Effect of Genetic Variation on Salt, Sweet, Fat and Bitter Taste

Andre Dias

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Nutritional Sciences University of Toronto

Andre Dias

Doctor of Philosophy

Department of Nutritional Sciences University of Toronto

2014 Abstract

Background: Taste is one of the primary determinants of food intake and taste function can be influenced by a number of factors including genetics. However, little is known about the relationship between genetic variation, taste function, food preference and intake.

Objective: To examine the effect of variation in genes involved in the perception of salt, sweet, fat and bitter compounds on taste function, food preference and consumption.

Methods: Subjects were drawn from the Toronto Nutrigenomics and Health Study, a population of healthy men (n=487) and women (n = 1058). Dietary intake was assessed using a 196-item food frequency questionnaire (FFQ) and food preference was assessed using a 63-item food preference checklist. Subsets of individuals were phenotyped to assess taste function in response to salt (n=95), sucrose (n=95), oleic acid (n=21) and naringin (n=685) stimuli. Subjects were genotyped for Single Nucleotide Polymorphisms (SNPs) in candidate genes.

Results: Of the SNPs examined in putative salt taste receptor genes (SCNN1(A, B, D, G),

TRPV1), the rs9939129 and rs239345 SNPs in the SCNN1B gene and rs8065080 in the TRPV1 gene were associated with salt taste. In the TAS1R2 gene, the rs12033832 was associated with sucrose taste and sugar intake. The rs1077242 SNP in the bitter taste receptor gene TAS2R19 was

ii associated with naringin taste and both grapefruit intake and preference. In the putative fat taste receptor CD36 the rs1761667 and rs1984112 SNPs were associated with intake of total, polyunsaturated and monounsaturated fats as well as oleic acid taste.

Conclusions: Our findings demonstrate that genetic variation is associated with differences in taste function, food preference and intake across a number of taste modalities.

iii

Acknowledgments

It has been more than a decade since I first arrived at the Department of Nutritional Sciences. Still in high school, I was amazed by the work that was being carried out, the dedication of the staff and students to their discipline, and the people who seemed larger than life. I am still amazed. The department has been my home for the last 10 years. It is where I have had some of my most joyous, and most humbling, experiences. To the hundreds of people who have built it, and created this place of learning, thank you.

This thesis would not have happened without the patience and guidance of my supervisor, Dr. Ahmed El-Sohemy. Ahmed, I first heard you speak long before I contemplated grad school. At the time I was advised to, if at all possible, be more like you. It was sound advice ;) Learning with you during these last 4.5 years has been an incredible experience. Your passion, belief in your principles, and strategic approach in all situations is inspiring. Additionally, your mentorship style is quite simply phenomenal. Your willingness to give your students as much help as necessary, while granting them the freedom and independence to chart their own course, is commendable. Few professors get this right and, in my life, this has made all the difference. In our time together you have allowed me to accomplish so much, both in and outside the lab. I will forever be in your debt for this. Thank you for a truly incredible experience.

From grade 11 till the end of my undergraduate studies I had 3 parents; my mom, my dad and Dr. Vladimir Vuksan. Dr. Vuksan, my days working with you at the Risk Factor Modification Center are by far some of my fondest. Aside from my parents, no one has had a greater impact on my development than you. You taught me to think on my feet, pushed me beyond my limits, and gave me the learning experience of a lifetime. I cannot thank you enough. Looking back, you were part of almost all my major life decisions over the last decade. You introduced me to research, were one of the primary reasons I decided to join U of T, pushed me to pursue graduate school, and have supported me as a mentor and friend ever since. Thank you. I look forward to staying in touch over the next decade and many more afterwards.

Moira and Winnie, thank you for your hard work and friendship. Without your help, and the help of your students, I don’t think my PhD work would have been possible. Working with you both at George Brown College was probably the most fun I have ever had and I will always fondly

iv remember my time there. Chef Ian, thanks for putting up with all of the inane restrictions of my work and for keeping me well fed at all times. I can’t tell you how jealous the other graduate students were when they found how well I dined. Candace, Lara, Catherine, Pia, Geremy, and Cecilia, thank you so much for all of your hard work and the numerous 5AM mornings you endured with me. I couldn’t have done it without you.

Mom and Dad, no one could ask for better parents than Nandita and I. As I get older I am able to more fully appreciate the immense effort you took to provide us with the best opportunities and push us to our full potential. From teaching me math and driving me to an array of activities, to giving me the confidence to start my own business, you have been integral in building all of the skills I have today. Thank you. Nandita, thank you for being the best big sister in the world. You have been helping me through school and expanding my horizons for as long as I can remember. Without you, I definitely wouldn’t have got here. Fidel, thanks for being such an awesome and supportive brother in law. You have been there for me since the day I met you. Rohit, we shared most of this journey together and I very grateful I got to do it with you. Thanks for being an awesome roommate, friend, and cousin.

Komal, you are by far the best girlfriend anyone could ask for. I can't count the number of times you opted to study with me in the library, rather than have a night out. You were beside me the entire way, pushing me when I was unmotivated, calming me when I was frazzled and cheering me up whenever I was down. I couldn’t have done it without you. Thank you for being so incredible and so supportive.

To my lab mates, thank you for your help, feedback, and friendship. Daiva, you’re my homegirl! Thank you for always being there for me in a pinch. Bibiana, thanks for your help and friendship. Joanne and Laura, I miss you. You made my first two years as a grad student truly amazing; the lab hasn’t been the same since you left.

To all of the people who made my time at U of T so enjoyable, thank you. Dennis, we have supervised each other, been colleagues, and, run a company together. You are one of my closest friends and definitely played a huge role in getting me here today, thank you. Jovana, you are both incredibly capable and down to earth, a rare combination. Thanks for being a great colleague and an even better friend. Elsa, you are one of the most caring and kind hearted people

I have ever come across. Thank you for your friendship and please continue to inspire the rest of us to change the world. Julie and Shannon, you opened my eyes to world I never knew existed. Thanks for your friendship and organizing LTS Sandy Lake. Chuck, you were personality of our department. No one did more to make everyone feel welcome. Conrad, Dave, Sonali, and Jake, thank you for your training and support. Without you, I would never have had the opportunity to transition to a new career.

Lastly, I would like to thank He Song. You are one of the kindest, most helpful people I know. I could never have finished this thesis without your guidance and help. Thank you.

Table of Contents

Abstract…...... ii

Acknowledgments...... iv

Table of Contents...... vii

List of Tables ...... xi

List of Figures...... xii

List of Abbreviations ...... xiii

Chapter 1 Introduction...... 1

Chapter 2 Literature Review ...... 3

2.1 General taste...... 3

2.2 Taste physiology ...... 4

2.3 Taste and food intake...... 6

2.4 Salt taste...... 7

2.4.1 Salt taste receptors ...... 8

2.4.2 Salt taste and food intake ...... 9

2.4.3 Summary...... 11

2.5 Sweet taste ...... 11

2.5.1 Sweet taste receptors...... 12

2.5.2 Sweet taste and food intake...... 15

2.5.3 Summary...... 16

2.6 Fat taste ...... 16

2.6.1 Fat taste receptors ...... 17

2.6.2 Fat taste and food intake ...... 20

2.6.3 Summary...... 22

2.7 Bitter taste ...... 22

vii

2.7.1 Bitter taste receptors ...... 23

2.7.2 Bitter taste and food intake ...... 24

2.7.3 Summary...... 25

2.8 Food intake assessment...... 25

2.9 Food preference assessment...... 28

2.10 Taste function assessment...... 29

2.11 Genotyping techniques...... 30

2.12 Rationale, Hypothesis and Objectives ...... 31

Chapter 3 Genetic variation in putative salt taste receptors and salt taste perception in humans...... 33

3.1 Abstract...... 34

3.2 Introduction...... 34

3.3 Methods...... 36

3.3.1 Subjects...... 36

3.3.2 General Protocol ...... 37

3.3.3 Dietary Assessment...... 38

3.3.4 Genotyping...... 38

3.3.5 Statistical Analysis...... 45

3.4 Results...... 45

3.5 Discussion...... 54

Chapter 4 Variation in TAS1R2, sweet taste perception and intake of sugars ...... 57

4.1 Abstract...... 58

4.2 Introduction...... 58

4.3 Methods...... 60

4.3.1 Subjects...... 60

4.3.2 Dietary Assessment...... 62

viii

4.3.3 Sensory Analysis Protocol ...... 62

4.3.4 Threshold Testing ...... 62

4.3.5 Suprathreshold Testing ...... 63

4.3.6 Laboratory Analysis and Genotyping ...... 63

4.3.7 Anthropometrics and Physical Activity...... 64

4.3.8 Statistical Analysis...... 64

4.4 Results...... 65

4.5 Discussion...... 73

Chapter 5 Genetic variation in TAS2R19 affects naringin taste sensitivity and is associated with grapefruit preference and intake...... 76

5.1 Abstract...... 77

5.2 Introduction...... 77

5.3 Methods...... 78

5.3.1 Subjects...... 78

5.3.2 Food Preference Assessment ...... 79

5.3.3 Dietary Intake Assessment...... 81

5.3.4 Bitter Taste Assessment...... 81

5.3.5 Genotyping...... 81

5.3.6 Statistical Analysis...... 85

5.4 Results...... 85

5.5 Discussion...... 90

Chapter 6 Genetic variation in CD36 is associated with oleic acid taste sensitivity and habitual fat intake...... 93

6.1 Abstract...... 94

6.2 Introduction...... 94

6.3 Methods...... 96

6.3.1 Study 1: CD36 and habitual fat consumption...... 96

6.3.2 Study 2: CD36 and Oleic acid taste ...... 99

6.4 Results...... 101

6.5 Discussion...... 107

Chapter 7 Summary, Limitations, Future Directions and Implications...... 110

7.1 Summary...... 110

7.2 The relationship between variation in putative salt taste receptors, salt taste and sodium intake...... 112

7.3 The relationship between variation in the TAS1R2 sweet taste receptor, sucrose taste and sugar intake ...... 114

7.4 The relationship between variation in the TAS2R19 bitter taste receptor, naringin taste, and both grapefruit preference and intake ...... 116

7.5 The relationship between variation in the putative fat taste receptor CD36, fat taste and fat intake...... 118

7.6 Implications...... 118

References...... 121

List of Tables

Table 2.1 Receptors responsible for each taste sensation and the potential mechanism of signal transduction ...... 6

Table 3.1 Allelic distribution of single nucleotide polymorphisms (SNPs) extracted from the Affymetrix 6.0 SNP chip within the study population ...... 40

Table 3.2 The association between candidate gene polymorphisms and both salt taste thresholds and suprathreshold taste sensitivity...... 46

Table 3.3 Sodium intake and SCNN1B/TRPV1 genotype...... 53

Table 4.1 General subject characteristics...... 61

Table 4.2 The association between TAS1R2 polymorphisms and both sucrose taste threshold and suprathreshold taste sensitivity ...... 66

Table 4.3 Dietary intake and rs12033832 genotype stratified by BMI...... 71

Table 4.4 Dietary intake and rs3935570 genotype stratified by BMI...... 72

Table 5.1 General subject characteristics...... 80

Table 5.2 TAS2R19 SNP characteristics ...... 82

Table 5.3 Odds of perceiving naringin as “high intensity” by TAS2R19 genotype ...... 86

Table 5.4 Odds of disliking grapefruit and grapefruit juice by TAS2R19 genotype ...... 88

Table 5.5 Odds of not consuming grapefruit and grapefruit juice by TAS2R19 genotype ...... 89

Table 6.1 General subject characteristics...... 97

Table 6.2 Dietary intake of by CD36 genotype ...... 103

Table 6.3 Dietary intake by combined rs1761667 and rs1984112 genotype ...... 104

Table 6.4 Oleic acid taste thresholds and suprathreshold taste sensitivity by combined rs1761667 and rs1984112 genotype ...... 106

List of Figures

Figure 3.1 Linkage disequilibrium (LD) in 541 (165M/376F) individuals for SNPs with a minor allele frequency > 15% in a) SCNN1A, b) SCNN1B, c) SCNN1G, and d) TRPV1...... 42

Figure 3.2 Comparison of a) incremental area under the taste sensitivity curve (iAUC) and b) taste intensity ratings between individuals homozygous for the A allele and carriers of the T allele for the rs239345 SNP ...... 50

Figure 3.3 Comparison of a) incremental area under the taste sensitivity curve (iAUC) and b) taste intensity ratings between individuals homozygous for the T allele and carriers of the C allele for the r rs3785368 SNP...... 51

Figure 3.4 Comparison of a) incremental area under the taste sensitivity curve (iAUC) and b) taste intensity ratings between individuals homozygous for the C allele and carriers of the T allele for the rs8065080 SNP ...... 52

Figure 4.1 Comparison of a) incremental area under the taste sensitivity curve (iAUC ± SE) and b) taste thresholds (mmol/L ± SE) between individuals homozygous for the A allele and carriers of the G allele for the rs12033832 SNP, stratified by BMI...... 68

Figure 4.2 Comparison of a) incremental area under the taste sensitivity curve (iAUC ± SE) and b) taste thresholds (mmol/L ± SE) between individuals homozygous for the T allele and carriers of the G allele for the rs3935570 SNP, stratified by BMI...... 69

Figure 5.1 Linkage disequilibrium (LD) for SNPs with a minor allele frequency > 5% in TAS2R19 ...... 83

Figure 5.2 Linkage disequilibrium (LD) between TAS2R19 SNPs (rs10772420, rs4763235) and 6 SNPs from TAS2R50...... 84

xii

List of Abbreviations

5-HT 5-hydroxytryptamine AFC Alternative Forced Choice Arg Arginine ATP Adenosine triphosphate cAMP Cyclic adenosine monophosphate BMI Body mass index CD36 Cluster of differentiation 36 CHO Carbohydrates CI Confidence interval Cys Cysteine ddNTPs Dideoxy nucleoside triphosphates DHQ Dietary history questionnaire DRKs Delayed rectifying potassium channels EDTA Ethylenediaminetetra-acetic acid ENaC Epithelial sodium channel ER Endoplasmic reticulum FFA Free fatty acids FFQ Food frequency questionnaire FPC Food preference checklist GLM General linear model gLMS General labeled magnitude scales GPCR G protein-coupled receptor GPR40 G protein-coupled receptor 40 GPR120 G protein-coupled receptor 120 HDL High-density lipoprotein iAUC Incremental area under curve Ile Isoleucine IP3 Inositol trisphosphate IP3R3 Inositol 1,4,5-trisphosphate receptor type 3 xiii

LCFA Long chain fatty acid LD Linkage disequilibrium LDL Low-density lipoprotein MAF Minor allele frequency MALDI-TOF MS Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry MET Metabolic equivalent MUFA Monounsaturated fatty acid NST Nucleus of the solitary tract ORs Odds ratios PCR Polymerase chain reaction PLCβ2 Phospholipase C β2 PROP 6-n-Propylthiouracil PTC Phenylthiocarbamide PUFA Polyunsaturated fatty acids SCNN1 Sodium channel, nonvoltage-gated 1 SE Standard error SNPs Single nucleotide polymorphisms SPINK5 Serine protease inhibitor Kazal-type 5 STS Suprathreshold sensitivity STT Suprathreshold taste sensitivity TAS1R2 Taste receptor type 1 member 2 TAS1R3 Taste receptor type 1 member 3 TAS2R19 Taste receptor type 2 member 19 TAS2R38 Taste receptor type 2 member 38 TAS2R50 Taste receptor type 2 member 50 TC Total cholesterol TRCs Taste-receptor cells TRPV1 Transient receptor potential cation channel subfamily V member 1 TRPM5 Transient receptor potential cation channel subfamily M member 5 USDA United States Department of Agriculture Val Valine WC Waist circumference xiv

Chapter 1 Introduction

Food intake regulation is a complex process mediated by a number of biological and environmental factors (Glanz et al., 1998). With increasing prevalence of diseases related to over nutrition there is considerable interest in identifying genes that predispose individuals to such disease by influencing dietary decisions. Taste is the number one modifier of individual food selection making it an important determinant in understanding how diet influences the risk of chronic disease (El-Sohemy et al., 2007). Taste is categorized into five taste modalities, which are sweet, bitter, sour, salty and umami (or savory) (Bachmanov and Beauchamp, 2007; Roper,

2007). Growing evidence suggests that fat may be the sixth taste modality (Laugerette et al.,

2005; Mattes, 2009c). Few studies to date have examined how genetic variation modifies salt, sweet, fat and bitter taste perception in humans and its potential effect on diet.

Salt taste largely refers to the taste sensation elicited by sodium chloride. Studies in rodent models have led to the identification of the Epithelial Sodium Channel (ENaC) and the

TRPV1 nonspecific cation channel as putative taste receptors (Bachmanov and Beauchamp,

2007). As such, an examination of the effect of genetic variation on salt taste should consider variation in candidate genes for these known ion channels; SCNN1(A,B,G,D) and TRPV1.

Sweet taste sensation can be stimulated by a number of compounds including natural sugars, artificial sweeteners, d-amino acids and sweet proteins (Chandrashekar et al., 2006).

However, in the context of diet, sweet taste is primarily linked with carbohydrate consumption, specifically sugars. First identified as the Sac locus, the sweet taste receptor has been shown to be a heteromeric protein structure made up of the TAS1R2 and TAS1R3 protein subunits.

TAS1R2 is unique to sweet taste sensation while the TAS1R3 protein is also involved in umami

2 perception. Single nucleotide polymorphisms (SNPs) in the TAS1R2 and TAS1R3 genes, which code for these two proteins, have been associated with sweet taste perception and habitual sugar consumption (Eny et al., 2010; Fushan et al., 2009). However, no studies to date have examined the two concurrently within the same population, leaving the link between changes in taste and intake unexplored. An examination of the effect of genetic variation on sweet taste should focus on the TAS1R2 gene with the aim of seeing how variants modify both taste and intake.

Bitter taste can be stimulated by thousands compounds that interact with at least 25 dedicated receptors (Meyerhof et al., 2010). Multiple ligands have been identified for a number of the receptors but some still remain orphaned (Meyerhof et al., 2010). The impact of variation in TAS2R genes on perception of bitter compounds has received considerable interest but the majority of research has focused on the TAS2R38 gene and its ligands, phenylthiocarbamide

(PTC) and 6-n-propylthiouracil (PROP) (Kim et al., 2003). Recent work has shown that the

TAS2R19 may be associated with grapefruit liking and perceived bitterness though little is known about how this affects intake and what ligand may be involved (Hayes et al., 2011).

Additional work is required to better characterize this relationship.

Recent studies have presented a strong case for the existence of a “fatty taste”(Mattes,

2009a). The postulate that free fatty acids (FFAs) are detected by gustatory cues is well supported and the CD36 trans-membrane protein has been shown as the likely sensor of long chain fatty acids in a number of rodent studies (Mattes, 2009a). Our lab has observed that two

SNPs in the CD36 gene are significantly associated with intake of polyunsaturated fatty acids

(PUFA) in Caucasians (Toguri, 2008). This association may be caused by differences in PUFA taste, elicited by changes in gustatory responses attributed to these SNPs and warrants further investigation.

Chapter 2 Literature Review

2.1 General taste

The gustatory system regulates taste perception and helps individuals evaluate the nutrient content of food and discriminate between safe and harmful foods (Bachmanov and Beauchamp, 2007; Chandrashekar et al., 2006). Taste may also aid in reflexes such as salivation, pre-absorptive insulin release and oral-motor reflexes such as gagging, chewing or swallowing (Bradley et al., 2005; Ding et al., 2003; Teff, 2000). In humans, taste contributes to the overall enjoyment of a meal.

There are five established taste modalities in humans: sweet, savoury (umami), salty, bitter and sour (Bachmanov and Beauchamp, 2007; Roper, 2007). Recently, several studies have provided evidence of a sixth taste, fat (Laugerette et al., 2005; Mattes, 2009c; Stewart et al., 2010). Each taste is thought have developed in response to different evolutionary pressures. Sweet taste allows for the detection of calorie-rich foods; savoury confers the ability to identify foods rich in amino acids; salty taste helps maintain dietary electrolyte balance; sour taste may provide warning against spoiled or unripen foods; bitter taste helps identify potential toxic or poisonous substances primarily found in plants; fat taste allows the detection of high-energy “fatty” foods (Bachmanov and Beauchamp, 2007; Chandrashekar et al., 2006; Kim et al., 2004b).

The ability to detect and differentiate between these different taste sensations is conferred by the anatomical units of taste detection, taste-receptor cells (TRCs), whose activities may be heavily impacted by variation in receptor genes (Chandrashekar et al., 2006). Genetics is an important determinant of variation in perceived taste intensity and food preference, both of which have been hypothesized to affect disease risk by augmenting satiety and influencing what people choose to eat (Glanz et al., 1998). Consequently, an understanding of the genetic causes of these variations may allow us to predict individual taste function and potentially dietary patterns followed by such individuals.

2.2 Taste physiology

Taste is mediated by TRCs that cluster in groups of 50 -100 to form taste buds (Chandrashekar et al., 2006). These buds are distributed inside specialized folds and protrusions across the tongue called papillae. The three types of papillae on the tongue are circumvallate, foliate and fungiform papillae and they can also be found on the palate, oropharynx, larynx, epiglottis and upper esophagus (Bachmanov and Beauchamp, 2007; Chandrashekar et al., 2006; Sugita, 2006). There are approximately 3 to 18 circumvallate papillae with a V-shaped distribution at the root of the tongue, each containing about 250 taste buds (Chandrashekar et al., 2006; Meyerhof, 2005). Approximately 2 to 19 alternating ridges and crevices containing about 120 taste buds per ridge make up the foliate papillae, which are located on the back edges of the tongue (Chandrashekar et al., 2006; Meyerhof, 2005). Approximately 200-300 fungiform papillae with 1 to 3 taste buds each are present on the tongue and the highest density exists at the tip of the tongue (Chandrashekar et al., 2006; Meyerhof, 2005).

Contrary to the popular belief that different tastes are detected by different regions of the tongue, molecular and functional studies have shown that sweet, savoury, salty, bitter and sour can be detected by the same regions of the tongue (Chandrashekar et al., 2006). Taste receptors are expressed in microvilli on the apical side of taste receptor cells, though for the sour and salt taste it is thought that ions can traverse tight junctions between TRCs and enter cells via receptors expressed on the basolateral membrane in addition to those on the apical side of the cell (Bachmanov and Beauchamp, 2007; Chandrashekar et al., 2006; DeSimone and Lyall, 2006). A given taste is perceived when taste inducing molecules (tastants) come into contact with various receptors specific for each type of taste expressed on TRCs and cause cell depolarization and an influx of extracellular cellular calcium via voltage gated channels on the basolateral cell membrane (Bachmanov and Beauchamp, 2007; Chandrashekar et al., 2006). The influx of calcium and cellular depolarization is thought to cause the exocytosis of vesicles containing ATP (adenosine triphosphate) and other neurotransmitters, which can bind to afferent nerves causing innervations (Chaudhari and Roper, 2010). These innervations transmit sensory information to the gustatory cortex of the brain through the chorda tympani branch of the afferent facial nerve (VII), the glossopharyngeal (IX) and the vagus nerve (X) (Bachmanov and Beauchamp, 2007).

Cells present in the taste bud can be further divided into 3 categories based on the proteins they express (Chaudhari and Roper, 2010). Type I cells are the most abundant cell type in the taste bud and they express a membrane bound nucleotidase that hydrolyzes extracellular ATP, a neurotransmitter in the taste bud (Bartel et al., 2006; Finger et al., 2005). As a result, Type I cells may be responsible for controlling the length and location of synaptic transmissions (Chaudhari and Roper, 2010). Type II cells contain taste-specific receptors in their plasma membrane that can bind sweet, bitter and savory compounds (Chaudhari and Roper, 2010; Tomchik et al., 2007). Moreover, each Type II cell contains G protein–coupled receptors specific for only one type of taste sensation (Tomchik et al., 2007). Currently known receptors for each taste and the mechanisms by which they transduce signals are summarized in Table 2.1. Type III cells express synaptic proteins and form synapses with nerve terminals directly. They have been proposed to participate in sour taste detection directly as well as receive signal inputs from other cell types to respond to sweet, bitter, savory and salty tastes (Chaudhari and Roper, 2010; Tomchik et al., 2007). It is interesting to note that even though Type III cells show many characteristics of neurons, they are not a pure population of one cell type (Roberts et al., 2009; Tomchik et al., 2007).

Taste Receptor Signal transduction Sweet T1R2, T1R3 GPCR (heterodimer): elevates intracellular Ca2+ through TRPM5 channel leading to ATP release Bitter T2Rs (a family of GPCR (multimer state unclear): elevates ~25 proteins) intracellular Ca2+ through TRPM5 channel leading to ATP release Savory T1R1, T1R3 GPCR (heterodimer): elevates intracellular Ca2+ through TRPM5 channel leading to ATP release Salty ENaC ENaC Na+ ion channel: allows Na+ entry into cell causing cellular depolarization Sour Intracellular Proton-sensitive K+ channel: becomes blocked by acidification increased intracellular H+ due to increased organic acid and causes cellular depolarization Fat GPR120, GPR40, GPCR (putative mechanism): CD36 binds to and CD36 delivers FA to GPR120 or GPR40, which may transduce signal similar to other GPCRs.

Table 2.1 Receptors responsible for each taste sensation and the potential mechanism of signal transduction.

2.3 Taste and food intake

Taste is modulated by both the environmental and genetic factors and is thought to be one of the most important determinants of food preferences and consumption (El-Sohemy et al., 2007; Garcia-Bailo et al., 2009b; Glanz et al., 1998). Level of industrialization and economic conditions are strong environmental factors that influence food consumption between populations, and account for the large differences in food intake seen between developed and developing countries (AICR, 2007). At the same time, gene variation has also been shown to be associated with differences in taste function and potentially food preference and consumption patterns. SNPs in many taste receptor genes have been linked to variability in taste perception, food preferences, dietary habits, nutritional and health status and risks for chronic disease (Garcia-Bailo et al., 2009b).

Two broad categories of tastes exist; appetitive and aversive. Sweet taste is inherently recognized as pleasant, possibly due to evolutionary pressures to select foods high in energy-rich carbohydrate content (Hladik et al., 2002). Similarly, salt taste is also generally viewed as appetitive. Given the indispensable role that NaCl plays in various physiological processes, it has

7 been proposed that evolution may favour the ability to taste and consume varying amounts of salt in response to environmental and dietary conditions (Wise et al., 2007). It is important to note that at high enough concentrations of given tastant both sweet and salt tastes can be aversive. Unlike sweet and salt taste, bitter taste is generally aversive and has been shown to negatively affect food preferences in both children and adults (Anliker et al., 1991; Dinehart et al., 2006; Drewnowski et al., 2000; Drewnowski et al., 1999; Intranuovo and Powers, 1998). The nature of fat taste is more complex. While it has been shown both in animal models and among children that a spontaneous preference for high-fat foods exists, suggesting an evolutionary pressure to select these foods, foods containing high levels of FFAs are generally repulsive to human, likely due to their involvement in food deterioration (Hansen and Rose, 1996; Heilmann and Hummel, 2004; Manabe et al., 2010; Rankin and Mattes, 1996; Refsgaard et al., 2000). Given that evidence suggests that FFAs are the tastants that allow for the detection of fat by the gustatory system, it is unclear whether fatty acids are desirable at some concentrations and not others or if the appetitive nature of high fat foods is driven by other characteristics (eg, texture)(Mattes, 2009a).

2.4 Salt taste

Sodium chloride, also known as salt, is an ionic compound with the chemical formula NaCl. When salt enters the oral cavity, sodium and chlorine dissociate within an aqueous environment and free sodium ions passively flow through sodium-specific and non-specific channels into TRCs causing cell depolarization and a cascade of steps leading to salt taste perception (Chandrashekar et al., 2006; DeSimone and Lyall, 2006). Patch clamp studies have suggested that Type I cells play a role in the salty taste sensation; however, the exact cell type(s) responsible for salt taste remains unclear (Chandrashekar et al., 2010; Chaudhari and Roper, 2010; Vandenbeuch et al., 2008). In addition, a number of other chlorides are also known to cause salty tastes (Boughter and Bachmanov, 2007). Salty taste’s hedonic attribute appears to be the most important factor driving its consumption in humans (Henney et al., 2010). Salt can improve the palatability of many foods; however, adding salt beyond a certain point begins to reduce its pleasantness (Henney et al., 2010). Substantial individual differences exist in the preferred level of saltiness for each particular food (Henney et al., 2010). Since humans consume substantially more sodium in their diets than is physiologically required, the preference for salt is

8 considered primarily due to learning and some even argue that the current human salt consumption is akin to addiction (Morris et al., 2008).

2.4.1 Salt taste receptors

Two channels have been associated with salty taste, the Na+-specific epithelial sodium channels (ENaC) and the non-specific cation channel TRPV1. In rodents, ENaCs located on apical TRC membranes in fungiform papillae are made up of two alpha, one beta, one delta and one gamma subunit and are thought to play an important role in the perception of Na+ (Heck et al., 1984). Na+ flows passively through ENaCs on the apical membrane of TRCs causing cell depolarization, release of neurotransmitters and a salt taste sensation (DeSimone and Lyall, 2006). Convincing evidence in support of ENaC as a sodium specific taste receptor protein is that, in various species, taste nerve responses to NaCl are significantly inhibited by amiloride, a known ENaC blocker without an effect on responses to stimuli of other taste modalities (Stewart et al., 1997). Amiloride suppresses salt taste intensity in humans reportedly by about 20% (Feldman et al., 2003; Smith and Ossebaard, 1995). This is in stark contrast to what is found in rodents where the majority of salt taste is amiloride sensitive. It would seem that there is a degree of variability in the human population with respect to amiloride sensitivity of salt taste as reported by Feldman et al., where the amount by which amiloride reduced lingual surface potential with increasing NaCl concentrations was on average approximately 20% but had a range of 5-40% (Feldman et al., 2003). The cause of these inter-individual differences are not known but it is possible that genetics may play a role. In rats, Shigemura et. al. showed that a SNP in the alpha subunit coded by the gene SCNN1A is responsible for a modification in amiloride sensitive salt taste indicating that polymorphisms in this gene might affect the degree to which ENaCs regulate salt taste and making it interesting to examine the effect of SNPs in this gene within human populations (Shigemura et al., 2008). More recent studies in mice have demonstrated that salt sensation is mediated by a dedicated population of cells that express ENaCs. ENaC knockdown mouse models lack sodium attraction and salt behavioral responses, which provides definitive evidence that ENaC is the primary receptor for salty taste sensation among mice (Chandrashekar et al., 2010).

Since the exact cell population responsible for salty taste has yet to be fully identified, the precise molecular signal transduction mechanism for this taste is currently unclear (Chaudhari

9 and Roper, 2010). It is currently postulated that Na+ entry into the TRCs leads to cell depolarization, which in turn causes the release of neurotransmitters into the intercellular space to activate afferent nerves (Chaudhari and Roper, 2010; Henney et al., 2010). Interestingly, recent study shows that ENaCα appears to be expressed in cells containing Car4, a marker for sour-sensing cells, while it is excluded from cells expressing TRPM5, the marker for the bitter, sweet and savory-sensing cells (Chandrashekar et al., 2010). ENaCβ, on the other hand, is expressed in TRPM5-positive cells while being excluded from Car4-expressing cells. Adding to the complexity, ENaCβ can also be observed in cells lacking both Car4 and TRPM5 markers (Chandrashekar et al., 2010).

The amiloride-insensitive vanilloid receptor, TRPV1 responds to various cations, including Na+, K+, NH4+ and Ca+2, and has been proposed to play a role in salty taste perception in humans (DeSimone and Lyall, 2006). Part of the amiloride insensitive NaCl response is inhibited by TRPV1 inhibitors and the effects of temperature and vanilloids on the amiloride- insensitive NaCl responses are additive, lending evidence to the fact that TRPV1, which is heat and vanilloid sensitive, is involved in salt taste (DeSimone and Lyall, 2006). The strongest evidence that TRPV1 is involved in salt taste sensation was obtained when TRPV1 knockout mice were shown to lack amiloride insensitive NaCl responses, while control mice displayed normal levels of afferent nerve innervations (Lyall et al., 2004). However, in absolute threshold detection tests, TRPV1 knockout and wild-type mice detected NaCl at similar concentrations and preference tests of NaCl solutions with amiloride versus water revealed that TRPV1 knockout mice actually preferred the salty tasting solution over water compared to wild-type animals (Ruiz et al., 2006; Treesukosol et al., 2007). This indicates that there are likely other channels involved in sodium taste in addition to TRPV1 and ENaCs.

2.4.2 Salt taste and food intake

Over the last 50 years it has become increasingly clear that excess consumption of dietary sodium causes increases in blood pressure and raises an individual’s risk of developing cardiovascular diseases (Dahl, 2005; Havas et al., 2007; Turnbull, 2003). However, sodium is also an essential micronutrient and is required for maintaining physiological electrolyte balance as well as is needed for blood pressure/volume regulation and water homeostasis (Chandrashekar et al., 2006; Dahl, 2005). The question of how much sodium is required in the diet is often

10 contentious but Health Canada states an adequate intake is between 1000-1500 mg (1g sodium is found in approximately 2.5 grams of salt) a day with an Upper Limit of 2300 mg (Health Canada, 2009). On the other end of the spectrum it has been shown that individuals consuming as little as 250-375 mg of sodium per day remain healthy for observation periods of as long as 2-5 years (Dahl, 2005). It is worth noting that in Canada 85 % of men and 60-80 % of women exceeded the upper limit of consumption with no deficiency reported (Dahl, 2005). Given that physiological need is rarely an issue, the pertinent question becomes how much salt do individuals consume and why they consume these levels. Salty taste preference exceeding physiological need has been proposed to be primarily or exclusively due to learning and some even suggest that this is an addiction (Dahl, 1972; MacGregor and de Wardener, 1998). However, others have argued that while learning plays a role, there is an evolutionary pressure to cause humans and other animals to have an inherent propensity for salty taste even when sodium is in excess (Beauchamp, 1991; Denton, 1982). Consistent with the learning theory, infants that consume low sodium diet during the first 6 months of life have lower blood pressure after 15 years, perhaps indicationg that early exposure determines future intake (Geleijnse et al., 1997). Moreover, there is evidence that true sodium depletion during late fetal period or early infancy may permanently increase preference for sodium later in life (Beauchamp, 1991; Leshem, 2009). It has also been reported that children on average have stronger preference for salty taste than adults; however, the underlying mechanisms of this phenomenon is currently unknown (Beauchamp et al., 1990; Beauchamp and Cowart, 1990). Interestingly, preferences for salt can be modulated later in life by shifting dietary sodium content. For instance, human subjects on low sodium diet initially dislike low sodium food; however, they eventually accept the low sodium diet and find previous sodium content too salty (Blais et al., 1986; Mattes, 1997). As expected, this shift can also go in the other direction where the subject consumes increasing amounts of sodium (Bertino et al., 1986). Together these findings suggest significant plasticity exist in salty taste preference, which can be drastically re-shaped by the environment.

