Characterization of an amiloride binding region in the -subunit of ENaC - 《American Journal of Physiology》美国生理学杂志干细胞之家



免疫细胞治疗专区	欢迎关注干细胞微信公众号

12 3 4 5 6 7 8 9 10 ... 25 下一页

返回列表

查看: 425204\|回复: 246	go Characterization of an amiloride binding region in the -subunit of ENaC [复制链接]

轻羽

新手上路

Rank: 1

积分: 0
威望: 0
包包: 0

楼主

发表于 2009-4-21 13:48 |只看该作者 |倒序浏览 |打印

作者：Ollie Kelly, Chaomei Lin, Mohan Ramkumar, Nina C. Saxena, Thomas R. Kleyman,  Douglas C. Eaton作者单位：1 Center for Cell and Molecular Signaling, Department of Physiology, and Graduate Program in Biochemistry, Cell, and Developmental Biology, Graduate Division of Biological and Biomedical Sciences, Emory University School of Medicine, Atlanta, Georgia 30322; and Department of Medicine, University of  7 U9 b$ T; e9 t) J. s
   # Y# o4 @& {( m( M4 v



         . @+ O* L- o. I7 e/ k9 r



                  8 H7 r0 R6 l# Q



      【摘要】
   One of the defining characteristics of the epithelial sodium channel (ENaC) is its block by the diuretic amiloride. This study investigates the role of the extracellular loop of the -subunit of ENaC in amiloride binding and stabilization. Mutations were generated in a region of the extracellular loop, residues 278-283. Deletion of this region, WYRFHY, resulted in a loss of amiloride binding to the channel. Channels formed from wild-type -subunits or -subunits containing point mutations in this region were examined and compared at the single-channel level. The open probabilities ( P o ) of wild-type channels were distributed into two populations: one with a high P o and one with a low P o. The mean open times of all the mutant channels were shorter than the mean open time of the wild-type (high- P o ) channel. Besides mutations Y279A and H282D, which had amiloride binding affinities similar to that of wild-type -ENaC, all other mutations in this region caused changes in the amiloride binding affinity of the channels compared with the wild-type channel. These data provide new insight into the relative position of the extracellular loop with respect to the pore of ENaC and its role in amiloride binding and channel gating.   [- q" w6 {3 L* [" _
      【关键词】 open probability extracellular loop channel pore sodium channel singlechannel recording
   EPITHELIAL SODIUM CHANNELS (ENaC) play a critical role in the control of blood pressure and regulation of total body sodium balance. Although the original work in which the channel components were cloned suggested that functional channels consist of three subunits,,, and ( 5 ), under appropriate circumstances, -subunits alone can form sodium-permeable channels ( 12, 13, 17 ). It has been proposed that some of the diversity in conductance, gating, and selectivity of functional ENaC may be due to different subunit combinations. The expression of -subunits alone, rather than the expression of all three subunits together, depends on the environmental conditions to which renal epithelial cells (A6) ( 7 ) or alveolar type II cells ( 14 ) are exposed. However, regardless of which channel type is expressed, one of the defining characteristics of all these channels is their block by the diuretic amiloride, a substituted pyrazinoylguanidine. The tertiary structure of all the subunits is similar: each subunit is predicted to span the membrane twice, to have a large extracellular loop, and to have short intracellular NH 2 and COOH termini ( 2, 6 ). Because amiloride blocks the channel from the extracellular surface of the channel protein, it is possible that one or more of the extracellular loops could play a role in amiloride's interaction with and block of the channel by stabilizing amiloride in the channel pore.
; B9 Z+ {. O. r( p  N7 Q+ h
About 70% of each ENaC subunit is extracellular. Structure-function studies of the extracellular loop have focused primarily on the region immediately preceding and including the second transmembrane (M2) domain of -ENaC. These data have implicated these regions in binding to amiloride and also play a role in other biophysical characteristics of the channel, including selectivity and gating ( 9, 15, 16, 26, 27 ).
  \; U9 G: V3 X# C5 X8 X
However, previous studies have identified a region in the extracellular loop of rat -ENaC between amino acids 278 and 283, with the sequence WYRFHY, that interacts with amiloride ( 12 ). When this stretch of amino acids, located just COOH terminal to the first transmembrane-spanning domain within a cysteinerich domain, is deleted or mutated, amiloride binding of these channels expressed in a bilayer system is dramatically reduced. The inhibitor constant ( K i ) for amiloride increased from 169 nM for the wild-type -subunit of rat ENaC (rENaC) to 26.5 µM for 278-283 -rENaC. Besides the deletion mutation, this study also demonstrated that the histidine residue in this tract could affect the amiloride sensitivity of the channel. Although expression of -ENaC in planar lipid bilayers has been used previously, there are questions about the expression and characteristics of -ENaC in the absence of associated proteins found in cellular expression systems. Therefore, we wished to examine the interaction of amiloride and ENaC in a more well-defined system expressing a simple form of the channel consisting of -subunits only. To this end, we examined amiloride binding of wild-type -ENaC subunits and -ENaC subunits that have been modified within the 278-283 region of the extracellular loop after expression of the constructs in Chinese hamster ovary (CHO) cells. Examining channels formed from -subunits alone allowed us to determine the particular residues in the extracellular loops of -ENaC that contribute to amiloride binding without having to consider the roles and influence of - and -ENaC and their extracellular loops.6 g4 @; M. u) V3 N$ c# r6 y* O0 W

The purpose of this study is to further examine mutations in this particular region in the extracellular loop of -ENaC. Channels formed from wild-type and mutant -subunits were examined by single-channel recording methods with and without amiloride. The single-channel characteristics of the channels formed from -ENaC subunits with point mutations were compared with those formed from wild-type -ENaC.2 l; M+ ^7 s% y% t  A. W5 j' m: I# X- O/ M
; [9 B3 K2 H, c0 d: N# {0 ]
MATERIALS AND METHODS! Y3 F4 J) y: u/ r& f

