Cystic fibrosis transmembrane conductance regulator differentially regulates human and mouse epithelial sodium channels in Xenopus oocytes.

The cystic fibrosis transmembrane conductance regulator (CFTR), in addition to its well defined Cl- channel properties, regulates other ion channels. CFTR inhibits murine or rat epithelial Na+ channel (mENaC or rENaC) currents in many epithelial and non-epithelial cells, whereas murine or rat ENaC increases CFTR functional expression. These regulatory interactions are reproduced in Xenopus oocytes where both the open probability and surface expression of wild type CFTR Cl- channels are increased when CFTR is co-expressed with alphabetagamma mENaC, and conversely the activity of mENaC is inhibited after wild type CFTR activation. Using the Xenopus oocyte expression system, differences in functional regulatory interactions were observed when CFTR was co-expressed with either alphabetagamma mENaC or alphabetagamma human ENaC (hENaC). Co-expression of CFTR and alphabetagamma mENaC or hENaC resulted in an approximately 3-fold increase in CFTR Cl- current compared with oocytes expressing CFTR alone. Oocytes co-injected with both CFTR and mENaC or hENaC expressed an amiloride-sensitive whole cell current that was decreased compared with that observed with the injection of mENaC or hENaC alone before CFTR activation with forskolin/3-isobutyl-1-methylxanthine. CFTR activation resulted in a further 50% decrease in mENaC-mediated currents, an approximately 20% decrease in alpha-T663-hENaC-mediated currents, and essentially no change in alpha-A663-hENaC-mediated currents. Changes in ENaC functional expression correlated with ENaC surface expression by oocyte surface biotinylation experiments. Assessment of regulatory interactions between CFTR and chimeric mouse/human ENaCs suggest that the 20 C-terminal amino acid residues of alpha ENaC confer species specificity regarding ENaC inhibition by activated CFTR.

Cystic fibrosis (CF) 1 results from mutations in the gene encoding the cystic fibrosis transmembrane conductance regu-lator (CFTR) (1). In addition to functioning as a cAMP-activated, ATP-dependent Cl Ϫ channel, CFTR influences the transepithelial transport of other solutes, including Na ϩ via the epithelial sodium channel (ENaC), Cl Ϫ via an outwardly rectifying Cl Ϫ channel, K ϩ via Kir1.1, HCO 3 Ϫ , and ATP (2)(3)(4)(5). Functional interactions between CFTR and rodent ENaC have been observed in epithelial as well as non-epithelial cells (6 -12). The activation of CFTR is generally associated with an inhibition of ENaC (7,8,10,11,13,14), although activation of CFTR leads to activation of ENaC in the sweat duct (12) suggesting that the regulatory interactions between these two transporters are complex. The co-expression of CFTR and rodent ENaC in Xenopus oocytes results in regulatory interactions that mimic the airway where there is a decrease in ENaCmediated Na ϩ transport in the presence of CFTR (9 -11, 14). Furthermore CFTR-mediated Cl Ϫ conductance is increased in presence of ENaC in oocytes (4,(7)(8)(9)(10)(11)14). In contrast, the ⌬F508-CFTR mutation, the most prevalent mutation found in North American Caucasian patients with CF, does not inhibit the functional expression of rat (15) or mouse ENaC (mENaC) (11), and mENaC does not enhance the functional expression of ⌬F508-CFTR in Xenopus oocytes (11). The isoflavone genistein, which activates the Cl Ϫ conductance of both wild type and mutant CFTRs (16 -19), can restore these interactions (11).
The mechanism by which this interregulation of CFTR and ENaC occurs in oocytes is unclear and somewhat controversial. Others have suggested that the decrease in ENaC-mediated current after CFTR activation may result from a series resistor error (20) and have presented data in abstract form that suggest that activation of CFTR in oocytes does not result in a further decrease in hENaC-mediated current (21). Our previous data (11) are inconsistent with the hypothesis that the apparent decrease in mENaC-mediated current with CFTR activation is due to a series resistor error. Furthermore such a mechanism cannot account for the increased CFTR functional expression observed with co-expression of rodent ENaC (9 -11, 22).
