Activation of volume-regulated Cl⁻ channels by ACh and ATP in Xenopus follicles

Abstract

1. Osmolarity-dependent ionic currents from follicle-enclosed Xenopus oocytes (follicles) were studied using electrophysiological techniques. Whole follicle currents were monitored using a two-electrode voltage clamp and single-channel activity was measured using the patch-clamp technique.

2. In follicles held at -60 mV two chloride currents were activated in external hyposmotic solutions. One was the habitual volume-regulated current elicited by external hyposmolarity (I_Cl,swell), and the second was a slow and smooth current (S_in) generated by ACh or ATP application.

3. In follicles, the permeability ratios for different anions with respect to Cl⁻ were similar for both I_Cl,swell and S_in, with a sequence of: SCN⁻ > I⁻ > Br⁻≥ NO₃⁻≥ Cl⁻ > gluconate ≥ cyclamate > acetate > SO₄²⁻.

4. Extracellular ATP blocked the outward component of S_in. Also, extracellular pH modulated the inactivation kinetics of S_in elicited by ACh; e.g. inactivation at +80 mV was ∼100% slower at pH 8.0 compared with that at pH 6.0.

5. Lanthanides inhibited I_Cl,swell and S_in. La³⁺ completely inhibited I_Cl,swell with a half-maximal inhibitory concentration (IC₅₀) of 17 ± 1.9 μm, while S_in was blocked up to 55% with an apparent IC₅₀ of 36 ± 2.6 μm.

6. Patch-clamp recordings in follicular cells showed that hyposmotic challenge opened inward single-channel currents. The single channel conductance (4.7 ± 0.4 pS) had a linear current-voltage relationship with a reversal membrane potential close to −20 mV. This single-channel activity was increased by application of ACh or ATP.

7. The I_Cl,swell generation was not affected by pirenzepine or metoctramine, and did not affect the purinergic activation of the chloride current named F_in. Thus, I_Cl,swell was not generated via neurotransmitters released during cellular swelling.

8. All together, equal discrimination for different anions, similar modulatory effects by extracellular pH, the blocking effects by ATP and La³⁺, and the same single-channel activity, strongly suggest that I_Cl,swell and S_in currents depend on the opening of the same type or a closely related class of volume-regulated chloride channels.

Follicle-enclosed Xenopus oocytes (follicles) generate ionic current responses when they are exposed to hyposmotic external media. These osmolarity-dependent current responses are carried principally by potassium and chloride ions (Arellano & Miledi, 1993, 1995). The ionic currents carried by Cl⁻ (I_Cl(osm); Arellano & Miledi, 1995; Arellano et al. 1996) seem to be very similar to the volume-regulated chloride currents, named I_Cl,swell, that have been characterised in many cell types and which may contribute to the regulatory volume decrease, a process activated in most cells in response to an hyposmotic challenge (Strange et al. 1996; Nilius et al. 1997; Okada, 1997; for consistency, this follicle osmolarity-dependent Cl⁻ current will be denoted here as I_Cl,swell). In follicles, the volume-regulated currents are principally generated in the membrane of the follicular cells. It is possible to monitor these currents with electrodes inserted into the oocyte, due to the electrical coupling via gap junction channels that exist between the oocyte and its follicular cells (van den Hoef et al. 1984; Woodward & Miledi, 1987; Arellano & Miledi, 1993). I_Cl,swell is not the only chloride current regulated by reduction of external osmolarity in follicles. Perhaps as important is a potentiation of currents that from their time course were named slow and smooth inward currents (S_in); these are elicited by several neurotransmitters, hormones, and intracellular increase of cAMP (Arellano & Miledi 1993, 1994). S_in originate also in the membrane of the follicular cells and are carried mainly by chloride ions. Several of the S_in characteristics show that they are distinct from Ca²⁺-dependent chloride currents that arise in the membrane of the oocyte itself (Kusano et al. 1982), and which are also activated by several neurotransmitters and other substances (Miledi et al. 1989). Moreover, S_in are different from another type of follicular chloride current activated by ACh or ATP, which has been named F_in from its fast time course (Arellano & Miledi, 1993; Arellano et al. 1998, 1999). The S_in elicited by hormones such as follicle stimulating hormone (FSH), or neurotransmitters such as noradrenaline and dopamine, are mimicked and potentiated by an intracellular cAMP increase, suggesting that those substances activate S_in via cAMP acting as an intermediary. Also, other studies suggest osmolarity-dependent current modulation through a cAMP increase in different cellular systems (e.g. Meng & Weinman, 1996; Carpenter & Peers, 1997). On the other hand the mechanisms of S_in activated by ACh or ATP remain poorly understood (Arellano & Miledi, 1994; Arellano et al. 1998).

In general, not much information is available on the signal transduction pathways involved in controlling the gating of the volume-regulated channels; but it seems clear that G protein-dependent mechanisms and phosphorylation pathways are involved (Doroshenko et al. 1991; Voets et al. 1998; Nilius et al. 1999). The molecular mechanisms of the gating and modulation of the volume-regulated channels are complex and our understanding of these phenomena is incipient. It is very probable that those mechanisms involve several cellular processes that might interact in order to maintain or control the cell volume, which is clearly a fundamental process for cellular functioning (Okada, 1999).

Due to the similarities between both follicle osmolarity-dependent currents it is tempting to suggest that they may be due to activation of the same type of chloride channel, even though they might be activated (or modulated) via distinct mechanistic pathways (Arellano et al. 1996). In this work we investigated whether the follicular-based I_Cl,swell and S_in (elicited by ACh or ATP) involve activation of a single type of channel.

In addition, various ionic channels and membrane receptors have been described in the ovarian follicle of many species; for several of them their function can be directly related to follicular development (e.g. gonadotropin hormone receptors). For several other molecules located in the ovarian follicle their functions are still not clear, such is the case for muscarinic and purinergic receptors and several of the ionic channels that they activate or inhibit (Arellano et al. 1996, 1999). Although the stimulation of muscarinic receptors and ionic channels have been implicated as modulators of important physiological processes, such as maturation (Dascal et al. 1984; Skoblina & Huhtaniemi, 1997), the cellular mechanisms involved are poorly understood. The presence of muscarinic and purinergic receptors in follicles is not particular to Xenopus frogs, and it has been shown that other species contain similar molecules (Eusebi et al. 1984; Kamada et al. 1994; Fritz et al. 1999). Moreover, several studies have shown possible sources for neurotransmitters within the ovarian follicle of other species (Sporrong et al. 1985; Dees et al. 1995; Mayerhofer et al. 1998). Thus, knowledge regarding the effects provoked by their stimulation might help to elucidate their role in ovarian follicle physiology.

