Pseudohypoaldosteronism type 1 (PHA‐1) is an inherited disease characterized by severe neonatal salt‐wasting and caused by mutations in subunits of the amiloride‐sensitive epithelial sodium channel (ENaC). A missense mutation (G37S) of the human ENaC β subunit that causes loss of ENaC function and PHA‐1 replaces a glycine that is conserved in the N‐terminus of all members of the ENaC gene family. We now report an investigation of the mechanism of channel inactivation by this mutation. Homologous mutations, introduced into α, β or γ subunits, all significantly reduce macroscopic sodium channel currents recorded in Xenopus laevis oocytes. Quantitative determination of the number of channel molecules present at the cell surface showed no significant differences in surface expression of mutant compared with wild‐type channels. Single channel conductances and ion selectivities of the mutant channels were identical to that of wild‐type. These results suggest that the decrease in macroscopic Na currents is due to a decrease in channel open probability (Po), suggesting that mutations of a conserved glycine in the N‐terminus of ENaC subunits change ENaC channel gating, which would explain the disease pathophysiology. Single channel recordings of channels containing the mutant α subunit (αG95S) directly demonstrate a striking reduction in Po. We propose that this mutation favors a gating mode characterized by short‐open and long‐closed times. We suggest that determination of the gating mode of ENaC is a key regulator of channel activity.
The amiloride‐sensitive epithelial Na channel (ENaC) is a highly selective Na channel found at the apical membrane of salt‐reabsorbing tight epithelia of tissues including the distal nephron, the distal colon, the salivary and sweat glands and the lung. In these polarized epithelia, the ENaC‐mediated entry of sodium into the cell represents the rate‐limiting step for the vectorial movement of sodium from the mucosal to the serosal side. In the kidney, ENaC activity is controlled by aldosterone, serving to maintain salt homeostasis and blood pressure (Rossier and Palmer, 1992; Rossier et al., 1994).
The primary structure of ENaC has been elucidated by functional expression cloning in Xenopus oocytes. It is made up of at least three homologous subunits (α, β and γENaC) which share ∼35% identity in their amino acid sequences (Canessa et al., 1993, 1994b). Each subunit has two transmembrane domains with short cytoplasmic N‐ and C‐termini and a large extracellular loop (Canessa et al., 1994a; Renard et al., 1994; Snyder et al., 1994). The biophysical properties and the pharmacological profile of ENaC expressed in Xenopus oocytes are similar to those for the native channel in the distal nephron (Palmer and Frindt, 1986; Canessa et al., 1994b).
The importance of ENaC for the regulation of sodium balance, blood volume and blood pressure has been demonstrated by the finding of mutations in either the β or the γ subunits that cause increased ENaC activity and Liddle's disease, a rare, inherited form of salt‐sensitive hypertension (Shimkets et al., 1994; Hansson et al., 1995). Recently, it has been shown that pseudohypoaldosteronism type 1 (PHA‐1), an inherited disease characterized by severe neonatal salt‐wasting, hyperkalaemia, metabolic acidosis and unresponsiveness to mineralocorticoid hormones, is also caused by mutations in ENaC subunits (Chang et al., 1996), further confirming the important role of ENaC for salt and water homeostasis. These mutations lead either to a frameshift or a premature stop codon in the α subunit or to an amino acid exchange in a conserved region at the N‐terminus of the β subunit (Chang et al., 1996). The functional consequences of this mutation in the β subunit (G37S) were investigated by co‐expression of the βG37S subunit with wild‐type α and γ subunits in Xenopus oocytes. Oocytes expressing mutant channels showed an ENaC activity which was reduced to ∼40% of the activity of oocytes expressing wild‐type channels. Oocytes expressing the mutant β subunit, however, still had significantly higher activity than oocytes expressing only α and γ subunits (Chang et al., 1996). This finding demonstrated that this mutation does not result in complete loss of function, leaving unresolved the mechanism by which this mutation leads to a reduction in ENaC activity. Here we report that introduction of the homologous mutation into the α subunit (αG95S) drastically reduces the open probability (Po) of the single channel, suggesting that the N‐terminal domain containing the conserved glycine controls the gating of ENaC.
