INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

LUMINAL SODIUM ENTRY THROUGH amiloride-sensitive, highly-selective sodium channels (ENaC) in the apical membrane of distal nephron epithelial cells is the rate-limiting step for hormonally stimulated sodium reabsorption. This implies that a thorough understanding of the mechanisms regulating these channels is fundamental to our understanding of total body sodium homeostasis and control of blood pressure.

Previous reports by us and others have shown that ENaC in A6 distal nephron cells are regulated by several different kinases. Ling and Eaton (36) have shown that protein kinase C (PKC)-mediated phosphorylation inhibits ENaC by reducing the open probability (P_o) in native cells, but when expressed in oocytes activation of PKC apparently stimulates ENaC by increasing the number of channels in the surface membrane (56, 57). Studies by Matsumoto and co-workers (40) showed that tyrosine kinase-mediated phosphorylation stimulates channel activity by increasing the number of channels per unit area of membrane. Finally, arginine vasopressin promotes protein kinase A (PKA)-mediated phosphorylation of the channel or some closely associated protein (30). In patch-clamp experiments by Marunaka and Eaton (39), arginine vasopressin, acting through PKA, increases the apical membrane density of sodium channels without affecting P_o. More recently, there have been several observations of the role of serum- and glucocorticoid-dependent kinase (SGK) in regulating ENaC activity (3, 11, 17, 42, 47). The serine/threonine kinase activity of SGK appears to determine the number of ENaC in the surface membrane of transporting cells. There are also reports of the effects of phosphorylation of ENaC, particularly the COOH-terminal domains of the - and -subunits (13, 18, 46), although there is at least one report that suggests that, at least in native cells, ENaC subunits are not constitutively phosphorylated (59) and another which suggests that direct phosphorylation of channel subunits only changes activity under special conditions (13).

Changes in protein function by phosphorylation involves not only regulation of a kinase that phosphorylates but also regulation of a phosphatase that dephosphorylates the protein. Given the substantial evidence for regulation of sodium channels by protein kinase phosphorylation, we decided to explore the phosphatase-mediated dephosphorylation step. Several different phosphatases have been described and cloned from Xenopus laevis (but not A6 cells). Two isoforms of protein phosphatase 2A (PP2A) have been reported (16), and the sequence of X. laevis PP2B (calcineurin), 2C, and 1 are all available in GenBank (accession nos. AB037146, AF019569, and L17039, respectively).

In this study, we investigated the effect of phosphatase antagonists on amiloride-sensitive, highly selective sodium channels (ENaC) in A6 cells. Okadaic acid, a monocarboxylic polyether isolated from marine sponges, is known to inhibit several types of phosphatases (28). It is a very potent blocker of PP1 and PP2A, two of the major serine and threonine phosphatases present in mammalian cytoplasm (14). At least in mammalian cells, PP2A is completely inhibited by 1 nM okadaic acid, whereas the half-maximal inhibitory concentration for PP1 is ~15 nM. On the other hand, inhibition of PP2B is measurable only at concentrations >1 µM, and other phosphatases (including 2C) or kinases are not affected by okadaic acid (14). In addition, we examined the effect of two other phosphatase inhibitors, calyculin A and microcystin. Both are strong inhibitors of PP1 and PP2A, with little effect on other phosphatases (14). In mammalian cells, calyculin blocks both phosphatases with similar potency, but microcystin is more than 40 times as effective at blocking PP2A (IC₅₀ = 0.04 nM) than PP1 (IC₅₀ = 1.7 nM).

We show that pretreatment of A6 cells with 100 nM okadaic acid, 20 nM calyculin A, or 20 nM microcystin all produce an increase in amiloride-sensitive, short-circuit current (I_sc). In addition, our results suggest that pretreatment of aldosterone-stimulated A6 cells with the phosphatase inhibitors affects the P_o through an increase in the mean open time for single sodium channels in cell-attached patches, with a time course similar to the okadaic acid-induced increase in amiloride-sensitive I_sc. On the other hand, the inhibitors do not seem to affect the number of channels present in the apical membrane. This study implies the existence of a phosphatase whose activity tonically reduces the P_o of apical sodium channels, without affecting channel density. This is an interesting result because, based on previous examinations of PKC and PKA (36, 39), dephosphorylation of a serine/threonine phosphorylation site would be expected to increase P_o (36) or reduce channel density (39), respectively. Thus it would appear that there is an additional kinase activity, not mediated by PKC or PKA, that is important in regulating sodium channel activity. This has been suggested in reports from several other laboratories (13, 56, 57).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