Generation of -ENaC Mutants* L$ d3 f$ H& \) X" j8 e- a8 b# \
# L. ~0 Z1 y2 \5 W# _9 v
Rat -ENaC mutants were generated as previously described ( 12 ) by designing PCR primers to amplify the 5' and 3' regions of -rENaC flanking the section to be deleted and introducing an Xho I restriction enzyme site 3' of the deletion so that the two PCR products corresponding to the 5' and 3' ends of -rENaC could be ligated after digestion with Xho I. The Xho I restriction site was generated by mutating nucleotide C942 to G and nucleotide C945 to G without altering the amino acid residues at positions 287 and 288. The 5' end, digested with Sma I and Xho I, along with the 3' end, digested with Xho I and Nsi I, were ligated into pSPORT, which had been digested with Sma I and Nsi I. The single amino acid mutations were generated using Altered Sites In Vitro Mutagenesis System (Promega, Madison, WI). Wild-type -rENaC was excised from pSPORT-1 by digestion with Sal I and Sph I and ligated into pALTER-1 at Eco RI and Sph I sites with an Eco RI- Not I- Sal I adaptor. The mutant -subunits generated by these methods and used in expression were W278G, Y279A, R280G, F281A, H282D, H282R, and Y283A. Wild-type and mutant -subunits were subcloned into Eco RV and Not I sites, respectively, of pCDNA3 (Invitrogen, Carlsbad, CA), a mammalian expression vector. Correct insertion was verified by restriction digests and sequencing. Additional mutations, W278A, Y279F, and R280K, were generated via PCR using primers across the mutation site and subcloning the product back into a wild-type -rENaC/pCDNA3 backbone. cDNA used for transfection was prepared using the Concert high-purity plasmid Maxiprep system (Invitrogen).) T; h; u* {" u- d3 l, i
2 X  `  T  o  P$ }  H2 i2 g8 u0 t
Cell Culture and Transfection
- h, R  v  F+ b- f- e
CHO cells (CHO-K1, American Type Culture Collection, Manassas, VA) were cultured in complete medium, F-12K nutrient mixture (Kaighn's modification), 10% FBS, 2 mM L -glutamine, and 1% penicillin-streptomycin (Invitrogen), in plastic tissue culture flasks (Corning, Corning, NY) at 37°C in a humidified incubator with 5% CO 2 in air and plated on 35-mm tissue culture dishes (Becton Dickinson Labware, Franklin Lake, NJ) at a density of 1 x 10 5 for transfection. Cells were transiently transfected with a total of 20 µg of cDNA, either wild-type or mutant -ENaC subunit, along with green fluorescent protein in a ratio of 2:1. A calcium phosphate transfection system (Invitrogen) was used to introduce the cDNAs into the cells 24 h after plating. After 4-6 h of incubation, the cells were treated with 15% glycerol and allowed to recover overnight in fresh complete medium.

Single-Channel Recordings and Analysis

Transfected cells were identified by their fluorescence using a Diaphot Epi-Fluorescence-2 attachment (Nikon Instruments, Melville, NY) on a Nikon Diaphot inverted microscope equipped with Hoffman modulation contrast (Modulation Optics, Greenvale, NY). Borosilicate glass pipettes (World Precision Instruments, Sarasota, FL) were pulled and fire polished using a glass microelectrode puller and microforge (Narishige Scientific Instrument Laboratory, Tokyo, Japan) to resistances of 5-8 M, and cell-attached patches with G seals were formed. The pipette and bath solutions were the same and contained (in mM) 150 NaCl, 2.8 KCl, 2 MgCl 2 ·6 H 2 O, 1 CaCl 2 ·2 H 2 O, and 10 HEPES (pH 7.4, 330 mosmol/kgH 2 O). The pipette solution was used to dilute the stock solution of 5 mM N -amidino-3,5-diamino-6-chloropyrazine carboxamide hydrochloride (amiloride hydrochloride; Sigma, St. Louis, MO) in ethanol to the appropriate working concentration (2 nM-10 µM).

To examine the effects of amiloride on channel activity, cell-attached patches were made with pipettes in which the tips were filled with regular saline by capillary action and then backfilled with an amiloride-containing pipette solution. The estimated diffusion time for amiloride within the pipette was calculated on the basis of Fick's second law of diffusion. By modeling the saline-filled tip of the pipette as a cone, the solution of Fick's equation is

( 1 )7 k8 i5 Z$ M/ `! Y
9 p7 |3 q& t7 T
where C tip is the concentration at the tip at time t, C pipette is the concentration in the pipette, D is the diffusion constant for amiloride, and l is the saline-filled length of the tip ( 2 mm). On the basis of diffusion constants for other molecules of amiloride's approximate molecular weight, modeling of amiloride diffusion suggests that, in a typical experiment, 10-15 min should be required for amiloride to reach the surface of the cell.
  d. G, u. J: u( c; e
We made practical observations of the mean open time of channels as amiloride approached the surface of the cell from within the backfilled pipette. Mean open time began to decrease 12-15 min after preparation of a pipette and patch formation and reached a steady state 15-17 min after filling of the pipette and patch formation. Because the average time to form a seal and begin recording was 4 min, data collected before 8 min and after 12 min were used as data before and in the presence of amiloride, respectively. This approach allowed amiloride-treated channels to act as their own controls. Data were obtained with an Axopatch-1B patch clamp (Axon Instruments, Union City, CA), filtered at 5 kHz with a low-pass Bessel filter (model LPF-100A, Warner Instrument, Hamden, CT), and digitized at 20 kHz utilizing a TL-1 DMA interface and Axotape 1.2 (Axon Instruments). Data were analyzed using pClamp 6 (Axon Instruments). For additional analysis and compilation of data, SigmaPlot (SPSS, Chicago, IL) and Freelance Graphics (Lotus Development, Armonk, NY) programs were used.
: Y$ \( s5 M" l# X% R5 y7 s
Statistics
. v$ n/ o  Y0 a4 q+ Q
Statistical analysis was performed with SigmaStat 2.0 (SPSS). Unless explicitly stated to the contrary, values are means ± SE. For comparisons of groups of values, one-way ANOVA was used to detect significant differences, and then Dunnett's post hoc test was applied to determine significant differences from wild-type values. For comparison of the values for two mutants, Student's t -test was used. In either case, P statistically significant. Curve fitting of dose-response curves was performed with the nonlinear curve-fitting module in SigmaPlot 8.0 (SPSS) using a Levenberg-Marquardt algorithm that provides estimates of the error of the fit parameters. Fits to interval histograms were based on the method of Sigworth and Sine ( 29 ) and used a simplex algorithm in pClamp 6 (Axon Instruments, Burlingame, CA).* w( j( `7 s" c# r6 P* |1 H; D

