesium, Mg mg 35 143 Phosphorus, P mg 108 264 Potassium, K mg 86 268 Sodium, Na mg 1 4 Zinc, Zn mg 1.16 2.02 Thiamin mg 0.07 0.413 Riboflavin mg 0.048 0.043 Niacin mg 1.6 4.308 Vitamin B-6 mg 0.145 0.509 Folate µg 9 20 Fatty acids, total saturated g 0.158 0.536 Fatty acids, total monounsaturated g 0.181 0.971 Fatty acids, total polyunsaturated g 0.155 0.959 Source: USDA-National Nutrient Database for Standard Reference Release 27 (2014)

2.3 Rice Starch

Normal rice starch contains about 70-80% amylopectin and 20-30% amylose (Pérez and

Bertoft 2010) and its crystalline structure exhibits A-type X-ray diffraction pattern. The granule size of rice starch has been reported to be among the smallest in cereal grains

(Vandeputte and Delcour 2004). Rice starch possesses some unique characteristics such as bland taste (flavor), smooth texture, creaminess, spreadability, acid resistance, freeze-thaw stability of pastes, hypoallergenicity and a wide range of amylose:amylopectin ratios.

These qualities make it valuable in the formulation of food products (Mitchell 2009; Lawal et al. 2011).

2.3.1 Morphology of Rice Starch Granule

Rice starch exists as compound polyhedral, oval, irregular, angular, or smooth shaped granules and measures between 3 and 8 µm in size. Starch granules are tightly packed in the endosperm cells within an amyloplast (Poochinya et al. 2008). The granule size of starch can be measured by a variety of techniques, which include laser light scattering, field flow fractionation and microscopy (light microscopy and scanning electron microscopy

(SEM) (Lindeboom et al. 2004), but SEM seems to be the most frequently used technique because, in addition to size it also provides information on the morphology and the characteristics of the surface of the granule (Chmelik 2001). Wang et al. (2012) studied the granular characteristics of starches separated from ten Chinese rice cultivars using scanning electron microscope and reported that the rice starch granules were angular and polygonal, with characteristic dimensions in the range of 3-8 µm. Additionally the granule size distribution of the different rice starches exhibited a unimodal shape with a mode diameter 9

of about 10 µm. Zhang et al. (2010) also used SEM to investigate the morphology and physicochemical properties of mechanically activated rice starches from three non-waxy rice cultivars (indica, japonica and glutinous rice from China) and reported an average granule size of 4.16-5.00 µm. The granules had polyhedron shapes with pronounced edges and their surfaces were smooth.

Singh et al. (2006) also reported rice starch granular size ranging between 1.5 and 5.8 mm when the relationships between physicochemical, morphological, thermal and rheological properties of rice starches starch separated from 19 different indica rice cultivars was studied using SEM. Some rice starch samples from Indian rice cultivars studied by Sodhi and Singh (2003) showed the presence of hexagonal granules in addition to the normal pentagonal shaped granules having size in the range of 2.4-5.4 µm. Scanning electron microscopy was used to determine the morphology of the starch granules isolated from five improved rice varieties from West Africa. Generally, the starches were polyhedral in appearance with irregular shapes and the granule sizes varied from 1.5 to 6.1 µm. These values are similar to those reported by Singh et al. (2006) and Sodhi et al. (2003), but slightly smaller than that reported by Li and Yeh (2001) for Taiwan rice starches where lazer particle size analyzer was used to study the granule size distribution.

2.3.2 Starch Structures

Most starch granules are made up of alternating amorphous and semi-crystalline shells which are between 100 and 400 nm thick at the lowest level of structure (Gallant et al.

1997). These structures are termed “growth rings” and are approximately 400 nm apart in rice starch as measured by atomic force microscopy (Dang and Copeland 2003). X-ray 10

diffraction investigations have indicated a periodicity of about 9-10 nm within the granule and the periodicity is interpreted as due to crystalline and amorphous lamellae found within the semi-crystalline shells (Blanshard et al. 1984). Most cereal starches give A-type crystalline pattern, while some tubers, such as potato rhizomes and cereal starches rich in amylose, yield the B-pattern while legumes give C-patterns (Pérez and Bertoft 2010).

Starch is made up of two types of macromolecules namely, amylose and amylopectin.

Figure 2.1. Structure of Amylose and Amylopectin

Amylose forms about 15-35% of the starch and it is a linear molecule composed of D- glucose units connected through α-(1,4)-linkages and the molecular size, shape and structure vary with the botanical origin (Tester et al. 2004). Amylose crystallizes readily in solution, in the form of left-handed double helices that are packed in a parallel fashion forming either A- or B-type allomorphs (Imberty et al. 1991). Amylose also forms complexes with ligands such as iodine, fatty acids and alcohols (Bail et al. 1999). The dimensions of the single stranded helices that are formed depend on the type of ligand molecule and the crystalized single helices are known as V-types. Amylopectin on the other hand is a highly branched molecule with α-(1,4)- linked D-glucose backbones and about 5% α-(1,6)-linked branches (Pérez and Bertoft 2010; Hizukuri and Takagi 1984).

There are two main types of chains in amylopectin, namely short chains with CL 6~36 and long chains with CL > 36 (Bertoft et al. 2008). Chains that do not carry other chains are known as A-chains, whereas chains that carry other chains are B-chains. The C-chain carries the sole reducing end-group of the macromolecule (Peat et al. 1952). All A-chains are external chains (completely situated outside the outermost branches) while all B-chains consist of an external chain segment and an internal part. On the basis of size-distribution of the internal unit chains, (the internal part of the B-chains) Bertoft et al. (2008) divided the structure of amylopectin was into four characteristic types. Type 1 amylopectin has very little long B-chains and a broad size-distribution of short B-chains while type 2 has considerably more long B-chains (especially B2-chains) and a narrower size distribution of the short chains. Type 3 has a little more long chains compared to type 2, whereas type 4 has considerably more long B-chains (particularly B3-chains). Cereals characteristically belong to type 1 (e.g., barley, oat, and rye) and type 2 (e.g., maize and rice), whereas legumes, roots and tubers are of type 3 (e.g., cassava and mung bean) and type 4 (e.g., potato and canna) (Bertoft et al. 2008). These structural groups have been shown to correlate with thermal properties of starches (Vamadevan et al. 2013). In addition to the unit chain profile of amylopectin, its internal chain structure, which describes the organization of chains in amylopectins, has also received considerable attention. Internal chain structure is obtained from limit dextrins, mainly β-limit and φ,β-limit dextrins. β- amylase cleaves the external chains from amylopectins resulting in very short A-chain stubs in the form of maltosyl or maltotriosyl residues. Phosphorylase from rabbit liver produces φ-limit dextrins that results in the production of glucose 1-phosphate. Perez and

Bertoft (2010) reported that in φ,β-limit dextrins, all A-chains are reduced into 4 or 2

residues irrespective of the length of the original chain length. In φ,β-limit dextrins, DP 3-

7 was also classified as fingerprint B-chains based on the fact that they were unique to certain samples.

The internal structure of amylopectins from diverse botanical sources have been reported on over the past years (Kalinga et al. 2013; Bertoft 2013; Zhu et al. 2011; Kong et al. 2009;

Laohaphatanaleart et al. 2009; Bertoft et al. 2008; Hanashiro et al. 2005; Klucinec and

Thompson 2002; Bertoft and Koch 2000; Takeda et al. 1987 and Hizukuri 1985).

However, so far no information on the structural characteristics of African rice starch has been found in literature.

In the study of the cluster structure of amylopectin over time, different models have emerged. The Hizukuri model (Hizukuri 1985), also known as the "classical" model is the most referred one. In the cluster model, the clusters are interconnected by long chains which participate in the formation of the crystalline and amorphous lamellae. In the two- directional backbone model suggested by Bertoft et al. (2012a), the cluster chains are anchored perpendicularly to the long chains which are mainly present in the amorphous region and do not participate in the formation of the crystalline lamellae. A cluster refers to a group of chains, in which the branches are separated by internal chain segments shorter than nine glucosyl residues (Bertoft, 2007). In addition to the large units of clusters, building blocks of clusters have also been investigated in recent times. Building blocks are basically tightly branched units found in clusters. The branching density in building blocks is known to be different for starches of different botanical origin (Bertoft, 2012b).

2.3.3 Cooking characteristics

Characterization of quality attributes of different types of rice varieties is important for the improvement in rice quality and rice-based products. The eating and cooking qualities of rice is very essential to consumers and breeders because most rice is cooked and consumed directly. Rice eating and cooking qualities are highly related to some easily measurable physico-chemical properties. These include gelatinization temperature and pasting viscosity which reflect the starch functionality of the grain and apparent amylose content

(Singh et al. 2005; Bao 2012). However, amylose content is widely recognized as the most important determinants for different rice products (Juliano 1998). Rice with similar amylose content can have differences in their eating qualities but this can be explained by differences in their amylopectin structures (Ong and Blanshard 1995). High-amylose rice cultivars are common in indica rice cultivars and are usually dry, fluffy and separate when cooked and become hard on cooling. Low-amylose cultivars which are common among temperate japonica rice cultivars are soft and sticky whereas intermediate-amylose rice is soft but not sticky (Bao 2012). Apart from the amylose content, the cooking quality of rice can also be influenced by components such as: proteins, lipids or amylopectin (Cai et al.

2011). The texture and physical appearance of cooked rice may vary depending on the method of cooking (Leelayuthsoontorn andThipayarat, 2006). The texture of cooked rice is related to its amylose content and the fine structure of amylopectin. The intra- and/or inter- molecular interactions of starch with other components in rice such as proteins, lipids and non-starch polysaccharides affect the texture of rice (Prasert and Suwannaporn, 2009).

Mohapatra and Bal (2006) reported that the degree of milling, amylose content and grain thickness affected the cooking qualities as well as textural attributes of three indica rice 14

cultivars. Additionally, the grain thickness of the rice cultivars was found to be negatively correlated with optimum cooking time, adhesiveness and positively with water uptake ratio, volume expansion ratio, length expansion ratio, hardness, and cohesiveness. Singh et al. (2005) worked on the physico-chemical, cooking and textural properties of 23 milled

Indian rice cultivars and reported that cooking time showed a negative correlation with amylose content and a positive correlation with bulk density of the milled rice. Gruel solid loss showed a significant positive correlation with the amylose content and was negatively correlated with cooking time. The rice cultivars with higher cooking time showed lower gruel solid loss. Textural parameters had a positive correlation with amylose and negative correlation with cooking time. After studying the effect of processing on the nutritionally important starch fractions in six rice varieties, Rashmi and Urooj (2003) concluded that cooking methods (pressure cooking, boiling, straining and steaming) influence the nutritionally important starch fractions (rapidly digestible starch, slowly digestible starch, resistant starch) in rice varieties.

2.3.4 Thermal Properties

Starch undergoes an irreversible transition termed gelatinization when it is heated in the presence of excess water. The process is endothermic and it results in the disruption of molecular order within the starch granules. Double helical and crystalline structures are disrupted in starches during the process of gelatinization. Increase in temperature causes the crystallites to break apart, and then to undergo hydration resulting in several changes such as granular swelling, native crystalline melting, loss of birefringence, and starch solubilization (Atwell et al. 1988). When cooled, the starch molecules tend to recrystallize

(retrogradation), and the process may be associated with water separating from the gel

(syneresis) (Yeh and Yeh 1993; Hoover and Manuel 1995). Differential scanning calorimetry (DSC), X-ray scattering, light scattering, optical microscopy, thermo- mechanical analysis (TMA) and NMR spectroscopy are some techniques that have been employed to study structural changes underlying gelatinization (Jenkins and Donald 1998).

The use of the DSC to measure melting transition temperatures and enthalpy of gelatinization involves the heating of starch (or flour) in excess water. The temperature range and heating rate are programmed and thermograms are recorded with an empty pan as a reference. Thermal properties typically reported using DSC include onset gelatinization (To), peak (Tp), and conclusion (Tc) temperatures, peak height index (PHI), gelatinization range (R), and enthalpy (ΔH). Crystallite quality and the overall crystallinity of the starch are measured by the peak temperature (Tp) and the enthalpy of gelatinization

(ΔHg) respectively (Tester and Morrison 1990). Onset temperature (To) and conclusion temperature, (Tc) determine the boundaries of the different phases in a semi-crystalline material like starch (Biliaderis et al. 1986). The enthalpy of gelatinization (ΔHg) is estimated by integrating the area under the endothermic peak produced as a result of the melting of amylopectin crystallites in the starches. ΔHg is expressed as J/g and it gives an indication of the rupturing of the links between and within the amylopectin double helices

(Cooke and Gidley 1992). Double helical and crystalline structures are disrupted in starches during gelatinization. This phase transition involves melting of crystals which is illustrated by DSC endotherms for various native starches (Jacobs et al. 1995). Li and Yeh

(2001) studied the relationships between thermal, rheological characteristics and swelling power of starches from different botanical sources that included starch isolated from

Taiwanese rice. They reported 57.7 ºC and 65.1 ºC for To and Tp respectively and ΔHg of

11.5 J/g. Starches isolated from five indica rice cultivars with varying amylose content, were studied by Sodhi and Singh in 2003 using DSC. Temperature ranges of 66.33-67.26

ºC, 69.74-71.94 ºC and 74.08-78.04 ºC representing To, Tp and Tc were reported while ΔHg ranged from 8.16-11.88 J/g. A high degree of crystallinity, which provides structural stability and makes the granule more resistant to gelatinization accounts for the high transition temperatures (Barichello et al. 1990). They also suggested that the differences in transition temperatures and ΔHg of the starches from the different rice cultivars may be attributed to differences in amylose contents and their granular structures. Additionally, they suggested that the starch samples with higher To, Tp, Tc values may be attributed to the compact nature of their small starch granules and higher degree of molecular order of the granules (Krueger et al. 1987). Further to this the presence of amylose lowers the melting point of crystalline regions and the energy for starting gelatinization because amylopectin plays a major role in starch granule crystallinity. More energy is needed to initiate melting in the absence of amylose-rich amorphous regions (Kreuger et al. 1987).

Recrystallization of amylopectin branch chains has been reported to occur in a more ordered manner in native starches than in stored starch gels and this may be the reason for the occurrence of ΔH (retrogradation) at a lower temperature range than ΔH

(gelatinization) (Ward et al. 1994). Varying ΔH (retrogradation) values of different starches may be due to differences in amylose-amylopectin ratios, granular structures and phosphate esters (Hizukuri et al. 1983; Kasemsuwan et al. 1995). Yamin et al. (1999) also reported that amylopectin and intermediate materials play a significant role in starch retrogradation during refrigerated storage and that the intermediate materials having longer 17

chains than amylopectin may also form longer double helices during re-association under refrigerated storage conditions.

2.3.5 Pasting Properties

When starch is heated continuously in excess of water with stirring, the granules swell irreversibly and burst as a result of breakage of the native starch structure. Amylose then leaches out and the granules disintegrate to form a viscous material called paste (BeMiller

2007). Pasting properties of starch are key pointers to how a particular starch sample will behave during processing and they can be used to assess the suitability of starch as a functional ingredient in food and other products (Wani et al. 2012). The physical and chemical characteristics of starch such as amylose:amylopectin ratio, granule size distribution, average granule size and mineral content influence starch properties (Madsen and Christensen 1996; Wani et al. 2010). Amylose and amylopectin play significant roles in the gelatinization and pasting properties of rice starch with amylose being the major factor influencing the physicochemical properties. While starch swelling is a property of amylopectin, amylose tends to restrict it (Tester and Morrison 1990). Starch pasting characteristics are usually determined by employing a Brabender visco amylograph , rapid visco analyzer (RVA), rotational rheometers or other viscometers capable of continuously recording change in viscosity with temperature (Wickramasinghe and Noda 2008;

Tukomane and Varavinit 2008; Lin et al. 2009; Park et al. 2007; Li et al. 2008). The important parameters commonly measured include pasting temperature, peak time, peak viscosity (PV), trough viscosity, final viscosity (FV), setback viscosity (SB), and breakdown viscosity (BD). Pasting temperature refers to the temperature at the onset of

increase in viscosity and it provides an indication of the minimum temperature required to cook a given sample while peak viscosity reflects the water-binding capacity of starch and often correlates with final product quality because the swollen and collapsed granules relate to the texture of cooked starch (Wani et al. 2012). Breakdown viscosity is a measure of the ease of disrupting swollen starch granules and gives an indication of the degree of stability during cooking of starch (Adebowale and Lawal 2003). Trough viscosity (hot paste viscosity) is the viscosity recorded at the maximum temperature of the RVA test. Setback viscosity, a measure of the tendency of starch paste to retrograde is an indicator of final product texture and it is linked to syneresis (weeping) during freeze-thaw cycles (Wani et al. 2012). Setback is also used to describe the increase in viscosity that occurs on cooling a pasted starch (Fisher and Thompson, 1997). Final viscosity indicates the ability of the material to form a viscous paste or gel after cooking and cooling. The rheology and functional properties of starches isolated from five improved rice varieties from West

Africa was studied by Lawal et al. (2011). The peak viscosities of the starches were reported to range between 147.48 and 209.17 RVU while hot paste (trough) viscosity was from 123.83-157.08 RVU. The high values of setback 58.83-122.42 RVU and breakdown viscosities 23.75-65.75 RVU observed in the study were attributed to the high amylose content of NERICA starch samples. Similarly, upon studying the functional and physicochemical properties of flours and starches of African rice cultivars, Falade et al. (2014) also reported high setback 150-174 RVU and breakdown viscosities 115.6-162.3 for two

African rice starches and attributed it to high amylose content of the starches. When the breakdown in viscosity is high, the ability of the starch sample to withstand heating and shear stress during cooking is low (Adebowale et al. 2005). Peak viscosity, trough

viscosity, breakdown, final viscosity, setback and gelatinization temperature range of values of 258.1-325.1 RVU, 160.6-178.2 RVU, 80.9-153.6 RVU, 307.7-322.3 RVU,

136.2-158.1 and 78.2-80.2 ºC respectively, were reported for a study on grain qualities and their genetic derivation of seven New Rice for Africa (NERICA) varieties (Masahiro et al.

2008).

2.3.6 Swelling and Solubility Characteristics

The swelling capacity is a measure of the ability of starch to hydrate under specific conditions such as temperature and water availability (Lawal et al. 2011). Starch swelling occurs alongside loss of birefringence and precedes solubilization (Singh et al. 2004).

Amylose-lipid complexes inhibit swelling, and the swelling behavior of cereal starch is primarily related to amylopectin structure Tester and Morrison (1990a). It has been reported that swelling power and solubility can be used to measure the extent of interaction between starch chains and within the amorphous and crystalline domains of the starch granule (Ratnayake et al. 2002). A greater swelling capacity is an indication of weaker binding forces in the starch granules (Hoover et al. 1996). After studying five waxy and ten normal rice starches for structure-function relationships, Vandeputte et al. (2003) reported that the swelling power of one waxy and three normal rice starches was a function of temperature. In the first swelling step, (55-85 ºC) amylose did not influence the swelling of normal rice starches but granule swelling was increased by short amylopectin chains

(DP 6 to 9) between 55 ºC and 85 ºC. The second swelling step (95-125 ºC) was influenced by amylose leaching (decreased swelling power). Lawal et al. (2011) reported of increase in swelling capacity and solubility with an increase in temperature for starches

isolated from five improved rice varieties from West Africa. When temperature was increased in the presence of water, the mobility of starch molecules increased and as a result the binding forces were weakened. This process enhanced water diffusion into the starch granules and leaching of the soluble components of the starch granules thereby leading to improved starch solubility. Amylose has been reported to constitute the larger percentage of the amorphous component of the starch granules where water penetration into the granule is more pronounced. Consequently, the highest swelling and solubility of

NERICA starch sample in this study was attributed to its high amylose component. In a study on functional and physico-chemical properties of flours and starches of African rice cultivars, Falade et al. (2014) reported swelling power values of 12.35 and 18.56 mL/g and solubility values of 5.72 and 8.24% for two African rice cultivars. For sixty-three non waxy starches studied, Tan and Corke (2002) reported a swelling power range of 11.4-23.5 mL/g.

2.4 In vitro Digestibility and Glycemic Attributes of Rice

The glycemic index (GI) and resistant starch (RS) content are two important indicators of starch digestibility. Glycemic index is a numerical value, which indicates the postprandial

(after meal) blood glucose response of a test food rich in carbohydrate (i.e. a scale that ranks carbohydrate-rich foods by their ability to raise blood glucose level) (Jenkins et al.

1987). A GI value greater than 70 is considered high while values between 56 and 69 are regarded as medium. Values less than 56 are low. Glycemic load (GL) is a function of carbohydrate intake and Glycemic Index. GL is the product of the grams of carbohydrates in a serving of food and the GI of the food (Willett et al. 2002). For one serving of a food,

a GL greater than 20 is considered high while a GL of 11-19 is considered medium. A GL of 10 or less is considered low. Using in vitro digestion as a basis of classification, Englyst et al. (1992) categorized starch into three types, namely rapidly digestible (RDS), slowly digestible (SDS) and resistant starch (RS) categories. Rapidly digestible starches refer to starches that are digested after 20 minutes while slowly digestible starches are those digested between 20 minutes and 120 minutes. Resistant starches are those that remain undigested after 120 minutes. Enzymatic digestibility is affected by the morphological, textural and rheological properties of food. Other factors that affect starch digestibility include presence of proteins, lipids, anti-nutrients/inhibitors and non-starch polysaccharides

(Singh et al. 2010). Non-starchy substances such as proteins or lipids may limit the rate of enzymatic hydrolysis by blocking the adsorption sites and therefore influence binding of enzyme (Oates 1977). Hamaker and Bugusu (2003) suggested that protein fractions such as albumins, globulins and glutenins, attach to the proteins bodies and form a matrix around starch granules thereby acting as a barrier to amylases. Lehmann and Robin (2007) reported that the crystalline structure of starches influences their hydrolysis. Amylose forms complexes with lipids which alter the behavior of starches considerably. These amylose-lipid complexes are known to be resistant to enzymatic attack (Holm et al. 1983).

Chung et al. (2011) studied the in vitro digestibility of rice starches differing in amylose content and reported that rice starch from long grain having amylose content of 27% had

RDS content of 39% while glutinous rice starch having 4% amylose had 71% RDS. Miller et al. (1992) reported the GI of white and brown types of three commercial rice varieties

(Doongara, Calrose and Pelde). Doongara with amylose content of 28% had GI of 64 for the white rice and 66 for the brown rice. The other two varieties Calrose and Pelde had GI

values of 83 and 93 for white rice and 87 and 76 for brown rice, respectively. Sato et al.

(2010) reported a GI of 89 for Japanese Koshikari rice and Frei et al. (2003) reported higher GIs (>90) for waxy rice cultivars compared to non-waxy cultivars. Ranawana et al.

(2009) reported that basmati rice cooked for longer times caused greater glycemic response.

2.5 Conclusion

It is believed that O. glaberrima has not been fully exploited in terms of the many traits it possesses for varietal improvement. It has been reported that there is a high level of genetic diversity in O. glaberrima accessions of in both apparent amylose content and pasting properties (Gayin et al. 2015). This diversity suggests that there may be a valuable resource of potentially very important gene pool that may be of benefit in rice breeding programs in this region. A detailed molecular analysis of promising accessions from such as those with a high tendency to retrograde (high setback and potentially slower rate of digestion due to retrogradation) or form viscous paste/gel after cooking and cooling (high final viscosities) will ensure the use of the diversity in the accessions to assist in future development of acceptable commercial cultivated rice varieties that have the desired amylose content and pasting characteristics to meet consumer preferences and benefit resource poor farmers and to ultimately reduce the over-dependency on rice imports by

African countries.

2.6 Significance of project

This research work will contribute to knowledge on rice by generating information on the cooking quality and glycemic characteristics and physico-chemical properties and molecular structure of starches from African rice accessions. The goal of this research is to ultimately promote and add value to African rice and to identify African rice accessions exhibiting promising traits for further improvement.

2.7 References (Chapters 1 and 2)

Adebowale, K. O., and Lawal, O. S. 2003. Microstructure, physicochemical properties and

retrogradation behavior of Mucuna bean (Mucuna pruriens) starch on heat moisture

treatments. Food Hydrocoll. 17: 265-272.

Atwell, W. A., Hood, L. F., Lineback, D. R., Varriano-Marston, E., and Zobel, H. F. 1988.

The terminology and methodology associated with basic starch phenomena. Cereal

Foods World 33: 306-311.

Bail, P. L., Bizot, H., Ollivon, M., Keller, G., Bourgaux, C., and Buléon, A., 1999.

Monitoring the crystallization of amylose-lipid complexes during maize starch

melting by synchrotron Xray diffraction. Biopolym. 50: 99-110.

Bao, J., Corke, H., and Sun, M. 2004. Genetic diversity in the physicochemical properties

of waxy rice (Oryza sativa L) starch. J.Sci.Food Agric. 84: 1299-1306.

Bao, J., Jin, L., Xiao, P., Shen, S., Sun, M., and Corke, H. 2008. Starch physicochemical

properties and their associations with microsatellite alleles of starch synthesizing

genes in a rice RIL population. J.Agric.Food Chem. 56: 1589-1594.

Bao, J. S. 2012. Toward understanding the genetic and molecular bases of the eating and

cooking qualities of rice. Cereal Foods World 57: 148-156.

Barichello, V., Yada, R. Y., Coffin, R. H., and Stanley, D. W. 1990. Low-temperature

sweetening in susceptible and resistant potatoes - Starch structure and composition.

J. Food Sci. 55: 1054-1059.

BeMiller, J. N. 2007. Carbohydrate chemistry for food scientists. 2nd ed. St. Paul, Minn.:

AACC Intl.: 389.

Bertoft, E. 2013. On the building block and backbone concepts of amylopectin structure.

Cereal Chem. 90: 294-311.

Bertoft, E., Koch, K., and Åman, P. 2012a. Structure of building blocks in amylopectins.

Carbohydr. Res. 361:105-113.

Bertoft, E., Koch, K. and Åman, P. 2012b. Building block organisation of clusters in

amylopectin from different structural types. Int.J.Biol.Macromol. 50: 1212-1223.

Bertoft, E., Piyachomkwan, K., Chatakanonda, P., and Sriroth, K. 2008. Internal

unit chain composition in amylopectins. Carbohydr. Polym. 74: 527-543.

Bertoft, E. 2007. Composition of clusters and their arrangement in potato amylopectin.

Carbohydr. Polym. 68: 433-446.

Bertoft, E., and Koch, K. 2000. Composition of chains in waxy-rice starch and its

structural units. Carbohydr. Polym. 41: 121-132.

Bhattacharya, P., Gosh,U., Chowdhuri, U. Roy, Chattopadhyay, P., and Gangopadhyay, H.

2004. Effects of different treatments on physico-chemical properties of rice starch.

J.Sci.Indus.Res. 63: 826-829.

Biliaderis, C. G., Page, C. M., Maurice, T. J., and Juliano, B. O. 1986. Thermal

characterization of rice starches: a polymeric approach to phase transitions of

granular starch. J. Agric. Food Chem. 34: 6-14.

Blanshard, J. M. V., Bates, D. R., Muhr, A. H., Worcester, D. L., and Higgins, J. S. 1984.

Small-angle neutron-scattering studies of starch granule structure. Carbohydr.

Polym. 4:427-442.

Bryant, R. J., Kadan R. S., Champagne, E. T., Vinyard B. T., and Boykin, D. 2001.

Functional and digestive characteristics of extruded rice flour. Cereal Chem. 78:

131-137.

Cai, Y., Liu, C., Wang, W., and Cai, K. 2011. Differences in physicochemical properties

of kernels of two rice cultivars during grain formation. J. Sci. Food Agric. 91:

1977-1983.

Carney, J. A. 2001. Black rice: the African origins of rice cultivation in the Americas.

Harvard University Press, Cambridge, Massachusetts.

Champagne, E.T. 2004. Rice: chemistry and technology, 3rd Edition. St Paul, MN, USA,

American Association of Cereal Chemists: 25-32.

Chmelik, J. 2001. Comparison of size characterization of barley starch granules

determined by electron and optical microscopy, low-angle laser light scattering and

gravitational field flow fractionation. J. Inst. Brewing 107:11-17.

Chung, H.J., Liu, Q., Lee, L., and Wei, D. 2011. Relationship between the structure,

physicochemical properties and in vitro digestibility of rice starches with different

amylose contents. Food Hydrocoll. 25: 968-975.

Cooke, D., and Gidley, M. J. 1992. Loss of crystalline and molecular order during starch

gelatinisation: Origin of the enthalpic transition. Carbohydr. Res. 227: 103-112.

Cuevas, R. P., and Fitzgerald, M. 2007. Linking starch structure to rice cooking quality.

IREC Farmers’ Newsletter. No. 177. Summer 2007-08: 16-17.

Dang, J. M. C., and Copeland, L. 2003. Imaging rice grains using atomic force

microscopy. J. Cereal Sci. 37: 165-70.

Diagana, B., Akindès, F., Savadogo, K., Reardon, T., and Staatz, J. 1999. Effects of the

CFA Franc devaluation on urban food consumption in West Africa: overview and

cross-country comparisons. Food Policy. 24: 465-478.

Englyst, H. N., Kingman, S., and Cummings, J. 1992. Classification and measurement of

nutritionally important starch fractions. Eur. J. Clin. Nutr. 46:33-50.

Falade, K. O., Semon, M., Fadairo, O.S., Oladunjoye, A. O., Orou, K. K. 2014. Functional

and physico-chemical properties of flours and starches of African rice cultivars.

Food Hydrocoll. 39: 41-50.

FAO. 1954. Rice and rice diets-a nutritional survey, revised edition. Rome, FAO: 78.

FAO. 2012. Rice Market Monitor - Volume XV-Issue No. 2 April 2012 Ed.

FAO. 2014. Rice Market Monitor - Volume XV1I-Issue No. 1 April 2014 Ed.

Falade, K. O., Semon, M, Fadairo, O.S., Oladunjoye, A, O., and Orou, K. K. 2014.

Functional and physico-chemical properties of flours and starches of African

rice cultivars. FoodHydrocoll. 39: 41-50.

Fisher, D. K., and Thompson, D. B. 1997. Retrogradation of maize starch after thermal

treatment within and above the gelatinization temperature range. Cereal Chem. 74:

344-351.

Frei, M., Siddhuraju, P., and Becker, K. 2003. Studies on the in vitro starch digestibility

and the glycemic index of six different indigenous rice cultivars from the

Philippines. Food Chem. 83: 395-402.

French D. 1984. Organization of starch granules. In: Whistler, R. L, Bemiller, J. N.,

Paschal, E. F. editors. Starch chemistry and technology. New York, N.Y.:

Academic Press. : 183-247.

Fuller, D., Sato, Y., Castillo, C., Qin, L., Weisskopf, A., Kingwell-Banham, E., Song, J.,

Ahn, S., and Etten, J. 2010. Consilience of genetics and archaeobotany in the

entangled history of rice. Archaeol. Anthropol. Sci. 2: 115-131.

Gallant, D. J., Bouchet, B., and Baldwin, P. M. 1997. Microscopy of starch: evidence of a

new level of granule organization. Carbohydr. Polym. 32: 177-191.

Gayin, J., Chandi, G. K., Manful, J., and Seetharaman, K. 2015. Classification of rice

based on statistical analysis of pasting properties and apparent amylose content:

the case of Oryza glaberrima accessions from Africa. Cereal Chem. 92: 22-28.

Ghesquière, A., Albar, L., Lorieux, M., Ahmadi, N., Fargette, D., Huang, N., McCouch, S.

R., and Notteghem, J. L. 1997. A Major Quantitative Trait Locus for Rice Yellow

Mottle Virus Resistance Maps to a Cluster of Blast Resistance Genes on

Chromosome12. Phytopathology 87: 1243-1249.

Hanashiro, I., Matsugasako, J., Egashira, T., and Takeda, Y. 2005. Structural

characterization of long unit-chains of amylopectin. J. Appl. Glycosci. 52: 233-

237.

Hirsch, R. 1998. La riziculture dans les pays de l’UEMOA: de la devaluation a la

liberalisation 1993-1997, AFD, Paris.

Hizukuri, S., Shirasaka, K., and Juliano, B. O. 1983. Phosphorus and amylose-branching

in rice starch. Starch/Stärke 35: 348-350.

Hizukuri, S., and Takagi, T. 1984. Estimation of the molecular weight for amylase by the

low angle laser-light-scattering technique combined with high-performance

chromatography. Carbohydr. Res. 134: 1-10.

Hizukuri, S. 1985. Relationship between the distribution of the chain length of

amylopectin and the crystalline structure of starch granules. Carbohydr. Res.

141: 295-306.

Holm, J., Bjorck, I., Ostrowska, S., Eliasson, A., Asp, N., and Larsson, K. 1983.

Digestibility of amylose-lipid complexes in-vitro and in- vivo. Starch/Stärke. 35:

294-297.

Hoover, R., and Manuel, H. 1995. A comparative study of the physicochemical properties

of starches from two lentil cultivars. Food Chem. 53: 275-284.

Hoover, R., Sailaja, Y., and Sosulski, F. W. 1996. Characterization of starches from wild

and long-grain brown rice. Food Res. Int. 29: 99-107.

Imberty, A., Buléon, A., Tran, V., and Pérez, S. 1991. Recent advances in knowledge of

starch structure. Starch/Stärke. 43: 375-384.

Jacobs, H., Eerlingen, R. C., Clauwaert, W., and Delcour, J. A. 1995. Influence of

annealing on the pasting properties of starches from varying botanical sources.

Cereal Chem. 72: 480-487.

Jenkins, P. J., and Donald, A. M. 1998. Gelatinization of starch: a combined

SAXS/WAXS/DSC and SANS study. Carbohydr. Res. 308: 133-147.

Jenkins, D., Thorne, M.J., Wolever, T., Jenkins, A.L., Rao, A.V., and Thompson, L.U.

1987. The effect of starch-protein interaction in wheat on the glycemic response

and rate of in vitro digestion. Am. J. Clin. Nutr. 45: 946-951.

Jones, M. P., Dingkuhn, M., Aluko, G. K., and Semon, M. 1997. Interspecific Oryza sativa

L. × O. glaberrima Steud. progenies in upland rice improvement. Euphytica 92:

237-246.

Juliano, B.O. 1998. Varietal impact on rice quality. Cereal Foods World 43: 207-222.

Juliano, B.O. 1992. Structure, chemistry and function of the rice grain and its fraction.

Cereal Foods World. 41: 827-832.

Juliano, B. O., and Goddard, M. S. 1986. Cause of varietal difference in insulin and

glucose responses to ingested rice. Plant Foods Hum. Nutr. 36: 35-41.

Juliano, B. O., and Bechtel, D. B. 1985. The rice grain and its gross composition. In Rice

chemistry and technology, St Paul, MN, USA, Am. Assoc. Cereal Chem.: 17-57.

Juliano, B. 0. 1972. The rice caryopsis and its composition. In D.F. Houston, ed. Rice

chemistry and technology, St Paul, MN, USA. Am. Assoc. Cereal Chem.: 1674.

Kalinga, D. N., Waduge, R., Liu, Q., Yada, R. Y., Bertoft, E., and Seetharaman, K. 2013.

On the differences in the granular architecture and starch structure between pericarp

and endosperm wheat starches. Starch/Stärke. 65: 791-800.

Kasemsuwan, T., Jane, J., Schnable, P., Stinard, P., and Robertson, D. 1995.

Characterization of the dominant mutant amylose-extender (Ae1–5180) maize.

Cereal Chem. 75: 457-464.

Khush, G.S. 1997. Origin, dispersal, cultivation and variation of rice. Plant Mol. Biol. 35:

25-34.

Klucinec, J.D., and Thompson D.B. 2002. Structure of amylopectins from ae-containing

maize starches. Cereal Chem. 79: 19-23.

Kong, X., Corke, H., and Bertoft, E. 2009. Fine structure characterization of amylopectins

from grain amaranth starch. Carbohydr. Res. 344: 1701-1708.

Krueger, B. R., Knutson, C. A., Inglett, G. E., and Walker, C. E. 1987. A differential

scanning calorimetry study on the effect of annealing on gelatinization behavior of

corn starch. J. Food Sci. 52: 715-718.

Laohaphatanaleart, K., Piyachomkwan, K., Sriroth, K., Santisopasri, V., and Bertoft, E.

2009. A study of the internal structure in cassava and rice amylopectin.

Starch/Stärke. 61: 557-569.

