The FASEB Journal • Research Communication
Mechano-sensitivity of epithelial sodium channels (ENaCs): laminar shear stress increases ion channel open probability Mike Althaus, Roman Bogdan, Wolfgang G. Clauss, and Martin Fronius1 Institute of Animal Physiology, Justus-Liebig-University Giessen, Giessen, Germany Epithelial cells are exposed to a variety of mechanical forces, but little is known about the impact of these forces on epithelial ion channels. Here we show that mechanical activation of epithelial sodium channels (ENaCs), which are essential for electrolyte and water balance, occurs via an increased ion channel open probability. ENaC activity of heterologously expressed rat (rENaC) and Xenopus (xENaC) orthologs was measured by whole-cell as well as single-channel recordings. Laminar shear stress (LSS), producing shear forces in physiologically relevant ranges, was used to mechanically stimulate ENaCs and was able to activate ENaC currents in whole-cell recordings. Preceding pharmacological activation of rENaC with Zn2ⴙ and xENaC with gadolinium and glibenclamide largely prevented LSS-activated currents. In contrast, proteolytic cleavage with trypsin potentiated the LSS effect on rENaC whereas the LSS effect on xENaC was reversed (inhibition of xENaC current). Further, we found that exposure of excised outside-out patches to LSS led to an increased ion channel open probability without affecting the number of active channels. We suggest that mechano-sensitivity of ENaC may represent a ubiquitous feature for the physiology of epithelia, providing a putative mechanism for coupling transepithelial Naⴙ reabsorption to luminal transport.—Althaus, M., Bogdan, R., Clauss, W. G., Fronius, M. Mechano-sensitivity of epithelial sodium channels (ENaCs): laminar shear stress increases ion channel open probability. FASEB J. 21, 2389 –2399 (2007) ABSTRACT
Key Words: DEG/ENaC family 䡠 patch-clamp 䡠 outside-out 䡠 Xenopus oocyte
Epithelial sodium channels (ENaCs) are a subfamily of ion channels within the degenerin/ENaC (DEG/ ENaC) superfamily (1, 2). These ion channels are found in different sodium-absorbing epithelia, including the epithelium of the colon, lung, and distal nephron; their activity represents the rate-limiting step for sodium uptake, and thus transepithelial water movement. Sodium uptake in these epithelia occurs via a mechanism summarized as the “two membrane hypothesis,” where sodium enters the epithelial cell via apical ENaCs, then is extruded by the basolateral Na⫹/K⫹0892-6638/07/0021-2389 © FASEB
ATPase (3). This electrogenic absorption of Na⫹ creates transepithelial osmotic gradients that represent the main driving force for water movement across epithelia. In accordance with this, ENaC plays a major role in hereditary forms of hypertension as identified in patients with Liddle syndrome (4, 5) and is crucial for removal of fluid from the lungs at birth (6). The ENaC multiprotein complex consists of three homologous subunits called ␣, , and ␥ (7), and it is suggested that the ion channel is assembled with a stoichiometry of 2␣, 1, and 1␥ subunits (5). Each ENaC subunit includes intracellular NH2 and COOH termini, two membrane-spanning domains, and a large glycosylated extracellular loop with cysteine-rich domains (5). This conserved structure, along with substantial sequence homology, is a common feature of all DEG/ENaC superfamily members (1). The similar structure of these proteins contrasts with their functional heterogeneity and broad tissue distribution (2). Although there is substantial evidence implicating DEG proteins in touch sensation and mechano-transduction (8, 9), the influence of mechanical forces on ENaC activity remains controversial. Different studies indicate that ENaC responds to mechanical forces—for example, those induced by osmotic stress (10, 11), hydrostatic pressure (12, 13), and laminar shear stress (14, 15). Although this issue is still under debate, with controversial results (compare refs. 10, 11 and ref. 16), there is growing evidence that at least laminar shear stress (LSS) represents an adequate stimulus to mechanically activate ENaC (14, 15). Since different Na⫹-absorbing epithelia are exposed to mechanical forces—for example, airway epithelia during breathing (17) and the cortical collecting duct epithelium by the tubular flow (14, 18)—it seems feasible that shear stress is of considerable physiological relevance. The suggestion that ENaC could be regulated by mechanical stimuli is further supported by the finding that ENaC subunits are expressed in vascular tissue and contribute to mechano-sensory structures involved in the control of blood pressure (19, 20). 1
Correspondence: Institute of Animal Physiology, JustusLiebig University Giessen, Wartweg 95, D-35392 Giessen, Germany. E-mail:
[email protected] doi: 10.1096/fj.06-7694com 2389
The aim of our study was to determine whether laminar shear stress might directly interfere with the activity of ENaCs. For this purpose we used the Xenopus oocyte expression system to investigate the effect of LSS on ENaCs cloned from the rat colon (7) and a Xenopus distal nephron cell line (A6 cells) (21). LSS was generated by a gravity-driven fluid stream using a pipette placed in close proximity to the oocytes. The ENaC activity was recorded in response to activated flow, which produced shear forces in physiologically relevant ranges at the oocyte surface. We obtained LSS-induced ENaC currents in whole-cell and single-channel recordings, independent of tissue and species origin. Our data suggest that mechano-sensitivity of ENaC might be a ubiquitous regulatory mechanism to control Na⫹ reabsorption and thereby water homeostasis in vertebrates. MATERIALS AND METHODS Heterologous expression of ENaCs Defolliculated oocytes of stages V and VI derived from female South African clawed frogs (Xenopus laevis) were injected with cRNA encoding Xenopus (xENaC, 5 ng ␣, 2.5 ng  and ␥; total volume 41.4 nl) and rat epithelial Na⫹ channels (rENaC, 1.1 ng ␣, 0.6 ng , ␥; total volume 9.2 nl). After injection, oocytes were kept in low sodium solution containing (in mM): 10 NaCl, 80 NMDG-Cl (N-methyl-d-glucamine), 1 KCl, 2 CaCl2, 5 HEPES, 2.5 Na⫹-pyruvate, 0.06 penicillin, 0.02 streptomycin, pH 7.4. Measurements were performed 1– 4 days after injection. For each experimental approach, oocytes from at least two different donors were used. Whole-cell recordings by the two-electrode voltage clamp technique (TEVC) Oocytes were clamped at a membrane potential of ⫺60 mV using a Warner-TEVC amplifier (Warner Instruments, Hamden, CT, USA) and transmembrane currents (IM) were recorded via a strip chart recorder. The oocytes were continuously superfused with oocyte Ringer’s solution (ORi, containing in mM: 90 NaCl, 1 KCl, 2 CaCl2, 5 HEPES, pH 7.4) driven by gravity (flow rate 2.5 ml/min). In some measurements, NaCl was substituted by equimolar amounts of LiCl. To avoid side effects induced by flow due to normal bath superfusion, a Plexiglas barrier was placed in front of the oocyte (see Fig. 1A). LSS was initiated by the additional superfusion of Ringer’s solution through a Pasteur pipette (inner diameter 1.0 mm) placed near the oocyte (⬃1 mm distance). Under these circumstances, we determined a Reynolds number (Re, see below) of 63.7 representing an adequate value for laminar flow (22, 23). The application of shear stress was also gravity driven. The flow rate was 3.0 ml/min producing shear forces (Fshear; see below) of ⬇5.1 dynes/cm2. To ensure that ENaCs also respond to lower shear forces, we performed some experiments using flow rates that produced shear forces of ⬇0.2– 0.5 dynes/cm2. All recordings were performed with continuous bath perfusion, since this did not affect the LSS responses (data not shown). Patch-clamp recordings Single-channel recordings were performed on excised patches derived from devitellinized oocytes in the outside-out 2390
