SARS-CoV proteins decrease levels and activity of human ENaC via activation of distinct PKC isoforms

Hong-Long Ji, Weifeng Song, Zhiqian Gao, Xue-Feng Su, Hong-Guang Nie, Yi Jiang, Ji-Bin Peng, Yu-Xian He, Ying Liao, Yong-Jian Zhou, Albert Tousson, Sadis Matalon


Among the multiple organ disorders caused by the severe acute respiratory syndrome coronavirus (SARS-CoV), acute lung failure following atypical pneumonia is the most serious and often fatal event. We hypothesized that two of the hydrophilic structural coronoviral proteins (S and E) would regulate alveolar fluid clearance by decreasing the cell surface expression and activity of amiloride-sensitive epithelial sodium (Na+) channels (ENaC), the rate-limiting protein in transepithelial Na+ vectorial transport across distal lung epithelial cells. Coexpression of either S or E protein with human α-, β-, and γ-ENaC in Xenopus oocytes led to significant decreases of both amiloride-sensitive Na+ currents and γ-ENaC protein levels at their plasma membranes. S and E proteins decreased the rate of ENaC exocytosis and either had no effect (S) or decreased (E) rates of endocytosis. No direct interactions among SARS-CoV E protein with either α- or γ-ENaC were indentified. Instead, the downregulation of ENaC activity by SARS proteins was partially or completely restored by administration of inhibitors of PKCα/β1 and PKCζ. Consistent with the whole cell data, expression of S and E proteins decreased ENaC single-channel activity in oocytes, and these effects were partially abrogated by PKCα/β1 inhibitors. Finally, transfection of human airway epithelial (H441) cells with SARS E protein decreased whole cell amiloride-sensitive currents. These findings indicate that lung edema in SARS infection may be due at least in part to activation of PKC by SARS proteins, leading to decreasing levels and activity of ENaC at the apical surfaces of lung epithelial cells.

  • Xenopus oocytes
  • voltage clamp
  • cell-attached patches
  • amiloride-sensitive currents
  • severe acute respiratory syndrome coronavirus
  • surface epithelial sodium channels
  • H441 cells

the fluid that fills the alveolar spaces in the fetal lung is cleared shortly after birth, mainly as a consequence of active transport of sodium (Na+) ions across the alveolar epithelium. This transport establishes an osmotic gradient that favors reabsorption of intra-alveolar fluid (18). Studies that demonstrate the reabsorption of intratracheally instilled isotonic fluid or plasma from the alveolar spaces of adult anesthetized animals and resected human lungs, and the partial inhibition of this process by amiloride and ouabain, indicate that adult alveolar epithelial cells are also capable of actively transporting Na+ ions (reviewed in Refs. 34, 35).

A variety of studies have clearly established that active Na+ transport limits the degree of alveolar edema under pathological conditions in which the alveolar epithelium has been damaged. For example, intratracheal instillation of a Na+ channel blocker in rats exposed to hyperoxia increased the amount of extravascular lung water (51). Conversely, intratracheal instillation of adenoviral vectors expressing Na+,K+-ATPase genes increased survival of rats exposed to hyperoxia (14). Moreover, patients with acute lung injury who are still able to concentrate alveolar protein (as a result of active Na+ reabsorption) have a better prognosis than those who cannot (47). Results from electrophysiological studies across both confluent monolayers of alveolar type II (ATII) cells mounted in Ussing chambers and alveolar epithelial cells patched in the whole cell or cell-attached modes indicate that Na+ ions diffuse passively into ATII and ATI cells through apically located amiloride-sensitive cation and sodium-selective channels (16, 19, 26, 52) and are extruded across the basolateral cell membranes by the ouabain-sensitive Na+,K+-ATPase (36). The cation channels on the apical surface usually constitute the rate-limiting step in this process, offering more than 90% of the resistance to transcellular Na+ transport in either ATI or ATII cells (25).

Acute respiratory viral infections cause significant morbidity and mortality in both adults and children. For example, respiratory syncytial virus (RSV), a member of the pneumovirus genus of the Paramyxoviridae, is the most common cause of lower respiratory tract infections in infants and children worldwide and also causes community-acquired lower respiratory tract infections among adults (39). Influenza viruses (types A and B) account for more than 50% of all viral pneumonias in adults. Influenza has a high morbidity, affecting 10–20% of the U.S. population, accounting for up to 40,000 deaths annually. There is also a continuing risk of more severe influenza pandemics. Both of these viruses have been shown to impair Na+ transport, albeit by different mechanisms: RSV inhibits Na+-dependent alveolar fluid clearance in Balb/c mice and amiloride-sensitive currents across human airway (H441) cells via increasing levels of UTP, which exit alveolar cells via volume-activated anion channels and act on purinergic receptors (5, 1013). Viral replication is essential for the inhibition of Na+ transport. On the other hand, nonreplicating influenza viruses inhibit epithelial Na+ channels (ENaC) by activating PKC (6, 29). In the case of RSV, amelioration of the decrease of Na+-dependent alveolar fluid clearance in vivo prevented both the RSV-induced hypoxemia and pulmonary edema (10, 11).

Among the multiple organ disorders caused by the newly emerged severe acute respiratory syndrome coronavirus (SARS-CoV), acute lung failure following atypical pneumonia is an often fatal event. The pathology of SARS virus-infected lung tissues includes acute lung injury, characterized by hypoxemia and lung edema. SARS virus has been detected in 100% of the lung tissues from infected individuals. SARS-CoV shares a high degree of sequence identity to group 1 coronaviruses and encodes one polyprotein for virus replication, four structural proteins (spike protein S, envelop protein E, membrane protein M, and nucleocapsid protein N), and eight additional polypeptides, such as the 3C-like protease (43). The S protein, located at the viral surface, plays a key role in cell-viral binding and membrane fusion. The E protein spans the viral shell and is involved in viral envelope formation as well as viral replication.

