Lead Review Article June 2001: 163-169

The Science and Complexity of Bitter Taste

Adam Drewnowski, Ph.D.

Food choices and eating habits are largely influ- (caffeine), sulfimides (saccharin), and organic and inor-
enced by how foods taste. Withoutbeing the domi- ganic s a k 3 The fact that such structurally diverse com-
nant taste sensation, bitter taste contributes to the pounds can elicit a single bitter taste suggests that mul-
complexity and enjoyment of beverages and foods. tiple mechanisms are responsible for the perception and
Compounds that are perceived as bitter do not transduction of b i t t e r n e ~ s . Some
~ , ~ of these mechanisms
share a similar chemical structure. In addition to may be common to the perception of both bitter and sweet.
peptides and salts, bitter compounds in foods may Small changes in chemical structure can convert bitter
include plant-derived phenols and polyphenols, compounds to intensely sweet or vice versa. Bitter and
flavonoids, catechins, and caffeine. Recent stud- sweet tastes in solution can enhance or suppress each
ies have shown that humans possess a multitude other, with the interplay between bitter and sweet occur-
of bitter taste receptors and that the transduction ring at the neuronal
of bitter taste may differ between one compound
The ability to perceive some bitter tastes varies greatly
and another. Studies of mixture interactions sug-
across individuals. In some cases, it can be an inherited
gest further that bitter compounds suppress or
trait.7 The only recorded instance of taste polymorphism
enhance sweet and sour tastes and interact with
in humans is the genetic “blindness” to the bitter taste of
volatile flavor molecules. Caffeine, a natural ingre-
dient of tea, coffee, and chocolate, has a unique phenylthiocarbamide (PTC) and 6-n-propylthiouracil
flavor profile. Used as a flavoring agent, it enhances (PROP). Sensory studies have linked the ability to taste
the sensory appeal of beverages. Research de- PTC/PROP, a dominant trait, with heightened sensitivity
velopments on the genetics and perception of bit- to such bitter compounds as caffeine*.9and naringin.”.”
ter taste add to our understanding of the role of Whereas the phenotypic taste responses to PTUPROP
bitterness in relation to food preference. are well studied, the gene responsible for this trait has not
been described and its exact location is ~ n k n o w n . ’ ~ . ’ ~
The focus of human taste research now includes bit-
Introduction ter as well as sweet tastes. So far, studies of taste genetics
in humans have only explored bitter The most
Taste is the main influence on food choices.’ Generally, recent studies on candidate taste receptors in humans
people like sweet and dislike bitter tastes, yet not all bitter and mice also involve the perception of bitter.l4.I5Studies
tastes are unpleasant to the consumer. In certain foods, a on mixture psychophysics have likewise focused on in-
limited degree of bitterness is expected and e n j ~ y e d . ~ . ~ teractions of bitter with sour and salty tastes.I6.l7Interac-
Without being the dominant sensation, bitterness helps tions of bitter with food flavor components have long
to balance the flavor profile of beverages and foods. Mix- been an interest of the food i n d ~ s t r yThe
. ~ contribution of
ture interactions among sweet, bitter, and sour tastes, and these research developments to a better understanding of
between taste and volatile flavor elements, add to the com- the role of bitterness in beverages and foods is the main
plexity and to the enjoyment oftea, coffee, chocolate, fruit topic of this review.
juices, and other beverages.
Of the four basic tastes-sweet, sour, salty, and bit- Mechanisms of Bitter Taste Perception
ter-bitter is the most complex and perhaps the least un-
derstood. Among bitter compounds in foods are amino Taste transduction begins when a stimulus comes into
acids and peptides, esters and lactones, phenols and contact with a taste receptor cell. Whereas each taste bud
polyphenols, flavonoids and terpenes, methylxanthines may contain 75-1 50 epithelial cells, only a few are exposed
at the taste pore at any one time.4 Most taste buds are
Dr. Drewnowski is Professor of Epidemiology and clustered in fungiform, foliate, and circumvallate papillae
Medicine, and Director of the Nutritional Sciences on the tongue surface, though some can also be found on
Program, University of Washington, Seattle, WA 98195- the soft palate, epiglottis, or even pharynx. As a result,
3410. USA. taste sensation can differ with the proportion of taste buds

