Intracellular Free Potassium, Sodium and Chloride Measured with Ion-Selective Microelectrodes from Salivary Gland Cells of the Snail Planorbis Corneus

The ionic mechanisms underlying the secretion of salt and water by exocrine glands are only partially understood. It is generally agreed that the interaction of agonists with secretory cells leads to an increase in intracellular free Ca²⁺ ([Ca²⁺]₁) in the cytosol of the gland cells and also to an increase in the conductance of the basolateral membrane to one or more species of ions (Ginsborg & House, 1980). These processes produce fluxes of ions which are believed to stimulate secretion by activating ion pumps in the gland cell membranes. However, the ions involved in the responses of gland cells to agonists have not always been identified unambiguously (Marty, Tan & Trautmann, 1984) and the exact nature of the pumps or transport mechanisms responsible for secretion are not entirely clear (Martinez & Cassity, 1984).

It is apparent that the measurement of intracellular ions in gland cells could offer some insight into the process of secretion. In the present study, double-barrelled, ion-selective microelectrodes (ISMEs) have therefore been used to make direct measurements of the intracellular free concentrations of K⁺, Na⁺ and Cl^- ([K⁺]₁, [Na⁺]₁ and [Cl^-]₁, respectively) in the cytosol of salivary acinar cells of the pond snail Planorbis corneus. In addition, estimates of the permeability of the resting membrane to these ions were made. The ionic mechanisms underlying the response to acetylcholine (ACh) were also examined. Some of the results given below have been presented in a preliminary form to a meeting of the German Physiological Society (Barber & ten Bruggencate, 1985a).

Pairs of salivary glands were isolated from specimens of Planorbis corneus as described previously (Barber, 1985) and transferred to a recording chamber which had a volume of 0·2 ml. The chamber was perfused continuously at a rate of 7·5 ml min^-1 with snail salines at room temperature (20-28°C).

Standard physiological saline contained (in mmoll^-1): NaCl, 39; KC1, 1·3; CaCl₂, 4·5; MgCl₂, 1·5; NaHCO₃, 7·0; and was bubbled to a pH between 7·5 and 7·2 with a CO₂/O₂ gas mixture. K⁺-free saline was made by replacing KC1 with an equimolar amount of NaCl, while Ca²⁺-free saline was made by substituting MgCl₂ for CaCl₂ and adding 1 mmol 1^-1 EGTA. High-K⁺ saline, in which 25 mmol 1^-1 of the NaCl content of standard saline was replaced by an equimolar concentration of KC1, was also Ca²⁺-free. Cl^--free saline was made by using sulphate salts of Na⁺, K⁺, Ca²⁺ and Mg²⁺. It should be noted that this type of saline had a different ionic strength compared to standard snail saline. Similar results were, however, obtained with low-Cl^- saline, which contained 39 mmol 1^-1 sodium glucuronate instead of NaCl. Nominally bicarbonate-free saline was made by substituting Hepes buffer for sodium bicarbonate (pH adjusted with NaOH) and was gassed with 100% O₂. The production of metabolic CO₂ means that this solution is unlikely to have been completely free of bicarbonate ions. Acetylcholine bromide (Sigma) and ouabain (Serva) were dissolved in these salines and perfused through the chamber as required.

Intracellular recordings were made from salivary gland acinar cells with doublebarrelled ISMEs which had very fine tip diameters (<0·5μm as measured under a light microscope). These electrodes were constructed as described by Grafe, Rimpel, Reddy & ten Bruggencate (1982; see also Barber & ten Bruggencate, 1985b ; Barber, 1986). The liquid ion-exchanger resins used in the tips of the K⁺-, Na⁺-and Cl^--selective electrodes were Corning477317, ETH227 (Fluka) and IE-170 (W-P Instruments), respectively. The shank and barrel of the ion-selective side of the electrode were filled with 200 mmol l^-1 KC1, NaCl or KC1 for K⁺-, Na⁺-and Cl^--selective electrodes, respectively.

The reference barrels were filled with 1 moll^-1 magnesium acetate in the case of K⁺-and Na⁺-ISMEs and Imoll^-1 magnesium sulphate for Cl⁻-ISMEs and had resistances of 50—100 MΩ measured in physiological saline. Both barrels of the electrodes were connected via chlorided silver wires to the input stage of a high impedance amplifier.

