Sensory receptor cells that project antenna-like hairs are located within

I Introduction

A number of sensory receptor cells appear to be modified epithelial cells and to share many properties common to epithelial tissues. These properties include the existence of two different membrane regions—an apical and a basolateral region—separated by membrane specializations, the intercellular junctional complexes, that bind adjacent cells together. The net result is that the two different regions of epithelial cells and sensory receptor cells can be exposed to vastly different milieu. This is especially the case in gustatory and olfactory receptor cells. The apical membrane tips are bared to an external environment that can vary profoundly in chemical composition, yet the basolateral membranes are bathed in a protected, relatively constant medium.

As a first generalization about olfactory and gustatory chemoreceptors, it is important to note that the apical membrane is the site where chemosensory transduction occurs. The intercellular junctional complex restricts or confines the chemical stimulation to the specialized apical membrane. Consequently, this membrane surface is exposed to gustatory and olfactory stimuli dissolved in an ionic medium that is quite different from that of the tissue spaces surrounding the basolateral membrane. Because chemosensory transduction is often associated with a receptor current (that is, a translocation of ions) across the stimulated membrane, the unique ionic environment at the apical region often imposes important constraints on the transduction process.

The basolateral membrane of olfactory and gustatory receptor cells represents the surface across which flow the receptor currents that are generated in the apical, chemosensitive tips. The basolateral membrane spreads the currents throughout the receptor cells and, in some cases, modifies the chemosensory signals before transmitting them to higher centers in the brain. In the case of olfactory receptors, signal transmission to the brain is a one-step process. Olfactory receptor cells communicate directly to the CNS via axons. In the case of gustatory receptor cells, signal transmission is a two or more step process. Taste cells communicate via synapses with other taste cells and with sensory afferent fibers that secondarily transmit signals to the CNS.

2 Physiological Responses of Chemoreceptor Cells: Spectral and Temporal Tuning

The task of sensory receptor cells is to interface efficiently with the physical environment and provide the brain with information that can be processed for adaptive behavior. Two key elements of this information are stimulus identification and stimulus localization. Important constraints on receptor cells are the amount and speed of information processing, sensitivity, and noise rejection. Indeed, beyond truly random noise, sensory systems determine what is signal and what is biological noise. Receptor cells act as filters with both spectral and temporal properties. Analogous to acoustic and visual senses, one may consider chemoreceptor cells for their spectral and temporal properties. The chemical spectrum is composed of the vast number of chemical compounds that exist in the environment. Each compound stimulates to various degrees some of the hundreds of thousands of relatively narrowly tuned receptor cells. Thus, each compound or mixture elicits a characteristic spatial pattern of excited receptor cells. The spectral tuning of a chemoreceptor organ results from the organ-specific assemblages of different receptor cells. Cell processes, such as receptor desensitization, second messenger release, and ion channel dynamics, give each receptor cell its temporal properties. These temporal filter properties vary from cell to cell and form a substrate for the discrimination of rapid concentration fluctuations, such as those found in odor plumes. Homarus americanus has contributed considerably to our understanding of such chemoreception principles.

