Pore-lining residues of MEC-4 and MEC-10 channel subunits tune the Caenorhabditis elegans degenerin channel's response to shear stress

The Caenorhabditis elegans MEC-4/MEC-10 channel mediates the worm's response to gentle body touch and is activated by laminar shear stress (LSS) when expressed in Xenopus oocytes. Substitutions at multiple sites within the second transmembrane domain (TM2) of MEC-4 or MEC-10 abolish the gentle touch response in worms, but the roles of these residues in mechanosensing are unclear. The present study therefore examined the role of specific MEC-4 and MEC-10 TM2 residues in the channel's response to LSS. We found that introducing mutations within the TM2 of MEC-4 or MEC-10 not only altered channel activity, but also affected the channel's response to LSS. This response was enhanced by Cys substitutions at selected MEC-4 sites (Phe715, Gly716, Gln718, and Leu719) between the degenerin and the putative amiloride-binding sites in this subunit. In contrast, the LSS response was largely blunted in MEC-10 variants bearing single Cys substitutions in the regions preceding and following the amiloride-binding site (Gly677–Leu681), as well as with four MEC-10 touch-deficient mutations that introduced charged residues into the TM2 domain. An enhanced response to LSS was observed with a MEC-10 mutation in the putative selectivity filter. Overall, MEC-4 or MEC-10 mutants that altered the channel's LSS response are primarily clustered between the degenerin site and the selectivity filter, a region that probably forms the narrowest portion of the channel pore. Our results suggest that pore-lining residues of MEC-4 and MEC-10 have important yet different roles in tuning the channel's response to mechanical forces.

Proteins of the ENaC/degenerin family share a common topology: short cytoplasmic N and C termini and two transmembrane helices connected by a large, highly organized extracellular domain (17,18). Several lines of evidence (19 -22) suggest that the channel's gate resides in the outer part of the pore, in the vicinity of the "degenerin" site. The introduction of bulky residues to this site in ENaC subunits, MEC-4, and acid-sensing ion channel (ASIC) subunits result in a dramatic increase in channel open probability (P o ). Strikingly, a large number of touch-deficient alleles have mutations within the pore-lining region of MEC-4 (16 alleles at 13 sites) and MEC-10 (five alleles), highlighting the importance of the channel's gate and pore in sensing mechanical stimuli (23). Additionally, pore-lining residues within the second transmembrane domain (TM2) of ENaC subunits are important for modulating channel gating in response to flow-mediated laminar shear stress (LSS) (24 -26). However, the role of the pore-ling residues of MEC-4 or MEC-10 in mechanosensation has not been systemically examined in worms.
The present study used the LSS response to examine the role of TM2 domains of the two pore-forming subunits, MEC-4 and MEC-10, in channel gating by a mechanical force. We found that the LSS response was reversible in oocytes expressing both MEC-10 and MEC-4d (A713T, a gain-of-function mutation at the MEC-4 degenerin site). The LSS response was largely diminished by touch-deficient mec-10 alleles that introduce charged residues into the channel's pore, as well as by Cys substitution at five sites within TM2 of MEC-10. In contrast, we observed an enhanced LSS response with five MEC-4 mutants, as well as one MEC-10 mutant. In addition, MEC-4 or MEC-10 mutants that altered the LSS response are primarily clustered between the degenerin site and the selectivity filter, a region that forms the narrowest portion of the channel pore in the related channel, ASIC1 (17,27). Together, our data support the hypothesis that pore-lining residues within the TM2 domain of MEC-4 or MEC-10 have important yet different roles in accommodating conformational rearrangements within the channel pore during channel gating in response to mechanical forces.