The present studies were aimed at better understanding the regulatory interactions between CFTR and ENaC and whether there are species-specific differences in these interactions. In this regard, our data are consistent with the decrease in mENaC functional expression that occurs with CFTR activa-tion being due, in part, to decreased mENaC surface expression, while hENaC (with Ala in position ␣663) functional and surface expression does not decrease following CFTR activation. These observations led us to test two additional hypotheses. First we tested the hypothesis that the C terminus of the ENaC ␣ subunit, which has limited homology between mouse and human, may influence the species-related differences in this interaction. We also tested the hypothesis that a naturally occurring polymorphism in the C terminus of ␣ hENaC, substitution of Ala at residue 663 for Thr (T663A), which we have recently shown to decrease the functional and surface expression of hENaC in oocytes, 2 would influence regulatory interactions between CFTR and hENaC. Our data suggest that the C terminus of ␣ hENaC confers species specificity regarding ENaC inhibition by activated CFTR, whereas the ␣T663A polymorphism has a modest effect on the response of hENaC to activated CFTR.

EXPERIMENTAL PROCEDURES
Materials-Forskolin and IBMX were purchased from Sigma. All other reagents were purchased from Fisher.
Electrophysiological Analyses-Whole cell current measurements were performed 24 -48 h after injection using the two-electrode voltage clamp method as described previously (10,11). Oocytes were placed in a 1-ml chamber containing modified ND96 (96 mM NaCl, 1 mM KCl, 0.2 mM CaCl 2 , 5.8 mM MgCl 2 , 10 mM Hepes, pH 7.4) and impaled with micropipettes of 0.5-5-megaohm resistance filled with 3 M KCl. The whole cell currents were measured by voltage clamping the oocytes in 20-mV steps between Ϫ140 and ϩ60 mV adjusted for resting transmembrane potential. Whole cell currents (I) were digitized at 200 Hz during the voltage steps, recorded directly onto a hard disk, and analyzed using pClamp 8.1 software (Axon Instruments, Foster City, CA). To reduce error due to series resistance, the voltage clamp (Axon Geneclamp 500B) was configured to clamp the bath potential to 0 mV. In this configuration, we independently monitored the oocyte membrane potential during our clamp protocol and routinely observed membrane potentials that were Ͻ5% depolarized from our target holding potentials.
The difference in whole cell currents measured in the absence and presence of 10 M amiloride was used to define the amiloride-sensitive Na ϩ current that was carried by ENaC. CFTR was activated by perfusion of the oocyte with buffer containing 10 M forskolin and 500 M IBMX (forskolin/IBMX) for 25 min (10,11). CFTR Cl Ϫ current was defined as the difference between amiloride-insensitive current measured before and after perfusion with forskolin/IBMX. Whole cell currents were recorded at a clamp potential of Ϫ100 mV for comparisons. All measurements were performed at room temperature.
Immunoblot Analysis-Streptavidin-precipitated proteins or whole oocyte lysates (prepared as above) were incubated in Laemmli sample buffer, resolved by SDS-PAGE, and transferred to nitrocellulose as described previously by our group. 2 ␤-V5 was identified by immunoblot using an anti-V5 monoclonal antibody (Invitrogen), an anti-mouse horseradish peroxidase-conjugated secondary antibody, and ECL (Amersham Biosciences) visualization essentially as described previously by our group (24,25). Fluorogram density was quantitated using an AlphaImager 2200 system and version 5.5 software (Alpha Innotech, San Leandro, CA), and the intensity of ␤-V5 was expressed relative to that of oocytes injected with ENaC alone without forskolin/IBMX stimulation. As in our previous work, 2 less than 20% of intracellular ␤-V5 epitope was biotin-labeled in oocytes expressing only ␤-V5 and ␥ ENaC subunits (data not shown).
Statistical Analyses-All data are presented as mean Ϯ S.E. Statistical comparisons were performed using the Student's t test. A pairwise t test was used for pre/post-treatment in experiments using an individual oocyte. A two-tailed t test was used when comparing currents obtained from oocytes injected with a cRNA for a single transporter (i.e. ENaC or CFTR versus oocytes co-injected with cRNAs for both ENaC and CFTR). p values Յ0.05 were accepted to indicate statistical significance. Statistical analyses were performed using SigmaStat version 2.03 software.
