Influenza matrix protein 2 alters CFTR expression and function through its ion channel activity

Abstract

The human cystic fibrosis transmembrane conductance regulator (CFTR) is a cyclic AMP-activated chloride (Cl⁻) channel in the lung epithelium that helps regulate the thickness and composition of the lung epithelial lining fluid. We investigated whether influenza M2 protein, a pH-activated proton (H⁺) channel that traffics to the plasma membrane of infected cells, altered CFTR expression and function. M2 decreased CFTR activity in 1) Xenopus oocytes injected with human CFTR, 2) epithelial cells (HEK-293) stably transfected with CFTR, and 3) human bronchial epithelial cells (16HBE14o−) expressing native CFTR. This inhibition was partially reversed by an inhibitor of the ubiquitin-activating enzyme E1. Next we investigated whether the M2 inhibition of CFTR activity was due to an increase of secretory organelle pH by M2. Incubation of Xenopus oocytes expressing CFTR with ammonium chloride or concanamycin A, two agents that alkalinize the secretory pathway, inhibited CFTR activity in a dose-dependent manner. Treatment of M2- and CFTR-expressing oocytes with the M2 ion channel inhibitor amantadine prevented the loss in CFTR expression and activity; in addition, M2 mutants, lacking the ability to transport H⁺, did not alter CFTR activity in Xenopus oocytes and HEK cells. Expression of an M2 mutant retained in the endoplasmic reticulum also failed to alter CFTR activity. In summary, our data show that M2 decreases CFTR activity by increasing secretory organelle pH, which targets CFTR for destruction by the ubiquitin system. Alteration of CFTR activity has important consequences for fluid regulation and may potentially modify the immune response to viral infection.

ion transporters maintain the proper thickness and ionic composition of lung airway surface liquid (ASL) (35). The cystic fibrosis transmembrane conductance regulator (CFTR) is the primary ion channel regulating apical chloride (Cl⁻) transport in the lung epithelium. Inhibition of CFTR function has been shown to reduce ASL thickness, ciliary beat frequency, and mucociliary clearance, resulting in decreased clearance of respiratory pathogens (16). Bicarbonate secretion by CFTR plays an essential role in maintaining the pH of the lung epithelial lining fluid and enhancing the killing of bacteria that would otherwise colonize the lung (40). In addition, CFTR alters the function of other ion channels (9, 31), plays a role in maintaining tight junction integrity (53), regulates antioxidant defenses, modifies cytokine signaling, and controls gene expression (15, 42).

Influenza infection is responsible for 3–5 million cases of severe illness and ∼30,000–50,000 deaths in the United States annually and has the potential to reach far greater numbers in the event of a pandemic (37). The influenza A virus contains three trans-membrane proteins [hemagglutinin (HA), neuraminidase, and matrix protein 2 (M2)] that are incorporated into the virus from the cellular plasma membrane during budding. Influenza M2 is a 97-amino acid integral membrane protein that forms a homotetrameric proton (H⁺) channel activated by acidic pH (29). During viral entry, the M2 protein is activated by the acidic pH in the endosome, causing acidification of the virion, which facilitates M1 uncoating, fusion, and release of ribonucleoprotein (RNP) into the cytoplasm for nuclear transport. This step is essential for viral replication (46) and can be blocked by treatment with the M2 ion transport inhibitor amantadine.

After transport of RNP to the nucleus, transcription of the influenza gene segment seven produces an mRNA encoding matrix protein 1 and a spliced mRNA encoding influenza M2 (29). After nuclear egress of the M2 transcript ∼4–6 h postinfection, the M2 protein is translated into the endoplasmic reticulum (ER) and transits through the secretory pathway to the plasma membrane for incorporation into budding virions (22). During transit, influenza M2 is activated by the low pH of the secretory organelles and transports H⁺ out these acidified compartments. Alkalinization of the secretory pathways enhances the production of viable virus by protecting hemagglutinin from conversion to its low-pH form (21, 43, 47). However, because alteration of vesicular pH by M2 ion channel activity is known to interfere with protein trafficking (21, 43) and clear evidence exists that CFTR is sensitive to changes of vesicular pH (38), we hypothesized that an increase of the secretory pH by M2 may compromise CFTR expression and activity.

In these experiments, we determined that M2 inhibits CFTR activity in heterologous expression systems and in cells expressing native CFTR in a specific and concentration-dependent manner. Since M2 ion channel activity was necessary for CFTR inhibition, and since agents that increased the secretory pathway pH also decreased CFTR activity in this system, we concluded that M2 alters CFTR activity by modifying the pH of the secretory pathway.

METHODS

Reagents.

We used IBMX (Sigma-Aldrich, St. Louis, MO); DIDS (Sigma-Aldrich); amantadine hydrochloride (Sigma-Aldrich); ammonium chloride (NH₄Cl) (Sigma-Aldrich); CFTR inhibitor II GlyH-101 (EMD Millipore, Billerica, MA); CFTR inhibitor IV PPQ-102 (EMD Millipore); concanamycin A (Cayman Chemical, Ann Arbor, MI); EUK-134 (Cayman Chemical); forskolin (Sigma-Aldrich); glutathione ethyl ester (GSH ester) (Sigma-Aldrich); lanthanum (Sigma-Aldrich); and PYR-41 (Sigma-Aldrich).

Harvesting of Xenopus oocytes.

Oocytes were obtained from anesthetized adult female Xenopus laevis toads as previously described (23, 30). After isolation, oocytes were incubated in OR2 solution (82.5 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 5 mM HEPES, pH 7.6; supplemented with 5% equine serum) until the experiments were performed.

Cell lines and cell culture.

HEK-293 stably transfected with human wild-type (wt) CFTR (HEK-293 CFTR-wt) were kindly provided by Dr. Zsuzsanna Bebok (6). Cells were maintained in Dulbecco's modified Eagle's medium and Ham's F12 medium (50:50, vol/vol) (Lonza, Walkersville, MD) supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, 1% penicillin-streptomycin, and 2 μg/ml puromycin (Thermo Scientific, Rockford, IL). 16HBE14o− cells, a human bronchial epithelial cell line (17), were maintained in MEM (minimum essential media; Cellgro, Manassas, VA) supplemented with 10% FBS, 2 mM l-glutamine, and 1% penicillin-streptomycin. All plates were coated with 30 μg/ml collagen (BD Biosciences, San Jose, CA), 10 μg/ml human fibronectin (BD Biosciences), and 100 μg/ml bovine serum albumin (BSA) prior to seeding. HEK-293 CFTR-wt cells were used between passages 8–20 whereas 16HBE14o− were used between passages 10–25.

