Skin from larval bullfrogs was mounted in an Ussing-type chamber in which the apical surface was bathed with a Ringer solution containing 115 mM K+ and the basolateral surface was bathed with a Ringer solution containing 115 mM Na+. Ion transport was measured as the short-circuit current (I sc) with a low-noise voltage clamp, and skin resistance (R m) was measured by applying a direct current voltage pulse. Membrane impedance was calculated by applying a voltage signal consisting of 53 sine waves to the command stage of the voltage clamp. From the ratio of the Fourier-transformed voltage and current signals, it was possible to calculate the resistance and capacitance of the apical and basolateral membranes of the epithelium (R a andR b,C a andC b, respectively). With as the anion,R m decreased rapidly within 5 min following the addition of 150 U/ml nystatin to the apical solution, whereasI sc increased from 0.66 to 52.03 μA/cm2 over a 60-min period. These results indicate that nystatin becomes rapidly incorporated into the apical membrane and that the increase in basolateral K+ permeability requires a more prolonged time course. Intermediate levels ofI sc were obtained by adding 50, 100, and 150 U/ml nystatin to the apical solution. This produced a progressive decrease inR a andR b whileC a andC b remained constant. With Cl− as the anion, I sc values increased from 2.03 to 89.57 μA/cm2 following treatment with 150 U/ml nystatin, whereas with gluconate as the anionI sc was only increased from 0.63 to 11.64 μA/cm2. This suggests that the increase in basolateral K+permeability produced by nystatin treatment, in the presence of more permeable anions, is due to swelling of the epithelial cells of the tissue rather than the gradient for apical K+ entry. Finally,C b was not different among skins exposed to Cl−, , or gluconate, despite the large differences inI sc, nor did inhibition of I scby treatment with hyperosmotic dextrose cause significant changes inC b. These results support the hypothesis that increases in cell volume activate K+ channels that are already present in the basolateral membrane of epithelial cells.
- cell volume regulation
- potassium channels
- epithelial transport
potassium channels in the basolateral membrane of epithelial cells allow K+ that is transported into the cell by the Na+-K+pump to exit and establish a negative potential within the cell (14). Basolateral K+ channels are also involved in the regulation of cellular volume. For example, the basolateral membrane of turtle colon has three types of K+ channels, one of which is opened under conditions that promote cell swelling (5, 8, 9). The volume-sensitive K+ channel was also found to be blocked by lidocaine. Dawson et al. (6) subsequently showed that substitution of a less permeant ion, gluconate, for Cl− inhibited the lidocaine-sensitive component of the basolateral K+ conductance and used fluctuation analysis to demonstrate that lidocaine inhibition under swelling conditions (i.e., Cl− as the anion) obeyed pseudo-first-order kinetics, as would be expected from a direct channel blockage.
The above studies of basolateral K+ permeability were conducted with epithelial tissues that were mounted in an Ussing-type chamber with K+ Ringer bathing the mucosal surface of the tissue and Na+Ringer bathing the serosal surface. The mucosal solution also contained the ionophore amphotericin B, which formed K+-conducting channels in the apical membrane so that the basolateral membrane could be effectively voltage clamped and the short-circuit current (I sc) could be related to the rate of basolateral K+ transport. We have used a similar approach to study basolateral K+ channels in the larval frog skin (12, 25), using the ionophore nystatin to increase apical K+ conductance. We have found that a period of ∼1 h following nystatin treatment is required forI sc to reach a maximum value. This suggests that the activation of basolateral K+ channels is a relatively slow process that might result from cellular swelling as K+ and permeant anions enter the cells across the apical membrane. Alternatively, the prolonged time course for the stimulation ofI sc could result from a gradual insertion of nystatin into the apical membrane. The present study was initiated to determine how quickly nystatin forms K+-conducting channels in the apical membrane relative to the time course for the development of maximal K+ transport across the basolateral membrane. Impedance analysis (24) was used to measure the resistance and capacitance of the apical and basolateral membranes (R a andR b,C a andC b, respectively) as I sc increased following nystatin treatment. An increase inC b would suggest that the increase in basolateral K+ transport was due to the addition of membrane vesicles, whereas lack of change inC b would suggest that channels had become activated in situ. The larval bullfrog skin is particularly useful for this study because the limiting barrier to ion transport is a single layer of apical cells (20), the apical membrane resistance can be easily reduced by nystatin (4), and the basolateral membrane contains K+ channels that are activated by conditions that result in cell swelling (12).