In the present study we have sought to determine whether SARS-CoV proteins alter ENaC-mediated amiloride-sensitive Na+ transport and, if so, to identify putative mechanisms responsible for this action. Because of well-known interactions between the cystic fibrosis transmembrane regulator (CFTR) with ENaC (3, 7, 21) and possible effects of SARS proteins on either transporter, we opted to coexpress ENaC and SARS proteins in Xenopus oocytes, which normally do not express CFTR. Thus we coinjected human α-, β-, and γ-ENaC (hENaC) cRNAs and cRNAs for either the SARS-CoV S or E proteins into Xenopus oocytes and measured the following variables: 1) whole cell and single-channel amiloride-sensitive currents, 2) total and plasma membrane levels of α- and γ-ENaC by Western blotting, and 3) rates of ENaC endocytosis and exocytosis. We also tested whether expression of either the S or E proteins increased the permeability of oocytes to a variety of cations, as previously suggested (48). Our results indicate that the coexpression of either S or E protein with ENaC significantly decreases both the whole cell and single-channel amiloride-sensitive currents and ENaC protein levels at the plasma membrane and that these changes are partially due to activation of PKC. Furthermore, our data reveal that, in contrast to previous observations in Escherichia coli (48), SARS-CoV S and E proteins do not form cation channels in oocyte membranes.

In the second series of experiments, to demonstrate the relevance of these findings in mammalian cells, we transfected H441 cells, a human airway Clara cell line that expresses highly selective amiloride-sensitive Na+ channels (31), with a plasmid expressing a cDNA for the SARS-CoV E and green fluorescent protein (GFP) fusion protein and then patched H441 cells expressing green fluorescence in the whole cell mode, and current-voltage relationships were measured before and after perfusion with 10 μM amiloride. Our data indicate that similar to our findings in Xenopus oocytes injected with hENaC, transfection of H441 cells with SARS-CoV E protein significantly decreased their amiloride-sensitive currents.


Plasmid construction and cRNA preparation.

Wild-type and mutant cRNAs for human α-, β-, and γ-ENaC as well as SARS-CoV S and E proteins were prepared as previously described (24, 45). γG536C-ENaC was generated by site-directed mutagenesis using the Chameleon double-stranded mutagenesis kit (Stratagene) and sequenced as described previously (20). For some experiments, human α-ENaC cDNA was Flag-tagged at the COOH terminus by PCR and cloned into pcDNA3.1 at KpnI and XbaI sites. An XhoI site was placed between ENaC and Flag-tag.

The open reading frames of SARS-CoV E and GFP cDNA were amplified by PCR from pET-24-SARS-E plasmid and pEGFP-C1 (Clontech, Mountain View, CA), respectively, and cloned in frame into pcDNA3.1 at BamHI and XbaI with an XhoI site between SARS-CoV E and GFP.

Voltage-clamp recordings.

A detailed description of the two-electrode voltage-clamp technique has been published previously (24). In brief, defolliculated oocytes were injected with cRNAs encoding for wild-type α-, β-, and γ-hENaCs and either S or E proteins, dissolved in 50 nl of RNase-free water per oocyte and incubated in half-strength L-15 medium (200 mosmol/kgH2O). Whole cell cation currents were measured using the two-electrode voltage-clamp technique. Briefly, oocytes were impaled with two electrodes with resistances of 0.5–2 MΩ when filled with 3 M KCl. A TEV 200 voltage-clamp amplifier was used to clamp oocytes with concomitant recording of currents. Two reference electrodes were connected to the bath. Sampling protocols were generated using pCLAMP 10.1, and currents at −120, −60, and +40 mV were simultaneously recorded in a chart recorder and stored electronically. Oocytes were perfused with ND-96 medium containing (in mM) 96 NaCl, 1 MgCl2, 1.8 CaCl2, 2 KCl, and 5 HEPES, pH 7.5. In experiments designed to assess permeability to various cations, NaCl was replaced with equal molar concentrations of the appropriate cations. Amiloride-sensitive currents were calculated as the difference currents before and after perfusion of oocytes with amiloride (10 μM). Data were sampled at the rate of 1 kHz and filtered at 500 kHz.

Single-channel patch-clamp recordings.

We recorded single-channel currents from oocytes patched in the cell-attached mode (45). Depolarizing potentials were applied to the patch by an Axopatch 200 amplifier (Molecular Devices, Union City, CA). The patch pipettes were pulled from fire-polished borosilicate glass (WPI Instruments, Sarasota, FL) by using a multistepped micropipette puller (model P97; Flaming/Brown, Berkeley, CA). The pipette tips were fire polished. When filled with pipette solution (100 mM NaCl, 10 mM HEPES-Na, pH 7.5), their resistances were 5–10 MΩ. Currents were collected using the CLAMPEX 10.1 software at a sampling interval of 500 ms. The current traces were filtered at 0.1 kHz with the built-in low-pass filter of CLAMPEX 10.1 and digitized by DigiData 1200 (Molecular Devices). Resting membrane potentials were recorded in the current-clamp mode.

Whole cell patch-clamp recording.

H441 cell culture and patch-clamp recordings were performed as previously described (31). When H441 cells reached ∼80% confluence, they were transfected with GFP-tagged SARS-CoV E cDNA plasmid using the FuGENE transfection reagent (Roche, Basel, Switzerland). Cells were used for imaging and patch-clamp recording 48 h after transfection. Control cells were incubated with transfection reagent without the plasmid. During patch-clamp recording, cells were continuously perfused with a solution containing (in mM) 145 Na-gluconate, 2.7 KCl, 1.8 CaCl2, 2 MgCl2, 5.5 glucose, and 10 HEPES (pH 7.4, 300 mosmol/kgH2O; “bath solution”). Pipettes were made from B150 capillary glass with a two-stage vertical puller (Sutter, Pfalz, Germany). They were back-filled with a solution containing (in mM) 135 K-gluconate, 10 KCl, 6 NaCl, 2 MgCl2, 2 Mg2ATP, 2 Na2GTP, and 10 HEPES (pH 7.2; “pipette solution”). The pipette resistance varied from 5 to 10 MΩ. Cells mounted on the stage of an inverted fluorescent microscope (Olympus IMT-2) were voltage clamped with an Axopatch 200B patch-clamp amplifier (Molecular Devices). H441 cells exhibiting green fluorescence were considered to be successfully transfected with SARS-CoV E protein. Recordings from these cells were compared with those from cells without green fluorescence (controls). Currents were digitized with digital-to-analog and analog-to-digital converters (DigiData 1200A; Molecular Devices), filtered through an internal four-pole Bessel filter at 1 kHz, and sampled at 2 kHz. Inward and outward whole cell currents were elicited by employing a step-pulse protocol from −120 to +80 mV in 20-mV increments every 10 s for 500-ms duration from a holding potential of −40 mV before and after perfusion with amiloride (10 μM). Amiloride-sensitive currents were calculated by subtracting the latter from the former.