Studies of distinct populations report that Inuit people observing a traditional diet consume an average of less than 4 g/d of salt, Marshall Islanders about 7 g/d, white male Americans about 10 g/d, Southern Japanese farmers and laborers about 14 g/d and Northern Japanese farmers about 26 g/d (Dahl, 2005). This wide variation in consumption could be dismissed as a cultural or environmental phenomenon had it not been for equally large disparities between individuals within the same communities (Dahl, 2005). The fact that some people

11 habitually eat very little salt while others consume it in copious amounts leads to the question of what causes these inter-individual differences. Taste has been shown to be a mediator of consumption and subsequently it has been hypothesized that part of the variation in sodium intake might be caused by genetic differences associated with sodium induced salty taste (Chandrashekar et al., 2006; Garcia-Bailo et al., 2009b).

2.4.3 Summary

When examining the effect of genetic variation on salt taste, it would seem that there are two channels, ENaCs and TRPV1, which are known to have a role in taste. As such, an examination of genetic variation on salt taste should look at variation in candidate genes for these two known channels, SCNN1A, SCNN1B, SCNN1G, SCNN1D and TRPV1.

2.5 Sweet taste

A number of compounds, including sugars, artificial sweeteners, certain amino acids and proteins can be perceived as sweet (Boughter and Bachmanov, 2007; Nelson et al., 2001). Sweet tasting natural sugars include glucose, fructose, sucrose; artificial sweeteners include saccharin, acesulfame-K and aspartame (Garcia-Bailo et al., 2009b). Certain amino acids, including glycine, D-phenylalanine, D-tryptophan, L-proline and L-glutamine can also be perceived as sweet (Boughter and Bachmanov, 2007). The physiological relevance and evolutionary importance of being able to perceive non-sugar compounds as sweet remains unclear (Garcia-Bailo et al., 2009b). Sweet substrates are inherently perceived as pleasant and are clustered separately from bitter taste on taste receptor cells in humans. It is thought that the development of sweet taste sensation occurred to allow individuals to be able to pick out foods high in carbohydrates, which are usually rich in energy (Hladik et al., 2002). A number of lines of evidence exist that indicate that sweet taste is modified by genetic difference between individuals. Examining facial responses among newborn infants in response to sweet solutions shows that a pleasurable response to sweet stimuli is present shortly after birth, suggesting a genetic contribution to sweet taste perception (Berridge and Robinson, 2003). Looking at sweet taste, there is a long history of documented differences in detection threshold between individuals (Blakeslee and Salmon, 1935; Henkin and Shallenberger, 1970). Work in groups of monozygotic and dizygotic twins has

12 shown that the additive genetic contribution to the inter-individual differences in the discrimination thresholds of sweet solutions is approximately 33%. Looking at intake, work in the same groups has shown that the additive contribution to consumption frequency of sweet foods is about 53% (Keskitalo et al., 2007). Furthermore, it is possible that the use of a high concentration (20%) sucrose solution in this study prevented investigators from examining individuals at other extreme of taste function, detection thresholds, potentially ameliorating their findings and underestimating the true contribution of genetics to sweet taste.

The sweet-sensing taste receptors TAS1R2 and TAS1R3 form heterodimers on the plasma membrane of Type II cells, bind to sugars, sweeteners and sweet tasting proteins, transduce the sweet taste signal, cause cellular depolarization and release of neurotransmitter, ATP (Chaudhari and Roper, 2010). In addition to the tongue, the sweet taste receptors are also present in the gastro-intestinal tract (Mace et al., 2007; Young et al., 2009), the pancreas (Nakagawa et al., 2009) and the hypothalamus (Ren et al., 2009), which are energy homeostatic tissues and may secrete hormones to modulate food intake (Zheng and Berthoud, 2008).

2.5.1 Sweet taste receptors

The existence of a sweet-specific taste receptor was proposed in the early 1970’s. In 1974, the Sac locus was hypothesized to be responsible for sweet taste and saccharin preference (Fuller, 1974). The discovery of the gene responsible for the association between the Sac locus and sweet taste did not occur until 25 years later when a novel G-protein coupled receptor family, TR1, was discovered in rats and humans (Hoon et al., 1999). In 2001, the T1R3 receptor (TAS1R3 in humans, and Tas1r3 in rat) was identified as the receptor responsible for the saccharin preferring phenotype by a number of groups (Bachmanov et al., 2001; Kitagawa et al., 2001; Max et al., 2001; Montmayeur et al., 2001; Nelson et al., 2001; Sainz et al., 2001). The role of the TAS1R3 gene in sweet taste perception in general was tested by introducing the TAS1R3 gene into non- taster mice. The transgenic mice expressing the TAS1R3 gene showed similar preferences for saccharin and sucrose as the control taster mice, while the non-transgenic siblings did not show preference to the sweetener or sucrose (Nelson et al., 2001). Moreover, in vitro heterologous assays showed that sugars can cause receptor activation in cells co-expressing T1R2 and T1R3 as measured by increased intracellular calcium levels, strongly suggesting that the T1R2/T1R3 heterodimer is the major receptor involved in sweet taste sensing (Nelson et al., 2001).

Interestingly, knockout mice lacking either TAS1R3 or TAS1R2 were still able to detect natural sugars at very high concentrations while double-knockout mice lacking both genes cannot sense natural sugars even at high concentrations (Zhao et al., 2003). In a similar knockout study, mice lacking TAS1R3 showed no preference for artificial sweetener and reduced but measurable preferences and nerve responses to sugar compounds (Damak et al., 2003). Together, these results suggest that there are TAS1R3-independent pathways for sugar detection. Based on these studies, it has been proposed that either heterodimerization is not necessary for sweet taste sensing and TAS1R2 can function on its own or there is a currently unknown receptor to facilitate the taste detection in the absence of TAS1R3 in conjunction withTAS1R2 (Chaudhari and Roper, 2010; Damak et al., 2003; Zhao et al., 2003).

Type II cells co-expressing TAS1R2 and TAS1R3 are found in the fungiform papillae on the human tongue. Interestingly, some cells naturally only express TAS1R2 but not TAS1R3, which could explain why in TAS1R3 knockout mice the ability ofr sweet taste detection is not completely abrogated (Liao and Schultz, 2003). Upon binding to a sweet tastant to the T1R2/T1R3 complex, GPCR (G protein-coupled receptor )activates G proteins, such as α- gustducin, which may transduce the signal into the cell through two distinct mechanisms (McCaughey, 2008). It has been proposed that the activated G protein turns on adenylate cyclase to generate elevated levels of cyclic AMP (cAMP), which can cause inhibition of K+ channels via phosphorylation by protein kinase A (McCaughey, 2008). Previous studies have shown that taste bud cells do express both adenlate cyclases, which synthesize cAMP, and phosphodiesterases, enzymes that degrade cAMP, indicating TRCs are capable of regulating cAMP levels (Abaffy et al., 2003; McLaughlin et al., 1994; Moriyama et al., 2002; Spickofsky et al., 1994; Trubey et al., 2006). Indeed, a more recent study using isolated mouse vallate taste cells showed that elevated cAMP level can indeed cause Ca2+ influx and cellular depolarization (Roberts et al., 2009). Moreover, treatment with protein kinase A inhibitor or removal of extracellular Ca2+ prevented this depolarization, demonstrating the importance of protein kinase A and Ca2+ in this signal transduction process (Roberts et al., 2009). The second signal transduction pathway involves the release of G protein subunits from the GPCR upon binding to sweet substrate. The released G protein subunits interact with and activate a phospholipase,

PLCβ2 (phospholipase C β2), which stimulates the synthesis of inositol trisphosphate (IP3) (Rossler et al., 1998). Knockout mice lacking PLCβ2 have severely reduced sweet taste sensitivity, even though the ability to detect sweet taste was not completely abrogated (Dotson et

al., 2005; Zhang et al., 2003). IP3 produced by PLCβ2 can open up IP3R3 (Inositol 1,4,5- trisphosphate receptor type 3) channel on the ER causing the release of Ca2+ into the cytosol (Roper, 2007; Simon et al., 2006). The intracellular Ca2+ can cause the opening of TRPM5 channel on the plasma membrane, which can cause cellular depolarization in TRCs (Liu and Liman, 2003). The increased Ca2+ and depolarization cause the release of ATP and possibly other neurotransmitters into the extracellular space through hemichannel pores (Huang and Roper, 2010; Huang et al., 2007; Romanov et al., 2007). Interestingly, even though it is widely accepted that these two pathways are involved in sweet taste sensing, the relative importance of each pathway remains controversial. Some argue that the cAMP pathway is the dominant path in sensing natural sugars and the PLC pathway is only important for detecting sweeteners (McCaughey, 2008). Others reason that because removal of PLC substantially reduces sweet taste sensitivity in mice, the PLC pathway is dominant (Chaudhari and Roper, 2010; Dotson et al., 2005; Zhang et al., 2003). Regardless of how the cellular depolarization is generated, the released neurotransmitters cause a signal to be sent to the brain through cranial nerves VII, IX or X (Reed et al., 2006). The brain areas that these nerves target include the thalamic gustatory areas, nucleus of the solitary tract (NST) and parabrachial nucleus (Reed et al., 2006).

Interestingly, sweet taste perception can be further modulated by hormones, such as leptin (Kawai et al., 2000). Leptin is a hormone produced by adipose tissues and it can modulate sweet taste sensitivity in both human and animal models (Horio et al., 2010). Both neural and licking responses to sucrose become attenuated after injection of leptin in mice, while the same injection had no effect in mice with dysfunctional leptin receptor (Kawai et al., 2000). Injection of leptin into mice that cannot make endogenous leptin also reduced licking reposes to sucrose (Shigemura et al., 2004). In human subjects, higher levels of leptin have also been associated with higher recognition threshold (less sensitive) for sweet taste (Nakamura et al., 2008). This negative impact on sweet taste sensitivity has been found to be caused by leptin’s ability to open outward K+ channels and cause taste cell hyperpolarization, which prevents the them from releasing neurotransmitters to elicit a detection signal (Kawai et al., 2000). In addition, glucose may also indirectly contribute to the modulation of sweet taste perception by decreasing the expression level of TAS1R2 in the hypothalamus and jejunum (Ren et al., 2009; Young et al., 2009).

2.5.2 Sweet taste and food intake

The innate human preference for sweet-tasting food is prevalent across all age groups and cultures (Popkin and Nielsen, 2003). Evolutionary pressures likely drove humans to perceive foods with high sugar contents as pleasant, because these foods are more energy rich (Drewnowski et al., 2012). This is shown by the fact that infants and young children innately prefer sweet taste. Interestingly, sweet taste may have pain-reducing effects in infants, as sweet- tasting solution administered in infant’s mouth can reduce responses to painful stimuli, further reinforcing its value (Segato et al., 1997). The low cost, ready availability and high hedonic appeal of added sugars in our food and beverage supplies led to the rising concern that high consumption of added sugars was the cause of obesity epidemics around the world (Drewnowski et al., 2012; Popkin and Nielsen, 2003). Consistent with this, the consumption of added sugars in various foods have risen among all age groups in the United States, which was accompanied by rising incidences of obesity in children, adolescents and adults (Drewnowski et al., 2012; Guthrie and Morton, 2000; Wang et al., 2008). In the United States, children in the 2-5 age group on average consume an estimated 60g/day added sugars (16% of daily energy intake) and the 6-11 age group consumed an estimated 90g/day added sugars (19% of daily energy intake). A significant portion of added sugars are ingested in the form of sugar-sweetened beverages, including sodas and flavored milk (Bellisle and Drewnowski, 2007). In Canada, men and women on average consume 115 g/d and 92 g/d of sugars respectively, according to the Canadian Community and Health Survey (CCHS, 2004).

Total consumption of sugars is influenced by a number of contributing factors. Studies have shown that even though preference for sweet taste is universal for humans, the preferred intensity is modulated by many factors such as gender, age, race and genetics (Collaku et al., 2004; Drewnowski et al., 2012). On average, adults have lower preferred sweet intensity than children and adolescents; men have higher preferred sweet intensity than women (Hayes and Duffy, 2008; Monneuse et al., 1991; Pepino and Mennella, 2005). The phenotype that a person prefers high sweet intensity for a large number of foods and beverages over savory taste has also been reported (Reed and McDaniel, 2006). It is currently still under debate whether adiposity and body size is associated with preference for high sweet intensity. Some studies suggest that body mass index (BMI) is related to sugar consumption among Canadian adults (Merchant et al., 2009); however, others provide evidence that obese individuals prefer high-fat foods, which may

16 or may not contain high sugar contents (Drewnowski et al., 1985; Salbe et al., 2004). Members of the Felidae (cat) family, which are obligate carnivores, do not have neural responses to sugars and have no preference for sweet tasting solutions (Li et al., 2005). This is due to a micro- deletion resulting in an early stop codon in the TAS1R2 gene, which led to the lack of TAS1R2 expression in taste tissue (Li et al., 2005). Both the TAS1R3 genomic sequence and expression were normal when compared to other animal models (rodent and dog) and humans (Li et al., 2005). In addition, given the observation that TAS1R2 is down regulated by glucose, the TAS1R2 gene is an intriguing candidate that may contribute to the modulation of sugar intake (Ren et al., 2009; Young et al., 2009).

Two studies thus far have examined the effect of gene variants in the TAS1R2 on sweet taste or sugar intake. Fushan et al., used DNA sequencing to look at the effects of all detected SNPs within the TAS1R2 gene on sweet taste in 144 individuals (Fushan et al., 2009). Interestingly, they did not find any significant variants in the TAS1R2 gene that effect sweet taste outcomes. In a recent study, Eny et al examined the effect of two missense SNPs, Serine9Cysteine (rs9701796) and Isoleucine191Valine (rs35874116), in the TAS1R2 gene on sweet taste using sugar intake as a surrogate measure of taste (Eny et al., 2010). Grouping subjects by BMI, they found that in those with a BMI greater than 25, carriers of the Valine191 amino acid consumed significantly lower amounts of sugars than the homozygous Isoleucine group. A possible explanation for this is that the Isoleucine191Valine mutation affects taste perception and subsequently intake though this has not yet been tested in animal or human models.

2.5.3 Summary

Together, the studies in animals and humans suggest that genetic variation in the sweet taste receptor TAS1R2 may influence both taste and intake outcomes. Therefore, work presented in this thesis aims to identify the effects of variation in these genes on sweet taste and sugar intake.

2.6 Fat taste

Fat is the most energy-dense component of the human diet and it is an extremely important contributor to the texture, aroma and flavour for a large variety of foods. Foods with

17 high energy and high fat content are generally perceived as the most palatable (Drewnowski, 1997a; Drewnowski, 1997c). Rodents prefer a high-fat diet over a low-fat one, when the choice is provided, suggesting animals have a natural preference for dietary fat (Hamilton, 1964). Fatty food preference tests have shown that humans may have a comparable preference for high-fat diet (Mela and Sacchetti, 1991). The taste, smell and hedonic attributes of fat contribute to its appeal, which leads to the close association between palatability and energy density of foods (Drewnowski, 1997c). Traditionally, fat perception is thought to be achieved through smell (olfactory), texture sensation (mechanical) and post-ingestive cues (Drewnowski et al., 1989; Laugerette et al., 2007). Initial detection of fat is through olfactory perception of fat-soluble molecules through the nose in a process termed orthonasal olfaction (Heilmann and Hummel, 2004; Mattes, 2005). Once inside the mouth, olfactory detection can still occur through a process called retronasal olfaction, which may have different brain pathways than orthonasal olfaction (Mattes, 2005). The mechanical perceptions of food can be classified based on many criteria, such as smoothness and viscosity, which may be modulated by fat content (Brandt et al., 1963; Montmayeur and le Coutre, 2010). Removal of these sensations did not completely abolish the detection of dietary fat, suggesting additional fat-sensing mechanisms exist (Fukuwatari et al., 2003). Indeed, studies have shown that fat can be detected through chemoreception and a number of potential fat receptors, such as CD36, GPR40 (G protein-coupled receptor 40) and GPR120 (G protein-coupled receptor 120) have been identified in TRCs (Cartoni et al., 2010; Gilbertson et al., 2005; Laugerette et al., 2005; Sclafani et al., 2007a; Wellendorph et al., 2009). CD36 was found highly expressed in the circumvallate papillae and a wide range of the other cell types (Degrace-Passilly and Besnard, 2012; Fukuwatari et al., 1997; Laugerette et al., 2005). Exposure to dietary fatty acids can lead to increased intracellular Ca2+ in TRCs containing CD36 (Abdoul- Azize et al., 2013; Laugerette et al., 2005; Sclafani et al., 2007a).

2.6.1 Fat taste receptors

In 1997, Fukwatari et al showed that CD36 was present on the apical surface of taste bud cells (Fukuwatari et al., 1997). CD36 is a glycosylated integral membrane protein with a wide range of functions (Degrace-Passilly and Besnard, 2012; Silverstein and Febbraio, 2009; Su and Abumrad, 2009). It has long been recognized that CD36 plays a role in facilitating the transport of FFAs across the plasma membrane. Interestingly, a recent study has suggested that CD36 may enhance fatty acid uptake through increasing intracellular esterification instead of directly

18 catalyzing the translocation of fatty acid across the plasma membrane (Xu et al., 2013). In 2005, Laugerette et al proposed that CD36 is responsible for detecting fat taste in the oral cavity (Laugerette et al., 2005). CD36 contains a large extracellular hydrophobic loop flanked by two transmembrane domains and two short cytoplasmic tails (Vega et al., 1991). It is believed that this molecular structure would allow CD36 to facilitate sensing of extracellular FFA (Gilbertson et al., 1997). The heavily glycosylated extracellular domain of CD36 has been proposed to be important for its wide range of functions in different cell types (Armesilla and Vega, 1994; Vega et al., 1991). Genetic knockout experiments in mice suggested the putative role of CD36 in sensing fat taste. Wildtype and CD36 knockout mice (targeted deletion) were tested in a two bottle preference test over a 48 hour time period (Laugerette et al., 2005). The treatment bottle contained 2% linoleic acid solution emulsified with 0.3% xanthan gum, while the control bottle contained 0.3% xanthan solution to mimic the texture of fat (Laugerette et al., 2005). The wildtype mice preferred the bottle containing 2% linoleic acid over the control bottle, which have solution with comparable texture (Laugerette et al., 2005; Takeda et al., 2001). The CD36 knockout mice on the other hand consumed solutions equally from both control and fatty acid- containing bottles, indicating CD36’s importance in fat taste preference (Laugerette et al., 2005). To control for the effects of post-ingestive cues, wildtype and CD36 null mice were 1h water- restricted or 12h fasted. Consistent with the previous experiment, wildtype mice showed immediate preference for the fat-containing bottle over the control bottle, while the CD36 knockout mice did not. To confirm that this CD36-dependent fat preference was not a liquid- specific effect, tests using solid foods containing 5% linoleic acid or 5% paraffin (control) were conducted on wildtype and CD36 null mice. Similar to the liquid experiments, wildtype mice showed clear preference for the linoleic acid-containing diet over the control diet, which had a similar texture, while CD36 mice did not (Laugerette et al., 2005). To ensure that CD36 knockout did not disrupt taste-sensing in general, wildtype and CD36 null mice were tested for preferences for other tastes including sweet and bitter. Both mice groups showed preference for sweet and avoidance for bitter solutions, indicating the taste function in general was intact and the observed difference caused by the CD36 knockout was specific to fat taste. Together, these results indicate that CD36 is essential for FFA detection and greatly impact intake. Lingual lipase is also very important for the fat taste sensation, because majority of the dietary fat is in the form of triglyceride (Kawai and Fushiki, 2003). Pharmacological inhibition of lingual lipase, an important enzyme responsible for converting triglyceride into fatty acids, leads to a

19 substantial decrease in preference for lipids (Kawai and Fushiki, 2003). Recent biochemical studies have shown that CD36 can cause intracellular Ca2+ increase by causing the release of 2+ Ca from the ER and this process is dependent on IP3 production and mediated by PLC (Abdoul-Azize et al., 2013; El-Yassimi et al., 2008; Gaillard et al., 2008). LCFA (long chain fatty acid) can cause the phosphorylation and activation of Src family kinases including Fyn and Yes in CD36-positive mouse TRCs (El-Yassimi et al., 2008). Src family kinases are known to regulate intracellular Ca2+ concentrations through modulation of the ER Ca2+ channel. Moreover, inhibition of Src family kinase activities with an inhibitor leads to abrogated intracellular Ca2+ accumulation in CD36-positive TRCs (El-Yassimi et al., 2008). TRPM5 channel may also play a role in fat taste sensing downstream of the Ca2+ release, similar to its role in sweet and salt signal transduction (Sclafani et al., 2007b). Knockout mice lacking TRPM5 show no preference for soybean oil emulsions initially, while after long acclimatization, knockout mice eventually developed preference for the fat containing solution (Sclafani et al., 2007b). Cellular depolarization induced by the presence of fatty acid can cause the release of neurotransmitters, 5- hydroxytryptamine (5-HT, serotonin) and noradrenalin (El-Yassimi et al., 2008). Both of these neurotransmitters are known to be released by Type III TRCs (presynaptic cells) in the taste bud; however, it is not clear whether inter-cellular coordination is necessary before afferent nerve becomes activated (Abdoul-Azize et al., 2013; Chaudhari and Roper, 2010). The fat taste signal is transmitted to the brain through gustatory nerves, including chorda tympani and glossopharyngeal, targeting the NST (Gaillard et al., 2008).

In addition to the well-studied CD36, a number of other potential receptors have been associated with fat taste detection. The delayed rectifying potassium channels (DRKs) allows the movement of K+ ions into the extracellular space, thus increasing the intracellular electronegativity and prevents depolarization (Mattes, 2009a). FFAs can block DRKs and can potentially cause depolarization directly or reduce the stimulus threshold for depolarization (Mattes, 2009a). As low as 10 µM of fatty acids is sufficient to block this channel and cause depolarization within milliseconds (Hirasawa et al., 2005). GPR40 and GPR120 are found primarily in Type I and Type II cells in the taste bud. Knockout mice lacking GPR40 and GPR120 show reduced preference and nerve responses to fatty acids (Cartoni et al., 2010).

2.6.2 Fat taste and food intake

Studies have shown that fat can promote overeating due to its energy density and the strong positive palatability of high-fat foods (Blundell and MacDiarmid, 1997). The average fat contribution to total energy consumption is 30-40% in developed countries and 15-30% of total energy consumption in developing countries (AICR, 2007). In order to prevent chronic diseases, the energy contribution from fat should be limited to 15-30% of total energy consumed (WHO, 2003). Fat intake varies greatly between individuals and even within the same individual on a day to day basis (Hunter et al., 1996; Willet, 1998). The drastic inter-individual variability in fat taste and preference has been proposed to be influenced by both genetics and acquired physiological conditions (Manabe et al., 2010). For instance, PROP tasters and non-tasters have been shown to have different sensitivity to dietary fats (Montmayeur and le Coutre, 2010; Tepper and Nurse, 1997). Individuals highly sensitive to PROP were able to distinguish between salad dressings containing 10% and 40% fat, while non-tasters were unable to detect the difference (Tepper and Nurse, 1997). Interestingly, the non-tasters were able to identify the high-fat dressing as more hedonically appealing, while the tasters were not (Tepper and Nurse, 1997). The underlying mechanisms for this hedonic perception difference is currently unclear; however, it could be due to other food attributes unrelated to fat taste (Montmayeur and le Coutre, 2010). Moreover, it was also demonstrated that tasters had higher numbers of fungiform papillae, which could explain the increased sensitivity to tastes in general (Montmayeur and le Coutre, 2010). In a separate study, an indirect association between the ability to perceive bitter taste of PROP and preference for high-fat foods was found in women and a direct association was found in men (Duffy and Bartoshuk, 2000). PROP tasters and non-tasters gave the same ranking of creaminess of dairy products, even though they used different set of criteria to reach the conclusion (Kirkmeyer and Tepper, 2003). The greater sensitivity for fat taste in PROP tasters has been proposed to be due to increased nerve innervation and tactile sensitivity rather than direct taste sensitivity (Kirkmeyer and Tepper, 2003).

Genetics may also contribute to the ingestion behavior and BMI of individuals in response to high fat consumption (Montmayeur and le Coutre, 2010). Even though obese individuals favor high-fat foods more than their lean counterparts, high-fat diets do not always lead to obesity (Blundell et al., 2005). Some individuals do remain lean even with high-fat intake (Blundell et al., 2005). Studies looking at the differences between high BMI individuals and

21 normal BMI (lean) individuals, who are both on high-fat diets, revealed that the high BMI group on average had stronger preferences high fat foods, and a weaker satiety response to fatty foods. Moreover, the high BMI group showed sustained preference for high-fat food during the post- ingestive period, suggesting the palatability could override the satiety signal from fat ingestion (Blundell et al., 2005; Erlanson-Albertsson, 2005). It has also been reported that lean individuals under a high-fat diet have elevated levels of diet-induced thermogenesis and leptin levels, suggesting that some individuals can sustain a high-fat diet without gaining weight due to their genotype (Blundell and Cooling, 1999).

Adding to the complexity of fat taste and intake behavior, FFAs, which can be detected by fat taste receptors such as CD36, generally taste repulsive for humans (Schiffman and Dackis, 1975). This is likely due to the presence of FFAs in decomposing foods and evolutionary pressure has drove humans to dislike the taste of FFAs in order to detect and avoid potentially toxic foods (Heilmann and Hummel, 2004; Montmayeur and le Coutre, 2010). For example, FFAs, such as docosahexaenoic acid and linoleic acid, contribute to the bitter and metal taste associated with taste deterioration in salmons due to long storage time (Refsgaard et al., 2000). The presence of fatty acid levels beyond 3.5-5.6 mg/g of fresh milk are avoided by participants (Nasser et al., 2001). Taste, orthonasal and retronasal olfaction tests of linoleic, oleic and stearic acids conducted on 22 healthy adult individuals showed that orthonasal detection threshold was significantly lower, consistent with the theory that evolution has pushed humans to detect FFAs, that are commonly found in spoiled foods, before the foods are ingested (Mattes, 2005). As a result, the hedonic value of fat taste has been proposed to be the result of the olfactory perception of fat-soluble flavor molecules (Montmayeur and le Coutre, 2010). Appetite and salivation can be elevated by these odors, which can also promote the release of gastric acid and insulin (Yeomans et al., 1997). Recent studies provide evidence that metabolic processes involved in digestion and energy storage may be influenced by oral exposure to fat, which could also affect long-term health status (Mattes, 2002; Mattes, 2005).

To date, few studies have investigated the effect of CD36 genotypes on human ingestive behaviors or fat taste. Keller et al found that, among African Americans, the rs1761667 SNP was associated with differences in the perceived creaminess of salad dressing and liking of certain types of fats (Keller et al., 2012). Pepino et al found that the rs1761667 SNP modified individuals ability to detect oleic acid and triolein where those homozygous for the G allele had

22 an 8 –fold lower detection threshold than those who were AA homozygous (Pepino et al., 2012). No effect of this genotype was seen on fat intake, though the study’s sample size (n=21) limited this analysis. A previous study from our lab investigated the effect of four SNPs (rs1984112 A>G, rs1761667 G>A, rs1527483 G>A, rs1049673 G>C) in the CD36 gene on fat intake (Toguri, 2008). Within Caucasians it was found that rs1984112 and rs1761667 both significantly affected the percent energy consumed from fat. Further differences in the amount of energy consumed from monounsaturated fatty acids (MUFAs) and PUFAs were observed. These differences in fat intake may be attributed to taste perception mediated by CD36.

2.6.3 Summary

Studies in humans and animals suggest that CD36 is a fatty acid receptor involved in fat taste perception. Further, it has recently been shown that variation in CD36 may result in differences in taste perception, preferences and intake. Work presented in this thesis aims to evaluate the effects of genetic variation within CD36 on fat intake and to see if variants that affect intake also are associated with fat taste.

2.7 Bitter taste

Bitter taste is thought to serve as a warning sign to protect animals from potentially toxic or poisonous compounds usually found in plants (Drewnowski, 2000; Tepper, 2008). This is supported by the fact that large numbers of naturally occurring noxious compounds taste bitter (Drewnowski and Gomez-Carneros, 2000). These potentially dangerous compounds, such as glycosides and alkaloids, are produced by plants for defense (Ames et al., 1990). Extremely bitter foods are generally considered unpalatable and are frequently avoided, even though some nutritious foods, such as grapefruit, can taste rather bitter due to the presence of phytonutrients (Drewnowski and Gomez-Carneros, 2000; Meyerhof, 2005; Tepper, 2008). Individuals with high phytonutrient diets have been found to have a lower risk of heart disease and cancer (Lee and Chan, 2011; Steinmetz and Potter, 1996). Individuals, who perceive bitter compounds more intensely, may be more likely to avoid the consumption of bitter foods, which may compromise their health. Bitter taste sensation is highly variable both within and between populations (Basson et al., 2005; Drewnowski et al., 1997; El-Sohemy et al., 2007). Genetic variations in T2R bitter receptor may contribute to the observed high variability in bitter taste (Bachmanov and Beauchamp, 2007; Kim et al., 2003). There are approximately 25 human T2R genes which

23 are highly polymorphic; however, it is currently unclear exactly how these genetic differences impact bitter taste sensation. Many of the T2R genes have not been well-studied, limiting our current understanding of their role in bitter taste sensation (Behrens and Meyerhof, 2009). Recent evidence suggests that variation in the TAS2R19 gene can modulate the liking of grapefruit juice (Hayes et al., 2011). Since naringin is the major bitter-tasting compound in grapefruit, it is possible that the perceived bitterness of naringin is modulated by variation in TAS2R19 gene, which can in turn affect grapefruit consumption.

2.7.1 Bitter taste receptors

The human bitter taste receptor TAS2R family consists of approximately 25 genes, which mediate bitter taste sensation (Chandrashekar et al., 2006). The TAS2R genes encode for GPCR proteins that have seven transmembrane domains (Bachmanov and Beauchamp, 2007). They are expressed in Type II TRCs in the taste bud (Chaudhari and Roper, 2010). T2R expressions have also been detected in non-gustatory cells of the digestive and respiratory tracts, suggesting that T2Rs may have functions beyond taste sensing (Finger et al., 2003; Wu et al., 2002). The possible gustatory involvement of all T2R genes were validated by reverse-transcriptase PCR (polymerase chain reaction) and in-situ hybridization in the circumvallate papillae (Behrens et al., 2007). The expression pattern of T2Rs in the TRCs appears to be heterogeneous, as studies have shown that different bitter-sensing TRCs express different subsets of T2Rs and no TRCs appear to express all T2Rs (Behrens et al., 2007; Roper, 2007). The number of TRCs expressing T2Rs and the expression level of different T2Rs in each cell differs based on the in-situ hybridization assay in use (Behrens et al., 2007). Moreover, a recent study using a heterologous expression system demonstrated that the compound specificity of the T2Rs can be variable (Meyerhof et al., 2010). Meyerhof et al expressed T2Rs in human embryonic kidney 293T cells stably expressing the chimeric G protein subunit Gα16gust44, the subunit necessary to induce intracellular Ca+ increase (Ueda et al., 2003). Following an incubation period of 24-26h for transfection, the increase in intracellular Ca2+ level is monitored using Ca imaging after treatment with various natural and synthetic bitter compounds, all of which have been previously shown to be bitter (Meyerhof et al., 2010). A number of broad spectrum T2Rs have been identified through this screen. TAS2R14 was activated by 33 (out of the 104 compounds assayed); TAS2R10 was activated by 32 compounds; and TAS2R46 was activated by 28 compounds (Meyerhof et al., 2010). Interestingly, TAS2R48 (renamed to TAS2R19) did not appear to respond to any of the

24 compounds tested including naringin; however, the authors did note that the expression level of this protein was marginal (Meyerhof et al., 2010).

Upon binding to a compatible compound, the TAS2R receptors are believed to activate α- gustducin and other G proteins, releasing their βγ subunits to cause PLC-mediated production of

IP3. The increased IP3 level leads to the opening of IP3R3 ion channels on the ER, releasing Ca2+ into the cytosol of TRCs (Behrens and Meyerhof, 2009; Chaudhari and Roper, 2010; Roper, 2007). The increased intracellular Ca2+ leads to the activation of TRMP5 channel, which helps generate a cellular depolarization and release of neurotransmitter ATP into the extracellular space (Chaudhari and Roper, 2010; Roper, 2007). This causes a neuronal excitation signal to be sent to the gustatory cortex of the brain through afferent facial (VII), glossopharyngeal (IX) and vagus (X) cranial nerves (Bachmanov and Beauchamp, 2007).

