The epithelial sodium channel (ENaC) is responsible for Na+ and fluid absorption across colon, kidney, and airway epithelia. Short palate lung and nasal epithelial clone 1 (SPLUNC1) is a secreted, innate defense protein and an autocrine inhibitor of ENaC that is highly expressed in airway epithelia. While SPLUNC1 has a bactericidal permeability-increasing protein (BPI)-type structure, its NH2-terminal region lacks structure. Here we found that an 18 amino acid peptide, S18, which corresponded to residues G22-A39 of the SPLUNC1 NH2 terminus inhibited ENaC activity to a similar degree as full-length SPLUNC1 (∼2.5 fold), while SPLUNC1 protein lacking this region was without effect. S18 did not inhibit the structurally related acid-sensing ion channels, indicating specificity for ENaC. However, S18 preferentially bound to the βENaC subunit in a glycosylation-dependent manner. ENaC hyperactivity is contributory to cystic fibrosis (CF) lung disease. Unlike control, CF human bronchial epithelial cultures (HBECs) where airway surface liquid (ASL) height was abnormally low (4.2 ± 0.6 μm), addition of S18 prevented ENaC-led ASL hyperabsorption and maintained CF ASL height at 7.9 ± 0.6 μm, even in the presence of neutrophil elastase, which is comparable to heights seen in normal HBECs. Our data also indicate that the ENaC inhibitory domain of SPLUNC1 may be cleaved away from the main molecule by neutrophil elastase, suggesting that it may still be active during inflammation or neutrophilia. Furthermore, the robust inhibition of ENaC by the S18 peptide suggests that this peptide may be suitable for treating CF lung disease.
- neutrophil elastase
the epithelial sodium channel (ENaC) is the apical conduit for Na+ absorption across a wide range of epithelia including renal, gastrointestinal, and airway (22). The airway surface liquid (ASL) is composed of a mucus layer, which is responsible for trapping both inhaled particles and pathogens such as bacteria and a periciliary layer, which facilitates ciliary beating and serves as a cell surface lubricant (32, 36). Airway epithelia can both absorb Na+ through ENaC and secrete Cl− through the cystic fibrosis transmembrane conductance regulator (CFTR). A precise balance of Cl− secretion and Na+ absorption across the airway is critical for regulating ASL volume. Normal human bronchial epithelial cultures (HBECs) regulate ASL height to an optimal depth of ∼7 μm while cystic fibrosis (CF) HBECs consistently show a much lower depth of ∼4 μm (32). ENaC is negatively regulated by CFTR (44), and in CF patients, the absence of CFTR results in ENaC hyperactivity, leading to uncontrolled absorption of Na+ and ASL volume depletion, which slows or abolishes mucus transport. This lack of mucus transport leads to a failure to physically clear the lungs of accumulated mucus and inhaled pathogens and causes chronic lung infections, which eventually leads to their destruction (40).
ENaC is made up of three subunits, α, β, and γ, which share ∼30% sequence homology (12). Structurally, each subunit is made up of two transmembrane domains, short NH2- and COOH-terminal cytoplasmic tails and a large extracellular loop that contains numerous sites for N-linked glycosylation (11, 37). Activation of this channel occurs through proteolytic cleavage of the extracellular loops of the α- and γENaC subunits by furin-type convertases (14, 25), membrane bound channel activating proteases (CAPs), such as prostasin (CAP1) and TMPRSS4 (CAP2). and/or soluble proteases including the serine proteases trypsin and neutrophil elastase (39). When these proteases are blocked by specific protease inhibitors, such as aprotinin for trypsin-like proteases, ENaC activation is attenuated (8). Alternatively, the cleaved segments of α- and γENaC may bind back into the channel and serve as inhibitory peptides (9, 13). Little is known about the physiological regulation of these key ENaC proteolytic processes. However, we recently hypothesized that a soluble modulator of ENaC existed in the ASL and designed a proteomic screen to identify it (20, 46). Our data indicated that the short palate lung and nasal epithelial clone 1 (SPLUNC1) was the soluble modulator of both ENaC activity and ASL volume and knockdown of SPLUNC1 abolished ENaC regulation in normal HBECs, leading to CF-like ASL volume depletion (20). SPLUNC1 is endogenously secreted into the ASL and functions as an ASL volume sensor: as ASL volume increases, SPLUNC1 becomes diluted, removing the inhibition of ENaC and signaling for absorption to begin; conversely, when ASL volume is low, SPLUNC1 is concentrated, causing less ENaC activity.
SPLUNC1 is a 256 amino acid protein that belongs to the bactericidal permeability-increasing (BPI)-fold containing family A and is also known as BPIFA1, LUNX, PLUNC, and SPURT. SPLUNC1 is expressed in the upper airways and nasopharyngeal regions and may also be expressed in Na+ absorbing tissues including the colon and kidney (20). Based on sequence similarity with BPI-like proteins, SPLUNC1 was hypothesized to be an innate defense protein and indeed, SPLUNC1 has been shown to be both antimicrobial and to reduce surface tension (3, 15, 19, 49). More recently, SPLUNC1 has been proposed to be a multifunctional defense protein since its knockdown in vivo has been shown to decrease mucus clearance (33) as well as to increase Mycoplasma pneumoniae infection (19). Due to the wide variety of functions assigned to SPLUNC1, we set out to identify its ENaC inhibitory domain to better understand how this protein functions and how it interacts with ENaC.