Lawal, O. S., Lapasin, R., Bellich, B., Olayiwola, T. O., Cesàro, A., Yoshimura, M. 2011.

Rheology and functional properties of starches isolated from five improved rice

varieties from West Africa. Food Hydrocoll. 25: 1785-1792.

Leelayuthsoontorn, P., and Thipayarat, A. 2006. Textural and morphological changes

of Jasmine rice under various elevated cooking conditions.

Food Chem. 96: 606- 613.

Liu, L., Lee, G., Jiang, L., and Zhang, J. 2007. Evidence for the early beginning

(c. 9000 cal. BP) of rice domestication in China: a response. The Holocene

17: 1059-1068.

Li, J., and Yeh, A. 2001. Relationships between thermal, rheological characteristics and

swelling power for various starches. J.Food Eng. 50: 141-148.

Lin, Q., Xiao, H., Zhao, J., Li, L., and Yu, F. 2009. Characterization of the pasting, flow

and rheological properties of native and phosphorylated rice starches. Starch/

Stärke 61: 709-715.

Lindeboom, N., Chang, P. R., and Tyler, R. T. 2004. Analytical, biochemical and

physico-chemical aspects of starch granule size, with emphasis on small granule

starches: a review. Starch/Stärke. 56: 89-99.

Luh, B.S., ed. 1991. Rice Utilization,Volume II. AVI, New York.

Ma, J., and Bennetzen, J.L. 2004. Rapid recent growth and divergence of rice nuclear

genomes. Proc. Natl. Acad. Sci. U.S.A. 101: 12404-12410.

Masahiro, K., Keitaro, S., Sumiko, N., and Ken’ichi, O. 2008. Grain qualities and their

genetic derivation of 7 New Rice for Africa (NERICA) varieties. J. Agric. Food

Chem. 56: 4605-4610.

Maji, A. T., Gana, A. S., and Ukwungwu, M. N. 2010. Responses of Oryza glaberrima

accessions to rice stresses and their morphological characteristics. African

J. General Agric. 6: 229-234.

Maji, T., Singh, B. N., and Akenova, M. E. 2001. Vegetative stage drought tolerance in O.

glaberrima Steud and O. sativa L. and relationship between drought parameters.

Oryza 38: 17-23.

Marshall, W. E., Normand, F. L., and Goynes, W.R. 1990. Effects of lipid-protein removal

on starch gelatinization in whole grain milled rice. Cereal Chem. 67: 458-463.

Madsen, M. H., and Christensen, D. 1996. Changes in viscosity properties of potato

starches during growth. Starch/Stärke 48: 245-249.

Masahiro, K., Keitaro, S., Sumiko, N., and Ken’ichi, O. 2008. Grain qualities and their

genetic derivation of 7 New Rice for Africa (NERICA) varieties. J. Agric.

Food Chem. 56: 4605-4610.

Miller, J. B., Pang, E., and Bramall, L. 1992. Rice: a high or low glycemic index food?

Am. J. Clin. Nutr. 56:1034-1036.

Mitchell, J. R. 2009. Rice starches: production and properties. In: BeMiller, J., and

Whistler, R. Editors. Starch chemistry and technology. New York: Academic

Press. : 569-579.

Mohapatra, D., and Bal, S. 2006. Cooking quality and instrumental textural attributes of

cooked rice for different milling fractions. J. Food Eng. 73: 253-259.

Nanri, A., Mizoue, T., Noda, M., Takahashi, Y., Kato, M. and Inoue, M. 2010. Rice intake

and type 2 diabetes in Japanese men and women: the Japan Public Health Center-

based Prospective Study. Am. J. Clin. Nutr. 92: 1468-1477.

Nayar, N. M. 2010. The history and genetic transformation of the African rice, Oryza

glaberrima Steud. (Gramineae). Curr. Sci. 99: 1681-1689.

Ndjiondjop, M. N., Albar, L., Fargette, D., Fauquet, C., and Ghesquiere, A.1999. The

genetic basis of high resistance to yellow mottle virus (RYMV) in cultivars of the

two cultivated rice species. Plant Dis. 83: 931-935.

Oko, A. O., and Ugwu, S. I. 2010. The proximate and mineral compositions of five major

rice varieties in Abakaliki, South-Eastern Nigeria. Int. J. Plant Physiol. Biochem. 3:

25-27.

Ong, M. H., and Blanshard, J. M. V. 1995. Texture determinants in cooked, parboiled rice.

I: Rice starch amylose and the fine structure of amylopectin. J. Cereal Sci. 21: 251-

260.

Park, I. M., Ibanez, A. M., Zhong, F., and Shoemaker, C. F. 2007. Gelatinization and

pasting properties of waxy and non-waxy rice starches. Starch/ Stärke 59: 388-396.

Peat, S.,Whelan,W. J.,and Thomas,G. J. 1952. Evidence of multiple branching in waxy

maize starch. J. Chem. Soc.: 4546-4548.

Pérez, S., and Bertoft, E. 2010. The molecular structures of starch components and their

contribution to the architecture of starch granules: A comprehensive review.

Starch/Stärke 62: 389-420.

PlowRight, R. A., Coyne, D. L., Nash, P., and Jones, M. P. 1999. Resistance to the rice

nematodes Heterodera sacchari, Meloidogyne graminicola and M. incognita in

Oryza glaberrima and O. glaberrima × O. sativa interspecific hybrids.

Nematology 1: 745-752.

Pomeranz, Y., and Webb, B.D. 1985. Rice hardness and functional properties. Cereal

Foods World 30: 784-788.

Poochinya, P., Onanong, N., Yupadee, P., and Patcharee, U. 2008. Morphological changes

of rice starch during grain development. J. Microscopy Soc. Thailand. 22: 76-79. 36

Prasert, W., and Suwannaporn, P. 2009. Optimization of instant jasmine rice process

and its physicochemical properties. J. Food Eng. 95: 54-61.

Preecha, P. 2008. Variation of the waxy microsatellite allele and its relation to amylose

content in wild rice (Oryza rufipogon Griff.). Asian J. Plant Sci. 7: 156-162.

National Nutrient Database for Standard Reference Release 27. 2014. Agricultural

Research Service, United States Department of Agriculture. Proc. Nat. Aca.

Sci.101: 12404-12410.

Ranawana, D. V., Henry, C. J. K., Lightowler, H. J., and Wang, D. 2009. Glycaemic index

of some commercially available rice and rice products in Great Britain. Int. J. Food

Sci. Nutr. 60:99-110.

Rashmi, S., and Urooj, A. 2003. Effect of processing on nutritionally important starch

fractions in rice varieties. Int. J. Food Sci. Nutr. 54: 27-36.

Ratnayake, W. S., Hoover, R., and Warkentin, T. 2002. Pea starch: composition, structure

and properties-a review. Starch/Stärke 54:217-234.

Sano, Y., Sano, R., and Morishima, H. 1984. Neighbour effects between two occurring

rice species, Oryza sativa and Oryza glaberrima. J. Appl. Ecol. 21: 245–254.

Sarla, N., and Swamy, B. P. M. 2005. Oryza glaberrima: A source for the improvement of

Oryza sativa. Curr. Sci. 89: 955-963.

Sato, S., Fukumura, K., Nishiyama, A., Inoue, Y., and Konishi, T. 2010. Glycemic index

and glucose utilization of rice vermicelli in healthy subjects. Biolog. Pharm. Bull.

33: 1385-1393.

Sauphanor, B. 1985. Some factors of upland rice tolerance to stem borers in West Africa.

Insect Sci. Appl. 6: 429-434.

Shobana, S., Malleshi, N. G.,Sudha, V., Spiegelman, D., Hong, B., Hu, F. B., Willett, W.

C., Krishnaswamy, K., and Mohan, V. 2011. Nutritional and sensory profile of two

Indian rice varieties with different degrees of polishing. Int. J. Food Sci. Nutr. 62:

800-810.

Singh N, Sandhu KS, Kaur M. 2004. Characterization of starches separated from Indian

chickpea (Cicer arietinum) cultivars. J. Food Eng. 63(4): 441-449.

Singh, N., Kaur, L., Sodhi, N. S., and Sekhon, K. S. 2005. Physicochemical, cooking and

textural properties of milled rice from different Indian rice cultivars. Food Chem.

89: 253-259.

Singh, N., Kaur, L., Sandhu, K. S., Kaur, J., and Nishinari, K. 2006. Relationships

between physicochemical, morphological, thermal, rheological properties of rice

starches. Food Hydrocoll. 20: 532-542.

Singh, J., Dartois, A., and Kaur, L. 2010. Starch digestibility in food matrix: a review.

Trends Food Sci.Technol. 21:168-180.

Sodhi, S., N., and Singh, N. 2003. Morphological, thermal and rheological properties of

starches separated from rice cultivars grown in India. Food Chem. 80:99-108.

Sweeney, M., and McCouch, S. 2007. The complex history of the domestication of rice.

Ann. Bot. 100: 951-957.

Takeda, Y., Hizukuri, S., and Juliano, B. O. 1987. Structures of rice amylopectins with

low and high affinities for iodine. Carbohydr. Res. 168: 79-88.

Tan, Y., and Corke, H. 2002: Factor analysis of physiochemical properties of 63 rice

varieties. J. Sci. Food Agric. 82: 745-752.

Tester, R. F., and Morrison, W. R. 1990. Swelling and gelatinization of cereal starches I.

Effects of amylopectin, amylose and lipids. Cereal Chem. 67: 551-557.

Tester, R. F., Karkalas, J., and Qi, X. 2004. Starch composition, fine structure and

architecture (review). J. Cereal Sci. 39: 151-165.

Tsai, M., Chin-Fung, L., and Cheng-Yi, L. 1997. Effects of granular structures on the

pasting behaviors of starches. Cereal Chem. 74: 750-757.

Tukomane, T., and Varavinit, S. 2008. Classification of rice starch amylose content from

rheological changes of starch paste after cold recrystallization. Starch/Stärke 60:

292-297.

Vamadevan, V., Bertoft, E., and Seetharaman, K. 2013. On the importance of organization

of glucan chains on thermal properties of starch. Carbohydr. Polym. 92: 1653-

1659.

Vandeputte, G. E., Vermeylen, R., Geeroms, J., and Delcour, J. A. 2003. Rice starches.I.

structural aspects provide insight into crystallinity characteristics and gelatinization

behavior of granular starch. J. Cereal Sci. 38: 43-52.

Vandeputte, G. E., and Delcour, J. A. 2004. From sucrose to starch granule to physical

behaviour: a focus on rice starch. Carbohyd. Polym. 58: 245-266.

Vernet, R. 2002. Climate during the late holocene in the Sahara and the sahel: evolution

and consequences on human settlement. In droughts, food, and culture: ecological

change and food security in Africa’s later prehistory (ed. Hassan, F. A), Plenum

press, New York: 47-63.

Wang, L., Xie, B., Xiong, G., Du, X., Qiao, Y., and Liao, L. 2012. Study on the granular

characteristics of starches separated from Chinese rice cultivars. Carbohydr.

Polym. 87: 1038-1044.

Wani, A. A., Singh, P., Shah, M. A., Schweiggert-Weisz, U., Gul, K., and Wani, I. A.

2012. Rice starch Diversity: Effects on structural, morphological, thermal, and

physicochemical properties-A review. Comprehensive Reviews in Food Science

and Food Safety 11: 417-436.

Wani, I. A., Sogi, D. S., Wani, A. A., Gill, B. S., and Shivhare, U. S. 2010. Physico-

chemical properties of starches from Indian kidney bean (Phaseolus vulgaris)

cultivars. Int. J. Food Sci. Technol. 45: 2176-2185.

Ward, K. E. J., Hoseney, R. C., and Seib, P. A. 1994. Retrogradation of amylopectin

from maize and wheat starches. Cereal Chem. 71: 150-155.

Willett,W., Manson, J. and Simin, L. 2002. Glycemic index, glycemic load, and risk of

type 2 diabetes. Am. J. Clin. Nutr. 76: 274-280.

Yamin, F. F., Lee, M., Pollak, L. M., and White, P. J. 1999. Thermal properties of starch

in corn variants isolated after chemical mutagenesis of inbred lines B73. Cereal

Chem. 76: 175-181.

Yeh, A., and Yeh, L. 1993. Some characteristics of hydroxylpropilated and cross-linked

rice starch. Cereal Chem. 70: 273-276.

Zhang, Z., Zhao, S., and Xiong, S. 2010. Morphology and physicochemical properties of

mechanically activated rice starch. Carbohyd Polym 79: 341-348.

Zhu, F., Corke, H., and Bertoft, E. 2011. Amylopectin internal molecular structure in

relation to physical properties of sweetpotato starch. Carbohydr. Polym. 84: 907-

918.

CHAPTER 3: CLASSIFICATION OF RICE BASED ON STATISTICAL ANALYSIS OF PASTING PROPERTIES AND APPARENT AMYLOSE CONTENT: THE CASE OF Oryza glaberrima ACCESSIONS FROM AFRICA

Abstract

The diversity of 1,020 Oryza glaberrima rice accessions being kept at the Genetic

Resources Unit of the Africa Rice Center, with a varied range of apparent amylose content and pasting properties, was explored using cluster analysis. Rice cultivars are usually characterized according to grain dimensions, apparent amylose contents and gelatinization temperatures, however, this work focused on grouping African rice accessions based on their pasting properties and apparent amylose content. Using Ward method of hierarchical cluster analysis, 1,020 rice accessions were initially distributed into five major clusters and further into twenty-three sub clusters. The distribution pattern indicated that Clusters I, II,

III, IV and V formed 27.6%, 10.2%, 15.8%, 23.7% and 22.6% of the entire population, respectively. Though some of the groups had similar apparent amylose content, their pasting properties were very different, making it imperative for further investigations.

Peak viscosity highly correlated (p<0.01) with trough, breakdown and final viscosities in all the five clusters while correlation between peak viscosity and apparent amylose content was not significant within clusters II and IV. Additionally, this categorization serves as a tool for exploring materials that can be employed in the development of rice cultivars for specific end uses.

Key words: rice, oryza, glaberrima, sativa, apparent amylose, cluster, accessions, pasting

3.1 Introduction

Rice is one of the main cereal crops cultivated worldwide on more than 148 million hectares in different agro-ecological areas and it serves as a primary source of food for more than two-thirds of the world’s population (Singh et al. 2013; Ndjiondjop et al. 2010).

The genus Oryza comprises of twenty-two wild and two distinct types of domesticated rice

- Oryza sativa (Asian rice) and Oryza glaberrima (African rice). While O. sativa is distributed globally but highly concentrated in Asia; O. glaberrima is mainly grown in the

West Africa sub region (Sweeney and McCouch 2007; Futakuchi et al. 2008). The advantages attributed to O. glaberrima over O. sativa include a preferred taste locally, early maturity, better weed competitiveness, resistance to biotic stresses such as diseases, insects and nematodes and abiotic stresses such as drought, flooding and salinity.

However, O. glaberrima is characterized by low yields as well as susceptibility to shattering and poor resistance to lodging. Its red colored pericarp reduces the perception of the overall quality when the grains are whitened after dehusking, resulting in a low value on conventional markets. Additionally, it breaks easily during milling resulting in low head rice yield of milled rice (Nayar 2010; Watanabe et al. 2002b; Sarla and Swamy 2005).

Although it is still popular in some areas because of its preference for the preparation of some traditional foods as well as its tolerance to stresses in the field, many farmers in the

West African region are replacing O. glaberrima with O. sativa for want of higher yields and better grain qualities (Linares 2002; Chakanda et al. 2013; Jones et al. 1997).

In the early 1990s, some O. glaberrima lines with desirable qualities namely, high tillering ability, rapid seed growth and early maturity were selected from a screening exercise and subjected to interspecific crosses with O. sativa (high yielding elite japonica upland lines)

leading to the development of NERICA (NEw RICe for Africa) lines for lowland and upland ecologies at the Africa Rice Center. The concept underlying this breeding process was to exploit the adaptability of O. glaberrima to local growing conditions and the high yield of O. sativa. The NERICA varieties developed and disseminated have impacted farmers’ livelihoods and contributed to poverty reduction in sub Saharan Africa. In spite of this success, it is believed that O. glaberrima has not been fully exploited in terms of other traits that were not focused on during the varietal improvement. These include amylose content which is a major factor that influences texture and taste of cooked rice (Futakuchi and Sié 2009; Zhang et al. 2010). To this end, a collection of Oryza glaberrima accessions are being systematically characterized by Africa Rice Center (AfricaRice) for resistance against major diseases and environmental stresses. To contribute to the broader objective of this characterization process, this study seeks to classify 1,020 O. glaberrima accessions from West Africa based on data of their pasting characteristics and apparent amylose content using hierarchical cluster analysis technique. Accessions with potentially different properties will be selected for further research on their functional, thermal and molecular characteristics. This study may contribute to the selection of improved material for the development of suitable O. glaberrima rice cultivars to meet diverse consumer preferences.

3.2 Materials and Methods

3.3 Materials

The O. glaberrima accessions used for the study were supplied by the Africa Rice Center based in Cotonou, Benin. All the 1020 accessions were planted in July, 2010 and harvested between October, 2010 and January, 2011. The samples were dehusked in a Satake

dehusker (THU 35B, Satake Co. Ltd. Tokyo, Japan) and brown rice recovered polished in a single pass friction rice pearler (BS08A Satake Co. Ltd. Tokyo, Japan). The samples were milled into flour with a UDY cyclone mill (Fort Collins, Co., USA) fitted with a sieve of

0.5 mm mesh size.

3.3.1 Measurement of Pasting Properties

The pasting properties of rice flour samples were measured using Rapid Visco Analyzer

(RVA- SUPER 4, Newport Scientific Pty Ltd., Warriewood, Australia). About 3 g of flour

(12% moisture basis) from each rice sample was weighed directly into the aluminium canister and mixed with 25 mL of distilled water resulting in a 9.5% suspension. The test profile involved holding the sample at 50 °C for 1 min, heating to 95 °C at a rate of 12

°C/min, holding at 95 °C for 2.5 min, cooling to 50 °C at a rate of 12 °C/min and holding at

50 °C for 2.5 min. The rotating speed of the paddle was maintained at 160 rpm throughout the run except for the first 10 s where paddle speed was held at 960 rpm. The pasting parameters were determined from the RVA curve included peak viscosity (maximum viscosity attained during heating), hot paste viscosity (trough) and final viscosity or cold paste viscosity (viscosity at end of cooling period). Breakdown was calculated as the difference between peak and trough viscosities while setback was calculated as the difference between the cold paste or final viscosity and the peak viscosity. All analyses were conducted in duplicate.

3.3.2 Measurement of Apparent Amylose Content

The apparent amylose content of the rice flour was determined in duplicate by using AACC approved method 61-03 (AACC 2000).

3.4 Statistical Analysis

Hierarchical cluster analysis technique (Ward method) was applied to the data set for the classification of the rice accessions. The apparent amylose content and the pasting parameters of the flour suspensions were used to classify 1,020 rice accessions into several groups and sub groups. The values of the apparent amylose content and pasting parameters were transformed into Z scores to facilitate their representation on one scale. Hierarchical clustering method is based on the average linkage clustering which is a method of calculating Euclidean distance between elements of each cluster. This statistical method converts various characteristics of objects to quantitative measures (similarity distance) and correspondingly clusters them at relatively close distances into a category (Lee et al. 2012).

The clustering was carried out using SPSS software IBM® SPSS® Statistics 21 (Windows) for the descriptive analysis and Pearson correlations between pasting parameters and apparent amylose content within the generated clusters were observed.

3.5 Results and Discussion

3.5.1 Pasting Properties of Rice Flour

The pasting properties of the glaberrima accessions are presented in Table 3.1. The pasting temperature, representing the temperature of initial viscosity increase (beginning of gelatinization) ranged from 86 to 96 °C with an average of 92 °C for the population. It was observed that the pasting temperatures in this study were far higher than those reported by

Bao et al. 2006 which ranged from 67.5 to 81.5 °C with an average of 75 °C for some O. sativa varieties. Peak viscosity ranged from 36.4 RVU (TOG 6481) to 230.7 RVU (TOG

16803) averaging 100.2 RVU for the entire population. The peak viscosity range was 47

narrower than earlier reports by Tan and Corke 2002 (92 to 319 RVU), Bao et al. 2006

(90.1 to 305.4 RVU) and Roferos et al. 2006 (105 to 308 RVU). Trough viscosity, representing the lowest viscosity upon heating was highest at 183.5 RVU for (TOG 16803), while the lowest value of 36.5 RVU was observed in accession IRGC 80809. The average for the entire population was 208.5 RVU. The accession IRGC 80809 had the lowest final viscosity (70.8 RVU), while TOG 16803 recorded the highest final viscosity (421.5 RVU).

The final viscosity range observed in this study is broader than earlier reports by Tan and

Corke 2002 (99 to 365 RVU) and Bao et al. (2006) (111.4 to 412.6 RVU). Breakdown values for the accessions ranged from -0.5 RVU (IRGC 86808; TOG 12105) to 49.5 RVU

(TOG 10664). Accession TOG 5817 exhibited the lowest setback value (33 RVA) while accession TOG 7507 had the highest setback value (224.8 RVU). The mean setback viscosity of 108.4 RVU of the population was higher than the mean of 69.2 reported by

Bao et al. 2006. This difference suggests that O. glaberrima accessions have very different pasting properties from O. sativa accessions making them an interesting subject for further investigation.

3.5.2 Amylose Content of Rice Flours

The apparent amylose content (AAC) of the rice flours of the entire 1,020 population of the

Oryza glaberrima accessions ranged between 15.1 (IRGC 86741) and 29.6% (TOG 12440) with a mean of 26.2% (Table 3.1). The results of this study showed a wider range of AAC as compared to the findings of Watanabe et al. (2002a, b) where 157 O. glaberrima lines screened for grain quality showed AAC ranging from 25 to 27%. However, in the analysis of genotypic diversity in the starch physicochemical properties of 499 non-waxy Oryza

sativa lines, Bao et al. 2006 reported a much wider range of AAC (7.9-32.6%). Other studies on Oryza sativa by Tan and Corke 2002 and Champagne et al. 1999 reported AAC ranges of 6.3-28.2% and 10.3-24.9%, respectively. The sativas in comparison to the glaberrimas seem to have a broader range of AAC which include waxy types of rice. No waxy types were found in the glaberrimas analyzed in this study. Roferos et al. 2006 also reported on the clustering of rice based on physical and physicochemical properties of milled rice where AAC ranged from 12 to 33%. An earlier study applied linear correlation, multiple regression, principal component and cluster analysis to show that sensory scores of raw and cooked rice were determined by physical quality characteristics such as grain translucency and chalkiness and pasting properties such as AAC and RVA pasting viscosity respectively (Roferos and Julianao 1997). Table 3.2 presents details of the distribution as per the five clusters and 23 sub clusters. AAC within clusters ranged as follows; cluster I = 24.7 to 29.4%, cluster II = 20.6 to 29.6%, cluster III = 21.7 to 26.8%, cluster IV = 15.1 to 28.6%, cluster V = 21 to 28.8%. AAC is related to the cooking quality of rice and other physico-chemical properties such as gel consistency, pasting viscosity and gelatinization temperature. It is also considered to be one of the most important predictors of the eating quality of cooked rice and various rice products (Juliano 1998; Bao 2012).

Milled rice cultivars may be generally classified on the basis of their AAC into waxy (1-

2%), very low (2-12%), low (12-20%), intermediate (20-25%) or high (>25%) (Bao 2012).

The AAC of the entire population of O. glaberrima accessions analyzed suggests that they fall into the categories of low, intermediate or high amylose types of rice. Cluster IV had the widest range of AAC followed by cluster II, while cluster I had the narrowest range.

Clusters I, II, III and V had both intermediate and high amylose types of rice but cluster IV

included low, intermediate and high amylose types of rice. The varied AAC of the glaberrimas suggests that when cooked, they will possess diverse textural characteristics since cooked rice kernels with high AAC are dry, fluffy, separate and hard whereas those with low AAC become glossy, soft and sticky (Bao et al. 2006).

3.6 Classification of Rice Accessions

The pattern of distribution of the accessions into various clusters (C) and sub clusters (SC) is shown in Figure 3.1. Using AAC and RVA pasting parameters of the rice flours as the basis, the accessions were hierarchically clustered into five groups - C I, C II, C III, C IV and C V. The distribution pattern indicates that the maximum number of accessions (282) were included in C I, followed by C IV (242), C V (231), C III (161) and C II (104). All five clusters had accessions with a large diversity of AAC and pasting properties. (Table

3.2; Figure 3.2)

3.7 Comparison of Clustering Pattern

3.7.1 Cluster I

The 282 accessions in cluster I formed 27.6% of the entire population. These accessions were subsequently grouped into 5 sub clusters - SC IA, SC IB, SC IC, SC ID and SC IE - with a population of 44 (15.6%), 84 (29.8%), 107 (37.9), 26 (9.2%) and 21 (7.4%), respectively. The AAC of the accessions in all the sub clusters under cluster I put them into a category of high amylose rice. All the sub clusters had very similar ranges of AAC.

However, their pasting characteristics differed greatly as shown in Figure 3.3 and Table

3.2. While peak and trough viscosities depicted strong diversity in all the 5 sub clusters,

SC IB, SC IC, SC ID and SC IE were close with respect to breakdown viscosity, but that of

SC IA was relatively very distinct. SC IA, SC IC, SC IE were likewise similar with respect to final and setback viscosities, but SC IB and SC ID showed diversity.

3.7.2 Cluster II

Cluster II with 104 accessions comprised of the lowest proportion (10.2%) of the entire population. Further grouping gave rise to three sub clusters SC IIA, SC IIB, and SC IIC with populations of 26 (25%), 54 (52%) and 24 (23%) accessions, respectively. The range of AAC for accessions in C II puts them into intermediate and high amylose types of rice.

C II had the broadest ranges of peak, trough, setback and final viscosities. However, it had the second broadest AAC range and the narrowest pasting temperature range. SC IIA, and

SC IIB had similar ranges of AAC with SC IIC having the broadest range (Table. 3.2). SC

IIC also had the broadest range with respect to peak, final, setback and trough viscosities while SC IIB had the narrowest range as regards peak, final and trough viscosities. Pasting characteristics (Fig. 3.4) showed diversity in all the sub clusters in peak and final viscosities, peak time and pasting temperatures.

3.7.3 Cluster III

Cluster III was made up of 161 accessions (Table 3.2) and formed 15.8% of the entire population of the O. glaberrima analyzed and had five sub clusters namely SC IIIA, SC

IIIB, SC IIIC, SC IIID and SC IIIE making up 23%, 24.2%, 26.7%, 11.2% and 14.9% of the accessions, respectively. Cluster III had the narrowest ranges of peak, trough, setback and final viscosities. The AAC range of this cluster puts the accessions into a category of intermediate to high amylose rice. While SC IIIA had the broadest apparent amylose range, SC IIID had the narrowest range and that of SC IIIB and SC IIIE were similar. 51

Additionally SC IIIA had the broadest ranges with respect to peak, trough, breakdown and final viscosities (Table 3.2). The sub cluster with the second broadest ranges of apparent amylose, peak viscosity and breakdown viscosity was SC IIIC but it had the narrowest final viscosity range. Figure 3.5 depicts the diversity within the five sub clusters as regards pasting characteristics notably peak, trough, final and setback viscosities. This observation is similar to the accessions in C I. Notwithstanding, a close similarity was observed in SC

IIIA, SC IIIB, SC IIIC and SC IIIE with reference to breakdown viscosity and peak time.

3.7.4 Cluster IV

23.7% of the entire population analyzed (242 accessions) were in cluster IV and the five sub clusters SC IVA, SC IVB, SC IVC, SC IVD and SC IVE constituted 33.1%, 38.8%,

17.4%, 5.4% and 5.4% of the accessions, respectively (Table 3.2). Cluster IV had the broadest range of apparent amylose, the second broadest peak viscosity range and the fourth broadest ranges of both final and setback viscosities but the narrowest breakdown viscosity range. The apparent amylose range of this cluster means that the accessions fall under low, intermediate and high amylose rice. Peak, trough, breakdown, final and setback viscosities depicted strong diversity in all the five sub clusters except in SC IVE (Figure

3.6), which though was distinct with respect to breakdown viscosity, showed very close similarity to SC IVB with regards to peak, breakdown, final and setback viscosities. SC

IVA, SC IVB and SC IVC shared some similarities in AAC and peak time.

3.7.5 Cluster V

The 231 accessions in cluster V formed 22.6% of the total number of accessions analyzed.

As in clusters I, III, and IV there were five sub clusters in cluster V namely SC VA, SC

VB, SC VC, SC VD and SC VE which constituted 25.5%, 15.2%, 32.9%, 9.1% and 17.3% of accessions, respectively (Table 3.2). The AAC of accessions in Cluster V ranged from

21 to 28.8% making them intermediate to high amylose rice types. All the sub clusters had very similar AAC range except for SC VE. Cluster V was with the broadest range of final viscosity and second broadest range of setback and breakdown viscosities and the third broadest AAC range. The nature and extent of diversity in the pasting parameters among the sub clusters in C V (Figure 3.7), generally varied from the other four clusters. Peak, trough, breakdown, final and setback viscosities depicted diversity in all the five sub clusters except in SC VE (Figure 3.6) which showed very close similarity to SC VA with regards to peak and trough viscosities. Additionally, SC VA and SC VD shared some similarities in AAC as well as SC VB and SC VC.

3.8 Correlations

Table 3.3 shows correlations between pasting parameters and AAC within the 5 clusters of

O. glaberrima rice accessions analyzed. Peak viscosity highly correlated with trough, breakdown and final viscosities (p<0.01) in all the five clusters. Similarly, peak viscosity correlated positively with setback viscosity (p<0.01), but correlated negatively with peak time and pasting temperature (p<0.01) except CII. Within clusters II and IV correlation between peak viscosity and AAC was not significant (p<0.05). Trough viscosity correlated positively with peak and final viscosities but exhibited relatively lower correlations with 53

breakdown and setback viscosities (p<0.05) in the five clusters. However, trough viscosity correlated negatively with peak time and pasting temperature except for clusters II and V for peak time. As with peak viscosity, correlation with AAC at p<0.01 was not significant, except for clusters I and III (p<0.05). Breakdown viscosity correlated negatively (p<0.01) with both peak time and pasting temperature in all the five clusters but on the contrary, correlated positively with peak and trough viscosities (p<0.01) except in clusters II and V with trough viscosity (p<0.05). Correlations between breakdown viscosity and setback viscosity were not significant in clusters I, III and V (p<0.05) and there was also no significant correlation with final viscosity in clusters II and V. Breakdown viscosity correlated negatively with AAC in clusters IV and V but positively in clusters I and III, however it was not significant for cluster II. Final viscosity correlated highly (p<0.01) with setback viscosity in all the five clusters while correlation between setback viscosity and apparent amylose was significant only in cluster I. Additionally, peak time also correlated positively with pasting temperature in all five clusters.

3.9 Conclusion

This study has shown that there is a high level of genetic diversity in the accessions of O. glaberrima from the West Africa sub region as evidenced by the variation in both AAC and pasting properties in and within the clusters and sub clusters. This diversity suggests that there may be a valuable resource of potentially very important gene pool that may be of benefit in rice breeding programs in this region. Additionally, a detailed molecular analysis of a selection of interesting accessions from the clusters such as those with a high tendency to retrograde (high setback and potentially slower rate of digestion due to

retrogradation) or form viscous paste/gel after cooking and cooling (high final viscosities) will ensure the use of the diversity in the accessions to assist in future development of acceptable commercial cultivated rice varieties that have the desired amylose content and pasting characteristics to meet consumer preferences and benefit resource poor farmers who still depend on rice cultivation for their livelihood.

3.10 Acknowledgements

The Global Rice Science Partnership (GRiSP) and Africa Rice Center (AfricaRice) are hereby acknowledged for providing funding for this study. Many thanks to the staff of the

Grain Quality Laboratory of the Africa Rice Center, Cotonou, Benin for assisting in running experiments.

