Physicochemical Characterization, Structural Determination, and Molecular Dynamic Modeling of Proso Millet Proteins for Enhanced Food Functionality

University of Kentucky UKnowledge

Theses and Dissertations--Biosystems and Agricultural Engineering Biosystems and Agricultural Engineering

2020

PHYSICOCHEMICAL CHARACTERIZATION, STRUCTURAL DETERMINATION, AND MOLECULAR DYNAMIC MODELING OF PROSO MILLET PROTEINS FOR ENHANCED FOOD FUNCTIONALITY

Felix Akharume University of Kentucky, [email protected] Author ORCID Identifier: https://orcid.org/0000-0003-3426-2669 Digital Object Identifier: https://doi.org/10.13023/etd.2020.354

Right click to open a feedback form in a new tab to let us know how this document benefits ou.y

Recommended Citation Akharume, Felix, "PHYSICOCHEMICAL CHARACTERIZATION, STRUCTURAL DETERMINATION, AND MOLECULAR DYNAMIC MODELING OF PROSO MILLET PROTEINS FOR ENHANCED FOOD FUNCTIONALITY" (2020). Theses and Dissertations--Biosystems and Agricultural Engineering. 74. https://uknowledge.uky.edu/bae_etds/74

This Doctoral Dissertation is brought to you for free and open access by the Biosystems and Agricultural Engineering at UKnowledge. It has been accepted for inclusion in Theses and Dissertations--Biosystems and Agricultural Engineering by an authorized administrator of UKnowledge. For more information, please contact [email protected]. STUDENT AGREEMENT:

I represent that my thesis or dissertation and abstract are my original work. Proper attribution has been given to all outside sources. I understand that I am solely responsible for obtaining any needed copyright permissions. I have obtained needed written permission statement(s) from the owner(s) of each third-party copyrighted matter to be included in my work, allowing electronic distribution (if such use is not permitted by the fair use doctrine) which will be submitted to UKnowledge as Additional File.

I hereby grant to The University of Kentucky and its agents the irrevocable, non-exclusive, and royalty-free license to archive and make accessible my work in whole or in part in all forms of media, now or hereafter known. I agree that the document mentioned above may be made available immediately for worldwide access unless an embargo applies.

I retain all other ownership rights to the copyright of my work. I also retain the right to use in future works (such as articles or books) all or part of my work. I understand that I am free to register the copyright to my work.

REVIEW, APPROVAL AND ACCEPTANCE

The document mentioned above has been reviewed and accepted by the student’s advisor, on behalf of the advisory committee, and by the Director of Graduate Studies (DGS), on behalf of the program; we verify that this is the final, approved version of the student’s thesis including all changes required by the advisory committee. The undersigned agree to abide by the statements above.

Felix Akharume, Student

Dr. Akinbode Adedeji, Major Professor

Dr. Donald Colliver, Director of Graduate Studies

PHYSICOCHEMICAL CHARACTERIZATION, STRUCTURAL DETERMINATION, AND MOLECULAR DYNAMIC MODELING OF PROSO MILLET PROTEINS FOR ENHANCED FOOD FUNCTIONALITY

______

DISSERTATION ______A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the College of Agriculture, Food, and Environment and the College of Engineering at the University of Kentucky

By Felix Umaizimede Akharume Lexington, Kentucky

Director: Dr. Akinbode Adedeji, Associate Professor of Food Process Engineering Biosystems and Agricultural Engineering Lexington, Kentucky 2020

ABSTRACT OF DISSERTATION

PHYSICOCHEMICAL CHARACTERIZATION, STRUCTURAL DETERMINATION, AND MOLECULAR DYNAMIC MODELING OF PROSO MILLET PROTEINS FOR ENHANCED FOOD FUNCTIONALITY

More than one-third of Americans today incorporate plant-based protein into their diet and about 40% believed that plant-based protein is healthier than animal protein, especially Millennials. The increasing global demand for plant-based proteins driven by the high cost of animal proteins, consumers’ desire for lean protein, vegetarianism, and the need for more sustainable green protein products have necessitated research into alternate emerging and underutilized sources of protein to complement or supplement the major plant protein in the market- soy, pea, and gluten. Therefore, this dissertation is focused on the valorization of the proteins in proso millet. Specifically, this work focused on the identification and structure-function characterization of the protein fractions in proso millet to include the determination of the three-dimensional structure of its glutelin fraction isoform (glutelin-type B 5-like protein) and finally, on the application of molecular dynamic modeling simulation to elucidate the effects of simulated processing stresses on the behavior of glutelin-type B 5-like protein at the molecular level. This dissertation is made up of eight chapters with four major objectives from chapter three through to chapter seven. In objective one, four major proteins fractions from two cultivars (Dawn and Plateau) of proso millet flour (albumin, globulin, prolamin, and glutelin) were identified and characterized for their physicochemical properties and functionalities. Prolamin and glutelin were identified as the major protein in proso millet with respective percent composition of 47.2, 39.1 for Dawn cultivar and 50.8, and 34.5 for Plateau cultivar. The average denaturation temperature of 82.1±3.5°C requiring an average enthalpy of 0.1±0.06 J/g was reported for all fractions. Most of the protein fractions showed the highest solubility at pH 9 ranging from 5.7 to 100%; however, these protein fractions showed poor solubility at pH 7 and below (less than 40%). Emulsifying activity index of less than 25 m2/g was recorded for most fractions, while the highest emulsion stability index recorded was about 60 min. In objective two, the effects of three levels ultrasound treatments (50%, 75%, and 100% amplitude and constant 20kHz for 5- and 10-min treatment times) were applied to the two major protein fraction from proso millet flour of Dawn and Plateau cultivars and selected functionality (solubility, emulsion, foam, thermal properties, and invitro-digestibility) of the two major fractions (prolamin and glutelin) were evaluated. It was observed that the ultrasound treatment (US) increased the solubility of the prolamin and the glutelin protein significantly (p<0.05) for both cultivars. For instance, Dawn prolamin showed a protein

solubility from 8.21±0.13% to 22.1±0.78%, 23.9±0.41%, 27.5±6.57%, 24.3±5.03%, 31.1±5.03% and 49.2±1.80% for 50, 75, 100% amplitude for 5 and 10 min, respectively. Additionally, the pepsin digestibility of both prolamin and glutelin also improved compared to their native protein. Dawn prolamin increased by about 30% for ultrasound treatment 100% amplitude for 10 min and Plateau glutelin pepsin digestibility increased by 48.6% for the same level of treatment. Foaming capacity was improved by the US at a higher treatment time of 10 min but there was inconsistency in the emulsion activity index. Objective three presents the results on work carried out to determine the three-dimensional structures of glutelin-type B 5-like protein, a type of glutelin protein using comparative homology, and the attempt made at an empirical approach. The result showed that the structure of glutelin-type B 5-like is a trimer in its natural state with amino acid residue length of 256 and conserved regions (one jelly-like β-barrel and two extended helix domains) with the globulin proteins (11s or pro-11s globulin proteins) of pea, soybean, pumpkin, amaranth, and rapeseed and between 35 - 45% structural similarity with the globulins of these proteins. The attempt made at purifying the glutelin-type B 5-like for X- ray crystallography revealed that the protein was aggregating in solution. The obtained three-dimensional structure contains 4.6% alpha-helix, 2.5% 3/10 helix, 23.8% beta-sheet and 69.1% coils/turns/bends/bridges. The structure was deposited in the protein model database (PMDB) (http://srv00.recas.ba.infn.it/PMDB/main.php) with PMDB identifier PM0083241. Finally, objective four and five reveal the effects of temperature levels (300, 350, and 400 K) combined with electric field levels (0, 0.1, 1 and 3 v/nm) as well as the effects of temperature levels (300, 350, and 400 K) combined with pressure levels (1b, 3kbar, and 6kbar) on the three-dimensional structure conformations of glutelin-type B 5-like in silico. The results showed that the root mean square deviations (RMSD) increased as the intensity of the stresses increased, except for increasing pressure where the RMSD decreased when the temperature was held constant. Moreover, we showed that the amino acid at the terminals of the protein fluctuates more with stresses and the alpha-helix fluctuates more than beta-sheet. While some losses of the amino residue that make up the secondary structure of the glutelin type-B 5-like protein were observed obvious disruption to this secondary structure were not noticed which may suggest that higher processing stress intensity and/or higher simulation time may be needed to cause a major and irreversible disruption of the protein secondary structure.

KEYWORDS: Proso Millet Proteins, Food Processing Stresses, Structure-Function Properties, Three-Dimensional Structure, and Molecular Dynamic Modeling.

Felix Umaizimede Akharume 07/30/2020 Date

PHYSICOCHEMICAL CHARACTERIZATION, STRUCTURAL DETERMINATION, AND MOLECULAR DYNAMIC MODELING OF PROSO MILLET PROTEINS FOR ENHANCED FOOD FUNCTIONALITY

By Felix Umaizimede Akharume

Akinbode Adedeji Director of Dissertation

Donald Colliver Director of Graduate Studies

07/30/2020 Date

DEDICATION

This dissertation and all accompanied achievements are dedicated to Almighty God, the one who loved me, even when I didn’t know myself. To Him be all praise, all glory, and adoration

ACKNOWLEDGEMENTS It goes without saying, that he whom much gift is given and appreciates not, is no different from the robber that steals such a gift. It is to this end that I want to immensely appreciate my advisor and mentor-Dr. Akinbode Adedeji and the rest of my committee members, Dr. Sue Nokes, Dr. Jian Shi, and Dr. Konstantin Korotkov. Personally, I want to thank my advisor for believing in me and constantly encouraging me to stay focused and push through this dissertation project while at the same time pushing me to take on new challenges. I am indebted and full of appreciation for Dr. Konstantin, whose expertise in structural biology and contribution to my project was very pivotal to its success. Particularly, in the design and cloning of my protein and the determination of its three- dimensional structure via homology modeling. Additionally, I will like to thank the duo of Dr. Sue Nokes and Dr. Jian Shi for their contribution to my committee and their encouragement and critical evaluation of my work. This feat would have been impossible without the support and backing of my beautiful and loving wife, who provides the emotional support when the going gets tough and keep the home in great shape for me to thrive. Thank you, sweetheart, for giving me all of you. I would also like to thank my parents (Mr. and Mrs. Akharume) and my siblings Benedicta, Justina, and Celestine who were constantly checking on me to give their moral support and love. As the saying goes, paraphrasing, a child may belong to one family but is raised by multiple people. With this in mind, I would like to thank my friends and lab mates that helped all the way in this journey- Michael Omodara, late Francis Agbali, Joe Woomer, Abuchi Okeke, Kenny Bababode and Aniet Kyomuhangi to mention a few. I will also like to thank the Biosystems and Agricultural Engineering department and staff that welcomed me and gave me a place to call home- Dr. Michael Montross, Ms. Julie Tolliver, Ms. Jayne White, Dr. Alicia Modenbach, and Donnie Stamper my office mate. I would also like to thank members of Dr. Konstantin’s lab that supported me on this project- Catherine T. Chaton, Svetlana Zamakhaeva, Zachary Williamson, and Muna Shakhashiro. Finally, I am most grateful to Almighty God, my help in ages past, and my hope for years to come for his unreserved love and care for me.

iii

TABLE OF CONTENTS ACKNOWLEDGEMENTS……...……………………………………………………….iii TABLE OF CONTENTS…………………………………………………………………iv LIST OF TABLES ……………………………………………………………………..viii LIST OF FIGURES...... x LIST OF ADDITIONAL FILES…..……………………….…………………………...xiv Chapter I: Introduction ...... ………..1 1.1 Rationale..…………………………………………………………………………...4 1.2 Project Goal…………………………………………………………………………5 1.2.1 Specific objectives ...... 5 1.2.2 Hypotheses ...... 6 Chapter II: Literature Review ...... 7 2.0 Introduction ...... 7 2.1 Modification of Plant Proteins by Various Processing Techniques ...... 15 2.1.1 Mechanism of protein modification ...... 15 2.1.2 Physical techniques for modification of plant-based proteins ...... 16 2.1.3 Chemically modified proteins ...... 28 2.1.4 Biological/Enzymatically modified proteins ...... 37 Chapter III: Objective 1 ...... 45 Physicochemical, Functional and Structural Properties of Proso Millet Storage Protein Fractions ...... 45 Abstract ...... 45 3.0 Introduction ...... 45 3.1 Material and Methods ...... 47 3.1.1 Materials ...... 47 3.1.2 Extraction and isolation of protein fractions ...... 49 3.1.3 Physicochemical properties characterization ...... 51 3.1.4 Functional properties ...... 52 3.1.5 Structural properties ...... 53 3.1.6 Experimental design and statistical analysis ...... 54 3.2 Results and Discussion ...... 54 3.2.1 Extraction and isolation of protein fractions ...... 54 3.2.2 Physicochemical properties characterization ...... 57

iv 3.2.3 Functional properties ...... 61 3.2.4 Electrophoresis pattern (SDS-PAGE) ...... 64 3.3 Conclusion ...... 66 Chapter IV: Objective 2 ...... 67 Effects of High-power Ultrasound on the In vitro Digestibility, Physicochemical and Functional Properties of Proso Millet Prolamin and Glutelin Protein ...... 67 Abstract ...... 67 4.0 Introduction ...... 67 4.1 Materials and Methods ...... 69 4.1.1 Materials ...... 69 4.1.2 Extraction and isolation of protein fractions ...... 69 4.1.3 Ultrasound treatment of protein fractions ...... 70 4.1.4 Physicochemical and functional properties characterization ...... 71 4.1.5 In-vitro protein digestibility...... 73 4.1.6 Experimental design and statistical analysis...... 73 4.2 Results and Discussion ...... 73 4.2.1 Protein solubility ...... 73 4.2.2 Particle size analysis ...... 75 4.2.3 Thermal denaturation Profile………………………………………….………77 4.2.4 Emulsion properties ...... 80 4.2.5 Foam properties ...... 81 4.2.6 Pepsin digestibility....…………………………………………………………82 4.3 Conclusion ...... 83 Chapter V: Objective 3a & 3b ...... 84 Main Obj: Determination of the Three-dimensional Structure of Glutelin type-B 5-like Protein ...... 84 Obj 3A: Cloning, Expression, and Purification of His-tagged Glutelin type-B 5-like Protein Isoform from Proso Millet in Escherichia coli ...... 84 Abstract ...... 84 5.0 Introduction ...... 84 5.1 Materials and Methods ...... 85 5.1.1 Construction of glutelin type-B 5-like subunit clone ...... 85

v 5.1.2 Expression of recombinant glutelin type-B 5-like protein and isolation on inclusion bodies ...... 86 5.1.3 Unfolding, refolding and purification of glutelin type-B 5-like from inclusion bodies ...... 86 5.1.4 Fast protein liquid chromatography (FPLC) purification and analysis…....….87 5.1.5 Western blot ...... 87 5.1.6 Dynamic light scattering (DLS) analysis ...... 87 5.2 Results and Discussions ...... 87 5.2.1 Cloning and expression of the recombinant glutelin type-B 5-like ...... 87 5.2.2 Purification of glutelin type-B 5-like protein from inclusion bodies ...... 88 5.2.3 Dynamic light scattering analysis ...... 91 5.3 Conclusion ...... 92 Obj 3B: In silico Analysis and Homology Modeling of Three-dimensional Structure of Glutelin type-B 5-like and Proteins from Proso Millet ...... 93 Abstract ...... 93 5.4 Introduction ...... 93 5.5 Materials and Methods ...... 95 5.5.1 Protein sequence identification and analysis ...... 95 5.5.2 Comparative homology modeling ...... 95 5.5.3 Geometry minimization and model re-assessment ...... 97 5.5.4 Physico-chemical and functional characterization ...... 97 5.6 Results and Discussions ...... 97 5.6.1 BLAST analysis of protein sequence alignment ...... 97 5.6.2 Model building and assessement…………………………………………….100 5.6.3 Model geometry minimization, re-assessment, and validation ...... 104 5.6.4 Physico-chemical and functional characterization ...... 106 5.7 Conclusion ...... 107 Chapter VI: Objective 4 ...... 108 In-silico Modeling of Glutelin type-B 5-like from Proso Millet Seed Storage Protein: Effects of Temperature and Electric Field ...... 108 Abstract ...... 108 6.0 Introduction ...... 108 6.1 Materials and Methods ...... 110

vi 6.1.1 Molecular dynamic simulations...... 110 6.2 Results and Discussions ...... 112 6.2.1 Root Mean Square Deviation (RMSD) ...... 112 6.2.2 Root Mean Square Fluctuation (RMSF) ...... 116 6.2.3 Secondary structure analysis ...... 118 6.2.4 Total Dipole Moment ...... 121 6.2.5 Radius of gyration (Rg) ...... 123 6.2.6 Solvent Accessibility Surface Area (SASA) ...... 125 6.2.7 Volume and density ...... 127 4.0 Conclusion ...... 129 Chapter VII: Objective 5 ...... 130 In-silico Modeling of Glutelin type-B 5-like from Proso Millet Seed Storage Protein: Effects of Temperature and Pressure ...... 130 Abstract ...... 130 7.0 Introduction ...... 130 7.1 Materials and Methods ...... 132 7.1.1 Molecular dynamic simulations...... 132 7.2 Results and Discussions ...... 134 7.2.1 Root Mean square Deviation (RMSD) ...... 134 7.2.2 Root Mean Square Fluctuation (RMSF) ...... 138 7.2.3 Secondary structure analysis ...... 140 7.2.4 Radius of gyration (Rg) ...... 143 7.2.5 Solvent Accessibility Surface Area (SASA) ...... 145 7.2.6 Volume and density ...... 147 7.3 Conclusion ...... 149 Chapter VIII: General Conclusion and Future Work…………………………………... 150 8.0 General Conclusion ...... 150 8.1 Future Work ...... 152 General References ...... 153 Vita……………………………………………………………………………………...183

vii

LIST OF TABLES Table 2.1. General classes of functional properties of proteins important in food applications……………………………………………………………………………….9

Table 2.2. Summary of common plant-based protein ingredients and their functionality……………………………………………………………………...………11

Table 2.3. Techniques for the modification of plant-based proteins for improved functionality...... 15

Table 2.4. A summary of commonly acylated protein ingredients, acylating agents and target functionality……….…………………………………………………………...…32

Table 2.5. A summary of commonly hydrolyzed protein, hydrolysis condition, enzymes, and target functionality…………………………………………….…………………….38

Table 2.6. A summary of common protein ingredients, type of cross-linking enzymes and the target functionality impacted……………………………………………………...…43

Table 3.1. Proximate composition of the proso millet flours used in this study...... 48

Table 3.2. Recovery yield of crude protein extract of and its corresponding protein content of proso millet protein fractions from Dawn and Plateau flour cultivars………...…...…...56

Table 3.3. Thermal denaturation properties of the protein fractions of two proso millet flour from two different cultivars (Dawn and Plateau) ….…………………………....…58

Table 4.1. Thermal denaturation properties of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour……………………………..78

Table 5.1. Summary of purification of recombinant His-tagged glutelin type-B 5-like protein from E. coli……………………………………………………………………….89

Table 5.2. A summary of the glutelin type-B 5-like protein details used for the homology modeling …………………………………………………………………………………96

Table 5.3. A summary of the glutelin type-B 5- like homology templates used for the homology modeling…………………………………………………………………….99

viii

Table 5.4. Summary of glutelin type-B 5-like protein models quality assessment scores……………………………………………………………………………………103

Table 5.5. Summary of glutelin type-B 5-like protein geometry minimized model’s quality assessment scores…………….…………………………………………………………105

Table 5.6. A summary of physicochemical properties and secondary structure assignment for glutelin type-B 5-like protein…………………………...……………….………… 107

Table 6.1. Summary of simulation conditions used in the study……………………….112

Table 6.2. Average root mean square deviation (RMSD), radius of gyration (Rg), solvent accessibility surface area (SASA), volume, and density of glutelin type-B 5-like protein after1 ns simulation under different temperature and electric field conditions.….……...114

Table 6.3. Percentage of secondary structure element per the three-dimensional structure of glutelin type-B 5-like protein molecule after1 ns simulation under different simulation conditions………………………………………………………………………………120

Table 7.1. Summary of Simulation conditions used in the study ………………………..134

Table 7. 2. Average root mean square deviation (RMSD), the radius of gyration (Rg), solvent accessibility surface area (SASA), volume and density of glutelin type-B 5-like protein after1 ns simulation under temperature and pressure conditions……………….136

LIST OF FIGURES Figure 2.1. Protein solubility profile of cowpea flour as affected by pH………….………17

Figure 2.2. Schematic of a typical Ultrasound system………...……………….…………20

Figure 2.3. Scanning electron microscope micrographs of untreated, pressure-treated, and heat-treated 24 g/100 g pea protein concentrate solutions………………...………………24

Figure 2.4. Schematic of a cold plasma system...…………………………………....……27

Figure 2.5. Mechanism of glycosylation reaction between the lysine residue of a protein and glucose leading to N-fructoselysine…………………………………………………29

Figure 2.6. Mechanism of phosphorylation showing the reaction of sodium tripolyphosphate (STP) and sodium trimetaphosphate (STMP) with serine and lysine residues…………………………………………………………………………….…….36

Figure 3.1 Flow chart of sequential extraction of proso millet protein fractions from defatted proso millet flour at room temperature…………………………………………50

Figure 3.2. Surface hydrophobicity index of protein fractions of two proso millet flours from two different cultivars (Dawn and Plateau) at pH 7……………………………...... 60

Figure 3.3. Solubility profiles of protein fractions from two proso millet flour cultivars (Dawn and Plateau) at various pHs (3-9).………………………………………………61

Figure 3.4 Emulsifying activity index (EAI) and emulsifying stability index (ESI) of protein fractions from two proso millet flour cultivars (Dawn and Plateau) at various pH 7………………………………………………………………………………………..…62

Figure 3.5. Foaming capacity (FC) and foaming stability (FS) of protein fractions from two proso millet flour cultivars (Dawn and Plateau) at various pH 7……………...…………64

Figure 3.6. SDS-PAGE profile of proso millet protein fraction from two proso millet flour cultivars (Dawn and Plateau) ……………………………….…………………………...65

Figure 4.1. The solubility of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. Solubility was carried out at pH 7…...………74

Figure 4.2. Dynamic light scattering analysis of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour……………………………76

Figure 4.3. Emulsion activity index of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour………………….……….…………80

Figure 4.4. Emulsion stability index of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour………………………….……………81

Figure 4.5. Foam capacity of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour…………………………………………………82

Figure 4.6. Foam stability of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour…………………………………………………82

Figure 4.7. Pepsin digestibility of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour……………………………….………83

Figure 5.1. SDS and Western Blot analysis of recombinant Glutelin type-B 5-like protein. ……………………………………………………………………………………………89

Figure 5.2. FPLC analysis of glutelin type-B 5-like protein eluted from Ni-NTA resin column loaded on a size exclusion column………………………………………………91

Figure 5.3. Dynamic light scattering analysis of glutelin type-B 5-like protein eluted from the size exclusion column……………………………………………………………...…92

Figure 5.4. Visualization of the superimposed monomers of the glutelin type-B 5-like protein models generated based on five different templates……………………………101

Figure 5.5. Visualization of the monomers of the glutelin type-B 5-like protein showing the variable regions………………………………………………………….………….101

Visualization of the three-dimensional structure of the glutelin type-B 5-like protein model geometrically minimized using PHENIX……………………………………………...102

Figure 5.7. Ramachandra plots of geometry minimalized glutelin type-B 5-like protein models…………………………………………………………………………………..106

Figure 6.1. A snapshot of the three-dimensional structure of glutelin type-B 5-like protein molecule (A) in vacuum (B) in neutralized water……………………………………...111

Figure 6.2. Root mean square deviation of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time…115

Figure 6.3. Root mean square fluctuations of glutelin type-B 5-like protein per residues under different simulation conditions (temperature and static electric field) after 1 ns simulation time………………………………………………………………………….117

Figure 6.4. The numbers of residues present in the secondary structure of glutelin type-B 5-like protein molecules under different simulation conditions (temperature and static electric field) after 1 ns simulation time…………………………………………………118

Figure 6.5. Snapshots of glutelin type-B 5-like protein molecule after1 ns simulation under different simulation conditions…………………………………………………………121

Figure 6.6. Total dipole moment of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) after 1 ns simulation time…………....122

Figure 6.7. Radius of gyration of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time………………124

Figure 6.8. Solvent accessibility surface area of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time……………………………………………………………………..………………126

Figure 6.9. Volume changes observed for glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time…128

Figure 7.1. A snapshot of the three-dimensional structure of glutelin type-B 5-like protein molecule (A) in vacuum (B) in neutralized water………………………………………133

Figure 7.2. Root mean square deviation of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time……………137

xii

Figure 7.3. Root mean square fluctuations of glutelin type-B 5-like protein per residues under different simulation conditions (temperature and pressure) after 1 ns simulation time……………………………………………………………………………………..139

Figure 7.4. The numbers of residues present in the secondary structure of glutelin type-B 5-like protein molecules under different simulation conditions (temperature and pressure) after 1 ns simulation time………………………………………...……………………...140

Figure 7.5. Snapshots of glutelin type-B 5-like protein molecule after1 ns simulation under different simulation conditions………………………………………………………….142

Figure 7.6. Radius of gyration of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time…………………………144

Figure 7.7. Solvent accessibility surface area of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time….146

Figure 7.8. Volume changes observed for glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time……………..148

xiii

LIST OF ADDITIONAL FILES 1. Structural data for glutelin type B 5-like protein

xiv

Chapter I: Introduction

The global plant protein product sales were estimated at $35 billion in 2018 and projected to be $45 billion by 2023 at a compound annual growth rate (CAGR) of 4.93% over a forecasted period from 2018-2023 (Cassity, 2019). The market has been dominated by animal-based protein ingredients. However, over the last three decades, there has been growing interest in plant-based protein ingredients, partly prompted by the rising cost of animal-based protein ingredients, consumers growing preferences for lean protein, increasing dietary knowledge, and population growth (Henchion, Hayes, Mullen, Fenelon, & Tiwari, 2017). It is well documented that reliance on animal sources alone for our dietary proteins, will require more land, more water and become difficult to support environmentally due to associated greenhouse gas emission (Henchion et al., 2017); hence, a crucial motivation for the growing interest in a more sustainable source like plant-based proteins. The plant-based portion of the protein ingredient market is fast expanding. The retail market sales amounting to $ 5 billion was reported in 2019 (GFI, 2020).

Cereal grains are particularly a plant-based protein source of interest because they are a widely grown, high yielding crop and an important source of the world’s food supply (Saleh, Zhang, Chen, & Shen, 2013; Shewry & Halford, 2002). For example, three of the cereal crops (rice, corn, and wheat) provides about two-third of global dietary energy intake (Gillespie & van den Bold, 2017); sorghum and millet are widely consumed in East and West Africa, some parts of India and Asia, and in limited amount in the United States of America (Henchion et al., 2017; Nirgude et al., 2014). While cereal grain storage protein may be attracting a lot of research interest, some cereal grain storage proteins of wheat, rye, and barley have been identified to trigger an allergic response, an autoimmune reaction called celiac-diseases (CD) in certain consumers (Gulati et al., 2017; Mesa‐Stonestreet, Jhoe, Alavi, & Bean, 2010; van den Broeck et al., 2009). This restricts celiac patients to a life-long gluten-free diet.

In the light of this development, there has been a significant expansion in research on the use of gluten-free cereal grain, such as corn, rice, sorghum, and millet as alternatives to wheat, barley, and rye in food applications in the past 25 years (Taylor, Taylor,

Campanella, & Hamaker, 2016). Millet, a gluten-free plant-based protein source is one of such alternative cereal grains attracting growing research interest owing to the fact that it is richer in the amount of essential amino acid (leucine, isoleucine, methionine), compared to wheat protein, and its protein concentration is higher than those of corn, wheat and other cereal grains (Kalinova & Moudry, 2006; Lorenz, Dilsaver, & Bates, 1980). Additionally, from a sustainability perspective, millets are a short season crop, disease-resistant, require low water for propagation to maturity, and adaptable to many soils and climatic conditions (Baltensperger 2002; Sheahan 2014; USDA 2012). Which connotes they can be easily grown, even on marginal soils. However, millet storage protein have low digestibility compared to wheat and sorghum, even at elevated temperature (> 70 °C) (Annor, Tyl, Marcone, Ragaee, & Marti, 2017; Gulati et al., 2017). More so, millet flour forms a poor visco-elastic dough compared to wheat flour, due to discontinuous cross-linking behavior of its storage protein (Lorenz & Dilsaver, 1980; Taylor et al., 2016). This means that stable and expanded leavened dough cannot be formed from 100% millet flour; some parts of the millet flour will have to be substituted with wheat flour to achieve improved dough quality (Lorenz & Dilsaver, 1980; Taylor et al., 2016). Thus, to achieve millet storage proteins with better functionality in food application, structural modification of the proteins is essential. It is well established that the structural conformation of a protein dictates its functionality (Withana-Gamage & Wanasundara, 2012).

The structural modification of millet storage proteins among other gluten-free cereal storage proteins, for improved functionality and food application, has been a subject of ongoing research since the early 1990s (Lawton, 1992; Taylor et al., 2016). Several approaches such as physical, chemical, enzymatic, and genetic modification of gluten-free cereal storage proteins have been applied to alter the structural conformation of the storage proteins. Nazari and others (2018) applied high-intensity ultrasound to modify proso millet protein concentrate (MPC) using an optimized acoustic power intensity of 73.95 W/cm2 for 12.5 min and they reported a decrease in the molecular sizes of the treated MPC compared to untreated MPC which resulted in a significant increase in solubility, foaming capacity, emulsifying activity index (EAI) and emulsion stability index of the ultrasound treated MPC (Nazari, Mohammadifar, Shojaee-Aliabadi, Feizollahi, & Mirmoghtadaie, 2018). In another study by Sainani and others (1983), eight different amino acid residues

(histidine, tyrosine, methionine, cysteine, serine, lysine, glutamine, and tryptophan) of pennisetin, a major storage protein in pearl millet was chemically modified to alter the overall structure of the protein. The chemical modification includes treating the pennisetin separately with N-Bromosuccinimide (NBS), Diethylpyrocarbonate (DEPC), phenyl- methylsulphonyl fluoride (PMSF), 2,3-Butanedione, Iodoacetamide, N-ethyl-5-phenyl isooxazolium 3´-sulphonate, N-Acetylimidazole and 2,4,6-Trinitrobenzene sulphonic acid (TNBS). Their result showed that modification of serine led to a decreased intrinsic viscosity from 16.8 to 14.9 mLg-1, while the modification of histidine, tyrosine, methionine, cysteine led to increase in intrinsic viscosity from 16.8 to 20.5 mlg-1 (Sainani et al., 1993). Enzymatic modification of foxtail millet protein via hydrolysis was reported to increase the solubility, foaming, and emulsion properties and digestibility of the proteins compared to untreated ones (Kamara, Amadou, Tarawalie, & Huiming, 2010). Not much information is found on genetic engineering for the modification of millet storage protein; however, a few research works on genetic engineering of other gluten-free cereal storage proteins is well discussed by Taylor and others (Taylor et al., 2016). In all, these approaches measures extrinsic responses in the protein functionality as a result of applied modification stress without much recourse to how this stresses transcends to the molecular level or the understanding of how this stresses cause the perturbation and severance of the intra and intermolecular bonding that modifies the secondary and tertiary structure of the protein and consequently its functionality.

Molecular dynamic modeling (MD) is one approach that offers this possibility to study and understand the structural modification at the molecular level as it relates to its functionality. Molecular dynamic modeling is well utilized in the field of molecular and structural biology, pharmaceutics, medicine, and chemistry, to develop de-novo protein, engineer new protein folds, modify protein conformations and study protein-protein interactions (Kortemme et al., 2004; Lippow & Tidor, 2007; Borgo & Havranek, 2012). The use of MD in studying the modification that could occur to the structure of cereal storage protein is a gray area, yet to be adequately explored. One of the challenges of using MD in understudying structural modifications of food proteins is the limited availability of the three-dimensional structure of most food proteins, including proso millet. Molecular

3 dynamic modeling relies on the determination of the three-dimensional structure of a protein, which is sequestered to the simulation environment, in silico.

Thus, the overarching goal of this dissertation research was in two-fold. First, is to extract, identify, characterize, non-thermally modify the storage proteins of two proso millet cultivars (Dawn and Plateau) for improved functionality in food application. The second goal was to determine the three- dimensional structure of the proso millet protein and investigate its behavior at the molecular level using a molecular dynamics simulation approach. Generally, in protein chemistry, the storage protein in plants are grouped along their solubilities in different solvent as albumins (water-soluble), globulin (dilute saline soluble), prolamin (soluble in aqueous alcohol) and glutelin (soluble in dilute alkaline and acid) fractions. Glutelin fractions are the predominant protein faction in protein concentrates from cereal grains. In this project, two proso millet cultivars, the non-waxy and waxy variety documented to have the highest protein content (15.14±0.01 and 14.79±0.02% dry basis, respectively) were used (Singh, Adedeji, & Santra, 2018). The Dawn is the most widely grown proso millet cultivar in the US, and Plateau is a newly developed waxy cultivar just released for market assessment in 2014 (Santra et al., 2014). For the MD simulation part, glutelin type-B 5-like, a type of glutelin protein fraction previously identified (Shi et al., 2019) was used as a model representative of proso millet protein. Thus, the specific objective of this dissertation research work was to (1) identity, extract, and functionally characterize the various protein fractions (albumin, globulin, prolamin, and glutelin) from the two proso millet cultivar, (2) modify the functionality of the major protein fractions (prolamin and glutelin) of the two cultivars of the proso millet using high power ultrasound of constant frequency of 20kHz at three levels of amplitude 50%, 75%, and 100% (3) determine the crystal structure of glutelin type-B 5-like subunit of glutelin fraction, and (4) evaluate the structural conformation of the glutelin type-B 5- like subunit of glutelin fraction during the application of simulated processing stresses (temperature, pressure, and static electric field) using MD.

1.1 Rationale In line with the quest to develop additional plant-based protein ingredients to meet the protein demand gap, which is projected to increase by about one-third of the current protein

4 demand by the year 2050 (Henchion et al., 2017). It has become imperative to not only increase plant productivity but to ensure that the available ones are well developed into protein ingredients with enhanced functionality for better food application. Thus, the research into proso millet protein seeks to address this protein challenge in a twofold objective: (1) to valorize the proso millet protein ingredients by elucidating its nutritional, functional, and structural information that will become more useful in the food industry, especially for consumers interested in the gluten-free product (2) to provide fundamental knowledge on the behavior of the protein during application of processing stresses at molecular that could become useful at the downstream macroscale food application.

1.2 Project Goal The main goal of this project was, to modify the structure of proteins from two proso millet cultivars (Plateau and Dawn) for enhanced functionality in applications as food ingredients by applying both computer-simulated and un-simulated conventional non-thermal processing techniques.

1.2.1 Specific objectives The specific objectives of the research are to:

Determine the physicochemical, functional, and structural properties of proso millet storage protein fractions extracted and isolated from waxy (Plateau) and non-waxy (Dawn) cultivars.

Determine the effects of ultra-sonication on the in-vitro digestibility, physicochemical, and functional properties of the major proso millet protein fractions (prolamin and glutelin) from Plateau and Dawn cultivars.

Determine the three-dimensional crystal structure of the glutelin type-B 5-like subunit of glutelin fraction using either X-ray crystallography or homology modeling approach.

Determine the effects of processing stresses (temperature – 300K, 350K, and 400K; static electric field –0.1, 1, and 3 V/nm) on the root mean square deviation (RMSD), root mean square fluctuations (RMSF), the radius of gyration (Rg), solvent accessibility surface area (SASA), volume and secondary structure changes of glutelin type-B 5-like in a molecular dynamic modeling simulation environment.

Determine the effects of processing stresses (temperature – 300K, 350K, and 400K, pressure – 1 bar (0.1 Mpa), 3 kbar (300 Mpa) and 6 kbar (600 Mpa)) on the root mean square deviation (RMSD), root mean square fluctuations (RMSF), a radius of gyration (Rg), solvent accessibility surface area (SASA), volume and secondary structure changes of glutelin type-B 5-like in a molecular dynamic modeling simulation environment.

1.2.2 Hypotheses Proso millet proteins are not widely used in the food industry. Various reports have ascribed this reason to its low digestibility and poor physicochemical functional properties. However, there is a dearth of information on protein functionality, structure, or conformational changes with processing stresses. Thus, we seek to fill this gap in knowledge by testing the following hypothesis:

There are differences in the physicochemical properties (denaturation temperature, surface hydrophobicity, and solubility), functional properties (emulsion and foaming), and structural particulars between the protein fractions from Dawn and Plateau proso millet cultivars. The effect of different levels of ultrasound treatments (50, 75, and 100% amplitude for 5 and 10 min) will increase the digestibility, increase the physicochemical properties (denaturation temperature and solubility), and increase the functional properties (emulsion and foaming) of prolamin and glutelin fractions from Dawn and Plateau cultivars proso millet cultivars. The three-dimensional crystal structure of the glutelin type-B 5-like will be homologous to the crystal structure of globulins from other legumes. Simulated processing stresses will unfold the glutelin type-B 5-like secondary structure and increase the RMSD, RMSF, Rg, SASA, and volume.

Chapter II: Literature Review 2.0 Introduction Protein ingredients are widely used in the food industry for their nutritional benefits and their functionality in food formulations. The amphipathic nature of proteins makes them versatile structural components in food matrices because they act as dynamic surface-active agents for interfacial protein-to-carbohydrate, protein-to-lipid, protein-to-air, and protein- to-water interactions. The recent and continuing growth in popularity of plant proteins has been quite significant; this trend is propelled by consumer’s increasing knowledge and awareness of food ingredients, their desire for clean label and lean protein as well as growing interest in natural, eco-friendly and sustainable food sources. The global plant protein product sales were estimated at $35 billion in 2018 and projected to be $45 billion by 2023 at a compound annual growth rate (CAGR) of 4.93% over a forecasted period from 2018-2023 (Cassity, 2019).

From a nutritional perspective, numerous research works have highlighted the benefits of proteins for overall good health and wellness, to include weight management, satiety, reduced glycemic index, heart health, muscle maintenance and even sports performance (Livesey, Taylor, Hulshof, & Howlett, 2008; Paul, 2009; Phillips, 2004; Soenen & Westerterp-Plantenga, 2008). It is reported that about 60% of Americans consider proteinaceous foods when making dietary choices (Cheatham, 2013). Plant proteins have hitherto been considered to be nutritionally inadequate because they are deficient in some essential amino acids. For example, cereal and pulse proteins are deficient in lysine and methionine, respectively (Sun-Waterhouse, Zhao, & Waterhouse, 2014). Besides, plant proteins are gastro-intestinally less bioavailable when compared to animal proteins (Joshi, Shah, & Kalantar-Zadeh, 2018). However, recent studies in the field of medicine have proven that differences between plant and animal proteins are clinically insignificant (Joshi et al., 2018), especially if a variety of plant-based diets is consumed. For example, a right combination of plant-based diets (e.g. cereal + pulses) will provide a complementary mix of the essential amino acids needed for body maintenance.