Derivation of Equations

Equation 9. Amiloride block of ENaC is usually considered a simple block of open channels and, therefore, can be represented by the following kinetic scheme
. l; p2 v/ {4 r7 O# q4 G' `$ x; Y) v8 u
( 2 )
; ^2 R4 G- D- T  c% w% }! |
Calculation of the open probability ( P o ) from this kinetic scheme depends on the mean duration of the three states and also on their relative frequency. Thus, for every open event, there are k b / k b k -1 blocked events and k -1 / k b k -1 closed events. Therefore, P o in the presence of amiloride ( P o ') is
8 M9 _  `/ ?' P9 R+ G, }
( 3 )2 T' S' e1 _* F7 P# {# q
9 a- ~& @) j* H2 h9 h, n! M
The mean durations in any state ( open, closed, and blocked ) are the reciprocals of the sums of the rates of leaving a state. Therefore, the mean durations can be replaced with the equivalent rate constant expressions to give. U  \! x6 o1 _. e0 {- U0 m2 V

( 4 )

or after multiplying the numerator and denominator by k b k -1

( 5 )
5 b! ~& [* @/ x& \7 i# G, g# [9 s
but because k 1 /( k 1 k -1 ) is the P o in the absence of amiloride, then

( 6 )# K5 c2 U1 B6 s( u0 V
+ g* m9 l& J9 _' U
Usually, when the effect of a drug is examined, the effect is expressed as a fraction of the response in the absence of drug, in this case, P o '/ P o. Dividing the expression above by P o gives- D  b2 ^. A* I$ R

( 7 )

The blocking rate ( k b ) is a pseudo-first-order rate constant that can be expressed as a second-order rate constant times the concentration of the blocker, amiloride, so
( P7 [1 t( L: V0 O! q
( 8 )

but k -b / k b is the equilibrium block of amiloride ( K Amil ), so that the final expression for the half-fractional block of P o is
" T8 T% {" E2 x/ t2 n4 @
( 9 )3 v$ M& d7 c0 V# o" j; r* o

Fractional block of P o is represented as K 0.5 values, indicating the concentration of amiloride () needed to obtain half block of P o.+ Q# e& v- W- J* t, _, H7 w0 n- B0 w
3 n: U0 c0 i! W# }
Equation 11. In the presence of amiloride, the kinetic scheme for the channel states is given in Eq. 2, where the mean open time ( open ) is composed of the two rates for leaving the open state ( k -1 and k b ), which is represented by the following equation
2 c9 n: v* a7 [: P# Y9 O
( 10 ), i0 `) |, B. f" e7 n+ q

The rate represented by k b is a pseudo-first-order rate constant that depends on the concentration of amiloride. Therefore, the expression for the rates of leaving the open state in the presence of amiloride is represented by

( 11 )6 O0 P, z9 U5 P* c+ j
+ |, d% O$ X% c% X/ r- d
which when plotted yields a line where the slope is k b, the blocking rate of amiloride, and the y -intercept is k -1, the closing rate in the absence of amiloride.

RESULTS

Properties of Wild-Type -ENaC

We initially examined the single-channel characteristics and amiloride sensitivity of channels formed from wild-type -ENaC alone. Figure 1 A illustrates the amiloride sensitivity of a channel consisting of only -ENaC subunits. In addition, from the current-voltage relation ( Fig. 1 B ), the conductance of the channel is 25 ± 3.2 pS, which is not statistically different from the conductance of the 21-pS nonselective cation channels formed from -ENaC subunits that we previously described in rat lung alveolar epithelial cells ( 13 ).

Fig. 1. A : wild-type -subunit epithelial sodium channel ( -ENaC) expressed in Chinese hamster ovary (CHO) cells in the absence or presence of amiloride. -ENaC can form channels that are sensitive to increasing concentrations of amiloride, as evidenced by decreasing open probability ( P o ) with increasing amiloride concentrations. Data are from cell-attached patches held at -50 mV. Records were software filtered at 500 Hz. B : current-voltage relation for wild-type -ENaC channel indicates that the conductance of these channels is 25 ± 3.2 pS.' G$ |, y6 I# _1 _( s! Y
# X- k. u5 S+ Q
Mutations in -ENaC

Table 1 includes point mutations generated at each position within the 278-283 tract of -rENaC. Besides these sets of mutations, this entire region, WYRFHY, was deleted.
$ {% ]9 \3 c( k6 O. }. T
Table 1. Point mutations generated at different amino acids within the 278-283 region of rat -ENaC. d0 \; A6 Z: F1 Z4 P7 a, k7 l
% N( h& x' F3 ^2 @& {6 L
Single-Channel Records From Mutant Channels( B* Y" F; j# N% z
; {% ]6 z) t2 N; o, w$ h
The total amount of transmembrane sodium current is represented by the expression in Eq. 12
4 n+ w7 j7 r) t% |' |' h
( 12 )
+ {1 d/ m% I7 J/ C
where I Na is the total current, i is the unitary current, N is the total number of channels, and P o is the mean open probability of one channel. Figure 2 shows that although there are no changes in the unitary current or the number of channels per patch, there are changes in the P o and kinetics of most of the mutant -ENaCs compared with wild-type channels. Two of the mutants, F281A and Y279A, which produced the most dramatic effects, caused a clear change in the way the channel transitioned from the open state to the closed state compared with wild-type channels. These observations imply that before we could examine the amiloride block of any of the -ENaC mutants, we first needed to characterize the biophysical properties of these mutant channels in the absence of amiloride.
& A- o6 f) H0 D) x* _
Fig. 2. Single-channel traces from wild-type (WT) and mutant -ENaC expressed in CHO cells. Data are from cell-attached patches held at -50 mV. Records were software filtered at 500 Hz. See MATERIALS AND METHODS for bath and pipette solutions, which were the same. Number and current amplitude of the channels remained the same between wild-type and mutant -subunit channels. There are differences in gating of the channels between open (o) and closed (c) states.8 ^+ D% f; p, ^% F

Kinetic Properties of Mutant and Wild-Type -ENaC Are Different

The kinetics of wild-type channels fall into two distinct groups, high P o ( 0.75) and low P o ( 0.25), and the means of these two groups are statistically different from one another ( Table 2 ). As might be expected, the mean open times and mean closed times for the two groups were also different ( Table 2 ). Because of the differences in P o, data from channels formed from -subunits with mutations were compared with the values for high- and low- P o wild-type groups. The P o for wild-type (high- P o ) and all mutant channels are different, except for channels formed by the Y283A mutation ( Fig. 3, Table 3 ). The mutant channels that are different from the wild-type (high- P o ) channels had lower P o. Interestingly, the P o of mutant channels was not different from that of wild-type (low- P o ) channels, suggesting that one effect on mutations in the 278-283 region was to switch channels from high- to low- P o forms. This change in P o could be due to a decrease in the open time or an increase in the closed time of the channels. The mean open times for the mutants are smaller than those of the wild-type (high- P o ) channels ( open = 0.495 s; Fig. 4 A, Table 3 ). The mean open times are important when examining amiloride block, because amiloride is predominantly an open channel blocker. There were also variations in the mean closed times. The mean closed times for 278-283, W278G, Y279A, and R280G were larger than those of wild-type (high- P o ) channels ( closed = 0.031 s). The mean closed time was smaller for mutation F281A than for wild-type (low- P o ) channels ( Fig. 4 B, Table 3 ).; v# H; l& d: h$ H: E% V/ b

