Phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) and phosphatidylinositol 3,4,5-trisphosphate (PI(3,4,5)P3) are physiologically important second messengers. These molecules bind effector proteins to modulate activity. Several types of ion channels, including the epithelial Na+ channel (ENaC), are phosphoinositide effectors capable of directly interacting with these signaling molecules. Little, however, is known of the regions within ENaC and other ion channels important to phosphoinositide binding and modulation. Moreover, the molecular mechanism of this regulation, in many instances, remains obscure. Here, we investigate modulation of ENaC by PI(3,4,5)P3 and PI(4,5)P2 to begin identifying the molecular determinants of this regulation. We identify intracellular regions near the inner membrane interface just following the second transmembrane domains in β- and γ- but not α-ENaC as necessary for PI(3,4,5)P2 but not PI(4,5)P2 modulation. Charge neutralization of conserved basic amino acids within these regions demonstrated that these polar residues are critical to phosphoinositide regulation. Single channel analysis, moreover, reveals that the regions just following the second transmembrane domains in β- and γ-ENaC are critical to PI(3,4,5)P3 augmentation of ENaC open probability, thus, defining mechanism. Unexpectedly, intracellular domains within the extreme N terminus of β- and γ-ENaC were identified as being critical to down-regulation of ENaC activity and Po in response to depletion of membrane PI(4,5)P2. These regions of the channel played no identifiable role in a PI(3,4,5)P3 response. Again, conserved positive-charged residues within these domains were particularly important, being necessary for exogenous PI(4,5)P2 to increase open probability. We conclude that β and γ subunits bestow phosphoinositide sensitivity to ENaC with distinct regions of the channel being critical to regulation by PI(3,4,5)P3 and PI(4,5)P2. This argues that these phosphoinositides occupy distinct ligand-binding sites within ENaC to modulate open probability.
Ion channels play a critical role in cellular function and physiology. As such, they serve as important effectors for many intracellular signaling cascades, including those using phosphatidylinositide second messengers. Phosphoinositides regulate channel activity both indirectly through the actions of intermediary proteins and more directly by acting as ligands interacting specifically with intracellular portions of channel effectors (Hilgemann et al., 2001; Ribalet et al., 2005; Pochynyuk et al., 2007b; Voets and Nilius, 2007). Phosphoinositide binding often results in channel activation through a molecular mechanism involving increases in open probability. This is the case for phosphoinositide regulation of KCNJ, KCNQ, and KCNK family K+ channels, TRP channels, epithelial Na+ channel (ENaC), and Cav2 channels (Shyng et al., 2000; Dong et al., 2002; Ma et al., 2002; Wu et al., 2002; Du et al., 2004; Gamper et al., 2004; Tong et al., 2004b; Li et al., 2005; Lopes et al., 2005). Binding and direct channel regulation by phosphoinositides is physiologically important for its disruption can lead to disease. This is true for loss of function mutations in Kir2.1, Kir6.2, and KCNQ1 channels resulting in decreased PI(4,5)P2 affinity/sensitivity leading to Andersen-Tawil, Bartter's, and long QT syndromes, as well as congenital hyperinsulinism (Lopes et al., 2002; Donaldson et al., 2003; Park et al., 2005; Lin et al., 2006; Ma et al., 2007). In several instances, disease-causing mutations modify the basic residues involved in forming electrostatic interactions with the negative-charged head groups of phosphoinositides. While diverse types of channels directly bind and, thus, are sensitive to phosphoinositides, details regarding sites within ion channels involved in this regulation and binding remain obscure. Moreover, the molecular consequences of phosphoinositide binding to ion channels, in many cases, remain conjecture. Here, we continue probing phosphoinositide regulation of ENaC to address some of these questions. ENaC is believed to interact with both PI(4,5)P2 and PI(3,4,5)P3 with direct interactions influencing channel activity (Ma et al., 2002; Yue et al., 2002; Tong et al., 2004b; Kunzelmann et al., 2005; Pochynyuk et al., 2005; Tong and Stockand, 2005).
The epithelial Na+ channel is a nonvoltage-gated, noninactivating, highly Na+-selective channel localized to the luminal plasma membrane of epithelial cells (Garty and Palmer, 1997; Benos and Stanton, 1999; Alvarez de la Rosa et al., 2000; Kellenberger and Schild, 2002). ENaC is common to a number of different epithelial tissues, including those lining the distal renal nephron, distal colon, ducts of exocrine glands, and pulmonary airways and alveolar sacs. ENaC activity is rate limiting for Na+ (re)absorption across these epithelial barriers. Because of this function, ENaC serves as a critical modulator of epithelial hydration, setting osmotic gradients driving fluid movement. Moreover, ENaC in the kidneys and colon is positioned to influence systemic Na+ balance and, thus, blood pressure. Indeed, ENaC is a final effector of the renin-angiotensin–aldosterone system, which is the primary negative feedback pathway governing blood pressure. As such, gain of function mutations in ENaC and its upstream regulatory pathways are causative for several hypertensive diseases associated with improper Na+ retention (Snyder et al., 1995; Abriel et al., 1999; Hummler and Horisberger, 1999; Bonny and Hummler, 2000; Lifton et al., 2001). Conversely, loss of function mutations in ENaC and its upstream regulatory pathways cause renal salt wasting diseases and the inability to clear the neonatal lung of fluid after birth.
