Evidence for Na⁺ Influx via the NtpJ Protein of the KtrII K⁺ Uptake System in Enterococcus hirae

Abstract

The ntpJ gene, a cistron located at the tail end of the vacuolar-type Na⁺-ATPase (ntp) operon of Enterococcus hirae, encodes a transporter of the KtrII K⁺ uptake system. We found that K⁺ accumulation in the ntpJ-disrupted mutant JEM2 was markedly enhanced by addition of valinomycin at pH 10. Studies of the membrane potential (ΔΨ; inside negative) by 3,3′-dihexyloxacarbocyanine iodide fluorescence revealed that the ΔΨ was hyperpolarized at pH 10 in JEM2; the ΔΨ values of the parent strain ATCC 9790 and JEM2, estimated by determining the equilibrium distribution of K⁺ or Rb⁺ in the presence of valinomycin, were −118 and −160 mV, respectively. ΔΨ generation at pH 10 was accomplished by an electrogenic Na⁺ efflux via the Na⁺-ATPase, whose levels in the two strains were quite similar. Na⁺ uptake driven by an artificially imposed ΔΨ (inside negative) was missing in JEM2, suggesting that NtpJ mediates Na⁺ movement in addition to K⁺ movement. Finally, the growth of JEM2 arrested in K⁺-limited high-Na⁺ medium at pH 10 was restored by addition of valinomycin. These results suggest that NtpJ mediates electrogenic transport of K⁺ as well as Na⁺, that it likely mediates K⁺ and Na⁺ cotransport, and that Na⁺ movement via NtpJ is the major Na⁺ reentry pathway at high pH values.

All living cells show Na⁺ circulation across the cell membrane. This circulation is driven by active transport systems, which extrude Na⁺ and maintain the Na⁺ concentration gradient directed inward (30, 32, 34). In animal cells, the familiar Na⁺, K⁺-ATPase expels sodium ions, to which K⁺ uptake is tightly coupled. Bacteria have evolved diverse mechanisms for active sodium extrusion. Secondary Na⁺/H⁺ antiporters are widely distributed (31), and some bacteria have been found to have primary sodium pumps coupled with chemical reactions such as decarboxylation (5), electron transport (38), and ATP hydrolysis (8). Na⁺ reenters the cells via the Na⁺ gradient-consuming systems, with Na⁺-coupled secondary cotransporters being the most widespread route (41). The Na⁺ gradient is utilized for ATP synthesis and flagellar motion in some bacteria (5, 10).

The gram-positive bacterium Enterococcus hirae lacks a respiratory chain; the electrochemical concentration gradient of proton (proton potential) is generated by proton expulsion via the F_oF₁, H⁺-translocating ATPase (1). This bacterium has two sodium extrusion systems: the NapA Na⁺/H⁺ antiporter (11, 40) and a vacuolar-type Na⁺-translocating ATPase (16). The Na⁺-ATPase is encoded by the ntp operon, which consists of 11 ntp genes (ntpFIKECGABDHJ) (36). It is now clear that all nine subunits of the vacuolar Na⁺-ATPase complex are encoded by genes ntpF to ntpD (28); ntpH is tentatively considered not to be an open reading frame. Since the activity of the H⁺-ATPase is optimal around pH 6.5, the proton potential generated is significant at low pH values but is minimal at high pHs (13). Therefore, the Na⁺/H⁺ antiporter operates for Na⁺ extrusion only at low pH values. The Na⁺-ATPase is important for Na⁺ extrusion under conditions in which the proton potential is dissipated, such as at high pHs (13). There was no clear evidence of the presence of Na⁺ gradient-consuming systems in this bacterium. Therefore, it has been speculated that the physiological role of sodium extrusion systems may be the elimination of sodium ions from the cytoplasm to make room for K⁺ accumulation (6, 7).