Touch-deficient mutations of MEC-10
C. elegans MEC-4/MEC-10 channels are activated by flowmediated LSS in Xenopus oocytes, and MEC-10 is required for a robust LSS response (14). In oocytes expressing MEC-4 A713T (MEC-4d) with MEC-10, MEC-2, and MEC-6, the LSS response is reversible, as the currents returned to the baseline when the vertical flow was stopped (Fig. 1a). Worms bearing mec-10 alleles that introduce charged residues within the TM2 of MEC-10 (see Fig. 1b, G676R (u20), L679R (u390), G680E (u332), and G684R (e1715)) are defective in their response to gentle body touch (16). We therefore examined whether these touch-deficient mutations of MEC-10 also blunt the channel's LSS response when expressed in Xenopus oocytes. An individual MEC-10 mutant or WT MEC-10 was co-expressed with MEC-4d as well as MEC-6 and MEC-2. Whole-cell Na ϩ currents were measured just before the initiation of LSS (I basal ) and at 40 s after initiation of LSS (I LSS ). The applied shear stress (0.12 dynes/cm 2 ) is comparable with the external force required to evoke mechanoreceptor currents in the worm's touch receptor neurons (13). In response to LSS of 0.12 dynes/ cm 2 , I LSS /I basal was 2.0 Ϯ 0.4 in oocytes expressing WT MEC-10 ( Fig. 1d). The LSS response was significantly diminished by each of the four MEC-10 mutations (Fig. 1). In addition, time constants of the LSS response () were also reduced by each MEC-10 mutant (Fig. 1d), suggesting a faster but diminished channel activation by LSS relative to the WT channel.
The blunted response and faster kinetics ( Fig. 1) recapitulate the LSS response of channels lacking MEC-10 (i.e. MEC-4d homomeric channels) (14). This observation raises the question whether the individual MEC-10 mutants are, like WT MEC-10, present in the functional channel complex. It has been shown that the association of MEC-10 with MEC-4d in the channel complex reduces channel activity and enhances the channel's sensitivity to pore blockers (28). We therefore examined the inhibitory effect of benzamil on channels containing WT or mutant MEC-10. As shown in Fig. 2, co-expression of WT MEC-10 with MEC-4d resulted in a left shift in the benzamil dose-response curve, indicating enhanced sensitivity to benzamil. We then used the difference in benzamil sensitivity as a readout to determine whether MEC-10 mutants with blunted LSS responses were present within functional channel complexes. The dose-response curves for G676R and G680E fell between those for MEC-4d alone (no MEC-10) and MEC-4d with MEC-10 (WT MEC-10) curves (Fig. 2). The dose-response curves for L679R and G684R were similar to that of channels with MEC-10 ( Fig. 2). Our data showing that these touch-deficient MEC-10 mutants were able to confer higher benzamil sensitivity as compared with MEC-4d alone channels (no MEC-10) suggest that the mutant MEC-10 subunits are a part of the functional complex. In addition, co-expression of these MEC-10 mutants reduced baseline whole-cell Na ϩ currents, when compared with either homomeric MEC-4d channels (no MEC-10) or channels with WT MEC-10 (Table 1). Taken MEC-4d/MEC-10 channel were exposed to repetitive stimulation by LSS (0.12 dynes/cm 2 ; gray bars). Benzamil was added to the bath perfusion at the end of each recording to determine the leak current. b, sequence alignment of the pore-lining residues within MEC-4 and MEC-10. The degenerin site, proposed amiloride-binding site, and selective filter region are noted. MEC-10 sites where mutations cause a touch-deficient phenotype in worms are underlined. c, representative traces of oocytes expressing MEC-10 touch-deficient mutants. The LSS response (0.12 dynes/cm 2 , 2 min) was examined in oocytes expressing MEC-4d/MEC-2/MEC-6 (no MEC-10), with WT MEC-10 or MEC-10 touch-deficient mutants (as indicated). Dashed lines indicate the zero of the y axis. d, LSS response of oocytes expressing MEC-10 touch-deficient mutants. The magnitude (I LSS /I basal , top) of the LSS response was determined as a ratio between the peak response of the whole-cell current following the initiation of LSS (I LSS ) and the basal whole cell current before initiation of LSS (I basal ). The time constants of channel activation by LSS (Tau) (bottom) were determined by fitting the current increase within the first 40 s following LSS stimulation as described under "Experimental procedures." The experiment was repeated in two batches of oocytes for the mutant channels, and the number of recordings (n) for each group is indicated. Significant changes when compared with WT MEC-10 (***, p Ͻ 0.001; ****, p Ͻ 0.0001) or homomeric MEC-4d channels ( † † † †, p Ͻ 0.0001) were analyzed with one-way ANOVA followed by a Bonferroni test.

