Modeling and simulation of ion channels and action ...

Science in China Series C: Life Sciences © 2009

www.scichina.com life.scichina.com

SCIENCE IN CHINA PRESS

Article

Modeling and simulation of ion channels and action potentials in taste receptor cells CHEN PeiHua1, LIU Xiao-dong2, ZHANG Wei1, ZHOU Jun1, WANG Ping1†, YANG Wei3 & LUO JianHong3 1

Biosensor National Special Laboratory, Key Laboratory of Biomedical Engineering of Ministry of Education, Department of Biomedical Engineering, Zhejiang University, Hangzhou 310027, China; 2 Calcium Signals Lab, Department of Biomedical Engineering and Neuroscience, Johns Hopkins University, Baltimore, MD 21205, USA; 3

Department of Neurobiology, Institute for Neuroscience, Zhejiang University School of Medicine, Hangzhou 310058, China

Based on patch clamp data on the ionic currents of rat taste receptor cells, a mathematical model of mammalian taste receptor cells was constructed to simulate the action potentials of taste receptor cells and their corresponding ionic components, including voltage-gated Na+ currents and outward delayed rectifier K+ currents. Our simulations reproduced the action potentials of taste receptor cells in response to electrical stimuli or sour tastants. The kinetics of ion channels and their roles in action potentials of taste receptor cells were also analyzed. Our prototype model of single taste receptor cell and simulation results presented in this paper provide the basis for the further study of taste information processing in the gustatory system. +

+

taste receptor cell, action potential, modeling and simulation, Na channel, delayed rectifier K channel

1 Introduction Taste receptor cells are epithelial cells, bearing the properties of neurons. They can realize taste information processing, such as the chemical-electrical signal transformation in taste sensation. Studies indicated that taste receptor cells could produce action potentials, elicited by sour[1], sweet[2,3], bitter[4] and salty[5] tastants. The analysis of the features extracted from the elicited action potentials demonstrated that the firing patterns might be utilized in taste coding[6]. Recently, some similarities were found between the responses from taste receptor cells and taste nerve fibers. Thus, the action potentials produced by taste receptor cells may be the key components of taste information transferred to the taste fibers[7,8]. Overall, these pieces of evidences showed that the electrical activities from taste receptor cells play an important role in taste information processing at periphery.

Ion channel study can help reveal the function and mechanism of the excitability of taste receptor cells at the molecular level. There are several types of voltagegated ion channels on the membrane of taste receptor cells, including Na+ channels[9], L- and T-type Ca2+ channels [4,10−12] , outward rectifier K + channels [3] , transient outward K+ channels, Ca2+ activated K+ channels, inward rectifier K+ channels[13] and Cl− channels[14]. Taste receptor cells can produce action potentials upon current stimuli with varying durations and intensities. Corresponding data indicated that Na+ and Ca2+ currents depolarize the membrane potential; while the outward K+ current plays a role in repolarization and after-hyperpoReceived October 21 2008, December 19, 2008 doi: 10.1007/s11427-009-0138-9 † Corresponding author (email: [email protected]) Supported by the National Natural Science Foundation of China (Grant Nos. 30627002 and 60725102) and the National Key Basic Research Program of China (Grant No. 2009CB320303).

Citation: CHEN P H, LIU X D, ZHANG W, et al. Modeling and simulation of ion channels and action potentials in taste receptor cells. Sci China Ser C- Life Sci, 2009, 52(11): 1036-1047, doi: 10.1007/s11427-009-0138-9

larization[15]. Such evidence further shows that voltagegated ion channels are crucial in taste information encoding and processing as the molecular basis. The responses elicited by tastants from taste receptor cells can be recorded with specific electrophysiological techniques. For example, cell-attached patch clamp technology could be utilized to record the responses upon bitter, sour and salty tastants in mouse taste buds cells[4]. Extracellular recording technology recorded the action currents elicited by tastants[7,8]. However, so far we have little knowledge about taste information conveyed by firing patterns of action potentials and the corresponding ionic mechanism, partially due to the limitations of experimental tools. It has been suggested that the taste sensation system is more complicated than the other sensation system[4,16,17]. Thus, it is necessary to utilize multiple approaches, including the computational model and simulation, to study the taste function and mechanisms at cellular and sub-cellular levels. Quantitative study of taste mechanism can be traced back to the study of ion transportation across the membrane of taste receptor cells and simulations of membrane potentials upon salty tastants[18,19]. According to these models, the membrane potential of taste receptor cell was related to the intensity of the taste stimuli. However, these models were constructed based on the ion transportation in the lipid bilayer and renal epithelial, which had intrinsic limitations, hardly related with the molecular mechanism and experimental data of the taste sensation and taste information processing. With applications of electrophysiology and molecular biological technology in taste, it is feasible to construct a computational model of taste receptor cells, including ion channels and taste receptors. In this paper, we propose a computational model of taste receptor cells based on Hodgkin-Huxley models. The construction of Na+ current and outward rectifier K+ current model is described in detail. The simulations of biophysical properties of ion channels and action potentials elicited by current stimuli and sour tastants are performed. We obtain the quantitative relation between stimuli and electrical activities of single taste receptor cells. The kinetic parameters, like activation and inactivation parameters of Na+ and K+ current, under the resting and stimulated condition, are also calculated based on our model. This model provides a basis for the further study of taste receptor cells, especially the acid-sensing mechanism.

