Electrogenic Antiport Activities of the Gram-positive Tet Proteins Include a Na+(K+)/K+ Mode That Mediates Net K+ Uptake*

Two Gram-positive Tet proteins, TetA(L) from Bacillus subtilis and TetK from aStaphylococcus aureus plasmid, have previously been suggested to have multiple catalytic modes and roles. These include: tetracycline (Tc)-metal/H+ antiport for both proteins (Yamaguchi, A., Shiina, Y., Fujihira, E., Sawai, T., Noguchi, N., and Sasatsu, M. (1995) FEBS Lett. 365, 193–197; Cheng, J. Guffanti, A. A., Wang, W., Krulwich, T. A., and Bechhofer, D. H. (1996) J. Bacteriol. 178, 2853–2860); Na+(K+)/H+ antiport for both proteins (Cheng et al. (1996)); and an electrical potential-dependent K+ leak mode for TetK and highly truncated segments thereof that can facilitate net K+ uptake (Guay, G. G., Tuckman, M., McNicholas, P., and Rothstein, D. M. (1993) J. Bacteriol.175, 4927–4929). Studies of membrane vesicles from Escherichia coli expressing low levels of complete and 3′-truncated versions of tetA(L) or tetK, now show that the full-length versions of both transporters catalyze electrogenic antiport and that demonstration of electrogenicity depends upon use of a low chloride buffer for the assay. The K+ uptake mode, assayed via 86Rb+ uptake, was also catalyzed by both full-length TetA(L) and TetK. This mode does not represent a potential-dependent leak. Such a leak was not demonstrable in energized membrane vesicles. Rather, Rb+uptake occurred in right-side-out vesicles when the intravesicular space contained either Na+ or K+ but not choline. If an outwardly directed gradient of Na+ or K+ was present, Rb+ uptake occurred without energization in vesicles from cells transformed with a plasmid containing tetA(L) or tetK but not a control plasmid. Experiments in which a comparable exchange was carried out in low chloride buffers to which oxonol was added confirmed that the exchange was electrogenic. Thus, the K+ uptake mode is proposed to be a mode of the electrogenic monovalent cation/H+ antiport activity of TetA(L) and TetK in which K+ takes the place of the external protons. Truncated TetK and TetA(L) failed to catalyze either Tc-metal/H+ or Na+/H+ antiport in energized everted vesicles. Truncated TetK, but not TetA(L), did, however, exhibit modest, electrogenic Na+(K+)/Rb+ exchange as well as a small, potential-dependent leak of Rb+. The C-terminal halves of the TetA(L) and TetK proteins are thus required both for proton-coupled active transport activities of the multifunctional transporter and, perhaps, for minimizing cation leakiness.

Tc 1 enters bacterial cells in a non-carrier dependent fashion that is promoted by a transmembrane pH gradient, acid out (1). The antibiotic thus enters the cell best under neutral and acidic pH conditions and could inhibit cell protein synthesis strongly in sensitive cells in this pH range. Both Gram-positive and Gram-negative Tet efflux proteins catalyze similar exchange reactions which prevent cytoplasmic accumulation of the antibiotic. Tc is actively extruded, as a complex with a divalent cation that bears a single positive charge, in exchange for external H ϩ (2,3). The smaller (12-transmembrane segments) Gram-negative Tet proteins and the larger (14-transmembrane segments) Gram-positive Tet proteins share sequence similarity largely in the N-terminal six transmembrane segments regions (4,5) but at least some motifs and/or residues in the C-terminal halves of each type of Tc efflux protein cannot be modified without loss of activity (6,7). Both the Gram-negative and Gram-positive Tet protein families contain examples that have further been shown to complement K ϩ -uptake deficient mutants of Escherichia coli (8 -11), but this net K ϩ uptake mode is not taken as a general property of Tet proteins. It has been attributed to an electrical potential-dependent K ϩ leak that could also be conferred by truncated forms of proteins that exhibit the property (10 -12). Recently, studies in this laboratory have shown that the chromosomally encoded Bacillus subtilis TetA(L) protein and closely related TetK from a Staphylococcus aureus plasmid catalyze Na ϩ (K ϩ )/H ϩ antiport (13)(14)(15)(16)(17) in addition to Tc Ϫ -Me 2ϩ /H ϩ antiport (2,13,14). These exchanges were evidently electrogenic, as assayed via energy-dependent Tc-cobalt or Na ϩ uptake by everted vesicles of E. coli that expressed a cloned tetA(L) gene from a weak promoter (14). The exchanges were not inhibited by low nigericin concentrations that reduce the ⌬pH but were significantly inhibited by valinomycin in the presence of K ϩ , a combination