The Journal of Neurophysiology Vol. 78 No. 4 October 1997, pp. 2079-2085
Copyright ©1997 by the American Physiological Society

Ionic Selectivity of Mechanically Activated Channels in Spider Mechanoreceptor Neurons

Ulli Höger, Päivi H. Torkkeli, Ernst-August Seyfarth, and Andrew S. French

Zoologisches Institut, J. W. Goethe-Universität, D-60054 Frankfurt am Main, Germany; and Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 4H7, Canada

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Höger, Ulli, Päivi H. Torkkeli, Ernst-August Seyfarth, and Andrew S. French. Ionic selectivity of mechanically activated channels in spider mechanoreceptor neurons. J. Neurophysiol. 78: 2079-2085, 1997. The lyriform slit-sense organ on the patella of the spider, Cupiennius salei, consists of seven or eight slits, with each slit innervated by a pair of mechanically sensitive neurons. Mechanotransduction is believed to occur at the tips of the dendrites, which are surrounded by a Na⁺-rich receptor lymph. We studied the ionic basis of sensory transduction in these neurons by voltage-clamp measurement of the receptor current, replacement of extracellular cations, and application of specific blocking agents. The relationship between mechanically activated current and membrane potential could be approximated by the Goldman-Hodgkin-Katz current equation, with an asymptotic inward conductance of ~4.6 nS, indicating that 50-230 channels of 20-80 pS each would suffice to produce the receptor current. Amiloride and gadolinium, which are known to block mechanically activated ion channels, also blocked the receptor current. Ionic replacement showed that the channels are not permeable to choline or Rb⁺, but are partly permeable to Li⁺. The receptor current was inward at all membrane potentials (200 to +200 mV) and never reversed, indicating high selectivity for Na⁺ over K⁺. This situation contrasts strongly with insect mechanoreceptors, vertebrate hair cells, and mechanically activated ion channels in nonsensory cells, most of which are either unselective for monovalent cations or selective for K⁺.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Mechanically activated ion channels with a variety of ionic selectivities have been found in animal, bacterial, fungal, and plant cells (French 1992; Morris 1990; Sachs 1988; Sackin 1995). The detection of internal and external mechanical stimuli by mechanoreceptor neurons is presumably based on such channels, but they are also important in other cellular functions, including cell volume regulation, growth, and migration. Single mechanically activated channel recordings have been made from nonsensory cells, but not from the sensory endings of specialized mechanoreceptor cells such as vertebrate skin mechanoreceptors, vestibular hair cells, or insect cuticular mechanosensilla. The locations of the sensory endings of most mechanoreceptor neurons, as well as their small sizes, make intracellular or patch-clamp recording difficult. Only vertebrate hair cells have so far been accessible for whole cell patch-clamp recording (e.g., Kros et al. 1992), and single mechanically activated channels have been recorded from crustacean stretch-receptor neurons (Erxleben 1989), although it is not certain that they were present at the tips of the dendrites that respond to mechanical stimulation.

The anterior lyriform organs VS-3 of the tropical wandering spider (Cupiennius salei Keys.) lie on the anterioventral side of the leg patella, where they detect strain acting on the exoskeleton (Barth 1985). VS-3 consists of seven or eight cuticular slits, each of which is innervated by a pair of spindle-shaped bipolar neurons. In response to step mechanical or electrical stimulation, one neuron in each pair produces a single action potential, and the other produces a short burst of action potentials with decaying amplitudes (Seyfarth and French 1994).

The first recordings of mechanically activated currents in spider slit-sense organs were obtained by Juusola et al. (1994). Although the distance from the recording site in the soma to the dendritic tip, where the mechanically activated current is generated, is ~100 µm, the receptor current can be recorded reliably under voltage-clamp conditions when the slits are stimulated with step displacement. Juusola et al. (1994) showed that the receptor current is carried mainly by Na⁺ at the normal resting membrane potential, but permeability to other ions or the conductance at different potentials was not investigated.

The dendritic tips of the slit-sense neurons, containing the mechanically activated ion channels, are located in a receptor lymph space that has a high concentration of Na⁺ (Rick et al. 1976). Other studies have shown that the concentration of K⁺ in spider receptor lymph does not differ significantly from hemolymph, but the concentration of Ca²⁺ in hemolymph is about three times higher than that of receptor lymph (Grünert and Gnatzy 1987). This situation is different to the analogous organs of insects, the campaniform sensilla (Barth 1985), which have up to 10 times higher concentration of K⁺ in the receptor lymph than in the hemolymph (Küppers 1974). Vertebrate mechanosensory hair cells also have their mechanically activated ion channels facing a relatively high K⁺ concentration (Corey and Hudspeth 1979).

