Protein kinase C isoform antagonism controls BNaC2 (ASIC1) function.

We explored the involvement of protein kinase C (PKC) and its isoforms in the regulation of BNaC2. Reverse transcriptase PCR evaluation of PKC isoform expression at the level of mRNA revealed the presence of alpha and epsilon/epsilon' in all glioma cell lines analyzed; most, but not all cell lines expressed delta and zeta. No messages were found for the betaI and betaII isotypes of PKC in the tumor cells. Normal astrocytes expressed beta but not gamma. The essential features of these results were confirmed at the protein level by Western analysis. This disproportionate pattern of PKC isoform expression in glioma cell lines was further echoed in the functional effects of these PKC isoforms on BNaC2 activity in bilayers. PKC holoenzyme or the combination of PKCbetaI and PKCbetaII isoforms inhibited BNaC2. Neither PKCepsilon nor PKCzeta or their combination had any effect on BNaC2 activity in bilayers. The inhibitory effect of the PKCbetaI and PKCbetaII mixture on BNaC2 activity was abolished by a 5-fold excess of a PKCepsilon and PKCzeta combination. PKC holoenzymes, PKCbetaI, PKCbetaII, PKCdelta, PKCepsilon, and PKCzeta phosphorylated BNaC2 in vitro. In patch clamp experiments, the combination of PKCbetaI and PKCbetaII inhibited the basally activated inward Na(+) conductance. The variable expression of the PKC isotypes and their functional antagonism in regulating BNaC2 activity support the idea that the participation of multiple PKC isotypes contributes to the overall activity of BNaC2.

The recent molecular identification of a class of proton-sensitive ion channels (ASIC; acid-sensitive ion channel (1,2); also called BNC and BNaC; brain Na ϩ channel (3,4)) belonging to the degenerin (DEG)/ENaC superfamily of ion channels (5) added a new molecular entity to the already complicated field of nociception. Even though their participation in nociception is controversial, ASICs might underlie some properties of native proton-induced currents (6 -8) and could contribute to the function of nociceptive transduction with many other key constituents including nociceptor-specific voltage-gated Na ϩ channels, ATP-gated channels, and capsaicin receptors (9 -13).
The tissue distribution of the ASIC members is not limited to the nervous system but also includes many other tissues such as the lung, testis, and intestine (20,21,26,27). Sensory neuron-specific expression of DRASIC has been reported (18), but Chen et al. (16) have found low level transcripts in superior cervical ganglia, spinal cord, and brain stem. ASICs have been characterized extensively in heterologous expression systems, and besides being implicated in nociception (1,2), a role in mechanotransduction (28,29), in the cellular response to an ischemic offense (27,30,31), and synaptic plasticity (32,33), have been proposed.
Despite significant characterization of ASICs, the possible role of second messenger regulation of ASICs has not been reported. Moreover, Bubien et al. (34) reported an amiloridesensitive Na ϩ current in malignant brain tumor cells and the presence of BNaC2 message in these cells. Also, human glioma cells show a differential expression of specific PKC 1 isoforms compared with normal astroglia (35). With this in mind, we explored the role of PKC and its isoforms in the regulation of BNaC2. We found 1) expression of PKC␣, PKC⑀, PKC␦, and PKC in most cell lines and no expression of PKC␤ in all glioma cell lines compared with normal astrocytes; 2) separately, PKC␤I and PKC␤II lacked a channel inhibitory effect, but in combination PKC␤I and PKC␤II inhibited channel activity in bilayers, which was comparable with the inhibitory effect of whole PKC; 3) PKC⑀ and PKC individually and in combination did not inhibit BNaC2, but a 5-fold excess of a PKC⑀ and PKC combination abolished the otherwise inhibitory influence of the PKC␤I and PKC␤II mixture; 4) whole PKC, PKC␤I, PKC␤II, PKC␦, PKC⑀, and PKC phosphorylated BNaC2 in vitro; 5) PKC␤I plus PKC␤II inhibited inward Na ϩ currents in human U87-MG glioma cells. Our findings of disproportionate expression of PKC isotypes in glioma cell lines and their antagonism with respect to influencing BNaC2 activity in bilayers suggest that different proportions of PKC isoforms differentially regulate BNaC2 activity. Also, dysregulation of BNaC2 resulting * This work was supported by National Institutes of Health Grants DK37206 and CA71933 and the Brain Tumor Foundation for Children. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ from an altered expression of the PKC isoforms could be responsible for an activated amiloride-sensitive Na ϩ current seen in glioma cells (34).