Macroscopic amiloride‐sensitive sodium currents of the G37S mutation in βENaC and of the corresponding mutations in α and γENaC
Gly37 of βENaC is conserved in all known subunits of this gene family and is located in a highly conserved stretch of ∼20 amino acids (Figure 1A) preceding the first transmembrane‐spanning region (Figure 1B). According to the established ENaC membrane topology, this N‐terminal segment is facing the cytoplasmic side of the membrane (Canessa et al., 1994a; Renard et al., 1994; Snyder et al., 1994). As the overall homology between different ENaC subunits is low (30–35% at the amino acid level), the high degree of identity of the segment containing Gly37 suggests an important role for normal ENaC function. Elucidation of the mechanism by which the Gly37→Ser (G37S) mutation decreases ENaC activity could thus help to define the role of this domain.
Gly→Ser mutations were introduced at homologous positions in α, β and γENaC. As an index of ENaC channel activity, we measured the macroscopic amiloride‐sensitive Na+ current (INa). As shown in Figure 2A a highly significant decrease in ENaC activity was seen with the three mutants. αG95S showed the most drastic effect, reducing INa to ∼15% of wild‐type activity (P < 0.001); γG40S reduced INa to ∼55% (P <0.001), similar to the effect of βG37S (62% of wild‐type activity; P < 0.02). Thus, the conserved Gly residue of all three subunits is essential for normal channel activity. Co‐expression of more than one mutant subunit resulted in a cumulative effect. When βG37S was co‐injected with γG40S, ENaC activity decreased to 11% of wild‐type activity (P <0.001, Figure 2B) and co‐expression of three mutant subunits (αG95SβG37SγG40S) resulted in a complete loss of ENaC activity (P <0.001, Figure 2B).
As the measure of INa would be invalid if the mutations changed the affinity of the channel for amiloride, we verified for the βG37S and the αG95S mutant subunits that these point mutations do not change the affinity of ENaC for amiloride (data not shown).
INa depends on several variables given by the following relationship:
where gNa is the unit conductance for a single ENaC channel, N is the number of conductive ENaCs per unit area of membrane, Po is the open probability of a single ENaC, E is the membrane potential and ENa is the reversal potential for sodium. To keep the driving force for the sodium ion (E − ENa) in our whole cell experiments largely constant, we always strongly hyperpolarized the oocyte membrane to −100 mV. Moreover, we observed that oocytes expressing ENaC wild‐type channels had a significantly more depolarized resting potential [−15.7 ± 1.6 mV (mean ± SEM), n = 78] than oocytes expressing mutant βG37S channels [−28.9 ± 1.8 mV (mean ± SEM), n = 84, P <0.001]. This observation suggests that βG37S ENaC mutants were less charged with sodium than oocytes expressing wild‐type ENaC, consistent with the observed reduction of channel activity. This would tend to increase the driving force for Na entering the cell (thus INa) and cannot account for the observed decrease in INa with mutant ENaC. Therefore, the decrease in INa caused by Gly→Ser mutations can either involve a change in single channel conductance reflecting an impaired permeation of Na+ ions through the channel, a decrease in channel open probability, or fewer channels expressed at the cell surface. The later decrease in cell surface expression of channel mutants can be related either to a misassembly of channel subunits or to a defective targeting of the ENaC channel to the cell surface.
Synthesis and assembly of mutant subunits
The highest degree of homology within the region containing Gly37 is found between αENaC and a recently cloned new member of this gene family, δENaC (Figure 1A), both of which are characterized by their ability to associate with β and γ subunits and to enhance ENaC activity (Canessa et al., 1994b; Waldmann et al., 1995), raising the possibility that this domain is involved in the assembly of ENaC subunits.