3.11 Tables and Figures

Table 3.1. Descriptive Statistics of Entire Population of 1,020 O. glaberrima Accessions AAC and Pasting Parameters Minimum value Maximum value Mean Std. Deviation Apparent Amylose Content (%) 15.1 29.6 26.2 1.33

Peak Viscosity (RVU) 36.4 230.7 100.2 29.31

Trough Viscosity (RVU) 36.5 183.5 95.4 23.33

Breakdown Viscosity (RVU) -0.5 49.5 4.8 7.45

Final Viscosity (RVU) 70.8 421.5 208.6 55.94

Setback Viscosity (RVU) 33.0 224.8 108.4 29.30

Peak Time (min) 5.7 7.0 6.4 0.48

Pasting Temperature (°C) 86.4 95.6 92.0 1.71

Table 3.2. Distribution of 1,020 O. glaberrima Rice Accessions Obtained Through Cluster Analysis Using SPSS# Cluster Range

Distribution (N) AAC (%) PV (RVU) TV (RVU) BDV (RVU) FV (RVU) SBV (RVU) PT (min) PTemp (°C)

C I (282) 24.7 - 29.4 67.3 - 123.8 66.7 – 124.2 -0.4 – 6.6 149.6 – 296.8 71.7 – 173.0 5.9 - 7.0 89.6 -94.5

SC IA (44) 25.3 - 29.4 76.8 - 115.6 73.3 – 113.4 1.0 – 6.6 160.1 – 225.4 71.7 – 124.5 5.9 - 6.5 90.7 - 93.6

SC IB (84) 24.8 - 27.9 67.3 - 87.4 66.7 – 87.3 -0.4 – 2.2 149.6 – 188.3 75.0 – 104.3 6.1 - 7.0 91.9 - 94.0

I SC IC (107) 24.7 - 28.5 76.3 -101.6 76.6 – 101.8 -0.4 – 0.8 170.4 – 234.9 75.8 – 139.5 6.3 - 7.0 91.9 - 94.5

SC ID (26) 26.0 - 29.0 92.9 - 123.8 93.0 – 124.2 -0.3 – 2.2 217.7 – 296.8 118.0 – 173.0 6.3 - 7.0 91.5 - 93.6

SC IE (21) 25.4 - 27.7 89.4 – 107.9 89.7 – 105.8 -0.2 – 2.2 176.3 – 217.1 86.8 – 116.8 6.4 - 7.0 89.6 - 92.8

C II (104) 20.6 - 29.6 152.2 – 230.7 113.0 – 183.5 6.0 – 49.5 246.5 – 421.5 94.9 – 224.8 5.7 - 5.9 86.4 - 91.9

SC IIA (26) 23.0 - 29.2 137.8 – 182.7 125.3 – 160.9 6.0 – 32.4 308.1 – 394.8 54.8 – 224.8 5.7 - 6.2 88.3 - 91.9

II SC IIB (54) 23.9 – 29.6 135.5 – 165.8 115.9 – 143.0 6.9 – 34.2 246.5 – 327.3 95.3 – 170.7 5.7 - 5.9 87.2 - 91.0

SC IIC (24) 20.6 - 28.5 152.2 – 230.7 113.0 – 183.5 21.9 – 49.5 247.0 – 421.5 94.9 – 190.8 5.7 - 5.9 86.4 - 89.2

C III (161) 21.7 - 26.8 80.9 – 125.5 80.4 – 122.2 -0.2 – 12.8 159.9 – 258.8 68.3 – 149.8 5.9 - 7.0 89.6 - 94.0

SC IIIA (37) 21.7 - 26.8 89.2 – 125.5 88.2 – 122.3 -0.2 – 12.8 228.3 – 258.8 120.8 – 149.8 5.9 - 7.0 90.7 - 93.6

SC IIIB (39) 23.2 - 26.6 89.8 – 110.8 89.8 – 103.1 -0.2 – 9.2 212.1 – 232.8 107.8 – 132.2 5.9 - 7.0 90.7 - 93.6

III SC IIIC (43) 22.6 - 26.5 80.9 – 110.0 80.4 – 104.1 -0.2 – 11.8 194.3 – 211.5 92.1 – 119.1 5.9 - 7.0 89.9 - 94.0

SC IIID (18) 24.6 - 26.4 108.2 – 121.3 100.7 – 115.5 2.4 – 10.1 213.6 – 236.1 103.3 – 117.2 5.9 - 6.2 89.6 - 92.0

SC IIIE (24) 23.4 - 26.4 82.9 – 107.9 81.0 – 105.0 0.1 – 6.5 159.9 – 196.6 68.3 – 101.2 6.0 - 7.0 90.3 - 92.4

C IV (242) 15.1 - 28.6 36.4- 93.1 36.5 – 93.0 -0.5 – 1.2 70.8 – 205.8 33.0 – 122.0 6.3 - 7.0 91.5 - 95.6

SC IVA (80) 22.4 - 27.8 53.6 –73.4 54.0 – 73.8 -0.5 - 0.0 119.8 – 158.4 57.9 – 89.8 7.0 - 7.0 92.8- 95.2

SC IVB (94) 15.1 - 28.6 63.6 – 83.9 63.7 – 84.0 -0.4 – 0.2 135.8 – 180.4 62.3 – 105.0 6.8 - 7.0 91.5 - 94.7

IV SC IVC (42) 23.0 - 27.6 36.4 – 63.3 36.5 – 63.4 -0.4 - -0.1 70.8 – 118.7 33.0 – 67.8 7.0 - 7.0 92.4 - 95.6

SC IVD (13) 21.3 - 27.4 74.5- 93.1 74.3 – 93.0 -0.2 – 0.4 171.0 – 205.8 81.2 – 122.0 6.6 - 7.0 92.0 - 94.4

SC IVE (13) 21.7 - 25.6 50.8– 83.3 50.4 – 82.9 -0.4 – 1.2 123.4 – 176.0 60.4 – 98.0 6.3 - 6.7 91.9 - 94.0

C V (231) 21.0 - 28.8 96.3 – 152.3 87.1 – 141.4 0.8 – 22.2 194.2 – 343.5 84.8 – 197.7 5.7 - 6.4 87.1 - 92.4

SC VA (59) 25.3 -28.8 117.3 – 152.3 104.9 – 139.3 6.3 – 22.2 212.4 – 275.3 84.8 – 137.7 5.7 - 6.0 87.1 - 91.1

SC VB (35) 25.7 - 28.8 114.3 -151.1 108.5 – 141.4 1.6 – 14.8 255.7 – 343.5 135.2 – 197.7 5.8 - 6.3 89.9 - 92.4

V SC VC (76) 25.5 - 28.5 100.8 – 131.6 198.3 – 124.8 0.8 – 10.1 209.7 – 268.1 91.3 – 153.2 5.9 - 6.4 89.1 - 92.4

SC VD (21) 25.6 – 28.1 96.3 – 117.2 87.1 – 104.2 3.1 – 22.0 194.2 – 229.4 90.5 – 117.3 5.7 - 6.0 89.3 - 91.5

SC VE (40) 21.0 - 27.5 108.3 – 141.5 95.9 – 127.5 5.7 – 17.0 237.5 – 306.7 111.4 – 167.7 5.7 - 6.2 88.7 - 92.0

#Ranges of values

C – Cluster SC – Sub Cluster PV-Peak Viscosity TV – Trough Viscosity BDV – Breakdown Viscosity FV – Final Viscosity SBV – Setback Viscosity PT – Peak Time PTemp – Pasting Temperature AAC - Apparent Amylose Content

Table 3.3. Correlations between Pasting Parameters and Apparent Amylose Content within the 5 Clusters of 1,020 O. glaberrima Accessions# Correlations Break- Apparent Cluster, Peak Trough down Final Setback Peak Pasting Amylose Parameter Viscosity viscosity viscosity viscosity viscosity time Temp. Content N=1,020 Cluster I (N=282) 1 0.992** 0.339** 0.765** 0.449** -0.320** -0.526** 0.426** Cluster II (N=104) 1 0.846** 0.688** 0.660** 0.217* NS -0.500** NS Cluster III (N=161) 1 0.955** 0.527** 0.646** 0.269** -0.382** -0.500** 0.264** Cluster IV (N=242) 1 1.000** 0.400** 0.915** 0.749** -0.258** -0.701** NS

Cluster V (N=231) Peak Viscosity 1 0.924** 0.501** 0.723** 0.408** -0.275** -0.384** -0.175** N=1,020 Cluster I (N=282) 0.992** 1 0.218** 0.775** 0.467** -0.220** -0.484** 0.398** Cluster II (N=104) 0.846** 1 0.195* 0.780** 0.472** NS -0.247* NS Cluster III (N=161) 0.955** 1 0.252** 0.669** 0.323** -0.174* -0.400** 0.194* Cluster IV (N=242) 1.000** 1 0.385** 0.915** 0.748** -0.245** -0.697** NS

Cluster V (N=231) viscosity Trough 0.924** 1 0.132* 0.776** 0.519** NS -0.227** NS

N=1,020 ** ** * ** ** ** Cluster I (N=282) 0.339 0.218 1 0.145 NS -0.838 -0.460 0.330 Cluster II (N=104) 0.688** 0.195* 1 NS -2.43* - 0.553** -0.584** NS Cluster III (N=161) 0.527** 0.252** 1 0.190* NS -0.751** -0.486** 0.308** viscosity ** ** ** ** ** ** ** Cluster IV (N=242) Breakdown 0.400 0.385 1 0.369 0.304 -0.810 -0.479 -0.244 Cluster V (N=231) 0.501** 0.132* 1 NS NS -0.740** -0.481** -0.319**

N=1,020 Cluster I (N=282) 0.765** 0.775** 0.145* 1 0.919** -0.130* -0.194** 0.335** Cluster II (N=104) 0.660** 0.780** NS 1 0.877** 0.218* NS NS Cluster III (N=161) 0.646** 0.669** 0.190* 1 0.909** NS NS NS Cluster IV (N=242) 0.915** 0.915** 0.369** 1 0.952** -0.249** -0.526** NS Final viscosity Cluster V (N=231) 0.723** 0.776** NS 1 0.926** NS 0.131* -1.46*

N=1,020 Cluster I (N=282) 0.449** 0.467** NS 0.919** 1 NS NS 0.204** Cluster II (N=104) 0.217* 0.472** -0.243* 0.877** 1 0.397** 0.420** NS Cluster III (N=161) 0.269** 0.323** NS 0.909** 1 NS 0.394** NS Cluster IV (N=242) 0.749** 0.748** 0.304** 0.952** 1 -0.214** -0.335 NS

Cluster V (N=231) Setback Viscosity 0.408** 0.519** NS 0.926** 1 0.215 0.384** NS N=1,020 Cluster I (N=282) -0.320** -0.220** -0.838** -0.130* NS 1 0.538** -0.340** Cluster II (N=104) NS NS -0.533** 0.218* 0.397** 1 0.693** -0.289** Cluster III (N=161) -0.382** -0.174* -0.751** NS NS 1 0.509** -0.427** Cluster IV (N=242) Peak time -0.258** -0.245** -0.810** -0.249** -0.214** 1 0.360** 0.236** Cluster V (N=231) -0.275** NS -0.740** NS 0.215** 1 0.628** 0.164*

N=1,020 Cluster I (N=282) -0.526** -0.484** -0.460** -0.194** NS 0.538** 1 -0.298** Cluster II (N=104) -0.500** -0.247* -0.584** NS 0.420** 0.693** 1 NS Cluster III (N=161) -0.500** -0.400** -0.486** NS 0.394** 0.509 1 -0.247** Cluster IV (N=242) -0.701** -0.697** -0.479** -0.526** -0.335** 0.360** 1 NS Pasting Temp. Cluster V (N=231) -0.384** -0.227** -0.481** 0.131* 0.384** 0.628** 1 NS N=1,020 ** ** ** ** ** ** ** Cluster I (N=282) 0.426 0.398 0.330 0.335 0.204 -0.340 -0.298 1 Cluster II (N=104) NS NS NS NS NS -0.289** NS 1 Cluster III (N=161) 0.264** 0.194* 0.308** NS NS -0.427** -0.247** 1 Content Amylose Amylose Cluster IV (N=242) Apparent NS NS -0.244** NS NS 0.236** NS 1 Cluster V (N=231) -0.175** NS -0.319** -0.146* NS 0.164* NS 1 # ** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed)

O. glaberrima accessions

(N= 1,020)

C I C II C III C IV C V (N=161) (N=282) (N=104) (N=242) (N=231)

SC IA SC IB SC IC SC VA SC VC (N=44) (N=84) (N=107) SC IIIA SC IIIB SC IIIC SC VB (N=37) (N=39) (N=43) (N=59) (N=35) (N=76)

SC ID SC IE (N=26) (N=21) SC IIID SC IIIE SC VD SC VE (N=18) (N=24) (N=21) (N=40) SC IVA SC IVB SC VIC (N=80) (N=94) (N=42) SC IIA SC IIC SC IIB (N=26) (N=24) (N=54) SC IVD SC IVE (N=13) (N=13)

Figure 3.1. Graphical Representation of Distribution of 1,020 African Rice Accessions into 5 Clusters (C) and 23 Sub Clusters (SC).

Figure 3.2. Distribution of Entire Population of 1,020 O. glaberrima Accessions .

Based on Apparent Amylose Content and Pasting Properties. C= cluster

Figure 3.3. Distribution of O. glaberrima Accessions in Cluster I Population Based on Apparent Amylose Content and Pasting Properties. SC=sub cluster

Figure 3.4. Distribution of O. glaberrima Accessions in Cluster II Population

Based on Apparent Amylose Content and Pasting Properties. SC= sub cluster

Figure 3.5. Distribution of O. glaberrima Accessions in Cluster III Population Based on Apparent Amylose Content and Pasting Properties. SC= sub cluster

Figure 3.6. Distribution of O. glaberrima Accessions in Cluster IV Population Based on

Apparent Amylose Content and Pasting Properties. SC= sub cluster

Figure 3.7. Distribution of O. glaberrima Accessions in Cluster V Population Based on

Apparent Amylose Content and Pasting Properties. SC= sub cluster

3.12 References

AACC International. Approved Methods of Analysis, 11th Ed. Method 61-03.01.

Amylose content of milled rice. First approval October 15, 1997; Re-approval

November 3, 1999. AACC International: St. Paul, MN.

http://dx.doi.org/10.1094/AACCIntMethod-61-03.01.

Bao, J., Shena, S., Sun, M., and Corke, H. 2006. Analysis of genotypic diversity in the

starch physicochemical properties of non-waxy rice: Apparent amylose content,

pasting viscosity and gel texture. Starch/Stärke 58: 259-267.

Bao, J. S. 2012. Toward understanding the genetic and molecular bases of the eating and

cooking qualities of rice. Cereal Foods World 57: 148-156.

Chakanda, R., van Treuren, R., Visser, B., and van den Berg, R. 2013. Analysis of genetic

diversity in farmers’ rice varieties in Sierra Leone using morphological and AFLP

markers. Genet. Resour. Crop Evol. 60: 1237-1250.

Champagne, E. T., Bett, K. L., Vinyard, B. T., McClung, A. M., Barton, F. E., II,

Moldenhauer, K., Linscombe, S., and McKenzie, K. 1999. Correlation between

cooked rice texture and Rapid Visco Analyser measurements. Cereal Chem.

76:764-771.

Futakuchi, K., Fofana, M., and Sié, M. 2008a. Varietal differences in lodging resistance of

African rice (Oryza glaberrima Steud.). Asian J. Plant Sci. 7: 569-573.

Futakuchi, K. and Sié, M. 2009. Better exploitation of African Rice (Oryza glaberrima

Steud.) in varietal development for resource-poor farmers in West and Central

Africa. Agric. J. 4: 96-102.

Jones, M.P., Dingkuhn, M., Aluko, G. K. and Semon, M. 1997. Interspecific Oryza sativa

L. x Oryza. glaberrima Steud. progenies in upland rice improvement.

Euphytica 92: 237-246.

Juliano, B. O. 1998. Varietal impact on rice quality. Cereal Foods World 43: 207-222.

Lawal, O. S., Lapasin, R., Bellich, B., Olayiwola, T. O., Cesàro, A., and Yoshimura, M.

2011. Rheology and functional properties of starches isolated from five improved

rice varieties from West Africa. Food Hydrocoll. 25: 1785-1792.

Lee, I., Gyoung, We, G. J., Kim, D. E., Cho,Y.-S., Yoon, M-R., Shin, M., and Ko,

S. 2012. Classification of rice cultivars based on cluster analysis of hydration and

pasting properties of their starches. Food Sci. Technol. 48: 164-168.

Linares, O. F. 2002. African rice (Oryza glaberrima): History and future potential. Proc.

Natl. Acad. Sci. U.S.A. 99: 16360-16365.

Nayar, N.M. 2010. The history and genetic transformation of the African rice, Oryza

glaberrima Steud. (Gramineae). Curr. Sci. 99: 1681-1689.

Ndjiondjop, M. N., Manneh, B., Cissoko, M., Drame, N.K., Kakai, R. G., Bocco, R.,

Baimey, H., and Wopereis, M. 2010. Drought resistance in an interspecific

backcross population of rice (Oryza spp.) derived from the cross WAB56-104 (O.

sativa) × CG14 (O. glaberrima). Plant Sci. 179: 364-373.

Roferos, L. T., Felix, A. dR., and Juliano, B. O. 2006. The search for the grain quality of

raw and cooked IR 64 milled rice among Philippine Seed Board rice varieties.

Philipp. Agric. Sci.: 89:58-70.

Roferos, L. T., and Juliano, B. O. 1997. Chemometrics of grain quality of raw and cooked

Philippine milled rice. Philipp. Agric. Sci.: 80: 211-236.

Sarla, N., and Swamy, B. P. M. 2005. Oryza glaberrima: a source for the improvement of

Oryza sativa. Curr. Sci. 89: 955-963.

Singh, A., Singh, B., Panda, K., Rai, V. P., Singh, A. K., Singh, S. P., Chouhan, S. K., Rai,

V., Singh, P. K., and Singh, N. K. 2013. Wild rices of Eastern Indo-Gangetic plains

of India constitute two sub-populations harbouring rich genetic diversity. Plant

Omics J. 6: 121-127.

Sweeney, M., and McCouch, S. 2007. The complex history of the domestication of rice.

Ann. Bot. 100: 951-957.

Tan, Y., and Corke, H. 2002. Factor analysis of physicochemical properties of 63 rice

varieties. J. Sci. Food Agric. 82:745-752.

Watanabe, H., Futakuchi, K., Jones, M. P., Teslim, I., and Sobambo, B.A. 2002a.

Brabender viscogram characteristics of interspecific progenies of Oryza glaberrima

Steud. and O. sativa L. Nippon Shokuhin Kagaku Kogaku Kaishi. J. Jpn. Soc.

Food Sci. Technol. 49: 155-165.

Watanabe, H., Futakuchi, K., Teslim, I., and Sobambo, B. A. 2002b. Milling

characteristics and grain quality traits of interspecific progenies of Asian rice

(Oryza sativa) with African rice (Oryza glaberrima). Jpn. J. Tropical Agric. 46:

47-55.

Zhang, C.-H., Li, J.-Z., Zhu, Z., Zhang, Y.-D., Zhao, L., and Wang, C.-L. 2010. Cluster

analysis on Japonica rice (Oryza sativa L.) with good eating quality based on SSR

markers and phenotypic traits. Rice Sci.17: 111-121.

CHAPTER 4: PHYSICOCHEMICAL PROPERTIES, COOKING QUALITY AND

GLYCEMIC ATTRIBUTES OF AFRICAN RICE (Oryza glaberrima)

Abstract

Seven dehusked and polished African rice accessions (ARAs) were evaluated for their grain quality attributes based on physicochemical, cooking quality, textural, pasting, thermal and in vitro digestion properties. They were compared with two O. sativa (Asian rice) varieties; Koshihikari (spp. japonica) and WITA 4 (spp.indica) and NERICA 4 (an inter-specific cross between sativa x glaberrima). Based on amylose content the ARAs examined were intermediate to high amylose rice type while NERICA 4 and the Asian rice varieties were low amylose type. The thousand kernel weight and bulk density of the

ARAs were comparable to the Asian rice varieties which can be classified as medium- shaped kernels based on ISO classification system except for TOG 12440 which was bold.

Volume expansion ratio of the cooked kernels ranged between 2.5 and 3.5 for all samples except for NERICA 4, which was significantly (p<0.05) higher (4.4). The ARAs had significantly (p<0.05) lower cooking solids loss and harder texture compared to the Asian rice varieties which were softer and more adhesive. The peak viscosities of the Asian rice flours were significantly (p<0.05) higher but showed a higher breakdown tendency as compared to the ARAs. Gelatinization transition temperatures (To, Tp, Tc) and enthalpy were significantly (p<0.05) higher in ARA flours as compared with Asian rice. However, melting transition temperatures of retrograded flour gels, were similar for all rice varieties, while the melting enthalpy of the ARAs were higher than that of Asian rice. The expected glycemic indices of both African and Asian rice were higher than 70. The study results show differences in cooking characteristics between African and Asian rice varieties which may be partially due to the high amylose content of the ARAs.

Keywords: amylose, texture, pasting, African rice, glycemic index 73

4.1 Introduction

One of the key cereal crops that is put under vast cultivation worldwide and serves as a primary source of food for more than two-thirds of the world’s population is rice

(Ndjiondjop et al. 2010). Oryza sativa (Asian rice) and Oryza glaberrima (African rice) are the two main types of domesticated rice that are cultivated by farmers globally (Linares

2002) but African rice which is indigenous to the West African sub region has seen less development (Sarla and Swamy 2005; Sweeney and McCouch 2007). One of the main advantages attributed to African rice in comparison to Asian rice is its resistance to a number of biotic stresses such as diseases, insects and nematodes and abiotic stresses such as drought, flooding and salinity. Other advantages cited by producers include its suitability for weaning foods, high protein content and good keeping quality of cooked grains (Sarla and Swamy 2005). Nonetheless it is characterized by low yield and high tendency to shatter, a situation that leads to high post-harvest losses. Conventional milling also leads to high levels of broken rice because it breaks easily (Nayar2010; Watanabe et al. 2002; Sarla and Swamy 2005). In spite of these challenges, some African farmers still crop African rice because of its adaptability, ceremonial and cultural importance (Falade et al. 2014). African rice can therefore be a valuable genetic source for breeding (Aluko et al

2004; Li et al. 2004; Watanabe et al. 2004) and a potential source of genes to enhance the milling, cooking and eating quality of Asian rice (Sarla and Swamy 2005).

In the past, a rice variety (NERICA) was developed successfully by crossing high yielding

Asian rice lines with African rice lines that had some desirable characteristics. The adaptability of African rice to local climatic conditions and the high yield of Asian rice

were the underlying factors for the success of the new variety. Notwithstanding this success, it is believed that African rice possesses other traits that can be tapped for future grain improvement programmes. These include amylose content, which is a major factor that influences texture and taste of cooked rice (Futakuchi and Sié 2009; Zhang et al.

2010). The Africa Rice Center (AfricaRice) has been systematically characterizing a collection of African rice accessions that there may be a valuable resource of potentially very important gene pool that may be of benefit in rice breeding programs. To support the broader objective of this characterization process, this study seeks to evaluate the cooking quality attributes of some African rice accessions that were selected from 1,020 rice accessions based on their amylose and pasting data using Ward method of hierarchical cluster analysis. Rice grain quality has been reported to be usually classified into four components namely: milling efficiency, grain shape and appearance, cooking and edibility characteristics and nutritional quality (Li et al. 2004). Grain quality characteristics vary remarkably with regards to different subspecies and varieties of rice. This provides the basis for categorising rice types that are preferred for ceremonial purposes or for use in specific products such as flour and noodles (Li 2003). Additionally, consumers in different areas of the world prefer different types of rice for general consumption purposes.

Characterizing quality attributes of different types of rice varieties is important for the improvement in rice quality and rice-based products.

Two essential indicators that are commonly used to describe starch digestibility in relation to health benefits are glycemic index (GI) and resistant starch (RS) content. Starch can be categorized into three nutritional fractions, namely rapidly digestible (RDS), slowly digestible (SDS) and resistant starch (RS). Rapidly digestible starches refer to starches that

are digested after 20 minutes while slowly digestible starches are those digested between

20 minutes and 120 minutes. Resistant starches are those that remain undigested after 120 minutes (Englyst et al. 1992). GI is a scale that is used to rank carbohydrate-rich foods by their ability to raise blood glucose level when ingested compared to a reference food

(glucose) or white bread (Jenkins et al. 1987). The predominant type of rice consumed worldwide is the polished or white rice (Nanri et al. 2010) and it is obtained by subjecting paddy rice through a series of mechanized unit operations. The glycemic index value of a specific white rice variety is influenced by the degree of processing, cooking time, and amylose content (Foster-Powell et al. 2002). Different varieties of rice may therefore have different GIs (Miller et al. 1992; Atkinson et al. 2008). GIs of cooked rice varieties ranging from 47 (Ranawana et al. 2009) to 109 (Frei et al. 2003) have been previously reported. In this study, the GIs of the samples are being investigated to ascertain farmers’ indirect claim of African rice having low GI. In addition, the selected African rice accessions were investigated in terms of physical and physicochemical properties of grains and flours. Information generated from this study may contribute to the selection of material for the development of suitable African rice cultivars to meet diverse consumer preferences.

4.2 Materials and Methods

4.3 Materials

Ten dehusked and polished rice samples were provided by the Africa Rice Center based in

Cotonou, Benin for this study. Seven of the samples were African rice accessions (ARAs)

- IRGC 86741, IRGC 80787, IRGC 103759, RAM 123, TOG 7682, TOG 12440 and TOG

14365. The others are two Asian rice varieties – cv Koshihikari (spp. japonica) and cv. WITA 4(spp. indica) and NERICA 4 (an interspecific cross between African rice x

Asian rice). The seven ARAs were chosen from a population of 1,020 African rice accessions from West Africa based on data of their pasting characteristics and apparent amylose content using hierarchical cluster analysis technique. Portions of the samples were milled into flour using UDY cyclone mill (Fort Collins, Co., USA) fitted with a sieve of 0.5 mm mesh size.

4.4 Methods

Whole rice kernels were separated for evaluation of physicochemical, thermal, cooking and in vitro digestibility properties.

4.4.1 Physicochemical Properties

4.4.2 1000-Kernel Weight

The method of Singh et al. (2005) was used for this determination.

4.4.3 Bulk Density

Whole rice kernels were transferred into a cup of known volume from a fixed height. A straight-edge was thereafter used to level off the rice at the top of the cup. The bulk density was calculated as the ratio of the weight of rice in the cup to the volume of the cup and reported as g/mL.

4.4.4 Length-Width Ratio

Length-width ratio was measured using a digital caliper with an accuracy of 0.01mm. Fifty head rice kernels of each sample were randomly selected and measurements of their lengths and widths were manually taken.

4.4.5 Amylose Content

The apparent amylose content of the rice flour was determined in duplicate by using AACC approved method 61-03 (AACC 2000).

4.4.6 Nitrogen

The Dumas combustion method was employed to measure the nitrogen content of rice flour samples. Weighed rice flour samples (0.2 g) were combusted in a LECO Nitrogen/Protein

Analyzer (LECO, FP-528 Lakeview Avenue, St. Joseph, Michigan, United States) after ethylenediaminetetraacetic acid (EDTA) standard was run to calibrate the equipment.

Nitrogen was converted into protein by multiplying the nitrogen determined by a factor of

5.7.

4.4.7 Total Starch

The total starch kit supplied by Megazyme (Wicklow, Ireland) was used for the determination of total starch in the rice flour samples. Total starch was quantified in g/100 g of flour on dry weight basis.

4.4.8 Colour Measurement of Rice Kernels

Rice kernel colour was measured with a Minolta Chroma Meter (CM 3500-d, Konica

Minolta Sensing Inc., Mahwah, NJ, USA). The Minolta Chroma Meter, allocates each sample a colour co-ordinate within the CIE (Commission Internationale l'Eclairage) L-a-b three-dimensional (L*a*b*) colour space. CIE L* values represent brightness on a 0 (pure black) -100 (pure white) unit scale, CIE a* values range from -60 (pure green) to +60 (pure red), and CIE b* values range from -60 (pure blue) to +60 (pure yellow). Each CIE L*, a*, b* value printed is averaged from three readings.

4.4.9 Cooking Characteristics

4.4.10 Optimum Cooking Time

The optimum cooking time (OCT) was determined using the method outlined by Juliano and Bechel (1985) and modified by Mohapatra and Bal (2006). Approximately 100 mL of distilled water was boiled in a beaker and 5 g of head rice samples was placed in it. Ten minutes into boiling and every minute thereafter, 10 grains were removed and pressed between two clean microscope glass slides. The appearance of a chalky/opaque core of the grain was an indication of partially cooked grain. When at least 90% of the grains no longer had uncooked cores, the sample was deemed to be optimally cooked. The sample was then allowed to simmer for an additional 2 min to allow for the core of all grains to gelatinize. OCT included the 2 min simmering time. Volume expansion, length expansion and water uptake ratios (VER, LER, WUR respectively) were determined by placing 1 g of head rice in 15 mL of boiling water till the OCT was reached. VER was reported as the ratio of volume of cooked grains to the volume of uncooked grains while LER was the ratio 79

of the length of cooked grain to length of uncooked grain. WUR was the ratio of water absorbed by the grains during cooking to the initial weight of uncooked rice. Solids loss

(SL) was determined by optimally cooking 2 g of head rice in 20 mL of distilled water and transferring the gruel into a dry pre-weighed aluminium dish. The dish and gruel were placed in an oven (100 °C) for 24 h to remove moisture. SL was calculated as the ratio of increase in weight of dish to weight of rice sample. The average of three replications was reported.

4.4.11 Textural Profile Analysis

Textural profile analysis (TPA) of the cooked rice was performed using a texture analyzer

(TA Exponent 32 Stable Micro Systems version 5,0,3,0, Texture Technologies Corp.,

Scarsdale, New York, USA) with a 5 kg load cell and a two-cycle compression according to the procedure by Mohapatra and Bal (2006). The rice samples were prepared by optimally cooking 10 g in 200 mL of distilled water (till white core in grains disappeared).

At the end of the cooking, excess water was completely drained by employing a strainer and the residual moisture on the grains blotted out. While the samples were still hot, grains from each sample were selected onto the base of the equipment for testing and a two-cycle compression force versus time program was used to compress samples. The probe was programmed to compress the samples till 90% of the original cooked grain thickness, return to its original position and then compress samples again (Juliano et al. 1984;

Meullenet et al. 1997). The probe was made to compress 5 grains arranged parallel to each other with a pre-test and post-test speed of 1 mm/sec and a test speed of 0.5 mm /sec. The parameters recorded from the test curves were hardness, adhesiveness and springiness.

Hardness is defined as the maximum peak force during the first compression cycle while adhesiveness is the negative force area for the first bite and represents the work required to overcome the attractive forces between the surface of a food and the surface of other materials with which the food comes into contact. Springiness is defined as how well a product physically springs back after it has been deformed during a first compression. The spring back is measured at the down stroke of a second compression. The values reported are means of ten replications.

4.4.12 Pasting Properties

The pasting properties of rice flour samples were measured using Rapid Visco Analyzer

RVA-4, (Newport Scientific Pty Ltd., Warriewood, NSW, Australia). 3 g of starch (12% moisture basis) from each rice sample was weighed directly into the aluminium canister and mixed with 25 mL of distilled water resulting in a 9.5% suspension. The test profile involved equilibrating the sample at 50°C for 1 min, heating to 95°C at a rate of 12 °C/min, holding at 95 °C for 2.5 min, cooling to 50 °C at a rate of 12 °C/min and holding at 50 °C for 2.5 min. The rotating speed of the paddle was maintained at 160 rpm throughout the run except for the first 10 s where paddle speed was held at 960 rpm. The under listed pasting parameters were determined from the RVA pasting curve: peak viscosity

(maximum viscosity attained during heating), hot paste viscosity (trough) and final viscosity or cold paste viscosity (viscosity at end of cooling period). Breakdown was calculated as the difference between peak and trough viscosities while setback was calculated as the difference between the final viscosity and the peak viscosity.

4.4.13 Differential Scanning Calorimetry

TA Instruments, Q2000 differential scanning calorimeter (DSC) equipped with a thermal analysis data and recording software facility (TA Instruments, Universal Analysis 2000,

DE, USA) was used to measure the gelatinization properties of the rice flours. Flour dispersion in water (1:5) was allowed to equilibrate for at least 3 h at room temperature before sample was loaded. In all measurements, the thermogram was recorded with an empty aluminum pan as a reference and the scanning temperature range was programmed from 20-120 °C at a heating rate of 10 °C/minute. Scanned samples were stored at 4 °C for

14 days and then re-scanned using the same temperature profile. The onset (To), peak (Tp) and conclusion (Tc) transition temperatures were reported while the enthalpy of gelatinization (ΔH) was estimated by integrating the area between the thermogram and a base line under the peak and was expressed as J/g of dry flour. Determinations were done in duplicates.

4.4.14 In vitro Digestibility and Expected Glycemic Index of Optimally Cooked Rice

In vitro digestibility was conducted on optimally cooked rice kernels using the method outlined by (Englyst et al. 1992). The ten rice samples were optimally cooked as per their

OCTs, freeze-dried and milled using UDY mill (Fort Collins, CO, USA) fitted with a 0.5 mm mesh. A fresh blend of three enzymes namely pancreatin from porcine pancreas

(Sigma-Aldrich P-1625, with activity 3x USP/g), invertase from baker’s yeast -

Saccharomyces sereviciae (1450, Sigma-Aldrich) and amyloglucosidase (Megazyme

International, Ireland with activity 200 U/mL p-nitrophenyl β-maltoside) was prepared according to Englyst et al. (1992) protocol for the hydrolysis of cooked rice flour. Based 82

on their dry weight the flour samples weighed for the hydrolysis process contained 0.7 g of starch. 10 mL of 0.1 M sodium acetate buffer (pH-5.2) was initially added to the weighed samples and incubated in a water bath (37 °C) for 5 min after which 5 mL of the enzyme mix was added. The sample-enzyme mixture was then incubated for 2 h and from start, 0.1 mL aliquots were pipetted from the reaction mass every 20 min into Eppendorf tubes containing 0.9 mL of 80% ethanol to halt the hydrolysis reaction. The amount of glucose in these aliquots was determined by glucose oxidase/peroxidase assay (GOPOD-

Megazyme) and converted to starch by multiplying by a factor of 0.9. The hydrolyzed starch (expressed on dry weight basis) was categorized into rapidly digestible starch

(RDS), slowly digestible starch (SDS) and resistant/residual starch (RS) according to the description of (Englyst et al. 1992). RDS represents glucose released at 20min x 0.9; SDS as (glucose released at 120min - glucose released at 20min) x 0.9 and RS as total starch -

(RDS+SDS). The kinetics of hydrolysis of the flour samples was described by the non-

-kt linear first-order equation C= C∞ (1- ℮ ) developed by Goni et al. (1997). C is the starch concentration hydrolyzed at a time t; C∞ the equilibrium concentration at the final time (120 min); k is the kinetic constant and t is the chosen time. The hydrolysis index (HI) was obtained by dividing the area under the hydrolysis curve (AUC) of the samples by that for the white bread as a reference sample as reported by Goni et al. (1997). The AUC was governed by the equation:

-k (tf-to) AUC= C∞ (tf-to) - (C∞/k) [1-e ]. Where tf is the final time and to is the initial time. The eGI was calculated from the equation: eGI= 8.198+0.862 (HI) as described by Granfeldt et al. (1992).

4.5 Statistical Analysis

The statistical programme Statgraphics Centurion XV, Version 15.1.02 (StatPoint,

Warrenton, VA, USA) was employed to conduct statistical analysis. One-way analysis of variance was performed to determine significant differences between any pair of means at the 95 % confidence level and multiple range tests were done to determine which means were different.

4.6 Results and Discussion

4.7 Physicochemical Properties

The physicochemical properties of the rice samples studied are presented in Table 4.1. The thousand-kernel weight which is a measure of seed/grain size ranged from 15.9 to 20.0 g.

NERICA 4 had the biggest seed size (20.0g) while TOG 7682 had the smallest seed size

(15.9g). The bulk densities of the samples were quite similar but that of Koshihikari

(0.93 g/mL) was the highest. The length-width (L:W) ratio which signifies grain size and shape ranged from 1.71 to 3.56 and corresponded to the two Asian rice varieties

(Koshihikari and WITA 4 respectively). The apparent amylose content of the samples showed significant variations (p<0.05) among the samples with the ARAs and NERICA 4 having significantly higher proportions (p<0.05) than the Asian rice varieties (Table 4.1).

Milled rice cultivars may be generally classified on the basis of their amylose content into waxy (1-2%), very low (2-12%), low (12-20%), intermediate (20-25%) or high (>25%)

(Bao 2012). Thus, by this classification the ARAs belong to intermediate to high amylose rice types while NERICA 4 and the Asian rice varieties are low amylose types. It was similarly observed that the crude protein content which ranged from 5.1 to 11.3% were 84

higher in all the ARAs and NERICA 4 than the Asian rice varieties. Total starch content of the ARAs and the Asian rice varieties were similar and ranged from 82.5 to 87.3 %. The L index of the color measured signifies lightness (whiteness) and values indicate that the appearances of the ARAs were comparable to the Asian rice varieties.

4.8 Cooking and Textural Properties

The cooking and textural characteristics of the ten rice samples are presented in Table 4.2.

The optimum cooking time (OCT) of the samples varied from 16 to 23 min. TOG 7682 which recorded the least bulk density (0.86 g/mL) and least 1000-kernel weight (15.9g) also had the least cooking time of 16 min. Similarly, Koshihikari which had the highest bulk density (0.93 g/mL) recorded the longest cooking time of 23 min. Significant differences were found among the samples in the loss of solids during cooking but it was generally observed that the Asian rice varieties had higher solids loss. Significant variations were observed among the rice samples with respect to water uptake ratio, volume expansion ratio (VER) and length expansion ratio (LER). Hardness in texture test was observed to be significantly higher (p<0.05) in the ARAs which interestingly had higher amylose concentration. In contrast to hardness, the adhesiveness values were found to be higher in the Asian rice samples. Springiness values were marginally higher in most of the ARAs and NERICA than in Asian rice varieties.

4.9 Pasting and Thermal Characteristics

Figure 4.1 and Table 4.3 depict the RVA profiles and pasting characteristics of the rice flour samples. The samples had a peak time ranging from 5.5 min to 7.0 min. The peak

viscosity of the flour samples ranged from 48.5 to 271.6 RVU and were significantly

(p<0.05) lower in the ARAs. The trough viscosity, which is a measure of the ability of the paste to withstand breakdown during the heating phase, varied widely among the samples while final viscosity was significantly higher in the Asian rice samples than most of the

ARAs. Peak viscosity indicates the water-holding capacity of a sample, while final viscosity, the most commonly used parameter to define a particular sample’s quality, indicates the ability of the material to form a viscous paste or gel after cooking and cooling.

While the ARAs showed stability, breakdown viscosity was higher in the Asian rice varieties. The setback viscosity, which is a measure of the tendency of starch paste to retrograde, varied widely among the samples.

DSC thermograms of flour from the rice samples are shown in Figure 4.2 and melting parameters of the same are presented in Tables 4.4 and 4.5. The onset of gelatinization temperatures of the ARAs and NERICA 4 were significantly higher than Asian rice varieties. A similar trend of the ARAs and NERICA 4 having significantly (p<0.05) higher melting temperatures than the Asian rice varieties was observed for both Tp and Tc. The gelatinization temperature range (Tc-To) of the samples was narrow and the gelatinization enthalpies (ΔH) defined as the energy required to unwind the double helices in amylopectin range was also narrow (6.1-10.1) J/g. The ΔH values of the ARAs were significantly higher than NERICA 4 and the two Asian rice varieties suggesting that the amylopectin crystals of the ARAs are more stable in comparison to the Asian rice (Vamadevan et al.

2013). Table 4.5 shows the melting temperatures of the retrograded flour gels. To, Tp, Tc and ΔH values were considerably lower than their corresponding flours. In addition, To, Tp and Tc showed very narrow ranges but with regards to the retrograded flour gels ΔH values 86

of the ARAs and NERICA 4 were significantly higher than the two Asian rice samples suggesting more stability even in the retrograded state.

4.10 In vitro Digestibility of Cooked Rice Samples

Figure 4.3 shows the in vitro starch hydrolysis profiles of the optimally cooked and freeze- dried rice samples. The hydrolysis trend was similar for all the samples and TOG 14365 had the highest rate of hydrolysis compared with the other samples until about 100 min while TOG 7682 had the lowest rate. Approximately 76-86 % of the starch in the samples had been hydrolyzed by 120 min of hydrolysis. This is comparable to 70 % that was reported for rice for the same time of hydrolysis (Zhang et al. 2006). The quantum of RDS in the cooked rice samples during hydrolysis ranged from 41.3 % to 51.4 % and corresponded to IRGC 80787 and RAM 123 (Table 4.6). The levels of SDS in the samples were generally lower than the levels of RDS and similarly RS levels were lower than the

RDS. Table 4.7 shows the kinetic constant, hydrolysis index and eGI of the cooked and freeze-dried rice. The hydrolysis constants and the eGI of the ARAs were comparable to the Asian rice varieties despite the variations observed.