2.7.2 Bitter taste and food intake

Bitter taste perception may have impact on health and nutrition; genetic differences among individuals may account for differences in health outcomes related to nutrient intake (Reed et al., 2006). Bitter compounds elicit a natural rejection response and this is critical for avoiding potentially toxic compounds in foods, such as alkaloids, decomposing fat and hydrolyzed proteins (Drewnowski and Gomez-Carneros, 2000). Even though many bitter-tasting foods should be avoided, some bitter foods contain compounds that can promote health, including phenols (in tea, citrus fruits and wine), organosulfur compounds (cruciferous vegetables such as broccoli) and phytonutrients (in fruits and vegetables, such as grapefruit). It has been proposed that it is challenging to increase fruit and vegetable intake partly because of the aversive bitter taste of some of these healthy foods. Many vegetables are often disliked by children likely due to their bitter taste (Drewnowski, 2000; Goldstein et al., 2005; Lamb and Ling, 1946; Tepper, 1998).

Differences in bitter taste sensation may modulate the risk associated with a number of diseases diseases, including obesity, cancer and heart disease; possibly through modification of dietary habits (Basson et al., 2005; Goldstein et al., 2005). Naturally occurring genetic variability in TAS2R38 gene is responsible for a well-documented individual difference in the sensitivity to phenylthiocarbamide, PTC and its related compound PROP. Population studies revealed that there is a wide distribution of taste threshold for PTC and PROP and there is a dip in the middle

25 of the distribution, dividing people into tasters and non-tasters (Guo and Reed, 2001). Supertasters of PROP also exist in the natural population and they perceive PROP as more intensely bitter than normal tasters (Tepper, 2008). Female supertasters have been shown to have lower acceptance ratings to naringin solutions produced from grapefruit peel and lower preferences for grapefruit juice (Drewnowski et al., 1997). Study done on preschoolers using a 50:50 grapefruit-orange juice blend found that taster children had lower preference than the non- taster children (Tepper, 2008).

A recent study by Hayes et al has identified TAS2R19 gene as a potential modifier of grapefruit juice liking (Hayes et al., 2011). The rs10772420 A>G coding SNP in the TAS2R19 gene results in a Cys299Arg substitution. Individuals homozygous for the Cys299 allele perceived grapefruit juice as twice as bitter as Arg299 homozygous individuals or heterozygous individuals. Moreover, the Cys299 homozygous and heterozygous individuals had lower preferences for grapefruit juice than the less sensitive Arg299 individuals (Hayes et al., 2011). Interestingly, TAS2R19 is still an orphan receptor and the possibility exists that naringin is one of its ligands, given its association with grapefruit preference.

2.7.3 Summary

Bitter taste is mediated by the TAS2R family of receptors. The compound naringin is the primary bitter tastant in grapefruit and variability in individual’s sensitivity to it may modulate intake. Recent work has shown that variation in the TAS2R19 gene is associated with differences in grapefruit preference. The aim of the work presented in this thesis is to understand how variation in the TAS2R19 gene modifies sensitivity to naringin and if changes in sensitivity impact preference for and intake of grapefruit.

2.8 Food intake assessment

In order to measure food consumption over a specific period of time, different assessment methods have been developed. Food choice is a complex behaviour affected many factors, including physiological, environmental, sociocultural and economic factors (Raine, 2005). Tools used to monitor or measure dietary intake include diet histories, diet records, 24-hour recalls and food frequency questionnaires (FFQs). Diet histories do not have a well-defined template; as a result, the information collected initially is not well-organized. However, information about

26 individuals’ past dietary habits can be derived (Thompson and Byers, 1994). Diet records require trained individuals to record all food and beverages as they are consumed over a period of time. The record time can range over a series of 3 to 5 days, which should include at least one weekend day (Willet, 1998). This method can allow the documentation of detailed information such as eating time, food and beverage size, portions and brand name (Thompson and Byers, 1994). A 24-hour recall is conducted by a trained interviewer who asks participants to systematically recall all food and beverages consumed with in a 24 hour time period (Willet, 1998). This method can also collect information regarding preparation method, brand names and portion sizes. It is possible to use the 24-hour recall method to estimate usual consumption over a particular study period; however, multiple recalls must be conducted in order to get an accurate measurement of individuals’ intake due to the fact that daily diet can be highly variable (Willet, 1998). FFQs allow for the documentation of eating habits by recording the frequency of food and beverage consumption from a pre-determined list over a particular period of time, which typically ranges from 1 month to a year (Thompson and Byers, 1994; Willet, 1998). The drawback for FFQ is that it does not collect as much detailed information on food and beverages consumed, such as portion sizes and preparation methods. It has also been criticized for its dependence on individuals’ memory (Subar et al., 2001). Nevertheless, FFQ can be extremely useful because it can be self-administered and allows quick processing, which can substantially reduce both time and cost (Thompson and Byers, 1994). Burden felt by study participants has been found to be lower for FFQs than multi-day dietary records and 24-hour dietary recalls (Thompson and Byers, 1994).

A number of FFQ formats exist and three of the most popular formats are: the Block FFQ, the Willet FFQ and Dietary History Questionnaire (DHQ) developed by the National Cancer Institute (Subar et al., 2001). Subar et al compared the 3 dietary FFQ formats and provided some great insights into the proper use and interpretation of these questionnaires (Subar et al., 2001). Based on the study, the 3 tools created similar results; therefore, using any one of the 3 questionnaire formats should generate similar intake values. Moreover, in the same study, results obtained using the 3 FFQ formats were compared with 4 separate 24-hour recalls conducted approximately 3 months apart. The correlation of energy adjusted fat intake between results obtained through FFQs and 24-hour recalls ranges from 0.60 to 0.67, indicating a strong correlation between the different methods (Subar et al., 2001). On the other hand, however, due to the fact that they are similar types of tools, intrinsic errors associated with this type of

27 measurement may limit their effectiveness (Willett, 2001). Indeed, similar attenuation efficiency suggest that the maximum effectiveness for FFQs to measure nutrient intake may be limited (Willett, 2001). FFQs validation are commonly achieved by comparing FFQ results to results obtained from multiple 24-hour recalls or multi-day food records (Thompson and Byers, 1994). For instance, energy-adjusted fat intake measured using FFQs has a correlation value of approximately 0.6 with 24-hour recalls and multi-day food records (Munger et al., 1992; Subar et al., 2001; Willett, 2001; Willett et al., 1985). Among women, energy adjusted fat intake measured by FFQs had a correlation value of 0.62 with results from 5 separate 24-hour recalls (Munger et al., 1992).

Even though FFQ results correlate well with results from 24-hour recall and food records, they may have some common areas that are prone to errors. For instance, all of these methods will likely have reporting errors. As a result, more objective biological markers have also been used to validate FFQ and other self-reporting results. This type of correlation study is particularly important because the source of error will not originate from dietary reporting itself. Fatty acids collected from subcutaneous adipose tissues have been used as a proxy for certain dietary fatty acid intakes (Arab, 2003; Bingham, 2002). The percentage of energy ingested in the form of PUFA, determined through Willett FFQ, had a significant correlation value of 0.3 with total n-3 and n-6 fatty acids found in the adipose tissue samples of African-American males (Holmes et al., 2007). In addition, percent energy values from PUFA, measured through Willett FFQ, also correlated with PUFA markers found in adipose tissues in a Boston population (Hunter et al., 1992). Together, these results suggest that FFQ measurements can accurately reflect true intake results. In addition to fat, meta-analysis correlation studies have shown strong correlation between FFQ results and results from multi-day food records and multiple 24-hour recalls, for protein (correlation coefficient: 0.56), carbohydrate (correlation coefficient: 0.58), alcohol (correlation coefficient: 0.80), calcium (correlation coefficient: 0.63), Vitamin C (correlation coefficient: 0.64) and dietary fibers (correlation coefficient: 0.53) (Molag et al., 2007). FFQs have also been used in studies measuring consumption of bitter taste food and results from FFQ also had strong correlation with 24-hour recall results (Duffy et al., 2010). Given its comparable accuracy with other methods, substantially reduced time and cost, FFQ is particularly well-suited for assessing habitual dietary intake in large epidemiological studies.

2.9 Food preference assessment

Food preference assessment aims to measure the liking or disliking of a particular food (Drewnowski, 1997a). Even though taste perception plays a large role in food preferences, it is not always the only determining factor for preference (Drewnowski, 1997a). Additional food attributes can also influence preference, including texture and appearance. Moreover, individuals’ age, gender, BMI and health consciousness are also may be associated with preferences. Two commonly used methods can be employed to conduct food preference assessments: food preferences can be measured directly using taste tests of real foods of interest under controlled laboratory conditions or it can be measured using a questionnaire (Drewnowski, 1997a). Food preference questionnaires ask participants to rate the liking and/or disliking levels of foods based on a hedonic scale. The criticisms of the food preference questionnaire are based on the fact that it relies on the participants’ memory of past experiences and may also reflect test subjects’ opinions of the food in addition to the actual taste of the food (Frank and van der Klaauw, 1994). However, the food preference questionnaire method is particularly well suited for large epidemiological studies, because it is relatively easy to use, has low time burden and cost requirement and can measure a large number of foods simultaneously.

The most commonly used scale in food preference measurement is the 9-point hedonic scale (Hein et al., 2008). This hedonic scale ranges from 1, being extremely dislike, to 9, being extremely like, with a the neutral point of 5, being neither like or dislike. This scale is criticized for the fact that it cannot exhibit ratio properties, may reduce of variability in responses due to ceiling effects and the interpretation of the descriptors may be different between participants (Hein et al., 2008). However, results obtained using this 9-point scale have been shown to be reliable, contain high degree of variability and have the sensitivity to discriminate preferences between samples (Lawless and Malone, 1986). The 9-point scale can sometimes even exceed the performance of other food preference scales (Lawless and Malone, 1986). Similar preference results were obtained from both the 9-point scale and the labeled affective magnitude scale, a category scale with possible ratio properties and extreme anchors to avoid ceiling effects (Hein et al., 2008).

Using this 9-point scale, a food preference checklist (FPC) was developed in 1999 by Drewnowski and Hann (Drewnowski and Hann, 1999). The list consisted of 171 foods

29 representing all major food groups such as: grains, vegetables, fruits, fats, sugars, desserts and dairy and beverages. The foods included were selected from commonly used FFQs at the time as well as from a more comprehensive checklist from the US Army (Block et al., 1986; Meiselman et al., 1974; Willett et al., 1985). Foods on this checklist were intentionally indicated as specifically as possible and generic terms such as “other fruit” were excluded. Participants could then indicate their preferences using the 9-point scale with additional options for “never tried” and “would not try” (Frank and van der Klaauw, 1994; Meiselman et al., 1974). In the original study, results obtained from FPC strongly correlated with the consumption data derived using the Block FFQ (Drewnowski and Hann, 1999). Since its initial development, this FPC has become a commonly used tool (Drewnowski et al., 2007; Drewnowski et al., 2000; Kaminski et al., 2000).

2.10 Taste function assessment

The ability to detect taste compounds at low concentrations and sensitivity to high concentrations contribute to an individual’s food preferences, which in turn could influence consumption (Bartoshuk et al., 2004; Glanville and Kaplan, 1965; Kaminski et al., 2000). Sensory testing can be used to obtain measures of an individual’s ability to taste a given tastant. Taste thresholds and suprathreshold taste sensitivity are distinct measures of taste function that have been proposed as optimal techniques to utilize when examining the genetic basis of taste (Galindo-Cuspinera et al., 2009).

There are two main kinds of threshold measurements, detection and recognition thresholds. Detection thresholds measurements allow for the assessment of lowest concentration at which detection of a given tastant is possible. Recognition thresholds allow for the assessment of the lowest concentration at which an individual can characterize a given tastant. A recent study showed that one method of detection threshold measurement, the staircase model, is superior to tested recognition threshold models in assessing the impact of genetic variation on taste(Galindo-Cuspinera et al., 2009).

There are numerous methods used to assess suprathreshold sensitivity (STS). Suprathreshold measurements examine an individual’s perceived taste intensity at either a number of concentration points or at a single concentration. Tastants are often presented in solutions or are infused on filter papers that are applied to the tongue. Subjects can assess the intensity of sampled compounds on a variety of scales, for example 1 to 9 hedonic scales or

30 general labeled magnitude scales (gLMS), or can rank solutions against a standard suprathreshold solution to determine the point at which they can no longer differentiate between the two, the Just Noticeable Difference (Galindo-Cuspinera et al., 2009). In a recent trial examining different methods of suprathreshold taste measurement, intensity ratings measured with gLMS were found to be the most reliable way to identify phenotypes that corresponded with expected genotypes(Galindo-Cuspinera et al., 2009).

The importance of suprathreshold measurement is illustrated by the discovery of supertasters of PROP. Individuals can be categorized based on their threshold detection levels of PROP into tasters or non-taster groups (Bartoshuk, 2000). However, among the tasters there were large variations in bitter taste perception (Bartoshuk, 2000). Tasters, who had similar detection thresholds, sometimes could experience drastically different bitter taste intensity in response to suprathreshold solutions (Bartoshuk, 2000). Supra-threshold testing allows for the sub categorization of these individuals and led to the discovery of super-tasters, who perceived PROP as much more bitter than normal tasters (Bartoshuk, 2000).

2.11 Genotyping techniques

SNP genotyping can be achieved through a number of techniques based on vastly different principles with varying degrees of high-throughput capabilities and cost (Edenberg and Liu, 2009). Currently, some of the most efficient methods include: single-base extension method with readouts of mass differences, which can be detected by Sequenom MassARRAY; hybridization method with readouts that allow detection of oligonucleotide binding on microarrays; and cleavage of hybridized oligonucleotides using Taq-Man.

The Sequenom MassARRAY technology allows the simultaneous measurement of up to 40 markers per reaction on genomic DNA of individuals, which makes it ideal for projects where a large number of individuals need to be genotyped with medium-sized SNP (Bradić et al., 2011). This method can distinguish different alleles based on the different masses of primer extension products (Bradić et al., 2011; Tang et al., 1999). Regions of interest are first amplified using PCR to increase the quantity of target region DNA. Unincorporated dNTPs are inactivated with shrimp alkaline phosphatase (PCR cleanup) and primer extension is subsequently carried out with ddNTPs (dideoxy nucleoside triphosphates) with modified masses. Different ddNTPs are incorporated into the extension product depending on the allele that is present immediately

31 downstream of the 3’ end of the primer. The reaction is cleaned using resin, which facilitates optimal detection, and the product is spotted into a chip. The reaction products from various regions of interest on the chip are analyzed by matrix-assisted laser desorption/ionization time- of-flight mass spectrometry (MALDI-TOF MS). The specific ddNTP incorporated can be identified based on their unique masses. The current MALDI-TOF MS mass detection range and precision limit the number of assays per reaction to 40. Peaks from the MALDI-TOF analysis can be analyzed with software that can match the primer masses with the alleles.

Microarrays, such as Affymetrix 6.0 SNP and Copy Number Variation chip, are also a powerful tool in SNP genotyping (Ding and Jin, 2009; Edenberg and Liu, 2009; Trachtenberg et al., 2012). The Affemetrix 6.0 microarray chip is based on the principle of differential hybridization based on matched or mis-matched probe binding to target DNA sequences (Ding and Jin, 2009; LaFramboise, 2009). For each SNP, 6-8 probes (25 base pair oligomers each) are designed and immobilized on the microarray chip. For the Affymetrix 6.0 microarray, every probe is designed to be perfectly complementary to a portion of the DNA sequence containing the SNP site (LaFramboise, 2009). Due to its high probe density, Affymetrix 6.0 microarrays can hold millions of different probes on the same chip and allow the screening of close to 1 million SNPs simultaneously. Fluorescently-labeled PCR-amplified DNA from test subjects are then added to the chip and allowed to hybridize to the probes under the same temperature, time and buffer conditions. The strength of DNA binding, which is determined by the degree of complementary matching between the probes and subject DNA, is measured through fluorescence intensity readout. Probes that are perfectly complementary to the subject DNA sequence generate a brighter signal than mis-matched probes. The fluorescence intensity map on the chip can subsequently be interpreted and translated into a SNP profile by the microarray software (Ding and Jin, 2009; LaFramboise, 2009).

2.12 Rationale, Hypothesis and Objectives

Rationale: Few studies have examined the effect of genetic variation on taste function and food preferences or intake concurrently. To date no studies have directly examined the role of genetic variation in salt taste perception in humans outside of those presented in this thesis. Only two studies have looked at the effect of variation in the TAS1R2 sweet taste receptor on sweet and sugar intake. No SNPs have been identified that are associated with taste but this may be because

32 studies completed thus far have focused only on one aspect of taste (recognition threshold) and did not adjust for the potential effects of BMI (Eny et al., 2010; Fushan et al., 2009). Recently Pepino et al examined the effect of variation in the CD36 gene on fat taste but also only focused on threshold taste and were not able to show any effect of identified taste modifiers on intake (Pepino et al., 2012). The role of genetic variation in bitter taste perception has received considerable attention but most has focused on the relationship between the TAS2R38 gene and its agonists PTC and PROP. Many other bitter taste receptor genes, like TAS2R19, are relatively unstudied and little is known about their agonists, what affect variation within them has on taste function and whether they are important in determining consumption behaviors. Given recent findings that show variation in TAS2R19 is associated with grapefruit preference, it represents an interesting target for further study (Hayes et al., 2011). With increasing prevalence of diseases related to over nutrition there is considerable interest in identifying genes that predispose individuals to such disease by influencing dietary decisions. It is important that we understand how gene variation affects taste function and if these changes in taste perception translate to differences in consumption.

Hypothesis: Variation in genes associated with salt, sweet and fat and bitter taste will affect measures of taste perception, food preference as well as intake of sodium, sugar, fat and foods containing bitter compounds, respectively.

Objectives: 1. Examine how variation in putative salt taste receptors ENaC (SCNN1A, SCNN1B, SCNN1D, SCNN1G) and TRPV1 (TRPV1) affects salt taste thresholds, supra-threshold taste sensitivity and sodium intake. 2. Examine how genetic variation in sweet taste receptor TAS1R2 (TAS1R2) affects sucrose taste thresholds, supra-threshold taste sensitivity and intake of sugars. 3. Examine how variation in bitter taste receptor TAS2R19 (TAS2R19) affects naringin sensitivity and both grapefruit/grapefruit juice preference and intake. 4. Examine how genetic variation in putative fat taste receptor CD36 (CD36) affects habitual fat intake and taste perception.

Chapter 3 Genetic variation in putative salt taste receptors and salt taste perception in humans

Adapted from: Dias A.G. et al. Chemical Senses. 2013 Feb; 38 (2):137-45.

3.1 Abstract

The objective of this study was to determine whether SNPs in the SCNN1A (3), SCNN1B (12), SCNN1G (6) and TRPV1 (10) genes affect salt taste perception and sodium intake. Participants were men (n=28) and women (n=67) from the Toronto Nutrigenomics and Health study aged 21-31 years. Taste thresholds (TT) were determined using a 3 alternative forced choice staircase model with solutions ranging from 9 x10-6 to 0.5 mol/L. Suprathreshold taste sensitivity (STT) to 0.01-1.0 mol/L salt solutions was assessed using general Labeled Magnitude Scales. None of the SNPs in the SCNN1A and SCNN1G genes were significantly associated with either outcome. In the SCNN1B gene 2 SNPs in intronic regions of the gene modified STT (mean iAUC ± SE). Those homozygous for the A allele of the rs239345 (A>T) polymorphism and the T allele of the rs3785368 (C>T) polymorphism perceived salt solutions less intensely than carriers of the T or C alleles, respectively (rs239345 70.82 ± 12.16 vs. 96.95 ± 3.75, p=0.02; rs3785368 57.43 ± 19.85 vs. 95.57 ± 3.66, p=0.03) In the TRPV1 gene, the rs8065080 (C>T, Val585Ile) polymorphism modified STT where carriers of the T allele were significantly more sensitive to salt solutions than the CC genotype (98.3 ± 3.8 vs. 74.1 ± 8.3, p = 0.008). Our findings show that variation in the TRPV1 and the SCNN1B genes may modify salt taste perception in humans.

3.2 Introduction

Sodium intake varies widely both between and within populations (Dahl, 2005). Taste is one of the primary determinants of food intake and variation in an individual’s ability to taste salt might partially explain the variation observed in sodium (El-Sohemy et al., 2007; Garcia-Bailo et al., 2009b). Previous studies with bitter and sweet taste have shown that genetic variation can alter an individual’s taste function (Eny et al., 2010; Fushan et al., 2010; Kim et al., 2003), but no studies have yet identified any genetic determinants of salt taste in humans.

Two lingually-expressed cation channels have been identified as potential salt taste receptors; the sodium-specific and amiloride-sensitive epithelial sodium channel (ENaC) and the TRPV1 channel (DeSimone and Lyall, 2006). ENaC is the primary mediator of salt taste in a number of herbivores and is thought to be responsible for the appetitive behavioral responses elicited by salt taste (DeSimone & Lyal, 2006; Chandrashaker et al, 2010). Mice genetically

35 engineered to lack expression of functional ENaC in taste cells have a complete loss of both sodium attraction and salt taste responses, providing evidence for the role of ENaC as the primary determinant of salt taste in mice (Chandrashaker et al, 2010). However, the role of ENaC in humans and other mammals is less clear. In rats, amiloride sensitive mechanisms have been estimated to be responsible for only 70% of salt taste perception based on chorda tympani nerve responses (DeSimone and Lyall, 2006). In humans, lingual surface potential in response to salt stimuli is reduced by only 5-40% in the presence of amiloride (Feldman et al, 2003). This suggests that other cation channels may be involved in salt taste in humans.

The TRPV1 channel responds to a variety of cations and is amiloride-insensitive (Stewart et al., 1997). This receptor responds to vanilloids such as capsaicin and resiniferatoxin, which are both known to modify amiloride-insensitive nerve activation by salt stimuli (Lyall et al., 2004). The strongest evidence that TRPV1 is involved in salt taste was observed when TRPV1 knockout mice were shown to lack NaCl chorda tympani nerve responses in the presence of amiloride, whereas control mice displayed normal levels of nerve innervations (Lyall et al., 2004). This indicates that by disrupting the function of ENaC and TRPV1 channels concurrently, one may eliminate chorda tympani mediated salt taste.

Studies involving mice have shown that heritability plays a role in the inter-strain differences in sodium intake and preference (Bachmanov et al., 2002; Tordoff et al., 2007). Using a candidate gene approach, Shigemura et al. demonstrated that variation in the gene for the alpha subunit of ENaC may explain the inter-strain variation seen in salt taste perception within mice (Shigemura et al., 2008). This work indicated that inter-individual differences in salt taste acuity may be mediated by genetic variation in putative salt taste receptors. The potential effect of variation in the TRPV1 gene in mice has not been explored.

In contrast to evidence from mouse models, no studies have directly examined the role of genetic variation in salt taste perception in humans. Twin studies examining salt preference and salt taste thresholds did not find a heritable component for either outcome (Greene et al., 1975; Wise et al., 2007). However, it has recently been suggested that the method used to determine taste thresholds in such studies might not have been ideal for taste phenotyping (Galindo- Cuspinera et al., 2009). Moreover, previous studies relied only on heritability estimates to

36 determine the effect of genetic variation on salt taste and preference, which does not provide information on any specific genes.

Given the current dearth of knowledge on the role of genetic variation on human salt taste perception and sodium intake, the objective of this study was to determine whether genetic variation in ENaC (SCNN1A, SCNN1B, SCNN1G, SCNN1D) and TRPV1 (TRPV1) is associated with salt taste thresholds, suprathreshold taste sensitivity and sodium intake in humans.

3.3 Methods

3.3.1 Subjects

Subjects were Caucasian women (n= 377) and men (n= 165), ranging in age from 20-29 years, from the Toronto Nutrigenomics and Health Study, a cross-sectional study aimed at investigating the effects of gene-diet interactions on biomarkers of chronic disease as well as identifying genetic determinants of food preference and intake in a population of young adults. Subjects were excluded from the study if they were pregnant or breastfeeding. Further, subjects were removed from the dietary analysis if their reported caloric intake was below 800 kcal/day or exceeded 3500 kcal/day for females and 4000 kcal/day for males. Informed consent was obtained from all participants and the study was approved by the University of Toronto Research Ethics Board.

Of these individuals 28 men and 67 women between the ages of 21-31 years were enrolled in the present study. Individuals who were smokers, pregnant or breast feeding, experienced marked weight changes in the last year (greater than 15 pounds), were diagnosed with chronic sinusitis or chronic obstructive bowel disease, had lost their sense of smell, often experienced severe dry mouth, were diagnosed with diabetes or any other chronic disease or were diagnosed with a psychological disorder were excluded from the study. This study was approved by the University of Toronto and George Brown College ethics boards. Written informed consent was received from all participants.

3.3.2 General Protocol

3.3.2.1 Experimental Condition

Subjects were asked to attend two separate visits at the study center at 10:00 or 11:30 AM following at least a two hour fast. Subjects were also asked to refrain from drinking coffee or eating strong tasting foods the morning of the test and were requested to repeat the same breakfast before each of the two visits. Each participant visit occurred at the same time of day and on the same weekday, one week apart. Taste thresholds were assessed on visit one, while suprathreshold taste sensitivity was examined on visit two.

3.3.2.2 Threshold Testing

Detection thresholds for NaCl were assessed using a 3 Alternative Forced Choice up down method. Participants were asked to rinse their mouths with distilled water prior to the start of the test and between trials. A 10 mL taste solution or distilled water was dispensed into covered plastic cups labeled with 3-digit codes for blinding. During each trial the subject was presented with one cup containing a taste solution and 2 containing distilled water only. The order of the presentation of cups was randomized. Individuals were instructed to sample all three solutions and identify the cup that was different from the other two. The participants were required to always make a selection and were informed that guessing was acceptable if the solutions seemed indistinguishable.

The concentration of the subject’s first taste solution exposure was 0.028 mol/L NaCl during the initial week of testing and for all subsequent weeks the initial solution concentration was selected based on the previous week’s average threshold for a given gender. Subsequent increases or decreases in taste solution concentration were determined by the participant’s choices. If an individual provided a correct response, the stimulus concentration was lowered; an incorrect response resulted in an increase in concentration. The taste solution concentration at which the concentration sequence changed from decreasing to increasing or increasing to decreasing was considered to be a reversal. Five reversals were allowed to occur and an individual’s threshold was calculated as the geometric mean of the final 4. Test solutions were 0.25 log cycles apart in concentration, ranging from 9 x 10-6 to 0.5 mol/L of NaCl dissolved in distilled water. Control solutions contained only distilled water. All solutions were prepared within 24 hours of testing.

3.3.2.3 Suprathreshold Testing

Suprathreshold taste sensitivity to NaCl was assessed using gLMS (Bartoshuk et al., 2004). Five solutions 0.5 log cycles apart in concentration, ranging from 0.01 to 1.0 mol/L NaCl, were presented to subjects in a randomized order using block randomization. Subjects were given a glass of distilled water and were asked to rinse between solutions. All solutions were prepared within 24 hours of testing. Prior to beginning suprathreshold testing, subjects were asked to complete a scaling exercise to confirm their ability to rank stimuli of different intensities. Subjects were asked to order the following sensations in order of decreasing intensity: 1) conversation, whisper, loudest sound experience 2) well-lit room, a dimly lit room, staring at the sun. Subjects who could not complete these tasks (n=0) correctly were excluded from the test.

3.3.3 Dietary Assessment

All subjects completed a 1-month, 196-item semi-quantitative FFQ, which has previously been described (Eny et al., 2008). In the FFQ, subjects are offered nine possible responses to indicate how many times in the past month they consumed a specified portion of each food or beverage: never, less than once per month, 1-3 times per month, once per week, 2-4 times per week, 5-6 times per week, once per day, 2-3 times per day and 4 or more times per day. Subject responses to the individual foods are converted to average daily intake for each item. The average daily intakes of all items are combined to compute a total daily intake of all major macro- and micro- nutrients for each subject using content values from the Harvard FFQ Database and the United States Department of Agriculture National Nutrient Database for Standard Reference.

3.3.4 Genotyping

Each participant had venous blood drawn after a 12-h overnight fast, and DNA was isolated from whole blood using the GenomicPrep Blood DNA Isolation kit (Amersham Pharmacia Biotech Inc, Piscataway, NJ). Genotyping was completed for each subject using an Affymetrix 6.0 chip. Genotypes for SNPs in the SCNN1A, SCNN1B, SCNN1G, SCNN1D and TRPV1 genes were extracted from the chip (Table 3.1). SNPs with a minor allele frequency (MAF) of less than 15% were excluded from the analysis considering the sample size of the current study population (Table 3.1). Linkage between the remaining SNPs in each gene was

39 assessed (Figure 3.1). LD between SNPs was calculated from a group of 541 (165M/376F) genotyped individuals from which participants of this study were drawn. SNPs deleted due to linkage disequilibrium (LD) are identified in Table 3.1.

Table 3.1: Allelic distribution of single nucleotide polymorphisms (SNPs) extracted from the Affymetrix 6.0 SNP chip within the study population.

SNP (rs Gene number) XX (n) XY (n) YY (n) MAF (%) SCNN1A rs4149570 23 38 33 60 rs11064145 11 50 33 47 rs7956915 34 38 22 44 rs3782724* 81 11 2 8 rs4149621* 2 2 90 3 SCNN1B rs6497659 22 51 19 48 rs7195627 25 50 17 46 rs4967999 35 40 17 40 rs239345 8 32 52 26 rs4511539 52 35 5 24 rs11074555 58 25 9 23 rs889299 6 26 59 21 rs8044970 4 29 59 20 rs8044984 a 59 29 4 20 rs63982 a 1 33 58 19 rs3785368 3 27 62 18 rs152733 60 31 1 18 rs17256727 a 1 28 63 16 rs7205273 1 26 65 15 rs9931113 67 19 4 15 rs16939978* 68 21 2 14 rs12447134* 71 18 3 13 rs3785361* 3 17 71 13 rs9939129* 3 12 77 10 rs17199599* 1 9 82 6 rs239350* 1 9 82 6 rs8055868* 0 9 83 5 rs250563* 1 7 84 5 rs7190829* 85 6 1 4 rs152734* 87 4 1 3 rs12596831* 88 4 0 2 rs4967994* 0 3 89 2 rs4967951* 89 3 0 2 rs41362849* 89 3 0 2 rs9930640* 0 3 89 2 rs16940013* 0 2 90 1 rs10492792* 0 1 91 1 rs11865186* 0 1 91 1 rs4421986 40 43 9 33 SCNN1G rs12923648 a 6 44 42 30 rs4480807 6 44 42 30 rs7404739 a 43 43 6 30 rs13331086 51 34 7 26 rs4421986 52 32 7 25

rs4297681 a 5 35 52 24 rs4073930 a 6 33 53 24 rs4073291 6 33 53 24 rs4499238 2 32 58 20 rs4297681a 51 34 7 27 rs12447053* 74 17 1 10 rs4967948* 77 15 0 8 rs17740896* 0 14 78 8 rs161386 31 47 16 42 SCNN1D rs6303788* 83 9 2 6.9 TRPV1 rs150908 12 53 28 41 rs879207 17 39 38 39 rs4790522 13 46 35 38 rs9902581 35 46 13 38 rs8065080 16 38 39 38 rs224546 39 40 15 37 rs7217945 13 42 39 36 rs1018187 a 38 43 12 36 rs150846 13 39 42 35 rs224536 7 44 33 35 rs161393 a 42 45 7 31 rs161364 5 35 52 24 rs222745* 75 18 1 11 rs222748* 75 18 1 11 rs17633288* 92 2 0 1 rs16953199* 0 2 92 1 rs224551* 94 0 0 0

MAF denotes minor allele frequency. * denotes that this SNP was not analyzed further due to low (<15%) MAF. a denotes that this SNP was not analyzed further as it was closely linked with another analyzed SNP.

Figure 3.1: Linkage disequilibrium (LD) in 541 (165M/376F) individuals for SNPs with a minor allele frequency > 15% in a) SCNN1A, b) SCNN1B, c) SCNN1G, and d) TRPV1. A representation of the gene (not to scale) along with the relative position of SNPs to one another is shown above each LD plot. The SNPs are listed from left to right at the top of the figure according to their position in the gene. Each square shows the pairwise comparison between the two SNPs on either side of the square on the diagonal. LD is assessed in this figure by r2. For SNP pairs with an r2 > 0.80 only a single SNP was analyzed. Eliminated SNPs are identified in Table 3.1 with *.

3.3.5 Statistical Analysis

Block randomization was completed using Random Allocation Software. LD plots were created using Golden Helix SNP Variation Suite (Version 7.4; Golden Helix Inc, Bozeman, Montana).

Taste thresholds were calculated by computing the geometric mean of the concentrations at which the final four reversal points in the staircase procedure occurred. Individual ratings of the intensity of suprathreshold solutions were plotted and the iAUC for each taste sensitivity (mean ± SE) was computed using GraphPad Prism (Version 5; GraphPad Software Inc, La Jolla, CA).

The effect of genotype on both threshold and suprathreshold taste was assessed using Statistical Analysis Software (version 9.2; SAS Institute Inc, Cary, NC). The general linear model (GLM) procedure in SAS was used to perform a one-way analysis of variance with appropriate adjustments, to test for differences in salt taste thresholds, iAUC and sodium intake across genotypes. Non-normally distributed variables were log-transformed for analysis, and their anti-logs are reported. Genotypes that were significantly associated with iAUC were assessed for mode of inheritance and, using a GLM adjusted for age and sex, the effect of genotype was assessed on individual intensity ratings at each concentration point in the taste sensitivity curve. SNP(s) that were found to be significantly associated with either taste outcome were further examined to see if they modified sodium intake. Adjustments for age, sex, BMI, caloric intake and physical activity were made to the dietary analyses.

3.4 Results

Table 3.2 shows the association between each SNP, salt taste thresholds and iAUC. None of the SNPs in the SCNN1A and SCNN1G genes were associated with either outcome. Two SNPs in the SCNN1B gene and one in the TRPV1 gene were found to be associated with suprathreshold taste.

Table 3.2: The association between candidate gene polymorphisms and both salt taste thresholds and suprathreshold taste sensitivity.