Laevis oocytes were harvested and injected as previously described (18). Oocytes were studied 24 h postinjection using the two-electrode voltage-clamp technique as previously described (20). Where appropriate, oocytes were incubated with S18 or a control peptide, ADG (described below), for 1 h before recording. ENaC activity was determined by measuring the amiloride-sensitive current. In some experiments βENaCS518C was used, which forms ENaCs that are locked into a fully open state with an open probability near 1.0 by exposure to the sulfhydral reactive reagent [2-(trimethyl-ammonium)ethyl]methanethiosulphonate bromide (MTSET). MTSET was added at a concentration of 1 mM to the oocyte bath as previously described (43). Oocyte buffer solutions were used exactly as described previously (20). The University of North Carolina (UNC) Institutional Animal Care and Use Committee approval was obtained for all Xenopus oocyte studies.
Peptides were synthesized and purified by the UNC Microprotein Sequencing and Peptide Synthesis Facility. The sequence of the S18 peptide was GGLPVPLDQTLPLNVNPA. A control peptide of S18, ADG, was made by alphabetizing the sequence. The sequence of ADG was ADGGLLLLNNPPPPQTVV. Both peptides were used with either a free or biotinylated NH2 terminus as needed. Biotinylation had no effect on the ability pf S18 to inhibit ENaC (n = 6).
Electrophysiological measurements of acid-sensing ion channels.
Previously described cell lines expressing human acid-sensing ion channel (ASIC)1a, human ASIC2a, and rat ASIC3 were used in these experiments (34). Electrophysiological measurements were carried out with an EPC10 patch-clamp amplifier (HEKA Electronics, Lambrecht, Germany) as previously described (6).
HEK293T cells were cultured in DMEM/F-12 medium containing 10% FBS, 1× penicillin/streptomycin, 0.2 μg/ml puromycin, and 0.1 mM hygromycin at 37°C, 5% CO2 on six-well plastic plates. Cells were transfected according to the manufacturer's instructions using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) using the DMEM/F-12 media. Cells were transfected when 90–95% confluent with 0.5 μg of plasmid DNA per construct per well and incubated with 5% CO2 at 37°C overnight. Chinese hamster ovary (CHO) cell lines stably expressing human ASIC1a, human ASIC2a, and rat ASIC3 were used in the electrophysiological measurements of ASICs (34).
Human excess donor lungs and excised recipient lungs were obtained at the time of lung transplantation, and cells were harvested by enzymatic digestion as previously described under a protocol approved by the UNC Institutional Review Board (46). HBECs were maintained at an air-liquid interface in a modified bronchial epithelial growth medium with 5% CO2 at 37°C and used 2–5 wk after seeding on 12-mm T-clear inserts (Corning-Costar, Corning, NY; Ref. 35). For experiments performed out of the incubator, both HEK293T cells and HBECs were maintained in a modified Ringer solution as described previously (20).
Peptide pull-down assay and Western blotting.
HEK293T cells were plated in standard plastic six-well plates (Corning Costar) transfected with double-tagged human ENaC subunits with HA and V5 epitopes at the NH2 and COOH termini, respectively, in combination with wild-type untagged subunit cDNA when 90% confluent. When all three ENaC subunits were expressed, 0.5 μg of each subunit were transfected. When expressed individually, 0.75 μg of the subunit were transfected. The transfected cells were lysed 24 h later using NP-40 buffer with 1× complete EDTA-free protease inhibitor (Roche, Basel, Switzerland). The lysate was centrifuged at 16,300 g for 15 min at 4°C, and the supernatant was collected. Protein concentration was determined using the BCA assay, and 500 μg of protein plus 0.25 mg peptide and 100 μl of neutravidin were added to a spin column and rotated end-over-end at 4°C for 24 h (all ThermoFisher Scientific, Rockford, IL). Flow-through was collected by centrifugation at 1,000 g for 30 s. The beads were then washed five times with NP-40 buffer. Bound protein was eluted by boiling at 95°C for 10 min in 75 μl of two times LDS NuPAGE sample buffer with 1× sample reducing agent followed by centrifugation at 16,300 g for 2 min. Samples were resolved on 4–12% Bis-Tris gels in MES and transferred to a nitrocellulose membrane using iBlot, setting P3 for 8 min (Invitrogen, Carlsbad, CA). The membrane was probed using 1:1,000 anti-V5 antibody (Invitrogen) overnight at 4°C in 3% fish gelatin in TBS-T. The blot was then incubated for 1 h at room temperature with an ECL sheep anti-mouse IgG secondary antibody and detected by ECL reagent (ThermoFisher Scientific) or by incubation with a goat anti-mouse IRDye secondary antibody and analyzed by an Odyssey infrared imaging system (LI-COR Biosciences, Lincoln, NE). Pull-down was performed using an NH2-terminally biotinylated version of S18 or ADG. Biotinylation had no effect on the ability of S18 to inhibit ENaC (n = 12).
Peptide pull-down assays were performed as described above. Samples were eluted by the addition of 100 μl of 0.1 M sodium citrate pH 5.5, and 0.1% SDS to the beads and incubating at 100°C for 2 min, followed by centrifugation at 16,300 g for 2 min. The samples were divided equally, and one-half was treated with 1 μl of endoglycosidase H (EndoH) and incubated at 37°C for 2 min. After incubation, all samples were lyophilized and then reconstituted in 30 μl LDS NuPAGE sample buffer with 1× sample reducing agent (Invitrogen). Western blots were completed as described above. A concentration of 5 μg/ml tunicamycin (Sigma-Aldrich, St Louis, MO) was added to the cell transfection media, and the cells were incubated overnight with 5% CO2 at 37°C. The following day the protocol for the peptide pull-down assay was performed as described above.
ASL height measurements.