Nutrition Reviews@,Vol. 59, No. 6 163

 that are stimulated, depending on whether the stimulus is PROP with a chromosome locus at 5p 15, with a modifier
swallowed or not.16 locus on human chromosome 7.I2-I3Thought to be a domi-
Taste stimuli influence receptor cells in a variety of nant trait, PROP tasting is shown by 70% of Caucasians.'
ways. Direct interaction of taste stimuli with ion channels The proportion of tasters among Asians and African Ameri-
on the cell membrane is most important for the perception cans is estimated to be 90% or higher. Recent studies also
of salty and sour. By contrast, the perception of sweet identified a separate subgroup of extremely sensitive PROP
and bitter involves specialized taste receptors coupled supert taster^."^ Supertasters, most of whom are women,
with protein^.^.'^ Such proteins exert their effects through tend to have more fungiform papillae and a higher density
second messengers, cyclic AMP (CAMP)or inositol triph- of taste buds per papilla. Opinions differ as to whether
osphate (IP3), that act on targets within the cell.19Accord- PROP supertasters are also more sensitive to sweetness,
ing to current thinking, there are at least two-ifnot more- salt, and to the oral sensation of fat.'.'
transduction mechanisms for bitter taste. One pathway Strains of mice are differentially sensitive to bitter
may involve a G-protein that stimulates enzyme-activated tastes, suggesting that multiple genes may be i n ~ o l v e d . ' ~ . ' ~
IP3, leading to a release of calcium from cell stores. An- Studies in humans have linked PTCPROP tasting with
other mechanism for the perception of bitter and sweet heightened sensitivity to such bitter compounds as caf-
tastes may involve G-protein, a-gustducin, that activates feine, saccharin, and quinine hydrochloride, but not
enzyme phosphodiesterase to decrease intracellular urea.7*11.12Taste detection thresholds for PTC and caffeine
cAMP.*O Direct blocking of K' channels by bitter tastants were correlated in some early studies.* PROP tasters were
may also O C C U ~ . ~ also more sensitive to the aftertaste of caffeine measured
Bitter taste perception may involve not only multiple for up to 4 minutes in a time-intensity study.22More recent
transduction mechanisms, but also a large number of re- work on the relationship between taste responsiveness to
ceptors. The number of different bitter taste receptors in PROP and the perceived bitterness of caffeine9 showed
humans that are linked to gustducin is estimated at 4&80, that most PROP tasters also gave high bitterness ratings
far more than previously thought.14 These candidate taste to caffeine solutions. By contrast, caffeine-insensitive re-
receptors (T2Rs) are organized in the genome in clusters spondents were more likely to be PROP nontasters. These
and are genetically linked to loci that influence bitter per- data, summarized in Figure 1 , are consistent with past re-
ception in humans and mice.14The T2Rs are expressed in ports that the PROP gene may confer both a specific abil-
all taste buds of circumvallate and foliate papillae, and in ity to taste PROP and a more general sensitivity to other
palate taste buds.14 Whereas T2Rs are rarely expressed in bitter tastes, including caffeine.
fungiform papillae, those fungiform taste buds that do Whether PROP tasting is associated with altered sen-
express T2Rs have a full repertoire of different receptors, sitivity to sweet is debatable. In some studies, PTC/PROP
suggesting that each cell may recognize multiple bitter tasters gave higher intensity ratings to dilute solutions of
tastants. A follow-up study showed these receptors to be sucrose, saccharin, and neohesperidin dihydro~halcone.~
highly specialized. A human bitter taste receptor (hT2R-4)
responded only to denatonium and PROP, whereas a mouse
receptor (MT2R-5) responded only to cy~loheximide.'~ The
perception and transduction of bitter tastes is a complex
and specialized system.
The perception of bitter and sweet tastes may share
some common pathway^.^,^^ Small structural changes con-
vert some compounds from bitter to sweet. Neohesperidin,
a bitter flavonoid, converts to neohesperidin dihydro-
chalcone (DC), an intense sweetener. By contrast, sucrose
v) A A
esters such as sucrose octaacetate are intensely bitter. v)
Saccharin and many other intense sweeteners have a bit-
ter aftertaste, especially at high doses.21This array oftaste
receptors and transduction mechanisms suggests that the
responses to sweet and bitter tastes may have a major role 3
in food selection.
Bitterness ratings - PROP
Genetics of Bitter Taste Perception
Figure 1. Summed bitterness ratings for seven PROP solutions
Behavioral genetic studies of taste perception in humans of increasing concentrations plotted against summed bitterness
have focused on two bitter compounds, PTC and PROP. ratings for seven caffeine solutions, by PROP taster status (.
Genetic linkage studies have linked the ability to taste tasters; A nontasters). From reference 9, with permission.