Two K⁺ uptake systems, KtrI and KtrII, have been reported to exist in E. hirae. KtrI recognizes K⁺ as well as Rb⁺ with an apparent K_m of 0.2 mM, and it is likely to be constitutive. KtrI K⁺ uptake requires both generation of proton potential and ATP (or a related high-energy compound) (3); the activity of this system is optimal around pH 6 to 6.5 (22). KtrII selectively recognizes K⁺ with a K_m of 0.5 mM, and it has a pH optimum of around 9 to 10. KtrII K⁺ uptake was independent of the proton potential (12, 22). Although these K⁺ transport systems have not been well characterized at the molecular level, we recently found that the ntpJ gene, a cistron located at the tail end of the ntp operon, encodes a component of the KtrII K⁺ transport system (27); the KtrII K⁺ uptake activity was missing in an ntpJ-disrupted strain, and growth of this mutant strain in K⁺-limited alkaline medium was impaired. NtpJ is the membranous component of KtrII, resembling various K⁺ transporters such as Trk1p and Trk2p in the yeast Saccharomyces cerevisiae and the TrkG and TrkH subunits of the E. coli Trk system (36). Interestingly, expression of the ntp operon is regulated at the transcriptional level by changes in the intracellular Na⁺ concentration (26). Therefore, we assumed that KtrII is linked in some manner with the Na⁺ electrochemical gradient generated by the action of the Na⁺-ATPase (27).

In this study, we found that NtpJ mediates electrogenic translocation of Na⁺ as well as K⁺. The ΔΨ of E. hirae was generated by the action of the Na⁺-ATPase at high pH values, and it was hyperpolarized in an NtpJ mutant. NtpJ is the major Na⁺ reentry pathway of this bacterium at high pHs.

MATERIALS AND METHODS

Strains and growth conditions.

E. hirae strains used were ATCC 9790 (wild type), obtained from the American Type Culture Collection; a mutant, JEM2, in which the ntpJ gene was disrupted by insertion of an erythromycin resistance cassette (27); and a 9790-derived mutant, Nak1, defective in the Na⁺-ATPase (14). Cells were cultured at 37°C in a standard complex medium, NaTY (1% Bacto Tryptone, 0.5% Bacto Yeast Extract, 1% glucose, and 0.85% Na₂HPO₄) (12); the Na⁺ and K⁺ concentrations of this medium were 120 and 15 mM, respectively. In some experiments, mNaTY (21), a modified NaTY medium in which the concentration of yeast extract was reduced from 0.5% to 0.025%, was used; the K⁺ concentration of mNaTY was less than 1 mM. If necessary, Na₂CO₃ was added to these media to increase the pH. Erythromycin (10 μg/ml) was added to the media for culture of JEM2. The growth of cells was monitored by measuring the optical density at 540 nm and by counting viable-cell numbers.

Transport experiments.

For determination of net K⁺ uptake, cells were harvested at the mid-exponential phase and loaded with Na⁺, as described previously, using 2,4-dinitrophenol (3). Sodium-loaded cells were suspended in 0.1 M Na⁺-2-(cyclohexylamine)ethanesulfonic acid (CHES; pH 10) at a cell density equivalent to 1 mg of protein per ml. After incubation of the cell suspension with 10 mM glucose for 10 min, the reaction was initiated by addition of 2 mM KCl. At intervals, samples (0.3 ml) were obtained and filtered through Nuclepore polycarbonate membrane filters (pore size, 0.4 μm; Costar Scientific Co., Cambridge, Mass.), and the cells were washed twice with 2 mM MgSO₄. The potassium and sodium contents were determined by flame photometry after extraction of the cells with hot 5% trichloroacetic acid. For determination of Rb⁺ accumulation, the reaction was initiated by the addition of ⁸⁶RbCl (0.37 MBq/mmol) at various concentrations. The samples (0.3 ml) were filtered through cellulose acetate membrane filters (pore size, 0.45 μm; Toyo Roshi Co., Tokyo, Japan) and washed with the same buffer, and the radioactivity in the cells was measured with a liquid scintillation counter. For measurement of downhill Na⁺ uptake, potassium-loaded cells were prepared by the 2,4-dinitrophenol method (3) and suspended in 2 mM MgSO₄ at a density equivalent to 5 mg of protein/ml. The reaction was initiated by addition of 0.2 ml of cell suspension to 2 ml of a buffer composed of 200 mM Na⁺-CHES (pH 10), 180 mM N-methylglucamine-CHES containing 20 mM Na⁺-CHES (pH 10), or 200 mM N-methylglucamine-CHES containing 2 mM Na⁺-CHES (pH 10). For determination of ΔΨ (inside negative)-driven Na⁺ uptake, potassium-loaded cells were suspended in 50 mM N-methylglucamine-CHES (pH 10) containing 2 mM MgSO₄ at a density equivalent to 0.8 mg of protein per ml. Five minutes after addition of 2 mM NaCl, the reaction was initiated by adding 30 μM valinomycin. At intervals, samples (0.3 ml) of these Na⁺ uptake reaction mixtures were obtained and filtered through membrane filters, and the Na⁺ and K⁺ contents of the cells were determined by flame photometry.