4.11 Discussion

Grain size is an important quality trait in rice and also a major determinant of grain weight

(Fan et al. 2006). Grain weight is important to farmers because it is one of the most stable components of yield (Li et al. 2004). Grain weight, which is commonly denoted by

1,000- kernel weight in breeding applications, is determined by the length, width and thickness of the grain (Evans 1972). Preference for rice grain characteristics vary with consumer groups and rice grain shape has been reported to play a major role in consumer

preferences in different countries and populations (Kesavan et al. 2013). The range of

1000- kernel weights observed in this study was comparable to what was reported for a range of popular rice varieties in Malaysia (Thomas et al. 2103) and some Indian rice cultivars (Singh et al. 2005). The bulk densities observed for the samples in this study were also similar to the range reported by Singh et al. (2005) but the length:width ratios for the ARAs were lower. The grain length and shape of rice are important characteristics to consumers because they determine the physical appearance and affect the cooking quality of the grain. Based on the International Standards Organization (ISO), the rice samples in this study were classified as slender, medium or bold (short). While NERICA 4 and all the

ARAs except TOG 12440 were medium, Koshihikari was bold/short and WITA 4 was slender.

The ARAs were significantly higher in amylose content in comparison to NERICA 4 and the Asian rice varieties. Differences in amylose content of different rice cultivars may be attributed to environmental factors, genotype and cultural practices as well as climatic and soil conditions (Kim and Wiesenborn 1995; Morrison and Azudin 1987). Amylose is also considered to be one of the most important predictors of the eating quality of cooked rice and various rice products (Juliano 1998; Bao 2012). Additionally, the content of amylose in rice is a major factor that affects texture and taste of cooked rice (Zhang et al. 2010;

Futakuchi and Sié, 2009). It has been reported that rice that has high amylose concentration upon cooking exhibits high volume expansion, becomes less tender and harder when it cools while low amylose rice cooks moist and sticky (Juliano 1985). The optimum cooking time showed a negative correlation with amylose content (r=-0.6399, p<0.05) and a weak positive correlation with bulk density of milled rice samples (r=0.5192, 88

p<0.05) (Table 4.8). The relationship between bulk density and cooking time suggested that the rice samples with higher bulk density exhibited slower rate of water uptake which eventually caused a longer cooking time. The samples with higher amylose content required relatively shorter cooking time. Amylose content and bulk density thus influenced the optimum cooking time of the rice samples. This was consistent with what was reported by Singh et al. (2005). Amylose correlated positively with hardness and springiness but negatively with adhesiveness. The positive correlation between amylose and hardness of cooked rice is consistent with what was reported by Singh et al. (2005) and Yu et al.

(2009). Cooked rice that has high amylose concentration easily hardens whereas cooked rice with lower amylose content hardens at a much slower rate (Yu et al. 2009). Amylose has been reported to leach out during cooking to form a coated film on rice grains and the rapid retrogradation of amylose results in the hardness of cooked rice grains

(Leelayuthsoontorn and Thipayarat, 2006). High amylose rice types have been reported to have a dry, firm, and fluffy texture when cooked (Noda et al. 2003; Mohapatra and Bal

2006).

Peak viscosity, the maximum viscosity attained by gelatinized starch during heating in excess water, is determined by granular swelling, amylose leaching and friction between granules and it also signifies water binding capacity of starch granules (Shimelis et al.

2006). The granular swelling behavior of cereal starches is primarily a property of their amylopectin content (Tester and Morrison 1990), and since the Asian rice varieties had higher amylopectin content than the ARAs it was expected that the former will have higher peak viscosity. The final viscosity which indicates the ability of the material to form a viscous paste or gel after cooking and cooling was similar in NERICA 4 and the Asian rice

varieties but relatively lower in the ARAs except IRGC 80787 and TOG 12440.

Breakdown viscosity measures the susceptibility of the cooked starch to fragmentation and that the higher breakdown viscosity, the lower the ability of the sample in question to withstand heating and shear stress during cooking (Adebowale et al. 2005). The low or

(negative) breakdown of the ARAs seems plausible due to the high enthalpy of gelatinization (ΔH) indicating a stable structure (Table 4.4). This suggests that the high amylose content of the ARAs may have restricted the extent of granular swelling and that granules were not fully disintegrated, hence their lower peak viscosities. Setback viscosity is commonly used to describe the increase in viscosity that occurs on cooling a starch paste

(Fisher and Thompson 1997) and it determines the retrogradation tendency of the paste

(Zaidul et al. 2007). Total setback viscosity did not show any particular trend among the

ARAs and Asian rice samples. The differences in To, Tp, Tc and ΔH in starches from different rice cultivars may be attributed to differences in granular architecture and glucan composition (Vamadevan and Bertoft, 2014).

Based on studies that used rice starches with a wide range of amylose concentration, it has been established that the extent of digestibility of rice starch is related to amylose content

(Chung et al. 2010; Frei et al. 2003; Zhu et al. 2011). Additionally, factors such as extent of granule swelling after cooking, inhibition of granule swelling by amylose-lipid complex and the speed and extent of retrogradation of amylose molecules are important. However, it has been also reported that rice varieties with similar amylose concentration can possess different physicochemical (gelatinization) properties which can affect starch digestion rate and blood glucose. The ARAs due to their higher amylose content were expected to show a much lower eGI than the Asian rice varieties but this was not observed. Consequently

using amylose content alone to predict starch digestibility and glycemic response can be inaccurate (Panlasigui et al. 1991).

Table 4.9 shows the correlations between amylose and OCT and thermal and pasting properties of the flour samples. Correlation between OCT and the thermal parameters as well as the pasting parameters measured were not significant (p<0.05) except for peak and trough viscosities. Amylose correlated positively with the enthanlpy of gelatinization and all the transition temperatures. Correlation was not significant (p<0.05) between amylose and final and total setback viscosities. However, amylose correlated negatively with peak, trough and breakdown viscosities but positively with setback viscosity. Multiple linear regression models to describe the relationship between amylose, OCT and cooking and textural parameters are presented in figure 4.4.

4.12 Conclusion

This study established the African rice accessions as medium shaped high amylose rice type with high protein content. Their relatively higher amylose concentration influenced some of the cooking quality and textural parameters measured. This reflected in the ARAs having relatively harder texture of cooked grains. The high enthalpy of gelatinization (ΔH) of the ARAs indicates a more stable structure that was relatively more resistant to granule swelling to preserve the granule integrity and exhibit shear stability. This resulted in negligible breakdown. Additionally, the observed differences between the ARAs and the

Asian rice varieties in the properties are likely due the differences in the amylose concentration. The high amylose did not however influence the estimated GI of ARAs.

4.13 Acknowledgements:

A special appreciation is hereby extended to the Global Rice Science Partnership (GRiSP) and Africa Rice Center (AfricaRice) for the provision of funding and rice samples for this work.

4.1 Tables and Figures Table 4.1. Physicochemical Characteristics of African and Asian Rice Kernels and Flours# BD Colour Accession MC (%) TK W(g) L:W AC (%) P (%) TS (%) (g/mL) L* a* b* African rice IRGC 80787 5.07bcde 17.63e 0.87bc 2.15c 28.73d 8.18g 85.20c 65.3 2.2 13.8

IRGC 86741 4.09ab 16.73c 0.86ab 2.35d 25.47c 11.29j 84.34b 59.4 4.3 16.5

IRGC 103759 4.55bc 16.82c 0.88cd 2.14c 28.23d 7.79f 84.83c 65.8 3.8 17.0

RAM 123 5.36de 19.09g 0.90e 2.14c 29.72de 7.55e 85.06c 69.7 1.3 13.7

TOG 7682 5.18bcde 15.94a 0.86a 2.22c 28.51d 10.35i 82.49a 64.3 1.0 14.7

TOG 12440 5.59e 16.26b 0.86ab 1.91b 28.25d 7.22d 84.32b 61.9 3.7 14.7

TOG 14365 5.25cde 17.36d 0.86ab 2.16c 31.07e 9.42h 84.10b 59.9 4.7 17.0

O. sativa x O. glaberrima cross NERICA 4 4.58abc 20.02h 0.90de 2.75e 23.76b 6.96c 86.43e 66.4 0.9 16.3

Asian rice Koshihikari 4.69abcd 18.11f 0.93f 1.71a 17.00a 5.06a 85.92d 66.0 -1.3 12.7

WITA 4 3.59a 17.35d 0.87abc 3.56f 22.88b 6.16b 87.31f 69.5 -1.1 11.8

MC - Moisture Content TKW – Thousand Kernel Weight BD - Bulk Density TS – Total Starch L/W - Length/Width Ratio AC - Amylose Content P- Protein #Values with the same letters are not significantly different within columns 93

Table 4.2. Cooking and Textural Properties of African and Asian Rice Kernels# OCT SL Hardness Adhesiveness Springiness WUR VER LER (min.) (%) (g) (g.sec) (mm) African rice IRGC 80787 18.0 3.25abc 3.00b 3.00bc 1.43abc 6039ef 39.4b 0.74d

IRGC 86741 19.0 2.96a 2.61a 3.25cde 1.42ab 5937def 0.8a 0.50ab

IRGC 103759 19.0 3.78cd 3.19bc 3.50de 1.44abc 5229c 14.3a 0.72d

RAM 123 17.0 4.48ef 3.38cd 2.56ab 1.62e 5428cde 12.2a 0.79d

TOG 7682 16.0 3.01ab 2.60a 2.88bc 1.42ab 6122f 0.8a 0.55bc

TOG 12440 21.0 5.00f 3.43cd 2.38a 1.51bcd 4142b 12.3a 0.72d

TOG 14365 17.0 3.56bc 3.06b 3.25cde 1.55de 5303cd 0.7a 0.50ab

O. sativa x O. glaberrima cross NERICA 4 19.0 4.33de 3.50d 4.36f 1.52cde 3825b 22.8ab 0.60c

Asian rice Koshihikari 23.0 6.04g 3.40cd 3.67e 1.60de 2729a 149.1c 0.43a

WITA 4 17.0 5.76g 3.39cd 3.04bcd 1.41a 3761b 40.3b 0.56bc

OCT – Optimum Cooking Time SL – Solids Loss WUR – Water Uptake Ratio

VER – Volume Expansion Ratio LER – Length Expansion Ratio # Values with the same letters are not significantly different within columns

Table 4.3. Pasting Properties of Flours from African and Asian Rice Kernels# PKT PV TV FV BDV SB Sample (min) (RVU) (RVU) (RVU) (RVU) (RVU) African rice IRGC 80787 5.9 171.6h 156.5f 335.1g 15.1c 163.5i

IRGC 86741 7.0 60.9b 65.2b 136.2b 0*a 75.4c

IRGC 103759 6.8 99.6d 101.1d 206.2d 0*a 106.7g

RAM 123 6.0 104.9e 100.3d 203.5d 4.6b 98.6f

TOG 7682 7.0 48.5a 51.1a 94.5a 0*a 46.0b

TOG 12440 5.8 144.4f 129.6e 276.4f 14.8c 132.0h

TOG 14365 7.0 78.5c 80.4c 165.3c 0*a 86.8d

O. sativa x O. glaberrima cross NERICA 4 5.5 166.2g 126.5e 260.5e 39.7d 94.4ef

Asian rice Koshihikari 6.0 271.6i 153.2f 260.6e 118.5f -11.0a

WITA 4 5.8 173.0h 128.2e 262.7e 44.8e 89.7de

PKT= Peak Time; PT= Pasting Temperature; PV= Peak Viscosity; TV= Trough

Viscosity;

FV= Final Viscosity; BDV= Breakdown Viscosity; SB= Setback Viscosity #Values with the same letters are not significantly different within columns *0 indicates the stability of the paste, i.e. there was no breakdown

Table 4.4. DSC Gelatinization Temperatures and Enthalpy of Flours from African and Asian Rice Kernels#

Accession To (°C) Tp (°C) Tc (°C) ΔH (J/g) MR (°C) African rice

e de f d IRGC 80787 71.6 75.5 83.2 9.1 11.6

d cd d b IRGC 86741 71.1 75.1 82.5 7.4 11.4

e f h c IRGC 103759 71.6 76.5 83.8 8.3 12.2

c c e ef RAM 123 70.7 75.0 82.9 9.6 12.2

c e j fg TOG 7682 70.8 75.7 84.8 9.8 14.0

d de i g TOG 12440 71.1 75.5 84.1 10.1 13.0

e de g de TOG 14365 71.4 75.6 83.4 9.2 12.0

O. sativa x O. glaberrima cross f f c b NERICA 4 72.0 76.3 82.0 7.5 10.0

Asian rice b b b b Koshihikari 62.9 69.1 76.1 7.7 13.2

a b a a WITA 4 62.2 68.0 75.6 6.1 13.4

To - Onset of Gelatinisation Temperature Tp - Peak of Gelatinisation Temperature

Tc - Conclusion of Gelatinisation Temperature ΔH - Enthalpy of Gelatinization

MR= Melting Range (Tc-To) # Values with the same letters are not significantly different within columns

Table 4.5. Transition Parameters of Stored African and Asian Rice Flour Gels# 1 Sample To (°C) Tp (°C) Tc (°C) ΔH (J/g) MR (°C) % Retrgd African rice IRGC 80787 43.5abc 58.4c 73.0e 4.8f 29.5 52.7

IRGC 86741 42.6a 58.3c 71.7bc 3.8cd 29.1 51.4

IRGC 103759 43.3abc 58.1c 72.0cd 4.2de 28.7 50.6

RAM 123 44.4bc 58.0bc 71.7bc 4.1d 27.3 42.7

TOG 7682 43.8abc 58.1c 72.7de 4.8f 28.9 49.0

TOG 12440 43.0ab 58.1c 72.1cd 4.7ef 29.1 46.5

TOG 14365 44.6c 57.4bc 68.4a 3.3c 23.8 35.9

O. sativa x O. glaberrima cross NERICA 4 43.7abc 58.0bc 72.2cde 3.9d 28.5 52.0

Asian rice Koshihikari 43.4abc 55.7a 71.8bc 0.3a 28.4 3.9

WITA 4 44.1abc 57.0b 71.0b 0.90b 26.9 14.8

To = Onset Temperature; Tp = Peak Temperature; Tc = Conclusion Temperature;

ΔH = Enthalpy of Gelatinization. MR= Melting Range (Tc-To);

1Percentage Retrogradation (calculated as ΔH of retrograded gel/ΔH of starch granules * 100)

#Values with the same letters are not significantly different within columns

Table 4.6. Nutritional Starch Fractions during Hydrolysis of Cooked African and Asian Rice# Sample RDS (%) SDS (%) RS (%) African rice IRGC 80787 41.3a 36.8e 21.9e

IRGC 86741 44.5b 31.8d 23.8g

IRGC 103759 49.9cd 30.2c 20.0c

RAM 123 51.4e 27.7a 21.0d

TOG 7682 42.5a 39.7f 17.9b

TOG 12440 49.6c 30.0bc 20.6cd

TOG 14365 50.9de 28.5ab 20.6cd

O. sativa x O. glaberrima cross

NERICA 4 49.3c 29.0abc 21.8e

Asian rice Koshihikari 45.1b 41.3g 13.8a

WITA 4 49.5c 27.6a 23.0f

# Values with the same letters are not significantly different within columns

RDS =Rapidly Digestible Starch; SDS = Slowly Digestible Starch;

RS = Resistant/Residual Starch

Table 4.7. Kinetic Constant, Hydrolysis Index and Expected Glycemic Index of cooked African and Asian rice# Kinetic Hydrolysis Sample eGI Constant (k) Index (HI) African rice IRGC 80787 0.0376a 73.2a 71.3a

IRGC 86741 0.0438b 74.1ab 72.4b

IRGC 103759 0.0488c 79.7c 76.9ef

RAM 123 0.0525e 79.8c 77.1ef

TOG 7682 0.0364a 76.3abc 74.0c

TOG 12440 0.0490c 79.1c 76.4de

TOG 14365 0.0513de 79.8c 77.0ef

O. sativa x O. glaberrima cross NERICA 4 0.050cd 78.1c 75.8d

Asian rice Koshihikari 0.0369a 83.7d 77.5f

WITA 4 0.0514de 77.5bc 76.5de

#Values with the same letters are not significantly different within

Table 4.8. Correlations between Amylose; Protein and Cooking and Textural Properties#

Amylose Protein Bulk OCT Solids WUR VER LER Hardness Adhesiveness Springiness Density Loss Amylose

Protein 0.5872

Bulk Density -0.5914 -0.6814

OCT -0.6399 -0.5206 0.5192

Solids Loss -0.6661 -0.8886 0.5407 0.5212

WUR NS -0.8638 0.5865 NS 0.7750

VER -0.4826 NS NS NS NS NS

LER NS NS 0.5905 NS NS 0.5323 NS

Hardness 0.7577 0.8387 -0.5653 -0.6555 -0.8844 -0.7546 NS NS

Adhesiveness -0.8161 -0.7020 0.7201 0.6841 0.6439 NS NS NS -0.6907

Springiness 0.5255 NS NS NS NS NS NS NS NS NS

# Correlation is significant at the 0.05 level

100

Table 4.9. Correlations between Amylose and Thermal and Pasting Properties# To Tp Tc ΔH Peak Trough Final Breakdown Setback Total viscosity viscosity viscosity viscosity viscosity setback

Amylose 0.7572 0.7370 0.8277 0.6777 -0.7568 -0.4660 NS -0.9016 0.6418 NS

OCT NS NS NS NS 0.6803 0.5575 NS NS NS NS

# Correlation is significant at the 0.05 level

101

350

300 IRGC 80787

IRGC 86741 250 IRGC 103759

200 RAM 123

TOG 7682 150 TOG 12240 (RVU) Viscosity

TOG 14365 100 NERICA 4

50 KOSHIHIKARI

WITA 4 0 0 5 10 15 Time (min)

Figure 4.1. RVA Pasting Profiles of African and Asian Rice Flour samples.

102

-3 ––––––– IRGC 80787.001 ––––––– IRGC 86741.001 ––––––– IRGC 103759.001 ––––––– RAM 123.001 ––––––– TOG 7682.001 ––––––– TOG 12440.001 ––––––– TOG 14365.001 ––––––– NERICA 4.001 ––––––– KOSHIHIKARI .001 – – – – WITA 4.001

-4 Heat Flow (W/g)

-5 20 40 60 80 100 120 Exo Up Temperature (°C) Universal V4.7A TA Instruments

Figure 4.2. DSC Thermograms of African and Asian Rice Flour.

103

IRGC 80787 60 (%) IRGC 86741

IRGC 103759

RAM 123 40 hydrolyzed TOG 7682

TOG 12440 Starch Starch TOG 14365

20 NERICA 4

KOSH'RI

WITA 4 0 0 40 80 120 Time (min)

Figure 4.3. Hydrolysis Kinetics of Cooked and Freeze-Dried African and Asian Rice Kernels.

104

Plot of Fitted Model Plot of Fitted Model

24 6.9

22 5.9

20 4.9 OCT Solids Loss 3.9 18

2.9 16 16 20 24 28 32 16 20 24 28 32 Amylose Amylose

OCT = 27.0311 - 0.319824*Amylose Solids Loss = 8.87102 - 0.176512*Amylose r=-0.6399 r=-0.6661

Plot of Fitted Model Plot of Fitted Model

6600 6600

5600 5600

4600 4600 Hardness Hardness 3600 3600

2600 2600 16 18 20 22 24 16 20 24 28 32 OCT Amylose

Hardness = 11657.0 - 365.893*OCT Hardness = -721.243 + 211.394*Amylose r=-0.6555 r= 0.7577

Plot of Fitted Model Plot of Fitted Model

6.9 180 150 5.9 120

4.9 90 60 Solids Loss Adhesiveness 3.9 30

0 2.9 16 20 24 28 32 16 18 20 22 24 Amylose OCT

Solids Loss = -0.922698 + 0.276376*OCT Adhesiveness = 258.6 - 8.70015*Amylose r=0.5212 r=-0.8161

Figure 4.4. Multiple Linear Regression Models to Describe the Relationship between Amylose, OCT and Cooking and Textural Parameters.

105

4.15 References

AACC International. Approved Methods of Analysis, 11th Ed. Method 61-03.01.

Amylose content of milled rice. First approval October 15, 1997; Re-approval

November 3, 1999. AACC International: St. Paul, MN.

http://dx.doi.org/10.1094/AACCIntMethod-61-03.01.

Adebowale, K. O., Olu-Owolabi, I. O., Olawunmi, E. K., and Lawal, O. S. 2005.

Functional properties of native, physically and chemically modified breadfruit

(Artocorpus altilis) starch. Industrial Crops and Products 21: 343-351.

Aluko, G., Martinez, C., Tohme, J., Castano, C., Bergman, C., and Oard, J.H. 2004. QTL

mapping of grain quality traits from the interspecific cross Oryza sativa×O.

glaberrima. Theor. Appl. Genet.109: 630-639.

Atkinson, F. S., Brand-Miller, J. C., and Foster-Powell, K. 2008. International tables of

glycemic index and glycemic load values. Diabetes Care 31: 2281-2283.

Bao, J. S. 2012. Toward understanding the genetic and molecular bases of the eating and

cooking qualities of rice. Cereal Foods World 57: 148-156.

Chung, H. J., Liu, Q. A., Wang, R. L., Yin, Y. L and Li, A. K. 2010. Physicochemical

properties and in vitro starch digestibility of cooked rice from commercially

available cultivars in Canada. Cereal Chem. 87: 297-304

Englyst, H. N., Kingman, S., and Cummings, J. 1992. Classification and measurement of

nutritionally important starch fractions. Eur. J. Clin. Nutr. 46: 33-50.

106

Evans, L. T. 1972. Storage capacity as a limitation on grain yield. In: Banos Los (Ed) Rice

breeding. International Rice Research Institute, Manila: 499-511.

Falade, K. O., Semon, M., Fadairo, O.S., Oladunjoye, A. O., Orou, K. K. 2014. Functional

and physico-chemical properties of flours and starches of African rice cultivars.

Food Hydrocoll. 39: 41-50.

Fan, C., Xing, Y., Mao, H., Lu, T., Han, B., Xu, C., Li, X., and Zhang, Q. 2006. GS3, a

major QTL for grain length and weight and minor QTL for grain width and

thickness in rice, encodes a putative transmembrane protein. Theor.Appl.Genet.

112: 1164-1171.

Fisher, D. K., and Thompson, D. B. 1997. Retrogradation of maize starch after thermal

treatment within and above the gelatinization temperature range. Cereal Chem. 74:

344-351.

Foster-Powell, K., Holt, S. H. A., and Brand-Miller, J.C. 2002. International table of

glycemic index and glycemic load values. Am. J. Clin. Nutr.76:5-56.

Frei, M., Siddhuraju, P., and Becker, K. 2003. Studies on the in vitro starch digestibility

and the glycemic index of six different indigenous rice cultivars from the

Philippines. Food Chem. 83: 395-402.

Futakuchi, K., and Sié, M. 2009. Better exploitation of African Rice (Oryza glaberrima

Steud.) in varietal development for resource-poor farmers in West and Central

Africa. Agric. J. 4: 96-102.

107

Juliano, B. O., Perez, C. M., Alyo-Shin, E. P., Romanov, V. B., Blakeney, A. B., Welsh, L.

A., Choudhury, N. H., Lilia Delgado, L., Iwasaki, T., Shibuya, N., Mossman, A. P.,

Siwi, B., Damardjati, D. S., Suzuki, H., and Kimura, H. 1984. International

cooperative test on texture of cooked rice. J. Texture Stud. 15: 357-376.

Juliano, B. O. 1985. Criteria and tests for rice grain qualities in: Rice: Chemistry and

Technology, 3rd Ed. St Paul, MN, USA. Am. Assoc. Cereal Chem.: 443-524.

Juliano, B.O. 1998. Varietal impact on rice quality. Cereal Foods World 43: 207-222.

Kesavan, M., Song, J. T., and Seo, H. S. 2013. Seed size: a priority trait in cereal crops.

Physiol.Plantarum. 147: 113-120.

Kim, Y. S., and Wiesenborn, D. P. 1995. Starch noodle quality as related to genotypes.

J. Food Sci. 61: 248-252.

Lawal, O. S., Lapasin, R., Bellich, B., Olayiwola, T. O., Cesàro, A., Yoshimura, M. 2011.

Rheology and functional properties of starches isolated from five improved rice

varieties from West Africa. Food Hydrocoll. 25: 1785-1792.

Leelayuthsoontorn, P., and Thipayarat, A. 2006. Textural and morphological changes of

Jasmine rice under various elevated cooking conditions. Food Chem. 96: 606-613.

Li, J., Thomson, M., and McCouch, S. R. 2004. Fine Mapping of a grain-weight

quantitative trait locus in the peri-centromeric region of rice chromosome 3.

Genetics 168: 2187-2195

108

Li, J., Xiao, J., Grandillo, S., Jiang, L., Wan, Y., Deng, Q., Yuan, L., and McCouch, S.R.

2004. QTL detection for rice grain quality traits using an interspecific backcross

population derived from cultivated Asian (O. sativa L.) and African

(O. glaberrima S.) rice. Genome 47: 697-704.

Meullenet, J-F. C., Carpenter, J. A., Lyon, B. G., and Lyon, C. E. 1997. Bi-cyclical

instrument for assessing texture profile parameters and its relationship to sensory

evaluation of texture. J. Text. Stud. 28: 101-118.

Miller, J.B., Pang, E., and Bramall, L. 1992. Rice: a high or low glycemic index food?

Am. J. Clin. Nutr. 56: 1034-1036.

Mohapatra, D., and Bal, S. 2006. Cooking quality and instrumental textural attributes of

cooked rice for different milling fractions. J. Food Eng. 73: 253-259.

Morrison, W. R., and Azudin, M. N. 1987. Variation in the amylose and lipid contents and

some physical properties of rice starches. J. Cereal Sci. 5: 35-37.

Nanri, A., Mizoue, T., Noda, M., Takahashi, Y., Kato, M., and Inoue, M. 2010. Rice

intake and type 2 diabetes in Japanese men and women: the Japan Public Health

Center-based prospective study. Am. J. Clin. Nutr. 92: 1468-1477.

Ndjiondjop, M. N., Manneh, B., Cissoko, M., Drame, N. K., Kakai, R. G., Bocco, R.,

Baimey, H., and Wopereis, M. 2010. Drought resistance in an interspecific

backcross population of rice (Oryza spp.) derived from the cross WAB56-104 (O.

sativa) × CG14 (O. glaberrima). Plant Sci. 179: 364-373.

109

Noda, T., Nishiba, Y., Sato, T., and Suda, I. 2003. Properties of starches from several low-

amylose rice cultivars. Cereal Chem. 80: 193-197.

Ranawana, D. V., Henry, C. J. K., Lightowler, H. J., and Wang, D. 2009. Glycaemic index

of some commercially available rice and rice products in Great Britain. Int. J. Food

Sci. Nutr. 60:99-110.

Sarla, N., and Swamy, B.P.M. 2005. Oryza glaberrima: a source for the improvement of

Oryza sativa. Curr. Sci. 89: 955-963.

Shimelis, E. A., Meaza, M., and Rakshit, S. 2006. Physico-chemical properties, pasting

behavior and functional characteristics of flours and starches from improved bean

(Phaseolus vulgaris L.) varieties grown in East Africa. Agric Eng Int:

CIGR Journal. 8: 1-18.

Singh, N., Kaur, L., Sodhi, N. S., and Sekhon, K. S. 2005. Physicochemical, cooking and

textural properties of milled rice from different Indian rice cultivars. Food Chem.

89: 253-259.

Sweeney, M., and McCouch, S. 2007. The complex history of the domestication of rice.

Ann. Bot. 100: 951-957.

Thomas, R., Wan-Nadiah, W. A., and Bhat, R. 2013. Physiochemical properties,

proximate composition, and cooking qualities of locally grown and imported rice

varieties marketed in Penang, Malaysia. Int. Food Res. J. 20: 1345-1351.

Tester, R. F., and Morrison, W. R. 1990. Swelling and gelatinization of cereal starches I.

Effects of amylopectin, amylose and lipids. Cereal Chem. 67: 551-557. 110

Vamadevan, V., Bertoft, E., Seetharaman, K. 2013. On the importance of organization of

glucan chains on thermal properties of starch. Carbohydr. Polym. 92: 1653-1659.

Vamadevan, V., and Bertoft, E. 2014. Structure-function relationships of starch

components. Starch/Stärke. 66: 1-14

Watanabe, H., Futakuchi, K., Teslim, I., and Sobambo, B. A. 2002. Milling characteristics

and grain quality traits of interspecific progenies of Asian rice (Oryza sativa) with

African rice (Oryza glaberrima). Jpn. J.Tropical Agric., 46:47-55.

Watanabe, H., Futakuchi, K., Jones, M.P., and Sobambo, B.A. 2004. Grain protein content

of African rice (Oryza glaberrima Steud) lines and Asian rice (O. sativa L.)

varieties in West Africa. Oryza 41: 35-38.

Yu, S., Ma, Y., and Sun, D. 2009. Impact of amylose content on starch retrogradation and

texture of cooked milled rice during storage. J. Cereal Sci. 50: 139-144.

Zaidul, I. S. M., Nik Norulaini, N. A., Mohd Omar, A. K., Yamauchi, H., and Noda, T.

2007. RVA analysis of mixtures of wheat flour and potato, sweet potato, yam, and

cassava starches. Carbohyd. Polym. 69: 784-791

Zhang, G., Ao, Z., and Hamaker, B. R. 2006. Slow digestion property of native cereal

starches. Biomacromolecules. 7: 3252-3258.

Zhang, C.-H., Li, J.-Z., Zhu, Z., Zhang, Y.-D., Zhao, L., and Wang, C.-L. 2010. Cluster

analysis on Japonica rice (Oryza sativa L.) with good eating quality based on SSR

markers and phenotypic traits. Rice Sci. 17: 111-121.

111

Zhu, L.-J., Liu, Q.-Q., Wilson, J. D., Gu, M.-H and Shi, Y.-C. 2011. Digestibility and

physicochemical properties of rice (Oryza sativa L.) flours and starches differing in

amylose content. Carbohydr. Polym. 86: 1751-1759.

112

CHAPTER 5: THERMAL AND MOLECULAR CHARACTERIZATION OF

STARCHES ISOLATED FROM AFRICAN RICE (Oryza glaberrima)

113

Abstract

African rice (Oryza glaberrima) and Asian rice (Oryza sativa) are two distinct types of domesticated rice with the former having been less extensively studied. With the expansion of rice food products in Sub-Saharan Africa, characterization of starches from

African rice is needed for product development. This study was designed to investigate the physical and molecular properties of starches isolated from seven African rice accessions

(ARAs) in comparison with two O. sativa samples; cv Koshihikari (spp. japonica) and cv

WITA 4 (spp. indica) and an inter specific cross between sativa x glaberrima cv NERICA

4. The granule size distribution of the starches examined ranged from 2-22 μm and showed tightly packed polyhedral shaped granules. A wide range of amylose content (17-30%) was found among rice starches with the ARAs having relatively higher proportions. The energy required to gelatinize starches or melt retrograded starch gels varied among starches and was observed to be lower in NERICA 4 and the Asian rice varieties. Significant

(p < 0.05) differences in the peak, trough, final, breakdown and total setback viscosities were observed among the starches. The starches also showed significant (p < 0.05) differences in the ratio of the absorbance to scattering coefficient (K/S) values when exposed to iodine vapor, suggesting differences in the rigidity of the glucan chains in the granules. Starches from all the ARAs had higher K/S values than the other samples.

Additionally, all the starches displayed the typical 'A' type x-ray diffraction pattern. The significant differences in physico-chemical properties observed between the ARAs and

Asian rice varieties are likely due to differences in amylose content which may influence the cooking and eating quality of the ARAs. The results also suggest that the structure and granular architecture of the ARA starches may differ from the Asian rice starches.

Keywords: Rice starch, iodine, molecular characteristics, physical characteristics, x-ray, amylose

114

5.1 Introduction

Rice is one of the main cereal crops cultivated worldwide and millions depend on it as food staple. Oryza sativa (Asian rice) and Oryza glaberrima (African rice) are the two distinct types of domesticated rice with O. sativa having a global distribution particularly in Asia.

However, African rice, which has the ability to withstand a wide range of environmental stress conditions, is cultivated mainly in the West Africa sub region (Sweeney and

McCouch 2007). Factors such as consumer preferences, genetic character and variation in climatic conditions have led to the development of many rice cultivars and the diversity in these rice cultivars is associated with different cooking qualities and physical properties

(Wani et al. 2012). Of major interest is starch, the predominant component of rice (90% dry basis) and its physicochemical properties, including apparent amylose content, amylopectin structure, gelatinization temperatures and pasting viscosities that determine the eating, cooking and milling qualities of rice. Several benefits of rice starch over other starch sources (e.g. hypo-allergenicity, creaminess and spreadability, smoothness, bland flavor, small granules, acid resistance, and relatively good freeze-thaw stability of its gels) make it a vital ingredient in the formulation of a variety of food products and other applications (Lawal et al. 2011; Wani et al. 2012). These exceptional attributes make rice starch one of the most suitable for industrial purposes (Vandeputte and Delcour 2004).

Since the development of new rice cultivars affects physico-chemical properties of starch as reported by several researchers (Ahmed et al. 2008; Lawal et al. 2011; Wang et al.

2012), it is expedient that the characteristics and functionalities of starches from new cultivars are studied so that those with desirable functional and physico-chemical properties are selected for specific end uses.

115

The NERICA (NEw RICe for Africa) rice varieties (an interspecific cross between

O. sativa x O. glaberrima) developed in the 1990s and disseminated to farmers has contributed to poverty reduction in sub Saharan Africa. The development of this variety harnessed desirable qualities such as high tillering ability, rapid seed growth and early maturity. Despite the success of this project, it is believed that African rice has not been fully exploited in terms of other traits that were not focused on during the varietal improvement. These include amylose content (and for that matter starch), which is a major factor that influences texture and taste of cooked rice (Futakuchi and Sié 2009). To this end, a collection of African rice accessions is being systematically characterized by Africa

Rice Center (AfricaRice) for resistance against major diseases and environmental stresses.

Objective: Since the basic structure of African rice starches have not been characterized yet, this study focuses on investigating the basic physical, thermal and molecular characteristics of seven African rice accessions that were selected from a population of

1,020 through cluster analysis. The cluster analysis of the 1,020 accessions was based on apparent amylose and pasting data of their flour.

116

5.2 Materials and Methods

5.2.1 Samples

Rice samples were supplied by the Africa Rice Center based in Cotonou, Benin. They consisted of seven African rice accessions (ARAs) - IRGC 86741, IRGC 80787, IRGC

103759, RAM 123, TOG 7682, TOG 12440 and TOG 14365. In addition were NERICA 4

(an interspecific cross between O. sativa x O. glaberrima) and two O. sativa varieties – cv. Koshihikari (spp. japonica) and cv.WITA 4 (spp. indica) for comparison purposes

5.2.2 Methods

5.2.3 Starch Isolation

Starch was isolated from the rice samples using the method of Waduge et al. (2010) with some modifications. The rice kernels were frozen in liquid nitrogen for 2 min and milled for 1 min using a Coffee and Spice Grinder with stainless steel grinding blades,

(SmartGrindTM model CBG100SC - Black and Decker Corporation, Towson, Maryland,

USA) into flour. Fifty grams of rice flour was then transferred into 1 L sodium borate buffer (12.5 mM, pH 10 and containing 0.5% sodium dodecyl sulfate (SDS) {w/v} and

0.5% Na2S2O5 {w/v}). The flour-buffer suspension (1:20 {w/v}) was stirred at a moderate speed for 5 min to extract proteins. The suspension was subjected to centrifugation at

900 × g for 5 min and the residue recovered. The protein extraction step was repeated, and the residue was washed three times with double distilled water and recovered by centrifugation (900 × g for 5 min) each time. Thereafter, the residue was suspended in double distilled water and the starch slurry was first passed through four layers of 117

cheesecloth to trap coarse particles and then through nylon material with a mesh size of 70

μm. The coarse mass in the cheesecloth after the filtration was further ground into finer particles to recover more starch. The starch-water suspension was centrifuged at 1600 × g for 10 min, after which the brown protein layer that settled on top of the starch layer was carefully removed with a spatula. The starch was then suspended in double distilled water again and subjected to centrifugation. This step in the isolation process was repeated till all the particles that formed the brown layer were removed from the surface of the starch residue. The starch fraction was finally washed with ethanol (98%) followed by acetone and allowed to air dry in a fume hood. Starch yield was expressed as the dry weight per weight of rice flour.

