|Home | About | Journals | Submit | Contact Us | Français|
Eating is dangerous. While food contains nutrients and calories that animals need to produce heat and energy, it may also contain harmful parasites, bacteria, or chemicals. To guide food selection, the senses of taste and smell have evolved to alert us to the bitter taste of poisons and the sour taste and off-putting smell of spoiled foods. These sensory systems help people and animals to eat defensively, and they provide the brake that helps them avoid ingesting foods that are harmful. But choices about which foods to eat are motivated by more than avoiding the bad; they are also motivated by seeking the good, such as fat and sugar. However, just as not everyone is equally capable of sensing toxins in food, not everyone is equally enthusiastic about consuming high-fat, high-sugar foods. Genetic studies in humans and experimental animals strongly suggest that the liking of sugar and fat is influenced by genotype; likewise, the abilities to detect bitterness and the malodors of rotting food are highly variable among individuals. Understanding the exact genes and genetic differences that affect food intake may provide important clues in obesity treatment by allowing caregivers to tailor dietary recommendations to the chemosensory landscape of each person.
Whenever anything is introduced directly into the body, there is a risk that it will be harmful. The air we breathe, the water we drink, and the food we eat give us oxygen, water, and nutrients, but they also have the potential, because of the presence of poisons and pathogens, to make us very ill, and they may even kill us. Our sensory systems are designed to help us detect and avoid these outcomes, through vision (seeing contaminants in our food), touch, taste, smell, and the common chemical sense—a sense that is not well known (with no universally accepted name) and that includes the sting of carbon dioxide and the burn of hot peppers. Together, these senses help protect us from bad food. Yet whenever anything is deliberately eaten or drunk, there is an expectation that it will be beneficial. Beverages, foods, and chemical compounds all are ingested because we are motivated to do so for pleasure or for the relief of unpleasant states such as thirst, hunger, or tiredness (e.g., caffeinated drinks such as coffee). These senses help us to distinguish not only good from bad food but also the good from the great—the sweetest apple, the juiciest meat, and the freshest bread. Eating may be a risk, but it is also a pleasure, and our senses help us find the most desirable food and drink available.
However, what constitutes the best food and drink is often a matter of opinion. When attempting to generalize about what constitutes “good food,” counterexamples abound. For example, cheeses such as limburger are desirable to some but repellent to others. Thus, eating is a matter of taste, both in the biological sense and as a matter of individual opinion. Why there should be such diverse views about what constitutes the most desirable food is a worthwhile question and one that can be answered from different perspectives: cultural, social, developmental, and medicinal. People eat what others in their communities and families find desirable;1 children like different foods than adults do;2,3 and some people believe diets recommended by their doctors for health reasons are the best food choice.
In this chapter, we provide a genetic and evolutionary perspective on food perception and preference. Humans have changed over time and adapted to specific environments that contain some types of food but not others; this in turn has tailored our sense of taste and, by extension, our genome and individual genes. Nonhuman species provide evidence that the sense of taste has been shaped by evolution; for instance, cats and some other carnivorous species, in addition to chickens,4 have lost the function of their sweet receptor5—they no longer need to taste “sweet” because the foods they eat, the flesh of other animals or starchy grains, contain little sugar. The availability of food may have shaped nearly the entire genome—in yeast, and probably other organisms as well, most genes are involved either directly or indirectly with nutrition and metabolism.6 Some animals have specialized in eating only particular foods,benefiting from a refined ability to find and ingest those foods and limiting the number of competitors for that food source, whereas other animals, like humans, are generalists that can eat most types of food, which brings other benefits and risks.
At the heart of evolution is individual variation, and perhaps no single human trait has as many person-to-person differences as the abilities to taste and smell.7,8 However, to what extent these genetic differences in taste and smell affect the food preferences and food intake of contemporary humans is contentious. Information from these senses is one of many influences on decisions about what to eat,9 and its place on a hierarchy of determinants is unclear. Genes that affect energy metabolism rather than taste and smell might be more powerful determinants of food intake. For instance, the different feelings of satiety and hunger that people experience arise from genetic variation.10
Because human food intake will always be determined by many variables, which change in importance with time and circumstances, the controlled settings available with animal models are useful in untangling the relative contributions of taste and smell, metabolism, and experience.11 In addition, new methods—genome-wide association studies—have recently been developed to survey the contributions of all genes to taste, smell, and food intake. In the sections that follow, we review the genetics of taste and smell, as well as fat and sugar preference, drawing on data collected in humans and other animals, when useful studies are available. We also include results and interpretations of genome-wide studies of taste, smell, and nutrient preference and intake.
Before taste and smell can be studied, they must be measured. The field of science devoted to these measurements is called psychophysics. The scientists trained in this field try to understand relationships between physical stimuli (e.g., a tastant or odorant) and the psychological responses they elicit (e.g., taste or odor). Thus, an individual’s ability to taste or smell can be determined using psychophysical testing, such as measurement of detection threshold. Detection threshold is the lowest concentration at which a compound can be detected, and subjects often perceive this as only a hint of “something”—just enough to discriminate the stimulus from a blank, but not necessarily enough to recognize its type or quality (e.g., sweet). The lowest concentration at which a stimulus can be named for its quality is called the recognition threshold.