Table 2. High and low P o of wild-type -channels; \( `. E, L+ ^) N9 K, m; {

Fig. 3. Mean P o values for wild-type and mutant -subunit-only channels. Each bar represents mean values. P o values are significantly lower for all mutant channels, except Y283A, than for wild-type (high- P o ) channels ( P n 3, except for R280G, where n = 2.3 n3 A9 Z/ R4 j9 T

Table 3. Comparison of P o, open, and closed for wild-type and mutant -ENaC

Fig. 4. A : mean open times of wild-type and mutant -ENaC. Error bars, SE; n 3, except for R280G, where n = 2. Mean open times were significantly lower for all mutant channels than for wild-type (high- P o ) channels ( P 2 orders of magnitude. Values are given in Table 3. B : mean closed times of wild-type and mutant -ENaC. Error bars, SE; n 3, except for R280G, where n = 2. Vertical axis is logarithmic. Values are given in Table 3.
2 r+ ~) L7 c. T# c; X: s
It is clear that these mutations have an effect on the total sodium current by affecting the P o of channels formed from mutant -subunits compared with wild-type channels. These data may be useful in understanding the overall secondary structure of this region.

Amiloride-Induced Changes in P o Are Altered in Mutant Channels( b( i1 G, f- X/ O4 v6 ], b
8 k3 L. x* t3 H/ i: X. d* ~* T& B
Examination of the effect of amiloride on the P o of mutant channels addresses the role of the amino acids at 278-283, WYRFHY, in amiloride block of the channel. Fractional block of P o is represented as K 0.5, indicating the concentration of amiloride needed to obtain half block of P o according to Eq. 13. The model for the dose-response curves was4 X7 h) V- n1 l8 c
0 ^9 z1 C) a: q6 n# H
( 13 )6 H3 m$ d4 L% k# O  \' q# A" }

where F(c) is the normalized concentration function (e.g., the concentration dependence of P o or open ),  is the amiloride concentration, and K 0.5 is the half-blocking dose of amiloride based on the Michaelis-Menten equation.7 R; P: n. h, x6 T9 N8 ~- r1 {

The deletion of WYRFHY from the extracellular loop of -ENaC, when expressed in CHO cells, causes a loss of amiloride sensitivity in the expressed channels with K 0.5 20 µM. These data are consistent with published results for the 278-283 mutation expressed in planar lipid bilayers ( 12 ). This region was further examined for its contributions to amiloride sensitivity by expression in CHO cells of individual point mutations generated in this region. These mutations fell into two groups when their amiloride sensitivities were compared with those of wild-type -ENaC: those more sensitive to amiloride and those less sensitive to amiloride (Figs. 5 and 6 ). Channels formed by mutations W278G, R280G, and H282D in -ENaC are more sensitive to amiloride than are wild-type ENaC. Although the K 0.5 for amiloride of wild-type -ENaC is 0.029 µM, the K 0.5 values for amiloride of these mutations are 0.006 µM for W278G and R280G and 0.022 µM for H282D. Mutations that are less sensitive to amiloride include Y279A, F281A, and Y283A, with K 0.5 values for amiloride of 0.22, 0.26, and 0.63 µM, respectively ( Table 4 ).

Fig. 5. Dose-response curves of effects of amiloride on relative change in P o for wild-type and mutant channels formed from -ENaC subunits alone.
$ i; H' K, G2 ~" D  r$ V& ^
Fig. 6. Amiloride concentration needed to obtain half block of P o ( K 0.5 ) for wild-type and mutant -channels expressed in CHO cells. Solid bars, channels formed from -subunits with point mutations that are more sensitive to amiloride than are wild-type channels; gray bars, channels that are less sensitive than wild-type -ENaC to amiloride. Error bars, SE; n 3, except R280G, where n = 2. All mutations are different from wild-type channels ( P Table 4.8 ?! |% v0 v3 w8 r0 r) a& `7 `
  G+ x: w4 O) G$ s% O( ~& T
Table 4. K 0.5 of wild-type and mutant -ENaC in Fig. 6
5 [$ g9 W) ?, T, v. o% k( U
Additional Mutations and Effects on Amiloride Block of ENaC

The previous results with the first set of mutations prompted the generation of additional mutations to address the role of certain amino acid side chains in amiloride binding to the channel. We wanted to determine whether making conservative replacements within the WYRFHY tract would produce an amiloride sensitivity of the channels similar to that of wild-type channels. These new mutations, W278A, Y279F, and R280K, were expressed in CHO cells and compared with wild-type -ENaC and the original mutations generated at these positions (Figs. 7 and 8, Table 5 ). The restoration of amiloride sensitivity to wild-type levels by Y279F and R280K mutant channels supports the roles of these residues in the overall binding of amiloride to the wild-type channel.

Fig. 7. Dose-response curves of effects of amiloride on the relative change in P o for channels formed from wild-type -ENaC or -ENaC mutations R280G/K, W278G/A, and Y279A/F.
& x7 J5 X- `2 g4 b+ Y$ ?. h% x
Fig. 8. Mean K 0.5 values for wild-type, W278G/A, Y279A/F, and R280G/K channels formed from -ENaC subunits alone expressed in CHO cells. Solid bars, wild-type channels and the first set of mutations generated; gray bars, additional mutations generated at these positions. Error bars, SE; n 3, except for R280G, where n = 2. All mutations are different from wild-type channels ( P Table 5.