Vol. 21
August 2007
Figure 1. Effect of laminar shear stress (LSS) on ENaCs expressed in Xenopus oocytes. A) Experimental setup used to apply LSS at the surface of the oocytes. Oocytes were placed in a customized Plexiglas chamber that was continuously perfused (inflow, if; outflow, of). The chamber included a shield (sh) to protect the oocytes from flow contamination due to the bath superfusion. LSS was applied separately via a pipette (pt) placed in front of the oocyte (oc). In all experiments, identical solutions were included in the bath and the pipette. LSS responses were not affected by bath perfusion (data not shown). Intracellular microelectrodes (me) were used to control membrane potential via a voltage clamp amplifier. Experiments were performed at a holding potential of ⫺60 mV. B) Original recording of LSS-induced responses of rat ENaCs expressed in oocytes. Note that the LSS effects were repetitive, although the amplitudes gradually decreased. Black bars indicate the application of laminar shear stress. C) Summarized results of LSS effects on rat ENaC (n⫽23, **P⬍0.01) and Xenopus ENaC (n⫽41, **P⬍0.01). Measured currents (IM) are normalized to the values before application of LSS. cont: control; LSS: laminar shear stress; ⫺LSS: after LSS application;. D) LSS-induced currents were amiloride sensitive. Blocking of rat ENaCs by amiloride (10 M) abolished the LSS effect (ami., amiloride). E) Control experiments with LSS application on water-injected control oocytes. Neither amiloride nor LSS obviously affected membrane currents of water-injected oocytes. F) Dose-dependent inhibition of expressed rat ENaC by increasing amiloride concentrations after LSS-induced activation of the channels. Results of these experiments are summarized in Table 1. configuration. Injection of cRNA followed the same routine as described above. Patch pipettes were pulled from borosilicate glass capillaries with an outer diameter of 1.6 mm (Hilgenberg GmbH, Malsfeld, Germany), exhibiting resistances between 3 and 5 M⍀ using a two-stage vertical puller
The FASEB Journal
ALTHAUS ET AL.
(Narishige, Tokyo, Japan) and filled with intracellular analogous high K⫹ solution [containing in mM: 140 KCl, 1 CaCl2, 2 MgCl2, 10 HEPES, 11 EGTA (ethylene glycol-bis-N,N⬘tetraacetic acid); pH 7.2]. Bath solution contained (in mM): 90 NaCl, 1 KCl, 2 CaCl2, 5 HEPES, adjusted to pH 7.4. In some experiments NaCl was replaced by equimolar amounts of LiCl. Effects of LSS were recorded at membrane potentials of ⫺100 mV. Currents were amplified with a List LM-PC amplifier (List Electronics, Darmstadt-Eberstadt, Germany) and digitized by an Axon interface (1200 series; Axon Instruments, Union City, CA, USA). Prior to data acquisition, currents were filtered with 100 Hz using a low-pass filter (Frequency Devices, Inc., Ottawa, IL, USA). Data were acquired (2 kHz) and analyzed with the Axon Clampex software 8.0.3 using a 266 MHz Pentium personal computer. Laminar shear stress was applied by a gravity-driven perfusion system (ALA Scientific Instruments, Westbury, NY, USA) connected to a tube (inner diameter, 1.0 mm) placed inside the measurement chamber. The patch pipette was aligned at an angle of 90° to the perfusion tube (Fig. 4A). The flow rate was 0.3 ml/min, resulting in laminar shear forces (Fshear) of ⬇0.2 dynes/cm2 (Re⫽6.37). Reynolds numbers and shear forces were determined as described below. To quantify ENaC activity in response to LSS, the open probability (PO) was determined. For this purpose we initially injected lower cRNA amounts, but were unable to obtain recordings with only one active channel (with either Na⫹ or Li⫹ as conducting ion). Therefore, the open probability was determined by the following equation:
冉 冘 冊/ nP n
N
n
In this equation, n is the number of conducting levels at any given time, N the number of active channels in the patch obtained from the number of observed current levels, and Pn the probability that n out of N channels are open. To obtain reliable estimates of the total number of ion channels in the patch, the estimated Pn values were compared with Pn values derived from a theoretical binomial distribution (see representative Fig. 4E), with the assumption that N channels all have the same PO. Open probability was calculated solely from recordings where the agreement was satisfactory (Fig. 4E). For a detailed description of the procedure, see ref 24. In addition, the relative open probability (NPO) was determined by the equation: NP O ⫽
冘
R e ⫽ D/ represents the density of water, the flow velocity, D the diameter of the perfusion pipette/tube, and the kinematic viscosity of water as described in ref. 23. Laminar flow is predicted for Re ⬍ 1000 (see ref. 22). Effective shear forces (Fshear) are given by the relation Fshear ⫽ Fdrag/surface oocyte or membrane area (see refs. 15, 23). The effective drag force (Fdrag) was calculated from: F drag ⫽ 0.5A 2C d where ⫽ density of water; A ⫽ effective surface; ⫽ flow velocity, Cd ⫽ drag coefficient (⬃1 for Re in the range of 3– 80, see refs. 15, 23). Chemicals All chemicals used to prepare experimental solutions were obtained from Fluka (Deisenhofen, Germany), except for HEPES, EGTA, pyruvate, streptomycin, and penicillin, which were from Sigma (Deisenhofen, Germany). The pharmacological compounds amiloride, trypsin, and gadolinium chloride were purchased from Sigma. ZnCl2 was purchased from Riedl-de Haen (Seelze, Germany). All compounds were applied in ORi. Statistical analysis
N
PO ⫽
mental conditions used to generate laminar shear stress in the two-electrode-voltage-clamp and in the patch-clamp measurements. Reynolds numbers (Re) were calculated by the equation:
共t nn兲/T
Data are presented as means ⫾ se, n indicates the number of performed experiments. If not stated otherwise, a 2-tailed paired Student’s t test was used to estimate the significance of the LSS-induced effects by comparing dependent values from identical experiments (without and with LSS and before and after pharmacological treatment, respectively). For independent values, a nonparametric Mann-Whitney test was used.