METHODS

Cell preparation

Xenopus laevis frogs were obtained from Xenopus I (Ann Arbor, MI, USA) and Xenopus Express (Homosassa, FL, USA). Ovary lobules (4-8) were surgically removed under sterile conditions from frogs that were anaesthetised using 0.1 % 3-aminobenzoic acid ethyl ester and rendered hypothermic. After surgery, frogs were sutured, and allowed to recover from anaesthesia and hypothermia. Frogs were maintained for 15–20 days in individual tanks until healing was complete, and were then returned to a larger group housing. No further oocytes were taken for at least 2 months. After the final taking of oocytes the anaesthetised frogs were killed by decerebration and pithing. The institutional animal use committees approved this procedure. The lobules were placed in sterile unsupplemented modified Barth's medium containing (mM): 88 NaCl, 0.2 KCl, 2.4 NaHCO₃, 0.33 Ca(NO₃)₂, 0.41 CaCl₂, 0.82 MgSO₄, 0.88 KH₂PO₄, 2.7 Na₂HPO₄, with gentamicin 70 μg ml⁻¹ and adjusted to pH 7.4.

For studies on follicle-enclosed oocytes, the follicles (stage VI; Dumont, 1972) were dissected as epithelium-removed, where their inner epithelia, together with thecal blood vessels, were separated using sharp forceps. This procedure leaves the follicular cell basal membrane in place which thus provides protection and a natural environment for these cells (Arellano et al. 1998). Also, the follicle epithelium-removed dissection facilitates electrode insertion, improves the stability of electrophysiological recording and simplifies the interpretation of results by eliminating the possible participation of epithelium and other surrounding thecal tissues in the responses. Epithelium-removed follicles were incubated (18-20°C) in sterile modified Barth's medium supplemented with glucose (5 mM), and fetal bovine serum (0.1-0.2 %). In these conditions, follicular cell-oocyte electrical coupling and follicular responses can be maintained for more than 10 days.

For follicular cell isolation and culture, and patch-clamp recording (see below), the follicles (stage VI) were dissected as ‘unzipped’ follicles (Arellano et al. 1996), where the external layers were removed together with the follicular cell basal membrane. With this dissection, the follicular cells were exposed but remained attached to the vitelline envelope, allowing them to be dislodged for cell culture (Miledi & Woodward, 1989), or the unzipped follicle to be used directly for patch-clamping of follicular cells (Arellano et al. 1996). For cell cultures, unzipped follicles were enzymatically treated for 5 min in Hanks’ balanced salt solution (Gibco BRL) containing 0.05 % trypsin and 0.5 mM EDTA, then gently washed in medium containing 10 % fetal bovine serum. The treated follicles were then sucked repeatedly with a polished Pasteur pipette, and the dislodged follicular cells were recovered on glass coverslips prepared with collagen as the substrate. Follicular cells were maintained in L-15 culture medium (supplemented with 0.2 % fetal bovine serum) at 18°C (R. Reyes, S. Alshihabi, R. O. Arellano & R. Miledi, unpublished results) until their use in patch-clamp experiments.

For patch-clamp experiments on the denuded oocyte membrane, follicular cells from unzipped follicles were removed by rolling the follicles on a poly-L-lysine treated coverslip (Miledi & Woodward, 1989). After ∼6 h the defolliculated oocytes were incubated in normal Ringer solution with 50 mM sucrose for 20 min, and the vitelline envelope was removed with fine forceps to patch clamp the oocyte membrane.

Electrophysiological measurements and analysis

Whole currents in epithelium-removed follicles were monitored using a two-microelectrode voltage clamp (Miledi, 1982). Unless otherwise stated, follicles were voltage clamped at -60 mV, and superfused with normal Ringer (NR) solution containing (mM): 115 NaCl, 2 KCl, 1.8 CaCl₂, 5 Hepes, pH 7.0, or hyposmotic Ringer solution (HR₉₀), where the NaCl concentration in NR was decreased from 115 to 90 mM. HR₉₀ was used to potentiate osmolarity-dependent S_in activated via stimulation of muscarinic or purinergic receptors by ACh or ATP, respectively. Different hyposmotic Ringer solutions, with either 80 or 50 mM NaCl (HR₈₀ or HR₅₀), were used to activate I_Cl,swell. Unless otherwise stated, all follicles were injected with ethylene glycol-bis(β-amino-ethylether) N,N,N′,N′-tetraacetic acid (EGTA) (∼0.1 nmol per oocyte) and tetraethylammonium (TEA⁺) chloride (∼0.25 nmol per oocyte) by pneumatic pressure ejection, via a third micropipette inserted into the oocyte filled with a solution (100 mM EGTA plus 250 mM TEA⁺) made up in 5 mM Hepes, pH adjusted to 7.0 with KOH. The EGTA-TEA⁺ follicle loading effectively eliminated Ca²⁺-dependent Cl⁻ currents and blocked K⁺ currents elicited by muscarinic and purinergic stimulation in follicles. TEA⁺ also blocked osmolarity-dependent K⁺ currents generated in some follicles by hyposmotic medium (Arellano & Miledi, 1993). The injection apparatus was also used for extracellular delivery of ATP (1 μM in NR). For these studies ATP was ejected from a micropipette positioned ∼50 μm from the follicle, this method allowed us to generate repeatedly focal F_in currents in an area of the follicle and concurrently to apply a general stimulation through the bath superfusion (Arellano et al. 1998). In these cases the superfusion rate was lowered from 10 to 2 ml min⁻¹ in order to allow current activation by the ejected ATP.

Ionic substitutions were made in either NR or the hyposmotic solutions, replacing all the NaCl by any of the following salts: NaSCN, NaI, NaBr, NaNO₃, sodium acetate, sodium cyclamate, Na₂SO₄, or sodium gluconate. Each follicle was tested first in NR and then in one other anion solution. Changes in junction potentials generated by ion substitution (and osmolarity) using the different external solutions were measured directly. This was completed before follicle recording. The microelectrode tip potential was set to zero in the bath filled with NR, which was grounded through a salt bridge with a chlorided silver wire. Then the bath solution was replaced with each testing solution and the stable value for the potential generated was confirmed by returning to NR. The final reversal potentials (E_rev) were corrected for the respective junction potentials. Permeability ratios for the different anions (X) relative to Cl⁻ (P_X/P_Cl) were estimated using the formula:

(1)

where [Cl⁻]_o and [Cl⁻]_t are the external Cl⁻ concentrations in control and test anion-substituted solutions, respectively, ΔE_rev is the difference in reversal potential value between those two solutions, and [X⁻]_o is the external concentration of the substituting anion, while F and R are the Faraday and gas constants and T is the absolute temperature.

