Differential Effects of Protein Kinase C on the Levels of Epithelial Na+ Channel Subunit Proteins*

Regulation of epithelial Na+channel (ENaC) subunit levels by protein kinase C (PKC) was investigated in A6 cells. PKC activation altered ENaC subunit levels, differentially decreasing the levels of both β and γ, but not αENaC. Temporal regulation of β and γENaC by PKC differed; γENaC decreased with a time constant of 3.7 ± 1.0 h, whereas βENaC decreased in 13.9 ± 3.0 h. Activation of PKC also resulted in a decrease in trans-epithelial Na+reabsorption for up to 48 h. PMA activation of PKC resulted in negative feedback inhibition of PKC protein levels beginning within 4 h. Both β and γENaC levels, as well as transport tended toward pretreatment values after 48 h of PMA treatment. PKC inhibitors attenuated the effects of PMA on ENaC subunit levels and Na+ transport. These results directly show for the first time that PKC differentially regulates ENaC subunit levels by decreasing the levels of β and γ but not αENaC protein. These results imply a PKC-dependent, long term decrease in Na+ reabsorption.

Sodium homeostasis is essential to proper maintenance of total body water and electrolyte content, and thus, blood pressure control. The activity of luminal, epithelial Na ϩ channels (ENaC) 1 is rate-limiting for trans-epithelial Na ϩ reabsorption across the renal collecting duct and other Na ϩ reabsorbing epithelium. Thus, understanding regulation of ENaC activity is relevant to physiology as well as to treating disease with associated fluid imbalance.
ENaC is a heterotetrameric channel complex composed of at least three homologous but distinct subunits: ␣, ␤, and ␥ (1). Numerous results show that expression of ENaC subunit message and protein are differentially regulated in various tissues and species (reviewed by Refs. [2][3][4]. For example, Masilamani and colleagues (5) recently showed in rat collecting duct principal cells that ␣ENaC protein levels increase in response to aldosterone; however, we found in the amphibian A6 cell model of the collecting duct principal cell that ␣ENaC is not significantly influenced by aldosterone, but ␤ENaC protein levels are increased in response to steroid (Ref. 6; also refer to Fig. 1 of the present study). Besides aldosterone, Zentner et al. (7) recently showed in the rat parotid epithelial cell line, Pa-4, that expression of ␣ENaC mRNA and possibly protein was decreased within 6 h by protein kinase C activation.
Activation of PKC decreases Na ϩ reabsorption across renal epithelium by affecting ENaC (8 -11). Studies of single channel properties show that in amphibian, rat, and rabbit distal tubule cells, ENaC activity is decreased within 5-10 min after activation of PKC (12)(13)(14)(15). A rapid initial decrease in ENaC open probability is, in part, responsible for the early change in activity; however, it appears that PKC may also subsequently affect the number of functional channels (14,15). Although most studies are consistent with PKC decreasing ENaC open probability initially and then subsequently reducing channel number, Els et al. (16) showed with blocker-induced noise analysis in A6 cells that PKC activation, besides producing the initial decrease in open probability and a longer term decrease in channel number, might also lead to a small initial compensatory increase in channel number presumably in response to decreased open probability. Nevertheless, actions on channel open probability likely precede effects on number and both reduce overall sodium transport. In a provocative study, Ishikawa and colleagues (15) report biphasic actions of Ca 2ϩdependent processes (e.g. activation of PKC) on repression of ENaC activity with time constants ranging between 1-2 and 100 -160 min. The current results are consistent with the latter time constant representing a decrease in ENaC number.
Recently, Shimkets et al. (17) showed for the first time that PKC directly phosphorylates ENaC when this ion channel is overexpressed and PKC is substantially activated. It is unclear how PKC-mediated phosphorylation relates to ENaC kinetics and number and, thus, Na ϩ reabsorption. Moreover, the long term actions of PKC on transport and ENaC protein levels have not been studied.