Salt Taste Suprathreshold taste SNP Threshold sensitivity Gene (rs number) Genotype n (mMol/L ± SE) (iAUC ± SE) SCNN1A rs4149570 XX 23 4.16 ± 0.69 89.17 ± 7.41 XY 38 3.23 ± 0.54 101.99 ± 5.94 YY 33 3.59 ± 0.58 88.92 ± 6.22 rs11064145 XX 11 3.55 ± 1.00 100.42 ± 10.66 XY 50 3.49 ± 0.47 94.88 ± 5.29 YY 33 3.74 ± 0.58 90.43 ± 6.41 rs7956915 XX 34 3.75 ± 0.57 95.17 ± 6.19 XY 38 3.81 ± 0.54 91.39 ± 6.1 YY 22 2.95 ± 0.7 96.69 ± 7.74 SCNN1B rs4511539 XX 52 3.42 ± 0.52 89.57 ± 4.79 XY 35 4.32 ± 0.62 97.65 ± 5.84 YY 5 2.79 ± 1.68 120.43 ± 15.46 rs11074555 XX 58 3.87 ± 0.44 97.73 ± 4.61 XY 25 3.32 ± 0.66 88.92 ± 7.01 YY 9 2.54 ± 1.11 87.33 ± 11.69 rs7195627 XX 25 3.41 ± 0.76 97.77 ± 6.94 XY 50 3.98 ± 0.54 97.35 ± 4.91 YY 17 3.56 ± 0.87 80.35 ± 8.42 rs3785368 XX 3 4.19 ± 2.17 57.43 ± 19.85 a XY 27 2.98 ± 0.71 88.60 ± 6.62 b YY 62 4.06 ± 0.47 98.6 ± 4.37 b rs7205273 XX 1 2.7 ± 3.71 126.7 ± 35.1 XY 26 4.94 ± 0.72 89.98 ± 6.88 YY 65 3.27 ± 0.46 95.56 ± 4.35 rs239345 XX 8 2.42 ± 1.33 70.82 ± 12.21 a XY 32 3.71 ± 0.66 93.81 ± 6.10 b YY 52 3.96 ± 0.52 98.26 ± 4.78 b rs8044970 XX 4 1.76 ± 1.88 128.87 ± 17.00 XY 29 4.13 ± 0.68 83.83 ± 6.31 YY 59 3.67 ± 0.49 97.13 ± 4.42 rs152733 XX 60 3.13 ± 0.47 94.43 ± 4.57 XY 31 4.77 ± 0.65 94.32 ± 6.35 YY 1 7.97 ± 3.68 88.16 ± 35.36 rs889299 XX 6 3.15 ± 1.52 96.18 ± 14.41 XY 26 2.78 ± 0.72 96.93 ± 6.92 YY 59 4.14 ± 0.48 92.39 ± 4.60 rs4967999 XX 35 3.88 ± 0.64 96.70 ± 5.88 XY 40 3.96 ± 0.58 87.86 ± 5.50 YY 17 2.93 ± 0.92 104.64 ± 8.43 rs6497659 XX 22 3.74 ± 0.70 96.04 ± 7.41 XY 51 3.89 ± 0.47 88.31 ± 4.86 YY 19 2.64 ± 0.76 81.64 ± 7.97

rs9931113 XX 67 3.83 ± 0.40 96.61 ± 4.30 XY 19 2.83 ± 0.75 86.60 ± 8.07 YY 4 2.36 ± 1.63 82.03 ± 17.59 rs4480807 XX 6 3.16 ± 1.36 102.50 ± 14.41 XY 44 3.34 ± 0.5 93.93 ± 5.32 YY 42 3.92 ± 0.52 93.57 ± 5.45 SCNN1G rs4421986 XX 40 3.23 ± 0.53 92.45 ± 5.59 XY 43 3.97 ± 0.51 95.5 ± 5.39 YY 9 3.34 ± 1.11 97.04 ± 11.77 rs4073291 XX 6 2.45 ± 1.35 107.87 ± 14.3 XY 33 3.47 ± 0.58 97.43 ± 6.1 YY 53 3.79 ± 0.46 90.85 ± 4.81 rs4421986 XX 52 3.71 ± 0.46 90.91 ± 4.89 XY 32 3.52 ± 0.59 97.82 ± 6.24 YY 7 2.41 ± 1.25 104.59 ± 13.33 rs13331086 XX 51 3.49 ± 0.47 89.6 ± 4.85 XY 34 3.82 ± 0.57 97.3 ± 5.94 YY 7 3.18 ± 1.26 115.6 ± 13.1 rs4499238 XX 2 3.54 ± 2.36 106 ± 24.83 XY 32 3.41 ± 0.59 89.03 ± 6.21 YY 58 3.69 ± 0.44 96.83 ± 4.61 rs879207 XX 17 3.96 ± 0.80 104.02 ± 8.36 XY 39 3.60 ± 0.53 88.15 ± 5.52 YY 38 3.39 ± 0.55 96.14 ± 5.59 TRPV1 rs4790522 XX 13 4.33 ± 0.91 97.74 ± 9.68 XY 46 3.20 ± 0.49 92.15 ± 5.15 YY 35 3.83 ± 0.57 95.71 ± 5.90 rs224546 XX 39 3.73 ± 0.53 97.03 ± 5.58 XY 40 3.19 ± 0.53 90.95 ± 5.51 YY 15 4.24 ± 0.85 95.83 ± 9.00 rs161364 XX 5 3.58 ± 1.50 97.59 ± 15.62 XY 35 3.90 ± 0.58 97.66 ± 5.90 YY 52 3.40 ± 0.45 90.54 ± 4.84 rs8065080 XX 16 3.36 ± 0.83 74.15 ± 8.38a XY 38 3.58 ± 0.54 100.46 ± 5.44b YY 39 3.66 ± 0.54 97.53 ± 5.40b rs9902581 XX 35 3.72 ± 0.56 91.67 ± 5.90 XY 46 3.22 ± 0.49 95.86 ± 5.15 YY 13 4.50 ± 0.91 95.52 ± 9.68 rs150908 XX 12 3.07 ± 0.99 89.59 ± 10.10 XY 53 3.53 ± 0.45 92.48 ± 4.81 YY 28 3.68 ± 0.62 99.50 ± 6.61 rs224536 XX 7 3.02 ± 1.24 86.42 ± 13.15 XY 44 3.15 ± 0.50 97.78 ± 5.24 YY 33 4.12 ± 0.50 91.92 ± 5.31 rs150846 XX 13 3.83± 0.92 106.26 ± 9.60 XY 39 3.50 ± 0.53 93.71 ± 5.54 YY 42 3.59 ± 0.52 91.04 ± 5.34 rs161386 XX 31 3.18 ± 0.59 92.22 ± 6.26 XY 47 3.27 ± 0.47 93.45 ± 5.08 YY 16 5.26 ± 0.81 100.53 ± 8.71

rs7217945 XX 13 4.91 ± 0.91 109.19 ± 9.54 XY 42 3.31 ± 0.51 90.32 ± 5.31 YY 39 3.43 ± 0.53 93.50 ± 5.51

SE denotes standard error. b significantly different from a (p < 0.05)

Individuals homozygous for the A allele of rs239345 (A>T) had lower iAUC (mean ± SE) (70.82 ± 12.21) than those with either the AT (93.81 ± 6.10) or TT (98.26 ± 4.79) genotypes (p = 0.05). Because of the apparent dominant mode of inheritance, carriers of the T allele were grouped together and the association between genotype and individual’s intensity rating at each suprathreshold concentration was examined (Figure 3.2). As shown in Figure 3.2a, individuals with the AA genotype had a significantly lower iAUC than carriers of the T allele (70.82 ± 12.16 vs. 96.95±3.75, p=0.02). No significant differences in intensity ratings were observed at any one concentration (Figure 3.2b).

Individuals homozygous for the T allele of rs3785368 (C >T) had lower iAUC (57.43 ± 19.85) than those with either the CT (88.60 ± 6.61) or TT (98.60 ± 4.37) genotypes (p = 0.04). Because of the apparent dominant mode of inheritance, carriers of the C allele were grouped together and the association between genotype and individual’s intensity rating at each suprathreshold concentration was examined (Figure 3.3). As shown in Figure 3.3a, individuals with the TT genotype had a significantly lower iAUC than carriers of the C allele (57.43 ± 19.85 vs. 95.57 ± 3.66, p=0.03). Significant differences in intensity ratings were observed across genotypes at 1 mol/L (TT: 65.67 ± 25.68 vs. CT/CC: 125.44 ± 4.71) (Figure 3.3b).

In the TRPV1 gene, the rs8065080 (C>T) SNP was significantly associated with iAUC (mean ± SE) where individuals with the CC genotype (74.15 ± 8.38) had significantly lower iAUCs than those with either the CT (100.46 ± 5.44) or TT (97.53 ± 5.40) genotypes (p = 0.02). Because of the apparent dominant mode of inheritance, carriers of the T allele were grouped together and the association between genotype and individual’s intensity rating at each suprathreshold concentration was examined (Figure 3.4). Individuals with the CC genotype (74.14 ± 8.34) had a significantly lower iAUC than carriers of the T allele (98.3 ± 3.8) (p = 0.008, Figure 3.4a). As shown in Figure 3.4b, significant differences in intensity ratings were observed across genotypes at 1 mol/L (CC: 102.3 ± 11.1 vs. CT/TT: 128.3 ± 5.0), 0.32 mol/L (CC: 71.9 ± 9.2 vs. CT/TT: 99.6 ± 4.2) and 0.1 mol/L (CC: 42.8 ± 9.3 vs. CT/TT: 60.6± 4.2) (p<0.05).

Despite their effect on taste, none of the three SNPs identified was associated with differences in sodium intake (Table 3.3).

Figure 3.2: Comparison of a) incremental area under the taste sensitivity curve (iAUC) and b) taste intensity ratings between individuals homozygous for the A allele and carriers of the T allele for the rs239345 SNP. * p<0.05

Figure 3.3: Comparison of a) incremental area under the taste sensitivity curve (iAUC) and b) taste intensity ratings between individuals homozygous for the T allele and carriers of the C allele for the r rs3785368 SNP. * p<0.05

Figure 3.4: Comparison of a) incremental area under the taste sensitivity curve (iAUC) and b) taste intensity ratings between individuals homozygous for the C allele and carriers of the T allele for the rs8065080 SNP. * p<0.05

Table 3.3: Sodium intake and SCNN1B/TRPV1 genotype1,2

SNP Sodium Intake Gene (rs number) Genotype n (mg sodium ± SE) SCNN1B rs3785368 CC/CT 447 2320 ± 42 TT 83 2269 ± 94 rs239345 AA 43 2336 ± 158 AT/TT 482 2314 ± 40 TRPV1 rs8065080 CC 89 2190 ± 84 CT/TT 452 2327 ± 42

1Values shown are mean ± SE. 2 A general linear model adjusted for age, sex, BMI, and physical activity and caloric intake was used to test differences across genotype.

3.5 Discussion

The present study is the first to examine the association between variations in specific genes and salt taste perception in humans. Our findings show that polymorphisms in the genes that code for the ENaC ß subunit (SCNN1B) and the TRPV1 non-specific cation channel modify an individual’s suprathreshold salt taste perception.

The ENaC blocker amiloride significantly inhibits taste responses to NaCl without affecting responses to stimuli of other taste modalities in mice (Stewart et al., 1997). In humans, amiloride suppresses salt taste by about 20% indicating that human salt taste perception is mediated, at least in part, by ENaC (Feldman et al., 2003; Smith and Ossebaard, 1995). In the present study we found two SNPs within the SCNN1B gene modified suprathreshold taste, potentially supporting the role of this channel in human salt taste perception.

It has been suggested that ENaC facilitates appetitive salt taste, stimulated at low NaCl concentrations, making it more likely that a relationship would exist between variation in ENaC and threshold taste (Chandrashekar et al., 2010). In the present study, however, we did not find any significant associations between SNPs in the SCNN1A, SCNN1B or SCNN1G genes and threshold taste.

In the TRPV1 gene, the rs8065080 (C>T) SNP was found to be associated with an individual’s perception of salt at suprathreshold levels. Carriers of the T allele perceived salt solutions as significantly stronger than those homozygous for the C allele. Since its cloning in 1997, the TRPV1 channel has been studied extensively for its potential role in pain perception (Szallasi et al., 2007). The rs8065080 (C>T) polymorphism is a missense mutation causing an isoleucine (C) to valine (T) amino acid substitution at position 585 of the TRPV1 protein. The role of this polymorphism has been explored in vitro, but did not significantly modify cell activation in response to capsaicin, temperature or pH, suggesting that this polymorphism might not affect the structure or function of the protein (Hayes et al., 2000). However, two studies examining thermal sensation and risk of experiencing pain reported an association between genotype and sensation where the homozygous CC group experienced lower sensitivity or risk of pain than carriers of the T allele (Kim et al., 2004a; Lö tsch J, 2009; Valdes et al., 2011). These findings are consistent with results from the present study where those in the CC group were less

55 sensitive to salt stimuli than carriers of the T allele. These results provide evidence that this polymorphism modifies TRPV1 function in humans where the CC genotype is associated with a lower sensitivity, or decreased response.

Little is known about whether the TRPV1 channel affects appetitive or aversive salt taste and whether functional variation in it influences either salt preference or sodium intake. The present study shows that variation in the TRPV1 gene modifies an individual’s suprathreshold sensitivity indicating TRPV1 is associated with salt perception at higher, potentially aversive, concentrations. Given its role in nociception, and the wide range of stimuli that can lead to its activation, it is possible that the sensation individuals perceive mediated by this channel in response to a salt stimulus are not unique to salt, but rather a subclass of general noxious responses (Stewart et al., 1997). Future studies should aim to examine the attributes of salt taste at these suprathreshold levels and characterize whether individuals perceive the taste stimuli as aversive or pleasurable, salty or painful. If the TRPV1 channel’s role in taste is to prevent the consumption of highly concentrated cation solutions it may have little effect on preference for salt or sodium consumption.

Despite being associated with differences in taste, none of the SNPs examined were associated with differences in sodium intake. However, previous work has shown that correlation between intake assessed by an FFQ and actual sodium consumption is quite low, r = 0.43 for men and 0.11 for women (McKeown et al., 2001). Given the over representation of females within our study population, it is likely that we were not able to adequately capture sodium intake, preventing us from drawing any conclusions on the effect identified SNPs on sodium intake.

In addition to those identified in the measurement of sodium intake, there were a number of limitations to present study. Our coverage of polymorphisms within what is thought to be the most important ENaC associated genes, SCNN1A and SCNN1D was limited. Our study population was not large enough to examine rare variants and all SNPs with a minor allele frequency less than 15 % were excluded from our analysis; these could have had an effect on protein function and taste perception. Despite the use of a rigorous threshold testing method, shown to be stable and repeatable in the short term, it is known that this measure may fluctuate over time (Galindo-Cuspinera et al., 2009; Grzegorczyk et al., 1979). Such variability would

56 increase the noise in our threshold measure and could have masked some potential associations. Finally, the large number of SNPs examined in each gene increases the likelihood of a false positive association and further studies that aim to replicate these findings with TRPV1 and SCNN1B will be needed.

In summary, we examined the association between variation in the SCNN1A, SCNN1B, SCNN1G and TRPV1 genes on both threshold and suprathreshold salt taste as well as sodium intake. In addition to some modest associations between SNPs in the SCNN1B gene and suprathreshold taste, we identified the TRPV1 rs8065080 (C>T) SNP as a likely modifier of an individual’s perception of salt at suprathreshold levels. This finding is supported by a number of lines of evidence and is the first to show that variation in the gene that codes for the TRPV1 channel may explain inter-individual differences in salt taste perception. None of the SNPs identified was associated with differences in sodium consumption.

Chapter 4 Variation in TAS1R2, sweet taste perception and intake of sugars

Adapted from a manuscript by Dias A.G. et al., which will be submitted to The Journal of Nutrigenetics and Nutrigenomics.

4.1 Abstract

The impact of genetic variation in the TAS1R2 gene on sucrose taste perception and intake of sugars in humans was studied in Caucasian men (n=238) and women (n=458) aged 21- 29 years. Of these individuals, 95 participated in a sensory analysis study. Intake was measured using a food frequency questionnaire and 8 polymorphisms (rs12033832, rs12137730, rs35874116, rs3935570, rs4920564, rs4920566, rs7513755, rs9701796) were genotyped. Sucrose taste thresholds were determined using a staircase procedure (solutions: 9x10-6 to 0.5 mol/L). Suprathreshold sensitivity to 0.01-1.0 mol/L sucrose solutions was assessed using general Labeled Magnitude Scales. A significant genotype-BMI interaction was observed for the rs12033832 (G>A) SNP and both suprathreshold sensitivity (p=0.01) and sugar intake (p=0.003). Among those with a BMI ≥ 25, carriers of the G allele had lower sensitivity ratings (mean iAUC ± SE) (GG/GA: 54.4 ± 4.1 vs. AA: 178.5 ± 66.6; p=0.006), higher thresholds (mmol/L) (GG/GA: 9.3 ± 1.1 vs. AA: 4.4 ± 4.3; p=0.004) and consumed more sugars (g/day) (GG/GA: 130 ± 4 vs. AA: 94 ± 13; p=0.009). Among individuals with a BMI < 25, carriers of the G allele had lower thresholds (GG/GA: 8.6 ± 0.5 vs. AA: 16.7 ± 5.7; p=0.02) and consumed less sugars (g/day) (GG/GA: 122 ± 2 vs. AA: 145 ± 8; p=0.004). These findings show that the rs12033832 SNP in TAS1R2 is associated with sucrose taste and sugar intake with those who were more sensitive at detecting sugar consuming less.

4.2 Introduction

Taste is considered one of the primary determinants of food preference and intake (El- Sohemy et al., 2007; Garcia-Bailo et al., 2009a). Given the extent to which taste may influence dietary choices it is important to understand factors that mediate differences in taste function and how they impact food selection and consumption. Inter-individual differences in sucrose detection thresholds have long been recognized. Twin studies have shown that the genetic contribution to the discrimination threshold of a sweet solution was approximately 33% and the contribution to the consumption frequency of sweet foods was 53% (Blakeslee and Salmon, 1935; Henkin and Shallenberger, 1970; Keskitalo et al., 2007). To date, only a few studies have examined the effect of genetic variation in sweet taste receptor genes on either sweet taste perception or sugar consumption in humans and none have identified variants that affect both taste and intake (Eny et al., 2010; Fushan et al., 2010; Fushan et al., 2009; Mennella et al., 2012).

The TAS1R2/TAS1R3 heteromer is considered the major receptor involved in detecting sweet taste (Chandrashekar et al., 2006). The TAS1R2 component is specific to sweet taste perception as TAS1R3 is also involved in the detection of umami when it dimerizes with TAS1R1 (Chandrashekar et al., 2006). Both proteins are co-expressed on the fungiform papillae of the human tongue; however, some TAS1R2 expressing cells do not co-express TAS1R3 (Liao and Schultz, 2003). This is consistent with observations in knockout mice showing that TAS1R2 may act as a low-affinity receptor detecting high levels of natural sugars independent of TAS1R3 (Chaudhari and Roper, 2010). Once a sweet tastant binds to the sweet taste receptor, a G-protein coupled pathway becomes activated (McCaughey, 2008). Evidence from chimera studies suggest that TAS1R2 is specifically involved in G-protein activation (McCaughey, 2008).

One study to date has examined the effect of variation in the TAS1R2 gene on sweet taste (Fushan et al., 2009). Fushan et al. used DNA sequencing to look at the effect of single nucleotide polymorphisms (SNPs) within TAS1R2 on the ability of individuals to discriminate between sucrose solutions of different concentrations (Fushan et al., 2009). None of the 34 variants examined were found to affect sweet taste. However, the study did not account for body mass index (BMI), which may affect sweet taste perception (Ettinger et al., 2012). Further, the taste phenotyping method utilized relied on an individual’s ability to recognize differences between sucrose solutions rather than quantifying more traditional measures of taste like detection thresholds and suprathreshold sensitivity. The latter approaches have recently been shown to be the most appropriate phenotyping methods to examine the effect of genetic variation on taste function (Galindo-Cuspinera et al., 2009). Studies that account for BMI and utilize these taste phenotyping methods are required to better understand the role of variation in TAS1R2 on sweet taste.

Variation in TAS1R2 has also been associated with differences in sugar intake. Eny et al. examined the effect of two missense SNPs, Ser9Cys (rs9701796) and Ile191Val (rs35874116), in the TAS1R2 gene on sugar intake (Eny et al., 2010). There was a significant SNP-BMI interaction of the Ile191Val SNP on sugar intake. Among those with a BMI > 25, Val carriers consumed significantly lower amounts of sugars than Ile/Ile homozygotes. A possible explanation for this is that the Ile191Val variant affects taste perception and subsequent sugar intake. Thus, additional work is required to better understand the mechanism by which TAS1R2 influences sugar consumption.

The objective of the present study was to determine whether genetic variation in TAS1R2 is associated with sucrose taste threshold, suprathreshold taste sensitivity and sugar intake, and if these relationships are modified by BMI.

4.3 Methods

4.3.1 Subjects

Subjects were women (n= 524) and men (n= 251) aged 20-29 years from the Toronto Nutrigenomics and Health Study, a cross-sectional study investigating the effects of gene-diet interactions on biomarkers of chronic disease and genetic determinants of food preferences and intake in a population of young adults. Subjects were excluded from the study if they were pregnant or breastfeeding. Informed consent was obtained from all participants and the study was approved by the University of Toronto Research Ethics Board.

Of these individuals, 28 men and 67 women aged (mean ± SE) 23.7 ± 0.6 years were recruited to take part in a sensory test. Individuals who were smokers, had experienced marked weight changes in the last year (greater than 15 pounds), were diagnosed with chronic sinusitis or chronic obstructive bowel disease, had lost their sense of smell, often experienced severe dry mouth, were diagnosed with diabetes or any other chronic disease or were diagnosed with a psychological disorder were excluded from the study. This study was approved by the University of Toronto and George Brown College research ethics boards. Written informed consent was received from all participants.

Table 4.1: General subject characteristics1

BMI < 25 BMI ≥ 25 p n 542 171 Female 386 (71) 94 (55) <0.0001 Age (years) 23.1 ± 0.1 23.6 ± 0.2 0.04 BMI (kg/m²) 21.8 ± 0.1 28.2 ± 0.2 <0.0001 WC (cm) 72.5 ± 0.2 86.4 ± 0.7 <0.0001 TC (mmol/L) 4.2 ± 0.03 4.3 ± 0.1 0.07 LDL (mmol/L) 2.2 ± 0.03 2.4 ± 0.1 0.0006 HDL (mmol/L) 1.6 ± 0.02 1.4 ± 0.03 <0.0001 Glucose (mmol/L) 4.7 ± 0.01 4.8 ± 0.03 0.0001 Insulin (pmol/L) 39.7 ± 1.1 59.0 ± 4.5 <0.0001 Energy (kcal/day) 2040 ± 28 2120 ± 51 0.14 Fat (% energy/day) 30.6 ± 0.3 30.5 ± 0.5 0.96 CHO (% energy/day) 52.6 ± 0.3 51.1 ± 0.6 0.02 Protein (% energy/day) 16.5 ± 0.1 17.3 ± 0.2 0.002

1Values shown are mean ± SE or number (%).

4.3.2 Dietary Assessment

All subjects completed a one-month, 196-item semi-quantitative food frequency questionnaire (FFQ), which has previously been described (Eny et al., 2009; Eny et al., 2008). Subjects indicated how many times in the past month they consumed a specified portion of each food or beverage and responses were converted to average daily intake for each item. The average daily intakes of all items were analyzed to compute a total daily intake of all major macro- and micro- nutrients for each subject. Individuals who reported daily energy intakes of less than 800 kcal/day or greater than 3500 kcal/day for women and 4000 kcal/day for men (n=43) or reported being on a diet restricting carbohydrate, fat or energy (n=18) were excluded from all analyses to eliminate subjects who may have been ill, dieting, or whose nutrient intakes may be unreliable due to over- or under-reporting intake. Subject characteristics for those included in the dietary analysis are shown in Table 4.1.

4.3.3 Sensory Analysis Protocol

Subjects were asked to attend two separate visits at the study center at 10:00 or 11:30 AM following at least a two hour fast. Subjects were also asked to refrain from drinking coffee or eating strong tasting foods the morning of the test and were requested to repeat the same breakfast before each of the two visits. For each participant, the two visits occurred at the same time of day and on the same weekday, one week apart. Taste thresholds were assessed on the first visit, while suprathreshold taste sensitivity was examined on the second visit.

4.3.4 Threshold Testing

Detection thresholds for sucrose were assessed using a three alternative forced choice up down method (Dias et al., 2013). Participants were asked to rinse their mouth with distilled water prior to the start of the test and between trials. A 10 mL taste solution or distilled water was dispensed into covered plastic cups labeled with 3-digit codes for blinding. During each trial the subject was presented with one cup containing a taste solution and two containing distilled water only. The order of the presentation of cups was randomized. Individuals were instructed to sample all three solutions and identify the cup that was different from the other two. Participants were required to always make a selection and were informed that guessing was acceptable if the solutions seemed indistinguishable.

The concentration of the subject’s first taste solution exposure was 0.028 mol/L sucrose during the initial week of testing and for all subsequent weeks the initial solution concentration was selected based on the previous week’s average threshold for a given gender. Subsequent increases or decreases in taste solution concentration were determined by the participant’s response. If an individual provided a correct response, the stimulus concentration was lowered; an incorrect response resulted in an increase in concentration. The taste solution concentration at which the concentration sequence changed from decreasing to increasing or increasing to decreasing was considered to be a reversal. Five reversals were allowed to occur and an individual’s threshold was calculated as the geometric mean of the final four. Test solutions were 0.25 log cycles apart in concentration, ranging from 9 x 10-6 to 0.5 mol/L of sucrose dissolved in distilled water. Control solutions contained only distilled water. All solutions were prepared within 24 hours of testing and stored at 4 °C. Solutions were presented to participants at room temperature.

4.3.5 Suprathreshold Testing

Suprathreshold taste sensitivity to sucrose was assessed using general Labeled Magnitude Scales (gLMS). Subjects were asked to rate the intensity of five solutions that were 0.5 log cycles apart in concentration, ranging from 0.01 to 1.0 mol/L sucrose. Solutions were presented to subjects in a randomized order using block randomization. Subjects were given a glass of distilled water and were asked to rinse between solutions. All solutions were prepared within 24 hours of testing and stored at 4 °C. Solutions were presented to participants at room temperature.

Prior to beginning suprathreshold testing, subjects were asked to complete a scaling exercise to confirm their ability to rank stimuli of different intensities. Subjects were asked to rank the following sensations in order of decreasing intensity: 1) conversation, whisper, loudest sound experience 2) well-lit room, a dimly lit room, staring at the sun. All subjects completed these tasks correctly and as such, no exclusions were made based on these criteria.

4.3.6 Laboratory Analysis and Genotyping

Each Each participant had venous blood drawn after a 12-hour overnight fast. Plasma total cholesterol, HDL cholesterol, glucose and insulin were measured by Life Laboratories (Toronto, Canada). The Friedewald equation was used to calculate LDL cholesterol. DNA was isolated from whole blood using the GenomicPrep Blood DNA Isolation kit (Amersham

Pharmacia Biotech Inc, Piscataway, NJ). Tag SNPs in TAS1R2 were selected for examination using Haploview Tagger software using the HapMap CEU population (database release 27) under the following parameters: Gene boundaries (Tas1R2 position 19,038,680 to 19,058,742), pair wise comparisons > 500 kilobase pairs apart were ignored, individuals with > 50% missing genotypes were excluded, SNPs with a minor allele frequency < 10% were excluded. Eight SNPs were selected and genotyping was completed for each subject using the Sequenom MassArray Analyzer System (Sequenom, San Diego, CA).

4.3.7 Anthropometrics and Physical Activity

Height, weight, and waist circumference were measured and BMI was calculated (kg/m²). Physical activity over the past month was measured using metabolic equivalent (MET) levels. One MET is the energy expended at rest, and is approximately equal to 1 kcal/kg of body mass per hour at rest. This accounts for the intensity of activity, and thus includes contribution from light activity such as occupational activity, but excludes sleeping.

4.3.8 Statistical Analysis

Block Taste thresholds were calculated by computing the geometric mean of the concentrations at which the final four reversal points in the staircase procedure occurred. Block randomization to determine suprathreshold solution order was completed using Random Allocation Software (Department of Anesthesia, Isfahan University of Medical Sciences, Isfahan, Iran). Individual ratings of the intensity of suprathreshold solutions were plotted and the incremental area under the taste sensitivity curve (iAUC) was computed using GraphPad Prism (Version 5; GraphPad Software Inc, La Jolla, CA). Variables that were not normally distributed were log transformed. General linear models were used to test the effect of genotype on both sensory and dietary outcomes using Statistical Analysis Software (version 9.2; SAS Institute Inc, Cary, NC). Analyses on threshold and suprathreshold taste were adjusted for age, sex and BMI. The interaction between genotype and BMI on taste was determined for each SNP. For SNPs found to be significantly associated with either taste outcome (p<0.05), mode of inheritance was assessed and genotypes were grouped. Where appropriate, analyses were stratified by BMI (≥ 25 and < 25) and the association between grouped SNPs and taste outcomes was examined. For SNPs found to be significantly associated with either taste outcome, the association between genotype and sugar intake was examined. All dietary analyses were adjusted for age, sex, BMI,

65 alcohol intake and physical activity and were stratified by BMI status as previously described (Eny et al., 2010).

4.4 Results

The rs12033832 (G>A) and rs3935570 (G>T) SNPs in the TAS1R2 gene were associated with suprathreshold taste, where carriers of the G allele for each SNPs had a lower suprathreshold sensitivity (iAUC ± SE) than the respective minor allele homozygotes AA (GG 71.7 ± 4.6; GA 57.0 ± 6.4; AA 98.6 ± 11.9; p = 0.02) or TT individuals (GG 67.8 ± 4.8; GT 65.4 ± 5.8; TT 114.8 ± 15.1; p = 0.04) (Table 2). There was a significant SNP-BMI interaction on suprathreshold taste for both the rs12033832 (p = 0.01) and rs3935570 (p = 0.02) (Table 4.2). Mode of inheritance was examined for both SNPs and carriers of the G allele were grouped together for each SNP. Individuals were then stratified by BMI status (≥ 25 or < 25) and the analysis on taste outcomes was repeated with the grouped genotypes (Figures 4.1 and 4.2).

Table 4.2: The association between TAS1R2 polymorphisms and both sucrose taste threshold and suprathreshold taste sensitivity. SNP Genotype n Sugar Suprathreshold taste sensitivity Taste (iAUC ± SE) 1 Threshold (mmol/L ± SE)1 TAS1R2 rs12033832 AA 9 12.4 ± 1.8 98.6 ± 11.9 GA 29 8.2 ± 1.0 57.0 ± 6.4 GG 53 8.9 ± 0.8 71.7 ± 4.6 p-value2 0.77 0.02 p-value: SNP X BMI3 0.55 0.01 rs12137730 AA 39 9.5 ± 0.9 72.3 ± 5.9 CA 36 8.0 ± 0.9 65.7 ± 6.2 CC 10 8.9 ± 1.8 72.6 ± 11.6 p-value2 0.88 0.62 p-value: SNP X BMI3 0.86 0.66 rs35874116 CC 9 11.6 ± 1.9 96.1 ± 11.5 CT 47 9.3 ± 0.8 66.9 ± 5.2 TT 35 8.0 ± 1.0 66.0 ± 5.7 p-value2 0.40 0.15 p-value: SNP X BMI3 0.29 0.14 rs3935570 GG 50 8.9 ± 0.8 67.8 ± 4.8 GT 36 8.5 ± 0.9 65.4 ± 5.8 TT 5 14.2 ± 2.5 114.8 ± 15.1 p-value2 0.49 0.04 p-value: SNP X BMI3 0.28 0.02 rs4920564 GG 17 9.5 ± 1.4 66.6 ± 8.9 GT 47 9.7 ± 0.8 71.4 ± 5.2 TT 27 7.5 ± 1.1 67.8 ± 7.0 p-value2 0.71 0.94 p-value: SNP X BMI3 0.89 0.89 rs4920566 AA 34 9.7 ± 1.0 79.5 ± 6.2 AG 40 9.2 ± 0.9 67.4 ± 5.5 GG 15 6.8 ± 1.5 55.3 ± 9.0 p-value2 0.43 0.11 p-value: SNP X BMI3 0.54 0.18 rs7513755 GG 3 11.0 ± 3.2 76.7 ± 20.5 GT 31 8.4 ± 1.0 65.5 ± 6.5 TT 57 9.2 ± 0.8 71.2 ± 4.7 p-value2 0.73 0.86 p-value: SNP X BMI3 0.75 0.84 rs9701796 CC 57 8.4 ± 0.7 68.7 ± 4.8 GC 25 10.9 ± 1.1 72.1 ± 7.1 GG 9 8.1 ± 1.8 66.8 ± 11.9 p-value2 0.32 0.65 p-value: SNP X BMI3 0.43 0.69

1 Values shown are unadjusted means ± SE 2 results are adjusted for age, sex and BMI

3 p value for the effect of the interaction between genotype and BMI on threshold and suprathreshold taste, adjusted for age and sex

Among individuals with a BMI ≥ 25, carriers of the G allele of rs12033832 had significantly higher detection thresholds (mmol/L ± SE) (GG/GA 9.3 ± 1.1 vs. AA 4.4 ± 4.3, p=0.004) and lower suprathreshold sensitivity ratings (iAUC ± SE) (GG/GA 54.4 ± 4.1 vs. AA 178.5 ± 66.6, p=0.006) (Figure 4.1a and 4.1b). The effect on threshold taste was the opposite among those with a BMI < 25, where carriers of the G allele had significantly lower detection thresholds (mmol/L ± SE) (GG/GA 8.6 ± 0.5 vs. AA 16.7 ± 5.8, p=0.02) (Figure 4.1b). There was no effect on suprathreshold taste among individuals with BMI < 25 (Figure 4.1a). Among individuals with a BMI ≥ 25, carriers of the G allele of rs3935570 had significantly higher detection thresholds (mmol/L ± SE) (GG/GT 9.3 ± 1.1 vs. TT 4.4 ± 4.3, p=0.004) and lower suprathreshold sensitivity ratings (iAUC ± SE) (GG/GT 54.4 ± 4.1 vs. TT 178.5 ± 66.6, p=0.006) (Figure 4.2a and 4.2b). There was no effect on threshold or suprathreshold taste among those with a BMI < 25 (Figure 4.2a and 4.2b).