Blockage effects by ATP and La³⁺ were evaluated using a one-site blockade model which assumes that the effect of the compounds is fast (Woodhull, 1973; Garcia & Miledi, 1996). The model estimates the half-blocking concentration (IC₅₀(0)) of the compounds at a membrane potential of 0 mV, and the fraction of the membrane potential sensed by the binding site for the blocking agent, applying the following equation:

(2)

where I_Cl is the amplitude of the chloride current studied, either the I_Cl,swell or S_in in the absence of the blocker B (ATP or La³⁺), and I_Cl(B) is the current in the presence of the blocker. [B] is the concentration of the blocker, V_m the membrane potential, δ the electrical distance of the binding site measured from the extracellular side, z the valence of the blocker, while F and R are the Faraday and gas constants and T is the absolute temperature.

Membrane currents in either isolated follicular cells in culture, in freshly dissected (within the first 6 h) unzipped follicles, or denuded (i.e. defolliculated) oocytes, were recorded using the cell-attached configuration of the patch-clamp technique. Unzipped follicles and denuded oocytes were not loaded with EGTA-TEA⁺. Polished patch pipettes were made from borosilicate glass capillaries and had a resistance of 5–8 MΩ when filled with NR. Currents were monitored using a patch-clamp amplifier APC-8 (Medical Systems, Japan) and stored on-line in a DAT recorder, for their later digitisation and analysis using the pCLAMP6 and Axoscope software (Axon Instruments, Foster City, CA, USA).

Substances

ACh, ATP, EGTA, collagenase type I, fetal bovine serum, collagen, 3-aminobenzoic acid ethyl ester, and all salts were from Sigma Chemical Co. (St Louis, MO, USA). Pirenzepine dihydrochloride and metoctramine tetrahydrochloride were purchased from RBI (Natick, MD, USA). TEA chloride was from Baker Co. (Phillisburg, NJ, USA). Trypsin, gentamicin, and L-15 culture medium were purchased from Life Technologies (Gibco BRL, Rockville, MD, USA).

RESULTS

Characteristics and anion permeability of I_Cl,swell and S_in follicle currents

The follicles loaded with EGTA and TEA⁺ had a resting potential in NR of -29.3 ± 6.7 mV (all results quoted are means ±s.e.m. unless otherwise stated), and an input resistance of 0.37 ± 0.03 MΩ (165 follicles, 14 frogs). In general, the I_Cl,swell generated by HR₈₀ reached a peak amplitude in 4–6 min (Fig. 1A), and inactivated slowly when follicles were maintained in this solution. However I_Cl,swell deactivated rapidly, and in most follicles completely, upon returning to NR solution. Defolliculation (manual or enzymatic) completely eliminated the I_Cl,swell elicited by superfusion with hyposmotic solutions (HR₈₀ or HR₅₀), while EGTA injection (together with TEA⁺) failed to abolish it, thus confirming that this current is follicular cell based and Ca²⁺ independent (Arellano & Miledi, 1993, 1995). The I_Cl,swell had a reversal potential (E_rev) of -23.5 ± 0.7 mV (78 follicles, 14 frogs), close to the equilibrium potential for Cl⁻ ions in Xenopus follicles (Kusano et al. 1982) and very similar to that for other Cl⁻ carried currents in follicles (Arellano et al. 1998).

Figure 1. Osmolarity-dependent membrane currents elicited in Xenopus follicles.

A, I_Cl,swell elicited by HR₈₀ in a follicle superfused originally in NR, and F_in and S_in elicited by ACh and ATP in follicles bathed in HR₉₀. All different epithelium-removed follicles were held at -60 mV. In this and subsequent records test solutions and drugs were applied during times indicated by bars at top, and steps to -40 mV were applied periodically to monitor membrane conductance. B, current-voltage relationships for I_Cl,swell (•), and S_in elicited in HR₉₀ by 50 μM ACh (○) and 50 μM ATP (▵); steady-state currents at different potentials were measured by application of a voltage step series in control conditions and at the peak of the currents. Superimposed records are examples of currents during the voltage steps to +40 mV during the indicated conditions. Data (here and throughout are quoted as means ±s.e.m.) from follicles of 7–14 frogs.

Follicle osmo-dependent responses to ACh or ATP had the same characteristics described previously (Arellano & Miledi, 1993; Arellano et al. 1998). Briefly, in NR both agonists evoked principally a fast inward current (F_in) followed by a slow and smooth current (S_in). The S_in current was potentiated several (5-8) times when the same concentrations of ACh or ATP were applied in an external hyposmotic medium, such as HR₉₀ (Fig. 1A and B). S_in elicited by ACh or ATP were dependent on the presence of the follicular cells and their coupling with the oocyte, and were also Ca²⁺ independent (Arellano & Miledi, 1993). In HR₉₀S_in had an E_rev of -25.8 ± 0.6 mV (26 follicles, 11 frogs) and of -29 ± 1.4 mV (21 follicles, 7 frogs) when elicited by ACh or ATP, respectively (Fig. 1B).

Usually, follicles in HR₉₀ (a solution with a similar osmolarity to Barth's incubation medium) had a lower input resistance (0.14 ± 0.01 MΩ; n= 80) than that of the same follicles in NR (0.2-0.4 MΩ, cf. Arellano & Miledi 1995). Apparently, the lower resistance is mainly due to the presence of a steady inward current in this hyposmotic solution (Arellano et al. 1996). This current did not seem to determine the presence or amplitude reached by I_Cl,swell or S_in current, because these were not directly correlated with the amplitude of the steady current. Due to the fact that the later current was blocked by La³⁺ (1 mM, see below) and it seemed to be carried mainly by Cl⁻ ions (not shown), we suggest that this current corresponds with a population of I_Cl,swell channels that remains active in Barth's or other hyposmotic solutions.

The amplitudes of both I_Cl,swell and S_in were variable among follicles from different frogs, as has been shown in previous studies. Here, I_Cl,swell generated by HR₈₀ had amplitudes that ranged from a few nanoamperes to 3.9 μA, and S_in elicited in HR₉₀ activated either by ACh or ATP were also in the range of a few nanoamperes to 1.2 μA.

Given the similarity in the general characteristics of I_Cl,swell and S_in, we proceeded to compare their channel permeability to the following anions: Cl⁻, Br⁻, I⁻, NO₃⁻, SCN⁻, SO₄²⁻, acetate (acet), cyclamate (cycl), and gluconate (gluc). The permeability ratios for the different anions with respect to chloride (P_X/P_Cl) were calculated applying eqn (1). For this purpose, current-voltage relationships were obtained for both osmolarity-dependent responses by giving a series of voltage pulses (duration of 650 ms) applied in the range of -100 to +40 mV (20 mV steps) from a holding potential of -60 mV.

For I_Cl,swell the control voltage pulses were applied during superfusion of the testing solution, where all NaCl in NR was substituted by the corresponding testing salt (NaX; e.g. NaI, sodium cyclamate). Then voltage pulses were reapplied during the peak I_Cl,swell elicited by reducing the osmolarity of the testing solution from 115 mM NaX to either 80 or 50 mM NaX. For every voltage step, the control pulse currents were subtracted from those obtained during I_Cl,swell, and these values were plotted as in Fig. 2A.

Figure 2. Permeability ratios for I_Cl,swell and S_in.

Current-voltage relationships for A, I_Cl,swell generated by HR₈₀ or HR₅₀, and S_in elicited in HR₉₀ by application of 50 μM ACh, B, or 50 μM ATP, C, in Ringer solution containing NaCl (•), or in solutions in which chloride was substituted by SCN (▴) or cyclamate (□). D, permeability ratios for different anions of I_Cl,swell (•) and S_in elicited by ACh (▵), or ATP (○). Data are from 4–10 follicles (3-8 frogs) in each condition.

For S_in, follicles maintained in Barth's medium were initially superfused in HR₉₀ solution, thus most I_Cl,swell was inactivated. The follicles in HR₉₀ were superfused with the control testing solution where the NaCl was substituted by 90 mM NaX, and a series of voltage pulses was applied in order to obtain the control membrane current. Then 50 μM ACh or 50 μM ATP was applied in the same testing solution and voltage steps were reapplied at the S_in peak amplitude. At this time (∼1 min) the F_in also elicited by both neurotransmitters was usually completely inactivated. Therefore, the F_in did not contribute significantly to the peak current of S_in. Again, for every voltage step, control membrane currents in HR₉₀ with chloride substituted by the testing anion were subtracted from those obtained during S_in elicited by ACh or ATP and plotted as in Fig. 2B and C, respectively.

The values for E_rev for I_Cl,swell and S_in, the latter elicited by either ACh or ATP, were estimated from the current-voltage (I–V) relations obtained in each test solution. The ΔE_rev was then used to estimate the permeability ratios P_X/P_Cl that were plotted in Fig. 2D. All three osmolarity-dependent currents presented the same sequence of permeability for the different anions: SCN⁻ > I⁻ > Br⁻≥ NO₃⁻≥ Cl⁻ > gluconate ≥ cyclamate > acetate > SO₄²⁻.

These results suggested that I_Cl,swell and S_in currents elicited either by ACh or ATP, were driven through a similar type of anionic channel. To explore this further, we performed pharmacological studies on the osmolarity-dependent currents, and patch-clamp recording on the follicular cell and oocyte membranes to obtain direct information on the characteristics of the channels involved.

pH modulation and inhibition by ATP and lanthanides on I_Cl,swell and S_in

There are two pharmacological effects that have been consistently shown on the I_Cl,swell from several cellular types including Xenopus follicles (Ackerman et al. 1994; Jackson & Strange, 1995b; Voets et al. 1996), which are: (1) a blocking effect by extracellular ATP, and (2) modulation of the time course of I_Cl,swell inactivation at positive potentials by extracellular pH. Thus, these two aspects were studied here with respect to their possible effects on the generation of S_in elicited by neurotransmitters.

The S_in currents were elicited by ACh or ATP at different maximal concentrations from 10 μM to 1 mM, and voltage steps to different potentials were applied at the peak amplitude of the S_in generated (Fig. 3A and B). The inward peak current amplitudes were similar (at -30 to -120 mV) independent of the agonist and the maximal concentration used, but when 1 mM ATP was applied, the outward currents were potently inhibited (13 follicles, 3 frogs). This blocking effect was voltage dependent, being stronger at positive potentials (Fig. 3A), suggesting that the binding site for ATP is located inside the S_in channel pore. When this effect was analysed using a one-site blockade model (eqn (2)), the parameter δ that indicates the fraction of the membrane electrical field sensed by the binding site for ATP was estimated as 0.26 ± 0.03. And the IC₅₀ at a membrane potential of 0 mV was 0.57 ± 0.12 μM (see Fig. 5C). The inhibitory effect by 1 mM ATP was specific for this agonist because a similar concentration of ACh did not produce blockage (6 follicles, 2 frogs). This effect of ATP on the S_in is similar to that reported for I_Cl,swell in other cells (Strange et al. 1996).

Figure 3. Effects of extracellular ATP and pH on S_in.

A, current-voltage relations of S_in elicited by 50 μM ATP (•), 1 mM ATP (○), or 1 mM ACh (▵). Data are from 5–13 follicles (2-3 frogs) in each condition held at -60 mV. B, superimposed current traces representative of the S_in elicited by the agonists at membrane potentials from +40 to -120 mV. Horizontal lines indicate zero current. C, superimposed current traces obtained at the peak of S_in elicited by 50 μM ACh at the membrane potentials stepped to +80 and +60 mV from a holding potential of -60 mV. At each potential the set of current traces corresponds with S_in generated in extracellular medium buffered to pH 8, 7, or 6. The inactivation kinetics at +80 mV were fitted to simple exponentials and the mean of the time constants (τ) in each value of extracellular pH were plotted in the bar graph. The open bars are the mean τ inactivation for the I_Cl,swell generated by HR₈₀ (3 follicles, 2 frogs), while the filled bars are those for the S_in elicited by 50 μM ACh in HR₉₀ (7 follicles, 2 frogs) in follicles from the same frogs.

Figure 5. Voltage dependence of La³⁺ block on I_Cl,swell and ATP on S_in.

A, current-voltage relations of I_Cl,swell elicited by HR₅₀ in presence of 0 μM (•), 10 μM (○) or 25 μM La³⁺ (▴). Data are the means of 6 follicles (2-3 frogs) in each condition held at -60 mV. B, superimposed traces are examples of I_Cl,swell in each condition. C, comparison of the voltage dependency of La³⁺ block on I_Cl,swell (•, 10 μM; ○, 25 μM) and of 1 mM ATP on S_in (□, data from Fig. 3A); chloride currents in each potential in the absence (I_Cl) and the presence of the blocker (I_Cl(B)) were analysed using eqn (2) that was linearized. Parameters from the lines fitted to the data points were used to estimate the IC₅₀ at 0 mV and δ values.

The other characteristic of I_Cl,swell is that the time course of its inactivation process at positive potentials is accelerated by a decrease of extracellular pH. We have confirmed the effect of pH on the inactivation kinetics of I_Cl,swell and studied this for the S_in elicited by ACh. Although similar experiments were made for ATP, this agonist at lower pH had additional effects that seemed to involve modifications in the interaction of the agonist with the purinergic receptor, and further experiments will be necessary to completely evaluate them. Thus, the S_in was elicited by 50 μM ACh in HR₈₀, and at the peak current amplitude voltage steps to +60 and +80 mV were applied, first in HR₈₀ at pH 7.0. After a wash period of 15 min, 50 μM ACh was again superfused in solution adjusted to pH 8.0 or 6.0, and the depolarizing voltage steps were applied at the peak S_in. The S_in inactivation kinetics was accelerated by reducing the external solution pH from 8.0 to 6.0 at both membrane potentials (Fig. 3C), as was shown by the time constants of the inactivation time course estimated by fitting with single exponentials. The time constants of inactivation at +80 mV for the S_in currents elicited by ACh (7 follicles, 2 frogs) were 347 ± 18 ms for pH 6.0, 514 ± 39 ms for pH 7.0 and 760 ± 46 ms for pH 8.0. The strength of this effect is very similar to that observed on the inactivation rates of I_Cl,swell in follicles (cf. Voets et al. 1996), and experiments eliciting I_Cl,swell in follicles (n= 3 at each pH) from the same frogs used here, confirmed this result (Fig. 3C).

Thus these results indicated that S_in showed a similar sensitivity to extracellular pH and ATP to that demonstrated for I_Cl,swell, not only from Xenopus follicles but also that from other cell types. This strengthened the suggestion that both currents were carried through a similar channel.

One important characteristic of I_Cl,swell in the Xenopus follicle is its sensitivity to lanthanides, specifically La³⁺ (Ackerman et al. 1994; Arellano & Miledi, 1995) and Gd³⁺. This effect on volume-activated chloride channels has been reported also for other cells (Robson & Hunter, 1994). Two aspects were studied here. First it was of interest to know if S_in currents elicited by agonists were similarly blocked by lanthanides (Fig. 4), and second we investigated the voltage dependency of the blocking effect on I_Cl,swell in order to obtain some information about the inhibition mechanisms involved (Fig. 5).

Figure 4. Follicle osmo-dependent currents inhibition by La³⁺ and Gd³⁺.

Sample records of follicles held at -60 mV and superfused with solutions containing increasing concentrations of La³⁺ (0, 50, 100 and 250 μM) in A, and 0, 10, 25, 50, 100 and 250 μM in B, or Gd³⁺ (0, 10, 50, 100, 250 and 500 μM) in C, during the activation of S_in elicited by 50 μM ACh in HR₉₀ (A) or I_Cl,swell (B and C). Dose-response relations of the inhibition were obtained with similar experiments in which the currents in each concentration of the lanthanide were normalized with respect to the current obtained in its absence (I_max). Data points are the means of 2–23 follicles (2-8 frogs). Curves are fits to the equation:

where IC₅₀ is the concentration at which the lanthanide blocks 50 % of the total I_Cl,swell or the fraction of S_in current sensitive to La³⁺, n a slope factor, and [X³⁺] the concentration of La³⁺ or Gd³⁺ in the extracellular solution. Dose-response inhibition relations were obtained in follicles held at -60 mV for S_in, and for I_Cl,swell at -60 mV (▪), -100 mV (▿) and -40 mV (○).

The S_in currents were elicited in HR₉₀ by 50 μM ACh, in the absence or presence of different concentrations of La³⁺ that were first applied alone for ∼1 min. In its maximal effect, La³⁺ (around 500 μM) blocked more than 55 % of S_in, with an apparent IC₅₀ of 30 ± 2.6 μM (Fig. 4A). The effect of La³⁺ was completely reversible after washing in HR₉₀ for 10 min, and was specific on S_in because the ACh-elicited F_in was not affected by the polycation. This strongly suggested that lanthanides had similar effects on I_Cl,swell and S_in currents, and further supported the notion that both currents are carried via the same type of channel.

The mechanism of lanthanide inhibition of I_Cl,swell is unknown. Among other possibilities, it might be due to a direct blocking effect in the pore of the channel. Therefore, we obtained dose-response relations for the effects of La³⁺ and Gd³⁺ on I_Cl,swell in follicles held at different potentials. We also obtained current-voltage relations in the presence and the absence of La³⁺, which were analysed using a one-site blockade model (Fig. 5). Similar to that made for the analysis of S_in inhibition, follicles were first superfused with NR solution containing a known concentration of either La³⁺ or Gd³⁺, after ∼1 min in this solution, the follicles were then exposed to HR₅₀ containing the same testing concentration of the lanthanide. The I_Cl,swell generated in the presence of each concentration of the polycation was normalized with respect to the current generated in its absence. The dose-response relations of inhibition were constructed in follicles held at -100, -60 and -40 mV (Fig. 4B and C). At a holding potential of -60 mV the IC₅₀ for La³⁺ and Gd³⁺ were 17.1 ± 1.9 μM (23 follicles, 6 frogs) and 20.9 ± 1.4 μM (11 follicles, 4 frogs), respectively. Similar IC₅₀ values were obtained in follicles held at -40 mV; 9.1 ± 0.8 μM for La³⁺ (2 follicles, 1 frog) and 24.8 ± 2.3 μM for Gd³⁺ (2 follicles, 2 frogs). And at -100 mV where IC₅₀ was 14.2 ± 5.9 μM for La³⁺ (17 follicles, 8 frogs) and 6.7 ± 3.6 μM for Gd³⁺ (6 follicles, 3 frogs), indicating that in general the inhibition effect is not strongly voltage dependent (Fig. 4B and C, bottom).

Similar results were obtained when the La³⁺ blocking effect on I_Cl,swell was evaluated obtaining the I–V relations in the absence and the presence of the polycation. For this, I_Cl,swell were elicited by HR₅₀ superfusion and I–V curves were obtained by stepping the potential from +40 to -120 mV at the peak current; this procedure was repeated in the presence of 10 or 25 μM La³⁺ (6 follicles, 2 frogs). From the I–V curves obtained (Fig. 5A) it is clear that the potency of the inhibitory effect of La³⁺ on I_Cl,swell behaved linearly at the entire voltage range tested, a result that is different from that observed for the inhibition produced by ATP on S_in (Figs 3A and 5C). When the I–V curves were analysed using a one-site blockade model, the parameter δ was estimated to be very close to 0, and was 0.03 ± 0.013 for 25 μM La³⁺ and 0.016 ± 0.01 for 10 μM. These results suggested that the site for La³⁺ and Gd³⁺ actions did not sense the transmembrane voltage, and that the I_Cl,swell inhibition was not due to a blocking effect within the pore of the channel.