Rates of ENaC exocytosis and endocytosis.

To examine the effects of SARS-CoV S and E proteins on ENaC exocytosis, we coinjected oocytes with wild-type α,β-hENaC and γG536C-hENaC and either S or E cRNA. Previous studies have shown that α,β,γG536C channels are almost completely inhibited when oocytes are perfused with medium containing the cell-impermeable thiol reagent methanethiosulfonate bromide (MTSET; Toronto Research Chemicals, North York, ON, Canada), resulting in very low levels of amiloride-sensitive currents (42). The subsequent delivery of unmodified ENaC proteins from the cytosol to the plasma membrane results in a time-dependent increase of the amiloride-sensitive Na+ currents and is a measure of the rate of ENaC exocytosis. We thus incubated these oocytes with MTSET (1 mM), washed the oocytes with ND-96 solution, and measured amiloride-sensitive currents continuously during the next 15 min. To quantify rates of ENaC removal from the plasma membranes (endocytosis), we coinjected oocytes with hENaC and S or E SARS protein cRNAs and measured the amiloride-sensitive currents at 0, 1, and 3 h following addition of vehicle or brefeldin A (5 μM; an inhibitor of exocytosis) into the perfusion recording solution.

Immunoblotting and immunoprecipitation.

Forty-eight hours after injection with either α,β,γ-hENaC alone or α,β,γ-hENaC and S or E cRNAs, Xenopus oocytes were washed in PBS and homogenized with a RNase-free PBS, supplemented with one tablet of Complete mini EDTA-free protease inhibitor cocktail tablets (Roche Applied Science, Basel, Switzerland) per 10 ml of PBS in a Kontes pellet pestle tube (Kontes, Vineland, NJ). The lysate was passed through 20- and 27.5-gauge needles and centrifuged at 3,500 g for 15 min at 4°C. The supernatant was aspirated and centrifuged at 55,000 g for 40 min. The second supernatant was discarded, and the pellet was resuspended in 1× Laemmle sample buffer. Proteins were separated with 8% SDS-PAGE, transferred to nitrocellulose, and incubated with anti-α- or anti-γ-ENaC antibodies (1:1,000; Affinity Bioreagents, Deerfield, IL) in TBS-Tween plus 1% (wt/vol) milk powder for either 2 h at room temperature or overnight at 4°C, as previously described (17). Washed blots were then incubated with peroxidase-labeled antibody in the recommended dilution (1:20,000; Pierce, Rockford, IL) for 1 h. Bound antibodies were detected using chemiluminescence. Western blotting studies with anti-β-ENaC from a variety of commercial sources did not generate reproducible results, and thus these results were not included in this report.

Surface expression of ENaC was examined by a cell surface biotinylation assay as described previously (30). cRNAs were coinjected into Xenopus oocytes. After 48 h, oocytes were mechanically stripped of their vitelline membranes in hypertonic medium and surface proteins were labeled with membrane-impermeable sulfo-NHS-Biotin (Pierce). Oocytes (10 per group) were subsequently lysed and centrifuged at 13,000 × g for 15 min at 4°C. Biotinylated proteins were precipitated with streptavidin-agarose (Pierce) and subjected to SDS-PAGE, followed by Western blotting. Expression of α- and γ-ENaC in the biotinylated fraction was determined using anti-ENaC antibodies, as described above.

To detect possible association of SARS E protein with ENaC subunits, we injected oocytes with Flag-tagged α-hENaC, β- and γ-hENaC, as well as SARS-CoV E-GFP cRNAs. Forty-eight hours following the injections, 50 oocytes from each experiment group were washed three times with 1 ml of TBS and disrupted by being passed through a 20-gauge needle five times in 1 ml of lysis buffer (50 mM Tris·HCl, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, and 1% Nonidet P-40) supplemented with 1× protease inhibitor cocktail (Roche Applied Science). The samples were cleared by centrifugation at 16,000 g for 10 min. Polyclonal antibodies for GFP and tags (Rockland, Gilbertsville, PA) were used to pull down GFP-tagged SARS E protein and Flag-tagged α-ENaC, respectively. Briefly, 5 μg of the above antibodies were added to 500 μl of sample, incubated overnight at 4°C, and subsequently incubated with 50 μl of protein G-agarose for 4 h at 4°C. After being washed three times with PBST (1× PBS + 0.1% Tween 20), captured proteins were eluted by being heated with 100 μl of 1× SDS sampling buffer in the presence of β-mercaptoethanol for 5 min. Proteins were separated by denatured SDS-PAGE and transferred to polyvinylidene difluoride membrane. Membranes containing immunoprecipitated α-ENaC were probed with anti-GFP or anti-γ-ENaC antibodies (Affinity Bioreagents, Golden, CO). Alternatively, membranes containing immunoprecipitated E-GFP protein were probed with an anti-Flag ENaC antibody. Bound primary antibodies were detected with horseradish peroxidase-conjugated donkey anti-goat or goat anti-rabbit secondary antibodies (Pierce) and enhanced chemiluminescence and exposed to X-ray films.

Cell imaging.

H441 cells were transfected with a GFP-tagged SARS-CoV E cDNA plasmid for 48 h and then imaged under an Olympus IX 70 inverted microscope with epifluorescence optics (Olympus, Japan). Images were taken by a Retiga 1300 cooled charge-coupled device, fire wire, high-resolution, monochromatic camera (QImaging). The fluorescent filters used for visualization of green and blue fluorochromes were a 83000 Pinkel filter set (Chroma Tech). For immunofluorescent colocalization imaging of GFP-tagged SARS-CoV E with α- or γ-ENaC, GFP-tagged SARS-CoV E-transfected H441 cells were processed according to standard immunofluorescent staining protocol as described previously (45). Fixed and permeabilized cells were incubated with either a rabbit polyclonal anti-α- or anti-γ-ENaC antibody (10 μg/ml; Calbiochem, Darmstadt, Germany). The secondary antibody was a goat anti-rabbit IgG conjugated to AlexaFluor 594 (1:100; Molecular Probes, Eugene, OR). Confocal microscopy imaging was performed on a Leica DMIRBE inverted epifluorescence/Nomarski microscope outfitted with Leica TCS NT SP1 laser confocal optics (Leica, Exton, PA). The system was equipped with UV, argon-krypton, and helium-neon lasers for the imaging of a wide range of blue, green, red, and far red fluorochromes. Precise control of fluorochrome excitation and emission was respectively regulated by an acousto-optical tunable filter and a TCS SP1 prism spectrophotometer. Spectral separation of fluorochromes was ensured by sequential scanning with single lasers and detection channels for each fluorochrome imaged and by the use of tight band-pass emission windows that do not overlap with the emission spectra of the other fluorochromes imaged. En face optical sections (XY plane) through the Z-axis and side view or sagittal sections (XZ plane) through the Y-axis were generated using a stage galvanometer. During image acquisition, the format size was set to 1,024 × 1,024 pixels for high resolution, and the pinhole aperture was set to 150 μm for a high level of haze removal.