In our studies to date, has been used as the anion, since the basolateral membrane is highly permeable to Cl− (11) and the low resistance of the tissue complicates fluctuation analysis measurements. We have found that quinine, quinidine, lidocaine (26), and more recently verapamil (12) inhibit that fraction of theI sc that can also be inhibited by treatment with hyperosmotic sucrose solutions. A second goal of the study was to determine whether substitution of anions that are progressively less permeant (i.e., Cl−, , and gluconate) or the addition of hypersomotic solute concentrations results in a lower stimulation of basolateral K+ conductance and to measure R b andC b under these conditions. If the stimulation of basolateral K+ transport is the result of cell swelling due to the entry of more permeant anions, there should be a progressive decrease inI sc and increase in R b with the less permeant anions or the addition of an impermeant solute. If these changes in I scand R b are the result of removal of membrane vesicles,C b would be predicted to decrease in proportion to the decrease inI sc.
Larval bullfrogs, Rana catesbeiana, were obtained from commercial suppliers and maintained in aquaria at 15°C. The ventral skin was dissected with the abdominal musculature attached (4), using MS 222 anesthesia. The skin was mounted in an Ussing-type chamber that minimized edge damage and permitted continuous perfusion of the mucosal and serosal solutions (7). The mucosal solution contained 115 mM K+ and 1.0 mM Ca2+, with either Cl−, , or gluconate as the anion, and 2.5 mM KHCO3. The serosal solution contained 115 mM Na+ and 1.0 mM Ca2+, with either Cl−, , or gluconate as the anion, and 2.5 mM KHCO3. The pH of both solutions was 8.0. Once stable values forI sc had been obtained, nystatin (Sigma Chemical) was added to the mucosal Ringer to give the desired concentration. The appropriate amount of nystatin was dissolved in 100 μl dimethyl sulfoxide per 100 ml Ringer.
The macroscopicI sc was applied with a low-noise voltage clamp and was recorded on a chart recorder. The direct current (DC) resistance of the entire tissue (R m) was measured by applying a 10-mV pulse to the command stage of the voltage clamp and dividing by the change inI sc. For impedance measurements, the command stage of the voltage clamp was connected to a computerized voltage signal composed of 53 sine waves, and the voltage and current signals from the clamp were sampled consecutively. The voltage and current signals were subjected to a fast Fourier transform, and the impedance (Z) was calculated as the ratio of the transformed voltage and current signals (see Refs. 16 and 22 for detailed description of the synchronization of the voltage and current signals). Two impedance curves were merged and plotted as impedance loci on Nyquist plots. Before nystatin treatment, the Nyquist plots contained a single semicircle. This semicircle could be fitted withEq.1, which assumes a single membrane with parallel resistance (R) and capacitance (C) elements (Ref. 24, adapted from Ref. 3) Equation 1In this equation, j = and ω is the angular frequency = 2πf, wheref is frequency. The exponent α = 1 − 2φ/π, where φ is the angle, in degrees, between the line that connects the center of the semicircle with its intercept on the real (resistance) axis. This equation provides an accurate measurement of tissue capacitance (C m) that is equivalent to C abefore nystatin treatment, since the apical membrane is the dominant resistance element (24). It should be noted, however, that the resistance elements calculated by Eq.1 include that ofR a andR b in series and the paracellular pathway (R p) that is in parallel with R aand R b. Thus the resistance calculated by Eq.1, before the addition of nystatin, should be equal toR m and, since theR a is very high, should provide an approximate measure ofR p.