A6 cell culture preparation. For single channel experiments, A6 cells from the American Type Culture Collection (Rockville, MD) in passage 68 were prepared as described previously (36). Experiments were carried out on passages 70-80, with no discernible variation between cells from different passages. For transepithelial measurements, A6 cells (subclone 2F3 from Krahenbuhl and Rossier) from passages 96-97 were prepared following methods we have used previously (34, 35, 37). Cells were maintained in plastic tissue culture flasks (Corning) at 26°C in a humidified incubator with 4% CO₂ in air. The culture medium was a mixture of Coon's F-12 medium (3 parts) and Leibovitz's L-15 medium (7 parts) modified for amphibian cells with 104 mM NaCl-25 mM NaHCO₃, pH 7.4, with a final osmolarity of 240 mosmol/kgH₂O. Besides these components, 10% (vol/vol) FBS (Irvine Scientific), 1% streptomycin, and 0.6% penicillin (Hazleton Biologics) were added. Cells grown on plastic tissue culture dishes were detached when confluent by exposing them to divalent-free (calcium and magnesium) medium containing 0.05% trypsin and 0.6 mM EDTA (Irvine Scientific). The cells were then rinsed, centrifuged, repeatedly resuspended, and finally replated. When used for patch-clamp experiments, A6 cells were replated at confluent density on collagen-coated CM permeable filters (Millipore) attached to the bottom of small Lucite disks, and the disks were suspended in 35-mm petri plates, as previously described (36). This sided preparation forms a polarized monolayer with the apical surface oriented upward and net sodium transport moving from the apical to basolateral surface. The bathing medium was supplemented with 1.5 µM aldosterone and 10% FBS. Every 2 days, the cells were fed with fresh medium, and patch-clamp experiments were performed 10 days after replating.

Single channel recordings and data analysis. Before the forming of patches, the apical cell surface was washed carefully several times with our standard extracellular solution containing (in mM) 95 NaCl, 3.4 KCl, 0.8 CaCl₂, 0.8 MgCl₂, and 10 HEPES (Sigma), and pH was adjusted to 7.4 with a small amount of NaOH. Patch pipettes contained the same solution. Experiments were only performed at room temperature (22-23°C) within 45-60 min of removing the A6 cells from the incubator. Patch pipettes with a tip diameter 1 µm were fabricated from WPI TW150 glass (New Haven, CT) and fire-polished following the procedure of Hamill et al. (27). Single channel currents from cell-attached patches were measured with an Axopatch 1-B current-voltage (I-V) clamp amplifier (Axon Instruments, Burlingame, CA), low-pass filtered at 1 KHz, recorded on a digital video recorder (Sony), and then digitized at two times the corner frequency (f_c) using a Scientific Solutions analog-digital converter and IBM PC computer equipped with Axotape software (Axon Instruments). The data were subsequently transferred to a Vax computer (Digital Equipment) for single channel analysis.

Fluctuation analysis. Noise analysis of sodium channels in intact epithelia generally requires the addition of a channel blocker like amiloride or CDPC (6-chloro-3,5-diamino-pyrazine-2-carboxamide) (see, e.g., Refs. 5, 9, 22, 29) to induce an f_c. Fluctuation measurements on single ENaCs reveal a spontaneous corner without the need for addition of pharmacological blockers. This is possible, not because of special properties of single channel patches but rather because of limitations of recording noise fluctuations in intact epithelia. As one of us has discussed previously (20), there are two problems of recording spontaneous ENaC fluctuations from intact epithelial tissue. The first is the exceptionally low f_c of ENaC (because of the long mean open and closed times, the spontaneous f_c is expected to be ~0.01-0.1 Hz). This means that the preparation has to be stable with no change in transport for a very long period of time (15-30 min), which is often difficult to achieve. The second problem exacerbates the first: whole tissue preparations have a 1/f component to the noise in addition to the 1/f² Lorentzian component. Because this component is large at low frequencies, more samples must be acquired at low frequency to resolve the Lorentzian, making the stability problem more serious. Single channel recordings have no 1/f component and can be stable for relatively long periods. Also, records with similar products of N times P_o can be appended to achieve longer records.

Pretreatment of A6 cells with phosphatase inhibitors. For pretreatment experiments, phosphatase inhibitors (LC Services or Calbiochem) were added to both the basolateral and apical bathing medium, achieving a final concentration of about 10 times the IC₅₀ for the phosphatase of interest. Cells were then incubated for 30, 60, or 120 min at 26°C in a humidified incubator with 4% CO₂ in air before cell-attached patch experiments were performed.

Acute perfusion of A6 cells with okadaic acid. Cell-attached patches were obtained on untreated A6 cell monolayers, and baseline channel activity was measured. After the patches were established (5 min), the apical solution was replaced with standard extracellular solution containing 100 nM okadaic acid. Perfusions were accomplished using a 23-gauge hypodermic needle positioned in the apical bath chamber of the Lucite disk and connected to a peristaltic pump (Buchler) via thin silicon tubes. A perfusion rate of 1 ml/min was sufficient to fill the cell chamber in ~60-90 s without disturbing the gigaohm patch seal. The pump was then stopped. For acute perfusion experiments, we always chose patches that initially contained a low number of channels per patch (N  4). In this way, even a short recording time (4-5 min) before the beginning of perfusion was sufficient to obtain a good estimate of the number of channels in the patch. Marunaka and Eaton (39) have previously reviewed the conditions under which highly selective sodium channel number and P_o can be accurately determined using patch-clamp methods.

