Determinants of Cation Permeation and Drug Sensitivity in Predicted Transmembrane Helix 9 and Adjoining Exofacial Re-entrant Loop 5 of Na+/H+ Exchanger NHE1*

Background: NHE1 is a therapeutic drug target, yet knowledge of sites involved in cation translocation and drug binding remains incomplete. Results: Mutagenesis analyses identified residues in transmembrane helix M9 and exofacial re-entrant loop EL5 that affect substrate affinities and/or drug sensitivity. Conclusion: M9 and EL5 form part of the cation permeation pathway. Significance: These findings provide new insight into the structure-function domains of NHE1. Mammalian Na+/H+ exchangers (NHEs) regulate numerous physiological processes and are involved in the pathogenesis of several diseases, including tissue ischemia and reperfusion injuries, cardiac hypertrophy and failure, and cancer progression. Hence, NHEs are being targeted for pharmaceutical-based clinical therapies, but pertinent information regarding the structural elements involved in cation translocation and drug binding remains incomplete. Molecular manipulations of the prototypical NHE1 isoform have implicated several predicted membrane-spanning (M) helices, most notably M4, M9, and M11, as important determinants of cation permeation and drug sensitivity. Here, we have used substituted-cysteine accessibility mutagenesis and thiol-modifying methanethiosulfonate (MTS) reagents to further probe the involvement of evolutionarily conserved sites within M9 (residues 342–363) and the adjacent exofacial re-entrant loop 5 between M9 and M10 (EL5; residues 364–415) of a cysteine-less variant of rat NHE1 on its kinetic and pharmacological properties. MTS treatment significantly reduced the activity of mutants containing substitutions within M9 (H353C, S355C, and G356C) and EL5 (G403C and S405C). In the absence of MTS, mutants S355C, G403C, and S405C showed modest to significant decreases in their apparent affinities for Na+o and/or H+i. In addition, mutations Y370C and E395C within EL5, whereas failing to confer sensitivity to MTS, nevertheless, reduced the affinity for Na+o, but not for H+i. The Y370C mutant also exhibited higher affinity for ethylisopropylamiloride, a competitive antagonist of Na+o transport. Collectively, these results further implicate helix M9 and EL5 of NHE1 as important elements involved in cation transport and inhibitor sensitivity, which may inform rational drug design.

reduced the activity of mutants containing substitutions within M9 (H353C, S355C, and G356C) and EL5 (G403C and S405C). In the absence of MTS, mutants S355C, G403C, and S405C showed modest to significant decreases in their apparent affinities for Na ؉ o and/or H ؉ i . In addition, mutations Y370C and E395C within EL5, whereas failing to confer sensitivity to MTS, nevertheless, reduced the affinity for Na ؉ o , but not for H ؉ i . The Y370C mutant also exhibited higher affinity for ethylisopropylamiloride, a competitive antagonist of Na ؉ o transport. Collectively, these results further implicate helix M9 and EL5 of NHE1 as important elements involved in cation transport and inhibitor sensitivity, which may inform rational drug design.
Electroneutral countertransport of alkali cations such as Na ϩ , K ϩ , and Li ϩ for H ϩ across membranes of mammalian cells are catalyzed by a heterogeneous family of at least 11 secondary active solute carriers (SLC9 gene family) generically termed Na ϩ /H ϩ exchangers or antiporters (NHE/NHA) 3 (1,2). The activities of these transporters are tightly controlled and important for efficient execution of numerous physiological processes, ranging from cellular and systemic pH and volume homeostasis (3)(4)(5) to the regulation of cell shape (6), migration (7)(8)(9)(10), and mitosis (11), among others (12)(13)(14)(15)(16)(17). On the other hand, aberrant overactivation of NHE activity, notably the ubiquitous NHE1 isoform, occurs in several pathophysiological conditions and contributes to disease progression, including tissue injuries following ischemic and hemorrhagic stroke (18 -20), acute myocardial infarction (21)(22)(23), cardiac hypertrophy and failure (24 -27), and cancer metastasis and invasion (28 -31); damages that can be mitigated in animal models by pharmacological inhibition of NHE1 activity. However, in the case of heart disease, attempts to translate some of these promising experimental findings into clinical therapies have thus far proven inconclusive due to modest efficacy and adverse side effects of the tested compounds (32)(33)(34)(35). Despite these setbacks, an improved understanding of the molecular determinants that underlie the catalytic and pharmacological properties of NHE1 could assist in the development of more efficacious drugs and treatment regimens.
Information regarding the structural and functional properties of the mammalian NHEs has been derived mainly from analyses of the ubiquitous NHE1 and epithelial NHE3 isoforms. Early computational modeling and substituted-cysteine accessibility mutagenesis studies of human NHE1 (hNHE1) by Wakabayashi et al. (36) predicted a configuration of 12 mem-brane-spanning (M) helices at its N terminus (ϳ400 -500 amino acids) involved in cation translocation and drug binding, and a cytoplasmically oriented segment at its C terminus (ϳ350 amino acids) that confers responsiveness to various regulatory stimuli (illustrated in Fig. 1A). Biochemical (37)(38)(39)(40)(41)(42)(43)(44) analyses indicated that the exchanger most likely assembles as a homodimer, although higher ordered structures have also been proposed (45).