164 Nutrition Reviews", Vol. 59, NO. 6

However, other studies failed to replicate those finding^.^." can enhance weak intrinsic flavors or modify an existing
There was no evidence for the notion that PROP flavor p r ~ f i l e .For
~ ~ example,
.~~ certain intense sweeteners
supertasters were highly sensitive to the oral sensation of enhance each other at weak but above-threshold concen-
dairy fat.23Contrary to suggestions that PROP tasting trations.26 The degree of such enhancement can depend
might entail aversions to sweet beverages, no differences on the particular sweetener and its concentration. Above-
in soft drink preferences by PROP taster status were ob- threshold levels of caffeine enhanced sourness but masked
served. Furthermore, sweetening caffeine solutions with the perception of sweetness in water solutions.27
neohesperidin DC obliterated any differences in prefer- Moderate to strong concentrations are more likely to
ences for caffeine solutions between PROP tasters and show mixture-suppression effects.16.17Each component is
nontasters. perceived as less intense than when it is tasted separately.
For example, the bitterness of caffeine can be suppressed
Bitter Taste and Aging by sugar, acid, or salt. Natural sweeteners (sucrose) sup-
pressed the bitterness of caffeine more effectively than
The ability to detect very low concentrations of bitter and
aspartame or s a ~ c h a r i n .Caffeine
~ ~ . ~ ~ and acids could en-
salty tastes declines with age. By contrast, the perception
hance or suppress each other depending on concentra-
of sweet and sour remains relatively stable. Not all bitter
tion. At subthreshold levels, weak sourness of citric acid
compounds are equally affected. Whereas the sensitivity
was suppressed by caffeine.30Above threshold, the sour-
to PTCPROP, quinine, and caffeine declined with age,
ness of citric acid was enhanced by caffeine bitterne~s.~'
sensitivity to urea did not. In judging more concentrated
At high concentrations, mixture interactions could go ei-
taste solutions, elderly subjects found bitter, but not the
ther way. In one study, strong citric acid enhanced the
other three tastes, to be less intense than did young sub-
bitterness of a wide range of caffeine concentration^.^' In
jects.
related studies on sweet and sour, M ~ B r i d e ~found
~ . ~ )that
Age-related deficits in taste were most pronounced
sweetness suppressed acidity, whereas acidity suppressed
when testing was localized to specific areas of the tongue.
sweetness. Fructose was more susceptible to suppres-
Instead of whole-mouth tasting, the taste solution was
sion than sucrose, an important point in the flavor formu-
applied to localized areas of the tongue using a cotton
lation of soft drinks.
swab. Scientists believe that whole-mouth perception may
Caffeine bitterness was also suppressed by sodium
compensate for some of the regional deficits, such that
salts. Bitter compounds and sodium salts showed asym-
older respondents may not even be aware that they have
metrical suppression, such that bitterness was suppressed
experienced a taste loss.
to a variable degree, whereas saltiness was generally un-
The decline in the perception of bitter and salty tastes
affected.16 Though mixture suppression effects can be
may influence food choices and eating habits. Generally,
substantial, there are no instances of two tastes canceling
intensely bitter and salty tastes are perceived as unpleas-
each other, as can be the case with colors and tones. Be-
ant. There were no major deficits in the scaling of salti-
cause the effects of some taste mixtures may involve mul-
ness in mashed potatoes or tomato juice, and no evidence
tiple transduction mechanisms, it can be difficult to pre-
for an age-related rise in salt consumption. By contrast,
dict how a given mixture will behave. Optimizing mixture
the perceived bitterness of PROP solutions declined with
interactions for consumer acceptance has been the prov-
age. Older women expressed an increased liking for bitter
ince of flavor chemists.
cruciferous vegetables and salad greens. The perception
Studies in taste psychophysics suggest that mixture
of bitter intensity declines with age and may influence the
suppression effects are neurally mediated and do not in-
liking for bitter in beverages and foods.
volve tastant competition for the same receptors in the
oral cavity.6 The suppression of PTC bitterness by su-
Mixture Interaction Studies
crose was much greater among PTC tasters than
Beverages are complex mixtures of tastes, flavors, and tex- nontasters, suggesting that the effect depended on per-
tures. Some compounds are present at above-threshold ceived bitterness as opposed to PTC c~ncentration.~'~
levels, whereas others such as caffeine may be present at Analogous results were recently obtained by Ly and
below-threshold levels. These different tastants can en- Drewnowski9 for PROP solutions sweetened with
hance or suppress each other depending on their concen- neohesperidin DC. These data, summarized in Figure 2,
tration, the nature of the food or beverage, and the experi- suggest that mixture interactions may further depend on
mental methods involved.'6.'7 genetic taste factors and individual sensitivity to bitter
At near-threshold concentrations, mixtures can be taste.
tasted even when each of their components is too weak to The perceived intensity of a taste stimulus may also
be tasted separately.6 Individual detection thresholds de- increase when an odorant is added.35Whereas laboratory
cline as the number of mixture components increases. The studies have focused on mixtures of pure tastants in wa-
taste synergy ofmixtures is such that multiple ingredients ter, the taste of beverages involves multiple interactions