Measurement of membrane potential.

Generation of membrane potential was monitored by fluorescence quenching of 3,3′-dihexyloxacarbocyanine iodide [DiOC₆(3)] as described elsewhere (2, 19). Cells were cultured in NaTY medium at pH 10, loaded with Na⁺, and incubated in 100 mM Na⁺-CHES (pH 10) containing 2 mM MgSO₄ and 1 μM DiOC₆(3) at a cell density equivalent to 0.5 mg of protein/ml for 10 min at 25°C. Fluorescence was monitored with a fluorescence spectrophotometer (model MPF-4; Hitachi Co.) at an excitation wavelength of 470 nm and an emission wavelength of 510 nm. The membrane potential (inside negative) was estimated by determining the equilibrium distribution of K⁺ or Rb⁺ in the presence of valinomycin (30 μM) as described elsewhere (18). The volumes of the cytoplasmic water space of E. hirae ATCC 9790 and JEM2 cells, determined with [¹⁴C]inulin (35), were 2.0 and 2.1 μl per mg of protein, respectively. These values were used for calculation of the intracellular concentration.

Miscellaneous methods.

The cellular contents of K⁺ and Na⁺ in growing cells were determined as described previously (20); cells in mid-exponential phase were collected on filters (pore size, 0.4 μm; Nuclepore) and washed twice with 2 mM MgSO₄. The cell membranes were prepared by a standard procedure as described previously (12) and, if necessary, stored frozen at −80°C. The Na⁺-stimulated ATPase activity of the membranes was determined at pH 8.5 in the presence of 0.2 mM N,N′-dicyclohexylcarbodiimide with or without 25 mM NaCl by a procedure described elsewhere (12). Denaturing polyacrylamide gel electrophoresis was carried out with the system of Laemmli, using 10% polyacrylamide (23). Western blotting was performed as described elsewhere (27), and proteins of interest were visualized with goat anti-rabbit immunoglobulin G conjugated to alkaline phosphatase. Protein levels were determined by the method of Lowry et al. (24) with bovine serum albumin as the standard.

Materials.

⁸⁶RbCl was purchased from NEN Life Science Products, Inc. DiOC₆(3) was purchased from Sigma-Aldrich Co. All reagents used were commercial products of analytical grade.

RESULTS

Effect of valinomycin on K⁺ uptake by E. hirae at pH 10.

Figure 1 shows net K⁺ uptake at pH 10 by Na⁺-loaded ATCC 9790 and JEM2 cells (ntpJ::Em^r); the assay was performed with 2 mM KCl. Glucose-dependent K⁺ accumulation, which is attributed to the activity of the KtrII K⁺ uptake system (12), was observed for ATCC 9790 (Fig. 1A). Accumulation of K⁺ or Rb⁺ in the presence of valinomycin at a low concentration is a measure of ΔΨ generation (18). K⁺ uptake by ATCC 9790 was strongly inhibited by the K⁺ (Rb⁺) ionophore valinomycin (Fig. 1A), supporting the previous suggestion that the KtrII system is not a K⁺ uniporter driven by membrane potential (13). K⁺ uptake was not observed for JEM2 (Fig. 1B). However, interestingly, we found that K⁺ was markedly accumulated by JEM2 in the presence of valinomycin (Fig. 1B). The amount of K⁺ accumulated by JEM2 in the presence of valinomycin was about 10-fold higher than that accumulated by ATCC 9790. The level of accumulation of ⁸⁶Rb⁺ by JEM2 in the presence of valinomycin was as high (data not shown). These results suggest that a high ΔΨ was generated in this ntpJ mutant.