The membrane potential (E_M) of the impaled cell was registered by the reference barrel of the ISME (V_ref). V_ref was also subtracted electronically from the potential registered by the ion-selective barrel in order to monitor a pure ion signal (V_ion). Current pulses could be applied between the bath and ground in order to check that this subtraction was being performed accurately (see Fig. 7). The experimental chamber was normally grounded through an agar bridge, though broken singlebarrelled microelectrodes filled with 3 moll^-1 KC1 (resistance 1-2MΩ) were sometimes used in an attempt to reduce bath ground artefacts during the exchange of different salines.

Electrodes were calibrated immediately after being withdrawn from the gland cells by running calibration solutions through the experimental chamber. K⁺-selective electrodes were calibrated in solutions containing 5 mmol l^-1 NaCl together with different concentrations of KC1. The V_ion for these electrodes (V_K) increased by an average of 53 mV per 10-fold increase in K⁺ concentration, while their mean K⁺:Na⁺ selectivity ratio was 1:0·013 (N=30). Calibration solutions for Na⁺-electrodes contained 50 mmol l^-1 KC1, 1 mmol l^-1 EGTA and different concentrations of NaCl. V_ion for Na⁺-ISMEs (V_Na) changed by a mean of 53 mV per 10-fold change in Na⁺ concentration, and the mean Na⁺:K⁺ selectivity of these electrodes was 1:0·06 (N = 9). Cl^--electrodes were calibrated in pure KC1 solutions (Thomas, 1978) and V_ion for Cl^--ISMEs (Vci) changed by a mean of 51 mV per 10-fold change in Cl^- concentration (N = 15). Values of [K⁺]₁, [Na⁺]₁ and [Cl^-]₁ are given in terms of free ion concentrations since conversions to ion activities were not performed [see Thomas (1978) and Tsien (1983) for a discussion of this point].

The K⁺ ion-exchanger used in the present study is known to be sensitive to ACh (Kuramoto & Haber, 1981), its response consisting of a positive shift in V_K. It was assumed, however, that the fall in intracellular V_K produced by a brief (10 s) extracellular application of ACh (e.g. Fig. 6A) was not influenced by any direct response of the ISME to ACh.

Intracellular measurements were also made with conventional single-barrelled microelectrodes which were filled with 3 mol 1^-1 potassium acetate and had resistances of 10—50 MΩ measured in physiological saline. The input resistances of impaled salivary gland acinar cells were measured by injecting pulses of direct current through a balanced bridge circuit.

Estimates of ion fluxes and permeability coefficients were made from changes in [K⁺]_i, [Na⁺]_i and [Cl^-]_i measured from glands bathed in either K⁺-free or Cl^--free saline. These calculations were based on the following assumptions.

1. Planorbis salivary gland cells are cuboidal in shape with sides of 15 μm (see Barber, 1985) and that only the basolateral membrane is exposed to changes in solutions.

2. Changes in [K⁺]_i, [Na⁺]_i and [Cl^-]_i reflect net movements of ions across the cell membrane, and changes occur uniformly throughout the entire cell volume.

3. No significant intracellular binding or buffering of K⁺, Na⁺ or Cl^- takes place (Thomas, 1978).

4. The operation of the electrogenic Na⁺/K⁺ pump is the only process responsible for the active transport of Na⁺ and K⁺ and any Cl^- pumps are inactive in Cl^--free saline (but see Aickin & Brading, 1985).

5. Changes in extracellular ion concentrations have no effect on the permeability of the cell membrane (see Discussion).

6. Measurements of [K⁺]_i, [Na⁺]_i and [Cl^-]_i in these experiments are uninfluenced by elicited action potential activity (but see Edman, Gestrelius & Grampp, 1983; Hodgkin & Horowicz, 1959) or any basal or stimulated secretion.

   The net flux of any given ion, M_ion, was calculated from the relationship

   

where M is measured in mol cm^-2 s^-1, Δ [ion]_i is the net change of the intracellular ion concentration in mol 1^-1, Δ t is time measured in seconds, V is cell volume in litres and A is the surface area of the cell in cm².

The theoretical value of E_M was then calculated from the normally employed form of the Goldman (1943) equation,

An example of a successful impalement of a Planorbis salivary acinar cell with an ISME, in this case a K⁺-selective electrode, is shown in Fig. 1. It was found that stable intracellular recordings with ISMEs were usually not possible with electrodes which had tip diameters greater than around 0·5 μm. Another general observation was that even fine electrodes required periods of 5·20 min after impalement to ‘seal into’ the gland cell membrane, as judged by the time necessary to measure stable values of E_M and V_ion. The relationship between the measured values of EM and [K⁺]_i, [Na⁺]_i and [Cl^-], are shown in Fig. 2.