2 Modeling The membrane potential of taste receptor cells upon the external stimuli is quantified based on Hodgkin-Huxley theory (Eq. 1): I + I stim dV = − ion , (1) dt Cm where V is the potential across the membrane and Cm is the membrane capacitance. Istim is the external stimuli. Iion is the total ionic transmembrane currents, which include voltage-gated Na+ current, outward rectifier K+ current and leak current. The gating parameters of each ion channel were utilized to describe ion currents, which were obtained from ordinary differential equations[20]. The ion currents were recorded at room temperature, about 22°C. A nonlinear least-squares algorithm in MATLAB was used to fit the ion current recordings for parameter estimation. The model of taste receptor cells was created on the platform of NEURON software[21]. The kinetic properties of Na+ and K+ channels were described with NMODL module in NEURON. The activation parameter m and inactivation parameter h of voltage-gated Na+ channels were obtained from Na+ currents[9]. We got the steady-state inactivation h∞ from double pulse protocol. Inactivation time constants were derived from deinactivation (−110—−70 mV), inactivation (−50—−20 mV) and activation (−10—30 mV) protocol, respectively. Na+ currents in activation protocol were recorded from the holding potential −80 mV, and then stepped to testing levels varying from −30 to +30 mV. Thus, m0 and h∞ were close to 0. Eq. 2 was used to fit the activated Na+ current[20], where steady-state activation and activation time constants were obtained: I = GNa (V − ENa )m∞3 h0 (1 − e−t /τ m )3 e−t /τ h

(2)

Electrophysiological and molecular biological experiments indicated that Kv1.5 was the functional delayed rectifier K+ channel expressed in the anterior tongue of rats[3,22]. Hence, the formula describing ultrarapid delayed rectifier K+ channel in atrial myocyte was chosen to quantify K+ currents from taste cells of rats. The gating parameters for delayed rectifier K+ channels were also obtained from K+ currents of taste cells. In activation protocol, membrane potentials were stepped from −10 to +40 mV in 10 mV increments. m0 and h∞ for K+ channels did not approach 0. The currents from the activation protocol were fitted using Eq. 3 to

CHEN P H, et al. Sci China Ser C-Life Sci | Nov. 2009 | vol. 52 | no. 11 | 1036-1047

1037

get activation and inactivation time constants[20,23]. The peak current in activation protocol was used to quantify steady-state activation. In inactivation protocol, steady-state inactivation was derived from the normalized tail current at testing potential −40 mV. I = GK (V − EK )(m∞ − (m∞ − m0 )e

−t /τ m 3

)

(h∞ − (h∞ − h0 )e −t /τ h )

(3)

2.1 Geometry In patch clamp experiment, the membrane capacitance was around 3—6 pF[9]. In this model, we used 6 pF to quantify all the ion currents. Taste receptor cells are bipolar. The shape of the cell in the model is represented as a cylinder of about 40 µm in length and 15/π µm in diameter. The specific membrane capacitance is 1 µF/µm2, which is consistent with ref. [24]. 2.2 Standard ionic concentration Based on patch clamp experiment, the physiological condition of taste receptor cells was mimiced[3,9]. At room temperature 22°C, we set [Na+]i= 10 mmol/L, [Na+]o= 131 mmol/L, [K+]i= 140 mmol/L, and [K+]o= 5 mmol/L. According to the Nernst equation, the equilibrium potential of Na+ and K+ channels is 65.4 and −84.7 mV, respectively. 2.3 Ionic currents (1) INa: voltage-gated Na+ current. From patch clamp experiments, the mean maximum conductance is (7.29±0.29) nS (SE, n=125), ranging from 0.57 nS to 18.4 nS. Eq. 4 is used: I Na = GNa m3 h(V − ENa ) ,

(4)

where V is the membrane potential. m and h are the activation and inactivation parameters, respectively. GNa=18 nS denotes the maximum conductance of sodium channel. The Nernst potential for sodium channel, ENa=57.1 mV. The transition rates for m and h are shown below: (V + 33.4) am = − 0.135 ⋅ (V + 33.4)/ (exp( − ) − 1), 7.42 V β m = 0.075 ⋅ exp( − ), 13.2 V ah = 0.00248 ⋅ exp( − ), 35 V + 12.6 )). β h = 1.34/ (1 + exp( − 8.57 m∞ = α m /(α m + β m ) h∞ = α h /(α h + β h ),