Detection threshold (often referred to as “threshold” and assumed to mean detection threshold) is a frequently used measure in studies on genotype–phenotype associations in taste and smell. The threshold can be measured inseveral ways. For instance, in odor tests, often the subject’s task is to find one bottle with the odorant among a set of three (one bottle with the stimulus and two with blanks), presented in ascending order of concentration. The measurement is regarded as relatively objective because it does not require use of subjective rating scales, in contrast to perceived intensity and pleasantness, which require the subject to offer a judgment. However, measures of perceived intensity are also important because, although biased by a subject’s experience,12 they provide information about the range of concentrations of odorants or tastants most often encountered in everyday life. Likewise, pleasantness is a matter of opinion, but it is also a crucial piece of information because liking is often a prerequisite for consumption.
Identification of stimuli is another common measure of taste and smell. Odor identification tests frequently include between 8 and 40 stimuli, each typically accompanied by four alternative descriptors including one that is meant to be the correct quality of the perceived odor. For instance, amyl acetate (a “fruity” smell) would be offered and the subjects asked to choose among these descriptors: banana, kerosene, burning rubber, and cinnamon. Commercially available odor identification tests, such as the UPSIT,13 BSIT,14 and Sniffin’ Sticks,15 are designed for clinical purposes, but because they can be administered quickly, they have also been used in epidemiological studies. Comparable tests for the identification of taste qualities are not widely available.
Some tastants and odorants activate not only the olfactory system, but also the free endings of the trigeminal nerve located in the mouth and nasal cavity, thus contributing to the common chemical sense. This may add complexity to the measurement of odor detection thresholds. For instance, an individual who could not detect an odorant even at a high concentration may still detect the odorant by common chemical sense.16
Some people are born with a total absence of the sense of smell (general anosmia, sometimes known only as anosmia) or taste (general ageusia), but these conditions are rare. More common is the inability to detect a specific odorant or tastant, known as a specific anosmia17 or specific ageusia, the most classic example being the inability to taste the bitterness of sulfur-containing thyroid poisons.18 However, unlike some people’s inability to taste, the inability to smell specific chemicals is more a matter of degree: Specific anosmia often refers not only to total lack of ability to detect a specific odorant but also to reduced sensitivity to the odorant.19 Subjects are often considered to have a specific anosmia if they have a detection threshold two standard deviations or more above the mean.20 The most studied specific anosmia is for the odorant androstenone.21,22 Among its other functions, androstenone is sometimes found in meat from uncastrated male pigs, and it underlies, together with skatole, the characteristic (some say unpleasant) “boar taint” odor.21
This section has described ways that taste and smell can be measured (i.e., detection and recognition thresholds, the intensity of the stimulus, its quality, and the criterion for defining specific deficits of taste and smell). When we eat, we attend to the concentration of chemical stimulus in our food (e.g., judging whether our food is too salty or contains a hint of onion). For taste, the qualities of bitter, sweet, sour, etc., determine how much we like our food. Now we consider each quality in turn.
Bitterness has the simplest relationship with food intake: What is bitter is bad, and what tastes bad is not eaten. Because poisons can kill quickly, their detection in food is paramount. And many poisons are bitter, a taste quality that evokes a classic rejection response.23,24 This rejection is assumed to be inborn and unlearned because it is apparent in human infants and in nonhuman primates.25 Furthermore, because it is also present in rodents that have had their neural connection between the brain stem and cortex severed, the rejection of bitterness could even be considered a reflex.26
Many people assume that all poisons are bitter, but this viewpoint depends upon the definition of poison. Toxicologists view all chemicals as potentially poisonous—the key issue is determining the relationship between dosage and lethality. Because every chemical is a potential poison but not every chemical is bitter, not all poisons are bitter. From the viewpoint of taste and food intake, a poison is defined as a chemical in a food that is liable to cause illness or death when eaten in sufficient quantity. Even with this narrow and unconventional definition of poison, it is not known how many chemicals are poisons and what proportion of them are bitter. However, when people are offered a range of chemicals to taste, they are overwhelmingly accurate at guessing the toxicity of given compounds using only taste as a guide.27
The following are common poisonous plants: castor beans contain ricin, a compound that causes red blood cells to clump together; turnips contain progoitrin, which inhibits thyroid hormones; cassava contains cyanide, which interrupts the ability of cells to make ATP; soybeans contain saponin, which is poorly absorbed into the body but when present in the bloodstream causes red cells to burst. All these chemicals—ricin, progoitrin, cyanide, and saponin—are bitter.28
We can see from these examples that many poisonous plant compounds are bitter and that the taste system developed in part to detect and avoid them. However, the relationship between the detection of bitterness of a chemical and its lethality is a puzzle, because some bitter chemicals that are not harmful to humans can nonetheless be perceived at low concentrations.29 So the ability to sense bitterness may serve other purposes in addition to poison detection. For instance, when proteins are fermented, some of the protein products are bitter,30 so the bitter taste system may also detect decayed proteins.