that abolished the ⌬⌿ generated by respiration (14). Consistently, the antiports catalyzed by purified and reconstituted TetA(L) could be energized by an imposed potential (16). In addition, the important role of TetA(L) in acidifying the cytoplasm of B. subtilis relative to the external medium during growth at alkaline pH would require that the monovalent cation/H ϩ mode be electrogenic (15,18). Since TetK could substitute for TetA(L) in a mutant of B. subtilis that had a disrupted tetA(L) gene, TetK is presumed to catalyze an electrogenic antiport similar to TetA(L). By contrast, the Tc-metal/H ϩ antiport catalyzed by Gram-negative Tet proteins has been proposed to be electroneutral (19). Moreover, Yamaguchi and colleagues (3,20) have experienced difficulty in demonstrating the Na ϩ /H ϩ activity of TetK and indicate that the Tc-metal/H ϩ antiport activity of TetK appeared to be electroneutral in preliminary work. One of the goals of the current study, therefore, was to examine the Tc-metal/H ϩ and Na ϩ /H ϩ antiport activities of TetA(L) and TetK side-by-side in comparable preparations and to clarify their electrogenicity versus electroneutrality. The studies have strongly supported the multifunctional and electrogenic nature of both TetA(L) and TetK.
A second major goal of the studies was to test an alternate hypothesis to the putative K ϩ leak in explaining the ability of Tet proteins such as TetK to complement K ϩ uptake-deficient E. coli. The new hypothesis arises from the discovery that these Tet proteins are electrogenic monovalent cation/H ϩ antiporters, i.e. have a catalytic mode in which cytoplasmic Na ϩ or K ϩ is exchanged for a greater number of external H ϩ . If K ϩ were able to occupy the external H ϩ sites, then an exchange of cytoplasmic monovalent cation for a greater number of K ϩ could account for the net uptake of K ϩ (Fig. 1). Thus, the net uptake of K ϩ catalyzed by a Tet protein could be one of its normal catalytic modes. Experiments were designed to test this hypothesis with TetK and to examine whether TetA(L), to which this kind of activity had never been attributed, might nonetheless possess a comparable capacity. If the capacity to catalyze net K ϩ uptake was in fact a function of the monovalent cation antiport mode, then TetA(L) might well demonstrate it. Or, a capacity to catalyze net K ϩ uptake might be restricted to those Tet proteins with both monovalent cation/H ϩ exchange activity and a particularly high affinity for K ϩ . Both cloned TetA(L) and TetK were shown to restore Na ϩ exclusion capacity and resistance to a ⌬tetA(L) strain of B. subtilis. In such experiments, the Na ϩ exclusion capacity of TetK, but not of TetA(L), was markedly reduced by the presence of K ϩ . This suggested a higher K ϩ relative to Na ϩ affinity for TetK than for TetA(L) (15). The current studies support the hypothesis that net K ϩ uptake catalyzed by full-length forms of TetA(L) and TetK is a mode of the Na ϩ (K ϩ )/H ϩ exchange of both proteins.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Plasmids-The E. coli strains used in this study are listed in Table I. The bacteria were routinely grown in Luria broth (LB) with KCl substituted for NaCl (LBK) (23) and then supplemented with NaCl as indicated. The potassium transport-deficient E. coli strain TK2420 was grown on a defined medium (24) supplemented with various concentrations of KCl. The plasmid constructs containing full-length or truncated versions of tetA(L) or tetK were made by cloning polymerase chain reaction products into pGEM3Zf(ϩ) behind the T7 promoter (Table I). Expression of such constructs in E. coli cells, without concomitant expression of a T7 polymerase, results in levels of expression that are sufficient for phenotypic effects of membrane transport proteins without the toxicity that most often results from greater overexpression of such genes (13). The primers for the full-length version of tetA(L) were tetF1S (GGAGGGGGACATGCTGAATACGTCT-TATTCACAGTC) and tetR1B (TCACTCATGGGATCCATGTCCGC-GAACGTT). The primers for the truncated version of tetA(L) were tetF1S and tetR2B (GCAGCAAATAGGATCCATGGATATAATGAGC). The primers for the full-length version of tetK were tetKFB (CAAGTA-AAGAGGGATCCATGTTTAGTTTATAT) and tetKRB (AAATATAA-TATAAGGATCCAAACTGCTTTTCAG). The primers for the truncated version of tetK were tetKFB and tetKRB4 (GAAGTATAAGTAGGTAG-GATCCATGAATAT). The polymerase chain reaction products were blunt end ligated to pGEM3Zf(ϩ) that had been cut with SmaI. The hexahistidine-tagged full-length version of tetA(L), pJQ2, was made as described previously (16).