Most mechanically activated ion channels that have been studied so far are either unselective between the alkali cations and Ca²⁺, or are K⁺ selective. Some large-conductance, anion-selective mechanically activated channels have also been found in plant cells and bacteria (French 1992; Sachs 1988). One purpose of the present study was to explore the selectivity of the mechanically activated ion channels in spider slit-sense organs, and to compare their characteristics to other known mechanically activated channels. We also tested the sensitivity of the spider neurons to two specific blockers of mechanically activated ion channels. Our results indicate that sensory transduction in the spider slit-sense organ neurons is based on a few hundred mechanically activated channels that are highly selective for Na⁺ over K⁺ and are blocked by amiloride and by Gd³⁺.

	METHODS

Abstract Introduction Methods Results Discussion References

Experimental arrangement

Legs from adult and penultimate-stage female hunting spiders (C. salei Keys., raised in our laboratories) were autotomized and a concave piece of patellar cuticle containing lyriform slit-sense organ VS-3 (nomenclature of Barth and Libera 1970) was dissected free. The preparation was mounted with dental wax and petroleum jelly onto a custom-designed Plexiglas holder. Details of the dissection and mounting of the preparation were described previously by Seyfarth and French (1994) and Juusola and French (1995). The microelectrode was positioned with a Leitz-micromanipulator. A second micromanipulator was used to place two plastic tubes into the bath chamber for exchanging solutions during the experiment. The experimental apparatus rested on a gas-driven vibration isolation table (Technical Manufacturing, Micro-g), and all experiments were performed at room temperature (22 ± 2°C, mean ± SD).

Recording and stimulation

Current- and voltage-clamp measurements were performed using the discontinuous (switching) single-electrode method (Finkel and Redman 1984) with an NPI SEC-10 l amplifier. The conditions for successful single-electrode voltage and current clamp were described in detail by Torkkeli and French (1994), and the same methods were used previously in this preparation for recording currents and voltages activated either by mechanical or by electrical stimulation (Juusola et al. 1994; Juusola and French 1995). Borosilicate microelectrodes (1 mm OD and 0.5 mm ID) were pulled with a horizontal puller (P-2000, Sutter Instrument, Novato, CA). The electrodes were filled with 3 M KCl and coated with petroleum jelly to decrease stray capacitance (Juusola et al. 1997). Electrode resistances were 45-70 M with time constants of 1-3 µs in solution. These electrodes allowed amplifier switching frequencies of 20-23 kHz to be used. A duty cycle of 1/8 (current passing/voltage recording) was used in all experiments.

The cells were identified visually and then penetrated through a thin layer of spider saline. Penetration was achieved by high-frequency oscillation ("buzzing") or delicate tapping of the manipulator. Impaled cells were allowed to stabilize for 15 min before recordings began. Criteria for reliable recordings were stable resting membrane potentials of about 60 mV and action-potential amplitudes larger than 40 mV.

A piezoelectric stimulator, equipped with a position transducer and a controller (LVPZ translator, PZT controller; Polytec Physik-Instrumente, WaldGronn, Germany) was used for mechanical stimulation. The stimulator was mounted on a three-dimensional micromanipulator to position the tip of the stimulating probe beneath the outer surface of the VS-3 slits (Fig. 1). The probe was prepared by melting the tip of a normal recording electrode to produce a bead of ~50 µm diam. The probe tip could be clearly seen through the preparation and was placed on the slit that was innervated by the impaled neuron. A 2-µm displacement of the slit was adequate to evoke action potentials in the neurons. Displacement stimuli of 200 ms duration and 2- to 3-µm amplitudes were used to induce the currents, and three responses were averaged to produce each current trace. Series of recordings of mechanically activated currents were performed at different holding voltages to definethe current-voltage relationship produced by each displacementstimulus.

View larger version (49K): [in this window] [in a new window]   FIG. 1. Experimental arrangement for mechanical stimulation of the spider slits and for intracellular recording from the sensory neurons innervating the slits are shown left. The isolated cuticle preparation with slit-sense organ VS-3 was mounted with the interior facing upward. The exposed sensory neurons were impaled with glass microelectrodes from above. The trough formed by the leg cuticle was filled with spider saline, and the solution changes were made directly with syringe needles. A probe for mechanical stimulation made contact with the slits from below. Right: a cross-section through one slit. Each slit is innervated by 2 sensory neurons, but only 1 dendrite proceeds to the external membrane covering the slit. Both dendrites are located in a receptor lymph space that is separated from the hemolymph (adapted from Barth 1971).

All current- and voltage-clamp experiments were controlled by an IBM-compatible computer using software that was custom written in this laboratory. The computer provided voltage, current, or displacement stimuli via a 12-bit D/A convertor. Current output and membrane potential were recorded via a 12-bit A/D convertor. Membrane potential was low-pass filtered at 33.3 kHz and current at 3.3 kHz by the voltage-clamp amplifier. Membrane resistance was calculated from the linear part of the current-voltage curve at hyperpolarizing potentials.