Typically, infants and primates immediately and automatically reject bitter stimuli. But for adult humans, the decision about what to do when bitterness is perceived is more complex. Adults sometimes eat foods and drink beverages that are bitter because they contain chemicals that increase feelings of wellbeing; the most obvious examples are the psychoactive drugs caffeine and alcohol. How much the person likes the effects of the bitter drugs, despite their taste, largely determines whether people ingest them.31 Even for bitter foods and drinks that offer pharmacological incentives, people often mask the bitterness, for instance adding cream and sugar to coffee. But if adults ingest bitter foods and drinks only when they contain drugs, then we must explain the willingness of people to drink decaffeinated coffee, which is still bitter but contains much less caffeine than regular coffee.32 It is possible that the overall sensory qualities of coffee become associated with the effect of caffeine33 and that even during extinction (i.e., when the stimulus is no longer followed by the rewarding response), the association is sufficient to maintain the behavior. Or it could be that the small amounts of caffeine are enough to maintain the consumption of this bitter beverage. But even if bitter substances are willingly ingested for their pharmacological benefit, we still need to explain why some people eat bitter melon (a plant commonly eaten in Asia) or other bitter plants that have no obvious drug-like properties.
This paradox—people eat bitter foods that contain no known psychoactive drug—might be resolved if the bitter compounds make people feel better in other ways. Recent studies suggest that bitter melon may contain secondary chemicals that have favorable metabolic effects, including reducing blood sugar in people with diabetes.34 Thus, bitter foods might contain healthful compounds that blur the line between nutrient and drug. If bitter-tasting chemicals in plant foods have health benefits, then removing these compounds (through manufacturing of processed foods or selectively breeding plants for low bitterness) may have negative consequences. The harmful effect of increased sugar and fat in the modern human diet has been widely discussed, but the loss of bitter compounds may also contribute to diseases associated with the modern diet, such as obesity and diabetes. Our bitter detection system seems to balance rejection and acceptance for bitterness in order to avoid poisons and to get enough—but not too much—of the bitter substances that make us feel good.
Tests often ask people to sample bitter chemicals dissolved in water, and because these chemicals must be safe to ingest (even so, subjects are usually asked not to swallow the samples), the number of bitter chemicals tested in a laboratory does not reflect the wide range of bitter compounds we could potentially taste. The selection of chemicals tested is further biased towardthose that have been previously used for sensory testing, so that data can be compared across studies and because their safe use already has been documented. Some of the frequently tested bitter compounds include quinine (found in the bark of the cinchona tree and used to treat malaria), caffeine (found in coffee beans and widely consumed for its stimulant properties), epicatechin (found in tea), tetralone (found in hops and, by extension, in beer), l-phenylalanine (an amino acid), magnesium sulfate (a mineral found in Epsom salts), urea (a product of nitrogen metabolism), naringin (a compound found in grapefruit), sucrose octaacetate (an acetylated derivative of sucrose), denatonium benzoate (used in consumer products to discourage accidental poisoning), and propylthiouracil (a sulfur-containing drug used to treat hyperthyroid disease). People exhibit marked differences in perceiving these chemicals18,35,36: Some find bitter compounds to be very bitter, whereas others experience the same concentration of the same chemical as much less intense.
We know that the origin of these individual differences, at least for most compounds listed above, is partially genetic because people with genetic makeups that are very similar (e.g., identical twins) are more alike in bitter perception than people who differ (e.g., fraternal twins).37,38 For the least lethal bitter chemicals, which are the most studied in humans, genetic variation is a moderate to strong determinant of how well a person can perceive them. For the most lethal poisons, less individual variation might be expected because people who have lost their ability to taste these chemicals might experience more accidental poisoning, so their genes would be less represented in the population. On the other hand, sensory variation in the worldwide population might be greatest for poisonous chemicals in plants that are found only in some geographic regions. But whether there is greater or lesser individual variation in the perception of lethal bitter chemicals has gone unanswered—ethical concerns obviously prevent testing with these poisons in people. Cell-based assays with one or two human bitter receptors can be used to test the response to a wide range of poisons,39 but this method provides only a partial answer to the question because artificial systems may not recreate the human taste experience.
In at least one case, a gene’s participation in bitter perception is well understood. The inability of some people to taste phenylthiocarbamide (PTC) was discovered in the 1930s by a DuPont chemist named Arthur Fox.18 It was soon determined that the trait was heritable (i.e., transmitted in families),40 and 70 years later the responsible gene and allele were identified.41 The gene, called TAS2R38, is a member of the bitter taste receptor family TAS2R. Three alleles in TAS2R38 account for the bitter-blindness to PTC—they combine to form a haplotype that leads to reduced ability to perceive PTC (and its chemical relative propylthiouracil, one of the commonly studiedbitters listed above). The TAS2R38 haplotype determines most of the variation in people, but alleles in other genes,42,43 and even age44 and sex,45 also contribute to variations in PTC perception. The study of the genetics of this trait is useful because it straddles the divide between the single-gene mode of inheritance found in diseases such as cystic fibrosis and the interactions of many genes found in a complex trait like obesity. Thus, PTC genetics is a useful model for studying genotype/phenotype effects and the influences that modify them.