Data concerning the resting levels of free intracellular ions in Planorbis salivary gland cells are brought together in Table 1. Where appropriate these values of [ion]₁ have been corrected for inaccuracies introduced by cross-sensitivity of the ion-exchangers to other intracellular ions ([Na⁺]₁ and [Cl^-]₁ measurements; see Discussion). The labelling of the [ion]_i ordinates of the figures, however, has not been adjusted for interference.

The high resting level of [K⁺]_i and the very low resting value of [Na⁺]_i in Planorbis salivary gland cells (Table 1) imply the presence of active mechanisms for accumulating K⁺ and extruding Na⁺ against their respective electrochemical gradients. It is generally accepted that the resting distribution of these ions across cell membranes is due chiefly to the operation of an electrogenic Na⁺/K⁺ pump (Petersen, 1980). The role of the Na⁺/K⁺ pump in maintaining [K⁺]₁ and [Na⁺]_i in Planorbis salivary glands was confirmed by reversibly inhibiting the activity of this pump with K⁺-free saline. After a delay of 4—10 min, the cells began to lose K⁺ at an initial and maximum rate of 2·6mmoll^-1 min^-1 (s.D. ±0·53 mmol 1^-1 min^-1, N = 4) and to gain Na⁺ at an initial and maximum rate of 0·15 mmol 1^-1 min^-1 (s.D. ±0·03 mmol 1^-1 min^-1, N — 5) (Fig. 3).

Time lags in changes in [K⁺]₁ and [Na⁺]₁ were not observed when the Na⁺/K⁺ pump was blocked irreversibly by ouabain (e.g. Fig. 6B), which probably indicates that some time is required in K⁺-free saline to deplete extracellular K⁺ at the surface of the gland cells. The rate of decrease of [K⁺]₁ and increase of [Na⁺]₁ were, however, similar in glands treated with K⁺-free saline or ouabain (the rate of increase of [Na⁺]₁ in Fig. 6B is approximately 0·13 mmol 1^-1 min^-1), which presumably indicates that the Na⁺/K⁺ pump is fully blocked in K⁺-free saline once K⁺ is washed from the experimental chamber.

Measurements of changes in [K⁺]₁ are unlikely to have been influenced to any great extent by the parallel increase in [Na⁺],. However, the relatively large fall in [K⁺]₁ and the comparatively poor Na⁺:K⁺ selectivity ratio of the Na⁺-ISMEs (see Materials and Methods) mean that the rise in [Na⁺]₁ is likely to have been underestimated by around 0·15 mmol 1^-1 min^-1 (i.e. [Na⁺]₁ actually rises by 0·3 mmol 1^-1 min^-1 in K⁺-free saline). These values were inserted into equation 1 and the net efflux of K⁺ (M_K) and net influx of Na⁺ (M_Na) were calculated (Table 1). The permeability coefficients for K⁺ (P_K) and Na⁺ (P_Na) were calculated (Table 1) using equation 2.

Readmitting normal K⁺-containing saline into the experimental chamber resulted in a transient membrane hyperpolarization, presumably caused by the reactivation of the electrogenic Na⁺/K⁺ pump (see Poulsen & Oakley, 1979) and the rapid restoration of the original levels of [K⁺], and [Na⁺], (Fig. 3).

Cl^- was lost from salivary gland cells in Cl^--free saline (Fig. 4) at a mean initial and maximum rate of 1·0 mmol 1^-1 min^-1 (s.D. ±0·47 mmol 1^-1 min^-1, N=7), without the delay characteristic of measurements of [K⁺]₁ and [Na⁺]₁ in K⁺-free saline (Fig. 3). This rate of Cl^- loss slowed gradually until [Cl^-]₁ stabilized after 10—20 min at a mean value of 15·0 mmol 1^-1 (s.D. ±4 mmol l^-1, N = 5). This value of rate of change of [Cl^-], was inserted into equation 1 to calculate the net Cl^- flux in Cl⁻-free saline, M_Cl, and the Cl^- permeability coefficient, P_Cl, was then estimated from equation 3 (Table 1).