As the mechanism by which these alterations in mENaC functional expression upon co-expression and activation of CFTR occur is not clear, we performed surface biotinylation experiments to assess the amount of mENaC at the oocyte surface. In these experiments, we used mENaC where the ␤ subunit contained a C-terminal V5 epitope tag to increase sensitivity of our immunoblots; control experiments suggested that this ␤ mENaC modification did not influence regulatory interactions between CFTR and mENaC as measured by TEV (data not shown). As shown in Fig. 1C, mENaC(␤-V5) expression at the oocyte surface was unaltered by forskolin/IBMX in oocytes injected with mENaC alone. In contrast, mENaC surface expression decreased when mENaC was co-expressed with CFTR and was further decreased by activation of CFTR in oocytes expressing both channels. There was direct correspondence of these changes in surface expression of mENaC(␤-V5) and the mENaC-mediated currents by TEV (Fig. 1A), suggesting that alterations in mENaC functional expression are, in part, a result of changes in mENaC expression at the oocyte surface (or "N").
We also assessed whole oocyte expression of mENaC(␤-V5) (Fig. 1D). Expression of mENaC(␤-V5) was decreased by coinjection of CFTR but was not further decreased upon activation of CFTR with forskolin/IBMX. These data are consistent with CFTR activation causing an acute decrease in mENaC(␤-V5) surface and functional expression without altering the total amount of mENaC(␤-V5) present in the oocyte.
Co-expression of CFTR and hENaC-We next assessed regulatory interactions between CFTR and hENaC as well as the potential influence of the T663A functional polymorphism of ␣ hENaC described recently by our group 2 on these interactions. Similar to our data with mENaC ( Fig. 1A), forskolin/IBMX did not alter hENaC functional expression in oocytes injected with hENaC alone as is shown in Fig (Fig. 1A), and CFTR activation did not significantly decrease ␣-A663-hENaC functional expression (Ϫ0.78 Ϯ 0.15 versus Ϫ0.69 Ϯ 0.14 A, n ϭ 25, p ϭ ns, Fig. 2B), although the lack of significance of this decrease for ␣-A663-hENaC may be related to the larger standard errors in these data compared with those of ␣-T663-hENaC. CFTR functional expression was enhanced by co-injection of ␣-T663-hENaC (coinjected: Ϫ6.32 Ϯ 1.14 A, n ϭ 24, versus injected with CFTR alone: Ϫ2.83 Ϯ 0.50 A, n ϭ 19, p ϭ 0.009) or ␣-A663-hENaC A and B, CFTR and mENaC were expressed separately or together in oocytes, and TEV was performed as described under "Experimental Procedures." A, whole cell currents that were sensitive to inhibition by 10 M amiloride (Ϫ100 mV holding potential) were determined in oocytes expressing mENaC or co-expressing mENaC and CFTR prior to (closed bars) and following (open bars) stimulation with 10 M forskolin, 500 M IBMX. B, changes in whole cell currents (Ϫ100 mV holding potential) after stimulation with 10 M forskolin, 500 M IBMX that were not inhibited by 10 M amiloride are illustrated (gray bars). Data obtained from the same CFTR/mENaCco-injected oocytes are presented in A and B. Means Ϯ S.E. are illustrated. C and D, ␣␤␥ mENaC, where the ␤ subunit contained a C-terminal V5 epitope tag, was expressed in oocytes either alone or together with CFTR. 24 -48 h after injection surface biotinylation was performed (C) or whole oocyte lysates were prepared from oocytes (D) prior to or following treatment with 10 M forskolin, 500 M IBMX for 20 min as described under "Experimental Procedures." C, immunoblot of streptavidin-agarose-precipitated proteins probed with anti-V5 antiserum to detect mENaC ␤-V5. A representative immunoblot and densitometry (mean Ϯ S.E. of n ϭ 4 independent experiments) are shown. Data were normalized to levels of ␤-V5 surface expression in oocytes expressing ENaC alone and not treated with forskolin/IBMX. D, immunoblot of whole oocyte lysate from equal numbers of solubilized oocytes probed with anti-V5 antiserum to detect mENaC ␤-V5. A representative immunoblot of n ϭ 3 independent experiments is shown. Forsk, forskolin.