Genetic differences in bitter taste perception might modify food preferences and intake in a complex manner. Although PTC was first created in a chemistry laboratory and is probably not found in plants, there are many chemical relatives of PTC that stimulate the TAS2R38 bitter taste receptor.39,46,47 At least one of these compounds is found in plant food (turnips),48 and less similar but still related compounds are found in other plant species.49 People with taster and nontaster alleles of TAS2R38 differ in their perception of vegetables (like watercress) that contain these PTC-like compounds.50 From here, it is a short step to hypothesize that genetically insensitive people would eat more of these vegetables than would people who find them to be bitter.
If people differ in their intake of some vegetables, bitter perception might ultimately influence body weight, as suggested by some investigators.51-54 However, a direct relationship between TAS2R38 genotype, food intake, and body weight has not been detected in epidemiological studies55,56 or in genome-wide studies of association with body mass index, a measure of obesity.10,57,58 Thus, if alleles of this bitter receptor gene can directly affect food intake or body weight, they are too weak to be detected in the population as a group. Progress toward understanding genotype/phenotype relationships for PTC taste-blindness and food intake will require narrowing the focus to vegetables that contain these specific compounds. In addition, instead of relying on indirect information about the chemical constitution of vegetables, concentrations of these bitter chemicals should be directly measured in vegetables, because amounts can differ depending on which cultivar is tested or the composition of the soil in which it was grown.
A related point to consider is which aspect of receptor function is most affected by alleles of TAS2R38. While it is often referred to as a “bitter taste receptor,” this receptor and other bitter receptors are also found in the gut59-61 and in nasal airways, where they detect molecules secreted by bacteria.62 The expression of the gene TAS2R38 in the gut is regulated by the amount of cholesterol in the diet, and its expression is highest when cholesterol is low.63 The interpretation of this observation is that gene expression of bitter receptors is increased when plant foods are consumed, which is logical because bitter compounds are more concentrated in plants than in other foods, like meats. It is therefore reasonable to assume that bitter taste receptors are intimately related to vegetable intake, because vegetables taste bitter and gene expressionin the gut is tied to the intake of diets high in plant food. However, the regulation of bitter receptors in the tongue (or the gut) in response to changes in diet had not yet been studied. This is a gap in our knowledge.
A convincing argument can be made that a specific bitter receptor and its alleles might affect food intake, especially of vegetables. But it is important to put these details in context. Although people differ in their ability to taste many bitter chemicals,36 the complete loss of bitter perception for a particular chemical like PTC is probably rare. (It might be misleading to call this a complete loss because the nontaster form of the receptor might detect different bitter molecules.64) Current studies suggest that PTC is unusual because only one of the 25 known bitter receptors is strongly stimulated by it, so the loss of this one receptor (TAS2R38) is consequential.39 Other bitter molecules stimulate multiple receptors, and the loss of one may decrease but not eliminate the ability to detect that particular bitter molecule.30,39,47,65-73 The perception of PTC is probably an extreme case of individual variation in bitter perception.
Although the relationship between bitter taste and plant poisons is relatively simple (compared to other taste qualities), it is not the only one that signals a warning. Bacteria and fermentation can spoil food, and the end points of these processes are detected by using sour taste as a guide, along with smell, vision, and the common chemical sense. Sometimes bacterial activity in food is wholesome, such as the fermentation of milk, wheat, or grapes to make cheese, bread, or wine. But sometimes it is not, as when meat or vegetables rot. Like bitterness, which can signal either a poison or a beneficial compound, sourness can signal either good or bad bacterial processes. And like bitterness, the preference for sourness is a matter of degree; low concentrations of sourness (and bitterness) must be evaluated in a specific circumstance and a decision made about acceptance or rejection. Context is important: The tartness that is desirable in buttermilk (caused by lactic acid, a by-product of the fermentation process) would be undesirable in ordinary milk. Concentration is also important because there is a continuum, from lower concentrations (which have a pleasant taste) to middle-range concentrations (which may be rejected), to high concentrations (which evoke pain receptors and lead to tissue damage). In other words, we like lemonade, but we don’t drink more concentrated acids. From a developmental perspective, bitter and sour differ because sour taste is readily accepted by many children but bitter taste is not.74 So the sourness of food conveys mixed signals: very bad at very high concentrations, bad in some foods but good in others, and neither universally liked nor rejected by all.
Also like bitter perception, the ability of humans to detect sourness at low concentrations is partially determined by genetics.75,76 But unlike bitter perception, the genes and their alleles associated with individual differences in sour perception are unknown, because studies to identify them have not been conducted. Furthermore, unlike for bitter taste, studies of sour taste in animals are no more advanced than those in humans. Although genetically distinct strains of inbred mice are known to differ in sour preference,77 no genes have been identified that account for these differences. One candidate is the polycystic kidney disease 2-like 1 gene (Pkd2l1), which is involved in sour taste in mice78 and possibly in humans.79 Whether naturally occurring allelic variation in this gene is responsible for differential sour perception (in mice or humans) is not known.