FIG. 1.

Effects of valinomycin on K⁺ accumulation at pH 10. Strains ATCC 9790 (A) and JEM2 (B) were grown in NaTY medium (pH 10), loaded with Na⁺, and suspended in 0.1 M Na⁺-CHES buffer (pH 10) at a cell density equivalent to 1 mg of protein/ml. The suspension was (○) or was not (▵) supplemented with 10 mM glucose at 0 min. Valinomycin (30 μM) was added together with glucose at 0 min (●); K⁺ uptake was initiated by addition of 2 mM KCl at 10 min. The cellular K⁺ contents were determined by flame photometry.

The hyperpolarized membrane potential of the ntpJ mutant is generated by the Na⁺-ATPase.

ΔΨ generation by E. hirae was monitored by fluorescence quenching of DiOC₆(3). Na⁺ loading of ATCC 9790 and JEM2 cells cultured in NaTY medium (pH 10) was performed, and the cells were suspended in the same Na⁺ buffer as used in the experiment shown in Fig. 1 (Fig. 2). Fluorescence quenching of DiOC₆(3) equilibrated with ATCC 9790 cells was induced by addition of glucose. The quenching of the wild type was not affected by valinomycin but was suppressed by 10 mM KCl (Fig. 2A, left). Glucose-dependent quenching of fluorescence was more significant in JEM2 cells (Fig. 2B). The quenching of JEM2 was not affected by valinomycin either but was suppressed by the subsequent addition of 10 mM KCl (Fig. 2B, left). Since the cytoplasmic water space of 2.1 μl/mg of protein for JEM2 was nearly equivalent to that (2.0 μl/mg of protein) for ATCC 9790 cells, the ΔΨ generated for JEM2 was higher than that of ATCC 9790. The quenching of the wild type was suppressed quickly by 10 mM KCl only (Fig. 2A, right). However, the suppression by 10 mM KCl only was very slow for the quenching of JEM2 (Fig. 2A, right). Valinomycin, which alone did not affect quenching, rendered the added KCl much more effective (Fig. 2B, right), suggesting a role for NtpJ in electrogenic K⁺ movement. Glucose-dependent fluorescence quenching was not observed in an Na⁺-ATPase mutant, Nak1 (14) (Fig. 2C), suggesting that the ΔΨ is exclusively generated by electrogenic Na⁺ efflux via the action of the Na⁺-ATPase at high pH values in E. hirae.

FIG. 2.

Time course of changes in DiOC₆(3) fluorescence in E. hirae cells. Sodium-loaded cells of ATCC 9790 (A), JEM2 (B), and Nak1 (C) were incubated at a density of 0.5 mg/ml in 0.1 M Na⁺-CHES (pH 10) with 1 μM DiOC₆(3). Fluorescence quenching was initiated by addition of 10 mM glucose (glc) followed by addition of valinomycin (val; 30 μM) and KCl (10 mM). In the experiment shown in panel C, valinomycin and KCl were added simultaneously.

The magnitude of the ΔΨ generated in these cells was calculated by the equilibrium distribution of K⁺ or Rb⁺ in the presence of valinomycin (18). K⁺ accumulation in the presence of valinomycin was at various concentrations examined (Fig. 3). As shown in Fig. 2, the ΔΨ was eliminated by 10 mM KCl. Therefore, K⁺ accumulation at high KCl concentrations does not correspond to the size of the ΔΨ, probably reflecting the uptake for the internal charge compensation based on Donnan potential. The ΔΨ was estimated at less than 0.5 mM KCl, since the effect of K⁺ on ΔΨ generation was negligible (data not shown). The mean ΔΨ values ± standard deviations for ATCC 9790 and JEM2 cells, estimated by K⁺ accumulation, were −118 ± 1 and −160 ± 2 mV, and the values for ATCC 9790 and JEM2 cells determined by measuring Rb⁺ accumulation (data not shown) were about −115 and −165 mV, respectively. The potentials of ATCC 9790 and JEM2 were estimated to be −110 and −150 mV, respectively, based on the fluorescence intensity (2, 19). The ΔΨ of E. hirae at high pH values was thus clearly hyperpolarized by a defect in the NtpJ transporter.