While the nutritional appeal of plant proteins is important, of equal importance is their functionality, which facilitates use as ingredients in food formulation to confer appearance,

7 flavor, color, odor, texture, and even structure to food products. The term “protein functionality” can be ambiguous depending on the scale and area of definition. However, protein functionality is defined in terms of how proteins behave as a colloidal biopolymer with or without other food ingredients in a food system, based on their intrinsic and/or extrinsic physicochemical properties but not necessarily their biological roles/functions or nutritional bioactivities (Foegeding, 2015; Foegeding & Davis, 2011). Some of the common functional properties of proteins include solubility, gelation, emulsification, foam formation, water, and fat binding, viscosity, film formation, and others (Table 2.1). All these functionalities are governed by the molecular properties of the proteins (surface hydrophobicity, surface topology, molecular weight, isoelectric point, secondary structures, and tertiary structures) in their native, intermediate or denatured states (Foegeding & Davis, 2011; Kinsella & Melachouris, 1976). The structural changes, especially as it relates to the local environment (pH, temperature, and ionic strength of solution) influence protein functionality and the quality of the food products in addition to the effects on nutritional properties and gastrointestinal availability.

Table 2.1. General classes of functional properties of proteins important in food applications*.

General Property Specific functional term

Organoleptic/Kinesthetic Color, flavor, odor, texture, mouthfeel, smoothness, grittiness, turbidity, etc.

Hydration Solubility, dispersibility, wettability, water absorption, swelling, thickening, gelling, water holding capacity, syneresis, viscosity, dough formation, etc.

Surface Emulsion, foaming, aeration, whipping, protein/lipid film formation, lipid binding, flavor binding, stabilization, etc.

Structural /Textural/Rheological Elasticity, grittiness, cohesion, chewiness, viscosity, adhesion, network cross-binding, aggregation, stickiness, gelation, dough formation, texturability, fiber formation, extrudability, etc.

Other Compatibility with additives, enzymatic, inertness, modification properties

*Adapted from (Kinsella & Melachouris, 1976).

Plant protein ingredients are currently widely used in the food industry owing to consumers' positive reaction to plant-based products. For instance, some vegetable proteins have been used to replace meat (e.g. meat analogs), egg, milk (soy milk, almond milk) and other dairy products (vegan cheese) (Bergsma, 2019; Lipan et al., 2020; Schreuders et al., 2019). Plant protein ingredients such as isolates and concentrates are needed for specific functions during food product development and as such, it is important to understand how their functionality supports product formulation and quality. Knowledge of how food processing conditions modulate individual or food matrix protein functionality is also of critical

9 importance for designing new ingredients. For example, the solubility and viscosity of a plant-based protein are important for beverages (Sethi, Tyagi, & Anurag, 2016); water holding and fat binding properties are desired in dough making (Villarino, Jayasena, Coorey, Chakrabarti-Bell, & Johnson, 2016); emulsification properties are important in coffee whiteners (soy creamer); and foaming properties are useful in whip toppings (Lu, He, Zhang, Bing, & nutrition, 2019). Some plant protein ingredients used for their functionalities in different products are shown in Table 2.2

Table 2.2 Summary of common plant-based protein ingredients and their functionality in food application

Protein Ingredients Functional property Application Sources examples

Legumes

Soybean

Flour/grits (defatted) Emulsification, fat absorption, Frankfurters, (Cho, Jung, Auh, & Lee, viscosity, and water absorption sausages, cakes 2017; Farzana, Mohajan, bologna, meat Saha, Hossain, & Haque, 11 Concentrates Emulsification, fat absorption,

patties, pancakes, 2017; Krintiras, Diaz, Van viscosity, and water absorption bread, doughnuts, Der Goot, Stankiewicz, & Isolates Emulsification, fat absorption, gravies and soups Stefanidis, 2016; Nguyen, viscosity water absorption, Bhandari, Cichero, & gelation, film formation, and Prakash, 2015; Singh, aeration Kumar, Sabapathy, & Bawa, Hydrolysate Whip ability, and water Soymilk and baby 2008; Taghdir et al., 2017) absorption foods

Texturized soy protein Texture, meat extenders, water, Meat analogs, and fat binders meatballs, beef

patties hamburgers, sausages, pizza toppings, etc.

Pulses (pea, bean, and lentil)

Flour Whippability, water, and fat Bean curd, nuggets (Boye, Zare, & Pletch, 2010; binding, emulsion, and foaming, sausages, Gumus, Decker, & Concentrates gelling and texture meatballs, and McClements, 2017) Isolates cake doughnut

12 Cereals

Wheat gluten (including texturized Visco-elasticity, texture, adhesive Bakery, hams and gluten) and film formation, hydrophobic turkey rolls, pizza

properties (water insolubility) toppings sausages, whipping agents, gluten films, (Dacey & O'connor, 2016; adhesives, Elzoghby, Samy, & Elgindy, plasticizer, 2012; Pereira, Gh, Zhang, & protein-based Research, 2016; Pietsch, nanocarriers, meat

replacement, and Emin, & Schuchmann, fish feeds 2017; Wang, Gulati, Santra, Rose, & Zhang, 2018; Corn zein adhesive and film formation, Zein coating, Wouters, Rombouts, hydrophobic properties protein-based Fierens, Brijs, & Delcour, nanocarriers 2018; Xiao, Wang, Other prolamins hydrophobic properties Protein-based Gonzalez, & Huang, 2016). nanocarriers

Cereal concentrates/isolates Water and oil binding, texture Bakery products

Oilseeds

(Sunflower, peanut, rapeseed, and flaxseed)

Flour Oil absorption, emulsification, Meat emulsion, (Arntfield, Murphy, & Ross, foaming properties, whipping (if sausages, tofu, ice 2018; Grasso, Pintado, chlorogenic acid is removed) cream, bakery Pérez-Jiménez, Ruiz- products, gravy Capillas, & Herrero, 2020; Concentrates/Isolates/hydrolysates Water absorption, high solubility, Mohammed et al., 2018; fat absorption, oil emulsification, Yoshie-Stark, Wada, & whippability, and foam stability Wäsche, 2008) (FS).

While there is a burgeoning interest in plant protein ingredients, there are functionality challenges with some of these proteins that limit their replacement of animal proteins in food formulations. For instance, most cereals proteins have low solubility at neutral pH and most pulse proteins form weak gels (Kyriakopoulou, Dekkers, & van der Goot, 2019; Wouters & Delcour, 2019). The structure-function relationship of a protein is key to understanding the mechanism of protein functionality (Creighton, 1993; Fukuda, Maruyama, Salleh, Mikami, & Utsumi, 2008). Not only do we need to know structural conformations of protein in their native states but how these conformations change as the protein unfolds during denaturation and under different processing conditions. Fukuda et al., (2008) did a comparative study of the 7S globulin proteins from Adzuki beans (7S1, 7S3), soybeans (β-conglycinin β, β-conglycinin α’c), and mung beans (8Sα) by comparing their three-dimensional structures to understand the differences in their thermal stability and solubility. Their results showed that the presence of a cavity inside the molecules of each of these proteins imparts thermal stability.

The general strategy for the modification of plant protein functionality involves the application of physical, chemical, and enzymatic stresses/forces or a combination of these stresses at the molecular, meso- and macro-scale of protein ingredient development. Such stresses include temperature, pressure, shear, freezing and thawing, electric field, electromagnetic field, surface tensions, hydration, and solvent force. A summary of the common modification processes associated with plant proteins is presented in Table 2.3 Such stress-induced perturbation will cause a change in the thermodynamic state of the protein, including its structural and conformational characteristics. For example, a modification could change the size, surface charge, hydrophobic/hydrophilic ratio and molecular flexibility (Phillips, 2013). Overall, the modification could improve or create an entirely new protein functionality (Sun-Waterhouse et al., 2014).

Table 2.3. Techniques for the modification of plant-based proteins for improved functionality*

Modification Examples of processing techniques Approaches

Physical Isolation/enrichment, thermally induced denaturation, ultrasound-induced shearing, ultra-high-pressure homogenization, microwave heating, gamma radiation, extrusion, cold plasma, pulsed UV light, pulsed electric field, and tribomechanical activation,

Chemical Glycosylation, Acylation, phosphorylation, and deamidation

Enzymatic Hydrolysis, and cross-linking

Genetic Recombinant DNA engineering, site-directed mutagenesis

*Adapted from (Phillips, 2013)

Therefore, this review explores research progress on protein modification for improved functionality with a focus on the use of physical, chemical, and enzymatic processes for both common and emerging plant proteins.

2.1 Modification of Plant Proteins by Various Processing Techniques 2.1.1 Mechanism of protein modification It is difficult to propose a general model for the modification of various food proteins including those from plants. Proteins are complex biomolecules and their modification is not a “one size fits all” approach. Depending on the functionality of interest, it is important to know the nature (whether in isolation or bound with other polymers) and structural properties of the protein ingredient as well the mechanism required to achieve a modified target functional property. Thus, it becomes simple to develop an approach tailored to addressing such functional property bearing in mind the structural properties of the protein and the mechanism needed to achieve such modification. The mechanisms to achieve some functionalities are well understood. For example, the mechanism to cause an improvement in the solubility of protein ingredients may require size reduction, electrostatic repulsion,

and an increase in hydrophilicity by conjugation with a more hydrophilic polymer, manipulating the protein isoelectric point and net charge. In short, any approach (including pH shifting, homogenization, ultrasonication, proteolysis, hydrolysis, conjugation, glycosylation, acylation, and esterification) that could cause these changes to the protein molecules would most likely lead to increased solubility. Similarly, increasing the emulsifying and foaming activities of a protein may require mechanisms that promote a balance of its hydrophobic-hydrophilic property in addition to its solubility. Other mechanisms include conjugation with other polymers, disruption of agglomeration using ultrasonication, and modifications that reveal the hydrophobic core of the proteins. Therefore, this section discusses the state-of-the-art technologies and processes used to modify various plant-derived food proteins based on a broad mechanism of operation.

2.1.2 Physical techniques for modification of plant-based proteins

Protein modification approaches that involve the application of some force fields to change protein structure whether in isolation or as part of a food matrix can be classified as physical methods. Generally, these techniques lead to protein size reduction and size redistribution, unfolding, agglomeration, dis-aggregation, or permanent denaturation of the protein conformation. The following examples illustrate commonly used physical techniques.

2.1.2.1 Heat treatment

Heat treatment is one of the widely used processing techniques for modifying proteinaceous foods (Sharif et al., 2018). Heat energy is the primary processing stress in cooking, roasting, drying, and high-temperature extrusion. In summary, heat treatment causes thermal mobility of the peptide chains which could lead to a severance of the inter and intramolecular hydrophobic interactions in addition to electrostatic, hydrogen and disulfide bonds when performed at above the protein denaturation; as such there is an initial reversible unfolding but then the unfolding becomes permanent especially at higher temperatures. This permanent denaturation is accompanied by the loss of secondary and tertiary structures of the protein molecules (Davis & Williams, 1998; Sun-Waterhouse et al., 2014; Zink, Wyrobnik, Prinz, & Schmid, 2016). Usually, when this happens, the hydrophobic core of the protein is exposed and there is a new augmented crosslinking via

hydrophobic interactions, hydrogen, and disulfide/sulfhydryl bonding (Mession, Chihi, Sok, & Saurel, 2015; Zink et al., 2016). Therefore, controlled application of heat could be used to modify the structure of food protein ingredients and consequently to improve some functional properties (Davis & Williams, 1998; Peng et al., 2016; Sun-Waterhouse et al., 2014). These functionalities are dependent on a careful combination of the heating temperature, heating rate, ionic concentration, and pH of the protein solution (Zink et al., 2016). It is well documented that heat treatment of certain plant proteins can improve their protein gelation (Sun, & Arntfield, 2011), emulsifying properties (Peng et al., 2016), and digestibility (Rehman & Shah, 2005); but not protein solubility and no conclusive evidence on the positive effects of thermal treatment on allergenicity reduction (Davis, Smales, & James, 2001; Mondoulet et al., 2005; Nowak-Wegrzyn & Fiocchi, 2009; Sun-Waterhouse et al., 2014).

350 raw 300 250 200 150

supernatant 100

50 μg lowry protein / ml ofml / protein μglowry 0 0 2 4 6 8 10 12 14 pH

Figure 2.1. Protein solubility profile of cowpea flour as a function of pH (adapted from (Abbey & Ibeh, 1988).

2.1.2.1.1 Effect of heat treatment on the functionality of plant protein ingredients

The effects of heat treatment on the functional properties of various plant protein ingredients have been well documented in the literature. Ma et al. (2011) showed that the protein solubility of pulse flours was reduced after thermal treatment (roasting or boiling) while other authors showed a similar trend for cowpea flour (Fig. 2.1) (Abbey & Ibeh,

1988), soy and peanut flours (McWatters & Holmes, 1979) as well as other legumes protein isolates/concentrates (Ghribi et al., 2015; Shand, Ya, Pietrasik, & Wanasundara, 2007). They all attributed the reduction in solubility of heated legume protein ingredients to denaturation upon heating, which also explains the reduced solubility of commercially processed protein isolates/concentrate when compared to those extracted in the laboratory. A significant increase in the fat binding, water holding capacity and emulsifying activity index of the pulse flours were reported following thermal treatment (Ma et al., 2011). Similarly, the thermal treatment of soybean protein isolate (SPI) and pea protein isolate (PPI) at 95°C for 15 and 30 min, respectively led to increased emulsifying properties due to hydrophobic aggregation and ease of diffusion of the protein molecules to the oil-water interface (Peng et al., 2016; Shao & Tang, 2014). In addition to the compositional heterogeneity and concentration of a protein ingredient, salt concentration, pH, heating temperature and heating rate are among the factors that control gelation properties of most heat-induced legume proteins. (Zheng, Matsumura, & Mori, 1993) revealed that heating broad bean legumin below the onset temperature of denaturation did not result in gel formation while heating to the maximum denaturation and above the final denaturation temperature produced gels. However, the hardest gel was formed when the legumin protein was heated between the maximum and final denaturation temperature while gel hardness decreased with increasing temperature after heating beyond the final denaturation temperature. In a separate study, O’kane et al. (2005) observed no difference in the gelling properties of PPI (minimum concentration 18% w/v at pH 7.6) from five pea cultivars at a heating rate of 1.0°C/min and 0.5°C/min; but they reported a considerable increase in the gel strength when cooling at a lower rate of 0.2°C/min compared to 1.0°C/ min (O'Kane, Vereijken, Gruppen, & Van Boekel, 2005). In a similar but separate study, O’Kane et al. (2004) observed a similar trend for the effects of heating/cooling rate on the gelling behavior of pea legumin and soy glycinin. There was no effect of heating rate (1.0°C/min and 0.5°C/min) on gel strength but a significant increase in gel strength was achieved when cooling was done at 0.2°C/min compared to 1.0°C/min (O'Kane, Happe, Vereijken, Gruppen, & van Boekel, 2004). Campbell et al. (2009) reported that heated SPI used in an acid-induced gel showed a higher gel strength that unheated SPI. The effects of thermal treatment to modify cereal protein functionality are well documented in the literature.

Similar to legume seed proteins, thermal treatment has been reported to reduce protein solubility of wheat flours at different (50-90°C) temperatures, particularly by 38.9% at 90℃ for 30 min (Mann, Schiedt, Baumann, Conde-Petit, & Vilgis, 2014). Corn flour subjected to radiofrequency heating was reported to show a decrease in protein solubility (26%) at a temperature beyond 60°C (Hassan, Pawelzik, & von Hoersten, 2016). At a steam treatment temperature of 120°C for 15 min, random-coil amorphous zein film was converted to β-crystal zein film (Magoshi, Nakamura, & Murakami, 1992). Depending on the forces (hydrophobic interaction, ionic, disulfide and hydrogen bond) that contribute to gel formation in the cereal proteins, high temperature will favor gel formation where hydrophobic interactions are dominant contrary to hydrogen bonds. Corn germ protein gel was stabilized by hydrophobic interactions and by disulfide bridges at a temperature of 87°C and above but by predominantly hydrogen bonds when cooled below 87°C (Sun et al., 2015). As for wheat gluten gel, the contribution of hydrogen and ionic bonds on the gel formation decreased when protein concentration ranged from 4.133 to 2.733 g/L and 0.746 to 0.397 g/L as the heating temperature increased from 25 to 90°C while hydrophobic interactions increased at temperatures beyond 70°C, which contributed to gluten gel stability (Wang et al., 2017). Oat globulin also produced stable gels when heated to temperatures above 90°C (Ma, Khanzada, & Harwalkar, 1988).

2.1.2.2 Ultrasound treatment

The effect of ultrasound (US) treatment on human tissues has been well known in the field of medicine since the early works of Theodore Dussik and his brother Friederich in the 1930s and 1940s (Newman & Rozycki, 1998). However, US application to food materials started to emerge in the 1950s mostly for food tenderizing (Meehan, 1952; Simjian, 1959). US technology (Fig. 2.2) is a propagated acoustic wave beyond the threshold of human hearing (>16 kHz), such that it causes a longitudinal displacement of the medium in its parts leading to compression and rarefaction of that medium (O’Sullivan, Park, Beevers, Greenwood, & Norton, 2017). Ultrasound can be divided into two frequency categories: low-frequency US (16-100 kHz, power 10-1000 Wcm-2) commonly used for the physical and chemical modification of proteins and high-frequency US (100 kHz – 1 MHz, power <1 Wcm-2), commonly used for the evaluation of the physicochemical properties of foods

(Jiang et al., 2014; Nazari et al., 2018; O’Sullivan et al., 2017). Ultrasound treatment modifies protein functionality majorly through localized hydrodynamic shearing and heating of the protein molecules in solution and consequentially modifying its structure (O’Sullivan et al., 2017). The hydrodynamic shearing is the result of ultrasonic cavitation produced by the US sonotrode. Ultrasonic cavitation is characterized by rapid build-up and collapse of gas bubbles, generated by localized pressure differentials in wave propagation over a short period (O’Sullivan et al., 2017).

Figure 2.2. Schematic of a typical Ultrasound system

Review articles on the effects of US treatment on protein structural modification of animal and vegetable proteins have been published (Higuera-Barraza, Del Toro-Sanchez, Ruiz- Cruz, & Márquez-Ríos, 2016; O’Sullivan et al., 2017). Here we expound on the effects of US on the functionality of other vegetable proteins not covered in previous reviews.

2.1.2.2.1 Effects of ultrasound treatment on plant protein functionality

The only well-known mechanisms by which high power US waves modify protein structure is through hydrodynamic shearing of the protein particles in the native form into

small nanosized and well-dispersed particles, as well as simultaneous heating that causes thermal degradation of the protein molecules in some cases (O’Sullivan et al., 2017). Usually, the shearing does not affect the primary structure or molecular weight of the protein but the size of the protein particles and its distribution (Jiang et al., 2014; O’Sullivan et al., 2017; Zhu et al., 2018). In some cases, it is possible to observe new forms of aggregation with a particle size that can be greater than the untreated samples after US treatment, especially with low power or prolonged (> 20 min) treatment (Jiang et al., 2014; O’Sullivan et al., 2017; Ren et al., 2015). A study by O’Sullivan et al. (2016) showed that the average particle size (52800 ± 840 nm) of US treated rice protein isolate (RPI) was higher than the native RPI (51600 ± 920 nm). However, for an appropriate high power US treatment, reductions in the native protein size are due to the severance of non-covalent interactions (such as hydrophobic, electrostatic, and hydrogen forces) that favor protein aggregation (O’Sullivan et al., 2017). These shearing effect together with the localized heating causes partial unfolding of the proteins and the revelation of some hydrophobic residues that may promote new forms of aggregation in certain circumstances (Nazari, Mohammadifar, Shojaee-Aliabadi, & Mirmoghtadaie, 2016; Wen et al., 2019). Additionally, there is an accompanying reduction in pH, an increase in electrical conductivity and some cases the formation of free radicals (Jambrak, Mason, Lelas, Herceg, & Herceg, 2008) for most but not all plant proteins. No significant reduction in pH was observed for US treated RPI (O'sullivan, Murray, Flynn, & Norton, 2016).

Several studies have reported on the positive effects of US treatment on the functionality of most plant protein ingredients. Taha et al. (2018) showed that high-intensity US-assisted (50-55 Wcm-2, 20 kHz, 40% amplitude and on-time 2 s + oﬀ-time 2 s) formation of SPI oil-in-water emulsions for 12 and 18 mins was well dispersed and stable compared to the emulsions made with a high-shear probe homogenizer (19000 rpm) for 6 min. They reported a d4,3 of 0.6 ± 0.1, 0.5 ± 0.0 and 16.1 ± 3.7 µm particle size for 12 and 18 min US treatment and 6 min homogenization, respectively for the medium-chain triglycerides oil used (Taha et al., 2018). Similarly, high-intensity US (HUS) (20 kHz at 400 W for 5, 20 or 40 min was observed to increase the solubility, emulsifying activities, emulsion stability (ES) and surface hydrophobicity of soybean β-conglycinin (7S) and glycinin (11S) fractions (Hu, Cheung, Pan, & Li-Chan, 2015). Additional US studies have shown

improvement in the functionality of leguminous proteins such as increased emulsion performance of PPI treated at ∼34 W cm−2 for 2 min (O'sullivan et al., 2016); increase in the solubility of black bean protein isolates after treatment for 12 and 24 min at 150 W, 300 W and 450 W (Jiang et al., 2014); and improved foam properties and solubility of faba bean after treatment for 17.29 min at an amplitude of 72.6% (Martínez-Velasco et al., 2018). Other cereal proteins have shown improved functionality following US treatment. Zhang et al. (2011) showed increases in the FC (~ 72, 132, 150, and 162%), FS ( ~ 40, 56, 76, and 84% after 60 min.), EAI (~ 36, 54, 60, and 78 m2/g), and ES (~ 12, 24, 30, 36) properties of wheat gluten with increasing US treatment power (0, 540, 720, and 900 W) for 10 min treatment (Zhang, Claver, Zhu, & Zhou, 2011). Similarly, HUS increased the solubility (~ 65, 78, 84, 90 %) , FC (271.03, 148.37, 435.37, and 716.03 ml) FS ( 4.37, 18.37, 10.70, and 25.7 ml after 10 min), EAI (27.92, 39.17, 45.83, and 52.07%) and ES (10.97, 22.42, 33.97, and 41.19%) of millet protein concentrates at treatment conditions of 0, 18.4, 29.58, and 73.95 W/cm2 for 20 min, respectively (Nazari et al., 2018).

2.1.2.3 High-pressure treatment

High static pressure between 200 – 700 MPa has been widely applied to modify some plant proteins (Farkas, 2016; Messens, Van Camp, & Huyghebaert, 1997). The level of applied pressure, duration of applied pressure, the temperature at treatment, and the condition of the protein solution (pH and ionic strength) combine to induce structural changes that subsequently affect protein functionality (especially gelation) during the high-pressure processing (HPP). The mechanisms of protein modification through high-pressure treatment are based on the ability to induce a volume change in the protein molecules in solution that leads to rupturing, denaturation, and aggregation of the protein molecules. This idea is based on the principle of Le Chateliers, that relates pressure to volume in an inverse relationship, i.e. decrease in volume is promoted by increased pressure and vice versa (Messens et al., 1997). The volume of a protein in solution consists of the volume of its atoms, the volume of its cavities from imperfect folding, and the volume change associated with its interaction with the dissolution solvents (Yang & Powers, 2016). Thus, application of high pressure leads to compression of the protein cavities, rupturing of the non-covalent interactions (intramolecular hydrophobic and electrostatic interactions) as

well as the formation of new non-covalent/semi-covalent associations (Galazka, Dickinson, & Ledward, 2000; Messens et al., 1997; Yang & Powers, 2016). However, hydrogen bonds are insignificantly impacted except at extremely high pressures (>1000 MPa).

2.1.2.3.1Effects of high-pressure treatment on plant protein functionality

High pressure does not affect covalent bonds so the primary structure of the proteins is intact, but does affect the secondary, tertiary, and quaternary structures of the protein (Yang & Powers, 2016). The effects of HPP on protein functionalities vary with the type of plant-based protein. Some studies have confirmed that HPP reduced the solubility (~2.5% less than control at 600 Mpa) of soybean protein isolate (SPI) or have an insignificant effect on its solubility compared to the native protein within HPP range of 200 – 600 MPa for 1-5 % (w/w) SPI concentrations at pH 6-8 under room temperature, but increases the surface hydrophobicity (about 32% more) with the level of applied pressure (Floury, Desrumaux, & Legrand, 2002; Puppo et al., 2004; Wang et al., 2008). Similarly, HPP was observed to cause a reduced solubility (~3.4%) at pH 3-9 for 1% (w/v) of PPI and cumin protein isolate (Chao, Jung, & Aluko, 2018; Chen, Mu, Zhang, & Goffin, 2019), increased surface hydrophobicity (about four-fold) in rapeseed protein isolate and sweet potato protein (He, He, Chao, Ju, & Aluko, 2014; Zhao, Mu, Zhang, & Richel, 2018). HPP has been well reported to affect the gelation and rheology properties of many plant-based proteins. Sim et al. (2019) established gel formation at 16 g protein/100 g at 250 MPa for PPI and observed an increased gel strength and rheology (1.6 Pa for untreated and 14432 Pa for 15 min at 550Mpa). with increased pressure level for a 24 g/100 g concentration as shown in Figure 2.3 compared to 250 MPa, HPP treatment at 550 MPa led to the formation of gels with thicker and more defined network structure comprised of fibrillar aggregates (Fig. 2.3). The storage modulus of PPI gel (16g protein/100g) increased about two-fold while that of PPI gel (20 g protein/100 g and 24 g protein/100 g) increased by four-fold for HPP treatment at 350 MPa and above. In another experiment, He et al. (2014) showed that HPP reduced the least gelation concentration of rapeseed isolate from 15 to 6% after 600 MPa and the hardness of the gel was consistently increased with an increase in HPP treatment level. Similarly, HPP treated SPI and its major globulins fractions (7S and 11S)

formed gel at 20% protein concentration within 300 – 700 MPa. The hardness of the gel increased with high-pressure levels (Molina, Defaye, & Ledward, 2002).

Figure 2.3. Scanning electron microscope micrographs of untreated, pressure-treated, and heat-treated 24 g/100 g pea protein concentrate solutions. Greater extents of aggregation and network are observed after treatment at higher pressures (Sim, Karwe, & Moraru, 2019).

Emulsifying activity index (EAI) and emulsion stability index (ESI) were reported to be increased by HPP treatments (200 - 600 MPa) for potato protein at pH 6-9 (Khan, Mu, Sun, Zhang, & Chen, 2015). The EAI of red kidney bean protein isolate, similar to its ESI increased significantly from 24.2 to 40.4 m2/g with increasing pressure from 0.101 to 400 MPa and then decreased to 33.9 m2/g with further increase in pressure to 600 MPa while it's foaming capacity (FC) and foam stability (FS) decreased with pressure levels (200 – 600 MPa) (Ahmed, Al-Ruwaih, Mulla, & Rahman, 2018). However, for 1-5% (w/v) SPI concentrations, the EAI was observed to increase after only 200 MPa treatment compared to untreated SPI but no significant increment after 200 MPa was observed while the ESI decreases consistently with HP treatments (200 – 600 Mpa) (Wang et al., 2008). The effects of HPP on the EAI and ESI have been attributed to the unfolding of the proteins, degree of

aggregation of the unfolded proteins and the level of molecular flexibility after the HPP treatments, which differ with different proteins and solution systems and may be responsible for the discrepancies in the emulsifying behavior of different HPP-treated proteins (Wang et al., 2008).

2.1.2.4 Extrusion cooking

Extrusion cooking is a thermo-mechanical process that combines high heat, high-shear, and high pressure to cause cooking, sterilization, drying, melting, conveying, kneading, puffing texturizing and forming of food product (Berk, 2013). In protein application, extrusion cooking is particularly applied for texturizing plant-based proteins (SPI, Wheat gluten, and PPI) known as textured vegetable proteins (TVP) for use as meat analogs (Chiang, Loveday, Hardacre, & Parker, 2019; Pietsch et al., 2017). The short time and high intense extrusion conditions impact the functionality of plant proteins such as solubility, texture, emulsion, and gelation properties. Most plant proteins are extruded (twin screw) at a cooking temperature (zone before the die) between 130 - 170℃, medium to high moisture content range of 30-70% (wet basis) and a screw speed of 150-160 rpm (Chen, Wei, & Zhang, 2011; Chiang et al., 2019). A cooling die is used in protein extrusion to avoid expansion of the extrudate following cooking (Chen et al., 2011; Samard, Gu, & Ryu, 2019). High moisture extrusion is preferable to produce TVP with meat-like characteristics while low moisture (< 30 %) extrusion have been reported to cause a harder and less soluble protein (Chen et al., 2011)

2.1.2.4.1Effects of extrusion cooking on plant protein functionality

During extrusion, the proteins are denatured, and the hydrophobic residues are revealed due to high shear and temperature. Each protein molecule unfolds and aligns in the direction of the flow of the material in the barrel towards the die-end to form a fibrous structure held by the formation of new intermolecular bonds and aggregations (hydrogen, disulfide, hydrophobic) (Samard et al., 2019)). High moisture protein extrudate matrices are stabilized by hydrogen and disulfide bonds while low-medium moisture extrudate (30- 40%) are stabilized by hydrophobic and disulfide bonds (Chiang et al., 2019; Lin, Huff, & Hsieh, 2000). The effects of cooking temperature, screw speed, feed rate, and moisture content combine to affect the functionality of plant proteins following extrusion cooking,

especially how the chemical bonds are modulated in meat analogs to confer the desired texture. Lin et al. (2000), observed a reduction in the solubility of SPI extrudate compared to raw SPI following a high moisture extrusion (60-70%) at 150 rpm screw speed at cooking temperature range of 137.8 - 160℃. They showed that only moisture content had a significant effect on the protein solubility and not cooking temperature; while the cooking temperature and moisture content had significant effects on the protein textural attributes. Texture profile analysis attributes were only impacted by the cooking temperature at low moisture (Lin et al., 2000). Similarly, the solubility of PPI was reduced by 50-90% compared to raw PPI following low moisture extrusion (26-35%), barrel temperature of 130 -170℃, and 400-700 rpm screw speed with expansion as the target objective (Beck, Knoerzer, & Arcot, 2017). The barrel temperature insignificantly affected the secondary structure of PPI, however, the extrusion process reduced the proportion of β-sheet and α- helical structures in the extrudates compared to the raw materials and increased the β-turns from 9.7 ± 0.01% in the raw material to up to 13.36 ± 0.13% in the extrudates at 400 rpm (Beck et al., 2017). Chen et al. (2018) reported that the hydrolysate of extruded peanut protein isolate showed higher emulsion properties than the hydrolysate of raw peanut protein isolate. They concluded from their experiment that low moisture (15%) extrusion conducted at 130℃ increased the degree of hydrolysis and solubility of hydrolysate compared to non-extruded peanut protein. The results were due to the ability of the extrusion process to cause expansion and pores in the extrudate, which later became reduced and extrudate hardened upon increasing the extrusion temperature to 160℃ (Chen, Chen, Yu, Wu, & Zhao, 2018).

2.1.2.5 Cold plasma technology Cold plasma technology (Fig. 2.4) is one of the emerging non-thermal technology being applied for food processing particularly for sterilization but also protein functionality modification. Cold plasma technology creates a state of matter that contains a cocktail of • • • • * * - - - + + reactive oxygen and nitrogen species (O , OH, N , HO2 , N2 , N , OH , O2 , O , O2 , N2 , + + N , NO, O , O3, and H2O2) and ultraviolet radiations generated when the energy supplied to a gaseous environment dissociates the gas molecular bonds into fully or partially ionized gases called plasma (Attri et al., 2015; Venkataratnam, Sarangapani, Cahill, & Ryan, 2019). The type of gas, energy discharge source (electrical, thermal, optical,

electromagnetic, etc) and type of electrode are the factors that influence properties of the generated plasma. Dong et al. (2017) showed that cold plasma treatment caused the depolymerization of zein molecules with a particle size that decreased with the applied voltages (50, 75, 100, and 125 V) over a 2 min treatment time.

Figure 2.4. Schematic of a cold plasma system, adapted from (Dong, Gao, Zhao, Li, & Chen, 2017)

2.1.2.5.1 Effects of cold plasma technology on plant protein functionality

The reaction between protein molecules and reactive species is one keyway by which cold plasma treatment modifies protein structure and subsequent functionality. It has been reported that the high energy of the plasma leads to the breakdown of covalent bonds within the protein molecules, and, the reactive species cause sulfur amino acid oxidation, which could lead to cleavage of disulfide (S-S) bonds. Cleavage of the S-S bonds could lead to the formation of SH or SO groups that can further distort the conformation of the protein and break the polypeptide (Dong et al., 2017). For example, Dong et al. (2017) showed that cold plasma treatment led to an increase in the SH group concentration up until the 75 V treatment before decreasing. The solubility of modified proteins increased from 0 to 75 V and then decreased. Cold plasma increased the solubility, water and fat binding capacity of protein-rich pea flour and led to the structural modification of protein structure as indicated

by the increase in fluorescence emission intensity (Bußler, Steins, Ehlbeck, & Schlüter, 2015). In another experiment in which cold plasma treatment of 35 V and 2 ± 0.2 A for 1, 2, 3, and 4 min was applied to peanut protein isolate, the results showed increase in solubility up till 3 min treatment with accompanying decrease in pH (from 6.92 to 6.80) while the formed emulsion containing 75% oil was stable for up to 7 hr for all treated samples but was best for the 2 min treatment (Ji et al., 2018).

2.1.3 Chemically modified proteins Chemical modifications are obtained from protein reactions with chemical agents in which there is breaking or forming of new bonds that alter the integrity of the original protein structures. Usually, this approach exploits the reactivity of the protein side chains (such as amino, carboxyl, disulfide/sulfhydryl, imidazole, indole, phenolic, and thioester) in chemical reaction to modulate protein biophysical properties and functionality. The objective is to change the net charge on the protein by substituting the ɛ-amino group or other amino and hydroxyl groups of some amino acid residues (Panyam & Kilara, 1996). Chemically modified proteins have reportedly shown improved functionality compared to native molecules. However, the commercialization of chemical modification techniques is limited by the production of toxic chemical by-products, cost, consumers, and regulatory concerns (Sun-Waterhouse et al., 2014; Zhang et al., 2018). Some examples of chemical modification techniques widely used for plant proteins are discussed below.

2.1.3.1 Chemical glycosylation Glycosylation is a protein modification technique that involves the attachment of carbohydrate moieties to mainly the amino acid side chain lysine residues or the N- terminus of protein molecules. This reaction usually occurs during Maillard reaction and involves the formation of a covalent bond between a ɛ-amino group of lysine, guanidino group of arginine, thiol group of cysteine, imidazole group of histidine, the indole group of tryptophan or any N-terminal amino group of amino acids and the carbonyl group of any reducing sugar (mono-, di- or polysaccharides) to form stable protein-polysaccharide conjugates or other glycoconjugates as illustrated in Figure 2.5 (Akıllıoğlu & Gökmen, 2016; Oliver, Melton, & Stanley, 2006; Zhang et al., 2018). The Maillard reaction, however, is a complex non-enzymatic browning known to occur during food processing

such as drying, cooking, roasting, microwave heating or during any dry or wet heating with optimum reaction at a water activity between 0.5-0.8 (Bielikowicz et al., 2012; Oliver et al., 2006). Maillard's reaction is divided into three stages: the early, intermediate, and final stages. The early stage is usually regarded as glycosylation, which involves the initial condensation between an amino group and a carbonyl-containing compound to produce a Schiff base followed by irreversible Amadori rearrangement to form ketoamine (Akıllıoğlu & Gökmen, 2016; Bielikowicz et al., 2012; Oliver et al., 2006). Thus, modification of protein biophysical properties via glycosylation involves the careful control of the reaction conditions for the early stage of the Maillard reaction. The factors that affect protein modification by glycosylation include reaction temperature and time, pH, relative humidity, type and ratio of glucans to protein, as well as the degree of glycosylation (Akıllıoğlu & Gökmen, 2016; Solá, Rodriguez-Martinez, & Griebenow, 2007; Zhang et al., 2018). The two methods of chemical glycosylation (wet and dry heating) have been comprehensively reviewed by Zhang et al. (2018).

Figure 2.5. Mechanism of glycosylation reaction between the lysine residue of a protein and glucose leading to N-fructoselysine. Adapted from Akıllıoğlu & Gökmen (2016).

2.1.3.1.1 Effects of glycosylation on plant protein functionality Glycosylation has been well reported to improve the functionality of various plant proteins by altering their biophysical properties. The effects of glycosylation on protein biophysical properties have been well documented by Solá et al. (2007). In summary, glycosylation

increases the thermodynamic stability but substantially decreases the structural dynamics of glycoprotein without distorting their original structural fold, irrespective of the size and amount of the attached glycan. The reduction in the glycoprotein structural dynamics could be due to the reduction of the solvent-accessible surface area of the protein, and reduced mobility due to steric crowding (Solá et al., 2007). These changes in the biophysical properties are responsible for the improved functionality of glycoproteins used in food formulation. Various works of soybean protein glycosylation have been reported: SPI- lactose conjugate (Wang, Zhang, & Jiang, 2013), SPI-dextran conjugate (Boostani, Aminlari, Moosavi-Nasab, Niakosari, & Mesbahi, 2017), SPI-chitosan oligosaccharide (Xu, Huang, Xu, Liu, & Xiao, 2019), SPI-carboxymethyl cellulose conjugate (Diftis & Kiosseoglou, 2003) and soy β-conglycinin–dextran conjugate (Zhang et al., 2012). In all these works, glycosylation led to improved thermal stability, viscosity, solubility, emulsification, foaming, and water holding capacity. Glycosylation was carried out mainly by dry heating at varying reaction times (usually above 24 h or as much as 8 days), varying relative humidity (between 75-80%), varying heating temperature (usually 60°C or as high as 80°C), with the protein-polysaccharides solution, maintained at pH 7.0 - 8.5 before lyophilization and the ratio of protein-polysaccharides usually 1:1 to 1:4. Dry heating has been reported to give superior products and requires a longer reaction time than wet heating (Zhang et al., 2012). Similarly, there have been many works on glycosylation of various pulse proteins. In two separate studies, gum arabic was conjugated with pea protein isolate or concentrate at 1:4 protein-polysaccharide ratio, 60°C dry heating temperature, and 79% relative humidity. The results showed that conjugation of gum arabic with pea protein concentrates improved solubility and emulsification capacity after 3 days of dry heating while PPI had increased solubility and ES after 1 day of dry heating (Zha, Dong, Rao, & Chen, 2019a, 2019b). Rapeseed protein isolate solubility, emulsifying property, and thermal stability were improved with conjugation with dextran via wet heating at 90°C for 1- 3 h. The increase in functionality was attributed to an increase in the surface hydrophilicity and an unfolding of protein structure (Qu et al., 2018). Rice hydrolysate was conjugated (1:1) with glucose, lactose, maltodextrin, and dextran under wet heating conditions (20 min at 100°C). The results showed a lower glycosylation rate among high molecular weight saccharides and an increase in solubility, emulsifying activity, and

stability, which correlated with the degree of glycosylation and was independent of the molecular weight of the saccharides (Li et al., 2013).

2.1.3.2 Acylation and Succinylation Acylation is a chemical derivatization technique that has been widely applied to modify the functional properties of plant proteins and the reaction involves the transfer of an acyl group to the amino or hydroxyl groups of amino acid residues (Aryee, Agyei, & Udenigwe, 2018; Ferjancic-Biagini, Giardina, & Puigserver, 1998). Protein acylation is expressed with different nomenclature depending on the acylating agent and the amino/hydroxyl group to which the acylating agent is attached. Acylation that involves the transfer of the succinyl group, usually from succinic anhydride, to the lysine residue (ɛ-amino group) of the protein is regarded as Succinylation (Qazi, Jan, Ramazan, & John, 2019). When acetyl group (usually from acetic anhydride) is attached to the protein residues, then the process is described as Acetylation while the use of maleic anhydride is called Maleylation (Das Purkayastha et al., 2016). Fatty acid acylation that involves the attachment of acyl-CoA derived from short, medium, and long-chain fatty acid covalently to the protein residues, also takes their process name from the fatty acid used. For instance, Palmitoylation refers to the attachment of palmitic acid (C16:0) to the cysteine residue of proteins (Rioux, 2016).