a. Spectral Tuning and Mixture Discrimination The spectral tuning properties of several appendages have been determined by measuring the sensitivity of individual receptor cells to various single compounds. At biologically relevant concentrations, single receptor cells often respond to only one or a few stimuli, that is, they are narrowly tuned (Fig. 8). Effective stimuli include amino acids and amines, but not sugars and alcohols. This is a common feature of chemoreceptors of marine decapod crustaceans. Furthermore, chemosensitivity toward other compounds has been established (secondary plant compounds, Derby et al., 1984; collagens, Johnson et al., 1988; oligopeptides, Merrill, 1992; fractions of fuel oil, Atema, 1980, Derby and Atema, 1981a), often with remarkably strong and selective responses by a few receptor cells. Narrow tuning may allow the identification of key stimuli without (complete) cross-adaptation by chemicals in the background (Borroni and Atema, 1988; Derby and Atema, 1988; Atema et al., 1989). Most chemoreceptor studies of the various lobster species, including Homarus americanus, have focused on lateral antennules (Ache, 1972; Shepheard, 1974; Johnson and Atema, 1983; Weinstein et al., 1989) and anterior walking legs (Derby and Atema, 1982a; Johnson et al., 1984; Bayha et al., 1993), but a few have investigated the chemosensory properties of the medial antennules (Tierney et al., 1988), maxillipeds (Corotto et al., 1992), and antennae (Derby, 1982; Voigt and Atema, 1992). Among the five chemoreceptor organs so far investigated in H. americanus, only the second antennae are dominated by one narrowly tuned cell population: hydroxyproline-best cells make up 85% of the population. By comparison, hydroxyproline cells account for 46% of all cells tested in the lateral antennules, 26% in the medial antennules, 13% in the maxillipeds, and 16% in the walking legs (Fig. 8). The organ-level tuning is similar within the cephalic appendages (lateral and medial antennules and the antennae), which differ from the thoracic appendages (walking legs and maxillipeds). Each organ represents a differently tuned spectral filter that allows the animal to extract different information from its chemical environment. Resemblance in spectral tuning could indicate a functional overlap of chemoreceptor organs, and the tuning differences between antennal and leg organs may reflect the behavioral separation of olfaction and taste (Atema, 1977a). Some compounds are most likely the preferred long-distance signals (olfactory) because of their longevity in seawater and their signal-to-back-ground ratio, while other compounds could allow a refined evaluation of food quality near the source of release and/or upon contact (gustatory); behavioral evidence for this notion is still sparse. The great variety of differently tuned chemoreceptor cells in each of the five chemoreceptor organs points to the idea that these receptor cells are given their tuning properties by different blends of receptor sites. These sites are themselves distributed in different proportions in the various chemoreceptor organs, thus giving the organs their specific “view” on the chemical world (Atema et al., 1989).

In its natural environment, Homarus americanus must extract important chemical cues from a fluctuating chemical background. Mopper and Lindroth (1982) found spatially averaged coastal marine concentrations of amino acids in the submicromolar range; data on naturally occurring concentrations in microodor patches are lacking (Atema, 1988; Manahan, 1990). Background adaptation represents the change in responsiveness of a cell exposed to a constant background concentration of chemical stimulus. This effect becomes evident in seconds and does not fully recover for at least 30 seconds after the cell leaves that background (Gomez and Atema, 1994). Self-adaptation, in which stimulus and background are identical, results in a shift in the stimulus–response function and is thought to extend the cell’s momentary working range, from two log steps of stimulus concentration to five or more log steps (Borroni and Atema, 1988). Cross-adaptation, in which stimulus and background are different, results in a change of slope in stimulus-response functions (Borroni and Atema, 1989). Individual receptor cells show a wide variety of effects under experimental conditions (Atema et al., 1989; Johnson et al., 1989).

The simultaneous presentation of two or more stimuli can result in mixture suppression or enhancement of response in comparison to the response to the cell’s best single compound (Johnson et al., 1985, 1989; Merrill, 1992). Also here, individual cells show the wide variety of effects necessary for a discriminating nose (Atema et al., 1989).

It has been proposed that the brain encodes stimulus quality as an across-neuron pattern, whereas stimulus intensity might be coded by the response magnitude of a receptor cell population. However, recent experiments suggest that intensity may also be encoded by an across-neuron pattern (Johnson et al., 1991; Merrill et al., 1994). Furthermore, a population of unreliable single receptor cells can encode stimulus intensity reliably using either code (Merrill et al., 1994).

b. Temporal Tuning and Signal Processing for Chemotaxis The patchy nature of underwater odor plumes appears to a chemoreceptor organ as a series of bursts of odor in varying strength and duration. These pulse patterns reflect the spatial–temporal distribution of odor patches, which contains directional information that can be used for orientation and distance estimates (Atema, 1985, 1988; Moore and Atema, 1991; Moore et al., 1991b). The lateral antennules and associated brain areas, in addition to extracting information regarding quality and intensity of the odor source described above, may also extract information regarding the spatial distribution of odor concentration based on temporal analysis. This information would be useful for chemotaxis (see Section II,E).