Chemicals and ion-replacement procedures

The spider saline used in these experiments was modified from Maier et al. (1987) and contained (in mM) 223 NaCl, 6.8 KCl, 8 CaCl₂, 5.1 MgCl₂, and 10 N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES), pH 7.8. After a 15-min stabilization period, mechanical stimuli were applied to produce and record action potentials. Subsequently, the spider saline in the bath (total volume 1 ml) was replaced with a saline supplemented by 5 µM tetrodotoxin (TTX) to block the voltage-activated sodium currents. Complete blockade of action potentials took 5-15 min. It was then possible to record the mechanically activated currents. After recording in TTX-saline, the solution in the bath was replaced by various other solutions. Blocking agents, 1 mM amiloride hydrochloride or 1 mM gadolinium (III) chloride (Gd³⁺), were added to regular saline. For ion replacement experiments, total NaCl was replaced by equimolar concentrations of choline chloride, LiCl, or RbCl. The 1-ml volume of the bath was washed out with at least 5 ml of the new solution for complete replacement of the previous agent. All chemicals were purchased from Sigma (St. Louis, MO), and all solutions were either freshly prepared or used from frozen stock.

	RESULTS

Abstract Introduction Methods Results Discussion References

A total of 18 intracellular recordings from VS-3 neurons was selected for analysis. All recordings were made from neurons numbered 2 to 6, which are large enough to visualize and penetrate reliably (Seyfarth and French 1994). Two recordings were from type A neurons and 16 from type B neurons. Type A fire only a single action potential in response to a step depolarization, whereas type B fire a rapidly adapting burst of action potentials (Seyfarth and French 1994). No significant differences in receptor currents were seen between type A and type B neurons. The average membrane potential after the 15-min stabilization period was 61 ± 6 mV (n = 18), the mean membrane resistance was 174 ± 91 M, and the action-potential amplitudes varied between 45 and 70 mV. Before recording mechanically activated currents, the voltage-activated Na⁺ currents were blocked with 5 µM TTX, because it was not possible to completely clamp the action potentials and this would have hindered the recording of the relatively small receptor current. With all muscle tissue dissected away from the patella, the cell bodies and axons of the neurons were relatively easily accessible to externally applied solutions, and the TTX effect usually occurred in 5-15 min. However, the mechanically activated ion channels are believed to be located on the dendritic tips of these cells (Fig. 1), surrounded by a receptor lymph that is isolated from the normal extracellular fluid of the spider. This arrangement limits the exchange of ions between these two fluid spaces. Therefore all experiments involving changes in extracellular solutions took a significant time (up to 1 h).

Current-voltage relationship

After successful penetration, the mechanical stimulator was positioned on the slit innervated by the neuron being recorded. Mechanical stimulation produced a receptor potential under current clamp and a receptor current under voltage clamp. Typical recordings of voltage and current responses to three different displacement stimuli are shown in Fig. 2. At 70 mV the average peak receptor current amplitude from all recordings was 217 ± 92 pA (n = 18). Activation and inactivation of the receptor current were fitted by two exponentials (Fig. 3 legend) (see Torkkeli and French 1995) with time constants of 3.3 ± 2.2 ms and 29.5 ± 6.8 ms(n = 7). However, the activation time course was primarily determined by the rise time of the stimulator (~4 ms). The inactivation time constant did not vary significantly with holding voltage, but the amplitude of the inward receptor current increased with more hyperpolarizing potentials as shown in Fig. 3. The direction of the current flow via mechanically activated channels was inward at all holding potentials. Even at strongly positive potentials the current never reversed but remained close to zero. The current-voltage relationship could be fitted by the Goldman-Hodgkin-Katz (GHK) current equation for Na⁺ (Hille 1992) provided that an offset voltage was added to the equation. The mean offset was 46.7 ± 19.9 mV (n = 7). Possible reasons for this offset are described in the DISCUSSION. GHK-fits gave a value of 4.6 ± 1.5 nS (n = 7) for the asymptotic conductance of the mechanically activated current at negative potentials.

View larger version (17K): [in this window] [in a new window]   FIG. 2. Typical voltage and current responses of a VS-3 neuron to 200-ms mechanical step stimuli are shown in A and B, respectively. Each trace represents the average of 3 recordings. A: voltage responses in physiological saline. A 1-µm deformation stimulus elicited only a receptor potential, but 2- and 3-µm stimuli produced 1 action potential and a graded receptor potential that persisted over the entire stimulus duration. B: current responses after tetrodotoxin (TTX) application. Response to a 3-µm step stimulus decayed with 2 time constants, in this case 1.6 and 63 ms. C: output signals from the stimulator position transducer during stimulation. The rise time of the stimulus was 4.5 ms, which was too slow to make accurate measurements of the time constant for receptor current activation.