Data analysis.

Amiloride-sensitive currents were calculated by subtracting current values following addition of amiloride from immediately preceding controls using Clampfit 10.1 software. Both whole cell and single-channel current traces carried by cations moving from the external to the intraoocyte sides were depicted as inward (negative) currents, and vice versa. NPo (products of numbers of channels × their open probability) of single channels were calculated from single-channel recordings at least 30 s in duration.

Statistical analysis.

Data are means ± SE unless otherwise stated. Statistical analysis was performed using one-way analysis of variance combined with the Bonferroni test for variance and mean of unpaired data. A P value of 0.05 or less between experimental groups was considered statistically significant.


SARS-CoV proteins decrease ENaC activity.

To test the hypothesis that SARS-CoV structural proteins alter ENaC activity, we coexpressed S and E proteins with α,β,γ-hENaC in Xenopus oocytes. Consistent with the biophysical properties of ENaC, the higher values of the inward current is a hallmark of this inwardly rectifying Na+ channel (22). In oocytes coexpressing hENaC and SARS-CoV proteins, both the total and amiloride-sensitive currents (IENaC) were considerably smaller than those in oocytes injected with α,β,γ-hENaC alone (Fig. 1A). For α,β,γ-hENaC-injected oocytes, IENaC at −120 mV were −9,220 ± 3,332 nA (n = 34), whereas in oocytes coexpressing ENaC with SARS-CoV S and E proteins, IENaC were −2,143 ± 1,004 nA (n = 29, P < 0.001) and −343 ± 518 nA (n = 19, P < 0.001; Fig. 1B), respectively. These results suggest that ENaC activity was decreased by the presence of either S or E proteins in Xenopus oocytes.

Fig. 1.

Coexpression of severe acute respiratory syndrome coronavirus (SARS-CoV) spike (S) and envelop (E) proteins reduce α,β,γ-epithelial Na+ channel (ENaC) activity in Xenopus oocytes. A: representative current traces recorded across Xenopus oocytes injected with α,β,γ-human ENaC (hENaC) cRNAs alone (top), α,β,γ-hENaC and SARS-CoV S protein (middle), and α,β,γ-hENaC and SARS-CoV E protein (bottom). Currents were measured following a change in the membrane potential from the holding potential (−40 mV) to −120 and +100 mV for 500 mS. After the inward and outward currents became stable, oocytes were perfused with 10 μM amiloride, as indicated. Amil, amiloride. B: amiloride-sensitive Na+ currents (IENaC) at −120 mV in oocytes expressing α,β,γ-hENaC alone (αβγ) or α,β,γ-ENaC and either SARS-CoV S protein (+S) or SARS-CoV E protein (+E). IENaC were calculated as described in materials and methods. Values are means ± SE; no. of oocytes is as follows: αβγ, n = 21; +S, n = 29; +E, n = 7. **P < 0.001 compared with the IENaC of αβγ alone. C: resting membrane potentials (in mV) were measured at the current-clamp mode for water (H2O)-injected oocytes and oocytes expressing α,β,γ-hENaC with or without SARS-CoV S or E protein. Bars represent means ± SE; no. of oocytes is as follows: H2O, n = 20; αβγ, n = 21; +S, n = 29; +E, n = 7. Normal resting membrane potential of uninjected oocytes is approximately −25 mV in ND-96 medium. **P < 0.001 for the indicated comparisons.

SARS-CoV proteins depolarize resting membrane potentials of hENaC expressing oocytes.

Measurements of membrane potentials with the current-clamp mode showed values of +7 ± 9 (n = 20) and −24 ± 4 mV (n = 21; Fig. 1C) in oocytes injected with α,β,γ-hENaC or H2O, respectively (P < 0.001). Oocytes coinjected with α,β,γ-hENaC and S or E proteins had membrane potentials of −12 ± 12 mV (n = 29, P < 0.001) and −23 ± 2 mV (n = 7, P < 0.001), respectively. These data are consistent with the higher inhibition of IENaC by E compared with S protein (Fig. 1B).

ENaC protein expression level.

To investigate the possibility that decreased ENaC activity by SARS-CoV proteins may be due to lower levels of ENaC expression, we measured α- and γ-ENaC proteins levels by Western blotting. As shown in Fig. 2A, α-ENaC and γ-ENaC were detected in oocyte lysates as single bands just above 75 kDa, consistent with the molecular mass of the subunits cleaved by proteases (27). Densitometric analysis of these bands (Fig. 2B) revealed that α-ENaC levels were significantly reduced by the coexpression of the SARS-CoV E (51 ± 7%; P < 0.001) but not by the S protein (104 ± 2%, n = 8; values normalized to that of oocytes expressing α,β,γ-ENaC alone). In contrast, γ-ENaC protein expression was markedly decreased in oocytes coexpressing either the SARS-CoV E (73 ± 5%; P < 0.001) or S protein (67 ± 8%; P < 0.001; Fig. 2B). The decrease of both α- and γ-ENaC levels by SARS-CoV E protein is consistent with the higher inhibition of IENaC as well as the higher polarization of membrane potentials of ENaC-expressing oocytes by this protein (Fig. 1, A and B).

Fig. 2.

SARS-CoV proteins reduce total ENaC protein expression in oocytes. A: representative Western blots for α-hENaC (top) and γ-hENaC (bottom) from 4 similar experiments. Twenty oocytes each, expressing α,β,γ-hENaC with or without SARS-CoV S or E protein, were lysed, and proteins were separated with SDS-PAGE, transferred to nitrocellulose, and probed with anti-α- or anti-γ-ENaC antibodies as described in materials and methods. Equal amounts of proteins were loaded in each lane. B: densitometric analysis of Western blots. Values are means ± SE; no. of measurements is as follows: α-ENaC, n = 8 for each bar; γ-ENaC, n = 7 for each bar. **P < 0.001 compared with the corresponding αβγ value for each group.

