Taste receptor cells (TRCs) respond to acid stimulation, initiating perception of sour taste. Paradoxically, the pH of weak acidic stimuli correlates poorly with the perception of their sourness. A fundamental issue surrounding sour taste reception is the identity of the sour stimulus. We tested the hypothesis that acids induce sour taste perception by penetrating plasma membranes as H+ ions or as undissociated molecules and decreasing the intracellular pH (pHi) of TRCs. Our data suggest that taste nerve responses to weak acids (acetic acid and CO2) are independent of stimulus pH but strongly correlate with the intracellular acidification of polarized TRCs. Taste nerve responses to CO2 were voltage sensitive and were blocked with MK-417, a specific blocker of carbonic anhydrase. Strong acids (HCl) decrease pHi in a subset of TRCs that contain a pathway for H+ entry. Both the apical membrane and the paracellular shunt pathway restrict H+ entry such that a large decrease in apical pH is translated into a relatively small change in TRC pHi within the physiological range. We conclude that a decrease in TRC pHi is the proximate stimulus in rat sour taste transduction.
- hydrogen ion conductance
- intracellular signaling
- chorda tympani
- acid stimulation
sourness is a primary taste sensation (25). Stimuli evoking a sour sensation yield dissociable H+ ions. It is natural to assume that taste receptor cells (TRCs) are extracellular pH (pHo) detectors, and that sourness should, therefore, be a graded function of stimulus pH. However, this is not generally true. The poor correlation between sourness and stimulus pH has been amply demonstrated in both human (1, 12, 17, 21, 24, 29) and animal (2, 3, 27) studies. At the same pH, acetic acid is a more potent sour stimulus than HCl (12, 27, 29). Moreover, the sourness of acetic acid (and other weak acids) is essentially the same as that of a buffer consisting of the acid plus its conjugate base, even though the latter has a higher pH (2,12). It has been speculated that weak acids might induce sour taste perception by penetrating plasma membranes as undissociated molecules and decreasing intracellular pH (pHi) upon dissociating within the TRCs (13, 27, 35). We have previously demonstrated that isolated rat (23) and hamster (34) TRCs respond to changes in pHo with parallel changes in pHi. This suggests that changes in TRC pHi may indeed be the appropriate stimulus variable for sour taste transduction. If so, it would explain the paradoxical correlation between stimulus pH and sourness. However, before a firm conclusion can be drawn regarding the role of pHi in acid sensing, measurements of TRC pHi must be made in an intact taste bud maintained with its natural polarity, and a direct correlation must be established between changes in pHi of polarized TRCs and the taste nerve responses to acids. Finally, pharmacological agents that specifically inhibit the normal acid-induced decrease in pHi should also inhibit the acid-induced neural response. We have made these measurements, and the results demonstrate unequivocally that changes in pHi in TRCs represent the proximate sour taste stimulus in rat.
MATERIALS AND METHODS
Chorda tympani recording.
Female Sprague-Dawley rats (150–200 g) were anesthetized by intraperitoneal injection of pentobarbital sodium (60 mg/kg), and supplemental pentobarbital (60 mg/kg) was administered as necessary to maintain surgical anesthesia. Body temperatures were maintained at 36–37°C with a circulating-water heating pad. The left chorda tympani (CT) nerve was exposed laterally as it exited the tympanic bulla (9, 34, 39) and placed onto a 32-gauge platinum-iridium wire electrode. An indifferent electrode was placed in nearby tissue. Neural responses were differentially amplified with a custom-built, optically coupled isolation amplifier. For display, responses were filtered with the use of a band-pass filter with cut-off frequencies of 40 Hz-3 kHz and fed to an oscilloscope. Responses were then full-wave rectified and integrated with a time constant of 1 s. Integrated neural responses and current and voltage records were recorded on a Linseis TYP7045 chart recorder and also captured on disk with Labview software and were analyzed off-line (40). Stimulus solutions were injected into a Lucite chamber (3 ml; 1 ml/s) affixed by vacuum to a 28-mm2 patch of anterior dorsal lingual surface. The chamber was fitted with separate Ag-AgCl electrodes for measurement of current and potential. These electrodes served as inputs to a voltage-current clamp amplifier that permitted the recording of neural responses with the chemically stimulated receptive field under current or voltage clamp (39,40). The clamp voltages were referenced to the mucosal side of the tongue. The anterior lingual surface was stimulated with a rinse solution (10 mM KCl) and with acetic acid, citric acid, or HCl solutions at pH 3.0 or with solutions containing 10 mM of each of the above acids. To measure the CT response to dissolved CO2, we stimulated the lingual surface with a solution containing 72 mM KHCO3 buffered to pH 7.4 with a 10% CO2-90% O2 mixture. In these experiments the rinse solutions contained 72 mM KCl buffered to pH 7.4 with 10 mM HEPES. To evaluate the role of carbonic anhydrase in dissolved-CO2-induced CT responses, we studied the effect of topical application of 50 mM MK-417 (Merck, Rahway, NJ), a cell-permeable blocker of carbonic anhydrase (11), on CT responses. In some experiments the CT responses were recorded at the slower perfusion rate of 1 ml/min to obtain flow rates comparable to those required in pHimeasurements in isolated single fungiform papillae.