2.1.3.2.1 Effects of acylation on plant-based protein ingredients functionality Acylation modifies the protein-protein structure and consequently its functionality by causing a reduction in the protein net surface charge, dissociation of the protein spatial structure (molecular weight), unfolding of the polypeptide chain and an increase in the aromatic-aliphatic residue balance without much compromise to the protein amino acid profile except for lysine (Gruener & Ismond, 1997). The attachment of acyl moiety increases protein flexibility, lowers surface tension, thus increasing foaming capacity (Das Purkayastha et al., 2016). Table 2.4 summarizes the effects of the different acylation processes on common plant protein ingredients.

Table 2.4. A summary of commonly acylated protein ingredients, acylating agents, and target functionality.

Substrate Acylating Agent Target functionalities Reference

Flaxseed protein isolate Acetic, succinic anhydride Decreased FC, and fat binding (Wanasunda ra & Shahidi, 1997)

Mung bean protein succinic anhydride Increased EAI and a slight increase in (Charoensu isolate ESI and solubility. k, Brannan, Chanasattru , &

Chaiyasit, 2018)

Oat protein Acetic, succinic anhydride A slight increase in solubility increased (Ma, 1984) EC and EAI. Formability was increased but FC decreased.

Pea protein isolate succinic anhydride n-octenyl succinic Increased solubility, FC, and ESI. (Shah, K.V, anhydride and dodecyl succinic & Singhal, anhydride 2019)

Increased solubility, EC, and FC (up to a (Das moderate level of acylation). Decreased Purkayastha Maleic anhydride ES, and FS. et al., 2016) Rapeseed protein Isolate Butanedioic anhydride Decrease the least gelation concentration (Wang, and increased gel strength. Increased Zhang, water holding. Zhang, Ju, & He, 2018)

Soy protein isolate, 7S, Succinic Anhydride Increased ES, reduced thermal (Wan, Liu, and 11 S, aggregation & Guo,

2018)

*FC- foaming capacity, FS- foam stability, EAI- emulsion activity index, ESI- emulsion stability index, ES- emulsion stability, and EC- emulsion capacity.

2.1.3.3 Deamidation Chemical deamidation refers to the reaction of proteins with either strong or mild acid/alkaline at elevated temperature accompanied by a loss of ammonia molecule (Hamada & Swanson, 1994). Generally, deamidation involves the conversion of δ- (asparagine) or γ- (glutamine) amide groups to carboxylic groups (α- and γ- aspartic and glutamic acids) with the release of ammonia (Li, Lin, & O'Connor, 2010). The rate of deamidation is influenced by temperature, pH, water activity, amino acid sequence, and the presence of a non-ionic catalyst (Riha, Izzo, Zhang, & Ho, 1996). The degree of deamidation is measured by the ratio of ammonia released during deamidation reactions to the total amide in the protein (Cabra, Arreguin, Vazquez-Duhalt, & Farres, 2007; Hamada & Swanson, 1994). Mild (dilute) alkaline or acid deamination is usually preferred because protein denaturation and high peptide hydrolysis are associated with strong alkaline/acidic deamidation and more importantly because of the difficulty of commercialization (Hamada & Swanson, 1994). Typically, the acid concentration range of 0.01 N – 0.04 N HCl, acetic or citric acid and alkaline concentration 0.1 - 0.5M NaOH, incubation temperature of 65- 100 ℃ for 0.5 - 48 h have been used for plant protein deamidation (Chan & Ma, 1999; Liao et al., 2010; Qiu, Sun, Cui, & Zhao, 2013; Zhao, Tian, & Chen, 2011).

2.3.3.1 Effects of deamidation on plant protein functionality The modification effect of deamidation on protein functionality has been attributed to an increase in the net negative charge, decrease in protein-protein interactions, and unfolding (Cabra et al., 2007; Riha et al., 1996). Deamidation has been widely applied to increase the functional properties (such as solubility, emulsifying, and foaming properties) of cereal protein because of the prevalence of asparagine and glutamine amino acid, which are important reactants during deamidation (Cabra et al., 2007). However, it has been reported that extensive deamidation may be detrimental to protein functionality. Thus, controlling the degree of deamidation is key to improving the functionality of plant proteins. Cabra et al. (2007) reported that the deamidation of α-zein with alkaline is best compared to enzymatic and acid deamidation. Additionally, α-zein deamidated using 0.5 M NaOH for 12 hr at 70 ℃ produced the most stable emulsions (80% of the emulsified oil), which reduced with increased incubation time and alkaline concentration. In a similar but different experiment, Zhao et al. (2010) reported increased solubility, foaming capacity

(FC), and emulsion stability (ES) of barley hordein at pH 3-7 compared to unmodified hordein. Flores et al. (2010) reported an increased emulsification activity index of corn gluten meal from 6.8 to 16.8 m2/g protein and ES from 0 to 90.6% oil retention for a 0.1 N HCl deamidation for 6 h at 70℃. Most data on cereal protein deamidation showed the best improvements in functionality when the degree of deamidation was kept below 45% with a minimal (< 3%) degree of hydrolysis (Cabra et al., 2007; Flores, Cabra, Quirasco, Farres, & Galvez, 2010; Zhao, Tian, & Chen, 2010). However, an increase in solubility, EAI, ESI, and foamability but a decrease in FS were recorded for up to 72.1% degree of deamidation when soymilk residue protein was deamidated at 65℃ for 48 h using 0.01 - 0.3 N HCl (Chan & Ma, 1999).

2.1.3.4 Phosphorylation Chemical phosphorylation was a very popular protein modification technique in the ’80s but it is now receiving renewed attention especially for the modification of cereal proteins. - It involves covalent attachment of phosphoryl group (PO3 ) to the protein molecule at specific reactive amino acid residues (Fig. 2.6). Amino acid residues with -NH, -OH, or - SH side groups such as serine, threonine, tyrosine asparagine, tryptophan, cysteine, glutamine, asparagine, histidine, lysine, arginine, aspartic acid, and glutamic acid are capable of being phosphorylated (Frank, 1987; Hu, Qiu, Sun, Xiong, & Ogra, 2019). The type of protein, phosphorylating agent, and reaction conditions are the main factors that affect the degree of phosphorylation. Three commonly used phosphorylating agents include sodium tripolyphosphate (STP), sodium trimetaphosphate (STMP) and POCl3.

According to the United States Food and Drug Administration, STP, STMP, and POCl3 are generally regarded as safe (GRAS) additive when used per good manufacturing practice (21CFR172.892, 2019). Most studies have shown that STMP and STP are best at alkaline condition (>pH 9.0) and 35-70℃ while POCI3 can be used at mild conditions, though protein cross-linking also occurs that caused reduced solubility (Hu et al., 2019; Liu et al., 2019; Sánchez-Reséndiz et al., 2018). The phosphate linkages have been reported to withstand high temperature (120℃) and pH 2.0 -10.0 thereby, making the phosphorylated proteins very useable in food applications (Li, C.-P. et al., 2010).

Figure 2.6. Mechanism of phosphorylation showing the reaction of sodium tripolyphosphate (STP) and sodium trimetaphosphate (STMP) with serine and lysine residues. Adapted from (Hu et al., 2019; Liu et al., 2019)

2.3.4.1 Effects of phosphorylation on plant protein functionality The attachment of a phosphoryl group to a protein molecule increases its hydrophilicity by deprotonating the protein, which increases the negative charge on the protein surface and predictably its solubility. Phosphorylation of many plant proteins, particularly cereal proteins is receiving new interest nowadays. STMP-phosphorylated rice bran protein at pH 9.0 showed a considerable increase in solubility (58.4 ± 2.61%) compared to untreated protein (6.67 ± 1.46%) and a significant increase in emulsion activity index (13.01 m2/g) compared to untreated protein (1.70 m2/g ) with a significant increase in the emulsion stability index (83.6% ) compared to untreated protein (50.0%) (Hu et al., 2019). An increase in the α-helix structure and not the β-sheets were observed following phosphorylation. In another experiment, the highest phosphorylation (30% for peanut and 23% for soybean protein) occurred at 2% STMP and pH 12.5 under specific reaction conditions (55°C, 5 h for peanut protein and 35°C, 3 h for soybean). The phosphorylation led to an increase in the protein functionality: solubility was 83.53% and 82.15% for the modified and unmodified pea protein while it was 29.75% and 24.07%, for the modified and unmodified soy protein (Sánchez-Reséndiz et al., 2018). Another study showed that STP-phosphorylated pea protein showed increased in nitrogen solubility index (20.26% versus 7.47% for control), EAI (139.87 m2/g vs 85.24 m2/g for control), ESI ( 87.48 min vs

51.74 min for control), FC (107.14% vs 50% for control) and FS (22.14% vs 0% for control) (Liu et al., 2019).

2.1.4 Biological/Enzymatically modified proteins The controlled application of proteolytic and non-proteolytic enzymes to alter the structural properties of a protein constitute the enzymatic modification techniques. Depending on the functionality of interest, enzymatic modification can be applied to breakdown or build up a protein structure to achieve the desired functionality. Proteolytic enzymes (such as pepsin, papain, trypsin, and alcalase) are applied to cleave peptide linkages in the protein primary amino acid sequence to modify functionality in the process of hydrolysis while non-proteolytic enzymes (such as transglutaminases) are used in the enzymatic modification of protein via protein-cross linking to build up protein structure and increase textural properties (Buchert et al., 2010). Generally, enzymatic modification is usually preferred to chemical modification because of the fast reaction time, specificity of the enzymes, and mild reaction conditions (Buchert et al., 2010; De Eslie & Cheryan, 1981).

2.1.4.1 Enzymatic hydrolysis Enzymatic hydrolysis of proteins involves a catalytic reaction between proteolytic enzymes and protein substrates that leads to cleavage of peptide bonds and splitting of the substrate into short-chain peptides and amino acids with lower molecular weights (Bučko, Katona, Popović, Petrović, & Milinković, 2016; Eckert et al., 2019). The process of enzymatic hydrolysis is best carried out at the optimal pH and the temperature of the test enzyme and substrate must be monitored/controlled for pH and temperature in order to achieve the desired results, i.e. improved functionality or yield. For instance, excessive reduction in pH during hydrolysis that could inactivate the enzyme could be prevented by adding the appropriate base to adjust the pH to the optimal level (Aluko, 2018). The various factors that influence enzymatic hydrolysis of proteins include the type of enzyme, nature of the protein substrate, the enzyme to substrate volume ratio, process conditions (pH, temperature, and pressure), and availability/absence of proteolytic inhibitors (Ahmadifard, Murueta, Abedian-Kenari, Motamedzadegan, & Jamali, 2016). Table 2.5 gives a summary of the most commonly used enzymes, optimal conditions, substrates, and target functionality.

Table 2.5. A summary of commonly hydrolyzed protein, hydrolysis condition, enzymes, and target functionality.

Substrate Enzymes Process condition Target functionalities Reference

Chickpea Alcalase 5% (w/v) substrate, pH 8.0, Increases solubility with DH. ES & FS (Yust, protein isolate temp 50℃ time 24 h, DH 1- decreased with DH. FC increased with DH. Pedroche, 10% EAI increased till DH 2.9 % and then Millán- decreased. Linares, Alcaide- Hidalgo, & Millán, 2010)

Quinoa Protein Alcalase 5% (w/v) substrate, pH 9.0, Solubility, ESI, & FC increased. FS & EAI (Aluko & concentrates temp 50℃ time 4 h, DH 48% decreased Monu, 2003)

Rapeseed Alcalase 5% (w/v) substrate, pH 8.0, Whippability, EAI, ES FC, water (Vioque, Protein temp 50℃ time 1 h, DH 1- absorption, and oil absorption increased Sánchez- concentrate 10% with increasing DH. FS decreased with Vioque, increasing DH Clemente, Pedroche, & Millán, 2000)

Neutrase 6.67% (w/v) substrate, pH (Pan et al., 7.0, temp 50℃ time 4 h, DH 2019)

4.23%

Trypsin 6.67% (w/v) substrate, pH Rice protein 8.0, temp 50℃ time 4 h, DH Emulsion droplet size of 0.27, 0.27 & 0.34 isolate 6.36% µm after 1 day to 0.47, 0.31 & 0.45 µm for Neutrase, Trypsin and Alcalase respectively Alcalase 6.67% (w/v) substrate, pH 8.0, temp 50℃ time 4 h, DH 9.81%

Soybean Isolate subtilisin 6.67% (w/v) substrate, pH Gelation occurred at higher pH 7.6 and with (Kuipers et

Carlsberg. 8.0, temp 40℃-time 30 min, increasing DH, but with a softer gel texture al., 2005) DH <10%

Soy Flour Alcalase 5% (w/v) substrate, pH 7.0, FC increased & FS decreased (Hrckova, temp 40℃ time 8 h, DH 35.1, Rusnakova, Flavourzyme 39.5, 33.3 & Novozym Zemanovic, 2002)

Wheat gluten Alcalase, 5% (w/v) substrate, pH 8.5, Solubility was over 60 % for all enzymes at (Kong, Pancreatin, 8.5, 2.0, 8.5, 6.5, & 7.0 all pH Zhou, & Pepsin, PTN respectively. Temp 60, 37, Qian, 2007) 3.0S, 37, 47, 50, & 50℃ Protamex, respectively time 6-24 h 30 and Neutrase min, DH 35.1, 39.5, 33.3

*FC- foaming capacity, FS- foam stability, EAI- emulsion activity index, ES- emulsion stability, EC- emulsion capacity, and DH- degree of hydrolysis.

Following enzymatic hydrolysis, the structural properties of the protein are modified and some hidden structural particulars such as buried hydrophobic residues are revealed (Eckert et al., 2019). Further hydrolysis into much smaller peptides can cause bitterness of the peptide due to the accumulation of hydrophobic residues (Wei, Thakur, Liu, Zhang, & Wei, 2018). However, it has been reported that limited hydrolysis (DH < 10%) can improve the functionality of the protein and avoid the bitterness associated with extended hydrolysis (Adler-Nissen, Eriksen, & Olsen, 1983; Eckert et al., 2019). The extent of hydrolysis is controlled by the degree of hydrolysis (DH) which is a measure of the percentage of peptide bonds cleaved (Adler-Nissen et al., 1983; Mokni Ghribi et al., 2015).

2.1.4.1.1 Effects of enzymatic hydrolysis on plant protein functionality

Generally, because of the size reduction of the protein polypeptides, enzymatic hydrolysis has been reported to increase certain functionality of food proteins such as solubility, emulsifying and foam properties (Galante, De Flaviis, Boeris, & Spelzini, 2020; Mokni Ghribi et al., 2015) It is obvious that the solubility of plant protein hydrolysates increases across a wide range of pH values, irrespective of the protein substrate and protease as can be seen in Table 2.5. Additionally, The FC and EAI tend to increase mostly for hydrolyzed protein while the FC and ES decrease. This behavior is exacerbated by increasing the DH of the protein. The increased functionality such as solubility, emulsion and foam capacity of hydrolyzed plant proteins has been attributed to the increase in protein molecule solvation while the challenges with emulsion and foam stability have been attributed to the inability of the short-chain peptides to be flexible enough to form stable interfacial films around the oil droplets or air bubbles (Aluko & Monu, 2003; Mokni Ghribi et al., 2015)

2.1.4.2 Enzymatic cross-linking Transglutaminase (TG) and other oxidative enzymes are some of the enzymes used to induce cross-linking of food proteins. Usually, the goal of crosslinking is to improve the textural properties of the protein by building up the polypeptides into stronger structures. Transglutaminase (EC 2.3.2.13) is the only commercially available food-grade cross-linking enzyme and it been reported to improve texture, viscosity, emulsion stability, gelation, foam and

increase hydrophobicity properties of food proteins (Djoullah, Husson, & Saurel, 2018; Glusac, Isaschar-Ovdat, & Fishman, 2020; Nivala, Mäkinen, Kruus, Nordlund, & Ercili-Cura, 2017; Sun, Xiang Dong & Arntfield, Susan D., 2011). Xiang et al. (2011), reported that TG increased the gel strength of PPI 8 times and SPI gel by 2 times in comparison to the gel made from untreated protein; gel strength also increased with higher TG levels. TG enzyme works by catalyzing acyl transfer reactions and crosslinking (polymerization) between protein intra- or inter-chain glutamine (acyl donor) and lysine (acyl acceptor) amino acid side chain residues in the food protein (Gaspar & de Góes-Favoni, 2015). A list of other crosslinking enzymes, their optimal conditions, substrates, and target functionality is presented in Table 2.6.

Table 2.6. A summary of common protein ingredients, type of cross-linking enzymes, and the target functionality impacted.

Substrate Enzymes Process condition Target functionalities Reference

Soy protein TG 10 unit/ml (10%, w/v) enzyme, Increases gel strength & elasticity. Reduced Sun & isolate and pea 10.5% (w/v) substrate, 0.3 M minimum gelation concentration. Arntfield, protein isolate NaCl pH 7.0, 40℃, 24 h 2011

Soy glycinin TrT 0.005% (w/v) enzyme, 0.5% Less stable emulsion, & high cream velocity (Isaschar- (w/v) substrate, pH 7.4, temp compared to untreated sample. Ovdat, 37℃ time 0 - 4 h + 1 mM Rosenberg, caffeic acid/ p-coumaric Lesmes, &

43 acid/chlorogenic acid Fishman,

2015)

Faba bean TG & TrT 10, 100, or 1000 nkat/g of TG decreased solubility from 83-60%. TrT (Nivala et protein isolate protein enzyme, 10 mg/ml decreased solubility from 83-75%. TG al., 2017) substrate, pH 7.0, 40℃, 20 h decreased foam height by 14% and TrT reduced it by 5%.

Denatured pea TG 20 units of TG/g protein at Albumin did not form a gel with TG. (Djoullah albumin and 40 °C and pH 7 at incubation Globulin with TG formed gel at 8% and 3% et al., 2018) globulin times 10- 300 min. protein concentration for chemical and thermally denatured globulin respectively.

Wheat flour TrT & LA 12g flour in 7.08 ml H2O Harder and less extensible dough. Larger (Selinheim dough substrate, 5, 10, and 30 nkat/g bread pore size and increased bread volume. o, Autio, of flour for 3.5 min Kruus, & Buchert, 2007)

Oat protein TG & TrT 10, 100, or 1000 nkat/g of TG increased solubility from 16-19%. TrT (Nivala et isolate protein enzyme, 10 mg/ml decreased solubility from 16-6%. TG al., 2017) substrate, pH 7.0, 40℃, 20 h increased foam height by 29% and TrT reduced it by 40%.

Potato protein TrT 1 % (w/v) substrate, enzyme to Stable emulsion for over a month and higher (Glusac,

44 and zein the substrate (1:30, w:w), pH 7 storage modulus Davidesko-

for potato protein and pH 10 for Vardi, zein, time of 30 min Isaschar- Ovdat, Kukavica, & Fishman, 2018)

*TG-transglutaminase, TrT- Tyrosinase, and LA- laccase

Chapter III: Objective 1 Physicochemical, Functional and Structural Properties of Proso Millet Storage Protein Fractions Abstract The understanding of the protein structure-function relationship is very important to the study of protein chemistry. In this research, the physicochemical, functional, and structural properties of the storage protein fractions from two defatted proso millet cultivars (Dawn and Plateau) were determined and reported. The results show that the protein recovery efficiency of 53.5% and 60.1% was recorded for Dawn and Plateau, respectively. The average denaturation temperature of all fractions was about 82.1±3.5°C. Surface hydrophobicity values for Dawn fractions were 11781, 10594, 316, and 2225 for albumin, globulin, and glutelin, respectively, and 3415, 2865, 353, and 456 for Plateau fractions, respectively. Most of the protein fractions showed the highest solubility at pH 9 and the lowest solubilities at pH ≤ 7 with solubility range from 5.7- 100%. Emulsifying activity index (EAI) of less than 25.0 m2/g was recorded for most fractions, while the highest emulsion stability index (ESI) recorded was about 60 min. Prolamin fractions showed three major peptide bands of 11, 14, and 24 kDa while glutelin fraction revealed only a major band of 15 kDa and several minor bands of 11, 22, 24, 78, 209 kDa. No differences in the electrophoresis pattern were observed for the fraction with or without a reducing agent.

3.0 Introduction There is an increasing global interest in plant-based protein partly due to the rising cost of animal-based protein ingredients, consumer’s preferences for lean protein as well as their desire for clean, natural and sustainable plant-based food (Aschemann-Witzel & Peschel, 2019; Henchion et al., 2017). Consequently, the plant-based protein market is fast expanding, estimated at about $8 billion in 2016, and projected to be $9.5 billion by 2024 at a compound annual growth rate (CAGR) of 7.0% over a forecasted period from 2019-2024 (Mordor_ Intelligence, 2017, 2019). With this trend, the search for non-traditional plant- based protein sources such as millet, sorghum, quinoa, hemp, and water lentil is attracting a lot of research interest in order to meet the high demand for protein foods expected to be in excess of one-third of the current demand by the year 2050 (Henchion et al., 2017).

Millet is an important emerging plant-based protein source owning to its short crop season (60-90 days), its drought and disease resistance potential and it’s adaptability to different soil and climatic conditions (Baltensperger, 2002; Sheahan, 2014). Millet is regarded as the sixth most important cereals in the world, consumed by more than one-third of the world’s population (Habiyaremye et al., 2016). Proso millet is the species of millet widely grown in the United States of America, majorly in states of Colorado, Nebraska and South Dakota, with a production of as much as 9.1 million bushels (202,475 metric tonnes) in 2011 and 14.6 million bushels (324,850 metric tonnes) in 2017 (AgMRC, 2018; Michael, 2012; Saleh et al., 2013). From a nutrition perspective, proso millet protein is gluten-free, a key interest for its use in gluten-free products (Kalinova & Moudry, 2006). Additionally, proso millet protein is richer in some essential amino acid (leucine, isoleucine, methionine), compared to wheat protein, and its protein content is higher and/or similar to wheat and some other grains (Kalinova & Moudry, 2006). Generally, the protein content of proso millet has been reported to range between 9.49-16% on a dry basis, depending on the variety (Kalinova & Moudry, 2006; Yu, Qiuxia, & Lizhen, 2012)

However, the application of millet proteins or its protein fractions as food ingredients is limited, probably due to scarce information on its physicochemical, structural, and functional properties. The knowledge of these properties allows the elucidation of proso millet proteins interaction with other biomolecules in a food system. For example, the absence of cysteine residues in the sulfur-poor prolamin of some cereals (wheat ɷ-gliadins, rye ɷ-secalin, and barley C-hordein) is responsible for their non-incorporation in a disulfide-bonded polymer (Greenfield et al., 1998). The acidic subunits (AS III) from glycinin plays an important role in soy protein isolate gel formation and increases gel hardness (Utsumi & Kinsella, 1985). Additionally, deamidated Z19 α-zein showed an increase in emulsifying capacity and stability compared to native Z19 α-zein because the process of deamidation causes the unfolding of Z19 α-zein revealing its buried hydrophobic residues and increasing its charges so that there is a better hydrophilic-lipophilic balance in the oil-water interface (Cabra et al., 2007).

Proteins are complex biomolecules and some of them are made of several subunits with different physicochemical and functional properties. Based on their solubility, plant storage

proteins are classified as albumin (water-soluble), globulin (saline soluble), prolamin (alcohol-soluble), and glutelin (alkaline- soluble). The knowledge of the physicochemical properties (e.g. solubility, surface hydrophobicity, and thermal denaturation temperature) and functional properties of proteins (e.g. emulsifying properties, foaming, water binding, and water holding properties) is very important in food application and these properties are influenced by the structure, conformation, and the ambient solution condition of the protein (pH, ionic solution, and temperature). Thus, the objectives of this research were to (1) extract the protein fractions of two important proso millet cultivars along the lines of their solubility (2) determine some of the physicochemical and functional properties of the various fractions (such as the thermal denaturation temperature, solubility, surface hydrophobicity, emulsion, and foaming properties) (3) determine the structural particulars of the protein fractions.

3.1 Material and Methods 3.1.1 Materials Two commercial proso millet cultivars (Plateau and Dawn) were received as a gift from Panhandle Research & Extension Center, University of Nebraska–Lincoln, USA. The samples were dehulled using a lab-scale dehulling machine (Glenn Mills Inc., Clifton, NJ) with some modification (the stationary disc was replaced with a rubber disc). The dehulled samples were washed to remove dirt and air-dried in an oven at 35°C for 1 day. Samples were then milled using Quadrant Junior Mill (C.W. Brabender Instruments Inc., South Hackensack, NJ). These samples are composed of waxy and non-waxy starch types, respectively with their proximate composition shown in Table 3.1. Details of proximate composition are shown in (Singh et al., 2018).

Table 3.1. Proximate composition of the proso millet flours used in this study

Proso Moisture Crude protein Crude fat (%) Crude fiber Ash (%) Carbohydrate Amylose millet content (%) (%) (%) (%) (%) cultivar Dawn 9.40±0.08a 15.14±0.01a 3.51±0.03a 0.59±0.30a 0.77±0.01a 80.50±0.06a 25.10±0.28a

Plateau 9.71±0.13b 14.79±0.02b 3.63±0.13a 0.80±0.08a 0.74±0.00a 80.66±0.11a 3.10±0.28b

The values are means ± standard deviations of two replicate determined on a dry basis. Crude protein, crude fat, crude fiber, ash, and carbohydrate are calculated on a dry basis. Carbohydrate

= 100% - (protein% + Fat% + Ash %). Means with a different letter in a column are significantly different (P<0.05). Source: Singh et al. (2018).

3.1.2 Extraction and isolation of protein fractions Proso millet flour was first defatted using hexane solvent (1:6) for 6 hours to remove lipids and a mixture of chloroform and methanol (40:60) to remove carotenes and xanthophylls. The process was repeated thrice, and the fat extracts were filtered through Whatman filter paper No. 4, then the residue was air-dried at room temperature for 24 h. The defatted proso millet flour samples were extracted into their corresponding fractions (albumin, globulin, prolamin, and glutelin) using different extracting solvents (Figure 3.1) according to Adebiyi & Aluko, (2011) with some modifications. In this case, to obtain the albumin and globulin fractions, defatted proso millet flour was mixed with 0.5 M NaCl solution and stirred continuously for 4 h followed by centrifugation at 5000 x g for 30 min at 4°C. The supernatant was collected and dialyzed (MWCO 10kD) extensively against Nano pure water for 5 days at refrigeration (≈ 4°C) condition with the intermittent replacement of the Nano pure water. The dialysate was further centrifuged at 5000 x g for 30 min at 4°C; the supernatant was collected as albumin fraction while the precipitate collected was globulin fractions. Both albumin and globulin fractions were freeze-dried and kept at -20°C until use. The prolamin and glutelin were extracted using 70% Isopropyl alcohol (Kohama, Nagasawa, & Nishizawa, 1999) and 0.05 M NaOH, respectively. The extracts were dialyzed extensively followed by centrifugation of the dialysate and the precipitate was freeze-dried and kept at -20°C until use. The protein content of all fractions was determined using the Dumas combustion method in duplicates (Jung et al., 2003).

Defatted proso millet flour (100 g)

Albumin: water soluble fraction (supernatant) Extraction with 0.5 M NaCl for 4 hr, centrifuged at Salt-soluble extract Dialyze against Nano 5000 x g for 30 min pure water and centrifuge Globulin: salt soluble fraction (precipitate)

Sediment-1 Freeze dry

Extraction with 0.05 M NaOH for 4 hr, centrifuged Alkaline-soluble Dialyze against Nano Glutelin: alkaline 50 at 5000 x g for 30 min extract pure water and soluble fraction

centrifuge (precipitate)

Sediment-2

Extraction with 70% Dialyze against Nano Prolamin: alcohol Isopropyl Alcohol for 4 hr, Alcohol-soluble pure water and soluble fraction centrifuged at 5000 x g for extract 30 min centrifuge (precipitate)

Residue

Figure 3.1 Flow chart of sequential extraction of proso millet protein fractions from defatted proso millet flour at room temperature

3.1.3 Physicochemical properties characterization

3.1.3.1 Thermal properties The thermal properties of proso millet protein fractions were determined using a differential scanning calorimeter (DSC - Q20, TA Instruments, New Castle, Delaware, USA) (Ju et al., 2001) with some modifications. The protein fraction weighing 50 mg was dissolved in 1 ml of 0.01 M phosphate buffer (pH 7) in an Eppendorf tube for 30 min at 4°C with occasional vortexing. The protein solution was then centrifuged at 16000 x g for 5 min at 4°C and the supernatant pipetted out leaving a solid residue at the bottom of the tube. About 10 mg of the residue protein was transferred into a stainless-steel pan and hermetically sealed. Stainless steel pans containing the samples were scanned against an empty pan (reference) from temperatures 20°C to 120°C at a heating rate of 20°C/min. Onset temperature, maximum denaturation temperature, and change in enthalpy were then computed from the thermogram generated by the DSC software. The assay was run in triplicates and the average of runs was reported.

3.1.3.2 Surface hydrophobicity (So)

The surface hydrophobicity (So) of the protein fractions was determined by the method of (Gulati et al., 2017; Nwachukwu & Aluko, 2018) with slight modification. About 10 mg/ml of protein fractions were dissolved in 0.01 M phosphate buffer solution for 30 min with occasional vortexing and the solution was then centrifuge (16000 x g for 15 min at 4°C) to collect the supernatant. The protein content of the supernatant was determined using the Bradford method (Bradford, 1976) and the supernatant diluted to a final concentration range of 0.01% to 0.1 % using the same phosphate buffer. To determine the fluorescence intensity (FI), 20 µL of 1-anilino-8-naphthalene sulfonate (ANS) solution (8 mM in 0.2 M phosphate buffer) was added to 4.0 ml of the protein solutions, incubated for 2 h at room temperature in the dark and the fluorescent intensity (FI) was measured using a spectrofluorometer (FluroMax®-3, Jobin Yvon Inc. Edison, NJ. USA) at an excitation wavelength of 365 nm and an emission wavelength of 520 nm. The initial slope of FI against sample concentration was used as a measure of surface hydrophobicity index.

3.1.3.3 Solubility Proso millet protein fractions solubility was determined along with the different pH ranges (3, 5, 7, and 9) by the method of (Wu et al., 1998) with slight modification. 20 mg of protein sample was dissolved in 10 ml of 0.01 M phosphate buffer and the protein solution was adjusted to the desired pH using either 1N HCl or 1 N NaOH with continuous stirring for 30 min at room temperature. The protein solution was then centrifuged (16000 x g for 10 min at 25°C) and the protein content of the supernatant determined by the Bradford method (Bradford, 1976) with BSA as standard. To determine the total soluble protein in the samples (control), twenty milligrams (20 mg) from each sample was dispersed in a 10 ml 0.1 M NaOH solution and the procedure to determined protein content followed as before. Protein solubility was expressed as a ratio of the protein content in the supernatant to the protein content in the original sample (control) equation 3.1 below.

protein content in the supernatant Solubility (%) = x 100 (3.1) protein content in sample

3.1.4 Functional properties 3.1.4.1Emulsifying properties Emulsifying properties of protein fractions were measured by the methods of Pearce (Pearce & Kinsella, 1978) as described in Wu et al., (1998) with modification. In summary, 1.5 ml of pure canola oil was mixed with 4.5 ml of 1% w/v protein solution dissolved in 0.01 M phosphate buffer (pH 7) and homogenized with a Kinematica Polytron homogenizer (Model PT 10/35 GT, Brinkmann Instruments Inc., Westbury, NY, USA) set at 16000 rpm for 1 min. About 50 µL of the portion of the homogenized emulsion was pipetted from the bottom of the container at 0 and 10 min after homogenization. Each aliquot sample was diluted with 5mL of 0.1% SDS solution. The absorbance of the diluted emulsions was measured at 500 nm with a spectrophotometer (UVmini-1240 Shimadzu, MD USA). Emulsifying activity index (EAI) and emulsifying stability index (ESI) were calculated from the equations below (Wu et al., 1998).

2 EAI (m ⁄g) = 2T (A0 x dilution factor/(C x Ф x 10,000)) (3.2)

Δt ESI (min) = A0 x ⁄ΔA (3.3)

Where Δt = 10 min andΔA = A0 − A10 ; Turbidity, T = 2.303, A0 = Absorbance measured immediately after emulsion formation; dilution factor =100, C = weight of protein/unit volume (g/mL) of aqueous phase before emulsion formation; and Ф = oil volume fraction of emulsion.

3.1.4.2 Foaming properties Foaming capacity (FC) and foam stability (FS) of the protein fractions were assayed according to the method of (Mohamed et al., 2009) with minor modification. About 20 ml of 0.5% w/v of protein solution prepared in 0.01 M phosphate buffer (pH 7) was homogenized in a 100 ml of plastic cylinder using a Kinematica Polytron homogenizer equipped with a high-foam PTA-20SM generator (Model PT 10/35 GT, Brinkmann Instruments Inc., Westbury, NY, USA) at speed setting “5” (approximately 12825 rpm) for 1 min at room temperature. The total volume of foam in the measuring cylinder was measured immediately by subtracting the new volume of liquid from the initial volume, and FC was computed according to (Motoi et al., 2004). The foam was allowed to stand undisturbed, and the volume of liquid drained from the foam after 10 min was measured and used as the indicator of foam drainage stability (Mimouni, Azanza, & Raymond, 1999)

volume of foam after agitation FC(%) = x 100 (3.4) volume of liquid before agitation

residual volume of foam after 10 min FS(%) = x 100 (3.5) volume of foam after agitation

3.1.5 Structural properties 3.1.5.1 Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-Page)

To identify the polypeptides subunits in proso millet protein, SDS-PAGE analysis of the protein fractions was done using the Laemmli buffer system (Laemmli, 1970). Protein samples were dispersed in Tris-HCl buffer pH 6.8 with or without (10% w/v) β- mercaptoethanol (β-ME) for non-reducing and reducing condition to make a final protein concentration of 4 mg/ml. The protein solution was heated at 100°C for 10 min. cooled, centrifuged and the supernatant collected. A 20 µL aliquot sample of the supernatant was then loaded onto a pre-cast gel well (12%) using a Bio-Rad Mini PROTEAN® 3 system (Bio-Rad Laboratories, Hercules, CA, USA). Molecular marker (2 - 250 kDa) was run on

the same gel as standard and the protein bands after electrophoresis were stained with Coomassie Brilliant Blue R-250. The molecular weight of separated polypeptides was estimated from the plot of the log standard (molecular marker) against the reciprocal of relative migration.

3.1.6 Experimental design and statistical analysis A completely randomized block design was used in this experiment in which proso millet cultivar (Dawn and Plateau cultivar) is the blocking factor. Four treatment factors were considered (albumin, globulin, prolamin, and glutelin fractions) and five response variables (denaturation temperature, solubility, surface hydrophobicity, foaming properties, and emulsion properties). The average of three replicates was analyzed by one-way-ANOVA using the GLM procedure of SAS (SAS Institute, 2004) to determine the differences in means at a significant level of p < 0.05. Tukey’s test was used to access the difference among specific treatments means at p < 0.05.

3.2 Results and Discussion 3.2.1 Extraction and isolation of protein fractions The extraction of the different protein fractions was carried out according to the method of Osborne (Osborne, 1916) and the fractions were isolated by centrifugation and filtration, followed by a dialysis purification step. Isolation using trichloroacetic acid (TCA) in cold acetone or just pure acetone alone resulted in protein factions with very low solubility (data not shown). Approximately, about 54.02% and 60.70% protein recovery yield were obtained for Dawn and Plateau cultivars, respectively following the purification step. Kohama et al., (1999) recorded a higher recovery yield (95.7%) of proso millet proteins because the prolamin was extracted at a higher temperature of 60°C which gave a more than three-fold increase in prolamin yield compared to ours extracted at room temperature. Additionally, the remaining prolamin and glutelin fractions in the proso millet flour were extracted in the presence of a reducing agent, β-Mecarptoethanol (β-ME) which facilitated additional protein yield. However, all of our extraction was carried out at room temperature and we avoided using β-ME because it is not a food-grade chemical. All these might have caused the lower recovery yield recorded in our study. From the extracted fractions, albumin accounts for 9.5%, globulin accounts for 4.1%, prolamin accounts for 47.2%, and

glutelin accounts for 39.1% of the total extracted crude protein of the Dawn cultivar. While for Plateau cultivar, albumin accounts for 8.6%, globulin for 6.1%, prolamin for 50.8%, and glutelin account for 34.5% of the total protein content. These results (Table 3.2) are in agreement with the percentage of protein fractions in most cereals, particularly with the result reported by other authors for the protein fractions in millet (albumin-18.2%, globulin-6%, prolamin-33.9% and glutelin-41.8% of the total protein respectively) (Kohama et al., 1999; Parameswaran & Thayumanavan, 1995; Wouters & Delcour, 2019). Consistently, prolamin and glutelin make up the majority of the protein content in most cereals, except for rye, triticale, and oats where albumin and glutelin account for the most protein fractions (Wouters & Delcour, 2019). The differences in the yield of the protein fractions are usually due to differences in the protein content of various millet cultivar and in the extraction and purification method used. The protein content of prolamin and glutelin is much higher than albumin and globulin.

Table 3.2 Recovery yield of crude protein extract of and its corresponding protein content of proso millet protein fractions from Dawn and Plateau flour cultivars.

Protein fractions Dawn cultivar Plateau cultivar Yield of crude Percent of Protein content Yield of crude Percent of Protein content protein extract total protein of crude extract extract (g/100g total protein of crude extract (g/100g flour) extracted (%) flour) extracted (%) Total extracted

Flour 13.72±0.01a 100 13.72±0.01a 13.35±0.02a 100 13.35±0.02a Albumin 0.70±0.03b 5.10 9.45±0.20cb 0.70±0.01b 5.21 10.95±0.55b

Globulin 0.30±0.00c 2.20 20.29±0.07c 0.48±0.04c 3.62 19.25±1.45c

Prolamin 3.51±0.17d 25.56 77.30±0.59d 4.13±0.24d 30.90 64.32±2.74d

Glutelin 2.90±0.011e 21.16 37.60±1.10e 2.80±0.11e 20.98 64.84±0.59e The values are means ± standard deviations of two replicate determined on a wet weight basis. Means with a different letter in a column are significantly different (P<0.05).

3.2.2 Physicochemical properties characterization 3.2.2.1 Thermal properties When proteins are subjected to thermal treatment there is a transition from their native state to a denatured one, in which there is an irreversible (mostly) unfolding of the protein into a disordered conformation without a break in the peptide backbone. This transitioning occurs at the thermal denaturation temperature and provides more information on the thermal properties, stability and cooking behavior of the protein. The thermal denaturation temperatures and heat capacities of the different protein fractions from proso millet are presented in Table 3.3. A single enthalpy peak was observed for all the fractions, however, prolamin fractions showed the least prominent peak. On average, the denaturation temperature of all the fractions for both cultivars is about 82.1 ± 3.5°C and the heat absorbed to cause denaturation is 0.1 ± 0.06 J/g. There was no significant difference (P≥0.05) in the maximum denaturation temperature for all the fractions for both cultivars but there was a significant difference in their enthalpy (△H). The difference in △H for the protein fractions could be attributed to the molecular changes from the unfolding of these proteins. Consistent lower△H values of prolamin could be attributed to the contribution from exothermic reactions from the disruption of hydrophobic interactions which may be predominant in prolamin fractions due to a preponderance hydrophobic amino acid residue. These values are consistent with those obtained for yellow (88.98°C, 0.01 J/g) and white millet (86.79°C, 0.10 J/g) concentrates by (Mohamed et al., 2009). Also, Ju et al., (2001) reported that no observable peak for rice prolamin but reported an average denaturation temperature of 78.1°C and a change in enthalpy of 3.27 J/g for all other fractions. Thermal treatment is applied to protein for enhancement of texture, flavor, digestibility, microbiological safety and sometimes to reduce allergenicity (Davis & Williams, 1998).