RESULTS Basic effects of laminar shear stress in whole-cell recordings
The following equations were used to calculate the approximate shear forces, which were predicted under the experi-
For the setup shown in Fig. 1A, we determined shear forces of ⬇5.1 dynes/cm2, which represent magnitudes of physiological relevance (17, 18). By activating the pipette perfusion, we obtained an increase in the measured membrane current (IM, Fig. 1B). The LSSinduced current (ILSS) was obtained in oocytes expressing either rENaC or xENaC (Fig. 1C). Although the expression rates of xENaC were lower than those of rENaC, the LSS-induced effects were within similar ranges (ratio of membrane currents with LSS and before LSS; rat ENaC: 1.23⫾0.05, n⫽23; xENaC: 1.23⫾0.02, n⫽41). Some experiments (n⫽6 for rENaC and xENaC) were performed with lower shear forces (⬃0.2– 0.5 dynes/cm2). In these experiments, we still obtained significant activation of ENaC currents (ratio with LSS and before LSS: rat ENaC 1.13⫾0.02, P⬍0.05; xENaC 1.06⫾0.02, P⬍0.05). In all experiments, LSSinduced activation was completely reversible, since the
LAMINAR SHEAR STRESS INCREASES ENAC OPEN PROBABILITY
2391
where t represents the open time of a channel obtained at a distinct level (n) and T the total recording time. Mean open (to) and closed (tc) times were determined by the equations: t o ⫽ P ONT/
冘
En
and t c ⫽ 共1 ⫺ P O兲NT/
冘
En
where N is the number active channels verified by binomial distribution, E is the number of binned events for a distinct level (n), and T the total recording time. Determination of shear forces
currents returned to baseline levels after the pipette perfusion was stopped (Fig. 1C). We also found that the LSS-induced effect was repetitive, but the subsequent current amplitudes were decreased after repeated LSS applications (Fig. 1B). The finding that the LSS-activated current was sensitive to the diuretic amiloride (Fig. 1D, E) and that there were no LSS-induced effects detectable in water-injected control oocytes (Fig. 1E) clearly indicates an activation of ENaCs due to LSS. These observations agree well with studies performed by Kleyman and colleagues with mouse ENaC expressed in oocytes (14, 15). To ascertain whether LSS may affect amiloride binding kinetics, we performed dose-response experiments with and without activated shear stress (Fig. 1F). Neither half-maximal blocker concentrations (IC50) nor Hill coefficients were significantly changed by activated shear stress (Table 1). Further, we did not find significant differences between LSS-induced responses when comparing devitellinized and vitellinized ENaC-expressing oocytes (ratio Iwith LSS/ Ibefore LSS in devitellinized oocytes rENaC: 1.16⫾0.04, n⫽5; xENaC: 1.34⫾0.01, n⫽4). Our data together with observations made by others (14, 15) clearly demonstrate that LSS in physiological relevant ranges provides an adequate stimulus to mechanically activate ENaC. Modulation of LSS response by ENaC activators Further evidence that LSS induces Na⫹ currents via ENaC activation included experiments with the trivalent gadolinium cation (Gd3⫹), glibenclamide, the divalent zinc cation (Zn2⫹) and trypsin. Gadolinium and glibenclamide, for example, are described to stimulate amphibian ENaCs by increasing ion channel activity (25–28) due to an increased relative open probability (NPO, 26, 27). Regarding the assumption that LSS may increase ENaC activity rather than provide new ENaCs from cellular stores, we anticipated that the effects of Gd3⫹ and glibenclamide would not be additive to the LSS effect. Therefore, we compared the effect of LSS under control conditions (superfusion with oocyte Ringer’s solution, ORi) with the LSS effect after preincubation with Gd3⫹ or glibenclamide. TABLE 1.
Addition of gadolinium (100 M) to the bath increased amiloride-sensitive xENaC currents. Subsequent application of LSS via the pipette (perfusate also contained 100 M gadolinium) did not further activate xENaC (Fig. 2A, B). Similar results were obtained with the sulfonylurea receptor inhibitor glibenclamide (100 M, Fig. 2B). Neither glibenclamide nor Gd3⫹ showed any reaction in water-injected control oocytes (data not shown) or in the presence of amiloride, respectively (data not shown and refs. 26, 27). Because rat ENaC was not activated by either gadolinium or glibenclamide, we used Zn2⫹, which is an activator of mouse ENaC, where it abolishes self-inhibition (29). We found that Zn2⫹ was also able to activate amiloride-sensitive rENaC currents (⫹24⫾5%, n⫽12, P⬍0.001). The effect of Zn2⫹ was not detected in the presence of amiloride or in water-injected control oocytes (data not shown). To measure the LSS effects on Zn2⫹-activated rENaC, the same experimental procedure was used as with Gd3⫹ or glibenclamide on xENaC. We found that pre-exposure of rENaC to 10 M zinc was sufficient to decrease the laminar shear stress-induced activation (Fig. 2C, D). Since we found that LSS stimulation is mimicked by Gd3⫹, glibenclamide, and zinc, these findings provide evidence that LSS may activate membrane-located ENaCs by changing ion channel activity rather than recruiting new channels from intracellular stores. In another approach, we used trypsin to evoke rENaC and xENaC currents. Proteases like trypsin, in contrast to ENaC openers, should evoke ENaC currents by increasing the number of active ion channels. In our experiments, trypsin (20 g/ml) increased amiloridesensitive currents as described previously (30). In contrast to the experiments with gadolinium, glibenclamide, and zinc, we found that the trypsin-evoked current was obviously affected by subsequent LSS application (Fig. 3A, C). Notably, the LSS-induced current (ILSS) was potentiated by trypsin in rENaC-expressing oocytes (Fig. 3B) compared with the LSS-induced currents under control conditions (cells perfused with ORi). In contrast, xENaC currents were inhibited by LSS after trypsin-induced activation (Fig. 3D). The LSS responses observed after trypsin application—stimulation of rENaC and inhibition of xENaC currents—were
Effect of laminar shear stress on dose-dependent inhibition by increasing amiloride concentrationsa rENaC Control n⫽6
IC50 (nM, amilo.) P Hill coefficient P
99 ⫾ 15 0.87 ⫾ 0.06
0.53
xENaC LSS n⫽7
Control n⫽5
81 ⫾ 5
196 ⫾ 15
0.82 ⫾ 0.03
0.63
0.82 ⫾ 0.05
LSS n⫽6
170 ⫾ 15 0.33
0.88 ⫾ 0.02
0.54
a Amiloride was added in concentrations from 0.001 to 100 m to the bath (control) or the fluid stream (LSS). The values of each experiment were fitted to the Hill equation in order to determine half-maximal inhibitory concentrations (IC50) and Hill coefficients for amiloride. For statistical analysis, IC50 values and Hill coefficients determined in the experiments without and with laminar shear stress were compared. Values are means ⫾ se; n represents the number of experiments; P estimated by nonparametric Mann-Whitney U test.