Construction of M2 mutants.

M2 Udorn wild-type pcDNA3.1 and M2 Udorn pVitro green fluorescent protein (GFP) plasmids were constructed as described previously (30). M2 mutants M2-G34V, M2-V27F (5), and M2-ER (endoplasmic reticulum) in pcDNA3.1 and pVitro GFP plasmids were constructed by use of Genetail or Geneart site-directed mutagenesis kits (Life Technologies, Carlsbad, CA) and verified by sequencing. M2-ER contains a missense mutation replacing amino acids 94 and 95 with a dilysine (KK) sequence as previously described by Sakaguchi et al. (43) that prevents trafficking beyond the ER.

cRNA.

cRNAs encoding the human α, β, and γ subunits of the epithelial sodium channel (ENaC), CFTR, M2, and M2 mutants were generated previously as described (30). Briefly, plasmids [CFTR in pcDNA3 (45), αβγ ENaC, M2, and M2 mutants in pcDNA3.1] were linearized by restriction digestion and sense RNA was in vitro transcribed from an upstream T7 promoter by using the mMESSAGE mMACHINE kit (Ambion Applied Biosystems, Austin, TX) according to the manufacturer's instructions. cRNA encoding the transient receptor potential vanilloid 5 (TRPV5) ion channel was generously provided by Dr. Ji-Bin Peng (58).

Microinjection of oocytes with cRNAs.

Oocytes were injected as previously described with minor variations (24, 30, 32). Oocytes were injected with 0.5 ng of human CFTR cRNA followed 24 h later by injection with either 1) 50 nl of varying concentrations of M2 cRNA or 2) 50 nl of RNase-free water. All experiments were performed 72 h post-CFTR injection.

Detection of whole oocyte Cl⁻ currents.

Oocytes were transferred to an experimental chamber containing ND96 with the following composition (in mM): 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl₂, and 5 HEPES (pH 7.6). Whole oocyte currents were measured by using an oocyte voltage clamp amplifier (OC725C; Warner Instruments, Hamden, CT) as previously described (30). Sampling protocols were generated by pCLAMP 8.0 (Axon Instruments, Union City, CA). Currents were sampled at the rate of 1 kHz, filtered at 1 kHz, and were simultaneously stored electronically and displayed in real time on a chart recorder. CFTR channel activity was induced by addition of 10 μM forskolin and 100 μM IBMX in the perfusing solution. In a number of cases oocytes were preincubated with forskolin and IBMX for 20 min prior to the measurements.

Current-voltage (I-V) relationships were determined by applying a pulse protocol ranging from −140 mV to +60 mV in 20 mV increments for 600 ms from a holding potential of −40 mV. In some experiments continuous recordings of currents were obtained while holding oocyte membrane potential at −40 mV and stepping from −140 mV to +60 mV every 5 s. Currents were inhibited by addition of GlyH-101 (20 μM), a specific CFTR inhibitor (36). GlyH-101-sensitive current was calculated by subtracting remaining current after GlyH-101 from forskolin- and IBMX-stimulated currents at +60 mV membrane potential.

Transfection of cells with cDNAs.

Both HEK-293 and 16HBE14o− cells were transfected with wild-type and mutant M2-GFP or GFP cDNAs by using XtremeGene HP transfection reagent (Roche Applied Science, Indianapolis, IN) at a 1:1 ratio of DNA to transfection reagent according to manufacturer's instructions.

Measurement of whole cell currents in cells.

Forty-eight hours posttransfection with cDNAs, individual cells expressing green fluorescence were patched in the whole cell mode (30) by using pipettes with an electrical resistance of 3–5 mΩ. For HEK cells we used pipette and bath solutions of the following ionic compositions (in mM): pipette: 135 KCl, 6 NaCl, 1 MgCl₂, 0.5 EGTA, 10 HEPES (pH 7.2); bath: 135 NaCl, 2.7 KCl, 1.8 CaCl₂, 1 MgCl₂, 5.5 glucose, and 10 HEPES (pH 7.4). For 16HBE14o−: pipette: 135 CsCl, 10 KCl, 2 MgCl₂, 0.1 EGTA, 5.5 glucose, 4 MgATP, and 10 HEPES (pH 7.2); bath: 145 CsCl, 2 MgCl₂, 2 CaCl₂, 5.5 glucose, 10 HEPES (pH 7.4). Potassium and Na⁺ were replaced with Cs in these experiments to block contributions of K⁺ channels to whole cell currents. Cells were perfused with bath solutions containing forskolin (10 μM) and IBMX (100 μM). I-V relationships were obtained by applying a voltage from −80 mV to +60 mV in 20-mV increments for 500 ms from a holding potential of 0 mV. Time course recordings were obtained by applying consecutive voltages of −60 mV and +60 mV every 5 s from a holding potential of 0 mV. Inhibitor-sensitive currents in HEK-293 CFTR-wt were calculated by subtracting remaining currents after perfusion with recording solutions containing forskolin, IBMX, GlyH-101 (20 μM), and PPQ-102 (10 μM) (49) from forskolin (10 μM)- and IBMX (100 μM)-activated currents.

Isolation of total and membrane proteins from Xenopus oocytes.

Whole oocyte lysate (WOL) and total membrane proteins (TMP) were obtained as reported elsewhere (30). Briefly, for WOL, 10–20 oocytes were washed three times in ND96, followed by mechanical homogenization in 1 ml RIPA buffer (Thermo Scientific) supplemented with 1 tablet of Complete Mini EDTA-free Protease Inhibitor Cocktail Tablets (Roche Applied Science, Indianapolis, IN) per 10 ml RIPA. For TMP purification, 40–60 oocytes were washed three times in ND96 and mechanically homogenization in 1.2 ml ND96 supplemented with one tablet of Complete Mini EDTA-free Protease Inhibitor Cocktail Tablets per 10 ml ND96. After centrifugation of homogenized oocytes at 4°C for 10 min, the yolk at the surface was aspirated and the supernatant was transferred to a new tube. The solution was then centrifuged at 4°C for 30 min to pellet membrane fragments. The supernatant was then discarded and the pellet was resuspended in RIPA buffer. Both WOL and TMP samples were rotated in a Labquake Rotisserie Shaker (Thermo Scientific) for at least 1 h and centrifuged again at 17,000 g at 4°C for 10 min, and the supernatant was collected.