After nystatin treatment, two semicircles frequently appeared in the Nyquist plots. To fit the two semicircles,Eq. 1can be used separately for each (24), and the resistance values calculated by this equation approximateR a andR b. Alternatively, Margineanu and Van Driessche (17) developed an expression for transepithelial impedance [Z(t)] that fits two semicircles, with the assumption that they represent two membranes in series, each with a parallel resistance and capacitance (e.g., R a,R b,C a,C b). This model also incorporates an expression for the paracellular shunt resistance (R p) that is in parallel with the apical and basolateral membranes Equation 2InEq.2, sis the product jω, and A, B, C, and D contain terms that relate toR p,R a,R b,C a, andC b A, B, C, and D were obtained by nonlinear curve fitting ofEq. 2to the impedance data as described by Margineanu and Van Driessche (17). Values forR p were taken to be the resistance values (R m) calculated from the DC voltage pulses before nystatin treatment. As indicated above, the apical membrane of the larval skin has a very high DC resistance (4), and it was assumed that the resistance of the skin with an intact apical membrane approximatesR p.
To follow the time course of impedance changes asI sc was stimulated, Nyquist plots were generated at 5-min intervals for 60 min following the addition of nystatin at a concentration of 150 U/ml, which we have found to produce maximalI sc values (25). In a second series of experiments,I sc was more gradually stimulated by sequentially adding 50, 100, and 150 U/ml nystatin. Impedance was determined whenI sc had reached stable values at each concentration. The above experiments were conducted with as the anion.
The effect of different anions was assessed by conducting separate sets of experiments using Cl−, , or gluconate as the anion. Skins were treated with 150 U/ml nystatin, and impedance was determined after stable I sc values had been achieved.
Finally, the effect of increasing the osmotic concentration was evaluated by adding dextrose, in concentrations between 10 and 50 mM, to the mucosal and serosal solutions of preparations having as the anion. Dextrose was added after I sc had equilibrated following mucosal treatment with 150 U/ml nystatin.
Statistical comparisons between current, resistance, and capacitance values were made with Student’st-test.
The time course for the stimulation ofI sc by the addition of nystatin to the mucosal Ringer is shown in Fig.1. Note the immediate rise inI sc followed by a slower but larger increase inI sc over a 1-h period. Note also that the nystatin effect could be reversed after removal from the mucosal Ringer. Nyquist plots obtained before nystatin treatment and after stimulation ofI sc with 150 U/ml nystatin are shown in Fig. 2. The magnitudes of the resistive and capacitive components of the plot obtained after nystatin treatment were multiplied by a scale factor of five to illustrate the presence of a second semicircle, and the points were fitted by Eq.2. The plot obtained before nystatin treatment was fitted with Eq.1.
R m decreased to minimal values within 5 min following nystatin treatment. It was found that resistance values calculated from the fit of the pretreatment plots by Eq.1 were almost identical withR m (Fig.3 A). After nystatin treatment, the fit withEq. 2showed that the smaller semicircle nearer the origin of the Nyquist plot (the higher-frequency impedance loci) represented the apical membrane, sinceC a values were the same as those seen before nystatin treatment (see below) andR a values declined in parallel withR m as would be expected when nystatin formed channels in the apical membrane. The larger semicircle represented the basolateral membrane, which constitutes most of the resistance (R b) following 150 U/ml nystatin treatment.
Apical capacitance was calculated fromEq. 1before nystatin treatment and by Eq.2 following nystatin treatment (Fig.3 B). The two fits gave values for apical capacitance that were not significantly different. The presence of the second semicircle following nystatin treatment allowed the calculation of C bby Eq.2. It was also possible to useEq. 1to fit the larger semicircle during the time course of nystatin treatment and obtain a measure ofC m when the basolateral membrane was the primary resistance element of the tissue. To demonstrate that the single and double fits gave similar values for basolateral membrane capacitance, a series of 22 preparations was evaluated afterI sc had equilibrated with 150 U/ml nystatin in the mucosal Ringer. The values for C b were 12.73 ± 0.77 and 13.41 ± 0.74 μF/cm2 withEqs.1 and2, respectively, and were not significantly different.
Both C m andC b increased during the first 15 min following nystatin treatment as the basolateral membrane became the major component of the Nyquist plot. Also, I sc was rapidly increasing during this period as the cells responded to the rapid increase in K+ entry across the apical membrane. Given the sampling time required to obtain values at the lower frequencies (∼150 s), it was desirable to measure resistance and capacitance at stable values ofI sc that were intermediate between pretreatment and maximal values. This was accomplished by exposing the apical membrane to intermediate concentrations of nystatin. It can be seen in Fig.4 that sequential treatment with 50, 100, and 150 U/ml nystatin in the mucosal Ringer produced progressively higher I sc values that were stable throughout the sampling period required for impedance measurements. Mean values for the stableI sc andR m values for five such preparations are given in Table1. Impedance measurements at these intermediate I scvalues showed the appearance of a second semicircle that became the dominant component at the highest nystatin concentration (Fig.5). Results from 13 preparations showed thatR a decreased to a small fraction of the tissue resistance; however,C a remained relatively constant (Table 2). The decrease inR a with increasing nystatin concentrations corresponds to the increase inI sc and a decrease in R b. The values forC b, however, were not found to be significantly different.