Whole cell IENaC, recorded by voltage clamp, reflect the total ionic charges flowing through ENaC channels located in the plasma membrane. To detect corresponding changes in the ENaC expression level at the oocyte plasma membranes, we labeled plasma membranes with the membrane-impermeable biotin reagent sulfo-NHS-biotin, captured biotinylated proteins with streptavidin beads, and immunoblotted them with anti α- and anti-γ-ENaC antibodies. We were unable to detect surface α-hENaC with this technique. However, results shown in Fig. 3 indicate that expression of SARS-CoV S or E proteins decreased surface γ-hENaC protein levels to 37 ± 6 or 66 ± 7% of control, respectively (n = 3 for each group), in agreement with the data reported in Fig. 2.

Fig. 3.

SARS-CoV proteins decrease ENaC protein levels at the plasma membrane. A: biotinylated plasma membrane proteins were isolated as described in materials and methods and probed with an antibody to γ-hENaC. Representative Western blots for γ-ENaC are from 3 independent experiments. B: average density (means ± SE) of the γ-hENaC band. **P < 0.001 compared with αβγ alone. See Fig. 2 legend for details.

SARS-CoV protein expression alters ENaC trafficking.

In our next series of experiments, we determined the effects of coexpression of SARS-CoV proteins on ENaC endocytosis and exocytosis rates using previously described methods (15, 42). Specifically, IENaC were continuously monitored in α,β,γG536C-hENaC-injected oocytes following incubation with ND-96 containing MTSET (1 mM), which decreased IENaC to zero. As shown in Fig. 4, A and C, rates of ENaC exocytosis, determined by slopes of the IENaC, were 51 ± 4 nA/min (n = 5) for α,β,γG536C-hENaC alone, whereas in oocytes coexpressing ENaC and either S or E proteins, the rates of ENaC exocytosis were 25 ± 1 and 16 ± 3 nA/min, respectively (n = 5 for each group; P < 0.01 compared with α,β,γ-hENaC alone). These data suggest that S and E proteins significantly decreased the rate of α- and γ-hENaC delivery and insertion into the plasma membranes of oocytes.

Fig. 4.

SARS-CoV proteins alter rates of ENaC exocytosis and endocytosis. A: ENaC exocytosis. IENaC of oocytes injected with α,β,γG536C-hENaC and incubated with methanethiosulfonate bromide (MTSET; 1 mM). The residual MTSET was washed out, and the recorded IENaC reflect newly synthesized ENaC proteins delivered to the plasma membrane. Values are means ± SE for the indicated time points; n = 5. B: ENaC endocytosis. IENaC were measured at time 0 and at 1 and 3 h following pretreatment of oocytes with 5 μM brefeldin A to disrupt delivery of newly synthesized ENaC to plasma membrane. Data are expressed as the ratio of amiloride-sensitive Na+ current at each indicated time point (IENaCt) divided by the corresponding value just before treatment with brefeldin (IENaC0). Data shown in A and B were fitted with linear regression; slopes indicate rates of ENaC exocytosis and endocytosis, respectively. Values are means ± SE for the indicated time points; n = 8. C and D: values are means ± SE of slopes (rates of exocytosis, n = 5; endocytosis, n = 8) shown in A and B, respectively. *P < 0.05 compared with αβγ alone.

We then measured IENaC across oocytes injected with α,β,γ-hENaC following incubation with brefeldin A (5 μM), which interferes with delivery of newly synthesized ENaC to the plasma membrane by disrupting the Golgi network (15). Thus the rate of IENaC over time is indicative of the rate of ENaC endocytosis. Data shown in Fig. 4 indicate that expression of SARS-CoV S protein had no effect on the rate of ENaC endocytosis (α,β,γ-ENaC: 0.18 ± 0.01 IENaC0/h; α,β,γ-ENaC + S: 0.16 ± 0.04; n = 8; P > 0.05; Fig. 4B). On the other hand, expression of SARS-CoV E protein decreased the rate of ENaC endocytosis (0.07 ± 0.05 IENaC0/h, n = 8; P < 0.05 compared with α,β,γ-ENaC alone). Thus we concluded that S and E proteins decrease the cell surface ENaC levels by mainly decreasing ENaC exocytosis.

SARS-CoV proteins decrease ENaC by activating PKC.

PKC is involved in the inhibition of native epithelial Na+ channels in influenza virus-infected ATII cells and intestinal epithelia (6, 28). PKC phosphorylation regulates the activity of cloned ENaC in heterogonous expression systems (1, 6, 44, 50). To examine whether expression of the S and E proteins decreased ENaC activity by activating PKC, we incubated oocytes coinjected with ENaC and SARS-CoV cRNAs in L-15 medium containing Gö-6976 (5 nM), an inhibitor of PKCα/β1 isoforms, at 16°C for 24 h before measuring IENaC. Consistent with data shown in Fig. 1, coexpression of S and E proteins decreased IENaC at −120 mV from −7,579 ± 992 to −3,693 ± 129 (n = 14) and −1,094 ± 158 nA (n = 12), respectively (n = number of oocytes; *P < 0.05 compared with α,β,γ-ENaC alone in both cases; Fig. 5A). Pretreatment of oocytes with Gö-6976 increased IENaC coexpressing either SARS-CoV S or E proteins by ∼50% (from −3,693 ± 129 to −6,304 ± 549 nA, n = 14, and from −1,094 ± 158 to −2,309 ± 226 nA, n = 12, respectively). Similar results were obtained when oocytes coexpressing SARS-CoV S or E proteins with hENaC were incubated with the 20-80 peptide (Fig. 5B), another PKCα/β1 inhibitor. These results indicate that activation of PKCα/β1 isoforms play an important role in the regulation of ENaC by SARS-CoV proteins.

Fig. 5.