Nutrition Reviews@,Vol. 59, No. 6 165

 9-
t k n t a s t e r s : PROP
8- +Nontasters: PROP + NHDC
+Tasters: PROP
7-
.-0
In
+Tasters: PROP +NHDC
7-1

E 6-
0,
.-c
c

In
In 5-
0)

E
2rn

1i
0.003 0.01 0.03 0.1 0.3 1 3.2 0.003 0.001 0.03 0.1 0.3 1 3.2
PROP concentration (mmolA) PROP concentration (mmolA)
Figure 2. Bitterness intensity (left panel) and hedonic ratings (right panel) for 7 PROP solutions, before and after the addition of
neohesperidin dihydrochalcone (NHDC), by PROP taster status. From reference 9, with permission.

among different tastes and between taste, aroma, viscos- similarities among groups and classes of odors. The im-
ity, and temperature. Studies of coffee-sucrose as opposed portant odor attributes are intensity, quality, and the he-
to caffeine-sucrose mixtures showed that mixture suppres- donic tone. Caffeine forms complexes with volatile flavor
sion was affected by f l a ~ o r At
. ~ similar
~ . ~ ~ perceived inten- molecules, altering their solubility, and modifying their
sities, caffeine bitterness or coffee flavor were suppressed perceived flavor impact. Studies reported that caffeine al-
by sucrose but the perception of sweetness was not af- tered the solubility of such compounds as ethyl benzoate
fected by coffee or caffeine. Follow-up studies using ter- and anisole, as well as terpenes and f~rans.~O
nary mixtures of caffeine, sucrose, and a vehicle (water, Odor molecules may stimulate trigeminal nerve end-
carboxymethylcellulose, or gelatin) showed that both ings as well as olfactory receptors. Carbonated beverages,
sweetness and bitterness were suppressed even more.38 alcohol, or the sensation of menthol are perceived through
the trigeminal nerve. Capsaicin, the active ingredient of
The Flavor of Foods and Beverages chili peppers, stimulates pain fibers as opposed to taste
receptors. Recent studies showing that PROP tasters may
Taste, aroma, and mouthfeel all contribute to the flavor of
be more responsive to hot peppers and alcohol suggest
foods. Whereas the four basic tastes are sweet, sour, salty,
that genetic taste markers and trigeminal perception may
and bitter, the range of taste experiences is far more exten-
well be linked.7
sive.$ Humans describe caffeine and quinine as purely
Mouthfeel refers to texture, as it is perceived in the
bitter, calcium chloride as bitter-salty, and urea as bitter-
mouth in the course of drinking and swallowing. Among
sour. The quality and the temporal profile of the taste
terms used to describe the mouthfeel of beverages are
experience may also vary. Saccharin has been described
smooth, viscous, and creamy, as well as foamy, clean, cool,
as sweet with a bitter aftertaste, whereas catechins are
or lit~gering.~’Pure caffeine has been described as having
bitter compounds with a sweet aftertaste. Caffeine has a
a “cottony” mouthfeel and it produces a drying effect.
unique temporal profile that builds up faster than quinine,
Beverages containing caffeine are sometimes said to have
has a slower rate of decay, and shows a much more pro-
a “clean” bitter taste and tea containing caffeine has been
longed a f t e r t a ~ t eThe
. ~ ~ time to maximum bitterness can be
described as
as long as 13 seconds. This breadth and duration of taste
quality that makes caffeine unique for beverage applica-
The Bitter Taste of Caffeine
tions results from the parallel activation of a broad array
of ion channels, specialized taste receptors, and second Caffeine, a methyl xanthine, is present in coffee, tea, and
messengers associated with taste cell membrane^.^ chocolate. It is generally present at low (mmol/L) concen-
Much of food’s flavor is perceived through the olfac- trations and need not be the major bitter ingredient. The
tory impression. Humans can distinguish between sev- taste of coffee-bitter and astringent-is largely due to
eral thousand odors that are sometimes detected at re- phenolic acids as opposed to caffeine. It is roasting that
markably low concentrations. The current thinking is that determines coffee flavor: the rich dark roast is actually
the olfactory system recognizes patterns and searches for somewhat lower in caffeine content than the more acid