When the Ringer solutions contained Cl− as the anion, theI sc values obtained after treatment with 150 U/ml nystatin were greater andR b was lower than those observed with (Fig.6). In contrast, lower values forI sc and higher values for R bwere noted when gluconate was substituted for . Values forC b obtained with the three anions were not significantly different, despite the large differences inI sc andR b. We also observed, as seen with in Fig.3 B, that the elevation ofC b recorded during the first 10–20 min following the addition of 150 U/ml nystatin was seen regardless of the anion and the resulting degree to which I sc was stimulated (not shown).
The addition of dextrose produced a dose-dependent inhibition ofI sc in conjunction with an increase inR m(Fig. 7). The values forC m (fitted withEq.1) that were obtained when stableI sc values were obtained at each dextrose concentration were not significantly different from that obtained before dextrose treatment.
The decrease in tissue resistance (R m or byEq. 1or Eq.2) within 5 min after the addition of 150 U/ml nystatin to the mucosal Ringer shows that nystatin inserts rapidly into the apical membrane. It is also evident from Fig. 1 that nystatin washes out of the apical membrane when removed from the mucosal Ringer. The development of maximalI sc values during the 1 h following treatment with the highest nystatin concentration is thus the result of a more prolonged time course for the stimulation of basolateral K+ permeability. After the initial decline inR a,R m values were essentially identical withR b but did not decrease further, despite the continued increase inI sc. This could be a result of other ions becoming permeable across the basolateral membrane as R abecame maximally reduced or possibly is a result of the increasingI sc interfering with the impedance measurements.
In contrast, when the nystatin concentration was increased incrementally, R bdid decrease asI sc values increased when R awas lowered by the ionophore, suggesting that K+ permeability was a major component of basolateral resistance under these conditions. In either case, an increase in the entry of K+ into the cells was compensated for by an increase in K+permeability across the basolateral membrane. It is unlikely, however, that the K+ gradient per se is the determining factor for the reduction ofR b, since the increase in I sc(Cl− > > gluconate) and decrease inR b(Cl− < < gluconate) that were observed with anion substitution occurred with the highest nystatin concentration and with comparable gradients for K+ transport across the tissue. Gluconate substitution for Cl− is known to reduce cellular swelling in turtle colon cells after treatment with amphotericin B (6). The increase in basolateral K+ permeability of the larval frog skin bathed either with more permeant anions or with increasing nystatin concentrations is probably regulated as part of a volume regulatory decrease, as is the case for most cells faced with conditions that promote swelling (15, 25). The reduction ofI sc and the increase in R mthat accompanied the addition of dextrose to the bathing solutions further support the hypothesis that the cells are responding to the osmotic rather than an ionic effect of anion substitution.
There was an apparent increase inC b andC m during the 15 min following the addition of 150 U/ml nystatin. This occurred during the transition from the pretreatment conditions, in which the apical membrane is the dominant component of the Nyquist plots, to the equilibrium condition, in which the basolateral membrane is the major component and I scvalues are increasing rapidly. The collection of impedance data requires that two curves be generated with fundamental frequencies of 3.2 and 0.5 Hz (17). Multiple sweeps are taken for each record, so that a total of ∼150 s is needed for data acquisition. Changes inI sc that result from increasing K+ conductance across the membrane during this sampling interval could result in errors in measurement of capacitance. This transient increase inC b was comparably seen in experiments with all three anions, even though the increase inI sc was markedly different among gluconate, , and Cl−. Thus, if the increase in capacitance represents an addition of membrane vesicles to the basolateral membrane, it does not appear to be associated with the magnitude of the increase in K+permeability. In contrast,C b values that were calculated after stableI sc values had been obtained, either from preparations treated with 150 U/ml nystatin or with stepwise addition of 50, 100, and 150 U/ml nystatin, were not different, even thoughR b was lowered by a factor of eight. Values forC b that were calculated from either Eq.1 orEq. 2were not significantly different, and the value forR p chosen forEq. 2was assumed not to change asR b decreased. The similar values forC b obtained by both calculations suggest that changes inR p, if they occur, are small.