Table 3.3. Thermal denaturation properties of the protein fractions of two proso millet flour from two different cultivars

(Dawn and Plateau)

Protein fractions Dawn Cultivar Plateau Cultivar Onset Max. Change in Onset Max. Change in temperature temperature enthalpy, ΔH temperature temperature enthalpy, ΔH (°C) (°C) (J/g) (°C) (°C) (J/g) Albumin 81.25±1.37a 83.17±0.18a 0.06±0.01a 85.03±7.42a 87.53±6.97a 0.10± 0.02a

a a b a a b 58 Globulin 80.35±0.39 83.12±0.74 0.12±0.04 79.45±0.93 81.83±0.45 0.05±0.01

Prolamin 72.63±5.96a 74.21±6.82a 0.07±0.01c 80.10±0.44a 81.77±0.90a 0.07±0.07c

Glutelin 79.51±0.68a 83.82± .55a 0.22±0.03d 80.43±0.30a 81.21±0.21a 0.03±0.00d The values are means ± standard deviations of three replicates. Means with different letter superscript in a column are significantly different (P<0.05).

3.2.2.2 Surface hydrophobicity (So) The surface hydrophobicity of protein indicates the extent to which the hydrophobic patches of protein molecules are exposed. The quantification of surface hydrophobicity can help in predicting protein functionality such as emulsifying and foaming ability (Nakai, 1983). The surface hydrophobicity data of the different proso millet protein fractions are presented in Figure 3.2. It was observed that the albumins fraction showed the highest hydrophobicity index (11781 for Dawn and 3415 for Plateau) and the prolamins (316 for Dawn and 354 for Plateau) showed the least. This result was within the range reported for corn glutelin (710.60) (Zheng et al., 2015). Cabra et al., (2007) reported a much higher So for Z19 α-zein from corn. The differences in the So of the different protein fractions within and across cultivars of the same cereal grain or other cereal may be due to differences in the level of hydrophobic residues and how well these residues are exposed following processing or denaturation. Although prolamin and glutelin proteins fractions from cereals generally contain mostly hydrophobic amino acid residues compared to albumin and globulin, the lower So value recorded in this study could be due to the fact that most of the hydrophobic residues of the prolamin and glutelin fractions are buried in the core of the protein at neutral pH. In consonance with our observation, rice globulin fraction was reported to show a higher surface hydrophobicity than glutelin fractions at room temperature (Ju et al., 2001). Exposure of the buried hydrophobic residues is usually achieved via denaturation that will unfold the protein. Ju et al., (2001) saw an increase in glutelin hydrophobicity with heating from 65 to 95°C while and increase in the hydrophobicity of Z19 α-zein was increased following alkaline deamidation (Cabra et al., 2007). Consistently, it was observed that the Dawn cultivar protein fractions showed significantly (P<0.05) higher surface hydrophobicity than those of the Plateau cultivar protein fractions, except for prolamin. The difference is cultivar may have imparted this observation.

Figure 3.2. Surface hydrophobicity index of protein fractions of two proso millet flours from two different cultivars (Dawn and Plateau) at pH 7. Means with different letter superscript across fractions are significantly different (P<0.05). While mean with different numerical superscript within a fraction are significant (P<0.05).

3.2.2.3 Solubility at different pH values The solubility profiles (Figure 3.3) of the various proso millet protein fractions followed the typical “U-shaped” curve exhibited by most seed storage proteins rather differently. There was an initial low solubility around pH 3 for albumins, followed by a steady increase in solubility from pH 3-9 for both cultivars, except that Dawn albumins decreased again from pH7-9. For globulins fractions, the initial decrease was from pH 3 to 5, followed by a steady increase to pH 9. The prolamin and glutelin fractions exhibit similar solubility profile, an initial increase from pH 3 to pH 5, and then a decrease at pH 7 before a greater increase at pH 9. In general, less than 40% solubility was recorded for the proso millet protein fractions at pH 7 and lower, except for Dawn albumin (85.8% at pH 7) and Plateau (43.8% at pH 7). Most protein fractions showed the highest solubilities at pH 9, except for Dawn albumin (69.1%) and Plateau glutelin (5.7%). Notwithstanding, prolamin and glutelin showed the lowest solubility (mostly less than 20%). Similar results by another author show that the average nitrogen solubility of proso millet flour from six different varieties (BR7, Heen Mineri, IPM 1006, MS 2420, MS 4872, and Raum 1) was 21.4% at pH 8 or above and 12% for pH 6 and below (Ravindran, 1992).

Figure 3.3. Solubility profiles of protein fractions from two proso millet flour cultivars (Dawn and Plateau) at various pHs (3-9).

This could corroborate why we observed low solubilities of prolamin and glutelin since these proteins form the major proteins in proso millet flour. There were significant differences (p<0.05) in the solubility results of all the proso millet fractions across different pHs irrespective of the cultivar. Consistently, there were significant differences between the globulin fractions of Dawn and Plateau cultivar, but not prolamin fractions across pHs. Also, there were consistent significant differences in the glutelins except at pH 5. Similarly, albumin fractions were significantly more soluble than globulin fractions at neutral pH irrespective of the cultivar. This could be largely due to the presence of lower molecular weight subunits and the attachment of large carbohydrate moieties to the albumins compared to globulins that are dominated by larger molecular subunits as can be seen later in the electrophoresis pattern (Figure 3.6). The presence of large hydrophobic amino acid residue may explain the poor solubility profile of prolamins and glutelins.

3.2.3 Functional properties 3.2.3.1Emulsifying properties The emulsion activity index (EAI) and emulsion stability index (ESI) of the various protein fractions from proso millet are presented in Figure 3.4. The EAI is an indication of the

ability of proteins to act as an interfacial active agent in an oil-in-water interface due to its ambivalent nature.

Figure 3.4 Emulsifying activity index (EAI) and emulsifying stability index (ESI) of protein fractions from two proso millet flour cultivars (Dawn and Plateau) at pH 7. Means with different letter superscript across fractions are significantly different (P<0.05) for a measured property. While mean with a different number of superscripts within a fraction are significant (P<0.05) for a measured property.

The emulsifier (proteins) forms a visco-elastic film around the dispersed oil droplets in the water phase (Lam & Nickerson, 2013). It can be observed that albumins and globulins fractions significantly (P<0.05) showed a higher EAI (17.41, 16.36 m2/g for Dawn albumin and globulin; 19.71, 15.37 m2/g for Plateau albumin and globulin respectively ) than prolamins and glutelins (2.12, 13.31 m2/g for Dawn prolamin and glutelin; 2.96, 15.05 m2/g for Plateau prolamin and glutelin respectively); particularly, prolamins showed the least EAI. Similar EAI values have been reported for other cereals proteins like freeze-dried rice bran isolate (0.19 m2/g) (Tang, Hettiarachchy, Horax, & Eswaranandam, 2003) and oat

protein isolate (49.0 m2/g) (Mirmoghtadaie, Kadivar, & Shahedi, 2009). It is not unexpected that prolamin and glutelin showed a lower EAI compared to the albumins and globulins since they also showed the least surface hydrophobicities and solubilities (Figures 3.2 & 3.3). It has been reported that EAI could be affected by surface hydrophobicity and solubility (Du et al., 2012; Lam & Nickerson, 2013). Although prolamin and glutelin are highly hydrophobic, good emulsifiers require a balanced hydrophobic–hydrophilic ratio. Also, it may be possible that the hydrophobic residues of prolamin and glutelin are buried in the core of the protein. Consequently, some form of modification to expose these residues is necessary, so that the protein can re-align and position themselves with their surface hydrophobic nature in the oil phase and hydrophilic nature in the water phase (Lam & Nickerson, 2013).

The emulsion stability index reflects how well an emulsifier can keep the dispersed oil droplet in solution stable to withstand perturbation to its structure (e.g. coalescence, creaming, flocculation, and sedimentation) over a period of time (Boye, JI et al., 2010). It can be seen from Figure 3.4 that the prolamins showed the most stable emulsions (64 and 35 min for Dawn and Plateau cultivar, respectively) while the globulins showed the least stable emulsions (average of 15 min each for either cultivar). Mirmoghtadaie et al., (2009) reported an ESI of 63 min for oat protein isolates and 26 min for rice bran protein concentrates (Zhang, Zhang, Wang, & Guo, 2012). Albumin fractions stabilized emulsion better than globulin, possibly due to the lower molecular weight of the fractions and the presence of attached polysaccharides (Lam & Nickerson, 2013). The effect of cultivar was not consistent for both EAI and ESI across the various fractions possible due to different physicochemical properties (surface hydrophobicity and solubility).

3.2.3.2 Foaming properties The foaming capacity (FC) and foam stability (FS) of the various proso millet protein fractions are presented in Figure 3.5. Foaming capacity measures the ability of the protein to trap air (or foam dispersed air bubbles) in a continuous liquid or semi-solid phase. Usually, the protein molecules will have to diffuse to the air/water interface, where they unfold, concentrate, and form a thick visco-elastic film around the gas bubbles. It can be observed that glutelin fractions had the highest foaming capacity (87% for Dawn and 60%

for Plateau) while prolamin showed the least foaming capacity (18% for Dawn and 14% for Plateau). Mohamed et al., (2009) reported an FC of 137 and 124 g/ml for white and yellow foxtail millet concentrates and an average FC of less than 40% was reported for rice bran protein concentrates at pH 7 (Bera & Mukherjee, 1989). There were insignificant differences (P>0.05) in the FC and FS of the fractions from the two cultivars of proso millet, except for FC for albumin and glutelin fractions. Foam stability is an indication of the formation of a continuous thick viscoelastic and air-impermeable film trapping air bubbles. Glutelin fractions showed the highest FS (97% for Dawn and 88% for Plateau) while prolamin recorded the lowest FS (56% for Dawn and 58% for Plateau). Other cereal proteins like oat protein isolates showed an FS of 100% (Mirmoghtadaie et al., 2009).

Figure 3.5. Foaming capacity (FC) and foaming stability (FS) of protein fractions from two proso millet flour cultivars (Dawn and Plateau) at various pH 7. Means with different letter superscript across fractions are significantly different (P<0.05). While means with a different number of superscript within fractions are significant (P<0.05).

3.2.4 Electrophoresis pattern (SDS-PAGE) The SDS-PAGE profiles of the various fractions of proso millet cultivars are presented in Figure 3.6. There was no considerable difference in the electrophoresis band pattern and

intensity of the fractions from the two proso millet cultivar with or without the addition of a reducing agent. This shows that the protein subunit may be devoid of disulfide bonds. A similar electrophoresis pattern was reported by Kohama et al. (1999), where they showed no difference in the pattern even after the addition of dithiothreitol (DTT) to prolamin fractions. The SDS_PAGE analysis showed that albumin fractions composed of polypeptides ranging from 11 to 60 kDa with major peptides of 11 and 24 kDa and four minor peptides of 13, 15, and 42 kDa.

Figure 3.6. SDS-PAGE profile of proso millet protein fraction from two proso millet flour cultivars (Dawn and Plateau). (A) without reducing agents (β-Mecarptoethanol), (B) with a reducing agent (β-Mecarptoethanol). 0- Molecular marker, 1- Dawn albumin fraction, 2- Dawn globulin fraction, 3- Dawn prolamin fraction, 4- Dawn glutelin fraction, 5- Plateau albumin fraction, 6- Plateau globulin fraction, 7- Plateau prolamin fraction, and 8- Plateau glutelin fraction.

Globulin on the other hand composed of several polypeptides with six major peptides of 11, 19, 22, 27, 34, and 42 kDa as well as three minor ones of 12, 14, and 15 kDa. The polypeptides in prolamin range from 11-24 kDa, with three major peptides of 11, 14, and 24 kDa in addition to minor peptides of 20 and 22 kDa. Kohama et al., (1999) observed a major band of 24 kDa and two minor bands of 17 and 14 kDa in prolamin. However, when the author extracted prolamin-like fractions using 70% Isopropyl alcohol containing 0.6% β-ME, they observed the additional higher molecular weight bands of 48 and 72 kDa,

which may suggest that the 24 kDa peptides of prolamin may be a monomer of these peptides (Kohama et al., 1999). Similarly, Parvathy and co-authors observed a range of prolamin peptides between 13 to 27 kDa for proso millet (Parameswaran & Thayumanavan, 1995). Glutelin fractions revealed only a major band of 15 kDa and several minor bands of 11, 22, 24, 78, 209 kDa.

3.3 Conclusion In summary, the results from this study provide information on the physicochemical properties, functional characteristics, and structural particulars of the protein fractions from two cultivars of proso millet, an emerging non-conventional source of plant-based protein ingredient. Additionally, it is clear from the study that proso millet protein fractions had poor solubility, especially the prolamin and glutelin fractions which made up over 75% of the total protein. This suggests a possible reason for their poor functional properties (poor emulsion and foaming properties) since proteins are required to be soluble to facilitate diffusion to the interface of oil, water, and air systems. It was also observed that the solubility of proso millet can best be improved at alkaline pHs ≥ 9. For the most part, there were insignificant differences (p≥0.05) in the physicochemical, functional, and structural properties of the two cultivars of proso millet protein fractions used in this study, albeit a few exceptions with glutelin fractions, which shows significant (p<0.05) differences in solubility. Finally, the ability of proso millet protein fraction to stabilize emulsion could be a potential area of application in the food system owning to their high hydrophobic nature. Further study should consider methods to improve the poor solubility of the proso millet proteins.

Chapter IV: Objective 2 Effects of High-power Ultrasound on the In vitro Digestibility, Physicochemical and Functional Properties of Proso Millet Prolamin and Glutelin Protein Abstract We investigated the effects of high-power ultrasound (50, 75, and 100% amplitude) which corresponds to a power level of 29.29, 38.08, and 52.72 W for 5 and 10 min on the selected physicochemical, function and invitro-digestibility of two proso millet protein fractions (prolamin and glutelin) from the Dawn and Plateau cultivars. The solubility, foam properties, and digestibility were significantly (p < 0.05) higher compared to the native protein. The solubility increased from 8.21 ± 0.13 to 49.20 ± 1.80% for Dawn prolamin, the foam capacity for Plateau glutelin increased from 22.50 ± 4.33 to 34.33 ± 2.75% and the invitro-digestibility of Dawn prolamin increased from 53.71 ± 7.57 to 71.48 ± 1.40%. There was inconsistency in the emulsion properties and particle size distribution.

4.0 Introduction Protein plays various functional and important nutritional roles in food matrices such as acting as a stabilizer in emulsion systems, or as a textural and structural agent in product formulations, and in terms of nutrition, it provides the body with the essential nutrient for growth and development. Proso millet is one of the emerging sources of plant-based proteins, with protein content generally in the range of 11.3 to 15.14% and an abundance of hydrophobic rich proteins having colloidal stability (Devisetti, Yadahally, & Bhattacharya, 2014; Kalinova & Moudry, 2006; Singh et al., 2018; Wouters & Delcour, 2019). Additionally, the consumption of proso millet protein concentrates has been reportedly linked to the elevation of plasma level of high-density lipoprotein (HDL) cholesterol without an increase in low-density lipoprotein (LDL) cholesterol, and effectively reduces glucose and insulin levels in different mice studies (Nishizawa & Fudamoto, 1995; Park, Ito, Nagasawa, Choi, & Nishizawa, 2008; Shimanuki, Nagasawa, & Nishizawa, 2006). The application of proso millet proteins in food is limited because of inadequate information on their functional properties and even where there are, the available information is not related to food product development. Only a few studies are available on the functional characteristics and application of proso millet protein ingredients. Wang et al. (2017) explored the lipophilic attributes of the ethanol-soluble

protein fraction (prolamin) from proso millet in the encapsulation of curcumin and reported the prolamins were better than its water-soluble fractions (albumins) in encapsulation (Wang, L. et al., 2018). Prolamin and glutelin are the major proteins fractions in proso millet making up about 33.9% and 41.8% of the total protein respectively (Wouters & Delcour, 2019). According to Gulati et al. (2017), about 51% of proso millet protein is made up of hydrophobic amino acid residues. Generally, poor viscoelasticity, low- solubility, low digestibility (especially when cooked) and sometimes low colloidal stability have been identified as some of the reasons limiting the application of proso millet protein ingredients in product formulation (Annor et al., 2017; Murugesan, 2015; Wouters & Delcour, 2019).

Various processing techniques are generally applied to modify food proteins to improve their functionality. The application of ultrasound is a widely adopted physical technique to improve the functional properties of food proteins and to generate nano-sized emulsion (O’Sullivan et al., 2017). Ultrasound technology is a propagated acoustic wave above the threshold of human hearing (>16 kHz), such that it causes a longitudinal displacement of the medium in its parts leading to compression and rarefaction of the medium (O’Sullivan et al., 2017). Ultrasound (US) can be divided into two frequency categories: Low- frequency ultrasound (16-100 kHz, power 10-1000Wcm-2) commonly used for the physical and chemical modification of proteins; and high-frequency ultrasound (100 kHz – 1MHz, power <1Wcm-2), commonly used for the evaluation of the physicochemical properties of food (Jiang et al., 2014; Nazari et al., 2018; O’Sullivan et al., 2017). Ultrasound technology modifies protein functional properties majorly through localized hydrodynamic shearing and heating (up to 5000°C) of the protein molecules in solution and consequentially modifying its structure (O’Sullivan et al., 2017). The hydrodynamic shearing is the result of ultrasonic cavitation produced by the ultrasound sonotrode. Ultrasonic cavitation is characterized by rapid build-up and collapse of gas bubbles, generated by localized pressure differentials in wave propagation over a short period of time (O’Sullivan et al., 2017). Several authors have reported the benefits of ultrasound application in improving protein functionality. They reported it improves the solubility of proso millet concentrate (Nazari et al., 2018); lower viscosity, interfacial surface tension, zeta potential and higher solubility of faba bean protein isolate (Martínez-Velasco et al., 2018); and improves the

foaming capacity of soy protein isolate (Morales, Martínez, Ruiz-Henestrosa, & Pilosof, 2015). Therefore, the objectives of the study were to evaluate the effect of high-power ultrasound and application time on (1) the digestibility and (2) some functional properties (solubility, foaming, and emulsion) of the major proso millet proteins (prolamin and glutelin).

4.1 Materials and Methods

4.1.1 Materials Two commercial proso millet cultivars (Plateau and Dawn) were received as a gift from Panhandle Research & Extension Center, University of Nebraska–Lincoln, USA. The samples were dehulled using a lab-scale dehulling machine (Glenn Mills Inc., Clifton, NJ) with some modification (the stationary disc will be replaced with a rubber disc). The dehulled samples were washed to remove dirt and air-dried in an oven at 35°C for 1 day. Samples were then milled using Quadrant Junior Mill (C.W. Brabender Instruments Inc., South Hackensack, NJ).

4.1.2 Extraction and isolation of protein fractions Proso millet flour was defatted using hexane (1:6) for 6 hours to remove lipids, and a mixture of chloroform and methanol (40:60) to remove carotenes and xanthophylls. The defatted proso millet flour was extracted into the prolamin and glutelin fractions (Akharume, Santra, & Adedeji, 2019). The defatted proso millet flour was mixed with 0.5 M NaCl solution and stirred continuously for 2 h followed by centrifugation at 5000 x g for 30 min at 4°C to get rid of albumin (water-soluble protein) and globulin (salt-soluble protein). The residue then washed twice with nanopure water to get rid of the salt content. Glutelin and prolamin were extracted sequentially from the residue using 0.05 M NaOH and 70% Isopropyl alcohol respectively. The different extracts were dialyzed extensively at refrigeration temperature (4°C) followed by centrifugation of the dialysate. The precipitate after centrifugation was freeze-dried and kept at -20°C until use.

4.1.3 Ultrasound treatment of protein fractions 4.1.3.1 Protein fractions Forty milliliters aqueous solution (5% w/v) each for prolamin and glutelin fractions were prepared in a 100ml plastic beaker by dispersing the protein powder in a nanopure water, with continuous stirring for 2h at refrigeration condition (4°C). After stirring, samples were treated with ultrasonication using a Qsonica Ultrasound processor (model 700, Newtown CT, USA) fitted a ½ inches (12.7 mm) diameter sonotrode probe that provides a continuous 20 kHz frequency wave at 100%, 75%, and 50% amplitude for 5- and 10-min treatment time. The pulse duration of (on-time: 4s and off-time: 2s) was used and kept constant throughout the experiment. Samples were kept in ice during treatment to prevent the protein solution from overheating. For control, the sample was only stirred for 2 h with no ultrasound treatment. All samples were then lyophilized and kept at -22°C until use.

4.1.3.2 Determination of ultrasound power and intensity Since the ultrasound energy is lost in the form of heat through the medium during ultrasonication (O’Sullivan et al., 2017), the acoustic power applied to the sample solution was determined by the colorimetric method as described by Margulis & Margulis, (2003). The change in temperature of the aqueous sample with time was used to estimate the acoustic power from equation 4.1 below and was determined to be 29.29, 38.08, and 52.72 W for 50, 75, and 100% amplitude, respectively for a constant frequency of 20kHz:

dT P = M x Cp x ( ⁄dt) (4.1)

Where P is acoustic power (W), M is the mass of the sonicated liquid (g), Cp is the specific heat of the medium under constant pressure (the Cp of water 4.184 J/g/k was used) and dT ⁄dt was determined from the slope of the linear portion of the rate of change of temperature with time. The acoustic intensity is then calculated from equation 4.2 below determined to be 25.47, 33.11, and 45.84 Wcm-2

푃 퐼푎 = (4.2) 푆퐴

2 Where 퐼푎the acoustic power intensity (W/cm ) and 푆퐴 is the surface area of the tip of the transducers (ultrasound emitting surface).

4.1.4 Physicochemical and functional properties characterization 4.1.4.1 Solubility measurement Protein fractions solubility was determined at pH 7. Twenty milligrams of protein sample were dissolved in 10 ml of 0.01 M phosphate buffer and the protein solution was adjusted to the desired pH 7 using either 1 N HCl or 1 N NaOH with continuous stirring for 30 min at room temperature. The protein solution was centrifuged (16000 x g for 5 min at 4°C) and the protein content of the supernatant determined by the Bradford method (Bradford, 1976) with BSA as standard. To quantify the protein in the original sample, prolamin was dissolved in dilute 0.05M NaOH. Protein solubility was expressed as a percentage of the protein content in the supernatant to the protein content in the original sample.

푝푟표푡푒푖푛 푐표푛푡푒푛푡 푖푛 푠푢푝푒푟푛푎푡푎푛푡 푆표푙푢푏푖푙푖푡푦 (%) = 푥 100 (3) 푝푟표푡푒푖푛 푐표푛푡푒푛푡 푖푛 푠푎푚푝푙푒

4.1.4.2 Foaming capacity and foaming stability Foaming capacity (FC) and foam stability (FS) of the protein fractions were assayed according to the method of Akharume, Santra, et al., (2019). Twenty milliliters of 0.5% w/v of protein solution prepared in 0.01 M phosphate buffer (pH 7) was homogenized in a 100 ml plastic cylinder using a Polytron homogenizer equipped with PT-DA-12SM generator (PT 10/35 GT, Brinkmann Instruments Inc., Westbury, NY, USA) at a speed setting of 15400 rpm. for 1 min at 20 °C. The total volume of foam in the measuring cylinder was measured immediately and the residual volume of foam after the solution was left undisturbed for 10 min was measured. FC and FS were computed according to Motoi et al., (2004).

푣표푙푢푚푒 표푓 푓표푎푚 푎푓푡푒푟 푖푚푚푒푑푖푎푡푒푙푦 푎푔푖푡푎푡푖표푛 퐹퐶(%) = 푥 100 (4.4) 푣표푙푢푚푒 표푓 푙푖푞푢푖푑 푏푒푓표푟푒 푎푔푖푡푎푡푖표푛

퐹푆(%) = (푟푒푠푖푑푢푎푙 푣표푙푢푚푒 푎푡 10 푚푖푛/ 푡표푡푎푙 푓표푎푚 푣표푙푢푚푒) 푥 100 (4.5)

4.1.4.3 Emulsifying activity index (EAI) and emulsion stability (ES) Emulsifying properties of the protein samples were measured by the methods of Pearce & Kinsella (1978), with modification. In summary, 1.5 ml of pure canola oil was mixed with

4.5 ml of 1% w/v protein solution dissolved in 0.01M phosphate buffer (pH 7) and homogenized with a mechanical homogenizer (PT 10/35 GT, Brinkmann Instruments Inc., Westbury, NY, USA) for 1 min at 15400 rpm. Fifty microliters of the portion of the homogenized emulsion were pipetted from the bottom of the container at 0 and 10 min after homogenization. Each aliquot sample was diluted with 5 ml of 0.1% SDS solution. The absorbance of the diluted emulsions was measured at 500 nm with a spectrophotometer (SpectraMax M2, Molecular Devices, San Jose, CA). Emulsifying activity index (EAI) and emulsifying stability index (ESI) were calculated from equations 6 & 7 below (Wu, Hettiarachchy, & Qi, 1998).

2 퐸퐴퐼 (푚 ⁄푔) = 2푇 (퐴0 푥 푑푖푙푢푡푖표푛 푓푎푐푡표푟/(퐶 푥 Ф 푥 10,000)) (6)

훥푡 퐸푆퐼 (푚푖푛) = 퐴0 푥 ⁄훥퐴 (7)

Where 훥푡 = 10 푚푖푛 and훥퐴 = 퐴0 − 퐴10 ; T = 2.303, 퐴0 = Absorbance measured immediately after emulsion formation; dilution factor =100, C = weight of protein/unit volume (g/mL) of aqueous phase before emulsion formation; and Ф = oil volume fraction of emulsion

4.1.4.4 Differential scanning calorimeter The thermal properties of the protein samples were assayed using a differential scanning calorimeter (DSC - Q20, TA Instruments, New Castle, Delaware, USA). Fifty mg of protein was dissolved in 1 ml of 0.01M phosphate buffer (pH 7) in an Eppendorf tube for 30 min at 4°C with occasional vortexing. The protein solution was then centrifuged at 16000 x g for 5 min at 4°C and the supernatant pipetted out leaving a solid residue at the bottom of the tube. About 10 - 30 mg of the residue protein was then transferred into a stainless-steel pan and hermetically sealed. Pan containing the sample was scanned against an empty pan (reference) from temperatures 20°C to 140°C at a heating rate of 5°C/min. Onset temperature, denaturation temperature, and change in enthalpy were computed from the thermogram generated by the DSC software.

4.1.4.5 Particle size determination and particle size distribution The particle size and size distribution of the protein were measured using Zetasizer Nano ZS 90 (Malvern Instruments, Malvern WR14 1XZ, U.K.). A 0.5% (w/v) aliquot solution

of the protein sample was prepared and transferred into a measuring cuvette before measurement in the Zetasizer. The viscosity and refractive index (RI) values for the solvent used (Pbs) were 0.8872 cP and 1.330, respectively.

4.1.5 In-vitro protein digestibility Pepsin digestibility of the protein was carried out according to the method of (Gulati et al., 2017; Mertz et al., 1984) with slight modification. Here, 50 mg of the protein was suspended in 1 ml of pepsin enzyme solution (1.5 mg/ml of pepsin in 0.01 M phosphate buffer pH 2.0) and digested for 2 h at 37 ℃. The digestion was terminated by heating the solution at 100℃ in a heat block for 10 min. The protein solution was centrifuged to remove the supernatant and the protein content before and after digestion was determined. The difference in protein concentration before and after digestion expressed as a percentage of the initial protein concentration was recorded as the pepsin digestibility.

4.1.6 Experimental design and statistical analysis. A completely randomized design in which the effect of six treatments (50, 75, and 100% amplitude for 5 and 10 min) and a control (no ultrasound treatment) on nine response variables (solubility, pepsin digestibility, foam capacity, foam stability, emulsion activity index, emulsion stability index, onset temperature, maximum denaturation temperature and change in enthalpy) was evaluated for each of the two cultivars (Dawn and Plateau) and two protein types (Prolamin and Glutelin). The average of three replicates was analyzed by one-way ANOVA using the GLM procedure of SAS (SAS Institute, 2004) to determine the differences in means at a significant level of P< 0.05. Means comparison of significant main effect was carried out using the Tukey test of SAS at a significant level of P< 0.05.

4.2 Results and Discussion 4.2.1 Protein solubility The solubility of a protein is an important physicochemical property that impacts protein functionality. The solubility at pH 7 of the prolamin and glutelin protein fractions following ultrasound treatments is presented in Figure 4.1. We observed that the solubility of US treated protein increased greatly when compared to the native for all types of proso millet protein fractions considered. This increment is observed to increase as the US power increases (29.29, 38.08, and 52.72 W) and treatment time (5 and 10 min). For instance, the

solubility of prolamin from Dawn cultivar after US treatments were 8.21 ± 0.13, 22.1 ± 0.78, 23.9 ± 0.41, 27.5 ± 6.57, 24.3 ± 5.07, 31.1 ± 5.03, and 49.2 ± 1.80% for native, US power 29.29, 38.08, and 52.72 W for 5 and 10 min in that order. The effect of the US treatment on the increment in solubility was significant (p < 0.05) for both the prolamin and glutelin protein types while the type of cultivar only had a significant effect on the prolamin and not the glutelin protein. Other authors have reported that US treatment increased the solubility profile of proteins due to disruption of non-covalent aggregation, and hydrophobic and electrostatic interaction that may hold the protein together in solution (Nazari et al., 2018; O'sullivan et al., 2016) thereby exposing or freeing this aggregate for dissolution and solubilization. Nazari reported a significant increase in proso millet concentrate following the treatment with US power intensities of 18.4, 29.58, and 73.95 Wcm-2 for 5, 12.5, and 20 min. Jiang et al (2014) showed that US treatment improved the solubility of black bean protein by inducing changes in its secondary structure without an associated loss in the molecular weight. A decrease in the alpha helix of the black bean protein and an increase its the beta-sheet was observed (Jiang et al., 2014).

Figure 4.1. The solubility of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. Solubility was carried out at pH 7. All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

4.2.2 Particle size analysis The effects of the US on the protein particle size are depicted in Figure 4.2. US treatment has been well reported to cause disruption and shear protein particle size. We observed that for the prolamin from Dawn cultivar (Fig 4.2a), the native protein exhibited a unimodal distribution, US power of 29.29 for 5- and 10-min treatment exhibited a multimodal distribution, however with the increase in US power and treatment unimodal configuration was observed again except for US power 38.08 W for 10 min. In addition, this new unimodal peak showed a shift towards smaller particle size compared to the native protein. This may mean that at low US power for a short and extended time, the protein particle size becomes easily disrupted. Though we did not see smaller peaks as the US power and time increased, which is unexpected. This may mean that at the US of 38.08, and 52.72 W for 5 and 10 min new soluble aggregates with smaller particle sizes are being formed as can be seen with the shift of the peaks towards smaller particle size. When the power of ultrasound increases soluble aggregates are sheared into smaller size with wider distribution (Jiang et al., 2014). This can be observed for glutelin protein from both cultivars where the native protein showed a unimodal peak while the US treated glutelin showed a multimodal with a wider distribution. Though, the native prolamin from plateau cultivar happened to show a multimodal distribution.

Figure 4.2. Dynamic light scattering analysis of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. (a) Dawn prolamin (b) Plateau Prolamin (c) Dawn glutelin (d) Plateau glutelin All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

4.2.3 Thermal denaturation profile

The knowledge of protein thermal properties provides an understanding of the thermal stability of the protein as well as its cooking behavior. The onset temperature, maximum denaturation temperature, and heat capacities of the prolamin and glutelin proteins are presented in Table 4.1. Since US treatment is accompanied by a large amount of heat, it is expected that proteins subjected to US treatment may have been partly denatured or unfolded as the US induce changes to the protein secondary structure, thus less heat is required for the transition of the protein during DSC scanning. We observed a single peak around 67 - 85℃ for all the treatment except for Plateau prolamin treated at 52.72 W for 10 min, perhaps a total denaturation occurred during its US treatment. For the rest of the treatment, we observe a lower change in enthalpy for the treatments compared to the native protein, except for Dawn glutelin. The changes in onset temperature, maximum denaturation temperature was insignificant (P > 0.05) while the changes in enthalpy were significant for both prolamin and glutelin proteins from the two cultivars.

Table 4.1. Thermal denaturation properties of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour.

Dawn Prolamin Plateau Prolamin Dawn Glutelin Plateau Glutelin

Protei n Onset Max. Chan Onset Max. Chan Onset Max. Chan Onset Max. Chan fracti tempera tempera ge in tempera tempera ge in tempera tempera ge in tempera tempera ge in ons ture ture enthal ture ture enthal ture ture enthal ture ture enthal (℃) (℃) py, (℃) (℃) py, (℃) (℃) py, (℃) (℃) py, ΔH ΔH ΔH ΔH (J/g) (J/g) (J/g) (J/g)

Contr 79.51 ± 82.44 ± 0.025 61.58 ± 67.73 ± 0.147 78.13 ± 81.07 ± 0.02 ± 77.99 ± 81.93 ± 0.114 ol 1.51a 1.13a ± 0.09a 0.06a ± 0.06a 0.27a 0.00a 0.19a 0.44a ± 0.01a 0.02a 0.02a

50% 78.84 ± 80.45 ± 0.001 67.93 ± 74.08 ± 0.044 78.2 ± 82.15 ± 0.11 ± 78.9 ± 82.4 ± 0.062 amp, 0.00a 0.00a ± 0.26a 0.14a ± 0.25a 0.30a 0.01b 0.27a 0.26a ± 5 min 0.00b 0.00b 0.01ab

75% 78.03 ± 80.58 ± 0.016 72.35 ± 76.45 ± 0.026 78.14 ± 81.69 ± 0.074 79.09 ± 81.96 ± 0.04 ± amp, 0.66a 0.03a ± 5.61a 4.52a ± 0.20a 0.21a ± 0.07a 0.03a 0.00a 5 min 0.00ab 0.01b 0.00b

100% 80.96 ± 82.1 ± 0.003 69.51 ± 73.18 ± 0.025 78.11 ± 81.06 ± 0.029 77.96 ± 81.54 ± 0.071 amp, 0.66a 0.07a ± 2.65a 2.91a ± 0.17a 0.29a ± 0.18a 0.12a ± 5 min 0.00b 0.00b 0.01b 0.01ab

50% 78.17 ± 80.21 ± 0.013 76.62 ± 79.25 ± 0.031 78.12 ± 80.96 ± 0.046 78.7 ± 81.97 ± 0.042 amp, 0.46a 0.00a ± 1.01a 1.58a ± 0.70a 0.10a ± 0.26a 0.01a ± 10 0.00ab 0.02b 0.01b 0.00b min

75% 78.04 ± 80.25 ± 0.008 72.16 ± 74.42 ± 0.049 77.94 ± 80.76 ± 0.028 78.12 ± 81.32 ± 0.055 amp, 1.32a 2.23a ± 4.97a 6.30a ± 0.44a 0.22a ± 0.12a 0.04a ± 10 0.00ab 0.01b 0.01b 0.02b

79 min

100% 78.02 ± 80.82 ± 0.017 NA NA NA 75.08 ± 81.05 ± 0.25 ± 76.4 ± 78.6 ± 0.057 amp, 0.24a 0.34a ± 0.83b 0.37a 0.06b 2.11a 2.88a ± 10 0.00ab 0.02b min

*All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

4.2.4 Emulsion properties The emulsion properties were determined using the turbidimetry method by measuring the absorbance of the protein at 500nm. The emulsion activity index (EAI) and emulsion stability index (ESI) of prolamin and glutelin are presented in Figures 4.3 and 4.4. For prolamin protein, US treatment slightly increased the EAI at 5 min treatment compared to the native but as the treatment time increased to 10 min, the EAI was observed to decrease. The increase in the EAI at 5 min treatment may have been due to less aggregation that allows the protein to diffuse into the oil-water interface. Conversely, we observed a decrease in EAI for the glutelin protein when compared to its native protein. The changes in EAI were significant (P < 0.05) for both prolamin and glutelin. The ESI of the protein reflects how well a protein supports rigidly the oil-in-water interface once it’s able to diffuse there. From Figure 4, the US at higher treatment time increased the ESI of the prolamin compared to the native and the US treatment seems not to have much effects on the ESI of the glutelin protein. The changes in ESI were significant (p < 0.05) for both prolamin and glutelin.

Figure 4.3. Emulsion activity index of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

Figure 4.4. Emulsion stability index of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

4.2.5 Foam properties The foaming capacity (FC) and foaming stability (FS) of the prolamin and glutelin protein are presented in Figures 4.5 and 4.6. The foam capacity measures the ability of the protein to trap air bubbles in a liquid or semi-solid phase. US treatment significantly (p < 0.05) increase the FC of the prolamin and glutelin particularly at higher US power and treatment time. US increased the FC from 3.83% to 10.33% for Dawn prolamin and from 22.5% to 34.33% for plateau glutelin. Similar increasing effects of the US on FC of wheat gluten have been reported by (Zhang et al., 2011), where they opined that partial denaturation due to the US may have caused the exposure of hydrophobic residue of the protein aiding the FC. Foam stability of the prolamin and glutelin protein was observed to be significantly (P < 0.05) higher than the native protein for all the treatments, except for the US treatment of Dawn prolamin at 38.08 and 52.72 W for 10 min. For instance, FS increased from 57.5 to 100% after 5 min of 52.72 W treatment. This may have been due to the ability to form a continuous rigid viscoelastic and air-impermeable film to trap the air bubbles.

Figure 4.5. Foam capacity of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

Figure 4.6. Foam stability of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

4.2.6 Pepsin digestibility

The ability of the protease that aid protein digestion in the stomach was tested on the protein fractions from proso millet. The summary of the digestibility of the prolamin and glutelin

fractions are presented in Figure 4.7. In all the treatment considered, the pepsin digestibility was higher for the US treated protein compared to the native ones. For example, the digestibility of Dawn prolamin increased from 53.71 to 71.48 while Plateau glutelin increased from 45.5 to 67.6 from native to 52.72 W for 10 min. The effect of US treatment and variety on the pepsin digestibility was significant (P < 0.05). The US was reported to increase the digestibility of rapeseed napin protein as ultrasound increased from 10% to 40% (Pan et al., 2020).

Figure 4.7. Pepsin digestibility of the prolamin and glutelin protein fractions from two cultivars (Dawn and Plateau) of proso millet flour. All the ultrasound treatments were carried out at a constant frequency of 20kHz and varying amplitude. Amplitude is represented by amp in the figure legend.

4.3 Conclusion In summary, the US in an emerging technology that showed a lot of promise for modifying protein ingredients. Here in this study, we establish that US treatment of 29.29, 38.08, and 52.72 W increased the solubility, foam capacity, foam stability, and invitro-digestibility of the prolamin and glutelin from two cultivars of proso millet flour. We observed that the effect of the US treatment was inconsistent for the emulsion properties perhaps due to the inconsistent in the size distribution of these proteins. However, US treatment-induced partial denaturation to the protein and perhaps exposed it hydrophobic structure as reflected in the low enthalpy change observed with the DSC thermogram.