2392
Vol. 21
August 2007
The FASEB Journal
ALTHAUS ET AL.
outside-out configuration to monitor changes in ion channel activity at the protein level. After the outside-out configuration was established, the patch pipette was moved toward a perfusion tube placed inside the measuring chamber to apply LSS (Fig. 4A). We determined shear forces of ⬇0.2 dyn/cm2 for this setup, which represents the lower range of shear forces present during physiological processes (23, 31). Measurements lasted ⬃3 min following the same protocol: recording the basal ion channel activity for 1 min without perfusion, then recording ion channel activity with activated perfusion and thereby LSS (1 min), and finally perfusion with 10 M amiloride (Fig. 4B, C). From these experiments we initially found that the relative ion channel open probability (NPO) of rENaC and of xENaC was significantly increased (without/with
Figure 2. Gadolinium, glibenclamide, and zinc mimicked the shear stress-induced effect. A) Original recording from an oocyte-expressing Xenopus ENaC (xENaC). The laminar shear stress-induced effect (ILSS) was first determined under control conditions and after application of gadolinium (Gd3⫹, 100 M). Gd3⫹ increased the xENaC-conducted Na⫹ current and prevented further activation due to LSS. The Gd3⫹activated current exhibited a magnitude similar to that of the LSS-induced current (ami., amiloride, 10 M). B) Summarized data from multiple experiments (n⫽9 –10) conducted as described in panel A using Gd3⫹ or glibenclamide (100 M). Data indicate means ⫾ se. Plotted are the LSS-induced current amplitudes (ILSS) without (⫺) and after (⫹) pharmacological activation (gado., gadolinium; *P⬍0.05; glib., glibenclamide; **P⬍0.01). C) Similar experiments were performed with the divalent zinc cation (Zn2⫹) on rat ENaCexpressing oocytes. The LSS effect was first determined under control conditions and after preincubation with 10 M Zn2⫹. D) Summarized data concerning the effect of LSS in the absence (⫺) and presence of Zn2⫹ (⫹). Data represent means ⫾ se of the LSS-induced current (ILSS, n⫽12, *P⬍0.01).
proportional to the magnitude of the trypsin-induced current for each experiment, so that ratios of the LSS-induced effects were similar before and after trypsin application (Fig. 3E, ratio of ILSS and the current values before LSS application). From these results we conclude that the LSS-induced effects are rooted in an increased number of channels, which were also activated by trypsin and thereby provided more molecules accessible to LSS.
To estimate whether the LSS-activated ENaC currents resulted from an increased ion channel open probability, we performed single-channel experiments in the
Figure 3. Laminar shear stress-induced effects after proteolytic cleavage. A) Current recording showing the impact of trypsin (20 g/ml) on the LSS effect of rat ENaC (rENaC). This procedure was found to increase the LSS-activated current amplitude with respect to control conditions. B) The LSS-induced current (ILSS) of rat ENaC was significantly increased after trypsin application (tryps., n⫽10; *P⬍0.05). C) Effect of trypsin on LSS-induced current on Xenopus ENaC. D) Notably, LSS obviously inhibited IM after trypsin activation (n⫽9, *P⬍0.05). E) Comparison of the LSS effects under control conditions (⫺) and after trypsin application (⫹). Values represent absolute changes given as ratios of the LSS-induced current with respect to the measured current before LSS administration independent of current deflection (rENaC, n⫽10, P⫽0.6; xENaC, n⫽9, P⫽0.9).
LAMINAR SHEAR STRESS INCREASES ENAC OPEN PROBABILITY
2393
LSS increases the open probability (PO)
Figure 4. Activation of ENaC by laminar shear stress in outside-out patch clamp recordings. A) Schematic side view of the setup in order to apply laminar shear forces on excised membrane patches from Xenopus oocytes (patch pipette, pp; perfusion tubule, pt). B) Representative outside-out recording of rat ENaC (rENaC) exposed to laminar shear stress (LSS) and amiloride (ami.). In this recording, two active channels are visible, although it is obvious that activation of flow mainly increases the dwell open time of active channels. Singlechannel currents were recorded at ⫺100 mV. Amiloride (10 M) was added to the perfusion solution to reveal baseline current (dotted lines). C) Similar recording of a Xenopus ENaC (xENaC) containing outside-out patch. In this recording, four active channels were visible at the same time before LSS activation. LSS increased the number of visible active channels to six. Note that this was observed in only two of seven experiments performed, but this recording is convenient for demonstrating how the number of active ion channels was determined (also see panel E). D) The open probability PO was determined without (60 s) and with LSS (60 s). LSS significantly increased the open probability of rat ENaC, as well as of Xenopus ENaC. Values are means ⫾ se (rENaC: n⫽7, xENaC: n⫽7, *P⬍0.05). E) Representative figure comparing the measured and the predicted Pn obtained from a binomial distribution. Pn represents the probability of n channels being open. The measured and predicted Pn’s are related to the recording shown in panel C. In this recording, four active channels were observed without LSS (control), whereas six channels were visible after LSS administration. Note the close correlation between the measured and the predicted Pn. Binomial distribution in order to obtain predicted values for Pn was calculated with n⫽4 channels without and n⫽6 channels with activated LSS.
LSS: rENaC: 0.43⫾0.1/0.92⫾0.18, n⫽7, P⬍0.05; xENaC: 0.32⫾0.14/0.69⫾0.18, n⫽7, P⬍0.05). From this parameter, one cannot distinguish between an increased number of active channels (N) and/or an increase in the single-channel open probability (PO). We first tried to inject lower amounts of RNA to obtain patches with only one active channel. Since we were not able to obtain such recordings, we determined the PO and N by analysis, as recently described in detail (see ref. 24). We found the PO of rENaC (n⫽7) and of xENaC (n⫽7) to be significantly increased by LSS with respect to control conditions (no LSS; Fig. 4D and Table 2). An increased PO can be achieved by two TABLE 2.
events: either the dwell open time (to) of the channel is increased or the dwell closed time (tc) of the channel is decreased. We found that the to of rENaC was significantly increased and that the tc of xENaC was significantly decreased by LSS (Table 2). Accordingly, both events may possibly explain the overall increased open probability. Although in two of seven recordings on Xenopus ENaC the number of active channels (N) was visibly increased by LSS (Fig. 4C), in general the number of active channels was not significantly affected by LSS compared with control conditions (Table 2). The sodium conductance of the two orthologs was not affected by LSS (Table 2).
