|Home | About | Journals | Submit | Contact Us | Français|
The act of eating and drinking brings food-related chemicals into contact with taste cells. Activation of these taste cells, in turn, engages neural circuits in the central nervous system that help animals identify foods and fluids, determine what and how much to eat, and prepare the body for digestion and assimilation. Analytically speaking, these neural processes can be divided into at least three categories: stimulus identification, ingestive motivation, and digestive preparation. This review will discuss recent advances in peripheral gustatory mechanisms, primarily from rodent models, in the context of these three major categories of taste function.
Our understanding of peripheral gustatory mechanisms continues to advance at a rapid pace. Ultimately, these neurobiological processes must be linked to behavioral outcomes. At times, such efforts have produced seemingly paradoxical results; e.g., knocking out a taste receptor caused severe impairments in one behavioral task but not in another. To explain these apparent disparities, it is important to realize that there are at least three categories of taste processing. Stimulus identification is the detection or discrimination of sensory signals arising from taste cell activation. Ingestive motivation involves processes that promote or discourage ingestion Digestive preparation refers to feed-forward physiological reflexes that protect oral tissues, aid digestion, and facilitate homeostasis. It must also be recognized that behavioral responses to taste stimuli can also be influenced non-gustatory factors, including olfactory, somatosensory, and visceral signals. We propose that integrating these perspectives into studies of taste function will help establish more logical links between neural processes and taste-related behavior.
Stimulus identification refers to the ability of animals to discriminate between the gustatory signals generated by different taste stimuli. Such processes allow animals to learn about foods by associating particular tastes with other stimuli and/or outcomes, ultimately facilitating survival. In humans, stimulus identification can be assessed through verbal qualitative descriptors such as “sweet,” “sour,” “salty,” “bitter” and “umami.” In nonverbal animals, more objective approaches, such as operant and classical conditioning procedures, must be used to draw inferences about whether the subject can identify and discriminate among taste compounds. When conditioning techniques are used toward this end, the taste stimuli serve as cues for other events, such as reward or punishment. This ensures that responses are not driven by an animal’s natural preference or aversion for a particular taste stimulus.
In rodents, salt taste transduction appears to occur through at least two ion channel receptors. One is specific for sodium (and lithium) salts and is suppressed by the epithelial sodium channel (ENaC) blocker amiloride (e.g., [2-•4]). The other receptor is less cation-selective and is unaffected by amiloride (; see also ). It is widely believed that the amiloride-sensitive receptor is an ENaC [•7], and that it mediates the sodium-selective responses of the “specialist” neurons in the chorda tympani nerve and its ganglion (Fig. 1). Indeed, although amiloride treatment of the tongue only partially suppresses sodium responses in the whole chorda tympani nerve (because the amiloride-insensitive salt transduction remains), it severely attenuates or abolishes sodium responses in the sodium-specialist neurons in the geniculate ganglion [8–10] and eliminates the ability of rodents to recognize sodium and distinguish it from other cations (e.g, [11–13]). These findings provide a compelling link between peripheral gustatory mechanisms of transduction, neural signaling, and sodium identification (see ).
The identification of the ion channel receptor(s) underlying the amiloride-insensitive component of salt taste transduction has been more elusive, but recent reports in rodents implicate a variant of the vanilloid receptor TRPV1 . For instance, amiloride treatment eliminates the tonic portion of the chorda tympani nerve response to NaCl in TRPV1 knock-out (KO) mice; in WT mice it merely reduces the tonic response [15;16]. Unexpectedly, behaviorally assessed detection thresholds for NaCl [16;17] and KCl  in TRPV1 KO mice match those in WT mice, and amiloride treatment raises the NaCl threshold to similar degrees in both genotypes (although see ). Thus, even though the chorda tympani nerve response to NaCl+amiloride was more disrupted in TRPV1 KO mice than WT controls, a link to a behavioral outcome was not established. This is likely because there are other receptors in different taste bud fields (i.e., posterior tongue, palate, laryngeal epithelium), innervated by other nerves (i.e., glossopharyngeal, greater superficial petrosal, superior laryngeal; Fig. 1), which are sufficient to maintain stimulus detectability. It remains possible, however, that TRPV1 KO mice would display behavioral differences from their WT counterpart if other taste functions were measured, including salt discrimination or salt responsiveness in a brief access test. For instance, TRPV1 KO mice show altered preference-aversion functions to salts in long-term 2-bottle tests, but these tests are often influenced by postingestive factors . A definitive conclusion awaits further testing.