68 a)

Figure 4.1: Comparison of a) incremental area under the taste sensitivity curve (iAUC ± SE) and b) taste thresholds (mmol/L ± SE) between individuals homozygous for the A allele and carriers of the G allele for the rs12033832 SNP, stratified by BMI.

69 a)

Figure 4.2: Comparison of a) incremental area under the taste sensitivity curve (iAUC ± SE) and b) taste thresholds (mmol/L ± SE) between individuals homozygous for the T allele and carriers of the G allele for the rs3935570 SNP, stratified by BMI.

For each of these two SNPs associated with sucrose taste outcomes, the SNP-BMI interaction for total sugar intake was assessed next. There was a significant interaction for rs12033832 (p = 0.003), but not rs3935570 (p = 0.18). Among those with a BMI ≥ 25, carriers of the G allele of the rs12033832 SNP had a higher intake (g/day) of carbohydrates (GG/GA 277 ± 8 vs. AA 214 ± 23, p=0.03), total sugars (GG/GA 130 ± 4 vs. AA 94 ± 13, p=0.009), sucrose (GG/GA 50 ± 2 vs. AA 36 ± 6, p=0.008), glucose (GG/GA 27 ± 1 vs. AA 19 ± 3, p=0.03), and fructose (GG/GA 29 ± 1 vs. AA 19 ± 3, p=0.02) than individuals who were AA homozygous (Table 4.3). Consistent with the effects observed for threshold taste, the effect was the opposite among those with a BMI < 25. Among those individuals, carriers of the G allele had a lower intake (g/day) of carbohydrates (GG/GA 265 ± 4 vs. AA 292 ± 14, p=0.03), total sugars (GG/GA 122 ± 2 vs. AA 145 ± 8, p=0.004), sucrose (GG/GA 47 ± 1 vs. AA 58 ± 4, p=0.02), glucose (GG/GA 26 ± 1 vs. AA 31 ± 2, p=0.009), fructose (GG/GA 28 ± 1 vs. AA 33 ± 2, p=0.04), and lactose (GG/GA 18 ± 1 vs. AA 22 ± 2, p=0.02) compared to those homozygous for the A allele (Table). Notably, there were no differences in any other macronutrients tested for rs12033832. The rs3935570 had no effect on intake of carbohydrates or sugars (Table 4.4).

Table 4.3: Dietary intake and rs12033832 genotype stratified by BMI1,2

BMI BMI < 25 p BMI ≥ 25 p

Genotype GA/GG AA GA/GG AA n 494 44 152 17 Calories 2027 ± 29 2158 ± 97 0.05 2167 ± 54 1770 ± 160 0.11 Fat (g/day) 69 ± 1 73 ± 4 0.33 73 ± 2 65 ± 7 0.53 Protein (g/day) 84 ± 1 88.0 ± 5 0.27 93 ± 3 79 ± 8 0.27 Carbohydrates (g/day) 265 ± 4 292 ± 14 0.03 277 ± 8 214 ± 23 0.03 Total Sugar (g/day) 122 ± 2 145 ± 8 0.004 130 ± 4 94 ± 13 0.009 Sucrose (g/day) 47 ± 1 58 ± 4 0.02 50 ± 2 36 ± 6 0.008 Glucose (g/day) 26 ± 1 31 ± 2 0.009 27 ± 1 19 ± 3 0.03 Fructose (g/day) 28 ± 1 33 ± 2 0.04 29 ± 1 19 ± 3 0.02 Lactose (g/day) 18 ± 1 22 ± 2 0.02 22 ± 1 18 ± 4 0.27 Maltose (g/day) 2 ± 0.1 2 ± 0.2 0.77 2 ± 0.1 2 ± 0.3 0.89 Alcohol (g/day) 8 ± 1 6 ± 2 0.22 10 ± 1 6 ± 3 0.52

1Values shown are mean ± SE. 2 A general linear model adjusted for age, sex, BMI, physical activity and alcohol intake was used to test differences across genotypes for all nutrients except alcohol intake, which was adjusted for age, sex, BMI and physical activity.

Table 4.4: Dietary intake and rs3935570 genotype stratified by BMI1,2

BMI BMI < 25 p BMI ≥ 25 p

Genotype GT/GG TT GT/GG TT n 497 36 158 10

Calories 2031 ± 29 2124 ± 108 0.27 2148 ± 53 1838 ± 211 0.15

Fat (g/day) 69 ± 1 70 ± 4 0.88 73 ± 2 61 ± 9 0.23

Protein (g/day) 84 ± 1 88 ± 5 0.28 92 ± 3 85 ± 11 0.34

Carbohydrates (g/day) 266 ± 4 288 ± 16 0.13 275 ± 8 228 ± 30 0.12

Total Sugar (g/day) 122 ± 2 136 ± 9 0.11 128 ± 4 105 ± 17 0.17

Sucrose (g/day) 47 ± 1 53 ± 4 0.10 49 ± 2 39 ± 8 0.20

Glucose (g/day) 27 ± 1 30 ± 2 0.07 26 ± 1 22 ± 4 0.26

Fructose (g/day) 28 ± 1 32 ± 2 0.11 28 ± 1 24 ± 5 0.52

Lactose (g/day) 18 ± 1 18 ± 2 0.63 22 ± 1 17 ± 5 0.47

Maltose (g/day) 2 ± 0.1 3 ± 0.2 0.23 2 ± 0.1 2 ± 0.4 0.74

Alcohol (g/day) 10 ± 1 6 ± 3 0.52 8 ± 1 6 ± 2 0.22

1Values shown are mean ± SEM. 2 A general linear model adjusted for age, sex, BMI, physical activity and alcohol intake was used to test differences across genotypes for all nutrients except alcohol intake, which was adjusted for age, sex, BMI and physical activity.

4.5 Discussion

The present study is the first to identify a variant in a sweet taste receptor gene that affects both taste and intake. The rs12033832 SNP in TAS1R2 was found to affect both sweet taste and sugar intake, but the direction of the effect differed by BMI. Among individuals with a BMI ≥ 25, those who were carriers of the G allele had higher sucrose detection thresholds, lower suprathreshold taste sensitivity ratings and a higher intake of total sugars. Among those with a BMI < 25, individuals who were carriers of the G allele had lower detection thresholds and a lower intake of sugars. Though both the rs12033832 and rs3935570 had an effect on taste, only the former was also associated with differences in sugar intake. The effect seen with rs3935570 may be attributed to the five individuals who were homozygous for the minor T allele for rs3935570 in the sensory study also being homozygous for the A allele of rs12033832. As such, the association observed between rs3935570 and taste was likely driven by the overlap among individuals who possessed the minor allele of both SNPs. Indeed, the two SNPs were in moderate linkage disequilibrium in the entire population (r2 = 0.67) and in the subset participating in the sensory study (r2 = 0.64). The relationship between BMI, genotype and taste perception observed in the present study may be related to leptin, which is known to increase sucrose and glucose taste thresholds and higher levels of leptin are associated with higher BMI (Considine et al., 1996; Nakamura et al., 2008). In the present study, among those with a BMI ≥ 25, individuals who were carriers of the G allele of the rs12033832 SNP were less sensitive to sweet taste stimuli, with higher detection thresholds and lower suprathreshold sensitivity ratings. Having a lower sensitivity to sweet compounds may result in individuals choosing foods with more sugar to achieve a sufficient level of sweetness and may increase overall consumption of sugars. Indeed, carriers of the G allele, who were more sensitive to sucrose stimuli, consumed more sugars. Interestingly, among those with a BMI < 25, individuals who were carriers of the G allele were more sensitive, with lower detection thresholds. As expected, these individuals consumed less sugars. The reversal of the effect of being a carrier of the G allele in normal weight and overweight/obese individuals may be the result of differences in leptin sensitivity between these groups. Leptin is thought to interfere with signal transduction during sweet taste perception by increasing outward K+ currents in taste receptor cells, causing the cells to remain in a state of hyperpolarization and diminishing responses to sweet taste stimuli (Kawai et al., 2000). Obesity has been associated

74 with leptin resistance, so among normal weight individuals, leptin may modulate the effect of the AA genotype resulting in lower sensitivity among these individuals (Zheng et al., 2009). In overweight and obese individuals, who may be less sensitive or resistant to leptin, the effect of the AA genotype may be unmasked, allowing increased sucrose sensitivity within this group. Two other studies to date have examined the effect of variation in the TAS1R2 gene on sweet taste and sugar intake (Eny et al., 2010; Fushan et al., 2009). Fushan et al. previously examined the rs12033832 variant and reported no association with sweet taste (Fushan et al., 2009). However, the discrepancy between those findings and those reported in the present study is not unexpected. The previous study did not account for BMI, which was found to be a significant effect modifier of the relationship between genotype and taste in the present study (Fushan et al., 2009). Eny et al. previously reported that, among overweight and obese subjects, the rs35874116 SNP in the TAS1R2 gene was associated with differences in carbohydrate and sugar intake in two population, including the one examined in the present study (Eny et al., 2010). Individuals with the Val residue at amino acid position 191 of TAS1R2 protein consumed less sugar than those who were carriers of the Ile allele at this position (Eny et al., 2010). In the present study, the rs35874116 SNP was not associated with either taste outcome and there was no significant genotype-BMI interaction on either threshold (p = 0.29) or suprathreshold (p = 0.14) taste. Subsequently, it would appear that though this SNP is associated with differences in carbohydrate and sugar intake, this may not be driven by differences in taste perception. Given TAS1R2 is also expressed in the small intestine, it is possible that this SNP may elicit its effect on consumption via a post-ingestive mechanism (Young et al., 2009). The present study had a number of limitations. The method of SNP selection excluded any SNPs that occurred at a low frequency within the population (≤ 10 % MAF). It is possible that a SNP that is less common may affect the function or expression of TAS1R2 and modify taste and intake to a greater extent than those identified here. As such, additional studies that look at the effect of a greater number of variants in TAS1R2 on sucrose taste and sugar intake are required. Additionally, the limited sample size of the sensory study resulted in a small number of individuals in the minor allele groups. The size of groups was further reduced after stratification by BMI status, with only 2 individuals with a BMI ≥ 25 who were homozygous for the AA allele for the rs12033832 SNP. This remained an issue in the dietary analysis where there were only 17 individuals in the same minor allele group. Despite these limitations, we identified one SNP rs12033832 which was associated with two related but distinct indices of sweet taste; taste

75 perception and sugar intake. Importantly, this SNP was not associated with intakes of protein, fat or alcohol, demonstrating the specificity of our results. In summary, the present study is the first to identify a variant in the TAS1R2 sweet taste receptor gene that is associated with differences in both sucrose taste and sugar intake. For the SNP identified, rs12033832, a significant gene-BMI interaction was found for both suprathreshold taste and sugar intake. These findings suggest that variation in the TAS1R2 gene affects sucrose taste and sugar consumption and that this effect is modified by BMI.

Chapter 5 Genetic variation in TAS2R19 affects naringin taste sensitivity and is associated with grapefruit preference and intake

Adapted from a manuscript by Dias A.G. et al., which has been submitted to The Journal of Nutrition.

5.1 Abstract

Differences in an individual’s preferences and intake of grapefruit are influenced by their sensitivity to the bitter tastant naringin. The TAS2R19 receptor may be involved in naringin detection, but the effect of variation in the TAS2R19 gene on naringin taste sensitivity and grapefruit consumption is not known. In order to determine the effect of variation in the TAS2R19 gene on sensitivity to naringin and grapefruit preference and intake, male (n=487) and female (n=1,058) participants from the Toronto Nutrigenomics and Health study, aged 20-29 years, were selected. Intake was measured using a 196-item food frequency questionnaire and preference was assessed using a 9-point hedonic scale. A subset of individuals (n=685) rated their sensitivity to filter papers infused with 3 µg of naringin on a 9-point scale. Five TAS2R19 polymorphisms were genotyped using the Sequenom MassArray system. The rs10772420 (A>G) and rs4763235 (C>G) polymorphisms were associated with naringin sensitivity and grapefruit and grapefruit preference. The rs10772420 was also associated with grapefruit and grapefruit juice intake. For rs10772420, the odds ratios (ORs (95% CI)) of perceiving naringin as “high intensity” were 1.00 (reference), 0.64 (0.39-1.04) and 0.22 (0.13-0.38) for the AA, AG and GG genotypes, respectively. The ORs (95% CIs) of disliking grapefruit for the AA, AG and GG genotypes were 1.00, 0.47 (0.33-0.67) and 0.21 (0.14-0.32), respectively. The corresponding ORs (95% CIs) for disliking grapefruit juice were 1.00, 0.57 (0.41-0.81) and 0.37 (0.25-0.53). The polymorphism was also associated with grapefruit and grapefruit juice consumption where the ORs (95% CIs) of not consuming grapefruit for the AA, AG and GG genotypes were 1.00, 0.62 (0.44-0.86) and 0.63 (0.45-0.89), and for not consuming grapefruit juice was 1.00, 0.77 (0.52-1.13) and 0.64 (0.43-0.96). Our findings show that variation in TAS2R19 affects naringin taste sensitivity, and grapefruit preference and intake.

5.2 Introduction

Sensitivity to bitter taste is believed to have evolved in mammals to deter them from ingesting toxic compounds (Drewnowski and Gomez-Carneros, 2000). However, not all bitter compounds are toxic when consumed in amounts typically found in our diets (Garcia-Bailo B, 2008). While very strong bitter tastes generally cause rejection, milder ones are expected and can be enjoyed (Meyerhof, 2005; Tepper, 2008). Common sources of bitter tastants include widely

78 consumed plants (spinach, endives, cruciferous vegetables), fruits (grapefruit), beverages (beer, green tea and coffee) and numerous other foods (sharp cheeses, soy products) (Anliker et al., 1991; Drewnowski, 2001). In many cases the bitter tasting compounds present in these foods are potentially beneficial phytochemicals (El-Sohemy et al., 2007). Individuals who perceive these compounds as more intensely bitter may avoid consuming them, and this could affect their nutritional and health status.

Bitter taste is mediated by the taste receptor type 2 family of receptors (TAS2R). In humans, TAS2Rs are encoded by about 25 genes located on chromosomes 5, 7 and 12 (Drayna, 2005; Meyerhof et al., 2010). Previously it was thought that a large number of bitter taste receptors were required to enable the recognition of a wide range of compounds, but not necessarily differentiate between them (Drewnowski et al., 1997). However, this leaves it difficult to explain why some individuals can tolerate certain bitter foods, but not others. A number of studies have recently shown that response to bitter stimuli is diverse and compound- dependent (Hayes et al., 2011; Meyerhof et al., 2010). Further, variations in TAS2R genes may explain inter-individual differences in bitterness perception and preference (Hayes et al., 2011).

Recently, Hayes et al. showed that the rs10772420 SNP in the TAS2R19 gene, which results in an Arg299Cys amino acid substitution, is associated with perceived bitterness, sweetness and liking of grapefruit juice (Hayes et al., 2011). Cys homozygous individuals rated the juice as more bitter and less sweet/likable than Arg/Arg individuals. This indicates that grapefruit juice liking may be a function of an individual’s sensitivity to bitter compounds in grapefruit juice. As the primary bitter tastant in grapefruit, the compound naringin may be a ligand for the TAS2R19 receptor and naringin taste sensitivity may be modified by the rs10772420 SNP.

The aim of the present study was to examine the effect of variation in the TAS2R19 gene on naringin taste sensitivity and grapefruit/grapefruit juice preference and intake.

5.3 Methods

5.3.1 Subjects

The Toronto Nutrigenomics and Health Study is an ethno-culturally diverse population of men (n=487) and women (n=1,058) ages 20–29 years, recruited from the University of Toronto campus. All subjects completed a general health and lifestyle questionnaire, which included information on subject characteristics such as age, sex, medical history, smoking status and ethno-cultural group. Subjects were grouped into Caucasian, East Asian, South Asian or other (all other ethno-cultural groups and those reporting more than one ethno-cultural group) based on their responses. Height, weight and physical activity were measured as previously described (Cahill et al., 2009). Subject characteristics are shown in Table 5.1.

5.3.2 Food Preference Assessment

Subjects completed a 63-item FPC that included both grapefruit and grapefruit juice. The FPC was modeled after a 171-item FPC that is widely used to assess food preferences in North America (Drewnowski et al., 2007). The FPC requires subjects to rank their liking/disliking of common foods on an anchored 9-point hedonic scale ranging from “dislike extremely” to “like extremely”. Subjects were able to indicate if they “never tried” or “would not try” a food and they were excluded from the analysis if they selected either option for grapefruit or grapefruit juice. Food preference was grouped into two categories; dislike (1=dislike extremely, 2= dislike very much, 3=dislike moderately) and neutral/like (4=dislike slightly, 5=neither like nor dislike, 6=like slightly, 7=like moderately, 8=like very much, 9=like extremely).

Table 5.1: General subject characteristics Characteristics Female1 Male1 n (%) 1058 (68) 487 (32) Age (years) 22.6 ± 0.1 22.9 ± 0.1 BMI (kg/m2 ) 22.5 ± 0.1 23.9 ± 0.2 Physical Activity 12.2 ± 0.1 12.5 ± 0.1 (MET.hrs/week) Energy Intake (kcal/day) 1936 ± 24 2359 ± 45 Grapefruit Dislikers (%) 174 (16) 82 (17) Grapefruit Juice Dislikers (%) 217 (21) 97 (19) Grapefruit (% did not 640 (60) 299 (61) consume) Grapefruit Juice (% did not 830 (78) 352 (72) consume) Taste sensitivity study n (%) 463 (67) 224 (33) Naringin High Intensity 178 (38) 57 (25) Tasters (%)

1 Values are reported as mean ± SE or number (%)

5.3.3 Dietary Intake Assessment

Dietary intake was assessed by using the Toronto-modiﬁed Willett questionnaire, which is a 196-item semi-quantitative food frequency questionnaire. Participants were asked to report their average intake over the past month of speciﬁc foods, beverages, and dietary supplements by choosing among several frequency options. A commonly used portion size (e.g. 0.5 cup) was pre-assigned to each item. Estimates of intake were based on the USDA Nutrient Database for Standard Reference. Dietary intake of grapefruit and grapefruit juice was grouped into two categories; “did not consume” (never) and “consumed” (less than once per month, 1-3 times per month, once per week, 2-4 times per week, 5-6 times per week, once per day, 2-3 servings per day, 4 or more servings per day).

5.3.4 Bitter Taste Assessment

A subset (n = 685) of the study population underwent a naringin taste sensitivity test. Filter papers infused with 3 µg of naringin (Precision Laboratories, Inc., Cottonwood, AZ) were used to assess subjects’ perceived bitter taste intensity. Upon arrival at the study center subjects were asked to rinse their mouth with room temperature bottled spring water before filter paper administration. A control strip with no bitter substance was placed on the antero-medial surface of the tongue until completely moistened. Subjects were asked to circle the number corresponding to their perceived bitter taste intensity on a 9-point numbered scale ranging from “1=not at all bitter” to “9=extremely bitter” with a central point labeled “5=moderately bitter”. This process was repeated with a filter paper infused with naringin. Bitter taste intensity was grouped into two categories; medium/low intensity (1-6) and high intensity (7-9).

5.3.5 Genotyping

A total of 5 SNPs in the TAS2R19 gene (Table 5.2) and 6 SNPs in the TAS2R50 gene (Figure 5.2) were genotyped using the Sequenom MassArray Analyzer System (Sequenom, San Diego, CA). LD between TAS2R19 SNPs in this population was assessed and two SNPs, rs10772420 and rs12313469, were in significant LD (r2 = 0.98) (Figure 5.1). As a result only one of the two, the nonsynonymous rs10772420 variant, was analyzed further. LD between two TAS2R19 SNPs (rs10772420 and rs4763235), found here to be associated with naringin taste sensitivity, and SNPs in TAS2R50 was assessed (Figure 5.2).

Table 5.2: TAS2R19 SNP Characteristics Major/Minor SNP ID Mutation Type Allele MAF1 (%) rs10772420 2 Arg299Cys A/G 36 rs12313469 2 Ala84Ala A/G 36 rs12424373 Lys126Gln G/T 6 rs1868769 Leu140Leu A/G 20 rs4763235 Noncoding C/G 48

1 MAF: Minor Allele Frequency 2 indicates SNPs in significant LD with one another that will be collectively represented by rs10772420 SNP, the non synonymous variant, for the remainder of the paper

Figure 5.1: Linkage disequilibrium (LD) for SNPs with a minor allele frequency > 5% in

TAS2R19. Each square shows the pairwise comparison between the two SNPs on either side of the square on the diagonal. LD is assessed in this figure by r2. For SNP pairs with an r2 > 0.80 only a single SNP was analyzed. Excluded SNPs are identified in Table 1 with 2.

Figure 5.2: Linkage disequilibrium (LD) between TAS2R19 SNPs (rs10772420, rs4763235) and

6 SNPs from TAS2R50. Each square shows the pairwise comparison between the two SNPs on either side of the square on the diagonal. LD is assessed in this figure by r2.

5.3.6 Statistical Analysis

Statistical analyses were conducted using SAS version 9.2 (SAS Institute, Cary, NC). The associations between genotypes and naringin taste sensitivity, intake and preference were assessed using binary logistic regression. Naringin analysis was adjusted for age, sex, ethnicity and smoking status. SNPs that were significantly associated with naringin taste sensitivity were tested against preference and intake. Preference analyses were adjusted for age, sex, ethnicity, BMI, smoking status, and physical activity. The intake model was adjusted for age, sex, ethnicity, BMI, energy intake, smoking status and season.

5.4 Results

Table 5.3 shows the effect of each SNP on an individuals’ naringin taste sensitivity. Two SNPs were significantly associated with sensitivity, rs10772420 and rs4763235. Individuals who were carriers of the G allele for the rs10772420 (A>G) and the rs4763235 (C>G) SNPs were less likely to perceive the taste from the naringin infused filter paper as being “high intensity.” For rs10772420, the odds ratios (ORs (95% CIs) of perceiving naringin as “high intensity” were 1.00 (reference), 0.64 (0.39-1.04) and 0.22 (0.13-0.38) for the AA, AG and GG genotypes, respectively. For rs4763235, the ORs (95% CIs) for the CC, CG and GG genotypes were 1.00, 0.70 (0.47-1.03) and 0.30 (0.19-0.49), respectively.

Table 5.3: Odds of perceiving naringin as “high intensity” by TAS2R19 genotype Naringin Taste Intensity Medium High Odds Ratio and Low Intensity (OR (95% CI)) SNP ID Intensity Tasters Tasters Model 11 Model 22 n (%) n (%) rs10772420 n = 450 n = 235 AA 43 (10) 48 (21) 1.00 1.00 AG 175 (39) 123 (52) 0.63 (0.39-1.01) 0.64 (0.39-1.04) GG 232 (51) 64 (27) 0.25 (0.15-0.41) 0.22 (0.13-0.38) rs4763235 n = 448 n = 235 CC 101 (23) 84 (36) 1.00 1.00 CG 177 (39) 104 (44) 0.71 (0.48-1.03) 0.70 (0.47-1.03) GG 170 (38) 47 (20) 0.33 (0.22-0.51) 0.30 (0.19-0.49) rs1242437 n = 445 3 n = 233 TT 380 (85) 210 (90) 1 1 TG 65 (15) 23 (10) 0.64 (0.39-1.06) 0.68 (0.40-1.16) rs1868769 n = 451 n = 234 AA 279 (62) 161 (69) 1 1 AG 151 (33) 65 (28) 0.75 (0.53-1.06) 0.71 (0.49-1.02) GG 21 (5) 8 (3) 0.66 (0.29-1.53) 0.52 (0.22-1.22)

1 Unadjusted 2 Adjusted for age, sex, ethnicity and smoking status.

Both the rs10772420 and the rs4763235 SNPs were significantly associated with grapefruit and grapefruit juice preference ratings (Table 5.4). For rs10772420, the ORs (95% CIs) of disliking grapefruit for the AA, AG and GG genotypes were 1.00, 0.47 (0.33-0.67) and 0.21 (0.14-0.32), respectively. The corresponding ORs (95% CIs) for disliking grapefruit juice were 1.00, 0.57 (0.41-0.81) and 0.37 (0.25-0.53). For rs4763235, the ORs (95% CIs) of disliking grapefruit for the CC, CG and GG genotypes were 1, 0.54 (0.39-0.74) and 0.32 (0.21-0.49), respectively. The corresponding ORs (95% CIs) for disliking grapefruit juice were 1, 0.65 (0.48- 0.87) and 0.56 (0.39-0.81).

The rs10772420 SNP was found to be significantly associated with grapefruit and grapefruit juice intake where carriers of the G allele were less likely to have refrained from consumption (Table 5.5). The ORs (95% CIs) for not consuming grapefruit for the AA, AG and GG genotypes were 1.00, 0.62 (0.44-0.86) and 0.63 (0.45-0.89), respectively. The corresponding ORs (95% CIs) for grapefruit juice were 1.00, 0.77 (0.52-1.13) and 0.64 (0.43-0.96).

Table 5.4: Odds of disliking grapefruit and grapefruit juice by TAS2R19 genotype Neutral/ Odds Ratio Dislike SNP ID Like (OR (95% CI)) n (%) n (%) Model 11 Model 22 Grapefruit Preference

rs10772420 n = 1283 n=251 AA 154 (12) 71 (28) 1 1 AG 535 (42) 118 (47) 0.48 (0.34-0.68) 0.47 (0.33-0.67) GG 594 (46) 62 (25) 0.23 (0.15-0.33) 0.21 (0.14-0.32) rs4763235 n = 1282 n=251 CC 318 (25) 105 (42) 1 1 CG 546 (42) 99 (39) 0.55 (0.40-0.75) 0.54 (0.39-0.74) GG 418 (33) 47 (19) 0.34 (0.23-0.50) 0.32 (0.21-0.49) Grapefruit Juice Preference

rs10772420 n = 1225 n=309 AA 150 (12) 75 (24) 1 1 AG 513 (42) 140 (45) 0.55 (0.39-0.76) 0.57 (0.41-0.81) GG 562 (46) 94 (31) 0.34 (0.24-0.48) 0.37 (0.25-0.53) rs4763235 n = 1224 n=309 CC 305 (25) 118 (38) 1 1 CG 524 (43) 121 (39) 0.60 (0.45-0.80) 0.65 (0.48-0.87) GG 395 (32) 70 (23) 0.46 (0.33-0.64) 0.56 (0.39-0.81)

1 Unadjusted 2 Adjusted for age, sex, ethnicity, BMI, physical activity, and smoking status.

Table 5.5: Odds of not consuming grapefruit and grapefruit juice by TAS2R19 genotype Consumed Did not Odds Ratio SNP ID consume (OR (95% CI)) n (%) n (%) Model 11 Model 22 Grapefruit Intake rs10772420 n = 603 n = 931 AA 72 (12) 153 (16) 1 1 AG 272 (45) 381 (41) 0.66 (0.48-0.91) 0.62 (0.44-0.86) GG 159 (43) 397 (43) 0.72 (0.52-0.99) 0.63 (0.45-0.89) rs4763235 n = 603 n=930 CC 157 (26) 266 (29) 1 1 CG 276 (46) 369 (39) 0.79 (0.61-1.02) 0.77 (0.59-1.00) GG 170 (28) 295 (32) 1.02 (0.78-1.35) 0.97 (0.71-1.32) Grapefruit Juice Intake rs10772420 n = 363 n = 1171 AA 44 (12) 181 (15) 1 1 AG 151 (42) 502 (43) 0.81 (0.56-1.18) 0.77 (0.52-1.13) GG 168 (46) 488 (42) 0.71 (0.49-1.03) 0.64 (0.43-0.96) rs4763235 n = 362 n = 1171 CC 91 (25) 332 (28) 1 1 CG 158 (44) 487 (42) 0.85 (0.63-1.13) 0.83 (0.61-1.13) GG 113 (31) 352 (30) 0.85 (0.62-1.17) 0.79 (0.55-1.12)

1 Unadjusted 2 Adjusted for age, sex, ethnicity, BMI, physical activity, energy intake, season and smoking status.

5.5 Discussion

The present study is the first to identify naringin as a potential TAS2R19 ligand and our findings show that variation in the TAS2R19 gene is associated with differences in naringin taste sensitivity and both liking and intake of grapefruit and grapefruit juice. The rs10772420 and rs4763235 polymorphisms in the TAS2R19 gene were associated with naringin taste sensitivity and both grapefruit/grapefruit juice preference. The rs10772420 polymorphism was also associated with a significant difference in grapefruit and grapefruit juice intake.

Our findings suggest that naringin may be a ligand of the TAS2R19 receptor. The potential relationship between 25 TAS2Rs, including TAS2R48 (the previous name of TAS2R19), and naringin was previously examined (Meyerhof et al., 2010), but no association was identified. However, the authors of that study acknowledged that in transfected cells meant to express TAS2R48, only marginal expression was observed, potentially limiting their ability to identify ligands.

Two studies to date have examined the role of the rs10772420 SNP in bitter taste. The rs10772420 is an A>G polymorphism resulting in a Cysteine (Cys) to Arginine (Arg) amino acid substitution at position 299 of the TAS2R19 protein. Hayes et al. showed that those who are Arg299 homozygous (GG individuals) are less sensitive to the bitter taste of grapefruit and had higher liking ratings (Hayes et al., 2011). Those findings are consistent with the results of the present study where individuals with the GG genotype were less likely to dislike grapefruit or grapefruit juice, less likely to avoid grapefruit consumption and were less likely to perceive the taste of naringin as highly intense.

Reed et al. have previously demonstrated that quinine may be another potential TAS2R19 ligand and that the rs10772420 SNP is associated with quinine taste (Reed et al., 2010). Consistent with what we observed with naringin in the present study, the AA genotype (Cys) was associated with more intense taste perception. This supports the hypothesis that the Arg>Cys amino acid substitution caused by the rs10772420 SNP alters the sensitivity of TAS2R19 where AA (Cys/Cys) homozygotes are more sensitive to potential TAS2R19 ligands, like quinine and naringin, than the GG (Arg/Arg) genotype.

Bitter foods are thought to have a number of putative health benefits (Shiffman et al.,

2005). One of the outcomes of investigating the role of genetic variants on an individual’s preference and consumption of bitter foods is to help identify “bitter sensitive” sub-populations who may be at risk of under consuming bitter foods. The rs1376251 SNP in the TAS2R50 gene has been associated with risk of myocardial infarction in two studies (Shiffman et al., 2005; Shiffman et al., 2008). Given the SNP’s location in a bitter taste receptor it has been hypothesized that this SNP may influence health outcomes by modifying an individual’s bitter taste sensitivity and in turn dietary choices (Shiffman et al., 2005). However, the TAS2R50 ligands currently identified such as amarogentin and andrographolide, are not present widely in the diet making it difficult to attribute this reduction in disease risk to differences in taste perception of these particular compounds (Meyerhof et al., 2010). Interestingly, rs1376251 in TAS2R50 is in LD (r2 = 0.99, Figure 2) with the rs4763235 SNP in TAS2R19, identified here as being associated with grapefruit preference, grapefruit juice preference and naringin taste sensitivity. We examined the extent of LD between SNPs in TAS2R50 and TAS2R19 and found that both SNPs reported here to be associated with naringin taste sensitivity are in high LD with TAS2R50 SNPs (rs10772420: rs2900554 (r2 = 0.99), rs4763235: rs1376251 (r2 = 0.99), rs7312202 (r2 = 0.94)). This may indicate that the health benefits attributed to TAS2R50 SNP rs1376521 are driven by its LD with taste modifying SNPs in the TAS2R19 gene or that naringin can bind to the TAS2R50 receptor and variation in the TAS2R50 gene may result in differences in naringin taste sensitivity. The latter scenario would indicate that SNPs identified in the present study do not result in changes in naringin taste sensitivity directly, but rather are in LD with TAS2R50 SNPs that are directly related to the observed effects. Further work is needed to determine whether naringin is a ligand for TAS2R19, TAS2R50 or both.

A significant disparity was observed between the proportion of individuals who did not consume grapefruit (Female: 61%, Males: 63%) and the proportion of individuals who dislike grapefruit (Female: 16%, Male: 17%) in the total population. This may indicate that the primary drivers of intake are not preference and bitterness sensitivity, but rather availability or other extraneous factors. This may, in part, explain why, though both rs10772420 and rs4763235 SNPs were found to be associated with preference and naringin taste intensity, but only rs10772420 was associated with grapefruit and grapefruit juice intake.

The present study had a number of limitations. The scales used to assess naringin taste sensitivity and grapefruit and grapefruit juice preference have been criticized because they do not

92 exhibit ratio properties, may reduce variability in responses due to ceiling effects and may not be uniformly understood due to the use of specific descriptors (Hein et al., 2008). However, in the context of food preference, the degree of variability in response, reliability of the scale, ease of use and ability to discriminate preferences between samples when using the 9-point hedonic scale was comparable and occasionally superior to other food preference scales (line, scanner and magnitude estimation) (Lawless and Malone, 1986). The use of a FFQ to measure grapefruit and grapefruit juice consumption over the past month does not take into account seasonal variation in food availability and intake. Although we adjusted for seasonality in the analysis, this may, in part, explain the discrepancy in dislike and intake observed. Further, the FFQ also did not differentiate between grapefruit juice and juice blends, which likely increased the variability in our data, increasing the likelihood of false negative findings. Finally, the coverage of polymorphisms in the TAS2R19 gene was not exhaustive. The potential for other, possibly more influential, variants in this gene exist.