I_sc recording. For transepithelial electrical parameters, A6 cells were seeded on type I collagen-coated nitrocellulose filters (Costar) at a density of 1 × 10⁶ cells/cm². Mature monolayers were selected for study 7-20 days after seeding. A6 cells were placed in serum-free medium for 40 h immediately before study with the serum-free medium supplemented with 300 nM aldosterone for the last 16 h (overnight). Measurements were made on confluent A6 monolayers using an Ussing chamber specifically designed to hold the nitrocellulose filters (4.7 cm²). The apical and basolateral bathing solutions had the same composition as the standard extracellular solution used for patch-clamp experiments. Baseline I_sc and potential difference measurements were initially made, and then okadaic acid was added to the extracellular bath. Amiloride (10 µM) was added to the apical compartment to measure the amiloride-sensitive component of I_sc. Exposure of A6 cells to vehicle alone produced no effect on baseline transepithelial parameters.

Okadaic acid-induced phosphorylation of cellular proteins and sodium channels. Okadaic acid should promote protein phosphorylation and might promote phosphorylation of sodium channel subunits. To examine these questions, we examined total cellular phosphorylation and phosphorylation of the -, -, and -subunits of the ENaC. A6 cells were deprived of serum overnight. The cells were washed two times with phosphate-free DMEM adjusted to amphibian osmolarity and supplemented with 5 mM glutamine and 1.5 µM aldosterone. After being washed, the cells were incubated in phosphate-free DMEM for 2 h before replacement of the media with fresh phosphate-free DMEM and incubation for another 2 h. Finally, the media was replaced with new phosphate-free DMEM containing a total of 5 mCi NaH₂³³PO₄ or NaH₂³²PO₄ overnight. The next morning, okadaic acid at a final concentration of 100 nM along with phosphate-free DMEM plus ³³PO₄ or ³²PO₄ was added to one group of cells while only medium was added to the other, and the cells were incubated overnight. ³³P was used in experiments to examine all cellular proteins with the hope that its lower energy emissions would produce narrower bands when imaged: a hope that was not fully realized (see Fig. 11). The next day, the cells were chilled on ice and scraped in a solution containing (in mM) 100 KCl, 1 EGTA, and 10 KH₂PO₄ adjusted to pH 7.4 with KOH and containing protease inhibitors (100 µM antipain, 100 µM leupeptin, 100 µM N-tosyl-L-lysyl chloromethyl ketone, 100 µM N-tosyl-L-phenyl chloromethyl ketone, and 1 mM phenylmethylsulfonyl fluoride). The cells were centrifuged and lysed in 400 µl of 50 mM Tris buffer by douncing 40 times. The cells were centrifuged for 10 min in a tabletop centrifuge, and the precipitate was dounced again in 400 µl of 50 mM Tris buffer and centrifuged for 10 min. The combined supernatant was precipitated with 10× volume of cold acetone for 15 min at 4°C and then centrifuged in a tabletop centrifuge for 10 min. The precipitate was dissolved in 5× TBS. Part of the lysate was reserved to examine total cellular phosphorylation. Preimmune serum (20 µl) was added to the remaining lysate and incubated at room temperature for 1 h. Protein A beads (220 µl) were added for 30 min and centrifuged two times for 10 min at 13,000 revolutions/min. Polyclonal anti--, anti-, or anti- X. laevis ENaC (20 µl) antibodies were added to one-third of the lysate and incubated overnight. The antibodies have been characterized in previous work (32, 33, 35, 38, 41, 44, 51). The next day, 400 µl of protein A beads were added and incubated for 30 min before centrifuging two times for 10 min at 13,000 revolutions/min. The precipitate was washed three times with 150 mM NaCl, 5 mM EGTA, 1% (vol/vol) Triton X-100, and 50 mM Tris · HCl (pH 7.4) and washed two times with 50 mM EGTA, 0.4% (wt/vol) sodium deoxycholate, 1% (vol/vol) Nonidet P-40, and 10 mM Tris · Cl (pH 7.4). The precipitate was boiled with sample buffer for 10 min and resolved on a 7.5% SDS-PAGE, and the phosphorylation was determined by quantitation of the gel on a PhosphorImager.

PCR cloning of A6 cell phosphatases. Phosphatase clones were amplified by PCR from A6 cDNA prepared by using a Marathon cDNA amplification kit (Clontech). The PCR primers were based on consenus sequences for X. laevis phosphatases obtained from GenBank (accession nos. L17039, X62114, and Z50852). For PP1 phosphatases, we used AAGAGAATGAGATTCGAGGGC for the forward primer and ACGGGTCTGCTTGCGTTAGG for the reverse primer. For PP2A phosphatases, we used the degenerate sequences GTGGATCGAGCAGCTGAAYGA (Y = T,C) for the forward primer and CCAAAGTAAGSCCATTAGCATGG (S = C,G) for the reverse primer. The PCR reaction conditions were 30 s at 94°C and 30 s at 60°C followed by a 2-min extension at 68°C for 30 cycles. Products of the expected size were subcloned into pGEM vector, and the nucleotide sequence of the PCR product was determined by Sanger dideoxynucleotide sequence analysis.