Degenerin TM2 modifies channel gating
together, our results suggest that these touch-deficient MEC-10 mutants are components of channel complexes that are deficient in their response to LSS.

Pore-lining residues of MEC-10
As the LSS response was largely lost with MEC-10 touchdeficient mutations located within the channel's pore (Fig. 1), we decided to systematically investigate the role of the poreforming region of MEC-10 in the channel's response to LSS. We introduced Cys residues at individual sites within MEC-10 TM2, from the degenerin site (Ala 673 ) through the putative selectivity filter to Val 687 (Fig. 3a). We chose to introduce Cys substitutions to allow us to perform additional studies with Cys reactive reagents, if warranted. Each mutant MEC-10 subunit was co-expressed with MEC-4d, MEC-2, and MEC-6 in Xenopus oocytes, and the LSS response was examined 4 -7 days after cRNA injection. As shown in Fig. 3, the magnitude of channel activation was significantly reduced in five mutants (G677C, H678C, L679C, G680C, and L681C). These reduced LSS responses were characterized by faster kinetics (smaller values). Strikingly, LSS of 0.12 dynes/cm 2 elicited a 7.1 Ϯ 2.8-fold increase of whole-cell Na ϩ currents in oocytes expressing MEC-10 S683C, compared with the 2.9 Ϯ 1.8-fold increase of whole-cell Na ϩ currents with WT MEC-10 ( Fig. 3c). This Figure 2. The channel's sensitivity to benzamil is altered by select MEC-10 mutations. Benzamil dose response was measured in oocytes expressing MEC-4d homomeric channel (no MEC-10) or channels with either WT MEC-10 or individual MEC-10 mutants. a, oocytes were sequentially perfused with TEV buffer or TEV buffer containing each concentration of benzamil (10 Ϫ9 , 10 Ϫ8 , 10 Ϫ7 , 10 Ϫ6 , 10 Ϫ5 , and 10 Ϫ4 M). Whole-cell Na ϩ currents were measured at a holding potential of Ϫ60 mV. b-f, benzamil dose-response curves. Whole-cell Na ϩ currents measured in the presence of increasing concentrations of benzamil (I) were normalized to the basal current in the absence of benzamil (I 0 ) to determine relative currents (I/I 0 ). Dose-response curves were generated by plotting I/I 0 values as a function of benzamil concentrations (M) and fitting data to a sigmoidal equation (Equation 1). The IC 50 was defined as the concentration of benzamil that inhibits 50% of the whole-cell Na ϩ current (y). Each data point represents the mean value from 6 -10 measurements from at least two batches of oocytes. Error bars, S.D. in each panel. Comparisons between the dose-response curves were analyzed with the extra sum-of-squares F test of the best-fit log IC 50 values.