In summary, we examined the association between variation in the TAS2R19 gene and both intake and preference for grapefruit and grapefruit juice as well as naringin taste sensitivity. We identified two SNPs (rs10772420 and rs4763235 SNP) that were associated with naringin taste sensitivity and preference for grapefruit and grapefruit juice. The rs10772420 SNP was also associated with grapefruit and grapefruit juice intake. Our findings suggest that naringin may be a TAS2R19 ligand and that sensitivity to naringin may mediate grapefruit intake and preference. Additional research is required to understand the potential impact that differences in consumption, driven sensitivity to tastants like naringin, has on an individual’s health.

Chapter 6 Genetic variation in CD36 is associated with oleic acid taste sensitivity and habitual fat intake

Adapted from a manuscript by Dias A.G. et al., which will be submitted to Chemical Senses.

6.1 Abstract

The objective of this study was to determine whether variation in the CD36 gene influences fat taste perception and habitual fat intake in humans. Participants were Caucasian men (n=238) and women (n=458) from the Toronto Nutrigenomics and Health study aged 20-29 years, with 21 subjects participating in the sensory study. Dietary intake was measured using a food frequency questionnaire and CD36 genotypes were determined using the Sequenom MassArray system. Taste thresholds were determined using a 3-AFC staircase test (solution range: 0.00028 to 5 % oleic w/w). Suprathreshold taste sensitivity to oleic acid solutions (0.5 to 5 %w/w) was assessed using general Labeled Magnitude Scales. Two CD36 SNPs, rs1761667 (G>A) and rs1984112 (A>G), were significantly associated with the percent of energy consumed from polyunsaturated, monounsaturated and total fat. Combining genotypes associated with the highest (GG/GG) and lowest (AA/AA) consumption of fat intake enhanced the effect (mean % of energy from fat ± SE); polyunsaturated (AA/AA: 5.5 ± 0.1 vs GG/GG: 6.0 ± 0.2; p = 0.03), monounsaturated (AA/AA: 11.6 ± 0.3 vs GG/GG: 12.9 ± 0.4; p = 0.004), and total fat (AA/AA: 26.8 ± 0.4 vs GG/GG: 29.0 ± 0.6; p = 0.004). Sensory analysis on 9 AA/AA and 12 GG/GG individuals revealed that AA/AA individuals perceived the taste of oleic acid solutions as less intense (mean iAUC ± SE) (AA/AA: 80.9 ± 13.9 vs GG/GG: 208.0 ± 48.9; p = 0.001). These findings document that variation in CD36 is associated with differences in habitual fat intake and this may be driven by differences in perceived fat taste intensity.

6.2 Introduction

Variation in fat intake exists both within and between populations (Hunter et al., 1996). Taste is one of the primary mediators of food intake and potential differences in fat orosensory perception may explain differences in nutrient intake (Drewnowski, 1997a). Fat detection has traditionally been attributed to sensory qualities such as texture and olfaction (Drewnowski, 1997b). However, blocking the ability to detect these sensory cues does not prevent the recognition of dietary fat (Fukuwatari et al., 2003). This suggests that another orosensory mechanism contributes to fat detection and this has been referred to as ‘fat taste’ (Abumrad, 2005; Laugerette et al., 2007; Laugerette et al., 2005). A number of groups have shown that humans are able to detect trace amounts of free fatty acids (FFAs) on the tongue and these may

95 be oral chemosensory cues that contribute to fat detection (Chale-Rush et al., 2007; Mattes, 2009c). The addition of orlistat, an inhibitor of lingual lipase, to fat emulsions diminishes an individual’s detection thresholds for the triglyceride triolein, but not its FFA form oleic acid (Pepino et al., 2012).This provides evidence that the fatty acid is a signaling stimulus and that it must be in free from to interact with fat taste receptors. The CD36 integral membrane glycoprotein is one of a number of potential fat taste receptors and has been implicated in fat taste detection in humans (Garcia-Bailo et al., 2009a; Keller et al., 2012; Mattes, 2009a; Pepino et al., 2012). CD36 is expressed in the taste bud cells of a number of species including humans (Simons et al., 2011). The protein spans the cell membrane forming a large extracellular hydrophobic loop, which is the region that likely senses FFAs (Abumrad et al., 1993). This protein structure is consistent with an apical taste receptor. Although the precise mechanism by which CD36 contributes to FFA detection or scaling is not clear, but it has a higher affinity for long chain fatty acids (LCFAs) (Baillie et al., 1996). It plays a role in FFA transport across selected cell membranes, but this has not been established for taste receptor cells (Harmon and Abumrad, 1993). Both saturated and unsaturated LCFAs applied to CD36-positive cells can cause subsequent depolarization, increases in intracellular calcium and potentially taste perception. This response can be attenuated by a CD36-specific inhibitor such as sulfo-N- succinimidyl oleic acid ester (Gaillard et al., 2008). Direct evidence for a role of CD36 in mediating fat taste comes from studies in CD36- null mice. These mice have a diminished ability to select a diet containing FFAs over a texturally similar control diet, whereas their wildtype littermates are more likely to choose a diet containing FFAs (Laugerette et al., 2005; Sclafani et al., 2007a). The absence of preference for fat among CD36-null mice is not due to a general loss of taste, since these mice continue to prefer sweet and avoid bitter stimuli (Laugerette et al., 2005). This suggests that CD36 may contribute to oral FFA sensing and that this is associated with a predilection for fat. To date, few studies have investigated the effect of CD36 genotypes on human ingestive behaviors or fat taste. Keller et al found that, among African Americans, the rs1761667 single nucleotide polymorphism (SNP) was associated with differences in the perceived creaminess of salad dressing and liking of certain types of fats (Keller et al., 2012). Pepino et al found that the rs1761667 SNP was associated with an individual’s ability to detect oleic acid and triolein (Pepino et al., 2012). Those homozygous for the G allele had an 8 –fold lower detection

96 threshold than those who were AA homozygous. There was no effect of this genotype on fat intake, though the study may have been underpowered for this outcome (n=21).

The aim of the present study was to expand the investigation of relationships between CD36 genotypes, fat intake and fat taste indices (i.e., threshold sensitivity and suprathreshold scaling). The study was conducted in two parts. First, the effect of variation in the CD36 gene on habitual fat intake in humans was assessed. Variants that were significantly associated with fat intake were then tested in the second stage to determine if they affected an individual’s threshold or suprathreshold taste sensitivity to oleic acid.

6.3 Methods

6.3.1 Study 1: CD36 and habitual fat consumption

6.3.1.1 Population

Subjects were Caucasian women (n= 458) and men (n= 238) ages 20-29 years from the Toronto Nutrigenomics and Health Study and were recruited from the University of Toronto campus (Table 1). Informed consent was obtained from all participants and the study was approved by the University of Toronto Research Ethics Board. Subjects were excluded from the analysis if they were pregnant or breastfeeding, if their reported energy intake was below 800 kcal/day or exceeded 3500 kcal/day for females and 4000 kcal/day for males, or if they reported being on a diet restricting carbohydrate, fat or calories.

Table 6.1: General subject characteristics1

Male Female n 238 (34) 458 (66) Age (years) 23.4 ± 0.2 23.1 ± 0.1 BMI (kg/m²) 23.9 ± 0.2 23 ± 0.2 WC (cm) 81.6 ± 0.5 72.7 ± 0.3 TC (mmol/L) 4.0 ± 0 4.4 ± 0 LDL (mmol/L) 2.2 ± 0 2.3 ± 0 HDL (mmol/L) 1.3 ± 0 1.7 ± 0 Glucose (mmol/L) 4.9 ± 0 4.7 ± 0 Insulin (pmol/L) 36.8 ± 1.4 47.5 ± 2 Energy (kcal/day) 2314 ± 46 1937.9 ± 27.6 Fat (% energy) 30.2 ± 0.4 30.7 ± 0.3 MUFA (% energy) 12 ± 0.2 12.4 ± 0.2 PUFA (% energy) 5.5 ± 0.1 5.7 ± 0.1 SFA (% energy) 9.9 ± 0.2 9.9 ± 0.1 Protein (% energy) 16.5 ± 0.2 16.7 ± 0.1 Carbohydrates (% energy) 51.7 ± 0.5 52.6 ± 0.4

1 Values shown are mean ± SE or number (%).

6.3.1.2 Dietary Assessment, Anthropometrics and Physical Activity

Subjects completed a self-administered Toronto-modified Willett semi-quantitative 196- item food frequency questionnaire (FFQ) to measure habitual dietary intake over the past month. Subjects were instructed on how to complete the FFQ by using cups of standard portion sizes to help estimate their intake of each item. Responses were converted into average grams of daily fat intake and were sub-divided into type of fat consumed based on USDA database values. We previously demonstrated consistent results using this FFQ in our population and a 3-day food record in a second population in a study examining dietary fat-gene interaction on biomarkers of lipid metabolism and in a separate study examining the relationship between a genetic variant and sugar intake (Eny et al., 2008; Fontaine-Bisson et al., 2009). Height, weight, and waist circumference were measured and body mass index (BMI) was calculated (kg/m²). Usual physical activity over the past month was measured with a questionnaire and expressed as metabolic equivalent (MET) hours per week, which represents both leisure and occupational activity, not including sedentary hours of sleeping or sitting.

6.3.1.3 Laboratory Analysis and Genotyping

Venous blood samples were drawn after an overnight fast. Plasma FFA, triglycerides, total cholesterol, and HDL cholesterol were measured by Life Laboratories (Toronto, Canada). The Friedewald equation was used to calculate LDL cholesterol. Tag SNPs in the CD36 gene were selected using the Haploview Tagger software under the following parameters (Gene boundaries: Position 80,069,459 to 80,141,668 ± 10,000 kilobase pairs; minor allele frequency (MAF) ≥ 10 %; r2 ≥ 0.8). Eight SNPs were identified and genotyped using the Sequenom MassArray Analyzer System (Sequenom, San Diego, CA).

6.3.1.4 Statistical Analysis

Statistical analyses were performed using SAS Statistical Analysis Software version 9.1 (SAS Institute, Cary, NC). Variables that were not normally distributed were log transformed. General linear models were used to examine differences across genotypes. We used analysis of covariance to adjust for potential confounders such as sex, BMI, and physical activity.

Based on our findings with the individual polymorphisms, the two genotypes that were significantly associated with fat intake were grouped and the extreme genotypes compared to determine the effect of inheriting both genetic variations: rs1984112 (A>G)/rs1761667 (G>A): AA/AA vs GG/GG. P-values from non-parametric tests conducted for skewed variables that were log transformed are given in the tables. However, the mean ± standard error of the mean (SEM) are presented from the untransformed data. The α error was set at 0.05, and reported p- values are 2-sided.

6.3.2 Study 2: CD36 and Oleic acid taste

6.3.2.1 Population

Based on our analysis showing the combined rs1761667 and rs1984112 genotype (AA/AA vs GG/GG) was associated with the largest differences in energy intake from monounsaturated fat (MUFA), polyunsaturated fat (PUFA) and total fat, we recruited 9 AA/AA and 12 GG/GG individuals to take part in a sensory test (Table 4).

6.3.2.2 Study Protocol

Subjects were invited to the study center for a single visit at 10:00 or 11:30 AM following at least a two hour fast. Subjects were also asked to refrain from drinking coffee or eating strong tasting foods the morning of the test.

6.3.2.3 Threshold and Suprathreshold testing

Detection thresholds for oleic acid were assessed using a 3 Alternative Forced Choice (AFC) up down method. Participants were asked to rinse their mouth with distilled water prior to the start of the test and between trials. A 10 mL taste solution or control was dispensed into covered plastic cups labeled with 3-digit codes for blinding. During each trial the subject was presented with one cup containing a taste solution and 2 containing control (vehicle) only. The order of the presentation of cups was randomized. Individuals were instructed to sample all three solutions and identify the cup that was different from the other two. Participants were required to always make a selection and were informed that guessing was acceptable if the solutions seemed indistinguishable.

100

The concentration of the subject’s first taste solution exposure was 0.05 % w/w oleic acid during the initial week of testing and for all subsequent weeks the initial solution concentration was selected based on the previous test’s average threshold for a given gender. Subsequent increases or decreases in taste solution concentration were determined by the participant’s choices. If an individual provided a correct response, the stimulus concentration was lowered; an incorrect response resulted in an increase in concentration. The taste solution concentration at which the sample sequence changed from decreasing to increasing or increasing to decreasing was considered to be a reversal. Five reversals were allowed to occur and an individual’s threshold was calculated as the geometric mean of the final 4. The 5 highest concentrations prepared were used to examine supra-threshold sensitivity using a general limited magnitude scale (gLMS) (Bartoshuk et al., 2004). Solutions were presented in random order. Subjects were given a glass of distilled water and were asked to rinse between solutions.

Prior to beginning suprathreshold testing, subjects were asked to complete a scaling exercise, to confirm their ability to rank stimuli of different intensities. Subjects were asked to order the following sensations in order of decreasing intensity: 1) conversation, whisper, loudest sound experience 2) well-lit room, a dimly lit room, staring at the sun. Subjects who could not complete these tasks correctly (n=0) were excluded from the test. Additionally, subjects were asked to assign a rating to the “loudest sound ever heard” and “brightest light ever seen” on the gLMS scale to allow for data normalization (Bartoshuk et al., 2004).

6.3.2.4 Test Solutions and Study Environment

Test solutions were 0.25 log steps apart in concentration, ranging from 0.00028 to 5 % w/w. Oleic acid was suspended via homogenization in a vehicle of 0.01 % w/w ethylenediaminetetra-acetic acid (EDTA), 5 % w/w mineral oil, 5 % w/w gum acacia, and 89.9% distilled water to control for textural and lubricity cues as previously described (Mattes, 2009b). Control solutions were vehicle only. All solutions were prepared less than 24 hours before testing and stored at 4° C under nitrogen gas. Testing occurred under red light with participants wearing red tinted glasses to control for visual effects. Nose plugs were utilized throughout the test to control for olfaction effects.

101

6.3.2.5 Statistical Analysis

Block randomization to determine suprathreshold solution order was completed using random allocation software. Taste thresholds were calculated by computing the geometric mean of the concentrations at which the final four reversal points in the staircase procedure occurred. Individual ratings of the intensity of suprathreshold solutions were plotted and the incremental area under the taste sensitivity curve (iAUC), for each individual, was computed using GraphPad Prism (Version 5; GraphPad Software Inc, La Jolla, CA).

The effect of genotype on both threshold and suprathreshold taste was assessed using Statistical Analysis Software (version 9.2; SAS Institute Inc, Cary, NC). General linear models were used to examine differences across genotypes. Thresholds were adjusted for age and sex; iAUC and intensity ratings at each concentration point were adjusted for age, sex and individual’s ratings of the “loudest sound ever heard” and “brightest light ever seen”. Non- normally distributed variables were log-transformed for analysis, and their anti-logs are reported.

6.4 Results

Table 6.2 shows the association between CD36 tag SNPs, energy intake and the percent of energy consumed from fat (PUFA, MUFA and saturated fat), carbohydrates and protein. Two CD36 SNPs were significantly associated with the % of energy consumed from fat. The rs1761667(G>A) SNP was associated with fat intake where individuals homozygous for the A allele consumed less (mean ± SE) total (AA:26.7 ± 0.5, AG:28.0 ± 0.3, GG: 28.4 ± 0.5; p = 0.01) , PUFA (AA:5.5 ± 0.1, AG:5.5 ± 0.1, GG: 5.9 ± 0.1; p = 0.04) and MUFA (AA:11.6 ± 0.3, AG:12.5 ± 0.2, GG: 12.5 ± 0.3; p = 0.02) as a percent of total energy. The rs1984112(A>G) SNP was also associated with fat intake where individuals homozygous for the A allele consumed less (mean ± SE) total (AA:27.0 ± 0.4, AG:28.1 ± 0.4, GG: 28.7 ± 0.6; p = 0.02) , PUFA (AA:5.4 ± 0.1, AG:5.7 ± 0.1, GG: 6.0 ± 0.2; p = 0.03) and MUFA (AA:11.8 ± 0.2, AG:12.4 ± 0.2, GG: 12.7 ± 0.4; p = 0.04) as a percent of total energy. In the combined genotype that resulted from the combination of the rs1761667 and rs1984112 SNPs the magnitude of the difference in energy intake from fat between the extreme genotypes was greater than within either SNP individually (Table 3). Compared to GG/GG individuals, those who were homozygous for the A allele for both SNPs consumed significantly less total (AA/AA: 26.8 ± 0.4 vs GG/GG: 29.0 ± 0.6; p = 0.004), PUFA (AA/AA: 5.5 ± 0.1 vs

102

GG/GG: 6.0 ± 0.2; p = 0.03) and MUFA fat (AA/AA:11.6 ± 0.3 vs GG/GG: 12.9 ± 0.4; p = 0.004) as a percent of total energy.

103

104

Table 6.3: Dietary intake by combined rs1761667 and rs1984112 genotype1

AA/AA GG/GG p value N (% Female) 185 (61) 95 (74) Fat (% energy) 26.8 ± 0.4 29.0 ± 0.6 0.004 MUFA (% energy) 11.6 ± 0.3 12.9 ± 0.4 0.004 PUFA % energy) 5.5 ± 0.1 6.0 ± 0.2 0.033 SFA (% energy) 9.6 ± 0.2 10.1 ± 0.3 0.171 CHO (% energy) 52.9 ± 0.6 51.1 ± 0.8 0.058 Protein (% energy) 16.8 ± 0.2 16.6 ± 0.3 0.610 Energy (kcal/day) 2141 ± 50 2111 ± 71 0.723

1 Values shown are mean ± SE. A general linear model adjusted for sex, BMI and physical activity was used to test for differences in nutrient intake between genotypes. p < 0.05 was considered significant.

105

The combination genotype was not associated with threshold taste sensitivity (AA/AA: 0.8 ± 0.3 % w/w vs GG/GG: 1.1 ± 0.4 % w/w; p = 0.56). Assessment of suprathreshold taste responses revealed that AA/AA individuals perceived oleic acid solutions as less strong at all concentrations tested resulting in a significantly lower iAUC than GG/GG individuals (AA/AA: 80.9 ± 13.9 vs GG/GG: 208.0 ± 48.9; p=0.001).

106

Table 6.4: Oleic acid taste thresholds and suprathreshold taste sensitivity by combined rs1761667 and rs1984112 genotype1 AA/AA GG/GG p value n (females) 12 (8) 9 (2) Age (years) 22.7 ± 0.9 22.9 ± 0.7 0.855 BMI (kg/m²) 23.6 ± 1.4 20.7 ± 0.7 0.111 Oleic Acid Taste Threshold % w/w 0.8 ± 0.3 1.1 ± 0.4 0.561 Oleic Acid Supra-threshold intensity ratings 0.001 (iAUC) 80.9 ± 13.9 208 ± 48.9 0.50 % w/w (mm) 28.1 ± 7.3 51.8 ± 13.1 0.004 0.89 % w/w (mm) 18.8 ± 2.5 52.1 ± 12.9 <0.001 1.58 % w/w (mm) 12 ± 1.8 37.6 ± 10.6 0.004 2.80 % w/w (mm) 21.8 ± 5.4 34.7 ± 6.1 < 0.001 5.00 % w/w (mm) 17.1 ± 2.7 42.8 ± 10.3 0.041

1 Values shown are mean ± SE. A general linear model adjusted for age and sex was used to test for differences in taste thresholds between genotypes. For suprathreshold taste, a general linear model was used adjusted for age, sex and participant ratings of the strongest light and sound sensations. p < 0.05 was considered significant. Abbreviations: % w/w = % weight/weight, mm = millimeters

107

6.5 Discussion

This study revealed that genetic variation in CD36 is associated with differences in habitual fat consumption in humans. We found that individuals homozygous for the A alleles of both the rs1984112 and rs1761667 SNPs consumed less total fat, PUFA and MUFA, as a percent of energy intake, than GG/GG individuals. Furthermore, we demonstrated that this combined genotype is also associated with differences in rated taste intensity of oleic acid, such that those providing lower intensity ratings consumed less fat. Studies in mice show that deletion of the CD36 gene specifically decreases the ability of mice to differentiate between diets with and without FFAs while wildtype littermates prefer high fat meals (Laugerette et al., 2005). In the present study, we found that among certain genotypes, a greater proportion of energy from fat was consumed at the expense of energy from carbohydrate; however, overall energy intake was not different. The alleles associated with lower fat intake may cause a decreased function of CD36 that results in lower fat intensity sensation, similar to that seen in CD36-null mice (Love-Gregory et al., 2011). This could result in reduced acuity in fat sensing, which translates into reduced fat intake. This hypothesis is supported by our finding that those individuals who were found to consume less fat were also less sensitive to oleic acid taste solutions. Previous work by Stewart et al has shown that lower FA detection thresholds are associated with decreased consumption of fats; indicating that sensitivity at the threshold level is negatively correlated with intake (Stewart et al., 2010). This is opposite to the association between suprathreshold taste and intake observed in the present study. However, though fat intake was lower among those classified as hypersensitive to oleic acid in the Stewart trial, those individuals also consumed less energy and carbohydrates (Stewart et al., 2010). Fat, as a percent of energy intake, was not significantly different between hyper- and hyposensitive individuals. Further, hypersensitive individuals in the study had a significantly lower BMI, which has been associated with both higher FA taste sensitivity and lower caloric consumption (Fricker et al., 1989; Running et al., 2013). As the study did not account for potential effect modifiers andf confounders like caloric intake and BMI in its analysis it is difficult draw conclusions on the relationship between FA taste and intake from these findings (Stewart et al., 2010). Interestingly a second study by the same group found that, among lean individuals, exposure to a low fat/lower calorie diet resulted in increased sensitivity to oleic acid at the threshold level and a high fat/higher calorie diet diminished sensitivity (Stewart and Keast, 2012). This provides

108 strong evidence that dietary exposure modulates fat taste and indicates that diet may be responsible for the differences in oleic acid sensitivity observed in the initial trial (Stewart et al., 2010). CD36 genotypes were most strongly associated with MUFA and PUFA intake, suggesting that genetic variation in CD36 may influence the ability to sense these types of dietary fats. This specificity is consistent with the preferential affinity of CD36 for unsaturated FFAs (Abumrad et al., 1993; Baillie et al., 1996; Gilbertson et al., 1997). However, consumption of saturated fatty acids (SFA) followed a similar trend suggesting its intake may also be associated with CD36 genotype. This could be because a portion of the SFA may be coming from longer chain fatty acids such as stearic and palmitic acid. Previous work has shown that palmitic acid stimulates increases in intracellular calcium in CD36 positive taste bud cells similar to unsaturated LCFA (Gaillard et al., 2008). A number of studies have examined the relationship between rs1761667 genotype and fat ingestive endpoints and behaviors. Differences have been observed between studies in Caucasians and mixed populations and those exclusively with African Americans. In the former, the GG is considered the “at risk” genotype. Ma et al reported that men with the GG genotype had higher FFAs and triglyceride levels while Madden et al showed that these individuals were less likely to benefit from dietary supplementation with fish oil (Ma et al., 2004; Madden et al., 2008). These findings are in line with those of the present study where GG individuals consumed more energy from fat than AA or AG individuals. However, in an African American population, Keller et al showed that individuals homozygous for the A allele of the rs1761667 SNP perceived salad dressings with varying amounts of fat as consistently creamier than AG/GG individuals and that these individuals have a greater acceptance of oils and fats added to foods. This would indicate that AA individuals might be at a greater risk of over consumption of fats; a finding not supported by the results of the present study. A potential explanation for this disparity, outlined by Keller et al, is that the rs1761667 SNP may be in linkage disequilibrium with another SNP in African Americans, but not Caucasians, and this drives the observed effects (Keller et al., 2012). Interestingly, in African Americans, the A is the minor allele while in Caucasians the G is the minor allele, which supports the idea of different selective pressures driving allele prevalence across the two ethnic groups (Keller et al., 2012) One previous study, has examined the role of CD36 genotype on fat taste and intake. Pepino et al found that individuals homozygous for the G allele of the rs1761667 SNP had an 8-

109 fold lower oleic acid detection threshold than AA individuals, but no association was observed with fat intake (Pepino et al., 2012). The intake analysis was likely limited by sample size with only 6 AA, 7 AG and 8 GG individuals examined. Moreover, the standard errors reported for fat intake, as a percent of energy, were 5 to 8 times greater than those reported in the present study (3 % of energy intake vs 0.4 to 0.6 %) and were larger than the magnitude of the effect observed in the present study. The differences observed in the relationship between genotype and threshold taste in the two studies may be due to a lack of control for lubricity perception in the study by Pepino et al (Pepino et al., 2012). In the present study, vehicle solutions contained 5% w/w mineral oil while those in the previous study did not. The purpose of the addition of mineral oil is to normalize the lubricity cues between control (vehicle only) and oleic acid (vehicle + oleic acid) samples. Without controlling for this factor it is possible that the findings observed by Pepino et al were driven by differences in lubricity perception rather than taste, as has been previously suggested (Running et al., 2013). Given that CD36 has not been associated with lubricity perception it is difficult to explain the differences across genotype groups. However, in addition to genotype, BMI (kg/m2) differed between the groups (AA: 34.9, AG: 38.3, GG: 41.5) and was not adjusted for in the analysis within that study. Previous work has shown that individuals who are morbidly obese prefer stimuli with a higher fat composition and this has been attributed to differences in their perception of the mouth feel and viscosity of these stimuli (Drewnowski and Almiron-Roig, 2010; Drewnowski et al., 1985). Consequently, it is possible that the effect observed by Pepino et al was due to a lubricity/BMI relationship rather than one between genotype and taste.

In summary, the present study is the first to show that CD36 genotype is associated with differences in habitual fat intake. Our findings indicate that the rs1984112 and rs1761667 SNPs in CD36 are associated with fat intake where individuals homozygous for the A allele of both SNPs consume less fat. The identified genotypes were found to have the greatest effect on the consumption of MUFA and PUFA, supporting the proposed role of CD36 as a LCFA sensor. Those individuals who consumed the least fat also had lower sensitivity to oleic acid, potentially providing an explanation of how genetic variation in CD36 affects fat intake.

110

Chapter 7 Summary, Limitations, Future Directions and Implications

7.1 Summary

The overall aim of this thesis was to determine whether variation in candidate genes associated with salt (SCNN1A, SCNN1B, SCNN1D, SCNN1G and TRPV1), sweet (TAS1R2), fat (CD36), and bitter (TAS2R19) taste were associated with taste perception, food preference and intake. Objective 1: Examine how variation in putative salt taste receptors ENaC (SCNN1A, SCNN1B, SCNN1D, SCNN1G) and TRPV1 (TRPV1) affects salt taste thresholds, supra-threshold taste sensitivity and sodium intake. Results: None of the SNPs in the SCNN1A and SCNN1G genes were significantly associated with either threshold taste or supra threshold taste sensitivity. In the SCNN1B gene 2 SNPs located in intronic regions, rs239345 (A>T) and rs3785368 (C>T), were associated with suprathreshold taste. Those homozygous for the A allele of the rs239345 (A>T) polymorphism and the T allele of the rs3785368 (C>T) polymorphism perceived salt solutions less intensely than carriers of the T or C alleles, respectively. In TRPV1 gene the rs8065080 (C>T, Val585Ile) polymorphism was associated with suprathreshold taste where carriers of the T allele were significantly more sensitive to salt solutions than those who were homozygous for the C allele. Of the three SNPs found to affect taste, none were associated with sodium intake.

Objective 2: Examine how genetic variation in sweet taste receptor TAS1R2 (TAS1R2) affects sucrose taste thresholds, supra-threshold taste sensitivity and intake of sugars. Results: The TAS1R2 rs12033832 (G>A) SNP was associated with both sucrose taste and sugar intake. There was a significant genotype X BMI interaction for the SNP on both suprathreshold taste and sugar intake. As a result analyses were stratified by BMI Status (≥ 25 or < 25). For individuals with a BMI ≥ 25, carriers of the G allele were less sensitive to sucrose stimuli, with lower suprathreshold sensitivity ratings and higher taste thresholds, and consumed more sugars. Among individuals with a BMI < 25 the effect was reversed where carriers of the G allele were more sensitive to sucrose stimuli, with lower taste thresholds, and consumed less sugars.

111

Objective 3: Examine how variation in bitter taste receptor TAS2R19 (TAS2R19) affects naringin sensitivity and both grapefruit/grapefruit juice preference and intake. Results: Two SNPs in the TAS2R19 gene, rs10772420 (A>G) and rs4763235 (C>G), were found to be associated with both differences in sensitivity to the bitter tastant naringin and grapefruit/grapefruit juice preference. Additionally, the rs10772420 SNP was also associated with grapefruit and grapefruit juice intake. Among those who were homozygous for the A allele of the rs10772420 SNP, the odds of perceiving naringin as “high intensity” were greater, the odds of disliking grapefruit and grapefruit and grapefruit juice were greater, and these individuals were also more likely to abstain from consuming grapefruit or grapefruit juice. Among those who were homozygous for the C allele of the rs4763235 SNP, the odds of perceiving naringin as “high intensity” were greater and the odds of disliking grapefruit and grapefruit and grapefruit juice were greater. However, this did not result in a change in grapefruit intake.

Objective 4: Examine how genetic variation in putative fat taste receptor CD36 (CD36) affects habitual fat intake and taste perception.

Results: Two CD36 SNPs, rs1761667 (G>A) SNP and rs1984112 (A>G), were significantly associated with the % of energy consumed from fat. Individuals homozygous for the A allele of rs1761667 consumed less total, PUFA and MUFA as a percent of total energy. Individuals homozygous for the A allele of rs1984112 also consumed less total, PUFA and MUFA as a percent of total energy. Comparing individuals with both alleles associated with the lowest (AA/AA) and highest (GG/GG) fat consumption resulted in an increase in magnitude of the difference in energy intake from fat. AA/AA individuals also perceived the taste of oleic acid solutions as less strong than GG/GG individuals.

112

7.2 The relationship between variation in putative salt taste receptors, salt taste and sodium intake

Overall, the results of the present thesis show that variation in putative salt taste receptors ENaC and TRPV1 is associated with differences in salt taste perception but not sodium intake (Chapter 3).

Few studies to date have examined the effect of genetic variation on salt taste and sodium intake. Studies involving mice have shown that heritability plays a role in the inter-strain differences in sodium intake and preference (Bachmanov et al., 2002; Tordoff et al., 2007). More recently, Shigemura et al. demonstrated that variation in the gene for the alpha subunit of ENaC may explain the inter-strain variation seen in salt taste perception within mice (Shigemura et al., 2008). Like rodents, at least part of human salt taste perception can be attributed to ENaC. In humans, amiloride, an ENaC blocker, suppresses salt taste by about 20% (Feldman et al., 2003; Smith and Ossebaard, 1995). However, no published studies have directly tested the effect genetic variants in ENaC have on salt taste; outside of the work presented here. Since its cloning in 1997, the TRPV1 channel has been studied extensively for its potential role in pain perception (Szallasi et al., 2007). Evidence of its role in salt taste is drawn primarily from rodent studies and, outside of the present thesis, no studies have examined the role of genetic variation within it on salt taste (Lyall et al., 2004). No studies to date have examined the role of genetic variants in putative salt taste receptors on sodium intake though a recent study has shown that sodium intake is in part determined by genetic predisposition (Kho et al., 2013).

The work presented in this thesis is the first to examine the association between variations in specific genes, salt taste perception and sodium intake in humans. Our findings show that polymorphisms in the genes that code for the ENaC ß subunit (SCNN1B) and the TRPV1 non-specific cation channel (TRPV1) modify an individual’s suprathreshold salt taste perception but not sodium intake. Interestingly, previous studies examining the TRPV1 rs8065080 (C>T) SNP have found that this mutation results in modification in the performance of this receptor where the CC genotype is associated with reduced sensitivity (Kim et al., 2004a; Lö tsch J, 2009; Valdes et al., 2011). These findings are consistent with results from the present study where those in the CC group were less sensitive to salt stimuli than carriers of the T allele.

113

Limitations

There are number of limitations to the work presented in Chapter 3 of the present thesis examining genetic variation, salt taste and sodium intake. Our coverage of polymorphisms within the ENaC associated genes SCNN1A and SCNN1D was very limited. Further, our study population was not large enough to examine rare variants and all SNPs with a minor allele frequency less than 15 % were excluded from our analysis; these could have had an effect on protein function and taste perception. One of our taste phenotypes, threshold taste, is known to fluctuate over time (Galindo-Cuspinera et al., 2009; Grzegorczyk et al., 1979). Such variability would increase the noise in our threshold measure and could have masked some potential associations. The large number of SNPs examined in each gene increases the likelihood of a false positive association. Lastly, previous work has shown that correlation between intake assessed by an FFQ and actual sodium consumption is quite low, r = 0.43 for men and 0.11 for women (McKeown et al., 2001). Subsequently our measure of sodium intake may not have been appropriate to predict habitual intake and increased variability in this data may have masked some potential associations.

Future Directions

Preference tests of NaCl solutions with amiloride versus water revealed that Trpv1 knockout mice actually preferred the salty tasting solution with amiloride over water; despite the blockage of the ENaC channel (Ruiz et al., 2006; Treesukosol et al., 2007). This indicates that there are likely other channels involved in sodium taste in addition to TRPV1 and ENaCs. Future studies should utilize genetic techniques like Genome Wide Association Scans to identify other variants, and potentially receptors that may be associated with salt taste

As stated in the limitations section, the work in the present section did not provide good coverage of genetic variation in the SCNN1A and SCNN1D genes, both of which may be important in human salt taste perception. Further, due to a limited sample size of the sensory study, we were unable to examine SNPs with low MAFs. Studies in a larger group of individuals that use sequencing to identify all of the variants in genes of interest are required to fully understand the effect of genetic variation in putative salt taste receptors on both taste and intake.

114

Given the role of TRPV1 in nociception, and the wide range of stimuli that can lead to its activation, it is possible that the sensation individuals perceive mediated by this channel in response to a salt stimulus are not unique to salt, but rather a subclass of general noxious responses (Stewart et al., 1997). Future studies should aim to examine the attributes of salt taste at these suprathreshold levels and characterize whether individuals perceive the taste stimuli as aversive or pleasurable, salty or painful. If the TRPV1 channel’s role in taste is to prevent the consumption of highly concentrated cation solutions it may have little effect on preference for salt or sodium consumption.