Table 5. K 0.5 of wild-type and mutant -ENaC in Fig. 8, k  g% p1 |5 R& r: g

Correlation Between P o and Amiloride Block

Because amiloride is predominantly an open channel blocker, the K 0.5 of amiloride depends not only on the intrinsic affinity of amiloride but also on the relative time that amiloride has access to the channel, i.e., the mean open time and P o. Thus the difference in the effects of amiloride between the wild-type and mutant channels might be explained by a difference in the P o of the channels, rather than a true difference in amiloride affinity: low P o would result in K 0.5 higher than that of wild-type channels, even though K Amil was unchanged. If the amiloride binding affinity was the same for all the channels and equal to the affinity of the wild-type channel, a plot of P o vs. 1/ K 0.5 would produce a straight line showing an inverse relation between P o and K 0.5. For the examination of P o in the presence of amiloride, we defined the half-fractional block of P o by Eq. 9, where K Amil is the true affinity of amiloride for the open channel (the concentration of amiloride at which open is reduced to one-half of its control value). Equation 9 predicts that if the binding affinity of a channel for amiloride ( K Amil ) remains constant, there should be an inverse relation between P o and K 0.5. That is, for low P o, K 0.5 would be larger than the corresponding K Amil. Because the mutant channels have quite variable P o values, usually less than wild-type channels, we were concerned that the observed differences in K 0.5 might be due to changes in P o, rather than changes in K Amil. The plot of P o vs. 1/ K 0.5 shows that this is not the case ( Fig. 9 ). Clearly, the values for K 0.5 for most of the mutant channels do not fit this relation; therefore, the changes in K 0.5 must be due to structural changes in the amiloride binding site, rather than merely changes in P o. Therefore, although the determination of fractional block of P o is an easy estimate of the effects of mutations on amiloride block of ENaC, it can be misleading if the P o of the mutant channel in the absence of amiloride is not similar to that of the wild-type channel. A comparison of P o values shows that the P o values of the mutant channels are clearly not similar to those of the wild-type channels ( Table 3 ). To obviate this problem, we more closely examined amiloride block by determining the blocking rate ( k b ) and the true amiloride affinity constant ( K Amil ).2 l2 U) m3 d8 Z( o

Fig. 9. Initial P o values vs. 1/ K 0.5 for amiloride for channels formed from wild-type and mutant -ENaC subunits. Dashed line, relation that would be expected if binding affinities of mutant channels were the same as those of wild-type channels and the difference in K 0.5 were due only to changes in P o. On the basis of data from single-channel analysis and determination of K 0.5 for amiloride, most of the mutants fall away from the line, except possibly Y279A and H282D. (Some errors in 1/ K 0.5 are smaller than symbol sizes and, therefore, are not visible.)' e. A" k8 t' J" p; v. ?
6 a4 w; ~, @  F' ~+ B# \
Determination of the Blocking Rate

Because amiloride is predominantly an open channel blocker, measurement of the mean open time gives a direct measure of amiloride block, because the mean open time is inversely proportional to the rate of leaving the open state. Equation 11 is used to determine the blocking rate. From Eq. 11, we are able to determine the blocking rate for amiloride in wild-type and mutant channels. Comparisons of the blocking rates are represented in Fig. 10, and the values are given in Table 6.: D5 \; ^# N1 c( K* T6 W7 x
6 X" ^9 }/ b' Q; j) w0 m6 A
Fig. 10. Blocking rates ( k b ) for amiloride for wild-type and mutant -ENaC. Rates were calculated using Eq. 3. Values are given in Table 6. Error bars, SE. *Significantly different from WT ( P3 y; c2 x1 s; i7 a( j9 K4 }9 O4 u

Table 6. Rate constants for closed, open, and blocked states of wild-type and mutant ENaC
3 U. _( C5 X( U( D  h5 F
DISCUSSION: X* m* b0 ~" E( E  V9 x
& Z' U5 s  ~: g7 e1 y2 {
Amiloride Binding Domains in -ENaC: Interaction With the Pore Region
) v, M/ J$ C9 E( X9 M
The interaction between the two functional regions of the amiloride molecule, the guanidinium group and the substituted pyrazine ring, and ENaC may occur at two different sites on the channel ( 7, 21 ). The guanidinium moiety is thought to interact with the pore by coulombically blocking entry of sodium. Previous work has shown that amiloride block is voltage sensitive and that amiloride senses 10-15% of the membrane field ( 22 ). This interaction between amiloride and the pore region (p-loop) of the channel has been demonstrated by the use of mutations in this region just before the second transmembrane (M2) domain that reduced the channel's affinity for amiloride ( 9, 16, 26, 27 )./ I5 L/ K8 }; R& P3 Z2 x
$ S: P1 I( t5 e1 X9 |7 Z/ r+ E
A Conserved Domain in the Extracellular Loop
$ n% D+ s' `  {9 i9 m
Despite the prior work with the pre-M2 region, other areas are candidates for amiloride binding domains. The region of the extracellular loop of rat -ENaC, 278-283, WYRFHY, is highly conserved across many species and tissues. In addition to the sequence WY(R/S)FHY in -ENaC, there is also a similar sequence, W(Y/C)HFHY, located in -ENaC, a functional homolog of the -subunit ( Table 7 ). Besides the conserved sequence in the - and -subunits, the -subunit also has a highly homologous region that could contribute to amiloride binding in the heteromultimeric channel. This strong conservation argues for some role of this region of the molecule in ENaC function, and there is support for an amiloride binding site outside the pore. Data in support of a site before M2 comes from the work of Li et al. ( 19 ). They demonstrated that an alternatively spliced form of the -ENaC subunit, truncated before the M2 region, was still capable of binding amiloride and has a dissociation constant similar to the wild-type -ENaC ( 19 ). Because of the possible interference from a similar sequence found in the -subunit (which could also contribute to amiloride binding), we examined the effect of mutations in the -subunit in channels composed only of -subunits.
, p2 J  `1 G3 B, K9 Y; Q& k2 F9 t
Table 7. Sequence alignment of WYRFHY amino acids in various species+ u% {) `5 c+ m4 I7 j* X
! W$ Z; N4 S% c) d/ S' B$ _
Previous Studies of the 278-283 Region in -ENaC/ W% b% n/ g; @/ r; W
1 j1 ]- g6 C8 ]: h; J6 C" W& N0 v
The purpose of this study was to examine a region of rat -ENaC, residues 278-283, WYRFHY, to determine the role of these amino acids in amiloride binding to the channel formed from this subunit. Previous work showed that deletion of this region as a whole ( 278-283), as well as mutations of individual residues, R280G, H282D, and H282R, alters amiloride binding to the channel ( 12 ). This work was done by expression of ENaC subunits in Xenopus oocytes, fusing these oocyte membrane vesicles with planar lipid bilayers, and examining the resulting channels formed in the bilayers. In addition to being expressed alone, wild-type and 278-283 -ENaC subunits were expressed in oocytes, along with wild-type - and -ENaC subunits, and membrane vesicles from this expression were also fused into bilayers. There was a substantial decrease in amiloride sensitivity of resultant channels from the 278-283 mutation compared with the wild-type channel. Three additional mutations besides 278-283 were characterized in planar lipid bilayers and produced some alterations in amiloride sensitivity. However, the approach of reconstituting oocyte membranes expressing ENaC into lipid bilayers might not reflect amiloride binding to ENaC in native cells or cellular expression systems. Therefore, we wanted to provide an alternative mammalian expression system in which to characterize this region of the -subunit and understand the role of the amino acids in this region of the extracellular loop.