Because ENaC subunit levels can be differentially regulated in response to various factors and PKC is known to affect ENaC activity, perhaps, in part, through regulation of channel number, we tested the hypothesis that PKC differentially regulates subunit protein levels. This is the first report directly showing that ENaC subunit protein levels are differentially affected by PKC with both ␥ and ␤, but not ␣ENaC, decreasing in response to kinase activation. The decrease in subunit levels is consistent with the long term actions of PKC on ENaC resulting in decreased Na ϩ channel number and, thus, transport. national, Naperville, IL) avidly reabsorbing Na ϩ were used for all experiments.
Trans-epithelial Na ϩ current was calculated as the ratio of transepithelial voltage to trans-epithelial resistance under open circuit conditions using a Millicel Electrical Resistance System with dual Ag/AgCl pellet electrodes (Millipore Corp.) to measure trans-epithelial potentials and resistances. If the current-voltage relationship of the monolayer is close to linear, this measurement is exactly equivalent to short circuit current. However, trans-epithelial currents calculated in this fashion do not require long term short circuiting of the tissue, which alters intracellular ion concentrations and, therefore, may alter single channel properties. The open circuit measurements, in addition, ensured that larger numbers of monolayers could be routinely sampled, thus ensuring that adequate cellular material was available for subsequent biochemical analysis, as well as enabling the pairing of electrical and biochemical measurements. Movement of cations from the lumen to serosal fluids is represented as positive current. With this preparation, the majority of trans-epithelial current was amiloride-sensitive and carried by Na ϩ via ENaC from lumen to serosal fluid.
A6 cells were extracted with RIPA buffer: 10 mM NaPO 4 , 150 mM NaCl, 1% deoxycholate, 1% Triton X-100, 0.1% sodium dodecyl sulfate, supplemented with 1 M phenylmethylsulfonyl fluoride (pH 7.2). After clearing of cellular debris, standardization of total protein, and addition of sample buffer (containing 10% glycerol, 1% SDS, and 20 mM dithiothreotol), lysates were boiled at 85°C for 10 min. Proteins were separated by standard SDS-polyacrylamide gel electrophoresis (7.5% gels) and subsequently electrophoretically transferred to nitrocellulose. Western blot analysis was performed using standard techniques and appropriate antibodies (see below). 0.1% Tween-20 and 5% dried milk (Carnation) were used as blocking agents. All secondary horseradish peroxidase conjugates were purchased from Kirkegaard & Perry Laboratories (Gaithersburg, MD). Western blots were developed with the ECL detection system (Amersham Pharmacia Biotech) and quantified with densitometric scanning using Sigmagel (Jandel Scientific). When possible, the flood configuration was used to measure band density.
Protein kinase C was detected with the commercially available antibody PKC (MC5) purchased from Santa Cruz Biotech. (Santa Cruz, CA). This antibody identifies ␣, ␤, and ␥ isoforms of PKC. The anti-␥-xENaC antibody used in this study was a kind gift from John P. Johnson (Department of Medicine, University of Pittsburgh School of Medicine). This affinity purified, polyclonal antibody, developed against Xenopus laevis ␥ENaC (residues 608 -660) in chicken, has been described previously (19) and specifically identifies an intrinsic membrane protein of 85-95 kDa. The polyclonal anti-␣ENaC antibody used in this study was raised in rabbit (by Lofstrand Labs, Gaithersburg, MD) against residues 137-161 of X. laevis ␣ENaC linked to keyhole limpet hemocyanin through an amino-terminal cysteine. Antiserum was subsequently affinity purified against inoculating immunogen using standard protocols to produce the final anti-␣ENaC antibody, AB586. Development and use of this antibody has been described previously (6). This antibody specifically recognizes a membrane protein of appropriate size (85-90 kDa), as well as in vitro translated ␣ENaC. The polyclonal anti-␤ENaC antibody used in this study was developed with a similar procedure. The anti-␤ENaC antibody (AB592) was against residues 624 -647 of ␤-xENaC. Affinity purified AB592 specifically recognizes an intrinsic membrane protein of appropriate size (90 -95 kDa) and in vitro translated ␤ENaC. These three anti-ENaC subunitspecific antibodies have no detectable, inappropriate cross-reactivity with other subunits.