Isolation of total and biotinylated proteins from HEK-293 cells.

Cells were washed three times with PBS and lysed in RIPA buffer supplemented with protease inhibitors. Cells were then rotated for at least 30 min and centrifuged at 17,000 g for 10 min at 4°C, and the supernatant was collected. For biotinylation, cells were washed three times with PBS, and incubated with EZ-Link Sulfo-NHS-SS-Biotin (Thermo Scientific) according to manufacturer's instructions in PBS at pH 8. Cells were then incubated on ice for 15 min, and quenched three times with 50 mM Tris buffer at pH 7.4. After being washed with PBS three times, cells were lysed with RIPA buffer. Biotinylated proteins were captured with NeutrAvidin-coated Sepharose beads (Thermo Scientific) O/N at 4°C. Beads were then washed five times with RIPA to remove unbound protein and eluted with SDS sample buffer at 37°C for 30 min.

Western blotting.

Protein concentrations were measured using a bicinchoninic acid (BCA) assay kit (Thermo Scientific). Oocyte and cellular proteins were incubated at 37°C for 30 min in SDS buffer. Samples were then subjected to SDS-PAGE on Tris·HCl Criterion precast gels (Bio-Rad Laboratories, Hercules, CA) and transferred to polyvinylidene fluoride membranes (Bio-Rad Laboratories). Membranes were then blocked with 5% nonfat dry milk in PBS and probed with antibody against CFTR (596 provided by John Riordan, Ph.D., University of North Carolina-Chapel Hill, Cystic Fibrosis Foundation Therapeutics). M2 expression was measured by probing with an antibody against the NH₂-terminus of influenza A M2 (14C2; Novus Biologicals, Littleton, CO). Membranes were probed with a secondary antibody conjugated to horseradish peroxidase (HRP) and the signal detected by the addition of chemiluminescence substrates and exposure to Fuji medical X-ray films (Fujifilm Medical Systems, Stamford, CT). Densitometry was obtained by using AlphaView SA software (Proteinsimple, Santa Clara, CA); signals were normalized to total protein, as quantified by Amido black staining (Sigma-Aldrich), to control for loading.

Luminescence measurements for surface M2 detection.

Oocytes were fixed in 700 μl 4% paraformaldehyde for 15–30 min and blocked with ND96 containing 1% BSA overnight. Oocytes were then incubated with primary M2 antibody 14C2 for 1 h, washed with ND96 + 1% BSA blocking solution, incubated with secondary HRP-conjugated antibody for 1 h, and washed with blocking solution for 1 h. Oocytes were then placed into individual wells of a 96-well black round-bottom plate. After addition of substrate (Millipore Immobilon Western; Millipore), relative luminescent units were obtained by use of a 96-well Luminometer (BMG Labtech FLUOstar OPTIMA, Ortenberg, Germany.)

Statistics.

Values are presented as means ± SE. Student's t-test was used for statistical analysis for comparison of two groups. The comparison of statistical significance among three or more groups was determined by one-way analysis of variance (ANOVA) followed by pairwise comparisons with Tukey's test using OriginLab Data Analysis Software (Northampton, MA).

RESULTS

Influenza M2 protein expression decreases CFTR activity in Xenopus oocytes.

To determine the effect of M2 on CFTR, we injected oocytes with cRNA encoding CFTR followed 24 h later with the injection of either water or cRNA encoding M2. We then measured whole cell currents using the double-electrode voltage-clamp technique at 72 h post-CFTR injection. CFTR and M2 coinjected oocytes had a similar morphology to water-injected oocytes. Basal currents observed in CFTR and M2 coinjected oocytes were very small and similar to uninjected controls (Fig. 1A). When M2-injected oocytes were perfused with ND96 at an acidic pH (5.5), an amantadine-sensitive inward current specific to M2 (Fig. 1B) was observed. Addition of forskolin and IBMX in the recording solution activated a nonrectifying current that was inhibited by the CFTR inhibitor GlyH-101. Oocytes coinjected with M2 had significantly less forskolin-stimulated activity than water-coinjected control oocytes (Fig. 1A). Since GlyH-101 is a negatively charged molecule, inhibition was more effective at positive membrane potentials [Fig. 1C; nonlinear effects (open squares)] as previously demonstrated (36). Thus in subsequent experiments we calculated the GlyH-101-sensitive current at +60 mV membrane potential after stimulation with forskolin and IBMX. Coexpression of CFTR with influenza M2 decreased the GlyH-101-sensitive current in a dose-dependent manner (Fig. 1D).

Fig. 1.