τ m = 1/(am + β m ), τ h = 1/(ah + β h ) 1038

(2) IK: delayed rectifier K+ current. We adopted the formula of ultrarapid delayed rectifier K+ channel in atrial myocyte to characterize the biophysical properties of K+ current in taste cells (Eq. 5). I K = GK ua 3ui (V − EK ) ,

(5)

where ua and ui are respectively the activation and inactivation parameters for K+ channels. The maximum conductance GK=15 nS, while EK=−70 mV. The transition rates for ua and ui are as follows: V − 16.86 aua = 0.254/ (1 + exp( − )), 17.64 V βua = 0.00854 ⋅ exp( − ), 62.17 V − 198.25 aui = 0.002227/ (1 + exp ( − )), 53.3 V − 199.2 βui = 0.00306/ (1 + exp( − )). . 53.25 ua∞ = α ua /(α ua + β ua ), ui∞ = 0.4266 + 0.5414/ (1+ exp ((V + 36.23)/ 6.166))

τ ua = 1/ (aua + βua ), τ ui = 1/ (aui + βui ) (3) Ileak: linear leak current. A linear leak (ohm) current is assumed (Eq. 6), conductance of which is the membrane conductance excluding the ion channels. From patch clamp experiment, the membrane resistance is (6.88±0.58) GΩ[15]. We set Gl about 0.145 nS and El=−45 mV. I l = Gl (V − El ) (6) (4) IPKD/IASIC: acid-sensing inward current. According to the recordings from Oocyte cells expressing ASIC2a/ASIC2b channels upon the sour stimulus pH=4.5 at the holding potential −70 mV[25], we fitted the upward and downward phase of the current with the third order polynomial equation and logistic equation, respectively. Likewise, whole cell currents from HEK293T cells expressing PKD1L3 and PKD2L1 channels were elicited by 25 mmol/L (pH=2.8) citric acid at the holding potential −60 mV[26], and described using two forth order polynomial equations. The inward current for ASIC2 channels was about 127—212 pA, while it was around 9.28—20.2 pA for PKD channels. (5)

ASIC2

current.

I up = a + b ⋅ t + c ⋅ t 2 + d ⋅ t 3 ,

I down = e/(1+ f ⋅ exp( − g ⋅ t)) a=−0.09, b=0.35, c=−0.2, d=0.012, e=−1.1, f=−0.92, g=0.0258


⎧0 t ≤ 1.635 ⎪ I ASIC = ⎨50 ⋅ I up 1.635 < t ≤ 8.66 ⎪ 8.66 < t ≤ 128.168 ⎩50 ⋅ I down

where t is the time. Iup and Idown denote the upward and downward phase, respectively. (6) PKD current. I up = a + b ⋅ t + c ⋅ t 2 + d ⋅ t 3 + e ⋅ t 4 , I down = f + g ⋅ t + h ⋅ t 2 + j ⋅ t 3 + k ⋅ t 4 a=51.14, b=−168.8, c=199, d=−99.7, e=17.88, f=−2.53, g=0.326, h=−0.0157, j=0.00033, k=−0.0000025 ⎧0 ⎪ I PKD = ⎨4.6 ⋅ I up ⎪ ⎩4.6 ⋅ I down

t ≤ 1.38 1.38 < t ≤ 1.914 1.914 < t ≤ 48.85

where t is the time. Iup and Idown are the upward and downward phase, respectively. The equivalent circuit of acid-sensing taste receptor cells is shown in Figure 1A. Protons in acidic tastants activated the acid-sensing receptors and ionic channels, and then depolarized the membrane potential. Voltage-gated Na+ and K+ channels respond to this depolarization and cooperate to produce action potentials. The taste information coded by action potentials effects Ca2+ channels, and then is transmitted

Figure 1

to the nervous system, as shown in Figure 1B.

3 Simulation results 3.1 Na+ current

To examine the simulation of sodium currents INa, we set the model parameters, GNa=13.5 nS and ENa=57.1 mV. We simulated under the voltage clamp protocols and compared with experimental results[27]. In activation protocol, the membrane potential was stepped from the holding potential of −80 mV to the testing levels, varying from −50 to +80 mV in 10 mV increments for 7 ms. The characteristic voltage-dependent sodium currents were reproduced and compared with experiment recordings (Figures 2A and B). Peaks of the activated currents at each testing level were plotted against the experimental results for comparison, as shown in Figure 2C. In inactivation double pulse protocol, the prepulse was stepped from −100 mV to −20 mV with an increment of 10 mV for 100 ms. The test pulse was −10 mV with duration 50 ms. Cells were clamped at −80 mV. During prepulse, Na+ current is small. While in test pulse, there is obvious inactivation. We set GNa=5.6 nS, ENa=56 mV. The simulation results and experimental recordings are compared in Figures 2D and E. The

The equivalent circuit. A, signal transduction pathway; B, acid-sensing taste receptor cells with ion channels.