Chapter V Main Obj: Determination of the Three-dimensional Structure of Glutelin type-B 5-like Protein

Obj 3A: Cloning, Expression, and Purification of His-tagged Glutelin type-B 5-like Protein Isoform from Proso Millet in Escherichia coli Abstract Glutelin type-B 5-like protein is an isoform of glutelin, a major storage protein fraction in proso millet, having over 40% structural identity with the crystal structure of recombinant pro-11S globulin of pumpkin (2E9Q) and crystal structure of pumpkin seed globulin (2EVX) according to the protein data bank (PDB). Proso millet protein is drawing research interest in food applications for their protein quality and sustainability. In this experiment, we cloned the gene of glutelin type-B 5-like protein with attached six continuous histidine codons into a NcoI/HindIII-digested pRSF-NT expression vector and overexpressed in Rosetta Escherichia coli BL21(DE3) cells. The recalcitrant glutelin type-B 5-like protein was recovered from inclusion bodies by unfolding it with 8 M urea and refolding in 0.5 M Arginine with 20 mM CHES (2-(Cyclohexylamino) ethanesulfonic acid) buffer. The refolded and dialyzed protein purified in an immobilized metal affinity column resulted in ~ 90% pure protein based on an SDS-PAGE gel analysis. The protein band was observed at 22.1 kDa on the SDS-PAGE gel consistent with the predicted molecular weight of 22.4. Dynamic light scattering (DLS) analysis of the protein showed that it is monodispersed with a broad range of particle sizes (6.50 – 43.82 nm) while the size exclusion chromatography showed, a higher-order aggregation with molecular weight about 96 kDa.

5.0 Introduction Proso millet is the specie of millet widely grown in the United States of America, majorly in states of Colorado, Nebraska, and South Dakota, with production as much as 9.1 million bushels in 2011 (Michael, 2012; Saleh et al., 2013). The seeds of proso millet are higher in protein content compared to other common cereals (like wheat, rice, and maize) and are nutritionally balanced for human consumption than sorghum (Manley, 2011). Glutelin, the alkaline soluble storage protein fraction is the second-largest storage protein fractions in

proso millet accounting for about 34-39% of the total protein content (Akharume, Santra, et al., 2019). Glutelin from cereals are made into nanoparticles for the encapsulation and delivery of lipophilic drugs in the pharmaceutical industry as well as an encapsulant for fat-soluble bioactive compounds and as a stabilizer for Pickering emulsions in the food industry (Wang, L. et al., 2018; Wouters & Delcour, 2019).

Not much is known about the structural particulars of proso millet glutelin protein. Very recently, the genome of proso millet was sequenced, annotated, and reported (Shi et al., 2019). The genes of three glutelin fractions were identified: two isoforms of “glutelin-2- like” (156 and 161 amino acid residue) and one of “glutelin type-B 5-like” (256 amino acid residues), all of which amounts to a theoretical molecular weight of 16, 17.2 and 27.8 kDa as analyzed by the by ProtParam tool on the ExpASy web server (https://web.expasy.org/protparam/). Earlier experimental study of the glutelin protein reveals two major polypeptides: a 20 kDa polypeptides and another higher than that (Kohama, Nagasawa, & Nishizawa, 1999). Both of them are higher in glutamine and proline amino acid residues (Kohama et al., 1999).

A BLAST search of the glutelin type-B 5-like isoform amino acid sequence against the non-redundant protein in the PDB revealed over 45% identity with the crystal structure of recombinant pro-11S globulin of pumpkin (2E9Q) and crystal structure of pumpkin seed globulin (2EVX) as well as over 35% identity with the structure of 11s proglobulin storage protein from Amaranth (3QAC), the crystal structure of pea prolegumin (3KSC) and crystal structure of peanut major allergen (3C3V). These homologs could provide good structural information for resolving the diffraction data of the glutelin type-B 5-like protein, However, the first step to determining the crystal structure of the glutelin type-B 5-like protein is to obtain a pure protein. Therefore, in this study we identified, cloned, expressed, and purified glutelin type-B 5-like protein as a first step to obtaining its crystal structure.

5.1 Materials and Methods 5.1.1 Construction of glutelin type-B 5-like subunit clone The gene fragments that correspond to the glutelin type-B 5-like protein, residues 1-256 were identified from the genomic DNA of proso millet deposited in the National Centre for Biotechnology Information (NCBI) genomic database (Accession number:

RLM98351). The gene fragment was designed to encode an N-terminal His6-tag with a tobacco etch virus (TEV) protease cleavage site. The synthetic gene fragment was purchased from Life Technologies lab (ThermoFisher) and cloned into a NcoI/HindIII- digested pRSF-NT (EMD Millipore) expression vector (Korotkov, Delarosa, & Hol, 2013) using Gibson Assembly® Master Mix.

5.1.2 Expression of recombinant glutelin type-B 5-like protein and isolation on inclusion bodies The expression vector was transformed into a competent Rosetta BL21(DE3) cells (EMD Millipore) and expressed according to the method of (Wagner, Evans, & Korotkov, 2014) with some modification. The cells were grown in Luria broth (LB) containing kanamycin at 37°C to OD600 0.5 and induced with 0.5M isopropyl-thio-β-D-galactopyranoside (IPTG) overnight at 18°C. Then, the cells were harvested by centrifugation at 6371 x g (Fiberlite F10-6x500y rotor, ThermoFisher Scientific) for 15 min at 4 °C before resuspension in buffer containing 20 mM Tris-HCl pH 8.0, 500 mM NaCl, 5mM imidazole and 1mM 2- mercaptoethanol. The resuspended cells were lysed using an EmulsiFlex-C5 microfluidizer (Avestin) and the lysate was spin at 38828 x g (SS-34 fixed angle rotor, Thermo Scientific) for 50 min to isolate the inclusion bodies (pellet).

5.1.3 Unfolding, refolding and purification of glutelin type-B 5-like from inclusion bodies The pellet containing the inclusion bodies was resuspended in a cold buffer containing 20 mM Tris-HCl pH 8.0, 500 mM NaCl, 2% TritonTM X-100 and spin at 7670 x g (SS-34 fixed angle rotor, Thermo Scientific) for 15min. the recovered pellet is then solubilized and incubated for 1hr in a buffer containing 20 mM Tris-HCl pH 8.0, 8 M Urea, followed by centrifugation for 30 min at 7670 x g (SS-34 fixed angle rotor, Thermo Scientific). The supernatant is carefully pipetted dropwise into a refolding buffer (0.5 M arginine, 20 mM CHES buffer pH 8.5) on ice ratio 1:10, followed by overnight dialysis against a buffer containing 20 mM Tris-HCl pH 8.0, 300 mM NaCl. Glutelin type-B 5-like protein was purified from the soluble fraction of the dialysate by immobilized metal affinity chromatography using Ni-nitrilotriacetic acid (Ni-NTA) resin (Thermo Fisher Scientific)

column. The protein content of the elution was determined using UV-Vis Spectrophotometer (Thermo Scientific™ NanoDrop™ OneC) at an absorbance of 280.

5.1.4 Fast protein liquid chromatography (FPLC) purification and analysis His-tagged glutelin type-B 5-like protein was further purified in a size-exclusion chromatography using a Superdex 75 column (GE Biosciences) equilibrated in buffer containing 20 mM HEPES pH 7.5, 300 mM NaCl.

5.1.5 Western blot Glutelin type-B 5-like protein expression was confirmed by Western blotting using Thermo Fisher Apparatus. The expressed protein was analyzed by 15% SDS-PAGE and electro- transferred onto a nitrocellulose membrane (0.45µm, Biorad). The nitrocellulose membrane was blocked, and antibodies were applied in 5 % milk in PST (1x Phosphate buffered saline, pH 7.4, and 0.1% Tween). 6x-His Tag monoclonal antibody (4A12E4, 1:500 dilutions) was used as the primary antibody and peroxidase-conjugated goat anti- mouse IgG (1:12500 dilutions) was used as the secondary antibody. Immunoblotting images were acquired using a BioRad ChemiDoc MP system.

5.1.6 Dynamic light scattering (DLS) analysis To access the dispersibility of the glutelin type-B 5-like protein, the particle size of the protein in 20 mM HEPES pH 7.5, 300 mM NaCl was measured using a Zetasizer Nano ZS (Malvern Instrument) with particle size measurement range 0.3nm – 10 microns and concentration range between 0.1 µg/ml to 0.4 g/ml. Samples concentrations were maintained at 0.5 mg/ml and the refractive index of dispersant used was 1.33 which corresponds to that of water. The particle size distribution was characterized by the Sauter diameter, D4,3 (volume-weighted mean diameter), and intensity weighted average, Z-value.

5.2 Results and Discussion 5.2.1 Cloning and expression of the recombinant glutelin type-B 5-like The full length of the gene encoding for glutelin type-B 5-like protein (Accession no. RLM98351) synthetically made was cloned into a NcoI/HindIII-digested pRSF-NT (EMD Millipore) expression vector, a T7 promoter with kanamycin resistance. The gene was designed to encode for six continuous histidine

residues to the N-terminal of the protein with a tobacco etch virus (TEV) protease cleavage site. The open reading frame (ORF) consists of a 602 bp double-strand DNA sequence. The NcoI/HindIII-digested pRSF-NT vector was transformed into competent E. coli cells BL21(DE3) (EMD Millipore) and was expressed according to our previously optimized protocol: cells were grown in Luria broth (LB) containing kanamycin at 37°C to OD600 0.5 and induced with 0.5M isopropyl-thio-β-D-galactopyranoside (IPTG) overnight at 18°C. The expression construct was verified by DNA sequencing (Eurofins Genomics) and was identical to the putative protein from P. miliaceum genomic DNA in the NCBI genomic database. The identity of the for glutelin type-B 5-like protein was confirmed by Western blot (Figure 5.1).

5.2.2 Purification of glutelin type-B 5-like protein from inclusion bodies An immobilized metal affinity resin chromatography was used to carry out the purification glutelin type-B 5-like protein following the process of unfolding and folding from inclusion bodies as summarized in the purification steps and yield presented in table 5.1. The soluble and insoluble protein extract of overexpressed glutelin type-B 5-like protein was analyzed on 15% SDS-PAGE and Western blot (Figure 5.1). The band corresponding to the molecular weight of glutelin type-B 5-like protein was observed at 22.1 kDa. A higher band observed at 34.7 was due to the inability of the loading buffer (Sodium dodecyl sulfate- based) initially used to fully linearized the protein, which became clear with a different loading buffer (lithium dodecyl sulfate-based) was used as can be seen in figure 1E. It can be observed that most of the proteins are accumulated in the inclusion bodies, which may suggest irregular folding or incomplete folding or possibly the inability of the E. coli to assist in proper folding (Baneyx & Mujacic, 2004; Ventura & Villaverde, 2006). Therefore, the protein was isolated from the inclusion bodies by unfolding with 8 M urea and refolding in a cold buffer (0.5 M arginine pH 8.5, 20 mM CHES) prior to purification on Ni-NTA resin (Thermo Fisher Scientific) equilibrated in lysis buffer. The bound protein was eluted with 250 mM imidazole and was observed to have a purity of about 90% (as judged by SDS-PAGE) shown in Figure 5.1.

Table 5.1. Summary of purification of recombinant His-tagged glutelin type-B 5-like protein from E. coli

Purification step Total protein Step yield (%) Overall yield Total activity (mg)a (%) (U)

Crude extract 1000.0 100 100 -

Refolding & 20.0 2 2 - his-bind purification a Total protein was isolated from a 300 ml culture after incubation at 18℃ for 20 h.

Figure 5.1. SDS and Western Blot analysis of recombinant Glutelin type-B 5-like protein. (A) SDS-PAGE analysis of soluble and insoluble fractions: M- Molecular marker, 1- soluble, 2- insoluble (B) Western blot for soluble and insoluble fractions: M- Molecular marker, 1- soluble, 2- insoluble (C) Purification stages of Glutelin type-B 5-like protein lysate: M- Molecular marker, 1- uninduced, 2- induced, 3- insoluble, 4-soluble, 5- flow- through, 6- wash, 7-elute (D) Western blot for purification progression of lysate: M- Molecular marker, 1-soluble, 2-insoluble, 3- flow-through, 4- wash, 5-elute and (E) Purification stages of Glutelin type-B 5-like protein from inclusion bodies: M- Molecular marker, 1- insoluble, 2- soluble, 3- soluble in detergents (2% TritonTM X-100 ), 4-soluble in urea, 5- post dialysis, 6- flow-through, 7- wash, 8-elute. The arrow in the figure is pointing to the protein band. The higher band appeared due to the initial loading buffer used and the band became absent after loading buffer was changed as shown in Figure 5.1 E.

Furthermore, the eluted fraction was then purified on a size exclusion chromatography (SEC) using a Superdex 75 Increase 10/300 GL column (GE Healthcare) equilibrated in buffer containing 20 mM HEPES pH 7.5, 300 mM NaCl. Glutelin type-B 5-like protein eluted from the column in a single symmetrical peak (Figure 5.2). However, the elution volume for the protein was observed to correspond to a molecular weight of about 60-70 kDa which signifies some form of higher-order aggregation occurring in the protein solution. Additionally, the protein was observed to continually degrade with time as the fraction volume is collected and concentrated as shown by the various runs in Figure 5.2. The arrow in figure 5.2 points to the products degradation of the protein.

Figure 5.2. FPLC analysis of glutelin type-B 5-like protein eluted from Ni-NTA resin column loaded on a size exclusion column (Superdex 75 Increase 10/300 GL column) at 4℃. Run 1 is the initial sample loaded, run 2 is the concentrated sample volume from run 1 and rum 3 is the concentrated sample volume from run 2.

5.2.3 Dynamic light scattering analysis To evaluate the dispersibility of glutelin type-B 5-like protein in solution, the particle size distribution of the protein was measured using the DLS, and the PSD curve is presented in Figure 5.3. The particle size distribution showed a mostly monomodal distribution (peak 1 of Fig 5.3.) with a broad range of particle sizes (6.50 – 43.82 nm). The average hydrodynamic z-value was 29.26 ± 1.46 with an average polydispersity of 35.76 ± 2.51 % which shows that the particles are not homogenous (Moradian-Oldak, Paine, Lei, Fincham,

& Snead, 2000). Mean D4,3 of 3.4 µm was recorded which reflects some sort of an aggregation. Higher D4,3 is an indication of a PSD curve having a peak in larger particle range (Valencia‐Flores, Hernández‐Herrero, Guamis, & Ferragut, 2013) as can be seen in peak 2 of figure 5.3.

Figure 5.3. Dynamic light scattering analysis of glutelin type-B 5-like protein eluted from the size exclusion column (Superdex 75 Increase 10/300 GL column) at 4℃. Concentration 0.5 mg/ml.

5.3 Conclusions Recombinant glutelin type-B 5-like protein with a molecular weight of 22.1 kDa from proso millet was successfully cloned, expressed, and purified with approximately 90% purity using immobilized metal affinity chromatography and size exclusion chromatography. The glutelin type-B 5-like protein is highly insoluble, and the majority of the protein was only purified from its inclusion bodies following refolding with 8 M urea. The glutelin type-B 5-like protein showed a higher-order aggregation with a wide range of particle sizes within 6.50 – 43.82 nm.

Obj 3B: In silico Analysis and Homology Modeling of Three-dimensional Structure of Glutelin type-B 5-like and Proteins from Proso Millet Abstract The knowledge of the three-dimensional structure of a protein is an important step forward to investigating its structure-function properties. In this research we determined the three- dimensional structure of glutelin type-B 5-like protein, a type of proso millet protein glutelin using comparative homology and investigated its physicochemical properties computationally. The choice of glutelin type-B 5-like protein was informed by the existence of close homologs (over 30%). Our result showed that the structure of glutelin type-B 5-like protein is a protomer that comprises three monomers made up of one jelly- like β-barrel and two extended helix domains with over 35% of the remaining residue as a coil. Instability index value of 62.95 was predicted indicating that this protein may be unstable.

5.4 Introduction Proso millet is a widely grown grain crop in many parts of Africa and Asia where it serves as a staple to the population in this area for making different types of foods (Habiyaremye et al., 2017; Ravindran, 1992). Proso millet is considered a healthy food due to its nutritional quality compared to other cereals and the presence of essential amino acids, phytochemicals (phenols, tannins, alkaloids, flavonoids, and saponins) and micronutrients (Rao, Nagasampige, Ravikiran, & Sciences, 2011; Saleh et al., 2013). It was reported that a feed rich in glutinous proso millet protein concentrates caused an increase in the levels of high-density lipoprotein (HDL) cholesterol in mice fed without an associated increase in the concentration of low-density lipoprotein (LDL) cholesterol (Nishizawa & Fudamoto, 1995). In a similar study, Park et al. (2008) showed that proso millet protein concentrate (MPC) elevated the high-density lipoprotein (HDL) cholesterol and adiponectin levels in obese type 2 diabetic mice by up-regulating the expression of adiponectin and down- regulating tumor necrosis factor-α (TNF-α).

Glutelin is a group of storage protein in proso millet with the second-largest storage protein concentration accounting for about 34-39% of the total protein content (Akharume, Santra, et al., 2019). The glutelin group forms most of the proteins in proso millet protein

concentrates (MPC) if produced by acid/alkaline extraction methods; the MPCs are used as a protein ingredient for food application. Like other cereals proteins, wide applications of MPC in food applications is limited by their low functionality (solubility, emulsion and foam properties) compared to other premium ingredients like similar proteins from soy, pea and dairy (Akharume, Santra, et al., 2019; Wouters & Delcour, 2019). The knowledge of the structure-function relationship of protein is needed to understand how the functionality of MPC, like every other food protein, could be improved. For instance, Tandang-Silvas et al., (2012) compared the three-dimensional structure of 11s proglobulins of amaranth, procruciferin, pumpkin pro-11S, soybean proA1aB1b, soybean proA3B4, and pea prolegumin and showed that the smaller cavity (4167.1 Å3) of 11S proglobulin of amaranth seed was responsible for the higher midpoint denaturation temperature (Tm) values (98.85 – 102.85°C) and thermal stability compared to other 11S proglobulins (Tandang-Silvas et al., 2012). In another study, Pantoja-Uceda et al., (2004) revealed that the presence of a hydrophobic patch (a five methionine residues that partially hides Trp76) at the surface of the three-dimensional structure protein surface may have been responsible for the high emulsifying properties of a 2S albumin protein from seeds of sunflower (Pantoja-Uceda et al., 2004).

The three-dimensional structure of most food proteins including millet proteins are yet to be determined or at least still a daunting task to determine using the available techniques (X-ray crystallography, nuclear magnetic resonance, and cryo-electron microscopy). Nowadays, there are a plethora of computational tools that can be used to predict the three- dimensional structure of proteins, predict their physicochemical properties, and even investigate their interaction with other macromolecules like carbohydrates to some degree of accuracy. The first step to the structural prediction of such proteins begins with the identification of its amino acid (AA) sequence and in some cases, it’s akin homologs (usually AA similarity >30%). Very recently, the genes that code for the AA sequence glutelin storage proteins in proso millet were annotated and reported by (Shi et al., 2019). They identified three glutelin protein subunits: two isoforms of “glutelin-2-like” (156 and 161 amino acid residue) and one of “glutelin type-B 5-like” (256 amino acid residues). Owing to the difficulty, cost and time constraint associated with the empirical determination of the three-dimensional structure of the glutelin storage proteins from proso

millet, we identified the amino acid sequence of glutelin type-B 5-like protein (256 AA residues) of glutelin protein from proso millet and therefore, determined its three- dimensional structures using comparative homology and investigated its physicochemical and functional properties in silico.

5.5 Materials and Methods 5.5.1 Protein sequence identification and analysis The amino acid (AA) sequences corresponding to one of the glutelin type-B 5-like (256AA) proteins were retrieved from the National Centre for Biotechnology Information (NCBI, www.ncbi.nlm.nih.gov) in FASTA format with accession number RLM98351. The AA sequence was searched by protein-protein BLAST against the Protein Data Bank (PDB) proteins to identify templates for model building. The templates with over 35% AA identity were selected for glutelin type-B 5-like protein. A summary of the protein details used for homology modeling is presented in Table 5.2.

5.5.2 Comparative homology modeling Template-based protein structure modeling of glutelin type-B 5-like protein was carried using the Swiss‐Model server (https://swissmodel.expasy.org/)(Guex, Peitsch, & Schwede, 2009; Waterhouse et al., 2018) and the Robetta web server (http://robetta.bakerlab.org)(Song et al., 2013) using the five templates with over 35% sequence identified from the PDB database (IDs: 2E9Q, 3QAC, 3KSC, 3C3V, and 1UD1). The procedure for modeling using the Swiss model and the Robetta server has been documented by (Kim, Chivian, & Baker, 2004; Waterhouse et al., 2018). Unlike Swiss- Model server, the Robetta server returned only the monomer of glutelin type-B 5-like protein and since the active form of glutelin type-B 5-like protein is a trimer, the homo- trimers configuration of glutelin type-B 5-like protein was further modeled into a trimer with Galaxy Homomer web server (www.galaxy.seoklab.org/cgi- bin/submit.cgi?type=PEPDOCK) (Ko, Park, Heo, & Seok, 2012; Shin, Lee, Heo, Lee, & Seok, 2014). Structural comparison of all the models (total of five) generated for by the two structure-building server was evaluated by Partial Order Structure Alignment (POSA).

Table 5.2. A summary of the glutelin type-B 5-like protein details used for the homology modeling

Protein Length Amino acid sequence Ascension number type/ label Glutelin 256 MESSAANPRGRPPSGRGPAKSDEPRQHYAEHHVRARERERSENA RLM98351 type-B 5- LLRETESQQRRQGNDRDYQQADGRGHSPDDQDQMCRMKVTMNLQ like ARPWHQRRPALPAEDEGHEPDRPQLPILNSLQVSVERGTLSQDV AVPPLYYTAGAQSVVYAVRGSARLQLVDNIGAAVFDGELRRGQL LVVPEFFVVLTEAGKDGVEYIAFRNDASPVTSRIAGPGSVLRGL PVGVIAASYNVPAKDAMKLKDSGGGGGGGGTSESEL

server using the root mean square deviation (RMSD) as an index of similarity (Li, Natarajan, Ye, Hrabe, & Godzik, 2014). Assessment of the model quality was performed by the Swiss-Model structural assessment tool to identify the best model quality from all the five models. The best model adjudicated based on Qualitative Model Energy Analysis (QMEAN) (Guex et al., 2009) was selected and used for geometry minimization.

5.5.3 Geometry minimization and model re-assessment Geometry minimization of selected models for glutelin type-B 5-like protein one from each modeling server was carried out by the geometry minimization tool of Python-based Hierarchical Environment for Integrated Xtallography (PHENIX) (Liebschner et al., 2019) with 1000 maximum iteration and five macro-cycles. Evaluation of refined models was assessed by the Psi/Phi Ramachandran plot from PROCHECK (Laskowski, MacArthur, Moss, & Thornton, 1993). The overall quality of the models was assured using Z-scores of the ProSA-web tool (Sippl & Bioinformatics, 1993; Wiederstein & Sippl, 2007) and the Qualitative Model Energy Analysis with distance constraints (QMEANDISco) (Studer et al., 2019). Furthermore, the final model was submitted to the protein model database (PMDB) (http://srv00.recas.ba.infn.it/PMDB/main.php) with PMDB identifier PM0083241.

5.5.4 Physico-chemical and functional characterization For Physico-chemical characterization, theoretical isoelectric point (pI), molecular weight, instability index, aliphatic index, and grand average hydropathy (GRAVY) were computed using the Expasy’s ProtParam server (http://us.expasy.org/tools/protparam.html) (Gasteiger et al., 2005). Secondary structures prediction was done by using the Dictionary of Secondary Structure of Protein (DSSP) (Kabsch & Sander, 1983; Touw et al., 2015). Solvent accessibility surface area (SASA was computed using PyMol and the transmembrane prediction was carried out using TMHMM server (Krogh, Larsson, von Heijne, & Sonnhammer, 2001).

5.6 Results and Discussion 5.6.1 BLAST analysis of protein sequence alignment Five templates were selected for building glutelin type-B 5-like protein. The templates were obtained from the PDB by searching using glutelin type-B 5-like protein AA as the

query and for sequence alignment on the NCBI database. The search query returned about 16 hits (match) with sequence similarity above 30% and only five templates were considered based on sequence similarity above 35% and expect value (E-value). The E- value gives information on the likelihood of getting a hit or protein by chance so that the lower the score the better the template fits our model. A summary of the result from the search query for the templates is shown in Table 5.3. The crystal structure of coconut allergen cocosin (ID: 5WPW) with 38.35% similarity and an E-value of 8e-25 was not selected because its natural state was a homo hexamer which was different from the templates selected (homo-trimers) and the top four above it. The same consideration was used to screen out the crystal structure of Arachin Arah3 Isoform (3C3V).

Table 5.3. A summary of the glutelin type-B 5- like homology templates used for the homology modeling

Protein sequence alignment Identities E- value Positives Gaps Scores Glutelin type-B 5-like protein vs pro- 58/128(45%) 2e-32 90/128 (70%) 1/128 (0%) 124 bits (311) 11S globulin of pumpkin (2E9Q_A)

glutelin type-B 5-like protein vs 11s 48/128(38%) 4e-26 83/128 (64%) 1/128 (0%) 106 bits (265) globulin protein from Amaranthus (3QAC_A) glutelin type-B 5-like protein vs pro- 50/131(38%) 2e-25 82/131 (62%) 1/131 (0%) 104 bits (260) 11s seed globulin

99 from pea (3KSC_A) glutelin type-B 5-like protein vs 47/130(36%) 2e-24 83/130 (63%) 1/130 (0%) 102 bits (254) proglycinin mutant of soybean (1UD1_A) Glutelin type-B 5-like protein vs 47/130(36%) 2e-24 83/130 (63%) 1/130 (0%) 102 bits (253) proglycinin mutant of soybean (1FXZ_A) *Five letters in parenthesis in column 1 are the identity number of the protein in the protein database (PDB)

5.6.2 Model building and assessment

A total of ten models of glutelin type-B 5-like protein was developed (five for each server) using two of the top-ranked and widely used structural building servers (Robetta and Swiss- Model Servers) as ranked according to the community-wide Critical Assessment of Predicted Interactions (CAPRI) (Waterhouse et al., 2018). Unlike the Swiss-Model that begins model building by transferring structural information of a conserved atom from the template onto the model, the Robetta server begins model building by turning the protein sequence into a parsed domain and then predicts the model from the domain (Kim et al., 2004; Waterhouse et al., 2018). Protein residues that cannot be modeled are built by de novo structure prediction by both servers. Inputting both the target sequence (glutelin type- B 5-like) and the templates into the servers generated five models one for each template (Figure 5.4). Unlike the Robetta server, the Swiss-Model server returned models with truncated residues: residue 1-76 and 245-256 were removed by moving the N- and C- terminals to glutamine (77) and Glycine (244), respectively. This is not unexpected as all the templates have five variable regions (I, II, III, IV, V) that were removed from their original AA sequence because the residues in these regions caused disorder in the crystals during experimentation (Tandang-Silvas et al., 2010). These regions are mainly looped as can be seen in Figure 5.5 with all the AA sequences. Generally, the glutelin type-B 5-like protein model reveals a protomer composed of three monomers, which are linked non- covalently as can be seen in figure 5.6. The model tertiary structure contains a jelly-like β- barrel and two extended helix domains. These domains are highly conserved relative to the different 11s or pro-11s globulin proteins of pea, soybean, pumpkin, amaranth, rapeseed, and ara h protein of peanut except that these other proteins have two jelly-like β-barrel, not one (Adachi, Takenaka, Gidamis, Mikami, & Utsumi, 2001). (Tandang-Silvas et al., 2012; Tandang-Silvas et al., 2010). The structure also contains two major variable regions at the terminals.

100

Figure 5.4. Visualization of the superimposed monomers of the glutelin type-B 5-like protein models generated based on five different templates as shown in the legends (A) generated from the Swiss‐Model server and (B) generated from Robetta web server. The color code point to the different models and the arrows show the N- and C- terminal.

Figure 5.5. Visualization of the monomers of the glutelin type-B 5-like protein showing the variable regions. (A) generated from Swiss‐Model server with the variable region removed and (B) generated from Robetta web server with variable region intact.

101

Figure 5.6. Visualization of the three-dimensional structure of the glutelin type-B 5-like protein model geometrically minimized using PHENIX (A) Model generated from the Swiss‐Model server and (B) Model generated from Robetta web server. The color code point to each chain of the monomer.

To evaluate the ten models generated from the servers, the RMSD of the superimposed models (Figure 5.4) from each of the servers showed a good fitting of the protein main- chain with minimal deviations of 1.16Å and 2.86Å for models from the Swiss-Model and Robetta servers, respectively. The higher RMSD value for Robetta may have been contributed by the disordered variable region as can be seen in Figure 5.4. Furthermore, model quality assessment was done by QMEAN Z scores (Benkert, Biasini, & Schwede, 2011) from using the Swiss-Model structural assessment tool (Guex et al., 2009) as presented in Table 5.4. The closer this value to zero the more the predicted model fits the experiment structure of similar size. Thus, the model generated from 2E9Q template was further used for geometry minimization.

102

Table 5.4. Summary of glutelin type-B 5-like protein models quality assessment scores

Model Swiss-Model Server Robetta Server template

Mol Clash Ramachandra QMEAN Z Mol Clash Ramachandra QMEAN probity favored score score probity favored score Z score score (%) score (%)

2E9Q 2.27 9.61 94.01 -2.67 1.23 1.04 93.4 -0.81

3QAC 1.66 1.93 89.22 -4.06 1.27 0.78 90.87 -1.59

3KSC 1.82 1.69 92.42 -3.27 1.35 1.30 91.73 -1.10

103 1UD1 1.38 0.65 93.21 -3.53 1.32 1.30 92.44 1.25

1FXZ 1.82 4.66 88.89 -3.37 1.49 2.85 93.7 -2.00

5.6.3 Model geometry minimization, re-assessment, and validation To improve the structure of glutelin type-B 5-like protein, the two models following the initial quality assessment were subjected to geometry minimization using PHENIX (Liebschner et al., 2019) with 1000 maximum iterations and five macrocycles. Geometry minimization relaxes the backbone bond angle by allowing full flexibility of the bond angle and length bonds thereby improving the energy landscape of the protein (Conway, Tyka, DiMaio, Konerding, & Baker, 2014). The models following geometry minimization are presented in Figure 5.6 and were re-evaluated using Ramachandra plot, Z-scores, and QMEANDisCo Scores as presented in Table 5.5 and Figure 5.7. Although the Z-scores seem not to have improved, the bond and bond angles improved significantly as can been seen in the Ramachandra plots from 94.01% and 93.4% to 99.8% to 98% for the Swiss- Model and Robetta Model, respectively. Comparing the minimized structure to unminimized ones showed that there were zero bad angles and bad bonds for both minimized models and 6 bad bonds, 97 bad angles and 3 bad bonds, and 15 bad angles for both unminimized models respectively.

104

Table 5.5. Summary of glutelin type-B 5-like protein geometry minimized model’s quality assessment scores

Model Swiss-Model Server Robetta Server Ramachandra Plot (%) Ramachandra Plot (%) Allowed Generally Disa Z- QMEANDisCo Allowed Generally Dis Z- QMEA allowed llow scores scores allowed allo scores NDisC ed we o d scores Glutelin 99.8 0.7 0.0 -5.44 0.62 ± 0.05 98 1.4 0.5 -5.11 0.48 ± type-B 0.05 5-like

105

Figure 5.7. Ramachandra plots of geometry minimalized glutelin type-B 5-like protein models (A) model from Swiss-Model server (B) model from Robetta server.

5.6.4 Physico-chemical and functional characterization The result of the physicochemical properties of glutelin type-B 5-like protein is presented in Table 5.6. The instability index is above 40, which is an indication that the protein may be unstable. Our earlier attempt at purifying the recombinant protein of glutelin type-B 5- like confirms this instability (data not shown). We observed higher-order aggregation of the protein as well as degradation with time. The secondary structure prediction showed that about 35.81% of the protein is made up of coils. This is obvious when looking at the three-dimensional structure of the protein from Figure 5.5a and this coil may have contributed to the instability of the protein.

106

Table 5.6. A summary of physicochemical properties and secondary structure assignment for glutelin type-B 5-like protein

Physical and chemical parameters Secondary structure prediction (%) Molecular weight (kD) 27.80 Alpha helix 4.55

Theoretical isoelectric point 6.33 310 helix 2.47 Instability index 62.95 Pi helix 0.00 Aliphatic Index 70.12 Beta bridge 0.39 Grand average of hydropathicity -0.714 Beta sheet 23.82 (GRAVY) Extinction coefficient 15930 Turn 14.32 Number of negatively charged 34 Bend 18.62 residues Number of positively charged 32 Random coil 35.81 residues Solvent Accessibility Surface 84845.2 Ambiguous states 0.00 Area (SASA) Å2 Transmembrane regions 0 Other states 0.00

5.7 Conclusions In summary, the three-dimensional structure of glutelin type-B 5-like protein, a type of glutelin protein from proso millet was successfully predicted using the Swiss-Model server and the Robetta Server. The two models obtained from the servers showed a good agreement in their main backbone and showed conserved domains with the globulin group of proteins from pea, amaranth, rapeseed, and peanut. Ramachandra plot of the protein structure gave over 98% of residues in the favorable region. Physicochemical analysis of the proteins showed it may be unstable particularly because most of its secondary structures were coils. This is a great contribution to the understanding of the structure-function properties of proso millet glutelin proteins as the knowledge of its three-dimensional structure could elucidate the reason for its poor solubility.

107

Chapter VI: Objective 4 In-silico Modeling of Glutelin type-B 5-like from Proso Millet Seed Storage Protein: Effects of Temperature and Electric Field Abstract The effects of simulated temperature (300, 350, and 400 K) and static electric field (0.1, 1, 3 V/nm) on the changes in secondary structures, solvent accessibility surface area (SASA), the radius of gyration, root mean square deviation (RMSD), root mean square fluctuations (RMSF), and secondary structure analysis of glutelin type-B 5-like protein, a type of glutelin protein from proso millet, in solution was investigated using a molecular dynamics simulation approach. The result showed that the secondary structure of the protein was not disrupted at all temperatures and static electric field levels, although there was the loss of residues from the secondary structure elements and formation of new secondary structure elements. The RMSD increased significantly with temperature and static electric field levels and with simulation time while both the Rg and SASA decreases with simulation time. Besides, the SASA value decreased with temperature but not the static electric field.

6.0 Introduction Proteins are the most complex of the three major dietary macro-molecules. They are made up of basic amino acids that form into a three-dimensional structure with varying surface- active properties that govern their interaction with other molecules (carbohydrate, lipids, air, and water) in a food matrix. The three-dimensional structures of food proteins dictate their biological and sometimes processing functionality; and more importantly their conformational disposition (native, intermediate or denatured state) in response to applied processing stresses (temperature, pressure, shear and force fields) or ambient environmental conditions (pH, ionic strength of the solution, and presence chaotropes). The structure-function relationship has remained a central dogma to elucidating the mechanism of protein functionality at the molecular level; such understanding provides information to identify processes and technologies to improve protein functionality and/or suppress undesirable functionality (allergenicity and unwanted flavor). Pantoja-Uceda et al., (2004) reveals that the presence of superhelix motif, especially “hypervariable loop” in

108

the SFA-8 structure of sunflower protein, which is also common to other 2S albumin proteins may have been responsible for the allergenicity of the sunflower proteins (Pantoja- Uceda et al., 2004). Fukuda et al., (2008) compared the three-dimensional structures of 7s globulin proteins from Adzuki beans (7S1, 7S3), soybeans (β-conglycinin β, β-conglycinin α/c), and mung beans (8Sα) and revealed that the differences in their thermal stability and solubilities are related to the size of the cavity found inside the molecules of each of these proteins (Fukuda et al., 2008). Furthermore, they showed that the difference in electrostatic surface potential distribution may cause differential solubility pH 8- while 7S1, 7S3 and 8Sα were soluble at pH 8, β-conglycinin β and β-conglycinin α/c were not (Fukuda et al., 2008). While the three-dimensional structure of a few food proteins has been determined and deposited in the protein database (RCSB-PDB), the investigation of the structure- function relationship of food proteins is still elusive. One major reason for this is that most food protein three-dimensional structure is determined in only static conformation using X-ray crystallography, as such it may be difficult to know their dynamic properties or how these properties change with structural conformations due to applied processing stresses. Several empirical studies on the structure-function of food proteins have been limited observational changes in structure or functionality at the macro scale without a deep understanding of structural changes at the molecular level. Molecular dynamic (MD) modeling of food protein is gaining research attention. Unlike traditional empirical investigations, MD provides an elucidation of structural changes at the atomic and molecular levels under dynamics simulated processing stresses. Several authors have studied the structure-function relationship of food protein using molecular dynamics modeling. Vagadia and co-authors investigated the effects of temperature (300k to 393k) and oscillating electric field (0.5 V/nm and 2.45 GHz) of the Soybean Trypsin Inhibitor (STI) in-silico. The study showed that the effects of the external stresses disrupted the amino acid residues (31-41, 61-66) and a rearrangement of the secondary structure (coils and turns), but the core of the STI was still stable due to antiparallel β-sheet structure (Vagadia, Vanga, Singh, & Raghavan, 2016). In another earlier study, a myoglobin molecule in water exposed to pulsed and static electric fields (108 to 109 V/m) were examined using MD; the authors reported that a remarkable structural rearrangement in the myoglobin molecules can be achieved by the application of 109 V/m pulsed electric field

109

for a few picoseconds or static electric field for a few nanoseconds (Marracino, Apollonio, Liberti, d’Inzeo, & Amadei, 2013). It is confirmed by several authors that only a high field electric strength is capable of causing a major disruption in the secondary structure of protein molecules (Marracino et al., 2013; Singh, Orsat, & Raghavan, 2013; Wang, Li, He, Chen, & Zhang, 2014). In this study, we investigated the effects of temperature with a static electric field on glutelin type-B 5-like protein, a type of glutelin protein from proso millet. We evaluated three levels of temperatures (300, 350, and 400k) and three levels of static electric fields (0.1, 1, 3 v/nm) on the root mean square deviation (RMSD), root mean square fluctuations (RMSF) per residue, radius of gyration (Rg), secondary structure analysis, and solvent accessibility surface area (SASA).

6.1 Materials and Methods 6.1.1 Molecular dynamic simulations The MD Simulations of glutelin type-B 5-like in water was carried out by the Groningen machine for chemical structure (GROMACS) software package (Version 2018.1, Stockholm Center for Biomembrane Research, Stockholm, Sweden)(Van Der Spoel et al., 2005). The glutelin type-B 5-like protein used in this study was developed in a previous study and can be accessed in Appendix A of this dissertation or downloaded from the protein model database (PMDB) with the identity number of PM0083241 where we had deposited it. The protein is a protomer of three non-covalently linked monomers with each monomer chain comprising 5.0% alpha-helix, 2.5% 3/10 helix, 23.8% beta-sheet and 69.17% coils/turns/bends/bridges (Figure 6.1a). The protein-containing 11574 atoms from 768 residues were kept in a cubic box of a dimensional 11.683 x 11.683 x 11.683 nm as a periodic boundary condition and the protein was solvated with 146262 atoms of water molecules neutralized with 6 sodium ions (Figure 6.1b).