To label the ASL, 20 μl PBS containing 10-kDa rhodamine dextran (0.2–2 mg/ml; Invitrogen) were added to HBEC mucosal surfaces as previously described (45). When added, peptides with or without 100 nM neutrophil elastase, 1 U/ml aprotinin, activated neutrophil supernatant (ANS; Ref. 23), or 10 μM sivelestat (Sigma-Aldrich) were added to the mucosal surface along with the rhodamine dextran. In some cases, bumetanide (100 μM) was added to the serosal solutions for the duration of the experiment. The HBEC mucosal surfaces were washed with 500 μl PBS for 30 min before experimentation to remove accumulated endogenous SPLUNC1. SPLUNC1 recovery into the ASL was measured over time by lavaging with 200 μl PBS at timed intervals after the initial wash/volume load. To detect SPLUNC1, 20% by volume of the lavage was transferred onto a nitrocellulose membrane using a slot blot apparatus (GE Healthcare). The blot was processed as described above using an anti-hPLUNC primary antibody (R&D Systems, Minneapolis, MN).
Transepithelial potential difference studies.
A single-barreled transepithelial potential difference (Vt)-sensing electrode was positioned in the ASL by micromanipulator and used in conjunction with a macroelectrode in the serosal solution to measure Vt using a voltmeter (World Precision Instruments, Sarasota, FL) as described previously (20). Bumetanide (100 μM) was present for all measurements to inhibit Vt due to Cl− secretion. To calculate the amiloride-sensitive Vt, Vt was measured, amiloride was added as a dry powder in perfluorocarbon, and Vt was remeasured 10 min later. Thus the amiloride-sensitive Vt constitutes the difference between these values.
S18 (100 μM) was dissolved in a buffer containing 10 mM sodium phosphate, pH 7.4 in a 1-mm cuvette. With the use of a Chirascan-plus instrument (Applied Photophysics, Leatherhead, UK), five individual spectra from 185 to 280 nm were recorded at 25 ± 1.0°C and averaged. All measurements were corrected for buffer signal.
Expression and purification of human SPLUNC1.
Full-length SPLUNC1 cDNA was kindly provided by Dr. Colin Bingle. The SPLUNC1-Δ19 and -Δ44 constructs were created by cloning amino acid residues 20–256 or 45–256 out of the SPLUNC1 cDNA for entry into pMCSG7 for protein expression. BL21-CodonPlus competent cells (Agilent Technologies, Santa Clara, CA) were transformed with the expression plasmid and cultured in the presence of antifoam (50 μl), ampicillin (100 μg/ml), and chloramphenicol (34 μg/ml) in LB medium with vigorous shaking at 37°C until an OD600 of 0.6 was attained. Expression was induced with the addition of isopropyl-1-thio-d-galactopyranoside. Bacteria were collected by centrifugation at 4,500 g for 20 min at 4°C. Cell pellets were resuspended in buffer A (20 mM potassium phosphate pH 7.4, 50 mM imidazole, 500 mM NaCl, and 0.02% Na-azide) along with lysozyme, DNase1, and protease inhibitor tablets. Cells were sonicated and cell lysate was separated into soluble and insoluble fractions using high-speed centrifugation. The soluble fraction was filtered then flowed over a Ni-NTA His-Trap gravity column and washed with buffer A. The bound protein was eluted with buffer B (20 mM potassium phosphate pH 7.4, 500 mM imidazole, 500 mM NaCl, and 0.02% Na-azide) and separated using an S200 gel filtration column on an ÄKTAxpress. The COOH-terminal histidine tag was removed with Tev protease as described before experimentation (21). Another round of nickel and HPLC purification removed the tag from the purified SPLUNC1-Δ19. We confirmed that SPLUNC1-Δ19 purified in this fashion inhibited ENaC in CF airways (n = 6).
Neutrophil elastase cleavage of SPLUNC1-Δ19 and analysis by liquid chromatography-electrospray ionization tandem mass spectrometry.
Neutrophil elastase was added to SPLUNC1-Δ19 at a ratio of 1:1,000 (enzyme: protein) followed by incubation for 5, 15, and 60 min at 37°C. Digested samples (5 μl) were introduced to a Q-Tof micro mass spectrometer (Waters, Manchester, UK) via a nano-acquity UPLC (Waters) system as described previously (31). Briefly, the analytical system was configured with a PepMap C18 (LC Packing, 300 μm ID × 5 mm) preconcentration column in series with an Atlantis (Waters) dC18 NanoEase (75 m × 150 mm) nanoscale analytical column. Peptides were separated on the column with a gradient of 5% acetonitrile in 0.1% formic acid to 60% acetonitrile in 0.1% formic acid over 60 min and were directly analyzed without addition trypsin digestion. All data were acquired using MassLynx 4.1 software. The raw data acquired were processed using the ProteinLynx module of MassLynx 4.1 to produce *.pkl (peaklist) files, which are suitable for the tandem mass spectrometry (MS/MS) ions database search via search engines. The data processed were searched against the Uniprot protein database (release 2,011_09) using an in-house MASCOT (Matrix Science, London, UK) search engine (Version 2.2). MASCOT probability-based Mowse individual ion scores >40 were accepted as indicating identity or extensive homology (P < 0.05).
SPLUNC1 peptides in CF patients and donor sputum.
Human sputum samples were obtained as previously described (30). The study protocol was approved by the University of North Carolina Committee on the Protection of Rights of Human Subjects, and informed consent was obtained. Sputum from four CF patients and four healthy donors were pooled, respectively. One milliliter of pooled sputum from each group was diluted in 4 ml PBS. The samples were filtered through a 0.22-μm membrane (Millipore, Bedford, MA). The resulting filtrate were injected onto an Ettan LC chromatographic system (Amersham Pharmacia Biotech, Piscataway, NJ) with a Superdex 200 HR 10/30 chromatography column. The large proteins were separated from the low-molecular-weight peptides with PBS elution at a flow rate of 0.3 ml/min. The peptide pool was dried down 10 times by volume using a vacuum concentrator and then mixed 1:1 with 1% formic acid and subjected to nano-liquid chromatography-electrospray ionization-tandem MS (LC/ESI/MS/MS) and analyzed using the above parameters. N.B.: unlike conventional proteomic studies, the sputum samples were not trypsin digested and only peptides that were endogenously formed were detected while all other protein was excluded from the analysis.