SARS-CoV E and S downregulation of ENaC are mediated by activation of isoforms of PKC. A: oocytes injected with α,β,γ-hENaC with or without SARS-CoV S or E protein were incubated 24 h later with the PKC inhibitor Gö-6976 (Go; 5 nM or an equal volume of DMSO, a solvent of Gö-6976) for 24 h in L-15 medium at 16°C, at which time currents were measured before and after the addition of amiloride (10 μM) and IENaC were calculated and expressed as the fraction of IENaC of oocytes expressing αβγ alone (I/Iαβγ). Values are means ± SE at −120 mV; n ≥ 12. *P < 0.05 for the indicated comparisons. B and C: effects of PKC specific isozyme 20-80 peptide (pep; B) and myristoylated PKCζ inhibitory peptide (pep; C) on SARS protein-mediated downregulation of ENaC. Values are means ± SE at −120 mV; n ≥ 12.*P < 0.05 for the indicated comparisons.

Because of previous reports showing the involvement of PKCζ in the regulation of Na+,K+-ATPase (8, 46), we also incubated oocytes coinjected with α,β,γ-hENaC and either SARS-CoV E or S protein with a cell-permeable inhibitor of the PKCζ isoform (myristoylated PKCζ inhibitory peptide, 2 μg/ml). In contrast to what was seen with the inhibitors of the PKCα/β1 isoforms, inhibition of PKCζ had no effect on IENaC in oocytes injected with SARS-CoV S protein; however, inhibition of PKCζ returned IENaC to normal levels in oocytes injected with SARS-CoV E (Fig. 5C). These data clearly indicate that whereas SARS-CoV S activates only PKCα/β1, E activates both PKCα/β1 and PKCζ.

Single-channel activity.

To further assess the effects of S and E protein expression on ENaC function, we recorded single-channel currents from oocytes coinjected with α,β,γ-hENaC and S or E cRNAs using the cell-attached patch technique. As shown in Fig. 6, oocytes injected with α,β,γ-hENaC generally displayed multiple channels in each patch with a single-channel conductance of 4 pS. We therefore calculated the single-channel activity (NPo) in patches from oocytes expressing hENaC alone or hENaC and SARS-CoV proteins. As shown in Fig. 6A, recordings from oocytes coexpressing SARS-CoV S or E proteins had a significantly lower number of open channels (N) per patch. The NPo for α,β,γ-hENaC-expressing oocytes was 3.8 ± 0.1 (n = 12), whereas NPo from oocytes coinjected with α,β,γ-hENaC and S or E proteins were 1.9 ± 0.3 and 0.6 ± 0.1, respectively (P < 0.05 compared with α,β,γ-hENaC; Fig. 6C). Pretreatment of oocytes with the PKCα/β1 inhibitor Gö-6976 increased NPo of ENaC in oocytes coexpressing SARS-CoV S and E proteins to 2.8 ± 0.3 (n = 12) and 2.3 ± 0.3 (n = 12), respectively (P < 0.05 compared with corresponding values from vehicle-treated oocytes; Fig. 6, B and C). These findings, together with data shown in Fig. 5, indicate that S and E proteins inhibit ENaC by activating the PKCα/β1 isoforms. Studies evaluating the effects of PKCζ inhibitors were not attempted.

Fig. 6.

Expression of SARS proteins decreases single-channel ENaC activity. A: representative single-channel current traces digitized from cell-attached patches in oocytes expressing α,β,γ-hENaC with and without SARS-CoV S and E proteins. The patch potential was −80 mV. Dotted lines illustrate open (o) and closed levels (c). B: representative single-channel current traces obtained in oocytes pretreated with PKCα/β1 inhibitor Gö-6979 for 24 h. The scale bar for A is also applied to B. C: single-channel activity (NPo) of 5 oocytes shown in A and B. Values are means ± SE; n ≥ 12 for each group. *P < 0.001 compared with corresponding values of αβγ. #P < 0.05 compared with the corresponding value without Go.

SARS-CoV E and S proteins do not increase cation permeability.

Like other viral proteins, the SARS-CoV E protein has been shown to increase membrane permeability when expressed in E. coli (32). Furthermore, coronoviral E proteins form hydrophilic, amiloride-inhibitable cation channels when reconstituted in lipid bilayers (48). We first examined the hypothesis that SARS-CoV proteins might either function as a H+ channel or stimulate H+ currents when expressed in Xenopus oocytes, as is the case with influenza M2 integral protein(41). Oocytes expressing either S or E proteins were perfused with acidic ND-96 (pH 6.0) containing amiloride (10 μM) to minimize the contribution of Na+ currents. As shown in Fig. 7A, similar levels of H+ currents were recorded at various holding potentials from −140 to +100 mV among uninjected oocytes and oocytes expressing SARS-CoV S or E protein. These data indicate that these proteins do not act as H+ channels. We then examined possible differences in cation permeability in oocytes injected with either SARS-CoV S or E cRNAs. No significant differences in membrane permeability to Na+, H+, K+, Mg2+, Ca2+, or Zn2+ were found compared with corresponding values in H2O-injected oocytes (Fig. 7B). These data indicate that, at least under our experimental conditions, S and E proteins do not form cation-permeable channels in Xenopus oocytes.

Fig. 7.

Cation permeabilities of SARS-CoV proteins heterologously expressed in Xenopus oocytes. A: proton (H+) permeability. Oocytes expressing either SARS-CoV S or E protein were perfused with acidic ND-96 (pH 6.0) containing amiloride (10 μM). Uninjected oocytes or H2O-injected oocytes served as controls. Currents (I) recorded from −140 to +100 mV in an increment of 20 mV were plotted as a function of membrane potential (Vm). B: cation permeability. Oocytes were perfused with Na+ (ND-96 at pH 7.5), K+ (Na+ was substituted with equal moles of K+), Mg2+ (ND-96 + 20 mM MgCl2), Ca2+ (ND-96 + 20 mM CaCl2), and Zn2+ (ND-96 + 1 mM ZnCl2). Average (±SE) currents at −120 mV from oocyte expressing SARS-CoV S and E protein were compared with those from H2O-injected oocytes (n = 6–10).

SARS-CoV E protein decreases amiloride-sensitive currents in human airway cells.

To determine whether SARS-CoV E proteins decrease native ENaC activity in human airway cells, we transfected H441 cells with a plasmid containing a subcloned human SARS-CoV E protein and the GFP reporter gene as described in materials and methods. Approximately 5% of cells were successfully transfected with SARS-CoV E protein after 48 h as indicated by GFP fluorescence (Fig. 8). No obvious cell detachment or cell death was observed posttransfection.

Fig. 8.