FIG. 3.

K⁺ accumulation in the presence of valinomycin at various concentrations. Sodium-loaded cells of strains ATCC 9790 (A) and JEM2 (B) cultured in NaTY medium (pH 10) were suspended in 0.1 M Na⁺-CHES buffer (pH 10) at a cell density equivalent to 1 mg of protein/ml. The suspension was supplemented with 10 mM glucose and 30 μM valinomycin at −10 min; K⁺ accumulation was initiated by addition of KCl at various concentrations at 0 min. The established K⁺ concentration gradients are shown in parentheses. ○, 20 mM KCl; ●, 10 mM KCl; ▵, 5 mM KCl; ▴, 2 mM KCl; □, 1 mM KCl; ■, 0.5 mM KCl; ×, 0.2 mM KCl.

Na⁺ movement linked with the NtpJ transporter.

In all of the experiments described above, ΔΨ generation by the Na⁺-loaded cells in Na⁺ buffer was observed. Therefore, it is likely that ΔΨ hyperpolarization in JEM2 resulted from an increase in Na⁺ efflux and/or a reduction of Na⁺ reentry. The activities of Na⁺-stimulated ATPase of the membranes of ATCC 9790 and JEM2 cultured in NaTY medium (pH 10) were 0.14 and 0.14 μmol/min/mg of protein, respectively. Furthermore, Western blotting of the cell lysates of ATCC 9790 and JEM2 with anti-V₁-ATPase serum revealed that the amount of the ATPase in ATCC 9790 was nearly equivalent to that in JEM2 (Fig. 4). Alteration of the activity of an electrogenic Na⁺ extrusion system by Na⁺-ATPase was insignificant in JEM2. Instead, the permeability of the E. hirae cell membrane to Na⁺ may be altered by a lack of the NtpJ protein.

FIG. 4.

Western blotting of cell lysates after denaturing polyacrylamide gel electrophoresis. The cell lysates were prepared from strains ATCC 9790 (lanes 1 and 2) and JEM2 (lanes 3 and 4) grown in NaTY medium (pH 10) as described elsewhere (24). Lysates (5 μg, lanes 1 and 3; 10 μg, lanes 2 and 4) were electrophoresed, immunoblotted with antiserum against purified V₁-ATPase (dilution, 1:3,000), and visualized by the alkaline phosphatase system.

We have previously suggested that NtpJ mediates K⁺ translocation by the KtrII K⁺ uptake system (27), but it is not known if NtpJ-dependent K⁺ movement is linked with movement of other ions, such as Na⁺. Figure 5 shows Na⁺ movement in ATCC 9790 and JEM2 cells at pH 10. First, downhill Na⁺ influx at various external Na⁺ concentrations was measured in K⁺-loaded ATCC 9790 and JEM2 cells (Fig. 5A and B). In ATCC 9790 cells, Na⁺ influx rates were dependent on the external Na⁺ concentration (Fig. 5A); the Na⁺ influx rate at the external Na⁺ concentration of 200 mM was about 2 nmol/min/mg of protein. In JEM2 cells, Na⁺ influx also was dependent on the external Na⁺ concentration. The influx rates at individual Na⁺ concentrations for JEM2 were very similar to those for ATCC 9790 (Fig. 5B). It is unlikely that the downhill Na⁺ entry observed here contributed to the change in ΔΨ in the NtpJ mutant.

FIG. 5.