Rats were anesthetized with methoxyflurane and killed by cervical dislocation. The tongues were rapidly removed and stored in ice-cold Ringer solution, containing (in mM) 140 NaCl, 5 KCl, 1 CaCl2, 1 MgCl2, 10 Na-pyruvate, 10 glucose, and 10 HEPES, pH 7.4. The lingual epithelium was isolated by collagenase treatment (23, 34). A small piece of the anterior lingual epithelium containing a single fungiform papilla (Fig.1, a–d) was mounted in a special microscopy chamber (7). The tissue was intermittently perfused with Ringer solution containing 25 μM of the pH-sensitive fluoroprobe 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein-AM (Molecular Probes, Eugene, OR) at 4°C for 2 h. Before the experiment was started, the tissue was perfused on both sides with control solution for 15 min. The control solution was Ringer solution without Na-pyruvate. The tissue was perfused at the rate of 1 ml/min. The TRCs in the taste bud were visualized from the basolateral side through a ×40 objective (Zeiss; 0.9 NA) with a Zeiss Axioskop microscope and imaged with a setup consisting of a cooled charge-coupled device camera (Imago, TILL Photonics Applied Scientific Instrumentation, Eugene, OR) attached to an image intensifier (model VS4–1845; Videoscope, Washington, DC), an epifluorescent light source (TILL Photonics Polychrome IV), a 515-nm dichroic beam splitter (Omega Optical), and a 535-nm emission filter (20-nm band pass; Omega Optical). The cells were alternately excited at 490 and 440 nm and imaged at 10-s intervals. Small regions of interest (ROIs) in the taste bud (diameter 2–3 μm2) were chosen in which the changes in fluorescence intensity ratio (F490/F440) were analyzed using TILLvisION v3.1 imaging software. The background and autofluorescence at 490 and 440 nm were corrected from images of a taste bud without the dye. The changes in TRC pHi were calibrated by bilateral perfusion of high-K+ calibrating solutions between pH 6.5 and 8.0 containing 10 μM nigericin (23, 34). All experiments were performed at room temperature (22 ± 1°C).
Stimulus solutions were without HEPES and contained HCl, acetic acid, citric acid, and tartaric acid at the concentrations given in the text. Stimulus solution containing 8.9 mM acetic acid (pH 3.2) was titrated to pH 6.0 with 155 mM potassium acetate (155 mM potassium acetate replaced 150 mM NaCl and 5 mM KCl in the solution). Bicarbonate-buffered solution contained (in mM) 78 NaCl, 5 KCl, 1 CaCl2, 1 MgCl2, 10 glucose, and 72 NaHCO3 and was bubbled with a 10% CO2-90% O2 mixture (pH 7.4). In some experiments both HEPES- and bicarbonate-buffered solutions contained either 5 or 50 mM MK-417, a membrane-permeable blocker of carbonic anhydrases (11).
Results are presented as means ± SE of the number of ROIs in the taste bud. Student's t-test was employed to analyze the differences between sets of data.
Effect of acid stimulation on CT nerve activity.
The lingual application of acid stimuli increases CT nerve activity (Fig. 2). At pH 3.0 (Fig. 2 A), acetic acid gave a larger CT response than either citric acid or HCl, indicating that acetic acid is a stronger sour taste stimulus than citric acid and HCl (2, 12, 24, 27). In contrast, when each acid was applied at a concentration of 10 mM, all three acids produced CT responses of similar magnitude even though the pH values of the acid solutions varied between 2.0 and 3.4 (Fig. 2 B). Thus CT responses are not graded functions of the lingual surface pH (pHo) (2, 12, 24, 27).