Effect of laminar shear stress on single channel properties of rat and Xenopus ENaCa
Channel
Condition
n
N
PO
to (s)
tc (s)
g (pS)
rENaC
Control LSS Control LSS
7 7 7 7
2.4 ⫾ 0.2 2.4 ⫾ 0.2 2.6 ⫾ 0.4 3.0 ⫾ 0.6
0.19 ⫾ 0.05 0.38 ⫾ 0.05* 0.1 ⫾ 0.03 0.23 ⫾ 0.06*
1.19 ⫾ 0.4 2.33 ⫾ 0.5* 0.46 ⫾ 0.2 1.0 ⫾ 0.3
5.83 ⫾ 1.5 3.77 ⫾ 0.6 6.85 ⫾ 1.8 3.86 ⫾ 0.9*
3.9 ⫾ 0.2 4.2 ⫾ 0.2 4.7 ⫾ 0.1 4.9 ⫾ 0.1
xENaC
Measurements were performed in the outside-out configuration at ⫺100 mV with Na⫹ as charge carrier. n, number of performed experiments; N, number of active ion channels; PO, open probability; to, mean open time; tc, mean close time; g, single channel conductance at ⫺100 mV membrane potential. Values are means ⫾ se; *significant different with respect to control (paired t test, P⬍0.05). a
2394
Vol. 21
August 2007
The FASEB Journal
ALTHAUS ET AL.
In some experiments, especially when basic ion channel activity was very low (⬃20% of the recordings), no changes in PO were observed in response to LSS. These data are not included in the statistics presented in Fig. 4 or Table 2. LSS does not affect ENaC selectivity To address whether mechanical activation may affect ENaC selectivity, we performed whole-cell recordings and patch-clamp recordings using the rat clone. With Li⫹ as a conducting ion in the bath, we still observed LSS-induced ENaC activation in whole-cell recordings (Fig. 5A, B). Although the relative responses (% of control) were slightly lower compared to the experiments with Na⫹, the characteristics of the effect were identical (compare Fig. 1B and Fig. 5A).
To further clarify this issue, outside-out patch-clamp recordings with Li⫹ in the bath were performed. There were no changes in permeability (Fig. 5C) or in the Na⫹/Li⫹ selectivity ratio detectable due to LSS (rENaC: 1/1.94 without and 1/2.0 with LSS from at least six current deflections at ⫺100 mV from two different recordings).
DISCUSSION
Figure 5. ENaC selectivity was not affected by LSS. A) Wholecell recording of a representative experiment with lithium as conducting ion in the external bath. Rat ENaC-expressing oocytes were repetitively exposed to LSS. The LSS-induced effects were similar to those obtained with Na⫹ in the bath (compare Fig. 1B). B) Summarized data of experiments as shown in panel A (n⫽12, *P⬍0.01). Measured currents (IM) are normalized to the values before application of LSS. cont: control; LSS: laminar shear stress; ⫺LSS: after LSS application. C) Current-voltage relation of outside-out single-channel recordings on rat ENaC using Na⫹ or Li⫹ as charge carrier. Current amplitudes for Na⫹ as well as Li⫹ were measured under control conditions and with activated LSS. Values represent means of at least six current deflections for each voltage step obtained from at least two different recordings. In all of these recordings, an increase in open probability was observed as well (data not shown). Means were fitted by the Goldman-Hodgkin-Katz current equation. The calculated permeabilities were absolutely identical independent of whether ENaCs were exposed to LSS (Na⫹: 1.17⫻10⫺12; Li⫹: 2.22⫻10⫺12 cm⫺3 s⫺1).
Highly selective epithelial Na⫹ channels are expressed in various vertebrate epithelia where they are exposed to shear forces—for example, distal nephron (14, 18) and airway epithelia (17). In addition, there is growing evidence that ENaC subunits are expressed in vascular tissue (19, 20, 32) as well as in sensory nerve endings, indicating participation in mechano-sensory processes (33). It is well known that ENaC activity is controlled by different factors, including hormones (34), kinases (35), intrinsic Na⫹-dependent mechanisms (36, 37), -adrenergic agonists (38), and proteases (39 – 42). There is growing evidence concerning activation of ENaC by mechanical forces, and at least laminar shear stress seems to be an adequate stimulus of physiological significance (14, 15). Given the relationship of ENaC orthologs to degenerin (DEG) proteins (1, 2), it seems reasonable that ENaCs might respond to mechanical stimuli also, since there is substantial evidence that members of the DEG/ENaC superfamily are involved in mechano-sensitive processes (8, 9, 43). From this point of view, mechano-sensitivity of DEG/ENaC superfamily members may represent a fundamental ancestral feature of these proteins, supporting the idea that DEG/ENaC family members were derived from an ancestor protein early in animal evolution (44). This hypothesis is further strengthened by the findings that mechano-sensitivity of ion channels is already evident in bacteria and Archaea, and therefore is a basic, ancestral feature of ion channels (45). The concept of ENaC mechano-sensitivity was introduced in 1995 by the group of Benos (13). This study, among others, provided evidence of ENaC activation by hydrostatic pressure. These results were discussed by Rossier (16), who pointed out that for those studies, ENaC subunits were reconstituted in lipid bilayers forming ion channels with conductances of ⬃40 pS, which does not fit the biophysical properties of ENaCs in native epithelial cells (46). However, Kleyman and colleagues demonstrated for the first time mechanical activation of ENaC by exposing heterologous expressed mouse ENaCs to laminar flow (ref. 14). Our present aim was to further characterize the effect of mechanical forces on different ENaC orthologs and to clarify whether these effects were due to an increased ion channel open probability. For this purpose, we used laminar shear stress to mechanically stimulate ENaCs, since this was shown to be an adequate mechanical stimulus of physiological relevance
LAMINAR SHEAR STRESS INCREASES ENAC OPEN PROBABILITY
2395
for different types of ion channels (23, 31, 47, 48), including mouse ENaC (14, 15). Basic effects of laminar shear stress in whole-cell recordings The shear forces used represent magnitudes within physiological ranges, since shear forces of up to 20 dynes/cm2 are predicted to occur in blood vessels (31). We found that application of LSS on oocytes expressing ␣␥ENaC proteins cloned from rat colon (7) or Xenopus kidney (21) activated an inward current that was sensitive to amiloride. Further, the amiloride binding kinetics were not obviously affected by LSS activation. These observations indicate that LSS selectively activated ENaCs and are consistent with the results published by Satlin et al. (14) and by Carattino et al. (15) with ENaC from the mouse kidney. Mechano-sensitivity of ENaC may represent a basic feature of ENaC proteins independent of tissue and species origin. Further, this feature may correspond to the mechano-sensitivity of other DEG/ENaC family members (1, 2, 9). Modulation of LSS response by ENaC activators To reveal the putative mechanism of LSS-induced ENaC activation, we used chemicals that are known to influence ENaC activity. The trivalent gadolinium cation (Gd3⫹) and the sulfonylurea receptor antagonist glibenclamide stimulate xENaC (25–28). Both compounds are suggested to interfere with external domains of xENaC (25, 26) or closely associated proteins (27, 28), and were demonstrated to stimulate ion channel activity by increasing the relative open probability (NPO) of the channels. In our experiments we found that Gd3⫹ and glibenclamide activated amiloride-sensitive currents of Xenopus ENaC. Moreover, preincubation with either Gd3⫹ or glibenclamide significantly prevented the subsequent LSS-induced effect on xENaC. Since we were not able to activate the rat ENaC by Gd3⫹ or glibenclamide, the divalent zinc cation (Zn2⫹) was used. This cation was shown to activate heterologously expressed mouse ENaC by abolishing self-inhibition (29). In our experiments we were able to activate ENaC currents by application of Zn2⫹. What is more interesting is that the LSS responses under these conditions were largely reduced compared with control conditions, similar to results obtained with Gd3⫹ and glibenclamide on Xenopus ENaC. We are unable to account for the species-dependent differences concerning pharmacological activation of the used ENaC orthologs. However, these compounds, which are described to increase ENaC activity by changing the gating properties of the channels, were able to mimic LSS-induced ENaC activation. This further indicates that LSS may activate membrane-located ENaCs by changing ion channel gating properties rather than by providing new channels to the membrane. Changes in ENaC gating properties might be due to mechanical2396