The primary systemic regulators of ENaC are corticosteroids, including mineralocorticoids, such as aldosterone, and glucocorticoids (Verrey, 1995; Verrey, 1999; Stockand, 2002). ENaC also responds to endocrine and paracrine signals like insulin and ATP. Both aldosterone and insulin increase ENaC activity through a transduction cascade including phosphatidylinositide 3-OH kinase (PI3-K) and its product phosphoinositides (Record et al., 1998; Blazer-Yost et al., 1999; Wang et al., 2001; Staruschenko et al., 2004b). Activation of ENaC by PI3-K signaling involves, at least in part, direct interactions of PI(3,4,5)P3 with channel subunits (Tong et al., 2004a,b; Pochynyuk et al., 2005). In contrast to aldosterone and insulin, ATP, most likely via PLC linked to a Gq-coupled receptor, decreases PI(4,5)P2 levels in epithelial cells; consequently, decreasing ENaC activity (Ma et al., 2002; Kunzelmann et al., 2005). Again, this regulation involves direct interaction between the phosphoinositide and channel.
ENaC is a heteromeric channel comprised of three distinct but similar subunits: α, β, and γ (Canessa et al., 1993, 1994). The exact subunit stoichiometry of ENaC remains uncertain; however, all three subunits are required for normal channel activity, regulation, and function (Canessa et al., 1994; McNicholas and Canessa, 1997; Fyfe and Canessa, 1998). Each of these three subunits shares a common tertiary structure having intracellular N and C termini separated from a single large extracellular loop by transmembrane domains (Snyder et al., 1994).
While PI(3,4,5)P3 and PI(4,5)P2 may bind to modulate ENaC activity, little is actually know about putative binding sites for these ligands in this channel (Pochynyuk et al., 2007b). This lack of understanding is not unique to ENaC for little is actually known about putative phosphoinositide binding sites in most phosphoinositide-sensitive ion channels. What we do know comes primarily from study of PI(4,5)P2 binding to Kir and TRP channels (Zhang et al., 1999; Shyng et al., 2000; Soom et al., 2001; Cukras et al., 2002a,b; Dong et al., 2002; Lopes et al., 2002, 2005; Prescott and Julius, 2003; Rohacs et al., 2003; Du et al., 2004).
Channel–phosphoinositide interactions are thought to be primarily electrostatic in nature, where the negative-charged head groups of phosphoinositides interact with positive-charged residues within intracellular portions of the channel. In addition, polar and nonpolar residues, which do not directly interact with the charged phosphates, further stabilize this interaction. Importantly, intracellular regions of ion channels just following transmembrane domains, which often are rich in positive-charged residues, are particularly involved in coordinating phosphoinositide binding (Zhang et al., 1999; Shyng et al., 2000; Soom et al., 2001; Dong et al., 2002; Prescott and Julius, 2003; Bian et al., 2004).
To learn more about the molecular mechanism of phosphoinositide regulation of ENaC and to better understand the domains within ion channels important to phosphoinositide regulation and potentially important to binding, we investigated the effects of PI(3,4)P2 and PI(3,4,5)P3 on ENaC. Substitution mutations, resulting in charge neutralization, and deletions were targeted to intracellular residues/domains having characteristics of phosphoinositide bind sites. We find that intracellular regions of β- and γ- but not α-ENaC are necessary for PI(4,5)P2 and PI(3,4,5)P3 regulation. Unexpectedly, different regions of the channel were necessary to regulation by these two phosphoinositides with the extreme N terminus being involved in regulation by the former and regions just following the second transmembrane domains being important to regulation by the latter. Both phosphoinositides affected ENaC open probability and conserved positive-charged residues played key roles in the response.
Materials And Methods
Chemicals and cDNA Constructs
All chemicals were from Sigma-Aldrich, BioMol, or CalBiochem, unless noted otherwise. All materials used in Western blot analysis were from Bio-Rad Laboratories. Sulfo-NHS-LC-biotin and streptavidin-agarose were from Pierce Chemical Co. The mouse monoclonal anti-myc antibody was from CLONTECH Laboratories, Inc. and the rabbit anti-Fra-2 polyclonal antibody was from Santa Cruz Biotechnology, Inc. Anti-mouse and anti-rabbit HRP-conjugated 2° antibodies were from Kirkegaard-Perry Laboratories. ECL reagents were from PerkinElmer Life Sciences. Dioctanoyl (diC8), short-chain phosphatidylinositides were from Echelon Biosciences Inc. The expression vector encoding membrane-targeted PI3-K (pUSE-amp-p110α) was from Upstate Biotechnology. This plasmid encodes p110α fused to an N- terminal Src myristoylation sequence, which localizes PI3-K to the plasma membrane, where it is then constitutively active (Kristof et al., 2003). The mammalian expression vector encoding type I PI(4)P5-kinase α-isoform was a gift from L. Pott (Ruhr University Bochum, Bochum, Germany) and has been described previously (Bender et al., 2002). The PI(4,5)P2 reporter, GFP-PLC-δ-PH, is a chimera consisting of the PI(4,5)P2-binding pleckstrin homology (PH) domain from PLC-δ conjugated to GFP (Haugh et al., 2000). The cDNA encoding this reporter was a gift from the T. Meyer laboratory (Stanford University, Stanford, CA). The mammalian expression vectors encoding α-, β-, and γ-mouse ENaC with N-terminal myc- and HA-epitope tags have been described previously (Staruschenko et al., 2004b; Tong et al., 2004b; Pochynyuk et al., 2005). As noted in these earlier publications, channels comprised of epitope-tagged subunits exhibit functional behavior indistinguishable from those lacking tags. All ENaC mutagenesis was performed on the backbone of plasmids encoding myc-tagged subunits. Mutants were created in our laboratory with QuikChange (Stratagene) mutagenesis per the manufacture's instructions or outsourced to TOP Gene Technologies. Regardless of source, every plasmid encoding an ENaC subunit mutant was sequenced to ensure proper incorporation of the expected mutation and to confirm sequence identity, orientation, and reading frame. We used 12 mENaC mutants in this study: (1) three of α-ENaC, including the R98-K108 (α1D) and R614-R630 (α2D) deletions and the R98A + K103A (α1S) substitution; (2) four of β-ENaC, including the K4-K16 (βND), K39-K49 (β1D), and K552-R563 (β2D) deletions and the K39A + R40A (β1S) substitution; and (3) five of γ-ENaC, including the R42-R53 (γ1D) and Q573-R583 (γ2D) deletions and the K6A + K8A + K10A + K12A + K13A (γNS), R42A + R43A (γ1S), and R569A + R570A + K574A + K576A + R581A + R582A + R583A (γ2S) substitutions. The relative position of these mutations to the first and second transmembrane domains in ENaC subunits are shown in Fig. 1 A.