166 Nutrition Reviews@,Vol. 59, NO. 6

 light roast. Coffee aroma, a major factor in coffee enjoy- potentiated the taste of Ace-K, saccharin, and neohes-
ment, is due to several hundred volatile chemicals includ- peridin DC. The enhancement was reversed by adenosine
ing alcohols, ketones, aldehydes, and ester^.^ suggesting its potential involvement as a second messen-
The taste, pungency, and color of fermented teas also ger in both sweet and bitter perception. However, caffeine
derive from phenolic compounds, including catechin and did not affect the perceived sweetness of aspartame, su-
epicatechin, and their oxidation products. Depending on crose, fructose, or cyclamate.28Not all studies have repli-
molecular weight, catechins can be bitter or a ~ t r i n g e n t . ~ . ~ cated those effects. Mela49reported that caffeine did not
Epicatechin is generally more bitter than cate~hin.*<~ The affect the taste of sweeteners, and other studies found no
bitterness and astringency of teas have been ascribed to effects of caffeine or adenosine on the perception of sweet
the combination of catechins, saponin, amino acids, and taste.’O
caffeine.2 Caffeine provides the needed “briskness” and The complexity of interactions among different tastes
greatly contributes to the sensory appeal of teas. Com- and flavors makes the perception of individual mixture
plex interactions between caffeine and tea catechins ac- ingredients especially difficult. The perception of mixtures
count for the complexity of tea flavor.42 can involve analysis, synthesis, or fusion.33In an analytic
Fermented cocoa contains polyphenols, catechins, model, all individual components can be separately per-
anthocyanins, and caffeine.43 Bitterness of chocolate is ceived; in a synthetic model, no individual components
partly due to catechins in fermented cocoa that are vari- are separately perceived; and a fusion model allows the
ously described as bitter with sweet aftertaste or as bitter perception of more than one ingredient without the ability
and astringent.2 Additional bitter elements are provided to identify it. The perceived pleasantness of beverages
by caffeine, theobromine, and the interaction of theobro- and foods is independent of the ability to recognize and
mine and diketopiperazines during r o a ~ t i n g . ~ identify any one ingredient.
Tea, coffee, and chocolate are complex taste and fla- One recent study5’of 25 adults examined the taste of
vor mixtures that contain multiple bitter phytochemicals, caffeine-free cola and cola to which different concentra-
including caffeine. Their sensory appeal often depends tions of caffeine (range 0.2-8.2 mmol/L) had been added.
on the subtle balance of sweet, acid, and bitter tastes, Using a simple overall difference test with 20 exposures
combined with multiple flavor elements and sometimes and an arbitrary cutoff point of 75% correct, respondents
the texture‘ of fat. Chocolate, in particular is a complex correctly identified all caffeine-containing beverages ex-
mixture of flavor elements, bitter, sugar, and fat. Some of cept two with subthreshold concentrations of caffeine (0.2
the same flavor combinations are used in the formulation and 0.5 mmol/L). Above-threshold concentrations of caf-
of soft drinks. feine were described as bitter and ~npleasant.~’
The data were widely misinterpreted as showing that
Caffeine As Flavoring Agent caffeine could not have been a flavoring agent. As docu-
Flavor is the complex of sensations allowing us to identify mented above, however, caffeine can exert its effects at
the presence and identity of beverages and foods. These below-threshold or near-threshold levels. Furthermore, the
flavor sensations include not only taste and aroma, but study was seriously flawed from the sensory evaluation
also mouthfeel and even visual and auditory aspects of standpoint. First, tests for an overall difference in taste
foods. Caffeine is a widely used flavoring agent44 that are not the same as attribute tests. In the latter, respon-
contributes to the popularity and enjoyment of beverages dents concentrate on a single attribute, such as bitterness
and foods. It is generally used in soft drinks at levels of or a lingering aftertaste, and ignore all others. Studies
100 mg/L (0.5 mmol/L) with rare products containing up to show that thresholds for attended stimuli are often lower
200 mg/L (1 .O mmoVL). than for unattended s t i m ~ l i .An
~ ~attribute
, ~ ~ difference test
A concentration of 0.5 mmol/L caffeine represents a would have been a more appropriate procedure. Second,
near- or below-threshold level. Studies have placed caf- simple difference tests always use a placebo control, such
feine threshold in water at 0.5 mmol/L (or 94 mg/L). The that same-same pairs are presented along with same-dif-
observed threshold in fruit juice or custard was approxi- ferent pairs. That serves to compare the placebo effect
mately double that. Caffeine thresholds can be reduced with the treatment effect following the same number of
with sensory training: Pangborn30obtained caffeine thresh- exposure^.'^ Because the study failed to include a pla-
olds of 1.5 mmol/L with n a h e panelists and reduced them cebo control condition, no appropriate statistical tests
to 0.4 mmol/L following training.4sGenerally, the percep- could be conducted. That design flaw makes the study
tion of the bitter taste of caffeine declines with age.46 difficult to interpret. Finally, hedonic preferences as op-
In addition to the mixture effects and taste-flavor in- posed to intensity ratings are the major influence on food
teractions described above, caffeine may potentiate the preferences and food choices. In the published study,5’all
impact of some intense sweeteners. Schiffman et a1.47,48 but one subject reported prefening the flavor of caffeinated
found that adaptation of the tongue to methylxanthines cola to caffeine-free cola.

