The Relationship Between Single Nucleotide Polymorphisms in Taste

Home , TAS2R38

Receptor Genes, Taste Perception and Dietary Intake

Elie Chamoun

A Thesis

presented to

The University of Guelph

In partial fulfillment of requirements

for the degree of

Doctor of Philosophy

Human Health and Nutritional Sciences

Guelph, Ontario, Canada

i ABSTRACT

THE RELATIONSHIP BETWEEN SINGLE NUCLEOTIDE POLYMORPHISMS IN TASTE

RECEPTOR GENES, TASTE PERCEPTION AND DIETARY INTAKE

Elie Chamoun Advisor: University of Guelph, 2018 Dr. David W.L. Ma

Food preferences and dietary habits are heavily influenced by taste perception, and there is growing interest in characterizing taste preferences based on genetic variation. Genetic differences in the ability to perceive key tastes may impact eating patterns. Therefore, increased understanding of taste genetics may lead to new personalized strategies, which may influence the trajectory of chronic disease risk. Recent advances show that single nucleotide polymorphisms (SNPs) in taste receptor genes are associated with changes in psychophysical measures of taste as well as eating patterns. The current understanding of how genetic variation impacts taste function and dietary habits for sweet, fat, salt, sour, and umami taste is limited and warranted due to the role of these types of taste in detecting nutrients that pose health risks when overconsumed. Results from this thesis demonstrated that specific SNPs are associated with taste sensitivity, taste preference and eating patterns. In adults, associations between SNPs and psychophysical measures of taste were observed between rs4790151 (TRPV1) and salt sensitivity, rs2499729 (GRM4) and umami sensitivity, rs713598 (TAS2R38) and bitter sensitivity, and rs236514

(KCNJ2) and sour preference. In children, associations were observed between rs4790522 (TRPV1) and salt preference, and rs173135 (KCNJ2) and sour preference. The rs9701796 (TAS1R2) sweet taste receptor SNP was significantly associated with both sweet preference and dietary intake of added sugar in children, while calories from sugar in snacks among children were associated with the rs35874116

(TAS1R2) sweet taste receptor SNP. Work from this thesis also demonstrates that higher sensitivities to salt, sweet, and umami taste are associated with decreases in the preferences for these tastes. Overall,

SNPs in taste receptor genes have been demonstrated in this thesis to associate with psychophysical

ii measures of taste and, in some instances, affect food preferences and eating patterns. This emerging and active field of taste research shows great promise to enhance our fundamental understanding of how taste receptor SNPs contribute to the genetic basis of eating patterns, health and disease. Results from this thesis provide important new insights into the relationship between SNPs in taste receptor genes, taste perception and dietary intake in both children and adults.

iii ACKNOWLEDGEMENTS

Firstly, I would like to thank my advisor Dr. David Ma for his mentorship both during my

Master’s and PhD research projects. You have played a fundamental role in my personal and professional growth over the past few years. You’ve always had my best interest in mind, and this is why I have reached the end of my PhD carrying with me a wealth of truly enriching lessons and experiences. Your professionalism and thoughtfulness have had an immeasurable impact on the way I carry myself as a student, a colleague, but also as a human being. Thank you for having such enormous confidence in me because it helped me gain confidence in my own abilities and helped me to excel as an independent thinker. As a direct result of your trust in me, I have been allowed to take on responsibilities that have honed my skills as a writer, communicator, teacher, mentor, but most of all as a leader. I wish you all of the best with your research program and believe whole-heartedly that the Guelph Family Health Study will flourish under your direction.

Next, I would like to thank my Advisory Committee members Dr. Jess Haines, Dr. Alison

Duncan, Dr. David Mutch, and Dr. Lisa Duizer. Jess, thank you for your support from the very start of my

Master’s when the Guelph Family Health Study was still just an idea. Your expertise was immensely helpful to me, especially during my Qualifying Exam when I was able to learn a great deal from you.

Alison, thank you for being such an excellent role model from the very beginning of my Master’s program. I am very fortunate to have a professor like you on my side because you care so much about the quality of my experience as a graduate student. I also deeply appreciate your mentorship in preparation for the Qualifying Exam. David, thank you for your support in the development of the genetics program of the Guelph Family Health Study. Your careful and meticulous review of my work was instrumental to my development as a scientist and your passion for nutrigenomics is not only evident, but infectious as well. Lisa, I’m going to miss our chats in the hallways of Food Science. I’m clearly not a sensory scientist, but you challenged me to push myself and encouraged me to think very deeply about my study design and methodology. I also want to thank you for your support as I prepared for the Qualifying Exam.

iv I would also like to acknowledge Dr. Nina Jones who served on my Qualifying Exam committee. Thank you to Dr. Bill Bettger and Dr. Iwona Rudkowska for serving as examiners for my thesis defense, your expertise in evaluating the thesis is truly appreciated. Thanks to Jonathan Guss for playing a pioneering role in forming the GFHS and for being such a great mentor.

Next, I would like to thank Lyn Hillyer and Angela Annis. Lyn, around the time that we met, I never anticipated what a great friendship we would have. I’ll always cherish our lovely chats over coffee and my palate thanks you for having an almost constant supply of chocolate in the lab. I have learned so much from you both academically and personally. Over the years, I have admired your quiet confidence and unbelievably kind heart. You have taught me so much about perseverance and your strength has inspired me more than you know. I wish you peace, love, and all the golf that your heart desires. Angela, where do I begin? You were the first person who made me feel like I was in the right place when I started graduate school, and I cannot thank you enough for that. Besides professors, you and I were the first members of the GFHS team. From those humble beginnings to the vibrant and successful GFHS that we now know, I will never forget all of the experiences we shared. Your patience, professionalism, and incredible work ethic shaped the person I am today and helped me to succeed in executing my own human research studies. As my Guelph mom, you were always there to hear me out when I was discouraged and made me feel more positive. Your friendship has meant so much to me and will continue to for years to come.

I would like to acknowledge past and present members of the Ma lab for all of their direct or indirect support for my graduate work. A special thanks to Jessie Burns for being the best lab mate I have ever had and for dealing with all of my horrible dad jokes. Your future will be so bright in every way imaginable. Thank you to the GFHS team past and present members. In particular, I would like to thank research assistants Charlotte Jones, Jessica Watterworth, Tory Ambrose, Sarah Wedde and Sabrina

Douglas. I would also like to thank students who processed dietary intake data including Laurie

Matthews, Julia Mirotta, Dee Muszynski and Robyn Lightfoot. A very special thank you to Angel Liu and

v Nicholas Carroll for being the best study assistants I could have asked for—I could not have run my studies without you two. Thank you also to the GFHS investigator team for supporting me in a variety of ways throughout my graduate career. I would also like to thank Dr. Gerarda Darlington, Dr. Zeny Feng and Annetta Qi for all of their help with statistical analysis. A final thanks to Kaity Roke for helping me develop protocols for my genetics work and for being a brilliant role model.

I would be remiss not to thank the HHNS department for providing such an excellent place to go to work. Thank you to the A team for your administrative support and Becki for keeping our work areas clean. I have made some unforgettable friendships in the department over the years. Thank you to everyone who crossed paths with me for being there through all of the laughs and the tougher times. In particular, I would like to thank Adrian Taylor, Shannon Klingel, and Willem Peppler. I am so fortunate to have met you three. Not only can I laugh my face off with you, but you have always been there to talk through all the bumps in the road that is graduate school. Each of you have qualities that I am lucky to have picked up throughout our friendship such as excellent work ethic, incredible perseverance, and remarkable compassion. It would be difficult to imagine my graduate career without you.

I have the deepest appreciation for my family and all of the support they have provided me all throughout my life. To my siblings Sandra and Joe, thank you for always making your little brother believe he can do anything he sets his mind to. To my parents Jeanne and Toni, your love and encouragement is seemingly limitless. Everything I have achieved is a result of your unwavering support and of raising me to be who I am today. Final thanks go to my partner Dita Moravek. As a fellow HHNS

PhD student, I am so fortunate to have a partner who shares all of my triumphs and struggles in such an intimate way. Thank you for enduring all of my antics and for being my cheerleader when I really need one. I am so excited for our future and can hardly wait to see what it brings.

vi TABLE OF CONTENTS

ABSTRACT ...... ii

ACKNOWLEDGEMENTS ...... iv

TABLE OF CONTENTS ...... vii

LIST OF TABLES ...... xii

LIST OF FIGURES ...... xiv

ABBREVIATIONS ...... xvi

CHAPTER 1 ...... 1

REVIEW OF THE LITERATURE...... 1

1.0 Introduction ...... 2

1.1 Brief overview of human taste perception ...... 3

1.2 Eating patterns may differ in children and adults ...... 4

1.3 Fat taste ...... 6

1.4 Sweet taste ...... 12

1.5 Bitter taste ...... 14

1.6 Umami taste ...... 17

1.7 Sour taste ...... 21

1.8 Salt taste ...... 22

1.9 Summary ...... 24

CHAPTER 2: ...... 32

THESIS RATIONALE AND OBJECTIVES ...... 32

vii 2.1 Rationale and Overall Objectives ...... 33

2.2 Specific Study Objectives and Hypotheses ...... 34

CHAPTER 3 RATIONALE ...... 37

CHAPTER 3 ...... 38

THE RELATIONSHIP BETWEEN SINGLE NUCLEOTIDE POLYMORPHISMS IN TASTE

RECEPTOR GENES, GUSTATORY FUNCTION, AND TASTE PREFERENCES IN YOUNG

ADULTS ...... 38

3.1 Abstract ...... 39

3.2 Introduction ...... 40

3.3 Methods...... 41

3.3.1 Participants ...... 41 3.3.2 Anthropometry, Body Composition and Blood Pressure Measurements ...... 41 3.3.3 Psychophysical Measurements ...... 41 3.3.4 SNP Selection and Genotyping ...... 44 3.3.5 Statistics ...... 44 3.4 Results ...... 45

3.4.1 Participant Characteristics ...... 45 3.4.2 Genetics and Taste Function/Preference ...... 45 3.4.3 Genetics, Taste, and Overall Health ...... 46 3.5 Discussion ...... 46

3.6 Conclusion ...... 51

3.7 Acknowledgements ...... 51

CHAPTER 4 RATIONALE ...... 52

CHAPTER 4 ...... 53

viii TASTE SENSITIVITY AND TASTE PREFERENCE MEASURES ARE CORRELATED IN

HEALTHY YOUNG ADULTS...... 53

4.1 Abstract ...... 54

4.2 Introduction ...... 55

4.3 Methods...... 56

4.3.1 Participants ...... 56 4.3.2 Psychophysical Measurements ...... 57 4.3.3 Data Analysis ...... 58 4.4 Results ...... 59

4.4.1 Correlation Analysis of Detection Threshold, Suprathreshold Sensitivity, and Taste Preference ...... 59 4.4.2 Principal Component Analysis of Detection Threshold, Suprathreshold Sensitivity, and Taste Preference ...... 60 4.5 Discussion ...... 61

4.6 Conclusion ...... 66

4.7 Acknowledgements ...... 66

CHAPTER 5 RATIONALE ...... 67

CHAPTER 5 ...... 68

SINGLE NUCLEOTIDE POLYMORPHISMS IN TASTE RECEPTOR GENES ARE ASSOCIATED

WITH SNACKING PATTERNS OF PRESCHOOL-AGED CHILDREN IN THE GUELPH FAMILY

HEALTH STUDY: A PILOT STUDY ...... 68

5.1 Abstract ...... 69

5.2 Introduction ...... 70

5.3 Materials and Methods ...... 72

5.3.1 Study Design ...... 72

ix 5.3.2 SNP Genotyping ...... 73 5.3.3 Snack Frequency, Quantity, Quality and Composition ...... 73 5.3.4 Data Analysis ...... 74 5.4 Results ...... 74

5.4.1 Participant Characteristics and Snacking Patterns ...... 74 5.4.2 Association of SNPs at the CD36, T1R2, and T2R38 Genes with Snacking Patterns ...... 74 5.5 Discussion ...... 78

5.6 Conclusions ...... 81

5.7 Acknowledgments ...... 82

CHAPTER 6 RATIONALE ...... 83

CHAPTER 6 ...... 84

THE RELATIONSHIP BETWEEN SINGLE NUCLEOTIDE POLYMORPHISMS IN TASTE

RECEPTOR GENES, TASTE FUNCTION AND DIETARY INTAKE IN PRESCHOOL-AGED

CHILDREN AND ADULTS IN THE GUELPH FAMILY HEALTH STUDY ...... 84

6.1 Abstract ...... 85

6.2 Introduction ...... 85

6.3 Methods...... 86

6.3.1 Participants ...... 86 6.3.2 Anthropometry, Body Composition and Blood Pressure Measurements ...... 87 6.3.3 SNP Selection and Genotyping ...... 87 6.3.4 Psychophysical Measurements ...... 88 6.3.5 Dietary Intake of Children ...... 91 6.3.6 Statistics ...... 91 6.4 Results ...... 92

6.4.1 Participant Characteristics ...... 92 6.4.2 Genetics and Taste Function/Preference ...... 93 6.4.3 Multiple Trait Analysis: SNPs, Taste and Dietary Intake ...... 94

x 6.5 Discussion ...... 96

6.6 Conclusion ...... 100

6.7 Acknowledgements ...... 101

CHAPTER 7: ...... 102

INTEGRATIVE DISCUSSION ...... 102

7.1 Preamble ...... 103

7.2 Summary and discussion of major findings ...... 103

7.2.1 Sweet taste ...... 103 7.2.2 Fat taste ...... 105 7.2.3 Salt taste ...... 106 7.2.4 Umami taste ...... 107 7.2.5 Sour taste ...... 108 7.2.6 Bitter taste ...... 108 7.2.7 The relationship between taste sensitivity and taste preference ...... 109 7.3 Future directions ...... 111

7.4 Conclusions ...... 113

APPENDIX A ...... 135

xi LIST OF TABLES

Table 1.1: Three types of taste bud cells and their known functions...... 5

Table 1.2: Taste as a method of oral fat detection and the role of CD36 ...... 8

Table 1.3: Associations between CD36 gene SNPs fat taste, dietary intake, and health outcomes ...... 12

Table 1.4: Associations between TAS1R2/TAS1R3 gene SNPs and sweet taste, dietary intake, and health outcomes ...... 15

Table 1.5: Associations between TAS2R38 gene SNPs and bitter/sweet taste, dietary intake, and health outcomes ...... 18

Table 1.6: Associations between TAS1R1 and TAS1R3 gene SNPs and umami taste perception ...... 21

Table 1.7: Associations between ENaC and TRPV1 gene SNPs and salt taste perception ...... 24

Table 3.1: Range of tastant concentrations used for each psychophysical test ...... 43

Table 3.2: Participant characteristics in total and separated by sex ...... 45

Table 3.3: Associations between SNPs in taste receptor genes and detection threshold, suprathreshold sensitivity, and taste preference ...... 47

Table 3.4: Multiple trait analysis of detection threshold, suprathreshold sensitivity, and taste preference with SNP predictors ...... 48

Table 4.1: Range of tastant concentrations used for each psychophysical test ...... 59

Table 4.2: Participant Characteristics ...... 60

Table 4.3: Correlation matrices for psychophysical measures of taste ...... 61

Table 5.1: Participant Characteristics ...... 75

Table 5.2: Participant Snacking Patterns ...... 76

Table 6.1: Range of tastant concentrations used for each psychophysical test ...... 91

xii Table 6.2: Adult participant characteristics in total and separated by sex ...... 93

Table 6.3: Child participant characteristics ...... 93

Table 6.4: Associations between SNPs in taste receptor genes and suprathreshold sensitivity and taste preference in adults ...... 95

Table 6.5: Associations between SNPs in taste receptor genes and taste preference in children ...... 96

Table 6.6: Multiple trait analysis of SNPs in taste receptor genes, taste preferences and dietary intake in children ...... 97

xiii LIST OF FIGURES

Figure 1.3: Schematic representation of the effect of the C allele in the rs713598 SNP in the TAS2R38 bitter taste receptor gene on bitter taste perception, bitter and sweet food selection, sugar, fruit, and vegetable intake, and biomarkers of health such as body fat % and plasma glucose...... 28

Figure 1.4: Schematic representation of the effect of the G allele in the rs34160967 SNP in the TAS1R1 umami taste receptor on umami taste perception, umami food selection, umami food intake, and biomarkers of health...... 29

Figure 1.5: Schematic representation of the potential effect of the PKD1L3+PKD2L1 sour taste receptor gene on sour taste perception, sour food selection, sour food intake, and biomarkers of health...... 30

Figure 1.6: Schematic representation of the effect of the AA genotype of the rs239345 SNP in the ENaC salt taste receptor on salt taste perception, salty food selection, salt intake, and biomarkers of health...... 31

Figure 4.1: Spearman Correlation between Umami Suprathreshold Sensitivity and Umami Preference .. 62

Figure 4.2: Spearman Correlation between Salt Suprathreshold Sensitivity and Salt Preference ...... 62

Figure 4.3: Principal Component Analysis of Detection Threshold, Suprathreshold Sensitivity, and Taste

Preference...... 63

Figure 5.1: Least square means of energy density of snacks by rs1761667 genotype ...... 76

Figure 5.2: Least square means of the percent calories from sugar in snacks by rs35874116 genotype ... 77

xiv Figure 5.3: Least square means of squared evening snack frequency by rs35874116 genotype ...... 77

Figure 5.4: Least square means of total energy density of snacks by rs713598 genotype ...... 78

xv ABBREVIATIONS

The following abbreviations are used in this thesis:

AI: Amiloride-insensitive AMP: Adenosine monophosphate AS: Amiloride-sensitive ATP: Adenosine triphosphate AUC: Area under the curve BMI: Body mass index BP: Blood pressure CD36: Cluster determinant 36 CVD: Cardiovascular disease DNA: Deoxyribonucleic acid DT: Detection threshold ENaC: Epithelial sodium channel ESHA: Elizabeth Stewart Hands and Associates FFA: Free fatty acids GEE: Generalized estimating equations GFHS: Guelph family health study GMP: 5’ribonucleotides guanosine monophosphate GPCR: G-protein coupled receptors HDL-C: High-density lipoprotein cholesterol HCT-8: Human colon carcinoma cell line IMP: Inosine monophosphate LCFA: Long-chain fatty acid LD: Linkage disequilibrium LDL-C: Low-density lipoprotein cholesterol LSM: Least squares means MAF: Minor allele frequency MetS: Metabolic syndrome mGluR1: Metabotropic glutamate receptor type 1 mGluR4: Metabotropic glutamate receptor type 4 MSG: Monosodium glutamate NaCl: Sodium chloride PCA: Principal component analysis PKDL: Polycystic kidney disease-like PR: Preference PROP: 6-n-propylthiouracil PTC: Phenylthiocarbamide REB: Research ethics board RNA: Ribonucleic acid RT-PCR: Real-time polymerase chain reaction SD: Standard deviation SE: Standard error SAS: Statistical analysis system SL: Sweet liker SNP: Single nucleotide polymorphism ST: Supertaster ST: Suprathreshold sensitivity

xvi T1R2: Type 1 receptor member 2 T2D: Type 2 diabetes T2R38: Type 2 receptor member 38 TBC: Taste bud cell TBW: Total body water TG: triglycerides TRPV1: Transient receptor potential cation channel subfamily V member 1

xvii CHAPTER 1

REVIEW OF THE LITERATURE

This chapter has been published in the journal Critical Reviews in Food Science and Nutrition and is permitted by the publisher, Taylor and Francis, to be included in this thesis.

Chamoun, E., Mutch, D. M., Allen-Vercoe, E., Buchholz, A. C., Duncan, A. M., Spriet, L. L., et al. A review of the associations between single nucleotide polymorphisms in taste receptors, eating behaviors, and health. Crit Rev Food Sci Nutr. 2018,58:194-207.

1 1.0 Introduction

Similar to many species, humans are often driven to select foods based on their hedonic characteristics [1, 2]. As such, the taste of food is an important factor when parents create dietary habits for themselves and their children [3]. In modern society, it is easy to gain access to large amounts of highly-palatable foods containing saturated fats and added sugars. Obesity, cardiovascular disease (CVD), type 2 diabetes (T2D), and metabolic syndrome (MetS—may include dyslipidemia, elevated blood pressure, insulin resistance, abdominal obesity, and pro-inflammatory and thrombotic states) are chronic pathologies which have been partially attributed to adverse eating behaviours in humans compelled by the rewarding experience of taste perception [2]. A growing body of knowledge demonstrates the importance of understanding individual responses to medicine and nutrition [4]. Emerging research suggests that genetic predisposition and non-genetic factors such as life stage, eating behaviour, physical activity, and gut microbiota also determine taste perception differently for every individual [5, 6]. It is therefore possible that due to different perceptions of taste, individual food preferences are important determinants of chronic disease risk.

Taste receptors on the tongue can be characterized in terms of their genetic variation to determine their contributions to taste sensitivity, taste preference, dietary intake and potentially the development of chronic diseases. Variations in receptor function for sweet, fat, and bitter tastes form much of the basis for the known inter-individual differences in taste perception [7]. However, more work is needed to determine if SNPs in salt, umami, and sour taste receptor genes lead to changes in taste sensitivity as well.

The putative fat taste receptor cluster determinant 36 (CD36), the sweet taste receptors type 1 member 2

(T1R2) and T1R3, the bitter taste receptor T2R38, the umami fat taste receptors T1R1 and T1R3, and the salt taste receptors epithelial sodium channel (ENaC) and transient receptor potential cation channel subfamily V member 1 (TRPV1) contain single nucleotide polymorphisms (SNPs) which may alter taste perception, food preference, and consequently metabolic and health outcomes. The forthcoming sections

2 will give a brief overview of human taste perception as well as discuss previously reported relationships between SNPs in taste receptor genes and taste perception, dietary intake, and health.

1.1 Brief overview of human taste perception

Taste or gustatory perception is one of the five traditional senses. Humans are able to taste food and other substances with the tongue papillae, which each consist of hundreds of taste buds [8]. Between

2000 and 5000 taste buds exist along the surface of the front and back of the tongue, and each taste bud is lined with 50-100 taste bud cells (TBCs) [9]. Nutrients dissolved in saliva interact with taste receptors located at the apical side of a TBC.

Taste bud cells can be separated into four morphological subtypes: Types I, II, III and IV [10]

(Table 1.1). It is thought that Type I cells play a supporting role for other TBCs, similarly to glial cells of the central nervous system. As such, Type I cells express proteins capable of clearing or degrading neurotransmitters following a synaptic event [11]. Additionally, the Type I cells comprise membrane ion channels that allow for the perception of salty taste from sodium chloride [12-14]. Type II cells can be further categorized based on the expression of sweet, bitter and umami taste receptors [15, 16]. These receptors are seven-transmembrane G-protein coupled receptors (GPCR). Sweet taste is elicited by a heterodimer of T1R2 and T1R3, whereas umami taste is elicited by a heterodimer of T1R1 and T1R3.

Bitter taste can be initiated by a number of T2R proteins, depending on the particular bitter substance consumed. All Type II cells express the signaling enzymes phospholipase C β2, transient receptor potential channel melastatin 5, and transient receptor potential channel melastatin 4 downstream of the

GPCRs, while only a subset express the G protein subunit α-gustducin [17]. Type II taste cells do not form traditional synapses with afferent nerve fibers. However, Type III cells do display presynaptic features including synaptic vesicles and vesicle fusion machinery such as synaptosomal-associated protein 25 (e.g., [18, 19]). Some Type III cells are sensors for sour and salt taste, however it has been suggested that some type II cells may respond to salt taste as well [20, 21]. Type IV cells are basal cells which have been suggested to contain precursor populations that can differentiate into other TBC types

3 [22]. It is not known which TBC types contain the putative fat taste receptor CD36, but it has been hypothesized that Type II and/or Type III cells contain this receptor [23]. One study used the cell culture method to determine that CD36 is expressed in a taste bud cell line (HTC-8) similar to Type II cells [24].

Small molecule neurotransmitters signal between taste cells and/or between taste cells and intragemmal nerve fibers. These molecules and their functions in taste perception have been reviewed in detail by Chaudhari and Roper [10]. As mentioned, Type II cells do not exhibit typical synapses. Instead,

Type II cells respond to stimuli by releasing adenosine triphosphate (ATP) and acetylcholine through hemichannels [25-28] which, in turn, activate purinergic receptors on cranial nerve fibers [29-32]. Type

III cells, however, respond to stimulation by releasing serotonin, norepinephrine and L-aminobutyric acid

[33-39]. The primary neurotransmitter that allows for taste cells to associate with afferent nerves appears to be ATP [25], while other neurotransmitters may mediate important taste cell functions through autocrine and paracrine signaling. Thus, other neurotransmitters may also contribute to the output of the taste bud [10]. Fatty acids, sugars/sweeteners, bitter compounds, sodium chloride, monosodium glutamate, and protons are examples of stimuli which have the potential to initiate synaptic events leading to the sensation of taste.

1.2 Eating patterns may differ in children and adults

The maturation of taste and its consequences on eating patterns appear to be influenced early in life by genetics, as well as by environmental and cultural experiences. In adulthood, environmental and/or cultural experiences rather than genetic predisposition appear to be more correlated to taste perception and eating patterns [40-44]. While not consistently reported, several studies on the genetics of taste suggest that learned behaviours can override genetic predisposition. This may explain, for example, why adults consume bitter vegetables which confer health benefits despite the bitter taste that instinctively promotes avoidance [45-55].

4 Table 1.1: Three types of taste bud cells and their known functions.

Type I Type II Type III

Glial-like Cell Receptor Cell Presynaptic Cell

Taste Transduction: Surface glycoproteins and • T1Rs, T2Rs, mGluRs (GPCRs) Neurotransmitter ion channels clearance • Gα-gus and Gγ13 (G protein Function subunits) Excitation and transmitter Ion redistribution and • PLCβ2 (Synthesis of IP3) release transport • TRPM5 (Depolarizing cation current) Neurotransmitter synthesis

Excitation and transmitter release Sweet

Taste Bitter Salt Sour Modality Umami

Fat?

Twin studies provide evidence that eating patterns in adults is contributed in part by environment and genetics. In a longitudinal adult study investigating eating styles in 39 female and 45 male monozygotic Finnish twin pairs discordant for obesity or overweight, normal-weight twins were observed to have different eating styles than their obese siblings. Specifically, the normal-weight twins were less likely to report restrictive eating, overeating, and unhealthful food choices [56]. In contrast, a Swedish adult male twin study demonstrated that the heritability of eating patterns such as cognitive restraint, emotional eating, and uncontrolled eating were 59%, 60%, and 45% respectively [57]. Similarly, another adult male twin study from the United Kingdom and Finland showed that the heritability of cognitive restraint, emotional eating, and uncontrolled eating were 26-63%, 9-45%, and 45-69% respectively [58].

It is important to note that the heritability of obesity in these discordant twin studies does not refer to the heritability of specific genes, such as taste receptor genes [56-58]. None of the heritable factors investigated in the twin studies have been attributed to specific polymorphic variations in taste receptors.

Therefore, future studies in twins discordant for obesity should explore taste receptor genetic variation as a heritable factor which influences eating patterns. Furthermore, no genetic analyses have been done in

5 elderly participants in association with taste perception, taste preferences, nutritional status, or health outcomes. The reason for the lack of genetics studies could be that the loss of taste sensitivity, as a result of reduced TBC regeneration, in the elderly would strongly confound genetic components of taste as causes of change in taste outcomes and health outcomes [59].

In children, studies examining the association between genetics of taste and eating patterns are limited and may not be generalized from adult studies. Compared to studies with adults, research with children show stronger correlation between genetic predisposition for certain eating patterns and their actual consumption patterns [54]. This relationship is especially pronounced in the tasting of bitter compounds by young children. It is thought that aversion to bitter taste by young children is an evolutionary adaptation to prevent the ingestion of toxic plant compounds or other harmful bitter substances which occur in nature [55]. These few observations present a potential new line of investigation to explore ways to develop positive eating habits in children through a better understanding of the genetics of taste on eating behavior [54, 55].

1.3 Fat taste

Oleogustus is the most recently identified taste modality [60]. The taste perception of long-chain fatty acids (LCFA) was previously thought to only be dependent on the texture of the lipid, and the olfactory stimulus to a lesser extent [61-63]. There is now an abundance of evidence supporting taste as a method of oral detection of dietary fat, including saturated fat (Table 1.2) [60, 64, 65]. LCFA provide this stimulus in both rodents [66, 67] and in humans [68]. In the mouse, knockout experiments for LCFA receptor genes such as CD36 indicated a decrease in the spontaneous attraction for fat-enriched foods [69,

70].

The fatty acid translocase CD36 has been the subject of much of the research related to LCFA taste perception. CD36 is expressed in human TBCs [23], and has a strong affinity for LCFA [71]. CD36 knockout mice have been shown to lack the ability to detect LCFA in behavioral tests [69, 70]. The oral

6 perception of linoleic acid (an n-6 polyunsaturated fatty acid), α-linolenic acid (an n-3 polyunsaturated fatty acid), and oleic acid (an n-9 monounsaturated fatty acid) have been attributed to the CD36 TBC receptor [24, 72]. Additionally, a GPCR has been found to bind LCFA and influences the perception of lipids in the oral cavity. GPCR120-null mice show decreased preference for LCFA-enriched solutions in eating behavioral tests compared to wild type [73]. Thus, LCFA as gustatory stimuli are more complex than initially understood.

The consumption of energy-dense foods may be linked to enhanced palatability of dietary fats which predispose individuals to metabolic complications [74, 75]. The relationship between sensory fat perception and predisposition to weight gain has been shown in recent adult human studies [74-76]. This relationship outlines the remarkable difference in obese and lean subjects with respect to fat intake and taste sensitivity. In an obese rat model, an inverse correlation between gustatory sensitivity to LCFAs and dietary preference for fat was observed [77]. Correspondingly, a higher sensitivity to tasting LCFA has been correlated to reduced fat intake, total caloric intake, and body mass index (BMI) [76]. Similarly, a high-fat diet decreased the taste sensitivity to oleic acid in lean, but not obese patients [78]. These observations suggest that a lack of sensitivity to fat taste increases fat intake, and may partly explain why obese adult individuals seem to consume fatty foods more frequently than lean patients [74, 75].

Therefore, the implications of these findings on eating patterns and health have potential significance.

A study by Ma et al. investigated associations between common SNPs in the CD36 gene, lipid and glucose metabolism, and risk for cardiovascular disease [79]. This group genotyped 21 polymorphisms which evenly spanned the CD36 gene, revealing 2 linkage disequilibrium (LD) blocks represented by a haplotype. The haplotype comprised the following five SNPs: 1) -33137A/G

(rs1984112; MAF=0.3468), 2) -31118G/A (rs1761667; MAF=0.3904, 3) 25444G/A (rs1527483;

MAF=0.1018), 4) 27645 deletion /insertion (del /in) (rs3840546; MAF not reported) and 5) 30294G/C

(rs1049673; MAF=0.3832). In a population of adults of European descent, the 30294C polymorphism

7 Table 1.2: Taste as a method of oral fat detection and the role of CD36

Study subjects Outcome Reference

Rat recognizes oleate by a gustatory cue and fatty Anosmic rats [64] acids are important for gustatory recognition of fat.

Rats select LCFA from olfactory or gustatory cues Wistar rats that are related to both the carbon chain and [66] carboxylate group.

Mice find vegetable oils to be as palatable as sucrose Mice [67] solutions.