Vol. 21
August 2007
induced conformational changes in extracellular domains of the channels. Another approach to modulate the LSS-induced activation was achieved by the application of trypsin. Proteolytic cleavage of silent ENaCs is a discrete regulatory mechanism controlling transepithelial Na⫹ reabsorption (39 – 42) since membrane-bound proteases (channel activating proteases, CAPs) could modify silent, inactive ENaCs by proteolysis. It was further shown that trypsin is able to mimic the CAP-dependent activation of ENaC (39, 40). Given the fact that trypsin increases the number of active ENaC molecules, this should, in contrast to the experiments with Gd3⫹, glibenclamide, and zinc, affect the LSS responses. For this purpose, rat and Xenopus ENaC-expressing oocytes were exposed to LSS before and after trypsin application. Consistent with data published by Chraibi et al. (30), trypsin activated an amiloride-sensitive Na⫹ current. But in contrast to Gd3⫹, glibenclamide, and Zn2⫹, the trypsin-activated current was sensitive to subsequent LSS application (Fig. 3). The LSS-induced current amplitude (ILSS) was increased when trypsin was applied to rENaC-expressing oocytes (Fig. 3B). Trypsinactivated currents obtained from xENaC-expressing oocytes were significantly inhibited by LSS (Fig. 3D). Although we cannot explain the opposing LSS effects on rat and Xenopus ENaC after proteolytic cleavage, it is remarkable that the relative amplitudes of LSS-induced current activation were not changed by application of trypsin (Fig. 3E). From these results we conclude that the LSS effect is related to the increased number of channels, which were also activated by proteolytic cleavage (39 – 42). This in turn provides more molecules that are accessible to LSS and thus increases the amplitude of the LSS responses. Nevertheless, it seems reasonable that the divergent LSS reactions obtained for rat and Xenopus ENaC after proteolytic cleavage may be due to different amino acid sequences in the extracellular loops (21) resulting in different recognition sites for trypsin, and thus different cleavage patterns. LSS increases the open probability (PO) Finally, we performed single-channel recordings in the outside-out configuration, which enabled us to directly expose the excised membrane patches containing either rat or Xenopus ENaC to LSS. In contrast to other studies measuring the effect of LSS on ion channel activity (14, 15, 23, 31, 48, 49), we for the first time detected a direct activation of an ion channel at the protein level in response to LSS. From our single-channel recordings, we were able to show that laminar shear stress activates rENaC and xENaC currents, clearly indicated by the significantly increased open probability obtained after LSS exposure. Increased PO was associated with either a significant increase of to or a decrease of tc . There was no significant change in the number of active ion channels (N) observed by LSS (data summarized in Table 2). In
The FASEB Journal
ALTHAUS ET AL.
some instances LSS increased the number of active channels (see Fig. 4C ); in these cases, however, an increased PO was also observed. These findings further support our observations obtained from experiments with Gd3⫹, glibenclamide and Zn2⫹. It seems feasible that LSS activates ENaCs by increasing the open probability of the channels rather than increasing the number of active ion channels or recruiting new channels from intracellular stores. Further, singlechannel characteristics in terms of ion channel conductance, permeability, and selectivity were not affected by LSS. The observed ENaC activation is of considerable relevance because the applied shear forces in these single-channel recordings were ⬃0.2 dynes/cm2 and represent the lower range of physiological shear forces. Liu et al. (18), for example, predicted shear forces of 0.52 dynes/cm2 for the wall of cortical collecting ducts, and Olesen et al. (31) were able to initiate LSS-induced whole-cell K⫹ currents in endothelial cells with shear forces of 0.2 dynes/cm2. In some patch-clamp measurements with low basal ion channel activity (⬃20% of the performed measurements), no changes in PO were detectable. Although similar observations were published by Satlin et al. from two electrode voltage clamp measurements (14), we cannot explain this phenomenon. It is described that a population of ENaCs escape proteolytic cleavage in Golgi, and these channels are characterized by low NPO (NPO⬍0.03; ref. 42). Thus, it is possible that these unprocessed ENaCs do not respond to mechanical stimuli. Another explanation could be that these channels in the excised patch were insufficiently exposed to the activated flow. Our data strongly support the suggestion that ENaCs respond to mechanical forces, since we were able to show a direct activation of two ENaC orthologs by mechanical stimuli within physiological relevant ranges. However, we would like to emphasize that one cannot directly compare the ENaC responses derived from whole-cell recordings with those from the singlechannel recordings because, in general, different magnitudes of shear forces were used. Although we were able to obtain significant ENaC activation in whole cell recordings with low shear forces, such as those used for the single-channel recordings, one should realize that it is impossible to determine the effective shear forces at the surface of the oocyte membrane. In contrast to the single-channel recordings, only a fraction of the membrane-located ENaCs will be exposed to the effective shear forces as generated by the perfusion system. This is reasoned by the rounded shape of the cells as well as by the folded membrane surface of the oocytes due to microvilli (50). Nevertheless, with each setup used, we obtained increased ENaC currents. Pharmacological modulation of the LSS responses by gadolinium, glibenclamide, zinc, and trypsin are related to basic regulatory principles of ENaC activity like selfinhibition (36, 37) and proteolytic cleavage (39 – 42).
Although little is known about the relevance of shear forces concerning their impact on epithelial function, it seems reasonable that the effect of these mechanical forces may be deeply rooted in the function of different ENaC-expressing epithelia (e.g., lung, colon, and kidney). From this point of view, mechanical activation of ENaC provides a putative mechanism to couple the rate of transepithelial Na⫹ absorption to luminal transport (e.g., transport of feces in the colon, tubular flow in the kidney, and the movement of airway surface liquid). In addition, this ENaC feature may represent a reasonable explanation concerning the presence of ENaC subunits in non-Na⫹ absorbing epithelia such as arteries and sensory nerve endings (19, 20, 32, 33).