Nutrition Reviews@,Vol. 59,No. 6 167

 A recent "taste test" conducted by Consumer Re- and foods is self-limited by its bitterness threshold be-
portss5 likewise reported that only 24% of middle school cause at above-threshold concentrations caffeine is per-
students were able to identify the soda they had preferred ceived as bitter and unpleasant. Its flavor properties, in-
earlier-caffeinated or noncaffeinated. That test required cluding a unique temporal profile, and documented inter-
the students to come up with correct answers five times in actions with sweeteners, acids, and volatile flavor mol-
a row. That study took no account of probability statis- ecules contribute to the flavor profile of beverages and
tics. Assuming that the students identified a given soda foods. Caffeine is an integral part of many beverages, in-
with 75% certainty, the probability of five correct answers cluding soft drinks, and contributes to their sensory ap-
in a row would be 0.75 to the fifth power, or only 0.24. As peal.
in the previous study, the flavor of the caffeinated bever-
age was more preferred. 1. Drewnowski A. Taste preferences and food intake.
Annu Rev Nutr 1997;17:237-54
2. Drewnowski A, Gomez-Carneros C. Bitter taste,
Flavor Perception and Hedonic Response
phytonutrients, and the consumer: a review. Am J
The taste system identifies foods and beverages for hu- Clin Nutr 2000;72:1424-35
3. Bitterness in foods and beverages. In: Rousseff RL,
man consumption. Whereas most laboratory studies have
ed. Developments in food science. Vol 25. Amster-
focused on intensity, food choices are influenced by the dam: Elsevier, 1990
quality and perceived pleasantness of beverages and 4. Brand JG. Biophysics of taste. In: Beauchamp GK,
food^.^^.^' Consumers can rate overall palatability without Bartoshuk L, eds. Tasting and smelling. Handbook
being consciously aware of all food ingredients, espe- of perception and cognition, 2nd ed. San Diego,
cially those present at near-threshold levels. BreslinI6 CA: Academic Press, 1997
5. Schiffman SS. Taste quality and neural coding:
makes a point that though bread is not perceived as salty, implications from psychophysics and neurophysi-
bread without salt is unpalatable. Similarly, respondents ology. Physiol Behav 2000;69:147-59
who rated high-fat and low-fat dairy spreads as equiva- 6. Lawless HT. Evidence for neural inhibition in bitter-
lent in both fatness and creaminess, tended to prefer the sweet taste mixtures. J Comp Physiol Psycho1
high-fat ~ e r s i o n .Though
~' judged as equal in fat content, 1979;93:538-47
7. Bartoshuk LM. The biological basis of food percep-
the stimuli were hedonically different. In the same way, tion and food acceptance. Food Quality and Pref-
colas and caffeinated soft drinks provoke a wide range of erence 1994;4:21-32
taste and flavor experiences that directly contribute to 8. Hall MJ, Bartoshuk LM, Cain WS, Stevens JC. PTC
preferences. taste blindness and the taste of caffeine. Nature
Caffeine, in particular, exerts its effects at near-thresh- 1975;253:442-3
9. Ly A, Drewnowski A. PROP (6-n-propylthiouracil)
old levels, whether in tea or in selected soft drinks (Table
tasting and sensory responses to caffeine, sucrose,
1). The amount of caffeine that can be added to beverages neohesperidin dihydrochalcone, and chocolate.
Chem Senses 2000;26:41-7
Table 1. Caffeine Content of Foods and Beveraaes 10. Drewnowski A, Henderson SA, Shore AB. Genetic
Caffeine Content (mg) sensitivity to 6-n-propylthiouracil (PROP) and he-
donic responses to bitter and sweet tastes. Chem
Item (serving size) Typical Range Senses 1997;22:27-37
Coffee (250 mL) 11. Drewnowski A, Henderson SA, Shore AB. Taste
Brewed, drip 100 60-1 80 responses to naringin, a flavonoid, and the accep-
Instant 65 30-120 tance of grapefruit juice are linked to genetic sen-
Decaffeinated 3 ' 1-5 sitivity to 6-n-propylthiouracil. Am J Clin Nutr
Espresso (30 mL) 40 3 0-5 0 1997;66:391-7
12. Reed DR, Nanthakurnar E, North M, et al. Localiza-
Tea (250 mL) tion of a gene for bitter taste perception to human
Brewed tea-green, 60 25-1 10 chromosome 5p15. Am J Hum Genet 1999;64:
black, oolong 1478-80
Instant 28 24-3 1 13. Reed DR. Gene mapping for taste related pheno-
Iced 25 9-5 0 types in humans and mice. Appetite 2000;35:189-
90
Soft drinks (250 mL) 14. Adler E, Hoon MA, Mueller KL, et al. A novel family
Cola and citrus beverages 24 2040 of mammalian taste receptors. Cell 2000; 100:693-
702
Cocoa beverage (250 mL) 6 3-32
15. Chandrashekar J, Mueller KL, Hoon MA, et al. T2Rs
Dark chocolate (30 g) 20 5-3 5 function as bitter taste receptors. Cell 2000;lOO:
703-1 1
Milk chocolate (30 g) 6 1-1 5 16. Breslin PA, Beauchamp GK. Suppression of bitter-
ness by sodium: variation among bitter taste stimuli.
Adapted from reference 58. Chem Senses 1995;20:609-23