EXPERIMENTAL PROCEDURES
Phospholipids were purchased from Avanti Polar Lipids (Alabaster, AL). PKC, PKC isoforms, and PKC inhibitor peptide 19 -31 were purchased from Calbiochem. All other chemicals were reagent grade, and all solutions were made with distilled water and filter-sterilized before use (Sterivex-GS, 0.22 m filter; Millipore Corp., Bedford, MA).
RT-PCR Detection of PKC Isozyme mRNAs-Total RNA was isolated from human glioma cells and normal astrocytes using a modification of the method of Chomczynski and Sacchi (36). The integrity of the RNA was verified after electrophoresis through 1% agarose-formaldehyde denaturing gels. One-step RT-PCR was performed to detect PKC isozyme mRNA with a Qiagen OneStep RT-PCR kit. Total reaction mixture was 50 l, containing 0.2 g RNA, 0.4 mM of each dNTP, 30 M of forward and reverse primer, and appropriate OneStep RT-PCR enzyme mix and buffer. RT-PCR was carried out beginning with a single cycle of 50°C for 30 min (reverse transcription), 95°C for 15 min (initial PCR activation step), followed by cycles of 94°C for 1 min, 56°C for 1 min, and 72°C for 1 min, for a total of 35 cycles. This was followed by a single cycle of 72°C for 10 min to facilitate final extension. The primers utilized are listed in Table I. Primers were synthesized by Invitrogen. All primer sequences were searched in GenBank TM , and no similarities between primers and other human gene sequences were found except for the target gene we intended to amplify. To confirm this, we used full-length cDNAs (ATCC) for human PKC isoform ␣, ␤, ␥, ⑀, and as substrates in PCR reactions with each of the primer pairs. PCR products were analyzed by agarose gel electrophoresis and visualized by ethidium bromide staining. The primers we designed for each isoform generated amplicons only from the appropriate PKC isoform cDNA template and not from the other templates. Authenticity of each product was confirmed by size and digestion with three restriction enzymes, as well as by direct sequencing. Computer analysis of nucleotide and restriction enzyme mapping was done using the Genetics Computer Group Package (37) on a Unix computer and were provided through the University of Alabama at Birmingham Center for AIDS Research. Human lymphocyte RNA preps (volunteer donors) were used as positive controls (data not shown; see Ref. 13).
Western Blot-The protocol used for Western analyses for different PKC isoforms in glial tumor cell lines was identical to that described earlier (38). The antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).
In Vitro Phosphorylation by PKC and Its Isoforms-In vitro phosphorylation by PKC and its isoforms was assayed by measuring the incorporation of [ 32 P] into immunopurified protein (BNaC2) from [␥-32 P]ATP, in a reaction mixture containing 20 mM Tris⅐HCl, pH 7.5, 10 mM MgCl 2 , 20 M ATP, 15-50 kBq of [␥-32 P]ATP, and 1 microunit of PKC or its isoforms. Immunopurification of in vitro translated protein was performed as described previously (38). The incubation was carried out for 3 min at 30°C, and the phosphorylated proteins were separated by SDS/PAGE and visualized by autoradiography. Where indicated, PKC activity was measured in the presence of 0.5 mM CaCl 2 or 1 M PKC inhibitor (peptide 19 -31).