Linoleic, oleic, stearic, and oxidized linoleic acids are detectable in the oral cavity of humans with minimal Humans [68] input from the olfactory, capsaicin, and viscosity- assessing tactile systems. Comparison of CD36 with that of human muscle fatty acid binding protein suggested that a potential Rats binding site for the fatty acid on CD36 may exist in [71] its extracellular segment between residues 127 and 279. CD36 is the putative orosensory receptor for dietary Humans and LCFA in humans and may be involved in the [23] pigs preference for fatty foods. CD36 is involved in oral LCFA detection and raises Mice and rats the possibility that an alteration in lingual fat [69] perception may be linked to feeding dysregulation. CD36 deletion decreased fat consumption and Mice enhanced the ability of the mice to compensate for the [70] calories provided by their optional fat intake. Mice (GPR40 GPR40 and GPR120 play a role mediating fatty acid and GPR120 [73] taste. KOs)

was significantly associated with higher plasma free fatty acids (FFA), with a stronger association among men compared to women. A similar increase in plasma FFA levels was seen in men carrying the -33137A and -31118G polymorphisms. In the order of SNPs listed above, the haplotype of an individual can be represented by sequentially naming the nucleotide bases. Men with the AGGIG haplotype had 31% higher FFA and 20% higher plasma triglycerides (TG) than non-carriers. The AGGIG haplotype was also

8 correlated to an increased risk of coronary artery disease in T2D American participants (n=197) and

Italian participants (n=321). In the scope of this study by Ma et al., CD36 modulated lipid metabolism and CVD risk in adults of European descent. However, it is unclear whether this modulation occurred at the level of taste perception or metabolism. Further research is required to discern the effects of taste perception and lipid transport on the levels of FFA and TG in the plasma, or if this modulation would also be seen in other ethnicities.

CD36 plays an important role in lipid metabolism and polymorphisms within the CD36 gene have been linked to CVD risk factors [80]. The relationship between genotypes and haplotypes of five SNPs in the CD36 gene (-33137G, -31118A, -22674C (rs2151916; MAF=0.3339), 27645 del/ins and 30294C) with lipid levels have been examined in young normal-weight subjects [81]. High-density lipoprotein cholesterol (HDL-C) levels were lower in -22674C carriers than -22674T carriers, and TT homozygotes had lower oxidized low-density lipoprotein cholesterol (LDL-C) levels compared to -22674C allele carriers. LDL-C levels were higher in carriers of the CC genotype for the 30294C polymorphism compared to non-carriers. Subjects carrying the AATDC haplotype had 3.2 times higher risk of LDL-C >

100 mg/dL than those carrying the AGTIG haplotype, whereas subjects carrying the AATIC haplotype had 2.0 times higher risk of total cholesterol > 200 mg/dL than the AGTIC haplotype. Therefore, genetic variation in CD36 may modulate CVD risk by affecting lipid metabolism.

Single nucleotide polymorphisms in the CD36 gene have also been associated with metabolic disorders related to excess fat depots in adult populations [81-85], and a (TG)-repeat in intron 3 has been linked to elevated BMI in Korean patients with coronary heart disease [86]. Bokor et al. assessed the link between CD36 SNPs and the risk of obesity in a case–control study comprising 307 obese (age = 15.0 ±

1.1 years) and 339 normal-weight (age = 14.6 ± 1.1 years) adolescents [87]. Four SNPs (rs3211867, rs3211883, rs3211908, and rs1527483) were associated with increased obesity risk, higher BMI, and higher percent body fat. A haplotype consisting of the minor alleles of these SNPs was also linked to obesity and excess adiposity (i.e., higher BMI and percent body fat).

9 Studies examining correlative relationships between genetic variations in CD36 and chronic disease have focused on predispositions to specific eating patterns (Table 1.3). Three CD36 gene polymorphisms (-31118A and 25444A and 27645ins), which correspond to part of the haplotype studied by Ma et al., were investigated for their associations with changes in fat taste perception [72].

Homozygotes for the A allele at locus -31118 reported greater perceived creaminess, independently of fat concentration of salad dressings and higher mean acceptance of added fats and oils compared to other

CD36 genotypes. Participants with the CT or TT genotypes at 25444 also perceived greater fat content in the salad dressings, regardless of actual fat concentration. 27645 deletion homozygotes had higher waist circumference and BMI than ins/del or ins/ins individuals (P < 0.001, n=2), however this particular observation requires further replication due to small sample size. Furthermore, Pepino et al. reported findings on the -31118A SNP and fat preference [88]. Twenty-one obese participants with the locus variations AA (n=6), AG (n=7), and GG (n=8) were studied with respect to oleic acid and triolein orosensory detection thresholds. The major and minor alleles of this SNP vary depending on the population. In populations with African ancestry, the G allele is minor, whereas the opposite is true in individuals of European descent. G allele homozygotes had 8-fold lower oral detection thresholds (i.e. higher sensitivity) for oleic acid and triolein than the A allele which was associated with a decrease in

CD36 expression. Intermediate thresholds were observed in heterozygotes. Higher fat taste sensitivity is associated with a decrease in fat preference [76]. However, these findings conflict with other studies on the effect of the -31118 SNP on CD36 expression and function. For instance, Ma et al. hypothesized that the G homozygosity for the -31118 SNP is associated with decreased CD36 function, along with the other associated SNPs in LD [79]. This was hypothesized due to higher plasma FFA and TG in the presence of normal glucose and insulin levels in G allele homozygotes. Higher plasma FFA are indicative of defective

CD36 due to the role of this receptor to translocate FFA into the muscle and other tissues from the plasma

[89]. The increased FFA in the plasma is redirected to the liver which takes up FFAs independently of

CD36 [90]. The liver can then synthesize TGs and release them into the plasma, which may also explain why TGs are high in G allele homozygotes. If the G allele of -31118 is associated with decreased function

10 in the skeletal muscle, and the SNP is located in the promoter of the gene, one would hypothesize that the decreased function might be the result of a decreased expression of the receptor in skeletal muscle.

Decreased expression of CD36 on the tongue should lead to decreased taste sensitivity to fat, hence an increase in fat intake and higher plasma FFAs and TGs. However, the oral fat perception studies which have examined the -31118 SNP have associated the A allele, not the G allele, with decreased fat taste sensitivity and greater preference for fat. This apparent discrepancy could be due to differences in CD36 promoter activity between skeletal muscle and tongue. This difference can be investigated in the future by assessing the expression of CD36 in these tissues in a group of individuals genotyped for the -31118 SNP as well as other associated SNPs in LD. This information would be useful in consolidating current studies on CD36 genetic variation and its effects on fat taste perception and metabolic profile. It would also serve to inform individuals on how they can optimize their diets to improve their metabolic profile based on their CD36 genetic profile [91].

Lean mice and humans can be more sensitive to LCFA-rich foods such as butter, oil, and fatty meat while obese individuals have decreased taste sensitivity to fatty acids due to particular CD36 variants. Therefore, CD36 genetic variation may have the potential to influence fat intake and weight gain by decreasing the oral taste sensitivity of LCFAs in individuals. CVD and dyslipidemia may result from this eating pattern associated with fat overconsumption, especially when the dietary fat predominantly consists of saturated fats [92]. Although studies to date are limited, they provide promising data on the link between genetics, fat perception, and fat preference. It is important to characterize genetic variations in putative fat taste receptors in order to more clearly understand the role of fat taste perception in the development of chronic metabolic disease.

11 Table 1.3: Associations between CD36 gene SNPs fat taste, dietary intake, and health outcomes

SNP Polymorphis SNP ID Outcome Reference Type m Common CD36 SNPs are -33137A > G rs1984112 related to oral fat sensitivity in an obese -31118G > A rs1761667 population, MetS in a 25444G > A rs1527483 Puerto Rican population, lipid levels in young 22674 C > T rs2151916 [72, 79-81, 88] normal-weight subjects, 30294G > C rs1049673 fat ingestive behaviours in African Americans, and 71670C > T rs3211931 lipid metabolism in 27645del > ins rs3840546 individuals of European descent. 73946T > G rs3211938 32307A > G rs10499859 35721C > T rs13438282

Intronic 61882G > A rs1358337 58439T > C rs1054516 48952A > C rs1049654 CD36 variants may impact MetS 67612T > C rs3211909 pathophysiology and 56820A > G rs3211849 HDL-C metabolism in African Americans and [82, 83, 85] 68101A > G rs3211913 French individuals of 80248C > G rs13246513 European descent, both predictors of the risk of 59347T > C rs3173798 heart disease and T2D. 60706C > T rs3211870 56133G > A rs3211842 36498C > T rs9784998 60519T > C rs3211868 -178 A > C -

1.4 Sweet taste

The only receptors known to be involved in the perception of sweet taste are T1R2 and T1R3, which form a heterodimer [93]. These transmembrane proteins allow humans to taste a variety of sweet substances such as naturally occurring sugars (glucose, sucrose, fructose and sugar alcohols), D-amino

12 acids (D-tryptophan and D-phenylalanine) and glycosides (stevioside and glycyrrhizin), as well as alternative sweeteners such as sucralose, aspartame, neotame, saccharin sodium, acesulfame potassium, and cyclamate [20]. Moreover, naturally occurring sweet proteins, such as brazzein, thaumatin, and monellin, and naturally occurring taste-modifying proteins, such as neoculin and miraculin, also bind to

T1R2 and T1R3 [94-99]. Although both receptors are needed to elicit sweet taste, T1R2 is the subunit specific to sweet taste perception because T1R3 is also involved in the detection of the umami taste when it dimerizes with T1R1 [100].

Genetic variations in T1R2 and T1R3 are important to characterize because they pertain to changes in taste sensitivity to sugar [58, 101-109], summarized in Table 1.4. In African, Asian, European and Native American populations, all three T1R genes have a multitude of polymorphisms. These SNPs include non-synonymous SNPs which are located in the N-terminal extracellular domain, the part of the receptor where ligands for taste are thought to bind. Being particularly polymorphic in comparison with other human genes, T1R2 is within the top 5–10% of all human genes with regards to the reported number of polymorphisms. This increased polymorphic rate was hypothesized to be associated with variations in sweet taste perception [110]. One study sought to determine whether Ile191Val variations in

T1R2 were linked to differences in the consumption of sugars in 1037 diabetes-free young adults, and 100 individuals with T2D. [111]. They demonstrated an association between the Ile191Val SNP (rs35874116;

MAF=0.2670) in T1R2 and BMI in individuals with a BMI ≥25. In the diabetes-free individuals, Val carriers (homozygous or heterozygous) consumed significantly fewer sugars compared to homozygous Ile carriers. This finding was consistent in T2D individuals where Val carriers also consumed significantly less sugar compared to Ile homozygotes. The Ile191Val SNP was also confirmed to be associated with intake of sweets as well as higher plasma triglycerides in another cohort of adults [112]. Genetic variation in the other half of the heterodimer, T1R3, was explored in a separate study by Fushan et al. They showed that intronic SNPs in the T1R3 promoter were linked to sucrose taste sensitivity in humans by altering

T1R3 transcription levels [113]. These SNPs accounted for 16% of the population’s variability in sweet

13 perception [113]. The associations between genetic variations in sweet taste receptors and consumption of sweet foods are important findings for individuals who are at risk for metabolic complications due to their eating pattern.

In children, the risk for developing dental caries is associated with the consumption of sweet foods [114]. One study has investigated an association between the prevalence of dental caries and the

Ile191Val polymorphism (T1R2) in 80 young adults of European descent [53]. Val carriers of the T1R2

SNP, i.e., individuals who have been shown to have lower consumption of sweet foods, had lower risk of developing dental caries. A similar finding was observed in a study conducted in schoolchildren. A significant association between total caries was observed with the Ile191Val (T1R2) and the rs307355

SNPs (T1R3). Moderate number of caries (4-7 caries) was observed in rs307355 (MAF=0.2364) heterozygotes, while high-risk caries (>8 caries) was observed in Val homozygotes of the T1R2 SNP

[115]. Therefore, individuals who are genetically predisposed to prefer sweet foods should also consider that their eating pattern influences not only metabolic risk factors, but also dental health. The profundity of those implications must be verified with more studies which consider this T1R2 SNP and other polymorphisms in sweet-sensing genes. Nevertheless, these findings suggest a highly relevant outcome in children as well as young adults.

1.5 Bitter taste

Bitter and sweet taste profiles interact together to affect eating behavior, and this phenomenon is seen especially in children [7]. High concentrations of bitter substances generally result in food rejection, which corresponds to an evolutionary adaptation to avoid toxic substances such as rancid fat, hydrolyzed protein and plant alkaloids [116, 117]. Higher sensitivity to bitter taste may cause individuals to avoid consuming vegetables rich in anti-tumour and anti-oxidant compounds and may, consequently, lead to a higher consumption of sweet and fatty foods as substitutes. This eating pattern has the potential to increase the risk of CVD, obesity, and cancer [118].

14 Table 1.4: Associations between TAS1R2/TAS1R3 gene SNPs and sweet taste, dietary intake, and health outcomes

SNP Type Polymorphism SNP ID Outcome Reference

Genetic variation in T1R2 is associated with consumption of sugars in overweight and obese individuals in 2 distinct Non- 18854899T > C rs35874116 populations. T1R2 is also [53, 111, 115] synonymous (Ile191Val) associated with dental caries in 21-32 year-old individuals of European descent and 184 schoolchildren aged 7-12 years. Two SNPs are linked to human sucrose taste sensitivity. The T allele of each SNP results in reduced promoter activity in comparison to the C alleles. A -1572C > T rs307355 Intronic distal region of the T1R3 [113, 115] rs35744813 -1266C > T promoter has a strong silencing effect on promoter activity. rs307355 was linked to dental caries in 184 schoolchildren aged 7-12 years.

Although there are 25 different types of T2Rs for bitter taste perception on Type II TBCs [119], only T2R38 has been associated with a genetically predetermined bitter taste status [120]. Kim et al. identified the bitter receptor T2R38 as being responsible for tasting phenylthiocarbamide (PTC) and 6-n- propylthiouracil (PROP) [120]. Three common SNPs were found in this gene which results in 3 amino acid substitutions at residues P49A (rs713598; MAF=0.4952), A262V (rs1726866; MAF=0.4255) and

V296I (rs10246939; MAF=0.4794). The bitter-tasting phenotype is attributed to a combination of these three SNPs resulting in a PAV amino acid haplotype, while non-tasters carry the AVI amino acid haplotype [121]. Homozygosity for the PAV haplotype is considered to be a marker of “supertaster” status while those who are homozygous for the AVI haplotype are considered “non-tasters”, and heterozygotes have an intermediate phenotype. This phenotype is also observable by simply assessing the intensity (or lack thereof) of the bitter taste impregnated on a strip of filter paper covered in PTC.

15 A supertaster is a person who experiences the sense of bitter taste with far greater intensity than average [122]. This amplified oral perception appears to have a biological basis rather than being the result of response bias or a scaling artefact [1]. The underlying cause of this relatively heightened oral response is not known, although it is thought to be due to both an increased number of fungiform papillae and the T2R38 phenotype [121, 123]; however, those factors do not completely explain the supertasting phenomenon as it may be attributed to other orosensory phenotypes [124]. Among adults, the T2R38 genotype has been linked to avoidance of alcoholic beverages [123], increased prevalence of colon cancer through inadequate vegetable consumption [116], avoidance of cigarette smoking [125], and a preference for sweetness (discussed in detail below).

A variety of studies, some of which are summarized in Table 1.5, have linked PTC taste status with dietary intake. Two studies involving young, healthy women observed an inverse relationship between bitter sensitivity and an acceptance of brassica vegetables, spinach, coffee, and tart citrus flavours [41, 126]. Furthermore, supertaster women showed a decreased acceptance of sweet and fatty foods [127]. T2R38 taster status has also been associated with a preference for sweet-tasting foods in children, but not adults [7, 128, 129]; however, this result was not replicated in another study examining children [130]. Bitter tasters ate fewer soy products and drank less green tea, and rated these foods to be more bitter than non-tasters [131]. These eating patterns, which are influenced by the T2R38 phenotype, can lead to adverse health effects [118]. However, a study by Choi et al. did not observe any differences in food preference with varying bitter taste status based on a questionnaire [132].

Turner McGrievy et al. investigated the relationship of supertasting (bitter) and sweet preference with MetS and dietary intake [1]. In this study, 38% of the participants met criteria for MetS. After classifying adult study participants with respect to sweet liking (SL) or supertasting (ST), the study found a correlation between the combination of SL and ST statuses (SL+ST) and higher incidence of MetS.

Participants who were either SL or ST had a significantly reduced risk of MetS compared to those who were SL and ST. In addition, those who were SL and ST consumed less fiber than SL + non-ST

16 individuals. This study highlights an interaction between SL and ST statuses that may be linked with

MetS. This is not the only study to indicate a possible relationship between taster status and weight.

Indeed, Tepper et al. showed that those who are STs have a lower BMI than nontasters [133] while another study observed that mean stature of adult PTC non-tasters was higher than that of tasters, and tasters had higher body fat % than non-tasters [54]. Taken together, studies on bitter tasting phenotype suggest a complex relationship with other taste modalities. Despite the evidence presented thus far, the association between taster profile and health outcomes, weight, and dietary intake has been mixed, with studies showing variable results [134-136]. Given that certain adverse eating patterns may result from having a supertaster phenotype, further characterizing the effect of bitter taste receptor variation on food preferences has implications for metabolic and health outcomes.

1.6 Umami taste

The name “umami” originates from the Japanese language which means “delicious savory taste”

[137]. The GPCR heterodimer complex of T1R1 and T1R3 elicits the umami taste when interacting with amino acids, typically mono-sodium glutamate (MSG), and this interaction occurs synergistically with

5’ribonucleotides guanosine monophosphate (GMP), inosine monophosphate (IMP), and adenosine monophosphate (AMP) [100]. The TBC receptors metabotropic glutamate receptor 1 (mGluR1) and mGluR4 have also been implicated in this taste modality [138, 139]. Perception of umami serves as an indicator of purine-rich foods, in particular, the purine-derived metabolite uric acid. Elevated serum urate

(a salt derived from uric acid) has previously been considered benign except for carrying an increased risk for gout [140]. However, serum urate has recently been linked to a multitude of biologic effects including high blood pressure, increased hepatic lipogenesis, and insulin resistance [141, 142]. These adverse health effects suggest that the consumption of certain umami tasting foods, similar to sweet and fatty foods, may increase metabolic disease risk.

Table 1.5: Associations between TAS2R38 gene SNPs and bitter/sweet taste, dietary intake, and health outcomes

SNP Type Polymorphism SNP ID Outcome Reference

A direct molecular link between heritable variability in bitter taste [121] perception to non-synonymous T2R38 variations. Greater propylthiouracil (PROP) sensitivity was associated with lower acceptance of coffee, cruciferous [41, 126, 131] vegetables, tart citrus fruit, dark breads, soy products, green tea, and selected fats. In women, preference of sweet and high-fat food and beverage groups decreased with increasing perceived bitterness of PROP. In men, [127] 145G > C rs713598 preference of these foods and (A49P) beverages increased with increasing papillae densities.

T2R38 variations accounted for a 785T > C major portion of individual Non- rs1726866 (V262A) differences in PROP bitterness synonymous perception in both children and [7] adults, as well as a portion of 886T > C individual differences in preferences rs10246939 (I296V) for sweet flavors in children but not in adults.

TAS2R38 status was not an important determinant of CHD, related risk factors, or eating patterns [128] in the British Women's Heart and Health Study sample The P49A SNP of T2R38 could not serve as a predictor of anthropometric measurements and [129] aversion to vegetables or sweet/fat foods in Malaysian subjects. Neither PROP taster status nor T2R38 genotype alone had [130] significant impact on bitter vegetable preference or intake.

Table 1.5 (Continued)

Associations between TAS2R38 gene SNPs and bitter/sweet taste, dietary intake, and health outcomes

SNP Type Polymorphism SNP ID Outcome Reference

Perceived bitterness of PTC/PROP thresholds were significantly and negatively [54] correlated with body height and fat-free mass. No significant differences in the PROP status distribution between African Americans and Asian Americans and in food preferences between tasters and nontasters. Significant differences [132] in fat foods, sugar, and black coffee preference were observed

among African Caribbean, 145G > C rs713598 African black, East Asian, and (A49P) South Asian groups. There was a significant Non- interaction between ST and SL 785T > C rs1726866 status as associated with MetS. synonymous (V262A) This interaction was also significantly associated with fiber and caloric beverage intake. 886T > C Participants who were only an ST (I296V) rs10246939 [1] or SL appeared to have a decreased risk of having MetS compared with those who have a

combination or are neither taster groups (P = 0.047) and that SL + ST consumed less fiber than SL + non-ST (P=0.04).

Medium tasters and supertasters could discriminate differences in fat content between 40% fat and [136] 10% fat salad dressings (p < 0.005), but the non-tasters could not.

19 There is evidence that differences in the perception of umami are common among individuals with taste threshold concentrations of umami stimuli varying up to 5 fold [143]. Similar to bitter taste perception, a fraction of the population may carry a “nontaster” phenotype for umami [143], but more research is needed to characterize this phenotype. Despite the lack of research linking differences in umami taste sensitivity to health outcomes, some studies have correlated certain SNPs in umami taste receptors to differences in taste perception (Table 1.6). One such study was conducted where 17 T1R1 and T1R3 SNPs were examined in Japanese adults [144]. The purpose of the study was to demonstrate a link between SNPs and detection thresholds for IMP and MSG. Taste thresholds for three common SNPs:

Gln12His (rs75881102; MAF=0.0114) and Ala372Thr (rs34160967; MAF=0.1354) in T1R1 and

Arg757Cys (rs307377; MAF=0.0481) in T1R3 were investigated. The T1R1 SNP Ala372Thr

(rs34160967) appeared to be non-synonymous as the Thr allele was associated with increased sensitivity to umami. This association was significant for MSG as well as a mixture of MSG and IMP, but not for

IMP alone. The T1R3 SNP Arg757Cys also appeared to be non-synonymous for all umami tasters, as the

Cys757 allele yielded a reduced sensitivity for umami taste. Therefore, it is possible that the T1R1 and

T1R3 heterodimer have separate binding sites for IMP and MSG, and the T1R3 757 and T1R1 372 SNPs are located in the active sites for IMP and MSG binding, respectively. Shigemura et al. have also shown that that these two SNPs are not in LD, but that their functions may overlap to a certain degree. Thus, having the T1R1 Thr372 SNP and the T1R3 Arg757 SNP causes an overall decrease in umami taste threshold, and having the T1R1 Ala372 SNP and the T1R3 Cys757 SNP causes an overall increase. This was confirmed in another study where the Cys757 variant reduced stimulation of the heterodimer by

MSG [145]. More research is needed to better understand the influence of genetics on the preference or intake of MSG-rich foods. Further characterization of the genetic variation implicated in this taste modality would allow individuals to determine their predisposition to preferring umami foods. This knowledge may prove to be important given that moderating the consumption of umami foods may reduce the risk for metabolic complications.

20 Table 1.6: Associations between TAS1R1 and TAS1R3 gene SNPs and umami taste perception

Polymorphis SNP Type SNP ID Outcome Reference m

The 372Thr allele was linked to

increased sensitivity to umami. A372T rs34160967 This association was significant [144] for MSG and a mixture of MSG

Non- and IMP, but not for IMP alone. synonymous The Cys757 allele yielded a R757C rs307377 reduced function and is necessary [144] for IMP interaction.

1.7 Sour taste

Taste receptor function for sour tasting, which involves GPCR-mediated signaling, is not well- characterized. Studies on the variation at the genetic loci for taste receptor genes involved in sour taste have yet to be undertaken—however, some important advances have been made with regards to this taste modality in a thorough and recent review by Garcia-Bailo et al. [146]. Major findings were related to the biology of sour taste perception and the palatability of sour foods [20, 107, 147-149]. In general, sour taste perception is thought to be elicited upon the stimulation of taste buds by acidic substances, leading to depolarization of TBCs. Many animals find mildly acidic foods to be palatable, but most mammals reject strong sour stimuli as a means to avoid the consumption of spoiled foods. Ye et al. (2016) identified

Kir2.1 as an acid-sensitive K+ channel in type III TBCs using pharmacological and RNA expression profiling as well as a tissue-specific knockout of the KCNJ2 gene [150]. The identification of Kir2.1 in sour taste cells suggests a mechanism for amplification of sour taste and provides a rational for stronger intracellular acidification by weak acids (such as acetic acid) than strong acids [150]. The KCNJ2 gene carries coding SNPs which may influence sour taste perception. Exploring the effect of these SNPs on sour taste perception and subsequent food choices are important future directions for this taste modality.

21 1.8 Salt taste

Salt taste perception influences sodium intake, the excess for which corresponds to a major public health concern due to the risk of developing hypertension. Excessive sodium consumption may be attributed to greater individual preferences for salty foods, which may be linked to genetic factors such as salt taste receptor function. However, the source of inter-individual differences in salt preferences is controversial as environmental influences, such as cultural determinants of dietary intake, may be more important than genetic predisposition to salt preference [151, 152]. Thus, salt preference may be due to sensory habituation to high sodium foods and a subsequent increase in the preference of these foods.

Similar to sour taste, salt taste receptors have yet to be characterized in sufficient detail to draw definitive conclusions about differences in salt taste receptor function and possible implications for differences in eating patterns and health outcomes. Due to the fundamental importance of electrolyte consumption for mammalian physiology, the desire to consume sodium-rich foods could be attributed to physiological factors rather than to the search for the sensory pleasure of salt taste [153]. The physiological basis for seeking electrolytes is supported by periodic fluctuations which occur in women more than men, likely due to hormonal cycling [154].

ENaC and TRPV1 are the two putative salt taste receptors [155]. Studies have shown that nerve fibers responding to sodium can be classified according to their sensitivities to the ENaC blocker, amiloride, which distinguishes amiloride-sensitive (AS) and amiloride-insensitive (AI) groups [12]. The

AS fibers respond specifically to sodium chloride (NaCl) whereas AI fibers respond broadly to various electrolytes including NaCl. It has been proposed that while the AS mechanism of salt detection is dependent on ENaC, the AI mechanism is mediated by the function of TRPV1. Thus, it has been suggested that the appealing taste of sodium, representative of a lower taste concentration threshold, is mediated by the AS fibers and hence the ENaC protein while aversive reactions to tasting high concentrations of sodium, and other types of salts, are mediated by the AI fibers using TRPV1 [12]. These

22 data suggest that at least two different systems exist to facilitate salt taste transduction in taste cells via

AS and AI fibers.

In a study by Dias et al., SNPs in ENaC and TRPV1 receptors were associated with differences in salt taste perception among adults [155]. In the ENaC gene, two intronic SNPs modified salt taste sensitivity. Homozygotes for the A allele of rs239345 (A/T) (MAF=0.2642) and the T allele of rs3785368 (C/T) (MAF=0.1899) perceived salt solutions less intensely than carriers of the other respective alleles, (P = 0.02; P = 0.03, respectively). In the TRPV1 gene, the Val585Ile polymorphism rs8065080 (C/T) (MAF=0.3177) modified taste sensitivity where carriers of the T allele were significantly more sensitive to salt solutions than the CC genotype (P = 0.008). In a separate study conducted in young children aged 1.5-5 years, carriers of the T allele for rs239345 were found to consume significantly more sodium (1.69 ± 0.09 g) compared to the AA genotype (1.41 ± 0.07 g).

Further, carriers of the T allele were also observed to have higher systolic (106.6 ± 3.5 mmHg) and diastolic BP (66.0 ± 4.7 mmHg) compared to children with the AA genotype (systolic: 96.3 ± 2.3 mmHg, diastolic: 54.5 ± 2.3 mmHg). No differences were found in dietary sodium intake or BP with the rs8065080 SNP. These findings suggest that genetic variation in the ENaC and TRPV1 genes may contribute to inter-individual differences in salt taste perception.

Altogether, environmental influences and genetic factors play a role in defining a salt-preferring phenotype, but environmental influences may have a more dominant role. Nevertheless, new genetic findings associating TRPV1 and ENaC to salt preference suggests that there is more to learn about genetic influences [155] (Table 1.7). Due to the novelty of this type of research, the extent to which such variation influences preference or intake of salty foods is unknown. Future studies are needed to investigate the role of salt taste receptor SNPs in sodium consumption and across the lifespan.

23 Table 1.7: Associations between ENaC and TRPV1 gene SNPs and salt taste perception

Polymorphis SNP Type SNP ID Outcome Reference m

The A allele carriers A > T rs239345 perceived salt solutions less intensely [155] than carriers of the T allele. (P=0.02) Intronic

T allele perceived salt solutions less C > T rs3785368 intensely than carriers of the C allele. [155] (P=0.03)

Carriers of the T allele were Non- V585I significantly more sensitive synonymou rs8065080 [155] to salt solutions than the CC s (C > T) genotype. (P=0.008)

1.9 Summary

The relatively new and active field of taste research shows great promise to enhance our fundamental understanding of how taste receptors SNPs contribute to eating patterns, health and disease.

For illustrative purposes, a single candidate SNP is used to demonstrate the effect of genetic variation on taste perception, dietary intake and health outcomes (Figures 1.1-1.6). In particular, genetic variation in the CD36 LCFA receptor (Figure 1.1), the T1R2 sweet taste receptor (Figure 1.2), and the T2R38 bitter taste receptor (Figure 1.3) has important health implications because they may predispose individuals to the over-consumption of unhealthy foods and the under-consumption of healthy foods. Investigation into the roles of SNPs in taste receptor genes for umami, sour, and salt tastes (Figures 1.4-1.6) is also warranted due to their potential to influence taste preference, dietary intake, and health. An important consideration in future research requires attention to taste-driven food preferences in children and adults, which may be genetic in origin in children but uncoupled in adults due to cultural influences. Additional

24 studies on taste preferences in children are needed, which may help establish healthy eating habits unique to every child [156, 157]. As food palatability often involves a combination of different types of taste, such as salt and fat, studies examining combinations of SNPs across different taste modalities are warranted. Sex-specific differences in various ethnicities are also important factors to be considered in future studies. Overall, new research in taste may aid in tailoring food choices and reducing the risk for obesity, T2D, CVD, and other chronic diseases.

25 CD36 SNP rs1761667 AA Genotype

Biomarkers of Health

Taste Fatty Food Fat Intake Sensitivity Selection

Figure 1.1: Schematic representation of the effect of the rs1761667 SNP in the CD36 fat taste receptor gene on fat taste perception, fatty food selection, fat intake, and biomarkers of health such as blood TG levels, cholesterol levels, and anthropometrics The effect of the AA genotype on CD36 function in the tongue as a fat taste receptor and in the muscle as a fatty acid transporter is controversial. While taste perception studies have been able to consistently show that A allele homozygotes have lower taste sensitivities to fatty acids due to the decreased expression of the receptor, others have suggested that the G allele is associated with decreased function in the muscle as a fatty acid transporter. Therefore, there is a discrepancy between the expected decrease in plasma TGs due to taste effects and the observed increase in plasma TGs due to the lack of FFA transport into the muscle.

26 T1R2 SNP rs35874116 Major Allele (T)

Biomarkers of Health

Taste Sweet Food Sweet Food Sensitivity Selection Intake

? - Dental caries

- BMI

Figure 1.2: Schematic representation of the effect of the T allele in the rs35874116 SNP in the TAS1R2 sweet taste receptor gene on sweet taste perception, sweet food selection, sugar intake, and biomarkers of health such as dental caries and BMI. While this SNP has been related to increased preference as well as intake of sweet foods, the effect of the SNP on taste sensitivity has not been elucidated. This SNP has been linked to increased dental caries, and individuals with increase sweet preference and intake were also more likely to have higher BMIs. Importantly, a different SNP in the TAS1R2 sweet taste receptor (rs12033832) was found to be associated with altered sweet taste sensitivity.

27 T2R38 SNP rs713598 Major Allele (C) Biomarkers of Health

PTC Sweet Food Sweet Food Non-taster Selection Intake

Vegetable Vegetable Selection Intake

- Body fat %

- Plasma glucose

Figure 1.3: Schematic representation of the effect of the C allele in the rs713598 SNP in the T2R38 bitter taste receptor gene on bitter taste perception, bitter and sweet food selection, sugar, fruit, and vegetable intake, and biomarkers of health such as body fat % and plasma glucose. The C allele of this SNP is well established to be part of a haplotype which causes the PTC non-taster phenotype. This phenotype, unlike the ‘supertaster’ phenotype, is not prone to decreased vegetable preference and consumption or a compensatory increase in the preference and consumption of sweet food.

28 T1R1 SNP rs34160967 Major Allele (G) Biomarkers of Health

Umami Umami Food Protein Sensitivity Selection Intake

? ? ?

PKD1L3+PKD2L1 Biomarkers of Health

Sour Sour Food Sour Food sensitivity Selection Intake

? ? ? ?