Complementation and Resistance Studies-The various constructs were tested for their ability to enhance the growth of potassium transport-deficient E. coli TK2420 on various concentrations of KCl in a defined medium. Two-ml cultures were grown in 15-ml conical tubes with shaking at 37°C. They were inoculated with 10 l of an 8-h culture (late logarithmic phase); the absorbance at 600 nm was recorded after 15 h. The concentration of KCl that allowed growth to an A 600 of at least 1.0 was defined as the minimum concentration that permitted growth. For complementation of E. coli strains NM81 or DH5␣ with various concentrations of NaCl, the bacteria containing various plasmid constructs were similarly tested on LB medium that was modified as indicated. The MIC was defined as that concentration of NaCl that stopped growth completely. The MIC of Tc was determined in a similar manner on cultures grown in LBK medium.
Assays of Active Transport in Everted Vesicles-Everted membrane vesicles were prepared in either 50 mM MOPS-KOH buffer or 50 mM Tris-HCl buffer, pH 7.5, as described previously (14). The transport of 50 M [ 3 H]Tc or 20 mM 22 Na ϩ was performed as described (14). Controls without the energy source, NADH, were always performed to assess energy-dependent uptake. Binding controls were conducted in the presence of 5% butanol.
Assays of Exchange in Right-side-out Vesicles-Right-side-out membrane vesicles were prepared by the method of Kaback (25) by shocking spheroplasts in 10 mM Tris-HEPES, pH 7.5, plus 2 mM MgSO 4 . Where indicated, the vesicles were passively loaded by incubation for 4 h at 20°C with various concentrations of NaCl, KCl, or choline-Cl. Exchange experiments were performed by diluting the vesicles 100-fold into buffer that contained a 100 M final concentration of 86 RbCl-KCl. To measure a ⌬⌿, positive out, vesicles were incubated with 2 M [ 3 H]tetraphenylphosphonium (26) and accumulation was measured by filtration onto OE67 filters (Schleicher & Schuell) that were then washed with buffer, dried, and counted by liquid scintillation. An internal vesicle volume of 2.2 l/mg of protein (27) was used to calculate the ⌬⌿ from the Nernst equation.
Fluorescence Based Assays of Transport-dependent Generation of an Electrical Potential-The ⌬⌿-dependent fluorescence of oxonol VI was used to measure the generation of a positive inside potential during exchange reactions. The assay mixture contained 10 mM Tris-HEPES, pH 7.5, plus 500 nM oxonol VI. The excitation wavelength was set at 580 nm and emission was at 631 nm (28). The change in fluorescence upon a 100-fold dilution of right-side-out membrane vesicles into the reaction tube was recorded. The final concentration of vesicles was usually 100 g/ml unless otherwise stated. The fluorescence changes were quantitated by setting up various gradients of K ϩ , outside greater than inside, adding 100 nM valinomycin, and recording the fluorescence change.