View larger version (28K): [in this window] [in a new window]   FIG. 3. A: typical receptor currents at potentials of 100, 80, 60, 40, 20, 0, 50, and 100 mV (in order of decreasing peak amplitude). The membrane potential was held at each test potential while 3-µm step stimuli of 200-ms duration were applied to produce the receptor current. Ionic currents that were independent of the mechanical stimulus have been subtracted. Peak and steady-state current values (B,  and ) were obtainedby fitting the individual curves with the function: I = I + et/ + et/,where I was the steady-state current, and the 2 exponential components represented the growth and decay of the transient current, following a step displacement. B: magnitudes of the peak and steady-state (at 200 ms) receptor currents are shown as functions of membrane potential. Continuous lines were fitted to the data ( and ) using the Goldman-Hodgkin-Katz current equation. Note the failure of the current to reverse, even at potentials that should exceed the Na+ equilibrium potential. The external solution was spider saline.

Ionic selectivity of the mechanically activated channels

The mechanically activated current did not reverse at any holding potential (Fig. 3), indicating that the ion channels responsible for this current in the VS-3 cells are strongly selective for Na⁺ over K⁺. The extracellular Na⁺ concentrations of spider hemolymph and receptor lymph are high compared with their K⁺ concentrations (Grünert and Gnatzy 1987; Rick et al. 1976), and Na⁺ was expected to be the main charge carrier of the mechanically activated current (Juusola et al. 1994). However, it was surprising to find that high positive potentials did not cause K⁺ ions to leave the cell through these channels. The lack of outward current may also mean that the intracellular Na⁺ concentration of VS-3 cells is low compared with the extracellular concentration. To test whether the mechanically activated channels are permeable to other alkali cations, we superfused the cells with three different solutions in which all of the extracellular Na⁺ was replaced with either choline, Li⁺, or Rb⁺. The results of these experiments are shown in Figs. 4-6. For these, and subsequent experiments, only negative holding potentials were used, to make the experiments faster, and to reduce the probability of losing a cell, which seemed more common with strong depolarizations. Data were then fitted by linear regression to approximate the asymptotic conductance at large negative potentials.

View larger version (26K): [in this window] [in a new window]   FIG. 4. Receptor current was entirely abolished 50 min after extracellular NaCl was replaced with equimolar choline chloride. This result was completely reversed ~1 h after rinsing with normal saline. The holding potential in A was 110 mV. B: current-voltage curves of the peak inward currents at holding potentials between 120 and 10 mV. Lines in B are linear regression fits to the data. The mechanical stimulus used in each recording was 2 µm.

About 50 min after extracellular Na⁺ was replaced with choline (Fig. 4), mechanically activated currents could not be detected at any holding potential. When the cells were again superfused with normal spider saline, the receptor current recovered fully after ~1 h. The current amplitude after rinsing was actually larger than in the control situation, which was probably due to improvement of the seal between the recording electrode and the cell membrane. The input resistance in the recording shown in Fig. 4 increased ~40 M from the time of the control recording to the time of the recovery, which also indicates that the electrode was better sealed.

When extracellular Na⁺ was replaced with an equimolar concentration of Li⁺ (Fig. 5), part of the mechanically activated current remained, even 1 h after the solution change. However, Li⁺ could not be washed out from the receptor lymph space, because the receptor current was still reduced 90 min after return to normal Na⁺ concentration. The effect of Rb⁺ was also irreversible (Fig. 6), but it blocked the mechanically activated current rapidly, in <15 min. Rb⁺ has been shown to pass through many types of K⁺ channels in other preparations (Cook 1990). The rapidity of the Rb⁺ effect in our experiments, and the fact that it caused a significant reduction in resting membrane potential, indicates that it may actually have entered the cells via voltage-activated K⁺ channels. However, Rb⁺ did not pass through the mechanically activated channels, and it was not possible to rinse it from the receptor lymph space.

View larger version (23K): [in this window] [in a new window]   FIG. 5. Replacement of extracellular Na+ with equimolar Li+ reduced the receptor current significantly, but even 1 h after application a reduced current remained. The holding potential in A was 100 mV. B: peak receptor currents at holding potentials from 150 to 10 mV, and the fitted lines are linear regressions to the data. The mechanical stimulus was 2 µm.

View larger version (21K): [in this window] [in a new window]   FIG. 6. When total extracellular Na+ was replaced with Rb+, the mechanically activated current completely disappeared in <15 min. Holding potential in A was 95 mV. B: peak inward currents at holding potentials between 95 and 35 mV. Straight lines in B are linear regression fits to the data. The stimulus amplitude in all recordings was 3 µm.

Blockers of mechanically activated ion channels

Amiloride (1 mM) and Gd³⁺ (1 mM) were each added separately to the spider saline to test their blocking actions on the mechanically activated ion channels. Both of these agents blocked the receptor current almost completely (Figs. 7 and 8), but only the effect of amiloride was reversible. The recovery after amiloride treatment produced a larger receptor current than the control situation, similarly to the choline experiment (Fig. 4), with the seal between the electrode and the membrane improving during the 2 h of recording. Amiloride did not change the voltage-activated currents. Gd³⁺ blocked the mechanically activated current almost completely, but it also blocked all voltage-activated outward currents in the VS-3 cells and its effect was not reversible. Gd³⁺ obviously has some permanent effects that prevent it from being a useful blocker of transduction in the spider mechanoreceptor cells.