Unless otherwise noted, all data are presented as means ± SE for n experiments. Differences between means were tested for statistical significance using paired or unpaired t-tests when the variances were homogeneously distributed, or in the case of nonhomogeneity of variance, the Wilcoxon rank-sum or Mann-Whitney U-tests were used as appropriate. From such comparisons, differences yielding values of P < 0.05 were judged to be significant. HBECs derived from more than three donors were used per experiment, and experiments using cell lines were repeated on three separate occasions. All analyses were conducted using Instat software (GraphPad, San Diego, CA).
Identification of the SPLUNC1 ENaC inhibitory domain.
SPLUNC1 is a 256 amino acid protein that has a cleavable NH2-terminal signal sequence. We have recently resolved the crystal structure of SPLUNC1 lacking this signal sequence (Δ19-SPLUNC1; Ref. 21). Interestingly, the NH2-terminal region up to residue 43 was different from the rest of SPLUNC1 in that it lacked any discernible structure. However, a peptide, termed S18, which corresponded to this unstructured NH2-terminal region of SPLUNC1 (22GGLPVPLDQTLPLNVNPA39), prevented cleavage of γENaC in HBECs derived from both normal and CF subjects (21). To examine the effects of S18 on HBECs in more detail, we first examined its onset of action. Inhibition of ENaC can cause hyperpolarization of the apical membrane and increase the transepithelial voltage (Vt) caused by Cl− secretion (24). Thus we exposed HBECs to serosal bumetanide to inhibit Cl− secretion before measuring Vt, since under these conditions, the remainder of Vt is largely due to ENaC activity/Na+ absorption (Fig. 1A). The S18 peptide significantly inhibited the bumetanide-insensitive Vt in normal HBECs in a similar time as SPLUNC1Δ19, while an alphabetized version of S18 (ADG, control) was without effect. Furthermore, amiloride strongly inhibited Vt in ADG- and vehicle-treated but not S18- or SPLUNC1-treated cultures. When added mucosally for 1 h, both S18 and SPLUNC1Δ19 inhibited the amiloride-sensitive Vt. In contrast, Δ44-SPLUNC1, which lacks the S18 region (i.e., amino acids G22-A39), had no significant effect, suggesting that the SPLUNC1 ENaC inhibitory domain was located at the NH2 terminus of SPLUNC1 (Fig. 1B).
To confirm that S18 was acting on ENaC, we coinjected oocytes with α-, βS518C-, and γENaC subunits and incubated them with 10 or 100 μM S18 for 1 h. We subsequently observed an ∼2.5-fold decrease in INA (P < 0.05), indicating that we had identified the ENaC inhibitory domain of SPLUNC1 (Fig. 2A). To further explore the effects of S18, MTSET was added during recording to lock the channel into a fully open position and give an approximation of the number of active channels in the plasma membrane (38, 43). MTSET significantly increased ENaC activity approximately sixfold over basal current levels (Fig. 2B). In the presence of S18, MTSET still raised INA above basal levels. However, the increase was significantly less than under control conditions, suggesting that exposure to S18 resulted in a decrease in surface ENaC levels (Fig. 2B). In contrast, when incubated with the control peptide, ADG, no difference in the INA was observed (Fig. 2C).
Structurally related ASICs are not inhibited by S18.
We tested whether SPLUNC1 affected the function of the related ENaC/degenerin family members, the ASICs. To assess possible effects of S18 on the pH dependence of ASIC activation and on current amplitude, channels were activated by a change from pH 7.4 to a pH corresponding to the steep phase of their activation curve, pH 6.6 for ASIC1a and -3 and 4.0 for ASIC2a. At these stimulation pH values, a change in current expression or pH dependence would be readily detected as a change in the measured current amplitude. Stimulations of a 5-s duration were performed every 45 s to allow recovery from inactivation between stimulations. Three control values were obtained before switching to a pH 7.4 solution containing 10 μM S18. The S18 peptide was then present during three stimulation rounds in both the conditioning (pH 7.4) and the acidic stimulation solution before it was washed out. Figure 3A illustrates a typical experiment with an ASIC1a-expressing cell. Average current amplitudes were normalized to the first control value and plotted as a function of time (Fig. 3B). The peptide had no apparent effect on the activation properties of the different ASICs tested (n = 3 by channel type, P > 0.05). No changes in current kinetics were observed (not shown). When exposed to a pH that is less than pH 7.4 but is insufficiently acidic to activate the channel, ASICs inactivate directly from the closed state. This process is known as steady-state inactivation (SSI). Its pH dependence has been determined to have a midpoint of pH 7.2 for ASIC1a, pH 7.1 for ASIC3, and pH 5.6 for ASIC2a (6). To detect changes in the pH dependence of SSI due to the presence of the peptide, cells were incubated 40 s in a conditioning solution with a pH close to the midpoint of SSI in the presence or absence of the peptide (pH 7.1 for ASIC1a and -3 and pH 5.6 for ASIC2a). The conditioning period was followed by activation with an acidic stimulus (pH 6.0 for ASIC1a and -3 and pH 4 for ASIC2a). Figure 3C plots the average values of the current after the conditioning period normalized to the maximal current obtained with a conditioning pH of 7.4. The peptide does not modify the pH dependence of steady-state inactivation of the tested ASICs.