Complementation or Resistance Properties Conferred upon E.
coli Cells by Tet Constructs-The salient properties conferred by the constructs were first examined in whole cells of suitable E. coli strains. E. coli TK2420 (K ϩ uptake-deficient) cells transformed with full-length and truncated tetA(L) or tetK genes were examined for complementation, i.e. a reduction in the FIG. 1. A diagrammatic representation of the catalytic modes established for TetA(L) and TetK and two possible modes of Tet-dependent K ؉ uptake. The net uptake of K ϩ by certain Tet proteins might be a potential-dependent, electrogenic leak made possible by the presence of those particular proteins or it might be a catalytic mode of the electrogenic monovalent cation/H ϩ antiport in which K ϩ replaces the H ϩ . concentration of K ϩ required for growth relative to a vector control. The MIC for Tc was examined in wild type E. coli transformants and the MIC for NaCl was examined in both the wild type and in strain NM81 (Na ϩ -sensitive). The studies of Na ϩ sensitivity were conducted both in the presence and absence of added K ϩ . As shown in Table II, TetK strongly complemented the K ϩ -requiring phenotype of E. coli inasmuch as the concentration of K ϩ required to reach an A 600 of 1.0 after 15 h of growth was 1.8 mM as opposed to 29 mM in the control plasmid transformant of E. coli TK2420. TetA(L) also complemented significantly, albeit not as well as TetK, lowering the K ϩ concentration required to 6.5 mM. Truncated TetA(L) showed essentially no complementation while truncated TetK reduced the required K ϩ roughly by 50%. The full-length TetA(L) and TetK both strongly raised the MIC for Tc, whereas the truncated forms showed only modest positive impact upon resistance to the antibiotic. In both wild type and the NM81 strain of E. coli, only TetA(L) raised the MIC for Na ϩ significantly in LBK medium. In medium from which the added K ϩ was omitted, TetA(L), TetK, and the truncated TetK all showed some positive effect upon Na ϩ resistance.
It should be noted that in the pGEM vector used, and in the absence of the tet promoter region, the cloned genes were being expressed at low levels from the T7 promoter on the vector. It was not possible to use Western analyses to quantitate the amount of Tet protein in the different recombinant strains (or subsequent membranes therefrom) so that comparative effects may in part represent some difference in the ultimate amount of the particular protein that is actually found in the membrane. Under these conditions, none of the transformants showed a growth defect on LBK medium in the absence of high added Na ϩ concentrations as might be expected if they catalyzed a potential-dependent K ϩ leak. When the tet constructs were expressed more strongly from an inducible promoter in different plasmids, however, strong growth inhibition was observed, consistent with a generalized leakiness resulting from overexpression of a potentially toxic membrane protein (data not shown).
Energy-dependent Tc-cobalt/H ϩ and Na ϩ /H ϩ Antiport in Everted Vesicles of E. coli Expressing tet Genes-Energized antiport activities that had previously been shown for the fulllength TetA(L) and TetK proteins were assayed again for these proteins and their truncated forms. Tc-cobalt/H ϩ antiport was assayed in K-MOPS buffer. Na ϩ /H ϩ antiport was assayed in this buffer as well as in Tris-HCl buffer to test the earlier suggestion that TetK might not transport Na ϩ as well as TetA(L) in the presence of elevated K ϩ . Transformants of wild type E. coli were used as the starting material for vesicle preparations for these experiments. As shown in Fig. 2, TetA(L) and TetK both supported strong transport of Tc while neither truncated form did so. Similarly, as shown in Fig. 3, only the full-length forms catalyzed Na ϩ /H ϩ antiport. Although not shown, a hexahistidine-tagged version of TetA(L) expressed in pJQ2, that had been used in earlier reconstitution work (16), also exhibited these activities. The efficacy of TetA(L) was not affected by the presence of K ϩ , whereas the efficacy of TetK for Na ϩ transport was greatly diminished by added K ϩ consistent with the earlier indication that TetK has  a relatively higher preference for K ϩ over Na ϩ in its monovalent cation/H ϩ antiporter mode (15).