Planar Lipid Bilayers-Oocyte membrane vesicles were fused with planar lipid bilayers made of a 2:1 (w/w) diphytanoyl-phosphatidylethanolamine/diphytanoyl-phosphatidylserine solution in n-octane (final lipid concentration 25 mg/ml). Bilayers were bathed with symmetrical 100 mM NaCl, 10 mM MOPS-Tris, 100 M EGTA, 50 nM [Ca 2ϩ ] free , pH 7.4. The Bound-and-Determined computer program was used to calculate the level of free [Ca 2ϩ ] (41). Phosphorylation mixture contained 10 ng/ml of PKC or its isoforms, 5 M diacylglycerol, 100 M Mg-ATP. To verify the orientation of BNaC2 and its block by amiloride, at the end of each experiment 5 M amiloride was added to the presumptive extracellular side of the channel. Single channel currents were measured using a conventional current-to-voltage converter with a 10 gigaohm feedback resistor (Eltec, Daytona Beach, FL) as described previously (42). Single channel analyses were performed using pCLAMP 5.6 software (Axon Instruments, Burlingame, CA) on current records low pass-filtered at 300 Hz through an 8-pole Bessel filter (902 LPF; Frequency Devices, Haverhill, MA) prior to acquisition using a Digidata 1200 interface (Axon Instruments, Burlingame, CA).
Whole-cell Patch Clamp-Whole-cell patch clamp experiments were performed on cultured human U87-MG glioma cells as described previously (34). PKC isoforms were included in the pipette solution at a final concentration of 5 ng/ml.

RT-PCR and Western Blot Detection of PKC Isozymes-RT-
PCR using specific primer pairs for PKC␣, PKC␤, PKC␦, PKC⑀, PKC␥, and PKC (Table I) was performed on total RNA isolated from SK-MG1 glioma cells (Fig. 1A, top). A similar analysis was carried out for primary cultures of human astrocytes, three first passage cultures of glioblastoma multiform (GBM) tumor resections (PT1, PT2, and PT3) and ten established cell lines, nine of which were originally derived from GBMs and one (D32GS) from a gliosarcoma (Fig. 1A). These experiments revealed that PKC␤ mRNA was expressed by normal astrocytes but not by any of the tumor cells. The astrocytes expressed all of the PKC isoforms examined, except for PKC␥. Only PKC␣ and PKC⑀/⑀Ј were detected in all of the gliomas. Likewise, PKC␦ mRNA was expressed in all samples except D32GS. There was more variability in expression of PKC␥ and PKC. These results demonstrated that PKC␣, PKC⑀/⑀Ј, and PKC␦ were expressed reliably in all of the GBM glioma cell lines examined and that PKC␤I and PKC␤II were not expressed at all (the primer pairs used for the detection of PKC␤ spanned the common region of PKC␤I and PKC␤II). Western blot analysis was also performed to examine protein expression of PKC␤I, PKC⑀/⑀Ј, and PKC in the astrocytes and SK-MG1 and U87-MG cells (Fig. 1B). Similar to the results presented in Fig.  1A, only astrocytes expressed PKC␤I, all three cell types expressed PKC⑀/⑀Ј, and the astrocytes and SK-MG1, but not U87-MG, expressed PKC. Effects of PKC and Its Isoforms on BNaC2 in Bilayers-BNaC2 incorporated into planar lipid bilayers forms a functional amiloride-sensitive Na ϩ channel with a very low probability of being in an open state (P O ϳ0.08) (43). However, buffering [Ca 2ϩ ] free in the bilayer bathing solution to Ͻ100 nM significantly increases P O and thus provides the opportunity to investigate the effects of PKC and its isoforms on a wild-type active channel. Addition of the phosphorylation mixture (with holoenzyme PKC) to the bilayer bathing solution decreased BNaC2 P O from 0.89 Ϯ 0.09 to 0.45 Ϯ 0.06 without any effect on single channel conductance ( Fig. 2A). We next tested the hypothesis that specific PKC isoforms could affect BNaC2 activity in different ways. The rationale for this set of experiments was the following. First, PKC is a large family of related proteins with at least 11 isotypes, each