Degenerin TM2 modifies channel gating
greater LSS response was characterized by slower kinetics, as was the LSS response of the V685C mutant (Fig. 3c). Our results suggest that key sites within the pore segment of MEC-10 accommodate conformational changes during channel gating in response to LSS.
As shown in Fig. 3, the LSS response was blunted by mutations at a stretch of residues (Gly 677 -Leu 681 ) adjacent to the putative amiloride binding site (Gly 680 ) of MEC-10. It is possible that the blunted LSS response resulted from the absence of mutant MEC-10 in the channel complex. To test this possibility, we probed whether these MEC-10 mutants sensitized the channel to external cadmium (Cd 2ϩ ). Multiple endogenous Cys residues are present within the pore-lining segments of MEC-4 and MEC-10, which are potentially accessible to Cd 2ϩ . Increased concentrations of Cd 2ϩ were applied to oocytes expressing MEC-4d or MEC-4d/MEC-10 channels via basal perfusion (Fig. 4a). We found that the presence of WT MEC-10 in channel complexes increased the channel's sensitivity to Cd 2ϩ (Fig. 4a). For example, 0.1 mM Cd 2ϩ reduced whole-cell Na ϩ currents by 26.1 Ϯ 11.6% in oocytes co-expressing WT MEC-10 and MEC-4d (Fig. 4b). This inhibition was completely reversible (Fig. 4a). In contrast, perfusion with 0.1 mM Cd 2ϩ had minimal effect on whole-cell currents of homomeric MEC-4d channels (no MEC-10, 4.4 Ϯ 7.7% inhibition), suggesting that MEC-10 is required for higher Cd 2ϩ sensitivity. The difference in Cd 2ϩ sensitivity provided us with an approach to test whether certain MEC-10 mutants were present in the channel complex. Indeed, we observed that 0.1 mM Cd 2ϩ inhibited the currents of oocytes expressing each of the LSS-deficient MEC-10 mutants (G677C, H678C, L679C, G680C, and L681C) to a similar or greater magnitude than that seen with oocytes expressing WT MEC-10 (Fig. 4b). These data suggest that these MEC-10 mutants were present in channel complexes that failed to respond to shear stress. In addition, Cd 2ϩ also inhibited LSSactivated currents in oocytes expressing MEC-4d/MEC-10 channels (19.3 Ϯ 4.8%; Fig. 4, c and d).

MEC-4 degenerin site
LSS activates ENaC by increasing channel P o (29, 30). Accordingly, this response was diminished in ENaCs that were "locked open" (e.g. bearing a Lys substitution at the degenerin site of the ␤ subunit (␤S518K) or modification of an introduced Cys near the degenerin site of the ␣ subunit (␣S580C)) (29). There are three gain-of-function mec-4 alleles, e1611 (MEC-4 A713V), u231 (MEC-4 A713T), and u56 (MEC-4 A713D), that cause degeneration of touch receptor neurons and loss of touch sensitivity in worms by increasing channel P o , leading to excess Ca 2ϩ and/or Na ϩ influx (9,31). We therefore examined the LSS response of each gain-of-function MEC-4 variant (A713D, A713T, and A713V) as well as A713C. As shown in Fig. 5, the magnitude of the LSS response was not altered by A713C, A713T, or A713V but was nearly abolished by the A713D mutant, as compared with WT MEC-4. These results are consistent with previous reports that channels with an A713D mutation have the highest channel P o (ϳ0.76) at baseline among MEC-4 degenerin site mutants (22) and suggest that the blunted LSS response seen in this mutant (A713D) is due to a high intrinsic P o .