5.2.4 Physical Characteristics:

5.2.5 Starch Damage

AACC Method 76-31.01 was employed in measuring level of starch damage in the rice starch samples using the Megazyme Starch Damage Test Kit (Megazyme International Ltd,

Bray, Ireland).

5.2.6 Nitrogen

The Dumas combustion method for the quantitative determination of total nitrogen was employed to measure the nitrogen content of rice starch samples. Weighed starch samples

(0.2 g) were combusted in a LECO Nitrogen/Protein Analyzer (LECO, FP-528 Lakeview

Avenue, St. Joseph, Michigan, United States) after ethylenediaminetetraacetic acid

118

(EDTA) standard was run to calibrate the equipment. Nitrogen was converted into protein by multiplying the nitrogen determined by a factor of 5.7.

5.2.7 Particle Size Determination

Particle size distribution of starch samples was determined with Mastersizer 2000 (Malvern

Instruments Ltd, Worcestershire, UK). Before analysis was carried out, the motor speed was set to 2000 rpm and the dispersion unit was flushed with distilled water. The tank was completely flushed three times with tap water and once with distilled water. The tank was refilled with distilled water and it was ensured that no bubbles were left in the tubes to and from the sample chamber. Starch samples were added with care to avoid lump formation and in enough quantities to produce laser obscuration range between 12% and 14%. A refractive index of 1.52 was used for the starches in the dispersed phase, 1.33 for the aqueous continuous phase and sample absorption of 0.001. A 300 Reverse Fourier

(convergent beam) lens was employed for the analyses. The Mastersizer 2000 software controls the system during the measurement process and analyzes the scattering data to calculate a particle size distribution.

5.2.8 Morphology

5.2.9 Scanning Electron Microscopy

The rice starch granule characteristics was studied with Hitachi S-570 Scanning Electron

Microscope (Hitachi Scientific Instruments, Rexdale, Ontario,Canada). The starch samples were initially thinly spread on double-sided adhesive tapes attached to aluminum specimen stubs. The samples were then coated with Gold/Palladium dust to a thickness of 15 nm in

119

approximately 2 min. using Emitech K550X Sputter Coater (Quorum Technologies Ltd,

Ashford, Kent, United Kingdom). The working distance used on the scanning electron microscope was 15 mm with an accelerating voltage of 10 kV. Samples were placed onto the stage mount using stub tweezers and tightened with Allen key. Images were brought into focused by increasing the magnification slowly and adjusting the scan speed until the images were sharp and stretches removed. Digital images of the samples were then acquired.

5.2.10 Wide Angle X-ray Diffraction Scattering

Sample preparation was done by hydrating 0.1 g quantities of the isolated rice starches over a saturated solution of K2SO4 with aw 0.97 in an air-tight desiccator at room temperature for 30 days. The hydrated starches were then tightly packed by firmly pressing samples into the groove of a quartz plate sample holder. X-ray diffraction measurements were obtained with a Rigaku X-ray diffractometer (Rigaku-Denki Co., Tokyo, Japan). The operating conditions included a target voltage of 40 kV, a current of 44 mA, step time of

0.95 s and scanning range was from 3-35° at a speed of 1.00°/min. The divergence slit width, sampling width and scatter slit width were 0.5°, 0.5° and 0.03°, respectively. Jade software (Material Data Inc. CA, USA) was used to estimate the relative crystallinity by engaging a Gaussian function for curve fitting. The ratio of the area under the peaks to the total area under the curve (peaks + amorphous background) was reported as percent crystallinity of the starches.

120

5.2.11 Reﬂectance Spectra after Exposure of Starch to Iodine Vapor

The reﬂectance spectra and color {(L* (white/black), a* (green/red), and b* (blue/yellow)} of the samples was determined using the Kubetta and Munk equation:

( ) = 2 Equation 1 K 1−R S 2𝑅𝑅 where K is the absorption coefficient, S is the scattering coefficient and R is the reﬂectance expressed as a fraction between 0 and 1 (Billmeyer and Saltzman 1981). The K/S values of the samples (representing the absorption corrected to the light scattering) were measured over a wavelength range of 400 to 700 nm using a CM 3500-d spectrophotometer (Konica

Minolta Sensing Inc., Mahwah, NJ, USA) equipped with SpectraMagic NX CM-S 100 software at 10 nm intervals. Exactly 0.2 g of sample equilibrated to aw 0.97 and exposed to iodine for 24 h was spread over the sample cell as a thick layer and the measurements were taken. Sample equilibrated to the same aw but not exposed to iodine vapor was used as a reference.

5.2.12 Thermal Properties

5.2.13 Rapid Visco Analyzer

The pasting properties of rice starch samples were measured using Rapid Visco Analyzer

for 2.5 min. The rotating speed of the paddle was maintained at 160 rpm throughout the run except for the first 10 s where paddle speed was held at 960 rpm. The under listed pasting parameters were determined from the RVA pasting curve; peak viscosity

5.2.14 Differential Scanning Calorimetry

TA Instruments, Q1000 differential scanning calorimeter (DSC) equipped with a thermal analysis data and recording software facility (TA Instruments, Universal Analysis 2000,

DE, USA) was used to measure the gelatinization properties of the rice flours. Starch dispersion in water (1:3) was allowed to equilibrate for at least 3 h at room temperature before sample was loaded. In all measurements, the thermogram was recorded with an empty aluminum pan as a reference and the scanning temperature range was programmed from 20-120 °C at a heating rate of 10 °C/minute. Scanned samples were stored at 4 °C for

122

5.2.15 Molecular Composition of Rice Starches

5.2.16 Chain Length and β-Limit Dextrins

Size distribution of starch components and their β-limit dextrins were investigated using

Gel Permeation Chromatography (GPC). Rice starch samples (8 mg) were dissolved in

90% DMSO (200 μL), heated in warm water (80 °C) for 5 min and then left overnight to stir at room temperature (25 °C). Warm water (800 μL) was used to dilute the samples.

The size-distribution of the components of the starch samples was determined by fractionating them (400 μL) on a Sepharose CL-2B column (1.6 × 32 cm) (GE-Healthcare,

Uppsala, Sweden). The eluent (0.01 M NaOH) was pumped through the column at a rate of 0.5 mL/min. The total carbohydrate content of the fractions collected (1 mL) was determined by using the phenol-sulfuric acid reagent (Dubois et al. 1956). The wavelength maxima (λmax) of the glucan-iodine complex was determined with a WPA Spectrawave

S800 Diode Array Spectrophotometer (Biochrom Ltd., Cambridge, United Kingdom) by initially neutralizing the collected fractions with 1 mL of 0.01 M HCl and thereafter adding

100 μL of 0.01 M I2-0.1 M KI solution. β-Limit dextrins (β-LD) of the rice starches were prepared according to the method outlined by Bertoft (2004). Starch (8 mg) was dissolved in 90% DMSO (200 μL) and left overnight to stir. β-amylase (4 U/mg, 2 μL) in 100 μL

(0.01 M NaOAc buffer; pH 6) was thereafter added and the resulting mixture was again stirred overnight at room temperature (25 °C). Water was added to dilute the mixture to 1 mL and then 400 μL was injected onto the Sepharose CL-2B column.

123

5.2.17 Chain Length of Debranched Starches

GPC as described by Sargeant (1982) was used to analyze debranched rice starches.

Samples (4 mg) were dissolved in 90% DMSO (100 μL) by heating gently in a water bath

(80 °C) for 5 min. The dissolved starch was allowed to gently stir overnight at room temperature (25°C). The solution was diluted with water (80 °C) (750 μL) and thereafter sodium acetate buffer (0.01 M, pH 5.5, 100 μL) was added. Exactly 1 μL each of isoamylase (465 U/mL) and pullulanase M1 (925 U/mL) (Megazyme International,

Wicklow, Ireland) was added and the mixture incubated overnight with gentle stirring at room temperature (25°C). The reaction was halted by placing samples in a boiling water bath for 5 min. Fifty μL of 5 M NaOH was added after which the sample was diluted to

1.5 mL with deionized water and filtered. The filtered sample (1 mL) was applied on a

Sepharose CL-6B column (1 × 90 cm; GE Healthcare, Uppsala, Sweden) with NaOH

(0.5 M) as eluent pumped at a rate of 1 mL/min. The total carbohydrates content of the fractions collected (1 mL) was analyzed by using the phenol-sulfuric acid reagent (Dubois et al. 1956). The amylose and amylopectin portions of the samples were determined by dividing the chromatogram at the lowest point between the two peaks as reported by

(Sargeant 1982). The amylose fraction was further divided into short and long chains.

5.2.18 Amylose

The amylose concentration of the samples was determined using GPC as described under section 5.2.17 above.

124

5.2.19 Statistical Analysis

The statistical programme Statgraphics Centurion XV, Version 15.1.02 (StatPoint,

Warrenton, VA, USA) was employed to conduct statistical analysis. One-way analysis of variance was performed to determine significant differences between any pair of means at the 95.0% confidence level and multiple range tests were done to determine which means were different.

5.3 Results

5.3.1 Starch Yield and Composition

The recovery of starch based on the whole weight of the dehulled rice samples is presented in Table 5.1. The recoveries were in the range of 65.7-72.9%. Protein and fat contents of all the starch samples were below 1% attesting to their purity. The starch damage level in all the samples was 1.2% or less (Table 5.1) indicating that the milling step in the starch extraction process did not adversely affect the structural integrity of the starch granules.

5.3.2 Morphology and Structure of Rice Starch Granules

Scanning electron micrographs of all the samples revealed that the morphologies of their granules were similar. The granules were found to be of similar sizes, angular and polygonal/polyhedral in shape and possessed the same tight mode of packing (Figure 5.1).

Grooves that seem to be points of attachment between the granules were observed on the surfaces of the granules. The granule size distribution curves of the samples displayed unimodal shapes with peak size of about 5-6 μm (Figure 5.2). The granule sizes of the starch samples ranged from 2-22 μm with Koshihikari, WITA 4 and RAM 123 having

125

similar size distributions with peak size of about 5 μm. The other samples possessed a peak size of approximately 6 μm, with the exception of TOG 14365, which had a peak size of 6.6 μm.

The X-ray diffraction patterns of the native and iodinated rice starches are illustrated in

Figures 5.3 and 5.4 respectively. The diffractograms showed regions of noticeable peaks centered on 2θ 15°, a doublet on 17° and 18° and another single peak at 23°. This diffraction pattern is consistent with cereal starches generally classified as type “A”,

(Imberty and Perez 1988). All the starch samples in this investigation displayed the typical

‘A’ type x-ray diffraction pattern. Exposure of the starch samples to iodine vapor resulted in changes to some characteristics notably the relative crystallinity. There was a significant increase in the relative crystallinity of all the iodinated starches by 2-5%, except for TOG

7682 (Figure 5.4). The diffraction pattern was not affected by the interaction between the iodine and the starch samples, however, shifts in some of the peaks were observed. The reduction in the intensities of peaks at 15°, 17° and 18° 2θ° were more obvious in NERICA

4 and all the ARAs than in WITA 4 however, no such effect was seen with Koshihikari. At

20° 2θ there was an increase in the peak intensity for all the starches due to the formation of a single helical complex with iodine but the 23° 2θ peak remained intact for all the starch samples.

The reflectance spectra of all the iodinated starches showed λmax at 550 nm (Figure 5.5).

Significantly higher K/S intensity values (p < 0.05) were observed for all the ARAs in comparison to the other samples over the entire wavelength range that measurement was taken (400-700 nm). The differences in the K/S values signify variances in the population of glucan polymers available in the granules of the starches to form inclusion complexes

126

with iodine (Saibene et al. 2008). At λmax 550, the K/S intensity values were highest for

TOG 14365 (35.2) and lowest for Koshihikari (13.6) and, interestingly, all the ARAs were higher in K/S value than NERICA4, WITA 4 and Koshihikari.

5.3.3 Thermal Properties of Starch Granules

The RVA profiles and pasting characteristics of the starch samples are presented in Figure

5.6 and Table 5.2 respectively. The samples had a peak time ranging from 5.0 min (TOG

14365) - 5.7 min (IRGC 86741). Pasting occurs when the starch granules absorb sufficient water and swell after gelatinization. The pasting temperatures differed significantly (p <

0.05) and ranged from 71.1 °C (Koshihikari) to 78.6 °C (IRGC 80787). The peak viscosity of the starches ranged from 231.5 to 308.9 RVU corresponding to RAM 123 and

Koshihikari, respectively. The trough viscosity, which is a measure of the ability of the paste to withstand breakdown during the cooling phase, was from 134.8 RVU (RAM 123) to 187.6 RVU (IRGC 80787) while final viscosity, which is the change in the viscosity after holding cooked starch at 50°C, ranged from 266.3 to 332.3 RVU corresponding to

Koshihikari and IRGC 80787, respectively. Peak viscosity indicates the water-holding capacity of the starch, while final viscosity, the most commonly used parameter to define a particular sample’s quality, indicates the ability of the material to form a viscous paste or gel after cooking and cooling. Breakdown viscosity was lowest for IGRC 86741 (68.7

RVU) and highest for Koshihikari (124.8 RVU). The total setback viscosity, which is a measure of the tendency of starch paste to retrograde, was from 82.2 to 166 RVU corresponding to Koshihikari and IRGC 103759, respectively.

127

Table 5.3 and the bar graph in Figure 5.7 show the transition temperatures (To, Tp, Tc) and gelatinization enthalpies (ΔH) of the native, granular starches. The onset temperatures of the ARAs and NERICA 4 were significantly higher than WITA 4 and Koshihikari. While

WITA 4 recorded the lowest To of 59.8 °C, TOG 14365 had the highest To of 69.1 °C. A similar trend was observed for both Tp and Tc. Gelatinization enthalpy (ΔH) of starch granules ranged from 13.3-15.8 J/g (Table 5.3) with the two Asian rice starches having the lowest ΔH values. ΔH and the narrow melting range (Tc-To) of the ARAs suggests that their amylopectin crystals are more homogeneous and uniformly stable in comparison to the Asian rice (Ratnayake et al. 2001).

Figure 5.8 and Table 5.4 show the properties of the retrograded starch gels. To, Tp, Tc and

ΔH values were considerably lower than their respective native starches. To and Tp ranged from 42.9 to 44.1 °C and 52.5 to 54.6 °C, respectively. While ΔH for the retrograded starches ranged from 6.4-10.8 J/g, Tc was from 65.5 to 72.3 °C. While To, Tp, and ΔH of the gelatinization of the native granules were considerably higher for the ARAs than the

Asian rice starches (Figure 5.7), differences for the retrograded starch samples occurred mainly in Tc and ΔH (Figure. 5.8).

5.3.4 Molecular Structure of Rice Starches

The amylose concentration of the starch samples ranged from 17% for Koshihikari to

31.1% for TOG 14365 (Table 5.1). These values were obtained by analysing their chain profiles with GPC on Sepharose CL-6B (Figure 5.9). It was observed that all the ARAs and NERICA4 had significantly higher proportions (p<0.05) of amylose than the Asian rice varieties. The gel permeation chromatograms also showed the relative proportions of long

128

(LCAM) and short chains (SCAM) of amylose (Figure 5.9). The relative amounts and ratios of LCAM:SCAM of the starches are presented in Table 5.1. The ARAs generally had significantly higher amounts of both long and short chains of amylose (p< 0.05) than

Koshihikari and WITA 4. All starches, except Koshihikari, recorded LCAM:SCAM ratios of

0.6 to 0.74 due to higher amounts of short amylose chains than long amylose chains.

Koshihikari possessed a high LCAM:SCAM ratio of 0.87, because it had nearly equal amounts of LCAM and SCAM.

Gel permeation chromatograms of whole rice starches and their corresponding β-limit dextrins (β-LDs) fractionated on Sepharose CL-2B along with their carbohydrate contents and λmax values are presented in Figure 5.10. The elution profiles of the chromatograms showed a bimodal distribution with the first/ narrow peak, which corresponds to the large amylopectin molecules, eluted at the void volume of the column. The second/ broad peak representing amylose (smaller glucan molecules) eluted after the void volume of the column. The λmax values of the samples in the amylopectin portion ranged from about 560 to 585 nm. In this region the ARAs and NERICA 4 recorded higher λmax values than

Koshihikari and WITA 4. This reflects the presence of long external chains which are available to complex with iodine in the ARAs and NERICA 4 (Table 6.1). A similar observation was made in the amylose region and this suggests that the ARAs and NERICA

4 have longer amylose chains available for iodine complexing than Koshihikari and WITA

4. The λmax values of the β-LDs from all the samples were observed to be lower than the corresponding whole starches in the amylose region. However, in the amylopectin region higher λmax values were observed for β-LDs from Koshihikari. It is important to note that iodine can form complexes with external and internal chains of amylopectin (Chauhan and

129

Seetharaman 2013; Vamadevan et al. 2015) but after β-amylolysis, the relative abundance of long internal chains of amylopectin could shift the spectrum to higher λmax values.

Additionally, the availability of internal glucan chains for iodine binding could be influenced by the nature of branching pattern. The maltose peak was the result of the removal of external chains of the branched glucans and linear polymers of the starch by β- amylase.

5.4 Discussion

The alkaline extraction procedure used in this study resulted in recoveries between 65.7 and 72.9%. A wide range of variation in rice starch recovery from different rice cultivars

(59% to 71.6%) have been reported (Mohan et al. 2005; Wang and Wang 2004). Yields of

70.0% to 73.77% were reported for five newly released improved rice varieties from West

Africa (Lawal et al. 2011) and the yields in this investigation compares well. Lawal et al.

(2011) reported that starch recovery is a function of the genetic composition of the rice itself, the environment under which it was cultivated and the starch extraction method employed. The method of starch isolation can significantly affect the pasting, rheological and textural properties of starch paste (Zhong et al. 2009). Isolation of starch from rice is quite different from other starches due to its unique protein composition (Wani et al. 2012).

The separation of starch from protein, fiber and lipid are the key elements in the isolation process and a number of important steps have to be considered. These include avoidance of amylolytic or mechanical damage to the starch granules, effective deproteinization of the starch, minimizing loss of small granules and avoidance of starch gelatinization (Schulman and Kammiovirta 1991).

130

Starch damage commonly occurs during milling of cereal grains and examples of the damage are broken starch granules with exposed interior and granule fragments (Dhital et al. 2010; Hasjim et al. 2009). Reports indicate that different milling or grinding processes result in different degrees of damage to starch granules in flour depending on the mechanical forces and temperature during the grinding process (Nishita and Bean 1982;

Nowakowski et al. 1986; Hatcher et al. 2002). Starch swelling can be improved through water absorption when there is a low degree of damage to starch granules but excessive damaged starch granules have increased solubility in cold water than the intact starch granule and undergo less swelling during cooking. This is due to the degradation of starch molecules (mainly amylopectin) and starch crystalline structure that occurs during grinding, leading to the production of highly soluble, low molecular-weight molecules,

(Tester 1997; Morrison et al. 1994; Becker et al. 2001). Since milling is necessary to disrupt protein and cell wall structures to facilitate the release of starch granules from the grain matrices, it is essential that care is taken not to cause damage to starch molecules for better starch characterization (Gidley et al. 2010). Damaged starch is known to affect thermal characteristics of starch. High levels cause a decrease in gelatinization temperature and enthalpy (Morrison et al. 1994). Starch damage levels of less than 1.2% were obtained for all samples in this study.

The morphological characteristics of the starch granules analyzed in this study (Figure 5.1) were largely consistent with what has earlier been reported in literature (Li and Yeh 2001;

Singh et al. 2003; Raina et al. 2007; Lawal et al. 2011; Wang et al. 2012) except for the wider granule size distribution of 2-22 μm. Using a laser particle size analyzer, Wang et al.

2012 found no significant distinctions between various rice starches in terms of their size,

131

shape and distribution. The range of median size for indica and japonica rice starch granules, which were angular and polygonal in shape, were reported to be around 6.2-7.7

μm and 6.7-7.8 μm, respectively. Using the same procedure for determining size distribution, Li and Yeh (2001) reported a size range of 2-10 μm with an average granule size of 6.4 μm for Taiwan rice starch granules, which had polyhedral shape. Polyhedral shaped granules with polygonal edges having size ranging between 3.1 and 7.8 μm was reported by Raina et al. (2007) for Oryza sativa rice varieties using scanning electron microscopy (SEM). Similarly, Lawal et al. (2011) used SEM to study starch isolated from five improved rice varieties from West Africa and reported granules with irregular polyhedral shapes that had sizes ranging from 1.5 to 6.1 μm. The granular structure of rice starches vary in shape and size among different cultivars and rice types; non-waxy, waxy, and long-grain (Wani et al. 2012).

All the starch samples investigated in this study had similar diffraction patterns, A-type crystal. Lawal et al. (2011) reported X-ray diffraction patterns for NERICA that showed no doublet but a single peak at 2θ 17.1° and a less intense and broader peak at 2θ 23.1°.

This was inconsistent with the findings on the NERICA 4 sample in this study. The relative crystallinity (Figure 5.4), which was calculated as the ratio of the area under crystalline peaks to the total area under the diffractogram (representing both crystalline and amorphous structures), did not differ significantly among the samples (p<0.05). The relative crystallinities of the samples in this study were very similar to that of improved rice varieties from West Africa as reported by Lawal et al. (2011), but marginally lower than the crystallinity of indica rice starches with various amylose contents (He et al. 2006).

The observed decrease in the peak intensities of the X-ray diffractograms of the starch

132

samples at 2θ 15°, 17° and 18° after exposure to iodine vapor (Figure 5.4) have been reported in other cereal starches. Saibene et al. (2008) reported a decrease in peak intensity at 2θ 15°, 17°, 18° and 23° of corn starch and a loss of resolution in peaks around 2θ 22-

24° in potato starch. Similarly, Annor et al. (2014) and Cheetham and Tao (1998) found a decrease in peak intensities in millet starches and normal maize starches, respectively, with the disappearance of some peaks in high amylose maize starch. While the iodine did not seem to have any observed effect on 2θ 23° in all the starch samples, the decrease in intensities of the characteristic peaks 2θ 15°, 17°, 18° was more pronounced in NERICA 4 and the ARAs (Figure 5.4) than the two Asian rice starches (Koshihikari and Wita 4). The decrease in peak intensities is thought to be due to the absorption of X-rays by high molecular weight iodine molecules (Saibene et al. 2008).

The presence of glucan polymers in starch granules and their ability to form single helices in the presence of iodine determine the extent of interaction with iodine (Waduge et al.

2010). The capacity of iodine to form inclusion complexes depends on the flexibility and availability of the glucan polymers within the granule, while penetration of iodine into the starch granule most likely depends on surface features of the starch granules such as pores or channels (Kim and Huber 2008). The K/S intensity value (defined as the ratio of light absorbance coefficient to light scattering coefficient) is indicative of differences in the flexibilities of the glucan chains within the starch granules to form inclusion complexes with iodine (Saibene et al. 2008). The higher K/S intensities of the starches from the ARAs suggest that the population of flexible glucan polymers available to form inclusion complexes with iodine was substantially greater than in the Asian rice and NERICA 4 starches (Figure 5.5). The formation of iodine-glucan complex when the granular starches

133

were exposed to iodine vapor affected their crystallinity to different extents. K/S at λmax

550 nm correlated highly (0.9389) with amylose content at p < 0.05 (Table 5.5) and the equation of the fitted model is

K/S at λmax550 = -12.6625 + 1.53405*amylose.

Pasting properties of starch are important indicators of how the starch will behave in the course of processing and they are used in determining the suitability of starch in different foods and other applications (Wani et al. 2012). When starch is continuously heated in excess water with stirring, granules swell and burst as a result of the breakdown of the native starch structure. Amylose then leaches out of the disintegrated granules leading to the formation of a homogeneous mass called paste (BeMiller 2007). All the starch samples in this study exhibited similar pattern and trend but significant differences were found among the parameters that were measured (Figure 5.6). The temperature at the onset of viscosity increase (pasting temperature) was significantly lower in the two Asian rice starches than in the ARAs and NERICA 4 (Table 5.2). Adebowale et al. (2005) reported that breakdown viscosity measures the susceptibility of the cooked starch to fragmentation and that the higher the breakdown in viscosity, the lower the ability of the starch sample to withstand heating and shear stress during cooking. The high breakdown viscosity exhibited by Koshihikari suggests that it will have a low capacity to withstand heating and shear stress compared to other starches. Setback viscosity is commonly used to describe the increase in viscosity that occurs on cooling a starch paste (Fisher and Thompson 1997).

It is defined as the difference between the breakdown viscosity and the viscosity at 50 °C and it determines the retrogradation tendency of the paste (Zaidul et al. 2007). The high amylose content of the ARAs and NERICA 4 starches explains their high setback and

134

breakdown values. In this study amylose is correlated (p<0.05) with the pasting parameters measured. The correlations between amylose and peak, final, setback and total setback viscosities at (p<0.05) were -0.65, 0.51, 0.87 and 0.90 respectively (Figure 5.11, Table 5.5) and the equation of the multiple linear regression model to describe the relationships are

Peak viscosity = 358.423 - 3.70361*amylose, Final viscosity = 226.873 +

2.74866*amylose and Total setback = 3.98736 + 5.06745*amylose respectively.

The transition temperatures (To, Tp and Tc) and enthalpies of gelatinization (ΔH) of the starches in this study are in line with other published data. To values ranging from 53.3°C to 75.9°C have been reported for non-waxy rice starch (Wang et al. 2010; Wickramasinghe and Noda 2008), likewise Tp ranging from 61.8°C to 80°C (Wang et al. 2010;

Wickramasinghe and Noda 2008). Tc from 70.9°C to 85.4°C (Wang et al. 2010) and ΔH from 3.7 J/g to 17.7 J/g (Singh et al. 2007; Vandeputtee et al. 2003) have been reported.

Yamin et al. (1999) reported that variations observed in To, Tp, Tc and ΔH in starches from different rice cultivars might be attributed to differences in amounts of longer chains in amylopectins. These longer chains require a higher temperature to completely dissociate than that required for shorter double helices. Additionally, Vamadevan et al. (2013) also reported that the internal structure of amylopectin correlates with To of starches from different botanical sources. The higher transition temperatures and ΔH observed in the

ARA starches compared to the Asian rice starches may therefore be due to differences in their amylopectin internal structures. Chung et al. (2011) reported that the differences in

To, Tp, Tc and ΔH in starches from different rice cultivars may be attributed to differences in amylose content, granular structure, molecular weight, and amylose/amylopectin ratio.

135

A linear relationship was observed between the amylose content of the starches and their gelatinization temperatures. Correlation between amylose and To, Tp, Tc and ΔH at

(p<0.05) was 0.74, 0.73, 0.7 and 0.64 respectively (Table 5.5). Figure 5.12 shows the multiple linear regression models to describe the relationship between To, Tp, Tc, ΔH and amylose and the equation of the fitted models. When the gelatinized starches were stored for 14 days at 4°C to monitor retrogradation, To, Tp, Tc and ΔH for the retrograded gels were found to have reduced drastically when compared to the native granules. The percentage retrogradation (calculated as ΔH of retrograded gel/ΔH of starch granules x

100) was between 44.8 and 71.5% (Table 5.4) and the least retrogradation was observed in

Koshihikari starch possibly because it had the lowest amylose concentration. The amylose content of the starches reported in this study (Table 5.1) was consistent with the range reported by Lawal et al. (2011) for improved rice varieties in West Africa and within the range of values reported for rice starches from other sources (Peisong et al. 2004;

Wickramasinghe and Noda 2008; Wang et al. 2010; Yu et al. 2012). Generally, the ratio of amylose to amylopectin is crucial in affecting many physico-chemical parameters of starches, since the amorphous component of starches is made up mainly of amylose,

(Lawal et al. 2011).

5.5 Conclusion

This study sheds light on the structural and thermal characteristics of starches isolated from seven African rice accessions that were selected from a population being screened for further development at the Africa Rice Center. These characteristics provide important information for useful applications of the starches and may also influence the advanced

136

selection of accessions of interest for the development of suitable varieties for cooking quality and specific end use. The Asian rice starches, apart from having similar physical and morphological characteristics as ARA starches, differed in their molecular composition

(amylose content) and structure, which might have influenced their functional properties.

But, NERICA 4 in many ways had properties similar to the starches from the ARAs. The differences in physicochemical properties that were observed are likely due to differences in amylose content. The results also suggest differences between the structure and granular architecture of starches from the ARAs and Asian rice varieties.

5.6 Acknowledgements:

A special appreciation is hereby extended to the Global Rice Science Partnership (GRiSP) and Africa Rice Center (AfricaRice) for the provision of funding and rice samples for this work.

137

5.7 Tables and Figures

Table 5.1. Characteristics and Composition of Starches Isolated From African Rice Accession and Asian Rice Varieties# Starch Starch AMP AM LC : Sample yield damage LC SC AM (%) (%) AM AM SC (%) (%) AM African rice IRGC 80787 70.9 1.02a 71.3 28.7d 11.2cd 16.9c 0.67bcd

IRGC 86741 66.5 1.16d 74.5 25.5c 10.8c 14.7b 0.74de

IRGC 103759 68.2 1.18def 71.8 28.2d 10.7c 17.5cd 0.62ab

RAM 123 72.3 1.21f 70.3 29.7de 10.8c 19.0d 0.57a

TOG 7682 72.9 1.06b 71.5 28.5d 11.0c 17.5cd 0.63abc

TOG 12440 65.7 1.02ab 71.8 28.2d 10.8c 18.0cd 0.60ab

TOG 14365 67.2 1.11c 68.9 31.1e 12.2d 17.5c 0.70cd

O. sativa x O. glaberrima cross

NERICA 4 69.7 1.17de 76.2 23.8bc 10.5bc 13.2b 0.8ef

Asian rice Koshihikari 71.7 1.20ef 83.0 17.0a 7.9a 9.1a 0.87f

WITA 4 69.2 1.15cd 77.1 22.9b 9.5b 13.4b 0.71d

AM – Amylose AMP - Amylopectin LCAM – Long Chain Amylose

SC AM – Short Chain Amylose # Values with the same letters are not significantly different within columns

138

Table 5.2. Pasting Properties of Native Starches Isolated From African Rice Accessions and Asian Rice Varieties#

PKT PT PV TV FV BDV SB TSB Sample (min) (°C) (RVU) (RVU) (RVU) (RVU) (RVU) (RVU) African rice IRGC 80787 5.2b 78.6e 277.5e 187.6e 332.3d 89.9b 54.8de 144.7cde

IRGC 86741 5.7d 77.8cd 251.4cd 182.7e 315.0c 68.7a 63.7e 132.4bc

IRGC 103759 5.2b 78.1d 253.6cd 145.4ab 311.4c 108.2de 57.8de 166.0g

RAM 123 5.0a 77.8cd 231.5a 134.8a 274.0ab 96.8c 42.5c 139.3cd

TOG 7682 5.2b 77.8cd 263.3d 165.5cd 318.2c 97.8c 54.9de 152.7ef

TOG 12440 5.2b 78.9e 235.6ab 139.3ab 285.0b 96.3c 49.4cd 145.7de

TOG 14365 5.0a 77.9cd 253.5cd 149.8b 308.4c 103.7d 54.9de 158.6fg

O. sativa x O. glaberrima cross NERICA 4 5.3b 77.7c 284.5e 175.1de 307.7c 109.3e 23.2b 132.5bc

Asian rice Koshihikari 5.2b 71.1a 308.9f 184.1e 266.3a 124.8f -42.5a 82.2a

WITA 4 5.4c 73.1b 248.3bc 153.3bc 275.0ab 94.9bc 26.8b 121.7b

PKT= Peak Time; PT= Pasting Temperature; PV= Peak Viscosity; TV= Trough Viscosity;

FV= Final Viscosity; BDV= Breakdown Viscosity; SB= Setback Viscosity; TSB= Total Setback Viscosity # Values with the same letters are not significantly different within columns

139

Table 5.3. Gelatinization Parameters of Native Starches Isolated From African Rice Accessions and Asian Rice Varieties#

Sample To (°C) Tp (°C) Tc (°C) ΔH (J/g) MR (Tc-To) African rice IRGC 80787 69.0ef 73.6d 83.0g 15.1bcd 14.0

IRGC 86741 68.6cd 72.7c 81.5cd 15.1bcd 12.9

IRGC 103759 68.6c 73.6d 82.2f 15.8d 13.6

RAM 123 68.5c 72.9c 81.8e 15.2bcd 13.3

TOG 7682 68.8cde 73.3d 81.7de 15.3bcd 12.9

TOG 12440 68.9def 73.3d 82.2f 15.6d 13.3

TOG 14365 69.1f 73.4d 81.4c 15.4cd 12.3

O. sativa x O. glaberrima cross NERICA 4 70.3g 74.4e 82.3f 14.5bc 12.0

Asian rice Koshihikari 60.5b 66.9b 77.0b 14.3ab 16.5

WITA 4 59.8a 65.8a 75.0a 13.3a 15.2

To = Onset Temperature; Tp = Peak Temperature; Tc = Conclusion Temperature;

ΔH = Enthalpy of Gelatinization. MR= Melting Range, #Values with the same letters are not significantly different within columns

140

Table 5.4. Melting Parameters of Retrograded Starch Gels of African Rice Accessions and Asian Rice Varieties# MR RT Sample T (°C) T (°C) T (°C) ΔH (J/g) o p c (°C) (%) African rice IRGC 80787 44.1c 54.6e 71.3g 9.8bc 27.2 64.9

IRGC 86741 43.1ab 53.5bcd 71.1f 10.8c 28.0 71.5

IRGC 103759 44.0c 53.8cde 72.3h 10.8c 28.3 68.4

RAM 123 43.7bc 53.8cde 70.0d 9.2b 26.3 60.5

TOG 7682 43.8bc 53.2abc 70.1d 10.0bc 26.3 65.4

TOG 12440 42.8a 52.5a 70.0d 10.1bc 27.2 64.7

TOG 14365 43.1ab 52.9abc 70.4e 10.6c 27.3 68.8

O. sativa x O. glaberrima cross NERICA 4 43.1ab 54.2de 69.2c 9.2b 26.1 63.4

Asian rice Koshihikari 42.9a 52.8ab 65.5a 6.4a 22.6 44.8

WITA 4 42.9a 52.4ab 67.5b 7.3a 24.6 54.9

To = Onset Temperature; Tp = Peak Temperature; Tc = Conclusion Temperature; ΔH =

Enthalpy of Gelatinization MR= melting range, RT%=Percentage retrogradation after 14

days of storage at 4°C.

#Values with the same letters are not significantly different within columns

141

Table 5.5. Correlations between Amylose Concentration and Some Measured Parameters# Amylose TO TP TC ΔH Peak Final Breakdown Setback Total K/S@ CNS CIDS viscosity viscosity viscosity setback λmax550 Amylose

TO 0.7406

TP 0.7278 0.9953

TC 0.6960 0.9721 0.9824 ΔH 0.6429 0.7063 0.7232 0.7651 Peak viscosity -0.6547 NS NS NS NS Final viscosity 0.5130 0.6709 0.6812 0.6616 NS NS Breakdown viscosity NS NS NS NS NS 0.5343 NS Setback 0.8651 0.7040 0.6791 0.6360 0.5668 -0.6982 0.6547 -0.6951 Total setback 0.8998 0.7371 0.7385 0.6807 0.6800 -0.5962 0.6191 NS 0.8970 K/S@ λmax550 0.9389 0.7963 0.7765 0.7590 0.6891 -0.6239 0.5997 -0.5711 0.9041 0.8463 CNS NS NS NS NS 0.4726 NS NS NS NS NS NS NS

CIDS NS NS NS 0.4553 0.5069 NS NS NS NS NS NS NS

# Correlations significant at p<0.05

CNS = Crystallinity of Starch; CIDS= Crystallinity of Iodinated Starch; To = Onset Temperature; Tc = Conclusion Temperature;

Tp = Peak Temperature; ΔH = Enthalpy of Gelatinization

142

Figure 5.1. SEM Micrographs of Isolated Starch from African Rice Accessions and

Asian Rice Varieties. 143

14 IRGC 80787

IRGC 86741 12 IRGC 103759

10 RAM 123

TOG 7682

8 TOG 12440 Volume (%)

TOG 14365

6 NERICA 4

Koshihikari 4 WITA 4

0 0 5 10 15 20 25 30 35 Granule size (um)

Figure 5.2. Particle Size Distribution of Starches Isolated from African Rice Accessions and Asian Rice Varieties.