with a distinctive primary structure, expression pattern, and subcellular localization (44). Second, several groups (35,45) have shown that PKC⑀/⑀Ј and PKC are overexpressed in many glial tumor cell lines, whereas PKC␤ is reduced or even absent compared with normal human astrocytes (Fig. 1A). Third, amiloride-sensitive Na ϩ currents were observed in primary cultures of freshly resected tumors and established glioma cell lines (34) along with BNaC mRNA. We began by examining PKC␤I and PKC␤II effects on BNaC2 activity in bilayers. When added alone, neither PKC␤I nor PKC␤II had any effect on BNaC2 activity (Table II) Holding potential was ϩ100 mV referred to the virtually grounded trans-chamber. Dotted lines indicate zero current. Phosphorylation mixture (10 ng/ml of PKC, 5 M diacylglycerol, 100 M Mg-ATP) was added to the presumptive intracellular side of the channel. Records shown were filtered at 100 Hz with an 8-pole Bessel filter prior to acquisition at 1 ms per point using pCLAMP software (Axon Instruments, Burlingame, CA). Traces shown are representative of at least five experiments. B, inhibitory effect of PKC␤I and PKC␤II combination on BNaC2 incorporated into planar lipid bilayer. Recording and acquisition conditions were the same as for A. Phosphorylation mixture contained, instead of whole PKC, a combination of PKC␤I (5 ng/ml) and PKC␤II (5 ng/ml). Traces shown are representative of at least five experiments. C, lack of inhibitory effect of PKC␤II and PKC␤II combination on BNaC2 incorporated into planar lipid bilayer in the presence of 5-fold excess of PKC⑀ and PKC combination. Recording and acquisition conditions were the same as for A. Phosphorylation mixture contained, instead of whole PKC, a combination of PKC⑀ (25 ng/ml) and PKC (25 ng/ml). Traces shown are representative of at least five experiments. observed (Fig. 2B). This inhibition of BNaC2 activity by the PKC␤I and PKC␤II combination was equivalent to that of whole PKC. These results support the hypothesis that PKC␤I and PKC␤II are essential, at least in planar lipid bilayers, for the inhibitory effect of PKC on BNaC2 activity. Because of the reported up-regulated levels of PKC⑀ and PKC isotypes, and the expectation that these isoforms could have their own effects on BNaC2 activity, we explored the effect of PKC⑀ and PKC on BNaC2 activity. We found that PKC⑀ and PKC added alone or in combination, did not have any effect on BNaC2 activity (Table II). This outcome prompted us to imitate the differential levels of PKC isoform expression in gliomas in our bilayer experiments. A 5-fold excess of PKC⑀ and PKC relative to PKC␤I and PKC␤II was added to the bilayer bathing solution following incorporation of BNaC2. This maneuver abolished the otherwise inhibitory effect of the PKC␤I and PKC␤II combination on BNaC2 activity in bilayers (Fig. 2C). We did not observe this effect with a 1:1 ratio of isoforms (Table II). Also, a 5-fold excess of PKC␦ or its combination with PKC⑀ or PKC neither affected BNaC2 activity nor interfered with the inhibitory influence of the PKC␤I and PKC␤II combination on BNaC2 activity (Table II).
In Vitro Phosphorylation of BNaC2 by PKC and Its Isoforms-Because of the functional effects of PKC and its isoforms on BNaC2 activity in planar bilayers, we hypothesized that BNaC2 should be a substrate for phosphorylation by this kinase and its isoforms. As illustrated in Fig. 3A, BNaC2 can indeed be specifically phosphorylated by PKC (lanes 1 and 2). Elimination of BNaC2 or inclusion of a PKC peptide inhibitor to the reaction mixture prevented phosphorylation of BNaC2 (lanes 3 and 4, respectively). Similarly, PKC isoforms (␤I, ␤II, ␦, ⑀, and ) phosphorylated BNaC2 (Fig. 3B). Elimination of BNaC2 protein from the reaction mixture prevented BNaC2 phosphorylation (data not shown).