Pore-lining residues of MEC-4
We also screened the TM2 of MEC-4 (residues Asn 710 -Thr 729 ) for sites that are important for the LSS response. We generated each Cys mutant in the MEC-4d background, as oocytes expressing WT MEC-4 exhibited very low currents even after prolonged incubation (14). Each MEC-4d mutant was co-expressed with WT MEC-10, MEC-2, and MEC-6 in Xenopus oocytes. Five MEC-4d mutations increased the magnitude of channel activation (I LSS /I basal ): F715C, G716C, Q718C, L719C, and L728C (Fig. 6c). Partially overlapping with this group, five MEC-4d mutations reduced the rate of the LSS response: N710C, Q718C, L719C, I725C, and L728C (Fig. 6d). The most striking changes were observed with Q718C, L719C, and L728C, where the magnitude of the LSS response was nearly doubled (Fig. 6c). These results contrast with the effect of mutating the sites equivalent to Gln 718 and Leu 719 in MEC-10 (His 678 and Leu 679 ), which blunted the LSS response (Fig. 3). Na ϩ currents were not detected in channels bearing Cys substitutions at Gly 720 , Trp 722 , Gly 724 , Ser 726 , or Thr 729 of MEC-4d, even after prolonged incubation ( Table 2). We therefore examined whether MEC-4d mutations near the putative selectivity filter affect the channel's permeability to Na ϩ . MEC-4d/ MEC-10 channels expressed in Xenopus oocytes are selective to Na ϩ , but also permeable to K ϩ (Table 3). We did not detect benzamil-sensitive currents in oocytes expressing MEC-4d G720C, W722C, G724C, or S726C when Li ϩ or K ϩ (100 mM) was the major carrying ion in the buffer. Only modest changes in cation selectivity were found for channels with the I725C or L728C mutant (Table 3). Table 1 Select MEC-10 mutations alter channel activity Xenopus oocytes were injected with cRNA mixtures of MEC-4 A713T (MEC-4d), MEC-2, and MEC-6 in the presence of WT MEC-10 or individual MEC-10 mutants. Basal currents (before the initiation of LSS stimulation) were measured at Ϫ60 mV by perfusing oocytes with NaCl-110 TEV solution. At the end of each recording, 20 M benzamil (50 M for channels without MEC-10) was added to the bath to determine leak currents. n indicates the number of oocytes assessed for each group. Statistical comparisons of whole-cell Na ϩ currents measured in oocytes expressing WT MEC-10 versus mutant channels were determined with one-way ANOVA followed by a Bonferroni test (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; ****, p Ͻ 0.0001; ns, not significant). Ϫ0.72 0.30 14 **** L679R

Discussion
Worms with mutations at key sites within the TM2 of MEC-4 or MEC-10 exhibit an abnormal gentle touch response (8,13,16,23), indicating that the pore-lining residues of both subunits have key roles in mechanosensation. In the present study, we examined the role of TM2 residues of MEC-4 and MEC-10 in the channel's LSS response. We found that both subunits contain key sites within the pore-forming domain for the channel's LSS response (Figs. 1, 3, 5, and 6). However, the LSS response of channels with mutations in the pore-forming regions of MEC-4 or MEC-10 clearly differed. For example, introducing Cys mutations at the corresponding sites within TM2 of MEC-4 (Gln 718 and Leu 719 ) and MEC-10 (His 678 and Leu 679 ) had opposite effects on the channel's response to LSS (Figs. 3 and 6). When Cys mutants were placed between the degenerin site and putative amiloride binding site of MEC-4 or MEC-10, we found that the LSS response was enhanced by four MEC-4d mutants but dampened by five MEC-10 mutations (Figs. 1, 3, and 6). These results are consistent with different physiological roles of MEC-4 and MEC-10 in the worm's touch response. MEC-4 is expressed exclusively in the six touch receptor neurons, whereas MEC-10 has been detected in other sensory neurons and thus also implicated in the worm's harsh touch response (8,15). Functional ENaC is also a heterotrimeric channel that adopts an asymmetric organization within its pore, and porelining residues from individual subunits contributed differently to the channel's activation by LSS (25,26,32,33). It is possible that TM2 residues of MEC-4 or MEC-10 do not align symmetrically within the pore of the C. elegans degenerin channel. In addition, we found the LSS response was largely blunted in MEC-4 A713D, a degenerin mutation with a high P o at baseline (Fig. 5). It was previously reported that the A713D mutant has higher channel P o than other MEC-4 degenerin site mutants (the baseline P o is 0.76 for A713D and 0.36 for A713T). It is reasonable to postulate that the intermediate P o of A713T (ϳ0.36) allows the channel to open in response to LSS, whereas the blunted LSS response seen in A713D is due to a high intrinsic P o (ϳ0.76) (22).
In the resolved open-state structures of ASIC1, there is a break in the ␣ helix near the putative selectivity filter (GXS belt) within TM2. This feature may facilitate extensive conformational changes during gating (27). It is not clear whether the TM2 of MEC-4 or MEC-10 is a continuous helix or adopts an analogous helical break as TM2 of ASIC1 does. We generated a homology model of the MEC-4d/MEC-10 channel pore based