The antennular chemoreceptor cells innervate thick tufts of aesthetasc hairs (Derby, 1982), which create thick boundary layers of water and thus a physical barrier to efficient odor access. To overcome this, antennules flick: the fast downward stroke of an antennule (100 msec) removes water, and therefore odor, caught between the aesthetasc hairs and replaces it with a new odor sample. During the slower upward stroke, odor concentration in the tuft changes only slightly due to diffusion and some wash-out (Moore et al., 1991a). Antennular flicking increases stimulus access (Moore et al., 1991a) and therefore enhances the response magnitude (Schmitt and Ache, 1979); the maximal flicking rate is 4–5 Hz (Berg et al., 1992; Leonard et al., 1994). This odor-sampling strategy is similar to vertebrate sniffing. The response fusion frequency (i.e., the limit of pulse frequency resolution, in which responses to individual stimuli begin to fuse) of antennular hydroxyproline cells is 4–5 Hz, perhaps not coincidentally (Gomez et al., 1994).

Cumulative adaptation, as a measure of recovery from the effects of prior stimulation, reflects the disadaptation rate following adaptation to a series of stimulus pulses. This recovery has been measured directly in antennular hydroxyproline and taurine cells and follows an exponential time course, reaching full recovery in 30 seconds, with a range of 20–50 seconds for individual cells (Voigt and Atema, 1990; Gomez and Atema, 1994). To avoid response reduction due to prior stimulation, most studies used interstimulus intervals of 1 minute or more (Merrill et al., 1994; Voigt and Atema, 1990).

Surprisingly, in leg chemoreceptors of Homarus americanus, interstimulus intervals of 5, 10, or 20 seconds do not affect the degree of cumulative adaptation of the glutamate-sensitive cell population (Voigt and Atema, 1990). This suggests that cumulative adaptation results from a change in the cell’s state, rather than a graduated, stimulus-determined process. As always, individual cells reveal great diversity in the time course of cumulative adaptation, regardless of the interstimulus interval. Legs are probably not involved in detailed temporal analysis, although they play a minor role in chemotaxis (Devine and Atema, 1982).

Receptor cells integrate several adaptation processes. Some of the temporal dynamics of this process have been modeled (Mountain and Atema, 1993) based on a series of intracellular transduction mechanisms (McClintock and Ache, 1989). The model was tested successfully, with published results on second messenger kinetics and in situ receptor cell responses to pulsed stimuli in various backgrounds.

Development

Gustatory cells of vertebrates, unlike all other sensory receptor cells and neurons, are unique in that they originate from local epithelial tissue elements of primary endoderm, and not from neurogenic ectoderm (neural tube, neural crest, or ectodermal placodes). Nevertheless, gustatory cells form synapses and are capable of generating receptor potentials, and like epithelial cells they have a limited life span and regularly replace themselves.

Ontogenetically, taste buds generally develop later than their counterpart olfactory system. For example, in rainbow trout the earliest taste buds are seen only in larvae 8 days post-hatching. Taste bud primordia appear first around the mouth and within the oropharyngial regions a few days prior to hatching, far ahead of those in the head region, in zebrafish (Danio rerio) and channel catfish (Ictalurus punctatus). The primordial cells accumulated beneath the single layer of epithelial cells begin to elongate from the basal lamina to the epithelial surface by 4 days post-fertilization. By 10 days post-fertilization, the primordia begin to erupt as mature taste buds, at which time basal cells and nerve fibers are clearly seen and the dermal papillae begin to form beneath the basal cells. Although details of the development of the innervation of taste buds are still unknown, taste buds are believed to arise from within the specified epithelium (early specification) rather than being induced (nerve induction) by peripheral nerves. Thus, the taste bud, while maintaining its entity, is a constituent of the epithelium without partition from the surrounding epithelial cells. The average life span of taste bud cells on the channel catfish barbel is 12–42 days. Unlike the olfactory epithelium, in which the basal cells function as progenitors for the receptor neurons, epithelial cells surrounding the taste buds divide and some of their daughter cells migrate into taste buds to form gustatory cells.