Additional evidence in favour of a substantial resting permeability to Cl^- was obtained from measurements of changes in EM and input resistance (RM) in CP-free saline. The introduction of CP-free saline into the experimental chamber produced a rapid transient depolarization of Planorbis gland cells accompanied by a powerful discharge of action-potential-like activity (Figs 4, 5), followed by a return of E_M to its original value. Similar transient depolarizations upon contact with CP-free saline have been reported previously in other preparations known to have a high resting CP permeability, such as frog skeletal muscle cells (Hodgkin & Horowicz, 1959). CP-free saline also produced a large increase in the RM of Planorbis salivary glands (Fig. 5), as in other cells, such as mouse liver cells (Graf & Petersen, 1978), where resting CP permeability is high.

The transient depolarization, the increase in R_M and the loss of CP in CP-free saline were also observed after pre-incubating glands for 10 min in Ca²⁺-free saline. This indicates that these effects did not arise from an indirect depolarizing action of Cl^--free saline on presynaptic elements (Ascher, Kunze & Nield, 1976) or the low Ca²⁺ activity in Cl^--free saline made with sulphate salts (Hodgkin & Horowicz, 1959). These changes in E_M, R_M and [Cl^-]_i were also observed in low-Cl^- solution in which glucuronate was substituted for part of the Cl^-. This means that these effects were not artefacts caused by replacing Cl^- with sulphate.

Transferring the glands back to normal saline led to the rapid recovery of the original level of [Cl^-]_i (Fig. 4), a transient depolarization and a transient increase in R_M (Fig. 5). The subject of the resting Cl^- permeability of the gland cell membrane will be considered further in the Discussion.

The application of ACh or nicotinic agonists onto Planorbis salivary glands is known to produce a dose-dependent biphasic electrical response from acinar cells (Barber, 1985). In the present study it was found that the ACh-induced depolarization occurs in parallel with a fall in [K⁺]_i (Fig. 6A) and a rise in [Na⁺]_i (Fig. 6B; Table 1). These data thus demonstrate that at least part of the K⁺ released into the bathing medium by ACh (Barber & ten Bruggencate, 1985b) originates from acinar gland cells rather than, or in addition to, presynaptic structures, and also confirm the occurrence of a Na⁺ influx during the ACh-induced depolarization (Barber, 1985).

In contrast to the results obtained for [K⁺]₁ and [Na⁺]_1, no significant change in [Cl^-]_i was detected following the application of ACh (Fig. 7). Any increase in Cl^- permeability would, however, be difficult to detect because the net driving force on Cl^- is comparatively weak (Table 1), and indeed reverses during the ACh-induced depolarization.

The possibility that ACh elicits an undetected increase in membrane permeability to Cl^- was investigated by applying ACh repeatedly onto glands bathed in Cl^--free saline (Fig. 8). Cl^--free saline produced an increase in the size of the depolarizing phase of the response to ACh and this enhanced response was maintained for as long as the preparation was bathed in Cl^--free medium (i.e. up to 10 applications over the course of 2h). This last finding is not a result which would be expected if ACh made the cell membrane more permeable to Cl (see Adams & Brown, 1975; Iwatsuki & Petersen, 1977). The increase in the size of the ACh depolarization is presumably a consequence of the increase in in Cl^--free saline (see previous section) rather than the result of an increase in the outward driving force on Cl^- following the removal of extracellular Cl^-.

The mechanism of action of ACh on Planorbis gland cells was further examined by comparing the effects of Ach on [K⁺]₁ and [Na⁺]_i with those of high-K⁺ saline (Fig. 9). High-K⁺ saline was applied only very briefly (30 s) to salivary glands in order to avoid the influx of KC1 and water which can occur over time in these salines (see Ascher et al. 1976; Iwasa, 1982). These experiments were also carried out in the absence of extracellular Ca²⁺ so that the depolarizing action of high-K⁺ saline on gland cells would not be influenced by any presynaptic excitatory effect of this saline. It was found, though, that good impalements with ISMEs were more difficult to obtain in Ca²⁺-free saline than in normal saline, with the result that recordings of [K⁺]_i and [Na⁺]_i were not optimal in these experiments.

Depolarizing salivary cells with high-K⁺ saline led to an influx of K⁺ (Fig. 9A), which is in accordance with the reversed driving force on K⁺ under these conditions.

Although the extent to which K⁺ may have leaked into salivary cells through a poor seal between the ISME and cell membrane in these experiments remains unknown, this result implies that a good part of the K⁺ lost from gland cells during ACh-induced depolarization leaves through voltage-dependent K⁺ channels and/or leakage pathways.

The brief application of high-K⁺ saline produced no change in [Na⁺]₁ (Fig. 9B), though it should be noted that the inward driving force on Na⁺ is reduced in this saline. Despite this difference it seems likely that the majority of Na⁺ which enters Planorbis gland cells after the application of ACh does so through channels activated directly by the ACh receptor.