The Effects of the PKC Isoform on Inward Na ϩ Currents in U87-MG Glioma Cells-Whole-cell patch clamp experiments were performed on cultured human U87-MG glioma cells to test the hypothesis that PKC␤I ϩ PKC␤II can inhibit the constitutively activated inward Na ϩ currents seen in these cells. As a prelude, Fig. 4 presents representative whole-cell patch clamp records from a U87-MG cell before and after treatment with amiloride. Amiloride effectively blocked inward currents. Fig. 5 represents the patch clamp results of the PKC experiments. Inclusion of 5 ng/ml PKC␤II in the pipette solution had no effect on the inward currents (middle panel). In contrast, PKC␤I ϩ PKC␤II abolished the inward currents (right panel). As an additional control, PKC also was without effect (data not shown), consistent with the bilayer findings (Table II).

Phosphorylation of BNaC2 by PKC and Its Isoforms-
The effects of PKC phosphorylation on ion channel function can either be stimulatory or inhibitory depending upon the type of ion channel and cell type (46 -51). The effects of PKC on amiloride-sensitive Na ϩ transporting pathways were even more complex and diverse. Activation of PKC greatly diminished Na ϩ transport in A6 cells (52,53) and in the LLC-PK1 epithelial cell line (54). PKC also attenuated the activity of purified renal amiloride-sensitive sodium channels (55). Contrary to these outcomes, PKC activation had no effect on Na ϩ transport in the rat cortical collecting duct (56) and even stimulated sodium transport across frog skin epithelium (57). These discrepancies may arise from reported biphasic effects of PKC on Na ϩ currents (58) when an initial increase was followed by inhibition of current. These differences may also reflect specific or nonspecific effects of PKC on Na ϩ transport (59). The inhibitory influence of PKC and its isoforms on BNaC2 activity (Fig.  2) was reminiscent of its effect on cloned ENaC, another degenerin (DEG)/ENaC member, where PKC treatment inhibited single sodium channel activity either in bilayers or following heterologous expression in Xenopus oocytes (60). Our bilayer findings favor the possibility of a direct effect of PKC and its isoforms on BNaC2. This is also supported by the fact that BNaC2 was subject to in vitro phosphorylation by PKC and its individual isoforms (Fig. 3) and by findings of Shimkets et al. (61) showing that ␤ and ␥, but not the ␣-ENaC can be phosphorylated in vivo by PKC. Moreover, Stockand et al. (62) showed that PKC inhibition of ENaC may involve differential regulation of subunit levels. PKC activation decreased protein levels of ␤ and ␥ but not ␣-ENaC. The conclusion of direct interaction of PKC and BNaC2 does not completely rule out the participation of other components in this interaction. Recent reports (63)(64)(65) demonstrated an interaction between members of ASIC family and the PDZ domain-containing proteins, which also interact with PKC. These protein-protein interactions may serve to coalesce the intracellular components necessary for PKC regulation of ASICs. Because of reported PKC isoform regulation of ion transport (66 -70), phosphorylation of BNaC2 by different PKC isoforms is of particular interest. First, in in vitro experiments PKC␤I and PKC␤II were able to phosphorylate BNaC2 individually, whereas only their combination was able to produce a functional inhibitory effect in bilayers. Moreover, this inhibitory effect was evident in the absence or the presence of an equal ratio of PKC⑀, and PKC isoforms. An excess of a PKC⑀ and PKC combination, compared with PKC␤I and PKC␤II, effectively abolished the inhibitory influence of the latter on BNaC2 activity. Also, addition of PKC␤I FIG. 4. Representative whole-cell patch clamp recordings from a single U87-MG cell: effect of amiloride. Cells were voltageclamped between Ϫ160 and ϩ 100 mV in 20-mV increments from a holding potential of Ϫ60 mV. Cells were superfused with RPMI 1640 medium, and the pipette contained the following (in mM): 100 potassium gluconate, 30 KCl, 10 NaCl, 20 HEPES, 0.5 EGTA, 4 ATP, and Ͻ10 nM free Ca 2ϩ , pH 7.2. After a basal recording (top), amiloride (100 M) was superfused over the same cell, and another set of voltage-clamp records were obtained (middle). Amiloride effectively abolished inward currents, as seen in the difference curve (bottom). This experiment was repeated four times with identical results.