Expression of Channels Composed of -Subunits Alone7 C; u$ _# o6 p

The primary focus of this study is to understand the role of the -ENaC subunit alone in amiloride binding to the channel. Originally, ENaC was characterized as a heteromultimeric protein composed of three subunits, -, -, and. The -subunit (unlike the - or -subunit), as well as the functionally homologous -subunit, is required for ionophonic activity ( 5, 30 ). In addition, the (or )-subunit can form a channel when expressed alone ( 3, 11, 13, 17, 30 ), and it has been suggested that the nonselective channels expressed in alveolar type II epithelial cells consist of -subunits only ( 14 ). In several preparations, -subunits are necessary for channel formation in vivo and in cells in culture ( 5, 14, 18 ), and the frequency of the presumptive -subunit-only channels can be reduced by antisense oligonucleotides targeting -ENaC that significantly reduce -ENaC protein.; n! L' m% j  I% S+ i1 C/ {5 y
$ X* }2 B; p1 R) B# Y, i. [5 V# m
Another reason to study channels composed of the -ENaC subunit alone is the possibility that our understanding would be confounded by contributions from the - and -subunits to amiloride binding. Differences in amiloride affinities have been observed between channels formed with wild-type -subunits and channels formed with wild-type -subunits. Channels formed with - and -ENaC subunits have a 20-fold higher affinity for amiloride than -channels, but these channels had similar K i values for guanidinium block ( 21 ). Channel block by amiloride is thought to be due to the guanidinium portion of the amiloride molecule interacting with the pore of the channel and the pyrazine ring being stabilized by a different region outside the pore. Because the K i values for guanidinium were equivalent for - and -channels, McNicholas and Canessa ( 21 ) concluded that the differences in amiloride inhibition between these two types of channels must be due to a second site of interaction between the pyrazine ring of amiloride and the channel and that the contribution to this interaction differs between the - and -subunits.% u) O* Z* }0 E! G' h

Although these data give some insight into identifying an amiloride binding region, the contribution of the -ENaC subunit alone remains unclear. Because - and -ENaC subunits contribute to the overall amiloride sensitivity of the channel, study of ENaC in its simplest form, a channel composed only of -subunits, provides useful information about the role of this individual subunit in amiloride binding.3 i% X0 c0 I2 ~  `( F9 b  B3 M/ [

Differences in Open Probability and Mean Open Times

The initial characterization of wild-type -ENaC expressed in CHO cells demonstrated that channels with properties similar to those seen in other expression systems and endogenous tissues could be expressed ( 6, 14 ). Another interesting observation was that the P o of the channels fell into two distinct groups. The wild-type -ENaC fell into a high- P o ( 0.75) or a low- P o ( 0.25) group. These groups were statistically different from each other in P o as well as mean open and closed times. Although the mean P o for all the wild-type channels was 0.42, there was no individual channel that had a P o near the mean value (because of the distribution into 2 different groups). The cause for this difference in opening and closing modes is not known. On the other hand, wild-type ENaC with two different kinetic populations have also been observed in cortical collecting tubule and A6 cells, a Xenopus kidney cell line ( 20, 23, 24 ) for channels composed of all three subunits. In one study, Palmer and Frindt ( 24 ) point out that the changes were not due to mechanical manipulations of the membrane. The modes of the P o did appear to be somewhat sensitive to changes in membrane voltages, with hyperpolarization favoring higher P o and depolarization favoring lower P o ( 24 ). A more detailed study of wild-type -ENaC in CHO cells may provide useful information about the bimodal distribution of P o in wild-type ENaC but was not pursued here.
1 {$ a+ M( z, m. k5 |: ?+ Y
Properties of Mutant -Channels

After expression of wild-type and mutant -subunitonly channels in CHO cells, we were able to compare the components of the single-channel current to determine whether any changes occurred as a result of mutations of amino acid residues 278-283 in the extracellular loop. Examination of the records ( Fig. 2 ) made it clear that there were differences in gating between the mutant and the wild-type channels. However, the unitary current and the unit conductance determined from current-voltage relations did not change between wild-type and mutant channels. In addition, the mean number of channels per patch ( 1-2) did not change. Yet there were differences in the P o between the wild-type and the mutant channels. The differences in P o between wild-type and mutant channels were primarily due to differences in the mean open times. The mean open times were smaller for all the mutant channels than for the wild-type high- P o channels (0.495 ± 0.215 s) and were as brief as 0.002 ± 0.0004 s at the limit of resolution of our recording system ( Table 3 ). The closed times for 278-283, W278G, Y279A, and R280G were longer than those for wild-type high- P o channels. These findings reinforced the idea that we needed to determine the kinetic properties of the mutant channels before we could reasonably examine the effect of amiloride on the kinetics of the mutant channels. Without a characterization of the kinetics of untreated mutant channels at the single-channel level, it would be very difficult to interpret changes in the amiloride block of mutant -channels.5 L" [  j5 E. _* G

Effects of Deletion and Point Mutations on Amiloride Block of ENaC$ F$ i3 ?" U% D) Y
" l9 }& |1 B: K9 _' m7 Q
Deletion of the region 278-283 resulted in a loss of amiloride sensitivity ( K 0.5 20 µM). These data were similar to the data obtained when 278-283 mutant channels were expressed in oocytes ( 12 ). These authors also expressed 278-283 -subunits, along with wild-type - and -subunits, in oocytes and still saw a reduction in amiloride sensitivity of the channels. Taken together, these data suggest the importance of this region in amiloride binding to -subunit-only and, possibly, -subunit channels.