Cell Lines and Tissue Culture
CHO cells were from American Type Culture Collection. These cells were maintained with standard culture conditions (DMEM + 10% FBS, 37°C, 5% CO2) as described previously (Staruschenko et al., 2004b; Tong et al., 2004b; Pochynyuk et al., 2005). Immortalized mouse cortical collecting duct (mpkCCDc14) principal cells were grown in defined medium on permeable supports (Costar Transwells; 0.4 μm pore, 24 mm diameter) as described previously (Bens et al., 1999). Cells were maintained with FBS and corticosteroids allowing them to polarize and form monolayers with high resistances and avid Na+ reabsorption.
Expression of Exogenous Protein
Recombinant ENaC and phospholipid kinases were overexpressed in CHO cells by transfecting the appropriate expression plasmids using the Polyfect reagent (QIAGEN) as described previously (Staruschenko et al., 2004b; Tong et al., 2004b; Pochynyuk et al., 2005). For electrophysiology studies, 0.3 μg/subunit/9.6 cm2 plasmid cDNA was used. Plasmid cDNAs encoding phospholipid kinases were added at 1.0 μg/9.6 cm2. For membrane labeling studies, 1.0 μg/subunit/314 cm2 plasmid cDNA was used. The plasmid encoding the PI(4,5)P2 reporter, GFP-PLC-δ-PH, was introduced into mpkCCDc14 principal cells within a confluent monolayer with a biolistic particle delivery system (Biolistic PDS-1000/He Particle Delivery System; Bio-Rad Laboratories). Use of this system has been described previously (Yuan et al., 2005; Gamper and Shapiro, 2006). We closely followed established protocols in the current studies. In brief, mpkCCDc14 cells were grown to confluence on permeable supports. After forming high-resistance monolayers avidly transporting Na+, cells were washed twice with physiologic saline, aspirated, and quickly bombarded (at the apical membrane) under vacuum with microcarriers coated with GFP-PLC-δ-PH cDNA. Medium was immediately returned to the cells, which where then placed within a tissue culture incubator for 2–3 d to allow expression of the PI(4,5)P3 reporter. Bombardment had little disruptive effect on cellular and monolayer integrity as established by maintenance of Na+ transport and a high transepithelial resistance.
Total Internal Reflection Fluorescence (TIRF) Microscopy
Fluorescence emissions from the PI(4,5)P2 reporter at the apical membrane of mpkCCDc14 cells within a confluent monolayer were collected using TIRF (also called evanescent-field) microscopy. TIRF generates an evanescent field that declines exponentially with increasing distance from the interface between the cover glass and plasma membrane, illuminating only a thin section (∼100 nm) of the cell in contact with the cover glass (Steyer and Almers, 2001; Axelrod, 2001; Taraska et al., 2003). For these experiments, GFP-PLC-δ-PH was introduced into polarized monolayers of mpkCCDc14 cells grown on permeable supports with the particle delivery system described above. Upon expression of the reporter, 5 × 5-mm sections of the support were excised, inverted, and placed onto cover glass coated with poly-l-lysine. This arrangement made it possible to visualize dynamic changes in the level of the PI(4,5)P2 reporter at the apical membrane in real time in living cells.
All TIRF experiments were performed in the total internal reflection fluorescence microscopy core facility housed within the Department of Physiology at the University of Texas Health Science Center (http://physiology.uthscsa.edu/tirf). We have previously described imaging the GFP-PLC-δ-PH reporter and other fluorophore-tagged proteins using this core facility (Tong et al., 2004b; Staruschenko et al., 2005; Pochynyuk et al., 2006). The methods used in the current study closely followed these published protocols. In brief, fluorescence emissions from GFP-PLC-δ-PH were collected using an inverted TE2000 microscope with through-the-lens (prismless) TIRF imaging (Nikon). Samples were viewed through a plain Apo TIRF 60x oil-immersion, high-resolution (1.45 NA) objective. Fluorescence emissions were collected through a 535 ± 25-nm bandpass filter (Chroma Technology Corp.) by exciting GFP with an Argon-ion laser with an acoustic optic tunable filter (Pairie Technology Inc.) used to restrict excitation wavelength to 488 nm. Fluorescence images were collected and processed with a 16-bit, cooled charge-coupled device camera (Cascade 512F; Roper Scientific Inc.) interfaced to a PC running Metamorph software. This camera uses a front-illuminated EMCCD with on-chip multiplication gain. Images were collected once a minute with a 100-ms exposure time. Images were not binned or filtered with pixel size corresponding to a square of 122 nm × 122 nm.
Membrane Labeling Experiments
Membrane labeling experiments closely followed those described previously (Booth and Stockand, 2003; Staruschenko et al., 2004a,b). In brief, CHO cells were transfected with myc- and HA-tagged ENaC subunits. The subunit of interest contained the myc tag and was followed with anti-myc antibody. 48 h after transfection, cells were washed three times with ice-cold PBS (pH 8.0) and subsequently incubated with 3 mM sulfo-NHS-LC-biotin (in PBS, pH 8.0) for 30 min at room temperature in the dark. Cells were washed again three times with ice-cold PBS plus 100 mM glycine to quench the reaction and then extracted in gentle lysis buffer (76 mM NaCl, 50 mM HCl-Tris, pH 7.4, 2 mM EGTA, plus 1% Nonidet P-40 and 10% glycerol; supplemented with the protease inhibitor 1 mM phenylmethylsulfonyl fluoride). Extracts were cleared and normalized for total protein concentration. Pre-equilibrated streptavidin agarose beads were agitated overnight at 4°C with 200 μg total protein. Agarose beads were then washed six times with gentle lysis buffer and subsequently resuspended in SDS sample buffer and 20 mM DTT, heated at 85°C for 10 min, run on 7.5% polyacrylamide gels in the presence of SDS, transferred to nitrocellulose, and probed with anti-myc antibody in tris-buffered saline supplemented with 5% dried milk and 0.1% Tween-20. Band intensity in developed blots was quantified by scanning the blots and then using SigmaGel (Jandel Scientific) to analyze digital images. We used a flood-above-threshold protocol with threshold set to the highest practical level.