Otolithic Organs (Macule and Sacule)

The macule consists of a sensory membrane containing the sensory receptor cells with a surface area less than 1 mm2 that supports a heavy otolithic membrane (specific gravity approximately 2.7) (Fig. 1). The otolithic membrane is composed of calcareous material embedded in a gelatinous matrix, with a mean thickness of approximately 50 μm. The position of the otolithic membrane on the sensory receptors depends on the magnitude and direction of the force acting on it. Even when the head is at rest, the calcareous material, because of its mass, exerts a force (Fg) on the receptor equal to the product of its mass and the acceleration due to the gravitational pull of the earth (g), which at sea level is 9.8 m/sec2. The distribution of Fg acting on the underlying sensory cells changes with different degrees of head tilt and with linear head accelerations. The utricular macule is located next to the anterior opening of the horizontal semicircular canal and lies mostly in the horizontal plane. In contrast, the saccular macule lies in a vertical plane in between that of the two vertical semicircular canals.

1.25.3.5.6 Sensory control

The stomach wall and the striated muscles that operate the gastric mill and pyloric pump contain sensory receptor cells sensitive to stretch (mechanoreceptors) and others inside the gut that respond to chemical stimuli. Not all of these have been investigated in the stomatogastric system but single sensory fibers have been shown to be able to affect the pattern on a cycle-by-cycle basis. Receptors in the posterior part of the stomach (PSRs) are stimulated by rhythmic movements and their action, via the CGs, is to activate both the gastric and pyloric rhythms (Nagy and Moulins, 1981). One feedback mechanism that has been described is a sensory neuron (AGR) close to the STG that receives stretch receptor innervation from the powerful GM muscles. The AGR can either enhance or restrain the output to these muscles depending upon the amount of force being sensed (Simmers and Moulins, 1988). In this sense, the AGRs act like both the Golgi tendon organs and muscle spindles of vertebrates which evolved quite separately to perform similar functions. The AGRs in the European rock lobster Hommarus gammarus synapse onto two pairs of command neurons called CG and GI which are located in the CG. When the gastric mill is active in vitro weak firing of the AGR occurs in time with the gastric mill rhythm, when all the power stroke neurons are synchronously active. However, with strong AGR firing, the phase relationships switch to a different pattern in which lateral and medial teeth motor neurons fire in antiphase (Combes et al., 1999). In this case, the feedback from a single mechanosensory neuron is able to specify two different motor pathways in an activity-dependent manner.

Another set of stretch receptors call the gastropyloric receptors (GPRs) do not act via the CGs but instead form a closed loop system in which the GPRs display endogenous rhythmicity (Katz and Harris-Warrick, 1990a). The GPRs found in crabs activate the pyloric rhythm via cholinergic mechanisms and cause some of the gastric mill neurons to fire in pyloric time. The later result appears to be due to the co-release of serotonin (Katz and Harris-Warrick, 1990a, 1990b). Finally, bilateral sensory neurons in the cardiac gutter of the crab stomach project to both the gastric and pyloric feeding circuits (Beenhakker et al., 2004). When stretched artificially, these neurons initiate chewing movements and modify the pyloric filtering movements via projection neurons in the CGs.