Degenerin TM2 modifies channel gating
on the structure of ASIC1 in the open state (Fig. 7). When we mapped the location of the functionally important residues identified in the present study, we found them in close proximity to the helical break in our model, clustering between the degenerin site (DEG in Fig. 7a) and the selectivity filter. For example, we observed a dramatic change in the LSS response (both magnitude and kinetics) when a Cys was introduced at MEC-10 Ser 683 , one residue before the selectivity filter GXS motif (Fig. 3). Meanwhile, the rate of the channel's LSS response was altered by mutating each site of the MEC-10 GXS motif (Gly 684 , Val 685 , and Ser 686 ) as well as MEC-4d Ile 725 , the equivalent site of MEC-10 Val 685 (Figs. 3, 6, and 7). We previously identified key residues for the channel's LSS response in the TM1 of MEC-10 (residues 130 -132 and 134 -137 (14)). Interestingly, these residues are predicted to neighbor the residues in the TM2 of MEC-10 that also blunt the channel's LSS response (Fig. 7b). Overall, we noticed that most mutants that altered the LSS response are clustered in a region that is predicted to form the narrowest portion of the channel pore, based on the resolved ASIC1 structure (Fig. 7). Together, our data support the hypothesis that the region preceding and including the GXS belt within MEC-4 and MEC-10 undergoes conformational changes during channel gating in response to LSS (27).
MEC-2 and MEC-6 were included in our experiments to enhance the functional expression of WT or mutant MEC-4d/ MEC-10 channels in oocytes. Neither MEC-2 nor MEC-6 contributes to the channel pore (11). Although we have previously  were examined in each experiment as a control. a, oocytes were exposed to increasing concentrations (0.1, 1, and 5 mM) of Cd 2ϩ dissolved in NaCl-110 TEV buffer for 40 s, followed by 40 s of washout with NaCl-110. At the end of each recording, benzamil was added to the bath to determine the leak currents. Tracings are representative of 3-4 recordings. b, the percentage of Cd 2ϩ (0.1 mM) inhibition on whole-cell Na ϩ currents was determined for each mutant channel as well as channels containing either WT MEC-10 or no MEC-10. Experiments were repeated in three batches of oocytes for each group with the number of recordings (n) indicated. Significant changes when compared with WT MEC-10 (***, p Ͻ 0.001; ****, p Ͻ 0.0001) or homomeric MEC-4d channels ( † † † †, p Ͻ 0.0001) were analyzed with one-way ANOVA followed by a Bonferroni test. c, the LSS was examined in oocytes expressing MEC-4d/MEC-10 channels, and Cd 2ϩ was applied through the vertical peptide to examine its inhibitory effect on whole-cell Na ϩ current in the presence of LSS. d, the percentages of Cd 2ϩ inhibition on LSS-activated currents were quantified. Experiments were repeated in two batches of oocytes (n ϭ 8). Statistical significance was analyzed with a one-sample Student's t test (****, p Ͻ 0.0001).

Degenerin TM2 modifies channel gating
shown that neither MEC-2 nor MEC-6 is required for the channel's response to LSS for WT MEC-4d/MEC-10 channels (14), we do not know whether the presence of MEC-2 or MEC-6 influences our results or the proposed channel pore model. Several Cys mutations generated in this study (MEC-4d G720C, W722C, G724C, or S726C) failed to generate functional channels in oocytes (Table 2), limiting our ability to test the role of MEC-4 TM2 in channel gating in response to LSS. Consistent with our observations, mutations introduced at Trp 722 , Gly 724 , or Ser 726 in MEC-4 cause a gentle touch defect in worms (23,34). It is not known whether these mutations affect the structure of the channel pore that influences channel conductance, ion selectivity, mechanical gating, or pore structure. As they are localized near the selectivity filter, we attempted to test whether these Cys mutants alter the channel's cation selectivity. However, our results suggest that these mutant channels do not carry Na ϩ , Li ϩ , or K ϩ currents. We suspect that these MEC-4d mutations reduce channel conductance, as ENaC subunits with specific mutations within the selectivity filter exhibited significantly reduced whole-cell currents and channel conductance (35,36).
In summary, we systemically examined the contributions of MEC-4 TM2 and MEC-10 TM2 in the channel's response to a mechanical force. Together, our results suggest that the poreforming segments between the degenerin site and the putative selectivity filter of both MEC-4 and MEC-10 have important but differing roles in accommodating conformational changes during channel gating in response to LSS. Future studies are needed to identify other sites and structural features within MEC-4 and MEC-10 that are involved in sensing mechanical stimuli and to determine whether similar sites or regions within other mechanosensitive members of the ENaC/degenerin family have roles in mechanosensing.