1039

Figure 2 Simulation results of sodium currents. Comparison of activated sodium currents between simulation results: A, and experiment recordings; B, (adapted from figure 3B[9]); C, peak values of sodium currents under various testing potentials. Open circles and solid circles represent simulation and experimental results, respectively; D, the simulation results and experimental recordings; E, of inactivated sodium currents (adapted from figure 7A[9]) were compared; F, steady-state inactivation; G, the simulation results and experimental recordings; H, of deactivated sodium currents (adapted from figure 5[9]); I, peak values of tail currents under various testing potentials, abscissa: testing potential, ordinate: peak of tail currents (data were adapted from figure 5[9]).

steady-state inactivation was derived from the peak tail currents (Figure 2F). The data points from simulation and experiment were fitted with Boltzmann function. For simulation, h∞=1/(1+exp((V+52)/6.9)), while for experiment, h∞=1/(1+exp ((V+55.3)/7.46), where h∞ is the steady-state inactivation, and V is the membrane potential. In deactivation protocol, double pulse is applied. The prepulse was stepped from holding potential −80 mV to 1040

−10 mV for 0.8 ms. Test pulse was then stepped from −70 mV to +80 mV with a 10 mV increment. Here, GNa=16.8 nS, ENa=57.5 mV. The comparison between simulation and experiment is shown in Figures 2G and H. Figure 2I shows the peak tail currents at various test potentials. Linear function was used to fit the data. For simulation, Itail=6.1V- 354.3, while Itail=6.1V- 364.8 for experiment, where Itail is the peak tail current and V is the membrane potential. Both the major kinetic and


steady-sate characteristics of the simulated INa readily agreed with the experimental recordings. 3.2 Delayed rectifier K+ current

Similar simulations were performed for IK. In activation protocol, GK =20.6 nS and EK=89 mV were set. Cells were clamped at holding potential −80 mV. Membrane potentials were stepped from −80 mV to +40 mV in 10 mV increments for 400 ms. Simulation results and ex-

perimental recordings are both shown in Figures 3A and B. The peak currents at each test potential are shown in Figure 3C. In inactivation protocol, the double pulse was adopted. Prepulse was stepped from −120 mV to +40 mV in 20 mV increment for 20 s. The test pulse was +40 mV for 5 s, and then stepped back to −80 mV. We set GK=23.1 nS, EK=-89 mV. Figures 3D and E show the comparison between the simulation and experimental results. Peak tail currents at test potentials are shown

Figure 3 Simulation results of delayed rectifier potassium currents. Comparison of activated potassium currents between simulation results: A, and experimental recordings; B, (adapted from figure 3B[3]); C, peak values of potassium currents under various testing potentials. Open circles and solid circles represent simulation and experimental results, respectively; D, the simulation results and experimental recordings; E, of inactivated potassium currents (adapted from figure 1E[3]); F, peak values of tail currents in inactivation protocol; G, the simulation results and experimental recordings; H, of deactivated potassium currents (adapted from figure 1G[3]); I, peak values of tail currents under various testing potentials in deactivation protocol. CHEN P H, et al. Sci China Ser C-Life Sci | Nov. 2009 | vol. 52 | no. 11 | 1036-1047

1041

in Figure 3F. In deactivation double pulse protocol, the prepulse was stepped to +40 mV from the holding potential −80 mV for 15 ms. During the test pulse, potential was stepped from −90 mV to −20 mV with 10 mV increment. GK=16.763 nS and EK=−73.5 mV were set. The results obtained from simulation and experimental recordings are compared in Figures 3G and H. The peak tail currents under each test potential are shown in Figure 3I. 3.3

Action potential under electrical stimuli

Action potential was closely related to the intensity and duration of the electrical stimulus. The modeled taste receptor cell started to fire from the resting membrane potential of −53.91 mV. A single action potential was elicited by the current stimulus 50 pA, with a short duration 2 ms (Figure 4A). The firing threshold was about −36.9 mV. Peak value reached 15.2 mV. With a sub-threshold current stimulus of 25 pA, 2 ms, no action potential was produced. The membrane potential decreased with RC exponential decay, as the dash line shown in Figure 4A. Trains of action potentials were elicited with the long-duration current stimulus of 5.2 pA, 320 ms. The amplitude of peaks decreased, as

shown in Figure 4B. The firing patterns, when current stimuli with different intensities and different directions (depolarization current 4, 6, 8 pA; hyperpolarization current −4, −6, −8 and 0 pA for 140 ms) were applied in the resting state, are shown in Figure 4C. The firing rate and dV/dt increased as the intensity of the current stimulus increased, but the latency decreased. Thus, the taste cell inclined to fire. With the application of the hyperpolarization current stimuli, membrane potential hyperpolarized. 3.4