To identify sites involved in Na ϩ binding and translocation, initial studies (46 -48) took advantage of pyrazine-(e.g. amiloride and ethylisopropylamiloride (EIPA)) or benzyolguanidinium-based (e.g. HOE642 and HOE694) compounds that inhibit NHE activity by acting as competitive or mixedtype antagonists of Na ϩ influx, suggesting that they bind at or near sites involved in Na ϩ binding. Random or site-directed mutagenesis of NHE1 followed by functional selection for altered drug sensitivity identified several residues located in the predicted second exofacial loop (EL2) (49) and two membranespanning segments, M4 (50 -54) and M9 (49,55,56), which are significant determinants of drug recognition and/or transport velocity. However, only mutations at Phe 162 in M4 of hNHE1 were found to appreciably reduce Na ϩ affinity (52). Additional mutations of conserved residues within the fifth intracellular loop (IL5) (i.e. Arg 440 ) and adjacent transmembrane helix M11 (Gly 445 , Gly 446 ) of hNHE1 decreased and increased, respectively, its sensitivity to intracellular H ϩ (57,58). Other mutagenesis analyses have confirmed the importance of M4 (59), M9 (60), and M11 (61), and further implicated residues in M6 (62) and M7 (63) as pore-lining elements.
Although there is general support for this model, the precise arrangement of transmembrane helices remains controversial (64 -66). High-resolution crystal structures of the bacterial Escherichia coli Na ϩ /H ϩ antiporter NhaA (67,68) that is structurally (ϳ10% amino acid identity) and kinetically (i.e. electrogenic) distinct from NHE1 also indicated a 12-transmembrane structure that assembles as a dimer. Such parallels prompted Landau et al. (65) and Nygaard et al. (69) to use the atomic coordinates of NhaA as a template to generate a three-dimensional model of hNHE1. However, whereas the model proposed by Nygaard et al. (69) closely mirrors that of Wakabayashi et al. (36), Landau and colleagues (65) derived a novel membrane topology that challenges this view (Fig. 1B). Notably, in their new model, the original transmembrane helices M1 and M2 were repositioned intracellularly and the catalytically important M9 helix was reconfigured to form two short intramembrane helices relabeled as M7 and M8, whereas the adjoining exofacial loop 5 (EL5, amino acids 364 -415) between M9 and M10 formed a new intracellular loop and a new transmembrane segment M9 that is largely inaccessible to external reagents. Although the computationally derived reorientation of M1 and M2 seems unlikely given that the intervening EL1 segment was shown previously to undergo N-and O-linked glycosylation and therefore places this segment extracellularly (70), the topology of the M9-EL5 region remains uncertain. This region is of particular interest because the putative EL5 loop was originally postulated to invaginate into the membrane (36) in a manner analogous to the pore-lining loops present in ion chan-nels (71,72), and therefore may contribute to cation permeation and drug sensitivity.
Based on discordances between these models, we have further probed the potential contributions of evolutionarily conserved as well as certain known drug-sensitive sites located in the originally designated M9 and adjacent re-entrant loop EL5 segments (as depicted in Fig. 1A) to cation translocation using the substituted-cysteine accessibility method (73). The data revealed additional amino acids in both M9 and EL5 that contribute to the affinities of the exchanger for substrates and pharmacological antagonists and hence are likely to line the cation permeation pathway. The relevance of these findings to the two different membrane topologies proposed for NHE1 is discussed.

Experimental Procedures
Materials-Carrier free 22 NaCl (radioactivity, 5 mCi/ml) was obtained from PerkinElmer Life Sciences. Amiloride, ouabain, and nigericin were purchased from Sigma. ␣-Minimal essential medium, fetal bovine serum, G418, and trypsin-EDTA were purchased from Invitrogen. Murine monoclonal HA antibody was purchased from BabCo (Richmond, CA), whereas horseradish peroxidase-conjugated goat ␣-mouse IgG was obtained from Jackson ImmunoResearch Laboratories Inc. (West Grove, PA). 2-Sulfonatoethyl methanethiosulfonate (MTSES) and 2-(trimethylammonium)ethyl methanethiosulfonate (MTSET) were purchased from Toronto Research Chemicals (North York, ON). All other chemicals and reagents used in these experiments were purchased from Fisher Scientific and were of the highest grade available.
Construction of Na ϩ /H ϩ Exchanger Mutants-The rat cDNA, engineered to contain a series of unique restriction endonuclease sites that provide convenient DNA cassettes for mutagenesis, was subcloned into a mammalian expression vector under the control of the enhancer/promoter region from the immediate early gene of human cytomegalovirus (pCMV), as previously described (56). A single copy of an influenza virus hemagglutinin (HA) peptide (YPYDVPDYA) was inserted at the C terminus of NHE1 using polymerase chain reaction (PCR) mutagenesis (the construct is referred to herein as NHE1 HA ). NHE1 HA ⌬C was constructed by substituting the eight endogenous cysteine residues (amino acids 117, 137, 216, 425, 481, 542, 565, and 799) with the structurally conservative residue serine. The substitution of serine did not generate any obvious consensus motifs for phosphorylation by known serine/threonine kinases. Single cysteine residues were introduced into this cysteine-less background using the QuikChange TM Site-directed Mutagenesis system purchased from Stratagene (La Jolla, CA). All mutant constructs were sequenced to verify their fidelity.