144

0.006

0.005 IRGC 80787

IRGC 86741

0.004 IRGC 103759

RAM 123

0.003 TOG 7682

TOG 12440 Intensity (cps)

0.002 TOG 14365

NERICA 4

0.001 KOSHIHIKARI

WITA 4

0 5 10 15 20 25 30 35 2θ (degree)

Figure 5.3. Wide Angle X-Ray Diffractograms of Starches Isolated from African Rice

Accessions and Asian Rice Varieties.

145

IRGC 80787 IRGC 86741 IRGC 103759 RAM 123 TOG 7682

CNS=31.3 CNS=32.1 CNS=31.2 CIS=35.4 CNS=30.5 CNS=32.5 CIS=36.7 CIS=36.1 CIS=32.9 CIS=32.2

TOG 12440 TOG 14365 NERICA 4 Koshihikari WITA 4 Intensity (cps)

CNS=30.7 CNS=31.0 CNS=29.5 CNS=31.6 CNS=29.5 CIS=35.6 CIS=35.1 CIS=31.2 CIS=34.5 CIS=31.2

Diffraction angle 2θ Figure 5.4. Wide Angle X-Ray Diffractograms of Starch and Iodinated Starches from African and Asian Rice.

(Intensity (cps) vrs. Diffraction Angle (2θ degree)). Black and blue lines represent starch and iodinated starches respectively; CNS = % crystallinity of starch;

CIS = % crystallinity of iodinated starch Arrows show the shift in peaks at 15°, 17° and 18° 2θ° observed in the iodinated ARAs.

146

35 IRGC 80787

30 IRGC 86741

IRGC 103759 25

RAM 123

20 TOG 7682

K/S valueK/S TOG 12440 15

TOG 14365 10 NERICA 4

5 KOSHIHIKARI

WITA 4 0 400 450 500 550 600 650 700 Wavelength (nm)

Figure 5.5. K/S Spectra of Starches Isolated from African Rice Accessions and Asian Rice. Varieties

147

350 100

IRGC 80787 300 90 IRGC 86741

250 IGRC 103759 80 RAM 123

200 ° C) TOG 7682 70 TOG 12440 (RVU) Viscosity 150 Temp. ( TOG 14365 60 100 NERICA 4

KOSHIHIKARI 50 50 WITA 4

0 40 Temp. 0 5 10 15 Time (min)

Figure 5.6. RVA Pasting Profiles of Starches Isolated from African Rice Accessions and Asian Rice Varieties.

148

To Tp Tc

WITA 4 ΔH= 13.3 J/ga

Koshihikari ΔH= 14.3 J/gab

bc NERICA 4 ΔH= 14.5 J/g

TOG 14365 ΔH= 15.4 J/gcd

TOG 12440 ΔH=15.6 J/gd

TOG 7682 ΔH= 15.3 J/gbcd

RAM 123 ΔH= 15.2 J/gbcd

IRGC 103759 ΔH=15.8 J/gd

IRGC 86741 ΔH= 15.1 J/gbcd

IRGC 80787 ΔH= 15.1 J/gbcd

55 60 65 70 75 80 85

Figure 5.7. Thermal Properties of Starches Isolated from African Rice Accessions and Asian

Rice Varieties.

149

To Tp Tc

WITA 4 ΔH= 7.3 J/g a

Koshihikari ΔH= 6.4 J/g a

NERICA 4 ΔH= 9.2 J/g b

TOG 14365 ΔH= 10.6 J/g c

TOG 12440 ΔH=10.1 J/g bc

TOG 7682 ΔH= 10.0 J/g bc

RAM 123 ΔH= 9.2 J/g b

c IRGC 103759 ΔH=10.8 J/g

IRGC 86741 ΔH= 10.8 J/g c

IRGC 80787 ΔH= 9.8 J/g bc

40 45 50 55 60 65 70 75 Temperature (°C)

Figure 5.8. Thermal Properties of Retrograded Starch Gels from African Rice Accessions and Asian Rice Varieties.

150

6 AMP LCAM SCAM

4 Carbohydrates (wt. %)

0 40 60 80 100 120 140 160 180

Fraction number

KOSHIHIKARI NERICA 4 WITA 4 IRGC 86741

TOG 7682 IRGC 103759 IRGC 80787 TOG 14365

TOG 12440 RAM 123

Figure 5.9. Gel-Permeation Chromatograms of Debranched Rice Starches

Fractionated on Sepharose CL-6B.

(LCAM = Long Chain Amylose SCAM = Short Chain Amylose AMP = Amylopectin Chains)

151

35 700 35 700 IRGC 80787 35 700 35 700 35 700 TOG7682 M IRGC 86741 M IRGC 103759 RAM 123 M M 30 650 M 30 30 30 30 650 650 650 650

25 25 25 600 25 25 600 600 600 600 20 20 20 20 20 550 550 550 550 550 15 15 15 15 15 500 500 500 500 500 10 10 10 10 10 450 450 450 450 450 5 5 5 5 5

0 400 0 400 0 400 0 400 0 400 15 25 35 45 55 65 75 15 25 35 45 55 65 75 15 25 35 45 55 65 75 15 25 35 45 55 65 75 15 25 35 45 55 65 75

35 700 35 700 35 700 35 700 TOG 12440 35 TOG 14365 700 NERICA 4 WITA 4 M Koshihikari M M M M 30 30 30 650 30 650 650 30 650 650

25 25 25 25 25 600 600 600 600 600 20 20 20 20 20 550 550 550 550 550 15 15 15 15 15

500 500 500 500 500 10 10 10 10 10

450 450 450 450 450 5 5 5 5 5

0 400 0 400 0 400 0 400 0 400 15 25 35 45 55 65 75 15 25 35 45 55 65 75 15 25 35 45 55 65 75 15 25 35 45 55 65 75 15 25 35 45 55 65 75

Figure 5.10. Gel-Permeation Chromatograms of Whole Rice Starches M=Maltose; Line graphs ( ------) represent carbohydrate and Their Corresponding β-LDs on Sepharose CL-2B. content (weight %) Symbols ( ● represent λmax values (nm)) Graph: Carbohydrate content (Weight %) on primary axis; λmax (nm) on secondary axis vrs. Fraction number. 152

Plot of Fitted Model Plot of Fitted Model

320 340

300 320 280 300 260 Final viscosityFinal Peak viscosity Peak 280 240

220 260 16 20 24 28 32 16 20 24 28 32 amylose amylose

Peak viscosity = 358.423 - 3.703*amylose Final vis. = 226.873 + 2.748*amylose r=-0.6547 r= 0.5130

Plot of Fitted Model Plot of Fitted Model

70 180

50 160 30 140 10

setback 120

-10 Total setback

-30 100

-50 80 16 20 24 28 32 16 20 24 28 32 amylose amylose

setback = -131.55 + 6.452*amylose Total setback = 3.98 + 5.067*amylose r=0.8651 r=0.8998

Figure 5.11. Correlation between Amylose Concentration and Pasting Parameters.

153

Plot of Fitted Model Plot of Fitted Model

71 75

69 73 67 71 65 69 peak temp onset temp 63

61 67

59 65 16 20 24 28 32 16 20 24 28 32 amylose amylose

Onset temp = 49.8519 + 0.657902*amylose Peak temp = 58.3673 + 0.517402*amylose

Plot of Fitted Model Plot of Fitted Model

84 16

82 15 80 14

78 enthalpy concl temp 13 76

74 12 16 20 24 28 32 16 20 24 28 32 amylose amylose

Concl. temp = 69.4606 + 0.43016*amylose Enthalpy = 11.6829 + 0.123362*amylose

Figure 5.12. Multiple Linear Regression Models to Describe the Relationship between Amylose

and Thermal Parameters of Isolated Starches.

154

5.8 References

Adebowale, K. O., Olu-Owolabi, I. O., Olawunmi, E. K., and Lawal, O. S. 2005.

Functional properties of native, physically and chemically modified breadfruit

(Artocorpus altilis) starch. Industrial Crops and Products 21: 343-351.

Annor, G. A., Marcone, M., Bertoft, E., and Seetharaman, K. 2014. Physical and

molecular characterization of millet starches. Cereal Chem. 91; 286-292.

Ahmed, J., Ramaswamy, H. S., Ayad, A., and Ali, I. 2008. Thermal and dynamic

rheology of insoluble starch from basmati rice. Food Hydrocoll. 22: 278-287.

Becker, A., Hill, S. E., and Mitchell, J. R. 2001. Milling-A further parameter affecting the

rapid visco analyser (RVA) profile. Cereal Chem. 78: 166-172.

Bertoft, E. 2004. On the nature of categories of chains in amylopectin and their connection

to the super helix model. Carbohydr. Polym. 57: 211-224.

Billmeyer, F.W., and Saltzman, M. 1981. The coloring of materials in industry. Principles

of color Technology, (Eds.), Wiley, New York: 133-172.

Chauhan, F., and Seetharaman, K. 2013. On the organization of chains in amylopectin.

Starch/Stärke. 65: 191-199.

Cheetham, N. W. H., and Tao, L. 1998. Variation in crystalline type with amylose content

in maize starch granules: An X-ray powder diffraction study. Carbohydr. Polym.

36: 277-284.

155

Chung, H.J., Liu, Q., Lee, L., and Wei, D. 2011. Relationship between the structure,

physicochemical properties and in vitro digestibility of rice starches with different

amylose contents. Food Hydrocoll. 25: 968-975.

Dhital, S.; Shrestha, A. K., and Gidley, M. J. 2010. Effect of cryo-milling on starches:

functionality and digestibility. Food Hydrocoll. 24: 152-163.

Fisher, D. K., and Thompson, D. B. 1997. Retrogradation of maize starch after thermal

treatment within and above gelatinization temperature range. Cereal Chem. 74:

344-351.

Futakuchi, K., and Sié, M. 2009. Better exploitation of African Rice (Oryza glaberrima

Steud.) in varietal development for resource-poor farmers in West and Central

Africa. Agric. J. 4: 96-102.

Gidley, M. J., Hanashiro, I., Hani, N. M., Hill, S. E., Huber, A., Jane, J.-L., Liu, Q., Morris,

G., Rolland-Sabaté, A., Striegel, A. M., and Gilbert, R. G. 2010. Reliable

measurements of the size distributions of starch molecules in solution: Current

dilemmas and recommendations. Carbohydr. Polym. 79: 255-261.

Hasjim, J., Srichuwong, S., Scott, M. P., Jane, J.-L. 2009. Kernel composition, starch

structure, and enzyme digestibility of opaque-2 maize and quality protein maize. J.

Agric. Food. Chem. 57: 2049-2055.

Hatcher, D. W., Anderson, M. J., Desjardins, R. G., Edwards, N. M., and Dexter, J. E.

2002. Effects of flour particle size and starch damage on processing and quality of

white salted noodles. Cereal Chem.79: 64-71.

156

He, G. Q., Song, X. Y., Ruan, H., and Chen, F. 2006. Octenyl succinic anhydride modified

early indica rice starches differing in amylose content. J. Agric. Food Chem. 54:

2775-2779.

Imberty, A., and Perez, S. 1988. A revisit to the three-dimensional structure of B-type

starch. Biopolym. 27: 1205-1221.

Kim, H.-S., and Huber, K. C. 2008. Channels within soft wheat starch A- and B-type

granules. J. Cereal Sci. 48: 159-172.

Lawal, O. S., Lapasin, R., Bellich, B., Olayiwola, T. O., Cesàro, A., and Yoshimura, M.

011. Rheology and functional properties of starches isolated from five improved

rice varieties from West Africa. Food Hydrocoll. 25: 1785-1792.

Li, J. Y., and Yeh, A. I. 2001. Relationships between thermal, rheological characteristics

and swelling power for various starches. J. Food Eng. 50: 141-148.

Mohan, B.H., Gopal, A., Malleshi, N.G., and Tharanathan, R.N. 2005. Characteristics of

native and enzymatically hydrolyzed ragi (Eleusine coracana) and rice (Oryza

sativa) starches. Carbohyd. Polym. 59: 43-50.

Morrison, W. R., Tester, R. F., and Gidley, M. J. 1994. Properties of damaged starch

granules II. Crystallinity, molecular order and gelatinisation of ball-milled starches.

J. Cereal Sci. 19: 209-217.

Nishita, K. D., and Bean, M. M. 1982. Grinding methods: Their impact on rice flour

properties. Cereal Chem. 59: 46-49.

157

Nowakowski, D., Sosulski, F. W., and Hoover, R. 1986. The effect of pin and attrition

milling on starch damage in hard wheat flours. Starch/Stärke 38: 253-258.

Peisong, H., Haijun, Z., Zhiying, D., Zhang, L., and Dianxing, W. 2004. Starch

digestibility and the estimated glycemic score of different types of rice differing in

amylose contents. J. Cereal Sci. 40: 231-237.

Raina, C. S., Singh, S., Bawa, A. S., and Saxena, D. C. 2007. A comparative study of

Indian rice starches using different modification model solutions. LWT-Food

Science and Technology. 40: 885-892.

Ratnayake, W.S., Hoover, R., Shahidi, F., Perera, C., and Jane, J. 2001. Composition,

molecular structure, and physicochemical properties of starches from four field pea

(Pisum sativum L.) cultivars. Food Chem. 74: 189-202.

Saibene, D., Zobel, H. F., Thompson, D. B., and Seetharaman, K. 2008. Iodine‐binding in

granular starch: Different effects of moisture content for corn and potato starch.

Starch/Stärke. 60: 165-173.

Sargeant, J. 1982. Determination of amylose: Amylopectin ratios of starches.

Starch/Stärke. 34: 89-92.

Schulman, A.H., and Kammiovirta, K. 1991. Purification of barley starch by protein

extraction. Starch/Stärke 43: 387-389.

Singh, N., Singh, J., Kaur, L., Sodhi, N. S., and Gill, B. S. 2003. Morphological, thermal

and rheological properties of starches from different botanical sources. Food

Chem. 81: 219-231. 158

Singh, N., Nakaura, Y., Inouchi, N., and Nishinari, K. 2007. Fine structure, thermal and

viscoelastic properties of starches separated from Indica rice cultivars.

Starch/Stärke 59: 10-20.

Sweeney, M., and McCouch, S. 2007. The complex history of the domestication of rice.

Ann. Bot. 100: 951-957.

Tester, R. F. 1997. Properties of damaged starch granules: Composition and swelling

properties of maize, rice, pea and potato starch fractions in water at various

temperatures. Food Hydrocoll. 11: 293-301.

Vamadevan, V., Bertoft, E., and Seetharaman, K. 2013. On the importance of organization

of glucan chains on thermal properties of starch. Carbohydr. Polym. 92: 1653-

1659.

Vamadevan, V., and Bertoft, E. 2015. Structure-function relationships of starch

components. Starch/Stärke. 67: 55-68.

Vandeputte, G. E.,Vermeylen R., Geeroms, J., and Delcour, J. A. 2003. Rice starches I.

Structural aspects provide insight into crystallinity characteristics and gelatinization

behavior of granular starch. J. Cereal Sci. 38: 43-52.

Vandeputte, G. E., and Delcour, J. A. 2004. From sucrose to starch granule to physical

behaviour: a focus on rice starch. Carbohydr. Polym. 58: 245-266.

Waduge, R., Xu, S., and Seetharaman, K. 2010. Iodine absorption properties and its effect

on the crystallinity of developing wheat starch granules. Carbohydr. Polym. 82:

786-794. 159

Wang, L. F., and Wang, Y. J. 2004. Rice starch isolation by neutral protease and high

intensity ultrasound. J. Cereal Sci. 39: 291-296.

Wang, L., Xie, B., Shi, J., Xue, S., Deng, Q., Wei, Y., and Tian, B. 2010. Physicochemical

properties and structure of starches from Chinese rice cultivars. Food Hydrocoll.

24: 208-216.

Wang, L., Xie, B., Xiong, G., Du, X., Qiao, Y., and Liao, L. 2012. Study on the granular

characteristics of starches separated from Chinese rice cultivars. Carbohydr. Polym.

87: 1038-1044.

Wani, A. A., Singh, P., Shah, M. A., Schweiggert-Weisz, U., Gul, K., and Wani, I. A.

2012. Rice starch Diversity: Effects on structural, morphological, thermal, and

physicochemical properties-A review. Comprehensive Reviews in Food Science

and Food Safety. 11: 417-436.

Wickramasinghe, H. A. M, and Noda, T. 2008. Physicochemical properties of starches

from Sri Lankan rice varieties. Food Sci Technol Res 14: 49-54.

Yamin, F.F., Lee, M., Pollak, L.M., and White, P.J. 1999. Thermal properties of starch in

corn variants isolated after chemical mutagenesis of inbred lines B73. Cereal

Chem. 76:175-181.

Yu, S., Ma, Y., Menager, L., and Sun, D. 2012. Physicochemical properties of starch and

flour from different rice cultivars. Food Bioprocess Technol. 5: 626-637.

160

Zaidul, I. S. M., Norulaini, N. A. N., Omar, A. K. M., Yamauchi, H., and Noda, T. 2007.

RVA analysis of mixtures of wheat flour and potato, sweet potato, yam, and

cassava starches. Carbohydr. Polym. 69:784-791.

Zhong, F., Li, Y., Ibáñez, A. M., Oh, M. H., McKenzie, K. S., and Shoemaker, C. 2009.

The effect of rice variety and starch isolation method on the pasting and rheological

properties of rice starch pastes. Food Hydrocoll. 23: 406-414.

161

CHAPTER 6: UNIT AND INTERNAL CHAIN PROFILES OF AFRICAN RICE

(Oryza glaberrima) AMYLOPECTIN

162

Abstract

African rice and Asian rice are two distinct types of domesticated rice. Little is known about structure of starch in African rice and its impact on rice quality. This work investigated the fine structure of the starches isolated from African rice in comparison to

Asian rice (globally consumed rice). The unit chain profiles of amylopectins and their

ϕ, β-limit dextrins from two African rice accessions (ARAs), namely TOG 12440 and

IRGC 103759, were compared with two Asian rice samples, namely cv Koshihikari

(spp. japonica) and cv WITA 4 (spp. indica) and NERICA 4 (an inter specific cross between sativa x glaberrima). The debranched starch samples were studied by high- performance anion-exchange chromatography. The average chain lengths (CL), internal chain length (ICL), external chain length (ECL), and total internal chain length (TICL) of the amylopectins showed narrow ranges among all the rice investigated. The short:long

(S:L) chain ratio ranged between 12.1-13.8, and the relative molar concentration of B- chains was approximately 50% in all samples, and thus the ratio of A:B-chains was 1.0.

Significant differences were, however, observed in the proportions of very short

"fingerprint" B-chains (Bfp, DP 3-7). The ARAs and NERICA 4 had higher levels of

Bfp-chains than the Asian rice but the major group of short B-chains (DP 8-25) was similar in all samples. Positive correlations (p<0.05) were observed between ECL and gelatinization transition temperatures and enthalpy. The study results strongly suggest that the structure of amylopectin of African rice is different from that of Asian rice and this may account for the higher gelatinization enthalpies observed in the ARAs. The differences observed also suggest possible dissimilarities in the fine structure (clusters and building blocks) of their amylopectins.

Key words: African rice, amylopectin, φ,β-limit dextrins, unit chain, internal chain

163

Introduction

Oryza sativa (Asian rice) and Oryza glaberrima (African rice) are the two distinct types of domesticated rice with Asian rice having a global distribution particularly in Asia. African rice is cultivated mainly in the West Africa sub region. The predominant component of rice is starch (90% dry basis) and its composition (including apparent amylose content and amylopectin structure) and physicochemical properties (i.e. gelatinization temperature and pasting viscosity) determine the eating, cooking and milling qualities of rice (Juliano 1998;

Bao 2012).

Starch contains three types of bio-macromolecules: amylopectin (highly branched molecule), amylose (almost linear but may have few branches molecule), and possibly an intermediate molecule, which is found in some kinds of starch as in oat and some mutant maize genotypes exhibiting a structure intermediate to those of amylose and amylopectin

(Wang and White 1994; Wang et al. 1993). Variations exist in the contents of these three components in starches from different sources, but amylopectin is commonly considered as the major component in storage starch and accounts for up to 85% by weight (Nakamura

2002).

Amylose is a linear polymer that is made up of almost entirely α-1,4-linked

D-glucopyranosyl units, however, several amylose molecules have between 0.3% and 0.5%

α-1,6-linked D-glucopyranose branches (Whistler and BeMiller 1997). Amyloses from rice starch have been reported to have average degree of polymerization (DP) = 920 to

1110 and they are slightly branched with 2 to 5 chains on the average (Takeda et al. 1986).

Amylopectin unit chains are broadly divided into A-, B- and C-chains. A-chains which do not carry other chains are attached through their reducing end to B-chains, while B-chains 164

are linked in the same way and carry one or more A-chains. Additionally, each molecule contains a single C-chain that contains the only reducing group of the amylopectin molecule and carries other chains (Peat et al. 1952).

The building block backbone model (Bertoft 2013) and the cluster model proposed by

(French 1972; Nikuni 1978) and modified by Hizukuri (1986) are the two key models of amylopectin structure that are discussed in recent times. In the two-directional backbone model, the short chains are anchored perpendicular to the long chains which are mainly present in the amorphous region and do not participate in the formation of the crystalline lamellae whereas in the cluster model, the clusters are interconnected by long chains which participate in the formation of the crystalline and amorphous lamellae. The amylopectin molecule consists of three types of unit chains, denoted as A-, B- and C-chains. Hizukuri

(1986) suggested that the A-chains are the shortest chains, B1-chains with a DP of 20-24 are short while B2-chains (DP 42-48) and B3-chains (DP69-75) are much longer. Exo- acting enzymes such as phosphorylase and β-amylase have been used to isolate the internal part of amylopectin through the removal of external chains. β-amylase produces maltose by acting on glucan chains from their non-reducing ends. Due to the inability of β-amylase to by-pass a branch point in the amylopectin molecule, the resulting β-limit dextrin with the branches denotes the internal part of the amylopectin molecule. Phosphorylase a from rabbit muscle produces φ-limit dextrins (Manners 1989). Practically, isolation of the internal part of amylopectin is done by treating amylopectin successively with both enzymes to generate φ,β-limit dextrins (φ,β-LD), in which the A-chains are reduced into maltosyl stubs, and the external segments of the B-chains into a single residue.

Debranching of the φ,β-LD reveals the internal chains, which can be analysed by

165

techniques such as anion-exchange chromatography (Bertoft 2004). Zhu and Bertoft

(1996) reported the use of limit dextrins in studying the internal structure and ratio of A:B chains in amylopectin molecules. After analyzing the internal unit chain distribution in amylopectins from different botanical sources, Bertoft et al. (2008) categorized the internal chains (which all are B-chains), into short (BS) and long chains (BL). The BS-chains were further divided into very short “fingerprint” Bfp-chains with a degree of polymerisation

(DP) 3-7 and BSmajor, with DP 8-25. On the other hand, BL-chains were divided into

B2-chains (DP 26–50) and B∼3-chains (DP > 50).

∼ Bertoft et al. (2008) and Manners (1989) classified amylopectin into external and internal chains where the external chains are found in the crystalline part of the amylopectin molecule and the internal chains are mainly present in the amorphous lamellae. The entire

A-chains are external since they do not carry other chains while the B-chains are involved in both external and internal divisions. From the internal chain profile the total internal chain length (TICL), which is defined as the average length of the B-chains without their external chains, can be estimated (Bertoft 1991). The length of the segments of the chains that extend from the outermost branch to the non-reducing end of the chains (ECL) and the average internal chain length (ICL), defined as the length of the chain segment between two branches in the amylopectin molecule, can be estimated from a comparison of the average chain length (CL) of the amylopectin and CL of the limit dextrin (Manners 1989).

The internal structure of amylopectins from diverse botanical sources have been reported over the past years (Bertoft 2013; Zhu et al. 2011; Kong et al. 2009; Laohaphatanaleart et al. 2009; Bertoft et al. 2008; Klucinec and Thompson 2002 and Bertoft and Koch 2000).

166

Bertoft et al. (2008) on the basis of internal chain profile categorized amylopectins into four structural types, of which medium amylose-containing Asian rice was found to be type

2. Since the basic structure of African rice starches have not been characterized yet, this study focuses on investigating the internal structure of two accessions of African rice amylopectins in the form of φ,β-limit dextrins. They were compared with two Asian rice varieties and an interspecific cross between African rice and Asian rice.

6.2 Materials and Methods

6.2.1 Materials

Five dehusked and polished rice samples were provided by the Africa Rice Center based in

Cotonou, Benin for this study. Two of the samples were African rice accessions (ARAs) -

IRGC 103759 and TOG 12440. The others are two O. sativa varieties – cv Koshihikari and cv. WITA 4 and NERICA 4 (an interspecific cross between African rice x Asian rice). All chemicals and solvents used were of ACS certified grade.

6.2.2 Methods

6.2.3 Starch Isolation

Starch was isolated from the rice samples using the method of Waduge et al. (2010) with some modifications as previously described in chapter 4.

6.2.4 Isolation of Amylopectin from Rice Starch

The method described by Klucinec and Thompson (1998) was employed with some modifications to isolate amylopectin from the rice starches. Forty mL of 90% dimethyl

167

sulfoxide (DMSO) was used to disperse 2 g of granular rice starch in a water bath (90 °C) for 3 h under nitrogen atmosphere. During this time the mixture was constantly stirred with a magnetic stirrer at low speed. The dissolved starch was gradually cooled to 80 °C after which 4 portions of ethanol (98%) (160 mL) were added drop-wise with the aid of a peristaltic pump (Easy-Load®Masterflex® Model 7518-12) to precipitate the starch out of solution. The starch-ethanol mixture was slowly cooled down to room temperature (25 °C) and then centrifuged at 8000 x g for 15 min at 4 °C. The supernatant was discarded and the precipitate (non-granular starch) was washed with ethanol. The non-granular starch was again dissolved in DMSO (90%) (56 mL) in a water bath at 90 °C while was constantly stirred for 1h under nitrogen atmosphere. The temperature was cooled to 80 °C and a vigorously stirred mixture of 1-butanol (23.5 mL) and isoamyl alcohol (23.5 mL) in 324 mL of water was added drop-wise using a peristaltic pump for a period of 1-1.5 h under nitrogen. Thereafter the mixture was continuously stirred for an additional 30 min under nitrogen. The mixture was allowed to cool gradually overnight to room temperature

(25 °C) in an insulated foam box. The mixture was gently agitated and centrifuged at

10,000 x g for 30 min at 4 °C after which the supernatant (which is the amylopectin portion) was cautiously decanted and concentrated to 50 mL in a rotary evaporator at

60 °C. A mixture of 1-butanol-isoamyl alcohol-water (1:14) was added to the concentrated supernatant to precipitate any residual amylose after which the centrifugation step was repeated. The supernatant (amylopectin) was decanted and concentrated to 50 mL in a rotary evaporator at 60 °C. Amylopectin was precipitated out of solution with 3 portions of methanol (150 mL) overnight at room temperature after which centrifugation at 22, 360 x g for 20 min at 4 °C was followed. The precipitate was collected and dissolved in warm

168

water (20 mL) and 3 portions of ethanol (98%) (60 mL) was used to re-precipitate amylopectin for 1 h at room temperature. The precipitated sample (amylopectin) was finally centrifuged at 22, 360 x g for 20 min at 4 °C. The precipitate was collected, dissolved in warm water (20 mL) and freeze-dried.

6.2.5 Determination of Purity of Amylopectin by Gel Permeation Chromatography

The purity of the isolated amylopectin fractions isolated from rice starch was determined by analyzing debranched samples by gel-permeation chromatography according to the procedure reported by (Laohaphatanaleart et al. 2009) on a column (1.6 cm x 90 cm) of

Sepharose CL-6B (Pharmacia, Uppsala, Sweden) eluted with 0.5 M aqueous NaOH at

0.5 mL/min. Purified amylopectin (2.0 mg) was dissolved in 90% DMSO (50 μL) by employing heat for about 15 min in boiling water bath with constant magnetic stirring. The amylopectin solution was then diluted with 400 μL warm water followed by the addition of

50 μL sodium acetate buffer (pH 5.5, 0.01 M). The solution was allowed to cool to room temperature after which 1 μL each of isoamylase and pullulanase were added to initiate debranching of the amylopectin. The debranching reaction lasted for 12 h with constant stirring at room temperature. Thereafter, 50 μL NaOH (5 M) was added to halt the reaction by inactivating the debranching enzymes. Aliquots (200 μL) were injected onto a column

(1.6 cm × 90 cm) of Sepharose CL-6B (Pharmacia, Uppsala, Sweden). The eluent used was sodium hydroxide (0.5 M) with a pumping speed of 0.5 mL/min. Fractions (0.5 mL) were collected analyzed for carbohydrate content by the phenol-sulphuric acid method

(Dubois et al. 1956). The column was calibrated with both linear and branched dextrins of known DP (Bertoft and Spoof 1989).

169

Production of φ,β-Limit Dextrin of Amylopectin

The procedure reported by Bertoft (2004) with modifications by Kalinga et al. (2013) was employed for the production of φ,β-limit dextrins (φ,β-LDs) from amylopectin.

Amylopectin (100 mg) was weighed and dissolved in 90 % DMSO (3 mL) in a beaker on a water bath for 10 min at room temperature and constantly stirred gently with a magnetic stirrer overnight. The dissolved amylopectin was then diluted with of warm water (~80°C,

32.7 mL), phosphate buffer-pH 6.8 (3.6 mL) and EDTA solution (1.7 mL). Freshly prepared solution of phosphorylase a from rabbit liver (0.9 U/mL - Sigma)

(9 mL) was added and incubated overnight at room temperature bringing the concentration of carbohydrate in solution to 2 mg/mL. The enzyme was inactivated by heating in boiling water bath for 5 min and thereafter, glucose 1-phosphate (G-1-P) and maltose formed in the reaction were removed by tangential flow filtration (Minimate™ TFF System, Canton,

MA), using a filter with a molecular weight cut-off limit of 10,000 Da. After the filtration process, phosphate buffer, EDTA solution and phosphorylase a were added to the filtrate and incubated overnight. Additional G-1-P and maltose generated were filtered out with the tangential flow filtration system after inactivating enzymes. β-Amylase from barley

(Megazyme 28, 400 U/mL) (2 µL) and phosphate buffer-pH 6.8 (6.6 mL) were added and incubated overnight. G-1-P and maltose generated were filtered out with the tangential flow filtration system after inactivating enzymes. The addition of β-amylase was repeated and the resulting φ,β-LDs were recovered and freeze-dried.

170

6.2.6 Analysis of Unit Chain Distribution

Two mg of amylopectin or φ,β-limit dextrin was dissolved in 90% DMSO (50 μL) with gentle and continuous stirring overnight. The solution was diluted with warm water (400

μL) (80 °C) after which 0.01 M sodium acetate buffer (50 μL) (pH 5.5) was added.

Pullulanase M1 (1 μL) and isoamylase (1 μL) (Megazyme International, Ireland) were then added to the mixture and stirred overnight at room temperature to debranch the structure of the amylopectin or φ,β-limit dextrin. The enzymes were inactivated by boiling for 5 min and the volume of the mixture was adjusted to bring the carbohydrate concentration to 1 mg/mL. The sample was further passed through a 0.45 μm nylon filter after which an aliquot (25 μL) was injected into the Dionex ICS 3000 high-performance anion-exchange chromatography (HPAEC) system (Dionex Corporation, Sunnyvale, CA, USA) equipped with a pulsed amperometric detector (PAD), CarboPac PA-100 ion-exchange column

(4 x 250 mm) and a similar guard column (4 x 50 mm). The samples were then eluted with a flow rate of 1 mL/min. The eluents used were A (150 mM sodium hydroxide) and

B (150 mM sodium hydroxide containing 500 mM sodium acetate). An elution gradient

(130 min) was employed by mixing eluent B into eluent A as follows: 0-9 min, 15-36%

B; 9-18 min, 36-45% B; 18-110 min, 45-100% B; 100-112 min, 100-15% B; 112-130 min,

15% B was used. The column was equilibrated with 15% eluent B for 60 min between runs. The areas under the chromatograms were converted to carbohydrate concentration using the method of Koch et al. (1998).

171

6.2.7 Statistical Analysis

The statistical programme Statgraphics Centurion XV, Version 15.1.02 (StatPoint,

Warrenton, VA, USA) was employed to conduct statistical analysis. One-way analysis of variance was performed to determine significant differences between any pair of means at the 95% confidence level and multiple range tests were done to determine which means were different.

6.3 Results and Discussion

6.3.1 Purity of Amylopectin

An example of the comparison of a purified amylopectin sample with its corresponding whole starch is shown in Fig. 6.1. In the chromatograms, the short chains of amylopectin eluted after the long chains of amylose from the whole starch and debranched amylopectin samples. Fractions 46-104 were considered to be the contribution of residual amylose while fractions 105-168 were attributed to the debranched amylopectin chains. The debranched amylopectin exhibited a major peak at fractions 134-138. In this study, the purity of the amylopectins was >95% for all samples. The shortfall of 100% purity could be attributed to the presence of trace amounts of amylose and/or long chains of amylopectin designated as “super long amylopectin chains”, which are reported to exist in some amylopectins, particularly in some rice starches (Koroteeva et al. 2007; Takeda et al.

1987). Laohaphatanaleart et al. (2009) reported that the super long chains of rice amylopectin constitute 3.4-8.6% of rice starches.

172

6.3.2 Unit Chain Length Distributions of Amylopectins

The profiles of the unit chain distribution of the debranched amylopectins of the rice samples analyzed with HPAEC are presented in Figure 6.2. Two sets of profiles were found for the samples. The first profile, which corresponded to the two ARAs and

NERICA 4, had chain distributions showing peaks at DP 12 and DP 13, respectively. The second profile, which corresponded to the Asian rice samples, had quite distinct patterns.

Both showed highest peaks at DP 12, just as NERICA 4 and the ARAs, but a peak at DP 14 was also present. In addition, they possessed a shoulder, which occurred at DP 19. This was consistent with descriptions of patterns for waxy and non-waxy rice amylopectins reported by Hanashiro et al. (1996) and Laohaphatanaleart et al. (2009). The short and long chains of the distribution could be divided at DP 35 for all samples except NERICA 4, for which it occurred at DP 34. Additionally, the peak DP of the long chains for

Koshihikari, WITA 4 and IRGC 103759 occurred at DP 43, while that of NERICA 4 and

TOG 12440 occurred at 42 and 44, respectively.

6.3.3 Unit Chain Length Distributions of φ,β-Limit Dextrins

Bertoft (1989) reported that the external chains in amylopectin are hydrolyzed consecutively by phosphorylase a and β-amylase to produce their corresponding φ,β-limit dextrins, and as a result of the debranching action of the enzymes, the A-chains appear as a maltose peak and the rest as B-chains if a chromatogram of the chain distribution is drawn.

Figure 6.3 shows the chromatograms of the B-chains and thus reflects the composition of the internal chains of amylopectin. In general, all internal unit chain profiles possessed the two major groups of short and long chains as found in the whole amylopectins. However,

173

while the short chains (S-chains) of amylopectin are mixtures of short A- and short

B-chains (BS- or B1-chains), the S-chains of the φ,β-limit dextrins represent only

BS-chains, (Bertoft et al. 2008). The A-chains in the form of maltose have been removed by filtration during the analysis and thus it did not appear in the chromatograms in Figure

6.3. At the lowest position in the chromatograms (weight-based) a division between the short (BS) and long chains (BL) was made at DP 22 except for Koshihikari which was at

DP 24. This was comparable to ~DP 21 that was reported for rice (Bertoft et al. 2008).

The peak-DP of the various B-chain categories was the same for all samples. All the rice amylopectin samples possessed a clearly distinguished group of Bfp-chains (DP 3-7) with a peak position at DP 5. It was also evident that the ARAs and NERICA 4 had the same peak-DP (31) of the long B-chain categories. However, the Asian rice samples had slightly different peak-DPs and it was higher (32) in WITA 4 but lower in Koshihikari (29).