Whole-cell macroscopic current recordings of ENaC reconstituted in CHO cells were made under voltage-clamp conditions using standard methods (Staruschenko et al., 2004b; Tong et al., 2004b; Pochynyuk et al., 2005). Current through ENaC was the inward, amiloride-sensitive Na+ current with a bath solution of (in mM) 160 NaCl, 1 CaCl2, 2 MgCl2, and 10 HEPES (pH 7.4) and a pipette solution of (in mM) 140 CsCl, 5 NaCl, 2 MgCl2, 5 EGTA, 10 HEPES (pH 7.4), 2.0 ATP, and 0.1 GTP. Current recordings were acquired with an Axopatch 200B (Axon Instruments) interfaced via a Digidata 1322A (Axon Instruments) to a PC running the pClamp 9.2 suite of software (Axon Instruments). All currents were filtered at 1 kHz. Voltage ramps (500 ms) from 60 to −100 mV from a holding potential of 40 mV were used to generate current–voltage (I-V) relations and to measure ENaC activity at −80 mV. Whole-cell capacitance was routinely compensated and was ∼9 pF for CHO cells. Series resistances, on average 2–5 MΩ, were also compensated.
For excised, outside-out patches made on CHO cells overexpressing wild-type and mutant ENaC, bath and pipette solutions were (in mM) 160 NaCl, 1 CaCl2, 2 MgCl2, and 10 HEPES (pH 7.4) and 140 CsCl, 5 NaCl, 2 MgCl2, 5 EGTA, 10 HEPES (pH 7.4), 2.0 ATP, and 0.1 GTP (in mM), respectively. Current recordings were made under voltage-clamp conditions using an Axopatch 200B. Currents were low-pass filtered at 100 Hz by an eight-pole Bessel filter (Warner Instruments) and digitized at 500 Hz and stored on a PC using the Digidata 1322A interface. For presentation, some currents were subsequently software filtered at 40 Hz. Gap-free current recordings were made at pipette potentials (Vp) of 0 or −60 mV with inward Na+ current downwards. Current data were analyzed using pClamp 9.2. Channel activity defined as NPo was calculated using the following equation: NPo = Σ(t1 + 2t2 +…iti), where ti is the fractional open time spent at each of the observed current levels. Po was estimated by normalizing NPo for the observed number of channels within a patch as established with all-point histograms. The error associated with this estimation of Po increases as patches contain more channels and as Po approaches either zero or unity (Kemendy et al., 1992). While there is an error associated with estimating Po in seals containing more than one channel, it can be lessened, as done here, by restricting this calculation to patches containing only five channels or fewer. Moreover, the error associated with this measurement was further limited by estimating Po over a span of 35–120 s for each condition. Calculations for channels in excised, inside-out patches were made using the number of channels observed before “rundown.” Single channel current–voltage relations were generated from unitary currents defined by all-point histograms at six different holding potentials from, at least, four independent experiments. Aqueous stocks of water-soluble, short-chain diC8 phosphoinositides were prepared at 1 mM by sonication for 30 min and stored at −70°C. Stock phosphoinositides were mixed just before use with an equal volume of a carrier solution containing histone H1 (0.2 mM; Echelon Biosciences Inc.) and sonicated again for 10 min. Phosphoinositide plus carrier was added to the bathing solution by direct pippetting close to the patched membrane.
For cell-attached patches made on the apical membranes of mpkCCDc14 principal cells, bath and pipette solutions were (in mM) 160 NaCl, 1 CaCl2, 2 MgCl2, and 10 HEPES (pH 7.4); and 140 LiCl, 2 MgCl2, and 10 HEPES (pH 7.4), respectively. Current recordings were made at a −Vp = −60 mV and collected and analyzed as above.
For excised, inside-out patches, bath and pipette solutions were (in mM) 140 CsCl, 0.5 MgCl2, 5 EGTA, and 10 HEPES (pH 7.4); and 140 LiCl, 2 MgCl2, and 10 HEPES (pH 7.4), respectively. Exogenous phosphoinositide was added in the presence of 0.1 mM GTP. Current recordings were made at −Vp = −60 mV, and collected and analyzed as above.
All summarized data reported as mean ± SEM. Membrane levels of ENaC normalized to total cellular pools. Emissions from GFP-PLC-δ-PH normalized to starting levels. Emissions were corrected for a modest, time-dependent photobleaching (<15%/15 min) as established with vehicle treatment. Macroscopic current density in the presence of PI3-K, PI(4)P5-K, and VO4 reported relative to control levels. Summarized data compared with either the Student's t test or a one-way ANOVA in conjunction with the Dunnett post-test where appropriate. P ≤ 0.05 was considered significant.
Characterization of ENaC Subunit Mutants
All three ENaC subunits have cytosolic tracks just preceding and following the first (TM1) and second (TM2) transmembrane domains containing clusters of conserved positive-charged residues. In addition, β- and γ-ENaC have clusters of basic residues in their extreme N terminus. We begin our investigation of phosphoinositide regulation of ENaC by characterizing channels containing charge neutralization and deletion of these conserved positive-charged residues. Fig. 1 A notes the relative position of substituted and deleted residues with respect to the transmembrane domains within ENaC.
