Role of the Charge Interaction between Arg70 and Asp120 in the Tn10-encoded Metal-Tetracycline/H+ Antiporter of Escherichia coli *

We reported that the positive charge of Arg70 is mandatory for tetracycline transport activity of Tn10-encoded metal-tetracycline/H+ antiporter (TetA(B)) (Someya, Y., and Yamaguchi, A. (1996)Biochemistry 35, 9385–9391). Arg70 may function through a charge-pairing with a negatively charged residue in close proximity. Therefore, we mutated Asp66 and Asp120, which are only two negatively charged residues located close to Arg70 in putative secondary structure of TetA(B) and highly conserved throughout transporters of the major facilitator superfamily. Site-directed mutagenesis studies revealed that Asp66 is essential, but Asp120 is important for TetA(B) function. Surprisingly, when Asp120was replaced by a neutral residue, the R70A mutant recovered tetracycline resistance and transport activity. There was no such effect in the Asp66 mutation. The charge-exchanged mutant, R70D/D120R, also showed significant drug resistance and transport activity (about 50% of the wild type), although the R70D mutant had absolutely no activity, and the D120R mutant retained very low activity (about 10% of the wild type). Both the R70C and D120C mutants were inactivated by N-ethylmaleimide. Mercuric ion (Hg2+), which gives a positive charge to a SH group of a Cys residue through mercaptide formation, had an opposite effect on the R70C and D120C mutants. The activity of the R70C mutant was stimulated by Hg2+; however, on the contrary, the D120C mutant was partially inhibited. On the other hand, the R70C/D120C double mutant was almost completely inactivated by Hg2+, probably because the side chains at positions 70 and 120 are bridged with Hg2+. The close proximity of positions 70 and 120 were confirmed by disulfide cross-linking formation of the R70C/D120C double mutant when it was oxidized by copper-(1,10-phenanthroline). These results indicate that the positive charge of Arg70 requires the negative charge of Asp120 for neutralization, probably for properly positioning transmembrane segments in the membrane.

Tetracycline resistance in bacterial cells is mediated mainly by the integral membrane protein designated TetA 1 (1), which catalyzes antiport of a divalent cation-tetracycline complex and a proton (2,3). The class B TetA protein (TetA(B)) encoded by transposon Tn10 is composed of 401 amino acids (4,5) and is believed to have 12 membrane-spanning segments (6 -8). Its structural and functional features are distinctly those of secondary transporters widely spread in many organisms (9,10). Therefore, this protein is a good model system for studying the molecular mechanism of membrane transport.
Several highly homologous amino acid sequence motifs are conserved in equivalent positions of proton-coupled symporters antiporters, and uniporters (9,10), suggesting the common role(s) of these motifs in transport phenomena. Three motifs, GXXXDRXGRR in loop 2-3, DXXXXXXR in loop 4 -5, and PESPR in loop 6 -7, are also found in the cytoplasmic face of the N-terminal half of TetA(B). We have already focused on the loop 2-3 motif itself (11) and on several charged residues in these motifs (12)(13)(14)(15), and we found Asp 66 and Arg 70 in loop 2-3, Asp 120 and Arg 127 in loop 4 -5, and Glu 181 in loop 6 -7 to be residues essential or important for function, which was based on the loss of transport activity and altered kinetic constants. However, we have not yet identified the precise role(s) of these residues, except that the negative charge of Asp 66 may first interact with a monocationic substrate, tetracycline-divalent cation complex (16).
Possible salt bridges in the transmembrane region were reported on Escherichia coli lactose permease (17)(18)(19)(20)(21) and a rat vesicular monoamine transporter (VMAT2) (22) and seem to be important for transport activity and substrate recognition. However, such salt bridges in the transmembrane region are not always found in all transporters, because two charged amino acid residues participating in them are not conserved in most of the secondary active transporters. Therefore, we focused on several charged residues (Asp 66 , Arg 70 , and Asp 120 ) in the cytoplasmic loops of TetA(B) and analyzed the salt bridge formation. Interestingly, because these three resides are all conserved in the secondary transporters, the results should be applicable to all secondary transporters belonging to the large family of proton-coupled symporters, antiporters, and uniporters.