168 Nutrition Reviews", Vol. 59,NO. 6

17. Breslin PAS. Interactions among salty, sour and pressive behavior of bitter-sweet mixtures. Percept
bitter compounds. Trends in Food Science and Tech- Mot Skills 1991;73:1216
nology 1996;7:390-9 38. Calvino AM, Garcia-Medina MR, Cometto-Muniz
18. Rosenzweig S , Yan W, Dasso M, Spielman Al. Pos- JE, Rodriguez MB. Perception of sweetness and
sible novel mechanism for bitter taste mediated bitterness in different vehicles. Percept Psychophys
through cGMF! J Neurophysiol 1999;81:1661-5 1993;54:751-8
19. Herness MS, Gilbertson TA. Cellular mechanisms 39. Leach EJ, Noble AC. Comparison of bitterness of
of taste transduction. Annu Rev Physiol caffeine and quinine by a time-intensity procedure.
1999;61:873-900 Chem Senses 1986;11:339-45
20. Spielman Al. Gustducin and its role in taste. J Dent 40. King BM, Solms J. Interactions of volatile flavor
Res 1998;77:539-44 compounds with caffeine, chlorogenic acid and
21. Schiffman SS, Booth BJ, Losee ML, et al. Bitterness naringin. In: Schreier R ed. Flavour '81. Berlin: de
of sweeteners as a function of concentration. Brain Gruyter and Co., 1981:707-16
Res Bull 1995;36:505-13 41. Drewnowski A. Dietary fats: perceptions and pref-
22. Neely G, Borg G. The perceived intensity of caf- erences. J Am Coll Nutr 1990;9:431-5
feine aftertaste: tasters versus nontasters. Chem 42. Wickremasinghe RL. Tea. Adv Food Res
Senses 1999;24:19-21 1978;24:229-41
23. Drewnowski A, Henderson SA, Barratt-Fornell A. 43. Arts ICW, Hollman PCH, Kromhout D. Chocolate as
Genetic sensitivity to 6-n-propylthiouracil (PROP) a source of tea flavonoids. Lancet 1999;354:488
and sensory responses to sugar and fat mixtures. 44. National Academy of Sciences, Committee on Food
Physiol Behav 1998;63:771-7 Chemicals. Food chemicals codex, 4th ed.
24. Stevens JC. Detection of tastes in mixtures with Washinton, DC: NAS Press, 199652-3
other tastes: issues of masking and aging. Chem 45. Pangborn RM. Influence of hunger and sweetness
Senses 1996;21:211-21 preferences on taste thresholds. Am J Clin Nutr
25. Stevens JC. Detection of very complex taste mix- 1959;7:280-7
tures: generous integration across constituent com- 46. Hyde RJ, Feller Rf? Age and sex effects on taste of
pounds. Physiol Behav 1997;62:1137-43 sucrose, NaCI, citric acid and caffeine. Neurobiol
26. Schiffman SS, Crofton VA, Beeker TG. Sensory Aging 1981;2:315-8
evaluation of soft drinks with various sweeteners. 47. Schiffman SS, Gill JM, Diaz C. Methyl xanthines
Physiol Behav 1985;34:369-77 enhance taste: evidence for modulation of taste by
27. Pilgrim F. Interactions of suprathreshold taste adenosine receptor. Pharmacol Biochem Behav
stimuli. In: Kare M, Halpern B, eds. Physiological 1985;22:195-203