Stable Transfection and Expression of the Na ϩ /H ϩ Exchanger cDNAs-Chinese hamster ovary cells (AP-1 cells), a chemically mutagenized cell-line devoid of endogenous plasmalemmal NHE activity (74), were transfected with mammalian expression plasmids containing either wild-type NHE1 HA or NHE1 HA ⌬C-based constructs using Lipofectamine TM . Cells were maintained in complete ␣-minimal essential medium supplemented with 10% fetal bovine serum, 100 units/ml of penicillin, 100 g/ml of streptomycin, and 25 mM NaHCO 3 , pH 7.4, and incubated in an humidified atmosphere of 95% air, 5% CO 2 at 37°C. Starting 48 h after transfection, the AP-1 cells were selected for survival in response to repeated (ϳ6 times over a 2-week period) acute NH 4 Cl-induced acid loads (i.e. H ϩ -killing technique) (75,76) to discriminate between NHE1 positive and negative transfectants. Six clonal isolates per mutant were routinely selected and the one exhibiting the highest amiloridesensitive H ϩ i -activated 22 Na ϩ o uptake was subjected to further analyses.
Measurement of Na ϩ /H ϩ Exchanger Activity and Covalent Modification with Sulfhydryl-reactive Reagents-Clonal cells were grown to confluence in 24-well plates and then acidified (to ϳpH i 6.0) using the NH 4 Cl pre-pulse technique (75,76). Briefly, the cell culture medium was aspirated and replaced by isotonic NH 4 Cl medium (50 mM NH 4 Cl, 70 mM choline chloride, 5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 5 mM glucose, 20 mM HEPES-Tris, pH 7.4). Cells were incubated in this medium for 30 min at 37°C in a nominally CO 2 -free atmosphere. After acid loading, the monolayers were rapidly washed twice with isotonic choline chloride solution (125 mM choline chloride, 1 mM MgCl 2 , 2 mM CaCl 2 , 5 mM glucose, 20 mM HEPES-Tris, pH 7.4). 22 Na ϩ influx assays were initiated by incubating the cells in isotonic choline chloride solution containing 1 Ci of 22 NaCl/ml (ϳ120 nM 22 NaCl, carrier-free). The assay medium was K ϩ -free to prevent 22 Na ϩ influx and efflux by the Na ϩ -K ϩ -Cl Ϫ cotransporter and the Na ϩ /K ϩ -ATPase, respectively. Influx of 22 Na ϩ was terminated by rapidly washing the cell monolayers three times with 4 volumes of ice-cold isotonic saline solution (130 mM NaCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 20 mM HEPES-Tris, pH 7.4). The washed cell monolayers were solubilized in 0.25 ml of 0.5 N NaOH and the wells were washed with 0.25 ml of 0.5 N HCl. Both the solubilized cell extract and wash solutions were added to vials and radioactivity was assayed using a liquid scintillation counter. Under the H ϩ -loading conditions used in this study, uptake of 22 Na ϩ was linear with time for 8 to 10 min (at low Na ϩ concentrations, 22°C). Therefore, 22 Na ϩ uptakes were measured after 5 min except when examining the kinetics of NHE activity as a function of the extracellular Na ϩ (Na ϩ o ) concentration. When assessing the K m for [Na ϩ o ], a modified version of the aforementioned protocol was applied. Previous measurements indicated that when [Na ϩ o ] is increased to 100 -125 mM, 22 Na ϩ uptake is linear for several minutes. Therefore, when conducting this kinetic analysis in which the [Na ϩ o ] concentration ranges from 1.25 to 120 mM, 22 Na ϩ uptake was terminated after 1 min. Measurements of 22 Na ϩ influx specific to the Na ϩ /H ϩ exchanger were determined as the difference between the initial rates of H ϩ -activated 22 Na ϩ influx in the absence and presence of either 2 mM amiloride or 100 M EIPA (concentrations sufficient to inhibit NHE1 under these experimental conditions) and expressed as amiloride-or EIPA-inhibitable 22 Na ϩ influx. To make quantitative comparisons of the intrinsic rates of transport of the various NHE constructs in stably transfected cells, the cellular rates of drug-sensitive H ϩ -activated 22 Na ϩ influx were measured under near maximal acid-load conditions and normalized per mg of total cellular protein per abundance of the fully glycosylated NHE1 protein present at the plasma membrane (assessed by Western blotting and densitometry), as described previously (49).