EXPERIMENTAL PROCEDURES
Materials-[7-3 H]Tetracycline was purchased from NEN Life Science Products. Restriction and modifying enzymes were obtained from Takara, Toyobo, and New England Biolabs. o-Phenanthroline was purchased from Sigma. All other chemicals were of reagent grade and were obtained from commercial sources.
Bacterial Strains-E. coli W3104 (23) was used for the preparation of inverted membrane vesicles and measurement of tetracycline resist-* This work was supported in part by a grant-in-aid from the Ministry of Education of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ance. CJ236 was used for the preparation of deoxyuracil-containing single-stranded DNA templates for site-directed mutagenesis according to the method of Kunkel (24). TG1 (25) was used for all other DNA manipulations.
Site-directed Mutagenesis and Plasmid Construction-Low copy number plasmid pLGT2 (26) was used for expression of the tet gene. The following mutant low copy number plasmids have been constructed previously: (a) pLGD66N (12), (b) pLGR70A (11), (c) pLGR70C (14), (d) pLGR70D (11), and (e) pLGD120N (13). Mutations D66R, D120R, and D120C were introduced into multicopy plasmid pER (26), which is a subclone of the first half of the tetA gene, using mutagenic primers 5Ј-tggaaaaatgAGtCGAcgatttggt-3Ј, 5Ј-gtcattgccCGtacGacctcagct-3Ј, and 5Ј-cggtcattgcATGCaccacctcag-3Ј (capital letters indicate mismatches). Single mutant, low copy number pLG plasmids were constructed by inserting the EcoRV-EcoRI fragment from mutant pER plasmids into the EcoRV-EcoRI vector fragment of pLGT2. The D66R mutagenic primer was also used for the construction of pERD66R/R70D, and the R70A primer (11) was used for the construction of pERD66N/R70A. pLG plasmids containing these four double mutations were also constructed by inserting the EcoRV-EcoRI fragment from multicopy plasmids into pLGT2. The Arg 70 /Asp 120 double mutant pLG plasmids were constructed by inserting the EcoRV-PshBI fragment containing the Arg 70 substitution and the PshBI-EcoRI fragment containing the Asp 120 substitution into pLGT2.
Preparation of Inverted Membrane Vesicles and Transport Assays-Inverted membrane vesicles were prepared by passing E. coli W3104 cells harboring pLG plasmids through a French pressure cell, and the vesicles were suspended in 50 mM MOPS-KOH buffer (pH 7.0) containing 0.1 M KCl and stored at Ϫ80°C (11). [ 3 H]Tetracycline uptake by inverted membrane vesicles was assayed as described previously (14). Inverted vesicles were energized with NADH and exposed to a solution containing 10 M [ 3 H]tetracycline and 50 M CoCl 2 for the indicated times. The effect of the Hg 2ϩ ion on tetracycline uptake was tested in the presence of 10 M HgCl 2 . Kinetic constants were measured in the presence of 1 mM CoCl 2 and various concentrations of [ 3 H]tetracycline. The effect of N-ethylmaleimide was measured as described previously (11).
Oxidative Disulfide Cross-Linking-The oxidative disulfide crosslinking experiment was based on the method described by Lynch and Koshland (27). Incubation at 37°C was executed for 2 h. The reaction was stopped by adding 40 l of 20 mM EDTA (pH 7.4), followed by the addition of 80 l of ␤-mercaptoethanol-free 2ϫ sample buffer or normal 2ϫ sample buffer (28). A portion (40 l) of reaction mixture was applied to SDSpolyacrylamide gel (10% gel), followed by electroblotting onto a nitrocellulose membrane. TetA(B) proteins were detected using rabbit anti-Ct14 antisera (29), horseradish peroxidase-conjugated goat anti-rabbit IgG, and enhanced chemiluminescence Western blotting detection reagents (Amersham).