The electrogenicity of the Tc-cobalt/H ϩ antiport catalyzed by TetK had not earlier been examined and was thus investigated side-by-side with TetA(L). Everted vesicles, with K ϩ on both sides of the membrane and energized as in the previous set of experiments, were treated either with low concentrations of nigericin (0.1 M), valinomycin (1 M), or both. The stimulation is consistent with nigericin minimizing ⌬pH-dependent loss of accumulated Tc from the energized, everted vesicles. If there is any inhibition of the antiport itself by the nigericin treatment it must be even smaller than the stimulatory relief afforded by the dampening of that Tc loss. By contrast, addition of valinomycin inhibited Tc uptake almost as much as the combination of nigericin and valinomycin which should together abolish the electrochemical proton gradient. Two different modifications of this protocol were also examined. Experiments had been conducted on the electrogenicity of the Gram-negative TetA(B) Tc-metal/H ϩ in which nigericin was used at 2 g/ml (19). This would be a concentration of 2.5 M as compared with the much lower 0.1 M used in our experiments. Since aberrant effects of nigericin, i.e. electrogenic exchange (29,30), have been reported at unusually high ionophore concentration, such a concentration could be problematic. Although not shown, use of that concentration in experiments of the type depicted in Fig. 4, A and C, resulted in complete inhibition of transport. Also, high concentrations of chloride have been reported in assays using vesicle preparations for assessment of the capacity of an imposed diffusion potential to drive Tc-metal/H ϩ antiport and no energization of antiport was observed (1). Since chloride is known to reduce and, in sufficient concentration, abolish a transmembrane potential across E. coli vesicle membranes (31,32), it was possible that only an ineffectively small potential was actually generated in such experiments. As shown for the Tc-cobalt uptake mediated by TetA(L) and TetK (Fig. 4, B and D, respectively), use of buffers containing substantial added chloride for the ionophore experiments totally changed the inhibition pattern. Valinomycin no longer inhibited and nigericin (at the standard low concentration) inhibited completely, as expected if a ⌬⌿ did not exist and the ⌬pH was now the sole driving force. The two ionophores were similarly examined, in both low and high chloride conditions, for their effects upon Na ϩ /H ϩ antiport as monitored by Na ϩ uptake (Fig. 5). In contrast with the experiments shown for Tc-cobalt uptake in Fig. 4 and for the Na ϩ uptake experiments on TetA(L), assays of Na ϩ uptake mediated by TetK were conducted in Tris buffers under both the low and high chloride condition. The chloride concentration used in the higher chloride condition was only sufficient to partially abolish the ⌬⌿ in the respiring vesicles because the attendant reduced K ϩ concentrations minimized the inhibition of Na ϩ translocation by TetK. The patterns of inhibition for both TetA(L) and TetK were the same as observed with the Tc-metal/H ϩ antiport in the low chloride buffer, consistent with the electrogenicity of Na ϩ /H ϩ antiport mediated by these proteins. Use of higher chloride concentrations reduced the total apparent dependence upon the ⌬⌿, completely for TetA(L) and partially with TetK where lower chloride concentrations had been used. 86 Rb ϩ Uptake by Unenergized or Energized Right-side-out Vesicles-The monovalent cation/H ϩ antiport mode of TetA(L) and TetK both appeared to be electrogenic and the complementation of TK2420 suggested that they could catalyze net K ϩ uptake. Experiments were therefore developed to directly test the hypothesis that the net K ϩ uptake was another reflection of the monovalent cation/H ϩ exchange mode. Right-side-out membrane vesicles from E. coli TK2420 transformed with the same set of plasmids used in the experiments above were loaded with choline, KCl, or NaCl and were then diluted into Tris-HCl, MgCl 2 buffer containing 100 M 86 Rb ϩ -KCl, such that a 100-fold outwardly directed gradient of choline, KCl, or  Values for uptake in the absence of the energy source were subtracted. The higher background in these experiments as compared with those in Fig. 2 are attributable to the Na ϩ /H ϩ antiporter activity of the wild type (E. coli DH5␣) strain used in these experiments, a choice made because of the stability of the tet-bearing plasmids. E and q, pGEM vector; ⌬ and OE, pJG2-Tet A(L); Ⅺ and f, pJG3-Tet K; छ and ࡗ, pJG4-Tet KT; ƒ and , pJG5-Tet A(LT). Open, K ϩ ; closed, no K ϩ . NaCl was produced. The vesicles were not energized. Measurements of the transmembrane potential using tetraphenylphosphonium indicated that no potential developed at this concentration of chloride upon dilution without energization by an electron donor. Although not shown, none of the choline-loaded vesicle preparations exhibited Rb ϩ accumulation nor did the preparations from the transformant with truncated TetA(L), under any of the three conditions. As shown in Fig. 6, the preparations from the control transformant exhibited no Rb ϩ accumulation either. On the other hand, both the full-length TetA(L) and the TetK vesicles exhibited significant, transient Rb ϩ accumulation upon dilution of the NaCl or KCl-loaded vesicles, and the truncated TetK vesicles exhibited a smaller but reproducible level of Rb ϩ uptake in those conditions. Consistent with the absence of a potential, addition of SCN Ϫ (to a final concentration of 500 M) did not stimulate Rb ϩ uptake. Importantly carbonyl cyanide p-chlorophenylhydrazone (to a final concentration of 10 M) did not inhibit the uptake, consistent with the lack of involvement of protons in the exchange being measured.