Figure 1.5: Schematic representation of the potential effect of the PKD1L3+PKD2L1 sour taste receptor gene on sour taste perception, sour food selection, sour food intake, and biomarkers of health.

30 ENaC SNP rs239345 AA genotype

Biomarkers of Health

Sodium Sodium Sodium sensitivity Preference Intake

Blood pressure

31 CHAPTER 2:

THESIS RATIONALE AND OBJECTIVES

32 2.1 Rationale and Overall Objectives

Food preferences and dietary habits are heavily influenced by taste perception. There is growing interest in characterizing taste preferences based on genetic variation, which may explain how taste impacts eating behavior and nutritional intake. Therefore, increased understanding of taste and genetics may lead to new personalized strategies which may prevent or influence the trajectory of chronic diseases risk. Recent advances show that single nucleotide polymorphisms (SNPs) in taste receptor genes are associated with differences in taste perception, dietary intake, and biomarkers of metabolic disease. For instance, SNPs in the CD36 fat taste receptor is linked to differences in fat perception, fat preference, and chronic-disease biomarkers. Genetic variation in the sweet taste receptor T1R2 has been shown to alter sweet taste preferences, dietary intake, and risk of dental caries. Polymorphisms in the bitter taste receptor

T2R38 have been shown to influence taste for brassica vegetables. Individuals that intensely taste the bitterness of brassica vegetables (“supertasters”) may avoid vegetable consumption and compensate by increasing their consumption of sweet and fatty foods, which may increase risk for chronic disease.

Emerging evidence also suggests that the role of genetics in taste perception may be more impactful in children due to the lack of cultural influence compared to adults. These recent findings demonstrate the potential of SNPs residing in taste receptors to influence food preferences and consequently health.

Overall, the objective of this thesis is to investigate associations between SNPs in taste receptors genes and taste perception and dietary intake. We hypothesized that some SNPs in taste receptor genes would be associated with taste perception and dietary intake in children and adults. To address the objective of this thesis, three independent studies were conducted, which are outlined as follows.

33 2.2 Specific Study Objectives and Hypotheses

Study 1 (Chapters 3 and 4): Taste sensitivity and taste preference measures are correlated and are each associated with SNPs in taste receptor genes in healthy young adults

The specific objectives of study 1 were to:

1. Investigate the associations between SNPs in fat, salt, sour, sweet, umami, and bitter taste receptor

genes with measures of taste function and taste preference in healthy young adults.

2. Examine if the effect of genetics on taste function is associated with differences in measures of

physical health and blood pressure.

3. Investigate correlations among measures of taste sensitivity, including detection threshold (DT),

suprathreshold sensitivity (ST) and preference (PR) within sweet, salt, sour, umami, and fat tastes.

4. Explore the underlying attributes of DT, ST and PR measurements using principal component

analysis.

Given the objectives of study 1, it was hypothesized that:

1. SNPs in fat, salt, sour, sweet, umami, and bitter taste receptor genes would be associated with

measures of taste function and taste preference within their respective taste modalities in healthy

young adults.

2. SNPs in fat, salt, sour, sweet, umami, and bitter taste receptor genes would be associated with

differences in measures of physical health and blood pressure.

3. DT and ST are negatively correlated for every taste modality.

4. DT is positively associated with PR and ST is negatively associated with PR for every taste

modality.

5. Measured PR will predict self-reported taste and food preferences.

Study 2 (Chapter 5): Single nucleotide polymorphisms in taste receptor genes are associated with snacking patterns in preschool-aged children in the Guelph Family Health Study

The specific objective of study 2 was to:

34 Determine the relationship between SNPs in CD36, TAS1R2, and TAS2R38 taste receptor genes and

snacking patterns measured in children aged 1.5-5 years in the Guelph Family Health Study pilot.

Given the objectives of study 2, it was hypothesized that:

1. Children with the AA genotype of the rs1761667 SNP in the CD36 fat taste receptor gene would

consume snacks with higher percent calories from fat than children carrying the G allele.

2. Children with the TT genotype of the rs35874116 SNP in the TAS1R2 sweet taste receptor gene

would consume snacks with higher percent calories from sugar than children carrying the C allele.

3. Children with the C allele of the rs713598 SNP in the TAS2R38 bitter taste receptor gene would

consume snacks with higher energy density than children with the GG genotype.

Study 3 (Chapter 6): The relationship between single nucleotide polymorphisms in taste receptor genes, taste function and dietary intake in preschool-aged children and adults in the Guelph Family Health Study

The specific objectives of study 3 were to:

1. Determine the relationship between SNPs in taste receptor genes and taste sensitivity and taste

preference in children and adults.

2. Determine if SNPs screened for associations in objective 1 are linked to measures of dietary intake in

children.

Given the objectives of study 3, it was hypothesized that:

1. SNPs in sweet taste receptor genes would be associated with differences in sweet taste sensitivity,

sweet taste preference, self-reported sweet preference, and dietary sugar intake.

2. SNPs in fat taste receptor genes would be associated with differences in fat taste sensitivity, fat

taste preference, self-reported fat preference, and dietary fat intake.

3. SNPs in salt taste receptor genes would be associated with differences in salt taste sensitivity, salt

taste preference, self-reported alt preference, and dietary sodium intake.

35 4. SNPs in umami taste receptor genes would be associated with differences in umami taste

sensitivity, umami taste preference, self-reported preference of high-protein foods, and dietary

protein intake.

5. SNPs in sour taste receptor genes would be associated with differences in sour taste sensitivity,

sour taste preference, self-reported sour preference, and dietary fruit intake.

6. SNPs in bitter taste receptor genes would be associated with phenylthiocarbamide taster status,

self-reported bitter and sweet preference, and an energy-dense diet.

The methodologies and results associated with each of these studies are presented in the forthcoming chapters.

36 CHAPTER 3 RATIONALE

With the knowledge that children establish important eating habits by the age of five, a key component of this thesis was to determine how genetic variation in taste receptor genes may contribute to the development of certain taste preferences and eating patterns in young children. However, examining this relationship first in healthy young adults provided some key advantages. Relative to children, recruitment and consent of adults is simpler and they are more capable of understanding instructions provided to them to perform a sensory task. Furthermore, the subjectivity and biases that adults have towards certain sensory measurements are more predictable due to the fact that the methodologies were developed and refined with adult respondents.

In addition to the methodological benefits of conducting sensory analyses in adults first, there are theoretical advantages of examining measures of taste function first in young, healthy adults rather than children from the GFHS. Taste function has been previously associated with body weight such that individuals with a higher BMI may be less sensitive to certain types of taste [74, 76, 158-160]. Adults who participated in Study 1 of this thesis were of a healthy body fat percentage on average, whereas children in the GFHS have diverse body fat mass measurements. Investigating taste function in a sample of healthy young adults would establish confidence that substantial inter-individual differences in taste occur despite there being small differences in body fat mass. Therefore, conducting a pilot sensory study in adults presented a special opportunity to become adept at administering sensory tests all the while limiting potential covariates. It was decided for the purpose of this thesis work that sensory studies would be piloted in adults first and then adapted for use in children and their families in the Guelph Family

Health Study.

37 CHAPTER 3

THE RELATIONSHIP BETWEEN SINGLE NUCLEOTIDE POLYMORPHISMS IN TASTE

RECEPTOR GENES, GUSTATORY FUNCTION, AND TASTE PREFERENCES IN YOUNG

ADULTS

This chapter is currently under review in the journal Physiology and Behavior.

Chamoun, E., Liu, A. S., Duizer, L.M., Feng, Z., Darlington, G, Duncan, A. M., Haines, J., and Ma,

D.W.L. The relationship between single nucleotide polymorphisms in taste receptor genes, gustatory function, and taste preferences in young adults. Physiol and Behav. Submitted: Feb. 2018.

38 3.1 Abstract

Taste is a fundamental mechanism whereby compounds are detected orally, yet it is highly variable among individuals. The variability in taste that is attributable to genetics is not well-characterized despite its potential role in food selection and therefore eating habits that contribute to risk of overweight and obesity. The primary objective of this study was to elucidate the relationship between single nucleotide polymorphisms (SNPs) in taste receptor genes and psychophysical measures of taste in healthy young adults. Sweet, salt, umami, fat, sour, and bitter taste receptor gene SNPs were genotyped in 49 participants (ages 24.6 ± 0.6 years) who completed testing to determine oral detection threshold, suprathreshold sensitivity and taste preference. Following a Bonferroni adjustment, the rs4790151 SNP in the TRPV1 putative salt taste gene was significantly associated with salt detection threshold and the rs2499729 SNP in the mGluR4 umami taste gene was significantly associated with detection threshold for monosodium glutamate + inosine monophosphate. No significant associations were observed between

SNPs in taste receptor genes and psychophysical measures of sweet, fat, sour, or bitter taste. These results demonstrate that SNPs in taste receptor genes may contribute to inter-individual differences in psychophysical measures of taste and taste preference. Future studies are warranted to investigate if these findings have consequences for habitual intake of salty and umami foods. (Support was provided by the

Ontario Ministry of Agriculture, Food, and Rural Affairs)

Keywords: Taste; sensory; genetics

39 3.2 Introduction

The growing rate of obesity is a serious international public health concern. As obesity is generally characterized by a chronic imbalance of caloric intake and energy expenditure [161], understanding the determinants of excess caloric intake is a critical aspect of developing effective dietary interventions to mitigate the risk of obesity. Previous research has demonstrated that approximately 50% of the risk of developing obesity is genetically inherited [162-164]; however, the genetic factors predisposing to obesity are poorly understood.

Taste is a fundamental mechanism whereby nutrients are detected orally, yet it is highly variable among individuals due in part to genetics [110, 165]. Understanding the genetic basis of taste function and its influence on taste preferences as well as food selection would be instrumental in tailoring dietary interventions to improve the eating patterns that lead to obesity [165]. To this effect, general nutrition advice can be supplemented with insight into which foods an individual may select based on unique taste preferences.

Psychophysical studies have previously been used to elucidate the role of genetics in taste function by examining single nucleotide polymorphisms (SNPs) in taste receptor genes [72, 88, 113,

158]. Moreover, including measures of physical and physiological health, such as body composition and blood pressure, in psychophysical studies provides insight into the impact that differences in taste function may have on health. The primary objective of this study was to investigate the associations between SNPs in fat, salt, sour, sweet, umami, and bitter taste receptor genes with measures of taste function and taste preference in healthy young adults. Secondly, this study examined if the effect of genetics on taste function is associated with differences in measures of physical health and blood pressure.

40 3.3 Methods

3.3.1 Participants

Fifty-three participants were recruited from the University of Guelph campus by e-mail and by word of mouth. Exclusion criteria included smoking, having been diagnosed with a taste dysfunction, and having undergone bariatric surgery. This study was approved by the Research Ethics Board at the

University of Guelph (REB#16-12-629).

3.3.2 Anthropometry, Body Composition and Blood Pressure Measurements

Participants arrived at the University of Guelph in the Body Composition Lab having fasted for at least two hours. Height was measured to the nearest 0.1cm using a wall-mounted stadiometer (Medical

Scales and Measuring Devices; Seca Corp., Ontario, CA). Body weight was measured while wearing tight-fitting clothing and no shoes using the BOD POD™ digital scale (Cosmed Inc., Concord CA, USA).

Body mass index (kg/m2) was calculated from the weight and height measurements. Waist circumference was measured using a Gulick II measuring tape (Country Technology Inc., Gay Mills, WI, USA) at the top of the iliac crest. The BOD POD™ was used to determine body composition of participants using air displacement plethysmography. Blood pressure and heart rate were measured from the right brachial artery using an automated oscillometric device (HBP-1300 OMRON, Mississauga, Ontario). Cuff size was determined based on arm circumference. Two rested measurements of blood pressure (systolic and diastolic) and heart rate were obtained via an automatic reading while participants were seated in an upright position. The average of the two measurements for each participant was used in subsequent analyses.

3.3.3 Psychophysical Measurements

All psychophysical tests were administered in sensory booths at the University of Guelph Sensory

Laboratory. Filter paper strips (Indigo Instruments – Cat#33814-Ctl; 47mm x 6 mm x 0.3 mm) immersed in varying concentrations of tastant solutions were used for all sensory tests. The tastants were: sucrose for sweet taste (Thermo Fisher; S5-500), monosodium glutamate (MSG) for umami taste (Thermo Fisher;

41 ICN10180080), inosine monophosphate (IMP) for umami taste (Thermo Fisher; AC226260250), MSG +

IMP, sodium chloride (NaCl) for salt taste (Thermo Fisher; S641-500), citric acid for sour taste (A940-

500), oleic acid for fat taste (A195-500) (Thermo Fisher Scientific), and PTC for bitter taste (Indigo

Instruments – Cat#33814-PTC). Oleic acid was homogenized in deionized water prior to immersing the filter paper, and all other tastants were dissolved in water at ambient temperature. Filter paper strips were immersed in the tastant solution for about one second before placing them on a drying rack to dry overnight at ambient temperature. This procedure was performed only once for all strips before the study commenced. Taste strips immersed in a solution with the same tastant and concentration were stored together at 4°C in a small plastic re-sealable bag. Each time a strip was tested, participants placed the taste strip in the middle of their tongue, closed their mouths, and allowed at least five seconds for the tastants to be sensed by taste receptors. Participants were asked to rinse and expectorate with distilled water before beginning and following each strip. Within each taste modality, the range of tastant concentrations tested is shown in Table 3.1.

Oral DT was measured using the two-alternative forced-choice staircase method [166, 167], ST was determined for a range of tastant concentrations by computing the area-under-the-curve (AUC) of intensity ratings on the general labeled magnitude scale (gLMS) [159], and PR was measured using the

Monell forced-choice paired comparison tracking method [168]. In the DT test, participants were first presented with a blank taste strip and were told that it is tasteless in order to establish a “negative control”. Participants were instructed that they may ask to receive the blank strip at any point during the test as a reminder of the sensation of the blank filter paper. Blank strips administered as negative controls during the test were isolated from the regular test procedure. Paired stimuli (blank strip and taste strip) were presented in random order one at a time to participants who were then asked to respond with “yes” or “no” to the question “Did you taste something?”. The first concentration tested was the median of the full range of concentrations used in the test. If a participant correctly identified a taste strip from a blank strip on two consecutive occasions, then the adjacent lower concentration was subsequently tested.

However, if a participant incorrectly distinguished a taste strip from a blank strip on one occasion, then

42 the adjacent higher concentration was subsequently tested. A change in the direction to a lower or higher concentration during the test was recorded as a reversal, and the threshold was calculated as the mean of the concentrations where the first four reversals occurred. In the ST test, participants were presented with a range of taste strips in random order and were asked to rate the intensity of the strips from 0-100 on a gLMS where 0=undetectable, 2=barely detectable, 6=weak, 18=moderate, 35=strong, 52=very strong, and

100=strongest imaginable sensation of any kind. The AUC of the intensity ratings for a tastant was computed to determine ST. For bitter taste, only one rating of PTC intensity was obtained. In the PR test, paired stimuli (two adjacent concentrations) were presented one at a time to participants. After the second stimulus was presented, participants were asked “Which of the two did you prefer?” and responded with

“the first one” or “the second one”. The subsequent pairing included the previously preferred concentration as well as the other adjacent concentration. Once the same concentration was chosen after being presented with two adjacent concentrations, it was deemed to be the PR. Oral DT for all taste modalities were measured in one session while ST and PR were measured in a separate session, on a different day.

Table 3.1: Range of tastant concentrations used for each psychophysical test

Taste modality (stimulus) Threshold/Suprathreshold (mM) Preference (mM) Sweet (sucrose) 2.5-500 6%-36% (w/v*) Umami (MSG) 3.13-200 3.13-200 Umami (IMP) 0.313-20 0.313-20 Umami (MSG+IMP) 3.13-200 MSG + 0.5 IMP 3.13-200 MSG + 0.5 IMP Salt (sodium chloride) 5-100 50-250 Sour (citric acid) 1-15 10-200 Fat (oleic acid) 30-100 50-100 Bitter (PTC) 3 μg/strip - Tastants were diluted in distilled water and filter papers were submerged in the solutions. *weight/volume

43 3.3.4 SNP Selection and Genotyping

A PubMed SNP search was conducted for the following genes previously implicated in taste detection: CD36, GPR120, GPR40, TAS1R1, TAS1R2, TAS1R3, TAS2R38, ENaC, TRPV1, mGluR4, and

KCNJ2. The resulting SNPs from each gene were filtered by global minor allele frequency (MAF), and

SNPs with a minor allele frequency below 5% were removed [169]. The resulting SNPs were filtered using HaploView 4.2 software to obtain tag SNPs (tSNPs). Each tSNP is considered independent due to low linkage disequilibrium (r2<0.05).

Saliva was collected at the health assessment using the Oragene•DNA (OG-575) collection kit for

Assisted Collection (DNA Genotek). Participants were fasted for a minimum of 30 minutes before the saliva sample was provided. Genetic material from saliva was extracted by ethanol precipitation according to the manufacturer’s protocol (DNA Genotek). The DNA samples were sent to The Centre for

Applied Genomics at The Hospital for Sick Children where they underwent genotyping using the Agena

MassArray System.

3.3.5 Statistics

Linear regressions were performed using R Statistical Software Version 3.4.0 (R Foundation for

Statistical Computing) to examine the associations between SNP predictors and taste outcomes, with ethnicity, age, sex, and BMI included in the models as covariates. Multivariate regressions were conducted to examine the associations between SNP predictors and multiple taste outcomes within one taste modality. No covariates were included in the multivariate regression as ethnicity, age, sex, and BMI were not significantly related to the outcomes in the single trait linear regressions. Regressions were only performed for SNPs located in a gene associated with the same taste modality as the taste outcome. A

Bonferroni adjustment was applied to p-values to determine if certain associations were independent of multiple testing. Statistical significance was set to p<0.05.

44 3.4 Results

3.4.1 Participant Characteristics

While fifty-three participants were recruited for the study, four participants relocated before completing the study and were therefore excluded. Thirty-five females and fourteen males with a mean age of 25 ± 3 years completed the study. The mean BMI of participants (22.3 ± 2.6 kg/m2) indicated a normal body weight. Measures of anthropometry, body composition, blood pressure, and heart rate were also considered normal and summarized in Table 3.2.

3.4.2 Genetics and Taste Function/Preference

In total, ninety-four tSNPs were genotyped from eleven taste receptor genes. Twenty tSNPs were genotyped from fat taste receptor genes (CD36, GPR120, and GPR40), eleven tSNPs were genotyped from sweet taste receptor genes (TAS1R2 and TAS1R3), one tSNP was genotyped from a bitter taste

Table 3.2: Participant characteristics in total and separated by sex

Characteristic Total Female Male n 49 35 14 Age (years) 25 (3) 25 (3) 24 (2) Systolic Blood Pressure (mmHg) 119 (10) 115 (6.2) 132 (8.2) Diastolic Blood Pressure (mmHg) 72 (6) 71 (5.2) 74 (7.5) Heart rate (beats/min) 68 (10) 68 (8.9) 67 (12) Height (cm) 169.3 (6.0) 166 (4.6) 177 (4.9) Weight (kg) 64.1 (10.3) 58.5 (6.1) 78.5 (8.2) BMI (kg/m2) 22.3 (2.6) 21.2 (2.0) 25.1 (2.5) Waist Circumference (cm) 76.7 (6.7) 73.4 (4.9) 84.9 (5.2) % Body Fat 20.2 (5.8) 22.2 (5.0) 14.8 (5.5) Ethnicity (%) Caucasian 75 - - South-east Asian 10 - - Latin American 10 - - Other 5 - - Means (SD) were computed for all characteristics except for ethnicities, which are presented as percents.

45 receptor gene (TAS2R38), nineteen tSNPs were genotyped from salt taste receptor genes (ENaC and

TRPV1), thirty-two tSNPs were genotyped from umami taste receptor genes (TAS1R1, TAS1R3, and mGluR4), and eleven tSNPs were genotyped from a sour taste receptor genes (KCNJ2).

As summarized in Table 3.3, many tSNPs were associated with a taste outcome before applying a statistical correction for multiple hypotheses. Following a Bonferroni correction for multiple hypotheses, two tSNPs remained significantly associated with a taste outcome. The A allele of the rs4790151 tSNP in the TRPV1 salt taste receptor gene was associated with a significantly higher DT for sodium chloride compared to the G allele. The C allele of the rs2499729 tSNP in the mGluR4 umami taste receptor gene was associated with a significantly higher DT for MSG+IMP compared to the T allele. The G allele of the rs4920564 tSNP in the TAS1R2 sweet taste receptor gene was associated with a higher PR for sucrose compared to the T allele, but not significantly.

A multivariate regression was conducted with DT, ST, and PR as outcomes and tSNP predictors, many tSNPs were associated with at least one of the three taste outcomes before applying a statistical correction for multiple hypotheses (Table 3.4). Following a Bonferroni correction, no tSNPs remained significantly associated with any taste outcomes.

3.4.3 Genetics, Taste, and Overall Health

No significant associations were observed between SNPs in taste receptor genes and measures of physical health including anthropometry, blood pressure, heart rate, and body composition.

3.5 Discussion

This study assessed the relationship between a comprehensive panel of SNPs in taste receptor genes and psychophysical measures across all known taste modalities. Overall, the findings in this study showed that only SNPs in salt and umami taste receptor genes are associated with inter-individual differences in psychophysical measures of taste. The role of genetics in taste perception is important to elucidate as it may be one of the determinants of excess energy intake, and therefore obesity.

46 Table 3.3: Associations between SNPs in taste receptor genes and detection threshold, suprathreshold sensitivity, and taste preference

SNP ID (gene) Taste Modality Outcome P-value

rs4920564 (TAS1R2) Sweet Preference 0.0058 rs10947476 (mGluR4) Threshold (MSG) 0.0369 rs10947476 (mGluR4) Threshold (MSG+IMP) 0.0125 rs11122100 (TAS1R1) Threshold (MSG) 0.0345 rs12080675 (TAS1R1) Threshold (MSG) 0.0255 rs1565356 (mGluR4) Threshold (MSG) 0.0432 rs2499729 (mGluR4) Threshold (MSG) 0.0185 rs2499729 (mGluR4) Threshold (MSG+IMP) 0.0003* rs35744813 (TAS1R3) Umami Threshold (MSG+IMP) 0.0124 rs2499711 (mGluR4) Preference (MSG) 0.0491 rs12080675 (TAS1R1) Preference (MSG) 0.0456 rs2499682 (mGluR4) Preference (IMP) 0.0398 rs11759763 (mGluR4) Preference (IMP) 0.0177 rs9469718 (mGluR4) Preference (IMP) 0.0314 rs11759763 (mGluR4) Preference (MSG+IMP) 0.0162 rs9469718 (mGluR4) Preference (MSG+IMP) 0.0446 rs222745 (TRPV1) 0.0182 rs4790522 (TRPV1) 0.0319 Threshold rs4790151 (TRPV1) 0.0003* rs4790522 (TRPV1) Salt 0.0643 rs8065080 (TRPV1) Suprathreshold 0.0316 rs4790151 (TRPV1) 0.0623 Preference rs4790522 (TRPV1) 0.0341 rs1527483 (CD36) Threshold 0.0248 Fat rs17108968 (GPR120) Suprathreshold 0.0197 Linear regressions were conducted each for detection threshold, suprathreshold sensitivity, and taste preference with SNP predictors and age, sex, ethnicity, and BMI as covariates (n=49). Regressions were only performed for SNP predictors located in a gene associated with the same taste modality as the taste outcome. *Remained significant following a Bonferroni adjustment for multiple hypothesis testing. The Bonferroni adjustment of the reported p-values accounted for the number of hypotheses equal to the product of the number of SNPs in genes associated with each taste modality and the number of taste outcomes (three).

While no significant differences in anthropometric or blood pressure outcomes were observed based on

SNP genotypes, this was not unexpected due to the fact our sample included only healthy young adults.

47 Moreover, examining taste perception in healthy participants was advantageous due to the potential moderating effect of BMI on taste perception [74, 76, 158-160].

Table 3.4: Multiple trait analysis of detection threshold, suprathreshold sensitivity, and taste preference with SNP predictors SNP ID (gene) Taste Modality Outcomes P-value

rs1873249 (mGluR4) Umami Umami (MSG+IMP) 0.017

rs224534 (TRPV1) Salt Umami (MSG) 0.025

rs224534 (TRPV1) Salt Umami (MSG+IMP) 0.047

rs4790522 (TRPV1) Salt Umami (MSG) 0.012

rs236512 (KCNJ2) Sour Sour 0.049

rs643637 (KCNJ2) Sour Sour 0.017 Multivariate regressions were conducted for multiple taste outcomes within one taste modality including detection threshold, suprathreshold sensitivity, and taste preference with SNP predictors (n=49). Regressions were only performed for SNP predictors located in a gene associated with the same taste modality as the taste outcome. A Bonferroni adjustment was applied to the reported p-values, with the number of hypotheses equal to the number of SNPs in genes associated with each taste modality. No p- values remained significant following Bonferroni adjustment.

Umami taste is thought to drive protein intake, and sensitivity to umami tastants has been implicated in a variety of sensory and health-related outcomes such as salivary secretion [170], taste enhancement [171], increasing appetite and satiety [172, 173], reducing fat mass in rats [174], and obesity

[159]. Furthermore, previous research has suggested that umami taste sensitivity is related to an increased liking of dietary protein and the nutritional need of protein in high-protein consumers [175]. Despite the potential importance of umami in the intake of dietary protein and health-related outcomes, studies that investigate the genetic determinants of umami taste sensitivity and preference are limited [144, 145, 176].

Similar to these previous studies, findings of this study indicate that SNPs in umami taste receptor genes are associated with umami taste. Notably, MSG and IMP were tested separately and together as these two compounds are known to elicit the umami taste synergistically when combined [177]. Moreover, the

MSG stimulus reproduces the sensation of dietary glutamate that is present in both animal and plant protein while the IMP stimulus reproduces the sensation of protein from animal sources [177]. While

48 some results were different depending on the umami stimulus, it was not within the scope of this study to determine how these differences relate to dietary intake of animal protein or plant protein. However, further characterization of SNPs implicated in umami taste may explain individual predisposition to preferring umami foods. This knowledge may prove to be important given that the consumption of umami foods is implicated in energy intake and therefore the risk of obesity [159, 172-175].

The genetic predisposition to salt sensitivity has been attributed to variation in genes related to homeostatic sodium regulation [178-181] and to genes which may alter hedonic responses to the taste of salt [182]. Sensitivity to the dietary intake of sodium is important to characterize due to its potential impact on the susceptibility to hypertension, a risk factor for the development of cardiovascular disease

[183-186]. This study demonstrates both novel and previously reported associations between SNPs in the

TRPV1 salt taste receptor gene and salt taste sensitivity as well as preference. To the authors’ knowledge, no associations have previously been found between the rs4790151 SNP in TRPV1 and salt taste.

Furthermore, the relationship between the rs8065080 SNP in the TRPV1 salt taste receptor gene and suprathreshold sensitivity to salt in this study is consistent with previous research [182], however this

SNP was not significantly associated following a correction for multiple testing. Due to the exploratory nature of this study, future research should consider SNPs with significant associations before adjustment for multiple testing as well as the rs4790151 SNP when investigating salt taste sensitivity and preference.

Studies with larger sample sizes are warranted to replicate these results in order to better understand the genetic basis for salt sensitivity, and therefore hypertension.

Genetic variations in TAS1R2 and TAS1R3 sweet taste receptor genes are important to characterize because they pertain to changes in taste sensitivity to sugar [58, 101, 102, 105-109], the excessive consumption of which is an established risk factor for obesity and chronic disease [187-189].

Previous research has implicated SNPs in TAS1R2 and TAS1R3 in inter-individual differences in sugar sensitivity [113, 158] and dietary intake [112, 158, 190-192]. However, this study is the first to find an association between a SNP in a sweet-tasting gene and sucrose preference in humans. However, this

49 association was not significant following a Bonferroni adjustment. While it is important to conduct studies pertaining to sweet taste sensitivity, the examination of sucrose preference and TAS1R2 and

TAS1R3 SNPs in this study offers a more plausible link between genetics and food selection. The rs4920564 SNP may be able to better predict dietary intake of sweet foods due to its association with sweet taste preference, although diet records were not obtained in this study and the sensory stimuli were not food. More research pertaining to this variant is warranted, particularly to assess its association with the consumption of sweet foods. By establishing these types of associations in future studies, genetic loci can be considered risk factors for the overconsumption of sweet foods and be used clinically to indicate the risk of developing obesity and other chronic diseases.

Acidic substances elicit sour taste through the depolarization of type III taste bud cells [35].

While most mammals reject strong sour stimuli as a means to avoid the consumption of spoiled foods, many animals find mildly acidic foods to be palatable. Törnwall et al. (2012) showed that genetic factors may be more important than shared environment to determine the pleasantness and intensity of sour taste, and that 34-50% of the variation in pleasantness and use-frequency of sour foods is attributable to genetics [193]. This demonstrated that there is a genetic basis to the preference for sour foods in humans.

Seminal work by Ye et al. (2016) proposed that intracellular acidification of type III taste bud cells inhibits an inwardly rectifying K+ current mediated by the potassium ion channel KIR2.1, encoded by the

KCNJ2 gene [150]. This work provided the rationale to select KCNJ2 as a candidate gene for analysis of

SNPs and sour taste perception in this study. Remarkably, two SNPs in the KCNJ2 gene were associated with at least one sour taste outcome in the multivariate analysis (DT, ST or PR) before applying the adjustment for multiple testing. This novel finding provides a foundation for future studies to investigate the genetic basis of sour taste as well as sour food intake.

There are some limitations to consider in this study. Firstly, the data obtained by assessing taste sensitivity and taste preference using isolated compounds on filter paper strips cannot be used to make direct associations between genetics and food intake. However, this can also be considered a strength of

50 the study as the observations made are accurate for specific taste modalities whereas using food would have introduced uncertainty due to the perception of texture, temperature, and other matrix-specific qualities. Future studies would benefit from performing hedonic tests with a food that is standardized with one matrix for all taste modalities in order to determine the relationship between well-studied SNPs in taste receptor genes and food selection. In addition, participants were tested for DT, ST and PR on only one occasion, but should be repeated to confirm validity. While this study was powered to observe differences in sensory outcomes, the sample size was small and the likelihood of making type II errors would be lower with a larger sample.

3.6 Conclusion

Overall, the findings in this study showed that SNPs in taste receptor genes may contribute to inter-individual differences in psychophysical measures of taste in healthy young adults. These findings are novel and unique additions to the body of literature about the genetic basis for variation in taste function. Genetic variation in taste is important to characterize as it contributes to differences in dietary intake of key macronutrients and micronutrients. Advances in our understanding about the genetics of taste may help to develop effective strategies to improve eating habits, and therefore risk of obesity through personalized nutritional recommendations based on unique taste preferences.

3.7 Acknowledgements

E.C. conceived and designed the experiments, collected and analyzed data, and wrote the paper; A.A.S.L. helped with data collection and analysis; L.D. helped to design the sensory experiments and provided critical revision for the paper; Z.F. helped with the statistical analysis and provided critical revision for the paper; G.D. helped with the statistical analysis and provided critical revision for the paper; A.M.D. helped with the dietary data and provided critical revision for the paper; J.H. provided critical revision for the paper; D.W.L.M. conceived and designed the experiments and provided critical revision for the paper. Support was provided by the Ontario Ministry of Agriculture, Food, and Rural Affairs (grant# 110- 029900-000000-030194)

51 CHAPTER 4 RATIONALE

The results from Chapter 3 established that specific SNPs in taste receptor genes are associated with measures of taste sensitivity and taste preference. Notably, SNPs exhibited associations with either a measure of taste sensitivity or taste preference, but rarely both types of measures. This observation suggests that sensitivity and preference may have distinct determinants. Moreover, this poses an issue when interpreting the results due to the lack of understanding of the relationship between taste sensitivity and taste preference. Developing this understanding would facilitate the interpretation of findings from sensory studies in the broad context of food preferences and eating patterns. While this contextualization is important for sensory studies in general, it is particularly important for this thesis as later work examines how SNPs in taste receptor genes may associate with eating patterns (Chapter 5). Therefore, the analyses in Chapter 4 were conducted to investigate potential correlations between measures of taste sensitivity and taste preference as well as the attributes of these measures.

