Sodium selectivity of Reissner's membrane epithelial cells

We first tested whether benzamil-sensitive currents, which was earlier observed as transepithelial currents with the current-density vibrating probe [5], could be detected under whole-cell patch clamp conditions (Series 1). Indeed, benzamil (1 μM) reduced the inward current when the pipette and bath solutions (P1, B1) approximated the physiological situation (ignoring differences in apical cation and intracellular Cl^- composition) (Additional file 1: Fig. S1 and Fig. S3; Table 1).

Indeed, benzamil (1 μM) reduced the inward current when the KCl-rich pipette solution (P1) mimicked the presumed intracellular composition (although Cl^- was higher than often observed) and when the NaCl-rich bath (B1) mimicked the basolateral (perilymph) composition, (Additional file 1: Fig. S1 and Fig. S3; Table 1). Benzamil was used throughout these experiments at a concentration (1 μM) that yielded a complete inhibition of the transepithelial current [5] and significantly inhibited the inward whole-cell current (at -100 mV; I_-100) by 48.5%; from -1305 ± 277 pA to -671 ± 146 pA (n = 7). These whole-cell data are consistent with the transepithelial measurements of Na⁺ absorption by Reissner's membrane.

We utilized gene array and RT-PCR to partially address the question of the participation of benzamil-sensitive nonselective cation (NSC) channels. Several isoforms of acid-sensitive ion channels (ASIC) and cyclic-nucleotide gated (CNG) channels were listed in our gene array of mouse Reissner's membrane (GEO GSE6196; Table 2) [6]. ASIC1a, 2a, 2b, 3 and 4 were listed; ASIC1a and 3 yielded a call of "Present", but ASIC1a was tested by RT-PCR and determined to be "Absent" (Table 2). ASIC3 was not determined by RT-PCR; however, ASIC3 is stimulated, rather than inhibited, by amiloride at neutral pH [7] and therefore would not be expected to contribute to the currents measured in the present study. CNGA1, 2, 4 and CNGB3 were listed in the gene array and all received a call of "Absent". CNGA3 and CNGB1 were not listed in the gene array and were not tested by RT-PCR. As mentioned above, ENaC can be a NSC channel under some subunit combinations and the α-,β- and γ -ENaC subunits were all expressed in RM (Table 2).

Whole-cell patch clamp currents were measured (Series 2) under conditions where the only major permeant ions were either Na⁺ or K⁺ (Figure 1, 2). Cl^- was replaced by methanesulfonate (Table 3). In symmetrical Na⁺-rich bath and pipette solutions (B2, P2), the current-voltage (I-V) relationship was nearly linear, with a slight inward rectification at large negative voltage. The slight inward rectification observed at large negative voltage in the first series of symmetrical Na⁺ experiments was at least partially due to the time dependence of the response. This series utilized 300 ms voltage steps and the second linear series utilized 200 ms steps. The strong rectification of the benzamil-sensitive current, however, may at least partially be due to the known voltage sensitivity of inhibition of ENaC by benzamil [8‐10]. Benzamil significantly inhibited the inward whole-cell current (carried mostly by bath Na⁺ at -100 mV) by 87.5%; from I_-100 = -2044 ± 552 pA to -255 ± 57 pA (n = 6). Some experiments displayed a slow rundown in channel activity, as has been observed by others [8, 11]. A representative experiment is shown in Figure 1 and a summary of similar experiments is shown in Figure 2 and Table 1.

In symmetrical K⁺-rich bath and pipette solutions (B3, P3), the current and conductance at -100 mV were significantly smaller than in Na⁺ (Figure 1, 2; Table 1). Benzamil had no significant effect on the inward whole-cell current (carried mostly by bath K⁺ at -100 mV) (mean decreased insignificantly by 9.8%; from I_-100 = -476 ± 112 pA to -429 ± 130 pA, n = 8). A representative experiment is shown in Figure 1 and a summary of similar experiments is shown in Figure 2 and Table 1.

The high Na⁺ selectivity of the benzamil-sensitive current over K⁺ suggested that ENaC might mediate part or all of that current. One salient characteristic of ENaC is a higher permeability to Li⁺ than to Na⁺. We compared (Series 3, 4) whole-cell currents in the presence of Li⁺ and Na⁺ in the bath and the effects of benzamil. The comparisons were conducted in two series of paired measurements from the same experimental data set in order to minimize the effects of rundown. In the first series of measurements (Series 3), inward Li currents were tested for sensitivity to benzamil. A representative experiment is shown in Figure 3A and its I-V releationship in Figure 3B. The I-V relationships were very similar for both Na⁺ (B2) and Li⁺ (B4) and benzamil markedly decreased the inward currents for both ions. Specifically, benzamil significantly inhibited the inward Li⁺ whole-cell current (at -100 mV) by 58.8%; from I_-100 = -1656 ± 126 pA to -682 ± 86 pA (n = 5) (Figure 3B, C and Table 1). Similarly, benzamil decreased (Series 3) the inward Na⁺ current at -100 mV by 66.8% (from I_-100 = -1922 ± 96 pA to -639 ± 151 pA), a decrease that was not significantly different from the decrease by benzamil seen in the previous series (Series 2) of experiments with Na⁺ in the bath (Figure 3B and Table 1).