Influenza M2 protein (M2) inhibits CFTR expression and activity. Oocytes were injected with cRNA encoding CFTR and coinjected 24 h later with either cRNA encoding M2 or vehicle (H₂O). In all oocyte experiments, recordings were obtained 72 h post-CFTR injection, 48 h post-M2 coinjection. Current measurements were performed with 2-electrode voltage-clamp technique. GlyH-101-sensitive currents were calculated by subtracting remaining current after GlyH-101 from forskolin- and IBMX-stimulated currents at +60 mV membrane potential as described in methods. A: representative traces of currents at membrane potentials of +60 and −140 mV in oocytes coinjected with cRNA encoding CFTR and either water (black trace) or M2 (gray trace). Oocytes were initially perfused with recording solution at neutral pH (7.6), and then perfused with 10 μM forskolin and 100 μM IBMX (FORSK) to activate CFTR. CFTR current was confirmed by perfusion with ND96 pH 7.6 + 10 μM forskolin, 100 μM IBMX, containing 20 μM CFTR-specific inhibitor GlyH-101 (GlyH); this GlyH-101-inhibited portion of the total current is referred to as GlyH-101-sensitive current (GlyH SENS). B: measurement of pH-activated M2 H⁺ currents in oocytes injected with CFTR and M2. Current (I)-voltage (V) plots of baseline currents of oocytes perfused with recording solution at neutral pH (7.6) (black triangles), pH 5.5 (gray inverted triangles), or pH 5.5 plus amantadine (100 μM) (open triangles) (n = 17). C: measurement of CFTR activity in oocytes coinjected with CFTR and either M2 or water. Current-voltage plot of forskolin- and IBMX-stimulated (black squares, n = 119) and GlyH-101-inhibited (open squares, n = 119) current in CFTR + water-injected oocytes and forskolin-stimulated current in CFTR + M2-injected oocytes (gray diamonds, n = 110). D: oocytes were coinjected with water or varying concentrations of M2 as indicated. GlyH-101-sensitive currents (I_GlyH-101) were expressed as percent of corresponding control values (H₂O injected) (% Cont) (n = 52, 10, 41, 52, 15). E: Western blot and densitometry of oocytes coinjected with CFTR and either M2 or water. Western blot: total membrane protein in oocytes that were 1) uninjected or coinjected with 2) CFTR and water or 3) CFTR and M2-wt. B, immaturely glycosylated CFTR band B protein; C, maturely glycosylated CFTR band C protein; M2, M2 protein. Densitometry: CFTR band B and band C protein in CFTR + water (n = 6)- and CFTR+M2 (n = 6)-coinjected oocytes. Values were normalized to CFTR+H₂O-injected controls as described in methods. All values are means ± 1 SE. **P < 0.01, ***P < 0.001, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups.

We next sought to analyze changes in CFTR expression by measuring TMP. Oocytes coinjected with M2 had decreased expression of mature, fully glycosylated CFTR protein (Fig. 1E). This effect was not due to a decrease in total protein expression since CFTR band B expression was actually increased (Fig. 1E). These results suggest that M2 causes a defect in CFTR maturation or degradation.

M2 inhibits CFTR in a specific manner.

To investigate whether the M2 inhibition of CFTR was specific, we coinjected oocytes with influenza virus matrix protein 1 (M1), which is encoded by the same viral gene segment as M2 (44). Whereas M2 significantly inhibited CFTR expression and activity, coinjections of identical concentrations of M1 cRNA had no effect on either of these values (Fig. 2, A and B), suggesting that saturation of the translational machinery is unlikely to be the cause of CFTR inhibition. To determine whether M2 inhibited the trafficking of any plasma membrane protein in this system, we injected oocytes with a cRNA encoding the calcium channel, TRPV5, and measured currents. By perfusing the oocytes with a divalent cation-free medium we were able to measure passage of monovalent cations through the channel (Fig. 2C) (58). We then measured the lanthanum-inhibited portion of these currents in the presence and absence of various concentrations of injected M2 cRNA. M2 had no significant effect on TRPV5 activity at concentrations shown to completely inhibit CFTR (Fig. 2D compared with 1D), suggesting that M2 protein did not have a general inhibitory effect on ion channel expression in Xenopus oocytes.

Fig. 2.

Alteration of CFTR activity by M2 is specific. A: GlyH-101-sensitive currents at 72 h post-CFTR injection in oocytes coinjected with CFTR and 24 h later with either water (n = 7), M1 (n = 7), or M2 (n = 7). B: Western blot of Xenopus oocytes total membrane protein in 1) uninjected oocytes and oocytes coinjected with 2) CFTR+water, 3) CFTR+M1, or 4) CFTR+M2. C: representative trace of TRPV5 current in Xenopus oocytes at +60 and −140 mV 72 h postinjection. Current was activated by the removal of divalent cations and inhibited by the calcium channel blocker lanthanum (Lanth; 100 μM). D: Xenopus oocytes were injected with TRPV5 and 24 h later were coinjected with either water or varying concentrations of M2. Lanthanum-sensitive current at −140 mV was recorded 72 h post-TRPV5 injection (n = 9, 9, 9, 9). All values are means ± 1 SE. ***P < 0.001, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups.

The M2 inhibition of CFTR is not dependent on generation of reactive intermediates.

Because of the M2-mediated inhibition of ENaC expression by reactive oxygen intermediates (ROS) (30), we tested whether ROS also played a role in M2 decrease of CFTR activity. ROS have been reported to either enhance (13) or inhibit (8, 10) native CFTR expression and activity in a concentration- and time-dependent manner. Although incubation of oocytes expressing ENaC with 500 μM GSH ester (a highly efficient source of reduced glutathione that is membrane permeable) rescued amiloride-sensitive currents in M2-coinjected oocytes (30), it did not prevent the inhibition of CFTR activity (Fig. 3A). Similarly, incubation of CFTR-injected oocytes with the SOD/catalase mimetic EUK-134 (100 μM) (4) prior to M2 injection did not mitigate the decrease of CFTR current, suggesting that M2-generated ROS were not responsible for CFTR inhibition (Fig. 3B).

Fig. 3.

Effect of antioxidants on M2-mediated decrease in CFTR activity. All currents were measured 72 h post-CFTR injection. A: total current in oocytes coinjected with CFTR and either water or M2 followed by incubation with either 500 μM GSH ester or vehicle (Veh) for 48 h (n = 15, 15, 15, 15). B: GlyH-101 sensitive current in oocytes coinjected with CFTR and either water or M2 followed by incubation with either 100 μM EUK-134 or vehicle for 48 h (n = 11, 11, 10, 9). All values are means ± 1 SE. ***P < 0.001, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups.

M2 H⁺ activity is necessary for the inhibition of CFTR.