It was important to assess whether the Rb ϩ uptake that was observed only upon dilution of K ϩ -or Na ϩ -loaded vesicles in the above experiments truly represented an electrogenic uptake. If so, a potential, positive-in, should be generated. Experiments with oxonol were conducted on K ϩ -versus cholineloaded vesicles to determine whether such a potential could be detected. The chloride content of the buffers was reduced by using 10 mM Tris-HEPES plus 2 mM KCl to load the vesicles so that any potential would not be immediately dissipated. Some chloride was retained because its complete elimination was found to impair the response of the probe. For each transformant type, K ϩ -or choline-loaded vesicles were diluted into buffer such that an outward gradient was generated (probe response designated F) or such that no gradient was generated (probe response designated F*). Control experiments showed that although less 86 Rb ϩ was accumulated under the gradientproducing conditions in this low chloride medium, its accumulation was easily demonstrable and remained dependent upon intravesicular K ϩ (data not shown). A lower level of accumulation was anticipated if a potential (positive in) is allowed to develop more fully during an electrogenic exchange. The F/F* ratios were calculated separately for the K ϩ -and cholineloaded vesicles. As shown in Table III, the response of the choline-loaded vesicles did not differ among the different transformant preparations. Moreover, every preparation elicited the same probe response whether or not a gradient was generated. By contrast, the K ϩ -loaded preparations from transformants with pJG2 and pJG3, encoding the two full-length Tet proteins, and pJG4, encoding the truncated TetK, elicited different probe responses upon generation of a gradient than its absence. That response, especially significant with the preparations in which full-length Tet proteins were expressed, was reflective of a significant potential, positive-inside, as calibrated by establishment of potentials of known magnitude. Preparations of plasmid control and pJG5 (truncated TetA(L)) preparations that were K ϩ -loaded both exhibited a F/F* ratio slightly below unity, perhaps reflecting a small potential, negative inside upon dilution.
Choline-loaded and KCl-loaded vesicles were then examined in a different protocol to test whether energized vesicles from the same transformants, without an outwardly directed gradient of cation, exhibited a Tet-mediated, ⌬⌿-dependent K ϩ leak. A leak was expected to be manifested as a potential-dependent (energization-dependent) but K ϩ -independent accumulation of Rb ϩ found in any of the preparations containing Tet constructs, but absent in the plasmid control preparations. It was also of interest to examine whether energization would stimulate Rb ϩ uptake by K ϩ -loaded vesicles as expected for an electrogenic antiport-dependent process. NaCl-loaded vesicles were not used because in preliminary experiments, the high level of FIG. 5. The effects of nigericin and/or valinomycin on Na ؉ uptake by everted vesicles from E. coli expressing either tetA(L) or tetK and assayed either in the presence of low or high chloride concentrations. Everted membrane vesicles of E. coli DH5␣ expressing tetA(L) were prepared as described for the comparable preparation used in Fig. 4, in either K-MOPS or Tris-HCl buffer (q). To some reaction mixtures, 1.0 M valinomycin (E) or 0.1 M nigericin (OE) or both (⌬) were added. The vesicles were preincubated for 1 min in the presence or absence of 2.5 mM K-NADH (for K-MOPS buffer) or 2.5 mM Tris-NADH (for Tris-KCl buffer). Uptake was initiated by adding either 10 mM Na 2 SO 4 (for no chloride conditions) or 20 mM NaCl (for high chloride conditions) plus 0.1 Ci of carrier-free 22 Na ϩ . The vesicles expressing tetK were prepared in either 50 mM Tris-HEPES, pH 7.5, or 50 mM Tris-HCl, pH 7.5, both containing 5 mM K 2 SO 4 . As described above for the tetA(L) vesicles, radioactive Na ϩ was added in the presence or absence of NADH to reaction mixtures containing no additions or various additions of ionophores. Values for the uptake in the absence of the energy source were subtracted from the values obtained in its presence. A, pJG2 in K-MOPS; B, pJG2 in Tris-KCl; C, pJG3 in Tris-HEPES; C, pJG3 in Tris-HCl.