Currents during the action potential

At 10 ms, the current stimulus of 5.2 pA, 250 ms was applied. Action potentials were produced as shown in Figure 5A. Na+ and K+ currents at the action potentials are shown in Figures 5B and C, respectively. Na+ current depolarized the membrane potential. Delayed rectifier K+ channels opened relatively late, which played an important role in the repolarization phase. Membrane potential increased rapidly with Na+ channels opening. Due to the fast inactivation of Na+ channels and the opening of K+ channels, membrane potential decreased. The magnitudes of the ion currents at the second and third action potential were smaller than that at the first

Figure 4 Action potential elicited by current stimuli. A, Supra-threshold (50 pA) and sub-threshold (25 pA) current stimuli with a short duration of 2 ms elicited all (solid line) and none (dash line) action potential, respectively; B, Action potential trains were produced by the current stimulus with long duration (5.2 pA, 320 ms); C, Action potentials and hyperpolarization membrane potential were elicited by a series of depolarization and hyperpolarization current stimuli.

Figure 5 The action potentials and corresponding ion currents. (A) Action potentials upon the current stimulus of 5.2 pA, 250 ms. Na+ (B) and delayed rectifier K+ currents, (C) during the action potentials.

1042


one, which indicated that partial ion channels were in the inactivation state. 3.5

Roles of ion channels at action potentials

Action potential has a voltage- and time-dependent refractory, which is induced by the kinetics of ion channels including Na+ and K+ channels. In simulation, with different holding potentials Vh, the action potentials elicited by the current stimulus of 50 pA and 2 ms are

shown in Figure 6A. When the taste receptor cell was clamped at −60 mV, the action potential with a large overshot was produced. While clamped at −50 mV, the amplitude of the overshot decreased. When the control potential equaled −40 mV, no action potential was obtained, which may be due to a large amount of Na+ channels in the inactivation state. The simulation results when using the double pulse protocol with different time intervals ∆t are shown in

Figure 6 The effects of the kinetic properties of Na+ and K+ channels on action potentials, with different resting holding potentials (A) and with different time intervals under the double pulse protocol (B).

Figure 7 The response of acid-sensing taste receptor cells. A, The firing pattern upon the sour stimulus with pH=4.5 was elicited by the transduction current through ASIC2 channels; B, The action potential response to citric acid through PKD channels. C, A single action potential from the response in B.


1043

Figure 6B. A number of Na+ channels were in the inactivation state, when the second pulse was applied after a short time interval (such as 20 and 50 ms). Consequently, in the second pulse, the action potential could not be elicited. When the time interval was around 60 and 80 ms, partial Na+ channels were recovered from the inactivation state and the second action potential could be obtained. However, the amplitude was significantly smaller than the first one. When the time interval reached 100 ms, the second action potential was a bit larger than the first one. In this case, partial K+ channels went into the inactivation state. When the interval increased to 180 ms, K+ channels were recovered and the elicited second action potential was equal to the first one. 3.6

Firing pattern upon acidic stimuli

Recent studies indicated that two possible acid-sensing ion channel candidates, ASIC2[24,28] and PKD[25,29,30] channels, played important roles in acidic sensation. Figures 7A and B show the simulated firing patterns of taste receptor cells with ASIC2a/ASIC2b channel and PKD channel, respectively. The firing rate was about 22 Hz and 12 Hz, correspondingly. The inward current through ASIC2 channel was about 212 pA, while for PKD channel, the current was 9.28 pA. From patch clamp recordings on native taste cells, the inward current was around 105 pA upon citric acid (pH=4.5)[31]. A single action potential from Figure 7B that was elicited by PKD transduction inward current is shown in Figure 7C.

4 Discussion 4.1

Ionic currents

Whole-cell patch clamp experiments demonstrated that there was no much apparent difference among three types of taste receptor cells (fungiform, foliate and circumvallate) in Na+ currents[32], which were in charge of the onset of action potentials. Evidence from electrophysiology and molecular biology showed that more than 95% taste cells expressed outward delayed rectifier K+ channels, which contributed to the modulation of action potentials and helped maintain resting potentials. Shaker Kv1.5 channel (KCNA5) is the major functional delayed rectifier K+ channel expressed in rat anterior tongue[3]. Therefore, computational models of Na+ and K+ channels for acid-sensing taste receptor cells in this work also provided a prototype for 1044

other types of taste cells. The model of single taste receptor cells in this paper was constructed based on the patch-clamp data of Na+ currents and delayed rectifier K+ currents. Simulation results indicated that cellular responses of taste receptor cells to electrical and chemical stimuli could be reproduced. According to our simulation, these channels formed a minimum set of mechanisms to reproduce the major characteristics of sour responses. Na+ current and delayed rectifier K+ current played dominant roles in action potentials. According to the conclusion drawn above and the practical considerations, Ca2+ current, other types of K+ current and Cl− current were not included in this model. Evidence from molecular biology and electrophysiology indicated that voltagegated channels were expressed in different taste cells and were closely related to specific taste receptors. Taste cells that expressed bitter, sweet and umami receptors had voltage-gated Na+ and K+ channels, but had no voltage-gated Ca2+ channels[33,34]. Some taste cells with salty-sensing EnaC channels were shown to have no Ca2+ channels, Whereas other taste cells had T-type Ca2+ channels[35]. ASIC2-expressing taste cells had voltage-gated Na+, K+ channels and Ca2+ channels[25]. Our further work will address on the biophysical properties of the channels mentioned above and the roles they play in taste information processing. 4.2 Action potential and taste information processing