52 CHAPTER 4

TASTE SENSITIVITY AND TASTE PREFERENCE MEASURES ARE CORRELATED IN

HEALTHY YOUNG ADULTS

This chapter is currently under review in the journal Chemical Senses.

Chamoun, E., Liu, A. S., Duizer, L.M., Feng, Z., Darlington, G, Duncan, A. M., Haines, J., and Ma,

D.W.L. Taste sensitivity and taste preference measures are correlated in healthy young adults. Chem

Senses. Submitted: Mar 2018.

53 4.1 Abstract

Taste is fundamentally important for food selection. While measures of taste sensitivity and taste preference have been refined over several decades, it remains largely unknown how these measures relate to each other and to food preferences. The objectives of this study were to examine, in healthy adults (age

24.6 ± 0.6 years, n=49), 1) correlations among measures of taste sensitivity, including detection threshold

(DT) and suprathreshold sensitivity (ST), and taste preference (PR) within sweet, salt, sour, umami, and fat tastes, and 2) the underlying associations among DT, ST and PR measurements using principal component analysis. As expected, DTs and STs were inversely correlated within each taste modality. Salt, sweet, and umami DTs and STs were positively and inversely correlated with PRs, respectively. No correlations were observed between sour and fat DTs, STs and PRs. Two principal components accounted for 41.9% of the variance and produced three clear clusters, where each cluster generally consisted of

DTs, STs or PRs from each taste modality. Sweet PR and fat ST deviated from the clusters and may therefore be driven by different factors. Overall, this study provides evidence that higher sensitivities only to salt, sweet, or umami taste are associated with a decrease in the preference for these tastes. These findings demonstrate the importance of investigating taste sensitivity together with taste preference to gain a more complete understanding of the determinants of food selection. (Support was provided by the

Ontario Ministry of Agriculture, Food, and Rural Affairs)

54 4.2 Introduction

With the growing prevalence of chronic disease in modern society, there has been a multitude of psychophysical studies to characterize the influence of taste in poor eating habits and overweight/obesity.

These advances in research on human taste perception have focused mainly on examining broad associations between measures of taste sensitivity, dietary intake, and physical measures of health such as body mass index (BMI) and blood pressure (BP) [194-206]. In general, these studies suggest that low salt, sweet, and fat taste sensitivity is associated with adverse eating habits and overweight/obesity. While studies that link taste perception to diet and BMI are important to establish the fundamental role of taste in health, their findings are not always congruent and more detailed analyses are needed to understand how differences in taste perception lead to changes in food selection. As taste has been shown to play a fundamental role in food selection [165], it is imperative that research in nutrition and sensory science aims to develop a complete understanding of taste function which includes both analytical and affective psychophysical properties.

While analytical measures of taste such as oral detection thresholds (DT) and suprathreshold sensitivity (ST) are important tools to characterize inter-individual differences in taste, understanding how these relate to affective measures of taste would facilitate the interpretation of psychophysical data in the context of food selection. Oral DT is a measure which indicates the lowest concentration of a compound that an individual can detect using taste receptors on the tongue, while ST is a measure of the perceived intensity of a range of concentrations above the DT. Pairing these analytical measures of taste with affective measures, such as taste preference (PR), is an important intermediate step in determining whether taste sensitivity influences food selection. A study by Bossola et al. (2007) tested the relationship between sweet, salt, sour and bitter taste sensitivity and hedonic response in gastrointestinal cancer patients as well as healthy controls. In both groups of participants, a linear negative correlation was reported between taste sensitivity and preference for salt, sour, and bitter whereas the opposite was true for sweet [207]. In a study that examined umami taste sensitivity and preference in adults, no association was found; however, self-reported liking and preference scores for high-protein foods were found to be

55 negatively correlated with umami DT [208]. An inverse correlation between fat taste sensitivity and the preference of fat in a tomato soup was reported in another sample of adults [209]. In a separate study by the same authors, fat taste sensitivity was negatively associated with the intake of high-fat meals in adults, but only in the presence of low salt [210]. While these previous studies provide insights for future work on the influence of psychometric functions on food selection, more research is needed to confirm their findings with different stimuli and study populations. In addition, studies examining taste sensitivity and preference would benefit from analyzing all the known types of taste modalities within one group of participants and by using a consistent delivery method of the stimuli.

The primary objective of this study was to investigate correlations among measures of taste sensitivity, including DT, ST and PR within sweet, salt, sour, umami, and fat tastes. Secondly, this study sought to explore the underlying attributes of DT, ST and PR measurements using principal component analysis. The third objective of this study was to examine associations between measured sweet, salt, umami, fat, and sour PRs and bitter intensity ratings with self-reported food preferences. It was hypothesized that 1) DT and ST are negatively correlated for every taste modality, 2) DT is positively associated with PR and ST is negatively associated with PR for every taste modality, and 3) Measured PR will predict self-reported taste and food preferences.

4.3 Methods

4.3.1 Participants

Fifty-three participants were recruited from the University of Guelph campus. Exclusion criteria included smokers, diagnosis a taste dysfunction, and bariatric surgery. Participants were required to be at least 18 years of age, and no other age restrictions were imposed. The age of participants ranged from 18-

44 years, and the median age was 24 years. This study was approved by the Research Ethics Board at the

University of Guelph (REB#16-12-629).

56 4.3.2 Psychophysical Measurements

All psychophysical tests were administered in sensory booths at the University of Guelph Sensory

Laboratory. Filter paper strips (Indigo Instruments – Cat#33814-Ctl; 47mm x 6 mm x 0.3 mm) immersed in varying concentrations of tastants were used for all sensory tests. The tastants were: sucrose for sweet taste (Thermo Fisher; S5-500), monosodium glutamate (MSG) for umami taste (Thermo Fisher;

ICN10180080), inosine monophosphate (IMP) for umami taste (Thermo Fisher; AC226260250), MSG +

IMP, sodium chloride (NaCl) for salt taste (Thermo Fisher; S641-500), citric acid for sour taste (A940-

500), oleic acid for fat taste (A195-500) (Thermo Fisher Scientific), and PTC for bitter taste (Indigo

57 were presented in random order one at a time to participants who were then asked to respond with “yes” or “no” to the question “Did you taste something?”. The first concentration tested was the median of the full range of concentrations used in the test. If a participant correctly identified a taste strip from a blank strip on two consecutive occasions, then the adjacent lower concentration was subsequently tested.

However, if a participant incorrectly distinguished a taste strip from a blank strip on one occasion, then the adjacent higher concentration was subsequently tested. A change in the direction to a lower or higher concentration during the test was recorded as a reversal, and the threshold was calculated as the mean of the concentrations where the first four reversals occurred. In the ST test, participants were presented with a range of taste strips in random order and were asked to rate the intensity of the strips from 0-100 on a gLMS where 0=undetectable, 2=barely detectable, 6=weak, 18=moderate, 35=strong, 52=very strong, and

4.3.3 Data Analysis

Inc., Cary, NC, USA). All DT, ST and PR variables were log-transformed prior to the correlation analysis to reduce the extreme magnitudes among the measurements. The ST variables were continuous, the DT and PR variables were not continuous, and survey responses were a combination of ordinal and binary variables. Therefore, Spearman correlations (rs) were used to examine the relationships among DTs, STs, and PRs. Weak, moderate, and strong correlations were considered to have correlation coefficients of

58 Table 4.1: Range of tastant concentrations used for each psychophysical test

rs<|0.30|, rs=|0.30|-|0.49|, and rs≥|0.50|, respectively [211]. Principal component analysis (PCA) was used to explain the structure of the DT, ST, and PR measures for all taste modalities. Binary logistic regression models were used to examine associations between PRs as well as PTC intensity ratings and binary survey response variables. Spearman correlations were used to examine associations between PRs as well as PTC intensity ratings and ordinal survey response variables. Statistical significance was set at p<0.05.

4.4 Results

While 53 participants were recruited for the study, four participants relocated before completing the study and were therefore excluded. Thirty-five females and fourteen males with a mean age of 25 ± 3 years completed the study. The mean BMI of participants (22.3 ± 2.6 kg/m2) indicated, on average, a normal body weight. Measures of anthropometry, body composition, blood pressure, and heart rate are summarized in Table 4.2.

4.4.1 Correlation Analysis of Detection Threshold, Suprathreshold Sensitivity, and Taste Preference

Correlation matrices for DT, ST, and PR within each taste modality are presented in Table 4.3.

Measures of DT and ST were consistently negatively correlated, however the negative correlation for sour

DT and ST was not significant. Moderately positive correlations were observed between DT and PR for umami and salt taste while sweet DT and PR had a weak positive correlation which was not significant.

Fat DT had a weak negative correlation with fat PR and this was not significant. No correlation was

59 observed between sour DT and PR. Moderately negative correlations were observed between ST and PR for umami (Figure 4.1) and salt taste (Figure 4.2) while sweet and fat ST and PR had weak negative correlations which were not significant. No correlation was observed between sour ST and PR.

Table 4.2: Participant Characteristics

Characteristic Total Female Male n 49 35 14

Age (years) 25 ± 3 25 ± 3 24 ± 2

Body mass index (kg/m2) 22.3 ± 2.6 21.2 ± 2.0 25.1 ± 2.5

Body fat (%) 20.2 ± 5.8 22.2 ± 5.0 14.8 ± 5.5

Ethnicity (%)

Caucasian 75 - -

South-east Asian 10 - -

Latin American 10 - -

Other 5 - - Characteristics that are not presented as a count or percentage are means ± SD.

4.4.2 Principal Component Analysis of Detection Threshold, Suprathreshold Sensitivity, and Taste

Preference

The result of the PCA for DTs, STs, and PRs in all taste modalities is presented in Figure 4.3. The two components that accounted for the most variance comprised 41.85% of the variance. Components 1 and 2 accounted for 26.13% and 15.72% of the variance, respectively. Subsequent components did not account for more than 10% of the variance and were therefore not considered for further analysis. Three clear clusters were formed as a result of the PCA including a DT cluster (blue), an ST cluster (red), and a

PR cluster (green) (Figure 4.3). One cluster is formed by sweet PR and DT for fat, sweet, salt, sour, MSG,

IMP, and MSG+IMP (top-right quadrant, Figure 4.3). A second cluster is formed by ST for sweet, salt,

60 sour, MSG, IMP, and MSG+IMP (top-left quadrant, Figure 4.3). Thirdly, a cluster is formed by PR for fat, salt, sour, MSG, IMP, and MSG+IMP (bottom-right quadrant, Figure 4.3).

Table 4.3: Correlation matrices for psychophysical measures of taste

Taste Modality Correlation Coefficients (r ) (stimulus) s DT PR DT - 0.22 Sweet (sucrose) ST -0.54* -0.20 DT - 0.46* Umami (MSG) ST -0.43* -0.59* DT - 0.36* Umami (IMP) ST -0.29* -0.32 DT - 0.44* Umami (MSG+IMP) ST -0.35* -0.62* DT - 0.36* Salt (NaCl) ST -0.44* -0.35* DT - 0.00 Sour (citric acid) ST -0.19 -0.09 DT - -0.17 Fat (oleic acid) ST -0.64* -0.17 Spearman correlations were calculated to determine the relationship between detection threshold (DT), suprathreshold sensitivity (ST), and preference (PR) for sweet, umami, salt, sour, and fat taste in healthy adults (n=49). *p<0.05

4.5 Discussion

This study provides evidence that measures of taste sensitivity and taste preference for some types of taste are correlated. In particular, higher sensitivities to salt, sweet, and umami taste are associated with decreases in the preferences for these tastes whereas fat and sour preferences were not associated with corresponding taste sensitivities. A linear negative correlation was reported between taste sensitivity and preference for salt and sour in a previous study [207], but that finding was only replicated for salt in this study. The linear positive correlation observed by Bossola et al. (2007) between sweet sensitivity and preference was the opposite trend from what was found in this study [207]. No association was reported

61 3 rs= -0.62 2.5 p<0.001 2 1.5 1 0.5 0 -0.5 -1 -1.5 Log MSG+IMP Preference (mM) Preference MSG+IMP Log 2.75 3.25 3.75 4.25 4.75 Log AUC of MSG+IMP Suprathrshold Intensity Ratings

Figure 4.1: Spearman Correlation between Umami Suprathreshold Sensitivity and Umami Preference Spearman correlation was used to determine the correlation between umami ST and umami PR. The variables were logged to reduce the magnitude of their scales. Umami ST was measured by the AUC of intensity ratings for MSG+IMP and umami PR was determined using the Monell forced-choice paired comparison tracking method for MSG+IMP. Umami ST and PR were negatively correlated with a Spearman correlation coefficient of -0.62 (p<0.001). AUC=area-under-the-curve; ST=suprathreshold sensitivity; PR=preference; MSG=monosodium glutamate; IMP=inosine monophosphate.

2.70 rs= -0.35 2.50 p<0.05

2.30

2.10

1.90

1.70

1.50 Log of NaCl Preference (mM) Preference NaCl of Log 1.50 2.00 2.50 3.00 3.50 4.00 Log AUC of NaCl Suprathreshold Intensity Ratings

Figure 4.2: Spearman Correlation between Salt Suprathreshold Sensitivity and Salt Preference Spearman correlation was used to determine the correlation between salt ST and umami PR. The variables were logged to reduce the magnitude of their scales. Salt ST was measured by the AUC of intensity ratings for NaCl and salt PR was determined using the Monell forced-choice paired comparison tracking method for NaCl. Salt ST and PR were negatively correlated with a Spearman correlation coefficient of - 0.35 (p<0.05). AUC=area-under-the-curve; ST=suprathreshold sensitivity; PR=preference; NaCl=sodium chloride.

62 Component 1 (26.13%)

ponent 2 (15.72%) 2 ponent Com

Figure 4.3: Principal Component Analysis of Detection Threshold, Suprathreshold Sensitivity, and Taste Preference Principal component analysis (PCA) was used to explain the structure of the detection threshold, suprathreshold sensitivity, and taste preference measures for all taste modalities (n=49). Values on the x- and y-axes represent the component scores. Variables were log-transformed to reduce the large differences in magnitudes between the measurements. The two components that accounted for the most variance comprised 41.85% of the variance. Component 1 accounted for 26.13% of the variance while Component 2 accounted for 15.72% of the variance. Subsequent components did not account for more than 10% of the variance and were therefore not considered for further analysis. Three clear clusters were formed as a result of the analysis including of a DT cluster (blue), a ST cluster (red), and a PR cluster (green). T=detection threshold; ST=suprathreshold sensitivity; P=preference; MSG=monosodium glutamate; IMP=inosine monophosphate.

in a previous study between umami DT and umami PR based on liking scores for umami foods [208]; however, a significant negative correlation was observed in this study. A negative correlation between fat taste sensitivity and fat taste preference has been suggested in previous work [209], but this result was not

63 replicated here. Altogether, there is some inconsistency between results reported in this work compared to previous studies. These differences may be due to the methodologies used and more research is needed to confirm all of these experiments. Bossola et al. (2007) tested cancer patients as opposed to young healthy adults, used solutions instead of filter paper strips, and measured pleasantness using a hedonic rating.

Bolhuis et al. (2016) tested the relationship between fat taste sensitivity and pleasantness of fat using a food stimulus and a ranking procedure instead of filter paper strips and a forced-choice paired-comparison tracking method. Nevertheless, the findings of this work advance the current understanding of the role that taste sensitivity plays in taste preference for specific taste modalities.

While correlations were useful to determine the magnitude and direction of associations between

DT, ST, and PR for each taste modality, the PCA provided additional insight into whether similar factors accounted for those correlations. Component 1 of the PCA is a factor that tends to reflect intensity perception whereby the ST data is clustered at the negative side of the Component 1 axis and DT data is clustered at the positive side of the Component 1 axis. Participants with higher detection thresholds are less sensitive to stimuli and therefore rate stimuli at lower intensities in the suprathreshold sensitivity test.

This pattern follows a logical and expected outcome for the relationship between DT and ST. Component

2 emerged as the factor that contributed to determining preference for every type of taste except for sweetness. While correlations of sweet sensitivity and preference were observed to have the same directionality as for other taste modalities, the underlying attribute of sweet PR was different than that of the other types of taste. The weaker positive correlation observed between sweet DT and PR was likely due to the lack of Component 2 influence on sweet PR. Component 2 is speculated to be a preference for high concentrations of the compounds tested. It can then be interpreted that, in general, the presence of a high concentration of sucrose in the samples tested elicited a hedonic liking response in the participants.

Consistent with the correlation analysis, this interpretation also suggests that high concentrations of all the other tastants elicited an aversive response in the participants.

Oleogustus is the most recently identified taste modality and refers to the taste of fatty acids

[212]. The taste of long-chain fatty acids (LCFA) was previously thought to only be dependent on the

64 texture and smell [61-63], but more recent evidence supports taste as a method of oral detection of dietary fat, including saturated fat [64, 212, 213]. Previous research has measured fat DT, ST, and PR using a variety of delivery methods such as deionized water, milk, pudding, custard, tomato soup, filter paper strips, edible strips, etc [214]. While the ideal technique and delivery method for determining fat sensitivity and preference is controversial [214], the strong inverse correlation between DT and ST for fat in this study coincides with what is expected for these measurements and therefore increased the validity of the measurement. The sensitivity-preference relationship for fat is opposite from that of all other taste modalities according to Components 1 and 2 from the PCA. While individual correlations did not indicate that fat DT or ST were significantly correlated with PR, it is worthwhile to consider in future research that the hedonic response to fatty acids as a stimulus may have a different mechanism than for other types of taste. It is also possible that the delivery method of fatty acids in this study was not suitable for measurement of preference due to the influence of other properties of fat on preference, such as texture, smell, and mouthfeel. Incorporating fatty acids into food matrices may be a more accurate method of measuring fat taste preference, despite the disavantage of variable perceptions of matrix-specific textures by participants.

There are some additional limitations to consider in this study. Firstly, the concentrations of the tastant solutions used to immerse the filter paper strips may not be an accurate representation of the amount of tastant dissolved on the tongues of participants. Moreover, the data obtained by assessing taste sensitivity and taste preference using isolated compounds on filter paper strips cannot be used to make direct associations between taste perception and food preference. However, this can also be considered a strength of the study as the observations made are accurate for specific taste modalities whereas using food would have introduced uncertainty due to the perception of texture, temperature, and other matrix- specific qualities. Future studies would benefit from performing hedonic tests with a food that is standardized with one matrix for all taste modalities in order to determine taste preferences. In addition, participants were tested for DT, ST and PR on only one occasion, but this procedure should be repeated to ensure test reproducibility.

65 4.6 Conclusion

Overall, this study provides evidence that taste sensitivity and taste preference measurements, when tested using a filter paper carrier system, are correlated in young healthy adults. In particular, higher sensitivities to salt, sweet, or umami taste were associated with a decrease in the preference for these tastes whereas this was not observed for sour and fat tastes. Furthermore, this study demonstrated that the determinants of sweet and fat taste preference may be different than those of other types of taste. These findings demonstrate the importance of investigating taste sensitivity together with taste preference to gain a more complete understanding of the determinants of food selection.

4.7 Acknowledgements

E.C. conceived and designed the experiments, collected and analyzed data, and wrote the paper; A.A.S.L. helped with data collection and analysis; L.D. helped to design the sensory experiments and provided critical revision for the paper; G.D. helped with the statistical analysis and provided critical revision for the paper; A.M.D. helped with the dietary data and provided critical revision for the paper; J.H. provided critical revision for the paper; D.W.L.M. conceived and designed the experiments and provided critical revision for the paper. Support was provided by the Ontario Ministry of Agriculture, Food, and Rural Affairs (grant# 110-029900-000000-030194)

66 CHAPTER 5 RATIONALE

Thus far, this thesis has shown that SNPs in taste receptor genes are associated with taste perception. What remains to be examined is how these relationships might impact food preferences and eating patterns. Ideally, genetics, taste perception, and eating patterns are analyzed in the same study and within the same sample to establish a strong connection between these factors. This has been possible to undertake in Chapter 6 of this thesis due to the availability of an adequate number of families recruited from the GFHS for a separate sensory study. In Chapter 5 however, the number of families was only high enough to conduct analyses with genetics and eating patterns as these measures are already collected by the GFHS and a separate study was not necessary. Furthermore, the GFHS only had resources at the time to genotype a limited number of SNPs. Therefore, only three SNPs were selected but that have been previously associated with taste perception and dietary intake (reviewed in Chapter 5 Introduction).

Chapter 5 importantly presents analyses that demonstrate proof-of-concept that SNPs in taste receptor genes have the potential to affect eating patterns, and this may be through their impacts on taste function. In particular, snacking patterns were analyzed as snacks constituted 32% of the children’s calories and represent an eating occasion which is typically representative of an individual’s taste preference. As previously discussed in this thesis, children’s taste and food preference may be more reflective of genetic predisposition due to their relative lack of exposure to cultural norms of eating.

Therefore, Chapter 5 was well-positioned to examine if SNPs in taste receptor genes are associated with eating patterns in preschool-aged children in the GFHS.

67 CHAPTER 5

SINGLE NUCLEOTIDE POLYMORPHISMS IN TASTE RECEPTOR GENES ARE

ASSOCIATED WITH SNACKING PATTERNS OF PRESCHOOL-AGED CHILDREN IN THE

GUELPH FAMILY HEALTH STUDY: A PILOT STUDY

This chapter has been published in the journal Nutrients.

Chamoun, E., Hutchinson, J., Krystia, O., Mirotta, J., Mutch, DM., Buchholz, AC., Duncan, AM., Haines,

J. Ma, DWL. and the Guelph Family Health Study. Single Nucleotide Polymorphisms in Taste Receptor

Genes Are Associated with Snacking Patterns of Preschool-Aged Children in the Guelph Family Health

Study: A Pilot Study. Nutrients. 2018,10:153.

68 5.1 Abstract

Snacking is an integral component of eating habits in young children that is often overlooked in nutrition research. While snacking is a substantial source of calories in preschoolers’ diets, there is limited knowledge about the factors that drive snacking patterns. The genetics of taste may help to better understand snacking patterns of children. The rs1761667 single nucleotide polymorphism (SNP) in the

CD36 gene has been linked to fat taste sensitivity, the rs35874116 SNP in the TAS1R2 gene has been related to sweet taste preference, and the rs713598 SNP in the TAS2R38 gene has been associated with aversion to bitter, green leafy vegetables. This study seeks to determine the cross-sectional associations between three taste receptor SNPs and snacking patterns among preschoolers in the Guelph Family

Health Study. Preschoolers’ snack quality, quantity, and frequency were assessed using three-day food records and saliva was collected for SNP genotyping (n=47). Children with the TT genotype in TAS1R2 consumed snacks with significantly more calories from sugar, and these snacks were consumed mostly in the evening. Total energy density of snacks was highest in the CC and CG genotypes compared to the

GG genotype in TAS2R38, and also greater in the AA genotype in CD36 compared to G allele carriers, however this difference was not individually attributable to energy from fat, carbohydrates, sugar, or protein. Genetic variation in taste receptors may influence snacking patterns of preschoolers.

Keywords: Genetics; taste; children; snacking

69 5.2 Introduction

The prevalence of overweight and obesity in Canada has been growing at an alarming rate, and children are not exempt from this epidemic growth. According to the World Health Organization, 31.5% of children and adolescents aged 5-17 years old — an estimated 1.6 million Canadians — were classified as overweight (19.8%) or obese (11.7%) from 2009 to 2011 [215]. The accessibility of palatable, processed foods high in fats and sugars has increased in modern society. Chronic diseases such as obesity, cardiovascular disease (CVD), type 2 diabetes (T2D), and metabolic syndrome (MetS) have been partially attributed to poor eating habits in those compelled by the hedonic qualities of food [216]. The taste of food is a particularly important factor for parents when selecting foods for themselves and their children

[3]. Through the individual selection of unhealthy foods, taste perception may therefore be an important determinant of chronic disease risk.

Research on the genetic predisposition to selecting specific foods based on taste perception may be significant in advancing our fundamental knowledge of genetic factors influencing habitual dietary intake and the development of chronic disease. Genetic variation in taste perception can be characterized in terms of single nucleotide polymorphisms (SNPs) in taste receptor genes. Taste receptor function for sweet, fat, and bitter tastes forms much of the basis for the known inter-individual differences in taste perception [7, 217]. The fat taste receptor cluster determinant 36 (CD36), the sweet taste receptors type 1 member 2 (T1R2) and T1R3, and the bitter taste receptor T2R38 contain common polymorphisms that may alter taste perception, food preference, and consequently food selection [216].

Research related to fat taste perception has focused on the fatty acid translocase CD36 gene.

Decreases in fatty acid taste sensitivity have been associated with the AA genotype of the rs1761667 SNP in the CD36 gene [88]. A lower oral sensitivity to fatty acids has been found to associate with a higher preference and consumption of dietary fat [78]. However, little is known about the effect of this SNP on dietary intake in children.

70 The heterodimer T1R2 and T1R3 allows humans to taste a variety of sweet substances [20, 93], and genetic variations in TAS1R2 may pertain to changes in taste sensitivity to sugar [58, 101-109].

Specifically, the TT genotype of the rs35874116 SNP in the TAS1R2 gene has been linked to habitual sugar consumption in overweight and obese adults [111] and children [218]. In children, dental caries risk is a concern associated with the consumption of sugar [114]. The prevalence of dental caries was observed to be associated with the rs35874116 SNP [53]. Research to date has yet to show a relationship between the rs35874116 SNP and sugar consumption in children.

The perceptions of bitter and sweet tastes interact in such a way to affect eating habits, especially in children [7]. Consuming bitter compounds generally results in food rejection, an evolutionary adaptation to avoid toxic substances such as rancid fat, hydrolyzed protein and plant alkaloids [116, 117].

Sensitivity to bitterness may lead to the avoidance of Brassica vegetables rich in fiber, thereby potentially being replaced with the consumption of energy-dense foods rich in sugar [219, 220]. These eating habits have the potential to increase the risk of obesity, CVD, type 2 diabetes, and cancer [118]. Kim et al.

(2003) showed that the bitter receptor T2R38 mediated the sensitivity to PTC, a thiol compound chemically related to those found in green, leafy vegetables (Brassica vegetables) [120]. Three common

SNPs in the TAS2R38 gene result in amino acid substitutions at residues P49A (rs713598), A262V

(rs1726866) and V296I (rs10246939). The C allele of the rs713598 SNP is associated with the PAV haplotype while the G allele is associated with the AVI haplotype. Individuals carrying two copies of the

PAV haplotype typically have a strong aversion to the bitterness of PTC. Those who seldom taste PTC carry two copies of the AVI haplotype while heterozygotes (PAV/AVI) typically have an intermediate taste phenotype [121].

The prevalence of snacking, or the consumption of foods and/or beverages between meals, is increasing in children as 98% of U.S. children snacked daily in 2006 compared to 74% in 1977 [221].

Young children are especially inclined to snacking and typically consume small meals throughout the day rather than larger meals [222]. These snacks are frequently energy-dense and nutrient-poor choices such

71 as desserts, salty snacks, and sugary drinks which lead to excess energy consumption [222]. While existing evidence identifies snacking as a substantial source of calories in preschoolers’ diets, there is limited knowledge about the factors that drive snacking patterns, including snacking frequency and the quantity and quality of snacks consumed. The genetics of taste is one factor that may help better understand snacking patterns of children. Studies examining the relationship between the genetics of taste and snacking in children are limited despite existing evidence that taste receptor SNPs may impact taste perception, and therefore food selection [216]. Examining snacking patterns among children may provide important insight into the contribution of genetic variation in taste receptors to food selection.

The putative fat taste receptor CD36, the sweet taste receptor T1R2, and the bitter taste receptor

T2R38 contain common polymorphisms which may alter taste perception, food preferences, and consequently snacking patterns. Thus, the present study aimed to determine the relationship between

SNPs in CD36, TAS1R2, and TAS2R38 taste receptor genes and snacking patterns measured in children aged 1.5-5 years in the Guelph Family Health Study pilot.

5.3 Materials and Methods

5.3.1 Study Design

To investigate the effect of genetic variation in taste receptors on snack consumption in preschoolers, baseline data from the Guelph Family Health Study (GFHS) pilot were analyzed cross-sectionally. The

GFHS is a family-based cohort study that aims to understand early life risk factors for chronic disease and test ways to help families maintain healthy behaviours over many years. Families were eligible to participate if they had at least one child aged 1.5-5 years at the time of recruitment, lived in the

Wellington County area, and did not plan to relocate outside of this area within the first year of the study.

Forty-four families participated in the GFHS pilot study. Forty-seven preschoolers from 38 families attended a health assessment visit during which saliva was collected and body mass index (BMI) was measured. Preschoolers’ dietary intake from snacks was assessed using parent-completed three-day food records. All study procedures were administered after the parents of the participants gave written,

72 informed consent. The study was approved by the University of Guelph Research Ethics Board

(REB#14AP008).

5.3.2 SNP Genotyping

Saliva was collected using the Oragene•DNA (OG-575) collection kit for Assisted Collection (DNA

Genotek). Participants were fasted for a minimum of 30 minutes before the saliva sample was provided.

Genetic material from saliva was extracted by ethanol precipitation according to the manufacturer’s protocol (DNA Genotek). Real-time polymerase chain reaction (RT-PCR) was used to identify the genotypes of the participants with respect to the CD36 (rs1761667; Assay ID C___8314999_10), TAS1R2

(rs35874116; Assay ID C_____55646_20), and TAS2R38 SNPs (rs713598; Assay ID C___8876467_10).

Taqman fluorescent oligonucleotide probes were obtained for each SNP (Thermo Fisher Scientific, Cat.

#4351379). A BioRad CFX96 Touch™ RT-PCR Detection System in tandem with the Bio-Rad CFX

Manager Software was used to detect the fluorescent signals and to produce a graphical representation which allowed for allelic discrimination.

5.3.3 Snack Frequency, Quantity, Quality and Composition

Parents completed a three-day food record for their children, including two weekdays and one weekend day. Parents documented a detailed description of each food item (i.e., cooking method, brand name) and the amount of the food or beverage consumed. Food records were inputted into a nutrient analysis program (ESHA Food Processor, Version 11.0.110, Salem, Oregon, USA). ESHA contains a database of over 55,000 food items from over 1,700 reputable sources including the Canadian Nutrient

File. Using an average of the three-day food records, snacking data were analyzed based on quality (total energy density and nutrient composition), quantity (total energy intake), and frequency of snacks in the morning, afternoon, and evening. Foods and/or beverages (except water) consumed between meals were recorded as snacks, as previously described in the GFHS [223]. The average daily energy density of snacks (kcal/g) was calculated by dividing the total energy intake from snacks by the total weight of snacks. Percent caloric intake from fat, carbohydrates, and sugar was determined by dividing the calories

73 from each nutrient component in snacks by the total calories of snacks. Quantity was determined as total energy from snacks (kcal/day). Frequency was calculated as the number of snacks consumed per day, and these were divided between morning, afternoon (between lunch and dinner), and evening snacks.

5.3.4 Data Analysis

Generalized estimating equations were used to estimate the regression coefficients for linear models of snack measure outcomes and SNP predictors, with BMI z-score as a covariate. Frequency of evening snacking was normalized by a square transformation to obtain a normal distribution. All analyses account for correlated outcomes resulting from multiple siblings within some families. Results are presented as least square means (LSM) and were not adjusted for multiple testing due to the exploratory nature of the study. Statistical significance was set at p<0.05. Recessive, dominant, and additive models were used to analyze rs1761667 (AA vs AG/GG), rs35874116 (CC/CT vs TT), and rs713598 (GG vs CG vs CC) respectively in line with previous research [88, 111, 123].

5.4 Results

5.4.1 Participant Characteristics and Snacking Patterns

The mean age of the participants was 3.5 ± 1.2 years. Approximately half of the participants were male. The children were mostly of Caucasian origin and a wide range of family incomes was represented

(Table 5.1). Snacking patterns show that all of the children snacked daily, with snacks constituting 32% of total daily energy intake (Table 5.2).