Effect of acid stimulation on TRC pHi.
Acidic stimuli were applied to the apical side while changes were monitored in TRC pHi in situ from the basolateral side. Acidic stimuli decreased TRC pHi (Fig.3, A–C). Acetic acid at pH 3.1 induced greater decreases in TRC pHi than citric acid, tartaric acid (Fig. 3 A), and HCl (Fig. 3 B). In contrast, the changes in pHi produced by individual acids at 10 mM concentration were of similar magnitude (Fig.3 C). In each case, therefore, the decrease in pHi correlates with the respective taste neural responses, while pHo does not.
At pH 3.0, the acetic acid concentration is 58.3 mM. Of this, 57.3 mM is undissociated acid. Citric acid, a much stronger acid, has a concentration of 2.2 mM, of which 1.19 mM is undissociated. The corresponding parameters for tartaric acid are similar to those for citric acid. At pH 3.0, acetic acid induced a greater decrease in TRC pHi (Fig. 3 A) and a greater CT response (Fig.2 A) for two reasons: 1) it is present at higher concentration than either citric acid or tartaric acid, and2) a much higher proportion of the acetic acid is present as the membrane-permeable undissociated form (13). At 10 mM concentration, acetic acid has a pH of 3.4. Of this, 9.59 mM is undissociated, while 10 mM citric acid has a pH of 2.6 and 7.49 mM is undissociated. The concentrations of the undissociated forms of acetic acid and citric acid in the stimulus solutions are comparable and, therefore, induce comparable changes in pHi (Fig.3 C) and CT responses (Fig. 2 B) for both acids. Within the normal range of pHi (Fig. 3, A andC), diffusing citric acid delivers three equivalents of H+ ions to the TRCs. This suggests that the intrinsic permeability of acetic acid is about three times that of citric acid.
Effect of pHo on acetic acid responses.
If the permeability of the undissociated form of a weak acid is the critical factor in determining its intensity as a stimulus, the changes in pHi and the CT response should both be independent of pHo. To test this hypothesis, we obtained the CT responses to unbuffered acetic acid (pH ∼3) and to the same concentration of acetic acid buffered at pH 6 with potassium acetate. At a given concentration, the buffered (pH 6.0) and unbuffered acetic acid (pH 3.0) gave similar responses (Fig.4 A). In parallel experiments, perfusing solutions containing 8.9 mM acetic acid plus 150 mM potassium acetate (pH 5.9) and 8.9 mM acetic acid alone (pH 3.2) across the apical side of a single fungiform papilla induced changes in TRC pHi of similar magnitude (Fig. 4 B). Thus the CT response and the changes in TRC pHi are independent of pHo and the acetate concentration. Both the CT response and the changes in TRC pHi were found to be independent of apical Na+ and were unaffected by 1 mM α-cyano-4-hydroxycinnamate (4), a blocker of monocarboxylate transporter I (unpublished observations). The data suggest that undissociated acetic acid molecules enter TRCs by passive diffusion across the apical membranes (13).
Effect of CO2 on CT nerve activity.
Initially the lingual surface was suffused with a rinse solution containing 72 mM KCl buffered to pH 7.4 with 10 mM HEPES, and the CT activity was taken as the baseline activity. Suffusing the lingual surface with a solution containing 72 mM KHCO3 buffered to pH 7.4 with 10% CO2-90% O2 produced reversible increases in CT activity (Fig.5 A). Cell membranes are freely permeable to dissolved CO2, and its conversion to H2CO3 catalyzed by intracellular carbonic anhydrase (8, 16) represents penetration of acid equivalents across TRC membranes (23, 34). This was tested directly by recording CT responses in the presence of MK-417, a specific blocker of carbonic anhydrases (11). We applied 50 mM MK-417 topically on the lingual surface in the rinse solution for 15 min. This inhibited the dissolved-CO2-induced CT responses by ∼50% (Fig. 5 A, middle). Upon washout, the effects of MK-417 were completely reversible (Fig.5 A, right). In a similar study (20), topical application of MK-927, another membrane-permeable carbonic anhydrase inhibitor, to the rat tongue inhibited CT responses to carbonated water by 62%.