To examine NHE activity as a function of intracellular H ϩ concentration, pH i was set over the range of 5.4 -7.4 using the K ϩ -nigericin method as previously described (77,78). Briefly, the confluent monolayers were washed with isotonic N-methyl-D-glucamine (NMG)-chloride solution (140 mM NMG-Cl, 1 mM MgCl 2 , 2 mM CaCl 2 , 5 mM glucose, 10 mM HEPES-Tris, pH 7.4), and then incubated for 4 min at room temperature in NMG-balanced salt solutions specific for each pH i (pH i -clamp solutions). All solutions contained 2 mM NaCl, 1 mM MgCl 2 , and 10 mM HEPES-Tris, pH 7.4, varying concentrations of K ϩ (1.4 to 140 mM KCl/K ϩ -glutamate, adjusted as needed with NMG-glutamate to bring the final concentration to 140 mM), and the K ϩ /H ϩ exchange ionophore, nigericin (10 M). These pH i clamp solutions are designed to fix the pH i at a desired level by adjusting the extracellular K ϩ concentration as described (78). In essence, the desired pH i can be established according to the following equation: , assuming that the intracellular K ϩ concentration is 140 mM and the extracellular H ϩ concentration is set at 7.4. 22 Na ϩ uptake was initiated in the same pH i clamp solutions supplemented with 1 Ci/ml of 22 Na ϩ and 1 mM ouabain in the absence or presence of 2 mM amiloride. Uptake occurred for a period of 10 min and was terminated in the same fashion as described above. Protein content was assessed using the Bio-Rad DC protein assay kit as per the manufacturer's protocol. Rates were expressed as nanomole of Na ϩ /min/mg of protein.
To evaluate the effect of the methanethiosulfonate derivatives (73) on NHE1 activity, cell monolayers were incubated in phosphate-buffered saline containing 1 mM MgCl 2 and 0.1 mM CaCl 2 , pH 7.2, with or without 1 mM MTSET or 10 mM MTSES for 30 min at 37°C. Cells were washed twice with isotonic choline chloride solution and then intracellular pH was clamped at pH i 5.6 using the K ϩ -nigericin method. 22 Na ϩ uptake was initiated using the same clamp solution supplemented with 1 Ci of 22 Na ϩ and 1 mM ouabain in the absence or presence of 2 mM amiloride or 100 M EIPA. 22 Na ϩ uptake was terminated after 10 min using ice-cold isotonic saline solution in the same fashion as mentioned above. Residual NHE1 activity was determined as the ratio of amiloride or EIPA-inhibitable 22 Na ϩ uptake in the presence and absence of MTSET or MTSES.
Immunoblotting-Stably transfected cells were grown to confluence in 10-cm plates and were lysed using 1% Triton X-100. Total cellular protein extracts (30 g) were resolved by 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto PVDF Hybond-P TM membranes (Amersham Biosciences). The blots were rinsed briefly with PBST (1 ϫ PBS containing 0.1% Tween 20), blocked with 5% nonfat skim milk in PBST, and then incubated with a murine monoclonal anti-HA antibody (dilution 1:5,000). Following extensive washing with PBST, blots were incubated with goat anti-murine IgG secondary antibody conjugated with horseradish peroxidase (dilution 1:5,000). Immunoreactive bands were visualized using enhanced chemiluminescence (PerkinElmer Life Sciences) and recorded on x-ray film.
Statistics-The data were analyzed using Origin and Microsoft Excel. The data from the sodium affinity, drug resistance, and pH profile experiments was fitted with a Hill function, whereas the EIPA and pH profile data were fitted with a doseresponse function. Unless otherwise stated, error bars represent the mean Ϯ S.E. and statistical analysis was performed by using the one-way analysis of variance Tukey post hoc test, with a significance level of 0.05.

Results
Characterization of NHE1 HA ⌬C and Single-substituted Cysteine Mutants-To further explore the structure-function domains of NHE1 in the M9-EL5 region (the nomenclature used will follow the Wakabayashi model illustrated in Fig. 1A), we used the substituted-cysteine accessibility method (73). This method relies on covalent interaction between the ionized sulfhydryl group of cysteines located in a water-accessible environment, such as the pore, with membrane-impermeant sulfhydryl-reactive reagents that have the potential to irreversibly impair transmembrane ion fluxes by either steric blockage or charge attraction/repulsion. Such covalent modifications might also hinder critical conformation changes that occur during cation translocation either at the cation pore itself or at a more distant site that nevertheless impacts cation translocation.
To this end, a cysteine-less HA epitope-tagged version of rat NHE1 (NHE1⌬C HA ) was constructed by replacing the eight native cysteines in NHE1 (indicated in Fig. 2A) with serine, a conservative substitution that maintains amino acid polarity and side chain length and hence minimizes potential structural perturbations of the transporter. The NHE1⌬C HA was stably expressed in mutagenized Chinese hamster ovary AP-1 cells devoid of endogenous NHE1 as previously described (75). Levels of protein expression for NHE1 HA wild-type (WT) and ⌬C are displayed in the inset of Fig. 2B. Previous studies demonstrated that the slower migrating band at ϳ100 kDa represents the fully glycosylated form of the protein present at the cell surface, whereas the faster migrating band at ϳ75 kDa represents the partially processed or core-glycosylated form of the protein that is largely retained in the endoplasmic reticulum (79). Importantly, substitution of the cysteines residues with serine did not impair the biosynthesis and maturation of the transporter.