The Effect of Mutation of Asp 66 and Asp 120 on the Arg 70
Mutant-Asp 66 is located in the same loop region as Arg 70 , and Asp 120 is located at the adjacent cytoplasmic loop in the putative secondary structure of TetA(B) (Fig. 1). They are the only negatively charged residues in each loop. The site-directed mutagenesis of these residues was performed previously (12,13), and in this study, additional mutations were constructed as described under "Experimental Procedures." The resulting mutants were subcloned into a low copy number plasmid, pLGT2, and expressed in E. coli W3104 cells. The tetracycline resistance levels of these mutants are shown in Table I. The D66N and D66R mutants showed no drug resistance, consistent with the results of our previous study (12). Drug resistance was not recovered in either the charge-neutralized (D66N/ R70A) or the charge-exchanged (D66R/R70D) double mutant. On the other hand, the D120N and D120R mutants retained low but significant drug resistance (19 g/ml), indicating that Asp 120 is not mandatory for TetA(B) function. As reported in previous studies (11,14), the R70A mutant showed low drug resistance but almost no tetracycline transport, whereas the R70D mutant showed absolutely no drug resistance (Table I). When Asp 120 of the R70A or R70D mutant was replaced by site-directed mutagenesis, the resulting charge-neutralized (R70A/D120N) and charge-exchanged (R70D/D120R) mutants clearly recovered drug resistance in comparison with either the Arg 70 or Asp 120 single mutant (Table I). The resistance of the charge-exchanged mutant (50 g/ml) is higher than that of the charge-neutralized mutant (38 g/ml).
Tetracycline Transport Activities of the Arg 70 and Asp 120 Single or Double Mutants-Tetracycline transport activities were measured in the everted membrane vesicles prepared from E. coli W3104 cells expressing mutant tetA(B) genes. Under normal assay conditions (50 mM MOPS-KOH, pH 7.0, in the presence of 0.1 M KCl, 50 M CoCl 2 , and 10 M [ 3 H]tetracycline), the R70A mutant showed almost no transport activity, whereas the D120N mutant retained about two-thirds of the wild-type activity ( Fig. 2A). Surprisingly, the R70A/D120N double mutant showed significant transport activity (about 50% of the wild-type activity; Fig. 2A), indicating that the positive charge at position 70 is not essential for the transport function when Asp 120 was replaced with a neutral residue. The transport activity of the double mutant was slightly less than that of the D120N single mutant under these conditions. When Arg 70 was replaced with Asp, the transport activity was completely lost (Fig. 2B). On the other hand, the D120R mutant had significant transport activity (about 25% of the wild-type activity; Fig. 2B), which was less than that of the   Fig. 2, C and D, shows the tetracycline transport by everted membrane vesicles under the conditions in which TetA(B) was almost saturated by the tetracycline-Co 2ϩ chelation complex. Under these conditions, the R70A mutant showed very low transport activity (Fig. 2C), whereas the R70D mutant had absolutely no activity (Fig. 2D). The relative transport activity of the D120N and D120R mutants (Fig. 2, C and D), compared with the wild-type, was significantly lower than that under normal conditions (Fig. 2, A  and B). Both the charge-neutralized and charge-exchanged mutants showed clearly higher activity (Fig. 2, C and D) than either the Arg 70 or Asp 120 single mutant.
V max and K m values for tetracycline uptake by these mutant TetA(B) proteins were calculated from Lineweaver-Burk plots of the initial rate (30 s) of tetracycline transport versus the tetracycline concentration in the presence of 1 mM CoCl 2 (Table  II). V max values were in agreement with the results of Fig. 2, C and D. The order of magnitude of the V max values is as follows: The signs in parentheses indicate a positive charge (ϩ), negative charge (Ϫ), or no charge (0) at positions 70 and 120, respectively. Thus, it seems preferable for tetracycline transport function that the sum of the charges at positions 70 and 120 be equal to 0. Net negative or positive charge(s) in this pair hinders the transport function.
On the other hand, K m values reveal a perspective different from V max values as to the relationship between the side chains at positions 70 and 120. As shown in Table II, all of the Arg 70 single mutants and Arg 70 /Asp 120 double mutants showed K m values similar to those of the wild-type; however, only the Asp 120 single mutants showed significantly reduced K m values (values were reduced by a factor of about 4). The reduction of K m values in the D120N and D120R mutants is probably due to the apparent reduction of the dissociation constant of the substrate from the carrier caused by the decrease in the substrate translocation. If this is the case, why is the K m value of the R70A mutant not reduced, despite its very low translocation rate (V max value)? One reasonable answer to this question may be that the reduction of the substrate binding affinity that occurred in the R70A mutant covers the reduction of the apparent dissociation constant, resulting in the unaltered K m value.
The Effect of SH Reagents and Mercuric Ions on the Cys Mutants at Positions 70 and 120 -Single or double Cys mutants at positions 70 and 120 were constructed by site-directed mutagenesis in order to investigate the effect of site-specific chemical modification at these positions. Fig. 3 shows the Nethylmaleimide (NEM) concentration dependence of tetracycline transport activity mediated by the inverted membrane vesicles containing mutant TetA(B) proteins. In our previous study (12), we reported that the transport activity of wild-type TetA(B) was not affected by NEM at all. The R70C single mutant and the R70C/D120C double mutant were almost completely inactivated by NEM, whereas the D120C mutant was only partially inactivated (about 40% under the saturated NEM concentration). As described in our previous study (30), the chemical modification of the water-exposed Cys residue is saturated at about 2 mM NEM. Therefore, it seems that the partial inactivation of the D120C mutant is not due to the low reactivity of the Cys residue at position 120. The results indicate that each molecule of the ethylmaleimidyl-D120C mutant retains 60% of the activity before modification. A similar result was obtained when the S65C mutant was inactivated by methyl methanethiosulfonate (12). The incomplete inactivation indicates that the inactivation is due to steric hindrance but not to the masking of the functional group.
When Arg 70 was replaced with Cys, the resulting mutant retained low but significant activity, probably due to the fact that the weak mercaptide formation between Cys 70 and Co 2ϩ  confers a positive charge at position 70 (14). As shown in Fig.  4A, Hg 2ϩ significantly stimulated the transport activity of the R70C mutant by a factor of about 2.5 because Hg 2ϩ forms stable mercaptide with Cys 70 , which confers a positive charge that is fully functional during the transport process (14). On the other hand, the D120C mutant is partially inhibited by Hg 2ϩ by a factor of about 2 (Fig. 4B). As a result, the R70C and D120C single mutants showed similar levels of transport activity. The transport activity level of the D120C mutant in the presence of Hg 2ϩ was also similar to that of the D120R mutant (Fig. 2B).
On the contrary, the R70C/D120C double mutant was almost completely inactivated by Hg 2ϩ (Fig. 4C). The effect of Hg 2ϩ on the double Cys mutant was not due to the sum of the effects on each single mutant. If Cys 70 and Cys 120 are located in close proximity, Hg 2ϩ is expected to bridge these two Cys residues. The complete inactivation of the R70C/D120C double mutant by Hg 2ϩ is probably due to the mercaptide linkage between the side chains of positions 70 and 120.
The Formation of Disulfide Linkage between Cys 70 and Cys 120 -To examine whether or not the distance between the side chains of Cys 70 and Cys 120 is small enough to form a mercaptide linkage, we investigated the formation of disulfide cross-linking between these two SH groups when the proteins were oxidized. The everted membrane vesicles containing the single or double Cys mutants at positions 70 and 120 were oxidized with Cu 2ϩ /o-phenanthroline. The electrophoretic distance of the oxidized proteins was then analyzed by Western blotting (Fig. 5). Both the wild-type and single Cys mutants showed no alteration of the electrophoretic distance under all conditions examined. On the other hand, when the R70C/ D120C mutant was oxidized, the electrophoretic distance was significantly elongated, indicating that the apparent molecular volume shrunk. When the SH groups of the double mutant were modified by NEM before oxidation, the distance was not changed by oxidation. In addition, when the oxidized proteins were reduced with ␤-mercaptoethanol, the electrophoretic distance returned to the normal value. These results indicate that the SH groups of Cys 70 and Cys 120 make a disulfide linkage by oxidation, resulting in the conformational change of the proteins.