LAMINAR SHEAR STRESS INCREASES ENAC OPEN PROBABILITY
2397
Hypothetic model of mechanical ENaC activation Since most of the gating features of ENaC are still unknown, it is difficult to develop a mechanistic model concerning LSS-induced activation. Nevertheless, we postulate a mechano-sensitive component, which may be located at the extracellular region of the channel. This idea is supported by the finding that LSS led to (species-dependent) contradictory effects after proteolytic cleavage. We suggest that the mechano-sensor might either be part of, or associated with, the large extracellular loops of the ENaC molecules. Further, we propose that this mechano-sensor must somehow be coupled to the gating site of the channel. Based on the structural model developed by Kellenberger et al. (51), the DEG site (which is important for gating processes), the amiloride binding site, and the selectivity filter represent distinct parts of the protein structure. This assumption may explain why LSS alters channel gating, and the open probability in particular, without affecting the other parameters, including amiloride binding kinetics, single-channel conductance, and ion selectivity. The hypothesis of an extracellular mechano-sensor is further supported by recent published observations demonstrating that flow activation of ENaC is independent of membrane trafficking (52). However, additional experimental data are necessary to identify the mechano-sensor and to clarify this issue. Taken together, we have evidence that laminar shear stress directly activates ion channel activity by increasing the ion channel open probability. We suggest that mechano-sensitivity represents an additional drug-independent regulatory mechanism to control ENaC activity. This functional principle may be of general relevance in terms of epithelial function as well as mechano-sensation in nonepithelial tissues, and underlines the evolutionary conservation of mechano-sensitivity as a ubiquitous feature within the DEG/ENaC superfamily. We thank Mirjam Buss for excellent technical assistance and Prof. R. Lakes-Harlan, Dr. R. E. Morty, and J. Strauss for comments and improvements on the manuscript. Prof. B. C. Rossier generously provided the rat and Xenopus cDNAs; we thank Prof. C. Korbmacher and Dr. A. Diakov for their kind support concerning the establishment of
stable outside-out patches from oocytes. The present study was supported by Deutsche Forschungsgemeinschaft (DFG) grant FR 2124/1–1.
21.
22. 23.
REFERENCES 24. 1. 2. 3. 4.
5.
6.
7.
8. 9. 10. 11. 12. 13.
14. 15. 16. 17.
18.
19. 20.
2398
Mano, I., and Driscoll, M. (1999) DEG/ENaC channels: a touchy superfamily that watches the salt. Bioessays 21, 568 –578 Kellenberger, S., and Schild, L. (2002) Epithelial sodium channel/degenerin family of ion channels: a variety of functions for a shared structure. Physiol. Rev. 82, 735–767 Koefoed-Johnson, V., and Ussing, H. H. (1958) The nature of the frog skin potential. Acta Physiol. Scand. 42, 298 –308 Snyder, P. M., Price, M. P., McDonald, F. J., Adams, C. M., Volk, K. A., Zeiher, B. G., Stokes, J. B., and Welsh, M. J. (1995) Mechanism by which Liddle’s syndrome mutations increase activity of a human epithelial Na⫹ channel. Cell 83, 969 –978 Rossier, B. C., Pradervand, S., Schild, L., and Hummler, E. (2002) Epithelial sodium channel and the control of sodium balance: interaction between genetic and environmental factors. Annu. Rev. Physiol. 64, 877– 897 Hummler, E., Barker, P., Gatzy, J., Beermann, F., Verdumo, C., Schmidt, A., Boucher, R., and Rossier, B. C. (1996) Early death due to defective neonatal lung liquid clearance in alpha-ENaCdeficient mice. Nat. Genet. 12, 325–328 Canessa, C. M., Schild, L., Buell, G., Thorens, B., Gautschi, I., Horisberger, J. D., and Rossier, B. C. (1994) Amiloride-sensitive epithelial Na⫹ channel is made of three homologous subunits. Nature 367, 463– 467 Driscoll, M., and Chalfie, M. (1991) The mec-4 gene is a member of a family of Caenorhabditis elegans genes that can mutate to induce neuronal degeneration. Nature 349, 588 –593 Ernstrom, G. G., and Chalfie, M. (2002) Genetics of sensory mechanotransduction. Annu. Rev. Genet. 36, 411– 453 Ji, H. L., Fuller, C. M., and Benos, D. J. (1998) Osmotic pressure regulates ␣␥-rENaC expressed in Xenopus oocytes. Am. J. Physiol. 275, C1182–C1190 Awayda, M. S., and Subramanyam, M. (1998) Regulation of the epithelial Na⫹ channel by membrane tension. J. Gen. Physiol. 112, 97–111 Palmer, L. G., and Frindt, G. (1996) Gating of Na⫹ channels in the rat cortical collecting tubule: effects of voltage and membrane stretch. J. Gen. Physiol. 107, 35– 45 Awayda, M. S., Ismailov, I. I., Berdiev, B. K., and Benos, D. J. (1995) A cloned renal epithelial Na⫹ channel protein displays stretch activation in planar lipid bilayers. Am. J. Physiol. 268, C1450 –C1459 Satlin, L. M., Sheng, S., Woda, C. B., and Kleyman, T. R. (2001) Epithelial Na⫹ channels are regulated by flow. Am. J. Physiol. 280, F1010 –F1018 Carattino, M. D., Sheng, S., and Kleyman, T. R. (2004) Epithelial Na⫹ channels are activated by laminar shear stress. J. Biol. Chem. 279, 4120 – 4126 Rossier, B. C. (1998) Mechanosensitivity of the epithelial sodium channel (ENaC): controversy or pseudocontroversy? J. Gen. Physiol. 112, 95–96 Tarran, R., Button, B., Picher, M., Paradiso, A. M., Ribeiro, C. M., Lazarowski, E. R., Zhang, L., Collins, P. L., Pickles, R. J., Fredberg, J. J., and Boucher, R. C. (2005) Normal and cystic fibrosis airway surface liquid homeostasis: the effects of phasic shear stress and viral infections. J. Biol. Chem. 280, 35751–35759 Liu, W., Xu, S., Woda, C., Kim, P., Weinbaum, S., and Satlin, L. M. (2003) Effect of flow and stretch on the [Ca2⫹]i response of principal and intercalated cells in cortical collecting duct. Am. J. Physiol. 285, F998 –F1012 Drummond, H. A., Welsh, M. J., and Abboud, F. M. (2001) ENaC subunits are molecular components of the arterial baroreceptor complex. Ann. N. Y. Acad. Sci. 940, 42– 47 Drummond, H. A., Gebremedhin, D., and Harder, D. R. (2004) Degenerin/epithelial Na⫹ channel proteins: components of a vascular mechanosensor. Hypertension 44, 643– 648
Vol. 21
August 2007
25. 26. 27. 28. 29. 30.
31. 32. 33. 34. 35.