Fig. 1 B shows results from experiments confirming that each mutant subunit expresses, produces a peptide of the expected size, and localizes to the plasma membrane when coexpressed with complementary wild-type subunits. Typical results for each mutant are shown in the Western blots to the right with summary data compared in the graph to the left. For these experiments, ENaC in the plasma membrane (M) was separated from total cellular pools (T) of the channel. The inset in Fig. 1 B shows a complete Western blot representative of such experiments. The top blot in this inset contains whole cell lysates (T) and membrane fractions (M) from CHO cells overexpressing all three wild-type ENaC subunits and from those expressing the negative control, eGFP. This blot first was probed with anti-myc antibody to identify α-ENaC (top) and then stripped and subsequently reprobed with anti–Fra-2 antibody (bottom). The prior antibody identifies the channel subunit of interest and the latter confirms good separation of the membrane fraction from total cellular pools for Fra-2 is a cytosolic protein. In every case, mutant ENaC expressed normally, with mutation having no overt effect on the relative membrane levels of the channel.
We continued characterizing mutant ENaC with whole-cell voltage clamp experiments. Results are shown in Fig. 1 C. As reported previously (Booth et al., 2003; Staruschenko et al., 2004b; Tong et al., 2004b), CHO cells have little background current and no amiloride-sensitive current in the absence of overexpression of ENaC. Coexpression of all three wild-type subunits resulted in an activity of 210 ± 26 pA/pF. In contrast, ENaC-containing subunits with mutations in the regions just preceding TM1 were not active, except for that containing a subtle substitution in this region of γ-ENaC. In contrast, channels comprised of subunits having mutations in their N terminus and just after TM2 were active.
While the region just preceding TM1 in all three subunits appears to be critical for normal channel function, the scope of the current study is to identify regions of ENaC important to phosphatidylinositide regulation. In this context, mutations having little effect on basal activity, such as those following TM2 and those in the N termini of β- and γ-ENaC, are more useful compared with those that result in a complete loss of function.
Conserved Positive-charged Residues in the Cytosolic Regions just following TM2 in β- and γ-ENaC Are Necessary for PI3-K and PI(3,4,5)P3-sensitive Increases in Po
We now were positioned to test whether mutant ENaC, as wild-type ENaC, responds to PI3-K stimulation. To do this, mutant subunits were coexpressed with complementary wild-type subunits in the absence and presence of constitutively active PI3-K. Relative increases in activity in the presence of PI3-K measured in whole-cell voltage clamp experiments for wild-type and mutant channels are shown in Fig. 2.
Mutation of the region following TM2 in β- and γ-ENaC subunits completely abolished PI3-K sensitivity. In contrast, channels with α subunits containing mutation of this region and those containing β and γ subunits with N-terminal mutations had normal responses to PI3-K. Similarly, channels containing the only functional mutant of the area just preceding TM1, γ1S, had a robust response to coexpression of PI3-K not different than wild type. These results demonstrate that PI3-K increases ENaC activity in a manner dependent on the regions just following TM2 in β- and γ-ENaC. More specifically, the response to PI3-K depends on the presence of conserved positive-charged residues within these regions. This is the first indication that this region of β-ENaC plays an important role in PI3-K regulation. The observation that this region of γ-ENaC plays a role in the response to PI3-K is consistent with the same finding made previously by our laboratory and our finding that this area of γ-ENaC physically interacts with the phosphatidylinositide products of PI3-K, PI(3,4,5)P3, and PI(3,4)P2 (Pochynyuk et al., 2005).
To extend these results to better understand mechanism, we next investigated regulation of ENaC by PI3-K and its products at the single channel level. Fig. 3 A shows current traces (left) and corresponding all-point histograms (right) for ENaC in representative excised, outside-out patches from cells expressing wild-type channels (top) and channels containing either the α2D (top middle), β2D (bottom middle), or γ2S (bottom) mutant subunit.
As is clear in these representative traces and reported previously by us (Tong et al., 2004b; Pochynyuk et al., 2005; Tong and Stockand, 2005), ENaC activity is stable under these seal conditions. The representative current traces and histograms, as well as the summary single channel current–voltage (I-V) relations (Fig. 3 B) and open probability (C) graphs demonstrate that there is little difference at rest in the basal activity and unitary conductance of ENaC containing all wild-type subunits and those containing a single type of mutant subunit. All were highly selective for Na+ over K+, had single channel conductances at hyperpolarizing potentials ∼5 pS with Na+ as the charge carrier, and had open probabilities approaching 0.4 that were not significantly different. The preference for Na+ over K+, and Po and conductance values agree well with that reported previously for recombinant and native ENaC, and are consistent with the results in Fig. 1 C where channels containing β2D and γ2S mutant subunits had basal activities similar to wild-type channels.
The representative current trace of wild-type mENaC in an excised, outside-out patch in Fig. 4 records a typical response to exogenous PI(3,4,5)P3.
As is apparent in the continuous current trace shown at the top, as well as in the current traces and associated all-point histograms shown below at expanded time scales before (1. control) and after (2. PIP3) addition of exogenous PI(3,4,5)P3, treatment markedly increases channel activity by affecting Po. This patch contains, at least, five ENaC, which become easy to resolve following the slowing of gating and stabilization of the open state upon addition of PI(3,4,5)P3. This characteristic response to exogenous PI(3,4,5)P3 is consistent with that reported previously for wild-type mENaC (Tong et al., 2004b; Pochynyuk et al., 2005).
PI(3,4,5)P3 markedly increases Po of ENaC-containing α2D. In contrast, exogenous PI(3,4,5)P3 has little effect on the Po of channels containing β2D.
Fig. 6 (A–D) summarizes the effects of exogenous PI(3,4,5)P3 on ENaC Po for wild-type channels and channels containing either the α2D, β2D, or γ2S mutant subunits.