The hyperpolarizing phase of the ACh response is associated with the gradual recovery of [K⁺]₁ and [Na⁺]_i to their original levels (Fig. 6). The part played by the Na⁺/K⁺ pump in this recovery was confirmed in the present study by inhibiting the activity of this pump irreversibly with 10⁻⁴moll^-1 ouabain (Fig. 6B). Gland cells in which the Na⁺/K⁺ pump was blocked were unable to restore [Na⁺]₁ (Fig. 6B) or [K⁺]₁ to their original resting levels. These experiments therefore indicate that the transient undershoot of [K⁺]_e and overshoot of [Na⁺]_e which occur during the hyperpolarization (Barber & ten Bruggencate, 1985b) are produced at least in part by the electrogenic accumulation of K⁺ and expulsion of Na⁺ from the gland cells. The Na⁺/K⁺ pump is presumably activated by the increase in [Na⁺]₁ which occurs during the depolarization, rather than by a direct action of ACh on the pump.

ISMEs have been used to measure [ion]_i in a broad range of tissues (Walker & Brown, 1977) and probably represent the most useful technique currently available for determining the activity of K⁺, Na⁺ and Cl^- directly and continuously in living cells. One of the first studies to make use of ISMEs to measure intracellular ions in gland cells was made by Palmer & Civan (1977), who measured [K⁺]₁, [Na⁺]₁ and [Cl ]_i in the cytoplasm and nucleus of giant salivary cells of the insect Chironomus. The smaller salivary gland cells of other insects (e.g. Berridge & Schlue, 1978) and mammals (e.g. Poulson & Oakley, 1979) have since proved amenable to study with ISMEs. However, the results of the present investigation, together with those of Barber (1986), represent the first direct determinations of [ion]₁ in molluscan gland cells.

In the present study, no striking dependence was found between the presumed degree of injury to the cells, as measured by E_M, and the magnitude of [Na⁺]₁ or [Cl^-]₁ (Fig. 2A,B). This implies that estimates of [Na⁺]_i and [Cl^-]_i are unlikely to have been influenced greatly by impalement artefacts. In the case of [K⁺]_i it was found that cells with higher E_M values tended to have a higher [K⁺]_i and vice versa (Fig. 2A). This probably reflects genuine variation in [K⁺]₁ and the dependence of resting E_M on the transmembrane K⁺ concentration gradient rather than a correlation between the degree of cell damage and the size of [K⁺]_i.

Another potential difficulty in the interpretation of these measurements is that the ion-exchangers used in the tips of ISMEs are not perfectly sensitive to only one kind of ion. Na⁺ is the most likely interfering intracellular cation for the K⁺ ionexchanger used in the present study (Meier et al. 1982). However, the very low levels of [Na⁺]_i in Planorbis salivary gland cells, together with the relatively good selectivity of the K⁺-ISMEs against Na⁺ (see Materials and Methods), mean that no allowance for Na⁺ need be applied to measurements of [K⁺]_i.

With Na⁺-ISMEs the largest errors are likely to be caused by K⁺ and Ca²⁺ (Meier et al. 1982). The relatively high levels of [K⁺]_i imply that intracellular K⁺ may make a significant contribution to the V_Na signal. Calculations based upon the selectivity ratios of the Na⁺-ISMEs (see Materials and Methods) and the mean value of [K⁺]₁ (Table 1) indicate that [Na⁺]₁ maybe overestimated by around 2·6 mmol 1^-1. Levels of intracellular free Ca²⁺ ([Ca²⁺]_i) in Planorbis salivary gland cells are very low (Table 1) which suggests that Ca²⁺ probably does not interfere significantly with measurements of [Na⁺]_i.

In contrast to recordings with K⁺-and Na⁺-ISMEs, the identity and concentrations of ions interfering with measurements with Cl^--ISMEs are usually not known (Thomas, 1978). It was found, however, that apparent [Cl^-]_i in gland cells soaked in Cl^--free saline falls to stabilize around 15mmol1^-1. On the assumption that all intracellular Cl^- is eventually lost from cells bathed in Cl^--free saline (Thomas, 1978), this would mean that true [Cl^-]_i lies around 10·3 mmol 1^-1 (Table 1) while the total interference in these cells is equivalent to 15 mmol 1^-1 [C1-]_i.