36. 37. 38. 39. 40. 41.
42. 43. 44. 45. 46.
Puoti, A., May, A., Canessa, C. M., Horisberger, J. D., Schild, L., and Rossier, B. C. (1995) The highly selective low-conductance epithelial Na channel of Xenopus laevis A6 kidney cells. Am. J. Physiol. 269, C188 –C197 McNaughton, K. J., and Sinclair, C. G. (1966) Submerged jets in short cylindrical flow vessels. J. Fluid Mech. 25, 367–375 Hoger, J. H., Ilyin, V. I., Forsyth, S., and Hoger, A. (2002) Shear stress regulates the endothelial Kir2.1 ion channel. Proc. Natl. Acad. Sci. U. S. A. 99, 7780 –7785 Anantharam, A., Tian, Y., and Palmer, L. G. (2006) Open probability of the epithelial sodium channel is regulated by intracellular sodium. J. Physiol. 574, 333–347 Friis, S., and Nielsen, R. (2001) Effect of the putative Ca2⫹receptor agonist Gd3⫹ on the active transepithelial Na⫹ transport in frog skin. J. Membr. Biol. 184, 291–297 Fronius, M., Clauss, W., and Schnizler, M. (2003) Stimulation of transepithelial Na⫹ current by extracellular Gd3⫹ in Xenopus laevis alveolar epithelium. J. Membr. Biol. 195, 43–51 Chraibi, A., and Horisberger, J. D. (1999) Stimulation of epithelial sodium channel activity by the sulfonylurea glibenclamide. J. Pharmacol. Exp. Ther. 290, 341–347 Schnizler, M., Berk, A., and Clauss, W. (2003) Sensitivity of oocyte-expressed epithelial Na⫹ channel to glibenclamide. Biochim. Biophys. Acta 1609, 170 –176 Sheng, S., Perry, C. J., and Kleyman, T. R. (2004) Extracellular Zn2⫹ activates epithelial Na⫹ channels by eliminating Na⫹ self-inhibition. J. Biol. Chem. 279, 31687–31696 Chraibi, A., Vallet, V., Firsov, D., Hess, S. K., and Horisberger, J. D. (1998) Protease modulation of the activity of the epithelial sodium channel expressed in Xenopus oocytes. J. Gen. Physiol. 111, 127–138 Olesen, S. P., Claphamt, D., and Davies, P. (1988) Haemodynamic shear stress activates a K⫹ current in vascular endothelial cells. Nature 331, 168 –170 Jernigan, N. L., and Drummond, H. A. (2005) Vascular ENaC proteins are required for renal myogenic constriction. Am. J. Physiol. 289, F891–F901 Drummond, H. A., Abboud, F. M., and Welsh, M. J. (2000) Localization of  and ␥ subunits of ENaC in sensory nerve endings in the rat foot pad. Brain Res. 884, 1–12 Rossier, B. C. (2002) Hormonal regulation of the epithelial sodium channel ENaC: N or Po? J. Gen. Physiol. 120, 67–70 Diakov, A., and Korbmacher, C. (2004) A novel pathway of epithelial sodium channel activation involves a serum- and glucocorticoid-inducible kinase consensus motif in the C terminus of the channel’s ␣-subunit. J. Biol. Chem. 279, 38134 –38142 Turnheim, K. (1991) Intrinsic regulation of apical sodium entry in epithelia. Physiol. Rev. 71, 429 – 445 Garty, H., and Palmer, L. G. (1997) Epithelial sodium channels: function, structure, and regulation. Physiol. Rev. 77, 359 –396 Matthay, M. A., Folkesson, H. G., and Clerici, C. (2002) Lung epithelial fluid transport and the resolution of pulmonary edema. Physiol. Rev. 82, 569 – 600 Vallet, V., Chraibi, A., Gaeggeler, H. P., Horisberger, J. D., and Rossier, B. C. (1997) An epithelial serine protease activates the amiloride-sensitive sodium channel. Nature 389, 607– 610 Rossier, B. C. (2004) The epithelial sodium channel: activation by membrane-bound serine proteases. Proc. Am. Thorac. Soc. 1, 4 –9 Knight, K. K., Olson, D. R., Zhou, R., and Snyder, P. M. (2006) Liddle’s syndrome mutations increase Na⫹ transport through dual effects on epithelial Na⫹ channel surface expression and proteolytic cleavage. Proc. Natl. Acad. Sci. U. S. A. 103, 2805–2808 Caldwell, R. A., Boucher, R. C., and Stutts, M. J. (2003) Serine protease activation of near-silent epithelial Na⫹ channels. Am. J. Physiol. 286, C190 –C194 Tavernarakis, N., and Driscoll, M. (1997) Molecular modeling of mechanotransduction in the nematode Caenorhabditis elegans. Annu. Rev. Physiol. 59, 659 – 689 Tavernarakis, N., and Driscoll, M. (2001) Degenerins. at the core of the metazoan mechanotransducer? Ann. N. Y. Acad. Sci. 940, 28 – 41 Hamill, O. P., and Martinac, B. (2001) Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740 Palmer, L. G., and Frindt, G. (1986) Amiloride-sensitive Na channels from the apical membrane of the rat cortical collecting tubule. Proc. Natl. Acad. Sci. U. S. A. 83, 2767–2770
The FASEB Journal
ALTHAUS ET AL.
47.
Barakat, A. I., Leaver, E. V., Pappone, P. A., and Davies, P. F. (1999) A flow-activated chloride-selective membrane current in vascular endothelial cells. Circ. Res. 85, 820 – 828 48. Brakemeier, S., Kersten, A., Eichler, I., Grgic, I., Zakrzewicz, A., Hopp, H., Kohler, R., and Hoyer, J. (2003) Shear stress-induced up-regulation of the intermediate-conductance Ca2⫹-activated K⫹ channel in human endothelium. Cardiovasc. Res. 60, 488 – 496 49. Lieu, D. K., Pappone, P. A., and Barakat, A. I. (2004) Differential membrane potential and ion current responses to different types of shear stress in vascular endothelial cells. Am. J. Physiol. 286, C1367–C1375 50. Dascal, N. (1987) The use of Xenopus oocytes for the study of ion channels. CRC. Crit. Rev. Biochem. 22, 317–387
51.
Kellenberger, S., Gautschi, I., and Schild, L. (2002) An external site controls closing of the epithelial Na⫹ channel ENaC. J. Physiol. 543, 413– 424 52. Morimoto, T., Liu, W., Woda, C., Carattino, M., Wei, Y., Hughey, R., Apodaca, G., Satlin, L. M., and Kleyman, T. R. (2006) Mechanism underlying flow-stimulation of sodium absorption in the mammalian collecting duct. Am. J. Physiol. 291, F663–F669
LAMINAR SHEAR STRESS INCREASES ENAC OPEN PROBABILITY
2399
Received for publication November 9, 2006. Accepted for publication February 22, 2007.