Different firing patterns were elicited by current stimuli with different intensities and durations. Upon long-duration current stimuli, trains of action potentials were elicited and the amplitudes of action potentials decreased with time. When current stimuli with different intensities were applied, the firing rate increased with depolarization currents increasing. Meanwhile, the amount of the elicited action potentials increased and the latency shortened, which indicated that taste cells were prone to fire. With hyperpolarization current stimuli, membrane potentials hyperpolarized. The variety of firing patterns provided the electrophysiological basis for the coding mechanism of taste receptor cells. Voltage-gated Na+ channel was in charge of the onset of action potentials. K+ channels were responsible for modulation of action potentials and maintenance of resting potential. These basic functions of Na+ and K+ channels were consistent with those in other types of


neurons. The activation and inactivation of ion channels affected the firing threshold, the shape and the amplitude of action potentials. Moreover, the dynamics of ion channels were modulated by signaling molecules under physiological conditions. In patch-clamp experiments, evidence demonstrated that taste cells responded to tastants and released transmitters. For example, released 5-HT could act on the adjacent taste cells, through inhibition of Ca2+ activated K+ channels and voltage-gated Na+ channels. Therefore, the excitability of the adjacent taste cells was inhibited[36]. Neuropeptides, such as cholecystokinin (CCK), inhibited the outward K+ channels and inward rectifier K+ channels, to enhance the excitability of the adjacent taste cells[37]. Bitter tastants, like quinine, inhibited the outward K+ channels and Na+ channels of the adjacent cells to broaden the shape of action potentials[15]. Further study on the regulatory roles of ion channels during firing will be helpful to reveal the coding mechanisms of taste cells. Taste receptor cells were reported to respond to sweet, sour, salty and bitter tastants and produced action potentials. In this work, we simulated the firing patterns of taste receptor cells in response to sour tastants. For sour sensation, protons, as the major stimuli, interact with ion channels of taste receptor cells. Different species might have different transduction mechanisms responsible for acidic sensation. At present, it is controversial about the acid-sensing taste receptors and transduction mechanisms, as shown in Figure 1B. In mammals, ASIC 2a/2b (acid sensing ionic channels) channels have been raised as one plausible mechanism, which is permeable to H+. PKD channels permeable to Ca2+/H+ and HCN channels (hyperpolarization activated channels) were also proposed as the potential mechanism. From simulation, we compared and verified these main mechanisms preliminarily. ASIC2a/2b is expressed in rat taste receptor cells[25]. While in mice, it was demonstrated that PKD channels but ASIC2 channels were expressed in acid-sensing taste receptor cells[29,38]. Simulation results indicate that the firing rate is 22 Hz and 12 Hz, respectively, while the inward currents are produced by ASIC2 and PKD channels correspondingly. However, we cannot support or reject either mechanism based on our simulations. The firing rate in simulation is much higher than the recordings from single taste receptor cells (about 2 Hz). Nevertheless, our results are comparable to the recordings from chorda

tympani nerve[8]. HCN channels, which can be modulated by various signaling molecules including H+, are widely expressed in nerves and cardiac myocyte[39]. From electrophysiological experiments, it was found that Ih current (HCN current) could be recorded in mouse taste cells, which expressed Na+, outward K+, L- and T-type Ca2+ channels[17]. Immunohistochemistry indicated that HCN channels were expressed in rat taste cells and were potential acid receptors[40]. However, different conclusions could be drawn from other experimental results. For instance, Ca2+ imaging showed HCN did not work in acidic sensation[41]. HCN channels had dual impacts on cellular excitability[42]. And yet, no date support that HCN channel would induce the increase of the firing rate upon sour tastants. Based on these facts, the support for HCN channel is relatively weak. We did not simulate HCN channels and its mechanism. In our future work, HCN channel will be included to analyze its impacts on action potentials and to confirm the analysis mentioned above. The study on taste sensation could result in significant findings in basic biology. Moreover, the results can be applied to improve the design of biomimetic taste sensors. The taste sensor based on lipid membrane and the taste imaging sensor based on light addressable potentiometric sensor (LAPS)[43], could respond to the taste stimuli (acidic, bitter, sweet and salty). These sensors could potentially be applied in the field of food and environment measurement. The taste receptor cell-based LAPS chips recorded the extracellular membrane potentials and detected tastants[44]. However, the relations between action potentials and extracellular recordings were not well understood. The electrophysiological experiments and theoretic computational analysis, including the model and simulation in this paper, will contribute to the understanding of the features of extracellular signals and also the design of high-performance biomimetic taste sensors.