Effects on Po were quantified in paired experiments similar to the representative experiments shown in Fig. 5. Channels containing all wild-type subunits and those containing the α2D mutant were fully responsive to PI(3,4,5)P3. In contrast, those having either the β2D or γ2S mutant subunit failed to respond to PI(3,4,5)P3. These results are consistent with those in Fig. 2. Also consistent are the results summarized in Fig. 6 (E–H). Shown here is Po for wild-type channels and channels containing α2D, β2D, and γ2S mutant subunits in the absence and presence of coexpression of PI3-K. Open probability was quantified in unpaired experiments using excised, outside-out patches similar to those in Fig. 3. Channels containing all wild-type subunits and those containing α2D but not β2D or γ2S mutant subunits had significantly greater Po in the presence of PI3-K. These results as a whole emphasize the importance of the conserved positive-charged residues in the regions just following TM2 in β- and γ-ENaC to regulation of Po by PI3-K and its phosphoinositide products.
Conserved Positive-charged Residues at the Extreme N Terminus of β- and γ-ENaC Are Necessary for PI(4,5)P2 Regulation of Po
After identifying regions of ENaC critical to regulation by PI3-K and PI(3,4,5)P3, we wondered if these regions played a role in PI(4,5)P2 regulation. To begin addressing this question, we next quantified regulation of ENaC by chronic stimulation of PI(4)P5-K signaling. Our goal was to establish conditions favoring chronic elevations in PI(4,5)P2. As with overexpression of PI3-K, we first quantified ENaC activity in response to coexpression with PI(4)P5-K in whole-cell voltage-clamp experiments monitoring changes in macroscopic, amiloride-sensitive currents. Fig. 7 reports the relative increase in activity in response to PI(4)P5-K for channels containing all wild-type subunits and those containing a single type of mutant subunit coexpressed with complementary wild-type subunits.
Under these conditions, PI(4)P5-K increased the activity of all channels. We interpret this finding as meaning that either none of the regions probed here play a role in regulation by PI(4,5)P2 or that a response to chronic elevations in PI(4,5)P2 is complex, involving possible changes in either the number of channels in the membrane or channel Po or both. In consideration of earlier work (Staruschenko et al., 2004a; Pochynyuk et al., 2006; Pochynyuk et al., 2007a), we suggest that chronic elevations in PI(4,5)P2 increase the number of ENaC in the membrane and that the regions mutated in the current study do not have a role in this regulation. If this is correct, then possible effects on Po may have been masked during chronic elevation of PI(4,5)P2. Other possibilities that are not necessarily mutually exclusive also exist. For instance, PI(4,5)P2 could repress negative regulation of channel Po. If resting PI(4,5)P2 levels are saturating for this effect but not for effects on N, then disrupting the PI(4,5)P2 regulatory site relevant to Po would have little effect on basal activity and activity increases in response to chronic elevations in PI(4,5)P2. With such a scenario, PI(4,5)P2 effects on channel activity via Po would only become apparent at lower, nonsaturating levels of the phosphoinositide.
Thus, we were not completely satisfied by this initial probing of sites within ENaC potentially important to direct regulation of the channel by PI(4,5)P2. Since we and others have had previous success modulating ENaC activity and Po by decreasing PI(4,5)P2 levels (Ma et al., 2002; Yue et al., 2002; Kunzelmann et al., 2005; Tong and Stockand, 2005), we next turned to this approach. However, before performing these experiments in an expression system, we wanted to confirm that such regulation is observable for native ENaC in epithelial cells with the goal of using identical tools to manipulate PI(4,5)P2 and possibly ENaC Po in both systems. We and others have previously decreased PI(4,5)P2 levels indirectly by pushing the equilibrium between tyrosine kinases and protein tyrosine phosphatases toward the prior (Bezzerides et al., 2004; Tong and Stockand, 2005). This allows PLC-γ, a tyrosine kinase target, to predominate decreasing PI(4,5)P2 levels.
Fig. 8 A shows fluorescence micrographs of emissions from a PI(4,5)P2 reporter fused to eGFP in the apical membranes of mpkCCD14 mouse principal cells within confluent epithelial monolayers before (1) and 5 (2) and 15 (3) min after addition of vehicle (top) and 100 μM vanadate (bottom).
For these experiments, emissions from the apical membrane were optically isolated with TIRF microscopy. Only monolayers cultured in the presence of FBS and corticosteroids that exhibited robust Na+ transport, thus containing active ENaC, were used. As is clear in this representative experiment and in the summary graph in Fig. 8 B, addition of vanadate rapidly decreases apical membrane PI(4,5)P2 levels (n = 5). In contrast, PI(4,5)P2 levels within the apical membrane were not influenced by the negative control of inhibiting PI3-K with LY294002 (50 μM; n = 9). Vanadate decreases PI(4,5)P2 levels in CHO cells with a similar time course (Tong and Stockand, 2005).
As shown by the experiments in Fig. 9, decreasing PI(4,5)P2 levels with vanadate decreases ENaC Po in mpkCCD14 principal cells.
Shown in Fig. 9 A is a representative current trace from a cell-attached patch made on the apical membrane of an mpkCCDc14 cell within a monolayer before and after addition of vanadate. This patch contains a single ENaC. Fig. 9 B shows the all points histograms associated with this channel before and after vanadate. The summary graph in Fig. 9 C of similar paired experiments demonstrates that decreasing PI(4,5)P2 with vanadate significantly decreases Po of native ENaC. In contrast, vehicle had no effect on Po (0.29 ± 0.07 before and 0.27 ± 0.07 after; n = 4; unpublished data). A similar rapid decrease in ENaC Po is observed in paired cell-attached patch clamp experiments when PI(4,5)P2 levels are decreased with ATP activation of purinergic receptors (Po = 0.32 ± 0.09 and 0.02 ± 0.01 before and after ATP; n = 6; unpublished data), and in an isolated, split-open rat collecting duct preparation when vanadate is used to decrease PI(4,5)P2 levels (unpublished data).