The average chain length classifications of different chain categories in φ,β-limit dextrins and their parent amylopectins as obtained by HPAEC analysis are presented in Table 6.1.

The φ,β-limit value, average chain length (CL) and external chain length (ECL) of the original amylopectins were estimated (Bertoft 2004; Charoenkul et al. 2006). The average chain length (CL) of the rice amylopectin samples ranged from 16.87 to 18.27 glucose units. (Chain length is measured in glucose residues, albeit it is most often not written out).

This was predictable since cereal starches with type A X-ray diffraction pattern are known to have amylopectins with shorter CL in relation to amylopectins from roots and tubers, which have type B X-ray diffraction pattern and longer chains (Hizukuri 1985). The chain length of the two ARAs and NERICA 4 were significantly (p<0.05) longer than the chain lengths of the two Asian rice samples.

174

The short chains (SCL) of the ARAs and NERICA 4 were also significantly (p <0.05) longer than those found in the Asian rice samples, however, the length of long chains

(LCL) in both types of rice were not significantly different. The SCL and LCL ranged from 14.7 to 15.6 and 48.6 to 50.5, respectively. The φ,β-limit value of the rice amylopectins, which generally reflects the length of the external chains, ranged from

54.7% (Koshihikari) to 60.7% for IGCR 103759. This range is consistent with that reported for rice amylopectin (Laohaphatanaleart et al. 2009) and is also comparable to ranges reported for other plants (Bertoft et al. 2008). The average external chain length

ECL of the rice amylopectins ranged from 10.7-12.6 and was shorter in the Asian rice than in NERICA 4 and the ARAs. As external chain segments participate in the crystalline lamellae of the starch granules, the ARAs by virtue of their relatively longer external chain length may have a thicker lamella than the rest of the samples (Koroteeva et al. 2007). The

ECLs of the samples in this study were comparable to values reported for other cereals

(A-type starches) (Zhu et al. 2011; Bertoft et al. 2008).

The average internal chain length (ICL) of the samples were not significantly different

(p <0.05), albeit the ARAs and NERICA 4 tended to have shorter ICL suggesting that they have a more tightly branched structure in their amorphous lamellae. The total internal chain length (TICL) of the amylopectins (defined as the entire internal length of the B- chains excluding the external segment and corresponds in φ,β-LDs to the B-chain length minus one residue) ranged from 11.38-12.43 and was comparable to the range of 12.4-13.8 for three Asian rice varieties reported by Laohaphatanaleart et al. (2009). There were no significant differences among the samples with respect to CLLD defined as the average chain length of whole limit dextrin (including A-chains) and the CL of long B-chains

175

(BL-CLLD). However, the CL of short B-chains (BS-CLLD) of the ARAs were shorter than the Asian rice samples.

The distribution profiles by weight of the B-chains and their division into subgroups are shown in Figure 6.3 and the relative molar amounts (%) of various chain categories of the rice amylopectins in Table 6.2. The molar amount of A-chains ranged from 49.8 to 50.6% and hence the samples were not significantly different from each other. The samples therefore contained almost equal-molar amounts of A- and B-chains (Table 6.2).

Koizumi et al. (1991) reported that Afp chains of amylopectins with DP 6-8 have profiles that are characteristic to the source of plant and Bertoft et al. (2012) found that these chains are part of clusters. Nevertheless, the Afp-chains are considered to be too short to form double helices and for that matter do not effectively play a part in the crystalline lamellae of the starch granules (Gidley and Bulpin, 1987). Consequently, the remainder of the

A-chains are expected to be involved in the crystalline lamellae. Therefore, these chains are labelled as "crystalline A-chains" (Acrystal). The Afp chains in the two ARAs and

NERICA 4 (7.16%-7.18%) were somewhat close, but statistically different (p<0.05), and lower than that observed in the Asian rice (7.33-7.40%). However, all the samples showed similar amounts of Acrystal as values were not significantly different from each other. The molar amounts of Acrystal-chains reported in this study were lower than values reported by

Bertoft et al. (2008) for rice amylopectin, whereas values for Afp chains were higher.

Within the B-chain categories, the number of Bfp-chains tended to be elevated in ARAs, which was apparently reflected in the shorter ICL of these samples. Bfp-chains have been suggested to participate in tightly branched building blocks in clusters (Bertoft and Koch

176

2000). Long B2-chains tended to be somewhat more abundant in the African rice starches.

Otherwise, similar molar amounts of the rest of the B-chain categories (BSmajor-chains and

B3-chains) were found for all the samples and were comparable to molar amounts of the said chains reported for rice (Bertoft et al. 2008). Molar ratios of some particular different chain groupings of the rice amylopectins and their φ,β-limit dextrins are presented in Table

6.3. These ratios, in combination with the relative molar amounts of the different chain categories (Table 6.2), can be used to categorize the samples on the basis of their internal unit chain profiles into one of the four structural types of amylopectin suggested by Bertoft et al. (2008). Going by this classification, the range of values for the molar amounts of

BS-chains, Bfp-chains, BSmajor-chains, B2-chains and B3-chains of the samples place them under the category of type 2 amylopectins. Similarly the molar ratios of BS:BL, Acrystal:BS,

Acrystal:B, Bfp:BSmajor and S:L fit the samples into type 2. Thus, the amylopectin of African rice belongs to the same structural type as amylopectin in Asian varieties

(Bertoft et al. 2008; Laohaphatanaleart et al. 2009).

6.4 Correlations Between Chain Length Classifications and Some Functional

Properties

Table 6.4 represents the correlation between the chain categories of the rice amylopectin and some functional properties of the rice starch samples presented in Chapter 5. With regards to thermal properties, gelatinization transition temperatures (To, Tp, Tc) and enthalpy (ΔH) significantly and positively correlated with CL, SCL and ECL but negatively with BS-CLLD. On the other hand, correlations between To, Tp, Tc and ΔH and

LCL, ICL, TICL, CLLD and BL-CLLD were not significant (p<0.05). The external chains

177

participate in the formation of double helices and ΔH mainly reflects the unwinding of double helices during gelatinization (Cooke and Gidley 1992). Since ECL contributes to the amount of SCL and CL, it was observed that SCL and CL also positively correlated with ΔH. Longer external chains can also increase the crystalline stability by forming additional hydrogen bonds between glucan strands and/or double helices (Vamadevan et al.

2013). Since melting temperature (Tp) reflects the crystalline stability (Tester and

Morrison 1990) it was also observed that there was a positive correlation between ECL and melting transition temperatures (To, Tp, Tc). With regards to the pasting properties, final viscosity correlated negatively with ICL, TICL, CLLD and BS-CLLD but positively with CL,

SCL and ECL. Correlation with LCL and BL-CLLD were not significant. Similarly total setback viscosity correlated negatively with BS-CLLD but positively with CL, SCL and

ECL. Correlations with ICL, TICL, CLLD and BL-CLLD were not significant. It has been reported that starches with longer amylopectin chains retrograde faster than those with short amylopectin chains (Silverio et al. 2000; Fredriksson et al. 1998; Kalichevsky et al.

1990). Generally, the external chains of amylopectin can re-associate faster than internal chains and this recrystallization process can increase the final viscosity and setback values hence the observed positive correlation. Figure 6.4 shows the multiple linear regression models to describe the relationship between ECL, enthalpy, final and setback viscosities.

The equation of the fitted models are; Enthalpy = 2.70171 + 1.02258*ECL,

Final Viscosity = 38.9856 + 21.3679*ECL and Total setback = -281.426 + 35.1199*ECL

178

6.5 Conclusions

The structure of the amylopectins studied falls into the type 2 classification of amylopectins proposed by Bertoft et al. (2008). The profiles of the unit chain distribution of the debranched amylopectins of the ARAs and NERICA 4 were distinct from the Asian rice. The ECL of the ARAs and NERICA 4 were longer than the Asian rice varieties.

Since six glucose units are needed per turn of double helix the data suggest that the ARAs and NERICA 4 can possibly have two complete turns, while the Asian rice varieties can have less than two. This may explain the higher gelatinization enthalpies observed in the

ARAs. Additionally the observations suggest that more consideration should be given to the ECL chains categories of amylopectin to throw more light on the functional properties of starches. The observed correlations strongly suggest that the short chains of the amylopectins, especially their external segments were important in influencing the thermal and pasting properties of the starch samples. The differences observed in the amylopectin structures of African and Asian rice in this study suggest possible differences in the fine structure of the clusters and building blocks of their amylopectins.

6.6 Acknowledgements

This research work was undertaken with the kind assistance of the Global Rice Science

Partnership (GRiSP) and the Africa Rice Center (AfricaRice) in terms of financial support and supply of samples.

179

6.7 Tables and Figures Table 6.1. Chain Length Classifications and φ, β-Limit Values of Rice Amylopectin as obtained by HPAEC Analyses# φ, β-Limit Sample CL SCL LCL ECL ICL TICL value CLLD BS-CLLD BL-CLLD African rice IRGC 103759 18.27c 15.64b 50.36a 12.58b 4.69a 11.38a 60.65c 7.19a 8.41a 39.21a

TOG 12440 18.11c 15.45b 49.14a 12.14b 4.97a 11.90a 58.74bc 7.47a 8.76ab 39.07a

O. sativa x O. glaberrima cross NERICA 4 17.23c 15.47b 50.50a 12.00b 4.94a 11.90a 58.51abc 7.44a 8.69ab 38.75a

Asian rice KOSHIHIKARI 16.87a 14.76a 50.26a 10.73a 5.15a 12.43a 54.68a 7.65a 9.26b 39.26a

WITA 4 17.23b 14.71a 48.64a 11.08a 5.15a 12.29a 55.61ab 7.65a 9.03ab 39.53a

#Within each column values with the same letters are not significantly (p < 0.05) different

CL = Average Chain Length of Amylopectin SCL = CL of Short Chains LCL = CL of Long Chains ECL (External Chain Length) = CL x (φ, β-Limit Value/100) + 1.5 ICL (Internal Chain Length) = CL-ECL – 1

TICL (Total Internal Cain Length) = B-CLLD-1 φ, β-Limit Value = Computed from the Difference between CL of Amylopectin and its φ,β-LD CLLD = Average Chain Length of φ, β-Limit Dextrin

BS-CLLD = CL of Short B-chains BL-CLLD = CL of Long B-chains

180

Table 6.2. Relative Molar Amounts (%) of Chain Groupings in Amylopectins Isolated from Rice Starches#

1 2 3 4 5 6 7 8 Sample A-chains Afp Acrystal B-chains BS BL Bfp BSmajor B2 B3

African rice IRGC 103759 49.99a 7.16a 42.84a 50.01a 42.99a 7.03a 22.69b 19.46a 6.27ab 0.65a

TOG 12440 49.80a 7.17ab 42.63a 50.21a 41.80a 8.41b 22.14ab 21.22abc 6.18ab 0.68a

O. sativa x O. glaberrima cross NERICA 4 50.04a 7.18b 42.86a 49.97a 41.42a 8.55b 22.45b 20.11ab 6.91b 0.67a

Asian rice KOSHIHIKARI 50.61a 7.33c 43.28a 49.34a 41.71a 7.69ab 19.73a 23.41c 5.59a 0.67a

WITA 4 49.96a 7.40d 42.56a 50.05a 41.53a 8.52b 20.69ab 22.61bc 6.01a 0.71a

# Within each column values with the same letters are not significantly different (p < 0.05)

1Detected as Maltose after Debranching of φ, β-LD 2 “Fingerprint” A-chains at DP 6-8 in the Original Amylopectin Sample 3 Acrystal -Chains Calculated as all A-chains Less Afp 4B-Chains Correspond to DP >3 in φ, β-LDs, and were Divided into Short (BS) and Long (BL) Chains Chains (BL) at between DP 22–25 and above Respectively Depending on the Sample 5“Fingerprint” B-Chains at DP 3–7 in φ, β-LDs

6The Major Group of Short B-Chains at DP 8 to 22 (this may stretch to DP 24 depending on the sample in question) 7 Long Chains between DP 22 or 25 (depending on the sample) and 55 8 Long Chains at DP >56

181

Table 6.3. Molar Ratios of Different Chain Groupings of Rice Amylopectins and Their φ,β-Limit Dextrins#

Sample A:B S:L BS:BL Acrystal:BS Acrystal:B Bfp:BSmajor

African rice IRGC 103759 1.0a 11.6a 6.1b 1.0a 0.9a 1.1c

TOG 12440 1.0a 11.7a 5.0a 1.0a 0.9a 1.0abc

O. sativa x O. glaberrima cross NERICA 4 1.0a 11.8a 4.9a 1.0a 0.9a 1.1bc

Asian rice KOSHIHIKARI 1.0a 14.2b 5.4ab 1.0a 0.9a 0.8a

WITA 4 1.0a 12.4a 4.9a 1.0a 0.9a 0.9ab

#Values with the same letters are not significantly different (p < 0.05) Within Each Column S (Short Amylopectin Chains) L (Long Amylopectin Chains) The Other Abbreviations Have Been Explained in Table 6.2

182

Table 6.4. Correlation between Chain Groupings of Rice Amylopectins and Some Function Properties of Starch#

CL SCL LCL ECL ICL TICL CLLD BS-CLLD BL-CLLD

To 0.8998 0.8967 NS 0.8548 NS NS NS -0.7038 NS

Tp 0.9002 0.9033 NS 0.8653 NS NS NS -0.7148 NS

Tc 0.8767 0.9057 NS 0.8441 NS NS NS -0.6826 NS

ΔH 0.7714 0.8228 NS 0.7791 NS NS NS -0.6575 NS

Final viscosity 0.8251 0.7958 NS 0.8578 -0.6889 -0.6799 -0.6889 -0.8307 NS

Total setback 0.9386 0.7997 NS 0.9060 NS NS NS -0.8380 NS

# Statistical significance of the estimated correlations was conducted at p<0.05 NS=Non Significant

183

9 Purified amylopectin 8

Whole starch 7

5 Carbohydrate Carbohydrate (wt. %) 4

0 40 50 60 70 80 90 100 110 120 130 140 150 160 170

Fraction number

Figure 6.1. Gel-Permeation Chromatograms (Sepharose CL-6B) of Debranched

Whole Starch and Purified Amylopectin from Rice Starch (Koshihikari).

184

6 6 12 6 13 TOG 12440 12 IRGC 103759 NERICA 4

5 18 5 5 18 18 4 4 4

3 3 44 3 43 42 (%) Weight (%) Weight 34 (%) Weight 35 2 2 2 35

1 1 6 1 6 6 0 0 0 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 Degree of polymerization Degree of polymerization Degree of polymerization 6 6 12 12 KOSHIHIKARI 14 WITA 4 14 5 5 19 19 4 4 43 43

3 35 3 35

(%) Weight

2 (%) Weight 2

1 6 1 6

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Degree of polymerization Degree of polymerization

Figure 6.2. Unit Chain Outline of Debranched Rice Amylopectins Obtained by High Performance Anion-Exchange Chromatography (HPAEC). (Arrows and Digits Signify Degree of Polymerization) African Rice= IRGC103759; TOG 12440 Asian Rice= Koshihikari; WITA 4 O.sativa x O.glaberrima cross=NERICA 4 185

BS BS BS TOG 12440 IRGC 103759 NERICA 4 Bfp BL Bfp BL Bfp 5 BSmajor 5 5 8 8 8 4 4 4

BL 3 3 31 3 31 31 B2 22 (%) Weight

(%) Weight 22

22 (%) Weight 2 2 B3 2

1 1 1

0 0 0 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 Degree of polymerization Degree of polymerization Degree of polymerization BS WITA 4 BS KOSHIHIKARI BL Bfp BL Bfp 5 4 9 4 5 9

3 32 3 22 29 weight (%) 22 2 Weight (%) 2

1 1

0 0 20 40 60 80 100 0 0 20 40 60 80 100 Degree of polymerization Degree of polymerization

Figure 6.3. Unit Chain Outline of Debranched φ,β-Limit Dextrins of Rice Amylopectins Obtained by High-Performance Anion-Exchange Chromatography (HPAEC). (Arrows and Digits Signify Degree of Polymerization) African Rice = IRGC103759; TOG 12440 Asian Rice = Koshihikari; WITA 4 O.sativa x O.glaberrima cross =NERICA 4

186

Plot of Fitted Model Plot of Fitted Model

16 320

310 15 300

14 290 Enthalpy 280 13 Viscosity Final 270

12 260 10 10.5 11 11.5 12 12.5 13 10 10.5 11 11.5 12 12.5 13 ECL ECL

Enthalpy = 2.70171 + 1.02258*ECL Final Viscosity = 38.9856 + 21.3679*ECL r=0.7791 r=0.8578

Plot of Fitted Model

180

160

140

120 Total Setback 100

80 10 10.5 11 11.5 12 12.5 13 ECL

Total Setback = -281.426 + 35.1199*ECL r=0.9060

Figure 6.4. Multiple Linear Regression Models to Describe the Relationship between ECL, Enthalpy, Final and Setback Viscosities.

187

6.8 References

Bao, J. S. 2012. Toward understanding the genetic and molecular bases of the eating and

cooking qualities of rice. Cereal Foods World 57: 148-156.

Bertoft, E., and Spoof, L. 1989. Fractional precipitation of amylopectin alpha-dextrins using

methanol. Carbohydr. Res. 189: 169-180.

Bertoft, E., and Koch, K. 2000. Composition of chains in waxy-rice starch and its

structural units. Carbohydr. Polym. 41: 121-132.

Bertoft, E. 2004. On the nature of categories of chains in amylopectin and their connection

to the super helix model. Carbohydr. Polym. 57: 211-224.

Bertoft, E., Piyachomkwan, K., Chatakanonda, P., and Sriroth, K. 2008. Internal unit chain

composition in amylopectins. Carbohydr. Polym. 74: 527-543.

Bertoft, E., Koch, K., and Åman, P. 2012. Building block organisation of clusters in

amylopectin from different structural types. Int. J. Biol. Macromol. 50: 1212-1223.

Charoenkul, N., Uttapap, D., Pathipanawat, W., and Takeda, Y. 2006. Simultaneous

determination of amylose content and unit chain distribution of amylopectins of

cassava starches by fluorescent labeling/HPSEC. Carbohydr. Polym. 65: 102-108.

Cooke, D., and Gidley, M. J. 1992. Loss of crystalline and molecular order during starch

gelatinisation: origin of the enthalpic transition. Carbohydr. Res. 227: 103-112.

188

DuBois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., and Smith, F. 1956.

Colorimetric method for determination of sugars and related substances. Anal.

Chem.28: 350-356.

Fennema, O. R. 1985. Water and ice. In: Fennema OR, editor. Food Chem. New York, N.Y.:

Marcel Dekker.: 23-67.

Fredriksson, H., Silverio, J., Andersson, R., Eliasson, A.-C., and Åman, P. 1998. The

influence of amylose and amylopectin characteristics on gelatinization and

retrogradation properties of different starches. Carbohydr. Polym. 35: 119-134.

French, D. 1972 : Fine structure of starch and its relationship to the organization of starch

granules. J. Jpn. Soc. Starch Sci. 19: 8-25.

Gidley, M. J., and Bulpin, P. V. 1987. Crystallisation of malto-oligosaccharides as

models of the crystalline forms of starch: minimum chain-length requirement for the

formation of double helices. Carbohydr. Res. 161: 291-300.

Hanashiro, I., Abe, J.I., and Hizukuri, S. 1996. A periodic distribution of chain length of

amylopectin as revealed by high-performance anion-exchange chromatography.

Carbohydr. Res. 283: 151-159.

Hizukuri, S. 1985. Relationship between the distribution of the chain length of amylopectin

and the crystalline structure of starch granules. Carbohydr. Res. 141: 295-306.

Hizukuri, S. 1986. Polymodal distribution of the chain lengths of amylopectins, and its

significance. Carbohydr. Res. 147: 342-347.

189

Juliano, B.O. 1998. Varietal impact on rice quality. Cereal Foods World 43: 207.

Kalichevsky, M. T., Orford, P. D., and Ring, S. G. 1990. The retrogradation and gelation

of amylopectins from various botanical sources. Carbohydr. Res. 198: 49-55.

Kalinga, D. N., Waduge, R., Liu, Q., Yada, R. Y., Bertoft, E., and Seetharaman, K. 2013.

On the differences in the granular architecture and starch structure between pericarp

and endosperm wheat starches. Starch/Stärke. 65: 791-800.

Klucinec, J.D., and Thompson D.B. 2002. Structure of amylopectins from ae-containing maize

starches. Cereal Chem. 79: 19-23.

Koch, K., Andersson, R., and Åman, P. 1998. Quantitative analysis of amylopectin unit

chains by means of high-performance anion-exchange chromatography with pulsed

amperometric detection. J. Chrom. A. 800:199-206.

Koizumi, K., Fukuda, M., and Hizukuri, S. 1991. Estimation of the distributions of chain

length of amylopectins by high-performance liquid chromatography with pulsed

amperometric detection. J. Chrom. A. 585:233-238.

Koroteeva, D. A., Kiseleva, V. I., Sriroth, K., Piyachomkwan, K., Bertoft, E., and Yuryev,

P.V. 2007. Structural and thermodynamic properties of rice starches with different

genetic background part 1. Differentiation of amylopectin and amylose defects.

Int. J. Biol. Macromol. 41: 391-403.

190

Laohaphatanaleart, K., Piyachomkwan, K., Sriroth, K., Santisopasri, V., and Bertoft, E. 2009.

A Study of the internal structure in cassava and rice amylopectin.

Starch/Stärke. 61:557-569.

Laohaphatanaleart, K., Piyachomkwan, K., Sriroth, K., and Bertoft, E. 2010. The fine

structure of cassava starch amylopectin: part 1: organization of clusters. Int. J. Biol.

Macromol. 47: 317-324.

Manners, D.J. 1989. Recent developments in our understanding of amylopectin structure.

Carbohydr. Polym. 11: 87-112.

Nakamura, Y. 2002. Towards a better understanding of the metabolic system for amylopectin

biosynthesis in plants: rice endosperm as a model tissue. Plant Cell Physiol. 43: 718-

725.

Nikuni, Z. 1978. Studies on starch granules. Starch/Stärke. 30: 105-111.

Peat, S., Whelan, W., and Thomas, G. J. 1952. Evidence of multiple branching in waxy maize

starch. J. Chem. Soc. 4546-4548.

Silverio, J., Fredriksson, H., Andersson, R., Eliasson, A.-C., and Åman, P. 2000. The effect

of temperature cycling on the amylopectin retrogradation of starches with different

amylopectin unit-chain length distribution. Carbohydr. Polym. 42: 175-184.

Takeda, Y., Hizukuri, S., and Juliano, B. O. 1987. Structures of rice amylopectins with low

and high affinities for iodine. Carbohydr. Res. 168: 79-88.

191

Takeda, Y., Hizukuri, S., and Juliano, B.O. 1986. Purification and structure of amylose from

rice starch. Carbohyd. Res. 148: 299-308.

Takeda, Y., Tomooka, S., and Hizukuri, S. 1993. Structures of branched and linear

molecules of rice amylose. Carbohyd. Res. 246: 267-72.

Tester, R. F., and Morrison, W. R., 1990. Swelling and gelatinization of cereal starches I.

effects of amylopectin, amylose and lipids. Cereal Chem. 67: 551-557.

Vamadevan, V., Bertoft, E., and Seetharaman, K. 2013. On the importance of organization

of glucan chains on thermal properties of starch. Carbohydr. Polym. 92: 1653-1659.

Waduge, R., Xu, S., and Seetharaman, K. 2010. Iodine absorption properties and its effect

on the crystallinity of developing wheat starch granules. Carbohydr. Polym. 82: 786-

794.

Wang, L. Z., and White, P. J. 1994. Structure and properties of amylose, amylopectin, and

inter-mediate materials of oat starches. Cereal Chem. 71: 254-262

Wang,Y-J., White, P., Pollak, L., and Jane, J. 1993. Amylopectin and intermediate

materials in starches from mutant genotypes of the Oh43 inbred line. Cereal Chem.

70: 521-525.

Whistler, R. L, and BeMiller, J. N. 1997. Carbohydrate chemistry for food scientists. St. Paul,

Minnesota: Eagan Press.

Zhu, F., Corke, H., and Bertoft, E. 2011. Amylopectin internal molecular structure in

relation to physical properties of sweetpotato starch. Carbohydr. Polym. 84: 907-918.

192

CHAPTER 7: STRUCTURE OF CLUSTERS AND BUILDING BLOCKS IN

AMYLOPECTIN FROM AFRICAN RICE ACCESSIONS

193

Abstract

This study aimed at determining the fine structure of amylopectin from Africa rice accessions

(Oryza glaberrima) and Asian rice (Oryza sativa). The clusters in the amylopectins, which

were produced by hydrolysis using α-amylase from Bacillus amyloliquefaciens and isolated by methanol precipitation, were treated with phosphorylase a and then β-amylase to remove all external chains and obtain φ,β-limit dextrins. The clusters were further hydrolyzed extensively with α-amylase to characterize their building block composition. The clusters, building blocks,

and their unit chain length profiles were analyzed by gel-permeation chromatography and/or

high-performance anion-exchange chromatography with pulsed amperometric detector. While

NERICA 4, which represents a cross between African and Asian rices, possessed the largest

clusters (DP 89.6), the average DPs of the clusters from the African rice samples were slightly

larger (DP ~83) than in the Asian variety (DP ~81). On the basis of number of chains (NC), the

African rice samples represented both the smallest (11.8; IRGC 103759) and the largest clusters

(14.1; TOG 12440). The short CL of TOG 12440 accounted for the large NC of its clusters and

similarly, the clusters of NERICA 4, which had high DP, had intermediate NC due to long CL.

Overall, the result showed that the structure of the clusters in TOG 12440 was dense with short

chains and high degree of branching, whereas in NERICA 4 the situation was the opposite.

IRGC 103759 and the Asian rice (WITA 4) possessed clusters with intermediate characteristics.

The major type of building blocks in all samples was group 2 (representing between 40.3-

49.4%) followed by group 3 (27.0-29.3%). Similar amounts of groups 4 and 5 (9.3-14.4%)

were present in the clusters, whereas group 6 represented only between 1.0 and 4.4% of the

building blocks. The average number of building blocks (NBbl) in the clusters was significantly

194

larger in NERICA 4 (5.8) and WITA 4 (5.7) than in the two African rice accessions (5.1 and

5.3). Key words: African rice, starch, amylopectin, clusters, building blocks

195

7.1 Introduction

Oryza glaberrima (African rice), is one of the two main domesticated rice that is cultivated in

the West Africa sub region mainly due to its ability to withstand the harsh environmental

conditions and Oryza sativa (Asian rice) is the other distinct type of domesticated rice having a

global distribution particularly in Asia. The predominant type of rice consumed worldwide is

the polished or white rice (Nanri et al. 2010) of which the major component is starch (90% dry

basis). Starch contains two major types of biomacromolecules namely: amylose and

amylopectin, of which amylopectin is considered as the major component in storage starch and

accounts for up to 85% by weight (Nakamura 2002). Amylopectin is reported to be responsible

for the semi-crystalline growth ring structure in the starch granules, which consist of stacks of

alternating amorphous and crystalline lamellae (Jenkins et al. 1993). In the crystalline lamellae,

short A-chains (un-substituted chains) and external sections of B-chains (substituted chains)

form double helices that crystallize into the so-called A-allomorph crystal type (Bertoft et al

2008; Imberty et al. 1991).

When amylopectin is treated with exo-acting enzymes, such as phosphorylase or β-amylase, the internal part of amylopectin can be isolated by the removal of its external chains (Manners

1989). When phosphorylase and β-amylase are used successively, φ,β-limit dextrins are produced and debranching of these dextrins leads to the exposure of the internal chains, which can be analysed by anion-exchange chromatography (Bertoft 2004). The internal unit chain distributions in amylopectins from different botanical sources have been analyzed by Bertoft et al. (2008). They reported that the internal chains were all B-chains and could be divided into short (BS) and long chains (BL). BS-chains were divided into very short “fingerprint” Bfp- chains with a degree of polymerisation (DP) 3-7, and BSmajor, with DP 8-25 (the exact range

196 ∼

depends on the botanical origin of starch). The BL-chains were divided into B2-chains

(DP 26-50) and B3-chains (DP > 50). Based on the internal unit chain profile, Bertoft et al.

(2008)∼ categorized amylopectins from different plants into four structural types. Cereals belong to either type 1 or type 2 amylopectins, of which rice is type 2. This type of structure typically contains considerably more BL-chains and a narrower size distribution of BS-chains compared to type 1.

Branches of amylopectin combine to form clusters, which are defined as groups of branches located within a length of nine glucose residues along the backbone (Bertoft 2007). Chains of clusters that constitute the crystalline lamellae are external chains, while the chain segments forming the amorphous lamellae are internal chains and defined as segments between branches

(Pérez and Bertoft 2010). In the backbone model of amylopectin, the branches of the clusters are outspread along a backbone, which is mostly composed of longer amylopectin chains. In the amorphous lamellae, the branches form tightly branched regions inside the clusters known as building blocks, which have internal chain lengths of ~1-3 glucose residues. Building blocks occur in a range of sizes; the smallest consists of two chains and have a degree of polymerization (DP) of 5 to 9, whereas the largest blocks have more than 10 chains (DP > 45).

The average numbers of building blocks (NBbl) in clusters have been reported to vary from about 3 in the small clusters of potato to 9 in clusters of cassava (Bertoft 2007; Bertoft et al.

2010). Within the cluster, the building blocks are interconnected by inter-block chain lengths

(IB-CL) of ~5 to 8 glucose residues (Bertoft et al. 2012b).

α-Amylase of Bacillus amyloliquefaciens has been used in the past to isolate and characterise

clusters and building blocks from a range of different plant sources (Kong et al. 2009; Zhu et al.

2011; Bertoft et al. 2012a). Clusters are formed through a carefully controlled limited

197

hydrolysis (Bertoft 1999: Zhu et al. 2011), whereas building blocks are isolated by an extensive

hydrolysis of clusters into dextrins which are in practice considered to be α-LDs (Kong et al.

2009: Bertoft 2007a). The objective of this study is to investigate the structure of clusters and

building blocks in amylopectin from two African rice accessions. They are compared with one

Asian rice variety and a cross between African and Asian rice.

7.2 Materials and Methods

7.2.1 Materials

The Africa Rice Center based in Cotonou, Benin, provided four dehusked and polished rice samples for this study. Two of the samples were African rice accessions (ARAs) – IRGC

103759 and TOG 12440. The others were one Asian rice variety – cv. WITA 4 – and an

interspecific cross between African rice x Asian rice – NERICA 4. All chemicals and solvents

used were of ACS certified grade.

7.2.2 Enzymes

The enzymes used for the analysis were β-amylase from barley (EC 3.2.1.2, specific activity

705 U/mg), isoamylase from Pseudomonas amyloderamosa (EC 3.2.1.68, specific activity 210

U/mg), pullulanase from Klebsiella pneumoniae (EC 3.2.1.41, specific activity 699 U/mg), and

α-amylase from B. amyloliquefaciens (EC 3.2.1.1), supplied by Megazyme International,

Wicklow, Ireland. The method described by Bertoft et al. (1993) was used to determine the

activity of the α-amylase at pH 6.5, 25°C. Phosphorylase a from rabbit muscle (EC 2.4.1.1, ~25

U/mg protein or 9 U/mg solid) was from Sigma–Aldrich, Deisenhofen, Germany.

198

7.2.3 Methods

7.2.4 Starch Isolation

Starch was isolated from the rice samples as described in chapter 4.

7.2.5 Isolation of Amylopectin from Rice Starch and Determination of Purity

The protocols outlined in chapter 6 were used to isolate amylopectin from the rice starch

samples and test for purity.

7.2.6 Cluster Preparation - Time Course for α-Amylolysis of Amylopectin

Time course analysis of α-amylolysis for the production of clusters from amylopectin was executed as per the method described by Zhu et al. (2013) and Kong et al. (2009). Amylopectin

(20 mg) was dissolved in 400 μL of 90 % DMSO in a boiling water bath and thereafter stirred constantly overnight at room temperature. The dissolved sample was then diluted with hot double-distilled water (1.4 mL) and cooled to 25 °C in a water bath. 200 μL of α-amylase (0.9

U/mL) in sodium acetate (NaOAc) buffer (0.01 M; pH 6.5) equilibrated to 25 °C was added to hydrolyse the sample for more than 2 h with constant magnetic stirring. The enzyme and substrate concentrations were 0.09 U/mL and 10 mg/mL, respectively, during the incubation.

Aliquots (200 μL) were drawn from the reaction mixture at 20 min interval up to 160 min into

Eppendorf tubes containing 40 μL of NaOH (5 M) to halt the reaction by destroying the α- amylase. The samples were stored at –18 °C if not analyzed immediately. To determine the molecular size distribution, the samples were diluted with 200 μL of water and analyzed by gel-

199

permeation chromatography (GPC) on the column of Sepharose CL-6B as described previously

(chapter 6).

7.2.7 Production of Clusters and Their Corresponding φ,β-Limit Dextrins

Clusters from amylopectin and their corresponding φ,β-limit dextrins were produced by the

method reported by Kong et al. (2009). Amylopectin (100 mg) was dissolved in 1.6 mL of 90%

DMSO in centrifuge tubes by gently heating and then constantly stirring overnight at room

temperature. Hot double-distilled water (5.6 mL) was added and the sample was cooled to 25

°C in a water bath. 800 µL of α-amylase (0.9 U/mL) in NaOAc buffer (0.01 M, pH 6.5) was

added to start the reaction in the water bath with constant magnetic stirring. The reaction was

continued for 90 min (based on the time course experiment) before halting it by adding 160 µL

NaOH (5 M). The mixture was allowed to stand at room temperature for 1 h to ensure the

destruction of the α-amylase. Methanol (40 mL) was added to precipitate the clusters in the

form of α-dextrins. The sample was allowed to precipitate for 3 h at 4 °C before recovering the

clusters by centrifugation (4000 x g, 20 min) at 4 °C. The precipitate was washed twice with

pure methanol (10 mL) and allowed to dry overnight in a fume hood.

For the production of φ,β limit dextrins, samples of clusters were dissolved in hot MilliQ water

(30 mL) and the carbohydrate content was adjusted to ~3 mg/mL. For 1 volume of the solution,

0.1 volume of 1.1 M sodium phosphate buffer (pH 6.8), 0.05 volumes of 2.8 mM EDTA, and

0.25 volumes of freshly prepared phosphorylase a solution (25 U/mg 2.5 mg dissolved in 25

mL MilliQ water) were added. The solution was stirred overnight at room temperature, after

which the reaction was terminated in a boiling water bath for 5 min. The solution was

concentrated with a rotary evaporator to 8 mL. The carbohydrate on the walls of the flask was

200

recovered by washing with 2 mL of water. Removal of glucose 1-phosphate was performed on

two PD-10 columns (Sephadex G-25, Pharmacia, Uppsala, Sweden) coupled in series: 2 mL

sample was applied on the columns followed by 3 mL MilliQ water. The eluate was discarded and then 5 mL MilliQ water was added and the sample was collected. The carbohydrate content was adjusted to ~3 mg/mL and the phosphorolysis and purification procedures were repeated to

produce φ-limit dextrins of clusters (φ-LD).

The carbohydrate concentration of purified φ-LD was adjusted to ~3 mg/mL. For 1 volume of

the purified φ-LD solution, 1/3 volume of 0.01 M NaOAc buffer (pH 6.0) and β-amylase (~4

U/mg substrate) were added and the solution was stirred overnight at room temperature. After

termination of the reaction in a boiling water bath for 5 min, the solution was concentrated by

rotary evaporation and purified on PD-10 columns as described above. The β-amylolysis step

was repeated and the φ,β-limit dextrins of clusters (φ,β-LD) produced were finally purified two times on the PD-10 columns to completely remove maltose. The purified φ,β-LD was lyophilized or directly analyzed on the Sepharose CL-6B column for size distribution of the

clusters as described above.

7.2.8 Chain Length Distribution of Clusters

φ,β-LD (1 mg) was weighed and dissolved in warm water (900 μL). 0.01 M NaOAc buffer

(100 μL; pH 5.5) was added and debranching of φ,β-LD was done overnight at room

temperature with 2 μL pullulanase (36.3 U/mL) and 2 μL isoamylase (1000 U/mL). The unit

chain distribution was analyzed with high-performance anion-exchange chromatography with

pulsed amperometric detection (HPAEC-PAD) (Dionex ICS 3000, Synnyvale, CA, USA) on a

201

CarboPac PA-100 column (4 x 250 mm) (Bertoft et al. 2012a). PAD signals were converted to

carbohydrate content (Koch et al. 1998).