All chemicals were purchased from Sigma, BIOMOL (Plymouth Meeting, PA), or Bio-Rad unless indicated otherwise. Phorbol 12-myristate 13-acetate (PMA) and 4␣-PMA were prepared fresh in Me 2 SO prior to each experiment. Final concentrations of Me 2 SO never exceeded 0.05% and were without effect when applied alone. For time course experiments, PMA was replenished for the longer time points every 24 h.
Data are reported as the means Ϯ S.E. The t test and a one-way analysis of variance plus the Student-Newman-Keuls test were used to compare data where appropriate. p Յ 0.05 was considered significant. Changes in relative protein levels and electrical parameters were fit using least squares nonlinear minimization (SigmaPlot 5.0, San Rafael, CA). Decreases were fit to a single exponential of the form is the time constant, and n is a factor that is a measure of the delay before an increase occurs.

RESULTS AND DISCUSSION
A6 cells are models of the principal cell frequently used for the study of regulated Na ϩ reabsorption. These cells have been specifically used for investigating the contribution and modulation of ENaC during transport. A6 cell signal transduction and electrophysiological properties are believed to be similar to those of mammalian collecting duct principal cells. A6 cells express the epithelial Na ϩ channel in the luminal membrane with the this channel's gating kinetics and number influenced by cortico-steroids (2, 3, 10). In the current study, we used the A6 cell line to investigate the regulation of ENaC subunit levels by protein kinase C. We show directly for the first time with ENaC subunit-specific antibodies that activation of PKC with PMA results in a decrease in ␥ and ␤ but not ␣ENaC protein levels; ␥ENaC decreased prior to ␤ENaC. PKC also decreased Na ϩ reabsorption. Decreases in subunit levels and transport in response to PMA were reduced or eliminated with PKC inhibitors. Our results are consistent with activation of PKC resulting in long term (Ͼ4 h) depression of ␤ and ␥ENaC levels leading, in part, to a sustained (up to 48 h) decrease in Na ϩ reabsorption.
The Western blot analysis in Fig. 1 is consistent with A6 cells expressing all three ENaC subunits: ␣, ␤, and ␥. It is well documented both biochemically and electrophysiologically that A6 cells contain typical, fully functional (heteromultimeric) ENaC (2,3,10). In fact, the cDNAs encoding these subunits have been cloned from A6 cells (20). More importantly for the current study, results in Fig. 1 show that we have antibodies that specifically recognize each of the ENaC subunits.
For Western blot analysis (representative blots shown in Fig. 1), each lane within a gel contained lysate with the same amount of total protein (ϳ80 g). The blots of A and C in Fig. 1 were probed with anti-␣ENaC antibody AB586 and anti-␤ENaC antibody AB592, respectively. The first and third lanes in these blots contained lysate harvested form A6 cells serum and aldosterone starved for Ͼ72 h (Ϫ). The second and fourth lanes (ϩ) had lysate from cells treated with aldosterone for Ͼ72 h. The right two lanes (ϩpep) of A and C were probed with antibody preabsorbed with 0.1 mg/ml of the respective immunogens. The top blot of B was probed with AB586. This blot was stripped and reprobed with anti-␥-xENaC (lower blot of B, where Ϫ and ϩ have the same meaning). The effects of aldosterone on ␣ and ␥ENaC in A6 cells were inconsistent with steroid clearly increasing ␣ENaC in two of four experiments (an increase is shown in A, and no effect is shown in the top blot of B) and increasing ␥ENaC in one of four experiments (refer to bottom blot of B). In contrast, long term exposure to aldosterone (1.5 M; Ն 72 h), as shown by the Western blot of C and summarized in D, consistently and significantly increased ␤ENaC 20.8 Ϯ 5.5-fold (seven of seven experiments). For all experiments, aldosterone-treated cells had significantly more Na ϩ transport. These blots suggest interesting characteristics for ENaC subunits in A6 cells: 1) the relative molecular weight of subunits are ␣ Ͻ ␥ Ͻ ␤ and 2) ␤ and possibly ␣ are often observed as doublets perhaps indicating glycosylation. Both subunits contain putative glycosylation sites and previously have been reported to be glycosylated (5,6,21). Moreover, ␤ but neither ␥ nor ␣ENaC subunit levels correlated well with the actions of aldosterone to increase transport suggesting that ␤ENaC may be limiting for long term steroid-regulated reabsorption across A6 cells. However, these results are preliminary in this regard, and thus, this notion needs to be investigated further prior to making more definitive statements on this topic.
It is well documented in A6 cells that ENaC transcript and protein subunit levels are differentially regulated by aldosterone and other factors (reviewed in Refs. 2, 3, and 8). The results of Fig. 1 are consistent with findings we have published previously (6) showing that ␤ (detected with an antibody distinct from the one used in the current study) but not ␣ and ␥ENaC protein levels are reproducibly increased in response to long term administration of aldosterone. Similarly, J. P. Johnson and colleagues 2 also find in A6 cells that ␤ but not ␣ and ␥ENaC protein is increased by aldosterone. It is unclear why these results differ from those of May et al. (21) in A6 cells and Masilamani et al. (5) in rat collecting duct principal cells showing that only ␣ENaC is consistently increased in response to aldosterone. Nonetheless, our antibodies, as well as those of others, are useful tools for studying the regulation of ENaC subunit protein levels.
Although PKC is known to decrease ENaC activity (reviewed in Refs. 9, 10, and 22), the direct actions of PKC on ENaC subunit levels has not been investigated. In addition, a temporal correlation between PKC effects on ENaC subunit levels and Na ϩ reabsorption has not been established. Thus, we tested the hypothesis that PKC differentially regulates ENaC subunit levels and also temporally correlated effects on subunit levels with changes in transport.
Our findings that ␥ and ␤ but not ␣ENaC subunit levels decrease in response to PKC activation are different than the results of Zentner et al. (7) in which PKC was found only to decrease ␣ENaC transcript levels in this latter study. Also in this latter study, preliminary evidence was provided that PKC effects on ␣ENaC transcript levels translated into decreased subunit protein expression. Thus, even though they have not directly examined subunit protein levels, it seems likely that there is a difference in the response to PMA of salivary gland cells and A6 cells. There are many possible explanations of the difference. The simplest would be that the complement of transcription factors present in salivary cells and A6 cells is different and that, therefore, regulation of gene expression is different. A second simple possibility is that the signaling pathways activated by PMA are different in the two cell types so that the relative activation of different transcription factors would be different. Interestingly, the time course of PKC action on ENaC subunits agrees with the results of Zentner et al. (7). Nonetheless, more research is necessary to understand the regulation of ENaC subunits in different epithelia. addition. PKC levels remained substantially depressed throughout the course of these experiments (n ϭ 6). In contrast, 4␣-PMA had no effect on PKC levels between 0 -96 h (n ϭ 4; not shown). These findings show that PMA addition to A6 cells does in fact activate PKC because phorbol activation of PKC is well known to cause self-inhibition via a cellular negative feedback loop resulting in decreased kinase expression.
The effects of PMA on ENaC subunit levels (A) and calculated current (B), and trans-monolayer resistances (C) and voltages (D) are summarized in Fig. 4. The mean Ϯ S.E. (n Ն 3) describing the relative (compared with time 0) density of each subunit at the indicated times are shown with the short dashed (q), long-dashed (E), and solid (ƒ) regression lines fitting ␣, ␤, and ␥ENaC, respectively. ␥ENaC begins to decrease soon after PMA addition with a time constant of 3.7 Ϯ 1.0 h. The decrease in ␤ENaC levels is described by a time constant of 13.9 Ϯ 3.0 h. The change in ␣ENaC over 96 h was not significantly different from that at time 0 (p Ͻ 0.01). 4␣-PMA had no effect on ␣, ␤, and ␥ENaC levels at any time point (times detected ϭ 2, 6, 8, 12, and 24 h; not shown). The decrease in both ␥ and ␤ENaC were fit best by a single exponential, suggesting that PMA had only one major effect on ENaC protein levels. Currently, it is unclear whether the action of PMA results in a decreased subunit expression or increased subunit degradation. In the continued presence of PMA, presumably because of self-inhibition of PKC, ␤ and ␥ENaC subunits levels began to return to pretreatment levels after 24 -48 h (with time constants of 32 Ϯ 8.1 and 52 Ϯ 10 h for ␤ and ␥ENaC, respectively, which are not statistically different from one another; p ϭ 0.128). The levels of ␥ and ␤ENaC at 96 h were significantly greater then the levels of these subunits at 24 h. Importantly, the increase toward pretreatment levels for ␥ and ␤ENaC after 48 h demonstrate the reversibility of PKC actions on subunit levels. It is unclear why there is an extended lag time after PKC selfinhibition prior to ␤ and ␥ENaC levels returning toward pretreatment values. This observation needs to be more carefully examined but suggests some prolonged suppression of ENaC levels that is first triggered by PKC and then sustained after kinase down-regulation. Alternatively, the results imply that ENaC synthesis is quite slow.
The summary graph of relative (to time 0) Na ϩ current in Fig. 4B shows that PMA (100 ng/ml, Ⅺ) but not 4␣-PMA (100 ng/ml, ƒ), or 0.03% Me 2 SO (control, E) decreases transport within 30 min (n Ͼ 6). Reabsorption was fully suppressed by 1 h and remained suppressed for 24 h with a subsequent, gradual return toward normal values. PMA data were fit with an exponential decreasing function with a time constant of ϳ28 min. PMA (q) but not 4␣-PMA (E) in the same solvent caused a decrease in trans-monolayer resistances (C) and voltages (D). Similar to current, the decrease in voltage in response to PMA was immediate. In contrast, the major decrease in resistance began 8 -12 h and peaked 24 h after PMA treatment. After 48 h, monolayer voltages and resistances increased toward pretreatment levels. The decrease in ␥ and ␤ENaC as well as Na ϩ reabsorption all substantially preceded the decrease in monolayer resistance and were completely reversible. In fact, the resistance of 3.2 Ϯ 0.1 K⍀ (n ϭ 20) in the presence of PMA at 8 h (a time point where ␥ and ␤ENaC, as well as transport were suppressed) was not different than the 3.5 Ϯ 0.1 and 3.3 Ϯ 0.1 K⍀ for control (n ϭ 10) and 4␣-PMA (n ϭ 10), respectively. The fact that changes in ENaC subunit levels and transport preceded those in resistance suggest that these earlier parameters change independent of resistance. The cause of the resistance decrease after 10 h is currently unclear. Simple measurements of transepithelial current cannot distinguish between a PMA-dependent change in the paracellular versus transcellular resistance. However, we believe that the decrease in monolayer resistance resulting from PMA treatment did not simply arise from a decrease in total protein expression or cell viability, because ␣ENaC, as well as extracellular-regulated kinase and phosphitidylinositol 3-kinase (not shown; n ϭ 3) levels were not effected by PMA treatment. It is interesting that the PMA-reduced Na ϩ transport began to return to pretreatment levels only after ␤ and ␥ENaC levels began to return to pretreatment values (Ͼ48 h after PMA treatment). This suggests that PKC decreases Na ϩ reabsorption at later time points via a decrease in ENaC levels, which are independent and possibly in addition to actions on ENaC open probability. It is reasonable to presume that the increase in both ENaC levels and transport result from self-inactivation of PKC.
It is unclear from the results of Shimkets et al. (17) showing Our results do not distinguish between either direct or indirect actions but do support a decrease in ENaC number after 3 h resulting from either a decrease in subunit synthesis or an increase in subunit degradation. It is interesting that May et al. (21) showed the half-lives of ENaC subunit proteins to be about 1 h. If PKC inhibits ENaC subunit synthesis, as suggested by Zentner et al. (7), normal channel turnover in the face of suppressed synthesis could account for the decrease in both ␥ and ␤ENaC. Thus, the current results, as well as those of others, support the hypothesis that the initial early actions of PKC on ENaC are to decrease sodium transport by decreasing open probability, and then, subsequently, PKC decreases sodium transport by reducing the number of ENaC subunits capable of forming channels. In other words, PKC appears to have a biphasic action to decrease ENaC transport activity measured as the product of the number of channels and the open probability (NP o ), an early phase involving a decrease in ENaC activity resulting from a change in P o , and a later phase of action involving a decrease in the total cellular amount of ENaC (which alters N, the functional ENaC in the luminal membrane). Fig. 5 (A and B) shows that the PKC inhibitor GF 109203X attenuates the effects of PMA on ␤ and ␥ENaC levels and Na ϩ reabsorption, respectively. These results are consistent with PMA-dependent activation of PKC resulting in both effects. GF 109203X (200 nM) was added simultaneously with PMA (100 ng/ml) to A6 cell monolayers with current and lysate being assessed after 24 h. As shown in the Western blot of Fig. 5A, both ␥ and ␤ENaC were clearly present in cells treated with PMA plus GF 109203X but not PMA alone (n ϭ 3). GF 109203X alone had no effect (n ϭ 3, not shown). Relative transport (compared with time 0) was also protected by GF 109203X with 1.0 Ϯ 0.1, 0.05 Ϯ 0.02, and 0.3 Ϯ 0.02 for control, PMA, and PMA plus GF 109203X, respectively (n ϭ 6). It is interesting that GF 109203X blocked all effects of PMA on ENaC subunit levels but only attenuated PMA effects on transport by ϳ30%. This observation is consistent with the idea that PMA affects both channel number and open probability at 24 h and that the effective dose of GF 109203X on these two parameters is different. Alternatively, the effective dose of GF 109203X to counter PKC inhibition of other transport proteins, such as the serosal Na ϩ /K ϩ -ATPase, may be different, or PMA may have actions on transport independent of PKC. Shown in Fig. 5  (C and D) are the dose-dependent actions of the PKC inhibitor calphostin C on ␥ENaC levels (n ϭ 3) and transport (n ϭ 6), A, temporal effects of PMA (100 ng/ml; 162 nM) on ENaC subunit levels (␣ ϭ q, ␤ ϭ E, and ␥ ϭ ƒ). The data for ␣ (short dashed line), ␤ (long dashed line), and ␥ENaC (solid line) were fit to exponential functions as described in methods. B, temporal effects of active (Ⅺ) and inactive (ƒ) phorbol on Na ϩ transport (control is E). Only the data describing PMA action are fit with an exponential decay. Note that current begins to recover in the continued presence of PMA after 48 h. C, summary of the temporal effects of PMA (q) and 4␣-PMA (E) on trans-monolayer resistances. D, summary of the temporal effects of PMA (q) and 4␣-PMA (E) on trans-monolayer voltages. The data for PMA were fit with an exponential decay. respectively, in the presence of PMA. Calphostin C alone had no effect on either parameter (not shown, n ϭ 3). PMA and calphostin C were added simultaneously with transport and subunit levels being assessed 24 h later. For the relative current graph (Fig. 5D) the actions of calphostin C on PMA depressed transport (at 24 h) were normalized (against time 0) and then standardized to control (not shown). The action of calphostin C to reverse PMA-dependent suppression of ␤ and ␥ENaC levels was similar to its action on PMA-depressed transport. This finding is different from that of GF 109203X, suggesting that calphostin C inhibits PKC actions on open probability and number with a similar effective dose.
It is interesting that PKC inhibitors blocked both the early and latter actions of PMA on transport. The early actions are presumably manifested by changes in channel kinetics or number of channels in the luminal membrane with little dependence on total cellular ENaC levels (13)(14)(15). That we see a PKC-dependent decrease in ENaC subunit levels beginning 3 h after PMA addition strongly suggests that later actions of PMA on transport (Ͼ3 h) must result from a decrease in total cellular ENaC levels. Because this is the first biochemical characterization of PMA actions on ENaC subunit levels and no electrophysiological study has characterized the long term effects of PMA on ENaC kinetics, we are unable to definitively state the contributions or the temporal importance of either PKC-dependent changes in ENaC kinetics or levels to Na ϩ transport. However, our results are consistent with the hypothesis that PKC decreases ␥ and ␤ENaC subunit levels, which leads, in part, to a long term decrease in Na ϩ reabsorption.