Site-directed mutagenesis and Xenopus oocyte expression
Single mutations were generated with the QuikChange kit (Agilent Technologies, Santa Clara, CA), using WT MEC-10 or

Degenerin TM2 modifies channel gating
MEC-4 or MEC-4 A713T (MEC-4d) as templates. Standard DNA sequencing was performed to confirm the desired mutations. cRNAs of WT or mutant subunits were synthesized and purified using the T7 mMESSAGE mMACHINE kit (Ambion Invitrogen). Defolliculated stage V-VI oocytes of Xenopus laevis were injected with cRNA mixtures of MEC-6 (1 ng) and MEC-2 (5 ng), along with WT or mutants MEC-10 and MEC-4d cRNA (5 ng) as indicated. MEC-2 and MEC-6 cRNAs were co-injected to enhance the functional expression of WT or mutant channels in oocytes. We presume that each component of the cRNA mixture was expressed at similar levels. Following microinjection, oocytes were initially maintained at 18°C in modified Barth's saline (MBS: 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3 , 15 mM HEPES, 0.3 mM Ca(NO 3 ) 2 , 0.41 mM CaCl 2 , 0.82 mM MgSO 4 , pH adjusted to 7.4) supplemented with 10 g/ml sodium penicillin, 10 g/ml streptomycin sulfate, and 100 g/ml gentamicin sulfate. On the following day, surviving oocytes were transferred into fresh MBS solution supplemented with 50 M benzamil before functional assays. Oocyte survival following cRNA injection improves when a channel blocker (benzamil) is present in the bath solution (37). Oocytes were incubated at 18°C for an additional 4 -7 days for optimal channel expression. Excess benzamil was removed by rinsing oocytes in benzamil-free MBS before starting each recording (14). All chemicals were purchased from Sigma-Aldrich unless stated otherwise. The protocol for harvesting oocytes from X. laevis was approved by the University of Pittsburgh's institutional animal care and use committee.

LSS response
Oocytes were placed in a recording chamber (20-mm diameter and 6 mm deep) that was constantly perfused with the NaCl-110 solution (110 mM NaCl, 2 mM KCl, 1.54 mM CaCl 2 , and 10 mM HEPES, pH adjusted to 7.4 with NaOH) at a rate of 3.5 ml/min. Ca 2ϩ is included in the bath solution, as we found that lowering or eliminating bath Ca 2ϩ results in a large leak current. Oocytes were clamped at Ϫ60 mV, whereas whole-cell Na ϩ currents were continuously recorded. 0.12 dynes/cm 2 of shear stress was applied by perfusion (1.5 ml/min) through a vertical pipette submerged near the top of the oocyte (38). At the end of each experiment, 20 M benzamil (50 M for channels with reduced sensitivity to benzamil) was added to bath perfusion to determine leak currents. Three measurements were taken: (i) the current measured just before the initiation of LSS (I basal ), (ii) the current measured 40 s following the initia- Table 2 Mutations of the pore-lining residues of MEC-4 alter channel activity Xenopus oocytes were injected with cRNA mixtures of WT MEC-10, MEC-2, and MEC-6 in the presence of the indicated MEC-4 constructs. Whole-cell basal currents (before the initiation of LSS stimulation) were measured at Ϫ60 mV by perfusing oocytes with NaCl-110 TEV solution. At the end of each recording, 20 M benzamil was added to the bath to determine the leak current. n indicates the number of measurements for each group. Significant changes in whole-cell Na ϩ currents when compared with MEC-4 A713T (MEC-4d) are noted (**, p Ͻ 0.01; ***, p Ͻ 0.01; ****, p Ͻ 0.0001; ns, not significant), determined with one-way ANOVA followed by a Bonferroni test).   Degenerin TM2 modifies channel gating tion of LSS (I LSS ), and (iii) the current measured after the application of benzamil. Whole-cell currents were corrected for the benzamil-insensitive component when determining I basal and I LSS . The magnitude of the LSS response was expressed as a ratio between the LSS-stimulated current (I LSS ) and the basal current (I basal ). Time constants for channel activation () were determined by fitting the first 40 s of current increases following the initiation of LSS with an exponential equation as previously described (38,39). To examine whether the LSS response is reversible, oocytes expressing MEC-4d/10/2/6 were exposed to four sequential LSS stimulations. Each LSS stimulation lasted 120 s, alternating with 240 s of basal bath perfusion. Benzamil was added via the basal perfusion at the end of each experiment to determine the leak current.