Voltage dependence of CO2 induced CT nerve activity.
Data summarized in Fig. 5 B indicate that unlike CT responses to HCl (9, 34) and acetic acid (data not shown), the CT responses induced by dissolved CO2 are voltage sensitive. Compared with open-circuit conditions (i.e., under 0 current clamp), the CT responses were enhanced at −60 mV and suppressed at +60 mV, similar to voltage-clamp effects on responses to NaCl (39,40).
Effect of CO2 on TRC pHi.
The generation of intracellular H+ ion was confirmed directly by the observed decrease in TRC pHi produced by perfusing the lingual surface with a solution containing 72 mM NaHCO3 buffered to pH 7.4 with 10% CO2-90% O2 (Fig. 6 A). The changes in TRC pHi were significantly attenuated in the presence of 5 mM MK-417 (Fig. 6 B). In the presence of 50 mM MK-417, dissolved CO2 induced a maximum decrease in pHi of 0.05 ± 0.008 pH unit. These data suggest a direct role of carbonic anhydrase in CO2-induced sour taste transduction (20). Dissolved CO2 entry across the apical membrane of TRCs is independent of pHo and the HCO concentration in the stimulus solution. The fact that this acidic stimulus can be presented at an alkaline pHo is further evidence that the proximate acidic stimulus is the change in pHi and not pHo. Fully dissociated strong acids, such as HCl (Fig. 1), HNO3, and H2SO4, are also potent stimuli of the taste nerves (2, 3, 9, 24, 30) and decrease TRC pHi(Fig. 3, B and C), suggesting that H+ions also gain entry into TRCs but, in all probability, not by diffusing through the membrane lipid bilayer like dissolved CO2 and the neutral forms of weak acids.
Temporal relation between CT activity and TRC pHi.
Consistent with our previous studies (9, 34, 39, 40), CT nerve activity was monitored in vivo at normal physiological temperature while the lingual surface was suffused with stimulating solutions at the rate of 1 ml/s. In contrast, our in vitro pHi measurements were made in a small microscopy chamber at room temperature where the maximum rates of perfusion were 1 ml/min. It is desirable to control for these differences in the rate of stimulus application when comparing the CT recordings with the changes in TRC pHi. It was not possible, for technical reasons, to increase the flow rate in vitro beyond 1 ml/min. Therefore, we made some CT recordings with stimulus and rinse applied at 1 ml/min. Consistent with previous studies (10, 31), the phasic parts of the CT responses were strongly influenced by the flow rate (Fig. 7); however, the magnitude of the maximum CT response to either dissolved CO2 (Fig.7 A) or HCl at pH 3.0 (Fig. 7 B) was not affected by the flow rate. At the flow rate of 1 ml/min both acid stimuli, dissolved CO2 (Fig. 8,A and B) and HCl (Fig. 8, C andD) induced CT-response profiles that were similar to the pHi changes observed in vitro. The data show that under the additional constraint of comparable flow rates, the CT nerve response and the changes in TRC pHi are temporally well correlated.
Regional differences in HCl- and acetic acid-induced changes in pH within a taste bud.
No CT responses were observed with HCl at pH 5.0 (data not shown). A decrease in pHo from 7.4 to 5.0 induced an average decrease in TRC pHi of 0.11 ± 0.003 pH unit in a taste bud (n = 12). We compared the regional changes in TRC pHi in three taste buds exposed to apical acetic acid or HCl at pH 3.0. All 58 ROIs in taste buds responded with a decrease in pHi >0.11 pH unit. The mean decrease in pHiproduced by acetic acid (ΔpHi = 0.40 ± 0.01) was significantly greater than that produced by HCl (ΔpHi = 0.29 ± 0.01; P < 0.001; paired). In 56 ROIs, acetic acid produced a greater decrease in pHi than HCl did. The changes in pHi varied widely within ROIs. After HCl treatment (Fig.9 A), 35 of 58 ROIs responded with a decrease in pHi between 0.2 and 0.299 pH unit, and 20 of 58 ROIs (34.5%) responded with a decrease in pHi>0.3 pH unit. It is likely that TRCs in the ROIs that respond with the greatest decrease in pHi participate most in sour transduction. In contrast, after acetic acid treatment, 5 of 58 ROIs responded with a decrease in pHi between 0.2 and 0.299 pH unit, and 53 of 58 ROIs (91.4%) changed pHi >0.3 pH unit (Fig. 9 B). The data suggest that HCl-induced CT responses are elicited by a subpopulation of TRCs contained in different ROIs within the taste bud that contain either an H+ entry mechanism in their apical membrane or in their basolateral membrane accessible via paracellular shunts. In contrast for acetic acid, the influx of acid equivalents into TRCs is augmented by a significant flow of unionized acetic acid. This would account for the significantly greater decrease in TRC pHi, consistent with the greater CT response.