Bicarbonate ions certainly contribute to this interference since exchanging standard snail Ringer in the experimental chamber with nominally bicarbonate-free saline produced a rapid (complete within 2-3 min) and fully reversible reduction in apparent [Cl^-]_i by around 4·7 mmol 1^-1 (s.D. ±1·5 mmol 1^-1; N=3; A. Barber, unpublished observations). Interference from intracellular bicarbonate is generally assumed to be insignificant in bicarbonate-free saline (Bolton & Vaughan-Jones, 1977).

The sulphate ions used as a substitute for Cl^- would also be a source of interference if they entered the cells (Saunders & Brown, 1977), as would sulphate which diffused from the reference barrel of the ISME. Estimating the contribution made by cross-sensitivity to these measurements is further complicated by the fact that the selectivity of the C1^--ISME deteriorates when [Cl^-]₁ is low (Vaughan-Jones, 1979; Aickin & Brading, 1983). Finally, it can also not be ruled out that salivary cells in Cl^--free saline lose water and shrink (MacKnight, 1985), thus influencing measurements of [Cl^-]₁.

Tip potential artefacts recorded by the reference barrel of the ISMEs were another possible reason for inaccurate measurements of [ion]_i. Although no systematic investigation of such artefacts was carried out, it was observed that E_M values measured with the acetate-filled reference barrels of K⁺-₁ Na⁺-and Ca²⁺-ISMEs were generally higher than those measured with the sulphate-filled barrels of the Cl^--ISMEs (Table 1). Similar problems with reference barrels filled with sulphate salts have been reported by a number of other workers (e.g. Bolton & Vaughan-Jones, 1977; Berridge & Schlue, 1978; Gardner & Moreton, 1985). Any underestimation of E_M would lead to an overestimation of V_Cl (see Materials and Methods) which would contribute to the interference encountered with measurements of [Cl^-]₁.

The mean basal [K⁺]_i in Planorbis salivary gland cells was determined as 42·9 mmol 1^-1 (Table 1), which is comparable to the [K⁺]_i values of 53·4 mmol 1^-1 (Kostyuk, Sorokina & Kholodova, 1969) and 51·1mmoll^-1 (s.D. ±10·3 mmoll^-1, N= 15; A. Barber, unpublished observations) measured in unidentified giant neurones of Planorbis central ganglia. This value is also comparable to the 64 mmol 1^-1 estimated indirectly to be the [K⁺]₁ in salivary cells of the related snail Helisoma (Hadley, Murphy & Kater, 1980), when allowance is made for the osmolalities of the different Ringer salines (in Helisoma 64/130 = 0·49; in Planorbis 42·9/112 = 0·38).

The [Na⁺]₁ of Planorbis salivary gland cells (2·4 mmol 1^-1; Table 1), on the other hand, is considerably lower than the 13·8 mmol 1^-1 measured from Planorbis neurones with Na⁺-ISMEs of the protruding tip type (Kostyuk et al. 1969). This relatively high value of neuronal [Na⁺]_i may be due to the difficulties of properly inserting ISMEs of this kind intracellularly (Thomas, 1972) or could represent a genuine difference between [Na⁺]_i in Planorbis neurones and salivary gland cells.

The [Cl^—]_i and equilibrium potential for Cl^- found in the present study (Table 1) are within the range previously measured in a variety of molluscan preparations (Gardner & Moreton, 1985).

Apart from K⁺, Na⁺ and Cl^-, the net cellular charge balance and osmolality are probably made up mainly by amino acids, phosphorus compounds and the fixed negative charges of proteins (Burton, 1983).

The high basal value of [K⁺]_i and very low level of [Na⁺]_i in Planorbis salivary gland cells were found to be maintained by the activity of an electrogenic Na⁺/K⁺ pump. Inhibiting the activity of this pump reversibly with K⁺-free saline blocked the active accumulation of K⁺ and extrusion of Na⁺ and revealed the presence of a passive K⁺ efflux and Na⁺ influx (Fig. 3). The P_K calculated from the K⁺ efflux (7·6× 10⁻⁶cms^-1) is consistent with a high resting permeability to K⁺, as for example in leech neurones (8×10⁻⁶cms^-1; Deitmer & Schlue, 1981). The P_Na estimated from the Na⁺ influx (5·4× 10⁻⁸cms^-1) is very low and a P_Na/P_K ratio of 0·007, if correct, would indicate that the resting basolateral membrane is very selective against Na⁺ as compared to K⁺.