The studies needed are of two types: (1) linkage studies, in which the DNA of family members with a similar trait is searched for shared DNA that would contain the causal genes, and (2) genome-wide association studies, in which subjects are grouped by genotype at many loci throughout the genome and compared for a particular trait, such as intensity of sour perception. Useful future studies would measure individual abilities to: (1) perceive sourness at low concentrations, and (2) judge its intensity at a range of concentrations. It would also be useful to measure how much people like sourness. Genome-wide studies could be performed to find regions of the genome in common among people with similar sour phenotypes. These studies would fill gaps in our current knowledge.
One of the pleasures of eating comes from sweet taste, but perception of sweetness and the liking for highly concentrated solutions differ among people (for reviews see Refs. 80-84). Being more or less able to perceive the sweetness of sugar will interact with a person’s liking for it. Some people are sweetlikers—no concentration of sweet in food is too much85,86—and their enhanced ability to perceive sweetness makes them like the food all the more. Other people have a peak concentration of sweetness they prefer, and as sweetness rises above that point, food becomes too sweet and unpleasant. While this dichotomy (i.e., sweet-liker or disliker) is a useful concept, like most human traits, sweet liking is probably on a continuum and is context-specific87 (e.g., people who like very sweet ice cream might not prefer very sweet juice). Therefore, because people differ in how much they like sweetness at a given concentration, having a genotype that makes some people more sensitive to sweetness than others will not always result in increased liking. Sweet as a taste quality is complex (from a genetics perspective) because the relationship between perception and liking is complex.
In mice, alleles of the Tas1r3 gene,75 which codes for one of the subunits of the sweet receptor (the receptor is a heterodimer, or combination, of Tas1r2 and Tas1r3), determine in part both the sensitivity to and preference for sweet solutions.88 Alleles in the promoter region of this same gene predict how well people can sort a range of sucrose concentrations into the correct order,89 but it is not known whether these sensitivity alleles are related to preference. Other genes and their alleles probably also contribute to genetic differences in sweet perception (e.g., second messenger molecules like gustducin). We might expect poor agreement between genes that affect perception and those that predict the preference for and intake of sweet food,90-92 and in fact, this is the case. Thus far, there is no convergence on particular genomic regions associated with sweet sensory perception and the actual intake of sweet foods.
Until recently, there was no consensus about whether umami was a true taste quality.93 The concept of umami, which perhaps translates best into English as “savory” or “meaty,” was suggested by Japanese investigators as a unique quality exemplified by monosodium glutamate (MSG). Umami also has a synergistic property: When MSG is combined with ribonucleotides such as inosine monophosphate (compounds often found in meat), the perceived intensity of the mixture is higher than the intensity of either compound alone. Umami was better accepted as a taste quality when its receptor was discovered in taste cells.94-96 Like the sweet receptor, which is a heterodimer of TAS1R2 and TAS1R3, the umami receptor is a heterodimer of TAS1R1 and TAS1R3. Some people are specifically insensitive to MSG,97 which is partly caused by alleles of the umami receptor.98-100 The detection of a genotype/phenotype relationship implies that the trait is at least partially heritable, but we know of no published twin or family studies that estimate the contribution of genes to trait variation.
Since humans differ in their ability to taste MSG, it sparks curiosity about what role this differential response might play in the liking for meat or other foods such as cheese, tomatoes, and mushrooms, which contain glutamate. However, the role that individual differences in umami perception might play in human food intake is unknown and represents another gap in our understanding. That the umami receptor might be a key biological protein in determining meat-eating is speculative, but the idea does have empirical support from comparative studies. The giant panda, which eats only plant food, has lost the umami receptor during evolution.101 The recent observation that obese women prefer higher concentrations of MSG in soup suggests that this taste quality may be of importance in determining food intake and body weight.102
Salt is both a simple pleasure and a complex poison. It is a pleasure in that humans choose to consume more salt than they need and it is added to food to enhance flavor in almost every culture.103 It is a poison in the sense that it may increase blood pressure and exacerbate other health problems. But whether salt reduction should be a universal mandate is a debated public health position.104 The Institute of Medicine, a health policy advisory group, recently drafted a report calling for reduced salt consumption.105
While the current viewpoint from a biomedical perspective is that nearly everyone overconsumes salt relative to physiological need, there are few studies that concentrate on individual differences in salt perception (e.g., Refs.7,106,107). Even fewer studies have asked whether there are heritable genetic contributions to variability in salt taste perception75,108 or preference.109 Nevertheless, the few results available are consistent: There is no evidence for genetic effects on salt perception or liking. Instead, environment seems to be the major determinant. One’s history of sodium exposure can have a substantial impact on preference for, consumption of, and physiological processing of NaCl.110 Research suggests that time of day111 or even short-term exposure can have some (temporary) impact on salty taste.112 Evolutionary forces may have shaped the human ability to recognize salty taste in such a way as to make it very responsive to differences in the environmental mineral and water supply or habitual diet.113 Therefore, efforts to assess the impact of genetic variation within the salt receptor114,115 should focus on salt perception of people with similar environmental backgrounds (e.g., early exposure, recent exposure) and be attentive to the current state of the subject (e.g., time of day, thirst).