After confirming regulation of native ENaC in epithelial cells by manipulating PI(4,5)P2 levels with vanadate, we tested the effects of this agent on the activity of wild-type and mutant ENaC expressed in CHO cells. The summary graph in Fig. 10 reports the relative decrease in activity for wild-type and mutant channels in response to vanadate.
As expected, vanadate acutely decreased activity. Channels containing subunits with the N terminus of β-ENaC deleted (βND) or the conserved lysines within the N terminus of γ-ENaC neutralized to alanines (γNS) were resistant to the effects of vanadate, with the latter having significantly more activity in the presence of vanadate compared with wild-type channels. When the βND and γNS mutant subunits were incorporated into the same channel, decreasing PI(4,5)P2 levels with vanadate had no effect on activity, demonstrating that these regions of ENaC, in particular, their conserved positive-charged residues, are critical to deceases in activity in response to a reduction of PI(4,5)P2.
To better understand the molecular mechanism of action of PI(4,5)P2 depletion in response to vanadate on ENaC, we studied this phenomena with single channel resolution. Fig. 11 shows representative current traces of ENaC in outside-out patches from CHO cells expressing recombinant channels containing all wild-type subunits (A) and γNS (B) mutant subunits before and after treatment with vanadate.
As is clear in these representative experiments and in the corresponding summary graphs of similar paired experiments (Fig. 11 C = wt, D = α2D, and E = γNS), vanadate significantly and rapidly decreases the Po of channels containing wild-type and α2D mutant subunits but not those containing the γ1S mutant subunit. These results are consistent with those in Fig. 10 and support a role for the conserved positive-charged residues in the N terminus of γ- and β-ENaC in the dynamic regulation of Po by acute decreases in PI(4,5)P2.
The direct effects of exogenous PI(4,5)P2 on ENaC open probability were measured next in paired experiments using excised, inside-out patches. Results from these experiments are shown in Fig. 12.
As expected, ENaC activity quickly decreases following patch excision. Such rundown is well documented for ENaC and several other channels in the excised, inside-out patch configuration. The cause of rundown remains unclear; however, it may be related to loss of a factor necessary/permissive for channel gating. Nevertheless, as documented by the representative current trace in Fig. 12 A and summarized in C, addition of 20 μM PI(4,5)P2 to the intracellular face of wild-type ENaC quickly recovers activity. Channels containing γNS and βND mutant subunits, in contrast, do not respond to exogenous PI(4,5)P2. Fig. 12 B contains a representative current trace for these channels, with D containing summarized data. The rapidity with which rundown wild-type channels recover activity in response to exogenous PI(4,5)P2 in excised patches presumably inert with respect to trafficking is consistent with effects on Po. It is important to note though that this also could reflect activation of previously quiescent/inactive channels. Moreover, possible (additional) effects on N likely are missed with these experimental conditions. Nevertheless, if one assumes that ENaC rundown does not exclusively result from loss of PI(4,5)P2, then these findings agree with the VO4 results made above and are consistent with regions in the extreme N termini of β- and γ-ENaC playing a role in regulation by PI(4,5)P2.
We identify in this study regions of ENaC necessary for phosphoinositide regulation. The composition and position of these regions are consistent with them possibly playing a role in phosphoinositide binding. They contain multiple positive-charged residues conserved across all species. They are well positioned to influence channel gating. Our results are consistent with both PI(3,4,5)P3 and PI(4,5)P2 impacting ENaC open probability in a fast and dynamic manner. Increasing the former elevates ENaC open probability above basal levels, demonstrating that this phosphoiniositide is modulatory and that its effects are not saturated at rest. Chronic increases in PI(4,5)P2, in contrast, appear to have little effect on ENaC gating at rest. However, acute decreases in PI(4,5)P2 markedly reduce ENaC open probability. Being mindful of the limitations inherent to all patch clamp studies, these findings suggest to us that basal PI(4,5)P2 levels may be functionally saturating with respect to modulation of ENaC in CHO cells and that resting levels of this phosphoinositide exert significant action on the channel. If correct, then PI(4,5)P2 may be necessary for normal ENaC gating in a manner similar to its actions on other ion channels, specifically Kir and TRPM; binding of PI(4,5)P2 primes the channel to gate constitutively and in response to additional modulators (Huang et al., 1998; Zhang et al., 1999; Enkvetchakul et al., 2005; Ribalet et al., 2005, 2006; Rohacs et al., 2005; Nilius et al., 2006). We report here that the regions at the extreme N terminus of β- and γ-ENaC are critical to regulation by PI(4,5)P2 but not PI(3,4,5)P3. In contrast, the intracellular regions just following TM2 in β- and γ- but not α-ENaC are critical to regulation by PI(3,4,5)P3 but not PI(4,5)P2.
The Putative PI(3,4,5)P3 Binding Site
Compared with ENaC, more is known about phosphoiniositide binding sites and their associated molecular mechanisms in other ion channels comprised of subunits having two transmembrane domains. The regions of β- and γ-ENaC that are necessary for PI(3,4,5)P3 stimulation are very similar to the proposed phosphoinositide binding sites in all Kir channels (Zhang et al., 1999; Soom et al., 2001; Lopes et al., 2002; Ribalet et al., 2005; Ma et al., 2007). These contain several well-conserved positive-charged residues at the interface between a trans membrane domain involved in pore formation and possibly gating and the intracellular domains of the polypeptide. Neither the proposed phosphoinositide binding site in Kir channels nor that proposed here for PI(3,4,5)P3 in β- and γ-ENaC subunits fits the classic examples of phosphoinositide binding sites, including PH, PX, and FYVE domains, well established in many signaling proteins (Lemmon, 2003). So, much remains unknown about how such putative binding sites within ion channels coordinate phosphoinositide binding. Nevertheless, phosphoinositide binding to ion channels, in general, is believed to involve electrostatic interactions between basic residues within the binding domain and anionic phosphate head groups of the phospholipid.