To assess the proper association of ENaC subunits, we used an epitope (FLAG)‐tagged β or γ subunit to immunoprecipitate, with an anti‐FLAG monoclonal antibody, the αβγENaC complex from microsomes prepared from 35S metabolically labeled ENaC‐expressing oocytes. As shown in Figure 3, anti‐FLAG antibodies do not immunoprecipitate proteins from oocytes expressing three untagged ENaC subunits (Figure 3, lane 1). When one of the subunits (β or γ) carried the FLAG epitope, we co‐immunoprecipitated three predominant proteins from oocytes expressing α, β and γENaC (Figure 3, lanes 2 and 7). These proteins correspond to α, β and γENaC as previously shown in immunoprecipitations using antisera specific for each of these proteins (Duc et al., 1994). As unassembled epitope‐tagged subunits will also be immunoprecipitated, greater amounts of immunoprecipitated tagged βENaC as compared with tagged γENaC are probably due to a more efficient translation of βENaC in Xenopus oocytes. From oocytes injected with only two subunits we precipitated only two of the predominant proteins, demonstrating the specificity of the co‐immunoprecipitation (Figure 3, lanes 5, 6 and 9). From oocytes expressing one of the mutant subunits (αG95S, βG37S or γG40S) together with two wild‐type subunits, one of which was epitope tagged, we co‐immunoprecipitated all three predominant proteins (Figure 3, lanes 3, 4 and 8). This shows that the Gly→Ser mutation does not inhibit the association of a mutant subunit with the two wild‐type subunits and the formation of a heteromultimeric αβγENaC channel. Thus, the decrease in INa resulting from expression of Gly→Ser mutant subunits with wild‐type subunits does not result from the activity of ENaC channels containing only two subunits (αβ or αγ) which are characterized by a low level of amiloride‐sensitive current. Moreover, since similar amounts of channel subunits are co‐immunoprecipitated from oocytes expressing wild‐type and mutant channels (Figure 3), it is unlikely that altered ENaC assembly accounts for the loss of channel function caused by these mutations.
Cell surface expression of mutant channels
To determine if the decrease in INa is due to a lower number of ENaC channel molecules expressed at the cell surface, we used a quantitative assay based on the binding of an [125I]IgG anti‐FLAG monoclonal antibody directed against a FLAG epitope in the extracellular loop of the α and the β subunit. We have demonstrated that the introduction of the FLAG epitope in the proximal part of the extracellular loop of ENaC subunits does not change the physiological and pharmacological properties of the channel (Firsov et al., 1996). To compare cell surface expression directly with activity of ENaC, we first determined the specific [125I]IgG anti‐FLAG binding and we then measured the INa of the same oocytes.
As shown in Figure 4, mutant ENaC bearing the FLAG epitope on the α and β subunit showed INa indistinguishable from that found in channels not containing a FLAG epitope (Figure 2A) (5% of wild‐type activity for αG95S, P <0.001, 47% for βG37S, P = 0.009, and 63% for γG40S, P = 0.03). In addition, the absolute values for INa (μA/oocyte) were similar to those measured without tagged subunits. In contrast to the reduced INa of mutant channels, there was no significant reduction in the number of mutant channels at the cell surface, as reflected in quantitation of specific binding of anti‐FLAG IgG to channels carrying the Gly→Ser mutations in either the α, β or γ subunit. The mean of specific anti‐FLAG binding sites per oocyte ranged from 0.7 to 1.05 fmol/oocyte (Figure 4). This finding indicates that the observed reduction in INa resulting from these mutations must result from reduced sodium current per channel molecule present at the cell surface. This is best illustrated (Figure 4B) by the striking decrease in the ratio of the whole cell Na+ current (INa) to the bound anti‐FLAG IgG; the mean of this ratio for individual oocytes was 4.11 μA/fmol for the ENaC wild‐type, 0.15 μA/fmol for the αG95S (P <0.001), and 1.35 and 1.73 μA/fmol (P <0.005) for the βG37S and the γG40S, respectively.
Thus, the Gly→Ser mutations impair neither the assembly of ENaC subunits in a heteromeric channel nor the trafficking and expression of ENaC at the cell surface. This suggests that these mutations lead to an intrinsic channel dysfunction characterized by a reduction of the average Na+ current passing through channels present at the cell surface. As shown by Equation 1, these mutations can decrease the macroscopic sodium currents by reducing either the single channel conductance or the probability of the channel being in the open conformation. To distinguish between these possibilities, we analyzed the properties of single channels by measuring microscopic sodium current by patch–clamp.