FIG. 2. Expression of CFTR and hENaC in Xenopus oocytes. A, B, and C, CFTR and hENaC were expressed separately or together in oocytes, and TEV was performed as described under "Experimental Procedures." A, whole cell currents that were sensitive to inhibition by 10 M amiloride (Ϫ100 mV holding potential) were determined in oocytes expressing ␣-T663-hENaC or co-expressing ␣-T663-hENaC and CFTR prior to (closed bars) and following (open bars) stimulation with 10 M forskolin, 500 M IBMX. B, whole cell currents that were sensitive to inhibition by (co-injected: Ϫ7.78 Ϯ 1.32 A, n ϭ 25, versus injected with CFTR alone: Ϫ3.08 Ϯ 0.40 A, n ϭ 19, p ϭ 0.004, Fig. 2C) as it was by mENaC (Fig. 1B).
We next sought to correlate the functional expression of ␣-A663-hENaC by TEV with its surface expression by surface biotinylation of oocytes expressing ␣-A663-hENaC (where the hENaC ␤ subunit contained a C-terminal V5 epitope tag). Again there was direct correlation of ␣-A663-hENaC(␤-V5) surface expression (Fig. 2D) and ␣-A663-hENaC functional expression by TEV (Fig. 2B). Surface and functional expression of ␣-A663-hENaC was not altered by treatment with forskolin/ IBMX in oocytes injected with ␣-A663-hENaC alone. Co-injection of CFTR decreased surface and functional expression of ␣-A663-hENaC (Fig. 2, B and D) as well as whole oocyte expression of hENaC(␤-V5) (Fig. 2E). Activation of CFTR with forskolin/IBMX in co-injected oocytes did not further alter surface or whole oocyte expression of hENaC(␤-V5). These data suggest that the regulation of mENaC and hENaC by CFTR in oocytes differs and involves changes in mENaC surface expression in response to CFTR activation.
ENaC ␣ Subunit Determines Species Differences in ENaC Functional Response to CFTR Activation-Others have suggested that the C terminus of ␤ ENaC and N terminus of ␥ ENaC are critical in the regulatory interactions between rat ENaC and CFTR (9). To map the domain(s) within mENaC and ␣-A663-hENaC that are responsible for the differential regulation of these ENaCs by activated CFTR in oocytes, we expressed chimeric ENaCs comprised of combinations of mENaC and hENaC subunits and performed TEV experiments similar in design to those of Figs. 1 and 2. Fig. 3 demonstrates the interregulation of CFTR with a chimeric murine/human ENaC comprised of human ␣ and ␥ ENaC subunits and ␤ mENaC. Expression of this chimeric murine/human ENaC in oocytes resulted in ENaC-mediated currents similar in magnitude to the non-chimeric murine and human ENaC. Human ␣(A663)␥ murine ␤ ENaC-mediated currents were slightly but significantly decreased by forskolin/ IBMX in oocytes injected with chimeric ENaC alone (Ϫ2.15 Ϯ 0.31 versus Ϫ1.75 Ϯ 0.28 A, n ϭ 23, p Ͻ 0.001, Fig. 3A). Decreased functional ENaC expression was observed in oocytes co-injected with CFTR (Ϫ0.88 Ϯ 0.12 A, n ϭ 23) compared with oocytes injected with chimeric ENaC alone (Ϫ2.15 Ϯ 0.31 A, n ϭ 23, p Ͻ 0.001, Fig. 3A) as well as enhanced CFTR functional expression in co-injected oocytes (Ϫ8.80 Ϯ 1.69 A, n ϭ 23) compared with oocytes injected with CFTR alone (Ϫ3.55 Ϯ 0.82 A, n ϭ 22, p ϭ 0.009, Fig. 3B). Interestingly CFTR activation in co-injected oocytes did not result in a further decrease of functional expression of the human ␣(A663)␥ murine ␤ ENaC chimera (Ϫ0.88 Ϯ 0.12 versus Ϫ0.70 Ϯ 0.09 A, n ϭ 23, p ϭ ns, Fig. 3A). The functional expression of this chimeric ENaC, as measured by TEV, again correlated with the surface expression of the ␤ subunit of this chimera as assessed by oocyte surface biotinylation experiments (Fig. 3, C and D). These data suggest that the murine ENaC ␤ subunit is not, by itself, responsible for the different responses of mouse and human ENaC to activation of CFTR.