110

Figure 6.1. A snapshot of the three-dimensional structure of glutelin type-B 5-like protein molecule (A) in a vacuum (B) in neutralized water

The OPLS-AA/L all-atom force field (Robertson, Tirado-Rives, & Jorgensen, 2015) and SPC/E water model (Kusalik & Svishchev, 1994) were selected to provide potential energy function and water parameters to the system. Following neutralization of the systems to mimic the physiological state of the protein, the systems were energy minimized to converge at maximum force value < 1000.0 kJ/mol/nm using stepwise descent minimization algorithm for 100, 000 steps and the systems was further equilibrated to a constant number of particles, volume and temperature (NVT) and a constant number of particles, pressure, and temperature (NPT) for 200 ps at 300k and I bar. The MD simulations were run for 1 ns using a leap-frog integrator algorithm during which the temperature of the systems was maintained using modified Berendsen thermostat, and the pressure was maintained using the Parrinello-Rahman barostat (Berendsen, Postma, van Gunsteren, DiNola, & Haak, 1984; Parrinello & Rahman, 1980). A total of nine simulations was run for static electric field levels coupled with the temperature levels (Table 6.1.). The results of simulations such as RMSD, Rg, and SASA were analyzed using GROMACS inbuilt tools. GROMACS inbuilt Dictionary of Secondary Structure of Protein (DSSP) (Kabsch & Sander, 1983; Touw et al., 2015) was used for the secondary structure analysis

111

and virtual molecular dynamics (VMD)(Humphrey, Dalke, & Schulten, 1996) was used to visualize protein conformational change.

Table 6.1. Summary of simulation conditions used in the study

Temperature Pressure Static electric field Simulation length 0.1 V/nm 300 K 1 bar 1 V/nm 1 ns 3 V/nm 0.1 V/nm 350 K 1 bar 1 V/nm 1 ns 3 V/nm 0.1 V/nm 400 K 1 bar 1 V/nm 1 ns 3 V/nm

6.2 Results and Discussion 6.2.1 Root Mean Square Deviation (RMSD) The root mean square deviation is a measure of the changes in the conformation of the protein due to the application of external processing stresses such as temperature, pressure, and electric field. It compares the protein new conformation to its native or initial conformation as shown by the equation:

1 푅푀푆퐷 = √ ∑푁 | 푟 (푖) − 푟 (푖) |2 (6.1) 푁 푖=1 푓푖푛푎푙 푖푛푖푡푖푎푙

Where 푟푓푖푛푎푙 is the final coordinates of atom i, and 푟푖푛푖푡푖푎푙 is the initial coordinate of the atom i, and N is the number of atoms. The average of the RMSD for glutelin type-B 5-like protein after 1ns simulation from different combinations of temperature and static electric field levels are presented in Table 6. 2 and Figure 6. 2. The average values of RMSD for the glutelin type-B 5-like protein ranged from 0.28 to 0.61 across temperatures of 300- 400K for both static and no-electric field treatments. It can be observed that the RMSD increases with temperature (Figure 6.2a-c) and decrease with a static electric field, except at 300K. Both the temperature and static field has direct significant effects (P<0.0001) and

112

interaction effects on the RMSD of the protein (P < 0.001). Additionally, the RMSD was observed to increases with simulation time for all treatments. Static electric field caused a higher RMSD were compared to no electric filed for all levels of temperatures while electric filed at 0.1 V/nm resulted in the highest RMSD at 350K and 400K except for 300K where it showed the lowest RMSD. One could conclude, judging from the close values of RMSD at 300K for static and no electric field that the effects of low field static electric field is less pronounced at low temperature. This phenomenon has been reported by Singh et al., (2013) where they showed a static electric field of 0.002 V/nm and 0.004V/nm had no effects on the stability of the structure of soybean hydrophobic protein at 300K. Vanga et al., (2015) also reiterated this observation where they revealed that both static (0.05 V/nm) and oscillating electric field (0.05 V/nm, 2450MHz) at 300K had no significant effects on the RMSD of Ara h 6 peanut protein allergen.

113

Treatments RMSD (nm) Rg (nm) SASA (nm2) Vol (nm3) Density (g/l) 300K, 0 V/nm 0.28 ± 0.05a1 3.25 ± 0.02a1 495.70 ± 5.38a1 167.02 ± 1.28a12 829.07 ± 6.35a12 300K, 0.1 V/nm 0.29 ± 0.06b2 3.22 ± 0.01b1 490.94 ± 3.40b1 166.69 ± 1.12b2 830.72 ± 5.56b1 300K, 1 V/nm 0.32 ± 0.08c13 3.21 ± 0.02c1 503.71 ± 6.50c1 167.56 ± 1.19c2 826.44 ± 5.87c1 300K, 3 V/nm 0.30 ± 0.06d23 3.23 ± 0.01d3 492.81 ± 3.82d1 166.28 ± 1.33d1 832.79 ± 6.64d12 350K, 0 V/nm 0.40 ± 0.09a1 3.23 ± 0.02a1 495.55 ± 6.96a1 167.09 ± 1.54a12 828.75 ± 7.62a12 350K, 0.1 V/nm 0.46 ± 0.10b2 3.24 ± 0.04b1 484.91 ± 17.21b1 166.24 ± 2.07b2 833.07 ± 10.41b1 c13 c2 c1 c2 c1 114 350K, 1 V/nm 0.39 ± 0.09 3.18 ± 0.04 471.00 ± 13.43 164.72 ± 1.64 840.71 ± 8.37

350K, 3 V/nm 0.45 ± 0.12d23 3.20 ± 0.03d3 476.08 ± 15.57d1 164.98 ± 2.05d1 839.41 ± 10.39d2 400K, 0 V/nm 0.54 ± 0.12a1 3.23 ± 0.06a1 462.85 ± 26.15a1 163.65 ± 3.27a12 846.48 ± 16.88a12 400K, 0.1 V/nm 0.61 ± 0.16b2 3.22 ± 0.04b1 479.99 ± 15.84b1 165.33 ± 2.23b2 837.69 ± 11.24b1 400K, 1 V/nm 0.57 ± 0.13c13 3.23 ± 0.02c3 485.62 ± 15.03c1 166.18 ± 2.21c2 833.38 ± 11.11c1 400K, 3 V/nm 0.59 ± 0.14d23 3.30 ± 0.06d3 483.40 ± 20.29d1 165.57 ± 2.61d1 836.52 ± 13.11d32

114

115

Figure 6.2. Root mean square deviation of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time.

6.2.2 Root Mean Square Fluctuation (RMSF) The root means square fluctuation (RMSF) reveals information on the flexible and rigid regions of the protein structure. The RMSF of the protein per residue basis at the end of 1 ns simulation is presented in Figure 6.3. From Figure 6.3. It can be observed that higher fluctuations were recorded at a high temperature of 400k and the least at 300K and these fluctuations increase with temperature and not the static electric field. The N- and C- terminals of the proteins showed a high degree of fluctuations consistently with all the treatments for all the chains. The regions around the two distant alpha helices (R60 – A65 and A233 – K240) of the protein showed a higher RMSF compared to the beta-sheet regions (D103 – H106, S122 - L128, V132 – A133, Y138 – T140, S145 – V159, A165 – L171, Q175 – V179, T187 – R200, and T207 – R 209). The average fluctuations for all residues recorded for the treatments are 0.05 ± 0.02, 0.05 ± 0.02, 0.05 ± 0.02, 0.05 ± 0.02 for 300K, 0 – 3 V/nm; 0.06 ± 0.03, 0.06 ± 0.03, 0.06 ± 0.03, 0.06 ± 0.03 for 350K 0 – 3 V/nm; and 0.07 ± 0.03, 0.08 ± 0.04, 0.08 ± 0.03, 0.08 ± 0.03 for 400K 0 – 3 V/nm.

116

117

Figure 6.3. Root mean square fluctuations of glutelin type-B 5-like protein per residues under different simulation conditions (temperature and static electric field) after 1 ns simulation time

6.2.3 Secondary structure analysis The secondary structure of the protein molecule was analyzed using the inbuilt DSSP tool of GROMAC. The changes in the number of residues counts that contributes to the secondary structure profile of the protein after 1 ns of different treatment levels are presented in Figure 6.4 while the percentage distribution of the secondary structures is presented in Table 6.3. It is expected that following the application of processing stresses that cause unfolding and denaturation of the protein, the unstable structure such as coil, turn and bend should increase while the more stable structure such as helices and sheets should decrease in the count. From figure 6.4, it can be observed that the increase in the number of residues for coils and bends was more obvious at 400K with a corresponding decrease in the number of residues for β-sheets and α-helices. There was a progressive increase in the number of residues for bend with temperature and the effects of the static electric field were clearly observed. It is possible that short time simulation may not have been enough to cause an irreversible unfolding or permanent denaturation of the protein secondary structure which may explain the reason for the absence of a progressive or regressive trend in the behavior of the secondary structures of the protein.

118

The snapshot of the molecule after simulation (Figure 6.5) did not show an obvious disruption to the secondary structure when comparing all treatments. However, the changes in the percentage distribution profile of the secondary structure (Table 6.3.) points out that some of the residues that make up the secondary structure were lost without necessarily disrupting the secondary structure confirmation. In addition, the interchange between turns structure and helices was seen was observed with simulation time (data not shown). For instance, we analyzed the secondary structures of the protein at 300K temperature level with STRIDE (Heinig & Frishman, 2004) (data not shown) and observed that the alpha- helix in the region of P114 – L119 on chain A lost one residue to become a helix with region I115 – L119 after treatment with a static electric field of 0.1 V/nm. No loss of residue was observed with treatment 1 and 3 V/nm. Also, an alpha helix formed on chain B in the region V222-Y229 when a static electric field of 0.1 V/nm was applied to the protein molecule. This was disrupted to a 310 helices when the static electric field changed to 1 and 3 V/nm (both at region V222-224) losing 3 residues as turns.

119

Table 6.3. Percentage of secondary structure element per the three-dimensional structure of glutelin type-B 5-like protein molecule after1 ns simulation under different simulation conditions.

Treatments Coil (%) Beta-sheet Beta- Bend (%) Turns (%) Alpha helix 3-helix (%) (%) bridge (%) (%)

300K, 0 V/nm 37.30 23.10 0.85 17.6 11.66 6.53 2.92

300K, 0.1 V/nm 36.3 23.9 0.97 18.6 12.2 5.35 2.58

300K, 1 V/nm 37.27 22.59 1.06 17.65 12.84 4.80 3.78

300K, 3 V/nm 37.59 22.16 1.01 18.98 12.38 5.41 2.46

120 350K, 0 V/nm 36.88 23.07 0.85 19.04 11.43 5.67 3.01

350K, 0.1 V/nm 37.56 23.04 1.11 18.30 12.07 5.71 2.10

350K, 1 V/nm 36.71 23.65 0.80 19.28 11.03 5.75 2.72

350K, 3 V/nm 36.36 23.05 0.87 18.73 10.86 7.64 2.23

400K, 0 V/nm 36.44 22.03 1.30 20.96 10.41 5.91 2.82

400K, 0.1 V/nm 38.74 21.77 0.88 19.24 9.37 7.00 2.89

400K, 1 V/nm 38.15 22.13 1.17 20.32 10.96 4.40 2.84

400K, 3 V/nm 38.21 22.16 0.84 21.39 10.93 3.60 2.80

Figure 6.5. Snapshots of glutelin type-B 5-like protein molecule after1 ns simulation under different simulation conditions.

6.2.4 Total Dipole Moment The dipole moment is a good measure of the behavior of protein to the external electric field (Marracino et al., 2013). It gives information on the polarization response of the protein to the external electric field and any shift in the total dipole moment is an indication of protein denaturation or transition in the protein secondary structures (Marracino et al., 2013; Wang et al., 2014). The total dipole moment of glutelin type-B 5-like protein frames with simulation time is presented in Figure 6.6. It can be observed that as early as 100 ps, there were already deviations from the non-electric field treatment for temperatures 300K, 350K and 400K, indicating an early transition in the secondary structure of the protein, which became more pronounced between 400 to 600 ps and much more between 800 to 1000 ps. The static electric field of 0.1 V/nm was observed to show the least deviation at all temperature levels. Both temperature and static electric field have significant effects (p < 0.05) on the differences observed in the total dipole moment.

121

122

Figure 6.6. Total dipole moment of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) after 1 ns simulation time.

6.2.5 Radius of gyration (Rg) The radius of gyration provides information on the spread of the atoms of the protein molecule with respect to its center of mass. This information is useful for understanding if the proteins compact into a globular state or unfold. The higher the Rg the less compact the protein and vice-versa (Fenwick et al., 2019). The Rg is calculated from the equation below:

1 푅푔 = √ ∑푁 | 푟 (푖) − 푟 |2 (6.2) 푁 푖=1 푐푒푛푡푟푒

Where 푟 (푖) is the coordinate of atom i, 푟푐푒푛푡푟푒 is the coordinate of the center of mass of the protein, and N is the number of atoms. The Rg was observed to reduce with simulation time for all the treatments which indicate that the protein compacts over time with temperature and heat treatment. Similar phenomena of a decrease in Rg with simulation time were reported for peanut allergen (Ara h 6 protein) (Vanga, Singh, & Raghavan, 2015) and the Rg of soybean trypsin inhibitor protein was also reported to decrease because of heating. Heating of protein in solution at high temperature (80-100℃ for pea protein) has been known to cause new aggregation through hydrophobic interaction following the revelation of hydrophobic residue during heating (Chao & Aluko, 2018). This behavior was observed for all the treatments where there was an initial increase in the Rg between 0 – 200 ps and then followed by a downtrend in Rg for the rest of the simulation time (Figure 6.7). This could be explained by the initial unfolding of the protein leading to a high Rg followed by new aggregation for the rest of the simulation time. The average of Rg for all treatments is presented in Table 6.2. For no electric field the Rg decrease significantly (P<0.05) with increasing temperature and though there were significant differences in Rg with the static electric field and the interaction of the two factors, close observation of the data in Table 6.2 does not show a steady decrease or increase in Rg with the static electric field.

123

124

Figure 6.7. Radius of gyration of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time

6.2.6 Solvent Accessibility Surface Area (SASA) The solvent accessibility surface area is one of the surface-active properties of the protein that impacts its functionality such as surface hydrophobicity which in turn affects the protein solubility, emulsifying, and foaming ability (Akharume, Santra, et al., 2019; Nakai, 1983). Hydrophobicity of protein is inversely related to its SASA (Gromiha, Nagarajan, & Selvaraj, 2019). The values of SASA at the end of 1 ns simulation are presented in Table 2 while the trends with simulation time are presented in Figure 6.8. From Figure 6.8, the SASA trends downward as the simulation progresses, confirming that the protein compacts over time, as seen by the trend observed with the radius of gyration (Figure 6.7). Similar to the Rg, there was an initial increase in SASA about 0 – 100 ps which may indicate the initial unfolding due to applied treatment, causing an increase in SASA. Later a decrease in SASA with simulation time was observed, which may have been due to the compaction of the protein molecule caused by hydrophobic aggregation. Vanga et al., (2015) reported a similar downward trend in SASA with simulation time as a result of temperature and electric field treatment Ara h 6 peanut allergy. Increasing the temperature from 300K to 400K causes a decrease in the average value of SASA at 0 and 0.1 V/nm static electric field but at 1 and 3 V/nm static electric field, a decrease was observed from 300 to 350K and increased again at 400K. The changes observed due to temperature were significant at p <0.0001 but changes observed due to the static electric field were not at p < 0.05.

125

126

Figure 6.8. Solvent accessibility surface area of glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time

6.2.7 Volume and density The volume and density of the glutelin type-B 5-like protein was stable with simulation time at 300K for all static electric field treatment levels considered with slight variations. However, at a temperature of 350 and 400K, a downward trend in the volume with simulation time was observed for all static treatment levels (Figure 6.9b and 6.9c). At all levels of static electric field treatment, the average volume was observed to decrease from 300K to 400K while the density as expected increased in reverse order. This further confirms that the protein molecules compact as the temperature were ramped up from 300 to 400K. The changes observed in volume were significant (P < 0.05) for temperature, static electric field, and their interaction.

127

128

Figure 6.9. Volume changes observed for glutelin type-B 5-like protein under different simulation conditions (temperature and static electric field) for 1 ns simulation time.

4.0 Conclusions We successfully evaluated the effects of temperature in combination with three levels of static electric field on the behavior of glutelin type-B 5-like protein in a molecular dynamic simulation environment. The simulated stresses caused a significant increase in the RMSD (0.28 to 0.61 nm across temperatures of 300-400K) and a decrease in both the Rg and SASA, implying that the glutelin type-B 5-like protein is compacting under these simulated conditions. Additionally, while the stresses lead to the perturbation of the atomic trajectory of the protein with the evolution of time, it however did not cause a major disruption of the protein secondary structure. Longer MD simulation may be needed to cause the disruptions of the protein secondary structure, which can provide more information on how the surface- active properties (such as SASA and Rg) may be affected. The temperature had significant effects on the changes observed for all parameters measured (RMSD, SASA, and Rg) while the effects of the static electric field were insignificant for SASA.

129

Chapter VII: Objective 5 In-silico Modeling of Glutelin type-B 5-like from Proso Millet Seed Storage Protein: Effects of Temperature and Pressure Abstract Molecular dynamic (MD) simulation provides an insight into the behavior of a protein under applied processing at the molecular level. Here we studied the behavior of glutelin type-B 5-like protein, a type of glutelin protein from proso millet, in solution under different temperatures (300, 350, and 400 K) and pressure (1 bar, 3 kbar, and 6 kbar) levels using a molecular dynamics simulation approach. The combined treatment effect (400K, 6 kbar) increased the compaction of the protein compared to the level at (300K, 1 bar) as shown by the decreased radius of gyration values from 3.26 ± 0.07 to 2.92 ± 0.06 nm, decreased solvent accessibility surface area from 327.47 ± 2.66 to 311.06 ± 5.31 nm2 and decreased volume from 108.35 ± 1.00 to 105.04 ± 1.16 nm3. The root mean square deviation (RMSD) and root mean square fluctuations (RMSF) increased significantly with temperature but only the RMSD decreased with pressure. A snapshot of the three- dimensional structure of the protein revealed compression of its occluded cavities at higher pressure levels but no obvious disruption to the secondary structure elements of the protein was observed, except for the loss of a few amino acid residues that comprise the secondary structure element.

7.0 Introduction Molecular dynamic (MD) modeling simulation is gaining traction in elucidating changes in structural properties or conformation of food proteins as they subjected to external food processing stresses (Fenwick et al., 2019; Vagadia et al., 2016; Vanga, Singh, & Raghavan, 2018). Traditionally, food proteins are subjected to various physical (heating, high- pressure processing, ultrasound, and extrusion), chemical (glycosylation, and phosphorylation) and biological (hydrolysis, crosslinking, and fermentation) modification and/or processing techniques to improve their functionality or suppress undesired property, by exploring the bio-physicochemical and structural properties of these proteins. Application of such processing techniques leads to changes in the native conformation of the protein (secondary structures, bond angles and, bond length) however minute (Phillips,

130

2013), and these changes on the protein structure-function modification are still at the meso- and macro- levels (Foegeding, 2015; Withana-Gamage & Wanasundara, 2012). Additionally, the various techniques (such as Spectroscopic: Fourier transformation infrared, Raman, Circular dichroism, Florescence, and Ultraviolet. Nuclear magnetic resonance, X-ray crystallography, Cryo-electron microscopy, and Small-angle, x-ray scattering) used to quantify these alterations in food protein structural properties are limited to evaluating the static conformation (Akharume, Xiong, & Adedeji, 2019; Wang, Sun, Pu, & Wei, 2017) of the protein structures before and after stress application and not dynamic conformation that reveals the real-time behavior of the protein under applied stress.

Thermal treatment beyond the protein denaturation temperature (varies with ambient conditions- 82.1±3.5°C for proso millet protein fractions (Akharume, Santra, et al., 2019) and 73.3 - 82.2°C for rice protein fractions (Ju, Hettiarachchy, & Rath, 2001)) leads to rupturing of the protein’s intra- and inter-molecular bonds, and loss protein of the protein secondary and tertiary structures (Sun-Waterhouse et al., 2014). Similarly, high pressure (usually beyond 200Mpa) leads to the loss of the protein’s secondary and tertiary structure but not able to rupture its covalent bonds (Yang & Powers, 2016).

MD simulations provide an opportunity to study and evaluate food protein dynamics and changes in conformation under real-time applied stress at the atomic and molecular levels (Singh et al., 2013; Withana-Gamage & Wanasundara, 2012). A few authors have studied the effects of temperature and pressure in MD simulation on the structural changes of selected food proteins. Vanga et., al (2019) revealed from their study on MD simulation of Gly m 4 soy allergen protein under temperature and pressure, that there were significant changes recorded in the structure of the protein molecules, particularly with residues D-27, and T-51. For instance, the root mean square deviation (RMSD) of the protein molecule increased from 0.254 ± 0.03 to 0.324 ± 0.039 nm as the temperature moved from 300 to 373K while the RMSD decreased from 0.324 ± 0.039 to 0.257 ± 0.037 nm at temperature 373K as the pressure changes from 1 bar to 6kbar (Vanga, Wang, Singh, Raghavan, & chemistry, 2019). In another study, simulated temperature and pressure of soybean trypsin inhibitor (STI) lead to changes in the radius of gyration, RMSD, and solvent accessibility surface area (SASA) of the protein molecule. For instance, both temperature (300-373K)

131

and pressure (1 bar - 6kbar) combinations were observed to reduce the radius of gyration of SPI from 1.581 ± 0.01 nm to 1.567 ± 0.006 nm at 300K, 1.591 ± 0.01 nm to 0.1557 ± 0.008 nm at 345K, and 1.590 ± 0.012 nm to 1.564 nm ± 0.008 nm at 373K, as the pressure increased from 1bar to 6kbar, leading to a more compact STI molecule (Vanga et al., 2018).

Glutelin type-B 5-like protein is a type of glutelin protein from proso millet. Proso glutelin proteins are not widely used as food protein ingredients owing to their poor physicochemical and functional properties (Akharume, Santra, et al., 2019), although they have promising nutritional and health benefits (Saleh et al., 2013) as the glutelin has been reported to reduce the concentration of low-density lipoprotein (LDL) cholesterol and increase the concentration of high-density lipoprotein (HDL) cholesterol in several mice studies (Nishizawa & Fudamoto, 1995; Park et al., 2008). With the understanding that the changes in protein structure conformation may confer improved functionality, we evaluated the effect of selected temperature (300, 350, and 400K) and pressure (1bar, 3kbar, and 6kbar) levels on the secondary structure, root mean square deviation (RMSD), root mean square fluctuations (RMSF) per residue, the radius of gyration (Rg), surface electrostatic potential, and solvent accessibility surface area (SASA) of the glutelin type-B 5-like protein.

7.1 Materials and Methods 7.1.1 Molecular dynamic simulations The MD Simulations of glutelin type-B 5-like in water was carried out by the Groningen machine for chemical structure (GROMACS) software package (Version 2018.1, Stockholm Center for Biomembrane Research, Stockholm, Sweden)(Van Der Spoel et al., 2005). The glutelin type-B 5-like protein used in this study was developed in a previous study and can be accessed in Appendix B of this dissertation or downloaded from the protein model database (PMDB) with the identity number of PM0083241 where we had deposited. The protein (Figure 7.1a) is a protomer of three non-covalently linked monomers with each monomer chain comprising 16.2% alpha-helix, 2.4% 3/10 helix, 33.5% beta-sheet, and 48% coils/turns/bends/bridges. The protein-containing 7713 atoms from 507 residues were kept in a cubic box of dimension 12.019 x 12.019 x 12.019 nm as

132

a periodic boundary condition and the protein was solvated with 163155 atoms of water molecules neutralized with 3 sodium ions (Figure 7.1b).

Figure 7.1. A snapshot of the three-dimensional structure of glutelin type-B 5-like protein molecule (A) in a vacuum (B) in neutralized water

The OPLS-AA/L all-atom force field (Robertson et al., 2015) and SPC/E water model (Kusalik & Svishchev, 1994) were selected to provide potential energy function and water parameters to the system. Following neutralization of the systems to mimic the physiological state of the protein, the systems were energy minimized to converge at maximum force value < 1000.0 kJ/mol/nm using stepwise descent minimization algorithm for 100,000 steps and the systems was further equilibrated to a constant number of particles, volume and temperature (NVT) and a constant number of particles, pressure, and temperature (NPT) for 200 ps at 300k and I bar.

The MD simulations were run for 1 ns using a leap-frog integrator algorithm during which the temperature of the systems was maintained using a modified Berendsen thermostat, and the pressure was maintained using the Parrinello-Rahman barostat (Berendsen et al., 1984; Parrinello & Rahman, 1980). A total of nine simulations was run for both static and oscillating electric levels coupled with the temperature levels (Table 7.1.). The results of simulations such as RMSD, the radius of gyration, and SASA were analyzed using GROMACS inbuilt tools. GROMACS inbuilt DSSP (Kabsch & Sander, 1983; Touw et al.,

133

2015) was used for the secondary structure analysis and virtual molecular dynamics (VMD)(Humphrey et al., 1996) was used to visualize protein conformational change.

Table 7.1. Summary of Simulation conditions used in the study.

Temperature Pressure Simulation length 1 bar 300 K 3 kbar 1 ns 6 kbar 1 bar 350 K 3 kbar 1 ns 6 kbar 1 bar 400 K 3 kbar 1 ns 6 kbar

7.2 Results and Discussion 7.2.1 Root Mean square Deviation (RMSD) When a protein molecule is subjected to external simulated processing stress such as temperature and pressure, there are changes in their original conformation as a result of the applied stresses. The RMSD provides information on the deviation of the protein during simulations from its original conformation at the start of the simulation (time zero). Mostly the deviation from the protein backbone or main chain’s atom for an atom is usually considered as in the case in our experiment and such deviation is calculated using equation 7.1.

1 푅푀푆퐷 = √ ∑푁 | 푟 (푖) − 푟 (푖) |2 (7.1) 푁 푖=1 푓푖푛푎푙 푖푛푖푡푖푎푙

Where 푟푓푖푛푎푙 is the final coordinates of atom i, and 푟푖푛푖푡푖푎푙 is the initial coordinate of the atom i, and N is the number of atoms. Table 7.2 summarises the average RMSD for glutelin type-B 5-like protein after 1ns simulation from different combinations of temperature and pressure. The average values of RMSD for the glutelin type-B 5-like protein decreased significantly (p < 0.0001) with increasing pressure. When the temperature was constant at 300K, the RMSD increased with increasing pressure (1 bar – 6 kbar) from 0.51 ± 0.16 to

134

0.37 ± 0.10 nm while at a temperature of 350K, the RMSD increased from 0.64 ± 0.21 to 0.47 ± 0.18 nm. Similarly, at a temperature of 400K, the RMSD increased from 0.82 ± 0.29 to 0.55 ± 0.12 nm. In addition, the average RMSD increased significantly (p < 0.0001) with increasing temperature from 0.51 ± 0.16 to 0.82 ± 0.29 nm, from 0.39 ± 0.14 to 0.66 ± 0.18 nm, and from 0.37 ± 0.10 to 0.55 ± 0.12 for pressure levels of 1 bar, 3kbar, and 6 kbar respectively. Vanga et al. (2018) reported a similar trend for soybean trypsin inhibitor protein where the RMSD values decreased from 0.269 ± 0.026 nm to 0.225 ± 0.019 nm when pressure changed from 1 bar to 6 kbar and similarly at other temperature levels of 345 and 373K. Figure 7.2 presents the behavior of the glutelin type-B 5-like protein with simulation time. It can be seen from the figure that the RMSD increased with simulation time and for constant pressure levels (Figure 7.2a-c), simulation at 400K gave the highest RMSD and 300K the lowest while at constant temperature (Fig 7.2d-f), the simulation at 1 bar gave the highest RMSD value and 6 kbar the lowest. This is very much expected, as the temperature increases the atoms of the protein molecules gain more energy for mobility, breaking of bonds, unfolding, and eventual denaturation of the protein molecule. However, the effect of pressure on protein is based on the Le Chatelier’s principle where a change in volume is accompanied by a change in pressure, that is as the pressure increases the volume of the protein decreases because the pressure pushes out the occluded cavities in the protein molecules (Galazka et al., 2000; Messens et al., 1997; Yang & Powers, 2016) as can be seen later in Figure 7.5. This might have been responsible for the decrease in the RMSD molecules.

135

Table 7. 2. Average root mean square deviation (RMSD), radius of gyration (Rg), solvent accessibility surface area (SASA), volume, and density of glutelin type-B 5-like protein after1 ns simulation under temperature and pressure conditions.

Treatments RMSD (nm) Rg (nm) SASA (nm2) Volume (nm3) Density (g/l) 300K, 1 bar 0.51 ± 0.16a1 3.26 ± 0.07a1 327.47 ± 2.66a1 108.35 ± 1.00a1 835.27 ± 7.69a1 300K, 3 kbar 0.39 ± 0.14b1 3.07 ± 0.03b1 317.06 ± 2.98b1 105.84 ± 1.08b1 855.09 ± 8.71b1 300K, 6 kbar 0.37 ± 0.10c1 3.06 ± 0.03c1 313.45 ± 3.76c1 104.95 ± 1.05c 862.41 ± 8.58c1 350K, 1 bar 0.64 ± 0.21a2 3.26 ± 0.05a2 331.65 ± 3.18a2 108.39 ± 1.09a2 835.03 ± 8.39a2 350K, 3 kbar 0.40 ± 0.10b1 3.12 ± 0.05b2 317.33 ± 3.47a2 106.22 ± 1.11a2 852.05 ± 8.85a2 350K, 6 kbar 0.47 ± 0.18c1 3.14 ± 0.05c2 317.98 ± 3.82a2 105.33 ± 1.00a2 859.26 ± 8.01a2 400K, 1 bar 0.82 ± 0.29a3 3.35 ± 0.08a3 330.84 ± 4.32a3 108.82 ± 1.19a3 831.71 ± 9.12a3

136 400K, 3 kbar 0.66 ± 0.18b1 3.09 ± 0.06b3 313.23 ± 10.68a3 106.05 ± 1.44a3 853.52 ± 11.58a3

400K, 6 kbar 0.55 ± 0.12c1 2.92 ± 0.06c3 311.06 ± 5.31a3 105.04 ± 1.16a3 861.65 ± 9.44a3

137

Figure 7.2. Root mean square deviation of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time.

7.2.2 Root Mean Square Fluctuation (RMSF) Figure 7.3 presents the root mean square fluctuations per residue for all the three chains of glutelin type-B 5-like protein under different simulated temperature and pressure combinations for 1 ns. The RMSF provides information on the flexible and the rigid regions of the protein molecules. We observed that intensity of the fluctuation increases with temperature and pressure. For instance, at the N-terminal of chain A (Q77), the fluctuation recorded for 300K to 400K at 1 bar to 6k bar were 0.09, 0.32, 0.53, 0.75, 1.00, 1.21, 1.56, 1.73, 2.00 nm respectively and on chain B were 0.11, 0.27, 0.54, 0.80, 0.97, 1.22, 1.50, 1.70, 1.96 nm respectively. Similarly, for one the alpha helices region on chain A (A233 - K240), the average fluctuations recorded for all treatment combinations were 0.08, 0.29, 0.50, 0.80, 0.99, 1.20, 1.50, 1.70, and 1.90 nm, respectively. Additionally, we observed that the N- and C- terminals on each of the protein chains as well as the alpha helix regions of the protein showed higher fluctuations compared to the beta sheets. For instance, on chain C we recorded average fluctuation values of 0.05, 0.28, 0.50, 0.77, 1.00, 1.25, 1.51, 1.76, and 1.91 nm for alpha helix region (V222 -A226) ; 0.06, 0.28, 0.51, 0.81, 0.97, 1.27, 1.51, 1.71, and 1.93 nm for alpha helix region (A233 – K240); 0.04, 0.29, 0.50, 0.74, 0.97, 1.18, 1.44, 1.67, and 1.90 nm for the β-sheet region (S122 – T127); 0.04, 0.28, 0.50, 0.76, 0.97, 1.20, 1.44, 1.66, and 1.89 nm for β-sheet region (S122 – T127); and we recorded 0.03, 0.27, 0.55, 0.84, 0.98, 1.27, 1.65, 1.67, and 2.04 for the C-terminal (G245) for all treatment combinations (300K 1 bar to 400K 6 kba respectively).

138

139

Figure 7.3. Root mean square fluctuations of glutelin type-B 5-like protein per residues under different simulation conditions (temperature and pressure) after 1 ns simulation time. The red rectangle shows the N-, C-terminals, the green rectangle shows the alpha helix region and the other obvious peaks are coils, turns, or bridges.

7.2.3 Secondary structure analysis Figure 7.4 presents the average number of amino acid residues that make the secondary structure elements of the glutelin type-B 5-like protein molecules after a 1 ns simulation under different simulation conditions. We expect that as the proteins unfold the number of residues for helices and beta sheets should decrease and adds to the number are coils, turns, bends, or bridges. We observed that the number of residues in coils, bends, and turns (6 kbar only) increases with temperature at constant pressure while in the alpha-helix and β- sheets the residue decreases with the temperature only at a constant pressure of 1 bar. However, as the pressure ramped up from 1 bar to 6 kbar the number of residues in the alpha-helix and β-sheets increased with temperature which may be a result of the refolding of the secondary structure elements in the protein or new aggregation which are a common characteristic of high-pressure treatment.

The snap shorts of the protein molecule after simulation (Figure 7.5) showed a compression in the cavities with the chains of the protein with pressure. However, no obvious disruption

140

of the secondary structure elements of helices and beta sheets were observed in all the treatment combinations, perhaps the short time simulation may not have been enough to cause an irreversible unfolding or permanent denaturation of the protein secondary structure. Additionally, high-pressure processing is only known to rupture the non-covalent interactions (intramolecular hydrophobic and electrostatic interactions), and forms new non-covalent/semi-covalent associations without breaking of hydrogen bonds (Galazka et al., 2000; Messens et al., 1997; Yang & Powers, 2016).

141

142

Figure 7.5. Snapshots of glutelin type-B 5-like protein molecule after1 ns simulation under different simulation conditions.

7.2.4 Radius of gyration (Rg) The degree of spread or compaction of the glutelin type-B 5-like protein after stimulation was measured by the radius of gyration using the equation (7.2). The Rg provides information on the deviation of the atoms of the protein molecule with respect to its center of mass.

1 푅푔 = √ ∑푁 | 푟 (푖) − 푟 |2 (7.2) 푁 푖=1 푐푒푛푡푟푒

Where 푟 (푖) is the coordinate of atom i, 푟푐푒푛푡푟푒 is the coordinate of the center of mass of the protein, and N is the number of atoms. We observed (Table 7.2) that at a constant temperature the Rg values of the protein decreased significantly (p < 0.0001) with an increase in pressure from 3.26 ± 0.07 to 3.06 ± 0.03 nm, 3.26 ± 0.05 to 3.14 ± 0.05 nm, and from 3.35 ± 0.08 to 2.92 ± 0.06 nm for 300, 350, and 400K temperature levels, respectively. At constant pressure, the Rg values increases with temperature from 3.26 ± 0.07 to 3.35 ± 0.08 at 1 bar and from 3.07 ± 0.03 to 3.09 ± 0.06 at 3 kbar but decreased from 3.06 ± 0.03 to 2.92 ± 0.06 at 6 kbar. In summary, at low pressure (1 bar) higher temperatures lead to unfolding and spreading of the protein (Figure 7.6a), however at a higher pressure of 3 and 6 kbar (Figure 7.6b&c), the protein begins to experience some compaction even with high-temperature levels.

143

144

Figure 7.6. Radius of gyration of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time.

7.2.5 Solvent Accessibility Surface Area (SASA) The summary of the solvent accessibility surface area of glutelin type-B 5-like protein under different simulation conditions are presented in Table 7.2. The SASA provides information on the hydrophobicity of the protein. Hydrophobicity of protein is inversely related to its SASA (Gromiha et al., 2019). The average SASA decreased significantly (p < 0.0001) with pressure at a constant temperature as presented in Table 7.2. For example, at 300K we recorded SAS values of 327.47 ± 2.66 nm2 for 1 bar, 317.06 ± 2.98 nm2 for 3 kbar, and 313.45 ± 3.76 for 6 kbar. Similarly, at a constant pressure of 1 bar, the SASA values increased with temperature showing that the pressure effects were not strong on SASA at this level. However, as the pressure was increased to 3 and 6 kbar the SAS values decreased, except at 350K, 6 bar which points to the fact that the pressure compaction effect was less pronounced at this level. The trends in SASA with simulation time is presented in Figure 7.7 and it can be observed that not much fluctuations were observed in the SASA value at low pressure (Fig. 7.7a) at higher pressure (Fig 7.7b&c), the SASA showed initial upward trend up to 700 ps and then begins to decline greatly which may suggest initial unfolding and revelations of the buried hydrophobic residue followed by new hydrophobic aggregation the hides some of the protein residues form solvent accessibility.

145

146

Figure 7.7. Solvent accessibility surface area of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time.

7.2.6 Volume and density The volume of protein under pressure stresses is expected to decrease and the density is expected to increase according to Le Chatelier’s principle. The average values of the volume and density of the glutelin type-B 5-like protein are presented in Table 2 and the trends with simulation time are presented in Figure 7.8. The trends in the volume of the protein for all treatment combinations were relatively stable over the simulation time. However, the average volume at constant temperature levels decreased significantly (p < 0.0001) with pressure while at constant pressure, the volume increased significantly ( P < 0.0050) with temperatures. Conversely, the average density at constant temperature levels increased significantly (p < 0.0001) with pressure while at constant pressure, the volume decreased with temperature.

147

148

Figure 7.8. Volume changes observed for glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1 ns simulation time.

7.3 Conclusions In summary, we evaluated the effect of temperature and pressure combinations (300K, 350K, and 400K for pressure levels 1 bar, 3 kbar, and 6 kbar) on the behavior of glutelin type-B 5-like protein in molecular dynamic simulation environment successfully. We observed that at a low pressure of 1 bar and high pressure of 3 kbar the temperature effects on the glutelin type-B 5-like protein is well pronounced and the stress at this levels led to increasing RMSD, RMSF, SASA, Rg and volume, but decreased density which reveals that the protein may not be experiencing much aggregation or compaction at the initial stage of simulation (600 ps) but as the simulation proceeds to 1000 ps, there may be compaction at 3 kbar levels (at all levels of temperature) for at the rest of the simulation time as a result of the pressure effects. However, at a higher pressure of 6 kbar, the temperature effects became less impactful on the values of Rg, SASA, and volume of the glutelin type-B 5-like protein so that the value of Rg decreased from 3.06 ± 0.03 to 2.92 ± 0.06 nm, and SASA values decreased from 313.45 ± 3.76 to 311.06 ± 5.31 nm2 and volume decreased 104.95 ± 1.05 to 105.04 ± 1.16 nm3 which shows that the protein treated at this level of pressure and temperature may aggregate and compact. Secondary structure analysis reveals loss in the numbers of residues of the beta-sheet and alpha-helix with a corresponding increase in the number of the residue of coils, turns and bends with increasing temperature while the alpha-helix and beta-sheet only reduced with the temperature at a constant pressure of 1 bar and increased at higher pressure levels confirming aggregation or refolding at high-pressure levels. We opined that increasing the MD simulation length could be necessary for the disruptions of the protein secondary structure which can expose whether there is a permanent denaturation or aggregation at higher pressure.