Statistics. Unless otherwise stated, data are presented as means ± SE. Paired or unpaired Student's t-test or ANOVA with a Tukey posttest was used as appropriate to test for significance.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

A6 cells contain PP1 and PP2A isoforms. Using PCR primers to conserved regions of PP1 and PP2A that should amplify any of the respective phosphatase isoforms, we were able to isolate and clone three phosphatase isoforms from A6 cells: one for PP1 (PP11) and two for PP2A (PP2A and PP2A). The sequence of these clones was identical, at an amino acid level, to the previously reported sequences for phosphatases from other X. laevis tissues (16, 31, 55). We did not attempt to clone other phosphatases, since none of the agents we used will inhibit other categories of phosphatase.

Pretreatment with okadaic acid increases I_sc. The amiloride-sensitive component of I_sc was measured hourly (for up to 4 h) from aldosterone-stimulated A6 cell monolayers in either the absence or presence of 100 nM okadaic acid (5 monolayers for each condition) in the extracellular bath (Fig. 1). Exposure to okadaic acid increased the transepithelial I_sc, beginning ~1 h after addition of the phosphatase inhibitor and peaking to ~1.5 times baseline I_sc at ~2 h after addition. In contrast, the stimulatory effect of okadaic acid was quite small in A6 cells grown in the absence of aldosterone (data not shown). Lower doses of okadaic acid had no effect on transepithelial electrical parameters (data not shown).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Okadaic acid increases short-circuit sodium current (Isc) in A6 cells. Amiloride-sensitive Isc was recorded from A6 cells monolayers in either the absence (, n = 6) or in the presence of 100 nM okadaic acid (, n = 6). Isc is expressed as a fraction of the initial baseline Isc value in untreated cells (a value of 8.3 ± 0.91 µA/cm2). All values are means ± SE.

Pretreatment with okadaic acid decreases the f_c of power spectra. We applied fluctuation analysis methods to the single channel records of cell-attached patches that contained a large number of channels per patch (N > 4). Figure 2 shows typical current records from two such patches, one untreated and one pretreated for 1 h with 100 nM okadaic acid. These and other similar records were used to generate power spectra and amplitude histograms from the same current records. The amplitude histograms indicate that both patches contain eight or more channels. Figure 2, bottom, shows the power spectra calculated from the same two current records. The power spectra from a number of similar patches (all with no potential applied to the pipette) were used to calculate the kinetic characteristics for a large number of sodium channels (see MATERIALS AND METHODS). We found that the f_c of power spectra derived from control cells (0.35 ± 0.051; n = 8) were significantly higher than those from okadaic acid-pretreated cells (0.22 ± 0.0070; n = 4 after 30 min, and 0.27 ± 0.0090; n = 4 after 1 h), suggesting that phosphatase inhibition produced a change in the P_o of single sodium channels. Consideration of the mean current and unit current indicates that the P_o increases (Table 1). On the other hand, although the estimate of apical membrane channel density from power spectra was not very precise, our analysis suggested that, if channel number was increased by okadaic acid, the increase must have been small (<25%).

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Okadaic acid reduces the corner frequency of power spectra obtained from cell-attached patches with large numbers of channels. Left: data from a patch on an untreated cell. Right: a similar patch except that the cell was pretreated for 1 h with 100 nM okadaic acid. Downward transitions are inward current associated with channel openings. Top: current records in which it is difficult to observe discrete channel transitions. c, Closed state. Middle: amplitude histograms produced from the current records indicating that there are at least 8 channels in each patch. In the histogram from an okadaic acid-treated patch, there is a shift leftward toward occupancy of states with a larger number of channels open. Bottom: power spectra for the patches. The corner frequency of the okadaic acid-treated patch is shifted slightly to a lower frequency.

Pretreatment with okadaic acid increases NP_o for sodium channels. To further investigate the effects of okadaic acid on individual sodium channel kinetics, analysis of single channel recordings was also performed on the cell-attached patches that contained a small number of channels per patch (N  4). A representative example of baseline sodium channel activity in a cell-attached patch from an untreated A6 cell is shown in Fig. 3, top left. Current traces represent ~1.5 min of continuous recording at resting membrane potential, with downward transitions representing inward sodium current. The amplitude histogram generated from this recording shows that this patch membrane contained three active sodium channels, whose P_o was ~0.36 (Fig. 3, bottom left). This experiment can be compared with a representative cell-attached recording from an A6 cell after 1 h of pretreatment with 100 nM okadaic acid (Fig. 3, top right). Notice that the closed level, representing zero channel activity in the patch, was very rarely reached after phosphatase inhibition. From the amplitude histogram, this recording also contained at least three active channels, whose P_o was ~0.58 (Fig. 3, bottom right). Table 1 summarizes our measurements from patches with multiple channels on untreated cells and cells pretreated with okadaic acid. Okadaic acid pretreatment for 1 h increased mean NP_o (1.7 ± 0.46; n = 15) compared with controls (0.74 ± 0.12; n = 25).