View larger version (21K): [in this window] [in a new window]   FIG. 7. Fifty minutes after 1 mM amiloride was added to the spider saline, most of the receptor current was abolished. This block was removed ~1 h after the preparation was rinsed with normal saline. In A the holding potential was 90 mV. B: peak receptor currents from 90 to 60 mV. Straight lines in B are linear fits to the data. The stimulus amplitude was 2 µm.

View larger version (19K): [in this window] [in a new window]   FIG. 8. Application of 1 mM Gd3+ blocked most of the receptor current in ~50 min. The holding potential in A was 95 mV. B: peak receptor currents between 95 and 25 mV. Fitted lines in B are linear fits to the data. The mechanical stimulus was 3 µm.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Ionic selectivity of the mechanically activated channels

The receptor current in the mechanoreceptor neurons of the spider slit-sense organ did not reverse at any holding potential when Na⁺ was the main extracellular cation. This indicates that K⁺ cannot pass through the mechanically activated channels in these cells, since depolarizing potentials would otherwise cause a K⁺ current to flow outward. Replacement of Na⁺ by choline blocked the mechanically activated current completely and reversibly, confirming that this current is carried by Na⁺ under normal conditions.

Li⁺ passed through the mechanically activated ion channels, but its permeability was significantly smaller than Na⁺(Fig. 5). Li⁺ depolarized the cell membrane potentialby ~10 mV, reduced the amplitude, and slowed the time course of voltage-activated outward currents. In the squid axon and the node of Ranvier, Li⁺ blocked delayed rectifier K⁺ channels from inside (Cook 1990), and this could explain its action on the resting potential and voltage-activated outward currents in spider VS-3 cells. Voltage-activated currents in the spider neurons have not yet been characterized, but preliminary results indicate the presence of delayed rectifier and A-type K⁺ currents. Li⁺ acted slowly on the mechanically activated current, indicating that the ions diffused into the receptor lymph space, entered the neuron via mechanically activated channels, and then blocked the delayed rectifier K⁺ channels from inside. In agreement with this model, the reduction in outward current occurred after most of the effect on mechanically activated current. This blocking action could not be reversed, indicating strong binding of Li⁺ to K⁺ channels.

Rb⁺ has previously been shown to pass through delayed rectifier K⁺ channels in squid axons and A-channels in Helix neurons, and to block inwardly rectifying K⁺ channels (Cook 1990). In VS-3 cells, Rb⁺ depolarized the resting potential by 10-20 mV, and generated large currents with voltage steps, indicating a high permeation of Rb⁺ through K⁺ channels. However, no mechanically activated currents were observed when extracellular Na⁺ was replaced with Rb⁺. If the mechanically activated channels were permeable to Rb⁺, mechanical stimulation should have produced either inward current if Rb⁺ had been passing through these channels from outside or outward current if Rb⁺ had first entered the cell via K⁺ channels and then departed via the mechanically activated channels. Mechanically activated currents were eliminated a few minutes after the solution change, indicating that Rb⁺ replaced the Na⁺ on the outside of the membrane very rapidly.

These results contrast strongly with previous findings for mechanically activated ion channels in nonsensory cells, which are either unselective among the alkali cations and Ca²⁺, or are K⁺ selective (French 1992; Sachs 1988; Sackin 1995). Some mechanically activated channels even distinguish poorly between cations and anions (Morris 1990). At least two other types of mechanoreceptor cells also have mechanically activated currents carried mainly by Na⁺: the Pacinian corpuscle, a mammalian touch receptor (Diamond et al. 1958), and the stretch receptor neurons of crayfish (Rydqvist and Purali 1993). However, in both of these cases, other ions have also been shown to contribute to the current (Bell et al. 1994; Rydqvist and Purali 1993).

The dendritic parts of mechanoreceptor neurons, where the mechanically activated ion channels are believed to be located, are usually surrounded by a fluid with a different ionic composition from the extracellular medium. For example, the dendrites of cochlear hair cells and Pacinian corpuscles are bathed in a solution with more than twice the K⁺ concentration of normal extracellular fluid (Bell et al. 1994; Corey and Hudspeth 1979). Similarly, the cuticular mechanoreceptors of arachnids and insects have dendritic endings surrounded by a receptor lymph with different ionic composition than the hemolymph. However, arachnids and insects also have differences between their receptor lymph fluids. The K⁺ concentration of spider slit-sense organ lymph is not significantly different from hemolymph, but its Ca²⁺ concentration is almost three times lower (Grünert and Gnatzy 1987). In contrast, insects have up to 10 times higher K⁺ concentration in the receptor lymph than the hemolymph (Grünert and Gnatzy 1987; Küppers 1974), but the Na⁺ concentrations are usually similar. Using X-ray analysis, Rick et al. (1976) found that the Na⁺ concentration in spider receptor lymph was high compared with K⁺ or other ions but did not give absolute values or measure concentrations in the hemolymph. Therefore the relative concentrations of Na⁺ in spider receptor lymph and hemolymph are unknown, but the ionic milieu in spider receptor lymph is clearly different to insect receptor lymph or vertebrate hair cells, which are all bathed in high K⁺. In view of the high concentration of Na⁺ in the receptor lymph, the VS-3 mechanically activated channels were expected to pass Na⁺. However, the strong Na⁺ selectivity was unexpected, because most other mechanically activated channels are either selective for K⁺ or unselective among the alkali cations.