Planorbis salivary gland cells also have a relatively high resting [Cl^-]₁ (10·3 mmol 1^-1; Table 1) which implies the presence of some process, at present unknown, for actively accumulating Cl^- against its electrochemical gradient. At the same time the resting Cl^- permeability of the salivary cell membrane would appear to be quite high. Thus when glands were bathed in Cl^--free saline, the result was a rapid fall in [Cl^-]_i (Fig. 4), from which a P_Cl of 8·9× 10⁻⁷cms^-1 was calculated (Table 1). It should be noted, however, that cells bathed in Cl^--free saline do not always lose Cl^- only as a result of passive leakage through the cell membrane. It has been demonstrated recently that the Cl^-/bicarbonate exchange mechanism which normally keeps [Cl^-]_i high in mammalian Purkinje fibres (Vaughan-Jones, 1979, 1982) and smooth muscle cells (Aickin & Brading, 1985) reverses in Cl^--free saline to transport Cl^- out of these cells. If such an active extrusion of Cl^- in Cl^--free medium also takes place in Planorbis salivary cells, then this would clearly lead to a spuriously high estimation of P_Cl.

The increase in R_M and the transient decrease in E_M in salivary gland cells observed after transferring the glands to Cl^--free saline (Fig. 5) also appear consistent with a high resting P_C1 in these cells (Hodgkin & Horowicz, 1959; Graf & Petersen, 1978). But an increase in leak resistance due to the lower solution conductivity of Cl^--free saline (Adams & Brown, 1975), a decrease in electrical coupling between neighbouring acinar cells (Barber, 1985) caused by the removal of extracellular Cl^- (Asada & Bennett, 1971), or a decrease in membrane K⁺ conductance in Cl^--free medium (Carmeliet & Verdonck, 1977) may also contribute to the increased R_M in Cl^--free saline.

The transient depolarization and increase in R_M observed after transferring the glands back to normal saline (Fig. 5) were unexpected observations. If Pa were high then the return to normal saline should have elicited a transient E_Mhyperpolarization (Hodgkin & Horowicz, 1959; Iwatsuki & Petersen, 1977) and a gradual decrease in RM-These effects of returning Cl^--depleted cells to normal saline are not readily explained, and may be secondary to changes in intracellular pH (Thomas, 1982) or cell shrinkage (MacKnight, 1985).

Given the reservations mentioned above, it was reassuring that a theoretical E_M of — 71·2 mV was calculated by inserting data from Table 1 into the Goldman equation (equation 4). This estimate of is very close to the values measured with doublebarrelled ISMEs (Table 1) or single-barrelled conventional electrodes (— 72mV; Barber, 1985). It would seem, then, that E^ in Planorbis salivary gland cells can be explained in terms of the measured concentration gradients of K⁺, Na⁺ and Cl^- and their estimated resting permeabilities through the gland cell membrane. This conclusion does not exclude the possibility that electrogenic ion pumps also make a small contribution to the resting E_M. Interestingly, the ionic basis of E_M in salivary gland cells of the related snail Helisoma appears to be different from that in Planorbis. Helisoma salivary cell membranes are less selective against Na⁺ (P_Na/P_K ratio, 0·04) and Cl^- permeability makes no contribution to their E_M (Hadley et al. 1980).

It has been demonstrated in the present study that the stimulation of Planorbis salivary gland cells with ACh leads to a fall in [K⁺]_i and a rise in [Na⁺]_i (Fig. 6). In contrast, stimulation with ACh produced no sign of any change in [Cl^-]_i (Fig. 7), though some passive changes in [Cl^-]_i following E_M might have been expected, given the relatively high resting P_cl. While agonist-induced changes in [ion], have been recorded previously from gland cells in insects and mammals (e.g. Berridge & Schlue, 1978; Poulsen & Oakley, 1979), this study, together with a preceding investigation (Barber, 1986), represents the first measurements of such changes from molluscan gland cells.

These observations are probably relevant for salivary secretion in intact specimens of Planorbis. ACh is at present the most likely candidate for the transmitter between the central nervous system and the salivary glands (Barber, 1985), as it is in several other species of mollusc (Barber, 1983). As such it is suspected to mediate the large (approx. 30 mV) excitatory postsynaptic potentials (EPSPs) which trigger action potentials in these salivary cells (Barber, 1985). Although ACh-induced depolarizations are larger than individual EPSPs a volley of EPSPs (Barber, 1985) can produce a depolarization not dissimilar to the ACh response in size and duration. Finally, while it has not yet been shown whether Planorbis salivary gland cells actually secrete saliva following the application of ACh, it is known that gland cells which are electrically excitable do secrete when depolarized either by agonists (e.g. Wada et al. 1984) or by high-K⁺ saline (e.g. Suchard, Lattanzio, Rubin & Pressman, 1982).