The point mutations appear to fall into two categories: those more sensitive to amiloride (W278G, R280G, and H282D) and those less sensitive to amiloride (Y279A, F281A, Y283A, and H2828R). The data suggest that the histidine, arginine, and tryptophan residues interact with amiloride on the basis of the differences observed with the H282D/R, R280G/K, and W278G/A mutations and the level of disruption to amiloride binding caused by channels made from -subunits with these mutations. Mutations generated at positions Y279, F281, and Y283 did not change the amiloride block of the channel as dramatically. Although these mutations caused changes in P o in the presence of amiloride compared with wild-type channels, the blocking was also affected by these mutations. The deletion of this region, 278-283, and point mutations, W278G, R280G, F281A, and H282D, caused a change in the amiloride-blocking rate of the channel.
0 [; h5 }1 f) S+ Q: w
On the basis of our data, we predict that the secondary structure in this region may be a -pleated sheet ( Fig. 11 ). The amino acids tryptophan, arginine, and histidine would be on one side of the sheet, whereas tyrosine, phenylalanine, and tyrosine would be on the other side. Information from Laser Gene and secondary structural content prediction ( 8 ) predict that this region is a -pleated sheet, although some other programs suggest that it is an -helix or a random coil region. However, if it is a -pleated sheet, then the changes in amiloride sensitivity of the mutant channels can be explained by the positions of the residues within the sheet. The three residues, Y, F, and Y, would face away from the binding site for amiloride and possibly be involved in orbital interactions through their side chains for stabilization of the overall structure to form an amiloride binding pocket. The disruption of these side chain interactions was evident when alanines were introduced at these positions. The differences in the effects of the alanine substitutions of the tyrosines, phenylalanine, and tryptophan suggest that the tryptophan does not face the same binding environment as do the tyrosines and phenylalanine. The other residues, W, R, and H, would participate in providing the energy for direct interaction with amiloride.

Fig. 11. Schematic of WYRFHY as the residues would exist in a -pleated sheet.
0 d! `6 b  N# D6 _; G
One additional interesting point is that the mutations to this region not only affected the amiloride block of the channel but also affected the gating kinetics of the channels formed. These data suggest that we must rethink where this region of the extracellular loop of the channel is with respect to the pore and how it may play a role in the overall structure and function of the channel. Because there were changes in the gating properties of the mutant channels compared with the wild-type channels, it is possible that this region of the extracellular domain is also closer than the rest of the loop to the gating mechanism of the channel. This could explain the interaction of this region of the molecule with amiloride and the effects on gating. The effects of mutations in this region on gating of -ENaC have also been demonstrated by Sheng et al. ( 28 ). On the basis of studies involving nickel binding and a histidine residue in the extracellular loop, H282, the authors suggest that residues close to H282 in the extracellular loop were in close proximity to the channel pore.( [. W3 I  B1 X  m9 h- z- U# R
! j- M, J9 a' G/ v: @/ ]
In summary, this study has demonstrated the significance of this region, WYRFHY, in the extracellular loop of the -subunit of ENaC and its role in amiloride binding to the channel, at least when -subunits are expressed alone. On the basis of the data from the deletion mutation as well as the point mutations, we are able to gain information about the significance of this region in the overall structure of the channel and its role in amiloride binding and gating. We have demonstrated that these mutations, made to a region other than the pre-M2 region, which has previously been identified as an amiloride binding or interacting region, affect amiloride binding and, therefore, can be considered a different amiloride binding region. These data may be useful in elucidating the role of the -subunit in ENaC that contain all three subunits,,, and.
, X9 J% `8 n+ r. ~
DISCLOSURES1 t8 N" L3 G2 H+ S2 Y8 T/ z
/ ~7 M* |# v& V  ]; R  ?1 _% C
This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R37 DK-37963 (to D. C. Eaton) and R01 DK-054354 (to T. R. Kleyman).
      【参考文献】
Benos DJ, Awayda MS, Ismailov II, and Johnson JP. Structure and function of amiloride-sensitive Na channels. J Membr Biol 143: 1-18, 1995.: u8 g9 ]3 V, ?- J' C2 x6 }. J
6 @8 D/ @  x+ ^# V% @+ |1 U$ A

- }4 ?: }! g& ~2 P
Benos DJ and Stanton BA. Functional domains within the degenerin/epithelial sodium channel (Deg/ENaC) superfamily of ion channels. J Physiol 520: 631-644, 1999.

) h5 G% t3 z$ E# h$ a! j/ b

Canessa CM, Horisberger JD, and Rossier BC. Epithelial sodium channel related to proteins involved in neurodegeneration. Nature 361: 467-470, 1993.

0 K% S) C" p  k4 d
. ~/ I. _" c8 ~+ W3 g
Canessa CM, Merillat AM, and Rossier BC. Membrane topology of the epithelial sodium channel in intact cells. Am J Physiol Cell Physiol 267: C1682-C1690, 1994.
( H+ }: y, m) s, ]1 s
$ a0 Q8 ^, a3 P# o" ^/ ]1 T

Canessa CM, Schild L, Buell G, Thorens B, Horisberger JD, and Rossier BC. Amiloride-sensitive epithelial Na channel is made of three homologous subunits. Nature 367: 463-467, 1994.
* D# `8 D! S8 v% G2 M
) Y" I  K. l5 s9 T; g' }9 m4 o' k
. j" s& P" q2 U! n
Ciampolillo F, McCoy DE, Green RB, Karlson KH, Dagenais A, Molday RS, and Stanton BA. Cell-specific expression of amiloride-sensitive, Na -conducting ion channels in the kidney. Am J Physiol Cell Physiol 271: C1303-C1315, 1996.

# E+ F4 ~$ N: Q/ k9 ~- p- L

Eaton DC and Hamilton KL. The amiloride-blockable sodium channel of epithelial tissue. In: Ion Channels, edited by Narahasi T. New York: Plenum, 1988, vol. 1, p. 251-282.
1 I8 r. t# R& n& o
" i2 R  s: G4 d+ q$ L
. E8 ^: W$ x* C: R# W& T
Eisenhaber F, Imperiale F, Argos P, and Froemmel C. Prediction of secondary structural content of proteins from their composition alone. I. New analytic vector decomposition methods. Proteins 25: 157-168, 1996.
4 ^8 [+ {- y( j* [* Y* ]
- o' Y1 o( D. \; F8 o

Fuller CM, Berdiev BK, Shlyonsky VG, Ismailov II, and Benos DJ. Point mutations in -ENaC regulate channel gating, ion selectivity, and sensitivity to amiloride. Biophys J 72: 1622-1632, 1997.