We did not test the Ca²⁺ permeability of the mechanically activated ion channels in VS-3 cells, because it would have been difficult to control the Ca²⁺ concentration in an intact preparation, especially in the receptor lymph space, and the cells deteriorate rapidly in low calcium. However, elimination of the receptor current when Na⁺ was replaced by choline (Fig. 4) indicates that any Ca²⁺ contribution is small. For some mechanically activated channels, Ca²⁺ can be removed from both sides of the membrane without eliminating the mechanical sensitivity, but others are regulated by Ca²⁺ (Sackin 1995). In cochlear hair cells Ca²⁺ is more permeant than monovalent cations (Ohmori 1985), although its concentration is lower in endolymph than in plasma. It would be interesting to learn the role of Ca²⁺ in the VS-3 organ. Its concentration in receptor lymph is also lower than in hemolymph (Grünert and Gnatzy 1987) but could be raised by mechanical activity. A reliable single-channel preparation may be required to answer this question.

Conductance of the mechanically activated channels

We used the GHK current equation to approximate the current-voltage relationship of the mechanically activated current. This equation fitted the data well only when an offset of ~47 mV was added to the membrane potential. There are several possible reasons for this offset. One would be a potential difference between the receptor lymph space and the hemolymph, as has been found in a range of insect cuticular mechanoreceptors (Küppers 1974). However, the potential would have opposite polarity to that seen in insects, and no such potential was found in a previous investigation of spider lymph space (Grünert and Gnatzy 1987).

Another possibility would be an artifactual difference between the clamped potential in the soma and the actual potential in the sensory dendrite. The time constant of the cell membrane in both types of neurons is ~7 ms (M. Juusola and A. S. French, unpublished measurements from 97 neurons), and the sensory dendrite diameters are >2 µm (Seyfarth et al. 1995). Assuming a specific membrane capacitance of 1 µF/cm² and an intracellular resistivity of ~100  cm (Hille 1992) gives a specific membrane resistance of ~7 k cm² and a space constant of ~600 µm, so the potential at the tip of the dendrite should be ~85% of that in the soma. The recordings were made with petroleum jelly (Vaseline)-coated electrodes, which improve the clamp and yield larger receptor currents (Juusola et al. 1997).

To test whether the offset was due to attenuation of the clamp potential, we fitted the current-voltage relationships using a model in which the offset was zero and the difference between the potential in the tips and the resting membrane potential was a fraction of the difference between the clamp potential and the resting potential. However, this always predicted amplification of the potential shift, rather than attenuation. Some combination of clamp attenuation and offset potential at the tips might fit the data better, but too little is known about these phenomena to justify such speculation at the moment. A significant calcium component of receptor current could also cause an offset potential, although this seems unlikely (see above). Other possible sources for the offset would be local membrane surface charge at the dendrite tip, or fixed charges at the mouths of the mechanically activated channels.

The asymptotic conductance of the mechanically activated current in VS-3 cells at negative potentials was ~4.6 nS, close to the 5 nS reported for turtle hair cells (Crawford et al. 1989) and about one-half the 9.2 nS of mouse cochlear hair cells (Kros et al. 1992). A single-channel conductance of 20-80 pS, as reported for other cation-selective mechanically activated ion channels (French 1992; Sachs 1988), would require a VS-3 cell to have 50-230 open channels. This is consistent with some other preparations. For example, each vertebrate hair cells has <100 mechanically activated channels (Kernan and Zuker 1995).

Blockers of mechanically activated ion channels

TTX reduced the receptor current in Pacinian corpuscles (Bell et al. 1994) and blocked one of five types of mechanically activated channels in chick cardiomyocytes (Ruknudin et al. 1993). We used a rather high concentration of TTX to block voltage-activated Na⁺ current in VS-3 cells but still recorded large receptor currents, indicating that TTX does not block mechanically activated channels in this preparation.

Gd³⁺ is a widely used blocker of mechanically activated currents. It blocked unselective and K⁺-selective mechanically activated channels in oocytes (Yang and Sachs 1989), yeast spheroplasts (Gustin 1991), and cardiac myocytes (Ruknudin et al. 1993). Gd³⁺ blocked most of the receptor current in spider VS-3 neurons (Fig. 8), and all of the voltage-activated currents. This is not surprising, since Gd³⁺ blocks many types of Ca²⁺ channels (Biagi and Enyeart 1990), end-plate channels (Yang and Sachs 1989) and, less potently, voltage-gated Na⁺ and K⁺ channels, as well as voltage-independent leak channels in myelinated nerves (Elinder and Arhem 1994).