T1R3 KO mice display severely blunted unconditioned licking responses to L-amino acids and sweeteners [18;19], but normal detection thresholds for both classes of stimuli [••20]. The latter finding was unexpected because the T1R1+3 and T1R2+3 heterodimers are thought to be the principal taste receptors for L-amino acids and sweeteners, respectively [•21]. Thus, while the gene deletion appears to largely eliminate ingestive motivation for the stimuli, either the T1R2 subunit and/or other receptors are sufficient to enable the KO mice to detect L-amino acids and sweeteners (see ). This inference is consistent with the observation that the chorda tympani nerve in T1R3 KO mice still displays some responsiveness, albeit severely compromised, to high concentrations of sucrose [18;19;23]. Nevertheless, the finding that C57BL/6J mice did not discriminate sucrose from glucose or fructose [•24] supports the view that the T1R2+3 heterodimer, which has been shown to bind with all three of these sugars [22;25;26], is the principal receptor for “sweet-tasting” ligands. This finding also suggests that sucrose, glucose, and fructose generate a unitary qualitative perception, at least from the standpoint of stimulus identification.
The mere existence of separate molecular receptors does not guarantee that the ligands for those receptors will be behaviorally discriminable. For example, in rats and humans there are over two dozen taste receptors (T2Rs) that are thought to bind with bitter-tasting ligands. If a taste cell expresses one T2R, then it is likely that it expresses many others . The degree of co-expression of T2Rs, although extensive, may not be complete ([28;29], but see ). For example, in rats bitter-tasting ligands, such as quinine and denatonium, do not activate identical subsets of taste bud cells . Based on these cellular data, one might predict that rats should be able to behaviorally discriminate between quinine and denatonium. This, however, does not appear to be the case . Behavioral discrimination tasks suggest that the initial signals differentiating quinine from denatonium in the periphery converge downstream in the gustatory pathway, resulting in a unitary signal. Indeed, in rats, at the level of the parabrachial nucleus (Fig. 1), a brainstem gustatory relay to forebrain structures, quinine and denatonium appear to stimulate activity in a similar subset of taste-responsive neurons . Clearly, more behavioral work is needed with a broader array of compounds before definitive conclusions can be reached about whether rodents can discriminate bitter-tasting ligands ..
It was recently reported that rats can discriminate nicotine from quinine in an operant taste discrimination task . These latter findings do not necessarily conflict with those from the study testing discrimination of quinine and denatonium . Because nicotine can both stimulate nicotinic acetylcholine receptors and modulate certain ion channels , it may be activating a variety of taste receptor cell types, gustatory afferent fibers, and trigeminal free nerve endings. Accordingly, nicotine could generate a more complex oral sensation than quinine. This example illustrates two important points. When an animal successfully discriminates two taste stimuli, it is difficult to attribute the discrimination unambiguously to central taste processing. On the other hand, when an animal fails to discriminate two taste stimuli (provided that learning and concentration effects can be ruled out), this suggests that an equivalence exists between the indiscriminable stimuli somewhere in the nervous system.
One of the growing controversies is whether fats generate a distinct taste quality. Prior work indicated that detection of fat was based on its ability to both alter the tactile properties of foods and retain food-related odors. There is accumulating evidence, however, supporting the involvement of the gustatory system. For instance, the CD36 fatty acid translocater is expressed in murine taste cells, and may serve as a receptor. This protein is necessary for normal responsiveness to fatty acids at both the cellular and behavioral (i.e., 30 min or 24-hr intake) levels [37;••38;•39]. Likewise, there is evidence that long chain unsaturated fatty acids can block delayed rectifying potassium (DRK) channels, which, in turn may bring taste cells closer to threshold and make them more responsive to other taste stimuli such as sugars and salts .
From a behavioral standpoint, rodents and humans can detect long chain fatty acids through oral mechanisms [41–43]. If a taste aversion is conditioned to linoleic acid, both rats and mice will avoid ingesting the same compound on future occasions [39;42;43]. Further, transection of the chorda tympani nerve in rats impairs long chain unsaturated fatty acid detection in rats [42;44]. Although these findings indicate that information in the chorda tympani nerve is necessary for detection to take place, it is unclear why stimulation of the anterior tongue with linoleic acid fails to activate gustatory afferents in the chorda tympani nerve or geniculate ganglion [45;46]. Another conundrum is the exact definition of the perceptual qualitative nature of fatty acid taste. Although speculative, it is possible that there is no fat taste quality per se. Instead, the effects of fats on the gustatory system may be limited to activating the central neural circuits that subserve ingestive motivation and digestive preparation (see Fig. 1 and below).
The motivational function of taste has been referred to as affect, hedonics, palatability, and reward. All of these processes share the same fundamental property of facilitating or inhibiting ingestion. It is important to recognize that two taste compounds can be equally preferred or avoided but have distinct taste qualities. For instance, even though rats avoid high concentrations of quinine and NaCl, they can nevertheless discriminate the tastes of these stimuli.
Deleting the genes encoding the T1R2 or T1R3 receptor subunits leads to profound deficits in responsiveness to sweeteners, as measured in brief-access lick tests . Recently, however, it has become clear that not all carbohydrates require T1R2 or T1R3 to support normal taste-related acceptability. T1R2 KO and T1R3 KO mice each display normal concentration-dependent responsiveness to Polycose, a glucose polymer mixture, in brief access tests [•47;•48] (Fig. 2). This indicates that glucose polymers can either bind with T1R2 (or T1R3) alone, or bind with a yet-to-be-identified taste receptor(s), and stimulate intake. Tests with T1R2 and T1R3 double-KO mice would help distinguish between these possibilities.