We next compared the kinetic properties of the WT and ⌬C transporters. The initial rates of NHE1 activity as a function of the Na ϩ o concentration were examined using solutions containing tracer amounts of the radioisotope 22 Na ϩ added to Na ϩ o concentrations ranging from 2.5 to 120 mM following an imposed NH 4 Cl-induced intracellular acidification. The data for both WT and ⌬C were best fit to a Hill equation and revealed an apparent sigmoidal dependence on the Na ϩ concentration for both constructs, indicative of positive cooperative binding with comparable Hill coefficients of ϳ1.5 and 1.7, respectively ( Table 1). Replacement of the cysteine residues also did not alter its maximal velocity (V max , nmol/min/mg of protein: WT, 105.5 Ϯ 1.0; ⌬C, 105.0 Ϯ 0.7), but did cause a reduction (ϳ36%) in its affinity for Na ϩ o (K Na : WT, 29.5 Ϯ 0.1; ⌬C, 46.1 Ϯ 0.9; p Ͻ 0.05) ( Fig. 2B; Table 1). To determine their affinities for intracellular H ϩ , the H ϩ i concentration was adjusted by clamping pH i at specific levels within the range of 5.4 -7.4 using the K ϩ /H ϩ ionophore nigericin, and measuring the rate of 22 Na ϩ influx. As shown in Fig. 2C, the WT and ⌬C transporters exhibited similar affinities for H ϩ i (K H pH: WT, 6.14 Ϯ 0.05; ⌬C, 6.26 Ϯ 0.03; p Ͼ 0.05). On the other hand, ⌬C showed an approximate 8-fold decrease in sensitivity to the pharmacological antagonist EIPA (IC 50 , nM: WT, 5.4 Ϯ 1.2; ⌬C, 40.7 Ϯ 10.4; p Ͻ 0.05) (Fig. 2D). Based on these data, the endogenous cysteines moderately influenced, either directly or indirectly, the Na ϩ and drug affinities of the transporter, but they were not structurally essential. Hence, NHE1⌬C HA could serve as a suitable template for further cysteine-substitution mutagenesis and functional analyses.
To rationalize potential targets for cysteine substitution, certain sites in M9 (Glu 350 , His 353 , and Gly 356 ) (49,55,56) were selected based on their known involvement in conferring sensitivity to amiloride-or benzyolguanidinium-based compounds that act as either simple or mixed competitors of Na ϩ o binding and therefore are likely in close proximity to the permeation pathway. Additional amino acids in M9 and EL5 were chosen based on their conservation among the plasmalemmal NHE isoforms (NHE1-NHE5) (Figs. 3A and 4A), their relative predicted position in a transmembrane ␣-helix (Fig. 3B), and their hydrophilic nature, which likely orients them in an aqueous environment and potentially accessible to the thiol-modifying reagents. In total, 20 single cysteine substitutions were generated in M9 (S342C, Y343C, Y346C, S348C, E350C, H353C, S355C, G356C, and A359C) and EL5 (Y370C, N374C, S376C, T381C, Y385C, S392C, E395C, L397C, G403C, S405C, and W415C). The resulting constructs were then stably expressed in AP-1 cells based on their ability to confer cell survival followed repeated intracellular acid loads that are lethal to AP-1 cells lacking a functional transporter. To make initial estimates of the relative activities of parental and single cysteine-substituted mutants of NHE1 HA ⌬C in stably transfected cells, the cellular rates of amiloride-sensitive H ϩ i -activated 22 Na ϩ o influx were measured in the presence of nominal non-radioactive Na ϩ o (ϳ120 nM 22 NaCl, carrier-free; 5 min uptake) and then normalized for the level of fully glycosylated NHE1 protein at the cell surface as determined by immunoblotting and densitometry. Levels of protein expression for NHE1 HA ⌬C and the single cysteine-substituted mutants are displayed in Figs. 3C and 4B. Within M9, the relative activity levels of the single cysteine-substituted mutants were generally higher (i.e. ϳ2-fold) than the parental NHE1 HA ⌬C (Fig. 3C), with the notable exception of S348C which, whereas not highly expressed, exhibited an apparent dramatic 24-fold increase in its rate of transport. However, because of its low level of expression, the calculated rate of activity per unit NHE1 protein is more subject to error and hence the determined value may overestimate its actual rate of transport. By comparison, the majority of the mutants in EL5 exhibited lower rates of transport with the exception of L397C, which displayed a 2.5-fold increase, although its expression level was also quite low, which made an accurate assessment of its relative rate of transport more difficult (Fig. 4B). These initial findings implicate both regions as important elements in cation translocation.