Microscopic sodium currents: single channel properties of wild‐type versus mutant ENaC (βG37S and αG95S)
Single channel currents were obtained in cell‐attached patch configuration from oocytes expressing either ENaC wild‐type, or ENaC containing either βG37S or αG95S. For the wild‐type and mutants, the currents resulting from channel openings are in the pA range and the current–voltage relationship of the βG37S mutant shows the characteristic higher permeability for Li+ over Na+ as observed for the wild‐type channel (Canessa et al., 1994b). The single channel conductance of the βG37S mutant was 5 pS in the presence of Na+ ions and 7.5 pS for Li+ ions, values similar to those measured for the ENaC wild‐type expressed in Xenopus oocytes. Considering that the corresponding mutation in the α subunit also exhibits a similar high conductance for Li+ ions (gLi = 6.9 pS), these single channel data indicate that the channel's ionic conductance and selectivity are not altered by the Gly→Ser mutations. From these experiments, we conclude that the glycine mutation must alter the open probability of the channel.
The single channel recording of wild‐type ENaC shows long lasting opening and closing events whose duration exceed 1 s, interrupted by short closures or openings lasting <1 s. The average open probability of ENaC wild‐type is ∼0.5 but, as has been shown in the native tissue, this open probability is quite variable from channel to channel, ranging from 0.05 to 0.9. The individual values are not distributed normally but are mainly <0.25 or >0.75 (Palmer and Frindt, 1996). The gating behavior of wild‐type ENaC was interpreted as showing two different transitions between open and closed states (Palmer and Frindt, 1996). The recording of the αG95S mutant shows characteristics different from the gating behavior of the ENaC wild‐type: this mutant channel spends most of its time in the closed conformation, and the long closing events are interrupted by short duration opening events. The open probability of the αG95S mutant was significantly reduced to 0.04 ± 0.012 (mean ± SD, n = 14, P <0.001) compared with wild‐type [Po = 0.48 ± 0.26 (mean ± SD), n = 33, see Schild et al., 1995]. Kinetic analysis of channel recordings showed that the mean open time duration of the opening events averaged 74 ± 54 ms (mean ± SD, n = 14); during these channel recordings (up to 15 min), no single opening event lasting >1 s could be observed. Thus, the αG95S mutation affects channel gating in which the channel mutant opens for only a short period of time, as if the mutation renders the open conformation of the channel unstable compared with ENaC wild‐type.
Single channel recording of the βG37S mutant was not statistically different from wild‐type, exhibiting similar long and short transitions between open and closed states, with a Po of 0.58 ± 0.26 (mean ± SD, n = 7). A 40–50% decrease in whole‐cell INa measured for the β or γ mutant channels should correspond to a Po of 0.25–0.3, a difference from wild‐type channels that cannot be detected statistically (see Discussion).
The G37S mutation does not act via the phosphorylation of a PKC site created de novo by the mutation
The serine replacing the glycine generates in all the three mutant subunits a site (SXXR, Figure 1A) which could potentially be phosphorylated by protein kinase C (PKC) and consequently modify the channel open probability. To address this question, different amino acid substitutions of Gly37 of the β subunit were performed: (i) a conservative alanine substitution, (ii) a cysteine substitution that resembles the serine but is not amenable to phosphorylation, and (iii) an aspartate substitution mimicking a constitutive phosphorylation by introduction of a negative charge at this position. As shown in Figure 6, the introduction of the cysteine and the aspartate, but not of the alanine, resulted in a significant decrease in INa (65 and 1% of wild‐type activity, respectively). As the reduction of INa in oocytes expressing G37C was comparable with that in oocytes expressing G37S (57% of wild‐type activity), phosphorylation of Ser37 is unlikely to account for the observed decrease in channel activity. This was supported further by the observation that neither stimulation of PKC by phorbol 12‐myristate‐13‐acetate (PMA) nor indirect stimulation of protein kinase A (PKA) by IBMX or forskolin changed INa differentially in oocytes expressing either βG37S mutant or ENaC wild‐type (data not shown). The differential effect of our amino acid substitutions at codon 37 on channel function suggests that rather than abnormal phosphorylation, steric effects on the protein structure induced by mutations at position 37 are involved in the abnormal function of the channel.