FIG. 3. Regulatory interactions of CFTR and chimeric human ␣(A663)␥ murine ␤ ENaC.
A and B, CFTR and human ␣(A663)␥ murine ␤ ENaC were expressed separately or together in oocytes, and TEV was performed as described under "Experimental Procedures." A, whole cell currents sensitive to inhibition by 10 M amiloride (Ϫ100 mV holding potential) were determined in oocytes expressing chimeric ENaC or co-expressing chimeric ENaC and CFTR prior to (closed bars) and following (open bars) stimulation with 10 M forskolin, 500 M IBMX. B, changes in whole cell currents (Ϫ100 mV holding potential) after stimulation with 10 M forskolin, 500 M IBMX that were not inhibited by 10 M amiloride are illustrated (gray bars). Data obtained from the same CFTR/chimeric ENaC-co-injected oocytes are presented in A and B. Means Ϯ S.E. are illustrated. C and D, human ␣(A663)␥ murine ␤ ENaC, where the ␤ subunit contained a C-terminal V5 epitope tag, was expressed in oocytes either alone or together with CFTR. 48 h after injection surface biotinylation was performed in oocytes prior to or following treatment with 10 M forskolin, 500 M IBMX for 20 min as described under "Experimental Procedures." A representative immunoblot of streptavidin-agaroseprecipitated proteins probed with anti-V5 antiserum to detect mENaC ␤-V5 (C) and densitometry (D) (mean Ϯ S.E. of n ϭ 3 independent experiments) are shown. Data were normalized to levels of ␤-V5 surface expression in oocytes expressing human ␣(A663)␥ murine ␤ ENaC alone and not treated with forskolin/IBMX. Forsk, forskolin.
Again the lack of significance of the increase for ␣-T663-hENaC may be related to the larger standard errors in these data compared with those of ␣-A663-hENaC.
We also assessed the effect of co-injection of ⌬F508-CFTR on the whole oocyte expression of ␣-A663-hENaC(␤-V5). Unlike for wild type CFTR (Fig. 2E), co-injection of ⌬F508-CFTR did not decrease the whole oocyte expression of hENaC(␤-V5). Furthermore forskolin/IBMX stimulation did not alter hENaC(␤-V5) whole oocyte expression (Fig. 7D). These data suggest that the decrease in hENaC(␤-V5) whole oocyte expression with co-injection of wild type CFTR (Fig. 2E) is not a result of competition for cellular translational or biosynthetic machinery as injection of a similar amount of ⌬F508-CFTR cRNA did not decrease hENaC(␤-V5) whole oocyte expression. DISCUSSION Defects in or the absence of CFTR leads to increased ENaC activity in the CF airway epithelia, resulting in increased solute and liquid absorption, decreased airway surface volume, and decreased mucociliary clearance (26 -28). Regulatory interactions between CFTR and ENaC are complex, and our data address a number of issues regarding these interactions. We observed that CFTR-induced inhibition of ENaC functional expression correlates with changes in ENaC surface expression, suggesting that an element of the regulatory interactions between CFTR and ENaC occurs at the level of plasma membrane expression of ENaC and presumably at the level of intracellular trafficking of this channel. Furthermore the extent of ENaC inhibition following CFTR activation is species-specific, and the distal C terminus of the ␣ subunit of ENaC has a key role in determining species specificity.