It is clear that the orosensory characteristics of fats have motivational salience to rodents. Indeed, rodents will lick for fats in a concentration-dependent manner . An intact olfactory system is not required for those responses to be displayed , but the presence of CD36 and TRPM5 is necessary initially [39;•51]. Following repeated testing, however, CD36 KO or TRPM5 KO mice will begin to display responsiveness to fat stimuli presumably through an associative learning process by which the positive consequences of their ingestion are paired with some detectable oral cue associated with the stimuli [39;51].
When low concentrations of long chain unsaturated fatty acids (e.g. linoleic acid) were added to prototypical taste stimuli in brief-access tests, concentration-response functions of rats were shifted leftward, indicating that stimulus acceptability was enhanced . This supports the hypothesis that fatty acids sensitize taste cells to other taste stimuli through their action on DRK channels. In humans, however, long chain unsaturated fatty acids do not appear to alter threshold or suprathreshold sensitivity to taste stimuli . It would be instructive to determine whether (a) the addition of linoleic acid changes the hedonic ratings of taste stimuli so as to bring the human taste testing more in register with the brief-access lick tests in rats; or (b) linoleic acid improves detection thresholds for prototypical taste stimuli in rats when operant psychophysical procedures are used.
A third function of gustatory input is the activation of physiological reflexes that produce effects like delaying gastric emptying, protecting the oral cavity, facilitating digestion, and maintaining homeostasis. These are commonly referred to as cephalic-phase reflexes because they are triggered by stimulation of head receptors. For instance, a recent study documented that bitter taste alone can delay gastric emptying in human subjects [•54]. This could have adaptive value in that it would both slow the rate at which an ingested toxin was absorbed and allow more time for emetic processes to expel the chemical culprit. These taste-triggered physiological effects have the potential to alter behavior.
Many foods and beverages can damage the hard and soft tissues in the oral cavity. Taste-induced salivation plays a central role in mitigating these negative effects by reducing irritation to soft tissues, frictional wear of the tooth enamel, and acidic dissolution of tooth mineral. Damage to oral tissues is also reduced by secretion of salivary proteins (e.g., proline-rich, anti-bacterial and anti-fungal proteins) that help neutralize reactive chemicals and microorganisms in foods. A recent study  revealed that oral stimulation with chemicals representing different taste qualities stimulates secretion of different quantities and types of proteins. More work is needed to determine the neural basis and functional significance of this latter observation.
Taste activates several physiological reflexes, which serve both to facilitate digestion and maintain homeostasis. For instance, oral stimulation with sweeteners but not salts, monosodium glutamate, or quinine activates a cephalic phase insulin release (CPIR) in rats [•56] and humans . Likewise, lingual stimulation with fats or fatty acids activates a pancreato-biliary response in wild-type mice, but not CD36 KO mice , and in humans elevates serum triglyceride levels  and induces release of pancreatic polypeptide . Investigators recently demonstrated that stimulation of the anterior tongue with glucose increased sympathetic stimulation of the interscapular brown adipose tissue (BAT) depot in rats . Given that BAT stimulation increases energy expenditure, it may aid in body weight regulation.
The three categories of taste function discussed here must have dissociable neural substrates at some level in the gustatory neuraxis (Fig 1) . Although these substrates have yet to be clearly delineated, there are hints in the literature. For example, in rats, stimulus identification relies on input from the gustatory branches of the facial nerve, whereas taste signals carried by the glossopharyngeal nerve appear unnecessary to support this function [62;63]. Neurons in the geniculate ganglion of the facial nerve of the rat can be divided into those that contribute input to the ascending gustatory pathway and those that contribute input to local brainstem circuits through the reticular formation [•64]. Perhaps the former are involved largely stimulus identification and ingestive motivation and the latter are involved with oromotor [e.g., 65] and physiological reflexes [e.g., 66]. It remains to be seen whether neurons in the petrosal ganglion of the glossopharyngeal nerve can be similarly subdivided on the basis of their central anatomical fates. Finally, the ascending gustatory system of the rodent bifurcates at the level of the forebrain, and it has been a longstanding hypothesis that the thalamocortical branch is more involved with stimulus identification whereas the ventral forebrain branch is more involved with ingestive motivation . While there is growing evidence supporting the latter, the former awaits to be explicitly tested.
As more is revealed about the neurobiological hardware of the gustatory system, it will be important for investigators to consider the multidimensional nature of taste function in their analysis and interpretation. Accordingly the use of a variety of complementary experimental approaches to study gustatory function will optimize the formation of links between neural processes and taste-related behavior.
We would like to thank Clare Mathes and Yada Treesukosol for providing feedback on the article. A.C.S. would also like to acknowledge research support from the National Institute on Deafness and Other Communication Disorders (R01-DC-004574).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.