DISCUSSION
In our previous study (14), we raised the possibility that the ionic interaction of Arg 70 with an acidic residue contributes to the switching of the opening/closing of the entrance transport path. Therefore, in this study, Asp 66 or Asp 120 was mutated as a candidate in combination with an Arg 70 mutation. Our results supported the observation that the positive charge of Arg 70 interacts with the negative charge of Asp 120 , whereas the charge pairing between Arg 70 and Asp 120 is not essential for function. When both Arg 70 and Asp 120 were neutralized, and when the charges of Arg 70 and Asp 120 were exchanged, the tetracycline resistance level (Table I) and the V max value of the tetracycline transport (Table II) were much higher than those of the respective single mutants. It seems preferable that the net charge between positions 70 and 120 be equal to 0. One possibility is that the net negative charge between positions 70 and 120 prevents proper positioning of the transmembrane segments in the membrane. It was interesting that the tetracycline transport activity of the Cys 70 /Cys 120 double mutant was almost completely inhibited by Hg 2ϩ (Fig. 4). This suggests that Hg 2ϩ bridges two SH groups from positions 70 and 120. Oxidative disulfide cross-linking experiments using Cu 2ϩ /ophenanthroline clearly indicated cross-linking between the SH groups of Cys 70 and Cys 120 (Fig. 5). Thus, it is concluded that Arg 70 is located close to Asp 120 . It has been estimated that the average distance between the ␣-carbons of the cross-linked cysteine residues is 5-6 Å (31). However, in the normal (nonoxidized) state, Cys 70 and Cys 120 were not cross-linked because the effect of NEM on the tetracycline transport was additive (Fig. 3), and the electrophoretic mobility of the non-oxidized Cys 70 /Cys 120 mutant protein was the same as that of the NEMtreated Cys 70 /Cys 120 mutant protein (Fig. 5). Therefore, it is likely that the distance between Cys 70 and Cys 120 is greater than average. It is known that the maximum distance at which two SH groups can be cross-linked is 7 Å (31).
Both Arg 70 and Asp 120 of TetA(B) are well conserved in transporters of major facilitator superfamily (9, 10), a fact that is therefore suggestive of a common role in the mechanism of transport and/or protein folding in the membrane. We previously concluded that the positive charge of Arg 70 was important for TetA(B) activity because the Lys 70 mutant retained the highest activity (11) and the Cys 70 mutant was activated by Hg 2ϩ (14). In a mutagenesis study of E. coli ␣-ketoglutarate permease, Seol and Shatkin (32) reported that Arg 92 , which corresponds to Arg 70 of TetA(B), was necessary for activity. On the contrary, the corresponding Arg 73 residue of lactose permease does not seem to be critical for activity (33,34). However, the motif in loop 2-3 of lactose permease is slightly modified; that is, an extra Leu residue is inserted before the Arg 73 residue. The difference between lactose permease and TetA(B) (and also ␣-ketoglutarate permease) may be due to the difference in the amino acid sequence of the motif. However, TetA(B) and lactose permease share the property that the loop 2-3 region of two transporters is important for conformational change of the protein during the transport cycle (33,(35)(36)(37)(38)(39).
Asp 120 of TetA(B) was also important for activity because the kinetics of tetracycline transport was corrupted (Ref. 13 and this study). To date, the effects of mutation of the residue corresponding to Asp 120 of TetA(B) were tested on lactose permease (40) and the myo-inositol/H ϩ transporter from Leishmania donovani (41). Both Glu 126 of lactose permease and Glu 121 of the myo-inositol/H ϩ transporter played a critical role in transport function, whereas Glu 126 in lactose permease is thought to form a charge pair with Arg 144 . In any case, the steady-state level of substrate transport was drastically decreased.
We emphasize that the interaction between the positive charge of Arg 70 and the negative charge of Asp 120 may contribute to the conformational change of the protein and/or proper positioning of the transmembrane segments in the membrane. The latter idea is consistent with the conclusion of a recent study on Glut1 (42) that the conserved motif (RXGRR) is important for the correct topogenesis. This is thought to be the feature common in members of the major facilitator superfamily.
FIG. 5. Oxidation-induced disulfide cross-linking between Cys 70 and Cys 120 . Reaction conditions are described under "Experimental Procedures." After the reaction, membrane vesicles were mixed with ␤-mercaptoethanol-free sample buffer (A) or normal sample buffer (B) and subjected to SDS-polyacrylamide gel electrophoresis, followed by electroblotting onto nitrocellulose membrane. Proteins were detected with enhanced chemiluminescence Western blotting detection reagents (Amersham).