5.4.2 Association of SNPs at the CD36, T1R2, and T2R38 Genes with Snacking Patterns

There were no significant differences in BMI z-score among the participants based on any SNP genotype. The total energy density of snacks was greater in the AA genotype (1.39 ± 0.08 kcal/g) of the

CD36 SNP compared to G allele carriers (1.17 ± 0.03 kcal/g) (Figure 5.1). There were no significant differences in other snack measures by CD36 genotype related to snack frequency, quantity, or calorie composition.

74 Children with the TT genotype of the rs35874116 SNP in TAS1R2 consumed snacks with higher

% calories from sugar than children carrying the C allele (0.40 ± 0.02 and 0.33 ± 0.02, respectively)

(Figure 5.2). In addition, the square-transformed frequency of evening snacks was higher (0.86 ± 0.07) for children with the TT genotype compared to C allele carriers (0.52 ± 0.09) (Figure 5.3). Based on a qualitative categorization of snacks consumed in the evening, the type of snacks consumed were also significantly more likely to be sugary snacks for children carrying the TT genotype (p=0.04; data not shown). However, there were no significant differences in measures of snack quantity and snack frequency in the morning and afternoon.

The total energy density of snacks was highest in the CC genotype (1.31 ± 0.06 kcal/g) and CG genotype (1.24 ± 0.05 kcal/g) compared to GG genotype (1.06 ± 0.05 kcal/g) of the rs713598 SNP in

TAS2R38 (Figure 5.4). There were no differences in other snack measures by TAS2R38 genotype related to snack frequency, quantity, or calorie composition.

Table 5.1: Participant Characteristics

Sex Male n=22 Female n=25 Age (years, mean (SD)) 3.47 (1.15) BMI Z-Score (mean (SD)) 0.43n=22, (0.99) 47% Ethnicity (%) Female Caucasian n=25,87.5 53% Other 12.5 Parent annual household income (Canadian $), n=37* (%)

≤$49,999 18.9 $50,000 - $79,999 21.6 Male $80,000 - $99,999 21.6 ≥$100,000 37.8 *Total Female families that provided parent annual household income (missing data, n=1)

75 Table 5.2: Participant Snacking Patterns

Snacking measure Mean (SD) (n=47)

Total daily energy intake (kcals/day) 1407 (347)

Total energy intake from snacks (kcals/day) 456 (213)

% Daily energy from snacking 32 (14)

Total energy density of snacks (kcals/g) 1.13 (0.43)

Total snack consumption frequency (%) 78 (39)*

Morning (%) 100 (33)*

Afternoon (%) 100 (33)*

Evening (%) 67 (67)*

Overall participant snacking patterns were computed (n=47). *Variable was not normally distributed: median and inter-quartile range used as measures of center and spread

* 1.60

1.40

1.20

1.00

0.80

0.60

0.40

0.20 LSM Energy Energy LSM Density(kcals/g) 0.00 AA AG/GG Genotype (rs1761667) Figure 5.1: Least square means of energy density of snacks by rs1761667 genotype Statistical differences between the total energy density of snacks by genotypes for rs1761667 of the CD36 fat taste receptor were determined using generalized estimating equations (n=47; *p=0.007) and a recessive genetic model [88]. Children carrying the AA genotype (n=13), predicted to prefer fat more than the AG/GG genotypes (n=34), consumed snacks with a higher total energy density than children carrying the G allele. Values are reported as least square means + SE.

76 0.45 * 0.4 0.35 0.3 0.25

Snacks 0.2 0.15

0.1 LSM CaloriesLSM from % in Sugar 0.05 0 CC/CT TT rs35874116 Genotype (TAS1R2) Figure 5.2: Least square means of the percent calories from sugar in snacks by rs35874116 genotype Statistical differences between the relative calories from sugar in snacks for rs35874116 of the TAS1R2 sweet taste receptor were determined using generalized estimating equations (n=47; *p=0.008) and a dominant genetic model [111]. Children carrying the TT genotype (n=25), predicted to prefer sweet, consumed snacks with significantly more calories from sugar than carriers of the C allele (n=22). Values are reported as least square means + SE.

1 * 0.9 0.8 0.7 0.6 0.5

Frequency 0.4 0.3 0.2

LSM LSM SquaredEveningSnack 0.1 0 CC/CT TT rs35874116 Genotype (TAS1R2) Figure 5.3: Least square means of squared evening snack frequency by rs35874116 genotype Statistical differences between the square-transformed frequency of evening snacks for rs35874116 of the TAS1R2 sweet taste receptor were determined using generalized estimating equations (n=47; *p=0.004) and a dominant genetic model [111]. Children carrying the TT genotype (n=25), predicted to prefer sweet more than carriers of the C allele (n=22), consumed more snacks in the evening. Values are reported as least square means + SE.

77 # 1.60 * 1.40

1.20

1.00

0.80

0.60

0.40 LSM Energy Energy LSM density(kcals/g)

0.20

0.00 GG GC CC Genotype (rs713598) Figure 5.4 Least square means of total energy density of snacks by rs713598 genotype Statistical differences between the total energy density of snacks by genotypes for rs713598 of the TAS2R38 bitter taste receptor were determined using generalized estimating equations (n=47; *p=0.01, #p=0.001) and an additive genetic model [123]. Children carrying the CC (n=12) or CG (n=27) genotypes consumed snacks with an overall higher energy density than the children carrying the GG genotype (n=8). Values are reported as least square means + SE.

5.5 Discussion

The present study showed that genetic variation in taste receptors are associated with potentially unhealthy snacking patterns in children. Three SNPs were investigated within taste receptor genes for fat

(CD36; rs1761667), sweet (TAS1R2; rs35874116), and bitter taste (TAS2R38; rs713598). The prevalence of snacking in this cohort is similar to the prevalence of snacking in American children, and the calories from snacks are similar as well [221]. Also, in a Brazilian study, the prevalence of snacking in children can be divided into children who are light-snackers and heavy-snackers. The prevalence of snacking and the calories from snacks are similar between heavy-snackers in Brazil and children from North American cohorts. However, the light-snackers in Brazil consume less snacks which also comprise a smaller portion of calories than North American children [224].

Adults with the AA genotype of the rs1761667 variant have been shown to have lower oral sensitivity to fat and higher preference for fat [72, 88]. In the present study, no relationship was observed

78 between carrying the AA genotype and snack quantity, frequency, or calories from fat; however, children carrying the AA genotype consumed higher energy density of snacks than children carrying the G allele at this locus. While studies have shown differences in sensory outcomes based on this CD36 SNP, an effect on dietary intake has never been observed. The present finding suggests that a fat taste receptor SNP genotype may influence the intake of energy-dense snacks. This finding is potentially due to the decreased sensitivity to fat taste and subsequent higher preference for fatty foods; however, measures of taste sensitivity were not determined in this study. Although the relationship between fat taste sensitivity and energy density of snacks in this study is consistent with the energy density of fat as a macronutrient, no difference was found between the calories from fat in snacks for this genotype. This apparent contradiction may be due to the challenges of measuring the contribution of fat in a complex diet which contains foods simultaneously high in fat and carbohydrates. Future research studies would benefit from performing hedonic tests for fat taste with children in more controlled settings in addition to tracking food records. A genetic predisposition to prefer fatty foods due to lower taste sensitivity may be of concern for children in the long term as high fat diets may lead to metabolic complications such as obesity, MetS, and

CVD [225].

Carriers of the TT genotype at the rs35874116 locus (Ile191Val) in the TAS1R2 sweet taste receptor gene have been shown to prefer sweet foods [111]. Children in the GFHS cohort carrying the TT genotype were found to consume snacks with significantly more calories from sugar than children carrying the C allele. While the percent of calories from sugar was different, it is not clear why the energy density of snacks was not significantly different based on rs35874116 SNP genotype. As snacks contain a complex combination of different macronutrients, it is difficult to discern which macronutrients contribute a higher proportion of energy density. While consistent with previous research linking the

Ile191Val variant with habitual sugar consumption in overweight/obese adults and dental caries in children [53, 111], the present finding demonstrates that taste preferences and dietary intake may also be related in children. Upon further analysis, it was found that sugary snacks were most frequently consumed in the evening by children with the TT genotype. If the evening is a time of day when children with a

79 “sweet tooth” potentially consume more sugary snacks, this finding warrants future studies to assess the availability and accessibility of sugary foods in the home environment as a means to address the role of parents in the intake of sugary foods by young children during the evening. Future studies may also benefit from including genetic variants in the TAS1R3 sweet taste receptor gene to examine determinants of sweet food intake [113, 218]. Chronic overconsumption of sugar is a well-known risk factor for MetS and T2D [189]. With the advancement of research pertaining to this variant, there may be reason to use genotype at this locus as a risk factor for overconsumption of sugar in children. Adapting the Ile191Val

SNP genotype into a clinical biomarker of dietary intake patterns in children may be helpful to mitigate the risk of developing obesity, MetS, and T2D.

The bitterness of green leafy vegetables (Brassica vegetables), related to the taste of thiol compounds, may be stronger in those homozygous for the C allele at the rs713598 locus in the TAS2R38 taste receptor gene. Those who do not carry the C allele may not taste PTC, and this may then influence the perceived bitterness of Brassica vegetables [121]. We hypothesized that carriers of the CC genotype in our study would be more likely to consume energy-dense snacks due to their potential avoidance of bitter vegetables [7, 226]. In support of this hypothesis, CC genotype carriers and CG genotype carriers had higher total energy density of snacks compared to GG genotype carriers. No relationships were observed between rs713598 genotype and snack quantity, frequency, or calorie composition, in line with previous research [227, 228]. This finding contributes to a growing, controversial body of literature about this genetic locus, showing that this genotype may adversely affect quality of dietary intake [124]. Similar to the other variants investigated in the present study, this SNP may serve as a biomarker for at-risk eating patterns with the potential of developing into adverse metabolic and health outcomes.

The prevalence of combinations of the taste receptor SNP genotypes associated with snacking patterns are important to consider. Participants can be classified by the proportion of “at-risk” genotypes they carry based on the three SNPs included in the present analysis. The “at-risk” genotypes were AA for the CD36 SNP (fat preference), TT for the T1R2 SNP (sweet preference), and CC for the T2R38 SNP.

None of the children in this study had all three “at-risk” genotypes, 13/47 (28%) of the children had two

80 “at-risk” genotypes, 24/47 (51%) of the children had one “at-risk” genotype, and 10/47 (21%) of the children had no “at-risk” genotypes. Therefore, the prevalence of carrying at least one “at-risk” genotype was almost 80% within this cohort. This observation is an indicator that these genotypes are sufficiently common and warrant parents’ awareness of the potential influence on habitual dietary intake in young children. It is not known whether the potential disadvantage of carrying “at-risk” genotypes is additive when a combination of these genotypes is present. Moreover, it is difficult to assess whether one “at-risk” genotype can be cancelled out by a “healthy” genotype at a different locus within the same taste receptor gene. Analysis of the relative influences that each SNP has on habitual dietary intake requires a greater understanding of the individual effect on habitual dietary intake as well as a much larger study cohort.

This study’s limitations should be considered when interpreting the results. These data were obtained in a pilot phase of the GFHS, and the probability of type 2 error is higher with a small sample size. Due to the exploratory nature of this study, multiple testing was not accounted for and there is a possibility of false discovery in the findings. Furthermore, there is uncertainty that the quantity, quality, frequency, or composition of snacks consumed reflect the taste preferences of the children rather than parental feeding styles and practices. While it is probable that taste preferences are strong indicators of food selection in this study, the parents’ influence on food availability and accessibility need to be considered in future investigations.

5.6 Conclusions

Overall, this study suggests that genetic variation in taste receptors for fat, sweet, and bitter taste may be associated with snacking patterns in young children. The extent to which the SNPs affect dietary intake must be investigated in future studies by elucidating the mechanisms with which the SNPs alter taste receptor function, and how snacking patterns are subsequently influenced by changes in taste perception.

81 5.7 Acknowledgments

E.C. conceived and designed the experiments, collected and analyzed data, and wrote the paper; J.M.H. helped perform dietary data collection and analysis; O.K. helped perform dietary data collection and analysis; J.A.M. collected dietary data; D.M.M. provided critical revision for the paper, A.C.B. oversaw analysis of snacking and provided critical revision for the paper, A.M.D. helped with the dietary data and provided critical revision for the paper, G.D. helped with the statistical analysis and provided critical revision for the paper, J.H. provided critical revision for the paper, D.W.L.M. conceived and designed the experiments and provided critical revision for the paper.

82 CHAPTER 6 RATIONALE

The first two studies of the thesis examined associations between SNPs in taste receptor genes separately with taste perception and eating patterns. While Study 1 (Chapters 3 and 4) establishes that

SNPs in taste receptor genes are associated with taste perception and Study 2 (Chapter 5) establishes that

SNPs in taste receptor genes are associated with eating patterns, Study 3 (Chapter 6) investigates all of these factors together. As emphasized throughout the thesis, the purpose of investigating genetics as a determinant of taste preferences is to develop a better understanding of eating habits. Chapter 6 combines a comprehensive analysis of SNPs, measures of taste perception, and eating patterns in children to form an integrative analysis of genetics as a determinant of poor eating habits.

Analysis of parents in the upcoming chapter supplements work done in Chapter 3 on genetics and taste perception in adults, however dietary intake data was not collected from the parents in the GFHS at the time this study was conducted. Therefore, dietary analyses were only conducted in children.

Furthermore, dietary analyses in children in the upcoming chapter supplements work done in children’s snacking patters in Chapter 5. Altogether, the analyses undertaken in Chapter 6 represent a succession of the work from previous chapters and an integration of concepts examined thus far.

83 CHAPTER 6

THE RELATIONSHIP BETWEEN SINGLE NUCLEOTIDE POLYMORPHISMS IN TASTE

RECEPTOR GENES, TASTE FUNCTION AND DIETARY INTAKE IN PRESCHOOL-AGED

CHILDREN AND ADULTS IN THE GUELPH FAMILY HEALTH STUDY

This chapter has been published in the journal Nutrients.

Chamoun, E.; Carroll, N.A.; Duizer, L.M.; Qi, W.; Feng, Z.; Darlington, G.; Duncan, A.M.; Haines, J.;

Ma, D.W.; the Guelph Family Health Study. The Relationship between Single Nucleotide Polymorphisms in Taste Receptor Genes, Taste Function and Dietary Intake in Preschool-Aged Children and Adults in the

Guelph Family Health Study. Nutrients 2018, 10, 990.

84 6.1 Abstract

Taste is a fundamental determinant of food selection, and inter-individual variations in taste perception may be important risk factors for poor eating habits and obesity. Characterizing differences in taste perception and their influences on dietary intake may lead to an improved understanding of obesity development and a potential to develop personalized nutrition recommendations. This study explored associations between 93 single nucleotide polymorphisms (SNPs) in sweet, fat, bitter, salt, sour, and umami taste receptors and psychophysical measures of taste. Forty-four families from the Guelph Family

Health Study participated, including 60 children and 65 adults. Saliva was collected for genetic analysis and parents completed a three-day food record for their children. Parents underwent a test for suprathreshold sensitivity (ST) and taste preference (PR) for sweet, fat, salt, umami, and sour as well as a phenylthiocarbamide (PTC) taste status test. Children underwent PR tests and a PTC taste status test.

Analysis of SNPs and psychophysical measures of taste yielded 23 significant associations in parents and

11 in children. After adjusting for multiple hypothesis testing, the rs713598 in the TAS2R38 bitter taste receptor gene and rs236514 in the KCNJ2 sour taste receptor gene remained significantly associated with

PTC, ST and sour PR in parents, respectively. In children, rs173135 in KCNJ2 and rs4790522 in the

TRPV1 salt taste receptor gene remained significantly associated with sour and salt taste PRs, respectively. A multiple trait analysis of PR and nutrient composition of diet in the children revealed that rs9701796 in the TAS1R2 sweet taste receptor gene was associated with both sweet PR and percent energy from added sugar in the diet. These findings provide evidence that for bitter, sour, salt, and sweet taste, certain genetic variants are strongly associated with taste function and may be implicated in eating patterns. (Support was provided by the Ontario Ministry of Agriculture, Food, and Rural Affairs)

Keywords: Taste; genetics; diet; health; children; adults

6.2 Introduction

The prevalence of obesity and associated co-morbidities is rising internationally despite ongoing prevention and intervention efforts [229, 230]. Therefore, new strategies are warranted to promote the development of effective obesity prevention initiatives. About half of the risk of developing obesity is

85 heritable [231, 232], thus work characterizing the genetic component of obesity and incorporating this information into obesity prevention efforts may be a key part of the complex solution to this global problem. Excess intake of calories due to poor eating habits has been widely recognized as a major factor in the development of obesity, and these habits are established in the earliest years of life [233]. While the genetic basis of these adverse behaviours is not clear, taste preferences have been shown to vary due in part to genetics and to be associated with poor eating habits [192]. Characterizing the genetic factors that predispose to certain taste preferences may therefore provide a tool to tailor eating patterns to promote healthy eating habits.

The relationship between genetic variation and taste has previously been investigated by examining single nucleotide polymorphisms (SNPs) with outcomes of sensory tests. Studies have focused on the link between taste receptor gene SNPs and measures of taste sensitivity, taste preference, and dietary intake [72, 111-113, 155, 158, 176, 191, 220, 234]. However, previous studies typically analyzed very few SNPs and only measured sensitivity, preference or dietary intake related to one type of taste.

This narrow focus to date is limiting and may not provide a complete and accurate understanding regarding the complex relationship between taste and eating patterns. In this study, 93 SNPs spanning taste receptor genes that elicit fat, sweet, salt, sour, umami, and bitter tastes were examined for their associations with measures of taste sensitivity and taste preference. SNPs determined to be significantly associated with taste were then examined for potential associations with dietary intake in children. As a result of this comprehensive analysis, SNPs that are associated with taste perception can subsequently be assessed for their effect on the intake of dietary components related to that specific taste.

6.3 Methods

6.3.1 Participants

Forty-nine families, including 72 children and 81 adults, were recruited from the Guelph Family

Health Study. Exclusion criteria included smoking, diagnosis of hypogeusia or ageusia, and having

86 undergone bariatric surgery. This study was approved by the Research Ethics Board at the University of

Guelph (REB#16-12-629).

6.3.2 Anthropometry, Body Composition and Blood Pressure Measurements

Parents and their children arrived at the University of Guelph in the Body Composition Lab having fasted for at least two hours. Among both parents and children, height was measured to the nearest

0.1cm using a wall-mounted stadiometer (Medical Scales and Measuring Devices; Seca Corp., Ontario,

CA). Body weight was measured while wearing tight-fitting clothing and no shoes using the BOD POD™ digital scale (Cosmed Inc., Concord CA, USA). Body mass index (kg/m2) was calculated from the weight and height measurements. Waist circumference was measured using a Gulick II measuring tape (Country

Technology Inc., Gay Mills, WI, USA) at the top of the iliac crest. The BOD POD™ was used to determine body composition of adult participants using air displacement plethysmography. Fat mass % in children was determined using bioelectric impedance analysis. Trained research assistants used the

Quantum IV – Body Composition AnalyzerTM (RJL Systems, Clinton Township, MI) using single- frequency, with electrodes placed on the right hand and foot. Total body water (TBW) was determined using the Kushner equation [235], then TBW was divided by an age- and sex-specific hydration factor to obtain fat mass %. Among both parents and children, blood pressure and heart rate were measured from the right brachial artery using an automated oscillometric device (HBP-1300 OMRON, Mississauga,

Ontario). Cuff size was determined based on arm circumference. Among adults, two rested measurements of blood pressure (systolic and diastolic) and heart rate were obtained via an automatic reading while participants were seated in an upright position. The average of the two measurements for each participant was used in subsequent analyses. Three measurements were taken for children whereby the average of the last two measurements was used.

6.3.3 SNP Selection and Genotyping

A PubMed SNP search was conducted for the following genes previously implicated in taste detection: CD36, GPR120, GPR40, TAS1R1, TAS1R2, TAS1R3, TAS2R38, ENaC, TRPV1, GRM4, and

87 KCNJ2. The resulting SNPs from each gene were filtered by global minor allele frequency (MAF), and

Saliva was collected at the health assessment using the Oragene•DNA (OG-575) collection kit for

Applied Genomics at The Hospital for Sick Children where they underwent genotyping using the Agena

MassArray System.

6.3.4 Psychophysical Measurements

Psychophysical tests for adults were administered in sensory booths at the University of Guelph

Sensory Laboratory after a 2-hour fast (n=65). Filter paper strips (Indigo Instruments – Cat#33814-Ctl;

47mm x 6 mm x 0.3 mm) immersed in varying concentrations of tastants were used to determine suprathreshold sensitivity (ST) for the adults only. The tastants were: sucrose for sweet taste (Thermo

Fisher; S5-500), monosodium glutamate (MSG) (Thermo Fisher; ICN10180080) and inosine monophosphate (IMP) (Thermo Fisher; AC226260250) for umami taste, sodium chloride (NaCl) for salt taste (Thermo Fisher; S641-500), citric acid for sour taste (A940-500), oleic acid for fat taste (A195-500)

(Thermo Fisher Scientific), and PTC for bitter taste (Indigo Instruments – Cat#33814-PTC). Oleic acid was homogenized in deionized water prior to immersing the filter paper, and all other tastants were dissolved in water at ambient temperature. Filter paper strips were immersed in the tastant solution for about one second before placing them on a drying rack to dry overnight at ambient temperature. This procedure was performed only once for all strips before the study commenced. Taste strips immersed in a solution with the same tastant and concentration were stored together at 4°C in a small plastic re-sealable bag. Each time a strip was tested, participants placed the taste strip in the middle of their tongue, closed their mouths, and allowed at least five seconds for the tastants to be sensed by taste receptors. Participants

88 were asked to rinse and expectorate with distilled water before beginning and following each strip. Within each taste modality, the range of tastant concentrations tested is shown in Table 6.1. Oral ST was determined using filter paper strips for a range of tastant concentrations by computing the area-under-the- curve (AUC) of intensity ratings on the general labeled magnitude scale (gLMS) [159], and preference

(PR) was measured using a forced-choice paired comparison of hummus samples. Hummus was selected as a vehicle in this study due to its neutral taste and due to its suitability to added sweetness, salt, sourness, fat, and savouriness. Participants were presented with a range of taste strips in random order and were asked to rate the intensity of the strips from 0-100 on a gLMS where 0=undetectable, 2=barely detectable, 6=weak, 18=moderate, 35=strong, 52=very strong, and 100=strongest imaginable sensation of any kind. For bitter taste, only one rating of PTC intensity was obtained.

In the PR test for adults, paired hummus samples labeled with random three-digit codes were presented simultaneously to participants in a small translucent sample cup. Each pair of hummus samples looked identical and consisted of one sample with a standard study formulation and the other with an added ingredient to more strongly elicit a specific taste modality. The standard study hummus was formulated at the University of Guelph Formulation Laboratory. First, chickpeas (540mL – ARZ Fine

Foods) were rinsed in a strainer with cold water and poured into the three-quart polycarbonate bowl of the

Robot Coupe Food Processor (Model# R2NCLR). Distilled water (92 mL - President’s Choice), olive+canola oil mix (54 mL - Pur Oliva), lemon juice (10 mL - ReaLemon), tahini (35 mL - ARZ), and salt (Thermo Fisher; S641-500) were then added to the chickpeas. The mix was processed for 40 seconds, mixed with a spoon to allow chunks of chickpeas on the sides of the processor bowl to be re-incorporated, and processed again for 60 seconds. Five 150g quantities of hummus were then set aside for the preparation of hummus samples with added ingredients. To elicit stronger fat, salt, sour, sweet, and umami taste, olive + canola oil mix (15g), salt (0.5g), lemon juice (7g), sucrose (4g), and MSG (4g) were respectively added to a 150g quantity of the standard study hummus and mixed thoroughly with a spoon.

For each participant, ten sample cups containing five standard hummus samples as well as five hummus samples with added ingredients were prepared (8g each). A random number generator was used by a

89 research assistant to produce the three-digit codes with which to label the sample cups such that the sensory test administrator was blinded to the hummus formulations. In the PR test, each sample was tasted using a metal spoon following an oral rinse with distilled water. After the second hummus sample was tasted, participants were asked “Which of the two hummus samples did you prefer?” and responded by providing the sensory test administrator with the three-digit code of the preferred sample. Oral ST and

PR for all taste modalities were measured during the same study visit.

Children participated in a PR test and a PTC taster test only, following a 2-hour fast (n=60).

While the hummus formulations in the PR test were identical to the test with the adults, the forced-choice paired-comparison method was adapted for young children to ensure that the tasks of the procedure would be understood. Once the children provided verbal assent to participate, they joined the test administrator alone in a conference room that was void of any potential distractions. To confirm that the children understood the test, a visual forced-choice paired-comparison task was performed [236]. Using hair elastic bands of various colours and two containers labeled with either a happy face or a sad face, the children were asked to choose a “favourite colour” and report this color to the test administrator. The children were then presented with two bands, one of which was their favourite colour and the other was a different colour. The children were then instructed to choose their favourite hair band and place it inside the container labeled with a happy face. If the child placed the hair band with their favourite colour into the appropriate container, then they were deemed capable of performing the preference test with the hummus samples. When choosing a preferred hummus sample, the children simply had to point to their preferred sample and the three-digit code of this sample was recorded by the test administrator. Instead of providing an intensity rating on the gLMS for the strip of PTC paper, the children participated in a yes-no task to determine PTC taster status. The children responded with a “yes” or a “no” to the question “Does that taste bad or have no taste at all?”. If the children reported a bad taste, they were recorded as “PTC tasters” whereas children who reported no taste were recorded as “non-tasters”.

90 6.3.5 Dietary Intake of Children

Parents completed a three-day food record for their children, including two weekdays and one weekend day. Parents documented a detailed description of each food or beverage (i.e., cooking method, brand name) and the consumed. Food records were inputted into a nutrient analysis program (ESHA Food

Processor, Version 11.0.110, Salem, Oregon, USA). Calories from sugar, added sugar, total carbohydrates, fat, and protein were computed from an average of three days. Energy density of the whole diet as well as the relative contributions of energy density of sugar, added sugar, total carbohydrates, fat, and protein were also computed.

Table 6.1: Range of tastant concentrations used for each psychophysical test

Taste modality (stimulus) Threshold/Suprathreshold (mM) Preference (mM)

Sweet (sucrose) 2.5-500 6%-36% (w/v*)

Umami (MSG) 3.13-200 3.13-200

Umami (IMP) 0.313-20 0.313-20

Umami (MSG+IMP) 3.13-200 MSG + 0.5 IMP 3.13-200 MSG + 0.5 IMP

Salt (sodium chloride) 5-100 50-250

Sour (citric acid) 1-15 10-200

Fat (oleic acid) 30-100 50-100

Bitter (PTC) 3 μg/strip - Tastants were diluted in distilled water and filter papers were submerged in the solutions. *weight/volume

6.3.6 Statistics

R Statistical Software Version 3.4.0 (R Foundation for Statistical Computing) was used to calculate regression coefficients for linear models of psychophysical measures of taste and SNPs using generalized estimating equations (GEE). Secondly, SNPs significantly associated with a psychophysical measure of taste were then further analyzed using GEE to estimate the regression coefficients for logistic models of SNPs and trait pairs including one taste variable and one diet variable. GEEs were also used to

91 estimate the regression coefficients for linear models to examine the associations between diet variables and covariates including age, sex, and BMI due to the potential moderating effect of BMI on taste perception [74, 76, 158-160]. SNPs initially found to be significantly associated with a taste preference in children, prior to the Bonferroni adjustment, were subsequently assessed for associations with dietary intake using logistic regression (Table 6.6). Therefore, taste variables were only paired with diet variables for analysis whereby the nutrient elicits that type of taste. For example, SNPs significantly associated with sweet taste preference would only further investigated for associations with added sugar intake. As sour taste is not typically associated with sensing nutrients, it was paired with 1) percent energy from added sugar as sourness often accompanies sweetness in children’s candies, and 2) total energy density of diet to examine any potential global effects of sour taste preference on the diet. Analyses for both parents and children account for correlated outcomes resulting from multiple siblings within some families and from sharing the same household. Regressions were only performed for SNPs located in a gene associated with the same taste modality as the taste outcome. Statistical significance was set to p≤0.05.

6.4 Results

6.4.1 Participant Characteristics

72 children and 81 adults from 49 families were recruited for the study, 60 children and 65 adults from 41 families completed the study. Eight families did not complete the study due to discontinued communication with the research personnel following recruitment. Adult participant characteristics are summarized in Table 2 and child participant characteristics are summarized in Table 3. Mothers (n=41) and fathers (n=24) had a mean age of 36.3 ± 4.3 years while boys (n=27) and girls (n=33) had a mean age of 4.1 ± 1.2 years. The mean BMI of adults (27.1 ± 5.6 kg/m2) reflect an overweight status of parents and the mean BMI z-score (0.30 ± 0.99) indicated normal weight among children.

92 Table 6.2: Adult participant characteristics in total and separated by sex

Characteristic Total Female Male n 65 41 24 Age (years) 36.3 (4.3) 35.8 (4.5) 37.2 (4.0) Systolic Blood Pressure (mmHg) 118.4 (19.1) 113.9 (9.9) 130.1 (12.5) Diastolic Blood Pressure (mmHg) 72.9 (12.4) 70.3 (8.0) 79.6 (8.8) Heart rate (beats/min) 69.8 (7.9) 69.9 (9.6) 69.8 (7.9) BMI (kg/m2) 27.1 (5.6) 26.4 (5.6) 28.1 (4.9) % Body Fat - 34.3 (8.9) 26.8 (9.0) Ethnicity (%) Caucasian 85 - - Other 15 - - Means (SD) were computed for all characteristics except for ethnicities, which are presented as percents.

Table 6.3: Child participant characteristics

Characteristic Total N 60 Female 33 Male 27 Age (years) 4.1 (1.2) Systolic Blood Pressure (mmHg) 102.5 (13.6) Diastolic Blood Pressure (mmHg) 56.9 (11.2) Heart rate (beats/min) 91.0 (12.4) BMI z-score 0.30 (0.99) % Body Fat 29.3 (6.1) Ethnicity (%) Caucasian 81 Other 19 Means (SD) were computed for all characteristics except for sex and ethnicity, which are presented as frequencies and percents, respectively. Sample size for each characteristic may vary due to incomplete information from 6 children.

6.4.2 Genetics and Taste Function/Preference

In total, 93 tSNPs were genotyped from eleven taste receptor genes in both children and adults.

Twenty tSNPs were genotyped from fat taste receptor genes (CD36, GPR120, and GPR40), eleven tSNPs were genotyped from sweet taste receptor genes (TAS1R2 and TAS1R3), one tSNP was genotyped from a bitter taste receptor gene (TAS2R38), twenty tSNPs were genotyped from salt taste receptor genes (ENaC

93 and TRPV1), twenty-nine tSNPs were genotyped from umami taste receptor genes (TAS1R1, TAS1R3, and mGluR4), and twelve tSNPs were genotyped from sour taste receptor genes (ASIC1 and KCNJ2).

As summarized in Table 6.4, twenty-three tSNPs were associated with a taste outcome in parents before applying a statistical correction for multiple hypotheses. Following a Bonferroni adjustment for multiple hypothesis testing, the rs713598 and rs236514 SNPs remained significantly associated with taste outcomes in parents. The C allele of the rs173598 SNP in the TAS2R38 bitter taste receptor gene was significantly associated with PTC sensitivity. The A allele of the rs236514 SNP in the KCNJ2 sour taste receptor gene was significantly associated with sour preference. As summarized in Table 6.5, eleven tSNPs were associated with a taste outcome in children before applying a statistical correction for multiple hypotheses. Two tSNPs remained significantly associated with a taste outcome in children after applying a Bonferroni adjustment. The C allele of the rs4790522 tSNP in the TRPV1 salt taste receptor gene was associated with a significantly higher salt preference compared to the A allele in children. The T allele of the rs173135 tSNP in the KCNJ2 sour taste receptor gene was associated with a significantly higher sour preference compared to the C allele in children. In both parents and children, the C allele of the rs236512 SNP from KCNJ2 was associated with sour preference. In parents, the A allele of the rs150908 SNP in TRPV1 was associated with both higher salt taste sensitivity and a lower preference for salt.