Cleavage of ASIC1a channels by trypsin leads to a shift in the pH dependence of activation to more acidic values (34). ASIC1a was first activated three times by pH 6.6, every 45 s. The conditioning solution was then switched to a pH 7.4 solution containing 40 μg/ml trypsin with or without 10 μM S18, and currents were measured every 45 s at pH 6.6. Due to the trypsin-elicited time-dependent shift in the pH dependence of activation, we observed a gradual reduction in pH 6.6-evoked current in the presence of trypsin, as illustrated in Fig. 3D. This current decrease was not prevented by S18. Since S18 takes up to one h to fully inhibit ENaC (Fig. 1A), we then preexposed ASICs to S18 for 1 h before performing patch clamp studies. This maneuver had no effect on ASIC1a or ASIC3 current densities (Fig. 4). However, ASIC2a was significantly activated by this 1-h pretreatment (Fig. 4).
S18 may interact through the βENaC subunit.
Using coimmunoprecipitation, we have previously demonstrated that SPLUNC1 binds to αβγENaC (20). A pull-down assay was performed using biotin-labeled S18 to determine if the peptide was also able to interact with ENaC in a similar fashion. Three different transfections were performed with one of the three subunits V5-tagged and the other two untagged. The cell lysates were then incubated with biotinylated S18 bound to neutravidin beads. As observed with full-length SPLUNC1, S18 was found to pull-down all three ENaC subunits, although βENaC was more strongly pulled down than either α- or γENaC (Fig. 5A). To further characterize this interaction, each subunit was expressed individually and the pull-down assay repeated. In this case, only βENaC was detected in the elution, suggesting that under these conditions, S18 binds exclusively to the βENaC subunit and not to the α- or γ-subunits (Fig. 5B). When the pull-down assay of the βENaC subunit was performed with ADG, βENaC was not detected in the elution, confirming that S18 is specifically binding to the βENaC subunit (Fig. 5C).
βENaC glycosylation is required for the βENaC/S18 interaction.
The predicted molecular mass of a nonglycosylated βENaC subunit is ∼73 kDa. However, the extracellular loop of βENaC contains 12 possible sites for N-linked glycosylation and βENaC is typically observed at 94–96 kDa due to extensive N-linked glycosylation (25). As seen in Fig. 5, the molecular mass of the βENaC subunit that predominantly interacted with the S18 peptide was ∼94 kDa. To confirm that this was glycosylated βENaC, the elution of the S18/α-, βV5-, and γENaC pull-downs were deglycosylated with EndoH (Fig. 6A). As previously described (25), upon treatment with EndoH, the 94-kDa βENaC band shifted to ∼73 kDa, consistent with the deglycosylation of βENaC (Fig. 6A). This experiment was repeated with the βV5ENaC subunit expressed alone. Upon EndoH treatment, this band shifted to ∼73 kDa (Fig. 6B). EndoH treatment could only be performed once the pull-down assay had been completed; therefore, to test whether S18 was able to interact with nonglycosylated βENaC, we exposed cells expressing either αβV5γENaC or βV5ENaC to tunicamycin, an inhibitor of N-linked glycosylation (48). The pull-down assay was then performed on the tunicamycin-treated cell lysate (Figs. 6, C and D). As seen in the input lanes, treatment with tunicamycin reduced the molecular mass of the βENaC subunit to ∼73 kDa, confirming that the protein was deglycosylated. No βENaC was observed in the elution of the tunicamycin treated pull-downs (Figs. 6, C and D). Combined with the EndoH data, these data indicate that S18 is interacting with a specific, glycosylated form of βENaC.
S18 attenuates ASL hyperabsorption in CF HBECs through ENaC inhibition.
To determine if S18 was capable of inhibiting ENaC-dependent ASL absorption in native airway epithelia, we measured ASL height over time in both normal and CF HBECs after treatment with S18 (Fig. 7, A and B). SPLUNC1 is endogenously secreted by both normal and CF HBECs, which could affect ASL volume regulation, especially in normal HBECs (20). Thus we standardized the mucosal washing/volume-loading protocol accordingly so that every culture had endogenous SPLUNC removed before t = 0. Each culture was left with undisturbed ASL for 24 h. They were then incubated for 30 min with 500 μL PBS followed by two quick successive washes with 500 μl PBS. Then, 20 μl of PBS containing rhodamine-dextran was added as a volume challenge with or without peptide. Slot-blot analysis revealed that it took 24 h for SPLUNC1 levels in the ASL to recover to baseline values (Fig. 7B, inset). Vehicle-treated CF HBECs had a significantly lower ASL height than normal controls. However, the ASL height in CF S18-treated samples, 7.9 ± 0.6 μm, was significantly higher than the CF controls, 4.2 ± 0.6 μm. This increase in ASL height in S18 exposed CF HBECs was comparable to that observed in normal HBECs (∼8 μm). Next, a full dose response was completed to calculate the IC50 by measuring the ASL height of both normal and CF HBECs 6 h after addition of S18 (Fig. 7C). The IC50 was not significantly different between normal and CF HBECs, 0.29 ± 0.19 μM and 0.52 ± 0.23 μM, respectively.
To test whether S18 specifically affected ENaC in normal and CF HBECs, the 24-h transepithelial potential difference (Vt) was measured (Fig. 7, D and E). In the normal HBECs, a basal Vt of −6.6 ± 0.5 mV was observed and this decreased to −10.6 ± 0.7 mV following a brief exposure to trypsin, which is indicative of ENaC activation (46). S18 had little additional effect on the 24 h Vt in normal HBECs. Consistent with our previous observation that CF ENaC remains fully active and nonresponsive to trypsin (46), the CF vehicle-exposed HBECs had an elevated Vt of −15.2 ± 0.9 mV and trypsin had no further effect. In contrast, after 24-h exposure to S18, the CF HBEC Vt was significantly lowered to −8.4 ± 0.9 mV, and a 30-min exposure to trypsin changed the Vt to −11 ± 0.7 mV, suggesting that S18 works through ENaC in CF HBECs. In contrast, ADG had no effect on normal or CF HBEC Vt (both n = 6). To further confirm that S18 functions by inhibiting ENaC hyperabsorption, the ASL height of normal HBECs was measured over time in the presence of bumetanide with or without S18 (Fig. 7F). Serosal bumetanide significantly decreased normal ASL height toward CF levels (i.e., <5 μm). In contrast S18 was able to maintain significantly greater ASL height even in the presence of bumetanide, indicating that S18 increases ASL height by inhibiting absorption not secretion.