Expression of SARS-CoV E proteins in H441 cells. H441 cells were cultured on coverslips and transfected with a plasmid containing green fluorescence protein (GFP)-tagged SARS-CoV E protein at 80% confluence as described in materials and methods. Overlapping bright field and fluorescent images of H441 cells show nuclei stained with Hoechst 33258 and transfected cells (used for patch-clamp recordings) emitting green fluorescence.

Previously, we have shown the presence of 4-pS channels in cell-attached patches of H441 cells with the biophysical characteristics of ENaC (31). As shown in Fig. 9A, the basal and amiloride-sensitive current densities for cells not expressing GFP were −11.4 ± 2.0 and −6.0 ± 1.9 pA/pF, respectively (n = 5), whereas those in GFP-positive cells were −4.7 ± 0.9 and −0.6 ± 0.5 pA/pF, respectively (n = 6; P < 0.01). The linear macroscopic conductance of amiloride-sensitive currents (Fig. 9B) was about 4 pS, indicative of ENaC. SARS-CoV E protein expression reduced the whole cell conductance from 53.2 ± 3.4 to 3.7 ± 0.7 pS (P < 0.01). Transfecting H441 cells with GFP alone did not decrease either the total or the amiloride-sensitive currents (data not shown).

Fig. 9.

Downregulation of native ENaC activity by SARS-CoV E protein in H441 cells. H441 cells were transfected with a plasmid expressing SARS-CoV E-GFP proteins as described in materials and methods. A: basal and amiloride-sensitive (INa) whole cell current densities of H441 cells expressing SARS-CoV E (+E; as indicated by the presence of green fluorescence; see Fig. 8) or control cells (H441) recorded at −100 mV. Values are means ± SE; n = 5 for control and 6 for SARS-CoV E-transfected cells. **P < 0.01, current densities in SARS-CoV E protein-transfected cells compared with those in GFP-negative cells (control). B: current-voltage curves of INa in control cells (▪) and SARS-CoV E protein-transfected cells (•). Whole cell conductances were computed as the slope of the curves by fitting the raw data with linear regression. Values are means ± SE; n = 5 for control and 6 for SARS-CoV E-transfected cells.

SARS-CoV E protein does not coimmunoprecipitate or colocalize with ENaC.

We performed colocalization and coimmunoprecipitation studies to detect possible direct interaction of SARS-CoV proteins with α- and γ-ENaC. In the first series of experiments, we transfected H441 cells with GFP-tagged E cDNA; 48 h later, we fixed the cells and labeled them with anti-α or anti-γ ENaC antibodies, followed by secondary antibodies conjugated with fluorescent probes, and imaged them using confocal fluorescent microscopy. H441 cells expressed significant levels of α-ENaC (Fig. 10, A and C) and γ-ENaC (Fig. 10B). About 5% of H441 cells showed GFP fluorescence, indicative of SARS-CoV E protein expression. SARS-CoV E did not colocalize with ENaC subunits either in the cytoplasm (Fig. 10, A and B) or at the level of the plasma membrane (Fig. 10C). Because of the small numbers of H441 cells transfected with SARS-CoV E, coimmunoprecipitation of SARS-CoV E with ENaC was not attempted.

Fig. 10.

Confocal immunofluorescent imaging of ENaC and SARS-CoV E in H441 cells. H441 cells were cultured, transfected with GFP-tagged SARS-CoV E cDNA plasmid, and incubated with either anti-α-ENaC (A and C) or anti-γ-ENaC antibodies (B), followed by an AlexaFluor 594-conjugated secondary antibody 48 h later. Images were also taken in green channel to show cells expressing GFP-tagged SARS-CoV E protein. Bottom panels of A and B and right panels of C show higher magnification of the areas defined by the corresponding boxes. Merging of ENaC and SARS-CoV E images in either XY plane (A and B) or XZ plane (C) indicates little colocalization (yellow) of these 2 proteins. Scale bars in top panels of A and B and left panel of C represent 20 μm; scale bars in bottom panels of A and B and right panels of C represent 4 and 10 μm, respectively. Images are representative of 20 slides from 2 independent experiments.

In the second series of experiments, we injected Xenopus oocytes with GFP-tagged SARS-CoV E cRNA and Flag-tagged α-hENaC along with β- and γ-hENaC cRNAs. Forty-eight hours later, both SARS E and α-ENaC were detected in oocyte lysates by Western blotting using anti-GFP and anti-Flag antibodies, respectively (Fig. 11A). Although α-ENaC coimmunoprecipitated with γ-ENaC (data not shown), SARS-CoV E protein did not coimmunoprecipitate with α-ENaC, and vice versa (Fig. 11, B and C). Together, these data indicate lack of direct interaction among SARS-CoV E and ENaC.

Fig. 11.

Immunoprecipitation of α ENaC and SARS-CoV E proteins in α,β,γ-ENaC-expressing oocytes. A: detection of GFP-tagged SARS-CoV E and α-ENaC proteins expressed in oocytes. Oocytes were injected with GFP-tagged SARS-CoV E (E-GFP) and Flag-tagged α- and wild-type β,γ-hENaC cRNAs (ENaC) as indicated. Lysates were Western blotted with either anti-Flag (Flag-α-ENaC) or anti-GFP (E-GFP) antibodies to show expression of ENaC or SARS-CoV E protein. B: lysates from ENaC-expressing oocytes were immunoprecipitated (IP) with anti-GFP or anti-Flag antibodies in SDS-PAGE gel, and the blot was then probed with anti-Flag antibody (IB). C: oocytes lysates were immunoprecipitated with anti-Flag and anti-GFP antibodies, and then the blot was probed with anti-GFP antibody. No coimmunoprecipitation between ENaC and SARS-CoV E was seen in either case. Each experiment was repeated at least twice with the same results.


As discussed below, numerous studies have shown that viral infections downregulate lung epithelial cell ENaC activity, which leads to pulmonary edema. However, our results demonstrate for the first time that expression of two viral proteins (SARS-CoV S or E) in Xenopus oocytes injected with α,β,γ-hENaC decreases amiloride-sensitive whole cell currents and single-channel ENaC activities as well as plasma membrane γ-hENaC levels. These effects were caused, at least in part, by activation of PKCα/β1 and PKCζ isoforms. Furthermore, transfection of human airway cells expressing native ENaC and amiloride-sensitive currents with SARS E protein decreased whole cell amiloride-sensitive currents.