As far as the mechanism of action of ACh is concerned, the absence of any change in [Na⁺]i when Planorbis salivary glands were bathed in high-K⁺ saline (Fig. 9B) suggests that detectable amounts of Na⁺ do not enter gland cells through voltagedependent Na⁺ channels during the ACh-induced depolarization. Like Ca²⁺ (Barber, 1986), the majority of Na⁺ probably enters these cells through channels opened directly by ACh. An ACh-induced depolarization produced by an exclusive increase in membrane permeability to Na⁺ and Ca²⁺ would explain the finding that even large depolarizations of E_M are accompanied by only relatively small decreases in membrane resistance (Barber, 1985). The large inward driving force on Na⁺ and Ca²⁺ would mean that even a small increase in permeability to these ions would produce a large depolarization (see Bührle & Sonnhof, 1983).

Depolarizing salivary cells with high-K⁺ saline produced substantial changes in the level of [K⁺]_i (Fig. 9A). This means that large numbers of potassium ions are probably able to leave salivary cells during the ACh-induced depolarization either through leakage pathways, as a result of the increased outward driving force on K⁺, or through voltage-dependent K⁺ channels. Large amounts of K⁺ probably do not leave Planorbis salivary gland cells via either Ca²⁺-activated K⁺ channels or Ca²⁺-activated non-selective cation channels (Marty et al. 1984; Petersen & Maruyama, 1984) because the ACh-induced K⁺ efflux is also observed in Ca²⁺-free saline (Fig. 9A; Barber & ten Bruggencate, 19856). The possibility that ACh-activated channels in Planorbis salivary cells are also permeable to a small degree to K⁺ has not been ruled out by the present experiments, however. Whatever its mechanism, it appears likely that this efflux of K⁺ plays an important role in repolarizing the gland cells at the end of the depolarizing phase of the ACh-response.

The electrogenic Na⁺/K⁺ pump seems to have no part in this repolarization of the gland cell membrane, at least during the short exposures to ACh used in the present experiments. Salivary cells in which the Na⁺/K⁺ pump was inhibited by ouabain repolarized quite normally (Fig. 6B), though the repeated application of ACh under these conditions does lead to a gradual depolarization of E_M (Barber, 1985). The function of the Na⁺/K⁺ pump appears to be to maintain the basal values of [K⁺]_i and [Na⁺]₁ both in resting glands and following stimulation.

The suggested mechanism of action of agonist on Planorbis salivary gland cells, namely an increase in membrane permeability to Na⁺ and Ca²⁺, the subsequent release of K⁺ and the activation of an electrogenic Na⁺/K⁺ pump, is different from those proposed in a number of other exocrine glands (for reviews see Ginsborg & House, 1980; Petersen, 1980). In mammalian salivary glands, for example, stimulation with ACh leads to an efflux of K⁺ through Ca²⁺-activated channels, followed by a re-uptake of K⁺, along with Na⁺ and Cl^-, by a K⁺/Na⁺/Cl^- cotransport system (Petersen & Maruyama, 1984; see also Martinez & Cassity, 1984). This net uptake of NaCl appears to be sufficient to account for the secretion of NaCl in the primary saliva (see also Marty et al. 1984).

The ion fluxes produced by ACh in Planorbis salivary gland cells, though, do not seem to represent a balanced uptake of ions suitable for incorporation into a profuse, watery salivary secretion. It may be that since Planorbis is an aquatic animal the secretion of enzymes and mucus is a much more important function of these glands than the secretion of salt and water. The salivary glands may not be required to secrete very much fluid.

Previous accounts of the electrophysiology of molluscan gland cells have tended to emphasize the role of action potentials in secretion. Salivary cells in Helisoma (Hadley et al. 1980; Senseman, Horwitz, Cleeman & Orkand, 1985) and Ariolimax (Goldring, Kater & Kater, 1983) and cells of the pedal gland of Ariolimax (Kater, 1977) all display action potentials whose inward currents are carried by Ca²⁺ and Na⁺. The fact that action potentials and ACh produce similar changes in ion permeability suggests, however, that EPSPs and action potentials may complement one another in producing the ion fluxes which stimulate secretion.

Financial support for this work was provided by the University of Munich. I express my thanks to Professor G. ten Bruggencate for having made my stay in Munich possible.