S18 has no intrinsic structure.
To better understand how S18 may interact with ENaC, we next looked for intrinsic structure in this peptide. Since the S18 regions does not show up in the crystal structure of SPLUNC1 (21), we next used the program PSIPRED to predict its structure (10, 26). No structure was predicted with PSIPRED, and we also failed to detect any secondary structure by circular dichroism (Fig. 8A). To functionally test this hypothesis, we heat denatured S18 by incubating it at 67°C for 30 min, added it to the mucosal surface of CF HBECs, and then measured ASL height 2 h later. The heat-denatured S18 was able to prevent ASL hyperabsorption to the same extent as the nonheat-denatured peptide (Fig. 8B), suggesting that S18 performs its function with no intrinsic secondary structure.
S18 prevents CF ASL hyperabsorption in the presence of neutrophil elastase.
CF airways are characterized by chronic neutrophilia and display a significantly higher level of neutrophil elastase activity than normal airways (47). Since neutrophil elastase has previously been shown to activate ENaC by cleaving the γ-subunit (39), we tested whether S18 was still capable of inhibiting CF ASL hyperabsorption in the presence of this protease (Fig. 9A). Mucosal addition of S18, but not ADG, to CF HBECs resulted in a significant increase in ASL height. Similarly, aprotinin, which blocks trypsin-like proteases, was also able to maintain ASL height to a similar degree as S18. Interestingly, S18 maintained its ability to prevent CF ASL hyperabsorption in the presence of neutrophil elastase, while aprotinin was unable to prevent neutrophil elastase-induced ASL height depletion. ANS, which is derived from human neutrophils, contains high levels of neutrophil elastase activity (23) but did not diminish the ability of S18 ability to prevent CF ASL volume hyperabsorption. When the neutrophil elastase inhibitor sivelestat, also known as ONO-5046, was added to the ASL along with S18 and NE, no further effect on ASL height was observed (Fig. 8B). Together these data suggest that S18 may be able to prevent ASL hyperabsorption, even in the presence of the high proteolytic activity that is typically present in diseased CF lungs.
Cleavage of SPLUNC1 by neutrophil elastase.
To determine whether peptides corresponding to the NH2-terminal ENaC inhibitory domain of SPLUNC1 may be released from the main molecule by neutrophil elastase, the SPLUNC1-Δ19 recombinant protein, which lacks the NH2-terminal signaling sequence but contains the ENaC inhibitory domain, was exposed to neutrophil elastase for 5, 15, and 60 min and the resulting cleavage products were determined by mass spectrometry. Similar sequence coverage was observed at all three time points ∼80% (Fig. 9C and Table 1). Three peptides were observed by mass spectrometry that spanned the ENaC inhibitory domain (i.e., residues 23–45) at all three time points: two ions corresponding to peptide 23GLPVPLDQTLPLNVNPALPLSPT45 with a m/z of 1,183.742+ and 789.493+ were observed; two ions corresponding to peptide 27PLDQTLPLNVNPALPLSPT45 with a m/z of 1,000.592+ and 667.393+ were observed, and one ion corresponding to peptide 32LPLNVNPALPLSPT45 with a m/z of 723.412+ was observed. In addition to this, in the 15- and 60-min samples one ion corresponding to peptide 20QFGGLPVPLDQTLPL34 with a m/z of 797.992+ was observed. Several other peptides spanning this region were observed in the 60-min sample only including an ion corresponding to peptide 23GLPVPLDQTLPLNVNPA39 with a m/z of 879.532+. All peptides observed in the 5-min sample are listed in Table 1.
Next, we determined which, if any, SPLUNC1 peptides were present in human sputum. In sputum obtained from normal subjects, four peptides were observed that spanned the ENaC inhibitory domain: an ion corresponding to peptide 28LDQTLPLNVNPALPLSPT45 with a m/z of 952.032+, one ion corresponding to peptide 29DQTLPLNVNPALPLSPT45 with a m/z of 895.49+, one ion corresponding to peptide 32LPLNVNPALPLSPT45 with a m/z of 723.412+ and one ion corresponding to peptide 35NVNPALPLSPT45 with a m/z of 561.812+ were observed. Surprisingly, no peptides were observed in CF sputum before residue 46, consistent with our previous studying demonstrating that SPLUNC1's regulation of ENaC is defective in CF airways (21).
We have recently resolved the SPLUNC1 structure to 2.8 Å (21). As expected, SPLUNC1 displays the a “half boomerang” shape that is typical of the bactericidal permeability-increasing protein/lipopolysaccharide-binding protein (BPI/LBP) superfamily (4). In contrast, the NH2-terminal domain (residues 20–43) had no observable structure. Surprisingly, we found that a peptide that reprised this unstructured region, S18 (Residues G22-A39), inhibited ENaC to the same extent as Δ19-SPLUNC1, while SPLUNC1 lacking the S18 region (Δ44-SPLUNC1) was without effect (Fig. 1, A and B). We have previously shown that SPLUNC1 lowers ENaC surface densities (38). Since S18 and SPLUNC1 act over a similar time frame, it is possible that S18 also reduces plasma membrane ENaC levels, rather than acting as a rapid-onset, amiloride-like pore blocker.