5 Conclusions In this work, we present a computational model of rat taste receptor cells, based on patch clamp recordings of voltage-gated Na+ and delayed rectifier K+ currents. Combined with recent electrophysiological findings on taste receptor cells and the molecular mechanism of sour taste sensation, the firing patterns upon electrical and acidic stimuli are simulated. Modeling and simulated


1045

results demonstrate that Na+ channel has the characteristic features of fast activation and fast inactivation. For the K+ channel, it is fast activated and slow inactivated. Na+ and K+ channels cooperate to form the basis of the action potentials of taste receptor cells. Results further prove that Na+ and K+ channels play dominate roles in taste information processing in 1

We thank Drs. T. A. Gilbertson from Utah State University and M. S. Herness from Ohio State University for providing the data of voltage-gated K+ and Na+ currents.

Gilbertson T A, Avenet P, Kinnamon S C, et al. Proton currents

voltage-dependent currents in dissociated rat taste cells. Pflugers Arch,

through amiloride-sensitive Na channels in hamster taste cells. Role in

1997, 434: 215―226

acid transduction. J Gen Physiol, 1992, 100: 803―824 2

periphery. This model provides a platform for studying cellular response of taste receptor cells upon different kinds of tastants.

16

Cummings T A, Powell J, Kinnamon S C. Sweet taste transduction in

characterization of voltage-gated currents in defined taste cell types of

hamster taste cells: evidence for the role of cyclic nucleotides. J Neurophysiol, 1993, 70: 2326―2336 3

mice. J Neurosci, 2003, 23: 2608―2617 17

Liu L D, Hansen D R, Kim I, et al. Expression and characterization of delayed rectifying K+ channels in anterior rat taste buds. Am J Physiol

bilayer membranes. Biophys J, 1991, 59: 1218―1234 19

Modelling, 1994, 19: 119―133

Res, 1997, 776: 133―139 Avenet P, Lindemann B. Noninvasive recording of receptor cell action

20

Physiol, 1952, 117: 500―544

the tongue: the response to mucosal NaCl and amiloride. J Membr Biol, 1991, 124: 33―41

21

Hines M L, Carnevale N T. NEURON: a tool for neuroscientists.

22

Hu S, Wang S, Gibson J, et al. Inhibition of delayed rectifier K+

Neuroscientist, 2001(7): 123―135

Varkevisser B, Peterson D, Ogura T, et al. Neural networks distinguish between taste qualities based on receptor cell population responses.

channels by dexfenfluramine (Redux). J Pharmacol Exp Ther, 1998,

Chem Senses, 2001, 26: 499―505 7

8

287: 480―486

Yoshida R, Shigemura N, Sanematsu K, et al. Taste responsiveness of fungiform taste cells with action potentials. J Neurophysiol, 2006, 96:

23

canine atrial action potentials: rate, regional factors, and electrical

Yoshida R, Yasumatsu K, Shigemura N, et al. Coding channels for

remodeling. Am J Physiol Heart Circ Physiol, 2000, 279:

gustatory nerve fibers. Arch Histol Cytol, 2006, 69: 233―242

H1767―H1785 24

Hille B. Ionic Channels of Excitable Membranes (2nd ed). Sunderand

25

Shimada S, Ueda T, Ishida Y, et al. Acid-sensing ion channels in taste

26

Ishimaru Y, Inada H, Kubota M, et al. Transient receptor potential

MA: SinauerAssociates Inc. 1992

Herness M S, Sun X D. Voltage-dependent sodium currents recorded from dissociated rat taste cells. J Membr Biol, 1995, 146: 73―84

10

buds. Arch Histol Cytol, 2006, 69: 227―231

Behe P, DeSimone J A, Avenet P, et al. Membrane currents in taste cells of the rat fungiform papilla. Evidence for two types of Ca

family members PKD1L3 and PKD2L1 form a candidate sour taste

currents and inhibition of K currents by saccharin. J Gen Physiol,

receptor. Proc Natl Acad Sci USA, 2006, 103: 12569―12574

1990, 96: 1061―1084 11

Noguchi T, Ikeda Y, Miyajima M, et al. Voltage-gated channels

27

soft palates. Brain Res, 2003, 982: 241―259

28

sour-taste receptor channel." J Neurosci, 2003, 23: 3616―3622

populations of receptor cells and presynaptic cells in mouse taste buds. 13

14

29 30

LopezJimenez N D, Cavenagh M M, Sainz E, et al. Two members of

Physiol, 1996, 271: C1221―1232

the TRPP family of ion channels, Pkd1l3 and Pkd2l1, are

Huang L, Cao J, Wang H, et al. Identification and functional charac-

co-expressed in a subset of taste receptor cells. J Neurochem, 2006, 98: 68―77

terization of a voltage-gated chloride channel and its novel splice variant in taste bud cells. J Biol Chem, 2005, 280: 36150―36157 15