Though the FFQ, utilized here, is not ideal for the measurement of sodium consumption, measurements exist that do adequately capture habitual sodium intake. Urinary sodium excretion is thought to strongly correlate with actual consumption (McKeown et al., 2001). Subsequently, future studies should utilize better measures of sodium intake to understand the potential effect of variation in putative salt taste receptors on consumption.

7.3 The relationship between variation in the TAS1R2 sweet taste receptor, sucrose taste and sugar intake

Overall, the results of the present thesis show that variation in the TAS1R2 sweet taste receptor is associated with differences in sucrose taste perception and intake of sugars. Further, this relationship is modified by BMI (Chapter 4).

Two studies to date have looked at the effect of TAS1R2 variants on sweet taste and sugar intake. Fushan et al. used DNA sequencing to look at the effect of single nucleotide polymorphisms (SNPs) within TAS1R2 on individual’s ability to discriminate between sucrose solutions of different concentrations (Fushan et al., 2009). None of the 34 variants examined were found to affect sweet taste. However, the study did not account for body mass index (BMI) and the taste phenotyping method utilized may not have been optimal (Ettinger et al., 2012; Galindo-Cuspinera et al., 2009). Eny et al examined the effect of two missense SNPs, Serine9Cysteine (rs9701796) and Isoleucine191Valine (rs35874116), in the TAS1R2 gene on sugar intake (Eny et al., 2010). Grouping subjects by BMI, they found that among those with a BMI ≥ 25, Valine carriers consumed significantly lower amounts of sugars than those homozygous for the Isoleucine group. However, this study only examined 2 TAS1R2 variants

115 and did not asses their effect on sweet taste. Subsequently, additional studies looking at the effect of TAS1R2 variation on taste and intake are required.

The work presented in this thesis was the first to investigate the effects of variation in the TAS1R2 gene on both sucrose taste and sugar intake within the same population. Among those with a BMI ≥ 25, individuals homozygous for the A allele of the rs12033832 SNP had lower sucrose detection thresholds, higher suprathreshold taste sensitivity ratings and a lower intake of total sugars. Among those with a BMI < 25, individuals homozygous for the A allele had higher detection thresholds and a higher intake of sugars. The reversal of the effect of possessing the AA genotype in normal and overweight/obese individuals may be the result differences in leptin sensitivity between these groups. Leptin is thought to interfere with signal transduction during sweet taste perception by increasing outward K+ currents in taste receptor cells; obesity may cause leptin resistance (Kawai et al., 2000; Zheng et al., 2009). Subsequently, among normal weight individuals, leptin may modulate the effect of the AA genotype resulting in lowered sensitivity among these individuals. In overweight and obese individuals, who may be less sensitive or resistant to leptin, the effect of the AA genotype may be unmasked, allowing increased sucrose sensitivity within this group.

Limitations

There were a number of limitations in the work presented in Chapter 4 of the present thesis. Like the salt taste study, the method of SNP selection eliminated any SNPs that occurred at a low frequency within the population. It is possible that that a SNP that is less prevalent may affect the function or expression of TAS1R2 and modify taste and intake more than those identified here. Additionally, the limited sample size of the sensory study resulted in very small number of individuals in the minor allele groups. The size of groups were further diminished after stratification by BMI status, with only 2 people individuals, with a BMI ≥ 25, who were homozygous for the AA allele of the rs12033832 SNP. This remained an issue in the dietary analysis where there were only 17 individuals in the same minor allele group.

Future Directions

Recent work has shown that variation in other genes involved in sweet taste perception may be associated with differences in taste (Fushan et al., 2010; Fushan et al., 2009). Two SNPs

116 in the TAS1R3 gene, rs307355 and rs35744813, were found to significantly affected measures of sweet taste perception explaining about 16% of the observed variation between individuals (Fushan et al., 2009). Further 11 SNPs in the GNAT3, which codes for the taste specific G protein gustducin involved in the sweet taste signalling cascade, where also associated with differences in sucrose perception (Fushan et al., 2010). Interestingly, neither study adjusted for, or stratified individuals by, BMI, which was shown to be a significant modifier of the relationship between genotype and taste in the present thesis. Subsequently, studies examining the effect of variation in these genes on sweet taste, accounting for BMI status, are required.

Additionally, similar to the salt taste trial, the limited sample size of the sensory study excluded the possibility of examining SNPs with low MAFs. Studies in a larger group of individuals are required to fully understand the effect of these potentially important, but less prevalent, SNPs.

7.4 The relationship between variation in the TAS2R19 bitter taste receptor, naringin taste, and both grapefruit preference and intake

Overall, the results of the present thesis show that variation in the TAS2R19 gene is associated with differences in sensitivity to the bitter compound naringin found primarily in grapefruit. Further those individuals who were more sensitive to naringin were more likely to dislike and not consume both grapefruit and grapefruit juice (Chapter 5).

Bitter foods are thought to have a number of positive health benefits (Shiffman et al., 2005). However, individuals who are highly sensitive to bitter compounds may be at risk of under consuming such foods. To date, few studies have examined how variations in bitter taste receptors modify sensitivity to bitter tastants and potentially preference and intake of foods. Recently, Hayes et al. showed that the rs10772420 single nucleotide polymorphism (SNP) in the TAS2R19 gene, which results in an Arg299Cys amino acid substitution, is associated with perceived bitterness, sweetness and liking of grapefruit juice (Hayes et al., 2011). Cys/Cys homozygous individuals rated the juice as more bitter and less sweet/likable than Arg/Arg individuals. This indicates that grapefruit juice liking may be a function of an individual’s sensitivity to bitter compounds in grapefruit juice. As the primary bitter tastant in grapefruit, the compound naringin may be a ligand for the TAS2R19 receptor and naringin taste sensitivity may be modified by TAS2R19 genotype.

117

The present study identified two TAS2R19 SNPs, rs10772420 (A>G) and rs4763235 (C>G), were significantly associated with naringin sensitivity. For rs10772420, the odds ratios (OR (95% CI)) of perceiving naringin as “high intensity” were significantly higher among those who were homozygous for the A allele. These individuals also had greater odds of disliking grapefruit/grapefruit juice and were more likely not to consume either. A similar effect was seen in those who were homozygous for the C allele of the rs4763235 SNP on naringin sensitivity and disliking grapefruit/grapefruit juice preference, but not intake.

Limitations

There were a number of limitations to the work presented in chapter 5 of the present thesis. The scales used to assess naringin taste sensitivity and grapefruit and grapefruit juice preference have been criticized because they do not exhibit ratio properties, may reduce variability in responses due to ceiling effects and may not be uniformly understood due to the use of specific descriptors (Hein, 2008). Subsequently the magnitude in the differences in sensitivity or liking across the genotypes may be greater than reported here. The use of a FFQ to measure grapefruit and grapefruit juice consumption over the past month does not take into account seasonal variation in food availability and intake. Although we adjusted for seasonality in the analysis, this may, in part, explain the discrepancy in dislike and intake observed. Lastly, the coverage of polymorphisms in the TAS2R19 gene was not exhaustive. The potential for other, possibly more influential, variants in this gene exist.

Future Directions

The rs1376251 SNP in the TAS2R50 gene has been associated with risk of myocardial infarction in two studies (Shiffman et al., 2005; Shiffman et al., 2008). Given the SNP’s location in a bitter taste receptor it has been hypothesized that this SNP may influence health outcomes by modifying an individual’s bitter taste sensitivity and in turn dietary choices (Shiffman et al., 2005). Interestingly, rs1376251 in TAS2R50 is in LD with the rs4763235 SNP in TAS2R19, identified here as being associated with naringin taste. Additionally the rs10772420 SNP is also in LD with a TAS2R50 variant. Subsequently, it is difficult to determine which gene is driving the effect on naringin sensitivity and future studies should focus on clarifying this.

The relationship between TAS2R19 and naringin was previously examined by Meyerhof

118 et al. to determine if naringin is indeed a ligand of TAS2R19; no association was identified (Meyerhof et al., 2010). However, the authors of that study acknowledged that in transfected cells meant to express this gene, only marginal expression was observed. Future studies should conduct similar analysis on multiple variants of the gene, based on the genotypes identified here, and aim to achieve better expression of the gene in cell lines, to determine if naringin is a ligand of TAS2R19. A similar approach should also be taken to examine if naringin is a ligand of TAS2R50.

7.5 The relationship between variation in the putative fat taste receptor CD36, fat taste and fat intake

Overall, the results of the present thesis show that variation in the CD36 gene is associated with differences in habitual fat consumption. Further, among those individuals who consumed the most and least fat, there was a significant difference in their sensitivity to the FFA oleic acid (Chapter 6).

Direct evidence for a role of CD36 in mediating fat taste comes from studies in CD36- null mice. These mice have a diminished ability to select a diet containing FFAs over a texturally similar control diet, whereas their wildtype littermates are more likely to choose a diet containing FFAs (Laugerette et al., 2005; Sclafani et al., 2007a). This suggests that CD36 may contribute to oral FFA sensing and that this is associated with a predilection for fat. To date, few studies have investigated the effect of CD36 genotypes on human ingestive behaviors or fat taste. Keller et al found that, among African Americans, the rs1761667 single nucleotide polymorphism (SNP) was associated with differences in the perceived creaminess of salad dressing and liking of certain types of fats (Keller et al., 2012). Pepino et al found that the rs1761667 SNP was associated with an individual’s ability to detect oleic acid and triolein but not fat intake (Pepino et al., 2012). Lastly, a previous study from our lab (unpublished) found that rs1984112 and rs1761667 both significantly affected the percent energy consumed from (Toguri, 2008). To date, no studies have identified CD36 variants that effect both fat taste perception and intake.

The results in the present thesis are the first to show that genetic variation in CD36 is associated with both differences in habitual fat consumption and fat taste in humans. We found that individuals homozygous for the A alleles of both the rs1984112 and rs1761667 SNPs consumed less total, PUFA and MUFA, as a percent of energy intake, than GG/GG individuals.

119

Further this combined genotype is also associated with differences in oleic acid taste sensitivity, where those who consume less fat are less sensitive to oleic acid.

Limitations

There were a few of limitations to the work presented in chapter 6 of the present thesis. The sensory analysis study was limited in size with just 9 individuals in the GG/GG genotype group. Further, even at the lowest suprathreshold concentrations examined, there was a significant difference in sensitivity between the genotype groups. When establishing a sensitivity curve the aim is to examine a range of concentrations wide enough such that it is possible to establish at what concentrations differences are not detectable; this was not possible in the present study. Though the findings here provide evidence of an association between CD36 variation, fat taste, and fat intake the cross-sectional study design precludes drawing any conclusions about causality based on the observed associations. The study population examined consisted only of individuals of Caucasian descent. As previous studies have shown, the CD36 risk genotypes may be different among different ethnicities limiting the generalizability of the findings reported here (Keller et al., 2012; Pepino et al., 2012).

Future directions

In the present thesis we found that those who were more sensitive to oleic acid consumed more fat. Previous work has shown FFAs generally taste repulsive to humans (Schiffman and Dackis, 1975). Subsequently it is unknown why increased sensitivity is associated with increased intake. However other argue that at certain levels, FFAs may be desirable (Mattes, 2009a). Studies that examine the effect of CD36 variation on individual’s preference for oleic acid solutions, in addition to taste sensitivity, may improve our understanding of the relationship between taste and intake.

As discussed in the limitations section, the sensory study was limited its size. Further, all subjects in both the intake and sensory studies were Caucasian, and the majority had a healthy BMI. Given that previous work has shown that both BMI and diet may be modifiers of fat taste, and that the effect of genotype may differ across different ethnic groups, additional studies, in larger groups of individuals, from diverse ethnic backgrounds with different levels of obesity are required (Keller et al., 2012; Pepino et al., 2012; Stewart et al., 2010; Stewart and Keast, 2012).

120

7.6 Implications

There are a number of implications resulting from the findings in this thesis. Though examined extensively in a number of animal and cell line models, the receptors responsible for taste perception are not well studied in humans. The work presented shows that variation in a number of putative taste receptors is associated with differences in taste perception, providing evidence to support the role of these receptors in humans. For example, though studies in mice have shown that the TRPV1 channel may be involved in salt taste, there is little evidence to support this finding in humans (Lyall et al., 2004). Our findings are the first to provide evidence in humans that supports the relationship between TRPV1 and salt taste.

Taste is thought to be the number one modifier of individual food selection (El-Sohemy et al., 2007). However, few studies to date have examined how genetic variation may affect taste and the potential impact this has on diet. Understanding the determinants of dietary patterns may allow us to develop screening models that predict who is at risk of a nutritionally compromised diet. For example, findings from the present thesis could help shed light on genotypes that may result in greater affinity for foods high in fats and sugars or an aversion to potentially health bitter foods like grapefruit. Identifying individuals who may be genetically predisposed to an unhealthy dietary pattern could be a first step in targeted strategies to improve these individuals’ nutritional status and health.

Lastly, some of the work here may help shed light on the mechanism by which genetic variants identified in other studies affect health outcomes. Genome wide association studies have linked the chromosomal region in which the CD36 gene is located to characteristics of the metabolic syndrome (An et al., 2005; Arya et al., 2002). Further variation in the TAS2R50 gene has been associated with risk of myocardial infarction in two studies (Shiffman et al., 2005; Shiffman et al., 2008). Our work may help explain the mechanism by which these variants affect health status; by modifying taste and preference which influences intake of foods.

121

References

Abaffy, T., K.R. Trubey, and N. Chaudhari. 2003. Adenylyl cyclase expression and modulation of cAMP in rat taste cells. Am J Physiol Cell Physiol. 284:C1420-1428.

Abdoul-Azize, S., S. Selvakumar, H. Sadou, P. Besnard, and N.A. Khan. 2013. Ca signaling in taste bud cells and spontaneous preference for fat: Unresolved roles of CD36 and GPR120. Biochimie.

Abumrad, N.A. 2005. CD36 may determine our desire for dietary fats. J Clin Invest. 115:2965- 2967.

Abumrad, N.A., M.R. el-Maghrabi, E.Z. Amri, E. Lopez, and P.A. Grimaldi. 1993. Cloning of a rat adipocyte membrane protein implicated in binding or transport of long-chain fatty acids that is induced during preadipocyte differentiation. Homology with human CD36. J Biol Chem. 268:17665-17668.

AICR, W.A. 2007. Food, nutrition, physical activity, and the prevention of cancer: A global perspective. AICR, Washington DC.

Ames, B.N., M. Profet, and L.S. Gold. 1990. Dietary pesticides (99.99% all natural). Proc Natl Acad Sci U S A. 87:7777-7781.

An, P., B.I. Freedman, C.L. Hanis, Y.D. Chen, A.B. Weder, N.J. Schork, E. Boerwinkle, M.A. Province, C.A. Hsiung, X. Wu, T. Quertermous, and D.C. Rao. 2005. Genome-wide linkage scans for fasting glucose, insulin, and insulin resistance in the National Heart, Lung, and Blood Institute Family Blood Pressure Program: evidence of linkages to chromosome 7q36 and 19q13 from meta-analysis. Diabetes. 54:909-914.

Anliker, J.A., L. Bartoshuk, A.M. Ferris, and L.D. Hooks. 1991. Children's food preferences and genetic sensitivity to the bitter taste of 6-n-propylthiouracil (PROP). Am J Clin Nutr. 54:316-320.

Arab, L. 2003. Biomarkers of fat and fatty acid intake. J Nutr. 133 Suppl 3:925S-932S.

Armesilla, A.L., and M.A. Vega. 1994. Structural organization of the gene for human CD36 glycoprotein. J Biol Chem. 269:18985-18991.

Arya, R., J. Blangero, K. Williams, L. Almasy, T.D. Dyer, R.J. Leach, P. O'Connell, M.P. Stern, and R. Duggirala. 2002. Factors of insulin resistance syndrome--related phenotypes are linked to genetic locations on chromosomes 6 and 7 in nondiabetic mexican-americans. Diabetes. 51:841-847.

Bachmanov, A.A., and G.K. Beauchamp. 2007. Taste receptor genes. Annu Rev Nutr. 27:389- 414.

Bachmanov, A.A., G.K. Beauchamp, and M.G. Tordoff. 2002. Voluntary consumption of NaCl, KCl, CaCl2, and NH4Cl solutions by 28 mouse strains. Behav Genet. 32:445-457.

122

Bachmanov, A.A., X. Li, D.R. Reed, J.D. Ohmen, S. Li, Z. Chen, M.G. Tordoff, P.J. de Jong, C. Wu, D.B. West, A. Chatterjee, D.A. Ross, and G.K. Beauchamp. 2001. Positional cloning of the mouse saccharin preference (Sac) locus. Chem Senses. 26:925-933.

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

Bartoshuk, L.M. 2000. Comparing sensory experiences across individuals: recent psychophysical advances illuminate genetic variation in taste perception. Chem Senses. 25:447-460.

Bartoshuk, L.M., V.B. Duffy, B.G. Green, H.J. Hoffman, C.W. Ko, L.A. Lucchina, L.E. Marks, D.J. Snyder, and J.M. Weiffenbach. 2004. Valid across-group comparisons with labeled scales: the gLMS versus magnitude matching. Physiol Behav. 82:109-114.

Basson, M.D., L.M. Bartoshuk, S.Z. Dichello, L. Panzini, J.M. Weiffenbach, and V.B. Duffy. 2005. Association between 6-n-propylthiouracil (PROP) bitterness and colonic neoplasms. Dig Dis Sci. 50:483-489.

Beauchamp, G. 1991. Salt preference in humans, Encyclopedia of human biology. Academic Press, New York.

Beauchamp, G.K., M. Bertino, D. Burke, and K. Engelman. 1990. Experimental sodium depletion and salt taste in normal human volunteers. Am J Clin Nutr. 51:881-889.

Beauchamp, G.K., and B.J. Cowart. 1990. Preference for High Salt Concentrations among Children. Dev Psychol. 26:539-545.

Behrens, M., S. Foerster, F. Staehler, J.D. Raguse, and W. Meyerhof. 2007. Gustatory expression pattern of the human TAS2R bitter receptor gene family reveals a heterogenous population of bitter responsive taste receptor cells. J Neurosci. 27:12630-12640.

Behrens, M., and W. Meyerhof. 2009. Mammalian bitter taste perception. Results Probl Cell Differ. 47:203-220.

Bellisle, F., and A. Drewnowski. 2007. Intense sweeteners, energy intake and the control of body weight. Eur J Clin Nutr. 61:691-700.

Berridge, K.C., and T.E. Robinson. 2003. Parsing reward. Trends Neurosci. 26:507-513.

Bertino, M., G.K. Beauchamp, and K. Engelman. 1986. Increasing dietary salt alters salt taste preference. Physiol Behav. 38:203-213.

Bingham, S.A. 2002. Biomarkers in nutritional epidemiology. Public Health Nutr. 5:821-827.

Blais, C.A., R.M. Pangborn, N.O. Borhani, M.F. Ferrell, R.J. Prineas, and B. Laing. 1986. Effect of Dietary-Sodium Restriction on Taste Responses to Sodium-Chloride - a Longitudinal- Study. American Journal of Clinical Nutrition. 44:232-243.

Blakeslee, A.F., and T.N. Salmon. 1935. Genetics of Sensory Thresholds: Individual Taste Reactions for Different Substances. Proc Natl Acad Sci U S A. 21:84-90.

123

Block, G., A.M. Hartman, C.M. Dresser, M.D. Carroll, J. Gannon, and L. Gardner. 1986. A data- based approach to diet questionnaire design and testing. Am J Epidemiol. 124:453-469.

Blundell, J.E., and J. Cooling. 1999. High-fat and low-fat (behavioural) phenotypes: biology or environment? Proc Nutr Soc. 58:773-777.

Blundell, J.E., and J.I. MacDiarmid. 1997. Fat as a risk factor for overconsumption: satiation, satiety, and patterns of eating. J Am Diet Assoc. 97:S63-69.

Blundell, J.E., R.J. Stubbs, C. Golding, F. Croden, R. Alam, S. Whybrow, J. Le Noury, and C.L. Lawton. 2005. Resistance and susceptibility to weight gain: individual variability in response to a high-fat diet. Physiol Behav. 86:614-622.

Boughter, J.D., Jr., and A.A. Bachmanov. 2007. Behavioral genetics and taste. BMC Neurosci. 8 Suppl 3:S3.

Bradić, M., J. Costa, and I. Chelo. 2011. Genotyping with Sequenom. In Molecular Methods for Evolutionary Genetics. Vol. 772. V. Orgogozo and M.V. Rockman, editors. Humana Press. 193-210.

Bradley, R.M., H. Fukami, and T. Suwabe. 2005. Neurobiology of the gustatory-salivary reflex. Chem Senses. 30 Suppl 1:i70-71.

Brandt, M.A., E.Z. Skinner, and J.A. COLEMAN. 1963. Texture profile method. Journal of Food Science. 28:404-409.

Cahill, L.E., B. Fontaine-Bisson, and A. El-Sohemy. 2009. Functional genetic variants of glutathione S-transferase protect against serum ascorbic acid deficiency. Am J Clin Nutr. 90:1411-1417.

Cartoni, C., K. Yasumatsu, T. Ohkuri, N. Shigemura, R. Yoshida, N. Godinot, J. le Coutre, Y. Ninomiya, and S. Damak. 2010. Taste preference for fatty acids is mediated by GPR40 and GPR120. J Neurosci. 30:8376-8382.

CCHS. 2004. Canadian Community Health Survey. Vol. 1-3. Health Canada, Ottawa.

Chale-Rush, A., J.R. Burgess, and R.D. Mattes. 2007. Evidence for human orosensory (taste?) sensitivity to free fatty acids. Chem Senses. 32:423-431.

Chandrashekar, J., M.A. Hoon, N.J. Ryba, and C.S. Zuker. 2006. The receptors and cells for mammalian taste. Nature. 444:288-294.

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

Chaudhari, N., and S.D. Roper. 2010. The cell biology of taste. J Cell Biol. 190:285-296.

Collaku, A., T. Rankinen, T. Rice, A.S. Leon, D.C. Rao, J.S. Skinner, J.H. Wilmore, and C. Bouchard. 2004. A genome-wide linkage scan for dietary energy and nutrient intakes: the

124

Health, Risk Factors, Exercise Training, and Genetics (HERITAGE) Family Study. Am J Clin Nutr. 79:881-886.

Considine, R.V., M.K. Sinha, M.L. Heiman, A. Kriauciunas, T.W. Stephens, M.R. Nyce, J.P. Ohannesian, C.C. Marco, L.J. McKee, T.L. Bauer, and et al. 1996. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med. 334:292-295.

Dahl, L.K. 1972. Salt and hypertension. Am J Clin Nutr. 25:231-244.

Dahl, L.K. 2005. Possible role of salt intake in the development of essential hypertension. 1960. Int J Epidemiol. 34:967-972; discussion 972-964, 975-968.

Damak, S., M. Rong, K. Yasumatsu, Z. Kokrashvili, V. Varadarajan, S. Zou, P. Jiang, Y. Ninomiya, and R.F. Margolskee. 2003. Detection of sweet and umami taste in the absence of taste receptor T1r3. Science. 301:850-853.

Degrace-Passilly, P., and P. Besnard. 2012. CD36 and taste of fat. Curr Opin Clin Nutr Metab Care. 15:107-111.

Denton, D. 1982. The hunger for salt: An anthropological, physiological and medical analysis. Springer-Verlag, New York.

DeSimone, J.A., and V. Lyall. 2006. Taste receptors in the gastrointestinal tract III. Salty and sour taste: sensing of sodium and protons by the tongue. Am J Physiol Gastrointest Liver Physiol. 291:G1005-1010.

Dinehart, M.E., J.E. Hayes, L.M. Bartoshuk, S.L. Lanier, and V.B. Duffy. 2006. Bitter taste markers explain variability in vegetable sweetness, bitterness, and intake. Physiol Behav. 87:304-313.

Ding, C., and S. Jin. 2009. High-throughput methods for SNP genotyping. Methods Mol Biol. 578:245-254.

Ding, R., J.A. Logemann, C.R. Larson, and A.W. Rademaker. 2003. The effects of taste and consistency on swallow physiology in younger and older healthy individuals: a surface electromyographic study. J Speech Lang Hear Res. 46:977-989.

Dotson, C.D., S.D. Roper, and A.C. Spector. 2005. PLCbeta2-independent behavioral avoidance of prototypical bitter-tasting ligands. Chem Senses. 30:593-600.

Drayna, D. 2005. Human taste genetics. Annu Rev Genomics Hum Genet. 6:217-235.

Drewnowski, A. 1997a. Taste preferences and food intake. Annu Rev Nutr. 17:237-253.

Drewnowski, A. 1997b. Taste preferences and food intake. Ann Rev Nutr. 17:237-253.

Drewnowski, A. 1997c. Why do we like fat? J Am Diet Assoc. 97:S58-62.

125

Drewnowski, A. 2000. Sensory control of energy density at different life stages. Proc Nutr Soc. 59:239-244.

Drewnowski, A. 2001. The science and complexity of bitter taste. Nutr Rev. 59:163-169.

Drewnowski, A., and E. Almiron-Roig. 2010. Human Perceptions and Preferences for Fat-Rich Foods.

Drewnowski, A., J.D. Brunzell, K. Sande, P.H. Iverius, and M.R. Greenwood. 1985. Sweet tooth reconsidered: taste responsiveness in human obesity. Physiol Behav. 35:617-622.

Drewnowski, A., and C. Gomez-Carneros. 2000. Bitter taste, phytonutrients, and the consumer: a review. Am J Clin Nutr. 72:1424-1435.

Drewnowski, A., and C. Hann. 1999. Food preferences and reported frequencies of food consumption as predictors of current diet in young women. Am J Clin Nutr. 70:28-36.

Drewnowski, A., S.A. Henderson, and J.E. Cockroft. 2007. 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. 107:1340-1348.

Drewnowski, A., S.A. Henderson, C.S. Hann, W.A. Berg, and M.T. Ruffin. 2000. Genetic taste markers and preferences for vegetables and fruit of female breast care patients. J Am Diet Assoc. 100:191-197.

Drewnowski, A., S.A. Henderson, A. Levine, and C. Hann. 1999. Taste and food preferences as predictors of dietary practices in young women. Public Health Nutr. 2:513-519.

Drewnowski, A., S.A. Henderson, and A.B. Shore. 1997. Taste responses to naringin, a flavonoid, and the acceptance of grapefruit juice are related to genetic sensitivity to 6-n- propylthiouracil. Am J Clin Nutr. 66:391-397.

Drewnowski, A., J.A. Mennella, S.L. Johnson, and F. Bellisle. 2012. Sweetness and food preference. J Nutr. 142:1142S-1148S.

Drewnowski, A., E.E. Shrager, C. Lipsky, E. Stellar, and M.R. Greenwood. 1989. Sugar and fat: sensory and hedonic evaluation of liquid and solid foods. Physiol Behav. 45:177-183.

Duffy, V.B., and L.M. Bartoshuk. 2000. Food acceptance and genetic variation in taste. J Am Diet Assoc. 100:647-655.

Duffy, V.B., J.E. Hayes, A.C. Davidson, J.R. Kidd, K.K. Kidd, and L.M. Bartoshuk. 2010. Vegetable Intake in College-Aged Adults Is Explained by Oral Sensory Phenotypes and TAS2R38 Genotype. Chemosens Percept. 3:137-148.

Edenberg, H.J., and Y. Liu. 2009. Laboratory methods for high-throughput genotyping. Cold Spring Harb Protoc. 2009:pdb top62.

126

El-Sohemy, A., L. Stewart, N. Khataan, B. Fontaine-Bisson, P. Kwong, S. Ozsungur, and M.C. Cornelis. 2007. Nutrigenomics of taste - impact on food preferences and food production. Forum Nutr. 60:176-182.

El-Yassimi, A., A. Hichami, P. Besnard, and N.A. Khan. 2008. Linoleic acid induces calcium signaling, Src kinase phosphorylation, and neurotransmitter release in mouse CD36- positive gustatory cells. J Biol Chem. 283:12949-12959.

Eny, K.M., P.N. Corey, and A. El-Sohemy. 2009. Dopamine D2 receptor genotype (C957T) and habitual consumption of sugars in a free-living population of men and women. J Nutrigenet Nutrigenomics. 2:235-242.

Eny, K.M., T.M. Wolever, P.N. Corey, and A. El-Sohemy. 2010. Genetic variation in TAS1R2 (Ile191Val) is associated with consumption of sugars in overweight and obese individuals in 2 distinct populations. Am J Clin Nutr. 92:1501-1510.

Eny, K.M., T.M. Wolever, B. Fontaine-Bisson, and A. El-Sohemy. 2008. Genetic variant in the glucose transporter type 2 (GLUT2) is associated with higher intakes of sugars in two distinct populations. Physiol Genomics. 33:355-360.

Erlanson-Albertsson, C. 2005. How palatable food disrupts appetite regulation. Basic Clin Pharmacol Toxicol. 97:61-73.

Ettinger, L., L. Duizer, and T. Caldwell. 2012. Body fat, sweetness sensitivity, and preference: determining the relationship. Can J Diet Pract Res. 73:45-48.

Feldman, G.M., A. Mogyorosi, G.L. Heck, J.A. DeSimone, C.R. Santos, R.A. Clary, and V. Lyall. 2003. Salt-evoked lingual surface potential in humans. J Neurophysiol. 90:2060- 2064.

Finger, T.E., B. Bottger, A. Hansen, K.T. Anderson, H. Alimohammadi, and W.L. Silver. 2003. Solitary chemoreceptor cells in the nasal cavity serve as sentinels of respiration. Proc Natl Acad Sci U S A. 100:8981-8986.

Fontaine-Bisson, B., T.M. Wolever, P.W. Connelly, P.N. Corey, and A. El-Sohemy. 2009. NF- kappaB -94Ins/Del ATTG polymorphism modifies the association between dietary polyunsaturated fatty acids and HDL-cholesterol in two distinct populations. Atherosclerosis. 204:465-470.

Frank, R.A., and N.J. van der Klaauw. 1994. The contribution of chemosensory factors to individual differences in reported food preferences. Appetite. 22:101-123.

Fricker, J., F. Fumeron, D. Clair, and M. Apfelbaum. 1989. A positive correlation between energy intake and body mass index in a population of 1312 overweight subjects. Int J Obes. 13:673-681.

Fukuwatari, T., T. Kawada, M. Tsuruta, T. Hiraoka, T. Iwanaga, E. Sugimoto, and T. Fushiki. 1997. Expression of the putative membrane fatty acid transporter (FAT) in taste buds of the circumvallate papillae in rats. FEBS Lett. 414:461-464.

127

Fukuwatari, T., K. Shibata, K. Iguchi, T. Saeki, A. Iwata, K. Tani, E. Sugimoto, and T. Fushiki. 2003. Role of gustation in the recognition of oleate and triolein in anosmic rats. Physiol Behav. 78:579-583.

Fuller, J.L. 1974. Single-locus control of saccharin preference in mice. J Hered. 65:33-36.

Fushan, A.A., C.T. Simons, J.P. Slack, and D. Drayna. 2010. Association between common variation in genes encoding sweet taste signaling components and human sucrose perception. Chem Senses. 35:579-592.

Fushan, A.A., C.T. Simons, J.P. Slack, A. Manichaikul, and D. Drayna. 2009. Allelic polymorphism within the TAS1R3 promoter is associated with human taste sensitivity to sucrose. Curr Biol. 19:1288-1293.

Gaillard, D., F. Laugerette, N. Darcel, A. El-Yassimi, P. Passilly-Degrace, A. Hichami, N.A. Khan, J.P. Montmayeur, and P. Besnard. 2008. The gustatory pathway is involved in CD36-mediated orosensory perception of long-chain fatty acids in the mouse. FASEB J. 22:1458-1468.

Galindo-Cuspinera, V., T. Waeber, N. Antille, C. Hartmann, N. Stead, and N. Martin. 2009. Reliability of Threshold and Suprathreshold Methods for Taste Phenotyping: Characterization with PROP and Sodium Chloride. Chemosens Percept. 2:214-228.

Garcia-Bailo B, T.C., Eny K, El-Sohemy A. 2008. Genetic Variation in Taste and Its Influence on Food Selection. Omics : a journal of integrative biology.

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

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

Geleijnse, J.M., A. Hofman, J.C. Witteman, A.A. Hazebroek, H.A. Valkenburg, and D.E. Grobbee. 1997. Long-term effects of neonatal sodium restriction on blood pressure. Hypertension. 29:913-917.

Gilbertson, T.A., D.T. Fontenot, L. Liu, H. Zhang, and W.T. Monroe. 1997. Fatty acid modulation of K+ channels in taste receptor cells: gustatory cues for dietary fat. Am J Physiol. 272:C1203-1210.

Gilbertson, T.A., L. Liu, I. Kim, C.A. Burks, and D.R. Hansen. 2005. Fatty acid responses in taste cells from obesity-prone and -resistant rats. Physiol Behav. 86:681-690.

Glanville, E.V., and A.R. Kaplan. 1965. Food Preference and Sensitivity of Taste for Bitter Compounds. Nature. 205:851-853.

Glanz, K., M. Basil, E. Maibach, J. Goldberg, and D. Snyder. 1998. Why Americans eat what they do: taste, nutrition, cost, convenience, and weight control concerns as influences on food consumption. J Am Diet Assoc. 98:1118-1126.

128

Goldstein, G.L., H. Daun, and B.J. Tepper. 2005. Adiposity in middle-aged women is associated with genetic taste blindness to 6-n-propylthiouracil. Obes Res. 13:1017-1023.

Greene, L.S., J.A. Desor, and O. Maller. 1975. Heredity and experience: their relative importance in the development of taste preference in man. J Comp Physiol Psychol. 89:279-284.

