| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Institute of Animal Physiology, Justus-Liebig-University Giessen, Giessen, Germany
1Correspondence: Institute of Animal Physiology, Justus-Liebig University Giessen, Wartweg 95, D-35392 Giessen, Germany. E-mail: martin.fronius{at}bio.uni-giessen.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: DEG/ENaC family • patch-clamp • outside-out • Xenopus oocyte
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 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+-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 
, ß, and 
(7)
, and it is suggested that the ion channel is assembled with a stoichiometry of 2, 1ß, 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)
.
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 |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2.5 ng ß and ; total volume 41.4 nl) and rat epithelial Na+ channels (rENaC, 1.1 ng , 0.6 ng ß, ; 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. 1
A).
|
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 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 (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. 4
A). 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:
 |
), 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:
 |
Mean open (to) and closed (tc) times were determined by the equations:
 |
 |
Determination of shear forces
The following equations were used to calculate the approximate shear forces, which were predicted under the experimental 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:
 |
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:
 |
= 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
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 |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 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 LSS-induced 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, LSS-induced activation was completely reversible, since the 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
26
27
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.
Addition of gadolinium (100 µM) to the bath increased amiloride-sensitive xENaC currents. Subsequent application of LSS via the pipette (perfusate also contained 100 µM gadolinium) did not further activate xENaC (Fig. 2
A, B). Similar results were obtained with the sulfonylurea receptor inhibitor glibenclamide (100 µM, 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 µM 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 µg/ml) increased amiloride-sensitive 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. 3
A, 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 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.
|
LSS increases the open probability (PO)
To estimate whether the LSS-activated ENaC currents resulted from an increased ion channel open probability, we performed single-channel experiments in the 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. 4
A). 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 µM 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 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 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)
.
|
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. 5
A, 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 |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 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)
, ß-adrenergic agonists (38)
, and proteases (39
40
41
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 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 ßENaC 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
26
27
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 mechanical-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
40
41
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
). Trypsin-activated 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
40
41
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 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, single-channel 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 single-channel 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 self-inhibition (36
, 37)
and proteolytic cleavage (39
40
41
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)
.
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.
|
ACKNOWLEDGMENTS |
|---|
Received for publication November 9, 2006. Accepted for publication February 22, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
ß-rENaC expressed in Xenopus oocytes. Am. J. Physiol. 275,C1182-C1190[Medline] subunits of ENaC in sensory nerve endings in the rat foot pad. Brain Res. 884,1-12[CrossRef][Medline]-subunit. J. Biol. Chem. 279,38134-38142This article has been cited by other articles:
|
|
 |
|
  W. Yu, P. Khandelwal, and G. Apodaca Distinct Apical and Basolateral Membrane Requirements for Stretch-induced Membrane Traffic at the Apical Surface of Bladder Umbrella Cells Mol. Biol. Cell, January 1, 2009; 20(1): 282 - 295. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Bhalla and K. R. Hallows Mechanisms of ENaC Regulation and Clinical Implications J. Am. Soc. Nephrol., October 1, 2008; 19(10): 1845 - 1854. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Estilo, W. Liu, N. Pastor-Soler, P. Mitchell, M. D. Carattino, T. R. Kleyman, and L. M. Satlin Effect of aldosterone on BK channel expression in mammalian cortical collecting duct Am J Physiol Renal Physiol, September 1, 2008; 295(3): F780 - F788. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. A. Drummond, S. C. Grifoni, and N. L. Jernigan A New Trick for an Old Dogma: ENaC Proteins as Mechanotransducers in Vascular Smooth Muscle Physiology, February 1, 2008; 23(1): 23 - 31. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