Light Response of Rods and Cones

Rods and cones have a membrane potential of about –30 to 35 mV in darkness, and they respond to light with a membrane hyperpolarization, unlike most other sensory receptor cells, which depolarize in response to stimulation. They do not have an action-potential mechanism. Because of their relatively depolarized state in darkness, they continuously release glutamate from their synaptic terminals onto bipolar cells and horizontal cells, the second-order visual neurons in the retina. The light-triggered membrane hyperpolarization reduces the glutamate release. The brighter the light, the larger is the membrane hyperpolarization in rods and cones, and therefore the greater is the reduction in glutamate release. The bipolar cell, which is the second-order throughput neuron, can respond to this decrease in glutamate release with either a hyperpolarization (OFF-bipolar cell) or a depolarization (ON-bipolar cell), depending on the nature of the postsynaptic glutamate receptor it possesses.

The rod and cone responses are relatively slow (with the cone response being the faster of the two), reflecting a second-messenger-mediated transduction mechanism (see the section titled ‘Phototransduction mechanism in rods and cones’). For example, the rod response to a dim, brief flash lasts for over a second for cold-blooded animals at room temperature, though faster at the higher body temperature of mammals. The speeds of the rod and cone responses dictate how rapidly the visual system can detect image changes on the retina (temporal resolution), and therefore motion. Because cones have faster responses than do rods, they are better motion detectors. On the other hand, the slower response of rods allows these cells to sum signals slightly spread over time (temporal integration), thus enhancing their sensitivity in dim light, especially useful for static images.

Figure 2 shows the responses of an amphibian rod to a light flash at different intensities, recorded simultaneously with an intracellular voltage electrode and with a suction pipette that records membrane current at the outer segment. The suction pipette shows that an inward current flows in darkness across the outer-segment membrane (which accounts for the not-so-negative resting potential in darkness), and that light decreases this ‘dark current’ (which accounts for the membrane hyperpolarization) until its complete suppression by bright light (see also Figure 1). With a bright flash, the voltage response shows a fast re-depolarization from the peak hyperpolarization before settling to a plateau level. This transient ‘nose’ (especially prominent in rods) is not observed in the response of the membrane current at the outer segment. The nose reflects the effect of a hyperpolarization-activated cation current (Ih) carried by both Na+ and K+, with a reversal potential close to 0 mV. This Ih current is situated at the inner segment and cell body, which explains why it is not detected by the suction-pipette recording. The suction-pipette recording method therefore monitors the phototransduction event in the outer segment more faithfully, without much contamination by the signal from other parts of the cell. This recording method has been tremendously useful in examining the details of phototransduction.

With increasing flash intensity, the response increases progressively until saturation is reached. When the transient peak amplitude of the flash response is plotted against the corresponding flash intensity (intensity–response relation), a relation fairly close to the Michaelis relation in enzyme kinetics is found, especially with cold-blooded animals, namely, R = Rmax[IF/(IF + σ)], where R is the transient peak amplitude of the response, Rmax is the saturated response amplitude, IF is the flash intensity, and σ is the half-saturating flash intensity. While simple, this relation does not really convey the mechanism underlying the generation of the light response. In other words, the relation is purely empirical. For one thing, with increasing flash intensity, the transient peak of the response moves earlier in time (i.e., the time between flash and transient peak of the response, or time-to-peak, gets shorter), due to the decay phase of the response cutting into the rising phase earlier. This is a sign of active adaptation by the cell to progressively brighter illumination. If a long step of light is used for stimulation instead, the response rises to a transient peak and then relaxes to a lower plateau level, with the response relaxation again reflecting adaptation of the cell to steady light. If the flash–response amplitude in Figure 2 is measured instead at a ‘fixed’ early time in the rising phase of the response, before any light adaptation sets in, the flash intensity–response relation is described instead by the relation R = Rmax[1 – exp(–IF/ρ)], where the half-saturating flash intensity is given by ρ(loge 2). This relation bears a simple mechanistic interpretation, namely that each absorbed photon activates a spatially restricted region of the outer segment, within which the transduction effect essentially reaches saturation (i.e., all dark current in this region is suppressed).