Benzamil dose-response relationship
Whole-cell Na ϩ currents were continuously recorded, whereas oocytes were clamped at Ϫ60 mV. For the first 60 s, basal currents (I 0 ) in the absence of benzamil were measured by perfusing oocytes with NaCl-110. A benzamil stock solution (0.1 M in DMSO) was diluted to a series of concentrations (10 Ϫ9 , 10 Ϫ8 , 10 Ϫ7 , 10 Ϫ6 , 10 Ϫ5 , and 10 Ϫ4 M) with NaCl-110 and delivered to oocytes in sequence. Oocytes were perfused with each concentration of benzamil for 20 s to obtain stabilized currents (I). Normalized Na ϩ currents (I/I 0 values) were plotted as a function of benzamil concentrations (M), and dose-response curves were generated by fitting data to the following sigmoidal equation (14), where X is the concentration of benzamil (M). The IC 50 is defined as the concentration of benzamil that inhibits 50% of the whole-cell Na ϩ current (y). Dose-response curves were compared by analyzing the best-fit log IC 50 values with the extra sum-of-squares F test.

Cation permeability
Different cation solutions (100 mM, Li ϩ , Na ϩ , K ϩ ) were prepared with the appropriate chloride salts (36,40). These solutions also contained 0.82 mM MgCl 2 , 0.41 mM CaCl 2 , 10 mM HEPES. In addition, 5 mM BaCl 2 and 10 mM tetraethylammonium chloride were included in bath solutions to block endogenous K ϩ currents. The present study did not test whether the addition of Mg 2ϩ , Ca 2ϩ , Ba 2ϩ , and tetraethylammonium alter channel permeation or selectivity. The pH was adjusted to 7.4 with a hydroxide of the main permeating cation (Li ϩ , Na ϩ , or K ϩ )inthesolution.Whole-cellcurrentswerecontinuouslymeasured while switching the perfusion buffer to each predominant cation. At the end of each experiment, 20 M benzamil was added to bath solution containing each cation to determine leak currents. Benzamil-sensitive currents recorded at Ϫ60 mV were used to calculate the ratios of Li ϩ current (I Liϩ ) and K ϩ current (I Kϩ ) relative to Na ϩ current (I Naϩ ).

Data and statistical analyses
Experiments were repeated with a minimum of two batches of oocytes obtained from different frogs. Electrophysiology data were analyzed using Clampfit (Molecular Devices, Sunnyvale, CA) and OriginPro (OriginLab Corp., Northampton, MA). Data were expressed as the mean Ϯ S.D. throughout. Prism (GraphPad Software, La Jolla, CA) was used for graphing and statistical analysis. Scatter-dot plots were used to show every data point along with a horizontal bar for the mean. Statistical comparisons were obtained from analysis of variance (ANOVA) followed by a Bonferroni test or Student's t test. A p value of Ͻ0.05 was considered statistically significantly different.