Mutations of a conserved glycine in the N‐terminus of ENaC subunits that affect channel gating and explain disease pathophysiology
In order to investigate systematically by which mechanism a mutant Gly→Ser subunit present in the αβγENaC complex reduces INa, we have, according to Equation 1, measured INa, gNa and N. As discussed above, a change in the driving force (E−ENa) does not account for the observed decrease in INa. The results of our measurements of macroscopic sodium currents, INa, and the number N of expressed ENaCs on the same oocyte as well as the results of our measurements of gNa are summarized in Table I. According to Equation 1, it is possible to deduce from this relative values for Po. This demonstrates a significant reduction of Po for the mutant α as well as the mutant β subunit (Table I).
A drastic reduction in the open probability (Po = 0.04) could also be demonstrated directly for the mutant α subunit. This reduction in Po was accompanied by a reduction in the mean open time duration, which was >1 s for the ENaC wild‐type versus 74 ± 54 ms for the αG95S mutant. This 10‐fold decrease in the Po and the mean open time duration could account for the ∼10‐fold reduction of INa found in whole‐cell voltage clamp experiments. For the mutant β subunit, we could not demonstrate directly the expected ≈2‐fold decrease in Po. This is not surprising because the gating behavior shows a large variation between channels, and channels present in the membrane patch represent only a small fraction of the channels present at the cell surface. Moreover, the measurement of Po by patch–clamp critically depends on the confidence with which the number of channels present in the patch can be determined. As accurate assessment of this number is unfortunately impossible; the Po as inferred from patch–clamp can only be an estimation. In contrast, with our macroscopic binding assay, we quantify the total number N of channels independently of their function. Therefore, we conclude that it is very likely that the corresponding mutation in the β subunit—hlike the α subunit—also reduces Po. Our findings suggest that mutation of a conserved glycine residue in the N‐terminus of ENaC subunits changes ENaC gating and that this change in ENaC gating accounts for the reduced ENaC activity causing PHA‐1.
ENaC gating modes: αG95S mutant favors a gating mode showing short open and long closed times
The gating of ENaC is a complex molecular phenomenon which still remains poorly understood. This complexity is illustrated in Figure 5, where channel recordings show the presence of multiple channels and opening and closing events of variable duration. In general, channels that can be observed in membrane patches exhibit long openings and long closures with dwell times in the second range interrupted by shorter events lasting 10 to several hundred milliseconds. Such a complex gating pattern is likely to be responsible for the highly variable open probability measured for ENaC. The Po as well as the mean open and closed times as inferred from patch–clamp studies represent, in general, the mean of many different channels present in the same patch; the values for individual channels are very difficult to determine.
A refinement of our understanding of ENaC gating comes, therefore, from a recent study in native tissue. Palmer and Frindt (1996) analysed 19 patches collected over 4 years which contained individual ENaC channels. They identified at least two different gating ‘modes’, one with a low open probability Po<0.25 characterized by long closing events and short openings and the other one with a high Po>0.75 showing long openings and short closures. The cellular or molecular factors that determine the mode in which the channel is gating or regulate transitions between the different gating modes are not yet understood.
We recently have been able to correlate in individual oocytes the number of ENaC subunits expressed at the cell surface with the measured amiloride‐sensitive Na+ current (Firsov et al., 1996). This correlation shows that the overall open probability of the channels expressed at the cell surface is <0.05, which means that the channels exhibiting a Po of 0.25 or higher represent only a small fraction of the ENaC pool at the cell surface. The predominant pool of ENaC expressed at the cell surface with a very low Po will escape detection in standard patch–clamp experiments.