7.2.9 Preparation of Building Blocks and Determination of Structure

The production of building blocks from the φ,β-LDs of clusters was done as reported by Bertoft et al. (2012a). φ,β-LD (5 mg) was dissolved in warm water (540 μL) and thereafter hydrolysed with α-amylase (60 μL, 60 U/mL in 0.01M NaOAc buffer, pH 6.5) for 3 h at 35 °C. The reaction was terminated by boiling in a water bath for 5 min. Then NaOAc buffer (60 µL; 0.01

M, pH 6.0) and β-amylase (3 µL) were added to remove any remaining external chain segments. After 18 h of β-amylolysis the reaction was stopped by boiling in a water bath for 5 min. The size distribution of the building blocks was analyzed on a column (1.6 x 90 cm) of

Superdex 30 (Pharmacia, Uppsala, Sweden) with NaCl (0.05 M) as an eluent set at a flow rate of 1 mL/min. The column was calibrated using α-dextrins as described by Bertoft and Spoof

(1989). The carbohydrate content of the fractions (1 mL) collected was estimated by using the phenol-sulphuric acid method (Dubois et al. 1956).

7.2.10 Statistical Analysis

The statistical programme Statgraphics Centurion XV, Version 15.1.02 (StatPoint, Warrenton,

VA, USA) was employed to conduct statistical analysis. One-way analysis of variance was performed to determine significant differences between any pair of means at the 95.0% confidence level and multiple range tests were done to determine which means were different.

202

7.3 Results and Discussion

7.3.1 Time Course Analysis of α-Amylolysis of Amylopectin from Rice Starch

Figure 7.1 shows the degradation by time of amylopectin from rice starch hydrolyzed with α- amylase. The amylopectin isolated from WITA 4 starch sample was used as a model for the assessment of a suitable time of α-amylolysis for the production of clusters. The kinetics of the hydrolysis as shown by the change in molecular size distribution with time was consistent with earlier reports on amylopectin from sweet potato (Zhu et al. 2011), grain amaranth (Kong et al.

2009) and maize (Zhu et al. 2013). The size distribution of the dextrins produced was bimodal and fraction A (Figure 7.1) characteristically represented single clusters or groups of clusters

(domains), while fraction B represented small fragments mainly derived from external chains of

amylopectin (Bertoft 1989). The hydrolysis process was initially very rapid resulting in

increased number of dextrins, but the rate slowed down after approximately 90 min. α-

Amylase of B. amyloliquefaciens has been reported to have nine subsites, which bind the D- glucosyl units of its substrate. α-Amylase is known to react at a very fast rate when all of its nine subsites are filled. However, when all the subsites are not filled, the reaction slows down

(Robyt and French, 1963). α-Amylase attacks the external chains of amylopectin leading to the

formation of maltohexaose. Simultaneously, however, it also initially attacks the long internal

chains between the clusters of amylopectin, which results in the release of clusters (Bertoft

1989). In this study, the activity of α-amylase used was 0.09 U/mL and when the relative molar

amount of branched dextrins (fraction A) was plotted against time (Figure 7.2), it was obvious

that the branched dextrins became less susceptible to further α-amylolysis after approximately

203

90 min (i.e., at this time a high proportion of clusters existed in the reaction mixture). Hence 90 min of α-amylolysis was chosen for cluster production from all four amylopectin samples.

7.3.2 Characterization of φ,β-Limit Dextrins Produced from Clusters

Following the choice of 90 min as the time for hydrolysis, the four amylopectin samples were subjected to α-amylolysis in order to release clusters. The clusters produced in the form of branched α-dextrins (represented by fraction A in Figure 7.1) were precipitated with methanol.

These sets of clusters were further hydrolyzed with phosphorylase a and β-amylase to cleave external chains that remained after α-amylolysis. The molecular size distribution patterns of the resulting φ,β-limit dextrins of clusters (φ,β-LDs) were rather similar and are shown in Figure

7.3 and Table 7.1. However, the average DP of the clusters from the African rice was slightly larger (DP ~83) than in the Asian variety (WITA 4; DP 81). The largest clusters were, however, possessed by the variety representing a cross between African and Asian rices

(NERICA 4; DP 89.6). In a previous report, disparities were observed in the sizes of the clusters isolated from different botanical sources. Generally, starches that exhibit B-type crystallinity (e.g., potato and yam starch) have relatively small clusters compared to those with

A-type crystallinity such as cereals (Bertoft et al. 2012a). The sizes of the clusters in this study were larger than those previously reported for clusters isolated from several plant sources

(Bertoft et al. 2012a). Only the clusters from Andean yam bean starch, with an average size of

DP 82.7, were comparable to the sizes reported for rice in this study.

The φ,β-LDs of clusters were debranched by isoamylase and pullulanase before further analysis with HPAEC to obtain the unit chain distribution (Figure 7.4), from which the average chain length (CL) was estimated. CL ranged from 5.9 to 7.3 (Table 7.1). A number of structural

204

parameters for the characterization of cluster structures based on DP and CL were also calculated and are presented in Table 7.1. Generally, these data compared well with a previous report on Asian rice and fit with the structural features of clusters from group 2 amylopectins

(Bertoft 2012). The average number of chains per cluster (NC), which is another parameter that can be used to express the size of the clusters (Bertoft et al. 2012a), ranged from 11.8 to 14.1.

On this basis the African rice samples represented both the smallest (IRGC 103759) and the largest clusters (TOG 12440). The large NC of the clusters from TOG 12440 was due to their short CL. Similarly, the clusters of NERICA 4, which had high DP, had intermediate NC due to long CL. The internal chain length (ICL) ranged from 3.8 to 5.3 and was reflected in the degree of branching (DB), which ranged from 12.7 to 15.8% and corresponded to NERICA 4 and TOG 12440, respectively. The total internal chain length (TICL) of the clusters in TOG

12440 was also short (7.9), whereas NERICA 4 possessed the longest TICL (8.8). Overall, the result showed that the structure of the clusters in TOG 12440 was dense with short chains and high DB, whereas in NERICA 4 the situation was the opposite. IRGC 103759 and WITA 4 possessed clusters with intermediate characteristics (Table 7.1).

7.3.3 Unit Chains in Clusters

In the structure of amylopectin, A-chains refer to chains that do not carry any other chains by α-

(1,6)-linkages while B-chains carry other chains (Peat et al. 1952). To differentiate the new chains produced in a cluster from those that originally existed in amylopectin, the comparable chains in clusters are designed by lowercase letters (Bertoft et al. 2012a). The b-chains in isolated clusters have two origins; some are produced by α-amylase action, while some arise from pre-existing B-chains in the parent amylopectin (Bertoft, et al. 2012a). The b-chains were

205

further sub-divided into different categories (Figure 7.5) to describe their possible role in the

structure of the clusters (Bertoft and Koch 2000; Bertoft et al. 2012a). Chains of the categories

b1a, b1b, b2, and b3 are thought to play a role in interconnecting the building blocks in a

cluster. The major types of inter-block chains are b1- and b2-chains, which for the most part

correspond to the BSmajor -chains in the original amylopectin. The b1-chains contain one inter- block segment and are divided into b1a and b1b subgroups: b1a-chains have a DP from 7 to 10 and b1b-chains a DP from 11 to 18. The b2-chains (DP 19-27) and b3-chains (DP > 28) are involved in the interconnection of three and four building blocks and contain two and three inter-block segments, respectively (Bertoft et al. 2012a). b0-Chains, which are the smallest type of b-chains with DP 4-6, are fully embedded in the building blocks and have no inter-block segments. Substantial amounts of b0-chains are produced during α-amylolysis and are therefore remnants of inter-cluster segments of the parent amylopectin molecule (Bertoft et al.

2012a). a-Chains are also to a substantial degree, a result of the action of the α-amylase. These chains are represented by maltose in the debranched ϕ,β-LD chromatogram (Fig. 7.4). In addition, clusters possess typically a lot of chains with DP 3 (Bertoft et al. 2008; Bertoft et al.

2012b), and this was also the case with the rice samples in this study. These short chains represent a mixture of a-chains and the shortest possible b-chains (Bertoft et al. 2012a).

Table 7.2 shows the molar distribution of the chain categories in the clusters from the four rice samples. Significant differences were found in the amount of a-chains (DP 2) in the samples.

While sample IRGC 103759 had most a-chains (49.5%), TOG 12440 had the least of them

(43.5%). Since chains indicated by maltotriose are a mixture of a- and b-chains, it is a challenge to precisely quantify the proportion of each category. The values are, however, interesting to compare with the corresponding A-chains in the parent amylopectin molecules,

206

which all possessed close to 50% A-chains (i.e., the ratio of A:B chains was 1.0) (Table 6.2). In

the clusters, IRGC 103759 and NERICA 4 had similar levels of a-chains with DP 2, but b- chains with DP ≥ 4 were found in rather small number. If the ratio of a:b chains in the clusters were similar to that in their parent amylopectins, this suggests that the chains with DP 3 were almost entirely composed of b-chains. Chains with DP 2 in TOG 12440 and WITA 4 were well below 50%. However, in these samples chains with DP 3 were considerably more abundant than in the other two samples. Possibly, therefore, chains with DP 3 in TOG 12440 and WITA

4 contained a considerable number of a-chains. Another possibility is that the number of a- chains, and thus the ratio of a:b chains indeed did change in some of the samples but not in others. In any case, the differences in the composition of the short chains suggested that there existed specific differences in the length of the segments involved in the interconnection of clusters in the amylopectin samples.

Table 7.2 also shows the composition of the sub-categories of the b-chains. In all samples, b0-

chains predominated and the other b-chains were present in reverse number to their length.

Compared to the number of long B2- and B3-chains in the original amylopectin samples (Table

6.2), the number of long chains in clusters (b3-chains) was low, which shows that they had been

attacked by the α-amylase. There were no special difference in the chain composition between

the African samples and the Asian sample. In fact, all of the chain categories presented in

Table 7.2 possessed the most significant differences between the two African rice samples. The

most probable explanation for this is that these two samples were chosen based on their large

differences in pasting properties: Despite having about the same amylose:amylopectin ratio,

their pasting profile differed remarkably. While peak time was similar and the pasting

temperature of IRGC 103759 (78.1°C) differed only slightly from that of TOG 12440 (78.9 °C),

207

the peak and trough viscosities of IRGC 103759 (253.6; 145.4 RVU) were significantly higher than the corresponding values for TOG 12440 (235.6; 139.3 RVU, respectively). IRGC 103759 also possessed higher values of final, breakdown and setback viscosities. Specific differences in the length of the segments involved in the interconnection of clusters in the amylopectin samples, leading to differences in the packing of the crystallites in the starch granules

(Vamadevan et al. 2013), may have accounted for these disparities in pasting characteristics.

7.3.4 Characterization of Building Blocks in Clusters

Building blocks are defined as very small and tightly branched units in clusters and they can be produced by extensive α-amylolysis of the clusters at high enzyme concentrations (Bertoft et al.

1999; Bertoft 2007b; Bertoft et al. 2011). In addition, small quantities of linear dextrins with

DP 1-6 are formed (Bertoft 2007b; Bertoft et al. 2011) and of these dextrins, those with DP 4-6 are hydrolysed into DP 1-3 when β-amylase is added. This improves the separation of these small dextrins from the branched building blocks in GPC. Figure 7.6 shows the size distributions of building blocks in the clusters of the rice amylopectins analyzed on Superdex

30. The profile of building blocks shows that group 2 (the smallest building blocks which consist of two chains) was most common and group 3 (which consist of three chains) was the second most common type. Groups 4, 5 and 6 (i.e. building blocks with increasing number of chains) were found in sequentially decreasing amounts in all the rice amylopectin samples.

Group 1, which was comprised of glucose, maltose and maltotriose, was found in specific amounts in the different samples. Figure 7.7a shows an example of the building blocks distribution when analysed by HPAEC-PAD. This technique gives a much higher resolution of building blocks with individual DP. In addition, the peaks of glucose, maltose and maltotriose

208

included in group 1 are completely resolved. However, group 5 and 6 cannot be resolved by

HPAEC.

One example of the unit chain distribution of building blocks (from TOG 12440) is shown in

Fig. 7.7b. After subtraction of the pre-existing amount of glucose, maltose and maltotriose

(Fig. 7.7a) the unit chain composition of the branched building blocks can be estimated and a typical unit chain profile is shown in Fig. 7.7c as bar graph. The peak of maltose represented a- chains and, as in the clusters, the peak of maltotriose contained a mixture of a- and b-chains of unknown proportions. The rest of the chains with DP ≥ 4 were all b-chains and they typically possessed a peak at DP 5. High amounts of chains at DP 4, 6 and 7 were also generally present.

Chains longer than DP 10 were mostly only represented in trace amounts.

The structural characterisation of the branched building blocks in clusters is summarised in

Table 7.3. The branched building blocks denoted the major part by weight of the samples

(73.5-83.9%), and by mole they were between 30.0-41.9%. The average DP of these branched blocks ranged between 10.6-13.4 with the two African rice samples and NERICA 4 having significantly higher values than the Asian rice sample. As expected, the NC, CL and ICL were much lower and DB was higher in the building blocks compared to the clusters (Table 7.1).

The two African rice samples had building blocks with higher NC, longer CL and higher DB than the Asian rice sample (Table 7.3). The same can be said for NERICA 4, except that it had a somewhat lower DB than the Asian rice sample.

The composition of building blocks in the clusters is summarised in Table 7.4. The density of building blocks (DBbl), which is a parameter analogous to the density of branches (Bertoft et al. 2010), ranged from 6.1 to 7.0, of which the latter corresponded to WITA 4. Thus, African rice accessions and NERICA 4 possessed lower density of building blocks than WITA 4.

209

IB-CL corresponds to the average length of the chains between the branches in two adjacent building blocks in the cluster (Bertoft et al. 2010) and was similar in all samples. The average

number of building blocks (NBbl) in the clusters was significantly larger in NERICA 4 and

WITA 4 than in the two African rice accessions.

The relative distribution of the different groups of building blocks is also shown in Table 7.4.

Group 2 was the major type in all samples and it represented between 40.3-49.4% of the blocks.

Group 3 was the second largest group representing 27.0-29.3%, while group 4 was represented

by roughly 11.0-12.5% of the blocks in all cluster samples. Similar amounts of group 5 as of

group 4 were present in the clusters, whereas group 6 represented only between 1.0 and 4.4% of

the building blocks. The most significant difference between the African and the Asian rice

samples was a higher representation of the large building blocks of groups 5 and 6 in the former

samples. Instead, group 2 was present in lower amounts in African rice samples. Indeed, the

abundant presence of large blocks of groups 4, 5 and 6 in TOG 12440 resulted in the large

average DP (13.4) of the branched blocks.

Generally, the size-distribution of the building blocks in the rice samples was similar to that

found for most other starches previously (Bertoft et al. 2012; Zhu et al. 2013; Kong et al.

2008)). However, the size-distribution of TOG 12440 was rather unique, as it possessed the

lowest number of group 2 blocks so far recorded. In addition, building blocks of group 5 were

present in higher amounts than group 4, which also was a unique feature. More African rice

samples should be analysed for their building block compositions before it can be concluded if

these starches in general tend to possess larger building blocks than other starches.

210

7.3.5 Correlations

Despite the small sample size of the study, correlations between IB-CL and NBbl and some thermal and pasting properties of starch from the rice samples were performed to check the existence of any relationships. NBbl correlated negatively (-0.7522; p<0.05) with enthalpy but all other correlations were not significant (p<0.5).

7.3.6 Conclusions

The average DPs of the clusters from the African rice samples were slightly larger (DP ~83) than in the Asian variety (DP ~81) while NERICA 4 (DP 89.6), which represents a cross between African and Asian rices, possessed the largest clusters. The short CL of TOG 12440 accounted for its large NC and similarly, the clusters of NERICA 4, had intermediate NC due to its long CL. Generally, the results showed that the structure of the clusters in TOG 12440 were dense with short chains and high DB while clusters in NERICA 4 were less dense with long chains and a low DB. IRGC 103759 and WITA 4, however, possessed clusters with intermediate characteristics. The most significant difference between African and Asian rice appeared to be found in the composition of building blocks. The two African rice samples and

NERICA 4 had significantly higher average DP of the branched blocks than the Asian rice sample due to the abundance of large building blocks of group 6. This also resulted in higher average NC, longer CL and higher DB of the building blocks in the African rice samples. The major type of building blocks in all samples was, however, group 2 followed by group 3.

211

7.3.7 Acknowledgements

This research work was undertaken with the kind assistance of the Global Rice Science

Partnership (GRiSP) and the Africa Rice Center (AfricaRice) in terms of financial support and supply of samples.

212

7.3.8 Tables and Figures

Table 7.1. Characterization of φ,β Limit Dextrins of Clusters from Amylopectin# 6 Sample DP1 NC2 DB (%)3 CL4 ICL5 TICL African rice IRGC 103759 82.8b 11.8a 13.0a 7.1c 5.1c 8.5b

TOG 12440 83.3c 14.1c 15.8c 5.9a 3.8a 7.9a

O. sativa x O. glaberrima cross NERICA 4 89.6d 12.4b 12.7a 7.3d 5.3c 8.8b

Asian rice WITA 4 81.0a 12.0ab 13.7b 6.8b 4.7b 7.9a

#Values followed by the same letters in each column are not significantly different (p < 0.05)

1 Average Size of the Cluster (DP) as Measured by GPC (Fractions containing < 1 % Carbohydrate were excluded)

2Number of Chains = DP/CL

3Degree of Branching = (NC – 1)/DP x 100

4Average Chain Length Estimated by HPAEC 5 Internal Chain Length = (CL – ECL) x NC/(NC – 1) – 1. (ECL = 1.5)

6 Total Internal Chain Length = CL of b-Chains – 1

213

Table 7.2. Relative Molar Distribution (%) of Different Chain Categories in Clusters Isolated from Amylopectin# a a and/or b b b0 b1a b1b b2 b3 3 Sample (DP2)1 (DP3)2 (DP≥4) (DP4-DP6)4 (DP7-DP10)4 (DP11-DP18)4 (DP19-DP27)4 (DP>28)4 African rice IRGC 103759 49.5c 9.8a 40.7a 15.1a 11.3a 8.6ab 3.5b 2.3c

TOG 12440 43.5a 12.8b 43.7b 20.6c 12.4c 8.2a 1.9a 0.6a

O. sativa x O. glaberrima cross NERICA 4 49.1c 7.7a 43.1b 16.6b 11.9b 8.6ab 3.5b 2.6d

Asian rice WITA 4 46.0b 13.3b 40.7a 14.9a 11.8b 9.1c 3.3b 1.7b

#values followed by the same letters in each column are not significantly different (p < 0.05)

1a-Chains

2 Mixture of a- and b-Chains 3b-Chains 4Categories of b-Chains of Clusters (DP ≥ 4)

214

Table 7.3. Structure of Branched Building Blocks in Clusters of Amylopectin from African Rice and Asian Rice# 1 2 3 4 Sample Weight (%) Mole (%) DP NC DB (%) CL ICL African rice IRGC 103759 75.6b 32.1b 12.8b 3.2c 17.2ab 4.0ab 1.9ab

TOG 12440 83.9c 41.9d 13.4c 3.5d 18.3b 3.9a 1.7a

O. sativa x O. glaberrima cross NERICA 4 73.5a 30.0a 12.5b 3.0b 16.1a 4.2b 2.3b

Asian rice WITA 4 76.4b 36.3c 10.6a 2.8a 17.1ab 3.8a 1.8a

#Values followed by a different superscript in each row are significantly different (p< 0.05) 1Number of Chains = DP/CL 2 Degree of Branching = (NC – 1)/DP x 100

3Average Chain Length Estimated by HPAEC

4 Internal Chain Length = (CL – ECL) x NC/(NC – 1) – 1. (ECL = 1.5)

215

Table 7.4. Building Block Structures of Clusters of Amylopectin from African Rice Accessions and Asian Rice Varieties# Molar distribution of groups of building blocks in clusters4 Sample DBbl1 IB-CL2 NBbl3 Group 2 Group 3 Group 4 Group 5 Group 6 African rice IRGC 103759 6.4a 6.9a 5.3a 46.5b 27.0a 11.0a 11.8c 3.8b

TOG 12440 6.1a 6.5a 5.1a 40.3a 28.7b 12.5b 14.4d 4.4b

O. sativa x O. glaberrima cross NERICA 4 6.4a 6.8a 5.8b 46.6b 27.4a 11.2a 11.2b 3.8b

Asian rice WITA 4 7.0b 6.9a 5.7b 49.4c 29.3b 11.2a 9.3a 1.0a

# Values followed by a different superscript in each column are significantly different (P < 0.05)

1 Density of Building Blocks (DBbl) = (NBbl/DPcluster) x 100

2 Inter Block Chain Length (IB-CL) = Mole% Linear Dextrins x DPlinear dextrins / Mole% Branched Building Blocks + 4

3 Average Number of Branched Building Blocks (NBbl) = Wt % Bbl/100 x DPcluster/DPBbl

4 Group of Branched Building Blocks Based on Their Branches

216

Degree of polymerization 3 2 10 10 1 1

5 A 20 min

4 40 min B

60 min

3 80 min

90 min 2 Carbohydrate (wt. %) Carbohydrate 100 min

1 120 min

140 min 0 40 60 80 100 120 140 160 180 Fraction number Figure 7.1. Time Course for α-Amylolysis of Rice Amylopectin by Gel-Permeation

Chromatography on Sepharose CL-6B.

217

0.80 Rice amylopectin

0.70

0.60

0.50 ) number elative 0.40

0.30

0.20 Branched dextrins (r 0.10

0.00 20 40 60 80 100 120 140 Time (min)

Figure 7.2. Development of Relative Molar Amounts of Branched Dextrins in the Time Course for α-Amylolysis of Rice Amylopectin for the Production of Clusters.

218

Degree of polymerization

1000 500 100 10 1

IRGC 103759 7 TOG 12440 6 NERICA 4

5 WITA 4

Carbohydrate content (wt%) 2

0 80 90 100 110 120 130 140 150 160

Fraction number

Figure 7.3. Size Distribution of φ,β-limit Dextrins of Clusters Isolated from Rice Amylopectin by GPC on Sepharose CL-6B.

219

100 3

80 5

4 60 6 PAD signal 7 40 8 9

0 0 10 20 30 40 Time (min)

Figure 7.4. HPAEC Chromatogram of Unit Chain Distribution of Cluster from

TOG 12440 (African rice).

220

10 b3 b2 b1 b0 b1b b1a 8

4 Carbohydrate content (%)

0 0 10 20 30 40 50 60 Degree of polymerization

Figure 7.5. Internal Chain Length Distribution of Cluster from TOG 12440 (African rice). Chains with DP ≥ 4 are shown and different b-chain categories are indicated. Peak areas in Fig. 7.4 were corrected to relative carbohydrate content.

221

Degree of polymerization

50 25 10 5 2

7 2

6 3 IRGC 103759

TOG 12440 4 5 1 NERICA 4

WITA 4 4 5

3 Carbohydrate (wt %)Carbohydrate

2 6

0 50 60 70 80 90 100 110 120 130 140 Fraction number

Figure 7.6. Size Distribution Obtained by GPC on Superdex 30 of Building Blocks in Clusters of Rice Amylopectin. Numbers Indicate Different Groups of Building Blocks.

222

600 Impurity in enzyme preparation

400 1 2 PAD Signal

200

3 7 5 6 8 9 10 11 12

0 0 10 20 30 40 50 Time (min)

Figure 7.7a. Building blocks distribution of debranched TOG 12440 (African rice)

when analysed by HPAEC-PAD.

DPs 1-12 are indicated

1 2 600 Impurity in enzyme preparation

400

3 PAD signal 200 4 5 6 7 8 9 10 11 12 0 0 10 20 30 40 50 Time (min) Figure 7.7b. Unit chain distribution of building blocks (from TOG 12440).

DPs 1-12 are indicated 223

5 Carbohydrates (wt. %)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Chain Length

Figure 7.7c. Example of Unit Chain Compositions in Building Blocks of TOG 12440 (African rice) on Weight Basis.

224

7.3.9 References

Bertoft, E. 1989. Investigation of the fine structure of amylopectin using alpha- and beta-

amylase. Carbohydr. Res. 189: 195-207.

Bertoft, E. and Spoof, L. 1989. Fractional precipitation of amylopectin alpha dextrins

using methanol. Carbohydr. Res. 189: 169-180.

Bertoft, E. 1991. Investigation of the fine structure of alpha-dextrins derived from

amylopectin and their relation to the structure of waxy-maize starch. Carbohydr.

Res. 212: 229-244.

Bertoft, E., Manelius, R., and Qin, Z. 1993. Studies on the structure of pea starches. part 1:

initial stages in -amylolysis of granular smooth pea starch. Starch/Stärke. 45: 215-

220.

Bertoft, E., Zhu, Q., Andtfolk, H. and Jungner, M. 1999. Structural heterogeneity in waxy-

rice starch. Carbohydr.Polym. 38: 349-359.

Bertoft, E., and Koch, K. 2000. Composition of chains in waxy-rice starch and its

structural units. Carbohydr. Polym. 41: 121-132.

Bertoft, E. 2004. On the nature of categories of chains in amylopectin and their

connection to the super helix model. Carbohydr. Polym. 57:211-224.

Bertoft, E. 2007a. Composition of clusters and their arrangement in potato amylopectin.

Carbohydr. Polym. 68: 433-446.

225

Bertoft, E. 2007b. Composition of building blocks in clusters from potato amylopectin.

Carbohydr. Polym. 70: 123-136.

Bertoft, E., Piyachomkwan, K., Chatakanonda, P., and Sriroth, K. 2008. Internal

unit chain composition in amylopectins. Carbohydr. Polym. 74: 527-543.

Bertoft, E., Laohaphatanalert, K., Piyachomkwan, K., and Sriroth, K. 2010. The fine

structure of cassava starch amylopectin. Part 2: Building block structure of clusters.

Int. J. Biol. Macromol. 47: 325-335.

Bertoft, E., Källman, A., Koch, K., Andersson, R., and Åman, P. 2011. The cluster

Structure of barley amylopectins of different genetic backgrounds.

Int. J. Biol. Macromol. 49: 441-453.

Bertoft, E., Koch, K. and Åman, P. 2012a. Building block organisation of clusters

in amylopectin from different structural types. Int. J. Biol. Macromol. 50: 1212-

1223.

Bertoft, E., Koch, K. and Åman, P. 2012b. Structure of building blocks in amylopectins.

Carbohydr. Res. 361: 105-113.

Buléon, A., Colonna, P., Planchot, V., and Ball, S. 1998. Starch granules: structure

and biosynthesis. Int. J. Biol. Macromol. 23: 85-112.

Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., and Smith, F. 1956.

Colorimetric method for determination of sugars and related substances.

Anal.Chem. 28: 350-356.

226

Imberty, A., Buléon, A., Tran, V., and Péerez, S. 1991. Recent advances in knowledge of

starch structure. Starch/Stärke. 43: 375-384.

Jenkins, P. J., Cameron, R. E., and Donald, A. M. 1993. A universal feature in the

structure of starch granules from different botanical sources. Starch/Stärke. 45:

417-420.

Klucinec, J. D., and Thompson, D. B. 1998. Fractionation of high-amylose maize starches

by differential alcohol precipitation and chromatography of the fractions.

Cereal Chem. 75: 887-896.

Kong, X., Corke, H., and Bertoft, E. 2009. Fine structure characterization of amylopectins

from grain amaranth starch. Carbohydr. Res. 344: 1701-1708.

Koch, K., Andersson, R. and Åman, P. 1998. Quantitative analysis of amylopectin unit

chains by means of high-performance anion-exchange chromatography with pulsed

amperometric detection. J. Chromatogr. A. 800: 199-206.

Laohaphatanaleart, K., Piyachomkwan, K., Sriroth, K., Santisopasri, V., and Bertoft, E.

2009. A study of the internal structure in cassava and rice amylopectin.

Starch/Stärke. 61: 557-569.

Manners, D. J. 1989. Recent developments in our understanding of amylopectin structure.

Carbohydr. Polym. 11: 87-112.

Nakamura, Y. 2002. Towards a better understanding of the metabolic system for

amylopectin biosynthesis in plants: rice endosperm as a model tissue. Plant and

Cell Physiology 43: 718-725.

227

Nanri, A., Mizoue, T., Noda, M., Takahashi, Y., Kato, M., and Inoue, M. 2010. Rice

intake and type 2 diabetes in Japanese men and women: the Japan public health

center-based prospective study. Am. J. Clin. Nutr. 92: 1468-1477.

Peat, S., Whelan, W., and Thomas, G. J. 1952. Evidence of multiple branching in waxy

maize starch. J. Chem. Soc. 4546- 4548.

Pérez, S., and Bertoft, E. 2010. The molecular structures of starch components and their

contribution to the architecture of starch granules: A comprehensive review.

Starch/Stärke. 62: 389-420.

Robyt, J., and French, D. 1963. Action pattern and specificity of an amylase from Bacillus

subtilis. Arch. Biochem. Biophys. 100: 451-467.

Takeda, Y., Hizukuri, S., and Juliano, B. O. 1987. Structures of rice amylopectins with

low and high affinities for iodine. Carbohydr. Res. 168: 79-88.

Vamadevan, V., Bertoft, E., Soldatov, D. V., and Seetharaman, K. 2013. Impact of

molecular organization of amylopectin in starch granules upon annealing.

Carbohydr. Polym. 98: 1045-1055.

Zhu, F., Corke, H., Åman, P. and Bertoft, E. 2011. Structures of clusters in sweet potato

amylopectin. Carbohydr. Res. 346: 1112-1121.

Zhu, F., Bertoft, E and Seetharaman, K. 2013. Characterization of internal structure of

maize starch without amylose and amylopectin separation. Carbohydr. Polym. 97:

475-481.

228

CHAPTER 8: CONCLUSIONS AND FUTURE WORK

Due to the importance of starch in rice quality, cooking and processing, this study focused on the structural and functional characteristics of flour and starch from African rice (Oryza glaberimma). Samples of African rice accessions (ARAs) were objectively selected from a collection of accessions with a diverse range of apparent amylose concentration and pasting characteristics. The hierarchical cluster analysis of the 1,020 rice accessions using Ward method which resulted in five main clusters and twenty-five sub clusters, revealed that some of the groupings had similar apparent amylose content but interestingly, their pasting properties were very different; a situation that necessitated further investigations. No waxy types of ARAs were analyzed in this study and it is also worthy to note that in comparison to the O. sativas (Asian rice) the ARAs seem to have a narrower range of apparent amylose content (AAC). This observation is important since AAC is related to the cooking quality of rice and other physico-chemical properties such as gel consistency, pasting viscosity and gelatinization temperature. Additionally, AAC is also considered to be one of the most important predictors of various rice products and the eating quality of cooked rice.

The African rice population of 1,020 accessions showed pasting temperatures that were far higher than those reported on Asian rice in previous studies. In addition, their peak viscosity range was narrower, while final viscosity range was broader than earlier reports.

The peak viscosity highly correlated with trough, breakdown and final viscosities (p<0.01) in all the five clusters while breakdown viscosity correlated negatively (p<0.01) with both peak time and pasting temperature in all the five clusters. Further to these observations, final viscosity also correlated highly (p<0.01) with setback viscosity in all the five clusters while correlation between peak viscosity and apparent amylose content was not significant 229

(p<0.01) within clusters II and IV. This categorization could serve as a useful tool for selecting materials with promising pasting properties that can be employed in the development of new rice cultivars. Based on grain dimensions and apparent amylose contents, the seven ARAs (IRGC 86741, IRGC 80787, IRGC 103759, RAM 123, TOG

7682, TOG 12440 and TOG 14365) selected from the population of 1,020 through cluster analysis were medium shaped with intermediate to high amylose rice types and relatively high protein content. TOG 12440 was also a high amylose rice type however, it had bold shape (L/W=1.91). Their relatively higher amylose concentration was a major factor in influencing some of the cooking quality and textural parameters measured. The cooked kernels of the selected ARAs had relatively lower cooking solids loss and harder texture compared to the Asian rice varieties which were softer and more adhesive. The Asian rice flours were generally more viscous than the ARAs but showed a higher breakdown tendency. The in vitro digestibility of the cooked kernel flours showed significant differences in the expected glycemic indices of the samples but they were all >70.

Gelatinization transition temperatures (To, Tp, Tc) and enthalpy were significantly higher in the ARAs flours as compared with Asian rice. However, melting transition temperatures of retrograded flour gels were similar. The high enthalpy of gelatinization (ΔH) of the ARAs is an indication of a more stable structure that was relatively more resistant to granule swelling. The shear stability exhibited by the ARAs may have resulted in the negligible breakdown. The gelatinization transition temperatures (To, Tp, Tc) of the starches and flours were comparable but the ΔH of starch granules were significantly higher than the flours. A similar trend was observed for the melting transition temperatures of the retrograded starch and flour gels.

230

Scanning electron micrographs of the rice starch granules showed polyhedral shaped granules that were tightly packed and the granule size distribution of the starches ranged from 2-22 μm. Additionally, all the starches displayed the typical 'A' type x-ray diffraction pattern. The starches showed significant (p < 0.05) differences in the ratio of the absorbance to scattering coefficient (K/S) values when exposed to iodine vapor with the

ARAs starches recording higher K/S values than the Asian rice samples. This suggested that the glucan chains of the ARAs were less rigid and more available to form inclusion complexes with compared to the Asian rice. The Asian rice starches, apart from having similar physical and morphological characteristics as ARAs starches, differed in their molecular composition (amylose content) and structure, which might affect their functional properties.

Based on a system of classification of amylopectins proposed by Bertoft et al. (2008), the structure of the amylopectins of two ARAs samples (TOG 12440 and IRGC 103759) in comparison with two Asian rice samples (Koshihikari and WITA 4) and NERICA 4 (which is an inter specific cross between sativa x glaberrima) falls into type 2. The profiles of the unit chain distribution of the amylopectins of the ARAs samples and NERICA 4 were distinct from the Asian rice. The external chain length (ECL) of the ARAs and NERICA 4 were longer than the Asian rice varieties. The data suggest that the ARAs and NERICA 4 can possibly have two complete turns of glucosyl units in their double helices, since six glucosyl units are needed per turn, while the Asian rice varieties generally have less than two turns. This may explain the higher gelatinization enthalpies observed in the ARAs.

The positive correlations between CL, SCL, ECL and gelatinization transition temperatures

(To, Tp, Tc) and enthalpy (ΔH) strongly suggest that the short chains of the amylopectins, 231

especially their external segments were important in influencing the thermal and pasting properties of the starch samples.

The average DPs of the clusters from the African rice samples were significantly (p<0.05) larger than in the Asian rice sample while NERICA 4 possessed the largest clusters.

Generally, the results showed that TOG 12440 clusters were dense with short chains and high DB while clusters in NERICA 4 were less dense with long chains and a low DB.

However, clusters of IRGC 103759 and WITA 4 possessed with intermediate characteristics. The average DP of the branched building blocks of the African rice samples and NERICA 4 were significantly (p<0.05) higher than the Asian rice sample and the major type of building blocks in all samples was group 2 followed by group 3. Similar amounts of groups 4 and 5 were present in the clusters, whereas group 6 represented only up to 4.4% of the building blocks. Additionally, the African rice samples had building blocks with higher number of chains and degree of branching than the Asian rice sample.

The hierarchical cluster analysis of the 1,020 rice accessions showed a high level of genetic diversity in the accessions of O. glaberrima from the West Africa sub region as evidenced by the variation in both apparent amylose content and pasting properties in and within the clusters and sub clusters. The cooking quality, in vitro digestibility, physical, thermal and pasting characteristics of the selected African rice starches and flour were established.

Additionally, the positive correlation between the chain lengths of amylopectin (CL, SCL,

ECL) and gelatinization transition temperatures strongly suggests that the chains influenced the thermal and pasting properties of the starch samples. These findings were based on seven samples selected through the cluster analysis. It will therefore be expedient to work

232

on more African rice samples to affirm the findings in this study especially in the area of starch structure in relation to its functionality in cooking, processing and nutritional aspects

(digestibility).

Based on the information generated through this study, it will also be interesting to develop some products from African rice and test their acceptability in a bid to promote more patronage of African rice.

233