To determine the region of M2 essential for CFTR inhibition by M2, we utilized a previously generated cRNA encoding a 62-amino acid COOH-terminal truncated M2 mutant (30). This construct inhibited CFTR activity to the same extent as M2 (data not shown). To investigate whether the ability of M2 to transport H⁺ played a role in the inhibition of CFTR we incubated oocytes with the M2 ion channel inhibitor amantadine (41). Incubation with 500 μM amantadine 1 h after M2 injection rescued CFTR activity and the expression of fully glycosylated CFTR protein while having no significant effect on CFTR in water-coinjected oocytes (Fig. 4, A and B). To further confirm the role of H⁺ channel activity in the inhibition of CFTR, we injected oocytes with cRNAs encoding M2 with decreased ability to transport H⁺ (M2-G34V and M2-V27F) (1, 5). These constructs had similar levels of expression and trafficked to the plasma membrane to the same extent as the M2 wild-type protein (Fig. 4C). Whereas M2-wt-injected oocytes exhibited large inward H⁺ currents when perfused with pH 5.5, oocytes injected with M2-V27F or M2-G34V exhibited little or no current respectively (Fig. 4D). When CFTR was coexpressed with M2-G34V, there was little alteration in the GlyH-101-sensitive current, indicating that M2 channel activity is required for suppression of the CFTR activity. M2-V27F also inhibited CFTR activity to a much lesser extent than the wild-type M2 (Fig. 4E). To ascertain the role of M2 ion channel activity on the inhibition of ENaC, we coinjected Xenopus oocytes with αβγ ENaC and M2-G34V. In contrast to what was seen for CFTR, ENaC activity was inhibited to a similar extent by M2 and M2-G34V (Fig. 4F).

Fig. 4.

Role of M2 ion channel activity in CFTR inhibition. A: GlyH-101-sensitive current in oocytes coinjected with CFTR and either water or M2-wt. Two hours post-M2 injection oocytes were incubated in either vehicle, 250 μM amantadine (AMA), or 500 μM amantadine for a total of 48 h and whole cell currents were recorded at 72 h post-CFTR injection (n = 30, 35, 12, 14, 35, 32) B: Western blot and densitometry of Xenopus oocytes total membrane protein in oocytes treated with amantadine as described above. Western blot: CFTR+H₂O (lanes 1–3)- and CFTR+M2 (lanes 4–6)-injected oocytes incubated with vehicle, 100 μM amantadine, or 500 μM amantadine. Densitometry: CFTR band C protein normalized to CFTR+H₂O-injected controls as described in methods. H₂O: CFTR- and water-injected oocytes (n = 6); M2: CFTR- and M2-injected oocytes (n = 6); H₂O AMA: CFTR- and water-injected oocytes + 500 μM amantadine (n = 3); M2 AMA: CFTR- and M2-injected oocytes + 500 μM amantadine (n = 3) C: total M2 protein expression in oocytes injected with M2-wt, M2-G34V and M2-V27F (same gel, truncated for comparison); immunoluminescence measuring M2 protein at the plasma membrane of M2 and M2-G34V-injected oocytes 48 h postinjection (n = 8, 10, 10). RLU, relative light units. D: measurement of pH-induced M2 ion channel current at −140 mV applied voltage. Oocytes were injected with water (n = 16), M2-wt (n = 10), M2-G34V (n = 8), or M2-V27F (n = 11) and perfused with a pH 5.5 recording solution. pH-induced currents were calculated by subtracting current during pH 7.6 perfusion from current when perfused at pH 5.5. E: GlyH-101-sensitive currents in oocytes coinjected with CFTR and either water (n = 119), M2 (n = 110) M2-G34V (n = 77), or M2-V27F (n = 12). F: oocytes were coinjected with cRNAs encoding αβγ epithelial sodium channel (ENaC) and 2 h later with either water, M2, or M2-G34V. ENaC current is the initial current minus the remaining current after the addition of the specific ENaC inhibitor amiloride. All values are means ± 1 SE. ***P < 0.001, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups; #P < 0.05, ##P < 0.01, ###P < 0.001, significantly different from M2 by ANOVA.

Increasing the pH of the secretory pathway is sufficient to decrease CFTR activity.

Influenza M2 protein is known to raise the pH of the secretory pathway during transit to the plasma membrane. To determine whether increased pH in the secretory pathway is sufficient to inhibit CFTR, we incubated CFTR-injected oocytes with either ammonium chloride (NH₄Cl) (3) or concanamycin A (18), two agents that raise the pH in acidified organelles. Treatment of CFTR-injected oocytes with either of these agents over a 48-h period decreased CFTR activity in a dose-dependent manner (Fig. 5, A and B).

Fig. 5.

Altered CFTR activity due to increasing the pH of the secretory pathways. Twenty-four hours post-CFTR injection, various concentrations of NH₄Cl and concanamycin A (ConA) were added to the incubation media. Seventy-two hours after CFTR injection, whole oocyte currents were recorded. A: GlyH-101-sensitive current in CFTR-injected oocytes incubated with vehicle or 50 μM, 500 μM, 5 mM NH₄Cl (n = 15, 15, 15, 14). B: GlyH-101-sensitive current in CFTR-injected oocytes incubated with vehicle or 5 nM, 500 nM concanamycin A (n = 15, 15, 15). All values are means ± 1 SE. *P < 0.05, **P < 0.01, ***P < 0.001, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups.

M2 must traffic past the ER to alter CFTR activity.

Since M2 H⁺ channel activity was necessary to inhibit CFTR, and agents that increased secretory pH reduced CFTR activity, we reasoned that passage of the active M2 channel through the secretory pathway would be necessary for CFTR inhibition. To test this hypothesis, we constructed a previously characterized M2 mutant containing a dilysine sequence that causes the protein to be retained in the ER (M2-ER). When the M2-ER mutant construct was expressed in Xenopus oocytes, it produced a higher amount protein than the wild-type construct but did not reach the plasma membrane (Fig. 6A) and had no detectable H⁺ channel activity (data not shown). Oocytes coinjected with CFTR and M2-ER did not show any loss of CFTR activity (Fig. 6B).

Fig. 6.

Role of M2 trafficking in CFTR inhibition. A: total M2 protein expression (same gel, truncated for comparison) and immunoluminescence of plasma membrane M2 expression in oocytes injected with water, M2, or M2-ER (endoplasmic reticulum) 48 h postinjection (n = 8, 10, 10). B: GlyH-101-sensitive currents in oocytes coinjected with CFTR and either water, M2-wt, or M2-ER (n = 12, 12, 12). C: Western blot and densitometry of total membrane protein in oocytes coinjected with CFTR and M2 or M2 mutants. Western blot: oocytes were 1) uninjected or coinjected with 2) CFTR and water, 3) CFTR and M2-wt, 4) CFTR and M2-G34V, 5) CFTR and M2-ER. Densitometry: CFTR band C protein normalized to CFTR+H₂O-injected controls as described in methods. H₂O: CFTR- and water-injected oocytes (n = 6); M2: CFTR- and M2-injected oocytes (n = 6); M2-G34V: CFTR- and M2-G34V-injected oocytes (n = 3); M2-ER: CFTR-, and M2-ER-injected oocytes (n = 3). D: GlyH-101-sensitive currents in oocytes coinjected with CFTR and either water or M2-wt. Oocytes were incubated with vehicle or indicated concentrations of an inhibitor of the ubiquitin-activating enzyme E1 (PYR-41) for 24 h prior to recording (n = 19, 21, 8, 11, 6, 11, 12, 12). All values are means ± 1 SE. **P < 0.01, ***P < 0.001, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups; #P < 0.05, ##P < 0.01, ###P < 0.001, significantly different from M2-wt by ANOVA.