Recording of osmo-dependent ion channel currents in the follicular cells

More direct evidence for the similarity between I_Cl,swell and S_in channels was obtained by monitoring the membrane channel activity in patches of follicular cells either isolated and maintained in culture, or recorded from the surface of unzipped follicles. Patches from both preparations behaved similarly. Patch-clamp recordings, in the cell-attached configuration, were obtained from follicular cells maintained in NR. After a high resistance seal was formed, with pipettes filled with NR, superfusion of the cells with HR₉₀ activated inward membrane currents (Fig. 6). Single channel activity was resolved in most of the patches made (Fig. 6A and B), and the unitary conductance was 4.7 ± 0.4 pS (Fig. 6B). Unzipped follicles had an average membrane potential of -55 ± 8 mV (13 follicles, 3 frogs) in NR, as measured with an electrode inserted into the oocyte. Assuming a follicular cell resting membrane potential of -60 mV in NR, the activated channels had an E_rev of approximately -20 mV, as expected for osmolarity dependent and other currents carried by chloride ions in follicles (16 unzipped follicles, 3 frogs, and 32 isolated follicular cells, 4 frogs) (e.g. Kusano et al. 1982). Accordingly, potassium channels (Arellano et al. 1996) activated in some patches (not shown) had an E_rev of -95 ± 6 mV, which is also that expected for these currents.

Figure 6. Osmo-dependent single channel activity generated in follicular cell membrane.

A, cell-attached patch-clamp current recorded at a pipette potential (V_p) of +50 mV from a follicular cell (unzipped follicle) maintained in NR (a). In b subsequent superfusion with HR₈₀ solution opened inward membrane currents, that increased in activity when the hyposmotic solution was applied together with 100 μM ACh (c–d). Activity returned to basal levels when superfused back with NR (e). S0, S1 and S2 signal the opening level of at least 2 channels in this patch. Similar records were obtained in another 8 unzipped follicles and in 5 isolated follicular cells maintained in culture (7 frogs). B, current-voltage relationship for single-channel currents opened by HR₈₀ in an unzipped follicle. A similar relation was obtained for single currents in presence of ACh or ATP. Records are examples of currents at different V_p; at least three channels were present in this particular case. Data represent the means of single current in five different patches performed in cells from three frogs. All records were digitised at 1 kHz and filtered at 100 Hz.

The single inward current activity was potentiated when HR₉₀ was superfused with 50 μM ACh (Fig. 6Ac and d) or ATP. The increase in activity was evidenced by an increase in the open time of the channels (Fig. 6Ac), and also in several patches by the number of units activated. The channels frequently remained open for long periods (several seconds), and occasionally several channels gated in concert, producing larger openings or closings (not shown). The unitary conductance did not change in amplitude when the agonist was superfused (see Fig. 6Ac and d). Osmolarity-dependent current activity was almost completely reversible after some seconds of superfusion with NR (see Fig. 6Ae). It seems that the channels are present at high density since very few (2 from > 40) patches of follicular cells had a single channel, and they regularly contained three or more (> 10) channels.

In several patches from unzipped follicles some channels were active even in NR, or they opened infrequently in this condition (Fig. 6Aa). In follicular cells maintained in culture this phenomenon seemed to be increased and many membrane patches displayed a large basal activity, apparently due to the opening of various channels with characteristics similar to those activated by HR. However, their activity and/or their number were always potentiated upon application of hyposmotic solutions. Thus, it is clear that the increase in single-channel current activity is a consequence of hyposmotic conditions. Because the single-channel conductance was the same in the absence or presence of the agonists, the potentiation of S_in is probably due to an increase in the number and open-time of channels activated. In patches performed on the manually defolliculated oocyte membrane (10 oocytes, 3 frogs) no current activity was generated by HR₈₀, alone or together with ACh or ATP, even after periods of 15 min of continuous superfusion with hyposmotic solution.

Therefore, these results show that HR solution alone or together with ACh or ATP activated inward currents with similar unitary characteristics in the follicular cell membrane but not in the oocyte per se. Since single-channel currents were detected in practically all the follicular cell patches studied, it seems very likely that these contribute significantly to the generation of total I_Cl,swell and S_in.

The S_in and neurotransmitter release induced by cell swelling

Cell swelling induces the release of various substances from the intracellular medium; several of them participate in the regulatory volume decrease that tends to restore the cellular volume. One of the molecules that has been shown to be commonly released in this condition is cellular ATP (Taylor et al. 1998; Hazama et al. 1999). Since our results seemed to indicate that I_Cl,swell and S_in were carried via a channel type with similar characteristics, it is possible that the release of substances produced by cellular swelling might be related to the generation of I_Cl,swell or S_in. At least two mechanistic relations can be directly proposed: (1) stimulation by the agonists produces follicular cell swelling and the concomitant I_Cl,swell (S_in) activation, and (2) cell swelling elicited by hyposmotic solution produces the release of ACh and/or ATP with the consequent generation of S_in (I_Cl,swell). The first suggestion was not directly examined here; however, some previous indirect evidence (Arellano et al. 1996) seems to indicate that agonists did not produce or potentiate swelling of the cells (see discussion). The second possibility was assessed with the following experiments. For the case of ACh release and receptor activation we have compared the development of I_Cl,swell (activated by HR₈₀) in the presence of potent and selective antagonists of muscarinic receptors. It is known that follicles are endowed with at least two different muscarinic receptors very similar to M3 and M2 subtypes, which are strongly blocked by substances such as pirenzepine and metoctramine (Kusano et al. 1982; Arellano et al. 1999). However, the amplitude and rate of development of I_Cl,swell was not affected by addition of these muscarinic antagonists (10 μM) in the superfusion medium (7 follicles, 3 frogs; Fig. 7A). In follicles from the same frogs the typical muscarinic response in the oocyte is the production of an oscillatory current (I_Cl(Ca); activation of this current was made in follicles not injected with EGTA-TEA⁺) and the F_in. Both were potently inhibited by the antagonists (6 follicles, 3 frogs; Fig. 7B). This result suggested that if release of ACh exists during cell swelling produced by the hyposmotic challenge, the possible activation of muscarinic receptors in this condition did not importantly affect the generation of I_Cl,swell.