The rod and cone photoresponses are detectable as the a-wave in the electroretinogram (ERG), which is a mass recording from the retina with an electrode placed near the cornea of the intact eye. However, the a-wave does not provide the full time course of the photoreceptor light response because the b-wave (which largely originates from the activity of bipolar cells, especially ON-bipolar cells) of the ERG, which immediately follows the a-wave, obscures the decline phase of the a-wave. The ERG nonetheless provides a very useful clinical tool for assaying rod and cone function in human patients.

II Anatomy of the Inner Ear

The mammalian inner ear is composed of two diverse functional parts, the cochlea. which is the auditory portion of the ear, and the vestibule, which functions in detecting gravity and linear and rotational motion required for balance (Fig. 1). The sensory receptor cell that performs these many diverse functions is the hair cell. Hair cells. along with their associated supporting cells, make up a sensory patch. and six different sensory patches are located throughout the ear. Within the cochlea lies a single sensory patch. the organ of Corti, which is responsible for transducing sound waves into neuronal impulses. Two different types of hair cells are found in this organ, inner hair cells (IHCs) and outer hair cells (OHCs). which differ in both morphology and function (Fig. 1D), By sending information back to the brain. IHCs act as the traditional receptor cells in the organ of Corti. In contrast, the OHCs are motile cells that primarily receive in- put from the brain and are thought to function as a cochlear amplifier (Dallos, 1992; Davis, 1983).

Within the vestibule there are five sensory organs that can be of two types: cristae or maculae (Figs. 1B and 1C). Cristae are humplike organs that lie at the base of each of the three semicircular canals. Maculae, of which there are two–saccular and utricular–are fiat organs located in the central region of the vestibule. As in the auditory system. two types of hair cells are found in the vestibular organs, type I and type II (shown in Fig. 1C), although the functional significance of having these two cell types is not yet clear.

In addition to being located within the skull. the inner ear is itself encased in a bony shell known as the otic capsule. Within the enclosed epithelial compartments of the ear is the endolymph, a specialized fluid with the unusual ionic composition of high [K−] and low [Na −], an essential medium for normal hair cell transduction. Between the otic capsule and the central endolymphatic compartments lies a called perilymph, which has low [K+] and high [Na+], similar to most extracellular fluids.

Molecular Receptors for Taste Stimuli

The molecular basis for detecting taste stimuli has been determined, in part, for sweet, bitter, and umami, but details remain elusive for sour and salty. Sweet, bitter, and umami are transduced by GPCRs embedded in the apical chemosensitive tips of gustatory sensory receptor cells (∼type II cells). GPCRs for these tastes have been identified and are unique to gustatory sensory receptor cells. The GPCRs for sweet include the heteromeric dimer T1R2 plus T1R3. There may be additional sweet receptors, as suggested by ongoing investigations. Umami is transduced by the dimer T1R1 plus T1R3 and by additional receptors that have not been completely characterized. These additional receptors may include GPCRs that are modified synaptic metabotropic glutamate receptors (mGluRs), specifically a modified mGluR4 and mGluR1. Molecular receptors for bitter are named T2Rs. A family of approximately 25 T2R genes transduces bitter substances in the human. Current research is aimed at identifying the ligands for each of the T2R GPCRs. It is highly likely that there are additional, less-explored molecular mechanisms for bitter taste, including direct actions of bitter ligands on ion channels.

The molecular transducers for salty and sour (acid) tastes have eluded identification to date. There is considerable evidence that ion channels, not GPCRs, subserve these qualities and that the amiloride-sensitive epithelial sodium channel transduces Na+ taste, at least in rodents. Even less is known about mechanisms for acid taste. Recent research has implicated TRP-like channels, specifically PKD1L3 and PKD2L1, in contributing to sour taste transduction but the evidence is not yet conclusive. It is important to note that the proximate stimulus for sour (acid) taste appears to be an increase in intracellular (not extracellular) protons, that is, intracellular acidification. Accordingly, solutions of organic acids such as acetic and citric acids are more intensely sour than are mineral acids such as hydrochloric acid when all are presented at the same pH. The protonated (undissociated) molecules of organic acids (HA) readily penetrate the cell membrane and dissociate inside the cell to lower the cytosolic pH. In contrast, the extracellular protons of a fully dissociated acid (HCl → H+ + Cl−) only penetrate the plasma membrane (Figure 4). Intracellular acidification occurs throughout the taste bud and possibly the entire lingual epithelium during sour stimulation, but only a subpopulation of acid-sensitive taste sensory cells respond to this intracellular pH shift by activating transduction mechanisms specific for sour taste. Sour-responsive taste cells have recently been identified as type III taste cells, but the details of how acid taste is transduce by these cells is not yet known. It is not known what characterizes this population of acid-sensitive sensory cells or how they signal sour taste.

The principal receptor for sucrose and other sugars appears to be a dimer of class C GPCRs: T1R2 plus T1R3. Class C GPCRs are characterized by lengthy extracellular N-termini which form a hinged binding pocket, akin to the carnivorous leaves on the Venus fly trap plant; hence the name for this portion of the receptor is the Venus fly trap domain. This motif seems to be the primary binding site for sucrose and glucose. However, there are at least two other binding pockets elsewhere in the T1R2–T1R3 dimer, including a site for the artificial sweetener cyclamate, and another for the sweet protein brazzein (Figure 5). The cyclamate pocket is tucked between transmembrane (TM) regions 3, 5, and 6 of T1R3 and situated within the TM depths of the molecule, as is found for the binding pocket of other GPCRs, such as the β adrenergic receptor. The binding site for brazzein is found in a cysteine-rich region near where the large extracellular N-terminus joins the first TM domain of T1R3. Considerably less is known about the structural details of GPCRs for umami and bitter.

Sensory Afferent Fibers and Information Coding

Sensory afferent fibers running in the gustatory nerves summarized above distribute themselves throughout the oral cavity and penetrate individual taste buds at their bases at the boundary between the epithelium and the lamina propria. Studies in mice have shown that axons from relatively few (four or five) sensory ganglion cells innervate a single taste bud, at least for taste buds in fungiform papillae and sensory cells in the geniculate ganglion. Thus, several taste bud cells must converge onto a single sensory afferent fiber insofar as there may be up to 100 cells in a taste bud.

As stated earlier, the only synapses that are recognized at the electron microscopic level in taste buds are formed between type III synaptic cells and sensory afferent fibers. Type II gustatory sensory receptor cells make no morphologically identifiable synapses with sensory afferent fibers but secrete the neurotransmitter ATP at presumed en passant synapses with afferent fibers. The best evidence is that type II gustatory sensory receptor cells release ATP via unconventional, nonvesicular synaptic mechanisms. Specifically, ATP is secreted through large bore membrane ion channels that are triggered open by increases in intracellular Ca2+ in combination with membrane depolarization. These channels are believed to be comprised of the calcium homeostasis modulator 1 (CALHM1). ATP stimulates postsynaptic P2X2 and P2X3 receptors on primary sensory afferent fibers. Such synaptic interactions would not necessarily require recognizable structures, thus explaining the lack of synapses observed in electron micrographs of taste buds.

ATP secreted from type II cells also stimulates adjacent type III synaptic cells within the taste bud and excites them to release serotonin (Fig. 3). The function of these latter cells is under scrutiny, and the role of biogenic aminergic neurotransmitters in information flow within taste buds is currently being investigated.

The lack of discrete synaptic connections between gustatory receptor cells and sensory afferent fibers may help explain why many afferent fibers respond to two or more taste qualities, such as sweet plus bitter. That is, when the electrical activity in afferent fibers is recorded during taste stimulation, one often observes that individual fibers respond to two or more taste qualities. Such a recording might arise from a sensory afferent that terminates in the taste bud near a sweet-sensing gustatory receptor cell as well as near a bitter-responsive taste cell. The afferent fiber will be excited by activation of the sweet-sensitive taste cell, the bitter-sensitive taste cell, or both. How strongly the fiber responds to sweet versus bitter would be governed by the proximity of the fiber to the sweet- or bitter-responsive taste cell.

Such a lack of narrow tuning in individual sensory afferent fibers to taste quality (sweet, sour, salty, etc.) has given rise to the concept that information about taste is encoded by the concurrent excitation of multiple sensory afferent fibers. According to this scheme, information encoded by activity in any given single fiber would be ambiguous because a single fiber can be excited by more than one taste stimulus. However, when activity of several fibers is integrated—so-called combinatorial coding or cross-fiber coding—coherent information would emerge. Many investigators believe that this is how taste information is coded and transmitted to the brain, in contrast to the notion that taste is transmitted by highly-tuned, individual, specific lines for sweet, sour, bitter, and so forth—a notion called the “labeled line” theory of coding. Labeled line coding is not consistent with the large number of multiple-responding taste afferent fibers observed experimentally.

Development of Tastebuds on the Oral Jaws

The larger jaws of the cavefish result in larger mouths to maximize the opportunities to capture food. With their larger mouths, there is more space on the lips for expansion of the gustatory system, namely tastebuds. Schemmel (1967) showed that the numbers of tastebuds in cavefish were five- to seven-fold greater compared to the sighted morph. Furthermore, in the sighted morph, tastebuds are mainly found in the labial epithelium, whereas in the blind morph, they are also found in the skin of the maxilla, lower jaw, and ventral aspect of the head (Boudriot and Reutter, 2001).

Tastebuds in teleosts are typically pear- or onion-shaped intraepithelial sensory structures that consist of multiple cell types (light and dark receptor cells, basal cells, Merkel cells; Figure 11.4(A)). Surrounding the tastebuds but not part of the organ sensu strictu are marginal cells; these cells are situated between the sensory cells of the tastebud and the squamous epithelium of the skin. The light and dark sensory receptor cells make up the bulk of the tastebud. The apical ends of these receptor cells protrude into the oral cavity or the external environment, and provide feedback to the chemoreception area of the telencephalon. Although Schemmel (1980) showed that the morphology of the tastebuds do not differ between the two morphs, more recent evidence has shown using immunohistochemistry that the tastebuds in cavefish larvae contain more receptor cells and are innervated by more axon profiles (Varatharasan et al., 2009).

Tastebuds begin their development in the same manner as teeth do in the form of epithelial primordia. Soon after the induced epithelium has evaginated above the level of the rest of the epithelial tissue, the tastebud will enter the differentiation stage of development (Northcutt, 2004). The fully developed tastebud is buried within the epithelium, and only small, chemosensory microvilli are exposed to the superficial environment in the oropharyngeal cavity (Boudriot and Reutter, 2001). At the base of the tastebud is the nerve fiber plexus, where the tastebud is innervated (Torrey, 1940). At 5 dpf in both cavefish and surface morphs of A. mexicanus, fish have one row of tastebuds on the lips (Varatharasan et al., 2009; Figure 11.4(B)). A second (inner) row starts developing at 12 dpf from the central midpoint of the jaw, in the area of the mandibular symphysis. Varatharasan et al. (2009) reported that this second row of tastebuds appears quicker in the cavefish lower jaw than at similarly aged surface fish (Figure 11.4), possibly suggesting that there is a more rapid induction of tastebuds in the cavefish. By 22 dpf, there are significantly more tastebuds in and around the cavefish mouth (Varatharasan et al., 2009; Figure 11.4(B)). Thus there is an expansion of the gustatory system in both numbers of tastebuds and receptor cells, and in their innervation in cavefish.