Huang A L, Chen X, Hoon M A, et al. The cells and logic for mammalian sour taste detection. Nature, 2006, 442: 934―938

Sun X D, Herness M S. Characterization of inwardly rectifying potassium currents from dissociated rat taste receptor cells. Am J

Ugawa S, Yamamoto T, Ueda T, et al. Amiloride-insensitive currents of the acid-sensing ion channel-2a (ASIC2a)/ASIC2b heteromeric

DeFazio R A, Dvoryanchikov G, Maruyama Y, et al. Separate J Neurosci, 2006, 26: 3971―3980

Herness M S, Sun X D. Voltage-dependent sodium currents recorded from dissociated rat taste cells. J Membr Biol, 1995, 146: 73―84

involved in taste responses and characterizing taste bud cells in mouse 12

Ramirez R J, Nattel S, Courtemanche M. Mathematical analysis of

3088―3095 taste perception: information transmission from taste cells to 9

Hodgkin A L, Huxley A F. A quantitative description of membrane current and its application to conduction and excitation in nerve. J

potentials and sustained currents from single taste buds maintained in

6

Mierson S, Fidelman M L. The role of epithelial ion transport in taste transduction: a network thermodynamic model. Mathl Comput

single taste bud cells in the peeled epithelium of mouse tongue. Brain 5

Naito M, Fuchikami N, Sasaki N, et al. Model for the dynamic responses of taste receptor cells to salty stimuli. I. Function of lipid

Furue H, Yoshii K. In situ tight-seal recordings of taste substance-elicited action currents and voltage-gated Ba currents from

Romanov R A, Kolesnikov S S. Electrophysiologically identified subpopulations of taste bud cells. Neurosci Lett, 2006, 395: 249―254

18

Cell Physiol, 2005, 289: 868―880 4

Medler K F, Margolskee R F, Kinnamon S C. Electrophysiological

Chen Y, Herness M S. Electrophysiological actions of quinine on

1046

31

Lin W, Ogura T, Kinnamon S C. Acid-activated cation currents in rat vallate taste receptor cells. J Neurophysiol, 2002, 88: 133―141


32

Doolin R E, Gilbertson T A. Distribution and characterization of

38

functional amiloride-sensitive sodium channels in rat tongue. J Gen

channel-2 is not necessary for sour taste in mice. J Neurosci, 2004, 24:

Physiol, 1996, 107: 545―554 33

Adler E, Hoon M A, Mueller K L, et al. A novel family of mammalian

34

Clapp T R, Medler K F, Damak S, et al. Mouse taste cells with G

4088―4091 39

taste receptors. Cell, 2000, 100: 693―702 protein-coupled taste receptors lack voltage-gated calcium channels

rod photoreceptors. Biochim Biophys Acta, 2003, 1609: 183―192 40

2001, 413: 631―635 41

Richter T A, Caicedo A, Roper S D. Sour taste stimuli evoke Ca2+ and

42

Poolos N P. The Yin and Yang of the h-channel and its role in epilepsy.

pH responses in mouse taste cells. J Physiol, 2003, 547: 475―483

J Neurophysiol, 2007 36

Herness M S, Chen Y. Serotonergic agonists inhibit calcium-activated

Epilepsy Curr, 2004, 4: 3―6

potassium and voltage-dependent sodium currents in rat taste receptor cells. J Membr Biol, 2000, 173: 127―138 37

43

actions of cholecystokinin in rat taste receptor cells. J Neurosci, 2002,

He H Q, Li R, Ye X S, et al. The integrated taste LAPS image sensor based on LB films (in Chinese). Instr Tech Sens, 2001, 10.

Herness M S, Zhao F L, Kaya N, et al. Expression and physiological 22: 10018―10029

Stevens D R, Seifert R, Bufe B, et al. Hyperpolarization-activated channels HCN1 and HCN4 mediate responses to sour stimuli. Nature,

Bigiani A, Cuoghi V. Localization of amiloride-sensitive sodium current and voltage-gated calcium currents in rat fungiform taste cells.

Malcolm A T, Kourennyi D E, Barnes S, et al. Protons and calcium alter gating of the hyperpolarization-activated cation current (I(h)) in

and SNAP-25." BMC Biol, 2006, 4: 7 35

Richter T A, Dvoryanchikov G A, Roper S D, et al. Acid-sensing ion

44

Chen P H, Cheng G, Wang B Q, et al. Design of a novel acid-sensing biosensor based on taste neurons and LAPS. ICBBE, 2008


1047