Our results show that weak acids enter rat TRCs from the apical side as neutral molecules (acetic acid or dissolved CO2). H+ ions (HCl) also gain access to TRCs, but the pathways involved have not yet been fully elucidated. Irrespective of the mode of acid entry, however, the data indicate that the consequent decrease in TRC pHi serves as the proximate stimulus in sour taste transduction.
pHi as the proximate stimulus in acid detection.
The hypothesis that a decrease in TRC pHi is the proximate stimulus in rat sour taste transduction is borne out by our studies with acetic acid stimuli at pH 6.0 and dissolved CO2 as an acid stimulus. Even at pH 6.0, a pHo that is significantly above the threshold pH that gives a detectable CT response, acetic acid was as good a stimulus as at pH 3.0. Acetic acid responses were completely independent of pHo and K-acetate concentration. Our data indicate that the permeability of the undissociated acetic acid is the critical factor in determining its intensity as a stimulus (13, 27). Inside the cell, acetic acid is dissociated into H+ and acetate ions. Thus a decrease in TRC pHiis the proximate stimulus for acetic acid-induced increase in CT responses at pH 6.0.
At constant pHo, dissolved CO2 decreased TRC pHi and increased CT responses. A membrane-permeable carbonic anhydrase blocker, MK-417, attenuated CT responses and inhibited changes in TRC pHi. It is interesting to note that unlike CT responses to HCl (9, 34) and acetic acid (data not shown), the CO2 responses were voltage sensitive. Inside the cell, the dissolved CO2 is converted to H2CO3 in a reaction catalyzed by intracellular carbonic anhydrases. The H2CO3 yields free H+ and HCO ions. It is likely that HCO ions exit TRCs via conductive pathways and/or ion exchangers in the cell membranes (unpublished observations). Because CO2 permeability across cell membranes is not voltage sensitive, voltage changes most likely affect CT responses indirectly by modulating the HCO flux across TRC membranes. Because the hydration of CO2 leads to the formation of H+ and HCO in stoichiometrically equal proportions, alteration in the concentration of the latter under voltage-clamp conditions will produce immediate changes in pHi. If a decrease in pHi is the necessary precursor to a sour taste response, the time course of the CT response ought to follow that of the pHi when the sour stimulus is applied at a given fixed rate. The optimal means of making this determination would be to measure both pHi and the neural response in the same system simultaneously. For technical reasons, this was not possible. However, when we obtained the pHi profiles and the CT response for both HCl and CO2 in separate experiments but at the same stimulus flow rate, their time courses were quite comparable (cf. Fig. 8). This result is consistent with the conclusion that the change in pHi is the actual rate-limiting stimulus in acid taste transduction.
Similar studies in other mammalian pH-sensors have also identified pHi as the proximate stimulus in acid detection. A decrease in pHi induced by elevated levels of CO2depolarized pH-sensing neurons (28, 30, 38) and increased action potential frequency. In cultured ventrolateral medullary neurons  during NH4Cl pulses a decrease in pHi led to increased action potential frequency. A decrease in pHi induced by blocking the neuron membrane Na+/H+ exchanger resulted in increased cell excitability (38). Acid detection by carotid body type I cells also depends on a decrease in pHi (5,6). It appears, therefore, that mammalian pH sensory cells have the common property of responding to changes in pHi as the proximate stimulus in acid detection.
Proton interactions with TRCs.