We believe that our findings have significant translational implications. The global SARS outbreak of 2003 infected over 8,000 persons worldwide with nearly an 11% mortality from respiratory failure (40, 43, 49). The responsible agent was identified as a novel coronavirus named SARS-CoV. SARS-CoV virus produces a spike protein (S), an envelope protein (E), membrane glycoprotein (M), and a nucleocapsid protein (N) (43). The S protein, located at the viral surface, plays a key role in cell-viral binding and membrane fusion. The E protein spans the virus shell and is involved in viral envelope formation and duplication. SARS-CoV was much more lethal than other coronaviruses such as HCoV-OC43 and HCoV-229E, which were known to cause respiratory infections. Presently, there are no effective antiviral therapies, and only supportive treatment for infected patients is available. Specific therapies aimed at neutralizing S and E proteins may be of value in limiting alveolar edema in SARS-infected patients.

Xenopus oocytes have been used widely as a model to study ion channel structure function relationships (9) as well as to understand how the properties of these proteins may be altered by a variety of agents in the absence of other factors. For example, our previous studies have shown that coexpression of CFTR and ENaC results in decreased ENaC activity, whereas coexpression of CFTR and acid-sensing ion channels (ASIC) increases H+ currents in Xenopus oocytes (21, 23). Because of the aforementioned interaction among channel proteins (such as ENaC and CFTR) and possible effects of SARS-CoV E and S on multiple channels, definitive answers as to whether they downregulate ENaC activity can only be obtained in systems expressing ENaC alone. In this study, we have shown that expression of either SARS-CoV S or E proteins with ENaC alone (i.e., in the absence of other channel proteins that may also modulate ENaC) decreases ENaC activity, albeit to various extents (Figs. 1 and 2). These findings (along with the observed reversibility of this effect by PKC inhibitors) argue against the possibility that the observed decrease in ENaC activity was the result of inhibition of ENaC synthesis due to overexpression of viral proteins. To further exclude this possibility, we coexpressed SARS-CoV S or E proteins with transient receptor potential vanilloid type 6 (TRPV6), a Ca2+ channel protein in Xenopus oocytes. TRPV6 activity was not altered by the E protein and was significantly increased by the S protein (data not shown). Together, our data indicate that the SARS-CoV protein-mediated reduction in ENaC expression and activity was not due to competition for protein synthesis and trafficking.

Previous studies have shown that both native and heterologously expressed ENaC activity is regulated by PKC phosphorylation (44, 50). Our results confirmed that at least three PKC-specific isozymes, namely, α, β1, and ζ, are activated when S and E proteins are expressed in Xenopus oocytes, albeit to different extents: SARS-CoV S activated predominantly the PKCα/β1 isoforms, whereas SARS-CoV E activated the PKCζ isoform. Davis et al. (13) recently reported that respiratory syncytial virus activates PKCζ and, in turn, inhibits amiloride-sensitive Na+ uptake across the distal lung epithelium in RSV-infected Balb/c mice. The PKC pathway also plays a critical role in initiating a series of events leading to downregulation of ENaC following the binding of hemagglutinin residues of inactivated influenza virus to sialic acid residues of ATII cells and airway cells (6, 28).

Protein trafficking as well as phosphorylation can be regulated by PKC activity (33). Our results also demonstrate that both the S and E proteins decrease ENaC protein trafficking to the plasma membrane, resulting in reduced γ-hENaC expression at the plasma membrane. We were unable to detect α- or β-hENaC at the plasma membrane, most likely due to low affinity of commercially available antibodies against these subunits. However, Na+ single channels recorded from cell-attached patches of oocytes injected with either hENaC alone or hENaC plus SARS-CoV proteins had similar conductances (4 pS; Fig. 6). Since changes in ENaC stoichiometry are accompanied by changes in single-channel conductance (4, 35, 37), our data suggest that expression of SARS-CoV protein in oocytes did not alter ENaC stoichiometry at the cell surface. Thus we expect that changes in γ-hENaC were accompanied by similar changes in both α-hENaC and β-hENaC. It is unlikely that the observed changes in surface ENaC can account for the almost complete loss of the whole cell amiloride-sensitive currents noted in Fig. 1. Instead, we believe that decreased ENaC activity was the result of activation of PKC isoforms, as previously described (2, 6, 38).

Expression of the S and E proteins in Xenopus oocytes alone did not alter the cation permeability of the oocyte membrane to Na+, K+, H+, Ca2+, Mg2+ or Zn2+. Although the E protein can function as a cation channel in the planar lipid bilayer system (48), our results indicate that E protein may not act as a cation channel by itself in a model vertebrate cell, the Xenopus oocyte. Our observations are not consistent with a recent report in which E protein expression in E. coli cells altered membrane permeability (32).

Protein-protein interactions have been suggested to be involved in the regulation of ENaC in a variety of systems; for example, a reciprocal interaction between homomeric CFTR and ENaC has been reported, and the intracellular COOH- and NH2-terminal tails of ENaC have been identified as interactive domains regulated by CFTR (21). We also have reported that δ-ENaC subunit interacts with α- and γ-subunits modifying ENaC activity (24). Because of these findings, we sought to identify possible intramolecular interactions between SARS-CoV viral proteins and each ENaC subunit. Our results showed that SARS-CoV E protein does not physically interact with α- and γ-ENaC subunits. On the other hand, SARS-CoV S and E proteins activate PKC, which may contribute to their diverse effects on exocytosis and endocytosis rates of ENaC proteins, in turn, downregulating ENaC channel activity at the whole cell and single-channel levels.

In summary, our studies suggest that SARS-CoV S and E proteins downregulate ENaC expression and activity in both Xenopus oocytes injected with hENaC and human airway cells expressing native ENaC. Thus these studies provide new insight into pathogenesis of pulmonary edema in SARS infections.


This work was supported by National Institutes of HealthGrants HL31197 and U54 ES017218 (to S. Matalon) and HL87017 (to H. L. Ji).


We thank Dr. Peter Smith (Dept. of Physiology, University of Alabama at Birmingham) for providing ENaC cDNAs. We also acknowledge the editorial assistance of Terese J. Potter (Dept. of Anesthesiology, University of Alabama at Birmingham) for editorial assistance in preparing this article.


  • * H.-L. Ji and W. Song contributed equally to this work.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract