Mutation E252C increases drastically the Km value for Na+ and causes an alkaline shift of the pH dependence of NhaA Na+/H+ antiporter of Escherichia coli.

A single Cys replacement of Glu at position 252 (E252C) in loop VIII-IX of NhaA increases drastically the Km for Na(+) (50-fold) of the Na(+)/H(+) antiporter activity of NhaA and shifts the pH dependence of NhaA activity, by one pH unit, to the alkaline range. In parallel, E252C causes a similar alkaline pH shift to the pH-induced conformational change of loop VIII-IX. Thus, although both the Na(+)/H(+) antiporter activity of wild type NhaA and its accessibility to trypsin at position Lys(249) in loop VIII-IX increase with pH between pH 6.5 and 7.5, the response of E252C occurs above pH 8. Furthermore, probing accessibility of pure E252C protein in dodecyl maltoside solution to 2-(4'-maleimidylanilino)-naphthalene-6-sulfonic acid revealed that E252C itself undergoes a pH-dependent conformational change, similar to position Lys(249), and the rate of the pH-induced conformational change is increased specifically by the presence of Na(+) or Li(+), the specific ligands of the antiporter. Chemical modification of E252C by N-ethylmaleimide, 2-(4'-maleimidylanilino)-naphthalene-6-sulfonic acid; [2-(trimethylammonium)ethyl]methane thiosulfonate, or (2-sulfonatoethyl)methanethiosulfonate reversed, to a great extent, the pH shift conferred by E252C but had no effect on the K(m) of the mutant antiporter.

physically and functionally interact (12). Based on the twodimensional crystals, a three-dimensional map of NhaA was obtained (13), the first insight into the architecture of the protein.
One of the most interesting characteristics of NhaA is its dramatic dependence on pH; both in isolated membrane vesicles and when purified in proteoliposomes its rate of activity changes more than 3 orders of magnitude between pH 7 and 8 (3,6). Amino acid residues involved in the pH response of NhaA have been identified in both loops and TMSs. 1 For example, His 225 in loop VII-VIII was found essential for the pH response of the antiporter (14,15). Gly 338 of TMS XI also affects the pH response of NhaA; its replacement with serine (G338S) produced a transporter that, in contrast to the wild type protein, lacks pH control (16).
NhaA undergoes a conformational change upon its activation by pH. Monoclonal antibody 1F6 raised against the NhaA antiporter (17) identified that the N terminus of NhaA, its epitope, responds to pH (18). The antibody binds NhaA at pH 8.5 but not at pH 4.5. Furthermore, H3C/H5C, a double mutation in this domain, changes the pH profile of NhaA (18).
Probing with trypsin digestion showed that Loop VIII-IX is another domain that changes its conformation with pH (19). Both in everted membrane vesicles as well as in DM micelles, NhaA is completely resistant to trypsin below pH 6.5 and with increasing pH is progressively cleaved at Lys 249 , reaching a maximum at pH 8.5. Furthermore, two NhaA mutants (H225R (19) and G338S (16)) with a modified pH profile are susceptible to trypsin, in isolated membrane vesicles, at Lys 249 , only at the pH range where they are active and reflecting the level of activity.
Two pieces of evidence suggest that Loop VIII-IX not only responds to pH but is also required for pH regulation: (a) Loop VIII-IX is located in the interface between monomers of NhaA dimer (12). Cross-linking between these loops of the NhaA dimer with a rigid and short cross-linking agent caused a dramatic change in the pH response as opposed to no effect of a long and flexible cross-linking reagent (12). (b) Insertion mutation Lys 249 -IEG-His 250 and Cys replacement mutations E241C and V254C in loop VIII-IX cause an acidic shift in the pH profile of NhaA (19).
In addition to Glu 241 , loop VIII-IX contains only one other Glu (at position 252). In this paper, we therefore, replaced Glu 252 with Cys and explored the properties of the E252C mutant both in the membrane and after solubilization and purification of the protein in DM. We found that in contrast to E241C and the other two mutations in loop VIII-IX that induce an acidic shift in the pH profile of activity of NhaA, the single amino acid change, E252C, causes a dramatic alkaline shift in the pH dependence of the Na ϩ /H ϩ antiporter activity of NhaA and increases drastically the K m of the antiporter to both Na ϩ and Li ϩ . In parallel, the E252C mutation causes a similar alkaline shift in the pH dependence of the conformational change in loop VIII-IX. Remarkably, this conformational change at position E252C was found sensitive specifically to both Na ϩ and Li ϩ , the specific ligands of the NhaA antiporter.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Culture Conditions-EP432 is an E. coli K12 derivative, which is melBLid, ⌬ nhaA1::kan, ⌬ nhaB1::cat, ⌬ lacZY, thr1 (5). TA16 is nhaA ϩ nhaB ϩ lacI Q and otherwise isogenic to EP432 (6). The cells were grown in modified L broth (LBK) in which NaCl was replaced by KCl (67 mM, pH 7.5). When indicated, the media was buffered by 60 mM BTP, and pH was titrated with HCl. The cells were also grown in minimal medium A without sodium citrate (20) with glycerol (0.5%) as a carbon source. Thiamine (2.5 g/ml) was added to all minimal media. For EP432, threonine (0.1 mg/ml) was also added (4). For plates, 1.5% agar was used. The antibiotics and their concentrations were 100 g/ml ampicillin and 50 g/ml kanamycin. Resistance to Li ϩ and Na ϩ was tested as described previously (4).
Plasmids-The plasmids encoding wild type NhaA used in this study were as follows. pECO2, a derivative of pECO (21), was constructed by purification of the BstXI-BstXI fragment (5188 bp) obtained from pECO (21), teaming with T 4 DNA polymerase and blunt end self-ligation and destroying the two BstXI sites of the plasmid. pBSTX is a derivative of pECO2 that contains a silent mutation introducing a BstXI site at codon 250 of nhaA. For construction of pBSTX, two fragments of nhaA were PCR-amplified using pGMAR100 (16) as a template and the end and mutagenic primers shown in Table I. The two fragments were annealed and used as a template for PCR amplification with the same end primers. The BglII-MluI fragment (682 bp) of the amplicon was ligated with BglII-MluI fragment (4502 bp) of pECO2.
Site-directed Mutagenesis-Site-directed mutagenesis was conducted following a polymerase chain reaction-based protocol (23). pBSTX-E252C and pCL-BSTX-E252C were constructed using plasmids pBSTX and pCL-BSTX as a template and the mutagenic primers described in Table I. In each case the entire fragment originated by PCR and cloned in a plasmid was sequenced through the ligation junction to verify the accuracy of mutagenesis.
pAXH2-E252C and pCL-XH2-E252C were obtained by replacing the MluI-BglII fragments of plasmids pAXH2 and pCL-XH2 with MluI-BglII fragments excised either from pBSTX-E252C or pCL-BSTX-E252C to obtain the mutation in either wild type or CL genetic background respectively. The plasmids carrying the Cys mutations in loops were previously described (22).
DNA Sequence-Sequencing of DNA was conducted by an automated DNA sequencer (ABI PRISM TM 377; PerkinElmer Life Sciences).
Isolation of Membrane Vesicles and Assay of Na ϩ /H ϩ Antiporter Activity-Assays of Na ϩ /H ϩ antiport activity were conducted on everted membrane vesicles of EP432 cells (24,25) transformed with the respec-tive plasmids. A fluorescence assay of antiport activity was performed as described (25) using acridine orange to measure generation of ⌬pH (the pH difference across the membrane). Energization was achieved with either Tris-D-lactate or ATP (1.6 mM each).
Protein Determination-The protein was determined according to Ref. 26.
Detection and Quantitation of NhaA and Its Mutated Proteins in the Membrane-Detection and quantitation of NhaA and its mutated derivatives in membranes of EP432 were conducted by Western analysis as described previously (14) using the NhaA-specific monoclonal antibody 1F6 (17). The amount of affinity-purified NhaA and its mutants was determined by Coomassie staining of the gel after SDS-PAGE as described previously (12).
Overexpression and Affinity Purification of His-tagged Antiporters by Ni 2ϩ -NTA Chromatography-To overexpress the plasmids encoding the His-tagged antiporters, TA16 cells transformed with the respective plasmids were used as described (6), and high pressure membranes were prepared as described (6). His-tagged NhaA was affinity-purified on Ni 2ϩ -NTA-agarose resin (Qiagen) by miniscale purification and eluted either with imidazole (10,22) for probing digestion with trypsin or acid (8) for labeling with MIANS.
Probing Digestion with Trypsin-For trypsin treatment (19), 15 g of affinity-purified protein was resuspended in 0.5 ml containing 0.1 M KCl, 0.7 mM EDTA, 1 mM CaCl 2 , 0.1% DM, and 50 mM Tris-HEPES at the indicated pH values. After the addition of 0.3 g of trypsin (Sigma, type III), the suspension was incubated at 37°C for 1 h. The reaction was terminated by adding 0.9 g of trypsin inhibitor type II (Sigma) dissolved in 1 mM HCl (200 g/ml). Then the protein was precipitated in 10% trichloroacetic acid for 0.5 h at 4°C, centrifuged (Eppendorf, 14.000 rpm, 30 min, 4°C), resuspended in sampling buffer, titrated to neutrality with Tris base, and loaded on the gel for SDS-PAGE as described (27).
Treatment with SH Reagents-Everted membrane vesicles or high pressure membranes were isolated from EP432 cells transformed with the indicated plasmids. The membranes (0.5 mg of membrane protein) were resuspended in a reaction mixture (0.5 ml) containing 5 mM MgSO 4, 100 mM potassium phosphate (pH 7.5), 1 mM NEM (Sigma), and MTSES or MTSET (10 mM each) and incubated for 20 min at 25°C with gentle shaking. The reaction was stopped by the addition of 20 mM dithiothreitol and 3 ml of TSC solution containing 10 mM Tris/Cl (pH 7.5), 250 mM sucrose, and 140 mM choline chloride. The membranes were centrifuged (Beckman, TLA 100.4, 75000 rpm, 20 min, 4°C) and resuspended in the same buffer (5-10 mg of membrane protein/ml). For measurement of Na ϩ /H ϩ antiporter activity of the treated membranes, ATP was used to energize the membranes because NEM inactivates the lactate-dependent respiration.
To determine accessibility to NEM or other SH reagents, the membranes (0.5 mg of membrane protein) were resuspended in 1.15 ml of TSC supplemented with 14% glycerol, 1% DM, and 0.06 M MOPS (pH 7). The suspension was incubated for 20 min at 4°C and centrifuged (Beckman, TLA 100.2, 75000 rpm, 20 min, 4°C). The supernatant was added to 100 l of Ni 2ϩ -NTA-agarose (Qiagen) and incubated with agitation for 1 h at 4°C. The beads were then washed twice in binding buffer (10) at pH 7.4 and resuspended in 100 l of binding buffer containing 0.2 mM fluorescein 5-maleimide (Molecular Probes) with gentle tilting for 30 min at 25°C and washed in washing buffer (10) at pH 7.4. The protein was eluted in 20 l of SDS-PAGE sampling buffer supplemented with 300 mM imidazole, agitated for 20 min at 4°C, and collected in the supernatant after centrifugation (Eppendorf, 14.000 rpm, 2 min, 4°C). The affinity-purified protein was separated on SDS-PAGE. For evaluation of the fluorescence labeling, SDS-PAGE gels were photographed under UV light (260 nm) as described (22).
Labeling of NhaA Mutants with MIANS and Fluorescence Measurement-For treatment with MIANS (Molecular Probes) the protein was affinity-purified on Ni ϩ2 -NTA with acid elution. The reaction mixture (2.5 ml) contained 0.03% DM, 50 mM BTP/Cl at the indicated pH values

Mutation E252C Increases K m and Shifts the pH Response of NhaA
and MIANS (4 M). The reaction was initiated by the addition of the affinity-purified protein (20 g of protein). Fluorescence was monitored continuously at 22°C with a spectrofluorometer (PerkinElmer Life Sciences) using an excitation wavelength of 330 nm (8-nm slit) and an emission wavelength of 415 nm (4-nm slit).

Cys Replacements Mutants of Loop VIII-IX-The
Cys-less-NhaA (CL-NhaA) is expressed similar to the wild type protein and is as active (Fig. 1D and Refs. 10 and 22). We constructed the Cys replacement mutation E252C in loop VIII-IX of both wild type and CL-NhaA. The mutated NhaA proteins were designated E252C and CL-E252C, respectively. To characterize the phenotypic properties conferred by the NhaA mutations, the mutant plasmids were transformed into EP432, a ⌬nhaA-nhaB strain. This strain can grow in high Na ϩ -selective media (0.6 M NaCl at pH 7 or 8.3) only when it expresses a functional NhaA. Hence, when transformed with plasmids bearing NhaA mutations, this host allows characterization of the effect of the mutations on cell and membrane phenotype. Fig. 1D shows that both mutants of NhaA, E252C and CL-E252C, were expressed in EP432 cells, even better than the wild type and the CL-NhaA, their respective controls. As compared with E252C mutant, CL-E252C mutant showed somewhat lower expression (Fig. 1D).
When Na ϩ was not added to the growth medium, the growth at both pH 7 and 8.3 of EP432 expressing either E252C or CL-E252C was similar to that expressing the wild type antiporter or its CL derivative (data not shown). In the selective medium, the growth of EP432 expressing E252C was very similar to that expressing the wild type protein with respect to both number and size of the colonies (Table II). The colony size of both strains was slightly smaller at alkaline pH than at neutral pH (Table II). On the other hand, the mutant CL-E252C conferred preferential growth at alkaline pH; although at both pH 7 and pH 8.3, the number of colonies of all strains was similar, the colony size of cells expressing CL-E252C was very small at pH 7, but at alkaline pH, it was even larger than that of wild type, CL-E252C, and E252C (Table II). It is not clear why CL-E252C but not E252C exhibits a pH-shifted growth phenotype.
E252C Shifts the pH Dependence of the Na ϩ /H ϩ Antiporter Activity of NhaA to the Alkaline Range-To determine the Na ϩ /H ϩ antiporter activity, everted membrane vesicles were isolated from EP432 expressing either E252C or CL-E252C. The Na ϩ /H ϩ antiporter activity was measured at various pH values. The assay was based on the measurement of the ⌬pH maintained across the membrane by respiration or H ϩ /ATPase as determined from fluorescence quenching of acridine orange. The Na ϩ /H ϩ antiporter activity was assessed from the dequenching caused by the addition of Na ϩ . Membranes derived from EP432 cells transformed with the vector plasmid have no Na ϩ /H ϩ antiporter activity (data not shown and (5)). Transformation with a plasmid expressing wild type NhaA restores Na ϩ /H ϩ antiporter activity (5) 1. The Na ؉ /H ؉ antiporter activity conferred by mutation E252C at various pH values. EP432 cells transformed with plasmids pBSTX, pBSTX-E252C, or pCL-BSTX-E252C encoding wild type (WT) or mutated NhaA (E252C or CL-E252C, respectively) were grown on LBK, and everted membrane vesicles were isolated. The ⌬pH was monitored with acridine orange at the indicated pH in a reaction mixture (2.5 ml) containing: 50 g of membrane protein, 0.5 M acridine orange, 150 mM KCl, 50 mM BTP, and 5 mM MgCl 2 . At the onset of the experiment, lactate (1.6 mM) was added (arrow pointing down), and the fluorescence quenching was recorded until a steady state level of ⌬pH had been reached (100%). NaCl (10 mM) was then added (arrow pointing up), and the new steady state level of fluorescence obtained (dequenching) was monitored. All of the experiments were repeated at least three times, and the results were essentially identical. A-C, the data of typical experiments are shown. D, Western blots of the indicated membrane samples (50 g of protein) using NhaA-specific monoclonal antibody 1F6 (17).

TABLE II
Cell and membrane phenotype of E252C and CL-E252C mutations EP432 cells transformed with the indicated plasmids were grown on selective agar plates containing 0.6 M NaCl at either pH 7 or pH 8.3. -, no growth on selective agar. ϩϩϩ, wild type growth with respect to both number and size of colonies; ϩϩ and ϩ, number of colonies identical but colonies size progressively smaller. The Na ϩ /H ϩ antiporter activity was determined in everted membrane vesicles, and the K m values for Na ϩ and Li ϩ were determined. All of the experiments were repeated at least three times, and the results were essentially identical. ND, not determined.
as observed at pH 9 (Figs. 1, B and C, and 2). The results suggest that whether in a wild type or CL genetic background, the E252C mutation causes a drastic alkaline shift, of 1 pH unit, in the pH dependence of NhaA.
The Effect of E252C on the K m of NhaA-The K m values for Na ϩ of the Na ϩ /H ϩ antiporter activity of E252C and CL-E252C mutants as compared with those of the wild type were measured at both pH 8 (Table II) and pH 8.5 (data not shown). At both pH values, the apparent K m for the Na ϩ ion (10 mM) of both mutants was drastically increased as compared with the wild type (0.2 mM). The K m for Li ϩ was also increased by the mutations, but the effect was less drastic than that for Na ϩ (Table II). The results summarized in Fig. 2 show that the drastic shift in the pH profile of the Na ϩ /H ϩ antiporter activity of both E252C and CL-E252C mutants is maintained even at saturating Na ϩ concentrations. Hence, the mutation E252C has an effect on the pH dependence of NhaA that is independent of its effect on the K m of the antiporter.
E252C Causes an Alkaline Shift to the pH-dependent Conformational Change of NhaA as Probed by Trypsin-Although NhaA has many trypsin-cleavable sites, trypsin digests NhaA, either in the native membrane or when purified in DM micells, only at Lys 249 in loop VIII-IX (19,28). Furthermore, the pH dependence of the digestion was identical in both NhaA preparations and reflected the pH dependence of the Na ϩ /H ϩ activity of the antiporter (19,28). To explore the pH dependence of the digestion by trypsin of E252C mutant, E252C protein was affinity-purified in DM and subjected to trypsin at various pH values (Fig. 3B). The results show that in contrast to the wild type protein that is progressively digested between pH 6.5 to pH 8.5 (Fig. 3A), the mutant is resistant to trypsin up to pH 8.5 and only between pH 8.5 and 9 is digested (Fig. 3B). It is also apparent that the two proteolytic fragments (Fig. 3B, HF and LF) obtained from the mutant protein are identical in size to those obtained from the wild type protein (Fig. 3A and Ref. 19). Identical behavior was obtained with mutant CL-E252C (data not shown). Hence, the alkaline shift caused to the pH profile of the antiporter activity by mutation E252C is reflected in a similar shift in the pH dependence of the conformational change of NhaA at position 249, as probed by trypsin.
Accessibility of E252C to MIANS Reveals a pH-dependent Conformational Change at Position 252-Because both positions 252 and 249 reside in loop VIII-IX, it was of interest to inquire whether E252C itself undergoes a conformational change with pH or rather merely changes the pH-dependent accessibility of Lys 249 to trypsin. We therefore tested accessibility of E252C in isolated membrane vesicles and in the pure protein (in DM micelles) to various SH reagents. For control we used other single Cys replacements that also reside in NhaA loops but, in contrast to E252C, do not affect the pH response of NhaA.
Several SH reagents have been used to modify Cys replacement mutations to determine accessibility, including maleimides (10, 29 -31) and MTS reagents (32,33). Modification of Cys by these reagents depends upon the ionization of the Cys sulfhydryl to its thiolate form (34 -36). This is expected to occur preferentially in an aqueous environment and to depend on pH, given that the pK of ionization of the sulfhydryl residue falls within the tested pH range (37). An important control to validate the specificity of the reaction of these reagents is to ensure that a Cys-less mutant of the respective protein does not react with the reagent under the reaction conditions used. As shown previously, NEM, a small maleimide reagent has no effect and does not bind CL-NhaA up to pH 8.5 (10). However, CL-E252C in membrane vesicles is readily accessible to NEM as many other single Cys replacements introduced into loops of NhaA (22). We found no difference in the pH dependence of NEM accessibility of E252C (data not shown) as compared with A118C of loop III-IV (22) or H225C of loop VII-VIII (10), replacements that have no effect on either the pH dependence of activity or conformation of NhaA as probed by trypsin.
To further probe the accessibility of E252C, we next used MIANS, a larger and negatively charged maleimide reagent that like NEM reacts specifically and covalently with thiol groups (38, 39) but becomes fluorescent only when its maleim- ide group reacts with Cys (40). We first tested MIANS labeling of Cys replacements in loops that have no effect on either the pH dependence of activity or conformation of NhaA as probed by trypsin. The reaction of MIANS with purified CL-NhaA containing single Cys residues H225C in loop VII-VIII (Fig.  4A) can readily be followed fluorimetrically. Above pH 6.5 the rate of the reaction with MIANS (4 M) increased with pH reaching a maximal rate at pH 8.5 and maximal fluorescence within 200 -300 s (Fig. 4A). The results obtained at pH 9 were very similar to those obtained at pH 8.5, and increasing the concentration of MIANS had no effect on maximal fluorescence level, suggesting that all Cys residues had been modified. Very similar pattern of reaction with MIANS was obtained at position K221C of loop VII-VIII and N177C of loop V-VI (data not shown). Furthermore, the pH dependence of the reaction at these positions was very similar to that of the reaction with NEM (data not shown).
However, a striking difference was observed in the pH de-FIG. 5. Na ؉ increases specifically the rate of MIANS labeling of E252C mutant. A and B, TA16 cells transformed with either pCL-AXH2-K221C, pCL-AXH2-H252C, or pCL encoding His 6 -CL-K221C, His 6 -CL-E252C, or His 6 -CL, respectively, were grown, and the protein was overexpressed and affinity-purified (20 g of protein) as described in Fig. 4. MIANS labeling was measured as in Fig. 4 at pH 9, but various concentrations of ions were added to the reaction mixtures as indicated. C, the rates of MIANS labeling for each Na ϩ concentration (2-20 mM) were determined separately (data not shown). At these Na ϩ concentrations the initial rates of MIANS labeling were linear with time for at least 90 s. These rates (in relative units/s/mg protein) are plotted versus Na ϩ concentrations.

FIG. 4. Accessibility of E252C to MIANS.
A and B, TA16 cells transformed with either pCL-AXH2-H225C or pCL-AXH2-E252C encoding His 6 -CL-H225C or His 6 -CL-E252C, respectively, were grown, and the protein was overexpressed and affinity-purified as described in the legend to Fig. 3, but the protein was eluted from the column by acid elution. MIANS (4 M) was added to a reaction mixture (2.5 ml) containing 0.03% DM, 50 mM BTP/Cl, at the indicated pH values. The reaction was initiated by the addition of the affinity-purified protein (20 g protein), and the fluorescence (excitation wavelength, 330 nm, 8-nm slit and emission wavelength, 415 nm, 4-nm slit) was monitored at 22°C. Each experiment was repeated at least three times, and the results were essentially identical. C, for comparison, the rates of MIANS labeling (in relative units/s/mg protein) of proteins CL-H225C and CL-E252C are plotted versus the different pH conditions of the reaction.
pendence of the reactivity to MIANS at position E252C. The reactivity of E252C with MIANS, at various pH values, as a function of time, is shown in Fig. 4B. It is evident that at pH 8 the control almost completed the reaction with MIANS within 300 s (Fig. 4A). In contrast E252C had a very low reactivity up to pH 8, and only above it the reaction progressively increased with pH; at pH 9 the maximal fluorescence level had hardly been reached even after 600 s. MIANS did not bind to CL-NhaA up to pH 9 ( Fig. 5B and data not shown), implying that at all pH values tested, the reaction of MIANS was specific to the Cys replacement in NhaA. It is apparent that, similar to its effect on the pH dependence of the antiporter activity and the conformational change at position Lys 249 as probed by trypsin, the Cys replacement of Glu 252 causes an alkaline shift to the accessibility to MIANS at position E252C.
The Rate of Labeling of E252C by MIANS Is Increased Significantly and Specifically by Na ϩ and Li ϩ -Because E252C increases the K m of the antiporter to its specific ligands, Na ϩ and Li ϩ , we tested the effect of Na ϩ on the rate of labeling of the NhaA protein with MIANS. In these experiments the ion was added at the indicated concentrations to the reaction mixture used for MIANS labeling (Fig. 5). For controls we used Cys replacements in loops that affect neither the pH response nor the K m of NhaA. An example, CL-K221C of loop VII-VIII, is shown in Fig. 5A. It is evident that both Na ϩ and K ϩ (20 mM each) had a minor effect on the rate of labeling by MIANS of CL-K221C mutant. However, in marked contrast, MIANS labeling of E252C was specifically affected by the type of added ion. Whereas K ϩ (20 mM) had no effect, Na ϩ (10 -20 mM) (Fig.  5B) or Li ϩ (10 mM; data not shown) significantly enhanced the rate of labeling of E252C.
The increase in MIANS fluorescence caused by Na ϩ and Li ϩ could be interpreted as an increase in quantum yield of E252C bound MIANS or as an increase in accessibility of the protein to MIANS. We prefer the latter alternative because we did not find any change in the absorption or emission spectrum of the fluorophore.
Chemical Modifications of E252C Alleviate the pH Shift Caused by the E252C Mutation-Most likely because of the Cys replacement at position E252, a negative charge was removed from loop VIII-IX of NhaA. Therefore, we modified E252C membranes by various SH reagents that chemically modify Cys while introducing either a negative charge (MTSES) or positive charge (MTSET) or that cause neutral modification (NEM) and measured the effect of the chemical modification both on the Na ϩ /H ϩ antiporter activity of the protein and its sensitivity to trypsin as a function of pH. The results summarized in Fig. 6A show the Na ϩ /H ϩ antiporter activity at pH 8 of NEM-treated CL-E252C membranes as compared with the untreated mutant or wild type membranes. It is evident that alkylation by NEM suppressed significantly the phenotype conferred by the mutation. Thus, at pH 8 when the activity of the wild type is maximal and untreated CL-E252C is very low, NEM treatment increases the activity of CL-E252C by about 3-fold. Fig. 6B shows that the effect of NEM on CL-E252C was saturated at 1 mM NEM.
We then explored the pH dependence of the Na ϩ /H ϩ antiporter activity of the NEM-treated CL-E252C membranes (Fig.  7). It is evident that the alkaline shift conferred by E252C mutation on the pH dependence of activity was, to a great extent, alleviated in the NEM treated membranes. However, NEM treatment did not change the K m of E252C.
Most importantly, similar to the effect on the antiporter activity, NEM treatment alleviated the alkaline pH shift conferred by the mutation on the pH-dependent conformational change as probed by trypsin (Fig. 3C). The pH dependence of the digestion by trypsin of the NEM-treated CL-E252C protein was very similar to that of the wild type (Fig. 3A). We conclude FIG. 6. Effect of NEM treatment on the Na ؉ /H ؉ antiporter activity of E252C. A, everted membrane vesicles (1 mg of protein) were isolated from EP432/pCL-BSTX or EP432/pCL-BSTX-E252C cells expressing CL or CL-E252C antiporter and treated with NEM (1 mM). The Na ϩ /H ϩ antiporter activity of the treated membranes was determined at pH 8 as described in the legend to Fig. 1. B, the membranes expressing CL-E252C were treated with the indicated concentrations of NEM and assayed for the Na ϩ /H ϩ antiporter activity as described in the legend to Fig. 1 (100% ϭ dequenching at 5 mM NEM). Each experiment was repeated at least three times, and the results were essentially identical.
FIG. 7. Treatment with NEM relieves the alkaline shift caused by E252C to the pH dependence of activity of the antiporter. Everted membrane vesicles were isolated from EP432/pCL-BSTX-E252C cells expressing CL-E252C antiporter, treated with NEM (1 mM) and tested for antiporter activity at various pH values as described in the legend to Fig. 2. The percentage of dequenching as observed following the addition of 10 mM NaCl is shown versus pH (f). For comparison, the pH dependence of the antiporter activity of wild type (WT, ࡗ) and untreated CL-E252C (OE) are also shown. Each experiment was repeated at least three times, and the results were essentially identical. that NEM alkylation of E252C reversed to a great extent the effect of the mutation on the pH response of the antiporter as reflected in the pH dependence of activity and the conformational change of the protein.
Treatment by MTSET and MTSES reagents had the same effect on E252C as that of NEM treatment (data not shown). Therefore, changing the native structure by the mutation E252C had a dramatic effect on the pH response of NhaA that could be suppressed by a chemical modification without restoring a negative charge at that position. DISCUSSION The pH regulation of NhaA (41), as of other both eukaryotic (42)(43)(44) and prokaryotic Na ϩ /H ϩ antiporters (2,(45)(46)(47), involves pH sensors and conformational changes in different parts of the protein. Accordingly, many amino acids in various domains of the protein participate in this regulation. Hence, to understand the mechanism underlying the pH regulation of NhaA, it is essential to identify the amino acid residues and domains involved and to elucidate the pH-induced conformational changes. In the present work, we have identified a novel mutation, E252C in loop VIII-IX, that has a very drastic effect on the pH response of NhaA. Whereas other mutations in this loop (E241C and V254C (19)) and other loops (15,18) cause acidic shift in the pH dependence of the activity of NhaA, Cys replacement mutation E252C causes an alkaline shift of one pH unit. As yet, such an alkaline shift of the pH response of NhaA has previously been found only in mutations at position 127 in the middle of TMS IV (21).
Furthermore, we show here that position E252C undergoes a pH-induced conformational change together with position Lys 249 , revealing that the pH-induced conformational change of the NhaA antiporter extends along a segment of loop VIII-IX. Using accessibility to digestion by trypsin to probe pH-induced conformational changes in NhaA, we have previously found that position Lys 249 in loop VIII-IX changes its conformation with pH, in a pattern that reflects the pH profile of the antiporter activity (19,28). Using the same approach, we show here that mutation E252C shifts drastically, by one pH unit (to the alkaline range), the pH dependence of the conformational change probed by trypsin at position Lys 249 (Fig. 3). Because probing with trypsin is limited to the trypsin-cleavable sites that exist along the protein molecule, this result can be interpreted in two ways: (a) positions Lys 249 and E252 in loop VIII-IX undergo a similar pH-induced conformational change and (b) E252C affects indirectly the conformation at position Lys 249 rather than participating in the conformational change. It thus has become crucial for the study of the pH-induced conformational changes in NhaA to introduce a probe that monitors conformational changes in a site-directed fashion.
In the present work, we show that accessibility to MIANS of nested Cys replacement mutations is a very good tool to probe, in a site-directed manner, pH-induced conformational changes in loops of NhaA. MIANS, a sulfhydryl reagent that reacts relatively specifically and covalently with thiol groups, is fluorescent only when the maleimide group reacts (38 -40). It has previously been used to probe accessibility and conformational changes in various membrane proteins (see for examples Refs. 39 and 48 -51). In Fig. 4C, the labeling rate by MIANS of Cys replacement E252C is compared with that of Cys replacement H225C. In contrast to E252C, H225C is a Cys replacement mutation in loop VII-VIII that does not affect the pH response of NhaA (22). The pH profile of the reactivity of the two proteins with MIANS was found to be very different; the pH dependence of reactivity to MIANS of E252C was shifted to the alkaline range by one pH unit as compared with that of H225C. Two other Cys replacement mutations, K221C in loop VII-VIII and N177C in loop V-VI, that, similar to H225C, do not affect the pH regulation of NhaA, behaved similarly to H225C (data not shown). Most importantly, up to pH 9, CL-NhaA did not react with MIANS (Fig. 5B). Hence, the reaction with MIANS is specific to the Cys replacement mutations.
Because the chemical reaction with MIANS depends on the ionized form of the sulfhydryl residue, it is possible that the pK of the Cys in the E252C mutant is different from that of the other Cys replacements tested in the other loops, because of a difference in the immediate environment of E252C in the protein. However, at all pH values, we did not find any difference in either the absorption or emission spectrum (properties that are sensitive to the immediate environment of the fluorophore (39,52,53)) of the E252C-bound MIANS (data not shown). Although we cannot exclude the possibility of a change in pK or a difference in the stereospecificity of the reaction with MIANS at position E252C, we prefer an alternative explanation: E252C changes its conformation with increasing pH and moves from a MIANS occluded to a MIANS accessible site. Hence, the pH dependence of the accessibility to MIANS at position E252C reflects a pH-induced change in the conformation at position E252C.
The revealed pH response of mutant E252C strongly supports our contention that loop VIII-IX is involved in the pH regulation of the activity of NhaA by changing its conformation with pH (19). It also shows that the pH-induced conformational change extends along the loop. Thus, two sites (Lys 249 and E252C) in loop VIII-IX of mutant E252C are shown to change conformation with pH ( Figs. 3B and 4B). Furthermore, the pH dependence of the pH-induced conformational change of E252C reflects the pH dependence of the Na ϩ /H ϩ exchange activity of mutant E252C (Figs. 1 and 2).
Remarkably, in addition to its effect on the pH-induced conformational change, the mutation E252C increases dramatically, by 10-fold, the K m of NhaA to Na ϩ (Table II). A somewhat lower but significant increase in K m was also noted for Li ϩ . Hence, E252C is the first known mutation in a loop of NhaA that affects the K m of the antiporter. Most interestingly, addition of Na ϩ (or Li ϩ ) increases significantly the rate of accessibility to MIANS at position E252C (Fig. 5B). Four lines of evidence lead us to suggest that it is the binding of Na ϩ , the specific ligand, that affects the conformational change at position E252C: (a) the effect is specific to Na ϩ , the specific substrate of NhaA and is not a general effect of ionic strength, because K ϩ has no effect (Fig. 5B); (b) the concentration of the ion that gives half-maximum increase in the rate of MIANS labeling (Fig. 5C) is within the range of the respective K m value of the antiporter activity of E252C mutant (Table II); (c) the effect of Na ϩ is observed at the pH range (pH 8.5-9) in which E252C is active (Figs. 1 and 2), and hardly any effect of the ion was discerned at pH 8; and (d) accessibility to MIANS of Cys replacements in other loops (that do not affect either the K m or pH response of NhaA) is not sensitive to the presence of Na ϩ (Fig. 5A). Our results show that E252C affects both the K m and the pH-dependent conformational change of NhaA, and in turn, binding of Na ϩ changes the rate of the pH-induced conformational change at position 252. Hence, the study of E252C has provided the first direct evidence that binding of the ligands of the antiporter and the pH-induced conformational change of the antiporter are related.
Although tested previously, we could not find any effect of Na ϩ on the pH-induced conformational changes of wild type NhaA. This was most likely due to Na ϩ contaminants that persisted in all media to a level similar to that of the K m (around 0.2 mM; Table II) of NhaA. In addition the MIANS accessibility assay is much more sensitive than the other meth-ods used previously (trypsin accessibility (19) and antibody binding (18)) to probe conformational changes in NhaA. Hence, the increased K m of E252C above the Na ϩ contaminants and the MIANS-based assay allowed detection of the effect of Na ϩ binding on the pH-induced conformational change of the antiporter.
Our previous results showed that residues that affect the pH response of NhaA may or may not overlap with those affecting the K m of the antiporter. Thus, similar to E252C of loop VIII-IX, A127V and A127T in TMS IV affect both the K m and the pH profile of NhaA (21). On the other hand D133C, T132C, and P129L, also in TMS IV, affect only the K m of the antiporter (21). Residues that affect exclusively the pH response of NhaA were also found: G338S in TMS XI (16) and all the rest in loops: 3HC/5HC in the N terminus part (18); H225R, H225D and H225A in loop VII-VIII (14); and V254C and E241C in loop VIII-IX (19). Hence, E252C is the first mutation in a loop that affects both the pH response of NhaA and its K m to Na ϩ .
There are many examples including that of NhaA (21) of residues located in TMS that affect the apparent K m or K d for substrates of transporters. These have been implicated to reside in the translocation passage of the respective substrates (see for example Ref. 30). However, recently more and more examples (36,54) of residues in loops were found that, similar to E252C in NhaA, affect the kinetics and other parameters of the transporters. It is clear that atomic resolution of NhaA as of the other transporters is needed to get a glimpse as to how and, whether directly or indirectly, loops participate in the translocation mechanism.
The effect of E252C on the pH profile of NhaA cannot be ascribed to a missing single negative charge that changes the overall charge of loop VIII-IX because E241C, a mutation in another Glu in the same loop, causes an acidic rather than alkaline shift in the pH profile of NhaA (19). It can be argued that, specifically at position 252, the negative charge is required to obtain the wild type pH profile. This is also unlikely because chemical modifications of E252C by MTS reagents that add either a negative (MTSES) or a positive charge (MTSET) at position 252 or NEM that does not add any charge upon modification all partially shift the pH profile back to the wild type pH profile (Fig. 7). Hence, it is the presence of the Cys replacement at position 252 that causes the alkaline shift, whereas its modification by various agents reverses the pH profile. We therefore suggest that the size and/or stereospecificity of the residue at position 252 is crucial for the pH response of NhaA.