Movements of Na⁺ and K⁺ at pH 10. (A and B) Passive Na⁺ uptake. Sodium uptake was initiated by suspending potassium-loaded cells of strain ATCC 9790 (A) or JEM2 (B) into 200 mM Na⁺-CHES (pH 10) (●), 180 mM N-methylglucamine-CHES containing 20 mM Na⁺-CHES (pH 10) (▴), or 200 mM N-methylglucamine-CHES containing 2 mM Na⁺-CHES (pH 10) (■), respectively. (C and D) ΔΨ-driven Na⁺ uptake. Potassium-loaded cells of strain ATCC 9790 (C) or JEM2 (D) were suspended in 50 mM N-methylglucamine-CHES (pH 10) buffer containing 2 mM NaCl at −10 min, and the cellular Na⁺ (triangles) and K⁺ (circles) contents were monitored; 30 μM valinomycin was added at 0 min (closed symbols).

Uphill Na⁺ influx by ATCC 9790 and JEM2 was also examined (Fig. 5C and D). A ΔΨ (inside negative) across the cell membrane was imposed by addition of valinomycin to the K⁺-loaded cells suspended in N-methylglucamine buffer (at pH 10), which contributed about 0.3 mM contaminating K⁺. Glucose was omitted. ΔΨ-dependent Na⁺ influx, with a flux rate of about 35 nmol/min/mg of protein, which is much faster than the value for downhill Na⁺ uptake (Fig. 5A and B), was observed in ATCC 9790 (Fig. 5C). At steady state, an Na⁺ gradient of about 100 was established. However, Na⁺ uptake was missing in JEM2 even if a ΔΨ was imposed (Fig. 5D). In this case, instead of Na⁺ ions, protons were probably taken up into cells in exchange for K⁺ extrusion. It is thus clear that NtpJ participates in ΔΨ-dependent Na⁺ movement. Taken together, these observations suggested that ΔΨ hyperpolarization observed in the ntpJ mutant arose from a decrease in Na⁺ influx via the NtpJ K⁺ transporter.

Effect of valinomycin on the growth of E. hirae at pH 10.

Finally, the effects of valinomycin on the growth of ATCC 9790 and JEM2 in K⁺-limited (less than 1 mM K⁺), high-Na⁺ medium were examined (Fig. 6). Both strains grew well in this medium at pH 7.5. ATCC 9790 grew as well at pH 10 in this medium (Fig. 6A); the internal K⁺ and Na⁺ concentrations were 290 and 35 mM, respectively. The growth of ATCC 9790 at pH 10 was inhibited by valinomycin (Fig. 6A), as expected from its inhibitory effect on K⁺ uptake (Fig. 1A). The amount of internal K⁺ in ATCC 9790 in the presence of valinomycin was negligible. JEM2 did not grow at pH 10 (Fig. 6B); the internal K⁺ and Na⁺ concentrations were 20 and 280 mM, respectively. On the other hand, growth of JEM2 at pH 10 was evidently recovered by addition of valinomycin (Fig. 6B), in parallel with K⁺ accumulation stimulated by this ionophore (Fig. 1B); the internal K⁺ concentration of JEM2 under these conditions was about 125 mM. These results suggest that ΔΨ hyperpolarization induced a valinomycin-mediated accumulation of K⁺ that was large enough to allow growth of the ntpJ mutant.

FIG. 6.

Effects of valinomycin on the growth of E. hirae at pH 10. Strains ATCC 9790 (A) and JEM2 (B) were cultured in mNaTY (K⁺-limited NaTY) complex medium at pH 8 (○). At an optical density at 540 nm of 0.08, the medium pH was shifted to 10 by addition of 80 mM Na₂CO₃ (●). Valinomycin (30 μM) (▵) or KCl (20 mM) (▴) was added 2 min after the addition of the Na₂CO₃. The cell growth was monitored by cell density measurement.