MTS Inhibition of the M9 and EL5 Single Cysteine Mutants-To further probe the involvement of these regions in cation permeation, the cysteine-substituted transporters were subjected to chemical modification using the membrane-impermeant sulfhydryl-active reagents MTSET and MTSES that are positively and negatively charged, respectively, and assessed for their effects on NHE1 activity. The activities of the majority of the M9 mutants were unaffected by the MTS derivatives, suggesting that the sites are either inaccessible to these reagents or alternatively that they did react with the compounds but that this did not lead to inhibition (Fig. 3D). However, three mutants containing neighboring substitutions, H353C, S355C, and G356C showed ϳ40, 30, and 98% inhibition of activity, respec-    tively, following treatment with MTSET (Fig. 3D). The negatively charged MTSES caused a similar reduction (ϳ30%) in the activity of the S355C mutant, but had a considerably lesser effect on H353C and G356C. Because both reagents are similar in size, the differing degrees of inhibition caused by the respective MTS reagents are likely due to charge effects rather than steric hindrance. Collectively, these data suggest that amino acids His 353 , Ser 355 , and Gly 356 (equivalent to hNHE1 His 349 , Ser 351 , and Gly 352 ) comprise an accessible segment that faces an aqueous environment and are potentially involved in cation translocation. Likewise, the majority of the EL5 mutants were unaffected by exposure to the MTS derivatives, with the exceptions of G403C and S405C, which showed significant reductions (ϳ30 and 70%, respectively) in 22 Na ϩ o influx in the presence of MTSET, and to a lesser extent in the presence of MTSES (Fig. 4C).
Kinetic Properties of Mutants Sensitive to Thiol Modification-Having identified sites where cysteine substitutions confer sensitivity to thiol-reactive compounds, as revealed by reductions in transport activity, we next investigated whether the mutations alone (i.e. without MTS treatment) influenced the kinetic properties of the exchanger.
The majority of the other Cys substitutions in the predicted re-entrant loop EL5 did not confer sensitivity to the MTS compounds. Nevertheless, we explored whether some of these sites (i.e. Y370C, N374C, S376C, T381C, Y385C, S392C, E395C, and L397C) could contribute to substrate affinities and in some cases to their sensitivity to inhibition by EIPA. The majority of the mutations caused marked reductions in V max and cooperativity of Na ϩ binding (Table 1). We were unable to determine the kinetic properties of the N374C and L397C mutants because whereas their velocities increased as a function of the Na ϩ concentration, the values were erratic and did not display saturable binding kinetics (data not shown). Among these mutations, only Y370C and E395C exhibited marked reductions in Na ϩ o affinity (K Na mM: ⌬C, 46.1 Ϯ 0.9; Y370C, 92.8 Ϯ 2.0, p Ͻ 0.05; E395C, 65.7 Ϯ 3.6, p Ͻ 0.05) (Fig. 7, C and D; Table  1), but no significant shift in H ϩ affinity (Fig. 7, G and H; Table  1). When analyzed for sensitivity to inhibition by EIPA, Y370C exhibited a significant 4-fold increase (K 0.5 nM: ⌬C, 40.7 Ϯ 9.1, n ϭ 11; Y370C, 9.6 Ϯ 3.2, n ϭ 8; p Ͻ 0.05) (Fig. 8), whereas E395C was not significantly altered (K 0.5 nM: 67.6 Ϯ 17.3, n ϭ 8; p Ͼ 0.05). This suggests that whereas the Y370C and E395C mutants were unaffected by the MTS derivatives, the sites nevertheless contributed to cation permeation and partly to drug sensitivity.

Discussion
In this study, the substituted-cysteine accessibility method (73) was applied to rNHE1 to identify candidate amino acids in two segments of the transporter, the predicted transmembrane helix M9 (rNHE1-(342-363); equivalent to hNHE1-(338 -359)) and the adjoining exofacial re-entrant loop between M9 and M10 (EL5; rNHE1-(364 -415) or hNHE1-(360 -411)), which had previously been implicated in conferring sensitivity to competitive antagonists of Na ϩ o binding and therefore may potentially line the cation permeation pathway of the transporter (49, 56, 60). The data show that of the 20 cysteine substitutions made at a select number of known drug-sensitive or evolutionarily conserved sites, only a limited number of these in M9 (i.e. rat H353C, S355C, G356C; human H349C, S351C, and G352C) and EL5 (i.e. rat G403C and S405C; human G399C and S401C) caused marked reductions in NHE1 activity upon modification with MTS sulfhydryl reagents. Significantly, kinetic measurements of unmodified rNHE1 S355C, G403C, and S405C as well as two other substitutions in EL5 (rat Y370C and E395C; human Y366C and E391C) showed reduced affinities for Na ϩ o and/or H ϩ i . Interestingly, the Y370C mutant, whereas possessing a 2-fold decrease in Na ϩ affinity, exhibited a 4-fold increase in affinity for the competitive antagonist EIPA. This Tyr residue is highly conserved in the mammalian NHEs and therefore it is tempting to speculate that it may serve as a common binding site for Na ϩ o and side chain substituents of

Determinants of Cation and Drug Sensitivity of NHE1
amiloride derivatives. Other cysteine substitutions within EL5 also significantly impaired the maximal activity of the transporter, although they did not alter the apparent affinities for Na ϩ o or H ϩ i , suggesting that they may fulfill other structural roles in ion translocation. Collectively, these data implicate these amino acids as important elements involved in cation permeation and drug sensitivity.