and behavioral aspects of taste. Chicago, IL: Uni- 48. Schiffman SS, Diaz C, Beeker TG. Caffeine intensi-
versity of Chicago Press, 1959 fies taste of certain sweeteners: role of adenosine
28. Schiffman SS, Gatlin LA, Frey AE, et al. Taste per- receptor. Pharmacol Biochem Behav 1986;24:429-
ception of bitter compounds in young and elderly 32
persons: relation to lipophilicityof bitter compounds. 49. Mela DJ. Caffeine ingested under natural condi-
Neurobiol Aging 1994;15:743-50 tions does not alter taste intensity. Pharmacol
29. Schiffman SS, Gatlin LA, Sattely-Miller EA, et al. Biochem Behav 1989;34:483-5
The effect of sweeteners on bitter taste in young 50. Brosvic GM, Rowe MM. Methyl xanthine, adenos-
and elderly subjects. Brain Res Bull 1994;35:189- ine, and human taste responsivity. Physiol Behav
204 1992;52:559-63
30. Pangborn RM. Taste interrelationships. Food Re- 51. Griffith RR, Vernotica EM. Is caffeine a flavoring
search 1960;25:245-6 agent in cola soft drinks? Arch Fam Med
31. Kamen JM, Pilgrim FJ, Gutman NJ, Kroll BJ. Inter- 2000; 9: 727-34
actions of suprathreshold taste stimuli. J Exp 52. Marks LE, Wheeler ME. Attention and the detect-
Psycho1 1961;62:348-56 ability of weak taste stimuli. Chem Senses
32. McBride RL. Three models for taste mixtures. In: 1998;23: 19-29
Laing DG, Cain WS, McBride RL, Ache BW, eds. 53. Frank RA, van der Klauw NJ, Schifferstein HN. Both
Perception of complex smells and tastes. San Di- perceptual and conceptual factors influence taste-
ego, CA: Academic Press, 1989:265-82 odor and taste-taste interactions. Perception and
33. McBride RL, Finlay DC. Perceptual integration of Psychophysics 1993;54:343-54
tertiary taste mixtures. Perception and Psychophys- 54. Meilgaard M, Civille GV, Carr BT. Sensory evalua-
ics 1990;48:326-30 tion techniques, vol I. Boca Raton, FL: CRC Press,
34. Frijters JE, Schifferstein HN. Perceptual interac- 1987
tions in mixtures containing bitter tasting sub- 55. Consumer Reports. Kids and caffeinated sodas.
stances. Physiol Behav 1994;56:1243-9 Consumer Reports, March 2000:9
35. Schifferstein HN, Verlegh PW. The role of congru- 56. Mattes RD. Influences on acceptance of bitter foods
ency and pleasantness in odor-induced taste en- and beverages. Physiol Behav 199436:1229-36
hancement. Acta Psycho1 (Amst) 1996;94:87-105 57. Mela DJ, Mattes RD, Tanimura S, Garcia-Medina
36. Calvino AM, Garcia-Medina MR, Cometto-Muniz MR. Relationships between ingestion and gusta-
JE. Interactions in caffeine-sucrose and coffee-su- tory perception of caffeine. Pharrnacol Biochem
crose mixtures: evidence of taste and flavor sup- Behav 1992;43:513-21
pression. Chem Senses 1990;15:505-19 58. Barone JJ, Roberts HR. Caffeine consumption. Food
37. Calvino AM, Garrido D. Spatial and temporal sup- Chem Toxicol 1996;34:119-29

Nutrition Reviews", Vol. 59, No. 6 169