FIG. 6. 86 Rb ؉ uptake by unenergized right-side-out membrane vesicles from E. coli expressing various tet constructs in response to outwardly directed gradients of Na ؉ or K ؉ . Membrane vesicles prepared in 10 mM Tris-HCl, 2 mM MgCl 2 , pH 7.5, were passively loaded with 10 mM NaCl or 10 mM KCl (open symbols) or 100 M of the same salt (closed symbols). Uptake was initiated by diluting 10 l of the vesicles into 1 ml of Tris-MgCl 2 buffer containing a final concentration of 100 M 86 Rb ϩ -KCl. The dilution buffer for samples in which an outward cation gradient was generated (open symbols) contained no other added salt, but the dilution buffer for those samples in which no outward gradient was to be generated (closed symbols) contained NaCl or KCl at the same concentration (100 M) as was present in the intravesicular space. At intervals, 200-l samples were taken, filtered (on 0.45-m HAWP filters (Millipore)), washed with 2 ml of Tris-MgCl 2 buffer, and dried for scintillation counting. electrogenic Na ϩ /H ϩ antiporter that was initiated upon energization complicated the experiment. As shown in Fig. 7, rightside-out vesicles from the vector control and truncated TetA(L) transformant exhibited a low level of Rb ϩ uptake whether energized with D-lactate or not and whether loaded with choline or K ϩ . This level corresponded to a level in which the intravesicular Rb ϩ had equilibrated with the outside concentration. All the unenergized vesicles of the remaining three preparations exhibited this same equilibration but no Rb ϩ accumulation in this protocol (with no outwardly-directed cation gradient). Most importantly, energized vesicles of both fulllength Tet transformants exhibited significant, sustained Rb ϩ accumulation upon energization when loaded with K ϩ but not when loaded with choline. This indicated that the energization was stimulating an exchange but that full-length TetA(L) and TetK did not provide a leak pathway for K ϩ -independent potential-dependent accumulation of Rb ϩ down its electrochemical potential. That the potential had been generated was confirmed by measurements via tetraphenylphosphonium accumulation which indicated that the ⌬⌿ was at least Ϫ125 mV in the various preparations. Truncated TetK vesicles, by contrast, exhibited the same smaller stimulation of Rb ϩ accumulation by D-lactate whether loaded with K ϩ or choline indicating that this truncated form of TetK might provide a modest leak pathway. DISCUSSION The studies conducted here confirm and extend earlier work indicating that both TetA(L) and TetK are multifunctional antiporters that catalyze electrogenic Tc-metal/H ϩ and Na ϩ /H ϩ antiport. The successful demonstration of both of these activities and their electrogenicity clearly depends upon low expression levels of the proteins. Higher levels of expression make both cells and membrane leaky. For experiments in which ionophores are used to assess electrogenicity in an E. coli vesicle system, it is further important to use ionophore concentrations that avoid aberrant exchanges and to reduce the concentration of chloride sufficiently to avoid dissipation of the ⌬⌿ by the permeant anion alone. The totality of earlier experiments supports the conclusion by Kaneko and co-workers (19) that TetA(B) catalyzes a largely electroneutral Tc-metal/H ϩ antiport. However, these authors themselves indicate some discrepancies in their findings with the conclusion of complete electroneutrality, and the issue might merit re-examination. As discussed below, the specific catalytic properties and possible multifunctional features are important factors in the design of strategies to minimize the interference of antibiotic efflux systems with use of antimicrobial therapies.
The truncated versions of TetA(L) and TetK failed to exhibit any of the energy-dependent, proton-coupled activities of the full-length proteins, consistent with the evidence that residues in the C-terminal halves of TetK cannot be mutated without loss of active Tc efflux capacity (6). Nonetheless, there were some modest but reproducible protective effects of the truncated Tet proteins in the whole cell growth complementation experiments (Table I). Possibly the truncated forms retain the capacity to bind Tc, Tc-metal, and monovalent cations, and this accounts for those effects. Such a basis for modest complementation in similar experiments has previously been noted (33).