6.4.3 Multiple Trait Analysis: SNPs, Taste and Dietary Intake

The rs9701796 SNP in the TAS1R2 sweet taste receptor gene was associated with both sweet taste preference (p=0.02) and percent energy from added sugar in the diet (p=0.05). The rs9701796 SNP was also significantly related to sweet taste preference when included in a model with total energy density of diet (p=0.05), however total energy density of diet was not statistically significant in the model. While the rs173135 SNP in the KCNJ2 sour taste receptor gene was not significantly associated with sour taste preference, this SNP was significantly associated with total energy density of diet with sour taste preference included in the model (p=0.03).

94 Table 6.4: Associations between SNPs in taste receptor genes and suprathreshold sensitivity and taste preference in adults

SNP ID (gene) Taste Modality Outcome P-value

rs12137730 (TAS1R2) Sweet Suprathreshold 0.021 rs2499729 (GRM4) 0.031 rs3778045 (GRM4) Suprathreshold 0.007 rs4908563 (TAS1R1) 0.022 rs11759763 (GRM4) 0.007 rs2451328 (GRM4) 0.021 rs2451361 (GRM4) Umami 0.012 rs2499682 (GRM4) 0.020 Preference rs2499729 (GRM4) 0.036 rs7772932 (GRM4) 0.015 rs937039 (GRM4) 0.046 rs9380406 (GRM4) 0.007 rs150908 (TRPV1) 0.043 rs161386 (TRPV1) Suprathreshold 0.045 Salt rs222745 (TRPV1) 0.036 rs150908 (TRPV1) Preference 0.036 rs2301151 (GPR40) 0.016 Fat Suprathreshold rs3211816 (CD36) 0.014 rs713598 (TAS2R38) Bitter Suprathreshold 0.003* rs236512 (KCNJ2) 0.041 rs236514 (KCNJ2) 0.002* Sour Preference rs376184 (ASIC1) 0.019 rs643637 (KCNJ2) 0.011 Generalized estimating equations were used to estimate the regression coefficients of a linear model including suprathreshold sensitivity and taste preference with SNPs (n=65). Regressions were only performed for SNPs located in a gene associated with the same taste modality as the taste outcome. *Following a Bonferroni adjustment for multiple hypothesis testing, the rs713598 and rs236514 SNPs remained significantly associated with phenylthiocarbamide suprathreshold and sour preference, respectively. The Bonferroni adjustment of the reported p-values accounted for the number of hypotheses equal to the number of SNPs in genes associated with each taste modality.

95 Table 6.5: Associations between SNPs in taste receptor genes and taste preference in children

SNP ID (gene) Taste Modality P-value

rs7534618 (TAS1R2) 0.026 Sweet rs9701796 (TAS1R2) 0.013 rs4713740 (mGluR4) Umami 0.039 rs4790151 (TRPV1) 0.008 rs4790522 (TRPV1) Salt 0.001* rs877610 (TRPV1) 0.010

rs17108968 (GPR120) Fat 0.029

rs173135 (KCNJ2) < 0.001* rs236512 (KCNJ2) 0.007 Sour rs236513 (KCNJ2) 0.006 rs9890133 (KCNJ2) 0.006 Generalized estimating equations were used to estimate the regression coefficients of a linear model including taste preference with SNPs (n=60). Regressions were only performed for SNPs located in a gene associated with the same taste modality as the taste outcome. *The rs4790522 (TAS1R2) and rs173135 (KCNJ2) SNPs remained significant following a Bonferroni adjustment for multiple hypothesis testing. The Bonferroni adjustment of the reported p-values accounted for the number of hypotheses equal to the number of SNPs in genes associated with each taste modality.

6.5 Discussion

This study examined associations between a comprehensive panel of SNPs in taste receptor genes and psychophysical measures of taste across all known taste modalities in both parents and their children.

Overall, the findings in this study showed that SNPs in taste receptor genes from all of the different types of taste may contribute to inter-individual differences in psychophysical measures of taste. However, only four SNPs (rs173135, rs236514, rs4790522 and rs713598) were found to be significantly related to a taste outcome after applying a statistical correction for multiple hypothesis testing.

96 Table 6.6: Multiple trait analysis of SNPs in taste receptor genes, taste preferences and dietary intake in children

Taste P-value SNP (gene) Dietary Outcome Modality Taste Preference Diet Total energy density (kcal/g) 0.09 0.46 rs17108968 Fat Energy from fat (kcal) 0.10 0.69 (GPR120) % Energy from fat 0.09 0.65 rs4790151 (TRPV1) 0.92 0.30 rs4790522 (TRPV1) Salt Sodium (mg) 0.29 0.44 rs877610 (TRPV1) 0.58 0.71 Total energy density (kcal/g) 0.20 0.03* rs173135 (KCNJ2) % Energy from added sugar 0.39 0.49 Total energy density (kcal/g) 0.64 0.36 rs236512 (KCNJ2) % Energy from added sugar 0.80 0.35 Sour Total energy density (kcal/g) 0.34 0.11 rs236513 (KCNJ2) % Energy from added sugar 0.55 0.78 Total energy density (kcal/g) 0.34 0.11 rs9890133 (KCNJ2) % Energy from added sugar 0.55 0.78 % Energy from added sugar 0.47 0.11 rs7534618 (TAS1R2) Total energy density (kcal/g) 0.32 0.39 Sweet % Energy from added sugar 0.02* 0.05* rs9701796 (TAS1R2) Total energy density (kcal/g) 0.05* 0.98 Total energy density (kcal/g) 0.37 0.59 rs4713740 (GRM4) Umami % Energy from protein 0.37 0.99

Generalized estimating equations were used to estimate the regression coefficients of a logistic model including SNPs with trait pairs including a taste preference variable and a diet variable (n=60). Regressions were only performed for SNPs determined to be significantly associated with taste preferences in the initial linear regressions. Taste variables were generally only paired with specific diet variables whereby the nutrient elicits that type of taste. An adjustment for multiple testing was not applied to these p-values due to the small number of tests conducted for each SNP. *p<0.05

97 The rs4790522 SNP, located in the salt taste receptor gene TRPV1, was found to be significantly associated with preference for salt in children. Regulation of salt intake, or sodium, is due in part to variation in genes related to homeostatic sodium regulation [178-181] and to hedonic responses to the taste of salt [182]. Sodium intake is important to monitor due to its role in the development of hypertension, a risk factor for the development of cardiovascular disease [183-186]. The rs4790522 SNP has previously been shown to change the miRNA binding site of TRPV1, suggesting that this SNP may affect the stability of the mRNA precursor to TRPV1 and prevent translation into its functional protein

[237]. The potential decreased functionality of TRPV1 may reduce salt taste sensitivity and therefore increase the preference of salt in carriers of this SNP. To the authors’ knowledge, no associations have previously been found between the rs4790522 SNP and salt taste. Future studies should also consider examining the rs150908 SNP which exhibited significant associations with both salt sensitivity and salt preference in parents, increasing the potential relevance of this variant for salt taste. Studies with larger sample sizes are warranted to replicate these results in order to better understand the genetic basis for salt sensitivity, and therefore hypertension.

Sour taste is elicited by acidic substances through the depolarization of type III taste bud cells

[35]. While sourness is conventionally considered a means to avoid the consumption of spoiled foods, many animals find mildly acidic foods to be palatable. Moreover, genetic factors may be more important than shared environment to determine the pleasantness and intensity of sour taste as 34-50% of the variation in pleasantness and use-frequency of sour foods is attributable to genetics [193]. With the knowledge that there is a genetic basis for the preference for sour foods in humans, Ye et al. (2016) proposed that sour taste is mediated by the potassium ion channel KIR2.1, encoded by the KCNJ2 gene

[150]. The rs173135 and rs236514 SNPs in KCNJ2 were found to be associated with the preference for sour in children and parents, respectively. Moreover, the rs236512 SNP was associated with sour preference in both children and adults. Observing associations with sour preference in two different cohorts suggests that this association may pertain to changes in sour taste function. The genetic basis of human sour taste has not previously been explored through examining KCNJ2 SNPs. These novel

98 findings provide a foundation for future studies to investigate the genetic basis of sour taste as well as sour food intake.

Variants in TAS1R2 and TAS1R3 sweet taste receptor genes have previously been associated with changes in taste sensitivity to sugar [58, 101, 102, 105-109], the excessive consumption of which is an established risk factor for obesity and chronic disease [187-189]. Previous research has implicated SNPs in TAS1R2 and TAS1R3 in inter-individual differences in sugar sensitivity [113, 158] and dietary intake

[112, 158, 190-192]. However, this study is the first to find an association between a SNP in a sweet tasting gene with both sucrose preference and dietary sucrose intake. In an analysis of SNPs together with taste and diet, it was found that the rs9701796 SNP in the sweet taste receptor gene TAS1R2 was both associated with sweet taste preference and percent energy from added sugar in the children. In a previous study in children and adolescents, rs9701796 was associated with increased waist-height ratio as well as with a higher chocolate powder intake in obese children [234]. In another study of children aged 7-12, rs9701796 was not associated with dental caries, a marker often related to excessive sweet food consumption [115]. More research pertaining to this variant is warranted, particularly to assess its relationship with the consumption of sweet foods. By establishing these types of associations in future studies, genetic loci can be considered risk factors for the overconsumption of sweet foods and be used clinically to indicate the risk of developing obesity and other chronic diseases.

The bitterness of green leafy vegetables including Brassica vegetables is related to the taste of thiol compounds and may be stronger in those homozygous for the C allele at the rs713598 locus in the

TAS2R38 taste receptor gene. Non-carriers of the C allele may not taste PTC, and this may then influence the perceived bitterness of Brassica vegetables [124]. While parents in this study exhibited a strong association between rs713598 genotype and PTC tasting, no relationship was observed between rs713598 genotype and PTC taster status in children [230, 231]. Children would be expected to show a stronger genotype-phenotype relationship due to having less exposure to culture at their age; however, the lack of association in this study is likely an indicator of the poor reliability of measuring PTC taste sensitivity in this age group. Children between 3-8 years of age may not have an adequately developed understanding

99 of the quality of bitterness. While the study personnel administered a simple yes-no task to determine

PTC taster status in the children, this task may still have been too complex due to the unusual taste and paper format the stimulus.

There are some limitations to consider in this study. Firstly, the data obtained by assessing taste sensitivity in parents, using isolated compounds on filter paper strips, cannot be used to make direct associations between genetics and food intake. However, this can also be considered a strength of the study as the observations made are accurate for specific taste modalities whereas the use of hummus as a food matrix in this study may have introduced uncertainty due to the perception of texture, temperature, and other matrix-specific qualities. The matrix-specific qualities of hummus may have affected the measurement of taste preference, however the use of a food as a stimulus increases the relevance of these results to food preferences and food selection. In addition, participants were tested for sensitivity and preference on only one occasion, but this should be repeated to confirm validity. While this study was powered to observe differences in sensory outcomes, the sample size was small and the likelihood of making type II errors would be lower with a larger sample. Finally, the genetic heterogeneity due to the presence of more than one ethnicity in this sample may hinder the interpretation of the results as the minor allele frequencies of SNPs differ depending on the population. However, the statistical methods used in this work account for correlated outcomes as parents share a household and siblings share household and genetics.

6.6 Conclusion

This study demonstrated that SNPs in taste receptor genes may contribute to inter-individual differences in taste sensitivity, taste preference and dietary intake. These findings, based on a comprehensive panel of genetic variants in adults and young children, support the relevance of genetics in explaining variation in taste function. The genetic determinants of taste function are important to understand as they may predispose individuals to developing poor eating patterns. In the future, effective strategies can be developed to improve eating habits and therefore risk of obesity through personalized nutritional recommendations based on unique taste preferences.

100 6.7 Acknowledgements

E.C. conceived and designed the experiments, collected and analyzed data, and wrote the paper; N.A.C. helped with data collection and analysis; L.D. helped to design the sensory experiments and provided critical revision for the paper; Z.F. helped with the statistical analysis and provided critical revision for the paper; W.Q. helped with statistical analysis; G.D. helped with the statistical analysis and provided critical revision for the paper; A.M.D. helped with the dietary data and provided critical revision for the paper; J.H. provided critical revision for the paper; D.W.L.M. conceived and designed the experiments and provided critical revision for the paper. Support was provided by the Ontario Ministry of Agriculture, Food, and Rural Affairs (grant# 110-029900-000000-030194)

101 CHAPTER 7:

INTEGRATIVE DISCUSSION

102 7.1 Preamble

This thesis aimed to explore associations between SNPs in taste receptor genes, psychophysical measures of taste and dietary intake. Three specific objectives were explored: 1) to assess the relationship between SNPs in taste receptor genes and psychophysical measures of taste in adults (Chapter 3); 2) to determine the relationship between taste sensitivity and taste preference (Chapter 4); and 3) to assess the relationship between SNPs in taste receptor genes, psychophysical measures of taste, and dietary intake in children and adults (Chapters 5 and 6). The following is a discussion of the major findings of this thesis, suggestions for studies moving forward as well as overall conclusions.

7.2 Summary and discussion of major findings

7.2.1 Sweet taste

Genetic variation in TAS1R2 and TAS1R3 sweet taste receptor genes are important to characterize because they pertain to changes in taste sensitivity to sugar [58, 101, 102, 105-109], the excessive consumption of which is an established risk factor for obesity and chronic disease [187-189]. Previous research has implicated SNPs in TAS1R2 and TAS1R3 in inter-individual differences in sugar sensitivity

[113, 158] and dietary intake [112, 158, 190-192]. However, a study in this thesis was the first to find an association between a SNP in a sweet tasting gene with both sucrose preference and dietary sucrose intake. An analysis of SNPs together with taste and diet in children found that the CC genotype of the rs9701796 SNP in the sweet taste receptor gene TAS1R2 was both associated with higher sweet taste preference and percent energy from added sugar. In a previous study in children and adolescents, rs9701796 was associated with increased waist-height ratio as well as with a higher chocolate powder intake in obese children [234]. In another study of children aged 7-12, rs9701796 was not associated with dental caries, a marker often related to excessive sweet food consumption [115]. More research pertaining to this variant is warranted, particularly to assess its relationship with the consumption of sweet foods.

While it is important to conduct studies pertaining to sweet taste sensitivity, the examination of sucrose

103 preference and TAS1R2 and TAS1R3 SNPs in this study offers a more plausible link between genetics and food selection.

Ile191Val variant with habitual sugar consumption in overweight/obese adults and dental caries in children [53, 111], this finding demonstrates that taste preferences and dietary intake may be related in children. Upon further analysis, it was found that compared to children with other genotypes, children with the TT genotype consumed the most sugary snacks in the evening. If the evening is a time of day when children with a “sweet tooth” potentially consume more sugary snacks, this finding warrants future studies to assess the availability and accessibility of sugary foods in the home environment as a means to address the role of parents in the intake of sugary foods by young children during the evening. Future studies would also benefit from including genetic variants in sweet taste receptor genes to examine determinants of sweet food intake throughout the day [113, 218].

In adults, the rs4920564 SNP was associated with sweet taste preference. Particularly, the G allele of the rs4920564 tSNP in the TAS1R2 sweet taste receptor gene was associated with a higher PR for sucrose compared to the T allele, but not significantly. While diet records were not obtained in the same study where this association was observed, and the sensory stimuli were not food, more research pertaining to this variant is warranted. In particular, this variant should be examined for a potential association with the consumption of sweet foods. By establishing these types of associations in future

104 studies, genetic loci can be considered risk factors for the overconsumption of sweet foods and be used practically in a health care setting to help indicate the risk of developing obesity.

7.2.2 Fat taste

Adults with the AA genotype of the rs1761667 variant (CD36) have been shown to have lower oral sensitivity to fat and higher preference for fat [72, 88]. Chapter 5 of this thesis found no relationship between carrying the AA genotype and snack quantity, frequency, or calories from fat in children; however, children carrying the AA genotype consumed higher energy density of snacks than children carrying the G allele at this locus. While studies have shown differences in sensory outcomes based on this CD36 SNP, an effect on habitual dietary intake has never been observed. The present finding suggests that a fat taste receptor SNP genotype may influence the intake of energy-dense snacks. This finding is potentially due to the decreased sensitivity to fat taste and subsequent higher preference for fatty foods; however, measures of taste sensitivity were not determined in the Chapter 5 study. Although the relationship between fat taste sensitivity and energy density of snacks in this study is consistent with the energy density of fat as a macronutrient, no difference was found between the calories from fat in snacks for this genotype. This apparent contradiction may be due to the challenges of measuring the contribution of fat in a complex diet, which contains foods simultaneously high in fat and carbohydrates.

Future research studies would benefit from performing hedonic tests for fat taste with children in more controlled settings in addition to assessing dietary intake.

The G allele of the rs2301151 (GPR40) SNP and the G allele of the rs3211816 (CD36) SNP were associated with fat taste sensitivity in adults and the G allele of the rs17108968 SNP (GPR120) was associated with fat taste preference in children; however, these variants did not remain statistically significant following a Bonferroni adjustment. These variants may be of interest for future studies seeking to determine the genetic basis for fat taste preference. Psychophysical measures of fat taste are the most difficult to measure in isolation of food as fat is typically combined with another type of taste when consumed. This problem was partially addressed in this thesis by using hummus as a matrix for the added

105 fat. While the use of a food-based matrix adds uncertainty to the results due to the presence of its own sensory features (temperature, mouthfeel, texture, etc.), this method is more representative of a preference for fat in foods. Therefore, future studies are encouraged to use this technique to investigate the link between genetic variation in fat taste receptor genes and fat taste preference. A genetic predisposition to prefer fatty foods due to lower taste sensitivity may be of concern in the long term as high fat diets may lead to metabolic complications such as obesity, MetS, and CVD [225].

7.2.3 Salt taste

The genetic predisposition to salt sensitivity has been attributed to variation in genes related to homeostatic sodium regulation [180-183] and to genes which may alter hedonic responses to the taste of salt [184]. Due to its potential impact on the susceptibility to hypertension, a risk factor for the development of cardiovascular disease, sensitivity to the dietary intake of sodium is important to characterize [185-188]. Work from this thesis demonstrated novel findings and repeated previously reported associations between SNPs in the TRPV1 salt taste receptor gene and salt taste sensitivity and preference. In adults, rs4790151 and rs8065080 emerged as potentially important SNPs to examine further with salt taste sensitivity, preference and sodium intake. The A allele of the rs4790151 tSNP in the

TRPV1 salt taste receptor gene was associated with a significantly higher DT for sodium chloride compared to the G allele. No associations have previously been found between the rs4790151 SNP in

TRPV1 and salt taste, whereas the relationship between the rs8065080 SNP in the TRPV1 salt taste receptor gene and suprathreshold sensitivity to salt in this study is consistent with previous research

[155]. Future studies should also consider examining the rs150908 SNP which exhibited significant associations with both salt sensitivity and salt preference in adults, whereby the G allele was associated with lower sensitivity and higher preference for salt, increasing the potential relevance of this variant for salt taste.

The rs4790522 SNP, located in the salt taste receptor gene TRPV1, was found to be significantly associated with the preference for salt in children. The C allele of the rs4790522 tSNP in the TRPV1 salt

106 taste receptor gene was associated with a significantly higher salt preference compared to the A allele in children. The rs4790522 SNP has previously been shown to change the miRNA binding site of TRPV1, suggesting that this SNP may affect the stability of the mRNA precursor to TRPV1 and prevent translation into its functional protein [237]. To the authors’ knowledge, no associations have previously been found between the rs4790522 SNP and salt taste. The potential decreased functionality of TRPV1 may reduce salt taste sensitivity and therefore increase the preference of salt in carriers of the rs4790522

C allele. Studies are warranted to replicate these results in order to better understand the genetic basis for salt sensitivity and hypertension.

7.2.4 Umami taste

Umami taste allows for the oral detection of protein-rich foods through the sensation of amino acids, most typically glutamate (reviewed in Chapter 1), and taste sensitivity to umami is related to an increased liking of dietary protein and the nutritional need of protein in high-protein consumers [175].

Sensitivity to umami taste is also associated with other sensory and health-related outcomes such as salivary secretion [170], taste enhancement [171], increasing appetite and satiety [172, 173], reducing fat mass in rats [174], and obesity [159]. Research on the genetic determinants of umami taste sensitivity and preference are limited despite its potential importance in dietary protein intake and health-related outcomes [144, 145, 176]. Similar to previous research, findings of this study demonstrated that SNPs in umami taste receptor genes are associated with umami taste. Before applying a statistical adjustment for multiple testing, 6 SNPs were significantly associated with umami taste sensitivity and 5 SNPs were associated with umami taste preference. The C allele of the rs2499729 tSNP in the mGluR4 umami taste receptor gene was associated with a significantly higher DT for MSG+IMP compared to the T allele. If future studies can confirm that rs2499729 is associated with umami taste sensitivity, these findings should form a basis for exploring how this SNP may also be implicated in energy intake and therefore the risk of obesity [159, 172-175].

107 7.2.5 Sour taste

Acidic substances elicit sour taste through the depolarization of type III taste bud cells [35].

KCNJ2 gene [150]. This work provided the rationale to select KCNJ2 as a candidate gene for analysis of

SNPs and sour taste perception in this thesis. Remarkably, two SNPs in the KCNJ2 gene were associated with at least one sour taste outcome in the multivariate analysis (DT, ST or PR) before applying the adjustment for multiple testing. The rs173135 and rs236514 SNPs in KCNJ2 were found to be associated with the preference for sour in children and parents, respectively. The T allele of the rs173135 tSNP in the

KCNJ2 sour taste receptor gene was associated with a significantly higher sour preference compared to the C allele in children. In both parents and children, the C allele of the rs236512 SNP from KCNJ2 was associated with sour preference. Observing associations with sour preference in two different cohorts suggests that this association may pertain to changes in sour taste function. The genetic basis of human sour taste has not previously been explored through examining KCNJ2 SNPs. These novel findings provide a foundation for future studies to investigate the genetic basis of sour taste as well as sour food intake.

7.2.6 Bitter taste

108 the perceived bitterness of Brassica vegetables [121]. Carriers of the CC genotype in one study of this thesis (Chapter 5) consumed more energy-dense snacks, potentially due to their avoidance of bitter vegetables [7, 226]. In support of this explanation, a trend occurred where CC genotype carriers

(“supertasters”) and CG genotype carriers (medium tasters) both had higher total energy density of snacks compared to GG genotype carriers, and the CC genotype group had a higher total energy density of snacks than the CG genotype group. No relationships were observed between rs713598 genotype and snack quantity, frequency, or calorie composition, in line with previous research [227, 228]. This finding contributes to a growing, controversial body of literature about this genetic locus, showing that this genotype may adversely affect quality of dietary intake [124].

While parents in this study exhibited an association between rs713598 genotype and PTC tasting, no relationship was observed between rs713598 genotype and PTC taster status in children (Chapter 6)

[230, 231]. Children would be expected to show a stronger genotype-phenotype relationship due to having less exposure to culture at their age; however, the lack of association in that study is likely an indicator of the poor reliability of measuring PTC taste sensitivity in this age group. Children between 3-8 years of age may not have an adequately developed understanding of the quality of bitterness. While the study personnel administered a simple yes-no task to determine PTC taster status in the children, this task may still have been too complex due to the unusual taste and paper format the stimulus. Similar to the other variants investigated in this thesis, this SNP may serve as a biomarker for at-risk eating patterns with the potential of developing into adverse metabolic and health outcomes.

7.2.7 The relationship between taste sensitivity and taste preference

This thesis (Chapter 4) provides evidence that measures of taste sensitivity and taste preference for some types of taste are correlated. In particular, higher sensitivities to salt, sweet, and umami taste are associated with decreases in the preferences for these tastes whereas fat and sour preferences were not associated with corresponding taste sensitivities. A linear negative correlation was reported between taste sensitivity and preference for salt and sour in a previous study [207], but that finding was only replicated

109 for salt in this study. The linear positive correlation observed by Bossola et al. (2007) between sweet sensitivity and preference was the opposite trend from what was found in this study [207]. No association was reported in a previous study between umami DT and umami PR [208]; however, a significant negative correlation was observed in this study. A negative correlation between fat taste sensitivity and fat taste preference has been suggested in previous work [209], but this result was not replicated here.

Altogether, there is some inconsistency between results reported in this work compared to previous studies. These differences may be due to the methodologies used and more research is needed to confirm all of these experiments. Nevertheless, these findings advance the current understanding of the role that taste sensitivity plays in taste preference for specific taste modalities.

While correlations were useful to determine the magnitude and direction of associations between

DT, ST, and PR for each taste modality, the PCA provided additional insight into whether similar factors accounted for those correlations. Component 1 of the PCA reflected intensity perception whereby the ST data are clustered at one end of Component 1 and DT data are clustered at the other end of Component 1.

Participants with higher detection thresholds are less sensitive to stimuli and therefore rate stimuli at lower intensities. This pattern follows a logical and expected outcome for the relationship between DT and ST. Therefore, Component 1 may also be an indicator of accuracy in measuring DT and ST in this study. Component 2 emerged as the factor that contributed to determining preference for every type of taste except for sweetness. While measures of sweet sensitivity and preference were observed to have the same direction of correlation as the other taste modalities, the determinant of sweet PR was different than that of the other types of taste. The weaker positive correlation observed between sweet DT and PR was likely due to the lack of Component 2 influence on sweet PR. Component 2 is speculated to be a preference for high concentrations of the compounds tested. It can then be interpreted that, in general, the presence of a high concentration of sucrose in the samples tested elicited a hedonic liking response in the participants. Consistent with the correlation analysis, this interpretation also suggests that high concentrations of all the other tastants elicited an aversive response in the participants.

110 Fat is the most recently identified taste modality and refers to the taste of fatty acids [212].

Previous research has measured fat DT, ST, and PR using a variety of delivery methods such as deionized water, milk, pudding, custard, tomato soup, filter paper strips, edible strips, etc. [214]. While the ideal technique and delivery method for determining fat sensitivity and preference is controversial [214], the strong negative correlation between DT and ST for fat in this study (Chapter 4) increased the validity of the measurement. The sensitivity-preference relationship for fat is opposite from that of all other taste modalities according to Components 1 and 2 from the PCA. While individual correlations did not indicate that fat DT or ST were significantly correlated with PR, it is worthwhile to consider in future research that the hedonic response to fatty acids as a stimulus may have a different mechanism than for other types of taste. It is also possible that the delivery method of fatty acids in this study was not suitable for measurement of preference due to the influence of other properties of fat on preference, such as texture and mouthfeel. Incorporating fatty acids into food matrices may be a more accurate method of measuring fat taste preference, despite the disavantage of variable perceptions of matrix-specific textures by participants.

7.3 Future directions

These studies have important implications for future studies examining associations between

SNPs in taste receptor genes, psychophysical measures of taste, and dietary intake. Firstly, future studies should attempt to assess different types of taste together instead of only one at a time. By using a combined approach, this field of research would advance more quickly and a comprehensive understanding of the relationship between genetics, taste, and dietary intake would develop much sooner.

To be more inclusive of different taste modalities, researchers should choose a panel of SNPs that spans all of the different types of taste whenever possible. The work presented in this thesis offers useful gene targets to focus on for each type of taste and employs a method of analyzing only tag SNPs that allows for analyzing a large number of SNPs with fewer representative SNPs. While accounting for LD between

SNPs, this method was also cost-efficient due to the fact that less primers are purchased with less SNPs.

111 Secondly, future studies should analyze more than one psychophysical measure of taste and do so using different methods. Despite having been refined over several decades, sensory measurement techniques are limited by the subjectivity of responders. While some techniques involve less subjectivity among the responders than others, using different techniques to measure taste sensitivity and taste preference allows researchers to cross-validate the accuracy of the results. Moreover, measuring taste sensitivity together with taste preference allows for the correlation between these two to be evaluated.

The relationship between different psychophysical measures of taste may provide critical insight into how each of them individually are related to food selection and eating patterns. This is especially useful when comparing studies to one another which have used different methods for measuring taste sensitivity and preference.

It is also important for future studies to consider their sample sizes when conducting genetic analyses. While the work in this thesis was often limited by small sample sizes, the hypothesis-generating nature of the work sets a foundation for future research to focus on certain SNPs and reproduce the results with a larger sample. This field of research will not advance if we cannot have sufficient confidence in the results of each individual study. While studies with a larger number of participants are financially costly, currently established large cohorts should consider examining the genetics of taste as a biological determinant of poor eating habits and chronic disease.

Once an advanced understanding of the genetic basis of taste and its influence on eating patterns can be attained, the knowledge created may have important implications for personalized nutrition. These advancements may prove to be particularly beneficial for children due to cultural influences on their taste perception and food preference is more limited than in adults [40-44]. Parents empowered with information about their children’s taste preference and predispositions to poor eating habits can use this knowledge to tailor their feeding practices and promote healthful eating habits in their children. In the meantime, research studies should be conducted to determine if genetics-based nutrition advice is well- received in practice and leads to behavioural changes related to eating habits [156, 157]. Positive results

112 in such nutrigenomic intervention studies would provide necessary rationale for the continued development of knowledge in this field of research.

Currently, many companies are beginning to offer direct-to-consumer genetic tests which purport to provide personalized information on health-related traits. This relatively new trend has revealed that members of the general public have a remarkable interest in their unique genetic makeups. Personalized health advice based on genetics will be an important aspect of healthcare in the future, and the high interest and trust of the public in this type of information reflects positively on the outlook of its use.

Given the importance of the public’s trust in personalized genetic information, it is imperative that companies providing this information reciprocatively ensure that their testing is evidence-based and have their genetic test reports scrutinized in a peer-review process. Without standards of practice for direct-to- consumer genetic tests, individuals interested in learning about their DNA are at risk of being advised falsely on the presence or absence of health-related traits. As research linking genetics to complex traits remains in its infancy, genetic test services must be careful not to oversimplify the genetic determinants of polygenic traits such that the associations reported to the consumer are elaborated unreasonably. Taste preferences represent a relevant example of polygenic traits that are oversimplified in direct-to consumer genetic tests. While this thesis presented some important advancements in understanding associations between genetics and taste, it also outlined that research to date is still in its early stages. Until associations between genetics and taste can be reproduced with high confidence in peer-reviewed research, the confidence of consumers in genetic testing should be respected by omitting reports of taste preferences.

7.4 Conclusions

The modern food environment, especially in developed countries and increasingly in developing countries, is conducive to the habitual selection of highly-palatable foods containing saturated fats and added sugars. As a result, an alarming portion of modern populations are overweight or obese and develop metabolic complications leading to obesity, CVD, T2D and MetS. While factors influencing the

113 food environment are vast and complex, determining ways to empower consumers to make healthier choices remains an important target for research. With the knowledge that poor eating habits are partially attributed to the compelling experience of taste, research into the determinants of taste may support current knowledge as to the etiology of obesity. Studies assessing the role of genetics in taste and eating patterns, such as the work undertaken in this thesis, have demonstrated promising links that may have important implications in the development of eating patterns from a young age. This research has the potential to provide consumers with a unique incentive to alter their own eating patterns, or those of their children, based on genetic predisposition to eating habits. Personalized nutritional advice based on genetics represents the goal for a relatively new and exciting field of research that is certainly worthwhile to pursue.

114 REFERENCES

[1] Turner-McGrievy, G., Tate, D. F., Moore, D., Popkin, B. Taking the bitter with the sweet: relationship of supertasting and sweet preference with metabolic syndrome and dietary intake. J. Food Sci. 2013,78:S336-S42.

[2] Rolls, E. T. Taste, olfactory and food texture reward processing in the brain and the control of appetite. Proc. Nutr. Soc. 2012,71:488-501.

[3] Kral, T. V., Rauh, E. M. Eating behaviors of children in the context of their family environment. Physiol Behav. 2010,100:567-73.

[4] Gibney, M. J., McNulty, B. A., Ryan, M. F., Walsh, M. C. Nutritional phenotype databases and integrated nutrition: from molecules to populations. Adv. Nutr. 2014,5:352S-7S.