Among the three MTS-sensitive residues within M9, the G356C mutation was the most reactive. The activity of G356C was almost completely inhibited (ϳ98%) in the presence of positively charged MTSET. Likewise, transport activity was also significantly reduced by negatively charged MTSES, although to a lesser extent (ϳ40%). Mechanistically, thiol modification of G356C may block cation transport by sterically hindering critical conformation changes that occur during cation transport, either at the cation pore itself or at a more remote site that nevertheless influences cation translocation. However, the observed sizeable differences in the degree of inhibition elicited by the two oppositely charged, but similarly sized, MTS reagents also implicates the involvement of electrostatic forces on the flow of cations through the protein, suggesting that Gly 356 is more likely in close proximity to the cation translocation pathway. Thus, aside from possible steric effects, the positive charge of MTSET could further impede the flow of Na ϩ by electrostatic repulsion, whereas the negatively charged MTSES reagent could act as an attractant and retard passage of Na ϩ ions through the pore, but with less efficiency than the repelling effects of MTSET. We have also reported (49) that this residue is an important determinant of the high sensitivity of NHE1 to inhibition by both amiloride-and benzyolguanidinium-based compounds, further supporting an important structural and functional role for Gly 356 .
Aside from G356C, cysteine substitution of the two adjacent N-terminal residues, H353C and S355C, also rendered the transporter moderately sensitive to inhibition by the MTS reagents. The reactivity of S355C agrees with an earlier study showing that the equivalent mutation in a cysteine-less variant of hNHE1, S351C, is also sensitive to MTS (60). The lesser reactivity of these two residues to MTS reagents might be expected if they lie deeper within the membrane, as predicted by the Wakabayashi model (see Fig. 9A) and their thiol side chains would be directed more toward the lipid bilayer, assuming 3.6 amino acids per turn of an ␣-helix relative to position Gly 356 and thereby making them less available for modification by the MTS reagents. This might also explain why MTS reagents did not have a detectable effect on rNHE1 mutants containing cysteine substitutions N-terminal to His 353 , although one previous study (60) reported that E350C (equivalent to hE346C) was sensitive to MTSET. The basis for the variance is unclear.
Although His 353 and Ser 355 are hydrophilic residues, it is perhaps initially unexpected to find that a hydrophobic residue such as Gly 356 would also directly face the aqueous pore. However, because Gly contains a hydrogen atom as its side chain group, it is only moderately nonpolar and confers flexibility to the peptide backbone structure that may be important for conformational changes that ensue during cation translocation. Moreover, previous studies have demonstrated that hydrophobic residues often found to line the translocation pores of transporters and channels where they are postulated to provide an inert surface that facilitates ion diffusion (80 -82). Kinetic analyses, however, did not reveal significant changes in the cation affinities for H353C and/or G356C, but did show a marked reduction in Na ϩ affinity for S355C, which agrees with a previous observation (60). This finding was further corroborated in native NHE1, where substitution of Ser 355 with Ala also reduced Na ϩ affinity by ϳ2-fold. Taken together, these data are consistent with the notion that M9, and these three residues in particular, constitute an integral part of the cation translocation pore.  The red or green shading identifies sites that conferred sensitivity to MTS reagents or affected substrate affinities (Na ϩ and/or H ϩ ), respectively, when replaced with cysteine. Mutations of some sites affected both parameters and are dual labeled.
The adjacent putative EL5 R-loop between M9 and M10 was also found to contribute significantly to transport activity and drug sensitivity. Indeed, manipulation of sites in EL5 often had more profound effects on transporter function than sites in M9. The R-loop shares structural similarities to P-loops that are known to invaginate into the membrane bilayer and facilitate ion flow through channels and pumps (83)(84)(85). These structural elements are responsible for mediating antagonist binding, ion selectivity, and conductance. Regarding NHE1, the vast majority of single cysteine substitutions in EL5 significantly decreased maximal transport velocity. In most cases (i.e. Y370C, S376C, Y385C, E395C, G403C, and S405C), these changes were associated with a reduction in positive cooperativity of Na ϩ binding, as defined by the Hill coefficient. Previous kinetic analyses of the Na ϩ dependence of the NHEs have provided mixed results with some studies reporting simple Michaelis-Menten kinetics (86,87), whereas others describe cooperative activating effects of Na ϩ (37,39,88,89). Although the basis for these differences is not fully understood, part of these differences may relate to assay conditions and/or cell type. In a detailed study by Fuster et al. (89), they observed that the Hill coefficient for extracellular Na ϩ dependence was dependent on cytoplasmic pH. Specifically, the extracellular Na ϩ concentration dependence was sigmoidal at a cytoplasmic pH of 7.2 with a Hill coefficient of ϳ1.8, whereas this cooperativity was diminished at more acidic values. The extracellular Na ϩ dependences of NHE1 were explained equally by either a parallel or serial model of dimer coupling with a 2Na ϩ /2H ϩ stoichiometry of the monomer, with cooperative Na ϩ o binding dependent on the H ϩ i concentration. Our mutagenesis data suggest that EL5 may play an important role in the cooperative activating effects of Na ϩ o . In the case of four mutants in EL5, Y370C, E395C, G403C, and S405C, the changes in activity also correlated with marked reductions in Na ϩ affinity (Y370C, E395C, and S405C) and/or H ϩ affinity (G403C and S405C). Both the G403C and S405C mutants were also sensitive to modification by MTS reagents, suggesting that they were accessible from the extracellular medium. Furthermore, the Y370C mutation exhibited increased sensitivity to inhibition by EIPA. Although other mutations in this region did not appear to affect ion or drug binding, the impairment in V max suggests that these sites may nevertheless be important for the ensuing conformational changes required for optimal transport activity.