149

CHAPTER VIII: General Conclusions and Future Works 8.0 General Conclusions As the world continues to seek ways to feed its teeming population that is expected to be over 9 billion by 2050, research into ancient and underutilized crops to expand and sustain the agro-food system network will remain a critical factor. Proso millet is one of the emerging ancient crops with such a favorable agro-climatological potential and could provide a sustainable source of rich plant protein to consumers. However, to increase the applicability of proso millet in the food protein industry, a lot of research is required to provide the needed information on its physicochemical, functional, and structural properties, especially as it relates to different processing stresses and food formulation conditions. Thus, my dissertation project from objective one through objective five provided answers to some of the overarching questions being asked by plant protein processors and in food protein chemistry research field as it pertains to the utilization of proso millet proteins.

First, the question of physicochemical properties and functionality being asked by food processors - What is the solubility of proso millet proteins? Can proso millet proteins make good foam or emulsion? Are proso millet proteins highly digestible? Can these properties be improved? My research successfully furnishes this information that may be first of its kind. We showed that the different fractions of proso millet protein have a solubility of less than 40% at acidic pH and even much lower at pH 7 (under 20%, except for plateau glutelin and albumin) where it most desirable for processors. Emulsion capacity of the proteins average below 20 m2/g with glutelin and prolamin fraction even showing lower emulsion capacity (about 15 and 3 m2/g for glutelin and prolamin fractions) and foam capacity of the protein below 100% with the albumin, globulin, and prolamin fraction averaging about 40, 30 and 20% respectively while the glutelin average about 80%. Much of this information depicts a low functionality or physicochemical properties when comparing the proso millet protein to some gold standard ingredients like soybean and pea that have a solubility in the 80’s percent (Jung, Murphy, & Johnson, 2005; Lam, Warkentin, Tyler, & Nickerson, 2017). However, our findings showed that some of the functionality of proso millet protein can be improved by the application of ultrasound technology. We established that ultrasound power intensity of 45.84 Wcm-2 for 5 or 10 min can improve the solubility of

150

prolamin protein by up to three and five-folds, and the solubility of glutelin protein by four and eight folds respectively. Additionally, the pepsin digestibility of Dawn prolamin increased by about 30% for ultrasound treatment times of 5 or 10 min and for the plateau glutelin the pepsin digestibility increased by 48.5%

The second question by protein research chemists usually relates to the structure-function properties of protein- what is the three-dimensional structure of the proso millet protein? What is known about the changes in the conformation of its three-dimensional structure and how does it impart its functionality such as solubility foaming and emulsion? My dissertation provides useful information in this area. We successfully determine the three- dimensional structure of glutelin type-B 5-like protein, a type of glutelin protein from proso millet using homology modeling and submitted that it has conserved regions (one jelly-like β-barrel and two extended helix domains) with the globulin proteins (11s or pro-11s globulin proteins) of pea, soybean, pumpkin, amaranth, and rapeseed and between 35 - 45% structural similarity with the globulins of these proteins. Additionally, we found that the predicted instability index value of the glutelin type-B 5-like protein was high (62.95%) which may have been responsible for it forming aggregation during the preliminary attempt to purify the protein and to determine its three-dimensional structure using X-ray crystallography. Finally, my work on molecular dynamic modeling of the glutelin type-B 5-like protein evaluating the combined effects of temperature (300, 350, and 400K) and static electric field (0.1, 1, and 3 v/nm) as well as the combined effect of temperature (300, 350, and 400K) and pressure (1 bar, 3, and 6 kbar) provides some information on the conformational behavior of the glutelin type-B 5-like protein that was not known before. We showed that the root mean square deviations (RMSD) increases as the intensity of the stresses increases, except for increasing pressure where the RMSD actually decreased when the temperature was held constant. Furthermore, we showed that the amino acid at the terminals of the protein fluctuated more with stresses and alpha-helix fluctuated more than the beta-sheet. While some losses of the amino residue that make up the secondary structure of the glutelin type-B 5-like protein were observed, obvious disruption to this secondary structure was not noticed which may suggest that higher processing stress intensity and/or higher simulation time may be needed to cause a major and irreversible disruption of the protein secondary structure.

151

8.1 Future Work While my dissertation addressed its set objectives and provided novel information that supports both food protein processors and researchers, there is still an avalanche of research opportunities and gray areas that can be explored. Here I highlight some of the potential research areas. First, the application of proso millet protein in product development. Since it is possible to improve some of the millet’s functionality through ultrasound, one question that begs to be answered is how the millet behaves in food formulation such as cakes, bread, and other pastries. Also, it is important to know how other modification techniques such as high-pressure processing and fermentation improves the functionality of proso millet protein.

Second, in the area of molecular modeling to understand structural conformation, the area that might be practicably applied might be to establish a correlation between MD simulation results (molecular scale) and the results obtained from using other investigative techniques such as circular dichroism, Fourier-transform infrared spectroscopy (FTIR), spectrofluorometer, and Raman spectroscopy all of which are used to investigate structure- function behavior of proteins at the macroscopic scale.

152

General References

21CFR172.892. (2019). Substances generally recognized as safe. US: United States Food and Drug Administration Abbey, B. W., & Ibeh, G. O. (1988). Functional properties of raw and heat processed cowpea (vigna unguiculata, walp) flour. Journal of Food Science, 53(6), 1775- 1777. doi:10.1111/j.1365-2621.1988.tb07840.x Adachi, M., Takenaka, Y., Gidamis, A. B., Mikami, B., & Utsumi, S. (2001). Crystal structure of soybean proglycinin a1ab1b homotrimer. Journal of molecular biology, 305(2), 291-305. doi:https://doi.org/10.1006/jmbi.2000.4310 Adebiyi, A. P., & Aluko, R. E. (2011). Functional properties of protein fractions obtained from commercial yellow field pea (pisum sativum l.) seed protein isolate. Food Chemistry, 128(4), 902-908. Adler-Nissen, J., Eriksen, S., & Olsen, H. S. (1983). Improvement of the functionality of vegetable proteins by controlled enzymatic hydrolysis. Plant Foods for Human Nutrition, 32(3-4), 411-423. AgMRC. (2018). Proso millet. Retrieved from https://www.agmrc.org/commodities- products/grains-oilseeds/proso-millet Ahmadifard, N., Murueta, J. H. C., Abedian-Kenari, A., Motamedzadegan, A., & Jamali, H. (2016). Comparison the effect of three commercial enzymes for enzymatic hydrolysis of two substrates (rice bran protein concentrate and soy-been protein) with sds-page. Journal of Food Science and Technology, 53(2), 1279-1284. doi:10.1007/s13197-015-2087-6 Ahmed, J., Al-Ruwaih, N., Mulla, M., & Rahman, M. H. (2018). Effect of high pressure treatment on functional, rheological and structural properties of kidney bean protein isolate. LWT, 91, 191-197. Akharume, F., Santra, D., & Adedeji, A. (2019). Physicochemical and functional properties of proso millet storage protein fractions. Food Hydrocolloids, 105497. Akharume, F., Xiong, Y. L., & Adedeji, A. (2019). Effects of transglutaminase in product formulation on physico-textural properties of protein rich extruded snacks. Paper presented at the 2019 ASABE Annual International Meeting, St. Joseph, MI. http://elibrary.asabe.org/abstract.asp?aid=50277&t=5

153

Akıllıoğlu, H. G., & Gökmen, V. (2016). Kinetic evaluation of the inhibition of protein glycation during heating. Food Chemistry, 196, 1117-1124. Aluko, R., & Monu, E. (2003). Functional and bioactive properties of quinoa seed protein hydrolysates. Journal of Food Science, 68(4), 1254-1258. Aluko, R. E. (2018). 15 - food protein-derived peptides: Production, isolation, and purification. In R. Y. Yada (Ed.), Proteins in food processing (second edition) (pp. 389-412): Woodhead Publishing. Annor, G. A., Tyl, C., Marcone, M., Ragaee, S., & Marti, A. (2017). Why do millets have slower starch and protein digestibility than other cereals? Trends in Food Science & Technology, 66, 73-83. Arntfield, S. D., Murphy, L. A., & Ross, W. F. (2018). Canola based tofu product and method: Google Patents. Aryee, A. N. A., Agyei, D., & Udenigwe, C. C. (2018). 2 - impact of processing on the chemistry and functionality of food proteins. In R. Y. Yada (Ed.), Proteins in food processing (second edition) (pp. 27-45): Woodhead Publishing. Aschemann-Witzel, J., & Peschel, A. O. (2019). Consumer perception of plant-based proteins: The value of source transparency for alternative protein ingredients. Food Hydrocolloids, 96, 20-28. doi:https://doi.org/10.1016/j.foodhyd.2019.05.006 Attri, P., Kumar, N., Park, J. H., Yadav, D. K., Choi, S., Uhm, H. S., . . . Lee, W. (2015). Influence of reactive species on the modification of biomolecules generated from the soft plasma. Scientific reports, 5(1), 8221. doi:10.1038/srep08221 Baltensperger, D. D. (2002). Progress with proso, pearl and other millets. Trends in new crops and new uses, 100-103. Baneyx, F., & Mujacic, M. (2004). Recombinant protein folding and misfolding in escherichia coli. Nature Biotechnology, 22(11), 1399-1408. doi:10.1038/nbt1029 Beck, S. M., Knoerzer, K., & Arcot, J. (2017). Effect of low moisture extrusion on a pea protein isolate’s expansion, solubility, molecular weight distribution and secondary structure as determined by fourier transform infrared spectroscopy (ftir). Journal of Food Engineering, 214, 166-174. Benkert, P., Biasini, M., & Schwede, T. J. B. (2011). Toward the estimation of the absolute quality of individual protein structure models. 27(3), 343-350.

154

Bera, M., & Mukherjee, R. (1989). Solubility, emulsifying, and foaming properties of rice bran protein concentrates. Journal of Food Science, 54(1), 142-145. Berendsen, H. J., Postma, J. v., van Gunsteren, W. F., DiNola, A., & Haak, J. (1984). Molecular dynamics with coupling to an external bath. The Journal of chemical physics, 81(8), 3684-3690. Bergsma, J. (2019). Vegan cheese analogue: Google Patents. Berk, Z. (2013). Chapter 15 - extrusion. In Z. Berk (Ed.), Food process engineering and technology (second edition) (pp. 373-393). San Diego: Academic Press. Bielikowicz, K., Kostyra, H., Kostyra, E., Teodorowicz, M., Rigby, N., & Wojtacha, P. (2012). The influence of non-enzymatic glycosylation on physicochemical and biological properties of pea globulin 7s. Food Research International, 48(2), 831- 838. doi:https://doi.org/10.1016/j.foodres.2012.06.028 Boostani, S., Aminlari, M., Moosavi-Nasab, M., Niakosari, M., & Mesbahi, G. (2017). Fabrication and characterisation of soy protein isolate-grafted dextran biopolymer: A novel ingredient in spray-dried soy beverage formulation. International journal of biological macromolecules, 102, 297-307. Borgo, B., & Havranek, J. J. (2012). Automated selection of stabilizing mutations in designed and natural proteins. Proceedings of the National Academy of Sciences, 109(5), 1494-1499. Boye, J., Aksay, S., Roufik, S., Ribéreau, S., Mondor, M., Farnworth, E., & Rajamohamed, S. (2010). Comparison of the functional properties of pea, chickpea and lentil protein concentrates processed using ultrafiltration and isoelectric precipitation techniques. Food Research International, 43(2), 537-546. Boye, J., Zare, F., & Pletch, A. (2010). Pulse proteins: Processing, characterization, functional properties and applications in food and feed. Food Research International, 43(2), 414-431. Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical biochemistry, 72(1-2), 248-254.

155

Buchert, J., Ercili Cura, D., Ma, H., Gasparetti, C., Monogioudi, E., Faccio, G., . . . Selinheimo, E. (2010). Crosslinking food proteins for improved functionality. Annual review of food science and technology, 1, 113-138. Bučko, S., Katona, J., Popović, L., Petrović, L., & Milinković, J. (2016). Influence of enzymatic hydrolysis on solubility, interfacial and emulsifying properties of pumpkin (cucurbita pepo) seed protein isolate. Food Hydrocolloids, 60, 271-278. doi:https://doi.org/10.1016/j.foodhyd.2016.04.005 Bußler, S., Steins, V., Ehlbeck, J., & Schlüter, O. (2015). Impact of thermal treatment versus cold atmospheric plasma processing on the techno-functional protein properties from pisum sativum ‘salamanca’. Journal of Food Engineering, 167, 166-174. doi:https://doi.org/10.1016/j.jfoodeng.2015.05.036 Cabra, V., Arreguin, R., Vazquez-Duhalt, R., & Farres, A. (2007). Effect of alkaline deamidation on the structure, surface hydrophobicity, and emulsifying properties of the z19 α-zein. Journal of Agricultural and Food Chemistry, 55(2), 439-445. Cassity, J. (2019). The state of the global plant-based protein market. Retrieved from https://www.kerry.com/global-en/insights/kerrydigest/2019/the-state-of-the- global-plant-based-protein-market Chan, W. M., & Ma, C. Y. (1999). Acid modification of proteins from soymilk residue (okara). Food Research International, 32(2), 119-127. doi:https://doi.org/10.1016/S0963-9969(99)00064-2 Chao, D., & Aluko, R. E. J. C.-J. o. F. (2018). Modification of the structural, emulsifying, and foaming properties of an isolated pea protein by thermal pretreatment. 16(1), 357-366. Chao, D., Jung, S., & Aluko, R. E. (2018). Physicochemical and functional properties of high pressure-treated isolated pea protein. Innovative Food Science & Emerging Technologies, 45, 179-185. Charoensuk, D., Brannan, R. G., Chanasattru, W., & Chaiyasit, W. (2018). Physicochemical and emulsifying properties of mung bean protein isolate as influenced by succinylation. International Journal of Food Properties, 21(1), 1633- 1645. doi:10.1080/10942912.2018.1502200

156

Cheatham, R. (2013). Protein: A plant-based look at this power macronutrient. Retrieved from Phoenix AZ, USA: Chen, F. L., Wei, Y. M., & Zhang, B. (2011). Chemical cross-linking and molecular aggregation of soybean protein during extrusion cooking at low and high moisture content. LWT-Food Science and Technology, 44(4), 957-962. Chen, J., Mu, T., Zhang, M., & Goffin, D. (2019). Effect of high hydrostatic pressure on the structure, physicochemical and functional properties of protein isolates from cumin (cuminum cyminum) seeds. International Journal of Food Science & Technology, 54(3), 752-761. Chen, L., Chen, J., Yu, L., Wu, K., & Zhao, M. (2018). Emulsification performance and interfacial properties of enzymically hydrolyzed peanut protein isolate pretreated by extrusion cooking. Food Hydrocolloids, 77, 607-616. Chiang, J. H., Loveday, S. M., Hardacre, A. K., & Parker, M. E. (2019). Effects of soy protein to wheat gluten ratio on the physicochemical properties of extruded meat analogues. Food Structure, 19, 100102. doi:https://doi.org/10.1016/j.foostr.2018.11.002 Cho, H.-W., Jung, Y.-M., Auh, J.-H., & Lee, D.-U. J. K. j. f. f. s. o. a. r. (2017). Effects of jet-milled defatted soy flour on the physicochemical and sensorial properties of hamburger patties. 37(6), 840. Conway, P., Tyka, M. D., DiMaio, F., Konerding, D. E., & Baker, D. J. P. S. (2014). Relaxation of backbone bond geometry improves protein energy landscape modeling. 23(1), 47-55. Creighton, T. E. (1993). Proteins: Structures and molecular properties: Macmillan. Dacey, M., & O'connor, C. (2016). Ready-to-bake gluten-free pizza dough formulations: Google Patents. Das Purkayastha, M., Borah, A. K., Saha, S., Manhar, A. K., Mandal, M., & Mahanta, C. L. (2016). Effect of maleylation on physicochemical and functional properties of rapeseed protein isolate. Journal of Food Science and Technology, 53(4), 1784- 1797. doi:10.1007/s13197-016-2197-9 Davis, P., & Williams, S. (1998). Protein modification by thermal processing. Allergy, 53, 102-105.

157

Davis, P. J., Smales, C. M., & James, D. C. (2001). How can thermal processing modify the antigenicity of proteins? Allergy, 56(s67), 56-60. doi:10.1034/j.1398- 9995.2001.00918.x De Eslie, W. D., & Cheryan, M. (1981). Continuous enzymatic modification of proteins in an ultrafiltration reactor. Journal of Food Science, 46(4), 1035-1042. Devisetti, R., Yadahally, S. N., & Bhattacharya, S. (2014). Nutrients and antinutrients in foxtail and proso millet milled fractions: Evaluation of their flour functionality. LWT - Food Science and Technology, 59(2, Part 1), 889-895. doi:https://doi.org/10.1016/j.lwt.2014.07.003 Diftis, N., & Kiosseoglou, V. (2003). Improvement of emulsifying properties of soybean protein isolate by conjugation with carboxymethyl cellulose. Food Chemistry, 81(1), 1-6. doi:https://doi.org/10.1016/S0308-8146(02)00236-4 Djoullah, A., Husson, F., & Saurel, R. (2018). Gelation behaviors of denaturated pea albumin and globulin fractions during transglutaminase treatment. Food Hydrocolloids, 77, 636-645. doi:https://doi.org/10.1016/j.foodhyd.2017.11.005 Dong, S., Gao, A., Zhao, Y., Li, Y.-t., & Chen, Y. (2017). Characterization of physicochemical and structural properties of atmospheric cold plasma (acp) modified zein. Food and bioproducts processing, 106, 65-74. Du, Y., Jiang, Y., Zhu, X., Xiong, H., Shi, S., Hu, J., . . . Sun, W. (2012). Physicochemical and functional properties of the protein isolate and major fractions prepared from akebia trifoliata var. Australis seed. Food Chemistry, 133(3), 923-929. Eckert, E., Han, J., Swallow, K., Tian, Z., Jarpa-Parra, M., & Chen, L. (2019). Effects of enzymatic hydrolysis and ultrafiltration on physicochemical and functional properties of faba bean protein. 96(4), 725-741. doi:10.1002/cche.10169 Elzoghby, A. O., Samy, W. M., & Elgindy, N. A. (2012). Protein-based nanocarriers as promising drug and gene delivery systems. Journal of controlled release, 161(1), 38-49. Farkas, D. F. (2016). A short history of research and development efforts leading to the commercialization of high-pressure processing of food High pressure processing of food (pp. 19-36): Springer.

158

Farzana, T., Mohajan, S., Saha, T., Hossain, M. N., & Haque, M. Z. (2017). Formulation and nutritional evaluation of a healthy vegetable soup powder supplemented with soy flour, mushroom, and moringa leaf. 5(4), 911-920. doi:10.1002/fsn3.476 Fenwick, S., Vanga, S. K., DiNardo, A., Wang, J., Raghavan, V., & Singh, A. J. E. R. (2019). Computational evaluation of the effect of processing on the trypsin and alpha‐amylase inhibitor from ragi (eleusine coracana) seed. 1(4), e12064. Ferjancic-Biagini, A., Giardina, T., & Puigserver, A. (1998). Acylation of food proteins and hydrolysis by digestive enzymes: A review. Journal of Food Biochemistry, 22(4), 331-345. doi:10.1111/j.1745-4514.1998.tb00248.x Flores, I., Cabra, V., Quirasco, M. C., Farres, A., & Galvez, A. (2010). Emulsifying properties of chemically deamidated corn (zea mays) gluten meal. Food Science and Technology International, 16(3), 241-250. doi:10.1177/1082013210366750 Floury, J., Desrumaux, A., & Legrand, J. (2002). Effect of ultra‐high‐pressure homogenization on structure and on rheological properties of soy protein‐stabilized emulsions. Journal of Food Science, 67(9), 3388-3395. Foegeding, E. A. (2015). Food protein functionality—a new model. Journal of Food Science, 80(12), C2670-C2677. Foegeding, E. A., & Davis, J. P. (2011). Food protein functionality: A comprehensive approach. Food Hydrocolloids, 25(8), 1853-1864. Frank, A. W. (1987). Non-enzymic methods for the phosphorylation of proteins. Phosphorus and sulfur and the related elements, 29(2-4), 297-315. Fukuda, T., Maruyama, N., Salleh, M. R. M., Mikami, B., & Utsumi, S. (2008). Characterization and crystallography of recombinant 7s globulins of adzuki bean and structure− function relationships with 7s globulins of various crops. Journal of Agricultural and Food Chemistry, 56(11), 4145-4153. Galante, M., De Flaviis, R., Boeris, V., & Spelzini, D. J. L. (2020). Effects of the enzymatic hydrolysis treatment on functional and antioxidant properties of quinoa protein acid-induced gels. 118, 108845. Galazka, V. B., Dickinson, E., & Ledward, D. A. (2000). Influence of high pressure processing on protein solutions and emulsions. Current opinion in colloid & interface science, 5(3-4), 182-187.

159

Gaspar, A. L. C., & de Góes-Favoni, S. P. (2015). Action of microbial transglutaminase (mtgase) in the modification of food proteins: A review. Food Chemistry, 171, 315- 322. Gasteiger, E., Hoogland, C., Gattiker, A., Wilkins, M. R., Appel, R. D., & Bairoch, A. (2005). Protein identification and analysis tools on the expasy server The proteomics protocols handbook (pp. 571-607): Springer. Good Food Institute. (2020). Plant-based market Retrieved from https://www.gfi.org/marketresearch Ghribi, A. M., Gafsi, I. M., Blecker, C., Danthine, S., Attia, H., & Besbes, S. (2015). Effect of drying methods on physico-chemical and functional properties of chickpea protein concentrates. Journal of Food Engineering, 165, 179-188. doi:https://doi.org/10.1016/j.jfoodeng.2015.06.021 Gillespie, S., & van den Bold, M. (2017). Agriculture, food systems, and nutrition: Meeting the challenge. Global Challenges. Glusac, J., Davidesko-Vardi, I., Isaschar-Ovdat, S., Kukavica, B., & Fishman, A. (2018). Gel-like emulsions stabilized by tyrosinase-crosslinked potato and zein proteins. Food Hydrocolloids, 82, 53-63. doi:https://doi.org/10.1016/j.foodhyd.2018.03.046 Glusac, J., Isaschar-Ovdat, S., & Fishman, A. (2020). Transglutaminase modifies the physical stability and digestibility of chickpea protein-stabilized oil-in-water emulsions. Food Chemistry, 315, 126301. doi:https://doi.org/10.1016/j.foodchem.2020.126301 Grasso, S., Pintado, T., Pérez-Jiménez, J., Ruiz-Capillas, C., & Herrero, A. M. J. F. (2020). Potential of a sunflower seed by-product as animal fat replacer in healthier frankfurters. 9(4), 445. Greenfield, J., Ross-Murphy, S., Tamás, L., Békés, F., Halford, N., Tatham, A., & Shewry, P. (1998). Rheological properties of monomeric and polymeric forms of c hordein, a sulphur-poor prolamin of barley. Journal of Cereal Science, 27(3), 233-236. Gromiha, M. M., Nagarajan, R., & Selvaraj, S. (2019). Protein structural bioinformatics: An overview. In S. Ranganathan, M. Gribskov, K. Nakai, & C. Schönbach (Eds.), Encyclopedia of bioinformatics and computational biology (pp. 445-459). Oxford: Academic Press.

160

Gruener, L., & Ismond, M. A. H. (1997). Effects of acetylation and succinylation on the physicochemical properties of the canola 12s globulin. Part i. Food Chemistry, 60(3), 357-363. doi:https://doi.org/10.1016/S0308-8146(96)00348-2 Guex, N., Peitsch, M. C., & Schwede, T. (2009). Automated comparative protein structure modeling with swiss-model and swiss-pdbviewer: A historical perspective. 30(S1), S162-S173. doi:10.1002/elps.200900140 Gulati, P., Holding, D., Li, A., Rose Devin, J., Santra, D., & Zhang, Y. (2017). Heating reduces proso millet protein digestibility via formation of hydrophobic aggregates. Journal of Agricultural and Food Chemistry, 65(9), 1952-1959. doi:http://dx.doi.org/10.1021%2Facs.jafc.6b05574 Gumus, C. E., Decker, E. A., & McClements, D. J. (2017). Impact of legume protein type and location on lipid oxidation in fish oil-in-water emulsions: Lentil, pea, and faba bean proteins. Food Research International, 100, 175-185. Habiyaremye, C., Matanguihan, J. B., D’Alpoim Guedes, J., Ganjyal, G. M., Whiteman, M. R., Kidwell, K. K., & Murphy, K. M. (2017). Proso millet (panicum miliaceum l.) and its potential for cultivation in the pacific northwest, u.S.: A review. Frontiers in Plant Science, 7(1961). doi:10.3389/fpls.2016.01961 Hamada, J. S., & Swanson, B. (1994). Deamidation of food proteins to improve functionality. Critical Reviews in Food Science and Nutrition, 34(3), 283-292. doi:10.1080/10408399409527664 Hassan, A. B., Pawelzik, E., & von Hoersten, D. (2016). Effect of radio frequency heating on nutritional quality and protein solubility of corn. Food science & nutrition, 4(5), 686-689. He, R., He, H.-Y., Chao, D., Ju, X., & Aluko, R. (2014). Effects of high pressure and heat treatments on physicochemical and gelation properties of rapeseed protein isolate. Food and Bioprocess Technology, 7(5), 1344-1353. Heinig, M., & Frishman, D. J. N. a. r. (2004). Stride: A web server for secondary structure assignment from known atomic coordinates of proteins. 32(suppl_2), W500-W502. Henchion, M., Hayes, M., Mullen, A. M., Fenelon, M., & Tiwari, B. (2017). Future protein supply and demand: Strategies and factors influencing a sustainable equilibrium. Foods, 6(7), 53.

161

Higuera-Barraza, O., Del Toro-Sanchez, C., Ruiz-Cruz, S., & Márquez-Ríos, E. (2016). Effects of high-energy ultrasound on the functional properties of proteins. Ultrasonics Sonochemistry, 31, 558-562. Hrckova, M., Rusnakova, M., & Zemanovic, J. (2002). Enzymatic hydrolysis of defatted soy flour by three different proteases and their effect on the functional properties of resulting protein hydrolysates. Czech journal of food sciences, 20(1), 7-14. Hu, H., Cheung, I. W., Pan, S., & Li-Chan, E. C. (2015). Effect of high intensity ultrasound on physicochemical and functional properties of aggregated soybean β-conglycinin and glycinin. Food Hydrocolloids, 45, 102-110. Hu, Z., Qiu, L., Sun, Y., Xiong, H., & Ogra, Y. (2019). Improvement of the solubility and emulsifying properties of rice bran protein by phosphorylation with sodium trimetaphosphate. Food Hydrocolloids, 96, 288-299. doi:https://doi.org/10.1016/j.foodhyd.2019.05.037 Humphrey, W., Dalke, A., & Schulten, K. J. J. o. m. g. (1996). Vmd: Visual molecular dynamics. 14(1), 33-38. Isaschar-Ovdat, S., Rosenberg, M., Lesmes, U., & Fishman, A. (2015). Characterization of oil-in-water emulsions stabilized by tyrosinase-crosslinked soy glycinin. Food Hydrocolloids, 43, 493-500. doi:https://doi.org/10.1016/j.foodhyd.2014.07.007 Jambrak, A. R., Mason, T. J., Lelas, V., Herceg, Z., & Herceg, I. L. (2008). Effect of ultrasound treatment on solubility and foaming properties of whey protein suspensions. Journal of Food Engineering, 86(2), 281-287. Ji, H., Dong, S., Han, F., Li, Y., Chen, G., Li, L., & Chen, Y. (2018). Effects of dielectric barrier discharge (dbd) cold plasma treatment on physicochemical and functional properties of peanut protein. Food and Bioprocess Technology, 11(2), 344-354. doi:10.1007/s11947-017-2015-z Jiang, L., Wang, J., Li, Y., Wang, Z., Liang, J., Wang, R., . . . Zhang, M. (2014). Effects of ultrasound on the structure and physical properties of black bean protein isolates. Food Research International, 62, 595-601. Joshi, S., Shah, S., & Kalantar-Zadeh, K. (2018). Adequacy of plant-based proteins in chronic kidney disease. Journal of Renal Nutrition.

162

Ju, Z., Hettiarachchy, N., & Rath, N. (2001). Extraction, denaturation and hydrophobic properties of rice flour proteins. Journal of Food Science, 66(2), 229-232. Jung, S., Murphy, P. A., & Johnson, L. A. J. J. o. F. S. (2005). Physicochemical and functional properties of soy protein substrates modified by low levels of protease hydrolysis. 70(2), C180-C187. Kabsch, W., & Sander, C. J. B. O. R. o. B. (1983). Dictionary of protein secondary structure: Pattern recognition of hydrogen‐bonded and geometrical features. 22(12), 2577-2637. Kalinova, J., & Moudry, J. (2006). Content and quality of protein in proso millet (panicum miliaceum l.) varieties. Plant Foods for Human Nutrition, 61(1), 43. doi:10.1007/s11130-006-0013-9 Kamara, M. T., Amadou, I., Tarawalie, F., & Huiming, Z. (2010). Effect of enzymatic hydrolysis on the functional properties of foxtail millet (setaria italica l.) proteins. International Journal of Food Science & Technology, 45(6), 1175-1183. Khan, N. M., Mu, T.-H., Sun, H.-N., Zhang, M., & Chen, J.-W. (2015). Effects of high hydrostatic pressure on secondary structure and emulsifying behavior of sweet potato protein. High Pressure Research, 35(2), 189-202. Kim, D. E., Chivian, D., & Baker, D. (2004). Protein structure prediction and analysis using the robetta server. Nucleic acids research, 32(suppl_2), W526-W531. doi:10.1093/nar/gkh468 %J Nucleic Acids Research Kinsella, J. E., & Melachouris, N. (1976). Functional properties of proteins in foods: A survey. Critical Reviews in Food Science & Nutrition, 7(3), 219-280. Ko, J., Park, H., Heo, L., & Seok, C. J. N. a. r. (2012). Galaxyweb server for protein structure prediction and refinement. 40(W1), W294-W297. Kohama, K., Nagasawa, T., & Nishizawa, N. (1999). Polypeptide compositions and nh2- terminal amino acid sequences of proteins in foxtail and proso millets. Bioscience, biotechnology, and biochemistry, 63(11), 1921-1926. Kong, X., Zhou, H., & Qian, H. (2007). Enzymatic hydrolysis of wheat gluten by proteases and properties of the resulting hydrolysates. Food Chemistry, 102(3), 759-763. Korotkov, K. V., Delarosa, J. R., & Hol, W. G. J. (2013). A dodecameric ring-like structure of the n0 domain of the type ii secretin from enterotoxigenic escherichia coli.

163

Journal of Structural Biology, 183(3), 354-362. doi:https://doi.org/10.1016/j.jsb.2013.06.013 Kortemme, T., Joachimiak, L. A., Bullock, A. N., Schuler, A. D., Stoddard, B. L., & Baker, D. (2004). Computational redesign of protein-protein interaction specificity. Nature Structural and Molecular Biology, 11(4), 371. Krintiras, G. A., Diaz, J. G., Van Der Goot, A. J., Stankiewicz, A. I., & Stefanidis, G. D. J. J. o. F. E. (2016). On the use of the couette cell technology for large scale production of textured soy-based meat replacers. 169, 205-213. Krogh, A., Larsson, B., von Heijne, G., & Sonnhammer, E. L. (2001). Predicting transmembrane protein topology with a hidden markov model: Application to complete genomes. J Mol Biol, 305(3), 567-580. doi:10.1006/jmbi.2000.4315 Kuipers, B. J., van Koningsveld, G. A., Alting, A. C., Driehuis, F., Gruppen, H., & Voragen, A. G. (2005). Enzymatic hydrolysis as a means of expanding the cold gelation conditions of soy proteins. Journal of Agricultural and Food Chemistry, 53(4), 1031-1038. Kusalik, P. G., & Svishchev, I. M. J. S. (1994). The spatial structure in liquid water. 265(5176), 1219-1221. Kyriakopoulou, K., Dekkers, B., & van der Goot, A. J. (2019). Chapter 6 - plant-based meat analogues. In C. M. Galanakis (Ed.), Sustainable meat production and processing (pp. 103-126): Academic Press. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage t4. nature, 227(5259), 680. Lam, A. C. Y., Warkentin, T. D., Tyler, R. T., & Nickerson, M. T. (2017). Physicochemical and functional properties of protein isolates obtained from several pea cultivars. 94(1), 89-97. doi:10.1094/cchem-04-16-0097-fi Lam, R. S., & Nickerson, M. T. (2013). Food proteins: A review on their emulsifying properties using a structure–function approach. Food Chemistry, 141(2), 975-984. Laskowski, R. A., MacArthur, M. W., Moss, D. S., & Thornton, J. M. J. J. o. a. c. (1993). Procheck: A program to check the stereochemical quality of protein structures. 26(2), 283-291.

164

Lawton, J. W. (1992). Viscoelasticity of zein-starch doughs. Cereal Chem., 69(4), 351- 355. Li, C.-P., Enomoto, H., Hayashi, Y., Zhao, H., Aoki, T. J. L.-F. S., & Technology. (2010). Recent advances in phosphorylation of food proteins: A review. 43(9), 1295-1300. Li, X., Lin, C., & O'Connor, P. B. (2010). Glutamine deamidation: Differentiation of glutamic acid and gamma-glutamic acid in peptides by electron capture dissociation. Analytical Chemistry, 82(9), 3606-3615. doi:10.1021/ac9028467 Li, Y., Zhong, F., Ji, W., Yokoyama, W., Shoemaker, C. F., Zhu, S., & Xia, W. (2013). Functional properties of maillard reaction products of rice protein hydrolysates with mono-, oligo-and polysaccharides. Food Hydrocolloids, 30(1), 53-60. Li, Z., Natarajan, P., Ye, Y., Hrabe, T., & Godzik, A. (2014). Posa: A user-driven, interactive multiple protein structure alignment server. Nucleic acids research, 42(W1), W240-W245. doi:10.1093/nar/gku394 %J Nucleic Acids Research Liao, L., Zhao, M., Ren, J., Zhao, H., Cui, C., & Hu, X. (2010). Effect of acetic acid deamidation‐induced modification on functional and nutritional properties and conformation of wheat gluten. Journal of the Science of Food and Agriculture, 90(3), 409-417. Liebschner, D., Afonine, P. V., Baker, M. L., Bunkóczi, G., Chen, V. B., Croll, T. I., . . . McCoy, A. J. J. A. C. S. D. S. B. (2019). Macromolecular structure determination using x-rays, neutrons and electrons: Recent developments in phenix. 75(10), 861- 877. Lin, S., Huff, H., & Hsieh, F. (2000). Texture and chemical characteristics of soy protein meat analog extruded at high moisture. Journal of Food Science, 65(2), 264-269. Lipan, L., Rusu, B., Sendra, E., Hernández, F., Vázquez‐Araújo, L., Vodnar, D. C., . . . Agriculture. (2020). Spray drying and storage of probiotic enriched almond milk: Probiotic survival and physicochemical properties. Lippow, S. M., & Tidor, B. (2007). Progress in computational protein design. Current opinion in biotechnology, 18(4), 305-311. Liu, Y., Wang, D., Wang, J., Yang, Y., Zhang, L., Li, J., . . . Technology. (2019). Functional properties and structural characteristics of phosphorylated pea protein isolate.

165

Livesey, G., Taylor, R., Hulshof, T., & Howlett, J. (2008). Glycemic response and health— a systematic review and meta-analysis: Relations between dietary glycemic properties and health outcomes–. The American journal of clinical nutrition, 87(1), 258S-268S. Lorenz, K., & Dilsaver, W. (1980). Rheological properties and food applications of proso millet flours. Cereal chem, 57(1), 21-24. Lorenz, K., Dilsaver, W., & Bates, L. (1980). Proso millets. Milling characteristics, proximate compositions, nutritive value of flours. Cereal chem, 57(1), 16-20. Lu, Z., He, J., Zhang, Y., Bing, D. J. C. r. i. f. s., & nutrition. (2019). Composition, physicochemical properties of pea protein and its application in functional foods. 1-13. Ma, C.-y. (1984). Functional properties of acylated oat protein. Journal of Food Science, 49(4), 1128-1131. doi:10.1111/j.1365-2621.1984.tb10410.x Ma, C. Y., Khanzada, G., & Harwalkar, V. R. (1988). Thermal gelation of oat globulin. Journal of Agricultural and Food Chemistry, 36(2), 275-280. doi:10.1021/jf00080a009 Ma, Z., Boye, J. I., Simpson, B. K., Prasher, S. O., Monpetit, D., & Malcolmson, L. (2011). Thermal processing effects on the functional properties and microstructure of lentil, chickpea, and pea flours. Food Research International, 44(8), 2534-2544. doi:https://doi.org/10.1016/j.foodres.2010.12.017 Magoshi, J., Nakamura, S., & Murakami, K. I. (1992). Structure and physical properties of seed proteins. I. Glass transition and crystallization of zein protein from corn. Journal of Applied polymer Science, 45(11), 2043-2048. Manley, D. (2011). 10 - meals, grits, flours and starches (other than wheat). In D. Manley (Ed.), Manley’s technology of biscuits, crackers and cookies (fourth edition) (pp. 134-142): Woodhead Publishing. Mann, J., Schiedt, B., Baumann, A., Conde-Petit, B., & Vilgis, T. A. (2014). Effect of heat treatment on wheat dough rheology and wheat protein solubility. Food Science and Technology International, 20(5), 341-351.