The Putative PI(4,5)P2 Binding Site
As with the putative PI(3,4,5)P3 binding sites in the C termini of β- and γ-ENaC, those identified in the extreme N termini of these subunits involved in PI(4,5)P2 regulation contain several conserved basic residues, and are similar in composition to the proposed phosphoinositide binding sites in other channels. The site in the N terminus of γ-ENaC contains the [R/K]-X3-11- [R/K]-X-[R/K]-[R/K] motif (where x is any amino acid) common to PH domains. A similar sequence is apparent in the phosphoinositide binding site in the C terminus of TRPM4 (Nilius et al., 2006). Thus, phosphoinositide binding to ion channels may have more in common, then first suspected, with binding of these molecules to signaling proteins. Additional analysis of the specific charged residues in the N terminus of γ-ENaC involved in coordinating PI(4,5)P2 will determine whether this is coincidental and may provide a framework on which to map other putative phosphoinositide binding sites within ion channels.
The putative PI(4,5)P2 binding sites in ENaC at the extreme N termini in β- and γ-ENaC subunits are far removed from the proposed channel gate. This makes it difficult to predict mechanism of action. Nevertheless, it is possible that PI(4,5)P2 binding to this region of the channel immobilizes a negative regulator of the channel, allowing normal gating. While several alternatives exist that cannot be excluded with the current dataset, we favor this interpretation of our results.
A major difference exists regarding PI(4,5)P2 and PI(3,4,5)P3 regulation of ENaC. This being that the effects of PI(4,5)P2 appear saturating at rest and permissive for normal channel gating. Thus, we propose that occupation of the putative PI(4,5)P2 binding site in some manner removes a negative regulation of the channel. This is consistent with the current results; depleting PI(4,5)P2 rapidly decreases ENaC Po and charge neutralization/deletion of key basic residues at the N termini of β- and γ-ENaC render the channel insensitive to PI(4,5)P2 without affecting basal Po. Thus, it appears that the basic residues in the N terminus of β- and γ-ENaC in some manner restrict ENaC opening and that PI(4,5)P2 occupies these to counter this negative regulation.
This proposed mechanism shares features with phosphoinositide regulation of Kv channels (Oliver et al., 2004). For these latter channels, PI(4,5)P2 binds and immobilizes the cytoplasmic inactivation ball by gluing it to the plasma membrane, resulting in relief of N-type inactivation. This is not to say that ENaC gating is controlled by voltage or a ball and chain mechanism, but rather that PI(4,5)P2 occupies a negative regulator of gating. The position of this switch then is defined by PI(4,5)P2 binding and removal of this region of the channel is equivalent to binding. In contrast, removal of the PI(3,4,5)P3 binding site is not equivalent to phosphoinositide binding, and thus PI(3,4,5)P3 binding does not relieve inhibition but rather is stimulatory.
An important finding of this study is that these distinct phosphoinositides require different regions of the channel in order to affect open probability and activity. This finding is important for it explains how a channel, such as ENaC, sensitive to, at least, two different phosphoinositides, is capable of responding in the proper manner to each. Moreover, it provides the framework to understand how dynamic stimulation of ENaC open probability by PI(3,4,5)P3 is possible in the presence of a distinct phosphoinositide, PI(4,5)P2, that is relatively more abundant and exerting simultaneous regulation that appears to be functionally saturated at rest.
We propose that like PI(4,5)P2 interactions with Kir, interaction of this phosphoinositide with ENaC sets the stage for subsequent regulation of gating by other channel modulators. Thus, PI(4,5)P2 binding to ENaC is permissive for the channel to gate at a constitutive level, which then can be further modified by additional factors to include PI(3,4,5)P3 binding at a distinct site. Such regulation of ENaC by PI(4,5)P2 and PI(3,4,5)P3 would have many parallels with contemporary thinking about regulation of Kir channels by PI(4,5)P2 plus ATP and βγ G-protein subunits (Xie et al., 2006; Ribalet et al., 2005; Zhang et al., 1999; Huang et al., 1998). (We note that the basic residues in the N terminus of γ-ENaC, and likely those in β-ENaC, involved in a PI(4,5)P2 response are identical to those targeted for ubiquitinylation by Nedd4 ubiquitin ligases during down-regulation of channel activity (Staub et al., 1997). It will be interesting to determine whether these residues play a role in signaling convergence between PI(4,5)P2 regulation of Po and Nedd4 regulation of N.)
Similar to β and γ subunits, α-ENaC has several intracellular tracks containing clusters of basic residues that are reminiscent of proposed phosphoinositide binding sites in other channels. However, in contrast to the other two subunits, we find no role for α-ENaC in a response to phosphoinositides. It is not clear why this subunit would not have a role in this regulation, and it may have been missed. Similarly, it is not clear why deletion and charge neutralization of intracellular regions of the channel just preceding TM1 result in loss of function, and the current results cannot exclude a role for these or any other channel domain in regulation by phosphoinositides. What we can conclude is that distinct regions within β- and γ-ENaC bestow sensitivity to both PI(4,5)P2 and PI(3,4,5)P3, allowing the channel to differentially respond to these phosphoinositides.
This research was supported by the National Institutes of Health grants RO1DK59594 and R01DK070571, and the American Heart Association grant EIA 0640054N (to J.D. Stockand).
Lawrence G. Palmer served as editor.
Abbreviations used in this paper: ENaC, epithelial Na+ channel; PH, pleckstrin homology; PI3-K, phosphatidylinositide 3-OH kinase; PI(4,5)P2, phosphatidylinositol 4,5-bisphosphate; PI(3,4,5)P3, phosphatidylinositol 3,4,5-trisphosphate; TIRF, total internal reflection fluorescence.