) |# A* n7 h1 {9 R) Q% N
Ishikawa T, Marunaka Y, and Rotin D. Electrophysiological characterization of the rat epithelial Na channel (rENaC) expressed in MDCK cells. Effects of Na and Ca 2 . J Gen Physiol 111: 825-846, 1998.; C7 q6 K; X! T8 @3 R! U
# g- [' E7 W/ V

& U5 M- V! M) B8 V  s0 S
Ismailov II, Awayda MS, Berdiev BK, Bubien JK, Lucas JE, Fuller CM, and Benos DJ. Triple-barrel organization of ENaC, a cloned epithelial Na channel. J Biol Chem 271: 807-816, 1996.$ s" Q0 `. A5 t6 E* O4 E9 X( C9 U+ p
  `( _" n" U& d
% f. n& l0 L: }6 s2 l

Ismailov II, Kieber-Emmons T, Lin C, Berdiev BK, Shlyonsky VG, Patton HK, Fuller CM, Worrell R, Zuckerman JB, Sun W, Eaton DC, Benos DJ, and Kleyman TR. Identification of an amiloride binding domain within the -subunit of the epithelial Na channel. J Biol Chem 272: 21075-21083, 1997.' k( M$ E5 z( H$ P& _7 ^! x

7 K: R" E" s+ W+ L
# ?3 }' H9 s& u0 ]" J  X0 J5 K
Jain L, Chen XJ, Malik B, Al Khalili O, and Eaton DC. Antisense oligonucleotides against the -subunit of ENaC decrease lung epithelial cation-channel activity. Am J Physiol Lung Cell Mol Physiol 276: L1046-L1051, 1999.

5 C/ W0 Y. ?. s  x" s0 c- L

Jain L, Chen XJ, Ramosevac S, Brown LA, and Eaton DC. Expression of highly selective sodium channels in alveolar type II cells is determined by culture conditions. Am J Physiol Lung Cell Mol Physiol 280: L646-L658, 2001.$ e$ X+ n1 r, k  D0 r- {

Ji HL, Fuller CM, and Benos DJ. Intrinsic gating mechanisms of epithelial sodium channels. Am J Physiol Cell Physiol 283: C646-C650, 2002.
/ K1 Q$ b- b3 ]

8 V1 b8 q9 n5 t' A' q  p. B# S
Kellenberger S, Gautschi I, and Schild L. A single point mutation in the pore region of the epithelial Na channel changes ion selectivity by modifying molecular sieving. Proc Natl Acad Sci USA 96: 4170-4175, 1999.

' z. u/ V) N- O

Kizer N, Guo XL, and Hruska K. Reconstitution of stretch-activated cation channels by expression of the -subunit of the epithelial sodium channel cloned from osteoclasts. Proc Natl Acad Sci USA 94: 1013-1018, 1997.
/ c; {4 _3 ]" ~: D7 N) c& C
* A% \- f* V. l9 B& q7 Q& X+ C1 Q

Lazrak A, Samanta A, and Matalon S. Biophysical properties and molecular characterization of amiloride-sensitive sodium channels in A549 cells. Am J Physiol Lung Cell Mol Physiol 278: L848-L857, 2000.

1 E! q: m; g  R

Li XJ, Xu RH, Guggino WB, and Snyder SH. Alternatively spliced forms of the -subunit of the epithelial sodium channel: distinct sites for amiloride binding and channel pore. Mol Pharmacol 47: 1133-1140, 1995.; z" w( e* m: k
. x' [+ x8 y% Y9 @, O; O
& }" X3 F( X# C4 ]4 U" F

Ling BN and Eaton DC. Effects of luminal Na on single Na channels in A6 cells, a regulatory role for protein kinase C. Am J Physiol Renal Fluid Electrolyte Physiol 256: F1094-F1103, 1989.3 {* s/ L) ?6 A* |" C8 P# \9 `
3 i! o* i0 {  W! J8 v0 ^+ G

McNicholas CM and Canessa CM. Diversity of channels generated by different combinations of epithelial sodium channel subunits. J Gen Physiol 109: 681-692, 1997.

, c$ h3 H$ @2 B- A
Palmer LG. Voltage-dependent block by amiloride and other monovalent cations of apical Na channels in the toad urinary bladder. J Membr Biol 80: 153-165, 1984.

1 c2 [4 g# S5 N
/ V8 G3 y" q( y4 s5 {& Z! }1 L
Palmer LG and Frindt G. Conductance and gating of epithelial Na channels from rat cortical collecting tubule. Effects of luminal Na and Li. J Gen Physiol 92: 121-138, 1988.
6 {6 ^4 M$ L1 e. g, x/ Y
9 ^( B$ t6 C% r
& G6 G2 b: p) r2 Y
Palmer LG and Frindt G. Gating of Na channels in the rat cortical collecting tubule: effects of voltage and membrane stretch. J Gen Physiol 107: 35-45, 1996.
+ m, z8 O. N( O5 }5 R- X; N2 \
) j' c, P# k7 ~: |; x3 \8 W
: F' _+ i, H- P' I$ a+ l
Prince LS and Welsh MJ. Cell surface expression and biosynthesis of epithelial Na channels. Biochem J 336: 705-710, 1998.6 l# n0 a% P$ E* e) E2 J$ Q
. ~; N: i. O( y0 s! h
- B. o5 X. [) l! q  {
3 L& N, m0 f* i0 \; M% w" }
Schild L, Schneeberger E, Gautschi I, and Firsov D. Identification of amino acid residues in the,, and subunits of the epithelial sodium channel (ENaC) involved in amiloride block and ion permeation. J Gen Physiol 109: 15-26, 1997.
; F1 E5 i) F1 \+ J1 y; @, q' I# q

Sheng S, Li J, McNulty KA, Avery D, and Kleyman TR. Characterization of the selectivity filter of the epithelial sodium channel. J Biol Chem 275: 8572-8581, 2000.  Y) c% @0 Y& q- N! J

, o. ^: V* K; N5 Q

Sheng S, Perry CJ, and Kleyman TR. External nickel inhibits epithelial sodium channel by binding to histidine residues within the extracellular domains of and subunits and reducing channel open probability. J Biol Chem 277: 50098-50111, 2002.' K' [9 F' v3 Q0 l+ r7 ?- C7 Y9 s
- L- I2 l9 @0 K# _* P& |
1 l( T6 l; Q8 S. `# }

Sigworth FJ and Sine SM. Data transformations for improved display and fitting of single-channel dwell time histograms. Biophys J 52: 1047-1054, 1987.& ~$ k% F0 X- _& \

' N) ?6 \5 p& U% h: _

Waldmann R, Champigny G, Lazdunski M, and Voilley N. Molecular cloning and functional expression of a novel amiloride-sensitive Na channel. J Biol Chem 270: 27411-27414, 1995.

3 l" z6 t/ ^$ S( Z
Weber WM, Liebold KM, and Clauss W. Amiloride-sensitive Na conductance in native Xenopus oocytes. Biochim Biophys Acta 1239: 201-206, 1995.

收藏0 分享0 顶0 踩0