Grzegorczyk, P.B., S.W. Jones, and C.M. Mistretta. 1979. Age-related differences in salt taste acuity. J Gerontol. 34:834-840.

Guo, S.W., and D.R. Reed. 2001. The genetics of phenylthiocarbamide perception. Ann Hum Biol. 28:111-142.

Guthrie, J.F., and J.F. Morton. 2000. Food sources of added sweeteners in the diets of Americans. J Am Diet Assoc. 100:43-51, quiz 49-50.

Hamilton, C.L. 1964. Rat's Preference for High Fat Diets. J Comp Physiol Psychol. 58:459-460.

Hansen, L., and M.S. Rose. 1996. Sensory acceptability is inversely related to development of fat rancidity in bread made from stored flour. J Am Diet Assoc. 96:792-793.

Harmon, C.M., and N.A. Abumrad. 1993. Binding of sulfosuccinimidyl fatty acids to adipocyte membrane proteins: isolation and amino-terminal sequence of an 88-kD protein implicated in transport of long-chain fatty acids. J Membr Biol. 133:43-49.

Havas, S., B.D. Dickinson, and M. Wilson. 2007. The urgent need to reduce sodium consumption. JAMA. 298:1439-1441.

Hayes, J.E., and V.B. Duffy. 2008. Oral sensory phenotype identifies level of sugar and fat required for maximal liking. Physiol Behav. 95:77-87.

Hayes, J.E., M.R. Wallace, V.S. Knopik, D.M. Herbstman, L.M. Bartoshuk, and V.B. Duffy. 2011. Allelic variation in TAS2R bitter receptor genes associates with variation in sensations from and ingestive behaviors toward common bitter beverages in adults. Chem Senses. 36:311-319.

Hayes, P., H.J. Meadows, M.J. Gunthorpe, M.H. Harries, D.M. Duckworth, W. Cairns, D.C. Harrison, C.E. Clarke, K. Ellington, R.K. Prinjha, A.J. Barton, A.D. Medhurst, G.D. Smith, S. Topp, P. Murdock, G.J. Sanger, J. Terrett, O. Jenkins, C.D. Benham, A.D. Randall, I.S. Gloger, and J.B. Davis. 2000. Cloning and functional expression of a human orthologue of rat vanilloid receptor-1. Pain. 88:205-215.

Health Canada. 2009. The Issue of Sodium.

Heck, G.L., S. Mierson, and J.A. DeSimone. 1984. Salt taste transduction occurs through an amiloride-sensitive sodium transport pathway. Science. 223:403-405.

Heilmann, S., and T. Hummel. 2004. A new method for comparing orthonasal and retronasal olfaction. Behav Neurosci. 118:412-419.

129

Hein, K.A., S.R. Jaeger, B. Tom Carr, and C.M. Delahunty. 2008. Comparison of five common acceptance and preference methods. Food Quality and Preference. 19:651-661.

Hein, K.A., Jaeger, S. R., Carr, B. T., Delahunty, C. M. 2008. Comparison of five common acceptance and preference methods. Food quality and preference. 19:651-661.

Henkin, R.I., and R.S. Shallenberger. 1970. Aglycogeusia: the inability to recognize sweetness and its possible molecular basis. Nature. 227:965-966.

Henney, J.E., C.L. Taylor, and C.S. Boon. 2010. Strategies to Reduce Sodium Intake in the United States. J.E. Henney, C.L. Taylor, and C.S. Boon, editors. National Academies Press (US), Washington (DC).

Hirasawa, A., K. Tsumaya, T. Awaji, S. Katsuma, T. Adachi, M. Yamada, Y. Sugimoto, S. Miyazaki, and G. Tsujimoto. 2005. Free fatty acids regulate gut incretin glucagon-like peptide-1 secretion through GPR120. Nat Med. 11:90-94.

Hladik, C.M., P. Pasquet, and B. Simmen. 2002. New perspectives on taste and primate evolution: the dichotomy in gustatory coding for perception of beneficent versus noxious substances as supported by correlations among human thresholds. Am J Phys Anthropol. 117:342-348.

Holmes, M.D., I.J. Powell, H. Campos, M.J. Stampfer, E.L. Giovannucci, and W.C. Willett. 2007. Validation of a food frequency questionnaire measurement of selected nutrients using biological markers in African-American men. Eur J Clin Nutr. 61:1328-1336.

Hoon, M.A., E. Adler, J. Lindemeier, J.F. Battey, N.J. Ryba, and C.S. Zuker. 1999. Putative mammalian taste receptors: a class of taste-specific GPCRs with distinct topographic selectivity. Cell. 96:541-551.

Horio, N., M. Jyotaki, R. Yoshida, K. Sanematsu, N. Shigemura, and Y. Ninomiya. 2010. New frontiers in gut nutrient sensor research: nutrient sensors in the gastrointestinal tract: modulation of sweet taste sensitivity by leptin. J Pharmacol Sci. 112:8-12.

Huang, Y.A., and S.D. Roper. 2010. Intracellular Ca(2+) and TRPM5-mediated membrane depolarization produce ATP secretion from taste receptor cells. J Physiol. 588:2343- 2350.

Huang, Y.J., Y. Maruyama, G. Dvoryanchikov, E. Pereira, N. Chaudhari, and S.D. Roper. 2007. The role of pannexin 1 hemichannels in ATP release and cell-cell communication in mouse taste buds. Proc Natl Acad Sci U S A. 104:6436-6441.

Hunter, D.J., E.B. Rimm, F.M. Sacks, M.J. Stampfer, G.A. Colditz, L.B. Litin, and W.C. Willett. 1992. Comparison of measures of fatty acid intake by subcutaneous fat aspirate, food frequency questionnaire, and diet records in a free-living population of US men. Am J Epidemiol. 135:418-427.

130

Hunter, D.J., D. Spiegelman, H.O. Adami, L. Beeson, P.A. van den Brandt, A.R. Folsom, G.E. Fraser, R.A. Goldbohm, S. Graham, G.R. Howe, and et al. 1996. Cohort studies of fat intake and the risk of breast cancer--a pooled analysis. N Engl J Med. 334:356-361.

Intranuovo, L.R., and A.S. Powers. 1998. The perceived bitterness of beer and 6-n- propylthiouracil (PROP) taste sensitivity. Ann N Y Acad Sci. 855:813-815.

Kaminski, L.C., S.A. Henderson, and A. Drewnowski. 2000. Young women's food preferences and taste responsiveness to 6-n-propylthiouracil (PROP). Physiol Behav. 68:691-697.

Kawai, K., K. Sugimoto, K. Nakashima, H. Miura, and Y. Ninomiya. 2000. Leptin as a modulator of sweet taste sensitivities in mice. Proc Natl Acad Sci U S A. 97:11044- 11049.

Kawai, T., and T. Fushiki. 2003. Importance of lipolysis in oral cavity for orosensory detection of fat. Am J Physiol Regul Integr Comp Physiol. 285:R447-454.

Keller, K.L., L.C. Liang, J. Sakimura, D. May, C. van Belle, C. Breen, E. Driggin, B.J. Tepper, P.C. Lanzano, L. Deng, and W.K. Chung. 2012. Common variants in the CD36 gene are associated with oral fat perception, fat preferences, and obesity in African Americans. Obesity (Silver Spring, Md ). 20:1066-1073.

Keskitalo, K., H. Tuorila, T.D. Spector, L.F. Cherkas, A. Knaapila, K. Silventoinen, and M. Perola. 2007. Same genetic components underlie different measures of sweet taste preference. Am J Clin Nutr. 86:1663-1669.

Kho, M., J.E. Lee, Y.M. Song, K. Lee, K. Kim, S. Yang, H. Joung, and J. Sung. 2013. Genetic and environmental influences on sodium intake determined by using half-day urine samples: the Healthy Twin Study. Am J Clin Nutr. 98:1410-1416.

Kim, H., J.K. Neubert, A. San Miguel, K. Xu, R.K. Krishnaraju, M.J. Iadarola, D. Goldman, and R.A. Dionne. 2004a. Genetic influence on variability in human acute experimental pain sensitivity associated with gender, ethnicity and psychological temperament. Pain. 109:488-496.

Kim, U.K., P.A. Breslin, D. Reed, and D. Drayna. 2004b. Genetics of human taste perception. J Dent Res. 83:448-453.

Kim, U.K., E. Jorgenson, H. Coon, M. Leppert, N. Risch, and D. Drayna. 2003. Positional cloning of the human quantitative trait locus underlying taste sensitivity to phenylthiocarbamide. Science. 299:1221-1225.

Kirkmeyer, S.V., and B.J. Tepper. 2003. Understanding creaminess perception of dairy products using free-choice profiling and genetic responsivity to 6-n-propylthiouracil. Chem Senses. 28:527-536.

Kitagawa, M., Y. Kusakabe, H. Miura, Y. Ninomiya, and A. Hino. 2001. Molecular genetic identification of a candidate receptor gene for sweet taste. Biochem Biophys Res Commun. 283:236-242.

131

LaFramboise, T. 2009. Single nucleotide polymorphism arrays: a decade of biological, computational and technological advances. Nucleic Acids Res. 37:4181-4193.

Lamb, M.W., and B.C. Ling. 1946. An analysis of food consumption and preferences of nursery school children. Child Dev. 17:187-217.

Laugerette, F., D. Gaillard, P. Passilly-Degrace, I. Niot, and P. Besnard. 2007. Do we taste fat? Biochimie. 89:265-269.

Laugerette, F., P. Passilly-Degrace, B. Patris, I. Niot, M. Febbraio, J.P. Montmayeur, and P. Besnard. 2005. CD36 involvement in orosensory detection of dietary lipids, spontaneous fat preference, and digestive secretions. J Clin Invest. 115:3177-3184.

Lawless, H.T., and G.J. Malone. 1986. THE DISRIMINATIVE EFFICIENCY OF COMMON SCALING METHODS. Journal of Sensory Studies. 1:85-98.

Lee, J.E., and A.T. Chan. 2011. Fruit, vegetables, and folate: cultivating the evidence for cancer prevention. Gastroenterology. 141:16-20.

Leshem, M. 2009. Biobehavior of the human love of salt. Neurosci Biobehav Rev. 33:1-17.

Li, X., W. Li, H. Wang, J. Cao, K. Maehashi, L. Huang, A.A. Bachmanov, D.R. Reed, V. Legrand-Defretin, G.K. Beauchamp, and J.G. Brand. 2005. Pseudogenization of a sweet- receptor gene accounts for cats' indifference toward sugar. PLoS Genet. 1:27-35.

Liao, J., and P.G. Schultz. 2003. Three sweet receptor genes are clustered in human chromosome 1. Mamm Genome. 14:291-301.

Liu, D., and E.R. Liman. 2003. Intracellular Ca2+ and the phospholipid PIP2 regulate the taste transduction ion channel TRPM5. Proc Natl Acad Sci U S A. 100:15160-15165.

Lö tsch J, F.h.K., Neddermayer T. 2009. The consequence of concomitantly present functional genetic variants for the identification of functional genotype-phenotype associations in pain. Clin Pharmacol Ther. 85:25-30.

Love-Gregory, L., R. Sherva, T. Schappe, J.S. Qi, J. McCrea, S. Klein, M.A. Connelly, and N.A. Abumrad. 2011. Common CD36 SNPs reduce protein expression and may contribute to a protective atherogenic profile. Hum Mol Genet. 20:193-201.

Lyall, V., G.L. Heck, A.K. Vinnikova, S. Ghosh, T.H. Phan, R.I. Alam, O.F. Russell, S.A. Malik, J.W. Bigbee, and J.A. DeSimone. 2004. The mammalian amiloride-insensitive non-specific salt taste receptor is a vanilloid receptor-1 variant. J Physiol. 558:147-159.

Ma, X., S. Bacci, W. Mlynarski, L. Gottardo, T. Soccio, C. Menzaghi, E. Iori, R.A. Lager, A.R. Shroff, E.V. Gervino, R.W. Nesto, M.T. Johnstone, N.A. Abumrad, A. Avogaro, V. Trischitta, and A. Doria. 2004. A common haplotype at the CD36 locus is associated with high free fatty acid levels and increased cardiovascular risk in Caucasians. Hum Mol Genet. 13:2197-2205.

132

Mace, O.J., J. Affleck, N. Patel, and G.L. Kellett. 2007. Sweet taste receptors in rat small intestine stimulate glucose absorption through apical GLUT2. J Physiol. 582:379-392.

MacGregor, G., and H.E. de Wardener. 1998. Salt, diet and health: Neptune's poisoned chalice: The origins of high blood pressure. . University Press, Cambridge.

Madden, J., J.J. Carrero, A. Brunner, N. Dastur, C.P. Shearman, P.C. Calder, and R.F. Grimble. 2008. 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. 78:327-335.

Manabe, Y., S. Matsumura, and T. Fushiki. 2010. Preference for High-Fat Food in Animals.

Mattes, R.D. 1997. The taste for salt in humans. Am J Clin Nutr. 65:692S-697S.

Mattes, R.D. 2002. Oral fat exposure increases the first phase triacylglycerol concentration due to release of stored lipid in humans. J Nutr. 132:3656-3662.

Mattes, R.D. 2005. Fat taste and lipid metabolism in humans. Physiol Behav. 86:691-697.

Mattes, R.D. 2009a. Is there a fatty acid taste? Annu Rev Nutr. 29:305-327.

Mattes, R.D. 2009b. Oral detection of short-, medium-, and long-chain free fatty acids in humans. Chem Senses. 34:145-150.

Mattes, R.D. 2009c. Oral thresholds and suprathreshold intensity ratings for free fatty acids on 3 tongue sites in humans: implications for transduction mechanisms. Chem Senses. 34:415- 423.

Max, M., Y.G. Shanker, L. Huang, M. Rong, Z. Liu, F. Campagne, H. Weinstein, S. Damak, and R.F. Margolskee. 2001. Tas1r3, encoding a new candidate taste receptor, is allelic to the sweet responsiveness locus Sac. Nat Genet. 28:58-63.

McCaughey, S.A. 2008. The taste of sugars. Neurosci Biobehav Rev. 32:1024-1043.

McKeown, N.M., N.E. Day, A.A. Welch, S.A. Runswick, R.N. Luben, A.A. Mulligan, A. McTaggart, and S.A. Bingham. 2001. Use of biological markers to validate self-reported dietary intake in a random sample of the European Prospective Investigation into Cancer United Kingdom Norfolk cohort. Am J Clin Nutr. 74:188-196.

McLaughlin, S.K., P.J. McKinnon, N. Spickofsky, W. Danho, and R.F. Margolskee. 1994. Molecular cloning of G proteins and phosphodiesterases from rat taste cells. Physiol Behav. 56:1157-1164.

Meiselman, H., D. Waterman, and L. Symington. 1974. Armed Forces food preferences. U.S.A.N.D. Center, editor, Natick.

Mela, D.J., and D.A. Sacchetti. 1991. Sensory preferences for fats: relationships with diet and body composition. Am J Clin Nutr. 53:908-915.

133

Mennella, J.A., S. Finkbeiner, and D.R. Reed. 2012. The proof is in the pudding: children prefer lower fat but higher sugar than do mothers. Int J Obes (Lond). 36:1285-1291.

Merchant, A.T., H. Vatanparast, S. Barlas, M. Dehghan, S.M. Shah, L. De Koning, and S.E. Steck. 2009. Carbohydrate intake and overweight and obesity among healthy adults. J Am Diet Assoc. 109:1165-1172.

Meyerhof, W. 2005. Elucidation of mammalian bitter taste. Rev Physiol Biochem Pharmacol. 154:37-72.

Meyerhof, W., C. Batram, C. Kuhn, A. Brockhoff, E. Chudoba, B. Bufe, G. Appendino, and M. Behrens. 2010. The molecular receptive ranges of human TAS2R bitter taste receptors. Chem Senses. 35:157-170.

Molag, M.L., J.H. de Vries, M.C. Ocke, P.C. Dagnelie, P.A. van den Brandt, M.C. Jansen, W.A. van Staveren, and P. van't Veer. 2007. Design characteristics of food frequency questionnaires in relation to their validity. Am J Epidemiol. 166:1468-1478.

Monneuse, M.O., F. Bellisle, and J. Louis-Sylvestre. 1991. Impact of sex and age on sensory evaluation of sugar and fat in dairy products. Physiol Behav. 50:1111-1117.

Montmayeur, J.P., and J. le Coutre. 2010. Fat Detection: Taste, Texture, and Post Ingestive Effects. CRC Press, Boca Raton.

Montmayeur, J.P., S.D. Liberles, H. Matsunami, and L.B. Buck. 2001. A candidate taste receptor gene near a sweet taste locus. Nat Neurosci. 4:492-498.

Moriyama, K., M.M. Bakre, F. Ahmed, N. Spickofsky, M. Max, and R.F. Margolskee. 2002. Assaying G protein-phosphodiesterase interactions in sensory systems. Methods Enzymol. 345:37-48.

Morris, M.J., E.S. Na, and A.K. Johnson. 2008. Salt craving: the psychobiology of pathogenic sodium intake. Physiol Behav. 94:709-721.

Munger, R.G., A.R. Folsom, L.H. Kushi, S.A. Kaye, and T.A. Sellers. 1992. Dietary assessment of older Iowa women with a food frequency questionnaire: nutrient intake, reproducibility, and comparison with 24-hour dietary recall interviews. Am J Epidemiol. 136:192-200.

Nakagawa, Y., M. Nagasawa, S. Yamada, A. Hara, H. Mogami, V.O. Nikolaev, M.J. Lohse, N. Shigemura, Y. Ninomiya, and I. Kojima. 2009. Sweet taste receptor expressed in pancreatic beta-cells activates the calcium and cyclic AMP signaling systems and stimulates insulin secretion. PLoS One. 4:e5106.

Nakamura, Y., K. Sanematsu, R. Ohta, S. Shirosaki, K. Koyano, K. Nonaka, N. Shigemura, and Y. Ninomiya. 2008. Diurnal variation of human sweet taste recognition thresholds is correlated with plasma leptin levels. Diabetes. 57:2661-2665.

134

Nasser, J.A., H.R. Kissileff, C.N. Boozer, C.J. Chou, and F.X. Pi-Sunyer. 2001. PROP taster status and oral fatty acid perception. Eat Behav. 2:237-245.

Nelson, G., M.A. Hoon, J. Chandrashekar, Y. Zhang, N.J. Ryba, and C.S. Zuker. 2001. Mammalian sweet taste receptors. Cell. 106:381-390.

Pepino, M.Y., L. Love-Gregory, S. Klein, and N.A. Abumrad. 2012. The fatty acid translocase gene CD36 and lingual lipase influence oral sensitivity to fat in obese subjects. J Lipid Res. 53:561-566.

Pepino, M.Y., and J.A. Mennella. 2005. Factors contributing to individual differences in sucrose preference. Chem Senses. 30 Suppl 1:i319-320.

Popkin, B.M., and S.J. Nielsen. 2003. The sweetening of the world's diet. Obes Res. 11:1325- 1332.

Raine, K.D. 2005. Determinants of healthy eating in Canada: an overview and synthesis. Can J Public Health. 96 Suppl 3:S8-14, S18-15.

Rankin, K.M., and R.D. Mattes. 1996. Role of food familiarity and taste quality in food preferences of individuals with Prader-Willi syndrome. Int J Obes Relat Metab Disord. 20:759-762.

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

Reed, D.R., T. Tanaka, and A.H. McDaniel. 2006. Diverse tastes: Genetics of sweet and bitter perception. Physiol Behav. 88:215-226.

Reed, D.R., G. Zhu, P.A. Breslin, F.F. Duke, A.K. Henders, M.J. Campbell, G.W. Montgomery, S.E. Medland, N.G. Martin, and M.J. Wright. 2010. The perception of quinine taste intensity is associated with common genetic variants in a bitter receptor cluster on chromosome 12. Hum Mol Genet. 19:4278-4285.

Refsgaard, H.H., P.M. Brockhoff, and B. Jensen. 2000. Free polyunsaturated fatty acids cause taste deterioration of salmon during frozen storage. J Agric Food Chem. 48:3280-3285.

Ren, X., L. Zhou, R. Terwilliger, S.S. Newton, and I.E. de Araujo. 2009. Sweet taste signaling functions as a hypothalamic glucose sensor. Front Integr Neurosci. 3:12.

Roberts, C.D., G. Dvoryanchikov, S.D. Roper, and N. Chaudhari. 2009. Interaction between the second messengers cAMP and Ca2+ in mouse presynaptic taste cells. J Physiol. 587:1657-1668.

Romanov, R.A., O.A. Rogachevskaja, M.F. Bystrova, P. Jiang, R.F. Margolskee, and S.S. Kolesnikov. 2007. Afferent neurotransmission mediated by hemichannels in mammalian taste cells. EMBO J. 26:657-667.

Roper, S.D. 2007. Signal transduction and information processing in mammalian taste buds. Pflugers Arch. 454:759-776.

135

Rossler, P., C. Kroner, J. Freitag, J. Noe, and H. Breer. 1998. Identification of a phospholipase C beta subtype in rat taste cells. Eur J Cell Biol. 77:253-261.

Ruiz, C., S. Gutknecht, E. Delay, and S. Kinnamon. 2006. Detection of NaCl and KCl in TRPV1 knockout mice. Chem Senses. 31:813-820.

Running, C.A., R.D. Mattes, and R.M. Tucker. 2013. Fat taste in humans: Sources of within- and between-subject variability. Prog Lipid Res. 52:438-445.

Sainz, E., J.N. Korley, J.F. Battey, and S.L. Sullivan. 2001. Identification of a novel member of the T1R family of putative taste receptors. J Neurochem. 77:896-903.

Salbe, A.D., A. DelParigi, R.E. Pratley, A. Drewnowski, and P.A. Tataranni. 2004. Taste preferences and body weight changes in an obesity-prone population. Am J Clin Nutr. 79:372-378.

Schiffman, S., and C. Dackis. 1975. Taste of nutrients: Amino acids, vitamins, and fatty acids. Perception & Psychophysics. 17:140-146.

Sclafani, A., K. Ackroff, and N.A. Abumrad. 2007a. CD36 gene deletion reduces fat preference and intake but not post-oral fat conditioning in mice. Am J Physiol Regul Integr Comp Physiol. 293:R1823-1832.

Sclafani, A., S. Zukerman, J.I. Glendinning, and R.F. Margolskee. 2007b. Fat and carbohydrate preferences in mice: the contribution of alpha-gustducin and Trpm5 taste-signaling proteins. Am J Physiol Regul Integr Comp Physiol. 293:R1504-1513.

Segato, F.N., C. Castro-Souza, E.N. Segato, S. Morato, and N.C. Coimbra. 1997. Sucrose ingestion causes opioid analgesia. Braz J Med Biol Res. 30:981-984.

Shiffman, D., S.G. Ellis, C.M. Rowland, M.J. Malloy, M.M. Luke, O.A. Iakoubova, C.R. Pullinger, J. Cassano, B.E. Aouizerat, R.G. Fenwick, R.E. Reitz, J.J. Catanese, D.U. Leong, C. Zellner, J.J. Sninsky, E.J. Topol, J.J. Devlin, and J.P. Kane. 2005. Identification of four gene variants associated with myocardial infarction. Am J Hum Genet. 77:596-605.

Shiffman, D., E.S. O'Meara, L.A. Bare, C.M. Rowland, J.Z. Louie, A.R. Arellano, T. Lumley, K. Rice, O. Iakoubova, M.M. Luke, B.A. Young, M.J. Malloy, J.P. Kane, S.G. Ellis, R.P. Tracy, J.J. Devlin, and B.M. Psaty. 2008. Association of gene variants with incident myocardial infarction in the Cardiovascular Health Study. Arterioscler Thromb Vasc Biol. 28:173-179.

Shigemura, N., T. Ohkuri, C. Sadamitsu, K. Yasumatsu, R. Yoshida, G.K. Beauchamp, A.A. Bachmanov, and Y. Ninomiya. 2008. Amiloride-sensitive NaCl taste responses are associated with genetic variation of ENaC alpha-subunit in mice. Am J Physiol Regul Integr Comp Physiol. 294:R66-75.

136

Shigemura, N., R. Ohta, Y. Kusakabe, H. Miura, A. Hino, K. Koyano, K. Nakashima, and Y. Ninomiya. 2004. Leptin modulates behavioral responses to sweet substances by influencing peripheral taste structures. Endocrinology. 145:839-847.

Silverstein, R.L., and M. Febbraio. 2009. CD36, a scavenger receptor involved in immunity, metabolism, angiogenesis, and behavior. Sci Signal. 2:re3.

Simon, S.A., I.E. de Araujo, R. Gutierrez, and M.A. Nicolelis. 2006. The neural mechanisms of gustation: a distributed processing code. Nat Rev Neurosci. 7:890-901.

Simons, P.J., J.A. Kummer, J.J. Luiken, and L. Boon. 2011. Apical CD36 immunolocalization in human and porcine taste buds from circumvallate and foliate papillae. Acta Histochem. 113:839-843.

Smith, D.V., and C.A. Ossebaard. 1995. Amiloride suppression of the taste intensity of sodium chloride: evidence from direct magnitude scaling. Physiol Behav. 57:773-777.

Spickofsky, N., A. Robichon, W. Danho, D. Fry, D. Greeley, B. Graves, V. Madison, and R.F. Margolskee. 1994. Biochemical analysis of the transducin-phosphodiesterase interaction. Nat Struct Biol. 1:771-781.

Steinmetz, K.A., and J.D. Potter. 1996. Vegetables, fruit, and cancer prevention: a review. J Am Diet Assoc. 96:1027-1039.

Stewart, J.E., C. Feinle-Bisset, M. Golding, C. Delahunty, P.M. Clifton, and R.S. Keast. 2010. Oral sensitivity to fatty acids, food consumption and BMI in human subjects. Br J Nutr. 104:145-152.

Stewart, J.E., and R.S. Keast. 2012. Recent fat intake modulates fat taste sensitivity in lean and overweight subjects. Int J Obes (Lond). 36:834-842.

Stewart, R.E., J.A. DeSimone, and D.L. Hill. 1997. New perspectives in a gustatory physiology: transduction, development, and plasticity. Am J Physiol. 272:C1-26.

Su, X., and N.A. Abumrad. 2009. Cellular fatty acid uptake: a pathway under construction. Trends Endocrinol Metab. 20:72-77.

Subar, A.F., F.E. Thompson, V. Kipnis, D. Midthune, P. Hurwitz, S. McNutt, A. McIntosh, and S. Rosenfeld. 2001. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires : the Eating at America's Table Study. Am J Epidemiol. 154:1089-1099.

Szallasi, A., D.N. Cortright, C.A. Blum, and S.R. Eid. 2007. The vanilloid receptor TRPV1: 10 years from channel cloning to antagonist proof-of-concept. Nat Rev Drug Discov. 6:357- 372.

Takeda, M., S. Sawano, M. Imaizumi, and T. Fushiki. 2001. Preference for corn oil in olfactory- blocked mice in the conditioned place preference test and the two-bottle choice test. Life Sci. 69:847-854.

137

Tang, K., D.J. Fu, D. Julien, A. Braun, C.R. Cantor, and H. Koster. 1999. Chip-based genotyping by mass spectrometry. Proc Natl Acad Sci U S A. 96:10016-10020.

Teff, K. 2000. Nutritional implications of the cephalic-phase reflexes: endocrine responses. Appetite. 34:206-213.

Tepper, B.J. 1998. 6-n-Propylthiouracil: a genetic marker for taste, with implications for food preference and dietary habits. Am J Hum Genet. 63:1271-1276.

Tepper, B.J. 2008. Nutritional implications of genetic taste variation: the role of PROP sensitivity and other taste phenotypes. Annu Rev Nutr. 28:367-388.

Tepper, B.J., and R.J. Nurse. 1997. Fat perception is related to PROP taster status. Physiol Behav. 61:949-954.

Thompson, F.E., and T. Byers. 1994. Dietary assessment resource manual. J Nutr. 124:2245S- 2317S.

Toguri, C. 2008. Genetic Variation in CD36 and Dietary Fat Intake. University of Toronto.

Tordoff, M.G., A.A. Bachmanov, and D.R. Reed. 2007. Forty mouse strain survey of water and sodium intake. Physiol Behav. 91:620-631.

Trachtenberg, A.J., J.H. Robert, A.E. Abdalla, A. Fraser, S.Y. He, J.N. Lacy, C. Rivas-Morello, A. Truong, G. Hardiman, L. Ohno-Machado, F. Liu, E. Hovig, and W.P. Kuo. 2012. A primer on the current state of microarray technologies. Methods Mol Biol. 802:3-17.

Treesukosol, Y., V. Lyall, G.L. Heck, J.A. DeSimone, and A.C. Spector. 2007. A psychophysical and electrophysiological analysis of salt taste in Trpv1 null mice. Am J Physiol Regul Integr Comp Physiol. 292:R1799-1809.

Trubey, K.R., S. Culpepper, Y. Maruyama, S.C. Kinnamon, and N. Chaudhari. 2006. Tastants evoke cAMP signal in taste buds that is independent of calcium signaling. Am J Physiol Cell Physiol. 291:C237-244.

Turnbull, F. 2003. Effects of different blood-pressure-lowering regimens on major cardiovascular events: results of prospectively-designed overviews of randomised trials. Lancet. 362:1527-1535.

Ueda, T., S. Ugawa, H. Yamamura, Y. Imaizumi, and S. Shimada. 2003. Functional interaction between T2R taste receptors and G-protein alpha subunits expressed in taste receptor cells. J Neurosci. 23:7376-7380.

Valdes, A.M., G. De Wilde, S.A. Doherty, R.J. Lories, F.L. Vaughn, L.L. Laslett, R.A. Maciewicz, A. Soni, D.J. Hart, W. Zhang, K.R. Muir, E.M. Dennison, M. Wheeler, P. Leaverton, C. Cooper, T.D. Spector, F.M. Cicuttini, V. Chapman, G. Jones, N.K. Arden, and M. Doherty. 2011. The Ile585Val TRPV1 variant is involved in risk of painful knee osteoarthritis. Ann Rheum Dis. 70:1556-1561.

138

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

Vega, M.A., B. Segui-Real, J.A. Garcia, C. Cales, F. Rodriguez, J. Vanderkerckhove, and I.V. Sandoval. 1991. Cloning, sequencing, and expression of a cDNA encoding rat LIMP II, a novel 74-kDa lysosomal membrane protein related to the surface adhesion protein CD36. J Biol Chem. 266:16818-16824.

Wang, Y.C., S.N. Bleich, and S.L. Gortmaker. 2008. Increasing caloric contribution from sugar- sweetened beverages and 100% fruit juices among US children and adolescents, 1988- 2004. Pediatrics. 121:e1604-1614.

Wellendorph, P., L.D. Johansen, and H. Brauner-Osborne. 2009. Molecular pharmacology of promiscuous seven transmembrane receptors sensing organic nutrients. Mol Pharmacol. 76:453-465.

WHO. 2003. Diet, nutrition and the prevention of chronic diseases: Report of a joint who/fao expert consultation. World Health Organization.

Willet, W. 1998. Nutritional epidemiology. Oxford University Press, New York.

Willett, W. 2001. Invited commentary: a further look at dietary questionnaire validation. Am J Epidemiol. 154:1100-1102; discussion 1105-1106.

Willett, W.C., L. Sampson, M.J. Stampfer, B. Rosner, C. Bain, J. Witschi, C.H. Hennekens, and F.E. Speizer. 1985. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol. 122:51-65.

Wise, P.M., J.L. Hansen, D.R. Reed, and P.A. Breslin. 2007. Twin study of the heritability of recognition thresholds for sour and salty taste. Chem Senses. 32:749-754.

Wu, S.V., N. Rozengurt, M. Yang, S.H. Young, J. Sinnett-Smith, and E. Rozengurt. 2002. Expression of bitter taste receptors of the T2R family in the gastrointestinal tract and enteroendocrine STC-1 cells. Proc Natl Acad Sci U S A. 99:2392-2397.

Xu, S., A. Jay, K. Brunaldi, N. Huang, and J.A. Hamilton. 2013. CD36 enhances fatty acid uptake by increasing the rate of intracellular esterification but not transport across the plasma membrane. Biochemistry. 52:7254-7261.

Yeomans, M.R., R.W. Gray, C.J. Mitchell, and S. True. 1997. Independent effects of palatability and within-meal pauses on intake and appetite ratings in human volunteers. Appetite. 29:61-76.

Young, R.L., K. Sutherland, N. Pezos, S.M. Brierley, M. Horowitz, C.K. Rayner, and L.A. Blackshaw. 2009. Expression of taste molecules in the upper gastrointestinal tract in humans with and without type 2 diabetes. Gut. 58:337-346.

139

Zhang, Y., M.A. Hoon, J. Chandrashekar, K.L. Mueller, B. Cook, D. Wu, C.S. Zuker, and N.J. Ryba. 2003. Coding of sweet, bitter, and umami tastes: different receptor cells sharing similar signaling pathways. Cell. 112:293-301.

Zhao, G.Q., Y. Zhang, M.A. Hoon, J. Chandrashekar, I. Erlenbach, N.J. Ryba, and C.S. Zuker. 2003. The receptors for mammalian sweet and umami taste. Cell. 115:255-266.

Zheng, H., and H.R. Berthoud. 2008. Neural systems controlling the drive to eat: mind versus metabolism. Physiology (Bethesda). 23:75-83.

Zheng, H., N.R. Lenard, A.C. Shin, and H.R. Berthoud. 2009. Appetite control and energy balance regulation in the modern world: reward-driven brain overrides repletion signals. Int J Obes (Lond). 33 Suppl 2:S8-13.

140

Statement of Publications The research presented in this thesis has appeared or has been submitted as a series of original publications in refereed journals.

Chapter 3 Dias AG, Rousseau D, Duizer L, Cockburn M, Chiu W, Nielsen D, El-Sohemy A. (2013) Genetic Variation in Putative Salt Taste Receptors and Salt Taste Perception in Humans. Chemical Senses 38(2):137-45. Copyright Chemical Senses. http://chemse.oxfordjournals.org/content/38/2/137.long