In the presence of MTSET, S18 caused a significant decrease in ENaC activity. We have assumed the MTSET has its usual effect in our study, i.e., to increase ENaC Po to ∼1.0 (1). One can divide the MTSET current by the basal current to give estimate the degree of activation by MTSET. Due to the depression of basal currents, MTSET activation was significantly greater in the presence of S18 than in its absence (9 vs. 5.7 -fold, respectively). This ratio has previously been taken as an indicator of channel open probability (Po) and the higher fold increase may indicate that ENaC resides in a lower Po in the presence of S18 (Fig. 2) (1). Therefore, while it is likely that S18 decreases N, we cannot exclude the possibility that S18 may also have an additional effect on Po. Thus an S18-induced reduction in MTSET-activated ENaC currents is consistent with S18 predominantly affecting N. However, absent single channel data, which is not available to us at present, the possibility that S18 also affects Po still remains to be tested.
Despite being structurally related to ENaC, our experiments did not show any inhibitory effect of S18 on ASIC activity following either acute or 1-h exposure to this peptide (Figs. 3 and 4). In the absence of an endogenous regulation of ASICs by proteases in the cell system used (34), a long-term exposure to S18 was not expected to have an effect. However, we did find that the 1-h exposure to S18 significantly activated ASIC2a. ASICs are the target of peptide toxins and of short peptides and several small, charged peptides have been shown to acutely modulate ASIC function (2, 42). Currently, there is no evidence of endogenous regulation of ASICs by proteases, although it is known that ASIC1 pH dependence is changed by exposure to serine proteases (34). The inhibitory toxins and the inactivation-modifying peptides all contain Arg and/or Lys residues that appear to be important for their function (2, 17), while the S18 peptide does not contain Arg or Lys residues. As a caveat, since the experiments were performed with the S18 peptide, we cannot exclude the possibility that other parts of SPLUNC1 might still interact with ASICs.
We found that S18 is able to pull down all three ENaC subunits when they are coexpressed (Fig. 5A). However, when each subunit was expressed individually, only the βENaC subunit was pulled down (Fig. 5, B and C). These data suggest that a complex may form between the α-, β-, and γENaC subunits that is pulled down along with βENaC-S18. Alternatively, it may be possible that when coexpressed, parts of the α- and γENaC extracellular domains may fold into a different conformation that is conducive to α- and γENaC binding to S18 when βENaC is present.
The predominant band observed in the peptide pull-down assay for βENaC, both when all three subunits were coexpressed, and when only the βENaC subunit was expressed, was ∼94 kDa (Fig. 6, A and B), which corresponds to glycosylated βENaC (25). In contrast, the expected molecular mass for nonglycosylated βENaC is ∼73 kDa. When we pulled down βENaC and exposed the elution to EndoH, the 94 kDa βENaC band pulled down with S18 was shifted to ∼73 kDa, indicating that the form of βENaC binding to the S18 peptide is glycosylated with noncomplex, high mannose N-glycans (Fig. 6, A and B). To confirm this interaction, cells were grown and transfected in the presence of tunicamycin, which prevents N-linked glycosylation. Under these conditions, S18 was unable to pull-down the nonglycosylated (i.e., tunicamycin-sensitive) form of βENaC (Fig. 6, C and D). While we cannot exclude the possibility that tunicamycin pretreatment caused mis-folding of the αβγENaC protein, leading to altered S18 binding, taken together, our data lead us to conclude that the S18/βENaC interaction is strongly dependent on the glycosylation state of βENaC.
In addition to inhibiting ENaC in oocytes, S18, but not the control peptide ADG, was also capable of reducing ASL absorption, and over a period of 24 h, S18 restored ASL height in CF HBECs to ∼8 μm, which is comparable to that observed in normal HBECs (Fig. 7, A and B). The transepithelial voltage (Vt) was measured in these same cultures to confirm that S18 was in fact functioning by inhibiting ENaC (Fig. 7, D and E). The thin film Vt reflects a bumetanide-sensitive component (Cl− secretion through CFTR) and an amiloride/trypsin-sensitive component (Na+ absorption through ENaC) (46). In normal HBECs, the bumetanide-sensitive Vt is stable with time, while the trypsin-sensitive component changes with time as ENaC is inactivated by SPLUNC1 (20, 46). In contrast, in CF HBECs there is no bumetanide-sensitive Vt due to the lack of CFTR and these cultures are also insensitive to trypsin, unless pretreated with a protease inhibitor like aprotinin since ENaC is not spontaneously regulated (46). However, in both normal and CF HBECs, S18 caused a decrease in the 24 h Vt and in CF HBECs induced trypsin sensitivity, giving further evidence that S18 is acting through ENaC. Together these data indicate that S18 can serve to restore the regulation of ENaC that is lacking in CF HBECs.
S18 is located at the NH2 terminus of SPLUNC1 in a region that is intrinsically disordered. Using circular dichroism, we demonstrated that the synthesized S18 peptide also displayed no secondary structure (Fig. 8A). This lack of secondary structure was further confirmed by the observation that heat-denatured S18 still prevented ASL hyperabsorption in CF HBECs (Fig. 8B). Other investigators have recently shown that unstructured regions of proteins play an important role in protein:protein interactions and that their flexibility is essential to their function (7). It is also possible that the several proline residues followed by a hydrophobic residue in the S18 sequence are important for its function. Proline-rich motifs have been recognized to play an important role in protein:protein interactions (29), and the role of the prolines, as well as the other residues, in S18 ais currently being explored to further characterize and optimize its inhibitory function.