We next measured total CFTR protein in M2-wt- and M2-mutant-injected oocytes. As previously observed, oocytes coinjected with M2-wt had significantly less band C protein than water-coinjected oocytes. Band C levels in M2-ER-injected oocytes were similar to those in water-injected controls, confirming that inhibition of CFTR activity is likely due to alteration of CFTR maturation by M2 (Fig. 6C). Higher levels of CFTR (as compared with M2) were seen in oocytes injected with G34V, although the differences were not statistically significant.

CFTR activity partially rescued by ubiquitin inhibitor.

To determine whether ubiquitination plays a role in CFTR inhibition by M2, we incubated oocytes with an inhibitor of the E1 ubiquitin pathway (PYR-41) (57). Although the inhibitor had no significant effect on basal CFTR activity in control oocytes, treatment of CFTR-injected oocytes coinjected with M2-wt with PYR-41 for 24 h partially rescued CFTR activity in a dose-dependent manner (Fig. 6D).

Influenza M2 protein expression decreases CFTR activity in HEK-293 stably expressing CFTR.

To confirm the effect of influenza M2 on CFTR activity in a mammalian system, we transfected HEK-293 cells stably expressing CFTR (HEK-293 CFTR-wt) (6) with a bicistronic vector expressing either GFP alone or both GFP and M2 (Fig. 7A). We regularly achieved transfection efficiencies of ∼70–80%. Forty-eight hours posttransfection, we measured whole cell currents in GFP-positive cells using whole cell patch-clamping technique. We observed a pH-stimulated current in the M2-GFP but not in GFP controls (Fig. 7B). Stimulation of M2 prior to activation of CFTR had no effect on CFTR current. CFTR was activated by the addition of forskolin and IBMX and inhibited by the addition of CFTR-specific inhibitors GlyH-101 and PPQ-102, a novel inhibitor that acts on the ATP binding domain of CFTR locking the channel in an inactive state (49). Although both inhibitors decreased CFTR activity separately, the combination of both inhibitors blocked nearly 100% of forskolin-induced currents (Fig. 7C). Similar to our observations in Xenopus oocytes, M2-wt GFP transfection significantly inhibited total forskolin-stimulated current and CFTR inhibitor-sensitive conductance compared with GFP transfection alone (Fig. 7, C and D) at both 24 and 48 h posttransfection.

Fig. 7.

Inhibition of CFTR in HEK-293 CFTR-wt cells. HEK-293 cells stably expressing CFTR were transfected with a bicistronic plasmid expressing either green fluorescent protein (GFP) or M2 and GFP. Forty-eight hours posttransfection, GFP-expressing cells were patched in whole cell mode. Inhibitor-sensitive current was calculated by subtracting remaining current after the addition of specific CFTR inhibitors (GlyH-101 and PPQ-102) from forskolin- and IBMX-activated currents. Inhibitor-sensitive conductance is the mean value of inhibitor-sensitive current divided by the applied voltage. A: image of HEK-293 cells transfected with M2-wt GFP. B: representative trace of whole cell current at +60 mV and −60 mV in HEK-293 CFTR-wt cells transfected with GFP (black trace) or M2-wt GFP (gray trace). Cells were initially incubated in recording solution at neutral pH. Recording solution was switched to pH 5.5 (5.5) to measure M2 activity. Cells were then perfused with recording solution at neutral pH (7.4) + 10 μM forskolin and 100 μM IBMX to measure CFTR activity (FORSK). CFTR activity was confirmed by perfusion with forskolin, IBMX, and specific CFTR inhibitors GlyH-101 (20 μM) and PPQ-102 (10 μM) (INH). C: current-voltage plot of forskolin-stimulated current in GFP-transfected cells (black squares, n = 24), GlyH-101 + PPQ-102-inhibited current in GFP transfected cells (open squares, n = 24) and forskolin-stimulated current in M2-wt GFP-transfected cells (gray diamonds n = 11). D: whole cell CFTR inhibitor-sensitive conductance (G_inhibitor) 24 and 48 h posttransfection with GFP or M2-wt GFP (n = 11, 12, 24, 11). All values are means ± 1 SE. **P < 0.01, ***P < 0.001, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups.

Influenza M2 protein ion channel activity and trafficking are necessary for the inhibition of CFTR activity in HEK-293.

Cells transfected with M2-G34V GFP did not exhibit an increase in current at acidic pH despite similar expression levels to the wild-type protein (Fig. 8, A and B). As shown in Fig. 8C, M2 but not M2-G34V inhibited CFTR activity in HEK cells. To confirm the role of M2 trafficking in CFTR inhibition, we transfected HEK-293 CFTR-wt cells with ER-retained M2 (M2-ER GFP). Although the M2-ER construct was expressed at higher levels than M2, no pH-induced M2-ER activity was observed (Fig. 8D), and M2-ER could not be detected at the level of the plasma membrane (Fig. 8E). Consistent with the oocyte data, M2-ER was unable to alter CFTR conductance (Fig. 8F).

Fig. 8.