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Effect of okadaic acid on a patch containing a small number of sodium channels. Left: sodium channel activity from a cell-attached patch formed on an untreated cell. Right: activity from a cell-attached patch formed on a cell after 1-h pretreatment with 100 nM okadaic acid. Top: current traces at resting membrane potential (VP = 0 mV) from cell-attached patches containing at least three channels. Horizontal lines to left of each trace show closed level (c) with inward current downward. Currents were software filtered at 100 Hz. Bottom: amplitude histograms obtained from recordings at the top.

Pretreatment with okadaic acid increases P_o without affecting channel density. For the patches containing a small number of channels (N  4) and a P_o >0.1 and <0.9, N could be estimated with good accuracy, even for short recording periods (4-5 min), making it possible to calculate P_o (see MATERIALS AND METHODS). The mean P_o value measured from control patches was 0.29 ± 0.027 (n = 25), whereas the P_o after 1 h of pretreatment with okadaic acid increased by ~40% (0.38 ± 0.053; n = 15). Therefore, this increase in sodium channel P_o appeared to be sufficient to fully account for the 40-50% increase in amiloride-sensitive I_sc previously observed after okadaic acid exposure. Consistent with our fluctuation analysis results, the number of sodium channels per patch from cells pretreated with okadaic acid (3.7 ± 0.93; n = 15) was not significantly different from that measured in the controls (3.1 ± 0.27; n = 25).

Pretreatment with okadaic acid increases the mean time that channels are open. A change in P_o could be because of a change in either the mean open or closed time of apical sodium channels. Table 1 shows that 1 h of pretreatment with okadaic acid increased the mean time open (see MATERIALS AND METHODS) of sodium channels from 92 ± 18 to 170 ± 28 ms. The average open time for multiple channels is not equivalent to the mean residency time in any specific channel state (which requires information about the correct kinetic model for channel transitions) but is the average open time for all open states. The increase in average open time (~70%) produced by okadaic acid seemed sufficient to account for the observed changes in P_o. This result suggested that the activity of an okadaic acid-sensitive phosphatase decreases the fraction of time during which the sodium channels are open.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Effect of okadaic acid on a cell-attached patch containing a single sodium channel. Left: sodium channel activity from a patch on an untreated cell. Right: activity from a cell-attached patch formed on a cell after 1 h of pretreatment with 100 nM okadaic acid. Top: current traces at resting membrane potential (Vp = 0 mV) from a cell-attached patch containing only one channel. Horizontal lines to left of each trace show closed level (c) with inward current downward. Data were software filtered at 100 Hz. Bottom: amplitude histograms obtained from recordings at the top.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Okadaic acid increases the mean open time of sodium channels. Interval histograms were constructed from 10 min of data from single channel patches. A: open interval histograms. B: closed interval histograms. Left: cell not pretreated. Right: cell pretreated for 1 h with 100 nM okadaic acid. Only the mean open time increases after treatment with okadaic acid.

Relative frequency of observation of patches with differing number of channels is not altered by okadaic acid. We used a variety of analysis methods to determine that the P_o of channels was increased by okadaic acid, but we wished to assure ourselves that okadaic acid (or the other phosphatase inhibitors) was not altering channel density (N). Therefore, we tabulated the number of channels per patch for all the patches with channels used in this study (48 untreated, 33 treated with okadaic acid, and 10 treated with other phosphatase inhibitors). The probability of observing different numbers of channels in a patch is shown in Fig. 6 for untreated and phosphatase inhibitor-treated cells. There is no statistically significant difference between the distributions, although we cannot rule out a very small increase in N after inhibitor treatment. This distribution does not take into consideration the fact that 34% of the seals we formed had no detectable channels within the patch (and also did not appear to depend on treatment).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Relative frequency of observation of patches with differing number of channels. The no. of channels per patch was tabulated for all patches with channels used in this study (48 untreated, 33 treated with okadaic acid, and 10 treated with other phosphatase inhibitors). The probability of observing different numbers of channels in a patch is shown for untreated (filled bars) and phosphatase inhibitor-treated (open bars) cells. There is no statistically significant difference between the distributions, although we cannot rule out a very small increase in the no. of channels per patch after inhibitor treatment. This distribution does not take into consideration the fact that 34% of the seals we formed had no detectable channels within the patch (and also did not appear to depend upon treatment).

Two other phosphatase inhibitors alter channel activity like okadaic acid. When cells were pretreated with either 20 nM calyculin A or microcystin, channel P_o increased, with most if not all of the increase because of an increase in channel open time. Calyculin A pretreatment for 1 h increased mean P_o (0.47 ± 0.0049; n = 5) compared with controls (0.36 ± 0.019; n = 5) ~1.4-fold, similar to the results with okadaic acid. Calyculin A pretreatment also increased mean open time from 1.3 ± 0.075 to 2.3 ± 0.068 s, whereas mean closed time did not change significantly (2.3 ± 0.16 to 2.6 ± 0.072 s). Microcystin pretreatment for 1 h also increased mean P_o from 0.37 ± 0.013 to 0.49 ± 0.0064 (n = 5), again an ~1.4-fold increase. Microcystin pretreatment also increased mean open time from 1.3 ± 0.029 to 2.5 ± 0.053 s, whereas mean closed time did not change significantly (2.3 ± 0.16 to 2.6 ± 0.069 s). These results are summarized in Table 1.