The molecular aspects of human salt perception are not known, but evidence has accumulated that a sodium channel is important for one component of salt perception in mice.114 The genes that code for the protein subunits for this channel would be a candidate target for genotype–phenotype studies in humans.
One of the most widely debated aspects of human taste is its definition: Most schoolchildren learn that there are four basic tastes, sour, sweet, salty, and bitter. But we now turn to evidence that the list of taste qualities is expanding. One reason for this expansion is that the definition of a basic taste is changing: If there is a working receptor on taste cells, its ligand can be considered to have a “taste.” Umami was a controversial taste until its receptor was discovered, and a similar change has occurred with the mineral calcium. A taste receptor sensitive to calcium (Tas1r3) has been identified in mice,116 and it is possible that the same receptor acts in humans.117,118 In this regard, it has been suggested that another receptor, CASR, mediates kokumi taste,117 an orosensory quality recognized in Japan but unknown in the West. Although people differ in the perception of calcium solutions,119 there are no genetic studies to indicate whether this trait is heritable and which genes (including the genes coding the subunits of the receptor) might be involved. It may be that genotype has a potent effect on calcium perception as it does in mice,120 or it may be that individual differences in calcium perception are tied to the current diet or metabolic need, similar to sodium and salt (as described in the above section), or both could be true. This is an understudied area.
While the controversy of umami as a basic taste quality is largely resolved, the controversy about fat as a basic taste quality is approaching but has not attained resolution. The idea that fat is a basic taste was suggested as early as the 16th century, at which time it was called pinguis (Latin for fatty).121 The evidence has recently been reviewed,122 as has the heritable aspects of fat perception, liking, and intake.123 Fat as a taste quality is especially relevant to obesity because of the observation that obese people typically have greater fat preferences than do lean people (reviewed in Ref. 80). Thus far, some genes have been implicated in fat perception, specifically a gene coding a transmembrane protein found in taste cells (Cd36)124,125 and genes coding several G protein-coupled receptors that respond to fatty acids elsewhere in the body and that are also found in taste cells.126-128 To date, these studies have been conducted in mice, and it is unclear whether human fat perception occurs through the same mechanism and, if so, whether alleles of these genes might lead some people to be fat-blind in the same way some people are bitter-blind. Thus, as with sour taste, the role of sensory differences in fat perception is another gap in knowledge that can be filled once the influential genes and alleles have been identified.
Perhaps because the common chemical sense has not had a single name in widespread use, it has been poorly integrated into the study of food intake and obesity, but it contributes in several ways to the pain and pleasure of eating: The tingle of carbon dioxide dissolved in soda,129 the cooling associated with methanol,130 and the burn of chili peppers131 all arise fromreceptors in the mouth (and nose and throat) that convey this information to the brain. These compounds exist primarily in plants and are defenses against insects that would do harm. At low intensity, these defense compounds produce sensations that many people find pleasant; at higher concentrations, the compounds produce sensations that are unpleasant and even painful.
People differ in their perception of these sensations,132 but little systemic research has focused on whether the differences are heritable in humans. However, one study has reported that the liking for spicy food was highly heritable.133 From a comparative perspective, birds are indifferent to the main ingredient in hot peppers (capsaicin) that causes the burn, because they lack the receptor TRPV1.134 Everyday experience suggests some humans, too, may be indifferent to, or even like, the burn of hot peppers. The prevalence of individual differences in the many facets of the common chemical sense (cooling, burning, stinging) in humans, the degree to which genetic variation explains those differences, and their impact on food intake are unknown. But because most of the compounds that stimulate this sense are found in plants,135 and humans eat plants, sometimes as their only source of food, their effects on human health are probably direct. It is possible that these sensations have effects that are equal to or even greater than those of bitterness in determining individual differences in the liking of vegetables, spices, and condiments such as mustard or chili sauce.
Carbon dioxide also stimulates the common chemical sense and is a constituent of the modern human diet. It is commonly consumed in fizzy soda, but its taste perception may have evolved originally to detect the carbon dioxide produced from rotting food.136 How carbon dioxide might affect food digestion and metabolism is unknown. For instance, the obesity effects of sugar in soda are often studied (e.g., Ref. 137). Whether these effects are exacerbated or offset by the fizz of carbon dioxide has not yet been examined.
If taste is the gatekeeper, the sense of smell is the sentinel, evaluating the food for danger before it enters the mouth. When offered an unfamiliar food, we will smell it before we taste it, and smell is one of the key first defenses against spoiled food and an important source of eating enjoyment. Thus, the sense of smell and its loss can have powerful consequences for food intake and quality of life.138,139 Before addressing genetics of smell and its potential connections to food intake, we introduce the olfactory system for background information.
When working normally, the sense of smell, or olfaction, enables us to detect a large number of different odorants and to perceive these volatile compounds as odors. Although the stimuli (odorants) are also sometimes called odors, in psychophysics, odor refers to a percept, the result of the process of odor perception, whereas odorant refers to the chemical that elicited the odor.