DISCUSSION

We investigated Na⁺ and K⁺ movements in E. hirae at a high pH, and the results are briefly represented in Fig. 7. It has been reported that ΔΨ generation measured by tetraphenylphosphonium ion (TPP⁺) accumulation was clearly observed in a H⁺-ATPase mutant at a high pH (15). Furthermore, ΔΨ generation was dependent on the presence of internal Na⁺ at a high pH (15). In this study we showed that ΔΨ generation did not occur in an Na⁺-ATPase-defective mutant at a high pH (Fig. 2C). At a high pH, ΔΨ is thus mainly generated by an electrogenic Na⁺ extrusion via the action of the Na⁺-ATPase in E. hirae. It has been assumed that KtrII K⁺ uptake is in some manner linked with the Na⁺ gradient (27), since (i) NtpJ, a membranous component of the KtrII K⁺ transport system, is encoded as a cistron located at the tail end of the ntp operon (36); and (ii) this K⁺ uptake system functions together with a vacuolar-type Na⁺-ATPase at a high Na⁺ concentration and/or a high pH (20). We found that Na⁺-dependent ΔΨ generation at a high pH was enhanced in this mutant (Fig. 2 and 3) and that ΔΨ-driven Na⁺ movement was absent in an NtpJ mutant (Fig. 5). These results suggest that NtpJ participates in Na⁺ movement. We expected to detect alterations in the ATP hydrolytic activity of the Na⁺-ATPase of the membranes, influenced by its Na⁺ permeability. However, the activity of the vesicles of JEM2 was nearly equivalent to that of ATCC 9790 (data not shown); the membrane vesicles prepared here may be very leaky to Na⁺ ions. ΔΨ-driven Na⁺ influx in E. hirae has been previously examined (9), but its pathway has not been characterized. We inferred that ΔΨ-driven Na⁺ reentry, at least at a high pH, occurs via the KtrII (NtpJ) transporter. The lack of this transporter markedly influenced the magnitude of the ΔΨ at a steady-state level (Fig. 7B).

FIG. 7.

Sodium circulation and potassium accumulation in E. hirae at a high pH: an interpretation. (A) Wild type. Elements shown are the Na⁺-translocating vacuolar ATPase; the KtrII (NtpJ) (J) system for electrogenic accumulation of K⁺ as well as Na⁺ (cotransport), which presumably interacts with a KtrA-like component modulated by NAD(H); and a leaky pathway for Na⁺. (B) NtpJ mutant. Valinomycin (V)-mediated K⁺ accumulation was monitored. The low-affinity K⁺ transport system described in the text was omitted.

There is some debate over the validity of ΔΨ estimations obtained by indirect methods (18). It is possible that active K⁺ transport took place in ATCC 9790 under our experimental conditions as shown in Fig. 3A, even with adequate amounts of valinomycin, resulting in an increase in the internal K⁺ concentration. However, the value of −118 mV (Fig. 3A) was nearly equal to that (about −110 mV) estimated by measuring [³H]TPP⁺ accumulation (15), excluding this possibility.

The amino acid sequence of NtpJ closely resembles those of potassium transporters such as TrkG and TrkH of Escherichia coli, KtrB of Vibrio alginolyticus (29), Trk1p and Trk2p of S. cerevisiae, and HKT1 of Tricum aestivum (4). Among these K⁺ transporters, T. aestivum HKT1 has been suggested to be an Na⁺-linked K⁺ transporter (4). Site-directed mutagenesis of HKT1 indicated that the Tyr-463 and Glu-464 residues, which correspond to the Tyr-382 and Glu-383 residues of E. hirae NtpJ, are functionally indispensable for the transporter activity (33). In addition, the Asn-270 and Lys-362 residues of HKT1, important for recognition of Na⁺ and K⁺ ions (33), are also conserved in the sequence of the NtpJ protein. Recently, it was reported that the V. alginolyticus KtrB K⁺ transporter is Na⁺ dependent (37). Based on these results, we speculate that E. hirae NtpJ is also a K⁺-Na⁺ cotransporter (Fig. 7A). The ΔΨ-driven Na⁺ influx observed in a buffer containing 0.3 mM K⁺ (Fig. 5C) may reflect the cotransport of Na⁺ and K⁺, driven by ΔΨ, under our experimental conditions. It is also possible that Na⁺ entry is not coupled with the flux of K⁺, since it is unknown whether the NtpJ-mediated movements of Na⁺ and K⁺ are obligatorily coupled. We are now attempting to obtain direct evidence of K⁺-Na⁺ cotransport activity of NtpJ in the reconstituted proteoliposome system.