An inactive analog of okadaic acid, okadaone, has no effect on channel P_o. Okadaic acid has occasionally been reported to have effects unrelated to its inhibition of phosphatases. These effects can be controlled for by examining the effect of a structurally related compound, okadaone, that does not inhibit phosphatases (14). Therefore, in one batch of cells we examined single channels in untreated, okadaic acid-treated, and okadaone-treated patches. In five patches in each condition, okadaic acid-treated channels consistently had a P_o higher (0.41 ± 0.028) than untreated channels (0.23 ± 0.027, P < 0.05) and okadaone-treated channels (0.19 ± 0.016, P < 0.05), and there was no significant difference in P_o between untreated channels and okadaone-treated channels (Fig. 7). Therefore, we conclude that the effect of okadaic acid is specifically related to its ability to inhibit PPs.

View larger version (11K): [in this window] [in a new window]   Fig. 7.   An inactive analog of okadaic acid, okadaone, does not alter the open probability (Po) of sodium channels. Single channels were examined in patches from untreated, okadaic acid-treated, and okadaone-treated cells. In five patches in each condition, okadaic acid-treated channels consistently had a Po higher (0.41 ± 0.028) than untreated channels (0.23 ± 0.026, P < 0.05) and okadaone-treated channels (0.19 ± 0.016, P < 0.05), and there was no significant difference in Po between untreated channels and okadaone-treated channels.

Incubation with 10 nM okadaic acid or 0.4 nM microcystin does not affect sodium channel activity. To distinguish whether the okadaic acid effect was directed on PP1 or PP2A, we pretreated A6 cells for 1 h with 10 nM okadaic acid or 0.4 nM microcystin (remembering that the IC₅₀ of okadaic acid is 50-300 nM for PP1 and 0.1-2 nM for PP2A, and that of microcystin is 1.7 nM for PP1 and 0.04 nM for PP2A; see Ref. 14). This treatment did not produce any significant effect on sodium channel activity, suggesting that the inhibitor's effects on sodium channels are mediated by inhibition of a type 1 phosphatase. The mean sodium channel P_o after incubation with okadaic acid was 0.17 ± 0.050 (n = 6) compared with the control mean P_o of 0.18 ± 0.089 (n = 4). For microcystin treatment of a different batch of cells, the mean P_o was 0.29 ± 0.034 (n = 5) compared with the control mean P_o of 0.30 ± 0.049 (n = 5). Although these results appear to imply a type 1 phosphatase, some care should be used in interpreting these results, since the pharmacology is based on the responses in mammalian cells and the response of X. laevis phosphatases may be different.

Pretreatment with okadaic acid does not alter the I-V relationship. Figure 8 shows the I-V relationship for sodium channels from cell-attached patches obtained by plotting the single channel current amplitudes measured at various negative applied pipette potentials, V_p (negative values = hyperpolarization, positive values = depolarization). Even without knowing the actual apical membrane potential, V_p values ranging from 80 to +60 mV should cover the physiological range of transmembrane potentials. The I-V curves obtained from untreated cells and cells pretreated with okadaic acid for 30 min and 1 h () were perfectly superimposable. Importantly, they were also consistent with I-V curves previously obtained for highly selective sodium channels in A6 and mammalian distal nephron cells (24). It is thus clear that okadaic acid does not affect transepithelial sodium transport through alterations in the rectification properties, ion selectivity, or the unit conductance of apical sodium channels.

View larger version (16K): [in this window] [in a new window]   Fig. 8.   Okadaic acid does not alter the sodium channel current-voltage (I-V) relationship. Single sodium channel current amplitudes (pA) were recorded at various pipette potentials (Vp) from untreated cells () or cells pretreated with 100 nM okadaic acid for 30 min () or 1 h (). Data points represent means ± SE. Lines are the best least squares fits to the data, and the unit conductances (~3.8 pS) were calculated from the slope near the resting membrane potential.

Time course for okadaic acid-induced effects. From all of the data in Table 1 combined, we plotted the time course for the stimulatory effect of phosphatase inhibition on the P_o of sodium channels (Fig. 9). After 30 min of okadaic acid pretreatment, the sodium channel P_o was increased but was not significantly different from that measured under control conditions. Therefore, it appeared that the okadaic acid-induced effect was only detectable in A6 cells pretreated for >30 min and reached its maximal value between 1 and 2 h after phosphatase inhibitor addition. The time course for the okadaic acid-induced increases in amiloride-sensitive I_sc (Fig. 1) was similar to the time course for increases in individual sodium channel P_o (Fig. 9), although the I_sc response appeared to be delayed relative to the single channel response. Despite the differences in the two measurement techniques and the use of two different A6 cell clones, these results appear to be consistent.

View larger version (12K): [in this window] [in a new window]   Fig. 9.   Time course for okadaic acid-induced stimulation of Po. Po values relative to the Po at time 0 were measured from patches as a function of pretreatment time with 100 nM okadaic acid. Data points are taken by combining the values of Po for all patches represented in Table 1 and are means ± SE. At 30 min, there is an increase, but it is not statistically significant (P = 0.053). At 1 h, the okadaic acid increase is highly significant (P < 0.01).

Acute perfusion with okadaic acid affects sodium channel P_o. We and other groups have previously noted the large variability in P_o that exists among single, high-selectivity sodium channels from different cell-attached patches (24). Because such variability might have obscured our ability to detect a subtle effect of okadaic acid pretreatment before 1 h, we also examined individual cell-attached patches before and after acute perfusion with 100 nM okadaic acid. Immediately after a gigaohm seal was established, baseline channel activity was recorded for ~4-5 min. The phosphatase inhibitor was then perfused in the apical bath chamber for ~10 min (see MATERIALS AND METHODS). We then proceeded with a single channel recording for as long as high-resistance seals could be maintained. During the course of these experiments, all patches were held at resting membrane potential. To maximize our ability to detect a subtle, early, okadaic acid effect, these experiments were performed on patches initially containing low baseline sodium channel activity (low N, a low P_o, or both). In general, the acute effect of okadaic acid was not large, although P_o decreased in only two of the eight patches we examined (Fig. 10B). However, comparing channel activity before and after application of okadaic acid may be inappropriate, since we really should compare the effect of okadaic acid with the effect of no treatment. This is particularly so because highly selective sodium channels in A6 cell-attached patches often show a spontaneous decay in activity ("rundown"), which could obscure the stimulatory effect of okadaic acid. Therefore, for comparison, we examined 12 patches from the same batch of cells with initial P_o comparable to those treated with okadaic acid. We recorded for an initial 4-min period, perfused the patches with A6 saline for ~10 min, and then recorded for four additional minutes to produce recordings comparable to the time course for okadaic acid-perfused patches. Figure 10A plots P_o for the first (left) and second (right) 4 min of cell-attached recording and shows a 49% decrease in mean P_o from 0.25 ± 0.036 to 0.16 ± 0.033 (P = 0.006; paired t-test). This frequent spontaneous loss of channel activity in the cell-attached configuration strongly suggested that we could have underestimated the stimulatory effects of okadaic acid. The mean changes in P_o between the first and second 4 min of recording from untreated cells (0.10 ± 0.030; Fig. 10A and Table 2) and between the 4-min periods before and after okadaic acid perfusion (+0.047 ± 0.06; Fig. 10B and Table 2) were significantly different (P = 0.019; t-test). However, despite acute perfusion of A6 cells with okadaic acid, channel activity still eventually decayed (data not shown). Unfortunately, we were unable to maintain stable cell-attached patches for >30-40 min, preventing a direct comparison of acute okadaic acid-induced effects with that after more prolonged pretreatment.

View larger version (15K): [in this window] [in a new window]   Fig. 10.   Acute okadaic acid perfusion prevents increases in the Po of sodium channels above that for untreated patches. A: sodium channel Po for cell-attached patches at the resting membrane potential are calculated for the 4-min period after a gigohm seal is established and then for the subsequent 4 min after 10 min of saline perfusion. B: Po is calculated for the 4-min periods after a gigohm seal is established and the subsequent 4-min period after a 10-min perfusion with 100 nM okadaic acid. Lines connecting symbols represent data from the same cell-attached patch. In patches on untreated cells, mean Po values for the first period (0.25 ± 0.036) and the second period (0.15 ± 0.033) were significantly different by paired t-testing (P < 0.001). Mean Po values before (0.24 ± 0.059) and after (0.29 ± 0.051) okadaic acid perfusion were not significantly different by paired t-testing.

Okadaic acid alters the phosphorylation of some cellular proteins but does not alter the phosphorylation of -, -, or -ENaC. Because of the manner in which we performed our experiments, the effect of the phosphatase inhibitors could be due to either a change in the phosphorylation of the sodium channels themselves or the phosphorylation of another protein, which in turn regulates sodium channel activity. In an attempt to investigate these possibilities, we labeled A6 cells to equilibrium with [³³P]phosphate, washed out excess phosphate, and then examined the effect of a 1-h treatment of the cells with okadaic acid on the patterns of total cellular phosphorylation. The problem with this approach is that phosphatases, unlike kinases, have relatively nonspecific targets; thus, decreasing phosphatase activity will generally increase the phosphorylation of many proteins that were originally phosphorylated by quite specific serine/threonine kinases. This problem is obvious in Fig. 11A, where a few phosphoprotein bands (imaged on a PhosphorImager) are relatively distinct in untreated cells, but after okadaic acid treatment there are too many bands to resolve. This does indicate that there are many proteins that are differentially phosphorylated by okadaic acid; however, we could identify no single protein band as a candidate for producing the okadaic acid-induced increase in sodium channel P_o. We also wished to test whether sodium channel proteins themselves might be a target for okadaic acid-induced phosphorylation. Sodium channel proteins consist of three subunits, , , and . We could identify the phosphorylation of the subunits by immunoprecipitation of the subunit with antisubunit antibodies. When we immunoprecipitated the three subunits from cellular extracts of okadaic acid-treated or untreated cells labeled as above and examined the immunoprecipitate on SDS gels, we could find no evidence for differential phosphorylation of ENaC (Fig. 11B). However, the baseline phosphorylation, particularly of the - and -subunits, was high enough that we might not have detected as much as a 50% change in the okadaic acid-induced phosphorylation.

View larger version (53K): [in this window] [in a new window]   Fig. 11.   Okadaic acid does not alter the phosphorylation of epithelial sodium channel (ENaC) subunit proteins. We labeled A6 cells to equilibrium with [33P]- or [32P]phosphate, washed out excess phosphate, and then examined the effect of 1 h of treatment of the cells with okadaic acid on the phosphorylation of all proteins and the phosphorylation of ENaC. A: all phosphoproteins were labeled in the presence and absence of okadaic acid. This shows that okadaic acid does increase the levels of protein phosphorylation, but it also shows that examining all phosphoproteins is not very useful since, after okadaic acid, there are too many bands to resolve individual proteins. Mol wt, molecular weight. B: on the other hand, when we immunoprecipitated the three ENaC subunits from 32P-labeled cellular extracts of okadaic acid-treated or untreated cells labeled as in A, resolved the immunoprecipitate on SDS gels, and quantitated the phosphorylation on a PhosphorImager, we could find no evidence for differential phosphorylation of ENaC by okadaic acid. However, the baseline phosphorylation, particularly of the - and -subunits, was high enough that we might not have detected as much as a 50% change in the okadaic acid-induced phosphorylation. The bands for the - and -subunits run at 95 and 98 kDa, respectively, which is similar to that reported by us and others in previous work. The -subunit consists of two bands at 78 and 85 kDa, which we presume corresponds to core glycosylated and unglycosylated forms of , as previously reported by others.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We have demonstrated that there are potential targets in A6 cells for all of the phosphatase inhibitors we applied. However, despite the presence of both PP1 and PP2A, based on the pharmacology PP1 is a possible candidate to alter the activity of ENaCs.

On the basis of transepithelial current measurements alone, it is impossible to determine the specific mechanisms responsible for the okadaic acid-induced increase in sodium permeability, since total sodium current per unit area of apical membrane is the product of the unit conductance, P_o, N, and electromotive driving force. In theory, modification of any of these parameters could alter apical sodium flux. We know from our previous studies that PKC-mediated phosphorylation alters P_o, PKA alters N, and different tyrosine kinases may alter both N and P_o (36, 39). Therefore, to correlate phosphatase activity with a specific protein kinase pathway requires an examination of how dephosphorylation could affect the parameters that alter sodium transport. To do this, we supplemented transepithelial current measurements with single channel recordings from cell-attached patches and simple measurements of cellular phosphorylation. We used both standard single channel statistical analysis of records from patches containing one or a few channels (N  4) and fluctuation analysis of patches containing a large number of channels (N > 4). Use of these techniques allowed us to obtain consistent information on the effects of okadaic acid and other phosphatase inhibitors, both on the number of channels per patch (N) and on intrinsic channel properties like P_o and the rate constants for individual channel openings and closings.

Sodium channel activity is dependent on phosphorylation/dephosphorylation reactions. In mineralocorticoid-stimulated A6 cells, okadaic acid induces an increase in the P_o of highly selective sodium channels with little, if any, effect on apical membrane channel density. This is consistent with the presence of an okadaic acid-sensitive phosphatase that counteracts a stimulatory protein kinase-mediated phosphorylation step. Indeed, several different types of phosphorylation reactions are known to modulate sodium transport in distal nephron cells. However, the literature and our previous work preclude many of these protein kinase pathways from being responsible for the phosphorylation step linked to the phosphatase blocked by okadaic acid. Our results indicate that the serine/threonine kinase associated with okadaic acid-inhibitable phosphatase must stimulate sodium channels by increasing their P_o. Phosphorylation reactions mediated by PKC in native cells are inhibitory (23, 36 and see below).

The I-V relationship is not affected by phosphatase inhibition. Pretreatment of A6 cells with okadaic acid did not appreciably alter the single sodium channel I-V curve. Therefore, okadaic acid had no affect on the sodium channel unit conductance or ion selectivity. However, because the short-circuit sodium current and sodium channel P_o are increased by okadaic acid, one may wonder why the apical membrane potential did not appear to change with the increase in apical sodium entry (i.e., the curve of the I-V reversal potential did not shift). The most likely explanation for this observation is that the increased intracellular sodium activity after okadaic acid treatment contributed to an increase in the rate of Na⁺-K⁺-ATPase pump exchange, with a concomitant increase in transepithelial voltage, as is generally observed in tight epithelia (19, 26). Such an increase would hyperpolarize