In the context of our present knowledge of ENaC gating, how can we interpret the changes in Po induced by the G→S mutations? The low open probability Po and the short open dwell time duration and the long closing events characterizing the αG95S mutant are reminiscent of the low Po gating mode described for the ENaC wild‐type channels. For example, one of the channels reported by Palmer and Frindt (1996) had a Po of 0.093 and a mean open time of 77 ms, similar to the mean values for the αG95S mutant. This observation suggests that the G→S mutation might shift the equilibrium between two normal alternative gating modes, favoring a mode with short open times and long closed times. The fact that we never observed long openings in recording of αG95S mutant channels suggests that the αG95S cannot escape the low Po gating mode whereas the βG37S can. As the glycine is not only conserved between all the subunits of the ENaC gene family but is also located in a highly conserved segment at the intracellular N‐terminus of ENaC subunits, it is tempting to speculate that this segment is involved in the normal gating of ENaCs. Mutating the conserved glycine would thus impair the normal functioning of this ‘gating domain’, thereby changing the equilibrium between different gating modes.
Different gating modes of ENaC: physiological importance for aldosterone regulation of channel activity
The different gating ‘modes’ of ENaC may have a physiological significance. The mineralocorticoid aldosterone is known to increase the activity of the epithelial sodium channel in most salt‐reabsorbing tight epithelia. The mechanism by which aldosterone regulates ENaC activity is, however, still controversial. In a study investigating the effect of aldosterone on single channel characteristics of ENaC in the aldosterone‐responsive epithelial cell line A6 (Kemendy et al., 1992), the authors found a reduction of Po and the mean open time of ENaCs as cells were depleted of aldosterone, suggesting that the hormone regulates ENaC activity by changing its single channel characteristics. The mean open time changed from 1615 ± 224 ms to 40.6 ms as cells were deprived for 36–72 h of aldosterone; the mean closed time changed only slightly, if at all (Kemendy et al., 1992). This shows that one of the mechanisms by which aldosterone regulates ENaC activity might be by shifting channels already present at the apical membrane of epithelial cells in a different gating ‘mode’. As the mean open time of ENaC of A6 cells stimulated by aldosterone resembles the mean open time of wild‐type ENaC expressed in Xenopus oocytes (>1 s), whereas the mean open time of ENaC of A6 cells deprived of aldosterone resembles that of the αG95S mutant (74 ± 54 ms), it seems possible that the segment containing the conserved glycine is not only involved in the gating of ENaC but is also part of a functional domain of ENaC directly or indirectly regulated by aldosterone. Assuming that the G37S mutation in the β subunit also changes the equilibrium between gating modes of ENaC channels, this would explain why PHA‐1 patients with this mutation do not respond appropriately to their elevated plasma aldosterone.
Mutations in ENaC subunits either reduce or increase ENaC activity causing human genetic diseases, which underscores its biological importance in sodium balance and in the control of blood pressure
Mutations found in ENaC subunits provoking genetic diseases have underscored the importance of the amiloride‐sensitive sodium channel for salt and water homeostasis. These are mutations leading either to a gain‐of‐function characterized by increased sodium reabsorption in the distal nephron leading to Liddle's disease (Shimkets et al., 1994) or to a loss‐of‐function characterized by a decreased sodium reabsorption in the distal nephron leading to PHA‐1 (Chang et al., 1996). As can be expected, gain‐of‐function mutations are operating on a specialized mechanism and, therefore, are limited to a defined region; in the case of Liddle's disease the C‐terminus (Figure 1B) of the β and γ subunits (Shimkets et al., 1994; Hansson et al., 1995; Schild et al., 1995; Tamura et al., 1996). In contrast, loss‐of‐function mutations can, in principle, operate on any mechanism disturbing normal channel function and are, therefore, expected to be much more heterogeneous. In the case of PHA‐1, they lead either to a disruption of the protein structure of the α subunit (Chang et al., 1996) or, by the exchange of a conserved glycine in the β subunit (as we have shown in this report), very likely to the impairment of channel gating. The mechanisms involved in the other PHA‐1 pedigrees are not yet known, but are probably quite different from the mechanism reported here. The two mutations reported for the α subunit lead either to a frameshift prior to the putative first transmembrane domain or to a premature stop codon in the extracellular loop prior to the putative second transmembrane domain (Chang et al., 1996). As the α subunit is required for ENaC activity (Canessa et al., 1993), it is likely that the frameshift mutation (I68frameshift), which results in an α subunit containing only the intracellular N‐terminus, leads to a non‐functional channel protein. The premature stop codon in the extracellular loop (R508stop) results in a protein containing the intracellular N‐terminus, the first transmembrane domain and most of the extracellular loop, but lacks the second transmembrane domain and the intracellular C‐terminus. It thus resembles two alternative splice products of the α subunit which have been identified in rat taste tissue and which delete part of the coding sequence thereby introducing premature stop codons ∼50 amino acids upstream of the stop codon at position 508 (Li et al., 1995). Co‐expression of these alternative splice products with wild‐type β and γENaC in Xenopus oocytes did not induce amiloride‐sensitive sodium currents (Li et al., 1995), suggesting that the R508stop mutation also results in an inactive channel protein.