[5] Grimm, E. R., Steinle, N. I. Genetics of eating behavior: established and emerging concepts. Nutr. Rev. 2011,69:52-60.

[6] Fay, L. B., German, J. B. Personalizing foods: is genotype necessary? Curr. Opin. Biotechnol. 2008,19:121-8.

[7] Mennella, J. A., Pepino, M. Y., Reed, D. R. Genetic and environmental determinants of bitter perception and sweet preferences. Pediatrics. 2005,115:e216-e22.

[8] D.Chiras, D. Human biology: Jones and Bartlett Learning; 2005.

[9] Schacter, D. Psychology. 2 ed: Worth Publishers; 2009.

[10] Chaudhari, N., Roper, S. D. The cell biology of taste. J. Cell Biol. 2010,190:285-96.

[11] Bartel, D. L., Sullivan, S. L., Lavoie, E. G., Sevigny, J., Finger, T. E. Nucleoside triphosphate diphosphohydrolase-2 is the ecto-ATPase of type I cells in taste buds. J. Comp Neurol. 2006,497:1-12.

[12] Yoshida, R., Horio, N., Murata, Y., Yasumatsu, K., Shigemura, N., Ninomiya, Y. NaCl responsive taste cells in the mouse fungiform taste buds. Neuroscience. 2009,159:795-803.

[13] Vandenbeuch, A., Clapp, T. R., Kinnamon, S. C. Amiloride-sensitive channels in type I fungiform taste cells in mouse. BMC. Neurosci. 2008,9:1.

115 [14] Chandrashekar, J., Kuhn, C., Oka, Y., Yarmolinsky, D. A., Hummler, E., Ryba, N. J., et al. The cells and peripheral representation of sodium taste in mice. Nature. 2010,464:297-301.

[15] Nelson, G., Hoon, M. A., Chandrashekar, J., Zhang, Y., Ryba, N. J., Zuker, C. S. Mammalian sweet taste receptors. Cell. 2001,106:381-90.

[16] Adler, E., Hoon, M. A., Mueller, K. L., Chandrashekar, J., Ryba, N. J., Zuker, C. S. A novel family of mammalian taste receptors. Cell. 2000,100:693-702.

[17] Yarmolinsky, D. A., Zuker, C. S., Ryba, N. J. Common sense about taste: from mammals to insects. Cell. 2009,139:234-44.

[18] DeFazio, R. A., Dvoryanchikov, G., Maruyama, Y., Kim, J. W., Pereira, E., Roper, S. D., et al. Separate populations of receptor cells and presynaptic cells in mouse taste buds. J. Neurosci. 2006,26:3971-80.

[19] Clapp, T. R., Medler, K. F., Damak, S., Margolskee, R. F., Kinnamon, S. C. Mouse taste cells with G protein-coupled taste receptors lack voltage-gated calcium channels and SNAP-25. BMC. Biol. 2006,4:7.

[20] Huang, A. L., Chen, X., Hoon, M. A., Chandrashekar, J., Guo, W., Trankner, D., et al. The cells and logic for mammalian sour taste detection. Nature. 2006,442:934-8.

[21] Lewandowski, B. C., Sukumaran, S. K., Margolskee, R. F., Bachmanov, A. A. Amiloride-Insensitive Salt Taste Is Mediated by Two Populations of Type III Taste Cells with Distinct Transduction Mechanisms. The Journal of neuroscience : the official journal of the Society for Neuroscience. 2016,36:1942-53.

[22] Sullivan, J. M., Borecki, A. A., Oleskevich, S. Stem and progenitor cell compartments within adult mouse taste buds. Eur. J. Neurosci. 2010,31:1549-60.

[23] Simons, P. J., Kummer, J. A., Luiken, J. J., Boon, L. Apical CD36 immunolocalization in human and porcine taste buds from circumvallate and foliate papillae. Acta Histochem. 2011,113:839-43.

[24] Hochheimer, A., Krohn, M., Rudert, K., Riedel, K., Becker, S., Thirion, C., et al. Endogenous gustatory responses and gene expression profile of stably proliferating human taste cells isolated from fungiform papillae. Chem. Senses. 2014,39:359-77.

[25] Finger, T. E., Danilova, V., Barrows, J., Bartel, D. L., Vigers, A. J., Stone, L., et al. ATP signaling is crucial for communication from taste buds to gustatory nerves. Science. 2005,310:1495-9.

116 [26] Huang, Y. J., Maruyama, Y., Dvoryanchikov, G., Pereira, E., Chaudhari, N., Roper, S. D. The role of pannexin 1 hemichannels in ATP release and cell-cell communication in mouse taste buds. Proc. Natl. Acad. Sci. U. S. A. 2007,104:6436-41.

[27] Dando, R., Roper, S. D. Cell-to-cell communication in intact taste buds through ATP signalling from pannexin 1 gap junction hemichannels. J. Physiol. 2009,587:5899-906.

[28] Romanov, R. A., Rogachevskaja, O. A., Bystrova, M. F., Jiang, P., Margolskee, R. F., Kolesnikov, S. S. Afferent neurotransmission mediated by hemichannels in mammalian taste cells. EMBO J. 2007,26:657-67.

[29] Bo, X., Alavi, A., Xiang, Z., Oglesby, I., Ford, A., Burnstock, G. Localization of ATP-gated P2X2 and P2X3 receptor immunoreactive nerves in rat taste buds. Neuroreport. 1999,10:1107-11.

[30] Yang, R., Montoya, A., Bond, A., Walton, J., Kinnamon, J. C. Immunocytochemical analysis of P2X2 in rat circumvallate taste buds. BMC. Neurosci. 2012,13:51.

[31] Dando, R., Roper, S. D. Acetylcholine is released from taste cells, enhancing taste signalling. J. Physiol. 2012,590:3009-17.

[32] Ogura, T. Acetylcholine increases intracellular Ca2+ in taste cells via activation of muscarinic receptors. J. Neurophysiol. 2002,87:2643-9.

[33] Dvoryanchikov, G., Huang, Y. A., Barro-Soria, R., Chaudhari, N., Roper, S. D. GABA, its receptors, and GABAergic inhibition in mouse taste buds. J. Neurosci. 2011,31:5782-91.

[34] Huang, Y. A., Maruyama, Y., Roper, S. D. Norepinephrine is coreleased with serotonin in mouse taste buds. J. Neurosci. 2008,28:13088-93.

[35] Huang, Y. A., Maruyama, Y., Stimac, R., Roper, S. D. Presynaptic (Type III) cells in mouse taste buds sense sour (acid) taste. J. Physiol. 2008,586:2903-12.

[36] Huang, Y. A., Dando, R., Roper, S. D. Autocrine and paracrine roles for ATP and serotonin in mouse taste buds. J. Neurosci. 2009,29:13909-18.

[37] Huang, Y. A., Pereira, E., Roper, S. D. Acid stimulation (sour taste) elicits GABA and serotonin release from mouse taste cells. PLoS. One. 2011,6:e25471.

117 [38] Obata, H., Shimada, K., Sakai, N., Saito, N. GABAergic neurotransmission in rat taste buds: immunocytochemical study for GABA and GABA transporter subtypes. Brain Res. Mol. Brain Res. 1997,49:29-36.

[39] Cao, Y., Zhao, F. L., Kolli, T., Hivley, R., Herness, S. GABA expression in the mammalian taste bud functions as a route of inhibitory cell-to-cell communication. Proc. Natl. Acad. Sci. U. S. A. 2009,106:4006-11.

[40] Drewnowski, A., Specter, S. E. Poverty and obesity: the role of energy density and energy costs. Am. J. Clin. Nutr. 2004,79:6-16.

[41] Drewnowski, A., Henderson, S. A., Levine, A., Hann, C. Taste and food preferences as predictors of dietary practices in young women. Public Health Nutr. 1999,2:513-9.

[42] Connors, M., Bisogni, C. A., Sobal, J., Devine, C. M. Managing values in personal food systems. Appetite. 2001,36:189-200.

[43] Scheibehenne, B., Miesler, L., Todd, P. M. Fast and frugal food choices: uncovering individual decision heuristics. Appetite. 2007,49:578-89.

[44] Monge-Rojas, R., Mattei, J., Fuster, T., Willett, W., Campos, H. Influence of sensory and cultural perceptions of white rice, brown rice and beans by Costa Rican adults in their dietary choices. Appetite. 2014,81:200-8.

[45] Forrai, G., Bankovi, G. Taste perception for phenylthiocarbamide and food choice--a Hungarian twin study. Acta Physiol Hung. 1984,64:33-40.

[46] Mattes, R., Labov, J. Bitter taste responses to phenylthiocarbamide are not related to dietary goitrogen intake in human beings. J. Am. Diet. Assoc. 1989,89:692-4.

[47] EV, G., AR, K. Food preference and sensitivity of taste for bitter compounds. Nature. 1965,205:851- 3.

[48] A., N., M., K., M., S. Genetic influences on the selection of Brassica vegetables by elderly individuals. Nutr. Res. 1988,8:13-20.

[49] Mattes, R. D. Gustation as a determinant of ingestion: methodological issues. Am. J. Clin. Nutr. 1985,41:672-83.

118 [50] Lucas, F., Bellisle, F. The measurement of food preferences in humans: do taste-and-spit tests predict consumption? Physiol Behav. 1987,39:739-43.

[51] Rodin, J., Moskowitz, H. R., Bray, G. A. Relationship between obesity, weight loss, and taste responsiveness. Physiol Behav. 1976,17:591-7.

[52] Jerzsa-Latta, M., Krondl, M., Coleman, P. Use and perceived attributes of cruciferous vegetables in terms of genetically-mediated taste sensitivity. Appetite. 1990,15:127-34.

[53] Kulkarni, G. V., Chng, T., Eny, K. M., Nielsen, D., Wessman, C., El-Sohemy, A. Association of GLUT2 and TAS1R2 genotypes with risk for dental caries. Caries Res. 2013,47:219-25.

[54] Sharma, K., Kaur, G. K. PTC bitter taste genetic polymorphism, food choices, physical growth in body height and body fat related traits among adolescent girls from Kangra Valley, Himachal Pradesh (India). Ann. Hum. Biol. 2014,41:29-39.

[55] Zeinstra, G. G., Koelen, M. A., Kok, F. J., de, G. C. Cognitive development and children's perceptions of fruit and vegetables; a qualitative study. Int. J. Behav. Nutr. Phys. Act. 2007,4:30.

[56] Keski-Rahkonen, A., Bulik, C. M., Pietilainen, K. H., Rose, R. J., Kaprio, J., Rissanen, A. Eating styles, overweight and obesity in young adult twins. Eur. J. Clin. Nutr. 2007,61:822-9.

[57] Tholin, S., Rasmussen, F., Tynelius, P., Karlsson, J. Genetic and environmental influences on eating behavior: the Swedish Young Male Twins Study. Am. J. Clin. Nutr. 2005,81:564-9.

[58] Keskitalo, K., Tuorila, H., Spector, T. D., Cherkas, L. F., Knaapila, A., Kaprio, J., et al. The Three- Factor Eating Questionnaire, body mass index, and responses to sweet and salty fatty foods: a twin study of genetic and environmental associations. Am. J. Clin. Nutr. 2008,88:263-71.

[59] Fukunaga, A. [Age-related changes in renewal of taste bud cells and expression of taste cell-specific proteins in mice]. Kokubyo Gakkai Zasshi. 2005,72:84-9.

[60] Running, C. A., Craig, B. A., Mattes, R. D. Oleogustus: The Unique Taste of Fat. Chem. Senses. 2015,40:507-16.

[61] Tepper, B. J., Nurse, R. J. Fat perception is related to PROP taster status. Physiol Behav. 1997,61:949-54.

[62] Greenberg, D., Smith, G. P. The controls of fat intake. Psychosom. Med. 1996,58:559-69.

119 [63] Ramirez, I. Role of olfaction in starch and oil preference. Am. J. Physiol. 1993,265:R1404-R9.

[64] Fukuwatari, T., Shibata, K., Iguchi, K., Saeki, T., Iwata, A., Tani, K., et al. Role of gustation in the recognition of oleate and triolein in anosmic rats. Physiol Behav. 2003,78:579-83.

[65] Mattes, R. D. Is there a fatty acid taste? Annu. Rev. Nutr. 2009,29:305-27.

[66] Tsuruta, M., Kawada, T., Fukuwatari, T., Fushiki, T. The orosensory recognition of long-chain fatty acids in rats. Physiol Behav. 1999,66:285-8.

[67] Takeda, M., Imaizumi, M., Fushiki, T. Preference for vegetable oils in the two-bottle choice test in mice. Life Sci. 2000,67:197-204.

[68] Chale-Rush, A., Burgess, J. R., Mattes, R. D. Evidence for human orosensory (taste?) sensitivity to free fatty acids. Chem. Senses. 2007,32:423-31.

[69] Laugerette, F., Passilly-Degrace, P., Patris, B., Niot, I., Febbraio, M., Montmayeur, J. P., et al. CD36 involvement in orosensory detection of dietary lipids, spontaneous fat preference, and digestive secretions. J. Clin. Invest. 2005,115:3177-84.

[70] Sclafani, A., Ackroff, K., Abumrad, N. A. CD36 gene deletion reduces fat preference and intake but not post-oral fat conditioning in mice. Am. J. Physiol Regul. Integr. Comp Physiol. 2007,293:R1823-R32.

[71] Baillie, A. G., Coburn, C. T., Abumrad, N. A. Reversible binding of long-chain fatty acids to purified FAT, the adipose CD36 homolog. J. Membr. Biol. 1996,153:75-81.

[72] Keller, K. L., Liang, L. C., Sakimura, J., May, D., van, B. C., Breen, C., et al. Common variants in the CD36 gene are associated with oral fat perception, fat preferences, and obesity in African Americans. Obesity. (Silver. Spring). 2012,20:1066-73.

[73] Cartoni, C., Yasumatsu, K., Ohkuri, T., Shigemura, N., Yoshida, R., Godinot, N., et al. Taste preference for fatty acids is mediated by GPR40 and GPR120. J. Neurosci. 2010,30:8376-82.

[74] Stewart, J. E., Feinle-Bisset, C., Keast, R. S. Fatty acid detection during food consumption and digestion: Associations with ingestive behavior and obesity. Prog. Lipid Res. 2011,50:225-33.

[75] Liang, L. C., Sakimura, J., May, D., Breen, C., Driggin, E., Tepper, B. J., et al. Fat discrimination: a phenotype with potential implications for studying fat intake behaviors and obesity. Physiol Behav. 2012,105:470-5.

120 [76] Stewart, J. E., Feinle-Bisset, C., Golding, M., Delahunty, C., Clifton, P. M., Keast, R. S. Oral sensitivity to fatty acids, food consumption and BMI in human subjects. Br. J. Nutr. 2010,104:145-52.

[77] Gilbertson, T. A., Liu, L., York, D. A., Bray, G. A. Dietary fat preferences are inversely correlated with peripheral gustatory fatty acid sensitivity. Ann. N. Y. Acad. Sci. 1998,855:165-8.

[78] Stewart, J. E., Keast, R. S. Recent fat intake modulates fat taste sensitivity in lean and overweight subjects. Int. J. Obes. (Lond). 2012,36:834-42.

[79] Ma, X., Bacci, S., Mlynarski, W., Gottardo, L., Soccio, T., Menzaghi, C., et al. A common haplotype at the CD36 locus is associated with high free fatty acid levels and increased cardiovascular risk in Caucasians. Hum. Mol. Genet. 2004,13:2197-205.

[80] Noel, S. E., Lai, C. Q., Mattei, J., Parnell, L. D., Ordovas, J. M., Tucker, K. L. Variants of the CD36 gene and metabolic syndrome in Boston Puerto Rican adults. Atherosclerosis. 2010,211:210-5.

[81] Ramos-Arellano, L. E., Salgado-Bernabe, A. B., Guzman-Guzman, I. P., Salgado-Goytia, L., Munoz-Valle, J. F., Parra-Rojas, I. CD36 haplotypes are associated with lipid profile in normal-weight subjects. Lipids Health Dis. 2013,12:167.

[82] Love-Gregory, L., Sherva, R., Sun, L., Wasson, J., Schappe, T., Doria, A., et al. Variants in the CD36 gene associate with the metabolic syndrome and high-density lipoprotein cholesterol. Hum. Mol. Genet. 2008,17:1695-704.

[83] Lepretre, F., Cheyssac, C., Amouyel, P., Froguel, P., Helbecque, N. A promoter polymorphism in CD36 is associated with an atherogenic lipid profile in a French general population. Atherosclerosis. 2004,173:375-7.

[84] Corpeleijn, E., van der Kallen, C. J., Kruijshoop, M., Magagnin, M. G., de Bruin, T. W., Feskens, E. J., et al. Direct association of a promoter polymorphism in the CD36/FAT fatty acid transporter gene with Type 2 diabetes mellitus and insulin resistance. Diabet. Med. 2006,23:907-11.

[85] Lepretre, F., Linton, K. J., Lacquemant, C., Vatin, V., Samson, C., Dina, C., et al. Genetic study of the CD36 gene in a French diabetic population. Diabetes Metab. 2004,30:459-63.

[86] Yun, Y. M., Song, E. Y., Song, S. H., Song, J., Kim, J. Q. CD36 polymorphism and its relationship with body mass index and coronary artery disease in a Korean population. Clin. Chem. Lab Med. 2007,45:1277-82.

121 [87] Bokor, S., Legry, V., Meirhaeghe, A., Ruiz, J. R., Mauro, B., Widhalm, K., et al. Single-nucleotide polymorphism of CD36 locus and obesity in European adolescents. Obesity. (Silver. Spring). 2010,18:1398-403.

[88] Pepino, M. Y., Love-Gregory, L., Klein, S., Abumrad, N. A. The fatty acid translocase gene CD36 and lingual lipase influence oral sensitivity to fat in obese subjects. J. Lipid Res. 2012,53:561-6.

[89] Febbraio, M., Abumrad, N. A., Hajjar, D. P., Sharma, K., Cheng, W., Pearce, S. F., et al. A null mutation in murine CD36 reveals an important role in fatty acid and lipoprotein metabolism. J. Biol. Chem. 1999,274:19055-62.

[90] Goudriaan, J. R., Dahlmans, V. E., Teusink, B., Ouwens, D. M., Febbraio, M., Maassen, J. A., et al. CD36 deficiency increases insulin sensitivity in muscle, but induces insulin resistance in the liver in mice. J. Lipid Res. 2003,44:2270-7.

[91] Madden, J., Carrero, J. J., Brunner, A., Dastur, N., Shearman, C. P., Calder, P. C., et al. Polymorphisms in the CD36 gene modulate the ability of fish oil supplements to lower fasting plasma triacyl glycerol and raise HDL cholesterol concentrations in healthy middle-aged men. Prostaglandins Leukot. Essent. Fatty Acids. 2008,78:327-35.

[92] Colin-Ramirez, E., Castillo-Martinez, L., Orea-Tejeda, A., Zheng, Y., Westerhout, C. M., Ezekowitz, J. A. Dietary fatty acids intake and mortality in patients with heart failure. Nutrition. 2014,30:1366-71.

[93] Li, X., Staszewski, L., Xu, H., Durick, K., Zoller, M., Adler, E. Human receptors for sweet and umami taste. Proc. Natl. Acad. Sci. U. S. A. 2002,99:4692-6.

[94] Temussi, P. A. Why are sweet proteins sweet? Interaction of brazzein, monellin and thaumatin with the T1R2-T1R3 receptor. FEBS Lett. 2002,526:1-4.

[95] Jiang, P., Ji, Q., Liu, Z., Snyder, L. A., Benard, L. M., Margolskee, R. F., et al. The cysteine-rich region of T1R3 determines responses to intensely sweet proteins. J. Biol. Chem. 2004,279:45068-75.

[96] Walters, D. E., Hellekant, G. Interactions of the sweet protein brazzein with the sweet taste receptor. J. Agric. Food Chem. 2006,54:10129-33.

[97] Cui, M., Jiang, P., Maillet, E., Max, M., Margolskee, R. F., Osman, R. The heterodimeric sweet taste receptor has multiple potential ligand binding sites. Curr. Pharm. Des. 2006,12:4591-600.

[98] Nakajima, K., Asakura, T., Oike, H., Morita, Y., Shimizu-Ibuka, A., Misaka, T., et al. Neoculin, a taste-modifying protein, is recognized by human sweet taste receptor. Neuroreport. 2006,17:1241-4.

122 [99] ssadi-Porter, F. M., Maillet, E. L., Radek, J. T., Quijada, J., Markley, J. L., Max, M. Key amino acid residues involved in multi-point binding interactions between brazzein, a sweet protein, and the T1R2- T1R3 human sweet receptor. J. Mol. Biol. 2010,398:584-99.

[100] Nelson, G., Chandrashekar, J., Hoon, M. A., Feng, L., Zhao, G., Ryba, N. J., et al. An amino-acid taste receptor. Nature. 2002,416:199-202.

[101] Reed, D. R., McDaniel, A. H. The human sweet tooth. BMC. Oral Health. 2006,6 Suppl 1:S17.

[102] Reed, D. R., Tanaka, T., McDaniel, A. H. Diverse tastes: Genetics of sweet and bitter perception. Physiol Behav. 2006,88:215-26.

[103] McDaniel, A. H., Reed, D. R. Genomics and Proteomics in Nutrition. New York: Marcel Dekker; 2003.

[104] Reed, D. R., X, L., AA, B., K, M., GK, B. Progress in Obesity Research. London: John Libbey Eurotext Ltd.; 2003.

[105] Reed, D. R., Bachmanov, A. A., Beauchamp, G. K., Tordoff, M. G., Price, R. A. Heritable variation in food preferences and their contribution to obesity. Behav. Genet. 1997,27:373-87.

[106] Drayna, D. Human taste genetics. Annu. Rev. Genomics Hum. Genet. 2005,6:217-35.

[107] Kim, U. K., Breslin, P. A., Reed, D., Drayna, D. Genetics of human taste perception. J. Dent. Res. 2004,83:448-53.

[108] Keskitalo, K., Tuorila, H., Spector, T. D., Cherkas, L. F., Knaapila, A., Silventoinen, K., et al. Same genetic components underlie different measures of sweet taste preference. Am. J. Clin. Nutr. 2007,86:1663-9.

[109] Keskitalo, K., Knaapila, A., Kallela, M., Palotie, A., Wessman, M., Sammalisto, S., et al. Sweet taste preferences are partly genetically determined: identification of a trait locus on chromosome 16. Am. J. Clin. Nutr. 2007,86:55-63.

[110] Kim, U. K., Wooding, S., Riaz, N., Jorde, L. B., Drayna, D. Variation in the human TAS1R taste receptor genes. Chem. Senses. 2006,31:599-611.

[111] Eny, K. M., Wolever, T. M., Corey, P. N., El-Sohemy, A. Genetic variation in TAS1R2 (Ile191Val) is associated with consumption of sugars in overweight and obese individuals in 2 distinct populations. Am. J. Clin. Nutr. 2010,92:1501-10.

123 [112] Ramos-Lopez, O., Panduro, A., Martinez-Lopez, E., Roman, S. Sweet Taste Receptor TAS1R2 Polymorphism (Val191Val) Is Associated with a Higher Carbohydrate Intake and Hypertriglyceridemia among the Population of West Mexico. Nutrients. 2016,8:101.

[113] Fushan, A. A., Simons, C. T., Slack, J. P., Manichaikul, A., Drayna, D. Allelic polymorphism within the TAS1R3 promoter is associated with human taste sensitivity to sucrose. Curr. Biol. 2009,19:1288-93.

[114] WEISBERGER, D. A role of glucose in the production of artificial caries. J. Dent. Res. 1950,29:14- 22, illust.

[115] Haznedaroglu, E., Koldemir-Gunduz, M., Bakir-Coskun, N., Bozkus, H. M., Cagatay, P., Susleyici- Duman, B., et al. Association of Sweet Taste Receptor Gene Polymorphisms with Dental Caries Experience in School Children. Caries Res. 2015,49:275-81.

[116] JI, G. Is the bitter ejective response always adaptive? Physiol Behav. 1994,56:1217-27.

[117] CM, H., B, S. Taste percpetion and feeding behavior in nonhuman primates and human populations. Evol Anthropol. 1996,5:58-71.

[118] Goldstein, G. L., Daun, H., Tepper, B. J. Adiposity in middle-aged women is associated with genetic taste blindness to 6-n-propylthiouracil. Obes. Res. 2005,13:1017-23.

[119] Dong, D., Jones, G., Zhang, S. Dynamic evolution of bitter taste receptor genes in vertebrates. BMC. Evol Biol. 2009,9:12.

[120] Kim, U. K., Jorgenson, E., Coon, H., Leppert, M., Risch, N., Drayna, D. Positional cloning of the human quantitative trait locus underlying taste sensitivity to phenylthiocarbamide. Science. 2003,299:1221-5.

[121] Bufe, B., Breslin, P. A., Kuhn, C., Reed, D. R., Tharp, C. D., Slack, J. P., et al. The molecular basis of individual differences in phenylthiocarbamide and propylthiouracil bitterness perception. Curr. Biol. 2005,15:322-7.

[122] Bartoshuk, L. M., Duffy, V. B., Miller, I. J. PTC/PROP tasting: anatomy, psychophysics, and sex effects. Physiol Behav. 1994,56:1165-71.

[123] Duffy, V. B., Davidson, A. C., Kidd, J. R., Kidd, K. K., Speed, W. C., Pakstis, A. J., et al. Bitter receptor gene (TAS2R38), 6-n-propylthiouracil (PROP) bitterness and alcohol intake. Alcohol Clin. Exp. Res. 2004,28:1629-37.

124 [124] Hayes, J. E., Bartoshuk, L. M., Kidd, J. R., Duffy, V. B. Supertasting and PROP bitterness depends on more than the TAS2R38 gene. Chem. Senses. 2008,33:255-65.

[125] Cannon, D. S., Baker, T. B., Piper, M. E., Scholand, M. B., Lawrence, D. L., Drayna, D. T., et al. Associations between phenylthiocarbamide gene polymorphisms and cigarette smoking. Nicotine. Tob. Res. 2005,7:853-8.

[126] Drewnowski, A., Henderson, S. A., Shore, A. B., Barratt-Fornell, A. Sensory responses to 6-n- propylthiouracil (PROP) or sucrose solutions and food preferences in young women. Ann. N. Y. Acad. Sci. 1998,855:797-801.

[127] Duffy, V. B., Bartoshuk, L. M. Food acceptance and genetic variation in taste. J. Am. Diet. Assoc. 2000,100:647-55.

[128] Timpson, N. J., Christensen, M., Lawlor, D. A., Gaunt, T. R., Day, I. N., Ebrahim, S., et al. TAS2R38 (phenylthiocarbamide) haplotypes, coronary heart disease traits, and eating behavior in the British Women's Heart and Health Study. Am. J. Clin. Nutr. 2005,81:1005-11.

[129] Ooi, S. X., Lee, P. L., Law, H. Y., Say, Y. H. Bitter receptor gene (TAS2R38) P49A genotypes and their associations with aversion to vegetables and sweet/fat foods in Malaysian subjects. Asia Pac. J. Clin. Nutr. 2010,19:491-8.

[130] O'Brien, S. A., Feeney, E. L., Scannell, A. G., Markey, A., Gibney, E. R. Bitter taste perception and dietary intake patterns in irish children. J. Nutrigenet. Nutrigenomics. 2013,6:43-58.

[131] Gayathri, D. A., Henderson, S. A., Drewnowski, A. Sensory acceptance of Japanese green tea and soy products is linked to genetic sensitivity to 6-n-propylthiouracil. Nutr. Cancer. 1997,29:146-51.

[132] Choi, S. E. Racial differences between African Americans and Asian Americans in the effect of 6- n-propylthiouracil taste intensity and food liking on body mass index. J. Acad. Nutr. Diet. 2014,114:938- 44.

[133] Tepper, B. J. Nutritional implications of genetic taste variation: the role of PROP sensitivity and other taste phenotypes. Annu. Rev. Nutr. 2008,28:367-88.

[134] Drewnowski, A., Henderson, S. A., Cockroft, J. E. Genetic sensitivity to 6-n-propylthiouracil has no influence on dietary patterns, body mass indexes, or plasma lipid profiles of women. J. Am. Diet. Assoc. 2007,107:1340-8.

125 [135] Tepper, B. J., Koelliker, Y., Zhao, L., Ullrich, N. V., Lanzara, C., d'Adamo, P., et al. Variation in the bitter-taste receptor gene TAS2R38, and adiposity in a genetically isolated population in Southern Italy. Obesity. (Silver. Spring). 2008,16:2289-95.

[136] Tepper, B. J., Nurse, R. J. PROP taster status is related to fat perception and preference. Ann. N. Y. Acad. Sci. 1998,855:802-4.

[137] Ikeda, K. New seasonings. Chem. Senses. 2002,27:847-9.

[138] Toyono, T., Seta, Y., Kataoka, S., Kawano, S., Shigemoto, R., Toyoshima, K. Expression of metabotropic glutamate receptor group I in rat gustatory papillae. Cell Tissue Res. 2003,313:29-35.

[139] Toyono, T., Seta, Y., Kataoka, S., Harada, H., Morotomi, T., Kawano, S., et al. Expression of the metabotropic glutamate receptor, mGluR4a, in the taste hairs of taste buds in rat gustatory papillae. Arch. Histol. Cytol. 2002,65:91-6.

[140] varez-Lario, B., arron-Vicente, J. Uric acid and evolution. Rheumatology. (Oxford). 2010,49:2010- 5.

[141] Baldwin, W., McRae, S., Marek, G., Wymer, D., Pannu, V., Baylis, C., et al. Hyperuricemia as a mediator of the proinflammatory endocrine imbalance in the adipose tissue in a murine model of the metabolic syndrome. Diabetes. 2011,60:1258-69.

[142] Lanaspa, M. A., Sanchez-Lozada, L. G., Choi, Y. J., Cicerchi, C., Kanbay, M., Roncal-Jimenez, C. A., et al. Uric acid induces hepatic steatosis by generation of mitochondrial oxidative stress: potential role in fructose-dependent and -independent fatty liver. J. Biol. Chem. 2012,287:40732-44.

[143] Lugaz, O., Pillias, A. M., Faurion, A. A new specific ageusia: some humans cannot taste L- glutamate. Chem. Senses. 2002,27:105-15.

[144] Shigemura, N., Shirosaki, S., Sanematsu, K., Yoshida, R., Ninomiya, Y. Genetic and molecular basis of individual differences in human umami taste perception. PLoS. One. 2009,4:e6717.

[145] Raliou, M., Grauso, M., Hoffmann, B., Schlegel-Le-Poupon, C., Nespoulous, C., Debat, H., et al. Human genetic polymorphisms in T1R1 and T1R3 taste receptor subunits affect their function. Chem. Senses. 2011,36:527-37.

[146] Garcia-Bailo, B., Toguri, C., Eny, K. M., El-Sohemy, A. Genetic variation in taste and its influence on food selection. OMICS. 2009,13:69-80.

126 [147] Ishimaru, Y., Inada, H., Kubota, M., Zhuang, H., Tominaga, M., Matsunami, H. Transient receptor potential family members PKD1L3 and PKD2L1 form a candidate sour taste receptor. Proc. Natl. Acad. Sci. U. S. A. 2006,103:12569-74.

[148] LopezJimenez, N. D., Cavenagh, M. M., Sainz, E., Cruz-Ithier, M. A., Battey, J. F., Sullivan, S. L. Two members of the TRPP family of ion channels, Pkd1l3 and Pkd2l1, are co-expressed in a subset of taste receptor cells. J. Neurochem. 2006,98:68-77.

[149] Richter, T. A., Caicedo, A., Roper, S. D. Sour taste stimuli evoke Ca2+ and pH responses in mouse taste cells. J. Physiol. 2003,547:475-83.

[150] Ye, W., Chang, R. B., Bushman, J. D., Tu, Y. H., Mulhall, E. M., Wilson, C. E., et al. The K+ channel KIR2.1 functions in tandem with proton influx to mediate sour taste transduction. Proceedings of the National Academy of Sciences of the United States of America. 2016,113:E229-38.