Correlation of ENaC Functional and Surface Expression in Oocytes-A number of groups have suggested that regulatory interactions between CFTR and ENaC in airway epithelia are replicated in Xenopus oocytes. Functional expression of CFTR has been associated with an inhibition of functional ENaC expression (7, 9 -11, 15) that was due, in part, to a decrease in ENaC open probability (29,30). We observed that co-expression of CFTR decreased both the total cellular pool and cell surface pool of mouse and human ENaC, suggesting that CFTR may alter ENaC biosynthesis, trafficking to the oocyte membrane, or retrieval from the plasma membrane. A further decrease in mouse ENaC currents was observed following CFTR activation that correlated with a decrease in mENaC surface expression. One group has suggested that the inhibition of ENaC functional expression with CFTR activation may be due to a series resistor error that is potentially inherent to TEV experiments in oocytes (20). However, our previous data (11,22) and experiments presented here (Fig. 1) demonstrating that decreases in mENaC currents following CFTR activation correlate with mENaC surface expression suggest that the decrease in mENaC activity following CFTR activation does not reflect a series resistor error.
Species-specific Regulatory Interactions between CFTR and ENaC and the Role of the C Terminus of ␣ ENaC-Our data suggest that there are differences in the regulation of mouse and human ENaC by activated CFTR in oocytes. Co-expression of CFTR with mouse or human ENaC was associated with a reduction in ENaC activity and surface expression. Although CFTR activation led to a further decrease in functional activity of mouse ENaC, either no decrease (␣-A663-hENaC) or a mod-est decrease (␣-T663-hENaC) in human ENaC-mediated currents was observed following CFTR activation. Furthermore, when mouse ENaC was modified by replacing the C-terminal 21 residues of the mouse ␣ subunit with the corresponding distal C terminus of the human ␣ (murine 1-678, human 650 -669, A663), the resulting chimera was not inhibited following CFTR activation. These data suggest that the C terminus of ␣ ENaC is a critical determinant for the differential responses of mouse and human ENaC to activated CFTR and complement previous findings indicating that the C terminus of the ␤ subunit and the N terminus of the ␥ subunit of rat ENaC are required for functional interactions between CFTR and ENaC (9).
Human ␣T663A is a common polymorphism (31,32), and our recent data suggest that it is associated with differences in functional activity in oocytes. 2 However, the corresponding mouse ␣T692A polymorphism was not associated with differences in mENaC functional activity. 2 Replacement of the Cterminal 21 residues of the mouse ␣ subunit with the corresponding distal C-terminal 20 residues of the human ␣ (murine 1-678, human 650 -669, A663) restored the functional differences that were observed with the human ␣T663A polymorphism. 2 With the increasing interest in potential modifier genes for the CF phenotype and as ENaC function is clearly implicated in CF airway pathophysiology, hENaC functional polymorphisms are candidates for modifying the CF phenotype (26 -28, 33). One study presented in abstract form examined the prevalence of the ␣-A663-hENaC polymorphism in a group of 80 CF patients and found an allele frequency that was similar to the general Caucasian population. However, differences in pulmonary function between patients with ␣-A663and ␣-T663-hENaC were not observed (34), although this study may have been underpowered to detect small differences.
Regulatory Interactions between ⌬F508-CFTR and ENaC-We observed that co-expression of an appropriately folded and trafficked CFTR, prior to its activation by forskolin/IBMX, decreases whole cell, surface, and functional expression of mouse and human ENaC. In contrast, co-expression of ⌬F508-CFTR did not decrease hENaC whole oocyte or functional expression. ⌬F508-CFTR is poorly trafficked to the cell surface and likely does not effectively compete for trafficking machinery (in distinction from translational and biosynthetic machinery) with ENaC. These data are consistent with a model in which CFTR decreases ENaC functional expression, in part, by reducing ENaC trafficking to the plasma membrane, while absent or poorly trafficked CFTR (like ⌬F508-CFTR) does not hinder ENaC trafficking.
Summary-Our results and previous studies suggest that regulation of ion transport in the airway epithelia is complex. Understanding the regulatory interactions between wild type or mutant CFTR and ENaC will likely impact strategies designed to improve mutant CFTR function in CF. That hENaC behaves differently than mENaC in oocytes in response to activated CFTR and that a region of ␣ hENaC responsible for this differential response is within its C-terminal 20 amino acid residues will likely, with future investigations, provide additional insight into this critical interaction of epithelial ion transporters.