Figure 7. I_Cl,swell generation and its relation with a likely ACh or ATP release elicited by cell swelling.

A, current traces correspond to the control I_Cl,swell activated by HR₈₀ superfusion alone (upper trace) and in the presence of 10 μM pirenzepine (lower trace), a potent antagonist of muscarinic receptors in follicles. Similar results were obtained in 7 follicles more from 3 frogs. B, in follicles from the same frogs used in A, the muscarinic responses were completely inhibited by superfusion with the antagonist. The example traces showed that the typical oscillatory Ca²⁺-dependent chloride current elicited by 10 μM ACh was effectively eliminated by pirenzepine. C, F_in currents were elicited by short pulses of 1 μM ATP ejected periodically from a pipette positioned close to the follicle; each ejection pulse was applied after the voltage step of 20 mV (arrow). For the time indicated, 5 μM ATP in NR was superfused; while this solution is present in the bath and after its corresponding F_in was elicited, the F_in generated by the 1 μM ATP jets was potently inhibited. Similar results were obtained in 8 follicles from 3 different frogs. D, similar F_in currents generated by 1 μM ATP pulses were not decreased when I_Cl,swell was simultaneously generated; nevertheless, a transient potentiation of 30–60 % (*) was noted during the first minutes during I_Cl,swell activation in 4 experiments from 10 follicles tested (3 frogs).

Another possibility is the involvement of ATP release, but potent antagonists of follicle purinergic receptors are not known. However, it has been shown that F_in is strongly inactivated by several seconds in the presence of ACh or ATP (Arellano et al. 1998). If a swelling-elicited release of ATP occurs during HR superfusion one would anticipate that in this condition F_in might exhibit inactivation. In previous studies we have shown that F_in currents are not strongly affected during hyposmotic treatment (cf. Fig. 26 in Arellano et al. 1996), and the application of hyposmotic medium itself does not generate ionic currents with the characteristics of F_in (R. Miledi & R. O. Arellano, unpublished results). To confirm these observations, the follicles were stimulated focally by ejecting brief (0.2-0.5 s) jets of NR containing 1 μM ATP while the superfusion medium was set to 2–3 ml min⁻¹. This stimulation allows F_in to be generated repetitively; if the ejection interval is adequate (> 40–60 s), F_in recovers well from each stimulation and the amplitude is maintained for several minutes. In these conditions, when ATP was superfused in similar concentrations (1-10 μM) to those released by cells exposed to hyposmotic medium (Hazama et al. 1999), the F_in were strongly inhibited during ATP application and remained so for several seconds (Fig. 7C; 8 follicles, 3 frogs). Contrary to this, during I_Cl,swell activation by HR superfusion (Fig. 7D), F_in elicited by the ATP jets were not abolished or decreased (10 follicles, 3 frogs). In fact, consistently during a short period of I_Cl,swell activation the F_in seemed to be potentiated (asterisks in Fig. 7D).

These results suggest that any release of the neurotransmitters, ACh or ATP, associated with follicular cells or oocyte swelling did not strongly affect I_Cl,swell generation. Further studies will be necessary to evaluate the possibility that stimulation by neurotransmitters provoke swelling of the follicular cells and consequently the activation of I_Cl,swell.

DISCUSSION

In the present study we show that follicle I_Cl,swell and S_in are carried via anionic channels with similar permeability, unitary conductance, and pharmacological characteristics, including sensitivity to extracellular pH, ATP and lanthanides. All this strongly suggests that both follicle currents flow through the same type of Cl⁻ channels or a closely related class of volume-regulated channels. In addition, both follicle currents are similar to the I_Cl,swell from other cell types that are activated by external hyposmotic conditions, and which participate in the mechanisms of regulatory volume decrease by allowing a net efflux of Cl⁻, organic anions and osmolytes (e.g. Pasantes-Morales & Schousboe, 1988; Worrell et al. 1989; Lewis et al. 1993). The channel discrimination for different anions by I_Cl,swell (S_in) Xenopus channels is also similar to that reported previously (Arellano & Miledi, 1993; Ackerman et al. 1994; Voets et al. 1996), confirming that the selectivity sequence is lyotropic, as has been observed for many anionic channels. Once again, the differences in P_X/P_Cl for different anions discriminate between I_Cl,swell and other chloride channels which are present in the follicle, namely the Ca²⁺-dependent Cl⁻ channels present in the membrane of the oocyte itself, and the F_in channels located in the follicular cells (Arellano & Miledi, 1993; Arellano et al. 1998).

One conclusion of this work is that in Xenopus follicular cells I_Cl,swell channels are regulated by neurotransmitters via activation of muscarinic and purinergic receptors. Since the properties of S_in activated by follicle-stimulating hormone (FSH) and intra-oocyte injection of cAMP (Arellano & Miledi, 1993, 1994) are not different from those elicited by ACh or ATP, we suggest that these currents are also due to activation of I_Cl,swell channels. Indeed, preliminary results in follicular cells show that FSH generates single-channel activity, similar to that described here (R. O. Arellano & R. Miledi, unpublished results). The complete sequence of events involved in activating S_in still remains unknown. However, the possibility remains that stimulation by neurotransmitters (or hormones) provokes swelling of the follicular cells and hence the activation of S_in (I_Cl,swell). We have no direct evidence to support or reject this suggestion; nevertheless, some of our previous results seem to indicate that cellular swelling might not be the only mechanism involved in the activation of S_in by neurotransmitters. For example, assuming that I_Cl,swell generation maintains a close relationship with the time course of cell swelling, one would expect that potentiation of S_in might follow a similar time course for I_Cl,swell activation. However, this is not the case. Potentiation of S_in starts very early (seconds) after an hyposmotic medium is applied to the follicle, while I_Cl,swell is activated several seconds later and with a slower time course (cf. Fig. 26 in Arellano et al. 1996). Thus, S_in are near their maximum, 75–90 %, during the first minute in the hyposmotic medium, but the peak amplitude of I_Cl,swell is reached several minutes later, regularly 6–10 min. In conclusion, even though I_Cl,swell and S_in seemed to be carried via the same channel type, their generation seemed to occur independently, and the activation mechanisms seemed to differ, at least in some steps of the pathway used. An obvious possibility is that agonists modulate the S_in (I_Cl,swell) channels, producing facilitation upon a smaller change in osmolarity. Thus, activation of the channels might be through the same mechanism that activates I_Cl,swell, but the actions provoked by the agonists sensitise the channels to a smaller reduction in osmolarity (and cell-volume change); this mechanism might involve intracellular signalling pathways. In this context, it is well known that in other cells, G proteins and phosphorylation mechanisms might be involved in modulating I_Cl,swell. For example, it has been suggested that G proteins regulate the sensitivity of I_Cl,swell to cell swelling, and that this might be mediated through tyrosine phosphorylation (Voets et al. 1998; Nilius et al. 1999). It is possible that such modulation is not restricted to phosphorylation through a single type of protein kinase, and that in Xenopus follicular cells several signalling systems participate in modulating I_Cl,swell channel sensitivity. An example of this might be represented by the S_in currents activated via stimulation of receptors coupled to an increase in cAMP synthesis (Arellano & Miledi 1994; Arellano et al. 1996). Further studies are necessary to elucidate the mechanisms implicated in the actions of cAMP on I_Cl,swell modulation, as well as those activated by stimulation of muscarinic and purinergic receptors in follicular cells (Arellano et al. 1998) during S_in generation. Likewise, it will be important to study the effects of osmolarity-dependent current activation on volume control of the follicular cells and the whole follicle. The results presented in this study suggest that the activation of volume-regulated channels in follicular cells can be influenced by external signals, which might be part of a systemic mechanism for cell volume control via transmitters.