In a recent report (Strautnieks et al., 1996), another mutation leading to PHA‐1 has been identified in γENaC, introducing a splice site mutation which leads either to a disrupted protein structure or to the exchange of three highly conserved amino acids by a new one. The functional consequences of this mutation have not been reported so far, but it seems likely that it will also reduce ENaC activity. In spite of this heterogeneity of loss‐of‐function mutations, they might—as we have shown here—hvery well contribute to the elucidation of structure–function relationships and a better understanding of ENaC function.
Materials and methods
Point mutations were introduced into α, β and γrENaC by recombinant PCR using Pfu DNA polymerase (Stratagene). Briefly, two fragments were amplified with primers containing the mutation in a short overlapping region, joined by recombinant PCR, digested by appropriate restriction endonucleases (αENaC: EcoRI–XbaI, βENaC: ApaI, γENaC: PstI) and ligated into the corresponding rat ENaC cDNA in the vector pSD5. For αENaC, a cDNA clone was used in which the EcoRI site in the insert had been eliminated without changing the amino acid sequence, and for γENaC one in which the PstI site in the vector had been eliminated. PCR‐derived fragments were entirely sequenced.
Some of the ENaC subunits used for co‐immunoprecipitations and the binding assay (see below) were epitope tagged. α, β and γrENaC were tagged with the FLAG reporter octapeptide (DYKDDDDK) which is recognized by the anti‐FLAG M2 IgG1 use monoclonal antibody (International Biotechnology). The FLAG epitope was introduced into the proximal part of the extracellular loop of ENaC subunits (Canessa et al., 1994a; Renard et al., 1994; Snyder et al., 1994) using recombinant PCR. In αrENaC, the FLAG sequence replaced the sequence T194RQAGARR201, in βrENaC it was inserted between T137 and S138, and in γrENaC it replaced the sequence M142PSTLEGT149.
Expression of ENaC in X.laevis oocytes
Capped cRNA was synthesized by SP6 RNA polymerase from wild‐type and mutant α, β and γENaC cDNA which had been linearized by BglII (α and βrENaC) or PvuII (γrENaC). Equal saturating amounts of each subunit (3 ng of cRNA of each subunit) were injected into stage V–VI oocytes of X.laevis. Oocytes were either kept in modified Barth's saline (MBS: 88 mM NaCl, 1.0 mM KCl, 2.4 mM NaHCO3, 0.41 mM CaCl2, 0.33 mM CaNO3, 0.82 mM MgSO4, 10 mM HEPES, pH 7.2) and the macroscopic amiloride‐sensitive sodium current (INa) was determined 24 h after injection or in low‐sodium MBS (10 mM NaCl, 90 mM NMDG‐Cl, 5 mM KCl, 0.41 mM CaCl2, 0.33 mM CaNO3, 0.82 mM MgSO4) and INa was determined 48 h after injection. INa was determined using two‐electrode voltage‐clamp as described elsewhere (Canessa et al., 1993) in the presence of 5 μM amiloride.