The airborne molecules from food take one of two paths to sensory cells in the olfactory epithelium: the orthonasal route (through the nostrils, before eating) or the retronasal route (through the nasopharynx, while eating). Both paths are important in food intake, for defense and for pleasure. In the olfactory epithelium, the airborne odorants are detected by olfactory receptors. The receptors lie embedded in the membrane of olfactory sensory neurons, each of which accommodates only one type of receptor.140 Binding of an odorant molecule by an olfactory receptor initiates a signal transduction cascade, which ultimately leads to the transfer of the olfactory signal to the brain, where the odor percept is generated.
To be a potential odorant, a molecule has to be volatile enough to reach the olfactory epithelium with airflow. Although most volatiles are odorants, some small molecule volatiles, such as carbon monoxide and carbon dioxide, are odorless. In addition, structurally diverse molecules can elicit indistinguishable odors, while similar molecules, such as stereoisomers, can yield distinct odors (e.g., R(–)-carvone smells like spearmint, but S()-carvone smells like caraway).141,142 To date, the type of odor elicited by a volatile compound cannot be reliably predicted by the structure of the molecule.143
Humans have about 400 different olfactory receptor types, a number greatly exceeded by the number of potential odorants. Thus, it is unlikely that a particular receptor would bind only one type of odorant or that a certain odorant would attach specifically to only one type of receptor. Instead, the olfactory system is thought to make use of combinatorial receptor coding to gain the capacity to recognize the immense amount of odorants; several types of related receptors bind an odorant with varying affinities, and in turn, multiple related odorants can be detected by the same receptor.140 The combinatorial coding suggests that most olfactory receptors are selective (broadly tuned) rather than very specific (narrowly tuned). However, the breadth of tuning varies among olfactory receptors.144,145
Only a few of the human olfactory receptors have been linked with their odorant ligands (i.e., the molecules that they detect).146 Development of automated, high-throughput methods for matching the receptors and their ligands in cell-based model systems (or using computational models) will facilitate confirming the functionality of the receptors. These methods, however, cannot replace the measurement of actual human responses in studies of geneticinfluence on odor percepts. The psychophysical measurement of responses to odorous stimuli remains a time-consuming but essential step when the genetics of the sense of smell and its implications for food intake are studied.
Humans have nearly 400 potentially functional olfactory receptor genes (OR genes), making this gene family one of the largest in the human genome.147 In addition to these intact genes, which are thought to produce functional olfactory receptors, humans have at least a similar number of nonfunctional OR genes (pseudogenes) and about 60 genes of which both functional and nonfunctional variants are known to exist (segregating pseudogenes). The exact number of functional genes will be known only after the functionality of the corresponding receptors is demonstrated. However, it is obvious that far more genes encode receptors for smell than for taste. The larger number of olfactory receptors likely reflects the need to detect a wider variety of compounds than is the case for taste. Further, the large number of compounds detected by the sense of smell reflects the wider role of this sense: While the sense of taste serves almost exclusively ingestion, the sense of smell has other functions, too. These include sensing environmental dangers (e.g., smoke) and potential interpersonal chemosignaling (e.g., sexual selection).
The heritability of a trait makes the search for genes influencing that trait reasonable. If little or no heritability is found, the underlying genes, if any, are difficult, if not impossible, to locate in gene-mapping studies. While the ability to smell some odorants is heritable, for other odorants, it is not. For instance, the ability to smell food odors like chocolate or lemon is associated with little or no heritability.148,149 However, the pleasantness of cinnamon is heritable and has been mapped to chromosome 4 by linkage analysis.150 If the allelic genes that determine the pleasantness of odors like cinnamon are identified, studies of genotype and food intake might be worthwhile.
Individual variation in perception of some odors has been attributed partly to specific OR genes. The differences among people in the ability to smell androstenone are at least partially determined by genes,151,152 and an allele of an OR gene, OR7D4, contributes to this trait.153 However, unlike alleles of the taste receptor gene TAS2R38, which account for almost 70% of the person-to-person variation in perception of bitter taste from PTC,41 OR7D4 alleles account for only a small amount of variance in perception of androstenone.153 Two other OR genes have been associated with individual variation in the sense of smell: OR11H7 with isovaleric acid (sweaty odor)154 and OR2M7 with the smell of asparagus metabolites in urine.155 Association between the gene OR2J3 and detection of cis-3-hexen-1-ol (green leaf odor) has also been suggested.156 Why there is relatively little effect of the alleles of a singleolfactory receptor on perception lies in this sense’s complex nature: Many olfactory receptors combine to detect a particular odorant,140 and one odorant may stimulate many receptors, so if one is not working, others may compensate.
Systematic, repeated exposures to individual odorants have been demonstrated to lower detection thresholds (increased sensitivity) to these odorants, suggesting that genes do not entirely determine the perceptions.157,158 One possibility, yet to be proved, is that there are gene-environment interactions in odor perception (i.e., genes influencing the sense of smell are controlled differently in different environments). Whatever the mechanism, the flexibility of the sense of smell could have been evolutionarily appropriate. When first humans moved to new environments and encountered novel odorants from new threats (e.g., toxins) and opportunities (e.g., food sources), the flexible sense of smell may have helped the population to survive.