It will also be important to investigate whether I_Cl,swell in follicular cells might be regulated in a similar way as has been postulated in other cells. For example, caveolines have recently been postulated as modulators of the activity of volume-regulated channels (Trouet et al. 1999) in different cell lines. This protein is the main protein component of caveolae, which are invaginations that maintain microdomains within the plasma membrane (Anderson, 1998). Caveolaes have been proposed as structures associated with the activation mechanism of the volume-regulated channels (Okada, 1997, 1999), providing a scaffolding on which these channels can be assembled in a pre-activated form. In the plasma membrane microdomain formed by these invaginations are located a variety of membrane proteins which are involved in cell signalling, such as membrane receptors, G proteins, and other signal transduction molecules (Anderson, 1998; Okamoto et al. 1998). These molecules might be involved in the regulation of ionic channels.

Here, we confirm that I_Cl,swell is strictly dependent on the presence of follicular cells and their coupling with the oocyte (Arellano & Miledi, 1993, 1995). Osmolarity-dependent activated channels are not recorded from the membrane of denuded oocytes that were defolliculated manually; therefore, these currents seem to originate totally in the membrane of the follicular cells.

Single channels activated by hyposmotic solutions in follicular cells have a conductance of about 5 pS and their I–V relation is linear in the range of ±100 mV. This single-channel conductance seemed to be smaller that the one recorded in other cells (∼10 pS), and similarly its linear I–V behaviour does not seem to be in accord with the outward rectification of other I_Cl,swell channels. Further experiments will be necessary to confirm these characteristics. Nevertheless, it has been noted previously that rectification observed for the osmolarity-dependent currents of the whole follicles is weaker than that shown in several other cell types (Voets et al. 1996). A possibility is that the weak outward rectification of the I_Cl,swell in follicles may result from other factors such as changes in the number and open-time of the channels at positive potentials. A singular characteristic of the channels was their prolonged open-time, especially when ACh or ATP was applied. This characteristic is similar to that of volume-sensitive organic osmolyte and anion channels which switch abruptly from a closed to an open state, where their probability to stay in this conformation is near unity (Jackson & Strange, 1995a). Although the single-channel currents recorded here were not directly shown to be part of the whole I_Cl,swell or S_in, they were consistently seen to be activated by hyposmotic conditions. Moreover, their E_rev values and time course were coincident with that of osmolarity-dependent currents, and we believe that their participation is fundamental to the generation of these currents in the whole follicle. Studies detailing the electrophysiological and pharmacological characteristics of these ionic channels in follicular cells will help to determine their participation in generating total I_Cl,swell and S_in.

Lanthanum and Gd³⁺ appear to block I_Cl,swell through a voltage-independent mechanism, strongly suggesting that these ions are not acting within the pore of the channel. Contrary to this, the ATP blocking effect was voltage dependent and sensitive to the direction of the net Cl⁻ flux, similar to what has been shown in other cells (Jackson & Strange, 1995b), suggesting in this case that the binding site for ATP is located in the channel pore. The binding site for the polycations may still be on the channel protein, but located in a domain that does not sense the transmembrane voltage field, i.e. in residues that do not form the pore. A possible explanation is that the binding of the lanthanides to an external domain of the channel protein favours a conformation in the closed state. Alternatively, the inhibition effect might involve a binding site that indirectly interferes with the channel activation mechanism. However, some observations suggest that the effect of lanthanides is complex and probably involves more than one site of action. For example, there is some variability observed using different superfusion protocols. Application of La³⁺ directly on I_Cl,swell inhibits the current with an IC₅₀ of ∼100 μM (Arellano & Miledi, 1995), while here a short (0.5-1 min) preincubation of the follicle with the cation alone gave an IC₅₀ of 17 μM. Similarly, it has been reported that in some circumstances the La³⁺ blocking effect is irreversible (Ackerman et al. 1994). A mechanism of inhibition produced by lanthanides which involves actions on the activation mechanism versus one acting in the channel pore, might also explain the fact that lanthanum did not inhibit S_in as potently as I_Cl,swell. Gd³⁺ seemed to be relatively more potent on I_Cl,swell in follicles held at -100 mV than at -40 or -60 mV, and more experiments are needed to define a possible weak voltage dependency.

In summary, in this study we show results which strongly suggest that in Xenopus follicles the activation of I_Cl,swell and S_in involves the opening of a single type of ionic channel, alternatively, the two currents might involve different but closely related channels, which conform to a class of volume-regulated chloride channels. It appears that I_Cl,swell channels are opened, or sensitised to the activation mechanism (swelling), via stimulation by neurotransmitters and hormones that act on specific receptors in the follicular cell membrane. For example, the sensitisation mechanism may facilitate the activation of the channels producing a shift to a lower cellular volume threshold, as has been suggested (Voets et al. 1998). This mechanism of regulation of I_Cl,swell channels via signalling pathways may be important for volume control of the cells, and thus affect different aspects of ovarian physiology. In addition, we show that the endogenous osmolarity-dependent channels are mainly located in the membrane of the follicular cells, as has been previously suggested (Arellano & Miledi, 1993, 1995; Voets et al. 1996). Therefore, the oocyte expression system is clearly suitable for further experiments on the structural and functional details of volume-regulated chloride channels.