Amiloride blocks many epithelial Na⁺ channels (Kleyman and Cragoe 1988). It has also been found to block mechanically activated channels at the single-channel level (e.g., Xenopus oocyte) (Hamill et al. 1992), and the whole cell level (e.g., mammalian hair cells) (Rüsch et al. 1991). Amiloride blocked some mechanically activated ion channels only at negative potentials (Hamill and McBride 1996). Receptor current in VS-3 cells was only significant at negative potentials and was almost completely blocked by amiloride (Fig. 7). Blockade of mechanically activated channels by amiloride agrees with the discovery that the mechanotransducing degenerin proteins of C. elegans are homologous to epithelial sodium channels and probably belong to the same gene family (Canessa et al. 1994; Hamill and McBride 1996). The evidence obtained here supports these suggestions for the receptor channels in VS-3 neurons. Future efforts will be directed at exploring relationships between the mechanically activated channels of VS-3 neurons and the epithelial Na⁺ channel family.

	ACKNOWLEDGEMENTS

  This work was supported by the Medical Research Council of Canada, the Deutsche Forschungsgemeinschaft, and a Collaborative Research Grant from the North Atlantic Treaty Organization.

	FOOTNOTES

  Address for reprint requests: A. S. French, Dept. of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 4H7, Canada.