Although there may also be some innate preferences, smell is probably more flexible and amenable to learning159 when compared to taste. This point is particularly relevant when we consider olfaction as a sentry against spoiled food: The products of fermentation can be perceived as wholesome or harmful, depending on context. As an example, isovaleric acid has a pungent odor that people like if they are told it is from cheddar cheese and dislike if they are told it is from body odor.160 Likewise, people will eat food with a bad smell (e.g., durians or limburger cheese) if they know it is safe and they like the taste. In addition, the pleasant odor of food can stimulate appetite, but the potency of these genetic differences in determining food intake and obesity is unclear.
Ethanol (or in more colloquial terms, alcohol) is a commonly consumed drug that is also a food, and just over 50% of adults living in the United States are regular drinkers.161 Because of alcohol’s popular pharmacological effects, the attractiveness or off-putting taste and smell of alcohol can be overlooked.162 As a taste, ethanol has a complex quality: Indirect evidence, mostly from the study of mice and rats, suggests that it stimulates the sweet receptor.163 One explanation for the connection between sweetness and alcohol is that sweet fruits ferment and so this sweetness may help animals gauge the sugar/alcohol ratio in fruit and other fermented products.164 In addition to sweetness, genotype–phenotype studies in humans suggest that ethanol stimulates at least one bitter receptor.165 Alcohol may also stimulate receptors for the common chemical sense, at least in rodents.166 Alcohol also has an odor, andalthough the exact receptors are not known, based on other typical molecules, it is likely to stimulate several different receptors; the patterns of receptor activation may differ based on concentration.144
Individual differences in alcohol intake are studied intensively because of the role of dependence and addiction in human health, yet we are aware of no studies that have examined the heritability of alcohol perception in humans. Although it is reasonable to expect large individual differences that may be due in part to genotype, this is a current gap in scientific understanding.
Taste is one reason people report for why they eat the foods they do, but cost, social influences, and food availability all play a role in human food intake.167-169 What constitutes good food is subtle, but extremes of most taste sensations, including bitterness, sourness, and sting, as well as excessive sweetness, saltiness, and richness from fats, detract from the pleasant experience of food for most people. However, humans can tolerate and even like foods that go too far; for instance, we deliberately add tingling wintergreen oil to candies and drink carbonated sodas that can cause a burning sensation. Or we drink (and even prefer) very bitter coffee. The observation that tingling, burning, and bitterness are so popular deserves more research attention than it receives. The liking for sweetness and fat depends on concentration: For some people, there is no such thing as “too sweet,” while others find more than moderate amounts of sweetness to be cloying. And although few people would eat a meal solely of oil or butter, people differ in how much fat is just right or too rich.123
Some progress has been made in defining the genes and their alleles associated with the positive and negative aspects of food and flavor. Taking a ham and cheese sandwich as an example (Fig. 1), we might imagine that people with sensitive alleles might differentially detect the mild sweetness of onion (TAS1R3),89 the savory glutamate taste of tomato (TAS1R3),98,100,170 the bitterness of watercress (TAS2R38),50 the smell of cheese (OR11H7),154 or the boar taint odor of ham (OR7D4).153 We envision that a combination of allelic differences might contribute to the range of liking for this sandwich. People who can taste the pleasant components (and not the unpleasant ones) may experience the ham sandwich as more desirable because of its taste.
But how do differences in sensory experience translate to actual food consumption? Whether these individual differences in chemosensory experience affect food selection is the weak link in the chain of causality. People eat what they like, but they also eat for many other reasons. Simple explanations of the links between sensory perception and food intake are misguided: Just as people do not choose art or music based solely on how well they can hear orsee, we do not choose food based solely on the reactions of the tongue or nose. Although genetic differences determine what we can taste and smell (and at what concentration), our taste is ultimately determined by our experiences, learning, and culture, in an artistic sense, as well as in our likes and dislikes of food and drink. However, perception is the first step toward liking: What cannot be perceived cannot be liked or preferred. Therefore, it is worthwhile to pursue these questions.
This focus on perception and taste is especially important in the realm of human health because most of the chemicals discussed that give rise to bitter taste have metabolic and behavior effects and many are drugs (caffeine, alcohol). People are always urged to eat diets higher in plant foods like vegetables, but these foods are bitter to many. As another example, new medicines that need to be given in liquid forms can taste excessively bitter.171 And some bitter or stinging compounds are concentrated in plants to help them to fend off insects, but they also tickle our taste buds. Thus, to understand our greater desires for certain types of foods above others, as well as our avoidance of compounds we know we should consume, such as medicines or healthy but bitter vegetables, we must consider our genotype, which dictates our ability to perceive these compounds.
Michael G. Tordoff and Gary K. Beauchamp commented on earlier versions of the chapter. Discussions with Julie A. Mennella, Charles J. Wysocki, Johannes Reisert, and Alexander A. Bachmanov improved the quality of this work. The editorial assistance of Patricia J. Watson is gratefully acknowledged. Mary Leonard provided assistance with graphic preparation. This work was funded in part by the National Institutes of Health grant DK56797.