Presently, two different transmembrane topologies have been postulated for NHE1. The first model developed by Wakabayashi et al. (36) using the substituted-cysteine accessibility mutagenesis approach predicted 12 membrane-spanning helices with M9 encompassing amino acids 342-363 of rat NHE1 (human 338 -359) followed by a re-entrant loop structure extending from amino acids 364 to 415 (human 360 -411), as illustrated in Figs. 1A and 9A. Subsequently, Landau et al. (65) proposed a different helical arrangement (Fig. 1B and 9B) based on comparisons of evolutionary conserved sequences and secondary structure predictions, and then mapped the optimized alignment to the crystal structure of the E. coli NhaA antiporter. In this alternate model, the N-terminal segment encompassing the original helices M1 and M2 (residues ϳ1-150) was located instead in the cytoplasm and postulated to serve as part of a truncated signal sequence, with the first transmembrane segment commencing at the former M3 helix. However, this seems unlikely given that the predicted EL1 segment between M1 and M2 is known to undergo N-and O-linked glycosylation (70) as it transits through the endoplasmic reticulum and Golgi compartments and therefore should be oriented extracellularly when the transporter is inserted into the plasma membrane. The M9 helix was also rearranged to form two short helices renamed M7 (amino acids 335-348) and M8 (amino acids 353-366), whereas the neighboring extracellular re-entrant loop 5 (EL5) was predicted to form a new intracellular loop (amino acids 367-377) and new M9 helix (amino acids 378 -402) that was predicted to be largely inaccessible to external reagents, followed by residues that again were accessible to external reagents (amino acids 403-418).
The results from our analyses are partially consistent with both models, but overall are better accounted for by the Wakabayashi model. The susceptibility of both G403C and S405C to modification by MTS reagents places these residues in an environment that is accessible to the extracellular milieu, in agreement with both models. The MTS sensitivity of H353C, S355C, and G356C, with the latter residue being the most reactive, also positions these amino acids extracellularly and possibly facing the external funnel of the ion permeation pathway. Based on the hierarchy of their MTS reactivity, these amino acids could be positioned according to the Wakabayashi model, but seem at odds with the Landau model, which places Gly 356 deep within the membrane. We further found that mutations to Ser 355 affected substrate affinities, especially for Na ϩ , and maximal transport velocity, consistent with an earlier finding (60). This suggests that this residue faces the ion pore and is catalytically important. In the Landau et al. (65) model, the side chain of Ser 355 was oriented into the lipid bilayer which, in principle, would make it less accessible. However, Landau et al. (65) postulated that this helical segment could potentially rotate 180°a round its axis to allow this residue to participate in cation permeation. Notwithstanding, the new proposed orientation of this segment also placed Tyr 370 within an intracellular loop. Such a position would make it more difficult to account for our observation that Tyr 370 is involved in external Na ϩ o binding and drug sensitivity. These data are explained more simply by positioning this residue in the extracellular milieu, but which is unaffected by MTS reagents for reasons that remain obscure. Because Gly 356 is also known to confer sensitivity to NHE antagonists (49), these residues may be in close proximity to each other. More recently, Reddy et al. (60) determined the structure of a peptide representing amino acids 342-369 (human 338 -365) by high resolution nuclear magnetic resonance (NMR) in detergent micelles. The structure contained two helical regions (amino acids Met 344 -Ser 348 and Ile 357 -Ser 363 ) separated by a sharp, potentially flexible, segment (amino acids Ala 349 -Gly 356 ) that bends immediately N-terminal to Ser 355 , resulting in a kinked "L"-shaped structure. If such an arrangement exists in the native transporter, then it has the potential to position Tyr 370 as well as neighboring amino acids in an intramembranous, albeit externally facing, environment that constitutes part of the pore funnel involved in ion coordination.
In summary, residues in M9 and EL5 comprise important structural elements for cation translocation in NHE1. As a crystal structure has yet to be achieved, additional biochemical, biophysical, and structural analyses using approaches that combine substituted-cysteine accessibility mutagenesis with electron paramagnetic resonance spectroscopy and NMR may help to shed further structural and mechanistic insight into this catalytically important region of the transporter. Such information should prove valuable in the design and development of more efficacious NHE1-specific drugs to prevent or lessen tissue damage arising from certain cardio-and cerebrovascular diseases and cancer progression.