Carattino et al. (14) previously identified a 26 amino acid peptide that is excised from the αENaC subunit during proteolytic processing, residues D206-R231 (14). When these residues are added back to cleaved ENaC, they inhibit ENaC activity (14). This 26-mer was further refined to a region of 8 amino acids, residues L211-L218 (13). Both the 26-mer and 8-mer αENaC peptides caused a reduction in the NPo of ENaC, and they concluded that the reduction largely reflects a change in channel Po (13, 14, 27). Interestingly, the sequence of the 8-mer, LPHPLQRL, shares 50% sequence identity with residues L24 to T31 (LPVPLDQT) of the S18 peptide. However, the IC50 of the 8-mer was calculated to be 106 and 76 μM in normal and CF HBECs, respectively, which was much higher than the S18 peptide (0.29 and 0.52 μM in normal and CF HBECs, respectively). The difference in the IC50s could be due to the peptide:ENaC subunit interaction, and the αENaC 8-mer is thought to bind to residues within the finger and thumb domains of αENaC (28), while our data show that S18 targets the βENaC subunit. Further characterization of the S18 sequence will be needed to distinguish which residues play the largest role in the SPLUNC1 ENaC inhibitory domain.
High levels of proteolytic activity are typically present in the diseased CF lung, leading to potential protein degradation by neutrophil elastase. In the case of S18, this could lead to decreased efficacy and/or duration of action. However, a single dose of S18 prevented CF ASL hyperabsorption over a 24-h period (Fig. 7A) and was unaffected by either purified neutrophil elastase or activated neutrophil supernatant (Fig. 9A). The ability of S18 to function in a proteolytically active environment makes S18 a strong therapeutic candidate for restoring ASL height in CF patients. Since S18 activity was unaffected by neutrophil elastase, we speculated that a region corresponding to this peptide may be released from SPLUNC1 upon proteolysis. MS analysis revealed that cleavage of recombinant SPLUNC1-Δ19 by neutrophil elastase indeed resulted in the formation of peptides corresponding to the S18 region (Fig. 9C and Table 1), suggesting that release of this region may be a mechanism to make the ENaC inhibitory domain of SPLUNC1 more available during times of inflammation, i.e., when both SPLUNC1 and neutrophil elastase levels are increased. This release of a S18-like region may serve to inhibit ENaC, increase hydration of the mucus layer, and increase mucus clearance, which would be beneficial for innate lung defense. Interestingly, similar peptides to S18 were observed in normal human sputum samples but were undetectable in CF sputum. While this may suggest that endogenous S18-type peptides are formed in normal ASL, whether they are present at sufficient concentration to regulate ENaC in vivo remains to be determined. Free neutrophil elastase is absent from the lungs of healthy individuals. However, the peptides that we detected may arise from the normal breakdown of SPLUNC1 by the extracellular proteases that are present in the airways (30).
SPLUNC1 is detectable in CF HBEC ASL (21) (Fig. 7B) and may even be upregulated in CF airways, (5, 41). However, we do not detect autoregulation of ENaC in CF HBECs (46). CF airways are moderately acidic due to the lack of bicarbonate secretion through CFTR (16). We have recently demonstrated that SPLUNC1 fails to function in the acidic environment seen in CF ASL due to pH-sensitive conformation changes in SPLUNC1, which reduce SPLUNC1ENaC interactions (21). This is likely due to pH-sensitive regions on the main SPLUNC1 protein and the S18 peptide alone is pH independent (21). Consistent with this observation, we found that S18 inhibited Na+/ASL absorption in CF HBECs for >24 h following a single dose. This long duration of action may be due to the ability of S18 to bind directly to βENaC, unlike the small molecule inhibitor amiloride, which is rapidly transported across the epithelia and has a half-life in the ASL of 10 min (45). Furthermore, we speculate that since S18 binds to βENaC that it will not be actively transported across the airways and would avoid the off-target effects seen with amiloride and its analogs, such as inhibition of ENaC in the kidneys, which can lead to natriuresis and reduced blood pressure. In summary, S18 is heat stable, protease resistant, and inhibits ENaC/ASL hyperabsorption for up to 24 h in CF HBECs, suggesting that it has therapeutic potential for the treatment of CF lung disease.
Funding was provided by the Cystic Fibrosis Foundation Resource Development Program Grant R026 (CFF RDP-R026); National Heart, Lung, and Blood Institute Grants PPG-P01-HL-034322, 5 P30-DK-065988-08, R01-HL-103940, and R01-HL-108927; and UNC-Chapel Hill School of Medicine.
R. Tarran is a founder and holds 60% equity in Spyryx. M. J. Stutts is acting Head of Product Development for Spryx. All other authors have no conflict of interest, financial or otherwise.
Author contributions: C.A.H., M.G.B., S.K., S.B., R.C., M.K., A.G.H., M.R.R., M.J.S., and R.T. conception and design of research; C.A.H., M.G.B., O.A., C.D.T., S.B., R.C., W.G.W., A.G.H., M.J.S., and R.T. performed experiments; C.A.H., M.G.B., O.A., C.D.T., S.K., S.B., R.C., M.K., W.G.W., M.R.R., M.J.S., and R.T. analyzed data; C.A.H., M.G.B., C.D.T., S.K., S.B., R.C., M.K., M.R.R., M.J.S., and R.T. interpreted results of experiments; C.A.H., M.G.B., O.A., C.D.T., S.B., and R.T. prepared figures; C.A.H. and R.T. drafted manuscript; C.A.H., M.G.B., O.A., C.D.T., S.K., S.B., R.C., M.K., W.G.W., M.R.R., M.J.S., and R.T. edited and revised manuscript; C.A.H., M.G.B., O.A., C.D.T., S.K., S.B., R.C., M.K., W.G.W., A.G.H., M.R.R., M.J.S., and R.T. approved final version of manuscript.
We thank Dr. Rebecca P. Hughey for helpful discussions on the glycosylation experiments. We gratefully acknowledge the technical assistance of Michael Watson, Hong He, and Yan Dang. We also thank the UNC Cystic Fibrosis Center Molecular and Cell Culture Cores, as well as the UNC Macromolecular Interactions Facility.
- Copyright © 2013 the American Physiological Society