FIG. 5. Whole-cell patch clamp recordings from representative U87-MG cells. PKC isoforms were included in the pipette solution at a final concentration of 5 ng/ml. The chord conductance measured between Ϫ80 mV, and the reversal potential was (in pS) 5250 Ϯ 1700 (basal), 7318 Ϯ 2316 (PKC␤II), and 2916 Ϯ 987 (PKC␤I ϩ ␤II); n ϭ 4 for each. The mean conductance value for the PKC␤I ϩ ␤II group was significantly different from the PKC␤I (p Ͻ 0.005), the basal (p Ͻ 0.01), or the PKC-(5350 Ϯ 1891 pS; p Ͻ 0.01, not shown) groups. and PKC␤II to the cytosolic compartment of U87-MG cells effectively abolished inward Na ϩ currents. These results suggest, at least functionally, that at least two isoforms (␤I and ␤II) are required for phosphorylation to observe a functional effect. Second, although PKC⑀ and PKC individually phosphorylated BNaC2 in vitro, they were unable to produce functional effects on BNaC2 in bilayers individually or in combination. Their ability to phosphorylate BNaC2 in vitro was materialized functionally by abolishing the otherwise inhibitory effects of PKC␤I and PKC␤II. The nature of molecular events underlying PKC⑀ and PKC effects in preventing BNaC2 phosphorylation by PKC␤I and PKC␤II is puzzling, but it is possible that phosphorylation by PKC⑀ and PKC alters the BNaC2 protein in a such way that the phosphorylation site(s) of PKC␤I and PKC␤II becomes inaccessible as a result of the involvement of the same or completely different phosphorylation site(s). These observations raise the possibility that the differential expression of PKC isoforms acts as one of many determining factors of the end effect of PKC.
Effects of PKC and Its Isoforms and Glioma Cell Biology-PKC plays an important role in glioma cell biology. Malignant glioma cells have been reported to express two to three orders of magnitude higher total PKC activity than normal astrocytes (45). Non-isoform-specific PKC inhibitors such as calphostin C and staurosporine (71) can block proliferation of glioma cell lines. Moreover, elevation of PKC␣ levels in U87-MG cells increased invasiveness relative to control cells (72); this effect may be related to matrix metalloproteinases (73). Interestingly, inhibition of PKC␣ and PKC⑀ had no effect on glial cell proliferation, but PKC inhibition blocked proliferation (71). In another study, overexpression of PKC␣ in U87-MG cells increased the proliferation rate (74). Expression of a dominant negative PKC⑀/⑀Ј mutant in U373-MG astroglial cells also inhibited proliferation (75). Thus, it appears that there are many PKC isoform-specific effects on glial cells, and these effects may extend to an amiloride-sensitive Na ϩ current in malignant brain tumor cells, attributed to BNaC2 (34). In patch clamp experiments, only a combination of PKC␤I and PKC␤II in the pipette solution inhibited the inward currents (Fig. 5). The presence of only PKC␤II or PKC in the pipette solution had no effect on the inward currents. The same pattern of effects were observed in bilayer experiments: lack of any effect by an individual PKC isoforms, but presence of inhibitory effect of PKC␤I and PKC␤II combination. Our RT-PCR results (Fig. 1) demonstrated that only PKC␣, ⑀/⑀Ј, and were expressed in most all of the glioma cell lines that were examined and that PKC␤ was not expressed at all. Thus, we confirmed that PKC⑀ and PKC are present in gliomas (35,45) and show for the first time that high grade gliomas do not express PKC␤, as compared with normal astrocytes.
Our experiments show that total PKC inhibited BNaC2 activity in planar lipid bilayers. Also, we found that an overabundance of specific PKC isoforms, specifically ⑀ and in the face of reduced PKC␤, overcame any inhibitory effects of the other PKC isoforms on BNaC2 activity. If expression of BNaC2, and for that matter other members of the ASIC family, is specific to high grade glioma cells, then it is likely that PKC isoform-specific modulation of ASICs contribute to its constitutive activity in glioma cells. This effect may be important in conferring tumor-like characteristics, such as invasiveness and proliferation, to these cells.