166

Margulis, M., & Margulis, I. (2003). Calorimetric method for measurement of acoustic power absorbed in a volume of a liquid. Ultrasonics Sonochemistry, 10(6), 343- 345. Marracino, P., Apollonio, F., Liberti, M., d’Inzeo, G., & Amadei, A. (2013). Effect of high exogenous electric pulses on protein conformation: Myoglobin as a case study. The Journal of Physical Chemistry B, 117(8), 2273-2279. Martínez-Velasco, A., Lobato-Calleros, C., Hernández-Rodríguez, B. E., Román- Guerrero, A., Alvarez-Ramirez, J., & Vernon-Carter, E. J. (2018). High intensity ultrasound treatment of faba bean (vicia faba l.) protein: Effect on surface properties, foaming ability and structural changes. Ultrasonics Sonochemistry, 44, 97-105. McWatters, K. H., & Holmes, M. R. (1979). Influence of moist heat on solubility and emulsification properties of soy and peanut flours. Journal of Food Science, 44(3), 774-776. Meehan, J. J. (1952). Method for selective sterilization of food products: Google Patents. Mertz, E. T., Hassen, M. M., Cairns-Whittern, C., Kirleis, A. W., Tu, L., & Axtell, J. D. (1984). Pepsin digestibility of proteins in sorghum and other major cereals. Proceedings of the National Academy of Sciences, 81(1), 1-2. Mesa‐Stonestreet, D., Jhoe, N., Alavi, S., & Bean, S. R. (2010). Sorghum proteins: The concentration, isolation, modification, and food applications of kafirins. Journal of Food Science, 75(5). Messens, W., Van Camp, J., & Huyghebaert, A. (1997). The use of high pressure to modify the functionality of food proteins. Trends in Food Science & Technology, 8(4), 107- 112. Mession, J.-L., Chihi, M. L., Sok, N., & Saurel, R. (2015). Effect of globular pea proteins fractionation on their heat-induced aggregation and acid cold-set gelation. Food Hydrocolloids, 46, 233-243. Michael, B. (2012). Proso millet. Retrieved from http://www.agmrc.org/commodities__products/grains__oilseeds/proso-mi

167

Mimouni, B., Azanza, J. L., & Raymond, J. (1999). Influence of double enzymic hydrolyses on gluten functionality. Journal of the Science of Food and Agriculture, 79(7), 1048-1053. Mirmoghtadaie, L., Kadivar, M., & Shahedi, M. (2009). Effects of succinylation and deamidation on functional properties of oat protein isolate. Food Chemistry, 114(1), 127-131. Mohammed, K., Obadi, M., Omedi, J. O., Letsididi, K. S., Koko, M., Zaaboul, F., . . . Research, I. (2018). Effect of sunflower meal protein isolate (smpi) addition on wheat bread quality. 6(9), 159. Mohamed, T. K., Zhu, K., Issoufou, A., Fatmata, T., & Zhou, H. (2009). Functionality, in vitro digestibility and physicochemical properties of two varieties of defatted foxtail millet protein concentrates. International journal of molecular sciences, 10(12), 5224-5238. Mokni Ghribi, A., Maklouf Gafsi, I., Sila, A., Blecker, C., Danthine, S., Attia, H., . . . Besbes, S. (2015). Effects of enzymatic hydrolysis on conformational and functional properties of chickpea protein isolate. Food Chemistry, 187, 322-330. doi:https://doi.org/10.1016/j.foodchem.2015.04.109 Molina, E., Defaye, A. B., & Ledward, D. A. (2002). Soy protein pressure-induced gels. Food Hydrocolloids, 16(6), 625-632. doi:https://doi.org/10.1016/S0268- 005X(02)00028-0 Mondoulet, L., Paty, E., Drumare, M. F., Ah-Leung, S., Scheinmann, P., Willemot, R. M., . . . Bernard, H. (2005). Influence of thermal processing on the allergenicity of peanut proteins. Journal of Agricultural and Food Chemistry, 53(11), 4547-4553. doi:10.1021/jf050091p Moradian-Oldak, J., Paine, M. L., Lei, Y. P., Fincham, A. G., & Snead, M. L. (2000). Self- assembly properties of recombinant engineered amelogenin proteins analyzed by dynamic light scattering and atomic force microscopy. Journal of Structural Biology, 131(1), 27-37. doi:https://doi.org/10.1006/jsbi.2000.4237 Morales, R., Martínez, K. D., Ruiz-Henestrosa, V. M. P., & Pilosof, A. M. (2015). Modification of foaming properties of soy protein isolate by high ultrasound intensity: Particle size effect. Ultrasonics Sonochemistry, 26, 48-55.

168

Mordor_ Intelligence. (2017). Global plant protein market - by sources, industry, regions - market size, demand forecasts, industry trends and updates (2016-2022). Retrieved from https://www.researchandmarkets.com/reports/4342820/global-plant-protein- market-by-sources#relb1 Mordor_ Intelligence. (2019). Plant protein market - growth, trends, and forcast (2019 - 2024). Retrieved from https://www.mordorintelligence.com/industry- reports/plant-protein-market Motoi, H., Fukudome, S., & Urabe, I. (2004). Continuous production of wheat gluten peptide with foaming properties using immobilized enzymes. European Food Research and Technology, 219(5), 522-528. doi:10.1007/s00217-004-0986-2 Murugesan, R. (2015). Use of millets for partial wheat replacement in bakery products. McGill University Libraries. Nakai, S. (1983). Structure-function relationships of food proteins: With an emphasis on the importance of protein hydrophobicity. Journal of Agricultural and Food Chemistry, 31(4), 676-683. Nazari, B., Mohammadifar, M., Shojaee-Aliabadi, S., & Mirmoghtadaie, L. (2016). Applications of high intensity ultrasound to modify millet protein concentrate functionality. International Journal of Nutrition and Food Engineering, 10(9), 568- 571. Nazari, B., Mohammadifar, M. A., Shojaee-Aliabadi, S., Feizollahi, E., & Mirmoghtadaie, L. (2018). Effect of ultrasound treatments on functional properties and structure of millet protein concentrate. Ultrasonics Sonochemistry, 41, 382-388. doi:https://doi.org/10.1016/j.ultsonch.2017.10.002 Newman, P. G., & Rozycki, G. S. (1998). The history of ultrasound. Surgical Clinics of North America, 78(2), 179-195. doi:https://doi.org/10.1016/S0039- 6109(05)70308-X Nguyen, T. T. P., Bhandari, B., Cichero, J., & Prakash, S. (2015). Gastrointestinal digestion of dairy and soy proteins in infant formulas: An in vitro study. Food Research International, 76, 348-358. doi:https://doi.org/10.1016/j.foodres.2015.07.030 Nirgude, M., Babu, B. K., Shambhavi, Y., Singh, U., Upadhyaya, H., & Kumar, A. (2014). Development and molecular characterization of genic molecular markers for grain

169

protein and calcium content in finger millet (eleusine coracana (l.) gaertn.). Molecular biology reports, 41(3), 1189-1200. Nishizawa, N., & Fudamoto, Y. (1995). The elevation of plasma concentration of high- density lipoprotein cholesterol in mice fed with protein from proso millet. Bioscience, biotechnology, and biochemistry, 59(2), 333-335. Nivala, O., Mäkinen, O. E., Kruus, K., Nordlund, E., & Ercili-Cura, D. (2017). Structuring colloidal oat and faba bean protein particles via enzymatic modification. Food Chemistry, 231, 87-95. doi:https://doi.org/10.1016/j.foodchem.2017.03.114 Nowak-Wegrzyn, A., & Fiocchi, A. (2009). Rare, medium, or well done? The effect of heating and food matrix on food protein allergenicity. Current opinion in allergy and clinical immunology, 9(3), 234-237. Nwachukwu, I. D., & Aluko, R. E. (2018). Physicochemical and emulsification properties of flaxseed (linum usitatissimum) albumin and globulin fractions. Food Chemistry, 255, 216-225. O'Kane, F. E., Happe, R. P., Vereijken, J. M., Gruppen, H., & van Boekel, M. A. (2004). Heat-induced gelation of pea legumin: Comparison with soybean glycinin. Journal of Agricultural and Food Chemistry, 52(16), 5071-5078. O'Kane, F. E., Vereijken, J. M., Gruppen, H., & Van Boekel, M. A. (2005). Gelation behavior of protein isolates extracted from 5 cultivars of pisum sativum l. Journal of Food Science, 70(2), C132-C137. Osborne, T. B. (1916). The vegetable proteins: Longmans, Green and Company. O'sullivan, J., Murray, B., Flynn, C., & Norton, I. (2016). The effect of ultrasound treatment on the structural, physical and emulsifying properties of animal and vegetable proteins. Food Hydrocolloids, 53, 141-154. O’Sullivan, J. J., Park, M., Beevers, J., Greenwood, R. W., & Norton, I. T. (2017). Applications of ultrasound for the functional modification of proteins and nanoemulsion formation: A review. Food Hydrocolloids, 71, 299-310. Oliver, C. M., Melton, L. D., & Stanley, R. A. (2006). Creating proteins with novel functionality via the maillard reaction: A review. Critical Reviews in Food Science and Nutrition, 46(4), 337-350.

170

Pan, M., Xu, F., Wu, Y., Yao, M., Xiao, X., Zhang, N., . . . Wang, L. (2020). Application of ultrasound-assisted physical mixing treatment improves in vitro protein digestibility of rapeseed napin. Ultrasonics Sonochemistry, 67, 105136. doi:https://doi.org/10.1016/j.ultsonch.2020.105136 Pan, X., Fang, Y., Wang, L., Xie, M., Hu, B., Zhu, Y., . . . Li, P. (2019). Effect of enzyme types on the stability of oil‐in‐water emulsions formed with rice protein hydrolysates. Journal of the Science of Food and Agriculture, 99(15), 6731-6740. Pantoja-Uceda, D., Shewry, P. R., Bruix, M., Tatham, A. S., Santoro, J., & Rico, M. (2004). Solution structure of a methionine-rich 2s albumin from sunflower seeds: Relationship to its allergenic and emulsifying properties. Biochemistry, 43(22), 6976-6986. Panyam, D., & Kilara, A. (1996). Enhancing the functionality of food proteins by enzymatic modification. Trends in Food Science & Technology, 7(4), 120-125. Parameswaran, K. P., & Thayumanavan, B. (1995). Homologies between prolamins of different minor millets. Plant Foods for Human Nutrition, 48(2), 119-126. doi:10.1007/bf01088307 Park, K.-O., Ito, Y., Nagasawa, T., Choi, M.-R., & Nishizawa, N. (2008). Effects of dietary korean proso-millet protein on plasma adiponectin, hdl cholesterol, insulin levels, and gene expression in obese type 2 diabetic mice. Bioscience, biotechnology, and biochemistry, 72(11), 2918-2925. doi:10.1271/bbb.80395 Parrinello, M., & Rahman, A. (1980). Crystal structure and pair potentials: A molecular- dynamics study. Physical Review Letters, 45(14), 1196. Paul, G. L. (2009). The rationale for consuming protein blends in sports nutrition. Journal of the American College of Nutrition, 28(sup4), 464S-472S. Pearce, K. N., & Kinsella, J. E. (1978). Emulsifying properties of proteins: Evaluation of a turbidimetric technique. Journal of Agricultural and Food Chemistry, 26(3), 716- 723. Peng, W., Kong, X., Chen, Y., Zhang, C., Yang, Y., & Hua, Y. (2016). Effects of heat treatment on the emulsifying properties of pea proteins. Food Hydrocolloids, 52, 301-310.

171

Pereira, J., Gh, Z., Zhang, W. J. J. o. F., & Research, N. (2016). Effects of rice flour on emulsion stability, organoleptic characteristics and thermal rheology of emulsified sausage. 4(4), 216-222. Phillips, L. G. (2013). Structure-function properties of food proteins: Academic Press. Phillips, S. M. (2004). Protein requirements and supplementation in strength sports. Nutrition, 20(7-8), 689-695. Pietsch, V. L., Emin, M. A., & Schuchmann, H. P. J. J. o. F. E. (2017). Process conditions influencing wheat gluten polymerization during high moisture extrusion of meat analog products. 198, 28-35. Puppo, C., Chapleau, N., Speroni, F., de Lamballerie-Anton, M., Michel, F., Añón, C., & Anton, M. (2004). Physicochemical modifications of high-pressure-treated soybean protein isolates. Journal of Agricultural and Food Chemistry, 52(6), 1564-1571. Qazi, H. A., Jan, N., Ramazan, S., & John, R. (2019). Chapter 8 - protein modification in plants in response to abiotic stress. In T. A. Dar & L. R. Singh (Eds.), Protein modificomics (pp. 171-201): Academic Press. Qiu, C., Sun, W., Cui, C., & Zhao, M. (2013). Effect of citric acid deamidation on in vitro digestibility and antioxidant properties of wheat gluten. Food Chemistry, 141(3), 2772-2778. doi:https://doi.org/10.1016/j.foodchem.2013.05.072 Qu, W., Zhang, X., Han, X., Wang, Z., He, R., & Ma, H. (2018). Structure and functional characteristics of rapeseed protein isolate-dextran conjugates. Food Hydrocolloids, 82, 329-337. Rao, B. R., Nagasampige, M. H., Ravikiran, M. J. J. o. P., & Sciences, B. (2011). Evaluation of nutraceutical properties of selected small millets. 3(2), 277. Ravindran, G. (1992). Seed protein of millets: Amino acid composition, proteinase inhibitors and in-vitro protein digestibility. Food Chemistry, 44(1), 13-17. Ravindran, G. (1992). Seed protein of millets: Amino acid composition, proteinase inhibitors and in-vitro protein digestibility. Food Chemistry, 44(1), 13-17. Rehman, Z.-u., & Shah, W. (2005). Thermal heat processing effects on antinutrients, protein and starch digestibility of food legumes. Food Chemistry, 91(2), 327-331.

172

Ren, X., Wei, X., Ma, H., Zhou, H., Guo, J., Mao, S., & Hu, A. (2015). Effects of a dual‐ frequency frequency‐sweeping ultrasound treatment on the properties and structure of the zein protein. Cereal Chemistry, 92(2), 193-197. Riha, W. E., Izzo, H. V., Zhang, J., & Ho, C. T. (1996). Nonenzymatic deamidation of food proteins. Critical Reviews in Food Science and Nutrition, 36(3), 225-255. doi:10.1080/10408399609527724 Rioux, V. (2016). Fatty acid acylation of proteins: Specific roles for palmitic, myristic and caprylic acids. Robertson, M. J., Tirado-Rives, J., & Jorgensen, W. L. (2015). Improved peptide and protein torsional energetics with the opls-aa force field. Journal of Chemical Theory and Computation, 11(7), 3499-3509. doi:10.1021/acs.jctc.5b00356 Sainani, M. N., Gupta, V. S., Mishra, V. K., Lachke, A. H., Ranjekar, P. K., & Pillay, D. (1993). Effect of chemical modification on some structural and functional properties of pennisetin, a major seed storage protein from pearl millet. Phytochemistry, 34(4), 919-925. Saleh, A. S. M., Zhang, Q., Chen, J., & Shen, Q. (2013). Millet grains: Nutritional quality, processing, and potential health benefits. Comprehensive Reviews in Food Science and Food Safety, 12(3), 281-295. doi:10.1111/1541-4337.12012 Samard, S., Gu, B.-Y., & Ryu, G.-H. (2019). Effects of extrusion types, screw speed and addition of wheat gluten on physicochemical characteristics and cooking stability of meat analogues. 99(11), 4922-4931. doi:10.1002/jsfa.9722 Sánchez-Reséndiz, A., Rodríguez-Barrientos, S., Rodríguez-Rodríguez, J., Barba-Dávila, B., Serna-Saldívar, S. O., & Chuck-Hernández, C. J. F. c. (2018). Phosphoesterification of soybean and peanut proteins with sodium trimetaphosphate (stmp): Changes in structure to improve functionality for food applications. 260, 299-305. Schreuders, F. K., Dekkers, B. L., Bodnár, I., Erni, P., Boom, R. M., & van der Goot, A. J. J. J. o. F. E. (2019). Comparing structuring potential of pea and soy protein with gluten for meat analogue preparation. 261, 32-39.

173

Selinheimo, E., Autio, K., Kruus, K., & Buchert, J. (2007). Elucidating the mechanism of laccase and tyrosinase in wheat bread making. Journal of Agricultural and Food Chemistry, 55(15), 6357-6365. doi:10.1021/jf0703349 Sethi, S., Tyagi, S. K., & Anurag, R. K. (2016). Plant-based milk alternatives an emerging segment of functional beverages: A review. Journal of Food Science and Technology, 53(9), 3408-3423. Shah, N. N., K.V, U., & Singhal, R. S. (2019). Hydrophobically modified pea proteins: Synthesis, characterization and evaluation as emulsifiers in eggless cake. Journal of Food Engineering, 255, 15-23. doi:https://doi.org/10.1016/j.jfoodeng.2019.03.005 Shand, P. J., Ya, H., Pietrasik, Z., & Wanasundara, P. K. J. P. D. (2007). Physicochemical and textural properties of heat-induced pea protein isolate gels. Food Chemistry, 102(4), 1119-1130. doi:https://doi.org/10.1016/j.foodchem.2006.06.060 Shao, Y., & Tang, C.-H. (2014). Characteristics and oxidative stability of soy protein- stabilized oil-in-water emulsions: Influence of ionic strength and heat pretreatment. Food Hydrocolloids, 37, 149-158. doi:https://doi.org/10.1016/j.foodhyd.2013.10.030 Sharif, H. R., Williams, P. A., Sharif, M. K., Abbas, S., Majeed, H., Masamba, K. G., . . . Zhong, F. (2018). Current progress in the utilization of native and modified legume proteins as emulsifiers and encapsulants–a review. Food Hydrocolloids, 76, 2-16. Sheahan, C. M. (2014). Plant guide for proso millet (panicum miliaceum). Retrieved from USDA. Shewry, P. R., & Halford, N. G. (2002). Cereal seed storage proteins: Structures, properties and role in grain utilization. Journal of experimental botany, 53(370), 947-958. Shi, J., Ma, X., Zhang, J., Zhou, Y., Liu, M., Huang, L., . . . Lai, J. (2019). Chromosome conformation capture resolved near complete genome assembly of broomcorn millet. Nature Communications, 10(1), 464. doi:10.1038/s41467-018-07876-6 Shimanuki, S., Nagasawa, T., & Nishizawa, N. (2006). Plasma hdl subfraction levels increase in rats fed proso-millet protein concentrate. Medical science monitor, 12(7), BR221-BR226.

174

Shin, W.-H., Lee, G. R., Heo, L., Lee, H., & Seok, C. J. B. D. (2014). Prediction of protein structure and interaction by galaxy protein modeling programs. 2(1), 1-11. Sim, S. Y., Karwe, M. V., & Moraru, C. I. (2019). High pressure structuring of pea protein concentrates. Journal of Food Process Engineering, 42(7), e13261. Simjian, L. G. (1959). Method of treating food articles by application of sound energy: Google Patents. Singh, A., Orsat, V., & Raghavan, V. (2013). Soybean hydrophobic protein response to external electric field: A molecular modeling approach. Biomolecules, 3(1), 168- 179. Singh, M., Adedeji, A. A., & Santra, D. (2018). Physico-chemical and functional properties of nine proso millet cultivars. Singh, P., Kumar, R., Sabapathy, S., & Bawa, A. (2008). Functional and edible uses of soy protein products. Comprehensive Reviews in Food Science and Food Safety, 7(1), 14-28. Sippl, M. J. J. P. S., Function,, & Bioinformatics. (1993). Recognition of errors in three‐ dimensional structures of proteins. 17(4), 355-362. Soenen, S., & Westerterp-Plantenga, M. S. (2008). Proteins and satiety: Implications for weight management. Current Opinion in Clinical Nutrition & Metabolic Care, 11(6), 747-751. Solá, R., Rodriguez-Martinez, J., & Griebenow, K. (2007). Modulation of protein biophysical properties by chemical glycosylation: Biochemical insights and biomedical implications. Cellular and Molecular Life Sciences, 64(16), 2133-2152. Song, Y., DiMaio, F., Wang, R. Y.-R., Kim, D., Miles, C., Brunette, T., . . . Baker, D. J. S. (2013). High-resolution comparative modeling with rosettacm. 21(10), 1735-1742. Studer, G., Rempfer, C., Waterhouse, A. M., Gumienny, R., Haas, J., & Schwede, T. (2019). Qmeandisco—distance constraints applied on model quality estimation. Bioinformatics, 36(6), 1765-1771. doi:10.1093/bioinformatics/btz828 %J Bioinformatics Sun-Waterhouse, D., Zhao, M., & Waterhouse, G. I. (2014). Protein modification during ingredient preparation and food processing: Approaches to improve food processability and nutrition. Food and Bioprocess Technology, 7(7), 1853-1893.

175

Sun, X. D., & Arntfield, S. D. (2011). Gelation properties of salt-extracted pea protein isolate catalyzed by microbial transglutaminase cross-linking. Food Hydrocolloids, 25(1), 25-31. doi:https://doi.org/10.1016/j.foodhyd.2010.05.002 Sun, X. D., & Arntfield, S. D. (2011). Gelation properties of salt-extracted pea protein isolate induced by heat treatment: Effect of heating and cooling rate. Food Chemistry, 124(3), 1011-1016. Sun, X. D., Lan, Y., Shi, D., Lu, S. W., Liao, H., Zhang, R. Y., . . . Shan, H. (2015). Determination of molecular driving forces involved in heat-induced corn germ proteins gelation. Journal of Cereal Science, 66, 24-30. doi:https://doi.org/10.1016/j.jcs.2015.09.007 Taghdir, M., Mazloomi, S. M., Honar, N., Sepandi, M., Ashourpour, M., Salehi, M. J. F. s., & nutrition. (2017). Effect of soy flour on nutritional, physicochemical, and sensory characteristics of gluten‐free bread. 5(3), 439-445. Taha, A., Hu, T., Zhang, Z., Bakry, A. M., Khalifa, I., Pan, S., & Hu, H. (2018). Effect of different oils and ultrasound emulsification conditions on the physicochemical properties of emulsions stabilized by soy protein isolate. Ultrasonics Sonochemistry, 49, 283-293. Tandang-Silvas, M. R., Cabanos, C. S., Peña, L. D. C., De La Rosa, A. P. B., Osuna-Castro, J. A., Utsumi, S., . . . Maruyama, N. (2012). Crystal structure of a major seed storage protein, 11s proglobulin, from amaranthus hypochondriacus: Insight into its physico-chemical properties. Food Chemistry, 135(2), 819-826. Tandang-Silvas, M. R. G., Fukuda, T., Fukuda, C., Prak, K., Cabanos, C., Kimura, A., . . . Proteomics. (2010). Conservation and divergence on plant seed 11s globulins based on crystal structures. 1804(7), 1432-1442. Tang, S., Hettiarachchy, N., Horax, R., & Eswaranandam, S. (2003). Physicochemical properties and functionality of rice bran protein hydrolyzate prepared from heat‐ stabilized defatted rice bran with the aid of enzymes. Journal of Food Science, 68(1), 152-157. Taylor, J. R., Taylor, J., Campanella, O. H., & Hamaker, B. R. (2016). Functionality of the storage proteins in gluten-free cereals and pseudocereals in dough systems. Journal of Cereal Science, 67, 22-34.

176

Touw, W. G., Baakman, C., Black, J., Te Beek, T. A., Krieger, E., Joosten, R. P., & Vriend, G. J. N. a. r. (2015). A series of pdb-related databanks for everyday needs. 43(D1), D364-D368. Utsumi, S., & Kinsella, J. E. (1985). Structure-function relationships in food proteins: Subunit interactions in heat-induced gelation of 7s, 11s, and soy isolate proteins. Journal of Agricultural and Food Chemistry, 33(2), 297-303. Vagadia, B. H., Vanga, S. K., Singh, A., & Raghavan, V. (2016). Effects of thermal and electric fields on soybean trypsin inhibitor protein: A molecular modelling study. Innovative Food Science & Emerging Technologies, 35, 9-20. Valencia‐Flores, D. C., Hernández‐Herrero, M., Guamis, B., & Ferragut, V. (2013). Comparing the effects of ultra‐high‐pressure homogenization and conventional thermal treatments on the microbiological, physical, and chemical quality of almond beverages. Journal of Food Science, 78(2), E199-E205. van den Broeck, H. C., America, A. H., Smulders, M. J., Bosch, D., Hamer, R. J., Gilissen, L. J., & van der Meer, I. M. (2009). A modified extraction protocol enables detection and quantification of celiac disease-related gluten proteins from wheat. Journal of Chromatography B, 877(10), 975-982. Van Der Spoel, D., Lindahl, E., Hess, B., Groenhof, G., Mark, A. E., & Berendsen, H. J. (2005). Gromacs: Fast, flexible, and free. Journal of computational chemistry, 26(16), 1701-1718. Vanga, S. K., Singh, A., & Raghavan, V. (2015). Effect of thermal and electric field treatment on the conformation of ara h 6 peanut protein allergen. Innovative Food Science & Emerging Technologies, 30, 79-88. doi:https://doi.org/10.1016/j.ifset.2015.03.003 Vanga, S. K., Singh, A., & Raghavan, V. (2018). Changes in soybean trypsin inhibitor by varying pressure and temperature of processing: A molecular modeling study. Innovative Food Science & Emerging Technologies, 49, 31-40. doi:https://doi.org/10.1016/j.ifset.2018.07.015 Vanga, S. K., Wang, J., Singh, A., Raghavan, V. (2019). Simulations of temperature and pressure unfolding in soy allergen gly m 4 using molecular modeling.Journal of Agricultural Chemistry 67(45), 12547-12557.

177

Venkataratnam, H., Sarangapani, C., Cahill, O., & Ryan, C. B. (2019). Effect of cold plasma treatment on the antigenicity of peanut allergen ara h 1. Innovative Food Science & Emerging Technologies, 52, 368-375. Ventura, S., & Villaverde, A. (2006). Protein quality in bacterial inclusion bodies. Trends in Biotechnology, 24(4), 179-185. doi:https://doi.org/10.1016/j.tibtech.2006.02.007 Villarino, C. B. J., Jayasena, V., Coorey, R., Chakrabarti-Bell, S., & Johnson, S. K. (2016). Nutritional, health, and technological functionality of lupin flour addition to bread and other baked products: Benefits and challenges. Critical Reviews in Food Science and Nutrition, 56(5), 835-857. doi:10.1080/10408398.2013.814044 Vioque, J., Sánchez-Vioque, R., Clemente, A., Pedroche, J., & Millán, F. (2000). Partially hydrolyzed rapeseed protein isolates with improved functional properties. Journal of the American Oil Chemists' Society, 77(4), 447-450. doi:10.1007/s11746-000- 0072-y Wagner, J. M., Evans, T. J., & Korotkov, K. V. (2014). Crystal structure of the n‐terminal domain of ecca1 atpase from the esx‐1 secretion system of mycobacterium tuberculosis. Proteins: Structure, Function, and Bioinformatics, 82(1), 159-163. Wan, Y., Liu, J., & Guo, S. (2018). Effects of succinylation on the structure and thermal aggregation of soy protein isolate. Food Chemistry, 245, 542-550. doi:https://doi.org/10.1016/j.foodchem.2017.10.137 Wanasundara, P., & Shahidi, F. (1997). Functional properties of acylated flax protein isolates. Journal of Agricultural and Food Chemistry, 45(7), 2431-2441. Wang, K.-Q., Luo, S.-Z., Zhong, X.-Y., Cai, J., Jiang, S.-T., & Zheng, Z. (2017). Changes in chemical interactions and protein conformation during heat-induced wheat gluten gel formation. Food Chemistry, 214, 393-399. Wang, K., Sun, D.-W., Pu, H., & Wei, Q. (2017). Principles and applications of spectroscopic techniques for evaluating food protein conformational changes: A review. Trends in Food Science & Technology, 67, 207-219. doi:https://doi.org/10.1016/j.tifs.2017.06.015 Wang, L., Gulati, P., Santra, D., Rose, D., & Zhang, Y. (2018). Nanoparticles prepared by proso millet protein as novel curcumin delivery system. Food Chemistry,

178

240(Supplement C), 1039-1046. doi:https://doi.org/10.1016/j.foodchem.2017.08.036 Wang, X.-S., Tang, C.-H., Li, B.-S., Yang, X.-Q., Li, L., & Ma, C.-Y. (2008). Effects of high-pressure treatment on some physicochemical and functional properties of soy protein isolates. Food Hydrocolloids, 22(4), 560-567. Wang, X., Li, Y., He, X., Chen, S., & Zhang, J. Z. (2014). Effect of strong electric field on the conformational integrity of insulin. The Journal of Physical Chemistry A, 118(39), 8942-8952. Wang, X. B., Zhang, Y. H., & Jiang, L. Z. (2013). Improvement of emulsifying properties of soybean protein isolate through glycosylation modification. Paper presented at the Advanced Materials Research. Wang, Z., Zhang, C., Zhang, T., Ju, X., & He, R. (2018). Effects of acylation and glycation treatments on physicochemical and gelation properties of rapeseed protein isolate. RSC advances, 8(70), 40395-40406. Waterhouse, A., Bertoni, M., Bienert, S., Studer, G., Tauriello, G., Gumienny, R., . . . Bordoli, L. J. N. a. r. (2018). Swiss-model: Homology modelling of protein structures and complexes. 46(W1), W296-W303. Wei, C.-K., Thakur, K., Liu, D.-H., Zhang, J.-G., & Wei, Z.-J. (2018). Enzymatic hydrolysis of flaxseed (linum usitatissimum l.) protein and sensory characterization of maillard reaction products. Food Chemistry, 263, 186-193. doi:https://doi.org/10.1016/j.foodchem.2018.04.120 Wen, C., Zhang, J., Yao, H., Zhou, J., Duan, Y., Zhang, H., & Ma, H. (2019). Advances in renewable plant-derived protein source: The structure, physicochemical properties affected by ultrasonication. Ultrasonics Sonochemistry, 53, 83-98. doi:https://doi.org/10.1016/j.ultsonch.2018.12.036 Wiederstein, M., & Sippl, M. J. J. N. a. r. (2007). Prosa-web: Interactive web service for the recognition of errors in three-dimensional structures of proteins. 35(suppl_2), W407-W410. Withana-Gamage, T. S., & Wanasundara, J. P. (2012). Molecular modelling for investigating structure–function relationships of soy glycinin. Trends in Food Science & Technology, 28(2), 153-167.

179

Wouters, A. G., & Delcour, J. A. (2019). Cereal protein based nanoparticles as agents stabilizing air-water and oil-water interfaces in food systems. Current Opinion in Food Science. Wouters, A. G., & Delcour, J. A. (2019). Cereal protein based nanoparticles as agents stabilizing air-water and oil-water interfaces in food systems. Current Opinion in Food Science. Wouters, A. G., Rombouts, I., Fierens, E., Brijs, K., & Delcour, J. A. (2018). Enzymatically hydrolyzed wheat gluten as a foaming agent in food: Incorporation in a meringue recipe as a proof‐of‐concept. Journal of food sceince 83(8), 2119-2126. Wu, W., Hettiarachchy, N., & Qi, M. (1998). Hydrophobicity, solubility, and emulsifying properties of soy protein peptides prepared by papain modification and ultrafiltration. Journal of the American Oil Chemists' Society, 75(7), 845-850. Xiao, J., Wang, X. a., Gonzalez, A. J. P., & Huang, Q. (2016). Kafirin nanoparticles- stabilized pickering emulsions: Microstructure and rheological behavior. Food Hydrocolloids, 54, 30-39. Xu, Z.-Z., Huang, G.-Q., Xu, T.-C., Liu, L.-N., & Xiao, J.-X. (2019). Comparative study on the maillard reaction of chitosan oligosaccharide and glucose with soybean protein isolate. International journal of biological macromolecules, 131, 601-607. Yang, J., & Powers, J. R. (2016). Effects of high pressure on food proteins High pressure processing of food (pp. 353-389): Springer. Yoshie-Stark, Y., Wada, Y., & Wäsche, A. (2008). Chemical composition, functional properties, and bioactivities of rapeseed protein isolates. Food Chemistry, 107(1), 32-39. doi:https://doi.org/10.1016/j.foodchem.2007.07.061 Yu, N., Qiuxia, M., & Lizhen, Z. (2012). Analysis of protein content in panicum miliaceum l. Using near infrared transmittance spectroscopy. Agricultural Science and Technology.

Yust, M. d. M., Pedroche, J., Millán-Linares, M. d. C., Alcaide-Hidalgo, J. M., & Millán, F. (2010). Improvement of functional properties of chickpea proteins by hydrolysis

180

with immobilised alcalase. Food Chemistry, 122(4), 1212-1217. doi:https://doi.org/10.1016/j.foodchem.2010.03.121 Zha, F., Dong, S., Rao, J., & Chen, B. (2019a). Pea protein isolate-gum arabic maillard conjugates improves physical and oxidative stability of oil-in-water emulsions. Food Chemistry, 285, 130-138. Zha, F., Dong, S., Rao, J., & Chen, B. (2019b). The structural modification of pea protein concentrate with gum arabic by controlled maillard reaction enhances its functional properties and flavor attributes. Food Hydrocolloids, 92, 30-40. Zhang, H., Claver, I. P., Zhu, K.-X., & Zhou, H. (2011). The effect of ultrasound on the functional properties of wheat gluten. Molecules, 16(5), 4231-4240. Zhang, H. J., Zhang, H., Wang, L., & Guo, X. N. (2012). Preparation and functional properties of rice bran proteins from heat-stabilized defatted rice bran. Food Research International, 47(2), 359-363. doi:https://doi.org/10.1016/j.foodres.2011.08.014 Zhang, Q., Li, L., Lan, Q., Li, M., Wu, D., Chen, H., . . . Zhang, Z. (2018). Protein glycosylation: A promising way to modify the functional properties and extend the application in food system. Critical Reviews in Food Science and Nutrition, 1-28. Zhang, X., Qi, J.-R., Li, K.-K., Yin, S.-W., Wang, J.-M., Zhu, J.-H., & Yang, X.-Q. (2012). Characterization of soy β-conglycinin–dextran conjugate prepared by maillard reaction in crowded liquid system. Food Research International, 49(2), 648-654. Zhao, J., Tian, Z., & Chen, L. (2010). Effects of deamidation on structure and functional properties of barley hordein. Journal of Agricultural and Food Chemistry, 58(21), 11448-11455. doi:10.1021/jf102307f Zhao, J., Tian, Z., & Chen, L. (2011). Effects of deamidation on aggregation and emulsifying properties of barley glutelin. Food Chemistry, 128(4), 1029-1036. doi:https://doi.org/10.1016/j.foodchem.2011.04.009 Zhao, Z.-K., Mu, T.-H., Zhang, M., & Richel, A. (2018). Chemical forces, structure, and gelation properties of sweet potato protein as affected by ph and high hydrostatic pressure. Food and Bioprocess Technology, 11(9), 1719-1732.

181

Zheng, B.-A., Matsumura, Y., & Mori, T. (1993). Relationship between the thermal denaturation and gelling properties of legumin from broad beans. Bioscience, biotechnology, and biochemistry, 57(7), 1087-1090. Zheng, X.-q., Wang, J.-t., Liu, X.-l., Sun, Y., Zheng, Y.-j., Wang, X.-j., & Liu, Y. (2015). Effect of hydrolysis time on the physicochemical and functional properties of corn glutelin by protamex hydrolysis. Food Chemistry, 172, 407-415. doi:https://doi.org/10.1016/j.foodchem.2014.09.080 Zhu, Z., Zhu, W., Yi, J., Liu, N., Cao, Y., Lu, J., . . . McClements, D. J. (2018). Effects of sonication on the physicochemical and functional properties of walnut protein isolate. Food Research International, 106, 853-861. doi:https://doi.org/10.1016/j.foodres.2018.01.060 Zink, J., Wyrobnik, T., Prinz, T., & Schmid, M. (2016). Physical, chemical and biochemical modifications of protein-based films and coatings: An extensive review. International journal of molecular sciences, 17(9), 1376.

182

VITA EDUCATION:

MSC. West Virginia University, 2017

BSC. Obafemi Awolowo University, 2012

PROFESSIONAL EXPERIENCE:

Quality Control Supervisor, Sumal Foods Limited Ibadan Nigeria.

Teaching Instructor, Federal College of Agriculture, Ibadan Nigeria.

Research Scientist, MicroNature LLC Morgantown West Virginia

Award and Recognition

1. Best postal presentation American Society of Agricultural and Biological Engineers (PRS community), 2019

2. Bluegrass Institute of Food Technologist Clair Hicks Scholarship Recipient, 2019

3. Recipient, Carl Del Signore Foundation Graduate Scholarships, West Virginia University, 2016.

4. Outstanding MSc. Davis College of Agriculture, Natural Resources, and Design. West Virginia University. Morgantown, WV.

5. Best Graduating Student in Agricultural Engineering at Obafemi Awolowo University, Ile Ife, Nigeria, 2012.

Selected list of Publication:

1. Felix Akharume, Dipak Santra, and Akinbode Adedeji. "Physicochemical and Functional Properties of Proso Millet Storage Protein Fractions. In press at Food Hydrocolloids https://doi.org/10.1016/j.foodhyd.2019.105497. Impact factor 5.839.

183 2. Felix Akharume, Alexandra Smith, Litha Sivanandan, and Kaushlendra Singh (2019). Recent Progress on Osmo-convective Dehydration of Fruits- A Review. Journal of Food Science & Technology 4(9) p:956-96. Impact factor 1.265. 3. Samuel Ayobami Adeyemi, Surajudeen Olarewaju Obayopo and Felix Akharume. Numerical Simulations and Experimental Validation Study of a Mixed-mode Solar Dryer for Cocoa Bean. Journal of Postharvest Technology 7.3 (2019). 4. Felix Akharume and Omotayo Aregbesola. Moisture Dependent Physico-Mechanical and Aerodynamic Properties of Roselle Calyxes (Hibiscus sabdariffa). Journal of Postharvest Technology 7.1 (2019). 5. Felix Akharume, Kaushlendra Singh, and Litha Sivanandan. Effects of Liquid Smoke Infusion on Osmotic Dehydration Kinetics and Microstructural Characteristics of Apples Cubes. Journal of Food Engineering-Elsevier. 246 (2019) 51-57. https://doi.org/10.1016/j.jfoodeng.2018.10.030. Impact factor 3.625.

6. Felix Akharume, Kaushlendra Singh, Jaczynski Jacek, and Litha Sivanandan. Microbial Shelf Stability Assessment of Osmotically Dehydrated Smoky Apples. LWT-Food Science and Technology-Elsevier 90 (2018) 61- 69https://doi.org/10.1016/j.lwt.2017.12.012. Impact factor 3.714. 7. Felix Akharume, Kaushlendra Singh and Litha Sivanandan. Characteristics of Apple Juice and Sugar Infused Fresh and Frozen Blueberries. LWT-Food Science and Technology-Elsevier 73 (2016) 448-457. https://doi.org/10.1016/j.lwt.2016.06.041 Impact factor 3.714.

Conference presentations (* denotes presenter) :

2019

1. *Felix Akharume, Youling L. Xiong, and Akinbode Adedeji (2019). Effects of Transglutaminase in Product Formulation on Physico-Textural Properties of Protein Rich Extruded Snacks. ASABE Annual International Meeting, held July 7-10, 2019. Boston Marriott Copley Place, Boston, Massachusetts USA.

2. *Felix Akharume, Samuel Adeyemi and Surajudeen Obayopo (2019). A Study Numerical Simulations and Experimental Validation of a Hybrid Solar Dryer for Cocoa. 184 ASABE Annual International Meeting, held July 7-10, 2019. Boston Marriott Copley Place, Boston, Massachusetts USA.

3. *Felix Akharume, Litha Sivanandan and Kaushlendra Singh (2019). Effects of Liquid Smoke Infusion on Osmotic Dehydration Kinetics and Microstructural Characteristics of Apple Cubes. ASABE Annual International Meeting, held July 7-10, 2019. Boston Marriott Copley Place, Boston, Massachusetts USA.

4. *Felix Akharume and Akinbode Adedeji (2019). Denaturation and Hydrophobic Properties of Proso millet Protein Fractions. Institute of Food Technologists (IFT, 2017) Annual Meeting held on June 25-28, 2017, Ernest N. Morial Convention, New Orleans, Louisiana. USA.

5. Joseph Woomer, *Felix Akharume and Akinbode A. Adedeji (2019). Evaluation of bourbon spent grain as a source of dietary fiber in a millet-based extruded product. Institute of Food Technologists (IFT) Annual Meeting held on June 25-28, 2017, Ernest N. Morial Convention, New Orleans, Louisiana. USA.

2018

1. *Akinbode Adedeji, Manjot Singh, Felix Akharume, and Joseph Woomer (2018). Proso Millet Application in the Development of Gluten-Free Products. A paper presented at and published in the proceeding of the 3rd International Symposium on Broomcorn Millet held in Fort Collins Colorado from August 8 – 12, 2018. Session #:5.

2. *Felix Akharume, Akinbode Adedeji, and Konstantin Korotkov (2018). In-vitro Digestibility, Structural Particulars and Physico-chemical Properties of Proso Millet Flour Proteins. ASABE Annual International Meeting, held July 29 – August 1, 2018. Cobo Center Detroit, Michigan.

3. *Felix Akharume, Kaushlendra Singh, Jaczynski Jacek, and Litha Sivanandan. Characteristics of Liquid Smoke-Infused Osmotically Dehydrated Apples (2017). Institute of Food Technologists (IFT) Annual Meeting held on June 25-28, 2017. Las Vegas, Nevada. USA.

185