Results are reported as means ± SEM and represent the mean of n independent experiments in which the average amiloride‐sensitive sodium current INa was measured for 4–8 individual oocytes originating from the same frog. Data on graphs represent the percentage of INa measured for mutant channels relative to INa of wild‐type channels. Statistical analysis was done with the unpaired t‐test.
Xenopus oocytes injected with cRNAs coding for wild‐type or mutant α, β and γENaC were incubated for 14 h in MBS supplemented with 1.0 mCi/ml of [35S]methionine. Either β or γ subunits were tagged with the FLAG epitope. Oocytes containing no tagged subunit served as a negative control. Microsomal membranes were prepared as described (Geering et al., 1989), microsomal proteins solubilized in a digitonin washing buffer [DWB: 20 mM Tris pH 7.6, 100 mM NaCl, 0.5% digitonin, 1 mM phenylmethylsulfonyl fluoride (PMSF), Leupeptin, Antipain, Pepstatin A (Sigma) (LAP)], and the incorporated radioactivity determined. Equal amounts of counts were supplemented to a final concentration of 2% bovine serum albumin and 2 mM PMSF, and ENaC subunits precipitated using the anti‐FLAG M2 IgG1 monoclonal antibody o/n at 4°C. The immunoprecipitates were recovered with protein G–Sepharose, washed four times with DWB and twice with DWB not containing digitonin, and separated on a 5–8% gradient SDS–polyacrylamide gel. The gel was exposed for 4–7 days on Kodak XS film.
Anti‐FLAG M2 IgG1 monoclonal antibody was iodinated using the Iodo‐Beads Iodination reagent (Pierce) and carrier‐free [125I]Na (Amersham) according to the Pierce protocol. Iodinated antibody had a specific activity of 5–20×1017 c.p.m./mol and was utilized during 2 months after the synthesis. Binding of the iodinated antibody to oocytes expressing the FLAG‐containing ENaC subunits was determined 20–24 h after the cRNA injection. Twelve oocytes in each experimental group were transferred into a 2 ml Eppendorf tube containing MBS solution supplemented with 10% heat‐inactivated calf serum and incubated for 30 min on ice. The binding was started by the addition of 12 nM (Kd = 3 nM) of the iodinated antibody in a final volume of 100 μl. After 1 h incubation on ice, the oocytes were washed eight times with 1 ml of MBS solution supplemented with 5% calf serum and then transferred individually into tubes containing 250 μl of MBS. The samples were counted and the same oocytes utilized for two‐electrode voltage‐clamp studies. Non‐specific binding was determined on oocytes injected with untagged ENaC subunits.
Results are reported as means ± SEM and represent the mean of six independent experiments in which the average anti‐FLAG IgG binding and the average amiloride‐sensitive sodium current INa were measured for 5–8 individual oocytes originating from the same frog. Data on graphs represent the percentage of IgG bound and of INa measured for mutant channels relative to the IgG bound and of INa of wild‐type channels. Statistical analysis was done with the unpaired t‐test.
Channel recordings were obtained from defolliculated oocytes in the cell‐attached configuration. Pipet solution contained in mM: 110 NaCl or 110 LiCl; 2.5 KCl; 1.8 CaCl2; 10 HEPES–NaOH pH 7.2; bath solution was identical to pipet solution except that 110 mM KCl replaced NaCl. Data were sampled at 2 kHz and filtered at 200–500 Hz for analysis. Single channel conductance was obtained from amplitude histograms of open state current of at least 50 transitions. The open state probability (Po) was determined from 4 to 15 min segments of digitized single channel data using a 50% threshold criterion to distinguish open from closed states.
We thank Nicole Skarda‐Coderey for secretarial assistance and Hans‐Peter Gaeggeler for technical help. This work was supported by grants from the Swiss National Fund for Scientific Research to B.C.R. (# 31‐43384.95) and to L.S. (# 31‐39435.93). R.P.L. is an investigator of the Howard Hughes Medical Institute.
↵† S.Gründer and D.Firsov contributed equally to this work
- Copyright © 1997 European Molecular Biology Organization