Role of M2 ion channel and trafficking in CFTR inhibition in HEK-293 CFTR-wt cells. A: measurement of pH-activated M2 H⁺ currents in HEK-293 CFTR-wt cells transfected with M2 GFP and M2-G34V GFP. Current-voltage plots of baseline currents of cells perfused with recording solution at neutral pH (black triangles, n = 6) M2 (pH 7.4), and with a pH 5.5 recording solution (gray inverted triangles, n = 6) M2 (5.5). M2 ion channel activity of M2-G34V-transfected cells was measured in cells perfused with pH 5.5 recording solution (open triangles, n = 5) M2-G34V (5.5) B: Western blot of M2 protein in whole cell lysate of HEK-293 CFTR-wt cells transfected with M2-wt GFP or M2-G34V GFP. C: whole cell CFTR inhibitor-sensitive conductance in cells transfected with GFP, M2-wt GFP, or M2-G34V GFP 48 h posttransfection (n = 12, 12, 12). D: measurement of pH-activated M2 H⁺ currents in HEK-293 CFTR-wt cells transfected with M2 GFP and M2-ER GFP. Current-voltage plots of baseline currents of cells perfused with recording solution at neutral pH (black triangles, n = 6) M2 (7.4), and with a pH 5.5 recording solution (gray inverted triangles, n = 6) M2 (5.5). M2 ion channel activity of M2-ER-transfected cells was measured in cells perfused with pH 5.5 recording solution (open triangles, n = 10) M2-ER (5.5). E: Western blot of M2 protein in whole cell lysate and at the plasma membrane in HEK-293 CFTR-wt cells transfected with M2-wt GFP and M2-ER GFP. F: whole cell CFTR inhibitor-sensitive conductance in cells transfected with GFP (n = 24), M2-wt GFP (n = 11), or M2-ER GFP (n = 11). All values are means ± 1 SE. *P < 0.05, ***P < 0.001, Significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups. #P < 0.05, ##P < 0.01, significantly different from M2-wt by ANOVA.

Influenza M2 decreases CFTR activity in M2 transfected 16HBE14o− cells.

16HBE14o− were transfected with either GFP or M2 GFP and then seeded onto coverslips 48 h posttransfection. Whole cell currents were obtained from GFP-expressing cells between 1 and 8 h after seeding. 16HBE14o− exhibited a large basal current that was not additionally stimulated by forskolin and IBMX. Although this current was only minimally inhibited by the non-CFTR chloride channel inhibitor DIDS (33), it was inhibited to a large extent by the CFTR-specific inhibitors GlyH-101 and PPQ-102 (Fig. 9, A and B). Both basal currents and inhibitor-sensitive CFTR conductance were decreased in 16HBE14o− transfected with M2-wt GFP (Fig. 9, C and D, respectively) but not M2-ER (Fig. 9D).

Fig. 9.

Inhibition of CFTR activity in 16HBE14o−. 16HBE14o− cells were transfected with a bicistronic plasmid expressing either GFP or M2 and GFP. Forty-eight hours posttransfection, GFP-expressing cells were patched in whole cell mode. Inhibitor-sensitive current was obtained by subtracting remaining current after the addition of specific CFTR inhibitors (GlyH-101 and PPQ-102) from forskolin, IBMX-activated, DIDS-inhibited current. Inhibitor-sensitive conductance is the mean value of inhibitor-sensitive current divided by the applied voltage. A: representative trace of whole cell current at +60 mV and −60 mV in 16HBE14o− cells transfected with GFP (black trace) or M2-wt GFP (gray trace). Cells were initially incubated in recording solution at neutral pH (7.4). Cells were then perfused with recording solution at neutral pH (7.4) + 10 μM forskolin and 100 μM IBMX (FORSK) followed by perfusion in an identical solution with the addition of 100 μM DIDS (DIDS) to inhibit non-CFTR chloride channels. CFTR-specific current was measured by perfusion with forskolin, IBMX, and specific CFTR inhibitors GlyH-101 (20 μM) and PPQ-102 (10 μM) (INH). B: current-voltage plot of baseline (black squares), forskolin- and IBMX-stimulated (open diamonds), DIDS-inhibited (black pentagons) and GlyH-101, PPQ-102-inhibited (open squares) current in GFP transfected 16HBE14o− cells (n = 6). C: current-voltage plot of baseline current in GFP-transfected cells (black squares, n = 6), and M2-wt GFP-transfected cells (gray diamonds n = 15). D: whole cell CFTR inhibitor-sensitive conductance in 16HBE14o− cells transfected with M2-wt GFP, M2-wt, or M2-ER (n = 6, 15, 5). All values are means ± 1 SE. **P < 0.01, significantly different from water-injected control by Student's t-test, or by ANOVA for 3 or more groups; ##P < 0.01, significantly different from M2-wt by ANOVA.

DISCUSSION

Herein we demonstrate for the first time that M2, an integral influenza protein that acts as a H⁺ channel, inhibited CFTR activity and expression in three different systems: 1) Xenopus oocytes injected with human CFTR; 2) epithelial cell (HEK-293) stably expressing CFTR; and 3) human bronchial epithelial cells (16HBE14o−) expressing native CFTR, representative of the lung epithelia likely to be infected by influenza (51). In all cases, the M2-mediated inhibition of CFTR occurred at neutral pH, at which M2 is electrically silent (i.e., does not conduct H⁺). Activation of M2 in either oocytes or mammalian cells by perfusion with an acidic pH (5.5) solution did not enhance CFTR inhibition. These data are consistent with previous studies. For example, treatment of Xenopus oocytes injected with human CFTR with the H⁺ ionophore CCCP resulted in only a 20% loss of CFTR at pH 5.5 (27). Also, acidic pH either potentiated or resulted in a small decrease in CFTR single channel activity (also ∼20%) of inside-out patches in epithelial cells stably transfected with CFTR (11). Therefore, even large transient changes in cytosolic pH cannot explain the drastic alterations in CFTR activity observed in our system.

Instead, our data provide substantial evidence that the M2 inhibition of CFTR was due to alkalinization of secretory organelles that damaged CFTR and targeted the protein for ubiquitination and destruction by the proteasomal or lysosomal systems. The secretory pathway consists of a regulated series of organelles that maintain pH by H⁺ transport through the vacuolar-type ATPase (V-ATPase) (54). As M2 is trafficked through the secretory pathway, it moves through organelles whose pH is sufficiently low to activate the channel. The M2 H⁺ channel unidirectionally conducts H⁺ out of the acidified organelles, increasing their pH. During viral replication, this increase of secretory pH is essential for protecting HA from being prematurely processed (43). Increases in the secretory pH mediated by M2 have been shown to alter glycosylation patterns, slow protein trafficking, and decrease plasma membrane protein recycling (20). Consistent with this hypothesis, the M2 ion channel inhibitor amantadine almost completely prevented the M2-mediated inhibition of CFTR expression and activity in Xenopus oocytes (Fig. 4, A and B). Furthermore, two M2 mutants with attenuated ion channel activity (M2-G34 and V27F) had significantly smaller effects on CFTR activity (Fig. 4E). Finally, incubating CFTR-expressing oocytes with ammonium chloride or concanamycin A, which both have been shown to alkalinize secretory organelles (3, 18), dose dependently decreased CFTR activity (Fig. 5, A and B).

To analyze the necessity of M2 transport through the secretory pathway in the inhibition of CFTR, we expressed an M2 mutant retained in the ER (43) that therefore was unable to traffic into the late secretory compartments like the Golgi apparatus and endosomes (2, 43). It is unlikely that this mutation will inherently alter the ability of the channel to conduct ions since earlier studies have shown that alterations in this region of the COOH-terminus do not affect M2 activity (48). Also, since M2 assembles in the ER, these channels should be fully formed tetramers (7, 34). M2-ER did not alter CFTR activity or the expression of fully glycosylated CFTR despite robust expression (Fig. 6, A and B, and Fig. 8, E and F). On the basis of these results, we believe that influenza M2 has a destabilizing effect on CFTR protein leading to an accumulation of immaturely glycosylated band B protein in M2-transfected cells. This may cause the channel to be targeted for degradation via ubiquitination, as suggested by our experiments partially restoring CFTR activity by inhibiting the ubiquitin pathway. Importantly, the effects of M2 were not indiscriminant since M2 did not inhibit TRPV5 activity.

Although it appears that the mechanism of CFTR inhibition by M2 is similar in mammalian cells and Xenopus oocytes, the extent of CFTR inhibition depends on M2 levels. For example, injecting Xenopus oocytes with five times the amount of M2-G34V cRNA decreased CFTR activity to undetectable levels. The same results were obtained in HEK-293 and 16HBE14o− cells transfected with high concentrations of M2-G34V cDNA (data not shown). The mechanisms by which higher concentrations of M2-G34V inhibit CFTR are not known. It is possible that residual M2-G34V H⁺ channel activity may alter organelle pH at higher concentrations, or H⁺ channel-independent mechanism may be involved. Attempts to transfect 16HBE14o− cells with lower concentrations of M2-G34V cDNA were unsuccessful because of low transfection efficiency.

Finally, our studies show distinct differences between the regulation of CFTR by M2 and earlier studies examining the inhibition of ENaC by M2. Treatments of CFTR and M2 coinjected oocytes with GSH ester and a second ROS scavenger, the cell-permeable catalase/SOD mimetic EUK-134 (4), did not prevent the M2-mediated inhibition of CFTR currents (Fig. 3, A and B). Although studies suggest that CFTR is modified by ROS (8, 10), it is possible that M2-induced ROS did not reach sufficient levels to alter CFTR activity.

In addition to these experiments, and earlier studies directly tying M2 protein to ENaC inhibition, several other studies have examined the effect of influenza on ion transport in the lung. Kunzelmann et al. (28) and Chen et al. (14) showed a direct, PKC-dependent inhibition of the ENaC activity that was due to virus binding to sialic acid receptors in cells. There is also a more general inhibition of ENaC due to an increase in ASL ATP and uridine triphosphate (UTP) through the activation of P2Y purinergic receptors (55). Inhibition of ENaC by influenza resulted in decreased alveolar fluid clearance as well as significantly decreased alveolar gas exchange that occurred before the onset of major morphological changes in the lung (50).

Little is known about the role of CFTR during influenza infection. Kunzelmann et al. (28) studied alteration of forskolin- and IBMX-induced currents at 1 h following influenza infection (before the production of secondary influenza proteins) in tracheal epithelial cells and saw no differences from control values. Studies by Wolk et al. (55) showed that influenza infection decreased alveolar fluid clearance, which was partially reversed by the CFTR-specific inhibitor 172. Increased inflammatory signaling or direct damage to cells during influenza infection leads to increased release of ATP and/or increased breakdown of ATP into adenosine. Adenosine has been show to enhance secretion of chloride by CFTR through the A₁R receptor (19). Additional mechanisms may also play a role in CFTR modulation by influenza. For instance, influenza infection has been shown to enhance nitric oxide levels in the lung (39). Nitric oxide has also been shown to directly alter CFTR activity (8, 12, 13, 25). For these reasons, we chose to directly examine the role of a viral protein on CFTR activity. Our results show that influenza M2 protein expression inhibits CFTR activity by modifying the pH of the secretory pathway at the cellular level.

Interestingly, there is evidence that CFTR regulates transcriptional responses to influenza viral infection. Influenza-infected cystic fibrosis bronchial epithelial cells had fewer genes induced overall than infected normal human bronchial epithelial cells, including inhibited transcription of apoptotic genes. A notable exception was an increased transcription of interleukin 8 (IL-8) (56), a chemoattractant independently shown to be inhibited by CFTR expression (52). Modulation of IL-8 signaling may alter the recruitment of neutrophils and other immune cells in response to viral infection (26).

In summary, our experiments show that M2 decreased the activity of CFTR in exogenous CFTR expression systems as well as in cells expressing native CFTR. Inhibition was dependent on ion channel activity and required trafficking of the M2 protein past the ER through the secretory pathway. In addition, agents modifying pH in the secretory pathway mimicked the effect of M2 expression. Alteration of CFTR activity in the lung epithelium could have important consequences for fluid regulation and immune responses, potentially altering viral propagation and dissemination.

GRANTS

Funding was provided by HL031197 (National Heart, Lung, and Blood Institute), 5U01ES015676 (National Institute of Environmental Health Sciences), and DK60065 (National Institute of Diabetes and Digestive and Kidney Diseases).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

J.D.L., A.L., A.J., J.F.C., J.W.N., and S.M. conception and design of research; J.D.L. and A.L. performed experiments; J.D.L. and A.L. analyzed data; J.D.L., A.L., J.F.C., J.W.N., and S.M. interpreted results of experiments; J.D.L. prepared figures; J.D.L. and S.M. drafted manuscript; J.D.L., A.L., A.J., J.F.C., J.W.N., and S.M. edited and revised manuscript; A.L., A.J., J.F.C., J.W.N., and S.M. approved final version of manuscript.