Our data indicate that H+ ions rapidly enter TRCs. Within the taste bud, a subpopulation of TRCs appears to contain an H+ entry mechanism either in the apical membranes or in the basolateral membranes made accessible via paracellular shunts. During acid stimulation, this subset of TRCs responds with greater changes in pHi and most likely participates in CT responses elicited by strong, fully dissociated mineral acids. However, the exact nature of the H+ entry mechanism remains unknown principally because the effect of H+ ions on TRC membrane conductances varies considerably among species (9, 22, 26, 33). H+ entry through apical amiloride-sensitive epithelial Na+ channels (ENaC) has been suggested as a possible mechanism for acid transduction (14, 22, 33). However, HCl-induced CT responses are not affected by apical Na+, amiloride, or changes in the transepithelial potential of the receptive field under stimulation (9, 34). Similarly, HCl-induced changes in TRC pHi were unaffected by apical Na+, amiloride, or changes in basolateral K+concentration (unpublished observations). In mouse TRCs, acid responses were insensitive to amiloride but were blocked by 5-nitro-2-(3-phenylpropyl-amino)benzoic acid, a Cl−channel blocker (26). Both the mammalian brain Na+ channel (BNC1) (36) and the HCN family of channels (32) have been shown to be present in rat vallate TRCs. However, recent evidence suggests that BNC1 is most likely not involved in sour taste transduction (19). Alternatively, H+ ions may pass through paracellular shunts to reach TRC basolateral membrane sites (9). These sites could belong to the HCN channel family.
pHi regulation and pH tracking in TRCs.
The acid-induced changes in TRC pHi were sustained. A sustained change in pHi is a common feature of acid-sensing cells (30, 37, 38). This, however, does not imply that TRCs lack the ability to regulate pHi. We have previously (23, 34) shown that at constant pHo, isolated TRCs recover from intracellular acid loading. It is likely that during acid stimulation, the inhibition of pH recovery mechanisms contributes to a sustained decrease in pHi (Ref. 30 and unpublished observations). This ensures that for a given acid stimulus, the induced pHi changes will be maximal and graded.
In contrast to the apical membrane, the relationship between pHi and pHo across the basolateral membrane gave a slope of 0.69 (unpublished observations). A similar relationship between pHi and pHo was observed in isolated TRCs (23, 34). This suggests that the apical membrane of TRCs is significantly less conductive to H+ ions relative to the basolateral membrane and acts as a filter for H+ions. In concert with the paracellular shunt pathway (9,34), apical cell membranes of TRCs help to regulate H+ concentration, thereby protecting the sensory apparatus from hyperacidic conditions. Consequently, a large change in apical pHo is translated into a relatively small change in pHi. This ensures that variations in TRC pHiremain within the physiological range during the rigors of acid stimulation (5, 30, 38).
Role of pHi in transduction.
In the case of rat acid-sensitive taste receptors, transduction steps subsequent to the decrease in pHi are still unknown. In carotid body type I cells, decrease in pHi leads to a conductance increase, membrane depolarization, and an increase in intracellular Ca2+ concentration (5). In the acid-sensors in the locus coeruleus, the proximate stimulus is still a decrease in intracellular pH, but in this case depolarization results from proton-induced closure of K+ channels (28). Divergence in the subsequent transduction steps in these cases suggests the possibility of alternate mechanisms leading to depolarization in taste cells. To some extent, one should anticipate this given the diversity of mechanisms that have been proposed for sour taste transduction in various species. For example, in the case ofNecturus, a decrease in apical pHo depolarizes taste cells by closing apical K+ channels (18). Our preliminary data (unpublished observations) suggest that TRC pHi is regulated via intracellular second messengers (e.g., intracellular Ca2+ concentration and cAMP). This suggests that, similar to other sensory cells, sustained changes in TRC pHi during acid transduction must involve activation of intracellular signaling mechanisms that result in depolarization of TRC receptor potential and release of neurotransmitter. This is consistent with the studies of Gilbertson et al. (15) where the contribution of cAMP to the development of acid-responses is also indicated.
We thank Victoria Bickel for assistance with stimulation chamber fabrication, Dr. Niley Desai for assistance with imaging, and Dr. Steven Price for helpful suggestions.
This work was supported by National Institute of Deafness and other Communications Disorders Grants DC-02422 and DC-00122 and by the Department of Veterans Affairs. During this study, V. Lyall was partially supported by the Veterans Affairs Merit Review.
Address for reprint requests and other correspondence: V. Lyall, Dept. of Physiology, Virginia Commonwealth Univ., Sanger Hall 3002, 1101 E. Marshall St., Richmond, VA 23298-0551 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society