It was to be expected that there would be no change in the observed V_r between the presence and absence of benzamil since the reversal potential for both Na⁺-selectective and NSC channels are both = 0. By contrast, if Li⁺ is more permeable than Na⁺, V_r would be expected to become more positive (see second series, below).

In the second series of measurements (Series 4), Na⁺ and Li⁺ I-V relationships were compared immediately before and after the solution change (points 3 and 4' in Figure 3A). Rapid change of the bath solution from a Na⁺-rich solution (B2) to a Li⁺-rich solution (B4) led to small but consistent changes in the current. The average inward current and conductance at large negative voltage (-100 mV) were significantly larger for Li⁺ than for Na⁺, by 19.9% (I_-100 = -1699 ±149 pA vs -1416 ±181 pA, n = 5) and 14.9% (g_-100 = 16.2 ± 1.4 nS vs 14.1 ± 1.6 nS, n = 5). Further, the reversal voltage was significantly more positive (7.1 ± 1.2 mV vs 0.1 ± 0.4 mV) (Figure 4).

The shift in V_r was analyzed with the Goldman-Hodgkin-Katz (GHK) equation as applied to bionic substitutions to estimate the relative membrane permeability to Li⁺ and Na⁺[12]:

where ΔV_r is the change in reversal potential, R the universal gas constant, T the absolute temperature, z the valence, F the Faraday constant, [Li]_o the extracellular Li⁺ concentration, [Na]_o the extracellular Na⁺ concentration and P_Li and P_Na the permeabilities of the membrane for Li⁺ and Na⁺. The shift of V_r by 7.0 mV in the current study corresponds to a P_Li/P_Na ratio of 1.3. This result is within the range of reports by others for the ratio in αβγ-ENaC (1.3 [13, 14], 1.6 [11], 1.8 [15]).

We hypothesized that the residual current after inhibition by benzamil could be transported through nonselective cation channels, although a substantial contribution of NSC channels was unlikely since K⁺ conductance was relatively small. Alternatively, the residual current could be a patch leakage current.

It was first observed that transcripts of TRPV4, an amiloride-insensitive NSC channel, were present by both gene array and RT-PCR (Table 2). The amiloride-insensitive channels HCN1-4 are permeable to K⁺ and Na⁺ (P_K/P_Na ~4 [16]) and were found to be absent by RT-PCR, although some false positives in the gene array were found (Table 2).

We tested for the presence of benzamil-insensitive currents by comparing the inward current in the presence and absence of benzamil with a substitution of the impermeable cation NMDG⁺ for Na⁺ in the bath.

In this series of experiments (Series 5), the initial phase before benzamil did not quickly reach a stable condition so that parameters in benzamil were compared to those after washout. Nonetheless, it is clear that the inward current and conductance in benzamil were significantly less than the inward Na⁺ current and conductance after washout (Figure 5, Table 1). Importantly, there was no significant difference between the inward current in benzamil (I_-100 = -741 ± 65 pA, n = 5) and in NMDG⁺ (I_-100 = -802 ± 118 pA, n = 5, B5). These results demonstrate that there were no cation currents other than the benzamil-sensitive channels under these conditions.

Na⁺ in the pipette was reduced to 15 mM (P4) in order to confirm the Na⁺-selectivity of the current in this experimental series. The large decrease in V_r between Na⁺ (B2) and NMDG⁺ (B5) (from 29.8 ± 1.0 mV to -27.8 ± 3.6 mV, n = 5; Figure 5 and Table 1) is consistent with a large Na⁺ conductance. The small, but significant, decrease in V_r between ± benzamil could conceivably be due to the presence of an H⁺ conductance and the small pH gradient between intra- and extracellular solutions. The equilibrium potential for Na⁺ was +60 mV and for H⁺ was +6 mV. The absence of any significant difference (Table 1) in I₊₁₀₀ (14834 ± 342 pA vs 1666 ± 243 pA, n = 5) in NMDG⁺ and Na⁺ baths (B5 and B2) is consistent with the current at +100 mV being carried predominantly by the efflux of intracellular Na⁺, since the pipette concentration (P4) remained the same for both bath solutions.

Tissue was obtained, prepared and analyzed as described previously [22]. Mice (C57BL/6) were anesthetized with 4% tribromoethanol (0.014 ml/g body wt ip) and killed following a protocol approved by the Institutional Animal Care and Use Committee of Kansas State University. The lateral wall of the cochlear apical turn was isolated from the temporal bone, the stria vascularis removed and Reissner's membrane was gently pushed into the same plane as the spiral ligament. Reissner's membrane consists of an epithelial cell layer facing the endolymph and a mesothelial cell layer facing the perilymph, the two layers being separated by a basement membrane. The tissue was folded with the apical membrane of the epithelium on the exterior of the loop in order to form patch clamp seals with the apical membrane (Additional file 1: Fig. S3).