The current studies add a catalytic mode to the repertoire of the Gram-positive Tet proteins, i.e. a mode in which net K ϩ uptake is achieved via a full catalytic cycle in which more than one K ϩ is taken up in exchange for a single cytoplasmic Na ϩ or K ϩ . Clearly, the full-length TetA(L) and TetK do not confer a leakiness upon E. coli membranes to K ϩ that allows electrogenic K ϩ entry (even down its chemical concentration gradient) in response to energization and establishment of a sizeable ⌬, inside-negative. The generation of a potential, inside-positive, during Na ϩ (K ϩ )/Rb ϩ exchange by unenergized vesicles is consistent with the operation of the whole catalytic antiport cycle but with the external Rb ϩ substituting for H ϩ . Were only a partial cycle to be used for the exchange, the Rb ϩ accumulation  a Right-side membrane vesicles loaded with 10 mM Tris-HEPES plus 2 mM KCl or 2 mM choline Cl were diluted 100-fold into 10 mM Tris-HEPES, pH 7.5. b F fluorescence intensity with a 100-fold K ϩ (or choline) gradient, in to out. F* is fluorescence intensity in the absence of a K ϩ (or choline) gradient, K ϩ (or choline) equal in and out.
c The ⌬, positive in, was quantitated by measuring the fluorescence intensity increase upon addition of 100 nM valinomycin in the presence of various concentrations of K ϩ out Ͼ K ϩ in. d 100 g of membrane vesicle protein was used in most assays, but 200 g of vesicle protein was used for pJG4 in order to better quantitate the signal. would represent counterflow entirely, i.e. with the intravesicular cation transported outward down its gradient, released, and then replaced on the outside with the external Rb ϩ without use of the "H ϩ " sites. In that case, the exchange should have been electroneutral. The occupation of a cation site by either K ϩ or H ϩ has similarly been proposed for the complete catalytic cycle of the eukaryotic serotonin transporter (34). It will be important to confirm the modest exchange capacity of the truncated TetK (as well as the possible leak) and the lack of a comparable activity by truncated TetA(L) in a purified reconstituted system in which the amount of transporter protein incorporated into the proteoliposome can be made comparable for different versions of the proteins. If Tet-mediated, electrogenic Rb ϩ (K ϩ ) uptake depends upon the use of the H ϩ -binding site and translocation pathway by these cations, and if the C-terminal part of the protein is required for proton binding and/or translocation, then even modest net Rb ϩ accumulation by truncated TetK is unanticipated under non-leaky conditions. The finding that net K ϩ uptake by full-length Tet proteins is definitely a mode of the normal catalytic functions rather than a leak, is consistent with the robust growth of cells expressing low levels of these proteins. It is notable that TetA(L) behaved qualitatively similar to TetK although it had not earlier been implicated as having the capacity for net K ϩ uptake. As hypothesized at the start of the study, this capacity may be a correlate of possession by a Tet protein of Na ϩ (K ϩ )/H ϩ antiporter activity and the extent to which this property occurs broadly among Tet proteins has not been carefully examined. Another question of interest in connection with the net K ϩ uptake mode is whether it may have a physiological role, e.g. at particular pH values and/or K ϩ concentrations. It will be of importance to examine the possibility that the Gram-negative TetA(C) (e.g. from pBR322 or pACYC184) might catalyze a similar spectrum of activities to that shown here for the Grampositive Tet proteins. TetA(C) is among the Tet proteins that can complement K ϩ uptake-deficient mutants of E. coli (8 -10). Moreover, this gene has been shown to have a beneficial effect on the "fitness" of adapted E. coli in the absence of antibiotic (35); this could reflect enhanced Na ϩ -resistance and K ϩ retrieval under some conditions. Whether TetK confers such a benefit on S. aureus will also be of interest to examine. Considerable current effort is directed toward reducing the prevalence and further spread of antibiotic-resistance genes among pathogenic bacteria or other organisms that might then transfer these genes to pathogens. In assessments of those conditions that will minimize positive selection for antibiotic efflux genes of particular types, e.g. tet genes, the full panoply of roles for the given efflux protein will be important information. For example, it might be important to consider the pH, Na ϩ , and K ϩ concentration to which the organisms are exposed, rather than simply the exposure to Tc, when evaluating strategies for decreasing the prevalence of TetA(L) or TetK.