NtpJ is probably not a simple K⁺-Na⁺ cotransporter. A bacterial genome sequence database revealed that KtrII (NtpJ)-like systems are relatively widespread in bacteria (29). In V. alginolyticus, the ktrB gene forms an operon with the ktrA gene; the ktrA gene encodes a hydrophilic protein possessing a putative NAD binding domain (29). It has been speculated that the KtrAB system is a new two-component K⁺ transport system. As we have found a ktrA-like gene in E. hirae and Enterococcus faecalis (21), it is conceivable that the E. hirae KtrII system belongs to the KtrAB family. It is important to investigate the role of the ktrA-like gene product in NtpJ-dependent K⁺ transport activity (Fig. 7A).

We have no clear information regarding the pathways responsible for K⁺ and Na⁺ movements other than the Na⁺-ATPase and KtrII (NtpJ) at a high pH. Although Na⁺ extrusion by the action of the Na⁺-ATPase was normal in this NtpJ-defective mutant (27), this mutant did not grow in K⁺-limited, high-Na⁺ medium at pH 10 (Fig. 6B), in which the internal Na⁺ and K⁺ concentrations were 280 and 20 mM, respectively. Na⁺ reentry was faster than K⁺ uptake in mNaTY medium (21) when external Na⁺ and K⁺ concentrations were 280 and 1 mM, respectively. In other words, at a high pH, E. hirae has no high-affinity K⁺ uptake system other than KtrII (NtpJ). Since a high ΔΨ was generated in JEM2, the valinomycin-mediated K⁺ channel rendered ΔΨ-driven K⁺ accumulation much faster than Na⁺ reentry even at limited K⁺ concentration, reflecting the growth recovery of this mutant induced by valinomycin (Fig. 6B). On the other hand, it is noteworthy that the growth of JEM2 in K⁺-limited, high-Na⁺ medium at pH 10 was restored by 20 mM KCl even in the absence of valinomycin (Fig. 6B); the internal K⁺ level of this mutant was high in these media. We examined K⁺ uptake by JEM2 at a high pH (Fig. 1B) and found a low-affinity K⁺ transport system dependent on ΔΨ in this bacterium (M. Kawano, R. Abuki, K. Igarashi, and Y. Kakinuma, unpublished results); this system acts as a means of bypassing a defect in the KtrII system in order to achieve K⁺ homeostasis at a high pH.

The proton potential, or ΔΨ of fermentative bacteria is distinctly lower than that of respiring bacteria (17, 39); in E. hirae, the ΔΨ is maintained at a relatively low value at acid to alkaline pHs, especially in high-K⁺ medium (13). The difference presumably reflects the fact that aerobic cells extrude protons by redox reactions while anaerobic cells rely primarily on the hydrolysis of ATP by the F_oF₁-ATPase (17). On the other hand, we know that the magnitude of the potential at steady state is influenced by the permeability of the cell membrane to ions. In this study, we showed that the KtrII K⁺ transport system makes a substantial contribution to the size of ΔΨ at a high pH in glycolytic enterococci. ΔΨ hyperpolarization attributable to a defect in the K⁺ transport system was also reported in an S. cerevisiae Δtrk1 Δtrk2 mutant (25); the growth of this organism became highly sensitive to gentamicin. E. hirae grew very well when K⁺ homeostasis was ensured under conditions in which ΔΨ was hyperpolarized (Fig. 6). Strict control of ΔΨ is not always important for bacterial physiology.

This is the first report suggesting the presence of an Na⁺-coupled transport system in E. hirae. The physiological role of sodium extrusion in this bacterium is thus not simply to eliminate sodium ions from the cytoplasm and to make room for K⁺ accumulation but also to drive K⁺ uptake and perhaps also some other transport system(s) by cotransport. The ntp-encoded transport systems play a key role in the energetics of K⁺ and Na⁺ fluxes in enterococci at high pH values.