Currents were recorded using the whole-cell configuration of the patch clamp technique, similar to our previous study [22]. Patch pipettes were made from borosilicate glass capillaries (1B150F; World Precision Instruments, Sarasota, FL), pulled in three stages with an electrode puller (Narishige, East Meadow, NY). Inner diameter of the tip was approximately 2 μm and after heat polishing the pipettes had resistances of 3.3 - 9.6 MΩ (n = 31) in the different bath solutions.

Currents were recorded with an Axopatch 200A amplifier (Axon Instruments, Foster City, CA) and low-pass filtered at 1 kHz. Current signals were digitized at 5 kHz using a computer with a Digidata 1322A (Axon Instruments) and pCLAMP 9 software (clampex9, Axon Instruments) was used for voltage step protocol recordings. In addition, AxoScope software (Axon Instruments) with MiniDigi 1A (Axon Instruments) data acquisition hardware was simultaneously used for continuous trace recordings and current signals were digitized at 1 kHz. The temperature was maintained at 37°C on a glass-bottomed bath chamber by a continuous, warmed perfusion and supplemental chamber heater.

For the pipette solution containing high K⁺ (P3) and bath containing high Na⁺ (B2), the data were corrected for the liquid junction potential of 4.1 mV. For the pipette solution containing low Na⁺ and high NMDG⁺ (P4) and baths containing high Na⁺ (B2) and high NMDG⁺ (B5), the data were corrected for the liquid junction potentials of -6.9 mV and -6.2 mV (the latter includes the liquid junction potential change of 0.7 mV at the reference electrode). Liquid junction potentials in symmetric Na-Methanesulfonate, K-Methanesulfonate, NMDG-Methanesulfonate were near zero, since voltage offset was adjusted when the pipette was immersed in each bath solution. The Li⁺ data were corrected by adding the 0.3 mV junction potential change at the reference electrode.

Two voltage protocols were used. Protocol 1: A series of pulses were applied for 300 ms each from a holding potential of 0 mV to steps from +100 to -100 mV in 20 mV decrements. The protocol was repeated every 16.7 s. Current-voltage (I-V) relationships were determined from averaging 30 ms of steady-state currents near the end of each voltage step. Protocol 2: This series only differed from the first protocol by the step duration (200 mS) and the repetition rate (5.3 s). The I-V relationships were obtained from these current averages plotted against the command voltage, corrected for the liquid junction potential and an estimate of the voltage drop across the pipette resistance. These corrected curves were used to obtain the currents and conductances at ±100 mV (I_-100, I₊₁₀₀, g₊₁₀₀ and g_-100) and the reversal voltage, V_r (Additional file 1: Fig. S4). Data were plotted with Origin software, version 7 (OriginLab Software, Northampton, MA).

Table 3 shows the composition of the solutions of pipette and bath for electrophysiological recordings. The osmolarity of these solutions was about 311-318 mOsm. Benzamil (Sigma #B-2417) was predissolved (1 mM) in dimethylsulfoxide (DMSO), and was used at a final concentration of 1 μM benzamil and 0.1% DMSO. Pipette solutions (P1, P2, P3, P4) were buffered to 100 nM free Ca²⁺[23]. The KCl-rich pipette solution (P1) was adjusted to pH 7.3 at room temperature, which corresponds to about 7.2 at 37°C and alters the Ca²⁺/EGTA ratio at that elevated temperature. Pipettes using P2, P3 and P4 were backfilled with 15 mM Cl^- substituted for methanesulfonate in order to make stable contact with the Ag/AgCl electrode. The backfill solutions contained the dye fast green to visualize the interface with the tip solution, which filled the first 10 mm of the pipette. All pipette solutions for patch clamp were passed through 0.22 μm cellulose acetate filters (Corning).

Four mice (C57BL/6, 4-10 weeks old) were anesthetized with 4% tribromoethanol (0.014 ml/g body wt ip) and killed for each collection of total RNA. The two temporal bones were removed from each mouse after a transcardial perfusion of a phosphate-buffered saline (PBS) solution to reduce both cell swelling and contamination from blood cells. Reissner's membrane was dissected in PBS, which was changed twice during isolation of Reissner's membrane to reduce cross contamination from other tissues. All procedures were approved by the Institutional Animal Care and Use Committee of Kansas State University. Total RNA was isolated from the tissue using an RNeasy Micro Kit with carrier RNA (Qiagen, Valencia, CA). RNA used as positive control was from mouse kidney, mouse brain (Ambion, Austin, TX). The quality and quantity of RNA were determined with a Bioanalyzer (Additional file 1: Fig. S3B; Agilent, Palo Alto, CA). RT-PCR was conducted as described before [22].

Data were expressed as the mean ± S.E.M. (n = number of whole cell patches). Increases and decreases in current and conductance were determined by Student's paired or unpaired t-test as appropriate for the specific data sets. Differences were considered statistically significant at a level of P < 0.05 (Microsoft Excel).