  Received 3 March 1997; accepted in final form 13 June 1997.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• BARTH, F. G. Der sensorische Apparat der Spaltsinnesorgane (Cupiennius salei Keys., Araneae). Z. Zellforsch. 112: 212-246, 1971.[Medline]
• BARTH, F. G. Slit sensilla and the measurement of cuticular strains. In: Neurobiology of Arachnids, edited by and F. G. Barth . New York: Springer-Verlag, 1985, p. 162-188
• BARTH, F. G., LIBERA, W. Ein Atlas der Spaltsinnesorgane von Cupiennius salei Keys. Chelicerata (Araneae). Z. Morphol. Tiere 68: 343-369, 1970.
• BELL, J., BOLANOWSKI, S., HOLMES, M. H. The structure and function of Pacinian corpuscles: a review. Prog. Neurobiol. 42: 79-128, 1994.[Medline]
• BIAGI, B. A., ENYEART, J. J. Gadolinium blocks low- and high-threshold calcium currents in pituitary cells. Am. J. Physiol. 259: C515-C520, 1990.[Abstract/Free Full Text]
• CANESSA, C. M., SCHILD, L., BUELL, G., THORENS, B., GAUTSCHIL, Y., HORISBERGER, J.-D., ROSSIER, B. C. The amiloride-sensitive epithelial sodium channel is made of three homologous subunits. Nature 367: 463-467, 1994.[Medline]
• COOK, N. S. In: Potassium Channels. Structure, Classification, Function and Therapeutic Potential., . New York: Horwood, 1990, p. 412
• COREY, D. P., HUDSPETH, A. J. Ionic basis of the receptor potential in a vertebrate hair cell. Nature 281: 675-677, 1979.[Medline]
• CRAWFORD, A. C., EVANS, M. G., FETTIPLACE, R. Activation and adaptation of transducer currents in turtle hair cells. J. Physiol. (Lond.) 419: 405-434, 1989.[Abstract/Free Full Text]
• DIAMOND, J., GRAY, J. A. B., INMAN, D. R. The relation between receptor potentials and the concentration of sodium ions. J. Physiol. (Lond.) 142: 382-394, 1958.
• ELINDER, F., ARHEM, P. Effects of gadolinium on ion channels in the myelinated axon of Xenopus laevis: four sites of action. Biophys. J. 67: 71-83, 1994.[Abstract/Free Full Text]
• ERXLEBEN, C. Stretch-activated current through single ion channels in the abdominal stretch receptor organ of the crayfish. J. Gen. Physiol. 94: 1071-1083, 1989.[Abstract/Free Full Text]
• FINKEL, A. S., REDMAN, S. Theory and operation of a single microelectrode voltage clamp. J. Neurosci. Methods 11: 101-127, 1984.[Medline]
• FRENCH, A. S. Mechanotransduction. Annu. Rev. Physiol. 54: 135-152, 1992.[Medline]
• GRÜNERT, U., GNATZY, W. K⁺ and Ca⁺⁺ in the receptor lymph of arthropod cuticular mechanoreceptors. J. Comp. Physiol. [A] 161: 329-333, 1987.[Medline]
• GUSTIN, M. C. Single channel mechanosensitive currents. Science 253: 800-801, 1991.[Free Full Text]
• HAMILL, O. P., LANE, J. W., MCBRIDE, D. W. JR Amiloride: a molecular probe for mechanosensitive channels. Trends Pharmacol. Sci. 13: 373-376, 1992.[Medline]
• HAMILL, O. P., MCBRIDE, D. W. JR The pharmacology of mechanogated membrane ion channels. Pharmacol. Rev. 48: 231-252, 1996.[Abstract]
• HILLE, B. In: Ionic Channels of Excitable Membranes., . Sunderland, MA: Sinauer, 1992, p. 667
• JUUSOLA, M., FRENCH, A. S. Recording from cuticular mechanoreceptors during mechanical stimulation. Pflügers Arch. 431: 125-128, 1995.[Medline]
• JUUSOLA, M., SEYFARTH, E.-A., FRENCH, A. S. Sodium-dependent receptor current in a new mechanoreceptor preparation. J. Neurophysiol. 72: 3026-3028, 1994.[Abstract/Free Full Text]
• JUUSOLA, M., SEYFARTH, E.-A., FRENCH, A. S. Rapid coating of glass-capillary microelectrodes for single-electrode voltage-clamp. J. Neurosci. Methods 71: 199-204, 1997.[Medline]
• KERNAN, M., ZUKER, C. Genetic approaches to mechanosensory transduction. Curr. Opin. Neurobiol. 5: 443-448, 1995.[Medline]
• KLEYMAN, T. R., CRAGOE, E. J. JR Amiloride and its analogs as tools in the study of ion transport. J. Membr. Biol. 105: 1-21, 1988.[Medline]
• KROS, C. J., RÜSCH, A., RICHARDSON, G. P. Mechano-electrical transducer currents in hair cells of the cultured neonatal mouse cochlea. Proc. R. Soc. Lond. B Biol. Sci. 249: 185-193, 1992.[Medline]
• KÜPPERS, J. Measurements of the ionic milieu of the receptor terminal in mechanoreceptive sensilla of insects. In: Symposium in Mechanoreception, edited by J. Schwartzkopf. Opladen, Germany: Westdeutscher Verlag, 1974, vol. 53, p. 387-394.
• MAIER, L., ROOT, T. M., SEYFARTH, E.-A. Heterogeneity of spider leg muscle: histochemistry and electrophysiology of identified fibers in the claw levator. J. Comp. Physiol. [B] 157: 285-294, 1987.
• MORRIS, C. E. Mechanosensitive ion channels. J. Membr. Biol. 113: 93-107, 1990.[Medline]
• OHMORI, H. Mechano-electrical transduction currents in isolated vestibular hair cells of the chick. J. Physiol. (Lond.) 359: 189-217, 1985.[Abstract/Free Full Text]
• RICK, R., BARTH, F. G., VON PAWEL, A. X-ray microanalysis of receptor lymph in a cuticular arthropod sensillum. J. Comp. Physiol. 110: 89-95, 1976.
• RUKNUDIN, A., SACHS, F., BUSTAMANTE, J. O. Stretch-activated ion channels in tissue-cultured chick heart. Am. J. Physiol. 264: H960-H972, 1993.[Abstract/Free Full Text]
• RÜSCH, A., KROS, C. J., RICHARDSON, G. P., RUSSELL, I. J. Block of transducer currents in neonatal mouse outer hair cells in culture by amiloride and some of its analogues. J. Gen. Physiol. 98: 24a, 1991.
• RYDQVIST, B., PURALI, N. Transducer properties of the rapidly adapting stretch receptor neurone in the crayfish (Pacifastacus leniusculus). J. Physiol. (Lond.) 469: 193-211, 1993.[Abstract/Free Full Text]
• SACHS, F. Mechanical transduction in biological systems. Crit. Rev. Biomed. Eng. 16: 141-169, 1988.[Medline]
• SACKIN, H. Stretch-activated ion channels. Kidney Int. 48: 1134-1147, 1995.[Medline]
• SEYFARTH, E.-A., FRENCH, A. S. Intracellular characterization of identified sensory cells in a new spider mechanoreceptor preparation. J. Neurophysiol. 71: 1422-1427, 1994.[Abstract/Free Full Text]
• SEYFARTH, E.-A., SANDERS, E. J., FRENCH, A. S. Sodium channel distribution in a spider mechanosensory organ. Brain Res. 683: 93-101, 1995.[Medline]
• TORKKELI, P. H., FRENCH, A. S. Characterization of a transient outward current in a rapidly adapting insect mechanoreceptor neuron. Pfluegers Arch. 429: 72-78, 1994.[Medline]
• TORKKELI, P. H., FRENCH, A. S. Slowly inactivating outward currents in a cuticular mechanoreceptor neuron of the cockroach (Periplaneta americana). J. Neurophysiol. 74: 1200-1211, 1995.[Abstract/Free Full Text]
• YANG, X.-C., SACHS, F. Block of stretch-activated ion channels in Xenopus oocytes by gadolinium and calcium ions. Science 243: 1068-1071, 1989.[Abstract/Free Full Text]

0022-3077/97 $5.00 Copyright ©1997 The American Physiological Society