[151] Bertino, M., Beauchamp, G. K., Engelman, K. Long-term reduction in dietary sodium alters the taste of salt. Am. J. Clin. Nutr. 1982,36:1134-44.

[152] Wise, P. M., Hansen, J. L., Reed, D. R., Breslin, P. A. Twin study of the heritability of recognition thresholds for sour and salty taste. Chem. Senses. 2007,32:749-54.

[153] Wald, N., Leshem, M. Salt conditions a flavor preference or aversion after exercise depending on NaCl dose and sweat loss. Appetite. 2003,40:277-84.

[154] Hayes, J. E., Sullivan, B. S., Duffy, V. B. Explaining variability in sodium intake through oral sensory phenotype, salt sensation and liking. Physiol Behav. 2010,100:369-80.

[155] Dias, A. G., Rousseau, D., Duizer, L., Cockburn, M., Chiu, W., Nielsen, D., et al. Genetic variation in putative salt taste receptors and salt taste perception in humans. Chem. Senses. 2013,38:137-45.

[156] Nielsen, D. E., El-Sohemy, A. Disclosure of genetic information and change in dietary intake: a randomized controlled trial. PLoS. One. 2014,9:e112665.

[157] Nielsen, D. E., Shih, S., El-Sohemy, A. Perceptions of genetic testing for personalized nutrition: a randomized trial of DNA-based dietary advice. J. Nutrigenet. Nutrigenomics. 2014,7:94-104.

[158] Dias, A. G., Eny, K. M., Cockburn, M., Chiu, W., Nielsen, D. E., Duizer, L., et al. Variation in the TAS1R2 Gene, Sweet Taste Perception and Intake of Sugars. J Nutrigenet. Nutrigenomics. 2015,8:81-90.

127 [159] Pepino, M. Y., Finkbeiner, S., Beauchamp, G. K., Mennella, J. A. Obese women have lower monosodium glutamate taste sensitivity and prefer higher concentrations than do normal-weight women. Obesity. (Silver. Spring). 2010,18:959-65.

[160] Park, D. C., Yeo, J. H., Ryu, I. Y., Kim, S. H., Jung, J., Yeo, S. G. Differences in taste detection thresholds between normal-weight and obese young adults. Acta Otolaryngol. 2015,135:478-83.

[161] Hill, J. O., Wyatt, H. R., Peters, J. C. Energy balance and obesity. Circulation. 2012,126:126-32.

[162] Kopelman, P. G. Obesity as a medical problem. Nature. 2000,404:635-43.

[163] Allison, D. B., Kaprio, J., Korkeila, M., Koskenvuo, M., Neale, M. C., Hayakawa, K. The heritability of body mass index among an international sample of monozygotic twins reared apart. Int. J Obes. Relat Metab Disord. 1996,20:501-6.

[164] Allison, D. B., Faith, M. S., Nathan, J. S. Risch's lambda values for human obesity. Int. J Obes. Relat Metab Disord. 1996,20:990-9.

[165] Chamoun, E., Mutch, D. M., Allen-Vercoe, E., Buchholz, A. C., Duncan, A. M., Spriet, L. L., et al. A review of the associations between single nucleotide polymorphisms in taste receptors, eating behaviors, and health. Crit Rev Food Sci Nutr. 2018,58:194-207.

[166] Pribitkin, E., Rosenthal, M. D., Cowart, B. J. Prevalence and causes of severe taste loss in a chemosensory clinic population. Ann. Otol. Rhinol. Laryngol. 2003,112:971-8.

[167] Wetherill, G. B., Levitt, H. Sequential estimation of points on a psychometric function. Br. J Math. Stat. Psychol. 1965,18:1-10.

[168] Mennella, J. A., Finkbeiner, S., Lipchock, S. V., Hwang, L. D., Reed, D. R. Preferences for salty and sweet tastes are elevated and related to each other during childhood. PLoS. One. 2014,9:e92201.

[169] The International HapMap Project. Nature. 2003,426:789-96.

[170] Uneyama, H., Kawai, M., Sekine-Hayakawa, Y., Torii, K. Contribution of umami taste substances in human salivation during meal. J Med Invest. 2009,56 Suppl:197-204.

[171] Yamaguchi, S., Ninomiya, K. Umami and food palatability. J Nutr. 2000,130:921S-6S.

128 [172] Masic, U., Yeomans, M. R. Umami flavor enhances appetite but also increases satiety. Am J Clin Nutr. 2014,100:532-8.

[173] Miyaki, T., Imada, T., Hao, S. S., Kimura, E. Monosodium L-glutamate in soup reduces subsequent energy intake from high-fat savoury food in overweight and obese women. The British journal of nutrition. 2016,115:176-84.

[174] Kondoh, T., Torii, K. MSG intake suppresses weight gain, fat deposition, and plasma leptin levels in male Sprague-Dawley rats. Physiol Behav. 2008,95:135-44.

[175] Masic, U., Yeomans, M. R. Does acute or habitual protein deprivation influence liking for monosodium glutamate? Physiology & Behavior. 2017,171:79-86.

[176] Raliou, M., Wiencis, A., Pillias, A. M., Planchais, A., Eloit, C., Boucher, Y., et al. Nonsynonymous single nucleotide polymorphisms in human tas1r1, tas1r3, and mGluR1 and individual taste sensitivity to glutamate. Am J Clin Nutr. 2009,90:789S-99S.

[177] Kurihara, K. Umami the Fifth Basic Taste: History of Studies on Receptor Mechanisms and Role as a Food Flavor. Biomed Res Int. 2015,2015:189402.

[178] Gu, X., Gu, D., He, J., Rao, D. C., Hixson, J. E., Chen, J., et al. Resequencing Epithelial Sodium Channel Genes Identifies Rare Variants Associated with Blood Pressure Salt-Sensitivity: The GenSalt Study. American journal of hypertension. 2017.

[179] Yang, X., He, J., Gu, D., Hixson, J. E., Huang, J., Rao, D. C., et al. Associations of epithelial sodium channel genes with blood pressure changes and hypertension incidence: the GenSalt study. American journal of hypertension. 2014,27:1370-6.

[180] Rao, A. D., Sun, B., Saxena, A., Hopkins, P. N., Jeunemaitre, X., Brown, N. J., et al. Polymorphisms in the serum- and glucocorticoid-inducible kinase 1 gene are associated with blood pressure and renin response to dietary salt intake. Journal of human hypertension. 2013,27:176-80.

[181] Zhao, Q., Gu, D., Hixson, J. E., Liu, D. P., Rao, D. C., Jaquish, C. E., et al. Common variants in epithelial sodium channel genes contribute to salt sensitivity of blood pressure: The GenSalt study. Circulation. Cardiovascular genetics. 2011,4:375-80.

[182] Dias, A. G., Rousseau, D., Duizer, L., Cockburn, M., Chiu, W., Nielsen, D., et al. Genetic variation in putative salt taste receptors and salt taste perception in humans. Chemical senses. 2013,38:137-45.

[183] Strazzullo, P., D'Elia, L., Kandala, N. B., Cappuccio, F. P. Salt intake, stroke, and cardiovascular disease: meta-analysis of prospective studies. BMJ (Clinical research ed.). 2009,339:b4567.

129 [184] Cook, N. R., Cutler, J. A., Obarzanek, E., Buring, J. E., Rexrode, K. M., Kumanyika, S. K., et al. Long term effects of dietary sodium reduction on cardiovascular disease outcomes: observational follow- up of the trials of hypertension prevention (TOHP). BMJ (Clinical research ed.). 2007,334:885-8.

[185] Liu, K., Stamler, J. Assessment of sodium intake in epidemiological studies on blood pressure. Annals of clinical research. 1984,16 Suppl 43:49-54.

[186] Elliott, P., Stamler, J., Nichols, R., Dyer, A. R., Stamler, R., Kesteloot, H., et al. Intersalt revisited: further analyses of 24 hour sodium excretion and blood pressure within and across populations. Intersalt Cooperative Research Group. BMJ (Clinical research ed.). 1996,312:1249-53.

[187] Gross, L. S., Li, L., Ford, E. S., Liu, S. Increased consumption of refined carbohydrates and the epidemic of type 2 diabetes in the United States: an ecologic assessment. Am J Clin Nutr. 2004,79:774-9.

[188] Bray, G. A., Nielsen, S. J., Popkin, B. M. Consumption of high-fructose corn syrup in beverages may play a role in the epidemic of obesity. Am J Clin Nutr. 2004,79:537-43.

[189] Lustig, R. H., Schmidt, L. A., Brindis, C. D. Public health: The toxic truth about sugar. Nature. 2012,482:27-9.

[190] Eny, K. M., Wolever, T. M., Corey, P. N., El-Sohemy, A. Genetic variation in TAS1R2 (Ile191Val) is associated with consumption of sugars in overweight and obese individuals in 2 distinct populations. The American journal of clinical nutrition. 2010,92:1501-10.

[191] Han, P., Keast, R. S. J., Roura, E. Salivary leptin and TAS1R2/TAS1R3 polymorphisms are related to sweet taste sensitivity and carbohydrate intake from a buffet meal in healthy young adults. The British journal of nutrition. 2017,118:763-70.

[192] Chamoun, E., Hutchinson, J., Krystia, O., Mirotta, J., Mutch, D., Buchholz, A., et al. Single Nucleotide Polymorphisms in Taste Receptor Genes Are Associated with Snacking Patterns of Preschool- Aged Children in the Guelph Family Health Study: A Pilot Study. Nutrients. 2018,10:153.

[193] Tornwall, O., Silventoinen, K., Keskitalo-Vuokko, K., Perola, M., Kaprio, J., Tuorila, H. Genetic contribution to sour taste preference. Appetite. 2012,58:687-94.

[194] Tucker, R. M., Kaiser, K. A., Parman, M. A., George, B. J., Allison, D. B., Mattes, R. D. Comparisons of Fatty Acid Taste Detection Thresholds in People Who Are Lean vs. Overweight or Obese: A Systematic Review and Meta-Analysis. PloS one. 2017,12:e0169583.

[195] Hardikar, S., Hochenberger, R., Villringer, A., Ohla, K. Higher sensitivity to sweet and salty taste in obese compared to lean individuals. Appetite. 2017,111:158-65.

130 [196] Proserpio, C., Laureati, M., Bertoli, S., Battezzati, A., Pagliarini, E. Determinants of Obesity in Italian Adults: The Role of Taste Sensitivity, Food Liking, and Food Neophobia. Chemical senses. 2016,41:169-76.

[197] Bobowski, N. K., Mennella, J. A. Disruption in the Relationship between Blood Pressure and Salty Taste Thresholds among Overweight and Obese Children. Journal of the Academy of Nutrition and Dietetics. 2015,115:1272-82.

[198] Sayed, A., Sery, O., Plesnik, J., Daoudi, H., Rouabah, A., Rouabah, L., et al. CD36 AA genotype is associated with decreased lipid taste perception in young obese, but not lean, children. International journal of obesity (2005). 2015,39:920-4.

[199] Fogel, A., Blissett, J. Effects of sucrose detection threshold and weight status on intake of fruit and vegetables in children. Appetite. 2014,83:309-16.

[200] Kirsten, V. R., Wagner, M. B. Salt taste sensitivity thresholds in adolescents: are there any relationships with body composition and blood pressure levels? Appetite. 2014,81:89-92.

[201] Tucker, R. M., Edlinger, C., Craig, B. A., Mattes, R. D. Associations between BMI and fat taste sensitivity in humans. Chemical senses. 2014,39:349-57.

[202] Piovesana Pde, M., Sampaio Kde, L., Gallani, M. C. Association between Taste Sensitivity and Self-Reported and Objective Measures of Salt Intake among Hypertensive and Normotensive Individuals. ISRN nutrition. 2013,2013:301213.

[203] Ettinger, L., Duizer, L., Caldwell, T. Body fat, sweetness sensitivity, and preference: determining the relationship. Canadian journal of dietetic practice and research : a publication of Dietitians of Canada = Revue canadienne de la pratique et de la recherche en dietetique : une publication des Dietetistes du Canada. 2012,73:45-8.

[204] Joseph, P. V., Reed, D. R., Mennella, J. A. Individual Differences Among Children in Sucrose Detection Thresholds: Relationship With Age, Gender, and Bitter Taste Genotype. Nursing research. 2016,65:3-12.

[205] Stewart, J. E., Feinle-Bisset, C., Golding, M., Delahunty, C., Clifton, P. M., Keast, R. S. Oral sensitivity to fatty acids, food consumption and BMI in human subjects. The British journal of nutrition. 2010,104:145-52.

[206] Pepino, M. Y., Finkbeiner, S., Beauchamp, G. K., Mennella, J. A. Obese women have lower monosodium glutamate taste sensitivity and prefer higher concentrations than do normal-weight women. Obesity (Silver Spring, Md.). 2010,18:959-65.

131 [207] Bossola, M., Cadoni, G., Bellantone, R., Carriero, C., Carriero, E., Ottaviani, F., et al. Taste intensity and hedonic responses to simple beverages in gastrointestinal cancer patients. Journal of pain and symptom management. 2007,34:505-12.

[208] Luscombe-Marsh, N. D., Smeets, A. J., Westerterp-Plantenga, M. S. Taste sensitivity for monosodium glutamate and an increased liking of dietary protein. The British journal of nutrition. 2008,99:904-8.

[209] Bolhuis, D. P., Newman, L. P., Keast, R. S. Effects of Salt and Fat Combinations on Taste Preference and Perception. Chemical senses. 2016,41:189-95.

[210] Bolhuis, D. P., Costanzo, A., Newman, L. P., Keast, R. S. Salt Promotes Passive Overconsumption of Dietary Fat in Humans. The Journal of nutrition. 2016,146:838-45.

[211] Low, J. Y., McBride, R. L., Lacy, K. E., Keast, R. S. Psychophysical Evaluation of Sweetness Functions Across Multiple Sweeteners. Chemical senses. 2017,42:111-20.

[212] Running, C. A., Craig, B. A., Mattes, R. D. Oleogustus: The Unique Taste of Fat. Chemical senses. 2015,40:507-16.

[213] Mattes, R. D. Oral detection of short-, medium-, and long-chain free fatty acids in humans. Chem. Senses. 2009,34:145-50.

[214] Heinze, J. M., Preissl, H., Fritsche, A., Frank, S. Controversies in fat perception. Physiology & behavior. 2015,152:479-93.

[215] 2009 to 2011 Canadian Health Measures Survey. 2015.

[216] Chamoun, E., Mutch, D. M., Allen-Vercoe, E., Buchholz, A. C., Duncan, A. M., Spriet, L. L., et al. A review of the associations between single nucleotide polymorphisms in taste receptors, eating behaviors, and health. Crit Rev. Food Sci. Nutr. 2016,0:1-14.

[217] Reed, D. R., Xia, M. B. Recent advances in fatty acid perception and genetics. Adv. Nutr. 2015,6:353S-60S.

[218] Melo, S. V., Agnes, G., Vitolo, M. R., Mattevi, V. S., Campagnolo, P. D. B., Almeida, S. Evaluation of the association between the TAS1R2 and TAS1R3 variants and food intake and nutritional status in children. Genet. Mol. Biol. 2017,40:415-20.

132 [219] Pawellek, I., Grote, V., Theurich, M., Closa-Monasterolo, R., Stolarczyk, A., Verduci, E., et al. Factors associated with sugar intake and sugar sources in European children from 1 to 8 years of age. European journal of clinical nutrition. 2017,71:25-32.

[220] Hoppu, U., Laitinen, K., Jaakkola, J., Sandell, M. The hTAS2R38 genotype is associated with sugar and candy consumption in preschool boys. J Hum. Nutr Diet. 2015,28 Suppl 1:45-51.

[221] Piernas, C., Popkin, B. M. Trends in snacking among U.S. children. Health Aff. (Millwood. ). 2010,29:398-404.

[222] Davison, K. K., Blake, C. E., Blaine, R. E., Younginer, N. A., Orloski, A., Hamtil, H. A., et al. Parenting around child snacking: development of a theoretically-guided, empirically informed conceptual model. Int. J. Behav. Nutr. Phys. Act. 2015,12:109.

[223] JM, H., J, W., J, H., Duncan, A. M., J, M., DWL, M., et al. Snacking patterns of preschool-aged children: Opportunity for improvement. Can J Diet Pract Res. 2017,0:1-5.

[224] Duffey, K. J., Pereira, R. A., Popkin, B. M. Prevalence and energy intake from snacking in Brazil: analysis of the first nationwide individual survey. European journal of clinical nutrition. 2013,67:868-74.

[225] Willett, W. C. Dietary fats and coronary heart disease. J Intern. Med. 2012,272:13-24.

[226] Keller, K. L., Tepper, B. J. Inherited taste sensitivity to 6-n-propylthiouracil in diet and body weight in children. Obes. Res. 2004,12:904-12.

[227] O'Brien, S. A., Feeney, E. L., Scannell, A. G., Markey, A., Gibney, E. R. Bitter taste perception and dietary intake patterns in irish children. J Nutrigenet. Nutrigenomics. 2013,6:43-58.

[228] Feeney, E. L., O'Brien, S. A., Scannell, A. G., Markey, A., Gibney, E. R. Suprathreshold measures of taste perception in children - Association with dietary quality and body weight. Appetite. 2017,113:116-23.

[229] World Health Organization. Obesity and overweight. 2017.

[230] Caballero, B. The Global Epidemic of Obesity: An Overview. Epidemiologic Reviews. 2007,29:1- 5.

[231] Borjeson, M. The aetiology of obesity in children. A study of 101 twin pairs. Acta paediatrica Scandinavica. 1976,65:279-87.

133 [232] Stunkard, A. J., Foch, T. T., Hrubec, Z. A twin study of human obesity. Jama. 1986,256:51-4.

[233] Skouteris, H., McCabe, M., Swinburn, B., Newgreen, V., Sacher, P., Chadwick, P. Parental influence and obesity prevention in pre-schoolers: a systematic review of interventions. Obesity reviews : an official journal of the International Association for the Study of Obesity. 2011,12:315-28.

[234] Pioltine, M. B., de Melo, M. E., Santos, A. S., Machado, A. D., Fernandes, A. E., Fujiwara, C. T., et al. Genetic Variations in Sweet Taste Receptor Gene Are Related to Chocolate Powder and Dietary Fiber Intake in Obese Children and Adolescents. Journal of personalized medicine. 2018,8.

[235] Kushner, R. F., Schoeller, D. A., Fjeld, C. R., Danford, L. Is the impedance index (ht2/R) significant in predicting total body water? Am. J Clin. Nutr. 1992,56:835-9.

[236] Guinard, J.-X. Sensory and consumer testing with children. Trends in Food Science & Technology. 2000,11:273-83.

[237] Zhang, J., Zhou, Z., Zhang, N., Jin, W., Ren, Y., Chen, C. Establishment of preliminary regulatory network of TRPV1 and related cytokines. Saudi journal of biological sciences. 2017,24:582-8.

134 APPENDIX A

Taste receptor SNPs and primers used:

SNP ID Forward primer Reverse primer Extend primer rs10082458 acgttggatgcaccttctttactgatgccc acgttggatggaatgaggaatttggagccc ctcacccccaccaca rs10947476 acgttggatgctgggttcccctgaccttc acgttggatgagagagacaggcgcttagag ttccggcacaggcag rs1108923 acgttggatgcacacgtgatggttaagccc acgttggatggcctgatatcctgggatgac cctccaccttgccata rs11122100 acgttggatgacagcacaaagtggaggatg acgttggatgcttcccccatgaaatatccc gcgggcacgatgagct rs11187515 acgttggatgtgagcacctaccctttctac acgttggatgcagtcctggtgtggagtaag cttggtgcccctgtctc rs11187519 acgttggatgggaaatcagaggcaatgacc acgttggatgcacaaggacaaggaacttgc ggggcaaccacctcata rs11187529 acgttggatgcattacctgtggaaaacccg acgttggatgtgcagatgatcctgagagac tccgaacaagaagacga rs11187534 acgttggatggtttctgaccaatgccttcg acgttggatgacacagcagcacacaatcac atgtagtgaagaggcag rs11187536 acgttggatgaaatgcagggccttcagatc acgttggatgcgtttgcgttggaatccctg cccacacctattaggtct rs11759763 acgttggatgctctgtgttcacctactctgc acgttggatgatgtatacagtgccgtcagc ctctctccagtgcagatc rs12080675 acgttggatgtagtgccagagacttagcag acgttggatgttttcactgtcgtcttcctc tgaaaatgaagttgccct rs12137730 acgttggatgagatgtccctcttgtagacc acgttggatggcagaaaatcacctaccttc gttcacttcccctcataac rs12192574 acgttggatgacgcgcatcgcacgcaaat acgttggatgtatgctgtgcgcatgtgcgt cccctggtacacgcaaagc rs12243124 acgttggatgcaaagaggagtaaagcaccc acgttggatgaggaattctggtgccagttg acaggacatacagcaatac rs12415204 acgttggatgttactggagggcatctagtg acgttggatgtctcagtgcttggaatgctc ggtttctagaggcagcagg rs1414929 acgttggatgtggagtgtttttccttcacc acgttggatgtggaaggaacgcgtttgttg ttttttccttcaccttttcc rs1466650 acgttggatgtgaggtggcccacatttact acgttggatgcctgcctgatagtgtttgag gttactgggtgcctatttca rs150846 acgttggatgtcctgacgtcggccatgag acgttggatgtcagccgtgacagcatgagt aggccatgagcacatagtca rs150908 acgttggatgagcttccgcctgtgcgtca acgttggatgcatcacagccttctgctacg ggggttgcgtcaggccggag rs1527483 acgttggatggggcaggattctccttttag acgttggatggtttcctacctttgggagag cccgtctccttttaggcccta rs1565356 acgttggatgctctggtgtgagaggcattc acgttggatgtgtcacaatggatggtgctc cccaagaggcattcaaccact rs161364 acgttggatgttcttggaagaatctttgg acgttggatgatgcatctaacctatcaggg ttgcagaatctttgggcttcc rs161386 acgttggatgtcaactctggacagctgcac acgttggatgtacagtgcaagggactagacgc ggcactttggattatcacaat rs16953163 acgttggatggagtaggagggtgactttct acgttggatgcaatgtaacagggacagtgg ggagggtgactttctagatgg rs17108968 acgttggatgtggcttggtgctgaggatag acgttggatgtcctgttgctcacccggttt tagtatagcagagcccccagcc rs173135 acgttggatgacagctgtgtctaccatgtg acgttggatgggtcctgaggaagtgagata gtgtctaccatgtgctaattct

135 SNP ID Forward primer Reverse promer Extend primer rs1761667 acgttggatgtcactgtccccaagtgaatc acgttggatgcatgtcccagtagagactgacc tgtttactcactttactaaacagt rs17706630 acgttggatgactcactgactcttctatgcag acgttggatgtgcctagaaaagaaggaatc gttactacatatctagtaaaacct rs1873249 acgttggatgcttttccaattcaattctagagtc acgttggatgttgtaagctataaccagggg ttaaattctagagtctcataggtct rs1906953 acgttggatgggcccatagtaggtgctctc acgttggatggtcctgttcctctcttctccc acaggaagggaacctgccacctctca rs222745 acgttggatgcttcagcacacccatcaacc acgttggatgcatgtacatgctcagtggtg cccatccccgcttaa rs222749 acgttggatgcagtggcccagtcagtattg acgttggatgcacagagctagaacatggag tattgggccactcca rs2229901 acgttggatgggccaatcgcagcgcgcct acgttggatggcggcaggttccctccgct gaggatgctgcgagg rs224534 acgttggatgttcccatttctaaccacccc acgttggatggaaactaggttcatttccgc accgtttccatgctca rs224546 acgttggatgagatatgggtgacagcaagg acgttggatgatgaacacaaggaggctctg agtcctcacgtggtgc rs2274333 acgttggatgtccagagtgttctcaattcc acgttggatgagaggctggatctgaactcg agcagctcctagagaa rs2301151 acgttggatgaagaaactgatggcagaagc acgttggatgacacctgagaatcaaggtgg tagccccctcctgatcc rs236512 acgttggatgcaactgagtcacctgtcttc acgttggatggtgaagaaaggtgcatgctg cccccttgtgacccggt rs236513 acgttggatggtggatttacaactatacccag acgttggatggcccaattcaatacagcaac acccagaagtacagaca rs236514 acgttggatgaaccacatttttgtggaggg acgttggatgaggtaataagcaacgcggtc ccaggcaggtgctgtgg rs239345 acgttggatgatgggcacctgtgactaaga acgttggatggtacctactgaaatgcaggg gacctaataggtgtggg rs2451328 acgttggatgcaaagcttgggactcagcac acgttggatgcaggctcaaatccaactcgg ctttgagacactcactct rs2451361 acgttggatgtgttttagaataagtcatacatgtaca acgttggatgagtgtgagaaggctgatggg tgcacagtttgtactgtt rs2499669 acgttggatggcttctgggctaaatgtcac acgttggatggtgactgcacatcccacga attccagcgtccttcctca rs2499673 acgttggatgcatctcctcctcgaattacg acgttggatgtcgccaggtgctacaatact cccattcggtgtgcctctg rs2499682 acgttggatggctcctttttcacctagcag acgttggatgcaagaatgaagtgccttctc agcagtatatcatggtgct rs2499683 acgttggatgattcccagctttaaccctcc acgttggatgatccggtacctagtgcttag agcgggttacaataaaaat rs2499686 acgttggatgaccccacgccatacacatc acgttggatgggaaatgtggatggtgaccc ccccttgtgtggcacagtct rs2499694 acgttggatgcactgtaggactcaatccag acgttggatgcttcaccttagggaagatgc ggactgatcagaagagcaag rs2499697 acgttggatgtgcctatctcatgcagtcag acgttggatgcacaagccggctcctctta ccagtcctttcctttcagccc rs2499711 acgttggatggaacatctggccttctctac acgttggatgtttccagaaagccatatggg gggacctgggaggtgactaa rs2499729 acgttggatgggagttaccaacttctctgc acgttggatgaggatctctccagtaactcg ctccctttaagatggaaaaaa rs307355 acgttggatgactcacctcaggacttgact acgttggatggtaggcagctggtctaaatc tcttcaggacttgactccctca rs3211816 acgttggatgggctaagcttctttttgctc acgttggatgcaatattccacctacttcatatgac acgcttctttttgctctcagat rs3211828 acgttggatgacctggctgagatgtcaacg acgttggatggctgtaaaataacagactgcc tccccgttgctaggatgtgttc

136 SNP ID Forward primer Reverse promer Extend primer rs3211849 acgttggatgctttaatgtgacaggctctc acgttggatgcattcaactcccgaaaaacc ttctctcacttgaaggaaaaaa rs3211864 acgttggatggtgacatcctacatctttgac acgttggatggttgtaaaccgccaacactg gtggcagttgatttgtattatg rs3211908 acgttggatggagcctcaggctaatgagaa acgttggatgagactggagcataacacaac cctccgctgttttcacctacgca rs3211952 acgttggatgtaatgaccttccccacctcc acgttggatggtggcaggaaacagagctg ggcaaggaggagtgaatgataac rs35744813 acgttggatgtggctccatttaacaacactagt acgttggatgtgggtgaggcttttagtacgt atatccccaaaccgatatgaaaga rs35874116 acgttggatgagtttgcatctcctcccctg acgttggatgtgtgaattgtctgttattatgcc ttttacagagttatgtaaaaccaa rs376184 acgttggatgatccctcatgcttttccatc acgttggatgttacctgccctactggagtc caggggcttctggggtcttattct rs3778045 acgttggatggctgtgcaatgatcgaagag acgttggatgctgcccaccttctgcttct caatgatcgaagagagtgtgtattt rs3826503 acgttggatgtacagcctgcagaacagtat acgttggatgtccttgactgtagatgcacc tacagcctgcagaacagtatgcccat rs3837841 acgttggatgtcacattcttgaaagttactga acgttggatgaagccaattagaatcacttc gtttgaaagttactgaaacttaggtc rs3935570 acgttggatgaaaagaagaaagatgttacatca acgttggatggtggactacattgagatcgaga tacatcattaacactcactatcttatca rs4713740 acgttggatgaacgtggagaacctcgtatc acgttggatgaaggagaggaactgagaagc ctgcctcaggctgta rs4790151 acgttggatgctgagaatgttctctatcgg acgttggatgttcaggtgtatgacttgttg tatcggataccggca rs4790522 acgttggatgtagtgagggtttgtggttcc acgttggatgacaggaaacaacgacaggac aatgccccgctacccc rs4908563 acgttggatgcctttaatcatactccaggc acgttggatgcatccattgaagcccttctg caggctttgagcatgg rs4920564 acgttggatgcagaatatcaccaggatggc acgttggatgtcttcctggaatggcatgag ctgaggaagcccaggg rs4920566 acgttggatgtcttgcgggaatcgtgtctc acgttggatgaagagcggttcagcccaatc gtctctccccacagttc rs555607708 acgttggatgcatacgtctggcagatcttc acgttggatgcacgatcagttcaaggcaac ctctggcacctcgctctc rs643637 acgttggatgtcacagtcagccctgttatc acgttggatgtccagacgcggacttacctg cagaggccaggagacggc rs6662276 acgttggatgagagcttgtttacgatgggc acgttggatgctcaagacccctccaaattc gaggatgggctgtagacc rs706792 acgttggatggcatgcaaacggcacacat acgttggatgtgtgtgtaccatgcgcatgt cactgcacacatgcatgcc rs713598 acgttggatgcctaaactgaacacaggtgg acgttggatgagacatctgagaccccaacc ttatgagcccattgcagtt rs7305558 acgttggatgtgcctgataatgaagactcc acgttggatggcttggctgatatcggttct ccttctcctctaaatccatt rs7513755 acgttggatgggagcgtaaaagttgtactatgc acgttggatgagccccttacaacttttagaca tacagaaaacactctgacac rs7534618 acgttggatgcatgctaaatcactgttttg acgttggatgagagaaaccatttgggaggc tcactgttttgttgtaagaa rs76457020 acgttggatgccagtatcaaataacttccc acgttggatgcctcagcaggaatatttcag aagtgttttgttattgtggc rs7772436 acgttggatgtttgtgccctcttgtcctac acgttggatgttcaggccaagctgttcaag gacatcttgtcctactgtccc rs7772932 acgttggatgtcacagattttggaatggtt acgttggatgtcctttcaataaattctcgggga atctcttattttgttagctgc rs8065080 acgttggatgtcactccttaaatatggtggaac acgttggatgtggagatcacaaccaccagt aggaaatgaagcttttgaatc

137 SNP ID Forward primer Reverse promer Extend primer rs8181419 acgttggatgacaactgagatgcctctacc acgttggatgtggtgcatatgtgggcaatg gtaacagtggatgtttattga rs877610 acgttggatgtagtggcattgtccctcaag acgttggatgaataacaacagccaccctgc aatacctcaagctcccctctgc rs9368793 acgttggatgctagggtactcttaagcagg acgttggatgagaacaggagtttccagagc gaataggcatcaacctatagaa rs937035 acgttggatgagaagagcttccttaagtgc acgttggatgttgccatcaggtgtgtaccc ggggccttaagtgcatgaggaa rs937039 acgttggatgacagattctagctgcttccg acgttggatggtgagctgcagtctaagaag ttgatttcttcttctttaactca rs9380406 acgttggatggtcacataacattgagtcagaca acgttggatggactatctaaccccgtccat ggaggacccattttccacctcta rs9469690 acgttggatggtcagaaaatgctattgaatgtgt acgttggatgctgattacatttaggatgactt caaaaaggtttattcattttgac rs9469718 acgttggatgtcccactccatttctgactg acgttggatgcagttatgcatttagaatgtctga ggggaactatccaagtgcatcac rs9701796 acgttggatggcagcgtgagaacaactaac acgttggatgcaagagaggggatttgtttg gtctttctctttaaggtagaaata rs9890133 acgttggatgagatcctgtacatggacagc acgttggatgacctgcgtttggtgttgttg ttttttaagcaattatttcatcttag

138