That C b was not different despite large changes inR b andI sc provides evidence that the activation of basolateral K+ channels does not require the net addition of vesicles to the basolateral membrane and that channels are activated in situ by cell swelling. These results are consistent with observations by Cantiello et al. (2) that activation of volume-sensitive K+ channels in melanoma cells is dependent on actin-binding protein and their suggestion that volume activation of channels involves an interaction between the actin cytoskeleton and volume-sensitive channels in the membrane.
As noted above, there was a transient increase inC b asI sc values increased in the first 15 min following the addition of 150 U/ml nystatin in experiments with all three anions. If this does, in fact, represent the addition of basolateral vesicles (as opposed to being the result of increasingI sc values during data acquisition), it could be that insertion of new K+ channels is taking place but the activation of these channels requires the volume change associated with the more permeant anions. Experiments to date have shown that blocker-sensitive channels (i.e., verapamil, quinidine, quinine) are also inhibited by the addition of hyperosmotic sugar solutions (12) or gluconate as the anion (unpublished observations). In these experiments, a two-state model for blocker-induced noise analysis has been used, and measurements of open probability are not available. Furthermore, comparison of the kinetics of blocker action on current fluctuations and on the inhibition ofI sc indicates that the kinetics of basolateral K+ channels in larval frog skin and in turtle colon (6, 12) are complex, so that the specific mechanism of channel activation remains elusive.
The lack of change inC b does not completely rule out the possibility that membrane retrieval and insertion are occurring at comparable rates so that new channels are added as membrane vesicles with a lower channel density are retrieved. Increases in membrane capacitance have been shown to correlate with the addition of water-conducting aggrephores in the amphibian urinary bladder (13, 19, 22), and it has been suggested that the stimulation of apical membrane Na+ conductance of A6 cells by antidiuretic hormone is associated with the addition of membrane vesicles containing amiloride-sensitive Na+ channels (18). Harris et al. (10) found that conditions that favor cellular swelling result in greater rates of exocytosis and endocytosis of apical membrane vesicles by toad urinary bladder granular cells when osmotic water flux is stimulated by antidiuretic hormone.
Using noise analysis, Van Driessche and Hillyard (25) estimated the density of basolateral K+ channels in the larval frog skin to be 7.7 channels/μm2 of cross-sectional area. Because the capacitance measurements used to estimate the basolateral membrane area suggest that the actual basolateral membrane area is ∼10- to 15-fold greater than the cross-sectional area, the density of channels would be on the order of 0.5–1 μm−2. Thus a small turnover of vesicles having a high density of channels could result in substantial changes in K+transport as the channels became activated.
Another caveat in the interpretation of the capacitance values is the possibility that the dielectric behavior of cell membranes in this tissue is frequency dependent, so that capacitance values obtained over a range of frequencies would give rise to errors in capacitance measurements (1). With Eqs.1 and2, membrane resistance and capacitance values are determined over the same range of frequencies, and the resistance values,R a andR b, are very similar to the values forR m that were obtained with a single DC voltage pulse. Even with errors in the determination of absolute values forC b, the lack of observed differences in this parameter over the wide range of resistances andI sc values produced in the experimental treatments is most consistent with the activation of in situ channels and/or a very precise mechanism for the control of basolateral membrane area.
We gratefully acknowledge the technical assistance of J. De Beir Simaels.
Address for reprint requests: S. D. Hillyard, Dept. of Biological Sciences, University of Nevada, 4505 Maryland Pkwy., Las Vegas, NV 89154-4004.
S. D. Hillyard was supported in part by National Science Foundation Grant IBN-9215023 and National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-38977-01. H. F. Cantiello was supported in part by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-48040, and W. Van Driessche was supported by Grant G.023.95 from the Fonds Wetenschappelijk-Vlaanderen (Belgium).
- Copyright © 1997 the American Physiological Society